American Journal of Anatomy 22 (1917)

From Embryology
Embryology - 19 Sep 2019    Facebook link Pinterest link Twitter link  Expand to Translate  
Google Translate - select your language from the list shown below (this will open a new external page)

العربية | català | 中文 | 中國傳統的 | français | Deutsche | עִברִית | हिंदी | bahasa Indonesia | italiano | 日本語 | 한국어 | မြန်မာ | Pilipino | Polskie | português | ਪੰਜਾਬੀ ਦੇ | Română | русский | Español | Swahili | Svensk | ไทย | Türkçe | اردو | ייִדיש | Tiếng Việt    These external translations are automated and may not be accurate. (More? About Translations)

Historic Disclaimer - information about historic embryology pages 
Mark Hill.jpg
Pages where the terms "Historic Textbook" and "Historic Embryology" appear on this site, and sections within pages where this disclaimer appears, indicate that the content and scientific understanding are specific to the time of publication. This means that while some scientific descriptions are still accurate, the terminology and interpretation of the developmental mechanisms reflect the understanding at the time of original publication and those of the preceding periods, these terms and interpretations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)
Amer. J Anat.: 1 - 1901-02 | 2 - 1902-03 | 3 - 1904 | 4 - 1905 | 5 - 1906 | 6 - 1906-07 | 7 - 1907-08 | 8 - 1908 | 9 - 1909 | 10 - 1910 | 11 - 1910-11 | 12 - 1911-12 | 13 - 1912 | 14 - 1912-13 | 15 - 1913-14 | 16 - 1914 | 17 - 1914-15 | 18 - 1915 | 19 - 1916 | 20 - 1916 | 21 - 1917 | 22 - 1917 | 23 - 1918 | 25 - 1919 | 26 - 1919-20 | 27 - 1920 | 28 - 1920-21 | 29 - 1921 | 30 - 1922
Other Historic Journals: Amer. J Anat. | Anat. Rec. | J Morphol.
Links: Historic Embryology Papers | Association of American Anatomists | Wistar Institute of Anatomy and Biology | Wliey - Developmental Dynamics | Internet Archive | biodiversity library


Charles R. Bardeen

University of Wisconsin

Henry H. Donaldson

The Wistar Institute

Simon H. Gage

Cornell University

G. Carl Huber

University of Michigan

George S. Huntington

Columbia University

Henry McE. Knower,

Secretary University of Cincinnati

Franklin P. Mall

Johns Hopkins University

J. Playfair McMurrich

University of Toronto

George A. Piersol

University of Pennsylvania

VOLUME 22 1917



No. 1. JULY

George L. Streeter. The factors involved in the excavation of the cavities in the cartilaginous capsule of the ear in the human embryo. Twelve figures 1

L. BoLK. On metopism. Nine figures 27

Franklin P. Mall. On the frequency of localized anomalies in the human embryos and infants at birth. Eighteen figures 49

Rhoda Erdmann. Cytological observations on the behavior of chicken bone marrow in plasma medium. Two text figures and nine plates 73

Ivan E. Wallin. The relationships and histogenesis of thymus-like structures in Ammocoetes. Three text figures and four plates 127


Warren H. Lewis and Margaret R Lewis. Behavior of cross striated muscle in tissue cultures. Fourteen figures 169

J. A, Myers. Studies on the mammary gland. II. The fetal development of the mammary gland in the female albino rat. Twelve figures 195

Charles R. Stockard and George N. Papanicolaou. The existence of a typical oestrous cycle in the guinea-pig— with a study of its histological and physiological changes. One text figure and nine plates 225

H. E. Jordan and J. B. Banks. A study of the intercalated discs of the heart of the beef. Fifty-one figures (four plates) 285


Aimee S. Vanneman. The early history of the germ cells in the armadillo, Tatusia "novemcincta. Three plates and two text figures 341

E. A. Baumgartner. The development of the serous glands (von Ebner's) of the vallate papillae in man. Ten figures 365

James Crawford Watt. Anatomy of a seven months' foetus exhibiting bilateral absence of the ulna accompanied by monodactyly (and also Diaphragmatic hernia) Four text figures and four plates 385

Leslie B. Arey. The normal shape of the mammahan red blood corpuscle. One figure 439

Andrew T. Rasmussen. Seasonal changes in the interstitial cells of the testis in the woodchuck (Marmota monax). Twenty-six figures (three plates) 475



Department of Embryology, Carnegie Institution of Washington, Baltimore,



The main mass of the cartilaginous capsule of the ear matures into true cartilage when the human embryo reaches a length of 20 to 30 mm., at which time it has acquired what may be considered its adult form with characteristic chambers and openings. From this time on, throughout its whole cartilaginous period, and even after ossification has begun, it undergoes continuous growth, maintaining at the same time, however, its general form and proportions. Such a growth involves both an increase in the surface dimensions of the capsule and a gradual enlargement or excavation of its contained cavities. It is to the manner in which this excavation is accomplished that the T\Titer wishes to call attention and particularly to the factors concerned in its progress whereby a suitable space is always pro\'ided for the enlarging membranous labyrinth. The actual amount of increase in size of the labyrinth is graphically pictured in figure 1. The outlines are made so that they show on the same scale of enlargement a series of wax-plate models of the left membranous labyrinth of human embryos having a crown-rump length of 20, 30, 50, 85 and 130 mm., as indicated in the figure. This covers the period during which the otic capsule is in a cartilaginous state. Ossification begins when the fetus has attained a crown-rump length of about 130 mm. The growth from then until the adult condition is reached may be judged by comparing the above with the final stage, labelled




adult, which is taken from Schonemann's reconstruction^ and reprotlucecl here so as to be on the same scale of enlargement as the younger stages. Since the cartilaginous labyrinth corresponds closely in form to the membranous labyrinth, particu

20 mm

30 mm.


Fig. 1 Median views of wax-plate models of the left membranous labyrinth in human embryos having crown-rump lengths as indicated in the figure. The largest one is taken from Schonemann ('04) and represents the adult condition. They are all on the same scale of enlargement (4.4 diameters) and thus comparison of them shows graphically the amount of growth the labyrinth experiences during this period.

larly as regards the canals, one can see from figure 1 that there is a progressive increase in the size of the cartilaginous chambers throughout the whole embryonic period.

In addition to this increase in size, there is a change in the form of the cartilaginous labyrinth. The general proportions

1 Schoenemann, A. Die Topographic des menschlichen Gehororganes. Verlag von Bergmann, Wiesbaden, 1904. Plate 2, figure 20.


are maintained but there are alterations in the detailed form. As the canals become larger and longer they describe arcs of lesser cm'vature. If one compares the superior canal of an 80 mm. fetus with that of a 30 mm. fetus it will be found that in the former it has doubled its diameter and trebled its length. There is, moreover, a constant change in the relative position of the cartilaginous canals. The lateral canal, for instance, progressively recedes from the lateral wall of the vestibule. In studying this canal, therefore, one may know that it is steadily becoming larger by means of a process of excavation, but this is so managed that the canal as a whole moves in a lateral direction through the substance of the cartilaginous capsule. The topography of the cartilaginous labyrinth is so well provided with known landmarks that these changes in its size and form can be accurately followed. It is possible to determine deductively at what points new cartilage is being laid down and at what points it is being removed. On this account the cartilaginous capsule of the ear is a particularly favorable place for determining the histological features of the growth of cartilage.

As has been noted above the growth of the cartilaginous otic capsule resolves itself into an increase in its external dimensions with a simultaneous hollowing out and reshaping of its contained chambers. It at once becomes evident that this cannot be accounted for on the basis of a simple interstitial increase in the mass of cartilage together with its passive rearrangement to allow for the enlarging cavities, due for instance to a mechanical expansive pressure from the growing membranous labyrinth with its surrounding tissue and fluid. Such a passive rearrangement could only occur in a tissue that is very plastic, whereas cartilage is one of the least plastic of the embryonic tissues. Moreover the histological picture is not that of mechanical pressure. The cartilaginous chambers are always excavated slightly in advance of the space actually required by the membranous labyrinth, and there is no evidence of the labyrinth being cramped or of the creation of pressure grooves in the margins of the cartilage. Nor is the situation improved by the introduction of the conjectured activity of the perichondrium, either in


explanation of the deposit of new cartilage or of the excavation of the old, since the perichondrium, as will be shown, does not make its appearance until after a considerable amount of the growth and hollowing-out of the labyrinth had been already completed. Therefore there is involved in the development of the cartilaginous capsule something more than interstitial and perichondrial growth, in the ordinary sense of the terms. On account of its bearing upon this problem, it is the purpose of the present paper to call attention to the occurrence of dedifferentiation of cartilage in the human embryo, and to point out the important part which this process normally plays in the hollowing out and reshaping of the otic capsule during its development.

The term dedifTerentiation is applied here in the sense of a regression of certain areas of cartilaginous tissue to a more embryonic form, the same areas being subsequently rebuilt or redifferentiated into quite a different type of tissue. Dedifferentiation is defined by Child as a process of loss of differentiation, of apparent simplification, of return or approach to the embryonic or undifferentiated condition." In his noteworthy review of this subject he makes the assertion that the wide occurrence and significance of dedifTerentiation in the lower animals and plants "must at least raise the question whether similar processes do not occur to some extent in higher forms. "2 From the context it is evident that he refers to man as well as other mammals. The materialization of his prediction is here at hand in the development of the cartilaginous capsule of the ear. Before entering into this further it will be necessary to outline the earlier steps in the histogenesis of this particular tissue.


The cartilage of the otic capsule in its transition from embryonic mesenchyme to true cartilage passes through three fairly definite phases: firstly, the condensation of mesenchyme

^^hild, C. M. Senescence and rejuvenescence. University of Chicago Press, 1915. Page 293.


around the otic vesicle; secondly, the differentiation of the condensed mesenchyme into precartilage; and thirdly, the conversion of precartilage into true cartilage. These three histogenetic stages merge more or less diffusely into one another and one must bear in mind that such a subdivision is necessarily, arbitrary and tends to result in an exaggeration of the distinctness of the lines of their demarcation. Their points of difference, however, are here emphasized because the reversal of one state of development into a previous state is the feature to which it is desired to call especial attention.


When a human embryo is 4 to 5 mm. long the mesenchymal tissue surrounding the otic vesicle differs very little from that in other regions. The nuclei, however, are quite sparse in the regions ventral to the neural tube in the median line, and they become perceptibly more numerous as one explores laterally into the neighborhood of the otic vesicle. This sHght increase in the number of nuclei around the vesicle marks the beginning of the mesenchymal condensation that is to form the otic vesicle. A definite layer of such nuclei is not found until the embryo reaches a length of about 9 mm. ; it is then possible to recognize a fairly well outlined zone of mesenchyme which represents the otic capsule in its first stage of development. In figure 2 is shown a sketch indicating the relations which exist at that time. It represents a transverse section through the otic vesicle at the level of the attachment of the endolymphatic appendage. The zone of condensed mesenchyme forming the primordium of the otic capsule abuts directly against the lateral wall of the vesicle and extends from there to a point about one-half the distance between the vesicle and the ectoderm. On the median side of the vesicle this zone is lacking, although there is a considerable number of mesenchyme cells clustered around the vascular plexus ensheathing the central nervous system, and among the nerve rootlets of the acoustic complex. When this zone is analyzed under higher magnification it is found that it still consists essentially of a mesenchymal syncytium. It differs



morphologically from the adjacent mesenchyme, with which it is directly continuous, only in its more numerous and more C()nii)a('tly ari'anged nuclei and its somewhat richer network of internuclear processes. This is shown in figure 3 which is taken from an embryo a little larger than that in figure 2, but which in its general form is apparently in about the same stage of development.

Otic copsule

Ect o d e rm

G.petros f/-^^,


Med. oblo na.

Fig. 2 Section through the region of the otic vesicle in a human embryo 9 mm. long (Carnegie Collection, No. 721) enlarged 66.6 diameters. The primordium of the otic capsule, consisting of condensed mesenchyme, can be seen enclosing the vesicle on its lateral surface.

During the period of growth represented by embryos between 9 mm. and 13 mm. long, that is, up to the time when the semicircular ducts begin to separate from the main labyrinth through the apposition and absorption of the intervening membranous wall, the zone of condensed mesenchyme around the otic vesicle increases in extent and compactness, thereby forming a sharply defined capsule which completely encases the labyrinth. This capsule of condensed mesenchyme has the same openings and corresponds closely in form to the cartilaginous capsule into which it is destined soon to be converted.


' • m4S^ ^vM^^jf otic vesicle


► Capsule


Ad) ac e nt

m es en c hyme


Fig. 3 Camera lucida drawing of a portion of the otic capsule while it is the state of condensed mesenchyme. It is taken from a human embryo 13.5 mm. long (Carnegie Collection, No. 695). The section is 10 microns thick and is enlarged 950 diameters. The syncytial character of the capsule can be seen and also its relation to the epithelial wall of the otic vesicle and to the surrounding mesenchyme.


The histogenetic changes which initiate the conversion of the capsule of condensed mesenchyme into a cartilage-hke tissue make their first appearance just after the separation of the semicircular ducts from the main vestibular pouch. This occurs when the embryo is about 14 mm. long. The conversion of the



capsule into a true cartilage with a characteristic tinctorial reaction of its matrix is not completed until the embryo attains a length of 30 mm. Thus in embryos between 14 and 30 mm. long the otic capsule consists of a tissue in an intermediate condition between condensed mesenchyme and cartilage. This inter

Otic cap s u le

Ec t o de rm


—D. sc.p&st.

Sinus t r.

Med. ^^


... _

o b 1 o n Q . i-A^/Ss

^ o


-Cl\ sJV


Gana. nodos



Fig. 4 Section through the region of the otic capsule in a human embryo 15 mm. long, (Carnegie Collection, No. 719). Enlarged 66.6 diameters. The epithelial portions of the labyrinth are shown in solid black and it will be noted that they are in direct contact with the substance of the capsule; there is as yet no periotic reticular tissue. The section passes through the superior and posterior semicircular ducts and through the utricle near its junction with the crus commune.

mediate form is known as precartilage. It constitutes the second of our three stages of cartilaginous growth.

The general form and relations of the otic capsule at the beginning of its conversion from condensed mesenchyme into precartilage is shown in figure 4, which represents a horizontal section through this region in a human embryo 15 mm. long (Carnegie



Collection, No. 719). It will be noted that the capsule abuts directly against the epithelial wall on the labyrinth. Around the margins of the capsule there is a vascular network the branches of which, however, do not penetrate into its substance. In its form it is essentially the same as its antecedent capsule of condensed mesenchyme, but in structure it can be seen to be undergoing certain characteristic alterations. These do not occur uniformly throughout its substance but appear

Fig. 5 Camera lucida sketches showing characteristic fields in sections of the otic capsule while it is in the precartilage state. Enlarged 950 diameters. The groups labelled A are taken from an embryo 17 mm. long (Carnegie Collection, No. 576). Group B is taken from an embryo 18 mm. long (Carnegie Collection, No. 409).

earlier in some areas than in others. They consist of an increase in distance between the nuclei, together with an alteration in the internuclear protoplasmic network and its spaces. Whereas the capsule, as seen in prepared sections, has previously consisted of a mesenchymal syncytium, it now gradually loses its syncytial appearance. Most of the branching processes disappear and are replaced by a homogenous mass. Some of the processes, on the other hand, persist, and become thicker and more sharply outlined. These persisting larger processes usually exhibit a characteristic relation to the nuclei. Two or more of


them unite in the formation of a loop at one side or at one or both ends of a nucleus, thereby creating a perinuclear space which soon takes on a more transparent appearance than the surrounding homogeneous material that accumulates in the place of the disappearing processes. These changes can be seen in the sketches shown in figure 5, which represent characteristic areas in the otic capsule while in the precartilage stage in human embryos 17 and 18 mm. long. In the two sketches marked A the contrast beween the permanent and disappearing protoplasmic processes is already noticeable. In the sketch marked B the transition is more advanced although one can still recognize in the homogeneous matrix remnants of branching processes which have not yet disappeared. The persisting processes enclose characteristic capsular or perinuclear spaces. Similar spaces are shown in figure 6 which presents a series of isolated nuclei with their associated permanent processes such as are found in sections of maturing precartilage. In some of these (figure 6, C and figure 5, B,) there is a beginning accumulation of granular protoplasm at the margin of the nucleus which constitutes the so-called endoplasm and becomes enclosed with the nucleus in the capsule. After the formation of the spaces the endoplasm gradually accumulates and forms the cell body of the encapsulated nucleus. Thus in precartilage we find all stages in the transition, from a mesenchymal syncytium to a tissue consisting of partially encapsulated cell-islands separated from each other by a homogenous matrix.


The transition from precartilage into cartilage gradually takes place in the otic capsule when the embryo is between 25 and 30 mm. long. This maturation is characterized by an increase in the amount of matrix combined mth a more complete encapsulation of the nuclei, or cartilage-cells, as they may now be designated. With the increase in the amount of the matrix there is also a change in its chemical composition, so that it becomes possible to stain it differentially. This tinctorial reaction constitutes an arbitrary point at which it may be said


that the precartilage becomes cartilage. In embryos 30 mm. long the greater portion of the otic capsule reacts tinctorially and has the histological character of young cartilage. With this stage we reach the third and final phase of the process with which we are dealing. The further changes from younger cartilage to

K:m) \^a

Fig. 6 Characteristic precartilage cells showing the manner in which spaces become enclosed around them, eventually becoming encapsulated cells of true cartilage. Enlarged 950 diameters. Group A is from the otic capsule of an embryo 17 mm. long (Carnegie Collection, No. 296) ; Group B is from an embryo 24 mm. long (Carnegie Collection, No. 455) ; and Group C is from an embryo 23 mm. long (Carnegie Collection, No. 453).

older cartilage, and the conversion of cartilage into bone, are doubtless a continuation of the same general process but in the present paper they will not be taken into consideration.


It has been pointed out elsewhere by the writer^ that there is derived from the condensed mesenchyme surrounding the otic capsule not only the cartilaginous capsule but also the periotic

' Streeter, G. L. The development of the scala tympani, scala vestibuli and perioticular cistern in the human embryo. Am. Jour. Anat., vol. 21, 1917.


reticulum which eventually intervenes between the capsule and the epithelial labyrinth. The relation existing between this reticulum and the three stages of cartilage that have just been defined must therefore now be referred to. The formation of the periotic reticulum is first indicated by a cluster of deeply stained nuclei that can be seen along the central edge of the semicircular ducts in embryos soon after the ducts are formed, and at about the time the otic capsule begins to change from condensed mesenchyme into precartilage. These nuclei constitute a focus at which the development of the reticulum and its blood vessels takes origin. Here the tissue of the capsule gradually takes on an appearance less like a cartilage-forming tissue and more like embryonic connective tissue. Spreading from this focus a narrow area is established which soon encircles the semicircular ducts and becomes the open-meshed vascular reticulum which in embryos 30 mm. long everywhere bridges the space existing between the epithelial labyrinth and the surrounding cartilage.

While in the stage of condensed mesenchyme and in the earlier part of its precartilage period the tissue of the otic capsule to all appearances abuts directly against the epithelial wall of the labyrinth as shown in figures 2, 3 and 4. It is possible, however, that some of the cells directly adjacent to the epithelium do not properly belong to the tissue of the otic capsule. It is conceivable that such cells may represent indifferent mesenchyme and perhaps angioblasts which were originally enclosed, along with the otic vesicle, by the condensed tissue of the capsule where they remain in contact with the epithelial wall in a resting condition until the embryo attains a length of 20 mm. We might regard as an indication of their resumed activity the formation of the deeply stained foci along the central margins of the canals which have been described above. It might thus be maintained that the periotic reticulum is derived from a few predestined mesenchyme cells which after a latent period undergo proliferation and occupy the space vacated by the receding precartilage. On the other hand one may also maintain that the reticulum is derived from cartilage-forming tissue; that it is not


a predetermined tissue but is simply precartilage that has undergone dedifferentiation. In the early stages when only a few cells are concerned this matter cannot be so well determined, the histological difference between early precartilage and indifferent mesenchyme cells not being sufficiently great for their certain recognition. In the later stages, however, it is quite evident that precartilage tissue is actually converted into a reticulum, and that the replacement of precartilage by a reticular connective tissue is accomplished by a process of dedifferentiation. By identifying a special area through its relation to a particular canal, and comparing this selected area in a series of stages, it is possible to observe the conversion of precartilage into reticulum, and to trace histologically step by step the manner in which a space occupied by precartilage in a younger stage is replaced by a reticulum in an older stage. This is the same procedure which occurs in the conversion of cartilage into precartilage and in the latter case, on account of the more highly specialized structure of the tissues, the picture is even more striking, as will be seen in the following outline in which the main features of the process will be pointed out.


It has been noted that in embryos 30 mm. long the main capsular mass consists of true cartilage possessing encapsulated cartilage cells and an intervening matrix that is differentially stainable. A section passing transversely through the lateral semicircular canal of an otic capsule of this age is shown in figure 7. This, and figures 8 and 9, form a series showing at the same enlargement the same canal, i.e., lateral, cut in the same plane at three successive stages in its development. A direct comparison of these figures can thus be made and there is thereby seen the histological changes that occur with the growth of the canal. The successive figures may be superimposed upon each other and in this way the relative amount and position of the constituent tissues be determined. When this is done it is found that in the process of enlargement the true cartilage around the margin of the canal becomes replaced by precartilage



and the precartilage in its turn becomes converted along its inner margin into the reticular mesenchyme which finally becomes the periotic reticulum. In other words, cartilage of the third stage as above described, reverts or is dedifferentiated into cartilage of the second stage and this in turn is dedifferen

."Ductus semicirc. lai".



'. ' •' ' ■ .• ' ■' -Reticulum

Fig. 7 Section passing transversely through the lateral semicircular canal in a human embryo 30 mm. long (Carnegie Collection, No. 86), enlarged 100 diameters. The canal at this time is only slightly larger than the contained epithelial duct, but the zone of temporary precartilage marks out an area that is soon to be excavated by the process of dedifferentiation through which it becomes converted into a reticular connective tissue.

tiated into a tissue approximating the first stage. It is this retrogressive adaptability of its tissues combined with their progressive development which render possible the enlargement of the otic capsule and the alteration in form and position of its contained cavities.

In the 30 mm. embryo shown in figure 7, the first of these three figures, it will be seen that the epithelial duct is sep



arated from the main cartilaginous mass of the capsule by a surrounding zone of precartilage and intervening between the latter and the duct is a narrow zone of mesenchymal tissue which is somewhat reticular in character. This zone of reticulum has attained its greatest width on the median side of the duct, toward the right, being at this point about twice as wide as the

rDu.ctus semiclrc. lat.


-^r-^. — Reticuki m

' . ■ .■ ■ " A!'. ' ■ . ■ ■

Fig. 8 Section passing transversely through the lateral semicircular canal in a human fetus 43 mm. long (Carnegie Collection, No. 886). Enlarged 100 diameters. The epithelial semicircular duct is larger in diameter than the one in figure 9, but that is the accidental result of its having been fixed while in a distended condition. The size of these ducts cannot be compared without taking account of this variation in their distension.

thickness of the duct wall. It is characterized by its reticular arrangement and by the presence of small blood vessels which are not found in the precartilage, although they lie closely against its inner margin. The area of precartilage stands out conspicuously in material that has been intensely stained in hematoxylin without any counter-stain. A series of this kind is represented by No. 199 in the Carnegie Collection. In that series


the true cartilage is deep blue on account of the avidity with which its matrix takes the stain, whereas the precartilage shows only a nuclear stain and therefore is only faintly colored, as compared with the sharply demarcated and almost opaque cartilage surrounding it. The negative of this picture is presented in material where there has been an intense nuclear stain with subsequent decolorization of the cartilaginous matrix. Such a condition exists in figure 7 but is more marked in specimens where the stain is more intense, such as the series No. 972 of the Carnegie Collection. Under such circumstances the area of precartilage appears as a dark field in the midst of the faintly stained true cartilage. Depending upon the management of the technique it is thus possible in embryos about 30 mm. long to display the future cartilaginous canals; that is, the precartilaginous areas which approximately correspond to them, either as dark fields in a light background or as light fields in a dark background.

In the second figure of the series, figure 8, the area representing the future cartilaginous canal, is appreciably larger. Its perimeter, compared with that of the canal in figure 7, is in the proportion of 152 to 115, which are measurements in millimeters made on photographs taken at 100 diameters. By comparing the two figures it will be seen that the increase in size is obtained by the encroachment of the precartilaginous area upon the surrounding cartilage. The amount of this encroachment represents the amount of true cartilage which has reverted or dedifferentiated into precartilage. In a similar manner the reticular zone surrounding the membranous duct has enlarged at the expense of the precartilage. The reticular zone as shown in this figure, taken from a human embryo 43 mm. long, forms a distinct and characteristic eccentric vascular field, but it undergoes its greatest expansion soon after this period.

In the 50 mm. embryo, as can be seen in the third figure of this series, figure 9, the area of the reticular zone is about the same in size as the whole precartilage area in the 30 mm. embryo of figure 7. On comparing these two figures it becomes apparent that there is just as much, and even more, precartilage


in the latter but it has moved outward into the area that was previously true cartilage. At this period the outer perimeter of the precartilage is 192 mm. as compared with 115 mm. in figure 7. As the old area of precartilage disappeared, a new and more peripheral area became established. Thus it may be seen

Ductus semicirc. lat.



,^: - " . ;!vv?.i' ./'/'•'; ' T ?eticulum

Fig. 9 Section through the lateral semicircular canal in human fetus 50 mm. long (Carnegie Collection, No. 95). Enlarged 100 diameters. This section is taken at the same relative position and at the same enlargement as those in figures 7 and 8, so that they may be directly compared. It will be seen that the area of precartilage in figure 7 is now entirely replaced by reticulum, whereas a new and more peripheral area of precartilage has formed at the expense of surrounding cartilage. This more peripheral precartilage likewise in the end becomes reticulum.

that true cartilage has been dedifferentiated into precartilage and this in turn into the periotic reticulum. It is in this way that the enlargement of the canals is accomplished, a process of excavation based on the dedifferentiation of a specialized tissue into a more embryonic type, followed by a readjustment of redifferentiation of this simpler form into a tissue adapted to the new conditions.



In addition to the excavation of cartilage there occurs, in connection with the growth and alteration of form of the otic capsule, the deposit of new cartilage. As the lateral cartilaginous canal, for instance, enlarges it also moves laterally, so that the distance between it and the cartilaginous vestibule increases, thereby producing a lateral migration of the space as a whole. Such a migration must involve an excavation of the established cartilage on its lateral margin and the formation of new cartilage on its median margin. Therefore on the lateral margin we find true cartilage being dedifferentiated into precartilage and on the median margin precartilage being differentiated into true cartilage. The margins of the cartilaginous canals throughout the whole embryonic period are in an unstable condition and are constantly undergoing changes. These are either in the nature of a uniform excavation throughout their whole contour, resulting in a simple enlargement of the canal, or of an excavation in certain parts combined with a deposit of additional cartilage in others resulting in a change of form and position of the canal. On account of the well defined landmarks that characterize the labyrinth, it is possible to orient points at which excavation and new deposit respectively are occurring. Thus one can follow the histological phenomena of these two processes with great accuracy. Where new cartilage is being deposited, the tissue shows all the stages of development from an embryonic connective tissue on its central margin through an aj^ea of precartilage to a true cartilage on its more peripheral margin. These different grades merging into one another repeat stages which characterized the whole capsule in embryos between 14 and 30 mm. Where the cartilage is undergoing excavation the same transitions exist, but the changes are more abrupt and there is a sharper line of transition between the different zones. The \\ddth, however, and the sharpness of the zones vary somewhat, being relatively ^\'ider and less abrupt in youngef stages and becoming narrower and more abrupt in their transition in older fetuses. It is quite possible that these changes occur in waves and when the zones are wider and less abrupt it is due to the greater acti\aty of this process of dedifferentiation and when the zones are nar


rower and more sharply outlined, as is common in older fetuses, the alteration is then proceeding more slowly.

The dedifferentiation of cartilage into precartilage involves first of all changes in its matrix including the loss of its tinctorial reaction, a decrease in its amount and an alteration in its structural appearance, in that it becomes less homogeneous and begins to show the presence of branching processes. As a result of these changes in the matrix, the encapsulated cartilage cells come to lie closer together, pressing to some extent directly against each other. The combined edges of the overlapping margins of flattened capsules give the appearance of wavy refractile lines running through the transition zone parallel to the margin of the canal. With these changes the capsules of the cartilage cells rapidly become incomplete and take on the appearance of branching processes. With the disappearance of the capsules the tissue assumes the appearance of a mesenchymal syncytium which then takes on a reticular character and becomes part of the general periotic reticulum. The question as to whether there is an active proliferation of the nuclei in the tissues subsequent to their alteration from cartilage to precartilage has not been definitely detennined. The material at hand is inadequate for a satisfactory solution of this point, although in some specimens there seems to be an increase in the number of nuclei in the transition zones of precartilage, over and above the apparent increase associated with the absorption of the intervening matrix, which could only be explained in that way. It would seem very probable that with its dedifferentiation there should be associated a renewed proliferative vitality of a given embjTonic tissue, sufficient at least for its reconstruction into the newer form.


In studying the cartilaginous canals one must take into consideration the perichondrium and its relation to the continual transformations occurring along their margins. Reasoning from the prevaihng conceptions, concerning the activity of periosteum in bone growth, one might expect to find in the perichondrium


an important factor in the growth and changes in the cartilage. In later periods its influence on cartilaginous changes cannot be easily determined, but fortunately for the solution of this point it happens that the perichondrium is late in making its appearance and therefore cannot take any part either in the deposit of new cartilage or in the excavation of the old until after a considerable part of this transformation is already completed.

The zone of precartilage surrounding the margins of the canals in embryos about 50 mm. long might be mistaken for perichondrium, such for instance as is shown in figure 9. If this area, however, is followed to a slightly older stage it will be found to be converted almost entirely into reticulum. The section shown in figure 10 is through the posterior semicircular canal of an embryo of the same length, 50 nnn., but a little older in development. It is just at this age that precartilage very rapidly reverts to reticulum, much more rapidly than the surrounding cartilage in reverting to precartilage; and therefore in sections at this period we find only a thin rim of precartilage around the margins of the canals. The real perichondrium makes its first appearance when the fetus has reached a length of about 70 mm. A photograph of a section of the posterior semicircular canal of a fetus 73 mm. long (Carnegie Collection, No. 1373) is shown in figure 11. Examination of this section reveals along the outer margin of the periotic reticulum a condensation of its trabeculae resulting in the formation of a thin fibrous lamina or membrane near the margin of the cartilage. This is the perichondrium in its early form. It does not abut directly against the cartilage but is separated from it by a zone of transition tissue which consists partly of precartilage and partly of reticulum. This transitional precartilage-reticular zone, becomes narrower and more abrupt in later stages. In all of the specimens studied, however, it was found intervening between the perichondrium and the surrounding cartilage. It will thus be seen that the perichondrium is a derivative of the periotic reticulum. It forms an outer limiting membrane along the cartilaginous margin of the latter in a manner somewhat similar to that in which the membrana propria forms an inner one along its epithe



lial margin. The reUition of the perichondriuiu to the reticular tissue surrounding the labyrinth, as seen under higher magnification, is shown in figure 12. The section is a portion of the one shown in figure 11 and includes the successive strata

Ductus semicirc. post.




.• •• •^^^"^"7— R e t i c u 1 u nn


P^ig. 10 Section through the posterior semicircular canals in a human fetus 52 mm. long (Carnegie Collection, No. 96). Enlarged 100 diameters. Here the replacement of precartilage by reticulum has been more active than that of cartilage by precartilage so there remains only a narrow zone of the latter. The reticulum begins to show an alteration in its trabeculae. Due to the retraction and rearrangement of the protoplasm of some of the trabeculae there results a coalescence of adjacent intertrabecular spaces. There are thus formed larger fluid spaces that are devoid of traversing trabeculae. As yet there is no perichondrium.

from the epithelial wall of the labyrinth to the true cartilage. It will be seen that the membrana propria consists of a narrow meshed syncytium, such as is found in embryonic fibrous connective tissue, and constitutes a supporting coat for the epithelial wall of the semicircular duct. The main part of the periotic connective tissue consists of a wide-meshed reticulum and arbor



iziiifi; through it arc Mic loops of small blood vessels. The pcrichoiulriuiu forms in tiu' outer part of this reticuhim as a compact fibrous uieiubrano. l\^rii)luvral to the ])('rich()n(lrium tlio tissue is still of a roticuUir ty]io but passes hi rapid transition hito precartilage and then into a, tiiie cartilaginous tissue.

After making its first a|)i)eai'ance, the ix'richondrium ra])idly becomes more consi)icuous. In fetuses 80 mm. CU length (Car


V\\l,. II l'li(it(ifir;i|)li (if seel ion lliiou^li I lie posi ciior scinicircul;! r cm nal in a liunian fclus 7.'{ imn. loii^ (C'aincfiic ( 'olloction, No. 1873). JMilarficd 100 diameters. It siuiws I lie pcticlioiKlriiiiii in its cai'liest form.

negie Collection, No. 172) it consists of a dense hbrous coat more than twice as thick as that shown in Hgiu'e 12. It is clearly separated trom the cartilage by a narrow zone of transitional precartihige-i"(>ticular tissu(\ In slightly older fetuses, S5 nun. CH length, (CaTn(>gi(> ( \)ll(H'tion, No. 14()()-3()) it has become a (Umisc broad zone sei)arat(Hl from lh(> surroundhig cartilage only by a, narrow cleft of transitional tissue which still, however, can be recogniz(Hl as reticular in character. In fetuses 130 nun. CR long (Carnegie Collection, No. 1018) the perichondrium presents



a relatively inatinx^ a])pearance. As observed under lower magnifications, one is apt to conclude that the perichondriini is in direct contact with the true cartilaj2;e. Under higher powers, however, a narrow zone of transitional precartilage can be seen intervening between them. In this dcHlifferentiating zone the matrix has largely disappeared and the cartilaginous capsules have collapsed and ar(^ flattened out. Thus the elongated endoplasm of adjacent cartilage cells is brought into contact, being separated only by



^,^ Epithelium

® ^' -^•'..:j^'®f^^^ .. r, Membr.prop.

i^.^^ '^" :M




-.fc ■. 7^->— ." \ ^ff'- Reticulum


^ « 


^■'^i^^^f -m^^ Perichondrium

Precart.- retic.


Fig. 12 Detailed drnwinf-- of n port ion of flic sniiic sect ion shown in figure 11. Enlarged 500 diameters. It, can he seen liere that the perichoTulriuni is a condensation of the meshes in the peripheral part of the periotic reticulum and that it separated from the true cartilage by a transitional area of precartilage and reticulum.


the reiuiuints of tlu^ capsiihir margins. The appearance of activity in this zone corresponds to the unstable condition of the margin of the cartilage which is still undergoing gradual excavation.


From a study of the development of the cartilaginous capsule of the ear in human embryos it is found that the changes in size and form which it undergoes during its development are accomplished in part by a progressive and in part by a retrogres.sive differentiation of its constituent tissues. Throughout the entire period of growth, as far as material was available for study, it was found that the margins of the cartilaginous cavities "undergo a process of continual transformation. They exhibit a state of unstable equilibrium, in respect to the opposing tendencies toward a deposit of new cartilage on the one hand and toward the excavation of the old on the other. The margins thereby are always either advancing or receding and in this way are produced the progressive alterations in their size, shape and position. In this manner a suitable suite of chambers is always provided for the enlarging membranous labyrinth.

The general tissue mass of the otic capsule during the period represented by embryos from 4 mm. to 30 mm. long passes through three consecutive histogenetic periods, namely, the stage of mesenchymal syncytium, the stage of precartilage and the stage of true cartilage. In the subsequent growth of the capsule it is found that in areas where new cartilage is being deposited the tissues of the areas concerned follow the same progressive order of development. In areas, however, where excavation occurs, where cartilage previously laid down is being removed, it is found that the process is reversed. The tissue in such areas returns to an earlier embryonic state, that is it undergoes dedifferentiation. Tissue that has acquired all the histological characteristics of true cartilage can thus be traced in its reversion to precartilage and from precartilage in turn to a mesenchymal syncytium. In the latter form it redifferentiates into some


more specialized tissue — in this case for the most part into a \-asciihir reticuhmi.

The perichondrium is a derivative of the periotic reticuhmi and forms an outer Umiting membrane along its cartilaginous margin. During the foetal period the perichondrium does not rest dhectly against the true cartilage but is separated from it by a zone of transitional tissue consisting partly of precartilage and partly of reticulum. This transitional zone intervening between the perichondrium and the surrounding cartilage was observed in all of the specimens that were studied, which includes fetuses up to 130 mm. CR length. Owing to the fact that the perichondrium is late in making its appearance, being first seen in fetuses about 70 nun. long in can take no part in the early changes in the cartilaginous capsule either as regards deposit of new cartilage or the excavation of cartilage that had been previously laid down.



Director of the Anatomical Institute, University of Amsterdam


It is a well-known fact that in man the two frontal bones in a certain number of individuals do not coalesce. In normal circumstances the frontal or metopical suture begins to disappear during the last quarter of the first year, and is completely closed before the end of the second year, the anterior fontanelle disappearing during the third year. The phenomenon of a persisting frontal suture generally is designed as metopism.

Many publications on metopism are contained in the anthropological and anatomical literature. Several reasons have induced me to add the present paper to them. Firstly, I am able to deal with data unknown till now regarding the numerical occurrence of the phenomenon in Dutch skulls. Such a communication is not wholly superfluous because the frequency of metopism varies not inconsiderably among different peoples or races. The second reason for the publication of this paper is given by the fact that in many points the results of my investigations contradict those of other investigators, and, as to the etiology of the phenomenon, I differ from the current opinion. Commonly an increased intracranial pressure, caused by the somewhat more strongly developing frontal brain, is regarded as the mechanical factor preventing the fusion of the two frontal bones. So Martin in his Manual of Anthropology says:

All this shows that a more considerable growth of the frontal cerebrum, as occurring in some brachy cephalic groups, is to be considered the cause of metopism. By the internal pressure the normal concrescence of the frontal bones is prevented, likewise in hydrocephalic skulls, in which regularly the metopical suture persists.


28 L. BOLK

After having- cominuuicated the results of my own investigation I will enter into some critical remarks ujion this opinion.

The above mentioned explanation of metopism gives rise to a more extended point of view. Some authors believe that a large brain indicates intellectual superiority. And it is easy to understand that to such a metopical suture too, should be a symptom of such a superiority, being a suture caused by a strongly developed brain. This opinion has in fact the approval of Schwalbe. In an investigation into the occurrence of a frontal sutiu-e in apes and monkeys this author, after having mentioned the current opinion with regard to the etiology of metopism, says: "This hypothesis agrees with the idea that persons with metopical crania are to be considered as being intellectually on a higher level."

The partisans of this hypothesis surely may advance the following anthropological fact, in favor of their view. It is incontestable that metopism occurs more frequently in culture races than in those possessing a lower degree of civilization. The differences are sufficiently pointed out in the following table, most of whose data are taken from Martin's Manual of Anthropology.

Frequency of metopism

per cent

Australian 1.0

Negroes 1.2

Malayan 2.8

Papuan 4.3

Slaves 6.4

Alsatians 6.5

Bavarians 6.4

Swiss 7.1

Hamburgher 9.5

Scotchman 9.5

Parisian 9.7

The difference between the civilized and uncivilized people is a very obvious one. And even when rejecting the hypothesis of any relation between metopism and intellectual development, this difference still retains its anthropological significance to the full.


Furthermore it is clear that even among the Europeans the ratio is not at all constant in crania of the inhabitants of the MidEuropean region (Bavarians, Alsatians and Swiss), and in the Slavs the frontal bone seems to oe divided less frequently than n crania of the inhabitants of the North-European regions (Hamburgher, Scotch, and, as will be demonstrated further on, also Amsterdamian). I draw special attention to this fact, which does not agree with the not seldom expressed contention, that metopism occurs more frequently in brachycephalic than in dolichocephalic skulls. As far as I am aware, it was Welcker who first pointed out this idea. And it is found in most treatises on metopism. But I think in most of these it is a mere statement of a current opinion, and not a result of definite investigation. The results of research do not confirm this hypothesis. This will be demonstrated by my own research in the course of this paper, and the investigations of Bryce on Scottish crania give similar results. As is well known these are very dolichocephalic, and yet the author found 9.5 per cent metopical skulls among them. Therefore among the dolichocephalic Scotchmen the metopical skulls are more numerous than is the case among the more broad-headed inhabitants of the Mid-European region. This contradicts the assumed prevalence of metopism in brachycephalic skulls.

Before finishing these introductory remarks it is necessary to give a brief account of some of the principal points in the comparative anatomy of the frontal suture. A knowledge of these points is necessary for the thorough understanding of my explanation of metopism, which, as already mentioned, differs from the current one. That the frontal bone in the human embryo arises by two points of ossification situated symmetrically is due to the fact that originally this bone was a paired one. As a rule this condition persists not only in the lower vertebrates, but even among mammals there are many groups in which the metopical suture does not disappear. In Prosimiae as a rule the frontal suture persists as long as the other sutures of the skull. In case of an early closure of the system the frontal suture also disappears earl}^, in case of a persistence of the system till an advanced age, the frontal suture also persists. There is considerable va

30 L. BOLK

riability as to the age at which the skull bones unite in Prosimiae. In monkeys the ossa frontalia unite and a persisting metopical suture is an individual and rare exception. Finally, in Anthropoids a metopical suture in an adult skull has never been seen.

The history of the metopical suture therefore is a somewhat complicated one. Originally the suture was always present, later it disappears, and finally in man it reappears as a not infrequent variation.

I wish to emphasize, that in consequence of this behavior of the frontal suture in the course of evolution, two possibilities must be taken into consideration when trying to account for its reappearance in man. Firstly this reappearance can be explained as due to a quite new influence acting only in man, namely the increased development of the brain which prevents the two frontal bones from uniting. But there is another point of view of a more physiological nature, claiming our full attention in no lesser degree. In primitive Primates the metopical suture persisted. In the further course of evolution certain causes, to which I intend to return, exerted their influence in such a way that both frontal bones were compelled to unite and the metopical suture disappeared. Now, I believe, the possibility presents itself that the metopical sutm-e in man reappears, just because the factor, which once caused its disappearance in monkeys, no longer exerts its influence in the human skull. From this point of view the problem has not yet been examined.

In the foregoing it is made clear that the metopism of the human skull is the starting point of some very interesting problems, to which I will shortly refer in the order in which they are treated on the next pages. Firstly the question about the frequency of the anomaly in Dutch skulls will be discussed, then the question whether the metopical suture occurs more frequently in brachycephalic skulls, and whether it is true that a persisting frontal suture is of some influence upon the shape of the skull. Thereupon we will examine if there exists any relation between metopism and intellectual development, particularly if it is true that the anomaly is more frequent in large skulls, containing a


heavier brain than usual, and finally we will enter into the question of the aetiology of metopism in men.

The material I used for this research consists of 1400 adult skulls of inhabitants of Amsterdam who died during the second half of the last century. It was gathered from one of the cemeteries of this town.

In this collection I found 134 skulls with a persisting metopical suture, that is 9.5 per cent. This relation equals that found by Bryce in Scottish skulls and by Simon in Hamburghian skulls, and agrees nearly with that found by Broca among the old Parisian skulls.

As mentioned in the introductory remarks, it is often claimed in the literature that the metopical suture occurs more frequently in brachycephalic than in dolichocephalic skulls. Now, we will examine in the fii^st place whether this statement agrees with the results of my own research. As a dolichocephalic skull I mean in the following pages all those with an index cephalicus lower than 80, omitting therefore a more detailed classification in mesocephalic, hyperdolichocephalic, etc.

The number of brachycephalic crania present in the whole collection of 1400 skulls, amounted to 420, or just 30 per cent, and among the 134 metopical skulls, there were 55 or 41 per cent brachycephalic. The number of brachycephalic skulls among metopical crania surpasses, therefore, that among the collection as a whole and the difference of 11 per cent really seems to be very considerable. Only the fact merits mention that the absolute number of metopical skulls (134) is a relatively small one, and hence a few skulls more or less exert a perceptible influence upon the percentage. Altogether the above described relation proves that the majority of the metopical skulls is not brachycephalic. And therefore I do not agree with the statement of Anntchin that "metopical dolichocephalic skulls are relatively rare." This conclusion, moreover, does not agree with the results of the investigation of Bryce who, among his material of Scottish skulls, only met with two brachycephalic crania. Yet in another way the eventual influence of a persisting metopical suture upon the shape of the skull may be verified, namely in

32 L. BOLK

comparing the average index cephalicus in normal and metopical skulls. In doing so the following averages were found. That of the total number of 1400 skulls amounted to 78.3 and that of the 134 metopical skulls, 78.9. This difference is such an insignificant one that it does not prove anything as to a supposed more brachy cephalic character of metopical skulls. And the average index cephalicus is such a low one that it by no means justifies the opinion that brachycephaly is a characteristic of metopical skulls, or that metopism in general is favorable to the formation of brachycephalic skulls.

Finally I wish to advance still another proof of the absence of any relation between the shape of the skull and the persistence of a frontal suture. Among the 1400 skulls there were 23 with the very low index cephalicus of 71, an indication of a very narrow skull. And among the 134 metopical skulls, five were found with the mentioned low index. This fact demonstrates clearly that metopism occurs even frequently in skulls which are dolichocephalic in high degree.

It is well known that for the characterization of a skull its index cephalicus is a very insufficient indicator, because for instance the height of two crania with quite the same index can differ considerably, or the curvatures of the calvarium can be very dissimilar. And finally this index furnishes not a single indication as to the absolute dimensions of the skull, a very large and a very small skull may have an equal index cephalicus. Hence a comparison of this index in regard to persisting metopical sutures is a very insufficient means of recognizing the existance of an eventual relation between the shape of the skull and the frequency of metopism. It is necessary to prosecute our investigation in still another direction.

First we will examine whether the three principal dimensions of the skull in average are different in normal and metopical crania. A comparison of the sum of these averages in both groups of skulls will enable us moreover to answer the question whether it is true that metopical skulls commonly are larger, including a heavier brain than nonmetopical crania.



In the next table the averages are dealt with of the three principal dimensions of the 134 metopical skulls and of the total number of 1400 skulls.




134 metopical skulls

1400 skulls.:

182 183.3

144.8 143.8



The height of the skull was measured from the bregmapoint to the casion.

As is clearly shown by the table, the height of metopical skulls does not differ from the usual measure, for a decrease of 0.2 mm. is of no consequence. Regarding this dimension it is certain that there exists no preponderance in metopical skulls. And also the two other dimensions scarcely testify in favor of such a supposition. For though it is true that metopical skulls average 1 mm. broader than normal skulls, their length, on the contrary, is a somewhat smaller one. The metopical skulls seem to be shorter and broader than normal skulls. But the differences are so insignificant that the capacity of metopical skulls equals that of crania with united frontal bones. And an equality of capacity includes an equality of brain weight.

Thus it is obvious that neither in the shape, nor in the absolute dimensions is there a striking difference between the two groups of crania. In this regard the results of my investigation does not agree with that of some other authors. The metopical skulls which I examined were not more brachy-cephalic and were not larger than the normal skulls with which they were compared. And not without reason I consider the results of my own researches to be of a greater value than the contradictory results of some other investigators. For the 134 metopical skulls belonged to the same group as the non-metopical with which they were compared, the whole collection originating from one source. And this was not always the case with the material used hitherto by other investigators.

The result of my research does not harmonize with the alreadjmentioned views upon the cause of metopism. I summarize that


34 L. BOLK

a heightened intracranial pressure during growth due to a greater development of the brain, is considered to be the cause of metopism. Now, I cannot agree with this opinion, for as clearly shown in the foregoing pages, metopism is independent of the shape as well as of the size of the skull. And if there really existed some relation between the degree of development of the brain and the frequency of metopism, one should expect among the largest skulls an increased number of metopical specimens and higher average values of the mean dimensions in metopical skulls. This is not at all the case. The averages of the three dimensions in metopical skulls are nearly the same as in the nonmetopical. Therefore a noticeable difference between the capacity of both groups of skulls cannot be accepted, and consequently the average weight of the brain must be the same.

An objection of more general theoretical nature against the current opinion about the etiology of metopism may be adduced. Is it really true that an increase of the intracranial pressure may prevent the coalescence of two bones of the skulls whose normal fate is to unite together? Martin, the renowned anthropologist, accepts this view, founding his opinion upon hydrocephalic skulls, in which, as he says, metopism is a common phenomenon.

I do not know how far this statement of the painstaking investigator is based upon observations by himself, or is merely the expression of a doctrine propagated in craniological literature. I am inclined to believe the latter. For the experience gained by myself upon this matter is in contradiction with the idea mentioned. There is no concurrence of hydrocephaly and metopism, hydrocephaly being not at all a condition propitious for the persistance of the frontal suture. I have examined carefully the hydrocephalic skulls present in the anatomical Museum of Amsterdam, and the results of this investigation are dealt with in the next table. This table informs us of the state of the frontal suture the horizontal circumference of the skull and the age. With regard to the circumference it may be remarked, that in normal Dutch skulls it amounts to 516 mm.







m m .













Existing. Sut. sagittalis entirely closed













By this table it is clearly shown that the assertion that hydrocephaly regularly is accompanied by metopism, is a false one. Only in the third case the suture was still open. But it is a question whether in this case the presence of the suture was due to supposed mechanical influence of .the hydrocephaly. For as mentioned in the table, in this case the sagittal suture was already entirely closed. And this fact justifies the supposition that in this case the skull was a metopical one by inheritance, in which therefore the suture also should have persisted, if the development of the brain had been quite normal. But, I admit, this to be a mere supposition, although I believe that this case may scarcely be accepted as a proof that metopism is caused by hydrocephaly. It seems better to disregard this case in a discussion of this matter. Furthermore the other data of the table afford a strong proof against the existence of such a casual relation. The first two crania are of an extraordinary size, with a circumference met with rarely, even in hydrocephalic skulls. Surely in both individuals the intracranial pressure must have been an excessive one. And notwithstanding this circumstance the frontal sutures vanished without leaving a single trace. And the same occurred in the other cases mentioned in the table.

I believe the data of this table to be sufficient to justify my statement, that hydrocephaly by no means produces, as a rule, metopism. Hence it seems to me an error to pretend that an increased intracranial pressure — caused by a marked development of the brain — is the cause of metopism. For, if the considerable increase of this pressure, as surely occurred in the skulls

36 L. BOLK

of the first two individuals of the table, was unable to prevent the coalescence of the two frontal bones, it is wholly unthinkable that a somewhat increased development of the brain will suffice to prevent these bones from uniting.

One may advance still another more weighty question with regard to the influence of the growing brain upon the skull. It is assumed that the pressure exercised by the growing brain upon the inner surface of the skull rises, when the brain is developing in a greater degree. Is this assumption true? I do not believe it. It seems to me more probable that with regard to the expansion due to their grow^th, the brain and the skull form one entity, the same hereditary factors determining the growth intensity of the brain as well as of the cranium. I do not believe that the dilatation of the latter is a mere mechanical phenomenon, depending on the pressure exercised by its contents. To some degree this may be the case in pathological circumstances, as in hydrocephaly or in premature closure of some suture or other, but under normal circumstances, I believe the intracranial pressure always to be the same, varying only between its physiological limits.

As a further argument in favor of the assumed influence of the growing brain upon the expansion of the skull, the fact is advanced that the forehead in metopical skulls is broader than in those with normal closure of the frontal suture, this increase of the transverse frontal measure being another result of the more strongly developing frontal lobes of the brain. Without doubt, the observation made f. i. by Welcker and Papillante is right, and I am able to confirm the same, the metopical skulls of my collection having an average breadth of the forehead of 99.7 mm. and the nonmetopical one of 96.5 mm. But I cannot agree with the interpretation of the phenomenon given by the above mentioned authors. I think in this matter they are confusing cause and effect. The difference may be elucidated in the following way. If the frontal suture does not disappear during the second year, the apposition of bony tissue in it is continued during a longer space of time than in case of its disappearance in the normal way, and therefore there is a very favorable


opportunity for the forehead to grow broader than usual. It seems therefore quite reasonable that in metopical skulls the forehead is broader, this being the natural consequence of the fact that the growth-centrum remains longer in an active state.

With regard to the problem of metopism, observations as well as theoretical considerations have convinced me, that the common opinion about the aetiology of this phenomenon is an erroneous one. As to the facts, I have been unable to confirm the existence of any relation between metopism and a particucular shape of the skull, the frontal suture persisting in dolichocephalic crania as frequently as in brachycephalic ones, and the index cephalicus being in average equal in metopical and nonmetopical skulls. Furthermore, the metopical crania of my collection were not larger than the normal specimens, consequently the average of the brain-weight should be equal in both groups. There is but one fact which I was able to confirm, namely the greater breadth of the forehead in metopical skulls, a phenomenon easily understood as a logical consequence of the protracted activity of the frontal suture.

And as to the theoretical side of the problem, I do not agree with the current opinion that metopism is caused by an increased intracranial pressure, the result of a greater development of the brain. First, because the least indication of such an increased development is wanting, and secondly because in pathological cases, as in hydrocephaly, in which undoubtedly the intracranial pressure had considerably risen, the frontal suture disappears as in normal circumstances.

Before entering into an explanation of my views upon the aetiology of metopism, I wish to discuss briefly the argument that metopism is less frequent in the lower races. As mentioned in the introduction to this paper, this fact is utilized as a proof that metopism, caused by a larger expansion of the brain, should be a symptom of higher intelligence. I think this opinion cannot withstand a serious analysis. If one accepts the principle that metopism is a symptom of intellectual superiority as true, because it is more frequent in culture races, than in uncivilized ones, one must accept also the consequence of this principle, that

38 L. BOLK

amongst the culture nations those are psychically the most favored in which metopism is the most frequent. Now in the middle region of the continent and in Russia metopism occurs in about 6.5 per cent, according to Schwalbe, Ranke, Gruber and others. In the northern part of Europe, the phenomexion is more frequent, and attains 9.5 per cent according to Bryce, Simon and myself. In Frisians, occupying the northern region of the Netherlands, metopism amounts even to 11.4 per cent. Though the acceptance of the principle should be very flattering for the Dutch people, I do not accept its exactness, metopism having nothing to do with intelligence. I think the interpretation of the different frequency of metopism in the inhabitants of the central and the northern region of the continent to be this : that it is simply a racial difference, the phenomenon occurring more frequently in the Homo nordicus than in the Homo alpinus.

• The opinion that the difference in frequency of metopism in the human race is a mere physical anthropological character also holds good with regard to a comparison of civilized and uncivilized races. In the former, metopism is commonly very rare. What may be the reason of it? The authors, who hold that the metopism is the result of an increased intracranial pressure, caused by a somewhat hypernormal growth of the brain, adduce this difference as a proof of the exactness of their doctrine, obviously supposing that such a hypernormal growth does not occur in uncivilized races. In this argument there is a very obvious mistake. Surely the average weight of the brain is a lower one in uncivilized races. But the individual weight of the brain differs in uncivilized races as well as in culture races. Not only among white men, but also among Negroes and Papuans there are individuals with sub-normal, normal and hypernormal volume of their brain. And if really a strongly developed brain should cause an increased pressure upon the. inner surface of the skull, this condition is realized as well in a Papuan with, a hypernormal development of his brain, as in an European. Nevertheless in Papuans and Negroes, metopism is rare. I consider this a further proof that the persistence of the frontal suture


has nothing to do either with brain development, or with the higher or lower degree of intellectual evolution.

Now I wish to express my opinion upon the aetiology of metopism. In the introduction to this paper a brief account is given of the phylogenetic history of the frontal suture, principally in Primates. I summarize that among the Prosimiae in some families the frontal suture, as a rule, persists, while in others, on the contrary, it disappears. In monkeys both frontal bones unite together at a very early stage of development, but in some individuals the suture may persist. In Anthropoids till now the suture has never been seen in an adult specimen. This summary shows that in the course of the phylogenetic evolution of man, originally both frontal bones remained separated; thereupon in the higher degree of evolution the bones coalesced, and finally in man the primitive state presents itself again in a number of individuals. These facts form the basis for a conception of the aetiology of metopism differing from those previously advanced. For it seems to me necessary to begin by discovering the cause which caused the suture to disappear in monkeys. Having elucidated this point, we have approached more closely to the solution of the metopical "problem in man. For the possibility must be taken into consideration that the influences which were acting on lower Primates and caused the concrescence of the two frontal bones, have lost their significance and activity in man. If this really happened, it is quite comprehensible that the frontal suture reappears. For in each individual both frontal bones arise separately, the bilateral condition being the rule in the younger stages of development even in such forms in which the individual is born with an already single frontal bone. The metopical suture in an adult individual hence represents no new condition, no alteration of a primitive state, but simply the continuation of an original condition. There must be a special cause for a union of the bones whereas there is no new factor required for the explanation of the fact that they may remain separated. Let us therefore try to find out the primary cause of the concrescence of the frontal bones in monkeys, afterwards we can examine whether this cause became inactive in man or not.

40 L. BOLK

It is a well established fact that the shape of a bone and especially its internal structure, are the results of the mechanical and muscular forces acting upon it. In accordance with the mechanical principle of securing the maximum of strength with the minimum of material, the cancellous tissues of each bone is so arranged as best to withstand the strains and stresses to which the bone is usually subjected. So the internal architecture of each bone is quite in accordance with the fundamental laws of physics; systems of 'pressure lamellae' running in definite direction are crossed by sets of 'tension lamellae.' A great number of investigators have tried with good results to analyze the structure of the different bones of the human skeleton from this point of view. Only in regard to the skull in general, and particularly the cranial vault, are we without definite knowledge as to the structure of the bony framework of the different bones of the skull and the relation between the statical and dynamical external forces to which it is subjected. The whole of our knowledge is confined to the fact that the structure of the plate-like bones of the cranial vault exhibits the following appearance : the outer and inner surfaces are formed by two compact layers, having sandwiched between them a layer of cancellous tissue.

Nevertheless concerning the cranial vault we find ourselves under relatively favorable circumstances, because the general conditions are so very simple here that the problem can be elucidated sufficiently from a mere theoretical standpoint. For the function of the cranial vault being principally a protective one, the number of mechanical stresses to which the frontal half of the skull is subjected is slight. There are but two factors to take in consideration, namely the weight of the facial cranium with the soft parts of the face as a constant working factor, and the pressure effectuated by the temporal muscle during its contraction. The weight of the facial cranium is transferred surely for the greatest part by means of the zygomatic arches to the middle of the base of the cranium, and so there remains as the only important external force acting upon the anterior and lateral part of the skull, the pressure of the temporal muscles, when the jaws are firmly closed. Surely this


stress will determine the arrangement of the cancellous tissue in the frontal bone. And the variations in the arrangement and the course of the pressure and tension lamellae in different animals, without doubt is caused by the variable relation between the frontal bone and the Musculus temporalis. If the muscle arises largely from the frontal bone the internal structure of the anterior region of the cranial vault will be largely influenced by the same. It is obvious that in such a case the frontal and sagittal suture are primarily subject to this influence, as their course is perpendicular to that of the fibers of the muscle.

I think this idea is sufficient to demonstrate why in lower Primates the frontal suture persists, while in the higher Primates it regularly disappears. For the stress of the masticatory muscles tends to compress the skull in a transverse direction and the vault of the skull will withstand this force by a system of trajectories, running on a frontal plane. Now it is not difficult to understand that it is of advantage that the trajectories do not meet with an open suture in their course. And so the fate of the metopical suture in Primates will depend upon the topographical relation between the temporal muscle and the frontal bone. If the muscle arises from the frontal bones a system of pressure and tension lamallae will be developed in it crossing the median line and hence necessitating the union of the tw^o primary frontal bones. If on the contrary, the bone remains free from the dynamical influence of the muscle, there is no reason for the union of the two bones.

