Talk:Journal of Morphology 29 (1917)

From Embryology

Skip to main content


Search metadata

Search text contents

Search TV news captions

Search archived web sites

Advanced Search

PEOPLE Full text of "Journal of morphology" See other formats

I. )iS^f^ :-'-m









Founded by C. O. Whitman



University of Illinois Urbana, 111.

with the collaboration op

Gary N. Calkins Edwin G. Conklin C. E. McClung

Columbia University Princeton University University of Pennsylvania

W. M. Wheeler William Patten

Busaey Institution, Harvard University Dartmouth College

VOLUME 29 1917



U7 ^S



By the Williams & Wilkins Company

Baltimore, Md., U. S. A.


No. 1. JUNE

B. H. Buxton. Notes on the anatomy of arachnids. Eight diagrams and

three plates 1

N. E. McIndoo. The olfactory organs of Lepidoptera. Ten figilres 33

Waro Nakahara. On the physiology of the nucleoli as seen in the silk-gland

cells of certain insects. Nine figures (two plates) 55

Jesse LeRoy Conel. The urogenital system of Myxinoids. Eighty-five figures

(twelve plates) 75

Jay Arthur Myers. Studies on the syrinx of Gallus domesticus. Eighteen

figures 165

S. Saguchi. Studies on ciliated cells. One text figure and four plates 217


Robert J. Terry. The primordial craniima of the cat. Thirty figures (twelve

plates) 281

Naohide Yatsu. Note on the structure of the maxillary gland of Cypridina

Hilgendorfii. Four figures 435

Benjamin Harrison Pratt and J. A. Long. The period of synapsis in the egg

of the white rat, Mus Norvegicus Albinus. Two text figures and one

plate 441

Harold Heath. The early development of a starfish, Pateria (Asterina)

Mineata. Five figures 461

D. H. Wenrich. Synapsis and chromosome organization in Chorthippus

(Stenobothrus) Curtipennis and Trimerotropis Suffusa (Orthoptera).

Three plates 471

Clarence E. McClung. The multiple chromosomes of Hesperotettix and

Mermiria (Orthoptera). Eight plates 519

Caroline M. Holt. Multiple complexes in the alimentary tract of culex

pipiens. Thirty-three figures (four plates) 607






This paper is a continuation and extension of one published in the Zoologische Jahrbiicher of July, 1913, since which time I have had an opportunity of adding to the number of species examined.

It was shown in the previous paper that there are two very distinct groups of the arachnids as regards the construction of the coxal glands. In Group I may be placed the scorpions, pedipalps and spiders, which have been more particularly studied. The Phalangids certainly, and the Pseudo scorpions most probably, also belong to this group, but as they have not been made a subject of special study, will not be described in detail. In Group II were placed the Solifugae, and to these can now be added the Palpigrades.

The component parts of the coxal glands of Group I consist of (1) the terminal sac, or saccule, lined with a very delicate cubical or flattened epithelium which has the property of excreting solid particles such as carmine when injected subcutaneously. Solid particles, probably urates, can also frequently be observed normally in the cells lining the walls of the saccule. From (2) the saccule, a short duct, or collecting tubule, leads into (3) the labjTinth, which consists of a single, usuall}^ coiled, tube, the walls of which are lined with excretory epithelium having the usual striated base indicative of excretory functions, but solid particles are never excreted by these cells. At the distal end of the labjTinth there is sometimes a vesicular swelling which can be regarded as (4) the bladder; and from the




bladder (5) , a short exit tubule lined with cells derived from the hypodermis leads to (6) the outlet^ on the external surface, either on segments III or V, or on both of these segments, just posterior to the appendage.

In Group II the same arrangement holds good except that between the collecting tubule and the labyrinth there is a large sac lined with secretory epithelium, and the outlet is on segment II.

It was pointed out in the pre\ious article that the coxal glands of Group I appear to be homologous with the large nephridia on segments VI and VII of Peripatus, whereas the coxal glands of Group II show homology with the salivary gland of Peripatus, the outlet of which is situated on segment II.


Several genera, both of scorpions and of Theraphosid spiders from the East Indies, have been recently examined, but there is nothing further to add to the description which has already been published of the coxal glands of certain American, European and African genera.

It may be recalled that the coxal gland of the scorpions lies m segments V and VI with an outlet at the base of the posterior surface of appendage V.

The coxal glands of the spiders show considerable differences. That of the Theraphosid spiders has two saccules and two outlets, one of each on segment III and the other on segment V with a large and complicated labyrinth apparently common to both saccules and outlets. The Araneae verae have lost the coxal gland of segment V, retaining only that of segment III; and, with the exception of the six-ej^ed spiders, the Sicariids and Dysderids (to which can now be added the Oonopids), the coxal glands of the higher spiders show various stages of degeneration, or perhaps simplification.

^ There is always an outlet on the external surface, both in immature and adult specimens. The arrangement of special muscle fibers around the exit tubule indicates that it can be closed or opened at will.



Pocock has divided the Pedipalps into two distinct orders, the Uropygi and the Amblypygi. The former group consists of the Thelyphonides and Tartarides, and the latter of the PhrjTiides. This separation seems to be justified on examination of the coxal glands, which differ very widely in these two groups.


The description of the coxal glands of the Phrynides given in the pre\ious article was based upon two species of Tarantula from Central and South America. Since then I have had an opportunity of examining several species of Phrynicus and of the Charontini family from India and the Malay Peninsula.

Phrynini PhrjTiicus ceylonicus, Ceylon.

Phrynicus nigromanus (Gravely), India. Charontini Charon grayi (Simon), Manila.^

Stygophi-jTius, caves in Malay Peninsula.

Sarax, caves in Malay Peninsula.

Phrynicho-Sarax singapurae (Gravely), Langkawi Island, Kedah.

Charinides bengalensis (Gravely), Bengal. The coxal glands of Phrynicus are precisely similar to those of Tarantula as previously described; i.e., there is a saccule on segment III from which a very extensive coiled labyrinth tube leads backwards to segment VI, where the tube loops forward and runs anteriorly as a long straight tubule to the outlet on segment III, just below the saccule and posterior to appendage III (diagram 1, fig, 1), The labyrinth is peculiar in that the central portion is lined with a basophil epithelium which appears to be secretory, and it was suggested that the secretion might be used for salivary purposes as there is no distinct sali ^ Specimens collected and fixed for me by Mr. F. H. Gravely of the Indian Museum, Calcutta.

^ Specimens given by M. Eugene Simon of Paris. They had been kept in alcohol for twenty years, but were in fairly good condition.


vary gland in Tarantula such as is found in the scorpions and the spiders.

The Charontini, like Tarantula and Phrynicus, have a large coxal gland with saccule and outlet on segment III, the labyrinth of which also possesses special secretory epithelium in its middle portion; but the labyrinth of this large coxal gland does not extend quite so far back as in Tarantula and Phrynicus.

Diagram 1 Coxal glands of the Amblypygi. Fig. 1 Tarantulini and Phrynini. Saccule on segment III from which the labyrinth extends posteriorly to segment VI, from which point it loops forward and runs as a single tubule to the outlet behind appendage III.

Fig. 2 Charontini. In addition to the large coxal gland on segment III there is a small one, complete in itself, on segment V. S, saccule; CL, labyrinth of coxal gland; ET, exit tubule.

On segment V there is a second, very much smaller gland, complete in itself, with saccule, lab5a-inth and outlet behind appendage V, but the labyrinth possesses no secretory epithelium (diagram 1, fig. 1).

The presence of this small, apparently disappearing, coxal gland indicates that the Charontini are more primitive than Phrynicus, before which they are probably giving way. The


larger genera of the Charontini: Stygophrynus, Charon and Sarax, are found only m caves, while those found in the open, such as Charinides bengalensis and Phrynicho-sarax, some of which I found on Langkawi Island, are very small, insignificantlooking animals. Langkawi Island lies off the coast of Kedah in the Malay Peninsula, and seems to possess a rather primitive arachnid fauna, as I shall also have occasion to remark in a note on the distribution of the Uropygi. Phrynicho-Sarax is quite abundant in the jungle on Langkawi Island, but no Phrynicus at all were found there.


Thelyphonides. The carapace of the older Thelyphonides is very hard and difficult to section, but these animals live very well n captivity, and among a number of specimens, kept in small cages and supplied with grasshoppers, one will moult now and then, immediately after which the chitin is thin and soft, and even the mature animal can be readily sectioned. At this stage, also, there is much turgescence of the tissues; the organs are all actively functioning and therefore in a favorable condition for observation.

Species of Thelyphonides examined:

Thelyphonus linganus Malay Peninsula

Thelyphonus sepiaris Ceylon (Plains)

Labochirus crassimanus Ceylon (Hills)

Hypoctonus kraepelini Langkawi Island, Kedah

The coxal glands of all of the above representatives of this family are precisely alike and resemble those of Tarantula in that the coils of the labyrinth extend well back into segment VI, and are not concentrated as in the scorpions; but nevertheless, the gland differs greatly from that of Tarantula, for there are two saccules — one on segment IV and- another on segment V. The saccules, moreover, are elongated and flattened, not roughly spherical like those of Tarantula and the other Amblypygi. In the case of the Amblypygi (and this applies also to the more or less spherical saccule of the scorpions and spiders) the surface


of the saccule is increased by blood capillaries which push in the wall of the saccule and break up its lumen into tortuous channels (figs. 1 and 2). The elongated and flattened saccules of Thelyphonus have sufficient surface of themselves and are not invaded by capillaries in this way. The saccule of Thelyphonus appears tubular in sections, and the tubules can be distinguished from those of the labyrinth only by the nature of the epithelium lining the walls. The cubical epithelium of the saccule has very delicate outlines and the cytoplasm often contains solid particles — ^probably urates — ^which are never found in the coarser striated epithelium of the labyrinth. On injecting carmine the particles are taken up by the cells of the saccules, but never by those of the labyrinth.

Diagram 2 Coxal gland of the Thelyphonides. Two saccules, one on segment IV, and another on segment V, with a labyrinth common to both saccules and an outlet behind appendage III. S, saccule; CL, labyrinth; ET, exit tubule.

From each of the two saccules of Thelyphonus a short but wide collecting duct leads into a tubule of the labyrinth directed posteriorly, so that the tubule from the saccule on segment V extends posteriorly into segment VI where it forms coils which ultimately pass forward again. Similarly the tubule from the saccule on segment IV extends backwards into segment V where it joins the main coils. The coils coalesce at some hitherto undetermined point, and the coiled labyrinth tubule passes forward to a point opposite the anterior saccule, where the coiling comes to an end and the labyrinth runs forward as a single straight tubule to its outlet just posterior to appendage III (diagram 2).

The outlet, therefore, is in the same position as that of Tarantula, but the saccules, instead of being in segment III only, or in III and V as in the Charontini, are in segments IV and V. Moreover, the entire coxal gland is evidently formed by fusion


and modification of three originally distinct glands on segments III, IV and V, whereas the two coxal glands of the Charontini are entirely distinct organs belonging to segments III and V, one of which has been lost in Tarantula and Phrynicus.


Species examined :

Schizomus peradeniyensis, Peradeniya, Ceylon. Schizomus vittatus (Gravely), Peradeniya, Ceylon. Schizomus perplexus n.sp. (Gravely), Plains, Ceylon. Schizomus modestus (Hansen), Kedah, Malay Peninsula.

It is only within the last four or five years that any collector has been able to obtain large numbers of the Tartarides. Messrs. Green and Gravely, in the botanical gardens of Peradeniya in Ceylon, used a sieve with which they sifted fallen leaves and debris over a large sheet of stiff paper, the Tartarides and small insects falling through the meshes. iVdopting this method I was fortunate enough to secure, not only at Peradeniya, but also on the plains of Ceylon and in the Malay Peninsula, nearly 500 specimens of these interesting little animals, of which more than 50 were fixed on the spot and sectioned later. The type species — S. crassicaudatus — can not be collected in this way, as it occurs under stones and bricks. A few specimens were secured in Ceylon but they were not sectioned.

Owing to their small size and the minuteness of their organs and individual cells it is not easy to determine the details of the coxal glands, but from a comparison of a large number of specimens one may conclude that the fundamental arrangements of the organ are precisely similar to those obtained in the Thelyphonides. There are two flattened tubular saccules, one in segment IV and another in appendage V, from each of which a collecting duct, consisting of five or six minute cells, leads into a labyrinth tubule running posteriorly to segments V and VI respectively. The labyrinth, lined with striated cells, coils very little and is much simpler in construction than that of Thelyphonus, but nevertheless it has not been found possible to determine just where coalescence of the two tubules takes place be


fore the labyrinth runs forward as a single tubule from segment IV to the outlet on segment III.

Note on the distribution of the Uropygi. Gravely ('12) suggests that Hypoctonus of Burma is hemmed in and is being gradually displaced by the more highly specialized Thelyphonus sepiaris of India on the one hand, and Thelyphonus linganus of the Malay Peninsula on the other; the latter being closely allied to other species found in the Malay Archipelago, where he supposes that this branch of Thelyphonus originated.

The Hypoctonus kraepelini found on Hangkawi Island is rather small, with red legs; the species having been previously unknown to M. Eugene Simon, to whom I gave some specimens. There is one specimen in the Indian Museum at Calcutta, from the northern (Siamese) region of the Malay Peninsula, found and named by Annandale some years ago, but I do not know of any other, nor is any other species of Hypoctonus known to occur outside of Burma.

In Perak and Kedah, British controlled States of the Malay Peninsula, Thelyphonus linganus is very common; but I found lonly one specimen of Hypoctonus kraepelini, in the jungle far removed from any settlement, near the Siamese border. On Langkawi Island, only a few miles from the Kedah coast, there are no Thelyphonus, but everywhere in the primaeval jungle one can collect numbers of Hypoctonus kraepehni, although they do not occur where the jungle has been cleared. The .'species seems to have been left undisturbed on the island, but on the mainland it has become very scarce, owing to pressure by Thelyphonus linganus.

It may also be noted that Tartarides (Schizomus modestus) are very common in Kedah and Perak on the mainland, but during a month's stay in Langkawi I did not find a single specimen.



The coxal gland of the Solifugae was fully described in the previous monograph, and a study of a considerable number of additional specimens goes to confirm the results already ob


tained. Stress was laid on the peculiar nature of the saccule, the lumen of which is filled with a pulpy mass of cells instead of the walls being merely lined with a single layer of dehcate epithelium, which is the usual arrangement. This part of the organ had never previously been described, and the opinion expressed, with some reservation, that it is actually the saccule, has since been proved by the injection of carmine into a number of specimens during a visit to Biskra in 1912. The carmine was injected in minute quantities under the thoracic carapace. The animals very quickly recovered from the effects of the inoculation and were killed afterwards at varying intervals of time — from six hours to one month. The carmine is very quickly taken up by the cells of the saccule, and seems to remain there for a 'ong time — even up to a month. No carmine is ever found in any other part of the organ.

The saccule is situated in segment II, the connecting tubule opening into a long narrow sac (labyrinth sac) which extends posteriorly into segment IV, where, after making several coils, it ends blindly. The walls of this part of the gland are fined with secretory epithefium. At a point in proximity to the saccule there is an opening from the labyrinth sac into the true labyrinth, which consists, as in group I, of a long single coiled tubule lined throughout with striated epithelium. The labyrinth extends back as far as segment VI, where it forms several coils and then runs forward again to the exit tubule and outlet on appendage II (diagram 3 fig. 1). In connection with the exit tubule there is a squirting apparatus, apparently intended to force a stream of the secretion towards the prey held in front of the mouth by the cheficerae, but this part of the organ need not be described again in detail. The probable homology of the coxal gland of the Solifugae with the salivary gland of Peripatus was considered and discussed in detail, and I have since seen no reason for changing my opinion.


Koenenia mirabilis. In the spring of 1914 I was able, with the help of Mr. C. Bonier of Metz, to collect in the olive groves



of Palmi in Calabria about sixty specimens of Koenenia mirabilis, almost all of which were fixed and sectioned for microscopic study. In addition I was able to examine five specimens which had been sectioned and described by Mr. Borner some years previously. The coxal gland had aheady been observed by Grassi, the original discoverer of Koenenia, ('85), and Borner ('04) described it as a single tube running anterior^ from the second

Diagram 3 Coxal gland of group II. Fig. 1 Coxal gland of the Solifugae. Saccule on segment II opening into the long labyrinth sac. From the latter there is an opening into the true labyrinth, running back and then forward again to the outlet on appendage II. In the neighborhood of the saccule the labj^rinth sac is pouched and the labyrinth coils a little. For the sake of clearness these pouches and coils are omitted in the diagram.

Fig. 2 Coxal gland of Koenenia. Saccule on segment II opening into the long labyrinth sac. The true labyrinth is represented only by the vesicle, the outlet from which is on appendage II. S, saccule; CSL, labyrinth sac; CL, labyrinth; ET, exit tubule; V, vesicle.

abdominal (genital) segment to an outlet just posterior to appendage II. This description is correct so far as it goes, but Borner overlooked the saccule and did not appreciate the significance of the tube being single and not double. He was not acquainted with the anatomy of the Solifugae and his efforts were chiefly directed towards grouping Koenenia with the Pedipalps.


As Koenenia is exceedingly minute, its organs and the cells composing them are all on a correspondingly minute scale, so that it is not by any means easy to make out the different parts of the gland, and it is only by studying a large number of specimens that one can come to satisfactory conclusions. There is a small circular saccule in segment II consisting of eight or ten fairly well defined cells enclosing a lumen, from which a short collecting duct of five of six cells leads into a long narrow sac which extends posteriorly into the abdomen to segment VIII (genital), where it forms a short loop and ends bhndly. As already observed by Borner, the cells of this posterior loop differ from those of the straight part of tlie sac, and he quite rightly remarks that none of the cells, either in the loop or in the straight part of the sac, have a striated base; i.e., they are secreting, not excreting cells such as one finds in the true labyrinth of the arachnids. Just below the point at which the collecting tubule or duct changes to the long sac, there is an opening leading into a dilated vesicle, and from the vesicle there is a short exit tubule leading to the outlet just posterior to appendage II (diagram 3, fig. 2).

It seems obvious that this organ must be homologous with the coxal gland of the Sohfugae rather than with that of the scorpions and spiders. The saccule and outlet are on the same segment (II), and we also find the same long sac with a coil at its posterior extremity extending far back and lined with secreting cells. In Koenenia, however, the true labyrinth has practically disappeared as such, being represented merely by the vesicle, which probably only acts as a reservoir for the products of secretion, having no excretory functions, since the walls of the vesicle do not appear to be lined with striated epithelium. - It may be further remarked that if this long sac represented the true labyrinth, not only should it be lined with striated cells but it would necessarily have to loop forward on itself and run forward again to the outlet, thus forming a double tubule on cross section at any point ; but it is a single tube throughout and has been correctly so described by both Grassi and Borner.



The coxal glands of the Arachnids have always been described as being located in such and such a segment, but it is probable that fundamentally each coxal gland represents two segments instead of one. There seems little doubt that in the saccule we have the last traces of the coelom, which would belong to the segment anterior to that of the main part of the coxal gland. One may reflect that in the worms the coelom of one segment is connected with the segment next posteriorly to it by means of a funnel represented by the collecting tubule of the arachnids. The funnel leads through the partition into a coiled nephridium situated in the next segment ; the coiled tubule of the nephridium

Diagram 4 Four segments of a worm, showing the nephridia.

having an outlet near the anterior partition of the segment (diagram 4). When, therefore, dealing with the arachnids, we say, as for instance with Phrynicus, that the saccule lies in segment III and the gland has its outlet on segment III just behind appendage III, it should be borne in mind that the saccule and collecting duct represent the coelom and funnel of segment III, while the labyrinth and outlet belong in reality to segment IV ; the labyrinth having encroached upon and suppressed the coxal glands immediately posterior to itself — a process of encroachment which is still in progress in the Charontini, the small posterior gland of which is composed of the coelom and funnel of segment V and the nephridium of segment VI.



If the cephalothorax of an arachnid such as the scorpion or Thelyphonus be sectioned sagitally through the median line, the gangUa composing the large suboesophageal ganglion can be clearly distinguished, the suboesophageal ganglion being mapped out into a number of neuromeres, separated from each other by a small artery whose course can be very readily determined. Each of these neuromeres represents one of the individual ganglia of which the whole is composed. In the same way also it can sometimes be recognized that the abdominal ganglia are composed of more than one individual ganglion. By enumerating the ganglia in this way it is determined that there are eighteen ganglia in each of the four orders which go to form group I of the arachnids: scorpions, spiders, Uropygi and Amblypygi. One of these eighteen gaiiglia (the cheliceral) has moved up and fused with the supraoesophageal ganglion; the other seventeen being found in the suboesophageal ganglion and in the abdomen.


The nervous system of the scorpions is less concentrated than that of the other arachnids. The suboesophageal ganglion consists of nine neuromeres, the ninth neuromere supplying the abdomen as far as the first lung (fig. 5). In the abdomen itself there are three single ganglia supplj'ing lungs 2, 3 and 4. In the post-abdomen are four ganglia, of which the last is double, having two neuromeres. The entire ventral nerve chain, therefore, consists of seventeen ganglia which, with the cheliceral ganglion, fused with the brain, would make eighteen.




The ganglia are all concentrated in the cephalothorax, the suboesophageal ganghon containing seventeen neuromeres. The photograph is taken from the only section I have which shows the whole of the neiu'omeres in one field, but in a large number of specimens it has been found easy to detect the whole seventeen by making a composite drawing from several sections in series. The Tarantulini Phrynini and Charontini are precisely shnilar as regards this arrangement of the ganglia (figs. 3 and 4) .


Thelyphonides. Both in young and adult specimens the suboesophageal ganglion contains only twelve neuromeres, but at the posterior end of the abdomen there is a large additional ganglion with five neuromeres, representing the last five abdominal segments.

The eggs of Thelyphonus (linganus, the species studied) are carried in a packet attached to the underside of the abdomen of the mother. Just after hatching out, the larvae, white and inert, are carried wrapped around the body of the mother to which they are firmly attached by a gummy substance. At this stage the suboesophageal ganglion consists of nine neuromeres; the other eight ganglia being paired and strung out separately along the abdomen. In a few days, however, and before the next moult, the ganglia have become concentrated; the three anterior pairs becoming fused in the median line and passing forward to join the suboesophageal ganglion, while the five posterior pairs also fuse in the median line and with each other to form the abdominal ganglion as in the more mature animals. Strubell ('92), for Thelyphonus Caudatus, described six pairs of ganglia in the cephalothorax and ten pairs in the abdomen — his specimens probably representing a slightly earlier stage than I had an opportunity of observing; but there are certainly seventeen ganglia altogether, and not sixteen as he described.


Tartarides. The arrangement of the gangha in the Tartarides at first sight appears to differ fundamentally from that in the Thelyphonides, since there is, in addition to the suboesophageal, a large ganglion situated in the anterior part of the abdomen in the neighborhood of the genital segment instead of at the posterior end of the abdomen. Microscopical examination, however, shows that there are nine neuromeres in the suboesophageal ganglion and eight in the abdominal ganglion, making seventeen in all — the same number as in the Thelyphonides — and it may be presumed that the Thelyphonides and Tartarides became differentiated from a common ancestor at a stage when there were still eight separate paired ganglia in the abdomen; a stage now represented in the larval Thelyphonides. On further specialization the ganglia in the Tartarides became concentrated into a single anterior abdominal ganglion, while in the Thelyphonides two groups of three and five ganglia respectively were formed, the first group passing forward to join the suboesophageal ganglion and the latter remaining in the posterior part of the abdomen.

It may be recalled that, apart from their external affinities, which have led to placing these two groups in the same order, the close relationship of the Thelyphonides and Tartarides is clearly established by the similarity of the coxal glands — a fact which makes it easier to accept this explanation of the apparently fundamental difference in the arrangement of the ganglia.


In all spiders there appear to be twelve neuromeres in the suboesophageal ganglion, although with the Araneae verae it is not often possible to make them out except in larval and quite young specimens, but the Theraphosid spiders show the neuromeres more clearly, always to the number of twelve.

In young and adult spiders there are no abdominal ganglia, nor have I ever found any trace of them in larval specimens of the Araneae verae; but in 1913 I was fortunate enough to find in Sumatra a cocoon of a large Chilobrachj's (Theraphosid) containing nearly a hundred larvae just after hatching, and


these I was able to keep alive for two months, fixing four or five for sectioning every few days. In the larval stage the spiders were very inert, with distended abdomen containing a large amount of yolk. They were not pigmented and remained quiescent in the cocoon. At the end of a month they all moulted simultaneously and left the cocoon, passing from the larval to the immature stage. They were now pigmented, longlegged, with small abdomen, and very active, remaining so until the last of them had been killed for sectioning a month later. These points are merely mentioned to show that the spiders were in a normal condition, at any rate in the early stages.

In the youngest specimens five abdominal ganglia were clearly distinguishable, but the ganglia rapidly disappeared and in about a week there were no traces left. I do not think that any observation of these transient abdominal ganglia in spiders has ever been recorded.

In specimens killed at once and two days later, the suboesophageal ganghon contains twelve neuromeres very clearly mapped out. From its posterior end two parallel nerve cords pass through the pedicle, beyond which they separate, running laterally and ventrally in the abdomen. On either side of the abdomen are three very clearly defined, large ganglia, connected together by the nerve cord. The nerve cord can be traced a little beyond the third ganglion, where it is lost, but more posteriorly there are two very much smaller ganglia ying in the neighborhood of the intestinal tract, which has so far only partially been formed. The rest of the abdomen is filled with yolk (diagram 5) .

In specimens fixed three days later (five days after finding the cocoon) the nerve cord was no longer traceable into the abdo-. men; the two posterior ganglia had disappeared and the three anterior were greatly diminished in size.

On the seventh day there were some traces of one or two ganglia, but it was difficult certainly to identify them, and this was the last trace of them that could be found. If the series had begun at this stage the gangha would certainly have been overlooked or not recognized as such.



With the twelve neiiromeres of the suboesophageal ganghon these five paired evanescent gangha in the abdomen make up the usual number of seventeen for group I of the arachnids, which may here be tabulated.




Ambl ypygi




1 1 1


6 segments



7 segments 5 segments


The Solifugae being tracheate animals, there are no blood channels to divide the neuromeres, although in the cephalothorax their place seems to be taken to some extent by tracheae. By means of the tracheae and the nerves proceeding from the suboesophageal ganghon, one can determine that the suboesophageal ganglion is composed of five neuromeres supplying appendages II to VI. The nerve supply to appendage Yl, on which the racquet organs are situated, is very abundant, and is supplemented by a number of secondary ganglia at the roots of the racquets themselves.

With the neuromeres supplying appendage \l, however, the suboesophageal ganglion appears to come to an end, and from this point there is a large nerve cord leading into the anterior p^rt of the abdomen where it swells into a large single ganglion in the region of the genital segment. The nerve cord emerges from this ganglion posteriorly, and can be traced a short distance after which it becomes distributed into small filaments and gradually disappears. There is no trace of any ganglion in the abdomen other than the one mentioned, nor does this ganglion show any evidence of being divided into neuromeres. It is simply one solid ganglionic mass (diagram 6, fig. 1).





Koenenia, the coxal gland of which has been shown to be closely related to that of the Solifugae, also shows a very similar arrangement of the ganglia. In the suboesophageal ganglion there are five swellings on either side, one for each con-esponding appendage, and from the posterior end of the ganglion a nerve cord passes to the abdomen, in the anterior part of which there

Diagram 5 Evansecent abdominal ganglia of Cliilobrachys. Five ganglia on either side of the abdomen, the first three of which are connected by a nerve cord running through the pedicle from the suboesophageal ganglion. Ent, intestine.

are tlii'ee distinct ganglionic swellings, beyond which the nerve cord can not be traced. These three ganglionic swellings in the genital region appear to correspond to the abdominal ganglion of the Solifugae, in which they have become concentrated into a single mass (diagram 6, fig. 2) .

Note on the segmentation of the arachnids. To return to group I, we find in each case that there are eighteen ganglia, and the question naturally arises — to what extent do these ganglia represent segments?



The supraoesophageal ganglion, with its sense organs, need not be considered as having any segmental value except in so far as the ganglion for the chelicerae is concerned — a ganglion which in the earliest stages is obviously suboesophageal and on later development moves over and becomes fused with the brain.

Diagram 6 The ganglia of group II. Fig. 1 Frontal view of the ganglia of the Solifugae. The five neuromeres of which the suboesophageal ganglion is composed are indicated, and below is the large single abdominal ganglion, connected by a nerve cord with the suboesophageal.

Fig. 2 Frontal view of the ganglia of Koenenia. Resembling that of the Solifugae, but showing three abdominal ganglia instead of a single one.

The five anterior ganglia of the suboesophageal mass supply appendages II to VI, which, with the chelicerae, represent the six segments of which the cephalothorax is composed. This number of segments of the cephalothorax obtains in all orders of this group and leaves twelve abdominal segments to be accounted for, each with its attendant ganglion.



In the Pedipalps, both Uropygi and Amblypygi, the relations are clear enough. There is a small first abdominal segment (VII), after which comes the second abdominal, or genital segment (VIII) ; posterior to the genital segment are ten more segments, making twelve abdominal segments in all; the number of segments therefore corresponding to the number of ganglia — six for the cephalothorax and twelve for the abdomen.


In the ordinary spiders all trace of segmentation of the abdomen is lost, but the pedicle is taken to represent a first abdominal segment (VII) after which comes the genital organ representing segment VIII. Liphistius has nine tergal plates on the abdomen, each of which is presumed to represent a segment, so there are still three segments wanting. McBride ('14), quoting Kishinouye, says that in the earlier stages of development of Theridion ten segments are marked out in the abdomen, of which the tenth is not always distinct, nor does it always develop a coelomic cavity. There is no proof, therefore, of correspondence between the number of ganglia and segments in the spiders, although, since the rudiment of the tenth abdominal segment is not always laid down, it is possible that an eleventh and a twelfth segment may have originally existed but have now entirely disappeared.


There is not complete correspondence between the ganglia and the segments in the scorpions, although the total number of eighteen is the same. The cephalothorax is composed of six segments as usual, and the abdomen of twelve, but the first abdominal segment (VII) has disappeared, leaving the genital as the first abdominal segment, posterior to which are the segments bearing the pectens, the four lungs and the last abdominal segment — making seven in all — while beyond this comes the post-abdomen with five segments forming the tail. There are,


therefore, eleven post-genital segments instead of ten as in the Pedipalps.

The first abdominal segment has disappeared, although Brauer ('95) found a transient rudiment corresponding to it in the embryo scorpion.

On examination of the suboesophageal ganglion it is found that the neuromeres 1 to 5 supply appendages II to VI; the seventh neuromere supplies the genital organ, the eighth neuromere the pectens and the ninth the first lung — leaving the sixth neuromere, which should supply the missing segment VII, to be accounted for. This sixth neuromere is clearly mapped out and is quite normal in size, but I have not been able to trace nerves preceding from it. It seems probable, however, that it may be used as supplementary to the pectens, for the nerve supply of which there is a secondary paired nervous mass lying ventrally to the suboesophageal ganglion as shown in diagram 7.

Just ventrally to the eighth neuromere, on either side of the median line, there is a large mass of nerve fibers from which a cord runs posteriorly and joins the nerve from the eighth neuromere, and another cord runs anteriorly to a point opposite the sixth neuromere where there is another large swelling. I have not been able to detect any direct connection between the sixth neuromere and the nerve mass just ventral to it, but it seems probable that there must actually be some such connection and that the sixth neuromere, after the disappearance of segment VII, became utilised as a secondary source of supply for the pectens.

In the Buthidae, but not in the Scorpionidae, Chactidae, or Vejovidae, there is a further extension of this ventral nerve mass anteriorly to about opposite the second neuromere, where there is another small swelling but no direct connection with the suboesophageal ganglion itself. (This extension is shown in diagram 7.)

Whatever may be the true function of the sixth neuromere, it is clear that it is interpolated between the fifth and seventh without having any corresponding segment to supply. Segment VII has disappeared, but the ganglion which originally supplied it remains.



Turning now to the abdomen we find twelve segments, of which the first is the genital, but there remain only eleven ganglia to supply them since we have already disposed of seven ganglia (segments I to VII) (VII being transitory). The ganglia of the genital organ (VIII), pectens (IX) and first lung (X) are fusedj^with the subo esophageal ganglion. Next comes a

Diagram 7 Sagittal view of the su'ooesophageal ganglion of the scorpion, showing nine neuromeres, the destination of the nerves proceeding from them, and the ventral nerve mass secondarily supplying the pectens. It is understood that all these structures can not be seen in one section. The neuromeres can only be distinguished in sections through the median line, whereas the nerves running to the appendages and the secondary nerve mass for the pectens are paired structures which appear in sections on either side of the median line.

series of three single ganglia for lungs 2, 3 and 4 (XI, XII, XIII), posterior to which is the last abdominal segment (XIV). In the post-abdomen there are five segments and four ganglia, of which the last, situated in the fourth segment, is double (having two neuromeres) and clearly suppUes segments 4 and 5. The last segment in the abdomen, therefore, appears to be without a ganglion. The first ganglion of the post-abdomen is cer


tainly single,* but it lies in the anterior part of this segment and extends anteriorly through the constriction into the last segment of the abdomen in the egg just before hatching. I have not had an opportunity of examining earUer stages. The suggestion may be put forward that the last abdominal and first postabdominal segments represent in reality a single segment which has become secondarily divided into two in order to give to the base of the tail greater strength than could be obtained by utilizing the di\'ision between two primary segments. This suggestion seems to receive some support from Brauer's figures of the development of the scorpion, a part of one of which is reproduced in diagram 8. Brauer, in his text, treats segments XIV and XV (first post-abdominal) as if they are two distinct segments, but it appears from his figm^es as if segment XIV folds over on itself to form the base of the tail, without at first any segmentation at the point of the fold.

That such a division is possible is indicated by the secondary division between the fifth post-abdominal segment and the poison gland, which corresponds to the telson of Limulus and the Euiypterides. The telson is not considered as representing an original segment, since it always lies behind the anus. There is no ganglion for the poison gland in the scorpion. Similarly the Uropygi and Koenenia have a tail jointed to the last abdominal segment, but the tail is not recognized as a segment, nor has it any corresponding ganglion.

The scorpion-like Eurypterides of the Silurian have ten postgenital segments — one less than the scorpion; the pre-genital segment (VII) being recognizable on the dorsal surface, but on the ventral surface it is fused with the genital plate. On the other hand, the Silurian scorpion, Palaeophonus, appears to have the same number of segments as the modern scorpions.

The general conclusion is that in group I of the arachnids the number of ganglia corresponds to the number of segments — eighteen of each — obvious in the Pedipalps; not so clear, but probable, in the scorpions; in the spiders possible but not proven.

■* I have looked in vain through scores of specimens for some indication of division of this ganglion into two neuromeres.



In regard to group II, since the large abdominal ganglion of the Solifugae is a single, solid mass, without any e\'idence of neuromeres, nothing definite can be said about it. It may represent a fusion of all the original abdominal ganglia, or it may be composed of one or more of the original ganglia which have suppressed and taken over the functions of others. The same may be said of the three abdominal ganglia of the Palpi

Diagram 8 From Brauer. Development of the tail of the scorpion. Between segments XIV (last abdominal), and XV (first post abdominal), there is no evidence of segmentation. Segment XIV seems to double over on itself to form the base of the tail. In the tail itself only three out of the five segments have been cut out at this stage.

grades. In neither of these two orders can any relation between the abdominal segments and ganglia be determined. In this respect they differ again from group I; and a still further difference between the two groups can be pointed out; in each of these two orders of group II there are nine post-genital segments instead of ten, as is probable in all the orders composing group 1.



Tabulation of differences between group I and

group II



Segment II Segment II Present 7 or 9 Doubtful 9

Segments III, IV, V

Segments III, V

Coxal gland labyrinth sac


Absent Always 18


My best thanks are due to Professor Pruvot for permission to work in his laboratory of comparative anatomy at the Sorbonne, and also to M. Eugene Simon and Mr. F. H. Gravely for specimens and suggestions. The diagrams were made for me by Miss G. E. Carver.


BoRNER 1904 Beitrag zun Kenntniss der Pedipalpen. Zoologica.

Brauer 1&>95 Zeit. fiir wiss. Zoologie, vol. 59, p. 360.

BuxTox 1913 Zoologische Jahrbiicher Abt. Anat. July Supplement.

Grassi 1885 Naturaliste Siciliano.

Gravely 1912 Records of the Indian Museum, Calcutta.

McBride 1914 Textbook of Embryology, vol. 1, p. 224.

Strubell 1892 Annals and Mag. Xat. Hist., vol. 10.



1 Tarantula palmata. Saccule of the coxal gland. X 175. The drawing is outlined from a photograph, the cells being drawn in semi-diagramatically, to show more clearly than is possible in photograph the differences between the cells of the saccule, the collecting duct and the labyrinth. The lumen of the saccule is broken up into channels by blood capillaries or rather lacunae) which push in the walls, and thus increase the surface of the saccule. In the diagrams this feature of th saccule is not shown. S, saccule; CD, connecting duct; CL, labyrinth; ET, commencement of exit tubule; M, muscle and connective tissue indicated by horizontal broken lines; BC, blood capillaries indicated only by vertical broken lines; GL, suboesophageal ganglion.

2 Vejovis flavescens. Photograph of coxal gland of scorpion. X 100. With the help of figure 1 the constituent parts can be made out. In the center is the saccule; its lumen, broken up by inhanging capillaries, opening on the right into the labyrinth, whose tubules in this case completely surround the saccule except at one point (CA) \vhere the blood supply enters. At ET is the commencement of the exit tubule, but the tubule itself is not shown. The coxal gland rests posteriorly (right) upon the diaphragm D which separates the cephalothorax from the abdomen. C, caeca of the intestine; .V, nerve supplying appendage \T; M, muscle; CA, coxal artery; ET, exit tubule; D, diaphragm.





1 Tarantula palmata. Neuromeres of the suboesophageal ganglion. X 60 Sagittal secton through median line of cephalothorax outlined from the photograph showing 17 neuromeres. Anterior to the suboesophageal ganglion the tissues are broken up on account of extraction of the chelicerae, which are too hard to section. ART, artery, the branches from which run between the neuromeres; CH, chitin of external surface; CT, muscle and connective tissue indicated by broken horizontal lines; A^^4, nerve to abdomen.

Fig. 2. Photograph from which the drawing was outlined.




'i .'.'v^b^V



1 Immature scorpion. Buthus occitanus. Neuromeres of the suboesophageal ganglion. X 40. Sagittal section through median line of cephalothora.s' outlined from the photograph showing 9 neuromeres, of which the 6th is well defined and normal in .«;ize. A, artery from which branches run between neuromeres; BR, brain; C, caeca of the intestine; CT, connective tissue and muscles; GSO, suboesophageal ganglion; GAi, 1st abdominal ganglion; HT, heart; INT, intestine; OES, oesophagus; 0, median eye.

2 Photograph from which the above drawing was outlined.






N. E. McINDOO Bureau of Entomology, Washington D. C.



Introduction and methods 33

The olfactory pores 35

Disposition 35

a. Bombyx mori 9 35

b. Other species 38

c. Generic, specific and sexual variations 41

Structure 44

a. External structure 44

b. Internal structure 44

The antennal organs 50

Disposition 51

Structure 52

Summary 54

Literature cited 54


In the investigation herein recorded a careful study of the morphology of the olfactory pores of Lepidoptera has been made in order to determine whether these organs are better adapted anatomically than the antennal organs to receive olfactory stimuli.

The investigators who have performed experiments on butterflies and moths with mutilated antennae have concluded that these appendages bear the olfactory organs, regardless of whether or not the antennal organs are anatomically fitted to receive olfactory stimuli. Since these investigators failed to study sufficiently the behavior of the insects investigated, it is possible that the responses observed misled them in determining the seat of the olfactory organs (see the author's paper, '14 c).



34 N. E. McINDOO •

In 1857 Hicks discovered pore like organs on the wings of Lepidoptera and called them vesicles. He says: In moths they are very apparent, being greatest in the noctuae [Noctuidae] and Bombycidae. There are about 100 vesicles on the upper surface of the posterior wing, and half that number beneath, besides some few on the nervures [veins]. In the butterfly they are smaller, but arranged in more definite groups, about three in number." In 1860 Hicks discovered them on the legs of Lepidoptera, and the present writer ('14 a and b, '15 and '16) has made a comprehensive study of these organs in Hymenoptera and Coleoptera, and since he proved experimentally that they receive olfactory stimuli, they were called olfactory pores. On account of other investigations, the writer has not had time to determine the physiology of these organs in Lepidoptera and for this reason the present paper deals with only their morphology. To obtain material for the study of the disposition of the olfactory pores, dried museum specimens were mostly used. In regard to preparing the specimens with caustic potash and to bleaching them with chlorine gas, the reader is referred to the writer's work on Hymenoptera ('14 b, p. 295).

To obtain material for the study of the internal anatomy of the organs herein discussed, moths emerged a short time from the cocoons and a butterfly were used. The material was fixed in Carnoy's fluid (equal parts of absolute alcohol, chloroform, glacial acetic acid, with corrosive sublimate to excess) and was embedded in celloidin and paraffin. The sections were cut five and ten microns in thickness and were stained with Ehrlich's hematoxylin and eosin.

All the drawings were made by the writer and all are original except figure 10, which represents the antennal organs of a moth and was copied from Schenk ('02). The drawings were made at the base of the microscope with the aid of a camera lucida.

The museum specimens were secured through the courtesies of Dr. L. O. Howard and Mr. August Busck, and the writer is also grateful to Mr. Carl Heinrich for verifying the identification of all the Lepidoptera used.



Before making a study of the anatomy of the organs, called the olfactory pores by the writer ('14 a), the distribution and number of them were first investigated.


In making a comparative study of the disposition of the olfactory pores in Lepidoptera, 40 species, belonging to 36 genera and representing 19 families, were used. In most cases onlyone specimen of each species was employed, and whenever a portion of an appendage or an entire appendage was missing or was badly mutilated in being prepared for study, the supposed number of pores on this portion or entire appendage was regarded the same as the number found on the corresponding portion or entire appendage on the opposite side of the body. Since the pores on only one specimen for each species were counted, the total number of pores recorded can not be a fair average. Besides this error, there is also another small probable error for each species, because a few of the pores were probably overlooked, and often, as on the tibiae, it was impossible to distinguish the olfactory pores from hair sockets. Only the legs, wings and mouth parts were examined, although in two or three instances the chitinous parts of the reproductive organs were also examined, but no olfactory pores were seen on them. The mouth parts of more than one-half the specimens were either missing or were so mutilated that the pores on them could not be counted; nevertheless, regardless of all the errors, the total numbers of pores recorded are probably not far from being accurate, but they are perhaps always shghtly less than the numbers actually present. The sex of the species, except in tour cases, was not determined.

a. Bombyx mori 9 . Since the silkworm moth, Bombyx mori, is conveniently studied and , as its olfactory pores are typical foi most of the moths examined, the disposition of them will be described in detail, and then the variations found in the other species will be given.



The wings have dorsal and ventral surfaces, and the legs may be divided for description into two surfaces. The inner surface faces the body of the insect and the outer surface is directed away from it. On the specimen examined, four groups of pores, besides many scattered pores, were found on each front wing; three groups, besides several scattered pores, on each hind ^\dng ; and four groups, besides two isolated pores on each leg.

Fig. 1 Wings of silkworm moth, Bombyx mori 9 , showing location of groups of olfactory pores as indicated by the numbers 1 to 7, and disposition of scattered pores on the veins as represented by dots, X 3. A and B, dorsal and ventral surfaces of right front wing respectively; and C and D, dorsal and ventral surfaces of right hind wing respectively. 1 to S A, first to third anal veins; C, costa; cl to 2, first and second cubital veins; M, media; ml to S, first to third medial veins; R, radius; rl to 5, first to fifth radial veins; Sc, subcosta.

For descriptive purposes in locating the olfactory pores on the wings, the veins at the bases of the wings may be called costa (fig. 1 A-D, C), subcosta (Sc), radius (R), me(^ia (M), and anal vein (1 to 3 A) ; the names, given by Comstock (manual, '01) to the distal ends of these veins, are also appended, so that no confusion may arise. The groups and scattered pores are located as follows: Groups nos. 1 to 7 lie on the bases of the wings, nos. 1 to 4 being on the front wing and nos. 5 to 7 on the hind wing; nos. 1, 2, 3, 5 and 6 are found on the dorsal surfaces while nos. 4 and 7 lie on the ventral surfaces. Nos. 1 to 3 (fig.


1 A) seem to lie on the union of the radius and media, but cross sections (figs. 6 A and 9 C) show that nos. 1 and 2 He on the media while no. 3 lies on the radius. No. 4 (fig. 1 B) lies on the broad portion of the subcosta which unites with the radius and media. No. 5 lies on the radius and no. 6 (fig. 1 C) on the media, but perhaps no. 6 lies more correctly on the union of the radius and media (fig. 6B). No. 7 (fig. 1 D) is located on the broad portion of the subcosta.

The number of pores in groups nos. 1 to 7 on the right wdngs are as follows: No. 1 — 11, no. 2 — 52, no. 3 — 56, no. 4 — 67, no. 5- — 128, no. 6 — 44, and no. 7 — 12; on the left wings: no. 1 — 10, no. 2 — 50, no. 3 — 52, no. 4 — 74, no. 5 — 143, no. 6 — 51, and no. 7 — 11. On the dorsal surface of the right front wing, the scattered pores are located as follows : 2 at the distal end of the costa (fig. 1 A), 10 extending the full length of the subcosta, 2 near the base and 5 at the distal ends of the radial branches, 6 at the distal ends of the medial branches, and 2 at the distal end of the anal vein; on the ventral surface of the same wing: 8 extending one-half the length of the costa (fig. 1 B), 5 near the middle of the radius and 8 at the distal ends of its branches, 6 at the distal ends of the medial branches, and 2 at the distal end of the radial vein. On the dorsal surface of the right hind wing, the scattered pores are as follows: 3 near the base of the subcosta (fig. 1 C), 1 at the distal end of a radial branch, 6 near the base of the media, and 6 on the medial branches; on the ventral surface of the same wing: 1 near the base of each the costa and subcosta (fig. 1 D), 1 at the distal end of a radial branch, 4 on a medial branch, and 2 at the distal end of the second branch of the anal vein. The disposition of the scattered pores on the left wings is very similar to that just given for the right wings.

Groups nos. 8 and 9 of the olfactory pores lie on the outer surface of the trochanter (fig. 2 E) of each leg near the anterior margin, while no. 10 is located on the inner surface of the same segment (fig. 2 F). No. 11 lies at the proximal end of the femur (fig. 2 F) of each leg on the inner surface near the posterior margin, and the two isolated pores (fig. 2 E, a) lie at the same position on the outer surface of each leg.



The number of pores in each group on the right leg is as follows : Front leg, no. 8—10, no. 9—8, no. 10—9, no. 11—9; middle leg, no. 8—11, no. 9—6, no. 10—6, no. 11—8, hind leg, no. 8—11, no. 9 — 6, no. 10 — 8, and no. 11 — 8. The variation in the number of pores on the left legs is very similar to that just given.

Fig. 2 Portions of legs of Bombyx mori 9 , showing location of groups nos. 8 to 10 of olfactory pores on the trochanters, no. 11 and a on the femurs, X 192. A and B, outer and inner surfaces of right front leg respectively; C and D, same of right middle leg; and E and F, same of right hind leg.

All six legs of the specimen of Bombyx mori 9 examined bear 218 olfactory pores; the front wings carry 423 pores, and the hind wings carry 420 pores; no pores were observed on the mouth parts; all of these combined make 1061 olfactory pores.

h. Other species. The greatest variation found in the olfactory pores of the other species examined is in regard to the total numbers of the pores. On the average, butterflies have only twothirds as many pores as have moths ; this difference is due solely


to the smaller number of pores on the wings of butterflies, and chiefly to the smaller number on their hind wings. The preceding conclusion was derived from the following data: The total number of pores on the legs of moths vary from 71 to 240, with

140 as an average; on the legs of butterflies from 73 to 196, with

141 as an average; on the front wings of moths from 52 to 662, with 399 as an average; on the front wdngs of butterflies from 206 to 404, wdth 303 as an average; on the hind wings of moths from 45 to 663, with 382 as an average; on the hind wings of butterflies from 130 to 228, with 197 as an average; on the mouth parts of moths from to 59, and on the same appendages of butterflies from to 34. The total number of pores on moths vary from 222 to 1422, with 930 as an average; and on butterflies from 514 to 784, with 645 as an average.

As a general rule, all the other insects hitherto examined for olfactory pores by the writer show^ed that the larger the species the greater was its total number of pores. Relative to Lepidoptera this ruling is not true, because the larger species have about the same number of pores as have the smaller species (table 1, p. 43).

The other variations, most of which are small, pertain chiefly to the distribution of the olfactory pores. For sake of brevity, instead of using the long scientific names of the Lepidoptera examined, the species will be numbered from 1 to 43, and those interested in associating the names of the species with the variations described may do so by referring to the names and numbers of the species in the table on page 43.

No wdng was found devoid of olfactory pores, although they are reduced in number on the rudimentary wings of the females of Hemerocampa and Alsophila (nos. 10 and 16). These wangs will be discussed under sexual variations on page 41. The disposition of the pores on the wings of the other specimens is more or less similar to that already described for Bombyx mori. The number of groups of pores on a wing depends on how closely the pores he to one another. At a given place on one wing the pores may be scattered and therefore do not constitute a group, wliile on another wing at the same place, the

40 N. E. McINDOO

pores are well grouped. This fact chiefly explains the variation in the number of groups of pores. Two specimens examined (Nos. 16, 18) have only two groups on each front wing and one group on each hind wing; three specimens (nos. 15, 17, 26) have three groups on each front wing and two on each hind wing; 11 specimens (nos. 1, 3, 5, 6, 8, 9, 11, 12, 19, 20, 43) have three groups on each front wing and three on each hind wing; one specimen (no. 25) has four groups on one front wing and two on the other, two groups on one hind wing and one on the other hind wing; two specimens (nos. 10, 22) have four groups on each front wing and two on each hind wing; six specimens (nos. 13, 14, 21, 23, 24, 28) have four groups on each front wing and three on each hind wing; ten specimens (nos. 7, 27, 31, 33 to 37, 41, 42) have four groups on each front wing and four on each hind wing;, two specimens (nos. 39, 40) have four groups on each front wing and five on each hind wing; three specimens (nos. 29, 30, 32) have five groups on each front wing and three on each hind wing; two specimens (nos. 2, 4) have five groups on each front wing and four on each hind wing; and one specimen (no. 38) has seven groups on each front wing and four on each hind wing. Of the 15 specimens of butterflies examined it is thus seen that each of 15 has 16 or more groups of pores on both pairs of wings, while of the 28 specimens of moths examined each of 24 has 14 or less groups on both pairs of wings. This indicates that the olfactory pores in butterflies are the more highly developed.

Practically every specimen examined has at least a few scattered pores on the wings, and, as a rule, the fewer the groups of pores the greater is the number of scattered pores. It is common for the pores to extend the full length of one or more veins and to terminate at the distal ends of the veins in pairs as shown in figure 1.

Every leg of the specimens examined bore pores, but the more the legs are reduced in size the fewer the pores they bear. The disposition of the pores on the trochanters and femurs of a few of the species is similar to that of the honey bee, but only occasionally are pores found on the proximal ends of the tibiae and


never on the tarsi, as observed in the Hymenoptera. A few pores, usually near the distal ends of the tibiae, were seen in 21 of these specimens (nos. 1, 2, 3, 5, 7, 8, 11, 12, 19, 22 to 26, 33 to 35, 37, 38, 40, 43), and pores were observed in the tibial spines of 12 individuals (nos. 1, 2, 6, 15, 16, 22 to 26, 28, 43). Relative to the isolated pores on the femurs at the position marked a in figure 2 E, one pore was found at this position in 28 specimens (1, 2, 3, 6, 10 to 12, 15 to 17, 19, 23, 26, 28, 30 to 41, 43); two pores were found at this position in three specimens (nos. 5, 13, 14); four pores at this position in one specimen (no. 42).

c. Generic, specific and sexual variations. The generic and specific variations are considerable when the total number of pores is considered. The noctuids (Nos. 4 to 9, table 1, p. 43) well illustrate the generic variation and the three species of Pontia (nos. 35 to 37) illustrate the specific variation. The total number of pores of the noctuids vary from 852 to 1422, with a difference of 570 pores, and the species of Pontia from 632 to 784, with a difference of 152 pores.

Excluding the females with rudimentary wings, the sexual variation is insignificant. The male each of Bombyx (nos. 13 and 14) and of Sanninoidea (nos. 19 and 20), the peach-tree borer, has 13 more pores than has the female of the same species.

The male of the tussock moth, Hemerocampa leucostigma, was not examined, but the number of pores on the rudimentary wings of the female (no. 10) does not seem to be greatly reduced. The front wings are plainly visible to the unaided eye, but the hind wings are not, and both pairs are nothing less than thick pads in which the veins are not distinct. As in Bombyx, three groups of pores (nos. 1 to 3) lie on the dorsal surface of each front wing (fig. 3 A) and one (no. 4) on the ventral surface; nos. 5 and. 6 are present on the dorsal surface of the hind wing (fig. 3 B), but no. 7, usually present on the ventral surface of the hind wing, is absent. The scattered pores on the front wings are twice, and those on the hind wings are three times as large as the ones in the groups; ordinarily the scattered pores are little, if any, larger than those of the groups.



Both pairs of wings of the female geometric! moth, Alsophila pometaria (no. 16), are invisible to the unaided eye, and are so greatly reduced that the front wing (fig. 3 D) is only about threefourths as large as the tegula {Tg) and no larger than the patagia (fig. 3 C) on the prothorax. The hind wing (fig. 3 E) is about one-third as large as the front wing. No pores are present on the ventral surfaces of the wings, and only two groups lie on the dorsal surface of the front wing and only one on the hind wing;

Fig. 3 Wings, tegulse {Tg) and patagia of so-called wingless female moths, showing location of groups of olfactory pores and scattered pores on dorsal surfaces of wings. A, tegula and front wing and B, hind wing of tussock moth, Hemerocampa leucostigma, X 21; C, patagia, D, tegula and front wing and E, hind wing of the geometrid, Alsophila pometaria, X 5.3.

a few scattered pores were also seen on each wing. All of these pores are larger than usual.

To determine the individual variation in the total number of pores, three specimens of the lesser wax moth, Achroia grisella (no. 22), were examined. The total numbers of pores on the legs and wings of these moths are 1112, 1168 and 1238. A part of this variation may be attributed to sex, because the sex of each individual was not determined.

The following table (p. 43) includes the family, number and name of the species, the number of olfactory pores on the legs,

TABLE 1 Number of olfactory pores on legs, wings and mouth parts of Lepidoptera



c « 





o S

a ^







Phlegethontius quinquemaculata






Ceratomia catalpae








Apantesis sp.







Prodenia arnithogalli







Lycophotia margaritosa






Agrotis unicolor







Cirphis unipunctata







Alabama argillacea






Caenurgia erechtea







Hemerocampa leucostigma 9








Tolype velleda






Malacosoma americana








Bombyx mori cf







Bombyx mori 9








Alsophila pometaria cT






Alsophila pometaria 9






Synchlora aerata







Thyridopteryx ephemeraeformis







Sanninoidea exitiosa cf







Sanninoidea exitiosa 9








Loxostege obliteralis






Achroia grisella







Ephestia cautella








Oxyptilus tenuidactylus







Laspeyresia pomonella







Eucosma scudderina







Atteva aurea








Tinea pellionella







Argynnis cybele






Grapta interrogationis






Euvanessa antiopa







Junonia coenia







Heodes hypophlaeas







Everes comyntas







Pontia monuste






Pontia protodice






Pontia rapae






Callidryas eubule






Eurema nicippe







Eurema euterpe







Papilio polyxenes







Papilio troilus







Eudamus proteus

122 71















44 N. E. McINDOO

wings, mouth parts, and the total number of pores on each of the 43 specimens examined. In the preceding pages the insects are usually referred to in this table by their respective numbers. The blank spaces in the fourth column mean that the mouth parts were either missing or so badly mutilated that the pores on them could not be counted. Owing to the rudimentary condition of the mouth parts, no attempt was made to identify accurately the various mouth appendages, but most of the pores recorded were found on the bases of the palpi, as is best illustrated in the lesser wax moth (no. 22). The only pores found on the mouth parts of Euvanessa antiopa and Papilio polyxenes (nos. 31 and 41) lie in two groups at the base of the proboscis, on the dorsal surface, in the same position as recently stated for the honey bee by the writer ('16).


The preceding pages deal with the disposition of the olfactory pores, and a discussion of the anatomy of these organs is given in the following pages.

a. External structure. When the superficial ends of the olfactory pores are examined under a high-power lens with a strong transmitted light, the pores appear as small bright spots, each of which is surrounded by darker chitin, the pore border (fig. 4 E, PorB) and by the pore wall (PorW). The pore aperture (PorAp) is usually oblong, but may be round; its size depends upon the focusing level of the microscope, showing that it is funnel-shaped. The size of the pores vary considerably, as may be seen by referring to figure 4.

b. Internal structure. The olfactory pores have been called dome-shaped organs, but the domes are not always present as is shown in figures 5 C and 8 C; in those sections in which the domes are not visible the microtome knife probably passed through the organ too far from the pore aperture. The domes (fig. 5 B and D, D) in the wings and legs of moths, and in the wings (fig. 9 A) of a butterfly rise slightly above the surface of the surrounding chitin, while in the legs of the same butterfly



the domes (fig. 9 B, D) lie below the surface of the surrounding chitin. Neither Guenther ('01) nor Freiling ('09), who have studied the anatomy of these organs, saw the sense fibers pass through the domes, and consequently they speculated about the function of these pores. Owing to the small size of the pore apertures and to the great thickness of the domes the present

® &




@ © ® C

§ @

© < '





/ © ^


(2) A®
















® © F









Fig. 4 External view of olfactory pores of Bombyx mori 9 , showing variations in size, X 320. A, group no. 1; B, 6 of group no. 2; C, 10 of group no. 3; D, 12 of group no. 4; E, 15 of group no. 5; F, 8 of group no. 6; G, group no. 7; H, group no. 8; I, group no. 9; J, group no. 10; K, group no. 11; L, group a and a hair socket (HrSk) ; and M, a hair (Hr), 4 hairs sockets and 4 pores from tip of hind wing on ventral side. Hr, hair, called S. trichodea on antenna; HrSk, hair socket; PorAp, pore aperture; PorB, pore border; PorW, pore wall.

writer has not been able to find a pore which well illustrated the sense fiber running into the pore aperture, nevertheless four pores (figs. 5 A and D, 8 A and E) were found, each of which had a light streak passing through the dome (D). A few sections cut obliquely clearly show the sense fibers (figs. 7 and 8B) connecting with the pore apertures, so that in the present writer's opinion there can be no doubt about the peripheral ends of the sense fibers coming in direct contact with the external air.



All the olfactory pores studied are more or less flask-shaped structures, although the width of the flask is often equal to its height, and the mouths of such flasks are quite flaring; this is particularly true for the pores in the legs (fig. 5 D and E) of Bombjrx, but in the wings (fig. 5 A and B) of the same insect the pores are more flask-shaped.

Fig. 5 Cross sections, showing internal anatomy of olfactory pores of Bombyx mori, X 506. A, from front wing; B, from hind wing, showing dome (D); C and D, from trochanter; and E, from femur. Con, chitinous cone of olfactory pore; D, dome; Hyp, hypodermis; N, nerve; NB, nerve branch; NF, nerve fiber; PorAp, pore aperture; SC, sense cell; SF, sense fiber; Tr, trachea.

Cones are usually present and are of three types : The external boundary of the most common type (fig. 5 E, Con) is convex, that of the second type (fig. 8 C) is almost flat, and that of the third type (fig. 5 A) is concave. The first type is found both in the legs and wings while the other two types are found in only the wings; only the first and second types are present in the wings of Achroia, the second type being the more common.

The sense cells (fig. 5 A and E, SC) are more spherical and therefore less spindle-shaped than those found hitherto in other



insects. The nerve fibervS (fig. 5 A, NF) run directly to the nerves which he surrounded by blood and the tracheae (Tr.) The sense fibers (fig. 5 D and E, SF) are easily traced to the cones, but seldom through them. Sometimes they are surrounded by


Fig. 6 Semidiagrammatic drawings of cross sections of wings of Bombyx mori, showing internal anatomy at bases of wings and groups nos. 2, 4, 6 and 7 of olfactory pores, X 80. A, from front wing and B, from hind wing. BlSin, blood sinus; C, costa; M, media; N, nerve; NF, nerve fiber; P, isolated olfactory pore on costa; R, radius; SC, sense cell; Sc, subcosta ;.Tr, trachea.

the hypodermal secretion (fig. 8 A, C and D, HypS) which has formed the cones.

To understand the anatomy of the wings at the position where the olfactory pores are located, the reader is referred to figure 6 A and B, which are semidiagrammatic drawings taken from sections across the bases of the wings, A being from the front



and B from the hind wing. At once it is seen that instead of the surfaces of the wings being smooth, as is generally beheved, thej^ are more or less corrugated, the elevations being formed by the veins and the depressions usually by the flexible chitin between the veins. The position of a vein on one side of a wing seldom corresponds exactly to the position of the same vein on the other side of the wing, and consequently it is difficult to identify

Fig. 7 Semidiagrammatic drawing of an oblique section through femur, trochanter and coxa of Bombyx mori, showing internal anatomy at this location in leg, and groups nos. 8, 10 and 11 of olfactory pores, no. 10 being shown partially from a superficial view, X 106. BlSin, blood sinus; Hr, hair, called S. trichodea on antenna; Me, muscle; A'^, nerve; NB, nerve branch; SC, sense cell; Tr, trachea.

the veins in cross-sections; but after making a study of the serial sections this difficulty was alleviated. The identification of the veins then makes it easy to recognize the groups of pores. The sections illustrated passed through groups nos. 2 and 4 on the front wing" and nos. 6 and 7 on the hind wing. The sense cells {SC) lie in thickened portions of the hypodermis and are constantly bathed with blood. The nerve fibers {NF) spread out fanlike from the nerves and unite with the sense cells. It is to be noted that a nerve (iV) and one or more tracheae {Tr)



lie in each vein, and that one of the sections passed through an isolated pore in the costa of the hind wing (fig. 6 B, P).

To understand the anatomy of the legs at the position where the olfactory pores lie, the reader is referred to figure 7, which is a semidiagrammatic drawing taken from one section cut obliquely across the femur, trochanter and coxa. In dead insects the femur

Fig. 8 Cross sections, showing internal anatomy of olfactory pores of other moths, X 506. A to G, from wings; H, from trochanter; and I, from femur. A, B and H, from catalpa sphinx, Ceratomia catalpse. B is cut obliquely; C, D and I, from lesser wax moth, Achroia grisella; E, from Prodenia ornithogalli; F, from Atteva aurea; and G, from front wing of tussock moth. HypS, hypodermal secretion forming cone; A"^, nerve; NB, nerve branch; PorAp, pore aperture; Tr, trachea.

and trochanter form a right angle with the coxa, and this fact explains how an oblique section may be cut passing through all three segments. This section passed through groups nos. 8, 10 and 11, no. 10 being shown partially in cross section and partially from a superficial view. Most of the muscles (Me) are cut longitudinally, but those near the sense cells (SC) run transversely, leaving a space, the blood sinus (BlSin), in which the sense cells lie. The nerve branch (NB) leaves the nerve




(N), runs to the hypodermis (figs. 7 and 9 B), and then divides into nerve fibers which unite with the sense cells.

Figure 9 C is similar to figure 6 A, but represents a section cut obliquely through the base of the front wing of a butterfly,


Fig. 9 Cross sections of wing and leg of cabbage butterfly, Pontia (Pieris) rapae, showing internal anatomy of these appendages at positions where olfactory pores are located. A, from front wing and B, from trochanter, X 506. C is a semidiagrammatic drawing from an oblique section cut in direction of line a in figure lA and B, showing groups nos. 2, 3 and 4, of olfactory pores, X 106. C, costa; D, dome; HrSk, hair socket; Hyp, hypodermis; M, media; N, nerve: NB, nerve branch; R, radius; SC, sense cell; Sc, subcosta; Tr, trachea.

Pontia, in the direction of the line a in figure 1 A and B. The section passed through groups nos. 2, 3 and 4 and longitudinally through the nerve (N) for a short distance.


Several investigators have studied the morphology of the antennal organs in Lepidoptera, but since Schenk's work ('02)



seems to be the latest and perhaps the best, most of the preceding discussion will be taken from his paper.


Schenk has carefully studied the antennal organs in both sexes of the following moths: One geometrid (Fidonia piniaria), two bombycids (Org>da antiqua and Psyche unicolor), and one zygaenid (Ino (Atychia) pruni). He found the following five sense organs on the antennae of both sexes: (1) Pit pegs (Sensilla coeloconica), (2) pegs (S. basiconica), (3) end pegs (S. styloconica), (4) bristle-like hairs (S. saetica) and (5) ordinary hairs (S. trichodea). In regard to the disposition of the above five types of sense hairs on the feathered antennae of the males and on the non-feathered ones of the females, the following maj^ be said : These sense organs not only he on the shafts of both types of antennae, but also on all the barbs of the male antennae. The total number of sense hairs of each type found on the antennae of the four moths examined by Schenk is tabulated as follows :

Number of sense organs on antennae of moths (after Schenk)














Pit pegs

End pegs

Bristle like


Ordinary hairs




117 Numerous

100 16






80 Numerous






Present Numerous


120 Numerous


120 Scarce

Hicks ('59) and Hauser ('80) saw the same organs, particularly the pit pegs, in the antennae of the butterflies, Argynnis and Vanessa. Hauser says that each segment of the antennal knob of Vanessa carries about 50 pits bearing pegs.

52 N. E. McINDOO

It is thus seen that in Lepidoptera pore plates, common to most insects, are entirely absent, while the pegs are practically wanting; however the pegs seem to be replaced by the end pegs. Of the five sense organs on the antennae of Lepidoptera, only the pit pegs and end pegs are regarded as olfactory organs. Since the end pegs are totally absent in Psyche and Ino, it does not seem reasonable that they can be olfactory organs for any lepidopteron, and, providing, there is no sexual difference in the size of the pit pegs, the male of Fidonia smells three and onehalf times as well as the female of the same species; the male of Orgyia eight times as well as the female of that species, and while the male of Psyche smells well, the female of the same species can not smell at all. It also seems doubtful that the pit pegs can function as olfactory organs.

Now let us inquire if any of these organs are really adapted anatomically to receive olfactory stimuli.


The ventral side of the shaft (fig. 10 A) of the antenna of Fidonia bears more sense organs than does the dorsal side, and the terminal segment bears more pit pegs (PPg) than does any other segment. Observed from the side under a high-power lens, the pit pegs look like pits lined with hairs, but when viewed from the tops of the pits, the organs resemble small wheels in that the hairs form the spokes and the base of the peg the hub. Two of the pit pegs thus viewed are shown on the segment third from the last one. ,

A longitudinal section of a pit peg shows that a peg (fig. 10 B, PPg), surrounded by a crown of pseudo-hairs (Hr^), arises from the bottom of the pit and that the base of the peg is connected with a sense cell group (SCG). The end pegs (fig. 10 C, E Pg) are nothing more than short, stubby hairs supported on stout projections, called styles (St), which are innervated. The bristle-like hairs (fig. 10 D), ordinary hairs (fig. 10 E, Hr) and pegs are also innervated.

Other authors who have studied the anatomj^ of these organs describe them as similarly to the preceding account, except the


-EPg PPg


Fig. 10 Antennal organs of the geometrid moth, Fidonia piniaria, copied from Schenk ('02). A, ventral view of terminal segment of male antenna, showing disposition of pit pegs (PPg), end pegs (EPg), supported on the styles (St), bristlelike hairs (Br) and ordinary hairs (Hr) in solid black, X 145; B, longitudinal section through a pit peg, X 570; C, two end pegs, supported on a style, X 220; D, section of a bristlelike hair, X 220; and E, longitudinal section of three ordinary hairs, one-half schematic. Br, bristlelike hair (Sensilla saetica) on antenna; EPG, end peg (S. styloconica) on antenna; Hr, hair, called S. trichodea on antenna; Hr\ pseudo-hair in pit peg; iV, 'nerve; PPg, pit peg (S. coeloconica) on antenna; SC, sense cell; SCG, sense cell group in antenna; St, style supporting end peg.

pit pegs and end pegs vary slightly in structure. The pit pegs sometime consist of compound pits instead of single pits, and the shape of the styles supporting the end pegs vary more or less.

Since the peripheral ends of the sense fibers are covered with the hard chitin forming the walls of all the antennal organs, it does not seem reasonable that the outside air carrying odoriferous particles can pass through the chitin in order to stimulate the nerves within.

54 N. E. McINDOO


The disposition of the olfactory pores of Lepidoptera is similar to that of Hymenoptera and Coleoptera, but resembles the former more closely. The structure of these pores is also similar to that of those found in other insects, but differs slightly in three following respects: The external ends of those in Lepidoptera are dome-shaped and the only other dome-shaped ones found by the writer are in the lady-beetle, Epilachna borealis. The external .boundary of the cones may be convex, almost flat, and concave, but hitherto it has always been convex. The sense cells are more spherical than usual.

Compared with the antennal organs, the olfactory pores are better adapted anatomically to receive olfactory stimuli, because the peripheral ends of their sense fibers come in direct contact with the external air, while those in the antennal organs are covered with hard chitin.


Freiling, Hans H. 1909 Duftorgane der weiblichen Schmetterlinge nebst Beitragen zur Kenntnis der Sinnesorgane auf dem Schmetterlingsflijgel und der Duftpinsel der Miinnchen von Danais und Euploea. Zeitsch. f. wiss. ZooL, Bd. 92.

GuENTHER, KoNRAD 1901 tjber Nervenendigungen auf dem Schmetterlingsfliigel. Zool. Jahrb., Anat., vol. 14.

Hauser, Gustav 1880 Physiologische und histologische Untersuchungen liber das Geruchsorgan der Insekten. Zeitsch. f. wiss. ZooL, Bd. 34.

Hicks, J. B. 1857 On a new organ in insects. Jour. Linn. Soc, London, ZooL, voL 1.

1859 On a new structure in the antennae of insects. Trans. Linn. Soc, London, vol. 22.

1860 On certain sensory organs in insects, hitherto undescribed. Jour. Linn. Socy., London, vol. 23.

McIndoo, N. E. 1914 a The olfactory sense of the honey bee. Jour. Exp. ZooL, vol. 16.

1914 b The olfactory sense of Hymenoptera. Proc. PhiUi. Acad. Nat. Sci., vol. 66.

1914 c The olfactory sense of insects. Smithsonian Misc. Collec, vol. 63, no. 9.

1915 The olfactory sense of Coleoptera. Biol. BuL. vol. 28.

1916 The sense organs on the mouth-parts of the honey bee. Smithsonian Misc. Collec, vol. 65, no. 14.

ScHENK, Otto 1902 Die antennalen Hautsinnesorgane einiger Lepidopteren und Hymenopteren, mit besonderer Beriicksichtigung der sexuellen Unterscheide. Zool. Jahrb., Bd. 17.



Cornell University, Ithaca, iV. }'.



1 . Int roduction 55

II. Material and methods 57

III. Observations and considerations 57

A. Morphological 57

a. Pieris rapae Linn 57

b. Neuronia postica Walker 61

c. Summary 63

B. Chemical 64

IV. Conclusions 66

Bibliography 67


The peculiarly shaped nucleus and the unusually large number of the nucleoh in the cells of silk-glands of certain insects have attracted the attention of previous workers on the histology of these organs, but detailed cytological vrork done in connection with the process of silk secretion is rather limited. . Marshall and Vorhies ('06) published a comparative study of the cell-structure of the silk-glands of Platyphylax designatus, a caddis-worm, under different physiological conditions; Vorhies ('08) demonstrated the fact that the multiple nucleoli are derived from a single nucleolus of the ordinary tj^pe in the nucleus of the same insect, and noted the increase of the nucleolar material with the growth of the nucleus. These two works have shown that the nucleoli may take some important part in the metabohsm of the silk-gland cell, but the exact r61e they


play has remained unsuggested until the appearance of Maziarski's ('11) paper, in which he stated that the nucleoli migrate in certain forms into the cell-body, and he regarded these as giving the material for the silk secretion.

Considering the cytological aspect of the phenomenon of secretion in general, the cases in which the secretion granules are derived from the nucleoli are very few, if any. Such authors as Garnier COO), for the saUvary gland of the rat, Maximow ('01), for the similar gland of the dog, claim that the migrated ' Nucleolekorper' are metabolized into secretion products, but this theory does not seem to be as yet satisfactorily established. The observations on the nucleolar migration by Carlier ('99, '06) in the stomach and liver cells of the newt, by Page and Walker ('08) in the mammalian nerve cells, by Walker and Embleton ('08) in the cells of Hydra, and by Walker and Tozer ('09) in vegetative cells of various plants and animals seem opposed to the view that the nucleolar material can be considered playing a role in the formation of the secretion substance. On the other hand, although not completely established, there is a considerable body 'of evidence that the chondriosomes, which are widely distributed in the cells of secretory nature, give rise to the secretion granules, as shown by Arnold ('05), Hoven ('10, '11), and Schultze ('11).

In view of the fact that the nucleoli are now generally consid'ered as passive by-products of the nuclear activity, and that, so far as the more reliable of previous observations indicate, the active role of the nucleoli in the formation of the secretion products is much to be doubted, it has seemed very advisable to study the phenomenon of the migration of this nuclear element in the silk-gland cells of some insects in greater detail.

The present paper gives a condensed account of my observations, both morphological and chemical, upon the subject. Before proceeding further, I wish to express my sincere thanks to Professor William A. Riley for his stimulating encouragement and valuable advice rendered me during the progress of this work.



As the material for this study, I used the larvae of the common cabbage butterfly, Pieris rapae Linn., and those of a large caddis-fly, Neuron ia postica Walker.

Usually the silk-glands were taken out by the decapitation method before being placed in the fixing agent; sometimes a dissection in normal salt solution was made and the silk-g)ands obtained by this method also proved to be as good material. Some other specimens, especially of very young stages, were fixed in toto. Flemming's chromo-aceto-osmic (strong formula) and Hermann's platino-aceto-osmic mixtures were used with good result for fixing the material. Gilson's mercuro-nitric also proved to be a very good fixative. Sections were cut from 3 to 5 micra thick, and the majority of them were stained on the slide with Ehrlich's triacid or Flemming's triple stains. A number of other stains were tried to bring out the changes in the staining reaction of the migrating nucleoli.


A. Morphological

a. Pieris rapae Linn (Plate 1). The silk-glands of Lepidoptera have been studied by Helm ('76), Van Liadth de Jeude ('78), Joseph ('80), Carnoy ('84), Blanc ('89), Gilson ('90), Tanaka ('11) and Ito ('15) from different standpoints. More cytological are the papers of Korschelt ('96, '97) and Meves ('97), and especially the interesting cytological paper on the process of silk-secretion by Maziarski ('11).

In his earlier work, Korschelt ('96) noticed two stainable materials within the nucleus, and called them macrosomes and microsomes. He described the former as large round, rregularly angular, or spindle-shaped bodies, and the latter as small particles which always stained darker than the former. The macrosomes, according to this view, represent the chromatin, and the microsomes are nothing but the nucleolar material.


However, Meves ('97), employing various stains, has shown that the microsomes are chromatin and the macrosomes are to be considered as nucleoh, and, although Korschelt ('97) adhered to his previous view in his later paper, the results of subsequent researches by Flemming ('97) and Henneguy ('04) seem to support the view of Meves.

Maziarski ('11) in the silk-gland cells of different lepidopterous larvae, observed that the nucleoli migrate from the nucleus into the cell-body, either as separate bodies or in the form of droplets of 'prosecret,' being dissolved within the nucleus and accumulated in vacuoles. This, although' he does not give any evidence except that this material and the silkfiber in the lumen of the gland show similar reaction to certain stains, he regards as the source of the silk-secretion.

I found the following conditions in the silk-glands of larvae of Pieris rapae. In the larva just hatched (figs. 1 and 2), the nuclei of the silk-glands are more or less round, showing almost no sign of branching. Such nuclei contains a number of the nucleoli and chromatin granules. The nucleoli are of varying sizes and some of them show such a shape as to suggest that they are in the process of division. They are scattered in the nucleus or sometimes grouped together into a loose mass. Not rarely one or two of them lie close to the border of the nucleus. The chromatin granules are imbedded in the linin reticulum. The nuclear membrane can easily be demonstrated in most cases. However, at places where the nucleoli are approaching closely to the border of the nucleus, it has been impossible for me to detect a nuclear membrane.

The cytoplasm is homogenous; but granules of different sizes are to be noted more or less surrounding the nuclei. While most of the granules are somewhat scattered, I find a few almost always occur at the outer border of the nucleus, or contiguous to the nuclear membrane, when this is demonstrable. Now, taking into account the facts of the close approach of the nucleoli to the border of the nucleus and of the partial disappearance of the nuclear membrane, we can conceive that a portion of the nucleoli migrate from the nucleus into the cell-body, forming


the granules we have just observed. The granules of the migrated nucleoh increase in amount as the de^'elopment advances, and, in a little older larva, the entire of the cytoplasmic area is filled with the granules (fig. 3).

In the larva a Httle older than those considered above (fig. 4), a mass of the secreted substance occurs in the lumen of the gland. It is apparent, therefore, that the gland is already functional at this stage. Here the nucleus begins to show its branching, and the amount of the chromatin and of the nucleolar material seem to have been increased. The process of the multipHcation of the nucleoh is apparently proceeding rapidly. Along the border of the nucleus and the cell-body are many masses of the nucleolar material, indicative of the rather rapid process of nueleoli migration occurring (fig. 5).

The cytoplasmic area shows some elongate vacuoles of varying sizes (fig. 4). Gilson ('90), working on the silk-glands of various lepidopterous larvae, found vacuoles present in the cellbody and also in the nucleus of the cell, and he interpreted them as containing secretory material. Matheson and Ruggles ('07) stated in their work on the siinilar glands of Apanteles glomeratus that "numerous vacuoles are present in the cytoplasm, becoming most abundant during the time of glandular activity," and that 'Hhe contents of the vacuoles remain unstained by any of the coloring agents used." Such vacuoles have been observed also in the silk-glands of caddis-worms, not only by Gilson ('96), but also by Marshall and Vorhies ('06), but the latter authors do not think that they contain secretory material. Their observations would show that the cytoplasm in the normal gland cell presents an appearance free from vacuoles," while after the activity of the gland for two and one-half hours, "a number of fairly large vacuoles are seen along, the outer margin of the cell, and such vacuoles persist in the cell-body of the gland which has been active for longer periods (at least up to 250 hours)."

It is rather inconceivable that the amount of the secretion material (supposed to be contained in the vacuoles) in the cell increases, instead of decreases, after the cell has discharged,


and that the normal gland cell, which must naturally be loaded preparatory to discharging, contains no 'secretion material.' These together with the fact that the vacuoles contain no stainable material seem to make Gilson's opinion untenable.

According to Maziarski's ('11) observations, at the height of the secretory activity of the cell, the nucleolar material is discharged from the nucleus, not only as separate bodies^ but also as accumulations in vacuoles in the form of droplets of 'prosecret,' being dissolved within the nucleus. I have gone over my slides with special attention to this point, but the result was negative. The appearances shown in his figures 13 to 27 do not occur in any of my slides, except those that apparently show artificial conditions. I am, therefore, rather doubtful as to whether the conditions observed by Maziarski, which led him to conclude that on certain occasions the nucleoli become dissolved within the nucleus, and carried by vacuoles to the cellbody, were normal. Even if they be normal conditions, we should think that they indicate phenomena that occur very rarely.

The granules of migrated nucleoli appear to be somewhat reduced in number, and some of them begin to show more or less elongated masses at this stage. These granules may naturally be supposed to hav6 something to do with the secretion of the cell, since Gilson's ' vacuole-therory' is to be discarded, and as far as my observations go, there are no other special granules to be detected in the cell-body. Maziarski ('11) claims that basophile granules, which he interprets as derived from chromatin, are also present in the cytoplasm. I have observed such granules only in the degenerating glands obtained from prepupae!

Maziarski ('11) considers the migrated nucleoli as the source of the secretion of the cell, because these and the silk materials in the lumen of the gland show similar color reaction. The fact of the increase in their amount preparatory to the discharging of the cell and its reduction as the cell discharges, also suggest that the granules derived from the nucleoli may be the material source of the silk-secretion.


As the stage advances, the nucleus shows more and more the feature of branching; its contents, both the nucleoU and the chromatin, are apparently being increased in amount with it. The migration of the nucleoli is to be demonstrated as occurring very frequently.

b. Neuronia postica, Walker (Plate 2). The silk-glands of caddis-worms have been discussed by Carnoy ('84), Gilson ('96) and Vorhies ('05) to a certain extent. It remained, however, for Marshall and Vorhies ('06) to attempt a minute comparative study of the gland-cells under different physiological conditions. Quoting from their conclusions, Marshall and Vorhies found that:

1. The even optical structure of the cytoplasm in the cells of most normal glands become decidedly changed after activity.

2. The activity of the gland causes the membrane on that surface of the nucleus Which lies nearest the outer margin of the gland to become irregular, the most noticeable feature of its irregularity being the pointed processes extending into the cytoplasm.

3. This nuclear membrane may often become indistinct.

4. The secretory activity does not cause the nuclei to become swollen.

5. No plasmasoma or other structure is formed in the nucleus during secretion.

6. As a result of activity the 'nucleolus' becomes very irregular in shape.

As has been said in the preceding section, there has been a considerable difference of opinions as to the nature of the two stainable materials in the nucleus of the silk-gland cells of insects, since Korschelt ('96, '97) insists that the larger granules are the chromatin, and nucleoli are represented by smaller particles, while Meves ('97), Flemming ('97), Henneguy ('04) and Marshall and Vorhies ('06) express opinions that are exactly opposite to that of Korschelt.

Vorhies' ('08) work on the development of the nucleus has finally settled the question in favor of the view that the nucleoli are represented by smaller particles, and at the same time, his studies enabled him to suggest that the nucleolar material bears a direct relation to the glandular activity," since he observed the


increase of the material with the growth of the nucleus. He, however, did not enter into a detailed discussion of the matter.

The nucleus contains many nucleoli and chromatin granules. The former are of different sizes, distributed in most cases, fairly evenly throughout the nuclear area and sometimes containing vacuoles. The chromatin granules are very fine, and he imbedded in the linin reticulum. As has already been said by Marshall and Vorhies, the changes occurring in the nucleus during the secretion are not great, and the nucleus retains the appearance here described for the entire history of the secretory activity of the gland. In the cytoplasmic area, however, some remarkable changes are noted to occur.

Figure 6 shows a cross section of a resting gland, i.e., the gland which has not yet discharged the secretion material, and figure 8 one which has apparently done some discharging. One will notice that the granules of different sizes fill the entire area of the cell-body in the resting gland, while in the case of the other one, the amount of the granules is much reduced.

The granules in question are apparently identical with those in the case of Pieris, and we can naturally expect them to be derived from the nucleolar material. Close examination of the nuclei with special reference to the nucleoli indicates that this supposition is not incorrect.

As has been observed by Marshall and Vorhies, the nucleus become very irregular in shape, especially on the side facing the outer margin of the cell. Not infrequently, the condition as shown in figure 9 is to be observed, in which the nuclear membrane apparently has disappeared at some part of the surface of the nucleus, and some of the nucleoli are apparently migrating into the cell-body.

Marshall and Vorhies described the appearance of small dark colored areas scattered irregularly in the cell-body, after glandular activity of five hours. They say that after twentyfour hours of activity, small and somewhat elongated masses of dark colored cytoplasm, running parallel to the vacuoles appeared, while in some cases the masses were also noticed along the inner border of the nucleus. Other very significant statements from their paper are the following :


In many places, the nuclear membrane is quite difficult to distinguish, being lost in the adjacent cytoplasm; this is especially true in the nuclei along whose outer border the dark cytoplasmic areas are numerous" (after one hundred and twenty hours of activity).

". . . . These (the areas of darkened cytoplasm) are here mostly in close proximity to the outer boundary of the nucleus, but exceptionally present along the inner border of the nucleus" (after an activity of two hundred and forty hours).

Their so-called "darkened areas of cytoplasm" are practically identical with what I consider as the nucleolar material extruded into the cell-body or remaining in the nucleus but ready to migrate. They attempted no explanation of the significance of the appearance of the "darkened areas of cytoplasm." If my interpretation be correct, there is rio doubt that ten years ago they observed the phenomenon of the migration of the nucleoli in the silk-gland cell of caddis-worms.

The case of the migration of the nucleoli in the silk-gland cells of caddis-worms supports the theory that the extruded nucleolar material is metabolized into the secretion products, since the granules, which are the migrated nucleoli, are in such large number as to render it rather improbable that they represent a degenerating product, and since, moreover, the quantitative relation of the granules to the different physiological conditions of the cell is such as to be naturally expected for the material prepared for secretion.

c. Summary. The foregoing observations on Pieris and Neuronia seem to justify the following statements as regard the morphological changes of the nucleoli in the silk-gland cells of the insects studied:

1. The nucleoli multiply by division of the preexisting ones, and they increase in amount as the gland becomes more functional.

2. Before the gland becomes functional, a portion of the nucleoli begins to migrate from the nucleus in considerable number into the cell-body.


3. The migration of the nucleoh is continued throughout the entire history of the functional cell.

4. The number of the nucleolar masses in the cell-body decreases after some amount of secretion has been discharged from the cell.

The facts enumerated above seem to lead to the conclusion that the nucleoli constitute at least a part of the source of the secretion products of the cell dealt with.

B. Chemical

Perhaps no one has paid more attention to the staining reaction of the nucleoli and the changes in the reaction after the material has passed out into the cell-body, than Walker and other students of his school.

In their paper on the migration of the nucleoli in the nerve cells of mammals, May and Walker ('08) stained the material with (A) basic fuchsin, followed by methylen blue and Unna's orange tannin, or (B) safranin, followed by metheJyn blue and Unna's orange tannin. In method A the nucleoli within the nucleus stained blue or violet. They stain purple or red as they pass through the opening in the nuclear membrane, bright red or pink as they come to lie definitely outside of the nucleus, and, as they travel away from the nuclear membrane, they are always stained pink or red. Using the second method, they observed that the nucleoli within the nucleus stain brilliant scarlet, become reddish orange as they migrate, and turn pale orange or yellow when they are completely extruded into the cell-body. "This suggests strongly," the authors say, "that some important chemical or physical change takes place in the nucleolus when it passes into the cytoplasm."

Walker and Embleton ('08) also made a similar observation on the nucleolus of the cells of Hydra, employing the same combinations of stains, as Page May and Walker did.

Working on the nucleoli of the vegetative cells of different plants and animals. Walker and Tozer ('09) have made more general statements as regard the staining reaction of the nucleoli as compared to that of other elements of the cell.



The contents of the nucleolus seem always less susceptible to the basic stain than is the chromatin. While in the nucleus, however, the nucleolar contents show a more basic reaction than the cytoplasm and this tendency remains for some time after it has left the nucleus. But in most cases the extruded nucleolus takes less of the basic and more of the acid stain, untU it is quite as deeply or even more deeply colored by it than is the surrounding cj-toplasm.

Studying the phenomenon of the migration of the nucleoli in the silk-gland cells of lepidopterous larvae, Maziarski paid no special attention as we might desire, to this sort of changes.

In order to throw some additional light upon this phase of the subject, I stained the secretions of the silk-glands obtained from the larvae of Neuronia postica and Pieris rapae, with different combinations of the stains. From this I got results similar in the main to those obtained by Walker and others in the case of various other cells.






Delafield's or Mayer's haematoxylin,

Salmon Black Red Red





Iron haematoxylin, orange-G

Safranin, licht grtjn

Dark orange Greenish


Delafield's or Mayer's haematoxylin,

Pale clay yellow Pink

Flemming's triple stain


Thus w^e see that:

(1) The nucleoli within the nucleus are stained more or less energetically by the acid as well as certain of the basic stains, while

(2) The migrated nucleoli always stain with acid stains but they have no or very little affinity, if any, for the basic stains.

In the field of the cellular chemistry, it has been ah*eady made known by Miescher, Kossel, Altmann, Hoppe-Seyler, etc., that the ' nucleins' form a series leading downward from the pure


nucleic acid according to the higher percentages of phosphorus and the lesser percentages of albumen contained in those compounds, and the fact that it is the nucleic acid that determines the staining of the nuclear substance is shown by Lilienfeld, Kossel, and others (Wilson, '00, Mann, '02, Jones, '14).

In 1893, Zacharias (later, Heidenhain, '94) showed that in staining the preparations of the nucleins containing different amounts of phosphorus, with alcoholic solution of acid-fuchsin and methyl green, the nucleic acid takes a pure green color, but that the nuclein poorer in phosphorus and that poorest in the same element stained bluish violet and pure red respectively.

Heidenhain ('94) applied this to the case of certain granules in the nucleus of the leucocytes and demonstrated that these granules may show different color reactions by combining with or giving off phosphorus, although they are all exactly aUke in morphological characters. The very interesting case of the changes in color reactions (as well as in sizes) of chromosomes in the eggs of Pristiurus, first described by Riickert ('93) has been beautifully explained by Wilson ('00) in similar manner.

Applying these principles to the case under discussion, and taking into account the interpretations of somewhat similar changes in chromatic bodies by the previous authors, we may say that the nucleoli originally contain some amount of phosphorus, but as they migrate into the cell-body, the phosphorus seems to be given off from their composition. The migrated nucleoli may, therefore, be considered as albuminous granules, almost or entirely free from phosphorus.

Considering the chemical composition of the silk-fiber, we see that this statement on the chemistry of the migrated nucleoli is perfectly acceptable as that of a constituent of the former, provided that there may be some other substance given off, in addition to the nucleoli, for the formation of the silk-fibers.


1. In the silk-gland cells of insects studied, a portion of the nucleoli migrates into the cell-body, and it forms at least a part of the secretion products of the cell.


This shows that although the nucleolus may originally be a passive by-product of the nuclear activity, it may also take an important part in the secretory activity of the cell in certain cases.

2. As the nucleoli migrate from the nucleus, they seem to give off phosphorus to form themselves one of the lowest members of the nuclein series.

This statement on the chemical change may hold true for the migrating nucleoli in different other cells.


Arnold, J. 1905 Uber Bau und Sekretion der Driisen der Froschhaut; zugleich

ein Beitrag zur Plasmosomen-Granulalehre. Arch. f. mikroskop.

Anat., 65. Blanc, L. 1889 Etude sur la secretion de la sole et la structure du brin et de

have dans le Bombyx mori. Rap. Lab. d'etud. soie, Lyon. Carlier, E. W. 1899 Changes that occur in some cells of the newt's stomach

during digestion. La Cellule, 16.

1905 Concerning the secretion of ferments by the liver cells and

some of the changes observable in them during digestion. La Cellule,

22. Carnoy, J. B. 1884 La biologic cellulaire. Etude comparee de la cellule dans

les deux regnes. Paris. Flemming, W. 1896 Zelle. Merkel u. Bonnet's Ergebn. d. Anat. u. Entw. 5. Garnier, C. 1900 Contribution a 1' etude de la structure et du fonctionnement

des cellules glandulaires sereuses. Du role de I'ergastoplasme dans

la secretion. Journ. I'Anat. et la Physiol., 36. GiLSON, G. 1890 Recherches sur les cellules secretantes. I. Lepidoptera. La

Cellule, 6.

1896 Recherches sur les cellules secretantes. II. Trichoptera. La

Cellule, 10. Heidenhain. 1894 Neue Untersuchungen iiber die Centralkorper und ihre

Beziehungen zum Kern- und Zellenprotoplasma. Arch. f. mikroskop.

Anat., 43. Helm, F. E. 1876 tjber die Spinndriisen der Lepidopteren. Zeit. f. wiss.

Zool., 26. Henneguy, L. 1904 Les Insectes. Paris.

HovEN, H. 1910 Contribution a I'etude du fonctionnement des cellules glandulaires. Du role du chrondriome dans la secretion. Anat. Anz., 37.

1911 Du role du chrondriome dans I'elaboration des produits de

secretion de la glande mammaire. Anat. Anz., 39. Ito, H. 1915 On the metamorphosis of the silk glands of Bombyx mori. Bull.

Imp. Tokyo Sericultural Coll., 1. Jones, W. 1914 Nucleic acids, their chemical properties and physiological

conduct. London, New York.


Joseph, G. 1880 Vorlaufige Mitteilungen liber Innervation und Entwicklung der Spinnorganen bei Insekten. Zool. Anz. 3.

KoRSCHELT, E. 1896 tJber die Struktur der Kerne in den Spinndrlisen der Raupen. Arch. f. mikroskop. Anat., 47.

1897 liber die Bau der Kerne in den Spinndrlisen der Raupen. Arch, f. mikroskop. Anat., 49.

VAN LiDTH DE Jeude, T. W. 1878 Zur Anatomie und Physiologie der Spinndrlisen der Seidenraupe. Zool. Anz., 1.

Mann, G. 1902 Physiological histology, method and theory. Oxford.

Marshall, W. S. and Vorhies, C. T. 1906 Cytological studies on the spinning glands of Platyphylax designatus Walker. Intern. Monatschr. f. Anat. u. Physiol., 23.

Matheson, R. and Ruggles, a. G. 1907 The structure of the silk-glands of Apanteles glomeratus L. Amer. Nat., 41.

Maximow, a. 1901 Beitrage zur Histologie und Physiologie der Speicheldrlisen. Arch. f. mikroskop. Anat., 58.

Maziarski, S. 1911 Recherches cytologiques sur les phenomenes secretoires dans les glandes filieres des larves des Lepidopteres. Arch. f. Zellforsch., 6.

Meves, F. 1897 Zur Struktur der Kerne in den Spinndrlisen der Raupe. Arch. f. mikroskop. Anat., 48.

Montgomery, T. H., Jr. 1899 Comparative cytological studies, with especial regard to the morphology of the nucleolus. Jour. Morph., 15.

Page, May W. and Walker, C. E. 1908 Note on the multiplication and migration of the nucleoli in the nerve cells of mammals. Quart. Journ. Exp. Phys., 1.

RiJCKERT, J. 1893 Zur Entwicklungsgeschichte des Selachereies. Anat. Anz., 7.

Schultze, O. 1911 tJber die Genese der Granula in den Driisenzellen. Anat. Anz., 38.

Tanaka, Y. 1911 Studies on the anatomy and physiology of the silk-producing insects. I. On the structure of the silk-gland and silk formation in Bombyx mori. Journ. Coll. Agr. Tohoku Imp. Univ., 4.

Vorhies, C. T. 1905 Habits and anatomy of the larva of the caddis-fly, Platyphylax designatus Walker. Trans. Wisconsin Acd. Sci. Arts, Letters, 15.

1908 The development of the nuclei of the spinning gland cells of Platyphylax designatus Walker. Biol. Bull., 15.

Walker, C. E. and Embleton, A. L. 1908 Observations on the nucleoli in the cells of Hydra fusca. Quart. Journ. Exp. Phys., 1.

Walker, C. E. and Tozer, F. M. 1909 Observations on the history and possible function of the neuleoli in the vegetative cells of various plants and animals. Quart. Journ. Exp. Phys., 2.

Wilson, E. B. 1900 The cell in development and inheritance. Revised ed. New York.




All the figures are drawn with camera lucida. Pieris rapae, L.

1 A cross section of a silk-gland of a larva just hatched. X 250.

2 A cross section of a silk-gland of a larva just hatched, a part of the cell. X 1200.

3 A cross section of a silk-gland of a little older larva. X 250.

4 A cross section of a silk-gland of a little older larva, representing the •condition after a moderate discharge of the secretion. X 250.

5 A part of the gland-cell of a fairly grown larva. X 1200.





'^^•ti..„ \



m ^,.*'




Neuronia postica, Walker

6 A cross section of a normal silk-gland of a young larva. X 250.

7 An enlarged portion of a cell shown in figure 6. X 1200.

8 A portion of a cross section of an active silk-gland of a full grown larva. X 250.

9 An enlarged portion of a cell shown in figure 8. X 1200.







. ^•.






