Talk:Journal of Morphology 2 (1888-89)

From Embryology






With the Cooperation of


Vol. II.







No. I. — July, 1888.


I. Frederick Tuckerman, M.D., Amherst, Mass. Observations on the Structure of the Gustatory Organs of the Bat {Vespertilio subulatus) . . . . . . 1-6

II. E. D. Cope.

On the Tritubercular Molar in Human

Dejitition . . . . . . 7-26

III. C. O. Whitman, Director of the Lake Labora tory, Milwaukee, Wis.

The Seat of Formative and Regenerative Energy ....... 27-49

IV. Prof, Henry Fairfield Osborn, Princeton Col lege.

A Contribution to the Internal Structure of

the AmpJiibian Brain .... 5 1-96

V. Dr. William Patten, Assistant in the Lake Laboratory, Milwaukee, Wis.

Studies on the Eyes of Arthropods . . 97-190

No. 2. — November, 1888.

I. Henry V. Wilson, Fellow of the Johns Hopkins University.

On the Development of Majiicina Areolata, 191-252

II. John M. Clarke.

The Structure and Development of the Visual Area in the Trilobite, Phacops Rana, Green ....... 253-270



III. R. W. Shufeldt, M.D., C.M.Z.S.

Further Studies on Grammicolepis Bra chiusculus, Poey 271-296

IV. E. D. Cope.

Oji the Relation of the Hyoid and Otic Elements of the Skeleton in the Batrachia 297-310

V. R. W. Shufeldt, M.D., C.M.Z.S.

On the Affinities of Aphriza Virgata . 311-340

No. 3. — April, 1889.

I. Charles-Sedgwick Minot, Harvard Medical School, Boston, Mass.

Uterus a7id Embryo : \. Rabbit ; 11. Man, 341-462

II. Edward Phelps Allis, Jr.

The Anatomy and Development of the Lateral Line Systern in Amia Calva. 463-568

III. Prof. A. E. Dolbear, College Hill, Mass.

On the Organization of Atoms and Molecules 569-585

IV. C. O. Whitman.

Some Neiv Facts about the Hirudinea . 586-599

V. Dr. William Patten.

Segmental Sense-Organs of Arthropods , 600-602

Volume IT. yuly, 1888. Number





GUSTATORY ORGANS OF THE BAT {Vesper tilio S2ib?ilatus).


Amherst, Mass.

The present paper contains a description of the anatomy of the taste organs of a single species of Chiroptera. It is highlyprobable that further study of these organs in other species of this interesting group of animals will reveal important variations, respecting both position and structural characters, from the results embraced in this short memoir.

It will be of interest first to notice briefly the form and general appearance of the tongue of this mammal.

General Description of the Tongue.

The organ measures 13.5 mm. in length, its greatest transverse diameter is 5 mm., and at its thickest part it measures 4 mm. Anteriorly, it is free from the floor of the mouth for 6 mm., or nearly half its length. The upper posterior surface is slightly convex, and has a nearly uniform breadth. In the anterior half of the organ the lateral margins gradually converge, blending at the tip in a slightly rounded or pointed extremity. The upper surface of this portion of the tongue is marked by several subparallel, transverse rugae or folds, with corresponding depressions between them. These folds decrease in size as they approach the anterior extremity of the organ, and cease altogether at 1.5 mm. from its apex. The dorsal surface is unmarked by any


median groove or raphe, except at the extreme posterior region. Here there is a wide and rather deep mesial groove, 2 mm. in length, beginning in front of the epiglottis and terminating midway between the two circumvallate papillae.

In one of my specimens, near the line of union of the posterior and middle third of the tongue, is a rounded eminence, showing a tendency to a raised posterior part, as seen in the tongue of the Rodentia. This feature, however, is wanting in other tongues of Vespcrtilio which I examined.

The upper surface, including the lateral margins, is covered with closely-set tactile and mechanical papillae, the points of which are directed backwards and inwards.

The fungiform papillae are only fairly numerous, and are distributed with some degree of uniformity over the dorsal surface and upon the sides of the tongue. Posteriorly they terminate in front of the gustatory area, and they cease anteriorly a short distance from the tip. Those scattered over the anterior third of the dorsum are usually larger than those occurring elsewhere.

On each lateral half of the tongue, 2 mm. from the base, is situated a circumvallate papilla. The two papillae are placed quite near the median line, the distance betv/een them being only 0.6 mm. They are oval in form, of nearly equal size, and are placed obliquely to the long axis of the tongue, their anterior extremity being directed outwards.

A papilla of similar type to those just mentioned, but less developed and apparently in a transitional stage, is present at the posterior limit of each lateral border. Further investigation will be necessary to determine whether these papillae are constant or not. In the specimens which I have examined I have always found them, although exhibiting striking variations in form, general appearance, and structure, from normal circumvallate papillae. No papilla foliata was found.

The under surface of the tongue is perfectly smooth. Anteriorly there is a median ridge, with sloping sides, extending from the frasnum to the tip.

Gustatory Structures.

TJie Circumvallate PapillcB. — These papillae show no indications of lobation. Their upper surfaces are rounded, and they measure 0.30 mm. in their transverse diameter, and are 0,22 mm.


in height. Where they join the tongue, the transverse diameter is only 0.12 mm. Each papilla is encircled by a rather shallow and very wide trench. In some sections this extreme width of the trench (as shown in Fig. i) is confined to its upper part, the lower portion curving beneath the papilla and becoming quite narrow. The ridge surrounding the trench, and forming its outer wall, has elongated tactile papillae projecting from its surface (Fig. i). The general surface adjoining this gustatory area is covered with large and small papillae, quite symmetrical in arrangement, but presenting a great variety of forms. Serous glands are fairly numerous in the gustatory area, but none were found within the papillary body itself. The ducts of the serous glands open into the trench at its base and sides. The papilla at its upper part bears many secondary papillae, the depressions between which are filled by the epithelium. The nerves are chiefly non-medullated and ramify throughout the papilla, but I was unable to trace their terminal branches with any distinctness. The large ganglion described by Poulton ^ in the circumvallate papilla of Peramelcs, and observed by me^ in the circum vallate papilla of Fiber, I failed to detect any indications of here. •

The two lateral circumvallate papillae are asymmetrical, the right one being much less developed than the left. The latter, as seen in vertical section, is elliptical in shape, and joins the tongue by a narrow pedicel. The trench which surrounds this papilla is very wide at its upper part, and narrow and of uniform breadth at its lov.'er. Serous glands are sparingly scattered through this region, and are entirely wanting within the papillary body.

The taste-bulbs are not very numerous in the circumvallate papilla of Vespertilio. They are disposed at the sides in a girdle of seven or eight tiers, the uppermost tier being nearly on a level with the top of the trench. From horizontal sections, made at different levels, I estimated the average number of bulbs in a tier at fifty. If we allow for eight tiers, we shall have four hundred bulbs for each papilla. I did not succeed in finding bulbs in the epithelium investing the upper surface of the papilla, nor was I able to detect them in the outer wall of

1 Quart. Journ. Micr. Sci., Vol. XXIII., 1S83, p. 73.

2 Journ. of Anat. and Physiol., Vol. XXII., 18S8, p. 136.

4 tuckerman: [Vol. ii.

the trench. The bulbs vary somewhat in shape, and in point of size they are the smallest I have yet observed, the nearest approach to them in this respect being those of the circumvallate papilla of the mouse. They vary in length from 0.025 to 0.030 mm., and their breadth is about 0.015 mm. Usually in a bulb of this area, the diameter of the peripheral end of the bulb exceeds that of the central. The latter is also curved slightly downwards. This form of bulb I have not observed before, although it is seen in those figured by Loven in the circumvallate papilla of the calf. I did not succeed in finding a bulb with the peripheral processes of its gustatory cells projecting beyond the pore. One bulb (in vertical section) shows a fissure, 0.015 mm. long and 0.0015 mm. wide, caused by a separation of the edges of two adjoining peripheral cells.

I was unable to isolate the central or gustatory cells of the bulbs sufficiently well for study, but the peripheral cells do not differ materially from those already described in the taste organs of other mammals. They are elongated, slightly flattened, nucleated cells, with their two extremities tapering gradually to a point.

The number of bulbs in the left lateral circumvallate papilla could not be estimated from my sections with any degree of accuracy. The most noteworthy thing about them in this region is their very unusual arrangement {Fig. 3). They occur only on one side of the papilla ; but here they form a continuous chain, seventeen tiers deep, extending from the base of the papilla nearly to its summit. Bulbs are likewise present in the lower half of the outer wall of the trench. Here I counted eight tiers. The bulbs of this region measure 0.024 ri'^"^- i^i length and 0.015 mm. in breadth, being thus a little smaller than those of the normal circumvallate papilla.

TJie Fungiform PapillcB. — These papillee are distributed quite regularly over the dorsum and sides of the tongue, from the gustatory area nearly to the tip. Interspersed among those of the posterior part of the dorsum are a few which appear to be undergoing transition to the circumvallate type of papilla (Fig. 6).

In several instances taste-bulbs were present in the epithelium at the upper part of these papillae. They are usually placed vertically, directly in the long axis of the papilla. By


far the most interesting specimens (which are shown in Fig. 6) were found in a papilla from the posterior region of the tongue. In this papilla there are two well-formed bulbs, and placed between them is a third, which is either of a low order or undeveloped.^ The largest bulb of the three measures 0.036 mm. in length and 0.016 mm. in breadth, and its apex appears to reach the free surface of the epithelium, its base penetrating the mucosa. Some of the isolated bulbs met with elsewhere in these papillae, particularly those of the anterior dorsal surface, are even larger than those shown in Fig. 6. Neither serous nor mucous glands were observed near the fungiform papillae.

The entire upper surface of the tongue is covered with papillae of mechanical and tactile ij) function. They are quite closely set, except at the basal region, are largest at the posterior part of the dorsum, and gradually decrease in size as they approach the anterior extremity. These papillae, when near the tip, enlarge slightly again. One from the posterior third of the tongue measured o. 11 mm. in height and 0.04 mm. in breadth. Behind the circumvallate papillae, and also about the tip, are numerous rather coarse, retroverted, conical papillae. Each papilla is seated upon a single papillary upgrowth of the mucous membrane, and is invested by epithelium of a uniform thickness. The outer layers of epithelium covering the upper surface and sides are usually partly, and occasionally wholly, cornified. These papillae vary much in shape and general appearance. Many of them are cone-shaped, while others resemble, in external structure, minute fungiform papillae. The upper surface is now and then flat or slightly convex, but usually the papilla terminates in a retroverted, horny sjoinule.

1 Figure 7 represents a vertical section through a fungiform papilla, from the anterior dorsal surface of the tongue of a pig, containing eight taste-bulbs.


EXPLANATION OF PLATE I. List of Reference Letters.

c. e. Columnar epithelium, g. p. Gustatory pore. gl. d. Duct of serous gland. in. in. Mucous membrane, p. Elongated papilla. /. e. Pavement epithelium. /. p. Papillary processes of mucous membrane, r. Ridge, s. e. Stratified epithelium. s. p. Secondary papillae, t. Trench, t. b. Taste-bulb. i. b'. Taste-bulb of outer wall of trench, t. b." Taste-bulb undeveloped or of low type.

Fig. I. — Vertical section through one of the circumvallate papillos. (x lOO diam.)

Fig. 2. — Vertical section through the base of the same papilla, showing overhanging side with four lowermost tiers of taste-bulbs. (X 480 diam.)

Fig. 3. — Vertical section through the left lateral circumvallate papilla. (X 125 diam.)

Fig. 4. — Horizontal section through the lower part of one of the circumvallate papillae, representing the taste-bulbs arranged in a zone. gl. d. Ducts of the serous glands which open into the trench at this level. (X 160 diam.)

Fig. 5. — Vertical section through a fungiform papilla of the mid-dorsal surface of the tongue, showing a single taste-bulb at its upper part, (x 400 diam.)

Fig. 6. — Vertical section through a fungiform papilla of the posterior dorsal surface of the tongue, which is probably undei-going transition to the circumvallate type. (X 480 diam.)

Fig. 7. — Vertical section through a fungiform papilla of the anterior lateral dorsal surface of the tongue of a pig. (X 200 diam.)

Jour/i Moi/ih Vol. II.

gti. ,4


\ /

'> )





Descriptions of the molar teeth of man, given by anatomists, differ in important respects. Thus, F. Cuvier (" Dents des Mammifers ") states that, while the crown of the first superior true molar consists of four tubercles, those of the second and third superior true molars consist of but three tubercles. In the American edition of " Sharpey and Ouain's Anatomy " it is stated that the crowns of the superior true molars of man consist of four tubercles ; and the same statement is made in Allen's late work on human anatomy.

My observations having shown me that both of these descriptions apply correctly to certain types of dentition, I determined to examine for myself, to ascertain, if possible, the extent and value of the variations thus indicated. My interest in the subject had been especially stimulated by the researches among the extinct mammalia, and the results which I had derived from them. These are, in brief, as follows : first, the quadritubercular type of molar crown, illustrated by the first superior true molar of man, belongs to the primitive form from which all the crest-crowned (lophodont) molars of the hoofed placental mammals have been derived ; and second, this quadritubercular type of molar has itself been derived from a still earlier, tritubercular crown, by the addition of a cusp at the posterior internal part of it. This tritubercular molar in the upper series has given origin directly to the superior sectorial teeth of the creodonta and carnivora. In the inferior series, I have shown that in known placental mammalia at least, the primitive molar crown is quinquetubercular, or tritubercular with a posterior heel ; that this form gave origin to the inferior sectorial tooth of carnivora by modification, and to the quadritubercular type — corresponding to the superior quadritubercular crown — by a loss of the anterior inner cusp and

8 COPE. [Vol. II.

connecting crest. And from the quinque- and quadritubercular types of molar crown, the various specialized types of the ungulates have been derived.

Considerable significance, therefore, attaches to the question as to whether the superior true molars of Homo sapiens are quadritubercular or tritubercular. The inferior molars are also either quadritubercular or quinquetubercular ; but less significance attaches to this modification than to that of the superior true molars. This is owing to two facts; viz., the fifth tubercle is not the anterior inner which completes the anterior triangle of the primitive inferior molar, but is a median posterior, such as is not uncommon in mammalia of Puerco and Eocene age ; and second, because this tubercle is of quite small size, and is therefore more liable to variation from insignificant causes.

In the nearest allies of man, the anthropoid apes, the superior true molars are quadritubercular, the posterior internal tubercle of the last or third molar being usually smaller than in the other molars in the chimpanzee. The inferior molars are quinquetubercular, in the human sense, the gorilla not infrequently adding a sixth lobe on the external posterior margin of the crown. The molars of both series are quadritubercular, with an occasional posterior fifth in the inferior molars in the Cercopithecidae and Cebidae, excepting the genus Pithecia of the latter, where the superior molars are tritubercular. The superior molars of the Hapalidse are tritubercular. In the Lemuridae the second and third, and frequently the first, superior true molars are tritubercular. In the Tarsiidse the superior true molars are tritubercular throughout. The superior molars of the extinct lemuroids differ, like those of the recent forms. Thus, in Adapis and its allies they are quadritubercular, but in Necrolemur they are tritubercular. In Chriacus (whose reference to the Lemuroidea is uncertain) they are tritubercular, as is the case, also, with Indrodon. In Anaptomophus they are of the true tritubercular type. This is the genus of Lemuroidea, which in its dental character most nearly approaches the anthropoid apes and man. I have elsewhere^ pointed out that the formula is I. -; C. -: Pm. -: M. -. The

'■ 2 ' I ' 2 ' 3

canines are small, and there is no diastema in either jaw.

1 Report U. S. Geol. Survey, Terrs., F. V. Hayden, Vol. III., 18S5, p. 245, Fl. XXV. fig. 10, and PI. xxiv^. fig. i.

No. I.]


It may be readily seen, in consideration of these facts, that the appearance of tritubercular superior molars in the genus Homo constitutes a reversion to the lemurs, and not to the anthropoid apes or to the monkeys proper. And among lemurs the reversion is most probably to that type which presents the closest resemblance to Homo in other parts of the dentition. The genus which answers most nearly to this requirement among those at present known, is Anaptomorphus.

Figure i. — Two species of Anaptomorphus. Fig. a, A. ainidiis, lower jaw from above X 2. Figs, b, c, d, A. hoinuiuulus ; b, c, natural size; d, ^ natural size. Both from Eocenes of the Rocky Mountains.

In Studying the dentition of man, I have examined the crania contained in the following six collections : those of the Academy of Natural Sciences of Philadelphia ; of the Army Medical Museum of Washington ; of the College of Physicians of Philadelphia; of the University of Pennsylvania; of the Boston Society of Natural History; and of my own museum. The first of these is especially valuable on account of the negro, Egyptian, and Hindu crania it contains. My acknowledgments are due to the Board of Curators, of which Professor Leidy is chairman, for the opportunity of studying it. I am also indebted to the Boston Society of Natural History, and its learned curator. Professor

lO COPE. [Vol. II.

Hyatt, for the opportunity of examining Hindu and Chinese crania in their museum. The collection of the Army Medical Museum at Washington is especially important on account of the Kanakas, Esquimaux, Peruvians, and North American Indians which it possesses. I am under great obligations to its distinguished director. Dr. J. S. Billings, for the facilities which he placed at my disposal. The museum of the Philadelphia College of Physicians contains the collection made by the late Professor Hyrtl of Vienna, of crania of Eastern and Mediterranean Europeans. In this department it is unrivalled, and I am greatly indebted to the council of the college, and its curator, Dr. Guy Hinsdale, for the opportunity of examining it. Some French skulls in the University of Pennsylvania were of value in the investigation. My own collection, though small, contains a number of Maoris, Australians, Tahitians,^ and North American Indians, which have proved to be of importance. Of English and Europeo-American crania, I have been able to examine but few of what might be termed the thoroughly amalgamated race. Of the latter there are probably many crania in the war collection of the Army Medical Museum, but how free the race of each may be from foreign intermixture, of course it is impossible to know. In selecting such as are supposed to be " stock Americans," those of persons with English names have been preferred, although many now true Americans are of German ancestry. In order to increase the list of this class of examinations, I have imposed on the forbearance of my friends by frequent inspections of their dentitions in ore aperto.

I suspect that the characters thus obtained will prove of importance in a zoological and ethnological sense. They have been already found to be of great fixity, and hence significance, in the lower mammalia. The only reason why they should be less so in man is, that the modification in reverting to the tritubercular molar is a process of degeneracy, and may be hence supposed to be less regular in its action than was the opposite process of building up, or addition of the posterior internal cusp. Some justification for a light estimate of its value may

1 For the Maori and Australian skulls, I am indebted to Mr. Speechey Gotch of Melbourne ; and for the Tahitians to Dr. Chassaniol, Chef de la departement de Sante de Taiti, of Paris. My best acknowledgments to these gentlemen are hereby expressed.


be found in the following tables. But it must be remembered that it is not always possible to determine exactly the race of the person represented by a skull, even when care in its identification has been exercised. Emigration and war have constantly rendered races impure, and transplantation on a large scale has in some parts of the earth produced hybrid races. The results of a study of human crania are sure to be more or less vitiated by these circumstances. We obtain averages rather than exact definitions. Nevertheless, the extremes of the series of variations are likely to be found to be characteristic of established forms of man, and will thus justify my belief in the value of the characters presented. To ascertain the relation of these variations to the races is the object of the present inquiry.

The cause of the tritubercular reversion belongs to the class of agencies active in evolution of organic types, of whose real nature we know little. It cannot be said to be due to a contraction of the maxillary arcade, for the Esquimaux and some other peoples which display the tritubercular dentition are not deficient in this respect. Nor do tritubercular molars require less space than the quadritubercular, for the external width of the crown is the same in both cases. They generally require less material however than a quadritubercular crown, since a triangle is smaller than a square drawn on the same base line ; however, in some men of the lower races who present the tritubercular molars, their outline is nearly square. The hypothesis advanced to account for the reduction of the number and quality of human teeth observed in the higher races, as well as for the replacement of the prognathous jaw by the orthognathous, is that such changes are due to a transference of material and of growth energy from these parts to the superior part of the skull and its contents. The relative superiority of the dimensions of these parts in the higher races is thus accounted for.

In the following tables the tubercular formulas are represented by numbers. Only the last three, or the true molars, in each jaw are considered. Tubercles of reduced size are represented

by fractions. Thus ^~^~'! indicates that each superior molar is quadritubercular, and each inferior molar quinquetubercular. This represents the extreme of the series represented by the

lowest races. The formula \~^~^ indicates that the true molars

4 — 4—4


COPE. [Vol. II.

have four, three, and three, tubercles respectively, and that the inferior true molars have four each. This represents the extreme common among the higher races. In the table which follows, the numbers attached to crania in the respective collections are appended, and initials indicating the collection follow. Thus A. M. M. refers to the Army Medical Museum, Washington ; C. P., the College of Physicians of Philadelphia ; U. P., the University of Pennsylvania ; B. S., the Boston Society of Natural History; E. D. C, my private collection; no initials follow the numbers of the Academy of Natural Sciences of Philadelphia.


5-5-5 Malay^ of Madura, *i339 ; Malay, 425 ; Negroes, d^ and 258, C. P. ; S. Sea Islander, 86, A. M. M.

4-4-4 5 — 4-4

Malays, 47 and 433 (latter of Sumbawa) ; Negroid Egyptians, 43, 798, 852*, 869.

4-4-4 4-4-4 Malay, 1338 (of Amboyna).

4-4-4 ? Tahitian, E. D. C. ; Kanakas, 143 and 1308, C. P.; Malay, 425 (Java) ; Negro, 963 ; Hindu, 1070, B. S.

4 - 4 - 3i 5-5-5 Marquesas Ids., 1531, C. P. ; Greek of the Morea, 55, C. P.

Peruvian, 932 (Arica).

4-4-37 5-5-4

4 - 4 - 3i

5-4-4 Malay, 430 (Amboyna) ; Italian, 114, C. P. (Elba) ; Peruvian, 2300, A. M. M.

1 Just what is meant by the "Malays" of the Philadelphia Academy collection I do not know.


4 - 4 - 3i

5-4-5 Negro, 4122, C. P.; Kaffir, 1358; Pessahs, io95-*7 ; Gipsy, 31, C. P. ; Montenegrin, 29, C. P. ; Tablunka, 106, C. P. ; Germans, 1063-4 (Tubingen) ; Feejees, 292-3, A. M. M.

4-4-31? Australian, E. D. C; Tahitian, E. D. C; Negroes, 901, 903, 914, 920, 921, 927, 964; American Negro, 9S0, A. M. M. ; Chatliam Id., 1557, A. M. M.; Peruvians, 68, 452; 2 N. Amer. Indians, E. D. C.; Ponka India, 487, A. M. M. ; Comanche, 6563, A. M. M. ; Sicilian, no, C. P.; Esquimaux, 1859, A.M. M. ; French, 1579, A. M. M. ; Anglo-Americans (Wilderness Battle-Field, Va.), 6305-6-7, A. M. M. ; Kanakas, 286, 434, and 584, A. M. M. ; Hottentot, B. S.



Madagascar, two (Hovas) ; Samoan Id. ; N. Amer. Indians, 1345

(Lipan), and Hudson's Bay Ind. 3 Mexican, 714 (Ancient) ; Peruvians,

72, 1373, 1465 ; Italian, 118, C. P.; Lombard, 117, C. P.; Czech, 131,

C. P. ; German, 239, C. P.

4-4-3 5-5-4 N. Amer. Indians, 530 and 1804 (Pawnees), A. M. M.

4-4-3 5-4-5 Italians, 2 (Giurgevo and Piedmont), C. P.; Gipsy, 18, C. P.; Ural, 33, C. P. ; Bulgar, 94, C. P. ; Slav, 84, C. P.


5-4-4 N. Amer. Indian (Spokane), E. D. C. j Italians, 46 and 47 (Terracina and Ragusa), C. P. ; Albanian, 123, C. P. ; Viennese, 9 59, C. P.



Negroid Egyptian, 885 ; Swede, 1549.

4-4-3 ?

2 Tahitians, E. D. C. ; Kanakas, 291, 451, 455, 570, 588, 844, 1057,

A. M. M. ; Chatham Id., 5721, A. M.M. ; Esquimaux, 1232, A. M. M. ;

N. Amer. Indian, 380 (Chippewa), A. M. M. ; ? Anglo-Americans, 5721,









14 CC/'^. [Vol. II.

3257, 6785, A. M. M. ; Jew, 3, C. P.; Rouman, 96, C. P.; Chinese, 873, B. S.; Hindu, 137, B. S.

4 - 4- ? ?

Kanakas, 79, 280, 283, 421, 423, 427, 431, 457, 463, 599, 709, 862,

A. M. M. ; Chinese, 958, 5130, 5139, A. M. M. ; Peruvian, 1765 (An con), A. M. M. ; Peruvian, 2302, A. M. M. ; S. Sea Islander, 344,

A. M. M. ; Indian of Yucatan, 629, A. M. M. ; Mojave, 209, A. M. M. ;

Hindu, 123, B. S.

Circassian, 762.

5-5-5 Negro, 1343 (Archipelago) ; American Negro; S. German, 1289.

4 - 3i - 3i 5-4-5 ■ N. Amer. Indian, 1794 (Ute), A. M. M. ; Hindu, *i344, Bengal; Gipsy, 98, C. P. ; S. German, 1158; Italians, 2, 8, and 109 (Tessino and Roveredo), C. P.; Greek, 53 (Cephalonia), C. P.; Bosniak, 100, C. P. ; Cossack of Don, loi, C. P.

4 - 3i - 3i 5-4-4 Patagonian, 1232 (last infer, mol. 4^) ; Negro lunatic, 55.

4 - 3? - 3i ?

African Negroes, 580, 685, 902, 917; Kanaka, 861, A. M. M. ; American Negro, 411, A. M. M. ; Peruvian, 1326; 2 N. Amer. Indians, E. D. C; Mound Builders, 1049-1123, A. M. M. ; Piegan, 6486, A. M. M. ; Cheyenne, 5560, A. M. M. ; ? Anglo-Americans, 2, A. M. M. ; Esquimaux, 1182, 1194, 1250 ; Hollander, 9 10, C. P. j Hindu, 124, B. S.

4 - 3i - 3

5-5-5 Malay, 1340 (Macassar), 1341 (Java) ; Finn, 1539 ; Czech, 79, C. P. ; Viennese, 61, C. P. ; French, U. P.


5-4-5 Algerine, 3900, C. P. ; Uskoke of Banjaluka, 90, C. P.; Styrian, 85, C. P.; Hindu, 432; Latin, 115, C. P.; Albanian, 97, C. P.; German of Siebenbiirgen, 76, C. P.

4 - 3i

- 3

4 - 4

■ 4

4 - 3i



4-37-3 5-4-4 Hungarian, 14, C. P. ; Russian, 35, C. P. ; Australian, (1327) ; Negroes (Benguela), 421, (Mozambique), 423; Hottentot, 1351; Aztec; Circassian; Moravia, 9 135, C. P.; Croat, 5134, C. P.; Galician, 139, C. P. ; Serb, 96, C. P. ; " Pizgau," 66, C. P. ; Saltzburg, 65, C. P.

Swedes, 1487-50.

? ? ? Negro (Dey), iioo ; Peruvian ; Kanaka, 576, A. M. M. ; Dutch, 434 ; Malay, 39; Bengal, 20, A. N. S. ; Hindu, 1330; Fellah, 999; Kabardine Caucasus, 1%, C. P. ; Ital. Piedmont, 45, C. P. ; N. Amer. Indians (Flathead) , E. D. C, (Cheyenne) , 65 25, A. M. M., (Ponka) , 486, A. M. M., (Sioux), 9 2049, A. M. M. ; Mound Builders, 169, A. M. M. ; Esquimaux, 1183, 1189, 1245, A. M. M. ; ? Anglo-Americans, 6847, 5922, A. M. M. ; German, 1066 ; Chinese, 879, B. S.

4 - 3i - ? ? Kanakas, 564, 569, 571, 591, A. M. M. ; Peruvian, 2308, A. M. M. ; Arucanian, 970, A. M. M. ; Chuktchi, 277, A. M. M. ; Esquimaux, 1779, A. M. M.

4 - 3 - 3i ?

Chukchi, 263, A. M. M.

4 - 3¥ - 3i ? ? German- American, 6445, A. M. M.


5-5-5 Dalmatian, 132, C. P. • Greek (Trieste), 56, C. P.; Indians (Santa

Barbara, Cal.), 3.

4-3-3 5-5-4 Gipsy, 130, C. P.; Trieste, 120, C. P.; Apache, 1168, A. M. M.


5 - 4i - 4i Modoc, E. D. C.

l6 COPE. [Vol. II.

4-3-3 5-4-4 Druse, 122, C. P.; Idria, 89, C. P.; Camiola, 87, C. P.; S. Tyrol, 121, C. P.; Paris, 3916, U. P. ; 4 Anglo-Americans, E. D. C; French, 867 (Paris), U. P., 1620, A. M. M. Kurd, 28, C. P.; Thug Hindu, 128 (A. N. S.) ; Esquimaux, 676 (trace of 5 on 2 and 3) ; Italian (Calabria), 48, C. P.; Greeks (Candia), 54 and 51, C. P.; Slovak, 73, C. P.; Czech, 4, C. P.; Lithuania, 9 21, C. P.; Bukovina, 9 78, C. P. ; Tyrol, 9 67, C. P. ; Upp. Austria, (>i, C. P. ; Moravia, 5, C. P. ; Hollander, 5 9, C. P.


5-4-5 " Krakuse " (Czech or Pole, large), 82, C. P. ; " Kumanie " ( ? Hungary), 71, C. P.; Pole, 24, C. P.; Volhynian, 36, C. P.; Podolian, 37, C. P.; Hindu (Bengal), 1344; Tatra (large), 83, C. P.; Wallach, 95, C. P. ; French, 105, U. P.


4-4-4 Tchuktchi; Zagrebin, 126, C. P.; Croat, 77, C. P.; Mixed German and Croat, 17, C. P.; French, 1082; do., 1801, A. M. M. ; AngloAmerican, 1 108.



Peruvian, 1478; do., 2299, A. M. M. ; Turkish grave, 127, C. P.; Crimean, 34, C. P.; Cossack, 133; Ruthenian, 103, C. P.; Armenian, 99, C. P. j Esquimaux, 675, 679, A. N. S. ; Esquimaux, 1211, 1218, 1219, 1206, 1221, 1225, 1226, 1230, 1231, 1241, 1244, 1251, 1278, 1781, 1782, 2113, A. M. M. ; Chuktchis, 256-7, A. M. M. ; N. Amer. Indians, 5550 (Pawnee), 6561, 6565 (Comanches), 7023 (Cheyenne), 170, (Mound Builder), A. M. M. ; Kanakas, 580, 841, A. M. M. ; Indians of Yucatan, 626-628, A. M. M. ; American Negroes, 981-983 (Richmond, Va.), 6615 (S. Carolina; m. 2 on one side is 3^) ; 6 AngloAmericans, E. D. C. ; ? Anglo-Americans, 95, 6305-6-7-8,, 6640, 739, 2699, 1768, 5 1 16, 5145, 57089, 5940, 6550, A. M. M.; Hindu, 425, B. S.


4-4-5 Hungarian, 15, C. P.


Neolithic Man.

Through the kindness of Mr. Thomas Wilson, honorary Curator of the Department of Prehistoric Anthropology in the United States National Museum, I have had the opportunity of examining a number of dentitions of men of the Neolithic period of France. The best preserved of these includes both series in place in both jaws, the last two inferior molars of one side and the last inferior of one side being absent. This man was disinterred from a cemetery on the island of Thinic, on the coast of Brittany. Teeth of a considerable number of men, in a separate condition, were obtained by Mr. Wilson from the dolmen of Poulzongue, near the town of Gramat, in the department of Lot, Central France, and have also been examined by me.

The dental tubercular formula of the man of Thinic is \-- 4 4 5

The second superior molar is a perfect quadritubercular, while the third is an equally perfect tritubercular, of somewhat reduced size. This dentition is, then, exactly intermediate between extremes. It is further advanced than that of the lowest existing races, but has not reached the final modification of the most specialized. The dental characters of the men of Poulzongue vary between the type of the man of Thinic and

the formula - - -. Several of them have the second superior 444

molar plainly tritubercular, and perhaps as many have the fourth tubercle imperfectly developed. But no indication of a tritubercular first superior molar is observed, such as occurs in a few Esquimaux. The resemblance of perhaps half of the superior molars of Poulzongue to those of Esquimaux and the more specialized Indo-Europeans is distinct. Pointing in the direction of the latter, are the small size of the teeth and the frequent occurrence of caries.

The Accessory Anterior Internal Tubercle.

This tubercle is characteristic of the genus Lemur (Fig. 11) and some of its extinct allies. Such are the Chriacus pelvidens and C. triincatus Cope, the former figured in Tertiary Vertebrata (Report U. S. Geol. Survey Terrs. III.), Vol. XXIII., d, Fig. 7.

1 8 COPE. [Vol. II.

I have found it well developed in the Malay (Fig. 2, aai), No. 1339. It is present in three Tahitian crania in my collection^ and I have observed it in another Micronesian cranium, and in some others. This character is decidedly lemurine, but whether to be properly termed a reversion or not may be questioned in view of the fact that it always occurs, so far as yet known, in dentitions most remote from that type in other respects. It might be rather regarded as a survival or as a character which has persisted from the " protanthropos " which was itself immediately derived from lemurine ancestors.

Anomalous dentitions.

I have observed the following examples of aberrant molars. In an Esquimaux (A. M. M.) the formula of the superior molars of one size is 3—4 — 0; in another (B. S.) it is 3 — 3— 4. In a French skull (A. M. M.) we observe on one side 4—4—3 ; on the other side, 3J— 3 — 3J. The posterior superior molar is, by reason of its late and retarded development, the most liable to exhibit abnormalities. In a Tahitian skull in my collection, it is tritubercular on one side and obtusely haplodont (conic) on the other.

General Conclusions.

For clearer understanding of these characters, they are arranged in the form of a table. Only the principal races are represented, and hybrids, when determinable, are omitted.^ The characters of the superior molar teeth only are referred to. These are classified under four heads, viz. : first, tubercles 4—4 — 4; second, tubercles 4 — 4 — 3 J or 4— 3j — 3j; third, tubercles 4— 3| — 3; fourth, tubercles 4—3 — 3. As already remarked, the extreme types of the series give the most precise indications of race, while the intermediate conditions have a various range.

In the first table the most obvious results are, that only the three lowest races present four tubercles on all the superior molars, and that of those with tritubercular second and third molars, Europeans and their American descendants greatly predominate. Also that of uncivilized races, the Malays and

1 Mulattoes, Mestizoes, Half-breed Indians, Gipsies, etc., are omitted.

No. I.]



Negroes never, and Micronesians very rarely, present this type of dentition, while in the Esquimaux it considerably predominates, examples of tritubercular first molars even occurring.







'3 «;


rt m



5 (5

S §

c .




2 'a. b;3 >^



a .1-1


^ 6

IS !=^

3 .-■

b 6







w ^










- ?)








-3i] -3i >













- 3 J


- 3




















Total . .











I now give a table of the characters of the superior molars in the Europeans and Europeo-Americans examined. The number is not sufficient for final conclusions ; nevertheless there are some indications of value. Some of the one hundred and nineteen dentitions examined are referred with doubt to their respective races. Thus the Europeo-Americans may have been in many instances immigrants, as many such left their bones on the battle-fields of the American civil war, where many of the crania were picked up. The supposed Germans are largely Austrians, so that some of them may be more or less Slavic or Magyar. Allowance for these would reduce the number of tritubercular molars.

1 Ten crania in the Museum of the Cincinnati Soc. Nat. History, mostly JSIoundBuilders, have the formula 4 — 4—3.



[Vol. II.


-73 a







!« rt


2 in




F C Oc/3



1 g

4-4 -3i







4 - 3i - 3i







4-4 -3






4 - 3i - 3










4-3 -3









Total . .










As results we have the following : the tritubercular dentition appears in 11 out of 25 Slavs; in 7 out of 23 Greeks and Italians ; in 6 out of 22 Germans and Scandinavians ; in 6 out of 8 French ; and in 20 out of 30 Europeo-Americans. The only great race which presents a similar high percentage of tritubercular molars is the Esquimaux, where they occur in 21 out of 30 dentitions. The tendency is most marked in Slavs, French, and Europeo-Americans, and is least marked in Greeks and Italians and in Germans. The former subrace stands in the series between the intermediate type of the North American Indians and the other Europeans. In the Germans the number with tubercles 4 — 3i- — 3 is large. If these be added to the number with 4 — 3 — 3, we have 16 out of 22.

It is important to remember in this connection that the distinguished ethnologist and archaeologist, W. Boyd Dawkins, affirms that the earliest inhabitants of Britain and some other parts of Europe were Esquimaux. He refers especially to the men of the caves, whose implements and arts he declares to be identical with those used by the Esquimaux of the present day.^ As it is evident that the lemurine or tritubercular reversion commenced with the Esquimaux, it may be that in some instances at least its appearance in men of Anglo-Saxon and other European races is due to inheritance alone. But it is also reasonable to suppose that in this case as in other evolutions, the

1 Perhaps improperly included in this table.

2 Early Man in Britain, 1880, p. 233.


cause which produced this modification of the Esquimaux dentition is still active, and its recent appearance in the most civilized races must be due to this unknown cause. The progressive character of the French dentition in this respect is in broad contrast with the primitive character of that of Italians and Greeks. The characters seen in the latter go far towards sustaining Professor Huxley's hypothesis, that the dark Mediterranean sub-races consist of a mixture of Egyptian with the IndoEuropean stock.

In conclusion it may be stated, that the tritubercular superior molars of man constitute a reversion to the dentition of the Lemuridas of the Eocene period of the family of Anaptomorphidae. And second, that this reversion is principally seen among the Esquimaux and the Slavic, French, and American branches of the European race. Observations on some of the races of the Indo-Europeans are yet so imperfect that some additions to the above list yet remain to be made, as for instance, probably, the English sub-race. The neolithic dentitions examined are intermediate between the two extremes, thus showing an advance over the lowest existing races.

The Origin of the Qiiadritubcrcjilar Molar.

This question has an interest beyond the history of human dentition. I will now inquire whether any mechanical cause can be assigned for the retention of the quadritubercular structure.

It may be recalled that I have shown that the development of a fourth tubercle on the posterior side of the internal tubercle of the tritubercular superior molar has been the origin of the dentition of the non-carnivorous types, which are principally Ungulata. The mechanical action of the inferior against the superior molars cannot have been very different in the early Ungulates from what it was in the early flesh-eaters, since the canine teeth, which partly direct this motion,^ are equally developed in both. But the history of the canines in the development of the Ungulates is exactly the reverse of what it has been in the flesh-eaters. In the former the canine has grov/n successively smaller, and has been in most of the lines completely

1 See mechanical origin of the dentition of the carnivora, Proceedings American Association Advanced Science, 1887.

22 COPE. [Vol. II.

aborted. Whence, then, has come this different history.^ It has been probably partly due to the substances used as food. The softer and often tougher animal tissues permitted the shearing motion through their elasticity and extensibility carrying the friction beyond the opposing transverse edges of the crowns, vertically along their sides. The grains and vegetable substances, on the other hand, possess no such elastic qualities, and are cut or broken by the approach of the edges themselves. The pressure would be direct and brief, and not a continued shear. Just how this would result in the development of the fold on the posterior side of the crown of the superior molar is a question of nutrition not yet explained by actual observation ; but it is generally observable, however, that in dentition, folds of the surface have resulted from ordinary use, not too severe. In the case of the shearing which developed the carnivorous dentition, the animal used force to cut the meat in the manner necessary to do it. This shearing force is so great as to wear the crown, rather than to encourage growth. The temporary force required by the act of crushing vegetable matters (excepting such as approach flesh in their characters) is more of the nature of impact, and is of very brief duration. The fourth tubercle has been the result, and I have described its various complex derivatives elsewhere.-^ Can it be possible that a largely, or exclusively meat diet has been the mechanical cause of the development of the tritubercular molar in man . Its great predominance in the Esquimaux suggests this explanation. The lower and quadritubercular races are largely granivorous and frugivorous, but whether so predominantly so as to restrain the modification in question, I do not know. It is probable that the tritubercular molar expresses a change which is both phylogenetic and physiological.^

^ In 1874, in the Journal of the Philadelphia Academy of Natural Sciences, I homologized the various parts of the mammalian tooth-crown structures and traced their phylogeny. The same was done about the same time by Kowalevsky. Not being then familiar with the capacity of dense substances, as dentine and enamel, to yield their form to continued strains, I did not pursue the question of the origin of these forms through use in mastication, although I suspected such origin. This was done in 1878 by Prof. J. A. Ryder, in an able paper on "The Mechanical Genesis of Tooth-Forms," Proceedings Acad. Phila., p. 45. With minor exceptions, I have adopted the views there set forth.

2 After I had nearly completed this investigation, I received during October of this year (18S6) the admirable monograph of Dr. Wortman on the dentition of the ver


tebrata. He describes human dentition more thorouglily than previous authors, and refers to the tritubercular modification in the following language (p. 444) : " The superior molars, like those in the lower jaw, are three in number, and have quadritubercular crowns normally, but many examples can be found in which the posterointernal cusp, the last one added in the quadritubercular molar, is little more than a cingulum, and is scarcely entitled to the appellation of a cusp." And in a footnote he observes, " It is probable that this condition, of which I have seen a number of examples in the higher races, is a degenerate one, and is an effort to return to the tritubercular stage."

Dr. Wortman has, at my suggestion, examined a large number of Esquimaux crania contained in the Army Medical Museum at Washington, which were not accessible at the time of my visit to it. He confirms the value of the tritubercular second superior true molar as a race character.

The conclusions described in this paper are mostly embraced in a preliminary one which appeared in the American Naturalist for November, 1886.

Note. — The discussion of the functional relation of the tritubercular to the quadritubercular molar dentition on a preceding page, has reference to the early mammalian types, and to the question of origin of the quadritubercular at that period. The relations of the parts of opposite jaws are different in the tritubercular races of men, since the interaction of the crowns in mastication is no longer alternate but opposite.

24 COPE. [Vol. II.



Fig. I. Chimpanzee, a superior, b inferior molars. Mus. Academy Philadelphia.

Fig. 2. Malay of Madura, No. 1339, Mus. Academy Philadelphia. Superior molars all quadritubercular, and m'^ with the lemurine (Pig- 13) accessory anterior internal tubercle (badly represented by artist).

Fig. 3. Negroid- Egyptian (Morton), No. 852, Mus. Academy Philadelphia.

Fig. 4. Tahitian, without lower jaw. From Dr. Chassaniol, the colonial physician of Tahiti. Mus. E. D. Cope.

4 4 37 Fig. 5. Pessah, No. 1097, Mus. Academy Philadelphia. Formula————^ Fig. 6. Neolithic man from Isle Thinic, Brittany; from collection of Thos. Wilson, Esq.

Journ. Morph.Vol. 11

Plate II


26 COPE, [Vol. II.


Fig. 7. Spokane from Rock Lake, Witman County, Washington Territory, of the

4 4 3 Kamaka tribe; collected by C. H. Sternberg. Mus. E. D. Cope; yTa

4 3^ 3i Fig. 8. Hindu of Bengal, No. 1344, Mus. Academy Philadelphia; " " •

Fig. 9. Swede, No. 1487, Mus. Academy Philadelphia. From Professor Retzius;

4 37 3 440'

Fig. 10. Esquimaux of Greenland, No. 1221, A.M.M.

Fig. II. Alaskan Esquimaux, No. 2669, A.M.M.

4 3 3

Fig. 12. Europeo-American, No. iioS, Mus. Academy Philadelphia; •

Fig. 13. Lemur collaris, Madagascar. Mus. E. D. Cope. Superior molars tritubercular with accessory anterior internal tubercle.


ae, anterior external tubercle; pe, posterior external; ai, anterior internal; pi, posterior internal; aai, accessory (lemurine) anterior internal tubercle.

Journ. Morph.Vol. II

Plate III.





The question as to the role of the cytoplasm, presents itself under two forms :

1. Is the cytoplasm a passive body, moving only as it is acted upon by external forces, or in response to influences emanating from the nucleus ?

2. Or does it behave rather like an organized body, endowed with subtle powers of its own, and capable of automatic as well as responsive action ?

There is a strong tendency at the present time to refer all kinetic changes in the cytoplasm to the agency of the nucleus, and to ascribe to the former the passive role of a nutritive substance. The kinetic phenomena of the egg during maturation and impregnation have already been considered in their bearing on this important question. ^ A number of decisive proofs of pure nuclear action were pointed out, and at the same time an attempt was made to support the opinion that the cytoplasm is capable of automatic as well as responsive action. The present paper is chiefly devoted to the consideration of phenomena displayed in the cytoplasm, and to the discussion of the question, whether the regenerative and formative power of the cell resides in the nucleus or in the cytoplasm, or in both taken as a highly complex physiological unit.

TJic Doctrine of Isotropy. — Pfluger's interesting experiments ^ with the amphibian egg to determine the influence of gravitation upon the direction of cleavage-planes, led him to conclude that the entire ^^g is " isotropic In other words, to quote from the

^ Ookinesis. Journ. Morph., I., 2, p. 227, December, 1887.

2 Ueber den Einfluss der Schwerkraft auf die Theilung der Zellen. Arch.f. d. ges. Physiolcgie, XXXI., pp. 311-318, and XXXII., pp. 1-79. 1883.

28 WHITMAN. [Vol. II.

author, " the fertilized egg possesses absolutely no essential relation to the later organization of the animal, no more than a snowflake stands in any essential relation to the size and form of the avalanche which under certain conditions develops from it. That the germ always gives rise to the same form, is due to the fact that it is always brought under the same external conditions." Immediately after the appearance of Pfliiger's papers, it was pointed out by Agassiz and Whitman ^ that, *' if gravitation were the sole guiding agency in cleavage, its effect ought to be instautaneoiis, and it should be possible to change the direction of a cleavage-plane already in progress." It was also shown that the time required to bring about a transposition of the third cleavage-plane, suggested a corresponding internal transposition of the active protoplasmic matrix of the ovnm, including of cow'se the nnclei, '* If a body constituted like the ovum is restrained by artificial means from taking its normal position, a redistribution of material must immediately set in and continue until the equilibrium is restored. The active portion of the ovum, having a lower specific gravity than the passive nutritive elements, would eventually recover its normal position, and thus tJie virtual axis of the ovum zuould inevitably j'ight itself in spite of the inability of the ovum to rotate bodily,^^ Later observations have fully verified these suggestions.

As now maintained by Born,^ Hertwig,^ Weismann,^ Kolliker, and others, the cytoplasm alone is isotropic, while the nucleus is the seat of the directive and form-giving power in development.

In this modified form, the doctrine of isotropy makes a much nearer approach to truth, but I believe that it is far from correct in its estimate of the functional importance of the cytoplasm.

The logical consequences of this view are clearly presented by Oscar Hertwig^ (p. 306) in the following words : '^ An die Kernsubstanz also sind die Krdfte gebujiden, durch zvelche die

'^ On the Development of some Pelagic Fish Eggs. Proc. Amer. Acad. Ai'ts &' Sciences, XX., p. 40, 1884.

  • Biolog. Untersuch. Arch.f. mik. Anat., XXIV., 1885.

5 Das Problem der Befruchtung und der Isotropie des Eies. yenaische Zeitschrift, XVIIL, 1885.

^ Die Continuitiit des Keiniplasma's als Grundlage einer Theorie der Vererbung. 1885.

"^ Die Bedeutung der Zellenkerne fiir die Vorgange der Vererbung. Zeitschr. f, zaiss. ZooL, XLII., 1885.


organisation dcs Thicres bestimmt wird. . . . £s erscheint gleicJigiiltig, ob bei dcr erstcn TJicilung der eine Kern sick mit der sogcnanntcK a^mialcn, dcr andere mit der vegetativcn Dottersiibstanz timhiillt oder ob beide Kerne sick in vegetative und a7iimale Dottcrsubstanz in dicser oder jener Weise thcilen. . . . Der Dotter ist nicJit so organisirt, dass aus einer bestiimnten Portion dessclben ein bcstininites Organ hervorgehen milsste. It is conceded (p. 304) that the cytoplasm may have a low grade of organization ; but it is an organization that changes from moment to moment (p. 309), not a "feste Organisation " bearing fixed relations to the future organism.

The opposing view finds the differentiating and formative principle either in preformed elements of the cytoplasm {" physiological moleadcs" Lankester, '^polarized molecules Pfliiger, " idioplasvia " Nageli), or in a definite organization of the cytoplasm itself (Van Beneden and others). "Though the substance of a cell," says Lankester^ (p. 14) "may appear homogeneous under the most powerful microscope, excepting for the fine granular matter suspended in it, it is quite possible, indeed certain, that it may contain already forrned and individnalized, various kinds of physiological molecules." And again, "The development of one kind of cell from another kind is dependent on internal movements of the physiologipal molecules of the protoplasm of such cells."

Van Beneden ^ has made a thorough study of the structure of the egg of Ascaris megalocephala with a view to ascertaining if " les plans de symetrie de I'embryon ne se trouvent pas deja preformes dans I'oeuf lui-meme et si I'un des traits les plus caracteristiques de I'organisation de I'esp^ce, la symetrie qui la distingue, ne se trouve pas deja indiquee dans I'oeuf. L'oeuf d'un animal a symetrie bilaterale aurait-il, comme I'animal dont il provient et qu'il doit devenir, une extremite anterieure, une extremite posterieure, une face ventrale, une face dorsale, une droite et une gauche } les materiaux qui doivent servir a edifier la moitie droite du corps siegent-ils dans la moitie droite de I'oeuf et la substance

8 Notes on Embryology and Classification. London, 1877.

9 Recherches sur la maturation de I'oeuf et la Fecondation. Arch, de Biologic, IV., 2 & 3. 1883.

30 WHITMAN". [Vol. II.

de la tete ne se trouve-t-elle pas concentree en un point determine du corps ovulaire .■* "

Both views recognize the necessity of assuming that the course of ontogenetic development is in some way predetermined in the ^^g ; but while one finds the force motrice in the nucleus, the other would locate it in the cytoplasm. The advocates of the former appeal to the so-called isotropy of the cytoplasm, to the conspicuous part played by nuclear bodies in fecundation and cleavage, to the incapacity for regeneration shown by enucleate fragments of infusoria, etc. ; while the supporters of the latter insist on the constancy of premorphological relations (axial relations, relation of first cleavage-plane to the median plane of the future embryo), the remarkable structural features exhibited in some eggs, cleavage in planes not previously marked by karyokinetic division, etc. The truth appears to me to lie on both sides, the error consisting only in unduly exaggerating the relative importance of one or the other factor. Just now the weight of authority seems to be turning in favor of the first view, a result which must be attributed very largely to the influence of recent discoveries and theories respecting the nature of fecundation. The question is one of such fundamental importance, that it seems desirable to analyze closely the facts bearing on the subject. It is for this reason that I have dwelt more at length on the movements of the germinal vesicle and the pronuclei.

Especially important is the study of the structure of the ^g%, and the modifications which it undergoes during the period of maturation. One of the most important contributions in this direction is unquestionably Van Beneden's great work on Ascaris. No other biologist has yet gone so deeply and thoroughly into the subject, nor has any one discussed it with a keener appreciation of its theoretical importance. Such well-marked structural features as are claimed to exist in this ^gg are inconsistent with the idea that the cytoplasm is isotropic.

Polar Ri?tgs. — Among the more extraordinary examples of cytokinesis may be mentioned the polar rings in the tgg of Clepsine, first described by Grube^*' (pp. 15-16), and recently more in detail by Robin " (pp. 97-105) and Whitman 12 (pp. 20-29, 39-41) 1 Untersuchungen ueber die Entwicklung der Clepsinen. Konigsberg, 1844. ^^ Memoire sui- le Developpement etnbryogenique des Hirudinees. Paris, 1875. ^2 The Embryology of Clepsine. Quart. Jour. Mic. Sci., July, 1878.


Although it is by no means certain that the hyaline protoplasm of these rings is not in part derived from the germinal vesicle, it is quite clear that the phenomena are very different from the polar phenomena attending the division of the first cleavage-nucleus. These remarkable exhibitions of polarity in the cytoplasm appear early in the pronuclear stage, and continue not only during the centripetal march of the pronuclei but even after the first cleavage-nucleus has entered upon its kinetic phases of division. Thus we have tzvo distinct scries of polar phenoDiciia in progress at the same time, one displaying itself in the cytoplasm, the other in the nnclcns. We cannot suppose that the cytokinetic series is dependent upon the karyokinetic series for two reasons : first, because the former begins earlier than the latter ; and second, because such cytokinetic displays are unknown in other eggs. The second ground would also hold against referring them to pronuclear influences. Allowing that it may yet be possible to demonstrate that these movements originate in response to pronuclear influence, it would still be very difficult to believe that they are sustained throughout by the continuous action of the same agency. It would be altogether more probable, as a little reflection vv'ill show, that the movements once started are capable of maintaining themselves, independently of the inciting cause.*

Cytokinetic Phenomena. — The cytoplasm exhibits a great variety of changes and conditions, variously described as 'polar concentration,' 'radiating bands,' 'waves of contraction,' 'zonal constrictions,' 'automatic cortical layer,' 'amoeboid movements,' ' phases of segregation,' * astral radiations,' ' rhythmic contractility,' 'migratory movements,' 'crown of folds' (Faltenkranz), ' autonomic movements ' (rotation, circulation, pulsation), etc.

While some of these phenomena might, with considerable reason, be claimed as purely cytokinetic, most of them are so intimately associated with karyokinetic activity that they must be explained, either as the direct result of the latter, or as the effect of impulses generated by the interaction of nucleus and

  • The polar and parapolar circles described by Van Beneden (No. 9) in the egg of

Ascaris, are not comparable with the polar rings of Clepsine. For the ' disque polaire' arises before, and disappears with, the penetration of the spermatozoon; the polar rings (Clepsine), on the contrary, appear after the penetration of the spermatozoon, and are not wholly dissipated until after the completion of the first cleavage.

32 WHITmMAN. [Vol. II.

cytoplasm. The hypothesis of reciprocal action is not incompatible with the opinion that the conditions of this action are furnished, in the first instance, if not continuously, by changes of a chemical or molecular nature, which arise quite independently, either in one factor alone, or in both. The source of the initiatory impulse would still be an open question. Our knowledge of the phenomena above designated is too incomplete to furnish a key to the solution of this problem. For the purpose we have in view, it will be sufficient, therefore, to refer to a few of the more important examples, in which each factor may be supposed to play a more or less important part, deferring the discussion of the main question until we come to consider phenomena of a more decisive nature.

Following the penetration of the spermatozoon into the ovum, various forms of contraction in the vitellus have been observed ; and these are generally regarded as the effect of impulses generated by the spermatic element. The usual sequence of events certainly accords very well with this view, but there are one or two facts which should make us hesitate to accept it. In some aquatic animals, in which the sexual cells unite before ovipositing, the time of these contractions bears no constant relation to the time of union, but does bear such a relation to the time of contact of the Qgg with water, whether this contact be brought about artificially or in the natural course of events. Still another cogent reason for not ascribing these contractions to the independent action of the male pronucleus is found in the fact that similar, though more sluggish, movements may, in some well ascertained cases, be induced by placing tmfertilized eggs in water. The most general of these movements is the flattening of the pole, and the gradual contraction of the whole vitelline sphere, resulting in the formation of a perivitelline space.

The Constriction Attending the Exit of tJie Polar Globules. — The flattening of the pole is attended, or followed, in some cases, with a very remarkable constriction, which, beginning in the equatorial zone, travels towards the animal pole, finishing up with a nipple-like protuberance, from which the first polar globule is expelled. The exit of the second polar globule is sometimes preceded by a similar but weaker constriction. This constriction has been observed (No. 12, p. 18) in the eggs of different species


of Clepsine ; and the same, or a closely analogous constriction, has been described by Kupffer and Benecke^^ (p. 19) in the Qg% of Petromyzon, and by Ransom ^* (pp. 463, 464, 477, 479) in the Q.g% of the Stickleback and some other fresh-water teleosts.

This constriction has been confounded by the last mentioned authors with yolk-contraction, and brought into connection with the formation of the perivitelline space (' breathing-chamber ' of Ransom, * Eiraum ' of Kupffer and Benecke). This space probably results from contraction of the vitellus as well as from expansion of the ^g^ membrane, but the constriction is a special act of the vitellus to expel the polar globule. The elimination of polar globules is thus a process involving co-operant actions of both factors ; and if the part performed by the polar amphiaster is karyokinetic, the associated act of the vitellus may be characterized as cytokinetic. I have before referred to the centrifugal movement of the germinal vesicle as an instance of repellant action, and I regard this constriction as a part of the same action. It is thus a phenomenon of maticration, not of hnpregJiation.

Ransom, as well as Kupffer and Benecke, explains the phenomenon as a result of the penetration of the spermatozoon, and hence has failed to distinguish it from other phenomena of a similar, though not identical, nature. Bearing this fact in mind, we are enabled to find in their descriptions — especially that of Ransom — an exact parallel of the special constriction which always accompanies the formation of polar globules in Clepsine. Ransom's account is extremely interesting, and has attracted so littk; attention from later embryologists, that it seems worth while to introduce a portion of it here. After stating that slow contractions begin from the first moment of entry of the spermatozoa, causing first a flattening of the germinal pole, and afterwards slight changes of outline due to ' travelling waves ' at other parts of the surface, he proceeds as follows : —

" Gradually more vivid contractions commence, at various times after fecundation, according to the temperature. In warm weather they have been noted in six minutes, in cooler weather

1^ Der Vorgang der Befruchtung am Eie der Neunaugen. Konigsberg, 1878. i* Observations on the Ovum of Osseous Fishes, Proc. Roy. Soc. London, VII., 1 85 6.

34 WHITMAN. [Vol. II.

in fifteen or twenty minutes after impregnation. They cause a flattening of one side of the yolk-ball, to see which it is often necessary to roll the ^gg over. The flat sitrface gradually becomes a stilcits, giving a renifonn outline to the yolk. It then extends all roimd, giving rise to a dumbbell shape. This sulcus, which may be termed equatorial, travels ivith considerable but variable rapidity tozvards the germinal pole, producing as it passes on, tJie flask form. The sulcus is lost by passing forivards to the germinal pole, not by relaxatioii. It is seen for a brief space affecting the thickness of the gcrniinal disc only, to zvhich it gives a nipple-like form, zvhile the food-yolk is round. WJicn ejfaced, the whole yolk-ball is globular and at rest, the gcnninal disc being no longer prominent. This series of forms recurs luith more or less of regularity, and with some variations both of time and form, about fifteen or tzoenty times, each series being the result of a travelling zvave (No. 14, pp. 463-464).

Had Ransom succeeded in connecting the formation of polar globules with the more regular and prominent 'wave,' which he has so vividly described, he would doubtless have seen the necessity of distinguishing this wave from the movements which follow it, precisely as Kupffer and Benecke distinguished in the &^g of Petromyzon a 'zonal constriction' which invariably accompanies the appearance of the second polar globule. They did not succeed in tracing the origin of the first polar globule ; but they have described a constriction (p. 15) around the germinal pole (Fig. 7, /), which appears im.mediately after the spermatozoa come into contact with the ; and this, I would suggest, may have the same relation to the first polar globule that the ' zonal constriction ' has to the second.

Polar Aggregation. — In the formation of the germinal disc of many pelagic fish ova, we meet with very remarkable cytoplasmic movements. In the fresh-laid egg, the germinal protoplasm forms a cortical layer of uniform thickness around the yolk. But this condition lasts only for a few seconds, during which the spermatozoon finds an entrance into the egg. This event is followed at once by a polar concentration of the peripheral layer of protoplasm, which results in the gradual formation of the germinal disc with its centrally placed pronuclei.

Is it in one or both of the pronuclear bodies that we are to look for the cause of the polar aggregation of protoplasm } Is


the protoplasm a passive mass, moved at the expense of nuclear energy alone, or has it motor energy of its own ? In the latter case, are the conditions necessary to action supplied by the nuclei or by the protoplasm, or by both ? Although we are yet a long way from a solution of these questions, it may be possible to show that the protoplasm is an active rather than a passive factor in the movements we are considering. Any view which represents the germinal protoplasm as a passive body, moving only as it is impelled by nuclear forces, appears to me irreconcilable with the following facts : —

1. In most meroblastic vertebrate ova (including those of many teleostei), the germinal disc is already formed before fecundation takes place. The male pronucleus cannot therefore be a necessary factor in the formation of this disc.

2. In many pelagic fish ova, where the disc forms after fecundation, the polar amphiaster is formed before polar concentration begins. The cause of concentration cannot therefore, in this case, be referred to the centrifugal movement of the germinal vesicle, nor to any changes which this body undergoes prior to the formation of the polar amphiaster.

If this conclusion holds equally in the first class of eggs, we are fully warranted 'in affirming that the germinal disc forms independently, not only of the male pronucleus, but also of the germinal vesicle and its derivatives, since in these eggs the disc is formed before fecundation and before the polar amphiaster divides.

The validity of these conclusions may be disputed by those who hold with Weismann (No. 6, pp. 90-122) that the two pronuclei are identical in their molecular structure, and that both act alike upon the protoplasm, but in proportion to their mass. It might be argued that a definite quantity of karyoplasm (* Keimplasm ') is requisite in order to concentrate the protoplasm in the form of a polar disc. If the mass of the germinal vesicle, or of its pronuclear elements, be large enough, it would form a germinal disc without the aid of the male pronucleus ; if it fall short of the requisite mass-measure, it would have to be reinforced by the male pronucleus before it could accomplish the work. It will be time to accept this view when it has been shown that there arc such quantitative relations as the theory postulates. Such an explanation of the disc-formation would take no account

36 WHITMAN. [Vol. II.

of the polarity of the ^%^, and would leave inexplicable the difference between telolecithal and controlecithal eggs. Besides, this view assumes that the pronuclei are homodynamous, a point which cannot be conceded, since male pronuclei do not behave towards one another as they do towards female pronuclei.

A special feature in the polarity of the fish &^g, noticed by Kupffer^^ and by Hoffman ^^ (p. 88), is the formation of a temporary discoidal thickening (' Gegenhugel,' Kupffer) at the vegetal pole. Here, then, is a disc-formation at the point farthest removed from the nuclear bodies, and this fact appears to be fatal to the above theory. We are reminded of the rings in the Qgg of Clepsine, and their concentration into two polar discs. It appears not improbable that the two sets of phenomena are similar in nature, and determined by like forces. In the fish Q.^g the disc-formation is not preceded by a ring-formation ; and the nearest approach to the ring-rays are the ' radiating bands ' or 'beaded streams' described by Ryder i (p. 17).

Raffaele (Mitth. d. Zool. Station z. Neapel, VIII., i, 1888) has recently described a very singular phenomenon in the ^gg of Labrax. The ^gg has a single large oil globule at the pole opposite the germinal disc. This oil globule is enveloped with protoplasm, which, on the side facing the germ, thickens up until it takes the form of a long club-shaped body. This body elongates in an axial direction, and the distal portion, which is gradually constricted off, eventually assumes a globular form and rests on the inner face of the germ. Ryder has noticed protoplasmic bodies in the egg of Gadus (' segmenting corpuscles ') which, as Raffaele suggests, may have a similar mode of origin. It occurs to me that this body of protoplasm may correspond to Kupffer's * Gegenhugel,' and that it is diverted from its usual peripheral track by the presence of the oil globule.

The Artificial Division of Infusoria. — The artificial division of infusoria has been resorted to as a means of deciding, experimentally, the question of the relative importance of the nucleus. M. Nussbaum^^ was the first to establish

15 Die Entwickelung des Herings im Ei. Jahresb. d. Comm. z. wiss. Untersuckung der detitschen Meere in Kiel, IV.-VI., 1874- 1876.

18 Zur Ontogenie der Knochenfische, Amsterdam, 1881.

1^ The Embryography of Osseous Fishes. Report of the Commissioner of Fish and Fisheries for 1 882.

18 Die spontane und kUnstliche Theilung der Infusorien. Arch.f. mik. Anat.,


the general fact, that enucleate pieces of an infusorium are incapable of regenerating lost parts, while nucleate pieces soon regain the specific form. " The micleiis is tJuis indispensa' blc to tJic preservation of the formative energy of the eell." Gruber,^^ whose experiments were begun at about the same time as those of Nussbaum, has reached the same general result. But one of Gruber's experiments, which was at first supposed to show that regeneration is possible without the presence of a nucleus, brings out a fact of considerable importance. A Stentor ccemlcus was selected, in which spontaneous fission had already begun, as indicated by the concentration of the rosary-formed nucleus into a simple oval form, and by the beginning of a new peristome at the middle of the body. A transverse section was made just in front of the new peristome, and so close upon the nucleus that it nearly all escaped from the cut surface of the posterior half of the Stentor. The anterior portion retained no part of the nucleus. The two parts were isolated, and on the following day each was found to have become a complete Stentor. As the plane of division was nearly the same as that which would have been followed if the process of spontaneous fission had not been interfered with, Gruber (No. 19, p. 13) finally concludes, contrary to his first interpretation, that this was not a case of regeneration, but of reproduction. The process of reproduction once initiated by nuclear action may go on, so he thinks, to completion without further assistance from the nucleus. He insists, however, that the first impulse to division is given by the nucleus, since in all cases where artificial division is executed before the process of spontaneous fission begins, enucleate parts are incapable of regeneration. While finding no reason to doubt the accuracy of Gruber's observations, I must contend that they do not warrant the conclusion so forcibly stated in the following words : " Auf rein empirischen Wege werden wir hier vor die unumstossliche Thatsache gestellt, dass der Kern der wichtigste, dass er der arterhaltende Bestandtheil der Zelle ist" (No. 19, p. 16). Allowing that regeneration is impossible in the absence of a nucleus, that is no proof that the nucleus is the sole seat of regenerative power, nor is it a proof that the nucleus is

XXVI., January, 1886, p. 485. Cf. also Sitz-Ber. d. Niederrh. Ges. f. Natur- u. Heilkunde, 1884, p. 262.

^^ Beitrage zur Kenntniss der Physiologic und Biologic der Protozoan. Berichte d. Nattirforsch. Ges. zu Freiburg, I., H. 2, 1886. Also Biolog. Centralbl., IV., p. 717, and v., p. 137.

38 whitman: [Vol. II.

a more important factor than the cytoplasm. There is not a single observation to prove what is so confidently asserted, that the nucleus gives the ' Anstoss ' to division. This may be so and it may not. The observations prove, (i) that, in some of the higher Protozoa, tJie zvhole process of reproduction by fission, with exception of the ijiitiatory steps, may be accomplished independently of nuclear actiott ; a7id (2) tliat the initiatory steps cannot take place if the nucleus and the cytoplasm are artificially separated. The whole truth is well stated by Nussbaum (No. 18, p. 516) : "Kent zmd Protoplasma sind nur vei'eint lebensfdhig: beide sterben isolirt nach kurzerer oder langerer Zeit ab." It is clearly impossible therefore, by any such experiments as Gruber has carried out, to settle the question of the precise locus of the regenerative energy.

On general theoretical grounds, as well as by the fact that enucleate forms are found among the Protista, we are compelled to accept the generally received view, that the nucleus is secondary in origin. It may be true, as suggested by Gruber 2*^ (p. 151), that these so-called enucleate forms contain nuclear substance in solution, and that the first step in the phylogeny of the nucleus consisted in the formation of scattered granules, the coalescence of which would give rise to the single nucleus. But it is hardly necessary to add that we do not see how this (or any other) mode of explaining the origin of the nucleus as a secondary body can be brought into harmony with the idea that it embodies the whole regenerative energy of the cell.

Interesting and instructive as are these experiments in the artificial division of the higher Protozoa, they do not alone furnish a satisfactory basis for general conclusions. They require to be supplemented by similar experiments on the simpler forms of Protozoa, and by much more complete observations than have yet appeared on the normal processes of fission and coalescence exhibited in the Heliozoa. In proof of this we have only to refer to Gruber's own observations on ActinopJirys sol'^^ (pp. 6^6"/) and Actijiosphcerium eichhorjiii'^'^ (pp. 381-382). Actinophrys

"^^ Ueber Kern und Kerntheilung bei den Protozoen. ZeitscJu-. f. wiss. Zool., XL., I, p. 121.

2^ Untersuchunge nueber einige Protozoen. Zeitschr. f. wiss. Zool., XXXVIII., p. 45, 1883. Also Zool. Aiizeiger, No. Ii8, 1S82, p. 423.

22 Ueber Kerntheilungsvorgange bei einigen Protozoen. Zeitschr. f. wiss. Zool. XXXVIII., p. 372, 1883.


is capable of breaking up into parts without the concurrence of any visible changes in the nucleus. The enucleate individuals — if such they are entitled to be called — may agree perfectly with the nucleate individual in outward form and behavior. They live and grow, and coalesce not only with one another, but with the nucleate form. Whether they are capable of generating a nucleus or not was not ascertained. The multinucleate ActinospJiicriuvi breaks up in a similar manner without the intervention of the karyokinetic process ; but here the individual parts generally contain one or more of the original nuclei. Two or more individuals may coalesce, but the coalescence extends only to the cytoplasm, the number of nuclei being the sum of those contained in the separate individuals before fusion. These remarkable facts appear to bear out Gruber's conclusion : " That the nucleus has no importance for those functions of the cell which do not stand in direct relation to reproduction ; such as locomotion (pseudopod-formation), inception of food, excretion (pulsation of contractile vacuoles), and growth. Even on the external form it may be without influence " (No. 21, p. 66). Gruber (No. 19, p. 12) still maintains the accuracy of this view in every particular except that relating to the influence of the nucleus on the form of the cytoplasm. He now holds, in common with Weismann, Hertwig, Kolliker, Strasburger, and many other German biologists, that the form-creating and form-conserving principle is confined to the nucleus. How, when, or where the nucleus manifests its form-moulding power remains a mystery ; but that it gives the first impulse to the regenerative act must be inferred — so the argument runs — from the fact that the violent separation of nucleus and cytoplasm destroys the regenerative power in all cases, except where the process of reproduction is begun before separation is executed. This is the focal point of the question with which we set out.

Aside from experiments in the artificial separation of nucleus and protoplasm, the principal arguments in support of this view are drawn from the phenomena of fecundation and cleavage, and are involved with certain theories of heredity which cannot be dealt with here. The observations of Gruber on Actinophrys and ActinosphcBrhivi^ if the phenomena are not of a pathological nature, — and such an interpretation seems to be precluded, — ■ should certainly make us hesitate to ascribe all the form-regu

40 WHITMAN. [Vol. II.

lating power ot the cell to the nucleus. This reserve is rendered imperative by other facts yet to be mentioned, for one of which we are again indebted to Gruber^^ (p. 221). In summing up the results of his " Studies on Amoebae," he states, " that two very closely related species of Amoeba may have quite unliketormed nuclei ; while species differing widely in external form may have quite similar nuclei." Plainly this is the contrary of what might be expected if the formative power lay exclusively in the nucleus.

No Form-Cot'relation between Nucleus and Cytoplasm. — It may be put down as an indisputable fact that no formcorrelation exists between nucleus and cytoplasm. Except during the process of division, the nucleus seldom departs from its typical spherical form. It divides and subdivides, ever repeating the same steps and ever returning to the same round or oval form. So far as can be seen, its influence upon the cytoplasm is equal in all directions ; and hence it would seem that its formative power, if it have any, could only contribute to the maintenance of the spherical form of the cytoplasm. How different with the cell ! It preserves the spherical form as rarely as the nucleus departs from it. Variation in form marks the beginning and the end of every important chapter in its history. While the nucleus repeats over and over again its little cycle of form-changes with mechanical regularity, the cell marches straight on from form to form, never returning, and never repeating, differentiating, developing, and adapting itself at every step to its environment and to the work it is destined to perform. From the egg onward through all the stages of histogenesis and form-evolution, we search in vain for a single intimation anywhere that either the form of the organism or the forms of the individual cells are moulded by direct nuclear influence. At the beginning of any ontogenetic series, when we get the most rapid and vivid displays of nuclear energy, we see that the environment of each cell is much more potent in determining its form than the nucleus. True, certain conditions of the environment may be said to be largely the result of nuclear activity ; and to this extent the nucleus may be said to determine, indirectly, the form of the cell. But this is very different from saying that the nucleus has a direct controlling

23 " Studien iiber Amoeben." Zeitschr. f. wiss. Zool., XLI., p. i86, 1884.


power over the specific form of the cell, as claimed by Gruber, Weismann, and others. When the end of the ontogenetic series is reached, we find the reproductive power of the nucleus greatlydiminished, and its influence over the form of the cell proportionally reduced. Indeed, in the majority of cases the form of the cell now appears to be maintained entirely independently of the nucleus, and whatever modification of form the latter exhibits appears to be the result of mechanical pressure. The form of the nucleus is now determined by that of the cell rather than the reverse.

If we study the more varied form-changes of the nucleus which occur in the life-cycle of one of the higher Protozoa, we are struck with the fact that the external form of the infusorium is, to all appearance, completely independent of what transpires in the nucleus. A single nucleus may divide a hundred times or more without the slightest effect on the form of the infusorium. The products of these many divisions may undergo various changes of form, and ultimately coalesce to form a single nucleus, and yet no change of form in the cytoplasm. The multinucleated form may break up into parts, some without, others with one or more nuclei, and the enucleate, uninucleate, and plurinucleate individuals all agree in presenting the same specific form. The nucleus may pass from the oval form to that of a long rosary; then, after a period of vegetative life on the part of the cytoplasm, return to the original oval form, and undergo the regular changes of division, the whole cycle of transformations coming to a conclusion without producing any discernible effect on the cytoplasm beyond that of simple fis-. sion. Each part carries with it the power to resume at once the form which characterized the original whole. The nucleus gives no evidence at any time of holding the formative power ; but we have seen from Gruber's experiment on Stentor, above referred to, that the cytoplasm does exercise this power, and that // does so even in the absence of a nucleus.

Gruber does not attempt to deny this ; but he thinks it necessary to assume that the power exhibited by the cytoplasm in the case mentioned, was communicated to it in the form of molecular motion, the original impulse being given by the nucleus. As we have seen, there is nothing in his experiments which makes such a conclusion necessary, and the burden of proof properly falls to him who makes the assumption.

._ WHITMAN. [Vol. II.

One needs only reverse the case to see the illogical nature of the position. Let us suppose that it has been ascertained by numerous experiments that the nucleus of an infusorium is incapable of karyokinetic division when separated from the cytoplasm, except in those cases where the division is already in progress at the time of separation. Would there be anything either in the general rule or in these exceptional cases, from which to conclude that the performances of the nucleus are first set in motion by some impulse from the cytoplasm } Would the incapacity of the nucleus to divide when placed in abnormal conditions demonstrate its inability to divide autonomically under normal conditions ? If artificial isolation were found insufficient to arrest a series of complex movements already begun in the nucleus, would the death of the nucleus, shortly after the completion of these movements, furnish any ground for denying that they were automatic } These questions appear to present the matter in a just light, and to carry with them their own answers. The point would hardly seem to deserve the attention given it, were it not for the great importance of the question under consideration, and the fact that we are dealing with the opinion of a high authority, — an opinion for which experimental evidence of a crucial nature is claimed, and an opinion fully indorsed by so eminent a thinker as Weismann.

ScJmeider s Experi'ment. — Schneider's ^^ (p. 509) experiment with Thalassicolla, although not affording any decisive evidence as to the precise location of the formative power, is yet of some interest in this connection. The extracapsular or cortical protoplasm was removed, leaving the central capsule free. At the end of twelve hours the whole surface of the capsule showed delicate pseudopodial extensions, and soon after appeared a distinct extra-capsular layer (" matrix " of the pseudopodial rays), which gradually grew to normal thickness. The experiment was repeated three times in succession on the same individual, and each time with the same result. The extra-capsular envelope gave no evidence of sharing the regenerative power, but died shortly after isolation. This envelope, according to Brandt ^^

^ Zur Kenntniss des Baues der Radiolarien. Miil. Arch. 1S67, p. 509. 2S Koloniebildende Radiolarien (Sphaerozoeen). Fauna und Florad es Golfes von Neapel. XIII. Monographic, 1885.


(p. 84) takes no share whatever in the production of spores ; and thus it appears that the central capsule is the seat of the reproductive as well as the regenerative energy. During "vegetative life " the various functions (digestion, locomotion, sensation, etc.) all appear to be performed by the extra-capsular cytoplasm ; but the whole functional activity, as supposed by Brandt, is under the regulative influence of the central capsule. In the Sphnsrozoa colonies the division of labor is carried still farther ; for while the pseudopodial rays have their special functions, certain definite areas of the matrix (" Klumpen von Assimilationsplasma ") provide for the digestion of the starch granules which are produced by symbiotic algae ; and different layers or zones are distinguishable even in the central capsule. That there should be such a complete separation of functions between the central core and the cortical layer of the same body of cytoplasm is no less instructive than it is remarkable. It is a capital illustration of the possibilities of organization and the physiological division of labor without any corresponding division into distinct morphological units.

Recent studies tend to show that the only important substance conveyed into the &g^ by the spermatozoon is that which takes the form of the male pronucleus. The unavoidable conclusion would appear to be that the pronuclei are the sole bearers of hereditary tendencies. This is unquestionably a point of cardinal importance, and it furnishes the strongest argument that has yet been advanced in favor of regarding the nucleus as the seat of the formative power. This side of the question could not be fairly dealt with within the limits of the present paper, as it would lead to a consideration of the whole problem of heredity. It is my purpose, however, to return to this subject at no very distant date.

TJie Idea of a Formative Power. — Let us now consider whether any rational basis can be found for the idea of a formative power as a resultant from, and an expression of, physiological unity. I am fully conscious that the subject is one of profound mystery, the solution of which appears to lie as far beyond our grasp to-day as at any time in the past. We draw nearer to the problem, but the effect is rather to enhance than to reduce its apparent magnitude. Every step in advance only brings us to a keener sense of the subtle and

44 WHITMAN. [Vol. II.

incomprehensible nature of the force or forces contemplated. We see the effects only imperfectly, and are baffled in every attempt to understand the mode of action. For the present we must be content to search for the direction in which answers lie ; and herein is found the chief value of theories.

The more important speculations on this subject have taken the form of theories of heredity. In most of these theories, at least those of recent date, we find a fundamental idea which must be accepted as true ; namely, that the sexual cells reflect in some way in their chemico-physiological constitution all the typical structural features of the parentorganism. How all the hereditary tendencies can be contained in a single cell, and with such completeness that the developing organism repeats step for step the chief form-phases of a genealogical history stretching through countless myriads of generations, back from the present into the very dawn of life, and ultimately unfolds every detail of structure and feature of the parent-organism, is a mystery that transcends our understand\ng. The preformationists of last century took refuge in the celebrated inclusion (Einschachtelung, emboitement) theory, which made the real mystery unapproachable by hiding it behind an endless series of miracles. The triumph of the epigenesists brought with it the reclamation of the problem, but left us with the indefinable vis essentialis of Wolff, the nisiis formativiis of Blumenbach. In more recent times we have seen various metaphysical hypotheses of extra-organic agents or forces supplanted" by physiological hypotheses which seek the cause of the phenomena in intra-organic forces. But the reaction which followed the fierce struggles with vitalism has left its impression on most of the theories now in vogue. Biological problems have been brought more and more under the influence of mechanical conceptions, which regard all phenomena from an objective standpoint. Science has vindicated thi§ method, and as a method it is unassailable. It is no less indispensable to research in the organic than in the inorganic world ; but the biologist is reminded at every turn that the method is not exhaustive, and from the nature of the case it never can be. The biologist does not hesitate to follow the shibboleth of " molecular motion " to its final goal, but he must beware of being blinded by any artifice of method to distinctions which lie


beyond it. He accepts the authority of the chemist and the physicist for the fact that the primary elements of the organism are identical with those found in inorganic matter, and with them repudiates the notion that life is " a force having no cojinection with primary energy or motion^ But no one disputes the fact that the living organism represents special combinations of matter and force, and displays phenomena which find no parallel in non-living matter. Hence it does not appear at all irrational to conclude that vital phenomena are the manifestations of special forces, resultants of course, and yet quite unlike the elementary forces front which they are derived.

If from the same elements by different chemical combinations, we get new substances, differing widely inter se in their chemico-physical constitution, and totally unlike their primary constituents, then why not new forces in the same way } The chemist and the physicist agree in referring the differences of substances to a dynamical cause, and their mechanical conceptions do not prevent, but compel them to ascribe a specific energy to each different atom and molecule. In spite of the tendency of physical thought to regard " all matter as one and all energy as one," chemistry and physics are built up on the assumption that chemical elements are unlike, and that in different modes of combination is given the basis for that infinite variety of substances with which we meet. If the primary energy is in each case called by the same name, "polarity," it is nevertheless understood that these polarities are as unlike as the elements which manifest them. It is just here that we see the foundation for those qualitative distinctions which, in the mind of the biologist, must ever overshadow in importance the physicist's factors of quantity and motion.

All physical explanations, no less than biological, lead ultimately to the conception of intrinsic forces. The chemist's unit, like the physicist's, is the embodiment of energy. From a comparatively few atom-energies an endless number of molecule-energies are built up ; these aggregate in units of a higher order, some statical, others dynamical ; and so on through what Nageli calls micelles and micellar aggregates, until we arrive at the living cell. In this ascending series each new aggregate represents a unit, the individuality of the parts being merged in that of the whole. The grounds for distinguishing various

46 WHITMAN. [Vol. II.

organic, physiological, and biological units, are thus of the same general nature as those which compel us to discriminate between physical and chemical units. Of course these higher units combine both atomic and molecular structure ; but they have superadded to, and including this, a structure as a whole, which is entirely ignored in the expression "molecular aggregates." As they result from the union, not of simple or complex molecules, but of complex molecular groups, their structure may be said to be at least as widely separated from the molecule as this is from the atom. The power which such a unit represents as a whole is not the same as the powers represented by its constituent elements when uncombined, nor is it the sum of these several powers. Derived from them, and yet wholly unlike them, as water is something totally unlike its chemical elements or any simple mechanical addition of these elements.

It is precisely this point which is so persistently ignored in all so-called physico-chemical theories of heredity. And yet all analysis and all observation leads to the conclusion, that molecular structure is not directly responsible for vital phenomena. In claiming that " physiological units " have something higher than molecular structure and power, I am not treading on ultra-scientiiic ground, but following the course already sanctioned by chemistry and physics, and the only one which can ever reconcile physico-chemical and biological conceptions.

Admit — what no one denies — that a molecule is totally unlike its constituent elements, that its energy is unlike that of its atoms, taken individually or collectively ; and further, that simple molecules, without losing their structural integrity, may unite to form complex molecules, and we have only to carry the same process a few steps farther to reach those units whose integral structure is no longer adequately described as m.olecular. If analysis fails to discover a physiological bond which is capable of binding molecular aggregates into units of the vital order, its failure must be attributed to imperfection of methods, for observation bears constant and unvarying testimony to its existence. If analysis and observation combine to show that whatever force an organism expends is the correlate and equiva'lent of force taken into it, and if chemical and physical processes underlie all vital phenomena, it does not by any means


follow that there is any identity between vital activity and physico-chemical forces. The power which causes chemical elements to combine is not identical with the power which results from their combination ; nor is the power which breaks down a chemical compound identical with the powers of its separate elements.

The views here enunciated are not contradicted by the longestablished fact, that the laws which regulate the formation of chemical compounds are the same for both organic and inorganic bodies. The position taken does not affirm that organic compounds differ from inorganic either in material constituents, or in the forces which hold these constituents together. What we do affirm is this : We cannot stop with the most complex molecules revealed or revealable by chemical or physical research ; we must pass from organic to living, organized matter, not by the supervention of new laws, but by ultra-chemicophysical, or chemico-organic combinations, which are absolutely beyond the highest possibilities of chemical analysis. Inability to define these higher modes of combination is no reason for doubting the testimony of all our senses to their existence. And why should we expect chemical research to bring any positive confirmation of their reality, when all chemical analysis presupposes conditions which are the absolute negation of vital conditions } Does self-stultification ever become more complete than in the assumption that vital forces and conditions are discoverable precisely there where they confessedly do not exist .'* Or is it rational to conclude that, because vital conditions have arisen from non-vital, the exclusive study of the latter will reveal the former . So long as the chemist's methods debar him from the study of physiological modes of aggregation will he be impotent to divine the links which connect molecular motion with sensibility, and just so long will "physiological chemistry" remain a delusive misnomer.

A complex aggregate of atoms, so bound together by mutual affinities as to represent a physical unit, possessing, as a whole, properties and powers derived from but unlike those of its constituent elements, and existing by virtue of, and only during the maintenance of, the chemical connexus of these elements, is a conception which may be carried straight forward up to the cell. The living cell may be regarded as a system of very com

48 whitman: [Vol. II.

plex chemico-organic units, bound together by subtle chemicophysiological bonds, and displaying in their collective capacity functions and powers which are entirely foreign to them as individual and isolated elements, and which are therefore indissolubly identified with the physiological connexus or consensus.

Vague and unsatisfactory as such a view may appear, and as the best possible view must be from the limitations of our knowledge, it may yet contribute something towards a clearer conception of what we have called the formative power of the cell. It will be sufficiently clear now that we have not in mind a phantom-form which, like a mould, impresses its shape upon plastic material ; but a power which represents the resultant of the consentient reactions of indwelling forces. Such a power declares itself in every living organism and in every developing germ.

The action of the formative power has often been likened to the architectural power displayed in crystallization ; and if the essential distinctions are kept in view, such a comparison is justified by one or two very instructive analogies. If the physicist is not compelled to recognize a special crystallizing force, he is at least unable to deny that a crystalline aggregate reacts upon the parts in such a manner as to determine the direction of that marvellous "constructive power" with which the molecules are endowed. When we see a crystal reproduce its lost apex ; or, as in the oft-cited experiment of Lavalle, an angle of an octohedral crystal spontaneously replaced by a surface, as the result of an artificially produced surface at the corresponding angle, we have no alternative but to infer a physical correlation of parts, under the influence of which the direction of forces is determined. So in the development of a germ, in the repair of injured parts, and in the regeneration of lost parts, the fact is irresistibly forced upon us, that the organism as a ivJwle controls tJie formative processes going on in each part. The formative power then belongs only to the organism as a physiological whole ; and it does not represent a sum or aggregate of atomic, molecular, or other forces, but results from special combinations of ultra-molecular units, and disappears as such the moment the physiological connexus is destroyed.

This idea m.ay appear, at first sight, to stand in contradiction with the fact that parts of an organism, resulting from sponta


neous or artificial division, possess the same formative power as did the undivided organism. But it must be remembered that most organisms do not admit of such division, and that in those that do admit of it, everything depends on how the division is made. The extra-capsular portion of a Radiolarian does not reproduce the central capsule, nor does the non-nucleated fragment of an infusorian regain its lost parts. Even here, then, it is not permissible to disregard the physiological correlation of parts, since both nuclear and cytoplasmic elements are indispensable to the preservation of the formative power. We still have to regard such organisms as physiological wholes, although the physiological connexus may be representable in aliquot parts.

The principle holds true of every organism, irrespective of whether the mass is divided into cells or not. The fact that physiological unity is not broken by cell-boundaries is confirmed in so many ways that it must be accepted as one of the fundamental truths of biology.



Princeton College.


This contribution to the structure of the Amphibian brain is the result of observations which began in 1879 and have extended at wide intervals over several years. The most systematic of these studies, beginning in Munich in 1886, were directed to the nerve-fibre courses, and interrupted by the discovery of the corpus callosinn, which naturally led off to the comparative study of the encephalic commissures in the vertebrata. They form the chief portion of the present paper, and have long been withheld from publication in the hope that an opportunity would arise to make them more thorough. As this seems now more distant than ever, it appeared to me best to publish them in their present shape, with the intention of attaching to each new point the degree of certainty to which the evidence entitles it. Only by carefully distinguishing between fully and partially established facts can a contribution of this unfinished character acquire any permanent value. The most carefully observed region is the medulla oblongata of Cryptobranchtis, with reference especially to the cranial nerve tracts and nuclei. The structure of the cerebellum in the Urodela has been well worked out, also the region of the optic lobes, thalami, and posterior portion of the hemispheres in the Urodela and Aunra. Much remains to be done in respect to the peripheral distribution of the component parts of the several cranial nei^ves. Only after this has been thoroughly worked out can we certainly determine the homologies of the cranial nerves and their segmental relations in the Amphibia. The present results go far enough to show that the determination of definite nuclei corresponding to definite peripheral sensory and motor areas is well within the range of possibility. In fact, the provisional character which I have given to some of the conclusions here reached is chiefly

52 OSBORN. [Vol. II.

due to the close connection between several of the cranial nerves at, or close to, their exit, which makes it necessary to follow each component bundle in continuous sections peripherad to a point where their further distribution can be traced macroscopically. This I believe is possible with several of the nerves, but has not as yet been successfully accomplished.

The genera studied include AnipJiiitma, CryptobTancJms, Necturiis, Sircdon, Proteus, Raiia, and Sire^i.

The following are some of the most important results arrived at : —

1. The determination of the chief motor and sensory nuclei of the 5th to lOth pairs, which enables me in some degree to homologize the intra-axial elements of the Vagus and Trigeminal systems, and demonstrate the independence of the Auditory system and system of the motor nerves of the eyeball.

2. The discovery of a new sensory tract and nucleus, the fasciculus communis, common to the loth, 9th, and 7th (or 8th) pairs.

3. The determination of the relations of the posterior longitudinal fasciculus (uncrossed Miillerian fibres), to the 8th, 6th, 4th, and 3d nerve tracts, and of the nucleus of the latter nerve with the fibres of the posterior commissure.

4. The passage of a portion of the descending trigeminal tract through the cerebellum and the direct connections of this bundle with the large mesencephalic nucleus.

5. In the encephalon, the determination of the direct motor tract to the prosencephalon, and of direct sensory tracts to the mesen- and diencephalon.

The principal characteristics of the Amphibian brain, as compared with that of the Sauropsida and of the Fishes, excluding the Dipnoi, are the following : The olfactory lobes are not sharply defined from the cerebral hemispheres. The corpora striata are not prominently developed. The cerebellum is always small, and is either in a primitive or degenerate condition in the Urodela. The pineal eye pierces the skull, but does not develop further.^ Finally, an important distinction is that the

^ According to Spencer it is more rudimentary in the Urodela than Anura and in the latter forms a vesicle which subsequently degenerates.


gray matter of the brain immediately surrounds the ventricles, the central gray, and with the exception of the layers of cells in the optic lobes of the A?mra, there is nothing developed in the nature of the cortical gray matter. The Urodela are further distinguished from the Annra by the small size of the optic lobes, which are barely marked off from the thalami, by the frequently extensive development of the superior commissure, and by the primitive condition of the ventriculus communis or unpaired cavity of the prosencephalon.


The division of the brain into five large segments is invariably well marked. The olfactory lobes do not appear as fully distinct segments, but form the anterior portion of the prosencephalon, sometimes separated from the hemispheres by a prominent lateral swelling or median indentation. The line of separation is much more clearly indicated in horizontal sections. The olfactory nerves arise either from the edge or base of the lobes. The lobes are broad and truncate anteriorly in the Urodela} but taper in Protetis as in the Frog. The diencephalon and mesencephalon are subequal in size in the genera with small functional eyes, such as Siredon, Nectimis, and Siren, these tv/o segments presenting the appearance of an elongate cylinder slightly compressed opposite the posterior commissure. In the blind Proteus the mesencephalon is smaller than the

' The classification adopted is from Baur's valuable memoir, " Beitrage zur Morphogenie des Carpus und Tarsus der Vertebraten, I Theil, Batrachia," Jena, 1 888. The recent Amphibia are arranged in three orders : —

Proteida. Urodela. Anura.

Fam. Proteida. Proteus. Fam. Crypiobranchidce, Cryptobranchus. Necturus. Sirenidce, Siren.

Pseudobranchus. AmphiurnidcE, Amphiuma.

Amblystomatince, Amblystoma

(Axolotl), etc. Also the CaciliidcE, Plethodontidce. DesrnognathidcE, Salamandridce. Pleurodelidce.

Baur considers the Crypiobranchidce the most central family of the Urodela, from which the Sirenidiz, Avtphittmida, and Amblystomidct: may have sprung, p. 55. I find that the brain structure fully supports this conclusion.

54 OSBORN. [Vol. II.

diencephalon. In the SalamandridcB the mesencephalon is larger than the diencephalon, and it is of much greater proportionate size in the Anura, owing to the large optic nerve nuclei. The size of the roof of the mesencephalon is in direct proportion to the functional perfection of the eyes. The cerebellum is invariably small in the Urodela and Protcida, but attains a segmental size and importance in the Amira. The metencephalon, or Medulla oblongata, forms the fifth, and invariably the largest segment.

The proportions of these segments are very similar to those observed in the Dipnoi (see Fulliquet, ^^6), but differ widely from those in the lower fishes. The great relative size of the prosencephalon in the Amphibia is, however, deceptive, since the walls are comparatively thin and the ventricles much larger than in the two mid-segments.

Surface characters of the brain, Plate IV. In carefully prepared specimens many of the more important structures of the superficial encephalic walls may be seen upon the surface.

In the hemispheres of Siren, Fig. 5, are observed nerve tracts from the postero-lateral regions. Between the hemispheres posteriorly lies the large supraplexus which covers the roof of the ventriculus communis, Fig. 8, and in most of the Urodela extends far forwards. The inner surfaces of the hemispheres are in close contact with each other in the Urodela and Proteida, but I have observed no actual union of the olfactory lobes such as is found in the Amira. The lamina terminalis, t, Fig. 3, extends well forwards, and is sharply defined in the ventral aspect.

The characters of the roof of the Diencephalon are very well marked. Just behind the supraplexus are two oval or round swellings, which represent the ganglia habenarum with distinct median contours, but closely applied to each other in forms in which the supracommissura is well developed, such as Menopoma, Amphiuma (Osborn, '83 and '84, Fig. 4), and Axolotl, Fig. i. Arching between them and spreading beyond, into the thalami, is a grayish band which probably represents the supracommissura. This commissure underlies the pineal process, and is not seen upon the surface in Nectiirus, Figs. 2 and 8, but is apparently present in Proteus. Immediately behind this the whitish plates connected with the ganglia diverge,


leaving a slit-like, oval or triangular space, in the centre of which lies the proximal portion of the pineal stalk or processus pineal is, /;/. Close behind this a white or grayish streak represents the postcommissura. Upon the floor of this segment is an oval space, bounded by a whitish area which represents the recessus opticus, or proximal portion of the primitive optic stalk, which recess in Nectiiriis and Proteus extends directly into the centre of rudimentary optic nerve. There is a striking similarity, which may be merely superficial, between this area and that surrounding the processus pinealis. The most prominent features of the floor are the infundibular lobes, probably homologous with the lobi inferiores of the Teleosts, which terminate in the large hypophysis. These lobes extend anteriorly into the flattened floor of the 3d ventricle, and laterally into the sides of the thalami. Above them, the cerebral peduncles diverge around the infundibulum into the lower portions of the thalami.

The roof of the Mesencephalon is distinguished by a longitudinal grayish line, probably caused by the thinning of the wall ; on either side of this is a whitish tract the meaning of which is unknown. Posteriorly the roof thins out into the valvula, a triangular area from which arises the 4th nerve. This nerve is apparently atrophied in Proteus. The slender band representing the cerebellum extends across behind this, continuous at the sides with the lateral fold of the medulla.

The figures do not represent the actual condition of the Metencephalon, since the metaplexus has been removed The metacoele, or 4th ventricle, is shallow and widely open in the Urodela, and varies considerably in shape. In Proteus the lateral edges are approximated opposite the loth nerve. On either side of the longitudinal sulcus, in the centre of the floor, are observed two long white tracts shining through and giving off lateral branches opposite the exit of the VII., VIII. nerv^es. This tract is the posterior longitudinal fasciculus, and can be very clearly seen in well-preserved specimens. On the ventral surface is seen the anterior sulcus continuous with the fissure of the cord. It diverges at two points in the medulla of Proteus, the meaning of which is not known. I have not observed in the floor of the 4th ventricle of the Urodela any trace of the foldings seen in the medulla of Rana


(Osborn, 'S6, Fig. 2, text), which I have suggested may be remnants of the embryonic neuromeres. The origin of the cranial nerves is described elsewhere.

The Ventricles of the Brain. A comparison of the encephalic ventricles of these genera, in longitudinal section, brings out some interesting facts, Figs. 7, 8, 9. CryptobrancJius approaches the piscine type, and Rana the Sauropidan type : they are at opposite ends of a series. The roof of the Cryptobranchiis brain is nearly straight in the mid-region ; in Rana it is much folded, so that the cerebellum and postcommissura are brought together ; these folds and the lateral outgrowth of the optic lobes produce the diverticulum of the mesocoele, the optocoele (Wilder), and the tori-semiculares in the walls. The position of the trigeminal nucleus affords a means of comparing the optic lobes of the Urodela and Anura (see below). A still more important difference is in the fore-brain. In Rana the cerebral commissures lie in the lamina terminalis proper; there is no extended ventri cuius communis. In the lower Urodela (tht SalamandridcB have not been examined), the commissures traverse a fold of the floor, and there is a large ventriculus communis in front of this, a, Fig. 7, Necturiis shows a remarkable development of the ventricular plexuses, which extend well into the lateral ventricles and backwards nearly to the cerebellum. The elevation of the floor which contains the corpus callosum and anterior commissure is very similar in position to the commissural band between the corpora striata in the Teleosts. (See Appendix, Note 7.)