In figure 1 an attempt is made to elucidate the above described idea by means of a very simple scheme. It represents a frontal section of the anterior part of the vault of the skull, wdth the temporal muscle on both sides. The direction in which the vault will be narrowed by the stress of the contracting muscle is indicated by two arrows. It is obvious that in order to withstand this stress pressure trajectories will be developed in the vertical parts of the vault, under the direct influence of this force. The compression in the indicated direction will produce a tension in the top of the vault. And while in the vertical parts of it the cancellous tissue will arrange itself in pressure



lamellae, on the top a system of tension lamellae will arise. In the figure both systems are represented by some simple lines. I admit it is a purely theoretical construction, which I have tried, however, to bring in accordance with the principles of mechanics. The point upon which I will lay some stress, is that the tension lamellae necessarily must cross the median plane. And because an interruption in their course by a suture would be contrary to their mechanical function, the two frontal bones unite together. Now we will examine in how far the anatomical conditions in the different Primates agree with the principles worked out above.

It is needless to give a long description of the anatomical conditions in several specimens, for the inspection of some few

crania suggests the regularity in the special groups of the Primates. I will confine myself therefore to treat each group as a whole.

The examination of the prosimian skull shows that in this lowest group of Primates the frontal suture is a constant element in the system of sutures, disappearing nearly at the same time as the other sutures. I regret to have at my disposal only a small number of skulls of prosimiae. Hence it is impossible for me to give a summary of the age at which the metopical suture disappears in the different genera of this group of Primates. The small number of skulls in my possession indicate that a considerable variability exists as to this point in the different genera of the Prosimiae. So I found among five adult skulls of Lemur only one specimen with the system of sutures still wholly


intact, including the metopical suture. In the others the system had completely disappeared. In three adult crania of Avahis on the contrary, apparently of old individuals, all sutures including the metopical, were still present, and so it was in two old crania of Nycticebus. It thus seems that the sutures in Prosimiae close at a very different stage in the different genera of this family. But for the present it suffices to know that the disappearance of the metopical suture takes place simultaneously with that of the other elements of the system. There is no special factor necessitating the same to close at an earlier period than the other sutures. In this respect the Prosimiae differ from the monkeys and apes in which the closure of the frontal suture always precedes those of the other sutures, and often very considerably. From this we may conclude that the influence compelling the metopical suture in monkeys and apes to disappear, is absent in Prosimiae. Now a comparison of the topographical relations between the temporal muscle and the frontal bone in the lower and higher Primates, reveals that in Prosimiae the muscle does not arise from the frontal bone at all. The reason for this is obvious. In Prosimiae the lateral wall of the orbit is a very incomplete one, and frequently also the floor of this fossa is restricted to a foremost part. As a rule the outer wall only is represented by an arch extending from the facial root of the zygomatic arch to the parietal margin of the frontal bone. This insertion of the orbital arch at the hindermost border of the frontal bone causes the latter to be situated completely in front of the temporal fossa, hence the temporal muscle cannot extend its origin forward upon the frontal bone. In monkeys, as in apes and man, the outer wall of the orbit is a complete one, formed partly by the orbital surfaces of the zygomatic bone and the great wing of the sphenoid bone. By this outer wall the orbit is separated almost completely from the temporal fossa and the plane of entrance of the orbit is considerably turned. In Prosimiae the inclination of the latter is more a lateral than a frontal one, the axis of the orbit making a more open angle with the median plane. But in monkeys the plane of entrance is turned, being directed principally forward and but slightly out

44 L- BOLK

ward. The axis of the orbital fossa is making therefore a more acute angle with the median plane. In consequence of the rotation of the plane of entrance of the orbit, the insertion of the primitive orbital arch at the frontal bone was shifted from the hindermost border of the bone forward, so that a part of the outer surface of the frontal bone is added to the temporal fossa. By this enlargement of the fossa the temporal muscle was enabled to arise to a smaller or greater extent from the frontal bone.

The differences between Prosimiae and the higher Primates are clearly shown by the figures 2 to 9. In these figures the lateral and superior view of some prosimian and simian skulls is drawn. The course of the main sutures and also the extension of the temporal muscle is indicated. Figures 2 and 3 represent lateral views of the cranium of Avahis sinavensis and of Stenops gracilis respectively. In both it is obvious that the frontal bone is completely excluded from the temporal fossa, and that there are no fibers of the temporal muscle arising from this bone. Hence it is easy to understand that in those crania the frontal suture persists, as is shown in figure 4, representing the superior view of the skull of an Avahis niger. The frontal bone remains free from the dynamical influence of the temporal muscle, its anatomical significance is a restricted one. It functions only as roof of the orbits and the foremost narrow part of the cavity of the skull. In consequence of the absence of forces acting upon this bone, its system of trajectories cannot be strongly developed. Hence there is no reason for both frontal bones to unite.

Quite the contrary happens in the skulls of monkeys from the Old and New World, as is illustrated by figures 5, 6, 7, and 8. Figure 5 represents a side view and figure 6 a superior view of the skull of Chrysothrix, a platyrrhinic monkey, figure 7 a side V ew of the skull of Macacus, and figure 8 such a one of a female Gorilla. The extension of the temporal muscle and the course of the sutures in the cranial vault are drawn. These figures require but little comment. In all it is clear that the frontal bone participates in the formation of the temporal fossa, and that no small part of the temporal muscle takes origin from this bone.



46 L. BOLK

In Macacus and Gorilla the origin of the muscle reaches to the median line, so that there is but a small triangular part of the outer surface of the bone uncovered by the muscle, while in Chrysothrix a narrow strip on both sides of the median line remains free from the origin of the muscle. It requires no special argument to show that the forces executed by the contracting muscle upon the frontal bone must give rise to a system of trajectories in it, able to withstand the strains on its outer surface. And it is important to draw attention to the fact that the fibers of the muscle are directed perpendicularly to the median line and consequently also with regard to the frontal suture, the forehead being directed horizontally immediately behind the superciliary arch. This condition surely favors the formation of trajectories crossing the median line and causing the frontal suture to disappear, as really occurs in all monkeys and apes. In man the condition is greatly changed, though a small part of the frontal bone is still participating in the formation of the temporal fossa, as shown in figure 9. There are two circumstances by which the relation between the temporal muscle and the frontal bone became altered from that obtainmg in monkeys. Firstly, the frontal bone in man is much larger, and the surface of it occupied by the origin of the temporal muscle is considerably smaller in man than in apes. The pressure of the muscle upon the outer face of this bone in man cannot be a very strong one, hence its influence upon the inner structure surely is of little importance. In this respect the condition in man is getting closer to that in Prosimiae.

The second circumstance peculiar to man is the well-pronounced curve of his frontal bone. By this curve the greater part of this bone rises vertically above the orbits. In apes, as pointed out, the fibers of the temporal muscle are directed perpendicularly to the whole length of the fronta^ suture. In man th's condition is altered, for in consequence of his strongly curved forehead the greater part of the frontal suture is situated in front of the anterior border of the temporal muscle, and moreover is directed nearly parallel to this border.


By these two circumstances the frontal suture in man becomes independent from the dynamical influence of the temporal muscle. Hence in man there is a return to the conditions as met with in Prosimians, though the anatomy of the skull and the muscles is quite different. The mechanical cause for the disappearance of the suture in monkeys having fallen out, the circumstances become very favorable to the persistence. Now it is obvious that these conditions act most favorably in individuals with a more prominent forehead and a less pronounced development of the masticatory musculature. In the white race therefore, the possibility for the persistance of the suture is far greater than in the races with a more flattened forehead, a higher development of the dentition and of the temporal muscle. And this may be considered the cause, accounting for the fact that commonly metopism is more frequent in Europeans than in Negroes or Australians.




111 a paper published nine years ago on the causes underlying the origin of human monsters, I made the assertion that localized anomalies were more common in embryos obtained from abortions than in the full term fetus, without, however, adducing conclusive evidence in support of this theory. ^

In a footnote on page 27 of that publication I gave a list of embryos with their chief defects, comparing them ^\dth the percentage of frequency of monsters born at full term. An objection to be raised to such a statement is the fact that there is not a complete correspondence between anomalies in the embryo and those found in the fetus at the end of pregnancy. For instance, spina bifida in young embryos is always complete while at full term the open canal is covered over with skin. Cyclopia and exomphaly are the same in the embryo as at birth, but the deformities of the head and neck of the embryo are of such a nature that it cannot live long enough to admit of comparison with like malformations found at term. With these difficulties clearly before me, I have made an effort to define sharply the anomalies in embryos, so that a satisfactory comparison might be made with those found in monsters at the end of pregnancy, as described in the literature.

I shall mention first cyclopia, for it seems to me that it is the type of monster which is now best understood. This clearer conception is due largely to the excellent experimental work of Stockard, and partly to the fact that the cj^clopean state can exist quite independently of other marked deformities of the

^ Mall, F. P. A study of the causes underlj^ing the origin of human monsters. Jour. Morph., vol. 19, 1908.




onihryos. T haAo pi-fniously discussed tho rjuostion of cyclopia in a sr])ai'at(^ puhlicatiou, and it is not therefore necessary for me to dilate further ui)on it at ])resent."- Hare lip is also sharply defined in the embryo and is as readily recognized as exomphaly. Other anomalies, however, are more difficult to recognize as sharply defined malformations in the embryo.

We have in our collection about 2000 embryos. The pathological specimens of the first 400 were reported in my paper on the origin of human monsters mentioned above. Since the collection was taken over by the Carnegie Institution of Washington, it has gTown at a very rapid rate, about 400 specimens being added to it each year. I have in preparation a more extensive study of pathological embryos, and during the past year have practically completed a careful study of the first thousand. While this was in progress, another thousand specimens were added to the collection. At present, however, only the first thousand will be considered, the remainder not having been sufficiently tabulated to be of statistical value.

We have introduced and are gradually perfecting a system of classification of the embryos which will enable us to locate any specimen in our collection and the record thereof by means of a card catalogue. Reasons for adopting this S3stem were given in a circular recently published.-^ The specimens can clearly be divided into two groups according to their origin, i.e., uterine and ectopic. In both of these, the embryos which are normal in form are catalogued according to their sitting height, which we call crown-rump (CR). All embryos therefore which are apparently normal, say 10 mm. long, are entered upon one card. What happens to these specimens subsequently, whether they are dissected, sectioned or preserved permanently as whole specimens, may also be entered upon this card without interference with the system of classification. The chief difficulty is to determine what constitutes a normal embryo, and

- Mall, F. P. Cyclopia in the human embryo. Contributions to Embryology, vol. 6, Publication No. 226, Carnegie Institution of Washington, 1917.

^ Mall, F. P. Embryological collection of the Carnegie Institution, Circular No. 18, 1916.


here we must rely largely upon our experience in human and in comparative embryolog3^ A sharply defined, well formed white embryo, with blood vessels shining through its transparent tissues, is considered normal. If it is partly stunted and opaque or disintegrating, it is considered pathological. A further study of the normal embryo, however, shows that in many of these specimens the membranes are decidedly pathological. For instance, the villi may be deformed, diseased, atrophic or hypertrophic, or the contents of the amnion and the exocoelom may be unusual. Nevertheless, in all of these cases we still classify the embryos as normal, although fully cognizant of the fact that the surrounding membranes are pathological; otherwise it would be difficult to account for the great number of spontaneous abortions. The theory is that the embryo was developed under pathological conditions, but that the chorion was not sufficiently affected to cause any apparent change in the embryo. If an embryo included in this group is apparently normal in all respects save one, we still consider it normal with a localized anomaly. In fact we are gradually forced into this position, as an embryo, considered at first to be normal, may later on prove to have a localized anomaly, such as spina bifida or cyclopia. As far as we can determine, such an embryo would have been able to sur\'i\'e longer had not something happened to its membranes, thus causing its expulsion. I am inclined to believe that pregnancies of this sort, if carried to term, would produce the ordinary monsters described by teratologists. As the study of our collection of specimens is continued by different members of the staff, localized anomahes, wlien found, are recorded in our card catalogue, without, as stated above, necessitating any rearrangement. WTien these anomalies are present in normal embryos, the embryos are classed as normal, with localized anomalies.

The second group of specimens, which are termed pathological, are in a way more interesting, and their study justifies our method of classifying localized anomalies with normal embryos. We have in this group a variety of changes ranging from those found in fetus compressus down to complete disin


tegratioii of the ovum, leaving only a few villi. Having made numerous efforts to classify these s]ieeimens, I have finally lesolved them into seven groups which I shall consider in their reverse numerical order.

The seventh group, shown in figure 7, is composed mostly of larger specimens which are either dried up and deformed, or macerated and soft. These, of course, apparently merge into each other, and for this reason we have had to consider them as a single group. We hope, however, in the course of time to be able to subdivide them, for it is well known that fetus compressus is extremely rare in pigs and other lower animals, w^hile edematous and macerated embryos are quite common. It appears that the type of fetus in this group develops as a normal embryo during the first portion of pregnancy, and then dies slowly, either undergoing maceration, or being transformed into a fetus compressus. In the latter the cord is long, thin and greatly twisted The structures of the embryo show that there has been a slow^ tissue growth w^hich has not been sufficiently rapid to allow the normal development of the extremities. Instead the hands and feet are club-shaped, and in several instances there are adhesions beween the extremities and the body We also find very pronounced and quite characteristic changes in the placenta of the fetus compressus, there being bcw^een the villi lage masses of chromatin substance presenting much the same picture as the photograph of a comet, a central nucleus with scattered granules extending from it. Generally in our notes we speak of this substance as nuclear dust.

The sixth group of specimens we term stunted (fig. 6). The form of the embryo is easily recognized, but the head is atrophic as are also usually the extremities. At the time of the abortion the tissues are quite transparent, giving every appearance of a living embryo, but with, increasing knowledge concerning tissue cultures and growing isolated cells, we can see in specimens of this sort an active but circumscribed tissue culture of a clump of differentiated tissues. In other words we have a tissue culture of the entire embryo, w^hich on account of faulty or arrested circulation, growls in an irregular manner. Changes


of tliis sort ill an embryo I have designated in my paper on monsters as a dissociation of the tissues. I picture to myself something Hke the following sequence: when the ovum comes into the uterus which is more or less diseased, it becomes somewhat poisoned and consequently does not implant itself well. This naturally results in an irregular growth of the chorionic villi; in turn the embryo is affected and it is only natural to infer that the most direct influence would be through the vascular system, soon ending in poisoning of the heart and frequently in the interruption of the circulation. In such specimens the nutrition would reach the embryo through the exocoelom. In fact one of the earliest indications of a pathological specimen is an increased amount of magma in the exocoelom. Embryos, which are thus cut off from the chorion, continue to grow in an irregular manner; the tissues are more or less dissociated, and the specimen as a w^hole is stunted. Hence the designation.

In the fifth group (fig. 5), the process of stunting has progressed to such an extent that the extremities are almost entirely lacking and only the head end can be recognized with certainty. On account of their shape, due to this extreme stunting, we speak of these specimens as cylindrical embryos. Falling frequently into this group are embryonic remnants which, however, really do not belong there, since a primary examination with a binocular microscope does not permit of a sharp differentiation between this and other cylindrical forms of stunted embrj^os. Close examination wdth a microscope reveals specimens of this sub-group to be, composed of a naked umbilical cord belonging to an older embryo which had disintegrated, or as seen in a few instances the embryo has been torn off by mechanical means during abortion. As rapidly as the sub-type is recognized, it is labeled in the card catalogue in parenthesis (cord) so that in stud^dng these specimens we may distinguish between the naked cords and the true cylindrical forms of pathological embryos.

\Mien the process of dissociation of the embryo begins in still earlier stages than those belonging to the older groups (Xos. 0, 6, and 7), the result is a nodular body representing the embryo,


Imt tho cliango in it is so complete that it is difficult to recognize the diffei'cnt ])arts of the embryo except in a general way (fig. 4). The coelom, heart and central nervous system can readily be made out. Sometimes there are pigmented spots in one or two of the sections, marking the position of the eyes. This group again divides into two quite sharply circumscribed sub-groups: first, those with an umbilical cord to which the dissociated embryo is attached together ^vith the umbihcal vesicle; and second, a vesicular group composed of specimens in which there is only the remnant of the umbilical vesicle, the embryo being nearly or entirely destroyed. Had it been possible in every instance to differentiate between these two types of specimens in the primary examination, they would, of course, have been recorded as separate groups; but this could not be done without sections and a microscopic examination. Therefore, for the present we must consider them together. In our ordinary laboratory parlance we speak of them as the nodular group.

In the third group, both embryo and umbilical vesicle are completely destroyed, but we can see within the degenerated chorionic sac a more or less complete amnion. This group is designated as the one in which the specimens are composed only of the chorion and the amnion (fig. 3).

In the second group the amnion is destroyed and there remains only the chorionic vesicle containing the coelom. This is usually filled with reticular magma and scattered cells, which may represent all that is left of the embryo (fig. 2).

Finally in the first group the form of the ovum is destroyed and the specimen consists only of the \dlli which have undergone fibrous or mucoid degeneration. Sometimes only a few of the \dlli are found, at other times there is a large cluster clinging to a single stem, and some specimens are composed of large masses of villi which form malignant hydatidiform moles. Such a mass may weigh a kilogram (fig. 1).

It can be readily seen that the above classification into subgroups is arranged somewhat in the order of the age of the ovum when it began to degenerate. Generally these changes are so pronounced that the embryo cannot live through the duration of pregnancy and this accounts for the abortion.



As far as localized anomalies are concerned, we naturall do rot find them in the first four groups, while in the remaining three groups we encounter only such as are very pronounced and stand out clearly in spite of other changes in the emtryo.


Fig. 1 lllustratinji (Jroup 1, composed exclusively of villi. Specimen Xo. 749 received from Dr. G. C. McCormick. Sparrows I'oint, Md. X 2. These hypertrophic villi came from a hydatiditorm mole weighing over a kilcgr: m.

Fig. 2 Illustrating Group 2, chorion with coelom. No. 12S9 from Dr. J. R. Cottell, Louisville, Ky. X 2. The picture shows the coelom filled mostly with granular magma.

Fig. 3 Illustrating (jroup 3. chorion with amnion. Xo. 813 from Dr. H. D. Taylor, Baltimore. X i The cavity of the ovum is filled with a dense mass of granular magma.



Thus, for instance, with fetus compressus we frequently recognize club-foot; in stunted forms, hare lip and spina bifida, and in cyUndrical forms, spina bifida. Of course, if cyclopia is encountered in any of these forms, it is looked upon as a localized anomaly in a pathological embryo. On the other hand, a single anomaly in an embryo called normal can easily be recog


1 ig. 4 Specimen illustrating Group 4. The ovum contains a nodular embryo; Ko. 1140b from Dr. George T. Tayler, Greenville, S. C. X U Fig. 5 Illustrating Group 5. Ovum containing a cylindrical embryo; No. 839 from Dr. W. S. Miller, Madison, Wis. X li

nized, and it is from this group that we should expect the development of monsters had the pregnancy progressed to term.

A few illustrations of localized anomalies are given in the figures in order to show that they are identical with those found in infants at birth. Figures 8 and 9 are specimens of cyclopia and double monsters in normal embryos. Figure 10 a and



10 b show an embryo and a fetus with hare lip. Figure 11, 12 and 13 have pronounced localized anomalies and need no further explanation. Finally figures 14 to 18 show anomalies of the hand the first and last are of the hereditary variety, and figures 15 and 16 show acquired anomalies, that is, they were subsequentl}^ formed in an embryo which started its development normally. It is proper to remark here that these illustrations are mostly from specimens from the second thousand of our


Fig. G Group 6. Three stunted embrj-os to illustrate this group. 6a, No. 1295d from Dr. B. T. Terry, Brooklyn, N. Y. X 4. 6b, No. 1523 from Dr. G. B. Ward, Gilman, Iowa. X 2. 6c, No. 1477 from Dr. H. B. Titlow, Baltimore. X 3.

collection but this is for the reason that recently we have made many more photographs and secondly, many of the specimens in the first thousand have already been figured in my paper on monsters.

In order to render possible a comparison between localized anomalies found in pathological and those found in normal embryos, the following six tables have been constructed. Table



1 gives the classified distribution of the first 1000 embryos in the Carnegie Collection. The primary division comprises two classes — pathological and normal. The pathological in turn is arranged in the seven groups just described. The normal are arranged in groups to correspond as nearly as possible with the



Fig. 7 Croup 7, giving two specimens of fetus compressus. 7a, No. 996 from Dr. H. B. Titlow, Baltimore, X r. 7b, No. 868 from Dr. E. H. Egbert, Washington. X 2.

ages of the embryos in lunar months. In order to define clearly which embryos belong to a given month, I have inserted their probable lengths for each month in table 6. Thus, for instance, the second month includes all specimens from 2.6 mm. to 25 mm. in length, etc. (Data upon the estimated age of embryos



may be found in my chapter on the age of embryos, contained in the Manual of Human Embryology.)^

It will be noted in these tables that the specimens are arranged in centuries; that is, each line in the table includes exactly 100 specimens. The first century includes specimens Nos. 1 to 98, the second, Nos. 99 to 205, and so on. This adjustment was necessary for the reason that frequently a single number is given to two or more specimens. Sometimes the


Fig. 8 Normal embryo with cyclopia; in front of the eye is seen the Cyclopean snout. No. 559 from Dr. B. J. Merrill, Stillwater, Minn. X 5.

Fig. 9 Normal double monster. No. 249 from Prof. L. Hektoen, Chicago. Natural size.

first is called a and the second, b; or the first may be given the number, and the second the letter a, etc. The second century passing from Nos. 99 to 205 includes more than 100 numbers, because specimens which are given a number are frequently found upon further examination not to contain any remnants of the ovum, and for this reason they are to be discarded. In our catalogue they are later marked as 'no pregnancy.' Finally the full 1000 ends with embryo No. 900g. The individual entries are percentage records. Thus in the fifth century, there

^ Determination of the age of human embryos and fetuses. Human Embryology, Keibel and Mall, vol. 1, Chap. 8. 1910.




10 b

Fig. 10 Two specimens of hare lip. 10a, No. 364 from Dr. B. .1. Merrill, Stillwater, Minn. X 3. There is also exencephaly in this specimen. 10b, No. 9S2 from Dr. G. C. ^NlcCormick, Sparrows Point, Md. X 2.




Fig. 11 Stunted fetus with a large hernia in umbilical cord, also spina liifida. No. 1330 from Dr. A. R. Mackenzie, Capitol Heights, Md. X H.

Fig. 12 Normal embryo with exencephaly and spina bifida (the latter opposite the arrow). No. 1315 from Dr. J. C. Bloodgood, Baltimore, X 2.

Fig. 13 Normal fetus with hernia of midbrain. No. 1690 from Dr. P. F, Williams, Philadelphia. X 9.10.



are 41 normal specimens of the second month; that is, of this hundred, 41 per cent of the specimens are normal embryos of the second month, whereas the total for the full 1000 has been brought down this percentag;e to 24.5.



Fig. 14 Anomaly of the left hand in which only the thumb and little finger are normal. No. 306a from Dr. F. A. Conradi, Baltimore. X f.

Fig. 15 Left hand which is club-shaped from a fetus compressus. No. 230, CR 57 mm., from the late Dr. J. P. West, Bellaire, Ohio. Natural size.

Fig. 16 Deformed wrist with atrophic radius in a normal embryo. No. 789, CR 50 mm., from Dr. H. F. Cassidy, Roland Park, Aid. X 2. The same kind of wrist is seen in the specimen illustrated as figure 11.

Fig. 17 Right hand with six fingers from macerated specimen. No. 1749 from Dr. S. M. Wagaman, Hagerstown, Md. X 2. This specimen had six digits on all four extremities.

Fig. 18 Double little finger of the left hand of the same specimen. X 2.




(iin'ntj llic ilixirllmtioti (if 1001) specinicin




























Ed u
















1- {'8



























































































































730-8 16b


















9 36

8 71


4 51

7 75

4 80

6 62

38 396

1 11

14 245



7 93

8 41





3 6



In determining the normality of specimens, the criterion used was the shape of the embryo, judging this as best we could by our own knowledge of human and comparative embryology, as well as by the experience of other students of human embr^^ology, and we have used freely the atlases of His, Hichstetter and Keibel and Else in making our decisions on this point. How

table 2

Specimens obtained from the uterus





1 1



3 1









EO E 24

1 3















1- 98


















206- 295

























































































































2 12

4 39




5 67



5 55

21 313

1 11

11 213

18 170

7 91

8 41

6 18




3 6





ever, many of thevse specimens are enclosed in membranes which have undergone very marked changes. Thus, an embryo, normal in form, may be found surrounded by an excessive amount of magma, and the chorion may have undergone very pronounced changes; but for purposes of classification we have found it necessary to arrange them all according to the shape of the embryo. A fairly large number of our specimens were


Ecotopic specimens



a o







< o










< o


1- 98


































































730-8 16b

































Specimens showing localized anomalies (a be compared with table 1












7 1

"3 2

] 1

2 4











1- 98























































































obtained from hysterectomies, and we believe with Hochstetter that we shall ultimately have to determine what constitutes a normally formed human embryo from specimens obtained in this way. However, even by this method we have found among about 25 specimens 3 markedly pathological ones undergoing abortion.

The second table includes all specimens that were obtained from the uterus, and the third, all ectopic specimens. Thus, in making a comparison of these three tables it will at once be noted that among the entire 1000 nearly 40 per cent are pathological embryos and ova. Of this number, 31 per cent were obtained from the uterus alone while slightly more than 8 per cent were ectopic. The comparative frequency of pathological and normal embryos can be ascertained, however, by comparing them within a given century, or for the whole thousand together. In the uterine specimens about one-third of the ova and embryos are pathological, as compared to two-thirds in the actopic. In other words, pathological specimens are twice as frequent in ectopic as in uterine pregnancy.

The fourth table includes all the specimens in which there are pronounced localized anomalies. The character of the anomaly is given with the individual specimens which are recorded in tables 5 and 6. It is interesting to note that these tables show that there are about as many anomalies among the normal as among the pathological specimens, but when these figures are compared with the total of specimens both normal and pathological, it becomes e\ddent that localized anomalies occur about twice as frequently in the pathological as in the normal embryo. Thus, there are 38 locahzed anomalies among 396 pathological specimens or about 10 per cent, while the occurrence of localized anomalies in 604 normal specimens is about 6 per cent. The table shows further that the 38 pathological specimens with locahzed anomahes abort in the early part of pregnancy and only one of them (No. 649) grew to a sitting height of 90 mm., that is, about the middle of the fourth month.

Among the normal embryos, those with localized anomalies usually disappear before the fifth month, there being but one in the sixth, one in the eighth, and four in the tenth month or

TABLE 5 Localized anomalies in pathological embryos




w « 

<«2 •a «  o o

S o 3



m m .

m m .






Spina bifida









Spina bifida






Group 5









Spina bifida







Eye detached from brain




Amyelia-Ectopia of heart








Anencephaly — Spina bifida










Spina bifida




Spina bifida




Spina bifida





Anencephaly i








Rounded head — Club leg

Group 6





Exencepahaly — Hare lip —


Exomphaly — Spina bifida





Hare lip — Spina bifida








Club foot and hand




Clul) foot and hand



Spina bifida — Exomphaly — Without radii and without thumbs



Head defective — Spina bifida







Head atrophic







Spina bifida





Anencephaly — Spina bifida

Group 7




Spina bifida





Club foot

and fetus



Club hand and foot. Hand


adherent to head. Skin nodulas




Club hand and Club foot



Club foot




Club hands and feet



Club foot







TABLE 6 Localized anomalies in normal embryos




(26-68 mm.

(69-12 mm.)

-i D

o o

K »
































































2 °

W 6,



10x9 x8 18x18x18

16x14x12 17x17x10 25x20x15

24x18x8 40x28x28 35x20x17 30x20x15





50x50x70 50x50x70 40x40x40














Ancncephaly — Spina bifida

Spina bifida

Spina bifida

Anomalous tracheal diverticulum


Spina bifida

Deformed tail

Spina bifida

Spina bifida

Leg hypertrophic — Head atropic


Spina bifida


Spina bifida

Constricted cord

Spina bifida

Double monster

Double monster

Cyst of spinal cord

Hernia of liver

Hernia of liver

Atrophic head

Hernia of liver

Double monster Double Monster Extremities deformed — Left

radius probably absent.

Head acbophic

Stub coccyx

Left forearm and hand

wanting Only 2 fingers on right

hand Pounded head — Thickened





T.VBLE 6— Concluded.




o c X « 







(122-167 mm.)

No sp



(168-210 mm.)





(211-245 mm.)

No specimen


(246-284 mm.)



Spinia bifida


(285-316 mm.)

No specimen


370 After b


Enormous tail

10 1

862 At birth

Ectopia of bladder

(317-3.36 mm.) |

862a At birth

Spina bifida


862b At birth

Stunted eyes

at the end of pregnancy. In other words, all pathological specimens, either with or without localized anomalies, abort in the first half of pregnancy ; while nearly all so-called normal embryos with slight malformations are also aborted before the middle of pregnancy, very few of them reaching full term.

We have made an especial effort to collect specimens of full term monsters as well as abortion material from all months of pregnancy. Only the first 100 specimens of the collection show an unusually large percentage of normal embryos. Although at first an effort was made to collect only good normal specimens the last 900 specimens, including all sorts of material, of the collection carry about the same percentage of normal specimens throughout. The first 1000 specimens of our collection is short of fetuses from the second half of pregnancy, but we are now endeavoring to collect material covering all months of


pregnancy. One monster at term, a sympus belonging in about the third hundred, was not recorded in our catalogue, and should be added to the four full term specimens given in table 4. This means that among 1001 specimens there were five monsters at term, while among 1000 specimens there were 71 with localized anomalies, most of which w^ere aborted early in pregnancy.

According to the table on the frequency of abortions given in my monograph on monsters,^ there are 80 full term births for each 20 abortions; therefore, the 1000 abortions under consideration were probably derived from 5000 pregnancies.

As we have calculated that there should be approximately 30 full term monsters in 5000 pregnancies, and as 5 of these were observed in our 1000 specimens, it is apparent that the remaining 25 should be encountered in 4000 additional full term births. When these figures are compared with the fact that 75 localized anomalies occurred in 1000 abortions — 7.5 per cent, it becomes apparent that in any similar numbers of abortions, localized anomalies should be noted twelve times as frequently as monsters at term. A similar result is obtained if the number of localized anomalies of the tenth month, as given in table 4, is compared wdth all of the localized anomahes of pre\dous months, as given in the same table."

Our studies seem to justify the conclusion that pathological embryos, as well as those which are normal in form, are very frequently associated with localized anomalies and that abortion usually follows as a result of serious lesions in the chorion, as well as in its environment. Should the alterations in the embryo and in the chorion be very slight, and the condition of the uterine mucous membrane, which may be expressed by the term inflammation, be overcome, the pregnancy in all probability would go to term and end in the birth of a monster or an infant presenting a well recognized malformation.

^ Also in a resume of the paper on monsters in the article entitled: Mall, F. P. Pathology of the human ovum. Chapter 9, Human Embryology, Keibel and Mall, vol. 1, 1910.

® Records are now being made of about 50,000 births in Baltimore, including the frequency of abortions for each mother. When these are completed, the above mentioned ratio of 1 to 4 will probably be changed.


I have already pointed out the difference in, frequency of malformations and destructive changes as observed in the ovum in tubal and in uterine pregnancies. Since the pubhcation of my monograph on monsters, I have reconsidered the question of tubal pregnancy, and the specimens mentioned in the present paper are recorded in detail in a book on tubal pregnancy recently published.^

It seems to me that the studies based upon our collection of embryos as well as recent investigations in experimental embryology, set at rest for all time the question of the causation of monsters. It has been my aim to demonstrate that the embryos found in pathological human ova and those obtained experimentally in animals are not analogous or similar, but identical. A double monster or a cyclopean fish is identical with the same condition in human beings. In all cases, monsters are produced by external influences acting upon the ovum; as, for instance, varnishing the shell of a hen's egg or changing its temperature; traumatic and mechanical agencies magnetic and electrical influences, as well as by alteration of the character of the surrounding gases, or by the injection of poisons into the white of an egg. In aquatic animals, monsters may be produced by similar methods. Whether in the end all malformations are brought about by some simple mechanism, such, for instance, as alteration in the amount of oxygen or some other gas, remains to be demonstrated. The specimens under consideration show such marked primary changes in the villi of the chorion and in the surrounding decidua that the conditions in the human may be considered equivalent or practically identical with those created artificially in the production of abnormal development in animals.

It would have been quite simple to conclude that the poisons produced by an inflamed uterus should be viewed as the sole cause, but when it is recalled that pathological ova occur far more commonly in tubal than in uterine pregnancy, such a theory becomes untenable. Moreover, monsters are frequently

' Mall, Franklin P. On the fate of the human embryo in tubal pregnancy. Publication No. 221, Carnegie Institution of Washington, 1915.


observed in swine and other animals without any indication of an inflammatory environment. For this reason I have sought the primary factor in a condition buried in the non-committal term fault}^ implantation. It would seem to be apparent that lesions occurring in the chorion as the result of faulty implantation, can and must be reflected in the embryo. For example, before circulation has developed, in a human embryo, pabulum passes from the chorion to the embryo directly through the exocoelom, and probably on this account we always encounter, as a first indication of pathological development, a change in the magma. In older specimens before any other changes are noticeable in the ovum, the magma become markedly increased, and a variety of changes are found between the villi. I shall not dwell further upon magma as I have recently dealt with the subject in detail.^

It is perfectly clear that monsters are not due to germinal and hereditary causes, but are produced from normal embryos by influences which are to be sought in their environment. Consequently, if these influences are carried to the embryo by means of fluids which reach it either before or after the circulation has become established, it would not be very far amiss to attribute these conditions to alterations in the nutrition of the embryo. Probably it would be more nearly correct to state that change in environment has affected the metabolism of the egg. Kellicott, who has recently discussed this question, seems to be disinclined to accept such an explanation, but I do not see that he has added materially to it by substituting the word disorganization for nutrition as one might as easily say that the altered nutrition causes the disorganization. ^

In my paper on monsters I stated that on account of faulty implantation of the chorion the nutrition of the embryo is affected, so that, if the ovum is very young the entire embryo is soon destroyed, leaving only the umbilical vesicle within the

Mall, Franklin P. On magma reticule in normal and in pathological development. Contributions to Embryology, vol. 4, Publication No. 224, Carnegie Institution of Washington, 1916.

3 Kellicott, W. E. The effect of lower temperature upon the development of Fundulus. Am. Jour. Anat., vol. 20, 1916.


chorion, and this also soon disintegrates, leaving only the chorionic membrane which in turn collapses, breaks down and finally disappears entirely. In older specimens, on the other hand, the process of destruction takes place more slowly and thus we account for a succession of phenomena which correspond with the seven groups of pathological ova recognized and given in the various tables appended.

In my original study, I really went, I believe, a step farther than Kellicott in his discussion of monsters, as he dropped the subject by stating that the embryo is a monster simply because it is disorganized. I attempted to analyze the process of disorganization more thoroughly and demonstrated that when disorganization begins it is accompanied by cytolysis, but as it progresses more rapidly it results in histolysis, and that these two processes do not act with equal severity on all parts of the embryo. When we consider the whole ovum, it is the embryo itself which is first destroyed; while within the embryo the central nervous system or the heart is the portion which is first affected. Morphologically, these changes are accompanied by a destruction of certain cells and tissues, leaving other portions which continue to grow in an irregular manner. For this reason I speak of the tissues which are first affected as more susceptible than the others. The entire process of disorganization, resulting in an irregular product, I have termed dissociation. In a general way this explanation is accepted by Werber in his recent studies, but he employs the term blastolysis instead.^"

At the time I prepared my paper on monsters, Harrison was just beginning his interesting experiments in tissue culture in our laboratory. Since then this method of study has given us clearer insight into the independent growth of tissues. I was fully convinced from the study of pathological embryos that tissues continue to grow in an irregular manner, thus arresting normal development; but since we are more familiar with the

10 Werber, E. I. Experimental studies aiming at the control of defective and monstrous development. A survey of recorded monstrosities with special attention to the ophthalmic defects. Anat. Rec, vol. 9, 1915. Also: Blastolysis as a morphogenetic factor in the development of monsters. Anat. Rec, vol. 10, 1916.


growth of tissues, as revealed by Harrison's method, we can understand a Uttle better the process of dissociation. In fact we have in our collection two striking examples of tissue culture in human embryos. In one, the cells had formed an irregular mass which is growing actively, but the contour of the organs has been entirely lost. In the other, from a tubal pregnancy, for some unknown reason, the ovum had been completely broken into two parts, which in turn had cracked the embryo, and from each piece had been a vigorous independent tissue growth, or, as we may now say, a tissue culture. Accordingly, when an embryo through changed environment is profoundly affected, the development of one part of the body may be arrested, while the remaining portion may continue to grow and develop in an irregular manner. In very young embryos tissues or even entire organs become disintegrated, as can easily be recognized by the cytolysis and histolysis present, and the resultant disorganized tissue cannot continue to produce the normal form of an embryo. If this process is sharply localized, for instance, in a portion of the spinal cord or in the brain, spina bifida or anencephaly results. To produce a striking result, as in cyclopia, a small portion of the brain must be affected at the critical time, and I think the work of Stockard has shown clearly that this is before the eye primordia can be seen. Consequently, in order to produce a human monster, which is to live until the end of gestation, the altered environment must be reflected from the chorion to the embryo, so that the tissue to be affected is struck at the critical time in its development. It is inconceivable that cyclopia should begin in an embryo after the eyes are once started in normal development. Moreover, the same is true regarding hare lip, for after the upper jaw has once been well formed, the abnormality cannot develop. We may extend this statement to include club-foot, spina bifida occulta, and other types of malformation. In fact, in discussing the origin of merosomatous monsters, hardly more has been stated by most authors than that there has been an arrest of development, but I have attempted to point out that the primary cause is in the environment of the egg and that the arrested development is associated with destruction of tissue.



Osborn Zoological Laboratory, Yale University, New Haven, Connecticut, and

Rockefeller Institute for Medical Research, Department of Animal

Pathology, Princeton, Neiv Jersey


The writer, employing the bone marrow of the chicken for attenuating the virus of cyanolophia (Erdmann '16^), by culture of the marrow and the virus in a medium of chicken plasma, has observed some interesting facts concerning the cytological changes in the bone marrow cells.

The morphology and development of chicken bone marrow and its relation to blood formation have been described by few authors. Dantschakoff ('09, pp. 859-65) gives an extensive review of the literature on these questions and establishes our knowledge of the origin of the different elements of chicken bone marrow\

In studying the cells of bone marrow in plasma culture medium, we must take into consideration the fact, that we add to the plasma in which the tissue culture is cultivated a heterogeneous mixture of highly differentiated cells. Chicken bone marrow has a loose framework of slender connective tissue cells, in the meshes of which blood and fat cells are scattered. The blood cells — eosinophils, erythrocytes, and myelocytes — form, according to Foot ('13, p. 45) strands and circles between and around the fat cells. The blood islands represent collections of cells of microlymphocytic and macrolymphocytic types, of more or less ripe erythrocytes and of young connective tissue cells. It must be clearly kept in mind that all these different

^ Received for publication March 14, 1917.

2 Erdmann, Rh. 1916 Attenuation of the living agents of cyanolophia, Proceedings of the Society for Experimental Biology and Medicine, vol. 8, pp. 189-193.



elements behave differently in the tissue cultures and may, after having undergone important changes in the plasma, offer some difficulties in interpretation.

The only observations of normal chicken bone marrow in plasma are those made by Foot '12 and '13. In the first series of experiments he studied especially the behavior of the fatty elements of chicken bone marrow, recording the following results. Six hours after implantation numerous cells leave the tissue center. They form rays of cells liquefying the plasma. These rays are formed by polymorphous leucocytes wdth eosinophile granules and by eine Art von mononuklearen basophilen Zellen" (p. 450). Foot gives the latter the name of X cells; they are the most important and they contain only fat according to his observations of 1912. They form, he says in 1912, the bulk of all cells migrating into the surrounding plasma. These X cells, the origin of which Foot tries to elucidate, are true phagocytes They include small fatty droplets and other particles which are dispersed in the cytoplasm. On the fourth day, these cells, after having been enlarged by the amount of fat which they have taken up during the first three days in the culture, form either syncytial masses or a widely spread network of anastomosing cells. The former may divide, after having lost most of their fatty granules, and form the final 'ruhende X Zelle' (Foot '12, fig. 8, pi. 22) : or the latter, after having been highly vacuolized, as stated by Foot '12, may form fibrils (fig. 18, pi. 22). If these X cells do not form resting X cells or cells which produce fibrils, they take the shape of 'Riesenzellen.' These ' Riesenzellen' are not identical, in Foot's opinion, with the 'giant' cells of the normal bone marrow. They are represented in his figures 11, 16, 17, 19. They are only X cells which have fused together, form no fibrils, and may later break up in small cells (figs. 12 to 14), which have generally one nucleus. Das Ergebnis der Aufteilung der Riesenzellen ist sozusagen eine neue Zellrasse" (p. 460) — cells adapted to the condition of the medium.

Foot believes that the X cells are transformed cells of the 'mesenchyme' and "Zwar indifferent gewordene Mesenchymzel


len" (p. 4()G). He reasons as follows: Because these cells have the potentiality of forming fibrils they must belong to those cells which can form connective tissue, and therefore these X cells without any intermediate stages take their origin from mesenchymal or endothelial cells. In a postscript to this paper he changes his opinion entirely and says (p. 475): "Was die Herkunft der X Zellen betrifft, so scheint es als ob die Hauptmasse derselben entweder direkt oder indirekt von den lymphocytaren oder myeloblastischen Elementen des Knochenmarkes abstammte," promising to give the reasons for this change of opinion in his second communication.

After a careful study of Foot's second publication ('13), which is rather difficult to understand because he does not very often connect his first publication with the second, I restate in his own words his revised opinion of the origin of those cells which form X cells ('13, pp. 46-47). The deductions as to the transformation of the lymphocytes from one form to another, which form the basis of the following descriptions, were made from the observation of transition forms. The later transformations of these cells into forms resembling fat and giant cells or cells of the connective tissue have been considered in my previous article." So it appears that the so-called X cells of this author ('12) — the name does not often appear in the paper of 1913 — are not directly transformed cells of the mesenchymal type but are said to be of lymphocytic origin. He observes that as early as three hours after implantation of the bone marrow a considerable number of microlymphocytes emigrate from the tissue particle. Their transformation occurs in the following way:

The small microlymphocytes are first transformed into macrolymphocytes, later into large mononuclear forms, then into myelocytes. At last the polymorphonuclear leucocytes appear, after having undergone difi"erent changes in the form and structure of the nucleus. The nucleus is at first horseshoe-shaped, later polymorphonuclear and even polynuclear. Finally the cells, by rounding off and dechromatization of the nucleus coincident with the rarification and a change in the staining


properties of the plasma, are transformed into the cell culture type (p. 56). This cell culture type (see his fig. 2, pi. 3, and his fig. 3, pi. 4) represents small polymorphonuclear leucocytes (p. 49) which have undergone the transformation, but not only does the cell culture type originate from lymphocyte forms, but this 'stem cell' can also be transformed through the transition stage of amoeboid forms into 'giant cells,' syncytia, and, as said before, into the cell culture type (table 1, p. 56).

Thus it is clear that, according to this author's view, all the different forms described by Foot in 1912 and 1913 originate from the microlymphocytes. Until the present time ('16) this important fact lacked verification, but by the cultivation of the virus of cyanolophia in chicken bone marrow an opportunity was afforded of observing the changes which Foot describes. A careful study of the morphological and cytological characters of the cells figured in the above mentioned papers, soon showed a lack of transition stages, which are needed as proof of Foot's final theory. Further, the nuclei of cell forms which are said to be transformed into each other do not show close resemblances, e.g., the cells in figures 1 and 3, 1913, which are said to be eosinophil leucocytes at different stages of incubation, have different nuclear structure as well from each other and from the cell of the cell culture type (fig. 2, left side, 1913). The nuclear structure of this particular cell (fig. 2, left side, 1913), however, has a certain resemblance to the nuclei shown in 1912, figures 5 and 6. These cells are considered by Foot as stages connecting the ' Riesenzellen' with "eine Art von monnuklearen basophilen Zellen" (1912, p. 450). But here, as far as could be judged from the draA\ings, the cytoplasm of the cells in figures 5 and 6 is very different. Figure 5 has granules, figure 6 does not show them; only traces of digested nuclei of other cells are visible. These contradictory facts present a priori difficulties in accepting the views of Foot. But they appeared far more disconcerting on examining the cells themselves.




It is not necessary to describe in detail the technique of these cultures, since the writer followed the same methods as those used by Harrison ('10), Burrows ('11), and particularly Foot ('12 and '13). For storing the plasma it was deemed important to use the methods described by Walton ('12) for keeping mammalian plasma in good condition for long periods of time. Great stress was laid on the study of the living cells, and a warm stage was used to follow out the transitions of one cell form into another. The bone marrow of very young chickens, those of medium age, and of old individuals was studied; observations were also made on bone marrow which contained a very small amount of fat, as well as that which had a large amount of fat.

The method described below gave the best results in identifying and showing the stages of the individual cell types in stained preparations. A small particle of bone marrow was put into the plasma medium. The cells in the tissue were then allowed to migrate out of it. At periods of either 2, 4, 6, 12, or 24 hours, the original particle of bone marrow was extracted, and the fate of those cells which had emigrated was studied. The writer found that from the original particle of tissue numerous cellforms had been sent into the surrounding plasma clot. Having thus extracted the bone marrow, it could be determined with absolute exactitude which cell-forms emigrated first, and the history of those cell types which had emigrated after 2, 4, 6, 12, or 24 hours, or at any given period, could be recorded. The extracted particle of bone marrow was now transferred to a new plasma medium and the cell forms which emigrated after the transfer were also observed. This was repeated several times, until practically all emigration of cells into the surrounding plasma had ceased. The structure of the remaining particle of bone marrow was of course studied. Smears and sections were made at every stage of the emigration process and a more complete history of this complicated process was thus obtained.

In staining the pieces of bone marrow, the methods used by Foot in 1912 and 1913 were followed and other methods for the


discovery of fat were added (see descriptions of plates, page 118. Besides these, the Giemsa stain after moist fixation according to the prescription of Giemsa proved to be very satisfactory. No dry smears of bone marrow were used.



The experiments from which the drawings on plates 1 and 2 were made were started on December 25, 1915, and on January 3, 1916. The bone marrow was taken from a full-grown chicken which had a large amount of fat, so that the pieces of marrow have a yellowish-white appearance. The first cells to leave the tissue after 40, 60, and 90 minutes incubation are, as Foot rightly remarks in his publication of 1913 (p. 49), small mononuclear or larger polymorphonuclear leucocytes (fig. 1). The forms have a very dark, granulated cytoplasm and are actively amoeboid (fig. 1). Pale mononuclear forms without granulations but with their characteristic vesicular nucleus, follow closely the emigrating polymorphonuclear leucocytes. The fourth cell from the left (fig. 1) represents an erythroblast. The structure of the nucleus makes this evident. Besides these forms figured in figure 1, red blood corpuscles and a few fat cells were present in those parts of the plasma clot which surround the implanted bone marrow particle. The network of the bone marrow was injured by the process of cutting and tearing the particle into small pieces, and it is therefore not surprising that a large number of red blood corpuscles and some fat cells were scattered into the surrounding plasma clot. They are not figured in figure 1.

After 24 hours various other cell types have migrated into the surrounding plasma.

Figure 2 shows bone marrow which has been in the plasma for 24 hours, from January 3 to January 4, 1916. We can easily distinguish two different kinds of granulocytes: big cells which have round, shining granules, the nucleus nearly half as big as the cell and half-moon shaped; and smaller forms, with very dark granules, the latter not rounded but more rod-shaped, the


nuclei spherical and very often dividing. It is impossible to define without doubt the exact type of these granulocytes before the relation of their granules to basic or acid stains develops the true character of these cells. Therefore we do not venture any interpretation of the bigger type of these granulocytes but point out only that the smaller forms must be eosinophil leucoC3'tes after their morphological structure, though their granules appear rather darker than those in non-incubated leucocytes of chicken-bone marrow. Also they have less distinctly round or less rod-shaped granules. These two observations are important. The big cell in the center of the figure 2 does not contain any granules but is from the large nongranular mononuclear lymphocyte type. Very often these cells break into pieces during observation.

Two other cells, one on the right, the other on the left side of figure 2 are of a different type They contain large shining droplets, the fatty nature of which seems doubtless. Their nuclei have a vesicular structure and appear at this stage of the culture as often dividing. They are less numerous than the eosinophil leucocytes which form, in the first 24 hours, the bulk of all cells migrating into the surrounding plasma medium.

Figure 3 represents bone marrow which has been incubated for 48 hours (January 3 to January 5, 1916). Here a 'Riesenzelle' is rapidly moving; its cytoplasm is spread over a great area on the cover-glass and contains fat droplets and glistening granules. This ' Riesenzelle' shows in its cytoplasmic structure a close resemblance to the fat droplet containing cells on figure 2. To account for the larger size, we can either suppose that several of these cells have fused together or the cytoplasm of a single cell is thinned out by the method of cultivation.

The structure of the granulocytes is not very much changed. The larger forms with glistening granules and half-moon-shaped nucleus have diminished in number but smaller cells of the same type can be discovered now and then. In these forms sometimes fat droplets are visible. The eosinophil leucocytes are still abundant, but are surpassed in number by small ungranulated cells. These form now the bulk of the cells migrating into


the surrounding plasma clot from the implanted tissue particle. They have either vesicular, less refractive or very shining and highly refractive nuclei.

In plate 2 we can follow in detail the further changes of the 'Riesenzellen.' The bone marrow (fig. 4) has been implanted 72 hours, from January 3 to January 6, 1916. Three round cells with big fat droplets can be seen, which seem to protrude out of the cell or cover its surface. The nuclei are therefore very seldom visible. When visible, they appear dark. A few granules are contained in the cytoplasm besides round or irregularly shaped masses, which seem to be remnants of other cells. On the third day after implantation these cells immediately attract the attention of the observer. They seem to have taken the place of the 'Riesenzellen;' this could be demonstrated by observation of the living cells. Some ' Riesenzellen' break apart, take on a round shape and completely extrude the fat droplets. These may be small or larger (fig. 5, second cell, left side) and show very fine pseudopodia. They are round cells which can survive an indefinite time in the plasma medium, the so-called 'cell culture type.'

Many 'Riesenzellen' however (fig. 5), the similarity of which to the round cells seen in figure 4 can be easily discovered, show all signs of degeneration. The cytoplasm has a 'curdled' appearance and is torn. The fat droplets have been thrown out into the plasma clot, and the granules have acquired a dark appearance. This regressive process takes place on the fourth or fifth day after implantation. These decaying cell masses are surrounded by small ' granulated and ungranulated cells and seem to be able to phagotise, because their cytoplasm shows in some places 'curdled granules.'

During the next days of incubation, no striking changes take place. The number of living cells diminishes and few tjrpes of cells are in healthy condition.

Fig. 6 shows cells which have been incubated in the same plasma medium 216 hours (from December 25 to January 3). They have small distended nuclei which do not seem to contain much chromatin, and the cytoplasm is filled with shining


droplets. They belong to the so-called 'cell culture' type. Besides these cells we find others with oblong nuclei and elongated cytoplasmic bodies full of glistening fine granules. These move slowly and show fine pseudopodia formed by their delicately granulated cytoplasm.

To summarize: Fat containing bone-marrow of chicken when incubated for 9 days in a plasma medium, undergoes the following changes which can be observed in the living preparation: The signet-like fat cell disappears, it is transformed to 'Riesenzellen' and finally to the 'cell culture' type. This type includes round cells with coarsely granulated cytoplasm, big shining droplets and oblong, less refractive nuclei. The other prevailing cell-form is distinguished by its finel}^ granulated cytoplasm, elongated or round cell body, and oblong nucleus.

These two cell types (not widely different in their morphological bearing) are always to be found among the cells which have migrated from the implanted bone-marrow particle into the plasma clot. Besides these cell forms, — capable as it seems of metabolism for long periods, — we see all forms of disintegrated cells. The cytoplasm and nucleus separate and the preparation is filled with debris. Fat droplets of different sizes which are freed from the cell fill the preparation. Nuclei of small granulocytes and lymphocytes without cytoplasm are often seen. Also shadows of blood corpuscles and granulocytes of all sizes are present.

It is certain that in non-renewed tissue culture retrogressive and progressive processes take place. It will be necessary to investigate the more intimate phenomena of these changes in stained preparations specially adapted to the study of each different cell type by different methods of cultivating and staining.




While describing the changes of the living bone-marrow cells after they had been 1, 24, 42, 72, 96, and 216 hours in the plasma medium, — the present author could give little or no definite



interpretation of the changes observed in the different types. Some exact knowledge could be acquired only by comparing and combining the phenomena observed in bone-marrow cells in preserved and stained preparations after they had been in the plasma medium for, well defined periods.

In figure 7, an exact microscopic field of a bone marrow preparation, after 36 hours incubation, is shown. The implanted tissue particle would be (if shown on the drawing) on the left side of the preparation. The cells shown have migrated to the zone next to the implanted bone-marrow tissue particle whichwas taken from a full-grown chicken and contained fat

Eosinophil leucocytes in various developmental stages are numerous. They are in rapid amoeboid movement, and by continued fragmentation diminish in size and multiply in number. Their plasma is slightly basophil. The nuclei are strongly chromophil and the nuclear leucocytic structure in most forms is indistinctly developed. By comparing the nuclear structure wdth that of eosinophil leucocytes which have been 24 hours in cultivation (fig. 9) we can better distinguish the typical leucocytic network of chromatin particles and threads. The plasma of these leucocytes and of those figured in figure 8, which have been only one hour in the plasma medium, is acidophil and the round granulations are very distinctly recognizable.

Besides the changes in the cytoplasm of the leucocytes from acidophily to basiphily, other phenomena are noticeable. After one hour and still more after 36 hours incubation, the leucocytes of all sizes are losing and expelling the granulations. The nuclei of these forms have either become pale and indistinct (fig. 7, right side, below) or condensed and strongly chromatic (figs. 12 to 14). They may fade out to mere shadows and disappear.

The farther the polymorphonuclear eosinophil leucocyte advances into the plasma clot, the more its cytoplasm spreads out in the tissue culture. The granulations in consequence no longer appear lying closely together, but seem widely scattered in the cytoplasm. The leucocytes finally lose their power of cytoplasmic division. This happens generally on the margin of the plasma clot where the culture medium is thinly spread. The horseshoe — or kidney-shaped nuclei separate, become


pyknotic and form round, chromatic bodies " (figs. 11 to 19). The acidophil granules become more and .more indistinct, the cytoplasm is again acidophil, and partly vacuolized. In this stage, long chains of these forms closely lying together cover the outer zones of the preparation, gi\ing it a reddish halo. Later these cells without granules flatten out entirely, lose their nuclei or their chromatic particles, and undergo total destruction.

To summarize: most mononuclear and polymorphonuclear eosinoiphil leucocytes with either round, kidney-shaped, or lobulated nuclei, during the first hour of their emigration (fig. 8, and fig. 43) into the surrounding plasma, divide rapidly. They form smaller cells with fewer granules and a more basophil cytoplasm. Later by dividing and moving to the outskirts of the plasma clot, they finally form rays and layers of partly acidophil, vacuolized 'cells' without nuclei and granules. Another group of these eosinophil leucocytes, before diminishing in size in the zone near the implanted bone-marrow particle, had extruded its granules at a very early period. They fade out and leave their more basophil cell bodies in the plasma clot. The mononuclear or polymorphonuclear eosinophil leucocytes undergo a regressive development in tissue cultures.

These conclusions agree with the writer's own observations of the cells in living preparations. On the first and second day of incubation the eosinophil leucocytes are numerous and of normal size (fig. 2, left side, above). On the fourth and the fifth day the few forms, which have not undergone the flattening-out process and which have not changed their character, are small, with fine granules and an ellipsoid nucleus (fig. 5, left side, below). Foot ('13, pp. 49-51), in his account of the changes of the eosinophil leucocyte in the culture medium, reports that these cells finally take on the same form as that assumed later by the large mononuclear lymphocytes, and cannot be distinguished from them. With this conclusion the present MTiter cannot agree. In figure 8, the emigration of small leucocytes is shown. The lean, almost fat-less bone-marrow orginated from a young, not full-grown chicken. After an hour in an identical preparation the tissue was extracted and only the emigi-ated cells were allowed to develop. All cell types which


are pictured in figures 11 to 26 are cells which have emigrated early from the bone-marrow particle, advanced to the border of the plasma medium, and changed in different ways.

Figin-es 11 to 19 show the regressive development of the polymorphonuclear leucocyte which is inserted in the plasma, either as a younger form, with spherical nucleus, or as an older form with kidney — or horseshoe-shaped, or lobulated nucleus always recognizable because of its acidophil granules. The long chains of these deformed cells in all transitions are easy to identify in preparations, where only a few cell types have been allowed to emigrate into the plasma. Here they never take on the character of the 'cell culture type' (Foot).

When bone marrow is taken from a young, poorly fed chicken and treated as above described, few ' mononucleare basophile Zellen' emigrate in the first half hour, and the bulk are only eosinophil leucocytes (fig. 43). If these preparations are allowed to develop two or three days the rays of cells consist for the most part of these eosinophil leucocytes and few X cells or forms of the cell culture type are visible. If the process of extracting and again implanting the bone-marrow particle is repeated and the cells of the succeeding emigrations are controlled, few eosinophil leucocytes are observed in the second and third stage and after the third implantation approximately no eosinophil leucocytes are to be seen.

Therefore, no new formation of this cell type from a stem cell could be observed in the plasma clot, but only a process of emigration, multiplication, transformation and degeneration of those forms which were implanted with the bone marrow in the plasma clot.


The general rule for the behavior of cells in tissue culture: the more they are differentiated, or adapted to certain functions,

^Erdmann, Rh. 1917 Some observations concerning chicken bone marrow in living cultures, Proceedings of the Society for Experimental Biology and Medicine, vol. 14, pp. 109-112.


the quicker they undergo destruction holds true in the case of erythrocytes. The red blood corpuscles appear often without a nucleus or without a shadow of a nucleus. The plasma seems perforated. This indicates that the haemoglobin has disappeared. Those cells in which we can trace only the shadow or a faint remainder of the nucleus are apt to deceive the observer. The remainder of the nucleus appears like a small parasite but is nothing more than the nucleus of the cell, as can be proved by numerous intermediate forms. These bodies resemble the Cabot's bodies which are described by Juspa ('14, p. 429) in certain diseases of men. Also the nuclei may become pyknotic in other forms and the plasma may disappear. Foot ('12, p. 461, and 1913, p. 46) notes these same two different ways of degeneration in erj^^throcytes. Their dead nuclei or their plasma is often incorporated into phagocytic cells (figs. 34 and 35) the origin and types of which will be discussed later.

The non-elongated round or irregularly shaped erythroblasts have a pale yellowish or colorless plasma (figs. 1 to 3). Well developed erythroblasts are distinguished when stained by their wheel-like, highly chromatic nucleus. Unstained cells show a whitish appearance of the nuclear membrane which seems crowded with the content of the nucleus and ready to break. Figures 3 and 7 represent erythrocytes and erythroblasts in various stages of their retrograde development. Their plasma-less nuclei cover the microscopic field and are often seen incorporated into cells of phagocytic character. Unripe, young erythroblasts are figured in figure 8. They have larger nuclei in proportion to their basophil plasma than the erythrocytes and are scattered, through the tearing apart of the bone-marrow network, in large quantities into the surrounding plasma. They are recognizable in stained preparations by the smooth surface of their plasma and their chromatic nuclei and cannot be confused with eine Art von basophilen mononuclearen Zellen" which, according to Foot '12, form the X cells and the cell culture type.

But the difficulty begins when very young, i.e., small cells characterized in the first day of incubation by their situation near the bone-marrow network, are to be isolated and cultures


from young living erythroblasts and from young basophil cells with vesicular nuclei are necessary, for deciding different questions. My experiments only proved, after isolating young cells near to the bone-marrow network that they underwent no transformation into erythroblasts but showed the phenomena fully described later on page 94-100 the transformation into cells of connective tissue cell type. It is naturally not excluded that erythroblasts — when they are already erythroblasts in a strict sense- — divide in the tissue cultures, but I never could isolate this cell type with any certainty just at the point in being transformed from its 'stem cell' into erythroblasts. This phenomenon seems not to take place in tissue cultures.