Introduction 75

Materials and methods 78

Observations 79

1. Excretory system 79

Pronephros 79

Mesonephros 108

Bdellostoma ' 108

Myxine 118

2. Reproductive system 120

Myxine 120

Female 120

Male 126

Bdellostoma 128

Female 128

Male 130

Protandric hermaphroditism in Myxinoids 131

Summary 134

Bibliography 136


As is well known, Johannes Miiller was the first to investigate the urogenital system of the Myxinoids, describing briefly both male and female of Bdellostoma forsteri and the European Myxine glutinosa. Since his day much has been written regarding this subject. The investigators of the excretory organs include W. MuUer (75), Kirkaldy ('94), Semon ('96), Spengel ('97), and Maas ('97), all of whom worked upon the Myxine found in European waters.

Weldon ('84) and Price ('10) are the only ones who have pubHshed a description of the excretory apparatus of Bdellostoma.

^ Contribution from the Zoological Laboratory of the University of Illinois. No. 92.



The former investigated Bdellostoma forsteri, the latter Bdellostoma stouti. Both of these descriptions are confined almost entirely to the structm-e and probable function of the pronephros, and are comparatively incomplete. Scarcely anything is said about the mesonephros. Price ('96 a, '96 b, '97, '10) has, however, given a detailed account of the development of the excretory system in Bdellostoma stouti, which throws much light upon the adult condition. Semon ('96) devotes one paragraph to a comparison between the pronephroi of Mjrxine glutinosa and Bdellostoma bischofi.

After J. Mliller, Thompson ('59) and Streenstrup ('63) were the next writers upon the generative organs of the Myxinoids, each giving a short account of the mature ova of the European Myxine. W. MiiUer ('75) contributed many points regarding oogenesis in Myxine and gave the first description of the minute structure of the testis. In 1886 Cunningham added a few more points regarding the reproductive elements of Bdellostoma forsteri and in 1887 he published the first detailed description of the sexual organs and products of Myxine glutinosa, introducing the subject of protandric hermaphroditism. Nansen ('87) supplemented Cunningham's description with the results of further investigations, which led him to agree with Cunningham that Myxine glutinosa is a protandric hermaphrodite. Cunningham ('92) next contributed a rather lengthly discussion of spermatogenesis in Myxine. Ayers ('93) published the fii'st description of the reproductive system of Bdellostoma stouti, and soon after Dean ('99) gave an account of the structure of the eggs and development of the embryo. The latest articles which have appeared regarding this subject are those by A. and K. E. Schreiner ('04, '05, '08) which consist of a description of the generative elements of Myxine glutinosa and a very thorough discussion of spermatogenesis in Myxine glutinosa and Bdellostoma burgeri. Besides the foregoing investigators, there are several who have made minor contributions to our knowledge of the genital apparatus in Myxinoids.

The literature contains only a very meagre description of the mesonephros and the reproductive elements of Bdellostoma and


practically nothing regarding the urogenital system of the NorthAmerican Mjodne. The latter animal differs from the European Myxine glutinosa sufficiently to cause some writers to regard the two as different species. Girard ('58) found specific differences in the external aspect of the snout and buccal aperture, the insertion and proportional development of the tentacles, the form of the body, and .... the presence of a membraneous fin-hke expansion along the abdomen," and proposed the name Myxine limosa for the North- American species. Putnam ('73), who compared specimens obtained off the north-eastern Atlantic coast, from the straits of Magellan, and in the Museum of Comparative Zoology, collected off the Engfish coast, however, considers the American animal as merely a variety of Myxine glutinosa. The writer has been unable to obtain specimens of the European animal for comparison. Judging from the literature, there is a great difference in the sizes of the two animals. Putnam's specimens from Liverpool measured 10 1 to 12 inches long (26 to 30 cm.) . Nansen, using animals caught near Bergen, and Schi-einer, who caught hundreds in Drobaksfjord, reported the largest adults as 35 cm. long. From Cunningham's account it is to be inferred that the largest adults caught by him off the Enghsh coast and near Bergen were from 25 to 42 cm. long, most of them being from 33 to 35 cm., these containing the largest eggs. None of the investigators of the European Myxine report animals longer than 42 cm. The single North-American specimen used by Girard was 11^ inches long (about 29.25 cm.), and the one examined by Putnam was 12 inches (about 30.5 cm.). The writer has been unable to find in the literature any other measurements of the North-American Myxine. The adult animals used in the present study range from 50 to 79 cm. in length, averaging 62 cm., which is almost twice the length reported for the adult European Myxine. Three specimens, measuring in length 31, 33, and 36.5 cm., respectively, are in a very j^oung stage, the genital ridge being only 2 to 3 mm. wide along its entire course and the eggs being nothing more than dots. Size alone, of course, is not a specific character, but such a great difference as the above is suggestive.


In view of the fact that only two comparatively short descriptions of the pronephros of Bdellostoma have been published, and that the mesonephros of Bdellostoma and the entire urogenital system of the North-American Myxine are still undescribed, it seems desirable :

First, That the pronephros of Bdellostoma be re-examined, in order, if possible, to collect additional data which may illuminate some of the uncertain or disputed points in regard to its structure ;

Second, That the mesonephros of Bdellostoma be described more fully;

Third, That the entire excretory and reproductive systems of the North-American Myxine be described and compared with those of Bdellostoma and of the European Myxine, in order to ascertain whether they present any specific differences or add anything to our comprehension of the urogenital system in Myxinoids.


The Myxine used in this study consist of 20 specimens which Were caught at South Harpswell, Maine, during the summers of 1914 and 1915. Three of the animals are immature females, nos. 7, 17 and 18, measuring 36.5, 33, and 31' cm., respectively. All the other specimens are adults ranging from 50 to 79 cm. and averaging 62 cm. in length. Only one specimen is 50 cm. long and the largest eggs in it are 8 mm. long. The next longest specimen measures 55 cm. and its largest eggs are 22 mm. long.

Twenty-four specimens of Bdellostoma stouti, 9 males and 15 females, which were caught at Pacific Grove, California, comprise the Bdellostoma material. The smallest is an immature male 33 cm. long. The other males are adults, ranging from 36 to 42 cm. in length, and all the females are adults which range from 34 to 40 cm.

All the animals were killed and preserved in 10 per cent formahn, and later were transferred to 70 per cent alcohol.

The study of the pro- and mesonephros was made from sections stained in Delafield's haematoxyhn and eosin. Toto mounts of these organs, cleared in xylol and imbedded in damar, were also used.


The work was done in the Zoological Laboratories of the University of Ilhnois under the guidance of Dr. J. S. Kingsley. I wish to acknowledge my sense of obligation to Dr. Kingsley for his kindly interest and helpful suggestions, always ^^dllingly given.


The investigation of the excretory apparatus is based upon Bdellostoma stouti, and the results will be presented by describing the organs of this animal and by interjecting, from time to time, comparative notes regarding the same organs in the American Myxine. Unless, therefore, it is specifically stated otherwise, the content of the following discussion refers to Bdellostoma stouti.


The right pronephros lies in the right pericardial cavity immediately dorsal to the portal heart and along the dorso-lateral surface of the alimentary canal, parallel to the dorsal aorta. In figures 1 and 3, which illustrate the position of the right pronephros of Bdellostoma and MjTcine, respectively, in the pericardial cavity, the lateral wall of the cavity has been cut away, but the edges of the pericardium which form the pericardoperitoneal foramen are shown. The pronephros lies along a large diverticulum extending posteriorally from a vein which is called by Jackson ('01) the anterior portal and by Cole ('14) the right anterior cardinal vein. I have had no opportunity to examine injected specimens, but from a study of serial sections the writer agrees with Price ('10) that this vessel is merely a diverticulum with no posterior outlet. In Myxine glutinosa, however, Cole ('14) has shown that this vein extends posteriorly after leaving the pronephros and enters the supra-intestinal vein. In a single poorly injected specimen of the North- American Myxine the wTiter found a small twig extending posteriorly from this vein toward the supra-intestinal vein and a larger


branch going to and entering the right postcardinal vein. In Bdellostoma the vein varies in diameter from 0.35 by 0.40 mm. to 0.65 by 0.85 mm. at a point about the middle of the pronephros. The vein is always between the two folds of pericardium, by means of which the pronephros is attached to the dorsal aorta and to the wall of the alimentary tract (fig. 5).

The pericardium completely envelopes the pronephros. It encircles each tubule where the latter extends into the pericardial cavity. At the distal end of each tubule the pericardium is continuous with the columnar epithehum which lines the tubule. This enveloping pericardium extends farther down between the tubules in Bdellostoma than in Myxine.

In all the specimens examined the right pronephros was located in somites 31 and 32, counting the somite in which the eye is located as the first. In only one specimen, a young female, did the pronephros occupy but one somite. In all the other specimens it began in somite 31 and projected into somite 32.

The left pronephros lies in the left pericardial cavity immediately dorsal to the auricle and along the dorso-lateral surface of the alimentary tract, parallel to the dorsal aorta. The left pronephros, also, is connected to the aorta and to the alimentary tract by a fold of pericardium (fig. 5). The fine of attachment of this fold of pericardium embraces from one-fourth to one-half of the width, and about three-fourths of the length of the medianlateral surface of each pronephros. In all the specimens the left pronephros was slightly posterior to the right, usually lying in somites 32 and 33. Both right and left pronephroi lie about one millimeter to the right and left of the dorsal aorta, as shown in figures 45 and 46.

The left pronephros lies along a vein extending posteriorly from the left anterior cardinal. From his sections the writer could not trace this vein to any posterior connection in Bdellostoma, and concludes, mth Price ('10), that it is merely a diverticulum. In Myxine, however, the writer found that this vein continues posteriorly and enters the large vein formed by the union of the two posterior cardinals (fig. 4), the left ductus Cuvieri of Cole ('14)'. Like the right vein, this vein along the



left pronephros lies between the folds of pericardium which attach the pronephros to the dorsal aorta and to the digestive tract. Since no name has been given to these veins, hereafter in this discussion they will be referred to as the right and left pronephric veins. The position of the left pronephros in the pericardial cavity is represented in. figures 2 and 4.

In shape the pronephros is never exactly the same in any two specimens nor on both sides of the same specimen. In most cases, however, it is somewhat three-sided, though it is sometimes flat on the attached side only and the remaining surface is rounded. The ends always taper more or less, the anterior end usually being the more pointed. This is due to the fact that at its anterior end the pronephros always invariably consists of only one or two small lobules, while the posterior end is usually composed of several lobules crowded together in a compact mass. Figures 7 to 11 represent pronephroi taken at random and illustrate the variation that may occur in the shape. Figure 11a represents a cross section of a pronephros and illustrates the three-sided form. When three-sided the head-kidney is more or less wedge-shaped, flattened dorso-ventrally with the sharp edge of the wedge opposite the point of attachment. In this condition the pronephros has the appearance of having been subjected to pressure, probably due to the distended portal heart and auricle pushing it against the dorsal body-wall.

The size of the pronephros is also quite variable, as the following few measurements will indicate:













































2.0 1.0




2.0 0.5




It was observed that the right pronephros is usually slightly larger than the left. In none of the specimens was a pronephros found even one-half as large as those of Bdellostoma forsteri, as reported by Weldon ('84), namely 20 to 25 by 5 to 7 mm. There is no correlation between the size of the pronephros and the length of the animal in Bdellostoma. Weldon does not give the length of the animals examined by him, hence we do not know whether the great difference in the sizes of the pronephroi of Bdellostoma stouti and Bdellostoma forsteri may be explained by a difference in the length of the two animals. The adult Bdellostoma used in the present investigation range from 33 to 42 cm. long, averaging 38 cm. The pronephroi represented by the above table are all from adult animals.

The writer has been able to find only one statement regarding the size or shape of the pronephros in the adult European Myxine. W. Muller (75) mentions that they are 'langlich,' 3 to 4 mm. long. Following is a table of dimensions of a few pronephroi of the North-American Myxine:





































































































As is indicated by the above tables, the pronephroi of Myxine are considerably larger than those of Bdellostoma. This is to be expected from the great difference in the size of the adult ani


mals. The writer found that there is a correlation between the jSize of the pronephros and the length of the animal in Myxine, but not a sufficiently large number of specimens was available to prove this to be constant. The largest pronephroi (those of Myxine no. 10) are in an animal which has the appearance of being quite old. The skin is darker than usual, very thick, WTinkled and hard, the muscle fibers in the body w^all are large and coarse, the walls of the alimentary canal are very thick, the liver and gall bladder are very large and coarse. In this specimen the pronephroi are a very dark brown color; in all the other specimens and in all the Bdellostoma examined the pronephros is gray. It is probable that the size of the pronephroi is more correlated with the age than with the length of the animal. The right pronephros of Myxine was almost always larger than the left.

Price ('04) found that at some stage of embryonic development of Bdellostoma stouti an excretory tubule appears in each somite, from the 11th to 13th to the 79th to 82d. That is to say, in the embryo pronephric tubules are present in the same segments with gills. As the gills shift posteriorly (Dean, '97) these pronephric tubules are pushed along behind them and eventually become crowded together in a small, compact body, the pronephros of the adult, in the region a little posterior to the thirtieth segment" (Price, p. 132). Price does not mention anything in regard to the history of the veins connected with the tubules in this movement posteriorly.

As w^ould be expected, there is a limitless variety in the minute structure of the pronephroi. It is well known that each pronephros is composed of a large number of small tufts or lobules of tubules which open separately into the pericardial cavity. Usually there are fewer of these lobules at the anterior than at the posterior end, and one or two lobules may be entirely separated anteriorly from the rest of the pronephros. Figure 12 represents a reconstruction on millimeter paper of the right pronephros of Bdellostoma no. 15. This pronephros was cut in transverse sections, each 10 microns thick, and it will be used as a


basis for the description of the minute structure of the headkidney of Bdellostoma, as follows :

Histology. The tubules in each of the two most anterior lobules anastomose inwardly and form a single tubule which unites with one similarly formed from the other lobule, and the singles tube thus formed constitutes the central duct of the pronephros. The first appearance of the central duct occurs in the 10th section from the anterior end of the pronephros. All of the pronephroi of Bdellostoma stouti examined by the writer have a central duct, which is so evident that it can not possibly be interpreted as anything else. Price ('10) found a central duct in the head-kidney of Bdellostoma stouti, and Weldon ('84) described one for Bdellostoma forsteri. Semon ('96) thought that Weldon was mistaking the venous sinus (pronephric vein) for a central duct, but Weldon's description of the duct coincides so nearly with the condition in Bdellostoma stouti that it seems very probable that Bdellostoma forsteri has a central duct in the pronephros. In the 10th section from the anterior end of the pronephros the lumen of the duct is 0.02 mm. wide, but t enlarges immediately and is 0.15 mm. at the widest part. The duct proceeds posteriorly, entering the pronephric vein in section 20 and ending blindly in the vein in section 48. From its beginning this short segment of the central duct is surrounded by a mass of large, deeply stained nuclei, which will be referred to hereafter in this discussion as the 'central mass' and which will be described in detail later. This mass is distributed somewhat regularly around the central duct, now more on one side, now more on the other. Some idea of its gross appearance may be had from figure 12. At its widest part the mass measures 0.21 mm. from its outer edge to the lumen of the duct, and it extends 0.14 mm. beyond the posterior end of the duct.

No trace of the central duct is again seen until section 52 is reached. Here the tubules of one lobule unite into a single tube, and the latter forms the beginning of the central duct. The lumen of this segment of the duct is 0.05 mm. in diameter at first, but measures 0.09 mm. at the widest part. The duct proceeds posteriorly to section 60, receiving in its course the tube


from another lobule. In section 60 the duct enters the pronephric vein, and remains in it to section 80 where it ends blindly. From its beginning this segment of the duct is also surrounded by a central mass like the one around the first segment of the duct. In transverse section the mass is circular in outline and is distributed somewhat regularly around the lumen of the duct. At its widest part the mass measures 0.16 mm. from its outer surface to the lumen of the duct. Only two lobules of the pronephros are connected with this segment of the central duct.

The third segment of the duct appears in section 90, and is formed by the main tube of one lobule. In this section the lumen of the duct is 0.05 mm. in diameter, but becomes as wide as 0.20 mm. This segment of the duct extends to section 150 and there breaks up to form the tubules of three lobules. Altogether there are four lobules connected with this segment. The duct runs parallel to the pronephric vein, and up to section 115 is separated from the latter by a wide layer of connective tissue. In this distance one vein 1 mm. wide and another 3 mm. wide permit blood to flow into the pronephric vein from capillaries which run between the tubules of the pronephros. At section 115 the duct widens abruptly and extends downward in the pronephric vein. In the vein the duct becomes sheathed in a central mass which mbles in all respects the masses which surround the first and second segments of the duct. In section 117 the duct opens through the central mass into the lumen of the pronephric vein. This opening is 0.05 mm. wide. Another such opening 0.09 mm. wide occurs in section 250, and a third 0.04 mm. wide in section 320. These openings are not breaks or tears, but are natural apertures lined by a single layer of squamous epithelium which is continuous with the columnar epithelium of the duct on the one hand and with the endothelium of the pronephric vein on the other. Figure 20 represents one of these openings. At least one, often more, such opening connecting the lumen of the central duct with the lumen of the pronephric vein was found in every pronephros of Bdellostoma examined. These openings occur only through the central mass, never through the walls of the central duct when not surrounded by the central mass.


To return to the description, the duct remains in the central mass to section 137 where it leaves the vein and breaks up into the tubules of three lobules. The central mass, however, continues in the lumen of the vein independent of the duct, and in section 147 the fourth segment of the central duct appears in it. The central mass projects into the pronephric vein as indicated in figure 12 and has a cylindrical shape. Between sections 137 and 147 it is 0.30 by 0.40 mm. in transverse section.

The fourth segment of the central duct begins as the result of the anastomoses of the tubules of four lobules whose main tubes unite near section 150 and form the central duct. In this section the duct is 0.05 mm. wide. At first the duct lies in the connective tissue at the base of the pronephros, but in section 157 it bends toward and enters the central mass in the vein. Here the lumen of the duct -is 0.27 mm. wide. In this segment one side of the duct remains closely apphed to the connective tissue at the base of the pronephros and retains its columnar lining, while the other side extends along the central mass to section 200 where it bends toward its fellow, thus narrowing the lumen of the duct considerably, and the segment ends by the duct branching into the tubules of a single lobule. There are eleven lobules connected with this segment of the duct. The central mass continues without interruption in the pronephric vein to section 240 where the fifth segment of the duct enters it. That is to say, the central mass extends from section 200 to section 240 (a distance of 0.40 mm.) entirely disconnected from the central duct. In the region of the fourth segment the central mass resembles in every respect that around the preceding segments of the duct. In the portion between sections 200 and 240, however, the mass changes its shape in transverse section from circular to elongated, measuring 0.15 by 0.55 mm. in the largest portion. Figure 20 represents the shape of the mass at this place. This figure also indicates the relation of the size of the mass to the lumen of the vein. In this particular pronephros the duct and the mass together in the largest parts never occupy more than about one-half of the lumen of the vein as viewed in transverse section. In one specimen, however, the central duct was


SO large that it, together with the central mass, almost entirely obstructs the lumen of the vein. Between sections 205 and 230 the pronephric vein opens directly between the tubules as a large sinus (fig. 20), and another such sinus lies between sections 255 and 300. At all other places, however, there is a strip of connective tissue from 0.01 to 0.05 mm. wide, or even wider, between the pronephros and the vein. In one specimen, though, the vein opens directly into the spaces between tubules along the entire base of the pronephros. But even here the pronephros stands upon the vein as illustrated in figure 20, so that in no sense can the head-kidney be said to he buried within the vein, a misleading statement which occurs very frequently in the Uterature. Where the vein opens between the tubules as a sinus, the capillaries, which otherwise carry the blood between the tubules, have broken down and only the pericardial sheath and the endothelium of the sinus confine the blood in the intertubular space. These vascular spaces between the tubules are' not 'sinusoids' according to Minot's use of the term. The pronephric tubules do not invade the vein, for, as Price ('10) has shown, the branching to form new tubules always occurs in a direction away from the pronephric vein, just back of the nephrostomes." The fifth segment of the central duct begins in section 212 where the tubules of one large lobule unite into a single tube which widens to form the duct. The latter extends posteriorly in the connective tissue at the base of the pronephros parallel to the pronephric vein to section 240, increasing gradually in width from 0.04 mm. to 0.15 mm. In this section the duct bends downward so that one side enters the central mass, which still remains in the vein. The wall of the duct which lies toward the pronephros retains its columnar structure, but the wall which lies in the central mass breaks down completely soon after entering the mass and presents a jagged appearance, and the nuclei bordering the lumen have no definite arrangement. In sections 243 to 255 th6 central duct opens through the central mass into the lumen of the vein. The mass is similar in all respects to that surrounding the preceding segments of the duct. In section 260 the duct divides, one branch entering a lobule and breaking up


into tubules, the other continuing posteriorly in the central mass and ending blindly in it at section 274. In section 276 the duct begins again as a part of the sixth segment, just opposite the blind ending of the fifth segment. There are six lobules connected with the fifth segment of the duct.

The sixth segment of the duct arises in section 273 from the union of the tubules of one lobule. The duct enters the central mass in the vein in section 300, and is here 0.12 mm. wide. A small twig extends anteriorly in the mass to section 276 where it ends blindly and is separated by only one section from the end of the fifth segment of the duct. Undoubtedly at one time the fifth and sixth segments of the duct were connected at this point. The sixth segment extends posteriorly through the central mass to section 380 where it ends as a single tubule which runs out of the vein and opens into the pericardial cavity. In sections 318 to 323 the duct opens through the central mass into the pronephric vein, and in this region the duct and neighboring tubules contain many blood corpuscles. In section 350 the duct leaves the central mass and the columnar cells of the walls of the duct become double their usual length and are thrown into longitudinal folds like those of the mesonephric duct, to be described later. This folded condition of the epithehal walls of the duct continues to section 370; then the cells diminish in length to the normal size for the tubules, and the duct here ends as a single tubule, as explained. There are six lobules connected with the sixth segment of the duct.

The seventh segment of the central duct begins in the central mass in section 340 beside the sixth segment. It continues in the mass, which still lies in the pronephric vein, to section 365 where it leaves the mass and vein and divides into two branches, one of which ends by branching into the tubules of a lobule and two single tubules, the other breaks up into the tubules of the four most posterior lobules of the pronephros. The lobules connected with this segment of the duct are in such a compact mass that they can not be counted with any degree of certainty.

The eighth and last segment of the central duct arises in the central mass in section 370 beside the se\'enth segment. The


mass continues around the duct to section 380 and there ends, remaining in the lumen of the vein. The duct proceeds posteriorly, and in section 385 divides into three tubules, each of which leaves the pronephric vein and enters the Bowman's capsule of the Malpighian body of the pronephros, opening into the capsule by a nephrostome. There are no lobules connected with the eighth segment of the duct.

Price ('14) suggests that perhaps each lobule of the pronephros represents a single original tubule and its secondary branches. This is scarcely probable, for there is a total of 31 lobules connected with the first six segments of the duct in the above described pronephros, while those of the seventh segment could not be counted with accuracy but there are at least four. If the most anterior original tubule lay in somite 11, then there could be only 21 tubules at most to and including somite 32, the definitive position of the tubules after they have been pushed posteriorly by the gills. In the pronephros under discussion, however, there are at least 35 lobules.

In Bdellostoma none of the tubules in any of the pronephroi examined by the writer show any signs of degeneration, and the lumen of each tubule opens into the pericardial .cavity at one end and is continuous, either directly or indirectly, with the lumen of the central duct. The lumen of almost every tubule is filled with a coarsely granular coagulum which is more or less shredded into long processes resembhng flagella (fig. 16). These processes are attached at one end to the columnar cells, usually one to a cell, and the free end extends inward toward the base of the tubule, that is, toward the central duct. The lumen of the central duct sometimes contains a comparatively small amount of this coagulum. The writer was unable to demonstrate cilia in the phonephric tubules, but it is probable that his material was not suitably preserved to permit these to be seen if present.

The tubules are very simple in structure. Each is a cyhnder consisting of a single layer of columnar cells, which are continuous at the mouth with the pericardium and, at the base, with the columnar cells of the central duct. The mouths of the tubules are sometimes funnel-shaped (fig. 13), but are usually constricted


(fig. 14), The columnar cells which constitute the wall of the tubules are approximately the same size in all the tubules, varying somewhat around 0.0135 mm. high by 0.004 mm. wide. The nucleus is always at the base of the cell and occupies the entire width and from one-half to two-thirds its length. At the base of the cells the cell-walls fuse to form a thin basal membrane, the outer limit of the tubule. When connective tissue occupies the space between the tubules, it lies next to the basal membrane, and when a blood sinus fills the inter-tubular space, the endothelium of the sinus usually lies against the basal membrane, completely surrounding the tubule (fig. 15). The lumen is practically the same in diameter in all of the tubules, and the outside diameter of the tubules is likewise approximately the same for all tubules.

The specimens of Bdellostoma examined were within a few centimeters of being the same length, and the dimensions of the tubules are about the same in all the pronephroi examined. At the mouth of the tubule the lumen narrows slightly, but otherwise is approximately the same diameter throughout. Often, however, a tubule is constricted in one or two places. The lumen is approximately 0.0216 mm. wide, and the outside diameter of the tubule is about 0.054 mm.

The protoplasm of the columnar cells of each tubule is coarsely granular, and the boundary of the cell facing the lumen is usually marked by many granules and one or two long processes which consist of granules (fig. 16). The nuclei are more or less oval and always contain many large deeply stained bodies. The cells of the tubules are continuous with those of the central duct at the proximal end, and with those of the pericardium at the distal end.

The cells of the central duct are also columnar and are usually longer than those of the tubules. The character of the cytoplasm and nuclei in the duct cells is practically the same as in the tubules, and the basal membrane of the latter is continuous with the basal membrane of the former. The cells of the duct in a given pronephros vary in length at different points, but any increase or decrease is gradual. They may be as long as 0.0324 mm. in


the midregion of the pronephros. The diameter of the lumen of the duct varies widely in an individual specimen. It may be as small as a tubule in places, then increase to several times that width. The duct and its lumen may be continuous throughout the entire length of the pronephros or may be broken up into disconnected segments. The most continuous duct was found in a young male. The duct in this specimen also has the least amount of central mass connected with it, and the mass is almost entirely confined to the posterior end of the duct.

The structure which has been referred to as the 'central mass' was first described by Kirkaldy ('94) for Myxine, and the ^\Titer has adopted his name for it. It is possible that Weldon's lymphatic tissue may have been the same thing as this central mass, but his short description does not enable us to be positive. This mass of tissue has been the subject of much discussion by investigators of the pronephros of Myxinoids. Weldon ('84) described a mass of lymphatic tissue which he found at the posterior end of the head-kidney of Bdellostoma forsteri. He states that the central duct ends posteriorly " in a mass of tissue .... resembling the trabecular supporting tissue of a lymphatic gland." Into this mass strands of blood vessels pass from the glomerulus which lay beside it. Upon the basis of this lymphatic tissue Weldon thinks the pronephros becomes transferred into and functions as a suprarenal body. Kirkaldy found a difference in the mass, in animals without ova and those with them. In the former the mass was divided into separate parts, and each part was regarded as a glomerulus supplied with a capsule, inside of which was a characteristic loop (his fig. 2). In an animal with ova he found no central duct or glomerulus in the pronephric vein, but their former position was occupied by a mass of small cells with small nuclei and larger cells with large, round, and deeply stained nuclei (his fig. 7). Capillaries were numerous in this mass. Kirkaldy regards the mass as the degenerating central duct of earher stages, hence Myxinoids may be considered as representing a stage in the phylogenetic reduction of the head-kidney, and the latter may represent the mesoblastic part of the supra-renal bodies.


Semon ('97) regards this mass as the glomeruU of the headkidney. Eventually, however, the pronephros becomes transformed into a suprarenal body. He found the mass rich in blood vessels.

Spengel ('97) thinks this mass was never glomerular, but that it is either a suprarenal body or a lymph organ, and that it is the result of a metamorphosis of the inner ends of the pronephric tubules.

Maas ('97) thinks the mass differs distinctly from a glomerulus. Blood vessels occupy only a little space in it, and it is greatly dissimilar to tubule epithehum. It resembles stages in the development of the Miillerian duct of higher animals, especially reptiles. It is also not unlike the medulla of suprarenals.

Price ('10) evidently saw this mass, but does not describe it. He states that '4n places the wall of the duct may become greatly thickened by an increase in the number of epithelial cells. These change their shape, and become much more loosely arranged, so that the tissue loses entirely the structure of columnar epithelium." Price adds that these thickenings are not suppUed with arteries.

In specimens with a duct which is continuous throughout the entire length of the pronephros, the central mass appears only in the posterior third of the pronephros where it surrounds the duct and lies in the pronephiic vein. In most of the specimens, however, the duct is broken up into segments which are not connected with each other, and that part of each segment which lies in the vein is surrounded by the central mass. The mass is found only in the pronephric vein (a very small quantity may be connected with the duct immediately before it enters the vein). A typical distribution of the mass is represented by figure 12. It may project into the vein like a glomus, one side remaining in contact with the base of the pronephros, or it may lie in the lumen of the vein, completely surrounded by blood, and may be attached to the base of the pronephros at one end only. The central mass is found only in connection with the central duct and is always attached to the latter, though one end of the mass ma}^ extend in the j^ronephric vein some distance (as much as


40 sections, 0.4 mm.) beyond the end of the lumen of the duct. The mass is more or less cyhndrical, though in places it may be considerably widened as if bent upon itself, and then it is elongated in transverse section. In none of the specimens does it entirely fill the lumen of the vein. In those pronephroi in which the duct is abnost continuous, however, the lumen is so great in diameter, that the duct, together with the central mass surrounding it, almost entirely fills the lumen of the vein, leaving only a very narrow space between the mass and the vascular endothelium. The endothehum of the pronephric vein is continuous around the mass (figs. 17 and 22), but is here much thinner than when Uning the vein, and its nuclei can be observed only now and then. The surface of the mass is smooth and convoluted.

Figure 18 represents a transverse section through the central mass, and illustrates the typical condition in Bdellostoma. It is composed of a large number of cells, the nuclei of which are more or less oval in shape. The latter are of various sizes, the largest and most numerous averaging 0.0054 by 0.0108 mm. These nuclei are not distributed in any regular manner, large, small and intermediate sizes are mingled in all parts of the mass. Also, there is no regularity in the manner of their arrangement ; some are isolated completely from surrounding neighbors, others are collected in groups of two or three, and again six or eight may be closely crowded together in a mass. There is no definite cellular structure observable in the central mass. Most of the nuclei contain many large, deeply stained granules, but many have only a nuclear membrane 'V'vdth no granules inside. Many of the nuclei have no cytoplasm surrounding them, while others are imbedded in what resembles the granular cytoplasm in the cells of the central duct and tubules. This granular cytoplasmic ground-work is scattered throughout the central mass and, in places, is drawn out into slender processes which resemble those found in the tubules. There is no connective tissue in the central mass. In an occasional transverse section of the mass there are one to three or four giant nuclei, but these are not frequent in Bdellostoma. About one hundred sections of the mass were examined before a section containing one of these giant nuclei


was found. Occasionally such nuclei occur in the walls of the duct or of the tubules near the central mass. These giant nuclei are much more deeply stained than are the other nuclei of the mass, and the cytoplasm, which always surrounds them, stains hke that of blood corpuscles, but much more deeply. These giant cells are interpreted as greatly enlarged blood corpuscles. They do not have the elongated characteristic shape of the blood corpuscles, but are usually rounded and quite irregular in outhne. In the central mass they are surrounded by a very thin membrane which is interpreted as endotheUum, although no nuclei were observed in it. In the walls of the duct or tubules, however, nuclei occur in the membrane surrounding the giant blood corpuscles and here it is undoubtedly endothelium. These giant cells are also frequently seen in capillaries which occur in the connective tissue surrounding the mesonephric duct, and here there is no doubt that they are in blood vessels.

Excepting the giant corpuscles just described, the writer rarely found a blood corpuscle or blood vessel in the central mass. An attempt was made to find blood vessels entering the mass, but without success. The blood corpuscles have a distinctly characteristic shape and appearance, and the cytoplasm stains more deeply in eosin than that of any of the other cells in the pronephros, hence the corpuscles are easily discerned in any tissue. There are many spaces in the central mass, but they are not lined by endothelium and never contain blood corpuscles. Furthermore, the writer found no vessels leaving the central mass to enter the vein or any surrounding tissue. Also, natural appertures in the endotheUal wall of the mass were diligently but vainly sought ; the mass is completely shut off from the blood in the vein. One series of sections is especially well stained to demonstrate this fact. The blood corpuscles are stained a deep, yellowish-red, which is strikingly different from the pale blue of the central mass. The blood corpuscles surround the latter as a dense mass which almost fills the lumen of the vein. On account of the striking contrast in stain, a single blood corpuscle inside the central mass would be instantly recognized if present. But


none was found. Price ('10), also, found no blood vessels in the central mass in Bdellostonia.

When not surrounded by the central mass, the central duct retains its columnar walls, but loses them upon entering the mass. The walls break down completely and the cells become scattered in the mass. The lumen of the duct may continue in the mass, but it has no organized lining and is bordered by nuclei and strands of cytoplasmic ground-work (fig. 29). The columnar lining of the central duct is never continuous with the epithelial sheath around the central mass. In view of this fact, it is difficult to regard, as Semon ('96) does, the central mass as a series of glomeruU into whose Bowman's capsules the tubules of the pronephros enter through their Tnnentrichter,' for Semon ('96, '97) himself has shown that, in the mesonephros, the columnar wall of the tubule of a Alalpighian body is continuous with the squamosal epithelial wall of the Bowman's capsule. Furthermore, in Bdellostoma, the writer never found a tubule entering the central mass. The tubules enter the central duct and the latter alone enters the central mass, hence in Bdellostoma Semon's Tnnentrichter' of the pronephric tubules really open into the central duct. Therefore, the inner ends of the pronephric tubules cannot correspond to that end of the tubule of a mesonephi'ic Malpighian body which opens into the Bowman's capsule, and hence the central mass of the pronephros cannot correspond to the glomerulus of a Malpighian body of the mesonephros.

In every pronephros of Bdellostoma examined by the wTiter the central mass has one or more openings, through which the central duct communicates with the lumen of the pronephric vein (figs. 12, 20, 21, 22). These openings are natural, and the endothelial sheath which surrounds the central mass lines the sides of the openings. In the lumen of the duct near these openings, or even projecting into them, there is almost always a rounded collection of nuclei and granules, which is more deeply stained than the central mass. The nuclei and granules do not resemble blood corpuscles, but have the appearance of waste particles. These openings do not occur in the central duct at


any place where the latter is not surrounded by the central mass, nor do they occur in all those segments of the duct which are ensheathed by the central mass. There is no regularity in the location of these openings, except that they are always confined to the posterior half of the pronephros. As is shown by figures 20, 21 and 22, by means of these openings the lumen of the pronephric vein is in actual communication with the pericardial cavity through the central duct and the tubules. Near the openings blood corpuscles are numerous in the lumen of the central duct, and they extend far up the tubules toward the openings of the latter into the pericardial cavity. In one specimen blood plasma extends from the central duct two-thirds of the entire length of four or five tubules. The pericardial cavity communicates with the peritoneal cavity through the large pericardo-peritoneal foramen, and the peritoneal ca\dty opens into the cloaca through the genital pore, hence we have in Myxinoids the strange condition of the vascular system being open to the exterior of the body. The writer found no traces of blood in the pericardial cavity. Since there are many hundreds of tubules in each adult pronephros, it is difficult to explain why blood is not poured into the pericardial cavity. Price ('10) found that carmine grains injected into the peritoneal cavity through the genital pore of living animals were present later in abundance in the blood taken from different parts of the body. He presumed that the carmine grains gained admission to the blood stream through the ciliated pronephric tubules, that is to say any current in the tubules which is caused by cihary action is from the exterior toward the interior. This inference is corroborated by the fact that, in almost all the tubules examined by the writer, the free ends of the granular processes extending from the epithelial cells of the tubules are directed inward toward the central duct.

In his description of the dievelopment of the pronephros Price ('04, p. 137) states that the manner in which the central duct becomes shortened as the tubules are crowded together is a point which has not been worked out; bending of the duct will not account for all of it. Kirkaldy ('94, p. 356) thinks the duct breaks down and becomes the central mass. Morphological evidence


supports this hypothesis somewhat. The diffused condition of the cells of the duct where the latter is in contact with the central mass suggests that the duct is breaking down, and the enlarged nuclei in the mass, which are without granules, are suggestive of disintegrating cells, while the rounded masses of small nuclei and granules found in the lumen of the duct may be interpreted as nuclei which have been broken down. These may either be ejected into the bod}^ cavity thi'ough the tubules and pericardial cavity, or, what is more probable, they may be swept into the blood stream and be engulfed by leucocytes. These masses occur only in or near the openings in the central mass which connect the central duct with the pronephric vein. In one specimen, however, a small ball of these nuclei was observed in a tubule. Furthermore, in those pronephroi in which the central duct is most continuous there is no central mass in the anterior part of the duct where its columnar walls remain entire, but the mass is limited to the posterior region where the walls of the duct have lost their columnar structure and the lumen of the duct is surrounded by the mass of nuclei. Also in those specimens in which the duct is broken up into segments and the palisade wall has broken down, each segment, anterior as well as posterior, is surrounded by the central mass, but even in these specimens the bulk of the central mass is found at the posterior end of the pronephros, where the crowding of the original tubules is the greatest. On the other hand, the quantity of the central mass in the older animals seems greater than would be the case if it were all derived from merely a crowding-together of the central duct, especially if it were disintegrating all the while.

The central mass is the same in appearance and structure wherever it occurs in Bdellostoma stouti. No lymphatic tissue such as Weldon ('84) describes is present at the posterior end of the pronephros of Bdellostoma stouti, unless he refers to the central mass around the duct.

The pronephros of Myxine differs strikingly from that of Bdellostoma in the fact that it contains much more connective tissue. The pericardial sheath around the tubules is farther from the wall of the tubule than in Bdellostoma, and the intervening



space is usually filled with connective tissue. Also, there is much more connective tissue at the base of the pronephros, separating the latter from the pronephric vein, than in Bdellostoma, and bands of connective tissue even extend into the vein and into the central mass. In the older animals there is often considerable connective tissue between the walls of the tubule and the surrounding sheath of pericardium (figs. 23 and 24). The pericardial sheath is heavier and its nuclei larger than in Bdellostoma. Figure 25, a cross section of one lobe of a pronephros of an adult female, shows the proportion of connective tissue with reference to the size of the tubules which it surrounds. Not all of the tubules in the pronephroi of this animal, however, have connective tissue between their walls and the pericardial sheath. In the youngest specimen examined, a young female 31 cm. long (no. 18), there is very little, and often no connective tissue at all, near the distal ends of the tubules. In the oldest specimen (Myxine no. 10) the connective tissue around the tubules and in the entire pronephros is more abundant than in any of the other animals and contains many blood vessels. In this specimen there is, in transverse section of almost every large tubule, a single large vessel which partially or entirely surrounds it (fig. 26). A blood sinus, which completely fills the space between the tubules, as is frequent in Bdellostoma, is rare in Myxine. The blood around the tubules is almost always confined in vessels which lie in the connective tissue surrounding the tubules.

Unhke Bdellostoma, the tubules in an adult Myxine are not all of the same size (fig. 27) . They are practically of the same size in the youngest specimen examined (Myxine no. 18), but in the older animals there is a gradually increasing number of tubules which are slightly larger than others, and in the oldest specimen (Myxine no. 10) there are present, in the pronephros, tubules which range in size from the smallest to very large ones. In Myxine no. 18 the average dimensions of the tubules are: Width of lumen 0.0162 imn., width of each wall 0.0189, outside diameter 0.054 mm. Only a few of this size are present in the pronephros of Myxine no. 10, many approximating the following dimensions: Width of lumen 0.108 mm., width of each wall 0.027, out


side diameter 0.1620 mm. One tubule in Myxine no. 10 is verylarge, the outside diameter being 0.675 by 1.1475 mm. before it breaks up into smaller branches. The smaller tubules have practically no connective tissue between the walls and the pericardial sheath, the medium size tubules a httle, while the large ones usually have a comparatively large quantity of connective tissue around them. The smaller tubules have the appearance of being recent branches from the larger ones.

Kirkaldy ('94) states that, in a Myxine with large eggs, the tubules at the posterior end of the pronephros are entirely different from those of a Myxine without eggs in that the former are considerably degenerated. The wTiter is of the opinion that this degeneration of the tubules is not so much a matter of being with or without eggs as one of age. Kirkaldy does not state whether the animal without eggs was young or old. In all the pronephroi of Myxine examined by the writer the inner ends of ahnost all the tubules have disintegrated, only occasionally is one found whose walls remain intact to its opening at the inner end. Only in the oldest specimen (Myxine no. 10), which does not have any eggs, are there tubules which have the appearance of degeneration. Some of the largest tubules alone are degenerating, and answer to the description given by Kirkaldy. The nuclei are enlarged, here and there, in the walls, while some of them are attenuated and almost fibrous, and connective tissue is replacing the palisade cells. These degenerating tubules are not numerous and are not hmited to the posterior end of the pronephros. They are surrounded by many smaller tubules which give no evidence whatever of degeneration, but on the other hand look like recently formed branches. The cells of the latter are more uniform in size and shape and stain more deeply than those of the large tubules.

One of the most striking differences between the pronephroi of Bdellestoma and Myxine is that in the latter the disintegrating process of the interior has proceeded further than in the former. In Bdellostoma the inner ends of the tubules are always entire, while in Myxine they are seldom so. In places only dense lines of nuclei indicate where the inner ends of tubules have been, and


these lines are always streaming toward a more or less large, loosely arranged mass of nuclei, bordering or lying in the pronephric vein.

Investigators do not agree as to the existence of a central duct in the pronephros of Myxine. W. Miiller and Kirkaldy affirm, while Semon, Spengel and Maas deny the presence of the duct. After studying the pronephros of Bdellostoma, it is inevitable for one to conclude that a central duct at one time existed and that fragments of it still exist in Myxine. It is not at all prominent and might easily be overlooked except in a comparative study. Along the entire extent of the pronephros the inner ends of the tubules, as stated above, end as a loosely arranged mass of nuclei in which a cavity frequently appears (fig. 28). This cavity is the remains of the lumen of the central duct. In occasional places it opens into the lumen of the pronephric vein, and blood corpuscles occur in it and in the tubules near it (fig. 29). Here and there the central duct is cut in longitudinal section and is then unmistakable (fig. 30). The youngest specimen (Myxine no. 18) shows the duct but imperfectly (fig. 31). Likewise, in the oldest animal examined (no. 10) the duct has almost entirely disappeared, the lumen being visible for only a short distance in the posterior part of the pronephros and having but a small piece of columnar wall (fig. 32) . Extending along the wall of the vein next to the pronephros is a large number of heavily stained nuclei, which are interpreted as remains of the central mass; or, in other words, . remains of the central duct and the inner ends of the tubules. Waste particles, nuclei, and granules are occasionally seen in the lumen of the duct and in the tubules. The tubules are filled with a coagulum which is more or less drawn out into granular processes, and these are almost always directed inwards. In only a few places do the walls of the duct still remain entire (fig. 33). Where the duct has entirely disappeared, the tubules often border directly on the vein and their inner ends are degenerated.

The central mass in the pronephros of Myxine is decidedly different from that in Bdellostoma. In the young, immature animals (Myxine no. 18) it approaches in structure and appearance


the condition existing in Bdellostoma (fig. 34). Here the mass contains the more or less oval nuclei, no blood corpuscles, but the cytoplasmic groundwork is replaced by connective tissue. Also, one side of the mass projects into the lumen of the pronephric vein and has a more or less rounded form, while the other side remains attached to the base of the pronephros, but is not as compact as in Bdellostoma. The mass is distributed along the entire length of the vein, from the anterior to the posterior end of the pronephros, and extends behind the posterior end of the latter about 0.18 mm. in one specimen. This posterior extension of the mass is cylindrical, has a tubule which ends blindly in it, and contains several large blood corpuscles in capillaries.

In the older animals the central mass is much more unlike that of Bdellostoma. In the anterior part of the pronephros it consists of a loose collection of nuclei which extends along the base of the tubules, embracing their inner ends, either lying outside the vein or projecting slightly into it. The mass is not at all compact, and has no definite form. Many of the nuclei are distributed in an irregular manner along the sides of the vein. At the base of the pronephros the mass is more or less exposed to the blood in the vein, even though it does not project into the lumen of the latter. In occasional places the lumen of the vein opens through the mass (fig. 32) and communicates with the lumen of the central duct, and here blood corpuscles are present in the duct, but are rarely found in the tubules. In the posterior part of the pronephros the central mass becomes cylindrical and compact, and projects into and extends posteriorly in the lumen of the vein, completely surrounded by blood. It is bound to the sides of the vein by broad bands of connective tissue which extends from the side of the mass here and there (fig. 35). Blood vessels run from the connective tissue at the base of the pronephros into these bands of connective tissue. The fibers of the latter continue to the center of the cyhndrical mass, and enlarged nuclei and blood capillaries are arranged in a cellular structure around this central core of fibers. The giant blood corpuscles are rounded in shape and are contained in capillaries. They are very numerous in the central mass in Alyxine (as many as fifteen in one sec


tion taken at random) and average 0.0081 mm. in diameter. Besides the giant corpuscles, there are many large nuclei in the central mass which are stained about the same as the nuclei of the tubules and which contain very many small granules. The giant blood corpuscles resemble those in Bdellostoma, but, in the latter animal, they are not nearly as numerous as in Myxine. In this posterior region of the pronephros of Myxine the central mass resembles a lymphatic structure.

Another striking difference between the pronephros of Myxine and that of Bdellostoma is the presence of broad bands of connective tissue which extend across the lumen of the pronephric vein from side to side, and partially or completely divide the lumen longitudinally in two parts (fig. 36). These bands are present in all the pronephroi of Myxine, from the youngest to the oldest, and are not confined to any particular region of the pronephros. In one place, near the Malpighian body of the pronephros, a large, isolated, thick-walled duct begins blindly in such a band, continues in it for 16 sections and ends bhndly. This duct has, on one side, the columnar cells characteristic of the central duct, on the other the appearance of a Malpighian body. From its posterior end a small tubule is given off which extends alongside the duct for five sections then ends blindly. The entire structure measures 0.24 mm. long by 0.195 nma. wide. In another Myxine a similar duct arises in one of these bands of connective tissue, extends posteriorly in it for 24 sections (0.36 mm.) and ends blindly at the side of a structure represented in figure 38, without connecting with a tubule which extends from the posterior end of the latter. These structures in the bands of connective tissue are interpreted as degenerating Malpighian bodies. Bowman's capsule is still present, and the tubules are connected with this, but the glomerulus is almost entirely gone.

Malpighian body of the pronephros. At least one Malpighian body is always found lying beside the posterior end of the pronephros of Bdellostoma and Myxine. Investigators of the headkidney of Myxinoids differ as to what part of the excretory system this Malpighian body belongs. W. Miiller's figure 2 shows two of them in the pronephros. He failed to state directly


whether he thought they belong to the pronephros or to the mesonephros. Kirkaldy seems to consider the Malpighian body as belonging to the head-kidney. Semon ('96) thinks it is the first Malpighian body of the mesonephros. Spengel ('97) is of the opinion that this is, in fact, the glomus of the pronephi'os in Myxine. From his investigation of very young Myxine Maas ('97) concludes that it is the glomus of the pronephros formed by the 'concentriren' of the most posterior 'Gefassnetze' which surround the original segmental pronephric tubules. Price ('04, '10) regards this Malpighian body in Bdellostoma as belonging to the 'pronephros,' and from embryological study concludes that it arises by the fusion of glomeruh which are formed in connection with some of the most posterior original tubules of the pronephros before they are crowded together by the posterior movement of the gills. Since, in the embryo, only two or three of the original tubules which take part in the formation of the head-kidney have glomeruh, then the definitive pronephros represents the fusion of not more than two or three glomeruli.

The adult condition of the Malpighian body in Myxine and Bdellostoma supports the conclusions of Maas and Price in regard to the manner in which it is formed. The pronephros of both shows evidence of being the result of fusion of two or three glomeruh. The following description is for both Bdellostoma and Mjodne, and any variations which either shows will be noted.

The Malpighian body is always located beside the posterior half of the pronephros, usually at the extreme posterior end. It is always imbedded in the connective tissue beside the pronephric vein and is separated from the pronephros by the sheath of its Bowman's capsule and the pericardium, as well as by a more or less wide strip of the pericardial cavity (fig. 39). Quite frequently the Malpighian body is followed immediately by a second, the anterior end of which may lie beside the last few tubules of the pronephros (fig. 12). Ordinarily, however, this second Malpighian body is located back of the posterior limits of the pronephros. The capsules of the two Malpighian bodies shown in figure 12 are connected by a very narrow duct, the short columnar cells of which do not have as much cytoplasm as those of the


pronephric tubules and they absorb much stain. The second Malpighian body is interpreted as being the first Malpighian body of the mesonephros. It is connected by a tubule to a short piece of the segmental mesonephric duct which lies in the lumen of the pronephric vein. The vein ends blindly with this piece of the mesonephric duct a few sections posterior to the Malpighian body. Furthermore, two pronephric tubules, which open into the capsule of the Malpighian body of the pronephros, extend into the pronephric vein, unite and enter the piece of mesonephric duct. Surrounding the base of these tubules, just before they enter the mesonephric duct, is a small amount of the central mass. This short piece of duct contains waste particles throughout its entire extent. The continuous duct of the mesonephros, in this specimen, begins 5 mm. posterior to this short piece and there is no connection whatever between them. When there is but one Malpighian body it usually ends blindly, although the capsule may be drawn out posteriorly into the very small duct which ends in the connective tissue.

In shape the Malpighian body of the pronephros of both Bdellostoma and Myxine is usually an elongated oval, although it may be almost round or flattened somewhat laterally. In Myxine one end is, as a rule, more pointed than the other.

The glomerulus of the Malpighian body in Bdellostoma is comparatively compact and uniform in structure. The surface is smooth but shghtly convoluted, and is covered by the epithelial lining of Bowman's capsule. There are lines of division which separate the glomerulus here and there into lobes (fig. 40), and in almost every specimen are one or more small cavities, inside of which are blood corpuscles or blood plasma.

In appearance the glomerulus of the youngest Myxine (no. 18) is very similar to that of Bdellostoma. In the older specimens, however, the lobed condition is more apparent, it is much less compact and has a more or less shriveled appearance, and the spaces in the glomerular mass are more numerous and larger. Webs and strands of connective tissue appear everywhere in the glomerulus and around the outside of the lining of Bowman's capsule. In the oldest Myxine (no. 10) these changes are most


advanced (fig. 41). As this figure shows, one large blood sinus has appeared in the glomerulus. This sinus contains corpuscles and blood plasma, and extends almost the entire length of the glomerulus. It has an epithelial lining, outside of which are several concentric layers of connective tissue. On the opposite side of the glomerulus there is a large space with comparatively few nuclei, but entirely filled with slender fibers of connective tissue. This space also extends almost the entire length of the glomerulus. The lining of Bowman's capsule is an epithelium surrounded by a broad band of concentric layers of connective tissue.

In one Bdellostoma and one young Myxine the glomerulus of the pronephros is double ; two distinct glomerular masses are contained in the same Bowman's capsule. In the Bdellostoma the capsule is constricted to one-half its width at the point of union between the two glomerular masses, while in the young Myxine there is but a slight constriction in one side of the capsule (fig. 42).

In all the pronephroi of both animals, one or more pronephric tubules open into the cavity of Bowman's capsule of the Malpighian body. The capsule in every specimen of Bdellostoma is connected with the pericardial cavity by a more or less long, very narrow duct (approximately one-half as wide as a pronephric tubule). In addition to this duct there may be one or more very small openings through the wall of the capsule which connect the cavity of the latter with the pericardial cavity. The' capsule of the specimen represented by figure 12 has seven such openings. In none of the specimens of Bdellostoma are these openings into the pericardial cavity greater than two or three one-hundredths of a millimeter in diameter.

Spengel ('97) considers the glomerulus of the pronephros to be really a glomus because it hangs freely in a cavity which he found communicated with the pericardial cavity through a very large aperture; in one specimen the opening extended through nineteen sections each 30 microns thick, or 0.57 mm. Semon ('97) never saw in his preparations a communication as wide as Spengel described. Maas ('97) also considers the glomerulus a


true glomus. In each of his youngest Myxine he found a very large opening connecting the capsule with the pericardial cavity; it extended through almost as many sections as the glomus itself. But in advanced stages, by a folding of the epithehal wall, a capsule is formed around the glomus, and only a slit-like communication with the pericardial cavity is left, and even this may be completely closed. The glomerulus of the youngest Mjocine examined by the writer (no. 18) lies directly exposed to the pericardial cavity through an opening 0.156 by 0.192 mm. and resembles a glomus (fig. 43). The opening in Myxine no. 15, an adult, measures 0.0195 by 0.105 mm., and that of Myxine no. 10, the oldest specimen, is 0.195 by 0.120 mm.

The glomerulus of the head-kidney is not always as large in comparison with the glomeruli of the mesonephros as would be expected if it were formed by the fusion of two or three glomeruli, as the following measurements will indicate; see opposite page.

The figures 1, 2, 3, 4 in the column under 'Glomerulus' represent the number of the glomerulus, counting that of the headkidney as 1 ; glomeruh 2, 3, and 4 are the first, second and third glomeruli, respectively, of the mesonephros immediately following the glomerulus of the pronephros. All the glomeruli which lie posterior to the glomerulus of the pronephros are considered as belonging to the mesonephros. The glomeruli of the latter, posterior to the most anterior three or four gradually diminish in size. None of the anterior glomeruli of the mesonephros of Myxine were sectioned, therefore no measm'ements of these are given in the above table. As the table shows, the glomerulus of the pronephros does not always exceed or even equal in size some of the most anterior glomeruU of the mesonephros.

As has already been noted, the Malpighian body of the pronephros of one Bdellostoma (fig. 12) is connected by tubules to a fragment of the segmental duct of the mesonephros, which is entirely disconnected from the continuous part of the duct. No other case of a communication between the Malpighian body of the pronephros and the duct of the mesonephros was found in either Bdellostoma or Myxine. In a young male Bdellostoma (no. 22), however, the mesonephric duct on each side continues






Bdellostoma no. 4.

Bdellostoma no.

Bdellostoma no. 10.

Bdellostoma no. 15.

Bdellostoma no. 16

Myxine no. 10

Myxine no. 11

Myxine no. 15

Myxine no. 18









Right Left Left Left


0.46 0.34 0.60 0.48

0.50 0.39 0.49 0.47







0.41 0.27 0.2925 0.41

0.39 0.40 0.40 0.65

0.4485 0.39 0.40 0.53

0.41 0.2925 0.2925 0.39

1.2675 0.8775 0.6825 0.4095

0.78 0.75 1.04 0.92

0.95 0.65 1.03 0.785

0.66 0.60 0.66

0.675 0.705 0.4875

0.60 0.2925 0.405 0.60

1.10 0.30 0.50 0.65

0.88 0.40 0.60 0.84

0.60 0.45 0.45 0.51

1.305 0.975 0.84 0.348


to the posterior end of the pronephros, as is shown in figures 1 and 2. When the anterior ends or the ducts of this animal were sectioned, it was found that the lumen of the right duct ends beside the posterior end of the Bowman's capsule of the Malpighian body of the pronephros, but there is no communication between them. Some of the sections of the left duct were scraped off the slide in this region, hence the duct could not be followed. Immediately posterior to the pronephros the lumen of the right duct of this animal abruptly enlarges, becoming a cavity which measures 0.4875 by 0.6240 by 1.335 mm. This cavity is almost entirely filled with a rounded, loose mass of nuclei which resembles a disintegrating glomerulus. This mass is connected here and there to the lining of the duct by narrow strands of nuclei, but no blood vessels could be seen in them. The lining of the duct here does not resemble that of a Bowman's capsule, and is composed of short columnar cells hke those of the pronephric. tubules. If this be a disintegrating glomerulus, it is difficult to explain why it should be in the lumen of the mesonephric duct. Immediately posterior to this enlarged cavity the duct becomes narrowed to a diameter of 0.117 mm.


Bdellostoma. Mesonephric ducts. Except in specimen 22, the mesonephric ducts of the Bdellostomae examined begin from 2 to 10 mm. back of the posterior end of the pronephros. Small strands of tissue, resembling in structure the outer wall of the duct, extend from the anterior end of the duct toward the pronephros. Sometimes a lumen is present in one of these strands for a varying, but short, distance. These strands may be the only trace of pro- or mesonephric elements between the posterior limit of the pronephros and the anterior end of the ducts, or, in addition to these, there may be one or more isolated tubules or traces of glomeruli in this intermediate region.

The anterior ends of the right and left ducts in an individual are seldom opposite each other. The duct does not start abruptly, but is always narrowest at the anterior end, and presents the


appearance of having at one time extended farther forward. There is no uniformity in the diameter of the duct at its anterior end, except that it is always smallest here, varying in width from a mere thread to 1 mm.

The left duct usually bends abruptly laterad at its anterior end, until it assumes a position close beside the left postcardinal vein, and retains this position to the end of the body cavity. The right ducts bends only sUghtly laterad at its anterior end to assume a similar position beside the right postcardinal vein. The right duct is invariably closer to the dorsal aorta than the left.

In adults of the same size, the ducts of the males are larger and longer than those of the females. There is not much variation in the size of the two ducts of a given animal, nor in the ducts of all the adult animals of the same sex. Each duct is smallest at its anterior end, and gradually increases to its largest size at the posterior end of the body cavity, where it is from 1.5 to 2 mm. wdde in females and from 3.5 to 4 mm. in males.

Each duct is flattened dorso-ventrally throughout its entire length, being most flattened at the posterior end of the body cavity. The outer surface of the ducts is smooth. When a duct is stained, cleared and examined under the microscope the surface appears longitudinally striated, due to the occurrence of ridges formed in the lumen of the duct by the columnar epithelial Uning.

The ducts lie immediately ventral to the notochord, separated from the latter by a small cavity which extends the entire length of the ducts. By cutting the peritoneum along the dorsal body wall at the sides of each duct, the ducts, together with the postcardinals and the dorsal aorta, can easily be removed.

The mesonephric ducts of the adult male are unlike those of the adult female in some respects. Those of the former are not only larger, but are longer, by reason of lateral bendings or convolutions. Figure 45, a dorsal view, represents the typical appearance of the male ducts, and figure 46 those of the female. At their anterior ends the ducts of the male are only slightly convoluted, but, beginning about ten somites posterior to the ante


rior ends, the convolutions become quite pronounced and regular. There is but one convolution in each somite; the ducts bend laterad at the interseptal lines and mediad between these Unes. The convolutions continue to the posterior end of the body cavity, but in the last five or six somites they diminish in extent. The greater size of the male duct is caused principally by the fact that its lumen is larger than that of the female. The walls of the male duct, however, are somewhat coarser than those of the female.

No such convolutions occur in the ducts of the females. The ducts bend laterad more or less slightly at the interseptal lines, but there is no suggestion of the convoluted condition of the male ducts.

At their posterior ends, the ducts of both males and females have the same gross structure and appearance. The two ducts leave the peritoneal cavity at its posterior end, approach each other and continue a short distance (5 to 10 mm.) along the dorsal surface of the genital chamber, then converge sharply. At the point of convergence the ducts bend ventrad and laterad, and become abruptly narrowed to very small tubes (figs. 47 and 48). The latter continue posteriorly, bend sharply ventrad at the posterior end of the cloaca, and open in this region of the cloaca on a papillary enlargement on a prominent ridge which extends along the dorsal wall of the genital chamber and cloaca. The two openings lie alongside, but one is usually slightly posterior to the other.

The ducts of young males (fig. 49) are not convoluted at any place, and resemble in appearance and gross structure those of adult females.

Histology. Figures 50 and 51 are camera lucida drawings of transverse sections of a male and female duct, respectively, taken from the mid-region of the body cavity. The section of the male duct was selected where there is the smallest possible effect on the internal structure because of convolution. As the figures show, the ducts consist of an inner epithelium and an outer envelope of connective tissue.



The epithelium is composed of colmiinar cells, and is arranged in a series of longitudinal folds or ridges. Wax models show that these ridges branch and anastomose freely, but are always longitudinal and the branching is dichotomous. The two branches may run parallel to each other for a short distance and then unite, or may unite with another ridge, or may end without union to other ridges. A ridge may be only a fraction of a millimeter long and be entirely unattached to any other ridges. When free, the ends of the ridges rise gradually from the normal height of the epithelium.

The ridges are present in all parts of the duct, from the anterior to the posterior end, but at the former end, where the duct is small, they may be only two or three in number (fig. 52) . As the duct becomes larger, the ridges increase in number, and are most numerous at the posterior end where the duct reaches its maximum width. The ridges vary in number in a transverse section in corresponding regions in different adult specimens, as is indicated by the following table:





Bdellostoma no. 15, female

Bdellostoma no. 17, female

Bdellostoma no. 16, male. .

0.546 X 1.0725 0.351 X 1.365 0.780 X 2.4375

20 29



0.0780 0.390

There is considerable variation in the height of the ridges in the same transverse section and in different parts of the same duct. Figures 50 and 51 show the variation in the same section. For a given duct the ridges are lower at the ends than in the mid-region.

The columnar epitheUal cells which form the walls of the duct are shortest between the ridges. The nucleus is always located about one-third the length of the cell from the basal membrane. The cytoplasm is very granular throughout the entire cell body, the granules being especially numerous and large at the distal ends of the cells. Along the surface of the distal ends of the cells are agglutinations of granules which project into the lumen of the duct and resemble the mass of waste material which is


seen in the center of the lumen of the duct throughout its entire length. Toward the distal ends of the cells are numerous small and large, yellowish, round or oval bodies of homogeneous structure, which resemble oil droplets. Some of these are minute granules, while others are as large as the nucleus of the cell. They are especially numerous and large in the cells on top of the ridges, and are distributed in no definite manner. Only a few are found in the cells between ridges. These bodies are probably what Yl. Miiller described as yellowish pigment granules. The largier bodies do not resemble pigment granules, however.

The long, columnar cells are always arranged on the ridges in the shape of a fan, as illustrated in figure 53. To form a ridge, the entire layer of epithelial cells bends into the lumen of the duct, and the connective tissue outside the basal membrane of the epithelium makes a core which fills the concavity beneath each ridge (fig. 53).

The ridges are more numerous in the adult male ducts than in the adult female. They are also higher, but the cells are no longer than those of the female. The greater height is due to the fact that the epithelial layer extends farther into the lumen of the duct in the males. Wide ridges are more numerous in the male than in the female, some of them being flat on top. The concavities formed by the infoldings of the epithehal layers are larger in the male than in the female, and are also filled by a core of connective tissue.

Maas ('97) has suggested that the mesonephric ducts are not simply excretory, but they may have a secretory function as well. The distribution of the blood vessels, the presence of ciha, the folded epithelium, the enlarged condition of the ducts in comparison to the small tubules of the Malpighian bodies of the mesonephros have been mentioned in the literature as indicative of such a possibihty. The many large foldings or ridges of epithelium certainly increase the epithelial surface far beyond what would seem necessary for excretion alone. The enlarged, convoluted condition of the ducts in the males is very peculiar. The presence of the small and very large, yellowish bodies near


the distal ends of the epithelial cells, especially in the large cells of the ridges, is suggestive of secretion.

The envelope of connective tissue surrounding the duct consists of two parts, a loose web containing blood vessels and a more or less compact stratified layer. The loose web lies next to the basal membrane of the epithelium. It follows closely the outline of this membrane, and fills the concavities under the ridges, as shown in figure 53. Some of the blood vessels in this connective tissue contain giant blood corpuscles, like those found in the central mass.

The compact stratified layer consists of fibers arranged in concentric rings around the duct, and is usually thicker along the median surface of the duct. The peritoneum Ues next to the stratified layer.

The writer was unable to identify in Bdellostoma the three layers which W. Muller describes as forming the connective tissue envelope around the duct of Myxine. Nowhere is there a layer that can be called the 'membrana propria.' The two layers observed by* the writer correspond to the ' adventitia' and the layer which contains blood vessels, as designated by Muller.

Malpighian bodies. Price has shown that at one time in the development of Bdellostoma the mesonephric duct is continuous with the central duct of the pronephros, but that, later, a short piece of the duct posterior to the pronephros degenerates, together with its Malpighian bodies. The amount which degenerates is not always the same, hence in the adult there is considerable variety in the structure of the anterior ends of the ducts and in the location of the first few Malpighian bodies of the mesonephros. The most anterior Malpighian body of the Mesonephros may be located immediately posterior to the Malpighian body of the headkidney and may be attached to the latter and to a short, isolated piece of the mesonephric duct (fig. 12) ; or it may be completely isolated in the space between pro- and mesonephros. Not only one, but two or three of the mesonephric Malpighian bodies may be in the intermediate space, entirely disconnected from each other and from the mesonephric



duct. Again, the first Malpighian body of the mesonephi-os may be located at the anterior end of the duct, and its tubule may be widened to form the beginning of the duct, or the latter may extend slightly anterior to the Malpighian body, in which case the tubule enters the duct at the side. Two or three Malpighian bodies may be crowded close together around the anterior end of the duct, but their tubules always enter the lumen of the duct separately and in consecutive order (fig. 54, a drawing of the anterior ends of the ducts of Bdellostoma 4). On the left side Malpighian bodies 1, 2, and 3 are very close together. The first two are in one somite, 3 occupying the next segment. When the Malpighian bodies are thus crowded together at the anterior end of the duct, the arteries which supply them arise from the aorta at successive points close to the posterior end of the pronephros, a condition which suggests that these Malpighian bodies have been pushed back from a more anterior location. For example, the point of origin of the artery which supplies the first Malpighian body of the right duct of Bdellostoma no. 15 is 3.5 mm. anterior to the Malpighian* body, and the latter is only 5 mm. posterior to the caudal end of the pronephros.

Except where the Malpighian bodies are occasionally crowded together at the anterior end of the duct, there is but one in each somite to somite 60-65. The ducts usually begin in somite 33-35, and the body cavity extends to somite 75-79, hence the posterior third of the ducts have no Malpighian bodies. Price learned that in embryos excretory tubules are present in the most posterior somites, but they later degenerate in the last 19 or 20 segments. The most posterior Malpighian bodies in the adult show signs of degeneration. They are small, often not visible to the unaided eye, frequently in a shriveled condition, and parts of them may be lacking ; for example, in one case the tubule only is "present. This degeneration is of unequal extent in different animals and on both sides of the same animal.

Except an occasional one at the anterior end, all the Malpighian bodies lie along the median side of the duct. There is no regularity in the orientation of the bodies with reference to the somites; they may be located at either anterior or posterior side


or near the middle of the somite. Those of a, side are not separated from each other by equal intervals, but, as a rule, are arranged in pairs, a greater distance occurring between two pairs than between the two bodies of a pair (figs. 55 and 56). This paired condition is in some way connected with the arterial supply, as is shown by figure 56. Each of the Malpighian bodies represented in this figure is in a separate somite. The two arteries which supply the two Malpighian bodies of one of these pairs are usually branches from, a somatic artery, as shown by figure 56, but sometimes one of the arteries arises directly from the dorsal aorta, close beside the somatic artery. When a Malpighian body is not one of a pair, its artery comes directly from the dorsal aorta. Seldom are the two Malpighian bodies of corresponding somites on each side of the body opposite each other, as is shown by figure 55.

The Malpighian bodies and their tubules usually extend anteriorly to the point where the tubule empties into the mesonephric duct. When, however, each Malpighian body in two adjacent somites is supplied by a branch from the same somatic artery, the anterior one is doubled back upon its tubule so that the distal end points posteriad toward the origin of the artery which supphes it.

Each Malpighian body has a tubule which enters the mesonephiic duct. There are two types of these tubules which differ distinctly and which will be designated as 'plain' and 'compound' tubules. All the tubules of a given animal are of one type only.

The plain tubules are composed of very short, columnar cells which have very little cytoplasm, as shown in figure 57. These cells are all approximately the same size. Their nuclei stain much more deeply than either the nuclei of the connective tissue or those of the mesonephric duct, hence the tubules are very distinctly differentiated from surrounding tissues. At the proximal end the cells of the tubule are directly continuous with those of the mesonephric duct, and at the distal end with the cells of the single layer of squamous epithelium which lines BowTiian's capsule. The basal membrane of the mesonephric


duct is continuous with that of the tubule. The tubules vary in length, the shortest averaging 0.07 mm., the longest 0.2 mm. \\Tien long, they are more constricted than when short, the outside diameter of the narrowest tubule measured being 0.04 mm. The size of the opening where the tubules enter the duct varies, but approximates 0.12 mm. in diameter. Figure 58 is a diagrammatic reconstruction of a plain tubule.

The compound tubule is strikingly different from the plain, as is shown by figure 59, which is a reconstruction of a Malpighian body with a tubule of this type. It consists of two parts, a neck and a trunk. All of the Malpighian bodies of all the specimens examined, except the one described above, have this type of tubule. The neck is that part of the tubule which joins the Bowman's capsule, and in structure and appearance it is exactly like the plain tubules. It consists of short, columnar cells of approximately equal height and which have very little cytoplasm. The nuclei stain much more deeply than those of surrounding tissues, so that the epithelium of the neck is easily distinguished from that of the trunk. The cells of the neck gradually decrease in height as they approach Bowman's capsule, and are continuous with the cells of the single layer of squamous epithelium which Hnes the capsule. The neck is variable in width and length within narrow limits in different specimens and in different tubules of the same specimen. An idea of the dimensions may be had from one specimen : Outside diameter at entrance to trunk of tubule, 0.35 mm., and at entrance to capsule, 0.078 mm.; length, 0.4 mm. The neck is sometimes as much as twice this length, however. At the proximal end of the neck the epithelial cells and basal membrane are continuous with the cells and basal membrane, respectively, of the trunk of the tubule. The transition from the cells of the neck to those of the trunk is gradual, as shown in figm'e 60.

The trunk of the compound tubule is exactly like the mesonephric duct in structure. It consists of an epithelium of high columnar cells which are arranged in ridges, and cells and ridges are of approximately the same height as those of the duct. At the anterior end of the duct this portion of the tubule is often as large in diameter as the duct itself. Farther posteriorly, how



ever, where the duct is larger, the trunk of the tubules is only from one-fourth to one-thii'd as large as the duct. At the opening of the trunk of the tubule into the duct, the epithelial cells of both are continuous with each other, including their basal membranes (fig. 61). When the trunk of the tubule lies close beside the duct, each has its own envelope of connective tissue, but the two are bound together by a compact band of connective tissue arranged in concentric layers. The trunk of the compound tubules varies somewhat in length, but is not more than a fraction of a millimeter, approximating 0.75 mm. The trunk and neck together are about 1.5 mm. long. In some cases the neck may be longer than the trunk, but it is usually much shorter.

Waste granules resembling those in the duct are present in large quantity in the lumen of both the plain and compound tubules along their entire length. The short columnar cells of the plain tubules and of the necks of the convoluted tubules lack the yellowish bodies which are found in the cells of the duct, but the long columnar cells of the trunk of the compound tubules have more or less of them.

Each glomerulus almost entirely fills its Bo\vman's capsule. The squamous epithelial lining of the capsule continues around the glomerular mass. The capsule is usually oval in shape. The glomeruli are usually largest at the anterior end of the duct, and diminish in size gradually toward the posterior, the most posterior being the smallest. For instance, the following are measurements of glomeruli from different regions of the duct:









■ I

Right Left

0.640 0.600




/ I

Right Left

0.500 0.480

0.630 0.645



Right Left

0.345 0.331





The tubules ofj the Malpighian bodies, also, are smallest and shortest at the posterior ends of the duct. In one specimen the most posterior trace of a Malpighian body is a compound tubule, the base of which opens into the duct, and the neck ends blindly. The neck is 0.06 mm. long by 0.105 mm. wide, and the trunk is 0.15 mm. long by 0.21 mm. wide.

Each Malpighian body is surrounded by a compact stratified connective tissue band which is arranged in concentric rings and which is closely applied to the epithelium of the capsule. Outside this band of connective tissue there may be a more or less loose web of connective tissue which envelopes the Malpighian body and the mesonephiic duct. In one section, which shows the short plain tubule opening into both the duct and the capsule, the broad band of stratified connective tissue surrounding the duct continues around the tubule and the Bowman's capsule. Here the tubule is very short and narrow. Figure 62 shows the relation of these tissues when both glomerulus and the basal portion of the tubule are present in the same section with the duct. From this figure it will be observed that the glomerulus and the distal end of the tubule are not entirely separated from the mesonephric duct, but that they are bound to the latter by either a band or a web of connective tissue which envelopes both duct and Malpighian body.


Mesonephric Ducts. The mesonephric ducts of Myxine are like those of Bdellostoma in so many respects that a detailed description of them would be needless repetition, hence only points in which they differ will be noted.

The ducts are longer in Myxine, of course, and are a fraction of a millimeter narrower than the female ducts of Bdellostoma. The ducts of the latter are gray and opaque, while those of Myxine are approximately of the same color as the flesh and are somewhat transparent.

The -anterior ends of the ducts are farther removed from the pronephros in Myxine, being 10 to 20 mm. posterior to it. No


bendings or convolutions were present in the ducts of any of the specimens. The later ad bending at the inter septal Unes is somewhat less than in Bdellostoma, which may be due to the fact that the somites are larger in Myxine. In the adult Bdellostoma the somites are approximately 5 mm. wide, while in Myxine they are about 6.5 mm.

The histology of the ducts of Myxine is practically the same as those of Bdellostoma. The long columnar cells are arranged in longitudinal ridges. In the youngest specimen studied (Myxine no. 18) the ridges are not formed in the same manner as in Bdellostoma. The basal membrane of the epithelium does not bend inward, toward the lumen of the duct, at the base of the ridges, hence no concavities are formed under the latter. Also, the nuclei of the epithelial cells remain at the same level around the entire circumference of the duct and turn inward only very sHghtly at the base of the ridges. The ridges are formed solely by the elongation of the epithelial cells. The longest cells are approximately 0.0675 by 0.0081 mm. The majority of the nuclei are round, although many are short ovals. There are approximately fifteen ridges in each duct in the mid-region of the body.

In all the adult Myxines the ridges in the ducts are more numerous, shorter and narrower than in BdeUostoma. Their shape is shghtly different from those of the latter animal, and much more uniform. They are usually narrower in the middle than at the ends, and never have a wide, flat surface such as are frequent in Bdellostoma. In one specimen, which has eggs 12 mm. long, the duct is 0.39 by 1.2675 mm., and has 54 ridges, each approximately 0.039 by 0.117 mm. The duct of another animal measures, at one place in the mid-region selected at random, 0.2925 by 0.975 mm., and has 53 ridges, each approximately 0.029 by 0.105 mm. The ridges present a finger-like appearance, as shown in figure 63.

The epithehal ridges in the ducts of adult animals do not have concavities beneath them, but the nuclei at the base of the ridges are arranged in a small heap (fig. 63) .


The nuclei of the epithelial cells in the ducts of Myxine are much nearer the bases of the cells than in Bdellostoma. Most of them are round, but some are oval. All the cells have much cytoplasm, and as a rule they are narrower at the middle than at the ends.

A very striking difference between the epithelial cells of Myxine and those of Bdellostoma is the quantity and distribution of the yellowish bodies. These are present in every ridge in Myxine, more in some ridges than in others of the same section, and more in some sections than in others. They are almost entirely confined to the center of the ridge, forming a narrow core which extends from the nuclei at the base of the cells to the top of the ridge (fig. 64), but not to the very distal ends of the cells. At the top of the ridges these yellowish bodies are in rows, which spread out in the shape of a fan, following the arrangement of the cells. The bodies vary in size from tiny granules to a diameter exceeding the width of a cell, and large and small are intermingled in an irregular manner in all parts of the core.

There is much less connective tissue around the mesonephric ducts and Malpighian bodies in Myxine than in Bdellostoma. The Malpighian bodies are confined to the anterior two-thirds of the duct, one in each somite. They are approximately the same size as those of Bdellostoma, the largest at the anterior end of the ducts and the smallest toward the posterior. Only the compound type of tubules were found in Myxine, and these are usually shorter than the compound tubules of the Malpighian bodies in Bdellostoma.



Fernale. The following is a general description of the female generative apparatus in Myxine. It is not strictly applicable to any one specimen, for there is much variation within certain limits. Schreiner (/04) has given a very detailed account of the ovary in the pAiropean Myxine, and examination shows that


the North-American Myxine does not present any striking differences.

The single ovary occurs on the right side, and extends from the region of the gall bladder to the posterior end of the coelomic cavity, and is approximately (within 2 or 3 cm.) one-half the entire length of the animal. It is suspended in the body cavity by a single mesovarium, which is attached along its proximal margin to the mesentery where the latter joins the dorsal surface of the alimentary canal, to the right of the supra-intestinal vein. In young animals the mesovarium presents a perfectly flat surface, but in specimens which have eggs 5 mm. long or more it becomes folded transversely, being most folded in the oldest females. The distal margin of the mesovarium does not take part in this folding, but remains straight, therefore the two margins are much shorter than the rest of the mesovarium, a condition which causes the mesovarium to bulge out laterally.

For about 10 mm. at the anterior and posterior ends, the mesovarium is merely a hne along the dorsal surface of the digestive tract, then increases more or less abruptly to its average width. At the posterior end it decreases more or less gradually to a Une on the dorsal surface of the aUmentary canal, and ends at the genital pore. Even in the youngest females, where the mesovarium is simply a flat sheet, its width varies at different points along its course (fig. 65). In successive older stages these inequalities in width are more and more pronounced, the widest points being where the eggs are attached. In the youngest specimens (no. 18) the mesovarium is 2 mm. wide at the widest points, and in Myxine no. 2, which has eggs 22 mm. long, it is as wide as 30 mm. where the large eggs are attached. Between these large eggs the mesovarium is from 10 to 20 mm. wide. The weight of the eggs doubtless stretches the mesovarium, for it is drawn out into an elongated strand where each egg is suspended. Frequently these strands are intertwined and even tied in knots. In specimen no. 20 the mesovariun is 25 to 30 mm. wide and the strands to which the corpora lutea are attached extend only 3 or 4 mm. beyond the general width of the mesovarium. In one of the oldest females the


mesovarium is approximately 15 mm. high along its entire course, except at the ends, and its distal margin is not folded, but is comparatively straight. Another old specimen has a mesovarium which is 20 to 25 mm. wide, and its distal margin is also straight. It is probable that, after losing the eggs, the distal margin of the mesovarium tends to become straight again and the excessive width caused by stretching where the eggs were suspended is taken up by much transverse folding. As stated above, the distal margin of the mesovarium does not take part in this folding. The eggs seldom occupy more than the distal third of the mesovarium, never extending entirely to the proximal margin. In the adult animals they are distributed throughout the distal half of the mesovarium, even to the outermost margin, but in one of the young specimens (fig. 65) the most distal milhmeter of the mesovarium is entirely without eggs. The smallest eggs are always most distal, successively larger stages extending proximally, the largest being most proximal (fig. 66).

The eggs are comparatively evenly distributed along the length of the mesovarium, except that there are few, if any, in the most posterior 20 to 30 mm. Eggs larger than 10 mm. long, when present, usually occur at fairly regular intervals along the mesovarium from within 2 to 5 cm. of its anterior end to 3 to 10 cm. from the posterior end. Often from three to four or five large eggs are found together in a cluster, due to the intertwining of the long strands of mesovarium by which they are suspended.

In each female which has normal eggs there are all gradations of sizes of these from mere dots to those 2 mm. long. When eggs longer than 2 nun. are present there are no intermediate sizes between the 2 mm. eggs and the large ones, and all of the latter are within 1 or 2 mm. of being of equal size. The largest eggs found in any of the specimens measure 7 to 8 by 24 to 25 mm. In immature females the smallest eggs are crowded close together along the entire distal margin of the mesovarium. In older specimens the smallest eggs are much less numerous; clusters of them are separated by more or less wide intervals. Eggs 2 mm. long average approximately one for every 5 to 10



mm. of mesovariun in young females, and even less in adults. The following table shows the number of larger eggs present in a few individuals :







1.5-2.0 X 6.0-7.0




3.0-3.5 X 10.0-11.0




3.0-4.0 X 13.0-15.0




6.0-7.0 X 18.0-19.0




8.0-8.5 X 19.0-20.0




7.0-8.0 X 21.0-22.0




8.0-9.0 X 21.0-22.0




7.0-8.0 X 24,0-25.0


None of the eggs are so far advanced as to have hooks at the ends.

In shape, the eggs are round until they reach a diameter of about 1.5 mm., then they begin to elongate, and become oval. Eggs 2 mm. long a e 1 to 1.5 mm. wide. The eggs are developed between the two layers of peritoneum which form the mesovarium, as shown in figure 74. When the eggs begin to assume the oval shape, the pointed ends project beyond the place of attachment to the mesovarium, and one end evidently grows more rapidly than the other for the largest eggs are usually attached between one end and the equator, leaving the ends free (fig. 67).

Other structures besides eggs may be present in the mesovarium. Specimen no. 20 has 33 empty egg envelopes, the 'corpora lutea' (fig. 68). The largest eggs in this animal are approximately 1 by 1.5 mm. Specimen no. 2 has, besides the 27 large eggs, two corpora lutea, each of which is the same size as the envelopes around the eggs (8.5 by 20 mm.), and which has been opened along one side (fig. 69). The corpora lutea in Myxine no. 20 are much smaller (4 by 6 mm.) and' are shrunken into a compact mass. Myxine no. 5 has 25 or 30 of these corpora lutea which are still smaller (2 by 3 mm.) than those of specimen no. 20, as shown in figure 70. The largest eggs of specimen no. 5 are 1.5 to 2 by 0.5 to 1 mm. Specimen no. 1 has several even


smaller corpora lutea. The writer is unable to explain what is the final fate of the corpora lutea, unless they form some of the small brown oval bodies which will be described later. It seems improbable that they are completely absorbed. The fact that, as noted above, there are no intermediate stages between eggs about 2 mm. long and the large ones, and that also in animals which have corpora lutea, the eggs present do not exceed 2 mm., is interpreted to mean that as soon as some eggs exceed 2 mm. in length all the other eggs are arrested in development until the larger ones have matured and have been passed from the body, and their corpora lutea are well along in the process of degeneration.

Distributed comparatively uniformly along the entire mesovarium, wherever eggs occur, are numerous brown oval bodies which measure approximately 0.5 by 1 mm. They are somewhat flattened laterally, and are located proximally to the smallest eggs. None of these brown bodies are present in the youngest specimens, but they occur in all adult females, being most numerous in the older ones. In the mesovaria of some of the latter, for example specimen no. 9, there are no eggs, but many of these brown bodies. Since only comparatively a few eggs become larger than 2 mm., many eggs do not attain complete development, but degenerate. The brown bodies are the degenerated eggs and their envelopes. Intermediate stages of degeneration between the brown bodies and the normal eggs occur occasionally. It is possible that some of the brown bodies represent degenerated corpora lutea, though none were found whose structure would indicate this. As shown in figures 71, 72 and 73, representing sections of three stages of these, the mesovarium envelopes the brown bodies in the same manner as it does the small eggs (fig. 74), but the walls of the envelope around the former are much thicker. Also, numerous blood capillaries occur between the envelope and the former membranes of the egg, which have been converted into convoluted strands of connective tissue. The center of the brown bodies is filled with round nuclei and dark-yellowish granules which may be the remains of the yolk and which cause the brown color.


Along the posterior 25 or 30 mm. the mesovarium of young animals has a narrow band (0.5 mm. wide) of testis lobes, as shown in figure 65, but there are no traces of such testis lobes in any of the nine adult females which have normal eggs. A few small eggs are occasionally found in the mesovarium proximad to these testis lobes, but only at the anterior end of the testis band. The posterior end of the mesovarium in adult females with normal eggs may contain a few small eggs or brown bodies, but it is usually without any reproductive elements.

Sections were made of portions of the bands of testis lobes found in two of the young animals, viz., no. 18, no. 7. Figure 75 is a transverse section of the band in Myxine no. 18. The structure is similar to that of the young European Myxine as described by Nansen ('87), Cunningham ('87 and '92), and Schreiner ('05). The testis band consists of a large mass of stroma cells and many primitive germ cells which lie among the stroma cells. The band is attached to the digestive tract by the continuation of the mesovarium, though here it may perhaps be called more properly the mesorchium. Here and there, in a transverse section of the band, is a small follicle formed by a single layer of stroma cells, inside of which are from four to two or three dozen spermatogonia. No mitotic figm'es are present in any of the spermatogonia. On the contrary, the cells are in a resting stage, the nucleolus being visible in most of them. The entire mass is surrounded by the squamous epithelium of the mesorchium.

Myxine no. 7 is older than no. 18, and its testis lobes are more advanced in development, as is shown by figure 76. The follicles are more numerous, are larger, and contain more spermatogonia. The stroma cells are much less numerous. No mitotic figures are present, but the chromatin of many of the spermatogonia is scattered throughout the cell and the nucleolus has disappeared. The epithelium which surrounds the testicular mass is thicker, and here and there it turns inward, thus cutting the testis band into lobes. Small eggs occur along the entire distal margin of the mesorchium anterior to the testis band. The follicles of neither no. 1 8 nor no. 7 contain any spermatozoa. These specimens are


not sufficiently advanced in development to indicate whether they will eventually be male or female.

Male. Myxine no. 15, an animal 62 cm. long, has a testis band 5 mm, wide along the distal margin of the posterior 9.5 cm. of the mesorchium. Anterior to this band the mesorchium is of comparatively uniform width (about 7 mm.) all the way to the anterior end, and contains along its distal margin numerous brown bodies (degenerated eggs) but no normal eggs. No brown bodies, however, are present in the portion of the mesorchium which is occupied by the testis band. The mesorchium is not folded, but is straight and flat (fig. 82). There are no indications that this animal ever produced large eggs, and it is considered as an almost mature male. Only about one-half of the folUcles contain cells which show mitotic figm-es, but no spermatozoa were found. A transverse section of the testis band (fig. 77) shows that, in this specimen, the large follicles are larger than the largest ones in specimens no. 18 and no. 7, but they range in size from small to large and are closely crowded together, so that the stroma cells are limited to narrow strips between the follicles or between the latter and the mesorchial sheath. Almost every stage of mitosis is represented. Not all the cells of an individual follicle show mitotic figures, but what figures are present are approximately in the same stage (fig. 78).

A striking difference between the testis band of Myxine no. 15 and those of specimens no. 18 and no. 7 is the fact that the single layer of squamous mesorchial epithelium which envelopes the entire band has been here and there converted into a columnar epithelium. The long, spindle-shaped cells of the mesorchial sheath become shortened and arranged in palisade order. These columnar cells have not entirely surrounded the testis band, but in places the squamous epithelium is still present and forms the sheath of the band.

In two other specimens, no. 12 and no. 13, this process of conversion of the mesorchial sheath into columnar epithelium has advanced until the entire testis band is enveloped by it. The band in these two animals is very different from that of any other specimen. In the first place, the band extends the entire


length of the mesorchium (fig. 83). Specimen no. 12 has the general appearance of being older than no. 13. At the anterior end of the genital fold the width of the mesorchium and the testis band together is 4 mm., and it increases at once to 5 mm., then fluctuates between 4 and 5 mm. to a point about 7 cm. from the extreme posterior end. At this point the combined width of the band and the mesorchium increases to 7 mm. and remains this to the posterior end of the mesorchium. The testis band is suspended from the distal end of the mesorchium, and at the anterior end it is 1 nam. wide by 0.5 mm. thick. The band increases gradually in both width and thickness as it proceeds posteriorly, and at mid-region it is 2 mm. wide by 0.75 mm. thick. At a point 7 cm. from the posterior end it reaches its maximum size, 6 mm. wide by 1 mm. thick, and retains these dimensions to the extreme posterior end, immediately in front of the genital pore. The testis band of specimen no. 13 is neither as high nor as thick as that of no. 12, the widest part being 5 mm. Also, it is not as lobulated as the band of no. 12, and it is yellowish in color instead of reddish as in no. 12. In the latter the small lobules give a gi-anular appearance to the band, which is especially pronounced in the posterior region. There is not the slightest trace of eggs, normal or degenerated, at any place in the mesorchium of either no. 12 or no. 13. Myxine no. 12 is 67 cm. and the testis band is 36 cm. long. Specimen no. 13 is 63 cm. long and its testis band measures 35 cm.

Transverse sections, cut in all parts of the testis bands of these two animals, reveal the fact that the bands are of uniform and peculiar structure. They do not resemble the testis bands of younger Myxine, nor those of Bdellostoma, as figures 80 and 81 show. The squamous mesorchial epithelium has been completely converted into the palisade layer of columnar cells. In transverse section the band is cone-shaped, and deep notches in the sides cut it into triangular lobes, leaving a central core of stroma. The latter has been entirely converted into fibrous connective tissue which contains many blood capillaries. Nowhere in the bands is there the slightest trace of primitive germ cells or follicles. Figure 81, a transverse section from the posterior end


of the band of Myxine no. 12, shows another peculiar condition. At the base of the band, in a pocket formed between the lobes, there is a tangled skein of columnar epithelium which is of the same width as the columnar epithelium around the band. Here and there this skein is continuous with the epithelium which envelopes the band. It extends through all the sections cut in the posterior region of the testis band of Myxine no. 12, but occurs nowhere else. The writer has been unable to explain what this tissue may be.

Myxine no. 12 and no. 13 are old animals, as is indicated by their general conditions, viz., the thick, hard skin, the reddish, coarse fibers of the muscular tissue, the enlarged, coarse, darkbrown liver, the great diameter of the body, etc. The peculiar condition of their testis . bands is interpreted as due to their having become completely sterile with old age. Myxine no. 1*8 and no. 7, young, immature males, have many primitive germ cells, a few follicles and very little connective tissue in the testis band. Myxine no. 15, and adult male, has few germ cells, many follicles, and much connective tissue around and between the follicles in the testis lobes. In places the squamous epithelial covering of the band has been converted into a columnar epithelium, but sufficient squamous epithelium remains to permit easy ruptiu-e of the folhcles when they are ripe. The band is notched only slightly here and there. In specimens no. 12 and no. 13 the germ cells and follicles have entirely disappeared, and the bands are completely filled with fibrous connective tissue and blood capillaries, and all the squamous mesorchial sheath has been converted into a columnar epithelium. The testis bands are deeply notched in all parts so that they are cut up into lobes. The testis bands in these two animals have lost, their power of functioning and are completely sterile.


Fe7nale. The generative organs of Bdellostoma closely resemble those of Myxine in many respects. The ovary is single, occurring on the right side only, and is suspended in the body



cavity from a mesovarium which extends from the region of the gall bladder to the posterior end of the body cavity, a distance which is approximately one-half the entire length of the body. Unlike Myxine, however, the proximal margin of the mesovarium is always attached to the mesentery about one-third the width of the latter from the alimentary canal. The mesentery and mesovarium are more delicate than in Myxine.

In the females which have eggs 2 mm. long or longer, the mesovarium is approximately 10 to 15 mm. wide. Wliere the large eggs are attached, the mesovarium is stretched into strands which are 25 to 30 mm. long, and between the eggs it is 10 to 20 mm. wide.

The distribution of the eggs is very similar to that in Myxine. There are all gradations of sizes of young eggs from mere dots to those 2 mm. long in adult females, but there are no intermediate sizes between the latter and any larger eggs or corpora lutea that may be present.

The following table will give some conception in regard to the number of large eggs which may occur in an adult female :





7nim. 1.0-1.5 X 3.0- 6.0



2.5-3.0 X 10.0-11.0



3.0-3.25 X 12.0-13.0



3.0-3.5 X 13.0-14.0



3.0-3.5 X 14.0-15.0



3.0-3.5 X 14.0-15.0



5.0-6.0 X 20.0-21.0


Two of the specimens of Bdellostoma contain corpora lutea, one having eleven, the other twenty-four, and all show signs of degeneration. Small brown ovals, representing degenerated eggs, are scattered irregularly among the eggs.

In all the females, the eggs extend to within 1 or 2 cm. of the posterior end of the mesovarium (fig. 84) , and there are sometimes a few very small eggs in this most posterior part. No testis



lobes are present in any part of the mesovarium of any of the females, not even in the posterior region.

Male. No female elements occur in any of the males, and none of the mesorchia presents the appearance of ever having contained eggs. The general structure and appearance of the mesorchium is very similar in all the specimens. It extends from the region of the gall bladder to the caudal end of the body cavity, and is attached to the mesentery along a line which is about one-third the width of the latter from the alimentary canal (fig. 85). At the anterior end the mesorchium is merely a line along the dorsal surface of the digestive tract. At a point opposite the posterior end of the gall-bladder it leaves the alimentary canal and gi-adually ascends in the mesentery to its definitive position. Immediately behind the gall-bladder it is from 1 to 3 mm. wide, and retains this width to the testis lobes at the posterior end, usually decreasing slightly in width just before reaching the lobes. In most of the males the mesorchium, which is very delicate, contains no genital elements in the anterior and mid-regions, but in some individuals very small testis lobes are scattered along the distal margin of the mesorchium from the extreme anterior end to the band of lobes at the posterior end. One male has a single lobe about 0.5 mm. in diameter at the anterior end of the mesorchiimi and no others excepting those in the testis band at the posterior end. Other specimens have two or three such small lobes, still others a dozen or more, while one animal has so many that they look as though at one time they formed a continuous narrow band along the distal margin of the mesorchium.

The large mass of testis lobes is confined to the posterior end of the mesorchium, and consists of a great many lobules closely crowded together into a band, which varies in length, in the seven males examined, from 3 to 8 cm. and in width from 2 to 7 mm. There is no correlation between the length of the testis band and the length of the animal, as is shown by the following table :







































In the longer, i.e., the older, animals the testis band is wider and thicker than in the younger males.

The testis bands are folded more or less in the young males and considerably in the older animals. In all, the bands have a very granular appearance (fig. 85). The bands in the younger males are hght gray in color, and yellowish in older animals.

Transverse sections of the testis bands show that in minute structure they are very similar to those of Myxine. Single germ cells and follicles of all sizes, from very small to large, are present. In a single section all stages of mitotic figures may occur, and many of the follicles contain spermatozoa (fig. 79). Each follicle has for its wall a single row of stroma cells, arranged end to end, and outside this are two or more concentric layers of fibrous connective tissue. Capillaries occur here and there in the connective tissue and stroma between the follicles. The primitive germ cells are limited to the distal end of the band, and here the follicles are smaller, their walls are not so prominent and there is less connective tissue than in the proximal part, where the large follicles are located. The entire band is covered by a single layer of squamous epithelium which is a continuation of that of the mesorchium. In no place is there columnar epithelium around the testis band as is described for Myxine.

Cunningham ('87) and Nansen ('87) have advanced the theory that Myxine is a protandric hermaphrodite, functioning as a male when young (28 to 32 cm. long) and as a female when older


(32 to 35 cm.)- Nansen thinks the change in sex occurs when the animal reaches a body length of about 32 or 33 cm. Out of hundreds of specimens examined by each of these investigators, only a few males were found (Cunningham 8, Nansen 'very few') and most of these were immature. Nansen regards the males as merely transformed hermaphrodites and found all transition stages between males and common hermaphrodites. Cunningham found that in "nearly all specimens with very immature eggs the posterior portion of the sexual organ had the same structure as the testis," and that "in all specimens with well-developed ovarian eggs .... with one exception, no testicular portion was present in the sexual organs," then he concluded "that in the young state the females are nearly, but not quite always hermaphrodite, and that the testicular portion normally disappears as the eggs become more mature." Nansen states that ' ' on opening large specimens of Myxine, we gen•erally find well developed ova in their sexual organs. If we, Jiowever, take smaller specimens, of about 28 to 32 cm. in length, and examine their sexual organs, we generally find that the anterior portion is but slightly prominent, and contains very small and young ova, whilst the posterior portion is often very broad and prominent, is lobate, and has a distinct whitish colour along its margin, and has, in all respects, the appearance that we would expect to find in a testis; and that it really is." He concludes that "Myxine is generally, or always (?), in its young state, a male; whilst at a more advanced age it becomes transformed into a female. Indeed, I have not yet found a single female that did not show traces of the early male stage."