Segmentation of the Brain.

Under this head we may consider the larger encephalic segments and pass by the neuromeres, which, as recently demonstrated by Orr {%'j, p. 335), have a special significance with relation to the origin of the cranial nerves, exclusive of the first and second pairs. The neuromeres and central nuclei, when fully investigated, will undoubtedly enable us to establish a closer comparison between the brain and spinal cord segments than can ever be drawn from the peripheral distribution of the cranial nerves.

As above pointed out, the chief points of contrast between the brains of the Aniira and Urodela are the large optic lobes


and cerebellum of the former. The cerebellum is generally found to have a direct relation to the size and activity of the limbs; it is extremely small in the limbless forms, intermediate in the Salamanders, and so on, increasing to the Aniira. Are we to consider the Urodele cerebellum as in a degenerate or in a primitive condition? The limbs of CryptobrancJius (Mcnopomd) are well developed and fully functional ; they are, moreover, in the most primitive condition known among the Amphibia (Baur, '88, p. 55). If the relation above mentioned holds good here, we may infer that the cerebellum is also in a primitive condition in the central Urodela phylum. It may be considered in a degenerate condition in some of the limbless types.

This correlation has a direct bearing not only upon the question of the number of the larger encephalic segments in the Amphibia, but upon the origin of this segment. By many authors the cerebellum is considered a distinct segment equivalent, for example, to the mid-brain. By others it is considered as a portion of the roof of the hind-brain. The hypothesis I offer is that the cerebellum is primitively intersegmental, and secondarily acquires a functional importance equivalent to that of the other segments, (Osborn, 'Zj, p. 940).

The cerebellum of Cryptobj'anchiis (see also p. ) is chiefly composed of decussating tracts, passing on the one side into the lateral regions of the medulla, on the other into the mesencephalon. It may even be questioned whether we have here the essential elements of the cerebellum, the structure is so extremely simple. It has been stated that the posterior commissure, which invariably marks the dorsal boundary between the dien- and mesencephalon, is not a commissure in the strict sense of the word, but consists of fibres from the two tegmental tracts ^ decussating to the opposite side of the brain. Similarly, the superior commissure, described independently by Bellonci ('81) and myself ('84, p. 268), consists of fibres passing across the roof of the third ventricle from the diencephalon to the opposite prosencephahc segment {see scs in all the fig ^ With special relations to the tegmental tracts (Pawlowsky) and the nucleus of the third nerve (Darkschewitz).

58 OSBORN. [Vol. II.

ures). There is thus a striking similarity in the fibre courses of these three dorsally decussating tracts. In Cryptobranc/ms, in which we have the most primitive type of brain thus far observed among the Amphibia (Plate IV., Fig. 7), these three tracts are nearly subequal, the superior commissure containing the largest proportion of fibres.

This anatomical evidence for the serial homology of these commissures is supported by the facts of their embryological development. I have followed these stages in the frog, in which the superior commissure is extremely reduced ; it would probably be much more clearly shown in CryptobrancJms embryos. These commissures develop nearly, if not quite, simultaneously with the anterior commissure, at the period immediately following the constriction of the neural tube into four vesicles. In the dorsal median line this constriction is clearer between the two posterior segments than between the first and second (compare cbl and pern in Cut i with scvi)} But in horizontal sections these three constrictions are equally great, Cut 2. It is noteworthy that the floor of the neural tube, which evidently has no relation to these dorsal commissures, is also the only region in which from the first there is no constriction between these vesicles, the constrictions occurring, first, at the point of the cranial flexure ; second, opposite the anterior commissure, acm.

The inference to be drawn from these facts depends largely upon the question whether there is really a serial homology between these commissures in their primitive condition. If not, there remains considerable ground for the supposition that the intersegmental folds are lines of retarded growth in the sides and roof of the neural tube to be traversed at an early period by the commissures.

In the last number of this Journal, Orr ('87, p. 347) has contributed valuable observations on the lizard's brain, which certainly support, and possibly extend, this hypothesis to embrace also the anterior commissure. He finds, what I had not observed, that the fibres of these commissures are possibly continuations of the primitive lateral longitudinal fibres, although not positively observed to enter them, concluding as follows : "The superficial position of these three commissures, anterior, superior, and posterior, their similar connections with the lateral

1 See Appendix. Note I.


bands, and their relation to the constrictions of the brain, suggest at this period a striking homology between them." (See Plate XVI., Fig. 62, F.) This does not include the cerebellum, the tracts of which he has not followed. It is true that the fibres beneath the fore-brain branch from the lateral longitudinal band in much the same manner as do those passing to the region of the other commissures ; but I cannot at present adopt his view that they represent the anterior commissure, because the development of this commissure, as I have found it in the Amphibia and Mammalia, indicates that it is strictly commissural, and not the decussation of a longitudinal tract, as must be inferred if its fibres pass directly into such a tract. Immediately beneath the anterior commissure in the Amphibian forebrain, I have observed fibres decussating from longitudinal bundles to the opposite hemisphere (op. cit., Fig. 11, Plate XIV.), which probably represent those attributed to the anterior commissures by Orr (op. cit., p. 346).

Upon these grounds, pro and con, the hypothesis may be restated as originally: that the early constriction of the brain roof which gives rise to the four vesicles is for the accommodation of three nerve-fibre tracts decussating dorsally, viz., the superior and posterior commissures and the cerebellum, which in their primitive condition have a serial homology.

If in some of the lower vertebrates, e.g., the Urodela, the cerebellum is intersegmental, and in other vertebrates, lower and higher, it becomes equivalent to the other segments, it merely accords with what we may conjecture as to the probable evolution of the encephalic segments, that they were not primary features of the vertebrate brain, but were defined secondarily, with the concentration of certain groups of func• tional centres at certain points. The neuromeres are probably remnants of the true primitive segmentation.


Exit from the Brain.

I. The olfactory nerves arise from the lateral or infero-lateral portion of the olfactory lobes and are distributed in the usual manner in the olfactory sac, as shown in the figures of Proteus. They vary little in size throughout the Amphibia.

6o OSBORN. [Vol. II.

II. The optic nerves are, as a rule, very much reduced in the Urodcla, and are in a degenerate condition in the Proteida. Both in the Proicits and Nectiiriis the primitive epiblastic stallc of the optic vesicles is persisent. In Proteus and in some examples of Necturus the lumen of the stalk is persistent in the adult condition.^ This lumen opens into the ±-like expansion of the bottom of the 3d ventricle, which is almost invariably found just in front of the chiasma, and demonstrates that this recessus opticus is the proximal portion of the primitive optic stalk.

III., IV., VI. These nerves require no special comment. Their development is naturally parallel with that of the 2d pair. I have made a doubtful observation of the presence of the 3d pair in Proteus^ but have seen no trace of the 4th and 6th, although it is quite possible that their rudiments will be found in sections. The 3d and 4th pairs pass off independently. The 6th pair, according to Fischer ('64, p. 125), unites with the Trigeminus. I have not been able to determine whether this passes into or through the Gasserian ganglion or unites with the 5th beyond it. The more difficult problems arise with, regard to the homologies of the Trigeminal, Facial, and Acoustic nerves, which can only be settled by obtaining serial sections through the nerves, brain, and skull in situ. This arises from the fact that the Trigeminus is always reinforced by a branch of the Facial, and that the latter and Acoustic are given off so close together that it is difficult to determine into what bundle the minor roots of each pass. A clear idea of the exit of these nerves, as found in Cryptobranchus, is given in Fig. 6, which is reconstructed from the sections, while the relation of the nerves to the encephalic tracts is given in Fig. 21 ; see also Figs. 15 to 18.

v., VII. The main division of the Trigeminus is given off as a single root from the antero-lateral region of the medulla, V. i, 2, and it is reinforced by two smaller roots, V. 3.^ The Facial

1 This fact when observed was communicated to Professor Kupffer, who made use of it in a paper upon the relation of the growth of the optic nerve to the stalk. The epithelial stalk in long sections extends outwards from the recessus opticus as far as the nerve can be followed, and is peripheral, the fibres lying at the side.

2 Fischer described four branches from the Gasserian ganglion, instead of the usual three, in Siredon (op. cit., p. 128) : i. To the skin of the nose region j this he be


arises from two closely applied roots, V.-VII. ti and /, which may be described as the upper and lower roots. The upper root apparently passes directly forwards as the reinforcing branch of the Trigeminus, I may anticipate by saying that this is probably to be regarded as a detached sensory root of the 5th, and not as properly a part of the 7th. But it is, thus far, somewhat uncertain whether some of the fibres of the lower root do not also join the 5th. The lower root passes directly outwards, and either wholly or for the most part unites with the 8th. For a short distance these roots form a single trunk, and it is at this point that the chief uncertainty arises as to their distribution. A short distance outwards the 8th turns back into the auditory capsule, and the 7th extends and divides into two branches, R. mentalis and R. alveolaris, then into four (Fischer, op. cit.,

p. 135)-'

VIII. The 8th springs from four roots. The most anterior root, VIII. I, is small, and is given off immediately below and behind the lower root of the 7th. Behind this is the large root of the 8th, VIII. 2, and this is reinforced by another small root, arising somewhat internal and posterior to it, VIII. 3 and 4, by two small branches. This root may possibly join the lower division of the 7th, this uncertainty being expressed in Fig. 21 by its double designation, VII. -VIII.

IX., X. The Glossopharyngeus arises by three roots,

IX. I, 2, 3, which unite to form a single branch, passing back into the ganglion of the Vagus. Behind these, two more roots,

X. a and b, unite to form a single branch, also passing back to the ganglion. Somewhat internal to these two roots, and extending at about equal intervals backwards, arise five small roots, which unite to form the third main branch of the Vagus system, X. 1-5. Some distance behind these is given off the 1 2th pair, on either side of the anterior fissure, much nearer the median line than any of the preceding nerves.

The above description of the exit of the 5th to loth pairs, applies especially to Cryptobranchus, but so far as the main branches are concerned holds good of the other genera as well.

lieved is formed from the fibres coming from the 7th. 2. Ramus nasalis. 3. Ramus maxil. superior. And 4. Ramus max. inferior. I have observed but three main branches from the ganglion. See Appendix, Note 2. ^ See Appendix, Note 2.

62 OSBORAT. [Vol. II.

Proteus presents the widest departure from the common type, inasmuch as the Trigeminus and upper root of the Facial are given off very close together, and, so far as observed, but two branches spring from the Gasserian ganglion. The 9th and loth pairs arise some distance behind and are followed by the 12th, which springs from two small roots and has the same position as the succeeding nerve, as an anterior spinal nerve root (Fig. 3).

The Medulla Oblongata of Cryptobranchus,

General Structure. I have not made a close examination of the entire section of the medulla at different levels, having devoted most attention to the lateral regions, which are especially connected with the tracts and nuclei of the cranial nerves. As we would naturally anticipate, in an animal of such low organization as CryptobrancJius, the transition from the spinal cord to the medulla is a gradual one, but it by no means follows that we can readily homologize the medulla and spinal cord sections.

As we ascend from the level of the loth pair (Fig. 10), at which the general arrangement of the gray and white matter of the cord is fully preserved, we find the tracts and nuclei of the posterior nerves, and the deep ascending tracts, which are at the extreme lateral limits of the medulla, are gradually thrust downwards by the nuclei and tracts of the more anterior nerves. Thus the nuclei of one pair of nerves are overlapped from above, and then replaced by the nuclei of the pair in front of them, and so on. This superposition of the posterior by the anterior nuclei is exclusively in the lateral portions of the medulla, and none of the nerves between the 12th and 6th pairs arise from the region corresponding to the anterior horn of the gray matter of the cord.

The central region of the medulla is in general composed of five main areas, i. At the floor of the ventricle on each side is a mass of sensory cells. ^ 2. Below these, scattered along the whole medulla, are multipolar ganglion cells, which are especially large and numerous opposite the exit of the Acusticus. 3. On either side of the sulcus are the large fibres of the posterior longitudinal fasciculus, and below these are the Miillerian fibres. 4. Below this, in the median line, are decussating fibres and the

^Nucleus centralis, Stieda.


decussating processes of the ganglion cells. ^ 5. The remaining area, which makes up the main central region of the medulla, is composed of the sensory and motor tracts.

The area of large ganglion cells, as shown in Figs. 10, 11, and 12, is continuous with the cells of the anterior horn. It is also clear that the main motor and sensory tracts which occupy the anterior and lateral columns, at the level of Fig. 10, are thrust down, with the ascending Trigeminal tract, to form the area 5. With the exception of the posterior longitudinal fasciculus, none of the nuclei or tracts of the cranial nerves below the 6th pair occupy the central region of the medulla. It follows that the lateral region of the medulla is exclusively composed of the cranial nerve tracts and nuclei and the central region is composed of the main sensory and motor tracts of the cord and special sensory and motor centres. These regions are pretty sharply defined from each other. The lateral region above the level of the Vagus is bounded by the ascending Trigeminus tracts.

The Tracts and Nuclei, Plate V. A general description of the fibre tracts and nuclei is necessary to introduce the special study of the origin of each nerve. In the section opposite the 12th nerve, we find the columns of the cord and horns of gray matter well defined. Between the posterior and lateral columns, at the extreme periphery of the cord, is a small bundle, the trigeminus ascendens, 5/.^ On the posterior horn is another small round bundle, not hitherto to my knowledge described, which from its common relations to a number of the cranial nerves we may call the fasciculus communis. Opposite this, upon the lower side of the posterior horn, is another round bundle, which from its relation to the Vagus is believed to represent the fasciculus solitarius of Lenhossek, fs. Above this section, Fig. 1 1, opposite the exit of X. 4 and 5, is seen a small nucleus contributing to the fibres of the fasciculus communis, /<:;/. ; then the nucleus and ascending bundle of the 9th, then the ascending trigeminal tract ; below this is the exit of a portion of the fasciculus communis, and the nucleus and a root of X. 4. In the centre of the medulla is the posterior longitudinal fasciculus, which is now a more compact and well-defined tract than in

1 Connected with the crossing Miillerian fibres, (Ahlborn).

64 OSBORN. [Vol. II.

the lower levels, and increases in distinctness as we ascend. The nucleus and bundle of the Glossopharyngeus increases in size, and the large ascending bundle of the Trigeminus is thrust downwards, Figs, ii and 12; at the level of the exit of the 9th pair, the sensory nucleus and upper and lower roots of the 7th come into view, as well as a third nucleus of the 9th, gn, which takes the place of the internal nucleus of the loth. The fasciculus solitarius has entered the loth, also a large bundle of the fasciculus communis,/^', and a fresh bundle of this fasciculus is now forming, yj;". It is noteworthy that although the two bundles of the Facial and the Trigeminal root are well formed, there is apparently no trace of the Auditory. Above this level, however, the auditory tract, 8/, forms rapidly between the ascending 5th and the lower 7th tracts. It has apparently two nuclei. Fig. 14, and encloses the fasciculus communis. The Auditory is given off by the union of four tracts, and above this the 7th arises by its two roots. The upper and lower nuclei of the 7th are then replaced by the large sensory nucleus of the 5th, which rapidly reinforces the already large ascending tract. A glance at the series of figures will show how the sensory nuclei of the 9th, 7th, and 5th pairs replace each other in regular succession, while the 8th, so far as observed at present, belongs to a separate system. A schematic arrangement of these nuclei is shown in Fig. 21, the level at which each transection is taken being indicated at the side.

GlossopJiaryngeiis and Vagus. These nerves are closely related to each other; the glossopharyngeus arises dorsally, contributing mixed fibres, and the vagus ventrally, also contributing mixed fibres. The definite homology of these roots with the 9th and loth pairs of the higher vertebrates is uncertain.

The posterior roots of the Vagtis, X. ab, 1-5, of this system, arise from at least three sources, i. The fasciculus solitarius. 2. The fasciculus communis. 3. A special nucleus in the floor of the 4th ventricle. The two ascending fasciculi are both found in the region of the posterior horn, (i) The fasciculus solitarius^ appears to arise from the

1 This bundle in the human medulla contributes also to the 9th pair, and extends to the level of the 8th Cervical (Krause). Hoffmann- Rauber, Lehr. d. Anat., 1886, p. 397


lateral columns of the cord ; at least, no special group of nerve cells has been observed contributing to it. The fasciculus communis, however, is closely applied to a nucleus of small sensory cells, _/i:«., from which it appears to be reinforced. The fasciculus solitarius enters the two posterior roots. (3) The three middle roots arise from the nucleus in the floor of the ventricle, 10 «, which is apparently composed of motor cells. (2) The two anterior roots, a, b, arise from the fasciculus communis, which is thereby reduced to a small bundle.

The GlossopJiaryngeiis arises from four sources, i. The fasciculus communis, 2. A large sensory nucleus. 3. A nucleus of doubtful motor cells. 4. A motor nucleus in the floor of the 4th ventricle. (2) The sensory nucleus and tract is found at a low level opposite the exit of the posterior roots of the loth. This tract 9 / rapidly increases in size from the nucleus in the dorsal folds of the medulla, which is undoubtedly sensory. Anteriorly, this nucleus is replaced by the sensory nucleus of the 7th, and below it, in the angle of the medulla, (3) is a small nucleus of cells of different character which are probably motor, 9 mn. These doubtful cells have the same position as the nucleus of the lower root of the 7th pair, which is unquestionably motor. The fibres from these two nuclei constitute a single tract, and form the anterior root of the 9th nerve. (4) The motor nucleus in the floor of the ventricle, 9 n, is in exactly the same position as that of the loth pair, and is composed of spindle-shaped, probably motor, cells. The fibres from this nucleus form the posterior root, (i) The fasciculus communis forms the middle root.

The 9th and loth pairs have in common, first, the continuous and closely similar sensory nuclei and tracts arising from them which constitute the main sensory supply of the former nerve ; second, the special motor nuclei in the floor of the ventricle ; third, the fibres of the fasciculus communis, also sensory. The element of doubtful motor cells which is present in the 9th and absent in the lOth, may subsequently be found in the latter. The 9th, however, apparently lacks the fasciculus solitarius. The special motor nuclei, 9 n and 10 11, are seen to correspond in position to the lateral portions of the anterior horn, i.e. to the region dorsal to the extremity of the anterior horn.

The Acusticus. The nuclei and tracts of the 8th pair,

66 OSBORN. [Vol. II.

as a nerve of special sense, do not in any degree repeat the features of the 9th and loth, which to a certain extent are witnessed in the 5th and 7th. Of the fo?ir distinct roots of the 8th, I am in some doubt whether the first and last may not be given off out of place and ultimately join the 7th. The last of these roots may, however, at present be described as a portion of the 8th. (4) The fibres of the fourth, or posterior root, are received from the posterior longitudinal fasciculus, as beautifully shown in the floor of the ventricle of Siren and in sections of this level. Fig. 15. They suddenly turn out from the ascending bundle of the fasciculus and traverse the medulla at an oblique angle to join the nerve by two small roots. Whether studied in transverse or horizontal sections, this is clearly an ascending fasciculus, and connects the nerve with the lower regions of the medulla or spinal cord, not with the encephalon. This is an important point. (3) Shortly above the exit of the 9th pair. Fig. 14, a nucleus of large pale ganglion cells appears, which is quite distinct from the underlying motor nucleus of the 5th. This is the second source of supply of the auditory fibres, and probably corresponds to Deiters nucleus; it is sharply defined from the surrounding cells. (2) In the lower angle of the medulla is also seen a group of small cells which appears to contribute to this tract of the 8th, 8 sn, but cannot be positively determined, (i) On the dorsal side of this tract is the bundle of the fasciculus communis, which is here double, one-half, VIII., passing outwards to form the anterior root of the 8th. (5) As the nerve is given off, a large tract is seen which rises external to the sensory nucleus of the Trigeminus and constitutes the 8th encephalic tract, which is indicated by a contour above /r'" in Fig. 17, and by dotted lines in Fig. 21. It is possible that this tract sends fibres to the mesencephalic sensory cells which may thus form a superior nucleus.^

It is seen that, whatever may prove to be the peripheral distribution of the fibres of the fasciculus communis and posterior longitudinal fasciculus, whether to the 7th or 8th, two facts remain : first, that the 8th arises ventral to the 7th, although a purely sensory nerve; second, it is inserted in the centre of the Facial-Trigem


^ It ascends to the midbrain, see Cerebellum


inal system, with no apparent homology in the arrangement of its nuclei to either.^

There is a strong ground for supposing that the fasciculus communis, although primarily given off with the Auditory, subsequently joins the Facial, since, first, it adjoins the lower motor bundle of the Facial, and second, it also joins the motor bundles of the loth, 9th, and probably of the 5th, nerves, which fall into the same general category as the 7th. Still, this question can only be settled by following this tract peripherad beyond the passage of the 8th into the auditory capsule.

The posterior longitudinal fasciculus is stated by some authors to send a bundle to the 7th ; I do not know that this has been positively determined. Spitzka regards it as highly improbable that this bundle enters the Auditory proper. It forms a close union with this nerve, VII.-VIIL, 3-4, soon after its exit, at which point the two facial and four auditory roots have the relations, shown in Fig. 16, to each other and to the large Auditory ganglion. It seems improbable, from these relations, that this inferior bundle should unite with the facial bundle, and further, both Ahlborn^ ('83, p. 262) and Fulliquet ('86, p. 81) have followed portions of these large fibres into the Auditory so that this tract is almost without doubt a portion of this nerve. There is also much similarity between the disposition of the 7th, 8th, and 5th tracts in CryptobraiicJms and Petromyzon (op. cit., Taf XIV.).

Facialis and Trigeminus. As already stated, the nuclei and tracts of the upper and lower Facial bundles arise at a low level, opposite the exit of the 9th pair, somewhat lower, in fact, than represented in the scheme, Fig, 21. There is no difficulty in recognizing the upper tract as sensory, since it springs from the outer fold of the medulla in precisely the same manner as

1 Dr. E. C. Spitzka, to whom I am indebted for some valuable suggestions as to the identification of these tracts, questions the determination of the upper bundles, 7 u-l, in Fig. 15, as parts of the Facial, on the ground that the ventral position of the Auditory reverses the usual order. There can, however, be no doubt that these belong to the Trigeminal system, from the fact that 7 u passes directly to the Gasserian ganglion (see Fig. 5).

2 Ahlborn distinguishes three groups of Miillerian fibres in Petromyzon, viz., lateral uncrossed, median crossed, and median uncrossed. The same groups are apparently present in Cryptobranchus. The first named corresponds in position with the posterior longitudinal fasciculus, and alone enters the Auditory root. See Appendix, Note 3.

68 OSBORAT. [Vol. II

the dorsal roots of the 9th, and from similar sensory cells, 7 S7i. The lower tract begins at about the same level, and is distinguished from the upper by its larger fibres and by a distinct contour, but there is some doubt whether it is a motor tract. The nucleus, 7 mn, is at the angle of the medulla in the same position as the supposed motor nucleus of the 9th, and is composed of slightly larger cells than the sensory nucleus, but they are not distinctly of the motor character. As represented in Figs. 15 and 16, some of them are bipolar; but they have not the characteristic multipolar shape and large nuclei which are seen in the motor nuclei of the 5th, 5 7iin, and loth. There can be little question, however, from the peripheral distribution of this tract that it is motor.^ The Facial is thus comparatively simple in its origin : unless it is reinforced by the fasciculus communis or posterior longitudinal fasciculus, or both, it has, so far as observed, no ascending bundle, and in this respect differs from the 9th, loth, and 5th. It has, however, a small descending tract, rising towards the cerebellum above the descending 8th.

The Trigeminus. Five, or possibly six, tracts are observed to enter the Trigeminus, as follov/s : i. The ascending tract from the cervical region, reinforced by, 2, fibres from the deep motor nucleus, representing two tracts. 3. Fibres from the sensory nucleus. 4. The descending tract from the mesencephalic nucleus. 5. The direct encephalic tract.

(i) The ascending trigeminus is first observed at the periphery of the cord, between the lateral and posterior columns, and increases rapidly in size, probably by accession of fibres from the lateral columns, so that at the exit of the 9th pair it is the largest of the tracts. (2) At this point we first observe fibres entering this tract from the motor nucleus, 5 inn, which immediately adjoins the motor nucleus of the 9th pair, and is external to the group of large ganglion cells which are continuous with those of the anterior horn. This nucleus increases in size to the level of the exit of this tract. (3) No sensory nuclear fibres enter the tract below the level of the exit of the 7th pair ; at this point the sensory nucleus appears, 5 sn, continuous with that of the 7th, Fig. 17, and the mgtor nucleus, which is throughout composed of typical motor cells, immediately adjoins it ventrally. The sensory nucleus is very large,

  • See Appendix, Note 2.


and extends forwards beyond the level of the cerebellum. At this upper level, Fig. 18, the motor nucleus extends towards the centre, and sends fibres directly into the nerve root. (4) The descending tract is a large bundle from the more central region of the medulla. It rises obliquely, 5 /,* from the root, Figs. 17 and 21, marked by its large, darkly stained axis cylinders, and is joined by another tract of similar fibres, the origin of which I have not observed, Fig. 21. The joint tract, thus formed, is very conspicuous and is followed without much difficulty. Opposite the cerebellum it splits into tv/o bundles (see Figs. 23 and 24, 5 t'^). One of these passes into the cerebellum. Fig. 19, and, without crossing, enters the roof of the optic lobe at one side of the median line. The second bundle, 5 t^\ passes forwards, and scatters into rays over the whole wall of the optic lobe, nearly as far forwards as the posterior commissure. These two bundles undoubtedly arise in the same manner from the large cells which constitute the nucleus.

The trigeminal mesencephalic nucleus has a remarkable extent in Necturus, reaching in two large masses of solidly packed cells, on either side of the median line of the roof, from the cerebellum to the posterior commissure. The posterior portion of this nucleus is shown, in Fig. 19, in which the cells lie, two or three deep, above the central gray of the ventricle. The cells are multipolar and spindle-shaped. Fig, \<^ a, with large nuclei, the main processes being directed upwards into the tectum opticum. The actual connection of these processes with the cerebellar bundle of the descending trigeminal can be observed. It is also probable that these fibres are reinforced from other cells of the roof of the optic lobes, a point which is discussed later. This trigeminal nucleus is found in all the Amphibia, but is much larger in Nectiirits than in the other Urodela, and larger in the latter than in the Amira. It does not, therefore, appear to be directly correlated with the power of sight, although it appears to have an indirect connection with the mesencephalic roots of the optic nerve.

(5) In horizontal sections a large portion, 5 /^ of the Trigeminus is observed passing forwards into the lateral portion of the medulla, but cannot be followed any great distance, owing to the similarity of its fibres with those of the main sensory

70 _ OSBORNi [Vol. II.

tract adjoining. This encephalic tract underlies the similar bundles from the Auditory and Facial, Fig. i8.

Abdticens, Trochlearis, and Oculo-moiorms. The most interesting facts observed in connection with these nerves is the close relation of their nuclei to the posterior longitudinal fasciculus and the relation of the oculomotor nucleus to the posterior commissure. Both of these observations have been made before, but in the Amphibian brain they stand out with unusual clearness. In vertical and longitudinal sections, Figs. 20 and 20 a, some of the large fibres of this fasciculus seem to pass directly into the larger processes of the cells of the oculo-motor nucleus, 3 n'. The larger processes seem to lie in the direction of these fibres; the smaller processes are partly directed towards the fibres of the posterior commissure, and actually extend into the lower portion of this commissure ; they are partly directed towards the exit of the nerve. This group is figured with the camera as observed in Ciyptobranchiis ; the posterior commissure also descends close to this nucleus in Nectums and Rami, as seen in vertical and transverse sections, Fig. 25.

The posterior longitudinal fasciculus also has close relations with the Abducens nucleus, as shown in Fig. 18, although I have not observed either in this or in the Trochlearis^ the actual continuity of its fibres with any of the cell processes. The Abducens nucleus is directly opposite the exit of the 5th pair, but unlike all the posterior nerves of the medulla, it passes out close to the middle line, more in the manner of an anterior spinal nerve root.

The Oculo-motor nucleus is double, consisting of two small groups of typical motor cells, 3 n and 3 n', the fibres of the posterior commissure descending between them. As found in NecUinis, Fig. 25, the fibres all arise upon the same side of the brain, and the nucleus is placed at the edge' of the central gray substance.

  • It is difficult to follow the Trochlearis from its nucleus to its exit. It is

stated to contain sensory fibres in the Selachians and Amphibia. Gegenbaur, op. cit. p. 49. I have only observed a motor nucleus.

No. I.]



Group B.

Group C,

VI., IV., Ill,


Special ventral motor

nuclei. Posterior longitudinal





Special sensory nuclei. (Fasciculus communis

from sensory nuclei ?.) Posterior longitudinal


General Conclusions.

A general survey of the distribution of the nuclei, shown in Fig. 21, as described in the preceding sections, points to the natural division of the cranial nerves into three groups : Group A embraces the Vagus, Glossopharyngeus, Facialis, and Trigeminus. Group B embraces the Abducens, Trochlearis, and Oculo-motorius. Group C embraces the Acusticus.

The known tracts and nuclei possessed in common by each of these groups are distributed below : —

Group A.

X.-IX., VII.-V.

1. Dorsal ^ sensory nuclei

( ? wanting in the loth) .

2. Lateral motor nuclei

( ? wanting in the i oth) .

3. Fasciculus communis 2

from sensory nuclei

(not observed in the 5th).

4. Ventral 1 motor nuclei

(wanting in the 7th) .

The nerves of Group A are apparently closely homologous, in so far as their mode of origin is concerned. The dorsal sensory nucleus of the loth, 9th, 7th, and 5th pairs shows the most marked continuity, and the nerve bundles are seen to issue directly from it. The ventral motor nuclei of the 9th and and I oth (10 « and 9 ;/) are continuous, and occupy the same region as the great ventral motor nucleus of the 5th. The fasciculus communis, first appearing in the posterior horn, and all along in close contact with a small group of sensory cells, is given off to the loth, 9th, probably to the 7th, and possibly to the 5th. The lateral motor nucleus, consisting of the cells in the angle of the medulla, contributes to the 9th, 7th, and probably to the 5th.

The absence of the ventral motor nucleus from the 7th nerve, and the large motor element in the 5th, may indicate that some

1 These adjectives, dorsal and ventral, are retained, since these nuclei are simply thrust laterad by the opening of the 4th ventricle.

2 This is upon the probability considered on p. CO, that the fasciculus communis joins the 7th instead of the 8tb.

72 OSBORN'. [Vol. II.

of the motor portion of the 7th, as found in the higher vertebrates, takes its exit with the 5th in the Urodela, while some of the sensory portion of the 5th takes its exit with the 7th, probably as the upper or sensory bundle which immediately joins the 5th.

The roots of the Glossopharyngeus seem to be in a similar manner complementary to those of the Vagus, since the former nerve receives a large bundle from the dorsal sensory nucleus, as well as from the lateral motor nucleus, neither of which bundles, so far as I have observed, enter into the loth.

Group B. The exit of the Abducens from the central region of the medulla, and the special connection of its nucleus and those of the other eyeball muscles with the posterior longitudinal fasciculus, seems to separate these three nerves sharply from the Vagus or Trigeminus system, and unite them closely into a group of their own. I am well aware that some grounds are found in the development and peripheral distribution of these nerves for considering the 4th or 6th as vagrant portions of the Trigeminal or Facial motor roots, but the internal origin certainly does not support this hypothesis ; at the same time it cannot be said to disprove it.

Group C. The Auditory has been shown to differ widely from the 9th, 5th and 7th nerves by deriving no fibres from the dorsal sensory nucleus, although a purely sensory nerve. This can be stated with considerable certainty. One of its special nuclei is also composed of cells of a unique character, viz., pale ganglion cells. The reception of fibres from the ascending tract adjoining the posterior longitudinal fasciculus further distinguishes it from the nerves of group A.^

These considerations lead to the following provisional conclusions in regard to the intra-axial origin of the cranial nerves in the Amphibia.

1°. That there is a close similarity between the disposition of the nuclei and tracts of the IX.X. and V.-VII. groups, the nerves of these groups being complementary to each other, and together apparently containing fibres from two sensory nuclei and from two motor nuclei. The extreme dor

  • See Appendix, Note 3.


sal and ventral nuclei are composed respectively of unmistakable sensory and motor cells. While one of the lateral or intermediate nuclei is composed of less distinctively motor cells of smaller size {e.g. ()mn, Jtmi), the other (fasciculus and nucleus communis) is clearly sensory.

2°. The nerves of the III.-IV.-VI. pairs form a special system with no apparent homology or connection with the motor elements of the Vagus or Trigeminus systems.

3°. The VIII. as a nerve of special sense has either no homology or an incomplete homology in the arrangement of its sensory nuclei and tracts with the sensory elements of the Vagus and Trigeminus systems.

If the fasciculus communis is found to enter the 8th instead of the 7th, this section must be modified to an incomplete homology.

I do not consider that these observations are sufficiently well tested to give sure support to theoretical deductions as to the homologies of the cranial and spinal nerve elements. It is clear that the Vagus and Trigeminus systems approach nearest the typical spinal nerve arrangement, and the apparent presence of two sets of sensory and two sets of motor nuclei in these systems is of great interest in its bearing upon Gaskell's theory of the compound nature, somatic and splanchnic, of the anterior and posterior roots. As frequently stated above, the motor or sensory character of some of the nuclei and tracts can only be determined by the study of their peripheral distribution, since the character and position of the cells themselves, as shown in the pale ganglion cells of the 8th and the small cells of the lower nucleus of the 7th, is an uncertain guide as to the nature of the fibres which spring from them. Even with this precaution it is a great advance to determine the presence of these two sets of nuclei for each of these systems.

III. THE ENCEPHALON. The Cerebellum.

As already observed, the cerebellum in the Vrodela is widely different from that in the Anura in its size and internal

74 OSBORN. [Vol. II.

structure. My observations relate principally to the former. The latter has recently been thoroughly investigated by Wlassak, in a memoir which I have not had an opportunity of studying, and independently by Koppen. /

In an earlier paper upon the Cerebellum of CryptobrancJms I showed that it receives two lateral systems of fibres on each side, one from the medulla, one from the mesencephalon. Also a central system entering the mesencephalon. The latter, which I mistakenly compared ('84, p. 266, Plate VI., Fig. 7) with the superior peduncles of the mammalian cerebellum, I now find is composed of the cerebellar branch of the descending Trigeminus tract. The cerebellum of Ajtiphiinna consists exclusively of these lateral and central tracts, containing no cells except the lining of ependyma ('83, Plate VIII., Fig. 3). The Cryptoh'anc/iHS cerebellum contains, in addition, a small mass of round cells of the same description as those composing the central gray substance of the optic lobes. It thus consists of the following elements, as sho\yn in Figs. 19, 23 and 28.

1°. Fine fibres from the'dxtreme lateral portions of the medulla. It is clear that these fibres, cbl.t', as shown in Fig. 18, do not arise from the central region of the medulla. They oan in part be followed to the point of exit of the 7th and 8th pairs of nerves. Of these the larger portion seem to come from the 8th pair.

2°. Coarse .fibres from the descending trigeminus tract. The course of these fibres has already been described on page 153, 5 t^.

3°. Fine fibres passing laterally into the mesencephalon, cbl.t; these can be followed some distance into the lateral area of the central gray of the mesencephalon.

4°. A nucleus of small rounded cells, cbl.n, in the ventral area.

The passage of the fibres from the 8th pair through the cerebellum has also been observed by Ahlborn. It is somewhat difficult to determine positively, owing to the close similarity between the adjoining trigeminal and facial fine fibre bundles. In horizontal sections the fine fibres of the Trigeminus, 5 fi, are, however, seen to pass directly forward into the lateral region of the mesencephalon. It is thus probable that the fibres entering from the medulla are mostly sensory fibres from the Auditory



nerve which decussate to the opposite side of the brain through the cerebellum, and either enter the main sensory tract or terminate in some of the cells of the mesencephalon. The descending Trigeminus tract does not decussate, but enters the .medulla on the same side.

The fine fibres of the Urodele cerebellum are in part decussating tracts of the Auditory nerve, and the coarse fibres are non-decussating descending tracts of the Trigeminus nerve.

This conclusion as to the connection with the Auditory nerve is supported by Koppen's observations upon Rana (op. cit., p. 13), and Ahlborn (p. 261). Ahlborn, however, considers that the fibres entering the cerebellum are commissural between the Auditory nuclei of opposite sides. This hypothesis is not supported by the fact that the superior and inferior tracts are subequal in diameter.

The passage from the medulla into the mesencephalon is accompanied by the compact disposition of the gray substance immediately surrounding the ventricle. Fig. 25. The fine fibres descending to the trigeminus and the great mass of fibres from the central region of the medulla become indistinguishable, and the only fibres which can be readily followed forwards in transverse sections are those of the descending trigeminal tract and the posterior longitudinal fasciculus. Anteriorly, in the region of the diencephalon, the so-called round bundle is sharply differentiated from the remaining fibres. In vertical sections, however, the tracts from the central region of the medulla can be followed with great ease, especially in the brains of Amphiiima, NectuniSy and CryptobrancJms, in which the midsegments are only slightly swollen and the sections fall in the plane of large bundles of fibres (see Fig. 30). These tracts have been carefully studied in Necturus, Cryptobranchus, and Rana, the former genera agreeing closely in all essential features.

The Nuclei of the Anterior Segments.

The main masses of fibres from the spinal cord and central region of the medulla sweep forwards either directly into the central gray of the mesencephalon, diencephalon, or prosen

76 OSBORAT. [Vol. II.

cephalon, while the secondary or more local tracts of the encephalon can in most cases be followed directly to certain nuclei. The latter are more numerous in the Amira than in the Urodela, at least they are much more sharply defined. The following nuclei have been observed in the brains of both classes {see

Fig. 31) A. The nuclei of the cranial nerves.

1°. The nuclei of the 3d, 4th, and 6th pairs, 3«, 4.11, and 6n, are nearly upon the same horizontal level in the Urodela.

2°. The great mesencephalic trigeminal nucleus extends along either side of the entire roof of the segment in many of the Urodela, but in the Aniira is much smaller and is mainly confined to the antero-lateral region (Fig. 26).

3°. Two special well-defined centres of the optic nerve fibres are observed : a. the roof of the mesencephalon on either side of the median line in the Urodela; in the A ?mra this extends into the widely expanding optic lobes, d. A distinct nucleus in the middle region of the walls of the diencephalon (corpus geniculatum).

B. The nuclei of the encephalic tracts.

1°. The most distinct of these are the ganglia habenarum, £-/i, just anterior to and partly traversed by the superior commissure. They are composed of small, rounded cells, like those of the central gray at the dorsal anterior angle of the thalami.

2°. In the floor of the mesencephalon, just posterior to the oculo-motor nucleus, is the ganglion interpeduncular e, £^, composed of very small triangular cells.

3°. Just posterior to the posterior commissure is a small dorsal nucleus, in the course of the optic tract (see Fig. 24), the connections of which are undetermined.

4°. The corpus striatum is not well defined in the Amphibia ; it consists of a mass of scattered cells, slightly anterior and ventral to the anterior commissure. Fig. 28, cs.

5°. Slightly posterior to the oculo-motor nucleus, on the same horizontal level, is a nucleus of pale bipolar ganglion cells quite distinct in character from the nucleus of the third pair, which consists of triangular deeply stained cells. This is figured, but not lettered, in Fig. 23, and is seen both in Rana and Cryptobranchus {? red nucleus).

6°. The most conspicuous nucleus in Rana is the nucleus


mag nils of Stieda. This is slightly below and in front of the cerebellum, nvi.

7°. Another smaller but conspicuous nucleus in Rana is in the wall of the diencephalon, behind the corpus callosum, n\ Fig. 29.

The homologies of some of these nuclei are rather uncertain. The nucleus in the wall of the diencephalon which gives rise to the fibres of the optic nerve corresponds probably to the middle or lateral geniculate body.

The nucleus just behind the oculo-motor ganglion corresponds closely in position to the red nucleus of the tegmentum. Stieda and Koppen have suggested that the nucleus magnus corresponds to the dentate nucleus of the cerebellum, but this is also in the position of the red nucleus and presents many points of agreement with it in its relations to the surrounding tracts ; its relations to the cerebellum are, however, by no means well ascertained.

The Main Sensory and Motor Encephalic Tracts.

There is no certain means of distinguishing the sensory from the motor tracts. In the ascending series of sections of the medulla oblongata, it has been shown that the posterior columns, which presumably contain unmixed sensory fibres, and the lateral columns of mixed fibres are thrust downwards by the superposition of the cranial nerve nuclei. As a result of this, we should expect to find the sensory tracts occupying the lateral portion and the motor tracts the median portion of the central region of the medulla, corresponding to the anterior columns of the lower levels of the cord. That such is actually the case is supported by two facts. First, it is found in Cryptobrancluis that the median region consists of slightly larger and deeply stained fibres, more of the nature of motor fibres than those of the lateral region ; and second, the fibres of the lateral region mostly terminate in the mesencephalon and diencephalon, while those of the median region, in large part at least, extend directly forwards into the prosencephalon. With this evidence I may at all events describe these lateral tracts as sensory and the median tracts as motor.

The Sensory Tracts. In successive sagittal sections of the brains of CryptobrancJms, Necturus, and Rana, we first pass

78 OSBORN. [Vol. II.

through the fibres of the extreme lateral portions of the medulla, entering the cerebellum, cblJ; then we observe large bundles of fibres from the lateral regions of the medulla, mst, ascending and spreading over the outer surface of the mesencephalon. Figs. 24 and 30. In sections slightly internal to these, the central gray of the mesencephalon comes into view, with rows of cells and bands of fibres alternating, and the continuation of the same medullary tract is observed spreading over the outer surface of the diencephalon in precisely the same manner, dst. These direct diencephalic and mesencephalic sensory tracts are beautifully shown in vertical sections of the brain of AmpJmima, in which these segments are very slightly differentiated from each other. In succeeding sections, still approaching the median line, the direction of the fibres is reversed ; from the postero-lateral region of the mesencephalon and the lateral region of the diencephalon, the main trend of the fibres is downwards and forwards, mst' and dst J The latter fibres pass directly forwards into the basal portion of the prosencephalon. The former, inst\ turn downwards, but their forward continuation into the prosencephalon cannot be so distinctly followed. The simplicity of these ascending and descending systems in the Urodcla is interfered with in the Amira by the expansion of the optic lobes, but the arrangement is the same.^

In transverse sections of the optic thalami, the dorsal portion of the cerebral peduncles is composed of a compact round bundle of fibres, (Osborn, '84, Fig. 8, /.). The origin of this is somewhat uncertain. From the fact that it is first differentiated in the anterior portion of the mesencephalon and becomes more distinct in the thalami, we may infer that this bundle is composed of the prosencephalic sensory tracts formed from these segments.

In Rana, in which the corpus callosum and ■ anterior commissure are somewhat separated, a portion of this bundle seems to pass between them (Fig. 29). There is room for error here in the fact that the basal fore-brain bundle from the medulla is reinforced by fibres both from the ascending diencephalic and mesencephalic tracts and from the infundibular tract.

^ See Koppen, op. cit., Taf. III., Figs. 27, 28.


The Motor Tracts, As the sections extend inwards the mass of fibres from the medulla does not ascend, but passes forwards directly into the basal portion of the prosencephalon as the basal prosencephalic tract (basal fore-brain bundle, Edinger). This is best followed in the Urodcla, but is also readily followed in successive sections in the Anura, Figs. 28, 29, Upon reaching the corpus striatum, some of the fibres of this tract enter this body, as described recently by Edinger (Fig. 28), while others pass directly into the inferior portion of the mantle of the hemispheres ; a third portion seems to terminate immediately below the anterior commissure : but in horizontal sections it is seen to pass to the other side.^

The tract thus consists of three parts : a, a direct bundle from the hemispheres ; b, a bundle from the corpus striatum ; c, a decussating bundle from the hemispheres. Two of these divisions were incidentally described and figured in my paper upon the Corpus Callosimt (loc. cit., Taf. XIV., Fig. %, pm, pi).

The Secondary Encephalic Tracts.

Under this head may be considered the tracts which, so far as observed, have no direct connection with the spinal cord or medulla.

Meynert's Bundle. This is a conspicuous tract in the brains of all the Amphibia, mb. It arises, in the usual manner, from the ganglia habenulse and descends beneath the superior commissure as a compact bundle of darkly stained fibres, to the ganglion interpedunculare. I have followed it slightly beyond this point in Rana, Fig. 29, but not into the medulla, as Ahlborn has succeeded in doing, in Petromyzon.

The Infundibular Tract. From the infundibular lobes this large tract of fibres ascends, it, beneath the basal prosencephalic tract, towards the hemispheres. Koppen has described it as entering the thalami. It has this appearance in Rana, but not in the Urodela, where it appears to pass directly forwards and not upwards.

The Posterior Commissure. The relations of this commissure are threefold ; first, to the oculo-motor nucleus and probably to the main sensory tract ; second, to the pale ganglion

1 This decussation is described by many authors as a portion of the anterior commissure, but in my opinion should be considered as entirely distinct.

8o OSBORN: [Vol. II.

cells behind this nucleus ; third, to the tectum opticum. As it descends, the fibres divide into two bundles (Fig. 25), of which the anterior surrounds the superior processes of the ganglion cells of the oculo-motor nucleus (Fig. 20, pes) : this connection is so close that some of these fibres seem to be actually continuous with the cells. The posterior bundle has a similar connection with the cell processes of the pale ganglion, which may in fact also belong to the oculo-motor nerve. None of the fibres of this commissure can be traced directly into the main (sensory) tracts adjoining these nuclei, as observed by Pawlowsky, although such a connection seems highly probable (tract us cruciatus tegmenti). Dorsally, the fibres of this commissure in Raiia can be clearly followed into the peripheral white substance of the tectum ojoticum, as shown in horizontal sections.

The Superior Commissure. This is much less constant in size and development than the foregoing. It is extremely small in the A)iura and apparently so in the Proteida} but is large in CryptobrancJuLS and Amphiiima. It divides into two distinct bundles, one of which descends into the inner mantle of the hemispheres, scs, and finally disappears, after bending around into the outer portion of the mantle. The second bundle descends directly along the outer wall of the thalami. These bundles are clearly seen where the commissure is well developed, and I have fully described them elsewhere ('84, p. 268, Fig. 8). One fact militates against our considering the commissure as a purely decussational system ; that is, the bundle entering the hemispheres is much larger than that entering the thalami. It forms either partly a commissural system between the posterior portions of the hemispheres and between the thalami, or partly a decussational system between the hemispheres and thalami.

Cerebral Commissures. I have seen reason to partly alter my views as to the nature of the commissures of the hemispheres which were described in detail in my paper on the corpus callosum. The more recent researches of Bellonci, with the aid of the Golgi method, upon these commissures, should be consulted.^ They show that with the purely commis ^ Necturtts.

2 I regret that I have not the opportunity at present to investigate thoroughly the interesting questions which Bellonci has raised in this valuable memoir in regard to


sural fibres, decussational fibres are intermingled. I have myself discovered that in the upper bundle or corpus callosum of Mciwbranchus there enter fibres from the diencephalon. Proteus agrees with Menobranchus in the entire separation of this bundle from the anterior commissure. I have seen no trace in the Amphibia of the fornix columns which I have found in the OpJiidia ('86, p. 533, Fig. 20), and Bellonci has figured in the Lacertilia, ('87).

Other important commissures are the infundibular commissures which connect the infundibular lobes, dorsally and ventrally, and the extensive commissure of the tectum opticum.

The Origin of the Optic Nerves,

It has recently occurred to me' that the presence of the two whitish bands seen upon the external dorsal surface of the optic lobes in many of the Urodcla is partly due to the underlying optic tracts. It is certain that these tracts in the Urodela are principally confined to the median portions of the tectum opticum. They can be followed as far forwards as the cerebellum, II. / I, Figs. 23, 24, and descend obliquely from this region to the nerve. In the Urodela there is no differentiation of cells into distinct layers.

The character of the cells of this region is shown in Fig. 2^a, which illustrates an interesting observation upon the apparent connection of the fibres either of the optic nerve or of the direct sensory mesencephalic tract with the fibres of the descending Trigeminus tract. The peripheral cells of the central gray substance, II. n, have long single processes. These processes branch, one of the finer branches entering the coarse fibres of the descending Trigeminus, the other passing outwards into the fine layer of fibres of the tectum opticum. This observation has been confirmed at several points. It shows first that the fibres of the Trigeminus may have a compound origin, partly from the large cells, 5 11, partly from the small cells ; second, that here is a possible centre between the optic and trigeminus nerves.

The second optic tract, II. t"^, arises from a mass of cells im the complete structure of these commissures. I do not therefore at present feel in a position to reply to his courteous criticism of my paper.

82 OSBORN. [Vol. 1 1.

bedded in the central gray of the thalamus, and not clearly differentiated as a distinct nucleus from the surrounding cells.

The third tract, II. t 3, enters the hemispheres directly. This observation, although made in both the Urodela and Anura, requires to be confirmed. The optic tracts have other sources of origin ; ^ the above are the main centres contributing the principal portion of the nerve, and are the only ones which can be clearly made out by the carmine method.^

In Rana the roof of the optic lobes is divided into eight distinct layers, the fibres of the superior portion of the optic tract entering the outermost layer, i Fig. 26, and ramifying to the interior cells. The inferior portion of the optic tract enters the second fibre layer, 3. The tracts surround the whole circumference of each of the lobes.

The relations of the encephalic tracts in the general architecture of the amphibian brain is shown in Fig. 21, in which only the well-determined tracts are introduced. The designation of the tracts expresses my present views ; as I have already stated, the efferent or afferent character of the main tracts is not by any means settled.

Comparison of the Dipnoan and Amphibian Brain.^

The recent memoir of Fulliquet upon the brain of Protoptenis is the first contribution to the histology of the Dipnoan brain, and enables us to make some interesting comparisons. In the first place, as observed by this author and others, the general external resemblance between the Dipnoan and Urodele brains is very striking. The principal external features in common are as follows : the olfactory lobes are not well distinguished from the hemispheres ; this is even more marked in Protoptenis than in the Urodela. The mesencephalon of Protoptenis passes imperceptibly into the diencephalon, as in AmpJduma, only a faint lateral constriction between these segments being evident.*

1 See the papers of Bellonci , Blaschko and Edinger.

2 In my paper upon Menopoma ('84, p. 267, Fig. 8), I described a portion of the optic nerve as non-decussating. I am now inclined to consider these supposed uncrossed tracts, b, as the mesencephalic tracts which cross at a different level, or the basal optic root of Edinger. The sections at this point are deceptive.

8 See also Appendix, Note 4.

  • In distinguishing between these segments internally, Fulliquet has failed to take


The optic lobes form single, unpaired bodies, as in many of the Urodcla. The infundibular lobes in both orders are large and functional. The cerebellum is small ^ and partly overhung by the mesencephalon.

In a few respects the brain of Protopteriis differs : the ganglia habenarii7Ji are much larger than in the Urodela, resembling those of Petromyzon. The anterior portion of the metencephalon is greatly expanded by the hypertrophied nuclei of the 5th-8th nerves and a diverticulum of the 4th ventricle. The olfactory nerves arise from the dorsal aspect of the rhinencephalon.

The similarity in the internal structure is very significant. The general arrangement of the encephalic gray substance, as the central gray, immediately surrounding the ventricles, is the same ; the dipnoan ependynia has the peculiarity, first observed by Stieda in Rana, of the thread-like extensions of its cells through the central gray into the white substance ; we also observe the peculiarly modified elongate ependymal cells (Fulliquet, PI. IV. Fig. 17, cde), in the region of the posterior commissure, so characteristic of the Urodele brain. As these observations were prior to the discovery of the true nature of the pineal gland, the author has naturally failed to identify this structure, mistaking the ganglia habenarum for it (Fig. 19). The relation of the dia- and procceliae seem to resemble those of Rana more closely than those of the Urodeles, since, as far as I can judge from Figs. 19 and 20, the prosencephalic commissures are in the lamina tcnninalis proper, and not in a projection of the floor of the ventriculus communis, as in the Urodela. The space marked TM in Fig. 19 represents this ventricle.

The encephalic commissures of Protopterns ^ apparently agree

advantage of the anterior and posterior boundaries of the Diencephalon (Entrenc^phale) , as defined by the posterior and superior commissures.

^ As there is no sagittal section of the Protopte7-ns brain given, it is somewhat difficult to determine the limits of the cerebellum. A portion of the brain designated Cervelet (Plate I.) is apparently a portion of the hypertrophied nucleus of the trigeminal nerve.

2 FuUiquet has not distinguished the posterior or superior commissures as such; nor has he identified the upper bundle in the lamina terminalis with the corpus callosum, as seems highly probable. It follows that my determination of these commissures is largely inferential from a comparison with similar sections of the Urodele brain.

84 OSBORAT. [Vol. II.

closely with those of CryptobrancJms. We first observe that the lobes of the infundibulum (Fig. 14) are united dorsally and ventrally by commissural fibres as in the Urodcla (Osborn, '84, Plate IV., Fig. 4, scm, icni). The two sides of the tectum opticum are also united by an almost continuous band of transverse fibres, Fig. w^fcm, as in the Amphibia, terminating anteriorly in a special enlargement, fee, Fig. 17, which I identify as the posterior commissure. The bande fibrillaire latcrale oiY\2^. 15, bfl, is probably either the posterior commissure or Meynert's bundle. The bundle connecting the ganglia habenarum, in Fig. i^,fgp, is probably the superior commissure. The upper bundle Jik, Fig. 19, and pfh, Fig. 20, in the lamina terminalis, is probably the corpus callosum; this is an important fact, if verified, since this commissure has not heretofore been definitely identified below the Amphibia.^ This can only be verified by following the course of its fibres. The lower bundle, Fig. 19, ca, occupies a somewhat unusual position, upon the floor of the brain.

Two differences in the internal structure may be noted : first, the optic chiasma is intra-axial or central, instead of peripheral, as in the Amphibia ; second, there is a single pair of large Mauthner fibres in the medulla oblongata, which are wanting in most of the Urodela^ although the posterior longitudinal fasciculus is apparently present,^, sp, Fig. i.

I cannot institute any satisfactory comparison in respect to the origin of the cranial nerves. Many minor points of resemblance and difference have been passed by in this resum6, the general conclusion being that there is a very close similarity between the Amphibian (Urodele) and Dipnoan brain, both in the external and internal structure.

Comparison with Petromyzon.

One of the chief features of Ahlborn's memoir upon the brain of Petromyzon is the thoroughness with which he has in 1 In concluding my paper upon the corpus callosum (Morph. Jahr., Band XII., p. 539), I had not seen Fulliquet's memoir, but anticipated from the embryology of the Ceratodiis fore-brain, that this commissure would be found in the DipnoL

'^ They have been observed by Stieda in Siredon, see Appendix, Note 5.


vestigated the medulla oblongata, especially the origin of the cranial nerves and the fate of the Mullerian fibres. The arrangement of the latter, I have reason to believe, is very similar to that in the Amphibia, but this system has not been as yet well investigated.^ His conclusions as to the origin of the cranial nerves are in many details supported and confirmed by my own, although I have discovered many additional structures which either do not exist or have been overlooked in the lamprey.

An important difference is seen as to the more primitive condition of the medulla in Petromyzon, in that the nerve nuclei are more central (Fig. 18), and not confined to the lateral regions of the medulla as in the Urodela. In the latter I have observed the large ganglion cells, gc, opposite the exit of the 8th, but cannot confirm the entrance of their processes into the Auditory roots as observed by Rohon and partly confirmed by Ahlborn. He places the 7th and 8th in one group, which I find cannot be done in the Urodela, the 7th being closely related to the 5th.

One chief point of agreement is that in Petromyzon the 7th nuclei and exit are dorsal to the 8th; this confirms my observations and undermines the hypothesis that the 8th is the sensory portion of a typical pair of nerves of which the 7th forms the motor element. Our observations agree further in respect to the connection of the posterior longitudinal fasciculus with the 8th (acusticus haubenbahn, p. 268); also as to the presence of a nucleus of pale ganglion cells near the exit of the 8th (p. 261), which I have found is one of the chief Auditory nuclei in the Urodela; further, the nuclei and tracts of the 8th have the same relation to those of the 5th, as I have described in Cryptobranchus ; finally, in both genera the 8th sends a tract into the cerebellum.^

The resemblance of the main Trigeminal system is also close. I do not find the nucleus of the transverse motor tract as large as he describes it, but the motor nucleus adjoining the ascending tract is similar in form and position (Figs. 1621). I do not find an upper and lower nucleus of the 8th,

^ See Appendix, Note 3.

2 This determination of the Auditory nuclei as ventral to those of the Facial is confirmed by both Stieda's and Koppen's observations upon Rana.

86 OSBORN. [Vol. II.

consisting of small cells such as Ahlborn describes (see Fig. 19). His upper 8th nucleus, VIII., corresponds exactly with my lower 7th nucleus, 7 mn ; this agreement raises a question whether Ahlborn's upper 8th root does not pass into the 7th nerve. If such is the case, the homology with CryptobrancJms is remarkably close.

In the upper portion of the encephalon in Petromyzon is found the posterior longitudinal fasciculus, approaching the Oculo-motor nucleus, and according to Ahlborn, not terminating with it, but decussating at this point to enter the thalami (p. 274). The descent of Meynert's bundle to the interpeduncular ganglion is also seen.^ The taenia thalami optici, p. 285, with fibres entering both the hemispheres and the thalami, is homologous with the superior commissure.


Ahlborn. Untersuchungen iiber das Gehirn der Petromyzonten. Zeit. f. wiss. Z06I.,

1883, p. 192. Bellon'CI. Sulle Commissure Cerebrali Anterior! degli Anfibia e dei Rettili. Bologna,

1887. Mem. del. Real. Accad. del. Sci. dell. Istituto di Bologna. Set. IV., Tom. VIII.

Intorno alia struttura e alle connessioni dei lobi olfattorii negli Artropodi

superiori e nei vertebrati. — Atti della R. Ace. dei Lincei, 1881-82. Edinger. Zehn Vorlesungen liber den Bau der Nervdsen Central-organe. Leipzig


Ueber die Verbindung der sensiblen Nerven mit dem Zwischenhirn. Ana tomischer Anzeiger. 1887. Fischer. Anatomische Abhandlungen iiber die Perennibranchiaten und Derotre men. Hamburg, 1864. FULLIQUET. Recherches sur le Cerveau du Protopterus Annectens. Geneve. 1 886. Gaskell. Journ. of Phys., 1886, Vol. VII., p. i. Gegenbaur. Die Metamerie des Kopfes und die Wirbeltheorie des Kopfskeletes.

Morpholog. Jahr. Band XIII., pp. I-112. Hill. The Grouping of the Cranial Nerves. Brain, Part XXXIX., 1887. KoPPEN. Zur Anatomic des Froschgehirns. Arch. f. Anat. u. Phys., Anat. Abth.,

1888, pp. 1-32.

Orr. The Embryology of the Lizard. Journ. of Morphology, Vol. I., No. 2, December, 1887.

OSBORN. Preliminary Observations upon the Brain of Antphiuvia. Proc. Phila. Acad., 1883, p. 177. Preliminary Observations upon the Brain of Menopoma and Rana. Proc. Phila. Acad., 1884, p. 262.

The origin of the Corpus Callosum, etc. Morph. Jahrb., Band XII., 1886. Part I., p. 223; Part II., p. 530.

^ Ahlborn has followed this bundle, asymmetrical in Petromyzon, into the medulla.

No. I.]



The Relation of the Commissures of the Brain to the Fonnation of the

Encephalic Vesicles. Proc. Amer. Assoc. Adv. Science, August, 1887; published March, 18S8, p. 262. Also, American Naturalist, October, 18S7, p. 941.

Rauber. Die Lehre von dem Nervensystem und den Sinnesorganen, (Anat. des Menschen), 1886.

Spitzka. Contributions to Encephalic Anatomy, in nine parts. Journ. of Nerv. and Mental Disease. New York.

Stieda. Ueber den Bau des centralen Nervensystems der Amphibien und Reptilien. Zeits. f. wiss. Zool. Band XX.

Ueber den Bau des centralen Nervensystems des Axolotls. Zeits. f. wiss. Zool. Band XXV.

Wilder. The Dipnoan Brain. American Naturalist, June, 1887.

" Anatomical Technology ; " also shorter papers upon the structure and nomenclature of the Brain.


1°. Figure 2. A vertical section of the Brain of a Rana embryo, at the period of the formation of the encephalic commissures, cbl, pcm, scm. Lettering as in the Explanation of Plates.

Fig. 3. A horizontal section of the same, composed from two levels, showing the relation of the commissures to the encephalic vesicles.

7^™ sen

2°. Fischer (op. cit., p. 135) upon the distribution of the Facial and Trigeminal Nerves in Cryptobranchus. The FACIAL divides into four main branches: R. palatinus; R. mentalis passes the mylohyoideus muscle and terminates in the skin of the lower jaw (sensory) ; R. alveolaris to the skin above the masseter muscle (sensory) ; R. jugularisto the mylohyoideus posterior and digastricus muscles (motor). The Trigeminus has three branches : 5", a, retractor bulbi, rectus externtis and superior, obliquus superior (motor, including A b d u c e n s) ; b, skin of forehead, olfactory chamber, snout and superior maxilla (sensory) ; 5"' to skin over premaxilla; 5", a, to masseter and temporal muscles (motor); /J, skin covering inferior maxilla (sensory). It will be observed that this distribution of the Facial is largely sensory and is consistent with the derivation of the main sensory elements of both nerves from the dorsal sensory nucleus, as I have described them. It also demonstrates that the lower bundle of the Facial is certainly motor.

3°. The Mvillerian fibres and posterior longitudinal fasciculus. As shoiivn in Fig. n (Plate V.), these two systems are apparently distinct. The Miillerian fibres are observed at the lowest medulla levels as a compact round


bundle immediately below the anterior commissure; as we ascend, they multiply and form a crescent in the anterior median columns; at the exit of the vagus they separate into two strands, one, dorsal, below the sulcus centralis, the other, ventral, in the anterior columns; the ventral strands cross and disappear between the exit of the 9th and 8th nerves, entering large ganglion cells. The dorsal ascend to a higher level without apparently crossing; they are continuous with and apparently constitute the main portion of the posterior longitudinal f a s c i c u 1 u s.

The Auditory tract of this fasciculus closely adjoins the dorsal Miillerian fibres, but may be distinguished from them by the slightly smaller diameter of its fibres; it is clearly distinguished opposite the exit of the loth pair^ and on all the higher levels; this leads me to the conclusion that this portion of the posterior longitudinal fasciculus is an ascending auditory tract apparently distinct from the Miillerian fibres. This further throws in doubt the supposed connection between this fasciculus and tract.

4°. Wilder upon the brain of Ceratodus. This description is macroscopic. In Ceratodus, like Protoptcrus, and unlike the lower Urodela, the ventriculus communis is very small. As in the Urodela there is a large szipraplexus. The olfactory lobes, unlike those of Protopterus and the Amphibia, are pedunculated instead of sessile. The anterior commissure is mentioned, but the corpus callosum is not observed. Finally, an important distinction is seen in the fact that the extension of the hemispheres (secondary forebrain) is ventral from the primary forebrain both in Ceratodus and Protopterus, instead of directly anterior, as in the Amphibia.

I may express here my indebtedness to Professor Wilder for much of the terminology which I have adopted from his papers, and my regret that he has found it necessary to change his former terms in so many instances.

5°. Stieda's observations upon Axolotl include several points in which, if correctly described, this genus differs from Cryptobranchus. i. Opposite the exit of the 1 2th nerve, on either side of the sulcus centralis, is a large Mauthner fibre, exactly as in Petromyzon. I think these fibres are not present in Cryptobranchus. 2. The ascending Vagus, or fasciculus solitarius, while arising in the same manner as in Cryptobranchus, is given off with the anterior roots, the Glossopharyngeus portion, instead of with the posterior. 3. A portion of the posterior longitudinal fasciculus makes its exit with the 8th pair, but is designated by Stieda as a portion of the Facial-Trigeminal system. This probably corresponds to the bundle which I have finally referred to the 8th nerve, Fig. 15, VII.VIII. 3, 4. Another important observation is that a portion of this same fasciculus enters the Trigeminus. I am inclined to doubt this, as it has no parallel elsewhere, and it would be easy to mistake this for the posterior bundle which I have observed uniting with the descending trigeminus (see 5 Z^, Fig. 17). 4. Stieda describes the ascending trigeminus as sensory, probably from its relations in human anatomy; it is true this tract is first found in the sensory column of the cord, but it subsequently lies in the motor region and is reinforced by motor cells. It probably consists of mixed fibres.

Stieda's and Koppen's observations upon the brain of Rana. It is difficult to give a critical review of Stieda's results, first, because from the limitations

1 I regret that none of the figures represent the relations of this tract to the Miillerian fibres with perfect accuracy. I am not sure that this tract can be clearly distinguished at the level of the 12th pair, as represented in Fig. 11.


of his histological methods his observations were fragmentary; second, because the brains of Rami and Cryptobranchus differ so widely.

Stieda designates as the nucleus centralis the group of sensory cells in the floor of the ventricle on either side of the sulcus centralis. I cannot support his suggestion that this nucleus is in any w^ay connected with the Vagus roots. As he has confused the Facial and Auditory nerves and nuclei, his results, so far as these are concerned, are invalidated.

Koppen, profiting by the Weigert method, has given a much more precise and full description of the medulla, although his description of the Facial-Trigeminal system is very incomplete. i. He recognizes the posterior longitudinal fasciculus, p. 6, and its probable connection with the Auditory nerve. This adjoins the dorsal part of the Miillerian fibre system, the ventral part of which, after crossing, disappears, as in Cryptobranchus, in the great ganglion cells opposite the exit of the Auditory nerve. 2. He describes also, p. 7, a nucleus of large cells as probably belonging to the Auditory; this corresponds to my pale ganglion, 8«, Fig. 14. At this point he fails to distinguish between the Facial and Auditory elements, for his dorsal Auditory root, p. 9, is probably the main portion of the Facial. This error, if error it be, arises from the fact that he expects to find the facial a purely motor nerve, p. 10. 3. The Trigeminus, p. 10, is traced to two ascending bundles, and a large motor nucleus. No sensory nucleus or mesencephalic descending bundle is described. 4. Koppen makes many important additions to the higher encephalic tracts and nuclei in Rana, among which are the Auditory tract to the cerebellum, p. 13; the interpeduncular ganglion, p. 16; the posterior longitudinal fasciculus to the region of the III nucleus, p. 17; the infundibular tract to the hemispheres; the mesencephalic and diencephalic tracts from the medulla, and the tracts from these segments to the hemispheres, pp. 30-31; the ganglion habenulse and Meynert's bundle. The hemispheres receive no direct sensory tracts from the medulla; these tracts first enter the optic lobes and thalami, from which fresh tracts rise to the hemispheres. With this conclusion, p. 31, I am inclined to agree, although I do not think it is absolutely demonstrated. 5. I differ from Koppen in his attempt to homologize the encephalic with the lower segments, p. 26; also, as I understand him, p. 30, he does not describe any direct motor tract, (basal forebrain bundle), to the anterior columns of the medulla. The agreement, in respect to the tracts mentioned above, is strongly confirmatory, since my observations were made independently, and the conclusions reached before the receipt of his paper.

6°. Methods. The methods of hardening and staining have been fully described in previous papers. The best staining results have been obtained with long in toto immersion in Ammonia Carmine. I have had the advantage, for purposes of comparison, of a full series of the brain of Salamandra, prepared after W e i g e r t ' s Method by my assistant, ]\Ir. J. Warne Phillips, in Dr. Edinger's laboratory. I find that brilliant as these preparations are, the carmine series give a fuller and more reliable field of observation. The late Professor Gudden, of Munich, told me, after long experience, that he had reached the same conclusion, and this is also the opinion of Dr. Spitzka of New York.

go OSBORN. fVoL. II.



Tke Encephalic Segments. — Rh, rhinencephalon, olfactory lobes, Pr, prosencephalon, cerebral hemispheres. Di, diencephalon, optic thalami. Me, mesencepha' Ion, optic lobes. Mi, metencephalon, medulla oblongata.

Ventricles. — a, aula, ventriculus communis, pre, prosocosle. p, porta, foramen of Monro, di, diacoele. nisc, mesocoele. mtc, metaccele. inf, infundibulum.

The Comviissures. — cto, commissure of the tectum opticum. cal, corpus callosum; c^/', posterior division of same, (?comm. cornu ammonis). zVw, inferior infundibular commissure, pes, posterior commissure, prcs, anterior commissure, scm, superior infundibular, scs, superior commissure.

Cranial Nerves. — I.-IV., as usual. V.', opthalmic; V.", superior maxillary; V.'", inferior maxillary, divisions of the Trigeminal. VII.', Facial, upper bundle, joining the Gasserian Ganglion; VII.", lower bundle joining the Auditory and sending off the Tt". wd-w/tf/u and J?. y«^M/<7rw. VIII., Auditory. IX., Glossopharyngeus. X., Vagus. XI., (XII.), Hypoglossal.

The Encephalic Nerve Tracts. — bpt, basal prosencephalic tract, bft, same. cblt, cerebellar tract entering optic lobes. cblt\ cerebellar tract from medulla oblongata, dst, sensory ( ?) tract to optic thalamus, dst', sensory tract from optic thalamus to hemispheres, fc, fasciculus communis, fs, fasciculus solitarius. it, tract from infundibular lobes to hemispheres, mst, sensory (?) tract to mesencephalon. msf, sensory tract from mesencephalon to hemispheres, mb, Meynert's bundle. plf, posterior longitudinal fasciculus.

Cranial Nerve Tracts. Trigeminus. — 5/', the ascending tract. 5 t^-^, the ascending tract reinforced by the tract from the great motor nucleus. 5 t^, the sensory tract. 5 t^, the descending (mesencephalic) tract. 5 t^, the encephalic tract.

Facialis. — T tl, lower, motor (?) tract. 7 /«, upper, sensory tract. 7 ?, encephalic tract.

Acustictis. — 8 /, tract from the pale nucleus, and special sensory nucleus. 8 ? , cerebellar tract.

Glossopharyngeus. — 9 /, sensory tract.

The Encephalic and Cranial Nerve Nuclei. — 3 «, 3 «', upper and lower Oculomotor nuclei. 4 «, Trochlearis, motor nucleus. 5 sn, dorsal sensory nucleus of Trigeminus. 5 77in, ventral motor ditto. 5 n, mesencephalic nucleus of Trigeminus. 6 «, Abducens nucleus. 7 sn, and mn, dorsal sensory and lateral motor nuclei of Facialis. 8 n. Auditory nucleus of pale ganglion cells. 8 sn. Auditory nucleus of small sensory cells. 9 mn, lateral motor nucleus ot Glossopharyngeus. 9 sn, dorsal sensory nucleus of same. 9 n, ventral motor nucleus of same. 10 n, ventral motor nucleus of Vagus, cbl.n, cerebellar nucleus, cs, corpus striatum. dn, superior diencephalic nucleus, dn', inferior diencephalic nucleus, fen, nucleus of fasciculus communis, gc, nucleus of large ganglion cells opposite Auditory root, gh, ganglion habenulse. gi, interpeduncular ganglion, nc, nucleus centralis, nm, nucleus magnus. np, pale nucleus.

92 OSBORN. [Vol. II.


The figures of the brains were outlined with a Nachet camera lucida, and the surface details inserted as studied in different lights. The vertical sections (Figs. 7-9) are reconstructions of the actual median plane by composition of a number of sections.

Figure i. Brain of Siredon (^Axolott) itiexicanus. The species is somewhat uncertain. Viewed from above. Enlarged \\ diameters.

Fig. 2. Brain of N^ecturus i^I\Ie)iobranchus) maculosus. Viewed from above. Enlarged 4V diameters.

Figs. 3 and 4. Brain of Proteus a^iguinetis. Viewed from below and above. Enlarged 4^ diameters.

Fig. 5. Brain of Siren laceriiiia. Viewed from above. Enlarged 4^ diameters.

Fig. 6. Medulla oblongata of Cryptobranchus alleghaniensis. Viewed upon the ventral surface, showing the exit of the cranial nerves, as reconstructed from transverse sections.

Fig. 7. Sagittal section of the brain of Cryptobranchus, through an ideal median plane, as reconstructed from sagittal sections.

Fig. 8. Sagittal section of the brain of Necttirus, drawn as above.

Fig. 9. Sagittal section of the brain of Rana, drawn as above.

Journ . Mor/ih . Vol II

94 OSBORN, [Vol. IL


Figs, io to i8. A series of transverse sections through the medulla oblongata of Cryptobranchiis alleghanieiisis, from the exit of the I2th pair (XI.) to the exit ot the 6th pair, enlarged about forty diameters. For the sake of clearness only a few cells of each nucleus are represented, and these are drawn to scale. Most of the sensory and motor cells of the Centi-al region are also omitted. The cell area is represented in dark gray. The contours of the tracts are usually well defined, but have been exaggerated in several figures. The level of the transaction represented by each of these figures is indicated in the diagram, Fig. 21 ( — f. 10 — f. 18).

Fig. 19. Sagittal section through the tectum opticum {op{) and cerebellum of Cryptobratichiis. This shows the posterior portion of the great trigeminal nucleus; the small nucleus of cells in the cerebellum, cbln ; the decussating cerebellar tracts, cblt; and the descending trigeminus, 5 /*. iqa represents the cells of the trigeminal nucleus enlarged.

Figs. 20 and 20a. Vertical and horizontal sections through the oculo-motor nucleus, as seen under a Zeiss D; camera drawing.

Fig. 21. Diagram of the cranial nerve nuclei of Cryptobranc/uis, as reconstructed from transverse and horizontal sections. The relative location of the 3d-6th nuclei, posterior commissure, Meynert's bundle, the main exits of the nerves and the posterior longitudinal fasciculus are all determined with the camera. The nuclei of the 5th-ioth pairs are reconstructed from transverse sections. They overlap each other less closely than in nature. The sensory nuclei are distinguished by oblique lines; the motor nuclei by transverse lines; the pale ganglion nucleus of the Auditory by crossed lines. The external row, 5, 7, 9, are the dorsal sensory nuclei. The internal row, 5, 9, 10, are the ventral motor nuclei. The external intermediate, 7, 9, are the doubtful lateral motor nuclei, adjoining 8 and / the lateral sensory nu cl ei.

The fascic ulu s com muni s/j/'" is represented in gray to indicate that it is apparently forming new bundles from a continuous sensory nucleus.

The dotted lines represent the descending, i.e. encephalic, tracts. The heavy lines, the descending trigeminus and the posterior longitudinal fasciculus. The pale lines, the tracts springing from the special nuclei.

.Joarn Mi'/yi/i i o/' il

96 OSBORiV. [Vol. II.


The figures, with the exception of Fig. 30, are composites of three or four sections, and thus present a larger number of fibres than are in view in any single plane.

Figs. 24 and 25. Sections of Ciyptobranchus. The former is through the outer regions of the encephalic segments; the latter is just external to the mass of central gray matter.

Fig. 25. Transverse section of the mesencephalon of N'ecturits at the exit of the 3d pair of nerves, 2^a. An enlarged actual section of the roof showing the connection of the descending trigeminal fibres with the cells of the optic tract. Camera drawing.

Fig. 26. Sagittal section through the outer portion of the left optic lobe of Rana. Eight layers may be distinguished as follows : i, the outer fibre layer which contains the superior portion of the optic tract; 2, the layer of scattered ner^'e cells; 3, the second fibre layer penetrated by the inferior portion of the optic tract; 4, the second layer of compact nerve cells ; 5, the third fibre layer ; 6, the third fine layer of compact nerve cells; 7, the fourth fibre layer; 8, the ependyma. The outer fibre layer is mostly homogeneous. The second and third fibre layers are traversed by radiating fibres which connect the three cell-layers.

Figs. 27 to 29. Outer, median, and inner sagittal sections of the brain of Rana. The latter is slightly external to the central gray substance.

Fig. 30. Sagittal section of the brain of Atnphiuma. The dotted lines represent the actual contours in the lateral plane of this section. The plain lines represent the contours of the median plane.

Fig. 31. A diagram of the main encephalic tracts and nuclei as obsei-ved in the Urodele Amphibia, a composite of the figures upon Plates V. and VI. The distribution of the tracts is a matter of actual observation. The sensory or motor character of the tracts is in a measure inferential. S, main sensory tract. M, main motor tract.

. Inurn. .'^for/i/i Vol II

y^y '^



WILLIAM PATTEN, Ph.D. II. Eyes of Acilius.

A COMPARATIVE study of the eyes offers, in my opinion, a promising field for the determination of the relationship between the different groups of Arthropods. To determine these relationships it is necessary to understand, (i) variations in the structure of the eyes, (2) their topography, and (3) the position and structure of the optic ganglia.

It is necessary first of all to obtain a thorough knowledge of some typical form. Comparison is then easy, for the problems solved in one case are already solved in others, while those left unsolved will torment the investigator in every succeeding study.

In following the development of the eyes of Acilius I was forced to consider their relation, and that of the optic ganglia, to the brain. This required a careful examination of the development of the whole head. But finding it impossible to treat of the head without considering the rest of the body, I decided to confine my attention to the eyes and optic ganglia. Some of the figures, however, will also serve to illustrate certain points in the "Development of Acilius," which will be the title of the paper I hope to publish in one of the next numbers of this Journal. I shall reserve till then a discussion of the relations which the eyes and optic ganglia bear to the head of Arthropods.

I. Topographical Relations of Optic Plate and Optic Ganglion to Cephalic Lobes and Brain.

The earliest stages of the eyes and optic ganglia are found before the appendages have made their appearance. The distal edge of each cephalic lobe is thickened to form the semicircular optic plate, distinguished on surface views by its opacity and the small size of its nuclei. The large ectoderm cells, on the inner

98 PATTEN. [Vol. II.

edge of the plate, are infolded, forming an almost semicircular depression, the beginning of the invagination which subsequently gives rise to a large part of the optic ganglion.

The optic plate is soon divided into three segments (PI. VII., Fig. i). The anterior, or first segment, forms a kind of corner to the anterior, lateral edge of the cephalic lobe ; the second occupies the median and larger part of the optic plate, with which its long axis is parallel ; and the third forms the posterior edge of the cephalic lobes, its long axis being nearly at right angles to that of the second segment. The distal edge, and in some cases the whole, of the plate is buried in the yolk ; it is then necessary, if one desires to obtain surface views, to cautiously pick the yolk away from the lobes with needles.

The ganglionic invagination soon deepens, and also shows traces of a differentiation into three parts, one of which lies on the inner edge of each segment of the optic plate (Fig. i o.g}-^).

From the median part of the cephalic lobes rise three pairs of thickenings, one pair opposite each ganglionic segment. They form the rudiments of the brain (Figs. 1-3, b^~^), and appear to be direct continuations of the segments of the ventral chord.

The embryo is divided by transverse folds into a certain number of post-oral terga, the first tJiree of which are gradually drawn tozuards, and united zvith, the cephalic lobes. The line surrounding the optic plate, after the rupture of the embryonic membranes, is not the old boundary of the cephalic lobes, but a new formation, which includes the cephalic lobes and the terga of the first three post-oral segments (PI. VIII., Fig. 8). The indentation, tgy', appears in surface views like a newly developed eye. It was quite enigmatical until a more careful study showed it to be the boundary between the posterior edge of the cephalic lobes and the terga of the first three post-oral segments. The last trace of this fold disappears soon after the stage shown in Fig. 8.

While the optic plate is still in its semicircular form, tv/o areas in each segment become visible in surface views, on account of the peculiar arrangement of their nuclei. From each of these areas is developed an ocellus, which I have numbered in the way shown in the plates. The growth of the embryo gradually forces the median, ventral part of the germ band forwards, leaving, as it were, the distal ends of the cephalic


lobes behind, so that they are turned first posteriorly and then ventrally. This makes the ventral row of eyes assume an anterior and then a dorsal position, and also brings the first segment behind the other two (Figs, i, 4, 8, 9, 10).

After the revolution of the embryo, eyes II. and IV, come to lie close together behind eye I., and eye VI. is wedged in between eyes I. and III. A comparison of the figures will show better than any description how the embryonic eyes develop into those of the larvae. The reader who would follow the rather confusing changes in the optic ganglion and its nerves must have a clear mental picture of the way in which the eyes shift their position.

When the optic plate first appears, it is conspicuous on account of its dark color. Sections show that at this period it is composed of a single layer of closely packed cells, whose small, darkly stained nuclei are crowded together several rows deep. The nuclei in the brain and optic ganglion are large and spherical, and do not stain deeply. The optic plate is everywhere sharply demarked from the surrounding tissues, except on its distal, inner edge, where there is a gradual transition to the wall of the optic invagination (PI. IX., Fig. 20, /.«.).

Occasionally one finds a large nucleus near certain clear areas of which we shall speak later. They are homogeneous and deeply stained, and, in some cases, contain a nuclear spindle (PI. IX., Figs. 20 and 21 ; PL XL, Figs. 60 and 61). They often appeared like naked nuclei projecting above the level surface of the optic plate. I do not remember seeing them anywhere except in those parts of the optic plate that give rise to the eyes.

Eye V.

We shall now describe, by means of sections and surface views, the development of each eye. It is only with the aid of these surface views that the earliest stages can be understood.

While the optic plate is but a semicircular thickening of the lateral edges of the cephalic lobes, eye V. appears as a round, dark spot, surrounded by a clear area, on the inner edge of the third segment (Figs. 1-3, V.). During the earliest stages, the central area is conspicuous on account of its size and dark color. Later, it is reduced to a narrow band composed of a double row of darklv stained nuclei.


A large nucleus appears in the central area, which increases in size as the latter becomes more and more elongated. It finally lies in about the middle of the dark area, between the two rows of nuclei (Fig. 6d).

The explanation of the surface views will be found on examining PI. XL, Figs. 61-63, which represent sections through the eye while in the stages shown in Figs. 2, 4, and 6. It will be seen that the median, dark area corresponds to a broad elevation of the ectoderm, in which the nuclei are crowded toward the surface, the superficial ones being slightly stained by my method of preparing surface views. On either side of the ridge the nuclei are deeply situated and unstained, hence the clear area surrounding the ridge. In order to save time and expense I have, in most instances, indicated the surface nuclei by shading. The elevation gradually narrows and sinks to the level of the surrounding ectoderm, or even below it, until it is reduced to a narrow ridge with a shallow furrow along its summit. On either side of this furrow a single layer of nuclei extends to the surface of the ridge (Fig. 63). The most superficial of these nuclei are the ones which appear as a double row in surface views.

The clear band surrounding the dark area soon loses its regular outline and becomes distinctly four-lobed (Figs. 6 and 6d), each lobe being a depression in the ectoderm, beneath which the nuclei are arranged much like those in the whole ocellus at a later period (compare Figs. 62 and 64), or like those found in separate sense organs. Each depression represents, in fact, a distinct sense organ, or eye. They form the greater part of the retina of the future ocellus. The two posterior pits are deeper and larger than the anterior ones, hence the whole clear area is somewhat conical.

There is a dark spot, with a clear area surrounding it, on the anterior dorsal edge, and one on the median ventral edge of the clear area (Fig. 6a, 5 and 6). These clear spots resemble the main clear area, with which they are continuous, in structure and general appearance.

All these clear areas and dark spots form one thickened patch of ectoderm, which soon becomes more regularly oval, and the pits and the dark streak disappear. The whole organ is being invagmated to form the floor of the deep depression seen in


Figs. 7, Ja, and 8c?. The opening of the depression, or optic cup, is pear-shaped, the broad, posterior end indicating the position of the two posterior pits, which are the last to be enclosed on account of their greater size.

During the earliest stages a striated cuticular thickening appears over each pit of the clear area, the striations being apparently continuations of those in the clear space beneath (PI. XL, Fig. 6i, ct). It terminates abruptly at the periphery of the central dark area, and as the latter decreases in width the edges of the cuticular thickening on either side of it come closer together, until only a narrow space, the median furrow, is left between them (Fig. 6% in.f.). When the whole sensory patch begins to invaginate, these edges unite, and the cuticula forms a continuous layer over the floor of the optic cup (Fig. 64).

From this cuticular layer the retinal rods are developed. In the stage shown in Figs. 62 and 63 it is apparently composed of stiff cilia, each of which has a minute, bead-like swelling at its base, while their outer ends are covered with a delicate membrane, which, in Fig. 64, is detached from the fibrous cuticula to form a delicate cover over the mouth of the optic cup.

The only evidence of the compound nature of the optic cup, after the disappearance of the median dark area, and the separate pits and cuticular thickenings, is to be seen in the optic nerve. In Figs. 62 and 64, the sections are at right angles to the line along which was developed the median ridge ; consequently one part of the nerve goes to the ventral, and the other to the dorsal half of the clear area (Fig. 6a). A longitudinal, horizontal section would show that each half of the optic nerve was divided into two much less distinct parts, one going to the anterior pit, and one to the posterior. Hence the optic 7ierve is composed of four, and perhaps more, nerve bundles.. This fact, together with the presence of fonr pits in the clear area, the cuticular thickenings over each pit, and finally the constant and remarkable arrangement of the nuclei beneath each thickening, shozvs clearly that this ocelhcs is composed of at least four primitive optic pits. The two clear areas, each with a dark spot in the centre, found on the dorsal and ventral side of the eye in Fig. 6a, form the two bands of inverted cells in the future ocellus. These patches resemble, both in sections and in surface views, the clear and dark areas in the centre of the

102 PATTEN. [Vol. II.

eye, and there is the same absence of nuclei beneath their cuticular thickenings. I have not been able, however, to detect any evidence that they are supplied with separate bundles of nerve fibres, but this will not seem strange, when we consider their small size and transitory nature. I see no reason why they, like the sensory patches composing the centre of the eye, should not be regarded as remnants of distinct sense organs. This supposition will appear in a more favorable light, perhaps, when we consider similar patches connected with the rudiments of the remaining eyes.

The sensory spots, or primitive eyes, of which the clear area is composed, are comparable with those on the mantle edge of Area, as described by me in " Eyes of Molluscs and Arthropods " (PI. 30, Fig. 42). It was there shown that some of the eyes arose as simple, pit-like depressions covered by cuticular thickenings, into which the everywhere-present intercellular nerve fibres extended, producing the appearance of vertical striations. Such eyes contain the lowest stages in the development of visual rods. In more specialized eyes it was shown that the cuticula had broken up into blocks, one overlying each cell. The nerve fibres arrange themselves around these blocks in various ways, and a true retinal rod is the result.

In the embryos of Acilius, the eye passes through the stages permanently represented on the mantle edge of Molluscs, for the cuticular thickening over each optic pit, at first fuzzy and non-refractive, soon changes into a layer of stiff and refractive cilia-like bodies, which finally form a dense and almost homogeneous cuticular layer. As soon as the retinal cells become distinctly outlined, this cuticula breaks up into a number of minute rods, two being formed over each retinophora. The arrangement of the nerve fibres about the rods will be described later. It is sufficient for the present to state that it corresponds in almost every particular with that found in the more highly developed rods of Molluscs. Moreover, m. Acilius as in Molluscs the cuticula overlying the sensory cells is divided into two layers : a thin outer membrane devoid of nerve fibres, the conical cuticula ; and a thicker, inner one, the retinidial ciiticula, so called because it contains nerve ramifications, and gives rise to the rods. The development of the rods in Acilius, therefore, is in perfect harmony with the view concerning the phylo


genetic development of visual rods, which I formulated from a study of the eyes of Molluscs. The harmony will be more perfect, when we consider the nerve endings, and compare the theory advanced to explain the intercellular nature of these nerve-ends with the origin of ganglion-cells.

The Large Nucleus of eye V., which is wedge-shaped and filled with dark granules, is at first situated very near the surface, and usually projects some distance above it. It remains in this position until the dark area begins to disappear (Fig. 6). Sections of the eye after this period show that as the sensory patch is invaginated, the large nucleus withdraws from the surface and takes up a position among the other nuclei in the middle of the optic cup. I have not been able to find it after the stage represented in Fig. 65, but I presume it remains practically unchanged throughout larval life. Such is undoubtedly the case with similar nuclei in eyes I. and IV.

The mouth of the optic cup is gradually reduced to a narrow slit, the long axis of which is parallel with the median ridge, above which the thickened lips finally unite (Fig. 66). Although there is no duplication to form separate layers, there are three regions that can be identified as the three layers of .the future eye. The closed lips of the optic cup are composed of radiating cells with deeply situated nuclei. Those over the centre of the eye are bent at right angles, their attenuated inner ends extending as far as the basement membrane on the sides of the eye. These bent cells, which give rise to the corneagen, are connected by intermediate forms with the short, straight ones of the surrounding ectoderm.

The Corneagen increases rapidly in thickness, and at the same time all its nuclei, except those on the periphery, become so indistinct that they seem to have disappeared. Even in very successful preparations it is only here and there that one can see, at the inner ends of the corneagen cells, the small, round nuclei with sufficient clearness to preclude all doubt as to their identity. In most cases, they are reduced to clear, colorless sacs, which can be recognized as nuclei only by a careful study of their position and the transitional stages by which they are connected with the undoubted nuclei of the periphery. In the fullgrown larva, the corneagen cells are large and wedge-shaped,

104 PATTEN. [Vol. II.

with distinct walls surrounding the coarsely granular, sometimes flocculent, cell contents.

The Iris contains pigment granules, varying in size from minute specks to large spherules, as large, or larger than the neighboring nuclei. The spherules are usually brown and contain a minute black dot in the centre. They are most abundant on the inner edge of the iris. The small granules are usually dull black.

When the pigment is dissolved by acids or alkalies, the iriscells are left quite empty and colorless, so that even after the removal of the pigment it is easy to distinguish them from the adjacent non-pigmentiferous cells of the corneagen, for the latter are filled with deeply stained granular protoplasm.

The Lens. — Immediately after the rupture of the embryonic membranes, and while the optic cups are still wide open, a delicate pellicle is formed on the outer surface of the hypoderm. This pellicle, which gives rise to the cuticula, usually covers the mouth of the optic cup (Fig. 65), but no part of the membrane is enclosed in the cavity of the eye, at least I have never been able to discover any traces of it in the newly formed vesicles. After the optic cups close, the pellicle is thrown into loose folds over the whole body, and is then cast off. At the same time, a new pellicle is formed beneath the old one, which is still present but widely separated from the surface of the embryo. The new skin is first visible above the unmodified hypodermis. Over the surface of the newly formed corneagen, it first appears as a thick, vertically striated layer much like that which gives rise to the rods (Fig. 6"]).

Soon after this stage, the corneagen cells increase in height and form an elevated cap to the optic vesicle, with a circular depression around it. The second pellicle becomes refringent and is thrown into minute folds everywhere except over the eye. There the non-refringent layer is transformed into a disc of clear, refractive cuticula (Fig. 6^.

After hatching, the dome-shaped layer of cuticula over the eye increases in density and thickness until it forms the strongly biconvex lens of the adult, the curvature of the inner surface being much greater than that of the outer.

In the full-grown larvae the cuticula surrounding the lens is composed of two layers : a dark brown, or quite black, outer one ;


and a clear, transparent, tangentially striated inner one. The whole layer is divided into thick, imbricating scales, the tips of which are raised a little above the surface, producing according to the location either a notched, wavy, or serrated outline. The lens is composed of the same kind of scales, but they are thinner and higher, and concave like the leaves of an onion. In most cases they overlie one another so closely, and their outer surfaces conform so accurately with the general curvature of the lens, that it is difficult to distinguish their boundaries. But it sometimes happens that their outer edges curl upwards, producing a serrated outline to the lens, like that in Fig. 70.

The Retina. — The thick cuticular layer upon the floor of the optic cup is finally interrupted by a second median ridge, which appears in exactly the same place as that occupied by the first. It is formed by two rows of cells, the enlarged, projecting ends of which are bent so that the tips of all the cells in one row face the tips of those in the other (Figs. 66-70). At the tip of each cell is a minute rod, which at first sight might appear to be a product of the lateral wall of the cell ; but this is not the case. These rods are terminal and horiaojital, as we can see from the shape of the cells, and from a comparison with other eyes, where the homologous cells and their rods are much larger and can be more conveniently studied.

After the closure of the optic vesicle, some of the cells near the inner surface of the corneagen are distinguished by their large, deeply stained nuclei. The horizontal ends of these cells, which ultimately give rise to the outer zvall of the optic vesicle, are frequently broken into loose, ill-defined, granular masses (Fig. 6"], o. zv.).

The innermost portion of the optic cup is composed of retinal cells, the greater number of which are placed so that their outer ends are at right angles to the median, vertical plane of the eye. Their rods are tJierefore Jiorizontal, and those on opposite halves of the eye face one another. Only a few retinal cells on either side of the ridge are upright and parallel with the optic axis. The free ends of the cells zuhich form the outer zvall of the optic vesicle are finally bent inzvards, and thus completely inverted (Figs. 68 and 69, 0. w.) ; at the same time they become inore sharply outlined, and zvell-defined rods are formed at their tips.

The ends of the peripheral retinal cells soon begin to draw

io6 PATTEN. [Vol. II.

away from the median plane of the eye, the cavity of the optic vesicle disappears (Figs. 67-70), the layer of rods becomes flattened out, and all the retinal cells and rods assume an upright position, except those of the median ridge.

The RETiNOPHORiE. — I have shown that in Molluscs, the essential elements of the retina were colorless cells, bearing double rods, and containing two nuclei and an axial nerve fibre. These cells, or retinophorje, I maintained were formed by the fusion of two rod-bearing cells, in one of which the nucleus degenerates. The nerve fibres between the two cells come to lie in the centre of the double one, to form its axial nerve. Each cell retains its own rod, hence the double rods of the retinophorae. Resting on these facts, and upon the supposition that the primitive nerve fibres were intercellular, I maintained that when sensory cells were found with axial nerves and traces of two nuclei, they should be classed as retinophorae, and must have arisen in the manner described above. In the Arthropods, I found the crystalline-cone cells had characteristics which led me to consider them as modified retinophorae. My supposition was based on the expectation of finding in the ocelli of Arthropods, retinal cells, either very similar to the retinophorae of Molluscs or intermediate between them and the crystalline-cone cells. A comparison of the text, and the diagrammatic figures in Plate 32 of " Eyes of Molluscs and Arthropods," will show that I expected to find the intermediate forms of retinophorae in the posterior eyes of Spiders, and in the eyes of Scorpions and Limulus, while the Molluscan type might be looked for in the ocelli of Insects and Myriapods, and in the anterior eyes of Spiders.

The retinophorae now to be described, show that these expectations have been realized, at least, as far as the ocelli of larval Insects are concerned.

The first stages in the formation of the retinophorae are found during the open-cup period of the eye. In Figs. 63 and 64, we see that the inner wall of the optic cup is closely packed with five or six rows of deeply stained nuclei ; all of which are alike except the median one already described, and those belonging to a few large ganglionic ones, to be described later. In the next stage (Fig. 65), they appear to be greatly reduced in numbers. This change is due to the fact that the cells are uniting


in tivos to form retinopJiorcB ; one half of the nuclei have degenerated into the non-stainable, apieal miclei of the retinopJiorcs, the other half form the deeply stained basal ones. The first stages of this process are necessarily difificult to follow, since both kinds of nuclei are mingled in one confused mass, where one can distinguish only uncertain differences in the color and size of the nuclei. But during, and after, the stage shown in Fig. 65, they begin to arrange themselves in two layers, most of the nonstainable nuclei lying above the normal ones. I fully expected to find in the embryos and young larvae, that the second nuclei would be, if not readily visible, at least much more distinct than in the adult. But my observations on Acilius do not support this expectation. Even as early as the stage shown in Fig. 65, the second nuclei are as faintly stained, and as difficult to see in section, as those of the adult eye. TJiis fact may indicate that the double cells are not the result of specialization in highly developed eyes, but that they are very ancient structtires which we should expect to find i^i the simplest, as well as in the most specialized, sense organs.

The large nuclei of the retinophoras arrange themselves in a single row on the inner surface of the retina. The position of the smaller nuclei varies somewhat; sometimes they are just above the large ones, sometimes just below the inner ends of the rods, but usually about half way between these two points.

During the stage shown in Fig. 64, and in younger stages it is impossible to distinguish the shape of the retinophoras as the cell walls are indistinct or absent, owing to the rapid division that is going on. In Fig. 65 the tissues are less dense, and the retinophoras can be seen as spindle-shaped cells drawn out to a point at either end. The clear striated area beneath the cuticular thickening is composed of the rod-like, or fibrous, outer ends of the retinal cells mixed with nerve fibres. They are soon transformed into strongly bent columnar cells.

Although it is sometimes possible to see the secojidary nuclei of the retinophorcB in sections, it is much easier to study their position and structure in isolated cells. It is also in this way that one obtains the most conclusive evidence of the double nature of the retinophorae, for by carefully turning and rolling them, either before or after the removal of the pigment, one always obtains one view which shows that they are composed of

I08 PATTEN. [Vol. IL

two twisted cells (PI. X., Fig. 58, a. b). The inner end of one of the component cells is swollen, and contains a round, fairly well-stained nucleus ; the outer end is quite small, and its contracted tip terminates in a flattened rod. The second cell is perhaps a little smaller ; its broad outer end contains a very faintly stained nucleus, and also terminates in a flattened rod. Its opposite extremity fuses with the first cell to form the inward prolongation of the retinophora. Both cells supplement each other's irregularities so perfectly that a symmetrical and apparently single cell is the result (Figs, ^"ja and 58(2).