The microlymphocytes in chicken bone marrow are found in great quantities. Their small protoplasmic brim and condensed, highly chromatic nuclei allow us to distinguish them easily from the small basophile round cells "with vesicular and achromatic nuclei, closely situated to the network of the bone marrow. The microlymphocytes seem to be present in the tissue cultures from the first day of the incubation of the bone marrow, mthout apparent changes, until the last day of cell life in the culture. But are those the same forms which were incubated or newly originated forms? The microlymphocytes implanted with the bone marrow particle must be capable of active movements, because they are no longer visible in the meshes of the bone marrow network after several days' incubation, but are always present in the plasma clot. In the preparations where only a few cells are allowed to emigrate and to stay several days in the plasma medium, the microlymphocytes are widely scattered. Their own cytoplasm expands in a star-like manner, often forming long cytoplasmatic raj^s. After a fortnight in the culture medium, they have the appearance of forms such as the cells pictured in figure 25. One cell appears normal; the other has a torn cytoplasmatic body. Figure 27 shows the remaining nuclei which will soon undergo complete destruction. Foot '13, page 43, believes that besides numerous microlymphocytes.


which die, a large number 'steadily increase in size' and either form cells of the macrolymphocytic type or of the large mononuclear lymphocytic type, after the latter has undergone nuclear enlargement and dechromatization." Foot presents no dra^^^[ngs of these highly important forms, but considers it sufficient to record the measurements of microlymphocytes of different sizes, measuring from 3.5 to 9.6 in diameter. The nuclear structure of these transition forms is not described by him. The present author has never seen cells with typical microlymphocytic condensed nuclei in all sizes, only cells with vesicular achromatic nuclei in every possible size. In the later discussion these contradictory reports of Foot and of the present author must be borne in mind.

Some authors hold the theory that microlymphocytes originated from the large mononuclear lymphocytes by multiple simultaneous divisions. Only in very recently incubated tissue cultures, as recorded on page 79 a breaking of large lymphocytic forms into pieces was observed. But the isolated cultivating of these small cells afforded no definite results. Multinucleated forms with ragged or torn cytoplasmic structure and nuclei with highly condensed chromatin may be observed in the case illustrated, of which three have a condensed chromatic structure (fig. 8). The younger the implanted bone marrow is, the more numerous these forms appear to be. They have a slight resemblance in their plasma to very young connective tissue cells, as, e.g., Maximow ('10) pictures them in figure 43, from a guinea pig, but they seem to have no connection with the formation of bone marrow lymphocytes.

To summarize: The microlymphocyte belongs to those cell types which undergo no progressive development in the tissue culture.


From the first to the sixth day after incubation large cell types can be observed in the tissue culture of bone marrow when the experiment is conducted with a full-grown, over a half year old ehicken. These cell types have, as described on page 79,


before staining and preserving, a half-moon shaped, or elongated nucleus, and their plasma is either granulated, or the granules are invisible dining cell life. The cells shown in figure 2, two granulocytes and one ungranulated large cell, have only been one day in the culture. The first type appears to divide; we can observe smaller forms on the following days, with larger granules than the eosinophil leucocytes possess. The other represented cell type is a large lymphocyte. These forms may break in pieces during observation. After six days incubation we discover in stained preparations the changed form of the myelocytes (figs. 39 to 42). The reddish ripened nucleus of these forms has all the characteristics of a myelocytic nucleus. But in eosinazur stains such nuclei are generally supposed to have a more bluish color. This must be explained by the rising acidity of the culture medium in growing tissue cultures (Rous, '13, p. p. 183-86). The cells in figures 39 and 41 must be considered eosinophil myelocytes, those in figures 40 and 42 mononuclear lymphocytes. In earlier stages of their degeneration process these large forms often have very fine acidophil granules in their cytoplasma when observed on the second or third day of incubation; but they are never seen to divide. Their plasma loses its granulations, flattens out, and vacuolizes. The eosinophil myelocytes and lymphocytes have only a regressive development in the tissue culture medium.



But one observation of the behavior of fat cells in tissue culture is given by Foot, who writes ('12, p. 447,) that the cultivation of subcutaneous or subepicardial adipose tissue was without success, growth of considerable amount could not be observed. The present writer repeated Foot's experiments. Adipose tissue of the omentum of the chicken showed, after three days incubation, almost a complete disintegration; further, the formation of few cells of the 'cell culture type' and the survival of connective tissue cells could be observed. It may be conceived that some connective tissue cells may have originated


from fat cells losing their fatty contents and assuming the character of the known type of connective tissue cells. Or the connective tissue cells, implanted together with the adipose tissue may have developed and multiplied. This is a separate question which has not been sufficiently studied in true adipose tissue.

The changes of the fat cells of bone marrow in tissue culture, though not considered by all authors to be real fat cells, have a great resemblance to phenomena seen in rapidly growing embryonic adipose tissue, as Foot remarks (p. 48, '12). But he himself, neither in 1912 nor in his later publication of 1913, states the ultimate fate of the implanted, so-called fat cells, which, together* mth the other cells of the bone marrow, are in the culture medium and are numerous in the white bone marrow of the adult chicken. The typical signet-ring cell may apparently remain unchanged for 24 hours in the plasma medium, as it is shown on a photograph (fig. 46, right side, above). But the observed facts do not agree in most cases with this view. After three hours incubation all fat cells show still their accustomed shape. The big fat -globule surrounded by a brim of cytoplasm flattens out and the large globule of fat separates into small droplets. Or the fat cell divides into two parts, and even a process of budding may be observed (figs. 29 and 30). If the cell has not divided up, the fat globule diminishes in size and does not fill the whole cell. With a specific fat stain it can be shown that the cytoplasm is filled with small fat droplets and strands (fig. 28). Later foamlike masses of cytoplasm, in the meshes of which the fat is easy to identify, protrude from the cell margin and separate themselves partially or totally from their 'mother cell.' Cells of this kind may offer the appearance of cells figure in figure 2, left side, in unstained preparations. In a tissue culture of 24 hours incubation, preserved with Orth's fluid and stained with Giemsa stain; they appear as cells with highly chromatic nuclei, and perforated cytoplasm (figure 7, right side and figures 33 and 34) ; also weblike masses, apparently without nuclei, are frequent (fig. 7) which are often surrounded by microlymphocytes and polymorphonuclear leucocytes. Text-figure A gives



the most striking phases of the activation of a fat cell. The original fat cell, the fat cell which has extended fine pointed processes, and the final stage that comprehends cells containing vacuoles which may still have traces of fat in them. (Compare cells on figure 2; figure 7, cell right side, above; and figures 45 and 46.)

Text fig. A. Fat cells after 6 and 12 hours incubation.

It must be kept in mind that these changes occur during the first 24 hours or 48 hours of incubation. Figures 45 and 46 show that in a 30 hours culture the dissolving of the big fat globules and the dividing up of the fat cells has been in progress. The cells form chains, typical for the stage of the culture of 24 to 48 hours of fat containing bone marrow. These cell chains flatten out, fine processes are extruded which cover great areas and may fuse with other cells in web-like masses. Figures 45 and 46 give a good surview of this process and such a cell is also represented in figure 33. We note its enormous size, its big vacuoles, its slender processes, its phagocytic capacity and its small nucleus. In short, we see a so-called 'Riesenzelle' of Foot


which is already present after 24 hours of incubation. Now Foot ('12, p. 459, fig. 5) gives the photograph of a preparation of bone marrow after 5 days of incubation in a plasma medium. This is a descrepancy for which no explanation could be found.

It is of importance to state that all vacuoles do not contain fat in such a condition as to make it visible by the osmium process. The cell (fig. 32) shows still some fine traces of fat, but in many preparations which were treated with Scharlach or Sudan stain after adequate fixation, the vacuoles were devoid of fat. It is conceivable that fatty acids or other products of related character fill the vacuoles, but even after trying the most complicated stains (Ciaccio, Benda) to elucidate the nature of the contents in the vacuoles, no final decision could be I'eached.

From the third to the fifth day, the number of 'Riesenzellen' has diminished; we see smaller round or oblong cells with one or several vacuoles, with oblong faintly chromatic nuclei (fig. 34). They are the products of the breaking up of the 'Riesenzellen' and seem to be identical with Foot's cell culture type. They are capable of phagocytosis and move slowly toward the periphery of the plasma clot.

How can we interpret these extraordinary changes in the fat cells? The only similar observation was made by Maximow ('04, p. 108), describing the changes occurring in the cells of inflamed connective tissue of the rat. There he gives a good description of the involution of the fat cells. The process shows the same phenomena in the involution of the fat cells in the connective tissue of the living animal after inflammation as are to be seen in tissue culture. The flattening out of the cytoplasm, the dividing up of the big fat globule into small droplets inside the cell (Maximow, plate 3, fig. 9; Erdmann, text-fig. A) and the transformation of the plasma in a honeycombed mass (Maximow, Plate 3, fig. 11; Erdmann, fig. 7, left side, above), are identical processes in both cases. Maximow believes ('04, p. 119) that some of these cells become fibroblasts. The present author ventures no opinion on the subject, though a striking similarity exists between the fibroblasts of Maximow (text-fig. B) and the cell in figure 7, right side above.



We find after the second day in our cultures: (1) cells of the fibrotjlast type; (2) cells of the 'Riesenzellen' type; (3) cells of the cell culture type, after Foot. All three types can originate from the implanted fat cell.

Besides these progressive changes we must state that many implanted fat cells undergo destruction. This is shown by the observation of the living cells as described on pages 79 to 81. Figure 4, shows such a disintegrating mass of fat cells from an unstained preparation, and fig-ure 7, shows the mass in a stained preparation. Here two cells of the honeycombed type are recogniza

Text fig. B Maximow, 1914, figure 8, plate 3. area of inflammation into a fibroblast.

Involution of a fat cell in an

ble (left side, above), one of which is intact, the other has expelled the contents of the plasma. Microlymphocytes are gathered around the disintegrating fat masses and the transformed fat cells. Maximow describes how his polyb asts, cells of the lymphocyte order, crowd around the fat cells and destroy them by phagocytosis (page 120). The same phenomenon occurs in the tissue culture; between the second and the fifth day the destruction and resorption of the dying fat cells is finished and the tissue culture gradually assumes a different aspect, as will be described later.


But together with these retrograde processes, easily observed in the Uving culture, small parts of the irregularly-shaped, large, disintegi-ating fat cells isolate themselves. They become spherical in shape and begin to wander away from their 'mother cells.' They can be recognized by their small nuclei, their coarse glistening plasma. They are identical with small fat cells. This 'rejuvenation' of the fat cell was only observed when bone marrow tissue of younger well-fed animals was implanted. Bone marrow from very young chickens and tissue from old hens seldom rejuvenate the fat cells, when such are present. In tissue from older hens the disintegration of the fat cells often obscures the observation of the other cell types.


\\'Tien implanted in the plasma medium, the bone-marrow particle itself appears basophil after preservation with Orth's fluid and staining with Giemsa stain. For a long period, up to 14 days, it shows a strong basophilic character. We have shown how fat cells and their derivatives generally have a strongly basophil nucleus and often a basophil plasma. Erythrocytes, erythroblasts, and eosinophil leucocytes, which show a strong basophily of the nucleus, emigrate or are washed out of the tissue particle and either perish or undergo the changes described. The eosinophil leucocytes, diminishing the size of their nuclei and acquiring an acidophil cytoplasm, later form, together with the erythrocytes, the reddish halo around the implanted particle.

After the first emigration or washing out of the cell types mentioned, the tissue particle consists almost solely of basophil cells, which are very young, small, unripe erythroblasts, small lymphocytes, connective tissue cells of the bone marrow network, and basophil cells of all sizes and forms, the character of which is not at first recognizable. The thickness of the tissue particle prevents the closest examination, but these cells have always ungranulated plasma. In figure 8, a general survey of these basophil cells is given, as they appear after one hour's in


cubation in bone marrow of a young nearly fat-less chicken. Two types besides the erythroblasts with their more or less pinkish plasma and their wheel-like nuclei are distinguishable — cells with crude irregular cell plasma, as if it has been torn They possess small, condensed, highly chromatic nuclei (fig. 8 left side, above), or their cytoplasm has well-rounded contours and a very big nearly chromatinless nucleus. This type and its changes will now be described.

In figures 11 to 27, different emigrated cell types of a similar bone-marrow particle are represented. The particle itself was twice extracted during an incubation period of 24 hours. The emigrated cells of each extraction stayed 12 days in the plasma until they were preserved and stained and later analyzed, so no new rear guard of eosinophil leucocytes and those mononuclear basophil cells, the fate of which Foot tried to elucidate, need be considered. According to this experiment, which was repeated several times, besides the eosinophil leucocytes the changes of which (fig. 11 to 19) have been fully treated on page 85, six different cell types are recognizable after the second extraction.

1. Cells which resemble fat cells (figs. 20 and 21).

2. Cells which, by their nuclear structure but not by their cell plasma, resemble true connective tissue cells (figs. 22 to 24).

3. Cells which are true connective tissue cells, from the type of endothelial cells (fig. 27).

4. Cells which are true connective tissue cells not shown in figures 20 to 27 but in figure 9, with star-like, fine protoplasmatic processes and elongated, often cone-like shapes, and a more mesenchymelike character.

5. Cells which are microlymphocytes (fig. 25 and also fig. 27).

6. Cells which are lymphocytes (fig. 26).

Cell types 3, and 6 are not often found in preparations made according to the prescribed method. The lymphocyte with its fine red granules (fig. 26) shows all signs of degeneration. It appears highly probable that in the plasma clot the normal ripening out of the large mononuclear lymphocyte began but could not be fully accomplished owing to the conditions of the culture medium. The endothelial cell and the elongated connective tissue cells


(figs. 27, 9, and 38) have not changed their characters. They abeady appear on the first day after incubation, because they could be observed in bone marrow culture of 24 hours incubation. The elongated connective tissue cell is highly amoeboid, and shows in its plasma, on the first days of incubation, fine and bigger fat droplets, which are coarser when stained with specific fat stains. Later their plasma looks as if pulverized with small fat droplets, still later they lose their fat and appear highly vacuohzed. They repeat on a smaller scale the changes of embryonic subcutaneous connective tissue that had been incubated 14 days in a plasma medium. Because these cells appear after the first day of incubation (the present author has observed them after but five hours' incubation) it appears highly improbable that they originated from the basophil spherical cells in question. They are cells of the bone marrow network or the vessels of the bone marrow, which have been torn apart by the cutting of the bone marrow. They can be also observed in tissue cultures of true adipose tissue and are distinguished by their rapid division rate.

In most cultures of connective tissue made by various authors these cells have been described. Lambert and Hanes ('11) mention the accumulation of fat and the vacuolization of the cytoplasm in cells of mesenchymal origin. They represent tumor cells in their publication of 1911, plate 66, figures 4 and 5, of this character. Lambert himself in 1912, on plate 72, figure 3 and plate 74 figures wandering cells from the chick spleen. Some of these forms are more related to the connective tissue cell type in question, some resemble more the cell type seen in bone marrow cultures when the fat cells have begun the disintegration. In 1914, plate 44, figure 6, he gives a good proof of this.

In figure 9, Carrel and Burrows, ('11), represent also fat storing cells of this type. They are said to be originated from an adult chicken spleen, while the first author must have seen the elongated vacuohzed type ('13, plate 17, figure 16), in cultivated connective tissue. Lewis, R. M., and Lewis, H. W., '11, show on their figure 20, left side, in a chicken liver culture, highly vacuohzed cells of the same type.


This comparison could be continued but the facts prove already that among connective tissue cells of the most varied parts of the chicken body these elongated, finely vacuolized, slender cells appear with a true connective tissue cell nucleus. They are all similar to the figures of Foot representing his X cells (cf. Foot '12, plate 22, figures 8, 16, 19). The connective tissue cell represented by the present writer in figure 9, is taken from a young chicken and is not of the same size as some of those cells which Foot shows. When cells, however, were taken from the bone marrow of a full-grown chicken, they were of the same dimensions as those given by Foot, '12, plate 22, figure 8.

Also, in the development of embryonic bone marrow tissue of the chicken, Dantschakoff, '09, depicts mesenchyme cells (plate 44, figures 5 and 6) which have a close resemblance to the above mentioned cell type (fig. 9). They are identical types, except that the latter may contain fat, the first are fatless. In this group must also be included the elongated forms of Foot's Riesenzellen which have pointed pseudopods.

To summarize: Though fat containing and often vacuolized the elongated cells with connection tissue like nuclear structure which appear in Foot's figures among his 'Riesenzellen' are true connective tissue cells. There can be no doubt that the granular lymphocytes, the elongated cells of connective tissue character, and the endothelial cells did not originate de novo in the tissue culture.

In studying the cells close to the connective tissue network of the bone marrow the present wTiter could only distinguish one well defined cell type (figs. 36 and 37). Small round cells with strongly basophil cytoplasm and large, faintly staining nucleus with two nucleoli are abundant. They are neither microlymphocytes nor mononuclear lymphocytes nor erythroblasts. They differ from the microlymphocytes by their vesicular nuclei, from the mononuclear lymphocytes by their size and their cytoplasm, from the erythroblasts by their nearly chromatinless nuclei and also by their size. In living cells the nuclei of erythroblasts appear whitish, the nuclei of these cells dark. If these cells, which migrate from the tissue particle after the leucocytes are


washed out by continued changing of the plasma, on the second incubation are allowed to develop we find after a fortnight two different types: figures 20 and 21, and figures 22 to 24. The cell represented in figure 21 differs from the basophil cells which had been implanted into the tissue culture (fig. 8, and figs. 36 and 37) onlj^ by its size and by the more chromatic contents of its nucleus. These forms are numerous; they later contain fat or vacuolize, forming chains, the cells of which are always to be distinguished by their nuclear structure from the eosinophil leucocyte. The nucleus has a close resemblance to that in fat cells; it is vesicular with round bulky, chromatic contents.

The next group (figs. 22 to 24) have a true connective tissue cell-like nuclear structure. The nuclei are elongated and fine threads of chromatin form a true connective tissue nucleus network. The cytoplasm is basophil in most cases, but in certain parts of the culture and in very old cultures it becomes acidophil. The basophily or acidophily of cells is no constant character in tissue cultures. Rous ('13, page 183) points out the changes in acidity of growing cells. The cells themselves become acid in the culture medium, after having been basophil. Later they may regain their basophil character. The cells in question are true phagocytes (fig. 23). They contain fat, blood corpuscles, dead nuclei, and other disintegrating particles. They are sometimes polynuclear; as the cell body does not divide they form also the so-called 'Riesenzellen' of Foot. They are more agile after the first days' of incubation. In older cultures they assume round, spherical and oblong shapes, and their enormous protoplasmatic body divides up. They then form the cell culture type (fig. 6) the nuclei of which are always vesicular and not very chromatic.

Therefore, in the group of Foot's 'Riesenzellen' do belong besides the products of the involution of the fat cells and the implanted elongated connective tissue celltype with its finely vacuolized plasma, these forms (figs. 22 to 24) in which the nearly fat-less bone marrow of a young chicken was used. This gave conclusive proof that the small mononuclear basophil cell



(figs. 8, 35, 36 and 37) after leaving the bone marrow network, can form 'Riesenzellen' which by their nuclear structure resemble connective tissue cells. They later become the cells to which Foot gave the name cells of the cell culture type." They are enlarged, fat-storing or vacuolized cells capable of phagocytosis.

The results here presented, i.e., the change of the small vesicular basophil cell into true phagocytes and later into 'Riesenzellen' or cells of the cell culture type — were attained by using the bonemarrow of a young, fat-less chicken and the washing out of the undesired cell types, as polymorphonuclear leucocytes. But even if we use the fatty bone-marrow of a full-grown chicken and control the daily changes, the same fact is demonstrated. The first day after incubation (fig. 7) we observe a large number of basophil mononuclear lymphocytes. Three are shown in one microscopic field. Their pale nuclei, often of a lighter blue than the plasma, the irregular shape of their plasmatic body in which sometimes a few fine -acidophil granules are visible, and their large size, make them conspicuous. Examining preparations of the same series a day later, the lymphocytes are very scarce. On the fifth day of incubation, when the disintegrated fat has been disposed of by the phagocytic acti\dty of these basophil cells, characterized by their close position to the network of the bone marrow, they are by far the most numerous types in our tissue cultures. In the following days they grow and divide rapidly forming 'Riesenzellen' which can store fat, become vacuolized, and end in rounding off and becoming cells of the cell culture type, their nuclei with a fine thread-work of chromatin becoming more like true connective tissue nuclei. They can even lose their basophily but may always be distinguished by their nuclear structure from the products of the regressive development of the eosinophil leucocyte in tissue cultures.

It might be possible to interpret Foot's text-figure 5, (page 459, '12) as representing a tissue culture preparation just in such a stage; because the time for formation of these features is the same. But then it is not explained why Foot does not describe the formation of the 'giant cells' and cells of 'cell culture' type after 24 hours' incubation.


In the above mentioned preparations the bulk of all cells, with their fat storing and phagocytic capacities, their vacuolized cytoplasm have now left the implanted bone marrow particle. They advance with their fine, pointed, plasmatic pseudopodia to the outskirts of the plasma clot. Their faintly chromatic nucleus has only two nucleoli. This character is evident in the youngest cells of that kind w^hich are close to the network of the bone marro^v (figs. 36 and 37) and is also found in 'Wanderzellen' after Dantschakoff (cf. Dantschakoff, '09, page 133), plate 7, figures 2 to 5. These 'Wanderzellen' w^hich originate from a mesenchyme or endothelial cell can, according to Dantschakoff, either be histiotypic or lymphocytic. They form in the embryonal development specific elements of the connective tissue or the hematopoetic apparatus, according to the conception of the monophyletic school. In older cultures nearly all basophil cells have nuclei of true connective tissue cell character, e.i., the chromatic granules of the nucleus are connected with fine threads. They are identical with those nuclei figured in figures 22 to 24. Not so frequent are types of nuclei figured in figures 20 and 21.

The 'Wanderzellen' in the tissue culture lose, in the later days of their existence, especially in unrenew^ed tissue cultures, their fine cytoplasmatic processes but are — by the structure of their nuclei and their cytoplasm — connective tissue cells of a more mesenchymelike character. They are transformed to cells of the cell culture type.

That these cells are descendents of the implanted cells, which were lying close to the bone marrow, is further proved by the following experiment. After all loose cells in the meshes of the bone marrow are w^ashed out by repeated changing of the plasma medium, cells of the type in figures 20 to 24, can be formed. After three changes of the culture medium, with a period of two days between, the cells close to the netw^ork formed vacuolized cells w^hich could be interpreted in no other way except as 'Wanderzellen.' Their nuclei had become nearly chromatinless, and their plasma acidophil; they sometimes assumed the character of fat cells, but were generally of the 'Wanderzellen' type.


No large mononuclear lymphocytes could be seen. It is, therefore, also evident that a new formation of this cell type, the niononuchar hir^e lymphocyte of the bone marrow, does not occur in i\\v tissue culture. The smaller and larger basophil cells with a vesicular nucleus near the bone marrow network, and the cells which later leave the network are 'Wanderzellen,' a type closely related to the mesenchymal cell. They can be kept alive for longer periods in renewed culture-medium.

The empty network of the bone marrow, consisting of slender connective tissue cells, has lost its power of sending new cells into the surounding plasma clot. The network cells remain living for long periods in renewed medium changing only their cytoplasma in the same manner as other connective tissue cells do in plasma culture. It becomes perforated with sieve-hke vacuoles which may store fat.


The growth of chicken bone marrow in chicken plasma may be divided into two distinct periods. The first period has a more regressive character. As process of this first period may be enumerated: — the degeneration of the erythrocytes and the nearly full-grown erythoblasts, the ripening of the granulocytes implanted with the bone marrow into the tissue culture; and the decay of the latter.

The eosinophil mononuclear or polymorphonuclear leucocytes after rapid multiplication lose their granules, are flattened out, and form cell chains of acidophil character which undergo slow destruction.

The myelocytes moving at first amoeboid-like in the plasma clot, and behaving like phagocytes, seldom divide, but ripen out until they assume a large size. Then their plasma vacuolizes and disappears, leaving only the nuclei.

The microlymphocytes show no signs of multiplying. They leave the meshes of the bone marrow particle; later lose their cytoplasm; and finally leave their condensed nuclei in the culture.


The large mononuclear lymphocytes of the type occurring in the flowing blood, present in great numbers after the first day of incubation, form now and then fine granules, but undergo no further development into myelocytes. They lose their nuclear chromatin, and their plasma becomes honeycombed and finely vacuolized, and they finally leave as the only trace of their existence faint shadows in the plasma clot.

The so-called fat cells of the bone marrow flatten out; the big fat globules divide into smaller droplets; their plasma either vacuolizes and forms long needle-like projections, or fibroblastlike cells with a central nucleus and honeycombed plasma. The first cell type is phagocytic. These cells represent 'Riesenzellen' in the first period of the tissue culture growth. Not all cells of this type are transformed into fibroblasts or 'Riesenzellen.' Some fat cells disintegrate filling the culture medium with degenerating fat particles. Now and then the nucleus, with a small amount of cytoplasm separates from the dying 'fat cell' and a young 'rejuvenated' cell of fat cell character appears. The so-called fat cells combine the first regressive period of bone marrow growth with the second of more progressive character. Some undergo destruction, some survive, later assuming Foot's cell culture type.

From the first day of incubation, connective tissue cells of elongated shape with very fine pointed projections migrate into the plasma clot. They store fine droplets of fat and partially vacuolize. They are also found in the second period of growth in the tissue culture.

The second period begins ^\dth the loosening up of the cells around the network of the bone marrow; the smaller, or larger basophil cells, with vesicular nucleus migrate into the surrounding plasma and the network sends new cells into the plasma clot till it is utterly devoid of cell forms. These cells represent an intermediate type between the 'histiotype Wanderzellen' (Dantschakoff, '09) and the embryonic mesenchyme cell. They do not resemble in all details the large nononuclear lymphocyte of the blood. They move into the surrounding plama, send out penetrating needle- and bristle-like projections; divide into phago


cytes; store fat; lose their projections and partially vacuolize, assuming the form of the "cell culture type."

The network of the bone marrow, having lost its cells, and no longer able to send out emigrating cells, consists of slender connective tissue cells. These show a remarkable paucity of chromatin, are strongly acidophil, and possess sieve-like vacuoles of the finest type.

The 'Riesenzellen' of Foot comprehend several cell types:

1. Transformed fat cells and elongated, vacuolized connective tissue cells.

2. Newly emigrated basophil cells of the bone marrow network, which are related to the histiotype Wanderzelle" of Dantschakoff.

3. Some few myelocytes and flattened out eosinophil monoor poly-morphonuclear leucocytes.

These two phenomena, the dying of the cell forms which are not adapted to the continued growth in tissue culture, and the adapting of a new character by those cells which are capable of living longer periods in the plasma medium, often overlaps. They appear more sharply separated in cultures of almost fatless bone marrow, where few 'Riesenzellen' appear in the first days of incubation. From the third to the fifth day, when the loosening of the bone marrow network and its content has begun, they become numerous. The duration of these periods may be stated as follows : The first period lasts from the first to the third day; the second period from the third day to the death of the culture. The surviving cells of the cell culture type (Foot) are modified fat cells and newly formed wandering cells of the mesenchymelike type. After fourteen days' cultivation, they are, except the elongated connective tissue cells the only living cells. They belong to the connective tissue cell type and may, when the medium is renewed, grow indefinitely.


As one of the first results of our analytic study, let us discuss the fact that the so-called X or 'Riesenzellen' of Foot represent several different cell types. The myelocytes and larger eosinophil leucocytes acquire, as shown, good dimensions in the tissue


culture of bone marrow. The myelocytes, capable of amoeboid moving, form few 'Riesenzellen.' They can easily be omitted in the following discussion, as they are always distinguished by their characteristic nuclei and the blunt form of their projections, when stained and preserved. They are just as unmistakable when living. The large mononuclear or polymorphonuclear eosinophil leucocytes only need be considered, as X or 'Riesenzellen' when they have flattened out and formed rays of cells. Then they are surrounded by the projections of the transformed fat cells or cell types of the 'histiotype Wanderzellen' order. Both cell types are true phagocytes, thus forming, chiefly in the first days after incubation, cell masses of X or 'Riesenzellen' of combined characters. The whole combination may even seem, judged only by its acidophil staining, to be from a different origin. But the daily observations reveal the facts of their development. It is questionable if any necessity exists for giving new names, as Foot did in 1912 and 1913, for the X cells 'Riesenzellen' and later forms. They are either transformed fat cells, or mesenchymelike wandering cells which have left their customary place and which assume in later life in tissue culture the characters of connective tissue.

The name 'Riesenzellen' or true giant cells has already been used for cells of the type represented in figure 10. This multinuclear cell was seen in a tissue culture of bone marrow from a two-months old chicken, and resembles in every particular the true giant cells figured and described by many authors.

To call the questioned basophil cells 'X cells' when their origin is known would be a contradiction. They are either 'fat cells' or mesenchymelike cells, and both types are transformed from their original type by our cultivation method. The present author would propose calling the latter simply wandering mesenchymelike cells, and the fat cells, transformed fat cells. Their close relationship to the mesenchymal cell type is again proved by their physiological behavior in tissue culture, so closely identical with that of the wandering mesenchymal type. It even became evident that some 'fat cells' may assume the character of fibroblasts when they are not transformed into


highly vacuoHzed or fat-storing cells of mesenchymal character with projections at first needle-like and later of a rounded or elongated shape. This twofold manner of development of the bone marrow fat cells is important, as it might probably be the result of a non-uniform origin.

In judging the transformations of cell types of mesenchymal origin in tissue culture, we already have established certain facts as a basis of comparison. The mesenchymal cells always grow more rapidly than any other known tissue; they have the ability to store fat; they can vacuolize and can emigrate out of the tissue clot. They can endure this highly artificial method of breeding indefinitely. The bone marrow particle, with its loose meshes, exhibits many 'Wundflachen' which are incited into new growth by the stimulus given by the cutting of the tissue. By repeatedly renewing the culture medium and transplantin ; the tissue particle, w^e stimulate the growth again and again, until we have exhausted the power of the network to send newly formed mesenchymelike cells into the plasma, and only a fine thread-like network with a few oblong, small nuclei remains. The pliability of the mesenchymal cell and its ability to undergo transformations is known in embryonic life and is here demonstrated in tissue culture life.

Two subjects of importance have not been touched. Can these w^andering mesenchymal cells form fibrils, and have they any relation to the formation of the different elements of the bone marrow? Throughout the whole description of the cell transformations in tissue culture, the writer has avoided Foot's conclusion of 1912, namely, since his X cells form fibrils, they must be of the mesenchymal type. The tissue particle of bone marrow has a fibril-forming connective tissue of its own. When, now and then, fibril-forming cells have been seen (as has been the experience of the \\Titer), they may either originate from cells already implanted in the tissue culture, with the bonemarrow particle, or the imbedded fibrils (Foot '12, plate 22, fig. 15) may represent fibrils or fibers formed by the fibrin-containing plasma of the culture medium (Baitsell '14, '15). Foot maintains that his X cells form fibrils but he does not prove it.


Proof could only be obtained by cultivating isolated cells of a certain known type in a medium which does not contain fibrin as the plasma does. This has never been done and still remains a subject for future investigation.

The author agrees with Foot's view of 1912, that X cells, or the conspicuous cells in tissue cultures of bone marrow, are of mesenchymal type, not because they contain fibrils, but because their origin could be traced and their cytological changes could be recorded. Foot's statement of 1913, must be refuted: that the transformation passed through the stages of small microlymphocyte, macrolymphocytes, large mononuclear forms, myelocytes, polymorphonuclear eosinophil leucocytes, X cells, cell culture type, omitting one or the other forms of this stage, so that directly a lymphocytic origin is considered. It was never observed that true microlymphocytes were transformed into macrolymphocytes in the tissue culture. The basophil cell with vesicular nucleus, pale cytoplasm of various sizes in the network of bone marrow, assumed the cell culture type, after wandering into the cytoplasm, forming point-like projections, displaying the capability of phagocytosis, storing fat, and being vacuolized. There was no stage observed in this transformation which resembled the large mononuclear lymphocyte or the 'lymphocytoid Wanderzelle' of Dantschakoff, though this type could be easily observed in chicken bone marrow when the bird had cyanolophia. The close resemblance with Dantschakoff's 'histiotype Wanderzelle' — cells which form ('09, page 177), after some changes, the 'ruhenden' wandering cells of the connective tissue — could, only be discovered when the basophil forms left the net-work and began to emigrate.

It appears highly plausible that in tissue culture the indifferent mesenchymelike cell in the bone marrow network does not show its supposed duality, either to form the known elements of the connective tissue or according to the views of the monophyletic school, the different elements of the hemato- and granulopoesis. In a medium, where circulation has ceased, where no oxygen renovation takes place, the potency to form the lymphocytic elements of bone marrow may not be strong enough to over


come the potency to form fat cells, fibroblast and 'histiotype' wandering cells. Therefore the present series of experiments does not prove anything concerning the views of the mono- or duophyletic schools, of the formation of blood and lymph in the bone marrow. Here renewed experiments should be made, the different cell types after emigration should be isolated and submitted to conditions reproducing either the condition of the blood or of the lymph. Only with still more refined methods would it seem possible to elucidate, outside the body, the complicated process of blood and lymph formation.

But this series of experiments proves that the latent quaUties of the basophil mononuclear cells in the meshes of the bone marrow can arise de novo in the adult animal, because their wandering phagocytic, fat-storing character has been made evident. This fact ought to be considered in dealing with the appearance of these, and related cell types in the blood and lymph during diseases.


Baitsell, G. a. 1915 The origin and structure of fibrous tissue which appears in living cultures of adult frog tissues. Jour, of Exper. Medicine, vol. 21, p. 455^479.

1916 The origin and structure of fibrous tissue formed in wound healing. Jour, of Exper. Medicine, vol. 23, p. 739-756.

Bell, E. T. 1909 On the histogeneses of the adipose tissue of the ox. Am. Jour. Anat., vol. 9, p. 412-438.

Burrows, M. T. 1910 The cultivation of tissues of the chick embryo outside the body. Jour, of the Amer. Med. Assoc, vol. 55, p. 2057. 1911 The growth of tissues of the chick embryo outside the animal body, with special reference to the nervous system. Jour. Exp. Zool., vol. 10, p. 63-83.

Carrel, A. 1913 Contributions to the study of the mechanism of the growth of connective tissue. Jour, of Exp. Med., vol. 18, p. 287-299.

Carrel, A. and Burrows, M. T. 1913 Cultivation of tissues in vitro and its technique. Jour, of Exper. Med., vol. 13, p. 387-397.

Dantschakoff, W. 1909 Untersuchungen iiber die Entwicklung von Blut und Bindegewebe bei Vogeln. Das lockere Bindegewebe des Hiihnchens im fotalen Leben. Arch. f. mik. Anat, Bd. 73. p. 117-182. 1909 Uber die Entwickelung des Knochenmarks bei den Vogeln und dessen Veranderungen bei Blutentziehungen und Ernjihrungsstorungen. Arch. f. mikr. Anat., Bd. 74, p. 855-929.



Emmel, V. E. 1914 Concerning certain cytological characteristics of the erythroblasts in the pig embryo and the origin of non nucleated erythrocytes by a process of protoplasmic condition. Am. Jour. Anat., vol. 16, p. 127.

Foot, N. C. 1912 Uber das Wachstum von Knochenmark in vitro. Experimenteller Beitrag zur Entstehung des Fettgewebes. Beitr. z. path. Anat. z. allg. Pathologie Bd. 53, p. 446-477.

1913 The Growth of chicken bone marrow in vitro and its bearing on hematogenesis in adult life, Jour, of Exper. Med., vol. 17, pp. 43-60.

Haerison, R. G. 1907 Observation on the living developing nerve fibre.

Proc. Soc. Exp. Biol, and Med., vol. 4, p. 140^6.

1910 The outgrowth of the nerve fiber as a mode of protoplasmic

movement. Jour. Exp. Zool., vol. 9, p. 787-848. JusPA, V. 1913-1914 Uber den Entstehungsmechanismus der Cabotschen

Korper und ihre diagnostische Bedeutung bei den experimentellen

Anamien oder den schweren Anamien des Menschen. Folia Haemat ologica, Bd. 17, p. 429-441. Lambert, R. A. 1912 The production of foreign body giant cells in vitro.

Jour, of Exper. Med., vol. 15, p. 510-516.

1914 The effect of dilution of plasma medium on the growth and fat accumulation of cells in tissue cultures. Jour, of Exper. Med., vol. 19, p. 398-405.

Lambert, R. A. and Hanes, F. M. 1911 Growth of sarcoma and carcinoma cultivated in vitro. Jour, of Exper. Med., vol. 13, p. 495-504.

Lewis Reed, M. and Lewis, H. W. 1911 Cultivation of chick embryo tissues. Anat. Rec, vol. 5, p. 277-288.

Maximow, a. 1904 tjber entzlindliche Bindegewebsneubildung bei der , weissen Ratte und die dabei auftretenden Veranderungen der Mastzellen und der Fettzellen. Beitr. zur path. Anat. und zur. allg. Pathologie. Bd. 35, p. 93-127.

1909 Untersuchungen fiber Blut und Bindegewebe. Die friihesten Entwickelungsstadien in Blut und Bindegewebszellen beim Saugetier embryo, bis zum Anfang der Blutbildung in dem Leben. Arch. f. mikr. Anat. Bd. 73, p. 380-444.

1910-1911 Untersuchungen fiber Blut und Bindegewebe. III. Die embryonale Histiogenese des Knochenmarkes der Saugetiere. Arch, f. mikr. Anat., Bd. 76, p. 1-114.

Oppel, A. 1912 Causal morphologische Zellenstudien. V. Mitteilung. Die active Epithelbewegung ein Factor beim Gestaltungs und Erhaltungs geschehen. Arch. f. Entw., Bd. 35 p. 371-456.

1912 tJber die Kultur von Saugetiergeweben ausserhalb des Organismus. Anat. Anz., Bd. p. 464-468.

1912 Kausal-morphologische Zellenstudien. IV. Mitteilung. Die Explantation von Saugetiergeweben ein der Regulation von seiten des Organismus nicht unterworfenes Gestaltungsgeschehen. Arch. f. Entw., Bd. 34, S. 132-167.

Rous, P. 1913 Growth of tissue in acid media. Jour, of Exper. Med., vol. 18, p. 183-186.


Walton, A. J. 1911 Variation in the growth of adult mammalian tissue in autogenous and homogenous plasma. Proc. R. S. L., B., vol. 87, p. 452-61.

1914 The effect of various tissue extracts upon the growth of adult mammalia cells in vitro. Journ. of Exper. Med., vol. 20. p. 554-573.


The drawings were made from total preparations, with Abbe camera lucida, Zeiss homogeneous immersion 2 mm. and compensating ocular 12, with drawing board level with stage of microscope. Magnification about 1500 diameters.



The bone marrow used for the preparations shown in figures 1 to 6 was taken from a well fed, full grown chicken containing a large amount of fat. It was incubated at a temperature of 38°C. in the chicken plasma medium.

1 The first cells emigrating from the particle into the plasma. Bone marrow one hour in plasma, January 3, 1916, 10 a.m. to 11 a.m. Two mononuclear eosinophil leucocytes, one lymphocyte, and one normoblast are visible.

2 Cells which have left the implanted bone marrow particle after twentyfour hours and emigrated into the plasma. January 3 to January 4, 1916. Mononuclear and polynuclear eosinophil leucocytes with rod-shaped granules and large granulocytes with rounded, highly refractile granules are visible. Two fat cells at the right and left side of the preparation have divided up their big fat globule into small fat droplets (compare plate 6). In the middle a large nongranular lymphocyte is to be seen.

3 Cells which have left the implanted bone marrow particle and have advanced to the border of the plasma clot after forty-eight hours' incubation. January 3 to January 6, 1916. One large 'Riesenzelle' and a small granulocyte with highly refractile granules are visible together with one small lymphocyte with vesicular nucleus. Red blood corpuscles with or without nuclei are present. One red blood corpuscle extrudes its nucleus.

6 Cells which have stayed two hundred and sixteen hours in the plasma medium December 25, 1915 to January 3, 1916. Cell culture types.



■"v •>■>>>

j; ■rf

^i- ■»


•■'4 -■» i- . ■* ) * >







4 Cells near the implanted tissue particle after seventy-two hours' incubation. January 3 to January 6, 1916. Extrusion of fat droplets and breaking up of the 'Riesenzellen.'

5 Cells on the outskirts of the surrounding plasma after ninety-six hours' incubation. Disintegration of fat cells. Note the verj' small leucocyte.









V^--. .^'T*f»








7 Total preparation: Bone marrow of a full-grown, well-fed chicken after thirty hours' incubation at 38°C. in the plasma medium. January 3 to January 4, 1916. Orth's fluid, Giemsa stain. (Compare for explanation pages 89 and 98-100.) Actual field represented.





. » • Ji' '








-"- Tiftf*^^^*^




N Total preparation: Bone marrow of a chicken not yet full grown, with a small amomit of fat, after ninety minutes' incubation at 38°C. in the plasma medium. June 7, 1916. Orth's fluid, stain. Small eosinophil leucocytes and many basophil cells with vesicular nuclei are present.

9 Total preparation: Bone marrow of a full grown, well fed chicken, after twenty-four hours' incubation at 38°C. in the plasma medium. December 14 to December 15. Orth's fluid, hematoxylin, eosin stain. The slender vacuolized cell with its nucleus of connective tissue cell structure is already visible after this short incubation period.










10 Giant cell from the Ijone marrow of a younji, but full grown, well fed chicken, after one day's incubation; to represent the type which is generally named giant cell and is not identical with Foot's 'Riesenzelle.'

11 to 19 White bone marrow of a young, nearly fatless chicken in tissue culture at 38°C. After one hour's incubation the tissue particle was extracted and the emigrated cells were allowed to develop further. February 11 to February 25, 1916. A detailed description of the changes of the eosinophil leucocytes is given on page 82-84.

20 to 27 The same bone marrow particle after having been freed from its eosinophil leucocytes by the above described process was implanted for one day again in a plasma medium and extracted again. The emigrated cells were allowed to develop from February 12 to February 25, 1916. Figures 20 and 21 represent a cell type more related to fat cells, figures 22 to 24 a type more related to connective tissue cells, figures 25 to 27 show known cell types which have not changed their character in the tissue culture. Note figure 24: a so-called form of the cell culture type. All cells on plate 5 are conserved in Orth's fluid and stained with Giemsa stain.









13 • *V. •



















28 to 32 Involution of the so-called fat cells of the bone marrow to ' Riesenzellen' in the plasma medium. White bone-marrow of a younger well-fed, fullgrown chicken in tissue culture from November 30 to December 1, 1915. Conservation: Formol. Osmium, Safranin stain.






2 9



A /





38 and 34 White bone marrow from a younger full-grown, well-fed chicken in tissue culture at 38°C. from February 29 to March 1, 1916. Foot's 'Riesenzellen' already present after one day's incubation.

35, 36 and 37 An identical piece of bone-marrow, as described above, was extracted after three hours and transplanted in a new culture medium. The next morning, again extracted and transferred in a new medium. After eight hours the preparation was conserved and the cells nearest to the network studied. A cell, 35, of this preparation having migrated from the network, showing phagocytosis.

38 A cell of the network which begins to become disconnected.

39 to 42 White bone marrow from a full-grown chicken in tissue culture from February 22 to March 2, 1916, at 38°C. Regressive changes of cells of the large mononuclear type or the myeolocytic type.

The cells figured on figures 33 to 38 were conserved with Orth's fluid, those on figures 38 to 42 with Schaudinn Sublimat Alcohol and stained with Giemsa stain.






■ '"^











43 Emigrated cells after three hours incubation (mono- and poly-nuclear "eosinophil leucocytes.

44 Emigrated cells after 24 hours incubation (mononuclear basophil cells) after the culture medium has been once changed after three hours.

Compare page 93-100 for detailed description.








45 Involution of the fat cells after 24 hours incubation near to the implanted bone marrow particle.

46 Involution of the fat cells after 24 hours incubation. Cells near the periphery of the plasma clot. Compare figures 33 and 34, plate 7, from the same series of experiments.














Department of Anatomij and Biology, Marquette Ufiiversity Medical School



Introduction 127

Literature 129

Material and methods 131

Comparative anatomy of the Ammocoete pharynx 13?

Organogenesis of the pharynx in the Ammocoetes 137

Histogenesis 1'13

Early development of the. pharyngeal wall in Petromyzon marinus uni color, Lampetra wilderi 1'13

Epipharyngeal placodes 146

Development of lymphocytes 152

Histogenetic comparisons between cells arising from placodes and ■

lymphocytes 154

Discussion 155

Summary 157

Bibliography 158


The question of the source and mode of development of the component structures of the thymus is one of the most difficult problems in anatomy. Although a vast literature has accumulated during the last fifty years some of the most fundamental phases of the problem still remain unsolved.

The thj^mus has been described in every group of vertebrate animals from the elasmobranchs up to and including man. The involvement of gill pouches in the formation of the gland has been established in almost all of the investigated forms. There

1 This thesis has been accepted by the Graduate School of New York University, in partial fulfillment of the requirements for the degree of Doctor of Science.



is still (lisagreenunit on the purely epithelial origin of the reticulum and ITassal's ('()r])usclos, and the cardinal ])rohleni, the source and nature of the de\'el()]nnent of the small thymus cells, has not been definitely settled up to the present time. Three fundamentally different views, each with its coterie of supporters, are held rej2;ar(lino; the source of these cells. A large number of investigators lielieve that the small thymus cells are true lymphocytes which are formed from the epithelium by a process of transformation. An equally large number believe that these cells represent true lymphocytes which have wandered into the epithelial anlage from the mesoderm. A remaining smaller group of investigators believe that the small thymus cells have an epithelial origin and are different from true lymphocytes.

Even in the most primitive animals in which the thymus has been established, the elasmobranchs, the formation of the small thymus cells does not occur until mesodermal tissue is present in the epithelial anlage. The source of a new type of cell which forms in a mixed tissue would not l)e difficult to determine if the two tissues entering into the formation had different morphological characters and retained them. The methods employed up to the present time have not shown sufficient morphological differences in the mesodermal and endodermal cells present in the thymus anlage to establish the source of the small thymus cells.

The sudden appearance of the thymus as a well defined structure in the elasmobranchs, together with the probability of finding a solution to the question of the source of the small thymus cells, has stimulated a number of investigators to search for a homologous structure in the more primitive types of chordate animals. While the search has been a fruitless one in the ascidians and amphioxus, various structures have been described for a thymus in the cyclostomes. The evidence offered in these descriptions has not been sufficient to establish the organ in this group of animals. The failure to find the thymus or its homologue in the cyclostomes especially in the Petromyzontes, may be attributed largely to the peculiar nature of the branchial region in this primitive group of animals.


The bearing on the interpretation of thymus histogenesis in higher animals suggested by the development of the organ in a primitive type, led the author to undertake a systematic study of the branchial region of, the petromyzon larva. The time and work which have been given to this study have, I believe, been amply repaid in the results obtained. Thymus-like placodes have not only been located in the position which makes them homologous with the thymus placodes of the elasmobranchs, but the placodes have also been found in a more primitive condition than they have been shown to exist in any other animal.


The search for a thymus in the most primitive chordate animals has been undertaken by a number of investigators Up to the present time the organ has not been established in any of these lower forms Willey ('94) suggests that the tongue-bars occurring in the gill-slits of amphioxus represents the thymus gland. The position of these structures is apparently the only basis for this suggestion Their gelatinous structure, however, would offset any argument that they were homologous with the thymus placodes of fishes. Stannius ('84) credits the discovery of the thymus in the myxinoids to Johannes Miiller. Later investigators, however, have been unable to verify this discovery Stockard ('07) in his study of the thyreoid in Bdellostoma Stouti was unable to find a thymus in this form M. Schultze ('56) described a tortuous sac in the ventral wall of the branchial cavity of Petromyzon planeri which he thought represented a thymus. Schneider ('79) showed that a part of this structure disappears in the development of the animal while the remaining part changes into a group of follicles which represent the thyreoid.

Schaffer ('94) described structures in the lateral branchial wall of a 51 mm. larva of Petromyzon planeri, which he thought represented thymus anlagen. He found in all twenty-eight anlagen, seven pairs on each side which consisted of ventral and dorsal portions. These anlagen were connected mth the epithelium of the branchial vestibules.



The minute structure of these buds is summed up in the following :

Was (Ion foineren Ban dieser Knospen anlangt, so gostattet mir die mangdhaftc histologische Conservierung (Alkohol) vorlaufig nur folgendc BernMTkungen zu machen: Von der kapsel dringen feine Bindogcwcbsbiilkchcn in das Innere ein, welche ein spiirliches, reticulumartiges Stutzgeriist fiir die zelligen Elemente biklen. Diese selbst sind kloine Rundzellen von lymphoiden Aussehen, kaum wahrnehmbarem Protopkisma, stark farbbarem Kern und Kerngeriist. Zwischen densolben find en sich ausserdem grossere., blasse Kerne mit deutlicher Kernmombran, und Kernkorperchen, welche dem Stiitzgewebe anzugehoren scheinen und rothe Blutkorperchen, von denen ich nicht sagen kann, ob sie frei zwischen den lymphoiden Zellen liegen oder eingeschlossen in Capillaren.

Die tymphoiden Zellen sind wahrscheinlich Abkommlinge des Kiemenepithels, wie ihr director tjbergang in das letztere vermuten lasst.

In a subsequent paper Schaffer ('06) withdrew his previous interpretation and said that he did not believe these structures represent thymus anlagen.

Giacomini ('00, 1 and 2) believed that 'Hhe lymphoid organ in the basalar region of the gill lamellae (in ammocoetes) might fulfill an analogous function to the thymus in the-fishes."

Castellaneta ('13) describes the structures which Schaffer found, but insists that these structures correspond to lymphoid organs in general and suggests the name 'lymphoid formations' for them.

He further calls attention to the fact that on the one side these lymphoid formations are in contact wdth the peribranchial vessels and on the other with the epithelium of the branchial sac. He does not consider these structures as thymus anlagen insofar that there is not a reciprocal penetration of epithelial and lymphatic elements which should occur in a thymus. Castellaneta calls attention to the abundance of lymphoid cells in the general branchial region. He suggests that these special lymphoid accumulations of Schaffer may represent a primitive condition of the thymus in which the epithelium participates only to the extent of attracting the lymphoid cells.

These lymphocyte accumulations do not occur in the part of the branchial cavity which would make them homologous with


the thymus placodes in the fishes. While lymphocytes are present in these situations the evidence brought out in connection with them is not sufficient to establish their origin or the cause of their presence in these places.

A contribution on the ganoid thymus (Lepisosteus, Amia) has been submitted by Ankarsvard and Hammar ('13). They found the organ a purely endodermal, unsegmented structure in a medial position in the dorso-caudal part of the epibranchial region. It nevertheless has a segmented origin and the epibranchial unsegmented thymus structure represents a secondary alteration from the branchial segments." In the older developmental stages there is a rich infiltration of lymphocytes into the sub-thymic and perivascular connective tissue which stands out in striking contrast to the conditions in an earlier stage. The authors discuss the question whether this condition represents an immigration into the thymus or an emigration from it They believe that the cells have migrated from the placode and represent the beginning of an accidental involution. A lobulization of the organ does not occur. In the adult Lepisosteus the thymus is strongly involuted.

The nature of the thymus in the ganoids as described by Ankarsvard and Hammar is so suggestive of the conditions I have found in the ammocoetes that it appears to me to represent but a very small advance beyond a primitive form in the phylogenetic development of the organ. The origin of the lymphocytes which were supposed to originally migrate into the epithelial anlage in the Lepisosteus does not appear to have been especially determined by the authors.


The material which is the basis for this work was generously supplied to me by Prof. Simon H. Gage. It consists of a series of specimens ranging from the segmentation sphere up to and including a transforming larva and the adult These specimens undoubtedly represent two species, Petromyzon marinus unicolor and Lampetra wilderi (the lake and brook lamprey of central New York).


A number of the 5 mm, specimens in my collection were kindly given to me by Prof. C. R. Stockard. They were collected in Naples.

1 wish to take this opportunity to express my thanks and appreciation to Professor Gage and Professor Stockard for this valuable material.

The specimens were fixed in various fixing fluids: Zenker, formol; Gilson's, Bouin's, picro-acetic, corrosive-absolute alcohol. After being imbedded in paraffin, transverse and frontal sections were cut from 4 micra to 15 micra in thickness. The sections weie stained by various staining methods; methylene blue and eosin, Weigert's haematoxylin and eosin. Giemsa's eosin-azur, and haematoxylin (Delafield's) and eosin. The best differentiation was obtained by the use of the ordinary haematoxylin and eosin method.

A paraffined blotting paper model- of a part of the branchial region of a petromyzon larva prepared by Mr. Warburton was used in this study. A clay model of a part of the branchial region was also prepared to facilitate the study of the arrangement of the ciliated bands in the pharynx.


The phylogenetic position of the lamprey is still a matter of speculation. Various hypothesis have been advanced in regard to its position. Some place this animal between the amphioxus and the elasmobranchs, others claim that it represents a degen 2 A description of the method of reconstruction referred to above may be of interest to workers. This method is a modification of the late Mrs. Gage's blotting paper method. Sheets of blotting paper are dipped in melted paraffin and dried. The drawings are transferred to the paraffined paper by the usual methods when wax plates are used. The cutting is also done in the same way, the knife used, however, must have a thin but strong blade. In stacking the sections bank pins were used to hold the sections together. Small screws were also used occasionally to give firmness. When the stacking has been completed the sections may be smoothed down by means of any rounded instrument. A hot iron may also be used to cement the sections together. To give the best stability the complete model may be immersed in hot paraffin a few minutes. Models made in this way have a great firmness and are admirably efficient for class room use where a great deal of handling is necessary.


eratc type. However, a comparative study of the branchial region of the lamprey larva with the same region in ascidians and amphioxus on the one hand, and with the elasmobranchs on the other, suggests that the branchial region of the lamprey larva represents a transitional stage between the amphioxus and the elasmobranch types.

A few comparisons between the pharyngeal region of the ammocoete and lower and higher forms may be found in the literature. Dohrn ('84, '85) discussed the homology of the thyreoid of ammocoetes with the endostyle of ascidians and the hypobranchial ridge in amphioxus and the circumoral ciliated ring in the ammocoetes mth that of the ascidians. Cunningham ('87) verifies the homologies Dohrn pointed out. Shipley ('87) calls attention to the homology of the dorsal ciliated ridge in ammocoetes and the dorsal lamellae of ascidians and the epipharyngeal groove of amphioxus.

The following considerations are based on my studies and include besides the homologies just quoted a comparison of the gills in these primitive animals:

The large branchial cavity with, its medial gill arches of the lamprey larva (text fig. 1) is very suggestive of the conditions in the ascidians and amphioxus. In the ascidians there is a central pharynx surrounded by a peribranchial cavity. The two cavities communicate by means of numerous small pores, the stigmata. It is an unsettled problem whether the peribranchial cavity is derived from ectoderm or endoderm. In amphioxus there is a central pharynx which is partially surrounded by an atrium (peribranchial ca\'ity). In this form the two cavities communicate by means of definite gill slits. The atrium of amphioxus is developed from ectoderm. In both forms there is a ventral endostyle and a structure homologous with the epipharyngeal ridge. The branchial cavity of the ammocoetes corresponds to a fusion of the two separate cavities in the ascidians and amphioxus. The primitive characters of these separate cavities, however, are still present. The central portion, that is the part bounded laterally by the gill arches (fig. 1, a.p.), corresponds to the pharynx of the ascidians and amphioxus.



ep r.


It differs from the conditions in amphioxus in that the gill clefts are very much wider. It is probable that the ammocoete gill cleft represents the fusion of two or more gill clefts of amphioxus. There is an indication of such a fusion in the formation of the tongue-bar or secondary gill bar of amphioxus. The larger respiratory part of the ammocoete pharynx (text fig. 1, r.p.) corresponds to the peribranchial cavity of ascidians and amphioxus. The entire branchial cavity of the petromyzon larva is entodermal in origin. The ammocoete thus represents a phylogenetic stage in which the respiratory cavity, originaliy of ectodermal origin, is derived from the endoderm as it is in most higher animals.

The elasmobranch pharynx, it seems to me, represents an advanced stage of a modification which is already indicated in the ammocoete. This modification consists of a lateral migration of the dorsal and ventral attachments of the gill arches, resulting in a lateral enlargement of the central portion of the pharyngeal cavity and a consequent reduction of the respiratory part. This lateral migration is indicated by the dorsal attachments of the second pair of gill arches in the lamprey larva.

Attention may also be called to the fact that the primitive elongated character of the pharynx in the ammocoete tends to obscure its relation to the elasmobranch pharynx, in which the length has been reduced with a consequent reduction of the number and size of the gill slits.

It is necessary to determine the character of the gills in ammocoetes in so far that it has been established that the thymus in all higher forms has a more or less definite relationship to the gill pouches and gill arches.

Text fig. 1 Model of a segment of the branchial region of a 15 mm. lamprey larva. Cephalic aspect. The model shows the relationship of the primitive thymus placodes to the epipharyngeal ridge and the ciliated epithelium, as well as the relationship of the epipharyngeal ridge to the gill arches and general branchial cavity, a., atropore; a.p., alimentary pharynx; c.c, ciliated epithelium; d.a., dorsal aorta; ep.r., epipharyngeal ridge; g.a., gill arch; g.L, gill lamellae; n., notochord; p.t.p., primitive thymus placode; /•./;., respiratory pharynx; S.C., spinal cord; v. a., ventral aorta; .r., position of accumulations of Ij-mphocytes in lateral branchial walls.


l)()hi-n CS4) made the statement:

the ^^rvixt (lilTcrcucc Ix'twccn th(> Selachian, Telcost and Ganoid branchial a])iiaratus and that of the ]M'troTnyz()n consists therein that the sill se])ta and lan)(>lhie (Kiem(n-l)l;i,tter und -l)lattchen) of the former ar(> directed outwai'd while in the latter they are directed inward.

He further states that this arrangement in the petromyzon exists from the beginning. This interpretation of the gills of petromyzon has been accepted in some of the textbooks on Comparative Anatomy of Vertebrates The basis for this interpretation is undoubtedly found in the position of the cartilaginous gill bars, which form a complicated branchial basket in the pharyngeal wall The branchial artery, however, is situated in the medial gill arch. From this medial gill arch the gill septum extends caudo-laterally to its attachment in the lateral wall. The gill lamellae are situated on the anterior and posterior walls of the septum. The picture of a frontal section of the gills in the ammocoetes is so much like the picture of a similar section of the elasmobranch gills that it is difficult to consider them as directed in opposite directions. The question resolves itself nto a choice between the cartilaginous branchial bars and the branchial aortic arches as a basis of interpretation. It is evident that the branchial basket of petromyzon is a special modification meeting the requirements of a specialized mode of breathing due to the life habits of the adult. The position of the cartilaginous gill bars must then be considered the result of a migration from a more medial position. Moreover, the presence of the ciliated bands in the medial gill arches point to a direct phylogenetic relationship to the gill arches of amphioxus. If we consider the gills of the ammocoete as directed inward it would be necessary to consider as the gill arch, the part of the respiratory portion of the lateral branchial wall to which the gill lamella is attached. This would be contrary to the arrangement of the gills in all other chordate animals.



A complete detailed description of the development of the branchial region in the lamprey larva is apparently not to be found in the literature. Separate structures and the condition in a single or in a limited number of developmental stages, however, have been described by various investigators. These descriptions have been accurate with the exception of minor details, but having been limited to a single stage in most cases they do not include the changes which occur with the growth of the larva. There are consequently contradictory statements in the litcT-ature on the pharynx of the ammocoetes and especially in the part dealing with the ciliated grooves and bands. Further, the formation of structures which I interpret as primitive thymus placodes is closely linked with the changes which occur in the ciliated bands in the branchial lining.

The following descriptions are based entirely upon my own material :

The transformations in the early larvae are very rapid so that in 6 and 7 mm. larvae gill lamellae have formed on the branchial septa, the pouches open to the outside, and the epithelium is represented by more than one layer. A system of ciliated epithelial grooves and bands are present in this stage of development of the pharynx. They form a connected system which may be looked upon as beginning in two rather deep diverticula in the caudal walls of the first pair of gill pouches. From each diverticulum two ciliated grooves originate, one passing ventro-caudally, the other dorso-caudally. These will be designated the ventral and doisal grooves respectively.