Schreiner found only 19 pure males out of 2500 specimens examined, and they were seldom over 33 cm. long, the majority of them being 30 to 31 cm. In a quite preponderating number of individuals which were longer than 33 cm. there was a more or less well-developed ovary and the testis part was sterile, and in the males there was a more or less well-developed testis and an abnormal or rudimentary ovary. In the females, the ovary was well-developed and the testis part was usually rudimentary or occasionally completely sterile. The testis was always ab


normal in both young and old females. In addition to functional males and females, Schreiner found individuals in which both testis and ovary were degenerated, and these he classed as sterile animals. He also found normal females 22 cm. long and less, and states that, lasst sich das haufige Vorkommen von jungen Weibchen unmoglich mit der Annahme eines protandrischen Hermaphroditismus bei diesem Tiere vereinigen." Schreiner beheves that Myxine is dioecious.

It is well known that Bdellostoma is an animal of separate sexes. The fifteen adult females examined by the writer have no trace of testis lobes in any part of the mesovarium, and no female elements occur in the mesorchium of any of the eight adult males. The females range in length of body from 34 to 40 cm., and the males from 36 to 42 cm. That is to say, in this particular lot of animals some of the functional males are larger than the adult females, hence it is certain that Bdellostoma is not a protandric hermaphi-odite. It is improbable, to say the least, that of two animals so closely related as Bdellostoma and Myxine, the one should have normal, separate sexes while the other should possess such a pecuhar method of reproduction as protandric hermaphroditism.

Furthermore, as Dean ('99) has suggested, the occurrence of large males would also discredit the theory that Myxine is a protandric hermaphrodite. Myxine no. 15, described above, is undoubtedly a normal male, in which the only female sexual products are small, completely degenerated eggs, and specimens no. 12 and no. 13 are males which have become sterile with age and which have no female elements whatever. Myxine no. 15 is 62 cm. long, no. 12 is 67 cm., and no. 13 is 63 cm. The average length of the females with large eggs, from 18 to 25 mm. long, is 59 cm., the longest being 64 cm. Here, then, we have one normal and two sterile males which are longer than the average female with large eggs. In view of these facts, it seems impossible that Myxine can be a protandric hermaphrodite.



1. A central duct is present in the pronephros of both Bdellostoma and the North American Myxine, but it is in a state of degeneration in both animals. This degeneration is more advanced in Myxine than in Bdellostoma, so that it includes the inner ends of the pronephric tubules and all parts of some of the largest tubules.

2. The 'central mass,' which lies in the pronephric vein, is not a series of glomeruli, but represents the disintegrated central duct (including the inner ends of tubules in Myxine).

3. The Malpighian body of the pronephros in both Bdellostoma and Myxine is located at tlie posterior end of the pronephros, and in appearance and structure closely resembles the Malpighian bodies of the mesonephros. It is formed by the fusion of two or more glomeruli.

4. The glomerulus of the pronephros in young Myxinoids is exposed to the pericardial cavity through a large opening, thus resembUng a glomus. This opening becomes constricted to a small tubule in adult animals.

5. The large vein which extends along the base of each pronephros communicates with the lumen of the central duct through large natural openings. The pronephric tubules connect the lumen of the duct with the pericardial cavity, and the latter opens into the peritoneal cavity through the large pericardoperitoneal foramen, while the body cavity opens into the cloaca through the genital pore. Thus, the vascular system is in communication with the exterior of the body.

6. The morphology of the pronephros does not reveal any positive indications as to its function. The granular coagulum in the tubules and duct suggests excretion. This question can probably be determined only by physiological methods.

7. The tubules of the pronephroi of adult Bdellostoma are approximately the same diameter, while those of Myxine vary in size.

8. Each tubule in the pronephros does not represent an original pronephric tubule.


9. The mesonephric ducts of Mjrxinoids probably possess a secretory as well as an excretory function, as is suggested by the long epithelial cells, the increased surface of these by ridges, the large granules at the distal ends of the cells, and the yellowish bodies in the cells. Also, isolated pieces of the duct contain waste particles hke those present in all parts of the continuous portion of the duct.

10. The mesonephric ducts of adult male Bdellostomae are larger than those of adult females, and the epithelial ridges within them are larger.

11. The mesonephric ducts of Myxinoids open separately into the cloaca.

12. The tubules of the mesonephric Malpighian bodies in Bdellostoma are structm'ally of two types.

13. \\Tien some eggs in the ovary of Myxinoids attain a size exceeding about 2 nmi. in length, all the smaller eggs are arrested in growth until the large eggs reach maturity and have been passed from the body.

14. ^lany of the small, young eggs in Myxinoids degenerate, forming the 'brown bodies.'

15. When a mature egg passes from the ovary it leaves attached to the mesovarium a 'corpus iuteum,' composed of the outer envelope of mesovarium which surrounded the egg. The corpora lutea degenerate.

16. After the testis of old males ceases to function, the peritoneal epithehum which covers the testicular band becomes changed from squamous to columnar; the empty testicular folhcles completely disappear, and the band becomes converted into a mass of connective tissue penetrated by blood vessels.

17. Neither Bdellostoma nor Myxine is a protandric hermaphrodite.

1&. The m'ogenital system of the North American ]\l}^xine does not present any specific differences from that of the Em'Opean animal. On the other hand the great difference in size of mature individuals from the two sides of the Atlantic would appear to support the distinctness of M. limosa of Girard from the M. glutinosa of Linneaus.



Ayers, Howard 1894 Bdellostoma dombeyi, Lac. Biological lectures at Woods

Hole in summer session of 1893. Ginn & Co. Beard, J. 1892 Notes on lampreys and hags. Anat. Anz., Bd. 8; pp. 59-60.

Also, Brit. Assoc. Rep., Edinburgh, 1892, p. 790. BuRNE, R. H. 1898 The 'porus genitalis' in the Myxinidae. Journ. Linn. Soc.

London, vol. 26; pp. 487-495. Cole, F. J. 1905 Notes on Myxine. Anat. Anz., Bd. 27; pp. 323-326.

1914 Notes on the vascular system of Myxine. Anat. Anz., Bd. 46;

pp. 478-485. CoLLiNGE, Walter E., and Swale, Vincent 1896 On the so-called suprarenal

bodies in Cyclostoma. Anat. Anz., Bd. 12; pp. 232-241. Cunningham, J. T. 1886 The reproductive organs of Bdellostoma. Trans.

Roy. Soc. Edinburgh, v. .33; pp. 247-250. Abstr. in Journ. R. Micr.

Soc. London, 1888, pp. 192-193.

1887 a On the structure and development of the reproductive elements

in Myxine glutinosa, L. Quart. Journ. Micr. Sci., vol. 27; pp. 49-76.

1887 b Herr Max Weber and the genital organs of Myxine. Zool.

Anz., Bd. 10; pp. 241-244.

1887 c The reproduction of Myxine. Zool. Anz., Bd. 10, no. 256;

pp. 390-392.

1892 Spermatogenesis in Myxine glutinosa. Quart. Journ. Micr. Sci.,

vol. 33; pp. 169-186. Dean, Bashford 1897 The Columbia University zoological expedition of 1896.

Trans. N. Y. Acad. Sci., vol. 16; pp. 33-42.

1897 On the development of the Californian hag-fish, Bdellostoma

stouti. Lock. Quart. Journ. Micr. Sci., N.S., vol. 40; pp. 269-279.

1899 On the embryology of Bdellostoma stouti. A general account of Myxinoid development from the egg and segmentation to hatching. Festschr. Kupffer, p. 221-276.

1900 The egg of the hag-fish, Myxine glutinosa, Linn. Mem. N. Y. Acad. Sci., vol. 2; pp. .33-45.

1903 The eggs of the eastern Atlantic hag-fish, Myxine limosa, Gir.

Science, N. S., vol. 17, no. 428; p. 433. Felix, W. 1897 Die Price'sche Arbeit, "Development of the excretory organs of

a Myxinoid (Bdellostoma stouti Lockington)," und ihre Bedeutung fiir

die Lehre von der Entwickelung des Harnssystems. Anat. Anz., Bd.

13; pp. 570-599. Girard, Chas. 1858 Icthyological notices. Proc. Phila. Acad., 223. Jackson, C. M. 1901 An investigation of the vascular system of Bdellostoma

dombeyi. Journ. Cin. Soc. Nat. Hist., vol. 20; pp. 13-47. Kirkaldy, J. W. 1894 On the head kidney of Myxine. Quart. Journ. Micr.

Sci., vol. 35; pp. 353-359. Lewis, F. T. 1904 The question of sinusoids. Anat. Anz., Bd. 25; pp. 261 279. Maas, Otto 1897 Ueber Entwicklungsstadien der Vorniere und Urniere bei

Myxine. Zool. Jahrb., Abt. Anat., Bd. 10; pp. 473-510.


MiNOT, C. S. 1898 On the veins of the Wolffian bodies in the pig. Proc. Boston Soc. Nat. Hist., v. 28; pp. 265-274.

1900 On a hitherto unrecognized form of blood circulation without capillaries in the organs of Vertebrata. Proc. Boston Soc. Nat. Hist., vol. 29; pp. 185-215.

MtJLLER, J. 18-43 Untersuchungen iiber die Eingeweide der Fische, Schluss der vergleichenden Anatomic der Myxinoiden. Abhandl. der konig. Akad. der Wissens. zu Berlin, pp. 109-170.

MuLLER, W. 1875 Ueber das Urinogenitalsystem des Amphioxus und der Cyclostomen. Jena. Zeit. f. Naturwiss., Bd. 9; pp. 94-129.

Nansen, Fridtjof 1887 A protandric hermaphrodite (Myxine glutinosa, L.) amongst the vertebrates. Bergen's AIus. Aarsber., p. 1-34.

Price, G. C. 1896 Some points in the development of a Myxinoid (Bellostoma stouti Lockington). Verh. Anat. Ges., Bd. 10; pp. 81-86. Zur Ontogenie eines Myxinoiden (Bdellostoma stouti, L.). Sitz. Ber. mathphys. CI. k. bayer. Akad. Wiss. Miinchen., Bd. 26; pp. 69-74. 1897 Development of the excretory organs of a Myxinoid, Bdellostoma stouti Lockington. Zool. Jahrb., Abth. Anat. Ontog., Bd. 10; pp. 205-226. Abstr. Journ. R. Micr. Soc. London, vol. 5; p. 316. 1904 A further study of the development of the excretory organs in Bdellostoma stouti. Am. Jour. Anat., vol. 4; pp. 117-138. 1910 The structure and function of the adult head-kidney of Bdellostoma stouti. Jour. Exp. Zool., vol. 9; pp. 849-864.

PRI^XE, E. E. 1915 The curious egg of the hagfish. Ottawa Naturalist, vol. 29; p. 14.

Putnam, F. W. 1873 Notes on the genus Myxine and genus Bdellostoma. Proc. Boston Soc. Nat. Hist., vol. 16; pp. 127-135, 156-160.

ScHREiNER, A. TJND K. E. 1905 Ueber die Entwicklung der mannlichen Geschlechtszellen von Myxine glutinosa. Arch. Biol., Bd. 21; pp. 183314, 315-355.

1908 Zur Spermienbildung der Myxinoiden. (Ueber die Entwicklung der miinnlichen Geschlechtszellen von Myxine glutinosa L.). Arch. Zellforschung, Bd. 1; pp. 152-231.

ScHREiNER, K. E. 1904 Ueber das Generationsorgan von Myxine glutinosa. Biol. Centralbl., Bd. 24; pp. 91-104, 121-129, 162-173.

Semon, Rich. 1896 Das Excretionssystem der Myxinoiden in seiner Bedeutung fur die morphologische Auffassung des Urogenital systems der Wirbelthiere. Festschr. C. Gegenbauer, Bd. 3; pp. 167-192. 1897 a Das Excretionssystem der Myxinoiden. Anat. Anz., Bd. 13; pp. 127-137. 1897 b Vorniere und Urniere. Anat. Anz., Bd. 13, pp. 260-264.

Semper, C. 1875 Das Urogenitalsystem der Plagiostomen und seine Bedeutung fiir das der iibrigen Wirbelthiere. Arbeiten aus dem zoologisch. Institut zu Wiirzburg, Bd. 2; pp. 167-509.

Spengel, J. W. 1897 a Die Excretionsorgane von Myxine. Anat. Anz., Bd. 13; pp. 49-60.

1897 b Semon's Schilderung des Mesonephros von Myxine. Anat. Anz., Bd. 13; pp. 211-216.


Weber, Max 1887 Erwiderung an Herrn Cunningham. Zool. Anz., Bd. 10,

no. 253; p. 318-321. Weldon, W. F. R. 1884 On the headkidney of Bdellostoma, with a suggestion

as to the origin of the suprarenal bodies. Quart. Journ. Micr. Sci.,

vol. 24; pp. 171-182. WoRTHiNGTON, JuLiA 1905 Contribution to our knowledge of the Myxinoids.

Amer. Natural., vol. 39; pp. 625-663.




a, dorsal aorta

ac, anterior cardinal vein

al, alimentary canal

an, anus

au, auricle

h, Bowman's capsule

fee, trunk of compound tubule

fern, basal membrane

fes, blood sinus

c, connective tissue

cb, band of connective tissue

cd, central duct

ce, columnar epithelium cl, corpora lutea

do, cloaca

cm, central mass

CO, blood corpuscle

cp, common portal vein

d, distal margin of mesovarium or mesorchium

dc, ductus Cuvieri

dm, degenerated Malpighian body

do, degenerated ova, 'brown bodies'

e, connective tissue envelope

eh, epithelium of base of compound

tubule of mesonephros el, epithelium of plain mesonephric

tubule em, epithelium of mesonephric duct en, epithelium of neck of compound

mesonephric tubule eo, endothelium ejp, epithelium /, funnel-shaped mouth of pronephric

tubule fl, testicular follicle g, glomerulus gh, gall bladder gc, giant blood corpuscle gp, process composed of granules i, intertubular space liv, liver

Ij), left postcardinal vein m, Malpighian body of mesonephros md, mesonephric duct

ini, mesoderm of peritoneum

mn, mesonephros

m-p, Malpighian body of pronephros

mr, mesorchium

ms, mesovarium

mt, tubule of Malpighian body

n, nucleus

nc, neck of compound mesonephric tubule

nu, nuclear mass

0, ovary

op, opening of central duct into pronephric vein

ov, ova

p, plain mesonephric tubule

pc, pericardium

?5cc, pericardial cavity

pf, pericardo-peritoneal foramen

pg, primitive germ cell

ph, portal heart

pn, pronephros

pnr, pronephros

pt, epithelium of peritoneum

pv, pronephric vein

ra, renal artery

rac, right anterior cardinal vein

97;, right postcardinal vein

rpv, right pronephric vein

s, stoma cells

sa, somatic artery

sc, skein of columnar epithelium

se, squamous epithelium

si', supra-intestinal vein

so, somite

ss, sheath of stroma cells around testicular follicle

st, septum transversum

t, pronephric tubule

th, testis band

tu, mesonephric tubule

V, blood vessel

w, waste granules

y, yellowish granules

z, pericardial attachment to dorsal aorta and alimentarv canal.



1 Right pronephros of Bdellostoma no. 6, showing its position in the pericardial cavity. Natural size.

2 Left pronephros of Bdellostoma no. 6, showing position in the pericardial cavity. Natural size.

3 Right pronephros of Myxine no. 5, showing position in the pericardial cavity. Natural size.

4 Left pronephros of Myxine no. 5, showing position in the pericardial cavity. X !§.

5 Transverse section through the right pronephros of Bdellostoma no. 6. Camera, X 25.

• 7 Left pronephros of Bdellostoma no. 16. X 4.

8 Left pronephros of Bdellostoma no. 18. X 4.

9 Left pronephros of Bdellostoma no. 7. X 4.

10 Right pronephros of Bdellostoma no. 16. X 4.

11 Right pronephros of Bdellostoma no. IS. X 4.










rac rpv


Pnr I






P^^ '^^^= rp^^ PC







11a Transverse section of the right pronephros of Bdellostoma no. 10, illustrating the three-sided shape.

12 Millimeter paper reconstruction of the right pronephros of Bdellostoma no. 15. X 25.

13 Pronephric tubule with funnel-shaped mouth.

14 Pronephric tubules in longitudinal and transverse section. Bdellostoma no. 15.

15 Transverse section through two pronephric tubules showing relation of the vascular endothelium to the tubules when a blood sinus occupies the space between the tubules. Bdellostoma no. 15.









16 Enlarged drawing of columnar epithelial cells of a pronephric tubule. Bdellostoma no. 15.

17 Transverse section through pronephros of Bdellostoma no. 10, showing the endothelium of the pronephric vein continuing around the central mass.

18 Transverse section through the central mass. Bdellostoma no. 15.

19 Enlarged drawing of the central mass. Bdellostoma no. 15.

20 Transverse section through pronephros of Bdellostoma no. 15, showing communication between the central duct and the pronephric vein, with nuclear mass in the opening; also shows pronephric vein opening into the space between the tubules, thus making a blood sinus. Not all the pronephric tubules are included in the drawing.

21 Transverse section in another part of pronephros of Bdellostoma no. 15, which also shows the central duct communicating with the pronephric vein and the opening filled with the nuclear mass. Not all of the tubules are shown.

22 Transverse section through pronephros of Bdellostoma no. 6, showing the central duct in communication with the pronephric vein through a large opening which is filled with a nuclear mass.

23 Transverse section through a pronephric tubule of Myxine no. 10, showing connective tissue surrounding the epithelium.









24 Transverse section through a pronephric tubule of Myxine no. 11, showing connective tissue.

25 Section through some tubules of pronephros of Myxine no. 15, showing the large amount of connective tissue around the tubules.

26 Transverse section through a pronephric tubule of Myxine no. 10, showing blood sinus around the tubule.

27 Transverse section through some tubules of pronephros of Myxine no. 10, showing difference in size of the tubules.

28 Transverse section through pronephros of Myxine no. 11, showing the central duct in the central mass.

29 Transverse section through pronephros of Myxine no. 11, showing central duct in communication with the pronephric vein. Camera lucida, X 50.




PLA l-E 4






30 Longitudinal section through the central duct in pronephros of Myxine no. n. Camera lucida, X 50.

31 Longitudinal section through the central duct of- one of the youngest specimens, Myxine no. 18. Camera, X 90. y^

32 Transverse section through pronephros of Myxine no. 10, showing the central duct. Not all the tubules are included in the sketch. Camera lucida, X 30.

33 Transverse section through pronephros of Myxine no. 11, showing the central duct with its columnar walls intact. Camera lucida, X 50.

34 Transverse section through pronephros of one of the youngest specimens. Myxine no. 18, showing the central mass. Camera lucida, X 50.

35 Longitudinal section through the central mass in pronephros of Myxine no. 11. Camera lucida, X 90.









36 Section pronephros of Myxine no. 15 which shows a broad band of connective tissue extending from the base of the tubules into the pronephric vein. Camera, X 50.

37 Transverse section through a degenerated Malpighian body immediately posterior to the left pronephros Myxine no. 15. Camera lucida, X 90.

38 Transverse section through a degenerated Malpighian body immediately posterior to the left pronephros of Myxine no. 11, with a blind tubule beside it. Camera, X 90.

39 Transverse section through pronephros of Bdellostoma no. 15, showing Bowman's capsule of the Malpighian body opening into the pericardial cavity. Camera lucida, X 50.

40 Transversesection through the Malpighian body of the left pronephros of Bdellostoma no. 16, showing lobulated structure of the glomerulus and the opening of Bowman's capsule into the pericardial cavity. Camera, X 90.






41 Section through the Malpighian body of one of the oldest specimens (Myxine no. 10), showing degeneration of the glomerulus. Camera lucida, X 19.

42 Section through the Malpighian body of pronephros of a young Myxine (no. 18), showing double glomerulus. Camera lucida X 50.

43 Section through Malpighian body of pronephros of a young Myxine (no. 18), showing the glomerulus exposed to the pericardial cavity through a large opening and resembling a glomus. Camera lucida, X 50.

44 Transverse section through mesonephros of Bdellostoma no. 17, cut in mid-region of body cavity, showing relation of the ducts to blood vessels and peritoneum.

45 Dorsal view of the excretory system of Bdellostoma no. 16. One-half natural size. Male.

46 Dorsal view of the excretory system of Bdellostoma no. 15. One-half natural size. Female.

47 Dorsal view of posterior ends of mesonephric ducts of Bdellostoma no. 16. Natural size.

48 Side view of posterior end of left mesonephric duct of Bdellostoma no 16. Natural size.




rpv a




pr liv


md -al

1 45



ov- tl!'^^

r-pr A

iiv pv-,f



i .,lr...t








pc y t pec




^ t







49 Ventral view of excretory system of a young male, Bdellostoma no. 6. One-half natural size.

50 Transverse section through the mesonephric duct of an adult male, Bdellostoma no. 16, cut in the mid-region of the body cavity. Camera lucida, X 19.

51 Transverse section through the mesonephric duct of an adult female, Bdellostoma no. 15, cut in the mid-region of the body cavity. Camera lucida, X 19.

52 Transverse section through the anterior end of a mesonephric duct of Bdellostoma no. 6, X 90.

53 Enlarged sketch of a ridge in the epithelial lining of a mesonephric duct. Bdellostoma.

54 Enlarged sketch of the anterior ends of the mesonephric ducts of Bdellostoma no. 4, mounted in damar.

65 Enlarged veneral view of the mesonephric ducts of Bdellostoma no. 8, showing location of the Malpighian bodies with reference to the somites.

56 Enlarged sketch of mesonephric ducts of Bdellostoma no. 16 showing the distribution of the arteries to the Malpighian bodies.

57 Longitudinal section through the tubule of a Malpighian body of the mesonephros, Bdellostoma no. 17. 'Plain' type of tubule. Camera lucida, X 90.









ra m.



58 Reconstruction of a 'plain' mesonephric tubule of Bdellostoma; diagrammatic.

59 Diagrammatic reconstruction of a 'compound' mesonephric tubule of Bdellostoma.

60 Sketch showing transition between the cells of the neck and those of the trunk in a 'compound' tubule. X 250.

61 Transverse section of a 'compound' tubule showing opening of the trunk into the mesonephric duct. Camera, X 50.

62 Transverse section through mesonephric duct of Bdellostoma no. 15, showing a Malpighian body and its tubule. Camera lucida, X 30.

63 Transverse section through mesonephric duct of Myxine no. 12. Camera lucida, X 50.

64 Enlarged sketch of epithelial ridges in mesonephric duct of Myxine no. 12, showing core of yellowish granules.

65 Sketch of ovary of one of the youngest specimens of Myxine (no. 18). Natural size.







66 Enlarged portion of ovary of an adult Myxine (no. 16), showing distribution of eggs. X 2.

67 Sketch of large egg in Myxine no. 2. Natural size.

68 Portion of mesovarium containing 'corpora lutea.' Myxine no. 20. Natural size.

69 A large corpus luteum from Myxine no. 2. Natural size.

70 Small corporea lutea from Myxine no. 5. Natural size.

71, 72, and 73 Sections of 'brown bodies' from Myxine showing three stages in the degeneration of eggs. Camera lucida, X 30.

74 Section through small eggs from Myxine no. 8. Camera lucida, X 30.








>IS6^^ ^



PLATE 10 -ms






ms .i




75 Transverse section through the testis band at the posterior end of the mesovarium of a young Myxine (no. 18). Camera lucida, X 50.

76 Transverse section through the testis band at the posterior end of the mesovarium of an older, but immature Myxine (no. 7). Camera lucida, X 30.

77 Transverse section through the testis band of Myxine no. 15, a normal male. Camera lucida, X 50.

78 Enlarged sketch of a transverse section through a follicle in the testis of Myxine showing the spermatogonia in the metaphase stage.

79 Transverse section through a follicle in the testis band of Bdellostoma no. 7, showing spermatozoa.

80 Transverse section through the anterior region of the testis band of Myxine no. 12. Camera lucida, X 30.

81 Sketch of a portion of transverse section through the posterior region of the testis band in Myxine no. 12. Camera lucida, X 30.




','f*f mr






Pg .^ _^^ -. mr










^ --5 2 J CQ s




— a



\ 'a

c "--^t


c ^ -"





{ ^J


1 ^1






•^" V

CO 00





Department of Anatormj, University of Minnesota



Introduction 165

Observations 1 67

I. Anatomj' of the adult syrinx 167

1 . Gross anatomy 167

a. Skeletal framework 168

b. Vibrating membranes • 172

c. Musculature 174

d. Sexual differences 175

2. ^Microscopic anatomy 176

II. Development 179

1. Mucous membrane 181

2. Cartilaginous skeleton 184

3. Submucosa _ 189

4. Musculature 190

5. Intercartilaginous membranes 191

6. Summary of development 192

Experiments on function 195

Conclusions 199

Structure 199

Development 200

Function 201


The songs of birds have served to inspire various classes of people for ages, but not until the last three or four centuries do we have any record of scientists becoming interested in them to the extent of studying the organs concerned in their production.

Competent authorities, such as Duvernoy, Cu\'ier, Johannes j\I idler, and others, early observed that birds possess two larynges




■ — an upper and a lower. From their observations these earlyinvestigators concluded that it was in the lower larynx that the voice is produced. Regarded as a voice organ, the syrinx has aroused much interest and has been the subject of numerous researches. These researches deal chiefly with its function, structure, and comparative anatomy. Its embryology has been quite neglected, and it was this fact that prompted the present investigation. At the outset it was thought best to confine this study to some single species of bird, and to work out the development of the voice organ, stage by stage, in this one form. It was found, however, that many anatomical details of the syrinx and its related structures had also been neglected. Therefore, before the embryological study could be undertaken, it became necessary to inquire more minutely into its anatomy. Somewhat incidentally the function of the syrinx, too, has been considered, and experiments performed to determine whether the lower larynx is, in truth, the voice-producing organ.

This investigation was undertaken on the suggestion of Prof. H. D. Reed of Cornell University, and was carried on, under his supervision in the Laboratories of Comparative Anatomy at that University until September, 1913, when it was transferred to the Anatomical Laboratories of the University of Missouri. To Dr. Reed of the former university the author is deeply indebted for many valuable suggestions and criticisms. At the latter university, the author is obligated to Dr. F. P. Johnson for his keen interest in the embryological division of these studies, and to Thos. J. Heldt for his aid and criticism.

For the early literature the reader is referred to L. Wunderlich ('84), who gives an excellent historical sketch of the work previous to his own. He includes a very complete resume of all work dealing with the syrinx as studied in numerous species of birds.

In his text book entitled "The development of the chick" Lillie '08, treats the respiratory system quite fully, but states that very little is known concerning the development of the voice organ in the chick.



In this investigation Galliis domesticus was used for two reasons. First, adults are accessible at all seasons of the year. Second, developmental stages are easily obtained and controlled.


1 . Gross anatomy

Birds differ from mammals in possessing two larynges, an upper and a lower. The latter, known also as the syrinx, was described by Johannes Mliller as occurring in three different positions. When the organ is located at the lower end of the trachea but above the bifurcation, he designated it 'syrinx trachealis.' If found in the bronchi below the bifurcation, 'syrinx bronchiahs.' But if the lower end of the trachea and the upper portions of the bronchi both took part in its formation he called it 'syrinx tracheo-bronchialis' (fig. 2). The syrinx of the domestic chicken belongs to the last type.

In the domestic chicken, the upper larynx (fig. 1) is located behind the tongue and on the floor of the pharynx. It is rudimentary when compared with that of other animals and is apparently incapable of producing voice. From the upper larynx the trachea courses caudad, ventral to the oesophagus, passes to the left of the ingluvies and pierces the pleura between the limbs of the furcula to enter the so-called thoracic cavity. Shortly after entering this cavity, the trachea bifurcates into the two primary bronchi. In the region of the bifurcation a great modification of the cartilaginous and intercartilaginous elements takes place. Here (fig. 2) a well developed bone, known as the cross-piece, or pessulus, lies transversely in the dorso-ventral plane between the two bronchi. Its dorsal and ventral extremities are greatly enlarged. The caudal end of the trachea is compressed. It is composed of rings which are enlarged and closely related to each other. The four most caudal rings are imperfectly fused. Immediately caudad to this fusion are four


highly modified cartilages which have free ends dorsally, but fuse with the pessulus ventrally.

The cartilages of the bronchi, unlike those of the trachea, are present only as halves of rings. These half -rings form the framework of the lateral bronchial walls. The first half-rings are modified, and are attached to the pessulus both dorsally and ventrally. The second are only slightly modified and are attached ventrally to the first by a small piece of cartilage. Their dorsal ends are free. Certain of these skeletal elements, so briefly referred to, serve as attachments for vibrating membranes. The external tympanic membranes (fig. 2) stretch between the first tracheal ring and the first half-ring. First, as used here in reference to the tracheal rings, designates the most caudal, as regards an enumeration of the total number of tracheal rings in the caudo-cephalic direction. First, as used here in reference to the bronchial half-rings, has a somewhat different meaning. It designates that bronchial half-ring which of all the bronchial half-rings lies in the most cephalic position. These designations are consistently adhered to throughout the paper. The internal tympanic membranes form that part of the medial wall of each primary bronchus which lies immediately caudad to the pessulus. They stretch, therefore, from the pessulus to the cephalic end of the medial membranous wall of each primary bronchus. In fact, the internal tympanic membranes may be regarded as specialized portions of the medial membranous walls. Still more caudad to the pessulus and between the caudal extremities of the two internal tympanic membranes, stretching transversely in the medio-lateral plane, is another membrane, the bronchidesmus. From its position, then, the bronchidesmus connects the medial walls of the two bronchi.

a. Skeletal framework. In the study of the bony and cartilaginous skeleton, three methods were pursued. First, numerous gross dissections were made. Second, cartilage and bone stains were used to stain the organ in toto. For staining bone 2 per cent alizarine in 96 per cent alcohol was used, while j per cent methylene blue in 67 per cent alcohol was employed as a cartilage


stain. These organs were then washed and cleared, after which the cartilaginous and the bony structures could be more easily studied with the binocular microscope. Owing, however, to the size and the shape of the organ as a whole, this method was not entirety satisfactory. Third, wax reconstructions were made according to the method described by Born.

The syringeal skeleton (fig. 4) consists of (1) the first four tracheal rings which may be designated, the cephahc sjTingeal skeletal elements; (2) the first three bronchial half-rings, or the caudal syringeal skeletal elements; (3) four modified cartilages, neither rings nor half-rings, which are found between the tracheal rings and the first bronchial half-rings. These modified structures will be designated the intermediate syringeal , cartilages; (4) a bony pessulus which hes at the summit of the bronchial junction in the dorso-ventral plane and transverse to the long axis of the trachea. Although the skeletal elements are similar in most respects in both sexes, the following description apphes, unless stated otherwise, only to the male.

The typical tracheal rings are broad flat bands of cartilage. They differ from the type found in most of the higher animals in that they are complete rings. In cross section the typical rings have a cephalo-caudal diameter which is much greater than the medio-lateral diameter, hence they appear very much elongated (fig. 5). But as the caudal end of the trachea is approached, the cephalo-caudal diameter of the rings diminishes while the medio-lateral diameter increases. The fifth, the sixth, and the seventh rings appear almost square when seen in cross section. These, as well as all other rings cephalad of them, are transverse to the long axis of the trachea and constitute the framework of its walls. The medio-lateral diameter of the fourth ring is greater than its cephalo-caudal diameter. The most ventral portion of this cartilage dips caudad to a considerable extent.

The first four tracheal rings (fig. 4) are very closely related to each other. They are partially fused along their sides as well as firmly bound together by dense fibrous tissue. Their ventral and dorsal extremities are free, but the spaces between these


extremities are very narrow. This arrangement gives a verystrong wall to this portion of the trachea, and because of this specialization this portion is known as the tympanum. The most caudal, or first, tracheal ring is smaller than the other three and is transitional in size and shape between the tracheal rings immediately above it and the intermediate syringeal cartilages immediately below it. As a whole, the tympanum arches distinctly cephalad, its ventral and dorsal extremities projecting caudad.

For the proper understanding of the arrangement of the remaining cartilages a description of the pessulus (fig. 3) here becomes necessary. In size, the pessulus far exceeds all other skeletal parts. It is a well developed bar of bone located at the junction of the bronchi, and lying dorso-ventrally in a plane transverse to the long axis of the trachea. Its ventral and dorsal extremities are large and serve for the attachment of some of the cartilages referred to above.

The bone as a whole may be described as consisting of a shaft and two large extremities (fig. 3). The shaft corresponds to a little more than the middle third of the bone and in shape resembles a three-sided prism with rounded borders. The cephalic porder projects into the lumen of the trachea, and marks the point of its bifurcation. The lateral surfaces of the shaft form the medial walls of the cephalic ends of the bronchi, between the diverging courses of which the basal surface lies.

The ventral extremity is very large and may be considered as being a pyi-amid whose apex points cephalad. On this pyramid three distinct borders, a dorsal and two lateral; three surfaces, a ventral and two dorso-lateral ; and a base, may be described. The dorsal border is continuous with the cephalic border of the shaft and projects into the lumen of the trachea. The two lateral borders give attachment to the intermediate syringeal cartilages which will be described presently. The ventral or anterior surface is broad and sUghtly convex and stands out so prominently that when the syrinx is viewed as a whole, this surface projects farther ventrad than any immediately neighboring skeletal element. The caudal portions of the dorso-lateral sur


faces, like the lateral surfaces of the shaft of the pessulus with which they are continuous, form the medial walls of the cephalic ends of the bronchi, while the cephalic portions of these same dorso-lateral surfaces form a part of the ventral wall of the trachea. The base, or basal surface, lies in the same plane as the corresponding surface of the shaft of the pessulus, and like the latter surface can be plainly seen in a \'iew of the caudal aspect of the entire syrinx.

The dorsal extremity of the pessulus, like the ventral, is a threesided pyramid with its apex pointing cephalad. A ventral border and two lateral borders, a dor.sal surface and two ventrolateral surfaces, and a base, or basal surface, are to be noted. With the exception of the dorsal surface, the surfaces and borders of this extremity have the same corresponding relations as those of the ventral extremity. The dorsal surface lies in close relation to the ventral wall of the oesophagus. To no part of this extremity, however, do the intermediate syringeal cartilages find attachment.

The first bronchial half -ring is attached at both ends to the lateral borders of the extremities of the pessulus. At its dorsal end this connection is composed of fibrous connective tissue. At the ventral end the union is made by means of cartilage. In old age, however, these attachments become ossified. In females of one or two years the half-rings are united to the ventral extremity of the pessulus by young cartilage which appears much lighter than other cartilaginous portions, when stained with hematoxyhn.

The four cartilages w^hich . occur between the tympanum and the first bronchial half-rings (fig. 4), and which have been designated above as the intermediate syringeal cartilages, appear to differ structurally from all the other cartilages in the entire respirator}^ tract. They are not complete rings nor are they halfrings. Ventrally they are continuous with the pjTamid of the pessulus, while dorsally they are free. The most cephahc ones are smallest and the most rudimentary. Ventrally they proceed from the very apex of the pyramid of the pessulus, extend along the sides of the syrinx as thin flat bands, and end before


reaching its dorsal wall. The second, enumerated caudad from the tympanum, are somewhat broader and extend farther dorsad. The first and the second cartilages are often fused for a part of their course. The third intermediate syringeal cartilages differ from the second only in being slightly broader. The fourth and the last of this series present more striking differences. They are larger than any of the preceding ones. Their free dorsal extremities lie laterad of the dorsal pyramid and gradually become very broad, but their medio-lateral diameter is not increased. These cartilages are arched slightly cephalad.

Caudad to the intermediate syringeal cartilages are the bronchial half-rings (fig. 4). The first of these differ from the typical half-rings in the following respects : first, they are longer and have much greater diameters; second, both ends are attached to the pessulus; third, well marked caudal prolongations extend from their ventral extremities and connect them with the second half -rings; fourth, their ventral extremities in old individuals become ossified; fifth, they are arched, with their convexities pointing caudad. .

Between the first half-rings and the last intermediate cartilages are very large, somewhat oval-shaped spaces, the upper thirds of which are directly opposite to the pessulus (fig. 4).

The second half-rings, which are somewhat smaller than the first, have enlarged ventral extremities which are attached to the cartilaginous prolongations of the first half-rings mentioned above. Their dorsal extremities are free.

The third bronchial half-rings are quite similar to the typical half -rings which lie caudad to. them. The only noticeable modification is an enlargement of their ventral extremities. They are somewhat flattened bands of cartilage which are free at both ends.

b. Vibrating membranes. Certain parts of the highly modified skeleton, just described, serve as attachments for specialized semitransparent membranes, which are the real voice producing elements. Two pairs of these are present in the chicken. They are known as the external and the internal tympanic membranes. Another membrane, not strictly a vibrating membrane, will also be discussed under this heading.


It should be recalled that the first half-ring of each primary bronchus and the last intermediate syringeal cartilage arch caudad and cephalad respectively, thus forming two (one on each side) oval spaces, occupied by the external tympanic membranes (fig. 2). Caudally these membranes are attached along the entire length of the cephalic borders of the first half-rings; cephalically they are attached along the entire length of the caudal borders of the last intermediate cartilages. Dorsally and ventrally the membranes narrow and are attached to the lateral borders of the corresponding pyramids of the pessulus. The internal tympanic membranes are situated more caudad (fig. 2), in the medial walls of the bronchi just below the tracheal bifurcation and stretch between the dorsal and ventral free ends of the half-rings. Cephalically they are attached to the lateral borders of the shaft of the pessulus. Caudally they extend to the level of the third half-rings where they are replaced by the heavy fibrous tissue of the medial bronchial walls.

Stretching across the interbronchial interval, just caudad to the internal tympanic membranes, is another membranous structure, the bronchidesmus (figs. 1 and 2) of Garrod. The middle portion of this is somewhat narrower than its ends, and may be said to he between the level of the fifth and the level of the eighth bronchial half-rings. Its ends, exceeding these levels and extending for a variable distance above and below them, are attached correspondingly to the dorsal third of the medial bronchial walls. The cephahc portions of these attachments show a modification. As these portions of the ends of the bronchidesmus extend to the attachment stated, they expand, dorsoventrally, and find further attachment as far ventral as the ventral third of the membranous medial walls of the bronchi and as far cephalad as the level of the third bronchial half-rings. This last attachment may be said to mark the caudal limits of the internal tympanic membranes. The caudal border of the bronchidesmus is broader than its cephalic border and presents two oval openings, one on each side of the median plane, which lead into two smaller irregular pockets within the bronchidesmus.


In the interbronchial region, in immediate relation to the caudal border of the bronchidesmus, numerous folds and pouches are developed. In the mid-line, the dorsal surface of the bronchidesmus is attached, by means of a thin sheet of fibrous tissue, to the ventral surface of the oesophagus and the dorsal portion of the cephalic part of the pericardium. From these attachments this same sheet of tissue extends ventrally, and is further attached to the caudal border of the bronchidesmus, which thus receives additional anchorage by the sheet of tissue also being attached to the ventral portion of the cephalic part of the pericardium. From these last named attachments the sheet of tissue extends cephalad in the median plane, and tapers to a narrow cord. The cord, after a short course, again expands, this time to gain attachment along the entire extent of the junction of the ventral surface and the base of the ventral pjTamid of the pessulus. Ventrally this cord and its expanded portions are attached throughout the entire length to a reflection of the fibrous pericardium surrounding the roots of the great vessels of the heart.

c. Musculature. In general the musculature of the syrinx of birds is composed of intrinsic and extrinsic muscles. Many birds have from three to seven pairs of intrinsic syringeal muscles. The frequent occurrence of these intrinsic muscles has been recognized of value in the classification of birds. Species possessing them are usually songsters. There are some exceptions, however, such as the crow and the jay, which, though possessing these intrinsic muscles, are no longer classified as singing birds. From the syrinx of the domestic chicken the intrinsic syringeal muscles are absent, and hence its voice organ is of course correspondingly simpler. Because of their absence from the syiinx of Gallus domesticus the intrinsic syringeal muscles merit no further description here.

The principal extrinsic syringeal muscles are the tracheal muscles known as the sterno-tracheales (fig. 2). In the chicken they are well developed muscles which take origin, one on each side, from the antero-lateral process of the sternum. Their fibers are directed obliquely ventrally and cephalad, and reach


the trachea at the level of the tenth ring (fig. 2). Here, as beyond, both muscles are firmly bound to the trachea by a strong common fascial sheath. But despite their fascial attachment, none of their fibers is inserted at this level. Springing from the medial surface of both muscles are several small bundles of fibers which, directed obliquely cephalad, are inserted on the tracheal rings, from the thirteenth to the twenty-first. Cephalad to these muscular bundles both muscles, now much diminished in size, proceed along the sides of the trachea to their final insertion on the ventral cartilages of the upper larynx. The exact extent and manner of insertion of the sterno-trachealis muscles offers a field for further investigation. The action of these muscles will be discussed in the section on function.

Two pairs of short bundles of muscle fibers are found on the caudal end of the trachea. One pair (fig. 2) lies on the ventrolateral aspect of the trachea, the other on its dorso-la*teral aspect. Caudally these bundles end at the levels of the fifth to eighth rings. Cephalically the majority of their fibers enter, and apparently become a part of, the ventral and the dorsal margins of the sterno-tracheales respectively. The fibers which do not enter the sterno-trachealis muscles are inserted on the respective ventral and dorsal walls of the trachea. Some of the fibers of the ventro-lateral bundles spread out cephalad to the level of the twelfth ring, to cover the ventral surface of the trachea, and certain fibers from the dorso-lateral bundles also spread out to cover its dorsal surface.

d. Sexual differences. Sexual differences are very marked in song birds, especially as regards the size of the labia and the syi'ingeal muscles. In male aquatic birds the tympanum is a large bony swelling projecting from the left side of the trachea, while in females the tympanum, though present, does not project beyond the tracheal walls. The main sexual difference to be noted in the s>Tinx of chickens is one of size (cf. figs. 5 and 6 for which males and females of the same varieties were chosen). Male birds are usually larger than females, and so some difference would naturally be expected in the size of their voice organs. But this difference in size of individuals is not sufficient


to account for all the differences observed. Quite naturally the sterno-tracheal muscles are noticeably smaller in females; but the bony rings above the tympanum, these, in the male (fig. 5) are entirely different in size and shape from the corresponding rings in the female (fig. 6) . Again, in the male the tympanum is composed of the first four tracheal rings, while in the female only the first three form this structure. The tympanic membranes, however, exhibit no marked difference.

It is not an uncommon thing to hear a female chicken try to crow. The sounds she produces are distinct, and cannot be mistaken for anything other than an attempted crow. Judging from the structure of the syrinx, there is no apparent reason why the female should not be able to crow perfectly, provided the instinct for it were properly developed. Hacker ('00) calls attention to the fact that the females of certain species of song birds, when kept in captivity, learn to sing as charmingly as the males. Barrington (1773) gives numerous examples of one species being trained to sing the song of another species.

2. Microscopic anatomy

For this phase of the investigation, cross, coronal, and longitudinal sections were made through the sj^inx of both male and female adult chickens. The sections were cut five, ten, fifteen, and twenty micra in thickness. The following stains were used : iron hematoxyhn, picrofuchsin, Malloru's aniline-blue connective tissue stam, Weigert's resorcm-fuchsin elastic tissue stain, Unna's orcein, and Mayer's mucicarmine.

In figure 5 all of the rings, the half-rings, the cartilages, and the pessulus are seen in cross section. It is to be noted that the tracheal rings cephalad to the second are completely ossified. Each consists of a thick peripheral layer of compact bone mthin which is a central area of cancellous bone. A distinct periosteum surrounds the whole. The cavities of the cancellous portions contain bone-marrow not unlike that found in other bones of birds.

As stated before, the first fom- tracheal rings are united to form the tympanum. The first two rings are composed of hya


line cartilage throughout; the third and the fourth are entirely of bone. The remaining rings of the trachea are also bone, but each is separated from the next by a small space, which is bridged by a heavy band of fibrous tissiie which, on reaching the rings, becomes continuous with their periosteum.

The first and the second intermediate syringeal cartilages are small, thin, cartilaginous bars, each having a very thin and indist'nct perichondrium. These cartilages are very rudimentary in every respect. The and the third fourthin termediate cartilages are much better developed and possess well marked perichondria.

Each intercartilaginous space, from the caudal border of the tympanum to the cephalic border of the fourth intermediate cartilage, is occupied by a strong thick band of fibrous tissue. As determined, after staining with Weigert's elastic tissue stain, these bands are composed almost entirely of elastic fibers.

As seen in figure 5, the fii'st and the second bronchial halfrings are cartilaginous, but as pre\dously mentioned, the ventral extremity of the first becomes ossified in old individuals.

The pessulus appears triangular in cross section, and is ossified throughout its entire extent. The marrow cavity is usually quite extensive, and is surrounded by a comparatively thin layer of compact bone. The lateral surfaces of the pessulus, as well as its cephalic and lateral borders, are covered by the mucous membrane of the respiratory tract. A distinct areolar tunic, or submucosa, especially marked over the cephalic border, connects the respiratory membrane to the periosteum of the pessulus. A somewhat different arrangement is found on the caudal surface of the pessulus. ^ Here an areolar coat, continuous with the bronchial submucosa, connects the periosteum of the pessulus with the membranous reduplications in the interbronchial interval iinmediately cephalad to the bronchidesmus.

Above the tj^mpanum the epithelium of the tracheal mucous membrane is of the stratified ciliated columnar varietj^, with a distinct basement membrane. The tunica propria is well developed and contains many lymph cells as well as numerous distinct lymph nodules. These nodules, which are most numerous


in the region of the tracheal bifurcation, are quite near the surface, since they he just beneath the epithehum. Where these lymph nodules occur the epithelial cells are non-ciliated.

A distinct submucosa is present throughout practically the entire extent of the mucous membrane of the trachea and the bronchi, and connects this mucosa to the interannu'ar membranes and the periosteum, or the perichondrium of the tracheal rings, •the intermediate syringeal cartilages, and the bronchial half-rings. There is no sharp line of demarcation between the submucosa and the tunica propria, but the former is easily distinguished from the latter in that it is less dense and contains fewer lymph cells. Over the tympanic membranes the tunica propria s thin and sparse and a submucosa is indistinguishable.

At the level of the caudal border of the third tracheal ring, the epithelial cells loose their cilia and immediately become columnar in shape. There is a gradual transition from stratified ciliated columnar epithelium to stratified squamous. The latter is made up of several layers at the lower border of the tympanum, but becomes thinner and thinner, as far as a point slightly caudad to the fourth intermediate syringeal cartilages, where it consists of only a single layer of flat cells, covering the internal, or medial, surfaces of the external tympanic membranes. The lateral surfaces of the internal tympanic membranes, or those surfaces of these membranes which face the lumen of each primary bronchus, are covered by a similar simple epithelium. Below the limits of the external tympanic membranes the epithelium gradually becomes thicker, and about midway between the first and second bronchial half-rings it again assumes the characteristics of a ciliated columnar epithelium . .

The disposition of the squamous epithelium suggests that the portions of the sja-ingeal walls covered by it are subject to considerable movement.

As stated above, the vibrating membranes are covered by a thin epithe'ium composed o a single layer of flattened cells; beneath this is the tunica propria which is thin and ess distinct than that found in other parts of the respiratory tract. Beneath the tunica propria is a thin, dense layer of tissue which comprises



the vibrating membranes proper. This layer is composed of numerous white and elastic fibers. It is to be noted that there is a tendency for all layers to become much thinner as the tympanic membranes are approached (fig. 5).

The external and the internal labia, which occur in song birds in connection with the tympanic membranes, were not observed in the chicken.

The semilunar membrane is poorly developed in the chicken. It is nothing more than a modification of the mucosa over the cephahc border of the shaft of the pessulus. The epithelium in this position is thickened and belongs to the stratified ciliated colunmar type. The submucosa too, as already mentioned, is especially thick. Thus it is evident that the semilunar membi ane in the chicken is but a poor representative of this structure as found in many song birds.


For this division of the present investigation the following selected stages were used.

Period of Incubation

68 hours

152 hours

272 hours

70 hours

164 hours

284 hours

72 hours

176 hours

296 hours

74 hours

188 hours

308 hours

78 hours

200 hours

.320 hours

80 hours

212 hours

332 hours

92 hours

226 hours

356 hours

104 hours

236 hours

380 hours

128 hours

248 hours

404 hours

140 hours

260 hours

452 hours

Twenty-four hours after hatching.

Serial cross sections were made of nearly all the stages and serial coronal sections were made of a large number of the stages. In the early stages series were made of the entire embryo whib later only the syringeal region was sectioned. For the most part the sections were cut ten micra in thickness. In the more important stages a drawing of the embryo was made before sectioning, which was utilized in obtaining the plane of sectioning


and in obtaining the curvature of the backUne of the embryo for reconstructions. Most series were stained with iron or alum hematoxylin, with picrofuchsin or eosin as counter stains. Some of the later stages, however, were stained with Mallory's anilinblue connective tissue stain, Weigert's resorcin-fuchsin elastic tissue stain, and Mayer's mucicarmine.

In transverse sections of a 68 hour embryo, the epithelial tube of the foregut is surrounded by a condensed mass of mesenchyma. This tube, which in its most cephaUc part is cylindrical, when followed caudad, enlarges and becomes triangular in shape. Extending ventrolaterally from this tube are two beginning diverticula, the anlagen of the trachea and the bronchi. In length the right is 0,05 mm. while the left is 0.03 mm.

Embryos of 74, 78, and 80 hours show a lengthening of the trachl^a and the bronchi, but the former is still relatively much shorter than the latter. Wunderlich called attention to this condition in Fringilla domestica and mentioned the fact that it represents a stage which is found throughout life in certain reptiles. This might be regarded as a homology, but such homology is not probable. It would seem that a logical explanation of this condition might be found in a study of the position and the relations of the organs of this region. Since the neck as such has not yet developed it is obvious that the trachea springs from the digestive tube and bifurcates at once at a point not far from the future position of the lungs. As the neck lengthens, the oesophagus and the trachea must keep pace. The trachea, therefore, becomes drawn out. This elongation goes on to such an extent in the chicken that the trachea eventually becomes much longer than the bronchi.

A 128 hour embryo shows a marked increase in the length of the trachea and the bronchi, and in a 140 hour stage these structures have nearly doubled in length. The right bronchus courses laterally and gives off seven sac-like branches which later become the so-called pipes of the lungs.

Since the tracheal bifurcation is the region of especial interest in this paper, the general development of the trachea and the bronchi will not be traced further.


1. Mucous membrane

In the early stages described above, the walls of the trachea and the bronchi consist of an epithelial tube surrounded by a loose mesenchyma. In the 68 hour stage the epitheUum of this tube may be said to belong to the stratified variety. The boundaries between the cells are not well defined, but two or three layers of nuclei can be observed. The cells lining the lumen are mostly columnar with their nuclei placed in the ends distal to the lumen. The basal cells are much shorter and more irregular in shape. Each bronchus presents a well defined lumen.

No noticeable differences occur in the structure of the tube in embryos of 72, 74, and 78 hours. In an 80 hour embryo, however, the beginning of the basement membrane is seen as a small differentiated line which extends around the outside of the epithelial tube. Immediately below its origin from the pharynx the tracheal tube has a lumen of oval outline. More caudally it elongates from side to side until immediately above the bifurcation it is little more than a transverse slit. At this level the dorsal epithelial wall becomes thicker, while the ventral wall appears to fold inward and to come into contact with the dorsal wall, thus dividing the one slit-like lumen into two. The lumina of the two bronchi and the trachea have a smaller diameter at 80 hours than at any of the earlier stages. The epithelial walls have not thickened at this stage, but the mass of mesenchyma surrounding the digestive and respiratory tubes has condensed.

In 128 and 140 hour embryos, the lumen of the tracheal tube is so small as to seem almost obliterated in some sections. The thickness of the dorsal wall is nearly double that of the ventral. About 0.1 mm. cephalad to the bifurcation, the tracheal tube becomes compressed dorso-ventrally, thus a slit similar to that described above is produced. It possesses, however, a somewhat larger lumen.

The lumen of each bronchus possesses a greater transverse diameter than that of the trachea, except in that region immediately above the bifurcation of the latter. The dorso-medial portion of each bronchial wall is much thicker than any other part.



This seems to be due to a greater number of cell layers rather than to the elongation of the cells, since four distinct layers of nuclei appear in the thickened part, while only 2 or 3 layers are present in the thinner portions of the wall. The same condition explains the thickening of the dorsal wall of the trachea. A very distinct basement membrane is present in these stages.

In 152 and 164 hour embryos the lumen of the trachea is still very small. At a point 0.24 mm. above the bifurcation it begins to enlarge, not only in the transverse diameter, but in the dorsoventral as well. A cross section of the lumen and walls of the tube at this level presents a figure which is almost a perfect square, and is much larger than that described in previous stages. Of the walls composing the square, the dorsal is thickest, th.e ventral thinnest, while the lateral are intermediate in this respect. The dorsal wall possesses four layers of nuclei, the lateral walls three, and the ventral wall two layers.

Figure 12, from an embryo of 176 hours, shows the relative thickness of the epithelium in the region of the tracheal bifurcation. The epithelium is thickest just cephalad to the bifurcation, where not less than three or four distinct layers of nuclei are visible. At the level of, and shghtly caudad to the bifurcation, the number of cells gradually diminishes, until only two layers of nuclei are present. The epithelium projecting into the tracheal lumen at the point of the bifurcation appears slightly thicker than that lining the medial bronchial walls with which it is continuous. At this stage, too, slight irregularities or evaginations become noticeable in the inner walls of the epithelial tube. These are due to encroachments upon the tube, from without, of growing mesenchymal condensations which lie immediately beneath and in direct contact with the basement membrane of the epithehum. Only two of these evaginations have developed to a stage worthy of mention. They are somewhat caudad to the tracheal bifurcation, one projecting into the lumen from the lateral wall of each primary bronchus. Caudad to the site of these evaginations the epithelium again thickens and for some distance three layers of nuclei can be distinguished.


In a stage 24 hours later (200 hours), as well as in a 212 hour embryo, the epithelium shows a noteworthy characteristic. The nuclei of the cells bordering upon the lumina have migrated from the basal to the free ends of the cells, thus giving rise to a very prominent superficial row of nuclei. Immediately beneath this the other nuclei show an irregular arrangement. Although this characteristic arrangement of nuclei is noticeable in the 176 hour stage, it does not become very striking until the present stage is reached. In this stage also a second evagination (fig. 13) appears in the lateral wall of each bronchus, just cephalad to the one described as present in the preceding stage. At 248 hours many of these evaginations may be observed in the walls of both the trachea and the bronchi. In the latter they are confined, however, to the lateral walls.

A cross section of a 332 hour embryo shows two or three layers of nuclei in the epithelium, except in the region of the future t5'mpanic membranes, where a single laj^er occurs. The cells in these regions are cuboidal. Throughout the region of the bifurcation of the trachea, a considerable number of goblet cells, as well as numerous cells exhibiting a somewhat differentiated clear area in their free ends, occur in the epitheUal layer nearest the lumen. Numerous cilia are present on the free borders of some cells.

In a 404 hour embryo a very distinct layer of nuclei appears at the very base of the epithelium. These are much smaller than those found nearer the lumen. As the regions of the future tympanic membranes are approached, all cells between the layer of small nuclei are left to cover the membranes. These cells differ from those described in the 332 hour embryo in being much less cuboidal, smaller and possessing much smaller nuclei.

The most noticeable differences between the last embryo and one of 452 hours are (1) the nuclei of the cells covering the tympanic membranes have become much more flattened and in appearance suggest the nuclei of endothelial cells; (2) in addition to individual goblet cells, small groups consisting of three or n.ore eecreting cells are found in the epithelium. Each of these groups is slightly invaginated, leaving a shallow pit connecting


with the lumen of the respiratory tract. Each cell appears light and presents a pore-like opening which communicates with the common pit. The nuclei of these secreting cells are situated nearer the basement membrane than those of the typical epithelial cells. These structures are not found in the epithelium, covering the tympanic membranes. In a chick 24 hours after hatching, such groups of secreting cells are much more numerous and show a higher degree of development in having become more invaginated. As determined by staining with mucicarmine it is evident that these structures are the anlagen of the simple mucous alveolar glands found in this region. Well developed cilia project from the free borders of the typical columnar cells, but they are not seen in the region of the tympanic membranes.

The vacuoles, which occur in such great abundance in the mucous membrane of the digestive tube in human embryos (Johnson '10), were observed in the respiratory tube of chick embryos, but they are neither constant or abundant. In a 280 hour embryo, buch vacuoles are quite numerous in the epithelium lining the medial bronchial walls, but in all other stages they are practically absent.

2. Cartilaginous skeleton

Ir. an 80 hour embryo the epithelial tubes constituting the trachea, the bronchi, and the oesophagus are all closely surrounded throughout their entire extent by a common area of mesenchyma which, in immediate relation to them, is much denser than it is a little beyond them. An inspection of figures 12, 13, and 7 will make clear this arrangement of the mesenchyma.

In the 128 hour stage that portion of the condensed area of mesenchyma surrounding the respiratory tract has become separated from that surrounding the digestive tube. They now appear as two distinct areas, that surrounding the trachea being the smaller. The condensed mesenchyma is, of course, surroimded in turn by loose, unmodified mesenchymal tissue.

In a 152 hour embryo the mesenchymal condensation, common to the entire respiratory tract, becomes more proininent in the region of the tracheal bifurcation than in any other portion.


Along the dorsal surface in this region the condensation is no thicker than in the preceding stages, but it is much thicker lateral and ventral to the bifurcation (fig. 7). It also extends between the bronchi, just caudad to their union with the trachea. In a 164 hour embryo, at the points in the lateral bronchial walls where the above mentioned evaginations of the epithelium occur, the aggregated mesenchymal cells arrange themselves into very compact areas. These areas appear quite round in cross section. In the center of each area are several unmodified cells which are immediately surrounded by others of the same type concentrically placed. This arrangement of mesenchymal cells is the beginning of the anlagen of the first bronchial half-rings which, of all the skeletal elements, are the first to appear. This anlage extends from the ventro-lateral to the dorso-lateral parts of each bronchus, but appears best developed in its middle portion. These anlagen are more marked in a 176 hour embryo and are represented in figure 7, as two rounded and elongated masses projecting dorso-ventrally around the lateral walls of the epithelial tubes of the bronchi a httle below the tracheal bifurcation. Twelve hours later the mesenchymal cells just cephalad to thes3 anlagen for the first bronchial half-rings become arranged so as to form two other areas of aggregated cells similar to the ones just described. These areas are the anlagen of the fourth intermediate syringeal cartilages, which at this stage he very near those of the first bronchial half-rings.

At 200 hours the first half-rings, and the fourth intermediate cartilages stand out much more prominently, since the mesechyma immediately surrounding them is now less dense (fig. 13). The fourth intermediate syringeal cartilages differ from the first half-rings only in extending farther ventrally. The anlage of the pessulus appears very faintly in the mesenchyma between the bronchi, and slightly caudad to the tracheal bifurcation. It occupies the ventral three-fifths of the distance from the ventral to the dorsal walls of the respiratory tube in the region of its bifurcation. In structure it does not appear different from the above described anlagen, but it is much larger than either of them. A V-shaped mass of condensed mesenchyma extends from


the cephalic border of the pessulus to the point of union of the two bronchi.

In a 226 hour embrj^o, the cells in the center of the anlagen have no definite arrangement, but are larger and more o^'al in shape than those in the preceding stages. The central cells, in enlarging, have crowded the outer cells together, thus causing the individual outer cells to appear flattened and the whole peripheral area to appear narrower. The first half-rings and the fourth intermediate syiingeal cartilages are now separated by a greater space than in the stages pre\dously described. The anlage of an additional cartilage appear just cephalad to the developing fourth intermediate syringeal cartilages. Though small, as yet, these anlagen, the beginnings of the third intermediate syringeal cartilages, are fused \Aith the cephalic extremity of the ventral pyramid of the pesulus. Dorsally these anlagen extend only slightly more than one-half of the distance to the dorsal extremity of the pessulus where they end unattached. These structures appeared in only one out of two embryos taken at this stage. These anlagen are shown in figures 8 and 9, which make clear the relation and the extent of these and other developing skeletal elements. The first bronchial half-rings extend around the lateral walls of the bronchi but do not reach the pessulus at either extremity. The fourth intermediate syringeal cartilages arch cephalad to some extent and are fused \nih. the ventral pyramidal extremity of the pessulus. Their dorsal ends are unattached and lie somewhat lateral to the dorsal pyramid of the pessulus. The developing pessulus has gradually assumed the same form as in the adult, except the shaft, which at this stage is round in cross section, rather than triangular. At this stage, too, the mesenchyma along the entire ventral surface of the trachea has become. more condensed. This condensation represents the primordium of the anlagen of the tracheal rings.

The 248 hour stage shows a marked advance in the development of the skeletal parts (fig. 14) All the tracheal rings are represented by circular (in cross section) condensations of mesenchyma which extend around the lumen, immediately external and in contact with the epithelial tube. The first and the second


intermediate cartilages are here represented by extremely small masses which lack the regular cell arrangement found in the others. Ventrally they fuse with the ventral pyramid of the pessulus. Immediately caudad to the first half-rings is a mesenchymal area, which is apparently condensing to form the anlage for the second half-ring. None of the other bronchial cartilages are represented at this stage.

At 260 hours the central cells of the anlagen for the first halfrings, the fourth intermediate syringeal cartilages, and the pessulus appear slightly more separated from each other. As would be expected from the order of their fu'st appearance this is most marked in the anlagen for the first half-rings, where true embryonal cartilage is closely approached, and least marked in the pessulus. The first seven half-rings are represented at this stage.

All the cartilaginous and the bony skeletal elements of the respiratory tract are represented in an embryo of 284 hours. The pessulus, the first and the second half-rings, the tracheal rings (except the fu'st and the second) and the fourth intermediate syringeal cartilages are composed of embryonal or primary cartilage. The perichondrium is represented by a thin layer of cells which are very much elongated and rather sparsely distributed. All skeletal elements, other than those above mentioned, are still in the mesenchymal stage. Figure 10, drawn from a wax reconstruction of the syrinx of a 284 hour chick, shows the extent, the size, the form, etc., of the cartilages, and their relation to the epithelial tube. All the intermediate syi'ingeal cartilages are fused to the ventral pyramid of the pessulus. The fourth are much the largest of these and closely approach the first half-rings in size. They present a decided cephalic arching in the middle portion of their course, and each shows a marked enlargement, which ends freely, but in close relation to the middle third of the dorsal pyramid of the pessulus. The third, with a diameter about one-half as great as that of the fourth, and placed slightly cephalad to the latter, pass around the epithelial tube and present free dorsal extremities, which he cephalad and lateral to the apex of the dorsal pyramid of the pessulus. The second are much smaller and take a course parallel to that of the


third. At this stage, the first are not fused with the ventral pyramid, but are continuous with a mass of condensed mesenchyma which may be said to be a part of the apex of the pyi'amid. They pass around the trachea and end freely on its dorsolateral aspect.

The first two tracheal rings are very small and, like the first three intermediate cartilages, are composed entirely of condensed mesenchyma. At both their dorsal and their ventral extremities they present a marked caudal dipping The second tracheal ring is, however, somewhat larger and more advanced in its development than the first. It should be noted, also, that the first four tracheal rings are not in direct contact with each other as in the adult ; but are separated by spaces which are occupied by embryonal fibrous tissue. The remaining tracheal rings call for no special description, since they differ from those of the adult only in size, in being circular in cross section, and in being composed entirely of cartilage. The ventral and dorsal extremities, or pyramids, of the pessulus, are cartilaginous throughout and are similar in shape to these parts in the adult. The shaft, however, as stated above, is round in cross section.

Second to the pessulus in size are the first bronchial half-rings. These half-rings in their middle portions, arch markedly caudad. Their ventral extremities are in close contact mth the base of the ventral pyramid of the pessulus, but fusion has not taken place. Dorsally they end unattached, caudad and lateral to the pessulus. The second half-rings, which are much smaller and immediately caudad to the first, are connected with the latter at their ventral extremities by means of a small mass of condensed mesenchyma. A wax reconstruction of the syrinx at this stage (284 hours) shows that the lower end of the trachea exhibits a slight compression (fig. 10). In the region of the intermediate syringeal cartilages this compression is well marked.

In a 320 hour embryo, all the skeletal elements are represented by primary cartilage. The first and the second intermediate syringeal cartilages are the last to be transformed into true cartilages, and at this stage are the smallest of all the skeletal


parts. These cartilages often appear fused into a single mass for at least a part of their extent.

In a 452 hour embrj^o, the last of the series, most of the tracheal rings cephalad to the tenth present two prolongations of precartilage, one of which is directed cephalad, the other caudad. The second to the tenth rings, inclusive, are now more or less perfect squares when seen in cross section. The third, the fourth, and the fifth, however, are much larger than the others (fig. 16). The first is roimd and still quite small. The intermediate cartilages also present a different appearance; being somewhat compressed, they now appear oblong in cross section. All of these cartilages are relatively much larger in all developmental stages than in the adult.

In cross section, the shaft of the pessulus is becoming slightly triangular, but it still does not possess the distinctly triangular shape of that of the adult. The first bronchial half-rings are only slightly compressed. At their dorsal extremities there are indications of fusion with the pessulus, while at their ventral extremities the two structures are separated onl}^ by their perichondria.

Wunderlich found the s;yTingeal skeleton of a 17 day chicken embryo to be composed entirely of cartilage. He states that these elements are cartilaginous in Fringilla domestica and Anas boschas at the time of hatching. Rathke states that at hatching ossification has begun in the sjTinx of the domestic chicken. In this investigation, however, no bone was observed in the s\Tinx of G alius domesticus at the time of hatching, nor in stages one day later.

3. Suhmucosa

In the earlier stages, the anlagen of the skeletal elements, except the pessulus, are in such close contact with the epithelial tube that they produce actual evaginations of the epithelium (fig. 13) . In a 284 hour embryo this is not true of the third and the fourth intermediate syringeal cartilages, and of the first half-rings; here, even typical mesenchymal cells occupy the considerable space between the above structures and the epithehal tube (fig. 15). The


mesenchyma referred to is continuous with, and seems to have migrated inward from, that occupying the space between the cartilages. The same condition exists between the tube and all the cartilaginous elements in the region of the tracheal bifurcation in a 296 horn- embryo. In a 332 hour embryo these intervening mesenchymal cells possess numerous protoplasmic processes which form a loose network and which, at the 356 hour stage, become smaller and resemble short connective tissue fibers. In these spaces between the developing cartilages and the epithelial tube small, but distinct, fibers appear in a 404 hour embryo. These are true connective tissue fibers since they stand out very prominently when treated with Mallory's connective tissue stain. At this stage, the tissue now forms a loose connection between the epithelium on the one hand and the cartilage and intercartilaginous structures on the other. It contains blood vessels and nerves and constitutes the submucosa. As in the adult, there is no sharp line of demarcation between the submucosa and the tunica propria.

4- Musculature

In the 164 hour embryo areas of differentiated cells may be observed between the ribs. Similar aggregates of cells extend from the antero-lateral process of the sternum to the trachea, somewhat above its bifurcation. The cells composing these aggregates have large oval nuclei, each of which possesses a distinct nucleolus ; and their cytoplasm, which has a strong affinity for eosin, appears to be drawn out into short strands. These structures are more prominent in a 176 hour embryo. Wunderlich states that the muscular system begins to differentiate on the tenth day in Fringilla domestica and on the twelfth day in Anas boschas.

In the 188 horn- stage numerous cytoplasmic strands have united to form long bundles which consist of minute fibrils. These bundles are present throughout the entire extent of the developing muscles, from the sternum to the cephalic end of the trachea. The nuclei are oval and several seem to belong to each bundle of fibrils. At the point where these bundles, or the


sterno-tracheal muscles, reach the trachea small additional masses of developing muscular tissue extend caudad along the dorsolateral and ventro-lateral aspects of the trachea. These small additional masses seem to arise from, and certainly do not differ from, the tissue found in the developing sterno-tracheal muscles.

In the succeeding stages the muscles continue to develop by the addition of more cytoplasmic bundles. The nuclei of the muscle cells are easily distinguished from those of the developing connective tissue cells, which are present between the muscular bundles, for the former are more elongated and regular in shape. No other structural changes call for special mention until the 296th hour is reached, when faint cross striations appear in certain fibers. These, however, are quite rare.

At 452 hours the fibers have arranged themselves into more definite bundles. The fibers of these bundles show distinct cross striations, and numerous elongated nuclei appear to lie on or between them. Certain of these fibers are attached directly to the perichondrium of the lower tracheal rings.

5. Inter cartilaginous membranes

When the aggregated mesenchymal cells arrange them.selves to form the anlagen of the skeletal elements, small spaces exist between these anlagen. These spaces, also, are occupied by typical mesenchymal cells, but here they are fewer and farther apart (fig. 13). This condition is particularly well marked in a 248 hour embryo, in which it occurs between all of the developing cartilages except the first three intermediate syringeal cartilages. The widest spaces are those between the first half-rings and the fourth intermediate sjTingeal cartilages. In these spaces the mesenchyma shows a slight condensation, suggesting the arjagen of the external tympanic membranes. In a 260 hour embryo the cells lying near the periphery of this developing membrane have begun to elongate. Differentiation becomes more marked in each of the succeeding stages (272, 284, 296, 308, and 320 hours), but it is not until the 332 hour stage that small connective tissue fibers can be disclosed by the use of Mallory's stain.


Extending caudad from the lateral borders of the shaft of the pessulus in the 404 hour stage, and taking part m the formation of the medial bronchial walls, a well marked strand of connective tissue fibers is ssen. A large number of these fibers become continuous with, and enter into the formation of, the bronchidesmus, which now stretches between the medial walls of the two primary bronchi.

On comparing figures 5 and 16 it will be seen that the tympanic membranes have not as yet reached a very high stage of development. From their structure and their sizs one would expect them to be quite inefficient as \'ibrating membranes, and capable of producing only very simple sounds. The simple sounds actually produced by the young chick, seem to verify this deduction.

At the time of hatching the various structures described are not essentially different, except for size, from those of the last stage (404 hours) discussed. It thus appears evident that the syrinx is very immature at the time of hatching. Between this time and the adult stage, the four most caudal tracheal rings unite, in the manner previously described, to form the tympanum; the pessulus and the tracheal rings cephalad to the second become ossified; and finally, the tympanic membranes become more like those of the adult. In addition to becoming thinner, these membranes become more extensive and their corresponding fibers become more longitudinally arranged.

6. Sum7nary of development

The first indication of the respiratory tract was observed in a 68 hour embryo in which the laryngeo-tracheal groove and the developing bronchi were present. The trachea is at first very shprt and, like the bronchi, is composed of an epithelial tube and a loose smTounding mesenchyma. The bronchi are relatively much longer than the trachea in the beginning. Owing, however, to the rapid development of the neck, the trachea lengthens very rapidly and is much longer than either of the bronchi in a 128 hour embryo. The epithelium is stratified and, at 80 hours, rests upon a well defined basement membrane.


Immediately above the bifm*cation, the lumen of the tube is at first slit-like in shape, but in later stages, 152 and 164 hours, it is almost square in cross section. Its dorsal wall is about two times as thick as its ventral wall.

The developing cartilages produce slight evaginations of the epithelium of the respiratory tube. In a 248 hour embryo these evaginations occur along the entire length of the trachea and the bronchi.

At 332 hom-s only a single layer of cuboidal epithelial cells covers the future tympanic membranes, while immediately cephalad and caudad to these membranes two or three layers of nuclei can be distinguished. A distinct layer of small nuclei appears at the base of the epithehum. In the tympanic regions this is the only layer of nuclei present. The cells possessing them are slightly cuboidal, but at 452 hours they become very flat.

Certain cells resembling goblet cells, and other cells with clear cytoplasm, are present in the 332 hour embryo in all portions of the epithelium, except that covering the tympanic membranes. They are quite numerous in the semilunar membrane. At 452 hours certain groups of secreting cells are found which, when stained with mucicarmine, and when traced through succeeding stages, were determined to be simple mucous alveolar glands.

Owing to the thickness of the sections it was difficult to detect the "first appearance of cilia, but one can be reasonably sure of their presence in 248 and 260 hour embryos. They are very abundant in and following the 332 hour stage.

Vacuoles in the epithelium appeared in the 284 hour stage, but they were not very numerous in this embryo and are practically wanting in all other stages.

The skeletal elements are first represented by a condensation of the mesenchyma, which in a 152 hour embryo is most marked ventrally in the region of the bifurcation. The anlagen of the first bronchial half -rings appear just caudad to the bifurcation in the lateral walls of the bronchi, and are the only skeletal elements represented in the 164 hour stage. The next anlagen to appear are those of the fom-th intermediate syringeal cartilages and of the pessulus which become detectable in 184 and 200 hour


embryos respectively. The third intermediate syringeal cartilages are represented at 226 hours. All of the tracheal rings, the intermediate cartilages, and the first two half-rings are represented in the 248 hour embryo.

Cartilage cells were first observed in the 284 hour embryo, in the first half-rings, in the fourth intermediate cartilages, and in the pessulus. The first tracheal ring and the first and the second intermediate syringeal cartilages are the last to differentiate. Differentiation for these cartilages begins in the 320 hour embryo. No bone has developed at the time of hatching, nor has the tympanum been fully formed.

The submucosa begins to develop ag 284 hours. At first it is represented by mesenchymal cells, which later produce white and elastic fibers. These fibers, however, do not take Mallory's connective tissue stain until the 404 hour stage. At this time the submucosa contains numerous nerves and blood vessels.

The two sterno-tracheal muscles are quite well differentiated at 176 hours. At 188 hours the long cytoplasmic strands or processes of the developing cells are collected into bundle-like masses. Faint cross striations were first observed in a 296 hour stage. At 452 hours the muscles are not essentially different in structure from those of the adult.

The intercartilaginous membranes in early stages are not essentially different in structure from the submucosa. In a 260 hour embryo, the cells nearest their external surfaces elongate and later develop strands of connective tissue fibers which extend between the cartilages and attach to these membranes. Prominent bundles of such fibers extend caudally from the pessulus and enter into the formation of the medial walls of the bronchi, in a 404 hour embryo.

In early stages the external and the internal tympanic membranes do not differ in structure from the smaller intercartilaginous membranes of the trachea and the bronchi. At the time of hatching the tympanic membranes are quite thick. It is not until after this time that they become thinner and appear as true vibrating membranes.



This division of the present paper may seem almost unnecessary since Duvernoy, Girardi, Cuvier, and others of the older investigators conclusively demonstrated that voice is produced in the lower larynx. Kitchner ('85), and other observers, however, have expressed their doubts concerning the correctness of the conclusions arrived at by these early investigators. Because of such doubts it was thought best to include these experiments.

Experiment I. The trachea of an adult cock was divided at its middle, after which the bird was set free with others. After the operation crowing occurred quite frequently, but the voice was somewhat modified. In order to study these modifications more carefully, phonographic records were made of the crowing before and after cutting the trachea. These records were made in the folloAving manner. A normal adult cock was placed in a small room. After becoming accustomed to the new environment he had periods of crowing which were quite regular. An Edison phonograph was arranged just outside, so that the horn projected into the room through a small opening. This allowed the experimenter to operate the machine unseen. The early morning hours were found best for making records, as crowing was more frequent and regular at that time.

It is interesting to note that such birds usually crow about five or six times at intervals varying from ten to fifty seconds, then after an interval of fifteen minutes to one hour, they again begin to crow.

After a sufficient number of records had been made, the bird was deeply anaesthetised and a small part of the trachea exposed (fig. 17). This caused no noticeable difference in the voice. After sufficient recovery the trachea was treated with a local anaesthetic and cut entirely across. The cephalic end was tied tightly so there was no possible chance for air to pass from it through the upper larynx and the mouth. The caudal end was left open and allowed to protrude through the skin of the neck (fig. 18). The operation had no marked effect on the well-being of the bird, for, not more than two hours later this same cock.


when placed with some other chickens, was eating and crowing as though nothing had happened. Records of the crowing were made on the following morning. These records were preserve^ and carefully compared "with those taken before the operation. This comparison showed that after the trachea had been cut, tones were produced as before but the pitch was noticeably higher. It must be admitted, therefore, that the trachea and upper larynx serve to modulate voice, just as the pharynx and cavities in the mouth serve the same function in mammals.

Similar experiments were performed on the domestic duck with precisely the same results.

Through the kindness of Dr. Max Meyer, professor in experimental psychology, University of Missouri, the author is able to give here the exact changes in pitch which result from dividing the trachea. Dr. Meyer determined the pitch of the voice of a normal adult male. In this particular individual he found the normal tone was one of about 375 double vibrations per second. The pitch is constant from time to time. It is interesting to note that there is also but little difference in the pitch from the beginning to the end of crowing. It was observed, however, that the pitch is slightly higher at the beginning, there being a fall of not more than twelve double vibrations.

After the trachea was divided the number of double vibrations was increased to about 500 per second.

The voice in crowing is not interrupted, but is produced by one continuous flow of air causing vibrations of the tympanic membranes. Ordinarily a single crow lasts for about three or four seconds. But in one individual it was observed that after the trachea was divided crowing extended over scarcely a single second of time. As explained by Dr. Meyer, there are two possible reasons for the time of crowing being so much reduced. First, when the neck is stretched pain may result from the wound. Second, the tone produced is so unnatural that it is soon discontinued.

Experiment II. Since the lungs of birds are not elastic structures as in mammals, but are more solid, it is obvious that other organs with a large air capacity must be present, and further,


that such organs must be capable of exhaling ah' in greater or less amount at the will of the bird. Such structures are found in the air sacs which are present throughout the thoracic and abdominal regions and are continuous with the cavities of neighboring bones.

A chicken was anaesthetized, the humerus sawed through at its middle and a tightly fitting piece of rubber tubing placed over the central stump. When air was forced through the tube into the air sacs, by way of the humerus, sound was produced. This varied in pitch with the amount of air pressure used, greater pressure resulting in higher pitch. The trachea was then divided, l)ut this produced no noticeable result, since the bird was unable to control the length of the trachea or the width of the glottis by muscular contraction. Finally, the cer\dcal air sacs were punctured after which it was impossible to produce voice in this artificial manner, thus proving Herissaut's statement that when these sacs were ruptured birds are unable to sing. He explains this inabihty to sing as follows: Since the air sacs form air cavities around the syrinx, it can be seen that there is a tendency to equalize pressure on each side of the tympanic membranes, but when sudden gusts of air are forced out through the bronchi the equilibrium is disturbed. Thus the membranes are set into vibration. It is evident that all structm'es through which the voice passes from the syrinx to the exterior act as resonators. So far as size and extent are concerned, the trachea forms a large part of this resonating system.

Experiment III. In that part of this paper dealing with structure, it was noted that the sterno-trachealis muscles extend along the sides of the trachea from the twelfth ring to the upper larynx. When a pair of electrodes were applied to this muscle in an anaesthetized chicken, it was seen by its contraction to shorten the trachea from one-fourth to one-third of its original length. It should be recalled that the typical tracheal rings are so arranged that they may overlap when the muscles contract. The trachea may thus act as a pipe or horn capable of being lengthened and shortened. This has a direct influence upon pitch as was shown by the following experiment : The so-called thoracic cavity was


opened and one of the sterno-tracheal muscles was dhnded at .4 (fig. 2). After this operation the tones produced by the chicken were of a somewhat different quality. When both muscles were cut the quaUtj^ of voice was even more distinctly modified. Obviously the di^dsion of these muscles was responsible for the changes produced. Some weeks later this same chicken was anaesthetized and electrodes applied both directly to the sternotracheales and to the nerves supplying them. It was observed, udth some surprise, that the division of these muscles apparently had but little effect upon the actual shortening of the trachea. On stimulation the muscles immediately contracted and produced a marked shortening. By way of explanation, it should be recalled that the tjmipanum is attached to the pessulus and the first intermediate syringeal cartilages only by elastic membranes; that, although the ventral ends of the intermediate cartilages are attached to the ventral pyramid of the pessulus, the dorsal ends of these cartilages are unattached, and that all of these cartilages are connected with each other by the intercartilaginous membranes. Further, the external tympanic membranes are attached to the caudal borders of the fourth intermediate cartilages. Now, if the tympanum be drawn cephalad it is evident that the external tympanic membranes will, indirectly, be made more tense. Under normal conditions stimulation of the sterno-tracheales serves to shorten the trachea. They tend to pull the caudal end of the trachea cephalad; especially is this the case if there be a simultaneous contraction of the dorso-lateral and ventro-lateral muscular bmidles pre\dously described. The extent of this shortening of the trachea is prevented, however, to some extent by the sternal attachments of the sterno-tracheales. Hence, when these attachments are severed, the contraction of the tracheal parts of these muscles tend to make the external tympanic membranes more tense and a modification in the quality of the voice is the result. It is of course probable that other factors, more physical than the above, also share in the production of the modification observed.

From their structure it is evident that the external and the internal tympanic membranes are vibrating structures. As in


mammals the stratified ciliated columnar epithelium of the larynx is transformed into the stratified squamous layer over the true vocal cords, and submucous glands are absent, so in case of the sjrrinx of the chicken the stratified ciliated columnar epithelium is changed into a squamous epithelium over the tjnnpanic membranes. This squamous epithelium, however, is composed of but a single layer. The prominent glands of the mucosa are also absent from these membranes.

The semilunar membrane was belie^^ed by Savart, Wunderhch, and others, to play an important part as a vibrating structure. It is said to be more prominent in songsters than in other birds, but Hacker ('00) pointed out that in black birds its epithelium is of the stratified cihated columnar type, and that it probably is not of much importance in the production of voice, and also that it is often as well developed in songless species as in the best songsters. This structure certainly does not act as a vibrating membrane in the chicken for it is covered with stratified ciliated columnar epithelium, and mucous glands are just as abundant in it as in any unmodified part of the epithelium.


1. The syi'inx of the domestic chicken belongs to the tracheobronchialis type, and is quite simple when compared with the voice organ of song birds.

2. No intrinsic muscles are present in the syrinx of Gallus domesticus. The extrinsic paired sternotrachealis with its caudal prolongations constitute the entire musculature of the syrinx.

3. The rigid skeleton is very highly modified. The first four tracheal rings are imperfectly fused to form the tympanum. The four intermediate syringeal cartilages are continuous ventrally with theii ventral pyramid of the pessulus, while dorsaUy they end unattached. The first bronchial half-rings are large and in adults are attached and fused at both ends to the pessulus. The pessulus is the largest of all skeletal parts and lies dorso- ventrally at the junction of the bronchi, in a plane transverse to the long


axis of the trachea. The tracheal rings, the pessulus, and the ventral ends of the first half-rings become ossified, while all other skeletal parts remain cartilaginous.

4. The external tympanic membranes appear between the fourth intermediate syringeal cartilages and the first half-rings, while the internal tympanic membranes extend from the caudal borders of the pessulus to the bronchidesmus and represent merely a modified part of the medial bronchial walls.

5. The syrinx is hned with stratified ciliated columnar epithelium containing numerous simple alveolar glands. Upon approaching the tympanic membranes this columnar epitheUum is transformed into a stratified squamous epithelium which becomes a single layer of flattened cells over the membranes proper.

6. The tympanum is attached to the remainder of the sjo-inx only by elastic membranes.


1. The first indication of the respiratory system was observed in a 68 hour embryo in which the laryngeotracheal groove and the bronchi were represented. At first the trachea is much shorter than the bronchi, but with the development of the neck, it becomes, after the 140 hour stage, relatively much longer than the bronchi. The walls of the trachea and the bronchi are at first composed only of epithelium which contains two or three rows of nuclei.

2. The mesenchymal condensation common to the entire epithelial tube first becomes markedly prominent at the tracheal bifurcation in an smbryo of 152 hours.

3. The anlagen of the first bronchial half-rings appear in a 176 hour embryo, those of the fourth intermediate syringeaL cartilages appear 12 hours later. The anlagen of the third intermediate syringeal cartilages and the anlage of the pessulus are present at 200 hours.

4. Distinct cartilage cells were first observed in the first bronchial half -rings.

5. The first four tracheal rings have not united to form the tympanum at hatching, nor have the other skeletal elements


taken the shape of those found in the adult. No bone is present at the time of hatching.

6. CiUated cells are present in stages beyond 248 hours but were not observed in the region of the future tympanic membranes.

7. Mucous cells were first observed in 332 hour embryos and only in later stages were they found arranged in the form of simple alveolar glands.

8. Muscular tissue is differentiated in the 176 hour stage. Muscle fibers showing faint cross striations appear at 296 hours. At 452 hours the muscles are well developed.

9. At the time of hatching the tympanic membranes are thick. They are covered, however, as in the adult, with a single layer of epithehal cells.


1. That the syiinx is the true voice organ of the chicken is evident from the following deductions :

First, structurally it is the only part of the respiratory tract capable of producing sound.

Second, when the trachea is divided and the cephalic portion tightly tied, the chicken is still able to crow.

Third, after di\dsion of the trachea, voice can be reproduced artificially by forcing air into the air sacs.

2. The upper larynx serves only to modulate the voice.

3. The sterno-tracheal muscles, by their contraction, shorten the trachea and modify pitch. They also draw the tympanum cephalad, thus indirectly varying the tenseness of the tympanic membranes.

4. The air sacs are necessary in voice production, for voice could not be produced artificially after puncturing the cervical




1 Dissection of ventral portion of neck and cephalic portion of thorax to show course of trachea, X h Adult male.

2 Dissection of syringeal region. The ventral body wall and the heart have been removed to show the syrinx in position, X 3. Adult male.

3 Pessulus, X I65. Young adult male.


a, apices of pyramids

an, B LB II., anlage of first and second bronchial half-rings

an.i. S.C.I, II, III, I v., anlagen of first, second, third and fourth intermediate syringeal cartilages

an.pes., anlage of pessulus

ant.l.p., antero-lateral process of sternum

br., bronchidesmus

b.v., blood vessels

b.w., body wall

B., typical bronchial half-rings

b., ossified portion of shaft

b.sur., basal surface of shaft and pyramids

BI,BII,BIII, first, second and third bronchial half-rings

b.m., basement membrane

c, crop

cor., coracoid

c.d.m., ventro lateral muscle bundles

c.mes., condensed mesenchyma

d.sur., dorsal surface of pyramid

dor. p., dorsal pyramid of pessulus

e.t.m., external tympanic membrane

ep.t., epithelial tube

ep., epithelium

inf. I., lower larynx or syrinx

i.m., interannular membranes

i.t.m., internal tympanic membrane

i.s.c, intermediate syringeal carti

i.c.m., developing interannular and

intercartilaginous membranes lat.b., lateral borders of pyramids Lb., lateral border of shaft m.c, marrow cavity vi.b.ui., medial bronchial wall mes., mesenchyma oe., oesophagus pes., pessulus

per., developing perichondria st.h.m., sterno hyoid muscle sup. I., upper larynx, sterno trachealis muscle s., shaft of pessulus sem.m., semilunar membrane sm., submucosa tr., tracheal rings T., tympanum V.C., vertebral column ven,p., ventral pyramid ven.s.p., ventral surface of pyramid







4 Wax reconstruction of syringeal skeleton, X 8. Adult male









5 Mid-coronal section of syrinx, X 8^. Adult male.

6 Mid-coronal section of syrinx, X 85. Adult female.

7 Transparent drawing of wax reconstruction of epithelial tube and condensed mesenchyma in region of tracheal bifurcation. Ventral aspect, X 50. 176 hour chick embryo.







8 Wax reconstruction of epithelial tube and related developing skeletal structures in the region of the tracheal bifurcation. Ventral aspect, X 50. 226 iiour chick embryo.

9 Same as 8. Dorsal aspect, X 50.

10 Wax reconstruction of epithelial tube and related skeletal elements. Yentral aspect, X 50. 284 hour chick embryo.

11 Same as 10. Dorsal aspect, X 50.






12 Coronal section of the trachea in the region of its bifurcation, X 176. 176 hour chick embryo.

13 Coronal section of the trachea in the region of its bifurcation, X 100. 200 hour chick embryo.

14 Coronal section of the trachea in the region of its bifurcation, X 104. 248 hour chick embryo.







•■■'"■f i.


an pes.

' *ii.^?V



mx^' 14




15 Coronal section of the trachea in the region of its bifurcation, X 64. 284 hour chick embryo.

16 Coronal section of the trachea in the region of its bifurcation, X 25. 452 hour chick embryo.




tr-.ll -^

ep. bm. tri











17 Photograph, after exposure of the trachea.

18 Photograph after division of the trachea.






S. SAGUCHI Medical Academy, Kanazawa, Japan



I. Introduction 218

II. Materials and methods of investigation 219

III. Structure of ciliated cells 220

A. Descriptions 220

B. General considerations 228

a. The manner of implantation of the cilia 228

b. The cytoplasm and rootlets of the cilia 240

IV. Function of the ciliary apparatus 243

V. Regressive metamorphosis of ciliated cells 246

A. Transformation of ciliated cells into cells of different nature . . . 246

a. Historical 246

b. Observations 247

c. Summary 249

B. Atrophy of ciliated cells 249

C. Elimination of ciliated cells • 251

VI. Regeneration of ciliated cells 251

A. Transformation of other kinds of cells into ciliated cells 251

B. Mitosis of ciliated cells 251

a. The centrosome of the ciliated cell 252

b. Summary 257

C. Amitosis of ciliated cells 258

D. Significance of mitosis and amitosis of ciliated cells 261

VII. Development of ciliated cells 262

A. Historical 262

B. Embryonic development of the ciliary apparatus 263

- C. Development of ciliated cells in the efferent tubule of testis of

the mouse and rat 265

D. Ciliogenesis in the daughter-cells produced by mitosis of the

ciliated cell 266

VIII. Henneguy-Lenhossok's hypothesis , . . . . 266

IX. Bibliography 268





Since ciliary movement was first noticed by Anton de Heide, in 1683, ciliated cells have been a problem for various investigators. Their structures were studied by Friedreich ('58), Marchi ('66), Eberth ('66), Engelmann ('68, '79, '80), Eimer ('77), Frenzel ('86) and others; above all Engelmann and Frenzel gave detailed accounts of the matter, so there has been left but little to add as regards their gross structure. According to these authors, ciliated cells are provided, on their distal borders, with rod-like corpuscles, named basal knobs (Engelmann 's 'Fusssttickchen') or basal rods (Frenzel's 'Fussstabchen'), with which the ciUa are connected. These corpuscles — later termed basal corpuscles by Apathy ('97) — ^were frequently recognized by several investigators and regarded as a constituent of great importance of the ciliary apparatus, which consists of cilia, basal corpuscles and rootlets.

The question also, whether the factor which produces the ciliary movement is to be sought in either of these three structures or in any other structure; especially the question as to the significance and development of basal corpuscles, were frequently discussed, and have not yet been decided definitely. Nearly about the same time ('98), but independently of each other, Henneguy g.nd Lenhossek announced their opinion as to these questions. They regard basal corpuscles as derived from central corpuscles and as representing a kinetic centre of the cihary movement, and point out, as evidence in support of this view, the morphological, topographical and chemical resemblance between these two structures, and the behavior of central corpuscles or similar bodies to the tails of spermatids of certain invertebrates and to those of antherozoids of certain plants. Whether or not this hypothesis (Henneguy-Lenhossek's hypothesis) expresses the truth, can be solved, I think, only by studies of the question of the existence or non-existence of centro somes and of mitosis in cihated cells, and especially by .hat of ciUogenesis. Benda ('00), Fuchs ('04), Ikeda ('06) and others believe they have found that the basal corpuscles take origin from the



dividing centrosome, whereas Gurwitsch ('Oi) and Wallengren ('05) assert, from iheir studies of the genesis of the ciha, both embryologically and after mitosis of ciUated cells, that the basal corpuscles are not derived from the centrosome, but differentiate in the cuticle or superficial layer of the cell-plasm, and that, accordmgly, there is no reason to accept the Henneguy-Lenhossek's hypothesis.

In the present paper I intend to give an account of the structure of cihated cells, their de- and regeneration, and the development of cilia, an account, which would throw some light on the problem above mentioned.


An extensive comparative study being necessary, I have fixed epithelia from various regions of invertebrates and vertebrates, which may be tabulated as follows:


(Anodonta Limax Helix Gill \

Foot /

Hepatic duct Helix






Efferent tubule of testis



Efferent tubule of testis


Efferent tubule of testis)

Epididymis . /

[Rana temporaria ) Rana esculenta



[ Trigonocephalus

I Eliaphis


J Mouse


Embryonic materials

Ciliated infundibulum] Tela chorioidea [

Pharynx f

Epidermis j

("Hyla, larvae J Rhacophorus, larvae \ Hynobius, larvae [Rana esculenta, larvae


Small pieces of the materials were preserved either in Meves's fluid or in sublimate-acetic mixture, the result ot which was in most cases favorable for the study of minute structures. With the former fluid it was passed through grades of alcohol, after washing thoroughly in running water, to 96 per cent alcohol; with the latter it was brought immediately into 70 per cent alcohol with iodine and then through grades of alcohol. Imbedded in paraffine through chloroform.

The sections were, for the most part, cut 4 /x thick and stained on the slide. For staining I have mainly employed Heidenhain's iron-haematoxyhn method, with or without counterstaining with plasma-dyes.


1. The foot of Anodonta. The epithelium consists of a roAv of columnar ciliated cells (figs. 1, 2), the distal ends of which are bordered with a well-defined, relatively broad cuticle which stains faintly with iron-haematoxylin. In profile it is noticed that the cuticle is striated longitudinally, each stria appearing to be continuous with a cilium, which is from six to seven dmes as long as the height of the cuticle. Below the cuticle is a narrow zone, consisting of a series of basal corpuscles, which may be called 'layer of basal corpuscles' (fig. 2). These are minute, oval bodies appearing, in surface view, as small granules arranged in parallel rows, which are, in most cases, at right angles to the larger side of the distal cell-end (fig. 3). Below the layer of basal corpuscles is another zone of dense nature, which stains more deeply with plasma-dyes. Through this zone, which provisionally may be designated 'transparent zone,' fine fibers pass vertically from the basal corpuscles towards "he cytoplasm proper, where they become lost to view. The cytoplasm contains a large number of tortuous mitochondrial filaments, stained black by iron-haematoxylin; they are mainly arranged vertically and are accumulated especially in the upper portion of the cell-body, under the transparent zone, without passing


into it. No centrosome and mitotic figure were seen in the cells in question.

2. Gills of Anodonta. The epithelium consists of ciliated and non-ciliated cells. Of the ciliated cells I have studied lateral-cells and corner-cells. Both these cell- varieties are provided with a cuticle carrying long cilia, which, in fixed preparations, are curved towards the external surface of the gill-torus, as noticed by Engelmann ('80); this is especially well marked in the lateralcells. The behavior of the laj^er of basal corpuscles, of the transparent zone and of the chondriocontes is analogous to that described above for the foot of the same species (fig. 4).

The layer of basal corpuscles, which seems to correspond to the zone found by Wallengren ('05) below the cuticle, contains basal corpuscles, in the shape and arrangement of which the lateral- and corner-cells differ widely from each other. Viewed from the surface (fig. 5 a), the lateral cells, each of which has a regular rectangular outline, are arranged in stone-wall-like rows running along the gill-torus. The basal corpuscles are spherical granules, which form parallel rows arranged in the direction of the short diameter of the cell-border, as is the case with the foot-epithelium of the same species. Engelmann ('80) has also described such a linear arrangement of basal corpuscles; he states that the lines are inclined about 45° to the longer margin of the cell-border, but I could not confirm this.

The corner-cells, separated from the series of lateral-cells by two rows of non-ciliated cells, are closely apposed on their flat faces, forming a series along the gill-torus (fig. 5 b); in surface view of this type of cell two parallel lines, stained black by ironhaematoxylin, run along the longer axis of the cell-border. T ese lines, named by Engelmann bandelets ('Leistchen') serve for the attachment of the cilia, and might conceivably be formed, as the author suggests, by the fusion of basal corpuscles.

3. The intestine of Anodonta {figs. 6-8, 16). The epithelium in this region consists of ciliated, non-ciliated and glandular cells. The first two of these are pro\dded with striated cuticular borders ; the clear space between the striae in lateral view correspond with pores seen in surface views; this suggests strongly that the


cuticular substance is perforated by canaliculae, filled with fluid; the striae are therefore nothing but the septa of these canaliculae viewed in profile. The ciUa pass through the striae and come into connection with the basal corpuscles underneath the cuticle. The basal corpuscles, spherical or oval in form, are here also arranged in linear series running, in most cases, parallel to the shorter diameter of the cell-border. Since it seems that the larger diameter of most epithelial cells is oriented in a direction, perhaps parallel to the axis of the intestine, it is natural that the rows of basal corpuscles all run in a direction, which is at right angle to the axis of the intestine, a condition which, I think, is of an important significance for the function ol ciliated cells in this situation. Under the layer of basal corpuscles follows a comparatively narrow, transparent zone, through which fine fibers, arising from the basal corpuscles, pass downwards.