When the cells are pigmented, the position of the second nucleus is often plainly indicated by a clear spot, but it would not be possible to identify it as a nucleus until the pigment had been removed.

The outer ends of the two component cells of the retinophorae are widely divergent (Fig. 58, g). A similar condition occasionally obtains in Area and Haliotis, and this fact affords excellent proof of the double nature of the retinophorae.

Pigment. — Just after the rupture of the embryonic membranes, one can readily distinguish with the naked eye the bright red ocelli of the living embryos. The pigment of these stages is readily soluble in alcohol, the reddish-brown pigment found in the succeeding stages, much less so. It is the latter pigment that is first seen in sections, distributed in coarse granules through the iris. At about the same time, the outer ends of the retinal cells assume a diffuse reddish-brown color. This coloring matter was probably in distinct granules, and became diffuse through the action of the alcohol.

In the next stages, thin sections show a row of minute, pigmented blocks, arranged in pairs (Fig. 69). The blocks are cross sections of a pigmented collar surrounding the outer end of each cell.

In the fully developed eye, the pigment varies in color from brown to jet black, according to the method of preparation. Chromic acid, bichromate of potash, and Muller's fluid, partly dissolve the pigment, leaving it a light brown color. Picro-sulphuric acid has a similar, but much less effect. Eyes treated with alcohol alone contain great quantities of intensely black pigment, which in all stages is most abundant around the outer ends of the retinal cells.


A careful study of sections and isolated cells shows that the pigment is distributed in granules along the nerve fibres that cling to the ivalls of the retinophora. Toward the inner ends of the latter, the course of the external nerve fibres is often distinctly marked out by the pigment granules deposited upon them (PL X., Fig. 58). At the outer ends of the cells, the external nerves are straight and close together. Where there is little pigment, one can see that the granules are arranged in rows around each of these nerve fibres (Figs. 56 and 57). Usually, however, the pigment forms a continuous and uniform envelop around the outer ends of the cells, where it terminates abruptly, while the nerve fibres are continued onwards over the outer surface of the rods. If there is any pigment at all inside the cells, it must be deposited in a very thin layer in, or just beneath, the cell wall.

Just below the inner ends of the rods, cross sections show a mosaic of deeply pigmented, hexagonal blocks, from which we might conclude that here pigment was lodged inside the cell walls. But isolated cells show that pigment is deposited between the flattened and diverging ends of the two cells composing each retinophora ; consequently we may not conclude from these solid blocks that the pigment is necessarily inside the retinophoras. Again, when the larva had died a so-called natural death, or perhaps owing to other conditions of which I was ignorant, the pigment dissolved, staining the retinal cells dark brown throughout. Of course, in cross sections of such material, pigment would appear to be deposited inside the cell wall.

There is proof in the optic nerves, that some of the pigment, at least, is an intercellular product. For in sections and in loosened or isolated fibres, it is evident that the pigment granules are scattered about between the fibres, and not in them or in any distinct cells.

In the peripheral cells of the retina, ami in those which constitute the outer wall of the optic vesicle, the pigment is coarser and less regularly arranged, resembling in general appearance that found in the iris. This fact enhances the deceptive appearance, already mentioned, that might lead one to think the retina was directly continuous with the iris and surrounding ectoderm. In the iris, the pigment is undoubtedly deposited inside the cell walls.

1 10 PATTEN. [Vol. II.

The Retinal Rods and Nerve Endings. — Since the retinophorae are closely packed in the retina and the rods are on the periphery of their outer ends, it follows that the rods of two neighboring rctinopJiorcB are placed side by side so that they often appear like one rod, while those of the same retinophorcB are separated by a zvide space.

In eye V. (Fig. 54a) there is a beautiful mosaic of brilliantly refractive rods which form regular hexagonal figures, in the centre of which are the axial nerve fibres, and on opposite sides, the two rods belonging to the same retinophora. On two of the sides no rods are developed ; only a thin membrane separates the adjoining hexagonal spaces at those points. Hence under a low magnifying power, the rods appear to be arranged in nearly parallel, zigzag lines.

It is impossible to see the axial nerves in the isolated retinophorge, owing to the twisting of their component cells. But they are visible in cross sections as a single bundle in each retinophora. In the clear space between the two rods, the nerve breaks up into three or four fibres arranged in a plane parallel with the sides of the rods. In one of the spaces in Fig. 54<7, they are at right angles to the normal position. Minute cross fibrillar arise from these axial fibres, which probably penetrate the rods and unite with the external nerves. The cross fibrillse constitute a retinidinm similar to that of Pecten. In the latter, each pair of rods has completely fused to form a hollow cylinder containing a single axial nerve, from which arise radiating cross fibrillas. In Acilius, the two rods are separate and flat, and the axial nerve is broken into smaller fibres placed in a row, so that a perfect radial arrangement of the fibrillse is impossible.

The clear space between the rods is apparently filled with a non-refractive fluid in which the retinidial fibrillas are suspended. It often contains fine granules, produced, I believe, by the coagulation or varicosities of the fibrillas. The space is exactly like that in the centre of the rods of Pecten, and is undoubtedly an ijitercellnlar one.

That the cross fibrillas penetrate the rods, I do not doubt, for they are marked with cross strias continuous with the fibrillas in the clear space. Neither do I see any reason to doubt that the fibrillas in the clear space, as well as those in the rods> are


equally essential elements. In Pecten, Acilius and Cephalopods, both the clear and the refractive parts of the rods are present. In the latter group, each retinal cell is probably double and has two rods separated by a clear space containing an axial nerve fibre. The rods are arranged in such a manner that four of them, each one from a different retinophora, come together to form compact groups, which Grenacher calls rhabdoms. Such rhabdoms, however, are different from those of the compound eye. They are more like the pairs of rods in Acilius. In neither Acilucs or CepJialopods have these groups of rods any moiphological significance, for they are incidental results of the arrangement of rods in pairs, and this arrangement varies greatly in the different eyes of the same individual. The important fact to be borne in initid is that the pairs of rods belonging to oite double cell are the units composing the layer of rods, and that the clear spaces containijig the axial 7icrvcs form the ce7itrcs of these miits.

In the reds of the convex eyes of Area, in the crystalline cones of Arthropods, and in the rods of the Vertebrate eye, if I am right in my interpretation of these structures, there is no clear space between the two rods of the same retinophora. On the other hand, in Mantis and probably in most Diptera, the hard refractive part of the rods has disappeared, and the retinidial fibrillse are suspended in a clear fluid. I think it is fair to conclude from these facts that the essential element of the rods is the system of cross fibrillar, or the retinidium, which may be suspended, in whole, or in part, either in a clear flidd or in a refractive and cuticula-like substance.

In Acilius, the clear space between the rods is not covered by a cap of cuticular substance, as in Pecten, although in some cases I have clearly seen that it is covered by an extremely delicate, arched membrane, through which one of the axial nerve fibres projects, and, bending at right angles, unites with a simi lar fibre from an adjacent pair of rods (Figs, ^ya and 59). Thus axial-nerve loops are formed much like those described in Pecten. In the latter case, the direction of the loops is more uniform than in Acilius, v/here they may be bent in any way, although there is a tendency to turn in some directions more than in others.

In Acilius, as in Molluscs, the retinophorse are supplied with numerous pigmented nerve fibres that extend along the surface

112 PATTEN. [Vol. II.

of the cells. At the outer ends of the cells, the fibres are straighter and slightly enlarged. The pigment stops at this place, but the fibres are continued onward, nearly parallel with one another, over the outer surface of the rods (Figs. 54 and 57).

The inner ends of the retinophoras are continuous with coarse fibres composed of the inward prolongations of the external and axial nerves. There are in rare cases smaller bundles which, on their way to the more peripheral portion of the retina, pass over the inner third of some of the retinal cells at a sharp angle. Some of these fibres impinge upon a retinal cell near the primary nucleus and join the other external nerve fibres belonging to that cell. Such nerve fibres occasionally cling to isolated cells, and one might erroneously infer that they entered the cell opposite the nucleus.

About the time the clear area begins to be invaginated a number of deeply stained granules appear on the periphery of each ocellar thickening. They increase in size from the surface of the ectoderm invv-ards, and each one is surrounded by a clear area (PI. XL, Figs. ^^-^6, and PI. XIIL, Figs. %2, Z^, ab. n). I have found similar products in the optic ganglia and in the ventral nerve chord. The manner in which they absorb coloring matter and their general appearance, together with the fact that they seem to be most abundant in tissues undergoing retrogressive changes, suggest that they may be degenerating nuclei.

Eye VI.

Eye VI. first appears in the posterior part of the third segment of the optic plate as a triangular depression, or clear area, the apex of which is directed dorsally and slightly forwards. On its ventral and posterior side, is a round dark area with a small clear spot in the centre. The depression soon becomes sickle-shaped, and finally circular, completely surrounding the dark area (Fig. 5(7). In a little later stage (Fig. 6a) the furrows separating the eyes from each other have disappeared. The clear space is divided into two parts by a dark ridge, on the dorsal side of which is that part of the clear area that appeared first; it is now somewhat rhomboidal, contains a dark ridge composed of a double row of nuclei, and is divided into four pits, l, 2, 3, 4. The ventral part of the eye is composed of an irregular clear


area or depression, 6, deepest at the anterior end ; it is connected with the dorsal part by a narrow furrow. The interpretation of these surface views is the same as that in eye V. The clear spaces represent slight surface depressions above a cup-shaped layer of deeply situated nuclei ; the deeper the depression, the lighter it is in surface views ; the darker parts are elevations where the nuclei come close to the surface (PI. IX., Figs. 33 and 34, VI). It is evident that the striictjire of eye VI. is much like that of eye V, for besides the two peripheral spots it contains four separate pits, or sense organs, divided by a transverse ridge, containing a large median nnclens, into two parts.

The whole sensory patch is soon changed into one uniformly clear area, which finally sinks below the surface to form the floor of a deep, circular-mouthed pit, or optic cup (PI. X., Fig. 47). The nuclei no longer show by their arrangement that the patch is composed of separate groups of cells ; the cuticular thickenings have united to form a uniform layer over the floor of the optic cup, and the median ridge has disappeared. I have not represented many sections of the earlier stages of eye VI., since with the exception of certain modifications readily understood from an examination of surface views, they are exactly like those of eye V.

There are considerable differences, however, between the later stages of the two eyes.

Just before, and during, the rupture of the embryonic membranes, the circular opening of the optic cup is converted into a transverse slit (Fig. 8c?). As soon as the lips meet, their cells, by assuming different curvatures, form the corncagen, and the outer wall of the optic vesicle. On the anterior, dorsal side of the eye, the retina is curled over in such a way that it forms a part of the middle layer of the eye, although it does not extend far enough to meet a similar, but larger fold on the opposite side (PI. XII., Figs. 72 and 73, o. zv.). In the still later stages, it is folded down on to the rods, and is apparently continuous with a very delicate membrane, — in which I believe I have seen nuclei, but so small and indistinct as to leave me in some doubt, — that extends over the outer surface of the retina till it meets the mass of inverted cells on the opposite side (Fig. 74, 0. w.).

When the sensory patch which formed the rudiment of eye

114 PATTEN. [Vol. II.

VI. was invaginated, the large clear spot, 6, (Fig. 6d) came to lie on the ventral and posterior edge of the optic cup (Figs. 8 and 9, 6). The cells of this sense organ become more and more bent and elongated as the cup closes, until they finally form a great tongue of cells projecting into the space between the corneagen and the retina (Figs. 71 and 72, /. i. c). As it increases in length, it reduces the optic cavity to an oval space in the dorsal and anterior portijn of the eye. The cavity is finally completely obliterated by the increase in length of the corneagen cells (Fig. 73).

The free ends of the cells bclo/igijig to the sixth sense organ bear mverted retinal rods, the same as those belonging to the upright 07ies, except that they are less regular in shape. Cross sections show that they are cylindrical with a clear central portion in which runs an axial nerve. The nuclei are at first situated near the free ends of the inverted cells, and are somewhat larger and differently stained from those in the remainder of the retina. When the pigment, which is scattered through these cells in coarse granules like that in the iris, is removed, they appear almost colorless, with here and there a few coarse granules.

Figs. 71-74, represent semi-transverse sections of the head, at right angles to the nerve n VI., Fig. 9, and parallel to the dark area of eyes III. and I., Fig. 8. When the sections are cut in most other cross planes, the retina appears perfectly symmetrical, its periphery being bent over and continuous with a very thin membrane in the same way that the left-hand edge is in Fig. 73. But if any of these cross sections should pass a little posterior to the median plane of the eye, the tongue of inverted retinophorae would be seen, which might readily be mistaken for the cut end of a large number of nerve fibres, such as one finds in retinas with inverted cells, as in Pecten for example.

That we should have eyes with both upright and inverted retinophorae, as in eye V., is remarkable and, so far as I know, without parallel. The condition in eye VI. is still more extraordinary, for there only a comparatively small cluster of rodbearing cells is inverted, while all the rest are upright.

The sense organ, 5, has either united with 1-4 to form the floor of the optic cup or disappeared ; at any rate after invagination it is no longer distinguishable.


In eye V., the two groups of inverted cells arise from the dorsal and ventral sense organs, 5 and 6, in the same way that the tongue of cells in eye VI. is derived from its ventral sense organ, 6.

In the full-grown larva, the group of inverted cells is proportionally as well developed and conspicuous as in the younger stages. It is not shown in Fig. 75, since the section is a semitransverse one passing through the optic nerve n VI., (Fig. \oci). The direction of the cells in the inverted sense organ is at right angles to that of the optic nerve, consequently it is impossible to cut a longitudinal section of both structures at once.

The Corneagen forms a thick cap to the eye, above which a lens is formed, as in eye V., from a striated layer of non-refractive cuticula.

In Fig. 73 there is a temporary indentation in the corneagen which seems to mark the place where the lips of the optic cup came together.

The Retina. — Just before the closure of the cup, the retinal nuclei suddenly appear to decrease in number, owing to the formation, in the manner already explained, of the retinophorae, which soon arrange themselves in a single layer, with all their primary nuclei at about the same level. The retinophorae are the sanie as those in eye V., except that they are a little shorter, and the primary and secondary nuclei are often situated so close together as to appear like one nucleus.

In the full-grown larva the retina is almost a hollow hemisphere, and consequently the rods on the periphery are very nearly at right angles with those in the centre.

The Rods are remarkably uniform in structure. On the posterior edge of the retina they are not quite so long as elsewhere. Each pair forms a thin-walled, hexagonal tube, which when isolated shows distinctly its two component rods. Cross sections show a mosaic of closely packed, hexagonal figures without the regularly arranged thin places so conspicuous in the rod-mosaic of eye V. (Fig. 45^).

Eye III.

Eye III. appears very early as an oval, dark area in the middle of the second segment of the optic plate. It is soon surrounded by a circular clear space, and then both parts become consider

ii6 patten: [

ably elongated, the latter assuming a figure-8 shape, and the former being reduced to an elongated ridge composed of a double row of nuclei (Figs. i-6). The explanation of surface views given in the description of eye V. will serve equally well here.

In both eyes I. and III. there are two broad and poorly defined light areas, one on the dorsal and the other on the ventral side of the eyes, which extend almost their whole length. These areas are well developed in the stages shown in Figs. 5 a and 4. But they soon disappear, and the eye assumes the appearance shown in Fig. 6a.

Although this eye is considerably longer and larger than eye v., it is evidently constructed on the same plan. The figure-8 shaped clear area is composed, as in eye V., of four sensory pits. In the centre is the elongated ridge, and on either side of its anterior end is a small, round, dark spot, surrounded by a faint depression. In surface views we see that the dark area is bent in the middle, and at the apex of the bend, which is directed dorsally, is situated the large nucleus. In some cases this configuration of the dark area is more conspicuous than that shown in Fig. 6a, which represents about the average condition. This bend is similar to that in the retinal furrow of eye IV., to be described later, but it does not, as in this last case, remain throughout life. The invagination and general structure of the eye is almost exactly like that of eye I., only it has no dorsal appendage or vertical retina. See i7ifra.

Eye I.

The early stages of this eye, which arises from the first segment of the optic plate, are much like those of eye III. After the light area has assumed a figure-8 shaped contour, it increases in extent at the anterior end until a large oval patch, in reality a shallow depression, is formed. This depression is finally separated from the remainder of the clear area by a dark ridge, leaving only a narrow connecting streak on the ventral side of the eye. Figs. 11-16 represent surface views of the first segment as seen from the anterior dorsal side of the embryo. In Fig. 13 a light streak, or furrow, is seen on the ventral side of the eye, united at its anterior end with the main clear area. At this point appears a round dark area {d. a} Fig. 14).


The main dark area, d. a}, which at first was very broad and surrounded by a narrow furrow, formed by the abrupt termination of the cuticular thickening overlying each sensory area (Figs. II, 12), is gradually reduced to a narrow ridge composed of two rows of nuclei, between which is the large median one, 7ic}

The dark area becomes strongly bent at the anterior end, and at the angle appears a second large nucleus, nc?, Fig. 14. Finally a third smaller and much less distinct nucleus appears where the area 7-8 joins the main eye.

After this stage, the round spot, d. a^, and the ventral clear area, 10, together with the nucleus, nc.^, disappear. The dorsal appendage grows smaller but more conspicuous, and contains an oblong dark area divided in halves by a narrow light streak, in the centre of which I thought I could detect a nucleus like that found in the dark areas of the other eyes (Fig. 18). The main part of the eye is reduced to a narrow, bent band, with a light or dark streak, according to the method of preparation, in the middle.

T/ie main part of eye I. differs frotn all others in containing at least tzvo large nuclei and eight sensory pits. The dorsal appendage represents at least one more pit, perhaps two, if %ve can place any reliance upon the indication of a retinal furroza as shown in the faint, light streak seen in surface views. As in eye V., the large nuclei are situated in the centre of a group of four sensory pits. The structure of the sensory area is seen from sections (PI. XIII., Figs. 82 and Zz) to be like that of eye V. The former section corresponds with Fig. 63, in which the cuticular thickenings are still separate ; and the latter, to Fig. 64, where they have united and the median furrow has disappeared. In describing the invagination, we will leave out of consideration the dorsal appendage c. e., of which we shall speak separately by and by.

Nothing like an optic cup is formed by the invagination. As fast as the median ridge sinks below the surface, the outer faces of the clear areas on either side are brought together like the leaves of a book, and the distal ends of the sensory cells meet one another in the median longitudinal plane of the eye.

But this is not the only direction in which invagination takes place. In Fig. 8 the sensory area is still seen lying lengthwise

Il8 PATTEN. [Vol. II.

on the surface ; but during the succeeding stage, its anterior end is pushed into the tissues of the head, as though the whole plate swung on a pivot at the posterior end, through an angle of ninety degrees. The long axis of the retina then lies at right angles to the surface, and in a semi-transparent preparation, like that shown in Fig. 9, we look along the retinal furrow, not down into it as in Fig. 8. It is not easy to form a clear mental picture of the way in which the retinas of eyes I. and III. are invaginated. To make matters clearer, let us suppose the head to be made of soft india-rubber. Then if a slender but inflexible rod be laid between the two rows of nuclei of the median ridge, as fast as it is pushed below the surface, the lips of the indentation that will tend to form, come together, conceal the rod, and prevent the formation of any cavity. If this process is continued until all the sensory cells on either side of the rod are invaginated, a long narrow pocket will be formed, composed of two thick walls, whose originally outer surfaces touch each other, bringing their cells end to end and at right angles to the opening of the pit (Fig. 85). There are no transitional cells connecting the inner edges of the two walls, so that, strictly speaking, there is no floor to the pocket. This completes the first stage of the invagination. Now, in eye III., suppose the posterior end of the rod is fixed and the opposite extremity swung inwards through an arc of ninety degrees ; the retina will then be brought with its long axis at right angles to the surface, and the second stage will be finished. Finally carry the rod in this position backwards a distance equal to the height of the future corneagen, and, without considering minor details, the third and last stage is completed.

In eye I. the pocket has a bend like that of the dark area (Figs. 14-16) which divides the retina into two unequal parts. Between the lateral walls of the posterior portion, or that which gives rise to the Jiorizontal retina, is situated the first large nucleus {nc\, Figs. 17 and 85), and in a corresponding, position in the anterior part, which gives rise to the vertical retina, is the second large nucleus, nc? The whole pocket is of nearly uniform depth, as shown in Fig. 17, which represents a digrammatic outline of the whole eye seen as a transparent object. In order to continue the invagination by means of our rod, it must be cut at a point a little back of the second nucleus, iic?, and


that "part of the rod lying in the future vertical furrow, removed. Now if, as in eye III., the posterior end of the rod is fixed, and its anterior extremity swung inwards through an angle of about ninety degrees, and tiien carried backwards a distance equal to the height of the future corneagen, the invagination of eye I. will have been completed, and a condition obtained like that in Fig. 18, PI. VIII. When we began to carry the posterior part of the furrow inwards and backwards, of course the anterior part was gradually pulled into a vertical position and extended along the inner lateral edge of the flattened optic cup, wliere it remains through life as the vertical retina.

Figs. 79-81, PI. VII., represent longitudinal vertical sections of eye I. (semi-transverse sections of the head, parallel with the long axis of the retinal furrow, see Fig. 8), showing three successive stages in the invagination of the retina.

Corneagen. — As soon as the early stages of invagination have carried all the clear area below the surface, the surround ing ectoderm cells are drawn in, in the same way, to form the corneagen. But these cells, instead of meeting end to end above the retinal furrow, take an outward curve, so that their long axes are parallel with the plane of invagination. The first direction of the invagination is towards the brain, and as the long axes of the corneagen cells are parallel with the direction of invagination, they naturally point inward and are at right angles to the long axis of the vertical and horizontal retinas (P'ig. 79). But when the invagination is directed backward, the inner ends of the corneagen cells are pulled backward also, so that they finally have their long axes at right angles to the horizontal retina, and parallel with the vertical one (Figs. 80 and 81).

The shape of the corneagen cells shows the way in which they were formed. In all cross sections of the horizontal retina, — sagittal sections of the head, — the nucleated ends of the corneagen cells bend away from the median plane to either side, and in such a way as to leave no conspicuous boundary between them and the retinal cells. For some time after the closure of the flattened optic vesicle there is a line in the middle of the corneagen marking the place where the lips of the invagination united. The difference in curvature between the cells belonging to the corneagen and to the retina gives rise temporarily to a small triangular space between these two struct

120 patten: [Vol. II.

ures (Fig. 85). After the lens has appeared, and even in the fully developed eye, the ends of the cells lying in the middle of the corneagen are so compressed and filled with fine granular protoplasm as to form a distinct, deeply stained core.

The nuclei of the corneagen are at first arranged in two great lateral masses (Fig. 85), but as it increases in depth, they form a single layer that at first sight appears to be a direct continuation of that formed by the retinal nuclei. But the inner ends of the median corneagen cells, although slightly bent away from the median plane, do not reach the periphery of the eye. One can see the nuclei at their expanded inner ends almost over the centre of the retina (PI. XIII., Figs. 90, 91). The inner ends of the remaining corneagen cells are situated at the periphery of the eye, and are filled with a dense layer of black or dark brown pigment granules which completely envelop and conceal the nuclei. In stained, depigmented sections, the formerly pigmented ends of the corneagen cells are quite colorless and empty, with the exception of a few coarse granules and the deeply stained nuclei.

Retina. — As we have already said, the horizontal retina is composed of two long strips of thickened ectoderm with their originally outer surfaces brought face to face (Fig. 85). In the early stages, the cells nearest the corneagen are indistinctly defined, and contain large oval nuclei with a few coarse, deeply stained granules. The position and general character of these cells is much like that of the inverted ones in eye V., and I believe them to be of the same nature. However that may be, they soon disappear. There is no evidence that, in the later stages, they form an outer wall to the optic vesicle.

Throughout the greater part of embr3^onic life the free ends of all the retinal cells are at right angles with the optic axis (Fig. 85) ; but the outermost cells finally draw away from the middle line and assume a more upright position. This process goes on until only a few rows of small horizontal cells are left overlying the gigantic ones which form the innermost walls of the furrow. In cross sections the layers of rods now form a Y, the diverging arms being composed of nearly upright rods, and the stalk, of the double rows of horizontal ones (Fig. 86).

In the full-grown larva the anterior wall of eye III. is quite straight, and most of the retinal cells of that side are horizontal


(Fig. 90). The posterior wall, however, is strongly convex, and the retina on that side is composed of a broad layer of nearly upright cells, whose short rods decrease in length from the furrow toward the posterior periphery of the retina. This remarkable asymmetry is correlated with the position and inclination of the lens, which looks upwards, inwards, and forwards, so that its optic axis falls about in the middle of this posterior layer of horizontal rods, and not, as one might expect, upon the retinal furrow. In eye I. the horizontal retina is also asymmetrical, but to a less degree, and in an exactly reversed manner, for the greatest expanse of horizontal rods is on the anterior side of the farrow instead of the posterior (Fig. 91). There is a similar asymmetry in the retinas of eyes II. and IV.

The gigantic retinal cells, arranged with great regularity in two rows, one on either side of the retinal furrow, are very broad and thin at their free ends. At their opposite extremities, there is a slight swelling containing a large oval nucleus. These cells are similar in nature and arrangement to those at the bottom of eyes II., IV., and V. (Figs. 70-78), except that they are much larger, and, owing to the way in which they receive their nerve supply, somewhat differently shaped. Those on the side of the furrow with the greatest number of upright rods are longer than those in the opposite part of the retina. The important point is, that, except in size and shape, they do not differ from the remaining retinal cells. They must be regarded as horizontal retinal cells with very wide and short rods ; compare the isolated cells d, e, and i, Fig. 58, PI. X. When these cells first become distinctly outlined (Fig. 85), there is only a small space between the primary nucleus and the rod, while the outer edge of the cell shelves steeply inwards to the level of the almost horizontal inner edge. They are also much darker than the smaller cells, and their free ends being concave, a space is formed between the two rows of rods. But the latter are soon brought more closely together, their terminal edges become as straight and ridged as though crystallized, and the narrow space is filled with a mass of densely pigmented nerve fibres.

In order to economize space without disturbing the arrangement of the rod-bearing portion, the thick nuclear ends of the cells are turned alternately toward and away from the median

122 PATTEN. [Vol. II.

plane. Hence it happens that in vertical cross-sections of the eye, four or five rows of large nuclei are seen near the middle of the retina, although the outer ends of the cells to which these nuclei belong are arranged with the utmost precision in two rows.

The giant cells have, in the early stages (Fig. ']6), short rods extending the whole length of their distal ends. In the fullgrown larva these rods are longer but narrower, the protoplasm of the cells having pushed its way under the inner edge of the rod to form a deeply pigmented heel that excludes the rays of light from the underlying nerve fibres (PL X., Fig. 56).

The large nucleus, nc^, conspicuous in surface views, remains throughout life unaltered, except that it is perhaps a little larger and flatter in the older stages. It lies a little inside the middle of the horizontal retina, between the deep edges of the two rows of large cells (Figs. 81 and 86).

The vertical retina differs principally in size from the horizontal one. It is deeply buried in the pigment at the periphery of the corneagen, and might be easily overlooked. It arises, as already pointed out, from the short bent portion of the dark area (Figs. 14, 15, 16). There is a second large nucleus that marks, in surface views, about the point where the horizontal and vertical retinas are continuous ; but when the eye is well invaginated, this nucleus has changed its position to about the middle of the vertical retina. In the next stage (Fig. 80), the vertical retina is reduced in size at the angle of the bend, so that it has apparently lost its connection with the horizontal one. At its outer end the retinal cells are well developed, and form a projecting mass that shelves off abruptly toward the dorsal appendage (PL XIII., Fig. 80).

In the still later stages the furrow is relatively smaller and more uniform in size. At the bottom of the furrow is a double row of large, flat cells, bearing short, broad rods placed end to end against those on the opposite side (Fig. 84). Outside -of them is a single row of smaller, cylindrical ones, bent in the form of a semi-circle and bearing short terminal rods. These retinal cells are filled with coarse, black pigment granules, so that only a narrow slit is left through which the light can pass to the circular space containing the large rods (Fig. 84).

Longitudinal vertical sections of the eye expose the arrange


ment of the cells on one side of the furrow (Fig. 81). It is there seen that the terminal edges of the broad rods trend toward the centre of the newly developed lens. The latter is placed so as to throw light upon the vertical furrow. But this condition is not permanent ; in the adult, the lens is horizontal, and its axis is directed toward the bottom of the eye. By this change, the amount of light that falls upon the vertical furrow is very much reduced.

The nucleated ends of the corneagen cells never lie in front of the furrow, but always to one side of it.

The 7'cthiopJiorcB of eyes I. and III. are like those of eye V. The secondary nucleus is always situated beyond the primary one, in the terminal quarter of the cell.

It is not so easy to detect the secondary nucleus in the gigantic cells as one might suppose. Their flattened ends make it difficult to roll them when isolated, so that one cannot view the same cell from different points. Moreover, they have a tendency to stick together in great flakes that appear, owing to the excessive thinness of the cells, to consist of but one rctinophora, when they may contain half a dozen or more. But in spite of these difficulties, I have seen enough to convince me that they are double cells and that the secondary nucleus is situated in their expanded and flattened ends about half way between the primary nucleus and the rod.

Although I have not made any extended observation on the structure of the retinophoras in the vertical furrow, their intimate connection with those of the horizontal retina, as shown by the early stages of development, and the presence of the double rods, leave no room to doubt that they have the same structure as the other retinophoree.

In only one instance have I seen anything like ganglion-cells in the retina. It was a small tripolar cell, with an outward prolongation extending along the sides of a retinal cell. Both cells were firmly united with each other, and were found among a great many isolated ones from all the eyes, so that it was impossible to determine to which retina it belonged. It was (PI. X., Fig. 58, Ji) very much like the tripolar cells found by me in the retina of Haliotis.

I have looked for these cells in sections, but have never found any trustworthy evidence of their presence. I regard

124 patten: [Vol. II.

them as half-formed ganglion-cells which have not succeeded in separating themselves from among the sensory cells, by transformation of which they arose.

Nerve Ends. — The manner in which the nerves terminate in the retina is practically the same as in eye V., while owing to the large size of the inner retinal cells, some points are brought out with great clearness. But the layer of pigmented nerve fibres between the rows of gigantic cells is a feature only found where these cells are especially well developed.

In Fig. 56, PI. X., is represented a section through the horizontal furrow of eye I. The sides of the cells are thickly covered with nerve fibres that converge toward the inner end of the cell to form a large bundle that might readily be mistaken for a continuation of the cell substance. The fibres follow, as nearly as possible, the contours of the outer and inner edges of the cell. This point is of importance, for it furnishes additional evidence that the rods belonging to these cells are terminal, and consequently developed at what is morphologically the free end of the cell ; for in the smaller retinal cells the nerve fibres always run parallel with the longitudinal axis ; hence the direction of the fibres shows in which way the cells have been bent.

At the distal end of the cell all the nerve fibres become roughly parallel, and develop refractive, spindle-shaped swellings, densely coated with pigment, the granules of which often form distinct lines, showing clearly the position and direction of the nerve fibres around which they are arranged ; or the granules may be collected in shapeless masses near the inner end of the rod, showing no traces of the arrangement of the nerve fibres they so effectually conceal.

At the opposite extremities of the cells the pigment granules are arranged in lines similar in direction to the nerve fibres ; they are also found on the nerve fibres after they have left the cell to form the optic nerve.

Beyond the spindle layer described above, the nerve fibres emerge from the pigment and extend over the surface of the rods in nearly parallel lines. At certain intervals these fibres expand into spindles, which were a sore puzzle to me before I had made a careful study of depigmented sections. They seemed to indicate quite clearly the division of the broad rods into segments, about the width of the smaller rods, a condition


that it was of course impossible to reconcile with the great size of the cells to which they belonged, unless we assumed that each of these cells possessed many rods.

In the smaller cells, the nerve fibres and pigment granules are also distributed in the manner just described, but their course and arrangement is less easily followed.

The small retinal cells contain an axial nerve which, between the two rods of each retinophora, breaks up into a fan-shaped bundle of fibres, that in cross sections appear like a row of dots midway between each pair of rods, and usually parallel with their broad surfaces (Fig. 55).

In the gigantic retinophorae, the axial nerves are arranged in a sheet as broad as the ends of the cells. In longitudinal vertical sections, they are consequently seen as a row of dots between each pair of rods (Fig. 55).

Most of the nerve fibres supplying the retinal cells come from two bundles that extend along the edges of the horizontal retina (Figs. 90 and 91). A third group extends along the middle of the under surface of the retina. Fro7n it arise two sheets of densely pigmented nerve fibres which extejid npzvard into the space bctzvecn the two rows of gigantic cells (PI. X., Fig. 56, v. n). These fibres are very straight and arranged with almost as much regularity as the large rods, with whose terminal edges they are parallel. I suspect that the number of these nerve fibres bears a pretty constant relation to the number of large rods. In some of my preparations I could see that a single nerve fibre lay in or near a shallow furrow which marked the point of union of each pair. From this I judge that in the living, normal condition, one of these vertical fibres extends along the terminal edge of each pair of large rods. But evidence of this condition is not often found in sections, owing to the action of the reagents, which cause a contraction that draws the fibres away from the rods towards the middle of the furrow. Hence in horizontal sections, these nerve fibres are seen in cross sections as two rows of dots, a little distance from the terminal edges of the rods (PI. X., Fig. 59, v. «.).

From the vertical nerve fibres arise a great many fine fibrillas which extend at right angles to the main fibres towards the rods and apparently become continuous with the external nerve fibres. Others extend in an opposite direction where, between the two

126 PATTEN. [Vol. II.

sheets of vertical fibres, they fonn a mass of the finest, interwoven fibrillcE, that presents the same appcai'ance as the medullary substance in the optic ganglia. It cannot be that this enigmatical mass of nerve fibres is nothing but a flake of coagulated serum. The latter is frequently present and is especially abundant after some modes of treatment, filling all the cavities of the head with a finely granular substance. At other times it is altogether absent, or at any rate the cavities appear quite empty. But under both these conditions, this mass of nerve fibres, unenclosed by any membrane and devoid of nuclei, maintains such a constant shape and uniform appearance as to preclude all thought of its being a product of coagulation. It may be remarked, also, that the coarse pigment granules, so abundant about this fibrous mass, are completely dissolved by acids, and consequently leave no granular residue that might be mistaken for nerve fibrillse. We have already pointed out that in the iris the pigment is also completely dissolved by these reagents.

One occasionally finds, among the fragments of cells isolated by maceration, what appear to be large flakes of pigment with parallel striations. They are fragments of the sheets of vertical nerve fibres covered with pigment, the fibres being united with one another by the innumerable fibrillas that arise from them.

Between the outer ends of the vertical nerve fibres is an oval space devoid of pigment, quite constant in shape and size (Figs. 90 and 91, PI. XIII. ). I have no suggestion to make concerning its significance.

The vertical fibres, and the pigment surrounding them, do not extend beyond the outer edges of the large rods.

The Rods in eyes I. and III. develop in the same way as those of eyes V, and VI. ; and although their arrangement in the fully developed eye differs in some important and interesting particulars from that in the last-mentioned eyes, their general structure is the same in both cases.

The horizontal rods are arranged with great precision and regularity edge to edge in vertical rows. The broad sides of two rods belonging to adjacent retinophorse lie so close together as to look like one rod with a faint vertical line in the middle. The rods belonging to the same retinophorae are separated by a clear space, a little smaller in diameter than that of the


retinophorse to which they belong. In the centre of the clear space is the vertical row of axial nerve fibres (PI. X., Fig. 55). At the lower end of each double row of rods are two pairs of much larger ones belonging to the large retinophorae at the bottom of the retinal furrow. They are arranged in pairs like those of the small retinal cells, except that the clear space between each pair is alternately wide and narrow.

The rods in the rows just above the gigantic ones differ from all others in that their free ends are united in pairs opposite the centre of the retinopJiorce (Fig. 60).

Toward the periphery of the retina, where the rods are upright, the rod-mosaic is composed of zigzag lines like those in eye V. (Fig. 54^).

Dorsal Appendage. — Besides the vertical, and horizontal retinas of eye I. there is a third part that up to this time we have left out of consideration. It is what in a former paper I have called the dorsal outgrowth of the large posterior ocellus. The observations recorded in this paper throw new light on that remarkable organ.

We have already described how the changes seen in surface views gave rise to an oval sensory patch constricted off from the anterior, dorsal end of eye I., but with which it was still connected by means of a narrow furrow (PI. VIII., Figs. 14-16). While the main portion of eye I. is being invaginated, the dorsal appendage is depressed, forming a deep pit, at first round, then oval, and finally slit-shaped, the long axis of the depression being at right angles with that of the main eye (comp. Figs. 8 and 14-16). In Fig. 79 is shown a longitudinal section of the main eye with a cross section through the middle of the dorsal appendage, during a stage corresponding to that shown in Fig. 16. In a similar section of a later stage, the pit is closed and three layers of cells are formed very much like those in eye V. (Fig. 80). The outer ends of the retinal cells soon lie close against the corneagen. The outer wall of the vesicle is flattened and probably forms an imperfect middle layer ; at least I have seen in several cases one or two nuclei just beneath the corneagen, which, I believe, must be referred to the outer wall of the vesicle. Fig. 16 shows that the sensory cells connecting the ocellus with its dorsal appendage are joined to the latter at its anterior edge, consequently only those sections that pass

128 PATTEN. [Vol. II.

through the anterior edge of both these structures — and not those shown in Figs. 79-81, which pass through the middle of the appendage — will show the continuity of their retinal cells. After the stage shown in Fig. 81, I have not been able to detect evidence of this continuity, although the vertical furrow contains retinal cells up to the very outermost edge of the eye.

Surface views indicate that the peculiar structures found in all the other eyes, such as the median rows of cells and the large nucleus, are probably present in the dorsal appendage, although imperfectly developed. My failure to detect them in sections may be due to the difficulties of observation.

In longitudinal vertical sections of eye I., the dorsal appendage is seen to be somewhat thickened in the middle, the rodbearing ends of the retinal cells converging toward an imaginary centre some distance above the corneagen.

The appendage is never sharply limited from the surrounding ectoderm ; it is less so in the older stages than during the period when it is an open cup or newly closed vesicle (Figs. 80 and 81). The basement membrane is continued without interruption from the surrounding ectoderm, over the inner surface of the appendage, and this tends to obscure still more the limits of the parts in question.

The retinophoras of the appendage are much like those of the other eyes. They are long, spindle-shaped cells with double rods, and hence probably contain two nuclei and an axial nerve fibre.

The observations recorded in this paper were made entirely upon material collected in a small pool near Milwaukee. In all these specimens the appendage was entirely devoid of pigment. ^ In the larvae that formed the basis of my preliminary note on the "Eyes of Acilius," in the first number of this Journal, and which were collected near Boston, Mass., the retinophorae of the appendage contained large blotches of pigment just below the rods.

Grenacher has described, without figures, the dorsal appendage in the larva of Acilius sulcata as being pigmented. He failed to recognize that it was composed of two distinct layers ; and while he thought it was in all probability some kind

1 In some of my sections I have noticed, since this was written, mere traces of pigment at the base of the rods in the Milwaukee species.


of a sense organ, perhaps one in course of development, he apparently had grave doubts about its being a visual organ ; at any rate he is careful not to call it such.

Eyes II. and IV.

Eyes II. and IV. arise from the distal edge of the first and second segments of the optic plate. They are so much alike that it will not be necessary to describe them separately. I have not been able to follow the very earliest stages in the development of these eyes, owing partly to their small size and partly to the fact that during these stages they are situated on the infolded edge of the optic plate, so that it is almost impossible to study them in surface views. Nothing in the sections indicates that they differ in any important points from the corresponding stages of eyes V. and VI.

The eyes are first seen in surface views (Figs, ^a and 5^), as round clear spaces containing an elongated dark area composed of a double row of nuclei nearly parallel with those of eyes I. and II. Between the rows of nuclei is a single large nucleus. The similarity between the structure of these eyes and the remaining ones makes it probable that the large nucleus, as in all the other eyes, is situated in the centre of four sense organs.

Eyes II. and IV. are not invaginated to form either optic cups or vesicles. Just as in eyes I. and III., as fast as the sensory areas on either side of the median ridge are tipped over, the ends of their cells meet in a vertical plane above the median ridge, and thus obliterate the cavity that tends to be formed.

Toward the close of invagination the cells on the outer edges of the sensory area curve outward and their sides meet in the median plane of the eyes to form the corneagen (PI. XII., Fig. 'jQ), whose nuclei form a row continuous with those of the retina, so that the latter appears to be directly continuous with the unmodified hyperdermis. This deceptive appearance is due to the fact that the nuclei of the corneagen cells over the centre of the eye stain very faintly, so that, in most cases, they escape, or even defy detection. In Fig. ^6 there is no trace of a middle layer of cells between the retina and corneagen. But in two cases, one or two cells were seen, with rather large and deeply stained nuclei, wedged in between the periphery of the

130 PATTEN. [Vol. II.

retina and corneagen. They were like the similarly placed cells described in eyes I. and III. (PI. XIIL, Fig. 85, o, w), and I do not doubt they are of the same nature, i.e. cells that have been pushed between the retina, and the corneagen, forming the rudiments of a middle layer. They never form groups of inverted sensory cells, as in eyes V. and VI., or a continuous layer of non-sensory ones, as in a part of eye VI.

TJie retinas of both eyes are asymmetrical, as in eyes I. and III., in that there are more upright rods on the dorsal side of the retina than on the ventral (Figs, "jy, 78).

Eye II. contains a greater number of horizontal rods than eye IV., and the median space in which they lie is considerably deeper.

The outer retinophorae are long and semi-circular, and nearly uniform in size throughout. The innermost ones are broad, scimetar-shaped cells, arranged with great regularity in two rows that extend the whole length of the longitudinal axis of the retina. Their free ends, which are bent at right angles, are almost as broad as the remainder of the overlying retina. In the younger stages, the terminal edges of the large rods are concave, and a considerable space is left between the two rows (Fig. 'jG). In the later stages they are rigidly straight, and the space is filled with a layer of densely pigmented, vertical nerve fibres, like those in eyes I. and III. (Fig. "]%).

In all essential particulars, the finer structure and arrangement of the retinophorae and their rods, and the distribution of nerve fibres, is the same as in eyes I. and III.

After the corneagen has formed a continuous layer over the retina, the large nucleus lies just below and between the inner edges of the two rows of giant retinophorae. It stains deeply on the periphery, and contains a few coarse, deeply stained granules. It is always oval, with its long axis parallel with the long axis of the retina, and is somewhat compressed, 'as though forceably flattened between the two rows of cells. It remains in this position, unaltered, through larval life.

Eyes II. and IV. are somewhat canoe-shaped, the retinal furrow being at first nearly parallel with that of eyes I. and III., below which they lie in the earlier stages ; but they finally change their position for one behind the deep end of eye I. During this change, they rotate on their optical axis through an


angle of about ninety degrees, so that by the time the embryo is ready to hatch, their retinal furrows are at right angles with those of eyes I. and III. {compare Figs. 5^, 8, and \oa). Surface views of eye IV., in its younger stages, show that its retinal furrow is bent, a condition which is retained through life (Fig. Za).

The Lens of eye II. (Fig. yf) is asymmetrical, in that the radius of curvature of its strongly convex inner surface is greater on the dorsal, than on the ventral side. The axis of the lens falls upon the layer of short, nearly upright rods on the dorsal side of the retinal furrow. The latter, which is the most specialized part of the retina, lies, therefore, very much to one side of the apparent optical axis of the eye. The same fact was observed in eyes I. and III. This extraordinary condition is difficult to account for satisfactorily. There is every reason to believe that the retina has not been moved by shrinkage from its normal position in relation to the lens. The greater depth of the iris on the dorsal side, the extension of the rods in that direction, the slight flattening of the dorsal, inner surface of the lens, and the whole history of development, show too clearly that these diverse parts are modified in this manner in order that the optical axis might have the direction indicated.

Although in eye IV. the lens is more symmetrical, its axis falls upon the retina some distance above the furrow.

In the larva, the optical axis of eye II. is directed upward and backward, that of eye IV. downward and backward.

II. The Optic Ganglion.

As soon as the cephalic lobes have assumed their characteristic shape, a slight depression, the future mouth, appears in the median line between them. On either side of this depression are three ectodermic thickenings from which the brain is subsequently developed ; they appear to be direct combinations of the segmental thickenings of the ventral nerve chord (Fig. 2, b^-^). On the lateral edge of the third pair of thickenings, l^, are the antennae, c^. On the inner edge of the second pair are two appendage-like outgrowths of the ectoderm, that finally unite in the middle line above the mouth to form the labrum, dS.

132 PATTEN. [Vol. II.

Between the brain and the optic plate (Fig. I.) is a broad expanse of ectoderm composed of faintly stained cells containing coarsely granular, or even flocculent, protoplasm and large spherical nuclei.

In PI. IX., Fig. 19, which represents a cross section through the middle of the cephalic lobes of an embryo considerably younger than that shown in Fig. i, there is no thickening to form the brain, although the distal edges of the lobes have already given rise to the optic plate, on the inner edge of which is a depression, containing large wedge-shaped cells quite different in appearance from those on either side of them {e.g.). This depression, the beginning of the invagination of the optic ganglion, is fairly uniform in depth, and in surface views appears like a semi-circular furrow on the inner edge of the optic plate.

The ectoderm that gives rise to the optic ganglion divides into tJiree segments, or lobes, each of which is Jinited 07i the one hand with a seg7nent of the brain, and on the other, with a segment of the optic plate (Fig. i, 0. g^^).

The semi-circular groove is soon deepened to form two distinct pits with slit-like openings (Fig. i, g. v}'"^). There is no infolding between the third segment of the optic plate and the third segment of the optic ganglion, but there is, as shown in sections, a distinct inward proliferation of ganglion-cells at this point. Each segment appears in some cases to be divided by a faint line into two parts, a condition that may have some connection with the fact that it belongs to a part of the optic plate provided with two eyes.

Figs. 20-23, PI- I^-» represent four cross sections of the cephalic lobes during the stage shown in Fig. i. The first section passes through the first ganglionic invagination, g. v}, the lateral wall of which is composed of a single layer of loosely connected cells, continuous, at the opening of the pit, with the edge of the optic plate. There is no infolding between the anterior edges of the optic plate and first ganglionic segment (Fig. i). The invagination appears just behind this point as a slight furrow that increases in depth backwards as far as the anterior end of the second segment of the optic plate.

The section shown in Fig. 21 passes through the bridge of ectoderm separating the first invagination from the second. Beneath the uninfolded layer is a V-shaped mass of ganglionic


cells, the cavity of which is continuous with that of the first ganglionic invagination.

In the next two sections there is a solid chord of cells in the place of the V-shaped mass of Fig. 21. It connects the cells of the first ganglionic invagination with those of the second.

The succeeding section (Fig. 22) passes through the second pit, ^. v.^

That shown in Fig. 23 passes through the base of the antenna and the point marked ^. v.^, Fig. i. It shows at g. v^ a cross section of a thick chord of ganglion-cells extending from the third ganglionic segment to the under surface of the third ocular plate. It is produced by an inward proliferation the direction of which is parallel with the long axis of the antenna.

The invaginated part of the optic ganglion now forms a continuous semi-circular mass of cells tucked beneath the optic plate. It is directly continuous along its inner edge with the distal inner edge of the optic plate, p. n.

This fact may possibly throw some light on the inexplicable presence of an optic, and ganglionic invagination side by side. My studies on Molluscs and Arthropods led me to suppose that ganglion-cells were derived from sensory cells that had wandered into the underlying tissues, leaving their outer ends, transformed into nerve fibres, sticking in the epidermis. Confirmation of this supposition was found in Pecten. Much better evidence will be given in describing the origin of the large tripolar cells in the optic ganglion of Acilius.

But the point I wish to emphasize now, is, that from the earliest stages, the inner surface of the optic plate is continuous with the optic ganglion (Figs. 20-23,/. n), and from what takes place later, it is highly probable that ganglion-cells are formed at this period by an inward proliferation of the optic plate.

My idea is that an increase in the width of the optic plate, since its distal edge is fixed, would produce a fold like that in Fig. 20. The main part of the optic ganglion of this period may be regarded as a formerly sensory area with an underlying plexus of ganglionic cells continuous with a similar plexus underlying the eyes. With the great development of the latter, all the cells of the sensory area were converted into ganglionic ones, which were then overgrown by the optic plate and added to the plexus arising directly from the eyes. The

134 PATTEN. [Vol. II.

general character of the invagination points to this conclusion, and it seems hardly possible to reconcile, by an explanation along any other line, the independent development of such dependent structures as the eye and the optic ganglion.

The sections shown in Figs. 24-27 belong to a little younger stage than those shown in Figs. 20-23. They were cut from an embryo somewhat abnormal, the antennae being lodged in a great depression, on the floor of which was the mouth. There was nothing abnormal about the parts belonging to the eyes. The first and second ganglionic invaginations are shown particularly well.

All these sections and those described above are instructive from the fact that there cannot be a shade of doubt that the ganglionic invaginations have nothing whatever to do with the formation of optic vesicles, for the peculiar way in which the eyes develop makes it possible to determine, even at this period, just what parts develop into the eyes and what into the optic ganglion.

These two series of sections show pretty clearly that there are two distinct, and one obscure, invagijiations to form the optic ganglion. This fact is of especial interest tvhe?t we consider that there are tJiree segments to the brain and three to the optic plate, and three distinct parts to the optic ganglion of the convex eye.

In Fig. 3 a the three ganglionic segments, 0. g^^ are still visible in surface views, and at the anterior ends of the first two are seen the openings of the ganglionic invaginations,^, v} and g. v?'

In a later stage (Fig. 4), nearly all the eyes, as well as the mouths of the two invaginations, are visible. Still later (Figs. 5 a and 6) the invaginations have disappeared, and all but a small part of the optic ganglion is concealed by the optic plate.

Just after the rupture of the embryonic membranes, the optic ganglion is completely shut off from the surface, and the ectoderm of the optic plate becomes continuous with that over the brain.

Figs. 28-34, PI- IX., represent seven sections of an embryo about the age of that in Fig. 4.

The optic ganglion now forms a great, complex mass of cells, the most of which is not yet overgrown by the optic plate. It requires careful attention, not only to recognize its parts, but to distinguish them from the brain.


The first section (Fig. 28) passes just above the anterior edge of the plate, and shows the proximal ends of each ganglionic segment, 0. g}'^

The next (Fig, 29), passes through the first invagination, g. v}, and the anterior end of eye V. ; compare Fig. 4. At this niveau, the third ganglionic segment, 0. g.^, is separated from the third lobe of the brain by the cephalic infolding, s.

In Fig. 30, the section passes through the centre of the third segment, which is not separated by any epithelial layer from the exterior. At x.f. is a section of the second segment of the optic plate, where it dips down towards the second ganglionic invagination. The underlying cells, 0. g.^ and 0. g.^, connect the main part of each ganglion with the segment of the optic plate to which it belongs.

In Fig. 31, the section passes through the upper edge of eyes V. and III., and in Fig. 32, through about the middle of these eyes. The scattering fibres uniting the eyes with the optic ganglion indicate the position of the future nerves.

In Figs. 33 and 34, the sections pass through the upper and lower edges of eye VI., and show how the optic ganglion is here reduced to a chord of cells continuous with the retinal cells of the eye.

After the rather confusing condition shown in the last series of sections, it will be a relief to find in the next (Figs. 37-39), a simpler arrangement. These sections are taken from a head a little older than that shown in Fig. 4. The first section (Fig. 37) passes through the site of the first ganglionic invagination. It is now closed, but there is a depression above an inwardly projecting core of cells, which marks its former position, g. v} It also cuts the anterior edge of the first ocular plate, showing a small part of the first ganglionic segment, o. g} Two or three sections below this we find one (Fig. 38) which passes through the point g. vP', Fig. 4, and consequently through about the middle of eye I. and the anterior edge of the second ocular segment. The third segment of the optic ganglion is still exposed to the exterior, 0. g.^ The outer wall of the ganglionic invagination has completely disappeared by this time, and the edge of the optic plate comes squarely against the outer surface of the third ganglionic segment. The only part of the brain seen at this niveau is a section of the anterior end of the antennary lobe.

136 PATTEN. [Vol. II.

It is separated from the optic ganglion by the invaginated ectoderm, z. After three or four more sections, we come to one (Fig. 39) that passes through the middle of the fifth eye and the lower end of the first. It should be remembered that the sections are cut from an embryo a little older than that in Fig, 4, and that the posterior end of eye I. is on a level with eye V. We see that the continuity of the optic plate with the optic ganglion is now interrupted everywhere except along three lines directly beneath, and parallel with, the median furrow of the eyes. These three connectives are short and nearly as broad as the optic plate. Compare Figs. 4 and ^a. They are the rudiments of the optic nerves, and are composed of a mixture of nerve fibres and ganglionic cells.

If we refer to Fig. i, and the sections of that age, Figs. 2023, we shall see that the first cross section passes through the first ganglionic invagination, with whose direction of ingrowth it is nearly parallel. Now if we suppose a section to be cut through the long axis of the optic plate at that stage, we should then see three bands of cells — the optic plate, the middle wall of the ganglionic fold, and the optic ganglion itself. Owing to the way the cephalic lobes have developed, a cross section of the head shown in Fig. 4 will give a picture something like our imaginary one, only the middle wall of the fold has disappeared, and the inner one, the optic ganglion, is bent so as to appear like a thick, imperfect ring (Fig. 39).

The brain at this period consists of a medullary core enclosed in a thick layer of ganglion-cells. The latter are continued, without any perceptible modification or break, into those which constitute the optic ganglion, which, up to this time, contains no medullary core. From this time up to the latest stages, there is no way of distinguishing the parts of the optic ganglion except by their position, and the nerves which go from them to their respective ocular segments,

I have represented two more sections from a head a little younger than that in Fig. 6. The first (Fig. 35) passes through the upper edges of the first and second ocular segments, and shows, besides eyes I. and III., the anterior edge of the optic ganglion. In the next section below, not represented in the plate, the optic ganglion is seen in section as a circular mass of cells, apparently cut off from the brain by the invaginated ecto


derm, s. It contains in the centre a round medulla, the downward continuation of the fibrous mass, m. b. in Fig. 35. The second section (Fig. 36) passes through about the middle of the first and second ocular segment, and just above the third. The ring of ganglion cells is now interrupted on its median edge, forming a narrow inlet into a central space lined with the cut ends of medullary fibrillae. The whole cavity is soon filled with a mass of fibres which develop into the medullary stalk of the optic ganglion, s. o. g., which afterwards serves to connect the different ganglionic centres with the brain.

In all sections of the optic ganglion during the younger stages, its inner face is sharply defined, as though bounded by a delicate membrane, something like the basement membrane on the inner surface of ectodermic layers. In the older stages this membrane disappears, and the optic ganglion becomes a mass of loosely connected cells.

In the section shown in Fig. 36, and in those of the same series below it, the short bundles of nerve fibres connecting the optic ganglion with the eyes, pass completely through the ganglionic layer, at the inner surface of which they bend upwards, and pass through the centre of the ganglion into the brain.

Medulla of the Optic Ganglion. — It is necessary to understand the position and arrangement of the larval medullae in order to comprehend the structure of the optic ganglion of the imago. They become definitely established soon after or during the rupture of the embryonic membranes, and may be studied in a series of sections (Figs. 42-47) cut from a head like that in Fig. 8.

The first section (Fig. 42) passes through the upper ends of eyes I. and III. and cuts through the optic ganglion at its junction with the brain, showing the cut ends of the medullary fibres as they bend from the brain down into the centre of the optic ganglion. After four more sections, we come to one (Fig. 43) passing through the upper end of the optic ganglion, in the centre of which is a circular mass of medullary substance bounded by a layer of minute dark cells, continuous with those surrounding the medulla of the brain.

In the fourth next section (Fig. 44) the optic ganglion appears as an oval mass of cells close beneath the optic plate. The

138 fatten: [Vol. II.

roots of nerves I. and III., seen in longitudinal sections, are

somewhat swollen and divided into two parts.

The fibres at the ends of the nerve roots turn upwards and pass into the stalk of the optic ganglion. Their cut ends do not lie on the inner surface of the ganglion, as during the earlier stage (Fig. 39), but in the centre. On the ventral side are seen medulla 3 and sections of the roots of the nerves supplying eyes V. and VI.

The next section (Fig. 45) passes through eye V., and shows its double nerve in longitudinal, and that of eye VI. in cross section. It is below the level of eyes I. and III., but shows eye II. and its minute nerve root with its adjacent large ganglion cell.

Fig. 47 represents a section passing through eye VI. and its nerve just as the latter bends at right angles. It is clothed with nerve cells up to the very base of the eye.

This series of sections shows the principal features of the optic ganglion at this stage. The parts are lettered in a way to show at once their relations to the eyes. They should be carefully examined and compared with Fig. 8, for a better understanding of the parts will be obtained in this way than by a long and tedious description.

Toward the close of embryonic life most of the nerve cells are situated on the dorsal surface of the ganglion, where the latter joins the brain.

In Fig. 48 is represented the optic ganglion as seen in a cross section of an embryo about ready to hatch. The section passed through the head between eyes II. and IV. showing a little of both. Fig. 8 represents a younger embryo than the one we have in mind, but it serves perfectly well to show the level of the section. As the head is bent at right angles to the body in such embryos, the section would be the same as a longitudinal, horizontal one of the larval head, Fig. 10. Again, the section we are about to describe would cut that shown, in Fig. 49, taken, however, from an older head, just below the crown of ganglionic cells on its dorsal surface. At this period, the shape of the medullae and the characteristic arrangement of the fibres in them can be studied to best advantage.

Suppose the side of the head to which belongs the optic ganglion we are describing, laid flat on the paper, opposite the

No. I.]



section of the optic ganglion. An outline drawing of the eyes in this position would then be almost the exact mirror-image or "Spiegelbild," as the Germans call it, of the medullae of the optic ganglion (Fig. i, wood-cut).



Figure i. — A. Projection of the retinas just affer the rupture of the embryonic membranes. B. Horizontal section of the optic ganglion during the same period. I.-VII. = eyes. 6 = cluster of inverted cells derived from sixth sense organ of eye VI. ; d. a.= dorsal appendage; k. r. = horizontal retina; v. r. = vertical retina; V. /.— vertical fibre; m^ = plate of medullary substance between the vertical fibres, corresponding to a similar plate in the retina, w. III. = medulla of eye III.; and m. d. a. = medulla of dorsal appendage, etc. The two cuts illustrate the correspondence between the structure of the medullae and that of the retinas.

Medulla 3 (Fig. 48) is long and flat in section and somewhat pointed at the outermost end. It is divided into two unequal parts by a double row of coarse fibres. Between them is a clear space containing a long line of extremely fine medullary substance, which I have not succeeded in dissolving into fibrillae, although I do not doubt it has such a structure from its great similarity to the medullary substance on either side of it. If we compare this part of the medulla with a horizontal section of the retinal furrow of eye III., we shall see (Fig. 59) that in both, there is a double row of coarse fibres seen in cross sections ; the rows are separated by about the same distance, and between them is a thin layer of interwoven fibrillae which in the retinal furrow are deeply pigmented.

On either side of the double row of cut fibres, there is a layer

I40 patten: [Vol. II.

of felted fibrillcC, that on the posterior outer side being much thicker than the other. If we turn again to the retina of eye III., it will be seen that the part on the posterior side of the furrow is much broader than that on the opposite, and exceeds it in width by about as much as the posterior layer of fibrillae in the medulla exceeds that on the anterior side. That this extraordinary correspondence between the structure of the eye and its medulla is no fanciful one, may be proved by referring to any other medulla. In each case, we shall find that the striictiwe of the medulla is like that of the retina to wJiich it belongs. For instance, in eye V. there are remnants of a median row of peculiar cells which undoubtedly represent the gigantic cells, such as those in eyes I. and III. In the embryos just ready to hatch, the medulla of eye V. is divided in the middle by a layer of coarse fibres, the cut ends of which are distinctly visible in cross sections, so that the outline of the medulla, with its median row of cross fibres, is the mirror-image of eye V. with its median row of cells. In the full-grown larva, the latter are much smaller and less conspicuous, and we find on turning to a section of the medulla that the median row of coarse fibres has degenerated in a corresponding degree.

In eye VI. the upright retinal cells are uniform in structure throughout, the median ridge seen in surface views of the younger stages having disappeared. But there is a remarkable bundle of inverted cells on the ventral side of the retina, and this peculiarity is expressed in the medulla of that eye by a small bundle of fibres in a corresponding position. Still again medulla I. shows by its configuration that it belongs to eye I. There is the narrow bent portion belonging to the vertical furrow, the broad inner part to the horizontal retina, and at the outermost end of the medulla is a small bundle of fibres belonging to the appendage. To complete the correspondence there are the three rows of fibres in the middle, just as described in medulla III., and on either side a layer of medullary substance, the thicker layer being on the anterior edge, corresponding to the greater width of the retina on that side of the furrow.

All the optic nerves are composed of coarse fibres arising from the medullae. So far as it is possible to determine, each fibre is composed of the prolongation of the external and internal fibrillae of the cell with which it is united. The latter are


probably continued into the medullae without losing their identity. There is reason to believe, from the similarity in structure of the retina and the medullar, that the fbrillcs of the latter are rearranged in a manner corresponding with that of the fibrillcB tn the retinal cells. Now if my supposition is correct, we ought to find in the medullae of eye I., for example, systems of fibrillae similar in number and arrangement to the retinidia of the retina. We find indications of such a condition, for there is a furrow in the medulla containing a double row of fibres, between which is a layer of very fine medullary substance. To make the comparison complete, the cross fibrillas arising from the fibres ought to unite with distinct clusters of fibrillae corresponding in composition and number with the retinidia on either side of the retinal furrow. But although we cannot distinguish these mednllary retinidia, since all are united to form one continuous mass, there is reason to suppose they really exist, for, as we have shown above, where there is a break or any marked peculiarity in the arrangement of the retinal retinidia, we find a corresponding change in the structure of the medullae. For example, if there are more rods and consequently more retinidia on one side of the retina, there is an increase in the thickness of the medulla on its corresponding side. If there is a bend in the retina, there is a similar one in the medulla. If there is a furrow in the retina, containing a double row of coarse fibres, there is a space in the medulla that also contains a double row of coarse fibres, and if the furrow is absent, the space is absent also ; and we might enumerate a number of other peculiarities in the structure of the retina that had their counterpart in the medulla.

I am not prepared to discuss the conflicting hypotheses that have been advanced concerning the significance of the medullary substance, but I desire to call attention to the fact that its peculiar structure in Acilius may throw some light upon its function.

I venture to suggest the idea that occurred to me. The medullae may be regarded as the retinas of the mind, the inner eyes, in which are reacted the nerve changes produced in the external ones. The whole apparatus may be compared to a telephone in which a vibration of the air produces a movement of the receiver, an intermittent flow of electricity follows, and

142 PATTEN. [Vol. II.

finally, at some distant point, perhaps, a second vibration is produced exactly like the first. We may suppose that in the eye, light causes some change — it may be vibrations of the retinidial fibrillae which are transmitted, not as vibrations, but as chemical changes, along the fibres of the optic nerve to the medullae, where they give rise to other vibrations exactly like those produced in the eye. The changes aroused in each retinal cell are transmitted along a bundle of wires to the medullse, where they are united in a proper sequence of time and place, which if we could see or hear, we should recognize as the same symphony of activities produced in the eye. An essential feature of the telephone is the similarity in structure and action of the two ends. It v/as the similarity in structure between the retina and the medullas which led me to infer that the activities aroused in one must produce similar activities in the other.

The fibrillae at the proximal ends of the medullae unite to form the m.edullary stalk to the optic ganglion, which is continuous with the medullary substance of the brain. The medullae of those eyes belonging to the same segment unite first, that is, medulla I. with II., III. with IV., and V. with VI., so that three stalks are formed, which almost immediately unite with one another to form the common stalk of the whole optic ganglion. Thus another intimation of the threefold structure of the optic ganglion is given.

On the dorsal surface of the young larval optic ganglion are some small, deeply stained cells, crowded together to form three ill-defined clusters, one on the side and dorsal surface of each of the three branches of the medullary stalk. The nerve cells send downwards single fibrous prolongations, which, after uniting with one another to form ill-defined bundles (Fig. 48, /. n.f}j, pass to the sides of the medullae where they appear to break up into fine fibrillae. The latter issue from the distal end of the medulla and immediately unite to form coarse fibres about the size of those which arise directly from the ganglionic cells. These fibres then unite in large bundles to form the optic nerves.

Between the medullae are scattered cells, which in the younger stages cannot be distinguished from other cells in the brain and optic ganglion.


In the larvae, however, after some modes of treatment, they appear to be much smaller and stain more deeply. They then resemble the small dark cells which immediately surround the medullae of the brain and the ventral nerve cord.

The optic ganglion of the late embryonic, or early larval, stages does not undergo any noteworthy change until toward the beginning of the pupal period. I never succeeded in finding or raising puppe, but feel confident that the oldest larvae obtained had about reached that stage. A view of the dorsal surface of part of the brain and optic ganglion at this period is shown in PL IX., Fig. 40; and sections of it in PL X., Figs. 49 and 53.

In Fig. 40 we see beneath the crown of ganglion cells the faint outline of the medullae, and still further below, the continuation of their distal ends into the optic nerves. The ganglion cells of the second segment, o. g?, now form a semicircular band that almost encloses the other two. They are recognized in sections by the clear protoplasm of their columnar cells (Figs. 49 and 50).

The cells of the third segment are crowded toward the brain, where they form a narrow collar around the stalk of the optic ganglion. Above the middle of the stalk, the band lies at the bottom of a cavity, the roof of which is formed by over-arching ganglion cells (Fig. 50, o. g?). Towards its anterior end it is wider and folded double (Fig. 49, o.g?). It is directly continuous on the one hand with the cells of the second segment, and on the other, with a broad layer of large scattering cells that fill up the space between it and the third segment (Figs. 40, 49 and 50).

Horizontal sections below the crown of nerve cells have a wedge-shaped outline in which the medullae have much the same arrangement as in Fig. 48. Fig. 51 represents such a section at the posterior edge of which is shown a part of the second ganglion segment. A section a little above this (Fig. 52) shows part of the second and first segments, and finally a still more dorsal section (P'ig. 53) shows the second and third.

In Figs. 49 and 50 the optic ganglion is seen in two vertical sections, the first, near the anterior edge, the second, near the middle. By comparing these sections with the surface view

144 PATTEN. [Vol. II.

in Figs. 40 and 41, a fair idea will be obtained of the structure of the optic ganglion at the close of the larval period.

Optic Nerves. — In the younger embryonic stages, the fibres connecting the eyes with the optic ganglia are so intermingled v/ith ganglion-cells that it is not easy to distinguish them from the optic ganglion. Even in the larval stages, when the optic ganglion is situated behind and some distance away from the eyes, the optic nerves are in places so ill-defined that it is difficult to follow them to the medullae from which they arise.

The nerve to eye V. descends from the outer anterior edge of the optic ganglion, and after reaching the inner lower edge of eye III. turns nearly at right angles and runs forward to join the eye on its posterior dorsal face.

Nerve 6 is the most compact of all. Its root lies on the inside of that belonging to eye V. The two nerve roots extend downwards nearly parallel with each other to the lower edge of the ganglion ; and then nerve 6 bends backwards, and sweeping around the lower posterior edge of eye I., terminates on the posterior edge of eye VI.

Nerves i and 3 extend downward and forward from the lower anterior edges of their respective medullae as two great sheets of loose fibres. At the lower, inner edge of eye I., those of nerve i unite to form a more compact mass which suddenly bends outward and divides into five bundles, two of which run along the under side of the eye on either side of the retinal furrow. Two other bundles extend upward on either side of the vertical furrow. The fifth bundle extends upward to the dorsal appendage (Fig. 18).

The nerve to eye III. is similar to that of eye I. in shape and in the direction it follows. On reaching the eye it divides into two branches which extend parallel with each other along the under side of the retina.

In both eyes I. and III., fibres extend along the median under surface of the horizontal retinas, and finally bend upward to form, between the rows of gigantic cells, the vertical fibres.

There is nothing noteworthy about the course of nerves 2 and 4.

The relative positions, in the full-grown larva, of the nerves, the medullae, and the eyes, throw some light on the relation of the eyes to one another.


It will be seen that in the later stages the medullas show more clearly their primitive relations to one another than do the eyes. At this period, it is impossible to recognize which eyes belong to the same segment, or that there is any paired arrangement at all.

Neurilemma. — The brain, optic ganglion, nerves, and eyes are suspended or enclosed in a common envelope continuous with the basement membranes and derived from the ectoderm. During the youngest embryonic stages, the inner surface of the ectoderm, especially of the thickenings that give rise either to the eyes or nervous system, is covered by a delicate membrane which in the majority of cases is quite devoid of nuclei. When the eyes are invaginated, the membrane is pushed in likewise, but it is unaffected by the shifting of the cells to form the corneagen and the outer wall of the optic vesicle, and is continued from the indifferent ectoderm over the bulb of the eye to the optic nerve.

The neurilemma of the brain and optic ganglion is formed in a similar manner. As the ectodermic thickenings that give rise to these organs separate from the parent layer, they are suspended in a sort of sling formed by the basement membrane, which is distinctly nucleated where it surrounds the developing organs. As the latter separate from the ectoderm, the niembrane surrounds, but does not completely enclose them, for at quite late periods one sees at certain places the two limbs of the membrane close together, suspending the brain, as the intestine in its mesentery, to the wall of the head (PL X., Figs. 43 and 44). Wherever the membrane is still attached to the surface ectoderm, the cells are drawn out into long fibres with nuclei at various heights. Some cells are entirely separated from the others, taking up a position on the outer surface of the membrane, in the formation of which they seem to take a part. As the outer surface of the basement membrane becomes the inner one of the neurilemma, we find that the nuclei of the latter, during embryonic life at least, are on its inner surface (Figs. 42 and 47).

In the adult larvae, the suspending membranes are reduced to cords of fibres mixed with nuclei extending from the posterior, median edge of each brain lobe to the roof of the head.

In Figs. 42-47, are shown different stages in the formation of

146 PATTEN. [Vol. II.

the neurilemma as seen in different parts of the head. In Figs. 42 and 43, its continuity with the basement membrane, and the drawn-out ectoderm cells are well shown, and in Fig. 45, the double membranes suspending each lobe of the brain.

The same membrane forms a common envelop for the distal end of the optic ganglion and the roots of the nerves, but beyond this point each nerve has a separate sheath.

In the full-grown larvae, the neurilemma is thicker and laminated, with small flattened nuclei in the middle of the layer. I have seen some cases when it appears to be composed of tv/o thin membranes with nuclei between.

Origin of Ganglion-Cells.

During the earliest embryonic stages, the whole inner, distal edge of the optic plate is connected with the optic ganglion by a mass of tissue composed of fibres mixed with ganglioncells. The connection is gradually broken everywhere except beneath each eye, where broad bands of tissue remain forming the rudiments of the optic nerves. It will be more convenient to follow the development of the ganglion-cells in eye V., since the different stages of this eye have been described and illustrated in most detail.

Near the connecting bridge of cells in PI. XL, Fig. 61, n. g. c, the ganglionic nuclei are smaller, cell boundaries have disappeared, and the tissue assumes an appearance more like that in the optic thickening. In fact, there is a gradual transition from the retinal cells to those of the optic ganglion. A careful examination shows that among the half-ganglion, half-retinal cells, are many fibres of varying size, some large enough to be parts of slender cells, others so minute that they are seen with difficulty. The larger fibres are the outer ends of newly formed ganglion-cells; the minute ones are similar parts of cells whose outer ends have been converted into true nerve fibres. Such a condition as this may be seen in sections of eyes from the stage shown in Fig. 4, PL VII., down to that in Fig. i. In the older stages the tissue connecting the eyes and optic ganglion has become almost devoid of cells and is composed of closely packed fibres. The migration of sensory cells from the eye has nearly ceased, but one or two cells may still be seen, which, owing to their enormous size, offer special facilities for following the pro


cess in detail. In Fig. 62 is shown one of these cells just as it has reached the inner surface of the eye. It arose from one of the large nuclei found among the closely packed retinal ones, from which they are distinguished by their large size and coarsely granular contents. As these nuclei descend into the clear space at the inner surface of the eye they increase rapidly in size, and a great mass of finely granular protoplasm collects around them so that they become very conspicuous objects in sections of the eye at this period. The cells are distinctly tripolar. One of the prolongations is always directed outwards. Its base, which is broad and filled with protoplasm like that in the cell, is continuous with a tubelike prolongation. Owing to its wavy course and to its clear contents, which cause it to stand out clearly against the dark protoplasm of the retinal cells, it is not difficult to follow this tube a considerable distance into the tissues of the eye, where it is finally reduced to a thin refractive fibre that disappears between the retinal cells. The two inner prolongations do not become visible until the cell has reached the innermost part of the eye. They are less conspicuous than the one just described, for they are smaller, and the dark protoplasm extends into them a short distance only. They usually arise from opposite sides of the inner surface of the cell and extend in opposite directions at right angles to the outer fibre. But there is considerable irregularity in the position of the fibres. In some cases it looks as though the nucleus moved inward faster than the rest of the cell, leaving the inner fibres behind. It frequently happens that the latter are directed at right angles to the section plane ; the cell then assumes a shape like that in Fig. 82. One often sees a cluster of four or five enormous cells midway between the eye and the optic ganglion (Fig. 63, n. g. c). Most of these cells are converted by rapid division into small pear-shaped ones which gradually move inwards along the optic nerve, forcing a way for themselves between the fibres, until they are finally lost among the similar cells of the optic ganglion. But one of the cells retains its great size. When it reaches the optic ganglion, it is pushed to one side, out of the nerve root ; its inner end is swung outward toward the periphery of the ganglion and the outer prolongation, which is now more conspicuous than ever, is bent almost double. Between the bend and the cell itself, the pro

1^8 PATTEN. [Vol. II.

longation is filled with dark granular protoplasm. Beyond the bend it is reduced to a coarse, colorless, and refractive fibre that may be followed outward into the eye. As development goes on, the bend of the fibre is carried more and more into the optic ganglion until it reaches the medulla, where it becomes hyaline and refractive and breaks up into a tuft of a dozen or more fibrillas that extend along the sides of the medulla and then turning inward, disappear (Fig. 48).

The outer end of the cell is somewhat flattened, and on either side is continued into a tube-like prolongation with clear, homogeneous contents. In the centre of the tube, I have seen on several occasions what appeared to be a small fibre, but cannot say positively that it was such. The protoplasm of the cell protrudes a little near the mouth of the tube, but is not continued into it.

These tubes, or fibres, are difficult to follow, and I have only succeeded in doing so in a few cases, and then not very far. They extend, at right angles to the great inner stalk, in opposite directions along the under surface of the neurilemma, and I should judge they served to connect the ganglion-cells with one another. I see no reason to suppose they extend into the medullse; they do not run in that direction, and no fibres were seen running into the medullae that could not be referred to the stalk-like prolongations of the ganglion-cells.

The gigantic ganglion - cells continue to divide, but with decreasing frequency, up to about the time of hatching. At that period, there is usually but one cell of enormous dimensions on the side of each medulla. Their position, size, and shape, as shown in Fig. 48, are very constant. In rare cases, there may be in the larv3e two large cells side by side, one a little smaller than the other. In the full-grown larvae, the cells are proportionally smaller and less conspicuous.

In Fig. 48 is shown a pair of ganglion-cells placed end to end, and easily distinguishable from the surrounding ones by their peculiar shape. There was an exactly similar pair in the optic ganglion on the other side of the head, and I have also seen them in one or two other cases, so that I do not doubt that they are usually present during these stages.

There is a singular modification of the sixth gigantic ganglioncell, in that its inward prolongation is divided into two branches,


one of which is continued half way across the optic ganglion to the fifth medulla.

There are some minute fibres scattered about among the cells of the cortical layer, that I cannot trace to distinct cells. They are seen running in an irregular manner from one cell to an other, or from the cortical layer to the neurilemma. They often extend over the surface of the large cells branching in all directions (PI. X., Fig. 48). They are not numerous enough to form a sheath or envelop about the cells, such as that so frequently described in the Annelids, although it is not improbable that they belong to the same category. They may possibly arise from the small, dark cells, sometimes called the inner neurilemma, that form an investment for the medulla, and which are specially well developed in the brain and ventral nerve chord.

The fact that a not inconsiderable number of the ganglioncells arise directly from the gigantic, tripolar ones might lead one to suppose that all the cells of the optic ganglion were tripolar and differed from the large ones only in size. All that one can actually observe points towards this conclusion.

In my paper on the eyes of Vespa, page 198, attention was called to certain conical or pear-shaped ganglion-cells at the inner edge of the middle lobe of the optic ganglion, and it was stated that some of the conical ones were possibly tripolar, for indications of two prolongations at the broad distal ends of such cells had occasionally beea seen. A re-examination has led me to believe that all of them are tripolar, like the large ones in Acilius. Their deceptive, unipolar habitus is due to the great size of the inner prolongation as compared with the very delicate outer ones.

As there is no reason to suppose that more than one kind of cell is present in the cortical layer of the optic ganglion, — we do not include the minute, dark cells surrounding the medullas, or those interspersed among the fibres of the optic nerves, concerning whose minute structure we are ignorant, — and in consideration of the facts we have presented concerning the origin of ganglion-cells, we have arrived at the following conclusion : The cortical layers of the optic ganglia are composed of large and small tripolar ganglion-cells. The large, inward prolongation of each of these cells is filled with granular protoplasm contijinotis with i/iat in the cell body, and represents a part of the otitcr end

I^o patten: [Vol.11.

of a formerly sensory cell lodged in tlie ectoderm, or of one derived by divisioti from, stick a cell. In passing tJirongJi the medtdla this prolongation breaks np ijtto a bundle of fibrillce. The other two prolongations are small, ajid contain no granular protoplasm ; they a7'ise from the opposite sides of the broad end of the pearshaped cells and, extending in opposite directions, probably serve to unite the ganglion-cells zvith one another.

Comparative Anatomy of the Optic Ganglion.

Although some work has recently been done on the development of the Arthropod eyes, the optic ganglion, both of the convex eyes and ocelli, has received little notice. It has been stated that it arises either as an invagination of the ectoderm near the eyes, or as an outgrowth of the brain. Both Bobretzsky and Reichenbach have failed to produce satisfactory evidence to show that what they regarded as the developing optic ganglion really was such. Reichenbach made the fatal mistake of neglecting to follow his optic invagination up to a point where its real nature would be apparent. Had he done so, he would not have mistaken the retinal ganglion for the layer of rhabdoms and retinulas, and he would have seen that his crystalline-cone layer was the whole ommateum. In my paper upon the eyes of Vespa, I maintained that Reichenbach had misinterpreted the facts. I have recently examined, by means of sections and surface views, the head of Astacus, and find my objections sustained.

It is not my intention to enter here into a long account of the comparative anatomy of the optic ganglion of Arthropods, for the subject is worthy of separate consideration. I merely desire to state the main conclusions to which I have been led by a study of the development of the optic ganglion in such groups of Arthropods as the Isopods, Decapods, Hymenoptera, Lepidoptera, and Coleoptera.

In Vespa, it was shown that on the dorsal edge of the optic thickening a great mass of cells was pushed inward to form the rudiment of the optic ganglion. This mass of cells, which from the very earliest stages is connected with the ventral edge of the optic thickening, after invagination divides into three lobes, which immediately arrange themselves in a line connecting the eye and brain. The middle lobe develops into the optic ganglion of the

No. i.l



adult ; the outer, into a layer of cells which, during the earlier stages, form, around the outer edge of the median lobe, a semicircular band, conspicuous on account of the peculiar character of its cells (Fig. 2, wood-cut). The inner lobe, which likewise disappears during the later stages, forms a characteristically shaped layer of dark cells on the proximal side of the median one. It is over-arched by the cells of the brain in such a way that it appears to form the floor of an oblong cavity. Each lobe, except the outer one, contains a great medullary core.


Figure 2. — Semi-diagrammatic views of the optic ganglion of: (//) Vespa, towards the close of the larval stage, (j9) Astacus, ( C) Cecropia, beginning of pupal stage; (Z?) Acilius, close of larval stage, — showing the position of the retinal ganglion and the three lobes of the imaginal optic ganglion.

E. Eye I. oi Aciliits; c. e. ventral half, c. e?- dorsal half of convex eye; ^^^^ lobes of optic ganglion; ni^~^ medullse of same; r.g. retinal ganglion.

The retinal ganglion arises at a comparatively late period, not by invagination, but by the formation of a layer of ganglioncells among the fibres connecting the optic ganglion with the eye.

In Astacus (Fig. 2, B.), the optic ganglion, after invagination, breaks up into three lobes. The outer one is formed of a folded

152 PATTEN. [Vol. II.

layer of clear cells, Reichenbach's " Atigenfaltey The shape, position, and structure of this fold, and the manner in which it degenerates and disappears, leave little room to doubt that it is homologous with the outer lobe of the optic ganglion of Vespa. The fact that in Astacus this layer is folded as if formed by invagination is an incidental and secondary condition. In Vespa, Cecropia, and Acilius the direction of the infolding is reversed.

The middle lobe is relatively smaller in Astacus than in Vespa; from it arises the epiopticon of Hickson, or the second optic ganglion of Carriere. This lobe contains an oval medulla, from the outer end of which arise decussating fibres that extend into the retinal ganglion. The inner lobe of Astacus, unlike that of Vespa, develops into a distinct ganglion, with a medulla and cortical layer of ganglion-cells like those in the median lobe, but darker colored. It is over-arched by a thick layer of cells arising from the cortical layer of the stalk of the optic ganglion.

The distal end of the stalk is enlarged, and apparently forms a third medulla, the minute structure of which is quite different from that of the other two.

On the outer side of the first lobe is a layer of small, dark ganglion-cells, which Reichenbach regarded as the outer wall of his ^'AugeiifalicJ' and which he maintained gave rise to the layer of retinulas and rhabdoms. But this layer of cells is really the retinal gajiglion, and is formed, as in Vespa, by the development, some time after the ganglionic invagination, of a layer of cells among the fibres connecting the optic ganglion with the eye.

Among Isopods I have examined the optic ganglion of Cymothoa, and find that in development and structure it does not differ from that of Astacus in any particulars that concern us at present.

No extended observations were made on the optic ganglion of the larvae of Cecropia. I can only say that it has a superficial resemblance to that of Acilius, but it is much smaller and less perfectly developed. Toward the beginning of pupal life, the optic ganglion increases enormously in size and assumes a structure very much like that of Vespa (Fig. 2, C). The resemblance between the optic ganglion of Vespa and Cecropia is sufficiently evident to render a detailed description unnecessary. In the later stages both the inner and outer lobes disappear or are reduced to mere insignificant appendages, and the cells at the


base of the optic nerve develop into a large retinal ganglion. The main portion of the optic ganglion arises from the middle lobe.

The similarity in the structure of the optic ganglion of these distantly related genera shows that the optic ganglion of the convex eye of Insects is homologous with that of Crustacea, and that in both groups its structure may be reduced to the same plan. TJic variations in the structure of the optic ganglion of Arthropods may be referred to the modification, development, or suppression of one or more of three lobes, to ivJiich may be subsequently added the retinal gatiglion. What is the meaning of the three lobes so constantly present, at one stage or another, in the optic ganglion of Arthropods } It would be difficult to form a fruitful hypothesis of their significance from a study of forms that do not pass through an active larval existence, for we now know that the optic ganglion of the imago is derived from a larval ganglion, which in some cases is highly complicated, and has itself a long story to tell. In Astacus, Cymothoa, etc., the convex eye and its optic ganglion appear at a period corresponding with the end of the larval stage. Hence we must look for the preliminary stages of the optic ganglion in the larvae and embryos of such forms as Acilius ; and there we may expect to find a solution of the three-lobed structure of the ganglion of the convex eye.

We have already seen how, in the earliest stages of Acilius, the optic plate is composed of three segments, each bearing a pair of eyes, and on the dorsal side of each segment there is a ganglionic invagination.

The three masses of ganglion-cells thus produced unite to form the larval ganglion. Toward the close of larval life, each segment increases rapidly in size ; the third segment grows around the proximal side of the ganglion to form the inner lobe, the second forms the fold around the anterior edge of the ganglion, or the outer lobe, and the first forms the posterior inner mass of the ganglion, or the middle lobe of the future ganglion of the convex eye.

At the beginning of pupal life, the ocelli are drawn towards the brain, and are replaced by the convex eye, which unites with the larval ganglion in a way that I do not yet understand.

From these facts we may infer that the three-lobed optic gatt

154 patten: [

glion of the convex eye of Arthropods is derived from a three-segmcjited, larval ganglio7i, each segment of which belongs to a pair of larval ocelli.

III. Summary and Comparison.

The great nucleus that figures so prominently in the early stages is not the least remarkable feature of the eyes of Acilius. So far as my knowledge goes, it is without parallel and inexplicable.

In eyes II., III., IV., and V., there is a single nucleus in the middle of the retina, between the two rows of gigantic cells, that remains unaltered through life. In eye I. there are at least two of these nuclei, one in the horizontal, and one in the vertical furrow. In eye VI. the nucleus is only present a short time during the earliest stages. In all the eyes the large nucleus seems to be situated in the centre of a group of four sensory pits.

Kleinenberg has described in Astcrope a gigantic, flask-shaped cell that secretes the vitreous substance filling the cavity of the optic vesicle. But there is probably only a superficial resemblance between this cell and that in the eye of Acilius.

The reader has no doubt been impressed by the presence of horizontal rods in the eyes of Acilius, as well as by the fact that tJie rod mosaic varies in a constant i?ianner, according to the inclination of the retinal cells. All the horizontal rods, large or small, are flat, and arranged in long, parallel rows. The semiupright ones form, in cross sections, zigzag lines, and the upright rods are bent so as to form the sides of regular hexagonal figures. Why are the horizontal rods invariably arranged in parallel lines . As the rods are mere frames for the support of nerve fibrillae, it must be, no doubt, because in this way the fibrillae are most advantageously arranged for the reception of light stimuli. In my paper upon the " Eyes of Molluscs and Arthropods," page 652, the statement was made that the perfection of a visual organ was dependent, among other things, upon the degree to which the retinidial fibrillae have become perpendicular to the rays of light. At that time I did not think it probable that there were horizontal rods, as stated by Grenadier, in the eyes of Myriapods. It must be evident now from my description of horizontal rods in Acilius that there is no longer


any reason for me to doubt the accuracy of that part of Grenacher's account.

When I became convinced that most of the rods of Acilius were horizontal, it seemed at first as though the above-mentioned view concerning the direction of the retinidial fibrillae would have to be abandoned, for here were undoubtedly highly specialized eyes in which as many retinidial fibrillae were parallel with the rays of light as at right angles to them, provided they radiated in all directions from an axial nerve. In Pcctcn, e.g., (Fig. 5, A.), we have cylindrical vertical rods containing axial nerves, from which radiate nerve fibrillar in all directions. All the fibrillae are consequently at right angles to the rays of light. But put the rod in a horizontal position, and just as many fibrillae will be vertical as horizontal. In order to bring these horizontal rods into harmony with my theory of nerve endings, it is requisite that the rods should be flattened, and of course so that the long diameter is vertical. Then if the axial fibres, instead of forming a single bundle, as in the cylindrical rods, should form a transverse row (Fig. 5, B.), fibrillse could arise from them, nearly all of which would be at right angles to the rays of light. If the rod were broad and very much flattened, and the upper and lower sides disappeared, then all the fibrillae would be at right angles to the rays of light (Fig. 5, C). In Acilitts all the horizontal rods fulfil the above requirements ; they have exactly the stritctm-e they ought to have, and the only one they could have, to bring their retinidial fibrillce at right angles to the rays of light ; therefore, instead of horizontal rods being fatal to my theory of nerve endings, they afford by their am^angement convincing evidence in favor of it. For those who believe that the nerve fibres described by me are the product of coagulation, or, perhaps of the imagination, may not be so ready to push lightly aside the evidence which the structure of the horizontal rods affords. It certainly is remarkable that in Acilius, whose eyes show as wide differences in structure as we could expect to find between any larval ocelli of Insects, the horizontal rods should invariably assume the only shape that would permit all the retinidial fibrillae to be at right angles to the rays of light.

A vertical section of a pair of broad horizontal rods would give an outline exactly like that formed by a similar section of



[Vol. II.

upright ones, except that in the first case between each pair would be the cut ends of a row of axial nerves, and in the second, a bundle of vertical nerve fibres (Fig. 5, C. and D.).

Some observations have been made upon the eyes of other insect larvae. With one or two exceptions, nothing was found that I care to speak of here. When enough material has been obtained, it is possible that I may be able to give a more extended comparative account of the larval ocelli of insects than I desire, or am able, to do now.



Figure 3. — Vertical (^) and horizontal (^E) section of a Hydrophilus larva about 7 mm. long. eg. corneagen; o. w. outer wall of the optic vesicle; rd. rods.

The eyes in the young larvae of both Hydrophilus and Dytiscus are much alike. In the woodcut (Fig. 3) is represented a section of the eye of a Hydrophilus larva about 8 mm. long. We perceive at once the fundamental differences in the construction of this eye and that shown in Grenacher's well-known figure of the young larval ocellus of Dytiscus. First, there are three distinct layers in my figure produced by the invagination of an optic vesicle. The outer wall of the vesicle is represented by a cluster of small, deeply stained nuclei over the centre of the retina. Th@ latter is composed, not of upright cells, as Grenacher has it, but of horizontal ones, bearing short horizontal rods. The latter form a thin oblong and vertical plate in the middle

No. I.]



of the retina. In this respect also, Grenacher's figure is incorrect, since it represents a broad layer of long upright rods, whereas no such structures are present.

The eye of Hydrophilus, as shown in the woodcut, is much like that of eyes II. and IV. of Acilius, even to the presence of the large retinal cells at the bottom of the furrow.


Figure 4. — Horizontal {A) and vertical (^) section of the larval eye of a neuropterous larva, Chauliodes (?). eg. corneagen; /i, /2, inner and outer lenses; 0. w. outer wall of the optic vesicle; rd. rods.

Figure j. — Diagrammatic views of the retinidial fibrillas in upright and horizontal rods. A. cross section of rods of Pecten ; B. same of small horizontal rods of Acilius ; C. same of large rods, and D. longitudinal section of upright rods in eye V. The arrows indicate the direction of the light.

In Fig. 4 is shown a section through the eye of what I took to be a full-grown larva of Chauliodes, but I am by no means certain that it is such. Expecting to obtain more material, all but the heads were thrown away, and now I have no means of identifying the larvae.

I have represented a section of the eye because it showed an interesting modification of the three-layered ocellus. There seem to have been present both upright and inverted retinal cells, as in eyes V. and VI. of Acilius, but the latter, instead of retaining their inverted rods have converted them into a cuticular, nerveless mass that has assumed the function of an inner

jc8 patten: [Vol.11.

lens something like that in Peripahis. On the inner surface of the lens, we can still see some of the inverted rods that have not quite lost their identity. The outer surface of the lens is surrounded by a membrane, between which and the lens are a few scattered nuclei. The latter are continuous with those of the retina. In vertical sections, the rods appear to be upright, but cross sections shov/ that they are horizontal and arranged in a double row in a nearly circular space surrounded by the pigmented ends of the retinal cells.

The optic vesicle is covered by a corneagen composed of a thick layer of colorless, columnar cells with distinctly stained nuclei. On the periphery of the corneagen is a thin layer of pigment.

Myriapods. — There are many important points in which the larval ocelli of Insects resemble those of Myriapods, and I venture to assert we shall find a still greater resemblance when a better knowledge of the structure and arrangement of the rods in the latter group has been obtained.

The material upon which Grenacher made his observations had been hardened in alcohol only, and as he himself had occasion to repeat, there was much to be desired, especially in the preservation of the rods. In only two instances was a clear picture of the rod mosaic obtained : in Cormocephalus, where they \vere cylindrical tubes ; and in Scutigera, where they were flattened much like those of Acilius. If, after re-examination of properly prepared material, the first case should still stand, it would offset the significance that I attached to the flattened rods in Acilius. But there are not inconsiderable reasons for doubting that it will stand, for in Acilius the flattened rods when hardened in alcohol sometimes appear like cylindrical tubes. In other cases the rods were so poorly outlined that Grenacher was always in doubt as to their structure ; indeed, in some instances he could go no farther than to assert that they were rods.

Besides the similarity in the possession of horizontal rods, one that is all the more striking since these remarkable structures are at present only known in the two groups we are discussing, there are many minor points in which the ocelli of Insects resemble those of Myriapods. The resemblance is striking when


we compare eyes II. and IV. of Acilius with those of luhis or of Heterostoma.

The retinal pigment is in both cases most abundant at the outer ends of the retinophoras, and is confined to the outside of the cell. At least on p. 437 Grenacher states that in the Scolopendriden, if we may regard the condition in Cormocephalus as normal, the pigment appears "in der Mantelflache der Zellen abgelagert zu sein, und die innern Theile derselben freizulassen." My sections show just such figures as he represents in Fig. 6, except that in my judgment the pigment in Acilius and Dytiscus is outside the cell wall instead of inside.

A peculiar feature of Acilius is that in the younger stages all the eyes are elongated in the direction of the retinal furrow, and this condition is more or less conspicuous in the larvae of Dytiscus, Hydrophilus, Colymbetes, Psephenus, and Gyrinnus. There is a similar elongation of the retina in lulus.

One of the most striking points in the less specialized eyes of Acilius is the great notch at the bottom of the eye containing the broad rods of the gigantic cells. Grenadier has shown in Scolopendra a similar notch, and there are strong reasons for believing that it contains a double row of large cells. I think it probable that in Scolopendra, as in the first four eyes of Acilius, all the originally horizontal retinal cells, except the large ones at the bottom, have gradually withdrawn from the median plane of the eye to form a concave layer of nearly upright rods.

When it is urged that Grenacher saw no such gigantic cells, and that there is no indication of them in his drawings, we shall answer that the same thing might be said of his description of the eyes of Acilius, and yet a casual examination of depigmented sections will show at once that such cells are there.

When we compare the retina of Heterostoma (Fig. 4) with that of eyes II. and IV., we cannot withhold the suspicion that in the retina of this Myriapod there is also a double row of large cells, and that their gigantic rods were provided with horizontal external nerve fibres which had been mistaken for the dividing lines between many small superimposed rods.^ If this be so, then Grenacher's " Haarzellen are probably ordinary retinal

1 At one time I also regarded these markings in Acilius as the divisions between small rods.

l6o patten: [Vol. II.

cells with double horizontal rods, and the so-called hairs are the external nerves of these rods. In sections parallel with the broad surface of the rods of Acilius, the external nerves are conspicuous, and in alcoholic material they appear like hairs, or at any rate might give rise to the impression that each cell was provided with a number of slender rods. That these external nerves are probably present in Myriapods is shown by the fact that Grenacher, p. 456, has himself described in the large flat rods of Scrutigera a set of cross striations which he says recall the Pldttchensiructiir so frequently described in the rods of Arthropods, and he even hints that there may be some connection between this fact and the hair-like rods in the eyes of Myriapod-s.

Finally, in his Fig. 14, is shown a semi-transverse section of the retina of Glomeris, where the hair-like rods are united in groups that would correspond very closely to the flattened rods of Acilius, and even in his vertical sections we see indications of the same thing.

The bodies projecting beyond the pigment layer in his Fig. I, look more like a layer of short rods than the ends of cells, as he calls them. There is no other instance to my knowledge where only the outer ends of retinal cells are free from pigment. I regard them as rods, and support this interpretation mainly on the fact that Grenacher represents them as extending into the notch at the bottom of the eye.

But what shall be done with the layer of rods in his first figure } I should regard it as a corneagen, which like that in Acilius had lost its nuclei, if in Fig. 2 both a corneagen and this striated layer were not present. If we do not accept Grenacher's interpretation, no course appears to be open except to consider the outer layer of Fig. 2 as a corneagen, and the middle one, if it is really a separate layer, as the outer wall of an optic vesicle, which forms in this case a vitreous, striated body comparable to that in Chauliodes (Fig. 3) or with that in the cavity of the optic vesicle of Peripatus.

That there is a tendency to form two kinds of rods in Myriapods is clearly shown in Grenacher's drawing of the eye of Heterostoma (Fig. 4). We might expect to find the outer rods short and upright as in the eyes II. and IV. of Acihus, although


Grenacher maintains that they are horizontal like the others, but much longer.

Eye I. of Acilius is the largest and most complicated, and as there are several instances in which the posterior dorsal ocellus of other Insect larvas is conspicuous on account of its greater size, it may be well to mention that a similar condition is to be found in Lithobius, for Grenacher says, p. 441, "Die Einselaugen sind unter sich nicht vollig gleich gross ; auffallig ist freilich nur die uberwiegenden Grosse der jenseite am meisten nach hinten gelegenen."