The two ventral grooves converge in a caudal direction as far as the third pair of gill pouches where they come to lie close together and parallel to each other near the median line. Between the two grooves is a median ridge of non-ciliated epithelium w^hich disappears in the fourth pouch where the two ciliated grooves fuse to form a single one. A tubular divericulum passes from the ventral groove into the thyreoid a short distance caudad of the point where the grooves fuse. A second diverti


('Ilium coiiiuM'ts tlu^ thyreoid witli the fifth pouch. In the seventh ])ou('li the cihated groove ends. It is (Hrc^ctly continued by non-cihatetl epithehuni which, in a few sections caudad, becomes a ridge. The cihated groove is a continuous groove from the fourth to the seventh pouches. The ridge which begins in the seventh pouch gradually becomes high and stalked in the eighth pouch. The median columnar epithelium becomes invaginated and is directly continuous into the floor of the oesophagus. Surrounding the junction it seems to me there is evidence of a vestigial eighth gill arch in w-hich an aortic arch is not present. The respiratory part of the branchial cavity extends a short distance caudad of the point of junction betw^een the pharynx and oesophagus.

The dorsal ciliated grooves arising in the diverticula follow the course of the first aortic arch to the median dorsal line of the pharynx. They fuse at this point to form a single ciliated band which extends caudally the whole length of the branchial cavity and w^hich is directly continuous into the roof of the oesophagus. A short distance caudad of the point of origin, this band forms a rounded ridge which extends to the seventh sac w^here it is converted into a groove. The aorta is lodged in the concavity of the rounded ridge. At the point where the tw^o dorsal grooves of the first pouch fuse a tongue-like piece of non-ciliated epithelium is pinched off (text fig. 2). Schaffer apparently mistook this for ciliated epithelium and considered it the end of the fused ciliated bands.

The first pair of gill arches come together dorsally in the median line. Their ventral extremities, however, are far apart and end in the ventro-lateral part of the respiratory pharynx. Gill lamellae are present only on the caudal surface of the first gill septum. The second pair of arches are farther apart and thus they differ from the remaining caudal arches. Their dorsal attachments are in the angle between the epipharyngeal ridge and the dorso-medial part of the respiratory pharynx. \'entrally, the second gill arches are attached about midway between the mid-ventral line and the ventro-lateral angle of the respiratory pharynx. The dorsal attachments of the third and remain



iiig pairs of gill arches is the ventro-lateral angle of the epipharj^ngeal ridge. Ventrally, they are attached near the ventral median line, a Uttle to the side of the endostyle. The second pair of arches are pecuHar in that they contain no cili

t h i

Text fig. 2 Camera lucida outline drawings to illustrate the course of the dorsal ciliated grooves (the ventral grooves are not shown) from the diverticula in the first pouch to the point where the grooves meet in the median dorsal line fo form a single band of ciliated epithelium, e.g., dorsal ciliated groove; d.a., dorsal aorta; e.p.. epithelial placode; «., notochord.


ated bands. The third and remaining arches have a broad ciHated band covering the medial and cephalic aspect. These bands are directly continuous with the ciliated band on the epipharyngeal ridge. They have no connection with the endostyle in this stage of development and I have been unable to determine whether such a connection exists or not in younger larvae.

The moving apart of the second pair of gill arches is very suggestive of an approach to the condition in fishes where the arches are attached in the lateral part of the roof of the pharynx. Accompanying this lateral migration there is a loss of the ciliated band on the arch.

The arrangement of the ciliated bands as described above does not persist in older larvae. This undoubtedly accounts for the contradictory descriptions given by Anton Schneider ('79) and Schaffer ('95, 1 and 2) and others. In a larva 9.5 mm. in length growth and differentiation of the epithelium of the gill arches has resulted in a new arrangement of the ciliated bands. This new arrangement has gained its permanent larval condition in a 15 mm. larva.

In the older larvae the median dorsal ciliated band which represents the fused continuation of the dorsal ciliated grooves of the first arch ends in the median dorsal line betw^een the second pair of gill arches. Immediately caudad of the dorsal attachment of the second pair of gill arches two ciliated bands appear on the ventro-lateral part of the epipharyngeal ridge. Tracing these bands in a caudal direction, they are seen to come together and fuse in the median ventral part of the ridge at the caudal end of the dorsal equivalent of the third gill pouches. From this single band a branch is given to each of the third pair of gill arches. In the median line the ciliated band ends as a pointed process in the angle between the dorsal attachments of the third pair of gill arches. This arrangement of the ciliated bands is repeated in the remaining arches and dorsal equivalents of the gill pouches (text fig. 3, and c.e. in text fig. 1). In the eighth pouch, however, the median A'entral fused part does not give off any lateral branches corresponding to the ones given off to each gill arch in the third to the seventh arches. This is



due to the nature of the dorsal attachment of what I have assumed must be the eighth vestigial gill arch. This arch, like the second, is not attached to the epipharyngeal ridge but to the angle between the ridge and the dorsal wall of the respiratory pharynx. Cilia, however, are present on the lateral side of the eighth gill arch. They undoubtedly represent the vestigial remains of a condition in which this arch had a dorsal attachment to the ridge or its equivalent. The median ciliated band

Text fig. 3 Diagram illustrating the arrangement of the ciliated bands on the epipharyngeal ridge and the gill arches in a 31 mm. larva. Ventral view. The gill arches which extend ventrally are here represented as extending laterally. The ciliated epithelium is represented in heavy black, the non-ciliated by the stippled part. Roman numbers indicate the gill pouches and Arabic numbers the gill arches, pi., primitive thymus placodes.


of the eighth pouch is directly continuous into the oesophagus. It is interesting to find that the band divides into two portions within the oesophagus. These perhaps represent the branches which were given off to the eighth pair of gill arches in an ancestral form.

Patches of ciliated epithelium are also present on the medial aspect of each gill arch from the third to the seventh inclusive. These ciliated patches undoubtedly were a part of the single ciliated band on the gill arches in the younger larva. With the growth of the non-ciliated epithelium of the gill arches, these patches were cut off from the ciliated band. In the older larva the ciliated band of the gill arch does not occupy the same relative position that it did in the young larva. The ventral end is situated on the lateral side of the gill arch. When traced in a dorsal direction it is found to take a slightly spiral course so that the dorsal extremity which is continuous into the ciliated band of the epipharyngeal ridge comes to occupy a cephalomedial position. This change in the course of the ciliated band naay also be looked upon as the result of the growth of the nonciliated epithelium. The ciliated patches on the medial side of the gill arch may acquire a sensory function as Schaffer suggested.

The arrangement of the ciliated epithelium on the epipharyngeal ridge in the older larvae is also the result of the growth of the non-ciliated epithelium. In a larva between 8 and 9 mm. in length the non-ciliated epithelium of the dorso-medial part of the gill arches begins to invade the ciliated epithelium of the epipharyngeal ridge. As a result of this invasion the continuity of the ciliated band on the epipharyngeal ridge is lost. The two cords of invading epithelium from the opposing gill arches fuse in the median ventral line of the ridge. The invasion is continued in a caudal direction dividing the ciliated band into two portions which are pushed laterally. This fused portion {pL, text fig. 3) becomes thicker and broader in a caudal direction and ends a short distance cephalad to the attachment of the next pair of gill arches. At the caudal end of this invading epithelium the subsequent growth does not divide the cihated epithelium fur


ther, but produced a tongue-like process which projects into the pharyngeal cavity (similar to the tongue-like non-ciliated epitheUal process represented in text figure 2).

The non-ciliated epithelium which has invaded the ciliated epithelium of the epipharyngeal ridge begins to show histogenetic actiWties in a larva between 20 and 30 mm. in length. The nuclei of the cells of these areas, or placodes, wander out into the underlying connective tissues and are transformed into lymphocyte-like cells. A study of the histogenetic processes in these areas in various developmental stages leads to the conclusion that these areas represent specialized regions of the branchial epithelium which are suggestive of primitive thymus structures.

From the foregoing description it is evident that there are in all seven placodes. The seventh and the first are smaller than the remaining ones but they take part in the histogenetic processes and are therefore to be considered true functional placodes. The placodes increase in size with the growth of the larva. In the mature larva, however, they show a depletion of cells.

When the larva undergoes metamorphosis the whole structural arrangement of the branchial-region is altered. In the single specimen of a transforming larva of my collection, the adult arrangement has been attained, so I am unable to describe the nature of this process. In this transforming specimen I have also been unable to find any remains of the epithelial placodes of the larva. Serial sections of the branchial region of an adult lamprey have also been examined but with negative results. It is evident that an involution of the placodal organ has taken place as one would expect of a thymus. This involution began in the maturing larva and was completed in the early stages of metamorphosis .


Early development of the pharyngeal wall in Petromyzon marinus unicolor, Lampetra wilderi

The search for a thymus in the lamprey larva has revealed an unusual accumulation of lymphocytes in the lateral walls of the branchial cavity. These accumulations were first observed by


Schaffer ('94) and later described more in detail by Giacomini ('00, 1 and 2) and Castellaneta ('13). The origin of these lymphocytes as well as the cause for their accumulation in these places has not been determined and consequently constitutes a problem to be solved in the consideration of a possible thymus structure in this animal. The descriptions in the literature have been limited to single stages of development and are consequently incomplete. The following descriptions are based on my own material and includes the essential developmental stages :

The branchial cavity of a 5 mm. larva represents a very simple condition. The gill pouches are present as mde evaginations extending to the ectoderm in the mid-lateral plane leaving the gills as simple projections of the lateral endodermal walls. Loose mesenchyma cells fill up the space within the gill. The aortic arches are forming near the free medial border of the gills, but the vascularization of the body of the gill has not begun as yet. Although the larva contains a great quantity of yolk in this stage the branchial region is quite free from it.

The epithelium lining the branchial cavity including the gills consists of a single layer of columnar cells. Dorsal and ventralto the mid-lateral plane of each gill pouch and corresponding to the position of the future lymphoid accumulations of Schaffer the endoderm shows a slight thickening (fig. 2). In these places which may be called placodes the cells have lost their columnar shape and their outlines have more or less disappeared. The area appears to be taking on a syncytial character in which the nuclei do not have any definite grouping.

Marked changes have occurred in these placodes in a 9.5 mm. larva. In general, the placodes have become enlarged both in thickness and area (fig. 3). Cell outlines are practically all obliterated. The cytoplasm is streaky in appearance suggesting a degeneration. The nuclei exhibit variations in character. Some are very dark with a large chromatin content while others are pale and contain a small amount of chromatin. Still others show amoeboid characters. I have not been able to determine whether all these nuclei are indigenous to the placode. The


general epithelium of the branchial cavity has also acquired a new character in this stage. Beneath the placodes and in direct contact with them the peribranchial blood channels are forming. It is difficult to distinguish an endothelial lining of these channels in all cases.

The degeneration of the placode has progressed farther in a 15 mm. larva (fig. 4). In places the cytoplasm has taken the stain very faintly.. Scattered about in the placode are streaks of cytoplasm which are very deeply stained. Vacuoles are also present. The nuclei appear to be fewer in number than in the earlier stages. They appear more constant in their general appearance and chromatin content. The amoeboid character of the nuclei has also become more prominent. Some nuclei have taken up positions at the surface of the placode and the cytoplasm appears to be cutting off a layer of flattened cell (fig. 4, s.L). The formation of a layer of flat cells at the surface of the general epithelium was begun in a much earlier stage of development.

The changes which occur in these placodes in older larvae approximate the character represented in the 63 mm. larva (fig. 1). It is a significant fact that cells in mitosis have not been seen in any stage of the development of these placodes. Furthermore, patches of epithelium giving the appearance of a degeneration are present in various other parts of the branchial lining, especially at the lateral attachments of the gill septa. Lymphocytes are present in the placodes in the older larvae, but they are also present in the general branchial epithelium. They do, however, occur in greater numbers at these placodes.

From these brief descriptions it is apparent that these placodes do not represent active anlagen of a future structure. Their development and structure do not suggest anything which might indicate their significance.

The lymphocyte accumulations in relation to the above described placodes are contained within vascular channels. These vascular channels contain red blood cells, and as has been shown by Mozejko ('10) and others they are in communication with



similar channels within the body of the gill as well as with the definitive blood vessels of the pharynx.

Lymphocytes begin to make their first appearance in the blood in larvae of about 9 mm. length. They increase in number with the growth of the larvae, but chiefly remain outside of the main blood vessels. They are especially abundant in the perivascular spaces of the gill arches.

The accumulations of lymphocytes in the lateral branchial walls are foreign to these situations so far as their origin is concerned. Furthermore, the epithelial placodes in these situations together with the lymphoid accumulations do not exhibit the characters which are essential in either a well-established or rudimentary thymus. I can offer no suggestion in regard to any special significance of these accumulations. They appear to be merely a part of a rich accumulation of lymphocytes in the connective tissue spaces of the branchial region. It is probable that the apparently degenerating placodes play a role of attraction for lymphocytes.

Epipharyngeal placodes

The placodes in the epipharyngeal ridge are present in an undifferentiated condition in a 15 mm. larva. They form distinct masses of cells in the mid-ventral part of the epipharyngeal ridge between the ciliated bands. They are very nearly circular in outline in a transverse section, producing a bulging into the interior of the ridge (figs. 5 and 6). A loose mesenchymatous tissue caps the dorsal surface of the placodes. Red blood cells are occasionally present in the spaces of the mesenchymal tissue. These spaces are apparently in communication mth the dorsal aorta by means of minute apertures and are directlj^ continuous with siixdlar spaces in the connective tissue of the gill arches. Through the spaces of the gill arches a communication is also made with the peribranchial sinuses and the perivascular spaces in the gills.

Within the placode the cells are in an active state of proliferation. The nuclei of the resting cells are rather clear


structures with the chromatin generally collected into small lumps situated next to the nuclear membrane. The nuclei are smaller than those of the ciliated epithelial cells. Cell outlines are more or less distinct in the placodes. A peculiar type of vacuolization is in progress in some of the cells in which the complete cell becomes vacuolated leaving the protoplasmic remains as free bodies within the hollow cell. The protoplasmic bodies in these cases are small lumps of nucleated protoplasm in which the nuclear material generally stains an intense black and the cytoplasm a light red. These protoplasmic bodies are not limited to the epithelial placodes, but may be found everywhere in the branchial epithelium, especially in the 15 mm. larva.

A layer of flat cells clothes the surface of the placode. This layer is not distinct in every section and may easily be overlooked. At the connective tissue border of the placode a basement membrane sharply marks off the epithelium from the mesoderm (b.rn., fig. 5). Lymphocytes have not been found in the placodes in a 15 mm. larva although they are present in the blood. They may be seen, however, in the general branchial epithelium and also in the ciliated epithelium (Im., fig. 6).

The placodes have increased considerably in size in a 31 mm. larva. The mesenchymatous tissue which was present above the placode in the 15 mm. stage has changed to connective tissue (fig. 7). Large spaces containing various kinds of blood cells are present in this connective tissue. Larger and smaller nuclei may be seen in the walls of these spaces. The smaller undoubtedly represent the nuclei of endothelial cells. The larger, however, are apparently derived from the placode and are in a stage of migration into the vascular spaces. It is doubtful whether these spaces should be considered true blood channels. While red blood cells are quite abundant in the spaces in this stage of development they are practically absent in them in the full grown larva. It is probable that they represent a primitive type of lymph vessels, as has been suggested by various authors. The connective tissue is of the fibrous variety in which the individual fibers are quite slender. The fibers interlace to


form a loose mesh work. Figure 7 represents a transverse section through the cephalic part of a placode with its neighboring connective tissue and the ciliated epithelium of one side. In such a region the nuclear elements are very scarce in the connective tissue when compared to the region above the central part of a placode. The basement membrane which is present only on the right-hand side in the section illustrated in figure 7 (b.m.) bridges across the entire placode a few sections cephalad of the one illustrated. In the central part there is no line of demarcation between the placode and the connective tissue. The cytoplasm of the placode in this place is directly continuous with the connective tissue.

The cells within the placode have greatly increased in numbers in this stage of development. Near the free surface of the placode they are loosened from each other, displaying their rounded outlines distinctly. Toward the deeper part of the placode the cells become oblong in shape. Near the connective tissue border the cell outlines are lost which gives the appearance of a syncytium. The nuclei of the cells in the placode are not unlike the nuclei of the ciliated and the general branchial epithelium in their general morphological characters, except in size. They are smaller than the nuclei of the ciliated and general epithelium. The chromatin of the nuclei, for the most part, is collected into a single lump which stains a reddish-purple with the haematoxylin-eosin stain. The nuclei also change from a circular to an oblong outline from the free surface of the placode to the connective tissue border. At the place where the cytoplasm of the placode is continuous with the connective tissue, the nuclei become quite elongated, having the appearance suggestive of a migration into the connective tissue. This migratory appearance is more prominent at the central part of the placode. Figure 10 represents a part of a transverse section from the central region of the placode. The two lower nuclei marked a in the figure lie in the cytoplasm of the placode. All the nuclei and cells above this level are in the connective tissue and spaces above the placode. The nuclei in the connective tissue show degrees of gradual variations in morphological


characters from the typical epithehal nucleus of the placode to the mature lymphocyte-like cell. Nuclei showing these degrees of variation may all be found in a single section. Figures 10 and 11 show the most obvious stages in this gradual variation. The nuclei marked a are typical placode epithelium nuclei apparently in a state of emigration. The nucleus marked b is in the connective tissue. The chromatin in this nucleus is apparently breaking up into a number of granules, a process which has proceeded farther in nucleus c. In nucleus d the chromatin gi-anules are apparently arranging themselves on the nuclear membrane, an arrangement which has been completely attained in nucleus e. Nucleus e further shows a tendency towards acquiring a circular outline which becomes more manifest in nucleus /. Nucleus / also shows a reduction in size. Nucleus g (fig. 11) has a circular outline and further shows a change in the character of the protoplasm. The nucleus h shows a still further reduction in size, the protoplasm stains darker as does also the chromatin. The chromatin, further, forms a continuous layer at the periphery. In i (figs. 10 and 11) the nuclei have acquired a thin covering of cytoplasm which is not visible at all points of the nuclear surface. The cytoplasm stains a grayblue. The nucleoplasm and the chromatin of these cells take the stain more intensely than the nucleus h. These cells have also become free from the connective tissue mesh work. In the cells j the nucleoplasm stains a deep purple. The chromatin appears to have left the nuclear membrane and is now present as granules scattered about in the nucleus. In some nuclei the chromatin granules are connected together by slender processes, in others this is apparently not the case. Still other cells show nuclei in which the chromatin is represented by a single large lump. These cells (j, figs. 10 and 11) represent the typical lymphocyte-like cell in this region of the 31 mm. larva. Some of these cells may be found in which the nucleoplasm stains a gray-blue {k, fig. 10). They are similar to the lymphocytes in older larvae and may either represent a final stage in the development of the cells, or they may represent cells foreign to this locality.


From the above described transitional conditions and from a study of the stained sections, I can draw no other conclusion than that, the nuclei of the epithehal placode transform into lymphocyte-like cells. It is a significant circumstance that the nuclei alone migrate from the placode, i.e., no cytoplasm is visible. Complete cells bearing epithelial characters may be found in the connective tissue spaces. However, I have never found them migrating from the placode while I have found migrations of the complete cell from the epithelium of other regions.

Cells in mitosis may be seen occasionally in the placodes of the 31 mm. larva. Cells in a state of amitotic division, however, are quite abundant in a 44 mm. larva, suggesting that cell-proliferation takes place chiefly by simple fission. Figure 15 shows the nucleus of a placode cell apparently in a process of simple fission. Mitotic cells are especially scarce in the connective tissue above the placode. A single instance has been found and is represented in figure 18. It is quite evident from the lack of mitotic or amitotic cells or nuclei in the connective tissue that cells or nuclei are not being formed in any significant quantities in this situation.

Transformation stages have not been found within the placode in the 31 mm. larva. Lymphocyte-like cells, however, are present in the placodes. Their presence may be accounted for by means of an immigration from the connective tissue.

The further development of the placodes is a repetition of the above-described processes except that the transformation is more rapid and begins within the placode. Figure 8 represents a part of a transverse section of the placode and the connective tissue above it in a 44 mm. larva. The illustration was drawn to the same magnification as figure 7. The nuclei in the placode are elongated and show amoeboid characters. They also appear to be in an active state of emigration. The transformation process appears to have begun in the placode in this stage. The nuclei near the connective tissue border have taken on characters which approach the characters of some of the nuclei in the connective tissue. This change is shown in the staining reaction, the condition of the chromatin, and the shape of the


nuclei. The nuclei at the border show a tendency to stain blue, the chromatin takes a darker stain, and in some cases is broken up into granules, and the nuclei approach the globular shape. The transformation of the nuclei in the connective tissue appears to be of the same character as in the 31 mm. stage, but apparently more rapid. Nuclei may occasionally be found in the placode which show phagocytic properties. Figure 14 represents a placode nucleus in the act of engulfing protoplasmic bodies.

The spaces in the connective tissue in the 44 mm. larva appear to be smaller than in the 31 mm. stage. Some of them have a distinct wall while others appear like transient spaces in the connective tissue. Red blood cells are only occasionally seen in the connective tissue spaces of this stage.

The placodes in a 63 mm. larva are larger in area but thinner than in the preceding stages. The nuclei of the epithelial cells of the placode have lost their original character. The chromatin is no longer represented by a single large lump, but is present in the form of granules corresponding to the chromatin in the nuclei which had migrated into the connective tissue in the 31 mm. stage. The number of lymphocyte-like cells has increased considerably within the placode. All the stages of transformation from epithelial nucleus to the mature lymphocyte-like cell may be found within the placode in this stage of development. A basement membrane is re-forming at the connective tissue border of the placode. The 'vascular spaces' of the connective tissue are now chiefly limited lo the peripheral part of the whole connective tissue within the ventral half of the epipharyngeal ridge. The mature lymphocyte-like cells are chiefly located in these channels, leaving the central connective tissue core quite free from cells. The central core consequently has a much lighter appearance. Some nuclei are present in the central core, the morphological characters of which are similar to the characters of connective tissue nuclei in other parts of the body. Other nuclei may occasionally be seen in which the characters agree with the various transformation stages of the lymphocyte-like formation shown in younger larvae.


The activities within the placodes of the full grown larva (120 mm.) have diminished and are apparently approaching a condition of cessation. The number of epithelial nuclei has been reduced considerably. Transitional stages may be found, but are quite scarce. Mature lymphocyte-like cells are also present, but not in great numbers. A definite basement membrane is now present at the connective tissue border of the placode. Figure 9 represents a portion of a transverse section of the placode and the tissue above it in a 120 mm. larva. The section is taken near the cephalic end of the placode. In such a region a peculiar formation has occurred in the connective tissue, the significance of which I am quite unable to explain. This formation consists of what appears to be red blood cells held in the meshes .of the connective tissue {x, fig. 9). The cells have the morphological characters of the red blood cells. The cytoplasm has a decided yellow tint, while the pale nuclei have a green tint. In some cases what appears to be the nuclei have morphological characters similar to the lymphocyte-like cells. These formations are present in the periphery of the whole connective tissue. A section through the central part of the placode would show the same character that was indicated in the 63 mm. larva, that is, a central core of connective tissue in which there are no 'vascular channels' surrounded by a 'vascular area.' The tissue between the 'vascular channels' in the 120 mm. larva consists entirely of the peculiar tissue just described.

Development of lymphocytes

A brief description of the general development of lymphocytes in the petromyzon larva is here given since the nature of this formation in the advanced larvae has a. direct bearing on the interpretation of the histogenetic processes in the above described placodes. My observations do not include the first appearance and development of the blood in the embryo, but begin with the development in the 5 mm. larva. The nature of the blood formation in this stage of larval development need not be described here for the reason that it occurs at a time when the placodes have not begun to form. However, in larvae


ranging from 9.5 mm. in length up to the mature individual, blood cells develop from the epithelial cells of the gills, gill arches, and probably the branchial wall by a process of transformation. It is the blood formation occurring in the gills and gill arches which is of especial interest in connection with the histogenesis in the placodes. The description of this formation will be limited to the formation of lymphocytes only, in the 31 mm. larva.

The similarity of the cytoplasm of the gill epithelium to the cytoplasm of some of the blood cells was early noticed. This similarity was found to be due to an actual relationship between the two kinds of cells and thus not a mere coincidence. This relationship was demonstrated when epithelial cells were found migrating through the walls into the lumen of the blood channels in the gill. Figure 13 represents a part of the gill epithelium and a blood vessel and shows an epithelial cell beginning its migration into the vessel. Figure 12 a shows another epithelial cell in the state of migration, almost half of the cell in this case is inside of the vessel. The cells to the left in figure 12 represent blood cells (in the vessel) in various stages of transformation. In this figure, the chief stages in the transformation of the epithelial cells to lymphocytes are represented. The lettering a to h in the figure shows the hne of transition from an epithelial to the mature lymphocyte.

In this formation of lymphocytes, it is noteworthy that the entire epithelial cell migrates from the 'epithelium' and takes part in the transformation. The transformation consists of a reduction in the size of the nucleus and also in the amount of cytoplasm. The cytoplasm retains its staining qualities through these changes so that even in the mature lymphocyte a cytoplasmic ring which stains red may be seen in many instances. It is very seldom that a lymphocyte containing a cytoplasm which stains a blue or gray-blue is. seen in these situations. All the transforming cells have a cytoplasm which stains red with the haematoxylin-eosin stain.

Although some of the epithelial cells in this transformation migrate directly into the blood vessels, the great majority wan


der into the perivascular spaces and undergo their transformation in these places. The sluggish character of the blood flow in these spaces must account for the retention of the large number of transforming and mature lymphocytes which are present in these situations. The entrance of these cells into the main blood vessels is of a slow nature.

The tall epithelial cells in the dorsal part of the epipharyngeal ridge also enter into the blood formation. Figure 16 shows a cell taken from the space in the connective tissue of the dorsal part of the epipharyngeal ridge. The nucleus has the morphological characters of the epithelial nuclei. It appears to be in a state of simple fission. The cell in figure 17 was taken from the same locality. Two nuclei are present in this cell which still show the epithelial character.

Histogenetic comparisons between cells arising from placodes and


In the study of the histogenetic processes in the placodes it was shown by means of various transitional stages that the epithelial cells of the placode transform into lymphocyte-like cells. Lymphocytes were shown to develop from the 'epithelial' cells of the gills and gill arches. The lymphocyte-like cells formed from the placodes do not have the same mode of development nor do the transitional forms have the same morphological characters as the lymphocytes and transitional forms developed from the gill and gill arch 'epithelium.' In the placode the nuclei alone migrate away from the original epithelial bed and the transformation occiu*s in the connective tissue meshwork. The complete cell migrates away from the epithelial bed in the gills and gill arches, the transformation occurs in the perivascular spaces and the blood vessels. A small amount of cytoplasm becomes visible in the placode 'lymphocyte' just before it attains its maturity. This cytoplasm stains a gray-blue. The cytoplasm of the gill and gill arch lymphocytes represent the original cytoplasm of the 'epithelial' cells and stains red. These important differences in the lymphocytes and lymphocyte-like cells


occur in the same section and thus cannot be attributed to difference of technique. A lymphocyte with red cytoplasm may occasionally be found in the epipharyngeal ridge just as a 'lymphocyte' wdth gray-blue cytoplasm may occasionally be found in the gill region. The great majority of 'lymphocytes' in the placode region, however, contain cytoplasm which stains gray-blue. It was also pointed out above that the lymphocytes in the gill region are chiefly the type which have red cytoplasm. The presence of the lymphocytes with red cytoplasm in the placode region and the type with gray-blue cytoplasm in the gill region may be accounted for by migration from their seats of origin. They may also be brought to these situations by the flow of the blood.

On account of the morphological difference of the developing lymphocytes in the gill region and lymphocyte-like cells of the placodes, the conclusion seems justifiable that the placodes are segregated portions of the 'epithelium' representing individual organs which produce cells of a lymphocyte appearance, but differing from the lymphocytes formed in the 'epithelium' of the gills.


The data submitted in the consideration of the lymphocyte accumulations in the lateral branchial wall of the lamprey larva does not supply any evidence that these formations represent primitive thymus anlagen. Although placode-like formations are present in the lateral branchial wall, similar formations are also present in other parts of the pharyngeal epithelium.

An important component of the thymus of higher animals is a reticulum. In my study of the thymus-like placodes in the lamprey larva, I have been unable to find any undisputable evidence of a reticulum in the placode. At the connective tissue border of the placode the epithelial cytoplasm apparently has a fibrous character (fig. 10). I have been unable to determine whether this represents connective tissue or transformed epithelial cytoplasm. Judging by its appearance and position it probably represents connective tissue which has been invaded by cyto


plasm from the placode. The connective tissue outside the placode plays the role of a reticulum insofar that the transformation of the epithelial nuclei occur within its meshes.

Hassal's corpuscles, or any structures comparable to them have not been found in the placodes or in the connective tissue outside of the placode.

The history of the placodes in the successive developmental stages indicates a gradual involution of the placodes. The maximum size of the placodes occurs in a larva of 50 to CO mm. in length. From this stage of development the placodes diminish in size so that in the mature larva very few lymphocyte-like cells remain. In the transformation of the larva, Nestler ('10) maintains that the oesophagus of the adult is formed by a transformation of ^'the under edge of the dorsal fold in the branchial chamber" (the epipharyngeal ridge). If such a process occurs, is is only after the histogenetic activities in the placodes have ceased and consequently does not affect the status of an earlier thymic function in these placodes.

An examination of the descriptions given in the preceding pages give the impression that the primitive thymus placodes and lymphocytes are formed from an endodermal epithelium. While I am not ready at this time to supply the evidence, the changes which occur in the general branchial epithelium in the early stages of development seem to point to a general fusion of the original endoderm with the underlying mesenchyma. The character of the epithelium in the more advanced larvae has such an important bearing on the interpretation of the histogenesis of the primitive thymus cells and lymphocytes that a separate and detailed study of this process seems warranted.

In a recent article on the Development of the Human Pharynx, Kingsbury ('15) discusses the intrinsic and extrinsic factors in thymus formation and challenges the view that the thymus is a branchiomeric organ definitely located in the branchial epithelium. The basis for this interpretation

is found in the recognition that it is a. structure whose appearance is determined by extrinsic factors of relation and position and not intrinsic factors located in any particular group of cells. In support of


such an interpretation and giving us, I believe, a better comprehension of its morphologic significance, we have the fundamental plan of its histogenesis.

The true natui'e of the endodermal-mesenchymal relationship in the ammocoete pharynx has not been definitely determined. Whatever these extrinsic factors may be, they are apparently of the same nature in the thymus-like placodes and the lymphocyte-forming 'epithelium' of the branchial arches. The products of these two regions, however, are not similar and it seems to me that this dissimilarity can only be explained on the basis of an intrinsic value or specificity of the 'epithelium' of the placode.

The nature of the formation of lymphocytes and the primitive thymus placodes in the lamprey larva point to an ontogenetic relationship in the histogenesis of thymus cells and lymphocytes. The branchial region of the lamprey larva may be looked upon as possessing general lymphocyte-forming properties in which the primitive thymus placodes represent specialized regions of the general lymphocyte-forming 'epithelium.'


From the evidence obtained in this investigation of the ammocoetes the following conclusions seem justified:

The placodes in the lateral branchial wall are apparently patches of degenerating epithelium and have nothing to do with a thymus structure. The collection of lymphocytes at these places are foreign to this situation so far as their origin is concerned.

The gills in the ammocoetes are homologous with and extend in the same direction as the gills in elasmobranchs.

The branchial 'epithelium' does not represent a pure endodermal epithelium. This 'epithelium' develops haemopoetic properties in the advanced larva.

'Epithelium' from the gill arches invades the ciliated epithelium of the epipharyngeal ridge and produces placodes. These placodes have a relationship to the gill arches and gill pouches which makes them homologous with the thymus placodes of


elasmobranchs and are to be considered primitive thymus structures.

The lymphocyte-hke cells which originate in the primitive thymus placodes have different morphological characters and have a different mode of formation than the lymphocytes which are formed in the gill arches and lamellae.

This investigation has been pursued in the laboratories of anatomy at Cornell University Medical School and Marquette University Medical School. While I hope to have established a primitive thymus structure in the ammocoetes, many of the important problems of the histogenesis of the lymphocytes and primitive thymus cells must be left undecided until more exhaustive investigations can be completed.


Ankarsvard, G., und Hammar, J. 1913 Zur Kenntnis der Ganoidenthymus.

Zool. Jahrb. Abt. f . Anat. u. Ontog. der Tierre, Bd. 36, p. 3. Castellaneta, V. 1913 Sulla questione del timio in 'Ammocoetes.' Monitore

zool. Italiano, Anno 24, pp. 161-174. Cunningham, J. Y. 1887 Dr. Dohrn's inquiries into the evolution of organs

in the chordata. Quart. Journ. of Micr. Science, vol. 27, pp. 265-266. DoHRN, A. 1884 Studien zur Urgeschichte des Wirbeltierkorpers. IV. Die

Entwickelung und Differenzierung der Kiemenbogen der Selachier.

Mitteil. a. d. Zool. Station zu Neapel, Bd. 5.

1885 Studien zur Urgeschichte des Wirebeltierkorpers; VII. Entste hung und Bedeutung der Glandula Thyreoids; VIII. Die Thyreoidea

bei Petromyzon, Amphioxus und den Tunikaten, Mitteil. d. Zool.

Station zu Neapel, Bd. 6, pp. 44-92. GiACOMiNi, E. 1900 a Sulla Struttrua dells branchiedei Petromyzonti. Monit.

Zool. Ital., Anno 11, Suppl. 9-10.

1900 b Ibid, (cited from Oppel '05). GoETTE, A. 1875 Die Entwickelungsgeschichte der Unke. Leipzig (cited from

Hammar, '10).

1890 Entwickelungsgeschichte des Flussneumauges. Hamburg and

Leipzig. MoLLiER, S. 1906 Die Entwickelung von Blut und Gefassen. In Hertwig's

Handbuch der vegl. u. exp. Entwickelungsgeschichte der Wirbeltiere,

Jena. MozEJKO 1910 liber die Injektion des Vascularsystems von Petromyzon fluvia talis. Zeitschr. f. wiss. Mikrosk., Bd. 27.

1911 tjber den Bau und den morphologischen Wert des Vascularsys tem der Petromyzon. Anat. Anz., Bd. 40.


Rabl, C. 1886 Zur Bildungsgeschichte des Halses. Prager Med. Woch., Bd.

11, p. 52 (cited from Hammar, '10). Rathke, H. 1827 Bemerkung uber den imeren Bau des Querder (Ammocoetes

branchialis) und des Kleinen Neunauges (Petromyzon Planeri (cited

from Oppel, '05). ScHAFFER, J. 1894 tjber die Thymusanlage bei Petromyzon Planeri Sitzungsb.

d. K. Akad. d. Wiss. Wien., Bd. 103, p. 3.

1895 a Zur Kenntnis des Histologischen und Anatomischen Baues

von Ammocoetes. Anat. anz., Bd. 10, pp. 697-708.

1895 b tJber das Epithel des Kiemensarmes von Ammocoetes nebst

Bemerkungen uber intraepitheliale drusen. Arch. f. Mikr. Anat.,

Bd. 45, pp. 294-338.

1906 Berichtigung, die Schilddriise von Myxine betreffend. Anat.

Anz., Bd. 28. Schneider, A. 1879 Beitr. zur Vergl. Anatomie und Entwickelungsgeschichte

der Wirbeltiere, Berlin. ScHULTZE, M. 1856 Die Entwickelungsgeschichte von Petromyzon Planeri.

Naturkundige Verhandel, van d. hollandsche Maatschappij . d. Weten schappen te Haarlem, II Versam, d. 12, p. 28 (cited from Hammer,

'10). Shipley, A. E. 1887 On some points in the development of Petromyzon fluvia tilia. Quar. Journ. of Mict. Sc, vol. 27, pp. 325-371. Stannius, H. 1854 Handbuch der Anatomie der Wirbelthiere (cited from

Schaffer, '94). Stockard, C. R. 1906 The development of the thyroid gland in Bdellostoma

Stouti. Anat. Anz., Bd. 29. Wheeler 1899 The development of the urogenital organs of the lamprey.

Zool. Jahrb., Bd. 13. Willy, A. 1894 Amphioxus and the ancestry of the vertebrates. New York.



All figures were drawn with the aid of the camera lucida. Higgin's carmine and true blue inks were used to reproduce the colors of the stained sections represented in the colored plate.

1 Lymphoid accumulation in the lateral branchial wall of a 63 mm. larva. Ep., Endodermal epithelium; End., endothelium of blood sinus; pb.b.s., peribranchial blood sinus; trab., connective tissue trabecula in blood sinus. dV oil immersion obj., ocular No. 3.)

2 Portion of a frontal section of a 5 mm. larva showing the epithelial placode in the lateral branchial wall. Ect., ectoderm; End., endoderm; Ales., mesenchyma. (iV oil immer. obj., ocular No. 3.)

3 Epithelial placode in lateral branchial wall of a 9.5 mm. larva. Frontal section. Am.n., amoeboid nuclei; b.c, red blood cells; bn., longitudinal muscle fibers; tm., transverse muscle fibers. (yV oil immer. obj., ocular No. 3.)

4 Epithelial placode in lateral branchial wall of a 15 mm. larva. Frontal section, pb.b.s., peri-branchial blood sinus; s.l., layer of flat cells forming at surface of placode. (^ oil immer. obj., ocular No. 3.)









5 Epipharyngeal ridge containing primitive thymus p'acode from a 15 mm. larva, d.a., dorsal aorta; tnes., mesenchyma containing vascular spaces; pi., primitive thymus placode; 6.m., basement membrane separating placode from mesenchyma. dV oil immer. obj., ocular No. 3.)

6 Primitive thymus placode in 15 mm. larva. Im., lymphocyte; 7nes., mesenchyma; v., vacuoles containing protoplasmic bodies, (jo oil immer. obj., comp. ocular No. 12.)

7 Primitive thymus placode in a 31 mm. larva (ventral surface to the left). ep., nuclei derived from the placode; c.n., connective tissue nuclei; bju.. remains of basement membrane. dV oil immer. obj., ocular No. 3.)








8 Primitive thymus placode in a 44 mm. larva. Ventral surface to the right. (,'3 oil immer. olij., ocular No. 3.)

9 Portion of primitive thymus placode in a 120 mm. larva, b.m., basement membrane; .r., cells which are apparently red blood cells held in the connective tissue. (iV oil immer. obj., ocular No. 3.)






^"^ J ®



y >^












10 A small portion of a primitive thymus placode and the connective tissue in relation to it in a 31 mm. larva. The lettering a to j shows the line of transition from the typical epithelial nucleus (a) to the completed lymphocyte-like cell (j). (fW oil immer. obj., comp. ocular No. 12.)

11 Transforming cells from primitive thymus placode of a 31 mm. larva. Cells g and h (not represented in fig. 10), complete the series of transforming cells shown in figure 10. c.t.c, connective tissue nucleus, {j-i oil immer. obj., comp. ocular No. 3.)

12 Portion of the epithelium of a gill in a 31 mm. larva showing the migration of an epithelial cell («) into a blood vessel. Cells in the left part of the figure (a to /;) show various stages of transformation of the epithelial cell to lymphocyte, ei-y., red blood cells; i, blood cell in vessel in which the nucleus appears to be dividing; j., blood cell in vessel in which there are two nuclei. (xV oil immer. obj., comp. ocular No. 12.)

13 Epithelial cell beginning migration into blood vessel. dV oil inmier. obj., comp. ocular No. 12.)

14 Nucleus in primitive thymus placode showing phagocytic properties. (iV oil immer. obj., comp. ocular No. 12.)

15 Nucleus in primitive thymus placode dividing by simple fission. dV oil immer. obj., comp. ocular No. 12.)

16 Epithelial cell found in a connective tissue space in dorsal part of epipharyngeal ridge. Nucleus beginning to divide. dV oil immer. obj., comp. ocular No. 12.)

17 Cell found in a connective tissue space in dorsal part of epipharyngeal ridge. Two nuclei in the cell which still retains epithelial characters. (jV oil immer. obj., comp. ocular No. 12.)

18 Connective tissue nucleus dividing by simple fission. (iV oil immer. obj., comp. ocular No. 12.)






authors' abstract of this paper issued BY THE BIBLIOGRAPHIC SERVICE.



FroDi Jdfuix Hopkins University and Department of Embryology, Carnegie Institution of Washington


Abundant outgrowth of skeletal muscles of chick embryos can readily be obtained by means of tissue cultures in Locke's solution, with or without the addition of other substances. The characteristic outgrowth can be recognized at a glance and presents features of unusual interest. That such a highly differentiated tissue as cross-striated muscle should grow out so abundantly in Locke's solution is somewhat surprising.

Harrison '10 noticed in cultures of tadpole tissues in frog lymph in a few instances, where the explanted myotome was thin, that the primitive myoblasts differentiated into crossstriated fibers. He did not find, however, that the myoblasts grew out into the culture medium. That amphibian embryonic tissue, where the amount of stored egg yolk supply is considerable should retain the power of differentiation outside the body agrees in general with what we know in regard to the power of self-differentiation exhibited by such tissues when they are transplanted to other parts of the same or different embryos. It indicates that muscle, or better, premuscle tissue can proceed along the path or at least a certain portion of its path of differentiation independently of any specific influences from the other tissues of the embryo. The possibilities of such self -differentiation are already inherent in the cells that are destined to form muscle in the wide open blastopore stage of the frog. For, as pointed out by Lewis, pieces of the rim of the blastopore when transplanted into older embryos continue to differentiate into muscle, notochord and nervous system.

169 .



Sinuhvall ('12) ohiaiiiod growth of nnisele tissue from the ciubiTos of guinea i)igs 2 cm. iu lougtli. He found three main types of eells, (a) elongated si)iii(lle forms, (I)) polygonal, and (c) giant cells. He also observed every gradation between thevse three types of cells. The elongated .spindle forms described by Sundwall evidently corresi)ond to the isolated fibers and myoblasts which are frequently abundant in our cultures. The polygonal and giant cell forms correspond perhap.s to the more irregular nmltinuclear pieces of muscle buds that we sometimes find when the connection between the muscle bud and the explant becomes broken, Sundwall does not seem to have found in his cultures the large muscle buds which are so characteristic of our cultures.

Congdon ('15) observed in plasma cultures the outgrowth of premuscJe cells from the limb buds of seven day chick embryos. The cells were in the form of much elongated spindles. The outgrowth was rather scanty and not nearly so abundant as are the spindle shaped myoblasts in our cultures in Locke's solution.

Levi ('10) has recently described in a few words the fact that he obtained the outgrowth of striated muscle fibers of chick embryos in plasma. He gives the impression that the outgrowth of the skeletal muscle corresponds more or less to that of the heart muscle with which his paper is more especially concerned.

Previous to this M. R. Lewis ('15) briefly described this outgrowth of skeletal muscle in Locke's solution in her paper dealing with the rhythmical contractions exhibited by some of the isolated skeletal muscle fibers found in these cultures.

It is possible that cross striated muscle fibers grow much better in Locke's solution than in other media since among the numerous contributions to tissue culture so little has been said of cross striated muscle by other observers who have confined themselves mostly to plasma diluted with water or Locke's solution as a culture medium, while we have used Locke's solution with or without the addition of other substances. The outgrowths of muscle in Locke's solution present such striking features, and they are so characteristic in shape as well as so


alniiulaut in quantity that they could not well be overlooked if present in plasma cultures.


The explants consist of small pieces of muscle a millimeter or less in diameter taken from the muscles ot the back, wing or leg of chick embryos of se^-en to eleven days incubation. The muscle fibers in the explanted pieces show somewhat varying degrees of differentiation of the cross-striations. The general character of the outgrowth, however, is much the same from pieces of muscle of the above ages, although no two cultures are exactly alike.

In the seven day chick the cross striations are but slightly developed in the myotomic muscles of the back and they are practically not developed at all in the limb muscles. In the nine day chick, however, the cross striations are very apparent in the muscles of both the back and the limbs, especially the muscles of the upper part of the wing and the leg.

The explanted pieces consist for the most part of a matrix of mesenchymal cells in which are embedded the young muscle fibers many of which are cut across at one or both ends. Huber has recently shown that in the adult rabbit muscle the fibers vary greatly in length even in the same fasciculus and probably the same condition holds in the young developing muscles of the chick embryo. We should expect then that the fibers in the piece at the time of transplantation would be of various lengths.

The variations in the size and the length of the fibers in the explanted pieces would explain in part at least the great difference in the length and the size of the outgrowing muscles buds. The medium does not, of course, afford all the necessary substances for growth. The muscle bud is probably derived for the most part from the substance of the old muscle fiber, the medium may furnish some food and the substances derived from the disintegration of cells within the explanted piece may also contribute.



■'■■■ ^¥•:■•' •' A;-M/ 1/. K-- ■ .••:.'^i-.;F'vV\-i-»-"'*-.'8 •>■

k-4w;i:#/;^p:--i ■■■#■■ '..^



In many of the cultures, owing to extensive migration, from the explanted piece, of the mesenchymal cells during the first two days the explanted piece often becomes thinned out so that one can observe the muscle fibers within it more clearly than at the time of the fii'st appearance of the muscle buds. In such cases the direct continuity of the muscle buds with the old fibers is definitely demonstrated m the living cultures and this continuity can also be observed after the culture is fixed and stained. Sections through the explanted piece and the culture in a plane parallel to the cover slip likewise show this continuity of the old muscle fibers and the new muscle buds.


The muscle outgrowths, though somewhat varied in details, have on the whole certain general characteristics that enable one to readily distinguish them from other tissues which grow out from the explanted piece (figs. 1, 2, 3, 4, and 5).

The muscle outgrowths occur either in the form of muscle buds that are continuous with the cut ends of the muscle fibers or as free fibers which wander out into the medium among the mesenchyme cells on the under surface of the cover slip. In most cultures both the attached buds and the free wandering fibers are found in abundance. The muscle buds vary in size from short, slender, pointed processes to large flat masses with many processes at the peripheral end and many nuclei. In practically all of the cultures the outgrowth of the mesenchymal cells begins earlier than that of the muscle fibers and forms a considerable zone of cells about the explanted piece before the muscle buds appear.

The muscle buds usually begin to appear around the edge of the explanted piece at the end of the first day or during the

Fig. 1 Muscle from the leg of a seven day chick embryo cultivated in ? Locke's solution plus i bouillon plus 0.5 per cent dextrose for forty-eight hours. Osmic acid vapor fixation, Benda stain. The long muscle buds radiate out from the explanted piece and are easily distinguishable from the mesenchyme cells. The explanted protoplasmic end of the muscle buds contain many nuclei. The muscle buds show branches and anastqmoses. X 100.


Fig. 2 Somewhat different character of muscle outgrowth from an exphmted piece of the same leg and cultivated in the same way as in figure 1. The enlarged protoplasmic ends are not so abundant. There are many isolated muscle fibers and myoblasts among the mesenchyme cells. X 100.



second day and do not reach their maximal gi'owth until the end of the thuxl or fomlh day. The buds even at the beginning of their growth appear to be less differentiated than are the fibers in the explanted piece from which they grow. This is especially true in the case of muscle buds from fibers where the cross striations are well marked, as in the explanted pieces taken from the older chicks (nine to eleven days).

The bud first appears projecting from the edge of the explanted piece as one or more pointed processes which adhere to the cover sUp. These processes are continually changing in length and size and slowly advance farther and farther out on the coverslip, pulling ))ehind them, as it were, a broad thin expanded mass of nmscle that retains its continuity with the end of one of the muscle fibers within the explanted piece. As the whole mass creeps out farther, nuclei begin to appear in the more proximal part of the mass (fig. 5). As the large flattened protoplasmic mass creeps still farther out on the cover slip, that part of the bud which connects it with the old piece in many cases becomes narrower or more slender and is apparently not so closely attached to the cover slip. The brush like protoplasmic tips with the slender connecting fibers are well shown in figure 1. The protoplasmic tips are evidently the actively migratory part of the bud (figs. 1, 3, 4, 5). The processes are at all times more or less active. They are often long and slender and usually are more numerous at the extreme end of the bud than along its sides.

As the protoplasmic end migrates farther and farther out on the cover slip it apparently exerts more or less of a pull on that part of the bud which connects it with the old j:)iece. It is not imcommon for the resulting slender part to break in two and for both ends to rapidly contract, as though the fiber had been under considerable tension. The entire muscle bud may contract back towards the explanted piece if the protoplasmic end becomes loosened from the cover slip.

There is a marked tendency for anastomoBes and fusion of muscle buds either directly or by branches. The muscle buds from neighboring fibers often fuse near the edge of the explant



«i ' 1

•^ ". - ' • ' ' » /V" «V* ^vv ♦'!'» > \^•♦'.\•^.**A■ Tk -,j>-x: <*» « * / V t <» ^»i\ • o^^i/,".' ■


and continue to grow out in this manner (figs. 1, 3, 4, 8). Buds widely sejmrated at their origin often fuse at some distance from the explanted piece when their direction of outgrowth is such as to bring them into contact with each other (fig. 3).

The muscle buds very often send off branches of different sizes, such branches project at various angles and often unite with other branches or buds. This may result in the formation of more or less complex networks (figs. 3, 4). In some cases the anastomoses are probably without direct continuity of the cytoplasm but in many cases there is undoubted continuity of the cytoplasm (fig. 8).

There is a very curious resemblance between the outgrowths of muscle and nerves in the tissue cultures. The formation of protoplasmic buds with numerous long processes that are continually changing and the migration of this mass away from the explanted tissue pulling out the muscle or the nerve fiber present somewhat similar phenomena. The two differ markedly in one important respect. The nerve outgrowths are entirely without nuclei while the muscle fibers contain many nuclei both in the protoplasmic buds and in the connecting fiber.

Different muscle buds, although they have the same general character, vary considerably in the more detailed appearances. Figures 1 and 2 show long slender outgrowths from a piece of the leg muscle of a seven day chick embryo. The two explanted pieces were from the same leg and planted in the same medium (one-half Locke's plus one-half bouillon plus 0.5 per cent dextrose). In figure 1 the ends are rather broad and fan-shaped while in figure 2 they are narrow or pointed. There are more anastomoses in the former culture than in the latter. In figure 2 there are to be seen many free fibers with one or more nuclei. Th se fibers are very slender, pointed at either end and have the

Fig. 3 Muscle outgrowth from an explanted piece of the leg of a nine day chick embryo cultivated in Locke's solution plus bouillon plus 0.5 per cent dextrose plus 2 per cent distilled water for four days. Osmic acid vapor, iron hematoxylin. The muscle buds have not extended out nearly as far as the mesenchyme. Several large isolated fibers are to be seen, also anastomoses of muscle buds. X 100.



same genera^ direction as the muscle buds from which they have p obabty separated.

The muscle buds from the explanted pieces of the seven day chick embryo are much slenderer than those shown in figure 4 from the leg of an eight day chick embryo. The latter culture was made in Locke's solution plus a little yolk. Whether the differences in the growth are the result of the difTerences in the media is not clear. They do not seem to depend upon the differences in the ages of the chicks fo:; we see in figure 3 the slender type of growth from a nine day chick embryo, somewhat similar to that from explants from the seven day chick embryo.

It is not uncommon for branches to split off completely from the outgrowing buds and to wander freely among the mesenchjaiie cells. 1^'uch isolated fibers may have one or two or several nuclei. Some seem to come directly from the explanted piece. The mononuclear and binuclear fibers are usually long and slender, very pointed at both ends and resemble young myoblasts. Others are somewhat irregular as in figure 13. The multinuclear ones vary somewhat in shape but are UvSually long and slender as in figure 12. Figures 2, 3, and 4 show various types of these free fibers. Some of them represent the entire peripheral end of a muscle bud and are more or less irregular, occasionally branched. They all have a cytoplasmic texture similar to that of the muscle buds and are easily distinguished from the mesenchyme cells by this as well as by their characteristic shape and by the nuclei.

Occasionally the more proxhiial part of the muscle bud becomes spread out into a thin veil-like membrane as in figure 14. Here two neighboring fibers are thus spread out against the cover slip and fused together to form an exceedingly thin membrane. The general appearance of the entire culture was similar to that shown in figure 4. The nuclei are abundant in this veil-like membrane.

Fig. 4 ^Muscle and mesenchyme outgrowth from an explanted piece of the leg of an eight day chick embryo cultivated in Locke's solution plus 0.5 per cent dextrose plus few drops of yolk for two days. The deeply staining muscle buds and smaller isolated fibers are easily distinguisluMl from the mesenchyme. Osmic acid vapor, iron hematoxylin. X 100.





Some of the muscle buds seem to consist o" chains of myoblasts which extend far out into the culture. Such buds tend to break up or give off the individual myoblasts.

The muscle buds do not degenerate in the cultures as a rule until after the mesenchyme cells.


The muscle buds from the eight or nine day chick embryos that arise from the cut ends of cross striated fibers are with very rare exceptions entirely devoid of cross striations. Sections through the explanted piece from a nine day chick embryo show even after two or three days in vitro well marked cross striations in most of the muscle fibers. The muscle fibers within the explanted pieces then do not seem to suffer any loss of differentiation. The area of transition between the cross striated muscle fiber within the explant- and the unstriated muscle bud covers a very short distance in which there is a gradual fading out of the cross-striations. In one or two instances we have seen in fixed specimens indications of cross-striations in the outgrowing muscle buds. Such cross-striations are not well marked antl only occupy a small portion of the bud, usually at the edge of the bud in the part of the liber connecting the protoplasmic end with the old fiber in the explanted piece. These crossstriations were not directly continuous with those in the old fiber. Rarely also cross-striations are seen in the isolated myoblasts but in no cases were they well developed. We are not prepared to state definitely whether such cross striations are

Fig. 5 Muscle l)uds with many nuclei from an explanted piece of the leg of an eight day chick embryo cultivated in Locke's solution plus bouillon plus 0.5 per cent dextrose plus 1 per cent distilled water for two days. Osmic acid vapor, iron hematoxylin. X 100.

P"ig. 6 Protoplasmic ending of muscle bud showing fine striae, spindles and processes. Osmic acid vapor, iron hematoxylin. Leg eight day chick embryo, cultivated in 80 per cent Locke's solution plus 20 per cent bouillon plus 0.5 per cent dextrose for two days. X 525.

Fig. 7 Another protoplasmic ending from the same specimen as the above.

Fig. 8 From the same specimen as above showing fusion of two normal buds.


duo to a r(Hlift'(>nMitiati()ii or are nMuiiaiits of the cross striations of the old libor.s which have been carried out hito the muscle bud. In regenerating luamniahan muscle fibers Waldeyer has pictured isolated groups of cross striations in the young muscle bud which were apparently cari'ied out into the muscle bud and so do not indicate the beginning of redifferentiation. From the work of Waldeyer, Volkman, Ziegler and others it is well known that the regenerating muscle buds in mammals are in the early stages entirely devoid of cross-striations except for such instances as quoted above.

The similarity between the muscle buds in tissue cultures and those pictured for the regeneration of muscle in mammals indicates that we have here in tissue cultures a process essentially the same so far as the initial stages are concerned.

The cytoplasm in the living cultures shows a very fine striation which has in general a longitudinal direction. This gives to the cytoplasm of the muscle bud a very characteristic appearance that distinguishes the muscle buds and the isolated fibers from other cells of the culture. One gets the impression that this cytoplasm has a firmer consistency than that of the mesenchyme cells. The cytoplasm is also somewhat more refractive than that of the mesenchyme. These longitudinal striae are much finer than the so-called sarcostyles or myofibrils seen in fixed normal muscle. The myofibrils are apparently wanting in the muscle buds of the tissue cultures and in the early buds of regenerating muscle.

Cultures fixed in osmic acid show the same characteristic fine longitudinal striations. This is evspecially well seen in the expanded ends of the muscle buds (figs. 6, 11).

In some of the fixed preparations it is not uncommon to find in the muscle buds especially in the enlarged ends, spindleFig. 9 Protoplasmic end of muscle bud from eleven day chick embrj'o cultivated in 90 per cent Locke's solution plus 10 per cent bouillon plus 0.5 per cent dextrose for two days.

Figs. 10 and 11 Protoplasmic ends from muscle bud of the wing of an eight day chick embryo cultivated in Locke's solution plus 0.5 per cent dextrose for three days. Figure 11 shows the striae and spindles.








shaped bodies. They stain dark with iron hematoxyhn and red with Mallory's stain. In favorable specimens, these spindles were seen to fray out in places into fine striae similar to those composing the cytoplasm (figs. 6, 7, 8, 10, 11). Such spindles have not been observed in the living buds.

Specimens fixed with acetic acid combinations and especially with acetic acid vapor give pictures of fibrils and other structures within the muscle buds, which are not present in the living cultures. Such methods are of course entirely useless so far as the study of the optical structure of the cytoplasm is concerned. The fibrils 'brought out' by the acetic acid are especially marked in that part of the muscle bud connecting the amoeboid end with the expl anted piece. This portion of the muscle bud is evidently under considerable tension as we have already noted. It is probable that coagulation of the cytoplasm when in a state of stress or pull takes place in lines parallel to this stress and hence the formation of the longitudinal fibers. Under such conditions the fibers brought to view are no indication whatsoever of their being differentiated structures in the cytoplasm.

The mitochondria are especially abundant in the muscle buds and are arranged longitudinally between the fine longitudinal striae. They are smaller than those in the mesenchyme cells and in the healthy fibers do not show the same irregular arrangement. It is rather difficult to make them out in the living buds. With Janus green, however, they usually appear as strings of minute granules of varying lengths and sometimes as long threads which seem to taper off at either end to the limits of visibility. The mitochondria are best seen in the enlarged protoplasmic end of the buds and undoubtedly contribute to the appearance of longitudinal striation.

Fig. 12 Isolated muscle fiber from the same culture as the above.

Fig. 13 Isolated myoblast from the cultures from the wing of an eight day chick embryo cultivated two days in 80 per cent Locke's solution plus 20 per cent bouillon; plus 5 per cent dextrose. X 525.

Fig. 14 Veil-like spreading out of the stem of a muscle bud from a two day culture of the muscle from the wing of an eight day chick. Locke's solution plus few drops yolk plus Co per cent dextrose. X 455.



Aside from the mitochondrial inclusions the cytoplasm contains varying numbers of neutral red granules. These are minute and not very abundant, and are usually situated in the neighborhood of the nuclei.


The nuclei appear in the young muscle buds soon after the protoplasmic ends begin to project from the explanted piece. They gradually increase in number as the bud increases in length and size. There is usually a large group of nuclei in the expanded end. They occupy the more proximal part of this expansion while the more distal part is usually free from nuclei. The narrow part of the muscle bud connecting the protoplasmic end with the explanted piece has a varying number of nuclei scattered along it. The isolated myoblasts and fibers contain varying numbers of nuclei from one to many. We have examined repeatedly both living and fixed cultures for indications of nuclear division but only in a few instances have we seen mitotic divisions and those occurred in the mononuclear myoblasts that were free in the culture. When the nuclei of the muscle buds were studied the condition of the mesenchyme in regard to the frequency of cell division was usually noted and it was not uncommon to see three or four mitotic figures in the mesenchjone cells in the neighborhood of the muscle buds in one field of the microscope. In spite of the fact that we have very little direct evidence of nuclear division in the muscle buds it seems probable that nuclear division does take place. Some muscle buds have thirty or forty or more nuclei and they must either have arisen by division from a few or more that came out from the old piece or have all migrated out from the old fiber as the muscle bud grew out from it on to the cover slip. The indirect evidence in favor of nuclear division is revealed through the staining of fixed specimens. In such specimens, stained either with iron hematoxylin or with Ehrlich's hematoxylin and eosin, it is seen that the nuclei vary considerably in their staining reaction. Some are darkly stained, others rather lightly, and


this holds even among the nuclei that lie side by side in the same group. We have often noticed similar differences among the nuclei of mesenchyme cells when active mitotic division is taking place. In fact everyone who has studied embryonic material has probably noted such differences in the staining reactions of nuclei. It is especially well marked, for example, in the cells of the neural tubes of young amphibian embryos where active mitotic division is taking place. We have been able to demonstrate in our cultures that the nuclei of the young daughter cells of the mesenchyme always stain deeper than the nuclei of the resting cell. This ability of the daughter nuclei to stain more deeply lasts for an hour or two after the mitotic division. If mitosis were taking place to any great extent in the muscle buds we should probably have observed it especially iii the expanded end of the bud. Yet here as well as elsewhere in the muscle bud the stainable differences in the nuclei are found in abundance. Of course it may be that the nuclei undergo mitotic division in the old piece out of range of direct observation in the living. On the other hand, there is, of course, the possibiUty of direct division. Direct division seems to be extremely rare in our cultures and Macklin, after an extensive series of observations, was able to observe but one case of direct division of the nucleus in the mesenchyme cells, and that without division of the cytoplasm. We have not observed direct division of muscle nuclei and have no data on the staining reaction of nuclei after direct division.