The chondriocontes in the cytoplasm are chiefly arranged in a vertical direction and are largely accumulated beneath the transparent zone (figs. 6, 8). They seem not to be floating in the plasma-sap, but to be suspended on or imbedded in the beams of the plasma-network. If a reagent which causes a shrinking of the cell, such as sublimate-mixture, be employed (fig. 7) there appears a sinuous or spiral fiber, stained deeply by ironhaematoxylin, and splitting upwards into several filaments, which end under the transparent zone. With Meves's fluid, on the contrary, no such feature appears (figs. 6, 8). From this it is highly probable that this fiber, found by various investigators and termed 'cone of rootlets,' is produced by the act of shrinking, which must result in the cohesion of beams of the plasma -network with the suspended chondriocontes. Centrosomes and mitosis are often met with in ciliated cells from this location.

4. The middle-intestine of Helix. The constituent elements of the intestinal epithelium of this species are the same as those of Anodonta. The ciliated cells (figs. 9, 15) bear comparatively short cilia, which pass through the cuticle and come into connection with basal corpuscles; these latter are very small granules, so closely applied to one another that they appear, at a certain


degree of differentiation of iron-haematoxylin, as a single dark line beneath the cuticle. The transparent zone is about one half the height of the cuticle and is marked off distinctly from the cytoplasm proper. Chondriocontes are also visible; their arrangement is similar to that described above for Anodonta. Centrosomes and mitotic figures are easily recognizable; twoor poly-nucleated cells also occur.

5. The hepatic duct of Helix. The epithelium in this region is composed of ciliated cells with interspersed glandular- and non-ciliated columnar cells; in approaching the gland proper non-ciliated cells increase in number. The cilia pass, as usual, through the cuticle in order to connect with basal corpuscles, which appear as minute granules and form, viewed from the surface, linear series running parallel with one another (figs. 10, 11). Sometimes the course of the chondriocontes is so regular, that they form a sort of palisade (fig. 10). In order to follow the fibrils which pass downwards from the basal corpuscles it is necessary to remove the chondriocontes, which often prevent close examination. This is realized practically in poorly preserved portions of the sublimate or Meves materials. Sections from such portions show that the fibrils in question (fig. 11) pass, in a slightly tortuous course, through the cytoplasm towards the distal end of the nucleus. These fibrils (corresponding with what are designated as rootlets of the ciha) appear not to be free, but to be connected with protoplasmic networks and are not to be confounded with the chondriocontes, which seem to be suspended on or embedded in the beams of the network and never pass into the transparent zone. Centrosomes and mitosis of ciliated cells occur not infrequently; two-nucleated cells also often are met with.

6'. The intestine of Helix. There exist no noteworthy differences in structures of cihated cells between the intestine of Limax and that of Helix, so that its description can be omitted.

7. The epidermis of Am-phihian larvae. The epidermis of young Amphibian larvae is composed of two layers ; in the upper or cuticular layer, there are two kinds of cells: non-ciliated and ciliated cells. The latter (figs. 27-32) are few in n amber and


are scattered over the epidermis. They are furnished, at the distal border, with striated cuticle; each stria corresponding with a cilium. With favorable staining the basal corpuscles appear as diplosome- or dumb-bell-shaped granules (figs. 31, 32); one of these granules is situated at the upper border, the other at the lower, of the cuticle; the former is always smaller and more easily decolorized than the latter, so that there often appears a single row of basal corpuscles below the cuticle. The axis of these diplosome-like basal bodies either corresponds with the vertical axis of the cell in question (fig. 32) or is inclined in a determinate direction (fig. 31). When studied in tangential sections, the basal corpuscles form parallel rows and are inclined towards one end of the row (fig. 33). Since the inclination of basal corpuscles occurs only in a determinate direction, it is evident that, when the cell in question is cut parallel to the rows, the feature as represented in figure 31 is produced, and, that, on the other hand, the same holds true for the cell with vertical basal corpuscles in the sections which are cut at right angles to that direction (fig. 32).

The chondriocontes in the cihated cells are more numerous than in neighboring non-ciliated cells and are collected especially beneath the cuticular border. In case the cilia are localized in a certain circumscribed region, the main mass of the chondriocontes drifts to that portion, which indicates a close relation between the two (fig. 27). In sublimate preparations, fibers which, arising from the basal corpuscles, pass downwards, can be readily seen (fig. 30); they are not to be identified with the chondriocontes, but must be regarded, from their staining reaction and the continuity with basal corpuscles, as rootlets of the cilia.

8. Tela chorioidea of Amphibia-larvae. The ciliated epithelium-cells from this structure (figs. 36-38) carry comparatively thick cuticular borders, which are not homogeneous, but are finely striated in the longitudinal direction. The cilia, which may be located in a circumscribed portion, especially in the middle of the cell-border, pass vertically through the cuticle and are connected with basal corpuscles. The latter are situ


ated in a layer corresponding to the transparent zone of the Meves preparates, and appear as diplosome- or dumb-bell-shaped granules; the distal granules are always small and readily decolorized. The axis of the basal corpuscles is not vertical, but is inclined to one side of the cell (fig. 38). Since, in other cases there are vertical basal corpuscles, it appears probable that here also there may be the same condition as in the ciliated epidermis cells of the same species. The corpuscles are, however, irregularly disseminated, without forming rows (fig. 39). In the cytoplasm there are numerous chondriocontes, which are often oriented towards the cilia, a condition which is especially striking in cases when the ciha are localized in a certain portion of bhe cell-border (fig. 36).

9. The ciliated infundihulum of Ainphibia-larvae. The ciliated cells (fig. 41) are short, cylindrical in shape and covered with thin structureless cuticular borders. The cilia are very long and often adherent into a thick curved bundle, the free end of which is always directed towards the peritoneal canal. The cilia penetrate the cuticle and come into connection with the basal corpuscles, which appear, in tangential sections of sublimate-materials, as very small granules scattered irregularly over the cellsurface, with the exception of a narrow peripheral zone. The chondriocontes are mainly accumulated in the distal portion of the cell-body, separated, however, from the layer of basal corpuscles by the transparent zone.

10. The pharynx and oesophagus of A^nphibia (figs. Jf2-5o). The cuticular border of the ciliated cells is longitudinally striated, each stria corresponding to a ciliary fiber. Since the basal corpuscles are very small and set closely side by side, it is difficult, in side view, to distinguish between them (fig. 46) . In tangential sections they form linear series, parallel to one another, the corpuscles in the same row being united by a slightly staining fiber (figs. 47, 50); the lines lie, in most cases, in a direction at right angles to the longer side of the cell-border. In the transparent zone there are fibrils which run from the basal corpuscles downwards and are to be identified with the rootlets of the ciHa. The chondriocontes are mainly arranged vertically and


are accumulated especially underneath the transparent zone. Ciliated cells with a centrosome and those with two or more nuclei often occur.

11. The gall-duct of Amphibia. The epithelium of the gallduct of Amphibia consists of ciliated and non-cihated columnar and basal cells. The ciliated cells (fig. 57) are columnar in shape and bordered with a cuticle, through which the cilia pass to connect with the basal corpuscles. These are small granules, situated underneath the cuticle. If, as in a certain degree of differentiation is the case, the points where the cilia are connected with the upper border of the cuticle appear as black granules, then the feature of diplosome- or dumb-bell-shaped corpuscles would be brought about. The behavior of the transparent zone and of the rootlets of the cilia is about the same as described in the pharynx of Amphibia. Ciliated cells with two nuclei and those with centrosomes are not infrequent here.

12. The oviduct of Amphibia. The oviducal epithehum is composed of ciliated and glandular cells. The former (figs. 63, 66) have thin cuticular borders; the basal corpuscles appear either as short rods in the cuticle, or as small granules underneath it. This difference of shape may perhaps be correlated partly with the degree of staining, partly with different animals. Thus I have found short, rod-hke basal corpuscles in the oviduct of Triton, whereas they appear in that of the frog as of minute granules underneath the cuticle. In tangential sections they are arranged in parallel rows, all of the granules of one row being united by a fine slightly-staining thread. The direction of the rows corresponds to that of the short side of the distal cell-border (fig. 67). The chondriocontes are mainly gathered under the cuticle, the transparent zone is illy defined; binucleate ciliated cells often occur.

13. Ciliated cells in the peritoneum of Rana temporaria. The occurrence of ciliated cells in the peritoneum of female frogs was first described by Thiry ('62), and then confirmed by SchweigerSeidel and Dogiel, Neumann, Nussbaum and others. According to these authors the ciliated cells in question are flattened and are smaller than neighboring endothehal cells ; they are either


scattered or form rows. The type of cell which I have found among the peritoneal cells near the oviduct is represented in figure 68. The ciha are comparatively thick and are connected with spherical basal bodies; a special cuticular border seems to be wanting in this cell.

14- The trachea of Trigonocephalus. The ciliated cells in this part (fig. 69) have on their free surface, cuticular borders and relatively long cilia. The basal corpuscles are differentiated wdth difficulty, and therefore appear in most cases as a dark fine underneath the cuticle. In tangential sections they appear as minute granules, closelj' appfied to one another, the whole, at first sight, resembfing a fine network. On close inspection, however, I had no difficulty in making out linear series of granules, arranged in a determinate direction. The transparent zone and the chondriocontes behave in the same manner as described in other species. Cifiated cells with two nuclei are not infrequent.

16. Efferent tubules of the testis of Reptilia. Of the reptiles, have studied snakes, adders and lizards. The epithelium consists of cifiated cells and cells with brush borders. The former are small in number in proportion to the latter. At the free extremity of the ciliated cells (figs. 72, 73, 76) there is a narrow zone, darkly stained by iron-haematoxyfin and presenting the appearance of a cuticular border. The basal corpuscles are either imbedded in this zone or situated beneath it. The chondriocontes are collected below the cuticle, being separated from it by a narrow transparent zone. Binucleat^ ciliated cells often occur in the efferent duct of these forms.

16. The oviduct of the lizard. The oviducal epithelium consists of cifiated and glandular ceUs. The former are few in number as compared with the latter and bear on their free surface thin cuticular borders. The basal corpuscles are in most cases rod-like granules imbedded in the cuticle; in other cases they appear as small spherical bodies beneath the cuticle. This difference maj^ perhaps be connected with the degree of staining. The transparent zone is illy defined; binucleate cifiated cells are not rare.


17. Efferent ducts of the rat and mouse. The epithelium consists of ciHated cells and cells with brush borders. The ciliated cells (figs. 84-88) taper gradually to their proximal extremity. Their nuclei are large and take a higher position than those of neighboring cells, as already noticed by Lenhossek (1898). The uppermost portion of the cell-body generally is of a dense character, and basal corpuscles, rod-like, diplosome- or dumb-bellshaped, are arranged here in parallel rows. Any well-defined cuticular border, such as seen in other places, is nob discernible here. There are centrosomes in the transparent zone. The chondriocontes are mainly collected below the transparent zone, without entering into it. The basal corpuscles, viewed in tangential sections, are not scattered ii-regularly, but are arranged in linear series, parallel to one another and at right angles to the larger diameter of the cell-border. There is always a single nucleus.

18. The trachea of the rat and mouse. The ciliated cells of the trachea of these forms (figs. 91-95) are bordered with a distinct cuticle through which the cilia pass downwards to connect with basal corpuscles. These are small spherical granules situated under the cuticle; in a surface view their linear arrangement is conspicuous (fig. 95). The chondriocontes never enter into the transparent zone. The cells are often provided with two nuclei, the centrosome is readily discernible.


a. The 7nanner of implantation of the cilia

Ciliated cells are either columnar, or cubical, or, in rare cases, flattened in shape, and, although they participate in the formation of the simple or stratified epithelium, it can hardly be said that there is any epithelium which consists of ciHated cells only, other kinds of cells, such as non-ciliated columnar or glandular cells almost invariably entering into its composition. And while the cihated cells, on the one hand, are partly eliminated after degeneration, partly converted into columnar or glandular cells, the columnar epithelial cells, on the other hand, become trans


formed into ciliated cells. This is a physiological change which is going on, not only in embryonic or 3^oung tissues, but also in those which are fully developed. Under these conditions, the differentiation of the superficial plasma-portion of a ciliated cell might be considered from two morphological points of view: first, from the structures derived from the normal columnar cell, and then from what are newly formed with the development of the ciliary apparatus; and it is natural that these differentiations vary according to the forms of epithelial tissues in which the ciliated cells are involved.

It is extremely rare that the ciliated cell exhibits no special differentiation in its distal border. Even in such cases as where a distinct cuticle seems to be lacking, as Lenhossek ('98) and Studnicka ('99, '00) state, the distal border of the ciliated cell where the basal corpuscles lie, is somewhat dense in character and stains more heavily with plasma-dyes. This fact might have, as Lenhossek remarks, a certain amount of importance for the physiological function of the cilia.

1. Relation between the cilia and the cuticle. Friedreich ('58) was the first to describe the cuticular border of the ciliated cell ; he found a striated cuticle in the ciliated cells of the trachea of man and bull, and of the ventricular wall of the human brain; and believed that the cilia pass into the cytoplasm along the striae. The structure in question has since been noticed by Eberth ('66), Marchi ('66), Engelmann ('68) and others.

I find in the literature of the subject five different opinions with regard to the structure of the cuticular border and the manner of implantation of the cilia:

1) Rabl-Riickhard ('68) asserts that, although he found a cuticle in the ciliated cells of gills of Annelida, the cilia never pass through it, but are attached to its upper surface by somewhat dilated bases.

2) Friedreich ('58), Stuart (67), Apathy ('97) and Gurwitsch ('01, '04) believe that the ciliated cell has a homogeneous cuticle, through which the cilia pass into the cytoplasm, and, in addition, it was shown by Gurmtsch that the cuticle must be soft, since the cilia pass through it irregularly.


3) According to Engelmann ('80), Gaule ('81), Carriere (82), Heidenhain ('99) and others, the cuticle is provided with rodlike corpuscles, which appear, in profile, as longitudinal parallel striae, with the upper ends of which the cilia are connected; a view which is not markedly different from that described under 2).

4) Marchi (66) and Studnicka ('00) admit sieve-like perforations of the cuticle, through which the cilia pass downwards into the cytoplasm.

5) This view admits an alveolar structure of the cuticular border. Gurwitsch ('01) finds that the cuticle, in surface views of the ciliated cells from the rabbit's oviduct, is made up of regularly arranged alveoh; at the nodes of the alveolar networks there are basal corpuscles, with which the cilia are connected. Studnicka ('00) also found an alveolar structure of the cuticle in the ependyma cells of Spinax niger.

I have, in turn, found, in all forms studied, th the superficial portion of protoplasm of the ciliated cell, that is to say, the part where the cilia are implanted in the cell, is always, of a denser character, appearing either as a crust or as a welldefined cuticle, which varies greatly in thickness.

Whether the longitudinal striation of the cuticular border, seen in profile, is a feature brought about by the passage of the cilia, or is owing to rod-like structures in the cuticle, or is an appearance due to canaliculi, or is produced by an alveolar structure, of the cuticular border, is a question which is very difficult to solve; at least when one attempts to draw any conclusion from the study of ciliated cells themselves, since the feature is more complicated by the implantation of the cilia. I believe that the clue to the solution of the problem is furnished by the study, either of those cases in which the ciha are limited in a certain circumscribed portion of the cuticle (e.g., in the epidermis and tela of Amphibia-larvae), or in the transformation of non-ciliated into ciliated cells, or in the redifferentiation of ciliated cells into non-ciliated. From such studies 1 have been able to find three types of differentiation of the superficial portion in the ciUated cells studied:


1) In the first type, such as in efferent tubules of the testis of Reptilia and mammals, I have never been able to recognize any well-defined cuticle; what can be seen is an ill-defined crust, in which the basal corpuscles He, thus giving rise to the appearance of a longitudinal striated cuticular border. That this appearance is owing to the existence of basal bodies in the crust, is readily explained by the study of cihogenesis in epididymis cells of mammals. In efferent tubules of the mouse, as will be afterwards described, it is evident that cells with brush borders, which are provided with a crust on their distal margin (fig. 89), become converted into ciliated cells, and that from the way that the newly-formed basal corpuscles arrange themselves in the crust. In one word, the striated cuticle in this case is nothing but a portion of the crust defined by the arrangement of basal bodies.

2) To the second type belong the ciliated cells of the epithelium of the pharynx and tela of Amphibian larvae. As the representative of this group I take the ciliated cells of the epidermis, the most closely examined. As will be afterwards described, these ciliated cells are developed from non-cihated, the cuticle of the latter becoming directly converted into that of the former. The cuticle, in profile longitudinally striated, exhibits, in surface view, a well-marked reticular appearance. If these two features are considered associated with each other, it becomes evident that the cuticle is composed of closely apposed aloeoli, the junction of the alveolar walls appearing, in side view, as striae (compare: F. E. Schulze '69, '88; 0. Schatze '07; S. Saguchi '15), through which the cilia pass into the cytoplasm.

The structure of the cuticle of ciliated ependyma cells is the same as that above described, except that the longitudinal striae are very closely applied, rendering close examination extremely difficult (figs. 36, 38). Studnicka ('99) describes the cuticle of ciHated cells in the tela epithelium of Petromyzon as provided with small canals through which the cilia pass into the interior of the cell. This seems, however to me to be an artificial result brought about by the fixation; for but one of the seven cells in his figure 5 has such canaliculi, and he says in his text that '^sie (canaliculi) sind, wie es scheint, viel grosser als es zu diesem Zweck (passing through of the cilia) notig ware "


His figure 5 to 7 and 8, rather, appear to me evidently to show an alveolar structure of the cuticular border.

3) The intestines of Helix, Anodonta, Lumbricus and Limax, and the hepatic duct of Helix belong to this group. Before proceeding to describe the ciliated cells it will be necessary to say a few words with regard to the structure of the cuticle of non-ciliated intestinal cells. There are three opinions with regard to this question: (1) the cuticle is perforated by small vertical canals (Kolliker, Funke) ; (2) it is provided with rods (R. Heidenhain) ; (3) it is composed either of coarser or of slender, finger-shaped rods (Studnicka). In intestinal cells of the invertebrates studied, I have found that the cuticle has, viewed 'in profile, an appearance of longitudinal striation, seemingly caused by a parallel arrangement of rods; the clear vertical narrow lines correspond to the spaces between them. In surface view, minute, round, clear pores appear scattered over the cuticle. These facts point to the canalization of the latter, as KoUiker and Funke maintain.

The structure of the cuticle of cihated cells is the same as that of the non-ciliated and can readily be observed in the functional change of the former, that is to say, in their transformation into non-cihated or glandular cells (figs. 8, 11). In these figures it can well be seen that the cilia never pass through the canaliculi, but through the axis of the rods themselves.

The intestinal cell has been considered by some to have a brush border, and there is much diversity of opinion in regard to the relation between the cilia and the hairs of the supposed brush border. Vignon ('00) found, in the intestine of Chironomus larvae, that the ciha are attached to the free extremities of the hairs of the brush border, and believes that there is no genetic connection between these two structures, whereas Prenant ('97-' 99) and Holmgren ('03) assert that the hairs of the brush border are nothing but atrophied cilia. According to Gurwitsch ('01), there exists no relation, either genetic or anatomic, between the cilia and the hairs, the former passing independently through the spaces between the latter.

2. Relation between the cilia and the brush border. Of the ciliated epithelia studied, I have found cells with brush borders in


only the efferent tubules of the reptihan and mammalian testis. The cells with brush borders in these places, as will be afterwards more particularly described, become converted into cihated. In this transformation the ciha are produced, not by the lengthening of the hairs, but from an intracellular constituent, which passes out through the axis of the hairs, so that there exists no genetic relation between the two. I will return to this question further on.

3. The basal corpuscles. Historical. By the basal corpuscle is meant a minute granular or rod-like body, situated at the proximal end of each ciUum and stained black by iron-haematoxyhn. It would seem never to be lacking in the ciliated cell proper; there is, however, no uniformity as to its shape and position; it also varies considerably according to the degree of coloration.

A review of the literature shows that there is a considerable difference of opinion as regards the shape, position and arrangement of basal corpuscles, which is briefly summarized as follows :

Shape: (a) spherical or elliptical granules (Lenhossek '98, Studnicka '99, Wallengren '05, Erhard '10); (6) short rod-like corpuscles (Studnicka '99, Ikeda '06); (c) dumb-bell-shaped or diplosome-like granules (Gurwitsch '00, '01; Henry '00; Fuchs '04; Kolacev '10; Tschassownikow '13).

Position: the basal corpuscles are situated (a) in the distal border of the ciliated cell destitute of the cuticle (Lenhossek '98) ; (6) in the cuticle itself (Heidenhain '99, Gurwitsch '01;) (c) underneath the cuticle (Apathy '97, Studnicka '99, Wallengren '05, Erhard '10); (d) more or less deeply in the cytoplasm below the cuticle, as in the ciliated cells in the tela of Petromyzon and Salamandra (Studnicka '99).

Arrangement of basal corpuscles; viewed from the surface: they are either (a) irregularly scattered (Lenhossek '98); or (6) gathered in the central part (Studnicka '00); or (c) arranged in linear series (Engelmann '80, Gurwitsch '01); or (d) fused into hues (Engelmann '80).

Summary aiid discussion. It is worthy of remark, that the result of staining with iron-haematoxyhn, which is usually applied



in order to exhibit the basal corpuscle, varies according to the grade of differentiation. This is an advantage, and, at the same time, a difficulty, of that technic. To speak more correctly: whilej on the one hand, it enables one to judge the intensity of its affinity for cell-constituents, the same constituent of the cell, on the other hand, often presents a good deal of variety accordin.g to the degree of staining, which may give rise to misinterpretation. Not only the manipulation of staining, but also the degree of matm'ation of haematoxylin-solution used has a marked effect upon the result. An especially important factor in this staining is the method of fixing and the quality of fixative. For example, in preparations from materials perfectly fixed in riemming's fluid the mitochondrial filaments stain deeply, whereas the basal corpuscles remain unstained ; in poorly-fixed Flemming and in the sublima+e-material, on the other hand, it shows an inverse staining capacity.

So far as can be seen from my preparations, the basal corpuscles may be classified according to their shape and position as follows:

1) Spherical or elliptical basal corpuscles which are situated beneath the cuticle; these forms are met with in the gill (fig. 5, lateral cells), foot (figs. 1-3) and intestine (figs. 6-8) of Anodonta, in the intestine (fig. 9) and hepatic duct (figs. 10, 11) of Helix, in the intestine of Limax and Lumbricus, in the pharynx of Rana, Bufo and Triton (figs. 46, 49), in the gall-duct of Rana (fig. 57), and in the trachea of Trigonocephalus, of the mouse and rat (fig. 92). Such cases as this are found in the description and figures of the following authors who studied the intestines of invertebrates: Apathy ('97, typhlosole of Anodonta, Lenhossek ('98, middle intestine of Anodonta), Heidenhain ('99, intestine of Helix), Gurwitsch ('01, intestine and mouth of Lumbricus), Holmgren ('03, pharyngeal pouch of Lumbricus), Erhard '10, typhlosole and hepatic duct of Anodonta), and Kolacev ('10, intestine of Anodonta).

The corpuscles found by Engelmann ('80) in the intestine of Cyclas and Unio, and named by him 'Fussstlicke' may, as already stated by Studnicka ('99), be identified with the basal corpuscles. Heidenhain ('99) describes and figures lozenge


shaped basal bodies at the distal margin of the cell; Brasil ('04) has found m ciUated cells of the middle intestine of Lagis coreni, rod-like corpuscles, termed 'La batonnet cilifere,' which, to my mind, are not special rods, but correspond with the intracuticular portions of the cilia.

As to the cihated cells of the gills of Anodonta, my observations are in accord with those of Wallengren ('05) and Erhard ('10).

The basal corpuscles in the ciliated cells of the Amphibian pharynx are, so far as can be seen from my materials, extremely small, often appearing in profile as a dark line beneath the cuticle, a feature which has been already noticed by Lenhossek ('98). In all of my preparations which are stained to varying extents, I have not been able to make out rod-like or dumb-bellshaped basal corpuscles in the cuticle, as described by Eimer ('77) and Studnicka ('99) in the pharynx of Salamandra, by Gurwitsch ('01) in the oesophagus of Bufo, and by Tschassownikow ('13) in the oesophagus of the axolotl. I am not quite sure whether this difference is due to the materials used or to other factors.

2) Dumb-bell-shaped or diplosome-like basal corpuscles. These are situated (a) in the crust, as in the cihated cells of efferent tubules of the mouse and rat (figs. 84-88); {b) in the cuticle, as those of the Amphibian epidermis (figs. 31-33); (c) beneath the striated cuticle, as in those of tela of Amphibia (figs. 37-38).

In the cases (a) and (b) the upper granules are easily decolorized, thus giving rise to the appearance similar to that described under 1).

Regarding the basal corpuscles in the ciliated cells of the efferent tubule, I refer to the papers of Lenhossek ('98), Henry ('00), Ach ('02), Fuchs ('02, '04), and Ikeda ('06). Of these authors, Lenhossek, Henrj^ and Ikeda have found oval or short rod-like basal corpuscles, while Fuchs noticed, in addition to these, diplosome-like bodies.

My observations on the tela epithelium are entirely in accordance with those of Studnicka ('99, '00) on the ciliated ependyma cells of Cyclostoma, fishes, Rana, Salanandra, and of man, and


with those of Gurwitsch ('01) on the cihated cells of the tela epithelium of Salamandra.

3) Short rod-like basal corpuscles, which are imbedded in the thin cuticular or crust-like, distal border of the cell. The ciliated epithelial cells of the oviduct of Rana temporaria, Triton, and the lizard, and those of efferent tubules of the lizard, snake, and Trigonocephalus belong to this type (figs. 66, 73, 76).

As regards the arrangement of basal corpuscles in surface view, I can distinguish the following three types:

1) The basal corpuscles are gathered, without special disposition, in the middle of the cell; such a form is found in the cihated cells of tela chorioidea of Amphibian larvae (fig. 39) and corresponds with what Studnicka ('00) noticed in the ciliated ependyma cells of Cyclostomata, Rana, and especially of man.

2) They are arranged in linear series, parallel to each other, and in most cases at right angles to the longer side of the cellborder. I have found such a type in the following situations: in the gill (fig. 5 a), intestine (fig. 16,) and foot (fig. 3) of Anodonta, in the hepatic duct of Helix, in the epidermis (fig. 33), pharnyx (figs. 47, 50), and o\dduct (fig. 67) of Amphibia, in the trachea (fig. 95) and the efferent tubule (figs. 83, 90) of the mouse and rat.

That the view of Engelmann ('80) as to the arrangement of the basal corpuscles in ciliated cells of gills of Anodonta is at variance with that of my own, has been already mentioned in the foregoing chapter.

For the surface view of ciliated intestinal cells of invertebrates I refer to the figures of Heidenhain ('99) from the hepatic duct of Helix, of Gurwitsch ('01) from typhlosoles of Lumbricus and Anodonta, and of Kolacev ('10) from the intestine of Anodonta. According to the figure of Heidenhain, the rows of basal corpuscles are nearly at right angles to the longer side of the cell, while Gurwitsch figures rows parallel to the latter. This latter type I have seen very rarely, but never such a type as Kolacev figures in which the basal corpuscles are irregularly scattered over the cell-surface.


The figures of Henry ('00) of the cihated cell of the epidermis, coincide in many points with mine, whereas in Lenhossek's ('98), Fuch's ('04) and Ikeda's ('06) figures we find no definite arrangement of the basal corpuscles.

3) It will be seen from the foregoing that the basal corpuscles show a marked tendency to form parallel rows and to undergo fusion with each other; and in fact, my attention was repeatedly attracted, in some places (in the oviduct of Triton, in the pharnyx of Amphibia, in the foot of Anodonta) by the presence of a faintly staining fibril, uniting the basal corpuscles of the same row, as Heidenhain ('99) shows in his figure from the hepatic duct of Helix. The case which may be regarded as the extreme of this process is found in the corner-cells of gills of Anodonta (fig. 5 b). In flat view of this cell as mentioned above, there can be easily found two parallel stripes, which are stained black by iron-haematoxylin, and from which the cilia arise. Regarding the significance of these stripes Engelmann ('80) expresses himself as follows: Die Cihen entspringen oben auf jeder Zelle von zwei, den langen Randern parallelen Leistchen, die nichts anders als die verschmolzenen, oder richtiger reihenweise aneinander gefiigten Fussstiicke der elementalen Cilien," a view with which I agree entirely. The stripes, for which it will be better to reserve the name 'basal stripes,' are not to be identified with 'Basalkorperfasern,' found by Kuperweiser ('06) in the ciliated cells of the corona of Cyphonautes-larvae; for the cilia never arise from these fibers.

4. The so-called basal rods {FrenzeVs F ussstabchen) . As early as 1877 Eimer noticed, on the upper border of the cihated cells from the pharnyx of Salamandra and from the gills of Siredon pisciformis, the existence of rod-like corpuscles, the juxtaposition of which gives rise to the appearance of a cuticle. Three years later Engelmann described strongly refracting rods on or in the distal border of the cihated cells from intestine of Cyclas and from the nasal mucous membrane of the frog, and termed them 'Fussstiicke.' These rods are, according to his description, attached by means of the intermediate segments (ZAvischenglieder) to the hair-bulbs. From these observations, the


author claims, with Eimer, that the cuticular border of the ciUated cell, as seen by Eberth, Marchi and many others, is nothing but the juxtaposition of the 'Fussstiicke.' Similar bodies were afterwards found by Gaule ('81) and Frenzel ('86). Especially the latter author recognized the presence of Engelmann's ' Fussstiicke' in ciliated cells from various classes of invertebrates and gave the name of ' Fussstabchen' to them. According to him, the length and the structure of these corpuscles vary in various animals and situations. In the simplest case it consists of a rod with a small knob at one or both ends, but the most complex one he found in the intestine of Littorina and Rissoa, which is composed of the following five segments: (1) the lower knob, (2) the basal block, (3) the side-knob, (4) the rod, and (5) the upper knob.

It is difficult at the present time to harmonize the 'rod theory,' urged by Engelmann, Frenzel and others, with the so-called 'cuticular theory' of Eberth and Marchi; for the same structure is materially influenced by the technique and by the character of the optical instruments used. Most of the recent investigators (Schiefferdecker '91, Apathy '97, Lenhossek '98, Heidenhain '99, Studnicka '99, '00, Gurwitsch '01, Wallengren '05,Kuperwieser '06, Erhard '10, Tschassownikow '13) have found a welldefined cuticle in the ciliated cell, through which the cilia pass. Lenhossek especially expresses himself as follows: i

Der Versuch von Engelmann und Frenzel, die Existenz einer eigentlichen Ciiticula in Frage zii stellen und das, was man als solche besehrieben hat, bloss als das Ergebiiiss der mosaikartigen Ziisammenlagerung der vertikalen basalen Telle der Flimmerhaare zu erklaren, muss als verfehlt bezeichnet werden. Die 'altere Ciiticiilartheorie,' wie sie Frenzel nennt, ist noch immer richtige.

Brasil ('04) still maintains, in the ciKated intestinal cells of Pectinaria, the existence of rod-like basal differentiations of the cilia, which correspond, according to him, to the basal rods of Frenzel. Studnicka ('99), on the other hand, finds the cuticle in some cases, in others not, but an arrangement of rod-like corpuscles.


1 have, in my turn, noticed in all ciliated cells studied, as the product of differentiation at the distal cell-border, a crust or striated cuticle pro\'ided with basal corpuscles or penetrated by the cilia, but I have never seen rod-like corpuscles situated on the outside of the cell-border. The cilia, from the basal corpuscles to their distal extremities, whether they are within or outside of the cuticle, show no trace of structure, but appear completely homogeneous. In the following I discuss the value of the 'rod theory,' comparing my results with those of Engelmann and Frenzel.

Engelmann ('80), (1) in his figures 11a and b, 12, 15, shows a well-defined, homogeneous border on the row of his basal knobs; (2) in his figures 5, 9, 18, and 19, a row of basal knobs, the upper and the lower ends of which are defined by straight lines, so that the appearance of a striated cuticle occurs; (3) in figure 20, basal knobs arranged on the surface of a cell from the olfactory pit of Rana temporaria ; the upper ends of these bodies are connected, by means of intermediate segments, with the hair-bulbs of the cilia. Thus Engelmann admits three types of implantation of the cilia. Is it rational to believe such a variety of structure in cells which fulfill the same function? I will try to account for this variety of structure in the following manner: In the first case Engelmann's conclusions are nearly in agreement with mine; for, in his figures, it is clearly seen that the cilia pass through the cuticle superposed upon the row of basal knobs which seem to correspond to my basal corpuscles. In the second case the author, in all probability, saw a striated cuticle with or without basal corpuscles. In the third case the cuticular substance, for some reason or other, escaped his obser^^ation, while the striae or the rod-like basal corpuscles, which stand out sharply, he takes to be independent rods upon the cell-border. The basal rods of Frenzel, on the other hand, seem to be distinct from the basal knobs of Engelmann in shape and refrangibility, as Frenzel says. I am rather, with Studnicka ('99), of the opinion that the basal knobs of Engelmann correspond to the lower knobs of Frenzel. The hair-bulb of the above authors might be identified with the node produced by


the accumulation of plasma at the point of union of thecilium with the upper border of the cuticle, a view which has been held by Lenhossek ('98), Wallengren ('05) and Erhard ('10). These nodes are not only characteristic of cliated cells, but also seen in ordinary epithelial cells; their existence in the epidermis cells of Amphibia-larvae has been shown by Leidig ('85) and Studnicka ('98).

From the above the inference would appear justifiable that the adherents of the 'rod theory' failed to notice any structural difference between the substance of the cuticle and that of the ciliary apparatus, but considered the more readily recognizable parts as a continuous structure, and overlooked the rest, a result either of poor fixation or of studying fresh tissues.

b. The cytoplasm and rootlets of the cilia

In the literature there are three types of special structure of the interior of the ciliated cell :

1) Valentin and Buhlmann first noticed the longitudinal striation of the cytoplasm, which has since been recognized by Friedreich ('58), Eberth ('66), Marchi ('66), Eimer ('77), and especially by Engelmann ('80), and named by the last author 'rootlets of the cilia.' And besides these, Lenhossek ('98), Fischer ('99), Peter ('99), Heidenhain ('99), Gurwitsch ('01), Joseph ('03), Kuperwieser ('06) and many others have found that the rootlets of the cilia are connected, above, with the basal corpuscles; below, they either lose themselves in the neighborhood of the nucleus or descend, forming a conical bundle, to the base of the cell. Apathy ('97) and Metalnikoff ('00), on the contrary, look upon this conical bundle as the termination of nerves which are not connected with the basal corpuscles, but which end freely between them.

2) According to Lenhossek ('98) and Fuchs ('02), the ciliated cells of the epididymis are devoid of rootlets of the cilia, but they contain, in the upper portion, granular substance which Fuchs takes to be mitrochondria.


3) Neither rootlets nor granules are visible (Lenhossek, '98) in the trachea of the rabbit; (Gurwitsch, '01), in the oviduct of the rabbit; (Tschassownikow, '13), in the oesophagus of the axolotl.

I have been able to distinguish the following two plasma divisions in all the cihated cells studied :

a) Under the cuticle there is a dense, transparent zone (figs. 6-9, 42-45), which is marked off below from the cytoplasm proper by a relatively distinct boundary-line. Apparently it never is lacking in ciliated cells, though it varies greatly in thickness. The rootlets of the cilia, arising from the basal corpuscles, pass through this zone (figs. 6, 54). Its existence was noticed by Gurwitsch ('01) in the cihated cells of the pharnyx of Bufo, of the intestine of Lumbricus and of the tela chorioidea of Salamandra, and we also find the same in the figures of Heidenhain ('99), and Erhard ('10). As stated above, the basal bodies are usually situated, either within the cuticle, or between this and the transparent zone; they may occur occasionally within the latter, as in the cihated cells of the tela of Amphibian larvae (figs. 37, 38).

b) The cytoplasm proper occupies by far the largest part of the cell-body; it exhibits a reticular or alveolar structure and contains a goodly number of mitochondria.

The size and arrangement of the meshes of the protoplasmic reticulum are not the same in different cells. We find a most striking instance of the longitudinal elongation of these protoplasmic meshes in the ciliated cells of the intestine of Anodonta and of the hepatic duct of Helix (figs. 6-11). In these cases the main beams run vertically and pass through the transparent zone to come into connection with the basal corpuscles. It is natural that these longitudinal striae, which correspond to what is described by various investigators as rootlets of the cilia, do not run independently, but are connected with each other by minute lateral fibrils. In this respect I agree entirely with Kolaev ('10), who says:

Aiif diese Weise entsteht das typische Bild eines etwas in die Lange gezogenen Netzes mit Verdickimgen an den Knotenpunkten, wobei die


Liingsseiten der Schlingen dieses Netzes, die sich durch ihre Massivitat imd ihre intensive Farbung auszeichnen, die Wurzeln der Flimmerhaare darstellen, wahrend die Querbalken schwacher ausgebildet sind, infolge dessen sie nicht immer wahrnehmbar ist.

I have found the so-called cone of fibrils, which is regarded as a bundle of rootlets of the cilia in sublimate-preparations only, but never in Meves-preparations. From this it may be inferred that the cone is not a real condition, but an artificial product due to the shrinking effect of sublimate-solution, by virtue of which the rootlets of the cilia adhere into a bundle.

The mitochondria in the ciliated cells were first noticed by Benda ('99) in the mid-intestine of Anodonta, in the hepatic duct of Helix, etc. He thinks that the rootlets of the cilia are formed of mitochondria or chondriocontes. Fuchs ('02) says that, in the ciliated cells of the epididymis of the mouse, the region between the row of basal corpuscles and the nucleus stains darkly and contains, in Benda preparations, a large number of mitochondria. Perhaps a dark granular zone found by Lenhossek ('98) in the upper portion of the ciliated epididymis cell is a similar structure. Recently, Meves and Tsukaguchi ('14) have found that, in the cihated cells of the small bronchi, the chondriocontes are accumulated for the mosi part in the superficial portion of the cell, some descending beside the nucleus.

This description of Meves and Tsukaguchi holds good for all the cihated cells I have studied. The chondriocontes course chiefly in the direction of the cell-axis and are especialh^ abundant beneath the transparent zone, without entering into it.

As regards the mutual dependence between the protoplasmic networks and the chondriocontes, I am fully justified in believing that the latter are not free, but are either imbedded in the trabeculae of the former or suspended on it; and, as the chondriocontes run mainly vertically and are suspended on the likewise vertical protoplasmic beams, that is to say, on the rootlets of the cilia, it often occurs that the latter are in a great measure hidden by the mitochondria; in fact, it would seem that these structures have not till now been distinguished from each other with certainty. Perhaps such granular or knotty appearance


of rootlets of the cilia, as are described by Eberth ('66), Gaule ('81), Lenhossek ('98), Benda ('99), Kolac6v ('10) and others are produced by a linear arrangement of mitochondria; on the other hand, however, there is no reason for believing the existence of such straight rootlets free from nodes, as are figured by Apathy ('97), Heidenhain ('99), and Erhard ('10).


Although this complex and difficult problem has received much attention, it is still far from a satisfactory solution. The review of the literature concerning this is given by Putter ('03), Erhard ('10) and Prenant ('14) to which the reader is referred for details. It may be briefly summarized as follows:

1) Some authors (Putter, '03; Gurwitsch, '04; Erhard, '10; Kolacev, '10) admit an active mobility of the cilia, others (Peter '99) a passive.

2) As to the function of the basal corpuscles, some look upon them as a. kinetic center of the ciliary movement (Henneguy, '98; Lenhossek, '98; Peter, '99; Joseph, '03), others as an endorgan of the nerve (Apathy, '97), and still others as a supporting organ (Eismond, '00); Maier, '03; Kuperwieser, '06; Erhard, '10; Kolacev, '10).

3) The rootlets of the cilia are regarded (a) as exhibiting contractibility and mobility (Stuart, '67; Simroth, '76; Benda, '99); (6) as primitive fibrillae of the nerve (Eimer, '77; Apathy, '97; Metalnikoff, '00); (c) as a nutritive organ (Engelmann, '80; Kolacev, '10); (d) as a supporting organ (Peter, '99; Eismond, '00; Maier, '03).

It is necessary to say here a few- words as to whether protoplasm or nucleus takes part directly in the ciliary movement. Of course, all protoplasm may be endowed with contractility, but there is no reason for supposing that it causes directly the ciliary movement, except that it may furnish a fresh supply to the ciliary apparatus; otherwise such special cell-organs, as the basal corpuscles and the rootlets of the cilia, would lose their significance. This is sufficiently demonstrated in the experiments of Peter ('99) which show that ev'en where the ciliated

244 . S. SAGUCHI

cell has lost its nucleus and the greater part of the cytoplasm, the cilia still continue to move, provided the ciliary apparatus be left intact.

The basal corpuscles, on the contrary, seem to play an important part in the ciliary movement and are never lacking in all true ciliated cells. Even when they have not been detected, it would be going too far to conclude that they are not essential to the movement of the cilia ; for it often occurs that they are not brought out by staining. Moreover I cannot so disregard the basal corpuscles which Kolacev ('10) describes: "Fiir die Mechanik der Flimmerbewegung . . . . ist das Vorhandensein eines fest fixierten Korperchens durchaus zweckentsprechend, wenn nicht gar physiologisch notw^endig." The view of Apathy ('97), postulating the nervous nature of the basal corpuscles and the rootlets of the cilia, is very curious and seems to have obtained few adherents amongst histologists.

Those investigators who accept an active mobility of the cilia themselves, claim to have observed, either the movement of broken ciha or of certain structures in them. Although I have studied sections from different places and with varying degrees of staining mth iron-haematoxylin, I have never been able to detect either axis-fibrils, as seen by Koltzoff ('06) and Erhard ('10), or the transverse striation described by Kolacev ('10). In the following I, first, summarize my observations and then express my opinion concerning the ciliary movement: (a) There can be seen no trace of structure in the cilium ; (6) at the basis of each cihum there is always a basal corpuscle, (c) beneath the row of basal corpuscles in most cases there can be seen a transparent zone, through which the rootlets of the cilia pass downwards from the basal corpuscles; (d) the rootlets of the cilia are only a part of the trabeculae of the protoplasmic reticulum; (e) the cilium, basal corpuscle and rootlet are a continuous structure; (/) viewed in tangential sections, the basal corpuscles are arranged in linear series, parallel to each other; the corpuscles in the same row show a marked tendency to undergo fusion with each other; (g) the dumb-bell-shaped or diplosome1 ke basal corpusc es form, in like manner, parallel rows. It often occurs that the upper extremity of every basal corpuscle



is incl'ned towards any one end of the row, and, at the same time, towards any one side of the cell; (h) in most cases these rows are at right angles to the longer side of the distal cell-border; (i) it seems that the direction of the rows agrees with that of the ciliary movement. These data lead us to the following conclusions :

1) The cihum itself has no active mobility.

2) The upper extremities of the rootlets are fixed in the compact, transparent zone, affording support to the c lia I have


no reason for supposing that the rootlets may cause the ciliary movement.

3) I am inclined to believe that the kinetic center of the ciliary movement is to be sought in the basal corpuscles, as Henneguy and Lenhossek maintain, and in the following manner: it might be conceived that the basal end of the cilium and the distal end of the rootlet are connected either directly or by means of a joint, around which is the basal corpuscle (s. text-figure). The latter, like muscle-substance, is endowed with contrac


tility, by virtue of which the ciHum is bent at that point ; hence the rootlet acts as a support of the ciUum, while the transparent zone holds the rootlet in its position; the restoration of the bent cilium may be effected either by the elasticity of the cilium itself, or by the antagonistic action of the contractile substance. It must be remembered, in this connection, that Eismond ('00) compared the ciliary apparatus to the skeleton of the fin of a bony fish, but I am far from agreeing with him in regarding both the basal corpuscle and the rootlet as supporting organs of the cilium. And, as the ciliary movement usually proceeds in a determinate direction, it is natural that there should be linear arrangement of the basal corpuscles in that direction. With these conditions, a stimulation which causes the movement of the cilia, is transmitted in a waving manner from one end to the other of the same row, and thus the regular succession of the ciliary movement is attained; this must be looked upon as being most completely effected when all of the basal corpuscles of the same row are fused, as in the corner-cells of gills of Anodonta; the movements in different rows, on the contrary, are conceivably independent of each other.




That ciliated cells become converted into glandular cells has been observed by Knauff ('67), Flemming ('85), Osawa (97), and others in the trachea, and by Lenhossek ('98) and Tschassownikow ('13) in the pharnyx of Amphibia. All of these authors state that droplets of secretion first appear above the nucleus, which becomes pressed downwards with the increasing accumulation of the former, so that at last there is formed a goblet or chalice cell. According to Lenhossek ('98) and Frenzel ('86), even in such glandular cells, the ciha can still be preserved and move actively. Next, the ciha are cast off and the secretion is discharged. The casting off of the cilia, according to Brasil


('04), results from the disappearance of the rootlets of the ciUa on account of the accumulation of secreted material ; concerning this he writes as follows: "La destruction du cone radiculaire deLruit les communication, quelle qu'en soit la nature, du cytoplasme et des cils. Isoles, ceux-ci ne peuvent plus que degenerer et disparaitre."

The transformation of ciliated cells into columnar has been observed by Henry ('00) in the epididymis, and by Gurwitsch ('00) in the intestine and the mouth of Lumbricus.


1 . The hepatic duct of Helix. The epithelium from this structure consists of ciliated, non-cihated columnar, and glandular cells. The ciliated cells become converted into non-ciliated columnar and the latter into glandular. This process can easily be followed in one and the same section. First, minute granules, stained yellowish by iron-haematoxylin, make their appearance above the nucleus (fig. 11, the cell to the left); they then increase in size and number and pass gradually upwards. Next, the ciha and the basal corpuscles disappear (fig, 11, the cell to the right) ; it seems that the cilia are not withdrawn, but are cast off in one of two ways: (a) they are broken off at the junction with the cell; the broken fragments are, in fact, met with in large number in the lumen; (6) they curl up and become converted into small drops, which at first are applied to the surface of the cell, but afterwards are separated from it. Thus the ciliated cells are turned into non-cihated columnar, and after the discharge of secreting mass into goblet cells.

2. The intestines of Anodonta and Lumbricus. The epithelium from these organs is composed of the same kinds of cells as those found in the hepatic duct of Helix. In many of the noncihated columnar cells there can be seen a row of basal corpuscles and a transparent zone, just as in the ciliated cells. On observing closely the cuticle of such cells, we often find some cilia (fig. 8) which are, as usual, connected with the basal corpuscles. Such cells might be interpreted as a form of transition between


the ciliated and non-ciliated cell, and as being on the road partly to the development, partly to the metamorphosis, of the ciliated cell.

3. The pharnyx of Amphibia. Here some ciliated epithelial cells transform themselves into goblet cells; first, the cilia disappear and then the upper portion of the cell becomes narrower, in consequence of the pressure of neighboring cells, which may cause a bulging of the cuticular border (fig. 53) ; on the surface of the latter there can often be seen degenerated remains of the cilia. As the process proceeds, the lower half of the cell enlarges, and elongates downwards; the chondi'ioncotes gather into a conical heap, above the nucleus, secretion granules first being formed at the top of the cone. Such cases as this are, however, not frequently seen; most of the goblet cells, rather, seem to be formed by the proliferation of basal cells.

Lenhossek ('98) and Tschassownikow ('13) report that ciliated cells in the pharnyx of Amphibia transform into goblet cells, and, that the cells may retain their cilia for a long time in spite of the accumulation of mucous droplets. Though I studied this point in considerable detail, I have not been able to confirm the latter opinion of these authors. I have frequently found some features which looked as if they might be due to the accumulation of mucous droplets in the upper portion of the ciliated cell, but, on closer inspection with higher power it became at once apparent that it was an appearance produced by the superposition of a tangential section of a goblet cell upon a neighboring ciliated cell, which may easily be understood when one recalls that the bulging goblet cell undermines neighboring cihated cells.

4. The oviduct of Amphibia. The oviducal epithelium is composed of ciliated and glandular cells. The former become converted into the latter by the accumulation of secreted material; they retain their cilia for some time (fig. 65, a and b) before the discharge of the contents sets in, as stated by Lenhossek and Tschassownikow.

5. The oviduct of the lizard. The above statement for Amphibia holds good for this form (fig. 78).



1. Ciliated cells are transformed into non-ciliated columnar and glandular cells.

2. In the transformation of the ciliated into the glandular cell, the cilia disappear, either before or after the accumulation of secreted material.

3. The disappearance of the cilia is not by withdrawal but by the'r being cast off.


That ciliated cells undergo atrophy, without being cast off, I have noticed in the gall-duct of Rana temporaria. The first change which occurs in the cell is nuclear hyperchromasy (compare my previous paper, 1915) : there appear in the nucleus numerous nucleolar granules, stained black by iron-haematoxylin, just as the nucleolus. With the increase of these corpuscles in number, the nuclear sap comes to stain more and more darkly, so that at last the nucleus becomes entirely a black mass. The cytoplasm also sooner or later grows black, especially in its upper portion (fig. 58, 59). In the course of time the black mass grows more and more pale (fig. 60) and then there can be seen in the cytoplasm numerous, larger or smaller round or oval black corpuscles, which often are so numerous as to hide the nucleus entirely from view; the chondriocontes which were found before, have now disappeared. I think that the corpuscles in question are derived partly from degenerated chondriocontes, partly from the nucleolar granules passed out of the nucleus. While the cytoplasm still is in the state of heavy nucleolar hj^perchromasy, the cell in question becomes constricted at the upper part, below the cuticle, two pieces, the upper smaller and the lower larger one, being thus produced (figs. 59, 60). Then the connection between the two becomes lost; the upper cihated one is, in all probabihty, cast off into the lumen.

The ciliated corpuscles, first found by Neumann and then recognized by Schmidt ('82) must be remembered in this connection. These corpuscles, according to Neumann, bear cilia,



either in a circumscribed portion, or over the whole of the surface. Schmidt says that, though occurring numerously in the epitheUum affected by catarrh, they are by no means infrequent in the normal epithehum; the author surmises their mode of development from the experiment, that, when ciliated cells are isolated, they undergo change in shape and constrict off spherical ciliated corpuscles. Since these bodies, according to his description, have no cuticular border, it must be fegarded as a portion of the cytoplasm which has taken cilia with it. From this description and his figures, it would seem that the ciliated corpuscle of Neumann corresponds with the upper piece which is produced by the constriction of the atrophying cihated cell. In contradistinction to the view of Schmidt, however, I cannot regard the corpuscle in question as derived from the cytoplasm only, but as composed mainly of the cuticular border. Similar corpuscles, but without ciha, were found by R. Heidenhain ('88) in the intestine of the rabbit, into which sulphate of magnesium was injected. From these facts it is evident that the constriction of the cell-body is a phenomenon which occurs in the degeneration, either of the non-ciliated columnar, or of the ciliated epithelial cell.

The lower piece which is left behind in the epithelium rounds itself off towards the basement membrane (figs. 61, 62), meanwhile the cytoplasm and nucleus become more and more clear and the black corpuscles are absorbed. The chondriocontes appear afresh in the cytoplasm (fig. 62), which indicates that the cell in question has not yet lost its vital activities. Thus the atrophied ciliated cell is transformed into a replacement cell; whether the latter returns to the ciliated cell, it is difficult to determine with certainty. Now the question arises, why are the cilia and the cuticular border cast off in the atrophy of the cihated cell? I am inclined to believe that this phenomenon is owing to their intense functional specialization which incapacitates the structures from undergoing redifferentiation.



In the epithelium of the intestine of Anodonta it is often seen that some cihated cells are being shed into the lumen. In the nuclei of such cells there appear numerous nucleolar corpuscles, while the nuclear sap becomes increasingly darker ; that is to say, the nucleus is in the state of nucleolar hyperchromasy ; in the cytoplasm, in addition, numerous larger or smaller vacuoles make their appearance, so that the whole exhibits a phenomenon of degeneration.


As the ciliated cells are reduced in number in the manner above mentioned, it is natural that they must be replaced by new ones. This is carried out either by the transformation of different kinds of cells into ciliated, or by the division of ciliated cells themselves.



That ciliated cells are derived from replacement cells, was noticed by Drasch ('81), Flemming ('85), Bockendahl ('85), Joseph ('05), Tschassownikow ('13) and others. According to these authors, the mitotic figures which often occur in the ciliated epithelium, do not belong to cihated cells themselves, but to cells of different nature, especially to replacement cells, which form a source of cihated cells. Schnitzler ('93), on the contrary, admits the transformation of goblet cells into ciliated.

I have observed, in the efferent tubule of testis of the adult mouse and rat, that ciliated cells are derived from neighboring non-ciliated, with which process I shall deal later in the chapter on the development of the ciha.


Although the question as to whether the ciliated cell multiplies by mitosis or not, has received much attention, it is still


far from a satisfactory solution. From tlue well-known fact that the centrosomg directs mitosis in general, it must be admitted that the mitosis of a cihated cell is also in close relation to the existence of the centrosome.

The centrosome of the ciliated cell

Upon this question, two distinct and opposed views are held by histologists. Henneguy ('98), Lenhossek ('98), Zimmerman ('98, in the uterus), Heidenhain ('99), Fuchs ('02, '04), Joseph ('03, '04) and others say that cihated cells lack centrosomes; many investigators (Studnicka '99, Fischel '00, Eismond '00, Henry '00, Benda 'OO-'Ol, Gurwitsch '00, '01, Ach '02, Wallengren '05, Ikeda '06, Erhard '10, and Tschassownikow '13), on the contrary, believe that they have found centrosomes in cihated cells.

It is a difficult matter to find centrosomes with certainty; for thej^ are minute corpuscles and often are apt to escape observation. Hence, it can hardly be said that they are absent when they cannot be detected; on the other hand, there are many other corpuscles which are similar to the centrosome in shape, position, and staining reaction, but of a different nature. Wallengren ('05) noticed such granules in ciliated cells, and said that they must be passed out of the nucleus, and to be closely related to the secreting process. We can only speak of centrosomes where the corpuscles show constant shape position and, staining powers and directly conduct mitotic division of the cell. It is very difficult, though not impossible, to follow such processes. Wallengren ('05) has already noticed that diplosomelike granules which are present between the basal corpuscles become centrosomes directing the karyokinetic cell-division.

I have noticed what are to be regarded as centrosomes in the following situations. Before describing my observations, it must be remarked that the centrosomes, in my conception, are to be recognized from the constancy of their shape, position and staining reactions, and, that whether or not these corpuscles, in reality, conduct the mitotic division, in other words, whether all


the ciliated cells which possess centrosomes always multiply by means of mitotic decision, is another matter, which will be dealt with separately in the next section.

1. The intestines of Helix and Anodonta. The ciliated cell from these organs contains a well-marked centrosome (figs. 6, 14, 15) which is commonly situated between the nucleus and the cuticular border, either near the cell-axis or at the periphery of the cell (fig. 6). It generally consists of two granules; the line of connection between them being in most cases parallel to the cell axis. They are surrounded with a clear halo, into which chondriocontes never enter. In the ciliated cells of the typhlosole of Anodonta, Erhard ('10) claimed to have found a centrosome in the form of double granules immediately below the cuticle, but the corpuscles in his figures 1, 11, and 12 are too indistinct to be distinguished from the basal corpuscles in this region.

2. The pharnyx of Hynobius larvae. Here the centrosome of the ciliated cell is a double granule (fig. 55) between the nucleus and the cuticle, especially near the latter and often near the periphery of the cell. The granules are either equal or unequal in size, and are surrounded by a halo. The line of connection between them is usually vertical, but an oblique or even a horizontal direction is not infrequent. I have also noted occasionally a centrosome containing one, three or four granules. There are descriptions of the central corpuscle of the ciliated cell in the pharnyx of Urodela, by Studnicka ('99), Eismond ('00), Fischel ('00), Gurwitsch ('01) and Tschassownikow ('13). The first three authors state that the centrosome occurs in the form of a diplosome between the nucleus and the cuticular border, just as I have found it. Studnicka, in addition, noticed, in the pharnyx of Salamandra, a centrosome consisting of three granules. On the other hand, Gurwitsch figures a centrosome in the superficial layer of a ciliated cell from the region intermediate between the pharnyx and the oesophagus of Salamandra* an observation with which those of the above authors and of mine are not in accord.

3. The pharnyx of Bufo. The centrosome is situated above and near the nucleus (fig. 49, the cell to the left), and consists of


two granules, which are usually arranged in a horizontal plane. Cases in which they are separated from each other or in which they seem to consist of a single granule are not rare.

If.. The gall-duct of Rana esculenta. The centrosome (fig. 57) surrounded by a clear halo is situated above the nucleus and consists of two granules, the axis of which either corresponds to the cell-axis or is more or less oblique to it.

5. The trachea of Trigonocephalus and the rat. The centrosome in both these forms, is situated between the nucleus and the cuticle, surrounded by a halo, just as described under 4 (fig. 69).

6. The efferent tubule of testis of the mouse and rat. The existence of centrosomes in cihated cells in the vas efferens was noted by Zimmerman ('98), Henry ('00), Ach ('02), Ikeda ('06) and Erhard ('10). According to Zimmerman, who has investigated ciliated cells of the human epididymis, the centrosome, surrounded by a clear halo, is situated immediately below the surface, and consists of two granules, the line of connection between them being either vertical or more or less oblique to the cell-axis. The observations of Ikeda and Erhard are in accord with those of Zimmerman. On the other hand, there is some difference of opinion as to the shape and position of the centrosome. In the epididymis of man, Henry found it above and near the nucleus, while Ach says that it consists of one, two or three granules, surrounded by a clear halo and that it lies either above or by the side of the nucleus or near the cell-basis.

In ciliated cells of the efferent tubule of the mouse and rat, I have also frequently seen centrosomes, which differ considerably in shape in the two species.

In the mouse, the centrosome of the ciliated cell (figs.- 82, 86) is situated immediately below the row of basal corpuscles, either in the middle or near the periphery of the cell, and consists of two granules which are always arranged in the horizontal plane. In order to distinguish them from the adjacent basal corpuscles with certainty, it is necessary to examine sections carried through the ciliated cell in an exactly vertical direction. Furthermore, I have often noticed, at the surface of neighboring cells with brush borders, a two-granulated centrosome, arranged


in the vertical direction ; the upper granule often bears a cilium, thus forming the so-called ' Zentralgeisselzelle' of Zimmerman. This is found not only in the efferent tubule of the mouse, but also in that of the rat (fig. 89).

In the efferent tubule of the rat, on the contrary, there are no such typical centrosomes as above described, but instead curious ring-shaped corpuscles in the transparent zone below the cuticle (figs. 87, 88). Most of these rings have a smooth internal and an uneven outer edge; in a certain sense they can be spoken of as being formed by the apposition of granules. Such rings are not always single, but there are many cells with two, either of equal or of unequal size (fig. 87) ; in either case they are arranged in a horizontal direction and connected with each other by a fiber.

I am inclined to beheve, from the position and staining of these rings, that they are derived from the centrosomes; a centrosome divides repeatedly and forms a ring by secondary fusion of separated particles; the connecting fiber representing the so-called 'centrodesmosis.' And, as a form of transition I have found a cell with a ring and a granule, of which the latter can be regarded as a component, still not divided, of the centrosome. Whether the ' Zentralkorperballen' found by Benda ('00) and Ikeda ('06) in the efferent tubule of the human testis and my ring are of the same character or not remains uncertain, but there are some differences between the two concerning the shape and position. Moreover, these authors admit that the 'Zentralkorperballen' take part in the formation of the cilia, a view with which I do not agree, for, on the one hand, the ciliated cell with such rings is always in the fully developed state and, on the other hand, it seems conclusive that the cihary apparatus never originates from the centrosome, but from a certain other constituent of .the cell, as will be afterwards described.

As above described, Henneguy ('98), Lenhossek ('98), Joseph ('03) and Fuchs ('02, '04) beheve that, in consequence of the absence of the centrosome, mitotic division does not occur in the ciliated cell. Ach ('02) and Tschassnowikow ('13) say that no mitotic figure occurs in cihated cells in spite of the existence


of the centrosome; Tschassownikow especially points out that the ciliated cell, though it shows no karyokinetic figure, can scarcely be said to be incapable of undergoing division, for it is provided with organs of cell-division, i.e., the nucleus and the centrosome. On the other hand, many of the investigators (Hammer '97, Gurwitsch '01, Maier '03, Wallengren '05, Erhard '10, '11, Gutheil '11) believe they have found mitotic figures in ciliated cells. Gurwitsch represents in his figure 19, a cell with a nucleus at the spireme-stage, which he takes as a ciliated cell; judging from his figure, however, it seems to me that the cell has no cilia and basal corpuscles, but a relatively high cuticular border. Just as little is certain whether the mitotic figures found by Maier in the epithelium of the gills of Triton and by Erhard in ependyma cells of an Acanthias embryo belong to ciliated or non-ciliated cells. On the contrary, the observation of Erhard in the typhlosoles, and especially that of Wallengren on gills of Anodonta appears to me to admit of very little doubt.

In the intestines of Anodonta and Helix and in the gall duct of Helix I have been able to follow the karyokinetic process of the ciliated cell, from the prophase to the anaphase. In the following I deal wdth the intestine of Anodonta as a representative of such cells.

Prophase (figs. 17, 18). The fu'st change of the nucleus consists, as usual, in the formation of the spireme; the cytoplasm becomes clear and the chondriocontes, which gradually grow pale, are scattered over the whole of the cytoplasm. At first the basal corpuscles still stain deeply, but afterwards they disappear. Contrary to the view of Wallengren ("05) I have found that the cilium and the basal corpuscle become lost simultaneously. All of the cilia and the basal corpuscles, however, do not disappear suddenly, but by degrees, so that it often occurs that a few basal corpuscles with cilia are left behind for some time (fig. 19) ; moreover I have noticed, in the intestine of Helix, cells with distinct cilia, even in the metaphase. The ciliated cell in the resting condition is long and narrow in shape; but on entering upon the karyokinetic process it becomes swollen where the


nucleus is situated, so that it takes the form of a pear. Although the distal border of the cell in question becomes diminished in extent, I have never been able to find that it becomes rounded off, after separating from the surface of the epithelium, as described by Wallengren. In some cases the distal border is either bulged (fig. 19) or indented, in consequence of the pressure of neighboring cells. The cuticular border, so far as can be seen from our materials, is left intact throughout the process of cell-division. Since the basal corpuscles are lost to view, the limits between the cytoplasm and the cuticle become indistinct, so that it looks as if the cuticle had disappeared, as Wallengren asserts.

Metaphase (fig. 19). The plane of division, so far as could be observed, is at right angles to the surface of the epithehum, so that the two resulting daughter-cells lie side by side.

Anaphase {fig. 20). The chromosomes send off lateral branches which join or anastomose together in a reticular manner while the nuclear membrane and nucleolus reappear. The cytoplasmic fission sets in at the inferior end of the cell and gradually proceeds upwards. Near and below the cuticle there appear the so-called intermediate corpuscle, from which protoplasmic filaments run towards the centrosomes. The chondriocontes, again becoming deeply stained, are arranged in the vertical direction, and are chiefly accumulated below the cuticle; the cell becomes in the meanwhile lengthened out downwards. Next, basal corpuscles and cilia make their appearance, which process will be referred to later on.

Although I have subjected various epithelial tissues of vertebrates to a careful examination, I have never been able to find karyokinetic division of the ciliated cell ; the mitotic figures which are often seen in the ciliated epithelium, do not belong to ciUated cells, but either to non-ciliated columnar, or to basal cells or to glandular cells.


1. The centrosome can be detected in ciliated cells of both invertebrates and vertebrates.


2. It can not be said that the centrosome is always lacking in case it can not be detected in the cihated cell.

3. The centrosome may be situated in every level between the nucleus and the cuticle.

4. I have been able to find mitosis of the ciliated cells in invertebrates only.


Henry ('99, '00), Ach ('02), Wallengren ('05), Jordan ('13) and others have noticed ciliated cells with two nuclei, which, according to these authors must be a result of direct nuclear division. I also found what appears to be amitosis of the ciliated cell in the following places.

1. The intestine and hepatic. duct of Helix. Here often occur ciliated cells with two nuclei (figs. 13, 14) which are oval in shape; these cells also may contain, as usual, a centrosome consisting of two granules, situated about midway between the cuticle and the nucleus (fig. 14). On close examination of preparations it would appear that these two nuclei are produced by direct nuclear division. In the middle of the nucleus there first appears a transverse furrow (fig. 12) ; by the deepening of which it becomes separated into two nearly equal parts. I have never observed cases in which the division was effected by the stretching of the nucleus. Cell-division follows nuclear division; the cell-boundary appears between the two nuclei; thus two superposed daughter-cells are produced, of which the upper alone bears cilia, while the lower one becomes transformed into a basal cell and then, by the accumulation of secreting granules in its interior, into a glandular cell, which elongates upwards between the ciliated cells until it reaches the surface of the epithelium.

Besides cells with two nuclei, there are occasionally seen those with three, four or five nuclei; even in these cases the internal structure at first shows no noticeable change. Whether these multinucleated cells are capable of division is doubtful; in all likelihood one or several of the nuclei undergo degeneration later on.

2. The pharnyx of Amphibia (Rana temporaria and esculenta, Bufo, Triton). In the normal condition, the surface of the


nuclei of the ciliated pharyngeal cells is smooth or shows at the most some irregularities. At the beginning of the nuclear constriction the nuclear membrane is characteristically thrown into folds, this is most marked in Rana temporaria. The cleavageplane of the nucleus is either vertical (figs. 42, 49) or oblique, or even horizontal (fig. 44). The furrow formed by a pushing in of the nuclear membrane appears, in Rana esculenta, as two parallel straight lines (fig. 42), but in Rana temporaria as zigzag lines which can be made out only by careful focussing (fig. 51). Between the structure of the amitotic dividing nucleus and that of the normal one there are no noteworthy differences, as will be seen in my figures. The daughter nuclei are at first closely apposed by their divided faces; later on, they are gradually separated from each other (figs. 43, 52). I have observed in Bufo a curious phenomenon in the behavior of the centrosome in the amitotic process : the ciliated cell in the pharyngeal epithelium of this animal, as already mentioned, contains a bigranulated centrosome lying above the nucleus (fig. 49, the cell to the left). With the separation of the dividing nuclei from each other, each of the granules accompanies each daughter-nucleus (fig. 49, the cell to the right), a fact which indicates that the centrosome is not entirely independent of amitotic nuclear division.

Cell-division follows nuclear division. Since the cytoplasmic fission always takes place along its longer axis of the cell, being inaugurated either at the upper end (fig. 45) or at the lower (fig. 48), it must be thought that the superposed nuclei, as it often occurs, undergo locomotion before the cell-division sets in. In the amitotic process there are no visible structural changes of the cytoplasm, the nucleus, or of the ciliary apparatus, except that the two former increase more or less in volume.

3. The gall-duct of Rana temporaria. Ciliated cells with two closely appHed nuclei occur, though not frequently, in the bileduct epithelium. The process of cell-cleavage, however, was not observed.

4. The oviduct of Rana temporaria and esculenta. In the oviducal epithelium are large ciliated cells with two nuclei (fig. 64), which cannot be interpreted except as having been pro


duced by amitosis, for mitosis of ciliated cells has never been observed here. Next, cytoplasmic division follows, the plane of which is always vertical. Sometimes the amitotic process seems to occur repeatedly; in figure 63 is a ciliated cell with six nuclei which are closely apposed on their flat side-faces; I am of opinion that the nucleus, in this case, has repeatedly undergone amitotic division, while cytoplasmic fission has not begun.

5. Efferent tubules of Reptilia. Ciliated cells with two nuclei often occur in the efferent tubules of Trigonocephalus and Elaphis (figs. 72-74), but rarely in the lizard. The direction of constriction of the nuclear membrane is not a constant ; the plane of cytoplasmic division, however, is always perpendicular.

6. The oviduct of the lizard. Binucleated ciliated cells are rarely seen here (fig. 77) ; the two nuclei which are evident must have been produced by amitosis, for mitosis has not been detected.

7. The trachea of Trigonocephalus. The two nuclei, either superposed or juxtaposed, are closely apposed by their flat faces; between these extremes there occur intermediate conditions.

8. The trachea of the rat. The arrangement of the two nuclei is the same as described under 7 (fig. 93). The cell-body gradually enlarges, especially increases in width. The cleavage of the cytoplasm begins either at the upper or at the lower end of the cell (fig. 94), and passes perpendiculraly between the nuclei. I have also often noticed that a granule accompanies each of the divided nuclei; from their position, shape and staining it is probable that these granules are derived from the pre-existing centrosome.


1. The ciliated cell may divide by amitosis.

2. This cell multiplication by amitosis occurs only in vertebrates.

3. The nucleus and the cell-body are constricted by pushing in of the nuclear membrane.

4. Cell-division follows nuclear division.

5. The ciliary apparatus remains unaltered in the amitotic process.


6. It seems probable that there exists some connection between the centrosome and the amitosis.


As mentioned above, there are two methods of division of the cihated cell, mitotic and amitotic. Mitosis occurs only in invertebrates; in vertebrates I have never been able to find it. Those who accept the hypothesis of Henneguy and Lenhossek, assert that the absence of mitosis in ciliated cells is a consequence of the lack of the centrosome. I do not agree with this, for the centrosome can easily be detected in ciliated cells in which mitosis does not occur. On the contrary, amitosis is the sole method of division of ciliated cells of vertebrates, in spite of the presence of the centrosome.

As to the significance of amitosis, two distinct and opposed views are held by histologists. According to one view, urged strongly by Flemming, Ziegler, and vom Rath, amitosis is not accompanied by the cytoplasmic division, but such a cell degenerates sooner or later. That the nuclei of degenerating cells may multiply by amitosis has received much attention (Nissen '86, Heidenhain '90, Plate '98, Dobell '07, Reichenow' 08) ; and, in fact, I have also noticed that the nucleus multiplies by direct division, in the degeneration of certain glandular cells in the larval epidermis of some Amphibia (Saguchi '15).

On the other hand, Child ('07), Patterson ('08), Maximow ('08), Des Cilleules ('14) and others affirm that amitosis is not always degenerative, but can be accompanied by the actual cellmultiplication; and, that mitosis follows amitosis and vice versa. According to Child and Patterson, amitosis is in close relation to the rapid nuclear multiplication and accordingly to the growth of tissues. Child, in addition, remarks as to the occurrence of amitosis as follows: Moreover, in several cases I have noted that in growing tissues where nuclei of different size are present, mitosis seems to occur more frequently in larger nuclei surrounded by considerable undifferentiated cytoplasm, while amitosis is more characteristic of the smaller nuclei with scanty cytoplasm." Recently, Jordan ('13) described ciliated cells


with two nuclei in the epididymis of various vertebrate animals, in the trachea of the cat, in gills of Unio, and he believes that these nuclei are produced by amitosis which is connected with the following cytoplasmic division. He expresses himself regarding the cause of amitosis in ciliated cells as follows: "The fundamental cause of amitotic cell division in cihated cells is the destruction of the centrosome in the formation of the basal bodies from which the ciha develop." Contrary to the view of this author, Henry ('00), Ach ('02), Ikeda ('06), ajid I have noted the existence of centrosomes in ciliated cells; and, on the other hand, there are many cases in which they may be detected in cells with two nuclei resulting from amitosis (Flemm ng, Maximow) .

My conclusions concerning the significance of mitosis and amitosis in ciliated cells may be summarized as follows:

1. The occurrence of amitosis in ciliated cells is not owing to the lack of the centrosome; for the latter can be detected in many cases in such cells.

2. The sole method of multipUcation of cihated cells in invertebrates is by mitosis, in vertebrates by amitosis; the cause bringing about this difference between these subkingdoms, must be due essentially to the degree of differentiation of the cell-plasm.


As early as 1875 Eichhorst described the development of cilia in the ependyma cells of the spinal cord of man; a transparent cuticle first appears at the distal border of the columnar cell; this afterwards becomes striated longitudinally, the cilia passing through this striated cuticle. According to Engelmann ('80), cilia regenerate by the elongation of their rootlets, while Fol ('96) described the formation of cilia by the prolongation of the distal cell-border. Gurwitsch ('01) distinguishes two types of cihogenesis: (1) first, the basal corpuscles appear at the nodes of the alveolar meshes of the cuticle and afterwards the ciUa become developed from them (in the oviduct of the rabbit and


in the pharnyx of Bufo) ; (2) first, the alveolar septa themselves become transformed into the cilia, while the basal corpuscles secondarily make their appearance at their basis (in the pharnyx of Salamandra larvae). On the other hand, the observations of Benda ('00), Fuchs ('04) and Ikeda ('06) on the efferent tubules of the human testis, had a marked influence upon the development of the so-called Henneguy-Lenhossek's hypothesis. According to these authors the centrosome, after multiplying by repeated divisions, gives rise to the basal corpuscles, from which the cilia develop. Wallengren ('05) studied ciliogenesis in the daughter cells produced by the mitotic division of the ciliated cell and found that, contrary to this view, the basal corpuscles are developed, independently of the centrosome, within the superficial portion of the cell-protoplasm, a view with which those of Heidenhain ('99) and Erhard ('10) are in accord. Ach ('02) asserts that the basal corpuscles are derived from granules situated above the nucleus, from which they are cast off, while Gutheil ('11) maintains that the cihary apparatus is developed from the microsomes suspended on protoplasmic networks.


1. The epidermis of Aviphihian larvae {Rhacophorus, Hynobius and Hyla). The development of the ciliated cell in the epidermis of these larvae begins at an early embryonic period in which the larvae are still enclosed within the gelatinous coat. The epithelial cell which is preparing to transform into ciliated (fig. 23, the cell to the left) is always provided with a well-marked cuticular border, the actual structure of which has been already referred to above in the section Relations between the cilia and the crust or cuticle." The cells in question are most commonly large, so that the lower ends often reach the basement membrane; most of these cells are laden with yolk spherules of various size (figs. 24, 26), in the intervals between which mitochondrial filaments (chondriocontes) course in different directions; they are also gathered in considerable numbers below the cuticle. At successive periods, these chondriocontes, after or without becoming vertically arranged, pass into the cuticle (fig. 24).


In this they do not pass into the alveoli, but ascend along the striae, which are nothing but nodes of the alveolar walls in the cuticle. Soon all the striae are occupied by the immigrant chondriocontes, so that deeply staining parallel striae occupy the same place as the alveolar ones (fig. 25). These, however, are not always parallel to the radii, sometimes they are more or less inclined in one direction, as seen in figure 25. This is, in all probability, due to the previous inclination of the alveolar walls. At successive periods the mitochondrial striae gradually grow pale; there are, however, cases in which it occurs only after completion of the development of the cilia. Next, they emit from their distal ends minute faintly stained prolongations (fig. 26), which are the young cilia. At the beginning they are relatively thick and short; later, they lengthen out (fig. 27). The chondriocontes are not used up in the formation of the cilia, but a number of them remain behind in the cell, especially gathered below the cuticle. From this manner of development it is possible that some of these chondriocontes are in continuity with the cilia, and give rise to the so-called rootlets.

2. The pharnyx of Rhacophorus larvae. The ciliogenesis begins in a larva about 15 mm. in length. The chondriocontes, growing more and more pale, ascend along the cuticular striae and project beyond the limits of the cuticle (fig. 56). At first they are short, and often curved in the shape of hooks, but afterwards they lengthen out.

3.. Tela chorioidea of Amphibian larvae (Hyla, Rhacophorus and Hynobius). In the early embryonic stage, the tela epithelium of Hyla consists of flattened cells, which contain numerous yolk-spherules and chondriocontes; the latter are accumulated near the upper border of the cell (fig. 34). In the next stage the cilia make their appearance (fig. 35). From the fact that the cilia are produced in that part which coincides with the accumulation of chondriocontes, the inference is warranted that there may be a genetic connection between the two.

4. The ciliated infundibulum of Hyla larvae. The chondriocontes are collected below the cuticle through which they pass out of the cell in order to form the cilia; the continuity between the two is readily discernible in favorable conditions (fig. 40).


Although, from these observations no definite conclusion as regards the development of the basal corpuscles and the rootlets, can be reached there is no indication of the formation of cilia from pre-existing basal corpuscles. The chondriocontes give rise to the cilia. The basal corpuscles and the rootlets are, in all probability, formed by the special differentiation of the chondriocontes remaining behind in the cell.


Contrary to the view of Hammer ('97), Henry^('99), Ach ('02) and Jordan ('13), I could not find any division figure, either mitotic or amitotic, in ciliated cells of the efferent tubules; the regeneration of these cells is, rather, effected, as mentioned before, by the transformation of cells with brush borders ; the ciliogenesis which occurs in this corresponds in all respects with that of the embryonal stage; first, the chondriocontes increase largely in number and are chiefly accumulated between the nucleus and the distal cell-border (fig. 79). It is certain that they are not derived from the centrosome, though nothing is known of the manner of their increase. They then proceed towards the distal cell-border and transform into rod-Uke corpuscles arranged in linear rows (fig. 80). These rods can scarcely be said to be basal corpuscles, for they stain in the same way as the chondriocontes. They emit, at successive periods, short initial cilia (fig. 81), which gradually lengthen (fig. 82). In figure 82 there is, as usual, a centrosome consisting of two granules, the existence of which indicates that the latter takes no part in the formation of the cilia. That the process of ciliogenesis takes place in cells with brush borders has been already referred to; the developing cilia are seen to pass, not through the interspace between the hairs of the brush border, but through their axes, as represented in figure 84 (the cell to the right). Since the cilia are longer than the hairs, it is evident that the former project l)eyond the distal extremity of the latter, which can no longer be detected in fully developed ciliated cells.



Whether or not the ring-shaped corpuscles which I found in cihated cells of the efferent tubule of the rat, are the same as the 'Zentralkorperballen' of Benda and Ikeda, I am not certain; I have not, however, sufficient evidence to show that they have any connection with ciliogenesis. It seems probable that these structures, as well as the granules or threads described by Fuchs, are similar to the chondriocontes.


As already mentioned, the ciliated cells in the intestine of Anodonta multiply by mitosis; at the beginning of the metaphase the basal corpuscles and the cilia disappear, and in the anaphase the ciliogenesis sets in; the process is, in the main features, similar to that of embryonic development of the cilia. The chondriocontes now give the previous staining reaction and are collected above the nucleus (fig. 21). At successive periods, some of these filaments proceed towards the cuticle, their upper ends often swelling out into bulbous enlargements, from which the cilia pass out through the cuticle (fig. 22). These granules are, in all probability, not identical with basal corpuscles, but mere local accumulations, of the mitochondrial substance.


1.. The ciliary apparatus is produced by the differentiation of mitochondria or chondriocontes, whether the process occurs in embryonic or in adult cells.

2. The centrosome takes no part in the formation of the cilia.


Henneguy and Lenhossek, at nearly the same time ('98), but independently, formulated an hypothesis that the ciliary apparatus, especially the basal corpuscles aie derived from the centrosome; an hypothesis which has since obtained some adherence among histologists (Benda '00, Flirst '00, Holmgren '03, Fachs '04, Joseph '05, Ikeda '06). The arguments in favor of this


hypothesis are: (a) the shape and position of basal corpuscles in the ciliated cell correspond with those of the centrosome in neighboring non-ciliated cells (Lenhossek, Holmgren) ; (6) viewed in unstained preparations, the basal corpuscle refracts light as strongly as does the centrosome (Lenhossek) ; (c) both the above structures show stain in the same way (Lenhossek) ; (d) the ciliated cell lacks centrosomes (Lenhossek, Fiisst, Fuchs, Jordan); (e) the ciliated cell bears resemblance either to a spermatid (Henneguy, Lenhossek) or to a ' Zentralgeisselzelle' (Joseph) ; (/) no mitotic figure is seen in the ciliated cell, because the latter lacks centrosomes (Henneguy, Lenhossek, Fuchs, Joseph) ; (g) the basal corpuscles arise from the centrosome.

There are many other investigators whose opinions are adverse to the above hypothesis ; they bring forward the following facts in support of their view ; (a) the granules which stain black with iron-haematoxylin cannot always be said to be centrosomes, for there are many other cell-constituents which give the same reaction (Studnicka '99, Fischer '99) ; (6) contrary to the view of Lenhossek, the centrosome is often situated deeply; (c) it can scarcely be said that the ciliated cell lacks centrosomes when it is not met with (Merkel '08) ; {d) the ciliated cell has a centrosome (Studnicka '99, Eismond '00, Fischel '00, Henry '00, Gurwitsch '00, Wallengren '05, Erhard '10, Tschassownikow '13); (e) sometimes mitosis is observed in ciliated cells (Gurwitsch '00, '01, Maier '03, Wallengren '05, Erhard '10, Gutheil '11); (/) the basal corpuscles are not derived from the centrosome, but from other cell-constituents (Gurwitsch '00, '01, Wallengren '05, Erhard '10, Gutheil '11).

My observations also are not in accordance with the hypothesis of Henneguy and Lenhossek; they are summarized as follows:

1. The basal corpuscle refracts light strongly and stains deeply with iron-haematoxylin, but these properties are not characteristic of these bodies.

2. I have found centrosomes in many ciliated cells.

3. The existence of the centrosome does not always signify the occurrence of mitosis, for the ciliated cells of vertebrates do not multiply by mitosis, though they contain distinct centrosomes.


4. The resemblance of ciliated cells to spermatids or to 'Zentralgeisselzellen' must be regarded as accidental.

5. The ciliary apparatus, especially the ciha, are not derived from the centrosome, but from the chondriocontes.

Kanazawa, Japan, April 14, 1916.


AcH, A. 1902 Beitrage zur Histologie des menschlichen Nebenhodens. Inaug.

Diss., Wiirzburg. Apathy, S. 1897 Das leitende Elemente des Nervensystems und seine topo graphische Beziehungen zu den Zellen. Mitt. zool. Stat. Neapel, Bd. 12. Benda, C. 1898-1899 Weitere Mitteilung liber die Mitochondrien. Verh. d.

phys. Ges. zu Berlin, Jahrg. 1898-1899.

1900-1901 tjber neue Darstellungsmethoden der Centralkorperchen

und die Verwandschaft der Basalkorper der Cilien mit Centralkorperchen. Verh. d. phys. Ges. zu Berlin, Jahrg. 1900-1901. BocKENDAHL, A. 1885 tJber die Regeneration des Trachealepithels. Arch.

mikr. Anat., Bd. 24. Brasil, L. 1904 Contribution a la connaissance de I'appareil digestif de

Annelides polychetes. Theses present a la Faculte d. sciences de

rUniversite de Paris. Carriere, J. 1882 Die Fussdrlisen der Prosobranchier und das Wassergefass System der Lamellibranchier und Gastropoden. Arch. mikr. Anat.,

Bd. 21. Child, C. M. 1907 Amitosis as a factor in normal and regulatory growth.

Anat. Anz., Bd. 30. Des Cilleules, J. 1914 Recherches sur la signification physiologique de I'ami tose. Arch. d'Anat. micr.. Tome 16. DoBELL, C. C. 1907 Physiological degeneration in Opalina. Quart. Journ.

micr. Science, N. S., N. 204. Drasch, O. 1880-1881 Zur Frage der Regeneration des Trachealepithels.

Sitz.-Ber. k. k. Ak. d. Wiss., Math.-nat. CI., Bd. 81. Eberth, C. J. 1866 Zur Kenntnis des feineren Baues der Flimmerepithelien.

Vixchow's Arch., Bd. 35.

1866 Zur Entwicklung d. Gewebe im Schwanzende der Froschlarven.

Arch. mikr. Anat., Bd. 2. EiCHHORST, H. 1875 Uber die Entwicklung des menschlichen Rlickenmarks

und seiner Formelemente. Virchow's Arch., Bd. 64. EiMER, Th. 1877 Weitere Nachrichten liber den Bau des Zellkerns, nebst

Bemerkungen liber Wimperepithelien. Arch. mikr. Anat., Bd. 14. EiSMOND, J. 1900 ijber die Natur d. sog. kinetischen Centrum der Zelle.

Verh. d. anat. Ges. zu Pavia. Engelmann, Th. W. 1868 tJber die Flimmerbewegung. Jenaische Zeitsch.,

Bd. 4.

1879 Physiologie der Protoplasma- und Flimmerbewegung. Hermann's Handbuch der Physiol., Th. I.


Engelmann, Th. W. 1880 Zur Anatomie und Physiologic der Flimmerzellen.

Pfliiger's Arch., Bd. 23. Erhard, H. 1909 Die Henneguy-Lenhossek'sche Theorie. Ergeb. Anat. u.


1910 Studien fiber Flimmerzellen. Arch. f. Zellforsch., Bd. 4.

1911 Dilplosomen und Mitosen im cilientragenden Ependym eines Haifischembrj'os. Anat. Anz., Bd. 38.

FiscHEL, A. 1900 Zur Histologie der Urodelen-Cornea und des Flimmerepi thels. Anat. Hefte, Bd. 15. Fischer, A. 1899 Fixierung, Farbung und Bau des Protoplasmas. Jena. Flemming, W. 1885 tjber die Regeneration verschiedener Epithelien durch

mitotische Zellteilung. Arch. mikr. Anat., Bd. 24.

1898 Morphologie der Zelle. Ergeb. Anat. und Entwickl., Bd. 7. FoL, H. 1896 Lehrbuch der vergleichenden mikr. Anat. Leipzig. Fhenzel, J. 1886 Zum feineren Bau des Wimperapparates. Arch. mikr. Anat.,

Bd. 28. Friedreich, N. 1858 Einiges iiber die Struktur der Cylinder- und Flimmer epithelien. Virchow's Arch., Bd. 15. FucHS, H. 1902 tlber das Epithel im Nebenhoden der Maus. Anat. Hefte,

Bd. 19.

1904 tJber Beobachtungen an Sekret- und Flimmerzellen. Anat.

Hefte, Bd. 25. FuRST, C. 1900 Haarzellen und Flimmerzellen. Anat. Anz., Bd. 18. Gaule, J. 1881 Das Flimmerepithel der Aricia foetida. Arch. Anat. u. Physiol., Physiol. Abt. Geist, H. 1912 Untersuchungen iiber die Histologie der Uterusschleimhaut.

Arch. mikr. Anat., Bd. 81. GuRwiTSCH, A. 1900 Zur Entwicklung der Flimmerzellen. Anat. Anz., Bd.


1901 Studien iiber Flimmerzellen. Arch. mikr. Anat., Bd. 57.

1902 Der Haarbiischel der Epithelzellen in Vas epididymis des Menschen. Arch. mikr. Anat., Bd. 59.

1904 Morphologie und Biologie der Zelle. Jena. Gutheil, F. 1911 iiber Wimperapparat und Mitose von Flimmerzellen. Zool.

Anz., Bd. 37. Hammer, A. J. 1897 tJber Sekretionserscheinungen im Nebenhoden des Hundes.

Arch. Anat. u. Physiol., Anat. Abt., Suppl. Heidenhain, M. 1890 Beitriige zur Kenntnis der Topographic und Histologie

der Kloake und ihrer driisigen Adnexa bei den einheimischen Tritonen.

Arch. mikr. Anat., Bd. 35.

1899 Beitriige zur Aufklarung des wahren Wesens der faserformigen Differenzierungen. Anat. Anz., Bd. 16.

1907 Plasma und Zelle. Jena. Heidenhain, R. 1888 Beitrage zur Histologie und Physiologie der Diinndarm schleimhaut. Pfliiger's Arch., Bd. 43, Suppl. Henneguy, L. F. 1898 Sur le rapports des cils vibratiles avec les centrosomes.

Arch. d'Anat. micr.. Tome 1. Henry, A. 1900 Etude histologique de la fonction s6cretoire de I'epididyme

chez les vertcbrcs superieurs. Arch. d'Anat. micr.. Tome 3.


Holmgren, N. 1903 Studien iiber Cuticularbildungen. Anat. Anz., Bd. 22. Ikeda, R. 1906 Uber das Epithel im Nebenhoden des Menschen. Anat. Anz.,

Bd. 29. Jordan, H. E. 1913 Amitosis in the Epididj'mis of the mouse. Anat. Anz.,

Bd. 43. Joseph, H. 1903 Beitriige zur FUmmerzellen und Cent rosomenf rage. Wien.

1905 tJber die Centralkorper der Nierenzellen. Verh. d. anat. Ges. KoLACEV, A. 1910 tJber den Bau des Flimmerapparates. Arch. mikr. Anat.,

Bd. 76. KoLTzoFP, N. K. 1906 Studien iiber die Gestalt der Zellen. I. Untersuch ungen iiber die Spermien der Dekapoden. Arch. mikr. Anat., Bd. 67 Knatjff 1867 Das Pigment der Respirationsorgane. Virchow's Arch., Bd. 39. KuPERWiESER, H. 1906 Untersuchungen iiber den feineren Bau und Metamorphose des Cyphonautes. Zoologica, Bd. 19. Langerhans, p. 1876 Zur Anatomie des Amphioxus hinceolatus. Arch. mikr.

Anat., Bd. 12. Lenhossek, M. v. 1898 Uber Flimmerzellen. Verh. d. anat. Ges. zu Kiel. Maier, H. N. 1903 Uber den feineren Bau der Wimperapparate der Infusorien.

Arch. f. Protistenk., Bd. 2. Marchi, p. 1866 Beobachtungen iiljcr Winiper-Epithel. Arch. mikr. Anat.,

Bd. 2. Maximow, a. 1908 tJber Amitose in den embryonalen Geweben bei Saugetiere.

Anat. Anz., Bd. 33. Merkel, Fr. 1908 Epithelium. Ergebn. d. Anat. u. Entwickl., Bd. 18. Metalnikoff, S. 1900 Sipunculus nudus. Zeitsch. f. wiss. Zool., Bd. 68. Meves, Fr. 1899 tJber den Einfluss der Zellteilung auf den Sekretionsvorgang,

nach Beobachtungen an der Niere der Salamanderlarve. Festsch.

zum 70. Geburtstag v. C. v. Kuppfer, Jena.

1901 iiber die sog. wurmfomigen Samenfiiden von Paludina und ihre

Entwicklung. Verh. d. anat. Ges. Meves, Fr. und Tsukaguchi, R. 1914 Uber das Vorkommen von Plastosomen

im Epithel von Trachea und Lunge. Anat. Anz., Bd. 46. Nissen, F. 1886 iiber das Verhalten der Kerne in den Milchdriisen wahrend

der Absonderung. Arch. mikr. Anat., Bd. 26. NussBAUM, M. 1877 Ein Beitrag zur Lehre von der Flimmerbewegung. Arch.

mikr. Anat., Bd. 14. OsAWA, G. 1897 Beitrage zur Lehre von den Eingeweiden der Hatteria punctata. Arch. f. mikr. Anat., Bd. 49. Patterson, J. Th. 1908 Amitosis in the pigeon's egg. Anat. Anz., Bd. 15. Peter, K. 1899 Das Centrum fiir die Flimmer- und Geisselbe'n'egung. Anat.

Anz., Bd. 15. Pfitzner, W. 1880 Die Epidermis der Amphibien. Morph. Jahrl)., Bd. 6. Plate, L. 1898 Uber regenerative Amitose, Degenerationserscheinungen und

Phagocytose in den Atemrohren der Janellen. Arch. mikr. Anat., Bd.

51. PoLowzow, W. 1904 Uber kontraktile Fasern in einer Flimmerepithelart und

ihre funktionelle Bedeutung. Arch. mikr. Anat., Bd. 63. Prenant, a. 1898 Sur le protoplasme superieure. Etude critique. Journ.

de I'Anat. et de la Physiol.


Prenant, a. 1914 Les appareils cilies et leurs derives. Journ. d'Anat. et de

la Physiol., Anne.e 48-50. Putter, A. 1903 Die Flimmerbewegiing. Ergebn. d. Physiol., Jahrg. 2. Rabl-Ruckhard 1868 Einiges liber Flimmerepithel und Becherzellen. Arch.

f. Anat., Physiol, u. wiss. Med. Reichenow, E. Die Riickbildungserscheinungen am Anurendarm wahrend der

Metamorphose und ihre Bedeutung fiir die Zellforschung. Arch. mikr.

Anat., Bd. 72. Saguchi, S. 1915 tjber Sekretionserscheinungen an den Epidermiszellen von

Amphibienlarven nebst Beitragen zur Frage nach der physiologischen

Degeneration der Zellen. Mitteil. d. med. Fakultiit Universitat zu

Tokyo, Bd. 14. Schiefferdecker, P. und Kossel, A. 1891 Gewebelehre. Braunschweig. Schmidt, C. 1882 tjber eigentiimliche, aus dem Flimmerepithel hervorgehende

Gebilde. Arch. mikr. Anat., Bd. 20. Schneider, K. C. 1902 Lehrbuch der vergleichenden Histologie der Tiere.

Jena. Schultze, O. 1907 ttber den Bau und die Bedeutung der Aussencuticula der

Amphibienlarven. Arch. mikr. Anat., Bd. 69. Schultze, F. E. 1869 iiber cuticulare Bildungen und Verhornung von Epi thelzellen bei den Wirbeltieren. Arch. mikr. Arch., Bd. 5.

1888 tlber die inneren Kiemen der Batrachierlaven. I. Mitteil. :

tJber das Epithel der Lippen, der Mund-, Rachen- und Kiemenhohle.

Berlin. Simroth, H. 1876 Zur Kenntnis des Bewegungsapparates der Infusorienthiere.

Arch. mikr. Anat., Bd. 12. Stuart, A. 1865 Uber die Entwicklung einiger Opistobranchier. Zeitsch.

wiss. Zool., Bd. 15. Studnicka, F. K. 1898 Uber die intercelluliiren Verbindungen, den sog. Cutic ularsaum u. d. Flimmerbesatz der Zellen. Sitzb. d. kgl. bohm. Ges.

d. Wiss., Mat.-naturw. CI.

1899 Uber Flimmer- und Cuticularzellen mit besond. Bercksichitigung der Centrosomenfrage. Sitzb. d. kgl. bohm. Ges. d. Wiss., Mat.naturw. CI.

1900 Untersuchungen iiber den Bau des Ependyms der nervosen Centralorgane. Anat. Hefte, Bd. 15.

TscHASsowNiKOw, S. 1913 Uber Becher- und Flimmerepithelzellen und ihre

Beziehungen zueinander. Zur Morphologic und Physiologic der Zent tralkorper. Arch. mikr. Anat., Bd. 84. Wallengrex, H. 1905 Zur Kenntnis der Flimmerzellen. Zeitsch. f. allg.

Physiol., Bd. 5. Wolff, G. 1889 Die Cuticula der Wirbelthierepidermis. Jenaische Zeitsch.,

Bd. 23. Zimmerm.\xx, K. W. 1898 Beitrage zur Kenntnis einiger Driisen und Epi thelien. Arch. mikr. Anat., Bd. 52.


All figures are drawn with the Zeiss camera under Zeiss ^-j oil-immersion objective and Zeiss compensating ocular 18, at the distance of 250 mm. tube length 160 mm. By reproduction they were reduced to three-fourths of the original. Figures 2, 3, 5, 7, 10, 14, 15, 30-33, 37-39, 46, 47, 49, 50, 55, 57, 69, 67, 73, 79-83, 86, 90, 92, 95 are from preparations stained by iron-haematoxylin, after fixation with sublimate; the others from Meves preparations.



ri to ^ Ciliated cells from the foot of Anodonta, figure 3 surface view. -4, 5 From gills of Andonta, figure 4 lateral cell, figure 5 surface view of the lateral surface of the gill-torus, a, lateral cell; b, corner-cell.

6 to 8, 16 to 20 From the intestine of Anodonta, figure 16 surface view.

9, 13 to 15 From the intestine of Helix.

10 to 12 From the hepatic duct of Helix.











'lTtI![nT!niIVBnsinlii»l;MKHIIl!iiniMl.l.liiniii::ri :::,i-ii.?7?





21, 22 From the intestine of Anodonta.

23 From the epidermis of the tail of 6 mm. Hyla larva.

24 to 27 From the epidermis of the tails of 12 mm. Rhacophorus larvae.

29 From the epidermis near the olfactory pit of 11 mm. Hyla larva.

30 to 32 From the epidermis near the olfactory pit of Hynobius larvae.

33 Surface view of a ciliated cell from the abdominal epidermis of 6 mm. Hyla larva.

34 to 36, 38 From the tela epithelium of Hyla larvae.

37 From the tela-epithelium of 10 mm. Rhacophorus larva.

39 Flat view of a ciliated cell from the tela of Hynobius larva (13 mm.).

40, 41 From the ciliated infundibulum of Hyla larvae.