We have tried to show above : TJiat there is a sinking resemblance between the eyes of Myriapods and those of larval Insects ^ (i) in the pigmentation of the retifial cells ; (2) ifi the shape of the retina ; (3) in the presence of a reti7ial furrow, and (4) and (5) in the probable presence of dimorphic retinal cells and rods, some of which are horizontal and others npright. This similarity is all the more important since we seek in vain for any special points of resemblance between the larval ocelli of Insects and those of Crustacea or Arachnids. Even if it should be shown that the retinal cells in Spiders and Scorpions are upright, there would still be a great difference between the ocelli of Insects and those of these groups, for the presence of upright rods in both instances is in itself too general a feature to be of much importance in determining affinities, and there is no trace whatever in Scorpions, Crustacea, or Spiders of those peculiar features, such as horizontal rods, retinal furrow, and dimorphic retinal cells, that are common to Insect larvae and Myriapods.

Spiders and Scorpions. — Although both Graber and Sograff maintain that the eyes of Myriapods and the ocelli of Insects and Spiders have essentially the same structure, it is probable that both writers based their opinions solely on the fact that in all these cases the eyes were composed of two layers. But granting that such a condition prevails, very little is gained thereby, for it leaves out of account altogether the fundamental differences in the structure of the retinas.

In spite of the recent contributions to our knowledge of the eyes of Arachnids, by Schwemkewitz, Bertkau, Locy, Mark, and Parker, we are still much in the dark concerning the structure and development of these eyes.

l62 patten: [Vol. II.

Locy maintained that the ocelli of Spiders were formed of invaginated vesicles, the outer walls of which became the retinas.

Mark, in discussing the various theories that had been advanced to account for the origin of the retina and its method of formation, says, p. 54, that Locy's observations " seem adequate to settle these conflicting views — so far at least as regards the spider-like type." Great stress is laid upon the inversion of the retina in Spiders, for he says (p. 54), "The inversion of the retina proper is a fact of broader significance than would at first sight appear, and it affords a satisfactory explanation of some of the points in the anatomy and histology of simple eyes which have been so earnestly discussed during the past few years." And a little farther on he intimates that it is doubtful if any Monomeniscus Arthropod eyes contain upright retinas. In fact, the main purpose of his paper is to show that the Arthropod ocelli are constructed on the " Spider-like type," that is, of three layers, the inverted middle one forming the retina. Wherever in the ocelli of Arachnids, Myriapods, and Insects there is evidence of a post-retinal layer and a " leiitigen " with a retina between. Professor Mark would infer that the eye originated as in Spiders, and he seems to have found evidence of one or more of these structures in the eyes of Myriapods and Scorpions and in the larval and frontal ocelli of Insects.

In the present state of our knowledge, the larval and frontal ocelli of Insects cannot be made to fit this theory ; certainly not in Acilius and Vespa, the two cases to which Professor Mark gives especial attention. In Vespa, I have shown since his paper was published that the frontal ocelli were derived from open-mouthed pits, the inner wall of which gave rise to the retina.

The eyes of Myriapods are so much like those of Coleopterous larvae that there is every reason to suppose they develop in the same way.

It would not be safe to conclude from Parker's observations that the retina of the median eyes of Scorpions is inverted, since they do not extend to sufficiently young stages. But, according to his observations, the lateral eyes are single-layered, and consequently the retinal cells are upright. Professor Mark regards the retinal cells in the median eyes as inverted, and since the rods and nerve fibres do not indicate any such inver


sion, he is forced to assume not only that the nerve fibres have shifted their attachment from one end of the cell to the other, but that new rods have been developed at those ends of the cells where, according to all experience, rods are never found. His supposition, which implies that inverted rods are not favorable to perfect vision, which pretends that, in some unexplained manner, upright rods get into an unfavorable position, and then forces us to -assume a fundamental upheaval of primitive relations in order to bring them back to what is practically their original condition, is, in my opinion, at present unnecessary. Even after making this sacrifice of preconceived ideas, the way is no clearer than before, since we have imposed upon ourselves the difficult task of explaining why these supposed inverted retinal cells are exactly like the upright ones of the lateral eye !

I may be permitted to call attention to one or two facts that will perhaps be of service to those who think it necessary to make further observations on the eyes of Spiders. My observations on the eyes of Astacus, Cymothoa, Vespa, and Acilius leave no room to doubt that there are two invaginations connected with the eyes : one to form the optic vesicle ; another, the optic ganglion. The presence of two such invaginations in types so widely separated suggest that a similar condition prevails in Arachnids. If this be so, where is the ganglionic invagination . Is it possible that the ganglionic and optic invaginations have been confused as in Astacus and Crangon, and the conclusions vitiated thereby 1

The second point I wish to mention is more of a theoretical nature, and has to do with the causes of inversion. Professor Mark has suggested that the primitive eye was a laterally flattened cup, with a lens over the slit-like opening, and that either the cup was bent to one side, so that the broad surface of one wall was folded against the lens, or a second lens had been formed, bringing light to the retina from a new direction, and this led to the development of the retina from that wall of the pit next the new lens. In either case the final result would be the formation of a three-layered eye, the middle layer of which developed into the retina. A simpler explanation is suggested by the condition in eye V. of Acilius, where there are both inverted and upright rods ; in such an eye it would be a simple

16^ PATTEN. [Vol. II.

matter to develop the inverted cells at the expense of the upright ones. The degenerated retinal cells could then be transformed into the tapetal matrix, the retinal cells producing tapetal scales instead of rods, as the inverted retinal cells in Chauliodes give rise to the inner lens. If this be true, then the tapetal slit might be compared to the median furrow in the retinas of Acilius.

A probable instance of such a method of inversion is found in Pecten, whose inverted rods I derived from the outer wall of an optic vesicle, by supposing that partially inverted and upright rods existed at the same time ; eye V. of Acilius furnishes us with a needed example of such an eye. Moreover, in Pecten there is evidence that the inner wall was derived from a retina, the retinophorae of which were transformed into the argentea or reflector, and the ganglion-cells into the red pigment layer. Such a transformation of the optic vesicle of Spiders would harmonize with Bertkau's description of the tapetal eyes, or " Nebenaitgen!' The anatomy of the non-tapetal ones, as far as known, would lead one to believe the retina is composed of upright cells, while embryology says they are inverted.^

In the earliest stages of Acihus, the eyes are composed of several sensory pits, each with its cuticular thickening and nerve. It is possible that in the Arachnids the numerous nerve bundles supplying the eyes, such as figured by Grenacher for Lycosa (PI. III., Fig. 22), might owe their existence to the fact that these eyes also were formed by the fusion of several sense organs.

We might change the eyes of Acilius into the Aranean type by suppressing the eyes of the third segment and inverting the retinas in all but the large posterior pair. The latter would then become the " Hmiptangeji " of Spiders, and the smaller ones with inverted retinas the " Nede;m?tgeu."

The failure to form, either by speculation or observation, an adequate notion of the origin of the retina in Arthropods has long stood in the way of a satisfactory explanation of the various structural forms these eyes assume. It was not possible to treat the subject comprehensively and systematically while still in

1 It is difficult to understand Bertkau's statement that his observations confirm those of Grenacher. There certainly is a great difference between his drawings of the tapetal eyes and those of Grenacher.


doubt as to whether the retina arose from the hypodermis or from the brain, or from neither ; or while that organ sometimes called the optic ganglion, in the absence of all evidence, might be regarded as a retina, a retinal ganglion, the brain, an outgrowth of the brain, or an optic nerve, or whatever else might commend itself to the imagination of the investigator.

There were the one-layered, the two-layered, and the threelayered eyes, one with upright, another with inverted, another with horizontal rods. The structure of one eye was as intelligible as that of another ; no more, no less. There was no unit of measure.

It was inferred in " Eyes of Molluscs and Arthropods " that the ground plan in all these confusing variations of structure was a three-layered eye, an invaginated optic vesicle, the inner wall of which became the retina, and an overlying layer of hypodermis the corneagen. Such an eye would be much like that of Peripatus ; flatten the vesicle vertically, reduce the outer wall to a thin membrane, and you have the ocelli of some Insects and Spiders. The observations recorded in this paper confirm the supposition mentioned above. We may now go still farther and say that a lateral flattening would produce the larval Insect, and Myriapod eyes, with horizontal rods ; that the outer wall of the vesicle may, in some cases, develop inverted retinal cells side by side with the upright ones of the inner layer ; and that the rods of these inverted cells may be converted into a lens inside of the optic vesicle, as in Chauliodes and perhaps Peripatus, or they may take the place altogether of the upright ones, as in the tapetal eyes of Spiders.

It can no longer be affirmed that there is a wide difference between the so-called Molluscan and Arthropod eye ; both belongto the same type, as I formerly maintained. This difference it has been urged was due to the presence in the former of an optic cavity filled with an inspissated, refractive substance. But there is just such a cavity in the early stages of eyes V. and VI. of Acilius and one in Chauliodes which is filled with a lens like that in Peripatus and some Molluscs and Worms.

Professor Mark maintains that " none of Locy's predecessors have in the least foreseen the true course of events" concerning the origin and method of formation of the retina of Arthropods.

l66 PATTEN-: [Vol. II.

But it cannot be denied that I foresaw in my paper on the " Eyes of Molluscs and Arthropods " not only the origin of the retina and its method of formation, but, to a great extent, that of the optic ganglion as well. The observations on the development of Vespa and Acilius now furnish a substantial support for the theoretical views advanced in that paper, and this confirmation does not lose any of its value in consideration of the inversion of the retina in the tapetal eyes of Spiders, for that can most readily be explained, as already shown, by 'referring to the condition in eyes V. and VI. of Acilius ; or by the fact that Reichenbach's and Kingsley's observations point to the conclusion that the sensory part of the convex eye is inverted, for in both instances the misconception undoubtedly arises from a confusion of ganglionic and optic invaginations ; and finally the inference lies close at hand that the supposed inversion of the retina in the eyes of Scorpions and in the non-tapetal ones of Spiders is due to a similar confusion of the two invaginations.

Convex Eye. — In my paper on the "Eyes of Vespa," an attempt v/as made to throw some light in the phylogeny of the "convex eye." It seemed to me that a solution of the problem might be obtained by explaining its double nature ; for as shown by the embryology of Vespa and by the permanent condition in such foi^ms as the Libelluliden, Ephemeriden, Gyrinnus, Astacus, Phronima, Schizopods, and others, it is probably composed throughout the Arthropods of a distinct ventral and dorsal part.

I attempted to explain this double condition by supposing that it was a modified larval ocellus like the posterior dorsal one of Acilius and Dytiscus, which I maintained was also composed of a dorsal and ventral part : the latter was the ocellus 'proper; the former, an appendage whose structure and general appearance indicated that it was a secondary and younger part subsequently added to the ocellus ; and it was maintained that if such an ocellus developed into the compound eye, a needed explanation would then be furnished of its double nature. But there was no evidence to show that the posterior ocellus with its dorsal appendage really did develop into the convex eye. I determined to obtain such evidence, if possible, by studying the history of the eye of Acilius during the larval and pupal stages. JBut all my efforts to obtain pupae were unsuccessful. I have


shown, however, that the appendage is not, strictly speaking, an outgrowth of the posterior dorsal ocellus, as I formerly supposed, and I do not now see any reason to suppose that the ocellus proper develops into any part of the compound eye. The appendage is, I now believe, one or two of the primitive sense organs of which the ocellus is composed, that have not completely united with the others, and have undergone a special modification in the direction we have already explained.

Toward the close of larval life, the convex eye appears as a thickening of the ectoderm immediately around the appendage. At this time it is difficult to distinguish any line of demarkation between the appendage and the thickening. In the latest stages I possess, it forms an enormous, thickened band that almost encircles the six ocelli. The band is narrow in the middle, but expanded and rounded at either end. Its ventral edge is deeply invaginated, especially near the appendage where the invagination first appears, and is connected with the adjacent ectoderm by a thin vertical layer of cells (Fig. i, wood-cut). At first it seemed probable that the appendage developed into the ventral half of the convex eye, and the thickened band into the dorsal half. If the band itself should be divided into two parts, this interpretation would be, so far as I can see, untenable, and I must admit that there is an indication of such a division, though so indistinct and unaccompanied by any difference in the development of the ommatidia on either side of it, that I am still in doubt as to its meaning.

It is difficult to believe that the appendage of the posterior ocellus has nothing to do with the convex eye, since the latter is, in the early stages, so intimately connected with the former. One thing is certain, that a great part of the compound eye arises suddenly at the close of larval life as a thickening in a previously indifferent layer of hypodermis; hence that part at least cannot be considered a modification of any functional larval organ.

The development of the frontal ocelli points to the conclusion that they are widely different from the larval ones, and perhaps closely related to the compound eyes.

One of the things that impressed itself most deeply upon me, after studying the embryology of Acilius, was the threefold structure of the head as shown in the three segments of the

l58 PATTEN. [Vol. II.

optic plate, optic ganglion, and brain. The same feature is also shown in the arrangement of the eyes in the imagines of those insects supplied with frontal ocelli, for in such cases it is evident, since I have shown in Vespa that the median unpaired ocellus is double, that the imagines have three pairs of eyes, and it at once suggests itself that there is some intimate connection between this fact and the presence of the three larval segments. It is possible that the solution of the problem, to which I shall return in my next paper on the "Development of Acilius," may lie in an explanation of the pupal stage.

Neuroepithel Cells. — My studies on the " Eyes of Molluscs and Arthropods " led me to believe that ganglion-cells were modifications of sensory ones. This belief was based upon the presence of intermediate stages between sensor}^ and ganglionic cells, upon the constant occurrence of intercellular nerve ends, and upon the embryological evidence afforded by the fact that in Pecten the ganglionic cells of the eye and sensory papillae were derived from the cells of the hypodermic thickening that gave rise to these sense organs. This supposition further led to the conclusion that the optic ganglion of Arthropods could not be an outgrowth from the brain toward the eye, but one from the eye toward the brain. This conclusion is now in a measure confirmed by the history of the development of the optic ganglion of Acilius.

When we review the semi-ganglionic cells described by me in Haliotis, " Eyes of Molluscs and Arthropods," PI. 30, Fig. 68, and in Acihus, PI. X., Fig. 58, h, and the myoepithelial cells of Coelenterates as described by Hertwig, we perceive that nearly all these cells have three prolongations, one of which is directed outwards, and terminates between, the cells of the ectoderm; the other two extend, from the opposite pole of the cell, inwards, and probably unite with similar prolongations from other ganglionic cells. There is no sharp line of demarkation between the tripolar cells described by me in the retina of Haliotis and the neuroepithelial cells of Coelenterates as figured by the Hertwigs. In the case of Haliotis and Pecten, the origin of some, at least, of the ganglion-cells is beyond question, for they still form a part of the ectodermic thickening that gave rise to the sensory part of the eye. In Acilius we have temporarily represented the condition that prevails in the sense organs


of Coelenterates and Molluscs ; for, while many ganglion-cells arise from the invagination on the side of the eye, others wander inward from the optic thickening itself. The details of the latter process are best studied in the great cells that are the last to form. While these cells are still in the optic thickening they are distinctly tripolar, one prolongation being directed outward to form the rudiment of a nerve end. Such cells would correspond to the tripolar neuroepithelial cells of Coelenterates, or those tripolar ones already mentioned in Haliotis and Acilius.

In Acilius there is no reason to suppose that a ganglion-cell once established in the above manner ever loses its direct connection with the ectoderm. According to Hertwig, those connected by an outward prolongation with the ectoderm are intermediate between fully developed ganglion-cells and sensory ones, and we are to infer that they are intermediate because of this outward prolongation. I fail to see the necessity for this conclusion, for there is reason to suppose that most, if not all, ganglion-cells, have some connection with the outer world, consequently the existence of such a connection would not indicate the age of the ganglion-cells.

The history of the giant ganglion-cells of Acilius, together with a consideration of the neuroepithelial cells of other groups, it seems to me, warrant the conclusion that the primitive ganglion-cells zverc tripolar, and zvere derived from tripolar neuroepithelial cells. The outer extremities of these neuroepithelial cells were reduced to intercellular nerve ends, tJie bases of which, in Acilius, became the protoplasmic prolongations of the ganglioncells, and are probably homologous with the axis-cylinders of Vertebrates.


The more important results of the foregoing study are as follows : —

(i) The laival optic ganglion is composed of three segments, each of which is united on the one hand with a segment of the brain, and on the other, with a segment of the optic plate.

(2) Each segment of the optic plate bears a pair of eyes.

(3) The ocelli are composed of four or more sensory spots, or pits, each pit being supplied with a separate cuticular thickening and nerve.

170 PATTEN. [Vol. II.

In the centre of each group of four sensory pits, is a single large nucleus of doubtful significance.

(4) The pits of each eye finally unite to form a thickened patch of ectoderm, with a median double row of gigantic cells and a common cuticular thickening.

(5) The thickened ectoderm is invaginated to form an optic vesicle, the inner walls of which form the retina, while the surrounding indifferent ectoderm forms a third layer of cells over each vesicle, thus producing a typical three-layered eye.

(6) In the embryonic stages of eyes I.-IV., the retinas of which are invaginated without the formation of a cavity in the optic vesicle (unless the space between the median row of gigantic cells can be called one), all the rods are horizontal.

(7) In the full-gro\vn larvae, the smaller outermost rods become upright ; the larger and deeper ones remain in a horizontal position.

(8) In eye V., there is at first a strong tendency to form horizontal rods. But the laterally flattened optic vesicle expands, forming a spacious cavity in the vesicle, and all the rods become upright except those of the median row of gigantic cells.

(9) In eye VI., which has no median row of giant cells, no horizontal rods are formed.

(10) The outer wall of the optic vesicle in eyes I.-IV. seem.s to be absent. In the embryos, its presence is indicated only by a few characteristic nuclei between the retina and the corneagen.

(11) In eye V., the outer wall of the optic vesicle is represented by two great masses of inverted, rod-bearing cells, probably derived from the tvv'o sensory spots, 5 and 6, seen in surface views of the eye before invagination.

(12) In eye VI., the outer wall is composed of a thin nucleated membrane, and a cluster of inverted retinal cells, derived from sense organ number 6.

(13) Eye I. is composed of at least nine sensory spots, four of which with their central nucleus and median row of giant cells give rise to the horizontal retina ; four more, exactly like the first, to the vertical retina ; and the ninth, to the appendage. All these sense spots unite to form a single homogeneous organ ; but, during the later stages, the three groups of sensory spots become greatly modified, so that in the adult eye the parts they give rise to, the vertical, and horizontal retinas, and the appendage, are widely different in structure.

(14) All the retinas are composed of retinophorae, formed by the union of two cells. They contain two nuclei and two rods, and are supplied with axial, and external nerve fibres.


(15) In very rare cases one finds ganglionic cells in the retina of Acilius.

(16) The rods are arranged in pairs, which form a mosaic of hexagonal figures when upright, and straight vertical lines when horizontal.

(17) In the horizontal, as well as in the vertical rods, the retinidial fibrillse are at right angles to the rays of light.

(18) All the larval ocelli of AciUus and Dytiscus contain more or less distinct dimorphic, retinal cells. The giant cells always form a double row along the bottom of the furrow. Their free ends are bent at right angles, and bear short but broad horizontal rods.

(19) The ends of the smaller retinal cells, and consequently their rods, may be horizontal, upright, or inverted.

(20) Between the two rows of giant rods are two sheets of coarse, vertical nerve fibres and a layer of medulla-like substance.

(21) The pigment granules are deposited on the surface of the retinophorse and around the external nerve fibres.

(22) All the eyes are developed from the optic plate, the thickened distal edge of the cephalic lobes. On the proximal edge of this optic plate is a semi-circular furrow, which gives rise to the optic ganglion. The furrow is deepened to form two distinct pockets, that give rise to the first and second segments of the optic ganglion ; the third segment is formed by an inward proliferation on the proximal side of the third segment of the optic plate.

(23) The innermost walls of the ganglionic segments are from the earliest stages connected with the inner face of the optic plate.

(24) Numerous ganglionic cells arise from the optic thickening, and wander along the optic nerves into the optic ganglion.

(25) Toward the close of this process, about the time when the invagination of the sensory areas begins, enormous, tripolar cells are formed in each eye, which pass along the optic nerve, from the eye to the optic gangUon, dividing rapidly on the way, and producing small, tripolar ganglion-cells. But one of the proliferating cells retains its great size throughout life, and finally takes up its position on one side of the medulla belonging to the eye from which it arose.

(26) The history of these cells affords excellent evidence in proof of the theory which explains the presence of intercellular nerve fibres, by supposing them to be the outer ends of sensory cells, now converted into ganglionic ones.

(27) The optic ganglion of the convex eye of Arthropods is composed of three lobes : the first always, and the third sometimes, disappears ; the second gives rise to the optic ganglion proper. The retinal ganglion is a secondary product, and is not formed by invagination.

172 PATTEN. [Vol. II.

(28) The three-lobed optic ganglion of the convex eye of Arthropods is derived from a three-segmented larval ganghon, each segment of the latter belonging to a pair of larval oceUi.

(29) The first, second, and third segments of the optic ganglion of Acilius larvae are respectively homologous with the second, first, and third lobes of the optic ganglion of the compound eye.

(30) Hence, from the first segment of the larval ganglion, or that segment which is united with the large, posterior, dorsal ocellus, is developed the optic ganglion proper of the compound eye.

(31) The optic ganglion contains six meduUse, each of which corresponds in structure to that of the retina to which it belongs, and this indicates that the arrangement of the medullary fibrilte is as near like that of the retinidial fibrillas of the retina, as the existing condition will allow.

(32) The structure of the retina in the larval ocelli of Insects is much like that of Myriapods, and the whole eye is constructed on the same plan as that of Peripatus and most Molluscs.

Milwaukee, June i, 1888.



Introduction 97

Topographical relation of optic plate and ganglion to cephalic lobes and brain, 97


Position; explanation of surface views; sense organs; optic cup; optic vesicles; cuticular thickenings; median ridge; median furro\v; comparison of sensory pits with eyes on mantle of Molluscs ; corneal and retinidial cuticula; the eye composed of a collection of sensory spots; large nucleus; closure of optic cup, and formation of the three layers; corneagen; iris; pigment; cuticular pellicle; lens; the retina; second median ridge; double row of cells with horizontal ends; retinophoroe; pigment; retinal rods and nerve endings; rods in Cephalopods; in Area, etc.; nerve bundles .....,,,., 99-112

Position; surface views; sensory pits; closure of optic cup; sixth sensory

spot forms tongue of inverted cells; corneagen retina; rods . . II2-II5


Position; surface views .......... II5-I16


Position; surface views; contain two large nuclei and nine sensory pits; invagination of eyes I. and III.; corneagen; retina; retinophorae ; nerve ends; rods; dorsal appendage ......... 1 16-129


Position; invagination; retina; shape; lens; asymmetry . . . 129-131


Origin OF Optic Ganglion. — Topography; three segments; invagination of first two segments and inward proliferation of the third; connection of inner surface of optic plate with the segments of- the optic ganglion; proliferation of ganglion cells; theory of invagination of optic gangli<in; inclosing of ganglionic segments by growth of optic plate ; optic ganglion before rupture of embryonic membranes 131-137

^Iedull^ of Optic Ganglion. — Position; shape; structure; structure is similar to that of retina to which it belongs ; signification of this similarity;

J 74 PATTEN. [Vol. II.

branches of medullary stalk; crown of ganglion cells; primary nerve fibres; optic ganglion of late larval stage; origin of the three lobes of imaginal ganglion from three segments of larval one . . . 1 37-144

Optic Nerves. — Position; course 144-145

Neurilemma. — Origin from basement membrane; structure . . . 145-146 Origin of Ganglion Cells. — Modified tripolar sensory cells; position and structure; passage from eye to optic ganglion; structure of outer prolongation; gigantic ganglion cells of larvae; position near medullte; connecting fibres; small ganglion cells, probably tripolar; Vespa . . . 146-150 Comparative Anatomy of Optic Ganglion. — Optic ganglion of Vespa; ofAstacus; Cymothoa; Cecropia; homology of optic ganglion of insects with that of Crustacea; the three lobes of optic ganglion of convex eye of Arthropods homologous with the three-segmented ganglion of the larvae, 150-154


Large nucleus; horizontal rods; mosaic of upright and horizontal rods; structure of horizontal rods supports the theory that retinidial fibrillse are at right angles to the rays of light; eyes of Hydrophilus; comparison with Grenacher's figure of larval eye of Dytiscus; eye of Chauliodes . 154-158

Comparison of Myriapod eyes with those of Acilius 158-161

Eyes of Scorpions and Spiders; absence of ganglionic invagination, in Arachnids; possible explanation of inversion of retina in anterior eyes of

Spiders 161-166

Convex Eye; origin; frontal oceUi of imagines 166-16S

Neuroepithel Cells 168-169

Summary 169-173

Contents i73-i74

Explanation of Plates 178-190

No. I.]




ab. n. a. I.

am. a?- n.

ap. ap. n. ax. n.



c. a.

ce. <5r= ce


eg. a. cl.


C. 0. t. C. 0. V.


d. «.!-"

d. n.


en. c.

ex. n. f-Sg>-^

g. c. r.

g. r. d.

g. v>-^.

h.g. c.

h. 7?i.




i. n.


first pair of antennce; labrum.

second pair of antennae.

nuclear-like bodies.

antennary lobe.


nerve to second antenna.

appendage to eye I.

nerve to appendage of eye I.

axial nerve.


bridge over ganglionic invagination.

basement membrane = outer neurilemma.

clear area. ,1 convex eye.


corneagen; median part with faint nuclei.

corneagen of appendage.

cuticula that gives rise to lens.

cavity of ganglionic invagination.

corneal cuticula.

cavity of the optic vesicle.

cuticula thickening = retinidial cuticula.

dark areas.

dividing nuclei.



cavity in endoderm.

external nerve fibres.

frontal ganglion.

great ganglion cells of optic ganglion.

double ganglion cell of optic ganglion.

ganglion cell of retina.

great rods.

ganglionic invagination.

heel of great retinal cells.

medulla of the horizontal retina.

horizontal retina of eye I.

horizontal rods.


inner neurilemma.

infra-oesophageal ganglion.


/. a. n.

I. n.


VIC. 7)1. c.

in. b.


VI d. n.

VI. h. m. i. c.

VI. f. VI. g. c. vi.g.r.

VI. o.

m. r.









n.g.c) nr. nr?n. s.


a', c.


0. n.

0. op. c.

0. p.

0. w.

0. %u. g.

pg. n.

p. n.

inverted rods of outer wall of

optic vesicle, lens.

loops of axial nerve, lateral nerve of horizontal retina, mouth, medulla of the appendage to

eye I. muscles, median chord, medulla of the brain, mandibles, mandibular nerve, medulla to horizontal retina, medulla to inverted cells. median furrow, median row of giant cells, median row of giant rods, oesophageal muscles, median ridge, mesoderm.

medulla of vertical retina, maxillae.

nerves to the eyes, nerves to the sense organs in

eyes, large nuclei, nerve fibres, nerve fibrillse.

newly formed ganglion cells, outer ends of ganglion cells that

have passed inwards, primary nuclei of retinophorK. secondary nuclei of retinophorre. nerve spindles, oesophagus.

oesophageal commissures, optic ganglion, outer neurilemma, opening of optic cavity, optic plate.

outer wall of optic vesicle, outer wall of optic ganglion, pigmented nerve fibres. place where optic ganglion is

continuous with the optic




[Vol. II.

/. n.f. primitive nerve fibres.

rd. rods. rdJ «5r= rdJ' rods of primary and secondary cells of the retinophorse. rn. retinidium.

s. segments of nerve chord. s. 0. segments of optic plate. s. o.g. stalk of optic ganglion. s. r. c. small retinal cells. J. rd. small rods. ty^ tentorium. tg?-"^ tergas.

/. I. c.

in. V. n. V. r.


y y.c.


X. y.

tongue of inverted cells.


vertical nerve fibres.

vertical retina.

dark nuclei in young eyes.


yolk cells.

first cephalic invagination.

second cephalic invagination.

sense organs of the eyes.

upper edge of eye III.

178 PATTEN. [Vol. II.


Fig. I. Surface view of the head of a young Acilius embryo. X no.

Fig. 2. The same. X no.

Fig. 3. The same. X no.

Fig. 3a. Side view of the same. X no.

Fig. 4. Side view of the head. X no.

Fig. 5. Ventral view of head. X no.

Fig. 5a. Side view of same head. X no.

Fig. 53. Dorsal view of dorsal edge of optic plate of same head. X 1 10.

Fig. 6. Ventral view of head. X no. Fig. da. Side view of optic plate of the embryo in Fig. 6, showing arrangement of

surface nuclei around the sense organ of each eye. X 190. Fig. 7. Ventral view of head just before rupture of embryonic membranes.

Fig. 'ja. Side view of same. X no.

Joiim.Morph Vol.1/.








't^""-- ■



r ■


1). 7




  • l^

^ ?


Sith-J^^t Ki^erxa-tWatter.Fraiif^^t^ff.

l8o PATTEN. [Vol. II.


Figs. 8 and 8^. Ventral and side view of head just after rupture of embryonic membranes. X 86.

Fig. 9. Side view of head of an embryo almost ready to hatch. Semi-transparent.


Figs. 10 akd 10^7. Dorsal and side view of the head of a full-grown larva.

Figs. 11-16. Surface views of the dorsal, anterior edge of the optic plate, showing successive stages in the formation of the first eye. In order to economize time and labor, the surface nuclei are shown only in Figs. 14 and 1 5. X 260.

Fig. 17. Semi-diagrammatic and transparent side view of eye I., in the stage shown in surface view in Fig. 16.

Fig. 18. The same, in a later stage, showing the bend in the middle of the sensitive area, to form the vertical and horizontal retinas, and the sense organs and nerves to each part.

Jnuni.Moriih Vol II

l^\ ^

Fi m.

md Wf ' ^


1.. ,>^ "


>^ rfa:

l82 PATTEN. [Vol. II.


Fig. 19. Cross section of the cephalic lobes of an embryo somewhat younger than that in Fig. I, PI. VII. X 360.

Figs. 20-23. A series of cross sections through the cephalic lobes of an embryo like that in Fig. I ; showing the three ganglionic invaginations. X 360.

Figs. 24-27. Cross sections of an embryonic head about same age, perhaps younger, as that in Fig. I. This head was abnormally shaped. It was deeply invaginated along the median line; on the sides of the invagination were the antennae, and at the bottom, the mouth. X 360.

Figs. 28-34. A series of cross sections through the head represented in Fig. 5. They show the structure of the brain, optic ganglion, and optic plate. X 360,

Figs. 35, 36. Cross sections of a head like that in Fig. 6, just after the optic ganglion is completely shut off from the exterior. X 190.

Figs. 37-39. Cross sections of a younger stage than that in the preceding figures, showing the last stages in the enclosure of the optic ganglion, etc. X 138.

Fig. 40. Semi-transparent surface view of the optic ganglion and one side of the brain; showing the three lobes of the optic ganglion, the medullie, and the nerves.

Fig. 41. Semi-diagrammatic horizontal section through the medullary portion of the optic ganglion; showing the position and shape of the medullte and the nerves that unite them with their respective eyes.

. Jo urn. . Vorjih I al II

»'... ^^<fSffp. ,,j-T— irK^l^

Wr^ yav: "«■■


'.•Ayfcjf rWtmrtWMB-^^^^vrt^H.

1 84 PATTEN. [Vol. II.


Figs. 42-47. A series of sections through the head of an embryo just after the rupture of the embryonic membranes. See Fig. 8. They show the structure of the brain and its envelop, optic ganglion and eyes, as well as the tracheal invaginations of the head. For explanation of last-named structures, see my next paper on the "Development of Acilius." X 138.

Fig. 48. Horizontal section of the optic ganglion, from an embryo just ready to hatch. The break in the outer neurilemma is on the side next the eyes. It is important to notice the remarkable similarity between the structure of the meduUse and the retinas to which they belong.

Figs. 49, 50. Two vertical sections through the optic ganglion of a full-grown larva; showing the medullar and nerves, and the three lobes of the ganglion. X 1 70.

Figs. 51-53. Three horizontal sections of the optic ganglion of a full-grown larva. Shows same as preceding sections. X 170.

Fig. 5417. Cross section of the rods in the fifth Eye; showing axial nerves, retinidia, etc.

Fig. 54/5. Same from the sixth eye; showing loops of axial nerves, etc.

Fig. 55. A vertical section through the furrow of eye I.; showing the ends of the giant, and small rods. Compare Fig. 56.

Fig. 56. Vertical section of the horizontal retina of the first eye; showing the double row of giant cells with their external nerve fibres, the layer of vertical fibres, with their fibrillce, and the ends of the small horizontal cells, etc.

Fig. 57rt. Semi-diagrammatic view of a retinal cell from the fifth eye; showing the double retinophorjE, and the position and shape of the rods belonging to each cell of the retinophora, etc.

Fig. 57^. Same, in side view.

Fig. 58. Retinophorse isolated by maceration.

a. A double cell, decolorized, from the fifth eye.

b. Isolated cell from the sixth eye. The primary and secondary nuclei are

situated close together.

c. Two double cells, depigmented, from the sixth eye; showing the nerve

fibres and the deceptive shape of the inner ends of the retinal cells.

d. Same.

e.f. Two views of a small retinal cell with horizontal rods, probably from the second or fourth eye. g. Decolorized retinophora that showed especially well the shape of its component cells. h. A fragment of a retinal cell with a tripolar ganglionic cell attached to it. i. Two giant cells from the first eye; showing how the light and dark cells overlap. Fig. 59. Horizontal section of the first eye; showing one of the rows of giant cells and the two layers of vertical fibres with the medulla-like substance between. Compare Fig. 56.

Fig. 60. Horizontal section of the row of small retinal cells just above the giant cells in the horizontal retina of the first eye. It shows how the rods of each retinophora unite with each other above the centre of the cells, and not with the rods of neighboring cells, as in all other cases. See Fig. 57a.

)!irn.Mi>rjiliA',)l J/


y%,^^^ ^'^'■


■' ^,<**'^'"*J7"^

it '.X'..


> , *t ' ^ 'j-»


--^ ag.'

r86 patten: [Vol. 1 1.


Development and Structure of. the Fifth Eye,

Fig. 6l. Cross section of the head in Fig. 3; showing the large median nuclei and two of the sensory pits of which the eye is composed, etc. X 430.

Fig. 62. Same, in a little older stage ; showing formation of a large ganglion cell.

X 560.

Fig. 63. Section of the eye in the stage shown in Fig. 6a. It shows a large cluster of newly formed tripolar ganglion cells, wandering from the retina to the optic ganglion. X 560.

Fig. 64. Section of the eye before the rupture of the embryonic membranes. The pits and their cuticular thickenings have united to form a single sensory layer, and the large nucleus has sunk into the retina. The formation of ganglionic cells from the retina has ceased. The last one formed is now seen in the optic ganglion on one side of the optic nerve. X 560.

Fig. 65. Section of the eye just after the rupture of the embryonic membranes.

Fig. 66. Section of the eye just after the closure of the optic cup. The median row of giant cells, although small compared with those in the other eyes, is very conspicuous on account of the manner in which it protrudes .ibove the level of the surrounding cells. X 560.

Fig. 67. Section of the eye in a later stage, showing the first steps in the differentiation of the corneagen, and the outer wall of the optic vesicle.

Fig. 68. Section from an embryo about ready to hatch. The outer wall of the optic vesicle is formed by two great masses of inverted rod-bearir.g cells, derived from the sense organs 5 and 6. See Fig. 6a. We can now see the minute rods of which the cuticular thickening is composed. Most of the reddish brown pigment has been dissolved by the alcohol. X 560.

Fig. 69. Section of the eye one or two days after hatching ; showing the distribution of pigment, and the well-developed inverted rods of the outer wall of the optic vesicle. X 520.

Fig. 70. The eye as shown in a longitudinal horizontal section of a fully developed larva. X 220.

Journ.Moqih. Vol.11


aim ^>^^^


■ eg \\\

"ii<-u- ■■;■ \i

iiujc 9---








nth Ak! r.Wtmrsi'JiSr FnuOcfurt'.f

1 88 PATTEN. [Vol. II.

EXPLANATION OF PLATE XII. Structure and Development of Eyes II., IV., and VI.

Fig. 71. Eye VI. as seen in a cross section of the head just after the closure of the optic cup. X 430 Fig. 72. Same in a little later stage. X 430.

Fig. 73. Eye VI. of an embryo about ready to hatch. Compare Fig. 9. The section is semi-transverse, cutting through the inverted sense organ t = t. i. c, and the longitudinal vertical axis of eye III. X 430.

Fig. 74. A similar section through the same eye of a one-day-old larva. X 430.

Fig. 75. Eye VI. of a full-grown larva, as seen in semi-transverse section. Compare Fig. \oa. The section passes through the optic nerve of this eye, but does not show the tongue of inverted cells. It passes just on the dorsal side of it. X 350.

Fig. 76. Eyes II. and IV. = A and B, as seen in a longitudinal horizontal section of an embryo, some time after the rupture of the embryonic membranes. Compare Fig. 9- X 430 Figs. 77 and 78. Eyes II. and IV. of a fully developed larva, as seen in a semitransverse section, nearly parallel with the vertical axia uf the hrst and third eyes.

X 430

.Jouni-MoriiiL Ibl.l!.


^&.Artit vVenurik^nlecBmk^^M

igo PATTEN. [Vol. II.


Fig. 79. Eye I. as seen in a semi-transverse section of an embryo just after the rupture of the embryonic membranes. Compare Figs. 8 and 16. X 520.

Fig. 80. The same, after the sensory area is differentiated into the vertical and horizontal retinas.

Fig. 81. The same, from an embryo some time before hatching.

Fig. 82. Transverse section of the sensory area of Eye I. in the stage shown in Fig. 14. X 560.

Fig. 83. Same, of Eye III. in a later stage. The median furrow has disappeared and the invagination of the sensory area has commenced. X 560.

Fig. 84. Cross section of the vertical retina of Eye I. in the full-grown larva.

X430Fig. 85. Eye I. as seen in a tangential section of the side of the head. The section is at right angles, and from an eye in the same stage, to that shown in Fig. 80.

X4I5 Fig. 86. Section, in same direction as the preceding, of Eye I. just before hatching. X 260.

Fig. 87. Transverse section of Eye I. in same stage as shown in Fig. 79. The section passes through the large nucleus of the horizontal retina.

Fig. 88. Section similar to the preceding one, through the vertical retina of the same eye. X 430.

Fig. 89. Horizontal section through the horizontal retina of Eye I. during a little earlier stage than that shown in Fig. 85. X 430.

Fig. 90. Eye III. as seen in a tangential section of the head of a full-grown larva; partly depigmented. X 240.

Fig, 91, Eye I. in same section as the preceding. X 240.

Jouiii.ViiijiJi. Vol.ll.

■an c<i




iia //iHKMm-<««K/J»rf*nW

Volutnc II. November, 1888. Number 2.




A thesis for the degree of Ph.D., accepted by the Board of University Studies of the Johns Hopkins

University, May, 1888.


Fellow of the Johns Hopkins University.

During the spring of 1887 the Marine Laboratory of the Johns Hopkins University was stationed on the island of New Providence, Bahamas. As soon as possible after my arrival, March 10, I endeavored to find out what corals were breeding. Of half a dozen species examined only one was with eggs. This was the so-called Chenille stone, or Manicma areolata, one of the commonest corals in the Bahama waters. The breeding continued to be very active until the middle of April, when it began to decline, coming to a close before the first of May.

The coral is found in large numbers in water of easy wading depth, lying with its pedicel buried in the sand. It lives very well in small aquaria, if care be taken to change the water twice a day. In rearing the larvae I was compelled to employ the same tedious method as Lucaze Duthiers : every morning and evening I transferred them, one or two at a time, by means of a pipette, to a dish of fresh water. When once the larvae had become attached, the matter was simplified by transferring the entire dish to a bucket of fresh water. For preserving the larvae, Perenyi's fluid proved the most generally useful reagent. Osmic acid specimens were also of very considerable value in elucidating special points. For the anatomy of the adult, abso



lute alcohol gave the most satisfactory results. I take this opportunity of thanking Professor Brooks, the director of the Marine Laboratory, for his advice and constant suggestions, which have been of the greatest value to me during the course of this work, both at the seaside and in the Baltimore laboratory. I am also indebted to Mr. C. S. Hodge, of this laboratory, who very kindly took upon himself a part of the task of cutting the sections.

I. Systematic Description of Manicina.

Milne Edwards makes Manicina the sixty-fourth genus in the family Astraeidze. Like Meandrina, this coral grows by incomplete fission, the calicles remaining connected so as to form meandering valleys, in which the limits of a calicle are not distinguishable. The shape of the colony ordinarily met with is subturbinate (see figures on Plate I), there being a well-marked pedicel. With increasing age the pedicel becomes less and less evident, until the corallum finally assumes the shape of a convex mass with a flat basal surface. The basal surface has two diameters, the longer about three inches, the shorter two inches, and the vertical height of the corallum is about equal to the shorter diameter. The distinguishing features of the genus, besides its growth from a pedicel, are as follows : The columella is spongy and very well developed. The septa are thin, closely set, and have strongly marked granulations on their sides. The edges of the septa, both within and over the exterior of the calicles (costae), are finely and regularly toothed. " The genus was instituted by Ehrenberg for certain of Lamarck's Meandrinae, characterized by growing from a pedicel or central point of attachment " (Dana). In the species Manicina areolata, the septa can be divided into three cycles, and each septum has in the neighborhood of the columella a large and rounded lobe.

The animal is of a brownish color, and when expanded, extends high above the skeleton. In this condition the tentacles are moderately long, and are closely set round the periphery of the oral surface. In the pedunculate forms, both in the expanded and contracted condition of the animal, the upper part of the lateral surface of the skeleton is covered by the polyp. The pedicel,


which includes the skeleton below this region, is usually disfigured by small annelid tubes and other incrustations.

The young, after the swarming life is over, affix themselves to some solid substratum, such as a piece of rock or a shell. To this they remain attached until they have reached a diameter of about half an inch. Until the diameter is one-third of an inch, the animal remains a single polyp, circular or oval in transverse section, and with a flat or irregular surface of attachment. The multiplication by fission then begins, and with it the formation of the pedicel. When the latter has become apparent, the coral is broken off from the rock to which it was attached, and henceforth lives free in the sand.

II. General Sketch of the Development from Surface


Lacaze Duthier's figures of Astroi'des (i) apply so well to Manicina, that I have thought it unnecessary to give a series of such views. The chief point of difference is that the larvae are red in the former genus, but colorless in the latter.

Manicina is hermaphrodite. Normally, it would appear, the mother gives birth to larvae, which pass out by the mouth. But the first two batches of corals I kept poured out an abundance of eggs and semen. Each batch numbered eight corals, was distributed into four aquaria, and was kept for only a day. The first set laid on the night of the 15th, the second, on the 17th of March. I was not able to try a third batch until the 20th, when I was surprised to find that all the corals had given birth to larvae a little more advanced than the planula stage. The throwing out of eggs and spermatozoa was probably abnormal, as the eggs which I watched underwent very irregular changes, and finally broke up. In connection with this ejection of eggs and semen, it may be interesting to note that though after March 20 the corals always ejected larvae when kept in aquaria, the stage of development in which the larvae were born became much more advanced as the season grew older. The larvae born March 20 and 21 were without cilia, and at the stage represented in Fig. 4, PI. II. Those born April 5 were ciliated and as far advanced as Fig. 20, PI. IV.

The larvae I obtained during March lay, for a day or two after

194 WILSON. [Vol. II.

birth, motionless on the surface of the water. They then acquired ciUa, and began to swim feebly about at the surface, often collecting in rows at the sides of the vessel or in groups away from the sides. Lucaze Duthiers noticed the same habit in Astro'ides. At the end of a week the larvae are rapid swimmers, now spending most of their time below the surface of the water. Often they creep along the bottom and sides of the dish like little worms. Though they are constantly lengthening or shortening their bodies, the general shape is that of a pear, with the mouth at the small end. The broad end is usually, though not invariably, directed forwards in swimming. While creeping over the glass they frequently fasten themselves by their broad ends for from one to ten minutes, becoming free again of their own accord. The larvae continued to swim about without any change in form until the middle of April. By this time, as subsequent examination showed, they had acquired six mesenteries (three pairs) and two long mesentrial filaments (Fig. 19, PI. III.). A large number now began to attach themselves. Supported on the broad end, they stood nearly erect, spinning slowly around. The long axis began to shorten, and by the time they had assumed the shape of an oblate spheroid, eight meridional constrictions had appeared, marking the number of mesenteries. The long or, rather, oral-aboral axis continued to shorten until the little coral had assumed a disk-like shape with a flat basal surface and a convex upper surface, the latter marked with the meridional constrictions (a transverse section of this stage is shown in Fig. 34, PL V., and a longitudinal in Fig. 38). The whole process of fixation occupied several hours. All the larvae did not become attached at the same time. On the contrary, many continued to lead a free life until the first part of May.

The lot of larvae I got on the fifth of April were born in a more advanced state than their predecessors, and developed after birth more rapidly. So much so, that by the middle of April as many of this batch became attached as of the larvae born March 20. Indeed, the first fixed larvae I obtained, April 9, were from this lot, and had taken only four days to go through the metamorphosis, which in all the other larvae required from two to four weeks. This great amount of difference amongst the larvae, as to the time required to reach a certain stage of development, made it a matter of some trouble to pick out a


consecutive series of stages. When I began to section, I found that inferences based on the length of the larva's life were onlytrue within wide limits.

The stage with eight mesenteries is followed in a day or two by the stage with twelve. At this period the first deposition of the skeleton makes its appearance as a circular patch of calcareous matter on the basal surface of the attached polyp. I was not able to rear the young corals beyond this point ; though I kept a few alive until the first of June. The death rate increased very much as soon as the larvae became attached.

III. Early Stages, including the Formation of Layers.

The diameter of the unsegmented tgg was about two-thirds the long diameter of the blastosphere (Fig. 2, PL II.). There was a very large centrally-placed nucleus, and the body of the Qgg was filled with vesicles, resembling in this respect the endoderm of the subsequent planula. The only observations on the segmentation were made on the eggs accidentally (.') discharged March 21. The spermatozoa played round each Qgg in large numbers. After fertilization the spherical egg became oval, and then divided into two equal blastomeres, which remained connected by a bridge of tissue. The division into four then followed. So far the segmentation was quite regular, but irregularities now began to crowd in, which led to the production of a grotesquely shaped mass comparable with the irregular planula described by Metschnikoff for Oceania (2). (I have observed a precisely similar segmentation and planula in an allied medusa, Turritopsis.) The irregular mass did not develop any further.

The normal segmentation which goes on in the body of the parent results in the formation of a blastosphere with a very large cavity (Fig. i). The blastosphere is markedly bilateral, and is without cilia. The cells contain a large number of very distinct vacuoles, pretty evenly distributed through the cell body. In older blastospheres, Figs. 2 and 3, the formation of the larval or primitive endoderm has begun. The blastosphere cells are columnar, the nucleus is peripherally placed, and the vacuoles are concentrated in the central end of the cell. The cells are evidently delaminating, the inner vacuolated ends being split off to form the endoderm. The nuclei of the endo

196 WILSON. [Vol. II.

derm segments are very hard to make out. The delamination takes place irregularly over the general surface of the blastosphere, and is of a peculiarly complex character in certain spots. In the cell groups m and m\ Fig. 2, for instance, the lines of the transverse divisions of the several cells are continuous, and seem to indicate that the whole group must have divided transversely while it was yet a single cell : before it had been split up by the longitudinal divisions. But the latter show that the transverse divisions did not precede them. Such a cell as n, Fig. 3, throws light on the matter. This cell has begun to divide longitudinally, but even in the peripheral part of the cell, though the longitudinal constrictions are evident enough, the segments are far more intimately connected than are, say 7i, and the adjacent cells. In another cell, n\ the longitudinal division is apparently perfect ; but on comparing the close apposition of the new segments with the marked space between ;/' and the adjacent cells, I am inclined to believe that the halves of «' are still connected by bridges of tissue, much as the blastomeres in many segmenting eggs remain connected after they are clearly marked off. In the group vi\ then. Fig. 2, it is probable that the longitudinal constrictions were the first to appear in the mother cell, but that before they became complete divisions a general transverse constriction took place.

Delamination appears to be the exclusive means by which the endoderm is formed. I looked in vain for cells migrating from the surface. Whenever there appeared to be such, examination of the next sections showed that the cells m question had been cut obliquely, the peripheral end with the nucleus having been severed from the rest of the cell body. The cell a, Fig. 2, for instance, is merely the central portion of a constricted cell like b in the same figure. The delamination and accompanying longitudinal division often give rise to spheroidal cells, such as X, Fig. 3. These cells are always at the surface of the blastospheres. Whether their fate is different from that of the general ectoderm cells, I do not know.

The cavity of the blastosphere having been filled up by the endoderm segments, a solid planula, Fig, 4, is formed. On studying this figure, a median longitudinal section, we see that the longitudinal division of the superficial cells, which was active in the blastosphere, has given rise to a layer of columnar


cells, the ectoderm. The inner ends of these cells are not distinctly marked off from the solid endoderm. The endoderm, when the living larva is compressed or when pretty thick sections are examined, appears made up of a quantity of vesicles, which look something like fat cells. When thin sections are examined (Fig. 4), it is found there exists a continuous protoplasmic matrix binding the vesicles together, which are now seen to be identical with the vacuoles present in the endoderm segments of the blastosphere cells. There is, however, a difference in their appearance, which is best shown in Figs. 9 and 10. In the blastosphere cell, the protoplasm in which lie the vacuoles is pretty uniform ; but in the endoderm of the planula, the protoplasm immediately round each vacuole is denser than the intervening stroma, and forms a more or less differentiated shell for the vacuole. The shell stains much more deeply than the stroma. What the contents of the vesicle is, I cannot say. Carmine and haematoxylin do not affect it, and in these young stages I did not try osmic acid.

The endoderm contains, besides the vacuoles just mentioned, a number of "yellow cells" and scattered nuclei. No cell boundaries can be made out, and there is every indication that the layer is a plasmodium. A stage between Figs. 2 and 4 would go far towards elucidating the intricate structure of the solid endoderm, but I did not succeed in obtaining such.

While the embryo is still solid, the oesophageal invagination makes its appearance. Fig. 4. The larva is very narrow except at its base, where it is slightly dilated. At this stage it swims feebly about, the cilia having commenced to develop.

The formation of the permanent layers is illustrated in Fig. 5. In the region d the layers are in the condition characteristic of the earlier larva. But in the neighborhood of c the ectoderm cells have become much more clearly marked off from the endoderm. Their inner ends form a ragged line which gradually becomes more uniform, until the ectoderm is bounded internally by a smooth limiting surface, as at a. Though the bounding surface is smooth in this region, the supporting lamella has not yet formed. A little later it appears as a delicate membrane between the two layers (from a towards the mouth). While it is still very thin it resembles an ordinary cuticle. It appears between the layers, and is not formed by

198 WILSON. [Vol. II.

the direct metamorphosis of the ends of ectoderm or endoderm cells. As to the much-discussed question of which layer secretes the membranes ; in early stage, such as Fig. 5, it is impossible to decide whether the secretion is the peculiar property of either layer. But in later stages, it is found in places where it can only be endodermic, and in others where it is evidently ectodermic. I shall return to this point further on. The various steps in the formation of the permanent layers can often be observed, with but few gaps, in a single section. Fig. 5 is such a section, with a few additions from another.

The permanent endoderm is formed from the larval endoderm by a differentiation of a peripheral layer from the central portion, Fig. 5. The peripheral layer is not formed continuously over the whole surface, but in spots here and there. It is at first marked off from the central portion only by its somewhat greater density, but it gradually breaks its connection with the latter by the acquisition of a smooth limiting surface. Figs. 5 and 10. Cell boundaries soon appear. The permanent endoderm as thus formed is a single layer of vacuolated cubical cells, except in the immediate neighborhood of the oesophagus, round the sides of which the cells are massed so as to fill up, more or less completely, the space between it and the body wall. This is especially noticeable on the left side of Fig. 5.

The oesophageal invagination as seen in Fig. 5 is much dilated at its base. Here the ectoderm has preserved its intimate connection with the endoderm. the ectoderm cells not even having acquired a smooth bounding surface. In Fig. 6 this even surface has been acquired, but at no time is a supporting lamella secreted over this area. The absence of the lamella clearly facilitates the absorption by the endoderm (or yolk) of the base of the invagination. The details of the absorption are as follows. The dilated bottom of the invagination, shown in transverse section in Fig. 10, is broken through at one point. Through this breach the yoke passes into the lumen, Figs. 8-10, and the base of the oesophagus thus enclosed above and below by yolk is absorbed. It is, I think, possible to recognize the base after it has passed into the yolk, though of course the histological structure is gone. In Fig. 8 is seen a curved mass of tissue extending from the left side of the oesophagus nearly to the right. Though in structure like the yolk, it stains a much


deeper hue, and is therefore well defined. On comparing this figure with Figs. 6 and 10, there seems to be no doubt that the mass in question really represents the ectoderm base of the oesophagus.

The central portion of the larval endoderm remains as a foodyolk. In Fig. 5 it still exists as a continuous structure, but as more and more of the protoplasm is drawn into the layer of permanent endoderm, the vesicles lose their connection with each other. By the time the layers are definitely established. Figs. 6 and 8, the yolk is a loose mass of vesicles, the shells of which have begun to disintegrate. The shell is extremely dense, and in this and subsequent stages seems to be fatty, as it stains very dark with osmic. Remnants of the yolk are found in stages as late as Fig. 19.

Though the formation of the supporting lamella and the differentiation of the permanent endoderm very often take place at about the same time, this is not always the case. In many larvae the supporting lamella is entirely formed, and the cesophagus has opened centrally, while the endoderm is still solid. Figs. 10 and 12.

The "yellow cells," which later are found in such abundance, appear for the first time in the planula. Fig. 4. I have not seen them actually entering the planula, but in this and slightly older stages. Fig. 5, a few occur in the surface ectoderm, whereas in older larvae and in the adult they are confined to the inner layer of the body.

Comparisons. a. Germinal Layers.

Among the many ways in which the germinal layers are formed within the Coelenterates, Metschnikoff (2) has decided that one represents the manner in which the earliest metazoa were formed. In this so-called "mixed delamination " a solid endoderm is built up both by delamination and by the migration of superficial cells into the interior. Polyxenia (Metschnikoff) and, better, Ajirelia (Gotte) are good examples of this process. Accepting this type as ancestral, Manicina has diverged along the same path as certain Trachomedusae (Geryonia, Liriope);

200 WILSON. [Vol. II.

i.e., none of the blastosphere cells migrate into the interior, but the endoderm is formed exclusively by delamination.

Though the way in which the solid larva was originally formed seems preserved in but a few species, the solid larva itself may fairly be considered as typical for the Coelenterates. It is especially well preserved in the Hydroids and Anthozoa. In the latter group it is nearly universal among Alcyonarians, and occurs in the majority of the Zoantheria. The Alcyonarian which comes nearest to Majiicijia in the formation of its layers is Renilla (3). In this genus, though the blastosphere has a very small cavity, the endoderm is formed by delamination, as in the coral. As regards the structure of its larval endoderm, however, Renilla differs from Manicina. In the former the endoderm is made up of a mass of cells, of which the peripheral layer becomes the permanent endoderm, while the central cells go to pieces, and are probably eaten, amoeboid fashion, by the peripheral cells. In Manicina, on the other hand, the larval endoderm is a Plasmodium, and in the entire process which leads up to the formation of the adult endoderm it would seem that a prominent physiological part is played by the "vesicles," which in all probability contain some kind of yolk. Owing to the plasmodial nature of the solid endoderm, the complete transformation of the latter into the permanent layer must require less time in Manicina than in Renilla. For in Manicina there is no large accumulation of yolk cells which must slowly be devoured, amoeba fashion, by the peripheral cells. On the contrary, when the time for the formation of the permanent endoderm has arrived, the general protoplasm is merely drawn towards the periphery, and after it has there broken up into cells, there remains but little nutriment in the loose yolk mass. Just how the yolk is ingested after the cellular endoderm is formed, I am unable to say. It is not devoured, amoeba fashion, by the endoderm cells at large, though in the region of the oesophagus. Figs. 6 and 8, connection is maintained for a considerable time between the yolk and the endoderm, which elsewhere is completely formed.

Amongst the corals and actinias, two or three forms have been described as undergoing invagination, notably CeriantJuis (Kowalevsky 4) and Actinia (Jourdan 5). While the mere recurrence of invagination in a coelenterate group can scarcely be said any longer to have a phylogenetic significance, these two


forms are especially interesting, as E. B. Wilson has remarked (3), because in each of them a yolk mass appears after the layers have been completely formed. I have not seen Kowalevsky's figures, but the yolk shown in Jourdan's figure 119 is precisely like the yolk of a young Manicina (making allowance for difference in thickness of the sections). ThQ a prioiH improbability that the endoderm would first secrete a yolk (Kowalevsky) and then swallow it again, taken together with the similarity of the yolk in question to that of Manicina, might tempt one to believe that both authors had mistaken a stage like Fig. 12, PI. I., for a true gastrula. But Kowalevsky's statements on this head are so definite as to preclude this supposition.

b. Supporting Lamella.

The only two authors who have described in detail the formation of the supporting lamella are Jourdan {l.c) and Wilson {l.c). Jourdan's observations were made on a coral, Balanophyllia, and an actinia, Actinia equina. He draws a sharp distinction between the "membrana propria" and the jelly.

The former answers to the German Stiitzmembran. It is a firm limiting membrane which appears between the two layers and extends into the mesenteries to form their axial bands. This is what I have spoken of as the supporting lamella (comp. any of figures on PI. IV.). Its origin Jourdan was unable to trace with any certainty. The jelly, on the other hand, which eventually becomes fibrous, is formed outside the membrana by the superficial ectoderm cells. The inner ends of these cells break off and fuse together to form a granular mass, in which fibres subsequently appear {I.e. Figs. 119 and 129).

The supporting lamella described by Wilson is evidently the same thing as Jourdon's membrana. Its origin according to Wilson is double. In the mesenteries it is a simple cuticular secretion of the ectoderm cells. But in the body wall it is formed in a manner similar to that just described for the jelly: the inner ends of the ectoderm cells become swollen, constrict off, and form a granular layer which condenses to a smooth membrane.

In Manicina there is first formed a very distinct membrane as described. In later larval stages, at various spots, especially in angles, a thin fluid jelly accumulates. This is noticeable in

202 WILSON: [Vol. II.

Figs. i6 and 17, beneath the right-hand mesenteric filament, and in Figs. 44 and 45 at the angles of the oesophagus. It is always to be found in the axis of the larval filament, Fig. 26. The distinction between the supporting membrane and the more fluid jelly, which is noticeable in larval stages, is lost later in life. In the older larvae, Fig. 39 for instance, the supporting membrane throughout its whole extent has become considerably thicker than in younger stages, though the jelly in the axis of the filament is still to be distinguished from the more membrane-like band of the mesentery. In the adult the axial band of the mesentery is so much wider, while the jelly in the axis of the filament is at the same time denser than in larval stages, that all distinction between the two structures is lost. Compared with an actinia the whole mesodermic skeleton of the adult Manicina is very scanty and membrane-like, but in certain places it reaches a more generous jelly-like condition, for example, in parts of the mesenteries. Fig. 51. Here the supporting substance is merely the thickened primitive membrane.

There indeed seems to be no difference between the supporting membrane proper and the jelly, except in the mere quantity of the secreted substance and in the percentage of water. What applies to the origin of one should explain the origin of the other. Now as regards the membrane proper, I am quite sure that in Manicina it is formed as a cuticular secretion, and not by the direct conversion of the ends of ectoderm cells into granular matter, which subsequently condenses to a membrane. Turning now to the question as to which layer secretes the membrane, we see from the figures (14, 17, 26, 45) that after the mesenteries are formed, the lamella is much more intimately connected with the endoderm than with the ectoderm : where the layers are forced apart, it always sticks to the endoderm. But even here the ectoderm is provided with a well-defined limiting membrane, which if thinner than the lamella, is essentially like it. I conclude from this and other facts to follow that when the lamella lies between the two layers, both layers share in secreting it. That the endoderm cells can secrete the lamella, is plain from its occurrence in the axis of the mesenteries, and in the axis of the genital bands of a Cubomedusa (6). On the other hand, the same argument can be used to prove the ability of the ectodermal cells to secrete this substance, for in the velum of Hydromedu


sae there is a well-defined lamella, which is formed in situ. In Cunocantha (7) I have pointed out that when the velum is developing, the ectoderm cells range themselves so that their bases will secrete a continuous membrane.

IV. Formation of the First Pair of Mesenteries and


I was led to pay especial attention to the filaments by E. B. Wilson's interesting discovery that the dorsal pair of filaments in the Alcyonarian polyps are ectodermal lobes (8). Relying on the histological similarity between this pair of filaments and the flivimerstreifcn of an actinian filament, Wilson suggested that perhaps in the latter the flimmcrstrcifen were ectodermal, and only the median nesseldmsejistreif \\z.s endodermal. I find, however, that the filaments of Manicina are as to origin like the dorsal pair in the Alcyonaria. They are simple ectoderm lobes which grow down from the oesophagus.

The oesophageal invagination in the earliest stages is symmetrically placed, Figs. 4 and 5. It very soon, however, begins to travel towards one side of the larva. This is shown in the transverse section, Fig. 7, and still better in the longitudinal section, Fig. 6 (the endoderm in these two larvae is in different stages of development). The lateral motion of the oesophagus has compressed the endoderm on its left into a compact mass, which completely fills the space between the oesophagus and body ectoderm. On the right hand the endoderm has been stretched, until it forms a single layer of cells. The narrow lumen of the oesophagus is bilateral (it is doubtful whether this is true in all larvae at this stage), and its dilated extremity has no longer the symmetrical bulb-like shape of Fig. 5. In Fig. 8 the lateral movement has gone a step farther, and meanwhile the oesophagus has opened centrally. The movement of the oesophagus is continued until in one meridian there is nothing left between the oesophageal and superficial ectoderm but the supporting lamella. Figs. 11 and 12. (In Fig. 11 the lamella has not yet appeared, but both stomadoeal and body ectoderm have smooth, limiting surfaces.) The intermediate stages show that the lateral movement of the oesophagus travels from above downwards, and that the endoderm has consequently been pushed

204 WILSON-. [Vol. II.

down in this meridian. This is proved by Figs. 9 and 10, transverse sections from a larva in which the oesophageal movement had gone farther than in Fig. 8. Fig. 9 is about at the level a in Fig. 8, and Fig. 10 is at the level b. In this larva then the oesophagus was closely pressed against the ectoderm at the level a, but was separated from it by endoderm at a lower level b. In the later stage, Fig. 1 1, the lowest part of the oesophagus has completed the journey. The meridian in which the oesophagus is thus pressed against the ectoderm, is that of the first mesentery. (Though Figs. 9 and 10 are from the same larva, both the supporting lamella and endoderm are much farther advanced in the lower section than in the upper.)

The larva from which the longitudinal sections, Figs. 12 and 13, were cut, was very backward in forming the permanent endoderm. Fig. 12 is taken through the line a and b in Fig. II, and is in the plane of the first pair of mesenteries. It is only on the left side that the oesophagus is in contact with the body ectoderm. But the right side is following suit, and in a slightly older stage. Fig. 12, is in the same manner applied to the surface ectoderm over a narrow tract. Fig. 13 is to one side of a and b in Fig. 11, and is consequently out of the plane of the mesenteries.

My next stage after Fig. 12, is the larva from which the series of transverse sections. Figs. 14 to 17, was taken. It is this stage which proves the meridians, along which the oesophagus is applied to the body ectoderm, to be really those of the first pair of mesenteries. Fig. 14 is the uppermost of the series, and is through the body of the oesophagus, the lumen of which is exceptionally large. On the right side the oesophagus is separated by jelly alone from the surface ectoderm. Following down the series of sections, we see that in this meridian the oesophageal ectoderm sends down a slender lobe, which, like the oesophagus above it, divides the endoderm and rests on an accumulation of jelly (right side of Figs. 15, 17). This lobe is a mesenterial filament. It is considerably shorter than the filament on the opposite side. Fig. 17, and is probably the second one of the first pair. The right side of this larva is practically in the condition shown in Fig. 12, except that the oesophagus has formed a filament.

On the left side of the larva, Fig. 14, matters are tnore


advanced. On comparing this side with Fig. 12, we see the oesophagus has moved away from the surface ectoderm, but while doing so has remained connected with it by a band of supporting lamella. Running through the series of sections, it is found that in this meridian also an ectoderm lobe has grown down from the oesophagus. This lobe is the primary filament. It is much wider at its start, near the oesophagus, than its fellow on the opposite side, but soon dwindles to about the same size. Like the oesophagus above it, this filament is in a more advanced condition than the filament on the right side. Since we know that the left side of the oesophagus itself has passed through the condition which exists on the right side, we are pretty safe in believing that the left filament has likewise passed through the condition in which the right filament is found. The lack of a fairly intermediate stage between Fig. 12 and the larva we are studying, is to a certain extent supplied by the larva from which Fig. 1 1 was taken. In this individual, which was sectioned transversely, the oesophagus extended so much farther down in the meridian of the first mesentery than round its general lip, that in this meridian it formed a very evident though not a very long lobe. This lobe, which becomes the primary filament, is shown in Fig. 11, and is to be found in the sections immediately below the one figured, growing smaller towards its end.

On comparing the first and second filaments in Fig. 17, it is clear that the endoderm, which at a higher level grew in between the oesophagus and body ectoderm, has likewise forced its way beneath the primary filament, and thus given rise tc the first mesentery. The mesentery is continued from the level of Fig. 17 downwards, as a very slight endodermic ridge on which rests the filament. The axial band of supporting lamella is continuous with a thinner lamella, separating the filament from the endoderm. Immediately beneath the opening of the oesophagus the mesentery is much more elevated than at a lower level. The elevation of the mesentery in this region is connected with the first appearance of an intermesenterial chamber.

In the larva we have been studying the first pair of intermesenterial chambers has been marked off. Round the oesophagus, however, they are still solid, Fig. 14, though at a lower level the hollowing out of the solid endoderm has begun. At this level, Fig. 15, which is just beneath the opening of the oesopha

2o6 WILSON: [Vol. II.

gus, the endoderm on one side of the first mesentery exhibits a small cavity. The cavity, when traced through the series of sections, is found to open into the coelom in Fig. i6, forming, as it does so, the cave or bay which underlies one-half the primary filament in the figure. It is thus (comp. Figs. 14 and 15) in the larger of the two primary chambers that the excavation of the solid endoderm begins, and, as I learn from other series, the excavation of this chamber is nearly completed before that of the smaller begins. The excavation starts, as we have just seen, in the immediate neighborhood of the first mesentery. From this spot it gradually extends across the chamber to the second mesentery, travelling all the while from the lip of the oesophagus upwards.

The second mesentery, which exists in its embryonic condition in Fig. 14, follows the example of the first. In Fig. 18 it is completely formed, and in the series of longitudinal sections, Figs. 20-23, which I shall describe later, the second mesentery is in the condition in which the primary is in Fig. 14 ; i.e., in the immediate neighborhood of the second mesentery the larger of the two intermesenteric chambers is solid, while the smaller chamber is entirely solid.

In the larva from which the transverse section. Fig. 18, was made, the various processes which have been described are now completed. The oesophagus is swung by two complete mesenteries. Both intermesenteric chambers — the larger on the left, the smaller on the right — have been hollowed out. Below the oesophagus, the mesenteric ridges extend the whole length of the larva, and the first pair of filaments about half the length. The section is slightly complicated by other features, of which I shall now speak.

In the larger of the two chambers, Fig. 18, the second pair of mesenteries, 3 and 4, has appeared. The axial bands of supporting lamella as yet cause no elevation of the endoderm, and at a level slightly below the oesophagus are entirely lost. In Fig. 18 a, above Fig. 18, the axial band numbered 4 stretches across to the oesophagus, and still higher up 3 does likewise. In this section the position of the oesophagus is eccentric. This is very often the case at the extreme upper limit (the appearance is not due to obHque sections), and hence one mesentery usually runs out before its fellow (2 before i). The mode of origin of


the second pair of mesenteries is thus entirely different from that of the first pair. All the subsequent mesenteries are formed after the fashion of the second pair.

The outer wall of the larger intermesenterial chamber in Fig. 18 is made up of unmistakable endoderm, but the inner or oesophageal wall has an ephithelium precisely like that of the oesophagus. Both are composed of very slender elongated cells with a median thickening, in which lies the nucleus ; the peripheral end is enlarged and flattened, so that by the juxtaposition of many such cells a continuous cuticle can be formed. The cells in question resemble the well-known "supporting cells " of the Hertwigs (9). Besides the supporting cells, glandular cells are found. These are slender and full of granules, the latter staining very deeply with haematoxylin. The epithelium of the oesophageal wall of this chamber is, moreover, sharply marked off from the rest of the epithelium. It is directly continuous with the oesophageal ectoderm round the lip of the oesophagus, and is evidently a tract of ectoderm. Any doubt which might cloud this point is removed by later stages, such as Figs. 44 and 45, where the epithelium, which is claimed as ectoderm, overlaps, at its upper limit, the endoderm. It is clear that in this stage, Fig. 18, the oesophageal ectoderm has been reflected round that portion of the lip (free edge of oesophagus) which belongs to the larger chamber, and has then run up along the outer surface of the oesophageal tube, driving the endoderm before it. The direction of growth is reversed, but otherwise the ectoderm is acting in precisely the way which it chose in forming the first pair of filaments. In the smaller chamber of Fig. 18, the epithelium forming the oesophageal wall does not differ from the rest of the endoderm. In a later stage, when the third pair of mesenteries has appeared, the ectoderm is also reflected round the lips of this chamber, and runs up along the oesophageal wall, Figs. 36 and 39.

The reflection of ectoderm, which leads to the condition shown in Fig. 18, commences as soon as the excavation of the primary (larger) chamber begins, and follows close on the heels of the latter process. The object of the reflection of ectoderm, as will be shown later, is to provide filaments for the young mesenteries before the latter are complete ; i.e., continuous from body wall to oesophagus. Bearing this in mind, it is prob

208 WILSON. [Vol. II.

able that the reason why the ectoderm is reflected into the primary chamber at such an early date, is that the second pair of mesenteries will be formed in this chamber. The series of longitudinal sections, Figs. 20-23, illustrates the early reflection of ectoderm round the lip of the larger chamber. The larva from which this series was made was at a stage intermediate between Figs. 14 and 18. The supporting lamella of both of the first pair of mesenteries were complete, as in Fig. 18 ; but the intermesenterial chambers were far from perfect. The smaller chamber was entirely solid, and the cavity of the larger was only partially formed. In spite of the difference in age, it will do to refer the planes of the longitudinal sections to Fig. 18. The section, Fig. 20, thus is taken through the mesenteries i and 2, and the corresponding filaments. The next three figures represent a series of sections to the left of this plane, the one farthest to the left lying in the plane a-b of Fig. 18. Finally, the left-hand mesentery in Fig. 20 is the primary mesentery, I in Fig. 18. With this orientation it is seen, on glancing through the series 20-23, that only in the neighborhood of the primary mesentery has the intermesenteric chamber been hollowed out. In all the sections the endoderm is solid on the right side of the oesophagus. Now, in Fig. 23, the intermesenteric chamber (to the left of oesophagus) is in the state in which it was formed. The ectoderm has not yet been reflected round the oesophageal lip. But in the sections nearer to the mesentery. Figs. 21 and 22, the ectoderm Jias been reflected, and has driven the endoderm before it. The reflection of ectoderm, as I have said, follows very closely the excavation of this chamber. As the latter proceeds from mesentery i to mesentery 2 (Fig. 18), so does the former; and by the time the chamber is completely established, the ectoderm is reflected all the way from i to 2, as is seen in Fig. 18.

In the larva which was used to show the earliest appearance of the primary intermesenteric chamber, Figs. 14-16, the reflection of ectoderm had already begun to take place round one edge of the primary oesophageal lobe (Fig. 16, left).

To sum up the more important events described in this section :

To form the first mesentery the whole oesophagus moves laterally, until in the meridian of the mesentery there is only sup


porting lamella between the oesophageal and surface ectoderm. The oesophagus now grows downward in this meridian as a lobe of ectoderm, which represents the primary filament, and which pushes the endoderm before it. On the opposite side of the animal, along the line of the second mesentery, the oesophagus becomes applied to the body ectoderm in the same manner, and a lobe grows down from it to form the second filament. The mesenteries as such are formed by the ingrowth of the endoderm between the body ectoderm and oesophagus above, and between the body ectoderm and the filaments below. The primary pair of intermesenterial chambers are at first solid. The larger chamber acquires its cavity before the smaller, the excavation travelling from the lip of the oesophagus upward, and from the first toward the second mesentery. The excavation of the primary chamber is closely followed by the reflection of ectoderm into this chamber, the reflected ectoderm running up the oesophageal wall, and driving the endoderm before it. The second pair of mesenteries appear in the larger chamber as longitudinal ridges of the supporting lamella, which cause no elevation of the endoderm.

V. Histology of the Larvae.

It may now be advisable to describe the histology of the larvae, going when necessary beyond the stages already studied.

a. The Surface Ectoderm.

The ectoderm at the time when the supporting lamella is formed, Figs. 5 and 6, consists of columnar cells, the protoplasm of which shows a great tendency to break into small polygonal balls. While the ectoderm is in this condition the mucus cells appear {m. Fig. 6, Figs. 7, 8), as pear-shaped bodies in the peripheral ends of the columnar cells. The contents of the pear-shaped body stains a deep blue with haematoxylin, and is thus distinctly marked off from the surrounding cells. It appears to be fluid from the start, as even in such young stages as Fig. 7, many of the mucus cells have poured out their contents, which adheres to the mouth of the cell as a little mass of a blue color. This tendency of the mucus cells to eject their contents, presumably when the killing fluid touches the larva, is very

2IO WILSON-. [Vol. II.

noticeable in older stages such as Fig. 27, which are often completely covered by a thin layer of mucus. The mucus cells increase in size and number, until they become the prominent feature of the ectoderm (Figs. 7, 20, 27). In their final condition, Fig. 27, they are large, clear sacs, in which a few strands of protoplasm may be seen in carefully prepared osmic specimens. The nucleus of the original cell in which the sac developed cannot be made out in sections.

The mucus cells develop over the general surface of the larva, but are not found at the aboral end. Here their place is taken by a slender elongated cell, full of granules which stain even deeper with haematoxylin than does the mucus cell. The granular gland cell is easily recognized in its earliest stage, and is formed from the embryonic ectoderm cell by a deposition of granules throughout the length of the cell. The granular cells, though especially grouped at the aboral end of the larva, are found here and there over the general surface.

The remaining cells of the embryonic ectoderm become for the most part transformed into slender "supporting cells." Thread cells, n.c, are first noticed in larvae at about the stage of Fig. 14. Whether the ectoderm contains any muscle cells, I do not know. The larvae can alter their shape to a great extent, but the fibres are probably all endodermal.

For the study of nervous elements, Manicina in all stages is a very unfavorable subject. Even the ganglion cells which show so plainly in sections of actinian larvae, I was not able to make out in the coral. There is, however, a finely granular stratum at the base of the ectoderm, which is very thin over the general surface. Fig. 26, but at the aboral end of the swimming larva is thick and easily seen. Fig. 20. When the living larva is compressed, this accumulation of granular substance is very noticeable. It is perfectly clear, and until I began to section I thought it was jelly. When carefully examined, the granular layer in this region is found to consist of a mass of fine fibrils. It is very probably nervous. The fact that the coelenterate planula in general swims with its aboral end in front, taken together with the occurrence of a bunch of long cilia on this end of many actinian larvae, suggests the existence of some such primitive nervous centre as I take this accumulation of granular matter to be.


When the various elements of the ectoderm have been completely differentiated, Fig. 26 (from a stage a trifle older than Fig. 29), a finely striated cuticle is secreted, in all respects like that described by the Hertvvigs for the actinias (9. Taf. III.). The cilia with which the body ectoderm is completely covered, and with which the oesophagus and filaments are likewise provided, I have not represented in the figures.

b. CEsophagiis and Filaments.

The oesophagus and young filaments. Fig. 20, have the same structure as the superficial ectoderm, except that mucus cells are absent, and the granular gland cells more abundant than over the general surface. As a rule it is only in the youngest stages that the reflected ectoderm. Fig. 18, contains gland cells, and even then they are rare. In later stages. Fig. 24, mucus cells are found in the upper part of the oesophagus immediately round the mouth, and very large thread cells, Figs. 24 and 27, appear in considerable numbers in the oesophageal epithelium, extending down into the filaments. The histology of the filaments will be treated in detail in a special section.

c. Endoderm.

The endoderm, after it has once formed, remains about the same during larval life. I am not sure whether it is ciliated. The cells are cubical or columnar, and contain one or more larsre and distinct vacuoles. The protoplasm is granular and coarsely reticular. The cell outlines can only be distinguished with a little care. The number of yellow cells steadily increases with age. A few very fine muscle fibres can be made out here and there, especially in the mesenteries, and a continuous layer probably exists.

In closing, the general similarity between the ectoderm of the coral larva (practically the same in the adult Manicina) and the actinian ectoderm, as described by the Hertwigs, may be noticed. In particular, the two kinds of gland cells in the actinias are exactly represented in the coral. The Hertwigs suggested that possibly the granular gland cell was but a stage in the development of the mucus cell ; but this appears not to be the case, as the two are distinguishable from the start, and have also a different distribution.

212 WILSON. [Vol. II.

VI. Mesenteries and Filaments from the Second to the

Sixth Pair.

Before proceeding to the detailed description of the reflection of ectoderm, which, as I conclude, leads to the formation of all the filaments except the first pair, it will be convenient to give an account of the order in which the mesenteries appear.

The first appearance of the second pair of mesenteries has already been described. Fig. i8. In a later stage, Fig. 27 (transverse section) this pair is complete at the level of the figure. At a lower level, just above the lip, Fig. 28, one of the pair, 3, is still complete, while the other, 4, is incomplete. The difference in growth between the two mesenteries of the second pair is an exception ; they usually develop at the same rate. Below the oesophagus. Fig. 29, both mesenteries exist as slight ridges, which extend the length of the larva.

The third pair of mesenteries have appeared in Fig. 27. They remain insignificant during the swimming life of the larva. After the larva has become attached, the third pair is prominent ; and the fourth pair is also found, in the position shown in Fig. 34 (a transverse section taken below the oesophagus — the primary mesenteries are supplied with large and coiled filaments). According to Lacaze Duthiers, the fourth pair appears between the first and second pairs. The Hertwigs suggested (i*) on general grounds of symmetry that the order of appearance was as I have figured. As regards the fifth and sixth pairs, however, the old account of Lacaze Duthiers holds for Manicina as against the figures given by the Hertwigs for Adamsia (Taf. i, f. 3). This is seen on referring to Fig. 39, a transverse section through the oesophagus of a stage with twelve mesenteries. The fifth and sixth pairs develop simultaneously on opposite sides of the primary mesenteries. In Adamsia, according to the Hertwigs, the two pairs of mesenteries appear on opposite sides of the long axis of the oesophagus, in the chambers between the first and second pairs.

Fig. 39 represents my oldest larval stage, and I was consequently not able to trace the development of the muscle plates and the rearrangement of the twelve primary mesenteries. According to the accepted account, the pairs 3 and 4, in Fig.


39, become the directive pairs, and on each side of the long axis of the cesophagus, 5 and i unite to form one pair, and 6 and 2 another pair.

In this connection it may be observed that after the larvae became attached, the lumen of the oesophagus was decidedly bilateral in the living animal, and bore the relations to the mesenteries shown in Fig. 39. The shape of the lumen is, however, susceptible of great changes, and in many of the sections through attached larvae the typical shape is not retained (Fig. 36), or the oesophagus is compressed in the (normally) long axis, and drawn out in the short (Fig. 45). In the swimming larva the small size of the mouth will not permit one to decide as to the shape of the oesophagus in the living animal. After examining a large number of sections I conclude that in stages as old as Fig. 27, the arrangements are as in the attached larva: the lumen is bilateral and has one of the first pair of mesenteries on each side of its long axis. In earlier stages, such as Figs. 7, 14, 18, I could not decide whether the oesophagus had acquired its ultimate shape.

We now come to the formation of filaments for the second and subsequent pairs of mesenteries. In Fig. 18 the reflected ectoderm which forms the oesophageal wall of the larger chamber extends horizontally from i to 2. The mesenteries of the second pair, 3 and 4, are extremely incomplete. In Fig. 27 these mesenteries are complete, and now the reflected ectoderm does not extend from i to 6, but is represented by the patch R.E. between 3 and 4. In sections of the same larva (Figs. 28 and 29, below the level of Fig. 27) a certain irregularity makes its appearance, which is connected with the formation of the third mesentery, and of which I shall speak later. In the typical larva of this stage the sections below Fig. 27, between it and the lip of the oesophagus, are precisely like Fig. 27. This means that the mesenteries 3 and 4 are both complete throughout the length of the oesophagus, and that the oesophageal wall of the chambers inclosed between 2 and 4, and 3 and i respectively, is endodermal. At the very lip of the oesophagus in some larvae the oesophageal walls of these chambers are ectodermic. ' In a section just below the oesophagus it is found that the mesenteries 3 and 4 have short filaments. Further, the tract of ectoderm R.E. only extends upwards for about

214 WILSON. [Vol. II.

one-third the length of the oesophagus, while it will be remembered that in this stage (Fig. i8) the reflected ectoderm extends nearly the whole length of the oesophagus. Now what has taken place in the transition of the stage Fig. i8 into the stage Fig. 27 . Plainly it is that the mesenteries 3 and 4, as they grew down and gradually became complete, carried along with them the reflected ectoderm, part of which came to lie along the mesenteries as short filaments, while the remainder was divided into three portions. Of these, the two lateral, which originally extended from 2 to 4 on the one side, and from i to 3 on the other (Fig. 27), were carried all the way down to the lip of the oesophagus, while the middle portion R.E. was not carried the whole way down.

A couple of longitudinal sections will further elucidate the series of changes which separate Fig. 18 from Fig. 27. The sections given in Fig. 25 were made from a larva at about the same stage as Fig. 27. The two mesenteries of the first pair had long filaments, and the mesenteries of the second pair were complete and equally advanced. The section <^ is a radial halfsection through one of the second pair of mesenteries, say 4 in Fig. 2"]. The mesentery is complete, and to its edge clings a short filament. The other mesentery of the second pair, with its filament, is the exact counterpart of the one figured (in cutting the larva, radial sections of one mesentery and equally true transverse sections of the other were obtained). The precisely horizontal plane in which the filament of b lies, is probably due to the sudden expansion of the larva, on being killed, in the direction of the shorter transverse axis of the oesophagus (the transverse section would be elongated in a direction at right angles to the long axis of Fig. 27). The half-section c in Fig. 25 is reversed so as to complete the oesophageal lumen. It is the second section to one side of b, and a is the second or third on the other side. Comparing all these figures with Fig. 27, b is through mesentery 4, a is on the far side of 4, and consequently cuts the reflected ectoderm R.E. ; ^ is on the near side of 4, and the ectoderm is not reflected round the free edge of the oesophagus.

Fig. 24 is a single longitudinal section from a stage slightly older than the one just described. On the right it is through one of the second pair of filaments, and on the left through an


intermesenteric chamber. Referring the section to Fig. 27, the right half is through mesentery 4, and the left through the chamber opposite at about the point x.

Having now described the ordinary way in which the first and second pairs of mesenteries and filaments are formed, I will take up an exception, which has more bearing on the relationship between the Anthozoa and Scyphomedusae. Figs. 30-33 are transverse sections, numbered from above down, of a larva in which the first pair of filaments extended about half the length of the body. The oesophagus throughout its vertical extent is apposed to the surface ectoderm over a wide tract, « to ^ in Fig. 30. On running through the series of sections it becomes evident that the tract a-b is not the meridian of a single mesentery, as it would be in a normal larva like Fig, 14, but is the space between two mesenteries. The first section below the oesophagus. Fig. 31, shows that a rather wide lobe of ectoderm is growing down, and also that between a and b this lobe has been forced apart from the body ectoderm. In Fig. 32 (two sections omitted between 31 and 32) a lobe of endoderm has grown in between a and b, and has thus given rise to two mesenteries, which are provided with a common filament. In the section below, Fig. 33, the lobe of endoderm is hollowed out, and the two mesenteries definitely established. In a section (not figured) below Fig. 33, the mesenteries a and b exist as separate ridges, and the common filament has split into its constituent parts, I and 3. The filament 3 extends a very short distance down, and the mesentery b only reaches the equator of the larva. The mesentery a, with its filament i, is the fellow of 2, on the opposite side of Fig. 32. These two mesenteries belong to the first pair. I have assumed that a is the primary mesentery, since its filament is larger than that of 2.

Bearing in mind the normal development as illustrated in Figs. 12 and 14, the exceptional character of this larva is due to the fact that one of the second pair of mesenteries is formed at the same time and in the same manner as the mesenteries of the first pair. This aberrant member of the second pair, b, Fig. 30, etc., may be called the third mesentery. We may suppose that the oesophagus, applying itself to the body ectoderm along the line not only of the first a, but also of the third mesentery b, was unable to force down the endoderm in these lines without

2i6 Wilson: [Vol. ii.

carrying down at the same time the endoderm between a and b. Thus came the condition shown in Figs. 30 and 31. When the time came for the mesenteries as such to be formed, the endoderm which had been pushed down between a and b was compelled to grow up again, becoming excavated so as to form the intermesenteric chamber. It is in this process that we see the endoderm in Figs. 32 and 33. The irregularity of the larva, it will be noticed, is confined to one side : the right side of the sections is normal, and doubtless the other member of the second pair of mesenteries would have developed in the usual way.

The larva just described was the only young specimen I observed, which exhibited this peculiar variation. I found a few older individuals, however, in which one of the filaments of the first pair was so intimately connected with one of the second pair, as to render it probable that the two were simultaneously formed from a common lobe in the manner shown in Figs. 30-33. The larva, Figs. 27-29, to which I have already referred several times, comes under this head. Though the section. Fig. 29, is an appreciable distance below the oesophagus, the filaments i and 3 are still united, and bear evidence of their common origin. The other mesentery of the second pair, 4 in Fig. 28, has itself been formed, and is gaining its filament, in the normal manner. The reflected ectoderm only extends from 2 to 3, and if the latter mesentery has been formed in the manner suggested, the ectoderm never was reflected between 3 and I.

Returning to the normal development we have now to trace the origin of the filaments for the third pair of mesenteries. These filaments are derived from a lobe of ectoderm, which is reflected into the smaller of the two primary chambers. The reflection takes place after the larvae have become attached. In a stage with twelve mesenteries. Fig. 39, the lobe is marked x. The section is taken at a level higher than that reached by the lobe belonging to the chamber c (marked R.E. in Fig. 27). In a section just above Fig. 39 the third pair of mesenteries bury themselves in the lobe x, much as 4 does in Fig. 28. In the uppermost sections, where the mesenteries are complete, x is not found. After a great many trials I succeeded in getting a radial section through one of this pair of mesenteries, which had forced down the reflected ectoderm so that the latter lay


along its free edge as a tiny filament (Fig. 38, left half; the right half is through chamber c, Fig. 39). The lobe for the third pair of filaments is almost always present in larvae with eight mesenteries : in Fig. 36, a section just above the lip, it is marked c. In this larva the mesenteries of the second pair are backward in development ; they are not yet perfectly complete, though on one of them a very small filament is seen.

As the fourth pair of mesenteries continue to increase in size, the tract of ectoderm, which belongs to the chamber c, in Fig. 39, extends farther upwards. In Fig. 39 it reaches about the same level as in Fig. 27 (R.E). In a slightly more advanced stage, Fig. 38 (right half), it extends nearly the whole length of the oesophagus ; and in the series of transverse sections. Figs. 43-45, it likewise extends to nearly the uppermost limit of the chamber. The latter series is a very good one, as the polyp was killed in its natural shape. In life, when the young coral is expanded, there is a very distinct oral cone, indicated by the line 7n-m in Fig. 38. In dying, the oral surface is almost always retracted, and the median longitudinal section is then as in Figs. 37 and 38. The transverse section, Fig. 43, is taken through the line x-y, in Fig. 38. Figs. 44 and 45 lie above it. In the larva from which the series was taken, both the third and fourth pairs of mesenteries were very well developed below the level of x-y, but above this level they were not perceptible. Judging from all my other sections I should say that this rapid development of the mesenteries along the sidewall of the polyp, as contrasted with their backwardness in the upper oesophageal region, was exceptional. In the sections figured the first and second pairs of mesenteries are complete. The chambers a and b, together with c, represent the larger of the two primary chambers. The walls of a and b are endodermal except for a short distance above the oesophageal lip, but in c and in d (the smaller of the two primary chambers) the reflected ectoderm extends very far up. As the third pair of mesenteries lie in chamber d, the ectoderm of this chamber, as might be expected, extends farther up than in c, in which lie the fourth pair. In the ectoderm of d there are a few large nettle cells, a rare occurrence. At the upper limit of each tract of ectoderm, the overlapping of the layers, previously referred to, is shown. This overlapping of the layers, while uncommon,

2i8 WILSON. [Vol. II.

is noticed in a good many specimens. In Fig. 37 the left half of the section, through an intermesenteric chamber, exhibits this phenomenon. The right half of the section is through one of the second pair of mesenteries.

The fifth and sixth pairs of mesenteries appear simultaneously, but it is convenient to speak of one pair as the fifth and the other as the sixth. They are still very small in my oldest larval stage. The filaments for the fifth pair are probably formed from the lateral portions of the lobe x, Fig. 39, after it has been divided by the completion of the third pair of mesenteries. The filaments for the sixth pair, it seems, will be formed from the tracts of ectoderm, which belong to the chambers a and b, Fig. 39. These tracts, it will be remembered, were in most larvae pushed completely back to the free edge of the oesophagus, where the second pair of mesenteries become complete. In stages with twelve mesenteries, however, such as Fig. 39, they have again appeared, though they usually extend but a very short distance above the lip. In a couple of larvae as old as Fig. 39, they were unusually well developed, reaching as far up as the tract for the chamber c.

In summing up the facts of the reflection of ectoderm it will be convenient to refer to Fig. 39.

The ectoderm reflected into the larger of the two primary chambers is pushed down by the growth of the second pair of mesenteries. From it are formed the filaments for these mesenteries, while the remainder of the original tract splits into three divisions. The middle division, chamber c, is not pushed entirely to the edge of the oesophagus ; later in life, when the fourth pair of mesenteries is well developed, this tract grows once more nearly to the upper limit of the oesophagus. The lateral divisions a and b are pushed to the edge, but after the sixth pair of mesenteries has appeared they begin to grow up again. When the mesenteries of the third pair are well advanced, the ectoderm is reflected into the smaller of the two primary chambers, and runs up the oesophageal wall nearly to the top of the chamber. The mesenteries, when they begin to grow down, carry a part of the ectoderm along their free edges as very slender filaments. The growth of the various tracts of reflected ectoderm is thus seen to follov/ in general the order of development of the mesenteries.


VII. Origin of the Filaments in the Adult.

After studying the larval development it seems very sure that the filaments of the first twelve mesenteries are ectodermal. Further, I think the stage with twelve mesenteries holds the key to the condition in the adult. In this stage, Fig. 39, etc., there are complete (first and second pairs) and incomplete mesenteries (third pair), both provided with ectodermal filaments. Both kinds of mesenteries are exactly comparable with the two kinds in the adult, and if the incomplete mesenteries of the larva are successively supplied with filaments by the reflection and upward growth of the ectoderm, it seems probable that the incomplete mesenteries of the adult are supplied in the same manner.

The gap between my series of larval stages and the adult is partially bridged over by the transverse sections, Figs. 41 and 42. The young Maniciiia from which these sections were made was one-eighth inch diameter. I found a couple of about the same size on a piece of coral rock. In hardening these two specimens I was not fortunate, and they were consequently of no value for the study of such fine points as the reflection of ectoderm. Fig. 41 is through the oesophagus, and disregarding the skeleton, shows six pairs of complete and six pairs of incomplete mesenteries. The twelve complete mesenteries represent the mesenteries present in the larva, now rearranged in pairs and with simple muscle plates (only shown in the directive mesenteries). The incomplete mesenteries have appeared according to the general law governing the mesenteries above twelve.

Manicina remains a single polyp until it has reached a diameter of about one-third of an inch. In this condition it has all the characters of the adult, except those dependent on asexual multiplication. It is, moreover, not sexually mature. Such a coral has, disregarding local irregularities, twelve pairs of mesenteries of the first order (complete), twelve of the second order (incomplete), and twenty-four pairs of the third order (much more incomplete). Fig. 50 gives a median longitudinal section of the coral at this age. The polyp was in a state of complete contraction, the oral surface or peristome, Pr, pulled down, the mouth, Mo, widely open, and the tentacles, 7", retracted. The

220 WILSON. [Vol. II.

section on the right is through a mesentery of the first order, on the left through one of the second order. The line 3 marks the position of the free edges of the tertiary mesenteries. The latter are not provided with filaments. On each side of the section, the mesentery is divided into a central and peripheral part (R.P.) by the calcareous theca, T/i. (When the animal is expanded, the peristome is lifted high above the level of the skeleton. It then embraces the whole width of the animal, the tentacles a dense ring round its edge, while the mouth is narrowed to a slit-like opening.)

On the right side of Fig. 50 the epithelium of the oesophagus, CE, is directly continuous with the filament i, the mesentery being complete. On the left side the mesentery being incomplete, the oesophagus has a free edge. Now if my view is correct, not only the lining epithelium of the oesophagus is ectodermal, but the epithelium x, which forms the gastric covering of the oesophagus and peristome between the complete mesenteries, is likewise ectodermal : the ectoderm here, as in the larva, is reflected round the lip of the oesophagus, and extends upwards until it reaches the secondary mesentery, down which it courses as the filaments. Fig. 55 is a more highly magnified view of the lower part of the oesophagus, as shown in the left half of Fig. 50. The lining epithelium of this part of the oesophagus, the lower third, is composed of slender supporting cells. The upper two thirds contain large nettle cells, and in the region of the mouth mucus cells. The epithelium on the outer surface of the oesophagus, is for some distance (compare Fig. 55 with Fig. 50) exactly like the lining epithelium, except that it contains a few yellow cells. Then comes a region of vacuolated cells, which is followed by a very low epithelium made up of exceedingly small cells, the exact shape of which I could not determine. The low epithelium is continued up the oesophagus and over the peristome, and is continuous with the filaments on the secondary mesenteries. These filaments, though of large size at a lower level, gradually become very small as they approach their upper limit, and by this means run without any break into the low epithelium covering the peristome. The filaments on the third pair of mesenteries in the larva were likewise very small in the upper part of their course, though, to be sure, in my oldest larval stage they as yet only existed in this part.


The view that the gastric Hning of the oesophagus and peristome is ectoderm can only be held by supposing the epithelium in question to have suffered a great histological change, (It is possible that a very careful histological examination of the epithelium, X, would show that it does not differ so much from the ectoderm as appears to be the case.) But though this tract of epithelium cannot be said to resemble the ectoderm, it differs quite as much from the undoubted endoderm. The endoderm of the adult. Figs. 51 and 52, is made up of large irregularly columnar cells, packed with "yellow cells," the vacuoles present in the larval endoderm being confined to special localities. The only apparent alternative to the view offered is that while the first twelve mesenteries are provided with ectodermal filaments, the filaments of all subsequent mesenteries are endoderm. Since the several orders of filaments in the adult differ only as regards size, and since even this difference is a transient one, owing to the constant transformation of incomplete into complete mesenteries during the growth of the coral, it seems improbable that such precisely similar organs should be formed by both layers.

On the other hand, E. B. Wilson (8) came to the conclusion that the dorsal pair of filaments in the Alcyonaria were ectodermal lobes, but that the remaining six filaments were purely endodermal. The two kinds of filaments in these polyps have, however, a very different histological structure, with which is associated a division of labor.

Von Heider (11) several years ago decided that the filaments of Cerianthjis were ectodermal. He reached his decision by a histological study of the adult, and though this method is mconclusive, I am not surprised after studying myself some immature specimens of Cerianthus, that he came to this view. The Hertwigs in their classical work on the actinias pointed out that embryological deductions based on adult histology are not very reliable, and also brought forward as an objection to von Heider's view, the existence of filaments in the actinias generally on incomplete mesenteries.

222 WILSON. [Vol. II.

VIII. Histological Structure of the Filaments.

The very young mesenteric filament is shown in cross-section in Fig. 17. In this larva, the first pair extend about half the length of the body. In Fig. 29, though the mesentery is elevated above the general endoderm, the filament retains its simple character. It is roughly hemispherical in section, and is separated from the mesentery by a thin sheet of supporting membrane. Besides supporting cells, there are present granular gland cells. In most larvae in which the first pair of filaments reach the aboral end of the body, the filament is no longer separated from the mesentery by supporting lamella. Fig. 19 gives a surface view of such a larva, and Fig. 26, a cross-section of one of the long filaments. In the latter figure the sheet of supporting lamella, on which the filament formerly rested, has given place to an accumulation of jelly at the apex of the mesentery. Though the cells of the filament are practically continuous round this gelatinous axis with the cells of the mesentery, the line of demarcation is very evident on each side owing to the different histological characters of the endoderm and ectoderm cells. In the filament there are now numbers of large nettle cells, and the gland cells are far more numerous than in earlier stages. Nervous elements are very probably present, as there is a granular stratum in the deepest part of the filament. As is shown in Fig. 19, the filament increases in size towards its lower end. Near the close of the swimming life the first pair of filaments begin to get slightly curved and twisted in the lower part of their course. The young filament before its cells become continuous with those of the mesentery, is very loosely attached to the latter ; in a number of cases I observed that the mesentery and filament had entirely separated from each other (Fig. 29, Mes. 3), owing, no doubt, to the contraction caused by the killing fluid.