The observations on the nuclei of muscle buds in the living are much more difficult than are those upon the nuclei of the mesenchymal cells and for the present at least many questions in regard to the origin of these nuclei must be left unsettled. It is often stated that direct as well as indirect division of the nuclei takes place in the regeneration of muscle in amphibia and mammals. Such statements are based not on direct observation of the living but on fixed preparations. It is evident from our studies on the living cells in tissue cultures that such observations on fixed and stained material in regard to direct division are no indication of what actually occurs in the living. Many


fixed specimens seem to indicate that the nuclei show all stages in the process of direct division while observations on similar cultures in the living fail to give evidence of a direct division.


The muscle buds from the explanted pieces of thie older embryos (nine to eleven days), which arise from the cut ends of the cross-striated fibers, appear to be less differentiated or more embryonic in type than normal muscle fibers of the same age. A process of dedifferentiation has evidentlj^ occurred in the formation of these muscle buds from the old fibers. Is this a true reversibihty or merely a breakdown with elimination or absorption of some of the more differentiated parts of the cytoplasm? Such unstriated buds are still capable of contraction and when portions of them become separated off they may undergo rhythmical contractions. It is then not necessarily loss of function which determined this dedifferentiation. Contractions occur however rather rarely. The fibers in the old piece are of course entirely severed from all nervous connections and there is no indication that they contract yet they retain their crossstriations.

This process of dedifferentiation or a return to a more embryonic condition probably underlies all types of regeneration. We doubt if there is ever any regeneration of differentiated tissue without a preliminary return of the cells involved to a more embryonic condition. In regeneration this preliminary stage of dedifferentiation prepares the way for growth and redifferentiation. The dedifferentiation in regeneration does not necessarily proceed to the extent in which the cells of the various tissues return to a common embryonic type, such as Champy maintains happens to practically all cells in tissue cultures. As we have seen this process of dedifferentiation does not proceed in our cultures to such an extent as to render the muscle cells indistinguishable from other types of cells. Prolonged cultivation might result in a return to a still more embryonic type of the outgrowing muscle tissue.


Champy, in a sei'ies of articles, has maintained that most of the cells in the body dedifferentiate in tissue cultures. They return, he claims, to a completely indifferent type of cell that no longer shows the imprint of its origin. In explants from late fetal stages he finds that cells of the kidney tubules, of the thyroid, of the parotid and of the submaxillary glands, of the smooth muscle, of the mesenchyme, etc. dedifferentiate into an indifferent embryonic type indistinguishable from each other. This dedifferentiation, he claims, is associated with the phenomena of cell division.

The rapidity of dedifferentiation is a function of the rapidity of the cell-division. Furthermore, according to Champy, all cells differentiated for a special function lose or tend to lose during mitosis, their characteristic function. In the animal organism they recover immediately after the telephase, since they are subject to the same functional excitation as before division. In the body, function does not maintain the differentiation but the function provokes and creates anew the differentiation after each mitosis. Champy's ideas are based in part on a law formulated by Prenant that a ceU during mitosis does not secrete. Among the tissues which do not dedifferentiate he finds the liver cells of the rabbit near term, the true gray substance of the central nervous system and striated muscle. Such tissues he finds do not grow out into his cultures and he reasons that since they do not grow and vegetate they are not susceptible of dedifferentiation. Maximow, on the other hand, takes exception to Champy. He finds that fibroblasts continue indefinitely as such through many generations of the culture and for this reason he calls them 'immortal* cells. Maximow also finds that the endothelial cells of blood vessels and of lymphatics as well as the mesothelial cells lining the serous ca\dties change into fibroblasts and become indistinguishable from those of connective tissue origin. This dedifferentiation is according to Maximow only apparent since he considers the endotheUum of blood vessels and lymphatics and the serosa but flattened-out fibroblasts.


The foregoing conclusions of Champy and the less general conclusions of Maximow in regard to the fate of endothelium and mesothelium are certainly in need of further substantiation. During the process of regeneration, in vertebrates at least, the dedifferentiation never proceeds to an indifferent stage; muscle is regenerated from muscle, nervous tissue from nervous system, bone or cartilage from bone or cartilage, ectoderm from ectoderm, etc.

In prolonged cultivation by means of frequent retransplantation of the culture such as was carried on first by Carrel, the fibroblasts seem to be the only cells which survive so that finally they are obtained in pure cultures. It is probably that both Champy and Maximow failed to realize that it is a question of the survival of the fittest and not complete dedifferentiation which is responsible for the appearance in cultures that have been carried on for many generations of but a single type of cell. Then too we must bear in mind the fact that even in the early stages of cultivation there is often great difficulty in distinguishing the various types of cells.

We are more especially concerned in this prehminary and essential process of dedifferentiation. That it should take place in a minute isolated piece of muscle outside the body, in an artificial medium, is of great significance. It makes possible an analysis of the process in a way that was not reahzable in the living organism. Attempts to get growth and regeneration from small pieces of muscle (one-half to one centimeter in diameter) in vivo have failed. Such pieces even when transplanted into muscle itself always degenerate (Volkman). It may be that pieces as small as those used in tissue cultures would have continued to live in vivo.

The nature of the changes in the organization of the cells of tissue cultures undoubtedly depends in part on the tissue explanted, in part on the age of the embryos or animal employed and in part on the culture medium and the peculiar conditions to which the cultures are subjected. Tissues of late fetal stages or of stages subsequent to birth in which differentiation is complete could remain either stationary or dedifferentiate, while


tissues of early embryonic stages might continue to differentiate, or remain stationary or dedifferentiate. In either case, pathological changes and degeneration may supervene. We know that the anlage of many tissues of amphibian embryos (central nervous system, the eye, otic vesicle, notochord, voluntary muscle, heart, etc.), when transplanted into strange environment of the same or another embryo will continue to differentiate. There is a period during which many young embryonic tissues are self-differentiating. It is not surprising then that Harrison should have obtained an outgrowth of the axis cylinders from young nerve cells and a differentiation of cross-striated muscle from young embryonic myoblasts in tissue cultures. On the other hand, it is perfectly evident that in older embrj^os (chick embryos of nine days, for example) cross-striated muscle as it grows out into the culture loses its cross-striations and assumes a more embryonic condition. The portion of the fiber which remains in the explanted piece retains, however, its crossstriations.

The muscle buds found in tissue cultures resemble in many ways the early stages of the regeneration of muscle in the higher mammals after injury or rupture of the muscle fibers as described by Waldeyer ('63) and Volkmann ('83) and Ziegler ('98). In mammals the buds which grow out from the cut ends of the fibers are more or less homogeneous and unstriated. There are often lateral buds as well. These buds elongate and extend between the connective tissue cells filling in the wound. Such buds are crowded with nuclei which are supposed to increase in number for the most part by direct division. Mitoses are also found. There are also found free myoblasts, long spindle cells with one or more nuclei, which come from the old piece. There is also a disappearance of the cross-striations in the old fibers near the cut ends. The process of regeneration is slow, extending over weeks. A redifferentiation occurs in these buds with the formation of longitudinal and cross-striations so that finally they come to resemble the old fibers. The free myoblasts also differentiate in a manner similar to that of embryonic myoblasts.


The experiments on the regeneration of muscle in amphibia also show that there is a return first to ah embryonic type of muscle cell followed by a redifferentiation in a manner similar to the differentiation of embryonic cells. These myoblasts come from the injured muscle fibers (Fraisse, Barfurth and Towle). According to Towle, the outer bundles of the cut nmscle disintegrate leaving nuclei surrounded by cytoplasm. The nuclei increase in number by amitosis. Some of the cells thus formed later divide by mitosis and from them are formed new muscle fibers. The inner bundles of the muscle do not disintegrate but spht longitudinally into myoblasts which later differentiate into muscle. Barfurth finds that in the very young larvae of Siredon, terminal and lateral buds grow from the injured fibers. The outgrowths contain nuclei and form sarcoblasts (myoblasts) and these differentiate into muscle fibers in the same way as do the myoblasts of the normal embryo. In the older larvae of the frog and in mature animals, there occurs a degeneration of the muscle with the accumulation of nuclei and the formation of giant cells. He also finds that there is a splitting of old fibers into myoblasts as well as sarcoblast-like outgrowths which form myoblasts which later become new muscle fibers.

The initial stages in the process of regeneration of muscle in mammals and amphibia are in many respects very much like the behavior of muscle in our cultures. In both there is (1) a formation of young myoblasts, a return to a more embryonic condition; (2) the formation of protoplasmic buds which grow out from the ends of the old fibers. Such buds contain many nuclei and lack cross-striations.

The factors involved in the formation of these muscle buds are probably the same in the tissue cultures and in regeneration and consequently are common to each. We can eliminate at the outset then various possible factors that are present where muscle buds are formed in the regeneration of muscle in the experimental animals, such as the influence of the nervous system, of substances brought by the blood or body fluids or of other influences that might come from the organism itself.


The formation of the muscle buds seems to be inherent in the muscle fiber itself and becomes manifested when the fiber is cut across or is injured. The peculiar form which they take as long narrow fibers is to be attributed to the specific complex of materials which compose the muscle substance and to the dynamic processes which occur there.

Although the initial stages are much the same in cultures and in regeneration, it is not to be expected even after prolonged cultivation in vitro there will be a redifferentiation of the muscle buds. Especially will this be true, if, as Morgan suggests, the same factors which affect the normal growth and differentiation of the embryo affect in the same way the regeneration of a part. In the healing of wounds a similar process of dedifferentiation followed b} a redifferentiation is involved.

The anastomoses between muscle buds suggests that in the normal muscle there may yet be found a syncytial hke condition even in the adult. It lends some support to Huber's suggestion that muscle may be syncytial in character which suggestion he makes in spite of the fact that he has succeeded in isolating fibers of various lengths. On the other hand, it may be that the pecuhar conditions found in tissue cultures produce conditions not normally present. We have in the past often observed anastomoses of nerve axones in cultures of sympathetic fibers. Even if such phenomena are the result of peculiarities of cultures that are not present in the hving organism, they serve to show us at least some of the potentialities of muscle and nerve protoplasm.



Adami 1908 Principles of Pathology, vol. 1.

Caruel, A. 1914 Present condition of a strain of connective tissue 28 months old. Jour. Expt. Med., vol. 20, 1914.

Champy, C. 1912 Sur les phenomenes cytologiques qui s'observent dans les tissues cultives en dehors de I'organisme. (tissue epitheliaux et glandulaires). Comptes rendus de la Soc. d. Biol., T. 72, p. 987.

1913 La dedifferentiation des tissues cultives en dehors de I'organisme. Bibliogr. Anat., T. 22.

1914 Quelques resultets de la methode de culture des tissues. I. Generalities. II. Le muscle tissue. Arch. Zool. Exper. et Gen., T. 53.

1914 Notes de biologie cytologique. Quelques resultats de la methode de culture de tissues. III. Le rein. Arch, de Zool. Exp., T. 54.

CoNGDON, E. D. 1915 The identification of tissues in artificial cultures. Anat. Rec, vol. 9.

Fraisse, p. 1885 Die Regeneration von Geweben und Organen bei den Weibelthieren, besonders bei Amphibien und Reptilien. Berlin und Kassel.

Harrison, R. G. 1910 The outgrowth of the nerve fiber as a mode of protoplasmic movement. Jour. Exp. Zool., vol. 9.

HuBER, G. C. 1916 On the form and arrangement in fasciculi of striated voluntary muscle fibers. Anat. Rec, vol. 11.

Levi, G. 1916 Migrazione di elementi specifici difTerenziati in colture di miocardio e di muscoli scheletrici. Archivio per le scienze Medichi., Ann. 40.

Lewis, M. R. 1915 Rhythmical contractions of the skeletal muscle tissue observed in tissue cultures. Am. Jour. Physiol., vol. 38.

Macklin, C. C. 1916 Binucleate cells in tissue cultures. Contributions to Embryology, No. 13. Publication 224, of the Carnegie Institution of Washington.

Maximow, a. a. 1916 The cultivation of connective tissue of adult mammals in vitro. Arch. Russes d'Anat., d'Hist. et d'Embry., T. 1.

Morgan, T. H. 1901 Regeneration.

SuNDWALL, J. 1912 Tissue proliferation in plasma medium. Bull. U. S. Hyg. Lab. and Mar. Hosp., vol. 81.

TowLE, E. W. 1901 On muscle regeneration in the limbs of Plethedon. Biol. Bull., vol. 2.

VoLKMANN, R. 1893 Ueber die Regeneration des quergestreiften Muskelgeweckes bein Menschen und Sfiugethier. Beitrage zur path. Anat. u. Path., Bd. 12.

Waldeyer, W. 1865 LTeber die Veriinderungen der quergestreiften Muskeln bei der Entziindung und den Typhusprozess sowie liber die Regeneration derselben nach Substanzdefecten. Arch, fur Path. Anat., Bd. 34.

ZiEGLER, E. 1898 Lehrbuch der allgemeinen Pathologie und. d. Path. Anat.. Bd. 1.





Institute of Anatomy, University of Minnesota


Henneberg ('00) made a careful study of the development of the mammary glands in the albino rat from the earliest appearance of the glands through the conditions found in sixteen day fetuses. Also the postnatal (birth to ten weeks) development of these glands has been investigated (Myers, '16). Heretofore the developmental conditions between sixteen day fetuses and newborn rats have presented a gap in our knowledge of the mammary glands. The object of the present investigation is to fill up this gap, thus completing the history of the mammary glands in the albino rat (Mus norvegicus albinus) to ten weeks after birth. An abstract of the results has already been published (Myers, 17).


No attempt is made to review all the literature pertaining to the development of the mammary gland, which is thoroughly discussed in the works of Bonnet ('97), Brouha ('05), Bresslau ('10) and Schil ('12). Henneberg's work ('00) in the early development of the mammary glands in the albino rat is here briefly reviewed, however, since the earlier stages must be kept in mind to make clear their relations wdth the later foetal stages described in the present paper.

Henneberg ('00) found in an albino rat embryo of eleven days, in the region of the dorsal limiting furrow (on only one side), some cubical cells in a single layer representing the anlage of the mammary streak. In an embryo of twelve days a mam 195

196 J. A. MYERS

mary streak is present on each side. Each streak consists of a single layer of cubical epithelium. The breadth of the streak has increased and now extends from a few cells dorsal to the dorsal limiting furrow ventrally to cover nearly half of the parietal zone. Its cephalic and caudal ends blend with the cubical epithelium of the limb anlages.

At twelve days and thirteen hours, the cells of the mammary streak are larger and in the region of the dorsal limiting furrow a second layer of cells is beginning to appear superficial to the cubical cells. Immediately beneath the mammary streak the mesenchymal cells have condensed. The mammary streak shows two distinct cell layers in embryos of thirteen days and one hour. The superficial layer — stratum corneum — consists of flat cells with oval nuclei with their long axes parallel to the surface. The deep layer — stratum mucosum — is composed of large round or cubical to cylindrical cells with oblong nuclei. The streak is separated from the mesenchyma by a distinct light line-^the basement membrane.

Henneberg found the first appearance of the mammary line in a rat embryo of thirteen days and fourteen hours. At this stage it is produced by a thickening of parts of the mammary streak. In some places a part of the mammary streak is converted into the mammary line by the appearance of a third layer of round cells between the superficial and deep layers. In other places the cells have slightly thickened thus producing the first appearance of the mammary line without the addition of a third layer. In other embryos of the same age the mammary line in the thoracic region is three to four layers of cells thick and its greatest breadth shows twelve to fourteen layers of cells. It disappears a short distance cephalad to the anterior extremity. In the inguinal region the line is still very indistinct and requires special technique for its study. In some embryos a complete interruption exists between the region of the future thoracic glands and the • adbominal gland. This is the first intimation of the future interspace between the glands of the thoracic region and those of the abdominal and inguinal regions. From this stage, Henneberg designated the cephalic part of the line


as the pectoral portion and the caudal part as the abdominal portion.

In rat embryos of fourteen days Henneberg found that the cephalic end of the mammary line has been transformed into a structure about the shape of a biconvex lens. This is the earliest appearance of the first pectoral mammary hillock. In other embryos of the same age the second and third pectoral and the abdominal hillocks are beginning to appear. The greater convexity of each hillock lies embedded in the mesenchyma. The remaining parts of the mammary streak and line represented by the space between the hillocks are beginning to atrophy. At this stage the mammary line for the inguinal glands resembles in structure the line for the pectoral and abdominal glands in the thirteen day and fourteen hour stage.

Henneberg found in fifteen day rat embryos that the mammary gland anlages are no longer elevated above the surface but that their deep surfaces have pressed deeper into the mesenchyma thus presenting the 'mammary point' stage. At this stage the inguinal glands are still somewhat retarded in their development. At sixteen days Henneberg states that the mammary gland anlages correspond to the club-shaped stage which Rein ('82) found in rabbit embryos. Henneberg did not investigate the later stages in the rat.


The fetuses for the present work were collected in the following manner. Adult males and females were placed in the same cage from six o'clock in the evening until six o'clock in the morning. As found by Danforth ('16) in case of mice, better results were obtained when the females were placed in the cage which the males occupy permanently. The females were then returned to their respective cages. In all cases of pregnancy semination was dated at the ninth hour after the females were placed in the cages with the males. The possibility of error in the age of the fetuses is plainly obvious. However, the error could only be a matter of a few hours. Sobotta and Burckhard ('11) estimated that spermatozoa of the albino rat do not

198 J. A. MYERS

live more than nine or ten hours in the reproductive tract of the female.

During 1914-1915 a large immber of observations were made on females with the hope of finding a definite way of knowing just when the animals is in heat or when copulation has taken place. No definite gelatinous plug was found closing the vaginal orifice after copulation as Sobotta ('95) observed in white mice. A yellow and somewhat viscid vaginal secretion appears at rather regular intervals. This secretion usually makes its first appearance shortly after the opening of the vagina which occurs about the eighth week. In young females it occurs thereafter at quite irregular intervals but later it may be seen about every fifth to eighth day. No definite relation has yet been established between the appearance of the vaginal secretion and the time of insemination. However, it was noticed that many of the females became pregnant while the secretion was present. The origin of the vaginal secretion and its relation to ovulation is still being studied with the hope of obtaining definite knowledge as to the time of ovulation in the white rat.

Some of the fetuses were fixed in Zenker's fluid, others in 10 per cent formalin. In the earlier (fifteen day and nine hours, sixteen day and twelve hours, and seventeen day and two hours) stages several fetuses were cut for each stage described while in the later (eighteen days and nine hours, nineteen days and six hours, and twenty days and six hours) stages only one fetus was entirely sectioned and merely the skin containing the mammary glands from several other individuals was sectioned. The mammary glands of other fetuses were studied macroscopically. In all 30 individuals were examined. A part of the material was cut at 5 M or 7 M and stained with iron hematoxylin; the remainder was cut at 10 /i and stained \vith alum hematoxylin and eosin or with Mallory's connective tissue stain. Weigert's elastic tissue stain was also applied to some of the fetuses of the latest stages. For a study of the varieties of white blood corpuscles Dominici's combination stain was used.

A few dissections and observations proved that in the late fetal stages the sex could be determined by the relative ano-genital


distance as described by Jackson ('12) in determining the sex of the newborn. In the earher fetal stages the sex was determined by studying the developing reproductive organs.

The' wax reconstructions w^ere made according to Born's method.


Henneberg states that in fifteen day and fourteen hour embryos the six pairs of mammary glands occupy their definitive positions. Since Henneberg made only a macroscopic study of the glands at this stage, a further account is here given of the condition found in embryos of nearly the same age.

Fifteen days. On the surface of the skin at this stage (fifteen days and nine hours) is a small eminence (fig. 7) over each developing gland. Such eminences are very prominent in fresh preparations. A cross section through a gland (fig. 1) shows that the epidermis in the neighborhood of the gland is composed of only two layers, a superficial layer (periderm) of flattened cells with their long axes parallel to the surface, and a deeper layer (stratum germinativum or Malpighian layer) of round or cubical cells. The nuclei of the latter layers are located toward the free end of the cells. The basal ends of the cells have a quite clear appearance and rest on a definite basement membrane

(% 1).

The basement membrane dips down into the underlying mesenchyma to surround the spheroidal mass of epithelial cells forming the gland anlage. Likewise the stratum germinativum of the epidermis passes deep around the same circular mass of cells and forms the basal layer of the mass. The cells of the spheroidal mass are differentiated and arranged so that they possess a characteristic appearance. The cells of the basal layer appear much more elongated than those in the stratum germinativum of the adjacent epidermis. The cells occupying the center are irregular in shape and closely packed.

Superficially the gland anlage projects somewhat producing the eminence \dsible from the surface. Around its deep surface the mesenchyma is condensed. The mesenchymal cells lying





.'SB^B*' c'^St) •S'^^^'^'*^


^ «s>


djc^*^ '^%'^u'^ •

■v*-^ '^(P^

^ " '!^

, >>^


f '



nearest the developing mammary gland are somewhat elongated and arranged in two or thi-ee very regular layers concentrically placed (fig. 1). Outside of the concentric layers, the condensed mesenchymal cells seem to have no definite arrangement. In the condensed mesenchyma is seen an occasional small blood vessel containing nucleated red blood corpuscles.

Wax reconstructions (fig. 10) show that the differentiated mass of cells which appears circular in cross section, forms an oblong ellipsoidal body which is attached to the epidermis by a very short, constricted neck (nk).

Sixteen days. In fetuses of sixteen days and twelve hours the mammary eminences still appear on the surface of the skin as slightly elevated areas which in fresh preparations have a somewhat lighter appearance than the surrounding tissue.

In microscopic sections the epidermis presents the two distinct layers of cells found in the preceding stage. In addition an intermediate layer of cells has appeared in some parts of the skin. In some places the epidermis is slightly thickened to f©rm hair anlages, but in no case were such anlages observed in the epidermis adjacent to the mammary gland anlages. The socalled basement membrane appears as a homogeneous band immediately below the stratum germinativum. Just beneath the basement membrane the mesenchymal cells are densely placed thus forming a fairly definite layer. Immediately beneath this layer the mesenchymal cells are less numerous and apparently have no regular arrangement. Mitotic figures are very com Fig. 1 Drawing of a section through the right second thoracic mammary gland region of an albino rat fetus of fifteen days and nine hours. X 300. Zenker's fixation; hematoxylin-eosin stain. Drawn with the aid of a camera lucida. b.m., basement membrane; cm., condensed mesenchyma; e.s., eminence (mammary hillock) on surface of skin produced by mammary gland anlage; m., loose, irregularly arranged mesenchyma; m.a., mammary gland anlage; p., periderm; s.g., stratvma germinativum.

Fig. 2 Drawing of a section through the left second thoracic developing mammary gland of a female albino rat fetus of eighteen days and nine hours. X 300. Zenker's fixation; hematoxylin-eosin stain. Drawn with the aid of a camera lucida. b.m., basement membrane; cm., condensed mesenchyma; m.p.. early appearance of mammary pit; p.d., deep portion of mammary gland anlajj;!' (primary duct); s.m.a., superficial part of mammary anlage, becoming cornificd.


202 J. A. MYERS

moil ill the mesenchyma immediately surrounding the gland. An occasional small blood vessel is seen coursing toward the mammary gland area.

The mammary gland anlages show about the same stage of development as in the preceding stage. A study of all six pairs shows that the inguinal mammary glands are slightly behind the others in their stage of development.

Seventeen days. At seventeen days and two hours the eminences described in the previous stages have disappeared. The gland areas instead appear as slight depressions or pits on the surface of the skin. These mammary pits represent the point of ingrowth of the epithelium. The epidermis is slightly thicker than in the preceding stage and in the regions of the mammary glands presents a very definite basement membrane. The gland anlages now measure only about 0.05 mm. in length.

Eighteen days. Fresh preparations, sections, and wax reconstructions from fetuses of eighteen days and nine hours show a definite mammary pit on the surface of the epidermis over each future nipple area (figs. 2 and 8, n.p.). In cross section the stratum germinativum is now depressed so as to form a shallow funnel-shaped outline. The mouth of the funnel is directed toward the surface and is partly filled with epithelial cells which show traces of cornification and desquamation. Intercellular vacuoles are also being formed. The outlet of the funnel extends into the corium and becomes continuous with the anlage of the primary mammary duct.

At this stage the gland anlage, which in the earlier stages was an oblong, ellipsoidal mass of epithelial cells, has increased in length. Its deep part now becomes the anlage of the primary duct, while its superficial portion is undergoing vacuolization, cornification, and desquamation, thus forming the pit superficial to the primary duct. The end of the primary duct anlage directly beneath the surface pit is attached to the epidermis and throughout this paper will be designated as the attached end. The opposite end of the anlage is unattached and throughout this paper will be known as the free end. The stratum germinativum of the adjacent epidermis continues over the mammary anlage as its future primary peripheral layer of cells. The


primary duct anlage is roughly L-shaped with its attached end perpendicular and its free end parallel to the surface (figs. 2 and 8). The anlage in elongating has pushed ahead of it the above mentioned layers of condensed mesenchyma representing the corium and tela subcutanea. These layers now completely surround the free part of the anlage. In the first thoracic gland the free end of the anlage is directed cephalad. In the second inguinal gland, the free end points caudad. Likewise the free end of each of the remaining ducts is directed toward the position which the future duct and its branches will occupy.

The anlages of the ducts are longer than in the seventeen day and two hour stage. In one of the first thoracic glands of one fetus and in one of the abdominal glands of another fetus the primary duct presents two secondary ducts (fig. 17, s.d.). All other glands observed at this stage possess a single undivided primary duct.

When seen in cross section at this stage, the primary duct of most of the glands possesses a basal layer of cuboidal cells with large oval nuclei. The basal ends of the cells rest on a somewhat indistinct basement membrane while the opposite ends are directed toward the center of the duct. The center of the duct is filled with cells of irregular shape. Somewhat nearer the free than the attached end of some of the ducts the cells occupying the center of the duct show a tendency toward separation from each other. In other ducts some of the central cells have entirely separated, thus producing small cavities or lacunae, the first appearance of a very indefinite lumen (fig. 4). Such a condition obtains in many of the thoracic and abdottiiinal glands examined, but is very rare in the inguinal glands of this stage. It is interesting to note that in the thoracic and abdominal glands which have already developed secondary ducts, only one of these ducts shows a slight indication of a lumen. The mesenchymal cells of the corium and tela subcutanea are somewhat condensed around the ducts. Those nearest the ducts are much elongated and are concentrically arranged.

In one abdominal gland about h^lf way between the outlet and mouth of the funnel the cells of the stratum germinativum have slightly elongated thus forming a low ridge which projects

204 J. A. MYERS

into the subjacent corium. The ridge extends entirely around the funnel and is the anlage of the epithelial hood, which was described in the postnatal stages of the albino rat (Myers '16).

Nineteen days. The funnel-shaped epithelial area corresponding to the mammary pit at nineteen days and six hours contains some cornified epithelium. This is apparently being cast off by the process of desquamation, thus deepening the mammary pit superficial to the attached primary duct.

The primary mammary ducts have made a rapid growth and present secondary ducts in all glands, while in most glands examined the secondary ducts present tertiary ducts. The two inguinal glands present lumina in about the same stage of development as was described in the thoracic and abdominal glands in the eighteen day and nine hour stage. The rudimentary lumina in all glands are slightly further developed toward the free ends of the ducts but are by no means confined to the free ends. Many of the cells near the developing lumina are undergoing mitotic divisions. There is no pyknosis or other eiddence of cell degeneration.

The anlage of the epithelial hood is composed of elongated cells of the stratum germinativum, but a second layer of cells deep to the layer described as forming a low ridge in the preceding stage is beginning to appear. The ridge now projects deeper into the subjacent corium. Numerous mitotic figures are seen in the epithelial cells in the region of the free edge of the hood.

The developing hair follicles have grown more deeply into the corium than those described in a preceding stage. Ordinarily the follicles are located a considerable distance from the mammary pits. No follicles were observed in the mammary pits.

Twenty days. At twenty days and six hours well defined mammary pits in the epidermis represent (as in the preceding stage) the regions of the mammary glands. Wax reconstructions, how^ever, show that at the bottom of each pit there is a rounded elevated portion of the epidermis (fig. 9, n.a.). This elevated part is the anlage of the nipple. In the preceding stage as noted the depression or funnel was partly filled with cells, which became cornified as age advanced, thus giving the integument


over the mammary glands a thickened appearance. Later the cornified cells were cast out and thus the funnel corresponding to the mammary pit was deepened. The anlage of the nipple in the present stage seems to have pushed from the bottom of the mammary pit toward the surface leaving a surrounding furrow or sulcus (figs. 3 and 9, s.). The superficial part of the epidermis over the nipple anlage now appears no thicker than that in adjacent regions.

The anlage of the epithelial hood has grown more deeply into the corium now encroaching upon the tela subcutanea. Cornified epithelial cells occupy the space between the inner and outer surfaces of the hood.

\'\Tien the stratum germinativum of the epidermis is traced toward the region of the mammary gland it is seem to pass deeply and form the outer surface of the epithehal hood. It then covers the free edge of the hood and turning back forms the inner surface of the hood. Next it covers the deep surface of the nipple anlage. Throughout its extent in the mammary region the cells of the stratum germinativum rest on a basement membrane (fig. 3).

The corium mthin the epithelial hood is composed of connective tissue cells the processes of which take a deep blue stain when treated with Mallory's connective tissue stain. Small blood vessels and nerves are also included. From the surface of the nipple anlage the primary duct is seen coursing through the corium in the center of the hood on its way to the subcutaneous tela w^here it turns at right angles after which it lies parallel with the surface of the integument. Soon after reaching the subcutaneous tela and turning at right angles the duct divides into secondary ducts each of which in turn divides into tertiary ducts. Quaternary ducts are beginning to arise from the tertiary ducts (fig. 12). The terminal ducts present small knoblike enlargements or end-buds. Not every gland observed at this stage presents all of the above mentioned branches. For example in the second inguinal gland of one specimen the primary duct has divided into two secondary ducts which remain undivided.



The time of formation of the lumen evidently is subject to considerable variation. While its first appearance was observed in eighteen day and nine hour and nineteen day and six hour

ep. m.

Fig 3 Drawn from a section through right first thoracic developing mammary gland of a female albino rat fetus of twenty days and six hours. X 3C0. Zenker's fixation; hematoxylin-eosin stain. Drawn with the aid of a camera lucida. c, irregularly arranged developing connective tissue cells ■,c.t., deve oping connective tissue forming sheath around duct ;, epithelial ingrowth or hood; n.a., nipple anlage; p.d., primary duct ; .s., sulcus surrounding nipple anlage.


fetuses there are still systems of ducts at twenty days and six hours which show absolutely no trace of a lumen. In other glands of this stage the lumina are much larger than in the preceding stages (fig. 5). When present, the lumina are better developed in the free ends of the system of ducts, i.e., in the terminal ducts and the ones from which they arise; however, quite frequently traces of lumina are observed in the primary and secondary ducts. In no part of any system of glands observed is there a definitive lumen present. The walls of all are irregular, but have quite sharp boundaries. In no case are degenerating cells found within the lumen.

Figures 4 and 5 show that the first indication of a lumen is the appearance of a few independent lacunae. In cross section of the ducts such lacunae are usually seen located near the center of the developing ducts ; however, they are not uncommonly found near the periphery, at the central ends of the peripheral layer of cells. When traced longitudinally any individual lacuna is found to extend only a very short distance; but in serial sections other lacunae are found forming more or less definite rows extending along the ducts. In some glands the lacunae are present from the end-buds well into the primary ducts.

The lacunae later increase in size and apparently flow together, thus forming the lumina found in some individuals of this stage (fig. 5, l) . The lumina at this stage are never continuous throughout the system of ducts. But a lumen may extend throughout a terminal duct, then with an interruption appear again in the tertiary or secondary ducts.

Owing to individual variation, it is possible to find all of the above described developmental stages of lumina in twenty day and six hour fetuses.

Several of the glands of this stage were stained with Dominici's combination stain. In blood vessels, the corium within the epithelial hood, the connective tissue immediately surrounding the ducts, and the ordinary connective tissue in the entire gland region were found various kinds of white blood cells including eosinophiles. In one gland a few lymphocytes were observed in the developing lumina of the ducts. None of the glands ex



c m




J .0

1-' 3

a J? j^




.^-« I




amined showed such an infiltration of leucocytes as Keiffer ('02) and others have described in the human newborn.

The processes of the connective tissue cells have elongated and when treated ^\ith Mallory's connective tissue stain many of them now appear as true white fibrous connective tissue fibers. Weigert's elastic tissue stain revealed no trace of elastic fibers at this stage. The developing connective tissue has so differentiated that the anlages of two adult parts may now be recognized. That part immediately adjacent to the ducts forms a thin sheath around them. This sheath is the anlage of the mantle layer. While the connective tissue between the ducts represents the anlage of the true stroma (fig. 6, m.l., s.t.).

Lobules have not yet formed in the mammary gland.

The masses of fat which are so conspicuous in the postnatal stages are not developed at this stage.

The foregoing stage at twenty days and nine hours brings the description up to the condition at birth which was the starting point in my previous paper (Myers, '16). In newborn rats the lumina were found to extend through the primary ducts (except the intraepidermal portion) into the secondary ducts and to terminate in the end-buds. In the primaiy ducts the lumina are small irregular slit-Uke spaces which become continuous with the more regular rounded lumina of the remaining ducts. One can

Fig. 4 Drawn from a section through the primary duct (near free end) of the left second thoracic gland of a female albino rat fetus of eighteen days and nine hours to show development of lumen. X 550. Zenker's fixation; hematoxylineosin stain. Drawn with the aid of a camera lucida. cm., condensed mesenchymal I.e., small cavities (lacunae) which later fuse to form lumen.

Fig. 5 Drawing of a tangential section through a secondary duct of the right first inguinal gland of a female albino rat fetus of twenty days and six hours to show developing lumen. X 550. Zenker's fixation; hematoxylineosin stain. Drawn with the aid of camera lucida. c, irregularly arranged developing connective tissue cells; c.t., developing connective tissue forming sheath around duct; I., lumen formed by fusion of small cavities (lacunae); I.e., small cavities (lacunae).

Fig. 6 Drawn from a section through four tertiary ducts of the left first thoracic gland of a female albino rat fetus of twenty days and six hours. X 175. Zenker's fixation; hematoxylin-eosin stain. Drawn with the aid of a camera lucida. t.d., tertiary ducts; m.l., developing connective tissue to form mantle layers; st., developing connective tissue forming true stroma; I.e., lumen.

210 J. A. MYERS

safely assume tha,t the process already begun in the eighteen to twenty day fetuses continues until the time of birth, thus producing the continuous lumina found in newborn rats.

The lumma at birth have not assumed their definitive form, however. In a later study the details of the further development of the lumina in the postnatal stages of the albino rat will bo described.


In the following discussion, the nipple, the milk-ducts, the epithelial hood, gland stroma, variation, and cephalocaudal sequence in development will be successively considered.

The nipple

A comparison of figures 1 and 7 is sufficient to show that in the albino rat fetus slight eminences occur in the region of the future nipples. These eminences evidently correspond to the mammary hillocks described in other forms. The mammary hillocks in the rat fetus (as in other forms) are temporary eminences, each being soon replaced, as has been shown, by a shallow depression, the mammary pit. At the bottom of this pit later occurs a slight elevation representing the nipple anlage. The true nipple reaches only a very rudimentary stage of development in rat fetuses. The latest stage studied (twenty days and six hours) shows the nipple anlage as a rather slight eminence at the bottom of the mammary pit (figs. 3 and 9).

The phenomena of development in the nipple and the associated hillock and pit are rendered more intelligible by a comparison with the conditions found in lower forms.

The mammary hillocks first appear in rat embryos of fourteen days (Henneberg, '00). The present work shows that they persist through the sixteenth day at which time they are less conspicuous than at the fifteenth day. These hillocks apparently occupy the positions of the future nipples. Because of their positions and resemblance to a nipple, Schultze ('92 and '93) in the pig and other species called them primitive nipples


('primitive Zitzen'), a misleading term since, as he observed, they are merely transient structures.

Similar hillocks have been observed in human embryos by Langer ('51), Rein ('82), Brouha ('05), Lustig ('16), and others. They have been described by Rein ('82), Schultze ('92 and '93), Bonnet ('92) and Brouha ('05) in the following species: pig, sheep, dog, fox, cat, rabbit, squirrel, rat, mouse, and mole. The name mammary hillocks ('Milchhiigeln') was applied to them by Bonnet ('92).

The depression or fossa (mammary pit) which forms over each developing gland resembles the pocket which contains the nipple in some marsupials and which Owen ('68), Gegenbauer ('73) and others believed to exist in Monotremes. Bresslau in 1908 proved the non-existence of such a pouch in echidna and ornithorhynchus. In an earlier work, however, Bresslau ('02) observed that a definite pocket ('Zitzentasche') developed in some marsupials in the region of the future nipple. Bresslau's findings in marsupials confirmed the work of Klaatsch ('84) and others who showed that in marsupials a fairly deep pocket is developed in the region of each mammary gland; and at the bottom of each pocket a small papilla-like eminence occurs which is believed to be the first appearance of a nipple in mammals. During the resting phases of the glands the nipples remain in the pocket, but they actually protrude from the pocket and may be drawn out to a considerable extent while the glands are active.

The ontogeny of the mammary gland nipple of the albino rat apparently repeats in most respects the above described conditions in the lower forms of mammals. In the rat we have seen the surface over the future nipple region excavated (chiefly by the processes of cornification and desquamation) so as to form a definite pocket (figs. 8 and 9), the mammary pit. At the bottom of this pit is seen in sections the proximal end of the primary duct. Later a papilla-like elevation (the nipple anlage) appears at the bottom of the pit. At this time the nipple is so small that it occupies only a part of the pocket. At birth the nipple has enlarged so as to fill the pit, with the

212 J. A. MYERS

exception of a shallow sulcus which still surrounds the nipple. The nipple in the newborn rat thus produces a slight eminence on the surface of the skin. In an earlier paper (Myers, '16) my low power drawings do not show the sulcus around the nipple in rats at birth and one week of age. This is due to the fact that over the sulcus the epidermis is slightly thickened, and also because the sulcus contains some cornified cells. Nevertheless under high power the sulcus is still very evident in these postnatal stages.

The mammary pit which develops before the appearance of the nipple is apparently homologous with the nipple pocket which Gegenbauer (73 and. 76), Rein ('82), Klaatsch ('84), Bresslau ('02), and many others observed especially in marsupials. Bresslau ('02) believed that the mammary pit is homologous with the marsupial pouch. Later, however (Bresslau '10), he regarded it as a homologue of the nipple pocket of marsupials.

The milk ducts

In the rat fetuses the anlage of the milk duct was first observed about the seventeenth or eighteenth day. At this time the deep part of each epithelial mammary gland anlage apparently elongates or sends out a single bud-like process which is the primary duct anlage. This stage may be said to correspond to Rein's ('82) period of bud formation ('Kjiospenbildung') in rabbits. It differs, however, from the findings of Langer ('51), Huss ('71), Kolliker ('79), Rein ('82), Profe ('98), Hamburger ('00), Brouh'a ('05), Lustig ('16), and many others in that they observed a variable number of buds (primary duct anlages) in man and other animal species including the horse, pig, cat and rabbit. On the other hand it agrees with the observations of De Sinety ('77), Gegenbauer ('76), Klaatsch ('84), and Brouha ('05) who reported the existence of a single primary duct in rodents and insectivorous mammals.

Between the eighteenth and nineteenth days each primary duct in the rat fetus presents two secondary ducts. The secondary ducts later present tertiary ducts. Quaternary ducts are


present at twenty days. Very rapid growth takes place between the twenty day stage and the newborn, as my reconstructions and cleared preparations (Myers '16) show that the ducts are much elongated and several new divisions have occurred in the latter.

The first few divisions of the milk-ducts in the twenty day fetus (fig. 12) follow the true dichotomous method of branching. The divisions farther away from the primary ducts, however, do not come ofT so regularly, yet they present a very irregular form of dichotomy. The same condition obtains in the newborn and later postnatal stages (Myers '16). Langer'('51), Kolliker ('79), and Lustig ('16) found that for the most part the milkducts of human fetuses branch dichtomously. Kolliker ('79) states that the human manmiary ducts branch two to eight times by the true dichotomous method after which the branching is somewhat irregular. The method of branching of the milk-ducts of the albino rat, therefore, appears to be similar to that of the human.

The terminal end of each milk-duct in all stages of the rat fetus studied presents an enlargement. Langer ('51) noticed such enlargements at the terminal ends of developing milkducts in the human, and they have since been reported by numerous investigators. Formerly such terminal swellings were believed to be true acini. The present work, however, as well as my previous study (Myers, '16), confirms the view that they are not true acini, but are merely growing end-buds.

The first indication of a lumen in the ducts was observed in a rat fetus of eighteen days and nine hours. The lumina appear, however, in only a part of the ducts observed at this stage, while at twenty days and six hours the majority of the ducts show lumina in an early stage of development. At birth the lumina extend from the intra-epidermal portion of the primary ducts to within 20 or 30 micra of the free extremities of the terminal ducts. Such lumina, however, have not yet reached their definitive state.

The time of development of the lumen in the mammary ducts is subject to considerable variation, not only in different species

214 J. A. MYERS

but in individuals of the same species. In the rabbit, Rein ('82) found the first vestige of a lumen in a very late fetus. At five days after birth canalization is not entirely complete, but at fifteen days the lumen extends to the tip of the nipple. It does not open on the surface, however, owing to the presence of cornified cells in the proximal end of the primary duct. Brouha ('05) in the rabbit four days old found two of the milk-ducts with lumina throughout, other ducts at the same age showing only faint traces of lumina. At twenty-five days he found the lumina completely formed for all of the ducts. In a kitten twelve hours after birth Brouha found a part of the ducts provided with lumina. In Vespertilio murinus he found a trace of a lumen in the milk-ducts of 20 mm. fetuses, while at birth the lumina are quite well represented throughout the ducts. De Sinety ('75) and Lustig ('16) found the lumina begin to appear in human milk-ducts about the sixth or seventh fetal month, but are not completely developed until birth or later.

From the present work on rat fetuses and the foregoing observations of De Sinety ('75), Rein ('82), Brouha ('05), and Lustig ('16) it may be concluded that the lumina of milk-ducts usually begin to develop during the later fetal stages, but the definitive lumen does not appear until birth or later.

The earUest appearance of the lumen has been reported in different parts of the milk-ducts. In the previously published abstract of the results of the present paper (Myers, '17) it was stated that the lumina make their earliest appearance in the free ends of the milk-ducts. This statement agreed with the findings of Rein ('82), Eggeling ('04), Raubitschek ('04), and Lustig ('16). Further observations on a larger number of albino rat fetuses, however, indicate that the lumina may appear first in the excretory or external portions of the milk ducts, as observed by Kolliker ('50) and Brouha ('05) in the glands of the mouse, rabbit, cat and man. We must therefore conclude that the first appearance of the lumina of the milk-ducts is variable and may occur in various parts of the ducts. In the rat, however, in the majority of cases the lumina show slightl}^ further progress in development toward the free ends of the ducts.


The manner in which the lumen is formed has hkewise been a subject of considerable controversy. It will be recalled that in the rat fetus of about the eighteenth or nineteenth day small irregular intercellular cavities or lacunae appear in the epithehum of the milk-ducts. The lacunae are chiefly confined to the center of the developing ducts, but may occur peripherally. The cells and their nuclei in the region of the lac;unae show no signs of degeneration. A little later the lacunae flow together, thus forming a lumen which is in a very incomplete stage of development at this age. The lumina are better developed at birth (Myers, '16), but are still incomplete. De Sinety ('75), Rein ('82), and Keiffer ('02) have described the formation of the lumina in human as a process of degeneration. They state that the central cells of the solid epithelial duct anlage degenerate, the debris being found in the developing lumina. My fetal stages show no such condition, but agree rather with the findings of Benda ('94) and Brouha ('05), who described the formation of the lumen in the mouse, rabbit, cat and man as a process of cell-rearrangement, rather than cell-degeneration.

The epithelial hood

The anlage of an epithelial ingrowth or hood was first observed in one of the abdominal glands of an eighteen day and nine hour rat fetus. Such anlages are present in most of the glands in ninteen day fetuses. These anlages were seen to bud off from the deeper epithelial surface funnel-shaped mammary pit. About the twentieth day the ingrowth forms a real hood around the proximal end of each primary duct. When examined microscopically, the part of the hood attached to the walls of the mammary pit is seen to be filled with a thin layer of cornified cells which is continuous with the mass occupying a part of the mammary pit. No cavity is yet present in the hood, although its attachment corresponds to the region of the sulcus between the nipple anlage and the wall of the manamary pit.

The epithelial hood has been observed by several investigators (Gegenbauer, '7(3, Rein, '82, Klaatsch, '84, in rodents and

216 J. A. MYERS

insectivorous mammals) some of whom believed it to be homologous with the marsupial pouch or the nipple pocket of marsupials. As to the significance of the epithelial hood in the albino rat I have as yet reached no definite conclusion.

Gland stroma

The majority of the investigators have observed and described the mammary gland stroma. In the rabbit and man at the end of the period of 'Knospenbildung,' Rein ('82) found the first appearance of the gland stroma. In the albino rat according to Henneberg, the mesenchyma deep to the first anlage of the mammary gland is condensed. The present work shows that in the rat fetus at fifteen days the mesench^aual cells lying nearest the mammary gland anlage are elongated and arranged in two or three distinct rows around the anlage. At about seventeen and eighteen days, as the primary duct buds out from the main gland anlage it becomes well surrounded with developing connective tissue cells, which at this stage present long fibrous processes. As many as three or four layers of the cells and their fibers surround each duct, while farther from the ducts the connective tissue cells and fibers are arranged parallel with the surface of the skin. A short time before birth, at twenty days and six hours, the ducts are covered with a sheath of fibrous tissue. The connective tissue external to this sheath is somewhat condensed (fig. 6). The sheath which intimately surrounds each duct corresponds to the part which Berka ('12) described as the mantle layer of young girls and older virgins. The condensed tissue external to the sheath he designated as the true stroma. In the true stroma, blood vessels and nerves are found, but the blood vessels are not as abundant as one might expect.

The fatty tissue enclosed by the gland stroma, which takes an important part in the later development of the gland, was not observed in the fetal stages.



Individual variations in the development of the mammary gland are so frequent that at least mention should be made of them. Moreover, no work on the mammary gland should be regarded complete until the conditions have been studied in a sufficient number of individuals to rule out all possibility of error from individual variations.

Rein ('82) found many individual fluctuations in the developing mammary gland of human. In one pig embryo of 1.5 cm. Schultze ('93) found only the milk line while in another embryo of about the same size he found the 'primitive Zitzen.' Henneberg ('00) found in one rat embryo of eleven days no indication of a manmaary streak while in another embryo of the same age a well developed streak appeared only on one side. Raubitschek ('04) states that probably no other organ is subject to such great fluctuations in its development as the mammary gland.

In the present study, it has been noted that in the eighteen day and nine hour stage of the albino rat fetus some of the glands possessed anlages of only the primary ducts while in others there were secondary ducts. Also the lumen began to appear in one individual of this stage while in others there was no trace of a lumen present. The lumen continued to develop until at twenty days and six hours it was represented by a considerable cavity in some part of most of the ducts. Yet even at this stage an occasional individual possesses a gland without the slightest manifestation of a lumen.

The number of mammary glands of the rat likewise is subject to indi-vddual variation. Schickele ('99) found that in 6.66 per cent of the rats examined, only 11 nipples were developing. In 80 per cent of his rats 12 nipples (the normal number) were present. WTiile in 13.33 per cent there were 13 nipples present. In no case did he find more than 13 nipples. Henneberg ('00), Myers ('16), also reported a variable number of glands in albino rats. Schultze ('93) in describing the mammary glands of a rat embryo of 1.2 cm. mentioned only two thoracic pairs of glands


218 J. A. MYERS

but found the usual number of abdominal and inguinal pairs. In the case of the rat, some authors fail to report the species studied, which should always be stated in order to avoid errors and confusion in making .comparisons. Lantz ('10) states that the female brown rat (Mus norvegicus) has usually 12 mammae • — 3 pairs of pectoral and 3 pairs of inguinal — although these numbers are not constant, one or more teats frequently being undeveloped. He also states that the black rat (Mus rattus) and the roof rat (Mus alexandrinus) have only 10 mammae — 2 pairs of pectoral and 3 pairs of inguinal — with but little tendency to vary. A variable number of mammary glands has also been reported in many other forms, including man. Therefore, in all morphological and histological work as well as experimental work on the mammary gland, individual variation must be considered before drawing any definite conclusions.

Cephalo-caudal sequence in development

Henneberg's ('00) work shows that in the early stages of development of the mammary gland the more cephalic or thoracic glands are better developed than the caudal or abdominal and inguinal glands. In fact the inguinal gland anlages remain considerably behind the thoracic anlages. In carnivora Schultze ('93) found the more cephalic mammary gland anlages earlier and better developed than the posterior ones at the same age. A similar condition was found in a part of the fetuses examined during the present work. However, when the twenty day and six hour stage is reached the difference is not so noticeable. The order of sequence is therefore in accordance with the general rule that those parts occupying a more cephalic position tend toward earlier development than those parts occupying a more caudal position.


1. In fetuses at fifteen days and nine hours the mammary glands of the albino rat are in the club-shaped stage, the epithelial anlage forming an ellipsoidal body attached to the epidermis by a constricted neck.


2 About the seventeenth or eighteenth day the deep portion of each anlage elongates into a long solid cord of epithelium — the anlage of the primary duct. At this time each anlage is only about 0.05 mm. in length. The free end of each primary duct is directed toward the position which the future system of ducts will occupy. At eighteen or nineteen days each of the primary ducts present two secondary ducts. About the twentieth day tertiary and quaternary ducts are present. The first few divisions are usually dichotomous, while the more distal ones become somewhat irregular. Growing end-buds are present on the free ends of the terminal ducts.

3. Between the eighteenth and nineteenth days an epithelial projection grows in from the stratum germinativum around each gland area. Each projection extends entirely around the primary ducts thus forming the epithelial hood.

4. The mammary pit first appears on the surface as a slight depression over each developing gland. It is apparently formed by the processes of cornification and desquamation of the thickened epithelium. The pit is well developed at nineteen or twenty days.

5. The nipple anlage was first observed in twenty day and six hour fetuses. At this stage it is a small papilla-like eminence lying at the bottom of the mammary pit. The nipple reaches only a rudimentary stage of development in the prenatal stages of the albino rat.

6. The lumina of the ducts were first observed in eighteen day and nine hour fetuses. They were not confined to any definite part of the system of ducts, but usually appeared shghtly better developed toward the free ends of the ducts. The lumina do not reach their definitive stage in the fetal state. In the fetusee examined, the lumina are apparently formed by a rearrangement of the cells, thus producing numerous lacunae which later flow together to form the incomplete lumina of the latest stage studied. No traces of cell degeneration were observed.

7. In the earliest stages studied the mesenchymal cells are condensed around the mammary gland anlage. Later these cells elongate and develop long fibrous processes. At twenty

220 J. A. MYERS

days and six hours these cells and fibers constitute the greater part of the gland stroma which may be divided into two parts: (1) the mantle layer which is a thin layer immediately surrounding the ducts; (2) the true stroma which lies between the ducts and outside of the mantle layer. The true stroma contains the larger blood vessels and nerves of the glands.


Benda, L. 1894 Das Verhfiltnis der Milchdriisen zu den Hautdriisen. Derma tolog. Zeitschr., Bd. 1. Berka, F. 1911 Die Brustdriise verschiedener Altersstufen und wahrend der

Schwangerschaft. Frankfurter Zeitschrift fiir Pathologie, vol. 8. Bonnet, R. 1892 Die Mammarorgane im Lichte der Ontogenie u. Phylogenie.

Ergebn. d. Anat. u. Entw., Bd. 2. »

1897 Die Mammarorgane im Lichte der Ontogenie ii. Phylogenie.

Ergebn. d. Anat. u. Entw., Bd. 7. Bresslau, E. 1902 a Beitriige zur Entwicklungsgeschichte der Mammarorgane bei den Beutelthieren. Zeitschrift f. Morphologic u. Anthropologic, Bd. 4.

1902 b Weitere Untersuchungen iiber Ontogenie und Phylogenie des

Mammarapparates der Saugetierc. Anat. Anz., Bd. 21.

1908 Die Entwickelung des Mammarapparates der Monotremen,

Marsupialicr und ciniger Placentalier. I. Entwicklung und Ur sprung des Mammarapparates von Echidna. Semon's Zoolog. V Forschungsreisen, Bd. 4, Lieferung 5.

1910 Der Mammarapparat (Entwickkmg und Stammesgeschichte).

Ergebn. d. Anat. u. Entw., Bd. 19.

1912 Ueber Hyperthelie. Miinchener Med. Wochenschr., Jahrg. 59,

S. 2793-2795. Brouha, Dr. 1905 Recherches sur les diverses phases du developpement et de

I'activite de la mamelle. Archives de Biologic, T. 21. Danforth, C. H. 1916 The use of early developmental stages in the mouse for

class work in embryology. Anat. Rec, vol. 10. De Sinety 1875 Recherches sur la mamelle des enfants nouveau-nes. Arch.

de physiol. norm, et patholog., T. 2.

1877 Sur le developpement et I'histologie comparee de la mamelle.

Comptes rendus des Seances de la societe de biologic. Eggeling, H. 1904 Ueber ein wichtiges Stadium in der Entwicklung der

foetale Mamma beim IMenschen. Anat. Anz., Bd. 24.

1905 a Ueber die Driisen des Warzenhofs beim Menschen. Jenaischc

Zeitschr. Naturwiss., Bd. 39.

1905 b Ueber die Stellung der Milchdriise zu den iibrigen Haut drusen. Semon's Zool. Forschungsreisen, Bd. 4, Lieferung 5. Gegenbauer, C. 1873 Bemerkungen iiber die IMilchdrusen-papillen der

Saugethiere. Jenaische Zeitschr. f. Med. u. Naturw., Bd. 7.


Gegenbauer, C. 1S7G Zur genauere.n Kenntniss der Zitzen der Siiugethiere. Morpholog. Jahrb., Bd. 1.

1884 Zur niihercn Kenntniss des Mammarorgans von P]chidna. jNIorpholog. Jahrb., Bd. 9.

Ha.\cke, W. 1885 On the marsupial ovum, the mammary pouch and the male milk glands of Echidna hystrix. Proc. of the Royal Society, London.

Hamburger, Clara 1900 Studien zur Entwickelung der Mammarorgane. I. . Die Zitze von Pferd und Esel. Anat. Anz., Bd. 18.

Henneberg, Briinno 1900 Die erste Entwickelung der Mammarorgane bei der Ratte. Anat. Hefte, Bd. 13.

Huss, M. 1871 BeitrJige zur Entwicklungsgeschichte der Milchdriise. Jenaische Zeitschr. fiir Naturw., Bd. 7.

Jackson, C. M. 1912 On the recognition of sex through external characters in the young rat. Biological Bulletin, vol. 23.

Keiffer, H. 1902 La glande mammaire chez le foetus et chez le nourrisson. Bull. Soc. Beige de Gyn. et d'Obstet., T. 13.

Klaatsch, H. 1884 Zur Morphologie der Saugethierzitzen. Morphol. Jahrb., Bd. 9.

1895 Studien zur Geschichte der Mammarorgane. I. Theil: Die Taschen und Beutelbildungen am Driisenfeld der Monotremen. Semon's Zool. Forschungsreisen, Bd. 2.

Kolliker, Th. 1'850 Mikroskopische Anatomic. Bd. 2.

1879 Beitrage zur Kenntniss der Brustdriise. Verh. d. phys.-med. Ges. zu Wtirzburg, N. F., Bd. 14.

Langer, Carl 1851 Ueber Bau und Entwicklung der Milchdriise. Denkschrift der Wiener Akad. d. Wiss., Bd. 3.

Lantz, David E. 1910 The natural history of the rat. In "The rat and its relation to the public health," by various authors. P. H. and M. H. Service, Washington, D. C.

LusTiG, Hilda 1916 Zur Entwicklungsgeschichte der menschlichen Brustdriise. Archiv f. mikr. Anat., Bd. 87.

Morgan, J. 1833 Description of the mammary organs of the kangaroo. Transact, of the Linnean Society of London, vol. 16.

Myers, J. A. 1916 Studies on the mammary gland. I. The growth and distribution of the milk-ducts and the development of the nipple in the albino rat from birth to ten weeks of age. Am. Jour. Anat., vol. 19. (Abstract also published in the Anat. Rec, 1916, vol. 10, p. 230.) 1917 The fetal development of the mammary gland in the female albino rat. (Abstract.) Anat. Rec, vol. 11, p. 390.

Owen, R. 1832 On the mammary glands of the Ornithorhynchus paradoxus. Philos. Transactions, vol. 122.

1865 On the marsupial pouches, mammary glands and mammary foetus of Echidna hystrix. Philos. Transactions, vol. 155. 1868 Comparative Anatomy and Physiology of Vertebrates, vol. 3.

Profe, O. 1898 Beitrage zur Ontogenie und Phylogenie der Mammarorgane. Anat. Hefte, Bd. 11.

Raubitschek, H. 1904 Ueber die Brustdriisen menschlicher Neugeborencn. Zeitsch. f. Heilkund. Abth. f. pathol. Anatomic, Heft 1.

222 J. A. MYERS

Rein, G. 1882 Untorsiichungen liber die enibryon:ile Entwicklunf^sgeschichte

der Milchdriise. Arehiv fiir mikr. Anatomic, Bd. 20 and 21. Ru(;e, G. 1895 Die Hautmusculatur der Monotremen und ihre Beziehiingen

zu dem Marsupial- und Mammarapparate. Semon's Zool. For schungsreisen, Bd. 2 (Jenaische Denkschriften, Bd. 5). ScHiCKELE, G. 1899 Beitriige zur Morphologic und Entwickelung der normalen

und i'lberzilhligen Milchdriiscn. Zeitschrift f. Morphologic und

Anthropologic, Bd. 1, Heft 3. ScHiL, L. 1912 Recherches sur la glandc mammairc, sur les phases qu'elle

prcscnte au cours de son evolution ct Icur determinismc. These,

Lyon. ScHULTZE, O. 1892 Ueber die erste Anlage des Milchdrlisenapparates. Anat.

Anz., Bd. 8.

1893 Beitrag zur Entwicklungsgeschichte der Milehdriisen. Ver handl. d. phys. med. Gcsell. in Wurzburg, Bd. 26. SoBOTTA, J. 1895 Die Bcfruchtung und Furchung des Eies der Maus. Archiv

f. mikr. Anat., Bd. 45. SoBOTTA, J., AND BuRCKHARD, G. 1911 Rclfung und Befruchtung des Eies der

wcissen Ratte. Anat. Hefte, Bd. 42.



7 External view of a wax model reconstructed from the right first thoracic gland of an albino rat fetus of fifteen days and nine hours. X 100. e.s., eminence (mammary hillock) on surface of skin produced by developing mammary gland.

8 External and part of internal view of a wax model reconstructed from the left first inguinal gland of a female albino rat fetus of eighteen days and nine hours. X 50. n.p., depression representing mammary pit; p.d., primary duct anlage.

9 External view of a wax model reconstructed from the left second inguinal gland and surrounding region of a female albino rat fetus of twenty days and six hours. X 50. n.a., nipple anlage; n.p., mammary pit; s., sulcus surrounding nipple anlage.

10 Internal view of a wax model reconstructed from the right first thoracic gland of an albino rat fetus of fifteen days and nine hours. X 100. m.a., ellipsoidal mass of cells (mammary gland anlage) connected to epidermis through a constricted neck (nk).