^^mmimm^' '^^MM






  • ■.• ^ ' '»»

mm - --. -^-

• ■' •' .• ' J ! '








42 to 47 From the pharynx of Rana esculenta, figure 47, surface view.

48 to 50 From the pharynx of Bufo, figure 50, surface view.

51 to 53 From the pharynx of Rana temporaria.

54 to 55 From the pharynx of 20 mm. Hynobius larva.

56 From the pharynx of 15 mm. Rhacophorus larva.

57 From the gall-duct of Rana esculenta.

58 to 62 From the gall-duct of Rana temporaria.




mil >W




63 From the oviduct of Rai;a.;esculenta.

64 and 65 From the oviduct of Rana temporaria.

66 to 67 From the oviduct of the triton, figure 67, surface view.

68 A ciliated cell from the abdominal endothelium of Rana temporaria.

69 to 71 From the trachea of Trigonocephalus.

72, 73 From the efferent tubule of testis of Elaphis.

74, 75 From the efferent tubule of testis of Trigonocephalus.

76 From the efferent tubule of testis of a lizard.

77, 78 From the oviduct of a lizard.

79 to 83, 85, 86 From the efferent tubule of testi.s of the mouse, figure 83 surface view.

84, 87 to 90 From the efferent tubule of testis of the rat, figure 90 surface view.

91 to 95 From the trachea of the rat, figure 95 surface view.




K ;l

1 s^ itt, -^




Department of Anatomy, Washington University Medical School



Introduction 282

Material and methods 285

Part I. Observations 287

Basal plate. Embryos of 23.1 mm 287

Occipital region 289

Atlas and epistropheus 293

Basal plate and occipital region in smaller embryos 294

Otic region. Embryos of 23.1 mm 298

Pars canalicularis. Outer form 299

Pars cochlearis 302

Cavities of the otic capsule ■ 305

Otic region of smaller embryos. 311

Nerves in the otic region 314

Blood vessels in the otic region. . ; '. 316

Orbito-temporal region in embryos of 23.1 mm 317

Orbito-temporal region in smaller embryos 325

Ethmoidal region in embryos of 23.1 mm 329

Ethmoidal region of smaller embryos 340

Part II. Discussion 342

Occipital region 342

Basal plate ■ 342

Atlas 351

Occipito-atlantal articulation 353

Plane of the foramen magnum 354

Basal fissures 355

Foramen magnum 356

Otic region 357

Position of the otic capsules 357

Origin of the cartilaginous otic capsule 359

Foramina acustica and meatus acusticus internus : 363

Foramen perilymphaticum 364

Cavum vestibulare 365

Lamina parietalis and tectum posterius 366

Facial and acustic nerves 369




Orbito-temporal region 370

Hypophyseal cartilage 370

Fenestra basicranialis posterior 371

Crista transversa and dorsum sellae 371

Foramen hypophyseos 374

Ala temporalis 374

Pterygoid 380

Carotid foramen 381

Ala orbitalis 384

Ethmoidal region 385

Region of the olfactory fenestra 385

Floor of the nose and Jacobson's cartilage 387

Tectum nasi 390

Paranasal cartilage 391

Lamina antorbitalis 394

Summary 396

Literature 401

Explanation of figures 405

Abbreviations. ; 406

Figures 409


Studies of the primordial cranium of mamcmals have in most instances dealt with but a single stage, and offer, therefore, almost no data, on developmental processes. Moreover, there seems to be a tendency, unfortunately, to pursue the study after this manner and so to continue limiting inquiry to the methods of comparative, descriptive anatomy. This criticism cannot be made of investigation of the crania of the ichthyopsida. It has been rather the rule than the exception for research in this group to include as complete a survey of developmental processes as the technique of the time permitted; that is, the studies proceeded toward the solution of cranial problems by comparative embryological methods. No doubt, certain physical conditions have in the one case been an obstacle to embryological work. The difficulty of securing material in control and of sufficient amount, and the great extent to which the investigation of even a few stages leads one, are factors which have determined in no small degree the methods and results of research on the crania of higher forms. However, these difficulties are not serious and should not stand in the way of gaining for the mara


malian cranium that knowledge of its development which is so much desired.

For the study of problems of the mammalian primordial cranium, the domestic cat offers material possessing certain advantages. In general, the of chondrification proceed sufficiently slowly to permit of easy determination of the origin of parts and of subsequent study of their individual histories. Also, the cat among the domestic mammals has remained standardized to a degree not exceeded, if equalled, by any other, and is on this account of value for anatomical study. Yet, notwithstanding these points in its favor, the house cat has not been utilized for cranial study to the extent that might be anticipated. There is at the present time no paper on the development of the cranium as a whole and no extensive description of a single stage in the course of its development. Not even Parker described the primordial cranium of the domestic cat. The only investigation which attempts to deal with the whole chondrocranium is that by Decker in which brief accounts of two rather advanced stages are presented. Wincza's communication discusses a number of questions, for the solution of which cat material has been employed, but it does not include all regions of the craniimi. This valuable paper is available to a relatively small number of investigators since it was pubhshed in the Polish language.

\ATiat has been said in regard to cat for the study of the whole chondrocranium does not apply in the case of investigations of special problems, although even in the latter case this material has been used to a more hmited extent than might be expected. Of the cranial regions usually recognized, the occipital in cat has received less attention than have the others. The otic region is represented by a number of papers concerned for the most part with late embryonic stages and adult conditions. Among these, the pubhcations of Spence, Bondy, van Kampen, and Denker have been consulted in the present work. Wincza's paper deals especially with the ala temporalis and carotid foramen; an article by Ai-ai includes the cat in a study of the craniopharyngeal canal, and a paper by Wilhams deals


with the notochord in its later history. In the ethmoidal region interest has been centered upon the skeletal structure of the floor of the nose. Harvey, Broom, and Zuckerkandl have each contributed valuable descriptions of the cartilaginous framework about Jacobson's organ and the incisive ducts in cat.

In the present paper the monographs on the anatomy of cat by Strauss-Durckheim, Wilder and Gage, Jayne, and Reighard have been very helpful, as have those works of a more general scope, but including the cat, by Chauveau, Paul Martin, and Weber. Finally should be mentioned those studies, in which cat has offered the material, of head structures other than the cranium which are inseparably connected with the problems of the skeleton, many of which have aided the present investigation. Among these are, specially, Retzius' researches on the ear; Goppert and Corning on the ocular muscles; Peter, Born and Seydel on the nose; Froriep and van Wijhe on the occipitospinal nerves.

This paper is concerned with some of the problems of cranial development in mammals, the chondrocranium (exclusive of the visceral skeleton) in cat serving as the material for study. It was the intention to give the descriptive portions of the several cranial regions, as far as possible equal treatment, but to limit the discussion to a few problems upon which some hght was thrown by the structure and development of the cranium in the particular form under consideration. The occipital region was looked to with keen interest as an unworked field for observation, and for any evidence of those processes revealed by Weiss, Gaupp, and Weigner in mammals which indicate a close resemblance to, if not an actual repetition of the steps in the development of the atlas. As stated above, the work on the otic region, hitherto, treats of late embryonic and adult conditions. In the present paper, attention has been directed mainly to conditions of early development, such as the origin of the cartilaginous otic capsule, the comparison of the embryonic vestibule with t'hat of reptiles and the theoretical questions concerned with the development of the cochlear skeleton. Consideration of questions relating to the development and signifi


cance of the ala temporalis and to the formation of the hypophyseal skeleton forms the main part of the study of orbitotemporal region. Concerning the origin and early history of the ethmoidal skeleton of mammals comparatively little is known, whereas, on the contrary, a great literatm-e exists on the development of the nasal conchae, floor of the nose and nasal sinuses. Cat material was found favorable for the investigation of many early processes of ethmoidal development, and to a discussion of these, practically the whole of the section on the ethmoidal region has been devoted.


The present study starts with a cat embryo of 23.1 mm., series no. 466 of the Harvard Embryological Collection. From this, a model of the chondrocranium was reconstructed by students of the Harvard Medical School under the direction of the author while a teaching fellow in histology in that institution.^ The model, enlarged thu'ty-three times, includes, in addition to the chondrocranium, the ossifications, main bloodvessels and nerves. It is an example of care and industry and as such has been on the one hand invaluable as material for study and on the other a reminder of the enthusiasm and earnest effort of those students, then in their first year of medical study, Messrs. G. D. Cutler, I. Gerber and R. D. Leonard, who responded to the suggestion to undertake the labors of cranial reconstruction.

Through the kindness of the late Professor Minot and Professor F. T. Lewis, the model, together with the following series of cat embryos from the collection at the Harvard Medical School, was placed at my disposal in St. Louis.

1 The model was reconstructed in 1907. The present investigation was undertaken in St. Louis several years later but on account of interruptions its progress was delayed. I take this opportunity to acknowledge the many courtesies of Prof. J. S. Kingsley and the staff of the Marine Biological Laboratory at South Harpswell, Maine, where during the summer of 1915 it was my privilege to work.


Cat 9.7 mm., frontal, Series 448

Cat 10.6 mm., frontal, Series 476

Cat 12 .0 mm., frontal, Series 404

Cat 12.0 mm., horizontal of head. Series 403 Cat 12.0 mm., sagittal. Series 400

Cat 17 . mm., sagittal. Series 492

I am indebted also to Professor Huntington for the use of series of cat embryos of the following stages, from the collection at the College of Physicians and Surgeons, Columbia University.

Cat 10 mm., Series 500

Cat 11 mm., Series 473

Cat 12 mm., Series 263

Cat 13 mm.. Series 262

From the collection in the Department of Anatomy of Washington University the following were studied:

Cat 12 mm., frontal. Series 50 Cat 12 mm., transverse. Series 52 Cat 15 mm., transverse. Series 56 Cat 17 mm.. Series 80

Cat 17 mm., sagittal. Series 81 Cat 17 mm., transverse. Series 82

Several special reconstructions were made of complex regions. One is of the left otic capsule of the 23.1 mm. embryo; a second represents the course and relations of the facial nerve in the same specimen; a third, the eye-ball and its muscles in relation to the chondrocranium; a fourth the cranial relations of the notochord. The atlas and epistropheus were included in the main model. The base of the cranium in the region of the carotid artery was reconstructed from a 12 mm. embryo; from the same a reconstruction of the hypophyseal cartilage was made. The nasal capsule and nasal sac have been modeled from an embryo of 17 mm.

The study of all stages of the chondrocranium was facilitated by a collection of preparations made by the van Wijhe ('02)


method. These are of great value for studies of the cartilaginous skeleton, especially when used in comparison mth the sections. Cat embryos were fixed in formalin, stained with methylene blue, decolorized, cleared and preserved in hquid Canada balsam. The last procedure, which departs shghtly from van Wijhe's final treatment, is of much advantage in permitting the free handling and turning of the specimen so that it may be easily observed from every aspect. The following stages were prepared and studied: Cat embryos of 10 nun., 12 mm., 15 mm.^ 17 mm., 20 nmi., 24 mm., 30 mm., and 35 mm.

As stated above, the present paper centers upon a particular stage, with which earlier and later stages have been compared. In accordance with this method the structures of each region of the chondrocranium as found in the stage of 23.1 mm., that is, the stage of the model, are first described, then younger stages are considered.


Basal plate. — Embryos of 23.1 mm.

The term basal plate is used here to designate that portion of the floor of the chondrocranium corresponding in cephalocaudal extent with the cranial part of the notochord, and in lateral extent to the level of the basal foramina and cochlear capsule.

The basal plate in the embryo of 23.1 mm. is represented in figures 1, 2 and 3. In the otic region it is very narrow; in the occipital region, on the contrary, very broad. Foramina are present at the margin of the basal plate, between it and the cranial side walls, but no openings exist in the plate itself. These marginal foramina are the fissura basi-cochlearis posterior^ the jugular and hypoglossal foramina.

The anterior, or otic portion of the basal plate, subcyhndrical in shape, Ues between the cochlear prominences, and, within


the cranium, forms the bottom of a deep sulcus whose sides are constituted by the median walls of the cochleae. This sulcus is occupied by no part of the brain, but is filled with a loose web of mesenchymal tissue, in which the basilar artery and abducent nerves run (fig. 22). Continuity between the cochlear prominence and basal plate is brought about by means of a thin strip of cartilage interrupted by the marginal foramina above mentioned. This strip is merely a septum between two oppositely placed grooves, one outside, the other within the cranium, which, in their semicircular courses, demarcate the limits between the ear capsule and cranial floor (figs. 1, 2). These grooves extend anteriorly on the outside of the cranium beyond the confines of the basal plate and include the carotid foramina; posteriorly, both within and without the cranium, they have a somewhat lateral direction between the occipital division of the floor and the otic capsule. Here are two large openings, the jugular foramen, to be presently described, and the fissura basicochlearis posterior (figs. 2, 12). The latter is sickle-shaped, broad laterally where it approaches the jugular foramen, from which it is separated by the commissura basi vestibularis. It gives passage to no nerve or large vessel.

The occipital portion of the basal plate forms the floor of the occipital region of the cranium. It presents a slightly concave intracranial surface and a convex face directed toward the nasopharyngeal duct ; at the sides it is continuous, at the level of the hypoglossal canals and jugular foramina, with the lateral occipital walls. The caudal free margin is concave from side to side, forming the ventral margin of the foramen occipitale magnum, where on each side is the basal, smaller part of the occipital condyle (fig. 19); between these is a shallow incisura occipitahs anterior.

The following observations were made on the position, relations and termination of the cranial portion of the notochord. As shown in figures 14 and 19, this structure passes through the middle of the body of the epistropheus, inclines ventrally on entering the dens, and holds a position ventral of the middle of the dens throughout the rest of its vertebral com-se. It leaves


the dens from the ventral surface, not at its extremity, and enters the hgamentum apicis dentis. It then passes over into the cranium, lying at first just beneath the perichondrium of the dorsal surface of the cranial floor, with the great thickness of the caudal edge of the basal plate beneath it. From this point the chorda dips into the cartilage of the floor as it extends forward, lying midway between the cranial and pharyngeal surfaces in the middle third of the plate. At no point does it sink nearer the pharyngeal surface nor was there observed any trace of a connection with the pharyngeal wall. Proceeding toward the sella turcica, the notochord gradually approaches the intracranial surface of the basal plate, until it attains the level of the dorsum sellae where it makes a rather abrupt bend so as to come to lie beneath the perichondrium of the caudal surface of the back of the saddle. Its terminal piece is marked by distortion to some extent and by a few irregular turns. Although shrunken in many places, and showing other evidences of degeneration, the head notochord is, nevertheless, continuous throughout. Within the apical Ugament it is considerably expanded.

Occipital region

From the basal plate, the side walls of the occipital region, the lateral occipital arches, continue laterally and dorsally, becoming confluent with the otic capsule, parietal plate and tectum posterius (figs. 1, 2, 3, 4). They lie in a plane transverse to the longitudinal axis of the posterior half of the cranium, but, owing to the flexure of the long axis of the whole cranium, this is at the same time, parallel with the plane of the floor of the nose.

The lateral occipital arches are connected with the pars canalicularis of the otic cap.sule by synchondrosis, marked in the sections by a narrow plane of young cartilage extending a considerable distance from the jugular foramen toward the parietal plate. Within the cranium a deep, wide groove, the sulcus sigmoideus, lodging the transverse sinus, lies opposite the ventral portion of the synchondrosis and leads to the jugular


foramen. Outside the wall a corresponding groove separates the paracondyloid process (figs. 2, 3) from the pars canalicularis of the otic capsule. The caudal free margin of the lateral occipital arch enters into the boundaries of the foramen magnum (figs. 1, 2, 3). Its ventral portion forms the larger part of the occipital condyle; its dorsal portion, a rough angular process directed medially. The processes of opposite sides mark the dorsal limit of the foramen magnum at this stage and afford attachment to the spino-occipital membrane. Dorsad of the level of these processes the lateral occipital arches are continued into the broad, curved posterior portions of the parietal plates, between which the tectum posterius extends in an arch from side to side.

Within the cranium, at the junction of basal plate and lateral occipital arch, is a rounded prominence, the cartilaginous precursor of the tuberculum jugulare, separating the entrance to the hypoglossal canal from the fossa occipito-canalicularis (fig. 1). At its cephalic end, this elevation, broadening considerably, becomes continuous with the basi-vestibular commissure (figs. 9, 10, 19). This commissure is united with the otic capsule at the boundary between the medial wall of the prominentia utriculo-ampullaris inferior and the cochlear capsule, and forms here a prominence making the posterior wall of the internal acustic meatus. The jugular tubercle, which presents much more the form of a ridge than of a tubercle, stands within the cranium opposite the paracondyloid process. Both processes contribute to the formation of the deep caudal wall of the jugular foramen, a relation observed by von Noorden in the human embryo. The occipito-canalicular fossa lies between the inferior ampullary eminence and the basivestibular commissure and connects the sigmoid sulcus with the jugular foramen.

The jugular foramen (figs. 1, 2, 12) when viewed from within the cranium is crescentic, the convex side being formed caudally by the tuberculum jugulare and medially by the commissura basivestibularis ; the concave margin, directed laterad, is constituted by the medial wall of the prominentia utriculo-ampullaris inferior. The vena jugularis interna and the accessorius, vagus


and glossopharyngeus nerves pass straight through the foramen. Although the paracondyloid process forms a deep wall for the foramen posteriorly, there is no inclination forward of this process nor a lamina alaris with attendant horizontal course of the nerves. On the extracranial surface, the foramen is divided into cephalic and caudal parts by an angular process of the cochlear wall (figs. 12, 19). The compartment posterior to the process is occupied by the jugular vein and the group of nerves; the anterior compartment, filled with connective tissue, lies within the fenestra perilymphaticum and forms a communication between the cavum cochleae and fossa occipito-canalicularis ; it is the beginning of the aquaeductus cochleae.

The hypoglossal canal (figs. 1, 2, 12) transmits the three previously united ventral roots of the hypoglossal nerve. It runs in a ventro-lateral direction, beginning caudad and medialward of the jugular foramen and terminating on the external surface at the level of the medial edge of the paracondyloid process.

The broad, square paracondyloid process (figs. 2, 3, 12, 19) projects widely from the ventral part of the pars lateralis of the occipital region, with ventral and lateral free edges, cephalic and caudal surfaces. The narrow cephalic surface enters into the wall of the jugular foramen; the ventral broad free edge is continuous medially with that part of the extracranial surface lying between the jugular foramen and the hypoglossal canal.

The foramen magnum (figs. 1, 2) is hexagonal with rounded angles, the sides consisting of the free caudal margins of the basal plate and lateral occipital arches and, dorsally, the edge of the spino-occipital membrane. The foramen is divisible into a large ventral part included between the condylar portions of the occipital arches, and a smaller dorsal region extending thence to the spino-occipital membrane. The lateral boundaries of these two divisions come together in a notch of the lateral occipital wall just dorsad of each condyle (figs. 1, 2, 3, 4). A number of small veins in the connective tissue of these notches, were connected with the sinus transversus. The plane of the whole foramen magnum is nearly transverse to the longitudinal axis of the posterior half of the cranium; however,


the positions of the planes of the dorsal and ventral divisions differ from one another, that of the ventral, condylar part is transverse, whereas that of the dorsal division is very obhque. The latter faces ventro-caudad and forms an angle, open caudally, of approximately 145 degrees with the plane of the ventral part of the foramen (fig. 3).

The occipital condyles (figs. 2, 3, 19) are a pair of oval eminences only slightly raised above the level of the ventro-lateral margins of the foramen magnum. Each condyle extends at first laterally from the incisura occipitahs anterior along the caudal margin of the basal plate to the level of the hypoglossal canal, then dorso-laterally upon the lateral occipital arch. The basilar portion is broader but shorter than the lateral which stretches out upon the occipital arch as far as the deep notch mentioned above (p. 291). The lateral and basal parts of the condyle are connected with the lateral mass and ventral arch of the atlas by intervening mesenchjrtna; an articular cavity is not present.

The tectum posterius (figs. 1 and 2) is a slender transverse bridge of cartilage arching over the brain and connecting the parietal plates from side to side. It is far from the auditory capsules caudally and dorsally, standing closer to the occipital side walls. The tectum expands laterally in joining the parietal plates, becomes narrow in its middle opposite the medulla oblongata. The caudo-ventral margin of the tectum is concave from side to side, bounds the incisura occipitalis posterior and affords attachment to the spino-occipital membrane; the cephalic, and at the same time dorsal, margin is straight, presenting no indication of a processus ascendens.

The term incisura occipitalis posterior (fig. 1) is given tentatively to the bay extending from the foramen magnum to the tectum posterius. Its lateral boundaries are the caudal margins of a pair of cartilaginous plates continuous with the occipital walls ventrally, and with the laminae parietales anteriorly. This notch is filled by the spino-occipital membrane.

Hypoglossal nerve. Three ventral roots unite to form a single nerve in the embryo of 23.1 mm. A dorsal root opposite to,


but not united with, the third ventral root, arises from the spinal cord between the ' atlas and occipital arch. In smaller specimens further evidences of a dorsal root and ganglion were observed. This component was best developed in an embryo of 15 mm. (W. U. C. ser. 52, si. 22-24) in which the dorsal root joined the third ventral root before reaching the hypoglossal foramen.

Atlas and epistropheus

In connection with the occipital region some observations on the epistropheus and atlas should be mentioned. These vertebrae have not reached their full chondrogenous development, the neural arches specially being very incomplete (fig. 9). The dens epistrophei is relatively very much -longer in the embryo of the present stage than in the adult. Its relations are also different in that it projects beyond the anterior arch of the atlas so far craniad as to enter a httle way into the cranial cavity through the occipital foramen (figs. 9, 14, 19). The atlas is remarkably massive. Just cephalad of the neural arches is a pair of processes of special interest. These project dorsad from the spot where the neural and anterior arches meet (the future lateral mass), inchning a little toward the median plane, and end in blunt extremities. These atlantal processes, which are the cartilaginous precoursors of the little bridges of bone of the adult atlas, completing the circumferences of the foramina atlantaha, form, with the neural arches, a notch on each side lodging the gangUon of the first spinal nerve and the vertebral artery. The articular surfaces of the atlas for the occipital condyles are formed at the meeting place of the neural and anterior arches. Two parts enter into the formation of the articular surfaces: (1) a thickening at the junction of the neural and anterior arches, which meets the basal portion of the occipital condyle; (2) the atlantal process which articulates with the lateral part of the condyle.


Basal plate and occipital region in smaller embryos

In van Wijhe preparations of embryos of 10 mm. (fig. 5) there appears in the floor of the occipital region a pair of faintly stained, elongate parachordal plates. These are united anteriorly by a commissure beneath the notochord, but are separated across the midline in the rest of their extent. Each plate is thickened at its caudal end, and is then continued laterally into the occipital arch. The lateral margin of the parachordal plate presents a prominent angle subdividing it into an anterior oblique part opposite the mesenchymal cochlear capsule, and a posterior straight part opposite the jugular vein. In the angle between the straight part of the margin and the lateral occipital arch are the three roots of the hypoglossal nerve. Between the parachordal plates and behind their commissure is a vacuity traversed by the notochord. This space is constricted at its middle by opposite projections from the medial edges of the parachordals. The notochord extends forward as far as the hypophysis and appears to lie in a tract of blue-stained tissue which terminates just behind the hypophyseal cartilage. Each lateral occipital arch is slender and cylindrical medially where it joins the parachordal plate, expanded and flat laterally at its free extremity. Where these two parts come together the cartilage is only slightly stained. The lateral arches at this stage lie far removed in a caudal direction from the otic capsules which are represented by parts of the anterior and posterior semicircular canals.

The atlas has the form of a transverse arch open dorsally. Each lateral extremity of the arch, deeply stained, is in the form of a broad plate presenting two processes, the neural and atlantal. The middle, slender hypochordal part of the arch is only slightly stained. The epistropheus likewise forms an arch, the ventral medial part of which includes the notochord. Here a darkly stained tract on each side is separated from the notochord medially and the deeply stained neural arch laterally by very lightly stained zones. The two tracts' near the notochord are the beginnings of the centrum of the epistropheus. Just cephalad of the epistropheal centrum and separated from


it by a transverse plane of unstained tissue is the broad conical mass of young cartilage surrounding the notochord, the beginning of the atlantal centrum. Between the apex of the latter and the commissure of the parachordals the notochord stands free of any chondrified tissue as evidenced by the absence of the blue stain about it.

Transverse sections of an embryo of 10.6 mm. (fig. 15) make clear the relations of the notochord, dens epistrophei and parachordal plates at the stage of cartilaginous structure of the latter. The parachordals are connected across the median plane, ventrad of the notochord by loose tissue. This tissue presents more and more the characters of mesenchyma when traced toward the median plane, and more and more the condition of cartilage when followed toward the parachordals into which it passes. An imperfect basal plate is thus formed in the occipital region which extends from side to side ventrad of the notochord and which passes laterally into the lateral occipital arches. Figure 15 shows the notochord covered dorsally and laterally by a layer of cellular tissue lying upon the dorsal surface of the stretch of me.senchyma connecting the parachordal plates. When followed caudad this layer increases in thickness, especially on the dorsal side of the notochord, and passes over into the mesenchyma of the centrum of the atlas.

In van Wijhe preparations of 12 mm. (fig. 6) the most important differences from the preceding stage in the posterior part of the cranium are the presence of a hypoglossal foramen, the beginning of the parietal plate and the slender cartilaginous basal plate of the otic region. The hypoglossal foramen has resulted from the development of a bar of cartilage uniting the lateral angle of the parachordal plate and lateral occipital arch outside the roots of the twelfth nerve. The beginning of the parietal plate appears standing free, dorsad of the posterior semicircular canal of the otic capsule. The basal plate of the otic region is continuous behind with the primary commissure of the parachordals, terminates in front in an expansion which extends nearly as far as the cochlear wall, and anteriorly projects somewhat dorsad of the hypophyseal cartilage.


In sections of 12 mm. embryos (fig. 17) the notochord, as it enters upon its cranial course, lies at first dorsad of the still imperfect basal plate, then sinks into the otic portion of the plate; it is now surrounded on all sides by cartilage at the level of the primary parachordal commissure. The notochord emerges from the basal plate at its free anterior end-, which is inclined dorso-cephalad, and terminates in the mass of mesenchyma which fills the interval (fenestra basicranialis posterior) between the expanded end of the basal plate and the hypophyseal cartilage. Where the notochord enters the cranium it is surrounded, as in the preceding stage, by a layer of mesenchyma which extends caudally to join with the chondrifying tissue forming the centrum of the atlas. The centrum stands in the same transverse plane as the arches of the atlas, from which it is separated' by a stratum of less compact mesenchyma. Sagittal sections show a plane of densely packed nuclei separating the centrum of the atlas from the centrum of the epistropheus already laid down in young cartilage. Chondrification of the hypochordal arch of the atlas is less advanced than that of the tissue beneath the notochord between the caudal margins of the parachordal plates; in the latter young cartilage is present, while in the hypochordal arch of the atlas mesenchyma alone is to be found.

Between the caudal margin of the parachordal plates and the lateral mass of the atlas is a transverse stretch of deeply staining tissue, dense laterally and thin medially where it meets the mesenchyma about the notochord. Its later history is not clear.

Two marked advances in development of the floor of the cranium appear in a van Wijhe preparation of 15 mm. (fig. 7). First, the basal plate is nearly perfected, the unchondrified part in the occipital region being now considerably reduced. The bar completing the boundaries of the hypoglossal foramen stands opposite the primary commissure of the parachordals and is thickened at its origin, just where the jugular tubercle and paracondyloid processes later are developed. At the ^ame spot a slight projection forward toward the cochlear capsules, medial to the vena jugularis, is probably the beginning of the basi


vestibular commissure. Second, the anterior expanded extremity of the basal plate is united with the hypophyseal cartilage forming the crista transversa. The carotid artery as it enters the cranium is pressed against the side of the expanded end.

The dorsal extremity of the lateral occipital arch is not. united anteriorly witfi the pars canalicularis of the otic capsule; it is, however, joined with the parietal plate which now presents a prominent dorso-medial angle, the beginning of the tectum posterius.

In the atlas the hypochordal arch is partly chondrified. It is composed of two bars, right and left, which are continuous at each side with the lateral masses and approach each other medially. The lateral mass has been formed from the expanded end of the hypochordal arch of the preceding stage. The centrum of the atlas has united with that of the epistropheus so that a dens epistrophei can be described. The anterior extremity of the latter projects cephalad beyond the transverse plane of the atlantal hypochordal arch. The lateral arches of the epistropheus are connected with the centrum by a lightly stained zone.

Sagittal sections of the embryos of 15 mm. show that the basal plate is chondrified and is continuous with the floor of the cranium in the hypophyseal region (fig. 18). Here, the anterior end of the plate is still recognizable as an elevation, the crista transversa, behind the hypophysis. The bent extremity of the notochord lies upon the transverse crest imbedded in a hill of mesenchyma. In the caudal region of the basal plate the notochord still occupies a position near the intracranial surface, covered by connective tissue continuous with the dens epistrophei. A superficial constriction marks the original plane of fusion of the centra of the first and second vertebrae. Chondrification of the dens is less advanced toward its extremity than at its base. Since the last stage described chondrification has extended forward in the mesenchyma about the notochord, in consequence of which the extremity of the dens has been lengthened and now lies upon the caudal margin of the basal plate.



The hjnpochordal arch of the atlas is composed of young cartilage, least developed toward the midplane and continuous, right and left, with the lateral masses.

In embryos of 17 to 20 mm. (fig. 8) the basal plate is chondrified from side to side. Boundaries of the jugular foramina are established by the presence of the basicochlear commissures and the more complete union of the lateral occipital arches and the otic capsules. In relation to the hypoglossal foramen, the jugular tubercle appears anterior, the paracondyloid process lateral; the one within the cranium, the other outside. The basal foramina are all relatively large. A long, wide basi-cochlear fissure is limited anteriorly by the basi-cochlear commissure which separates it from the carotid foramen. The tectum posterius is represented in cartilage by a spur springing on either side from the parietal plate. The hypochordal ai-ch of the atlas is complete from side to side. The dens epistrophei is further chondrified cranially.

Otic region. Embryo of 23.1 mm.

The otic capsule, externally, presents two well defined subdivisions, one serving as the framework of the semicircular canals, the other enclosing the cochlear canal (figs. 1, 2, 3). The vestibule, on account of its intimate relation both with the pars canalicularis and the pars cochlearis, will be included in the description of these subdivisions.

Regarding the position of the otic capsule at the stage represented by the model, a departure from the manmialian type is to be noted. The long axis of the whole auditory capsule of mammals is, as Gaupp has indicated, du-ected from behind and laterally, forward and medially; just the opposite of what occurs in lower vertebrates. The long axis of the otic capsule of the cat embryo of 23.1 mm. is inchned somewhat transversely (figs. 1, 2). It is true that the broader canahcular part reaches caudally a httle way beyond the cochlear division, but in the cephalic direction both attain almost to the same level.

Regarding the relation of the otic capsule to the walls of the cranium, the pai's cochlearis lies wholly in the floor and the pars


canalicularis stands nearly upright in the lateral wall. The external surface of this part rises straight toward the lamina parietalis; the internal surface incUnes from above, ventrad and a little medialward. The cochlear capsule, very large and rounded, lies beneath the brain in the base of the skull, approaching closely its fellow opposite, the two making up almost the entire cranial floor in the otic region.

At its circumference the otic capsule is connected with neighboring parts of the side wall and base of the cranium by broader or narrower stretches of cartilage with intervening vacuities. The several connections between the otic capsule and the occipital wall and basal plate as well as the related openings have already been described (p. 289)'. There remain to be noted the bonds between the capsule and parietal plate, and certain parts in the orbito-temporal region.

The dorsal, narrow margin of the pars canalicularis is, in large part, free, forming the ventral boundary of the foramen jugulare spurium (figs. 3 and 4). Cephalad and caudad of this narrow fissure it is connected with the lamina parietalis; in the former situation, by the commissura parieto-capsularis, in the latter place with that part of the parietal plate which is continuous ventrally with the occipital wall. Between the cephalic pole of the pars cochlearis and the ala temporalis is the commissura alicochlearis which hmits the carotid foramen laterally. Between this opening and the fissura basicochlearis posterior the cochlear capsule is united continuously with the basa. plate (p. 288).

The pars canalicularis. Outer form

. Oval and laterally compressed, the massive pars canahcularis stands in the side of the skull, contributing a large part of its lateral wall. It presents an extracranial lateral surface, anterior and posterior surfaces mainly within the cranium, and a narrow arched dorsal margin. A vestibular region unites this subdivision of the otic capsule with the pars cochlearis.

The lateral surface (figs. 3, 4) convex and smooth, shows only slight elevations corresponding to the semicircular canals.


Rounded prominences at the cephalic and caudal poles, which appear on both the medial and lateral aspects of the otic capsule, are identified as the prominentia utriculo-ampullaris superior and the prominentia utriculo-ampullaris inferior, respectively (fig. 10). The lateral surface of the pars canalicularis terminates ventrally in a broad irregular ridge, the crista parotica, which, projecting basally, extends from the superior utricular prominence the full length of this surface. In its caudal half the crista parotica (figs. 2, 20) gives rise to an oval elevation connected medially with the base of Reichert's cartilage. The elevation bears a close relation to the future processus mastoideus, which in the cat belongs exclusively to the otic region; but that it gives rise to it I am unable 'to say. The parotic crest, especially in its caudal two thirds, projects from that part of the ear capsule lodging the lateral semicircular canal. In its cephalic part is a slight depression of its lateral surface, the fossa incudis (fig. 3) where the cartilaginous incus is connected by a Ugament. A tegmen tympani at the stage represented by the model is not present. Further relations of the crista parotica with Reichert's cartilage, and the facial canal will be discussed under these subjects. A deep groove separates the crista parotica from the pars cochlearis. This can be traced anteriorly beyond the limits of the crest upon the inferior cochlear prominence toward the foramen faciale; posteriorly as far as the jugular foramen. The facial foramen (fig. 1) transmitting the N. facialis, is directed antero-laterally and lies without the plane of the membrana limitans (vide infra). Here is formed the geniculate ganglion and the origin of the great superficial petrosal nerve. In its anterior part the groove lodges the facial nerve and is identified as the sulcus faciaUs. At the base of Reichert's cartilage it becomes narrow through the approximation of this cartilage toward the crista parotica and the wall of the cochlea, here forming the promontory (fig. 12). Since, at this spot the nerve makes its exit from the groove, it is to be regarded as the foramen stylomastoideum primiti\Tim. Caudad of this foramen, the gToove is broad and deep, filled with connective tissue and occupied to a small extent by the hinder part of the


cavnm tympani. In it lies also the vena capitis lateralis. In the model the expanded posterior part of the groove communicates freely with the great foramen cochleae, opening at its medial side.

The intracranial surface of the pars canalicularis is depressed below the general surface of the cranium and presents an uneven contour (fig. 1). Ventrally it passes into the pars cochlearis. At the place of transition a great quadrilateral fenestra acustica transmits the divisions of the N. acusticus. A ridge, eminentia cruris communis, produced by the crus commune of the anterior and posterior semicircular canals, subdivides the intracranial surface into anterior and posterior faces. The eminence presents the foramen endolymphaticum and inferiorly passes into the prominentia utricularis. The latter, corresponding to the utricle and saccule, reaches below to the acustic window. The posterior surface of the pars canalicularis looks caudad, mesad and slightly ventrad, contrasting in its position therefore with the caudally and ventrally directed corresponding .surface in mammals generally. It passes caudally into the medial surface of the lateral occipital arch and ventrally into the prominentia utriculo-ampullaris inferior. The posterior surface is hollowed to form the sigmoid sulcus, occupied by the sinus transversus. The anterior surface of the pars canalicularis presents the fossa subarcuata anterior limited anteriorly and dorsally by the prominentia semicircularis anterior. This prominence begins at the dorsal pole of the pars canalicularis and curves anteriorly and ventrally to the prominentia ampuUaris superior. On this surface, also, the stout commissura suprafacialis^ is found, connecting the medial wall of the prominentia utriculo-ampullaris superior with the anterior part of the roof of the pars cochlearis, and participating in the ventro-caudal circumference of the spheno-parietal fenestra. It is continued dorsally into a

■ In van Wijhe preparations of embryos 17 to 30 mm. long, the suprafacial commissure appears deeply stained and separated from the pars canalicularis by a zone of lightly stained cartilage. It continues directly into the commissura orbito-parietalis where this is united with the otic capsule (commissura parieto-capsularis) (fig. 9).


prominent ridge which joins the orbito-parietal commissure (figs. 1, 3). The suprafacial commissure constitutes the roof of the short first portion of the facial canal, the floor of which is made by a cartilaginous septum between it and the cavum cochleae. The canal leads from the meatus acusticus internus to the foramen faciale. The dorsal circumference of the pars canaUcularis is partly free, partly continuous with the cranial wall. Posteriorly it passes into the occipital wall; anteriorly it is connected with the commissura parieto-capsularis. Between these connections the dorsal margin is free, forming the ventral boundary of the foramen jugulare spurium. Upon the medial side of the dorsal circumference is a groove occupied by the transverse sinus and its main anterior tributary.

Pars cochlearis

The most conspicuous object of the cranial floor as presented by the model is the huge, rounded bulging of the pars cochlearis of the auditory capsule (figs. 1, 2). The strong walls of this shell enclose the cochlear canal and the saccule. The form of the cochlear capsule is not quite round, but longer in a direction from behind, forward and medialward. Upon the ventral, extracranial surface a shallow groove, sulcus septalis, extends from behind and laterally, medially and anteriorly. Upon the caudal aspect is the large foramen cochleae. Where the cochlear and canalicular parts come together are three openings: the fenestra vestibuli, appearing outside the cranium, the fenestra acustica, within the cranium and the foramen perilymphaticum, appearing partly inside and partly outside the skull. The pars cochlearis shows no sharp boundary toward the pars canaUcularis ; as stated above, the two are continuous in the region of the vestibule. On the surface, this common vestibular zone corresponds to the intermediate stretch which includes the sulcus facialis laterally, and the region of the fenestra acustica, medially. The connections between the cochlear capsule and the floor of the cranium have already been described (p. 288).

Foramina acustica (figs. 10, 11). A large quadrilateral opening, fenestra acustica, the future porus acusticus internus,


is situated upon the medial aspect of the ear capsule, partly in the wall of the cochlea, partly in that of the vestibule. Its plane faces dorsad and mesad, the lateral margin made by the prominentia utricularis being considerably higher than the medial cochlear edge. The fenestra leads into a short, imperfectly formed meatus acusticus internus ; within is an incomplete floor presenting two openings, one transmitting the facial nerve^ the other the branches of the acustic nerve. The former, which lies anteriorly, is the beginning of the primary facial canal. It will be recalled that this short canal passes beneath the suprafacial commissure and ends in the facial foramen outside th^i chondrocranium. The opening for the acustic nerve is an irregular gap ; for convenience of description it will be referred to as the acustic fissure. It is shaped Uke a V with apex caudad, one limb lying almost wholly in the cochlear, the other in the vestibular region. The latter division, widened in the form of an oval notch, lies anterior and slightly dorsal to the cochlear division. It transmits the nerves to the utricle, and to the ampullae of the anterior and lateral semicircular canals. Its position and relations to the structures just named identify it with the superior acustic foramen of other mammahan embryos. The larger cochlear portion of the acustic fissure hes caudad, medialward and somewhat ventrad to the superior acustic foramen. It gives passage to the nerve of the cochlea and the nerve to the saccule, and is identified with the inferior acustic foramen. In the dorsal caudal corner of the meatus acusticus internus is a notch opening into the cochlear part of the fissure and leading to a groove, lodging the nerve to the ampulla posterior. The notch is the forerunner of the foramen singulare and will be designated the incisura singularis.

The position and relations of the facial and acustic foramina having been noted, it now remains to describe the walls of the internal acustic meatus. Figures 13 and 14 show the meatus from the medial aspect. Of the stretches of cartilage intervening between the facial and acustic foramina two are of special interest: one, constituting the primary floor of the meatus, separating the two named foramina; the other forming the incom


plete lateral wall of the meatus. The primary floor of the meatus appears in the model as an extension from the roof of the anterior part of the cochlear cavity. It can be followed laterally from the dorsal part of the cochlear capsule at the medial margin of the fenestra acustica to pass into the first turn of the septum spirale. This part is the forerunner of the longitudinal bony ridge in the floor of the meatus of the adult, the crista transversa of human anatomy, which separates the superior cribriform area and entrance to the facial canal above from the middle cribriform area and the spiral tract of foramina below (Jayne '98, p. 200). Anteriorly it continues as the floor of the prunary facial canal and at the same time the roof of the fore part of the cochlear cavity. Lateral to the facial foramen the primary fccr of ihe meatus passes into the floor of the prominentia utJ'icalo-r.-T.pullaiis superior and forms the ventral boundary of the superior acustic foramen. The posterior free edge of the floor constitutes the cephalic margin of the inferior acustic foramen. The lateral wall of the meatus is formed mostly of precartilage. Its free ventral margin bounds the acustic fissure dorsally. The wall itself covers the medial aspect of the saccule. The medial edge of the meatus, derived from the roof of the cdchlear capsule, stretches from the suprafacial commissure to the posterior wall and is very low. The anterior and posterior walls, on the contrary, tend to become very high toward the vestibular part of the capsule, and are formed respectively by the suprafacial commissure and an elevation made at the junction of the basivestibular commissure with the prominentia ampullaris inferior and cochlear capsule.

Within the meatus acusticus internus are found, anteriorly, the nerve of the vestibule and its ganglion, together with the facial nerve, and, posteriorly, the nerve of the cochlea.

In the model of the chondrocranium, there is a small flattened surface of the cochlear capsule, lateral to the suprafacial commis.sure (figs. 3, 20). This, the planum supracochleare of Voit, supports the seventh nerve as it passes out of the foramen faciale, the geniculate ganglion and the beginning of the great superficial petrosal nerve. Upon it rest also the caudal part


of the semilunar ganglion and the roots of the fifth nerve. As shown by the sections, these structures are embedded in mesenchyma which extends laterally to Meckel's cartilage and medially as far as the plane of the chondrocranial wall; here, in the form of a membrane, it becomes fixed to the margins of the fenestra sphenoparietalis. For this region the name cavum supracochleare has been proposed by Voit. The cavum supracochleare and the several structures therein are excluded from the cavity of the chondrocranium by the membrane of the sphenoparietal window.

Cavities of the otic capsule

A frontal section through the vestibular region shows a large irregular space containing parts of the membranous labyrinth (fig. 10). This space extends anteriorly into the pars cochlearis as the ca\'nm cochleae, lodging the cochlear duct; posteriorly into the pars canalicularis in the form of the canales semicirculares. The main room opened by the section, vestibulum, contains the utricle and saccule.

Cavum cochleae (figs. 19, 20, 21, 22). The great oval cavity of the pars cochlearis, traversed by a low septum spirale, has already begun to asssume the form of a winding canal. This may be followed from the vestibule, wdth which it is in wide communication, ventrally, then in a spiral course into the anterior part of the cochlear capsule. At its commencement the canal is quite wide, and its limiting cartilaginous walls incomplete through the presence of large openings; caudally, the periljaiiphatic foramen; laterally, the fenestra vestibuU; dorsally, the inferior acustic foramen. The cochlear duct makes a little more than one complete turn and occupies but a small part of the cavity which is elsew^here filled by young connective tissue (fig. 9). This tissue extends into the perilymphatic foramen and to the edge of the spiral septum where, by condensation of its elements, it presents the form of a spiral membrane (fig. 22). No indication of cartilage formation was observed in this membrane in the present or in a later stage. The cartilaginous septum spirale takes origin in the primary floor of, the internal


acustic meatus (figs. 10, 11, 19), makes a steep descent toward the floor of the cochlear cavity, and continues as a low ridge to the extent of somewhat more than half a turn from the beginning of the septum. Its position externally, is indicated on the ventral surface of the cochlear capsule by the sulcus septahs (p. 302).

The oval fenestra vestibuli (figs. 3, 20, 21) appears externally at the bottom of a deep depression of the surface of the capsular wall, standing in a coronal plane, above and lateral to the perilymphatic foramen. Its superior and medial boundaries are very prominent; its inferior limit obscure. Within this depression is the foot-plate of the cartilaginous stapes, embedded in mesenchyma which completely fills the window. Within the cavum cochleae, the fenestra appears on an eminence of the lateral wall which projects posteriorly toward the vestibule.

The large foramen perilymphaticum (figs. 2, 12, 20) extends upon both the caudal and dorsal surfaces of the pars cochlearis and is incompletely divided by a process of the cochlear capsule into two parts, the fenestra cochleae and the aquaeductus cochleae. The former is located at the caudal pole of the cochlear prominence, stands in a transverse plane and is occupied by mesenchyma. In figure 11 it appeiars completely separated from the aquaeductus cochleae, but the septum between these openings is chondrified only in its medial part; laterally it is composed of precartilage. The aquaeductus cochleae, lying in a frontal plane, opens into the fossa occipito-canahcularis above (p. 291), and communicates with the cavum cochleae below. It is filled with mesenchyma, through which small veins pass to the vena jugularis interria. The aqueduct of the cochlea at the present stage is an extension of the jugular foramen forward in the form of a deep notch between the commissura basivestibularis and the prominentia utriculo-ampullaris inferior. It constitutes a vacuity in the roof of the cochlear capsule. The aquaeductus cochleae is separated from the jugular foramen proper (i.e. the posterior part transmitting the vein and nerves) by the free angular projection of the cochlear wall above mentioned (indicated by * in figs. 2, 12, 19).


Vestibulum. A frontal section through the middle of the vestibular region reveals a space within the cartilage stretching antero-posteriorly and lying nearer the medial than the lateral margin of the section (fig. 10). This space, a part of the future vestibule, is narrowed in its middle by the encroaichment of a massive projection of the lateral wall and is expanded rather widely in front and behind this. The anterior wide room, containing in the section the utricle and recessus utricuU and the a,mDullae of the anterior and lateral semicircular ducts, will be described under the name of the cavum vestibulare anterius. The posterior room includes, in the section, the posterior part of the utricle, the ventral part of the ductus endolymphaticus, the beginning of the canalis utriculi-saccularis, the sinus inferior, ampulla posterior and the crus simplex of the ductus semicircularis lateraUs. This room will be further considered under the name of the cavum vestibulare posterior. The parts of the membranous labyrinth mentioned do not fill completely the space limited by the cartilaginous walls of the pars vestibularis, but leave an interval filled by mesenchyma representing the future perilymphatic space.

Cavum vestibulare anterius (figs. 10, 11). This large, irregularly shaped cavity occupies the dorsal part of the vestibular region. Its contents have been noted. Its anterior wall swells out as the prominentia utriculo-ampullaris superior. This corresponds to an anterior extension of the cavity called recessus ampullaris anterior containing the ampulla anterior. The floor is broad, slopes ventrad in a medial direction, then drops abruptly to continue into the lateral cochlear wall. At this spot is located the fenestra vestibuh (figs. 19, 20). The medial wall bulges slightly into the cranial ca\dty forming the prominentia utricularis, in conformity with the subjacent surface of the utricle and saccule. The ventral free edge of this wall forms the dorsal boundary of the superior acustic foramen as already described (p. 304). The lateral wall presents a dorso- ventral ridge which projects into the cavity, fitting into the angle between the ampulla laterahs and ampulla anterior. The considerable space behind the ridge, containing the greater part of the lateral am


pulla opens posteriorly into the canalis semicircularis lateralis. The posterior wall is formed by the mass of cartilage already referred to which projects medialward from the lateral wall of the otic capsule (fig. 11). This crista intervestibularis, as it may be termed, intervenes between the anterior and posterior vestibular ' caves, but fails to meet the medial vestibular wall ; there is left a communicating space between the two divisions occupied by a part of the utricle. Ventrally, the crista inter vestibularis enters into the roof of the lateral semicircular canal ; dorsally it becomes continuous with the massa angularis, the great body of cartilage encompassed by the semicircular canals. The dorsal wall of the cavum vestibulare anterius is made by that portion of the angular mass which fills the concavity of the canalis semicircularis anterior.

Cavum vestibulare posterius. In form, the posterior division of the vestibule is oval dorsally, narrow and tunnel-shaped ventrally, expanding into a roomy space in the region of the fenestra cochleae. The posterior wall bulges in adaptation to the adjacent recessus ampullaris posterior, forming on the surface of the ear capsule the prominentia utriculo-ampullaris inferior. The ventral part of the posterior wall presents the posterior orifice of the lateral semicircular canal. The dorsal part, at its junction with the roof of the cavum vestibuH posterius, shows the orifice of the posterior semicircular canal. The medial wall, continuous with the medial wall of the anterior cave, corresponds to that part of the otic capsule marked on the surface by the beginning of the prominentia cruris communis and the incisura singularis. It presents a ventral free edge looking into the inferior acustic foramen. The dorsal wall consists of that part of the massa angularis intervening between the canalis semicircularis posterior and the vestibule. In the highest part of the roof are two openings: — the great circular orifice of the cavum cruris communis, containing the crus commune of the anterior and posterior canals; the small irregular hole, foramen endolymphaticum, looking medialwards and transmitting the ductus endolymphaticus. A floor is present in the caudal half only, beneath the ampullary end of the posterior semicircular duct;


where the floor is wanting anteriorly there is a wide communication with the ca\iim cochleae.

The semicircular canals (figs. 2, 9, 10, 11, 19) are cyUndrieal tunnels of a diameter from two to two and one-half times that of the contained membranous ducts. The future periljanphatic space between the ducts and walls of the canals, is filled with mesenchyma. Within the compass of the canals is the body of cartilage to which Gaupp has given the name massa angularis. The lateral semicircular canal lies in a frontal plane; its floor, lateral and medial walls go into the cartilage of the crista parotica ; its roof is formed by the crista inter vestibularis. The plane of the anterior semicircular canal is dorso-ventrad inclining lateromesad from before backward. Within its concavity is that part of the massa angularis which forms the roof of the cavum vestibulare anterius and which corresponds to the fossa subarcuata of the intracranial vsurface of the pars canalicularis. The posterior semicircular canal also runs in a dorso-ventral plane, which, however, is inclined from behind and laterally, forward and medially. It includes within its concavity that part of the massa angularis forming the roof of the vestibulare posterius and which corresponds on the intracranial surface of the pars canahcularis to a From the bottom of this fossa, which is situated behind the prominentia cruris communis, a canal, transmitting some blood vessels, leads to a great vacuity within the angular mass.

The crista intervestibularis (fig. 11) is a projection of the base of the massa angularis. It begins just dorsad of the fenestra vestibuli as a ridge lying between the ampulla lateralis and the saccule ; it reaches its maximum height opposite the ductus utriculo-saccularis.

The massa angularis (figs. 2, 10, 11), filling the space between the three semicircular canals and the vestibule, is composed of hyaline cartilage, fully formed. Its relation to the walls of the canals and vestibule have already been described. Ventrally the mass becomes continuous with the crista intervestibularis. The angular mass is not continuous thi'oughout, being broken in its central portion by a space occupied by connective tissue


and blood vessels (indicated by ** fig. 2) . The latter find a passage to the cranial cavity by the opening in the fossa behind the eminentia cruris communis (vide supra); veins passing out of this opening join the sigmoid sinus. The space extends cephalocaudad and dorsad, approaching the anterior and posterior semicircular canals and the crus commune, and reaching ventrad as far as the level of the roof of the vestibular cavities and the crista inter vestibular is. It does not at any point communicate with the vestibule.

Lamina parietalis (figs. 1, 2, 3, 4). This plate of cartilage stands in a sagittal plane dorsad of the pars canahcularis. The concave extracranial surface gives origin to part of the temporal muscle; the convex intracranial surface is smooth. The parietal plate is broad anteriorly and posteriorly and narrow between where it lies dorsad of the pars canalicularis of the otic capsule. It goes over anteriorly into the commissura orbito-parietahs posteriorly into the lateral occipital wall, and, in connection with these two parts, is fixed to the pars canalicularis. Between the four elements named is the foramen jugulare spurium, narrow and curved around the dorsal circumference of the otic capsule. Some minute veins traversing the foramen connect the transverse sinus with extracranial veins. The foramen of the right side is subdivided by a cartilaginous connection of the parietal plate and pars canalicularis. The hinder of the two resulting foramina is located opposite the summit of the posterior semicircular canal. In van Wijhe preparations of embryos 24 to 30 mm. in length, union of the parietal plate and pars canalicularis was observed dividing the spurius jugular foramen into two parts. This connection was found on one side or the other, rarely on both, occurring, as in the stage modeled, at the summit of the anterior semicircular canal. The parietal plate at this level is very narrow, is notched in its dorsal margin and, compared with the regions anterior and posterior, is less chondrified.


Otic region in smaller embryos

In cat embryos of 17 mm. (H. E. C. ser. 492 and van Wijhe preparations, fig. 8) both pars canalicularis and pars cochlearis are largely chondrified and stand in connection with each other. At their margins they are united with the floor and walls of the cranium by commissures between which openings are found. The basal plate, sella turcica, parietal plate, orbito-parietal commissure and occipital wall, parts adjacent to the auditory capsule, are all formed of cartilage. The cartilaginous structure of the ventral portion of the occipital wall and that of the floor of the cranium are further advanced in development than the cartilage in the remaining parts.

In the pars canaUcularis, cartilage has formed upon its lateral surface, anterior, posterior and dorsal margins, around and between the semicircular ducts; its medial surface is chondrified about the circumference and crus commune. The prominentiae utriculo-ampullaris superior and inferior present cartilaginous walls, that of the last named being continuous ventrally with the cochlear wall. A chondrified medial wall of the vestibule forms the dorsal boundary of the future internal acustic meatus. At the margins of the pars canaUcularis the sections show that planes of prochondral tissue separate it from the parieto-capsular commitesure in front and lateral occipital arch posteriorly.

As shown by the sections through the cochlear capsule, precartilage is everywhere present excepting at the anterior and posterior poles. At these two spots the cochlea is united with the cartilage of the cranial base by cominissures, but between the commissures it is separated from the basal cartilage by an extensive basicochlear fissure filled with mesenchyma. The roof of the cochlear capsule is considerably further developed than the floor and is nearer the cartilaginous state toward the pars canaUcularis than in the direction of the median plane. At the posterior pole young cartilage can be followed from the prominentia utriculo-ampullaris inferior medially and caudally into the cochlea. This area lies between the fenestra acustica and the foramen perilymphaticum ; from it the basivestibular commis


sure, just caudad, can be differentiated by its more advanced state of development. At the anterior pole, cartilage formation is found in connection with the suprafacial commissure. Where this bar joins the cochlear capsule two regions of the capsule, distinguished by their different states of development, can be seen in the sections. One region, that next to the epithelial cochlear duct is composed of mesenchyma which passes peripherally into precartilage. The other is a circumscribed area superimposed upon the part just mentioned. Its form is somewhat circular in sagittal sections; it is composed of young cartilage. This area occupies the most cephalic part of the cochlear capsule. It can be followed medially through sections 198-192 (H. E. C. ser. 492) of the anterior portion of the pars cochlearis as a more or less distinct tract; further on it becomes completely fused with the deeper stratum of the wall of the cavum cochleae. In sections 191-186 the anterior pole of the cochlear capsule appears as a simple curved plate of young cartilage. In section 185 this is joined with the alicochlear commissure. The relations of the cochlear capsule to the latter and to the suprafacial commissure will be understood by reference to figure 8. • When traced in the lateral direction, the curved plate in question goes over into the suprafacial commissure, with which it conforms in the degree of cartilaginous differentiation, outward size and shape. A second connection between the anterior pole of the cochlear capsule and basal cartilage is presented at this stage by the cartilaginous basicochlear commissure, or synchondrosis, which forms the posterior boundary of the carotid foramen. The parietal plate is partly separated from the canalicular division by the foramen jugulare spurium. It is continuous through young cartilage or precartilage with the occipital arch and pars canalicularis posteriorly and with the commissura orbito-parietalis anteriorly. The tectum posterius is represented by a short process of the lamina parietalis.

Cartilage is found in the otic capsules of a cat embryo of 15 mm. (fig. 7) about the walls of the semicircular ducts and the vesibule, as indicated by the deep blue stain in van Wijhe preparations; the cochlear capsule is unstained. The basivestibular


commissure, the alicochlear commissure and the basicochlear commissure behind the carotid arterj^ are hkewise unstained; the commissura orbito-parietahs is represented by a faintly stained strip of cartilage independent of both the orbital plate and otic capsule. The parietal plate is a triangular cartilage surmounting the extremity of the occipital arch and joined to it; it is unconnected with, though close to, the pars canalicularis. Its medial and dorsal angle is the beginning of the tectum posterius. Sections (H. E. C. ser. 400) show -the material of the basal plate and floor of the sella turcica to be young cartilage. The basivestibular, ahcochlear and basicochlear commissures are composed of mesenchyma. The suprafacial commissure appears as a bar extending from the anterior pole of the pars canalicularis over the nervus faciahs to the anterior part of the cochlear capsule. It is composed of precartilage, easily distinguished from that of the otic capsule by the more abundant, clear ground substance between the nuclei. The relations of this commissure are of high interest. Followed laterally across the facial nerve it meets the pars canalicularis over the pronainentia utriculoampullaris superior. It then passes without the smallest difference in degree of chondrificiation, without boundary of any sort, into the commissura orbito-parietalis. Chondrification of these commissures is quite uniform, and represents a state of histogenesis different from that of the otic capsule. Between the orbitoparietal commissure and the otic capsule is a plane of mesenchymal tissue. The separation of the suprafacial commissure and otic capsule is not so sharp.

When now sections are followed in series from the spot where the suprafacial commissure joins the pars cochlearis, toward the Hiedian plane, an equally interesting connection becomes manifest. It was shown above that in embryos of 17 mm. the cochlear end of the commissure became continuous with a stretch of precartilage in the anterior part of the cochlear capsule, corresponding in shape and state of development exactly with that of the commissure itself. In the present stage this stretch again appears and can be followed medialward and cephalad almost as far as the point of union of the commissura alicochlearis with the cochlear capsule.



Figure 6 made from a van Wijhe preparation of 12 mm. shows the otic capsule from its medial side. Only those parts appear which are chondrified. The three semicircular ducts are in large part walled by cartilage. The blue stain is most intense upon the lateral aspect, faint on the medial side of the ducts. The parietal plate is a small triangular cartilage, lying dorsad of the interval between the posterior semicircular canal and the lateral occipital arch. It is entirely independent of other cartilages of the cranium.

The first trace of chondrification of the otic capsule was found in embryos of about 10 mm. (fig. 5). In van Wijhe specimens of this stage a narrow streak of blue follows the dorsal and lateral circmnferences of the anterior and posterior semicircular canals.

Nerves in the otic region

A brief description of the nerves and blood vessels which have been referred to in the description of the otic region will serve to correlate the many structures which have just been described. The roots of the trigeminal nerve pass forward over the suprafacial commissure and through the septum transversum (vide infra), to the semilunar ganglion, the caudal half of which rests upon the planum supracochleare. The abducent nerve runs along the medial side of the cochlear capsule, leaving the otic region by passing forward between this structure and the dorsum sellae. The seventh nerve, together with the vestibular and cochlear divisions of the acustic, enter the internal acustic meatus from before backward in the order named. The glossopharyngeal, vagus and accessory traverse the jugular foram_en, behind the otic capsule.

The facial nerve, including the pars intermedia (figs. 2, 4, 20, 21), is close to the ganglion of the vestibular nerve as it lies in the internal acustic meatus. It passes through the primary facial canal and the facial foramen to enter the cavum supracochleare and there, outside of the cavity of the chondrocranimn, presents upon its dorsal side the geniculate ganglion (fig. 20). Connected with the ventral surface of the latter is the great superficial petrosal nerve. Its position is entirely extracranial.


It stretches across the lateral wall of the cochlear capsule, beween the geniculate ganghon and a plexus of nerves about the internal carotid artery. In the plexus it joins the N. pterygoideus Vidiani (fig. 3). Leaving the region of the geniculate ganglion the facial nerve enters upon the second part of its cranial course. It undergoes first its characteristic bend and then passes nearly straight caudad, lying for a short distance upon the prominentia utriculoampullaris superior, then enters the sulcus facialis. Above the nerve is the lateral semicircular canal, separated by a floor of thick cartilage from the sulcus faciahs. Continuing its course, backward through the groove, the nerve crosses dorsad of the incudo-stapedial articulation, separated from the vestibular window and basis stapedis which he ventrad. Passing beneath and laterad of the origin of the stapedius muscle (originates from the roof and lateral wall of the sulcus facialis) to which it suppUes a twig, the facial reaches the base of Reichert's cartilage, where, bending sharply ventrad it leaves the groove. The term foramen stylo-mastoideum primitivum has been given by Broman ('99) to the exit from the sulcus facialis bounded by the upper, proximal division of Reichert's cartilage and the ear capsule. After passing the confines of this ill defined foramen (fig. 19) the facial nerve descends upon the caudal vside of Reichert's cartilage and leaves the otic region.

At the side of Reichert's cartilage the facial nerve gives off its chorda tympani branch (figs. 2, 22). This stout nerve winds about the lateral side of Reichert's cartilage, ascends somewhat and, tm-ning mesad and cephalad, enters the region of the first pharyngeal pouch. Here it lies at first upon the medial surface of the manubrium of the cartilaginous malleus, ventrad of the insertion of the M. tensor tympani and of the incudo-stapedial articulation. Proceeding forward it passes out of the t>anpanic region, following closely the medial and ventral surface of Meckel's cartilage.

Acustic nerve. Two chief branches were present, an anterior and posterior. The anterior ramus gives a twig to the recessus utricuU, one to the ampulla anterior, another to the ampulla externa. The posterior ramus gives rise to a ramulus sacculi, a


twig to the posterior ampulla (the two together may be regarded as a ramus medius), and continues as the ramulus basilaris. In addition to the ramulus sa^cculi, described by Retzius ('84), a small twig to the sacculus coming from the anterior ramus was observed.

The vagus, together with the glossopharyngeus and accessorius, form a bundle which occupies a position in the fossa occipito-canalicularis close to the lateral occipital arch and considerably posterior to the pars canalicularis of the otic capsule (figs. 1, 2, 19). The vagus hes medialward of the vena jugularis with the glossopharyngeus anterior and the accessorius posterior to it. These relations are maintained in passing through the jugular foramen, excepting that the accessorius becomes incorporated wdth the vagus. The jugular ganghon is found upon the roots of the vagus as it lies in the fossa occipito-canalicularis and therefore above the jugular foramen. Some distance outside the cranium the ganglion nodosum appears.

The ganglion superius of the N. glossopharyngeus (figs. 1, 2, 19 to 23) is situated partly in the recessus occipito-canalicularis, and partly upon the roots of the nerve as they lie in the cranial cavity above the recess. The ganglion petrosum, much larger, is located on the nerve below the jugular foramen and above the level of the ganghon nodosum of the vagus. The tympanic nerve arises from the upper end of the ganglion petrosum and proceeds in the direction of the foramen cochleae. This nerve runs in the thick mesenchyma between the medial epithelial wall of the first pharyngeal pouch and the foramen cochleae. After much branching, it reaches the ventrolateral surface of the posterior cochlear prominence where it enters a plexus of nerves, derived in part from the sympathetic around the internal carotid artery.

Blood vessels in the otic region

A great blood sinus runs upon the inner surface of the cranial wall in the sulcus which circumscribes the pars canalicularis. This, a forerunner of the sinus transversus system, becomes continuous at the fossa occipito-canalicularis with the vena jugularis


interna. Small tributaries enter from the connective tissue filling the foramen jugulare spurium; some of these are emissary veins. Behind the pars canaUcularis the sinus is joined by a small vein formed by tributary systems, one being the plexus of veins in the mesenchyma which fills the vacuity of the massa angularis (p. 309). A plexus of veins in the facial sulcus empties into a large vein which, running backward in company with the facial nerve, joins the internal jugular. An extensive network of veins within the mesenchyma of the cavum cochleae is drained in part by two veins which pass to the vena jugularis interna by way of the aquaeductus cochleae.

Vestiges of a stapedial artery are present in the embryo of 23.1 nrni., perforating the stapes (fig. 21). These can be followed anteriorly toward the terixdnal branch of the internal maxillary artery and posteriorly toward a plexus of small vessels on the ventral sm'face of the cochlear capsule, which is connected with an offset from the internal carotid artery. In an embryo of 15 mm. the stapedial artery is a relatively large branch of the A. carotis interna, extending forward through the mesench>^nal beginning of the stapes to the semilunar ganghon whence it continues as the mandibular artery.

The internal acustic artery arises from the basilar and enters the internal acustic foramen together with the acustic and seventh nerves, giving off several branches to the membranous labyrinth.

Orhito-temporal region. Embryo of 23.1 mm.

The basis cranii (figs. 1, 2, 3) of the orbito-temporal region is an unbroken bar of cartilage continuous with the basal plate of the otic region posteriorly, and with the nasal septum of the ethmoidal skeleton anteriorly. The extracranial surface, strongly convex from side to side looks toward the pharynx, from which it is separated by an interval filled by wide-meshed mesenchyma. The intracranial surface presents in its posterior one-third the sella turcica with the root of the ala temporalis on. each side, and in front of this, in successively higher planes, the two roots of the ala orbitahs, the optic foramen intervening. At the level of the


sella turcica the basis cranii is bent in a sagittal plane so that the otic and orbital portions form an angle, open ventrally. Also, where the orbito-temporal portion of the base goes over into the septum nasi, a similar angle is formed.

At the level of the sella turcica the base is flattened dorso-ventrally, becomes thicker opposite the optic foramina and finally triangular in its most anterior part. Here, the two extracranial surfaces inchne medially and ventrally to meet in a keel, forming a short interorbital septum which passes forward into the septum nasi.

Sella turcica (figs. 1, 23). The elongate, shallow pitiutary fossa is limited anteriorly by a low tuberculum sellae and posteriorly by a well defined wall. The base of the latter is made by the crista transversa extending nearly to the cochlear capsules, a slight interval remaining between these parts transmitting the abducent nerve. The dorsum sellae is represented by a conical elevation of the middle of the crista transversa. Remains of a hypophyseal canal are present and are occupied by epithelial vestiges of the stalk of the anterior lobe of the pituitary body.

Ala temporalis (figs. 1, 2, 3, 4, 23). The ala temporahs of the 23.1 mm. embryo extends from the processus alaris of the basis cranii, opposite the sella turcica, outward bfeneath the semilunar ganglion; it is limited by free margins in front and behind. Its structure is of young cartilage separate from that of the processus alaris, a stratum of mesenchyma standing between them. The temporal wing consists of a larger blade-formed lateral portion, the lamina ascendens, and a small, medial pterygoid process. A separate pterygoid cartilage is present, lying mesad of the pterygoid process below the basis cranii. In connection with the description of the ala temporalis the cavum epiptericum will be considered.

Processus alaris (figs. 2, 23). The alar process lies in front of the carotid foramen, supports the internal carotid artery in its forward course and forms the ventral boundary of the innermost part of the spheno-orbital fissure. Laterally it extends to the temporal wing, the layer of mesenchyma above mentioned limiting the two parts (fig. 23) . This layer takes an oblique direc


tion from before backward and laterally, on account of which the ala temporalis stands in front, as well as to the side of the processus alaris. Between the latter and the anterior pole of the cochlear capsule extends a cartilaginous bridge, the commissura aUcochlearis, forming the lateral boundary of the carotid foramen. It is quite distinct from the temporal wing, the two being separated by the continuation of the zone of mesenchyma just described. The processus alaris and commissura aUcochlearis are, however, continuous, and together form an arch around the front and side of the carotid foramen.

The lamina ascendens (figs. 1, 2, 3, 23) of the ala temporahs is a cartilaginous plate which extends laterally and dorsally toward the commissura orbito-parietahs, but remains widely separated from it. It is somewhat triangular in form, presenting free margins, anterior and posterior, which meet laterally in a rounded free point, and a base directed toward the obUque line of junction with the alar process and alicochlear commissure. The dorsal surface slopes from an intermediate longitudinal ridge, forwaj-d and backward, thus presenting two subdivisions: an anterior face entering into the floor of the spheno-orbital fissure ; and a posterior, supporting the cephaUc end of the semilunar ganglion and forming the floor of the epipteric cave. The longitudinal ridge itself lies opposite the lower margin of the ala orbitahs and enters into the ventral boundary of the spheno-orbital fissure. The ventral surface of the lamina ascendens overhangs the mandible and gives origin to part of the pterygoid muscle. A short canal, the forerunner of the foramen rotundum, traverses the lamina ascendens from behind forward, giving passage tothe maxillary nerve. The posterior margin of the ala is free and limits anteriorly a broad, deep fissure whose posterior boundary is the cochlear capsule. The fissure ends medially at the alicochlear conmiissure and, like the spheno-orbital fissure, is open laterally. To this gap, which has long been known, the name fissura aUcochlearis may be given. Through it pass the mandibular and the great superficial petrosal nerves, the former through the incisura ovalis in the posterior margin of the lamina ascendens, the latter traversing another notch, incisura lacera, and


joining beneath the lamina the great^ deep petrosal nerve to form the N. pterygoideus. Incisura lacera is the name proposed for the deepest part of the fissura alicochlearis, a part distinguished by certain important relations. It reaches medially toward the carotid foramen, from which it is separated by the commissura alicochlearis, and here adjoins the posterior end of the mesenchymal zone dividing the commissure from the lamina ascendens. A prominent spine of the caudal margin of the ascending plate forms its antero-lateral boundary and stands between it and the incisura ovalis. On the anterior margin of the ala temporalis, adjacent to the processus alaris, is the broad processus pterygoideus (fig. 1), in connection with which is later formed the insignificant bony lamina lateralis processus pterygoidei of the adult. The pterygoid process projects forwa'd and toward the midplane, lying medial of the foramen rotundum. Its medial end is separated from the adjacent pterygoid cartilage by a layer of mesenchyma. At this spot the Vidian nerve begins to turn from beneath the lamina ascendens to gain, eventually, a position dorsad of the pterygoid cartilage and within the spheno-orbital fissure.

The base of the ala temporalis is directed obliquely from behind, forward and medialward and for the most part corresponds to the synchondrosis between the ascending plate, alar process and alicochlear commissure. Two other regions, however, must be included in the base: posteriorly, the medial free margin of the spine separating the incisura lacera from the incisura ovalis, and anteriorly, the medial free margin of the processus pterygoideus. The relations into which the base of the ala enters with the rest of the chondrocranium are, therefore, the medial corner of the spheno-orbital fissure, processus alaris, commissura ali cochlearis and incisura lacera.

Ala orhitaUs (figs. 1, 2, 3, 4). This great sickle-shaped cartilage forms, on each side, the larger part of the medial wall of the orbit and the lateral wall of the fore part of the cranial cavity. Its lateral and medial surfaces are respectively concave and convex, apparently in adaptation to the eye-ball and its adnexa. The anterior extremity is joined to the nasal capsule by the com


missura spheno-ethmoidalis ; the posterior extremity is connected wdth the fore part of the basis cranii by two roots, named preoptic and metoptic, relative to the optic nerve which they embrace. The preoptic root is broad and straight whereas the metoptic root is narrow and bent so as to present a prominent angle of its caudal margin directed backward. The optic foramen, large and irregulary oval, is bounded medially by the basis cranii, on its other sides by the orbital wing and its two roots. Through the optic foramen a short, blunt process, processus orbitahs, extends into the orbit from the cephalic edge of the metoptic root. On the right side this process Ues close to the basis cranii. From -the roots of the orbital wing some of the ocular muscles take origin (vide infra). The short concave, anterior margin of the wing is free and separated from the nasal capsule by the fissura orbitoethmoidalis. This is filled by membrane and transmits the lateral branch of the nasal nerve. The extensive, convex, posterior margin, dorsally, passes into the commissura orbito-parietahs connecting the orbital wing with the parietal plate, and ventrally is free in the anterior boundary of the fenestra spheno-parietalis.

The spheno-orbital fissure (Wincza, '98) (figs. 1, 2, 3, 4, 23) in the present stage stands in marked contrast to the circular, completely walled opening of the adult cranium. It is a narrow, deep gap running in a frontal plane between the ala orbitalis above and the ala temporaUs and pterygoid cartilage below. It is open laterally, but closed medially by that part of the basis cranii lying between the metoptic root of the ala orbitahs and the processus alaris. It gives passage to the oculomotor, pathetic, ophthalmic and abducent nerves, together with a number of veins which go to join the plexus about the semilunar ganghon and carotid artery.

The carotid foramen (figs. 1, 2) is completely w^alled, as described by Decker and later by Wincza. It is circular, not large as compared with that in Lepus, stands opposite the middle of the sella turcica and traverses the floor of the skull straight ventro-dorsad. In front is the processus alaris; behind, the cochlear capsule; laterally, the aUcochlear commissure. Upon its medial side is the sella turcica.


Pterygoid cartilage (figs. 1, 3, 24). In the model of the cranium of the 23.1 mm. embryo the pterygoid cartilages appear as a pair of irregular but symmetrically formed bodies lying one on either side of the naso-pharyngeal duct and adjacent to the alae temporales. Each cartilage presents a short cyhndrical, transversely placed pedicle and, continuous with it and in a dorso-ventral plane, an elongate, bent plate. The pedicle consists of dense mesenchyma continuous with the perichondrium of the margin of the pterygoid process of the lannna ascendens; in the center bone has begun to form. The N. pterygoideus (Vidiani) turns dorsad upon the posterior aspect of the joint between the pedicle and pterygoid process to enter the spheno-orbital fissure. The plate-shaped part of the pterygoid cartilage presents a flat free surface directed forward, a medial concave margin tm-ned toward the naso-pharyngeal duct and a lateral convex margin giving attachment to the M. pterygoideus. In structure the lamina consists of mesenchymal syncytium, within which bone has formed in two regions. One center lies in the dorsal half of the plate ajid is continuous with the ossific center of the pedicle; the second is found in the ventral half and seems to be separated from the first by a stratum of unossified tissue, broad at the margin and surfaces of the plate, but becoming narrower toward the center. I find it impossible to state whether the two bony centers so clearly separated one from the other at the margins of the lamina are entirely distinct throughout. A mesenchymal condensation surrounds the pterygoid cartilage, extending forward, becoming somewhat less compact, and finally joining the mass of mesenchyma in which the palate bone is forming. This mesenchymal mantle extends dorsally toward the cranial base without, however, quite reaching it.

The region named cavum epiptericum by Gaupp ('05) and interpreted in its relation to the cranial cavity of mammals by this investigator ('02) as an acquisition of a space which, in reptiles, is extracranial, presents some interesting peculiarities in the embryos of cat. First, the epipteric cave with its contained structures is, in the absence of those bones which in the adult form the lateral wall of the skull in this region, wholly outside


the confines of the brain case of the 23.1 mm. embryo. The exclusion of the space from the cavmn cranii is brought about by a strong membrane, conspicuous in the sections, which arises from the base of the cranium and is apphed to the medial surface of the semilunar ganglion, excluding this structure, as it rests on the ala temporalis, from the cranial space occupied by the brain. Less dense mesenchyma surrounds the ganglion and extends laterally as far as the temporal muscle ; forward through the sphenoorbital fissure and about the ala temporalis; backward, into the cavum supracochleare. The medial limiting membrane calls for special description. We may tentatively refer to it as the membrana limitans (figs. 19 to 23). It fills the foramen sphenoparietale attaching itself to its margins, i.e., to the commissura orbito-parietalis, conm^iissura suprafaciahs, where it is perforated by the roots of the trigeminus, and to the metoptic root of the ala orbitalis; then upon the base of the cranium, from the anterior pole of the cochlear capsule along the commissura ahcochlearis, crista transversa and dorsum sellae, the lateral elevated margins of the hypophyseal fossa, and on to the tuberculum ephippii. In the bottom of the sella turcica and upon the dorsum sellae the membrane is continuous from side to side. At the saddle-back it is continuous with a stout meningeal membrane which ascends into the narrow space in the flexure of the mid-brain and diencephalon; this may be refered to as the septum transversum (see p. 327). The thickness of the limiting membrane is not the same throughout. It is densest at its broad attachment to the cranial base where it covers the carotid foramen. Also a definite band of condensed tissue extends through the septum transversmn from the crista transversa, crossing the N. abducens and the roots of the N. trigeminus running parallel with the commissura suprafaciahs and reaching laterally the parietal plate. It is along this stretch that a small cartilaginous bar is later developed (embryo of 7 cm.) over the dorso-medial surface of the semilunar ganghon and roots of the trigeminus where the latter cross the suprafacial conmaissure.

The floor of the epipteric cave is formed by the lamina ascendens of the ala temporahs. The following structures are found


within the cave : the ganglion semilunare and the three divisions of the N. v., the NN. Ill, IV and VI. The caudal half of the semilunar ganglion rests upon the supracochlear plane, in its cephalic half upon the alicochlear commissure and the posterior subdivision of the dorsal surface of the lamina ascendelis. Beneath it, the mandibular nerve passes to leave the cave through the incisura ovalis. The ophthalmic and maxillary nerves leave the region immediately in front of the ganghon, the one by way of the spheno-orbital fissure, the other by the foramen rotundum. Both the oculomotor and trochlear nerves run a long course in the loose tissue immediately surrounding the brain, proceeding from their origins ventrad and cephalad toward the sphenoorbital fissure, where they lie above the ophthalmic nerve. They pierce the membrana Imiitans just before entering the fissure, and their course in the epipteric cave is therefore very short (fig. 23). On the contrary, the course of the abducent nerve through the cavum epiptericum is very long ; it enters the space from behind, passing between the dorsum sellae and cochlear capsule, over the crista transversa and beneath that strand of condensed mesenchyma in the septum transversum which is the forerunner of the cartilage referred to above. Running cephalolaterad, it crosses laterally the carotid artery and gains the medial side of the semilunar ganghon and ophthalmic nerve; in the sphenoorbital fissure it crosses dorsally the N. maxillaris. The internal carotid artery enters the membrana limitans at the carotid foramen, turns forward in this tissue, crosses the processus alaris at the side of the hypophysis, then pierces the membrane in a dorsal and medial direction to enter the cavum cranii. At its entrance into the basal portion of the membrana hmitans the artery is ventrad of the sixth nerve, but it turns immediately mesad beneath it. Several veins accompany the nerves through the spheno-orbital fissure and go to form, by anastomoses, a plexus in the membrana limitans which surrounds the carotid artery. The vessels of this plexus are separated by mesenchymal tissue, altogether constituting the beginnings of the channels and walls of the later cavernous sinus.


The orbit (figs. 3, 4, 24, 25, 26) is relatively shallow and is illdefined in extent and boundaries. The planum antorbitale and prominentia lateraUs of the ethmoidal skeleton form its anterior wall; the ala orbitaUs and its two commissures enter into the posterior wall. The dorsal hmit is given by the frontal bone which extends along the spheno-ethmoidal commissure. The ventral limits are found in the zygomatic and maxillary bones. The ocular muscles arise in the following way. From the preoptic root springs the superior oblique; from the orbital process of the metoptic root arise the rectus superior, internus and externus. The origin of the inferior rectus is conjoined with that of the retractor oculi on the lateral surface of the cranial base, ventrad of the metoptic root and adjacent processus alaris; these muscles arise within the spheno-orbital fissure. The inferior oblique springs from the ventral margin of the planum antorbitale well forward of the origins of the other ocular muscles.

Orhito-tem'poral region in smaller embryos

Van Wijhe preparations of 10 mm. (fig. 5) present the first trace of chondrification in the orbito-temporal region as a small crescentic cartilage ventrad of the anterior hypophyseal lobe. The concavity of the crescent embraces the hypophyseal stalk. Behind this hypophyseal cartilage and occupying a dorsal plane is the termination of the notochord.

Sections of an embryo of 12 mm. (figs. 16, 17) show that the cranial floor beneath the hypophysis consists of a horseshoeshaped cartilage independent of other parts of the chondrocranium. The legs of the horseshoe lie on either side of the midline beneath the anterior pituitary lobe, reaching as far forward as the trabecular plate, but not united with it; the commissure of the horseshoe crosses behind the stalk of Rathke's pouch. Between the commissure and the terminal expansion of the basal plate is a small space, fenestra basicranialis posterior, occupied by mesenchyma in which the cephalic end of the notochord lies. In median sagittal sections of embryos of 12 mm. (fig. 17) the anterior expanded extremity of the basal plate is turned somewhat in the dorsal direction above the line of the posterior parts


of the hypophyseal cartilage. The trabecular plate makes its appearance in embryos of the present stage (fig. 17), consisting of a single thick mass of young cartilage continuous anteriorly with the beginnings of the septum nasi. Caudally it reaches almost to the hypophyseal cartilage. In the van Wijhe preparations the limits of the trabecular plate are clearer than in the sections (fig. 6). On its dorsal surface the sulcus chiasmaticus appears; on its sides, the beginnings of the pre- and metoptic processes.

In a frontal section passing through the hypophysis and cochlear canal of an embryo of 12 mm. (fig. 16) the carotid artery appears in its course through the cranial wall. It lies in an oval area of less dense mesenchyma than that a little wa> behind its position. The denser tissue anterior to the area forms a curved bar, continuous with the side of the hypophyseal cartilage medially, thence extending in a curve outside the artery toward the cochlear capsule. This will be referred to as the commissural element; it is the beginning of the processus alaris and commissura alicochlearis. The oval, hght area about the carotid is limited caudally and medially by the cochlear capsule and a bridge of mesenchyma (future basicochlear commissure) stretching from the latter to the hypophyseal cartilage.

In the mesenchyma laterad of the commissural element is a small condensation beneath the Gasserian ganglion, which proves to be part of the medial extremity of the future ala temporalis. This condensation will be refered to as the alar element. The less dense mesenchyma between the latter and the commissural element extends obliquely from before, backward and outward. The mesenchyma of the alar element becomes broader and denser as it extends laterally; its caudal concave margin, pressed against the mandibular nerve, forms the primitive incisura ovalis ; within its anterior part appears the maxillary nerve. The ophthalmic nerve (whose ganglionic cells are in part separate from the semilunar ganglion) passes over the alar element toward the eye. The van Wijhe specimens of 12 mm. (fig. 6) show a small darkly stained nodule far removed from the hypophyseal cartilages and lying immediately ventrad of the anterior end of the semilunar


ganglion. This center is located at the site of the anterior part of the alar element.

In sections the ala orbitaUs is a plate of condensed mesenchyma Umited sharply in front, above and behind, but connected with the trabecular plate by two roots. The preoptic root extends from the trabecular plate anterior to the sulcus chiasmaticus and passes laterally without interruption into the main body of the ala. The metoptic root is joined to the side of the caudal portion of the trabecular plate. This root terminates laterally in a pointed extremity connected by less dense mesenchyma with the ala orbitalis. The extremity lies beyond the optic foramen, immediately dorsad of the third nerve where this enters the orbit.

The cavum epiptericum in the 12 mm. embryo, is clearly defined toward the base of the neuro-crailium. The membrana Umitans (figs. 16, 17) can be followed over the medial aspect of the semilunar ganglion to the anterior end of the basal plate, to the conunissural element and over the carotid artery. Upon the lateral side of the membrane, opposite the vessel runs the abducent nerve. At its insertion into the anterior margin of the basal plate (the future crista transversa) the limiting membrane becomes continuous with the septum transversum. Within the latter immediately caudad of the hypophysis is a small, transversely placed precartilaginous rod (indicated by an * in fig. 17).

The pterygoid cartilage is represented by a condensation of mesenchyma lying anterior and ventral of the alar element and close to the epithelium of the pharynx. A constriction incompletely subdivides it into dorsal and ventral parts. Behind this mass is the first pharyngeal pouch; through its dorsal extremity runs the pterygoid nerve.

- In embryos of 15 nun. the hypophyseal cartilage has united anteriorly with the trabecular plate, forming the floor of the sella turcica, perforated by a foramen hypophyseos (figs. 7, 18). Union of the basal plate and hj^pophyseal cartilage has also occurred, but the confines of the originally separate cartilages are still evident (fig. 18). As a result of the incongruity in the meeting of these parts (indicated in the 12 mm. stage) there remains a transverse ridge in the basis cranii behind the pituitary fossa,


the beginning of the crista transversa. This is the stage of precartilaginous structure of the alar and commissural elements. The latter is continuous with the side of the sella turcica and separated from the alar element by an oblique layer of mesenchyma. The boundaries of the carotid foramen are well defined, the medial boundary being formed by the posterior part of the side of the sella turcica ; the caudal limit by the prochondral union of the basal plate and cochlear capsule, the commissura basicochlearis; the anterior and lateral by the prochondral commissural element. The alar element has a homogeneous structure of young cartilage and is now recognizable as the ala temporalis; both the mandibular and maxillary nerves pass through notches in its margins.

In van Wijhe preparations of 15 mm. (fig. 7) a continuous stretch of cartilage occurs in the floor of the cranium in the otic, hypophyseal, trabecular and nasal septal regions. The tuberculum sellae is apparently developed from the caudal and dorsal edge of the trabecular plate. The junction of the basal plate and floor, in the hypophyseal region, is marked by a broad expansion of lightly stained tissue (young cartilage). In a median section (fig. 18) the cartilage in the septum transversum behind the hypophysis, present in the preceding stage, appears as a small nodule. The ala temporalis is represented by a slender sickle-shaped cartilage standing ventrad of the anterior pole of the semilunar ganglion. It is laterally removed from the sella turcica and from the commissural element (the interval is exaggerated in figure 7). In the epipteric region the membrana limitans is clearly defined throughout. The origin of the transverse septum is' marked by a band-like thickening, extending from the crista transversa laterad over the semilunar ganglion and roots of the trigeminus, and terminating on the supraf acial commissure and parietal plate.

The pterygoid presents no important change from the preceding stage. Chondrification of the ala orbitahs has begun over the eye-ball in the form of a triangular plate, separate from all neighboring cartilages (fig. 7). Posterior to the orbital ala is the independently chondrifying commissura orbito-parietalis ; far removed anteriorly are the cartilages of the nasal, capsule. One


angle of the ala orbitalis is directed medially toward the trabecular plate and its two processes, preoptic and metoptic. The extremity of the latter presents the relation with the oculomotor nerve observed in the earlier stage. From the metoptic process a thick projection extends a short distance laterally, just behind the optic nerve, to end in the muscle mass about the nerve; this is the beginning of the orbital process of later stages.

Van Wijhe preparations of embryos of 17 mm. (fig. 8) show a lightly stained curved bar, the combined processus alaris and alicochlear commissure, extending around the carotid artery, from the side of the sella turcica to the anterior pole of the cochlear capsule. Just in front and laterad of the commissure, and separated by an interval of unstained tissue, is the triangular ala temporalis, with the maxillary nerve against its anterior, concave margin, and the mandibular nerve at its posterior side. The separate cartilage behind the infundibulum in the preceding stage is no longer apparent in the sections, but there is now at this spot a dorsal, median process, partly cartilaginous, largely mesenchymal, of the crista transversa. Union of the orbital plate with neighboring skeletal parts has occurred (fig. 8) : the medial angle, presenting a notch for the optic nerve, is connected with the pre- and metoptic processes of the trabecular plate; the anterior angle is prolonged to the paranasal cartilage as the commissura spheno-ethmoidalis, thereby completing the boundaries of the fissura orbito-nasalis ; the posterior angle is fused with the commissura orbito-parietalis. This commissure, which has united also with the parietal plate, has extended ventrally to the pars canalicularis (commissura parieto-capsularis), but, as sections show, is separated from the otic capsule by a thin layer of mesenchyma (p. 313). Continuity of the cartilage of the orbito-parietal commissure and suprafacial commissure has been described above (p. 301).

Ethmoidal region. E7nbryo of 23.1 mm. .

The parts of the chondrocranium included in this most clearly defined region of the skull are the paired nasal capsules (figs. 1, 2, 3, 4). These conform closely with the membranous walls of



the cavum nasi within, and are shaped externally somewhat Hke a pair of spindles pressed together, making a single strong framework, the cartilaginous ethmoidal skeleton, at the anterior end of the skull.

The length of the cartilaginous ethmoidal skeleton is equal to about one-third that of the entire cranium. Its anterior half projects free beyond the brain case; its posterior half, terminating on each side in the posterior cupola, is subcerebral in position. Here it is continuous with parts in the orbito-temporal region: the trabecular plate passing into the septum nasi and roof of the posterior cupola, the commissura spheno-ethmoidalis joining the lateral nasal wall. An extensive roof, tectum nasi, is in the anterior half of the ethmoidal skeleton; an incomplete one, lamina infracribrosa, in the posterior. The roof goes over on either side into the paries nasi. The interior room of the nasal skeleton opens into the cranial cavity by the fenestrae olfactoriae, and upon the face through the fenestrae narinae. The floor of the capsule, solum nasi, is very incomplete, presenting a great fenestra basalis, within the compass of which are the paraseptal cartilages standing next to the nasal septum. The side walls within, present the beginnings of the complex of turbinals in the form of simple inrollings and elevations.

The tectum nasi (figs. 1, 3, 27 to 30) is furrowed in the midline of its dorsal surface by the sulcus supraseptalis, deepest anteriorly, extending from the fenestra olfactoria to the level of the fenestra narina. On either side the roof is highly arched and passes laterally into the anterior region of the paries nasi.

The side wall (figs. 3, 4) presents three regions: an anterior, continuous with the tectum nasi, a middle region occupied by the rounded prominentia lateralis, and a posterior area, the planum antorbitale. Two grooves demarcate the prominentia lateralis from the anterior and posterior regions. One of them, the sulcus laterahs anterior, at first pursues a curved course from the olfactory window near the mid-line, cephalo-ventrad between the tectum nasi and prominentia lateralis. This part of the sulcus stands opposite the crista semicircularis within the nasal capsule and is interrupted by the epiphanial foramina of which


there are ^wo on each side. In its ventral one-third the groove is broad and shallow and adjoins the anterior region of the paries nasi and the lateral prominence. It occupies an area which passes ventrally into the maxilloturbinal and stands at a level anterior to the crista semicircularis. This part of the paries navsi corresponds to the lamina supraconchalis of Voit. The second groove, sulcus lateralis posterior, limits the lateral prominence posteriorly; it is broad and shallow and stands opposite ethmoturbinal I within the capsule. The anterior region of the paries nasi extends from the antero-lateral sulcus to the lateral margin of the fenestra narina. It is co-extensive with the roof dorsally, but becomes narrow ventrally where it joins the lamina transversalis anterior. The surface in this region is convex, corresponding to the internal concavity of this part of the lateral wall. The middle region, the prominentia lateralis, is narrow from before backward. Dorsally it is united with the commissura spheno-ethmoidahs and forms the lateral boundary of the olfactory window, between the posterior extremities of the crista semicircularis and ethmoturbinal I (vide infra). Ventrally it passes into the maxillo-turbinal. The posterior area, antorbital plane, corresponds to the region within occupied by the ethmoturbinals, and forms the antero-medial wall of the orbit. Where it adjoins the middle region of the paries nasi (sulcus lateralis posterior) there is present ventrally a rounded prominence, giving origin to the inferior oblique muscle of the eye. The dorsal margin of the antorbital plane at the orbito-nasal fissure, participates, anteriorly in the lateral boundary of the olfactory window, and posteriorly passes into the lamina infracribrosa. The ventral margin is straight and free in the lateral boundary of the fenestra basalis. Caudally, this division of the nasal capsular wall continues into the posterior cupola. This conical shell extends backward at the side of the nasal septum and trabecular plate beneath the preoptic root of the ala orbitalis. It is closed ventrally by the lamina transversalis posterior, which reaches to the septum nasi. The roof of the posterior cupola is formed by the unpaired lamina infracribrosa (figs. 1, 25). This plate presents a free edge anteriorly toward the fenestra olfactoria, con


tinues caudally into the dorsal surface of the trabeqjilar plate and laterally goes into the planum antorbitale. A slight median ridge, crista interorbitaUs, separates two shallow fossae lodging the olfactory tracts.

Ventrally, the nasal capsule presents the great fenestra basalis (figs. 1,2), the relations of which to the cavum nasi are somewhat complicated. It is limited laterally by the ventral margin of the planum antorbitale and by the maxillo-turbinal. In the region of the latter it is very narrow from side to side, and is partly covered ventrally by the paraseptal cartilage hanging down from the septum. Anteriorly the basal window reaches to the lamina transversalis anterior which stands between it and the fenestra narina. The caudal limit is formed by the lamina transversalis posterior. Further relations of the basal window will be better understood when considered in connection with the description of the nasal cavity.

The solum nasi is represented by the two transverse laminae and the horizontal portion of the paraseptal cartilage (figs. 2, 3, 13, 25, 26). The lamina transversalis posterior (figs. 2, 13, 25) forms the floor of the posterior cupola and appears in the sections and in the model as an inrolling of the caudo-ventral part of the planum antorbitale. Its medial margin is turned dorsad; in its caudal half it is applied closely to, but not actually joined with, the keeled trabecular plate behind the level of the vomer; in its cephalic half it is free, being separated from the nasal septum by a narrow stretch of the fenestra basalis. Here the lamina is overlapped by the caudal end of the vomer. The bandlike lamina transversalis anterior (figs. 2. 3, 13), incompletely chondrified, bounds the fenestra narina ventro-laterally, connecting the septum nasi with the paries nasi. It is continuous with both parts, that is to say, a zona annularis is present. Its horizontal, broader, septal end stands opposite the anterior half of the OS incisivum. Followed laterally, the lamina undergoes a curve in the dorsal direction, and comes to lie in a sagittal plane. ^

' In cat embryos of about 70 mm. the lamina transversalis anterior presents two backwardly directed processes which embrace the funnel-shaped depression of the floor leading into the incisive duct. These have been called by Zucker


Just ventrad of its junction with the anterior part of the paries nasi it is crossed by the naso-lacrimal duct. The caudal margin of the lamina is free and is separated by a wide interval from the paraseptal cartilage.

Behind the anterior transverse lamina, the tear duct lies upon the most ventral part of the paries nasi and is further supported by a cylindrical processus maxillaris anterior of the latter (figs. 3, 13, 30) directed forward. At this place the os maxillare overlaps the nasal capsule and, in the model, hides from view the maxillary process and tear duct.

The cartilago paraseptahs (figs. 2, 3, 13, 28, 29), scroll-formed in adaptation to Jacobson's organ, hangs from the septum nasi below the level of the plane of the fenestra basaUs. It is unconnected with other parts of the nasal skeleton. The medial half of each cartilage hes in a sagittal plane applied against the ventral part of the septum. The lateral half is rolled dorsally to form a gutter lodging the vomero-nasal organ. The cephalic margin is free and straight; the caudal pointed. The long axis of Jacobson's cartilage extends from behind cephalo-ventrad, the caudal extremity lying at the level of the basal window, the cephahc end reaching a position ventrad of the plane of the anterior transverse lamina. At this stage the paired naso-palatine cartilage is represented by a stretch of prochondral tissue upon the lateral side of each incisive duct.^ The U-shaped beginning

kandl the medial and lateral limbs of the ventro-lateral process of the septal cartilage. The incisive duct is partly enveloped on its lateral side by a troughshaped nasopalatine cartilage, which reaches from a spot near the oral end of the duct, to the level of the entrance of the duct of Jacobson's organ. The last named canal is surrounded completely by a cartilaginous tube which, as it ascends toward the cavum nasi, gradually loses its dorsal wall and becomes continuous with the anterior extremity of the paraseptal (Jacobson's) cartilage. The cartilage of Jacobson's duct lies medial of the process of the lamina transversalis anterior; neither it nor the nasopalatine cartilage is joined with the processes of the anterior transverse lamina.

■* The paraseptal cartilage in embryos of 70 mm. presents an anterior process, continuous with the cartilage of the duct of Jacobson's organ, and continues posteriorly into a slender cartilaginous paraseptal rod. The latter lies against the side of the ventral edge of the vomer and extends to the floor of the posterior cupola, the lamina transversalis posterior with which it is directly continuous.


of the vomer (figs. 2, 26, 27) embracing the ventral edge of the septum nasi, hes at some distance behind the paraseptal.