Fig. 35 is through one of the first pair of filaments, and its mesentery, of an attached larva with eight mesenteries. In the attached larva these filaments pursue a straight course for a short distance below the oesophagus. In the lower part of their extent, they are curved and twisted as in the adult, and in this region sections like Fig. 34 are obtained. Fig. 35 is through the straight portion of the filament. On comparing this figure


with Fig. 26 (magnified to the same degree) it is seen that the mesentery has become more elevated, and at the same time thinner ; also, that where the endoderm cells end and the filament cells begin there has been a pinching in, which, added to the actual bulging out of the filament, has very distinctly worked off the latter from the mesentery. The bulging out of the filament is due to the lateral expansion of the axial jelly shown in Fig. 26. The expansion of the jelly has so taken place that the filament. Fig. 35, is divided into three portions : the main body to which are confined the nettle and gland cells ; and two tracts, v.L, which may be called the ventro-lateral tracts. The latter are composed exclusively of slender supporting cells. As in the swimming larva, the filament cells are sharply marked off from the cells of the mesentery. In spite of the twisted condition of the lower part of the filament, sections show that the structure is the same as in the upper part.

In the adult the upper portion of the filament on a complete mesentery is comparatively straight, but the main portion is twisted. Fig. 50. The filaments of the incomplete, are less twisted than those of the complete mesenteries. The filaments are attached their whole length to the mesenteries, there being no free acontium ; they are, however, capable of extrusion both through the mouth and (though I could not find the apertures) through pores (cynclides) in the body wall.

Unlike the larval filament, that of the adult has a different structure at different levels. Figs. 52, 53, and 54 are sections through the different parts of a filament on a complete mesentery. Fig. 52 is through the upper third. The filament itself has almost exactly the shape shown in Fig. 35, but the mesentery is swollen out and forms two lateral lobes, in. I., between which the filament rests. The ventro-lateral tracts are much better marked in the adult than in the larval filament. This is due to the continuation of the pinching in process which had already gone some distance in Fig. 35, and to the outgrowth of the mesenteric lobes, m.l. By these means the slender "waist" is produced, which indicates the separation of filament and mesentery.

The ventro-lateral tracts of the filament, both in the upper portion of its course. Fig. 52, and lower down. Figs. 53 and 54, are made up exclusively of supporting cells. The main body of

224 WILSON. [Vol. II.

the filament in Fig. 52 contains a large number of granular gland cells and numerous nettle cells. The mesenteric lobes are composed of cells which do not differ essentially from the rest of the endoderm ; they are only much elongated and contain a number of very large vacuoles. The passage of the filament into the oesophageal epithelium is effected in the following manner. Immediately below the oesophagus, the " waist " in Fig. 52 becomes gradually wider until the ventro-lateral tracts no longer exist, and the filament cells are continuous with the mesentery cells round the horns, h, of the supporting lamella of the filament. At this level, the filament or oesophageal lobe is much flatter and wider than it is below (Fig. 52), and the supporting lamella of the filament is also nearly flat. The latter passes directly into the supporting lamella of the oesophagus, and the filament into the lining epithelium of the oesophagus.

Fig. 53 is through the middle third of the filament. The gland cells are absent in this region, but the nettle cells are very large and exceedingly abundant. The mesenterial lobes are not so well developed as in the region above.

The lower third of the filament. Fig. 54, contains neither nettle cells nor the typical gland cells. The body of the filament is here made up of very large granular cells, between which are scattered a few supporting cells. The granules are much more numerous in the peripheral than in the central halves of the large cells, and at the sides of the filament where the main body passes into the ventro-lateral tracts, they become gradually restricted to the peripheral ends of the cells. The granules are chemically different from those in the ordinary granular gland cell : they do not stain especially well with haematoxylin, but become dark brown with osmic acid. In a number of filaments these peculiar granular cells contained large, irregular concretions, which stained dark red with borax carmine, while the cell body stained but faintly. The mesenteric lobes in this region are slightly less pronounced than in the rest of the filament. In a transverse section through the filament of Sagartia, Von H eider (10) has figured the lateral parts of the simple filament (according to the Hertwigs, the section is through the acontium) as composed of just such large granular cells as I have described. The Hertwigs do not speak of these tracts.


From Fowler's brief description (15, II.) it would seem that the filaments of Madrcpora are essentially like those of Manicina. The condition of the specimens, however, as the author states, did not permit a detailed study.

The filaments on the secondary are somewhat smaller than those on the primary mesenteries. The mesenteric lobes, too, are less pronounced, not reaching a development greater than is shown in Fig. 54. It has been mentioned that in the upper part of their course the secondary filaments become very small. Though the diminution in size is so great in this region, that it was impossible for me to make out the histological structure, I was able to trace the filament as a darkly staining and compact mass of tissue into the epithelium of the peristome. The filaments of the young coral, from which Figs. 41 and 42 were made, were like those of the adult, except that the mesenterial lobes were less developed.

I did not study the living filaments, but from the histology it is evident that the function of the ventro-lateral tracts is that of ciliated bands, while the digestive functions and nettle cells are distributed over the three portions of the main body of the filament. The mesenteric lobes I regard merely as a device to support the filament.

Before comparing the mesenteric filaments of Manicina with the actinian filament as described by the Hertwigs, I will give a brief account of the filaments of Ceriattthus, this being the only actinia I have been able to study by way of comparison. In Nassau harbor the larval or free-swimming Cerianthus was common. All the individuals I obtained were at about the same stage of development. They were oval light brown bodies about one-fourth inch long, and had eight or nine very short, stubby tentacles. On sectioning the larvae, I found they confirmed the theory advanced by the Hertwigs {I.e.) as to the mode of origin of the mesenteries. Fig. 49 is a section through the upper part of the oesophagus. The number of complete mesenteries is fifteen. At one end of the oesophagus is the ventral or directive pair of mesenteries, D.M. At the other or dorsal end are two incomplete mesenteries, i and 2, of which the former is the younger. The mesentery 3 becomes incomplete not far above the oesophageal lip, and 4 follows its example just above the lip. From these four mesenteries it may be gathered

226 WILSON. [Vol. II.

that each new mesentery appears along the dorsal meridian between the two last formed, and in such a way that the successive mesenteries come to lie on opposite sides of the dorsoventral plane. The directive pair, D.M., corresponds to the directive pair which the Hertwigs have figured (Taf. I., Fig. 8). These authors state, however, that this pair is the longest of all, extending to the anal pore. In the young specimens I have studied, this pair disappears some distance above the pore, and is exceeded in length by two or three, mesenteries on each side of the dorso-ventral plane. Von Heider (i i) made the discovery, which the Hertwigs confirmed, that in the adult there is a much shorter pair enclosed between the two directive mesenteries. This short pair only extends the length of the oesophagus. It must appear in stages later than Fig. 49, so that besides the formation of mesenteries which takes place dorsally, at least one pair is formed ventrally.

Filaments were found only on the complete mesenteries. Fig. 47 is from a section just below the oesophagus. The mesentery is composed of very large cells with immense vacuoles.

At the edge of the mesentery, just beneath the filament, the tissue is becoming more compact. This process has gone much farther in Fig. 48 (from the same mesentery at a lower level). The whole mesentery, as may be seen by comparing my figures with Hertwigs' Taf. VIII., gradually suffers this change as the animal becomes adult. In Fig. 47 the filament has a central body provided with gland and nettle cells, and two lateral ciliated tracts in, composed of supporting cells, which are gradually assuming the shape of distinct lobes owing to the subjacent growth of the jelly. The central body passes directly into the epithelium of the oesophagus, the lateral tracts becoming indistinguishable from the body of the filament just before the transition. This figure is but slightly different from the section of the adult filament given by the Hertwigs. At a lower level. Fig. 48, the ciliated tracts are no longer found. On many mesenteries the filament has almost no connection with the endoderm, as in the figure. In such figures as Fig. 48, it need hardly be said that violent contraction has had much to do with producing such complete separation of filament and mesentery; but this only shows how loose the connection is in early life. On other mesenteries the filament is continuous at its sides


with the cells of the mesentery. None of the filaments in these young specimens of Ccrianthus ran the whole length of the mesentery. Most of them stopped at about the equator of the body.

In the actinias, studied by the Hertwigs, the mesenterial filament in the upper part of its course is in section tri-lobed. There is a median secretory tract and two lateral ciliated tracts. The mesoderm extends into each lobe. On all the mesenteries the ciliated tracts are lost in the lower portion of the filament. On the incomplete mesenteries the median tract disappears towards the upper limit of the mesentery, the lateral ciliated bands remaining. On the complete mesenteries the median tract merely gets smaller towards the oesophageal end, but does not disappear.

The description I have given of the filaments of Ccrianthus is, barring slight differences due to age, like the account the Hertw^igs give for this actinia. From this description it is seen that the filament of CcriantJms corresponds in the main to a typical Hertwig filament, such as is on a complete mesentery. But the lateral lobes, ;;/, Fig. 47, are very small even in the adult when compared with the lateral wings of the Sagartia filament (H., Taf. V., Fig. 10).

The filament of Mauicina is a much simpler structure. It is not trifid, though the mesenteric lobes, ni.L, Fig. 52, give such an appearance to the edge of the mesentery. But on referring to Hertwig's figure of Sagartia (Taf. V., Fig. 10), it is evident that the mesenterial lobes of Mauicina do not represent the ciliated bands of Sagartia. They are totally different histologically, and do not contain prolongations of the mesoderm. On the other hand, the ventro-lateral tracts of Majiicina do correspond histologically with the ciliated bands in the Hertwig filament, and the glandular lobe of the latter is in general similar to the median stripe of the Manicina filament. As I take it, all three tracts of the Hertwig filament, taken together, are homologous with the simple filament of Manicina. They must all, therefore, be ectodermal : in the young Ccrianthus, Fig. 47, the lateral lobes evidently belong to the median lobe, and are not modified parts of the mesentery ; and since the median lobe at a lower level. Fig. 48, shows its independence of the mesentery, the conclusion seems to be that the whole fila

228 WILSON. [Vol. II.

ment is ectodermal. Though the filament of the adult coral is a much simpler form than the trifid filament, it would be difficult to derive the latter from the former, owing to the presence of mesenteric lobes in the coral. It is easy, however, to derive the trifid from the larval filament. Fig. 35. In this form the ciliated tracts, v.L^ and the median secretory tract, are already differentiated ; to produce the trifid filament it is only necessary for these tracts to become separated by the division of the mesoderm into three lobes. By this division, while the median stripe gains but little, the ventro-lateral tracts are put in a position where it is possible for them to reach a high development.

The number of Actinozoa in which the mesenterial filaments have been carefully studied is very limited, but from the data at hand it seems probable that the ancestral filament, like the filament of the larval Manicina, was a simple undivided body, in which, however, the originally uniform ectoderm had become split up into three physiologically distinct tracts : a median tract, in which were concentrated the nettle and glandular elements, and which embraced most of the filament ; and two lateral ciliated portions which were but slightly developed. The belief that the division of labor in the Zoanthorian filament dates very far back, is supported by the existence of a similar division of labor in the Alcyonaria. But in these polyps, as Wilson (8) has shown, the functions are not distributed over different parts of the same filament, but amongst the several filaments. The dorsal pair are ciliated bands, having no gland or nettle cells, while the remaining six filaments contain gland and nettle cells, but have no ciliated tracts.

IX. The Skeleton.

Until the time of Von Koch's researches the skeleton of the Madreporaria was regarded as calcified mesoderm. The theca, or coral wall, according to this view, represented the supporting lamella of the lateral body wall of an actinia ; and where, as in most corals, the theca was largely uncovered by animal substance, the explanation was that the ectoderm had atrophied. In 1879 Von Koch (12) showed that the theca is independent of the lateral body-wall, and projects into the coelenteron of the


polyp in such a way as to divide the cavity and mesenteries into central and peripheral parts. This conclusion, though not universally accepted, is, I think, now incontestable. Von Heider in 1882 (14) published the important observation that the skeleton is not only covered by supporting lamella and endoderm, but that between the calcareous matter and the supporting lamella is another layer of cells, to which was given the name of calycoblasts. Von Koch in 1883 {13) made the whole matter clear by showing that in the young Asteroides the skeleton is secreted by the ectoderm and is at first entirely outside the body. The ectoderm of the basal surface of the attached larva secretes a calcareous basal plate. Radial folds of the basal ectoderm then grow up between the mesenteries into the cavity of the polyp, lifting up, as they grow, the supporting lamella and endoderm. The cavity of the fold is filled with calcareous matter, which is the septum. The secreting ectoderm becomes the calycoblast layer of Von Heider. The origin of the basal plate and septa is thus clear enough, but Von Koch's account of the development of the theca is not satisfactory. As far as his observations went they appear to have confirmed his belief, gathered from a study of adult corals, that the theca is formed secondarily from the septa ; the originally simple septa become bifurcate at their peripheral ends ; the lateral processes of adjacent septa grow towards each other, pierce the mesentery, and, fusing, form the theca. Von Koch's account of the relation of the skeleton to the soft parts in the adult has recently been confirmed on a number of genera by Fowler (15) and Bourne (16), though as regards the extra-thecal part of the coelenteron Fowler differs from Von Koch in the interpretation of his sections. From these papers it is also evident that the Madreporaria exhibit great variety in the details of the relation between the soft parts and skeleton. My own very incomplete observations on the skeleton of Manicina are for the most part a confirmation of Von Koch's statements.

In the newly attached larva. Fig. 37, the ectoderm of the basal surface is made up exclusively of supporting cells. The granular cells present at the aboral end of the swimming larva have all disappeared. Some time after attachment a small patch of calcareous matter is found on the basal surface, and sections

230 WILSON. [Vol. II.

through such larvse, after the lime has been removed by acid, give figures Hke Fig. 38, though in many individuals the ectoderm was torn in freeing the larva from the bottom of the dish. In Fig. 38 the cells of the basal ectoderm radiate toward a common centre. There seems to be an effort on the part of the more peripheral cells to share in the secretion of the central patch of homogeneous substance shown in the figure. This homogeneous substance is the animal basis of the nascent basal plate. It exhibits no structure, but stams deeply with hagmatoxylin, and in its general appearance impresses one as a very much thickened cuticle. I was not able to trace the development of the skeleton any further.

The fully formed skeleton may very conveniently be studied in young polyps, which have not begun to multiply asexually. Ground sections of the skeleton are of some use, but with a little care the polyps may be decalcified so that the skeletal layer of tissue retains with great exactness the shape of the skeleton. Figs. 46, 50, and 56 are from young Maiiicinas about one-third inch in diameter. Fig. 46 is part of a transverse section through the line a in Fig. 50 {pe. ec. is the ectoderm of the peristome ; ec. that of the lateral body wall). Were the section complete, it would show twelve primary, twelve secondary, and twenty-four tertiary pairs of mesenteries. At this level the secondary mesenteries are complete ; a little lower down they are incomplete.

In going down through a series of sections, the tertiaries run out before the lip of the oesophagus is reached. In Fig. 50 the line 3 marks the free edge of a tertiary mesentery. The theca, TJi., divides the mesenteries and coelenteron into peripheral and central portions. The septa, S., are all entocoelic ; i.e., lie between the two mesenteries of one pair, and not between two adjacent pairs. The size of the septa varies with the rank of the mesenteries between which they lie. On the outer surface of the theca are the longitudinal ridges, C, or costae, which appear to be merely the peripheral prolongations of the septa. The edges of the septa are finely and regularly toothed. This is shown in the radial section, Fig. 56, which is taken through one of the coral septa. In Figs. 50 and 56 the relations of the columella. Col., to the theca and septa are shown. In the former figure, through a mesentery on each side and consequently between two septa, a deep depression separates the


columella from the theca. The primary mesenteries extend to the floor of the depression ; the secondaries end some distance above it. The columella is circular in section, and the depression surrounds it as a trench. But the trench is not continuous, being completely divided in the radius of each primary septum, as is gathered from Fig. 56, which is taken through such a septum. In this figure the septum and columella are directly continuous with each other. The columella in reality is not solid, as it is drawn in the figure, but is spongy, being full of portions of the body cavity which it has cut off during its growth. Whether an originally simple and solid columella is formed as a central elevation of the basal plate, is an open question. But the subsequent growth of the columella takes place by the constant incorporation in it of the lower portions of the inner edges of the primary septa. The teeth on the edge of the septa, and the granulations found on their sides, sufficiently explain the spongy nature of the columella.

The skeleton inside the polyp is everywhere covered by the three layers of the body wall, of which the skeletogenous ectoderm or calycoblast layer is next the skeleton. In Fig. 46 and the radial sections these layers are not represented, but they are shown in Fig. 51, a more highly magnified portion of one of the peripheral entocoelic chambers of Fig. 46. The superficial ectoderm is marked cx.\ the peripheral parts of the mesenteries mes. The calycoblast layer covering the costa, or, more exactly, the costal tooth, is as elsewhere a layer of flattened cells. The skeletal endoderm is likewise very flat, and markedly different from the endoderm of the body wall and mesenteries. In radial sections the calycoblast layer is found to be continuous with the superficial ectoderm round the edge of the extra-thecal part of the polyp, or randplatte, R.P. in Fig. 50.

That portion of the polyp, R.P., that lies outside the theca, has been called by Von Heider the rmidplatte. It has been claimed (Moseley, 18; Fowler, 15) that this part of the polyp ought not to be regarded as normally on the outer surface of the skeleton ; but that when the expanded polyp, which extends high above the skeleton, contracts, while most of the body is drawn into the interior of the theca, a portion is pulled down over the outer surface of the skeleton. However plausible this belief may be in the case of those adults in which the extra

232 WILSON, [Vol. II.

thecal part of the polyp is confined to the upper edge of the theca, it is at once found to be untenable on examining a small number of young Manicinac of the size of Fig. 50. In the polyp, for instance, from which this section was made, the randplatte covered more than one-half of the lateral surface of the skeleton, and this was true both in the expanded and contracted condition of the animal. Further, in every dozen such young Manicinas, one or two asymmetrical ones will be found, in which, while the randplatte is confined to the upper half of the lateral surface on one side of the polyp, on the other side it covers the whole wall down to the surface of attachment. What puts the matter beyond dispute, in this genus at least, is that in the very young Ma7iicma, Figs. 41 and 42, the whole skeleton is practically inside the polyp. I also found two or three older specimens (single polyps) in which the entire lateral surface of the skeleton was covered by the randplatte. One of these which I sectioned, was oval in transverse section, long axis about | in.; short axis \ in. The skeleton had a flat surface of attachment, and was -1^ in. in height. The randplatte (on rather extra-thecal polyp) in this individual was unbroken, down to the piece of rock to which the coral was fastened; here its ectoderm turned in to form the calycoblast layer.

It appears, then, that up to a certain age, which varies much in individuals, the lateral surface of the skeleton is entirely covered by the polyp. The transformation of the originally large extra-thecal part of the polyp into the relatively small rand platte may possibly take place in many cases gradually, by the constant dying off of this part of the polyp at its free edge, and subsequent disappearance of the dead tissue. But in some instances this is not the case. I found a number of small single polyps in which I first thought the randplatte covered the entire lateral wall of the skeleton, but on looking again, I saw that a very definite line extended all round the lateral surface at about the level of the edge of the randplatte in Fig. 50. Above this line the soft parts looked perfectly healthy, but below it dull and sickly. On sectioning I found that above the line the randplatte was normal, but that the tissue below it was a membrane with scarcely a trace of cellular structure and made up for the most part of plant filaments. In some asymmetrical specimens the portion of the skeleton left uncovered by


the randplatte was covered by a precisely similar membrane, which was marked off from the randplatte by a deep furrow. In all such specimens the ectoderm of the randplatte was continuous at the edge of the latter with the calycoblast layer. These peculiar membranes gave every indication of having been origi nally continuations of the randplatte ; and I conclude that in the individuals possessing them, the extra-thecal part of the polyp had remained intact up to a certain time, but that then the whole lower portion of this part of the animal was cut off from the general gastric cavity. Deprived of its nutriment, this portion became membranous, and was infested by plant filaments. As the coral increases in size, the membranes are shed, leaving a large part of the theca bare. After the randplatte once becomes restricted to the upper part of the skeleton, there is no more shedding of large pieces of tissue. But as the skeleton is constantly growing in height, we have to suppose that the randplatte as constantly dies at its free edge, unless, indeed, we assume that the connection between the calycoblast layer and the skeleton is so slight that the randplatte is merely carried up with the growing skeleton.

In the young Maniciiia, Figs. 41, 42 (i in. diam.), the skeleton is very immature, though the various parts of the adult skeleton can all be recognized. In this specimen there was a thin, flat, basal plate, uncovered at its periphery by the body wall. The skeleton above the plate was internal except at the points a and a', where the lower edge of the extra-thecal part of the polyp was notched, so that at these points the skeleton was bare for a short distance above the basal plate. Fig. 41 is through the oesophagus, Fig. 42 is below, and strikes the apex of the columella. Though some of the septa are still independent at as low a level as Fig. 42, at a still lower level they all unite to form a theca, which on the right side of the directive mesenteries is very slightly developed, but on the left side is prominent. Two or three of the septa in the region where the theca is so slightly developed, exhibit the bifurcation of their peripheral ends which Lacaze Duthiers and Von Koch have described. In Fig. 42 indications of six primary septa are more or less evident, and in lower sections where the septa fuse with the columella they are easily distinguished. There are also six secondary septa to be made out in Fig. 42. These twelve septa,

234 WILSON. [Vol. II.

which are entocoelic, extend as high and higher than Fig. 41. Further, in Fig. 42 it is evident that twelve tertiary septa have started to develop, though only a few of them extend an appreciable distance above the theca. As only twelve pairs of mesenteries have been formed, the tertiary septa are temporarily exocoelic. Transverse and longitudinal sections through such a polyp as this show very clearly that the growth of the theca is continuous and from below upwards. Nowhere are two septa found which actually pierce a mesentery to form the theca, but everywhere the growing theca pushes the tissues of the mesenteries upwards. Where the theca is formed in such a manner as this, it is out of the question to believe, after Von Koch, that the radial cracks found in ground sections of the theca are due to the existence in these radii of the atrophied remains of the lower portions of the mesenteries.

The costae in Figs. 41 and 42 are very feebly and irregularly developed. Each septum has, however, three or four teeth along its outer edge, and also a few on its inner edge. All the teeth on the outer edge of the septa open directly to the exterior over the side wall of the coral. In Fig. 41 the septum included between the near pair of directive mesenteries is cut just below the point at which it thus opens. Fig. 40 gives a more highly magnified view of the same septum cut at the level of its opening. In Fig. 42 a tooth has opened at m, and another at 71 is sectioned just beyond its opening. The apertures are of good size, and through them the external ectoderm is continuous with the calycoblast layer. This peculiarity of the teeth is no longer found in Manicina after asexual multiplication has begun, and only a few teeth open to the exterior in single polyps with the full complement of septa. In one such polyp, however, all the teeth opened in this way. From this it would seem that the connection of the calycoblast layer with the surface ectoderm at these points is a characteristic of youth which is gradually lost. It is possible that this peculiarity may have been acquired in order to bind the little polyp and its simple skeleton firmly together.

X. Origin of the Anthozoa.

As is well known, the reigning view as to the origin of the Anthozoa is that advanced by Claus and Haeckel, according to



which the Anthozoa are descended from hydropolyps with gastric ridges. This hypothesis, which considers the distinguishing feature of the Anthozoa (and Scyphomedusae) to be the possession of taenioles, or endodermal ridges (mesenteric ridges), has recently been attacked by Professor Gotte in a very interesting paper (17) on the development of the Scyphomedusae (Aurelia). The author conclusively proves the Anthozoan nature of the Scyphostoma larva, showing it to possess four complete mesenteries and an ectodermal oesophagus. Unfortunately, no observations were made on the origin of the mesenterial filaments, which in all probability are ectodermal lobes. Regarding, then, the Scyphomedusae as an offshoot from the older Anthozoan

T7ci. /^


A^, 2

Ectoderm represented as made of columnar cells. Endoderm is dotted, ing lamella left untouched. After Gotte, PI. II., Figs. 20 and 24.


stem, Gotte argues that the manner in which the mesenteries and intermesenterial chambers are formed in the larva of the former, is directly opposed to the old view of the hydroid origin of the Anthozoa. The mesenteries and chambers are formed in the following manner : the planula is bilateral, and the permanent layers are formed before the invagination of the oesophagus takes place. The latter is invaginated in such a shape that in the shorter transverse axis of the larva the endoderm is pushed down, while in the other transverse axis two endoderm sacs are formed. The woodcut. Fig. i, is a transverse section of this stage, through the oesophagus, and Fig. 2 is a longitudinal section through the shorter transverse axis. In a subsequent stage.

236 WILSON. [Vol. II.

two endoderm sacs grow up from the coelenteron in the shorter axis, and the oesophagus is then surrounded by four sacs. The sacs are the intermesenteric chambers, and their partition walls form the mesenteries. Gotte concludes that in this development there is no stage which corresponds to the hypothetical hydropolyp ancestor. The young Scyphostoma itself, which has hitherto been regarded as a hydropolyp with taenioles, is in reality an Anthozoan with four complete mesenteries ; and the development previous to the Scyphostoma does not pass through a hydroid stage, but, on the contrary, jumps directly from a hollow planula to the larva provided with two (later four) endodermal sacs. This larva is called the Scyphula, and from Kowalevsky s account of the development of Cerianthiis (4), Gotte believes it to be common to both the Scyphomedusae and Anthozoa, and consequently an ancestral form. In Cerianthiis according to the abstract given by Hoffman and Schwalbe of Kowalevsky's paper, the oesophageal invagination does push down the endoderm along two opposite meridians ; but though the abstract is not definite on this point, the implication is that the meridians are those of future mesenteries. If this is so, Cerianthus agrees essentially with Manicina, and not with Aurelia.

The ancestral Scyphula form was derived, according to Gotte, directly from the hollow planula, the invagination of the oesophagus necessitating the simultaneous formation of endoderm sacs. The infra-oesophageal mesenteric ridges, from this point of view, are not of any phylogenetic importance, and have nothing to do with the endoderm ridges of Tubularian or Siphonophore polyps ; they have come into existence merely as the after-result of the formation of endodermal sacs. This theory contains in itself an obvious difficulty : the sudden and direct transformation of such a simple form as the planula into such a complex form as the Scyphula. What could have caused this complex group of changes, Gotte does not suggest. But aside from this objection it seems clear that the development of Aurelia is a highly modified form of the development of Manicina, and the manner in which this peculiarly symmetrical modification was brought about is suggested by the variations shown in Figs.

30-33 If Fig. 30 is compared with the woodcut Fig. i, it is seen that


as far as the left halves of the sections are concerned, they are identical in all essential respects. In each the oesophagus is opposed to the surface ectoderm along -the meridians of two adjacent mesenteries ((7 and b, i and 3), and also over the intervening tract. In a later stage an endoderm lobe grows up between a and b in Fig. 30, and between i and 3 in the woodcut, and in each this lobe becomes an intermesenteric chamber. If, now, in the larva. Figs. 30-33, the right side had followed the example of the left, that is, if the second and fourth mesenteries had been formed in the same manner as the first and third {a and h), there would have resulted an exact counterpart of the condition in the Aiirelia larva.

The four mesenteries of Scyphostoma would thus seem to correspond to the first and second pairs of mesenteries in Manicina. We may suppose that in the primitive Scyphostoma the mesenteries were usually formed in the gradual way which is normal in Manicina, but that the Scyphostoma had inherited from the parent stock (probably Anthozoa with a large number of mesenteries) a tendency towards the variation illustrated in Figs. 30-33, and that this variation gained ground and finally became the normal process. There is, of course, an alternative to this hypothesis, namely, to regard the variation found in Manicina as a case of partial reversion to the ancestral condition as presented by Aurelia. But the derivation of the Anthozoa from such an ancestor as the young Scyphostoma (or Scyphula with four endoderm lobes) is beset with the greatest difficulties ; for instance, the formation of the first before the second mesentery, and the very general occurrence of a primary tentacle in actinia larvae and in the Scyphostoma itself. Moreover, the Scyphula larva has not been found in any Anthozoa, unless, indeed, the case of CeriantJms be really, as Professor Gotte seems to have considered it, similar to Aurelia.

Having shown that it is possible to derive the so-called Scyphula larva from the larva of Manicina, and that it is consequently in all probability an instance of secondarily acquired symmetry, I consider Gotte's objection, based on the existence of this larva, to the hydropolyp ancestry of the Anthozoa, as no longer valid. The question whether or not the Anthozoa are descended from hydropolyps must be argued out on the ground of some more primitive Anthozoan development, such as that

238 WILSON. [Vol. II.

of Manicina. And here it is at once seen that, contrary to Gotte's idea, the invagination of the oesophagus does not necessitate the formation of endodermal sacs. In Figs. 5 and 7 the oesophagus is already formed, but is still surrounded on all sides by endoderm. The apposition of the oesophagus to the surface ectoderm along the lines of the first and second mesenteries takes place later; and though, since this process occurs in the Scyphomedusae as well as the Zoantharia, it must date very far back. I am inclined to believe it was secondarily acquired and was not a peculiarity of the primitive Anthozoa. This belief is supported by the entire absence of the process in the Alcyonaria (Wilson, 3). The explanation of the process is possibly connected with the early development of the first pair of filaments.

Baltimore, March 25, 1888.



1. Lacaze Duthiers. Developpement des Coralliaires. Arch, de Z06I. Exp. T., II.

1873 2. Metschnikoff. Embr. Studien an Medusen. Wien, 1886.

3. Wilson, E. B. Development of Renilla. Phil. Trans., 1883.

4. Kowalevsky. Unters. ii. Entwick. d. Coelenteraten. Nachrichten d. kais. Gesell.

d. Freunde, etc., Moskau. Vol. X. (Russian.) Abstract in Hoffman & Schwalbe, 1873.

5. Jourdan. Zoanthaires du Golfe de Marseille. Ann. des Sc. Nat. T., X. 1879 1880.

6. Glaus. Unters. ii. Charybdea marsupialis. Arb. aus dem Zool. Inst. Wien, 1878.

7. Wilson, H. V. Structure of Cunoctantha oct. in Adult and Larval Stages.

Studies fr. Biol. Lab. Johns Hopkins Univ. 1887.

8. Wilson, E. B. Mesenterial Filaments of the Alcyonaria. Mitt. Zool. Sta.

Neapel. Bd. V. 1884.

9. Hertwig. Die Actinien. Jena, 1879.

10. Von Heider. Sagartia troglodytes. Sitzungsber. d. kais. Akad. zu Wien. Bd.

75, Abth. I. 1877.

11. Von Heider. Cerianthus mem. Sitz. kais. Akad. Wien. Bd. 79, Abth. i. 1879.

12. Von Koch. Bemerk. U. das Skelet d. Korallen. Morph. Jahrb. Bd. V. 1879.

13. Von Koch. Entwick. d. Kalkskelettes von Astroides calycularis. Mitt. Zool.

Sta. Neapel. Bd. 3. 1882.

14. Von Heider. Die Gattung Cladocora. Sitz. kais. Akad. Wien, 1882.

15. Fowler. Anatomy of the Madreporaria, I., II., HI. Quar. Jour. Micr. Sc.

1885, 1886, 1887.

16. Bourne. Anatomy of Fungia. Anatomy of Mussa and Euphyllia. Q. J. Micr.

Sc. 1887.

17. Gotte. Entwick. d. Aurelia aurita und Cotylorhiza tuberculata (Abh. z. Entw.

d. Th., IV.). Hamburg and Leipzig, 1887.

18. Mosely. Remarks upon Some Specimens of Sections of Coral. Proc. Zool. Soc.

London, 1880.



In the larvse, the supporting membrane is represented as a heavy black line; the jelly, where present, as light brown. In the adult, and in the complete section of Ceriantkus, the supporting substance is uncolored. All the sections were drawn with the camera. Zeiss lenses are referred to, and the figures are reduced to one-half the size of the original drawings.

The following letters have been used uniformly in the figures : —


Calycoblast Layer.










Randplatte or Extra-thecal Part of


Directive Mesenteries.

















Ventro-lateral Ciliated Tracts.


Mesenterial Lobes.



Fig. a. Corallum of a Manicina colony, viewed from above. Natural size.

Fig. B. Side view of same, to show the pedicel. The randplatte extends but a short distance below the rim of the theca; its lower Hmit is marked by a slight accumulation of irregular laminae.

Journ. Morph.^ Vol. II.








Fig. I. Section through a blastosphere. C, 4.

Fig. 2. Section through delaminating blastosphere. C, 4.

Fig. 3. Part of another blastosphere such as Fig. 2. D, 4.

Fig. 4. Median long. sec. through a solid planula with commencing oesophagus.

C.4 Fig. 5. Long. sec. through a larva in which the permanent layers are forming.


Fig. 6. Long. sec. of slightly older stage. Position of oesophagus is eccentric. C, 4.

Fig. 7. Trans, sec, through cesophagus, of larva at about the stage of Fig. 5. The formation of the permanent layers is not so far advanced as in the latter figure. D, 4.

Fig. 8. Long. sec. through a larva in which the oesophagus has opened centrally. D, 4.

Fig. 9. Trans, sec. through a larva in which the oesophagus is still more eccentric than in Fig. 8. The section is through hne a in Fig. 8. D, 4.

Fig. 10. Trans, sec. of same larva at lower level, through line b in Fig. 8. D, 4.

foiirn. Mnrph.^ Vol. II.

H. K. WiUcn^ n^t^

A. Hocn tr Co. L!th., Balu.



Fig. II. Trans, sec, through lowest part of oesophagus, of a larva in which the oesophagus along its whole length is apposed to the surface ectoderm. D, 4.

Fig. 12. Long. sec. of larva at about the same stage as Fig. 11 in the plane of the first two mesenteries. D, 4.

Fig. 12'. Median long. sec. of a larva, in which the oesophagus is applied to the surface ectoderm along the lines of the first and second mesenteries. An older stage than Fig. 12. D, 4.

Fig. 13. A section to one side of Fig. 12, through future intermesenteric chambers. D, 4.

Figs. 14, 15, 16, 17. Series of transverse sections, numbered from above down (Fig. 14 is through the oesophagus), from a larva in which the first pair of filaments and the first mesenteric ridge are present. Along the line of mesentery 2, the oesophagus is still applied to the surface ectoderm. D, 4.

Figs. 18, \%a. Trans, sec, through oesophagus, of a larva with two complete mesenteries and a pair of filaments. Fig. 18 « is the upper of the two. D, 4.

Fig. 19. Surface view of balsam preparation. The larva has two long, and one (possibly two) short filaments. C, 2.

Journ. Morpii.^ t "(■ jI.


'ViUon, Dtl.

A. HinK &• Co. Lith., Baiti:



Figs. 20, 21, 22, 23. Series of long. sec. from a stage somewhat younger than Fig. 18. Fig. 20 is through the first pair of mesenteries and filaments; Fig. 23 is through the hne a, b in Fig. 18; and the other sections fall between. C, 4.

Fig. 24. Long. sec. through a larva such as Fig. 19; on the right, through one of second pair of filaments, on the left through an intermesenteric chamber. D, 4.

Fig. 25. From a larva with two long and two very short filaments, a, b, c, are three half-sections from the same side of the larva, b is through one of second pair of mesenteries with its filament; a and b are on opposite sides of this mesentery. D, 4.

Fig. 26. Trans, sec. through one of primary pair of filaments and mesenteries of a stage like Fig. 19. F, 4.

Figs. 27, 28, 29. Series of trans, sec. from a larva like Fig. 19. Fig. 27 is through the cesophagus, 28 is just above the lip, and 29 is below the oesophagus. The sections are very slightly oblique. The third pair of mesenteries exist as longitudinal ridges of the supporting lamella. Ji.£. is the reflected ectoderm. D, 4.

Joum. Morph., Vol. //•


"■ . WUton. Del.

A. Horn b' Co. Litk., SaJlo



Figs. 30, 31, 32, 33. Series of trans, sec, numbered from above down, from a larva with two pretty long (half the length of larva) and one very short filament. The first and third filaments have been formed from a common lobe, and the third mesentery {F) has been formed at the same time and in the same way as the first (a).


Fig. 34. Trans, sec, below oesophagus, of attached larva with eight mesenteries.


Fig. 35. Trans, sec, through upper part of one of the primary filaments of an attached larva. The large clear cells are nettle cells. F, 4.

Fig. 36. Trans, sec, through oesophagus, of attached larva with eight mesenteries. The second pair of mesenteries is not quite complete, and consequently at this level (just above the lip) the reflected ectoderm is found all round the oesophagus. B, 4.

Fig. 37. Long. sec. through attached larva; on the right through one of second pair of mesenteries, on the left through an intermesenteric chamber. C, 2.

Fig. 38. Long. sec. through attached larva in which the basal plate has appeared; on the left through one of third pair of mesenteries with its filament, on the right through an intermesenteric chamber. The line m, 771 indicates the outline of the polyp when expanded. C, 2.

Fig. 39. Trans, sec, through oesophagus, of attached larva with twelve mesenteries. B, 4.

Fig, 40. A more highly magnified view of one of the septa of Fig. 41 (that lying in the near "directive " chamber), sectioned at the level of its opening, c.b. is the calycoblast layer.

i Journ. Morph.^ Vol. II.

A. Horn &■ Co. Lillt.. Ba

2 so WILSON.


Fig. 41. Trans, sec. through oesophagus of young Manicina, \ in. diam. D.M., the directive mesenteries. X, 60.

Fig. 42. Section of same specimen, below oesophagus. Skeleton is bare at a and a' . The tip of the columella lies in the coelenteron. X, 68.

Figs. 43, 44, 45. Series of trans, sec, numbered from below upwards, through the oval cone of a larva like Fig. 38. Fig. 43 is through the line x, y in Fig. t^. Only the first two pairs of mesenteries extend into the uppermost part of the coelenteron. The overlapping of the reflected ectoderm and endoderm is well shown. D, 4.

Fig. 46. Part of a trans, sec. of adult single polyp, through line a in Fig. 50. The ectoderm of the peristome is marked ; that of the lateral body wall, ec. X, 30.

Fig. 47. Trans, sec. of filament and mesentery of a larval Cerianthtis. Level of section is just below the oesophagus, m is the commencing ciliated band (flimmerstreif). F, 2.

Fig. 48. Section of same filament lower down. The violent contraction has caused the halves of the mesentery to spring apart, leaving the coagulated supporting lamella partially free. F, 2.

Fig. 49. Trans, sec, through cesophagus, of a larval Cerianthus. D.M., the (larval) directive mesenteries. The numbers i, 2, 3, 4, mark the four youngest mesenteries, in the order of their age, i being the youngest. X, 60.

Journ. Morph., Vol. II.


II. K Wilson, Dil.

A. Hotn 6- Co. Lilh., Hallo



Fig. 50. Median long. sec. through an adult single polyp; on the right through a primary, on the left through a secondary mesentery. The line 3 marks the position of the free edge of a tertiary mesentery. The surface of attachment was irregular, but the corallum extended only a very short distance below the limit of the skeleton in the figure. X, 30.

Fig. 51. One of the extra- thecal entoccelic chambers of Fig. 49 more highly magnified. Mes. is the peripheral portion of a mesentery.

Figs. 52, 53, 54. Trans, sec. of an adult primary filament. Fig. 52 is through the upper, 53 through the middle, and 54 through the lower portion. The large clear cells in Fig. 53 are nettle cells. D, 4.

Fig. 55. A more highly magnified figure of the lower portion of the oesophagus, as shown in the left half of Fig. 50. is the lining epithelium of the cesophagus.

Fig. 56. One-half of a median long. sec. of an adult single polyp. The section is through one of the primary septa. X, 30.

Journ. Morph., Vol. II.


Fig. 53

Fig. 55

Fig. 54

H. V. Wilton, Dil.

A. Hon &■ Co. Lith., li,iltii.





To Students of the fossil Arthropoda it should be a matter of congratulation that so great success has been achieved by palaeontologists in solving the problem of anatomy and development in the extinct crustacean order, the Trilobita. Though much may remain to be done, too great esteem cannot be accorded to those who have contributed to what has been accomplished ; Eichwald, Burmeister, Volborth, Quenstedt, Richter,^ Barrande, Hall, Billings, Ford, Walcott, Woodward, Mickleborough, Matthew, Packard. To the labors of these men is due our knowledge of the development of the trilobite from the ovum to maturity ; of its delicate locomotive and respiratory apparatus ; and somewhat of its reproductive, alimentary, and muscular anatomy.

The present paper endeavors to throw some light upon the structure and development of the eye in a typical representative of an extensive group of trilobites, Phacops rana. Green. In the study of this organ abundant material consisting of several thousand specimens has been accessible, in the majority of instances only those being utilizable in which the lenses of the eye

^ It may be well to call attention to the fact that Richter's single observation upon the ventral anatomy of Phacops, which has been overlooked by later investigators, is of much greater significance than the author himself accorded it. In the Beitragzur Pal'dontologie des Thiiringer Waldes, 1848, PI. 2, Fig. 32, is given an enlarged view

of a transverse section through one-half the thorax of a De- ^ , ^^___^

vonian Phacops (species not given). As this work may not ^ — T^ /^^^ be generally accessible, the figure is here reproduced; and ^^..^JUt/^ although the author states {op. cit. p. 20) that the section

serves to establish Burmeister's conception of the ventral anatomy of the trilobite, in the light of Walcott's demonstration of the spiral branchise in Calyinene senaria it appears that what is represented here is a section of one of these appendages. I may add that I have also detected evidence of these spiral branchise in Phacops rana.

254 CLARKE. [Vol. II.

have been so perfectly retained as to allow of enumeration. These specimens have been derived from the shales and limestones of the Hamilton group at various localities in Western New York, those best adapted for the purpose of sectioning being from the basal limestones near Centerfield, Ontario County.

The character of the visual area in the trilobites is twofold ; (a) it may be covered by a smooth, continuous epithelial film or cornea, through which the lenses of the ommatidia are visible by translucence, and ib) the cornea may be transected by the protrusion of the sclera ^ and limited to the surfaces of the ommatidia. To the first group belong species of the genera AsapJius, Illcenus, Calymene^ Ho^nalonotus, Proetus, Cyphaspis, Acidaspis, Lichas, and others ; to the second, the single extensive family, Phacopidce, with its genera, Phacops and Dalmanites {} Harpes ; vide conclusion). The first group may be designated by the term Holochroal ; the second group by the term Schizochroal,

Phacops rana is one of the most abundant and characteristic species of the Hamilton faunas. Though widely distributed in the formations of this age throughout the United States and Canada, it is not known with certainty to have been present in faunas older than those of the Hamilton, and it does not appear to have continued its existence after the displacement of the Hamilton faunas. A detailed and very complete description of the species, accompanied by copious illustration, is given in the

1 1 have found the term otrwiatidium, proposed by Carriere for the little eyes or ocular elements in the compound eyes of Arthropods and Mollusks, a very convenient and significant term, but may use it with a little license, as I do not regard the eyes included in the second of the above groups as properly compound. The term sclera as here used may be open to some objection. It is applied to the interstitial test between the ommatidia, and is preferable to the expression cornee opaque of Barrande.

2 Professor Edward Orton, of Columbus, Ohio, has allowed me to examine a very young individual of Calymene senaria in which the lenses are relatively very large, and are strongly suggestive of the character of the lenses in Phacops, although in the adult of this species they are so small as scarcely to be detected. This specimen suggests the query whether in the holochroal eyes the lenses may, with the advancing growth of the animal, become apparently smaller, from close juxtaposition or other cause, and also indicates the possibility that the difference in the holo- and schizochroal eyes is not as great as it now appears.


monograph of North American Devonian Crustacea, constituting Vohime VII. of the Palaeontology of New York. In this place are also to be found satisfactory figures of the eye in various modes of preservation.

Composition of the Visual Node.

The eye is composed of the visual surface, which is' normally a lunate segment of the surface of a cone, but often in senile individuals is inclined to sphericity ; this surface is buttressed on the glabellar side by a strong palpebral lobe, which is produced to and slightly beyond the upper edge of the visual surface, forming a ^x'sAXxvqX palpebnnn. The lower edge of the visual surface is bordered by a ridge, which becomes broader and more conspicuous outwardly, and may be called the orbital ridge.

TJic lejises as seen from the upper surface are convex, sometimes being translucent, especially when they have been filled with crystalline calcite, or have been slightly separated from the matrix ; they are circular in outline, although the cavities in which they lie {lensal pits) appear in old individuals to be hexagonal. This appearance is due to the undiminished growth beyond maturity of the sclera, which crowds upon and overlaps the edges of the lenses on all sides, deepening the lensal pits. The general arrangement of the lenses is in alternating vertical rows or quincunx. For convenience, however, in the accompanying statements of enumeration, the lenses are regarded as arranged in diagonal rows parallel to the lower posterior margin of the visual surface, and are numbered from this line consecutively. It appears probable that this is also the order followed by nature in increments to the number of lenses, from the time of the formation of the primary row of ommatidia (^ide seg.) onward to maturity. Under this arrangement the last row in enumeration is that ending in the upper anterior angle of the visual surface.

The number of these rows is variable, in the majority of cases being nine, in comparatively few individuals of average size and richly supplied with lenses, being teji, in extremely rare instances eleven, only a single example of average size showing so many ; in very young individuals the number of rows is but eight. The youngest specimens observed show no less than this number, and it seems probable from other considerations that in still



[Vol. II.

earlier stages of growth the number of diagonal rows was not much less than eight.

The number of lenses constituting the visual surface of each eye is variable^ but not irregularly so. The smallest number noted is thirty-three, in a very young individual having but eight rows ; the greatest is eighty-eight, occurring in the single example mentioned, which bears eleven rows.

Again, tJie number of lenses in successive rows is variable, but only so zvithin certaiji well-defined limits. In order to make this point clear, the following ten enumerations are presented, taken at random from a list of above three hundred tabulated eyes.


< W































































































































It appears from this table that whatever may be the total number of rows in any eye, in the last four rows of lenses, and generally in the fourth from the last, the number of lenses is subject to very little variation, while the number in the first two or three rows may vary greatly. This variation is partially explained by the following fact : It will be found upon examination of an average eye that the anterior edge is normally nearly vertical ; the vertical row following this edge is composed of fotir lenses. It is impossible for any lenses to be added to the anterior extremities of the diagonal rows terminating at this anterior margin, for new lenses are added only from the lower and upper margins of the visual surface (see further on). In


immature eyes a greater variation is noticeable in the last rows, as seen in the following examples of surfaces bearing but eight rows of lenses.









































In these instances the eye has not attained its full growth in height, which would preclude variations in the last rows.

Conversely, as the first four or five rows of lenses terminate on the lower margin of the visual surface, and as additions to the number of lenses are made most abundantly from this area, the number in these rows is constantly varying. It may be here stated, that with the exception of the right and left eyes of the same individual, no two eyes in all the specimens enumerated have shown the same number of lenses in all corresponding rows.

A definite relation exists between the number of lenses of the eyes and the size {i.e. age) of the atiintal. This fact has been established by recording with each enumeration of lenses a single measurement which would serve as an index of the stage of development attained by the animal. The measurement taken is the basal width of the cephalon. Phacops rana is rarely found with all the parts in articulation, and still retaining the lenses with sufficient distinctness for enumeration. Detached cephala are abundant, and it serves every purpose to take the indicial dimension from this part of the animal ; it is, moreover, found that the peculiar form of the cheek renders this dimension of the head less liable to variation from flattening in the shales than the longitudinal measurement. Comparison of all the specimens enumerated gives the following results : —

The average number of lenses in individuals having a cephalic width less than

10 mm. is 44 Between 5 and 15 mm. is 56.5 " lo and 20 mm. is 69.5 " 15 and 25 mm. is 73 " 20 and 30 mm. is 71 " 25 and 35 mm. is 66

258 CLARKE. [Vol. II.

Between 30 and 40 mm. is 62.5 " 35 and 45 mm. is 62.1 From 40 mm. upwards, 58

The calculated average basal cephalic width in this species, deduced from measurement of 15 18 cephala, is 22.8 mm. The material from which this average is derived was unselected, much of it collected without reference to quality or size, and is fairly representative. I therefore venture the statement that the average PJiacops rana has a width across the posterior margin of the cephalon of approximately 22.8 mm.

It is, moreover, probable that 22.8 mm. is approximately the dimensional index for the average normal adidt of this species. In all specimens of the entire animal which have passed under observation, varying in axial length from 10 mm. to 100 mm., no evidence has appeared of any developmental change in the successive stages of growth, except in the increase and diminution of the number of corneal lenses. Save in this one respect the species assumed all the features of maturity at a very early point in its history ; and the data given above conclusively indicate that in this feature, also, maturity was attained with this stage of growth. The important conclusion here drawn is that the number of lenses increases from yoiitJi to maturity (dimensional index approximately 22.8 mm.), and decreases from maturity to senility.

Two questions immediately arise from this inference ; {a) how is the number of lenses increased ? and {b) how is it diminished .? These points will be adverted to in a following section.

Structure of the Lens.

Sections across the visual surface show that the lenses are unequally bi-convex, the curvature being greatest on the proximal surface. This inferior surface is perforated by a central circular aperture. Vertical sections of the lens when favorably preserved, also show this envelope as a simple, thin, distinctly black or brown corneous film ; and in natural casts of the internal surface of the visual area, the ommatidial cavities are represented by a series of shallow cups standing on short pillars, and each bearing at its centre a little ball, which is the filling of the interior of the lens. These lenses are consequently corneal and


hollow. It appears, also, in sections that in the mature lens the cornea is discrete from the sclera, lying in juxtaposition with and held in place by it, but in nowise continuous with it. This fact is also frequently apparent in specimens from which the sclera has been removed by solution, leaving the corneal lenses standing on little pillars of the matrix, which has filled the ommatidial cavities. Other specimens just as frequently show the converse, the lenses being removed, while the sclera is retained. There is evidence which I deem worthy of consideration, that these corneal lenses, during the life of the animal, were not empty, but filled possibly with some viscid or spissate humor. This evidence is of the following character: —

The cornea itself is thin, and of even calibre throughout its extent. The casts of the corneal cavities, such as are shown in Figs. 25 and 26, as little balls lying in cups, are not of sufficient size to have occupied all the space within the cornea. A specimen of Phacops of a species closely allied to rana {Ph. cristata var. pipa Hall), from the decomposed phtanite of the Corniferous limestone, seems to demonstrate this fact. This fossil was evidently originally preserved in calcic carbonate, which not only replaced the entire crust, but filled the ommatidial cavities, and the posterior cavity of the corneae as well. This calcic carbonate was subsequently removed from within, and its place partially taken by silica, and when exposed to more rapid decomposing agencies, the remainder of the calcic carbonate was removed, leaving the fossil so preserved that the cornea and a thin film over the entire external surface of the sclera has been taken away, the remainder of the test being replaced by silica. The external surface of each lensal cast remains convex, but on carefully removing a little of the decomposed rock from beneath the position of the cornea a vacant space appears, which corresponds to the corneal cavity as represented by the ball-in-cup casts. To elucidate this point, see Plate XXI., Fig. 5, and explanation thereof.

Again, certain well-preserved sections from the limestone show a distinct difference in the character of the matrix filling the outer and inner cavities of the cornea, that in the outer being of lighter color, and more translucent i^. subcrystalline), while that in the inner is the opaque mud of the sediment. More evidence upon this point is very desirable, but enough

26o CLARKE. [Vol. II.

has been seen to indicate the fact that the cavity of the cornea was not simple, but compoimd. (May the posterior cavity-fillings represent the position of the anterior extremities of crystalline cones ?)

Multiplication and Diminution in the Number of


The lenses of the visual surface are not all of the same size in any of the stages of growth observed. The size of the fully developed lens varies according to the individual development of the animal ; i.e., the larger the animal, the larger the lenses ; but in any given subject some lenses are to be found which are below the average of size for that eye. These small lenses are found at the extremities of the diagonal rows which terminate on the posterior portions of the upper and lower margins of the visual surface.

The inferior size of these lenses is due in part to unlike causes. Of these the principal cause is {a) that they are new and immature lenses added in regular order to the ends of the rows of older lenses. It has not been as yet satisfactorily determined whether the increment of new lenses may take place at either upper or lower extremity of the diagonal rows, although the small lenses occur indifferently at either end. There is no reason to doubt that this addition does take place at the lower extremities, but on account of the close juxtaposition of the palpebrum to the upper margin of the visual surface, it may be questioned if at any period of growth sufficient room is allowed in this region for additional lenses.

A secondary reason for the small size of the lenses is ib) the constantly increasing size of the interlensar sclera after maturity, which gradually envelops the lenses especially along the upper margin of the visual surface, where, by coming into contact with the increasingly prominent palpebrum, the lenses are often nearly concealed.

To what degree the small size of the lenses along this upper margin may be due to the overgrowth of the sclera, and how much to the possibility of their being newly developed, it has been impossible to ascertain, but that it is due to a certain degree to both causes, is shown by the following facts : (i) immature


eyes, in which the sclera has attained no excessive growth, very often show these small lenses along the upper margin, and they would, therefore, appear to be developed there ; (2) it has already been shown that after the average normal mature growth of the animal has been reached, the number of the lenses becomes less with advancing senility. This fact must be explained either by the gradual envelopment of the lenses of the upper margin by the sclera and palpebrum, and their entire concealment within the substance of the latter, unless it is possible that atrophy of the opimatidial nerve branches and concomitant reabsorption of the lenses takes place with advancing old age.

From the examination of eyes limited to eight rows of lenses, it appears that with this number of rows there may be considerable variation in the number of lenses, as seen on the plate (Fig. 9, thirty-one lenses ; Fig. 10, forty lenses). These figures also indicate the fact, that with a constant diminution in the number of lenses from the upper and lower extremities of the rows, the eight diagonal rows would ultimately be reduced to a single or double longitudinal row parallel to the margins of the visual surface. Hence, without overmuch hypothesis, the primary lenses probably appeared in a single or double row, a visual line parallel to the margins of the orbital node.

Development of the Lens.

There is sufficient evidence at hand for the statement that that portion of the ommatidial cavity which penetrates the test arises from an evagination from the internal surface of the test accompanied by a corresponding but very shallow invagination from the upper surface. Natural casts of the internal surface of the visual area not infrequently show minute lensar cavities at the ends of the rows of lenses, which appear not to have penetrated to the upper surface, and bear at their summits no impression of a corneal surface or corneal cavity, as do the other lenses in the same eye. It would be inferentially true that the cornea is developed from the attenuated integument (cuticular epithelium), and is the specialized film of the test left between the depressions from its lower and upper surfaces, eventually becoming: discrete.


Structure of the Sclera.

The interlensar sclera is continuous with the test, and its structure is in all points identical with that of the test.

The vertical tubules and smaller tubulipores, with which nearly every part of the test of PJiacops i-ana is densely perforated, are plainly visible in every section of the sclera, no difference in the structure of the parts being discernible, although the thickness of the sclera is somewhat less than that of the adjoining portions of the test ; however, the thickness of the test is of necessity very variable in different parts of the animal. Not infrequently eyes have been observed, preserved as casts in decomposed chert, in which the tubules of the sclera are represented by delicate rods traversing the vacant space left by the removal of the integument.

Abnormalities in the arrangement of the corneal lenses are of comparatively rare occurrence. They appear to be due, in every instance observed, to the failure of a lens to develop at the proper time and place in its own row, but in no case has a lens appeared so out of place as to be intercalated between rows. Marked abnormalities, such as that represented in Figs, II and 12, are usually confined to one of the two eyes. It is, however, not uncommon to find the right and left eye differing in the number of lenses in the corresponding rows, either with or without affecting the total number of lenses. In the following example the total is the same, although the arrangement differs : —

Right eye. 6. 8. 9. 9. 9. 7. 6. 4. 2 = 60 Left eye. 6. 8. 10. 9. 8. 7. 6. 4. 2 = 60

In another example both the number in the corresponding rows and the sum total differ: —

Right eye. 5. 7. 8. 7. 7. 6. 4. i. 2 = 47 Left eye. 5. 7. 9, 8. 7. 7. 6. 4. 2 = 55

These instances of irregular development may be due to pathologic or other organic conditions of the animal ; perhaps, also, in part to external influences.


Note. — No satisfactory evidence of crystalline cones within the ommatidial cavities has been ascertained, and it is not surprising that these bodies, which undoubtedly existed, were removed with the soft parts of the visual organ. With respect to this feature the sections of the eye of Phacops given by Barrande {Systeme Silurien du Centre de la Boh'eme, Vol. I., PI. 3, Figs. 15 and 16), and reproduced by Zittel (^Handbuch der PaUrontologie, 1885), are misleading. The fillings of the ommatidial cavities are so shaded as " to indicate prisms " (compound) " corresponding with each lens," and extending very far inward without diminution in width. Such structure finds no correspondence in the eye of the living Arthropod, and is probably to a large degree schematic and imaginary. The structure of the lens, as we have found it, is also essentially different from that represented by Barrande.

Modes of Preservation of the Visual Surface.

a) The cornea and sclera are normally preserved. (Fig. I.) This is the usual mode of preservation in the limestones where the original substance of the test has been preserved in calcic carbonate, though leaving so considerable a portion of the organic matter as to give a black and lustrous surface. Such specimens retain the minute structure of the test most perfectly and are most satisfactory for sectioning.

b) The cornea is removed and the sclera retained. (Fig. 2.) This is a rare mode of occurrence noticed only in specimens from the shales.

c) The sclera is removed and the cornea 7'etained. (Fig. 3.) In these exaniples, which occur in the shales and weathered limestones, the corneal lenses stand supported on the summits of pillars of matrix. It is not an uncommon mode of occurrence.

d) Both cornea and sclera are removed (Fig. 4), leaving pillars of the matrix with cup-shaped upper surfaces, each bearing a little ball at the centre. This condition is often observed in the decomposed limestones and phtanites.

e) An external film is removed from the entire visual area, destroying the cornea (Fig. 5). A single example has been observed in which the entire test was apparently originally preserved in calcic carbonate. Subsequently this was removed from within and gradually replaced by silica, with the exception of the thin outer film, which afterward was entirely removed, leaving the space it occupied vacant.

f) Silica deposited as a thin film npon, or replacing a thin film of the external and internal surfaces of the test, and all the rest of the substance of the test and the matrix removed (Figs. 6 and 7), In this condition the visual surface is a mere shell appearing as

264 CLARKE. [Vol. II.

when normally preserved, but the corneal lenses are hollow, and the sclera represented by a thin wall of silica. This condition is sometimes modified by the removal of the entire upper surface of the visual area, generally by its adherence to the outer part of the matrix, leaving only vertical tubes representing the ommatidial cavities.

While the foregoing observations and the essential conclusions therefrom in regard to the structure of this phase of the trilobite eye, agree in some respects with the opinions of earlier writers upon this topic, there are many important points of difference and various features of structure which have not before been noticed. I therefore give a brief historical review of the observations upon the subject.

Quenstedt, 1837 (Wiegmann's Archiv fiir Natiirgeschichte, Vol. I., p. 340), was the first to recognize two distinct types of structure in the eyes of trilobite, and divided them into

1. * Aggregated eyes having a facetted cornea.

2. Aggregated eyes having a smooth cornea.

In both groups the cornea was regarded as the direct continuation of the superficial layer of the test of the cheek. The first type of structure was represented by the eye of Phacops latifrons, Bronn {i.e., the Phacopidae). The second type was exemplified by Illcenus crassicauda, Dalman {i.e., holochroal eyes), in which the facets are said to be in relief upon the internal surface of the general corneal (visual) area, each facet being formed by a lens or crystalline body, behind which lies a vitreous body, penetrating deeply into the organ. In these the cornea was regarded as composed of two distinct layers, of which the outer is quite smooth, the inner very finely reticulate.

Burmeister, 1843 {Organization der Trilobiten, p. 19), regarded the structure of the holochroal eye as directly comparable in all respects to that of the eye in Bra?ichipus stagtialis, and indorsed Quenstedt's view of the compound corneal layer, while admitting but a single type of structure. He assumes with respect to Phacops that the cornea must have been more destructible than in the other trilobites, and by its removal the facetted surface exposed.

  • As I have not had access to this work, I am compelled to take the summary of

these observations as given by Barrande (^Syst. si/., p. 133).


Barrande, 1852 {Systeme silurien du centre BoJieme, Vol. I., p. 185), recognizes three distinct types of structure, two of which are similar in breadth to those of Quenstedt, though differently interpreted. The third type of structure is exemplified in the genus Harpcs, the eye of which is considered as an aggregation of ocelli {two or three in each eye). For the schizochroal eyes Barrande establishes the fact that the sclera (cornee opaque) of the visual surface is identical with that of the test of the head, and continuous with it. He also regards the existence of a transparent cornea covering the entire surface as suggested by Burmeister, as probable, but admits that he has been unable to assure himself upon this point. For the holochroal eyes the envelope of the visual surface is shown to be of different character than that of the test, and consists of a general cornea covering the entire visual area. His observations do not seem to have led to a decisive conclusion in regard to the simple or compound character of this cornea.

Packard, 1880 ^Fhe Structure of the Eye of Trilobites : American Naturalist, p. 503), quotes a resume of the status of the discussion as given by Gerstacker in Bronn's Classen und Ordnungen des Thierreichs, which is virtually a reiteration of Barrande's opinions. The "ocelli" of Harpes are regarded as aggregate eyes, not comparable with the simple eyes or ocelli of Limulus and the Merostomata. The eyes of Barrande's first and second groups are considered as true compound eyes and not aggregated eyes ; no essential difference is recognized in the form and arrangement of the corneal lenses of Phacops and Asaphus, and the distinctions pointed out by Quenstedt and Barrande are considered artificial.

The sections used by Professor Packard in the comparative study of the trilobite eye appear to have been altogether of holochroal forms. A close correspondence in structure is demonstrated between the eye of Asaphus and that of Limulus. In the Limulus eye the lenses are covered by a continuous corneal layer, which does not make itself apparent in Professor Packard's sections of Asaphus, although it undoubtedly exists. It may be questioned whether the conclusion drawn from this comparative study is not too broad, viz., that the trilobite eye is organized on the same plan as that of Limulus}

^ Professor Packard's observations upon the eye of Asaphus show no indication of the existence of interstitial epithelium between the lenses, an extremely important

266 CLARKE. [Vol. II.


The study of the eye in Phacops rana as here presented allows the statement of the following points : —

i) The schizochroal eyes of the Trilobites are aggregated and not properly compound eyes. The visual organs of Harpes may prove to be of similar character.

2) The scleral portion of the visual surface is of the same structure as the test, and is a direct continuation of it.

3) No evidence appears of any continuous corneal layer covering the entire surface.

4) The corneal lenses are wholly discrete from the epidermis, but are of epidermal origin. In the addition of new lenses to the visual surface, they appear to arise from a thinning of both surfaces of the integument.

5) The corneal lenses were hollow or filled with some matter not homogeneous with the cornea itself.

6) The corneal lenses, and therefore the ommatidia, are added to the visual surface with advancing age until the mature growth of the individual is attained ; thereafter they diminish in number with increasing senility.

7) The addition of corneal lenses occurs regularly at the extremities of the diagonal rows.

8) No evidence is preserved of crystalline cones in the ommatidial cavities, though they may have been removed in the decomposition of the soft parts of the eye.

In conclusion I wish to call attention to the primitive structure of the eye exhibited in a Devonian subtype which is provisionally referred to the order of the Phyllocarida. Dr. J. S. Kingsley in the final paragraph of a valuable paper on the '* Development of the compound eye in Crangon " {Jojirnal of Morphology, Vol. I., p. 63), has written : " The observations as yet recorded are not sufficient to throw any great light on the phyllogeny of the Arthropod eye ; still one or two points may be

feature of difference from the schizochroal eyes. My own study of the holochroal eyes has not been as careful as I hope to have the opportunity of making it, but it may be observed that sections of the eye in Proeius Roiui seem to indicate a very tenuous interlensar sclera; moreover, the immature Calymene senaria referred to in a previous footnote shows evidence of such interlensar integument.


spoken of. The mere fact of invagination ^ must be regarded as indicating an ancestral condition, but what this condition was is uncertain. The pit or groove must have had sensory functions and either wall" (retinogen and gangliogen) "must for a time have been like its fellow, as is shown by its having similar nuclei, and by the similar development of rows of nuclei."

In the species Mesothyra Oceani, Hall, a member of one of the faunas of the Portage group, and one of the largest known representatives of the Phyllocarida, the eye consists of a simple deep pit at the summit of the optic node. There is no evidence that this pit contained a series of lenses, but it is highly probable that it is an otherwise embryonic character retained at maturity, and may serve as the ancestral condition of the Decapod eye suggested by Dr. Kingsley. That there is Decapod blood in the Phyllocarida has been disputed by Packard, the author of the group term, but Decapod affinities are strongly indicated by the recently described Devonian genus of Phyllocarida, Rhinocaris, {Palceontology of Nezv York, Vol. VII., 1888).

1 This refers to the primary optic invagination in the embryo, for the formation of the entire visual surface, first pomted out by Locy {Bull. Ahis. Comp. Zo'ol. 1886), and verified by Kingsley {op. cit.). The term is not used in a sense similar to that in vehich it has been employed in this paper in referring to the formation of the corneal lenses.

268 CLARKE. [Vol. II.


Figs. 1-7. Schematic representations showing the different modes of preservation observed in the visual surface of Phacops. Each figure represents a single lens with the lensal or ommatidial cavity, and the adjoining sclera or its equivalent space.

1. The cornea and sclera normally retained, the lensal and corneal cavities being filled with matrix. The lens is represented with its posterior cavity having the relative size indicated by the specimens figured elsewhere on the plate, and the anterior corneal cavity is characterized by radiating lines which are intended to show the difference often apparent in the character of the matrix filling the cornea.

2. The corttea retnoved and the sclera retained, the matrix showing the position and size of the posterior corneal cavity, but retaining no indication of the anterior cavity filling, which is removed with the cornea.

3. The cornea retained and the sclera removed, the lens standing at the summit of a pillar of matrix which represents the ommatidial cavity.

4. Both cornea and sclera removed, leaving pillars of matrix with a cup-shaped summit, in the bottom of which lies a little ball. The pillar represents the ommatidial cavity; the concave summit, the lower surface of the lens, and the little ball, the posterior corneal cavity.

5. An external fib7i is removed from both corttea and sclera, destroying the former and leaving the filling of the anterior corneal cavity standing out prominently with almost the full size of the lens. In the single instance observed of an eye in this condition, the posterior corneal cavity is empty in all the ommatidia that were opened. The reason for this is not well understood. The sclera has been silicified, and subsequently decomposed, so that it is indistinguishable from the matrix.

6. The outer and inner walls of the visual area have been replaced by a film of silica, and the rest of the calcareous matter subsequently removed, leaving both cornea and sclera preserved as a mere shell.

7. The same condition of preservation, modified by the adherence of the cornea and outer wall of the sclera to the matrix outside the eye, leaving the walls of the ommatidial cavities adhering to the internal matrix as a series of short tubes. This mode of fossilization has not been observed in Phacops rana, but is not uncommon in specimens of Phacops cristata, var. pipa, from the decomposed Upper Helderberg phtanite.

Figs. 8-22. Schematic representations of the lenses of the visual surface. The curved surface is projected upon a plane, and the relative size and position of the lenses is retained. Whether the representation is from a right or left eye, the lower posterior margin is at the right of the figure, the diagonal rows being enumerated from this side, obliquely downward from right to left. All the figures are drawn to the same scale.

8. The visual surface of an extremely young individual measuring 6 mm. across the base of the cephalon; composed of 31 lenses in 8 rows, nearly all the terminal lenses being immature. The older lenses show a tendency to arrangement along a single or double transverse row, parallel to the margins of the visual surface.

9. An older eye, bearing 35 lenses in 8 rows, belonging to a young Phacops, having a cephalic width of 7 mm. At the upper extremities of the rows the lenses are all full-grown, and all immature at the lower extremities.

10. An eye composed of 40 lenses in 8 rows, and belonging to a young individual with a cephalic width of 9 mm.


11. An eye with 42 lenses in 8 rows, belonging to a large individual, having a width across the cephalon of 34 mm. The lenses are nearly all mature, but are abnormal in their arrangement, showing a failure to develop properly at the upper extremities of the rows after the third.

12. An eye of Phacops crista ta, var. pipa, composed of 50 lenses, and showing an abnormal arrangement, a single mature lens in the last row being situated high up in the base of the palpebrum. Had it developed in the vacant space in the last row but one, the arrangement of the lenses would have been normal for this variety.

13. An eye of the same variety bearing 52 lenses in 8 rows. Eight appears to be the normal number of rows in the mature eye of this form, the anterior vertical row, and the last diagonal row, consisting of three lenses. In the other diagonal rows there is apparently much greater variation than in the eye of Phacops rana.

14. An eye of Phacops rana composed of 57 lenses in 9 rows, the head having a basal width of 52 mm. This is the eye of a senile individual, and all the lenses are of mature size with the exception of those on the palpebral margin. The number of lenses in the anterior vertical row is three instead of four, as in the normal adult.

15. Another senile eye, with 60 lenses in 9 rows, belonging to a very large specimen, with a cephalic width of 70 mm. Here again the lenses are all of full size with the exception of those on the upper margin of the visual surface; the lenses of the anterior vertical row are also three in number.

16. An eye composed of 60 lenses in 9 rows belonging to an individual slightly above normal adult size, having a cephalic width of 28 mm. In the first row one interval has been skipped in the addition of the last lens.

1 7. An eye of normal size and development, bearing 70 lenses in 10 rows, the cephalon having a width of 24 mm. The first row consists of a single full-grown lens at a considerable distance from the posterior extremity.

18. An eye from an individual of the same size, having 70 lenses in 9 rows.

19. An eye composed of 71 lenses in 10 rows, the cephalon to which it belongs having a width of 27 mm. It essentially differs from the preceding only in the presence of the single lens constituting the first row.

20. An eye with 75 lenses in 10 rows, from an individual having a cephalon 28 mm. in width.

21. An eye composed of 88 lenses in il rows, from an individual measuring 17 mm. in cephalic width. This is the greatest number of lenses and rows of lenses noticed in this species.

22. An average eye of Dalmanites Boothi, var. Calliteles, Green, consisting of 206 lenses in 29 rows.

Fig. 23. A left eye of Phacops rana, enlarged to 3 diameters, showing the arrangement of the lenses, the smaller size of several of the terminal lenses, and the tubercles on the integument of the palpebrum and orbital ridge. The lensar cavities are represented as much too sharply hexagonal; they should be more rounded and excavate. The figure is copied from the Paleontology of New York, Vol. VII., PI. 8, Fig. 6.

Fig. 24. A portion of the visual surface of a very old eye, taken from an individual measuring 70 mm. across the base of the cephalon; showing the thick and deeply excavate sclera, the full-grown lenses along the lower margin, and the small lenses at the upper margin of the area. Enlarged to 3 diameters.

Fig. 25. A natural cast of the internal surface of a portion of the visual area in Phacops cristata, var. pipa, enlarged to 6 diameters. The specimen shows very beautifully the cup-shaped casts of the ommatidial cavities, each with a little ball at its centre representing the posterior corneal cavity. At the upper extremities of the last

2/0 CLARKE. [Vol. II.

two vertical rows at the left, are two casts of immature lenses, in one of which the fining of the corneal cavity is just discernible; in the other the ommatidial cavity appears not to have penetrated to the upper surface. Several features of similar character are to be seen on parts of the specimen not represented in the figure. The portion of the eye represented is the posterior one-third of the right eye.

Fig. 26. Two cavity-fillings from the same specimen enlarged to 15 diameters, showing the small size of the balls compared with the probable size of the entire cavity of the cornea.

Fig. 27. A section through the eye and adjoining parts of Phacops rana, showing the lenses and the interlensar sclera. Enlarged to 3 diameters.

Fig. 28. The same enlarged to 6 diameters, showing the continuity of the poriferous integument of the head with the interlensar sclera, the double convexity of the lenses, and the depth of the ommatidial cavities in the sclera. At the right of the first lens in the series is an internal depression on the integument which appears to indicate the position of a newly developing ommatidium. No evidence of a lens was visible at this point before the section was made.

Fig. 29. Three lenses with their interstitial integument, from the same specimen, enlarged to 10 diameters. There is a slight difference in the character of the matrix filling the cavity of the cornea, it being possible to distinguish the line between the anterior and posterior divisions of the space. The dark color of the lower portion of the sclera is due to an increase of pigment.

Fig. 30. A portion of the eye of Phacops rana, from which the sclera has been removed by natural causes, leaving the cornese standing on pillars of matrix. Enlarged to 12 diameters.

Fig. 31. A natural section of the eye oi Phacops cristata, w&r. pipa, enlarged to 8 diameters. The sclera has been removed, and the doubly convex surface of the lenses is well shown.

Fig. 32. The eye of Mesothyra Oceani, Hall, enlarged to 3 diameters, showing the strong optic node with a simple, deep pit at its summit. The figure is taken from the Paleontology of New York, Vol. VII., PL 32, Fig. 2.

/oiiri) . Mor/ih . Vol .11




i ^ A • A • i




•••••* ••*•••'



'••• '•••'




• •••»






••••• •••••







BMeisel.hOi Boston.



During the spring of 1872, Professor Poey of Havana, Cuba, came into possession of a very remarkable form of fish, which presumably was taken in Cuban waters.

Fortunate it was for science that it fell into such excellent hands, as that eminent ichthyologist promptly presented us (Ana/, de la Soc. Esp. de Hist. Nat., Tom. 11, 1873) with a very excellent account of this more than rare type, the duplicate of which, so far as I am aware, has yet to be found by naturalists.

This account of Professor Poey's, as will be seen, was published in the Spanish language, and it has given me much pleasure to make a translation of it, and present it here as an introduction to some subsequent examinations which I had the rare opportunity of undertaking upon the skeleton of the specimen in question. The osteology of this fish is very interesting, not only from the fact that it is the only specimen in the hands of science, but from its extraordinary peculiarities, and from the fact that it may some day be found upon our own coasts.

My translation of the original description just referred to reads as follows : —



Professor in the University of Havana, and Corresponding Member of the Academy OF Exact, Physical, and Natural Sciences of Madrid.

The length of this extraordinary fish is 470 millimetres. The head enters five times into the total length of the body, and two

272 SHUFELDT. [Vol. II.

and two-thirds times into its greatest depth. The body is much compressed, and quite deep. The very large eye is contained two and a third times in the length of the head, and lacks the membrana adiposa.

The branchial apertures are deeply cleft, but I fail to find more than four branchiostegal rays, without being able to assert that there may be a greater number of them. The snout is short. The prefrontal, the turbinal, and the anterior suborbital, are extremely hard, and covered with spiny rugosities. The preoperculum and interoperculum have rugose borders, while the remaining opercular bones are entirely so. The mouth is small, subvertically cleft ; the premaxillary process is large, and is lodged in a fossa of the cranium. The maxillary is complicated. The teeth are simply a narrow row of minute prickles ; they do not occur upon the vomer, nor the palatines.

D. 6-34; A. 2-33; V. I, 6 ; P. 15 ; C. 1-13-1. The leading spine of the first dorsal series is rugose, as is the first ventral, the two post-anals, and the external ones of the tail, which latter show the condition equally well in either one.

The rays of the pectoral, second dorsal, and the anal fins are compressed, and do not ramify at their extremities. The pectorals are very short and rounded. On the other hand, the vertical fins, the dorsal, and anal are well developed.

The tail was injured, and apparently cut ; the membrane which unites its rays had disappeared ; the peduncle which supports it is large, and capable of communicating a powerful impulse to the act of progression. The thoracic pectorals unquestionably possess a rugose spine and six flexible ones that are branched.

Aside from the frontal bones and the suborbitals where the skin abruptly terminates, and the nasal portion of the snout, all the trunk and the head is covered with scales, including the inferior mandible.

These scales in no way resemble those found among the acanthopterygean fishes. Their length greatly exceeds their width ; they have the appearance of parchment, — transparent, brittle when dry, — overlap each other, and are strengthened longitudinally by a raised lineal ridge.

Their contact with each other is so extremely intimate that it lends to the skin of either side a very smooth appearance — so


much so, that the rough borders of the scales would not be suspected without the aid of the Angers.

Thanks to the length of these scales, four, five, or six of them are sufficient to span the height of the trunk, one of such a series being crossed by the lateral line, where its presence is denoted by a raised ridge.

The leading scales on the body, above as well as below, are shorter, and where carried on to the head, are doubly as firm as those found at the base of the fin rays.

Without having done more than counted the scales in a longitudinal line, I calculate that the number is considerably above two hundred ; those of the head, although shorter, have the same form as those of the trunk. There are no scales upon the fins.

The caudal peduncle develops neither a cartilaginous nor an osseous plate at its sides. Posterior to the arms the ventral keel is rough.

The cranium is more cartilaginous in structure than it is osseous, except the frontals, which are rugose in lines in the supraorbital region, and bristly in front, as are the turbinals and suborbitals ; these latter are four in number, the last three being very slender. There are two supratemporals.

The inferior mandible is characterized by several rows of minute spines upon the dentary and articular elements. The vertebrae number 10+36.

The anterior neural spine is not excavated, being lofty and smooth ; the five that follow are short and inclinecl backwards. The remaining ones are slender, which applies also to their haemapophyses. The last vertebra is without lateral spines.

The pleurapophyses are inconspicuous, feebly developed, and have much the same size and shape as the epipleurals. I discover but one pseudo-interneural spine in front of the one that supports the first dorsal fin ray.

The specimen I described, when received by me was without gills and without abdominal viscera. Preserved as it was three days upon ice, its general color appeared to be white ; but we have reason to suspect that in the fresh condition this fish can easily assume a violet tint. The hard parts of the head were of an intense violet shade. The ascending border of the preoperculum, violet. The fins were white, changing to violet in cer

274 SHUFELDT. [Vol. II.

tain lights ; the caudal fin rays were of a reddish tint. Eye inclining to white.

Family : The characters which are presented us in this fish are of such an extraordinary nature that they will not permit us to place their possessor in any of the recognized families of fishes. Its nearest affinities are with the BcrycidcB and with the Carangidce, two families widely separated from each other; I am inclined to believe that its better place is along-side of the lastnamed one. Its resemblance to the Bcrycidce is seen in the large eye ; the asperity of the cranium ; the rugosities upon the fin rays ; the ventrals composed of more than five soft rays, over and above the spiny one ; its resemblance to the Carangidcs is seen in the two free spines which precede the anal fin, and especially to Scriolas for lacking the bony plate of the lateral line ; but in the number of its vertebrae it approaches the Sconibridce, as the shape of its ventral fins are in pattern analogous to those of the AcantJuiridce, and its unramified fin rays agree with the Balistida.

The character of the scale, to which ichthyologists have attached so much importance, separates it from all other forms known to me.

My examination, then, authorizes me in establishing the family Graminicolcpidce, based upon the following characters : Lateral line unarmed with bony plates ; ventral fins composed of more than five soft rays ; two free postanal spines ; caudal vertebrae numerous ; scales very long and narrow, without fan-like expansions or denticulations.

Genus : The genus Graimnicolepis has for its etymology <ypaix/xip(;o9, line ; XeTrt?, scale.

The characters, in addition to those I have already pointed out for the family, are : Body deep, compressed ; eye large ; mouth small ; head, in part, rugose, which also applies to the interoperculum and the preoperculum ; to all appearances a limited number of branchiostegal rays ; teeth mere asperities, the palatine arch without them ; two dorsals, the first short, the second very extensive, its height insensibly increasing; pectoral short and rounded ; the dorsal, anal, and pectoral fin rays do not ramify at their extremities.

History : I saw this fish for the first time in Havana, on the 5th of April, 1872, and I have not observed it since; neither


fishermen nor students of the class have been able to give me its name, because neither one nor the other have seen it to know it. It is, then, one of the rarest forms in existence. The skeleton I have sent to the eminent Professor Gill, who has it in his possession, though I do not know but that he has preferred placing it in the collection of the Smithsonian Institution in Washington."

This account is completed by a plate and its accompanying description, showing the fish one-third the size of nature, and various illustrations of its scales and other parts.

Now about a year ago. Professor Gill did me the great honor to place in my hands, for a little more extended illustration, not only the skeleton of this rare type, but a life-size outline drawing of the fish made by Professor Poey himself. In addition to these treasures, this eminent zoologist also placed at my disposal several crania of fishes, representing the genus Caranx and others, to be used in the present connection. Situated as I now am, at an outpost in New Mexico, notwithstanding the great value of these crania for comparison, I can only regret that the material at my hand is not still more extensive, as it might be, were I more favorably situated to undertake this kind of a paper. Especially would I like to examine specimens of Brama Raii} which, if I have recalled the proper form, possesses vertical linear scales something like those in Graminicolepis, though, I believe, very much smaller.

In order to give an idea of the external appearance of the subject of this article, I brought to my aid the two drawings of

• Since writing the above, a very valuable work upon ichthyology has appeared, viz. : The Fisheries and Fishery Industries of the United States. By G. Brown Goode, Asst. Direc. U. S. Nat. Mas. and a staff of Associates. Washington, 1884. On page 335 of the text of this book, we read of the Bramid.e that "The only member of this family of interest to us is the Braina Raii, called " Pomfiet " in Bermuda, where a few individuals were observed by the writer in 1876. In 1880, an individual was obtained on the Grand Bank of Newfoundland, and more recently the species has been found to be somewhat abundant on the coast of Washington Territory and Vancouver's Island. This species was described from the coast of South America, under the name Brama Chilensis." In the second volume of this work, we find an excellent figure of Brama Raii, Plate 1 1 2, which shows the fish possesses vertical linear scales, although they are much shorter than they are in the subject of this article.

The Pomfret also has its tail more deeply forked, and the dorsal fin is seen to be continuous. The eye is very much smaller, though otherwise there are some general external resemblances between the two forms. (R. W. S., 7 Aug., '86.)



[Vol. II.

Professor Poey, neither of which profess to be anything more than the merest outHne of Grainmicolepis, and the scales, fins, and other parts that accompanied the skeleton. These, taken in connection with the lucid description of the fish, and all care




fully compared, have resulted in my drawing presented in Fig. i of this memoir.

Owing to the fact that many parts of the skeleton, from long keeping and their delicate structure, have warped considerably out of shape, I propose to devote myself on the present occasion only to such as seem most important of them, and chief among these stands the cranium.

As I say, so far as I know, the specimen of the craniuvi of Grummicolepis before me is the only one in the hands of science, and a most extraordinary object it is. Three features strike us most forcibly when we first came to examine it : the enormous orbits, the truncate appearance of its anterior part, and the semi-transparency of its gelatinous-looking bones.

Figure 3. — Left lateral aspect of the cranium of Grammicolepis brachiusculus; life size, drawn by the author from the specimen. Fr, frontal; Pa, palatine; S.O., supraoccipital; Sq, squamosal; Ep.O., epiotic; Pt.O., pterotic; E.O., exoccipital; C/. (9., opisthotic ; j5. (9., basioccipital; ^i-, basisphenoid; Pr-C, prootic; /•(/", postfrontal; ^j, alisphenoid; C5, orbitosphenoid; j5^//i, ethmoid; jP/yj prefrontal; Pr.S., parasphenoid; Vo, vomer.

The peculiar rugose condition of di frontal bone, referred to by Professor Poey, is well shown in Figs. 2 and 3, Fr. It will be seen that these ruo-osities of the frontal radiate from a common centre on its superior aspect, this centre being found at about the middle of the bone, or what would be the middle of its oblong figure were its anterior internal notch completed, and we do not regard its postero-lateral prolongation. This latter part of the bone forms the superior periphery of the orbit, and is produced backwards as far as the squamosal {Sq). To the inner side of

278 SHUFELDT. [Vol. II.

this process, the posterior border of the frontal shows at least one conspicuous notch, while its free margin overlaps the supraoccipital, and is in turn overlapped by the parietal {Pa) more externally {Fig. 2). Mesially, its surface turns upwards, more particularly behind, where with the fellow of the opposite side it grasps in the middle line the anterior portion of the supraoccipital crest. Below this point the two frontals have their straight, free, mesial edges roughly in contact with each other, and slope gradually downwards to the margin of that concavity which is found in front (Figs. 2 and 3).

This extraordinary fossa on the anterior aspect of the cranium of Grammicolepis is entirely open above ; its rugose and subcircular margin being formed by the frontals ; while below it becomes conical with its apex in the middle line, and in the ethmoid. Above, where it is most capacious, it has its posterior wall formed by descending plates developed on the part of the frontals, the left one considerably overlapping the right. Below this, in the middle line, there is an opening of some size, which leads into a commodious chamber lying between the frontals above and the mesethmoid below.

A frontal is truncate in front, where it overhangs the corresponding prefrontal, and internally articulates with the curiously shaped ethmoid. Behind this, and on its under side, it forms the major share of the roof of the orbit. Then occurs a longitudinal keel, which separates this from that other part of its under surface which forms the roof of the mid-chamber described in the last paragraph. Viewed together from above (Fig. 3), it will be observed that the rugosities of the frontals are limited behind by a subparabolic curve with its arc anteriorly directed.

In this dried cranium a parietal {Pa) is represented by a thin, flake-like, semi-transparent piece of bone, of a form shown in Fig. 3. To the outer side of its mid-longitudinal line it develops for its entire length quite a prominent, though thin, crest, which is rugose all along its superior margin.

The anterior three-fifths of the under surface of this element simply rests upon the frontal and supraoccipital, while the remaining portion behind is more firmly attached, and really holds the bone in its position. Its outer free margin articulates principally with the inner border of the posterior prolongation of the corresponding frontal, though still more posteriorly it meets


to some extent the squamosal {Sq). With the epiotic {Ep.O.) it is connected simply by a feeble and thin band of bone. The parietals, then, seem to play the part here of binders, rather than their presence is at all essential to covering over any sizable vacuity in the cranial vault, that might exist were either of them removed.

The siipraoccipital is a very extensive ossification, and is characterized by a fairly prominent crest.



Figure 3. — Cranium of Gratnmicolepis brachiuscuhis, seen from above; life size, by the autiior. Lettering as before.

This crest is triangular in outline, with its apex above, and its base attached in the middle line, to the horizontal portion of the bone. From the middle point of this base, a narrow, fanshaped development springs upon either side, which is incorporated with the crest, to strengthen it, being carried nearly as high as its apex (Fig. 2).

, The broad and spreading horizontal portion of the bone forms the roof of the brain-case, while posteriorly, just before the termination of the crest, it is bent abruptly down to meet the exoccipitals. The flexure, being sharply defined by a transverse line, the outer end of which, either way, terminates in the apex

28o SHUFELDT. [Vol. II.

of a pyramid, the lateral and upper sides of each being also formed by the supraoccipital. The lateral aspect of this pyramid is overlapped by the epiotic {Ep.O), while outwardly its free margin articulates with the squamosal (Fig. 3).

Regarding one of these epiotics, we find that its fan-like portion is finished off behind by a semicircular piece, which is thickened below, where it becomes firmly attached to the neighboring bones. The blade portion is longitudinally fluted, but no rugosities are found upon it. This does not apply to the element at its outer side, the squamosal {Sq), which element develops very conspicuous rugosities upon its upper aspect in direct continuation with the longitudinal ones on the long, backward-extending process of the corresponding frontal (Fig. 3).

At the distal extremity of the squamosal I detect a small, flake-like piece of bone, thoroughly attached, though individualized by sutural traces, which I take to be the representatives of the pterotic {Pt.O.). Beneath and beyond, the squamosal seems to make the usual ichthyic articulations with the postfrontal {Ptf) and the prootic {Pr.O). At its under side we find a small hyomandibular facet (Fig. 4, /if).

A postfrontal of considerable size {Ptf) develops at its outer side, a sharp, descending, spicula-form process of bone, which is transversely pierced at its base by a small foramen. The inferior articular sutural trace of the postfrontal, as I make it out, is subcircular in outline, and closely meets corresponding marginal concavities offered by the prootic and alisphenoid.

Each alisphenoid is necessarily a very extensive ossification, forming, as they conjointly do, the major part of the bony wall of the posterior aspect of the immense orbit {As).

In front they articulate with far smaller orbitospJicnoids, which in their turn meet in the median line anteriorly {Os). Now above the basisphenoid {Bs), the alisphenoids and orbitosphenoids are separated from each other, mesially, by a vertical vacuity, broadest below, gradually tapering to a blunt apex above, which constitutes a great fenestra for the anterior wall of the cranial casket (Fig. 4).

As already stated, the hinder portion of the ctJnnoid {Etli) forms the mid-roof of the orbital space. This division of the bone is of an oblong outline, being encroached upon by the common, circular, anterior margin of the obitosphenoids behind,

No. 2.]



the two elements being completely united. Below, it is convex from side to side, correspondingly concave above, where it forms the floor of the interfronto-ethmoidal chamber, already alluded to above. I have previously described how now the ethmoid is deflexed, and becomes concaved in front to form the lower limits of that excavation on the anterior aspect of the cranium.

This, as we have already seen, terminates in a conical point, and even beyond this the bone is carried forwards as a median




Op. a


Figure 4. — Under side of the cranium of G^-aniutiiolepis brachiuscnhts ; life size; reference letters the same as in former figures, with Jif, hyomandibular facet, and gp, foramen for the exit of the glossopharyngeal nerve in the opisthotic {Op.O.).

triangular process, the apex of which rests upon the parasphenoid (Fig. 3, Eth. Pr.S.).

Upon either side of the ethmoid, the flat anterior aspect of this cranium is completed by a broad prefrontal {Prf). The form assumed by one of these elements can best be appreciated by referring to Fig. 5, which represents the cranium of Graniinicolepis seen from behind, while the anterior face is in the horizontal plane. In this position the posterior aspects of the prefrontals come into view.

Each one essentially consists of a thicker and vertical outer

282 SHUFELDT. [Vol. II.

column of bone, antero-posteriorly compressed, and an expanded inner portion, which latter is reenforced by radiating projections that converge to meet at a point at the lower p^rt of the inner margin of the columnar portion. It is hardly necessary to state that these prefrontals form the externo-lateral parts of the anterior orbital wall, the ethmoid completing it mesially.

Coming now to the vomer ( Vo), we find it to be a thin scalelike bone, of a form best shown in Fig. 4. It rests in the longitudinal excavation of the anterior and lower side of the parasphenoid, while its firmest attachment to that bone seems to be by the periphery of its anterior margin.

The parasphenoid (Pr.S.) is gently arched as it spans the orbital space below, having its convex arc downwards. The lower side of this part of the bone, as I have already intimated, is longitudinally scooped out, while the upper side presents lateral surfaces, which are inclined so as to meet in a median line. Posteriorly, the parasphenoid makes the usual teleostean connections with the basioccipital, basisphenoid, and prootic, being deeply cleft as it passes to cover the under side of the firstnamed element (Fig. 4).

Occupying its usual position, the basisphenoid (Bs) not only develops the median process (Fig. 2) seen in so many true teleosts, but furnishes a firm horizontal roof for the three-sided pyramidal eye-muscle canal, the lateral walls of which are completed by the prootic and parasphenoid.

As in the majority of osseous fishes, the prootic is a welldefined and important element at the lateral aspect of the braincase {Pr.O.). Its anterior margin is pierced by the foramen for the trigeminal nerve, from which point faint lines in the tissue of the bone are seen to radiate.

The basioccipital {Bs) has its thickened and longitudinal portion underlying the brain-case, as in most fishes, being completed behind the facet for the leading vertebra of the spinal column. This facet is comparatively rather small, with its conical depression very deep.

At either of its sides the basioccipital develops an upturned and semicircular plate of bone, similar in structure to the other flat bones of the lateral cranial walls, which articulates with the lower margin of the opisthotic and the posterior margin of the prootic (Fig. 2).


The opistJiotic [Op.O.) is large and occupies its usual position, as generally found, in the cranium of the teleosts. Its posterior margin is pierced by a conspicuous foramen for the exit of the glossopharyngeal nerve from the brain-case. The intersutural traces defining its borders are easily made out in the specimen, and this element contributes not a little to the lateral wall of the cranial cavity, — a large vacuity existing after its disarticulation.

Each exoccipital {E.O.) develops at the outer sides of the vertically oval foramen magnum, a fan-shaped, bony thickening (Fig. 5), which nearly meets at the middle point above.


Figure J. — Posterior aspect of the cranium of Gratnmicolepis bracMusculits ; life size. Letters have the same signiticance as in former figures. The cranium is represented as resting upon its anterior face on the horizontal plane, the line of sight being perpendicular to the latter, and passing through the imaginary centre of the foramen magnum.

Further, these bones spread out so as to complete the hinder cranial wall, where the supraoccipital and osseous elements at the lateral angles have failed to do so.

Now a number of the bones required to complete the skull of this fish have been lost, and, as I said before, the others in my possession are too much out of shape, from their fragile nature, for me to decide, with any degree of certainty, as to their several proper positions. This is much to be regretted, as I expect a complete skull of Grammicolepis would prove a very interesting and instructive object.

284 SHUFELDT. [Vol. II.

To return to the cranium for the purpose of taking a general glance at it : we are to note especially the almost entire absence of those parial and lateral crests, developed on the part of the parietals and squamosals, so manifest in some fishes, as for instance the genus Caranx ; we are to note, also, the very peculiar texture of the bone that composes this cranium, being more like the material that is found in ordinary fish scales rather than bone ; particularly are we to observe the relation between the anterior portion of the supraoccipital crest, and the upturned portions of the frontals.

There are but few striking features within the cranial case of this strange form of fish. For the most part, surfaces, convexities, and concavities on the outside give rise to similar surfaces on the inside, the last two being, of course, reversed. The fossae for the otoliths are ample and well defined, but the elements themselves have been lost.

I have already expressed my regrets at not having at hand more extensive material wherewith I might compare this extraordinary fish ; they only increased as my investigations proceeded, while the remaining consolation left me, is, that I feel I have added at least my mite to the labors of Professor Poey ; so should another specimen of Graniniicolepis fall into the hands of naturalists, we can, at least, meet it with drawings of its cranium and other skeletal parts, as well as with similar drawings of some of the forms to which it is supposed to be related.

Through the courtesy of the Smithsonian Institution, and the kindness of Dr. Gill for selection, I find before me the cranium of a specimen of Caraux hippos, with the spinal column of the same fish (No. 13,561 S. I. Coll.). There is also the cranium (No. 11,385) of another and still larger Caraux, the species being unknown. This last specimen presents some points of peculiar interest not so well shown in the first. I have also the cranium of a specimen of Tciithis cosrnlciis, which will be introduced to show certain points ; and finally, the cranium of Poinacanthiis pant (No. 12,770 S. I. Coll.) brought forward to illustrate still other points.

Professor Poey's investigations evidently led him to believe that Gratmnicolcpis was more nearly related to the Carangidce than any other family of fishes known to him. And in this opinion, so far as I can see or am able to judge, I must concur.


The cranium of the Ca^'anx No. 11,385, which bears a veryclose resemblance to C. hippos, shown in Fig. 6, although, be it known, it possesses marked differences, is composed of a bone tissue much more like that seen in the cranium of Graminicoh-pis than any of the other specimens before me. As much as it is unlike it, it evidently approaches the semi-transparent and brittle condition found in our subject. The next thing that our attention is directed to, is the strikingly large orbit of this Caranx, and the evident, though distant, similarity of the elements that go to form its walls. The chief difference we meet with here is the absence in the Caranx of the backward-extending plate of the ethmoid seen in Gj'aniniicolepis, while there is much to support a probable relationship of the forms, in the parasphenoid, the basisphenoid, and less so in the prefrontals, of the two.

Figure 6. — Left lateral view of cranium of Caranx hippos. Spec. No. 13,561, Collection of the Smithsonian Institution; life size, by the author. Letters signify the same as in the other hgures.

Again, in the Caranx, the ethmoidal mass, and parts, which of a consequence associate with it, are produced forwards, and we fail to find anything upon this aspect that in any way reminds us of the curiously truncate appearance of the front part of the cranium in Gramviicolcpis.

Another marked difference is seen in the vomerine element ; this bone, as we have observed, is in our subject merely a kiteshaped scale, in no way incorporated with the parasphenoid, being merely attached around its anterior rim. Now our sped

286 SHUFELDT. [Vol. II.

men of Caranx, No. 11,385, foreshadows in its vomer what eventually comes to pass, between this condition in Gj-anunicolepis and what we find in C. hippos. In this latter fish, as will be observed by referring to the figures Fig. 6, ei seq., the vomer is quite a solid bone, and is moulded upon the anterior end of the parasphenoid, forming a more or less massive termination of this end of the cranium.

In our unspecified specimen of Caranx, this general appearance is likewise maintained ; but upon a lateral view, we are enabled to look in between the vomer and parasphenoid, and the less solid formation of either can at once be appreciated as well as their less intimate relation to each other. It is a shame that this species is not known, nor was ever diagnosed when this specimen of cranium was taken, as this condition is very interesting in the present connection, as are several others, as we shall presently see.