11 Internal view of a wax model reconstructed from the right abdominal gland of a female albino rat fetus of eighteen days and nine hours. X 50. e.b., end-bud; p.d., primary duct; s.d., secondary ducts.

12 Internal view of a wax model reconstructed from the left first inguinal gland of a female albino rat fetus of twenty days and six hours. X 50. e.b., end-bud;, epithelial ingrowth (hood); p.d., primary duct; s.d., secondary duct; t.d., tertiary duct.


J. A. MTEliS







Dcpaiitiicnl of Aiidlniin/, ('unicll V iiivcr^llji Mcilicdl School, A^eir )'ork Clly



The existence of a more or less regular and definite oestrous cycle has been recognized in a number of mammals, particularly among the different classes of primates, carnivores, ungulates and insectivores. Yet very little is actually known or understood regarding the oestrous cycles and heat periods of a great many other very common mammals. Strangely enough, our knowledge of the sexual rhythm in the guinea-pig is much confused and not properly understood despite the great number of breeding experiments and the several studies of the sexual conditions which have been performed on this animal.

While conducting an extensive breeding experiment with guinea-pigs for the past several years it has become more and more desirable to know their exact oestrous periods. ^ A careful study of the existing literature bearing on this subject serves merely to produce uncertainty and confusion regarding their

1 Throughout this paper we have used the terminology proposed by Heapc, Quar. Jour. JMic. Sc, vol. 44, 1900, and adopted by Marshall and others. Anoestrous period or anoestrum, period of rest in the female; prooestrum, the first part of the sexual season; oestrus or oestrum, especial jx^rind of desire in the female; metoestrum, the short period when the activKy of the ii;cn<Ma( ivc system subsides and the normal condition is resumed in case concei)ti()n did not occur; dioestrum, the short period of rest which in some mammals lasts only a few days. Such a short cycle as we shall describe in tiie guinoa-i)ig consist in<i; of four periods the prooestnnn, oestrum, metoestrum and dio(>struiii is known as a (liocstrous cycle.



ovulation times aiitl heat seasons. The reason for such a lack of knowledge is that these small rodents do not reveal in a very evident manner the existence of their typical sexual rhythm as do many mammals of other classes.

The guinea-pig never, or only in rare cases, shows an external flow from the vagina, and there is no easily noticeable change in the appearance of the external genital organs during the different periods of sexual activity. The only expression generally obserA'ed of the sexual condition or heat jDeriod in the female is her willingness to accept the male, and this sign is, of course, only manifested when a male is present and a copulation takes place. The copulation then brings about the disturbing factor of pregnancy and the observation of the return of the heat period is prevented. The practical difficulties in observing successful copulation in these animals makes the study of their sexual conditions still more difficult.

Marshall ('10), in a recent smnmary has stated the case as follows :

It is difficult to detorniine the length of the prooestrum and (wstnis in rodents, since the external changes which characterize these conditions are comparatively slight. Heape says that the prooestrum in the rabbit lasts, probably, from one to four days. At this time the vulva tends to become swollen and purple in color, but there is no external bleeding. The same may be said of the rat and the guineapig; but, in the experience of the writer, it is generally impossible to detect the prooestrous condition in either of these animals with aljsolute certainty.

It must be recalled here that Marshall has devoted a great deal of study to this subject.

The difficulty in observing signs of heat in the guinea-pig has led a numbers of workers during the past fifty years to a study of the ovaries in order to establish the ovulation cycle. The results of such studies, as we shall point out beyond, are inaccurate and confusing in all cases.

Recognizing the above state of affairs, w^e determined to ascertain whether by a more minute examination of the genital organs of the female it might not be possible to observe an oestrous cycle. In order to examine the vagina thoroughly we have in


troduced a small nasal speculum which facilitates a clear view of the interior and a smear is made of any fluid that may be present.

A microscopic study of these vaginal fluids, to be described in the following pages, has shown that the guinea-pig possesses a perfectly regular and typical dioestrous cycle. And further, the surprising fact that the composition of the fluids is exactly comparable to the menstrual fluid taken from so high a mammal as the monkey. Heape, ('99), states that the menstrual fluid of the monkey contains a mucous secretion of the uterine glands, blood corpuscles, particles of stroma and epithelium from the uterus and the vagina and leucocytes. All of these elements are present in the fluid from the vagina of the guinea-pig during heat though the relative amounts differ from those in the monkey and the fluid is rarely sufficiently abundant to be recognized on the vulva.

The great advantage of this simple method of examination for the study of the oestrous cycle in these mammals which show no external signs of heat is evident, and we trust that the method may prove useful to those who find it necessary or desirable to know accurately the sexual periods in animals used f o ' experimental breeding.

Having begun a study of the vaginal smears from guinea-pigs we have been led to a more complete consideration of the uterine changes which alter the composition of these smears, and finally to an in^'estigation of the changes in the ovary and the process of ovulation and corpus luteum formation which accompany the activities on the part of the uterus. The present contribution comprises the results of these investigations.



It has been recognized for more than half a century that the guinea-pig comes into heat very quickly after giving birth to a litter of 3^oung. This period immediately following parturition has been the starting point for the great majority of studies on the sexual behavior of this animal and it has been demonstrated


frequently by such studies that ovulation takes place a few hours after parturition, the female accepting the male at that time. These facts are generally admitted but the most varied opinions prevail regarding the times of the su})se(iuent ovulations, when conception does not occur soon after parturition.

The question whether ovulation in the guinea-pig is spontaneous or dependent upon copulation has often been raised by various workers. The majority are of the opinion that ovulation is, or may be, spontaneous although influenced by copulation, and that there is no definite regularity or typical periodicity in the ovulation cycles.

Bischoff, was one of the oldest advocates of the theory of spontaneous ovulation. In a special paper devoted to the study of this problem in 1844, and later in a study of the development of the guinea-pig ('52), he defended the view^ that the guinea-pig, hke all other mammals, has a spontaneous ovulation. Bischoff states that the mature eggs reach the oviducts through the rupture of the greatly distended Graafian follicles during the first twenty-four hours following parturition. This fact, he points out, had previously been observed and was generally accepted by the earlier investigators with the exception of Schulz, 1829, who failed to recognize a heat period before the fifteenth day after parturition, and sometimes even to the forty-ninth day.

According to Bischoff copulation takes place within three hours after parturition. He agrees with the earlier statements of Aldrorandi, Legullois, Fraser and Schultz regarding the length of gestation, or period of pregnancy, as being about nine weeks, which is very nearly correct, sixty-two days being the normal length of time. He held that the return of the heat period did not follow any regular periodicity: "Wenn die Befruchtung unmittelbar nach der Geburt verhindert wird, so scheint die Wiederkehr der Brunst an keine ganz bestimmte zeit geknlipft zu sein, sondern von Umstanden der Individualitat, des Alters, der Jahreszeit, der Fiitterung, etc., abzuhangen." In fom* cases in which the females were prevented from copulating for some time after they gave birth to young a copulation occurred 40, 50, 51 and 51 days after the birth.


Reichert ('01), confirmed the observations of Bischoff regarding the existence of a heat condition and an ovulation process shortly after parturition — Reichert found manj^ fertilized eggs in the oviducts 18, 19, 20 and 22 hours after parturition which showed by their condition that copulation must have taken place many hours before. His opinion is that the Graafian follicles rupture about twelve to fourteen hours after copulation.

]\Iany recent authors have incorrectly stated Reichert 's position and assert that he claimed ovulation in the guinea-pig not to be spontaneous but to depend upon copulation. This is due to a misinterpretation of Riechert's ideas, originated by Bischoff in his second paper, 1870, which is chiefly an answer to Reichert's arguments. No doubt many of the incorrect notions regarding Reichert's position have resulted from authors reading this paper by Bischoff without referring to Reichert's own paper for his exact position.

Reichert explains his position very clearly as follows:

Es ware wiinschenswerth die Zeit genau angeben zu konnen, in welcher das Ei nach der Begattung aus dem Graaf'schen FoUikel ausgestossen \vird um die Einwirkung der Begattung auf das Austreten der Eichen bemessen zu konnen. Es ist zwar zu keiner Zeit auch nur wahrschelnlich gewesen, dass das bis zu den Eierstocken vordringende Sperma irgend wie direkt die Losung der Eichen oder richtiger das Bersten der Graaf'schen Follikel bewrken konne. Es ist ferner die bei anderen Thieren bekannte Tatsache, dass reife, selbst eingelvapselte Eier auch ohne vorausgegangene Begattung gelost werden durch Bischoff's Versuche auch fiir die Sdugethiere ausser Zweifel geseM. Das Bersten aber der Graaf'schen Follikel erfolgt unter vermehrtem Zudrang des Blutes zu denselben und in Folge der starken Vergrosserung ihres Inhaltes, des gallertartigen Fluidums und auch der Zellen der Membrana granulosa, sowie des Discus proligerus; das Eichen selbst vergrossert sich in der Brunstzeit wenig oder vielleicht gar nicht; dasselbe loset sich nicht, es wird, so zu sagen, von der Mutter ausgestossen. Daraus geht ferner hervor, dass die Begattung mit ihren aufregenden Wirkungen auf das Mutterthier, insbesondere auf den Zudrang des Blutes nach den geschlechtstheilen, einen sehr ivesentlichen Antheil am Bersten des Graff'scJwn Follikels und so also an der Befreiung des Eichens haben kann und haben muss.

This quotation shows that Reichert did not deny the existence of a spontaneous ovulation, but claimed that copulation had an important influence on the process of breaking the Graafian


follicle. He also admits that the existence of a spontaneous ovulation is pro^'cn for mammals by the experiments of PHschoff. The difference between the opinions of Reichert and Bischoff is not that the one denies and the other admits the existence of a spontaneous ovulation, but that the one believes copulation to exert an important influence over ovulation, while the other holds that such an influence, if it exists at all, is not really great. Leo Loeb ('11), who has studied the problem of ovulation in the guinea-pig very recently, still claims that copulation exerts an influence over the time of ovulation. That Bischoff also finally thought that there might be an influence on ovulation as a result of copulation is shown by the following remark from his second paper:

Sie meinen nur, es giibe doch auch noch Erscheinungen, welche zeigen dass die Mannchen und die Begattung auch einen Einfluss darauf ausiiben. Wenn dieser Einwurf so gehalten wird, dass er (name'y Reichert) zugesteht, die Erscheinung an und fiir sich ist vollkommen unabhiingig von dem Mannchen, dieses aber kann doch forderlich darauf einwirken, so wird dadurch nicht mehr gesagt, als wenn man sagen wlirde, eine gute Ernahrung, giinstige Verhiiltnisse der Temperatur und cles Klimas haben ebenfalls einen Einfluss auf die Reifung und Loslosung der Eier, und diese vielleicht einen noch grosseren als die Gegenwart des Mannchens und die Paarung. Und wirklich stecht auch gar Nichts entgegen, dem Mannchen in diesem Sinne einen Einfluss einzuraumen.

Hensen (76), also recorded that in the guinea-pig a copulation takes place shortly (about one hour) after parturition and six to ten hours later an ovulation follows. In cases where this first ovulation was not followed by pregnancy he recorded another ovulation 17, 18, 35 and 37 days later in the different cases. The duration of pregnancy he found to be 66 days — This along with Bischoff's record of an ovulation 43 and 44 days after parturition made it difficult to admit that the guinea-pig had regular periodical ovulations every eighteenth day. Hensen, therefore, beUeved that the guinea-pig probably did not have a sharply expressed periodicity — "Es scheint also die Brunstzeit der Meerschweinchen nicht scharf periodisch zu sein."

Rein ('83), again reports the existence of a condition of heat in the guinea-pig within twenty-four hours after parturition.


Regarding the occurrence of further heat periods Rein failed to obser\e any regular periodicity. Im Eintreten der Brunst habe Ich kcine Periodicitat bei den Versuchstieren bemerkt."

The foregoing studies are chiefly of historic interest yet they show that these earlier workers recognized the occurrence of ovulation shortl}^ after parturition and were uncertain or confused regarding the time or periodicity of subsequent ovulations. Little of definite value has ever appeared in the literature to further clear up the last point. We may now briefly consider the more recent contributions which bear on the subjects of ovulation and oestrous in the guinea-pig.

Rubaschkin ('05), gives a detailed description of the sexual conditions in the guinea-pig. He also recognized, as did the earlier observers, that a condition of heat followed shortly after parturition. In almost all females killed a few hours (up to fifty hours) after the birth of a litter an ovulation had occurred. He never observed o^^ulation as early as five hours after parturition though he found fertilized eggs in the oviducts as early as fifteen and seventeen hours after. Copulation occurs directly after ha\'ing given birth to young but for later heat periods Rubaschkin was unable to demonstrate any regular periodicity. Es ist mir nicht gelungen, eine bestimmte Frist fiir das Auftreten der Brunst festzustellen."

He did observe, however, that in some animals ten to twelve days after having given birth to young the entrance of the vagina showed some signs of heat activity. "Oeffnung der Vagina und Rothung der Vaginaloffnung." He claimed that heat ceased to recur after the month of October, at least when the animals were kept in a cold place. The duration of pregnancy was reported by Rubaschkin in three cases to be ten weeks.

Rubaschkin thus failed to recognize the regular oestrous cycles in these animals and also states the gestation period somewhat too long.

Konigstein in 1907 recorded the results of observations made on eighteen rats, one guinea-pig and five rabbits. He states that in the rodents heat occurs immediately after giving birth


to the young and lasts for twenty-four hours. Copulation only takes place during heat and if pregnancy fails to occur at the period just after parturition the next heat periods follow after intervals of three to four weeks.

Konigstein also examined sections of the genital tract giving some important histological descriptions based chiefly on the rat — we shall return to a consideration of these observations in connection with our findings on the guinea-pig.

Bouin and Ancel ('10), are of the opinion that guinea-pigs do not have a spontaneous ovulation, the process being dependent upon copulation. However, these workers seem to have reached this opinion from observations made on rabbits which were the chief objects of their study. Despite the striking classification which they make of animals having a spontaneous ovulation (monkeys, dogs, horses, cows) and those not having spontaneous o\ailation (rabbits, guinea-pigs, cats) they admit that rare exceptions are possible and that in any animal an ovulation might occur independently of a copulation.

C'est la im fait general, mais soumis a des exceptions rares. II peut arriver que des aniinaux a ovulation non spontanee operent la dechirure de leurs follicules murs en I'absence de tout rapprochement sexuel. Nous-meme et M. Villemin avons constate le fait chez le Lapin. M. Miilon vient egalement de I'observer chez le cobaye.

During the past several years Leo Loeb ('11 a, b) has contributed extensive and valuable studies bearing upon the sexual cycles in guinea-pigs, considering in particular the function and importance of the corpus luteum. Loeb examined a great number of ovaries at different periods, beginning with the time of the first copulation after parturition and concludes, as Rubaschkin 1905 and others had previously done, that the cycHc changes in the ovary take place independently of copulation. Loeb thought that the ovulations followed no exact and regular periodicity in all cases. The periodicity differed among the individuals and was influenced by certain external factors, particularly copulation. To quote:

The exact time at which the new ovulation occurs varies however somewhat in different animals, ovulation occurring earlier in some ani


mals than in others. In some cases it can be hastened through certain external factors, especially copulation, but in the large majority of cases it occurs sooner or later even without a preceding copulation.

He holds that eight days after ovulation large follicles are present in the ovary but sometimes ovulation may not occur for twenty or twenty-four days.

The 'sexual peziod,' period between two ovulations, according to Loeb lasts usually twenty to twenty-five days instead of being about two weeks, the time necessary for mature folhcles to appear. This delay in ovulation in spite of the presence of mature follicles within eleven to thirteen days, he beUeves is due to a mechanism in the ovary which prolongs the cycle, the corpus luteum begins this mechanism. The corpus luteum degenerates after a period of growth lasting from seventeen to twenty days and thus ovulation occurs about once in three weeks. We shall show beyond by a demonstration of the oestrous cycles, that Loeb's deductions drawn from studies of the histology of the ovary ai*e incorrect and, therefore, cannot be employed for determining the ovulation cycles in these animals.

Loeb further finds that when the corpus luteum is cut out immediately after an ovulation, the next ovulation occurs soon after mature follicles are developed— about thirteen to fifteen days. Under these conditions the normal sexual cycle is reestablished — but even here his periods are not exact being somewhat shorter than are actually normal.

The very varied time results obtained by Loeb may be given as follows: First, no ovulation has been found under normal conditions before the fifteenth day after the last copulation. Second, in a group of thirty-eight guinea-pigs killed fourteen days and eighteen hours and nineteen days and fifteen hours after the last copulation — one had o\ailated about the sixteenth day, another the eighteenth and another at the nineteenth day, while the remaining thirty-five had not yet ovulated. Third, in a lot of twenty-two guinea-pigs, twenty to twenty-six days after the last copulation, one had supposedly ovulated at the eighteenth day, four at the nineteenth day, one at the nineteenth to twentieth day, one at the twenty-third and one at the



twenty-fifth day and a half, while fourteen had not yet ovulated. J'ourth, in a lot of six animals killed twenty-six to thirty-four da3^s after the last heat period or copulation, only one had already ovulated.

A recapitulation of these results may be stated thus: under fifteen days no ovulation; sixteenth day, one; eighteenth day, two; nineteenth day, five; nineteenth to twentieth day, one; over fourteen days and eighteen hours and nineteen days and fifteen hours, thirty-five; twenty-third day, one; twenty-five and a half days, one; over twenty to twenty-six days, fourteen; twenty-sixth to thirty-fourth day, one ; over twenty-six to thirtyfour days, five. These figures as Loeb points out do not show any regularity in the occurrence of the ovulation process and, as we shall show beyond, they demonstrate how difficult or almost futile it is to attempt to solve the sexual cycles of an animal by a simple study of the ovarian conditions found on killing the animals at different periods. To anticipate slightly, the figures above show that Loeb entirely failed to discover the presence of a definitely regular periodicity in the ovulation process of the guinea-pig. Thus his examinations though much more thorough were as ineffective as those of the previous workers.

In 1913, Lams gave an instructive review of this problem. He again confirmed the long known fact that a heat period followed parturition in the guinea-pig. The copulation was found to take place within two to four hours after the delivery while ovulation occurred from twelve to seventeen hours after. Thus copulation generally preceded ovulation without being its cause. Lams gives no data on the occurrence of later ovulations but devotes himself to a detailed account of fertilization and the early development of the egg.

A consideration of the sum total of these various observations compels the admission that the opinions concerning the oestrous cycles in the guinea-pig are highly confused and totally unsatisfactory for application in exact breeding experiments. The one fact which presents itself was estabUshed by the earliest workers and confirmed by all subsequent studies — -that is, that a period of heat follows \vithin the first few hours after parturi


tion. In the literature only Schulz ('29), according to Bischoff ('52), denies this fact.

No typical rhythm has been established so far for the subsequent o\njlations in the guinea-pig. All observers who have examined a number of o\ailations found great differences in the supposed periods of time intervening between two ovulations as we have reviewed in detail above. The numbers give no evidence of a regular periodicity in the ovulation process but on the contrary would lead one to beheve that the greatest irregularity in time intervals was the rule.

On the other hand, really no observations exist to show anything like the occurrence of periodic changes in the uterus and vagina accompanying the retiurn of the heat periods. Such a thing as a regular oestrous or preoostrous flow is completely undiscovered in these animals.

Konigstein ('07), has examined sections of the uterus and vagina of a guinea-pig, and Blair-Bell ('08), has drawn comparisons giving many interesting observations, but they failed entirel}', or made no attempt, to observe the regular reappearance of a definite order of changes in either the uterus or vagina of this animal.


During the past six years we have been using guinea-pigs in an extensive breeding experiment and it has become more and more evident as om' work goes on that the existing notions of the ovulation periods in these animals are of no practical value, or are practically incorrect. In a number of the experiments it became important to know accurately when the females ' came into heat' and when ovulation took place. We had concluded, from numerous observations as well as theoretically, that the female guinea-pig very probably had a definitely regular and periodic sexual cycle if it could be worked out exactly. On account of the need of this exact information, we have studied the oestrous cycle in these animals during the past eighteen months.

Most other attempts at a solution of this problem have centered in a study of the ovary which necessitated either its removal


by operation or the killing of the animal. In either case the procedure brought to a conclusion the observation or experiments on the ovulation cycles in that specimen. Recognizing, on the other hand, that no thorough investigation of the uterus and vagina in the living female had been made, it occurred to us that possibly oestrous changes might take place even though they are so feebly expressed as not to be noticeable on casual observation. The absence of an apparent oestrous or prooestrous flow from the vagina of the guinea-pig has, as before mentioned, no doubt been the chief reason for the general lack of knowledge of the oestrous cycle. It was therefore determined to make a minute examination of the contents of the vaginae of a number of females every day for a long peroid of time, to ascertain whether a feeble flow might exist although insufficient in quantity to be noticed at the vaginal orifice or vulva.

The observations were made by using a small nasal speculum which was introduced into the vagina and the arms opened apart by means of the thumb screw. The speculum permits an examination of the entire surface of the vaginal canal. In this way the vaginae of a number of virgin females have been examined daily and smears made from the substances that happened to be present in the lumen.

By the use of such a simple method, it was readily determined after examining the first lot of animals for a few months that a definite sexual period occurs lasting for about twentyfour hours and returning with a striking regularity every fifteen or sixteen days. During this twenty-four hour period the vagina contains an abundant fluid which is for about the first half of the time of a mucous consistency. The vaginal fluid then changes into a thick and cheese-like substance which finally becomes slowly liquified and serous. This thin fluid exists for a few hours and then disappears. Occasionally toward the end of the process a sfight trace of blood may be present giving the fluid a bloody red appearance, otherwise it is milk-white or cream-color.

According to the changes in appearance and consistency of the vaginal fluid, one may distinguish four different stages. The


first stage having a mucous secretion, a second stage the cheesehke secretion, a third stage with the fluid becoming serous and a fourth stage, not always recognized, during which a bloody discharge is present. The duration of these several stages is subject in the different animals to individual variations. The first stage, however, is generally longest and lasts from six to twelve hours or even more and during this time there is a gradually increasing quantity of the mucous secretion which at its height is very abundant and fills the entire lumen of the vagina. The second stage is shorter, lasting from two to four hours, and passes gradually over into the third stage which lasts from four to six hours. The fourth stage is the shortest, only about one to two hours long, and for this reason it is often missed in examining the animals during the periods. It is also possible, as mentioned above, that the fourth stage may not typically exist in all individuals and the quantity of blood present is very different in the different specimens. The succession in which these stages follow one another is remarkably definite. We have never observed any change in the typical sequence of the stages and the time consumed by the entire process is generally as stated about twenty-four hours.

A macroscopical examination of the uterus and vagina during this period of sexual activity shows the entire genital tract to be congested. The vessels to the ovary, uterus and vagina are large and conspicuous, the uterine horns and the vagina are slightly swollen and inflamed. However, as soon as this short period of activity is over, the congestion disappears and the uterus and vagina take again their normal pale aspect. At the same time the vaginal fluid diminishes and the vagina, especially during the first week after this sexual acti\dt3^, is as clean as possible sho^\dng none of the secretion. The external vaginal orifice, which during the period of activity is more or less open actually shomng in a few cases a little fluid or some blood, closes and becomes less accessible after the period.

During the second week follo^Aing oestrus a little mucous discharge begins to appear in the vagina and increases progressively indicating that the new period of activity is nearer and nearer


approaching. The orifice of the vagina is sometimes open during this stage and thus explains why this sign, which was observed before, does not make it possible to detect the actual time of the regular oestrous activity. Rubaschkin has observed the opening of the vagina ten to twelve days after parturition, but this period of time is certainly too short ta indicate the return of heat. We agree with Rubaschkin in stating that during the ovulation the vagina is open, but we do not admit that the opposite is also true, that the opening of the vagina indicates unmistakably the return of the ovulation process.



A microscopical examination of the smears prepared from the vaginal fluid taken at the several stages separated above shows decidedly typical differences. The cellular character of a smear made at a given stage differs from the cellular make-up of all other stages. The relative numbers of various cell types in the fluid at different stages are so definite that one with a little experience may diagnose the exact sexual stage of the animal concerned solely by an examination of the smear.

A photomicrograph from a smear of the vaginal content during the first stage of mucous secretion is shown by figure 1. This mucous fluid is seen to contain an abundant mass of cells which, as shown in the figure, are of a squamous type with very small pycnotic nuclei sometimes broken into pieces. The cell protoplasm is also greatly degenerated having only a weak affinity for the plasma stains and exhibits a reticular structure. These cells derived from the wall of the vagina (fig. 17) characterize by their presence and great superiority in numbers this first stage. There are, however, to be seen particularly toward the end of the first stage a certain number of elongate, cornified viells without nuclei, which are desquamated from the more external portions of the vagina. These cells contrast in appearance with the fu'st type cells since in smears stained with haematoxylin and eosin they present a decidedly red color, while the abundant first type cells are almost grey. The red cells rather


serve to indicate an intermediate period between the first and second stages or periods of the flow, and may really be found during both stages but particularly at the end of the first and beginning of the second stage. In addition to these two kinds of cells other types may also be found in a first stage smear but they are never present in such abundance nor are they so typical as the two just mentioned. All of the cells float freely in the. mucus without assuming any definite arrangement.

During the second stage the vaginal fluid is filled with enormous numbers of cells which cause the cheese-like consistency of the discharge at this time. These cells illustrated by the photomicrographs, figures 2, 3 and 4 at three different magnifications, are derived from the upper portions of the vagina with a few from the uterus and they maintain to a higher degree the original or healthy architecture of an epithehal cell. The nuclei are fairly well preserved sho^vdng only slight signs of degeneration. The protoplasm has not greatly deteriorated and gives a good staining reaction thus differing from the grey-staining first stage cells. The cells are present in innumerable quantities forming the thick cheesy substance while the mucous secretion diminishes more and more until it almost disappears. This stage is of short duration.

The third stage begins with the hquefaction of the cheesy mass. A microscopical examination shows that the cells of the second stage become less and less numerous, while a great number of polymorphonuclear leucocytes appear among them (figs. 5 and 6). When the end of this process is reached almost every one of the cells has become isolated from others of its kind and lies in the midst of a number of leucocytes. The apparent action or effect of the leucocytes is to dissolve or digest the desquamated epithelial cells and this dissolving effect is not only noticed on cells surrounded by the leucocytes but in some cases the leucocytes dissolve their way into the interior of the cell-bodies (figs. 7 and 8). 'These appearances are not due, as might possibly be supposed, to the cells having devoured the leucocytes. This destructive influence of the leucocytes begins, as will be described latqr, before the desquamated epithehal cells have fallen


away from the wall of the uterus and vagina (figs. 15, 16 and 17). But it probably continues also after the cells are free in the lumen of the vagina. The dissolving power of the leucocytes, which probably causes the liquefaction of the cheesy mass of epithehal cells is shown very well when leucocytes are seen within a cell and the nucleus is beginning to dissolve. The nucleus is apparently digested and dissolved by coming in contact with the leucocyte without being at all engulfed or enclosed within the smaller body of the leucocyte.

As the third stage appoaches its end the material within the vagina is a thin fluid containing a great number of leucocytes as well as many epithelial cells of the second stage some of which contain leucocytes within their bodies. Such leucocyte containing cells are strikingly typical of the third stage. The leucocytes within these cells as would be expected very soon show signs of degeneration never staining so clearly as the free outside ones.

The fourth stage shows the same condition as the preceding but often at this time a slight hemorrhage takes place, though this does not always occur. A microscopical examination of the hemorrhagic fluid shows in addition to the great number of red blood corpuscles, a large number of leucocytes and also desquamated cells of the second stage, some of which are penetrated by leucocytes (fig. 9) . Sometimes red blood corpuscles are enclosed within the bodies of the leucocytes and digested, this is probably a truly phagocytic action and not entirely the same as their dissolving effect on the neighboring epithelial cells within the fluid.

The presence of the leucocytes is not alone confined to the heat period but an abundant quantity of them is also to be foiind in the lumen of the vagina during the dioestrum. The only time that leucocytes are absent from the vaginal lumen is during the first and second stage described above at the beginning of the oestrus. Throughout the first week after heat the little fluid which exists in the vagina contains chiefly leucocytes and a few atypical desquamated cells. During the second week the number of epithelial cells increases more and more and among these


atypical cells there may exist isolated cells of the first or of the second stage type.

At the fourteenth and fifteenth day the number of first stage cells already described begins to increase gradually and the groA\ing proportion of these cells indicates the approaching newperiod of heat.


The periodical return of a typical flow showing the above described macroscopical and microscopical details, was found to be very regular in twenty-six virgin females examined during different seasons of the year. Table I shows the results of this examination. As this table indicates, all the fema^les examined were virgin thus eliminating any chance of modification whichmight be due to the act of copulation. Their ages ranged betW'een three and a half and fifteen and a half months during the time of examination. The female guinea-pig is sexually mature at about three months old. Almost every animal, as the table shows, was examined for a length of time covering several oestrus periods. In the sixty-seven periods examined altogether the vaginal flow returned regularly every fifteen to seventeen days with an average of 15.73 days interval between the beginning of periods.

This table contains nine oestrus periods for operated animals from which one ovary was removed. The operation w^as done to determine w^hether any decided alteration in the oestrus would result after the loss of one ovary. The animals 1080 9 and 1102 9 were semi-spayed during the time of examination given in the table. In the animal 1080 ? the first heat period following operation came at the sixteenth day after the la^st period but in the animal 1102 9 the first heat period following operation came on the fourteenth day after the last heat, a little earlier than it should come inder normal conditions. The three heat periods following this came, however, very regularly every sixteenth day.

The animals 867 9 , 923 9 , and 1069 9 were semi-spayed a considerable time before the beginning of the examination and



r. o

h a z: o


1 uJ



I o

- -Q

(!) g









C£) CD







^ h

•oo c





















c O

aJ c o


c o










■oo d) ^ ^•^

< -Cx:



10 •-1









































































in if)




















to in'


'^ Sk














































o =





















1- S.















r2 >s



















s ^




















































■o •^



























15 £









































































00 0^













CO oo
















all of these showed an interval of seventeen days between the beginnings of the dioestrous cycles. The average of seven periods in animals with only one ovary is 16.57, this being much higher than the average of all the cases, which is 15.73 days. The average of the only two cases of first heat-periods after operation, on the other hand, is lower than the general average 15.0. The number of cases is, however, entirely insufficient to warrant a conclusion, though suggestive for further investigation. It is probable that when only one ovary exists, the period between ovulations is a little longer than under normal conditions. The two ovaries may alternate to a certain degree in their function or they may share the entire task in a less exhaustive way than one o\'ary is capable of doing. Semi-spayed females often have large litters which might indicate that the single ovary matured more foUicles than would have been its share should the other ovary have been present.

Eliminating from the general table the results obtained by the examination of the semi-spayed animals, one finds an average of 15.65 days for the length of time from the beginning of one heat period to the beginning of the next in all normal cases. This we befieve to be the length of the oestrous cycle of the guinea-pig under uniform conditions.

Table 1 further shows the months during which these observations were made. The animals were examined during early summer, fall, winter and spring and have shown at all seasons a perfect regularity in the return of the heat periods. Their oestrous cycle is certainly typically regular. The only months during which the animals were not examined are July, August and September. During the winter the guinea-pigs are kept in a fairly well regulated warm temperature running about 70° Falirenheit on an average. It may be possible that in the wild state under natural conditions when the weather is cold and food somewhat scarce, the heat periods may cease for a season or become less frequent. Rubaschkin claimed that heat ceased to recur after October when guinea-pigs were kept in a cold place. But under the steadily favorable conditions in which the guinea-pigs here considered are kept, it is certain that


they are sexually active throughout the entire year with an astonishingly regular return of their oestrous flow and breeding reactions.

A more careful consideration of the figures obtained during the different months indicates, however, that there probably is a small difference in the length of the sexual cycles during the warm and the cold seasons.

The curve shown in figure A indicates graphically this slight fluctuation, operated animals are excluded. The lowest average 15.50 days, or the shortest oestrous cycles, was found in the month of October, while the highest 16.14 days is shown during Januarj^ The heavy line at 15.82 days indicates the mean between these two extremes. It is probably not without significance that the averages during the months December, January, February, March and April fall above the mean line, while the averages during the months of May, June and October are below the line. From the cases considered this indicates that the length of the oestrous cycle is probably a little shorter during the warm time of the year and a httle longer during the cold weather. We must, however, admit that the number of considered cases, as given in table 1, is actually small and these slight seasonal variations may be more suggestive than demonstrative in importance, yet there is certainly a striking consistency in their arrangement.


After having determined the regularity of the dioestrous cycle in a number of virgin females, they were killed at different stages of the oestrous period and their ovaries as well as pieces of the uterus and vagina were carefully examined and then fixed and preserved for microscopical study. The uterus and vagina must be fixed in certain fluids to avoid shrinkage and a tearing away of the epithelium from the wall. Bouin's fixing fluid has proven most satisfactory for this purpose while the ovaries were generally fixed with Zenker's fluid.

During the dioestrum or resting period the uterus is lined by a layer of cuboidal ciliated epithelium. Figure 10 shows a sec



3 • ^^

" "S


£ ^ is « 

<D ^


'^ bC

- c

CO •-<

•3 §

■*^ ^ ^ bjD

m _ ""

-5 ^ ^^

bC •

05 ^ ^ -^J -^ «*bC O S <n

I 1 §

^ s

U S3


■< ^

• a>

c +^























sAva KJi gnoAo sno\Jis'30


tion through the uterus at four and a half days after the last oestrus. At this thne the epithelial cells present a normal and vigorous aspect. No loss or breaking down is to be noticed. A few leucocytes are occasionally seen among the cells of the stroma, but never in large numbers. Mitoses are not frequent at this time but they are to be seen now and then.

When the heat period begins, the epithelium loses its normal appearance (figs. 11, 12 and 13). The epithelial cells become tall and columnar and are fiUed with mucus which they begin to form in abundant quantity The nuclei of the columnar cells appear closely pressed one against the other and are pressed into different levels in the various cells so as to give an appearance of several rows of nuclei. The epithelium thus takes on a pseudo-stratified arrangement. At the same time, a large number of leucocytes begin to migrate from the capillaries through the stroma and towards the epithelium. The stroma itself is congested and possesses a more profuse circulation than usual.

These appearances are to be seen in animals killed during the first phase of their period, that is, when the vagina contains an abundant mucous fluid filled with desquamated epithelial cells. A smear of this fluid is illustraed in figure 1.

As soon as the second phase of the vaginal fluid appears (figs. 2, 3 and 4), the uterus shows another aspect. The leucocytes are accumulating in large numbers below the epithelium, forming in some places a perfect wreath of leucocytes under the epithelium or actually a separate layer of cells (fig. 14). The stroma shows a more advanced degree of congestion.

During the third stage, smears figiu'es 5 and 6, the leucocytes penet;rate more and more into the epithelium some of them making their way into the lumen of the uterus by passing between the epithelial cells. Other leucocytes actually enter the epithelial cells and penetrate into their interior (fig. 15). A stage more advanced in appearance corresponding to a late third stage though from the same animal as figure 15, is shown in figure 16, where the entire epithelium is almost completely disintegrated. A great number of leucocytes has already penetrated the epithelium the cell structure of which has become largely destroyed.


Large vacuoles are to be seen between the epithelial cells, and these are probably produced by the dissolving power of the leucfocytes. Under the destroyed epithelium haematomata are to be seen in several places, produced by the congestion of the peripheral capillaries in the stroma. A leucocytosis somewhat similar to the above has been described by Heape, Konigstein, Blair-Bell and others in the uteri of several mammals.

The vagina of the guinea-pig also shows analogous conditions as illustrated in figure 17.

The broken down epithelium remains until the regeneration process begins. The reparation starts from the necks of the uterine glands which have remained intact during the entire process of destruction. A few leucocytes are to be seen between the epithelial cells of the uterine glands but this small number apparently passes through the epithelium into the duct without injuring the epithehal cells. The stage of reparation corresponds to the fourth stage, that is, to the period when blood is sometimes seen in the vaginal fluid, see smear figure 9. This is not difficult to explain since regeneration and the falling off of the degenerated epithelium take place at the same time. Regeneration of the uterine epithelium before the oestrous flowhad ceased has been reported in other mammals.

After examining a number of specimens, one may get t;he impression that the new epithelium growing out from the neck of the glands tends to push off the old degenerate epithelium, as it becomes detached from the wall of the uterus. Figures 18 and 19 show this condition where the new and the old epithelium are still existing in close proximity, the one growing out from the gland, the other breaking away from the wall of the uterus. In figure 19, this condition is more adva;nced and one sees th.e old epitheUum partly detached from the wall of the uterus. Generally the epithelium falls off still connected with pieces of the stroma, which also seems to be destroyed to some extent during every heat period. These masses of epithehal cells are commonly found in the vaginal fluid. When the epithelium falls awav the haematomatia are uncovered and the blood con


tained in them passes into the lumen of the uterus. A similar bleeding may also occur into the lumen of the vagina.

The regeneration of the mucosa seems to take place very quickly. About six to ten hours after the above stage the new epithelium is already completely formed. The growth of the new and the falling off of the old epithelium seem to go hand in hand, so that no stage is to be found when the uterus is completely unlined by its epithelial layer. However, one may occasionally observe, during the above described fourth stage, limited naked regions from which the old epithelium has been detached before the new has formed.

The wall of the vagina undergoes somewhat the same destructive changes as the wall of the uterus except that the desquamation of the vaginal epithelium does not occur in cell clumps or groups at t.he end of t^e third st0,ge. The vagina merely sheds its epitheUal cells singly but in increasing numbers from the beginning of the heat period up to the third stage. The desquamation appears to proceed from near the entrance up into the inner portions of the vagina. The cells which appear during the first stage come from near the outer part of the vagina, while during the second stage the desquamated squamous cells are derived from the inner part of the vagina. This statement does not include the cornified cells from near the orifice, which are found as mentioned above, between the first and second stages. The vaginal epithelium is also invaded by the leucocytes. This migration is very \dgorous during the third stage, about the same time as in the uterus. An innumerable mass of polymorphonuclear leucocytes migrate into the vaginal epithehum and actually enter its more superficial cells by penetrating into their ceU bodies (fig. 17).

The beginning of the desquamation before the massive arrival of the leucocytes shows that the primary cause of the desquamation is not the presence of the leucocytes. But, on the contrary it is probably the presence of the altered and dying desquamated cells which induces the extensive migration of leucocytes to this epithelial surface. The large epithelial cells of the vagina photographed in figure 17 are the same cells which are


to be observed in the vaginal fluid during the third stage, see smears figures 6, 7 and 8. A congestion of the capillaries of the mucosa also takes place in the vagina, and slight hemorrhages may occur as in the uterus, when the destruction of the stratified epithelium chances to reach down to the tunica propria.

The leucocytes are chiefly attracted to that portion of the epithelium covering the outfoldings into the lumen and this part undergoes a greater destruction. In a similar way it is the epithelium covering the prominent folds of tjie uterus which is destroyed, while the ingrowths which form the uterine glands are preserved and thi'ough regeneration from their necks furnish the new material which is necessary for the restoration of the lost epithelium.

During the dioestrum or rest period the desquamation of epithelium from the vagina does not stop completely and the scant vaginal fluid always contains some desquamated cells. At the same time, and probably connected with the shedding process the exodus of the leucocytes also continues though in a less active way than during heat. The 'intermenstrual fluid' therefore always contains a considerable number of leucocytes.


A study of the ovaries fixed during different stages of the oestrous cycle has shown that every change taking place in the uterus and the vagina has its corresponding stage of change in the ovary. At the beginning of the first stage the ovaries possess large, ripe follicles, figures 20 and 21. The nuclei of the eggs contained in the follicles are in a resting condition. The theca folliculi shows the beginning of a slight congestion. As the first stage advances this congestion becomes more and more pronounced and by the beginning of the second stage it is highly developed, figures 22 and 23. This extreme congestion of the theca folliculi, which exist at about the same time as the congestion stage in the uterus (cf. fig. 14) indicates that the follicle is ready for rupture. Heape has pointed out that the rupture of the follicle is due to this congestion and if the ovarian



blood supply be tied off follicles do not rupture. During this time the nucleus of the egg is still in a resting condition.

llie ripe follicles break at about the end of the second or the beginning of the third stage. Figure 24 shows a follicle just broken at the conunencement of the third stage. It will be recalled that at this time the active leucocytosis begins in the uterus and the vagina, compare figures 15, 16 and 17. The ovaries are not omitted from this active migration of the leucocytes. A number of leucocytes are to be seen in the corpus luteum during its early development, but great numbers of leucocytes are to be found mainly in the atretic follicles, which are now becoming the seat of regressive and degenerative processes (fig. 25). The eggs in these disorganizing follicles show a peculiar activity expressed by the formation of the maturation spindle. Most of the eggs begin to degenerate before the formation of a polar body, though some of them succeed in completing their maturation divisions. Figure 25 shows an egg within a disintegrating follicle, the follicle containing a great number of leucocytes. This egg possesses a well formed polar body in the process of division. Kirkham has reported similar conditions in the ovary of the mouse, he notices that eggs degenerate after forming the first polar body and the second polar spindle, a condition closely similar to that shown in our figure 25. The outline of the polar body is clearly shown in the specimen. The photograph is not 'touched up.'

The chromatin of the nucleus is to be seen in the center of the egg in figure 25. In all the cases observed, the eggs of the atretic follicles degenerated, the nucleus breaking up into irregular pieces very soon after ovulation had taken place from the ruptured follicles. We failed to find anything to indicate a tendency toward parthenogenetic divisions in the many specimens which we have examined as Leo Loeb reported for these animals.

The ruptured follicles very quickly begin to undergo a reorganization resulting in the formation of the corpora lutea. Even during the third stage the corpus luteum is a well circumscribed body beginning its differentiation by the ingrowth of the vascular


tissue of the theca foUiculi into the hypertrophied foUicular epithehum (fig. 26). This condition is more advanced during the fourth stage, when reparation begins in the uterus. Figures 27 and 28 illustrate two corpora lutea from the same ovary during the stage of uterine hemorrhage, the two are cut in different directions. The ingrowth of the vascular tissue toward the central cavity is apparent in these two figures. A well formed mature corpus luteum is shown in figure 29, taken from a section through the ovary of an animal about four and a half days after the heat period when the uterus was in a typical resting condition (fig. 10).


After a review of the above described facts there are several problems of general importance which may be profitably discussed in connection with them.

A fact of considerable significance is that the development and the degeneration of the uterine and vaginal mucosa corresponds very closely to the development and degeneration of the corpora lutea in the ovaries. At the time when the corpora lutea are highly developed and apparently active the mucosae of the uterus and vagina show a normally vigorous and healthy condition (cf . figs. 10 and 29) . While on the other hand when the corpora lutea begin to degenerate during the second week after the 'heat period' the mucosae of the uterus and vagina also begin to show" signs of degeneration and the process of desquamation slowly commences. At about two weeks after the last 'heat period,' when the wholesale destruction of the mucosa begins, the corpora lutea are almost completely degenerated.

The breaking of the Graafian follicles occurs during the oestrus as a result of a congestion which began in the theca folliculi at about the same time as the congestion of the stroma of the uterus and vagina. And finally when the regenerative growth of the uterine mucosa sets in, the ovaries then possess new corpora lutea in an active state of differentiation which were derived from these recently ruptured follicles.


These occuiTences argue very decidedly against the theory ad\'anoed by Fraenkel ('03), and until recently supported by a number of other investigators. Fraenkel believed that the corpus luteum is the cause of the menstrual condition, producing through its secretion the destructive changes in the uterus and vagina. Such a supposition does not in any sense accord with the phenomena as they appear in the gTiinea-pig. If there is to be ascribed to the secretion of the corpus luteum an action upon the uterine and vaginal mucosae such an action is not of an injurious but of a protective nature. As we shall bring out further, the most plausible opinion of the action of- the corpus luteum in the ovary itself, may also be interpreted as of a protective nature since it seems to prevent rupture of the Graafian follicles and the discharge of the ova. The facts obtained in the present investigation might not fully warrant the position that the corpus luteum really exerted an actively protective influence over the uterine mucosa, but they certainly in no sense suggest, and actually speak against, any injurious action on the mucosa by the secretion of the corpus luteum.

At the same time it is difficult to maintain that the absence of the protective action of the corpus luteum is the only or actual cause of the oestrous activity. The cause of oestrous is very probably more complex and the definitely regular rhythmical changes which take place in the uterus and vagina of the guineapig can not be fully explained as due alone to the degeneration of the corpus luteum. The absence of the luteal secretion possibly merely permits the uterine flow to occur as it seems also to permit the rupture of the ripe Graafian follicles. While the real mechanism determining the uterine reaction is a more complex factor and relatively independent, but affected in its expression by a close inter-relationship with the ovaries.

The various theories, however, which attempt to localize the cause of the uterine changes in the ovary are not in any case fulh' in accord with all the facts. It is of course true that the existence of the ovaries is necessary for the normal development and function of the uterus and vagina, and also that the removal of both ovaries leads to a disappearance of the typical oestrous


changes in the uterus and finally to a degeneration of this organ. Yet the complete removal of the ovaries does not always prevent the menstrual periodicity from expressing itself in an atypicl but regular way for a considerable time afterwards (see Halban).

Our observations on three females from which both ovaries have been completely removed, show that such an operation does not fully abolish the return of the destructive menstrual changes as is generally claimed. But on the other hand, the absence of the ovaries promotes and prolongs the continuation of these destructive changes in such a way, that instead of a periodical menstruation, these spayed females have a long, continuous and atypical destruction of the uterine and vaginal mucosae, which leads finally to the degeneration of these organs. In some cases a distinct periodicity may be perceived, indicating that the rhythm of the menstrual activity may exist independently of the ovaries. The phenomenon that really is abolished and absent from the uterus after the removal of the ovaries is the return of any regenerative or reconstructive process which we believe is normally due to a secretion from the newly formed corpora lutea.

From such a view of these phenomena one may draw the following general conclusions: The oestrous changes in the uterus are regulated by two different factors, one direct and the other indirect. A secretion elaborated in the ovary apparently by the corpus luteum is necessary for the normal development and persistence of the uterine and vaginal mucosae. The absence of the secretion leads to regression and degeneration of the uterine tissue. Yet this control is not the entire explanation of menstruation. The regulation of this process and the return of definite changes in definite periods of time may possibly be due to the existence of a fixed mechanism somewhere outside the ovary. The role of the ovary and especially of the corpus luteum is not to produce but to permit and to stop the menstruation. Our conceptions correspond completely with the ideas of Halban, who has recognized the protective role of the ovaries upon the


uterus and the vagina and the existence of a separate causal factor of menstruation independent of the ovary.

Fraenkel's theory that the corpus luteum is an active factor producing menstruation does not correspond with our observations. Neither, on the other hand, does the assertions of Marshall and Runciman that the corpora lutea evidently exert no influence on the occurrence of heat" seem to us justified. Marshall and Runciman ('14), have advocated the importance of the interstitial cells in considering the ovarian factor concerned in the recurrence of the oestrous cycle as opposed to any active effect of the corpora lutea. They point out the evident incorrectness of the old views that the ovaries and uterus are related by a nervous connection. Transplantation experiments have shown the fallacy of such a notion and have demonstrated the presence of an internal secretion from the transplanted ovarian mass. Marshall then in arguing against the importance of the corpus luteum uses Heape's ('97), observations which showed that in monkeys menstruation might take place in the absence of either ripe follicles or newly formed corpora lutea. This observation, it seems to us, does not in any way point towards the interstitial cells as being important. Nor does it argue against our view that the absence of the corpora lutea permits menstruation and that their presence exerts a protective influence over the uterine mucosa. Heape's observation is perfectly in accord with this and it is to be expected that corpora lutea should be either degenerate or absent when menstruation occurs.

Marshall and Runciman performed operation experiments on lour bitches. At these operations they attempted to destroy the large Graafian follicles by pricking with a knife or needle. In the first fox terrier at least nine follicles were injured in this manner. But one who has operated on the dog's ovaries knows how difficult it would be to discover all of the ripe follicles and almost impossible to get those on the dorsal surface of the ovary which is often closely bound down and almost covered. Yet it is not necessary in this discussion to question the destruction of every ripening follicle since the photomicrographs, which the


authors publish, show that corpora lutea formed after the rupture of the folhcles, and they state that the folhcles artificially ruptured changed into structures almost identical with normal corpora lutea" — except that development was not sufficient to fill the central cavity.

In the first two animals, which were their best experiments, since the time of the expected 'heat period' was fairly accurately known, the 'heat' came on about the time, or perhaps a little later, than it was expected and was not greatly influenced by the operation. This is just what we should expect on our supposition of the function of the corpora lutea. The dog is a monoestrous animal with a long anoestrous period and the destruction of Graafian follicles a few weeks before the oestrus was expected would have no bearing on the probable function of the corpora lutea in bringing on this period. The old corpora lutea resulting from the last ovulation were not disturbed and were probably just about degenerating and thus permitted the oestrus to occur very near the normal time. While the newly formed corpora lutea resulting from the operation were not sufficiently vigorous in their action to do more than slightly delay the menstruation.

Marshall and Runciman concluded that it is evident that the occurrence of 'heat' in the dog is not dependent upon corpora lutea, and that "The ovarian interstitial cells are possibly conerned in the process, but cyclical changes in the condition of these cells have not so far been observed in the dog's ovaries."

These conclusions and Marshall and Runciman's discussion are directed chiefly against Fraenkel's idea regarding the way in which the corpora lutea act; that is, the corpora lutea by their secretion perform an active function in bringing on the oestrous condition. We also disagree on the basis of the evidence furnished by the guinea-pigs with Fraenkel's views and for these animals at least such opinions are entirely incorrect. It seems to us, however, that Marshall and Runciman's experiments do not in any way argue against the position that the corpora lutea exert a protective influence over the uterine mucosa, nor that the absence or degeneration of the corpora lutea and the dis


appearance of its secretion permits the uterine mucosa to vmdorgo the degenerative changes typical of the 'heat period.'

Therefore, we must object to their conclusion that the occurrence of heat is not dependent upon corpora lutea — and further we are unable to believe that their experiments, or any other so far recorded, indicate that the ovarian interstitial cells are possibly concerned in the process." The evidence to our minds does not in the least point in such a direction.

A most ingenious attempt at an explanation of menstruation and one of the first logical views regarding the function of the corpus luteum was advanced twenty years ago by Beard' in his monograph on the 'Span of gestation and the cause of birth.' According to Beard "Menstruation is comparable to an abortion prior to a new ovulation, and it is an abortion of a decidua prepared for an egg which was given off subsequent to the preceding menstrual period, and which had escaped fertihzation."

In the earlier mammals. Beard imagines that gestation extended over only one ovulation period or short dioestrum of Heape's terminology. Thus prior to each ovulation, a birth would take place provided pregnancy had ensued after the previous ovulation, and if not the ovulation would be preceded by an abortive birth act. In this connection it is interesting to recall the well known fact that in man and otjier mammals abortions occur with a far greater frequency at the times for regular menstrual periods than at other times. In the human the time of the first menstruation after conception is a most critical period, and the time when the third menstruation should occur is responsible for the great predominance of three month foetuses to be seen in most collections, and so on up to the tenth period when the normal birth takes place.

In the evolution of mammals Beard calls attention to the tendency to develop a longer gestation period and more fully developed offspring, but in all ca^es the length of the gestation period is a multiple of the primitive ovulation periods. A reminiscence of the earlier primitive conditions still exist in all of the polyoestrous mammals. The gestation period of the guinea-pig extends over four oestrous cycles making it about sixty-two days long.


During pregnancy in higher forms, according to Beard's scheme the corpus luteum exerts a protective function by preventing a new ovulation and an abortive birth. In non-pregnant females, however, this abortive process is not counteracted by the quickly degenerating corpus luteum spurium and the uterus undez'goes the changes of menstruation and a new ovulation occurs. This ingenuous theory aims to furnish an explanation of the periodically destructive changes occurring in the uterus and vagina of some mammals at the same time that the ovary is preparing to liberate its ova. And the chief virtue of the theory is that it points out the protective action of the ovary and especially of the corpora lutea on the uterine mucosa. Every menstruation process and every abortion reflex as well as every normal birth is the result of two different factors, one the condition produced by the absence of the luteal secretion and the other is the expression of a phylogenenetically and physiologically fixed rhythmical tendency within the uterus itself.

Beard's conception of the corpus luteum as an organ preventing ovulation has been adopted and further developed by many later investigators, Prenant, Sandes and Skrobansky, Leo Loeb, Ruge, Pearl and Surface, Halban and Kohler and others. All of these investigators have a,dded evidence in favor of Beard's corpus luteum theory partly by new observations and partly by experiments on the living animals.

To state Beard's ('98, p. 101) position in his own words:

The corpus luteum is probably a contrivance for the supression or rendering abortive of ovulation during gestation. The commencing degeneration of this structure some little time before the end of the gestation (like its rapid atrophy where fertilization has not taken place) allows of preparation being made for a new ovulation.

We are indebted to Leo Loeb ('11), for first putting these conceptions of Beard to experimental test. And Loeb showed that pregnancy as such does not prevent ovulation if corpora lutea are extirpated from the ovaries. Loeb also destroyed the corpora lutea in non-pregnant guinea-pigs and later examined the ovaries after different periods of time. In forty-two females the corpora lutea were destroyed by cutting them out completely


with the following results: In one case the next ovulation had already occurred at the twelfth to thirteenth day (by the next o^•ulation is meant the ovulation following the last copulation) in one case at the thirteenth day, in five cases at the thirteenth to fourteenth day, in twelve cases at the fourteenth to fifteenth day, in four cases at the fifteenth to sixteenth day, in one case at the sixteenth to sixteenth day and a half, in one case at the sixteenth to seventeenth day, in one- case after eighteen days, while in eight cases ovulation had not yet occurred at the time when the animals were killed.

Loeb also cauterized the corpora lutea in the ovaries of thirtyone guinea-pigs but the results, oAving to the inferiority of this method, were not so satisfactory. The ovulation in some cases came at the fourteenth to fifteenth day, in other cases later. Loeb interpreted these experiments to indicate that the removal of the corpus luteum hastened the next ovulation. Such a conclusion is in no way actually contradicted by our observations, yet the experiments of Loeb are not completely satisfactory in the fight of the present findings Loeb thought the usual sexual period, or time between two ovulations, in the guinea-pig was very much longer, and much more variable than it actually is. On such a basis it seemed that the ovulation period in the animals he examined had been considerably reduced. But as the present study shows the normal oestrous cycle in the guineapig is from fifteen to seventeen days, usually about sixteen days with very insignificant variations. So that the periods recorded by Loeb, after the operations are actually just about of normal duration. He found the greatest number of cases to ovulate after a period of fourteen to fifteen days (12 such cases or 28.57 per cent) and considered this much shorter than the normal condition, where as a matter of fact such a period differs only insignificantly from what we find to be the regular length of the oestrous cycle.

When we also take into account his method of calculating the days between the last copulation and the next ovulation, and especially the fact that he figured the ovulation time by the condition and probable age of the newly formed corpora lutea


found in the ovaries examined, the sHght variations are all very probably within the limits of error. We also believe that Loeb has been misled by the application of similar methods in calculating the normal sexual periods in these animals.

In order to test the influence of the removal of the corpora lutea on the following ovulation time, one must first definitely establish a normal o\iilation period. Since this was not done we are forced to acknowledge that Loeb's experiments do not demonstrate the importance of the corpus luteum in regulating the ovulation process, though he must be credited for having definitely attacked the problem experimentally. Some doubt will also exist in the minds of those who have attempted the operation as to whether all of the corpora lutea are often to be removed from the ovary while it is in position in the abdomen.

We are not at all opposed to admitting the probability that the removal of the corpus luteum may shorten the usual sexual cycle. In fact such a discovery would accord Avith our notions of the function of the corpus luteum. We feel further that the present study has established the existence of a definite normal oestrous cycle and this knowledge makes the experimental analysis of the influence of the corpus luteum much more readily approached.

The knowledge of a typical and regular sexual cycle in the guinea-pigs as here demonstrated, paves the way for a better and more uniform understanding of the oestrous conditions prevailing in the different classes of mammals. All cases that have been studied with sufficient care give evidence at least of some rhythmical activity. The absence of external signs of oestrus in a great number of mammals, one of which was the guinea-pig, is the most e\'ident cause of a lack of understanding of their sexual periodicity. It is to be hoped that the application of the simple method of examination of the vaginal fluid used in the present study may enable workers to readily obtain a clearer understanding of the sexual acti\'ities of other commonly used laboratory animals as well as mammals in general, since such information is of the greatest value in all exact experimental breeding.


The typical oestrous cycles are probably more regularly expressed among mammals living in a state of domestication, and consequently under steady environmental conditions, than among their relatives living in the wild, where the existence of great disturbing factors, especially variations in food and temperature conditions, may tend to modify their behavior. The evidence of such modification by these disturbing factors is the existence in most mammals of differences in their sexual behavior during the different seasons of the year. Such seasonal variations are frequently lost under uniform conditions of temperature and feeding as is the case with rabbits, and also with guinea-pigs if these show seasonal changes in their native wild.

It has been reported by some investigators, Rubaschkin and others, that guinea-pigs in captivity breed less frequently in winter than during the warmer months, though they may become pregnant at any season. Such results are probably due to a failure to keep the animals properly warm during winter.

Guinea-pigs under the uniform conditions of our experiments do not show any apparent changes in their sexual rhythm with the seasons, but as indicated on previous pages, it is probable that their sexual cycle is a little shorter during the summer than in winter, yet even this difference does nOt seem to be very definitely expressed.


The above description of the details of the oestrous cycle in the guinea-pig may be briefly summarized as follows:

1. Guinea-pigs kept in a state of domestication and under steady environmental conditions possess a regular dioestrous cycle repeating itself in non-pregnant females about every sixteen days throughout the entire year with probably small and insignificant variations during the different seasons.

2. During each cycle typically corresponding changes are occurring in the vagina, the uterus, and the ovary; a given stage in one of these organs closely accompanying parallel stages in the other two.

3. Each period of sexual activity lasts about twenty-four hours and is characterized by the presence of a definite vaginal


fluid, which is not sufficiently abundant to be readily detected on the vulva but is easily observed by an examination of the interior of the vagina.

4. The composition of the vaginal fluid changes with the several stages of change occurring in the uterus and vagina.

a. To begin with, during what we term the first stage, the fluid consists of an abundant mucous secretion containing great numbers of desquamated vaginal epithelial cells. At this time sections of the vagina show an active shedding or desquamation of its epithelial lining cells. The cells of the uterine epithelium are loaded with mucus, and an active migration of polynuclear leucocytes is taking place from the vessels of the vagina and uterus out into the stroma and towards the epithelial layer.

b. During the second stage the contents of the vagina become thick and cheese-like on account of the great accumulation of desquamated epithelial cells. The walls of the uterus and vagina become congested and the migration of leucocytes becomes still more active.

c. The leucocytes reach the epithelium and vigorously invade its cells and intercellular spaces during the third stage. These wandering cells become enclosed within and apparently dissolve the breaking-down dead cells of the epithelium. The vaginal fluid becomes thinner under the dissolving or digesting action of the leucocytes. The congestion in the uterus and vagina becomes still more pronounced giving rise to small blood masses or haematomata beneath the epithelium. The epithelium of the uterus is highly disorganized, vacuolized and richly invaded by the leucocytes, so that portions of it tall away en masse actually carrying with it in some cases cells of the stroma.

d. The fourth stage is merely a continuation or result of the acti\dties of the third. The falling away of the epithelial pieces and stroma cells permits the escape of the small haematomata or blood knots thus causing a slight bleeding into the lumen of the uterus and vagina. These traces of blood often give a redish aspect to the vaginal fluid. At this same stage a regeneration process begins from the necks of the uterine glands and also apparently from the epithelial infoldings in the vagina, so that


the lost epithelium becomes rapidly replaced almost before it has ceased falling away. If one may picture the epithelial surface of the uterus and vagina as consisting of innumerable prominences and depressions, it may be said that the destructive processes mentioned above are largely confined to the epithelium covering the prominences and that this epithelium is finally restored by regeneration from the epithelium lining the depressions, or in the case of the uterus from the epithelium of the uterine glands. The congestion with the diapedesis of corpuscles and the formation of the blood haematomata and the great accumulation of leucocytes all occur chiefly in the outpushed or protruding parts of the uterine wall.