The fenestra narina (figs. 2, 3, 4, 13), oval with its long axis dorso-ventral, faces rostro-laterad. Its medial boundary is formed by the septum, the lateral and ventral by the lamina transversalis anterior, the dorsal, by the paries nasi. The lateral margin presents a deep notch, the incisura lacrimaUs, where the naso4acrimal duct turns medialward toward its termination. The ventral margin of this notch is formed by the sagittal portion of the anterior transverse lamina; the dorsal margin by the rostral extremity of the atrioturbinale, or inrolled ventral margin of the paries nasi. The lacrimal notch is the meeting place of two divisions of the fenestra narina, a dorsal and a ventral. The former leads to the atrial region of the cavum nasi; the latter division receives the naso-lacrimal duct and opens into the inferor meatus. The names pars atrialis and pars lacrimalis, respectively will be used in further reference to these divisions of the fenestra narina. In the present stage the superior alar process is represented in dense mesenchyma. It is chondrified in the 7 cm. embryo.

Fenestra olfactoria (figs. 1, 13, 26). The paired, quadrilateral olfactory fenestrae offer wide communications, in the present stage, between the nasal and cranial cavities. Their position is in the caudal half of the nasal capsules, that is, in the subcerebral division, between the tectum nasi, anteriorly, and the roof (lamina infracribrosa) of the posterior cupola, posteriorly. The plane of each opening looks dorso-caudad and also somewhat mesad. The anterior boundary, formed by the caudal free margins of the tectum nasi and prominentia lateralis, is considerably wider than the posterior limit, made by the infracribrous plate. The medial side is formed by the dorsal concave margin of the septum nasi. The obhque lateral boundaries of the olfactory fenestrae are formed by the dorsal margins of the prominentiae laterales and the antorbital plates. On the anterior and lateral sides of the fenestra olfactoria of the model are three processes, the caudal extremities of the crista semicircularis, ethmoturbinal I and ethmoturbinal II. Between these processes, the


olfactory fenestra extends in the form of three bays or sinuses. In the present stage there is no cartilaginous cribriform plate; its place is occupied by mesenchyma perforated by the olfactory nerves. These are in two groups, one near the septum, the second represented by bundles passing through the sinuses at the side of the window. The sheet of mesenchyma perforated by the olfactory nerves does not quite fill the fenestra olfactoria. It is closely applied to the septum nasi, then crosses the midline to be continuous with the membrane of the opposite side; it is fixed at the anterior boundary of the window (tectum nasi) . Posteriorly the membrane passes over the lamina infracribrosa against which it is closely applied. In the lateral direction it becomes continuous with the membrane filling the sphenoethmoidal fissure. This is fixed to the dorsal but not to the ventral boundary of the fenestra, which, it will be recalled, is formed by the antorbital plate. In relation to this membrane is the lateral branch of the nasal nerve (fig. 26). The latter appears in the sections through the ajiterior part of the orbit, running toward the spheno-ethmoidal fissure. This it enters, passing ventrad of the membrane, a position it retains in its course forward to the olfactory fenestra. Here it enters the nasal capsule, pursues a short course through the crista semicircularis and then passes to the exterior by one of the epiphanial foramina.

Septum nasi (figs. 1, 2, 25 to 30). This median partition is the direct continuation of the trabecular plate forward into the ethmoidal region. It is low and broad cau dally where it lies between, and enters into the median walls of the posterior cupolae. As it extends forward the dorsal margin ascends, the height of the septum increasing gradually toward the fenestrae olfactoriae; between the olfactory fenestrae it increases rapidly and reaches its maximum height at the tectum nasi; then decreases gradually in approaching the level of the fenestrae narinae. The ventral margin of the septum presents a longitudinal concavity, and is thickened between the paraseptal cartilages and again opposite the vomer. Anterior to the paraseptais it is united on each side to the anterior transverse lam


ina, the line of junction being marked ventrally by a longirudinal groove. The anterior free margin is straight and reaches further rostrad than the lateral wall of the nose. Dorsally, the septum presents between the olfactory fenestrae a free concave edge (there is no evidence of a crista galli) and, in its precerebral portion, continues into the tectum nasi. The latter relation appears in the sections of the anterior third of the nose as a bifurcation of the septum into two laminae, which extend on either side into the roof of the nose (fig. 28). In this way the sulcus suprasep talis is formed.

Interior of the nasal capsule (figs. 13, 25 to 30). The conformation of the nasal capsular walls from within is extremely simple at the present stage as compared with that of the adult. Turbinal processes are in evidence in the form of ridges, the maxilloturbinal alone showing any degree of scroll form. The bones in relation to the nasal cavity are the vomer, incisivum and maxilla.

Lateral wall of the cavum nasi. In figure 13, which represents the left lateral wall of the nasal capsule from within, the following parts, already referred to, will be recognized: the tectum nasi extending from the fenestra narina to the fenestra olfactoria, the posterior cupola and lamina transversalis anterior, parts showing cut surfaces at their junction with the septum nasi; ethmoturbinale I and II presenting at the olfactory fenestra; the crista semicircularis and the nasoturbinal ; the inrolled ventral margin of the paries nasi forming the maxilloturbinal and atrioturbinal. The lateral wall of the cavum nasi presents two unequal divisions: a small region ventrad of the level of the maxillo- and atrioturbinals, and an extensive region dorsad of this level. The former includes the sagittal portion of the lamina transversalis anterior and the incisura posttransversalis; the latter corresponds to the rest of the lateral nasal wall.

Ve7itral region of lateral wall of the cavum nasi. The sagittal portion of the lamina transversalis anterior presents a smooth, concave surface toward the cavum nasi, becoming broad ventrally where it passes into the floor of the nose (frontal portion of the lamina transversalis anterior). Caudad of the lamina


the cartilaginous lateral wall of this region is deficient; here the great notch, the incisura posttransversalis is found (figs. 2, 3, 4, 13, 29). This space is entered from behind and ventrally by the • cartilaginous anterior maxillary process. The post-transverse notch is closed toward the nasal mucosa, by a layer of mesenchyma and by the incisive and maxillary bones which lie just outside the plane of the cartilaginous nasal wall (figs. 29, 30),

The dorsal region of the lateral wall of the cavum nasi corresponds on the exterior to the paries nasi as described on p. 380 and may, like it, be considered as presenting three divisions, anterior, middle and posterior.

In the anterior division (pars maxillo-nasoturbinalis) will be included, for convenionc, the inner surface of the lamina supraconchalis, although, as will be seen later, its primary relation to the components of the nasal side wall was not established. This division, then, extends from the level of the fenestra narina to the crista semicirculars, and is limited ventrally by the maxilloand atrioturbinals. The most cephalic portion reaches the atrioturbinal ventrally, and forms the lateral boundary of that part of the nasal cavity here designated the atrium. The atrioturbinal is in line with the maxilloturbinal but is separated from that process by a notch, the incisura maxillo-atrioturbinalis. The succeeding part of the anterior division stands opposite the area of the lamina supraconchalis and sulcus lateralis of the external surface of the nasal capsule, and further includes the maxillo-turbinal and crista semicircularis in its ventral and caudal limits respectively. The maxilloturbinal, triangular in general form, is continuous laterally by its base, with the paries nasi at the eminentia lateralis. It stands mainly in a frontal .plane; its medial margin, opposite the septum, inclines ventrocaudad to meet the caudal margin at an angle opposite the paraseptal cartilage. Anterior to the level of the paraseptal it participates with the paries nasi in the caudal boundary of the ■post-transverse notch; posteriorly to the paraseptal it forms, by its free margin, a wide notch or sinus with the ventral free border of the antoibital part of the paries nasi, both entering into the boundary of the fenestra basalis. The maxilloturbinal


reaches its greatest breadth opposite the base of the crista semicircularis; from this point caudad it rapidly diminishes in width. On the lateral wall, dorsad of the maxilloturbinal, is a low antero-posterior eminence produced by a slight inward bulging. of the paries nasi in the region of the lamina supraconchalis. This elevation is at the base of the great naso-turbinal body composed of mesenchyma in the present stage (figs. 13, 29). Posteriorly, the elevation in question and the nasoturbinal continue into the crista semicircularis (iigs. 13, 27, 28, 29). The latter, at present, forms the anterior boundary of the entrance into the recessus lateraUs (figs. 23, 27, 28, 29) ; it extends ventrodorsad to the tectum nasi, sweeping in a fine curve caudally, alongside the septum, and terminating at the olfactory fenestra as already described (p. 334, fig. 1). Where the crest approaches the tectum nasi, more or less complete discontinuity of the cartilage occurs in the line of the sulcus lateralis. A small fissure here, made by the foramina epiphanialia (figs. 28, 29), separates, for a short distance, the wall of the recessus lateralis from the tectum nasi. A wide meatus of the lateral nasal wall (here termed meatus supraconchalis) runs between the naso- and maxilloturbinal bodies, continuing anteriorly into the atrium. Posteriorly this meatus extends ventrad of the crista semicircularis (here it is very narrow) and enters into a wide space which opens into the fenestra basalis (fig. 13). Just where the meatus passes into this space, and at a spot anterior to the ventral end of the crista semicircularis, the lateral nasal glands lie under cover of the mucosa (fig. 27). In its posterior half the meatus supraconchalis is extended into a groove between the maxilloturbinal ventrally, and paries nasi (lamina supraconchalis) laterally, the sulcus supraconchalis of Voit.

The middle division corresponds to the prominentia lateralis of the exterior and presents the recessus lateralis to which reference has been made. This large cavity communicates with the general room of the nasal capsule by a wide opening behind the crista semicircularis. Anteriorly it undermines the crest for a short distance in the form of a bhnd pocket; caudally it reaches the first ethmoturbinal, and dorso-caudally opens into


the olfactory fenestra through the sinus between the crista semicircularis and ethmoturbinal I (p. 335). Although at this stage the cartilaginous parts fail to meet in forming a medial wall for the inferior portion of the lateral recess, the soft parts which stretch between ethmoidal I, the crista semicircularis and the floor shut off this part of the recess from the common meatus of the nose and convert this part of the cavity into a blind pocket. Similarly, the entrance from the common meatus into the recessus lateralis between ethmoidal I and crista semicircularis is very broad, as shown in the models, but by the presence of the soft parts over these elevations, it is reduced to a narrow fissure, the hiatus semilunaris of later stages. The lateral recess is incompletely subdivided into dorsal and ventral rooms, recessus lateralis superior and recessus lateralis inferior, by a slight frontal ridge springing from the lateral surface of ethmoturbinal I, and by an oppositely placed ridge of the lateral wall of the recess (fig. 27) . The superior lateral recess opens into the common meatus of the nose through the dorsal part of the hiatus semilunaris. Within it are two antero-posterior, curving ridges of the mucosa, one upon the roof, the other on the lateral w^all. Each includes a condensation of the mesenchyma which is in contact, but apparently not continuous by any transitional zone, with the cartilaginous wall. These ridges separate three outpocketings of the recessus lateralis superior, a lateral, a superior and an inferior groove (fig. 27). The recessus lateralis inferior is incompletely walled medially, as already explained. It lodges a blind sac of mucosa which conmiunicates with the connmon meatus through the ventral part of the hiatus semilunaris.

^ Posterior division. (Pars ethmoturbinalis) . This region of the dorsal part of the lateral nasal wall corresponds to the antorbital plane of the exterior. It is characterized by the presence of the bases of the ethmoturbinals. Ethmoturbinal I appears in the model as a massive irregular ridge, extending dorso-ventrally on the lateral nasal wall at the caudal margin of the entrance into the recessus lateralis (fig. 13). Anteriorly, the process goes over into precartilage in the form of a broad,


thick piece, triangular in section, extending cephalad toward the crista semicircularis (fig. 27). Its medial surface bounds the common meatus; a dorso-lateral surface enters into the floor of the recessis lateralis superior; the ventro-lateral face is turned toward the recessus lateralis inferior. Ethmoturbinal I reaches the olfactory fenestra in the thick layer of mesenchyma filling this space (fig. 1); ventrally it is separated by a wide stretch from the level of the solum nasi; anteriorly it enters into the caudal boundary of the hiatus semilunaris. Ethmoturbinal II (figs. 1, 13, 26) is represented by a small cartilaginous plate, with precartilaginous margins, jutting mesad from the lateral nasal wall between ethmoturbinal I and the posterior cupola. From the level of the olfactory fenestra its long axis extends ventro-rostrad. It terminates a considerable distance above the plane of the fenestra basalis.

The cavity of the posterior cupola lies behind ethmoturbinal I, occupying the caudal extremity of the nasal capsule. The walls formed by the septum nasi, lamina infracribrosa, lamina transversalis posterior and antorbital plane, are smooth. The cavity opens anteriorly and ventrally into the common meatus of the nose and the fenestra basalis.

Ethmoidal region in smaller embryos

The first evidences of chondrification in the ethmoidal region were found in embryos of 12 nun. in which the process was manifested in the septum nasi (fig. 17). In van Wijhe preparations the ventral part of the septum was stained blue (fig. 6), appearing in the form of two streaks extending from the trabecular plate forward, side by side and separated by a less deeply stained tract. In embryos of 15 mm. chondrification of the septum has extended dorsally, reaching its greatest height in front. It is now a single median cartilage (fig. 18). The septum is continuous with the trabecular plate caudally, and in front gives off a pair of arching processes from its dorsal margin (fig. 7). The latter, which may be called the parieto-tectal cartilages, are at this stage in relation to the roof and lateral wall of the anterior one-third of the nasal cavity. Whether these processes are pri


marily outgrowths of the septum is brought into question by the fact that they are most deeply stained in the van Wijhe preparations in their lateral parts and less so next to the septum. In the midline dorsally a deep groove, the beginning of the sulcus supraseptalis, lies between them. Besides the paired parietotectal cartilages, there is, in the nasal region of van Wijhe preparations of 15 mm., a mass of cartilage quite independent of other chondrifying tracts. This is a curved plate, overlying, on each side, the diverticulum of the cavum nasi, which later is included in the recessus laterahs of the cartilaginous wall. This cartilage may be referred to, tentatively, as the paranasal cartilage. The parietotectal and paranasal cartilages stand close together, the one in front of the other. It is of importance to note that the dorso-cephalic margin of the latter overlaps the dorso-caudal edge of the former. Van Wijhe preparations of 17 mm. exhibit still a third chondrifying tract in relation to the nasal wall (fig. 8). This is a small plate of cartilage at the very back of the cavum nasi, on either side of the nasal septum. It appears to be entirely free from the septum and paranasal cartilage. This lamina antorbitalis, as it may be called, lies in a plane anterior to that of the origin of the preoptic root from the trabecular plate, and behind the paranasal cartilage. It is curved about the caudal end of the nasal sac, thus indicating the beginning of the posterior cupola. The anterior margin of the antorbital lamina projects into the fold of ethmoturbinal I behind the diverticulum of the lateral recess, and is overlapped by the caudal edge of the paranasal cartilage. The latter is larger than in the preceding stage and presents anterior and ventral incurved margins, continuous with each other. The ventral edge projects into the fold of the maxilloturbinal and represents the base of the cartilaginous process of that name. The anterior incurved margin overlaps the lateral and, at the same time, caudal margin of the parieto-tectal, and fusion has occurred to some extent between them. A double layered curved ridge is thus formed, projecting into that fold of the nasal wall which bounds the lateral recess anteriorly; this is the beginning of the crista semicircularis. Where fusion has not occurred be


tween the paranasal and parieto-tectal cartilage, spaces remain, one of which is traversed by the lateral branch of the nasal nerve; this is the beginning of the foramen epiphaniale. The overlapping and fusion in the ventral region occui-s between the inroUed anterior and ventral margins, where, in later stages, the lamina supraconchaUs and nasoturbinal are found. The formation of these parts, however, was not observed. The dorsal margin of the parieto-tectal cartilage lies next the olfactorybulb and is joined with the commissm'a spheno-ethmoidahs. The triangular parieto-tectal cartilage has grown backward along the septum nasi nearly its full length, thus forming the roof of the nose. Its caudal and lateral obUque margin enters into the formation of the crista semicircularis. In its anterior part it reaches, in a ventral direction, as far as the naso-lacrimal duct. The lamina transversalis anterior is unchondrified. The paraseptal cartilages are represented by incompletely chondrified tracts, independent of other skeletal parts.

PART II. DISCUSSION Occipital region

Basal plate. As we have seen, the first evidence of a basal plate was in the form of a pair of small cartilages on either side of the notochord of the occipital region, united anteriorly by a hypochordal commissure. Wincza has aheady observed this early form of the base of the cranium and called the two component laminae the parachordal plates.

Regarding the chondrification of the basal plate in mammals, several authors have shown that the initial stage is characterized by the presence of a pair of cartilaginous centers or tracts, one on either side of the notochord in the occipital region. Parker refers to the basal chondrification in the embrj^o pig, first (75), as the 'investing mass,' later (77) as the parachordal


cartilages. It is not clear from the description that there are two separate elements present. The term "investing mass" is misleading with respect to the notochord ; Parker expressly states that the former Ues beneath the notochord and so represents the relations of these two parts in the figures. Froriep ('86) recognized a tendency to the formation of bilateral symmetrical anlagen of the caudal part of the occipital floor in the calf (p. 91, woodcut to fig. IV, fig. IV, 2). This author remarks on the striking difference between cervical vertebrae and the occipital vertebra presented by the bilaterality of the anlagen of the bodies. In the former it becomes more pronounced in the cranial direction, whereas in the occipital vertebra this condition is presented in a lesser degree. In reference to the anlage in the occipital floor, Froriep says: " In einem Querschnitt dagegen wie Fig. IV, 2, ist eine bilaterale Sonderung des Knorpelgewebes nicht zu bemerken, die Knorpellage ist hier ventralwiirts der Chorda fast ebenso machtig wie zu beiden Seiten" (p. 91).

In the fight of recent studies the condition represented in the figure might justly receive a different interpretation wdth respect to bilaterafity of the anlage; the figure shows two cartilages in the occipital floor, one on each side of the notochord and some distance removed from it, united by continuous cartilage across the midline ventrad of the notochord. Levi ('00) described a pair of precartilaginous and cartilaginous anlagen for the middle piece (basilar portion) of the occipital region in a human embryo of about 13 nmi. These were united across the midplane by connective tissue in which ran the notochord, and which was continuous in a caudal direction with the connective tissue of the first cervical vertebral arch (p. 355). Bardeen ('08) also observed the beginning chondrification of the base of the occipital in man' in two bilateral centers. Other investigators have recorded an unpaired chondrogenous beginning of the occipital basal plate, or, again, its development in connection with the lateral occipital arch. Weiss ('01) found, in the white rat, that chondrification of the floor of the occipital region was manifested, first, by centers in the hypochordal region, one of them in Froriep's apparently unsegmented portion, the other in the


region of Froriep's occipital vertebra. Noordenbos ('05) found the occipital basal anlage to be single in the mole, calf and pig and also in the rabbit. Regarding the parachordal plate in the latter, this author found that it was formed by the fusion of the opposed ends of the two free occipital arches which were first to develop and from this starting point, grew forward (p. 375) . Here are then, apparently, three different conditions presented in the origin of the cartilaginous basal part of the occipital in manmaals: the appearance of a hypochordal center; a pair of bilaterally placed masses; origin by growth and fusion of the apposed ends of the lateral occipital arches.

Regarding the relation of the notochord to basal plate, it was found that in cat embryos of about 10 mm., the former enters the occipital region between the parachordal cartilages and lies in the dorsal part of the mesenchymal sheet which unites these cartilages across the midplane. This mesenchymal sheet later becomes chondrified in connection with the parachordals, forming thus a hypochordal bridge closing the space which originally separated these cartilages (Terry, '13). Where the notochord lies between the parachordals it is surrounded by a layer of denser mesenchyma than that concerned in hypochordal arch formation; this specialized sheath is continuous with the mesenchyma about the notochord in the region of the atlas — that in which the atlantal centrum is later formed. In regard to a cartilaginous hypochordal layer in the occipital region of mammals generally, there seems to be no doubt of its constant occurrence. My own observation of the position of the notochord with reference to the basi-occipital region in cat is in agreement with that of Williams ('08), namely, that after chondrification is well established, the caudal portion of the basal plate is hypochordal. Recently Kernan ('15) has observed the hypochordal position of the basal plate in cat embryos. Parker, Froriep, Levi, Noordenbos, and Weiss have observed, in various mammals, that the caudal part of the occipital basal cartilage is ventrad of the notochord.

The parachordals in cat are united at their anterior ends across the midplane (embryos of about 10 mm.) by a hypo


chordal commissure of young cartilage. This lies at a level cephalad of that of the anterior root bundle of the hypoglossal nerve. Cartilage afterwards forms dorsad of the notochord in the region of the primary commissure, so that the former comes secondarily to he within the basal plate. I am aware of the fact that some observers have described this anterior commissure of the parachordal in other forms as lying dorsad of the notochord, but in cat I have found it primarily hypochordal. Weiss, however, found the primary process of cartilage formation hypochordal in the apparently unsegmented division of the basal plate. The next cartilaginous union to be estabhshed between the parachordal cartilages in cat is the hypochordal arch uniting their caudal extremities. This has formed somewhat earUer than the anterior arch of the atlas, next to which it stands. The posterior hypochordal commissure is in the same transverse plane as the primitive lateral occipital arch. Between the two hypochordal arches, anterior (primary) and posterior, there remains a sheet of mesenchyma, stretching from side to side beneath the notochord, and continuing into the medial edges of the parachordals. The direction in which chondrification proceeds in this tissue is lateromesad. In the middle of this tissue a third conamissural process is indicated where the medial edge of each parachordal sends a projection toward the midplane, producing a constriction in the vacuity between the parachordals. These two symmetrical processes are directed ventro-mesad to a hypochordal plane and lie at a level corresponding to the middle of the future hypoglossal foramen. In the smaller embryos (9 mm.) four membranous arches in the occipital region were noted by Kernan, who states that the two cranial have a tendency to fuse.

When, in the cat, the parachordal plates have formed, the three root bundles of the hypoglossal lie against their lateral margins, in the angle formed by the lateral occipital arch. The foramen is later completed by the formation of a cartilaginous bar in front of the nerve roots, which extends laterally and dorsally from the anterolateral corner of the parachordal, to unite with the primary lateral occipital arch beyond the



nerve roots. The three roots are united in passing through the hypoglossal notch of the primitive cartilaginous occipital element, and there is no indication of a subdivision of the margin into smaller notches for each root, as appears to be the case in an earlier stage and before cartilage has formed. This condition varies in maimnals; in rabbits for example, (Noordenbos, Voit) two canals are present on each side. That two canals occur occasionally in the human skull as an anomaly is well known. It will be recalled that the hypoglossal nerve in cat embryos shows a tendency to conform with a spinal nerve in presenting a posterior root and ganghon in relation to the third motor root. The presence of this ganglion was discovered by Vulpian ('62) and I can confirm this observation here.

The presence of more than two cartilaginous hypochordal commissures in the development of the occipital basal plate has, I believe, not been recognized. Also, the recognition of the primary hypochordal nature of the anterior commissure apparently rests on only one other observation, that of Weiss, in the rat. Until further distribution of these phenomena be discovered, or the present observation confirmed, it would seem of Uttle profit to attempt an interpretation of their significance. The antero-posterior succession of a series of transversely placed bar-hke structures in the base of the skull at once suggests the idea of segmentation. A relation of these bars laterally with the hypoglossal nerve roots may or may not be a primary one. But in utilizing the results of chondrocranial study in any discussion of the segmentation of the cranium, it must be borne in mind that such evidence can be of high value only when the relation of chondral to blastemal developmental processes is known. It seems highly probable that the cartilaginous commissures here described and the membranous bars observed by Kernan ('15) are two developmental phases of one and the same process, which subsequent investigation will elucidate.

In regard to the condensed tissue sheath about the occipital notochord found in cat, it may be said that a number of observations made on different animals point to the presence of such a layer as probably of general occurrence. Froriep ('86, p. 92)


records the presence in the calf of a connective tissue layer dorsad of the chorda. Weiss ('01) found, in embryos of the white rat, an exceptional development of the perichordal sheath in the occipital region, not separated from the horizontal plate of the primitive arch of the atlas (p. 511). The work of Gaupp ('06) on Echidna, and Weigner ('12) on man, support the view, first brought forward by Weiss, of a special development of perichordal tissue about the occipital notochord in mammalian embryos. Weiss described the notochord as occupying, in the blastemal stage, a position at the dorsal surface of the segmented portion of the floor of the occipital region; he found that the cranial end of the perichordal sheath grew to assmne a globular form; in the chondral stage, cartilage appeared in the perinotochordal sheath, quite independently of the process of chondrification which takes place lateral to and beneath the notochord in the formation of the lateral occipital arches and the basal plate. The more or less spherical cartilage so formed about the notochord is fused, in later stages of development, with the cartilaginous centrum of the atlas, and becomes eventually the extremity of the dens epistrophei. Weiss saw in the cartilage forming the end of the dens epistrophei, which is marked off from the cartilaginous centrum of the atlas by grooves, an element comparable with a vertebral centrum, and concluded that it represented the body of an occipital vertebra, or of a proatlas. Gaupp ('06) found the dens epistrophei in Echidna embryos to be composed of the centrum of the atlas and, in addition, of material lying cephalad (and perhaps derived from the basis cranii). This author is of the opinion that the dens epistrophei and the ligamentum apicis dentis represent the anterior reduced end of the vertebral column in which a number of vertebral centra lie imbedded. Weigner ('11) found in a human embryo of 13.5 mm. paired anlagen in the floor of the occipital region, presenting a notch laterally for the hypoglossal nerve, with a deep groove between their caudal extremities occupied by mesenchyma, in which the notochord lies. In the ventral part of this sheet, the hypochordal arch of the occipital region was observed. In the atlantal region, in the tissue dorsad of the notochord, the


paired anlagen of the atlantal centrum were noted. Evidence of the beginning of the occipital centrum was presented in an embryo of 14.5 mm.; the anlage was clearly defined in a later stage (15.3 mm.) as a pair of centers behind the notochord, which pass in a caudal direction gradually into the older centers of the body of the atlas. In man, according to Weigner, the body of the definitive epistropheus with its tooth-process, is developed from three vertebral bodies; those of the occipital vertebra, atlas and epistropheus. The dens itself is formed from the centrum of the occipital vertebra and of the atlas. In the present work no separate centers, in advance of that for the atlantal centrum, were seen, but the tissue about the occipital notochord became condensed next to the chondrifying center of the atlas, and eventually became the cartilaginous apex of the dens epistrophei. In the ossification of the dens epistrophei of cat (as in several mammals) there is, in addition to the bilateral atlantal centers, a single center for the apex of this process (Jayne '98).

Since Froriep's work on the development of the occipital region ('83, '86, '02), the interpretation which that investigator drew from his own results of the relation between cranial and vertebral development and structure has been, in general, sustained; some of the conclusions have, however, been modified by new evidence brought out by recent research. Noordenbos, in 1905, attacked the evidence which has been used in support of the vertebral theory of the skull, claiming in effect that it does not support the homology of parachordal plate and occipital arches with vertebral centra and vertebral arches. Noordenbos rightly objects to the comparison of vertebral bodies, arising as separate, rounded cartilaginous masses, with the parachordal plates, continuous unsegmented masses presenting no trace of special chondral centers. He states that vertebrae arise around the body notochord while the parachordal plate does not. This plate takes origin in one of the three ways mentioned above: as an independent center at the side of the notochord; in connection with the lateral occipital arch ; in connection with a hypochordal plate. By the van Wijhe method the vertebral


centra in the mole were found to arise in connection with intravertebral capsules, outside the sheath of the notochord. Furthermore, the relation of the neural arch of a vertebra to its centrum differs from that between the lateral occipital arch and parachordal; in the former the parts are primarily independent, in the latter continuous. On account of these discrepancies between vertebral and occipital chondrogenous development, Noordenbos finds weak support for the vertebral theory in the chondrocranium. He says (p. 373): Ich mochte aus diesen Erscheinungen, im Gegensatz zur allgemein giltigen Auffassung, schUessen, dass fiir das Chondrocranium die Wirbeltheorie nicht aufrecht zu halten sei." But the vertebral theory, notwithstanding the blows dealt it from the time of Huxley's attack to the present, has shown itself tenacious of Ufe, and the thought uttered by Oken more than one hundred years ago demands deference of the worker of today.

Recent investigations into vertebral development, (Bardeen, Weiss, Weigner) demonstrate the presence of a pair of chondrogenous centers, lying close to the midline and reacliing a somewhat higher level dorsally than the notochord, which grow to surround the latter, and so form the cartilaginous vertebral body. There is, in general, apparently a fundamental difference in the original relation to the notochord of the parachordal plate and vertebral centrum. At one spot, only, does the relation of the parachordal to the notochord approach that of a vertebral centrum, namely at the level of the anterior commissure; cartilage is here developed around the notochord. The discovery of processes of chondrification in the perichordal sheath of the occipital region, related eventually to the dens epistrophei and which fulfill the conditions of vertebral central development in general, have been mentioned. It would seem that the evidence presented by these investigators (Weiss, Gaupp, Weigner) justifies the interpretation which has been put forward of rudimentary centra in the occipital region. To this evidence must be added that given by cat. By the interpretation of Weiss of the hypochordal nature of the caudal part of the basal plate, the objection to comparing the lateral occipital arch with the


neural arch of an ordinary vertebra is largely overcome. The lateral occipital arch presents the same relations to a centrum as does the atlantal neural arch, and also the same relations to a hypochordal arch as obtains between the neural and ventral arches of the atlas. That the order of formation of the hypochordal and lateral occipital arches varies somewhat in different species is no obstacle to the interpretation of their equivalency to vertebral structures; the work of Weiss ('01) and Levi ('08) has shown that there is also great variability in the development of the ventral arch of the atlas among mammals.

The fundamental differences between occipital and vertebral development, which have been so clearly indicated by Noordenbos, must, I think, be recognized; but if the interpretation of Weiss be correct, that the occipital develops not as a typical vertebra, but like a specialized vertebra, namely, the atlas, we must admit that there is still evidence of vertebral structure in the skull. This interpretation is in accord with the physiological environment of the region: a transitional zone between a rigid (cranial) and a movable (spinal) division of the axial skeleton. Regarded from this aspect, the structural conditions are seen to change as we pass along the vertebral column toward the head, the form of the vertebrae becoming less typical, due to tendencies along certain definite lines (regression of centrum, development of hypochordal parts) . The occipital region shows by its structures the culmination of these tendencies; extremely rudimentary centrum separated from all connection with highly perfected arches, lateral and hypochordal; characters which, to a lesser extent, mark the atlas from a typical vertebra.

There is probably more than a superficial resemblance between the lateral occipital arch and neural atlantal arch in respect to the relations of these parts to nerves. The primary notch formed for the hypoglossal nerve bundles at the side of the parachordal plate, in front of the base of the lateral occipital arch (Noordenbos describes it in the occipital arch), and its subsequent conversion into a foramen, are phenomena which seem to parallel the development of the atlas in the region of the first spinal nerve. In cat embryos of 15 and 23 mm. the first spinal


nerve makes its exit through a notch at the side of the lateral mass of the atlas, between the base of the neural process posteriorly and a short but prominent atlantal process directed dorsad from the lateral mass. The articular regions of the atlas and occipital are comparable, in respect to their relations, to the components of these two skeletal elements. In cat the occipital condyles are formed at the spot where the lateral and hypochordal arches come together, that is on the parachordal plate. This is in agreement with Levi's and Weigner's observation of the relation of condyle and parachordal in man. Likewise, the corresponding articular surfaces of the atlas are formed where ventral arch and neural arch come together, namely at the massa lateralis. The early establishment of similar relations to nerve and articular surface offers a basis for a comparison of parachordal plate and lateral mass of atlas. Apparently these parts are not distinct elements in their relation to the occipital and atlantal arches, but mark a definite locus between the hypochordal arches on the one hand, and neural or lateral occipital arch on the other, constitute the region of articulation, and mark the ventral limit of the exit of the nerves. In this sense the parachordal plate could not include the cartilage ventrad of the notochord; this would fall into the category of the hypochordal arches. The terai occipital basal plate would include the originally paired parachordals and the hypochordal cartilage. Comparison of parachordal plate and lateral mass which is here proposed, while attempting to bring out the relations of the occipital and atlas in further detail does not at the same time close the way to a better understanding of a possible relation of atlantal centrum and lateral mass which has been advanced by authors (Hagen, '00).

Atlas. Reference has been made to a peculiar character of the atlas, namely the foramen in the neural arch of the adult bone and its relation to the notch in the arch of its cartilaginous predecessor. This atlantal foramen transmits the vertebral artery and the first spinal nerve in cat. It is present normally in most, if not all mammalian orders, holding a position nearer or farther from the cephahc margin of the neural arch, through which varia


tion of its position the bar of bone forming its cephalic Umit is narrow or broad. In man the first spinal nerve and vertebral artery normally traverse a notch in the superior margin of the neural arch; in some instances, however, they pass through a foramen produced by osseous bridging of the notch. Bolk ('99) has pointed out that the usual conditions are indicative of regressive processes in the formation of the human atlas, that in man this bone is reduced in mass as compared with the atlas of those animals in which an atlantal foramen obtains. The bony reduction goes hand in hand with the imperfect development, and probable loss to some extent, of muscle and nerve in the dorsal part of the neck at the level between epistropheus and occiput. The recognition of the human atlas as an atypical example of the form which generally prevails in mammalia is helpful in approaching problems of structure in the head-neck region. Out of the recently much discussed phenomenon of manifestation of an occipital vertebra or assimilation into the occiput of the atlas (Swjetschnikow '06, Kollmann '07, v. Schumacher '07, Smith '09, Glaesmer '10), the question has presented itself to me as to the development of parts in relation to the atlantal foramen. A number of observations on the development and comparative anatomy of the atlas have been made and will be reported in another place. At present I wish to discuss only those which have some bearing upon the occipital region.

The atlas of the smallest cat embryo studied (10 nmi.) was represented in cartilage by a pair of neural arches, a small centrum about the notochord, and the beginnings of the hypochordal arch. The expanded base of the neural arch (lateral mass) sends dorsad in front of the latter a small, blunt atlantal process. The notch between the process and arch lodging the first spinal nerve and vertebral artery, is the first step toward the formation of the atlantal foramen. In the stage represented by the model, the notch is relatively deeper, owing to the increase in length of the atlantal process, which is now in the form of a short bar. The base of this bar, anteriorly, participates in the articulation with the occipital condylar surface; its dorsal extremity is free. The boundaries of the foramen are


completed, in the full term fetus, by connective tissue stretching between the free extremity of the anterior bar and the neural arch, anterior to the first spinal nerve.

Observations on the sulcus in the atlantal neural arch for the first spinal nerve in man have been made by Macalister ('93), who gave the name post-glenoid tubercle to the process which rises from the lateral mass and hmits the sulcus anteriorly. This process was found to vary considerably in its extent, an observation easily verified in even a small series of specimens. MacaUster also noted the presence of an independent bony center in the ligament completing the atlantal foramen in the atlases of young skeletons. The question, whether this represents the typical mode of origin of the bony rod which, in man, completes the atlantal foramen, remains for future enquiry. Separate ossifications in the region of the neural arch of the atlas, between it and occipital, have several times been observed in man. Trolard ('92) found, in two instances, in the posterior ligament of the atlanto-occipital joint an osseous bar placed horizontally, in one case nearly reaching the midplane. The possibility of relation of these independent ossicles in man to the arch of the atlas has, I believe, not been considered. It is well known that an osseous element in the posterior occipito-atlantal ligament, compared with the proatlas of reptiles and extinct amphibia, has been described for Erinaceus (see Baur, '94). The possibihty of the proatlas being a component of the atlas in a persistent type was many years ago suggested by Osborn ('00). The question of special interest which I wish to mention here in regard to the post-glenoid tubercle, the bar of bone completing the atlantal foramen and the ossicles in the posterior occipitoatlantal ligament, is whether these structures may not possibly represent parts of one element which, in a primitive state, was separate from, but closely related to the atlas.

Occipito-atlantal articulation. The discovery by Fischer ('01) of a single, horseshoe-shaped surface at the ventral margin of the foramen magnum in embryos of Talpa, articulating with the ventral arch of the atlas would seem to lessen the gap between reptihan monoconydlic and mammalian dicondylic articulation.


That the condition in Talpa is primitive, and not secondarilyacquired, is supported by Gaupp's ('08) observation of a similar atlanto-occipital articulation in Echidna. Whereas, in cat, the occipito-atlantal articulation is apparently dicondylic primarily, the joint surfaces are located to a considerable extent upon the basal plate, as well as on the lateral arch, and meet corresponding surfaces of the ventral arch of the atlas and its lateral mass, including the atlantal process. The participation of the basal plate in the articular surface is evidence of greater proximity of the two condylar surfaces toward the median plane than in those mammals (Lepus) where the surfaces are restricted to the region lateral to the foramen magnum. This embryonic state of the condyles is, therefore, apparently intermediate between typical dicondylism and the condition observed by Fischer in Talpa. In the later development of cat embryos the condylar surface grows further dorsad by the side of the foramen magnum.

If we accept the comparison of occipital and atlas, we must throw aside the idea of that form of articulation between them, such as exists between the centra of typical vertebrae.

The basal plate in the cat is a derivative of the parachordal cartilages and hypochordal arches and therefore the development of an articular surface upon its caudal margin must involve either one or both of its constituent elements. The point to be emphasized now is the fact that a part of the occipital element, which apparently is as distinct from the centrum as the lateral mass is distinct from the body of the atlas, enters into the constitution of the condyle; that is to say, the condyle belongs to an arch structure.

Further study of that region dorsad of the condyle which is characterized by a notch traversed by veins is necessary before any conclusion can be drawn regarding its significance; it seems not improbable that it may have to do with the foramen condyloideum.

Plane of the foramen magnum. Although a marked flexure exists between the head and trunk of cat embryos, this does not explain the basal position of the plane of the foramen magnum


and the occipital condyles referred to on p. 292, which is in contrast with their caudal position in the adult. The immediate cause of this basal position is not difficult to find. Since the lateral occipital arches are quite narrow antero-posteriorly and are joined directly with the otic capsules, it is e\ddent that their basal inclination cannot be attributed either to growth, as in man, or to the presence of wide fissures between them and the ear capsules. The explanation is to be found in the flexures to which the whole chondrocranium is subjected. One result of these flexures upon the longitudinal axis of the cranium is to put its anterior and posterior halves nearly at right angles with each other. It follows that the plane of the foramen magnum stands parallel with that of the floor of the nose. It is of interest to find that, with the formation of the bony cranium, a nearly straight longitudinal axis is substituted for the primary angular axis and, with it, the plane of the foramen magnum becomes less obhque.

A primitive condition, and an exceptional one in the primordial skull of mammals, it seen at the stage represented by the model in the abrupt ascent of the lateral occipital walls from the basal plate. The mamjnahan chondrocranium, as contrasted with that of lower animals, shows a tendency toward lateral extrusion of these walls; as Gaupp ('06) says: "Die Seitenteile der Occipitahegion sind bei Saugern nicht mehr steil aufgerichtet, sondern nach hinten hin basalwiirts niedergelegt am starksten und vollkommensten beim Menschen."

Basal fissures. The fissura basicochlearis posterior of cat is apparently comparable with the opening of the same name in the chondrocranium of Talpa, first described and named by Noordenbos ('05). Mead has noted a probable homologue of the posterior basicochlear fissure in Sus. In Lepus this opening is not present in the stages of development studied by Voit, but an anterior basicochlear fissure was observed by that investigator. These fissures are filled with mesenchyma, which, in the cat embryo, passes into a zone of precartilage at the edges of the opening. The posterior basicochlear fissure, present in Talpa embryos of 14 mm., disappears in embryos of 19 to 20


mm. and older, being replaced by cartilage (Noordenbos). Parker ('85) and Fischer ('01) had already shown continuous cartilage in the region between basal plate and cochlear prominence in Talpa embryos of later stages. Tarsius exhibits still another phase of vacuity at the basicochlear junction, presenting an extensive slit separating the ear capsule and cranial floor (Fischer, '05). In the adult cat, a fissure separates the pars petrosa from the basi-occipital and it appears that in the adult pig the foramen lacerum anterius and the foramen jugulare are connected by a fissure median to the auditory bulla (Mead, '08). The persistence of an original fissure, growing larger as the cranium enlarges, is an interesting phenomenon calling for further study. The posterior basicochlear fissure of Talpa is one of several spaces which, as Noordenbos has shown, are derived from the original space separating the independently arising otic capsule from the basal and lateral parts of the chondrocranium. That part of the original space between the base of the skull and the ear capsule is broken up by the later formation of synchondroses, uniting the auditory capsule with the parachordal plate and with the basal region which later enters into the sphenoid; and so there arise a canaHs caroticus, an anterior basicochlear fissure and a posterior basicochlear fissure. The observations on cat embryos, presented here, show that the origin of the cartilaginous basal plate and otic capsule are independent and that the fissura basicochlearis posterior is derived from the original space separating the parts.

Foramen magnum. In Lacerta the dorsal boundary of the foramen magnum is the tectum synoticum. Fischer ('03) found the apparent foramen occipitale magnum of the Semnopithecus embryo larger than the future real foramen magnum, the hinder part of the former being closed by membrane. This author remarks on the probabihty of the membrana atlanto-occipitaUs in part undergoing ossification. Bolk ('03) described this region in human embryos, and applies the name incisura occipitalis posterior to a httle space filled with membrane made secondarily by the approximation of the dorsal extremities of the occipital side walls in the formation of the foramen magnum. Between


this notch and the tectum posterius is a broad space filled by membrane continuous ^\ith that of the incisura posterior. A pair of chondral centers hes in this membrane and a third center stretches through the cranial roof anterior to the tectum posterius. The latter disappears in further development; the pair of centers remaining help to complete the boundary of the incisura posterior, while the membrane between the latter and the tectum posterius undergoes ossification. Voit ('09) also found a space (incisura occipitalis posterior) in Lepus between the dorsally open foramen magnum and the tectum posterius, and since, as is explained, the latter belongs to the otic region, the side boundaries of the incisura occipitahs posterior are regarded as the dorsal portions of the occipital pillars. If we compare now the extent of the region between the dorsal confines of the foramen magnum and the tectum posterius in Hzard, rabbit, cat, ape and man, it at once appears that there is a progressively increasing area exhibited. This begins with Lacerta, where, as Gaupp observed, the tectum is so shaped as to give to the foramen magnum an angle in the dorsal median fine, and reaches the great expanse described by Bolk in the human embrj^o, in which the tectum is far removed from the foramen magnum. It is evident, also, that the term incisura occipitahs posterior has not the same value throughout its application; in the rabbit Voit apparently regards the space limited laterally by the dorsal portions of the occipital pillars as the incisura occipitalis posterior, while Bolk limits the term to only a small part of the region encompassed by the dorsal limits of the occipital walls.

Otic region

Position of the otic capsules. It has been noted (p. 298) that the approximately transverse position of the plane of the otic capsules in the stage represented by the model forms an exception to the general rule in mammals of obhquity of the principal otic axis toward the longitudinal axis of the skull. It will be recalled that the degree of differentiation exhibited by the stage of the model is one, expressed in terms of the skeleton, wherein bone formation has advanced but little; of the mem


brane bones several have still to make their appearance and none of the purely endochondral ossifications is present. Talpa of 27.3 mm., Semnopithecus of 53 mm., Homo of 8 cm,, in which the obhquity of the principal otic axis has been recorded, are all at stages of development in which bone formation is well established and therefore can hardly be compared with cat embryos in the stages under consideration. On the other hand, obliquity of the otic axis is present in the mole of much younger stages, if one can judge by the photographs of Noordenbos, and the same is true for the Sus cranium described by Mead, which is at approximately the same stage as the Felis cranium modeled. Again, Voit states that in Lepus the two canalicular parts stand parallel, and only a shght convergence of the cochlear parts is present; the stage is one in which osseous development is advanced (45 mm. gr. L.). Therefore, it appears that the definitive oblique direction of the long axis of the auditory capsule, typical of mammals, is attained at different periods in the species considered, the tendency being toward its early establishment. As in the case of the occipital region, so also with the otic capsule, the position changes with the development of the bony cranium. The shifting of the otic axis from a position at right angles to the cranial base in the chondral stage to one of marked obliquity (cephalo-ventro-mesad) in the osseous stage, can hardly be a result of the straightening of the cranial axis. The factors involved in influencing the change must remain for future inquiry.

Another characteristic of the position of the mammahan otic capsule is its location in the base of the skull rather than i i the lateral wall as is the case in increasing degree from this class back to lower vertebrates. In mammals, not only does the cochlear capsule hold a basal position, but the phylogenetically older canalicular part is rotated ventralward as well as caudalward. Toward the attainment of this mammalian peculiarity the cat cranium, in the stages under discussion, presents what seems to be the initial steps. Though the cochlea, a phylogenetically later acquisition to the ear, asserts itself early in claims for space, the relations established in the chondocranium


are readjusted in the bony cranium, the cochlear capsules being widely separated by a broad processus basilaris, and are themselves relatively smaller in the adult skull.

Origin of the cartilaginous otic capsule. In discussing the observations which were made on the chondrification of the otic capsule two questions of special interest present themselves: the relation which the origin of the cartilaginous capsule as a whole bears to the rest of the cranium; the original relation between the pars cochlearis and the cartilaginous basis cranii. The first question was raised by Huxley and there was sufficient evidence, notwithstanding the crude methods of his time, for the advancement of the theory of intrinsic skeletal capsules for each of the sense organs, ear, eye, and nose. Subsequent discovery has tended to confirm the truth of this theory, the evidence coming, as might be expected, chiefly from the lower vertebrates. What mammalian crania will show must wait until the study of successive developmental stages of species has been repeated by modern methods. By the van Wijhe method, Noordenbos has presented, recently, very strong evidence of the independent origin of the cartilaginous otic capsule in Talpa; similar results were obtained for rabbit, ox and pig.

The second question has arisen in connection with Gaupp's theory of the reformation of the cranial base to contribute a supporting wall for the cochlear duct. There appears to be in the reptiles a beginning development of the cochlear capsule at the expense of the basal plate. Is the theory supported by evidence from the mammalian chondrocranium and, what concerns us here, does the development of the cochlear capsule in cat throw any hght on the problem? So far as mammals are concerned, it is convenient to consider these two questions together. First, it should be borne in mind that only the circumstances of chondrogenous development will be reviewed; the concfitions of blastemal structure are here excluded.

In regard to the pars canalicularis the evidence afforded both by sections and van Wijhe preparations indicates that this component of the ear capsule arises independently of other parts of the cranium. There is some difference as to the form of the


cartilage first appearing in the pars canalicularis from what has been observed in other animals. I refer to the observation of a plate of cartilage upon the lateral side of the semicircular canals; in the cat, cartilage forms upon the lateral surfaces of the canals in more or less separate stretches for each canal. The canals are subsequently completely walled and the intervals between them filled through the development of cartilage, but whether this is by extension of cartilage formation from the canalicular walls already established or from independent chondrifying centers was not observed. The more or less compact mass constituting the pars canalicularis is secondarily connected with the lateral occipital arch and parietal plate, although its dorsal margin remains free in the fissura jugulare spurium, and the posterior margin is clearly indicated, even at the 23.1 mm. stage, in the stretch of young cartilage between it and the lateral occipital arch directed toward the jugular foramen. The pars canalicularis is also apparently formed independently of the suprafacial commissure, if one may judge this by the difference in degree of development of these two closely associated parts. In the pars cochlearis, cartilage was first observed in the region next the pars canalicularis and in the anterior and posterior poles, i.e. in the neighborhood of the suprafacial and basivestibular commissures. There was no actual separation of these chondrifying tracts from one another, no independent centers of cartilage formation. At the stage when the medial wall of the cochlear capsule is in precartilage, the capsule is separated from the chondrified basal plate by a fissure filled with mesenchyma. Union of the capsule by cartilage with the basal plate behind the carotid foramen, with the alicochlear commissure and basivestibular commissure, is brought about secondarily. Between the connnissura suprafacialis and anterior pole of the cochlear capsule a smaller degree of difference in development obtains than is the case between this commissure and the pars canahcularis ; the ventral end of the commissure, however, appears to be blended, if not actually continuous, with the cochlear capsule. In this discussion, the fact of the continuity of these two parts is important because of the possibility of pa


rietal if not actual basal relationship and significance of the commissure. The evidence of the parietal nature of the suprafacial conmiissure may be considered at once, since upon its interpretation depends to a large extent the question of the cochlear relations to the basal plate.

The relations of the suprafacial commissure in cat appear to differ somewhat from those described in other mammals. In Talpa, according to Fischer ('01), the roof of the facial canal is made by a thin lamella of cartilage stretching from the pars canalium semicircularium to the highest elevation of the cochlear capsule. Fischer regards the walls of the facial canal as made entirely by the ear capsule, and contrasts this condition with the formation of the canal in Lacerta, in which, as Gaupp ('00) has found, the foramen for the seventh cerebral nerve lies in the boundary zone between the basal plate and ear capsule. Noordenbos ('05) named the roof of the facial canal the tectum nervi fascialis, and found it connected with the medial wall of the anterior ampullary swelling of the pars canalicularis. Voit ('09) found, in the rabbit, that the suprafacial commissure stretched from the anterior end of the pars posterior of the otic capsule, beneath the prominentia utriculo-ampullaris posterior and right above the superior acustic foramen, to the roof of the anterior part of the pars cochlearis. In comparing the facial foramen in rabbit and lizard, Voit evidently agrees with Gaupp that, in mammals, the walls of the foramen are in part contributed to by the cochlear capsule. In the dog, the suprafacial commissure bridges over the facial foramen, from the borders of the pars utriculo-canahcularis and the pars sacculocochlearis, according to Olmstead ('11). De Burlet ('14) described the commissura praefacialis of Balaenoptera rostrata as a cartilaginous bridge between the pars cochlearis and pars canalicularis. In De Burlet's plates VI and VII, however, there is shown what appears to be a connection between the praefacial commissure and lamina parietalis. The relations of the lateral end of the suprafacial conmiissure in the cat differ from those in the mammals mentioned, with the exception possibly of Balaenoptera. Connection between the conmiissure



and pars canalicularis is likewise seen in cat, but the relation is apparently not the primary one. The primary lateral termination of the suprafacial commissure in the cat is not in the otic capsule, but in the commissura orbito-parietalis, where the latter meets the prominentia ampuUaris superior. Continuity with the orbito-parietal commissure (a part of the lateral cranial wall) and, at the same time, partial separation from the otic capsule, is to be regarded as evidence pointing toward the parietal nature of the suprafacial commissure. While I find no reference to a relation between the suprafacial commissure and cranial side wall in descriptions of mammalian chondrocrania, the constant proximity of these parts strongly suggests that the continuity seen in cat may obtain in other mammals. De Burlet's figures of the Balaenoptera chondrocranium are of great interest in this connection. Further support of the view of the parietal nature of the suprafacial commissure is offered by the relation of this structure to the exit of the trigeminus. As stated in the description of the orbito-temporal region, the roots of this nerve and the semilunar ganglion lie between the suprafacial commissure and a thickened band of the transverse septum in which is developed an accessory cartilaginous rod resting upon the dorso-medial surface of the ganglion. A similar structure was observed in the rabbit by Voit who interpreted the 'Restknorpel b' as possibly representing the pila prootica of Lacerta, a structure of the primary cranial wall. In Lacerta the trigeminus makes its exit by way of the prootic fenestra whose boundaries are the prootic pillar and prefacial coixanissure. In the cat the fifth nerve leaves the cranial cavity between the accessory cartilaginous rod and the suprafacial commissure. Continuity of the anterior part of the otic capsule with a portion of the lateral wall of the cranium in the cat should make us hesitate in assigning entirely otic boundaries for the facial canal. The evidence so far indicates possibly that the foramen faciale in the cat stands in the boundary zone between the pars cochlearis of the otic capsule and the cranial wall, represented by the suprafacial commissure. Such a relation is in harmony with that which obtains in reptiles, to the


extent that the foramen for the facial nerve stands between a sense capsule (otic) and cranial wall (basal plate).

In accordance with our observations, the cochlear capsule in the cat is developed in connection with the pars canalicularis and with the suprafacial commissure which gives evidence of being, in part at least, a parietal structure. The cochlear capsulfe chondrifies independently of the basal plate, with which it is secondarily united by commissures. Its precocious growth and great bulk encroach upon the broad region occupied in lower vertebrates by the basal plate. If we can, on the evidence given, interpret the suprafacial commissure as a parietal structure in the cat, it would appear that its relation to the cochlear capsule (continuity) affords support to the theory (Gaupp) of the latter having preempted the territory of the basal plate and developed at its expense.

Foramina aciistica and meatus acusticus internus. The acustic foramen or fissure and the entrance to the facial canal are established early in the development of the otic capsule. These openings he between the suprafacial commissure behind and, in the 23.1 mm. stage, are found at the bottom of a shallow internal acustic meatus. The single acustic fissure is constricted in its middle so as partly to separate two wide divisions accommodating the vestibular and cochlear nerves. Completely separated foramina were not observed in any of the stages studied. Comparable with, this form of acustic nerve entrance is the single dumb-bell shaped foramen which IVIead has described in Sus. In the rabbit (Voit) and the dog (Olmstead) there are separate foramina for the vestibular and cochlear nerves. In the cranium of Talpa of the stage described by Fischer there are also two separate acustic foramina; but in younger embryos Noordenbos found a round porus acusticus internus, in the bottom of which openings for the cochlear and vestibular nerves were not clearly separated on account of lack of chondrification. The persistence of a single opening in the cat speaks for tardy development of the medial wall of the ear capsule, a tendency characteristic of the lower vertebrates. The relative positions of the foramina, one to the other, or of the nerves, where a single


opening exists, while in general constant for mammals, shows an interesting variation in the cat. In mammals these openings lie approximately one above the other, so that a superior and an inferior acustic foramen, for the vestibular and cochlear nerves respectively can be spoken of. In the cat, at the stage of 23.1 mm. the vestibular nerve occupies the antero-lateral part of the fissure; the cochlear nerve the medio-caudal end. The anteroposterior order of the acustic rami recalls the condition in reptiles. Although a common meatus for the two divisions of the acustic nerve occurs in birds (Tonkoff, '00) this is separate from the exit of the facial nerve, and therefore is not comparable wilji the internal acustic meatus of manamals, which includes separate passages for the vestibular and cochlear rami and for the nervus facialis. The development of the internal acustic meatus begins very early and seems to be a result, not of a depression of the medial otic wall, but, as Voit remarks, of the elevation of the surrounding cartilaginous parts. In the cat, the elevations are made by the suprafacial commissure anteriorly, the otic extremity of the basivestibular commissure posteriorly and the prominentia utricularis dorsally and laterally. These three parts rise above the level of the cochlear roof and form three sides of the meatus acusticus internus and porus acusticus. The low, medial and ventral side is the roof of the cochlea. Of these three elevations, that of the suprafacial commissure is most prominent and is probably the chief factor in determining the presence of the meatus.

Foramen perilymphaticum. Since Fischer's ('03) description of the derivation of the aquaeductus cochleae and fenestra cochleae from the perilymphatic foramen, several observations have been made by other investigators concerning this interesting phenomenon. Fischer found in an embryo Semnopithecus that the downward directed opening at the basal side of the ear capsule was divided into two parts by a process which sprang from the anterior margin of the opening and extended backward. The larger lateral part looked out from the free outer surface of the skull and was closed by thick membrane ; it was identified as the fenestra cochleae. The medial smaller opening, on a


higher plane, looked inward, was traversed by the perilymphatic duct and was recognized as the aquaeductus cochleae. Following his observation, Fischer states that Hertwig's enlarged model of the otic region of man shows obscurely the process in question. I find in this model a free corner of the cochlear wall projecting forward from the posterior margin of the large foramen so as to divide the opening into a large lateral and a very small medial part. In the rabbit, Voit found the conditions exactly as Fischer saw them in Semnopithecus and named the dividing process the processus intraperilymphaticus. Macklin ('14) described the intraperilymphatic process in the human embryo as a short conical projection directed forward from the inferior utriculoampullary prominence." In the cat, a process of the caudal wall of the cochlear capsule, adjacent to the basivestibular commissure, projects laterally, and tends to separate the perilymphatic foramen into two parts. We find, therefore, that the perilymphatic process in the ape and the rabbit is directed backward, and in man forward, while in the cat the process which separates the perilymphatic foramen into intraand extracranial openings is directed laterally. If these be comparable processes they probably indicate, merely, differences in the place where chondrification begins in the septum dividing the perilymphatic foramen.

Cavum vestibulare. It is apparent from the foregoing description of the vestibular cavity that the conditions present in the embryo of 23.1 mm. are far from what obtains in the adult mammaUan temporal bone. Only on the most general lines are the form and relations seen in .-the bony walled vestibular spaces of adult cat referable to the conditions of the cartilaginous jotic capsule. Such characteristics of adult structure as the sharp deUmitation of special recesses for the utricle and saccule, apparently are not even indicated in embryos considerably further developed than those of the stage modeled. Between these early stages and adult conditions many processes of formation are involved, about which almost nothing is known.

If the vestibular cavities of the cat embryo of the stages described present but few indications of their adult form, they


do show a certain agreement with conditions observed in the lower groups. The tendency to subdivision of the vestibule into anterior and posterior rooms is strongly suggestive of the conformation of the cavum vestibulare in Lacerta, as described by Gaupp. The contents of the anterior and posterior vestibular cavities in the cat appear to be comparable with the parts contained in the cavum vestibulare anterius and cavum vestibulare posterius of lizard. The anterior and posterior acustic foramina in lizard and cat are similarly related to the vestibular cavities. Furthermore the relative positions of the two spaces, one to the other, and to the otic capsule as a whole, are strikingly similar in the two forms. Such a comparison must accept of necessity the homology of the boundary structure between the vestibular cavities. In the cat this is the medial ward projecting ventral part of the massa angularis, the part designated crista intervestibularis. In the lizard it is the septum inter vestibulare. Gaupp describes the latter as a transverse vertical partition between the two vestibular cavities, presenting a lateral opening filled with membrane and a medial foramen for the utriculus. The crista intervestibularis is hkewise transverse in position, but it cannot be called a partition since it fails to extend entirely across, between the anterior and posterior cavities. While it possesses no opening, yet, by its failure to reach the medial vestibular wall, a space is left between the latter and the free medial edge of the crista, by which the anterior and posterior cavities are put into communication and which is occupied by the utriculus.

Lamina parietalis and tectum posterius. Decker ('83) found the parietal plates in cat embryos of 5.5 to 6.1 cm. bent inward, which is not the case in the younger embryos of the present study; rather, the dorsal, irregular margins are turned a little laterally. Decker's observation is interesting in proving that the parietal plate participates in the cranial roof, to only a sUght extent, as his figure shows, but, nevertheless, marking an unusual development of this part of the cranial wall for mammals. The lamina supracapsularis of Echidna, remarkable for its great breadth and its continuity with the ear capsule, is, in its anterior part, com


parable with the lamina parietahs of placental mammals, also to the hinder portion of the taenia marginalis of Lacerta (Gaupp, '08, b).

Chondrification of the parietal plate independently of the ear capsule was not observed in the series of Echidna studied by Gaupp, but it was noted that the plate was separated from the occipital pillar caudally and from the ala orbitalis rostrally by stretches of unchondrified tissue. Noordenbos is apparently the first to have observed the independent origin of the parietal plate and its secondary union with other parts of the skull. In Talpa of 11 mm. this cartilage was dorsad of the ear capsule and immediately in front of the tectum interoccipitale; its union with the otic capsule occurred very soon after its appearance. Evidently the parietal plate in Talpa does not unite posteriorly with the occipital arch, but with the tectum interoccipitale. The latter appears in the mole at the same stage as does the parietal plate, in the form of an independent piece arched over the cerebellar region. Noordenbos observed its union, first with the occipital arch, then considerably later with the parietal plate. On account of the primary union with the occipital arch Noordenbos is inclined to adopt the term 'tectum interoccipitale' as more correctly expressing its relations than the terms 'tectum synoticum' and 'tectum posterius.' Whether the tectum arises by paired anlagen in the mole could not be stated; its origin in the rabbit was observed by Noordenbos to be paired.

In regard to the origin of the parietal plate in cat, I am unable to present a conclusion, owing to lack of material at a critical stage in the development of this part. Apparently primary conditions siiTiilar to those described for Talpa obtain in Felis. A cartilage, which I have called lamina parietalis, arises by paired beginnings in the form of triangular plates, above the interval between the occipital arch and pars canalicularis. At a later stage the parietal plate has united with the lateral occipital arch and presents a prominent angle toward the mid-dorsal line, which is the beginning of the tectum posterius. At this stage (15 mm.) there is a faintly stained tract of cartilage in the region of the future commissura orbito-parietalis, but extending caudad into


the otic region. There remains, however, a wide gap, equal in extent to the dorsal margin of the ear capsule, between this tract of cartilage, which I have referred to as the beginning of the commissura orbitoparietalis, and the parietal plate. Now in the mole, Noordenbos found a cartilaginous plate immediately anterior to the tectum interoccipitale and identified it as the parietal plate. It united subsequently with the otic capsule, opposite the junction of the anterior and middle thirds of the anterior semicircular canal (synchondrosis parieto-canalicularis) , with the anterior margin and ventral end of the tectum interoccipitale, and also with the ala orbitalis. The last connection came about through the synchondrosis orbito-parietalis, by extension of the lateral hinder angle of the ala orbitahs.

In sections of cat embryos of 17 mm. a parietal plate of young cartilage extends from the occipital arch to the commissura orbito-parietalis. The arch of the tectum posterius is represented merely by a prong, springing on either side from the broad plate of cartilage forming the caudal end of the parietal plate, and which is united by stretches of young cartilage with the pars canalicularis and lateral occipital arch. The anterior end of the parietal plate, while approaching the otic capsule at the level of the middle of the anterior semicircular canal, is still not united with it, but is separated by a plane of mesenchyma. In later stages (24 to 30 mm.) van Wijhe preparations show that chondrification is less advanced, in the narrow part of the parietal plate, where the inconstant union with the pars canalicularis occurs (p. 310), and where union has occurred with the occipital arch and otic capsule. There seems to be no doubt of the cartilaginous plates, identified as the beginnings of the parietal plate, going to form also the tectum posterius (the arch is complete in the 23.1 mm. stage) and also the posterior part of the parietal plate. The relation of the conamissura orbito-parietalis to the otic capsule anteriorly (the so-called parieto-capsular commissure) is evidence of the probable homology of this cartilage with the parietal plate of Talpa, as described by Noordenbos. There is no evidence of another center of chondrification between the termination of the orbito-parietal commissure and the tectum pos


terius, and it is doubtful, notwithstanding the great space between these parts, if another center does develop and enter into the formation of the parietal plate. In favor of this assumption is the brief period between the stage when the orbito-parietal commissure and tectum posterius are separate (15 mm.) and the stage when they are joined (17 mm.). Apparently, in the cat, the parietal plate is formed by the coalescence of two cartilages arising independently, one, mostly anterior to the otic region, which gives rise also to the larger part of the orbito-parietal commissure and to the parieto-capsular commissure; the other dorsad of the interval between the lateral occipital arch and pars canalicularis which unites with these parts, forms the broad caudal portion of the parietal plate and also gives rise to the tectum posterius.

Facial and acustic nerves. The suprafacial commissure, forming the roof of the primary facial canal, separates the facial nerve from the ganglion semilunare. As the seventh nerve (including the pars intermedia Wrisbergii) leaver the canal, the geniculate ganglion is formed on its dorsal side. This is in contact with the ganghon of the trigeminus, both structures lying outside the plane of the fenestra sphenoparie talis. The lateral opening of this primary facial canal should be compared with the foramen faciale of reptiles. A foramen or canal, traversed by the facial nerve beyond the ganglion genicuh, is a new acquisition for mammals and not to be found in the reptilian cranium. Such is the stretch which Fischer ('01) has described in Talpa, roofed over by the "ganz dlinne Knorpelspange" (p. 504), separating the proper facial opening from the hiatus spurius. Also the foramen faciale externum of the tegmen tympani of the rabbit forms an acquisition to the primary facial canal, whose lateral opening in the tympanum is the apertura tympanica. This conception of the facial canal is partly in accord with that of Vrolik ('73). In the cat, at the stage modeled, the exit of the primaiy facial canal is at the level of the ganglion geniculi (position of the future bony hiatus canalis facialis) and outside the cavity of the chondrocranium. In the bony cranium of cat, and probably in later stages of the chondrocranium, the exit from the cranial cavity is by the apertura tympanica.


Regarding the course of the seventh nerve in the sulcus faciaUs, it is to be remembered that in adult Fehs domestica the second part of the facial nerve traverses an open groove in the medial tympanic wall (the rule in mammals), whose lateral boundary is the ossified processus paroticus (crista parotica of van Kampen, '04). Neither in the embryo nor in the adult does the free, margin of the parotic crest incline toward the vestibular wall in the formation of a canal. In Felis pf),rdus, however, the nerve does run in a closed canal (Denker, '99). The groove in the adult domestic cat begins anteriorly at the apertura tympanica and terminates posteriorly opposite the level of the tympanohyale.

Finally, reference should be made to the discovery by Spence ('90) in the adult and new born cat of a bony or cartilaginous support of the chorda tympani, projecting from the tympanic bone. Bondy ('07) has confirmed this observation, finding the process not only in cat but in a number of other mammals. No evidence of the support was found in the stages described here; its formation takes place according to Bondy, late in fetal life.

Acustic nerve. Retzius ('84) described the nervus acusticus of the cat as dividing into two or three chief branches, preferring two in his account. I found this mode of branching in the cat embryo of 23.1 mm.., and it may be remarked that, in our present state of knowledge of the distribution of the acustic nerve in the cat, the nomenclature of Retzius seems preferable to one which attempts to represent the origins of the nerve fibers. For, in the case of the ramus posterior, cells of the vestibular and spiral ganglia are intimately associated, and no safe conclusion on the origin of fibers of this ramus can be reached without further neurological investigation. The small twig from the anterior ramus to the sacculus appears to correspond with the ramulus maculae sacculi pars superior found by Voit ('07) in Lepus.

Orbito-temporal region

Hypophyseal cartilage. Chondrification of the base of the cranium in the hypophyseal region has been observed in several mammals to take place independently of the rest of the chondro


cranium. Parker (75) observed a secondary growth of cartilage beneath the pituitary body" in Sus. A sphenoidal cartilage, independent of the occipital skeleton of the ox, is described by Froriep ('86) as lying beneath the hypophysis. This observation is confirmed by Noordenbos ('05) who found, moreover, that the cartilage was paired. Noordenbos has discovered, by van Wijhe's method, the origin of the middle piece of the cranial base by the fusion of several small islands of cartilage in the crania of mole, rabbit and pig. For these pieces, which surround the stalk of the hypophysis, Noordenbos has proposed the name 'insulae polares.' Wincza ('96) noted the independence of the cartilaginous basi-sphenoid and alisphenoid (ala temporalis) in cat. As we have seen, the cranial floor, beneath the hypophysis, is first represented by a crescentic cartilage, which soon grows around the stalk of the hypophysis, probably completely surrounding it, although this was not actually observed. The formation of the sella turcica is brought about by the union of the hypophyseal cartilage anteriorly with the trabecular plate and posteriorly with the basal plate, the former contributing the tuberculum sellae, the latter the crista transversa which is the beginning of the dorsum sellae.

Fenestra hasi-cranialis posterior. The existence of an opening in the basis cranii between the anterior end of the basal plate and the hypophyseal cartilage is merely temporary in the cat. In embryos of 12 mm. the fenestra basi-cranialis posterior has no lateral limits, since the cochlear capsule is not yet joined with the basal plate. The anterior boundary is made by the hypophyseal cartilage, so that the fenestra lies, not within the basal (parachordal) plate, but anterior to it, as Noordenbos insists.

- Crista transversa and dorsum sellae. The crista transversa represents the anterior, dorsally turned, free edge of the basal plate. The upward bend of this margin just prior to the fusion of the basal plate and hypophyseal cartilage, can be seen in sagittal sections of embryos of 12 nm:i. and its identity with the crista transversa proved by sections of 15 mm. specimens in which the line of fusion of the two plates is still distinct. Noordenbos found that the dorsum sellae of Talpa is formed from the anterior mar


gin of the parachordal plate. In the cat, the dorsum sellae is also in part at least, derived from it, but probably the greater part is derived from the mesenchyma above the crista transversa and in front of the end of the notochord, in which the small cartilaginous nodule was found in the smaller embryos.

Let us consider briefly some observations regarding the dorsum sellae. I believe no comparative study of its formation has been made. This structure reaches a high development in man and apes. Fischer ('03) found in a Macacus embryo the upper border isolated from the rest of the dorsum sellae, lying as a transverse bar, and terminating laterally in the posterior clinoid processes. This author also points out that in the adult human cranium a groove or ridge stands between the dorsal border and the clivus and puts forth the suggestion that this structure, isolated in Macacus, is probably genetically foreign to the basal plate. Voit ('09) supports this view by the discovery in Lepus of a partial separation of the dorsal part of the saddle-back by a wide foramen from the rest of that structure. Faw^ cett ('10) found in human embryos of 19 and 21 mm. a rounded mass of cartilage, behind the pituitary body, connected with the clivus region by a fibrous bridge; he concluded that the dorsum sellae arises independently in man. In his study of the primordial cranium of Talpa, Fischer ('01) found the hypophyseal fessa to be a slight depression and noted the absence of a dorsum ephippii. Noordenbos ('05) speaks of a weakly developed dorsum sellae turcicae in mole embryos and states that it is a structure of the parachordal plate. This investigator observed that in embryos of 11 mm. the caudal end of the polar plate, lay somewhat beneath the cranial end of the parachordal plate. With the union of these parts the hypophyseal fossa is formed, limited posteriorly by the projecting anterior end of the parachordal plate.

Fischer's and Voit's observations point to the origin of the upper part of the dorsum sellae as possibly distinct from the basal plate in Macacus and Lepus, and, according to Fawcett, the origin of the dorsum sellae in man is from an independent center of chondrification. Noordenbos found the dorsum sellae of Talpa as a product of the parachordal (basal) plate. These, observa


tions suggest the possibility of the dorsum sellae of different species not being strictly homologous. Gaupp says ('00, p. 538) "Dass die hinten begrenzende Crista sellaris, wie sie bei dem Chondrocranium der Saurier (und auch bei dem der Vogel) vorkonunt, der Sattelsehne des Saugercranium entspricht, ist allgemein anerkannt." But does the crista sellaris of reptiles compare with the dorsum sellae of mammals? And are the latter strictly homologous in the different orders of mammals?

This is not the time for a full discussion of these questions which requires a larger basis of observations, but a few comments may be offered. Regarding the homology of the dorsum sellae among the mammals the following may be noted. The saddleback of the chondrocranimn. of Echidna (Gaupp '08), Talpa (Noordenbos, '05), Caluromys and Didelphis (my own observation) is very low. As already stated, Fischer denies its presence altogether in the Talpa embryo he studied. In contrast wdth the insignificant low ridge-like dorsum ephippii of these species is the high saddle-back of the chondrocrania of man, apes, rabbit and cat. In all these species evidence is at hand indicating the presence of an element in the dorsal part of the saddle-back, more or less independent of the base of that structure. In embryos of cat smaller than the stage modeled, the notochord terminates in a mass of mesenchjana which surmounts the crista transversa, whereas in the latter stage it ends in the perichondrium of a cartilaginous tubercle which rises from the middle of the transverse crest. This tubercle is developed from the mesenchymal mass, and may possibly have its beginning in the prochondral nucleus observed in the embryos of earUer stages (p. 327). In the cat embryo of the stage modeled, the dorsmn sellae is formed to some extent also from the up-turned edge of the parachordal plate (crista transversa). It would seem, in respect to the contrast between the caudal hmits of the hypophyseal fossa in the two groups here presented, that there is an element present in the one which is not found in the other, or that in one group a simple crista transversa forms the back of the pituitary fossa, whereas in the other a crista transversa plus an additional element enters into the construction of a dorsum sellae.


There is nothing in the model of the cat embryo to indicate the presence of post-chnoid processes, which are found in the bony cranium and are therefore formed later. These processes in connection with the processus inter clinoideae, Voit regards as vestiges of the primitive cranial wall in this region. Inter chnoid processes have often been observed in the adult cranium of man, and Fischer has noted them in the chondrocranium of Macacus.

Foramen hypophyseos. The foramen hypophyseos is converted into a canal by the growth in thickness of the floor of the sella turcica. In embryos of 17 mm. the stalk of Rathke's pouch is still intact as it hes in the canal. In the floor of the sella turcica of the stage modeled, the hypophyseal canal is present, but only vestiges of the stalk are apparent in it. Aral ('07) has described a bony walled canalis cranio-pharyngeus in the cat, containing a vein and a hypophysis accessoria cranio-pharyngei ; also an epithehal-Uned bhnd canal interpreted as a possible vestige of the stalk of Rathke's pouch. Voit has criticised Aral's homology of the cranio-pharnygeal canal and the hypophyseal foramen in the rabbit, asserting that the former is a secondary development, occurring in a position caudad of the location of the hypophyseal foramen. This criticism is not pertinent in the cat; the foramen and the canal have the same location.

Ala temporalis. Wincza ('96) described the boundary line between the ahsphenoid and basisphenoid (properly Ungula) in embryos of the cat. My observations are in accord with this description; a zone of perichondral tissue standing between a chondrified processus alaris of the basis cranii and a broad cartilaginous plate, ala temporaUs, in relation to the Gasserian ganglion. The complete independence of the alisphenoid in the cat led Wincza to investigate its relation to the cranium in other mammals with the following results : In the dog embryo, a separating zone was found between the cartilaginous alisphenoid and basisphenoid; in embryos of the polar bear complete separation of the two parts, with a small wedge-shaped cartilage in the cleft; in man, a joint between alisphenoid and basisphenoid, recalling the relation between the head of the femur and the acetabulum; in hedgehog embryos a boundary between the basi


and alisphenoid behind, both parts united in front in three of the stages studied. Wincza's observations have since been confirmed for man by Levi ('00), Gaupp ('02) and Fawcett ('10); for the dog by Ohnstead ('11). Noordenbos ('05), Voit ('09) and Fuchs ('10) have seen the separating zone in Lepus. Noordenbos also found the ala temporahs a free process in the pig and horse. A characteristic of the synchondrosis between ahsphenoid and basisphenoid of the cat described by Wincza, is its obUque course from behind, forward and medial ward. Whether the obUquity is a constant feature of the separating zone of the ala temporalis in other forms studied by Wincza is not specfically stated. Figure 8 of Wincza's paper shows a cleft separating the two parts in the polar bear, having the same direction as the boundary zone in the cat. Mackhn ('14) found in a 40 mm. human embryo that the connection between the lateral portion of the ala and the processus alaris took place between the ventral surface of the latter and the subjacent ala.

In contrast to the type of independent ala temporalis there has been observed another type, characterized by its continuity with the basis cranii. Wincza saw no trace of a boundary between basisphenoid and aUsphenoid in the chondrocranium of embryos of the horse, pig, sheep and calf. Noordenbos, however, as just stated, disagrees with the observation on the horse and pig; he finds the ala temporalis in mole to be a process of the side of the sella turcica. In Echidna, Gaupp identified the ala temporahs in the small continuous process which springs from the side of the sella turcica, laterad of the carotid foramen.

From these records we learn that, in a ^^ariety of mammals, the ala temporahs is more or less distinct from the rest of the cranium, being separated from a basal process by a stratum of some tissue other than cartilage, or even by a cleft of greater or less extent ; whereas in others the ala is a simple process continuous with the cranial base. Fm-thermore, Noordenbos, from his own observations, recognizes two types of embryonic origin of the ala tero.porahs in mammals; one type represented by the mole, in which the ala as a lateral process of the pole-plate; the 'other type, represented by the rabbit, wherein the ala arises


independently and unites secondarily with a process (processus alaris) of the margin of the sella turcica.

Of the many questions yet unanswered regarding the significance and relations of the ala temporalis, one only will be discussed here. The cat belongs to the type in which the temporal wing is more or less separate from the rest of the cranium, and in which the connection is an indirect one through a processus alaris. The successive stages of development of the ala temporaUs in the cat seem to throw some light on the nature of the differences between the temporal wing of the continuous type and that of the separate type. The following discussion deals with this question.

We may first compare the development of the ala temporalis in the two types, i.e., the one in which the temporal wing is continuous with the rest of the cranium and that in which it is more or less separate. For the first, Talpa may be chosen as an example; for the second, the cat. The following results have been brought out by Noordenbos in his study of the mole. In embryos of 12-13 mm. three lateral processes at the side of the pole-plate (basisphenoid) are described. With the posterior one of these processes (basicochlear synchondrosis) we are not here concerned. The "intermediate process, knee-formed, extends from the poleplate in front of, and then bends laterally around, the carotid artery passing backward to join the cochlear capsule. The anterior process ends free. In embryos of 14 to 17 mm. the latter has grown so as to come into relation with the Gasserian ganghon. In still older embryos (17-19 mm.) the anterior limb and knee of the middle process grow out in connection with the anterior process to reach the under surface of the semilunar ganglion and thus is formed the ala temporalis. "This little plate is connected with the lateral margin of the basisphenoid through the processus anterior and the anterior limb of the processus intermedius." Noordenbos states that the largest part of the ala temporaUs is contributed by the anterior lateral process of the pole-plate ; a small part arises from the knee of the middle lateral process. Union of the two elements of the ala occurs in embryos of 20 to 25 mm., but there remains for a long time, a sJnall


fissure-like opening in the root of the ala as the last vestige of the original gap between the anterior and middle lateral processes of the pole-plate. The middle process, the synchondrosis sphenocochlearis lateralis, is, in its caudal half, the part named by Fischer in a later stage the 'trabecula alacochlearis.' Noordenbos has shown, however, that the trabecula alacochlearis genetically has its anterior connection, not with the ala, but with the basisphenoid, and only secondarily unites with the hinder margin of the root of the ala temporalis.

Let us consider now, the development of the ala temporalis in the cat and compare it with that of the mole, taking as a point of departure the stage of 17 mm. In the cat there arises from the side of the sella turcica, a precartilaginous bar, which bends about the carotid artery, its anterior limb (processus alaris) crossing transversely in front of the vessel, its posterior Hmb (commissura ahcochlearis) extending backward upon the side of the artery and joining the cochlear capsule. Here, then, are elements present in both mole and cat which are in agreement in several important relations. In order further to identify this combined alar process and alicochlear commissm^e of cat with the middle process of mole, it should be recalled that it is derived from the coiTonissural element of the earUest stage (12 m.m.) the relations of which to the base of the cranium, to the cochlear capsule and to the' carotid artery are equivalent to the relations of the middle process of mole at its first appearance (12 to 13 mm.). Comparing, next, the relations of the middle process of mole and the precartilaginous derivative of the commissural element in cat, with parts lying laterally, there is present in each case an element, separate from the process under consideration, related to the semilunar ganglion. In Talpa this element, derived from the anterior process, grows out alongside the anterior limb of the middle process, from which, however, it is at first separate. It is noteworthy that the -former extends somewhat caudally as well as laterally, and, in conformity therewith, the space intervening between it and the middle process extends from behind, forward and medialward. In embryos of 20 to 25 mm. these two^ pieces of the ala temporalis unite, the anterior limb of the



middle process being thereby taken into the root of the ala. In cat embryos of 17 mm., the ala temporalis, the derivative of the mesenchymal alar element in relation to the Gasserian ganglion, stands lateral to and in front of the anterior limb of the commissm^al element from which, however, it is separated by a layer of mesenchyma (p. 329). Caudally the ala temporalis is separated from the posterior limb of the commissural element by the space which later results in the incisura lacera. The lamina presents then an oblique Une of junction with the anterior limb of the commissural element, extending from behind, forward and medially. Later stages (23.1 mm.) as already described, show along this Une a persistent sign of the separation of ala temporalis and the arch made by the processus alaris and alicochlear commissure, in the perichondral boundary zone first described by Wincza. The anterior process, which in mole forms the greater part of the ala temporalis, is entirely separated by an interval from the anterior hmb and knee of the middle process for a time; the alar element, which in the cat forms the entire ala temporalis, is separated from the commisvsural element by an interval (incisura lacera) only in its cauda.1 part. It will be recalled that Wincza found in the polar-bear the alisphenoid separated from the basi-sphenoid by a cleft which, as the figure shows, runs from behind forward and medial ward. Also the same author found in the hedgehog a limiting zone in the posterior part only of the ala temporalis.

Apparently the alar element of the cat and the anterior lateral process of the pole-plate of the mole are, in several respects, comparable. The one real difference between the alar element of the cat and the anterior process of the mole lies in the relation of these parts to the side of the sella turcica : in the mole the element in question is connected medially with the sella; in the cat it passes medially into the commissural element, without independent connection, if any at all, with -the side of the sella turcica. There is a reduction, if not entire absence, of a medial part of the alar element, next to the hypophyseal cartilage comparable with the root of the anterior process of mole.


The differences between the independent ala temporahs of the cat and the continuous ala of the mole are apparent rather than real. A very simple explanation removes these apparent differences. 1) The alar element of the cat is wanting in a medial part, comparable with the origin of the anterior process from the pole-plate in the mole. Excepting in the undifferentiated mesenchymal continuity of the alar element and anterior limb of the commis.sural element, there is no indication of a connection between it or its derivative, the ala temporalis, and the side of the sella turcica. 2) The processus alaris is compound in the mole, being derived from the root of the anterior process plus the anterior limb of the middle process; it is simple in the cat, being composed of the anterior limb of the commissural element only. The processus alaris of the cat is therefore comparable with only the caudal part of the ala temporalis of the mole, i.e., with that part derived from the anterior limb of the middle process. 3) The alar element of the cat and its derivative fails to unite with the commissural element and its derivatives, and there remains throughout the cartilaginous stage of the ala temporalis, a synchondrosis representing the limiting margins of the two elements. In the mole, partial fusion takes place between the comparable elements, but a cleft-like vestige remains locating the original separating space. Perhaps the conditions in the hedge-hog and bear described by Wincza may be similarly explained.

The conclusions reached therefore are : 1) the reduction or the absence of a medial part of the alar element of the cat, comparable with the origin of the anterior process from the sella turcica of mole; 2) the presence of a simple processus alaris in the cat, of a compound one in the mole ; 3) the persistence of the original boundary between the lateral and commissural elements in the cat and the obliteration of the limits between the comparable anterior and middle processes in the mole. It follows that the ala temporahs of mole and cat are in the main comparable. It is probable that other examples of the types of continuous and independent temporal wings may be similarly explained.


Whether the lamina ascendens of the cat should be regarded as an independent element in origin, or continuous with the basis cranii, is a question which could be answered either way from the evidence here presented, and would be purely a choice of interpretations. The whole ala temporalis is shadowed in mesenchyma, continuous with the basis cranii; in this adumbration condensations appear which are distinct from one another or confluent at the edges, as you choose to describe the conditions. Comparison of the lateral element of the cat with the anterior process of the mole inclines me to regard this element and its derivative, the ala temporalis, as not an independent element genetically. There appear to be two parts under consideration in comparisons of the ala temporalis of different mammals, (1) a part which is related chiefly to the carotid artery, represented by the middle process of the mole and the commissural process of the cat; (2) a part which is related to the semilunar ganglion and the three divisions of the fifth nerve. These two parts are typically both connected with the cranial base, the one archlike, the other process-like. Both are typically originally separate from one another, the process standing in front and at the side of the arch. In some forms the process unites with the arch completel}^ (mole), in others partially (hedge-hog), instill others not at all (polar-bear, cat). Finally, this conception of the comparison of the anterior process of mole and the ala temporalis of cat, supports the homology of the ala temporalis of placental maminials with the ala temporahs of Echidna.

Pterygoid. The pterygoid cartilage is developed relatively late, not until after bone has begun to be formed in its dorsal part. It is represented by a rather ill-defined mesenchymal condensation, even at the time when the ala temporalis is w^ell chondrified. Only its caudal part is cylindrical and thus bears some resemblance to the cartilage of the medial lamella of the pterygoid process in Talpa. It is unlike this process in not reaching forward to the ethmoidal region. How^ever, as pointed out, it develops in the same morphological plane as the palate bone, lies next the ductus nasopharyngeus, in front of the first pharyngeal pouch and is crossed dorsad by the Vidian nerve.


Jayne ('98) has described, in the skull of the adult cat, a pterygoid process presenting a well defined internal pterygoid plate or process and a variable external plate or process. Union of the originally separate pterygoid bones wdth the alisphenoid takes place very early, and sutures or lines indicating their boundaries cannot be found in the adult. The pterygoid elements are represented therefore, as processes of the aUsphenoids. Reighard ('02) describes an external pterygoid muscle, taking origin from the external pterygoid fossa whose surface includes the lateral aspect of the external pterygoid plate of Jayne, and an internal pterygoid muscle springing from the internal pterygoid fossa ; the latter is bounded in part by the medial surface of the external plate. The internal pterygoid plate terminates in a hamular process, related, in the usual manner, to the tendon of the tensor palati. The origin of these two bony processes was found in the present study in cat embryos of 7 cm., and 23.1 nam., the external process being an extension of the endochondral ossification of the pterygoid process of the ala temporahs, the internal process consisting of an ossification, at first in membrane and subsequently in cartilage, in connection with the separate pterygoid cartilage. The latter accords with an early stage of the human internal pterygoid plate as described by Fawcett ('10), both in its early ossification (it is the first part of the sphenoid to ossify in the cat) and in the ossific process, proceeding primarily in membrane and later in cartilage. The external plate is feebly developed in the cat, but its ossification in relation to that of the ala temporalis is nevertheless, in principle, the same as in man. It will be remembered that the Vidian nerve runs along the mesenchymal junction of the pterygoid cartilage and pterygoid process of the ala temporalis; now, although no suture or fine can be found in the adult skull indicative of the original limit of the pterygoid bone toward the alisphenoid, as Jayne has stated, yet the course of the bony walled Vidian canal of the adult can be taken as marking this boundary.

Carotid foramen. The epipteric cavity in the cat embryos is limited, toward the primary cranial ca\dty, by a membrane whose relations to the base of the skull are of considerable interest as


affecting the question of the nature of the region about the carotid artery. In Echidna the carotid enters directly the primary cranial cavity, but the position of the foramen through which it passes lies within the lateral confines (trabecula basis cranii) of the hypophyseal fossa and therefore may be compared with the entrance into the cranium of this vessel in Lacerta. In placental mammals the carotid foramen lies laterad of the sella turcica, a position which has been explained by two assumptions: (a) lateral migration of the vessel, (b) non-equivalency of the mammalian and reptihan carotid. In accordance with the theory of lateral migration, the carotid foramen is supposed to have moved outward across that part of the cranial floor equivalent to the trabecula; or both the trabecula and artery have moved lateral while still retaining their primitive relations to each other. The trabecula cranii of Lacerta is represented in mammals, according to Gaupp ('02), by the ahcochlear commissure. The unity of this commissure, including both the processus alaris and commissura aUcochlearis, its distinction from the ala temporalis, and its relation to the base of the cranium were recognized by Gaupp. In one place ('06), he says: "The processus alaris of man appears as a process only through the decadence of the cartilaginous bridge which closes laterally the carotid foramen. It belongs to the median cartilaginous mass in the base of the orbito-temporal region," Voit, who has advocated the view of non-equivalency of the internal carotid in mammal and reptile, locates the carotid foramen in Lepus laterad of the trabecular region. He concludes that the commissura alicochlearis plus the processus alaris alae temporahs should not be compared with some part of the trabecula, but rather the processus alaris should be referred to the processus basipterigoideus of Lacerta and ala temporalis of Echidna. The commissura alicochlearis is regarded by Voit as a new structure, in mammals, in the floor of the epipteric cavity, a continuation of the floor of the cavum supracochleare. Fuchs ('10) also, compares the medial part (root) of the ala temporalis with the processus basipterigoideus of reptiles. In accordance with Voit's interpretation, the carotid artery first enters the epipteric cave, then, after traversing the medial limiting membrane


of the cave, comes into the primary cranial ca^dty. Recently, De Burlet ('13) has brought forward evidence of the carotid in Phocaena traversing the trabecula or its equivalent. In the porpoise the carotid passes directly into the primary cranial cavity. As we have seen, the luniting membrane in the cat is fixed to the basis cranii, neither to one or the other side of the carotid foramen, but broadly over the region where the vessel enters. It is attached both to the alicochlear commissure and to the lateral margin of the sella turcica. The former does not, in the cat, enter into the floor of the ca\^um epiptericum, nor does the carotid artery pass directly into the primary cranial cavity. The vessel first traverses that part of the membrana Umitans which is fixed to the cranial floor, before passing into the original cavity of the skull. As to the homology of the commissura alicochlearis, it is evident that if the processus alaris is not included as a part of it, the diflficulty of comparing the commissure with the trabecula cranii is much increased ; if the alar process be recognized as the continuation of the ahcochlear conmaissure forward, in continuity \\ith the median basal cartilage of the orbito-temporal region, the comparison is far less difficult. Reviewing the conclusions stated above, we may note, first, that the name commissura alicochlearis is a misnomer, since it does not express the true relation of the commissura, but implies a connection \\ith. the ala temporalis which does not exist in the mammals so far studied. Secondly the studies of Noordenbos on the mole and my own on the cat show that the so-called processus alaris may be simple or compound in different animals, but that it is primarily a part of the commissura ahcochlearis (its anterior end in fact), continuing into the median basal cartilage; relation of the processus alaris to the ala temporalis is secondary and obtains in those types (mole) where the ala springs from the basal cartilage; its root, at first separate from, later becomes fused with, the processus alaris to produce the compound form of alar process. When the ala temporahs is independent of the basal cartilage (cat) the former has nothing to do with the processus alaris, which then is the simple extension forward of the commissura ahcochlearis into the sella turcica.


A word may be said regarding the membranous structure here called septum transversum. The development, in its basal part, of a marked thickening which extends from the crista transversa to the parietal plate, together with the presence in it of a cartilaginous bar over the semilunar ganglion are suggestive of structures in the reptilia, such as Shino ('14) has described in the crocodile and compared with the pila prootica of Lacerta. A cartilage over the semilunar ganglion in Lepus recorded by Voit was interpreted as a vestige of the primary cranial wall; in the cat, a cartilaginous mass, having the same position and relations, was found within the septum transversum.

Ala orhitalis. The occurrence generally, in a wide range of types, of the independent origin of the ala orbitahs has been emphasized by Noordenbos. This phenomenon,' as shown above, is also characteristic of the cranial development of the cat. The idea of the association of the ala orbitalis, primarily with the optic nerve and eye-ball, seems to be well founded and receives further support from the early form and relations of this cartilage in the cat. The relations of the origins of the ocular muscles may be interpreted as further indicating a close connection between the orbital wing and the organ of sight. The definitive origins of these muscles are in the main related to the optic foramen, about as they are in the chondrocranium, but some shifting has evidently taken place as comparison of the two states shows (Wilder and Gage '86; Corning '02). Four definite separate spots in the chondrocranimn are occupied by, and one of them apparently specially adapted to, the attachment of the muscles of the orbit. How these attachments compare with those in the mammaUan types represented in the literature of the chondrocranium I cannot say, since this question does not seem to have received attention. Does the superior obhque constantly spring from the preoptic root; does the orbital process of the metoptic root function generally as a point of attachment of a definite group of eye-muscles; is the side of the basis cranii in the orbital fissure a special locus of origin of another group, and is the origin of the inferior oblique constantly the planum antorbitale? Regarding the metoptic root, it has been noted that, in contrast to the


straight direction of the preoptic, it presents a marked bend, convex caudally. This curve is apparent to a slight extent in the bony cranium. Judging from figures of the chondrocranium of several manmaals, the bend in question seems to be a characteristic of this cartilaginous rod. In cat the knuckle of the curve apparently mark the place of union of the metoptic process of the trabecular plate and the corresponding process of the ala orbitahs. Opposite this spot the oculo-motor nerve leaves the cranium by the orbital fissure. What significance there is in the form and relations of the metoptic root, which are early estabUshed and permanently retained, must await future inquiry.

Ethmoidal region

Region of the olfactory fenestra. The floor of the condrocranium of mammals, between the levels of the preoptic root of the ala orbitahs and the fenestra olfactoria, is made by the lamina infracribosa, which is the roof of the posterior cupola of the nasal capsule and an extension of the planum antorbitale. The lateral brancl^ of the naso-cihary nerve runs upon this lamina on its way toward the interior of the nasal capsule, having come through the spheno-ethmoidal fenestra from the orbit. The lamina cribrosa, a structure peculiar to mammals, develops over the fenestra olfactoria. Now, in reptiles the ethmoidal nerve and the • whole of the posterior cupola, with the olfactory fenestra, are extracranial. A study of these contrasting conditions of the mammalian and reptilian ethmoidal skeletons has led- Gaupp to the conclusion that a new region, named by him recessus supracribrosus, has been added to the mammalian brain case, and that this region is comparable, in reptiles, with the extracranial parts about the olfactory fenestra. In the cat, the form and relations of the posterior cupola and olfactory fenestra are essentially as in other mammals. The fenestra olfactoria opens directly into^ the chondrocranial cavity. As aheady stated, the lamina cribrosa was not present in the stages of the cat studied, consequently no data have been presented bearing on the question of its position relative to the walls of the chondrocranium. The


lamina infracribrosa is a part of the antorbital division of the nasal capsule, entering into the floor of the cranium. The course of the ethmoidal nerve is, strictly speaking, outside the chondrocranial cavity, since the layer of mesenchyma (probably in part giving rise to dura), which covers the lamina infracribrosa, stands between it and the cerebral cavity. In the cat, in the stages studied, there is no overhanging anterior wall of the chondrocranium such as is shown in the model of Echidna; a flat anterior margin of the fenestra olfactoria marks the anterior limit of the cranial cavity. Consequently, the term 'recessus' supracribrosus is hardly appropriate for the cat; the term 'supracribrous region' is preferable. From these data, it seems that the ethmoidal region of the cat, like the orbito-temporal and otic regions, gives evidence in support of Gaupp's assertion of the nonequivalency of reptilian and mammalian crania, and also, by its relations to the chondrocranial cavity, indicates that the latter is enlarged over that of reptiles by the acquisition of a region wholly extracranial in that class. A factor having a large share in bringing about this mammalian characteristic is the backward growth of the nasal cavity (Weber, '04; Gaupp, '08) whereby the interorbital septum is encroached upon by the posterior cupola of the ethmoidal skeleton. That this occurs to a slight extent in the ontogeny of the cat was indicated by the difference in position of the posterior cupola relative to the preoptic process in earlier and later embryonic stages.

In Lepus, according to Voit, the foramina of the lamina cribrosa are separated into two groups by a crista intercribrosa, olfactory filaments from the recessus lateralis traversing the foramina of the antero-lateral group, those from the recessus posterior (ethmoturbinal region) passing through the holes of the posteromedial group. Voit also found the- dorsal end of ethmoturbinal I continuous with the crista intercribrosa. Continuity between these two parts is present in Caluromys philander, as observed by Dr. Denison in ' this laboratory. In the cat, at the stage modeled, the dorsal end of ethmoturbinal I is very prominent and stands between the group of olfactory filaments coming from the epithelium of the recessus lateralis and those from the ethmo


turbinal region. The latter group is further subdivided by the dorsal end of ethmoturbinal II into two bundles coining, one from the area in front of, the other from the area behind this turbinal process.

The floor of the nose and Jacohson's cartilage. In cat embryos of the stage modeled, the secondary palate is nearly complete, forming a floor for the nasal cavities. The latter are separated from each other by the septum nasi, excepting for a stretch posteriorly, where the ventral margin of the nasal partition is free from and above the level of the palate. Between the maxilloturbinal dorsally and the nasal floor is a broad groove of the lateral nasal wall, which extends from before backward to the choanae. This gi'oove, the inferior meatus of adult anatomy, the lower furrow of Legal ('83) or the choanal passage of Fleischmann and Beeker ('03), besides differing from the other nasal passages by its origin from the primitive mouth, is specialized through its relation to the naso-lacrimal duct and Jacobson's organ. A striking feature of this ventral part of the cavum nasi is the variabihty among mammals of its cartilaginous walls and floor. This characteristic stands in marked contrast to the usual condition of stabihty of the dorsal part of the ethmoidal skeleton (Spurgat, '96).

The peculiar characteristics of the nasal floor cartilages in the. cat have been described by several investigators. Our observations must be regarded as confirmatory of Broom's ('96) description and of Zuckerkandl's ('08) more recent representation of the paraseptal and related cartilages. Both of these authors found an inner and outer division or process of the nasal floor cartilage (lamina transversalis anterior) and a cartilage related to the nasopalatine duct.