As representatives of the Carangidcc, neither of these specimens develop a spine-like process descending from the postfrontal, which is a very marked feature of that bone in Graminicolepis. It is, however, present in other teleosteans, as seen in one crania of Poniacanthiis and TeutJiis, Figs. 10 and 11.

Before leaving this region of the cranium, I would like to invite attention to the anterior aspect of it, in this very specimen of Poinacanthus (Fig. 10). It approaches to some degree the truncate appearance, so often alluded to in Gramviicolepis ; a closer resemblance, however, is vitiated by the extraordinary forward and upward projection of the vomer in PoinacantJms. Posterior to this bone, in the individual in question, an extraordinary concavity is seen, the sides of which are formed by the prefrontals and parasphenoid, being perforated on either side by a group of foramina. Its bottom is completed entirely by the latter bone.

TcTtthis cccrulcjis offers us in its ethmoid and vomer just the very reverse of this condition, as may seen by a reference to Fig. II.

It may be as well to note in passing that in Poinacanthus pane the parasphenoid is very deep in the vertical direction, being longitudinally excavated above and continuous with the capacious eye-muscle canal, while anteriorly and below it is sharply carinated. Posterior to this carination the bone develops a

No. 2.]



rounded and descending prominence, which is bifid, the two lamina being directed backwards and outwards. The behavior of the anterior end of the parasphenoid of P ojnacantJms has already been described above.

This bone is also wonderfully developed in our cranium of Teuthis (F\g. 11). Here the carination in front is exceedingly deep, while behind it, a distinct descending process is also seen.

Figure 7. — Superior view of the cranium of Caranx hippos, same specimen as in Fig. 6; life size.

Figure 8. — The same seen from below. Letters as in the former figures.

Further, this elegantly developed element makes another curious departure from all the specimens thus far examined. It is this : where it forks over the basioccipital behind, a large foramen is found between the diverging limbs ; this opening is in reality the apex of the eye-muscle canal in this fish, and consequently leads through that fossa to its continuation, which again is the longitudinal excavation on the upper side of the parasphenoid or the floor common to the orbits.

But to return to our cranium of the Caranx No. 11,385 : we find that the anterior and horizontal portions of the frontal bone

288 SHUFELDT. [Vol. II.

are quite transparent at their centres, while raised flutings radiate from their hinder points, forwards and outwards. The transparent areas are found to be even perforated in my specimen of Caranx Jiippos, so thin do they become. Now it will be remembered that in Grammicolepis, from the horizontal portion of either frontal, was developed an upturned, scroll-like projection, the free edges of the two bones meeting in the median line. There was thus formed sort of a conical prominence, the lower part and base of which was anterior, being terminated by the transverse rugosity in front, while the apex, or highest part, seized the free front margin of the supraoccipital crest. In Caranx hippos these vertical portions of the frontal bones are in close approximation, so that they appear to be the continuation forwards of the supraoccipital crest ; the sutural traces, however, have entirely disappeared ; while in the cranium of our other Caranx, the method of formation is very evident from the fact that the vertical frontal plates are not thus coossified, but plainly show their individual origin as well as their relation and connection with the. anterior free margin of the supraoccipital crest, which is wedged in between them. (Compare Figs. 2 and 6, as well as 7 and 3.)

Pomacanthns paru has very extensive rugosities upon its frontal bones, but these latter elements are exceedingly dense and thick, as is the anterior border of the supraoccipital crest in this fish, which measures at its widest part nearly five millimetres across.

Such forms as Pomacanthns parn do not develop conspicuous parietal and squamosal crests ; they are still less manifest in Tenthis. On the other hand, in the Carangidce these crests constitute the most striking feature of the cranium. As already stated above, they are but feebly produced in Grammicolepis, though they are plainly indicated.

All the lateral parts of the cranium, made up of the hinder portion of the parasphenoid, theprootic, opisthotic, exoccipital, basioccipital,- and below the squamosal line, are very much alike in Caranx and our unique subject, more particularly in our undiagnosed specimen of a Caranx. But in the latter the basioccipital enters far more extensively into the formation of the eyemuscle canal than it does in Grammicolepis, as in the Caranx we find a condition existing, as regards the opening between

No. 2.]



the hinder forks of the parasphenoid, very much the same as described for Tenthis coeruleiis.

Upon comparing the posterior views of the crania of Caranx hippos and Gravimicolcpis (Figs. 5 and 9), we find, indeed, but few points of resemblance between them. The occipital crest in the former comes almost down, as it does in PomacantJius, to the supero-median point of the foramen magnum. It is far above it in the latter fish. There the absence of the spreading lateral crests, seen in the Caranx, constitute a marked difference. Professor Poey's fish also has bony pillars developed by the exoccipitals, one being on either side of the foramen magnum. These are absent in the Carang'idcc.


Figure g. — Posterior view of the cranium of Caranx hippos, Spec. 13,561, Smithsonian Institution Collection. Life size, by the author from the specimen. Letters as before.

In C. hippos the facets on the exoccipitals for the first vertebra of the column meet in the middle line ; these parts, however, in G. brachiiisadus have been injured, probably during the first dissection, so that I am unable to say positively upon this point in regard to them. In P. pant the first vertebra of the column coossifies with the basioccipital, but this condition does not obtain in TeutJiis. In this latter form the supraoccipital crest also fails to reach the upper periphery of the foramen magnum at its middle point.

We find that both P. parn and Teiithis have the squamosal curled downwards and forwards in the most extraordinary manner, best marked in the latter type. This is well shown in lateral view in Fig. 11, though I am not sure but that the piece there indicated by sq may not be a separate ossification, in



[Vol. II.

which case it would be a pterotic. I would have to dissect a young fish to decide this point. These processes are veryconspicuous upon posterior view, and of course Graniniicolcpis can show nothing like them.

Pomaca7itJius parii has another condition present, not seen in any of the other forms alluded to above. Just posterior to the prootic, and above the basioccipital and parasphenoid in the cranium of this fish, on either side, we find a subelliptical foramen, with its major axis placed longitudinally, of no inconsiderable size, through which we can easily observe the movable otolith (Fig. lo, otl).


Figure lo. — Left lateral view of the cranium of Pomacanthics parn ; life size, by the author, from specimen 12,770 of the Smithsonian Institute. Lettering as l^efore, with otl, otolith, and c.v., first vertebra of the spinal column, which is here coossified with the basioccipital.

Before concluding our comparison of these crania, we must note another point in the cranium of Pojfiacanthns, and this is, the ossified orbitosphenoids and the ethmoid meet in the middle and interorbital line, immediately beneath the frontals,

We have already fully described above, the relation of these several elements in the subject of this paper, and how any such condition is entirely absent in Caranx. This latter forni, however, may have the ethmoid extended backwards in cartilage, which material may be missing in these dried preparations.

No. 2.]



Now my material on this occasion will not admit of such a thing as an analysis of characters for comparison. In my opinion, without a thorough examination of the entire organization of not only the forms at my command at the present writing, but several others, such a tabulated synopsis, made up at the best from such fragmentary material, would be of but little service to us. The structure of Seriola taken in the present connection would come handsomely into play. Naiicrates ductor would be another good form to examine.

From a comparison of the crania alone, I should say that the relation between Gninimicolepis and such a fish as PoinacantJms pane was very distant, while its affinity with Teuthis coendeus is still more remote. I should have liked, however, to have examined some of the SalistidcB, and perhaps glanced at one or two more of the Chcctodonts.


Figure 11. — Left lateral view of the cranium of a specimen of Tetitkis cceruleus ; life size. Kindly loaned the author by Professor Gill. Letters have the same significance as in the foregoing figures.

Its relationship with the Carangidce, as Professor Poey predicted, is far more evident, though this, too, is extremely indirect, and many forms still unknown to us are required to demonstrate the connection.

These forms must especially show an increased density in the cranial bones ; a decrease in the size of the eye and orbits ; a gradual disappearance of the rugose condition of some of the flat bones of the cranium, particularly the frontals ; a gradual protrusion of the snout ; and finally the development of the parial cranial crests.

Of the shoulder girdle. — Although a number of the bones are



lost, I am enabled at least to present a very good idea of the shoulder girdle in this fish. This is shown in Fig. 12, illustrating this paper, and if the reader happen to have at hand a copy of my osteology of Amia calva, published in the Annual Report of the Commissioner of Fish and Fisheries for 1883, it will be well to compare it with the figure I there gave of the shoulder girdle of Miavptcrus salmoides (PL 14, Fig. 35). It represents these parts as they appear in a typical teleostean fish. A glance is sufficient to satisfy us that the general form of the proscapula {Ps) of Gyaminicolepis is very much like that element in the Bass, differing principally in being slenderer and more sloping, and in its relations with some of the other bones. I am very sorry that I have not at hand the shoulder girdle of a Caranx, as it would be interesting to compare it in the present connection.

As is most usual in teleosteans, the hypercoracoid and hypocoracoid (Fig. 12, Hyp.c. and Hyo.c.) are fused together, and in the present instance, to the proscapula also. The hypercoracoid {Hyp.c.) is pierced by the usual foramen seen in this element among typical teleosts. The anterior projection of the hypocoracoid {Hyo.c.) is long and slender, almost reaching to the extremity of the proscapula {Ps).

It will be noted in Fig. 12 that each of these elements develop a backward, extending process, and the letters Hyp.c. stand between them. This recess harbors the actinosts of the pectoral fin, when these parts are in situ. These pectoral fins have been carefully wrapped up by Professor Poey in a separate little package, and I find three of these actinosts attached to each fin. It does not appear as though any of them had been lost, and I am led to expect that that is the correct number in life. They are composed of rather elementary bone, as is so much of the rest of the skeleton in this curious fish. Now the bone marked 2" in Fig. 12 I take to be th.Q tcleoteviporal, and designated by the same letter in my drawing of the shoulder girdle of the Black Bass. On this latter form, however, as it is also in Amia, the teleotemporal is very loosely attached to the rest of the girdle by ligament, while here in Gravwticolepis, it is represented (7^) by an exceedingly long and slender bone, which has its superior extremity moulded upon the side of the proscapula s, and firmly attached thereto. I fail to find in any of

No. 2.]



the little packages put up by Professor Poey anything that might answer for a loivcr teleotemporal, T, of my Bass figure. We would, however, hardly expect to find such an addition, where the superior element proves to be so very much elongated.

Figure 12. — Outer view of the principal elements of the left side of the shoulder girdle of Granifnicolepis brachiusculus ; life size, from the specimen. Ps, proscapula; Hyp.c. hypercoracoid; Hyo.c. hypocoracoid; T, teleotemporal. The actitiosts are attached to the vertical border of the recess in which the letters Hyp.c. are contained.

Whatever we may see in the cranium of Grammicolepis to remind us of like parts in any of the Carangidcs, such resemblances are certainly not borne out when we come to compare the vertebral column of our subject with the column of a specimen of Caranx hippos. Figs. 13 and 14 are presented to show the extraordinary departures that take place in this part of the skeleton.

294 SHUFELDT. [Vol. II.

Of all the bizarre structures that pertain to the organization of fishes which it has been my pleasure to examine, I cannot recall at this moment one that presents quite so supremely a fantastic arrangement as the eleven or twelve leading vertebr£e in the column of Graiiimicolcpis. These are represented in Fig. 1 3, but I have omitted to include the first vertebra, or that one which is found between the basioccipital and the one shown in the figure with the enormous neural spine. In it the neural spine is not developed, and its connections with the skull are very intimate.

Taken in connection with Professor Poey's account of these parts, this figure obviates the necessity of my presenting a verbal description of any great length, as all the details can be plainly studied without any such additional assistance.

I am inclined to think that the bony pillars, which I described in a previous paragraph, found on either side of the foramen magnum, and completely fused with the exoccipitals, are the halves of the neural spines of this first vertebra of the spinal column. To support this view, we find by placing this vertebra in position, that their pedicles spring from the centrum as in other vertebrae, and that, moreover, the sculpturing on the external surfaces of these pillars is precisely like that upon the sides of the neural spine of the second segment of the column.

This last process is very strong, and quite firmly attached to its centrum : it curves gracefully first backwards, then upwards, in a gentle curve, as shown in Fig. 1 3.

The succeeding four neurapophyses are inclined well backwards, each one, as we advance in that direction, becoming shorter, more slender, and with a gradual disposition to assume the vertical attitude. This is nearly accomplished by the neural spine of the next segment following, or the seventh vertebra. Fig. 13 shows, also, the eighth, ninth, tenth, and eleventh vertebrae, and, as will be seen, the neurapophyses of these segments actually lean forzvards. The one on the twelfth, not here shown, is nearly vertical again, while after that, they gradually incline backwards. The broken spines on the last two vertebrae of the figure I have restored by dotted lines.

Now a glance at Fig. 14 is enough to convey to us that the arrangement of these neurapophyses are entirely different in Caranx hippos. In this latter drawing the first vertebra is

No. 2.]



shown, and it has a vertical neural spine movably articulated with its centrum. The succeeding spines, firmly fused with their centra, gradually become slenderer, longer, and more inclined backwards, to again become nearly vertical in the midseries of the column, to incline once more as we approach the caudal end.

In GraniDiicolcpis the ventral parial apophyses at first support the freely articulated ribs with their attached epipleural appen


Figttre ij, — Left lateral view of the anterior end of the vertebral column of Gratnmicolepis brachitisculus, the first vertebra removed; life size from the specimen.

dages, but in the seventh vertebra the capitulum of the rib completely fuses with the apophysis, and as the latter lengthens, the two become indistinguishably blended. The first and second vertebrae do not support ribs, and in the third pair only, they articulate high up on the sides of the centrum, at the base of the neural arch.



[Vol. II.

The appendages have been lost in my specimen of Caranx hippos, so I am unable to say anything about them.

When we come to compare the existing differences in the latter halves of the spinal columns of these two fish, we find that they are quite as great as those shown in Figs. 13 and 14, Indeed, I can see there nothing to indicate that the forms in question have any relationship whatever and were they in the

Figure 14. — Left lateral view of the first eleven vertebrae of the spinal column of Caranx hippos ; specimen 13,561, of the Smithsonian Institution collection; life size, from the specimen.

class Avcs, I should feel justified in placing one at one end of the system and one at the other.

The absence of the Jiypiiral spine in Grammicolcpis has already been commented upon by Professor Poey in his memoir at the head of this article ; this apophysis is quite manifest in the genus Caranx, as in most true teleosteans.





The characters of the hyoid bones of many Batrachia have been described by Dr. Parker ^ and Professor Wiedersheim,^ and their otic elements have received much attention from the same authors. In both fields, however, much remained to be done. The otic elements of the Salientia have been extensively described by Parker, but no especial attention has been devoted to those of the Urodela by either author, except incidentally to other objects. In the present paper I desire to call attention to these elements in the Urodela, and to contribute thereby to the general theory of the morphology of the inferior arches of the skull. Whether light be thrown on the vexed question of the homologies of the suspensors of the inferior arches of the skull or not, it is desirable to see what this intermediate group of vertebrata contributes to its solution. I go over the types seriatim, commencing with the lowest.


In the genus Trimerorhachis (Cope) there appears to be a distinct opisthotic bone in addition to the intercalare, as in the fishes. In form it is much like the prootic, but in reversed position ; its anterior thin edge forming a suture with the posterior thin edge of the latter. The two together form on their inferior surface a spout-like groove, which extends outwards,

1 Read before the United States National Academy of Sciences, April, 1888. Abstract published in American Naturalist, 1888, p. 464.

2 On the Structure and Development of the Skull of the Common Frog, Transactions of the Royal Society, London, 1870, p. 137; On the Structtire ajid Development of the Skull in the Batrachia, loc. cit., 1875, p. 601 ; Ow the Structure atid Development of the Skull in the Urodelous Amphihia, loc. cit., 1876, p. 529.

^ Das Kopfskelet der Urodelen ; Morphologisches Jahrbuch, 1877, pp. 352, 459,


COPE. [Vol. II.

terminating at the narrow truncate extremity of the two bones. At the base of the opisthotic, and between it and the parasphenoid, is situated the fenestra ovale. This is closed by the extremity of a columella auris, whose proximal part at least may be homologized with the stapes, since no other element corresponding with the latter is visible. The columella is then directed outwards, backwards, and upwards to the notch which is formed between the adjacent borders of the intercalars and suspensorium, where it terminates without having displayed any segmentation. This notch, which is present in all the Permian Batrachia known to me except Acheloma, may have been occupied by a membranum tympani, and that the Ganocephala had, like the Salientia, distinct external organs of hearing, thus differing from the Proteida and Urodela.


The only genus in which I have observed ossicula auditus is Zatrachys (Cope). Here the parts resemble nearly those described in Trimerorhachis. The columella ^ is curved outwards and backwards, and terminates at the notch external to the os intercalare, which was, I suspect, covered by a membranum tympani as in the Salientia.


In Necturus (PL XXII., Fig. i) the stapes is osseous and has its columella directed abruptly forwards. It articulates with a corresponding process of the squamosal bone, which extends posteriorly to meet it from the posterior bone of the latter. The quadrate part of the suspensorium of the mandible is extensively osseous. The distal extremity of the ceratohyal is not articulated with anything, but is connected with the quadrate by the hyosuspensorial ligament, and with the angle of the mandible by the mandibulo-hyoid ligament, as has been pointed out by Huxley.2

In Proteus the relation of the stapes to the squamosal is similar in general to that in Necturus, but the connection

^ Transactions American Philosophical Society, 1886, p. 290.

- Proceedings Zoological Society, London, 1874, p. 192, PI. XXIX. Professor Huxley does not describe or figure the relation of the stapes to the squamosal bone.


between the anterior process of the former and the posterior process of the latter is accomplished by an intervening bit of ligament. The quadrate cartilage is little or not ossified in Proteus. The ceratohyal is not articulated at its distal end, but the latter is attached to the squamosal by two ligaments, the superior and inferior hyosuspensorials (PI. I., Fig. 2). The attachment is higher up than in Necturus, as the ceratohyal is longer.



In Cryptobranchus the columella of the stapes is directed forwards, and terminates in a cartilaginous stem. This is articulated with the suspensorium of the mandible at its proximal part, at the line of junction between the squamosal bone and the quadrate cartilage. The latter is not ossified. The distal end of the ceratohyal is not prolonged, and it is connected with the distal half of the posterior border of the quadrate cartilage by a wide hyosuspensorial ligament. This ligament is interrupted by a subtriangular cartilage, the hyosuspensorial cartilage.

I have not examined the genus Megalobatrachus.

A mphiiunoidea.

In Amphiuma the stapes is lateral in position, and its short columella is directed outwards. It is continued as a cartilage to the truncate posterior apex of the osseous quadrate bone, with which it articulates by a suture. The quadrate is extensively osseous. The distal extremity of the ceratohyal is long and free, and is connected with the middle of the posterior border of the quadrate by an elongate hyosuspensorial ligament.


Merrem Pseiidophidia De Blainville. Gymnophiona Miill.

In Typhlonectes (compressicaudus) the stapes is osseous, and is lateral in position. Its columella is short, and is directed forwards, and is connected by ligament with the posterior border of the quadrate. The distal end of the ceratohyal is entirely free from the manibular suspensorium.

300 COPE. [Vol. II.

In Dermophis (mexicanus) the stapes is lateral in position, and is osseous. Its columella is robust and osseous, and extends forwards, abutting against the posterior border of the quadrate, with which it forms a close movable articulation. The quadrate is completely osseous, and is freely articulated proximally with the cranium. The ceratohyal is free, and not connected with the suspensorium.


De Blainville ; Myctodera, J. Miiller.

This extensive group is most conveniently considered by families.

In the Amblystomidae the columella of the stapes is replaced by the stapedius muscle. This is directed posteriorly, and away from the suspensorium. The ceratohyal is short distally, and its extremity is articulated to the distal part of the posterior border of the quadrate cartilage. The quadrate cartilage is ossified distally, but not proximally.

In the Plethodontidae the stapes has the same character as in the Amblystomidae, except that in some specimens a slender cartilaginous process is seen to be directed towards the quadrate cartilage in Plcthodojt ghitinosus and Spelcrpcs mber. This is probably the persistence of the larval condition of both this family and of the Amblystomidae. In the larva (PI. I., Figs. 7, 9) the columella of the stapes is directed forwards, and is connected with the proximal part of the quadrate cartilage by a short cartilaginous rod. In the larva of CJiondrotus tenebrosus the connection is completed for a short distance by ligament. Thus the larvae of these two families present the character of the mature members of the suborders Trematodera, Amphiumoidea, and Pseudophidia. The quadrate in the Plethodontidae is ossified in its proximal part only. The ceratohyal has its distal extremity curved forwards, and articulated by distinct suture with the quadrate cartilage or bone ; in Plethodon ghitinosus it is inserted in a fossa (PI. I., Fig. 14).

In Desmognathidae the characters are as in Plethodontidae.

In Salamandridae the position of the stapes is as in the previous families, but the relations of the extremity of the ceratohyal are as in the larvae of those animals, or as in the


Amphiumoiclea. The extremity is connected with the quadrate bone by a hyosuspensorial hgament. In Salamandra the ceratohyal and the Hgament are of moderate length.

In the Pleurodelidae the arrangement is as in the Salamandridae. In some of the species the ceratohyal is greatly elongate. In the Diemyctylus torosus the free extremity of the ceratohyal extends to the inferior line of the occipital condyle, carrying with it the hyosuspensorial ligament. This ligament is elongate, and arises from the proximal part of the posterior border of the quadrate cartilage (PI. II., Fig. 3). In the D. viridescens this peculiarity is carried still further. The ceratohyal extends to the lateral crest of the exoccipital, and is received into a fossa of its inferior surface, as has been pointed out to me by my friend Dr. E. E. Gait. The hyosuspensorial ligament extends from the proximal part of the quadrate cartilage beneath the ridge mentioned to the apex of the ceratohyal (Plate II., Fig. 4).


In Siren the stapes is osseous. Its columella is replaced by the stapedius muscle, which extends posteriorly. It has no connection with the suspensorium. The quadrate is cartilaginous. The ceratohyal is large and is much produced distally. It is connected with the posterior part of the quadrate, the exoccipital, and the stapes by a wide hyosuspensorial ligament. It is inserted on the anterior side of the ceratohyal opposite the quadrate, and is interrupted by a hyosuspensorial cartilage, as in Cryptobranchus.


Laurenti. Amira Dumeril.

The stapes in this order resembles that of the Urodela. It is an oval disc without distinct process, and gives insertion to a stapedius muscle near its centre. But this order differs totally from the other existing orders, in the presence of a chain of ossiaila auditiis, which extends from the border of the stapes to the dermal membranum tympani. There are three of these, which have been named the interstapedial, the mesostapedial, and the epistapedial. The interstapedial is a bony style with a cartilaginous basis which originates alongside of the anterior

302 COPE. [Vol. II.

border of the stapes, in a shallow cup-like expansion which abuts against the cranial wall. The shaft is cylindric. The mesostapedial is a cartilage which is attached to the narrow extremity of the interstapedial much as an anther of a flower is attached to its filament. The proximal part of this element is shorter than the distal, and is connected with the superior part of the quadrate by a ligament, the mesostapedial. The distal part is deflected at a strong angle to the interstapedial, and is frequently somewhat spatulate by reason of an expansion distally. Its external face is flat and is applied to the internal face of the epistapedial. The latter is a cartilaginous disk which closes the tympanic chamber externally. It fits like a lid on the cartilaginous annulus tympanicus, which extends beyond it all round. The annulus tympanicus is a thin and wide cartilaginous ring with a thickened margin, which is not continuous, but is interrupted at its superior outline. This interruption is occupied by the distal end of the interstapedial, and the proximal part of the mesostapedial. The distal part of the latter extend vertically across the median foramen. These structures have been described and figured by Parker (I.e.), who has found them to be generally similar in all the families of the order. I have examined all the principal types, and give figures of them in the genera Xenopus, Discoglossus, Stereocyclops, Scaphiopus, Bufo, Hyla, and Rana (PI. II., Figs. 7-12; PI. III., Fig. i).

The ceratohyal is slender at the point of connection with the skull. This is just in front of and external to the cartilaginous base of the interstapedial. It is continuous with the cranial wall in some species. It is involved, just distal to its origin, in the annular ligaviciitiim tynipani, which forms the posterior wall of the tympanic chamber; but it has no structural connection with it.

The development of the Saliential skull has been studied by various authors, especially by Duges, Huxley, and Parker. I have examined series of Rana virescens, R. clainata, and R. catesbriana for the purpose of determining the homologies of the auricular bones of this order, by a study of their development. It has been shown that the ceratohyal cartilage is, for the greater part of the life of the tadpole, articulated with the quadrate cartilage, first on its inferior, and then later on its posterior face. Professor Parker believes that it is a dismemberment of the dis


tal half of the third ventral arch of the skull, and that the superior half of the same becomes fused later with the second arch, thus forming the quadrate cartilage as it exists in adult Salientia. On approaching maturity, the ceratohyal leaves this connection, and is attached to the base of the skull as above described. My observations coincide with those of Parker in that the ossicula auditus do not appear until a later period of larval life in the genus Rana. But they appear before the ceratohyal has abandoned its articulation with the quadrate cartilage. They then arise as follows : the epistapedial occupies from the first its normal position as a disk of cartilage at the flexure of the quadrate cartilage. The interstapedial, on the other hand, arises as a bud from the normal position of its base, and gradually extends itself anteriorly. It early appears as cartilage with a short, free membranous extremity. The latter becomes the mesostapedial cartilage. These elements gradually elongate until they reach the epistapedial. For a time they reach no farther than the quadrate cartilage, and they rest on it, as in the Proteida and larvae of Urodela.


From what has preceded, the following conclusions may be derived ; —

First. The relations of the stapes to the quadrate cartilage or bone are of two types in the Urodela. The one is possessed by the Proteida, Trematodera, Amphuimoidea, and Pseudophidia ; the other by the Pseudosauria and Trachystomata. The larval structure in the Pseudosauria, and inferentially in the Trachystomata, is identical with the structure characterizing the adults of the other division. This is confirmatory of the opinion which I have expressed ^ as to the origin of the genus Siren. This is to the effect that Siren is an animal which is descended from a land salamander, and that its immediate ancestor became aquatic again at a comparatively late period of geological time. My opinion was at first suggested by the condition of the branchiae in VQxy young anihials, where they are functionally abortive, and do not become respiratory organs until later in life, the largest animals having the best developed gills. The characters

"^American Naturalist, 1885, p. 1226.

304 COPE. [Vol. II.

of the stapes confirm this view, since they are those of land salamanders, as distinguished from those of aquatic habitat.

Second. There are also three types of relation between the ceratohyal arch and the skull. In the one there is no connection between the two, as in the Pseudophidia. Secondly, the connection is by ligament. This is seen in Proteida, Trachystomata, and all Pseudosauria except the Amblystomidas and Plethodontidae. The last two families embrace the third type, in which the ceratohyal is articulated by suture with the quadrate. This last type is the most specialized, since the larvae of those families display the connection between the ceratohyal and the skull similar to that seen in the types first and second. Thus the Salamandridae, which are superior to the Plethodontidse in their osseous carpus and tarsus, and opisthocoelous vertebrae, have the hyoid connected with the skull, as in the larvae of the latter.

Third. At a stage in the history of the development of the Salientia, the relations of the stapes and of the ceratohyal to the skull are the same as in a transitional stage of the Urodele family of Plethodontidae ; or, taken separately, the relations of the stapes are those of Proteida, Trematodera, and larval Pseudosauria, while the relation of the ceratohyal is as in adult Plethodontidae and Amblystomidae. This is when the interstapedial cartilage connects the stapes with the posterior face of the quadrate cartilage, and where the ceratohyal articulates with the posterior face of the quadrate at its distal part.

Fourth. It is not probable that the epistapedial forms an integral part of a single primitive element, which includes the other ossicnla auditiis, as it originates independently of the interstapedial and mesostapedial.

Fifth. The interstapedial and mesostapedial do not, at any time in the history of the development of the genus Rana, form any part of the ceratohyal or meckelian ventral arches. As the incus and malleus of the Mammalian ossicnla auditiis are segmented from the proximal parts of these arches, embryology indicates that they are not homologous with the ossicula of the Salientia. From this point of view, the latter form a special line of development, distinct from that displayed by the Mammalia, unless the developmental record has been greatly falsified by caenogeny. From the embryological standpoint, it


follows that the ossictda auditus of the Batrachia Salientia must be excluded from the discussion of the homologies of the Mammalian ossicul^i.

Sixth. But the characters of the Ganocephala and Rhachitomi permit the following reflections, since the latter order is the one from which the Salientia derive their descent. The existence of a well-developed columella auris, which is unsegmented, in the former orders, apparently like that of the Lacertilia, suggests that the segmentation seen in the Salentia is a specialization of later origin. This columella has also the position of the proximal part of the ceratohyal in the adult frog and the larval salamander. As the position of this element in all but the youngest tadpoles is a result of coenogeny, it may be inferred that the ossicula midihis of both the Rhachitomi and the Salientia represent the separated proximal end of that arch, and hence be truly homologous with the incus of the Mammal. The probability that this is the case is increased by the character of this element in the Pelycosaurian genus Clepsydrops,^ — where the columella extends to the cranial wall, leaving the stapes to one side. This is exactly comparable to the relation between the interstapedial and the stapes seen in the Salientia, except that the two elements are not actually connected, as in Clepsydrops. Palaeontology then modifies the evidence from embryology, and renders it probable that the columella auris of the Permian genera, the interstapedial, and the incus are homologous elements, and originated by segmentation from the proximal end of a ventral cranial arch, probably the ceratohyal.

Seventh. It follows from what has preceded, that the condition of the representatives of the ossicula auditus in the Urodela is one of degeneration.

Eighth. It becomes probable, but not certain, from the position of the tympanic disk in the Rhachitomi, at the proximal base of the quadrate bone, that the epistapedial cartilage has originated as a segmentation from the proximal extremity of the quadrate cartilage, and is therefore truly homologous with the Mammalian malleus. This is, however, nothing more than a probability. For a considerable part of the material described in the preceding pages I am indebted to the United States National Museum.

^ See Proceedings American Philosophical Society, 1884, p. 41, PI. I., Fig. 2.




The relations of the quadrate, stapedial, and hyoid apparatus. In Figs, i, 3, 9, 14, 15, and 76, the squamosal bone has been removed.

Figures twice natural size, excepting i, 3, 4, 7, and 8, which are natural size, and 10, II, and 12, which are three times natural size.

Fig. I. Nee turns macidatus ; squamosal removed.

Fig. 2. Proteus anguinus.

Fig. 3. Cryptobranehus allegheniensis ; the middle of the squamosal removed, the extremities remaining.

Fig. 4. Ainphitima means ; a, from behind.

Fig. 5. Typhloneetes cotnpressieaudus ; from the Belize.

Fig. 6. Dermophis mexicanus ; with the quadrate bone turned up, exposing its inferior face, and that of the quadrato jugal; 4«, the same with the quadrate iu normal position. From Mexico.

Fig. 7. Ckondrotus tenebrosus ; larva 250 mm.

Fig. 8. Ckondrotus tenebrosus ; adult.

Fig. 9. Ambly stoma tigrinum ; larva; squamosal removed.

Fig. 10. Amblystoma puneiatum ; adult.

Fig. II. He?nidaetylium scutatum.

Fig. 12. Batrachoseps attenuatus.

Fig. 13. Gyritiophilus porphyriticus.

Fig. 14. Plethodon glutinosus ; squamosal removed.

Fig. 15. Autodax higubris; squamosal removed.

Fig. 16, Spelerpes ruber ; squamosal removed.

louin.Morjih. Vol. II.




D Cofit dtL

ItOi.Anst. V WsTnerU WiniexFra.nkfun'^M.

3o8 COPE.


The relations of the quadrate, stapedial, and hyoid apparatus in Urodela and Salientia. Figures twice natural size, with separate details larger.

Fig. I. Desmognathus nigra; a, stapes separate and enlarged, the squamosal in place.

Fig. 2. Salaniandra maculata \ ; the squamosal separated.

Fig. 3. Diemyctylus (orosus, squamosal removed; a, separate squamosal.

Fig. 4. Diemyctylus viridescens, three times natural size; the squamosal removed. 2a, the squamosal, external side ; b, apex of ceratohyal, with hyoquadrate hgament.

Fig. 5. Siren lacertina \ ; squamosal in place.

Fig. 6. Discoglosstts pictus, partly posterior view; a, ear bones, and origin of ceratohyal, enlarged.

Fig. 7. Bufo lentiginosus americanus, the squamosal removed ; a, the squamosal separate.

Fig. 8. Scaphiopus hanunondii, the squamosal removed; a, the squamosal; b, the ear bones.

Fig. 9. Hyla gratiosa, the squamosal removed; a, the squamosal; b, the ear bones and cartilages in profile, the cartilages of the tympanum divided by vertical section; e, the ear bones and cartilages undivided, external view.

Fig. 10. Xenopus calcaratus, partly from behind, with squamosal in place.

Fig. 1 1 . Stereocyclops incrassatus, squamosal in place ; a, stapes and ear bones and cartilages.

Fig. 12. i?a«tz/r^/?o^a, squamosal in place; a, ear bones and vertically divided cartilages.




E.B.Cojxt del

liXh.Anstv'WeTmtli Winter, FraiikfiiTt'^K



FlG. I. Rana virescens, adult, X 2; a, squamosal bone; b, ear bones without epistapedial.

Fig. 2. Rana virescens, larva with hind legs, and developed fore legs concealed, the skull X 2; «, the hyoid apparatus from below, X 4 Fig. 3. Rana catesbeyana, larva further advanced than that represented in Fig. 2; showing the first appearance of the auditory cartilages at ST and AT.

Fig. 4. Trimerorhackis insignis of the Permian bed of Texas; part of skull fifom below, showing columella at St.; natural size.

Fig. 5. Trimerorhackis insignis basicranial axis from below, without stapes; natural size.

Fig. 6. Zatrachys serrattis, upper posterior part of skull from above; natural size.

Fig. 7. Zatrachys serrattis, inferior view of external part of posterior part of skull of individual represented in Fig. 6; showing columella.


A.T., Annulus tympanicus; B.O., basioccipital; C.Br., Ceratobranchial; C.H., Ceratohyal; CT^r., Cornutrabeculi; ^.5., Epistapedial; ^//i., Ethmoid ; ^jr.C, Exoccipital; Z'./*., Frontoparietal; i¥>«., Hyomandibular; H.Q., Hyosuspensorial ligament; Int., Intercalare; I.St., Interstapedial; y. Jugal; LI., Lower labial cartilage; Mk., Meckel's cartilage; AIx., Maxillary; Afn., Mandible; M.S., Mesostapedial ; O.C, Occipital condyle; P., Parietal; Far., Parasphenoid ; Pg., Pterygoid; Pm., Premaxillary; Pt>3., Postorbitalarch; ^., Quadrate; ^.C, Quadrate cartilage; S., St., Stapes; 6"^., Squamosal; 5/., Superior labial cartilage; T'., Trabeculum; 7/., Position of Membranum tympani.

Cartilage, blue; ligament and membrane, yellow; bone, white.

foiirn.Nofiili lol IJ.



Mx Pmx


D- Cojw (Ul

LitkAnst.vWemi-r&Wmer, Frankivri'^M.


\_Based tipon a comparative study of its skeleton i\ R. W SHUFELDT, M.D., C.M.Z.S.

During the latter part of November, 1885, Surgeon Thomas H. Streets of the United States Navy, then naturalist of the United States Exploring Steamer "Patterson," kindly sent me a skeleton of the Surf-bird, the subject of the present memoir ; and again in August, 1886, the same distinguished officer forwarded me a fine pair of skeletons of this species, representing both sexes, adult. All of this material was collected by Dr. Streets in Alaska, and I am indebted to him also for skeletons of CJiai'adriiis squatarola and Arenaria vielanoccphala, the former taken in San Francisco Bay, the latter at Port Townsend, Washington Territory, and placed at my disposal for comparison with the skeletons of Aphriza.

As additional material for the osteological comparisons the writer here proposes to make, I find I have at my command two skeletons of adult specimens of HcBmatopiis bachmani (Nos. 13,636 and 13,637), belonging to the Smithsonian Institution of Washington ; and finally, in my own cabinet several skeletal preparations of the CharadriidtB and Tringecs, all of which will be of assistance in the work now in hand.

Aphriza virgata constitutes but another one of those forms around which centres so much that is of interest to the systematic ornithologist, owing to the fact that even after more than a mere superficial examination we discover not a little in its anatomy that is inclined to puzzle one, when called upon to pronounce as to its kinship with more or less nearly related groups or types.

Of the first skin that he ever examined, Audubon wrote : " The remarkable bird here represented, which in form and size bears a considerable resemblance to the Knot \Tringa caniiUis\, was procured by Mr. Townsend on the shore of Cape



Disappointment, and proved to be a female." ^ And Dr. Coues remarks, in defining the genus, — " General character of plumage, in its pattern of coloration and seasonal changes, as in Tringecz. One species, a remarkable, isolated form, perhaps a plover, and connecting this family with the next by close relationships with Strepsilas, but with hind toe as well developed as usual in Sandpipers, and general appearance rather sandpiper-like than plover-like. ApJirizince might go under Hcsmatopodidce next to Strepsilas ; or, perhaps better, Aphriza and Strepsilas might together constitute a family AphriziDjE, next to, but apart from, HcematopodidcB!' ^

Two years after the publication of the doubts here candidly expressed by this eminent authority, the American Ornithological Union issued its official Check-List, wherein the arrangement proposed was adopted ; and we find the AphrizidcB with its one species, A. virgata, including the sub-family ArenariincB, the Turnstones, standing between the Plovers on the one hand, and the Oyster-catchers on the other,^ an arrangement which probably meets with the views of the majority of avian taxonomists of the present time.

Believing that under these circumstances a full account of the skeleton of Aphriza will prove an acceptable contribution to the literature of the family, as well as a useful work to the better understanding of the osteology of the LimicolcE generally, it was in view of this that I undertook the task. So far as my material at hand will admit of it, the study will be made thoroughly comparative, and in the end it may go to show that, for osteological considerations at least, the position selected in the system by ornithologists may be the most natural one to create for our Surf-bird, or a somewhat different light may be cast upon the problem ; in any event we are satisfied that more light in such matters is a thing most to be desired, and this alone morphology has the power of affording.

The Skull. — When we come to submit the skulls of adult specimens of certain true Tringea; to moderately prolonged maceration in water, we invariably find that the premaxillary bone detaches itself, as does the dentary of the mandible, leaving in

1 Birds of America, Vol. V., pi. 228.

2 Coues, E, Key to North American Birds, 2d. ed., p. 605. 1884.

^ Check-List of North American Birds, pp. i6o-i66. New York, 1886.


the first case the long, thread-Hke extremities of the palatines and median processes of the nasals freely projecting forwards, and in the latter instance, a similar condition of the mid-ramal elements of the lower jaw. Tringa, Actitis, and others have their skulls behave in this manner, under similar circumstances, and it is quite characteristic of them ; whereas, on the other hand, macerate the skulls as long as we may, no such detachment ever takes place in the Surf-bird, nor in Turnstones and Oyster-catchers, nor in such forms as Gallinago and Philohela. Indeed, this peculiar feature seems to be restricted to the more typical Sandpipers, and it is worthy of our notice that in Aphrisa it never happens, and I may add that one never observes it to be the case in any of the Plovers.

Aside from this fact, the general fades of the form of the superior osseous mandible in Aphriza, carried back as far as the rhinal chamber and frontal region, is more as we find it in such a species as Actitis, for example, or some of the genus Tringa, than it has any semblance to either Arenaria, or much less HcEinatopus. Even when we come to compare the individual parts, this assertion holds true, including in the observation the extensive outline of the triangular aperture, on either side, of the external nares ; and the superior ends of the fronto-maxillary processes of the nasals not overlapping, mesiad, the outer and juxtaposed margins of their premaxillary processes ; which latter condition obtains both in the Turnstones and in Oystercatchers. Again, in Aphrisa we note that the extremity of the superior mandible in the skull is somewhat tumefied, and marked over with minute pits, as we find it to be in Gallinago. This character is absent in such a form as Actitis and other Sandpipers, nor do we find it to be present in Charadriiis, Arenaria, and Hcematopiis.

In Bachman's Oyster-catcher {H. bachmani) the superior osseous mandible is about seven centimetres long, the remainder of the cranium being something over 3.5 cm. long, measured along the median line. Taking these same distances in the Black Turnstone, we find the first to be 2.5 cm., and the latter 2.3 cm.; but barring this difference in proportion between beak and cranium in these two species, the superior osseous mandible in the Turnstone is almost the perfect miniature of that part of the skull in an Oyster-catcher. We are to note, however, one

314 SHUFELDT. [Vol.11.

Other difference, and that is, in Hcematopus the anterior extremities of the palatines fuse across the median interspace much further back than they do in the Turnstone, but essentially, as I say, the plan of these parts in the two birds in question is the same, and quite different from the corresponding structure in Aphriza. Plovers in this particular seem to stand between the Surf -bird and Arenaria.

Taking next these several skulls upon their superior aspects, and comparing them with the same view in A. virgata (Fig. i), we find that the latter bird again agrees with such a Sandpiper as Actitis niaciilaria, or with such a species as T. minutilla. The cranium being very narrow in the frontal region between the superior orbital margins, while the vault of the brain-case is externally rounded, smooth, and ample, being occasionally marked in some of the Sandpipers by a median, longitudinal furrow.

In both, the lacrymals jut out about in the same proportion, and much in the same manner, as they do in a Turnstone. Aphriza and Actitis show scarcely any tilting upwards of the superior orbital peripheries, a feature so very conspicuous in Charadriics sqnatarola ; less so in Hcentatopus, and again quite absent in Arenaria.

Oyster-catchers exhibit their strong larine affinities in their prominent out-jutting lacrymal bones, and still more in their deep and sharply defined glandular fossae over the orbital roofs. These latter depressions are but fairly well marked in Turnstones, while in A. virgata, again agreeing in this respect with most, if not all true Tringece, they are entirely absent.

Passing to the basi-posterior aspect of the cranium, the Surfbird is seen to possess, in common with all North American Limicolcs, a. well-marked supraoccipital prominence, which shows one on either side, the supraoccipital vacuities being larger in this species than they chance to be in any Sandpiper at my hand. Plovers have them next largest, they being very minute in Arenaria, and not present in any of the skulls of the genus HcBmatopus before me, although these latter have large supraoccipital prominences.

In this last-named genus, too, the " occipital area " is sharply defined by a very strong and raised crest, of a subcircular outline. This crest is very much mellowed down in Charadrius ;


more manifest again in the Turnstones, in which genus the occipital area is proportionately much smaller ; and finally, in Sandpipers and ApJiriza the crest is present only as a raised line, which in them also bounds an occipital area of a subcircular outline, the extremities of the boundary line terminating, on either side, at the apex of a paroccipital process.

Most limicoline types have the plane of the foramen magnum nearly coincident with the horizontal plane ; and so it is in all the skulls of the species now before me. But the outline of this great foramen of the occiput differs considerably in the several genera, being perfectly cordate in the Surf-bird, with the prominent posterior apex standing well in between the pair of vacuities of the supraoccipital prominence, which condition is enjoyed by most Tringecs; while \n Ai-ejiaria and the Plovers there is a slight tendency for this aperture to become rounder, its circularity becoming quite evident in Hmriatopiis. In all these species, the condyle of the occiput is proportionately very small in each and every species, it being of a hemispherical form, and unnotched.

So in general pattern, taken as a whole, the basi-posterior aspect of the cranium is more nearly alike in Aphrisa and any true tringine form ; while in Turnstones the approach is towards the Oyster-catchers. And Plovers in this particular seem to maintain sort of a mid-position between the two latter groups. Hcsmatopiis is notorious for the restricted confines of its basi-temporal area at the base of the cranium,^ and this space is never very wide and deep in any of the Liniicol(2.

Figure 3 of my drawings in the plate shows the skull of A. virgata upon direct lateral view, and here we have a number of points to examine. First, it will be observed that \h.Q. pais plana is of a quadrilateral outline, thoroughly ossified, and with its superior and inferior borders not in contact with either the orbital roof above or the maxillary bar below. The out-jutting lacrymal sends down a spicula of bone, which diminutive process turns backwards to have its lower end anchylose with the pars

1 Had not the writer already fully figured the skeleton of the Black Oyster-catcher (which figure and its description has been accepted for publication by the yournal of Anatomy of London), I should have taken pleasure in presenting drawings of it in the present connection, as it constitutes a wonderfully interesting form for comparison in this way with species now under consideration.

3i6 SHUFELDT. [Vol.11.

plana, on its anterior aspect at a point below, and to the inner side of its externo-superior angle. I am thus particular in describing in detail the point of attachment of this extremity of the descending process of the lacrymal, for the reason that in some Sandpipers it is found to be attached to the very angle above referred to, while just before arriving there it sends forwards another small, pointed spine, the extremity of which latter is free, and directed somewhat inwards towards the rhinal chamber. No little variety, however, is to be met with in this particular among the Tringecs, and in such a form as the solitary Sandpiper {R. solitaries) the free end of the lacrymal is attached at a point corresponding to the site of its attachment in Aphriza. With the true Plovers the case is still different, for in them \.h.Q pars plajia is quite small and with rounded angles, while the lacrymal has a long, slender, descending process terminating in a free point below, and likewise has the anteriorprojecting spine coming off from a mid-point of its continuity.^

Areiiaria has a large square /rtTJ plana, and in this form the lacrymal soon fuses at its lower end with the supero-external angle of the osseous lamina in question ; while in Hcematopns the pars plana is again found to be small, and the lacrymals very large, with their descending processes thick and strong, being fully anchylosed with the outer border of the pars plana, and projecting a short distance below that plate, though not arriving at the maxillary bar.

From all this I should say, then, that in Aphriza, Arenaria, and most true Sandpipers thQ pars plana and the lacrymal bones were more or less alike, having much the same relation to each other; but in Plovers and Oyster-catchers these structures are essentially very different ; not only do they differ with each other, but with the aforesaid mentioned types, and in a manner set forth in the last paragraph.

Limicoline birds are notorious for the large vacuities that occur in their interorbital septa, rendering that osseous partition in the vast majority, if in not all the species, a very deficient bulwark between the orbits. Hcematopns has it most entire, while in Aphriza, the Turnstones, Sandpipers, and Plovers, it commonly

1 For an excellent figure of this arrangement of the lacrymal, see Huxley's figure of the enlarged skull of C. pluvialis in his paper, " On the Classification of Birds." P.Z.S., 1867, p. 427, Fig. 7, L.


presents two large deficiencies, or often one imperfectly divided by a backward-extending remnant of the septum springing from the mesethmoid (Fig. 3).

In the Sandpipers {Actitis) the entire anterior wall of the brain-case is, mesiad, wonderfully lacking in bony support ; on the other hand, in the black Oyster-catcher the foramina for the first pair, as well as for the optic nerves, are entire and completely encircled by bone. All the species that we have thus far mentioned in this memoir possess a straight and slender infraorbito maxillary bar of nearly uniform calibre throughout.

Figure i of the plate will show the peculiar orbital processes of the quadrates in the Surf-bird, and indeed they are nothing less than unusually enlarged, thin, and compressed lamina of bone, which in no small degree contribute to the floor of the orbit on either side. We also find the two remaining apophyses of a quadrate to be more than commonly large and stout, and each of these bones has a double mastoidal facet ; two mandibular facets, the inner one being as good as two in Aphriza, though less markedly divided in Turnstones, Oyster-catchers, or Sandpipers ; and finally, a proportionately large pterygoidal facet situated upon the internal aspect of the bone, just above mandibular articulation. Few departures are to be detected from this almost common pattern of the quadrate among the Limicolce; where, too, I may add, this bone is always found to be perfectly pneumatic, as is the greater share of the skull itself among these shore-birds. Still confining ourselves to the lateral aspect of the skull, we are to note the long, spicula-form, sphenotic process possessed by Aphriza, and the less conspicuous and bifurcated squamosal apophysis ; both of which features are far better marked in this species than they are in any of its allies now before me. Turnstones show quite deeply marked temporal fosscE that lack something less than a centimetre of meeting posteriorly over the supraoccipital prominence. HcBmatopus also show these, but here they are strictly confined on either side to the squamosal region, and do not encroach in the slightest degree upon the posterior aspect of the cranium. CJiaradrius likewise faintly displays the same feature, a feature which is so prominently seen in any of the skulls of the LaridcB and their kin.

Careful search through my material fails to detect the presence

3i8 SHUFELDT. [Vol. II.

of such fossae, either in the Surf-bird or among the Sandpipers. Actitus has a skull as free from any such depressions in its temporal region as the veriest Thrush that ever lived, and the same remark fully applies to our least Sandpiper and others of its genus. Another interesting point to notice in the cranium of A. virgata is the marvellous degree to which a tympanic cavity is exposed, as well as the external opening of the Eustachian tube within it. These parts seem to depend upon the large quadrate anteriorly, and the somewhat arching paroccipital process behind, for their protection, in so far as any osseous surrounding defence is concerned.

This condition obtains in varying degree in the allied forms we have under consideration ; it being notably the case in Turnstones and in Hcematopiis, though Plovers and Tringe<z are but little better off in this respect, perhaps the former of these last being the better provided of them all.

And now, finally, turning to the nether aspect of the skull in Aphriza, we are at first struck with the diminutive size of its pair of pterygoid bones. These are short, compressed from above downwards, sharpened all along their anterior margins ; while behind they are characterized by strong basipterygoid processes, which articulate in the usual fashion with the facettes designed for them at the base of the sphenoidal rostrum.

In the black Oyster-catcher, a pterygoid is no longer than the mandibular aspect of the corresponding quadrate is wide, and this statement holds true for the other skulls we have been examining in connection with ApJiriza.

For the Plovers the pterygoids are well shown in Professor Huxley's figure of Charadrins pluvialis} although they are comparatively much smaller in such a species, for instance, as C. sqiiatarola, or others of our North American Charadriidce.

Upon comparing t\\Q palatines of A. virgata with those bones in the allied forms before us, we meet with several excellent distinctive characters, especially when we include in this comparison the maxillo-palatines. In Plovers, Sandpipers, Turnstones, and Oyster-catchers, the postero-internal angles or heads of the

1 Ibid. P. Z. S., 1867, p. 427, Fig. 6. These bones are also represented in my figure of the base of the skull in the Mountain Plover (^. montana) ; Jozir. of Anat. and Phys. London, October, 1883, PI. V., Fig. Z,pg., where their relatively small size can at once be appreciated.

No. 2.] ' APHRIZA VIRGATA. 319

palatines, each turns outward as it articulates with the anterior end of the corresponding pterygoid ; nor do these palatine heads meet each other in the medium plane. With this common feature, ApJiriza agrees ; but when we come to examine the posteroexternal angles of these bones, we find them obliquely truncated in our subject, or at least the hinder ends of the palatines themselves thus obliquely truncated, forming obtuse postero-external angles, which angles are rounded off in Actitus, Cliaradriiis, and in Arenaria, while in Hcsmatopus a moderate, backward-extending and rounded process is found at their site. In this latter species, too, the internal and external laminae of a palatine bone are very conspicuous, as they are, though less so, in the Surfbird, Spotted Sandpiper, and the Black Turnstone. Plovers have this character much subdued. Mesiad, the superior and at the same time, internal borders of the palatines, beyond the "pterygoidal processes," along the rostrum, come in contact with each other in all of the enumerated species ; and in all completely fuse anteriorly with the hinder limbs of the vomerine bifurcation. Coalescence, again, is most thorough in all these forms among the anterior ends of the palatines and those parts of the maxillaries and premaxillaries with which they come in contact, upon either side (Fig. 2). From this description, then, of the palatines, it is clear that A. virgata possesses them in a form essentially its own, and further that the pattern varies in all the allied groups or species.

Not so, however, is this altogether the case with the maxillopalatines. One of these in Aphrisa virgata is seen to be a shell-like process ; extending backwards, concave externally ; the reverse being the case upon its mesial aspect ; perforated by a few foramina ; and finally, 7iot united along its inferior border with the palatiiie of the same side which extends beneath it. Charadriiis also possesses a maxillo-palatine almost identical with this, and in the last-named feature, the true Sandpipers also agree, but in these latter birds the bone is much narrower and more curved.

Passing to the Oyster-catchers and Turnstones, a very different state of affairs is to be met with ; for in Hcematop7{s, for instance, we find a maxillo-palatine to be quite a thick, laterally compressed and unperforated lamina of bone, which completely fuses with the corresponding palatine for the entire length of its inferior

320 SHUFELDT. [Vol. II.

maj'gin, appearing in the adult to be a part of the palatine itself. A similar arrangement is to be found in Arenaria, — so that in this character, Turnstones and Oyster-catchers agree, while the Surf-bird, Plovers, and Tringccs are fundamentally alike ; in the first, the maxillo-palatines have, in the adult individual, coalesced with the palatines, each on its own side ; while in the three last-mentioned families the maxillo-palatines are independent and freely projecting processes.

Aphriza agrees with all the allied types before us, in having a long, slender vomer, its posterior limbs anchylosing with the palatines behind, and a median carination extending for its entire length along its nether aspect. With the exception of Hcsmatopus, the bone in all is pointed anteriorly, and in the excepted genus it is at that end distinctly bifurcated.^

In the form of its mandible, A. virgata substantially agrees with the Plovers and Sandpipers. The upturned, slender process at the angle is well developed, and the inturned ends of the articular cups are quite prominent, showing the usual single pneumatic foramen at each extremity. Two vacuities occur in either ramus, a small surangular one, and a long, semiclosed, splenial one, the sides of the jaw in this region being vertical and rather deep. The symphysis is rounded beneath, concaved above, and equals in length about one-fifth the length of a ramus. In the Black Oyster-catcher the symphysis equals more than a third of the ramus, and is scarcely at all excavated above, while it is decidedly wedge-shaped below. Arenaria also has a deep symphysis to its mandible, but here again it is concaved above and rounded beneath, and the larger of the two ramal vacuities is represented by a very open, subelliptical foramen.

Among the more ordinary types of existing birds, the hyoidean apparatus, as we know, presents but few differences when we come to compare it for the various families. Our subject forms no exception to this statement, and little need be said here in reference to this part of the skeleton in Aphriza. We find the arches alluded to to be very delicately constructed, and have the basibranchials in two separate pieces, the second one being very short and finished off behind in cartilage. The cerato 1 Here we have a departure from the rule laid down by Huxley, who has said that in Schizognathous birds, " the vomer, sometimes large and sometimes very small, always tapers to a point anteriorly." P. Z. S., 1867, p. 426.


hyals largely coalesce, but the ossification proceeds but a limited distance into the glosso-hyal, which latter element in the adult is found to be in cartilage only, Thyro-hyals are notably slender and of considerable length ; they, too, being posteriorly tipped with cartilage throughout the life of the individual. Most Sandpipers possess a hyoidean apparatus similar to the one I have just described lox ApJiriza ; and, indeed, no material difference can be detected for the allied species from other groups.^

Upon carefully examining the special ossifications of the ear in a skull of A. virgata, I fail to detect anything worthy of detailed record ; the form of the stapedial plate and the coossified shaft of the columella have much the same pattern as we find these parts in the Fowl, and so clearly drawn for us by Parker.2 Dr. Streets, in preparing my skeletons of Aphriza for me, threw away the eye-balls, so I am unable to make any observations here upon the sclerotal plates in this species ; they have also been lost from my skeletons of Hcematopics, the Turnstones, and Plovers.

This completes what I have to say here in reference to the comparative osteology of the skull and its associate parts of the skeleton in Aphriza virgata, and such deductions as may be drawn from the study the writer proposes to defer making until the closing paragraphs of the memoir are arrived at, when the skeleton as a whole, in this species, will be passed in review, and the proper comparisons made with the others we have been considering.

Remainder of the Axial Skeleton. — Since the days that anatomists first took to counting the total number of vertebrae in the spinal column of any of our existing Carinatcs, there have been difificulties and differences upon that point ; and as to the differences, these have been but multiplied by the many attempts that have been made to state exactly the num ^ Unfortunately the hyoid arches of both my specimens of Hamatopus bachmani have been lost, and I am unable to add, by way of comparison, anything about them here. I am inclined to believe, however, that if any essential difference is to be hereafter detected in these arches for adult specimens of various representations of the LimicolcE, it may be that some form will show a fusion of the first and second basibranchials ; but I look for nothing more than some such a difference as this, and even it may never be found to exist.

2 Parker, W. K., Morphology of the Skull, p. 258, Fig. 74, and st.



ber of cervical vertebras, and the number of true dorsals, in any particular species. In the case of the total count for the vertebrae in any adult bird of the kind alluded to, the trouble is with the pelvic sacrum, and the skeleton of the tail. In nearly every example it is almost impossible to decide with certainty as to the number of vertebras that have been incorporated in its fusion ; while in the case of the tail, the pygostyle is the stumbling-block, for sometimes in representatives of the same species this bone in one individual will thoroughly appropriate to itself a terminal caudal vertebra, that will perhaps remain free in the other, thus making an additional segment for a bird of the same species. Then, again, it is pretty well agreed that when we come to define the line between cervical and dorsal divisions of the column, we look chiefly to the ribs for assistance ; yet these are by no means to be always relied upon ; as sometimes, in the same species, an additional pair may remain free at the further end of the cervical region, or an additional pair (always at the anterior part of the dorsal division) may connect with the sternum by a pair of hsemapophyses. So that one observer might state that the first pair of free ribs in a certain species of bird were to be found upon the 13th cervical vertebras, while another observer, equally careful, finds an individual of the same species wherein it is the 14th cervical that has the leading pair of free ribs. Thus it is that confusion and doubt creeps into the work.

After having paid no little attention to the osteology of birds for more than ten years, now, I am fairly convinced that there is but one safe or very nearly safe method of arriving at the total number of vertebrce in the spinal column of any particular species of the class, and to do it we should have the prepared skeletons of the young of the species, two, or, better, three, of them, made just at the time when fusion is about to take place among the vertebrae, which eventually go to form the pygostyle in the adult, and when the vertebras composing the pelvic sacrum are all still individualized. Then the count can be made with certainty, and we should not fail to have by us at the time several skeletons of the adult of the same species for comparison. Finally, our remaining difficulty, it is evident, cannot be settled in any such manner, and the only course left to pursue here, is to state, if possible, tJie general rule and tJie noted


exceptions. For instance, we may say in a certain species, it has been found as a general rule that free ribs occur upon the 13th and 14th vertebrae, and that upon the 15th, the ribs connect with the sternum by costal ribs, but examples have been found in the same species where free ribs were found upon the 12th, 13th, and 14th vertebrae ; or again, a specimen of the same species showed free ribs on the 12th and 13th vertebrae, while the ribs on the 14th connected with the sternum. This is by no means a fanciful description, as I have studied cases quite like it, and it only goes to show the utter hopelessness of laying down hard and fast lines in describing the cervical and dorsal divisions of the vertebral column in existing birds. On the other hand, I am inclined to believe that if the total number of vertebrae be counted in the skeleton of the chick of any particular species, the number will be found to be quite constant. Now I believe, after carefully re-counting the vertebrae in the skeleton of the tail in the Mountain Plover (^. montanus), there should be a quaere placed after the number I have given in my memoir upon the osteology of that species,^ as I believe it should be seven, not counting the pygostyle ; further, I am inclined to think I made another miscount, for some reason, in my specimen of Haviatopus, and too many vertebrae have been mentioned as forming the skeleton of its tail in another paper of mine.2

At the present writing I have not at hand the immature skeletons of any of the species we have here under consideration, and consequently am unable to carry out the suggestions made in the foregoing paragraphs, so I can candidly admit that the number of vertebrae given below as found in the sacra of the birds there enumerated must be taken cum grano salts ; although I will add that I counted them to the best of my ability : the process can often be assisted by holding the pelvis up to a strong light, with the dorsal aspect towards you.

In my female specimen of A. virgata, the " epipleural appendages" are present upon the pair of free ribs on the 15th vertebra, the reverse being the case in the skeleton of two males

1 Ibid., p. 97.

2 Shufeldt, R. W., " Osteology of Numenius longirostris, with notes upon the Skeletons of other American ZmzVo/^." yourn. of Anat. and Phys. London. Vol. XIX., October, 1884, pp. 71, 72.



[Vol. II.

of the same species. Charadriiis squatarola seems to have the usual cervical vertebrae (6th to 9th inclusive) modified beneath for the carotid canal ; while in a Killdeer before me, the series so modified includes the fifth cervical. Only \\\&free caudal vertebrae have been counted in the subjoined "Table," omitting in every case to include the pygostyle ; a bone that may include several more, diad perhaps vary for the species here given.










c B 3


  • ^



S ^

w "rt






U u >

u >


i 1






0. c


U >





4) 1






2 6

.Q ■5 (.3




[0 •





J3 3








-a s .


4-* V

° a 5 £i





"0 u



3 C

1^ =1






° 3


.a E








H S!


cS w






Aphriza virgata



15 ••







6th to 9th

Charadrius squatarola .










6th to 9th

Actitis macularia ....










6th to 9th

Hasmatopus bachmani .











6th to 9th

Arenaria melanocephala










6th to 9th

Rhyacophilus solitarius



15 •■







6th to 9th

We find nothing especially noteworthy in the vertebras of Aphriza virgata, as they in a general way are fashioned upon the usual ornithic pattern of those bones among ordinary birds. The axis, and the three cervical vertebrae that follow it, are especially conspicuous for their prominent neural and haemal spines, and their postzygapophyses, which latter extend upwards, outwards, and backwards, as strong processes in the second and third cervicals. This last feature is not so manifest in Aphriza but becomes more so in Arenaria, and in Hceinatopiis arrives at its maximum development, for in this species the postzygapophyses of the axis and the vertebra next behind it are nearly as lofty as the great neural spines upon these vertebrae, giving them the appearance of being tricornuted upon their dorsal aspects. Oyster-catchers have comparatively short parapophy


ses in the cervicals, while in both the Surf-bird and in Turnstones these are quite long and spicula-form in the fourth, fifth, sixth, and seventh vertebrae ; and these last species differ with Hcematopus, in that in this bird the vertebrae of the entire column are unusually large in proportion to the size of their owner. Free ribs in all the species before us possess both capitula and tubercula ; the leading pair being of diminutive size, and gradually increasing as they near the ribs of the dorsal region. The last pair usually do not develop unciform appendages, though they may do so, as in the skeleton of my female Surf-bird, and an Oyster-catcher before me. In the dorsal region only the first two, or at the most, three, leading vertebrae possess haemapophyses, while ossified metapophyses link their transverse processes together above. Aphriza has the neural spines of its dorsals very low, they being more lofty in the Turnstone, and still more so in Hcematopics. Plovers and TriiigecE have them of a medium height. We find few or no especial distinctive differences in the true dorsal ribs of the several species before us, to render any aid as pointing to near or remote kinship in the case of any two species compared. Agreeing with the majority of limicoline birds, these parts are for the most slender, curved as usual, and rather long, with uncinate processes that do not completely anchylose with the several ribs to which they belong. Elsewhere, as I have already said above, I have presented figures of the pelvis of the Black Oyster-catcher, and it can be said here now that this bone in Hcsmatopus very markedly differs from the pelvis either of Aphriza or of Arenaria. Indeed, in some respects, the pelvis of an Oyster-catcher reminds us not a little of the pelvis in Gallus bankiva ; it lacks, however, the propubic spine, and the antero-median borders of the ilia do not fuse with the sacral crista. Further, it is noted for having a certain solidity and thickness of the several bones that compose it, quite in keeping with the other parts of the skeleton of this powerfully built bird. It differs with the general plan of the pelvis in Aphriza (Fig. 21), and in the Turnstones, in that in Hcematopus the ilium forms a ledge, overhanging to some extent the ischium and the ischiadic foramen ; the postpubis is far separated from the lower ischiadic margin, and is truncated almost immediately after passing the ischium behind ; finally, anchylosis is very strong between all the several bones making

226 SHUFELDT. [Vol. II.

up the pelvis in the Oyster-catcher, so that prolonged maceration does not tend to separate them.

One skeleton of Hcsniatopiis at my hand shows a curious dilatation of the latter third, or less, of the postpubis of the left side, not present in any other specimen, nor upon the opposite side of the same individual. All these points and characters are very different in the pelvis of the Surf-bird, wherein this bone is remarkable for its breadth, and for being, as it were, flattened out in the vertical direction, — shallowed ; its sides narrow, and ilia more or less spread. As among Plovers, and most or all Sandpipers, it shows a double row of interapophysial foramina down the sacrum (Fig. 21) ; and the anterior ends of the ilia are subtruncated and embellished with a delicately raised rim, nearly always to be seen in limicoline birds. Between the hinder ends of ilium and ischium, a rounded notch occupies the entire margin of the bone, which notch in H(zmatopus is angulated. Postpubis is in contact, more or less, for its entire length with the lower border of the ischium above it ; and is carried beyond for some little distance to a pointed extremity, much the same being the case in the Turnstones.

The ischiadic foramen is of considerable size, and of a subelliptical outline; this vacuity being unusually large in an Oystercatcher, and such a species as the Killdeer Plover.

Without, then, further entering upon a fuller discussion of minor details, although they have all been carefully noted and weighed by the writer, I would state that in so far as the Hceniatopodid(2 are concerned, the pattern of the pelvis in that group, as represented by the bone from the skeleton of H. bachmaiti, is of quite a different style from the pelvis of either ApJiriza or Arenaria. That when we come to study the pelvis of A. virgata, and compare it with the bone in allied types, we discover that notwithstanding the fact that it exhibits many characters held in common by the limicoline forms generally, still it is impressed with a character of its own, being proportionately broader and shallower than the generality of either charadrine or tringine pelves. It is most nearly approached by the pelvis of a Turnstone, while this latter bird shows in its pelvis a marked shading towards the style of the bone in the Plovers.

Again, Actitis and y^gialitis vocifera possess pelves almost identically alike in their general form ; and in both these


species the postpubis is separated by more or less of an interval from the lower margin of the ischium of the same side, along the middle of its continuity.

Figure 20 of the Plate shows the form upon lateral view of the pygostyle in A. virgata, and that is not far departed from by the representatives of nearly affined Limicolce. In Oystercatchers it varies, because in them the ultimate caudal vertebra usually is coalesced with the bone, giving it the different shape. The Surf-bird and its allies show the diapophyses of the caudal vertebrae to be spreading in the first few leading ones ; then in mid-course they become shorter and bend downwards; then long and more flaring again, to terminate by an aborted one just before coming to the pygostyle. The ultimate two or three may develop bifid haemapophyses, which anteriorly stand between the joints of the centra as chevron bones ; this feature is well marked in H. bachmani.

Oyster-catchers have the neural spines of the caudal vertebrae notched in front, while ApJiriza, CJiaradrhis , and the majority of the TringccB possess these apophyses as plain points.

That time-honored standby of avian skeletologists, the sternum, presents among the birds we have been noticing in this memoir but few insignificant departures from a common pattern, except in the cases of R. solitarius and Actitis.

Aphriza has the bone (Figs 12 and 15) doubly notched on each side posteriorly, the outer notches being fully twice the dimensions of the inner ones, — the outer pair of xiphoidal processes thus formed having a tendency to flare outwards ; otherwise, the outline of the sternal body is quadrilateral, being much concaved above and correspondingly convexed upon its pectoral aspect. Either costal border shows six haemapophysical facettes for the costal ribs, while extensive " ^^^ostal processes " of a triangular outline rear above these in front.

The manubrium is small and sessile, being sharp beneath and blunt above, and directly between coracoidal grooves, thus preventing the coracoids in the articulated skeleton from infringing upon each other ; the reverse being the case in Oyster-catchers and Areiiaria, wherein these bones are in contact posterior to the manubrium, when the former are in situ. Returning to the sternum of Aphrisa, we are to note its ample carina, which extends the entire length of the sternal body, being thickened for

328 SHUFELDT. [Vol. II.

the upper two-thirds of its anterior border, sharpened below, and concaved throughout. Along its lower margin a thickened rim defines its edge, which is outlined by a long, gentle convexity from carinal angle in front to mid-xiphoidal process behind (Fig. 12).

Viewing the bone upon its pectoral aspect, a muscular line — the inter'pectoral muscular line we will call it here — extends, on either side, from the boundary of the costal fossa beneath, to the inner of the pair of xiphoidal notches behind, to the mesial edge of the same. This line occurs sharply defined in both Plovers and Oyster-catchers, where it is similarly drawn, as it is likewise in a Turnstone.

Arenaria has a sternum, which, excepting the fact of its being a size smaller, is the perfect counterpart of the bone in A. virgata, though in the latter species the sternal keel seems to be comparatively a trifle deeper in proportion.

When I published my Osteology of Ntmienius longirostris, I therein stated that the sternum to a skeleton of the Spotted Sandpiper {A. niacularid) possessed two notches upon either side, a large outer pair and a very small inner pair. Those skeletons were prepared and the species diagnosed by a very incompetent man, and at the present writing I am not prepared to vouch for the correctness of the statement. I have had to regret many times since, the confidence I placed in the identification of the skeletons which went to make up the collection of the vertebrate series in the Museum and Library of the Surgeon General's Office of the Army at Washington ; they were the cause of a number of oversights creeping into my avian osteology in early days, which now are being corrected in my more recent papers. So I am inclined to believe that the four-notched sternum of Actitis may be a specimen from the same category; it was "on file" in the same institution. At any rate, we have before us now an excellent skeleton of Actitis of my own collecting, and it possesses but a pair of notches in its sternum, otherwise the bone has the general pattern of the LimicolcB at large.

In a few days The Auk for July (1888) will be issued, and in there, in a letter to its distinguished editors, I point out the fact that the Solitary Sandpiper is another species which has but a pair of notches in its sternum, and suggest that its old


genus, Ryacopliilus, be restored for it (at the present writing it is a subgenus of Totanus). But the majority of Tringecs, as for example, all the true representatives of the genus Tringa, have the "four-notched" sternum, and the fact that Actitis and RyacopJiihis have each only a pair of notches, may be of more significance in so far as the true affinity of those two species is concerned, than some of their external characters (that are so very subject to change from minor causes), but which have alone been relied upon thus far by the systematist, to classify them.

Finally, then, the fact may be broadly stated, that with respect to the sterna of ApJiriza and Arenaria, they are to all intents and purposes the miniatures of that bone as we find it in Hcsmatoptis bachmatii ; and although this is very interesting, the circumstance goes to prove, in the light of other skeletal characters of these species, we would utterly fail were we to rely upon this bone alone as any indication of family, or much less, generic, affinities.

All the sterna that the writer has examined from the skeletons of North American Limicolce were non-pneumatic bones, and if it is ever found to be otherwise, we must believe that it is the exception. In a specimen of Tringa minutilla in my collection, the bone is so thin that a large perforation occurs on either side of the keel in the sternal body, and one through the keel itself, high up in front.

Speaking of the pneumaticity of the skeletons of the several species we are here considering, I am strongly inclined to believe that the skull is the only part in any of them where air gains access to the interior of the bones, and as a rule to no very great extent, comparatively, there.

Turning our attention next to the shoulder girdle (Figs. 11, 16, and 17) in ApJiriza virgata, we meet with a scapula having much the form that that bone assumes among the Shore Birds, generally, being shaped a good deal like the blade of a miniature cimeter, and when in situ in the articulated skeleton bearing the usual relations to the clavicle and coracoid of its own side. A coracoid is conspicuous for its short, sub-cylindrical shaft ; its head being tuberous and much crooked over towards the median plane, while the sternal extremity of the bone is not only thickened from before, backwards, but much expanded lateral-wise, showing at its outer angle a prominent lamelliform process, com

330 SHUFELDT. [Vol. II.

pressed in the antero-posterior direction, and curved somewhat upwards. Os furcida typifies the U-shape form of the bone, and has in all Limicolce, that the writer has ever studied, the merest apology for a hypocleidium, below. Its limbs are of nearly uniform calibre throughout, being slightly heavier above than below, and moderately compressed transversely for their entire lengths. Their heads, that is, the free ends of the clavicular limbs, are much elongated, and when articulated in sihi, reach far back to the heads of the scapulae. Viewed upon direct lateral aspect, it will be noticed that the os furcula is very considerably, though gracefully, curved in the antero-posterior direction, the convexity of the curvature being to the front.

In H(2matopiis the shaft of a coracoid is pierced by a small oval foramen, from before, backwards, and just below the scapular process, that does not occur in either the Surf -bird or among Turnstones.

Now, aside from the exceptions of this foramen, and differences as to the method of articulation of the coracoids with the sternum, already alluded to, and finally, the mere matter of size, — the description of the shoulder girdle here rendered for Aphriza, will apply almost equally well to the shoulder girdles of the TringecE, the Turnstones, and the Plovers, and no doubt to many other Shore Birds. Were we to depend for characters, upon which to base distinctive differences, on the shoulder girdle alone in Aphriza, Arenaria, and a true Plover of a corresponding size, we should surely be disappointed, for the bones of this arch do not offer them in the genera enumerated.

Of the Appendicular Skeleton.

The Pectoral Limb. — The majority of the bones of this part of the skeleton in Aphriza are shown by my drawings in the Plate, as already stated above. The corresponding parts for H(smatoptis and Plovers have been given in other connections.

It will be seen that in the humerus of our Surf-bird, both "radial crest " and "ulnar tuberosity" are strikingly well developed, while the articular humeral head between them is rather small (Figs. i8 and 19). The pneumatic fossa is an extensive excavation, although the bone itself is not pneumatic. As to the shaft, it shows but little curvature, is smooth, and subcylindrical in form. At the distal extremity of the bone, a conspicu


ous " ectocondyloid process" is to be observed; and at the extreme end, the oblique and radial tubercles, or condyles, stand out with more than their usual boldness, a condition which is enhanced by the fact that just above them there exists a fossa of some considerable depth. When the bones of this limb are articulated, and the arm tightly closed, the head of the radius accommodates itself to this latter depression at the distal end of humerus. Upon the anconal aspect of this extremity of the bone, we also find strong and powerfully impressed grooves for the passage of the tendons of muscles inserted upon it, and at the shoulder joint, during the life of the individual.

Neither radius nor ulna exhibit any very great degree of curvature, and as a consequence the interosseous space in the articulated arm is not very wide. The latter bone develops down its shaft a row of some eight or nine papillae, for the quill butts of the secondaries of the wing, a feature common to so many of our Shore Birds. Comparatively speaking, the distal extremities of these bones of the antibrachium in Aphriza, as in all Lwiicol(2 which the writer has examined, are small, though to this fact the Oyster-catcher seems to offer an exception.

The carpal segments are two in number, as usual, and present nothing, either in their size, form, or method of articulation, in any way different from what we usually find.

In the pinion (Figs. 13 and 14), the cai'-po-meiacarpal horiQ is long, large-headed, and with its coalesced shafts straight, and more than commonly close together. As a negative character, we miss the overlapping process at the proximal end of the midmetacarpal, — an apophysis which characterizes this bone in all true GallincE. Prominent and upturned, the pollex metacarpal supports a phalangeal joint that is clawless ; the same being the case with the distal phalanx of the digit of the succeeding finger.

Regarding the expanded and proximal phalanx of this last digit, it is to be noted that its somewhat narrow blade, although more or less scooped out upon either aspect, is never perforated as we find it in some of the Laridce. At its further end a little process protrudes, against which, in life, the head of its following phalanx plays ; this latter joint being very long and slender, as in some Plovers and Sandpipers. The little phalangeal joint of the remaining digit is crooked so as to accommodate itself to

332 SHUFELDT. [Vol. II.

the outline of the posterior border of the blade of the proximal joint of index finger.

Equal correspondence of characters in the pectoral limbs of Plovers, Sandpipers, and Turnstones, as compared with the bones of the skeleton of the arm, in Aphriza seem to prevail, as they were found to prevail among the shoulder girdles and sterna of the same species. Practically, the description I have just given for the arm bones of the Surf-bird will answer all purposes of exactness for the representatives of the genera Tringa, Arenaria, and most, if not all, true pluvialine types. This, of course, does not apply to the mere question of the difference of size, but to the salient features only. Our engaging little T. minutilla has a pectoral limb, the skeleton of which corresponds, in so far as its available taxonomic points are concerned, with the same parts in the Surf-bird, to a degree of nearness most exasperating to the comparative anatomist, and discouraging to one eagerly searching for strong diagnostic distinctions in the structure of such types. Turning to the skeleton of the black Oyster-catcher we find but one sure point of difference that can be relied upon, apart from its size, as distinguishing the fundamental plan of the structure of the skeleton of its arm from the above species. In this bird we find a claw upon the end of the pollex phalanx, which, by the way, is also found in some Curlews.

Aside from this character, then, we might, were it possible, reduce the arm bones of this Oyster-catcher until they equalled in size those of a Surf-bird, and it would puzzle the best of us to correctly decide to which species either belonged, after such a hypothetical reduction.

These limicoline birds all seem to lack the os hiimero scapulare at the shoulder joint, and I have yet to find in any of them the sesmoidal ossicles at the elbow, such as we do find, for instance, in some of the Puffins.

Of the Skeleton of the Pelvic Li? fib. — In Aphriza virgata the femur possesses only a moderately sized, semi-globular head, with barely any depression on the top of it for the insertion of the teres ligament. The trochanter is lofty and thick, reaching conspicuously above the articular summit of the bone, which latter is broad and smooth. Little or no curvature is met with as we descend the shaft, and this is quite cylindrical for its


middle third. As to the condyles, we find nothing worthy of special note, as they are fashioned much as we generally find them in ordinary birds ; the external one being the larger and lower, while it shows behind the usual vertical groove for the head of the fibula.

In the leg the fibula is free, not anchylosing with the tibiotarsus, being but a trifle longer than the half of this latter bone. Its lower two-thirds are almost hair-like in proportion, and as for that matter it would seem that this delicate osseous filament can add next to nothing to the skeletal support for the structures of the limb (Fig. 8).

Tibio-tarsiis also has very nearly a straight shaft, but commonly shows, as in Hcematopiis, a curvature in the transverse plane, being gently bowed outwards towards the fibular side. Its procnemial process is much the larger of the two laminar crests in front of the head of the bone, and it is pointed almost directly forwards, while the ectocnemial process is turned outwards, and curves downwards, as a pointed hook.

They both rear slightly above the articular summit of the proximal end of the shaft. Occupying the middle part of the upper third, the fibular ridge stands out quite boldly at the outer aspect, as in all birds. Proceeding to the distal extremity of the shaft we find a marked difference in the form of the two condyles, the outer one being decidedly uniform in outline, and the inner one more elongated and showing compression in the vertical direction. Between, and at the same time above them, in front, we detect the usual little osseous bar for tendinal confinement. A patella is absent in the Surf-bird, the case, I believe, in all Limicolcs — at least the writer is not aware of any exception to the rule.

Having a length just equal to that of the femur, the tarsometatarsits exhibits at its usual site at the upper end and back part of its shaft a stumpy hypotarsus, that appears to be both pierced and grooved, for the guidance of tendons. Two minute foramina in this locality also perforate the bone in the anteroposterior direction, which is also the case in an Oyster-catcher. A mid-section of the shaft of tarso-metatarsus would show a quadrilateral outline, but the calibre increases, and this form changes as we pass to its distal extremity, where we find three large and prominent trochleae, the middle one of which is the

334 SHUFELDT. [Vol. II.

biggest and placed the lowest on the bone. The ones upon either side of it are directed to the rear. At its usual site, above the innermost one, is faintly to be discerned the facette for the articulation of the os metatarsale accessorius, of hallux phalanx. Already the ornithologist has placed at our command the fact that the arrangement of the joints of pes in this species, as in Arenaria and many Tringecs, is upon the most common plan of 2, 3, 4, and 5 bones from hind toe to outermost digit, respectively, and a study of these in the skeleton reveals nothing peculiar in them.

The Black Turnstone being a somewhat smaller bird than Aphriza, it has a correspondingly smaller pelvic limb and skeleton ; aside from this fact, however, the several bones of that extremity are fashioned quite as they are in our Surf-bird, and present no decided differences in any particular. This statement applies with equal truth to various Sandpipers, and would to Chai'adrius were it not that in the latter genus a disproportion, and a disproportion merely, exists in the hallux, — that toe being much smaller in the Plover, though all the other bones of the limb are remarkably similar in all their essential features, bone for bone, as we have described and figured them for Aphriza virgata. Hallux is also missing in HcEmatopus, where we sometimes find a process of bone anchylosed to the shaft of tarso-metatarsus, at a point where the accessory metatarsal is commonly attached by ligament. In a specimen of H. bachmani before me, such a process is to be found upon the shaft of the tarso-metatarsus of the right pelvic limb, but is missing from the corresponding bone in the opposite limb of the same individual. Strange to say, this is precisely the state of affairs in the skeleton of a second specimen at my hand, — the process is present in the right, absent in the left. It may prove of interest to look further into this point with additional specimens. Returning to our comparison between the pelvic limb of this Oyster-catcher, then, with the same part of the skeleton in Aphriza, we may say in brief that beyond the few exceptions mentioned the limb in the former is simply an amplification of the limb in the latter species, — its bone for bone, character for character, — the same thing over again, only the mould happened to be several sizes larger wherein the skeleton of these parts in our Oyster-cr,l:cher were cast.


Having now briefly passed over the more salient characters to be found in the skeletons of ApJiriza, Arcnaria, Hcematopiis, and several representative Plovers and Sandpipers, we are, I think, in a position to offer our conclusions upon the affinities of the Surf-bird, in so far as they seem to be indicated through this study of its osteology.


Taking the skull as the part of the skeleton, I would invite attention to the peculiar form of the superior osseous mandible in H<zmatopiis; and in the cranium of the same bird to the deeply sculpt, supra-orbital, glandular depressions ; to its bold, outstanding lacrymal bones, and their mode of meeting \\\q pars plana on either side ; to its vomer, which is forked at its anterior extremity ; to the absence of the supra-occipital foramina ; to the complete coalescence of the maxillo-palatine, with the palatine of the same side ; to its cranial foramina in front for exit of the first and second pair, they being completely surrounded by bone ; and, finally, to the subelliptical outline of its foramen magnum. All these points, and many minor ones in its skull, are totally different in ApJiriza, and alone are sufficient to warrant us in at least drawing strong lines of family distinction between these two types. Equally clear is the difference between the skull in HcBmatoptis and Charadrms, but when we come to compare it in the former with the skull of Arenaria, a faint indication of affinity may be made out, which indication is principally seen in the Turnstone, having similarly formed nasals ; feebly developed supra-orbital glandular depressions ; a somewhat similar superior osseous mandible ; and, lastly, in that its maxillo-palatines unite along their entire lower margins with the corresponding palatine on either side below them.

The differences presented to us, however, in the skulls of Hmnatopus and Arenaria, amount at least to family distinction. So, too, when we come to compare the skulls of Arenaria and Charadriiis, they are very different, and undoubtedly belong to representatives of distinct families.

Comparing next the skull of a Turnstone with that of the Surf-bird, we are at first struct*, with the radical differences

336 SHUFELDT. [Vol. II.

exhibited in the superior osseous mandibles of the two species. Then in the former the maxillo-palatine fuses, in a manner already described, with the palatine, and finally, the " crotophyte fossae " are pretty well defined in Arenaria, whereas in ApJiriza not a vestige of them is to be seen. The posterior external angles of the palatines are obliquely truncated in A. virgata, rounded in the Turnstone ; and finally, the mandibles of these two types are essentially very different, both as to form and structure, as in the Turnstone a large vacuity persists between the forks of the dentary, which is nearly absent in the Surf-bird.

Taking it all and all, the skull, mandible, and associated parts of the skeleton of the head in ApJiriza agree more nearly with the corresponding structures in the skeleton of the head of some of the larger TringecB than with any other class of limicoline birds. I have arrived at this point by carefully comparing this part of the skeleton in the numerous species at my hand, and I am fully convinced of the truth of the assertion.

Passing next to the remainder of the axial skeleton, and without enumerating the differences that have been dwelt upon quite at length when under review, it is evident from the arrangement and number of the ribs that Hcematopus differs essentially from the several types we have been examining, and agrees only with Areftaria in the number of its free caudal vertebrae. If we were to be guided in this matter by the total number of vertebrae, as shown by these several species of birds, providing our count is correct, it would seem that in this respect HcsmatopJis, Arenaria, and Aphriza fall together, while some Plovers and some Sandpipers are thus associated. HcBinatopiis has a very different style of pelvis from the other species we have enumerated, while the pattern of this bone is not a little different when compared in ApJiriza and Arenaria. Were we allowed the expression, it might be said that in so far as the pelvis goes in the genera ApJiriza, CJiaradrins, Actitis, Arenaria, and some few others, the pluvialine design seems to be impressed with more or less strength upon them all. This statement will not apply to the pelvis in an Oyster-catcher.

The sternum and shoulder girdle do not materially assist us in this matter; we found it with but two notches in such

No. 2.] APHRIZA V/RGATA. 337

species as Actitis macularia and R. solitari?is, and in the other forms it is of a markedly pluvialine stripe throughout.

As to the appendicular skeleton, a marvellous sameness of style — barring the question of mere difference in size — seems to pertain to the skeleton of the limbs in Surf-birds, Turnstones, Sandpipers, Plovers, and others. Hcematopus, as we saw, possesses a claw upon its pollex phalanx, while some of the species possess but three toes, as in the last-named genus and in Himantoptis and others.

To sum up, then : my comparative studies of the osteology of Aphrisa virgata lead me to believe that in the first place its affinity with HcBmatopiis is by no means a close one, the Oyster-catchers forming a very well-marked and distinct family of limicoline birds by themselves, showing by the structure of certain parts of their skeletons strong larine derivation.

Secondly, the skull in Arenaria is in some respects more upon the plan of the skull in HcEmatopus than it in any way approaches the structure of that part of the skeleton in Aphrisa ; indeed, so far as the skulls are concerned, the skulls in CJiaradriiis, Arenaria, and Aphrisa might be represented as each occupying the angle of an equilateral triangle, each sharing certain characters in common and each removed from the other, equally distant by an equal set of others, not possessed by either of the other species. Were the skulls of HcBmatopiis and an average Trijiga introduced into this hypothetical figure, the skull of the Oyster-catcher would be, as far as the characters they hold in common are concerned and for no other reason, on the line joining the skulls of the Turnstone and Surf-bird, and considerably nearer the first species than the last. The Sandpiper's skull would be on the line joining the skulls of Aphrisa and the pluvialine type, and considerably nearer the former species than the latter. Aside from the skull, the remainder of the skeleton of Arenaria would answer fully for the skeleton of an equal-sized Plover, but hardly so for A. virgata.

To conclude : the sum total of the skeletal characters of Aphrisa virgata place this species nearer the TringecB — say some typical large-sized Tringa, with a four-notched sternum — than it does to the Plovers ; less so to Arenaria, and far less so to H(B7natopiis. Some pluvialine forms, however, are not so dis


tantly removed from ApJiriza, if we may be allowed to judge from osteological premises.

There is not the slightest doubt in my mind that had such a form as Charadn?ts sqnatarola been the sole representative of the Charadriidce in the entire world's avifauna, and been found taking the place, the habitat, of ApJiriza, modern systematists of the class would surely have grouped it with the Turnstones as forming a family ApHRiziDiE. We must believe that the fashion is far too prevalent, that when an isolated form like Aphriza is met with, to lay about and endeavor to hit upon a family, a corner, where best to stow it away, without creating, as in the majority of cases should be done, a separate family for it.

Now Aphriza is no more nearly related to Arejiaria than is CharadriiLS related to Arenaria, and as at present classified by American Ornithologists, by no means indicates its true position in the system. I would propose then, in view of the anatomical facts brought out in the present memoir, that the family Aphrizidte be made to contain the sole representative of it, Apkrisa virgata ; and that another family, the Arenarid^, be created to contain the Turnstones.

It would seem that ApJiriza in some way connects the Plovers with such TringecB as may possess the hallax and a four-notched sternum ; while Arenaria links the Plovers with H<£matop2is, albeit that the Turnstone is much nearer the Plover than the far more remotely affined Oyster-catcher would in any way appear to be.




Note. — The figures in this Plate were all drawn by the author from the skeleton I

of a single individual, an adult male, and are all of natural size. I

Fig. I. The skull of ^. z'zrg'a/a, seen directly from above; mandible removed. '

Fig. 2. The skull of ^. z^e>^<z/«, seen upon basal aspect; mandible removed. ,

Fig. 3. The skull of A. virgata, seen upon direct lateral view ; right side, includ- ing the mandible. .

Fig. 4. The mandible of A. virgata, direct superior view. '

Fig. 5. The right tarso-metatarsus oi A. virgata, anterior view. |

Fig. 6. The right radius of A. virgata, palmar aspect. !

Fig. 7. The right ulna of A. virgata, palmar aspect. ; Fig. 8. The right fibula of A. virgata, inner side.

Fig. 9. The right tibio-tarsus of A. virgata, mesial aspect. '.

Fig. 10. The right femur of A. virgata, anterior surface. 1|

Fig. II. The os furcula of ^. virgata, seen obliquely from the right side. }

Fig. 12. The sternum of A. virgata, seen upon right lateral aspect. j

Fig. 13. The right carpo-metacarpus of .,4. virgata, anconal surface; poUex pha- '.

lanx in situ.

Fig. 14. The right proximal phalanx of the index digit of A. virgata, anconal 1


Fig. 15. The sternum of .^. virgata, seen upon direct anterior view. 1

Fig. 16. The left coracoid and scapula of A. virgata, outer aspect, and articula- ]

ted in situ. |

Fig. 17. The right coracoid of A. virgata, seen upon direct anterior view. j

Fig. 18. The left humerus of ^. virgata, palmar aspect. j

Fig. 19. The left humerus of A. virgata, anconal aspect. \

Fig. 20. The four last vertebrae of the skeleton of the tail of A. virgata, and the |

pygostyle, seen upon the right lateral aspect. i

Fig. 21. The pelvis of A. virgata, dorsal view. \

Jonrn . Moi]ih . Vol. 11.


Jt WS},uMat,iJ-,ut iel

BMeisellilh Boston.

Jijihriz a x'ifcjata, ad.^.xl.

Volume II.

April, i88g.

Number S.




CHARLES-SEDGWICK MINOT, Harvard Medical School, Boston, Mass.



PAGE 341

§ I § 2.

§ 3 § 4.

§ 5 § 6.

§ 7 § 8.

§ 9 § 10.

§11. §12.

I. Rabbit.

Uterus at six days Uterus at seven days Uterus at eight days . Uterus at nine days . Embryo at nine days Uterus at nine days and

teen hours . . . Uterus at eleven days Embryo at eleven days Uterus at thirteen days Uterus and embryo at


Summary ....

page 343 343 345 348 353

357 359 365 369

372 376 Comparison with other rodents 378


IL Man.


§ 13. Allantois and umbilical cord . 380

§ 14. Amnion 385

§ 15. Chorion 389

§ 16. Uterus during menstruation . 413 § 17. Uterus one month pregnant . 416 § 18. Uterus seven months pregnant 421 § 19. Uterus after abortion . . . 428 § 20. Origin of decidual cells . . . 429 § 21. General considerations - . .431 § 22. Summary 435

23. Bibliography 438

This paper arose not as the result of a special investigation, but as the outcome of general studies undertaken in connection with the preparation of a Treatise on Human Embryology. For

342 MINOT. [Vol. II.

such a work, knowledge of the foetal envelopes seemed especially important. I was thus led to examine them, and found in so doing that the structure of the parts differed in many respects from what had been assumed. The rabbit's uterus was examined in the hope of obtaining light as to some of the changes in the human uterus, but the differences are so great that little help was gotten ; but on the other hand I was brought to a conception of the changes in the rabbit's uterus so fundamentally different from the views of previous writers that I was induced to carry my observations far enough to make sure of the essential alterations. The following communication is therefore in no sense monographic, but only supplementary to the work of others. My own work has been accomplished by the aid of a grant from the American Association for the Advancement of Science, without which it could not have been carried out. The recipient of such aid naturally wishes to publish what he has accomplished, since such a publication is the most fitting acknowledgment of the assistance enjoyed. I feel my obligations to the Association the more deeply because the grant is the first made from its Research Fund. May I express the hope that that fund will be largely increased, and the Association enabled to make numerous grants to other workers, for in so doing it will do more for the promotion of science than, I believe, by any other means whatsoever.


The observations here recorded were made upon pregnant uteri of the rabbit at various intervals from the sixth to the fifteenth day of gestation, both inclusive. The uteri were cut out carefully, stretched very slightly, and the ends of each uterus tied to an iron rod ; the specimens were then hardened in Kleinenberg's picrosulphuric acid, according to the directions given in Foster and Balfour's Embryology, 2d ed. pp. 425, 426. Although this reagent worked fairly well, and preserved the histological elements of the uterus and of the older embryos satisfactorily, it failed to preserve the blastodermic vesicles in uteri of six and seven days ; and in the older specimens, after hardening, the extra-embryonic foetal membranes were somewhat rumpled. Owing to the great difficulty of obtaining doe-rabbits


in Boston, and their consequent high price, I have been unable to experiment with other methods of hardening. The specimens after hardening were for the most part stained in toto with alum cochineal and eosine, imbedded in paraffine and cut into serial sections with an automatic microtome, made by Herr G. Baltzar of Leipzig.

§ I. Uterus at six days and three hours. — The position of the ova is recognizable externally, being marked by a very slight protuberance on the free side of the tubular uterus. Transverse sections show that there is a considerable dilatation of the uterine cavity, corresponding to the swelling ; the walls are considerably thinned out by stretching. The glands are much altered, otherwise there is no striking change in the uterine structure. The shape of the glands varies, but everywhere their cavities are very much expanded, and the epithelial linings of adjacent glands are separated only by very thin connective tissue partitions ; on the side of the mesentery the glands are distinctly tubular, and grouped on folds of the mucosa ; the relations of these folds are described in the next section. On the opposite side of the uterus, that is, away from the mesentery, the glands are short, with wide cavities, constituting a series of irregular ampullae with wide mouths. The epithelium is thickened everywhere ; it stains deeply, and has enlarged nuclei ; it has many intercellular vertical fissures, and therefore a good many of the cells are separated from their neighbors ; adjacent cells project unequally, rendering the surface of the epithelium irregular. The change in the epithelium is greatest opposite and least near the mesentery, but is everywhere similar in kind, though varying in degree.

I am unfortunately unable to state anything in regard to the relations of the ovum, owing to the failure of its attempted preservation in my specimens.

§ 2. Uterus of seven days and three hours. — The placental swellings are well marked as smooth, rounded bulbs only a little larger in diameter than the unaltered uterus between the swellings, and not projecting at all on the mesenterial side of the uterus. Transverse sections show at once the changes which have taken place. As is well known, the rabbit's uterus has six longitudinal folds, symmetrically disposed ; the line of insertion of the mesentery (mesometrium or broad ligament) corresponds


MINOT. [Vol. II.

to the space between two folds, which alone participate in the formation of the placenta ; accordingly we may designate them as the placental folds. In the region of the swellings the placental folds are already hypertrophied, and form a marked contrast to the opposite side, where the folds have completely disappeared, and their glands have become shorter and somewhat contorted ; the two lateral folds are intermediate in appearance. In the placental folds there is a great increase in the connective tissue, which consists solely of anastomosing cells, forming a loose meshwork of very granular protoplasm, of which only a small amount is accumulated around each nucleus. Described in other words, the cells are small, granular, with long processes continuous with those of adjacent cells. The glands extend only a short and nearly uniform distance down from the surface of the folds ; the glands themselves are somewhat dilated ; their epithelium stains deeply ; its free surface is quite irregular ; the nuclei are greatly increased in number, and lie crowded throughout the whole thickness of the layer ; the nuclei are round or oval in outline, with a well-marked reticulum densest superficially. In many places the nuclei are grouped, three, four, or five together, and sometimes one can distinguish a distinct outline around the group. These appearances I interpret as evidence that the nucleus of each cell proliferates, renderins: the cell multinucleate. The blood-vessels are likewise hypertrophied in the placental folds, and to a less extent in the adjacent folds, but not at all in the folds opposite the placenta. In the placental folds there are larger blood-vessels running for the most part longitudinally, and all situated in the zone next the muscularis ; between this zone and the glandular layer, the blood-vessels are, on the contrary, all of small calibre, most of them taking a more or less radial course, and lying approximately in the plane of the transverse sections. All the blood-vessels of the mucosa have, so far as I have observed, the character of capillaries, for they consist of merely an endothelium without adventitial or muscular envelope, although some of them are many times the diameter of ordinary capillaries. The blood-vessels of the placenta of the Guinea pig are stated by Creighton, 77a, p. 544, to have the same character. There is a single layer of connective tissue cells condensed around the vessels, and representing the commencement of the perivascu


lar decidual cells, though the cells themselves scarcely differ yet from the ordinary connective tissue cells between the vessels.

We find at this stage all the regions of the placental swellings, to be found in later stages, already definitely marked out.

These regions are as follows : —

1, Placental : subdivided into

A, glandular zone.

B, sub-glandular vascular zone with {a), sub-glandular

zone with small vessels (p), outer zone with large vessels.

2, Peri-placental.

3, Ob-placental.

Each of these regions comprises two folds of the mucosa, viz. : the placental, the two folds next the mesentery;^ the periplacental, the two lateral folds ; and the ob-placental region, the two folds opposite the mesentery. The three zones of the placental area persist and become much more marked in later stages, owing to the great divergence of the processes of histological differentiation in each zone. Of the two vascular zones, the siib-glandiilar is characterized later by its very large multinucleate cells, while the oicter zo7te is characterized by its crowded uninucleate decidual cells. All the regions and their subdivisions will be perhaps better understood by the descriptions and figures of the nine days' uterus. {See below.)