The regeneration process in the guinea-pig is very short, lasting only a few hours, from six to twelve in all.

5. Ovulation seems to occur spontaneously during every heat period without exception. The rupture of the follicles with the consequent ovulation takes place about the end of the second stage or the beginning of the third ; that is, during the presence of the thick cheese-like vaginal fluid.

6. During the dioestrum or intermenstrual period there is very little fluid to be found in the vagina. This scant fluid consists of mucus in which are some atypical squamous cells from the vaginal wall and many leucocytes. A number of the leucocytes are old but there are probably new ones arriving almost continuously from the wall of the vagina. The only time at which the vagina seems to be practically Iree of leucocytes is immediately before and during the first and second stages of the oestrous period described above.

7. A marked correlation exists between the oestrous changes in the uterus and the developmental cycle of the corpora lutea. When the corpora lutea are highly developed and apparently active the mucosae of the uterus and vagina show a normally vigorous and healthy condition. While, on the other hand, when the corpora lutea begin to degenerate during the second week after the 'heat period' the mucosae ot the uterus and vagina also begin to show signs of degeneration and the process of desquamation slowly commences. At about two weeks after


the last 'heat period,' when the wholesale destruction of the mucosa begins, the corpora lutea are almost completely degenerated. The breaking of the Graafian follicles occurs during the oestrus as a result of a congestion which began in the theca folliculi at about the same time as the congestion of the stroma of the uterus and vagina. And finally when the regenerative growth of the uterine mucosa sets in, the ovaries then possess new corpora lutea, in an active state of differentiation, which were derived from the recently ruptured follicles.

It, therefore, might be imagined that the secretion from the corpora lutea exerts a protective influence over the uterus and vagina while the absence of this secretion permits the breaking down and degeneration of the uterine epithelium typical of the 'heat period.'



I^KAUD, J. 1897 The span of gestation and the cause of birth. A study of the critical period and its effects in mammals. Fischer, Jena. 1898 The rhythm of reproduction in mammalia. Anat. Anz., vol. 14.

BiscHOFF, Th. L. W. 1844 Beweis der von der Begattung unabhjingigen periodischen Reifung und Loslosung der Eier der Siiugethiere und des Menschen. Giessen.

1852 Entwickelungsgeschichte des Meerschweinchens. Giessen. 1870 Neue Beobachtungen zur Entwickelungsgeschichte des Meerschweinchens. Abh. der Konigl. Bayer. Akad. der Wissensch., vol. 10.

Blair-Bell, W. 1908 Menstruation and its relationship to the calcium metabolism. Proc. Roy. Soc. of Med., Obstet. and Gync. Sec, July, p. 291.

Bouin, p. et Ancel, p. 1909 Sur les homologies et la signification des glandes a secretion interne de I'ovaire. C. R. Soc. Le Biol., vol. 67.

1910 Recherches sur les fonctions du corps jaune gestatif. Jour, de Physl. et de path, gener.

Corner, G. W., and Amsbaugh, A. E. 1917 Oestrus and ovulation in swine.

Anat. Rec, vol. 12. Fraenkel, L. 1903 Die Function des Corpus luteum. Arch, fiir Gynaekol.,

vol. 68. Halban, J. 1911 Zur Lehre von der Menstruation — Protective Wirkung der

Keimdriisen auf Brunst und Menstruation. Zentralbl. f. Gynaekol.,

vol. 35. Halban, J., und Kohler, R. 1914 Die Beziechungen zwischen Corpus luteum

und Menstruation. Arch, of Gynakol., vol. 103. Heape, W. 1899 The menstruation and ovulation of monkeys and the human

female. Trans. Obstet. Soc, vol. 40.

1900 The sexual season. Quart. Jour. Mic. Sc, vol. 44.

1905 Ovulation and degeneration of ova in rabbits. Pi'oc Roy. Soc

London, vol. 76 B. Hensen, V. 1876 Beobachtungen iiber die Befruchtung und Entwickelung des

Kaninchens und Meerschweinchens. Zeit. f. Anat. u. Entwick, vol. 1. Kirkham, W. B. 1910 Ovulation in mammals with special reference to the

mouse and rat. Biol. Bull., vol. 18. Konigstein, H. 1907 Die Veriinderungen der Genitalsschleimhaut wahrend

der Graviditat und Brunst bei einigen Nagern. Arch. f. Physiol.,

vol. 119. Lams, H. 1913 Etude de I'oeuf de Cobaye aux premiers stades de I'embryo genese. Arch, de Biol., vol. 28. Loeb, L. 1911 a tJber die Bedeutung des Corpus luteum fiir die Periodizitjit

des sexuellen Zyklus beim weiblichen Siiugetierorganismus. Deutsche

Mediz. Wochensch. No. 1.

1911 b The cyclic changes in the ovary of the guinea-pig. Jour. Morph., vol. 22.

1911 c The cj^clic changes in the mammalian ovary. Proc Am. Phil. Soc, vol. 50.


LoNGLEY, W. H. 1911 The maturation of the egg and ovulation in the domestic cat. Am. Jour. Anat., vol. 12.

Marshall, F. H. A. 1903 The oestrous cycle and the formation of the corpus hiteum in the sheep. Phil. Trans. B., vol. 196.

1904 The oestrous cycle in the common ferret. Quart. Jour. Mic. Sc, vol. 48. 1910 The physiology of reproduction. London, 1910.

Marshall, F. H. A., and Jolly, W. A. 1905 The oestrous cycle in the dog. Phil. Trans. B., vol. 198.

Marshall, F. H. A., and Runciman, J. G. 1914 On the ovarian factor concerned in the recurrence of the oestrous cycle. Jour. Physiol., vol. 49.

Pearl, R., and Surface, F. M. 1914 On the effect of Corpus luteum substance upon ovulation in the fowl. Jour. Biol. Chem., vol. 19.

Prenant, a. 1898 De la valeur morphologigue du corps jaune, son action physiologique et therapeutique possible. Rev. gener. d. Sciences pur. et appl., vol. 9.

Reichert, C. B. 1861 Beitrage zur Entwicklungsgeschichte des Meerschweinchens. Abh. d. Kgl. Preuss. Akad. d. Wiessensch, Berlin.

Rein, 1883 Beitrage zur Kenntnis der Reifungserscheinungen und Befruchtungsvorgiinge am Saugetierei. Arch. f. Mikr. Anat., vol. 22.

RuBASCHKiN, W. 1905 tJber die Reifungs- — und Befruchtungsprocesse des Meerschweincheneies. Anat. Hefte., vol. 29.

Ruge, C. 1913 Ueber Ovulation, Corpus luteum und Menstruation. Arch. f. Gynakol., vol. 100.

Sandes and Skrobansky (Quoted from Oppenheim's Handbuch der Biochemie III, 1, p. 377).

ScHULz, 1829 Observationes de cobayae. Hist. Nat. Diss. Berlin (quoted from Bischoff).

Stockard, C. R. 1912 An experimental study of racial degeneration in mammals treated with alcohcrl. Arch. Internal Med., vol. 10. 1913 The effect on the offspring of intoxicating the male parent and the transmission of the defects to subsequent generations. Am. Nat., vol. 47.

Stockard, C. R., and Papanicolaou, G. N. 1916 A further analysis of the hereditary transmission of degeneracy and deformities by the descendants of alcoholized mammals. Am. Nat., vol. 50.




The figures in all of the plates are photomicrographs made by Mr. Wm. Dunn of the Photographic Department of Cornell Medical School.

1 Squamous epithelial cells contained in the vaginal fluid during the first stage ()f oestrus from animal 1089 9 . The vaginal fluid at this time is mucus filled with abundant cells of this type.

2 Cells from the second stage vaginal fluid. The great majority are squamous epithelial cells from the wall of the vagina with a few uterine epithelial cells. From animal 1066 9 .

3 and 4 Cells of the second stage more highly magnified from 1066 9 .












5 A smear of the fluid during the third stage, from animal 1104 9 . This shows the arrival of myriads of leucocytes among the epithelial cells in the vaginal fluid. Such an appearance is characteristic of the third stage.

6 A more highly magnified view of the same stage showing in clearer detail the cell structures.


DIOESTROUS CYChH IN THE GUINEA-PIG rnABLEs n. stockaud and g. x. Papanicolaou





7 and 8 Highly magnified epithelial cells containing many leucocytes within their cell-bodies. A condition typical of the third stage^ — also from 1104 9 .

9 A smear showing the presence of red blood corpuscles in the vaginal fluid during the short period of hemorrhage, following the third stage. From animal 1099 9 . re, red corpuscles.





^ a ♦












■^ t#^ %.





10 A section of the resting uterus during dioestrum, four and one-half days "after oestrus, showing the normal cuboidal ciliated epithelium — animal 1074 9 .

11, 12 and 13 Sections showing the condition of the uterine epithelium during its active secretion of mucus and the beginning of the leucocyte migration, from animal 1089 9 in which the oestrus was just commencing — leu, leucocytes. Note the contrast with figure 10. A corresponding smear of the vaginal fluid from the same animal just before it was killed is shown by figure 1.







^5 -S-V

^"V^L'"' s& ■<«^ -<^ €> if^ -^Si











14 A section illustrating the condition of the uterine epithelium and the accumulation of large numbers of leucocytes below the epithelium during the second stage of oestrus, from animal 1066 9 . Corresponding smears of the vaginal fluid at this time are shown in figures 2, 3, and 4 from the same female.

15 and 16 Sections of the uterus during the third stage of oestrus showing the invasion of the epithelium by migrating leucocytes. The epithelium is partially destroyed and greatly vacuolized, as a result of the dissolving action of the leucocytes, but is still adherent to the underlying stroma which also contains leucocytes, leu, leucocytes. Both sections are from 1104 9 and corresponding smears of the vaginal fluid from this animal immediately before being killed are shown in figures 5, 6, 7 and 8.

17 A section of the wall of the vagina from the same animal, 1104 9 , during, of course, the same stage. The vaginal mucosa is also invaded by leucocytes in a manner similar to that of the uterus, several epithelial cells are seen to contain leucocytes within their bodies. The epithelium here is being desquamated or thrown off while the uterine epithelium is seen to be disintegrating before being shed, leu, leucocytes.






^*f ^:'



<^->.^ '





.. / ;. . N» or."--* ' K V "^ "



/ .1



I'LATl-; ()


18 A section of the uterus from animal 1099 9 during the fourth stage, the short period of slight hemorrhage. The beginning regeneration of new epithelium from the neck of a uterine gland is shown while simultaneously the breaking down of the old epithelium is still taking place, and other portions of this section show a loss of the old epithelium from the uterine wall. A smear of the vaginal fluid from the same animal just before killing is shown in figure 9.

19 A similar section from the uterus of another animal, 860 9 , during the same stage. This shows better the falling off of the old epithelium and the simultaneous formation of new epithelium.








20 A section of ovary from animal 1089 9 killed during the first stage of oestrus. A ripe follicle is shown a few hours before congestion of the theca begins. A smear of the vaginal fluid from the same animal is seen in figure 1 and sections of the uterus in figures 11, 12 and 13.

21 A higher magnification of the ovum and follicular wall shown in figure 20.

22 A section of the ovary from 1066 9 killed during the second stage of oestrus. The theca folliculi surrounding the ripe follicle has become highly congested, bv, blood vessels.

23 Shows at a higher magnification a clearer view of the congested condition of the follicle in figure 22, bv, blood vessels. The nucleus of the ovum is in a resting condition. Corresponding vaginal smears from this animal 1066 9 just before being killed are illustrated in figures 2, 3 and 4, and a section through the uterus in figure 1-1. All of these figures illustrate commonly seen second stage conditions.











24 A section of the ovary from 1086 9 showing a follicle shortly after rupture. The congestion in the theca folliculi is evident, bv, blood vessels. This animal was killed at the end of the second stage or early beginning of the leucocytosis, the third stage.

25 A degenerating atretic follicle from the same ovary as figure 24, the cells of the cumulus ociphorus are degenerating while the follicle is being invaded by leucocytes. The ovum shows the first polar body in process of division while the nucleus of the egg is represented by a small chromatic mass near the center.

26 An early corpus luteum from animal 1104 ? killed during the third stage. Near the corpus luteum is seen a degenerating atretic follicle invaded by leucocytes. Compare smears figures 5, 6, 7 and 8, and sections of uterus figures 15 and 16, and section of vagina figure 17, all from the same animal.

27 A somewhat older corpus luteum from 1099 9 killed during the hemorrhage stage. The vascularization of the corpus is apparent at the periphery and is growing toward the center, hv, blood vessels. Compare the smear in figure 9, and section of the uterus figure 18.












(• «t- V** '

3- ^: •■ V- ^ '^ - :; %.^"^'--^-- ■ "■ ' • -^ ■■■• ^






28 A higher magnification of another corpus luteum from the same ovary as figure 27. The ingrowth of the peripheral vessels is more apparent, hv, blood vessels.

29 A fully developed corpus luteum from animal 1074 9 , killed four and onehalf days after oestrus. The typical glandular structure is clearly shown, cords of cells surrounded by capillaries, cap, capillaries. A section of the wall of the resting uterus from the same animal is given in figure 10.









Department of Anatomy, University of Virginia



I. Introduction 285

II. Histologic methods 288

III. Descriptive 289

a. The ventricular myocardium 289

b. The atrial myocardium 298

c. The moderator band 301

d. The atrioventricular connecting bundle 303

e. The fibers of Purkinje 306

f. The fetal myocardium 308

IV. Discussion 314

V. Summary of results 324

VI. Literature cited 328


This investigation involves a detailed microscopic study of the fetal and adult myocardium of the beef, including the atrioventricular bundle of His, the moderator band, and the fibers of Purkinje. The end in view is to test, in the light of additional data, the four chief hypotheses regarding the significance of the intercalated discs: (1) That they are intercellular cement substance (Schwigger-Seidel,^ Eberth,^ Zimmerman (24), et al.); (2) that they are regions of muscle growth, that is, differentiating sarcomeres (Heidenhain (4) ) ; (3) that they are of the nature of tendons (Marceau (19) ) ; and (4) that they represent local modifications of the myofibrils, of the nature of irreversible contraction phenomena following unusual functional conditions or

1 For bibliography, and discussion of the early literature, see Jordan (5).



stresses, in essence, irreversible contraction bands (Jordan and Steele (14) ), Dietrich (3) has proposed what appears to be a modification of Marceau's original interpretation in terms of a tendinous structure, namely, that the intercalated discs are constant structures formed during later myocardial histogenesis to provide for the functional coordination of previously incoordinated myofibril-bundles in the branching trabeculae of the muscle-plexus.

The investigation was begun with a study of the ventricular myocardium of the adult heart. The object was to discover the different types of discs with respect of intrinsic structure, and their varying relations to the constituent elements of the fiber: the telophragmata, nuclei, sarcolemma, etc. Comparative observations were made between the right and left ventricles, between the atria and ventricles, and between the papillary muscle, the columnae carneae, and the general myocardium, in order to determine structural and numerical variations. Study was then directed to the atrioventricular bundle, and especially to the area of transition between its termination as Purkinje's fibers and the myocardium, with the expectation of finding here some further clue to the significance of the discs. This expectation was to a considerable degree realized as will be described below. The study was completed by an examination of young and fetal hearts, in an attempt to discover the time and mode of origin of the discs.

The least tenable of the above-mentioned hypotheses appears to be that of Heidenhain, namely, that the intercalated discs are developing sarcomeres. It fails by reason of the facts, chiefly, that the discs have a definite developmental history of their own, that they do not occur in their definitive condition during the stages of very rapid earlier fetal growth, that the sarcomeres of fetal myocardium do not resemble the initial discs, and that they do not disappear after the heart has attained its maximum physiologic development. Heidenhain' s explanation that the persistent discs in the lull-grown heart may be of the nature of developmental vestiges, somewhat like the epiphyseal lines of long bones, seems inapplicable. Moreover, the different types


of discs can not be ranged into a consecutive series leading to a completely differentiated sarcomere.

Marceau's interpretation is at first consideration more plausible, especially in view of the facts that the discs very generally divide areas of different physiologic states, and that, at least in certain arthropod muscles, e.g., leg muscle of sea-spider and scorpion (Jordan (10 and 12) ), the tendon fibrils apparently differentiate from original myofibrils; but it meets with the objection that the discs do not react to specific stains for tendinous tissue, e.g., van Gieson's stain (and the Bielschowsky technic; Dietrich), and that the discs frequently lie within either contracted or relaxed areas.

Dietrich's interpretation of the discs as coordination mechanisms is no more than a suggestion, and no direct evidence is given in its support.

The original interpretation of the discs as intercellular cementsubstance finds good support only in the fact that macerated myocardium dissociates into elements bounded by discs and sarcolemma. But these elements do not closely resemble the stellate and fusiform cell-areas of the original embryonic myocardial syncytium, nor the fusiform elements of the early fetal myocardium. Furthermore, the earlier fetal heart is composed of anastomosing, branched, cylindric trabeculae, forming a continuous network, apparently without sign of typical discs. After prolonged maceration the heart muscle fragments also along the telophi'agmata. The intercalated discs are always associated in some manner with the telophragmata, hence in fragmenting myocardium the plane of fracture must necessarily frequently involve a disc. When we add to these facts the probability that the discs as modified portions of the myofibrils are lines of relative weakness, the behavior of the rnacerating myocardium becomes readily comprehensible.

The conduct of the discs towards silver nitrate solutions also need not necessarily indicate an essential intercellular cementsubstance as a constituent of the discs. It may mean only that the discs are regions of relatively greater abundance of the more fluid portion of the interfibrillar sarcoplasm, which may precipi


tate the silver nitrate. Observations are recorded below which indicate that the j^recipitation of silver nitrate within the discs is incidental to the presence of tissue-fluid in the interstices of the fundamental bacillary elements of the discs, which fluid has penetrated via the telophragmata from the tissue spaces between the fibers. In myocardium treated with silver nitrate the telophragmata also precipitate the salt and appear more deeply colored. The ready passage of tissue-fluid along the telophragma is provided for by the close union between telophragmata and sarcolemma.

The three most widely prevalent hypotheses above discussed meet \vith such serious objections when thoroughly analyzed and strictly applied that they must be abandoned as complete interpretations of the intercalated discs. It will be the chief burden of this new investigation to further support the hypothesis first suggested by Jordan and Steele (14) that the discs are of the nature of irreversible contraction bands. The suggestion had frequently been made by various investigators that the intercalated discs are related in some manner to contraction phenomena, but their specific interpretation as modified irreversible contraction bands had not been previously proposed.


The tissues were in every case fixed in the nitric-acid-alcohol mixture of Zimmermann (24). Parallel series of sections were prepared according to Zimmermann's hemalum-staining method, and with the iron-hematoxylin-van Gieson combination. Dissociated tissues were also prepared for study by maceration with potassium hydroxid, and staining on the slide with a dilute solution of methylene blue. Ventricular tissue was treated also with silver nitrate solutions for study of possible intercellular cement. Beautiful and most instructive preparations were made also by teasing hemalum-stained blocks of tissue, and mounting the fragments in glycerin on the slide. This last technic may be very highly recommended as a simple routine laboratory method for class demonstration of intercalated discs. Not only


the intercalated discs, but also the telophraginata, and the isotropic and anisotropic substances, stand forth with almost the same sharpness and clearness as in sections.


a. The ventricular myocardium

There are no striking numerical or structural differences between the intercalated discs of the right and left ventricles. Nor do appreciable differences occur between the atria and ventricles, contrary to the opinion of Werner (23). As regards the ventricular wall, the intercalated discs appear somewhat more numerous in the papillary muscles and in the moderator band, than in the more peripheral myocardium. Moreover, the discs of the moderator band, and to some extent those also of the papillary muscles and the columnae carneae, are less complicated structures, that is, they are more generally of the simple band form. The numerical difference may inhere largely in the fact of less coarse and therefore relatively more abundant trabeculae in the papillary musculature. The structural differences are probably incidental to the generally different disposition of the branches at wider angles with respect to the coarser trabeculae, thus producing more oblique stresses during contraction, and to the spiral twistings of the muscle fibers during development and growth, in the ventricular myocardium. Neither the numerical nor the structural differences, however, have fundamental significance. Structural differences are largely the result of secondary modifications of originally very similar and simple discs. We may quite securely begin the description of the structural variations of the discs wdth the general proposition that they are in all parts of the heart-musculature essentially of the same nature, variety, and abundance. Such variations as occur in normal and pathologic hearts are incidental respectively to normal and modified functional activity.

The simplest type of disc in the adult heart is similar to those which first appear in the fetal heart, and resembles a peripheral


deeply staining band in series with the telophragmata, composed of modified bacillary segments of the included portions of the involved peripheral myofibrils (fig. 1). The modification shows itself chiefly in an enhanced tingibility in certain stains, e.g., hemalum. The modification is apparently, fundamentally, chiefly chemical. Such a disc is originally bisected by a telophragma. In the case of certain of the simpler discs in which the telophragma is not discernible, they appear to shade laterally into a telophragma or abut upon it at the lateral mid-point (fig. 2). That the discs are peripheral structures for the most part can be demonstrated by changing the level of focus, when the disc either disappears from the field or can be traced in a lateral or spiral direction to an underlying or overlying surface. The same fact can be even better demonstrated in transverse sections. In figure 7 is shown a simple disc involving only the peripheral myofibrils of the radial lamellae in approximately a quarter of the circumference. In figure 8 two discs appear, one internal to the peripheral element. In both sections the discs are at the same level as the nucleus. Figure 25 shows a scattering of smaller discs throughout the fiber, probably the result of a spiral twisting which caused an inturning of portions of originally peripheral discs.

In figure 1 are shown two successive discs in series with the telophragmata. Their location and general structure agrees with that of contraction bands. The upper two discs in figure 2 show the same structure and relationships. In figure 1 both discs cover the entire breadth (transverse) of the fiber. Discs may be of much lesser breadth, indeed including only a single fibril ; but narrower discs may also in certain cases represent transverse sections of broader discs, e.g., as in figure 9. Again, similar discs of lesser width (longitudinal) occur. These represent the original condition, both phylogenetically and ontogenetically (Jordan and Steele (14) ), the wdder discs being a modification resulting from a traction produced by the contracting myocardium. That the discs are subjected to the modifying influence of a traction is indicated also by the frequently constricted condition of the fiber in the region where the discs are located (fig. 1).


In figure 2 a terraced or step-form of disc is shown in connection with simpler discs of the character above described. That the latter type may be bisected by the telophragma is demonstrated by the manner of the attachment of the sarcolemma festoons. The same point is even more clearly demonstrated in teased preparations. The figure illustrates also another common feature in connection with the discs, namely, the division of a contracted from a relaxed area along the line of the discs. In the upper relaxed region are seen the telophragmata, the Q-discs and the J-discs; in the lower contracted region delicate contraction bands alternate with lighter discs. The contracted region stains more intensely than the non-contracted region. The discs evidently frequently act as barriers to the spread of a particular physiologic state; but occasionally the discs are crossed by functional phases, and so may lie in either contracted or relaxed areas.

It should be noted also that in the terraced portion of this complex disc the successive steps are so arranged that the upper border of any one is in line with the lower border of the next higher disc, and the left hand border of any one is in line with right hand border of the next higher disc ; that is, the arrangement is such as would result if the several steps had originally formed portions of the same continuous band at the upper level and had been divided into smaller sections, which were subsequently drawn to successively lower sarcomeric levels in a lateral progression. Moreover, certain terraces are united by a deeper-staining membrane or 'riser;' and the relation of the involved telophragmata is such that the membranes of opposite sides join opposite surfaces of the discs. Such discs are common (figs. 5, 9 and 13), and the more general condition of terraced discs with respect to the association of the steps and the included telophragmata is like the one here described. But several chief variations occur: (1) The terraces may ascend again foUownig a descent (fig. 27) ; (2) all of the levels need not be placed in the regular order above described (fig. 9); and (3) the telophragmata may be similarly placed on both sides of the discs. Illustration figure 2 shows further the usual location of the discs at points where the coarser


trabeculae branch. Terraced discs arise in at least two different ways: (1) As dislocations of original band-forms following functional or developmental stresses; (2) as concomitants of a fusion along an oblique surface of the two originally discrete portions of the myocardial plexus. The methods of the original formation of the several types of terraced discs will be further described and discussed below.

Figure 3 illustrates the opposite surfaces of the same fiber. At the upper level of focus (a) the disc appears of the usual simple band-form, composed of modified portions of the involved myofibrils, in series with the telophragmata. There is no evidence that this disc is. bounded on either side by a telophragma. In passing to the opposite surface the disc appears distorted, as if by opposed stresses, in such a manner as to form a two-step disc. A common telophragma bounds the lower border of the left segment and the upper border of the right segment.

In figure 4 is shown a similarly dislocated disc, the two segments having been moved somewhat farther apart, and having remained connected by a deeply-staining membrane, probably portion of a telophragma.

Another complex type of disc is illustrated in figure 5. The band elements shade into the telophragmata. The steps are interconnected by membranes. The different levels of location of the several portions are indicated by numerals. The disc as a whole has an interrupted spiral form, and bounds a wedge-shaped lighter-staining area at the left.

The occasional super-nuclear position of the disc is illustratred in figures 6, 7, 8 and 31. Figure 6 shows also the close union of the telophragmata with the sarcolemma and the nuclear wall.

Figure 9 illustrates clearly one manner of the formation of terraced discs. Here two originally discrete muscular trabeculae have fused. The 'risers' or connecting membranes of this complex disc have resulted from the fusion of the apposed sarcolemmae. An irregularly terraced disc resulted in consequence, the several segments having been contributed in part by one, in part by the other fiber. Since the fusion was such as to produce disaccordauce of the apposed sarcomeres, the discs became ar


ranged with respect to the telophragniata so that the opposite telophragniata joined opposite (upper and lower) borders of the disc-sections. In anticipation of the ensuing discussion it may be stated here that the fundamental causal factor in the formation of this terraced disc is believed to be the unusual stresses imposed upon the peripheral myofibrils in the region of the area of fusion, incidental to the functional re coordination required of the fibrils. The location of discs generally near the levels where branches arise also becomes comprehensible under this hypothesis.

The band-forms in figure 9 are located at telophragniata levels. In the upper portion of the field one lies superjacent to a nucleus. This same disc extends for some distance into the adjacent fiber. Such disposition of the broader discs, that is, a location across several fibers, is a common feature. It occvirs extensively even in the Limulus heart, w-here the discs are numerically rare and of the simplest 'comb' type (fig. 41). The condition indicates a local functional alteration influencing several adjacent fibers in a transverse plane. Such discs can be plausibly interpreted on no hypothesis involving growth phenomena, intercellular cement, or tendinous structures. They appear to signif}^ identical modifications resulting from identical functional phases at the same transverse level of the heart musculature.

Fusion of two adjacent trabeculae is further illustrated in figure 10. Here two groups of narrow band-discs occur, connected by the fused sarcolemmae. The formation of discs is ey\dently closely associated with the processes of fusion among fibers. But the location of the discs with respect to the surface of fusion is a matter of fundamental significance. The point is wtII illustrated in both of the figures 9 and 10. The discs do not lie in the line of fusion but at riglit (or oblique) angles to it, and in ahnement with the telophragmata. It is readily conceivable that the fusion of the fibers involved a functional recoordination of gi'oups of peripheral myofibrils. This produced unusual strains at certain levels. Such levels offer, theoretically, favorable sites for the formation of discs by process of modification of contraction bands (essentially an irreversibility) according to the hj^^othesis here adopted and discussed below.


Iroiii the standpoint of disc-fovniation fusions, however, are of two sorts: (1) such as furnish the causal factor; and (2) such as simply distort, displace, or modify in some way, discs already present in the fibers involved in the fusion. The second sort is illustrated semidiagrammatically in figure 11. Here two fibers have become fused in such a manner as to produce an an harmonic alinement of telophragmata in the apposed fibers, the result of the superposition of a mutual spiral twisting around a common axis of the tw^o fibers. Such spiral twistings and fusions are common. They have been described also in scorpion voluntary striped muscle (Jordan (12) ) and in human heart muscle (Heidenhain (4) ). According to Heidenhain a similar condition results from the spiral twisting of a single fiber about its central axis (Plasma und Zelle," p. 616). The festooned sarcolemma, according to Heidenhain's interpretation, would here represent inturned portions of the originally peripheral membrane. The same condition would result, how^ever, if tw^o adjacent fibers fused in such a manner that the crest of a festoon of one side alternated with the trough between tW'O successive festoons on the apposed fiber. The latter method appears to be more common, though the former probably also occurs. At any rate the spiral twisting of the myocardial trabeculae during development and grow^th is a characteristic of the mammalian heart (e.g., bulbo-spiral bundle of fibers). Under these conditions the definitive position and relationships of the original discs is secondary, a modification resulting from the twisting and fusion of the fibers.

In figure 12 is illustrated a rare type of disc. Two fibers seem to have fused in an oblique plane, end to end. The terraced disc is explicable on the basis of our hj-pothesis of strain effects following unusual stresses, and resulting in irreversible contraction bands. The form of this particular disc may also be in part the result of a spiral twisting of the fiber.

In figure 13 a long terraced disc separates a deeper-staining (contracted) region sharply from a lighter-staining region. One of the intervals here, as frequently in such discs, between successive terraces is of the length of two sarcomeres. On raising


the \e\v\ of focus from level 2 to kn^el 1 the discs in the upper lighter region come into view. The latter are more delicate, stain only relatively faintly and shade into the telophragmata. This comjilex of discs may be interpreted in the light of the evidence derived from the simpler conditions in figures 9 and 10. The terraced disc probably formed along the oblique surface of fusion of two distinct trabeculae. The band-discs in this region probably formed in connection with this same fusion as incidental strain effects.

Special note must also be taken of the alterations of the telophi'agmata in this region. Accessory telophragmata appear to pass obliquely between three successive primary teloplu'agmata. The condition is probably the result of a rearrangement of the telophragmata and subsequent fusion following a spiral twisting of the fibers in this region. This group of discs no doubt suffered a secondary alteration incidental to the spiral twisting. The connecting accessory membranes between the three successive telophragmata demonstrate the telophragma-nature of certain of the 'risers' of the step-discs.

The arrangement of the telophragmata in this trabecula demonstrates, moreover, the possibility of a realinement of telophragmata following gross morphological changes in the trabeculae. It emphasizes also the imperative necessity, for a complete interpretation of the intercalated discs, of dissociating the fundamental structure, relation, and forms of the discs from their secondary mechanical alterations follo^\ing distortion in the trabeculae.

Figure 14 shows an unusually large number of discs within a relatively small area. The discs lie at different levels as indicated by the numerals. The majority shade laterally into telophragmata. Such an area would seem to defy interpretation in terms of intercellular cement, tendons, coordination mechanisms, or gi'owth areas.

Before proceeding further with the description and inter]:)retation it may be well to emphasize the following cardinal facts: (1) The embryonic heart-musculature is a syncytium composed of anastomosing stellate and fusiform myoblasts with continuous


myofibrils; tlie 'cells' elongate into fusiform elements, the constituent myofibrils meanwhile increasing in number, and subse(juently by fusion form a close-meshed network of delicate trabeculae (fig. 15) with still more delicate branches, for the most part originating at very acute angles; coarser trabeculae arise by growth and further fusions, their coarser branches coming off at more obtuse angles; the myocardial plexus may suffer still further local fusions, and becomes meanwhile subjected to the functional stresses of opposed and oblique tensions in part the consequence of a spiral twisting of its constituent fibers and branches. (2) Intercalated discs develop gradually during fetal life; they are from the beginning closely associated with the telophragmata, having the appearance of thickened membranes or portions of telophragmata; at this time the only conspicuous stripes are the telophragmata, which appear very delicate and irregular; these apparently developed out of the spongioplasm of the myoblasts, while the sarcolemma develops from the cell membrane; the discs are at first granular, and only subsequently show the typical comb structure. (3) The discs increase in size and number coincident with the pre- and post-natal development of the heart, the results respectively of a longitudinal splitting of the fibrils with their intercalated discs and a new formation of discs, and persist under modification throughout life; once formed the discs are apparentl}^ persistent structures subject to growth and extensive mechanical alterations. (4) The discs are peripherally placed, always in association wdth telophragmata, and with the sarcolemma. (5) The discs are more commonly located in the regions where the coarser trabeculae branch, and freciuently dixdde areas of different physiologic states. (6) The myofibrils pass \\dthout interruption through the discs; the discs are essentially modified portions of the involved myofibrils, among the structural units of which a relatively more abundant tissue fluid occurs. (7) In the simplest condition they are similar to the contraction bands both in structure and in their relation to the myofibrils and the telophragmata; in sections stained with iron-hematoxylin the narrower band-forms of discs and complete contraction bands


appear practically identical; if contraction bands are conceived to be rendered incapable of reversion to the relaxed condition, and as such to have become permanent structures modified under unequal functional tensions incidental to the branched and syncytial condition of heart musculature and the spiral t\\dsting of certain bundles of fibers during development, and the lateral fusion of such adjacent and mutually twisted fibers during growth, the derivation of the various definitive types of intercalated discs becomes clear.

With the above general features of the origin, structure and relation of the discs in mind we may now more profitably proceed to the further description of the various types of discs. But before doing so a critical estimate should be made of the value of sectioned material for the study of the character of the discs. Sections of 10 microns' thickness include many complete fibers, the diameter of the fibers being on the average from 10 to 15 microns. Accordingly little likelihood remains of misinterpretation on account of a peculiarity of the plane of section, or by reason of partial views. In order to reduce the theoretical disadvantages of sections to a minimum, comparative studies were made with teased material. In teased hemalum-stained fragments of the ventricle, mounted in glycerin, the discs appeared exactly as in the sections. Occasional discs may be seen in which a bisecting telophragma is conspicuous. These bandforms of discs are peripheral in position; a certain number have the form of short spirals; some are located superjacent to nuclei. By raising or lowering the level of focus an apparently short disc can occasionally be followed as a complete band (crescent) across the fiber, and in some cases even as a more or less complete ring or spiral to the opposite surface. Step-forms also are abundant in the teased material ; they are therefore not generally only the optical expression in sections of transverse cuts of a series of band discs, as might have been suspected; but they must either be due to dislocation of original band discs or they are originally formed as terraces. This material shows clearly also the frequent phenomenon of a division on the part of a disc of a contracted from a relaxed area.- The teased tissue is quite



as favorable for a study of the discs and striations as are sections; it reveals clearly also the H-discs at certain stages in the process of contraction (fig. 32).

h. The atrial myocardium

The several new types of discs to be described for the atrium must not be thought to be characteristic of this portion of the myocardium. Similar discs are found also in about equal profusion in the ventricles. They are considered in this order because in some respects this selected group represents a more complex condition.

In figure 16 is illustrated a simple and common type of disc. In so far however as these discs are bounded on both borders by a telophragma they represent a secondary modification of the original disc which is in contact with only one telophragma, which generally bisects the disc. Both of these discs moreover separate areas of different physiologic condition. Both discs have the width of a complete sarcomere. The one on the left is a two-step form, and probably arose from a dislocation through one sarcomere of an original band disc. In terms of a contraction band such a disc might conceivably be the result of the fusion of adjacent halves of successive bands, or of a secondary modification of a single band in such a manner as to cause it to spread to the adjacent telophragma.

Appearances like that illustrated in figure 18 strongly suggest that the former interpretation of discs of the width of a complete sarcomere as the result of a fusion of adjacent halves of successive contraction bands, is, at least in some cases, the correct one. Here, in figure 18, two discs appear within a single sarcomere, each bounded on opposite sides by a telophragma. If the disccondition is conceived to spread between the two moieties, such discs as are illustrated in figure 16 would be formed. Whatever the fundamental modifying factor may be that operates to convert the myofibrils to the disc-condition — whether of the natm'e of an irreversible contraction following uncommon stresses or not — it acts along the line of a telophragma, causing the myo


fibrils here to be altered on either one or both sides, or (perhaps only secondarily) between successive telophragmata.

The peculiar disc illustrated in figure 17 is of the same nature as those in figure 16. Here, however, a central segment appears dislocated, the three segments remaining interconnected by a deeply-staining membrane, probably the telophragma about which the original band disc originated.

It is significant that in the common form of terraced disc, the lateral surfaces of successive segments are in myofibril series, that is, the several segments do not overlap. This phenomenon is illustrated in the semidiagrammatic illustration, figure 19. The relatively wider space between successive disc-bundles further strongly indicates that the terraced condition arose from a dislocation of an original band-disc at the higher level. Such interpretation involves the assumption of a realinement and fusion of the disconnected telophragmata, for which appearances illustrated in figure 13 give some basis of fact. Moreover, in the less highly differentiated condition of the trabeculae in the younger heart, when such dislocations more probably occur, the telophragmata are relatively less rigid and less firmly attached to sarcolemma and nuclei (fig. 15).

Figure 20 includes discs at various stages of development, presumably from contraction bands. The uppermost one shades laterally into a telophi-agma; this disc is practically a contraction band, and in so far as contraction involves a segregation of Q-substances about a telophragma the disc is in part of anisotropic nature (Jordan, (5) ). Similarly the two discs next below are bands, in width more like the unmodified regular contraction bands of the muscle (figs. 2 and 30), The diminution in width laterally to the character of the telophragma, illustrated also in figures 5, 13 and 14, corresponds with the condition of the discs as seen in transverse section (figs. 7, 8 and 25). The succeeding three series of discs represent modifications of contraction bands in that they are placed to one side of their respective telophragmata, perhaps irreversible halves of contraction bands. The wide, short disc to the left is again of the modi


fied type, a sarcomere in width, perhaps fused adjacent halves of successive contraction bands.

Figure 21 illustrates a rare type of disc. It is of the simple band type, bounded on both borders by a telophragma, and shading on the right into the lateral extension of these membranes. Viewing the sarcomeric section as a unit, the modified disc portion appears contracted. Such a disc may seem to support Heidenhain's interpretation in terms of a differentiating sarcomere; but it can equally well be interpreted as the result of the only partial relaxation of adjacent halves of successive contraction bands, the disc representing the irreversible fused product on the left.

Figures 22 and 23 illustrate peculiar atypical forms of discs, several sarcomeres in width. In figure 23 the central element is a simple band-disc with which are associated outlying modified portions of fibrils. The entire disc-area here stains more deeply than adjacent portions of the trabecula, indicating a different physiologic condition in the disc region. Such discs seem impossible of interpretation with any degree of plausibihty on a basis of growth phenomena, tendinous fibrils, or cement substances.

The complex disc in figure 24 likemse defies interpretation in terms of any of these hypotheses. It is one of the most complicated types, and perhaps involves secondary distortions coincident with a spiral twisting and concomitant dislocations. Figm*e 25 possibly represents a transection of a similar fiber, in which portions of the originally peripheral band-discs have been transferred centrally through spiral twistings of the fiber. Similar discs are shown in figures 26 and 27, where the irregular condition is associated with a secondary longitudinal splitting of the primary trabeculae into smaller bundles. Such splitting of trabeculae, followed by a partially independent and opposed functional activity, may be a factor in the formation of these types of discs. The connecting deeply-staining membrane may be the distorted telophragma of the original disc, or perhaps an inturned portion of the sarcolemma.


c. The moderator band

The moderator band is of importance in this connection because of the abundant areas of transition it furnishes between branches of the atrioventricular bundle and the myocardium. The transition area includes fibers of the nature of the subendocardial Purkinje fibers. Moreover, the myocardial meshwork here seems of a slightly lesser degree of modification; that is, the fibers branch more regularly, and mainly dichotomously (fig. 29), at very acute angles, and the discs are predominantly of the simple band-type. The step-forms generally lack the deeply-staining connecting membranes, which indicates, in combination with other appearances, absence of a spiral twisting of the muscle trabeculae.

In transection the moderator band has an oval form (fig. 28). It is enveloped by a dense areolar connective-tissue capsule. It contains at opposite surfaces, just beneath the capsule and within a looser areolar connective tissue, two unequal divisions of the main right branch of the atrioventricular bundle. The muscle tissue also is collected into two unequal bundles, surrounded by a perimysial connective tissue layer. The larger bundle contains centrally a relatively large arteriole with two opposite large periarterial lymphatics, and a relatively small, more peripheral venous comes.

In the muscular portion of the moderator band contracted areas alternate quite regularly with non-contracted areas (fig. 30). The contracted areas show relatively coarse, granular, deep-staining contraction bands which alternate with wider, lighter-staining discs. In the relaxed intervals the telophragmata, Q-discs and J-discs are conspicuous. In certain deeperstained fibers additional H-discs appear (fig. 32). These latter fibers are in the early phases of contraction. The contracted regions stain more deeply, and have a considerably greater diameter than the relaxed intervals (fig. 30).

The muscle nuclei of the moderator jDand are located in fusiform sarcoplasmic areas. The myofibrils are arranged peripherally, and they are relatively less abundant than in the ventricular


trabeculae. These conditions, combined with the less difierentiated character of tlie nuclei, the form of the muscular mesh, and the prevailing form of the discs, all indicate a relatively less highly differentiated musculature. The sarcolemma also is more generally festooned (figs. 29, 32, 33 and 34). The telophragmata are in intimate union with both the sarcolemma and serrations of the nuclear membrane (fig. 34). There is not the sUghtest indication in any condition of an additional membrane, the alleged mesophragma (Heidenhain). Since its presence can not be demonstrated in the relatively coarse musculature of the beef heart, its occurrence in cardiac muscle seems doubtful.

The discs are generally of the narrow band-type, in close association with telophragmata (figs. 29, 31, 32 and 33). Stepforms occur,' but connecting membranes ('risers') appear to be lacking. There is some evidence of persisting amitotic division of the nuclei in this region. The moderator band here described is relatively slender, and the heart is that of a young, almost full-grown, beef.

The musculature of the moderator band furnishes an exceptionally favorable opportunity for testing the conclusion that discs occasionally lie superjacent to the nuclei. The difficulty of establishing this fact in tissues where abundant branches arise from all surfaces of a main cylindric trabecula is fully appreciated. But the illustrations given in support are of examples where no doubt can remain (figs. 6, 7, 8 and 9). The fact is, if possible, still more certain in figure 31; here the supernuclear group of discs has no relation to anastomoses with extraneous branches. Identical evidence accrues also from a study of teased preparations. Finally, the group of discs shown in figure 31 admits of no interpretation except in terms of a supernuclear location within the 'cell-area' represented by this nucleus.

The possible suggestion that the discs represent an original intercellular substance (plus apposed cell-membranes), into which a nucleus has migrated, can have no value as an argument for the intercellular hypothesis of intercalated discs, since, aside from their peripheral location, they could not as true intercellular


cement-substances be normally and constantly pierced by nuclei.

The peripheral position of the discs is well illustrated also in the type shown in figure 33, where on the upper surface the disc is in series with the telophragmata, while on the lower surface it is bounded along both borders by these membranes, having here suffered a slight spiral distortion.

d. The atrioventricular connecting bundle

Before describing the transition from Purkinje fibers to the cardiac muscle, it becomes necessary briefly to describe the structure of the atrioventricular bundle. This has already been done more or less completely by Tawara (22), by De Witt (2), by Lhamon (17) and by King (15), and we shall touch only certain details which relate themselves to our investigation of the intercalated discs. The atrioventricular bundle is distinctly cellular in structure. This conclusion is in accord with the descriptions of all of the above-named investigators except De Witt, who regards the bundle as a syncytium. Moreover, all agree that certain of the myofibrils of the constituent 'cells' have an unbroken course through the intercellular spaces. Agreement is complete also with respect to the descriptions of the shape of the cells, as somewhat modified spherical or polyhedral elements with a crenated or serrated contour. The bi- tri- or quadrinucleated condition of the cells has also been noted. In general our findings agree closely with those of Tawara (22) and of King (15).

In stained sections the constituent cells of the atrioventricular bundle are conspicuous (fig. 35). Their borders appear serrated, the myofibrils pass directly without apparent modification through the intercellular spaces, and the majority of the cells are binucleated (figs. 35, 36, 37, 42, 43 and 50). The several divisions of the branches of the atrioventricular bundle are enveloped by a dense fibroelastic capsule, between which and the muscular columns occurs a space (figs. 43 and 45). This had been previously noted by Tawara (22), by Curran (1), by


Lhamon (17) and by King (15). It may represent a lymphoid space, but we can find no evidence of lining epithelial-cells, in which result we are in agreement with Lhamon and with King.

The latter two investigators demonstrated the continuity of the sheath and the enclosed lymphoid space' throughout the entire bundle by means of injections with india-ink and Prussian blue. Injection of silver-nitrate solution failed to reveal lining cells. Lhamon (17) concludes that in hearts of beef, calf and sheep the sheath does not simulate, except perhaps very remotely, a mucous bursa, as claimed by Curran; and that it is not a part of the lymphatic system of the heart.

The nuclei of the cells are located centrally, within a finelygranular sarcoplasmic area free of myofibrils. The two nuclei are almost invariably in very close apposition; frequently flattened along the apposed, surfaces. They arise chiefly by amitotic division of a single nucleus of the original cell, a process which can be observed in fetal hearts of from two to four months. A few nuclei were observed in the segmented spireme condition in the tw^o-month fetal heart, which would seem to indicate that mitotic division also may occur in the earlier stages. In this respect the bundle simply agrees with ordinary myocardium, where nuclear division is originally mitotic and subsequently becomes exclusively amitotic. The tri- and quadri-nucleated condition of the bundle cells follows a later similar amitotic event.

The myofibrils are relatively sparse, but are more closely aggregated peripherally. They are collected in smaller irregular bundles, which peripherally are generally arranged parallel with the borders of the cell. The telophragmata are conspicuous among the bundles; between the bundles they appear more delicate, distorted, and frequently interrupted.

The serrations in the enveloping sarcolemma are fixation artifacts, due to the close union between telophragmata and sarcolemma and the unequal shrinkage in fixation between the myofibril bundles and the sarcolemma. They are the homologues of the sarcolemma festoons of the myocardium. The serrated condition is rendered still more conspicuous in stained preparations by reason of the fact that the stain penetrates more pro


fuscly for a short distance the peripheral ends of the telophragniata (figs. 36 and 37). Between adjacent cells are larger and smaller intercellular spaces (figs. 36 and 50) ; through the intervening 'intercellular bridges' pass the myofibrils. Figure 43 shows the various shapes of the cells in cross-section. That the above interpretation of the serrations of the cell-borders is correct is further demonstrated by the appearance of the cells in macerated preparations. Here the cells have a sharp contour (fig. 44) . The nuclei appear homogeneous, the cytoplasm finely granular; the myofibrils are indistinctly visible and very irregularly distributed.

In figure 42 is illustrated a peculiar condition where an arteriole appears to lie within the cell. This definitive condition is probably the result of a secondary adaptation of the cell to the growing blood vessel.

The various histologic conditions above described for the cells of the atrioventricular bundle indicate a relatively slight differentiation, or an embryonic condition. Such interpretation has frequently been given to the cells and their slightly modified forms, the Purkinje fibers. But that they actually represent embryonic forms of myocardial fibers, that is, that they are similar to the elements from which the myocardium develops has been disputed by certain investigators, e.g., Moenckeberg^ who points to the fact that in the human embryo they are already clearly differentiated froin the myocardium at the fifth fetal month. In the beef heart they can be readily distinguished already at the end of the second month. But the very close structural correspondence between the cells of the atrioventricular bundle and the myocardium in the two-month fetal heart of the beef very strongly suggests the interpretation of the three elements in terms essentially of a difference in degree of progressive differentiation. This point will be further discussed below.

The presence of myofibrils and especially of telophragmata in the cells of the atrioventricular bundle characterizes them as

- Cited from Lange (16).


muscular in nature. They however lack intercalated discs; the intercellular spaces and cement substance haVe no resemblance to intercalated discs. The atrioventricular bundle is originally and definitively cellular. The myocardium is both originally and definitively syncytial, and the intercalated discs arise as secondary modifications at certain levels of the trabeculae (transiently fusiform elements in the early fetal heart) in relation to telophragmata. The closest points of resemblance between the cells of the atrioventricular bundle and the trabeculae of the myocardium are the presence of myofibrils, and their continuity through intercellular spaces and intercalated discs respectively.

e. The fibers of Purkinje

We may now more profitably return to the description of the transition area between the atrioventricular bundle and the ventricular myocardium. Here we encounter the fibers of Purkinje. Tawara was the first to describe the continuity of the atrioventricular bundle with the Purkinje fibers. This observation has been repeatedly confirmed by other investigators; vide, e.g., Retzer (21). The Purkinje fibers (cells) are essentially identical with the cells of the atrioventricular bundle, only somewhat modified by elongation, fusions into fibers, and a higher degree of differentiation. The latter consists in a relatively greater abundance of myofibrils, which are more regularly disposed and less distinctly aggregated into smaller bundles, coarser and more conspicuous telophragmata, and the presence of simple band- and step-forms of intercalated discs (figs. 46 to 48 and 51).

The Purkinje-fiber transition-area is definitively a syncytium. This conclusion agrees with De Witt's (2) description of these fibers as forming a syncytium in man, dog, cat, sheep and calf. In the dog embryo De Witt describes the Purkinje fibers as composed of 'single short clear cells.' The definitive condition in the beef heart still gives evidence of the originally cellular structure of these Purkinje fibers (fig. 51).


Figure 45 shows a transection of one of the terminal branches of the left atrioventricular bundle where it passes under the endocardium to unite with the ventricular myocardium. The fibers are still enveloped by a connective tissue sheath enclosing a subjacent lymphoid space. The nuclei lie in a finely-granular, delicately-reticular, central, sarcoplasmic area of fusiform shape (figs. 46 to 48). The myofibrils are peripherally arranged, in general in delicate radial lamellae (fig. 47). Simple discs occur sparsely, in close connection with the telophragmata (figs. 46 and 48). Some are located at nuclear levels. Deeply-staining connecting-membranes may occur in the step-discs (fig. 48). They probably represent the fused sarcolemmae along the line of union of two cells which have fused in the formation of a fiber.

Figure 46 would at first consideration seem to furnish incontrovertible proof of the inadequacy of the interpretation of intercalated discs as intercellular cement substances, for here we have a single elongated cell upon which appear several intercalated discs, at least one of which is supernuclear in position. But conditions like that illustrated in figure 51 rob this illustration of its apparent finality in this connection, since it indicates that these discs may actually be related to lateral surfaces of fusion. This possibility, moreover, in part at least explains the supernuclear position of intercalated discs. An attempt will be made below to harmonize the apparent discrepancies here suggested.

When we pass now again to the moderator band, we find the same series of events. The cells of the atrioventricular bundle (fig. 50) pass more or less abruptly into Purkinje fibers (fig. 49, above) , and the latter by more gradual stages pass into the myocardial meshwork (fig. 49, below) where simple band-forms of discs appear.

The difference in shape of the nuclei in these several regions is also noteworthy. In the cells of the atrioventricular bundle the nuclei are generally spherical or stoutly oval, and paired; in the Purkinje fibers they are still substantially of the same shape,


but scattered; in the myocardial trabeculae they are relatively larger and elongated elements (fig. 49).

Figm*e 51 illustrates conditions at the level of transition between the Pur kin je fibers and the myocardium of the moderator band. The three cells shown are histologically of the Purkinje type, and are in process of fusion to form a fiber. Various short discs occur all along the surfaces of fusion. The connecting membrane represents the fused sarcolcmmae of the adjacent fibers. The discs appear to have arisen peripherally in connection wdth these areas of fusion. These are, however, not in the line of fusion, but at various angles to it, and in connection with the teloplii'agmata. The appearance is such as to suggest a penetration of intercellular fluid along some of the telophragmata peripherally; the teloplii-agmata may furnish more favorable channels for the capillary imbibition of such fluids; the fluids might conceivably alter the myofibrils in the close vicinity of these telophragmata into the disc-structure.

The above interpretation of the discs in terms of a local chemical modification of the myofibrils by tissue fluid, which at first consideration seems plausible, is stated simply for purposes of sharper contrast "with the interpretation which seems to us, in the light of more inclusive e\adence, to be the correct one; namely, that the intercalated discs, many of which undoubtedly arise in connection Tvith surfaces of fusion, are the products of modifications, of the natm-e of irreversible contraction bands on the peripheral myofibrils, resulting from unusual strains upon the fiber at the points of fusion incident to a rearrangement and new coordination of the peripheral fibrils in accord with the new stresses imposed by the fusion of distinct cells into a unit fiber. This point ^\dll be further discussed below.

/. The fetal myocardium

This order of description follows the actual order of the investigation. It might at first seem a more logical procedure to have begun the description with the younger fetal material and then to have passed from that through later fetal and early


post-natal to the adult conditions. But in the actual investigation it was found necessary to pass in the reverse order, for only in this way were these earliest fetal conditions correctly interpreted. Conditions in the Purkinje fibers of the adult heart served as the connecting link, and the interpretative key. Once the intercalated discs were discovered and interpreted in the Purkinje fibers (fusing cells) and in the early fetal heart, these simple conditions threw much light upon their definitive structure and relationships. Fetal and adult conditions served mutually to disclose the correct interpretation of the discs.

The youngest fetal heart studied was that of the end of the second or the beginning of the third month. The ages specified for the fetal hearts can only be regarded as close approximations. The youngest heart measured 32 mm. from base to apex, and 25 mm. at its widest point. This heart takes us very close, if not actually to, the first beginnings of the intercalated discs. At this stage the ventricular myocardium consists of closely-compacted, slender, fusiform elements (fig. 38). The resemblance to smooth-muscle structure is striking. This resemblance has not to our knowledge been previously pointed out. It is significant from the point of \dew of comparative histogenesis that cardiac muscle should pass through a transient phase of development in which it resembles definitive smooth-muscle tissue. Striped voluntary muscle of vertebrates likewise passes very early in its histogenesis through a very similar condition. In their earlier embryonic condition smooth muscle and cardiac muscle both consist of stellate and irregular elements whose processes have anastomosed to form a syncytium. The general idea that smooth, cardiac and striped skeletal muscle represent essentially successively higher stages of differentiation receives additional support in the evidence that heart muscle and skeletal muscle pass through a smooth-muscle stage.

But the cardiac muscle even at this early stage contains simple intercalated discs. The question then arises as to why neither smooth nor skeletal muscle contain similar discs. The answer probably inheres in the functional differences. The rhythmic contraction of heart muscle even in early fetal conditions


probably underlies the formation of intercalated discs in cardiac muscle, presumably as the effect of strains to which neither smooth nor skeletal muscles are subjected, at least not at correspondingly early stages.

The close essential resemblance between discs of adult Limulus heart muscle (fig. 41), and of fetal mammalian heart and the simpler types of vertebrate hearts in general, is striking and significant. In the Limulus heart the disc is clearly a modification of the myofibrils about a level bisected by a telophragma. The structure of the intercalated discs in the Limulus muscle (9 and 11), their relation to telophragmata, resemblance to contraction bands, and their relative scarcity, seem to permit of no interpretation other than one in terms of a modified contraction band. This being so in Limulus myocardium, and also in hearts of lower vertebrates (e.g., teleost fish and amphibia; Jordan and Steele (14) ), the conclusion seems to follow logically that a very similar structure in the fetal mammalian heart has a similar origin, and that its later condition must be explained in terms of further additions and modifications.

It was formerly thought that in the mammalian heart intercalated discs did not appear until some time after birth. Jordan and Steele first described their occurrence in the heart of pre-natal life in the case of the guinea-pig. Here they were described as first appearing during the last week of the gestation period. Jordan and Steele (14) had studied also earlier fetal hearts but were unable to identify the beginnings of discs. It may be that discs actually did not occur earlier in the guineapig heait. Or it may be that on account of the relatively finer structural features they were not discernible. But our experience with the beef heart leads us to surmise that the reason for the failure to identify earlier the discs in the guinea-pig fetus lay in an unsatisfactory staining.

In looking for discs in the fetal beel-heart we studied first the four-month heart. Discs were not at first clearly identified though there seemed to be some vague and uncertain evidence of their presence. We then proceeded to a study of the sevenmonth heart. But meanwhile we had prepared tissue also from


a two-month heart for study of the origin of the cells of the atrioventricular bundle, and the manner by which they became bi nucleated. This tissue was deeply stained and at once clearly showed intercalated discs in the fusiform cells. The resemblance between the histologic features of the heart of the beef-fetus at two months and those of the adult toad-heart, for example (Jordan and Steele), is striking. Tissue from the four-month heart was then restained, when the discs became clearly visible. And in the seven-month heart the discs were abundant and of substantially identical structure and relationship, except as altered by a coarsening, and the extraneous mechanical factors incidental to development, which factors continue to operate as modifying influences through postnatal growth and development.

As concerns the myofibrillar elements the early fetal heart is syncytial. But at two months, fusiform cells are plainly distinguishable. They contain an oval vesicular central nucleus. Certain nuclei are in process of amitotic division. The myofibrils are . sparse and peripherally arranged. The telophragmata are delicate but conspicuous, and peripherally among the fibrils Q-discs are faintly discernible. Delicate, deep-staining, granular discs appear fairly abundantly peripherally, apparently as modifications of the telophragmata (fig. 38). The cells are beginning to fuse to form the coarser trabeculae characteristic of later developmental conditions. It seems probable that the roughly-dichotonous division of the trabeculae of succeeding earlier stages, characteristic also of the moderator band (fig. 29), results from a central fusion of such fusiform cells, the branches representing the more widely spaced and unfused distal pointed ends of these fusiform elements. These terminals fuse with other similar terminals in the formation of the more regular meshwork of the later fetal heart.

The discs appear at right angles (approximately) to the surface of fusion, rarely in the surface of fusion as when two cells fuse end to end. Where cells fuse in this manner, along oblique surfaces, a recoordination of the constituent myofibrils must be effected along such surfaces, and the stresses involved may effect the modification of the myofibrils which constitute the inter


eahited discs, perhaps in essence irreversible contraction bands. Such originally modified areas may become secondarily further modified through the influence of, or by the addition of, relatively more abundant tissue fluid upon which the reactions to silver nitrate depend. It should be emphasized that the telophragmata react similarly to silver nitrate, which indicates that the telophragmata are the more direct paths for the penetration of the tissue fluid, which fact further explains the presence of tissue fluid in the discs because of the intimate relationship to telophragmata. In adult ventricular tissue tested with silver nitrate, the latter is precipitated in the spaces between adjacent fibers, in the telophragma, and in the discs. The close union of the telophragma with the sarcolemma gives the mechanical explanation of the penetration of tissue fluid from exterior towards center via telophragmata.

In the two-month fetal heart we certainly come very close to the beginning of the discs. The embryonic heart is a syncytium, composed of anastomosing stellate and irregular cells. It is only when these have become altered into fusiform elements, and the latter begin to fuse to form the beginning of the secondary meshwork of myocardial syncytium (fig. 38) that the first discs appear. This probably occurs somewhere early in the second month, and the actual beginning is hardly very different from the one here described for the two-month heart.

The two-month heart shows, then, conclusively that the discs arise in connection with cellular fusions, as modifications of the myofibrils in lines corresponding to telophragmata, and approximately at right angles to surfaces of fusion. The evidence is not incompatible with our general interpretation in terms of contraction bands, more especially when the Limulus and lower-vertebrate hearts are kept in mind, but it furnishes no additional support to the hypothesis and it cannot be finally denied that the myofibril modification might possibly be of the nature of a splicing (for purposes of recoordination) of myofibrils of fusing adjacent cells. If end to end sphcing of myofibrils were the complete explanation of the discs, however, it is


very difficult to understand their originally sharp segregation along the telophragmata.

Study of the two-month fetal heart throws further light also on the nature of the cells of the atrioventricular bundle and of the Purkinje fibers. In this heart the resemblance between these elements and the fusiform cells of the myocardium is much closer than in later fetal hearts. The element in each case is a fusiform cell. The cell of the atrioventricular bundle is short and very stout, the cells of the Purkinje fibers are longer and less stout, that of the myocardium is still longer and relatively slender. Moreover each type multiplies its nuclei by amitotic division. The atrioventricular bundle cell more generally has only two nuclei; and an occasional nucleus may be seen in amitotic division.