The significance and homology of these parts of the ethmoidal skeleton have been much discussed. Division of the lamina transversaUs anterior, into medial and lateral cartilages (inner and outer parts of the nasal floor cartilages) embracing the nasopalatine duct, has been observed by Broom ('96) to occur with considerable constancy in the mammaUan series. The outer portion of the nasal floor cartilage stands lateral to the nasopalatine


canal. In Ornithorhynchus and Echidna this cartilage expands at its caudal end and unites with its fellow across the midline to form a broad plate beneath the organ of Jacobson and its cartilaginous capsule. The conditions in the MarsupiaUa were traceable to those of the monotreme type. In Lepus, as representing the rodents, Broom found an independent cartilage supporting the outer side of the nasopalatine canal, which was homologized with the outer nasal floor cartilage of the Monotremes, and was regarded as a much modified form, compared with the simple higher eutherian type. Voit has also described, under the name cartilago nasopalatina, the supporting cartilage for the nasopalatine duct in the rabbit, and, with Gaupp, compares it with the processus palatinus (outer nasal floor cartilage of Broom) of Echidna.

This comparison is founded by Voit mainly on the fact that in certain mammals (horse, pig and sheep) the processus palatinus (cartilago basalis laterahs as described by Spurgat, '96) is united with the lamina transversalis anterior (processus septi cartilaginei lateralis ventrahs) in the same relationship as is the case with the processus palatinus of Echidna; and fiuther, that in at least one form, Vesperugo noctula, described by Grosser ('02) the 'cartilagines posteriores laterales' (cartilagines nasopalatinae of Voit) are considerably broadened posteriorly and fused across the median line, as is true for the palatal processes of Echidna.

Two conditions, presented by the cat, seem to indicate that the significance of the nasopalatine cartilage and the processus posterior laterahs of the lamina transversahs, is not fully brought out by the above comparisons. These conditions are (a) the presence, at the same time, of two separate cartilages, one in relation to the nasopalatine duct (nasopalatine cartilage) the other in the floor of the nose next to the entrance of the nasopalatine duct (posterior lateral process of the transverse lamina) ; and (b) the independent origin of the former. Furthermore, in regard to the first condition, it seems evident, from Grosser's description of Vesperugo, that both paired nasopalatine cartilages (cartilago ductus incisivi) and posterior lateral cartilages are present and separate, the one pair from the other.


A cartilage, having the position and relations to the nasopalatine duct and to the anterior transverse lamina, such as presented by the posterior lateral cartilage of Vesperugo and the lateral process of the lamina transversalis anterior of FeUs, is of frequent occurrence in the mammalian series. It stands as a support to the floor of the nose at the outer side of the entrance to the nasopalatine duct and probably possesses, as Grosser says, a great significance as a supplementary structure to the palate. It is found in forms in which a nasopalatine cartilage (as here defined) does not exist, notably in mammals having a very short nasopalatine duct (Echidna) and also in forms where the nasopalatine cartilage is present, in which cases it may be separate from or united with the latter.

In accordance with this view, the processus palatinus of Echidna would be comparable with the posterior lateral cartilage of Vesperugo, the lateral process of the lamina transversalis anterior of Felis, and not with the cartilago ductus incisivi of these forms. The primary relation of the nasopalatine cartilage (cartilago ductus incisivi), on the other hand, appears to be directly with the nasopalatine duct, as in the cat. There is some evidence in the series of mammals of its correlation with the presence of a long nasopalatine canal. Absence of a nasopalatine cartilage in the lower mammals and its independent origin and late appearance in dsvelopment in the cat may be interpreted as indicating that it is phylogenetically a recent acquisition to the palatal skeleton, and also that its connection with the lateral posterior cartilage may be regarded as secondary.

Jacobson's cartilage, relatively short, arises independently, as recorded by Schwink ('88) and as observed in the present study, and remains ununited with other parts of the chondrocranium in embryos of the cat up to 23.1 mm. In later stages, as Zuckerkandl has observed, this cartilage has elongated and taken on connections with the lamina transversalis posterior, thi'ough the development of a slender paraseptal rod. Zuckerkandl believed that Jacobson's cartilage represents the anterior half of a cartilaginous ridge which is derived from the primitive floor of the nose. This is in accord with Seydel's ('99) conception of the


paranasal cartilage forming a part of the primitive nasal floor which, by reduction, came eventually to be represented by more or less separate parts (Jacobson's cartilage) in the mammals. In the light of this interpretation, the slender cartilaginous paraseptal strip which appears late in the development of the cat, and unites Jacobson's cartilage with the posterior transverse lamina, may be regarded as evidence of a partial restoration of the primitive floor.

Tectum nasi. Noordenbos has observed the beginning of the tectum nasi in mole embryos of 9 mm. as bilateral divisions of the anterior margin of the trabecular plate, forming curved lamellae. In the dorsal middle line a shallow groove stands between the two halves of the tectum. In the cat the nasal tectum arises apparently similarly in connection with the nasal septum, as the parieto-tectal cartilages, although, as already mentioned, some evidence of the independent origin of these cartilages was presented. The groove between them, dorsad of the septum, is the beginning of the sulcus supraseptahs. In the cat, the parieto-tectal cartilages gi-ow back, along the dorsal margin of the septal cartilage, as far as the fenestra olfactoria, becoming gradually narrower from side to side and assuming the form of triangular plates. These cartilages contribute the entire roof of the pre cerebral division of the nose and the side wall anterior to the recessus lateralis. What the relations of the parieto-tectal are to the development of the lamina transversalis anterior were not determined. Eventually a complete zona annularis is established behind the level of the fenestra narina, the roof being derived from the parieto-tectal cartilage, the floor from the lamina transversalis anterior, the lateral wall from the union of these two parts. The relations of the lateral margin of the parietotectal with the anterior free margin of the paranasal cartilage are at first similar to those in Talpa, as observed by Noordenbos. This cartilage does not seem to be concerned in the formation of the maxilloturbinal, although in the stage of 23.1 mm. this process is connected with the anterior part of the lateral nasal wall through the lamina supraconchalis. Within the nasal cavity the atrioturbinal appears as the inrolled ventral margin of the parietotectal cartilage.


Paranasal cartilage. Independent origin of a plate of cartilage lying in the anterior part of the lateral wall of the nose has been observed by Noordenbos in embryos of Talpa of 9 mm. length; also in the embryos of calf and pig. Following Mihalkovics, the cartilage was called cartilago paranasalis. In Talpa, the paranasal cartilage and the tectmn nasi are separated by a fissm-e, bounded by the anterior border of the former and the free margin of the latter. In the cat, also, an independent paranasal cartilage occurs which bears similar relations to the parietotectal cartilage. It is directly over that diverticulum of the cavum nasi which later is included in the recessus lateralis of the ethmoidal skeleton. The little cartilaginous plate, bent upon the convex bulging of the nasal sac, presents free margins: one toward the side of the olfactory lobe; one backward toward the eye; a ventral margin, and a long anterior side, opposite the free, oblique edge of the parieto-tectal cartilage ; relations comparable with those observed in the mole.

Great interest attaches to the ventral and anterior margins of the paranasal cartilage. The latter overlaps the parieto-tectal, and, whereas in younger embryos it is separated from that cartilage by a fissur'e, it is later fused with it. The result of this fusion is a curved, intranasal ridge, the crista semicircularis, perforated by the foramen epiphaniale, which is the remains of the original fissure. Thus is explained the position and extent, quite to the olfactory fenestra, of the semicircular crest. It follows from its mode of formation that, in later stages, the crista semicircularis may be taken as the boundary between two originally distinct territories of the nasal wall. During the period occupied by the process of fusion of the incurved anterior margin of the paranasal and the hinder margin of the parieto-tectal cartilage, and for some time thereafter, an interval of variable extent (derived from the space of the original separating fissure) is enclosed within the encompassing cartilaginous margins; that is, for a time the crista semicircularis presents double walls with an intervening space; this space is crossed by the lateral nasal nerve at the level of the future epiphanial foramen.


Now in the nasal capsule of Lacerta, at the place where the zona annularis joins posteriorly the broader, bulging portion of the side wall, a conchal fold is present, projecting into the nasal cavity; within the aditus conchae, or entrance to the conchal fold is the foramen for the lateral nasal nerve. The fissure-like aditus extends dorso-ventrad, its convex lateral margin overhanging the medial margin made by the zona annularis. Gaupp says of it ('00, p. 484) : An der Muscheleinstlilpung slnd zwei Lamellen, eine mediale und eine laterale, zu unterscheiden. Die mediale Lamelle der Concha hegt in der direkten Fortsetzung der Seitenwand der vorderen Nasenkapselhalfte, sie endet ventral mit

freiem Rande Die laterale Lamelle der Concha

geht ventralwarts in den Boden des Recessus extraconchalis tiber.

The aditus, in Lacerta, lodges the external nasal gland; its duct passes forward across the outer side of the zona annularis to enter (he nasal mucosa through a gap in the side wall of the nasal capsule, behind the processus alaris superior. Finally, the concha itself forms a ventral and anterior partial partition between the general space of the nasal capsule and a lateral diverticulum which Gaupp has named the recessus extraconchahs. The incomplete floor of the recessus is made by the ventral prolongation of the lateral lamella of the concha.

It seems evident that in the stage of development of the ethmoidal skeleton in cat when fusion is occurring between the parieto-tectal and paranasal cartilages, certain conditions are presented which reflect the structure of the reptihan nasal capsule, as shown by .the embryo of Lacerta. The general direction of the space between the parieto-tectal and paranasal cartilages on the one hand, and the aditus conchae on the other, is the same, and its boundaries are made by lamellae similarly placed in respect to each other; it is traversed by the lateral nasal nerve in both forms. The position of the external nasal gland in the aditus conchae of Lacerta and of the externally coursing duct is, apparently, impossible to reconcile, on our present knowledge, with the lateral nasal gland of Fehs and the intranasal course of its duct, if these glands be comparable at


all. Peter ('06) seems to regard the mammalian gland as a new acquisition. It will be remembered, however, that the lateral nasal gland in the cat embryo of 23.1 mm. is located just in front of the base of the crista semicircularis.

Next, we may attempt to compare that part of the nasal wall (paranasal cartilage), which stands just behind the fissure in the cat, and the aditus conchae in Uzard. We have seen that the commissura spheno-ethmoidahs unites, in cat embryos of 17 mm., with the dorsal margin of the paranasal cartilage, where the latter bounds the olfactory fenestra, thus presenting relations which recall those between the posterior half of the .side wall of the nasal capsule and the cartilago spheno-ethmoidalis of Lacerta. For, at the junction of these parts, in the cranium of the lizard, the lateral wall hmits by its upper margin the olfactory fenestra while, medially, it passes into the posterior wall, the planum antorbitale, as is also true for the cat. It has been noticed that the ventral and anterior margins of the paranasal, are continuous inrolled edges, which form the cartilaginous maxillo-turbinal and, with the free posterior margin of the parieto-tectal cartilage, form the beginning of the crista semicircularis. In Lacerta, the part of the side wall in question goes over anteriorly into the lateral overhanging plate of the bilaminar concha, the medial plate of which is the direct continuation of the side wall of the anterior half of the nasal capsule. Furthermore, the lateral plate passes ventrally into the floor of the recessus extraconchalis. Of great interest is the resemblance between the concha and floor of the recessus extraconchalis of Lacerta on the one hand and the crista semicircularis and maxilloturbinal of the cat on the other.

This comparison brings out a certain degree of similarity of form and relation between the skeletal parts of the region of the recessus extraconchalis of Lacerta (embryo of 31 mm. total length) and of the recessus lateralis of the cat (embryos of 1720 mm.). An opinion as to a possible homology between these parts should not be ventured on a comparison limited to the data here presented. Greater knowledge of the development of the ethmoidal skeleton in the hzard is required, and additional



data on the history of the paranasal cartilage in mammals is essential.

As to the origin of the lamina supraconchalis and the broad base of the naso-tm-binal, unfortunately little evidence was found. The region of the side wall occupied by these parts corresponds to that between the antero-inferior quarter of the paranasal and the adjacent region of the parieto-tectal cartilage. I cannot say to what extent, if at all, these two cartilages enter into the lamina supraconchalis and naso-turbinal.

One other matter should be mentioned with which, possibly, the paranasal cartilage may be related — the origin of the inferior oblique muscle. In the adult cat, this muscle arises from the orbital surface of the maxilla behind the lacrimal bone (Reighard, Jayne, Corning). In the embryo of 23.1 mm. its origin is from a prominent angle of the lateral nasal wall, at the junction of the paries nasi and plamun antorbitale. In the 17 m.m. embryo this angle corresponds with the ventral inferior corner of the paranasal plate, overlapping the anterior margin of the antorbital plate. Now, although no posterior maxillary process is developed in cat in the stages studied, yet the angle here referred to possibly represents this process. I have not noted in the literature the origin of the inferior obhque from the chondrocraniimi in mammals. Corning ('02), quoting Weber on its origin in Lacerta viridis, says (p. 122) : Der M. obliquus inferior 'entspringt von der Cartilago ethmoidalis, wo diese sich an den Knochen der vorderen Orbital wand ansetzt.' "

Lamina antorhiialis. Noordenbos is apparently the only one who has observed in mammals the isolated origin of the cartilage forming the posterior part of the nasal capsule. This element of the ethmoidal skeleton, found in mole embryos of 10.5 mm. and 11 mm., finds its parallel in the cat, where, as stated, a separate plate of cartilage forms the planum antorbitale and cupola posterior. Union between the antorbital plate and the paranasal in the cat must take place rapidly, as no trace of an interval between the two parts is to be seen in embryos of 20 mm. Complete union, that is, fusion, with the septal cartilage is not present in the embryo of the stage modeled. Of much in


terest is the relation of the antorbital plate to the ethmoturbinal bodies. A well developed, epithehal ethmotui'binal fold can be seen behind the recessus laterahs in embryos of 15 to 17 mm. We have seen that the posterior margin of the paranasal cartilage overlaps the anterior edge of the antorbital plate, and that the latter projects forward and medial ward into the base of the epithehal fold of ethmoturbinal I. Sections show that the anterior, free cartilaginous extremity of this ethmoturbinal, in embryos of 23.1 mm., is formed, in connection with mesenchyma, within the fold of epithelium and independently of the wall of the nasal capsule. The mesenchyma can be followed, however, backward to the base of the fold and to the ridge of cartilage derived from the antorbital plate of earher stages, which bounds posteriorly the opening into the recessus laterahs.

The development of the ethmoidal skeleton in the cat indicates a certain degree of correlation between its cartilaginous components and regions of the nose more or less sharply defined by particular functions. The dorsal, larger, and more completely walled division includes the olfactory and paranasal territories and the atrimn opening toward them; whereas the ventral, smaller and imperfectly formed division stands in relation to the respiratory air passage, nasopalatine and lacrimal ducts. Even the original components of the dorsal division present evidence of specific adaptation to functional territories. Thus, the antorbital plate forms the wall of the ethmoturbinal territory of distribution of the olfactory nerve, the paranasal cartilage incloses the diverticulum of the lateral recess, later the frontal sinus; and the parieto-tectal develops into the wall of the atrimn. I do not wish to imply that the evidence at hand justifies the conclusion of rigorous specificity of these components, it certainly does not; but facts are sufficient to indicate that each of these original cartilages is, in the main, developed in relation to recognized functional territories of the nose.



1. The basal plate of the occipital region in the cat is derived from a pair of parachordal cartilages and from two, probably three, hypochordal commissures or arches. The parachordals are continuous with the lateral occipital arches and, topographically, are to be compared with the lateral masses of the atlas. Chondrification of the perinotochordal sheath in the occipital region forms the apex of the dens epistrophei, which, in certain respects, is comparable with a vertebral centrum.

2. The observations on the developmental processes of the occipital region in the cat indicate that, in principle, they are comparable with those of an atypical vertebra of the atlas kind. The basal plate of the occipital region falls into the category of arch structures, not centra.

3. In the atlas, the beginning of an atlantal foramen is very early indicated by a notch, traversed by the first spinal nerve, the sides of which are formed by the neural arch and a process of the lateral mass. The possible significance of this atlantal process was discussed.

4. The occipital condyles are located primarily on that part of the basal plate derived from the parachordal cartilages and, in accordance with the interpretation of the form of the occipital element, belong to arch structures and not to centra.

5. The position of the lateral occipital arches and the plane of the foramen magnum, parallel with the floor of the nose in embryos of 23.1 mm., is correlated with flexures, in a sagittal plane, of the cranium as a whole.

6. The fissures around the otic capsule, between it and the basal plate and walls of the chondrocranium in the 23.1 mm. embryo, are vestiges of the spaces primarily existing between these independently arising cranial elements.

7. The incisura occipitahs posterior of the cat embryo is intermediate in relative size between that of the reptile (Lacerta) and that of man. There is apparently a tendency for this space to increase in area from lower to higher forms.


8. The upright position of the axis of the otic capsules in cat embryos of the stage modeled is exceptional for mammals. There is evidence in the literature that the obliquity of the otic axis characteristic of the mammahan chondrocranium is attained at somewhat different periods of development in different species.

9. The origin of the cartilaginous pars canahcularis of the otic capsule is independent of other parts of the cranium. The pars cochlearis arises in connection with the pars canahcularis and the suprafacial commissure. The latter is continuous with the orbito-parietal commissure, and is therefore, in part at least, to be regarded as a parietal structure. The cochlear capsule forms independently of the basal plate, which in the cat is reduced to a narrow bar in the otic region. The observations on the relationship of the cochlear capsule to the basal plate and suprafacial commissure are interpreted to support Gaupp's theory on the development of the cochlear capsule and reformation of the cranial base.

10. A single acustic fissure, transmitting the two divisions of the acustic nerve, is situated at the bottom of a shallow internal acustic meatus. The latter results, principally, from the elevations of the suprafacial commissure and commissura basivestibularis, these forming its anterior and posterior walls.

11. The fenestra cochleae and aquaeductus cochleae result from the division of the foramen perilymphaticum by a process extending from the cochlear capsule.

12. The cavum vestibulare of the cartilaginous otic capsule of the cat is comparable with that of the reptile (Lacerta) in possessing two subdivisions, a separating partition and similar relations to parts of the membranous labyrinth.

13. The lamina parietalis seems to be derived from two independently arising centers of chondrification, an anterior one, which also constitutes the orbito-parietal commissure and unites with the otic capsule in the parietocapsular commissure ; and a posterior center which unites, first, with the summit of the lateral occipital arch and, next, with the pars canahcularis. From the posterior center (called parietal plate in the descrip


tion) the tectum posterius apparently arises as a process. The definitive parietal plate seems to be the result of the fusion of the anterior and posterior elements.

14. The facial nerve makes its exit from the chondrocranium by the primary facial foramen, at which spot is located the geniculate ganglion and the origin of the great superficial petrosal nerve. The primary facial foramen is to be regarded as comparable with the foramen faciale of Lacerta.

15. The floor of the hypophyseal fossa is formed from a single crescentric cartilage arising independently.

16. In the cat a small space, filled with mesenchyma, existing for a brief period, between the anterior end of the basal plate and the hypophyseal cartilage, is identified as a fenestra basicraniahs posterior.

17. The dorsum sellae in cat is formed from the crista transversa and, in addition, by chondrification of the mesenchyma dorsad of it, probably by a separate center. There is evidence in the literature that the dorsum sellae of the lower mammals is a derivative of the crista transversa alone, while that of the higher mamjaals includes an additional element.

18. The development of the ala temporalis of the cat, representing the type of discontinuous ala, is comparable with that of mole, representing the type of ala temporalis continuous with the sella turcica. The difference between these forms of ala is an apparent one, resulting from failure of the root of the temporal wing of the cat to form in cartilage. The synchondrosis between the ala temporahs and alicochlear commissure and the incisura lacera of cat, the foramen or fissure between ala and conamissure in other forms, indicate the plane of junction between these originally independent and different parts. The ala temporahs of cat is comparable with the ala temporahs of Echidna.

19. A pterygoid element appears in the cat as a condensation of mesenchyma, in which ossification is occiu-ring at the stage of the 23.1 mm., embryo. Cartilage is subsequently developed, in which ossification proceeds. There is evidence of two ossific centers, and from these the medial pterygoid lamella and hamular process are formed.


20. A ca^iiin epiptericum and caviun supracochleare are recognized in the chondrocranium of cat. The names 'membrana hmitans' and 'septum transversum' are proposed for the mesenchymal • sheets connected with the walls of the orbito-temporal region. In the cat the internal carotid artery enters the membrana limitans, by the carotid foramen, and secondarily into the primitive cranial cavity. A bar of cartilage in the septum transversmii, lying upon the semilunar ganglion, is probably a vestige of the primitive chondrocranial wall.

21. The ala orbitaUs arises independently in the cat and unites secondarily ^^dth the paranasal cartilage, orbitoparietal commissure, and trabecular plate. It is apparently primarily adapted to the eyeball and its adnexa.

22. A region, extracranial in reptiles, is represented within the chondrocranimn of cat as the supracribrous region of the ethmoidal skeleton.

23. The structm-e of the anterior part of the floor of the nose is in agi'eement with the descriptions of Broom and Zuckerka-ndl. The nasopalatine cartilage of mammals is interpreted as a structure originally distinct from the lateral posterior cartilage of the nasal floor, and its connection with the latter secondary.

24. Jacobson's cartilage in the cat is the anterior, well developed, portion of a paraseptal cartilage, connected posteriorly with the lamina trans versaUs posterior.

25. The cartilaginoUvS ethmoidal skeleton of the cat is separated on developmental grounds and physiological relations, into a small ventral part and a large dorsal part. The former presents incomplete walls about the inferior meatus; the latter rather continuous walls about the paranasal sinuses and olfactory region of the nose.

26. The dorsal part of the ethmoidal skeleton in embryos to the stage of 30 mm. is formed from the septal cartilage, parieto-tectal cartilages, paranasal cartilages and the laininae antorbitales.

27. The parieto-tectal cartilages, paired, spring from the septal cartilage. They contribute the roof of the precerebral division of the ethmoidal skeleton, the side wall of the atrial region,


participate in forming the crista senaicircularis and probably, to some extent, give origin to the alar cartilages.

28. The paranasal cartilages, paired, arise independently and are primarily adapted to the diverticula of the recessus laterales. They participate with the parieto-tectal cartilages in the formation of the semicircular crests and in the origin of the maxilloturbinal processes. From this cartilage the inferior obhque muscle of the eye arises.

29. The foramen epiphaniale is the vestige of the original space between the parieto-tectal and paranasal cartilages.

30. The lamina antorbitahs, paired, arises independently, and unites secondarily with the paranasal and septal cartilages. From it are formed the posterior cupola and the bases of the ethmo-turbinals I and 11.

31. The ethmoidal skeleton of cat may be compared with that of Lacerta as follows : the region of alar cartilages in the cat and the anterior zone of the lizard; tectum nasi and lamina transversahs anterior of the cat with the zona annularis of the hzard ; the combined paranasal and antorbital parts of the cat with the posterior zone of the saurian.



Arai, H. 1907 Der Inhalt des Canalis craniopharyngeus. Anat. Hefte,

vol. 33. Bardeen, C. R. 1908 Early development of the cervical vertebrae and the

base of the occipital bone in man. Am. Jour. Anat., vol. 8. , Baur, G. 1894 Ueber den Proatlas einer Schildkrote (Platypeltis spinifer

Les.). Anat. Anz., vol. 10. Beecker, a. 1903 Vergleichende Stilistik der Nasenregion bei den Sauriern,

Vogeln und Siiugethieren. Morphol. Jahrb., vol. 31. Blendinger, W. 1904 Das Cribrum der Saugetiere. Morphol. Jahrb., vol. 32. BoLK, L. 1899 De Variaties het grensgebied tusschen hoofd en halwervel kolom. Neder. Tijdschr. voor Geneeskunds, vol. 1.

1904 Entwicklungsvorgange in der occipital Region des Primordialcraniums beim Menschen. Petrus Camper, vol. 2.

BoNDY, G. 1907 Beitrage zur vergleichenden Anatomie des Gehororgans der

Sanger. Anat. Hefte, vol. 35. Broman, I. 1899 Die Entwickelungsgeschichte der Gehorknochelchen beim

Menschen. Anat. Hefte, Abt. 1, H. 37. Broom, R. 1896 On the comparative anatomy of the organ of Jacobson in

Marsupials. Proc. Linnean Soc. of N.'S. W., Pt. 4. Chauveau. 1905 Comparative anatomy of the domesticated animals. New

York. Corning, H. K. 1902 tJber die vergleichende Anatomie der Augenmuskulatur.

Morph. Jahrb., vol. 29. De Burlet, H. M. 1914 Zur Entwicklungsgeschichte des Walschadels. III.

Das Primordialcranium eines Embryo von Balaenoptera rostrata.

Morph. Jahrb., vol. 49. Decker, F. 1883 Ueber den Primordialschadel einiger Saugetiere. Zeitschr.

f. d. wissensch. Zoologie, vol. 38. Denker, a. 1899 Vergleichend — anatomische Untersuchungen liber das

Gehororgan der Saugetiere. Leipzig. Dixon 1906 Some specimens showing indications of the presence of an occipital vertebra. Trans. R. Acad, of Med. in Ireland, vol. 24. Fawcett, E. 1905 On the early stages in the ossification of the pterygoid

plates of the sphenoid bone of man. Anat. Anz., vol. 26.

1906 On the development, ossification and growth of the palate

bone of man. Jour, of Anat. and Physiol., vol. 40. .Fischer, E. 1901 Das Primordialcranium von Talpa europaea. Ein Beitrag

zur Morphologic des Saugetierschadels. Anat. Hefte, vol. 17.

1902 Zur Kenntniss des Primordialcraniums der Affen. Anat. Anz., vol. 20.

1903 Zur Entwickelungsgeschichte des Affenschiidels. Zeitschr. f. Morphol. u. Anthropol., Bd. V.

1905 Das Primordialcranium von Tarsius spectrum. Vorlauf. Mitteilung, Abhandl. d. kgl. Akad. der Wissensch. Amsterdam.

Fleischman, a. 1904 Das Kopfskelet der Amnioten. II. Morphol. Jahrb., vol. 32.



Froriep, a. 1882 Ueber ein Ganglion des Hypoglossus und Wirbelanlagen in der Occipitalregion. Beitrag zur Entwickelungsgeschichte des Saugetierkopfes. Arch. f. Anat. u. Physiol., Anat. Abt. 1886 Zur Entwickelungsgeschichte der Wirbelsaule, insbesondere des Atlas und Epistropheus und der Occipitalregion. II. Beobachtung an Saugethierembryonen. Arch. f. Anat. u. Physiol., Anat. Abt.

1902 Zur Entwickelungsgeschichte des Wirbeltierkopfes. Verh. d. Anat. Geselsch. auf der 16 Vers, in Halle a. S. 1902 Einige Bemerkungen zur Kopffrage. Anat. Anz., vol. 21.

FucHS, H. 1910 Ueber das Pterygoid, Palatinum und Parasphenoid der Quadrupeden, insbesondere der Reptilien und Saugetiere. Anat. Anz., vol. 36.

Gaupp, E. 1898 Die Metamerie des Schadels. Merkel u. Bonnet, Ergebn, der Anat. u. Entwickelungsgesch. vol. 7, 1897.

1900 Das Chondrocranium von Lacerta agilis. Ein Beitrag zum Verstandnis des Amniotenschadels. Anat. Hefte, vol. 15.

1901 Alte Probleme und neuere Arbeiten iiber den Wirbeltierschjidel. Ergebn. d. Anat. u. Entwickelungsgesch., vol. 10.

1902 Ueber die Ala temporalis des Sougerschodels und die Regio orbitalis einiger anderer Wirbeltierschadel. Anat. Hefte, vol. 19. 1905 Neue Deutungen auf dem Gebiete der Lehre vom Siiugetierschiidel. Anat. Anz., vol. 27.

1905 Die Entwickelung des Kopfskeletes. Hertwigs Handb. d. vergl. u. exper. Entwickelungslehre d. Wirbeltiere, vol. 3.

1906 Ueber allgemeine und spezielle Fragen aus der Lehre vom Kopfskelet der Wirbeltiere. Verhandl. d. Anat. Gesellsch. auf. d. 20. Versamml. in Rostock.

1908 A Ueber Entwickelung und Bau der beiden ersten Wirbel und der Kopfgelenke von Echidna aculeata, nebst allgemeinen Bemerkungen fiber die Kopfgelenke der Amnioten, Jen. Denkschr., vol. 6. 1908 B Zur Entwicklungsgeschichte und vergleichenden Morphologie des Schadels von Echidna aculeata, var. typica. Semons zool. Forschungsreisen, vol. 3. (Jen. Denkschr., vol. 6.) 1910 Saugerpterygoid und Echidnapterygoid, nebst Bemerkungen iiber das Sauger-palatinum und den Processus basipterygoideus. Anat. Hefte, 127 Heft.

Glaesmer, E. 1910 Die Atlanto-Occipital-Synostose. Ueber ihre pathologischen oder morphologischen Ursachen auf Grund eines Weichteilpriipurates. Anat. Anz., vol. 36.

Grosser, O. 1902 Zur Anatomic der Nasenhohle und des Rachens der einheimischen Chiropteren. Morphol. Jahrb., vol. 29.

Harvey, R. 1882 Note on the organ of Jacobson. Quart. Jour. micr. Science.

Jacoby, M. 1895 Ein Beitrag zur Kenntnis des menschlichen Primordialcraniums. Arch. f. mikros. Anat., vol. 44.

Jayne, H. 1898 Mammalian Anatomy. The skeleton of the cat. Philadelphia.

Kernan, J. D., Jr. 1915 The development of the occipital region of the domestic cat with an interpretation of the paracondyloid process.


KoLLMANN, J. 1904 Der Canalis cranio-pharyngeiis. Verh. d. Anat. Ges.

Jena, 1904. Anat. Anz., vol. 25. KuNKEL. 1912 The development of the skull of Emys lutaria. Jour. Morph.,

vol. 23. Legal, E. 1883 Die Nasenhohle und der Thriinennasengang der Amnioten

Wirbeltiere (Schwein). Morphol. Jahrb., vol. 8. Levi, G. 1900 Beitrag zum Studium der Entwickelung des knorpeligen

Primordialcraniums des Menschen. Arch. f. mikros. Anat., vol. 55.

1908 Studi anatomici ed embriologici suU' osso occipitale. Arch.

Ital. d. Anat. e di Embr., vol. 7. Macalister, a. 1893 Notes on the development and variations of the atlas.

Jour. Anat. and Physiol., vol. 32. Macklin, C. C. 1914 Skull of a human fetus of 40 mm. Am. Jour. Anat.,

vol. 16. Mead, C. S. 1909 The chondrocranium of an embryo pig, Sus scrofa. Am.

Jour. Anat., vol. 9. Nicola, B. 1906 SuUo svelluppo, sui canali perforanti e sulle fessure della

porzione laterale dell' ala magna dell' os sfenoidale nella specie imana.

Mem. Accad. Sc. Torino, ser. 2, vol. 56. NooRDENBOS, W. Ueber die Entwickelung des Chondrocraniums der Sauge tiere. Petrus Camper, vol. 3. Olmstead, M. 1911 Das Primordialcranium eines Hundeembryo. Anat. Hefte,

vol. 43. OsBORN, H. F. 1900 A Intel-centra and hypophyses in the cervical region of

Mosasaurs, lizards and Sphenodon. Am. Nat., vol. 34.

1900 B Origin of the mammalia. III. Occipital condyles of reptilian tripartite type. Am. Nat., vol. 34. Parker, W. K. 1875 On the structure and development of the skull in the

pig. Philos. Trans. Roy. Soc. of London, vol. 164.

1885 On the structure and development of the skull in Urodeles.

Trans. Zool. Soc. of London, vol. 11. Parker and Bettany, G. T. 1877 The morphology of the skull. London. Peter, K. 1902 Anlage und Homologie der Muscheln des Menschen und der

Saugetiere. Arch. f. mikr. Anat. u. Entw.-Gesch., vol. 60. PoHLMANN, E. H. 1910 Die embryonale Metamorphose der Physiognomie

und der Mundhohle des Katzenkopfes. Morphol. Jahrb., vol. 41. Reighard, J. AND J. 1902 Anatomy of the cat. New York. Retzius, G. 1884 Das Gehororgan der Wirbelthiere. II. Das Gehororgan

der Reptilien, der Vogel und der Saugethierc. Stockholm. Seydel, O. 1891 Ueber die Nasenhohle der hoheren Saugethiere und des

Menschen. Morphol. Jahrb., vol. 17. Shino, K. 1914 Studien zur Kenntnis des Wirbeltierkopfes. Anat. Hefte,

vol. 50. Spence, L. 1890 A comparison of the external and middle ear of man and the

cat. Proc. Am. Soc. of Microsc, vol. 14. • Spurgat, F. 1896 Beitrage zur vergleichenden Anatomie der Nasen- und

Schnauzenknorpel des Menschen und der Tiere. Morphol. Arbeiten,

vol. 5.


Strauss-Durckheim, H. 1845 Anatomie descriptive et comparative du chat,

type des mammiferes en general, et des carnivores en particular.

Paris. SuscHKiN, P. 1895-96 Zur Anatomie und Entwickelungsgeschichte des Scha dels

der Raubvogel. Anat. Anz., vol. 11. SwjETSCHNiKow, W. A. 1906 Ueber die Assimilation des Atlas und die Manifestation des Occipitalwirbels beim Menschen. Arch. f. Anat. u.

Physiol. Anat. Abt. Terry, R. J. 1913 Development of cranium in mammals. Science, vol. 37,

(N. S.), p. 994. ToNKOFF, W. 1900 Zur Entwickelungsgeschichte des Hiihnerschadels. Anat.

Anz., vol. 18. Trolard, p. 1892 Note sur la presance d'un arc osseux dans I'epaiseur du

ligament atlanto-occipital posterieur. Comp. rendus, vol. 44. Van Bemmeln, J. F. 1900 Monotreme skull. Proc. kon. Akad. Wetensch.

Amsterdam. Van Kampen, N. 1905 Die Tympanalgegend des Saugetierschadels. Mor phol. Jahrb., vol. 34. VoiT, M. 1907 Zur Frage der Verastelung des Nervus acusticus bei den Sauge tieren. Anat. Anz., vol. 31.

1909 Das Primordialcranium des Kaninchens unter Beriicksichtig ung der Deckknochen. Anat. Hefte, vol. 38. Von Noorden, W. 1887 Beitrag zur Anatomie der knorpeligen Schadelbasis

menschlicher Embryonen. Arch. f. Anat. u. Physiol. Anat. Abt. Vrolik, a. J. 1875 Studien iiber die Verknocherung und die Knochen des

Schiidels der Teleostier. Niederland. Arch. f. Zool., vol. 1. Vulpian, a. 1862 Racine posterieur ou ganglionnaire du nerf hypoglosse.

Jour, de la Physiologic, vol. 5. Weber, Max 1903 Die SiJugetiere. Einfuhrung in die Anatomie und Syste matik der recenten und fossilen Mammalia. Jena. Weigner, K. 1911 Ueber die Assimilation des Atlas und liber die Variationen

am Os occipitale beim Menschen. Anat. Hefte, vol. 45. Weiss, A. 1901 Die Entwickelung der Wirbelsaule der weissen Ratte, be sonders der vordersten Halswirbel. Zeitschr. f. wissensch. Zool.,

vol. 69. WiJHE, J. W. van 1902 Eene niewe methode ter demonstratie van kraak beenige mikroskeletten. Kon. Akad. van Wettenschappen, Amsterdam, 30 April. Wilder and Gage 1882 Anatomical technology as applied to the domestic

cat. New York. Williams, L. W. 1908 The later development of the notochord in mammals.

Am. Jour. Anat., vol. 8. WiNCZA, H. 1896 Ueber einige Entwickelungsveranderungen in der Gegend des

Schadelgrundes bei den Saugethieren. Anz. d. Akad. d. Wissensch.

in Krakau. ZucKERKANDL, C. 1908 Uebcr den Jacobson'schen Knorpel und die Ossification des Pflugscharbeines. Sitzungsb. d. Wiener Akad. Wiss., math. naturw. Kl., vol. 117.


All figures are from preparations of embryos of the cat, the lengths of which are indicated. With the exception of figures 17 and 18, the drawings are the work of Mr. C. D. Jarrett.

In the model from which figures 1 to 5, 12, 13 have been made, cartilaginous parts only are represented on the left side; on the right side, in addition, are shown vessels, nerves and bones. In figures 1 and 2, the right otic capsule is represented as partly opened. In figure 3, the incudal cartilage and part of Meckel's cartilage have been omitted. Figures 15 to 30 are camera lucida drawings of the sections represented. Figures 19 to 30 are from sections of the 23.1 mm. embryo (H. E. C. Series 466); the number of the section is indicated in each fieure.

Fig. 1 From a model of the cranium of an embryo 23.1 mm. long (H. E. C. Series 466) Dorsal aspect, X -/.

Fig. 2 The same; ventral aspect.

Fig. 3 The same; left lateral aspect.

Fig. 4 The same; right lateral aspect.

Fig. 5 From a van Wijhe preparation of the head of an embryo 10.5 mm. long (W. A. C. No. 310). Dorsal aspect, X V Fig 6 From a van Wijhe preparation of the head of an embryo 12 mm. long (W. A. C. No. 311). Left side and base seen froni the right, X r.

Fig. 7 From a van Wijhe preparation of the head of an embryo 15 mm. long (W. A. C. No. 312). Dorsal aspect, X V. The breadth of this is somewhat exaggerated, the ala temporalis being too far toward the side.

Fig. 8 From van Wijhe preparations of the heads of embryos 18-20 mm. long (W. A. C. Nos. 313, 314). Dorsal aspect, X V.

Fig. 9 From a van Wijhe preparation of the head of an embryo 24 mm. long (W. A. C. No. 315). Right occipital and otic regions seen from the left. In this, as in all the van Wijhe preparations, the cochlear capsule is for the most part only lightly stained. X V^.

Fig. 10 From a model of the left otic capsule, of an embryo 23.1 mm. long (H. E. C. Series 466). The large, cut surface is in a frontal plane through the middle of the vestibule. Dorsal and medial aspect, X t".

Fig. 11 The same as figure 10, excepting the level of the cut surface, which is at the ventral part of the vestibule.

Fig. 12 Region of the foramen rotundum as it appears in the model of the 23.1 mm. embryo. Ventral and caudal aspect, X t".

Fig. 13 Ethmoidal skeleton; from the model of the 23.1 mm. embryo. Represents the left half from its medial aspect, the septal cartilage being omitted and the nasal tectum, laminae transversales and lamina infracribrosa sectioned just to the left of and parallel with the median plane. X -Y'

Fig. 14 From a model of the basal plate and epistropheus in median section showing the course of the notochord. Embryo 23.1 mm. long. X ¥.

Fig. 15 Transverse section through the occipital region. Embryo 10.6 mm. long (H. E. C. Series 476, sect. 300). X V' 405

Fig. 16 Frontal section through base of orbito-temporal region. Embryo 12 mm. long (H. E. C. Series 403, sect's. 200-207). X ¥-.

Fig. 17 Median sagittal section through base of cranium. Embryo 12 mm. long (H. E. C. Series 400). X ¥•

Fig. 18 Median sagittal section through base of cranium. Embryo 15 mm. long (H. E. C. Series 437, sect's. 158, 161, 169). X ¥.

Fig. 19 Frontal section through occipital and otic regions. (Sect. 246). X ¥.

Fig. 20 Frontal section through occipital and otic regions. (Sect. 253).

X -¥.

Fig. 21 Frontal section through the otic region. (Sect. 256). X -rFig. 22 Frontal section through the otic region. (Sect. 273). X V. Fig. 23 Transverse section through the orbito-temporal region. (Sect. 300).

X ¥-.

Fig. 24 Transverse section through the orbito-temporal region. (Sect. 339). X ¥•

Fig. 25 Transverse section through the orbito-temporal and ethmoidal regions. (Sect. 385). X V Fig. 26 The same as in figure 25 but further forward. (Sect. 404). X ¥•

Fig. 27 Transverse section through the ethmoidal region. (Sect. 457).

X ¥.

Fig. 28 Transverse section through the ethmoidal region. (Sect. 489). X ¥•

Fig. 29 Transverse section through the ethmoidal region. (Sect. 492).

X V-.

Fig. 30 Transverse section through the ethmoidal region. (Sect. 513). X Y.


VI, abducent nerve

XI, accessory nerve

Ac.fis., acustic fissure

VIII, acustic nerve

Al. orb., alsi orbitalis

Al.temp., ala temporalis

Alar., alar element

Amp. a., ampulla anterior

Amp. I., ampulla lateralis

Amp. p., ampulla posterior, antorbital plane

Aq.coc, aquaeductus cochleae, atlantal process

At., atlas

Atrt., atrioturbinale, basal plate

Can.sem,.ant., canalis semicircularis

anterior, canalis semicircularis


Can. sem. post., canalis semicircularis

posterior Car. for., carotid foramen Cav.epipt., cavum epiptericum Cav.coc, cavum cochleae Cav.nas., cavum nasi Cav. vest. ant., cavum vestibulare an terius Cav. vest. post., cavum vestibulare pos terius, centrum of atlas Cent.ep., centrum of epistropheus Ch.tym., chorda tympani Co.cap., cochlear capsule Co.d., cochlear duct Com. alio., commissura alicochlearis Com.bas., commissura basicochlearis Com.basiv., commissura basivestibu laris Com..el., commissural element


Co7n.orb.-par., commissura orbito-parietalis

Com.p.-cap., commissura parieto-capsularis

Com.sph.-eth., commissura spheno-ethmoidalis

Com.supr.f., commissura suprafacialis, crista interorbitalis

Cr.intervest., crista intervestibularis

Cr.par., crista parotica

Cr.sem., crista semicircularis

Cr.trans., crista transversa

Den., dens epistrophei

Dent., dentale, dorsum sellae, ductus cochlearis

D.sem.a., ductus semicircularis anterior

D.sem.l., ductus semicircularis lateralis

D.sem.p., ductus semicircularis posterior

Endl.d., endolymphatic duct

Ep., epistropheus

Eth. I, ethmoturbinal I

Eth. II, ethmoturbinal II

Ext.meat., external auditory meatus

Oc, eyeball

VII, facial nerve

Fen.acust., fenestra acustica

Fen. has., fenestra basalis

Fen. has, post., fenestra basicranialis posterior

Fen.nar., fenestra narina

Fen.olf., fenestra olfactoria

Fen.sphen.-par., fenestra spheno-parietalis

Fen.vest., fenestra vestibuli

Fis.alicoch., fissura alicochlearis

Fis. has. post., fissura basicochlearis posterior

Fis.orh.-eth., fissura orbito-ethmoidalis

For. CO., foramen cochleae

For.endl., foramen endolymphaticum

For.epiph., foramen epiphaniale

For.fac, foramen faciale

For.hyp., foramen hypophyseos

For. jug., foramen jugulare

For.jug.sp., foramen jugulare spurium

For.mag., foramen occipitale magnum

For. peril., foramen perilymphaticum

For.rot., foramen rotundum, fossa incudis

Fos.oc-can., fossa occipito-canalicularis

Front., frontal bone

Gen. gang., geniculate ganglion

IX, glossopharyngeal nerve

Gr.sup., great superficial petrosal nerve

Hi.sejn., hiatus semilunaris, hypochordal arch of atlas

Hyp. can., hypoglossal canal

XII, hypoglossal nerve

Hyph.can., hypophyseal canal

Hyph.cart., hypophyseal cartilage

Hyph., hypophysis

Inciv., incisive bone

Inciv.d., incisive duct

Inc.lac, incisura lacera

Inc.lacr., incisura lacrimalis

Inc.7nax.atr., incisura maxillo-atrioturbinalis

Inc.occ.ant., incisura occipitalis anterior

Inc. occ. post., incisura occipitalis posterior

Inc. oval., incisura ovalis

Inc.posttr., incisura postransversalis

Inc. sing., incisura singularis

Inc.cart., incudal cartilage

Inf. ac. for., inferior acustic foramen

Inf.ohl.m., inferior oblique muscle

Infr.n., infraorbital nerve

Int. car. art., internal carotid artery

J.O., Jacobson's organ

Jug. v., jugular vein

Lam. ant., lamina antorbitalis

Lam.asc, lamina ascendens

Lam.infracr., lamina infracribrosa

Lani.par., lamina parietalis

Lam.suprch., lamina supraconchalis

Lam. trans. ant., lamina transversalis anterior

Lam. trans. post., lamina transversalis posterior, lateral mass of atlas

Lat, lateral nasal glands

Lat.nas.n., lateral nasal nerve

Lat. occ, lateral occipital arch

Mand.cond., mandibular condensation of ala temporalis

Mand..n, mandibular nerve


M, massa angularis

Mast., masseter muscle

Max., maxillary bone

Max.cond., maxillary condensation of ala temporalis

Max.n., maxillary nerve

Maxturb., maxilloturbinal, meatus acusticus internus

Meat.inf., meatus inferior of the nose

Meats., meatus supraconchalis

Meek., Meckel's cartilage

Memh.lim., membrana limitans

Mel. opt., metoptic root

Mo., mouth

N.L.D., naso-lacrimal duct, nasopharyngeal duct

Nast^irb., nasoturbinal, neural arch of atlas neural arch of epistropheus

Note, notochord

Cond.occ, occipital condyles

III, oculo-motor nerve

Olf., olfactory nerve

Oph., ophthalmic nerve

Op.for., optic foramen

II, optic nerve

Pal., palate bone

Pal.max., palate process of maxilla, parachordal plate

Parna.cart., paranasal cartilage

Para.cart., paraseptal cartilage

Par.nas., paries nasi

Par.-tect., parieto-tectal cartilage

Pa. air.; pars atrialis

Pa.can., pars canalicularis, pars cochlearis

Pa.lac, pars lacrimalis

Phary.pch., pharyngeal pouch I

Pl.suprch., planum supracochleare

Post.cup., posterior cupola

P. opt., preoptic root, processus alaris

Pre. max. ant., processus maxillaris anterior

Prc.orb., processus orbitalis

Prc.para., processus paracondyloideus

Prc.pty., processus pterygoideus, prominentia lateralis

Prm.utr., prominentia utricularis

Prm.ut.inf., prominentia utriculo-ampullaris inferior

Prm.ut.sup., prominentia utriculo-ampuUaris superior

Pty.cart., pterygoid cartilage

Pty.n., pterygoid nerve

Pty.m., pterygoid muscle

Rec.amp.ant., recessus ampuUaris anterior

Rec. amp. post., recessus ampullaris posterior, recessus lateralis, recessus lateralis inferior sup., recessus lateralis superior

Reich., Reichert's cartilage

Sac, sacculus

Sel.tur., sella turcica

Sent. gang., semilunar ganglion

Sept., septal cartilage

Sep.sp., septum spirale, septum transversum

Sph.-orb.fis., spheno-orbital fissure

Sph.p.gang., sphenopalatine ganglion

Sp.N.I, Spinal Nerve I

Sp.N.II, Spinal Nerve II

Sq., squamosal

St., stapedial cartilage

Stp.m., stapedius muscle, sulcus chiasmaticus

Sulc.fac, sulcus facialis, sulcus lateralis anterior, sulcus lateralis posterior

Sulc.sig., sulcus sigmoideus

Sulc.sp., sulcus spiralis

Sulc.suprach., sulcus supraconchalis

Sulc.supr., sulcus supraseptalis

Sup. ac. for., superior acustic foramen

Tect.nas., tectum nasi, tectum posterius

Temp.rn., temporal muscle

Trab., trabecular plate

V, trigeminal nerve

IV, trochlear nerve

Tub.eph., tuberculum ephippii

Tub. jug., tuberculum jugulare

X, vagus

Vert. a., vertebral artery

Vo., vomer

Utr., utriculus

Zyg., zygomatic bone




Com. sph. etl

Hyp. can




O >■

















JL am. infracr. Lam. irans. pus

13 425




Cent, ep.

Hyp. ar. at

Hyph. cart.

Memb. lim.




In /

\ f!.-\K - .



\ vl

„: •



gang Gtn










VII 1 Men.l). l.m.


^P. 21


.'II Gen g:ing Memb ..\ ...... ,\...l

Pre. para. Jug- v, XII










Zoological Institute, Tokyo Imperial University, Japan


While at Misaki in the summer of 1900 I undertook, at the suggestion of Professor Watase, a study on the ostracods with especial reference to the cytology of the luminous glands. In the winter of 1900-1901 quite a number of specimens were cut into sections and several sketches were made. But my going to America obliged me to suspend the work. Last summer ('16) Dr. Harvey of Princeton University came over to Japan and carried out at the Misaki Biological Station biochemical experiments upon the luminous substance formed in the maxillary gland of Cypridina hilgendorfii. In connection with his study it may be of some interest to publish the results of my observation on the morphology of the gland carried out sixteen years ago.

In Cypridina hilgendorfii (Miiller '90, p. 228) the maxillary gland (Oberlippendriise) attains a considerable development, consisting of a number of large gland cells. These cells open to the exterior on five protuberances, one unpaired anterior and two pairs of posterior ones (figs. 1 and 2). The anterior process has five or six pairs of openings. The middle pair is shorter t-han the anterior one and each has only three or four openings. The hindmost pair is the longest of all, projecting over the anterior wall of the mouth cavity, and each has seven or eight openings along its side. The last pair of protuberances alone is provided with fine hairs.

The maxillary gland is a group of unicellular glands, as have often been described by previous writers. The longest cells measure nearly 0.7 nmi. in length and reach as far up as the




supraoesophageal ganglion. Their diameters vary naturally according to their secretory activity. The gland cells usually assume a club shape, tapering towards the openings. In this form as in other marine ostracods, the gland cells are not differentiated into the gland proper and the duct, contrary to those of fresh water forms (Bergold, '10). The secretion granules fill

Fig. 1 Upper lip of C. hilgendorfii seen in an oblique view from the anterior side. One of the middle pair of protuberances is behind the anterior one. X 115.

Fig. 2 Upper lip of C. hilgendorfii f 9 ) seen from the right side. Fresh, material treated with methylgreen acidulated with acetic acid. Mucous gland cells are stippled. X 90.

up a greater part of the cells, if specimens are killed quickly, that is, before the oozing out of the contents. But in individuals from which the secretion products have been discharged, the lower portion of the cells has a large space containing a coarse spongy coagulum. This space, especially in poorly fixed preparations, is liable to give a deceptive picture, as though there were a large reservoir common to all the gland cells.



The maxillary gland is composed of two entirely different elements, namely the mucous gland cells and the yellow gland cells. The posterior protuberances have the openings of the former alone, while both the anterior and middle ones have those of two kinds of cells (fig. 2). The secretion product of the mucous cells of the maxillary gland stains intensely green with methylgreen, as in the gland cells found in the appendages and the

gm -iMM

Fig. 3 A portion of a vertical section through the maxillary gland of C. hilgendorfii (6). Alcohol-acetic preparation. X 247.

Fig. 4 Cross section through the anterior process of the upper lip (9). Flemming's fluid preparation. X 247.

mantles. In sections the cytoplasm of the mucous cells is finely granular. The other kind of gland cells I shall call the yellow gland cells, since they are yellow in life, due to the color of the secretion product they contain. The yellow substance, which emits light, is in the form of coarse, somewhat angular, granules. The size of the granules varies considerably, often attaining the diameter of 10 to 15 m- In general it may be said, that the nearer the openings, the larger the granules. Their seat of formation is in all probability near the nucleus, inferring from the fact that a small number of minute granules of the same


nature is usually found there. The cytoplasm of the yellow gland cell is fibrillar, in contrast to that of mucous cells, which is granular as has already been mentioned. Figure 4 shows a section through the anterior protuberance near the openings. Here we see very clearly the relative position of both kinds of gland cells.

In comparison with Cypridina hilgendorfii, the maxillary gland of Pyrocypris japonica (Miiller '90, p. 233^ Was studied. The protuberances of the latter differ both in shape and number from those of the former, one unpaired one being added in front. But so far as the internal make-up of the glands is concerned, it is so similar that its description would be simply the repetition of what has been stated above. The only difference is the presence of pigment cells in the upper hp of P. japonica.

In conclusion, I would like to emphasize once more the two points, that is, the presence of two kinds of gland cells and the absence of a reservoir for secretion granules common to all the gland cells. It should be mentioned that Watanabe was the first to direct especial attention to the luminous glands of ostracods at Misaki. He read a paper on this subject before the meeting of the Tokyo Zoological Society on January 23, 1897, an abstract of which appeared in Japanese ('97). He states that the maxillary glands secrete colorless transparent fluid and yellow homogeneous substance. But he does not say whether each gland cells produces these two substances or whether there are two different glands for them. Miiller ('90, p. 248) clearly states that he saw two groups of gland cells of different nature, i.e., the secretion product of the upper group is found in the form of droplets and does not take carmine stain at all, while that of the lower group takes it. Furthermore he advanced the view that light is produced by the interaction of these two substances. He, however, seems to have failed to observe the correct topographical relation of these two kinds of cells.

1 The general outline of the shell of Pypocypris japonica is more like that of P. chierchiae (Miiller, '90, pi. 25, fig. 3) the posterior process being longer than in the figure for this species.


One other point, which I would like to call attention, is the absence of common reservoir for gland cells. Although all the students of ostracods maintain that the maxillary glands are unicellular, yet curiously enough those who have studied the luminous glands entertain an erroneous idea that there is a special cavity to store up the secretion product. Miiller ('90, p. 248), for instance, states that "die Ausfiihrunsgange sammtliche Driisenzellen vereinigen sich zu einem gemeinsamen Hohlraum." Doflein published a paper on the maxillary glands of a Japanese species of ostracod, which he provisionally calls Halocypris (?) and gives a semidiagrammatic figure ('06, p. 134). Since pigment cells are drawn in his figure, his material may have possibly been Pyrocypris japonica. At any rate he interpreted the section as though there were a spacious reservoir for secretion granules. Probably influenced by Doflein's description, Liiders ('09) a' so mentions the presence of a special reservoir in Gigantocypris agassizi. It should be mentioned that the above authors seem to have studied specimens from which a greater part of the secretion products had been discharged, and it is, I think, quite natural that they have come to such an interpretation. The lower part of each gland cell functions as a temporary reservoir of the secretion granules it is true, but this cannot be called a special organ at all. As a matter of fact, as I have expressly mentioned above, there is no reservoir in the sense of previous writers.

January 9, 1917



Bergold, a. 1910 Beitriige zur Kenntnis des inneres Baues der Siisswasser ostracoden. Zool. Jahrb., Abt. Anat., Bd. 30. DoFLEiN, F. 1906 tJber Leuchtorganen bei Meerestieren. SB. der Gesell. f.

Morph. u. Phys. Mtinchen, Bd. 22. LtJDERS, L. 1909 Gigantocypris agassizi (Miiller). Zeit. wiss. Zool., Bd. 92. Mangold, E. 1910 Die Production von Licht. Winterstein's Handbuch d.

verg. Physiol., Band 3, Heft 2. MtJLLER, G. W. 1890 Neue Cypriniden. Zool. Jahrbuch., Abt. Syst., Geogr.

u. Biol., Bd. 5. Watanabe, H. 1897 Umihotaru no Hakko ni tuite (On the phosphorescence of

Cypridina hilgendorfii). Dobutugaku-zassi 9 also in Annotationes

Zool. Japon. 1 and Zool. Jahresbr. fiir 1897, Arth., 24.

authors' abstract of this paper issued

BY the bibliographic SERVICE SEPTEMBER.



From the Laboratory of Zoology, University of California^



Introduction 441

Material and method 442

General considerations 442

The last oogonial division and early pre-leptotene changes 443

Contraction figures 446

The leptotene stage 447

The synaptene stage 447

The pachytene stage 448

The diplotene stage 450

The nuclei of the early growth period 451

The centrosome and chromatoid body 451

Discussion and conclusions 453

Literature cited 455


Of the very considerable literature on synapsis much relates to vertebrates, but only a small amount to mammals. In nearly all of the pubhcations on mammals synapsis in the male. is described, the process in the ovary being investigated in only a very few cases. In view of the importance of synapsis in its relation to theories of heredity and to the behavior of the chromosomes it has seemed worth while to extend the study to the female of another mammal. Of particular interest have been the problems of the presence of an accessory chromosome, of the identity of the chromosomes throughout the series of changes, and ot whether or not an actual pairing of chromosomes takes place.

1 Aided in the preparation of most of the figures by the Department of Anatomy of the University of California.




The rat was selected because of its interest in connection with other studies and because of the possibility of obtaining a chronological series of embryos of known ages, and, consequently, of securing a natural series of stages.


Since the period of synapsis in the egg of the rat is well along toward completion at birth, it was found necessary to obtain embryos of known age in order to complete the necessary series of slides. These were obtained in the following way. It having been clearly demonstrated that the female albino rat normally ovulates within twenty-four hom*s after having given birth to a litter (Long, '12), mother rats were taken from their young immediately after parturition and placed with sexually active males. Twenty-four hours later these females were taken from the males and left alone or with females only. Upon subsequent examination about 50 per cent of the rats so treated were found to be pregnant, and the age of their embryos was consequently known to within a fraction of a day. For the most part the sex organs were dissected out of these embryos before fixing, but in a few cases the entire embryo was fixed and sectioned. For those stages occurring after birth the ovaries of young of appropriate ages were used. The fixing solutions employed were Zenker's, Flemming's strong solution, Bouin's formol-picro-acetic, and sublimate acetic. The first two gave the best results and were used almost exclusively. The material was sectioned from 3 to 10 micra thick and for the most part stained in Heidenhain's iron haematoxylin. Usually a counterstain was employed, orange G or acid fuchsin, or a mixture of the two being most satisfactory. Erythrosin was used to some extent. A few slides were stained in safranin and gentian violet.


The process of synapsis in the rat lasts for approximately ten days, starting about seven days previous to birth and being practically completed by three days thereafter (table, p. 444). Following von Winiwarter ('00) it has been found possible to


distinguish clearly leptotene, synaptene, pachytene, diplotene, and dictye stages, together wdth more or less complicated transition changes to be mentioned later. These various stages succeed each other with considerable regularity throughout the entire ovary, so that at any one time most of the germ cells are in the same stage of the synaptic process. While it is true that those nuclei least advanced in the general process of change are most likely to be found at the periphery of the ovary, there is an absence of that clear distribution of stages in concentric zones so evident in the cat, rabbit, and man (Winiwarter '00, and Winiwarter and Sainmont '08). Because of these features of the process, the difficulty of determining the sequence of stages has been reduced to a minimum.


The last oogonial division occurs about seven days before birth. In the ovaries of embryos from female 389 (table, p. 444) there are a large number of division figures together with numerous nuclei which have not yet passed through the last oogonial division and also many that belong to the early stages of the oocytes. Some few division figures may still be ftjomd in embryos from female 451-6, at six days before birth. The nuclei of the oogonia in the typical resting condition are spherical. They are approximately equal to or perhaps a little small-er than the leptotene nuclei to be described later, and are supplied with a large amount of chromatin in irregular blocks. Often one or more of these blocks is comparatively large and conspicuous, somewhat suggestive of the chromatin nucleoli of the post-synaptic stages, though they never have the definiteness of outline characteristic of these later chromatin bodies. A heavy discontinuous spireme is formed during the prophase of the last oogonial division, and often the nuclei appear slightly oval at this time. The chromatin is very closely massed together during the equatorial plate stage as well as during the anaphase, so that it is extremely difficult to distinguish individual chromosomes. The last oogonial division may be distinguished from previous ones by the numerous



deeply staining granules found in the cytoplasm during the anaphase, and by the character of the chromatin in the daughter nuclei. In the very young daughter nuclei, when the nuclear wall is first appearing, the chromatin forms a somewhat flattened, closelypacked aggregation of blocks exceedingly varied in size and shape (fig. 1). No linin is evident in such nuclei. A conspicuous, very elongate, spindle-shaped remnant of the interzonal fibers and the zwischenkorper extends over the two adjoining margins of the two daughter cells containing such nuclei. These cells may be regarded as the very first stages of the primary oocytes.

Table showing time of occurrence of synaptic changes














Gonad barely distinguishable





Gonad well formed, multiplication division





Last oogonial division and pre-leptotene stages





Pre-leptotene and contraction figures





Contraction figures and early leptotene





Leptotene and synaptene





Synaptene and some pachytene





Synaptene and pachytene






22 .


















Pachytene and diplotene





Pachytene and diplotene





Diplotene and dicty6






In the same ovary with these late telophase nuclei and often immediately adjacent to them are other nuclei which, for the following reasons, may be considered as directly derived from them and as constituting the early stages of the preleptotene nuclei. Although these preleptotene nuclei may he side by side with the late telophase nuclei they are shghtly larger, almost spherical and contain chromatin still in the form of numerous


blocks of irregular size and shape (figs. 2, 3, and 4) which are distributed over the inner side of the nuclear membrane. The outline of the blocks is clear cut and definite. A few linin threads may be attached to one or more of the blocks at this stage (figs. 3 and 4) . As in the case of the late telophase nuclei, here, too, the preleptotene nuclei are evidently intimately associated in pairs, as is clearly proved where remnants of interzonal fibers and zwischenkorper persist between the cells of two such nuclei. It may be concluded then that the late telophase nuclei increase in size and become more spherical; and at the same time the chromatin blocks separate and become arranged about the inner surface of the nuclear membrane. In the early stages the number of these blocks of chromatin varies from twenty-two to thirty-five and seems therefore not to be significant, except in indicating a fragmentation of the original chromatin masses.

The comparatively simple condition of the contents of the nucleus represented by the peripherally arranged chromatin blocks changes gradually into an extremely complex, chromatinlinin network dispersed more or less regularly throughout the entire nucleus. The changes represented by these closely similar conditions involve a regularly progressive increase in the number of chromatin blocks, a gradual augmentation of the number and complexity of arrangement of the linin threads, and a more or less equal distribution of the chromatic material throughout the entire nucleus, (figs. 4, 5, and 6). In the earlier phases of this series of changes it may be readily seen that all of the linin threads are attached to chromatin bodies, or, that the many small, almost powdery particles of chromatin are arranged along linin threads. There is no distinct chromatin element constant in size or shape throughout this series of changes.

This whole series of changes, from the very first form of the primary oocytes to the most complex condition of the chromatinlinin network, is strongly suggestive of the migration of chromatin along linin filaments which later probably become incorporated in the leptotene threads. It is very difficult to explain in any other manner the gradual decrease in average size and the increase in number of the chromatin blocks, simultaneously with


the appearance of the Unin filaments. A study of the oldest nuclei alone would hardly lead one to such an inference, but the complete series seems very convincing. Such an elaborate series of transformations intervening between the last oogonial division and the leptotene stage has not been described in the ovary of any other mammal carefully studied, though there is a slight indication of such changes in the cat (Winiwarter '00, and Winiwarter and Sainmont '08).


Immediately preceding the leptotene stage, or possibly coincident with the earlier portion of it, there is a clearly defined contraction stage. This occurs about six days before birth. The chromatin-linin is eccentrically arranged in a confused mass closely packed against the nuclear wall (fig. 7). Such figures may be found throughout the entire ovary after either Zenker's or Flemming's fixation, fixation that is evidently good in all other portions of the same tissue. Since these figures are found in all portions of well fixed ovaries, and since there is no constancy whatever in regard to the direction of contraction, it is not possible to agree with Schaffner ('07), Von Hoof ('12) and others, who contend that this so-called contraction stage is an artifact due to poor fixation. It seems rather that this stage is quite as normal as those stages immediately preceding or following it. In these figures it is obvious that the eccentrically placed darkly staining mass is not of the same consistency throughout, but is composed of chromatin blocks, granules, and linin threads. Finely attenuated Unin threads often extend from the chromatin knot to the opposite wall of the nuclear cavity. The whole picture strongly suggests that the complex chromatin-linin network has somehow been concentrated at one side of the nuclear cavity, and that those strands of linin attached to the opposite side of the nucleus have not entirely shared in the process.



The leptotene stage (figs. 8 and 9) which follows the contraction stage, occurs about five to three and one-half days previous to birth. It may even persist to a later date, as illustrated by figure 9, from an ovary three days before birth, in which most of the nuclei had passed over into the pachytene stage. The chromatin threads are either regular in outline with a constant size in cross-section or slightly moniliform, exceedingly fine, and very numerous. Many definite visible ends suggest that a large number of individual threads are present. As indicated in figures 8 and 9 the greater part of the threads lie at one side of the nucleus. This arrangement may well be considered as determined by the previous contracted state of the chromatin and as leading to the subsequent polarized orientation of the pachytene. In fact, the above figures show a tendency for many of the ends of threads to reach the nuclear membrane in the region near which the so-called centrosome is found when demonstrable at this stage. In figure 9 are a number of confused clumps of chromatin characteristic of early leptotene nuclei, but no definite chromatin body, uniform in either size or shape could be found at this period. A tendency toward parallel arrangement of threads is very apparent in all nuclei closely examined, and in many cases (figs. 8 and 9) figures indicating the side by side approximation of individual threads are clearly demonstrable. Where the moniliform character of the threads is appreciable, the enlargements into which the chromatin is collected tend to he exactly opposite to each other in the more closely approximated parallel threads. These leptotene nuclei are not appreciably larger than those of the stages immediately preceding. It was not possible to determine the number of individual threads in these nuclei.


The synaptene stage may be found from three and one-half to about two days before birth. It often happens that some portions of the threads may show the bipartite nature at a much later period, but, for the most part, do not do so as late as a couple


of days before birth. The orientation of the threads first discernible in the leptotene stage is distinctly more advanced. Figure 11 is a lateral view while figure 10 is a polar view of a nucleus in the late synaptene stage. In figure 10 the nucleus is viewed from the pole opposite the so-called centrosome. It is clearly evident from figure 11 that the loops of the synaptene threads are of different lengths, some of them extending barely half as far out into the nucleus as do the others. Counts of the loops as seen in polar views or sections have not been made at this stage. Moreover, since it has not been possible to determine the number of leptotene threads, a comparison of the numbers in these two stages cannot be made as evidence of pairing. Nevertheless a consideration of the sizes and structures of the leptotene and synaptene threads give reasons for beheving that the latter are derived from the former. A comparison of figures 8 and 9 with 10 and 11 makes it evident that the synaptene threads are about double the width of the leptotene. The synaptene is further clearly composed of distinct moniliform halves separated by a longitudinal split (fig. 10). Each half of the double synaptene thread is very similar to the leptotene threads, and bears much the same relation to the other half which some leptotene threads have to other leptotene threads. That is, the spindle shaped and laterally flattened enlargements in each half are usually opposite and nearly or quite in contact with those of the other half. Also the portions of each moniliform half, between the deeply staining enlargements, are almost or quite devoid of chromatin. It seems then that each of these synaptene threads was formed by the side by side approximation of two leptotene threads.

Throughout this period as in the preceding ones a chromatin body in any way simulating an accessory chromosome is not evident.


The disappearance of the longitudinal division in the synaptene thread marks its transformation into the typical thread of the pachytene stage. This stage begins two days or thereabouts before birth and persists for about a day thereafter. The ori


entation of the previous stage is retained, though it becomes more indefinite (figs. 12, 14 and 15). The nuclei sometimes appear more oval, in contrast to the former spherical shape, and the thread is more regularly distributed throughout the entire nucleus (fig. 15), with a tendency, perhaps, toward a peripheral arrangement. The thread itself is of about the same thickness as that of the preceding stage. It is distinctly monihform, and in early pachytene nuclei it is clear cut and definite in outline (figs. 12 and 14). The enlargements of the thread stain heavily and show no evidence of their double origin except in very early threads (fig. 12). In the intervals of the thread between these enlargements, there is a paucity of chromatin, and in some cases it seems entirely wanting, only the linin framework being left. In the later stages of the pachytene nuclei, the outline of the thread becomes less distinct due to the extension laterally of numerous very fine linin strands connecting these enlargements of the thread.

Figure 13 represents a pole view of a nucleus similar to figure 11 or 12 cut through the equator. The threads are accurately drawn as they are seen near the sphere or centrosome. They show no sign of being double and each one undoubtedly represents one end of a loop. They number 40. Consequently there would be 20 loops; and, if each loop represent paired chromosomes, the diploid number of chromosomes is 40, the haploid 20. Too great reliance, however, must not be put on the number obtained by this single count.

For the first time in this series of stages, a distinct chromatin body becomes distinguishable in the nucleus (figs. 12 and 15). This chromatin body, however, is constant neither in shape nor in ^ize, and is often distinctly connected with one or more pachytene threads. It can hardly be regarded with reason as an accessory or sex chromosome, but is rather to be looked upon as the first appearance of the large chromatin nucleoli which characterize the nuclei of the early growth period.



Nuclei in the diplotene stage may first be distinguished at the time of birth, or possibly a little earlier, and persist in comparatively large numbers for a couple of days thereafter. At three days after birth they are but infrequently met with, practically all of the nuclei having passed over into the stage characteristic of the early period of growth and designated by Winiwarter as 'Noyoux dictye.' The early diplotene nuclei (figs. 16 and 17) show numerous contorted threads, more or less irregular in outline, but often distinctly parallel in arrangement or wound about each other in spirals. Fine strands of linin are often distinctly visible passing from thread to thread. The number of individual threads is clearly greater than in the pachytene stage, though the exact relations in this respect could not be definitely determined . The diplotene threads are also clearly only about half as wide as those of the previous stage. The orientation, so conspicuous in the threads of synaptene and pachytene nuclei, is almost entirely lost. A large chromatin body is present in each of these nuclei, and very often an additional one or two may be discovered. The chromatin threads very commonly are united to the chromatin body in large numbers (fig. 16). Some of the threads are always connected with these bodies. It seems from an examination of a large number of these diplotene nuclei, that the individual threads are formed by a longitudinal splitting of the pachytene threads, and that the chromatin is rapidly being assembled in the large chromatin bodies. The appearance of the later phases of the diplotene stage gives further evidence of the validity of these inferences: the individual threads are no longer distinctly visible as such; often rows of granules indicate their probable former course; the linin may still be apparent, in arrangement strongly suggesting the course of the former threads. The chromatin bodies, chromatin nucleoli, are more definite in outline and typically number one to three. These later stages gradate imperceptibly into the nuclei characteristic of the early growth period.



By three days after birth, most of the nuclei have passed over into the resting stage, so-called, and in this condition they remain, with certain modifications, up to the period of sexual maturity. Typically from one to three large definitely outlined chromatin nucleoli are present. Throughout the remainder of the nucleus a fine achromatic network is dispersed, with an occasional deeply staining block or particle suspended mthin it (text fig. 1). There is a decided increase in the size of the nucleus coincident with the beginning of the enlargement of the cell itself. The amount of achromatic material in these nuclei is strikingly greater than that of any preceding period. Its arrangement is without constant order and gives no hint of its relation to the previous stages. If it be the linin of previous stages, certainly there is here little evidence in support of that theory which postulates that the individuality of the chromosome is maintained through the constant discreteness of the linin framework.


While no attempt has been made to analyze carefully the cytoplasmic changes which occur simultaneously with the series of nuclear alterations just described, our attention has been drawn to two distinct structures which appear in the cytoplasm very constantly during a part of the process. These are the so-called centrosome or body of Balbiani and the chromatoid bodies. The centrosome appears as a roughly spherical or elongated condensed portion of finer texture than the remainder of the cytoplasm, lying never more than half way from the nuclear wall to the periphery of the cell. It has no definite membrane and its outline is uneven. It is first discernible in the early leptotene nuclei (fig. 9) where it may occur in only a comparatively small percentage of the cells. Throughout the synaptene stage it is more conspicuous, and only rarely at this time is there within it a darkly staining granule, perhaps the centriole. It is very evident in most of the cells in the pachy


tene stage, and the granule above referred to as the centriole here becomes more constantly apparent (fig. 15). At this period, in ovaries fixed in Flemming's solution, this internal granule seems to be surrounded by a narrow zone of less dense material. During the diplotene stage the centrosome can still


•••'■ 0/



Text fig. 1 One of the largest cells from the ovary of a rat three days old. Fixed in Flemming's fluid and stained in iron haematoxylin. X 1830.

Text fig. 2 One of the largest cells from the ovary of a rat nine days old. Fixed in Zenker's fluid and stained with iron haematoxylin and erythrosin. X 1830.


be distinguished in a fairly large percentage of the cells, but it has not been found in any of the nuclei of the later stages. It will be seen that this body is evident throughout the period in the process of synapsis, when there is a distinct orientation of the chromatin threads. This orientation seems to have a definite relation to the position of the centrosome, the bent portion of the loops being, in all clear cases, directed away from that side of the cell where it occurs (fig. 12). As has already been clearly pointed out (von Winiwarter and Sainmont, '08) it cannot be asserted on the evidence here available that this body actually is the centrosome and its enclosed granule the centriole. On the other hand the converse is just as difficult of demonstration.

The chromatoid body here referred to is a deeply staining body of small size appearing, seemingly, in any region of the cytoplasm. It is rarely to be found in the cells with synaptene nuclei. During the pachytene stage (figs. 12 and 15) it is common, and here not infrequently two such bodies may be found in the same cell. From this time on throughout the period involved in this paper, it is a constant feature of a large percentage of the cells. During later stages there may be quite a number of these bodies (text fig. 2). Neither the origin nor the fate of these two structures has been determined.


As a whole the series of changes associated with synapsis in the egg. of the rat rather closely resembles the same period of development in other mammals carefully studied (von Winiwarter, '00, and von Winiwarter and Sainmont, '08). There are, however, some very noticeable differences. Those changes taking place subsequent to the last oogonial division and just previous to the formation of the leptotene threads, are much more elaborate in the rat than in any of the other forms described. In the case of the cat there is some shght suggestion in the figures (von Winiwarter and Sainmont, '08) of such a series of changes but they seem much less complex than in the rat. Particularly in the rabbit and in man there is a complete absence of those transition nuclei with the chromatin in irregular


blocks peripherally arranged, and in none of the three forms described by these authors is there that finely graded series of nuclei showing the fragmentation of these blocks and the simultaneous appearance of linin threads. In the case of the rat, the contraction stage certainly must start earher than in these other forms. It would seem Irom the figures that this stage persists longer in the rabbit and in man (von Winiwarter, '00).

It has already been mentioned that distribution of nuclei in zones — the most advanced stages being in the interior and the least advanced at the periphery with the intervening stages in more or less definite concentric areas — is almost completely lacking in the case of the rat. It would seem that this phenomenon might be correlated with the comparatively shorter period over which these nuclear changes extend in this form.

In the case of the cat (von Winiwarter and Sainmont, '08) there is a clearly defined chromatin element, fairly constant in appearance, persisting throughout this entire series of changes. There is certainly no such body in the case of the rat; and von Winiwarter finds none in either the rabbit or man. In all other respects, particularly in reference to the gross as well as the more detailed appearance of the chromatin threads throughout the entire series of changes, the four forms are closely similar.

Of particular interest throughout this study has been the relation of the chromatin to the linin threads, and of the chromatin nucleoli of the later periods to the synaptene and diplotene threads. During the first part of the series of stages here described, the evidence is very strongly in favor of the conception that fragmentation of the original chromatin blocks and migration of these chromatin fragments along the hnin threads then evident, contribute largely to the formation of the leptotene threads. This process is almost exactly reversed at the other end of the process of synapsis. The chromatin nucleoli seem to be merely the closely compacted clumps of that same chromatin which has now migrated back along linin threads to one or more common points. The contact of the synaptene and diplotene threads with these chromatin bodies, and the persisting linin


threads in the remainder of the nuclei, point to this conclusion. Such a phenomenon has been elaborated upon by McGill in the case of the Dragonfly (McGill, '06), and has been described by numerous other authors.

In regard to those questions which attracted attention in the first place, it seems: (1) There is no evidence, during the period of synapsis, of the presence of an accessory chromosome or sex chromosome. (2) If it be assumed that the individual leptotene threads represent chromosomes, chromosomes seem to pair during synapsis; and this is a side by side pairing, or parasynapsis. (3) Under this same assumption, it would seem that individual chromosomes maintain their identity throughout this period up to and including the diplotene stage, although there is opportunity for mixture of material in the chromatic enlargements during the pachytene stage. There is, however, no satisfactory evidence that these individual leptotene threads do represent chromosomes.


Arnold, Geo. 1908 The prophase in the ovogenesis of Planaria lactea. Arch.

f. Zellforsch, Bd. 3. Browne, E. N. 1910 The relation between chromosome number and species in

Notonecta. Biol. Bull., vol. 20. Cardiff, I. D. 1906 A study of synapsis and reduction. Bull. Torrey Botan.

Club, vol. 33. Davis, H. S. 1908 Spermatogenesis in the Acrididae and Locustidae. Bull.

Mus. Comp. ZooL, vol. 53. DuESBURG, Jules 1908 Les Divisions des Spermatocytes chez le Rat. Arch.

f. Zellforsch., Bd. 1, pp. 399-449. Foot, K. and Strobell, E. C. 1910 Pseudo-reduction in the oogenesis of

Allobophora foetida. Arch. f. Zellforsch., Bd. 5. Von Hoof, Lucien 1912 La spermatogenese dans les mammiferes. La Cellule,

T. 27. Jordan, H. E. 1911 The spermatogenesis of the opossum, Didelphys virginica,

with special reference to the accessory chromosome and the chondrio somes. Arch. f. Zellsforsch., Bd. 7. King, Helen D. 1908 The oogenesis of Bufo lentiginosis. Jour. Morph.,

vol. 19. Long, J. A. 1912 The living eggs of rats and mice. Univ. of Calif. Pub.

ZooL, vol. 9, no. 3. McGill, Caroline 1906 Behavior of the nucleoli during the oogenesis of the

dragonfly with especial reference to synapsis. Zool. Jahrb. Abt. Anat.,

Bd. 23.


McIlroy, a. Lotjise 1910 The development of the germ cells in the mammalian ovary with special reference to the early phases of maturation. Proc. Roy. Soc. Edin., Bd. 31.

Regaud, Cl. 1901 Etudes sur la structure des tubes seminiferes et sur la spermatog^nese chez les mammiferes. Arch. d'Anat. Micr., v. 4.

ScHAPFNER, J. H. 1907 Synapsis and synizesis. Ohio Naturalist, vol. 7.

Skrobansky, K. 1903 Beitrage zur Kenntnis der Oogenese bei Saugethieren. Arch. f. mik. Anat., Bd. 62.

Stevens, N. M. 1911 Heterochromosomes in the guinea-pig. Biol. Bull., vol. 21.

Wilson, E. B. 1911 Studies on chromosomes, VIII. Observations on the maturation phenomena in certain Hemiptera and other forms, with considerations on synapsis and reduction. Jour. Exp. Zool., vol. 13.

Von Winiwarter, Hans 1900 Recherches sur I'ovogenese et I'organogenese de I'ovaire des mammiferes. Arch, de Biol., Bd. 17.

Von Winiwarter, Hans et Sainmont, G. 1908 Nouvelles recherches sur I'ovogenese et I'organogenese de I'ovaire des mammiferes (chat). Arch, de Biol., Bd. 24.

WoDSEDALEK, J. E. 1913 Spermatogenesis of the pig with special reference to the accessory chromosomes. Biol. Bull., vol. 25. 1914 Spermatogenesis of the horse with special reference to the accessory chromosomes and the chromatoid body. Biol. Bull., vol. 27.




Figures 1, 2, 3, 7, 9, 11, 12, 16 and 17 are drawn to the same magnification (X 2290) with a Zeiss 2 mm. 1.4 n. a. apochromatic objective; figures 13, 14 and 15 (X 2270) with a Zeiss 1.8 mm. achromatic objective; and figures 5, 6, 8 and 10 (X 2490) with a Leitz jV achromatic objective; a no. 12 Zeiss compensating ocular being used in all cases.

Figures 1 to 4, 6, 7, 9 to 11, 16 and 17 prepared by Dr. H. M. Gilkey; and figures 12, 18 and 19 by Mr. Willard Shepard.

1 to 6 Successive phases of the transition nuclei between the last oogonial division (fig. 1) and the contraction stage. From young of female 389, fixed in Zenker's fluid and stained with Heidenhain's iron haematoxylin and acid fuchsin.

7 Two nuclei in the contraction stage. Fixed in Zenker's fluid, from young of female 451-6 and stained with Heidenhain's iron haematoxylin and acid fuchsin.

8 Side view of nucleus in the leptotene stage. Material from female 454, fixed in Zenker's fluid and stained in Heidenhain's iron haematoxylin and orange G.

9 Lateral view of a nucleus in the leptotene stage. Material from female 472, fixed in Zenker's fluid and stained with iron haematoxylin and orange G.

10 and 11 Nuclei in synaptene stage, 10 polar view, and 11 a lateral view. Material from female 376, fixed in Zenker's fluid and stained with iron haematoxylin and orange G.

12, 13, 14 and 15 Nuclei in the pachytene stage. Material of 13 from female 376, and that of 14 from female 481, both fixed in Zenker's fluid and stained in iron haematoxylin and orange G. Material of 12 and 15 from young at birth, fixed in Flemming's fluid and stained with iron haematoxylin. Figure 12, an early pachytene, and figure 14 a later, both in side view. 13 is a pole view. 15 a very late pachytene in which polarized arrangement of threads is lost.

16 and 17 Nuclei in the diplotene stage. Material from young one day old. Fixed in Zenker's fluid and stained with iron haematoxylin and orange G.








\ vt





Frovi the Leland Stanford Junior University, California


During the past summer a small number of gastrulae were discovered in the plankton of Monterey Bay, California, which exhibited a unique combination of characters, relating them on one hand to the Echinodermata and on the other to the Enteropneusta. To discover their true relationship all were placed in balanced aquaria where they subsequently developed into bipinnariae. Ova of various species of shore-dwelling starfishes were then artificially fertilized and reared for a sufficient length of time to show that the first captured specimens were the young of Pateria (Asterina) mineata.

Judging from artificially fertihzed material the segmentation and early blastula stages do not exhibit any especially noteworthy features, but beyond this point certain structures arise that have no known counterpart among starfishes. The first of these unique organs is the apical plate. In the late blastula stage the cells about the animal pole commence to elongate and, in the gastrula, form a thickened area, more or less lensshaped in form, having approximately one half the diameter of the transverse axis of the embryo. As indicated in the drawings (figs. 3, 5) its center is exactly opposite to the blastopore and therefore is strictly apical.

Sections show the component cells to possess a height fully tKree times that of the average ectoderm cell of the animal half of the embryo. Passing outwardly the altitude diminishes at a fairly uniform rate until the outer hmits are reached. The nuclei are distally located and therefore similar in this respect




to the ectodermal elements generally. Furthermore all of the cells of the apical plate are packed with granules, evidently yolk, and thus are in marked contrast to the remaining ectoderm where the granules are relatively few and minute.

Generally speaking, this condition of affairs continues to the close of gastrulation. In specimens entering upon the initial stages of enterocele formation, where the walls of the blind end of the archenteron have grown thin, a few of the cells of the apical plate are usually to be seen in the act of migrating into the blastocele. As will be seen later these and other migrants from the apical plate form mesenchyme.

The formation of the enteroceles (hydroceles) proceeds in the typical fashion. In practically every case they arise independently of each other, and from the outset the left is almost invariably the larger of the two (fig. 1). Mesenchyme cells arise as usual from the walls of the enteroceles as well as from the intervening portion of the archenteric wall. Their numbers, when compared with those in Asterias ocracea for example, are relatively small, and, in comparison with the migrating cells of the apical plate, they are distinctly smaller and far more hyaline in appearance — features which enable one to distinguish the two types of mesenchyme (ectodermal and entodermal) throughout the stages described in this paper.

In fully 50 per cent of the specimens in hand a few of the mesenchyme cells arising from the blind end of the archenteron, between the enterocele pouches, unite to form a small vesicle (fig. 1, av). This usually occurs after the enterocele pouches are well differentiated though not completely cut off. In a relatively short space of time this anterior, unpaired vesicle is rapidly enlarged by the addition of other mesenchyme cells from the archenteron, and ultimately attains a diameter equal to that of the gut. In a few surface views it is clearly seen to be entirely enclosed, with walls everywhere complete, and, though it is closely attached to the distal wall of the archenteron, it is nevertheless distinctly separated from it. In other cas6s the vesicle is so closely applied to the archenteric wall that it is not possible to determine whether the vesicular walls are complete. In


every instance, however, the distal wall of the archenteron is continuous, and the evidence is perfect that this anterior vesicle is entirely formed from mesenchyme and not as an outpouching of the archenteric wall.

By the time that the anterior vesicle has attained a diameter equal to that of the archenteron the mesenchyme migration from the apical plate is at its height. Many cells in sections are seen to have become entirely free, while others are in the act of slipping past what are to become the final ectodermal elements, and are making theu' way into the blastocele. In several larvae they come in contact with the anterior vesicle (fig. 1), and cover it like a roof. Where the mesenchyme is more scattered, a relatively small number of cells rest upon the vesicle, while the others pass around it, especially in the region of the future oesophagus.

Beyond this stage in the development, the walls of the anterior vesicle usually disappear, the mesenchymal elements separating and migrating to other regions. In the oldest stage where the vesicle has been found to exist the stomodaeum had completely formed, and the enterocele pouches had severed their connection with the gut. The vesicle in question showed sHght signs of disintegration on its posterior face, but, as figure 4 indicates, it is fairly complete, and has a diameter approximately equal to that of the middle section of the body. Here it is a question whether it has a complete wall next to the gut, but there is no doubt whatever about its lack of communication with the digestive tract.

Surface views under fairly high magnification show the wall of the vesicle to be made up of a relatively small number of cells. These possess highly ramified pseudopodial processes which appear to form an extremely delicate granular meshwork. Whether the meshes are bridged by a non-granular ectoplasm or by some intercellular material has not been determined, but it is evident that the bounding wall, as a whole, is fairly complete, since neither the pseudopodia of other cells nor minute granules which appear to have escaped from some of the apical cells make their way into the cavity of the vesicle. In some


specimens the blastocele contains a ground substance distinctlystained by Delafield's haematoxylin, and is to that extent in contrast with the fluid contained in the vesicle, which has only a slight affinity for dyes of this character.

After cells of the anterior vesicle have separated and migrated into the blastocele, many of the mesenchyme elements from the apical plate move to the wall of the gut, especially in the region of the oesophagus. In the oldest larvae in my possession (about twenty hours later than the stage represented in figure 2) a few of these apical mesenchyme cells, more or less bipolar in form, rest upon the wall of the anterior half of the oesophagus at right angles to its long axis. Between this point and the stomach other mesenchyme cells of entodermal origin are taking up similar positions. Both types of mesenchyme doubtless become transformed into circular muscles.

The left hydrocele, after separating from the archenteron, is not only larger than the right, but its .subsequent growth is more rapid, and it soon develops a pore canal and hydropore which have no counterparts on the opposite side; at least none have been discovered in the material at hand.

A posterior enterocele pouch (fig. 4) arises from the left side of the gut at about the middle of the future stomach. It is a hollow outgrowth, nearly as thick-walled as the gut itself, and after severing its connection with the digestive tract, it becomes a flattened vesicle wedged in between the stomach and body wall. In a stage slightly older than the one represented in figure 2 the left hydrocele forms a diverticulum at its posterior end which ultimately comes in contact with the posterior enterocele pouch and fuses with it. Before this fusion takes place the posterior enterocele vesicle, in several larvae, assumes the form of a rather narrow ellipse with its long axis extending from a point slightly behind the left hydrocele to the base of the intestine. Nothing is known regarding its subsequent history.

Comparisons. Since the year 1869, when Metschnikoff first called attention to the resemblance between the echinoderms and Balanoglossus a vast amount of data has accumulated. This tends to support the original theory that the two groups in ques


tion are genetically related, but the interpretation of certain details is not completely satisfactory. For example there is a lack of agreement relating to the homologies of the coelomic cavities. It will be remembered that Balanoglossus develops five pouches, an anterior unpaired vesicle and two succeeding pairs which become the proboscis, collar and trunk coelom respectively. In Balanoglossus clavigerus, according to Heider ('09) the anterior vesicle arises as a diverticulum from the distal end of the archenteron, and after separating completely, forms the future proboscis pore, before coming in contact with the apical plate. After a time the vesicle withdraws from the plate, though retaining its connection by means of a muscular strand, while mesenchyme cells from the strand and undetermined neighboring regions migrate to the oesophageal region where they become transformed into circular muscles.

The resemblance in the behavior of the anterior vesicle in Balanoglossus clavigerus and Pateria mineata is decidedly striking. In the starfish the vesicle arises in the form of mesenchyme cells which subsequently unite; in Balanoglossus it originates as a direct outgrowth of the archenteron. In Asterias rubens and a few other species of starfishes Gemmill (14) has described the rudiments of posterior enterocele pouches which either fuse with the left hydrocele or break up into mesenchyme (and possibly. unite later with the hydrocele, though this is not indicated). In other words it is well known from this and numerous other instances that there is no fundamental difference between a vesicle formed as an outpouching of the gut and the precocious development of mesenchyme with subsequent fusion. In Pateria mineata the anterior vesicle soon disappears, the component mesenchyme cells wandering off into the blastocele, but before its disappearance it has come into close contact with wandering cells from the apical plate. No pore unites this anterior vesicle with the exterior, but its behavior is such that I am convinced it is the homologue of the proboscis coelom of the Enteropneusta.

There is, so far as I know, no other starfish which develops a distinct, independent anterior vesicle. In several species the blind end of the archenteron becomes expanded, thin-walled and


develops laterally the two well known enteroceles, so that the median section between the pouches may possibly be considered an incipient unpaired vesicle as certain authors have suggested. On the other hand it is equally possible that some of the mesenchyme cells migrating from the distal end of the archenteron represent the anterior vesicle though they may never actually fuse. In Pateria mineata, for example, not over 50 per cent of the embryos form a complete vesicle; in a few cases, relatively, it is imperfect, and it appears probable that in the others it is represented by isolated mesenchyme cells which arise in the proper position but never unite. Furthermore, it is a significant fact that in all of. these cases, even where the anterior vesicle is perfectly formed, the enterocele pouches are separated by a thin section of archenteric wall which therefore can scarcely be regarded as representing the anterior vesicle.

Gemmill's discovery of a pair of posterior enterocele pouches in the starfish is of the highest importance. They have the same origin as the trunk coelom in Balanoglossus, and as Gemmill states "We have here, I think, rudiments of a paired posterior enterocoelic outgrowth, which in the common ancestor of Balanoglossus and the Echinoderms gave rise to the coelom of the body or trunk." In some of the species studied by Gemmill "this rudiment takes no part, or only a small part, in the formation of the wall of the posterior coeloms, but it still retains the function of producing mesenchyme." In Pateria mineata, on the other hand, it is a relatively large vesicle, present on the left side only, and in the future an attempt will be made to determine its fate. It is possible that, in this species and in the others where it forms mesenchyme, this posterior outgrowth does play a more or less important part in the history of the left enterocele from which the hydrocele arises, and that some of the difficulties in homologizing this last named cavity are due to its compound character.

To sum up : I am strongly inclined to look upon the anterior vesicle in Pateria mineata as the homologue of the proboscis coelom of Balanoglossus, while the posterior outgrowth corresponds to the trunk coelom, and the intermediate pair of vesicles


in the echinoderms, often with two hydropores in certain species, is the equivalent of the collar coelom.

A word remains to be said regarding the apical plate. In certain crinoids and echinoids it becomes a well developed organ in fairly old larvae, in which definite ganglion cells appear. As the adult nervous system arises from the ectoderm in close proximity to the hydrocele, it would appear that these embryonic nerve cells play no part in tKe process. Whether they degenerate or become mesenchyme remains undetermined. It is fairly certain, however, that they furnish a strong bit of evidence in favor of the theory that the apical plate of the trochophore larva and of the echinoderm larva are homologous structures.

In Pateria mineata it has been shown that many of the cells of the apical plate completely lose their connection with the ectoderm and, as mesenchyme, migrate through the blastocele to the oesophageal section of the gut where they probably become transformed into circular muscles. All signs that they ever functioned as nervous elements appear to be wholly lost. In this connection one calls to mind the work of Katschenko ('88), Piatt ('94) and others who find that numerous cells leave the neural crest in the head region of various vertebrate embryos and become mesenchyme. It is possible that such cells represent ganglionic elements in the primitive ancestor which formerly innervated preotic structures now vanished or rudimentary. The matter, however, remains unsettled for the vertebrates; such is emphatically the case with the echinoderms where the species studied are few indeed and the data far too scanty to enable one to draw trustworthy conclusions.


Gemmill, J. F. 1914 The development and certain points in the adult structure

of the starfish Asterias rubens. Phil. Trans. Royal Soc, v. 205,

Series B. Heider, K. 1909 Zur Entwicklung von Balanoglossus clavigerus. Zool. Anz.,

Bd. 34, 1909. Katschenko, N. 1888 Zur Entwicklungsgeschichte des Selachierembryos.

Anat. Anz., Bd 3. Platt, J. B. 1894 Ontogenetische Differenzirung des Ectoderms in Necturus.

Arch. mikr. Anat., Bd. 43.



1 Late gastrula or early bipinnaria stage showing mesenchyme migrating from the apical plate (ap) and coming in contact with the anterior vesicle (av). The archenteron bears the usual two enterocele pouches and a single left posterior one.

2 Anterior end of bipinnaria showing the completion of the mesenchyme formation at the apical pole and its migration to the esophageal region. The left hydrocele is provided with pore canal and hydropore. Ip left posterior vesicle.

3 Section of gastrula showing apical plate with signs of mesenchyme formation.

4 Bipinnaria showing mesenchjTne migrating past the anterior vesicle (av) to the esophagus. The left hydrocele has severed its connection with the archenteron. Ip left posterior vesicle in process of formation.

5 Surface view of gastrula immediately before mesenchyme development from apical plate.








Department of Zoology, University of Pennsylvania



I. Introduction 472

II. Observations 474

A. Chorthippus (Stenobothrus) curtipennis 474

a. Introductory statement 474

b. The post-spireme chromosomes 475

c. Synapsis 476

d. Chromatid movements 479

e.. The planes of the first division 480

f. The apical constrictions and the chromomere-vesicles 481

g. The accessory chromosome 483

h. Summary 483

B. Trimerotropis suffusa 484

a. Introductory statement 484

b. The post-spireme chromosomes 484

c. The evidence on synapsis 486

d. Chromatid movements 487

e. The planes of the first division 490

f. Additional observations on organization 492

g. Summary 493

III. Discussion 494

A. Synapsis 494

B. Pre-reduction vs. post-reduction 498

C. Chromosome organization 505

a. Atelomitic chromosomes 505

b. The chromomere-vesicles 507

D. Summary of conclusions 509

IV. Literature cited 510




In a recent paper (Wenrich, '16) the writer presented what seemed to him conclusive evidence that parasynapsis occurs in at least three pairs of first spermatocyte chromosomes of Phrynotettix magnue. Although only these three chromosomes were studied in great detail, sufficient attention was given the other members of the complex to assure the writer that parasynapsis occurs in all. Phrynotettix, however, like many of the Oedipodinae, possesses only rod-shaped, or telomitic chromosomes. It seemed advisable, therefore, to investigate the conditions to be found in species of grasshoppers in which the chromosomes are V-shaped, or atelomitic. Chorthippus (Stenobothrus) curtipennis was chosen for this study because it is readily obtainable, gives clear figures when fixed and stained, and because various species of the genus have been the subject of investigation in the past. Sufficient study of this species, which has three pairs of atelomitic chromosomes, was made before the publication of the paper on Phrynotettix to assure the writer that the chromosome relationships in synapsis are essentially the same as in that form with respect to both types of chromosomes. This is indicated by the following statement (Wenrich, '16, p. 98):

I have recently made a study of the conditions in Stenobothrus and may say that I found parasynapsis for both forms of chromosomes, and that the V-shaped chromosomes divide reductionally in the first maturation mitosis as Davis ('08) described, but that the rod-shaped chromosomes divide equationally in the first division as I found that they did in Phrynotettix.

No drawings were presented to support these statements at the time, but the evidence (with additions) on which the conclusions as to parasynapsis were based is presented in the present paper. Unfortunately, more careful study of the material makes it impossible to be so confident as to which division is reductional for the two types of chromosomes. This point mil be discussed later.

The same conclusions as to parasynapsis in Chorthippus, together with the same kind of evidence, in part, has recently


been presented by Robertson ('16). In many respects the evidence and conclusions which have been independently reached by Robertson and myself will be mutually corroborative. In regard to some points, however, differences of interpretation exist which will be discussed in an appropriate place.

In view of the very interesting conditions in Trimerotropis and Circotettix, where telomitic chromosomes are paired with atelomitic ones, as found by Dr. E. Eleanor Carothers ('17) it seemed worth while to examine stages which would indicate what form of synapsis occurs between these chromosomes of diverse form. Dr. Carothers very kindly permitted the use of her slides for this purpose. It may be stated at once that parasynapsis was found to occur for all the chromosomes of Trimerotropis of whatever form, just as it does in Chorthippus and Phrynotettix. In addition to the subject of synapsis the topics of pre- and post-reduction and chromosome organization will be considered.