In my specimens, although the uterus seemed to be very well preserved, the blastodermic vesicles were completely shrivelled up ; hence I could make no observations as to the relations of the embryo to the uterine wall.

§ 3. Uterus of eight days and three hours. — It is unnecessary to describe the appearances at this age in detail, as they have been described already with admirable clearness and exactitude by Masquelin and Swaen, 114, 25-30. I have therefore only to confirm their account and refer to certain points on which my observations extend or differ from theirs ; it is also necessary to describe the extra-placental structures, which are left out of consideration by Masquelin and Swaen.

^ Although this use of the term mesentery is etymologically indefensible, it seems permissible, and not hkely to lead to misunderstanding any more than the etymologically indefensible terms cell, endothelium, terminology, etc., etc.

346 MINOT. [Vol. II.

The connective tissue is but little altered from the condition at seven days ; it has the same adenoid character ; the cells are elongated in directions more or less parallel to one another, and the appearance of the protoplasmic reticulum therefore varies according as the section passes at right angles or parallel with the long axes of the cells ; in the former case the meshes are smaller, in the latter larger and longer. The perivascular cells have grow^n ; Masquelin and Swaen trace their origin to a metamorphosis of the connective tissue cells, in doing which I entirely agree with them ; these authors likewise describe fibrillae in the connective tissue, but in my preparation I can find none, nor from what we know of the structure of the mammalian uterus is it probable that any are present ; in regard to this point I think that Masquelin and Swaen's account needs rectification.

The placental blood-vessels have increased in size, and, I think, in number ; their epithelium, particularly in the larger vessels, is decidedly thickened.

The uterine epithelium has entered upon its complex degenerative metamorphosis — most of the changes have been seen and correctly described by Masquelin and Swaen for the placental area. They conclude that the changes lead to a new formation of blood corpuscles out of the substance of the epithelium. My own observations oblige me to regard the changes as phases of a hyaline degeneration with hyperplasia of the degenerating elements, and having nothing to do, therefore, with blood formation. In all parts of the uterine dilatations the epithelium is considerably thickened (Plate XXVI., Fig. i). The thickening is due to the enlargement and fusion of the epithelial cells, and this enlargement of the cells is due to the proliferation of the nuclei and to the growth of the protoplasm, which begins later and continues longer (as later stages show) than the multiplication of the nuclei. That the nuclei multiply within each cell can be distinctly seen in my specimens of this age ; the same fact has been observed by Masquelin and Swaen. The growth of the protoplasm is more properly described as an enlargement, due to degenerative metamorphosis. As to the nature of this metamorphosis I am unable to speak with much precision. The substance presents a very granular appearance, and possesses a slightly greater affinity for coloring-matters


than the unaltered protoplasm. Examination with an apochromatic oil immersion shows in some parts of the degenerated epithelium a distinct network, the threads of which are rather coarse and hyaline in appearance. In default of chemical and further microscopic examination we may accept the hypothesis that the degeneration consists in direct change of the protoplasmic reticulum into hyaline substance, accompanied by thickening of the reticular threads. The degeneration of the epithelium has progressed much further over the non-placental area than elsewhere, and much less over the placental area : the peri-placental regions are in an intermediate stage.

It is also important to note that the deep portions of the glands are nowhere degenerated. The glandular layer may be divided accordingly into an upper degenerated zone and a lower not degenerated zone.

In the placental area there is no stretching of the tissues, and accordingly the glands retain their tubular character. The nuclei fill up most of the epithelial layer ; there are three, four, or even five, in each cell in the upper part of the gland ; the cells of the fundi are but slightly altered from their usual appearance. The embryo is attached to the maternal placental surface only by the ectoderm, without any participation of the other germ layers, direct or indirect, so far as I can observe. That portion of the ectoderm which is soldered to the uterus is very much thickened, in marked contrast to all other parts of the layer. As shown in Fig. i., PI. XXVI,, the placental ectoderm runs over the surface only of the placenta, and stretches straight across the mouths of the glands, shutting them completely ; it does not dip down into the glands at all, and possesses no villi whatsoever. On the surface, between the glandular orifices, the uterine epithelium, already degenerated, is clearly distinguishable.

In the other regions the stretching of the walls stretches the glands also, and of course proportionately to the extent of the strain ; hence, in the non-placental area the glands become slits running parallel to the surface, and in the peri-placental part become wide cavities. The upper zone of the peri-placental glandular layer has its epithelium changed into a very thick layer, and beginning to undergo resorption, as evidenced by the presence of cavities. As we follow round towards the non-pla

348 MINOT. [Vol. II.

cental area, the evidences of resorption are greater, and over the area itself a large part of the upper glandular zone has disappeared altogether. Similar relations are found in the uterus of nine days, from which the drawings have been taken. For the sake of greater clearness, and to avoid repetition, we pass at once to the next stage. The fact that I have found the uterus at eight days so much nearer in its stage of development to that of nine than to that of seven days, may be attributed to accidental variations.

§ 4. Uterus at nine days and three hours. — Fig. 2, PI. XXVI., represents a transverse section through a swelling. The attenuation of the walls everywhere, except in the placepta, is very marked, and affects both the outer and inner muscular layers, Im, cm, and the mucosa, 7mic. In the placental region, PI, on the contrary, the walls are thickened ; the placenta itself is formed chiefly by the hypertrophy of the connective tissue of the two longitudinal folds nearest the mesentery, mes : the superficial glandular layer, gl, owing to its deeper staining, is readily distinguished even by the naked eye ; each lobe of the placenta is imperfectly subdivided into two lobules ; the embryo, in the specimen figured, appears in transverse section over the right-hand lobe, directly above the furrow separating its lobules ; the actual disposition is shown in Cut i ; in Fig. 2 the embryonic structures are purposely omitted on account of the small scale ; to the consideration of the foetal membranes the next section (§ 5) is devoted.

The connective tissue of the placenta is already far advanced in its metamorphosis, which progresses as described by Masquelin and Swaen. It consists of a rich cellular network, Fig. 3, conn, of which the cell bodies are much larger than in previous stages ; these bodies are for the most part elongated, with very irregular surfaces, and are, therefore, perhaps best characterized as roughly spindle-shaped ; their long axes are more or less parallel with the blood-vessels ; the nuclei are round, oval, or elliptical, granular, but with a clearer cortical layer, as is usually the case in young connective tissue cells : compare Rollett's Figs. 4 and 5 in Strieker's Handbuch der Lehre von den Geweben, I., pp. 63 and 65. The processes of the cells are numerous and very fine, forming a meshwork, between the cells, of such delicacy that it can be followed out only with high powers


(400-500 diams.). The observation of the threads of this network has led certain investigators to assume the presence of connective tissue fibrillae. Scattered about in the connective tissue are a not inconsiderable number of leucocytes, /, /, /, easily recognized by their size and shape, their granular appearance, deep staining and characteristic nuclei. Around the blood-vessels is the perivascular layer of decidual cells, per v, which have already been amply described by Masquelin and Swaen, Ercolani, Godet, Creighton, and others. Ercolani's descriptions, of which the most important to us is that of the rabbit at fifteen days,^ 89, p. 278, is far from sufficient. Godet's paper I know only from an unsatisfactory abstract. Creighton's account of the perivascular layer in the Guinea pig, 77a, 544, is also good, and he agrees with the Belgian authors in tracing the origin of the cells to the metamorphosis of the connective tissue. Ercolani opposed this view and maintained that the uterine mucosa is completely destroyed, leaving the whole placental tissue of the mother to arise as a new formation. My preparations render it impossible to agree with Ercolani, since they show all the phases of the metamorphosis. It is only necessary to follow, in Fig. 3, the three series of cells numbered I, 2, 3, 4, each, and to find in all parts of the placenta the same appearances ; to see the perivascular layer at six and seven days, before it is much differentiated ; and finally to see the perivascular accretions at later stages, to render inevitable the conviction that the perivascular layer is modified connective tissue. Neither at this stage nor at any earlier or later one have I been able to detect any evidence whatsoever of the resorption of the connective tissue affirmed by Ercolani, 89. Masquelin and Swaen describe multinucleate cells, but I fail to find them until later stages.

The blood-vessels have their endothelial lining considerably thickened, each cell for itself, and to its individual degree. Fig. 3, Endo ; they are stained by alum-cochineal and cosine more deeply than the adjacent decidual cells, from which they are sharply distinguished. I am unable to recognize any cells which might be interpreted as intermediate stages between the endothelial and decidual cells, as we should anticipate, were Erco ^ The specimen described by Ercolani I consider to have been probably really only about thirteen days.

350 MINOT. [Vol. II.

lani's suggestion correct that the decidual cells arise from the blood-vessels. The contents of the blood-vessels are blood corpuscles and coagulum ; the blood corpuscles resemble the ordinary red globules of the rabbit, a point deserving notice in view of the change occurring later. There are occasional leucocytes, but they are nowhere numerous.

The two layers of the vascular zone are now distinguishable not only by the size of the vessels (not well illustrated in the section drawn as Fig. 2), but also by the much greater development of the perivascular cells in the outer zone ; in the subglandular zone the cells are now not far from their maximum development, forming a coat one or two cells thick around the vessels ; on the outer zone, on the contrary, the number of layers of cells has still to increase very much ; consequently as development progresses, the difference between the subglandular and outer zone becomes more conspicuous.

The epithelium and glands repay careful study at this stage. The degenerative processes are similar in certain essential respects in all parts of the uterine swelling. The likeness concerns five chief points : 1°, the deep portions of the glands show little change in the epithelium ; 2°, the upper portions are very far degenerated ; 3°, the protoplasm of the degenerated epithelial cells is fused into a continuous thick hyaline mass, the growth of which ultimately obliterates the cavity of the glands ; 4°, the nuclei of the degenerated epithelium multiply enormously; 5°, the degenerated tissue is absorbed by progressive vacuolization.

But although these resemblances are dominant, each of the three principal regions, the placental, peri-placental, and ob-placental, presents now a very distinctive appearance, and has its distinctive further history.

In the peri-placental region, with which we begin, because the relations are more obvious there than elsewhere, we find the appearances shown in Fig. 4. The line of demarcation from the placenta, though not definite or sharp, can be approximately determined, but the passage into the ob-placental region is very gradual. The most striking feature of the section is the degenerated and enormously thickened epithelium, Ji.ep, deeply tinctured by the eosine, and remarkable for the crowded band of nuclei. Within the area of the degeneration the former


gland cavities are closed ; the diameter of the glands has enormously increased, and in some places two adjacent glands have swollen until they have come in contact and fused, the glands then forming a network ; in the placental region the conversion of the glands into a network goes very far. The distribution of the nuclei as at a and b preserves in some parts the original grouping in opposite walls of the gland tube ; at other points they lie in irregular patches. Secondary cavities, vac, appear at various points ; they are irregular in size, shape, and position, and arise by the resorption of the degenerated tissue. There is probably a certain amount of resorption carried on upon the surface against the connective tissue, for that surface becomes jagged and irregular, presenting a corroded appearance, as can be seen at various parts of Fig. 4. The vacuolization is, however, the principal factor of destruction. As to the manner in which the spaces are produced in the heart of the very compact layer, my observations give no satisfactory information. There are no accumulations of leucocytes either in the epithelial layer nor even in the connective tissue (see Fig. 4), in which all the cells are copied with approximately entire accuracy from the preparation. The only material I have ever noticed in the vacuoles is broken-down fragments of the surrounding hyaline tissue (epithelium) itself. The hypothesis may be suggested that the resorption vacuoles are produced by liquefaction, but the suggestion calls for no further discussion since there are no direct observations to test the validity of the hypothesis.

The deep portions of the peri-placental glands. Fig. \,gl, are dilated transversely to an extent which has converted them from tubes into wide vesicles. Towards the ob-placental region the transverse stretching gradually increases. The epithelium differs but little from that of the resting utricular glands ; it is composed of cylinder cells with basally placed oval nuclei.

In the ob-placental region the mucosa is much thinner than elsewhere. As we proceed from the edge of the peri-placental region towards the pole furthest from the placenta we find that the layer thins out and is more advanced in its degeneration. Near the peri-placental thickening there is a wide superficial layer of degenerated epithelium with the characteristic central band of nuclei, but the prolongations corresponding to the degenerated gland ducts are short ; the deep portions of the

352 MINOT. [Vol. II.

glands are oval slits parallel with the muscularis, Fig. 5, gl. A little further along, the resorptive vacuolization begins, producing a curious irregular layer, Fig. 5. The degeneration and vacuolization is found still further along to have involved the inner adjacent wall of the gland vesicles, thus producing the appearances shown in the left-hand part of Fig. 5, where there are shallow cups, gl, of epithelium, each entirely separate from its fellows, and all overlaid by the hyaline stratum, h.ep. There is usually a dome-like hollow in the degenerated stratum above each cup. Since the processes described vary in rapidity, there is not a uniform, but only a general, progression of stages towards the centre of the ob-placental region. Moreover, the variability is great, and the images from different sections and different parts of the same section are correspondingly multifarious, but the general succession of changes is everywhere the same ; hence it would be profitless to expand the descriptions.

The placental glands have preserved their tubular character ; they are less degenerated than the uterine glands of the nonplacental parts ; their walls are less thickened and in most parts the glandular cavity is still present. The deep portion of the glands are tubes lined by columnar epithelium. For the rest, I may refer to the satisfactory description of Masquelin and Swaen, 114, 30-31, except as to one point. As shown in Cut i, p. 355, the ectoderm of the embryo is firmly soldered to the placental surface over certain areas. The nature of this connection and the accompanying structural changes in the uterus are illustrated in Fig. 7, which has been copied with great care from one of the sections. For the sake of clearness, only the nuclei of the connective tissue have been drawn in ; the perivascular cells are represented by their nuclei and outlines, and the nuclei and cells of the foetal ectoderm are given in outline ; but there is nothing diagi-ammatic in the drawing; of course in the figure the distinction between the foetal and maternal tissues is more marked, though not more real than in the section ; in fact there is scarcely a cell even on the line of junction of the ectoderm with the uterus about the assignment of which one could have any doubt, so distinct is the texture and the staining of the foetal and maternal tissues. This is a matter of importance, as it renders it possible to ascertain beyond question that there are no villi ; nevertheless their presence has been assumed not infrequently. To pass on :


where the ectoderm, Fig. 7, Ecto, touches the placenta, the active resorption of the degenerated glands is going on (see the part of Fig. 7, above bracket C) ; whereas in other parts the glands present the appearance shown in Fig. 7, A, gz, and described by Masquelin and Swaen. There is also an intermediate zone shown in Fig. 7, above bracket B, where the transition phases between the two states are found; the zone of transition lies immediately underneath the point where the ectoderm, Ecto, joins the uterine epithelium, h.ep\ here the glands are thickened and hypertrophied ; the lumen is obliterated, but the cylindrical shape is irregularly preserved ; where the distal end of the ectoderm leaves the placenta, there is again a similar transition : in other words, the resorption is less advanced around the periphery than in the centre of the area of ectodermal attachment. The resorptive process is essentially the same as outside the placenta, — superficial corrosion and internal vacuolization, — but the vacuoles formed are relatively small and consequently more numerous : moreover, the space left by the disappearing epithelium is at once occupied by connective tissue cells, so that there are no cavities. The resorption goes on principally in the superficial layers of the placenta, where it results, as later stages show, in the complete disappearance of the glands immediately underneath the ectoderm ; deeper down the glands at the present age are hypertrophied and without lumina, but even in the region of bracket C of Fig. 7 (PI. XXVII.) most of the glands show very few or no vacuoles.

§ 5. Embryo at nine clays and three hours. — It is not proposed to consider here anything but a few points bearing on the relations of the embryo to the uterine walls.

First, as to the position of the embryo: the dorsal surface of the embryo is turned towards the placenta ; the embryo may be situated over one or the other lobe of the placenta or across both ; its long axis may be either parallel or oblique or at right angles, to the long axis of the uterus. In the specimen represented in Cut I, both the uterus and embryo appear in transverse section. Similar variability appears in my specimens of uteri of eight days, and of nine days and seventeen hours. The statements of Van Beneden and Julin, 44, and other authors,^

1 Compare Bischoff, Enttvickelungsgesch. d. Kaninchens, p. 138, and von Baer, Entivickeliujgsgesch. II., 232.

354 MINOT. [Vol. II.

lead rather to the supposition that the embryo normally lies across the uterus; but this is true in my experience only of the later stages after the embryo is suspended more freely.

Next as to the extra-embryonic portions of the germ layers : my observations compel me to differ as to the composition of the walls of the blastodermic vesicle or yolk sack. Sections through the middle region of the embryo, Cut i, show an open medullary groove, Md, the thickened notochordal band, ncJi, of the entoderm; the mesoderm of the Stammsone is undivided; that of the parietal zone is split to form the coelom, Coe, which may be followed a considerable distance from the axis of the embryo, when the two leaves of the mesoderm again unite into a single distal plate, vies, in which the two layers can be distinguished for a certain distance ; the distal edge of the mesoderm is sharply marked ; the mesoderm is thickened around its margin. Beyond the margin the ectoderm and entoderm come into contact, a\ the former is the thicker layer, being composed of cubical cells, while the entoderm consists of very thin broad cells ; the two layers continue outwards until, having passed over the periplacental thickening, they reach the region of transition from the peri-placental to the ob-placental area, where the ectoderm changing to a thin flat-celled membrane is intimately conjoined to the degenerated epithelium of the uterus. Where the concrescence takes place, the union of the two layers becomes very intimate, so that it is difficult to satisfy one's self how much farther the ectoderm extends, but apparently it goes completely around, forming a closed vesicle, as is generally stated and as is found later. Apparently the ectoderm is not involved in the resorption of the uterine epithelium, which disappears — but further investigation is required. The entoderm of the other hand is readily followed, and although somewhat crumpled and torn in the obplacental region of my specimens, there is to my mind very little doubt that it forms a closed sack, corresponding to the entodermic lining of the yolk sack of other mammals. From the observations above recorded it is evident that two germ layers are readily traced for some distance from the embryo, but that beyond a certain line only one layer, the entoderm, is readily followed. Now this fact is carefully represented in van Beneden's and Julin's diagram, 44, PL XXIV., but the membrane which stops is there given as entoderm, which, as seen in my

No. 3.]



356 MI NOT. [Vol. II.

specimens, is continued around. It is probable that when embryos of nine days are removed, that a portion of the ectoderm remains attached to the uterine wall, and consequently the inferior portion of the vesicle is without any ectoderm. Being unaware of such a possibility, van Beneden and Julin have perhaps represented the single layer left as ectoderm on account of the theoretical necessity of an ectodermal covering on the external or apparently external surface of the ovum. The question, therefore, is to be posed : Have not observers found two layers up to a certain limit beyond the vetia termiiialis, and only one layer over the remaining inferior portion of the embryonic vesicle, and assumed the single layer to be ectoderm, whereas it is entoderm, and the true ectoderm is left upon the uterus, to which it is indissolubly attached. The view I advocate brings the further question whether a portion of the embryonic ectoderm disappears by being involved in the resorption of the ob-placental uterine epithelium. This I think is not the case. The intimate adherence of the extra-embryonic portions of the germ layers to the uterine walls has been carefully recorded by Bischoff, Eiitivickehings gescJi. Kaninchens, p. 131, " Vom dem Umkreise der Vena terminalis an ist das Ei [of ten days] in die in dieser Lage, etc. . . . und von hier an sind auch alle Eihaute so innig unter einander und durch den Uterus vereinigt, dass es nicht gelingt sie zu losen."

The attachment of the embryo takes place as described by van Beneden and Julin, p. 402, 403, by an area of thickened ectoderm ; the general arrangement is well shown in Cut i, while the fitting together of the foetal and maternal surface is better illustrated by Fig. 7, PI. XXVII. The foetal mesoderm does not participate even indirectly in this attachment, but runs along free from the outer germ layer. The ectoderm, as it nears its attachment (see Fig. 7), gradually thickens. Just where it joins the uterine surface there are several large cells with very large nuclei ; appearances which are probably connected \vith the growth of the layer, for beyond the line of the large cells the ectoderm is very much thicker. Extremely distended nuclei also occur very strikingly in the developing supra-renal capsules, and are also there connected presumably with cell proliferation. If these suppositions are correct, there is a modified form of cell division characterized by dilatation of the nuclei and which


deserves special study. Over the area of attachment the uterine epithelium, Fig. 7, Ji.ep, is degenerated as before described, its surface is extremely irregular, but the ectoderm, Ecto, is perfectly fitted to every irregularity, but the free surface (towards the mesoderm) is comparatively smooth ; the layer consists of two, three, or four strata of cells. Beyond the area of attachment the ectoderm again thins out.

As to how the tissues are held together, my observations afford no explanation. It seems to me possible that the two tissues actually grow together as a grafting unites with a bough ; but for aught we know it may be by some other process, perhaps simple agglutination. The thickening of the ectoderm I am inclined to regard as degenerative, and therefore somewhat comparable to the degenerative thickening of the uterine epithelium. I am brought to this view by no conclusive observations, but chiefly by two facts : 1°, that in later stages the ectoderm seems to have disappeared over the greater part of the placenta (see § 7. Uterus of eleven days) ; 2°, hyperplasia is often the commencement of degeneration, as is familiarly known to pathologists. To this evidence may be added the appearance of the ectoderm at nine days and seventeen hours, which I interpret as indicative of degeneration.

§ 6. Uterus of nine days and seventeen hours. — In my specimen there are not many changes from the previous stage last described, but of these changes the following deserve special mention : 1°, the commencing formation of perivascular decidual cells in the peri-placenta ; 2°, the reconstitution of the ob-placental epithelium ; 3°, the formation of the true chorion ; 4°, changes in the extra-embryonic ectoderm ; 5°, the contents of the placental blood-vessels.

1°. The peri-placenta is still only a small bolster at the side of the placenta ; its glands are still recognizable and its blood-vessels are more conspicuous ; the connective tissue cells are enlarged and have begun to form more or less distinct coats around the blood-vessels. I feel assured that the decidual cells arise here in the same way as those of the outer zone of the placenta ; the cells in the two parts appear to me identical in character as soon as they attain their full development, and to differ only in the period during which their development takes place ; later on, Fig. 8, PI. XXVIII., the peri-placenta forms, together with

358 MINOT. [Vol. IL

the outer zone of the placenta, a continuous layer of uninucleate decidual cells, extending over half the uterus.

2°. In the ob-placental region the degenerated portion of the uterine epithelium is almost completely resorbed around the pole opposite the placenta, (compare Figs. 4, 5, and 6). Fig. 6, taken from an older stage, in which the phase existing at nine days and seventeen hours at the pole is found near the peri-placenta, illustrates the manner in which the patches of unaltered epithelium, gl, of Fig. 5, grew together by the union of their edges into a continuous sheet of epithelium. Fig. 6, gl, forming a series of shallow cups, widely open.

3°. The chorion of mammals, as I have defined it elsewhere, is " the whole of that portion of the extra-embryonic somatopleitre which is not concerned in the formation of the amnion." ^ The term is not applicable until the mesoderm has united with the ectoderm in the region outside the embryo to form a single membrane : such a union has now taken place ; the thickened placental ectoderm is coated by a thin layer of flat cells, epithelial in character and with bulging nuclei. These cells represent the lining of the body cavity, or, as this lining is conveniently called, mesothelium. The mesothelium, and consequently the coelom, extend a slight distance beyond the edge of the placenta ; the mesothelium then bends over onto the yolk sack, of which it becomes the vascular coat, and then runs towards the embryo ; the vascular coat has a large vessel, sinus tertninalis, near the end of the coelomatic space, and the mesoderm stretches a short distance beyond away from the embryo. The ectoderm, on the contrary, extends beyond the end of the mesoderm away from the embryo over the rest of the yolk sack. Thus the yolk sack, as is well known, comprises two parts, one near the embryo with walls composed of entoderm covered by mesoderm, and away from or opposite the embryo, with walls composed of entoderm covered by ectoderm ; compare the clearly expressed summary of the relations in the rabbit given by Balfour in his Comparative Embryology, II., 199, 200.

4°. The ectoderm of the embryo presents the same general arrangement as at nine days. The area of thickened entoderm, however, which is attached to the placenta has changed in appearance ; at nine days and three hours each cell outline was

^ Wood's Reference Handbook of the Medical Sciences, II. , 143; article, Chorion.


distinct, and the protoplasm around each nucleus dense and finely granular ; now the cell outlines are hard to follow and the picture is confused by broader lines of hyaline matter, which is colored by the eosine ; the nuclei are enlarged, the protoplasm is more coarsely and more irregularly granular, and somewhat vacuolated. The characteristics enumerated concord with the idea that degeneration is going on, — an idea suggested, also, as before stated, by the failure to find this part of the ectoderm in later stages.

5°. The blood-vessels show increased hypertrophy of their epithelium, and the perivascular cells form two or three layers around them ; they are especially remarkable for containing a very large number of multinucleate leucocytes and comparatively few red corpuscles. The excessive abundance of white globules continues up to the oldest stage I have examined (sixteenth day). The predominance of nucleated corpuscles causes the contents of the maternal vessels to resemble foetal blood when examined with a low power ; with high magnifications the difference is evident. To the foetal blood in the placenta we shall have to recur.

§ 7. Uterus at eleven days and three hours. — Very great changes have taken place — so great that they cannot be understood completely until some of the intermediate phases are studied. Want of suitable material has hitherto prevented my doing this. At the present stage — the beginning of the twelfth day — the placenta is distinctly pedunculate, and there is consequently a circular cleft between its sides and the closely adjacent peri-placenta; in the middle of the placenta a deep fissure corresponds, of course, to the space between the two folds of the uterus, out of which the placenta is developed, and therefore runs lengthwise of the uterus. The allantois has acquired considerable size and is attached to the surface of the placenta, from which the ectoderm has disappeared. The glands of the placenta are very far degenerated and altered ; in the sub-glandular zone the multinucleate cells have appeared, and in the outer zone the perivascular cells have increased so as to occupy nearly all the space between the vessels. In the periplacental and ob-placental regions, the modifications are equally noteworthy. Such, in brief, are the more striking changes. Let us consider them with crreater detail.

36o MINOT. [Vol. II.

The diagram on PI. XXIX. will enable the reader to follow the ensuing descriptions. The general explanation of the diagram is given in the next section.

The placenta is shaped somewhat like a mushroom : it has a very thick stalk, with a somewhat broader top. The top is bilobate, there being a deep fissure between the two lobes ; this fissure persists at thirteen days (Fig. 8,/) ; its fundus is the subplacenta (Ercolani's cotyledonary organ). The sides of the fissure are, of course, part of the surface of the placenta, morphologically speaking, and bear glands. The three zones of the placenta are well marked. In the outer zone the blood-vessels are very wide, with thickened degenerated epithelium ; the perivascular cells occupy the entire space between the vessels in all that part of the zone towards the muscularis and most of the space in the part towards the glands. Next to the sub-glandular zone, therefore, we see the vessels surrounded, each by its separate thick perivascular coat, while the intervening tissue still consists of anastomosing cells, like those which in earlier stages occupied more of the space and which formed the only packing between the vessels at six days. The blood-vessels convert the layer, by their enlargement, into a spongy tissue, which has been described not only in the rabbit, but in other rodents ; the vessels themselves have been generally described as glands, but the study of their development renders doubt as to their true character impossible. The vessels are partially empty in my preparations, but they contain very numerous leucocytes, nearly all of which have several nuclei apiece, which are conspicuous from their dark staining : there are a few red globules and here and there a little coagulum. As the corpuscles of the embryo are large nucleated bodies, there is no difficulty in distinguishing the foetal from the maternal blood, even in the upper part of the placenta, where the two circulations are juxtaposed. The middle or sub-glandular zone has undergone greater changes still. In it, as likewise in the glandular zone, the perivascular cells have almost entirely disappeared,^ but they are, as it were, replaced

^ This statement is perhaps not correct. There are certain spaces surrounded by epithelial or epithelioid cells to be seen in the upper part of the sub-glandular and in the lower part of the glandular zone; these spaces I have interpreted as parts of the glandular system, but they are perhaps maternal vessels with perivascular cells. The uncertainty as to their character could be probably removed by the examination of the ten days' placenta, which presumably offers the intermediate stages.


by multinucleate cells (compare Fig. 14, PI. XXVIII., of these cells from an older placenta) ; their origin appears to be due to the development of clusters of connective tissue cells, which lie scattered about between the blood-vessels ; each cluster consists of from three to six cells lying together and connected on the one hand by short processes with one another, and on the other by longer processes with the cells of adjacent clusters. The larger clusters are separated by membranes from one another, and thus every cluster becomes enclosed in a membrane and appears as a multinucleate cell. The development of these cells would doubtless repay more accurate investigation. The multinucleate cells do not yet form a continuous bed under the placenta, but are divided into parts by masses of very loose connective tissue. At the base of the fissure between the two lobes of the placenta the glands have almost entirely disappeared, but we still find a few unresorbed fragments of their degenerated epithelium ; these fragments are conspicuous by their very deep staining, both of the hyaline substance and of the nuclei : the neighboring tissues are less colored. The fissure itself is like an inverted X ; that is, it is transversely expanded at the base ; the floor of the expansion is thrown up into folds and covered by a cylinder epithelium, which I feel some hesitation in designating as the regenerated uterine epithelium, although it resembles the epithelium on the peri-placenta, where the glands are resorbed and the epithelium reconstituted from its degenerated self. On the other hand, as shown in the next section, there is some proof that the foetal ectoderm penetrates, by villous growths, far into the placenta. It seems possible that the fissure is filled by villous excrescence of foetal origin and that the epithelium of the sub-placenta belongs to the villus. This view does not commend itself to me. Neither upon the upper wall of the expansion nor on the sides of the fissure have I recognized any epithelium. The upper part of the fissure is closed by an ingrowth of connective tissue. Hence the lower part is changed into a shut cavity in the centre of the placenta, and into this cavity the folds covered with epithelium project. So far as I am informed, this curious structure has not been described hitherto, but what appears to be clearly its homologue has been observed by Ercolani, 89, pp. 290, 291, Tav. IV., Fig. I., O, and specially studied by Creighton, 77b. Both of these

362 MINOT. [Vol. II.

authors examined late stages when the fissure is completely filled by connective tissue, so that there is no space — a condition found in the rabbit at thirteen days. It will be convenient to designate the structure as the sub-placenta. Its occurrence is confined to rodents so far as at present known. Finally, we have to note that at the edge of the placenta, toward the peri-placenta, the sub-glandular layer, which we are now considering, is characterized by the presence of deeply stained fragments of glandular epithelium irregularly scattered through the other tissues and similar in appearance to the remnants of the glands about the sub-placenta. These fragments appear to have been seen by Ercolani, Creighton, Masquelin, Swaen, and others, and variously interpreted, their true nature not being recognized. The disappearance of the glands at the centre and at the periphery of the placenta virtually increases the domain of the sub-glandular layer. The greatest changes have occurred in the glandular layer. Scarcely a trace of the perivascular cells can be found ; the space they formerly occupied is taken up by a very loose embryonic tissue ; the glands are completely altered ; they have lost their special affinity for eosine and cochineal, neither the hyaline substance of which they are composed nor the nuclei they contain being more stained than other tissues (compare Fig. 8) ; they are irregularly cylindrical in shape, very much contorted, and united with one another at irregular intervals, so as to constitute an actual network : they are very much vacuolated ; their deep portions (fundi) are somewhat wider than the upper parts ; here and there one sees a remnant of the original central lumen. The contorted masses, which I consider glands, are apparently the same as have been seen by Mauthner in the placenta at term, 115, p. 121. He describes these cords as consisting of the fused epithelium of adjacent foetal villi, and the spaces I have designated as vacuoles he describes as maternal blood-channels; he states explicitly that he has injected them from the maternal vessels, and in other cases found them gorged with maternal blood. These statements are irreconcilable with my own views, detailed in the present article. The uterine epithelium has entirely disappeared both from the top and the sides of the placenta. The top surface is covered by a very thin layer of flat epithelium. Fig. 8, msth, which is found, when followed out, to be continuous with the lining of the body cavity of the


embryo ; it is therefore mesothelium. Underneath this covering, and above the glands, there is a layer of varying thickness containing some large and a few small blood-vessels with embryonic blood in them, and consisting otherwise only of scattered anastomosing connective tissue cells,^ which can be followed without the slightest break on the one part until they pass directly into the mesoderm of the superjacent embryo; on the other part, down between the glands. Fig. 8, mes ; compare, also, later stages, Figs. 10 and 1 1. Between the glands, also, are blood-vessels containing embryonic blood. On the top surface of the placenta I can find nothing recognizable as even a trace of the foetal ectoderm, which formed a thick and conspicuous covering in the latest previous stage examined (nine days and seventeen hours). At the edge of the top of the placenta. Fig. 8, the relations change : the mesothelium, insth, bends up and leaves the placenta, and together with a few subjacent mesodermic cells joins a sheet of cylinder epithelium, Ecto, which is shown by its connections to be foetal ectoderm. The ectoderm from the point where the mesothelium, msth, bends on to the top of the placenta continues downward, Ecto, to clothe the side of the placenta which faces the peri-placenta. Between the placenta and peri-placenta, as shown in Fig. 9, there is a fissure ; the ectoderm can be followed to the bottom of this, and from there extends, — not on to the peri-placenta, — but turns abruptly back on to the side of the placenta, up which it stretches a minute distance and thereupon ends abruptly. The disappearance of the ectoderm is discussed in the next section.

The peri-placenta is now characterized by the enormous increase of the perivascular decidual cells and the accompanying expansion of the blood-vessels ; by the disappearance of its glands and by the reconstitution, in part, of its superficial epithelium. The peri-placenta appears like the continuation of the outer zone of the placenta, for it directly adjoins it, is of about the same thickness, and is histologically similar. The bloodvessels are wide with hypertrophied endothelium; the perivascular cells are disposed as in the sub-glandular zone of the placenta ; that is, in the half towards the uterine muscularis they completely fill the intervascular room, but in the half towards

1 A layer closely similar to this, and presumably homologous with it, exists in the Guinea pig (Creighton, 77a, p. 558), in the rat (Ercolani), and other rodents.

364 M/IVOT. [Vol. II.

the interior of the uterus they form a discrete envelope around each vessel, the spaces between the perivascular coats being occupied by simple connective tissue cells. The glands which at nine days, Fig. 4, were so bulky and conspicuous, have almost completely disappeared, being now represented only by remnants of multinucleate hyaline matter scattered superficially, and easily recognized by their distinctive and conspicuous coloration : some of these remnants are still united with the surface. The epithelium is in two forms : on the half of the peri-placental surface towards the placenta it is entirely in the phase of degeneration, while over the other half it is already reconstituted as irregular cylinder epithelium, the cells of which are more or less separated from one another, and somewhat variable in height ; this epithelium stops abruptly near the middle of the periplacenta and is replaced towards the placenta by a hyaline nucleated layer occasionally thickened into a lump, where the nuclei are clustered ; the cylinder epithelium is deeply stained by the cochineal ; the hyaline epithelium has a marked color from the eosine, and its nuclei are dark with cochineal. The glands are further resorbed under the cylinder epithelium than nearer the placenta.

The ob-placenta is now characterized by the disappearance of its degenerated epithelium, by the fusion of the epithelium of the deep portions of its glands into a new continuous layer, and by the development of peculiar monster cells in its central area facing the placenta. The resorption of the epithelium by vacuolization has already been described in the account of the nine days' uterus, § 3. Thje epithelium, Fig. 6, gl, is everywhere re-formed as a continuous layer ; portions. Fig. 6, h.ep, of the degenerated layer remain especially near the peri-placenta, but for the most part the new epithelium is entirely uncovered, and in the central region it has grown, so that the glands are already deepened. But the most remarkable feature is the accumulation, opposite the placenta, where the mucosa is much thickened, of the curious bodies, to which I apply the term monster cells. They are round or oval masses many times the size of any other histological element of the uterus or embryo, and possess huge nuclei. They are shown in Fig. 17, which represents them at a later stage, when they are further enlarged. I regard these bodies as detached epithelial cells, undergoing degenerative hypertrophy. In


spite of long searching for the phases representing their earlyhistory, I have failed to ascertain positively their origin. In the next section the question is recurred to. The monster cells vary in size: the smallest ones lie near the epithelium; the larger ones, for the most part, deeper down and even among the muscular fibres, but a few large cells lie next the surface. The body of the cell is evenly and coarsely granular and resembles the hyaline degenerated protoplasm of the epithelium ; its external outline is distinct, well-rounded, and without processes ; the nucleus, which often has a slight space around it, as if it had shrunk a little, has a clear regular outline, being apparently provided with a membrane ; it is well colored by cochineal, and contains an indistinct network with imbedded granules of various sizes ; in the smaller cells the nuclei have one or two granules much larger than the rest, and which may be spoken of as nucleoli ; the size of the nucleus increases with that of the cell, and at the same time the granulation becomes coarser.

The description of the placenta at ten and eleven days given by Masquelin and Swaen I have not been able to follow in all respects. Owing to their conclusion that the epithelium of the uterus gives rise to blood, they apply the term cavites hcsmatoblastiques to apparently all the cavities of the placenta except those of the maternal blood-vessels. I have compared their description very carefully with my own preparations : so far as this enables me to judge, their "cavites haematoblastiques " include the foetal blood-vessels, the vacuoles in the degenerated glands, the spaces included within the epithelial U's described in the next section, and which are supposed to be the tips of foetal villi, the multinucleate cells and perhaps also the sub-placenta. Why the multinucleate cells are included among the bloodforming organs the authors do not render clear. Their failure to recognize the variety of constituents in the glandular layer of the placenta must be ascribed to the want of the perfected methods at present at our disposal. With the means now at command there is no difficulty in obtaining preparations which show indisputably that the glands though degenerated persist intact, and do not give rise to blood cavities nor blood corpuscles as Masquelin and Swaen have erroneously believed.

§ 8. Embryo at eleven to thirteen days. — As known already, the embryo is completely separated from the yolk sack, and the

366 MINOT. [Vol. II.

allantois has grown forth and attached itself to the placenta. The relations of the extra-embryonic structures have been represented by Bischoff in the diagrams of PI. XVI. of his classical memoir on the development of the rabbit. These diagrams have since been reproduced again and again, sometimes with modifications as notably by Kolliker in his manual, and by Van Beneden and Julin. Guided by these and by my own preparations I venture to construct a new diagram, PL XXIX., which I hope will approximate more nearly to the actual relations of the parts, with which we are now concerned.

In the first place it is to be noted that most of the section is occupied by uterine tissue: — compare Fig. 9, PL XXVIII. The largest space is occupied by the placenta, on the surface of which is situated the embryo, lying upon its side. Opposite (above in the figure) the embryo is the ob-placenta, ob-fil, with its central area, containing the monster cells, mo cl; the inferior wall of the yolk sack is fitted upon, but not attached to, the obplacental surface. The peri-placenta, PP, appears as the continuation of the outer zone, oz, of the placenta ; it has no glands : its blood-vessels are enlarged, and all the space between them is filled with uninucleate decidual cells. This description of the peri-placental structure applies also to the outer zone, oz, of the placenta. A narrow space separates the surface of the peri-placenta from the side of the glandular zone, gl, of the placenta: the letters a and b lie in this space. The placenta consists of three zones : 1°, the upper glandular zone, gl, divided by a fissure,/!, into two lobes. This fissure is partly filled with an ingrowth of embryonic mesoderm, ines ; the transversely expanded bottom of the fissure forms the sub-placenta, ; the glandular zone as a whole constitutes a protuberant mass with top and sides clearly distinguishable. Below the sub-placenta is the sub-glandular zone, s.-glz, with dilated blood-vessels and multinucleate decidual cells.

The embryo lies upon the surface of the placenta. From its ventral side spring the allantois, all, and the stalk of the yolk sack ; for the sake of clearness the amnion and pro-amnion are entirely omitted, since they have no direct relation to the uterus.i The allantois expands upon the placenta ; the yolk

^ For diagrams of the pro-amnion, etc., see Van Beneden et Julin. Copies of their figures are given in Buck's Reference Handbook of the Medical Science, VI., 32.


sack expands over the ob-placenta. The cavity of the allantois, all, is of course lined by entoderm ; it is, however, quite small, and in my preparations by no means the spacious vesicle commonly represented, for instance, by Kolliker in his Grundriss (2te Aufl., Fig. 88), or by Balfour {Comparative Embryology, II., Fig. 148). The allantoic mesoderm, mes, on the other hand, spreads out, over the surface of the placenta, down its sides, down into the fissure, f, between the two lobes, and penetrates between the glands,^/, of the placenta; wherever it goes, the mesoderm carries foetal blood-vessels. The free, i.e., inner or coelomatic, surface of the mesoderm bears the mesothelium, visth; as the extra-embryonic coelom does not extend beyond the top of the placenta, there is, of course, no mesothelium upon the sides of the glandular zone (between a and b), but at the edge of the top of the placenta the mesothelium is reflected back, and after a short course joins the wall of the yolk sack near the sinus terminalis, v.t.

The yolk sack, as has been long known, consists of two parts :^ I St, the area vasciilosa bounded approximately by the sinus terminalis, vt ; within this area the entoderm is united with the mesoderm, which passes only a very short distance further out ; 2d, the remaining portion without mesoderm, excepting always the pro-amnion, which is included in the area vascidosa ; over this second region the entoderm, en, rests directly upon the outer germ-layer, ecto.

If we follow the ectoderm around, we find that it leaves the yolk sack, just before the sitms terminalis, vt, is reached, and being joined by the mesodermic lining of the coelom passes down b on to the lateral surface between the peri-placenta, P, and the glandular placenta, gl, where, as already described, it bends inwards, and turning back runs a minute distance upwards ; according to my hypothesis it continued earlier over the surface of the placenta, as indicated by the broad dotted line, d. The layer of embryonic epithelium upon the side wall of the rodent placenta has been seen by other observers, among whom may be mentioned Ercolani and Creighton ; the latter, 77b, 560, directs especial attention to it, in the Guinea pig, but refers it to the entoderm. I consider it probable that it is really ectodermal in the Guinea pig, as in the rabbit. Underneath the

^ Leaving the pro-amnion out of consideration.

368 MINOT. [Vol. II.

ectoderm, b, to be seen at eleven to thirteen days at the sides on the placenta, is a layer of mesoderm without any ccelom. Now, if my suppositions are correct, then the ectoderm forms at first an independent fold, ba, beyond the terminal vein, vt\ the mesoderm, but not the mesothelium, extends into this fold, which covers the sides of the placenta. The disappearance of the foetal ectoderm from the surface of the placenta, and the penetration of the foetal blood-vessels between the glands, are changes which take place during the eleventh day. How those changes occur, observations on the development at that age must decide. Meanwhile let us make shift with two hypotheses. The first is : The whole of the ectoderm attached to the placenta degenerates and is resorbed. Since the uterine epithelium, as observation indicates, has likewise disappeared from the placenta, the mesoderm, mes, of the allantois, all, is brought into direct and free contact with the connective tissue and degenerated glands of the placenta, and is thus enabled to carry by its own ingrowth the foetal blood-vessels into the very substance of the placenta. The second hypothesis is that the ectoderm and mesoderm have produced villi, which have grown into the placenta. In favor of this latter hypothesis there is certain evidence which I have not yet alluded to. In the deep portions of the glandular layer of the placentas of both eleven and thirteen one finds narrow loops of epithelium like a tuning-fork in shape ; the open ends of the U-loops are towards the top of the placenta ; the epithelium composing them is a cylinder epithelium, which gradually thins out towards the upper end of the legs of the U ; it differs altogether in appearance from the degenerate gland epithelium, the interiors of the U's contain vessels with foetal blood ; so far, then, these structures might be longitudinal sections of the ends of foetal villi. Towards the surface of the placenta the epithelium of the loops thins out, and I have not been able to follow them. If we have to do with villi, we must assume that the ectoderm has become exceedingly thin over their basal portions, but is preserved as a thicker layer over their tips, and my failure to trace the villi would be attributable to the imperfection of my preparations and observations. Balancing the pros and cons leads me to favor the second hypothesis. Let me add that the mesoderm of the embryo is continuous without a break with the interglandular connective tissue ;


this statement is correct beyond any doubt, for I have several sections, in each of which the direct passage is observable under the microscope without even displacing the slide. By hypothesis this mesoderm is, however, really separated by a very thin covering of foetal ectoderm from the uterine tissue, and the whole constitutes a system of villi which have grown down like roots into the placental soil.

That there is no communication between the foetal and maternal circulations must be deduced from the fact that the two bloods are never mingled in one vessel, although found side by side in adjacent vessels. The separation of the foetal and maternal blood has already been insisted upon. The full elucidation of the double placental circulation must be left for injections to bring.

In brief : The rabbit embryo is attached to the placenta by the ectoderm, which disappears from the surface of the placenta during the eleventh day ; the vascular connective tissue of the allantois grows probably by forming true villi into the placenta, and so comes close to the maternal circulation.

In other rodents the placenta contains foetal vessels ; its surface is covered after a certain stage by a thin epithelium like the mesothelial layer of the rabbit, and by a layer of vascular connective tissue. Hence it seems probable that the structure in the rabbit is typical of the class — compare § 12.

§ 9. Uterus at thirteen days and three hours. — The placenta and embryo are considerably bigger than at eleven days, but the structure of the parts is comparatively little changed.

A complete section is drawn in Fig. 9. The longitudinal muscles, l.m.., and the circular muscles^ form the external covering. They differ in microscopical appearance from the muscles of the resting uterus, but I have not investigated the change in them.

The placenta is very bulky. Its two lobes have begun to form separate protuberances, so that the top of the placenta is no longer a nearly plane surface. The placental surface is covered by the mesothelium, which is a little thicker than in the previous stage, the cells having a greater vertical diameter. Between the mesothelium and the glandular layer, gl, is the vascular mesoderm, several of the large vessels of which are shown in Fig. 9. The central fissure,/, of the placenta is very

370 MINOT. [Vol. II.

deep ; it is completely filled with the ingrowth of mesoderm and its accompanying large vessels. At the bottom of the fissure next the outer placental zone, o.z, is situated the sub-placenta, sb.-pl. The section drawn in Fig. 9 does not show the connection between the fissure and sub-placenta, which appears in sections 208-214 of the same series. The thickness of the mesodermic covering of the placenta has increased very considerably, and the larger vessels are now provided with well-marked muscular as well as endothelial walls. Many of the foetal vessels run in spaces which stretch down nearly vertically from the placental surface ; in some cases the vascular columns can be followed until they enter a cap of epithelium which forms a sort of U. These relations suggest the presence of a series of foetal villi covered in part by a very thin epithelium, and covered at their tips by a relatively thick epithelium. This interpretation has been discussed in the previous section. Beside the normallooking epithelium, we find the degenerated glands not much changed from eleven days. The sub-glandular zone, sgl.z, shows further enlargement of the blood-vessels, so that they are now larger than those of the outer zone, o.s, thus reversing the earlier relative proportions ; the multinucleate cells have increased in number and size, and contain more nuclei than at eleven days ; they occupy all, or nearly all, the room between the vessels. Towards the outer zone the vessels are surrounded by the uninucleate perivascular cells, but the intervening tissue consists of multinucleate cells, so that there is a boundary region which cannot be assigned strictly either to the subglandular or to the outer zone. The outer zone, o.z, is solidly packed with perivascular cells.

The sub-placenta,, lies still deeper than before, being now close to the outer zone. Its epithelium is undergoing hyaline degeneration, and accordingly is irregularly thickened, and its nuclei are multiplied : the substance of the layer stains deeply with eosine.

The peri-placenta, P, differs from that at eleven days, principally in having the perivascular cells as a solid packing throughout the whole of its extent, except just where it adjoins the glandular layer of the placenta. As at eleven days, its covering epithelium is reconstituted on the part towards the sub-placenta, and is in the phase of degeneration towards the placenta.


The ob-placenta,, shows everywhere a marked growth of its glands ; as illustrated by Fig. 10, the glands are follicular; their cavities wide. The glands are not branched or pouched, as the appearances in the sections suggest ; they are broad tubes closely packed, and are necessarily cut obliquely in most cases. The rather ragged-looking epithelium is composed of long cylinder cells (Fig. 10), with the nuclei at various heights, and the protoplasm a good deal colored by the cochineal. The connective tissue of the mucosa has also grown, and forms both thin inter-glandular dissepiments and a thickened sub-glandular stratum. In the centre of the ob-placenta the mucosa is still further thickened to make room for the monster cells, which lie for the most part below the glands, but are found also between the glands and in the superficial portion of the muscularis. At one point the ob-placenta is interrupted by a protuberant mass, X, resembling the peri-placenta in structure ; it consists of crowded perivascular cells with dilated blood-vessels, and is covered by epithelium. As I have seen nothing analogous to this mass in any other specimen of any age, it must be regarded as a singular sporadic variation from the normal processes of development.

The origin of the monster cells I am inclined to seek in the uterine epithelium, as stated in § 7. The appearance of their cell bodies, and of their nuclei at once suggest this origin on account of the similarity with the appearance of the degenerated epithelium elsewhere. We find, also, the smallest monster cells near the epithelium. In Fig. 1 1 portions of the epithelium of the peri-placenta are represented. The cells are all multinucleate, as seen both in vertical section, A, and surface views, B ; occasionally, but very rarely, there is a cell with the nuclei gathered together in a central mass, with an indistinct line enclosing the bunch, Fig. \i, c. These cells are larger than the rest, and their protoplasm is somewhat degenerated. If such a cell were to detach itself, and hypertrophy and the bunch of nuclei to break down, it would resemble a monster cell. Yet I can find no evidence that such a metamorphosis actually takes place in the ob-placenta. In the ob-placenta itself there appear a few epithelial cells with a single nucleus which are slightly enlarged, and are possibly the initial stages of monster cells, but between them and the youngest monster cells observed I have failed to

272 ' MINOT. [Vol. II.

discover any intermediate stages. The difficulty of finding the first stages of the monster cells indicates that their development must be extremely rapid, almost sudden.

§ 10. Uterus and embryo at fifteen days and four hours. — The swelling of the uterus has considerably increased ; the placenta is larger ; the cavity containing the embryo is very much larger ; the peri-placenta has grown but little. We notice now that of the six folds of the uterus, the two placental have expanded both in width and thickness to a far greater extent than the remaining four folds, except that the lateral expansion of the two ob-placental folds, by attenuation of their walls, has enabled the ob-placenta to occupy an extent of the circumference of the uterus which is about equal to that taken up by the placenta proper ; only about one-sixth of the whole circumference is allotted to the peri-placenta. With the naked eye one can see that the fissure of the placenta has opened so that the surfaces of the two lobes of the placenta now face each other like the sides of a V ; the surface of each lobe, though somewhat irregular, is as a whole arched. The glandular zone is perhaps slightly thicker than at thirteen days, but the diameter of the sub-glandular zone is markedly lessened, owing apparently to the opening of the interlobal fissure and the consequent flattening of the surfaces of the lobes. With a hand-lens one easily recognizes that the blood-vessels of the vascular zone of the placenta are of much greater diameter than at thirteen days, while the dissepiments between the vessels are not only relatively but absolutely thinner than before : this observation does not necessarily involve the conclusion that there has been an actual loss of tissue, for the placenta as a whole has increase in bulk. Let us turn now to the microscopical examination.

The placenta differs but little, except in the respects above mentioned, from the stage last described. The mesodermic covering of the placenta is well marked. Fig. 12, mes, and the foetal mesothelium, msth, is perfectly distinct ; it leaves the placenta at its edge to curl over on to the yolk sack, just as at an earlier stage. Fig. 8, msth. The side of the lobe next the periplacenta is clothed by ectoderm essentially as described at eleven days and partially shown in Fig. 8, ecto ; but the ectoderm is now more irregular than at earlier periods and is thrown into small folds near the point where it is reflected back on the pla


centa ; similar appearances are clearly indicated in Ercolani's memoir, 89, Tav. IV, Fig. i, i, i, for the Guinea pig. It is quite possible that the folds are more developed in the rabbit later. The placental glands are very much contorted, Fig. \2,gl,gl; very coarsely grandular, with numerous irregular vacuoles and with the nuclei lying for the most part against or near the outer surface of the gland, Fig. 13, gl '. the nuclei no longer stain deeply as they do during the first stage of the gland degeneration, Fig. 7. In the upper part of the placenta the glands are much narrower and more widely separated than in the deep part of the layer, as can be seen in Fig. 12, which takes in about half of the glandular layer from the surface down ; towards the surface the glands often form wide loops, Fig. 12, and join one another, making a network with closed meshes. As regards the supposed foetal villi, I find the columns of the foetal mesoderm running down more distinctly than at thirteen days, but as before, the only epithelium which I clearly distinguished, is that in the deepest part of the glandular layer disposed as if covering the tips of the villi. The blood-vessels are very numerous, and some of those above the glands in the foetal mesoderm are very large. Fig. 12, v. It will be remembered that these vessels belong to the foetal system and that the plexus of vessels, which is so conspicuous upon the surface of a freshly excised placenta, pertains therefore to the embryo. At certain points there rise thin membranes from the surface of the placenta, which carry good-sized vessels : whether these are accidental or constant, I am unable to say. Examined with a still higher power. Fig. 12, the glandular layer shows the peculiarities of its structure still more clearly ; the mesothelium, msth, upon the surface, though composed of flat cells, has considerable thickness ; the mesodermic cells, mes, are for the most part spindle-shaped and their processes anastomose ; the foetal bloodvessels, Vy V, come close against the glands, gl; if, therefore, there is a layer of foetal ectoderm separating the foetal mesoderm from the uterine tissues, it must be very inconspicuous from extreme thinness.

As to the relations of the sub-placenta, my preparations are unsatisfactory.

The sub-glandular layer shows the vascular endothelium advanced in degeneration, the cells projecting far from the surface.

374 MI NOT. [Vol. II.

Many parts of the vessels are filled with coagulum, suggesting thrombi formed during life, as has been asserted to occur normally in the human placenta. For the most part, the vessels contain normal blood, save that there is an excess of leucocytes ; in some vessels, however, there are large clear refringent bodies which look like vesicles. What these bodies are I am unable to say — possibly they come from breaking down of the endothelium. The multinucleate cells, Fig. 14, are large and very much crowded ; they contain each a dozen nuclei, more or less. I have nothing of importance to add to the previous descriptions.

In the outer zone we notice at once that the expansion of the blood-vessels is far less active near the muscularis than further in ; indeed, we might subdivide the zone into an outer compact and an inner cavernous layer. The vascular epithelium is far degenerated, Fig. 16 ; A is a surface view ; B and C vertical sections; each cell forms a more or less independent projection ; the cells vary extremely in size ; the nuclei are either single or multiple ; in the former case they may be small and comparatively regular, or large and very irregular in shape ; in the latter case they are of unequal sizes. The perivascular cells are innumerable ; their appearance is indicated by Fig. 1 5 ; but where the blood-vessels are wider, or, in other words, towards the glandular zone, they exhibit signs of breaking down ; the signs in question are indistinctness of outline, granular appearance of the protoplasm, and the difficulty of staining the nuclei. As the changes are slight, they are perhaps accidental. It must be left for future examination of later stages to show whether they do break down or not. I also think that there is a tendency for the multinucleate cells to invade the territory of their uninucleate neighbors.

The peri-placenta agrees with the outer zone of the placenta in its parenchymal structure, except as to two points : 1°, it is now invaded to a slight extent by the multinucleate cells, at the spot nearest the placental glands ; I have no reason to suppose that these cells actually migrate into the peri-placenta, but presume that they arise in siUi ; 2°, near the ob-placenta there are in some parts young monster cells lying close under the epithelium ; the evidence is better here than anywhere else I have observed that the monster cells arise from the epithelium.

The ob-placenta now has monster cells throughout almost its


entire extent, but the greatest accumulation is where they first are developed, directly opposite the placenta. In this region (Fig. 17) they occupy not only the connective tissue of the mucosa, a, b, but also the territory of the circular muscular coat, where they lie, c, between the bundles, imisc, of muscular fibres, which they have forced apart to make room for themselves. The smallest monster cells, a, are found nearest the lining epithelium, ep ; those at the base of the mucosa, b, are bigger, but the biggest of all are those which lie in the outer part of the muscularis, c ; if, therefore, the cells arise from the epithelium and migrate outwards, they must grow while they move. My preparations show in the nuclei of the monster cells certain large, deeply stained fragments which are perhaps chromatine, Fig. 18. Owing to the stretching of the uterine walls, the regenerated glands of the ob-placenta are no longer follicular as at thirteen days (Fig. 10), but are again stretched out, so as to approximate a second time to the form of shallow, open cups, which they had at eleven days (Fig. 6) ; but where the monster cells have accumulated most (Fig. 17), the only distinct trace of the glands is the irregularity of the free surface covered by epithelium, ep.

The embryo and its appendages do not show much alteration in the parts concerning us in the present article. We may, however, note especially two changes in the outer germ layer. 1°. On the strip of ectoderm between the vena terminalis of the yolk sack and the points where the ectoderm joins the placenta, there are a number of thickenings, which form small papillae upon the outer surface of the layer. These outgrowths are solid ectoderm, and like the buds of the villi of the human chorion contain no mesoderm. Whether these structures do become actual villi in later stages, I am unable to say. 2°. Over the yolk sack the ectoderm has become a cylinder epithelium, of which the outer surface is irregular, each cell projecting a little more or less than its neighbors. A similar modification occurs in the opossum according to H. F. Osborn, 61 A, 378-379, PI. XVII., Fig. 4, and Selenka {Entwickehmgsges. d. TJiiere, Taf. XXVIII., Fig. 5). It is probable that the ectoderm assumes this modification in other mammals, where it remains attached to the yolk sack owing to failure to form a complete chorion.

376 MINOT. [Vol. II.

§ II. Summary.^ — In the resting uterus of the rabbit there are six longitudinal folds. The ovum attaches itself on or between the two folds nearest the mesentery, and the placenta is there developed ; the two adjacent lateral folds form a cushion (the peri-placenta) about the placenta, but the two folds opposite the mesentery are flattened out by the stretching of the walls to form the swelling to contain the embryo ; they constitute the ob-placenta. In the region of the placenta the mucosa undergoes an enormous hypertrophy : there is likewise an enlargement, but much slighter, of the peri-placenta.

The entire epithelium lining the uterine swelling degenerates ; its nuclei proliferate, and its protoplasm hypertrophies, becoming at the same time hyaline and granular. The degeneration affects the glands also. The degenerated epithelium becomes vacuolated and in large part resorbed. The process goes on with distinctive features in each of the three primary divisions of the swellings.

The connective tissue increases by hyperplasia in the peri-placenta and to a still greater degree in the placenta, and is transformed for the most part into uninucleate perivascular decidual cells, but also in part, — namely, immediately below the glandular layer of the placenta, — into large multinucleate cells. In the placenta, and to a less extent in the peri-placenta, there is a new formation of blood-vessels, which subsequently enlarge to great size, although their only walls are an endothelium which undergoes rapid hypertrophic degeneration.

In the placental region the uterine epithelium degenerates and disappears, but the glands are preserved as irregular anastomosing rows of coarse granular matter, with numerous vacuoles and scattered nuclei, but without central lumina. Below the glands is a zone containing wide vessels and large multinucleate cells. The outer layer has wide blood-vessels, with numerous uninucleate decidual cells, which arise from the connective tissue cells and arrange themselves in successive coats around the blood-vessels until they occupy the entire room between the vessels.

The embryo is attached at first to the surface of the placenta

^ It will be remembered that the observations cover the period of from six to fifteen days only, and do not include the eleventh day, when several important developments occur.


only by the ectoderm, to which the mesoderm soon joins itself. As soon as the coelomatic fissure appears, we can speak of a foetal chorion adhering to the placenta. When the allantois grows out, it forms the stalk of connection between the embryo and the placental chorion. After the development of the chorion, the free surface of the placenta is, of course, covered by mesothelium (the epithelium of the coelom). Outgrowths of the chorion penetrate the glandular layer of the placenta ; whether these outgrowths are in the form of villi in the sense that they preserve a covering of foetal ectoderm was not ascertained, although the tips of the outgrowths appear to have such a covering, and there is no mingling of the foetal with the maternal circulation. The coelom of the embryo does not extend to the edge of the placenta next the peri-placenta, but the mesoderm does, and is covered by ectoderm.

In the peri-placenta, the glands degenerate and disappear completely, but the covering epithelium is reconstituted except on the part near the placenta. The blood-vessels and connective tissue change as in the outer zone of the placenta, though later. At the fifteenth day a few young monster cells were found near the surface.

In the ob-placenta the degeneration and resorption affect only the surface epithelium and the upper part of the glands ; the deep portions remain as a series of shallow cups, having been stretched transversely by the expansion of the ob-placenta ; the epithelium of the cups unites into a new continuous layer ; the glands grow up into follicles and are again stretched out by the expansion of the walls. Meanwhile there appear monster cells, which probably arise by the hypertrophy and migration of single cells of the epithelium ; they are characterized by the granular hyaline appearance of their bodies, by the coarse granulation and large scattered fragments of chromatine of their nuclei, and by their hugeness. The monster cells continue to enlarge and subsequently invade the whole thickness of the annular muscularis. They appear first and are always most numerous directly opposite the placenta, but they are ultimately present throughout the ob-placenta.

The relations of the embryo having been outlined in § 8, with the aid of Plate XXIX., it is not necessary to recapitulate them ajrain.

378 M/iVOT. [Vol. II.

§ 12. Comparison with other rodents. — The history of the rabbit's placenta elucidates also that of the Guinea pig, of which we possess descriptions by Bischoff, Ercolani, 89, Creighton, 77a, 77b, Tafani, 134, and others. These authors being unaware of the nature of the metamorphoses of the uterine glands, and not knowing the disappearance of the foetal ectoderm over the placenta, but, on the contrary, seeking for foetal placental villi, lacked the necessary basis for a correct interpretation. Ercolani was further misled by his erroneous belief that the placental tissues of the mother arise as new formations, not as metamorphosed constituents of the uterine mucosa, but coming after the assumed but non-occurrent complete destruction of the mucosa. Tafani' s work betrays gross inaccuracy, for he based his figures and descriptions upon schematic notions, based in their turn upon very superficial, and often entirely false, observations. To justify a judgment so severely unfavorable, it is necessary only to direct examination to some of Tafani's plates. His drawings of the human placenta, for instance, I.e. Tav. VII., leave a great deal to be desired ; in Fig. i the sections of the villi are altogether too large and too few ; the separate triangle of tissue at the edge of the placenta does not exist ; the decidua is represented without any compact layer, and its gland cavities are made into blood-vessels. The section of the rabbit's placenta (Fig. 2, Tav. IV.) is even more open to criticism, since it is impossible to determine the foundation of observation. Ercolani, on the other hand, was an observer of considerable ability, and his numerous memoirs on the placenta are valuable, although his hypothesis of iieoformazioiie led him to adopt an unfortunate terminology which makes it difficult to follow him. Creighton observed with more impartial objectivity. That Bischoff was a first-class observer every one knows ; he never leaves any confusion between what he saw and what he inferred ; for us he has the disadvantage of having written before the developments of recent histology. On the whole, we probably do best to turn to Ercolani, who figures 89, Tav. IV., Fig. I., a section of a placenta of a Guinea pig near full term. Let us compare it with the rabbit's placenta.

It is discoidal, pedunculate, and bilobed. The upper surface is covered by a thin epithelium beneath which is a layer of vascular connective tissue, Z, extending over the sides of the pla


centa, g, f, f, and down between the lobes, g : the epithelium therefore corresponds entirely with the placental mesothelium of the rabbit. The upper portion of the placenta, /, corresponds to the meshwork of degenerated glands in the rabbit's placenta. The layer of epithelium, i, i, m, covering the side of the placenta, corresponds to the foetal ectoderm in a similar position in the rabbit ; at an earlier stage it resembles very closely in appearance what I find in the rabbit (Creighton, 77a, p. 560, PI. XIX., Fig. b, c, c, c,). Deep down under the space between the lobes of the placenta comes the sub-placenta, Ercolani's cotyledonary organ, O, which was compared above with the sub-placenta of the rabbit ; the thick pedicle of the placenta, e, n, n, corresponds to the sub-glandular layers of multinucleate decidual cells, which has encroached upon and apparently replaced the outer zone of uninucleate decidual cells, which is present earlier, as it is in the rabbit. At the side of the placenta is the peri-placental thickening, g.d, and springing from it the so-called reflexa, c, which is probably only the peri-placenta hypertrophied. The reflexa is entirely absent in the rabbit. In regard to what I suppose to be the glands, P, neither the descriptions nor the figures of Ercolani suffice to indicate their character.

The interpretation offered differs in nearly every respect from Ercolani's own ; and yet though I have no preparations of the Guinea pig's placenta, and am acquainted with the organ only through the publications of others, I think the homologies drawn may be accepted with considerable security ; but let me add that I am well aware that their actual justification can come only from the specimens.

Sections of the rat's placenta near full term show that the structure in that species is strictly comparable to what exists in the rabbit. The surface is covered by a thin epithelium overlaying a vascular connective tissue layer ; the vacuolated tubular glands, very much degenerated, occupy the greater part of the placenta, leaving only a thin vascular zone from which the outer zone is lost, and which is therefore occupied solely by the much altered sub-glandular zone of multinucleate cells. There are many differences in details of structure from the rabbit, but the fundamental likeness is self-evident.

As the similarity of the placentae of various rodents has been noted by previous authors, it is probable that the type of placental organization is the same throughout the class.



[Vol. II.


The following observations are of a fragmentary character, but may serve to round out our information in certain respects. Some of the facts have already been recorded in the series of numerous embryological articles contributed to Dr. Buck's Reference Hajidbook of the Medical Scioices ; but as that work is for consultation rather than the publication of original observations, it will hardly seem a mistaken repetition if I include here some things already published there.

§ 13. AUantois and umbilical cord. — Prof. W. His has shown that the entodermal cavity of the allantois is the terminal stretch of the entodermal canal ; the posterior end of the body is prolonged into a mass to which he gives the name of "Bauchstiel" [Atiatoinie menschlicher Embryonen, III., 222226), and which develops in the same general manner as the


Cuts. — Diagrammatic section of the Bauchstiel of a human embryo, modified from W. His. Am, amnion; md, medullary groove; z/, z*, veins; /f, /4, umbilical arteries; All, allantois; coe, coelom.

body proper, having a rudimentary medullary groove, a somatopleure and splanchnopleure, Cut 2. It is morphologically the hind portion of the body. After its closure and separation from the amnion it appears as the umbilical cord. Its development requires that the umbilical cord should be covered, not by the amnion, as it is almost universally stated, but by an extension of the foetal epidermis. Histological examination shows that this is the case. The amnion is characterized by the ectoderm remaining a single layer of cuboidal or low cylinder cells, and by the matrix of mesoderm being distinct, owing to its high refran

No. 3-]



gibility. The foetal skin is characterized by the ectoderm becoming many layered, while the cutis remains for a long time undifferentiated from the mesoderm below, and the matrix is of low refrangibility. In comparing the ectoderm of the umbilical cord with the skin, therefore, we do not expect to find any differentiated cutis. The epithelium of the cord is at first, of course, single layered, the condition which is permanent over the amnion. In the cord of a three-months embryo, Cut 3, I find the two-layer stage. The outer layer is granular, and in some parts each cell protrudes like a dome. Dome cells also appear on the young epidermis, and as I learn from Dr. J, T. Bo wen, who has

'M /

Cttt J. — Epithelial covering of the umbilical cord of an embryo of three months. X 545 diams.

been investigating the subject in my laboratory, are probably the precursors of the epitrichium. The cells of the inner layer are larger and clearer than those outside. By the fifth month the epithelium is distinctly stratified, and the superficial layers consist of flattened cells similar to those of the horny layer of the skin at an early stage. The ectoderm of the cord agrees therefore entirely with that of the embryo proper in its general development, but the differentiation proceeds more slowly, so that at any given age the ectoderm of the cord is at a less advanced stage than that of the embryo.

The appearance of the cord in cross-sections is instructive. Cut 4, A, is a section through a cord of sixty days ; the right umbilical vein is already aborted ; the coelom, coe, is a large cavity, and contains the yolk stalk, Y.S, with its two vessels.



[Vol. II.

Cut 4. — Two sections of umbilical cord. A, at sixty days; B, at three months; V, vein; Ar, artery; All, allantois cavity; Coe, ccelom; Y, yolk sack ; X 22 diams.

and its entodermic cavity entirely obliterated. Near the embryo the coelom may become much enlarged, and is often found during the second month and even later to contain a few coils of the intestine, as has been long known. Above the body cav

Cm/ J. — Connective tissue of the umbilical cord of an embryo of 21™™; X 540 diams., stained with alum-cochineal, and eosine.

No. 3.]



ity is the duct of the allantois, All, Hned by entodermal epithelium ; and in this region are situated the two arteries and single vein ; the section is bounded by ectoderm.^ The further development of the cord depends upon three factors : 1°, the growth of the connective tissue and blood-vessels ; 2°, the abortion of the coelom, yolk stalk, and allantoic duct in the order named ; 3°, differentiation of the connective tissue and of the ectoderm.

^^^ /V/> Ui'k^ Vn-- ^'




(J •' ?, pvS^

1; ^•'^'"- #>f ($






C?// 6. — Connective tissue of the umbilical cord of a human embryo of about three months, X 511 diams. Stained with alum-cochineal, and eosine.

The growth and differentiation of the mesoderm proceeds rapidly, encroaching upon the coelom, which is obliterated (early in the fourth month). At first the connective tissue, Cut 5, is Composed merely of numerous cells embedded in a clear substance ; the cells form a complex network, of which the filaments and meshes are extremely variable in size ; the nuclei are oval, granular, and do not have always accumulations of protoplasm about

^ The ectoderm is often wanting, owing to its frequent destruction /(Jj/ mortem.

384 MINOT. [Vol. II.

them, forming main cell bodies. I notice, also, a few cells which I suppose to be leucocytes, but see no other structures. By the end of the third month the cells have assumed nearly their definite form ; the protoplasm has increased in amount, and forms a large cell body around each nucleus. Cut 6. The network has become simpler and coarser, the meshes bigger, and the filaments fewer and thicker ; in the matrix are numerous connective tissue fibrils, not yet disposed in bundles, except here and there ; as they curl in all directions many of them are cut transversely, and therefore appear as dots. In older cords there is an obvious increase in the number of fibrils, and they form many bundles. In the cord at term the matrix contains mucin, and may be stained by alum haematoxylin ; at what period this re

\ ^^'^ i.-TrTE3

Cut 7. — Section of the allantois from the umbilical cord of an embryo of three months; ent, entoderm; mes, mesoderm. X 340 diams.

action is first developed I have not ascertained. I have observed nothing to indicate the presence of special lymph channels in the cord at any period, but I have not investigated the point. Tait's lymph channels are merely the intercellular spaces. The tube of allantoic entoderm increases very little in diameter after the second month ; compare A and B, Cut 4. It is very persistent, appearing usually even in the cord at full term, at least in the proximal end, according to Kolliker {Entiuicklungsges. 2te Aufl., p. 34). After the second month it is a small group of epithelioid cells, with distinct walls, irregularly granular contents, and round nuclei ; around the cells, ent, which may or may not show a remnant of the central cavity, there is a


slight condensation, vies, of the connective tissue to form, as it were, an envelope. This structure has been regarded by Ahlfeld and others as the persistent yolk sack. I think the correct interpretation was first suggested by Kolliker.

§ 14. Amnion. — The tissues of the amnion do not progress beyond an early embryonic stage ; the ectoderm remaining at the one-layered stage, the mesoderm preserving much of the primitive matrix. Emery {Arch. Ital. Biol., III., 'i^']) has directed attention to the primitive homogeneous matrix of the vertebrate mesoderm, and especially to the separate sub-epidermal layer of the embryo, which contains no cells at first. In the human






Cut 8. — Two sections of the placental amnion : A, from an embryo of the eighth month; B, at term; ect, ectoderm; ??ies, mesoderm; a, layer of mesodermic cells.

X 340.

amnion there is a non-cellular layer under the epithelium, as is well shown in Cut 8, A and B. Sometimes this layer is invaded to a certain extent by connective tissue cells, B ; in other cases the portion of the matrix towards the chorion acquires a fibrillar character. A, as if partially resorbed, but in no case have I seen the matrix entirely altered from its primitive character. The cells of the mesoderm lie in lacunae ; they are flattened in the plane parallel to the surface, and hence in vertical sections. Cut 8, appear more or less fusiform. They present no special features, so far as I have observed, to distinguish them



[Vol. II.

from other embryonic connective tissue cells. Their bodies have little affinity for coloring-matters, hence it is difficult to follow the processes by which the cells are united. Their nuclei are at first round or oval. After the third month they often show a great variety of alterations in shape and size, Cuts 9, lo ; some of the nuclei are then very large, with a distinct net-work, d ; others are smaller and differ but slightly from the normal ; some are very irregular, b, and others again strangely elongated, a ; many other forms beside those represented in Cut 9 are to be found. The changes indicated I consider of a degenerative character, and in fact many of the nuclei are

Cu^ g. — A natural group of nuclei from the mesoderm of the amnion of a foetus of the fifth month. X 1225 diams.

breaking down, for one finds in some specimens every stage between a nucleus and scattered granules, — nuclei, nuclei with indistinct membranes, nuclei without membranes, masses of granular matter, clusters of granules crowded together, and finally other clusters more or less scattered. This degenerative process may be compared with that described by Phisalix (^Arch. Zool. Expt., Ser. II., T. III., 382) as occurring in the blood cells of the spleen of teleosts. Compare also the chromatine degeneration observed by Flemming to occur in ova of the vertebrate ovary {His and Bmunes Archiv. 1885, 221-244). In the human amnion the nuclear degeneration described is not always to be recognized so clearly, although the nuclei in all amnia older than three months, which I have observed, are more or less

No. 3.]



irregular and distorted. Finally it is to be added that not infrequently the cells form a distinct epithelioid layer upon the surface of the amnion next the chorion, as represented in Cut 8, B, a.

C?// /o. — Mesodermic nuclei of the amnion of an embryo of about four months. X 713 diam.

The epithelium of the amnion varies in appearance, as seen in transverse sections. Usually the cells are cuboidal or low cyl





Cut II. — .Surface view of the amniotic epithelium of an embryo of 144 days; stained with alum — hsematoxyline, and eo*ine. //, protoplasm; pr, intercellular processes; nu, nucleus. X 1225 diams.

388 M/NOT. [Vol. II.

inders, Cut 8, A, each one with a rounded top, in which is situated the more or less nearly spherical nucleus ; sometimes, however, the nuclei lie deeper down. Less frequently the epithelium is thin, Cut 8, B, and its nuclei, which are transversely elongated, lie further apart. It is probable that these differences are not structural, but conditional upon the greater or less degree to which the amnion is stretched. I have observed no constant differences between the placental and the remaining amnion. The most interesting peculiarity of the epithelium is best seen in surface views ; namely, the intercellular bridges. They display themselves with a clearness which I have never seen in other epithelia; see Cut ii.

The nuclei, ;///, are relatively large, rounded with distinct outlines ; they have a more or less well marked intra-nuclear network, with thickened nodes, and a small number of deeply stained granules, which are probably chromatine. Each nucleus is surrounded by a cell body,//; and the adjacent cell bodies are separated from one another by clear spaces. With high powers, as represented in the cut, one sees that these spaces are separated from one another by threads of material,/;-, stretching across as bridges, connecting neighboring cells. Examined attentively, the protoplasm of the cells exhibits a vacuolated appearance. One is thus led to view the epithelium as a sponge work of protoplasm somewhat condensed around each nucleus ; according to this interpretation the intercellular spaces are large meshes of the sponge work, and the intercellular bridges are protoplasmic. A recent paper ^ by M. Manille Ide, which I owe to the kindness of the author, brings a series of interesting observations to show that the intercellular bridges of the Rete Malpighi of the mammalian epidermis are not protoplasmic, but processes of the cell membranes. This paper has led me to reexamine my preparations of the amniotic epithelium, but I have been unable to find in them any indications of membranes around the cells or reasons for considering the intercellular bridges as other than protoplasmatic in constitution. Whether this result is due to the imperfection of my preparations, or is in accordance with the truth, must be decided by further investigation.

^ Manille Ide, La membrane des celljiles du corps muqueux de Malpighi. La Cellule, IV., 2me fasc., 1888.


Meola, 59, ascribes a much more complex structure to the amnion than his predecessors, in which he is followed by Viti, 21. Both of these authors subdivide the mesodermic stratum into three layers : a lamina cotinetivale, next the ectoderm, a sostanza intermedia, and a mcmbrana limitante. As to the histological details, Viti differs somewhat from Meola, but agrees with him in finding a histological distinction between the three layers enumerated. The extent to which I can distinguish three layers is indicated by the description of the mesoderm given above : I have been unable to find the marked structural differences affirmed by Viti. Viti's paper is to be commended for its excellent historical reviews, particularly for his summary of the various theories as to the origin of the amniotic fluid.

§ 15. Cborion. — The human chorion has been the object of greater misconception than perhaps any other organ of the body. Even at the present time there prevail numerous false conceptions concerning it, nor do I know of any text-book which gives a satisfactory or even tolerably correct account of its structure. This singular confusion is not due to deficiency of observations, for from the vast literature of the subject (by trusting the accurate observers, such as Coste, Farre, Kolliker, Turner, Langhans, Waldeyer, etc.), we may cull a fairly complete and exact history of the development of the chorion. But the literature of the chorion consists chiefly of papers of little value, and often remarkable for the gross crudeness of the observations they record, and for the proofs they are of their author's ignorance of other and better investigations. It appears that the anatomists and physiologists, by a species of tacit understanding much to be regretted, have regarded the uterus and placenta outside of their province, and have left the investigation of the anatomy and functions of these organs to gynaecologists and others, whose capacities have lain rather in the direction of medical practice than of original research, although among them are some notable exceptions. The majority of the practitioners who have written on the uterus and foetal appendages have done at least as much harm as good. It would be a sheer waste of time to subject this mass of literature to a critical revision in order to extract from it what little there may be of value. I have, however, read a large number of the articles, and studied those which seemed worthy of it. Upon

3go MINOT. [Vol. II.

this course of reading and a study of my own extensive material I have based the following history of the chorion, which passes briefly over what is known, and dwells upon what is founded on my own observations.

The human chorion as I have defined it ^ is " the whole of that portion of the extra-embryonic somatopleure, which is not concerned in the formation of the amnion." The human chorion is remarkable for its very early complete separation from the yolk sack, and for its precocious development of villi. Both of these developments had already taken place in His' youngest embryo, and in Reichert's ovum, which is supposed to be normal and the youngest known, there were chorionic villi, though no embryo was distinguished. Reichert's description is not satisfactory, his long memoir ^ being principally concerned with speculations.

The villi of the chorion, as shown long ago by the obser-«^ vations of Coste, are formed at first only

^ >' by the ectoderm. I reproduce here Fig.

^ 6, of PI. II., referring to the human

"j ' species from Coste's great work. The

\. , hollowness of the villi and their clumsy

shape are to be especially noted. The mesoderm grows into the villi subsequently. The branches of the villi grow out in a similar manner, the process being led, as it were, by the ectoderm. Orth, in a special paper, 118, has used ^j J these facts to argue against Boll's Priji cip des Wachsthunis. Kollman's obser Cut 12. — Portions of the ...

chorion of an embryo sup- vatlOnS ^ on the growth of Vllh dunng

posed to be about eighteen the fourth week are particularly instruc days; vies, mesoderm; ec, tJyg Jhg OUtgrOWth of the branches Iter Crs'te."' ^""^^ "' ^^ ^^^y ^^P'^ ^^^ ^^^^^^' "^'^^ ^^^"^ ^^ gree of participation of the connective tissue. The two extremes are: i°, a bud consisting wholly of epithelium, which may become a process with a long, thin pedicle, and a thickened free end remaining entirely with 1 Buck's Reference Handbook, Medical Sciences, — Art. Chorion.

2 Reichert, Berlin Akad. Abhandlungen, 1873.

3 Kollman, Arch. Anat. Physiol. Anat. Abth., 1879, 297.

No. 3.]



out mesoderm ; 2°, a thick bud with a well-developed core of connective tissue, and having a nearly cylindrical form. Between these extremes every intermediate state can be found. Other observers have noted this peculiar manner of growth, which I have found still going on in the placental chorion during the fourth month. Robin, 125, appears also to have crudely observed both the young hollow villi, and the solid epithelial buds. The blood-vessels he traces to the division of the cavity of the villi into an artery and a vein ; from the nature of things he offers no observations in support of this assertion.

Only the tips of the villi touch the surface of the decidua, either at first or subsequently, except of course, over the chorion laeve during the abortion of the villi. The tips of the villi are attached to the uterine surface ; they penetrate the decidua for a short distance, but even in the placental area at the close of gestation, the penetration is slight, and the villi make their way only into the surface stratum of the decidua serotina. There is no evidence of any sort that the villi penetrate the glands at any period. The relation of the villi to the decidua has now been so accurately ascertained, that there can be, I think, no longer any question whatsoever on this point. The best discussion is by Langhans, 110, p. 231 ff.

The shape of the villi varies according to the part of the chorion and the age of the embryo. They gradually abort over the chorion Iseve, and gradually grow over the chorion frondosum. Let us begin with the placental villi : At first they are short, thick-set bodies of irregular shape, as shown in Cut 1 2 ; at twelve weeks their form is extremely characteristic, Cut 13 ; the main stem gives off numerous branches at more or less acute angles, and these again, other branches, until at last the termi

Cut ij. — Isolated terminal branch of a villus from the chorion of an embryo of twelve weeks.



[Vol. II.

nal twigs are reached ; the whole of the space between the chorion and decidua is occupied by these ramifications ; the

branches and twigs, as the illustration shows, are extremely irregular and variable, although in general they may be described as club-shaped, being more or less constricted at their bases. The branches may be bigger than the trunk which bears them, or of any less size ; some of the smallest are merely slender outgrowths of the epithelial covering of the villus, such as have already been alluded to. Gradually there is a change. During the fifth month we find the irregularity, though still very marked, decidedly less exaggerated, Cut 14 ; the branches tend to go ^ . ,,.„ . . off at more nearly right angles :

Cut 14. — Villous stem from a y t> & >

placenta of the fifth month. X 9 o^c finds very numerous free ends, diams. as of coursc Only a small proportion

of the branches touch the decidual surface ; the branches, too, are less out of proportion to the stems, less constricted at their bases, or, in other words, less remote from the cylindrical form ; the awkward cucumber shapes of the twelfth week are no longer found except here and there. The change continues in the same direction ; that is, is towards

Cut /J. — Terminal villi of a placenta at full term. The little spots represent the proliferation islands of the covering epithelium.


greater regularity of configuration. It is hardly necessary to describe the intermediate phases that have been examined, but it will suffice to describe the form at full term, Cut 15, when the branches are long, slender, and less closely set, as well as less subdivided, than at earlier stages ; they have nodular projections, like bra;iches arrested at the beginning of their development ; there are numerous spots upon the surfaces of the villi ; microscopical examination shows that these spots are proliferation islands, as we may call them, or little thickenings of the ectoderm with crowded nuclei. It appears that not all the villi change to the slender form ; for some villi, having still the earlier, thicker form, are found even in the mature placenta, a fact already noticed by Jassinsky, 105, 346. These thick villi

^^ C^,



i'^ i^



Cu/ 16. — Section of the chorion at three weeks, a, layer of coagulum; h, mesoderm of chorion; Ep, epithelium, also extending over the villi; Vi and Vi' ; the mesoderm, b, contains a number of blood-vessels, nearly all in transverse section. X 65.

usually show also a distinct " cellular layer " in their ectoderm, a peculiarity to be considered below again. Seller, 131a, has given figures of the villi at various ages, but fails to show the characteristic forms. Langhans has observed the alteration in the villi, llO, 199, and even justly remarks that many of the villi in so-called "moulds" are not pathological, as they have been frequently considered, but normal young villi. The differences in the villi, according to age, are very conspicuous in sections. The sections should of course be made so that the fragments of the villi will remain in situ ; imbedding in celloidine is convenient for this purpose ; if this end be attained, one



[Vol. II.

finds below the chorionic membrane numerous sections of villi ; if the specimen be a young chorion, — first to third month, — the villi are large, with a good deal of room between them ; their outlines are very irregular, and there are relatively few small branches (Cut i6). The older the specimen, the larger the

proportion of small branches. In an old chorion — seventh to ninth month — the number of small villi of nearly uniform size is very striking (see the figure of a section through a placenta in situ, given in Cut 35).

The abortion of the villi of the chorion lasve takes place by an arrest of development and a subsequent slow degeneration of the tissues, which lose all recognizable organization in the protoplasm, and to a large extent of the nuclei ; at the same time they alter their shape (Cut 17), becoming more and more filamentous ; by the fourth month only a few tapering threads, with very few branches, remain. The villi disappear almost completely from the laeve, except near the edge of the placenta, where they are to be found, even in the after-birth, imbedded in

Cut 77. — Aborting villus from chorion of the second month.

Cut 18. — Section of the chorionic membrane of an embryo of three weeks; stained with osmic acid; nies, mesoderm; ect, ectoderm; a, outer, b, inner layer of ectoderm. From a section prepared by Prof Theodor Langhans. X 445 diams.



the degenerated epithelium of the chorion and the upper layers of the decidua, as shown in Cut 25, vi, the epithelium and decidua being so fused at this point that it is impossible to determine any line of demarcation between them.

The chorion, being a portion of the somatopleure, consists, of course, of two primary layers, the mesoderm and ectoderm. During the second half of the first month, the earliest period concerning which we have any accurate knowledge, the mesoderm is already a vascular layer of considerable thickness (Cuts 16 and 18, nies), and the epithelium (ectoderm) has two layers of cells (Cut 18, a and h) ; of which the outer is the darker in specimens stained with osmic acid, carmine, cochineal, or haematoxyline, and has also smaller and more granular nuclei. The same distinction exists in the two-layered stage of the ectoderm of the umbilical cord (Cut 3), and of the foetal skin. Hitherto most authors have entirely overlooked the inner layer at early stages. It was first clearly recognized by Langhans, who directed attention to it in a special memoir. 111, he having already described its later history, llO. In some earlier writers are allusions to the layer. Kastschenko, in his paper on the chorionic epithelium, has also described it, although he has not followed its history very far. The interpretation to be offered seems to me clearly to be, that the chorionic epithelium advances in its differentiation to a stage equivalent to the two-layered stage of the epidermis and there stops ; whatever further change occurs is degenerative.

The two primitive layers of the chorionic epithelium have been more or less clearly observed at later stages by several anatomists, and have been variously interpreted. Ercolani and Turner regard them as absolutely distinct, assigning the deep layer to the chorion as its true and only epithelium, and the outer layer to the uterus, thus enabling themselves to conceive the villi as covered by a maternal as well as a foetal epithelium, so that maternal blood found between the villi is still within the maternal tissue. After accepting the outer layer as maternal, the question as to its origin still remained. Some authors affirmed it to be the uterine epithelium, others to be the lining of expanded uterine blood sinuses. So far as I am aware, no one has made observations to show by the developmental history of the layer, that one or the other of the last mentioned hy

396 MINOT. [Vol. II.

potheses is correct. When we consider the precision and exactitude of Kastschenko's observations, which actual specimens enable one to verify, there is in my judgment no reason left for differing from the conclusion that both layers are parts of the foetal ectoderm.

Governed by the difficulty of accounting for the presence of maternal blood in the intervillous spaces, and therefore apparently outside the maternal tissues, several investigators have been led to seek for at least an endothelium outside the chorionic epithelium. Some authors, as for instance Winkler, have asserted the existence of such an endothelium, but after a prolonged and careful search, I fail to find anything of the kind, and in this result it seems to me the best observers are agreed.

The conclusion, I think, may now be safely formulated that the chorion is covered externally by the foetal ectoderm, and has no other covering in any part except, of course, where the chorion laeve rests upon the surface of the decidua, and where the tips of the villi touch the serotina ; but the morphological distinction holds, and the decidua is no more the covering of the chorion, than are clothes morphologically the covering of the body. I believe further, on grounds stated below, that the conclusion just formulated holds true of the chorion at all periods.

The further history of the chorionic mesoderm is so fully given by Langhans in his invaluable memoir, 110, and Kastschenko, 107, that there is little to be added. In the earliest » stage I have been able to examine, an ovum of the third week, the matrix of the chorionic connective tissue, in a preparation stained with cochineal or hagmatoxyline, and imbedded in paraffine for cutting, appeared hyaline and glistening, owing to its refrangibility (Cut 19) ; it has lacunae in which the cells lie; the cell bodies are either shrunken or colorless, so that lacunae, except for the staining of their contained nuclei, are clear and light. This appearance I find again in specimens a little older. The image is entirely distinct from that of the same layer later, for then the cells are stained darker than the matrix, which at the same time has lost its homogeneous character, and acquired a fibrillated look. Very different from my own sections are several which I owe to the kindness of Professor Langhans of Bern, and which that distinguished investigator informs me are from a three-weeks ovum, which had been preserved in osmic

No. 3.]



acid (see ante, Cut 18). In Professor Langhans' preparations the cells are all stained much deeper than the matrix ; they have an elongated form, and run in various directions more or less parallel to the epithelium ecto ; hence many of them are cut transversely or obliquely. Whether the differences noted are due to the methods of preparation must be decided by preserving the same chorion in part with osmic acid, in part with Muller's fluid or picrosulphuric acid, the latter being the reagents I have used. In specimens of the tenth week, the matrix of the chorionic mesoderm has quite altered in character, being no longer homo



Cut ig. — Section of tne chorionic membrane of an ovum supposed to belong to the third week; ect, ectoderm; mes, mesoderm; a, outer, b, inner layer of ectoderm; stained with alum-cochineal. X 445 diams.

geneous, and at the same time it has increased in thickness. For the most part the matrix stains lightly, and where it is lighter it contains fibrils of extreme fineness, and running curly courses ; there are also streaks of lightly stained matrix, giving the impression of fibres resulting from portions of the primitive colorable matrix being left. In other parts of the layer the primitive matrix is still present, and we find a homogeneous well-colored basal substance, the cell lacunae of which appear light by contrast, as in Cut 19. One can distinguish also the



[Vol. II.

commencement of the perivascular coats, at least of the larger vessels, the matrix being quite dense around them, and the cells elongated almost into fibres, and possessing a slightly increased affinity for coloring-matters. The larger blood-vessels and unmetamorphosed part of the layer occupy a middle portion between the two surfaces, but the smaller blood-vessels lie near the ectoderm (compare Cut 19, li), thus presaging the formation of Langhans' vascular layer {Gefdssschicht). The development of the mesoderm of the chorion Iceve stops at about this stage, or at the stage when the matrix has completely changed from its first state; in the region of the frondosum, however, develop

Cut 20. — Section of the amnion and placental chorion of the fifth month. Ep, amniotic epithelium; Am, amnion; Sir, stroma; Fib, fibrillae layer; fbr, fibrine layer; c, cellular layer; Vi, villi. (From a section cut in celloidine, and stained with Weigert's Hsematoxyline. The drawing is only approximately correct as to details. X 71 diams.).

ment proceeds much further by the production of fibres throug out the whole of the layer ; usually, but not invariably, the fibres become much more numerous near the ectoderm than in the inner part of the mesoderm, thus differentiating a well-marked sub-epithelial fibrillar layer. Cut 20, fib, from the deeper and wider stroma, Str. The fibrillar layer is that commonly spoken of as the connective tissue layer of the chorion : for details of its structure, including the Gefdssschicht^' see Langhans and


Kastschenko. The inner layer, Str, is called the GallertscJiicht by many German writers, and seems to be what Kolliker {E^itwickelungsgeschichte, 2te Aufl., p. 322) designates as " Gallertgewebe zwischen Chorion und Amnion " ; it usually contains a considerable number of large granular wandering cells. Jungbluth, 106a, describes a network of capillaries, which exist during the first half of pregnancy, apparently in the upper part of the stroma, — i.e. next the amnion — but I fail to find any. Where the amnion comes into contact with the chorion the adjacent parts of the two membranes are more or less loosened, forming a network of strands by which the membranes are united : most of the uniting strands appear to belong rather to the chorion than to the amnion. This loose tissue is perhaps that which Kolliker designates as a Gallertgezvebe distinct from the chorion.

Although the chorion bounds the coelom, I have observed no mesothelium upon its mesodermic surface ; but I have not made search for it by any special methods. In the rabbit, it will be remembered, the mesothelium is very evident over the placenta, but the rabbit differs from man by the absence of union between the amnion and chorion. Nor have I been able to find any basement membrane, properly so called, under the chorionic ectoderm. As to the appearance which suggests it, I accept Kastschenko's explanation, 107, 455.

The mesoderm in the villi is differentiated otherwise than that of the membrane of the chorion. In the youngest stage I have examined there is some of the primitive matrix present in the villi ; and I presume that earlier the whole mesoderm has the same character. In my specimen (three weeks) the change is progressing. I have not succeeded in satisfying myself as to the process of change which takes place, but I think it probably essentially as follows : The cells gradually develop large bodies and acquire a more decided affinity for coloring-matters ; meanwhile vacuoles appear in the matrix, presumably by its modification into a new substance; the vacuoles increase in size and number, transforming the matrix into a network and ultimately causing its total disappearance, leaving the intercellular spaces filled entirely with the new substance, which has come from a metamorphosis of the original matrix; probably this new substance is more or less fluid, since wandering cells are scattered freely through it. Leaving this half-hypothetical history, let us



[Vol. II.

pass on to direct observations. In the placental villi of embryos of four months and older, the mesoderm exists in two principal forms, — adenoid tissue and fibre-cell tissue around the bloodvessels. The adenoid tissue, Cut 21, is that of which the supposed development has just been sketched ; it may be considered as the proper tissue of the villus. It consists of a network of protoplasmic threads, which start from nucleated masses (cells). There are many large meshes, which are partly occupied by the coarsely granular wandering cells, /, /, which are scattered about, and are usually present in large numbers. About the capilla


^V 1.

Cut 21. — Adenoid tissue of a villus from a placenta of four months. /, /, /, wandering cells; v, v, capillary blood-vessels; J, finer meshwork from near a capillary. X 352 diams.

ries the network is much more finely spun. Kastschenko, 107, 454, found the wandering cells most abundant near the epithelium, but I have noticed no such peculiarity, except that they do not often enter the dense perivascular tissue ; and as the bloodvessels are centrally situated, the adenoid tissue and the wandering cells in it are of course more peripheral. It seems to me that the leucocytes are distributed more or less evenly throughout the adenoid tissue. I fail to recognize any intercellular substance. The abundance of nuclei deserves special mention. Around all the non-capillary vessels the mesoderm is very different, for it exhibits distinct intercellular substance, with a ten


dency to fibrillar differentiation in quite a wide zone around the blood-vessels ; in this zone the cells become elongated and irregularly fusiform ; around the larger vessels the cells are grouped in lamina, making the structure similar to that already described in the walls of the vessels of the umbilical cord ; after the perivascular coats acquire a certain thickness, the cells of the inner layers are more elongated, more regularly fusiform, and more closely packed than those of the outer layer ; the transition from the denser to the looser tissue is gradual. We are perhaps entitled to recognize in the denser inner layer the media, in the outer looser layer the adventitia, although neither of the layers has by any means the full histological differentiation characteristic of the like-named layers of the blood-vessels of the adult.

The epithelium of the chorion becomes differentiated in three different ways : 1°, upon the chorion frondosum ; 2°, upon the chorion laeve ; 3°, upon the villi. For a correct knowledge of the remarkable changes which the epithelium undergoes, particularly in the placenta, we are indebted to the remarkably exact investigations of Langhans, 110 and 111. This author left two points of importance unsettled ; namely, the origin of his ^^ ZellschichtJ' and of the " canalisirtes Fibrin!' Kastschenko has traced the cellular layer {Zellschicht) to the epithelium, as already stated : compare pp. 463-469 of his memoir, 107. My own observations show, I think conclusively, that the canalized fibrine arises through a degenerative metamorphosis of the epithelium, which begins in the outer layer and may invade the inner layer (Langhans' Zellschicht) . Let us consider separately the three series of modifications of the chorionic, ectoderm.

In the region of the chorion frondosum the inner layer of the ectoderm (the cellular layer of Langhans) becomes irregularly thickened in patches, which present every possible degree of variation as to number and as to their breadth and thickness. Although at first the cellular layer is more or less continuous and composed of uniform cells, this is not the case in later stages. We must assume that with the growth of the membrane the epithelium increases in area, but remains in many places single layered, developing no "Zellschicht." The patches of cells have been well described by Langhans, 110, and Kastschenko, 107, 466, and are represented with lower power in



[Vol. II.

Cut 20, Cf and with a higher power in Cut 22, c. They varymuch in appearance : the cells are more distinct in the small patches, but are less individual in the large patches, owing to the spread of the process of degeneration into the layer, Cut 22, c. The cell bodies are lightly stained, and the granular nuclei are not very sharply defined and vary in size and shape. The cellu





"^.:^S^ ^P

Cut 22. — Placental chorion of an embryo of seven months; vertical section through the ectoderm and portion of the adjacent stroma, mes, mesodermic stroma; c, cell layer; fb, fibrine layer; ep, remnant of epithelium. X 445 diams.

lar layer is always sharply defined against the stroma, although there is no true basement membrane, but towards the outer layer of the ectoderm its boundary is sometimes distinct, sometimes lost in a gradual transition.

The outer layer of the ectoderm of the frondosum is even more variable. As stated by Kastschenko, it is primitively a dense protoplasmic reticulum, with nuclei in a single layer and


without any cell boundaries. In the chorion frondosum at four months and after I find spots where this structure still prevails, either with or without an underlying cellular layer ; in other spots the layer is thickened and contains an increased number of nuclei, which are sometimes crowded in a bunch ; elsewhere the layer is thinned out and has no nuclei ; in still other spots the thickening has gone on much further, and usually, but not always, where the outer layer is much thickened the cellular layer under it is also thickened ; wherever it is thickened, and occasionally where it is thin, the outer layer of the ectoderm shows a marked tendency to degenerate into canalized fibrine. Cut 20, fbr, and Cut 22, fb. It is not difficult to assure one's self that the fibrine arises by direct metamorphosis of the ectoderm. I now think that its formation begins in the outer layer and thence spreads into the cellular layer ; for, in fact, when both layers are distinguishable, as in Cut 22, the fibrine layer, fb^ is always external, and the external layer of nucleated protoplasm has either totally disappeared or is represented by mere remnants, as in Cut 22, ep. The fibrine layer consists of a hyaline, very refringent substance permeated by numerous channels. Cut 22, fb; the substance has a violent af^nity for carmine and hasmatoxyline, and is always the most deeply colored part of a stained section ; the channels tend to run more or less parallel to the surface of the chorion and are connected by numerous cross-channels ; some of the channels contain cells or nuclei. This complex system of canals is by no means of uniform appearance in all parts of the placenta, both the spaces and dissepiments varying in size and shape. The fibrine often sends, as shown in Cut 22, long outshoots into the cellular layers upon which it seems to encroach. The frequency of these images in my preparations led me to the opinion ^ that the fibrine arises from the cellular layer only, and I concluded that the ectoderm was first transformed into the so-called cellular layer, which was then transformed into fibrine. It still appears to me that much of the degeneration goes by these stages ; but, on the other hand, it seems clear that the degeneration begins, as above stated, in the outer layer. Another appearance is presented by the ectoderm where it is thickened and wholly transformed into the cellular layer. In brief : the ecto ^ Anatom. Anzeiger, ii. 23.

404 MIXOT. [yoL. II.

derm of the placental chorionic mesoderm undergoes patchwise manifold changes ; it exists in three chief forms : i°, the nucleated protoplasm ; 2°, the cellular layer ; 3°, canalized fibrine. A patch of the ectoderm may consist of any one of these modifications, of any two or of all three, but they have fixed relative positions, for when the nucleated protoplasm is present, it always covers the free surface of the chorion ; when the cellular layer is present, it always lies next the mesoderm ; and when all three forms are present over the same part, the fibrine is always the middle stratum. In general terms it may be said that the amount of canalized fibrine increases with the age of the placenta, but it is very variable in its degree of development. The peculiar layer into which the ectoderm is transformed has long puzzled anatomists. E. H. Weber recognized the fibrine layer and described its appearance correctly ; it has probably been often seen, but generally regarded as either pathological or a blood coagulum. Robin, for instance, may be cited, 125, 70-71, as one who saw, without observing correctly and understandingly, the tissue in question. An important gain was made when Winkler recognized the modified ectoderm as a constant layer, and in 1872 directed especial attention to it under the name of " Schlussplatte," 152. Kolliker {EntwickcliingsgescJiicJite, 2te Aufl., 337) added essentially to our knowledge of its structure, but it is to Langhans that we owe the first clear light. Meanwhile, other writers, following the lead of Ercolani and Turner, 146, 551-553, have been influenced chiefly by the presence of the cellular layer, in the large size of the elements of which they found a resemblance to the decidual cells, which has guided them to the conclusion that the cellular layer is derived from the wall of the uterus. This error has been definitely corrected by Kastschenko, as already stated. In further support of the conclusion that the chorionic cellular layer is not decidual, may be brought forward the fact that there is a certain immigration of decidual cells into the placenta at its margin ; but they remain entirely distinct from the cells of the cellular layer. This is readily seen in radial sections through the margin of a placenta from a normal after-birth — compare below, the account of the ectoderm of the chorion laeve. The origin of the canalized fibrine from blood, which Langhans left in his first paper as an open possibility, and which

No. 3.]



even so recent a writer as Ruge, 129a, 123 and 130, has advocated, cannot be maintained. Of course, there may be a deposit of blood fibrine (coagulum), but it would be pathological, and therefore to be distinguished from the normal fibrine of ectodermal origin. Moreover, the microscopic appearance of a blood clot or thrombus is so extremely characteristic that one can readily distinguish it from the placental canalized fibrine.

The ectoderm of the villi of the placenta differs from that of the chorionic membrane in several respects : 1°, the cellular layer after the first month becomes less and less conspicuous, and after the fourth month is present only in a few isolated patches, known as the Zellkjioten, and carefully described by Langhans and Kastschenko ; both of these authors were impressed by the resemblance of the cells to those of the decidua serotina ; Langhans concludes that the Zellknoteii arise from the serotina, but Kastschenko, having traced their development from the chorionic epithelium, denies his predecessor's conclusion, but still clinging to the idea of a genetic connection between the Zellknoten and the decidua, reverses the reasoning, and concludes that the decidual cells arise in part at least from the Knoten. Neither of these authors have found the intermediate forms between the two types of cells, and when we examine their descriptions critically we find that they have really no evidence except the likeness of the cells to offer in favor of their genetic relationship, and accordingly Langhans expresses himself with characteristic caution. To me the resemblance appears altogether superficial ; hence my conclusion that the Zellknoten are remnants of the cellular layer. 2°, For the most part the villi remain covered by the nucleated protoplasm, which in many places is thickened. In the later stages these thicken

Cut 23. — Cross-section of a villus from a placenta of seven nnonths; three blood-vessels are shown; a, a, thickenings of the ectoderm; /, a thickening transformed into canalized fibrine. X 222 diams.

4o6 MINOT. [Vol. II.

ings are small and numerous, constituting the so-called " Prolifemtio7is-inseln" : compare Cut 15. Many of the little thickenings appear in sections of the villi, Cut 23, a, a, and here and there are converted into fibrine, /. I have interpreted them (Wood's Reference Handbook of the Medical Sciences, V., 695) as commencing buds, and consider that in earlier stages they grow into branches, but in later stages are in part at least arrested in their development. 3°, The proliferation islands are converted into canalized fibrine, and at the same time grow and fuse, forming larger patches, particularly on the larger stems : in this manner are produced the large areas and columns of fibrine found in the placenta at four months and after ; they have been well described by Langhans, and form a striking feature in sections of placentae. Some of the columns, as stated by Langhans, stretch along the villi from the chorionic membrane to the surface of the serotina as if to act as supports. Ercolani appears, if I understand his account, to have seen the fibrine columns, without, however, ascertaining either their structure or their origin. 4°, Over the tips of the villi, which are bent considerably where they are imbedded in the decidua serotina, the relations are not clear ; the epithelium is certainly not present in its original form over the imbedded ends of the villi, which are, however, surrounded by a hyaline tissue of the character of the canalized fibrine, except that the canals are often indistinct or even wanting ; the hyaline tissue forms an almost continuous coat over the decidual surface ; in earlier stages the ectoderm of the terminal villi is often considerably expanded. The natural interpretation of these facts is that the ectoderm of the villi expands over the decidua serotina and degenerates. In this manner we account for both the absence of any cellular ectoderm over the ends of the villi and the presence of canalized fibrine upon the serotinal surface — but the hypothesis must await the final test by observation.

The ectoderm of the chorion laeve loses by the seventh month all traces of the protoplasmic layer, and is without any canalized fibrine, except near the placenta ; cf infra. It is transformed into a ZcllscJiicJit. In a section of the lasve in sitn at seven months, Cut 33, the chorionic ectoderm, c, rests directly upon the decidua, which has no epithelium of its own. The ectodermal cells lie two or three deep ; they are described by Kolli

No. 3.]



ker and Langhans, the former designating them as the chorionic epitheHum, while the latter doubtfully traces their origin to the uterus. That Kolliker {Entwickelnngsgeschichtc, 2te Aufl., p. 322) is right, I am confident. It is easy to follow the layer of cells in question at the edge of the placenta, and see that it is directly continuous with the cellular layer of the frondosum, which it resembles in character. On the other hand, the ecto

Cut 24. — Placenta at full term. A, vertical radial section through the margin; D, decidua; vi, aborted villi outside the placenta; Cho, chorion; Si, circular sinus; Vi, placental villi; Fib, canalized fibrine. B, portion of A more magnified to show the decidual tissue near b\ v, blood-vessel; d.d' , decidual cells; d, with one, d' , with several nuclei.

dermal cells of the Iseve are distinct in character from the decidual cells next to them, Cut 34, having smaller and more darkly stained nuclei, and much more coarsely granular protoplasm ; the ectodermal cells are much smaller than the decidual. The ectoderm is sharply marked off from the decidua, but its surface



[Vol. II.

is often corrugated, and then the line of separation between the tissues is irregular, and in sections it may even appear that there is a true interpenetration and mingling of the decidual and ectodermal cells ; but it is only apparent, and the demarcation is always preserved.

At the edge of the placenta, as shown by examination of afterbirths, the relations of the layers are somewhat different. I reproduce with a few additions the descriptions given in my article on the Placenta'^ of a radial section through the margin of a normal placenta discharged at full term. Cut 24, A, from which the amnion had been removed. The chorion, Cho, and decidua, D, are in immediate contact at the left of


•>■■■ ■■■■■■,' '.T^-T^TE^i.-*

Cut 2^. — After-birth at full term; vertical section of the aninion, chorion, and decidua in their natural relations near the placenta, atn, amnion; r-4o, chorion; f, cellular layer or ectoderm; f, fibrine and decidual tissue, degenerated; D' , decidual tissue. X 125 diams.

the figure ; that is, outside of the placenta, though remnants of the aborted villi, vi, are still plainly recognizable ; but, as stated previously, they occur only in the immediate neighborhood of the placenta. These villi are surrounded by hyaline matter which resembles and can be followed into continuity with the canalized fibrine layer, Fib, covering the surface of the decidua serotina and the fibrine layer of the chorion frondosum. Below the aborted villi, vi, of the chorion laeve, the fibrine layer is broken down and penetrated by the decidual tissue, so that the demar

^ Wood's Reference Handbook Medical Science, V., 694, 695.


cation between the foetal and maternal tissues is here lost, and in fact, at the edge of the placenta the decidual cells make their way into the chorionic tissue, and for a certain distance towards the centre of the placenta they are found lying chiefly in the ectoderm. In other placentae the fibrine layer and the decidual tissue around the margin of the placenta have not only intergrown, but also undergone a common degeneration. Cut 25, in consequence of which all distinct structure is obliterated, and we find the villi, viy imbedded in a stratum, /", of more or less colored substance, without definite organization except irregularly scattered nuclei. Attentive examination shows that this layer,/", has unmistakable remains, c, of the cellular layer next the mesoderm of the chorion, and that it passes into an outer layer, D\ in which the traces of decidual structure are unmistakable ; the dark line at the lower edge of the decidua, D\ is merely detritus and coagulum, as is often found on after-births. If we follow the layers in this, or a similar specimen, in the direction away from the placenta, the layers gradually alter, losing their degenerated character, until we reach a point where the chorionic ectoderm and the uterine decidua both exhibit their normal features. Returning now to the placenta we were previously considering. Cut 24 : The placental chorion begins to exhibit its characteristic stratification a short distance within the margin. I have found, however, that the distinctness of that stratification varies considerably, not only in different placentae, but also in different parts of the same placenta. The decidua, D, outside the placenta is very thick, but at the edge of the placenta it begins to thin out, and as it passes over the under side of the placenta, rapidly becomes so much reduced as to be even less in thickness than the chorion, cho. The decidua is everywhere crowded with an immense number of decidual cells, but in some other specimens they are less crowded. The surface of the decidua serotina is covered by a layer of fibrine, easily recognized by its deep staining ; this coat of degenerated material has not yet received the attention it deserves, as a feature of the human placenta, which is quite constant, so far as my observations go ; as stated previously, I consider its origin to be the epithelium of the ends of the villi imbedded in mucosa. Up to the edge of the placenta the chorion laeve and decidua are united ; at the edge they separate, to make room for the



[Vol. II.

villi, Vi, Vi, of the frondosum. In the angle, Si, where the two membranes first separate there are very few villi, so that there is a comparatively clear space left, which is known as the circular sinus. It is not, as some of the older writers have believed, a distinct vessel, nor does it extend as a clear space completely

Cut 2b. — Placenta at full term, doubly injected by Dr. H. P. Quincy, to show the distribution of blood-vessels upon the surface; the arteries are drawn light; the veins dark. X 0.7 diams.

around the placenta ; but, on the contrary, it is interrupted here and there by an ingrowth of villi. In the cut, the spaces occupied of maternal blood are left white ; the foetal blood-vessels are drawn black.

The chorionic circulation is complete in itself. The single vein and the two arteries of the umbilical cord spread out over


the surface of the chorion, marking their course by projecting ridges. The insertion of the cord is always, so far as I have observed, obviously eccentric ; the degree of eccentricity varies from a nearly central position to the so-called velamentous insertion, — compare B. S. Schultze, 159; the degree of eccentricity is easily seen to be related to the distribution of the vessels, — a point not mentioned in current text-books. The arteries come down together from the cord, and are usually connected, but not invariably, by a short transverse vessel, situated about half an inch above the surface of the placenta, and which has been noted by many observers. I have never noticed any arterial or venous anastomoses on the surface of the placenta. The two kinds of vessels do not run together ; the arteries lie nearer the surface, the veins deeper. Cut 26 ; the arteries fork separately until they are represented only by small branches and fine vessels ; some of the small branches disappear by dipping down suddenly into the villi below; the veins are considerably larger than the arteries, and some of the larger branches disappear from the surface in the same abrupt manner as do the smaller arteries. There is the greatest possible variability in the vessels of the placenta ; I have never seen two placentas with the vessels alike. The more eccentric the insertion of the cord, the more do the vessels tend to distribute themselves symmetrically ; the more central the position of the cord, the less can any vascular symmetry be made out.

The two following paragraphs are copied without change from my article on the placenta (Buck's Reference Handbook of the Medical Sciences, V., 696, 697) : —

" To follow the course of the foetal blood-vessels within the placenta, the best method is by corrosion injections. These may be made either with fusible metal, wax, or celloidine. The first is specially suited for the study of the large trunks ; the latter, for that of the smaller vessels also. I have a very beautiful celloidine injection by Dr. S. J. Mixter, which, with others of wax and metals, has served as the basis of the following description : The veins leave the surface somewhat more abruptly than do the arteries, which gives off more small branches to the surface than do the veins. Cut 26. Both kinds of vessels leave the surface by curving downward for a short distance into the trunk of a villus ; the vessels then divide, and their branches



[Vol. II.

again take a more horizontal course ; the branches then curl over downward, and, after a second short descent toward the decidua, again send out horizontal branches. The result of this arrangement is a terrace-like appearance in the course of the vessels ; they approach the uterine side of the placenta in this very characteristic manner. The number of terraces is variable ; usually there are two or three, but sometimes there is only one, or they may number four or even five. Arrived at the end of its terraces, the main vessel takes a more nearly perpendicular

course, and rapidly subdi\p 1/ vides. Immediately after

entering the villi, the arteries and veins give off but few capillaries, but after a short course in the main stalk of the villus, the vessels give rise to many branchlets, and gradually the character of the circulation changes, until in the smallest villous twigs there are capillaries only, Cut 27. The vascular trunks have a marked tendency to dichotomous division, which is maintained within the villi to a certain extent ; the arterioles and veinlets in the mature placenta go from their trunks at wide angles for the most part, and subdivide in the same manner, so that they spread out through the whole substance of the placenta. The vessels next the decidua take a more horizontal trend, like the top branches of a wind-swept tree. As the vessels run in the villi, of course the way in which the latter branch out determines the paths of the former ; hence by following the distribution of the vessels we inform ourselves as to the ramifications of the villi. Thus the horizontal course of the vessels on the uterine side of the placenta corresponds to the wellknown fact that the ends of the villi attached to the uterus become bent and adhere by their sides to the decidual surface."

Cui sj. — Portion of an injected villus from a placenta of about five months; magnified 210 diams.

No. 3.]



"The capillaries of the villi are remarkable for their large size, and on this account have been described as arteries or veins by E. H. Weber, Goodsir, and other writers. Their calibre is often svifiEicient for from four to six blood-disks abreast. They are very variable in diameter, and also peculiar in exhibiting sudden restrictions and dilatations, Cut 27. In the short bud-like branches there is often only a single capillary loop, but as the branch becomes larger, the number of loops increases, and they form anastomoses. In branches large enough to serve as a stem, some one or two of the vessels may be enlarged, as may be seen in Cut 27 ; in the branches large enough to admit of it, there are two (or sometimes only one) longitudinal central vessels, an artery and vein, and a superficial network of capillaries. Cut 2^ a. Goodsir and other early writers laid great stress on the formation of the capillary loops, but this feature is a common one in the development of the foetal vascular system, as is also the width of the capillaries. In my opinion these peculiarities are characteristic rather of the foetus than

. Cut 21 a. Placenta of about

specifically of the placenta. In some of ^^^ ^^^^^^. p^^^j^^ ^^ ^ ^^.^ the older writers (Goodsir, Farre, et al) villus, to show the central ves it is asserted that the true capillary SyS- sels and superficial capillaries.

tem disappears toward the end of ges- ^ '°^ '^"^^' tation. I am unable to confirm this, but find instead that in the slender terminal villi of the placenta at term there is often only a single, sometimes long, capillary loop ; the capillary is very wide, and its width is probably the reason of its having been held formerly to be a vein or an artery."

§ 16. Uterus during menstruation. — I have little to add to the descriptions of previous authors, particularly those of Leopold, 36, and Kolliker.^ It is, however, worth while to present the accompanying illustration, Cut 28, since there is a lack

1 Kolliker's Haiidbiuh der Geiuebelehre, 5te Aufl., p. 563.



[Vol. II.

O R>


of figures. The cut represents a transverse section of the corpus uteri of a fine specimen, for which I am indebted to Dr. W. W. Gannett. The woman died from acute miliary tuberculosis ; the autopsy was made almost immediately after death, and within four hours from death the complete genitalia were placed in MuUer's fluid, the uterus having been first carefully opened by a single median ventral incision. Death is said to have occurred on the day of the regular period. The hymen was intact. There was no sign of pathological change in any of the genitalia. In one ovary, the right, there was a fresh corpus hcBinorrJiagiciim. These data afford a sufficient basis for the belief that the uterus was well preserved in a perfectly normal condition.

The mucous membrane is from 1.1-1.3 mm. thick; its surface is irregularly tumefied ; the gland openings lie for the most part in the depressions. In the cavity of the uterus there was a small blood-clot. The mucosa is sharply limited against the muscularis, Cut 28. In transverse sections one sees that the upper fourth of the mucosa is very much broken down and disintegrated, Cut 28, d\ the cells stain less than those of the deep portions of the membrane ; as represented in the cut the tissue is divided into numerous more or less separate small masses ; some of the blood-vessels appear torn through, but it is difflcult to make sure observation : Overlach, 39, considers it probable that the infiltration of blood takes place by diapedesin, not by rupture of the capillaries. The superficial epithelium, ep, is loosened everywhere ; in places fragments of it have fallen off, and in some parts it is gone altogether ; it stains readily with cochineal and its nuclei color well, the epithelium differing in this respect from the underlying connective tissue, which does not stain well ; the blood-vessels in the disintegrated layer are for the most part small.

The deeper layer of the mucosa is dense with crowded wellstained cells, which lie in groups separated by clearer lines ; in the cut this grouping shows less plainly than in the preparation ; the lighter channels are perhaps lymph-vessels, a suggestion which occurs to me, because in so-called "moulds" one sometimes finds similar channels crowded with leucocytes. The cells appear to be the proliferated interglandular tissue ; there are very few leucocytes, so far as I can distinguish ; the cells have

4l5 MINOT. [Vol. II.

small, oval, or elongated, darkly stained nuclei, with a very small granular protoplasmatic body each ; there is certainly no noticeable enlargement of the cells, but only a remarkable multiplication. The point is important ; I see nothing to suggest the presence of decidual cells, nothing even like definite enlargement of any of the cells. The image of the tissue is comparable to that of the connective tissue of the rabbit's placenta at six days, except that there the cells are widely separated, here closely crowded, but in each case the cells are small, with little protoplasm, and connected by their processes. In another specimen in my possession of a normal uterus at the close of menstruation, the condition of the mucous membrane agrees with that of the specimen we have considered, except, of course, that the disintegrated superficial layer is lost, and that the superficial layers stain poorly. In this second specimen, also, the interglandular cells are small and very crowded ; there are few leucocytes and no decidual cells. The two specimens further agree in having the glands distended and contorted ; each gland is surrounded by a distinct basement membrane or layer of connective tissue cells closely investing the epithelium, as has been observed by Leopold, 36. In my article on the decidua in the Reference Handbook, II., p. 390, is a summary of the changes occurring during menstruation, and stress is there laid upon two points emphasized by previous writers ; namely, the increase in the number of leucocytes and the presence of decidual cells. Since my own observations have failed to confirm these statements, I can no longer accept them. The proliferated connective tissue cells are those, probably, which become decidual cells when the decidua vienstriialis is changed into the decidua graviditatis — compare the account of the one month's uterus in the next section.

§ 1 7. Uterus one month pregnant. — The specimen to be described came from a woman who committed suicide by violence, not by poison, and I was informed that she was known to be about one month pregnant. Further information was not • obtained, and I was requested not to seek it. The specimen was received in very fresh condition, but it had been opened, the reflexa was torn and pretty much gone ; the embryo had been removed, and I was therefore unable to verify the age, or investigate the attachment of the villi of the chorion to the

No. 3.]



Uterus. There was a beautiful corpus hitciirn in one ovary, quite similar to that figured by Dalton in his Report on the corpus luteiim in the transactions of the American Gynaecological Society for 1877, Fig. 9. The surface of the uterus seemed uninjured. The specimen was hardened in Miiller's fluid, and found subsequently to be well preserved. It may be considered, I think, perfectly normal.










muse — ".:--• — -- " — — • — " — -^

Cut 3g. — Uterus one month pregnant ; outlines of the glands from a vertical section : to show the divisjon of the mucosa into an upper compact layer, D' , and a lower cavernous layer, D" ; gl',gl", glands; art, spiral artery; muse, muscularis.

My specimen enables me to confirm in most respects Turner's accurate description of two uteri of about the same age, 146, 546-548. The inner surface shows the hillocks {Inseln) de



[Vol. II.

scribed by Reichert in the uterus of two weeks, studied by him, which have been figured by Coste in slightly older specimens, and found by Turner also, 146, 540.

The three illustrations given herewith are all from sections through what I suppose to be the placental region.

There is an upper compact layer. Cut 29, D\ and a lower cavernous layer, Z>" ; the caverns, being gland cavities, which appear as rounded areolae lined with epithelium, are filled with broken-down epithelial cells. The drawing, reproduced in Cut 29, was obtained by drawing the outlines very carefully, stippling the areas occupied by the connective tissue, representing the blood-vessels by double outlines, and omitting the glandular

Cut 30. — Uterus one month pregnant; portion of the compact layer of the decidua seen in vertical section; coagl, coagulum upon the surface; d, d', decidual cells. X 445 diams.

epithelium altogether. It will be noticed that about threefourths of the diameter of the mucosa is occupied by the cavernous layer, D".

The upper or compact layer is shown in Cut 30. The surface


is without any trace of epithelium, but is covered only by a thin fibrous and granular coagulum, coagl ; the tissue itself consists almost exclusively of young decidual cells, d, d\ with a clear homogeneous matrix; here and there are leucocytes, but they are nowhere numerous ; the decidual cells are all quite large, with their bodies deeply stained by the eosine ; the nuclei are round, oval, or slightly irregular in shape, coarsely granular, and sharp in outline ; the cells themselves, though irregular and variable in shape, are all more or less rounded with processes running off in various directions ; scattered between the cells are many sections of their processes ; occasionally it can be seen that two cells are connected ; in fact, we have in this tissue evidently a modified embryonic or so-called anastomosing connective tissue. Now, as we know through the observations of Leopold, 36, which I have verified, the connective tissue of the uterine mucosa consists of anastomosing cells, and as stated in the previous section, the cells are found proliferated in the menstruating uterus ; we have therefore only to imagine the cells enlarged with certain accompanying modifications, to obtain the tissue figured in Cut 30. There is no special formation of cells around the blood-vessels, where, according to Ercolani, the decidual tissue arises by new formation. In Turner's specimens the upper part of the compact layer was imperfectly preserved, but according to his description there appears to have been a coagulum similar to that which I have found, but thicker. In the deep part of the layer the cells are less enlarged, and when the cavernous layer is reached, there occurs a rapid transition in the character of the cells, which become smaller and more fusiform, and their nuclei more elongate, smaller, and deeper stained by alum-cochineal. The gland openings upon the surface of the uterus lead into tubes. Cut ZO,gl', which run slightly obliquely through the compact layer, taking a more or less nearly straight course and joining the contorted gland tubes. Cut 30, gl'\ of the cavernous layer. The gland ducts are completely devoid of lining epithelium, which has disappeared except for a very few loose cells, occasionally found lying free in the ducts ; the cells have not fallen out from the sections, but were lost before the tissue was imbedded.^ The ducts then

^ The blocks to be cut were stained in toto with alum-cochineal and eosine, imbedded in paraffine, etc. The sections were fastened on the slide with celloidine, to keep the parts in place.



[Vol. II.

are wide tubes running nearly straight through the upper part of the decidua and bounded directly by the decidual tissue ; they communicate below with a contorted cavity. Similar tubes appear in later stages and have been described as blood-vessels — see the next section.

The cavernous layer contains numerous spaces, the areolae of Turner, 146, 547, who was uncertain as to their character,

though he ascertained that many of them belonged to the glandular system. In my specimen it is perfectly clear that all the larger areolae belong to the glands, which must be extremely distorted and distended to give the shapes shown in Cut 29. The thin dissepiments between the areolae are composed of connective tissue, the long dark nuclei of which. Cut 31, are strikingly different from those of the cells of the compact layer. Cut 30. The areolae present two extreme modifications and all intermediate phases between these two. The smaller areolae are lined by a well-preserved cylinder epithelium,- or by one in which the cells are separated by small fissures ; in other areolae the cells are a little larger, Cut 31, each for the most part cleft from its fellows, and some of them loosened from the wall and lying free in the cavity. The other extreme is represented in Cut 32 ; the size of the areolae is much increased, — compare Cuts 31 and 32, — both drawn on the same scale ; the epithelium is entirely loosened from the wall, and the cells lie separately in the cavity which they fill ; the cells are greatly enlarged, their bodies having three or four times the diameter of the cells in the small areolae ; they have not the cylinder shape, but are irregular in outline : their protoplasm is finely granular and stains rather lightly ; the nuclei are large, rounded, glandular, and with sharp outlines ; they are less darkly stained than the nuclei of the epithelium of Cut 31.

Cut ^i. — Uterus one month pregnant; section of gland from cavernous layer, with the epithelium partly adherent to the walls. X 445 diams.

No. 3.]



The obvious interpretation of the appearances described is, that the glandular epithelium is breaking down, that it is lost altogether from the ducts, but is still present in the deep portions of the glands ; in breaking down the cells separate from

Cut ^2. — Uterus one month pregnant; section of gland from cavernous layer, with the epithelium loosened from the walls; X 445 diams.

one another, and then from the wall, and falling into the gland cavity, there enlarge, the cavity enlarging also. Similar appearances are also found in "moulds" of the second month; very likely they have been often observed and mistaken for pathological changes.

The blood-vessels of course lie in the dissepiments between the glands. I observed nothing to correspond with the " colossal capillaries dilated into small sinuses," mentioned by Turner 146, 548. Were not these supposed capillaries gland cavities, from which the epithelium had fallen out } Occasionally the sections pass through a spiral artery. Cut 29, art, which is cut again and again as it twists around in its characteristic separate column of connective tissue.

§ 1 8. Uterus seven months pregnant, with the foetal membranes in place. — The specimen to be described was obtained for me through the kindness of Dr. W. W. Gannett. It is an apparently normal uterus, which contained a normal embryo weighing 11 50 grammes and having an umbilical cord 58 centimetres long, — probably about seven months old, or a little



[Vol. II.

more: there were no data as to the duration of gestation. The uterus was opened, and preserved in Miiller's fluid without disturbing the membranes.

A section through the amnion, chorion laeve, and uterine

U ^-*^..

mucosa, stained with haematoxyline, and viewed with a low power, is represented in Cut ii ; the dark spots are maternal blood-vessels, which have been shaded for the sake of clearness. The amnion, am, and chorion, cho, present the characteristics

No. 3.]



previously described, §§ 14, 15 ; the chorion is bounded against the decidua by an epithelium c, which I interpret as the chorionic ectoderm ; there is no trace of a second layer of epithelium ; so that the uterine epithelium must be considered lost, a conclusion agreeing with the observations of KoUiker, Turner, and myself upon earlier stages, and the statements of Ercolani. The decidua has eight or nine times the thickness of the chorion ; it has an upper compact and a lower cavernous layer ; the former contains numerous decidual cells, most of which are a little larger than those nearer the muscularis ; the compact layer contains a few blood-vessels of moderate calibre, and occa


Cut ^4. — Uterus about seven months pregnant; upper portion of decidua vera, with the chorion laeve in situ, mes, mesodermic layer of chorion; ep, epithelial layer of chorion; Z*', decidua. X 340 diams.

sionally a large vessel, v, surrounded by connective tissue containing no decidual cells. Examined with a higher power, the decidual cells — compare Cut 34, D' — are found to resemble quite closely those at one month. Cut 30, but they are much more numerous and closer together, and their processes are fewer ; they vary also more in size ; some of the larger ones are multinucleate ; it is probable that the cells are multiplying by division ; the matrix presents a fibrous look, but whether it contains actual fibres, I am not sure ; between the decidual cells are a



[Vol. II.

0</jj._ Section through a normal placenta of seven months, z« situ. Atn, amnion; Cho, chorion; Vi, villus trunk; vi, sections of villi in the substance of the placenta; A decidua; yiA, muscularis; /)', compact layer of decidua; / d-, uterme blood-vessel (or gland?) opening into the placenta. The foetal blood-vessels are drawn black; the maternal blood spaces are left white; the chorionic tissue is stippled, except the canalized fibrine, which is shaded by lines; the remnants of the gland cavities in Z?" are stippled dark. (Drawn from nature by J. H. Emerton.)


certain number of nuclei, some of which belong to leucocytes, others to blood capillaries, and still others, which I am uncertain about, which are few in number, and possibly belong to connective tissue corpuscles. The cavernous layer resembles now, in contrast to the first month, the upper layer of the decidua in histological constitution, but the decidual cells are smaller and at little wider intervals from one another ; the cavernous layer is especially characterized by the slit-like spaces in it ; some of these spaces, as indicated by the drawing, Cut 33, are undoubtedly blood-vessels or sinuses, but still others contain no blood, or at most three or four isolated corpuscles, although close to them are capillaries gorged with blood ; once in a while a few epithelioid cells can be seen adhering to the walls of the spaces. These spaces can hardly be assigned to the vascular system ; they have been held by Kundrat and Engelmann, ISO, and various subsequent writers, to be the gland cavities ; we have not sufificient observations to establish the actual metamorphosis of the areolae of the one month's uterus into the slits, ^/, of Cuts 33 and 35, D' , but there is no ground to question the occurrence of the change, which appears to be a necessary consequence of the stretching of the decidua due to the expansion of the uterus during pregnancy.

A complete section through the placenta in situ and uterus is represented in Cut 35, which has already appeared in my article, " Placenta " (Buck's Handbook, V., 696), and been sufficiently described. The chorion is separated by a dense forest of villi from the decidua, D ; the ends of some of the villi touch and are imbedded in the decidual tissue ; these imbedded ends are without epithelium, but their connective tissue is immediately surrounded by hyaline substance. The decidua is plainly divided into two strata — cf. infra. The section passes through a wide tube, Ve, which opens directly into the interior of the placenta and contains blood; in my article, I.e., this opening is referred to as that of a vein, the identification being in accordance with my understanding of the descriptions of Waldeyer,i 149. Professor Langhans has since informed me, that according to his own observations the opening of the arteries are characterized by the absence of villi projecting into their openings.

^ I am under much obligation to Professor Waldeyer for an opportunity to examine some of the injected specimens upon which his very important researches were conducted.



[Vol. II.

His pupil, Raissa Nitabuch, has since published a dissertation, 117, confirming this opinion, according to which the vessel shown in Cut 35, Vc, is not venous, but arterial. Another possibility has occurred to me, viz. : that it is a gland duct ; in fact, it resembles very closely the undoubted gland ducts of the one month's decidua: there is no reason apparent why the gland ducts, which pass nearly vertically through the compact layer, should be obliterated ; on the contrary, one might expect to find them widened by the stretching of the uterus ; as there is

Cut ^b. — Uterus of seven months, vertical section of the decidua serotina from near the margin of the placenta. 7nc, muscularis; Z>', Z)", decidua serotina; D' , cavernous or spongy layer; Z?", compact layer; Vi, scattered chorionic villi. The intervillous spaces were filled with blood, which rs not represented in the figure, X 50 diams.

blood in the intervillous spaces, it could easily make its way into the distended glands, and its presence there would not prove the glands to be blood-vessels. While, therefore, I accept Waldeyer's researches, 149, as well as those of Langhans and Nitabuch, 117, as verifying Farre's neglected account, 172, 722, of the placental circulation, I venture to express a note of caution as to the danger of mistaking glandular for vascular openings.

No. 3.]



The following additional points deserve notice : The serotina is about 1.5 mm. thick, and contains an enormous number of decidual cells, Cut 36. The cavernous, D\ and compact layers, D'\ are very clearly separated ; the mucosa is sharply marked off from the muscularis, mc, although scattered decidual cells have penetrated between the muscular fibres. The muscularis is about 10 mm. thick, and is characterized by the presence of quite large and numerous venous thrombi, especially in the part towards the decidua. The decidua contains few bloodvessels. Upon the surface of the decidua can be distinguished a special layer of mingled hyaline and decidual tissue, which in many places is interrupted by the ends of the chorionic villi, as is well shown in Cut 36. The supposed gland cavities of the spongy layer, D\ are long and slit-like ; they are filled for the most part with fine granular matter, which colors light blue with haematoxyline ; they also contain a little blood, sometimes a few decidual cells. I have seen in them also a few oval bodies several times larger than any of the decidual cells, and presenting a vacuolated appearance ; what these bodies are, I have not ascertained. In places the glandular epithelium is distinct ; its cells vary greatly in appearance, neighboring cells being often quite dissimilar ; nearly all are cuboidal, but some are flattened out ; of the former there are some with darkly stained nuclei, but the majority of the cells are enlarged, with greatly enlarged hyaline, very refringent nuclei.

The decidual cells are smaller and more crowded in the cavernous layer, and mostly larger in the compact layer — compare

Cut ^j. — Decidual cells from the section represented in part in Cut 36. a, b, d,f, various forms of cells from serotina; c, giant cell from the margin of the placenta; e, clear cells from chorion; at a, seven blood globules have been drawn in to scale. X 545 diams.

428 M/NOT. [Vol. II.

Cut 36. The largest cells are scattered through the compact layer, but are most numerous towards the surface. The decidual cells exhibit great variety in their features, Cut 37 ; they are nearly all oval disks, so that their outlines vary according as they happen to lie in the tissue ; they vary greatly in size ; the larger they are, the more nuclei they contain ; but I observe no cells with more than ten nuclei. The nuclei are usually more or less elongated ; the contents of the cells granular. Some of the cells present another type ; these are more nearly round, clear, and transparent, e ; the nucleus is round, stained lightly, and contains relatively few and small granules ; such cells are most numerous about the placental margin. ^

§ 19. Uterus twelve hours after abortion at six montlis. — For this specimen, also, I am indebted to Dr. W. W. Gannett. The woman was brought into the Boston City Hospital in a comatose condition ; the foetus, estimated to be about six months, was removed by the forceps ; the mother died twelve hours later ; the autopsy by Dr. Gannett showed death to have been caused by tubercular meningitis. The uterus is apparently normal ; I received it in a fresh state, and hardened it in Miiller's fluid. It was already very much contracted ; the mucosa measured about 2 mm. in thickness ; its surface was ragged and more or less covered with clotted blood, presenting very much the appearance so superbly figured by Coste {Devcloppeme7it des Corps organises, PI. X., Espece humaine).

Vertical sections, Cut 38, show that the surfaces of the mucosa are very uneven ; on the free surface there is a thin layer of clotted blood, coagl; the upper or compact layer of the decidua has entirely disappeared, leaving only the deep portion, Dy permeated by numerous large empty spaces, which I take to be in part gland cavities, in part blood sinuses, both changed from their slit-like form by the contraction of the uterus during and since the delivery of the child. Between the spaces are the brownish and hyaline cells, and a great many blood-corpuscles, which lie throughout the tissue itself as well as in the blood-vessels. In short, the conditions found agree with those described by Leopold as present in the uterus a short time after normal delivery at full term, 36, and accordingly, further details concerning my specimen may be omitted.

^This and the preceding paragraph are taken with sundry alterations from my article on the placenta, l.s.c.

No. 3.]



§ 20. Origin of decidual cells. — Besides the erroneous hypothesis of Ercolani, there are three views as to the origin of the decidual cells known to me, to wit : 1°, they are modified leucocytes (Hennig, Langhans, ct al) ; 2°, they arise from the connective tissue cells of the mucosa (Leopold, 36) ; 3°, they


are produced by the epithelium (Overlach, 39). The first view is not supported by observation, even by its advocates, and may be dismissed. Overlach's observations certainly favor the third view, but inasmuch as he has studied only one uterus with pseudo-menstruation from acute phosphorus poisoning, his the

430 MI NOT. [Vol. II.

ory cannot be accepted definitely until verified by further observations on normal uteri. Overlach found in the cervix of the uterus in question, the lining epithelial cells to contain an endogenous brood of small cells, one to fifteen in each parent-cell ; the daughter-cells begin as nuclei, around which there gathers a protoplasmatic body for each. The cells are like the young decidual cells just below, so that the latter may be assumed to have wandered forth from the epithelium. I may recall that in the normal menstruating uterus I find no true decidual cells, and consequently I must regard Overlach' s find as pathological.

The observations of Creighton, of Masquelin and Swaen, and of myself may be fairly considered to establish the fact that in rodents, at least, the decidual cells arise from the connective tissue cells of the mucosa. That they arise from the same cells in man is rendered extremely probable by the investigations of Leopold, which have been confirmed and extended by the observations recorded in §§ i6, 17, and 18, of the present article. Accordingly I assent to the second of the views above enumerated.

Ercolani erroneously regarded the decidual tissue as a new formation, arising after the total destruction of the mucosa. He observed the degenerative processes of the uterine epithelium, and the arrangement of the decidual cells around the vessels of the placenta in rodents and other mammals ; he inferred that the whole mucosa was degenerated and lost, but he never established the inference by observation ; he also inferred that the perivascular cells, being different from the surrounding tissues, were a new formation, but he never traced the actual genesis of the cells. In spite, however, of the absence of the observations necessary to establish his double thesis of the total destruction of the mucosa and the new formation of the decidua, he advocated his doctrine with the greatest earnestness, even to the last — see 91, 92. The failure of his hypothesis to find acceptance has been due not to any unreadiness to bestow merited acknowledgment upon his researches, but to the incompatibility of the hypothesis itself with the ascertained facts of the structure and development of the placenta. While, therefore, we utilize Ercolani's numerous and valuable observations, it will be a distinct gain for science to set aside his theory of the new formation of the decidua.


§ 21. General considerations. — We are now in a position to compare the changes in the uterus during menstruation and gestation. In both cases the processes begin with tumefaction and hyperaemia of the mucosa ; they continue with hyperplasia of the connective tissue (the decidual cells being regarded as modified connective tissue corpuscles) and with hypertrophy, accompanied by distention and contortion of the glands ; they both close with casting off the superficial layers of the mucosa, after which follows the regeneration of the membrane. The essential steps, then, are the same in both cases. The difference is, that during the long life of the decidiia graviditatis, changes supervene in the tissues which do not take place during the rapid menstrual cycle ; the mucosa of gestation is distinguished by the loss of both its surface and glandular epithelium, and by the enlargement of its connective tissue cells into so-called decidual cells. We must accordingly view the changes in the uterus during gestation as a prolonged and modified menstrual cycle. The relation in time between menstruation and the commencement of pregnancy is attributable to the menstrual process rendering the uterus receptive ; that is to say, capable of receiving and retaining the ovum. We must conceive that the ovum has no power of initiating the development of a decidiia, but only of modifying the menstrual process ; hence pregnancy can begin only at a menstrual period. The ovum, too, exercises this influence at a distance, for in all mammals, the earliest development of which is known, the ovum passes through its segmentation in the oviduct (Fallopian tube), and takes from three to eight days to reach the uterus ; but during this period the change in the womb is going on. The most plausible explanation of this action of the ovum at a distance is a reflex stimulus passing from the oviduct to the central nervous system of the mother, and thence back to the uterus ; the validity of this hypothesis must be tested by physiological experiment. That the nerves are able to effect morphological changes is already abundantly proven, not only by the influence of the secretory nerves upon gland cells, by the degeneration of muscular and other tissues, when their nerves are severed, but also by certain embryological observations tending to show that histological differentiation does not progress very far until the tissues are joined by the outgrowing nerves.

432 MINOT. [Vol. II.

When the ovum reaches the uterus, it appears to exert a more direct influence, for one set of changes occurs in the placental area, where there is concrescence of foetal and maternal parts ; another in the region around the placenta (peri-placenta, decidua reflexa), and still another in the rest of the uterus (decidua vera, ob-placenta). Whether the three zones enumerated can be distinguished in the pregnant uteri of all placental mammals, and w^hether they have more features in common than appear from a direct comparison between man and the rabbit, are questions to be decided by increased knowledge. However, it already seems very probable that the decidua reflexa and peri-placenta are homologous at least in rodents.

Concerning the evolution of the amnion nothing definite is known, nor do the speculations of Balfour {Comparative Embryology, II., 256) nor of van Beneden and Julin, 44, 425, seem satisfactory, although the view of the latter is suggestive. They say : —

" ' Dans notre opinion, la cause determinante de la formation de renvellope amniotique reside dans la descente de I'embryon, d^terminee elle meme par le pois du corps. C'est par une acceleration du developpement que la cavity amniotique en est venu a se former quand I'embryon ne possede encore qu'un pois insignifiant.' The chief objection to this theory is that it really gives no cause for the expansion of the somatopleure and chorion ; there is no proof that a mere strain of weight can cause the cells of a membrane to proliferate, and since such proliferation is the immediate cause of the growth of the amnion, van Beneden and Julin must assume for their theory that the strain of weight does cause proliferation ; but this assumption lacks support. Moreover, they give no evidence to show that the embryo /;/ utero is situated in the primitive amniota upon the upper side of the ovum, although it is probable such is the case." ^

Ryder's theory, 19, of the origin of the amnion, like that of van Beneden and Julin, to which he does not refer, is purely mechanical ; but Ryder seeks the cause in a rigid zona radiata, forcing the embryo down into the yolk. See his summary. I.e., p. 184. So far as we know, however, the embryo of the Sauropsida cannot be said to sink into the yolk, and so lead to the development of an amnion ; but, on the contrary, the amniotic

1 Quoted from Buck's Reference Handbook, I., 140.


folds rise up clear above the yolk. Moreover, the formation of the amnion is really a very complex process, part arising from the pro-amnion, part by a dilation of the pericardial cavity {ParictalJidhlc), and part as the extra-embryonic tail folds. These facts speak, in my judgment, unequivocally against the amnion having arisen by the sinking of the embryo into the yolk sack. Nor is there any justification, I think, for seeking these simple mechanical explanations, which are worthy of Herbert Spencer, since the formation of the amnion depends upon inequalities in the growth power of the germ layers, and only such explanation can be valid as explains that inequality — which Ryder's hypothesis fails to do, so far as I can see.

As regards the evolution of the placenta, we are in the dark. Contrary to prevalent opinion, it is not an organ of the allantois, nor is it an organ of the yolk sack. On the contrary, it is always, so far as we know, an organ of the chorion, and begins its development by a differentiation of that membrane. The allantois is a secondary and later structure. Its primitive role is apparently only that of a stalk of connection between the chorion and embryo. There is no evidence to show that the tissue of the allantois spreads out over the chorion to form the mesodermic layer thereof, but the mesoderm of the chorion is proper to it as much as to any part of the somatopleure the mesoderm thereof. When the allantois becomes a large sack, we have a subsidiary change, so that we are brought squarely to the conclusion that the foetal placenta is chorionic. From this premise phylogenetic speculation must start. Further, we know through the discovery of fundamental importance by His that the allantois cavity is at first a small entodermal tube lying in a posterior prolongation of the body {BaucJisticl), and that at this time the so-called allantoic vessels run to and branch out upon the chorion ; the placental differentiation of the chorion has already begun, without participation of the allantois, the enlargement of which, when it occurs at all, occurs at a later stage. To speak, therefore, of an allantoic chorion as do Balfour and Selenka {Studieu iibcr EntwicJiclungsges,, p. 135) is unjustifiable. Nor can we trace the origin of the placenta to the yolk sack, since in most mammals the mesoderm does not spread over the yolk until quite late, so that the yolk sack consists, as in the rabbit and opossum, in large part of ectoderm

434 MINOT. [Vol. II.

and entoderm only, and is without vessels, and therefore unable to form a placenta, which, however, is developing meanwhile from the chorion.

We seek nowadays, following the lead of Professor Cope, to deduce mammalia from the reptilia. Since the reptilia have a free allantois, it is a temptation for embryologists to seek to trace the placenta to a modification of the allantois ; but the placenta of mammals appears in the embryo before the allantois becomes free, and the great size of the allantoic vessels is connected primitively not with the allantois, but with the already important chorionic circulation. The placenta is interpolated in the ontogeny of mammals before the specialization of the allantois, which functions as the vascular pathway between the embryo and the chorion, both primitively and permanently. The enlargement of the allantois, which takes place in certain mammals, is a supervening change, probably a survival of reptilian ontogeny. The question is, not how is the connection of the allantois with the placenta (chorion) established in mammals, for it exists from the start, ^ but what becomes of it in reptiles and birds.

Ryder's theory, 128a, of the origin of the discoidal placenta 2 by constriction of the villous area of the zonary placenta, is difficult to accept. The placenta, being chorionic, cannot of course develop, except so far as the chorion is differentiated ; that is to say, so far as the ectoderm (exochorion) is underlaid by mesoderm. Now, in mammals, the chorion, as mentioned above, does not go at first but part way over the yolk sack, even at the period when the development of the placenta has begun. Accordingly, so far as our present knowledge enables us to judge, the discoidal is probably the primitive placental type. If the chorion is completed by the further extension of the mesoderm around the yolk sack, then the placental formation also may spread, and a diffuse type arise. At present, the whole subject is very obscure, but there is certainly no sufficient evidence to prove that the diffuse placenta is the primitive type.

In conclusion, let me point out that we have no satisfactory

^ This is beautifully shown by Selenka's investigations on the opossum, cited in the text.

^ The human placenta is 7iot discoidal, but metadiscoidal.


knowledge of the nutrition of the embryo. We know positively scarcely more than that the maternal and fcetal circulations are brought very close together in the placenta. We infer that there must be a transfer of nutritive material from one blood to the other. As to zvJiat material is transferred and hozv, we have only theories, but of them an abundance. Under these circumstances, the best beginning is undoubtedly a frank acknowledgment of our ignorance.

§ 22. Summary. — The following paragraphs attempt to give the more important of the conclusions reached in the second part of this paper.

§ 13. The umbilical cord is not covered by the amnion, but by an extension of the foetal epidermis. Its coelomatic cavity is completely obliterated during the third month, and a little later the stalk of the yolk sack is resorbed. The allantoic epithelium persists as a tube or cord of cells for a long period. The blood-vessels have specialized walls derived from the surrounding mesoderm, but have no true adventitia. Connective tissue fibres begin to develop during the third month.

§ 14. The amnion is covered by a single layer of ectodermal cells, which are connected by conspicuous intercellular bridges. It has no true stomata. Its mesoderm consists of anastomosing cells, with a dense matrix ; it is imperfectly divided into three strata, of which that next the ectoderm is without cells, that furthest from the ectoderm is often of a loose texture.

§ 15. The chorion consists of two layers, mesoderm and ectoderm, both of which are present over all parts of the chorion throughout the entire period of pregnancy. The mesoderm has at first a dense colorable matrix, with cells, which color very slightly. During the second month the matrix loses its coloring property, and subsequently the cells acquire a greater affinity for coloring-matters ; the matrix assumes a fibrous appearance, and ultimately in the region of the chorion frondosum connective tissue fibrils appear in it, most numerously next the ectoderm, so that the mesoderm is differentiated there into an outer fibrillar layer and an inner and thicker stroma layer. The ectoderm during the first month divides into two strata, an outer dense protoplasmic layer and an inner less dense cellular layer. In the latter part of pregnancy the whole ectoderm of the chorion laeve has acquired the character of the cellular layer, except close

^^Q MINOT. [Vol. II.

to the margin of the placenta ; at the same period the cellular layer forms a number of irregular patches over the chorion laeve, while the protoplasmic layer remains over the entire surface, both where the cellular is present and where it is absent ; the protoplasmic layer may undergo complete or partial degeneration into canalized fibrine, which is developed in irregular patches. The cellular layer remains on the villi only in a few patches {Zcllhtoteji) and over the tips of certain villi ; the protoplasmic layer of the villi remains everywhere and develops numerous nodular thickenings ; it changes partially into canalized fibrine. It is probable that the fibrine covering the surface of the decidua serotina is derived from the ectoderm of the ends of the villi imbedded in the decidua. The villi are at first of awkward and irregular forms, but their branching gradually becomes more regular, and the twigs acquire a slender and more uniform shape.

§ 1 6. The menstruating uterus is characterized by hyperasmia, by hyperplasia of the connective tissue of the mucosa, and by hypertrophy of the uterine glands ; the upper fourth of the mucosa is loosened and breaks off : there are no decidual cells.

§ 17. The uterus one month pregnant has lost its epithelium from its surface, and from the ducts of its glands ; owing to the dilatation and contortion of the deep parts of the glands, it is divided into a lower cavernous or spongy layer and an upper compact layer ; the connective tissue of the upper layer is transformed into decidual cells ; in sections the glands of the lower layer appear as crowded areolae, which are lined by a cylinder epithelium more or less disintegrated, or else filled with isolated enlarged epithelial cells.

§ 18. The uterus seven months pregnant is without epithelium either on its surface or in the glands, except a few isolated patches in the deep parts of the latter ; there is no trace of the decidua reflexa ; the decidua vera is covered by the epithelium of the adherent chorion laeve ; the decidual serotina is covered for the most part by a layer of fibrine, which is probably derived from the degeneration of the chorionic ectoderm covering the imbedded ends of the villi; the decidua is divisible into an upper compact and a lower cavernous layer, in which latter the gland cavities are reduced to slits ; the decidual cells are very numerous and crowded ; the larger ones lie near the chorion ;


the multinucleate decidual cells are found chiefly in the serotina; at the edge of the placenta decidual cells are found in the chorion.

§ 20. The decidual cells arise by direct enlargement of the connective tissue cells of the mucosa. All parts of the decidua and placenta arise in place by metamorphosis of the tissue ; the mucosa is preserved, and there is no production of placental tissues by new formation.

§ 21. The changes of the uterus during menstruation and gestation are homologous, the menstrual cycle being prolonged and modified by pregnancy ; hence it is that conception takes place only at the menstrual period, for the ovum can only modify the menstrual change, not initiate the formation of a decidua. No satisfactory explanation of the origin of the amnion has yet been offered. The placenta is an organ of the chorion ; its evolution cannot be traced to modifications of either the allantois or the yolk sack ; the allantois is originally the intestinal canal of the Baiichstiel, which serves as the means of vascular communication between the chorion and embryo ; the enlargement of the allantois is secondary. We possess no positive information as to how the placenta performs its nutritive functions.

Boston, Aug. 3, 1888.



[Vol. II.

§ 23. Preliminary Bibliography of Works and Articles specially relating to the foetal envelopes of mammals, exclusive of general works.

  • jit* I shall be much obliged to any who will inform me of errors in and omissions

from this list. I have other titles, but the following are all I have been able hitherto to verify.


(See also Fcetal Membranes.)

1. Ahlfeld, F.

Die Allantois des Menschen und

Arch. f. Gynak., X.,


ihr Verhaltniss zur Nabelschnur.

81-117. Taf. III.

2. Beddard, F.

Note on the presence of an Allan

Zool.Anz.,VII., 653



toic (Anterior Abdominal) Vein in Echidna.


3. Budge, A.

Ueber die Harnblase bei Vogel

Deutsche med. Woch



enschrift, VII., 6970. Gaz. med. Paris, 4

4. Cadiat.



Ser., VI., 97-98.

5. Dastre, A.

Recherches sur l'allantoide et le

Ann. Sci. Nat., 6 Ser.,


chorion de quelques mammiferes.

III., pp. 118. Pis. VII. -X.

6. Dobrynin,P.

Ueber die erste Anlage der Allan

Wiener Akad. Sitzb.,



LXIV, 185-192. Taf. I.

7. Krause, W.

Ueber die Allantois des Menschen.

Arch. f. Anat. u. Phys., 1875, 215; 1876, 204-207.


8. " "

Ueber die Allantois des Menschen.

Zool. Anz., IV., 185.


9. " "

Ueber die Allantois des Menschen.

Z. Z., XXXVL, 175179. Taf. IX.


10. Kuppfer, C.

Die Entstehung der Allantois und

Zool. Anz., II., 520


die Gastrula der Wirbelthiere.

522:593-597; 612617. Marburger Sitz., 70

11. LieberkUhn,

Querschmitte von der Anlage der



Allantois und der Harnblase von Meerschweinchen Embryonen.


12. Olivette,

Ein Beitrag zur Kenntniss der

Wiener med. Jahrb.



ersten Allantoisbildung.

447-450. Taf. XIII.

13. Preuschen,

Die Allantois des Menschen, eine

Wiesbaden, pp. 195.


Franz von.

entwickelungsgeschichtliche Studie auf Grund eigner Beobachtung.

Taf. I.-X.

14. Strahl, H.

Ueber die Entwicklung der Allan

Marburger Sitz., 47


tois der Eidechse.


15. " "

Die Allantois von Lacerta vividis.

Sitz. Marburger Ges.,



No 3.]




(See also Fcetal Membranes.)

16. Ahfeld, F.

17. Hoffmann, C.


18. Hotz, Anna.

19. Ryder, J. A.

20. Schenk, L.

21. Viti, Arnaldo.

22. Winkler, F.


23. " "


Zur Genese der Amnion-zotten (Antvvort zur Wienkler, F. N. 24).

Ueber das Amnion des zweitblattrigen Keinies.

Das Epithel des Amnions.

The Origin of the Amnion.

Beitrage zur Lehre vom Amnion.

L' amnois umano nella sua genesi e struttura ed in rapporto all' origine del liquido amniotico.

Die Zotten des menschlichen Am

Textur, Structur and ZelUeben in der Adnexen des menschlichen Eies.

Erwiderung and Berichtigung zu " Ueber die Zotten des Amnion" (Arch. f. Gynak., VI., 358).

Arch. f. Gynak., VII.,


Arch. f. mikr. Anat., XXIII., 530-536. Taf. XXV.

Bern (Inaug. Diss.).

Am. Nat., XX., 179 185. Arch. f. mikr. Anat.,

VII., 192-201. Taf.

XVIII. Siena, pp. 59. PI. I.

Jenaischr. Z. Med. u.

Naturu., IV., 535 540. Jena (Hermann Da bis), pp. 56. Taf.

I.-II. Arch. f. Gynak., VII.,



1871. 1886. 1868. 1870.



(See also Placenta and Uterus.)

25. Ahlfeld, F.

26. " "

27. Duncan, J.


28. Engelmann,

Geo. J.

29. Ercolani, G. B.

Ueber Befunde an der Decidua vera und reflexa reifer Eier.

Die Beschaffenheit der Decidua des Eies ein Zeichen der Reife oder Friihreife der Frucht.

Notes on the history of the mucous membrane of the body of the uterus.

The mucous membrane of the uterus, with especial reference to the development and structure of the decidua. (This is a translation, with some changes of Kundrat and Engelmann, 180).

Delle glandole otricolari dell' utero e dell' organo glandulare di nuova formazione, etc.

Arch. f. Gynak., XIII.,

290-292. Centralbl. f Gynec,


Edinburg Med. Journ., III., 688-697.

Amer. Journ. Obst., VIII., 30-87. PI. I.

Mem. Accad. Sci. 1st. Bologna. 4to. Ser. 2, VII., 133-207. Tav. I.-X.





[Vol. II.

30. Ercolani, G. B.

31. Friedlander, Carl.

32. Gervais, H.


33. Haussmann.

34. Heintze.

35. Langhans.

36. Leopold,


37. Moricke.

38. Oliver,


39. Overlach,


40. Pfluger,

41. Robin, Ch.

42. Rouvier,


43. Sutton, J.


The utricular glands of the uterus, etc., translated from the Italian under the direction of H. O. Marcy.

Physiologisch-anatomische Untersuchungen iiber den Uterus, I., Die Innenflache des Uterus frisch nach der Geburt und die Neugestaltung der Uterusschleimhaut wahrend des \Vochenbetts, II., Die UterinsinusThrombose an der Placentarstelle.

Sur un uterus gravide de Pontoporia Blainvillei.

Lehre von Decidua Menstrualis.

Ueber den feineren Bau der Decidua.

Die Losung der miitterlichen Eihaute.

Studien ueber die Uterusschleimhaut vfahrend Menstruation Schwangerschaft und Wochenbett.

Die Uteruschleimhaut in den verschiedenen Alters-Perioden und zur Zeit der Menstruation.

Menstruation — its nerve origin — not a shedding of mucous membrane.

Die pseudo-menstruirende mucosa uteri nach akuter Phosphorvergiftung.

Ueber die Bedeutung und Ursache der Menstruation.

Sur quelques points de I'anatoniie et de la physiologic de la muqueuse et de I'epithelium de I'uterus pendant la grossesse.

Recherches sur la menstruation en Syrie.

On menstruation in monkeys.

Boston, 8vo, pp. 305. Atlas, pp. 21. PI. I.XV. 4to.

Leipzig, pp. 57. Tafn. I.-II.

C. R. Paris, XCVII., 760-762.

Beitr. z. Geb. u. Gynak., Berlin, I., 155277.

Centralbl. f. med. Wiss.,XIII., 33-35.

Arch. f.Gynak., VIII., 287-297. Taf. IX.

Arch. f. Gynak., XL, 1 10-144. Tafn. I.III. Cont. XL, 443500. Tafn., XII.XV. Cont. XII., 169-210. Tafn. III.V.

Zeitschr. f. Geburtsk. u. Gynak., VII., 84 137J. Anat. and Physiol., XXI.,n.s.I.,Pt.IIL,

PP- 378-384A. f. m. A., XXV., 191-235. Tafn. X.XI.

Journ. Phys., L, 4671

Ann.Gynec.XXVIL, Mars, pp. 178-201.

British Gynec. Journ., VII., pp. 285-292. Med. Press and Circular, London, n. s. XLIIL, pp. 259-261,




1875. 1876. 1877.




1887. 1886.

No. 3.]




(See also Allantois, Amnion, and Placenta.)

44. Beneden, E.

V. and Julin, Chas.

45. Bischoff.Th.

L. W.

46. Blacher, K.

47. •• "

48. Caldwell,

W. H.

49. Chabry,L.,et

Boulart, R.


Chapman, Henry C.

51. Dohrn, H.

52. Edwards,

A. Milne.

53. " "

54. Hennig, C.

55. Hart, B. D.,

and Carter, J. T.

56. Klaatsch,


57. Kyburg,


Recherches sur la formation des annexes fcetales chez les mammiferes (Lapin et Cheiiopteres).

Beitrage zur Lehre von den Eihiillen des menschlichen Foetus.

Ein Beitrag zum Bau der menschlichen Eihiillen.

Noch ein Beitrag zuni Baue der menschlichen Eihiillen.

On the arrangement of the embryonic membranes in marsupial animals.

Note sur un foetus de dauphin et ses membranes.

On a fcetal Kangaroo and its membranes.

Ein Beitrag zur mikroskopischen Anatomie der reifen menschlichen Eihiillen.

Recherches sur les envellopes fcetales du Tatou ci neuf bands.

Sur la disposition des envellopes foetales de I'Aye-Aye.

Ueber die Eihiillen einiger Sangethiere.

A contribution to the sectional Anatomy of advanced extrauterine Gestation.

Die Eihiillen von Phocsena communis, Cuv.

Beschreibung von Foeten und peripheren Eitheilen einer Vierlingsgeburt nebst Musterung der Angaben iiber die Geschlechtsverhaltnisse der einem Ei entstammenden Foeten.

Arch. Biol.jV., fasc. 3, pp. 369-434- PIXX.-XXIV.

Bonn, pp. 112. Tafn.


Arch, f. Gynak., X.,

459-469. Taf. XI. Arch. f. Gynak., XIV.,

121-126. Tafn. I. II. Q. J. M. S., XXIV.,

pp. 655-658. PI.

XLIII. Journ. Anat. et Phys.

Robin et Pouchet,

XIX., 572-575. PI.

XXXIV. Proc. Acad. Sci., Philadelphia, pp. 468 471. PI. XX. Ann. and Mag. N. H.,

Ser. 5, IX., 338 340. Monatschr. f. Ge burtsk., XXVI.,

1 14-127. Tafn. I. IV. C. R. Paris,

LXXXVIII., 406 408. C. R. Paris, XCIX.,

265-267. Sitzb. nat.-forsch Ges.

Leipzig, I., 9-12. Edinburgh Med.

Journ., XXXIII.,

332-343. Pls.I.-III. Arch. f. mikrosk.

Anat., XXVI., 1-50.

Taf. I. -II. 4to (Diss.), pp. 27.

I Taf. Halle.

1834. 1876. 1879.







[Vol. II.

58. Lowe, L.

59. Meola, F.

60. Osborn, H.


61. " " 61a. " "

62. Owen, R.

63. " "

64. Ravn, Ed ward.

65. Ruge, Carl.

66. Turner, Wm.



69. Walker, A.

In Sachen der Eihaute jlingster

menschlichen Eier. Sulla struttura degli involucri del

feto umano.

Observations upon the fcetal membranes of the Opossum marsupials.

Upon the foetal membranes of the marsupials.

The fcEtal "membranes of the marsupials: the yolk-sac placenta in Didelphys.

Description of the membranes of the uterine fcetus of the Kangaroo.

Description of the foetal membranes and placenta of the Elephant (Elephas Indicus, Cur.), with remarks on the value of placentary characters in the classification of the mammalia.

Ueber die mesodermfrie Stelle in der Keimscheibe des Hiihnerembr>'os.

Die Eihiillen in der Geburt befindlichen Uterus. Bemerkungen ueber den Ort und die Art der Ernahrung des Kindes in demselben.

On the gravid uterus and on the arrangement of the foetal membranes in the Cetacea.

Note on the foetal membranes of the Reindeer (Rangifer tarandus).

On the foetal membranes of the eland (Oreas canna).

Der Ban der Eihaute bei Graviditatis abdominalis.

Arch. fGynak., XIV.,

190-196. Riv. Internaz di Med.

e Chir., I., 505-502;

582-609. Tav. I.

Abstr. Cbl. Gynak.,

pp. 584. Q. J. M. S. XXIII.,

473-484 PI. XXXIII. Zool. Anz., VI., 418 419. Journal of Morphol.,

I., 373-382. PI.

XVII. Mag. Nat. Hist., I.,


Philos. Trans., CXLVIL, 347-353

His u. Braune, Arch., pp. 412-421. Taf. XXI.

In Schroder, 129a, pp. 113-151.

Trans. Roy. Soc, Edinburgh, XXVI., 467-504. PI. XVII.XVIII.

Journ. Anat. and

Phys., XII., 601


Journ. Anat. and


Phys., XIV., 241


Virchow's Arch.,


CXVII., 72-99. Taf.



1883. 1887.




No. 3-]




(See also Fcetal Membranes, Amnion, Allantois, and Decidua.)

70. Baer, C. E. von.

71. Balfour, F. M.

72. Bassett, J.

73. Beauregard

et Boulart.

74. Cainven berghe, Charles van.

75. Chapman,

H. C.

76. " "

77. Chatellier, H.

77a. Creighton, Ch.

77b. " "

78. Dastre.

79. Delore.

80. Duges, A.

81. Duncan, J. M.

Untersuchungen iiber die Gefassverbindungen zwischen Mutter und FVucht in den Saiigethieren. Ein Gliickvvunsch zur Jubelfeier. Samuel Thomas von Sommerings.

On the evolution of the placenta and on the possibility of employing the characters of the placenta in the classification of the mammalia.

Utersuchungen der Placenta bei Drillinger.

Note sur la placentation des Ruminants (cites some earlier literature).

Sur L'Anatomie Physiologique and la Pathologic du Placenta.

Placenta of Macacus cynomolgus.

The placenta and generative apparatus of the Elephant.

Etude sur un point de I'anatomie du placenta chez les femelles du rat blanc.

On the formation of the placenta in the Guinea Pig.

Further observations on the formation of the placenta in the Guinea Pig.

Du placenta foetal des Pachydermes.

Etude de la circulation maternelle dans le placenta.

Lettre relative \ la placentation du Dasypus novemcinctus.

Note of a proof of the free intercommunication near the chorionic surface, between different parts of the system of maternal

Leipzig, pp. 30, I., col. pi. folio.

Proc. Zool. Soc, London, 210-212.

Obstetr. Trans., XXIIL, 129-130.

Robin's Journ., XXI., 93-99. Pl.V.

Gand (Diss.), pp. 203.

Proc. Acad. Nat. Sci., Philadelphia, 146147.

J. Acad. Nat. Sci., Philadelphia, 4to, Ser. II.-VIIL, 413422. PI. XLVIII.-L.

Nouv. Arch, d' obstetr. et de gynec, Paris, I., pp. 488-491.

Journ. Anat. and Physiol., XIL, 534-590. PI. XIX. -XX.

Journ. Anat. and Physiol., XIIL, 173-182. PI. XVI.

Ann. Gynec, V., 6667.

Ann. Gynec. Juin. Pis. I.-II.

Ann. Sci. Nat. Zool., Ser. IX., VI., Art. 3, pp. 2.

Edinburgh Obstr. Soc, III., 116-121.






1876. 1874. 1880.




[Vol. II.

82. Edwards,

A. Milne.

83. Ercolani, G.


84. " "

85. " "

86. " "

87. " "

88. " "

89. " "

90. " "

91. " "

92. " "

93. Fleischmann,


94. Garrod, A.

H., and Turner, Wm.

cells or blood caverns of the placenta, in the same and in different cotyledons. Observations sur la conformation du placenta chez le Tamandua.

Sul processo formative della porzione glandulare o materna della placenta.

Delle malattie della placenta.

De la portion maternelle du placenta chez les mammiferes.

Sulla parti che hanno le glandule otricolari dell' utero nella formazione della porzione materna della placenta.

Della struttura anatomica della caduca uterina nei casi di gravidanza extrauterina nella donna.

Della placenta nei mostri per inclusione e nei casi di gravidanza extrauterina nella donna e in alcuni animale.

Sulla unitil del tipo anatomico della placenta nei mammiferi e nella umana specie e suU' unit^ fisiologica della nutrizione dei fete in tutti vertebrate.

Nuove Ricerche sulla placenta nei pesci cartilaginosi e nei mammiferi e delle sue applicazioni alia tassonomia zoologica e all'antropogenia.

Nuove Ricerche di anatomia normale e patologica sulla placenta dei mammiferi e della donna (Tre Lettere al chiarissimo signore Prof. Alberto Kolliker).

Resume fran9ais par I'auteur de 91.

Ueber die erste Anlage der Placenta bei den Raubtieren.

The gravid uterus and placenta of Hyomoschus aquaticus.

Ann. Sci. Nat., Ser.

v., XV., Art. 1 6, pp.

4. PI. V. Mem. Accad. Sci. 1st.,

Bolonga, 4to, Ser.

II., IX. (fasc. 3),

363-432. Tav. I. VI. Mem. Accad. Sci.

1st., Bolonga, 4to,

Ser II., X., 491 554. Tav. VII. Journ. Zool., I., 472 480. PL XXIV. Mem. Accad. Sci. 1st.,

Bolonga, 4to, Ser,

3, III., 264-312.

Tav. I.-IV. Mem. Accad. Sci. 1st.,

Bolonga, 410, Ser.

3. IV., 397-414Tav. I. Mem. Accad. Sci. 1st., Bolonga, 4to, Ser. 3. v., 527-541 Mem. Accad. Sci. 1st., Bolonga, 4to, Ser. 3, VII., 271-346. Tav. I.-V.

Mem. Accad. Sci. 1st., Bolonga, Ser. 3, X., 601-982. Tav. I.XI.

Mem. Accad. Sci. 1st., Bolonga, Ser. 4, IV., 707-782. Tav. I.III.

Arch. Ital. Biol., IV., 179-192.

Sitzb. d. Phys.-Med. Soc, Erlangen, pp. 3.

Proc. Zool. Soc. London, 682-686. Pis. XLIV.

1872. 1870.


1872. 1873 1874.





1883. 1886. 1878.

No. 3-]



95. Gervais, P.

96. Godet, R.

97. Goodsir, J.

98. Hegar.

99. Hening, C.

100. Hicks, J.


101. Harting, P.

102. HoU, M.

103. Homberger

104. Hyrtl, J.

105. Jassinsky,


106. Joly, N.

106a. Jungbluth.

107. Kastschenko, N.

108. Kolliker, A.

Addition au memoire de M. Turner, (135).

Recherches sur la structure intime de placenta du lapin.

Structure of the human placenta.

Die Placenta materna am Ende der Schwangerschaft.

Studien ueber den Bau der menschlichen Placenta und Uber ihr Erkranken.

Some remarks on the anatomy of the human placenta.

On the ovum and placenta of the Dugong (Abstract by Prof. Turner).

Ueber die Blutgefasse der menschlichen Nachgeburt.

Die nachtragliche Diagnose der Lagerung des Eies im Uterus aus den ausgestossenen Nachgeburtstheilen.

Die Blutgefasse der menschlichen Nachgeburt in normalen und abnormen Verhaltnissen.

Zur Lehre ueber die Structur der Placenta.

Etudes sur la placenta de I'Ai (Bradypus tridactylus, Linn.). Place que cet animale occupe dans la serie des Mammiferes.

Beitrage zur Lehre von Fruchtwasser und seiner iibermassigen Vermehrung.

Das menschliche Chorion epithel und dessen Rolle bei der Placenta.

Ueber die Placenta der Gattung Tragulus.

Journ. Zool. I., 323324. PI. XVII.

Neuveville, pp. 48. Tav. I. -I I. (Berner Inaug. Diss.).

Goodsir's Anat. and Pathol. Observations, Edinburgh, pp.

5^54Monatschr. f. Ge burtsk., XXIX., I II. Leipzig, Englemann,

pp. 39. Tafn. I. VIII. Journ. Anat. and

Phys., VII. (Ser.

II.), v., 405-410. Journ. Anat. and

Phys., XIII, 116 117. Sitzb. Akad. Wiss.,


3, April, pp. 1-41.

Tafn. I.-II. Freund Gynak. Klin ik, L, pp. 663-676.

Wien, 4to, pp. VIII., 152. Tafn. I.-XX.

Virchow's Arch., XL., 341-352. Taf. III.

C. R. Paris, LXXXIL, 283-287.

Inaug. Diss., Bonn.

His. Arch., 451-480. Taf. XXI. (also in Russian, 1884, Charkow, pp. 2>Z- Pis. I.-II.)

Verh.WUrzburg Phys.Med. Ges., N. F., X., 74-83. Tafn. IV.-V.

1872. 1877.


1867. 1872. 1872. 1879.

1870. 1867.

1869. 1885.




[Vol. II.

109. Langbans,


110. " "

111. " « 

112. Laulanie.

113. Marcy, H. •O.

114. Masquelin,

H., et Swaen, A.

115. Mauthner,


116. Meyer, L.

117. Nitabuch,


118. Orth, J.

119. Pyle, J. P.

120. Reidjohn.

121. Reitz.


Zur Kenntniss der menschlichen Placenta.

Untersuchungen ueber die menschliche Placenta.

Ueber die Zellschicht des menschlichen Chorion.

Sur la nature de la neoformation placentaire et I'unite de composition du placenta.

The placental development in mammals. A unity of anatomical and physiological modality in all vertebrates.

Premieres phases du development du placenta maternal chez le lapin.

Ueber den miitterlichen Kreislauf in der Kaninchenplacenta mit Riicksicht auf die in der Menschenplacenta bis jetzt vorgefundenen anatomischen Verhaltnisse.

Ueber die Blutmenge der Placenta.

Beitrage zur Kenntniss der

menschlichen Placenta. Das Wachsthum der Placenta

fetalis und Boll's Prinzip des

Wachsthums. An experimental research on the

utero-placental circulation.

On the anatomical relations of the blood-vessels of the mother to those of the foetus in the human species (also Appendix to Dr. J. Reid's paper on the anatomical relations of the blood of the mother to those of the fcetus) .

Beitrage zur Kenntniss des Baues der Placenta des Weibes.

Artikel Placenta.

Arch. f. Gynak, I.,

317-334. Taf. V. His. Arch., 188-267.

Tafn. VII.-VIII. Beitrage z. Anat. u.

Embryol. als Fest gabe. F. G. J.

Henle, 4to, 69-79.

Taf. IX. a. C. R. Paris, 4to, C,

651-653 Ann. Anat. and Surg., pp.

Arch. Biol., I., 2544.

Wien Sitzb., LXVIL, Abth. 3. 1 18-124. Taf. I.

Centralbl. f. Gynaec,

II., 220-222. Bern. (Diss.), pp. 39.

PI. I. Zeitschr. f. Geb. h. u.

Gyn.,11.,9-23. Taf.

III. Philadelphia Med.

Times, No. 433,

XIV., June 28, pp.

711-715Edinburgh Med. and Surg. Journ., LV., 1-12. PI. I., 135136.

Sitzb. K. Akad. Wiss.

Wien, LVIL, 1009 1012. Taf. I. Strieker's Handbuch

der Gewebelehre,

II., 1 18 VI 186.

1870. 1877. 1882.






xNo. 3.]



123. RibemontDessaignes, A.

124. Ritgen, F.


125. Robin, Ch.

126. Rolleston.

127. Romiti, Guglielmo.

128. Ryder, J. A.


129. Schatz, F.

129a. Schroder, Carl.

130. Schroeder van der Kolk, J. L. C.

131. Schultze,

B. S.

131a. Seiler.

132. Sym, A. C.

Des placentas multiples dans les grossesses simples.

Beitrage zur Aufhellung der Verbindung der menschlichen Frucht mit dem Fruchthiilter und der Ernahrung derselben.

Recherches sur les modifications graduelles des villosities du chorion et du placenta.

On the placental structures of the Tenrec (Centeles ecaudatus) and of those of certain other mammalia, with remarks on the value of the placental system of classification.

Placenta (Anatomia descriptiva e struttura).

The placentation of the two-toed Anteater (Cycloturusdidactylus).

A theory of the origin of placental types, and on certain vestigiary structures in the placentae of the mouse, rat, and field-mouse.

Die Gefassverbindungen der Placentakreislaufe eineiger

Zwillinge, ihre Entwicklung und ihre Folgen. Fortsezung.


Der schwangere und kreissende Uterus.

Waarnenungen over het maaksel van de menschelijke Placenta en over haren Bloeds-omloop.

Die Placentarrespiration des


Die Gebarmutter, etc.

On a case of vesicular placenta from a premature birth at the seventh month, the child being born alive.

Ann. Gynec, XXVII., Janvier, pp. 12-52.

Leipzig u. Stuttgart, pp. 78. Tafn. I.-III., fol.

Mem. Soc. Biol, Ser. 2, I., 63-75.

Trans. Zool. Soc, 4to, v., 285-316. PI. I.

Encicl. Med. Ital.,

Ser. 2, III., Part 3. Proc. Acad. N. 5,

Philadelphia, 115 120. American Naturalist,

XXL, 780-784.

Arch. f. Gynac,


XXIV., 337-399.

Taf. I.-V.

Arch. f. Gynac,


XXVIL, 1-72. Tafn.


Arch. f. Gynac,


XXIX., 41 9-443

8vo, Bonn.


Verhand. eerste Klas


se K. nederl. Institut., Amsterdam. Derde Reeks, 4to, 4 deel., 69-180. Tafn. I.-VI. Jena. Zeitschr. f. Med. n. Naturu., IV., 54I552• Fo.,pp. 38. XILPls. Edinburgh Medical Journal, XXXIIL, 102-109.

1887. 1835 1854. 1866.

1887. 1887.



1832. 1887.


133. Tafani, A.

134. " "

135. Turner,


136. " '

137. " "

138. " "

139. " "

140. " "

141. " "

142. 143.

144. 145.

146. 147.


Sulle condizione uteroplacentari della vita fetale.

La circulation dans le Placenta de quelques mammiferes (Translated from Lo Sperimentale, Agosta, 1885).

De la placentation des Cetaces comparee a celle des autres mammiferes.

Observations on the structure of the human placenta.

On the placentation of the Sloths (Choleopus Hoffmani).

On the placentation of Seals (Halichcerus gryphus).

Lectures on the Comparative Anatomy of the Placenta. First series.

Note on the placentation of Hyrax.

On the placentation of the Lemurs (see No. 144).

On the placentation of the Cape Ant-eater (Orcetropuscapensis).

Some general observations on the placenta, with especial reference to the theory of evolution.

On the placentation of the Lemurs (see 141).

A further contribution to the placentation of Cetacea (Monodon monoceros) .

On the placentation of the Apes, vfith a comparison of the structure of their placenta with that of the human female.

On the cotyledonary and diffused placenta of the Mexican deer (Cervus Mexicanus).

[Vol. IL

Arch. Scuola d' Anat. pathol. diretto dal Prof. Giorgi Pellizari, Firenze, IV., pp. 53-216. Tav.L-VIII.

Arch. Ital. Biol., VIIL, 49-57

Journ. Zool., L, 304323. PI. XVL

Journ. Anat. Physiol., Ser. 2, VI., 120133. PI. V.

Trans. R. Soc, Edinburgh, XXVII., 71104. Pis. III.-VI.

Trans. Roy. Soc, Edinburgh, XXVII., 275-304. Pis. XVIII.-XXI.

Edinburgh (Black), pp. 124. PI. I.-III.

Proc. Roy. Soc,

XXIV., 151-155. Philos. Trans. Roy.

.Soc, London,

CLXVL, 569-587.

PI. XLIX.-LI. Proc. Roy. Soc,

XXIV., 409. Journ. Anat. and

Phys., X., 693-706. Journ. Anat. Physiol.,

XL, 33-53 Journ. Anat. Physiol.,

XII., 147-153 Proc. Roy. Soc, Edinburgh, IX., 103 IIO.

Phil. Trans., LXIX.,


Pis. XLVIIL XLIX. Journ. Anat. and

Phys., XIII., 195 200.


1S72. 18731873. 1875 1876.

1876. 1876.

1876. 1877.

1877. 1877.

1879. 1879.

No. 3.]



148. Turner,


149. Waldeyer,


150. Watson, M.

151. Weldon,

W. F. H.

152. Winkler,

r. N.

On the placenta of the hog-deer

(Cervus porcinus). Ueber den Placentarkreislauf des

Menschen. On the female organs and placen tation of the Racoon (Procyon

lotor). Abstract.

Note on the placentation of Tetra ceros quadricornis. Zur Kenntniss der menschlichen


Journ. Anat. and Phys., XIII., 94-98.

Sitzb. K. Akad. Wiss. Berlin, VI., 83-93.

Proc. Roy. Soc. London, XXXII,, 272298. Pis. III.-VI.

Proc. Roy. Soc, London, XXXI., 325326.

Proc. Zool. Soc. London, No. I, pp. 2-6.

Arch. f. Gynak., IV., 238-265. Taf. V.

1879. 1887.


153. Koster.

154. Neugebau er, L. A.

155. Renaut, J.

156. Ruge, C.

157. " "

158. Schott, J.

A. C.

159. Schultze.B.


160. ' "

161. Stutz, Gus tav.

162. Tail, Law son.


(See also Allantois.)

Ueber die feinere Structur der

menschlichen Nabelschnur. Morphologic der menschlichen

Nabelschnur. Du tissu muqueux et du cordon

ombilical. Ueber die Gebilde im Nabel strang. Untersuchungen iiber den Dotter gang und iiber die Capillaren

im Nabelstrang. Die Controverse ueber die Nerven

des Nabelstranges und seine

Gefasse. Die genetische Bedeutung der ve lamentosen Insertion des Nabelstranges. Ueber velamentale und placentale

Insertion der Nabelschnur. Der Nabelstrang und dessen Ab sterbe-process. Preliminary note on the anatomy

of the umbilical cord.


WUrzburg, pp. 32, 2,





Arch, de Physiol.,


VII., 219.

Z. Gebiirtshilfe u. Gyn


ak., I., 1-21.

Z. f. Geburts. u. Gyn


ak., I., 253-259. ,.

Frankfurt a M.


Jena Zeit., III., 198


205. Zweiter Arti

kel. III.. 344-358.

Arch. f. Gynak.,


XXX., 47-56.

Arch. f. Gynak., XIII.,



Proc. Roy. Soc, Lon


don, 417-440. Pis.


Proc. Roy. Soc, Lon


don, XXIIL, 498




[Vol. IL


(See also Decidua and Placenta.)

163. Bonnet, R.

164. Cadiat, O.

165. Coblenz, H.

166. Delaunay,


167. Dohrn, H.

168. " "

169. " "

170. " "

171. Elisher, J.

172. Farre, A.

173. Friedlander


174. Gasser.

175. Gottschalk.

176. Hofmeier,


177. Imbert,


Die Uterinmilch und ihre Bedeutung fur die Frucht.

Memoire sur I'uterus et les Trompes.

Zur Entwicklungsgeschichte der inneren weiblichen Sexual-organe beim Menschen im Zusammenhange mit pathologischen Vorgangen.

La Fecondite.

Ueber die MuUer'schen Gange und die Entwicklung des Uterus.

Zur Kenntniss der Mullerschen

Gange und ihrer Verschmelzung.

Ueber des Entwickelung des Hy

Ueber die Gartnerschen Kanale beim Weibe.

Beitrage zur feineren Anatomie der Muskelfasern des Uterus.

Article "Uterus and its appendages."

Ueber die Innenflache des Uterus post partum.

Ueber Entwicklung der Miiller'schen Gange.

Ein Uterus gravidus aus der fiinf ten Woche der Lebenden ent nommen. Das untere Uterinsegment in an atomischer und physiologischer

Beziehung. Le Col du Segment interieur de

I'uterus ^ la fin de la grossesse;

documents anatomiques.

Beitrage zur Biologic (von BischofFs Festgabe, Stuttgart), 221-263, "^it einer Tafel.

Robin's Journ. Anat., 409-431. Pis. XXVI.-XXIX.

Giebel. Z. Naturwiss., 3 Folge, VI., 313329, Taf. II.

Revue Scientif., 3 Oct.,

433-437. et 10 Oct.,

466-470. Monatschr. f. Ge burtsk, XXXIV.,

382-384. Marburg Ges., IX.

Schriften. Ges. Nat. Wiss. Marburg, X., Suppl., Heft I., pp. 8. Tafn. I.-X.

Arch. f.Gynak., XXI.,

328-345Arch. f. Gynak., IX.,

10-21. Taf. II. Todd's Cyclop. Anat.,

v., Suppl. Arch, of Gynak., IX.,

22-28. Sitzb. Ges. Beford

Naturw. Marburg.,

1-3Arch. Gynak., XXIX.,

Heft 3, pp. 488-510.

Taf. IX. In Schroder, 129a,

21-73 Paris, pp. 79.




1876. 1858. 1876. 1872.



No. 3.]



178. Johnstone.

179. Kustner,


180. Kundrat,

Hanns, u. Engelmann G.J.

181. Leopold, G.

182. Lowenthal.

183. Paschen,D.

184. Patenko.T.

185. Sinety, de.


187. Tourneux,

F., et Legay, Ch.

188. Underbill.

189. Waldeyer, W.

190. Williams, J.


On the Menstruation Organ. Discussion.

Das untere Uterinsegment und

die Decidua cervicalis. Untersuchungen ueber die Ute ruschleimhaut. (Translation, see

Engelmann, 28).

Die Lymphgefasse des normalen nicht schwangeren Uterus (III. Die Lymphlamen der Schleimhaut, 30-48).

Eine neue Deutung des Menstrual-processes.

Beschreibung eines graviden Uterus aus dem funften Monat der Schwangerschaft.

Uber die Nervenendigungen in der Uteruschleimhaut des Menschen (Vorlanfige Mittheilung).

Sur I'histologie normale de la cavite uterine quelques heures apres I'accouchment.

Etude histologique sur la cavite uterine apres la parturition.

Memoire sur le developpement de I'uterus et du vagin envisage principalment chez le fcetus huniain.

Note on the uterine mucous membrane of a woman who died immediately after menstruation.

Medianschmitt einer Hochschwangeren bei Steisslage des Fcetus nebst Bemerkungen iiber die Lage und Formverhaltnisse des Uterus gravidus nach Langsund Querschmitten.

On the structure of the mucous membrane of the uterus and its periodical changes.

The mucous membrane of the body of the uterus.

British Gynec. J., Pt.

VII., S., pp. 292 301; 301-307. 8vo, 2 Tafln.

Strieker's Med. Jahrb., I35-I-77- Taf. I.

Arch. f. Gyn., VI., i54. Tafn. I.-III.

Arch. Gyn., XXIV., 168-261. Abstr.Cbl. Gyn., 306-312.

Marburg (Diss.), pp. 19.

Centralb. Gynak., IV., 442-444.

Ann. Gynecol., VL, 217-220,

Arch. Physiol. Norm, et Path., Ser. 2, III.,

342-352Robin's Journ. Anat., 330-386. Pis. XX.XXV.

Edinburgh Med. Journ., XXL, 132 133Bonn. (Cohens Verlag) Atlas, 5 Tafln.

Obstet. Journ. Gt.

Brit, and Ireland,

IL, 681-696. PI. I.;

753-767. PI. IL III. Obstet. Journ. Gt.

Brit, and Ireland,

III., 496-504.


1882. 1873 1874.









[Vol. II.

192. Wyder, Aloys Theodor.

Beitrage zur nomalen und pathologischen Histologic der menschlichen Uteruschleimhaut (I. Die mucosa uteri der Kinder. II. Menstruation Endometritis. III. Dyemenorrhea membranosa).

Arch. f. Gyn., XIII., 1-55. Taf. I.


193. Ahlfeld, F.

194. " "

195. Allen, W.

196. Kleimwach ter, L.

197. Schenk, S.

198. Schultze, B. S.


Ueber die Persistenz des Dotterstrangs in der Nabelschnur.

Ueber die Persistenz der Dottergefasse nebst Bemerkungen iiber die Anatomic des Dottcrstranges.

Omphalo-mesenteric remains in mammals.

Ein Beitrag zur Anatomic des Ductus omphalo-mcsentericus.

Der Dottcrstrang der Plagiostomen.

Das Nabelblaschen ein constantes Gebilde in der Nachgeburt des ausgetragenen Kindes.

Arch. f. Gynak., IX., 325. Only four lines.

Arch. f. Gynak., XL, 184-197.

J. Anat. and Phys.,

XVII., 59-61. Arch. f. Gynak., X.,

238-247. Taf. I. VII. Sitzb. K. K. Akad.

Wien., LXIX., 301 308. Taf. I. Leipzig, pp.

1876. 1877.

1883. 1876.






Nearly all the figures were drawn by Mr. E. Stanley Abbot under my supervision. The outlines were drawn with the camera lucida, and the details added free-hand. The drawings are all accurate representations of the preparations, and though of course not photographically exact, are not diagrammatic, except in the case of a few figures expressly specified below. I owe much to Mr. Abbot's patient skill.

Reference Letters.

all, allantois.

7nes, mesoderm.

cm, circular muscles., monster cells.

conn, connective tissue.

fnsth, mesothelium.

ecto, foetal ectoderm.

nine, mucosa.

emb, embryo.

fimsc, muscularis.

en, entoderm., ob-placenta.

endo, endothelium.

o.z, outer zone of placenta

ep, epithelium.

P, periplacenta.

f, placental fissure.

per.v, perivascular cells.

f.v, foetal blood-vessel., sub-placenta.

gl, gland ; glandular layer.

Sgl.z, subglandular zone.

h.ep, hyaline epithelium.

V, blood-vessel.

/, leucocytes.

vac, vacuole.

Im, longitudinal muscles.




Fig. I. Placenta of rabbit at eight days, with dilated glands,^/, and superjacent foetal ectoderm, ecto (X 125 diams.).

Fig. 2. Rabbit's uterus at nine days, transverse section of a swelling (X 7 diams.).

Fig. 3. Portion of the placenta of Fig. 2 (X 445 diams.), to show the connective tissue, conn, the perivascular cells, per.v, and the thickened endothelium, endo, of the blood capillaries.

Fig. 4. Portion of the periplacenta of Fig. 2 (X 175 diams.), to show the degeneration of the epithelium, h.ep.

Journal ofMorplioloc/ij. \'oill.


/ S 3 *


% \



^ a, :\'\

III %& 'i^



>. '














«")9/m jier.y Endo

hep ^.'^ otj'/'-'


.: A

°# <

/ 2 3 4'

"^^ %





/tf^ .-f/ii: vW!Trerxy^Lt:rT_.




Fig. 5. Portion of the ob-placenta of Fig. 2 ( X 175 diams.), to show the degenerated epithelium, h.ep, and the saucer-shaped glands, gl, gl'.

Fig. 6. Rabbit's uterus of eleven days; portion of the ob-placenta to show the degenerated epithelium, h.ep, and the regenerated glands,^/ ( X 175 diams.).

Fig. 7. Portion of the placenta of Fig. 2 ( X 175 diams.), to show the degeneration of the uterine tissue and the relations of the foetal ectoderm to the placental surface.

Fig. 10. Rabbit's uterus of thirteen days; portion of the ob-placenta (X 175 diams.), to show the regenerated glands.

Fig. II. Portions of the epithelium of the periplacenta at thirteen days. A, vertical section (X 175 diams.). B, surface view (X 175 diams.). C, single cell ( X 445 diams.).

Jounial of Moqiliolotfij ■ Vol.n







I, ,-,,





Fig. 8. Portion of a vertical section of the placenta at eleven days of a rabbit, to show the relations of the mesothelium, msth, to the top, and of the ectoderm, ecto, to the side of the placenta ( X 175 diams.).

Fig. 9. Complete transverse section of a rabbit's uterus at thirteen days, with the embryo, enib, in place ( X 7 diams.) ; the details are only approximately accurate ; X, mass of perivascular decidual cells, developed in the region of the ob-placenta.

Fig. 12. Rabbit's uterus at fifteen days; portion of a section through the placenta (X 90 diams.), to show the degenerated glands,^/, and the mesoderm, mes, and mesothelium, msth, covering the surface of the placenta ; the blood-vessels are drawn dark.

Fig. 13. Portion of upper part of a rabbit's placenta at fifteen days (X 34° diams.), to show the histological structure of the glandular layer of the placenta.

Fig. 14. Multinucleate decidual cells from the subglandular zone of a rabbit's placenta at fifteen days ( X 540 diams.).

Fig. 15. Uninucleate perivascular decidual cells from the outer zone of a rabbit's placenta at fifteen days ( X 540 diams.).

Fig. 16. Endothehum from the blood-vessels of the periplacenta of a rabbit at fifteen days ( X 240 diams.). A, surface view; B, C, in section.

Fig. 17. Ob-placenta of a rabbit at fifteen days (X 125 diams.), to show the monster cells, a, b, c, and the uterine epithelium, ep.

Fig. 18. Nucleus of a monster cell from the ob-placenta of a rabbit at fifteen days ( X 445 diams.).

, /ouniai n/Mor/t/iohf/!/. I of //.







■'^J ■i^





/ \^'










462 MINOT.


Diagram to show the relations of the embryo and uterus in the rabbit from the eleventh to the thirteenth day of gestation.

.JouTjtid of Morjilwlo(jij. \bl.l!.


a ectoderm , lost jmrf miisc. mnsciilaris

all aUcmtois lusth. mesfllheliiiw

l> ectodcnn at side of jilaecnta instij mcsenterij

ccto ectoderin ofijolk-sack. oh.jd oh placenta


entoderm of yolk- sack


oilier zone ofjxlaccnta


fissure betjveen lobes.






sub - pleicenta




sub-glemdnlar zone ;


monster cells


vena terminalis.





I. Adult Form.


General Description 465

Course ot the Canals 470

1. Infra-orbital Canal . . . .471

2. Supra-orbital Canal .... 472

3. Operculo-mandibular Canal . . 473

4. Supratemporal Cross-commis sure 473

Topography of the Peripheral Canalsystems 474

1 . Infra-orbital Canal .... 474

2. Lateral Line of the Body . . 483

3. Supratemporal Cross-commis sure 483

4. Supra-orbital Line .... 483

5. Operculo-mandibular Line . . 485

6. Summary 486

Surface Pores 487

1. Infra-orbital Line 487

2. Lateral Line of the Body . . 490

3. Supra-orbital Line . .

4. Operculo-mandibular Line

PAGB • 463

PACK 491


II. Post-Larval FoRMb.

Primary Pores 494

Number and Position of Canal

Organs 499

Sense-organs of the Spiracular

Canal 501

Lines of Pit Organs 502

a. Head Lines 505

b. Body Lines 507

Surface Sense-organs .... 509

Innervation .510

Review of Nerves and Organs . 520

III. Larval Forms.

1. Formation of the Canals . . .523

2. Origin of the Canal Organs . .531

General Summary .... 535

The general course and position of the lateral canals in fishes have long been known, and short descriptions and diagrams, showing what is commonly accepted as the typical arrangement of the main canals, are given in nearly all modern text-books of vertebrate anatomy. More complete descriptions are given in numerous special works, most of which deal more particularly with other subjects, such as the structure of the skull, the

464 ALUS. [Vol. II.

development and distribution of the cranial nerves, or the histoiogy and function of the sense-organs.

With a few exceptions, all these descriptions, so far as they relate to the cranial part of the lateral system, are of a general character only, giving little more than the course of the main canals. The development of the canals, the number and position of the organs, and their innervation, receive but scant attention. This has doubtless been largely due, as Merkel suggests, to the difficulties attending this part of the investigation before the introduction and perfection of modern methods of research ; but this cannot have been the only reason, for most of the work could easily have been done by any of the earlier writers. The purely descriptive part of the subject seems simply to have been neglected in the greater interest attaching to the histological and physiological sides, so that it is only within the last five or six years that the constant relations of the cranial canals to the dermal bones of the head, and their importance in determining these bones in doubtful cases, have been recognized. Both Sagemehl and Van Wijhe have called special attention to this, and Sagemehl further says (No. 12, p. 182, note) that the lateral canals seem to deserve a more careful study than has hitherto been given them.

In Amia calva the cranial canals have been oftener and more fully described than in any other form. Franque (No. 6), in 1847, in a dorsal view of the dermal bones of the head, shows some of the surface openings of these canals, but in the text he does not refer to them, these four words only being found, " Linea lateralis fere recta" (No. 14, p. 368). Bridge (No. 4, p. 620), thirty years afterward, in describing the skull, gave the course and position of the main canals, the connections they form with one another, and the bones they traverse, and the arrangement given by him agrees closely with that more fully detailed in the present paper. His work was mainly confirmed in 1882 by Van Wijhe (No. 17, p. 288), and in 1883 and 1884 by Sagemehl (No. 12, p. 183, and No. 13, p. 36) ; still later by Shufeldt (No. 15), in connection with his translation of Sagemehl's paper ; and finally Wright (No. 18) has called attention to the sensory tissue lying in the upper end of the spiracular canal, and belonging, by its innervation, to the general lateral canal system, though classed by him as hypodermal and exceptional.


Sagemehl's description is much the best and most complete of all ; but he does not give the smaller branches of the system, simply saying of them that the main canals have numerous branches, arranged in many longitudinal rows and leading to minute openings on the outer surface of the head. These small branches are similarly disposed of in his description of the cranial canals in the Characinidae, where he says they are too variable and irregular to have any morphological significance whatever.

A closer examination of Amia has shown that these small branches have as constant relations to the dermal bones as the main canals have, and hence may have as great significance. They also have constant relations to the sense-organs found in the canals, agreeing in this with the conditions described by Bodenstein in Cottus gobio (No. 3) and by Emory in Fierasfer (No. 5, p. 39).

The drawings for this paper have been made by Mr. Nomura, to whose patience and' skill the plates bear witness.

I. ADULT FORM. General Description.

If an adult specimen of Amia calva be examined, many hundreds of minute pores will be found, scattered to all appearance most irregularly over a large portion of the head (Figs. 20,