In this paper new terms will be used in accordance with those recently adopted by Dr. Carothers ('17), as follows:(l) Telomitic — chromosome with terminal fiber-attachment, i.e., rod-shaped; telomitic and rod-shaped will therefore be used interchangeably; (2) atelomitic — chromosome with non-terminal fiber-attachment = V-shaped; V-shaped and atelomitic will be used interchangeably; (3) heteromorphic — those pairs of chromosomes (tetrads) of which one member is telomitic and the other atelomitic, or, of which one member is noticeably different in form or size from its mate; (4) chromomere-vesicle — ^the plasmosome-like appendages formerly called 'vesicles' by Carothers ('13), and called appendages and plasmosome-like structures by the writer ('16). I shall also make use of McClung's ('14) term, euchromosome as equivalent to Montgomery's 'autosome.'

All drawings have been outlined with a pencil under the camera lucida and then inked while the cell remained under observation.


II. OBSERVATIONS A. Chorthippus (Stenobothrus) curtipennis

a. Introductory statement. Chorthippus (Stenobothrus) curtipennis belongs to the sub-family Truxalinae of the Orthopteran family Acrididae. The material from which nearly all the drawings were made consists of a single smear prepared during the summer of 1914 at Woods Hole, Mass., near which place this species was abundant. Through the kindness of Professor McClung I was afforded facihties for preparing and staining cytological material at that time. The smear was fixed in Flemming's stronger solution (at a temperature of about 4°C.). One part was stained with Zwaardemaker's safranin, the other with Heidenhain's haematoxylin. Figures 12, 13, and 14, plate 3, are from sectioned material, fixed in cold Flemming's fluid and stained with Heidenhain's haematoxylin.

While- this material was studied some during the summer of 1914 and at Harvard University during the winter of 19141915, it was carefully re-examined in 1915-1916 at the University of Pennsylvania and again at Woods Hole during the summer of 1916, when most of the drawings were made. I am indebted to the University of Pennsylvania and to the director of the Marine Biological Laboratory for facilities for pursuing this study at Woods Hole.

Although material from a number of individuals was studied, nearly all the drawings are made from a single smear derived from one animal as stated above. It was deemed desirable to represent conditions from a single specimen in order to avoid any possible confusion arising through individual variations, since recent studies on such forms as Phrynotettix, Trimerotropis, Hesperotettix, etc., have shown that individuals within a species may vary as to certain features of their chromosomal constitution, but the conditions in each indi\ddual remain constant. In the second place, a smear was used in order to insure the presence of whole complexes in each cell and thus to avoid the difficulties incident to the study of sections, where most of the cells are cut and distributed into two or more sections with frequent


disarrangements or cuts in the chromosomes. Some distortions natm-ally arise as a result of the smearing process, but the preparation used was more than usually free from such disturbances and in this case the advantage gained through the study of whole cells far outweighs the disadvantages arising from the smearing process.

In plate 1 all of the chromosomes in each cell are represented. It is believed, therefore, that, although identifications of any chromosome in any cell might sometimes prove doubtful, the chances for error, with all members present, is reduced to a minimum. This method also has its disadvantages, because only occasionally are all of the chromosomes in any particular cell in favorable positions for drawing.

b. The post-spireme chromosomes. The post-spireme stages have received the most attention and are represented on plate 1. As a glance at the plate will show, there are nine chromosomes in the reduced series. This number has been reported for all the species of the genus which have been studied except the doubtful member reported by McClung ('14) and which he now thinks may be Circotettix in which there were eleven. In plate 1, the chi'omosomes of eight cells are represented, those from each cell being arranged in a vertical column according to size, mth the smallest at the bottom. The cells are lettered A, B, C, etc. Each horizontal row therefore presents examples of a particular chromosome, and successive stages in development from the end of the spireme stage to early anaphase can be read from left to right. These horizontal rows correspond in developmental stages to figures 62 to 65, plate 6, of the paper on Phrynotettix.

Of the nine chromosomes in the first spermatocyte one is the unpaired accessory chromosome (no. 3) ; the others are tetrads, consisting of three pairs of atelomitic (nos. 7, 8, and 9) and five pairs of telomitic chromosomes (nOs. 1, 2, 4, 5, and 6). Column A represents the chromosomes of a cell at the stage when the spireme of the growth period has just become distinguishable as indi\ddual segments, each segment being the direct forerunner of a tetrad (except the acccessory, no. 3). Each of the segments is


an attenuated thread with a longitudinal cleft, and more or less coiled or bent, according to its length. The longitudinal cleft is what has been called the primary longitudinal spht.

In cell B, a second split is seen at some points, where four threads are visible (e.g., at a). In cell C all of the segments (except no. 1 and the accessory) have the four chromatids, indicating that the secondary longitudinal split has become complete. The other cells, D to H, represent the farther progress of chromatid transformation and condensation, through characteristic tetrad figures, to the metaphase or early anaphase of the first maturation mitosis. As nearly as possible, all of the tetrads in cells D to H have been given the same orientation; those in cells A, B, and C could not be definitely oriented, partly because the synaptic points could not be determined and, in the case of numbers 7, 8, and 9, partly because of lack of space on the plate.

c. Synapsis. The point to which especial attention should be directed is that all of the autosomes, or euchromosomes, show the same fundamental conditions, whether they be pairs of telomitic or atelomitic chromosomes. All come out of the spireme as elongated threads with one longitudinal cleft visible; all develop a second longitudinal cleft at right angles to the first, giving rise to tetrads, each with four chromatids. And since, as is generally believed, each tetrad represents a pair of spermatogonial chromosomes, it would appear that each of the spireme segments in cell A represents a pair of chromosomes which have been extended axially and united side-by-side throughout their length.

I have already shown ('16) how this side-by-side union of attenuated leptotene threads takes place in Phrynotettix. Evidence of the same nature for Chorthippus is presented in figures

1 to 4, plate 3. In figure 1 a nucleus is shown with one double thread (5) — that attached to the accessory — and another thread which is double at the end nearest the accessory (proximal end) but divides distally into two single threads. Numerous single threads are visible in the other parts of the nucleus. In figures

2 and 3 a considerable number of these partially conjugated pairs of threads are shown. In none of these drawings (figs. 1 to 4) was it possible to represent all of the threads because they


are so long and so extensively curved and coiled that they appear at first glance to be hopelessly tangled and to constitute a network. It was possible, however, by careful study, to follow some of the threads for considerable distances, but others only short distances. On this account these drawings are, in part, diagrammatic.

In these lepto-zygotene nuclei there is always a tendency toward an orientation of the threads just as was found at the corresponding stage in Phrynotettix ('16, figs. 29, 30, 31, plate 3). This orientation appears to be confined mostly to the region near that end of the thread to which the spindle-fiber is attached (proximal end) and therefore that side of the nucleus toward which these ends are directed is called the proximal side. From this proximal side the threads run more or less parallel toward the opposite side of the nucleus, thence turning to traverse the nuclear space in various curved or convoluted courses. The amount of curving depends, apparently, on the length of the threads, though even the shorter ones have a tendency to form a loop as is indicated by the short, deeply staining thread (5), attached to the accessory chromosome in figures 1 and 4. In general, only the proximal ends of the threads exhibit orientation and this orientation, I believe, facilitates th-e initiation of the pairing process and causes its inauguration to occur at the proximal ends, as shown in figures 1 to 3. In the distal part of these figures most of the -threads are single. The accessory (X), usually has one end attached at the proximal pole, as shown in figures 2 and 3. In figure 4 the position of the nucleus was such that the orientation could not well be shown without interfering with other features which it was desired to represent.

The nuclei in figures 2, 3, and 4 were chosen for drawing because they show, in addition to the partially conjugated threads already mentioned, certain others to which small plasmosomelike bodies, or chromomere-vesicles, were appended. These structures, (c), as will be pointed out later, mark the synaptic points on one of the pairs of V-shaped chromosomes.

The question has naturally arisen as to the behavior of these V-shaped chromosomes in synapsis. Does the process of conju JOURNAL OF MORPHOLOGV, VOL. 29, NO. 2


gation begin at the apices of the V's and proceed toward the ends, or does it begin at the free ends and proceed toward the apices? The drawings shown are not entirely conclusive on this point, though very suggestive. In figure 2 the two synaptic points are marked by deeply stained knobs to which the 'vesicles' are attached. These knobs are not situated at the proximal side of the nucleus, but appear to have been pulled away from this position which they undoubtedly held in the telophase of the last spermatogonial division (figs. 7 and 8). The knobs are at the apex of an angle made by a sharp bend in the threads at that point, and while the knobs themselves stand somewhat apart, the threads appear to be conjugated for a short distance on either side of them. When I attempted to follow these conjugated threads they soon became lost in the maze of unconjugated, single threads. The appearances, however, indicate that the only parts of these threads which had conjugated at the time were those immediately adjacent to the appendage-bearing knobs, and therefore at the apices of the V's.

In figure 3 a situation is presented somewhat similar to that in figure 2, except that here conjugation has taken place on only one side of the apical knobs, the remainder of this pair of threads apparently consisting of unpaired leptotene strands. In figure 4, conjugation of this pair has gone further, but even in this case the conjugated part is confined to the region on either side of the synaptic points. These cases, together with others which have not been drawn, tend to support the idea that conjugation of the atelomitic chromosomes begins at the apex of the V, which, as the point of spindle fiber attachment, is homologous with the proximal end of the telomitic or rod-shaped chromosomes.

As soon as conjugation is complete, all of the chromosome threads appear d.ouble, as shown in figure 5, which represents the pachytene stage. In this nucleus, the usual orientation of the chromatic elements has been lost. There is some tendency for the threads to be centered toward the side of the nucleus which is uppermost in the drawing, but the accessory chromosome (X), which is customarily attached at the proximal pole


of the nucleus, is here removed a considerable distance, and separate from the deeply stained loop (5), to which it is usuallyattached. This deeply-staining thread is found in the later stages to be that of chromosome 5 (see plate 1) and it very commonly lies near, and often attached to, the accessory. This association may be seen in figures 1, 4, and 6.

Figure 6 represents a nucleus at about the same stage as cell A on plate 1 and shows the separate spireme segments. Each of these has but one longitudinal split, but each represents a tetrad. The segments are numbered according to the arrangement on plate 1. A series of successively older stages may therefore be recognized, beginning with figure 1, extending through figure 6 on plate 3, and continuing through cells A to H on plate 1.

The series of stages just mentioned constitute, it seems to me, a sufficient body of evidence to indicate that in Chorthippus parasynapsis is the rule.

d. Chromatid movements. Attention may be directed to the different rates of change in the chromatid relationships experienced by different chromosomes in the . same cell and by the same chromosome in different cells. The typical form of the metaphase chromosome " of the telomitic type seems, in this species, to be the rod extended parallel to the spindle axis. Such a form is taken by chromosomes 1, 2, 4, and 6 of cell G on plate 1. Although number 5 has not been pulled out in the direction of the mitotic poles sufficiently to make it a rod, it is readily apparent that only a small amount of movement of the oppositely directed pairs of chromatids would produce the forms seen in chromosomes 4 and 6. As an example of the different degrees to which the chromatid movements have brought about an approach to the extended-rod condition of the metaphase, attention may be called to cell E, where chromosome 4 seems to be lagging much behind the other chromosomes, though in cells F, G, and H, such is not the case. In cell D, on the other hand, chromosomes 2 and 5 seem to be farther advanced toward the metaphase-rod condition than any of the others.


When we examine a particular chromosome as it appears in different cells we do not always see a consistent progress in development in the successively later stages. Chromosome 4 in cell E seems to be in about the same condition, so far as chromatid movement is concerned, as it is in cell C; that is, either one could, by the simple rotation of the free arms, assume the condition seen in cell G. However, it is just possible that in cell D a valid intermediate stage is represented and that this chromosome in the condition found in cell C would have to pass through a stage such as that in cell D before reaching the condition seen in F or G.

e. The planes of the first division. In the absence of wellmarked polar granules it is impossible to determine which are the proximal, or synaptic, ends of the chromatids, but if we may assume for the time being that the ends at the left in cells B to E are the synaptic ends, we may trace for chromosome 4 successive steps as follows: In cell B the longitudinal split on the right (the only one \dsible for two-thirds the length of the tetrad) is probably the primary longitudinal split. The split just forming at the left and continued to the point a is probably the secondary split. In cell C, then, we may consider that the formation of the secondary split has been completed and that the chromatids have separated along this split almost to the distal end. . Such an element might remain in this condition, condensing through a stage such as is shown in cell E and finally opening out to form an element in the metaphase like the one in cell G. In cell D we may have a stage in which the secondary split has become complete but the separation along its plane has not. If the progress of separation along this plane were to be impeded by gi^adual condensation, this stage might be followed by a cross formed by the proximal ends (left) of the two pairs of chromatids becoming extended in opposite directions in the same plane (as shown in chromosome 2, cell D), which cross might reach the metaphase in some such form as shown for chromosome 5 and 6, cell F, or even chromosome 4, same cell. A similar series of stages occurs for chromosome 2 in cells B, C, and D. Series of movements such as have been suggested would corre


spend in all essentials to the series of changes found to occur for chromosome A in Phrynotettix (see '16, fig. 62, plate 6), and would indicate an equational division. Other interpretations, indicating a reductional di\dsion are, of course, possible, and cannot be excluded.

In chromosome 6 a similar series of stages may be traced. In cell B the split running through most of the length of the element from the right is probably the primary split, while that just beginning at the left and extending to point a would be the secondary split. A continuation of the secondary split and a separation along its plane would give rise to such a stage as that seen in cell C, or, if the separation were delayed, to that in cell D. A rotation of the arms of the tetrad in cell C would give rise to a stage such as is shown in cells E and G, while a similar rotation of the arms of the example in cell D would give rise to the condition seen in F.

It is regrettable that the polar granules, which do occur to some extent, are not sufficiently differentiated by the staining method used to make the orientation of the tetrads certain. In sectioned material stained with iron-haematoxylin polar gi-anules are, in fact, better differentiated than in the safranin stained smear, but in this instance little study has been made of sections, where identification of all the chromosomes is not so certain.

In the case of the pairs of atelomitic chromosomes (nos. 7, 8, and 9) variations in rate and extent of chromatid movement are not so apparent in the examples drawn on plate 1. Such variations do exist, however, as will be seen by comparing number 8 mth three 'rings' in cell C, and two 'rings' in cells D, E, and F, and figures 9 to 14, plate 3.

/. The apical constrictions and the chromomere-vesicles. In these atelomitic chromosomes the synaptic point is in many cases very clearly marked by a constriction, as Robertson ('16), has emphasized. This constriction (point h in the figures, plate 1) is not visible in all cases, as is indicated in the figures. Chromosome 9 in cell E for example, shows scarcely any evidence of a constriction, although it is plain enough in chromosomes 7


and 8. In cell F chromosomes 7 and 9 show the constriction, while it is not apparent in chromosome 8. Again in cell G chromosome 9 did not reveal the constriction as did chi'omosomes 7 and 8. All three show evidences of a constriction in cell C.

In connection with this synaptic point, as already mentioned, there is a small plasmosome-like appendage, or chromomerevesicle (marked c in the figures), which seems, in the earlier stages at least, to be as constant in its occurrence as is the constriction just described. It disappears in the later tetrad stages as such structures usually do. Similar appendages are occasionally seen attached to the synaptic points of chromosomes 7 and 8, as is indicated for the former in cell D and for the latter in cell A, but their occurrence on these chromosomes is rather infrequent.

It will be noticed that only one appendage is to be seen on a smgle chromosome in the figures on plate 1. In earlier stages, however, there are two such bodies as is indicated in figures 2, 3, 4, 6 7, 8, and 9, plate 3. Figure 5 shows the two partially fused together and many times they appear to be entirely fused at these stages. It could not be determined whether the single appendage in plate 1 and in figures 10 to 14 on plate 3 represent two fused together, or whether one of them disappears before the other. The latter possibility is suggested by a constant difference in size in all cases where two are in evidence and by the fact that the single ones in figures 10 and 11 do not appear to be equal in bulk to two such as are shown in figure 9. In figures 12, 13, and 14, they appear larger, but these figures are taken from sections from another individual, which fact may account for their larger size. That these structures are constant elements of this pair of atelomitic chromosomes is well illustrated by the spermatogonial telophases shown in figures 7 and 8. In each case the 'vesicle' is attached at the apex of a V-shaped chromosome with a marked disproportion between the lengths of the arms, comparable to that seen for chromosome 9 on plate 1. In figure 8, the smearing process has resulted in a slight distortion of one of the atelomitic chromosomes (at the left) to which an appendage is attached. It seems fairly safe to


regard these chromomere-vesicles as diagnostic for the apices in this largest pair of V-shaped chromosomes.

In -the figures from sections (figs. 12 to 14, plate 3) the appendage is attached to only one member of the pair of knobs which mark the synaptic point, and, further, the attachment is always to the knob nearer the middle of the chromatid. In the smears this position is not so evident owing to the lack of definite knobs showing at the synaptic points. However, chromosome 9, cell A, does show such knobs and it will be observed that the appendage is attached to the one nearer the middle ot the chromatid. Chromosome 8, cell A also shows such a relationship. The possible evidence that these structures furnish toward an understanding of the planes of first spermatocyte di\dsion will be considered in the discussion.

g. The accessory chromosome. The accessory chromosome does not require description or discussion in this paper, but attention may be called to the vesicular appendage attached to it. This appendage is similar to those just mentioned which are attached to the atelomitic chromosomes. This structure seems to have been overlooked by all those who have studied this form, except Davis ('08).

h. Summary. In summarizing the observations on Chorthippus we may call attention to the follomng:

1. The number of chromosomes in the reduced series (first spermatocyte) is nine, consisting of one unpaired accessory chromosome and eight pairs of autosomes or euchromosomes. Of the latter there are tliree pairs of atelomitic and five pairs of telomitic chromosomes.

2. The leptotene threads conjugate two-by-two throughout their length during the zygotene, or lepto-zygotene, stages. They remain conjugated through the pachytene stage and until the tetrads are formed. This condition of parasynapsis holds for all the euchromsomes, both telomitic and atelomitic.

3. The process of parallel conjugation (parasynapsis) appears to be inaugurated at the proximal ends of the leptotene threads as was the case in Phrynotettix, and it appears to begin at the apices of the atelomitic chromosomes.


4. A constriction occurs at the synaptic points in most of the tetrads of the atelomitic chromosomes.

5. A small plasmosome-like appendage, or chromomerevesicle, is attached to the conjugants of the largest pair of atelomitic chromosomes (no. 9) at the apex of the V (synaptic point. A similar appendage is occasionally found at the homologous point on the other two atelomitic pairs, and occurs very constantly on the accessory chromosome.

B. Trwierotropis suffusa

a. Introductory statement. Trimerotropis belongs to the subfamily Oedipodinae of the family Acrididae. A discussion of some systematic problems connected with this genus will be found in the paper by Dr. Carothers ('17). The material from which the drawings were made consisted of a single smear made by Dr. Carothers and loaned to the writer for study. This indi\ddual corresponds to her number 34 (plate 6) (T7) and cell E on my plate 2 is the same as that drawn by her. Other individuals were studied but no drav/ings from them are used in the plates.

The reasons for using a smear preparation in this case are the same as those already stated in connection with Chorthippus. This smear was fixed in Flemirdng's solution and stained by Flemming's tri-color method. The chromatin was not so deeply stained in this smear as it was in the smears of Chorthippus, and this difference appears in the drawings.

b. The post-spireme chromosomes. As Carothers has found, the reduced number of chromosomes in this genus is twelve. Occasionally, however, supernumerary chromosomes occur, as she discovered, and one such supernumerary occurs in this individual, making the number of first spermatocyte elements thirteen instead of twelve.

The arrangement of the chromosomes in plate 2 is the same as for Chorthippus in plate 1. Each vertical column represents the complete series of chromosomes from a single cell arranged in the order of size with the smallest at the bottom.


The differentiation on the basis of size is not always easy to make. This difficulty is made more apparent by the differences between the order used on this plate, where the accessory is made number 3, and the order used by Carothers, who made the accessory number 4. Such differences, however, are without significance so long as other means of identification are available to obviate confusion. The supernumerary is placed at the top of the plate to avoid any possibility of its becoming confused with the smaller tetrads. From left to right are shown successive stages in the transformation of the spireme segments to tetrads, and their condensation, final orientation, and partial division in the spindle of the first maturation mitosis.

A glance at the metaphase figures shows that the complex of this individual consists of four pairs of atelomitic chromosomes (nos. 7, 10, 11, 12), five pairs of telomitic ones (nos. 2, 4, 5, 6, 9), two heteromorphic pairs (nos. 1 and 8), the atelomitic accessory (no. 3), and the supernumerary. In other individuals of the species the relative numbers of pairs of atelomitic, telomitic, and heteromorphic pairs may be different, as found by Carothers and described by her.

It is not possible in every case always to identify each chromosome in a cell. This is due to similarities of size and organization. Chromosomes 11 and 12, for example, appear very similar in size and organization in the earlier stages, so that it is not always possible to tell which is which. In the metaphase and anaphase, however, these two are distinguishable by the different proportions between arm-lengths of the Vs. Number 12 has the greatest disproportion between arm lengths; in number 11 the disproportion is less, and in number 10 still less. The proportions between arm-lengths for these atelomitic chromosomes can the more readily be determined by reason of the occurrence in many of the metaphase examples of constrictions or transverse clefts at the points of spindle-fiber attachment. Such clefts are to be seen, for example, in cells E, F, and G for chromosomes number 11 and 12, and cells E and F for number 10. Such clefts or constrictions (marked b in the figures) can sometimes be distinguished in the earher stages


as, for example, one cleft for number 12 in cell D, two clefts for number 11 ill cell D, two clefts for number 10 in cell C, and in the case of the atelomitic member of number 8 the cleft shows very plainly in cell D. Clefts are like\\dse indicated for number 7 in cells C and D, and still more plainly in the metaphase cells, E and G. These clefts or constrictions appear to correspond in all respects to those found in the ateloftiitic chromosomes of Chorthippus.

Chromosomes 5, 6, and 7 are somewhat similar in organization, each possessing polar granules at both ends. This gives rise to some uncertainty as to identity, especially in cells B and C. In cell A the differences in length are more apparent.

Number 4 in cells A, B, and C, and number 8 in cells A and B are readily identified by the pecuhar clusters of granules (c) attached to them. In the later stages these disappear and this gives rise to some difficulty in identification in the case of chromosome 8. Number 4, however, possesses an additional means of identification by reason of its staining power which is constantly gTeater than in the other euchromosomes.

In spite of the difficulties of identification, it is felt that with all the chromosomes present in each cell, and with a careful consideration of the size and structure relations of the various chromosomes, very few mistakes in serial arrangement have been made.

c. The evidence on synapsis. The evidence for parasynapsis, as presented on plate 2 is of the same nature as that for Chorthippus, and consists in the facts: (1) that all the euchromosome pairs, irrespective of the point of spindle-fiber attachment, come out of the growth period as longitudinally split, attenuated threads, and (2) that from these spUt thi'eads tetrads arise by the formation of the secondary longitudinal spht. Figure 21, plate 3, gives the evidence for parasynapsis as it occurs in the lepto-zygotene stages and the conditions shown are quite similar to those already described for Chorthippus and shown in figures 1 to 4. Figure 21 does not represent nearly all of the chromatin threads in the nucleus and is partially diagi-ammatic. However, the condition indicated — namely, the presence of double threads


with one end at the proxinaal pole of the nucleus and the other end separating distally into two single threads — was clearly seen for a number of those represented and indicated for the others. The deeply staining body near the accessory (X) is regarded as the supernumerary chromosome. In this stage in this material none of the atelomitic chromosomes could be identified and, while the tendency for conjugation to begin at the proximal ends of the threads is clearly indicated, the behavior of the atelomitic chromosomes could not be followed.

d. Chromatid movements. Differences in rates of development, or progress toward separation at anaphase, in different cells find some striking examples in the series drawn on plate 2. Chromosome number 1 seems to be lagging somewhat behind the others in cells B and C, though in cell D it is as far advanced as any. In cell D, number 2 is not so far along toward its typical metaphase condition as it is in cell C. The same relations appear to hold for chromosome 5 in these two cells, yet all of the chromosomes in cell D are more condensed and therefore presumably nearer metaphase than they are in cell C.

In the metaphase there are also some striking differences. Number 4 in cell G is very far behind the other members of the complex and seemingly little advanced in chromatid movement beyond the condition seen in cell A. In cell F, too, separation is not so far advanced for this chromosome as it is in cell E. Cell H is in early anaphase, and cells E, F, and G are at about the same metaphase stage, yet chromosome 9 is not so far advanced toward complete separation in F as it is in E, and in E it is not so far along as in G, yet in G it seems further along than in H. Again number 6 is lagging in cells F and G, as compared to the conditions in E and H. The examples of this chromosome in cells F and G are at about the same stage, but in G a ring exists through the contact of the ends of the arms at the right which are seen to be free in cell F.

All these differences seem to indicate that the extent of chromatid movement is variable, not only for different chromosomes but also for the same chromosome in different cells, so that in some cases the movement which ends in the separation accom


plished in the anaphase may be well advanced before the element reaches the metaphase, while in other cases these movements are largely deferred until after the element has been oriented in the metaphase spindle.

There are some pronounced differences in shape of the atelomitic chromosomes in the metaphase. In the case of chromosome 12, for example, a different shape is seen in each of the cells, E, F, G, and H. It is easy to see how a simple movement of the oppositely directed dyads in cell E toward their respective poles would give rise to the condition in cell H ; but no such simple movement would account for the shapes in F and G. The explanation of these various shapes, I think, may lie in the statement made above as to the variations in chromatid movement in the earlier stages. A search among the tetrad figures of these earlier stages should reveal the forerunners of the shapes seen in metaphase. In cell C, for example, is a condition of chromosome 12 which would readily give rise to that in cell E if the shorter arms of the V's are toward the right and the apices of the V's are pulled out toward opposite poles. Similarly for chi'omosome 11, the tetrad in cell C might easily take its place in the spindle in the form shown in cell E and the tetrad of number 10 in cell D could, by orientation, assume the shape in cell E. In cell D, on the other hand, occurs a condition of number 12 whic-h in many respects is similar to that in cell F. The chief difference is that in the latter the short arms of the V's have been reflexed toward the longer arms. The tendency for the shorter arms to bend back in this way when they are free is not infrequent among the atelomitic chromosomes, as shown by McClung ('14). A somewhat similar metaphase of number 11 or 12 from another cell is shown in figure 18, plate 3. Figure 17 shows an earlier tetrad which could give rise to it by the separation of the ends on the right and their flexure backwards toward the left.

None of the earlier stages for chromosome 11 or 12 as seen on plate 2 indicate a condition which could by any simple movement give rise to the shapes seen in cell G. Here the curious condition is found of the shorter arms remaining in contact,


while the longer arms have become free. From the standpoint of simple mechanics it would be difficult to understand how a tetrad such as is seen in cell C, for example, could give rise to the condition in cell G. It would be natural to suppose that the crossed longer arms (at the left) would require a greater amount of force for separation than the shorter arms in contact at the right; the pull of the spindle fibers would be expected to be exerted more strongly on the shorter ends. The conditions in cell G are the exception rather than the rule, and I have not found an earlier tetrad which showed the longer arms free as they are in this cell. However, in figure 19, plate 3, there is a ringshaped tetrad which, by the mere separation of the lightly attached ends of the long arms at the left, would pro\dde a condition analogous to that in figure 20, which resembles those in cell G, plate 2. A condition which is the reverse of that in figure 20 is shown in figure 16, where the short arms are free. Such a shape could readily arise from that in figure 15 by a little more separation along the plane between the arms at the right and their flexure at the point of spindle-fiber attachment.

It might be remarked that one cannot be certain whether chromosome 11 or 12 is represented in figures 15 and 17 because the synaptic points are not indicated by clefts as they are in figure 19. Figure 16 is doubtless number 12, with its great disproportion between arm lengths, while figure 20 is more probably number 11. Figure 18 is problematical.

In cells E and G chromosome 7 (plate 2) seems to have the shorter arms united and the longer arms free, while in F the longer arms are attached to each other and in H they have been the last to separate. In cell E the difference between arm lengths seems much less than in the other cells. This is to be accounted for by the foreshortening of the longer (free) arms and by the stretched conditions of the attached shorter arms, shown by their smaller diameter.

As a net result of such observations one may conclude that the movements of the constituent chromatids of a tetrad are subject to great variation and that with respect to the separation of the ends, there is sufficient e\'idence to show that either the


longer or the shorter ends may become free in advance of the other end. Just why the chromatids should assume the varying relationships to each other in similar stages of the tetrad transformation period is of course not easily determinable, but the metaphase figures seem to be derived from similar ones in the earlier stages wdthout much movement of the chromatids during the later condensed prophases, nor until the forces in metaphase and anaphase are at work. One may suppose that the movement of the chromatids takes place at different rates in different cells and that, as condensation in the later prophases progresses, such movements are retarded; the result being that whatever stage in transformation has been reached in a late prophase is the one in which the element enters the spindle. The further movements then must be the result of the forces in operation during metaphase and anaphase.

e. The planes of the first division. In Chorthippus, as described on pages 480 and 481 the evidence seemed to favor an equational division for chromosomes 2, 4, and 6 (plate 1) which are telomitic. E\ddence of equational division among the telomitic chromosomes of Trimerotropis can be found on plate 2. Chromosome 2, for example, as it appears in cell A, has a very characteristic arrangement for the parts of this element. It has very pronounced polar granules at both ends, those at the distal end being the larger. That, the larger granules represent the distal end is shown very well in the metaphase figures E, F, G, and H. A fm-ther indication that these larger granules are at the distal end is their behavior toward each other. In all of the cells studied which were at stages corresponding to cells A and B, these large granules stood widely apart from each other. The example in cell A shows this especially well. It would hardly be supposed that the proximal ends of a pair, being the first to conjugate, would behave in this manner, while the distal ends might easily fail of complete pairing. Now, since it is possible to determine which end is proximal and which distal, it is also possible, I believe, to trace the entire series of changes into the metaphase. In cell A the primary spht is visible at the distal end- (at the right in the figure) of this chromosome, but


not through the middle. A spUt is seen in the proximal end (at the left) but it cannot be determined whether this represents primary or secondary split; in cell B, however, both sphts are easily seen and it is further observable that the separation at the proximal end (left) is along the plane of the secondary split. The tendency for the proximal gi-anules to diverge with further separation along the secondary split is seen in cell C. In cell D the relation of the chromatids is much the same as in cell B but they are more condensed. In cells E, F, and H, as I have already noted, smaller knobs or enlargements occur at the proximal ends, while a bulging at the middle of the oriented tetrad indicates the larger distal granules. It seems fairly safe, then, to assume that a tetrad in the condition seen in cell C would become oriented in the metaphase without further change in the relationships of the chromatids, except their further separation along the plane of the secondary split. There seems to be good reason for concluding, therefore, that division is equational for this chromosome.

There is also some e\idence that chromosome number 4 divides equationally in this first division. In the metaphase cell E the horizontal components of the chromatids on the right side of the tetrad have stained differently from those on the left. They failed to retain the stain so well and appeared finely granular, as indicated in the dra\\dng, in contrast to the compact, deeply staining chromatids on the left. It is reasonable to suppose that those two chromatids which show this peculiarity belong to the same parent chromosome, and, therefore, that halves of chromosomes are being separated in this division, constituting an equational di\dsion. Attention may be directed to the cluster of granules attached to this chromosome. In the examples in cells A, B, and C, plate 2, it cannot be determined just what relation this appendage bears to the chromatids. In figures 22 and 23, plate 3, however, it is plainly indicated that the structure is attached to a chromomere of only one of the conjugated threads. Presumably this condition holds true for all stages. If that be true, then these appendages may be at the distal end, if it be permitted to associate the peculiar stain


ing condition of the example in cell E with the presence of this appendage in the earher stages.

In the cases of chromosomes 5, 6, and 9, there is no method of distinguishing distal from proximal end in the earlier stages and therefore no deductions can be drawn as to which plane of separation prevails in metaphase.

For the heteromorphic pairs of chromosomes (nos. 1 and 8) there seems to be little doubt that the first division is a segregating one, because of the striking difference in shape of the segregating parts. In this respect they are analogous to the unequal pairs found in Phrynotettix and other forms which divide reductionally in the first division.

/. Additional observations on chromosome organization. Some additional points of interest on the subject of chromosome organization may be indicated. In the first place outstanding chromomeres are much more abundant than in Chorthippus, and this is particularly true of those occurring at the ends of the chromosomes. These polar granules (chromomeres) have already been noted in chromosomes 5, 6, and 7, but they are also prominent in the earlier stages, at least, in numbers 2, 4, 8, and 9. The terminal chromomeres are less prominent but easily distinguishable in number 10, still less so in numbers 11 and 12. In most cases, with the gradual condensation ol the entire chromosome, these prominent chromomeres become difficult to detect in the later stages (cell D, e. g.). In the case of number 2, however, the polar granules are relatively very large and they continue to be recognizable even in the metaphase (cell E).

Secondly, the presence of the appended bunches of granules (c) on chromosomes 4 and 8, already noted, may be considered further. As characters for identification of these chromosomes they have been mentioned on a previous page. An examination of the pachytene stages shows these appendant structures to be prominent. Figures 22 and 23 (plate 3) show them (c) attached to their respective spireme segments. Chromosome 4 is further identified in these figures by its greater capacity for holding stain. In both of these figures the appendages appear to be attached to only one of the conjugants of each pair of threads.


and furthermore, they appear to be joined to single chromomeres in three of the four illustrations. In both cells, that end of the chromosome, near which the appendages are attached, is united with the end of another chromosome, the opposite ends being free in each case. In figure 23 the burdened ends of these two chromosomes are joined together, the appendages are in contact and, so far as could be determined, may possibly be fused at the point of contact. It will be noted that the chromomere to which the appendages are attached appears to be in about the same position with reference to the end of the segment in both cells.

A study of still earlier stages to determine whether these appendages may be double, that is, one for each conjugant, was not made. In the stages shown on plate 2 the bunches of granules were so disposed as to make it impossible to determine whether or not their attachments were restricted to one pair of chromatids. Structures which were interpreted as representing these appendages were seen in the nucleus from which figure 21 was made but were not drawn because to have done so would have confused the other features which were to be emphasized.

A third point on the subject of organization is the occurrence of the transverse clefts or constrictions at the points of spindlefiber attachment in the atelomitic chromosomes. These clefts have been described in another connection so that it will be unnecessary to repeat' a description of them.

In passing, it may be noted that the distribution of the accessory chromosome with reference to the members of the heteromorphic pairs follows the law of chance, as discovered by Carothers ('17). (See chromosomes 1, 3, and 8, plate 2.)

g. Summary. In this individual of Trimerotropis the following conditions have been found:

1. The reduced number in the first spermatocyte is twelve (plus one supernumerary). The complex consists of five pairs of telomitic chromosomes, four pairs of atelomitic chromosomes, two heteromorphic pairs, and the atelomitic accessory besides the supernumerary. These facts were all previously



determined by Carothers, and the present observations corroborate those made by her.

2. Synapsis consists in a side-by-side union of leptotene threads, which parallel association persists until the breaking up of the spireme into its constituent segments and the production of tetrads by the formation of the secondary longitudinal split.

3. The process of parallel conjugation appears to begin at the proximal ends of the leptotene threads, as was found in Chorthippus.

4. The great variety of metaphase shapes for the same tetrad appears to be correlated with variations in the extent and character of the movements of the chromatids with reference to each other in the earher tetrad stages.

5. At the point of spindle fiber attachment (synaptic point) in the pairs of atelomitic chromosomes there occurs a transverse constriction, or cleft, similar to those found for the atelomitic chromosomes of Chorthippus.

6. Chromosomes 4 and 8 (plate 2) each bears near one end a cluster of granules which is thought to be homologous with the chromomere-vesicles seen in Chorthippus and other grasshoppers.


On the basis of the foregoing observations it seems worth while to discuss briefly the following topics: (A) Synapsis, (B) pre-reduction vs. post-reduction, (C) chromosome organization.

A. Synapsis

In considering the subject of synapsis it may be well first to summarize the evidence gained from the study of Chorthippus and Trimerotropis, together with that previously recorded for Phrynotettix, and then to compare these evidences with the results recorded by others who have studied Chorthippus (Stenobothrus). An extended discussion of the literature on synapsis would be superfluous in view of the consideration given the subject by McClung ('14), Robertson ('16) and the writer ('16).


The evidence for parasynapsis as it has been presented for Chorthippus and Trimerotropis on the previous pages is very much of the same nature as that presented in the paper on Phrynotettix ('16). The chief differences are as follows: In Phrynotettix three particular first spermatocyte chromosomes were selected for intensive study because they possessed characters which made their identification possible at most of the stages concerned. For these three chromosomes stages in the transformation of the double-thread spireme segment to the tetrad in metaphase or anaphase were traced. In the present study similar stages have been traced for all of the chromosomes in both species. In Phrynotettix the leptotene and zygotene stages were carefully studied and one of the selected pairs of chromosomes traced through these stages. In Chorthippus and Trimerotropis, lepto-zygotene stages have been found which demonstrate the phenomenon of parallel conjugation, beginning at the proximal ends of the threads, just as was found for Phrynotettix.

In the latter species all of the chromosomes are of the rodshaped or telomitic type, while in Chorthippus there are both telomitic and atelomitic types and in Trimerotropis there are, in addition to the two types named, heteromorphic pairs with one member telomitic and the other atelomitic. I think the evidence presented shows very conclusively that parasynapsis occurs for all the chromosomes of whatever type in these three species.

When we consider the papers of others who have studied the male germ cells of various species of Chorthippus (Stenobothrus) we find that not all have expressed an opinion on the subject of synapsis, and among those who have considered the subject, there are some differences of opinion.

The work of Carnoy ('85), who studied the male germ cells of Stenobothrus along with those of a large number of other Arthropods, was of a pioneer kind and did not concern itself with the subject of synapsis.

Davis ('08) was the first among those who have studied Stenobothrus to consider seriously the subject of synapsis, and


on account of a misinterpretation of the looped condition of the pachytene threads he decided that synapsis consisted in an end-to-end union. He regarded the longitudinal cleft ot these threads as ha\dng arisen by a sphtting of the individual chromomeres of the single leptotene threads, instead of arising from a pairing of chromomeres, as I have shown. He failed to detect the formation of the secondary spht, confusing it with the primary, and also failed to notice the zygotene stages with their partially conjugated pairs of threads.

Gerard ('09), on the other hand, did notice the zygotene stage with the partially paired threads and correctly interpreted them as evidence of parallel conjugation. He regarded one of the conjugating threads as more prominent "principale" than the other secondaire" but I have failed to find evidence for such a distinction. Gerard also noticed that pairing began at the proximal pole and progressed distally, as I have described, and further that orientation was not complete nor entirely persistent in the pachytene stages.

Meek ('11), in his first paper on the spermatogenesis of Stenobothrus viridulus, does not commit himself on the subject of synapsis further than to point out the reduction in number of chromosomes in the first spermatocyte and then to say (p. 12): "This reduction must be effected before the breaking up of the spireme, for I have found no evidence of lateral association of filaments after this has occurred," and then (p. 15): Probably

. . . . a numerical reduction takes place as a result of lateral association of chromatin granules or masses on the reticulum of tlireads prior to the spermatocyte prophase of mitosis." In a later paper ('12) he concludes that conjugation in pairs actually does take place, though he offers no proof for the statement and it is apparent that he has not observed the critical stages.

McClung ('14) in considering a Stenobothrus-like form along with numerous other species of Acrididae admits the possibility of parasynapsis, but did not make a study of the earlier stages where conclusive evidence is to be found.


Robertson ('16) considers the subject of synapsis in Chorthippus at some length. His figures 163 to 180 represent about the same stages as those shown on plate 1 of the present paper. The evidence in the two sets of drawings is of the same general nature and consists in the facts that each of the early postspireme segments is a slender filament with one longitudinal split and that a second longitudinal split appears in a plane at right angles to that of the first, giving rise to the tetrads. So far as this evidence goes, therefore, my work, though done independently serves only to corroborate his.

In regard to the actual process of conjugation Robertson has made no observations, but assumes, in accordance with the work of the Schreiners ('06) on Tomopteris and Salamandra and the work of Janssens ('05) on Batrachoseps, that, in the V-shaped chromosomes at least, the process of pairing begins at the distal ends and proceeds toward the apices of the Vs. He is further led to this conclusion by finding interlocking pairs, the peculiar relations of which could readily be explained on such an assumption. He then carries the analogy over to the rod-shaped chromosomes and considers that they probably also begin pairing at the distal end.

My observations, as well as those of Gerard, show that, in a general way, the reverse process takes place, namely, that conjugation is initiated at the proximal end and proceeds distally. In the case of the atelomitic chromosomes the evidence is not quite so conclusive, but that shown in figures 2, 3, and 4, plate 3, indicate that here conjugation begins at the apices of the V's, and proceeds distally. However, I am inclined to believe that the progress of pairing may not always be uniform along the whole length of the threads, but that chance association of the distal ends of a pair would result in their union, even if the region between the ends and the apices had not completely united. Such a condition is shown very clearly in the case of chromosome A in Phrynotettix (see '16, figs. 73, plate 6, and figs. 75 to 78, plate 7), and the interlocking of chromosomes such as that shown by Robertson in his figures 163 and 177 could be explained by such behavior.


In conclusion it may be said that parasynapsis has been amply demonstrated for Chorthippus through the work of Gerard, Robertson, and myself. The evidence is just as conclusive for Phrynotettix and Trimerotropis as found by me, while Robertson found clear evidence in Syrbula. Besides these examples from among the Acrididae, parasynapsis was also found among Orthoptera by Otte ('07) for Locusta, by Morse ('09) for blattids, by Vejdovsky ('11-12) for locustids, by Stevens ('12) for Ceuthophilus, and by Robertson ('15) for Tettigidae.

B. Pre-reduction vs. post-reduction

The old controversy over pre-reduction and post-reduction which has agitated the minds of cytologists for so long a time seems still far from a settlement. Even without considering the subject of synapsis there has been little tendency to agreement and with the assurance that parasynapsis is the rule in a large number of animals (Wenrich, '16) the uncertainty becomes more acute. There is one thing, however, about which I believe we can be certain, and that is that there is no general rule which is followed by all chromosomes at all times, and a failure to appreciate this fact has had much to do mth the present unsettled state of this subject. I have shown for Phrynotettix ('16) how chromosome-pair C (type Ci) divided by pre-reduction half the time and by post-reduction half the time. In the case of chromosomes A and B the rule was post-reduction, and I believed that post-reduction was the rule for the remainder of the euchromosomes. In that paper (quoted on page 472) I was also bold enough to declare my belief that pre-reduction was the rule with the atelomitic chromosomes of Chorthippus. My greater familiarity mth the atelomitic chromosomes since my study of Chorthippus and Trimerotropis has somewhat shaken my faith in some of the criteria used to determine this point. I am strongly inclined to believe that the mode of division in the first spermatocyte cannot be determined with absolute certainty in the absence of recognizable differences between the two conjugants of a pair.


In this coimection it may be of interest to review the evidence on this subject as found in the three forms that I have studied (Phrynotettix, Chorthippus, and Trimerotropis) , and then to compare the opinions of some other investigators, especially those who have studied Chorthippus (Stenobothrus). It will be necessary to consider only the chromosomes of the first spermatocyte, for the behavior in this first division will govern the subsequent behavior in the second spermatocytes.

In the case of the telomitic chromosomes of Phrynotettix, unequal chromosome-pair C, type C, divided half the time reductionally and half the time equationally. Chromosome-pair B, unequal type, always divided equationally while the equal type behaved in precisely the same manner and presumably also divided equationally. Chromosome-pair A divided equationally with the possible exception of the ring-shaped forms, which were not traced to the metaphase and their behavior not determined.

In the case of the telomitic chromosomes of Chorthippus no very conclusive evidence is available, but a probable behavior leading to an equational division has been described on a previous page (480). Among the telomitic chromosomes of Trimerotropis, number 2 (plate 2) exhibited convincing evidence of an equational division, and good evidence for a similar type of division was found for number 4.

When we come to consider the pairs of atelomitic chromosomes, a decision as to pre- or post-reduction is not so easily reached. The evidence in relation to the chromomere-vesicles found in Chorthippus may be made to favor pre-reduction. In plate 1, it will be remembered, whenever a vesicle occurred it was always single and attached to but one pair of chromatids. This single one might be accounted for in one of two ways: either the two that are found in the earlier stages (figs. 2, 3, and 4, plate 3) have fused together to form a single body, or there may be a normal difference between the two causing one to disappear before the other. In all cases where two are seen, one is shghtly larger than the other. This is well shown, for example, in figure 9, plate 3, which presents one of the few cases encountered in which both vesicles appeared at so late a stage. Figures 7 and 8


also show this difference. In figures 10 and 11, the vesicle that is present does not appear to be any larger than the largest one in figure 9, and this fact lends support to the idea that the smaller one disappears early instead of fusing with the larger. The cases shown in figures 12, 13, and 14 are from section of another individual and can therefore not be discussed in this connection. If we assume, then, that the pair of chromatids to which vesicles are attached in chromosome 9, cells E and F, and in chromosome 7, cell D (plate 1), for example, represent one of the conjugants, then the evidence seems clear that a reductional division should take place.

On the other hand, if it ber assumed that the two vesicles have fused to form one, then one of them must have broken loose from its former attachments when the pairs of chromatids separated, as they do in the tetrads, whether or not the separation is along the primary or the secondary plane. When it is considered that these structures are eventually destined to disappear in the later tetrad stagles, as shown by the fact that many of the tetrads observed in the stages represented by cells C to E (plate 1) do not show any vesicles, "the assumption made above does not seem unreasonable. But, if these vesicles are to be regarded as shifting in their attachments, then it might be objected that such structures offer uncertain bases for deduction. On this account, deductions made on the basis of the behavior of these vesicles in the later tetrad stages may have to be considered as not entirely satisfactory. Nevertheless, because of the constancy of occurrence of these bodies throughout the pachytene stages, and the constant difference in size between them, as already mentioned, the most reasonable conclusion would appear to be that the one remaining vesicle, as found in plate 1, and figures 10 and 11 on plate 3, is attached to one conjugant and that therefore, a reductional division is foreshadowed.

We may attack this problem of reduction in the atelomitic chromosomes from a different angle. In chromosomes 2, 4, and 6, plate 1, a probable behavior of telomitic chromosomes was described which paralleled that of chromosome A in Phrynotettix,


and consisted in a first separation (though sometimes very shght) along the primary spUt, followed by a progressive separation of a more pronounced nature along the plane of the secondary split. When we attempt to find an analogous behavior in the atelomitic chromosomes we meet with difficulties. If, however, we follow the example of Robertson ('16) and consider each of the atelomitic chromosomes as composed of two telomitic ones joined at their proximal ends and then consider that the behavior of these component rods is analogous to that described above for the telomitic chromosomes, we would have to conclude that division of these pairs of V's was equational.

If we carry this idea of the compound nature of the atelomitic chromosomes over to Trimerotropis, further difficulties await us. Robertson uses the idea of compound chromosomes in Chorthippus to indicate that the real number of chromosomes is 23 (cT) instead of the apparent 17 (cf), and that, consequently, Chorthippus possesses the number typical of the Acrididae. If we assume that the atelomitic chromosomes of Trimerotropis also are compound then we should have 34 as the total number of telomitic chromosomes in this particular individual (see p. 506). However, if we are to consider the number of chromosomes in the Acrididae to be constant or nearly so we must conclude that the atelomitic chromosomes of Trimerotropis are not compound and therefore we cannot carry over the analogy of behavior from the possibly compound chromosomes of Chorthippus.

On the other hand the behavior of the chromatids may be considered as variable, so that at one time separation in any pair of atelomitic chromosomes could be along the primary plane and at another time along the secondary plane. In this connection, consider for a moment figure 11, plate 3. Here, there are three successive 'rings' with a potential fourth at the right — that is, if the two free ends were in contact, as they are at the left. Each succeeding ring is in a plane at right angles to those adjacent. If we may number these rings 1 to 3 beginning at the left and consider the incomplete ring on the right as number 4 it will facilitate discussion. We might consider either plane as


primary or secondary but let us assume that number 3 has the chi'omatids separated along the primary plane, with the vesicle attached to the chromatids of one conjugant. This tetrad is at a stage early enough so that it could be expected to undergo chromatid movements before it became condensed and oriented in the metapase. If ring 3 has arisen from a separation along the primary spht, then the free ends (no. 4) must have separated along the secondary split. So, also, number 2; while number 1 would have separated along the primary split. Now the question may be asked if there is any reason why the further separation of chromatids should prevail along one plane rather than along the other. Does not, for instance, the greater extent of the separation in ring 2 and the smaller size of ring 3 indicate that further movement is likely to be in the plane of separation of rings 2 and 4 which we have assumed to be the secondary plane? Would not such a process give us a condition similar to that in figure 10 with the long free arms at the right having opened out beyond the apical point? Again, could not separation proceed to the enlargement of ring 3 at the expense of 2 and 4 and give rise to such a form as occurs in figure 14?

Similar situations are presented in the series from Trimerotropis in figures 15 to 19. In figure 15, so far as can be observed, there would be just as much chance of separation along one plane as the other, since we cannot recognize the synaptic point in this element. In figure 17, the middle ring might become enlarged so as to produce a condition such as is shown in figure 19, or the ends at the right might be separated and refiexed toward the left as indicated in figure 18, as previously mentioned (p. 488). In any case the primary cannot be differentiated from the secondary plane of separation, and in many ways it seems as reasonable to assume one as the other as likely to prevail in anaphase.

From the heteromorphic pairs another line of argument may be developed. If we assume as established that the heteromorphic pairs in Trimerotropis always divide reductionally as they appear to do, we have a basis upon which to argue that the other atelomitic chromosomes should behave in the same man


ner and thus divide reductionally. As a matter of fact, chromosome 12 in cell D, plate 2, shows a constriction in one pair of chromatids (b) and not at the homologous point in the other pair. This might be considered as evidence that the constricted pair represented one conjugant which showed this pecuUarity without its being shown by its partner, and that therefore a reductional division was foreshadowed. The results of Carothers ('17) also go far to support pre-reduction for the pairs of atelomitic chromosomes. She finds that for each of the chromosomes 1, and 4 to 9 inclusive one of three conditions may exist in 'different individuals: (1) there may be a pair of atelomitic chromosomes, or (2) a heteromorphic pair, or (3) a pair of telomitic chromosomes. Since the members of the heteromorphic pairs separate in the first di\dsion, it is reasonable to suppose that in another individual the two members of a pair of atelomitic chromosomes, both of which were homologous to the one atelomitic meniber of the heteromorphic pair, would also separate in the same way. And further, it might be presumed that even the telomitic homologues, when both are present in the same tetrad, would have a similar behavior. The strongest evidence, then, in both Chorthippus and Trimerotropis seems to support the pre-reduction hypothesis for the atelomitic chromosomes.

I have gone into all these suppositions and analogies for the purpose of showing how difficult it is to arrive at any general conclusion with regard to the behavior of the chromatids and the consequent plane of separation at the first division. It is my belief that only in those cases where the two conjugants of a pair can be differentially recognized can the plane of separation in the first spermatocyte di\dsion be absolutely determined for these atelomitic chromosomes.

The method of analogy is perhaps too generally used in the discussions of the subject of reduction. Gregoii^e ('10) attempted, unsuccessfully, to align most of the results published up to the time of the completion of his paper into an agreement on the side of pre-reduction. Davis ('08), taking the V-shaped chromosomes of Stenobothrus as a type, concluded that pre


reduction occurred in them, and then by analogy, probably occurred in all the others. Gerard ('09), though somewhat uncertain, likewise concludes that the plane of separation of the atelomitic rings is along the primary plane and is therefore reductional and then says (p. 579) : Personne ne contestera cependant que tous les chromosomes doivent se comporter a ce point de vue de la meme fagon et que Ton est autorise a appliquer, a tous, ce que Ton a vu d'une fagon irrefutable chez quelques uns d'entre eux." I think there is danger in drawing such analogies.

Meek ('11) says he is not able to. discover whether reduction occurs at the first maturation division or the next but thinks possibly both divisions are equational and a numerical reduction takes place only as a result of lateral association of chromatin granules or masses on the reticulum of the threads prior to the primary spermatocyte prophase. Meek thus failed to see the parallel conjugation (though he assumes it may occur) and does not recognize the primary longitudinal split (along which one division takes place) as the space between conjugants.

McClung ('14) mentions the difficulty in determining the plane of division in the first spermatocyte and says, p. 665-666: "It is probably true that until the relations of the chromosomes during the synaptic phase are definitely determined it will not be possible to assert unequivocally that the longitudinal axes of the paired chromosomes of the first spermatocyte represents the coincident axes of the spermatogonial chromosomes constituting it, for if there be a parasynapsis at any period there is a possibility that the doubly split thread may open out along the plane of the equational cleavage instead of along the. space, between chromosomes. In either event the form of the resulting chromosome would be the same." However he believes that (p. 667) : "The sister halves of a chromosome remain closely united here as in other generations of cells, and that separation between parts of a tetrad is much more likely to occur along spaces between whole chromosomes." If one were to apply the reasoning in the last sentence quoted to the cases under discussion one would get variable results. Besides, I think I have sufficiently demonstrated parasynapsis and therefore the existence of the basis for the uncertainty mentioned.


Robertson ('16) is persuaded that all the autosome pairs divide reductionally in the first division. He bases this conclusion on the interlocking tetrads which he found in Chorthippus, the beha\dor of the 'bi-tetrads' in Jamaicana (Woolsey, '15), the segregation of unequal rods in the first division described by Hartman ('13), Carothers ('13), and Robertson ('15), and the pre-reductional division of the multiples in Hesperotettix and Mermiria (McClung, '05).

He says (p. 246) : "From the instances which I have here given it seems to me the inference may possibly be drawn that all autosomal tetrads will be found to divide reductionally in the first maturation division."

In Phrynotettix I found one unequal pair which apparently always divided equationally iii the first division and another which did so half the time. I was also able to show that Chromosome A in Phrynotettix, though the conjugants were of equal length, nevertheless divided equationally in the first division, and I believe that a similar behavior is clear for chromosome 2 in Trimerotropis. With these different facts established, the variability of behavior in the telomitic chromosomes \\ regard to reduction must be recognized, and, further, in the case of the atelomitic chromosomes it seems unwise to attempt to estabhsh a general rule.

C. Chromosome organization

a. Atelomitic, or V-shaped chromosojyies. The idea of multiple or compound chromosomes was first presented for Orthopteran material by McClung ('05). In a paper in this journal Dr. McClung considers this subject in a comprehensive manner so that it will be unnecessary to discuss here more than a comparison between the chromosomes of Chorthippus and Trimerotropis.

The occurrence of a constriction or transverse cleft at the apices of the atelomitic chromosomes has been noted several times and I have already mentioned the view that Robertson ('16) has taken in regard to the possibly compound nature of the atelomitic chromosomes of Chorthippus. In drawing his conclusion that these chromosomes are compound, Robertson makes


two assumptions, namely, (1) that the number of chromosomes for the Acrididae is constant (i.e., 23, diploid series) and (2) that the 'achromatic bridge' or constriction at the apex represents the point of union between two telomitic chromosomes joined at their proximal ends. Were we to generalize on these assumptions and apply them to Trimerotropis a difficulty would immediately arise. In the first place, this form possesses the number of diploid chromosomes typical for the Acrididae (23) in spite of the presence among them of V-shaped members. If we should assume that each of these V's represents two formerly separate rods, then the total number of rods in the individual studied would be 34, as follows: for the 4 pairs of atelomitic chromosomes, 16; for the 2 heteromorphic pairs, 6; for the 5 pairs of telomitic chromosomes, 10; and for the atelomitic accessory, 2, This number (34) would be a radical variation from the supposedly constant number of 23, and in other individuals the number might be more (or less). As shown by Carothers ('17), the range in number of rods on this basis could extend from 30 (individual no. 62) to 40 (individual no. 6). From the standpoint of constancy in number, therefore, the V's in Trimerotropis should not be considered compound.

On the other hand, if it .s to be assumed that the 'constriction' is a criterion for the recognition of compound chromosomes, then the majority, at least, of the atelomitic chromosomes-in the individual studied would have to be considered compound (plate 2). It is true that the assumption of a compound nature for the atelomitic chromosomes of Chorthippus enables one to recognize the number typical for the group (23), but a similar assumption for Trimerotropis only causes a wide divergence from the typical number which is already present. Perhaps the way out of the difficulty may lie in the suggestion made by Robertson ('16, p. 221) that there are two types of V-shaped chromosomes, one type representing two rods joined at their proximal ends and the other type a bent rod with spindle-fiber attachment in a non-terminal position.

On this supposition the typical Acrididian number may be conserved by assuming that the atelomitic chromosomes of


Chorthippus are of the compound type and that those of Trimerotropis belong to the other type. Considering the similarity of organization of these chromosomes in the two species, i.e., the presence of the apical constriction in the two types, it will be seen that such a constriction cannot be used as a criterion for the recognition of the compound type.

Robertson seeks to extend this compounding idea to the V-shaped chromosomes of other groups of animals. It seems to be unsafe to generalize so extensively, considering that such a generalization does not seem to hold in all cases, even among the Acrididae. It would appear, rather, that each group would have to be considered by itself and conclusions drawn only from a study of numerous species within a group. Such a study has been made by Metz ('14) in which case the compound nature of the V-shaped chromosomes of Drosophila seems to be established. But in this case, as Metz has pointed out, it is as admissible to consider two rods to have arisen from the transverse cleavage of a primitively V-shaped chromosome as to assume that a V-shaped type arose by an end-to-end union of two of the primitively rodshaped type. As Robertson has noted, Agar ('12) found very decidedly constricted chromosomes in Lepidosiren and that (p. 295): The constancy of the position at which transverse segmentation takes place indicates a constant differentiation of the chromosomes in a lengthwise direction." But he considered that : The presence of transverse constrictions or joints in chi'omosomes is not, without special evidence, to be taken as an indication of bivalency or of a future division plane." This note of caution ought, I think, to be especially emphasized.

b. The chromomere-vesicles. The plasmosome-like chromomere-vesicles were first called 'vesicles' by Carothers ('13). In '16 I called attention to their plasmosome-like appearance, at least in certain stages, and emphasized the point, already noted by Carothers, that they are related to definite regions of chromosomes. Dr. Carothers ('17) now calls them ' chromomere vesicles' (p. 473), a term which emphasizes their relationships to separate chromomeres and distinguishes them from the chromosome vesicles found in the spermatogonia where each vesicle


represents an entire chromosome. These structures will be considered here only as further evidences of the exact organization of each chromosome.

As pointed out in the observations, these chromomere-vesicles are attached to the chromomeres at the apices of the V's in Chorthippus — quite regularly in chromosome number 9, less so in numbers 7 and 8 — and near one end of the telomitic chromosomes 4 and 8 in Trimerotropis. In Chorthippus one is practically always attached to the accessory.

The relationship of the vesicle to a particular part of this accessory is not readily observable in the growth period, as shown in figures 1 to 6, plate 3, but on plate 1, it is readily seen that the vesicle (c) is attached to a point near the middle but always nearer one end than the other, indicating that in this case, also, the relationship is to some definite region or chromomere of the accessory.

Whenever the chromomeres at the apices of the V's in Chorthippus can be distinguished in the same tetrad which shows a vesicle, it is found that the vesicle is always attached to that one of the pair of chromomeres which belongs to the longer arm of the V. This is seen especially on plate 3 in figure 9 from the smear and in figures 12 to 14 from sections. It is also seen in in chromosomes 8 and 9, cell A, plate 1.

Similarly, in Trimerotropis, the relationship of the granular appendage is constant in the two chromosomes to which they are attached. In plate 2 the relationships are not so definitely ascertainable, because of the more condensed condition of the chromosomes and of the granules in the cluster, but it is noticeable that the cluster is closer to the end in chromosome 8 than it is in chromosome 4. This is better seen in figures 22 and 23 in plate 3. In both cells the clusters are attached to chromosome 8 at or near the last chromomere but one, near the end and in chromosome 4 to the fourth chromomere from the end. In figure 23 the ends of chromosomes 4 and 8 are in contact. The two appendages are also in contact, but it could not be definitely determined whether or not they had fused together. Taken all in all, this brief study of these slightly-understood,


plasmosome-like chromomere-vesicles adds considerable to the evidence indicating a constant and definite organization of each chromosome.

D. Summary of conclusions

1. Parasynapsis occurs in all the chromosomes of Chorthippus and Trimerotropis, whether they are pairs of rods, pairs of V's, or heteromorphic pairs.

2. The pairing begins at the proximal end of the rods and at the apices of the V's and proceeds distally. It is probable that the distalward progress is not uniform, but that the distal ends may occasionally become conjugated before intermediate parts have united.

3. In the telomitic chromosomes of Chorthippus and in chromosomes 2 and 4 of Trimerotropis there is evidence that the first matm^ation division may be an equational one. The heteromorphic pairs probably divide reductionally. In the case of the atelomitic chromosomes the bulk of evidence favors prereduction, but no definite conclusions on this point seems to be possible.

4. There are probably two types of atelomitic chromosomes, one, found in Chorthippus, consists of compound chromosomes, representing two rods joined at their proximal ends, the other, in Trimerotropis, is a bent rod. This is in agreement mth Robertson's ('16) suggestion.

5. Chromomere-vesicles are constant in position in the chromosomes to which they are attached, and go to support the idea that the chromosomes are constant in their internal organization.




Agar, W. E. 1912 Transverse segmentation and internal differentiation of chromosomes. Quart. Jour. Mic. Sci., vol. 58.

Carnoy, J. B. 1885 La cytodierese chez les Arthropods. La Cellule, tome 1.

Carothers, E. Eleanor 1913 The Mendelian ratio in relation to certain Orthopteran chromosomes. Jour. Morph., vol. 24. 1917 The segregation and recombination of homologous chromosomes as found in two genera of Acrididae (Orthoptera). Jour. Morph., vol. 28, no. 1.

Davis, H. S. 1908 Spermatogenesis in Acrididae and Locustidae. Bull. Museum Comp. Zool., Harvard College, vol. 43.

Gerard, Pol 1909 Recherches sur la spermatogenese chez Stenobothrus biguttulus (Linn.). Arch, de Biol., tome 24.

Gregoire, Victor 1910 Les cineses de maturation dans les deux regnes. La Cellule, tome 26.

Hartman, Frank A. 1913 Variations in size of chromosomes. Biol. Bull., vol. 24.

Janssens. F. a. 1905 Evolution des auxocytes mdles du Batrachoseps attenuatus. La Cellule, tome 22.

McClung, C. E. 1905 The chromosome complex of Orthopteran spermatocytes. Biol. Bull., vol. 9.

1914 A comparative study of the chromosomes in Orhopteran spermatogenesis. Jour. Morph., vol. 25.

Meek, C. F. U. 1911 The spermatogenesis of Stenobothrus viridulus with special reference to the hfeterotrophic chromosome as a sex determinant in grasshoppers. Linnean Society's Jour., vol. 32. 1912 A metrical analysis of chromosome complexes, showing correlation of evolutionary development and chromatin thread width throughout the animal kingdom. Philos. Trans. Roy. Soc. London, Series B, vol. 203.

Metz, Chas. W. 1914 Chromosome studies in the Diptera. I. A preliminary survey of five different types of chromosome groups in the genus Drosophila. Jour. Exp. Zool., vol. 17.

Morse, Max 1909 The nuclear components of the sex cells of four species of cockroaches. Arch f. Zellforsch., Bd. 3, pp. 482-520.

Otte, H. 1907 Samenreifung und Samenbildung bei Locusta viridissima. Zool. Jahrb. Abt f. Anat. u. Ontog., Bd. 24.

Robertson, W. R. B. 1915 Chromosome studies. III. Inequalities and deficiencies in homologous chromosomes: their bearing upon synapsis and the loss of unit characters. Jour. Morph., vol. 26. 1916 Chromosome studies. I. Taxonomic relationships shown in chromosomes of Tettigidae and other subfamilies of the Acrididae, Locustidae, and Gryllidae: Chromosomes and variation. Jour. Morph., vol. 27.

ScHREiNER, A. UND K. E. 1906 a Neue Studien liber die Chromatinreifung der Geschlechtszellen. I. Die Reifung der mannlichen Geschlechtszellen von Tomopteris onisciformis, Escholtz. Arch. Biol., tome 22.


ScHREiNER, A. UND K. E. 1906 b Neue Studien. II. Die Reifung der mann lichen Geschlechtszellen von Salamandra maculosa (Laur.), Spinax

niger (Bonap.), und Myxine glutinosa (L). Arch. Biol., tome 22. Stevens, N. M. 1912 Supernumerary chromosomes and synapsis in Ceutho philus sp. Biol. Bull., vol. 22. Vejdovsky, F. 1911-12 Zum Problem der Vererbungstrager. Bohm. Gesell.

Wiss., Prag. Wenrich, D. H. 1916 The spermatogenesis of Phrynotettix magnus with

special reference to synapsis and the individuality of the chromosomes.

Bull. Mus. Comp. Zool., vol. 60. WooLSEY, C. E. 1915 Linkage of chromosomes correlated with reduction in

nmnbers among the species of a genus, also within a species of Lo custidae. Biol. Bull., vol. 28.


All figures in the three plates were drawn under a camera lucida at a magnification of 3000 and reduced one-half in reproduction.



All drawings made from a smear from a single male of Chorthippus (Stenobothrus) curtipennis. Vertical columns represent separate cells, each horizontal row represents one chromosome in different stages. Figures 1 to 9, at the left, refer to the separate chromosomes arranged in order of size. The letters A to H at the top refer to individual cells, a, Refers to four strand condition, indicating two longitudinal clefts; b, refers to transverse constrictions in the atelomitic chromosomes; c, refers to the chromomere-vesicles.







^ d

6 «■


^^^s^^ «^S

I 2

^ *\ I


2 45*^^

I ^m





a t




All drawings from a smear of a single male of Trimerotropis suffusa. Figures 1 to 12, at the left, refer to the different chromosomes arranged in order of size. Letters A to H at the top refer to separate cells. The supernumerary chromosome in the row at the top of the plate is unnumbered, h, Refers to the transverse clefts or constrictions in the atelomitic chromosomes; c, refers to appendant clusters of granules (chromomere-vesicles) in chromosomes 4 and 8.




%#' H

9 £:::.*-v'.*, x%M

^ ^<

8 -9^ti^ "%S£;?«^'4t t^^ ^?;fe^^


6 i?^*^ r*^*^

2 *•.» -i?^




Small numerals in figures 1, 4, 5, and 6 correspond to the numbers of the chromosomes in plate 1. The small numerals in figures 22 and 23 correspond to the numbers of the chromosomes in plate 2. c, indicates the chromomere vesicles; X, indicates the accessory chromosome.

1 to 11 From the same smear of Chorthippus that was used for the drawings on plate 1. Figures 12 to 14 are from sections of another individual of the same species.

1 to 4 Zygotene stages, showing partially conjugated threads. In figures 2 to 4 the apices of the atelomitic chromosome no. 9 are marked by chromomerevesicles (c).

5 Pachytene stage. Chromomere-vesicles (c) partially fused.

7 and 8 Telophases of spermatogonia showing chromomere-vesicles (c) attached to chromosomes no. 9. Chromosome vesicle of accessory at A'.

9 to 11 Tetrads of chromosome 9, illustrating variation in chromatid arrangements at about the same stage.

12 to 14 Tetrads of no. 9 from sections showing chromomere-vesicles and variety of chromatid arrangement.

15 to 23 Trimerotropis; from same smear as that used for drawings in plate 2.

15, 17, .and 19 Prophase tetrads which may be considered forerunners of the metaphase figures 16, 18, and 20, respectively.

21. Zygotene stage showing partially conjugated threads.

22 and 23 Late pachytene stages showing chromomere-vesicles on chromosomes no. 4, and no. 8.





12 '^4



10 I


4, 17











Department of Zoology, University of Pennsylvania



I. Introduction 519

II. Chromosome conditions in the genus Hesperotettix 521

1. General observations 521

2. The complex of H. speciosus 522

3. The complex of H. pratensis 524

4. The complex of H. brevipennis 525

5. The complex of H. festivus 526

6. The complex of H. viridis 526

III. Chromosome conditions in the genus Mermiria 530

IV. General considerations 536

1. Chromosome numbers in the Acrididae 538

2. Chromosome numbers in general 545

3. Chromosome sizes 554

4. Chromosome forms 558

5. Chromosome behavior 57 1

6. Chromosome distribution 573

7. Chromosome individuality 577

8. Chromosome specificity 585

V. Summary 587

VI. Bibliography 589


Under the title The chi-omosome complex of Orthopteran spermatocytes" ('05) I described certain unusual conditions in these cells, among them the union between the accessory chromosome and particular euchromosomes to form multiple chromosomes. At the time my material was limited and I was able to present only a partial account. Since then I have been accumulating a large series of specimens which have been studied both



cytologically and taxonomically. In the course of this study of the more extensive series of preparations a number of facts have been determined which were obscure in the early stages of the investigation. Although it is not yet possible to present a complete study of the two genera, owing to failure to secure cytological material from all the species, there are some facts known which should be presented in ordqr to add what has definitely been determined and to correct errors in the earher description.

Since the appearance of my former paper ('05) upon multiple chromosomes, a number of observations upon similar structures have been reported by different investigators. Some of these, such as the ones of Voinov ('14), are not clear in their nature, but recent papers by Robertson ('16) and his student, Woolsey, ('16) have made distinct contributions to our knowledge of chromosome relations which will certainly prove valuable. These are strongly confirmatory of the view expressed in my former papers ('05, '08) regarding the persistent organization of chromosomes even in the face of apparent numerical reductions. Very striking is the discovery of a structure, the octad, which I was not then able to find, but whose existence I anticipated so strongly that I gave it a name in advance of its actual observation. The occurrence of such a multiple chromosome I am now abundantly able to confirm, as well as the steps in its formation reported by Woolsey. The underlying principles of chromosome organization, permitting the anticipation of yet unobserved conditions, stands in as striking contrast to the conception of chance association of undifferentiated masses as does the periodic law of chemists to the vagaries of the alchemists.

In this reexamination, advantage has been taken of the unusual opportunities for a study of the taxonomic characters of the two genera, especially Mermiria, offered by the splendid collections at the Philadelphia Academy of Natural Science under the charge of Messrs. Rehn and Hebard. Much good is sure to come from their active interest in all that concerns the group upon which they are specializing and from their cordial and generous cooperation with other students. By their very careful studies of large numbers of excellently preserved speci


mens they reach conclusions regardmg the relationships of individuals and groups which are soundly based upon well defined external characters and upon personal knowledge of the habits and distribution of the materials. Working quite independently of them I reach conclusions regarding the relationships of individuals and groups from the study of their germ cells and, in most cases, find that there is no difference in our estimate of these relationships. Since the discrimination between nearly related forms, upon cytological characters, sometimes reveals groupings that have been overlooked by earlier taxonomists, the feeling of confidence in the validity of the theories upon which such determinations are made is much strengthened. I feel confident that a full agreement between cytology and taxonomy depends only upon the quality of the criteria of differentiation and upon the accuracy of observation on the part of the followers of these two methods of gaining a knowledge of the organization of biological units of different degrees of complexity and extent.


1 . General observations .

When the observations on Hesperotettix were first announced I had only a few specimens, and, because of unfamiliarity with taxonomic characters, my assistants failed to distinguish the species and confused the sources of the material used for study. It was only when the germ cells were examined that it became apparent that all the specimens regarded as H. viridis did not belong to that species. Later and more careful collections enabled me to determine that H. viridis and H. pratensis were both represented. No difficulty is now experienced m discriminating between these species by germ cell characters. Together with H. speciosus, these represented the full extent of my acquaintance with the genus. Although the chromosome complex was, in its major features, consistently afike in the three species, I pubhshed no further observations, because I hoped to obtain preparations from the other North American forms in order to make the presentation complete. After waiting a number of


years I am now in a position to give a much fuller account of ihe chromosomal characters of the genus, although I have not yet seen all the species. Much to my surprise some of the new material shows marked departures from the uniformity prevalent in all my earlier shdes. As a result, some of my generaUzations are now^ rendered invalid, but it is hoped that fuller knowledge may make possible the formulation of principles having even wider application.

At present I have preparations from five different species, but in this article I shall discuss fully only the three of which earlier mention was made, for it is only in these that multiples have so far appeared. The number of specimens in each case, except brevipennis, is possibly sufficiently large to be representative.

2. The complex of H. speciosus

The observations recorded for H. speciosus are essentially correct and need only amplification here. It will be recalled that the accessory chromosome is united with one of the tetrads in the first spermatocyte to produce a hexad element. In the spermatogonium the union with one half of the tetrad also exists and is carried over into one of the second spermatocytes. This association was invariable in all the specimens studied, affecting always the same elements. So far as could be observed the unusual relations of the accessory chromosome did not modify its peculiar character and behavior in other respects. The striking appearance of a chromosome, one part of which is condensed and safraninophilous while the other portion is granular and tinged with the violet in Flemming's tricolor, is presented to our view in prophase and telophase of the first spermatocyte. As was pointed out at the time, an apparent reduction in chromosome number occurs without there being any real difference in this respect from other Acrididae. The diploid number of twenty-two becomes the normal twenty-three when it is noted that the accessory chromosome, instead of being free, forms one hmb of a V-shaped element, easily distinguished among the


twenty-one ordinary rods. That there is here a true preservation of the physical identity of each member of the complex is evident from the continued characteristic behavior of the members which are apparently fused together. The conception of genetic continuity is therefore not merely formal, but expresses the actual morphological conditions of the species. Aside from the presence of the multiple chromosome the germ cells show no marked difference from moBt other Acrididae.

To complete my former account of the chromosomal characters of this species the following facts may be given : In the spermatogonium appear twenty-one rod-shaped chromosomes which are generally characteristic of the Acrididae. Besides these there is a V-shaped element with arms of unequal length (fig. 9, pi. 3). That this is not a simple euchromosome is indicated by the observation that one member becomes highly vesicular during the prophases. Such a condition is uniformly characteristic of the accessory chromosome under these circumstances and is sufficient to identify this arm of the V as the accessory chromosome. During the synapsis stage there is joined to this heterogeneous pair a third member, the homologue of the euchromosome portion. In the late prophase the accessory chromosome is precocious in condensation and appears as a dense, homogeneous rod, more or less bent, joined to a granular tetrad by endwise union. The contrast is most evident in preparations by the tri-color stain. In the first spermatocyte metaphase this hexad element goes on the spindle ^\ith its long axis parallel to that of the spindle and having fibers attached at the ends of the tetrad. This brings one fiber to the point of union between the accessory chromosome and the tetrad. There are visible, at this time, eleven separate chromosomes, of which ten are quite like the ordinary Acrididaean type while the eleventh is the cane-shaped multiple (complex 6, pi. 1).

The anaphase groups resulting from this division each show eleven chromosomes, but in one there is a V which has no like mate in the sister group. This is constituted of the accessory chromosome and one half of the tetrad. In the second spermatocyte into which the V goes, the nature of this is again Hke that


of the spermatogonium — the member which was added during synapsis having been removed in the first spermatocyte mitosis. The second spermatocytes are accordingly of the usual dimorphic character, one with the accessory chromosome — here permanently joined to a euchromosome — and the other without it. Since the V is made up of a dyad plus the accessory chromosome dyad, it is, in effect, a tetrad, but of unusual type since its parts are not homologous. It divides here just as do tetrads in the first spermatocyte, and the anaphase shows a V going to each pole along with the ten rods. The union of the accessory chromosome with one euchromosome is therefore not lost during all the changes of the maturation period, but persists into the spermatozoon and, by it, may be handed on to the female line of the next generation from which it passes into the male line upon the following fertilization.

3. The complex of H. pratensis

In this species, as in H. speciosus, there is uniformly present a multiple chromosome which differs from the speciosus form only in the proportion of its parts. Without some more definite criterion for the homology of the euchromosomes than we now possess it is not possible to say that it is the same member of the complex with which the accessory chromosome unites in the two species. If size only were used as a basis for judgment, then it would be necessary to say that different tetrads are involved, because the euchromosomes are much larger in speciosus than in pratensis. From the fact that the accessory chromosome itself may vary considerably in different species, it would appear that size is not alone a safe indication of homology. Leaving out of consideration the question of whether it is a particular one of the eleven tetrads with which the accessory is joined, we face the concrete fact that so far as our observations have gone, there is always such a multiple element present in the germ cells of the two species.

AVhen we come to observe the character of the multiple element in the different generations of cells it is found to exhibit


throughout essentially the same features as mark its history in speciosus. The only observable difference is in the size of the parts. One indi\ddual, however, exhibits a variation in appearance of the multiple, which makes the conditions in Anabrus, already described ('05), directly comparable to those of Hesperotettix (fig. 23, pi. 4). Instead of the accessory being attached to the end of a rod-shaped tetrad it is joined to one extension of a ring. This relation appears in the prophase and is carried over directly to the metaphase, so that multiple chromosomes of very different shape from the ones commonly present are produced. Both forms appear, not only in the same animal, but also in a single cyst. It is probable on this account that the variation does not represent any fundamental difference, but only a divergence in some rate of movement of the parts.

The spermatogonial complex shows practically the same conditions which characterize the corresponding stage in H. speciosus. Minor differences of relative size of the elements are the only ones of note.

Figure 23, plate 4 exhibits the complex in the first spermatocyte. It is at once noticeable that the chromosomes are strongly of the Hippiscus type and in polar view appear as rings or V's lying in the equatorial plate. The only unusual feature is the tetrad, elongated in a plane of the spindle axis, with the accessory chromosome joined to one end at a more or less acute angle. The behavior of the chromosomes during division presents no unusual features and my general description of these types of chromosomes, given at length in a former paper ('14) may be he'd to apply here. The second spermatocyte mitosis likewise is typical for the euchromosomes, and the multiple chromosome behaves like the same structure in speciosus.

4- The complex of H. brevipennis

There is nothing in the specimens of this species, so far examined, to indicate the existence of any of the unusual conditions of chromosome association which occur in speciosus, pratensis and viridis. As may be seen by an examination of plate 3, figure



8, the complex is strictly of the Hippiscus type. There are neither multiple nor atelomitic^ 1 chromosomes. It is possible that a more extensive series might show some variation, but this is doubtful. If multiple chromosomes are characteristic of the genus — ^w^hich might fairly be said from their prevalence in the species so far studied — then brevipennis departs here from the type. Since this paper deals only with the question of multiple chromosomes, no further consideration will be given to species in which they are absent, except to note that the seriation in size is uniform.

5. The complex of H. festivus

This striking and characteristic species, of which numerous specimens have been studied, is like brevipennis and shows no trace of multiple chromosomes. The nature of the complex in the first spermatocyte may be observed in plate 1, complex 2. It seems therefore to have no immediate bearing upon the subject of the present discussion.

6. The complex of H. viridis

This species, like speciosus and pratensis, may possess a multiple chromosome, a hexad, constituted of a tetrad joined to the accessory chromosome dyad. The euchromosome portion is larger than those of the other species. This was the condition present in the first specimens which I studied and which served

1 H. viridis. In a former paper I found it convenient to refer to various forms of chromosomes and to note that their forms depend upon the fiber attachment. Following de Sinety, the position of this attachment was designated as terminal, subterminal, or median. In the case of the annular chromosomes terminal attachment apparently results in rings whose plane in metaphase coincides with that of the equatorial plate, while with subterminal, or median, attachment the rings are so placed as to lie parallel with the spindle axis. For convenience these were spoken of as the Hippiscus type and the Stenobothrus type respectivel}^ With the discovery of additional complexity of forms it has been found desirable to have general terms for the two conditions represented by the ring chromosome, and so it has been agreed in this laboratory to speak of chromosomes with terminal fiber attachment as 'telomitic,' while those with non-terminal attachment receive the name of 'atelomitic' These terms may therefore be used to designate chromosome conditions similar to, but more general than, the ones involved in the two types of ring formation.


as the basis for my earlier discussion, although at the time they were not fully identified indi\dduals. Now, with an extensive series secured for me by Miss Carothers, I find that the uniformity of association, so marked a feature of my early slides, has, for certain chromosomes, given way to an order characterized by multiples or non-multiples in equal proportions for the individuals studied. For any individual animal however there is no variation. But this species possesses added interest because of associations between euchromosomes to produce octad multiples in which the accessory chromosome does not occur. This is the first unquestionable instance of any such condition in the Acrididae, but its occurrence in a genus of the Locustidae has recently been reported by Miss Woolsey, working under the direction of Robertson. So clear and unequivocal is this condition in viridis that its interpretation can not be escaped, and that it should have failed to appear in my early material is most surprising. A careful restudy of my preparations makes it evident that no other multiples than those involving the accessor}^ chromosome are to be found there. In some more recently prepared sHdes of earUer lots I was however somewhat chagrined to find individuals with the accessory chromosome free. These various modifications of the complex are most interesting and suggestive and merit extended consideration.

The following conditions of the first .spermatocyte complex have been encountered so far (plates 1 and 2).

Class (1 )-12 separate chromosomes = 11 tetrads + the accessory dyad Class (2 )-ll separate chromosomes = 10 tetrads + one hexad Class (3 )-10 separate chromosomes = 8 tetrads + one octad + one hexad Class (4 )- 9 separate chromosomes = 6 tetrads + two octads + one hexad Class (5 )-10 separate chromosomes = 7 tetrads + two octads + one dyad Class (5a)-ll separate chromosomes = 7 tetrads + two octads + one dyad

and supernumerary Class (6 )-ll separate chromosomes = 9 tetrads + one octad + one dyad

While the number of chromosomes varies from nine to twelve, the number of chromatids, morphologically recognizable, remain in all cases forty-six. The particular form of chromosome which I called an 'octad' in my former study ('05) presents itself very conmionly in the new material.


A discussion of these various classes of complexes will serve to bring out most clearly the conditions prevailing in the species. Of these the first shows no unusual conditions whatever, and had specimens of this nature been the only representatives of the genus examined, they would have been regarded as coming from a group entirely typical of the family. There are the usual twelve chromosomes in the form of rings, V's, crosses and rods as found in Hippiscus, all of the telomitic type. The accessory chromosome is free and early passes undivided to one pole of the spindle. Of thirty-eight individuals studied, five had this chromosome constitution.

Seven of the thirty-eight belong to the second class and were the only kind of which I had preparations when my first study was made. Conditions here are as in class 1, with the striking exception that the accessory chromosome is joined permanently to one of the tetrads, forming a hexad element. The structure and behavior of this complex of eleven chromosomes instead of twelve, has been fully described and will not need further consideration here.

Very different is the aspect of the complex in the five individuals in class 3. Instead of twelve separate chromosomes as in class 1, or eleven like class 2 there are but ten. We recognize at once the hexad multiple of the second class and, in addition, eight telomitic tetrads of characteristic forms. The tenth element is much the largest and is otherwise strikingly different from its mates. In shape it is a much elongated ring of the Stenobothrus type, but, like the rings of Chloealtis, usually lacking any lateral extensions. Not infrequently there will be a separation on one side of the ring producing a C-shaped structure. The character of this element will be considered later in connection with other chromosome forms.

Members of class 4, of which I have seven indiAdduals, show still further modifications. The number of chromosomes is reduced to nine, among which are distinguishable six tetrads of usual types, two large Stenobothrus rings and a hexad multiple. The two differential rings are of unequal size. The larger is directly comparable to the similar element of class 3, but the


smaller is always recognizable by its minor size and also by its variability of form. In this Matter respect it may appear as a fully formed ring, or it may show varying degrees of separation at one of the polar angles, resulting in a large V-shaped structure. Any one individual animal however shows either a ring or a V and this does not suffer variation in any cells.

In class 5, represented by seven individuals, we encounter again the number ten as in class 3, but it is constituted in a different manner. The absence of the hexad multiple is at once apparent, for the accessory chromosome is free and occupies a characteristic polar position. Its sometimes associated tetrad appears among the seven chromosomes of this type and there are in addition, two large rings as in cells of class 4. So far as can be seen, these two elements are as directly comparable with the corresponding two of the preceding class as are similar structures within either of these classes.

To the complexity resulting from combinations of separate elements into multiples, thus reducmg the number of independent structures, there is added the opposite condition of a supernumerary chromosome in the animal found in class 5-a. This has a complex similar in its composition to the ones in class five to which is added a small, extra element with the usual characteristics of the supernumerary chromosomes. For the third time there appears the number eleven, but its make-up is different from the complexes of either class 2 or class 6.

The number eleven reappears in cells of the six individuals in class 6, but once again there is lacking the hexad multiple characteristic of classes 2, 3, and 4. Nine ordinary tetrads, a free accessory chromosome and one large Stenobothrus ring make up the complex in this group. In place of the smaller ring, or V, of classes 4 and 5 there are present two rod-shaped chromosomes of corresponding size. The large ring occurs in five of the six individuals and is directly comparable to the element faUing in the same place in complexes of classes 3, 4 and 5. The sixth specimen of this group has, in place of the ring, a large V of the same construction as the smaller octad in class 4. This is the only individual in which the larger tetrads were not imited at


both extremities and, until it appeared, I was inclined to believe that the unipolar union, rather common in the case of tetrads nine and ten, was lacking and that some modification of the association force existed. It is now evident that the conditions are essentially the same in both cases, except that the more frequent occurrence of the large ring and its complete character would suggest a longer history of multiple constitution.


At the time of my first report ('05) upon Mermiria I had studied aU the available Acrididaean species and, in every individual, found, only one type of chromosome — a telomitic rod in the spermatogonium, which, after synapsis with its mate, appeared in the form of a Hippiscus-type ring or in some modification of the type. Although familiar with the atelomitic form of chromosomes in other material, I had never seen it in the Orthoptera and had no reason to suspect its occasional appearance. Failure to consider the possibility of such chromosome forms was responsible for my misinterpretation of the multiple chromosomes in Mermiria, through which I made the serious error of reporting the segregation of whole tetrads. Although, with increased knowledge of the conditions in the group, it is now possible to determine with certainty the nature of the multiple chromosomes, at the time of my first acquaintance with the phenomenon of chromosome combinations the constancy of chromosome structure and behavior was so marked as to make quite unjustifiable any assumption of variation. Added to this is the fact, which seemed at the time very significant and which even yet is not satisfactorily explainable, that in certain individuals the multiple has a series of definite constrictions corresponding in number and position to what would exist in a decad element. Finally; actual separation in the first spermatocyte metaphase at the level of these constrictions in certain cells persuaded me of their value as an indication of chromosome boundaries. The theoretical difficulties involved in explaining the preservation of the complex on the assumption of tetrad segregation finally convinced me that I was probably mistaken, but


it was not until the discovery of the J-shaped chromosome in Trimerotropis by Miss Carothers that any explanation, consistent with the other known conditions in the group, was made possible. Meanwhile I had accumulated an extensive series of specimens and had studied the complex of the female and needed only the conception of the J-shaped tetrad to bring all the observations — with the exception just mentioned — into conformity.

In Mermiria, as in Hesperotettix, the fixity of combinations which seem.ed to mark the genus has failed to prevail throughout all its species. While such vai'iation as is shown by H. viridis does not obtain in any species of Mermiria studied, multiples are lacking in texana and neomexicana. These results emphasize strongly the necessity of an extensive series of specimens in any investigation — a requisite I have always appreciated and sought to meet. Our recent work in the Orthoptera has shown that numbers of considerable magnitude are required for extensive generalizations.

Mermiria hiviitata

It is not my purpse in the present paper to enter into a comparative history of the chromosomes in the various species of Mermiria. I wish merely to consider the multiple chromosome found in bivittata, with its modifications in certain groups of individuals which seem to have specific value. My former report on this species stated that the multiple chi'omosome consisted of two tetrads joined to the accessory chromosome. This is a mistake the occasion for which is discussed elsewhere. More extended study upon a large series of specimens has shown that in this genus, as in Hesperotettix, there is a hexad multiple instead of a decad. While there is a striking difference in configiu-ation of the element in the two groups, it is due entirely to the form of tetrad involved. In Hesperotettix this is an ordinary extended rod with the accessory chromosome joined at approximately a right angle on one end. Mermiria has, on the contrary, a tetrad alread}'^ of just this shape, to the straight end of which the accessory chromosome unites. Thus, while the composition is the same, the appearance is very unlike. How


this configuration is established, appears very definitely when the chromosome complex of the different cell generations is studied.

Conditions of the spermatogonia! metaphase are presented in figures 44, 46, 47, 49, pi. 6. First it may be noted that the number of chromosomes present is twenty-two instead of the usual twenty-three, and that, of these, two differ from the remainder and from the usual spermatogonia! rods in being atelomitic V's, with more or less unequal arms. Even a casual study of these differential elements reveals their divergence in both size and structure. One is distinctly larger, and, under favorable conditions, it may clearly be seen that its shorter arm is irregular in outline and more lightly staining. In these respects it conforms to the behavior of the accessory chromosome in many Orthoptera. So characteristic are these appearances that there can be no doubt whatever in the identification of this arm of the V chromosome as the accessory. The V chromosome, then, is a multiple, a tetrad, consisting of the two rod-shaped chromatids of the accessory chromosome joined to the two corresponding rods of a euchromosome, just as in Hesperotettix. At the angle of the V, where the fiber is attached and where the chromosomes join, there is, not infrequently, a clear break in outline. The compound nature of this V is clearly evident from these facts.

Inspection of the smaller V reveals no such indications of heterogeneity, and only in shape does it differ from the remaining elements of the complex. Its position in the metaphase plate is coimnonly near the multiple V. For convenience of comparison several of these pairs of V's in different stages are shown in figures 47 and 49. The number of chromosomes pres-, ent in the complex is therefore found to be twenty free telomitic rods of various lengths, two rod chromosomes joined at their inner extremities forming a V and one other V-shaped chromosome — a total of twenty-three, the number characteristic of the family. The internal morphological evidence of the complex in this generation is sufficient proof to establish this conclusion. In addition however we have two confirmatory lines of evidence which are of great value, viz., the subsequent history of these structures in spermatogenesis and the conditions of the female


diploid complex. Before following out the later spermatogenetic history I ^\dll briefly indicate the conditions in the female cell.

Unfortunately it has not been possible to work out the history . of the female germ cells, but the egg follicle cells show beautiful, clear chromosome groups, some of which are represented in figures 43, 45 and 48. It is at once observable that the resemblance to the male diploid group is very marked. Here again there are twenty-two chromosomes, among which are also two Vs. A more careful study however reveals a significant sexual difference in the case of the two Vs. Instead of being unlike in size they are of practically identical proportions. Owing to their somewhat sinuous course through the cell, with consequent foreshortening, it is difficult, if not impossible, to represent them accurately in drawings. Where they lie more nearly in the same plane, as in figure 43, pi. 6, their equivalence is clearer. Under the microscope there is no difficulty in appreciating the close resemblance existing. On comparing these two Vs with the ones of the spermatogonium (figs. 47, 49, pi. 6) it is seen at once that in size and proportion they agi-ee with the larger one of the male cells. In other words these are two multiples, the short arms of which are the accessory chromosomes. The count for the female complex, instead of the apparent twenty-two, is therefore twenty-fouT; corresponding to the conditions in other species. No differential beha\aor of the sex chromosomes was observed in these female somatic cells and but little e\ddence of separation at the point of fusion.

Other species of Mermiria show no multiple chromosomes in the first spermatocyte, and when the spermatogonial complex is observed there are found twenty-three rod shaped chromosomes of the usual type. Absence of a multiple chromosome in the spermatocyte is accompanied by the absence of V-shaped chromosomes in the spermatogonium. Animals of this type are not classified as bivittata, but, quite aside from their exact taxonomic disposition, it is plain that they must be very nearly related to bivittata. In such material, then, the direct relation between V chromosomes of the spermatogonium and multiple chromosomes of hexad nature in the first spermatocyte is


strongly suggested. Because of the constancy of chromosome organization, indicated among other ways by the constancy of fiber attachment, we would be justified in saying that the V's of the spermatogonium unite in synapsis to form the first spermatocyte multiple and that this should appear with two nonterminal fiber attachments. Such a condition is reahzed in the structure of the hexad in the first spermatocyte mataphase.

The conditions of the first spermatocyte complex in metaphase are readily determined, particuiaiiy if sections of sufficient thickness are used, and in smears are almost diagrammatic. Here there are clearly eleven chromosomes present, among which is a very large and distinctly different shaped one (figs. 55, 56, 57, pi. 7), This is characteristically in the form of a rod with the two ends of shghtly different length bent sharply back in the same plane. One or both of the ends may exceed, or fail to reach, the common angle, producing some variety in form. Greater or less extension, preparatory to division, may result in considerable variation in length, but commensurate and opposite changes in diameter show that the volume remains very constant. Indications of internal composition are afforded by the contour of the element, although, as will be shown later, they are not entirely trustworthy. At each of the bends, where the fibers attach, there is a constriction, and nearer the shorter bent end, at about its length down the shaft, there is a pronounced fissure.

At the time of division, separation occurs at this pomt, producing two unequal V's in the anaphase. Although the chromosomes are much shorter and thicker than in the spermatogonium, relative proportions are preserved, and if these two parts of the long chromosome (figs. 47a, 50a, pi. 6) be compared with the two V's of the spermatogonium they will be found to correspond almost exactly. That is to say, two V-shaped chromosomes of certain proportions found in the spermatogonium reappear in the first spermatocyte (united by one limb) and are there separated at this point and segregated into different second spermatocytes. Since one limb of one V is the accessory chromosome, it remains undivided, as usual. Attached to the accessory chromosome is the rod portion of a J-shaped chromosome which


has separated from its V-homologue to which it was joined in synapsis. The only differences between Mermiria and most other Orthoptera are (a) the multiple chromosome V and (b) the euchromosome V. That the union of the accessory to a euchromosome is not a fundamental change of the nuclear state is indicated by the fact that certain species of Mermiria lack the association. No change in the distribution of the accessory results from its union, and, as we have seen in Hesperotettix viridis, the combination may be so weak as not to occur in some individuals.

Evidence that the euchromosome V is of like transitional character is not wanting in certain well marked members of this loosely constituted species. In these (figs. 58, 59, pi. 7) the multiple chromosome of the first spermatocyte metaphase is much like that of Hesperotettix except in proportions. There is a pronounced bend at one end at the point of fiber attachment, but the other extremity is almost straight, only a slight subterminal flexure indicating the place of the other fiber insertion. In some instances this point is almost at the end of the chromosome. Individuals with this peculiarity are clearly distinguishable by somatic characters and, I believe, constitute a distinct species. It is possible that, with fuller representation, forms similar in the constitution of their multiple chromosomes to those of Hesperotettix might appear.

This type of hexad caused me much difficulty and led to the conception of a decad chromosome. As will be noticed in figures 58 and 59, pi. 7, there are a number of constrictions along the length of the chromosome quite constant in position, and so placed as to indicate that there are five divisions or parts. It will also be observed that the chromosome may be divided at more than one of these levels. I can not now find any explanation for these separations at various levels on the chromosome, but that they are evidences of unions between various tetrads is not indicated by the full history of the chromosome complex. Again, in the anaphase and telophase of the first spermatocyte, the composite nature of the larger V becomes marked. As may be seen in figure 13 of a former paper ('05), one member of the


V consists of two granular rods of the same character as the remaining chromatin elements, while the other arm is composed of two dense and homogeneous rods of smooth contour. All these parts tend to diverge widely except at the level where the unlike portions join, which is the point of fiber attachment in the preceding cell generations and its site in the one to follow. This difference in constitution is the reverse of the one in the spermatogonium, but corresponds to the relative degree of concentration of the nuclear elements in the first spermatocyte prophase. The interval betw^een the two spermatocyte mitoses is very brief and the chromosomes may be followed thi'ough their changes without any loss of identity.

The metaphases of all second spermatocytes show eleven separate chromosomes of which one is a V. On observing these V-chromosomes (figs. 63, 64, pi. 7), however, it is found that they are of two sizes which correspond in proportion to the two V's of the spermatogonium and to the two of the first spermatocyte anaphase. Upon division these are distributed to the spermatids which are accordingly of two classes, equal in number. It is clear from this very e\ddent history of the two V-chromosomes that they go into different spermatozoa and so, upon fertilization, are contributed to different individuals. Of some theoretical interest is the fact that the rod -shaped homologue joined to the accessory chromosome has a criss-cross inheritance while its V-mate is confined t