It is quite true that even at the two-month stage the atrioventricular bundle can be clearly recognized. But this fact is no proof that these cells and the Purkinje fibers are not actually less differentiated myocardial fibers. It seems probable, in view of the evidence from the two-month fetal heart, that originally the three types came from a similar tissue or syncytium. The atrioventricular bundle cells differentiate only to a certain early stage. This stage is characterized by a stout fusiform shape, much sarcoplasm, few fibrils slightly differentiated and peripherally arranged. The cells, moreover, frequently have only a single large central nucleus. The nucleus occasionally divides by mitosis, but more generally at this stage by amitosis, to produce a binucleated cell. The cells remain distinct, but are closely united by intercellular bridges and continuous myofibrils. The Purkinje fibers progress to a somewhat later stage characterized by an elongated fusiform shape, amitotic division of nucleus, and a fusion to form fibers, the fusion involving the formation of discs. The myocardium passes through very similar earlier stages, but progresses along the same lines to a higher degi*ee of differentiation. From this viewpoint it is quite correct to speak of the Purkinje fibers and the atrioventricular bundle cells as less highly differentiated myocardial elements.




It becomes a relatively easy matter to support, on the basis of observational data, any one of the earlier-proposed hypotheses regarding the significance of the intercalated discs if only discs of a certain tj^pe are selected as representing the original form, and others regarded as secondary modifications. For example, certain regions may be found in abundance in which nucleated areas of coarse trabeculae and related branches are clearly and sharply demarked from similar areas by a fairly uniform type of band- and terraced-discs. Such have been published almost exclusively in support of the intercellular nature of the discs by Palczewska (20) and by Werner (23). No interpretation is attempted by these investigators of such discs as are illustrated in figures 22, 23 and 24. If we combine with such evidence also the results of macerating cardiac muscle, when areas similar to those above-described for sections are isolated; and the further fact that silver nitrate is precipitated by the intercalated discs, the evidence at first seems complete that the myocardium is compounded of distinct cells. But such interpretation must ignore the facts that cardiac muscle is originally syncytial, that discs appear only gradually during fetal (from about the second month) and infantile life, and that the nucleated areas outlined by the discs and sarcolemma do not correspond closely with the original stellate myoblasts of the embryonic myocardium nor with their later fusiform and cylindric modifications. If we add to these countervailing facts the further facts that many discs are of very irregular character (e.g., figs. 24 to 27) and that they are only incomplete peripheral band-like structures (not membranes passing from surface to surface) occasionally supernuclear in position, an intercellular interpretation becomes untenable.

Discs like the one illustrated in figure 21 seem to support Heidenhain's opinion that they represent areas from which new sarcomeres arise. But the great majority of the discs are very different from such a structure, and cannot be interpreted in this manner. Nor do such discs resemble the differentiating sar


comeres of fetal muscle. Moreover, this disc may equally plausibly be interpreted as a partial relaxation of adjacent halves of two successive contraction bands. But the strongest countervailing evidence to any interpretation in terms of sareomeric differentiation are the facts above stated that trasitions between alleged differentiating discs and definitive sarcomeres are lacking, and that the discs are not abundantly present, nor at all in definitive form, when the heart grows most actively during fetal life, and that they do not disappear at the close of physiologic maturity. Moreover, the discs show a progressive increase and development from the sparse, delicate, simple types of early fetal life to the abundant, more robust, and complex types and modifications of the adult heart.

When one wishes to argue for the tendinous nature of the discs, he may point to the fact of the bulging of certain contracted areas which are bounded at both ends by discs. But again such appearances are relatively rare. Moreover, the alleged tendons (discs) do not react to specific stains for collagen fibrils, the discs frequently lie within contracted or relaxed areas, and the irregular varieties have no close structural resemblance to tendons. The discs only fortuitously bound such contracted bulging regions.

Similarly with respect to Dietrich's coordination-mechanism theory. The discs in general occupy the proximal regions of trabecular branches, and might perhaps serve well to coordinate the functional activity of the included myofibrils; but no evidence accrues that such is actually the case. Our evidence indicates rather that an attempt at coordination (or recoordination) is the cause of disc formation; not that the discs effect the coordination as Dietrich (3) believes.

In attempting an interpretation of the intercalated discs all the available evidence must of course be included. The correct interpretation must be able to comprehend in logical form all modifications of type and relationship of the discs in fetal, normal adult and pathologic hearts. We incline to believe that the interpretation of the discs as secondary modifications of the myofibrils at certain areas characterized by unusual functional


conditions, probably excessive stresses, causing an inability on the part of contraction bands to revert to the relaxed condition, and as such subsequently chemically and mechanically modified, can embrace more of the actual observational data and come nearer expressing the real significance of the discs than any hypothesis hitherto proposed.

In support of this hypothesis numerous comparative developmental and structural data must be considered. In the first place, the gi'osser developmental features and alterations must be kept in mind and brought into line. In the interpretation of the intercalated discs not enough attention has hitherto been given to the gross changes in the trabeculae: (1) The myocardium is originally and definitely a syncytium, both with respect to the grosser anastomoses between the trabeculae through their branches and with respept to the myofibrils; however, towards the end of the second fetal month, the myocardium consists of closely compacted slender fusiform elements, resembling adult conditions in smooth muscle, but the delicate intercellular bridges with the continuous myofibrils meanwhile still effect a syncytial structure. (2) The more delicate trabeculae of the later fetal myocardium lengthen and coarsen through fusion and intrinsic growth, and the originally more delicate branches undergo similar changes and in many instances alter their origin at sharp angles to origins at less acute angles. (3) In the processes of later development the ventricular fiber-bundles and their constituent trabeculae undergo spiral twistings (e.g., the bulbo-spiral band: Mall (18), Am. Jour. Anat., vol. 11, 1911, fig. 19, p. 262) which involve secondary fusions of adjacent trabeculae and distortions of branches, including in certain cases an inturning of a portion of the originally peripheral sarcolemma (see also Heidenhain's "Plasma und Zelle," figs. 297e, 298 and 300, and diagram fig. 353, p. 616; and Jordan's (12), "Studies in Striped Muscle Structure," No. Ill, Anat. Rec, vol. 6, 1917)..

With respect to the finer microscopic features, the following changes must be kept in mind: (1) The discs are invariably related spatially to the telophragmata, being bounded on one or


both sides or bisected by them, and, at least occasionally, shading laterally into these membranes. (2) The telophragmata are in close union with the myofibrils, the sarcolemma, and the nuclear wall. (3) The discs are peripherally placed and consist of associated local modifications of adjacent myofibrils.

The developmental features that require emphasis in this connection are: (1) The absence of discs in the embryonic myocardium; the discs appear only gradually in fetal life (beginning about the second month) as delicate peripheral bands, apparently as thickenings of parts of the telophragmata; they increase in number, size and variety during the period of the growth of the heart; these developmental changes recapitulate the phylogenetic history of the discs, as first pointed out by Jordan and Steele (14), who found them in hearts from teleost fishes to birds, and even in the Limulus heart (Jordan (9 and 11) ) where they are exclusively of the simple-comb type (sometimes in the shape of a two-step form), located at telophragmata levels. (2) The more complex tj^es of discs can all be refelred to the simpler band types, as mechanical secondary modifications of these simpler types. (3) The simplest discs consist of rows of bacillary modified foci on adjacent fibrils., (4) Hypertrophied and atrophied pathologic myocardia are characterized by definite types of discs, complex serrated forms and narrow comb forms respectively (Dietrich (3); Jordan (6, 7 and 13) ). • The close structural similarity of the original and simplest discs to contraction bands, and their identical location with respect to the telophragmata, suggested an origin of discs from modified contraction bands. A contraction band in a stained section of certain insect muscle fibers (leg or wing; Jordan (10) ) has the appearance of the simplest type of intercalated disc. If it be assumed that certain bands, on account of excessive strains, become incapable of reversion, then the possible beginnings of discs seem to be present, which simple discs are correctly conceived to be capable of modification through the operation of mechanical factors into the various types of discs above described.

It seems desirable at this point to trace the probable steps, as suggested in the histologic preparations, by which the more


complex types of discs originate from the simpler band-discs. The first types in the order of simplicity are the terraced forms. The same explanation that applies to a two-step form will apply also to multiterraced forms. Moreover, the explanation must hold as well for a terraced type in which the steps are only of one order (descending or ascending) as also for those in which the steps are of a double or compound order (descending combined with ascending). But as we shall see the same explanation need not apply also to the irregular terraced types.

Obviously one of two explanations might apply to the terraced types of regular order: (1) They might have resulted from a dislocation of an original band form; or (2) they might have resulted from the close allocation of originally disconnected short bands. The fact that the interval between successive steps may be one or several sarcomeric segments need not affect this conclusion. In the first case specified, connecting membranes or 'risers' would not be expected. Such step-forms appear abundantly. It should be noted also that generally in the case of step-forms the involved myofibrils are divided into bundles corresponding in width with the width of the step-segments (figs. 2, 13, 19 and 26). In the second case specified, the connecting membranes might conceivably be either a portion of the sarcolemma or a portion of a telophragma. The possibility of a contributory mesophragma need not be considered, since no evidence appears that such an alleged membrane (Heidenhain (4) ) actually occurs in the cardiac muscle of the beef. Where a secondary spiral twisting of the fiber is superimposed, the connecting membrane may very likely be an inturned portion of the sarcolemma. In the absence of a spiral twisting, in which case the membranes ('risers') are relatively dehcate, the connections may be formed by portions of an involved telophragma. But as we saw from a study of the first origin of discs in the fetal heart, step-discs may arise in relation to oblique surfaces of fusion, and in such instances the connecting membranes are also portions of the fused sarcolemmae of adjacent fibers. Whether the fasciculation of the trabeculae above mentioned in connection with terraced discs lacking connecting membranes is sec


ondary to the formation of the step-discs or a result of the dislocation of an original straight band-disc is uncertain, and not of fundamental significance.

If the terraced forms of discs were the result of a dislocation of a simple band disc, then it might seem to be required that the involved telophragmata should show a distortion. Such is not generally the case. Figure 13 shows an exception. But a careful consideration of the possibilities will explain the general absence of coincident telophragmatic distortion. If it be presumed, as seems necessary under the conditions postulated for the formation of a certain type of terraced disc, that the involved telophragmata are broken and the segments shifted in position, they could only shift to some place between two successive telophragmata or in series with them. The former may involve fusions between portions of successive telophragmata, a phenomenon indicated in figure 13; or a blending of telophragmata.

The available evidence seems to force the conclusion that many of the regularly terraced types of discs originate by a process of secondary dislocation of band discs, and a shifting of the resulting segments to successively lower levels in a lateral direction, due apparently to successively greater tensions laterally in the trabeculae, the result in part of the oblique tensions caused by the anastomosing branches, and in later stages in part probably also to the spiral twistings of certain groups of trabeculae (e.g., the bulbo-spiral band).

With respect to the irregular types of terraced discs, in which the coarser connecting membranes are invariably present, the processes of formation involve the fusion of apposed portions of the sarcolemmae of the adjacent trabeculae, caused to fuse by reason of a mutual spiral twisting. Such fusions are common in certain skeletal muscles, e.g., in the post-abdominal segments of the scorpion (Jordan, (12) ) and in human cardiac muscle (Heidenhain (4) ) . In cases where peripheral discs were present in the regions of the fusions of the involved fibers these would become arranged in irregular step-form due to their relation to the anharmonic telophragmata of the fused fibers. Similarly in cases where a single fiber is spirally twisted a portion of the sarco


lemma may become inturned (Heidenhain (4) ) and form a connecting membrane between the inturned portions of- peripheral discs. These discs are not formed as the result of fusions in the manner of those previously described, but are simply morphological modifications of discs already present by reason of a secondary t\visting and fusion.

From the above it seems clear that all varieties of discs can be explained in terms of a band-disc connected with a telophragma, as secondary modifications incident to the various tractions and tensions acting upon adjacent groups of myofibrils, or even adjacent single fibrils. The latter condition would result in the more delicate serrated types, which in the case of growing or hypertrophying fibers would involve also the telophragmata included among the splitting fibrils (figs. 24 to 27). The presence or absence of a delicate connecting membrane between the segments of a step-disc in the former condition might depend upon whether the elasticity of the telophragmata was sufficient in any given instance to withstand the strain- ol extension to the distance of one or several sarcomeres.

We may now return again to a consideration of the initial stage in the formation of discs. We appreciate the fact that the weakest link in the chain of argument in support of the interpretation of the original discs as modifications of contraction bands is the explanation of their inception. But the histogenetic data also seem to point to such an interpretation. When once formed in their simplest condition, all the various types of more complex discs can be readily explained by our hypothesis. An explanation is not equally easy on the basis of any other hypothesis previously proposed. The facts that the discs make their first appearance while the heart is actively growing (about the end of the second month in the beef) and persist thereafter in coarsened and modified forms, and that they are at first invariably peripheral in position, have also a special bearing in this connection. In the growth of the fiber, myofibrils are being constantly added centrally by process of splitting from the more peripheral older fibrils of the radial lamellae. The more peripheral fibrils are first formed and are consequently the


first to function; hence they support all the strains of contraction and recoordination at the very time when rearrangement of fibers (trabeculae), fusions and twistings are most active, and before they are reinforced by more central fibrils. These structural peculiarities of the trabeculae explain in a measure both the formation of the discs as possible irreversible contraction bands, and their peripheral location. As more centrally placed myofibrils develop in later cardiac histogenesis, they would be more likely to be modified in a similar manner at the levels of the earlier-formed more peripheral discs, and thus the discs would tend to grow coarser in a radial direction and wider in a transverse direction. But growth in radial width is probably more largely a matter of multiplication through fission of the discunits of dividing myofibrils. The initial simple discs apparently arise both in relation to surfaces of fusion and independently of fusions. The common factor in the production of these original discs is presumably a strain effect upon localized portions of myofibrils causing an irreversible condition of a contraction band.

Finally, .when one turns for evidence in support of this hypothesis to the Purkinje fibers one finds here a combination of the distinctive differential characteristics of the musculature of the atrioventricular bundle and the ventricular myocardium, namely both serrated cell-margins (in histologic preparations) and a few simple hand-discs. In other words, the Purkinje fibers at the level of transition from the cells of the atrioventricular bundle are still largely distinct cells, but they are drawn out into fibers towards the ventricular myocardium where they contain also a few discs, and where they are undergoing fusion. The Purkinje fibers of the adult heart are apparently at the histogenetic stage attained by the ventricular myocardium at about the beginning of the third fetal month. This structural condition is incompatible with an interpretation of the discs as intercellular cement substances or as tendons. An interpretation in terms of differentiating sarcomeres is likewise inadmissible on evidence already stated.


The only other hypothesis that has any appearance of plausibihty as suggested by certain conditions of the cells of the atrioventricular bundle, the appearance of the early fetal myocardium, and by adult myocardium treated with silver nitrate — is that at certain levels, for some unknown cause, intercellular tissue-fluids may penetrate via the telophragmata and modify the myofibrils in these regions. Such an interpretation has not to our knowledge been previously proposed, but it may at least be stated. Once formed in this manner, the discs could again be altered by the mechanical factors incident to development and function as above explained. But a complete interpretation on this basis would still demand an explanation of the original causal factor upon which the locally increased penetration of tissue-fluid depended.

The cause of such localized (selected) relatively more pervious regions is obscure, unless on some basis requiring a previous modification of the telophragmata concerned and as a concomitant result of a modification of the attached portions of the involved myofibrils. Such modification might again conceivably be a result of a local unusual functional requirement, possibly producing an excessive strain effect. The peculiar diffuse staining-reaction of the intercalated discs in general may be the result of a relatively more profuse collection of intercellular tissue fluid in the already modified portions of the myofibrils represented by the discs. This is indicated more especially by the appearance of cardiac tissue treated with silver nitrate: the intercalated discs are not sharply outlined, but their margins are vague and irregular, and the myofibrils appear masked by the granular precipitate and show no resemblance to the definite comb-discs of the hemahim-stained tissue. The relative increase of tissue-fluid in the discs is more probably the result than the fundamental cause of disc formation.

The unit of the original disc is a modified focus of a myofibril at the level of a telophragma. By transverse linear combinations of such units, and subsequent mechanical modifications all the types of discs may be readily conceived to be derived. Since this initial unit (a bacillary portion of the myofibrils,


bisected by a telophragma) is comparable, structurally, tinctorially, and in respect of relation to telophragma, to a contracted portion of a myofibril (see e.g., figure 6, illustrating contracted leg muscle of sea-spider; Jordan (12) ), an explanation of intercalated discs is suggested by the microscopic evidence in terms of a modified contraction band, possibly an irreversible band.

But this conclusion must be brought into harmony with the fact of the formation of discs in the Pur kin je fibers and in the early fetal hearts, in relation to surfaces of fusion among adjacent cells and trabeculae. The hypothesis that the intercalated discs of heart muscle are of the nature of irreversible contraction bands must be able to include and harmonize the evidence that in the PurkiAje fibers and the fusiform elements of the fetal heart the discs arise in relation to fusion-areas between elongating cells like those of the atrioventricular bundle. If it cannot do this it must be abandoned. It will be observed that the discs do not generally arise in the areas of fusion but at right angles (approximately) to such fusion areas. Where two fibers fuse along oblique surfaces, the peripheral myofibrils at least, must be brought into coordinated functional relationship. This conceivably involves special stresses and strains at the point or levels of recoordination. Since at these earlier stages when lusions are most extensively made, the myofibrils are relatively less abundant while peripherally arranged and in union with the telophragmata, and since only the most peripheral fibrils are probably involved in the new coordination, the peripheral location of the discs and their spatial relationship to the telophragmata is accounted for.

The above discussion would seem to bring into harmony with the newer hypothesis here accepted of the significance of the intercalated discs, two other hypotheses, namely the intercellular and the coordination-mechanism hypotheses. The intercalated discs originate along original intercellular surfaces but not generally in such surfaces; they accordingly in part outline more or less accurately original intercellular or inter-fiber region^. The discs may be conceived as the result of attempts


at coordination of functionally incooordinated myofibrils of fusing trabeculae (/cells') which involve unusual strains, but they do not themselves effect the coordination; they are effects of functional coordination not primarily causes of such coordination, as urged by Dietrich (3).


1. Intercalated discs are described in sections from the atria, ventricles, moderator band, and Purkinje fibers of the adult heart of the beef. No striking numerical or structural differences obtain between thd discs of the right and left ventricle, nor between those of the ventricles and atria. The types of discs include the simple band-forms, more or less complex terraced forms, and serrated forms. These occur in frequency in the order named, the serrated type being relatively sparse. Discs are somewhat more abundant in the papillary muscles than in the ventricular wall, and are more predominantly of the band-form. A similar statement applies also to the moderator band. Considered in toto many of the 'band-forms ' of disc are more or less complete rings or spirals. In the Purkinje fibers the discs are relatively less abundant than in the ventricular myocardium proper, and they are predominantly of the band-form, with occasional short step-forms. The several technics employed include maceration, treatment with silver nitrate solutions, and fixation by the Zimmermann nitric-acid-alcohol mixture \vith hemalum and iron-hematoxylin staining respectively. The stained tissues were studied in sections, and in teased condition mounted in glycerin. The investigation included further the study of hearts of fetuses of the second, fourth and seventh months, and of young calves' hearts.

2. Discs are present already towards the end of the second fetal month (ventricle) as delicate peripheral bands, apparently as local thickenings of the telophragmata. Subsequently to the second fetal month the discs become progressively more abundant and more robust, and after birth they become altered into more complex terraced and irregular forms.


3. In the adult heart the discs are still for the most part peripheral, as revealed both in transverse sections and in teased preparations. They never extend completely through a fiber. They are always intimately associated with telophragmata. The teloplu*agmata are in close union with the sarcolemma, the nuclear wall and the myofibrils. In their simplest form, the discs shade laterally into a telophragma, the latter apparently bisecting the disc. In the more highly differentiated types (mechanically modified discs) telophragmata frequently bound one and occasionally both surfaces of the disc.

4. The unit of structure of the simple band-disc is a modified bacillary portion of a myofibril at a telophragma level. Such units are grouped into bands of various widths (longitudinally) and breadths (transversely) to form the initial discs.

5. The more complex terraced, serrated and irregular types of disc are derived from the simple band-forms through the operation of secondary extensive mechanical and possibly also chemical factors. The fundamental mechanical factors are irregular tensions operating in opposed or oblique directions upon certain regions during the development and functional activity of the heart. The irregular direction of the stresses are determined by the syncytial (meshwork) character of the myocardium. The primary results of such stresses are further modified during development by spiral twistings of single fibers involving occasionally an inturning of portions of the sarcolemma, and by similar mutual twistings of two adjacent fibers resulting in lateral fusions.

6. Terraced or step-like discs result in part from a segmentation of the original band-discs and a secondary dislocation of the resultant segments consequent to dissimilar tractions upon successively more lateral segments; in part they arise also along (approximately at right angles to) the oblique surfaces of fusion of adjacent fibers (cells), presumably as strain effects (modified irreversible? contraction bands) resulting from a re coordination of the peripheral myofibrils of the fusing fibers. Terraced discs may be formed also as secondary modifications of original band discs by spiral twistings of single fibers or of two adjacent fibers.


In the first type connecting membranes ('risers ') are generally lacking; when present they are very delicate and probably represent portions of an involved telophragma. In the last two types connecting membranes are more robust and stain more deeply, and represent fused portions of the involved sarcolemmae. The serrated types of disc result from unequal functional tensions upon the units of an original band disc in a region where the myofibrils are undergoing a longitudinal fission in the process of growth of the fibers.

7. Evidence is presented in contravention of the pre\dously proposed hjqDotheses concerning the significance of the discs, namely as intercellular cement-substances (Zimmerman, et al.), as tendinous structures (Marceau), as developing sarcomeres (Heidenhain), and as originally coordination mechanisms for the myofibrils (Dietrich) .

8. Histologic data are presented in further support of the hypothesis proposed by Jordan and Steele that the simplest types of discs are local modifications of adjacent myofibrils at the level of a telophragma (possibly of the nature of strain effects producing a condition of irreversibility ot contraction bands), and that the more complex types are secondary mechanical modifications of the simpler discs. Additional evidence indicates further that the discs are incidentally modified through the infiltration of intercellular tissue fluid along the telophragmata, which accounts for the precipitation of silver nitrate in these regions. A concomitant chemical modification would account for the relatively greater ease with which the myocardium fragments in macerating fluids in the regions of the discs.

9. The atrioventricular bundle can be distinguished from the myocardium already at the second month. It is composed of short, stout, fusiform or polyhedral cells containing scattered myofibrillar elements continuous from cell to cell. The cells commonly contain two centrally located nuclei, closely apposed, the amitotic division products of an originally single nucleus. The bundle terminates distally in Pur kin je fibers, which connect with the myocardium of the ventricles and of the moderator band. The Purkinje fibers are elongated elements similar to


the cells of the atrioventricular bundle. Occasional band and short step-like discs occur on these fibers. The transition from the cells of the atrioventricular bundle to the Purkinje 'cells' is characterized by an elongation and fusion of the cells to form true fibers (trabeculae), with intercalated discs. The intercalated discs of the Purkinje fibers occur (arise) along the surfaces of oblique fusion of original cells (figs. 48 and 51). Such definite evidence of the origin of the discs in the Purkinje fibers gives the clue to their origin, in part at least, also in the general myocardium, namely in relation to surfaces of fusion of originally distinct elements (cells; trabeculae'). This can actually be demonstrated in the early fetal heart. The origin of the discs in regions of fusion between cells, approximately at right angles to the surfaces of fusion and in relation to telophragmata, is interpreted in terms of a strain effect resulting in a local modification of adjacent myofibrillae (essentially an irreversible contraction band in the initial condition) and incidental to a recoordination of the peripheral myofibrils of the fusing cells or trabeculae.

10. Intercalated discs occur in the fetal heart already towards the end of the second month as deeply-staining granular modifications of certain telophragmata in their lateral extensions among the peripheral myofibrils. The myocardial elements are long, slender, fusiform cells in process of lateral and terminal fusion to form the trabeculae and branches of the later syncytial musculature. The discs are located at angles to surfaces of fusion. The resemblance between early fetal myocardial elements and the Purkinje fibers of the adult heart is striking. The Purkinje fibers, as also the cells of the atrioventricular bundle of the fetal heart, are fusiform elements whose myofibrillar constituents are associated with telophragmata and extend from cell to cell via intercellular bridges. The relation of the very similar intercalated discs of the Purkinje fibers of the adult heart and those of the fusiform elements of the early fetal heart to surfaces of fusion is the same in both cases.

11. The new data disclosed in this investigation, namely, the origin of the intercalated discs in relation to surfaces of fusion of


previously distinct myocardial elements, need not be prejudiced by a forced association with the hypothesis that the discs are essentially irreversible contraction bands. But it may again be emphasized that the discs do not generally occur in the surfaces of fusion (hence not fundamentally intercellular in char* acter) but laterally to such areas of fusion. As regards the fetal myocardial elements and the Purkinje cells (fibers) this is indisputable fact. The hypothesis here supported is simply interpretative of this, in common with other, facts. Nevertheless w^e believe that the hypothesis can interpret more logically and consistently than any previously proposed the microscopic data concerning the intercalated discs. Moreover, sight must not be lost in evaluating the hypothesis, of the strong support it receives from conditions in the Limulus heart where the simpler discs appear, though sparsely, in a considerably coarser and clearer form. In the Limulus heart the discs seem to admit of no possible interpretation except in terms of modified contraction phenomena .


(1) CuRRAN, E. J. 1909 A constant bursa in relation with the bundle of His;

with studies of the auricular connections of the bundle. Anat. Rec, vol. 3, pp. 618-640.

(2) De Witt, Lydia M. 1909 Observations on the sinoventricular connecting

system of the mammalian heart. Anat. Rec, vol. 3, pp. 475^98.

(3) Dietrich, A. 1910 Die Elemente des Herzmuskels. Fischer, Jena, pp.


(4) Heidenhain, M. 1911 Plasma und Zelle, Fischer, Jena.

(5) Jordan, H. E. 1911 The structure of the heart muscle of the humming

bird, with special reference to the intercalated discs. Anat. Rec, vol. 5, pp. 517-529:

(6) 1912 The intercalated discs of hypertrophied heart muscle. x\nat. Rec, vol. 6, pp. 357-362.

(7) 1912 The intercalated discs of atrophied heart muscle. Proc Soc Exp. Biol, and Med., vol. 10, pp. 1-3.

(8) 1914 The microscopic structure of mammalian cardiac muscle, with special reference to so-called muscle cells. .\nat. Rec, vol. 8, pp. 423-430.

(9) 1916 A comparative microscopic study of cardiac and skeletal muscle of Lipiulus. Anat. Rec, vol. 10, pp. 210-213. (Proc. .\m. Assoc. Anat. 1915).


(10) 1916 The microscopic structure of the leg muscle of the sea-spider, Anoplodactylus lentus. Anat. Rec, vol. 10, pp. 493-508.

(11) 1917 The microscopic structure of striped muscle in Limulus. PuV). 251, Carnegie Institution of Washington, pp. 273-290.

(12) 1917 Studies in striped muscle structure. III. The comparative histology of cardiac and skeletal muscle of scorpion. Anat. Rec, vol. 6, pp. 1-20.

(13) Jordan, H. E., and Bardin, J. 1913 The relation of the intercalated

discs to the so-called segmentation and fragmentation of heart muscle. Anat. Anz., Bd. 43, pp. 612-617.

(14) Jordan, H. E., and Steele, K. B. 1912 A comparative microscopic

study of the intercalated discs of vertebrate heart muscle. Am. Jour. Anat., vol. 13, pp. 151-173.

(15) King, M. R. 1916 The sinoventricular system as demonstrated by the

injection method. Am. Jour. Anat., vol. 19, pp. 149-179.

(16) Lange, W. 1914 Die anatomischen Grundlagen flir eine myogene Theorie

des Herzschlages. Arch. mikr. Anat., Bd. 84, pp. 215-263.

(17) Lhamon, R. M. 1912 The sheath of the sinoventricular bundle. Am.

Jour. Anat., vol. 13, pp. 55-71.

(18) Mall, F. P. 1911 On the muscular architecture of the ventricles of the

human heart. Am. Jour. Anat., vol. 11, pp. 211-267.

(19) Marceau, F. 1904 Recherches sur la structure et le developpement

compare des fibres cardiaque. Ann. des. So. Nat. Zool., vol. 19.

(20) Palczewska, Irene von 1910 Ueber die Struktur des menschlichen

Herzmuskelfasern. Arch. mikr. Anat., Bd. 75, pp. 41-101.

(21) Retzer, R. 1908 Some results of recent investigations on the mammalian

heart. Anat. Rec, vol. 2, pp. 149-155.

(22) Tawara, S. 1906 Das Reizleitungssystem des Saugetierherzens. Fischer,


(23) Werner, Marie 1910 Besteht die Herzmuskulatur der Saugetiere aus

allseits sharf begrenzten Zellen oder nicht? Arch. mikr. Anat., Bd. 75, pp. 101-149.

(24) Zimmermann, K. W. 1910 Ueber den Bau der Herzmuskulatur. Arch.

mikr. Anat., Bd. 75, p. 40.


The illustrations were made with the aid of the Bausch and Lomb camera lucida. Unless otherwise specified all figures are of tissue fixed in the alcoholnitric-acid mixture and stained in hemalum according to Zimmermann's technic, and magnified 1500 diameters. The original magnification is reduced one-third in reproduction.

The drawings were made with water-proof India ink of various dilutions. As such they represent quite faithfully the appearance of sections stained with iron-hematoxylip. If shades of blue are substituted for the black and grays in the illustrations the appearance of the hemalum-stained sections is closely imitated.





1 Portion of a longitudinal section of a muscular trabecula from the myocardium of the right ventricle. The trabecula is in the relaxed condition. Two intercalated discs are shown, at the levels of successive telophragmata, having apparently displaced these membranes. They are of the simple band form, the constituent elements being modified rod-like portions of the included myofibrils. The fiber is somewhat narrower at the point of the location of the discs, as if stretched in this region. The discs are comparable, structurally, to contraction bands which have become stretched out by reason of the tension exerted by the adjacent contractile portion of the trabecula.

2 Portion of a longitudinal section of a branching trabecula of right ventricle. A terraced form of intercalated disc divides an upper relaxed from a lower (left) contracted region. The contracted stains more deeply than the non-contracted region. In the latter the telophragmata, Q-discs and J-discs are conspicuous; in the contracted region only relatively narrow contraction bands occur, alternating with lighter broader discs. At the left is shown the sarcolemma, thrown into arcades or festoons. These span the spaces between successive telophragmata. Where intercalated discs occur, the point of attachment generally corresponds with the midportion of the disc. In the case of the upper two elements of this step-like disc, the telophragmata at the right also pass to the mid-line of the discs. In so far, these elements correspond to slightly modified contraction bands. In teased preparations of stained myocardium, mounted in glycerin, the telophragma can occasionally be seen actually bisecting a disc. The lower four elements are of identical structure, and placed at successive sarcomeric segments in such a manner that their upper limits are on a level with the telophragmata at the right, and their lower borders correspond to the level of the delicate contraction bands at the left. The second and third element are connected by a slightly deeper-stainipg membrane, probablja sarcolemma remnant. Similar complex discs appear in teased preparations, hence not due to any peculiarity of section.

3 a and b Portion of a longitudinal section of a myocardial trabecula of the right ventricle at a higher (a) and a lower (6) level of focus. This disc is of the simple band-type, peripherally located, apparently displacing a telophragma. In lowering the level of focus the disc is seen to have become dislocated or distorted on the opposite surface.

4 A broad two-step disc from a trabecula of the right ventricle. The steps are connected' by a 'riser' in the form of a deeper-staining membrane. The disc may represent two successive bands, but is more probably a band-disc secondarily dislocated.

5 Complex terraced disc bounding a wedge-shaped, lighter-stained, relaxed portion of the fiber at the left. In passing from a lower to a higher level of focus the group of discs marked 1 and 2 come successively into view. These discs are peripherally located. Those marked {I') shade laterally into the telophragmata. The intermediate terraced portion {2) consists of 'steps' connected by deeplystaining membraneous 'risers.' A complete interpretation probably requires


the assumption of a fusion along the line of the terrace of originally separate portions of the fetal syncj'-tium. The disc portions of the 'terraces' appear at angles to the obUque line of fusion, in close association with the telophragma, as the expression of a modification of the myofibrils by reason of new stresses imposed incidental to a functional recoordination of myofibrils in the altered trabeculae. The modification may be of the nature of an irreversible contraction-band, subsequently modified both mechanically and chemically, possibly also by the accumulation of tissue fluid.

6 Portion of a longitudinal section of a fiber of the right ventricle, showing a disc overlying the end of the nucleus. The telophragmata are attached centrally to the serrated nuclear membrane and peripherally to the festooned sarcolemma.

7 and 8 Transverse sections of ventricular trabeculae at the level of the nucleus, showing the peripherally disposed intercalated discs. The myofibrils are aggregated into peripheral lamellae and central more or less irregular cylinders. Some of the lamellae show a peripheral longitudinal split, probably a growth phenomenon.

9 Portion of a longitudinal section of two adjacent trabeculae from the right ventricle. Both fibers show a single branch demarked by an intercalated disc. In the upper portion occur four discs of the simple band type. The upper one on the left extends from the branch across the nucleus, into the main trabecula, and into the lateral portion of the adjacent fiber. This disc can best be interpreted according to the hypothesis which relates it to a modified contraction band. In the lower portion of the field on the fiber at the left occur three discs at successive telophragmata levels. In the center where the two fibers have come into close apposition, involving a fusion of the sarcolemmae in such a manner as to produce an asymmetrical arrangement of the adjacent sarcomeres, a long step-disc appears. The 'risers' here consist of the fused sarcolemmae.

10 Two adjacent fibers with sarcolemmae fused medially. Several small band-like discs appear at the upper and lower levels of the central fusion at approximately right-angles to the line of fusion. The lower group of two discs lies within a contracted area. At the upper and lower terminals of the central fused area, the myofibrils have become modified to form simple discs, presumably a consequence of the change in the direction of function of the involved myofibrils.

11 Semidiagrammatic drawing of two adjacent fibers whose sarcolemmae have fused in such a manner as to produce an alternation of apposed telophragmata, showing the arrangement of two band-discs with relation to the common sarcolemmae an9 the telophragmata. The formation of the discs in such cases probably proceeded the spiral twisting to which the fusion of the sarcolemmae is due.

12 Irregular type of band-disc. Two trabeculae, or branches, have apparently fused end to end at an obtuse angle. The new functional requirements on the part of the myofibrils in this region of fusion effected a modification which resulted in this peculiar type of disc. The disc is readily interpretable in terms of modified irreversible contraction-bands.


13 Complex terriiccd discs from right ventricle. Groups 1 and 2 come successively into view as the focus is lowered, and are continuous around the left margin. The lower group separates a contracted from a relaxed region; the upper group lies in a relaxed area. The discs are peripherally placed, and form portions of a spiral, possibly the combined result of a fusion of adjacent trabeculae and a subsequent spiral twisting of the new fiber. A spiral twisting is indicated also by the fusion of the telophragmata above.

14 Portion of a longitudinal section of a myocardial trabecula of the right ventricle showing an extensive group of band-like discs. The discs are peripherally located and a number can be seen to be continuous across the lateral border as the level of focus is changed, thus revealing an annular or spiral form. The numerals indicate successive levels of focus at which the discs appear. The great number and considerable variety of discs in such a small area would seem to exclude interpretation in terms of intercellular cement, tendinous structures, or growth regions.

15 Transverse and longitudinal sections of trabeculae of right ventricle of four-month fetal heart (compare with figures 8 and 10). In the fetal heart of this stage the trabeculae have a relatively lesser diameter; more vesicular and more regular nuclei; fewer myofibrils, peripherally disposed; more widely-spaced, less robust, and less regularly arranged telophragmata.

16 Two adjacent fibers from the atrium, with typical band-discs bounded on both sides by telophragmata. The structural units are clearly modified portions of the myofibrils.

17 Disc from atrium, composed of three portions interconnected by a deeply-staining membrane, probably the result of an upward dislocation of a central portion of the original band disc.

18 Two apposed discs within the same sarcomere. They apparently represent apposed halves of successive contraction-bands, which failed to relax.

19 Semidiagrammatic sketch illusttating the possible origin of the type of terraced disc lacking coarser 'risers,' from an original band-disc by process of division and dislocation of the resulting segments to successively lower levels in a lateral direction; and a subsequent rearrangement of .the telophragmata at regular intervals.

20. Group of discs from atrial trabecula. The discs represent various degrees and combinations of myofibril modification.

21 Atrial disc bifurcating on the right to pass into the two successive telophragmata. This is the only t^pe of disc which gives plausible basis for interpretation as a region from which a sarcomere develops (Heidenhain), but it may equally well be interpreted as a partial failure of reversion of apposed halves of two successive contraction Ijands.

TXTKi?c.\r.ATi;n niscs of Tin-: iiKAirr ok h!;i:i'

II i:. JOHMAX AND .1. H. 1)\NKS






















22 A group of atrial discs, comprising modified portion of myofibrils of the extent of one and two sarcomeres.

23 Atrial disc of fundamental straight-band form, associated with which are other complete and partial sarcomeric modifications of myofil)rils. The disc-area here stains more deeply.

24 Very complicated atrial disc of the distorted serrated type, certain portions of which are interconnected by light-staining membranes, perhaps telophragmata.

25 Transverse section of an atrial trabecula with larger peripheral disc (above) and scattered sub-central smaller disc elements.

26 Portion of a longitudinal section of a secondarily dividing fiber from left ventricle. The telophragmata in the several branches are at different levels.

27 Irregular type of disc from left ventricle apparently produced by a splitting of an originally coarse trabecula into subdivisions, with dislocation of the discs due to imequal tensions during contraction in the resulting smaller fibers. The lightly-staining connecting membranes are perhaps remnants of distorted telophragmata.

28 Semidiagrammatic drawing of a transverse section of a slender moderator band, showing peripherally on opposite surfaces two branches of the right limb of the atrioventricular bundle (A.V.B.). The muscular tissue is grouped into two large bundles (M), the larger bundle containing a large central arteriole (A), and a small peripheral venule (F). X 13 diameters.

29 Portion of moderator-band musculature in longitudinal section, showing the simpler branched condition, and the abundant band-type of discs. X 535.

30 Portion of a longitudinal section of a trabecula from the moderator band showing a central relaxed area, and terminal contracted areas. The contracted areas stain more deeply and show only the contraction bands and alternating lighter-staining discs. The contracted areas have a considerably greater diameter than the relaxed portion. Iron-hematoxylin and van Gieson's stains.

31 Group of simple discs of fiber from moderator band, overlying a nucleus- — where neither tendons not intercellular cement could be expected to occur.

32 Portion of a longitudinal section of a fiber from the moderator band, stained in iron-hematoxylin, showing the festooned sarcolemma, an intercalated disc at the level of a displaced telophragma, the Q-, J- and //-discs.

33 A band-form of disc from the moderator l)an(l extending completely around the periphery of the fiber, but passing in opposite directions on the opposite surface, thus assuming a short spiral form. Prolxxbly similar to disc, figure 3, a and h.













f : '

■ ■'. •'■ :■■












«it n* «r *>«kfW«ai«i>M*»*« 






r:.-'Zi^:rc 30


tui." I's






PLAT1<: 3


34 Portion of a longitudinal section of a trabecula from the moderator band, showing the union of the telophragmata to the nuclear wall and to the festooned sarcolemma. No indication appears of an additional mesophragma. Ironheniatoxylin and van Gieson's stains.

35 Longitudinal section of a portion of the cellular network from the right limb of the atrioventricular bundle. The majority of the cells contain two nuclei very closely associated. Occasional cells contain three or four nuclei. The cells are polyhedral in shape, variously modified (figs. 37, 42, 43 and 44). The margins appear serrated in sections (in macerated preparations the cell meml)rane has a sharp contour; fig. 44). The serration in histologic preparations is an artifact due to the non-uniform shrinkage between the myofibrils and the sarcolemma, to both of which the telophragmata are attached. The sarcolemma stains more deeply, as do also the lateral attached portions of the telophragmata. X 108.

36 More highl}- magnified portion (.r) from figure 35. The intracellular myofibrils are seen to be aggregated into smaller groups, more abundant peripherally. The peripheral groups follow in general the cell contour; hence the myofibrilgroups are irregularly disposed with respect to each other, and the telophragmata are distorted and apparently in places interrupted. Numerous intercellular spaces occur. Certain fibrils of the myofibril-bundles are continuous from cell to cell, forming thus a syncytium in spite of distinct cell-walls. The sarcolemma and the attached portions of the telophragmata stain more deeply than the central fibrils. X 666.

37 More highly magnified cell, of modified polyhedral form, from the portion of the atrioventricular bundle shown in figure 35.

38 Small area of ventricular myocardium of fetal heart towards end of second month, showing several adjacent muscle cells in longitudinal section. The cells are long, slender, fusiform elements resembling definitive smooth-muscle cells. The myofibrils are apparently continuous from cell to cell. They are relatively meagre in amount and peripherally arranged. The telophragmata are conspicuous but very delicate. Peripherally, among the fibrils, Q-discs are also barely discernible midway between successive telophragmata. Simple, delicate, deep-staining, granular intercalated discs also occur peripherally at telophragma levels, apparently as modifications of this membrane. The discs do not occur in the areas of fusion between adjacent cells, but at right angles to such surfaces. X 1000.

39 Portion of longitudinal section of ventricular myocardium of fetal heart of fourth month; showing the process of fusion of the slender fusiform cells to form trabeculae, and a few simple band-like discs.

40 Portion of longitudinal section of ventricular myocardium of fetal heart of seventh month, showing numerous band-like discs.

41 Portion of a longitudinal section of a coarser trabecula from an adult Limulus-heart, including two intercalated discs, apparently forming a two-step type, separating an upper contracted from a lower uncontracted region. The discs are at telophragma levels and apparently represent modified contraction bands. The illustration shows clearly the close resemblance between the discs of the adult Limulus heart and the simplest type of disc of mammalian lioarts.

{Continued on page 338) 336


II. K. JORDAN AND .1. li. liANKS



i 3 3


— 39




38 z::*h

. ... (^-J^>i


m/^j ^\U~x











(Coidimied from page 336)

42 Transverse section of a central cell from a large strand of the right limb of the atrioventricular bundle, containing an arteriole. This condition vi^as probably attained by an adai)tation of the cell about a closely apposed blood vessel.

43 Transverse section of a smaller strand of the right limb of the atrioventricular bundle. This shows the variable shape of the cells, and the enveloping connective tissue capsule. Between the sarcolemma and the capsule a space invariably occurs, perhaps of the nature of a lymphatic channel, but not lined by endothelial cells. The serrated borders and the perinuclear clear spaces are fixation artifacts.


p:xplanation of figures

44 (a, b, c, d, e) Various forms of cells from a macerated preparation of the right branch of the atrioventricular bundle. The cells vary considerably in form and size, but can all be referred to a fundamental spherical or polyhedral form. In unstained preparations the paired nuclei are somewhat darker than the cytoplasm, and are homogeneous in appearance. The cytoplasm appears granular, and denser peripherally. The myofibrils are barely discernible. They occur in all portions of the cytoplasm, though more abvmdantly peripherally, are very irregularly disposed, and are continuous from cell to cell. X 535.

45 Transverse section of smaller terminal subdivision of the left limb of the atrioventricular bundle (Purkinje fibers) showing the pericellular lymphoid spaces and the connective-tissue capsular-stroma.

46 Cell from area of transition of left limb of the atrioventricular bundle to the Purkinje fibers. The cell, still enveloped by a capsule, has elongated into a fiber of Purkinje. The cell is apparently in a contracted condition. Centrally it contains a large sarcoplasmic area free of myofibrils. The cell contains a few intercalated discs at the levels of displaced telophragmata.

47 Transverse section of two adjacent Purkinje fibers. The myofibrils are arranged in the form of delicate radial lamellae. The cells are enveloped by a connective tissue membrane.

48 Longitudinal section of transition area from atrioventricular bundle to Purkinje fibers. The Purkinje fiber (cell) is surrounded by a lymphoid space and contains a step-form of intercalated disc. It probably represents a fusion product of two originally distinct cells.

49 Transition area between atrioventricular bundle and myocardium of moderator band. The upper portion of the illustration shows Purkinja cells, the lower, myocardium with band-discs. In figure 50 are shown the cells of the atrioventricular bundle which ends in the Purkinje fibers here shown. X 535.

50 Group of cells from the atrioventricular bundle of the moderator band. Intercellular bridges are conspicuous; these are composed of myofibrils passing irregularly between adjacent cells. X 535.

51 Portion of longitudinal section of muscle of moderator band, corresponding to the Purkinje fibers of the myocardium. Intercalated discs are forming in relation to the fused sarcolemmae of these three adjacent fibers. The forming Purkinje fiber is here represented by three originally distinct cells. The discs do not generally arise in the line of fusion but at angles to the fusion-area.





.3 '




m i






S*-' /







<'^- 5!





AIMEE S. VANNEMAN Contribution from the Zoological Laboratory of the University of Texas, No. 137



Introduction 341

Structure of germ cells 346

Germ cells in early stages 347

Migration of germ cells 350

Discussion 354

Summary and conclusions. 356

Bibliography 357


In addition to the interest attached to a problem of this sort, there are thi'ee reasons for undertaking the work of the present paper. Fu'st of all, the germ cells of the armadillo are remarkably conspicuous, even in young stages. Probably no other mammal so far studied for the point in question, offers such possibilities for the solution of a yet unsettled problem. The germ cells of the armadillo, besides being clear-cut are easily traceable thi'ough the tissues, without the characteristic details found necessary for recognition in forms investigated by other workers. Again, the armadillo presents a problem of unusual interest in being a polyembryonic form. Here, it is a question as to whether or not the germ cells of the embryos of a given blastocyst have a common origin. It is to be noted that no one has ever traced germ cells with certainty to a pre-embryonic stage in a polyembryonic form. The discovery by Swift ('14, '15), that germ cells in the chick migrate by way of the blood-vascular system, stimulates further investigation as to the path of ixiigration of




germ cells in mammals. The question arises whether the germ cells have been overlooked in the blood vessels, or whether they never occur in the vascular system of mammals. It is because of the before-mentioned distinctness of the germ cells that the armadillo has been chosen and found peculiarly favorable for this study. The species here used is Tatusia novemcincta.

Although in recent years a large amount of work has been done on the origin and history of the germ cells in various forms, yet in vertebrates, relatively little has been accomplished along this line. The question in this case is a far more difficult one, as is acknowledged by all investigators who have attacked the problem. The possibility of distinguishing germ cells from somatic cells in most vertebrates is rendered smaller by the absence of characteristic yolk substances and further, by the appearance only in advanced stages of the various so-called Keimbahn determinants, which serve so admirably for the definite recognition of the germ cells of certain invertebrates. Modern advances in cytological methods, however, give promise of a solution to the problem, as is evidenced by the recent results of Rubaschkin ('10), Tschaschin (10), Swift ('14) and others.

It is unnecessary here to discuss the earlier literature of germ cell history except to note that Eigenmann ('97) was the first to give a detailed account of the wandering of germ cells in vertebrates. He reached the conclusion that the sex cells in fish are probably set aside as far back as the thirty-two cell stage. This is the furtherest that any vertebrate germ cell has been traced.

Among those advocating the early origin of germ cells, may be mentioned Beard ('00) in his work on Raja bates. He claims that the germ cells may be found in a stage preceding the appearance of any real embryo, the size of the cells suggesting origin after about the thirteenth division. These, he finds, journey from outside the embryonic region between the blastodermic layers upwards through the yolk sac into the splanchnopleure and gut regions.

In 1902, Woods studying Acanthas, found that the germ cells appear in the entoderm before the differentiation of mesoderm


aud also in the yolk at the meeting point of the germ layers. Thence they migrate to thie splanchnopleure and into the epithelimn around the intestine, from which point they find their way to the germinal anlage.

Allen ('06) discovered in Chrysenis that the germ cells arise in the entoblast in the region between the area opaca and the area pellucida, posterior and lateral to the embryo and wholly without the zone of gastrulation. These migrate between the entoderm cells to the mesentery and finally to the peritoneum on either side of the latter.

Rubaschkin ('07, '09, '10, '12) investigated conditions in the chick, cat, rabbit, and guinea-pig again tracing back germ cells to the entoderm with further migration through the mesentery to the sex anlage.

Among those working on mammals may be mentioned Fuss ('12) who studied the rabbit, pig and man. He concludes that the germ cells first become apparent in the region of the primitive streak where no segmentation has yet taken place in the embryo. Scattered cells also may be seen out on the yolk, whence they migrate to the intestinal endoderm and mesoderm en route to the genital region.

Especially interesting are the conclusions reached by Swift ('14) in his work on the chick. He finds that the germ cells arise anterior and antero-lateral to the embryo in a specialized region of the germ wall entoderm at the margin of the area pellucida. They first appear during the primitive streak stage and continue to arise up until the three somite stage. Although first found between the entoderm and ectoderm, they later enter the mesoderm and the forming blood vessels of the mesoderm. By the blood, they are carried to all parts of the embryo, remaining distributed till about the twenty somite stage. At this time, the germ cells begin to concentrate in the vessels of the splanchnic mesoderm. After the twenty-five somite stage they appear to have left the blood having passed out through the vessel walls into the splanchnic mesoderm near the mesentery. In the thirty and thirty-three somite stages, the germ cells are seen in the root of the mesentery and in the coelomic


epithelium, soon migrating thence into the sex anlage. It will be found interesting to use Swift's work on the chick as a basis of comparison for the present paper on the Armadillo. For this reason, Swift's conclusions are cited more or less in detail.

Many other investigations might be cited in this connection, but space and time do not permit of further enumeration.

For the material of this investigation, I am indebted to Dr. J. T. Patterson, who has generously furnished me with an almost complete series. The embryos for the most part were fixed in Bouin's fluid, and the sections stained in part in iron haematoxylin, in part with Delafield's haematoxylin. In the latter, the germ cells show up exceptionally well, as Delafield's in coloring somatic cells relatively densely, permits the large, clear germ cells to be easily distinguishable in their surroundings. As most of the slides used for this work, had previously been prepared for a purpose other than that of germ cell study, obviously no selective technique could be employed for the present purpose. However, unlike the experience of other investigators who have studied the subject, I have found that whatever the stain or fixing fluid used, the germ cells have been perfectly clear in the tissues. The sections were cut from seven to ten micra thick. It is to be regretted that the earlier stages were not cut thinner as the germ cells of the earlier stages of development are not grouped, but are few in number and often widely scattered and for this reason, more difficult of detection and demonstration in thick sections.

The following are the stages examined for germ cells: (1) early ectodermic vesicle stages, (2) early and late primary bud stages, (3) secondary bud stages, (4) early and late primitive streak stages, (5) three somite stage, (6) seven somite stage, (7) ten somite stage, (8) fourteen somite stage, (9) 4-mm. embryo, (10) 6-mm. embryo and (11) the 10-mm. embryo. Beyond the 10 mm. embryo, the germ cells are well established in the indifferent gonad, preparatory to undergoing further development. The series is unusually complete, there being but one stage lacking which would, if obtainable, be of value in tracing the germ cells. This gap occurs between the 4- and 6-mm. embryo stages, but is such that it can easily be bridged over.


It will be recalled that in his paper on Polyembryonic development in Tatusia novemcincta," Patterson ('13) has given a thorough treatment of the early development of the armadillo. For those, however, who are not sufficiently familiar with the facts to follow the migration of cells, a brief resume may be profitable. After the usual cleavage stages and the formation of a typical mammalian blastocyst, consisting of one trophoblastic layer and an inner cell mass of embryonic cells, a process of differentiation sets in through the migration of entodermal mother cells from among the ectodermal cells. These cells, directly or after division migrating to the under surface of the cell mass, presently become transformed into a continuous layer which splits from the ectoderm. Following this, the embryonic ectoderm rounds up into a spherical mass which withdraws from the trophoblast, and pushes into the vesicle cavity, becoming included in a layer of entoderm. Through a process of vacuolization, the ectoderm sphere now becomes a vesicle. It is after this stage that the primary buds first appear from thickened areas which have arisen on opposite sides of the ectodermic vesicle through a shifting of cells. The primary buds show no signs of embryonic primordia, but each directly gives rise to two secondary diverticula, thus forming four buds which soon extend and begin to show the beginnings of four primitive streaks destined to give rise to the quadruplets. Each embryo derives its ectoderm from a portion of the lateral plate, while the entoderm arises in loco from the primitive entodermal sac. This description though incomplete is sufficient to present the main points of interest, and to show that a common point of origin for germ cells of the 4-embryos of a blastocyst might, under the conditions, be considered probable.


Before proceeding to the history of the germ cells in the Armadillo, a description of the form and structure of the cells under consideration might be of value. On the whole, the form of the germ cells is almost constant from the earliest stages up till the time of the indifferent gonad. The size is equally constant^at


least within certain limits. The primordial germ cell is large, being almost twice the size of an ordinary erythrocyte, and in contrast to the surrounding tissue cells, it takes a lighter stain. It varies in shape according to its location, but is typically spherical. At times the shape is very suggestive of amoeboid movement. Especially is this true in certain early stages to be referred to later. The cell outline is always definite. This is one of the surest criteria for the identification of germ cells. The cytoplasm is very pale and seems to be concentrated more or less closely around the nucleus, while the space directly within the cell membrane is practically clear. The nucleus too, is large in comparison to the nuclei of neighboring cells, and almost without exception is spherical in form. It may be noted that very frequently the nucleus is eccentrically placed. It is usually coarsely granular in appearance and always contains one definite dark staining nucleolus, frequently two, and sometimes more in younger stages. As the material used for this work was not fixed or stained for mitochondria, these bodies were not observed. Indeed, there appeared to be little need of using such criteria for the detection of germ cells, since the latter are distinct beyond suspicion without the aid of details. This statement may possibly have to be modified in regard to very early stages. The question will receive discussion later in the paper. Concerning other criteria used by the various investigators for distinguishing germ cells, there is little to be said. Of course, yolk substance is not present in this case. Neither is the 'attraction-sphere' of Swift apparent, although the proper fixation and staining might reveal such inclusions.


Specimen 256 is the earliest stage in which I have been able to detect germ cells. It represents a condition where the lateral plates, which are to be the beginnings of primary buds, are just becoming differentiated through a shifting of cells of the young ectodermic vesicle. The plates are merely in the process of forming. At this time the germ cells are extremely few in number. Indeed, not more than two could be found. They are


situated between the ectodermic and entodermic layers of the blastocyst at points where the layers are fairly widely separated (fig. 1 and la) It is decidedly more diflftcult to locate germ cells at this time than in later stages, since the neighboring cells are naturally larger now than they are later.

Next in order is the primary bud stage represented in specimen 247 (fig. 2a). Here the germ cells are similarly located. Text figure 1 represents a reconstruction of no. 247, point X indicating the location of the germ cells observed in this stage.

Text fig. A A reconstruction of 247, point x indicating the location of the germ cells at this stage. (This figure is taken from Patterson, '13.) It will be noted that the germ cells are not found in this specimen in the primary bud regions.

It will be noted that they appear outside of the points where the primary buds are forming. No germ cells were found in the vicinity of the primary buds. There is not a sufficient number of primary bud stages in my possession to permit of stating definitely the location of these cells on the blastocyst at this time, but I am inclined to believe that they probably lie Scattered very sparsely here and there in the entoderm, chiefly, it would seem, in the regions which do not give rise to the primary buds. What signifiance this fact may have, it is hard to explain. I question whether these few stray cells, found in stages before the embryos appear, play much part in the subsequent history. To definitely settle the matter of location, careful examinations of a number of primary bud stages is indispensable. In figure


2, two germ cells are portrayed, one in the process of division. They will be seen to be connected by a frail strand of tissue to the entodermal layer. A further examination of the series of sections reveals the fact that some of the entodermal cells are cytologically very much like the germ cells. A study of young stages suggests that germ cells undergo a considerable number of divisions up until the period when they are seen to enter the gut entoderm. After this and until they reach the gonad, they remain in a resting stage, evidenced by the fact that dividing cells rarely, if ever, are seen in advanced stages of development. There is a tendency, however, for the student of germ cells to overlook dividing cells and consider them ineligible to the category of germ cells, just because of the fact that they are dividing. This, I believe, is an explanation of the frequent low count of germ cells in earlier stages. That it is not the only explanation in the case of the Armadillo, however, will be shown later. Certainly no small number of divisions must occur in early stages, for the comparison in numbers of germ cells in early and late stages is striking. The small number of germ cells found in no. 247 may be explained partly by the fact that the series was cut 10 micra thick, thus obscuring some cells which in thinner sections might have been visible. Such a thickness in older stages is not so disadvantageous, because of the greater number of germ cells, permitting of just so many more chances of cutting through a cell instead'of just missing it. The stage just preceding that described for no. 256 was carefully scrutinized for germ cells, but results were fruitless. It might be remarked that the next stage younger in my possession is an early ectodermic vesicle before the shifting of any cells preparatory to lateral plate formation— a stage considerably younger in time, even if not in appearance, than no. 256. As a matter of fact it is known that the primary buds do not start to differentiate for a considerable time after the completion of the ectodermic vesicle (Patterson, '13). Unusual interest attaches itself to the study of such an ectodermic vesicle, because of the before-mentioned possibihty of discovering some common place of origin in the vesicle wall for the germ cells of


the four embryos which are to develop in the diverticula of this same vesicle. The failure to find germ cells at this time may be due to one or more of several causes. It is possible but not probable, that grerm cells at this early period, even though present, have not yet assumed the form which in future stages become so constant and reliable for identification. Moreover, the somatic cells at this time are larger than later, having undergone fewer divisions — thus making it less easy to distinguish, by size relationship, the germ cells from surrounding cells. Again, it may be questioned whether the germ cells arise at all before the appearance of the primary embryonic rudiments — such a suggestion excluding the possibility of a common origin for germ cells in a polyembryonic form. It will be remembered that vSwift ('14) in his study of the chick arrives at the (Conclusion that the germ cells arise at the time of the primitive streak in a specialized region of the germ wall. That is, he believes that certain entodermal cells of the germ wall at this time are producing, through division, germ cells which cytologically are similar to, the cells of the germ wall. Thus, according to Swift, earlier than the primitive streak stage, germ cells, as such, are not to be found. Whether or not this fact, unmodified, holds true for the armadillo, notwithstanding that a few germ cells may be seen before embryonic primordia appear, is a question. It is the desire of the writer to demonstrate that in all probability not only the time, but the mode ol origin of germ cells in the armadillo is similar in most respects to that described by Swift for the chick.


Although the germ cells in the stages just described may be in the act of migrating, it seems best to discuss them merely as in the condition found in early stages, and to describe the migration as beginning with the secondary bud stage, from which time the wandering may more surely be followed. In specimen 290, representing , an early secondary bud stage (fig. 3a), the germ cells have become more numerous and are located on the entoderm a little lateral to the primitive streak region which


is beginning to give off a few mesoderm cells. These cells are found in the neighborhood of each of the four embryonic areas, and must either have migrated to these points along the blastocyst entoderm, or else they are arising de novo from the yolk sac entoderm at the point of contact of the latter with the embryonic area. A further discussion of this point follows in the conclusions. In any case, some of the cells are slightly amoeboid in shape and become more so in late secondary bud stages when they appear to be traveling toward the ventral and central portions of the future embryo in the posterior primitive streak region (figs. 4 and 4a). In no. 290, some of the germ cells lie in the space between the entoderm and the ectoderm in a region between the embryonic areas as shown in text figure 2, a detail figure of the same being found in figure 3. Other germ cells are to be seen in the embryonic areas. These lie close upon the entoderm seemingly in the act of pushing their way into the layer destined to become the gut entoderm (fig. 4). A considerable number of divisions seem to occur among the germ cells just prior to their entrance into the gut entoderm. In late secondary bud stages where the diverticula are undergoing a process of further elongation, conditions are similar to those just described. Frequently, germ cells are found among mesoderm cells which have budded out a distance from the primitive streak. Specimen 226 shows such a condition (fig. 4). The germ cells probably have no relation to the mesoderm cells, but have only temporarily wandered among them. By the time the stage represented in no. 276 (fig. 5a) is reached a condition in which the embryonic rudiments lie as 'slipper shaped' structures each at the terminus of an elongated canal — the germ cells are becoming fairly well established in the entoderm of the embryos which are now in a relatively advanced primitive streak stage. The position of the cells may be seen from an examination of figure 5a.

As the primitive streak advances, the entoderm previously seen as a straight ribbon of cells, now commences to thicken and push down and inward to form the intestinal groove. The germ cells, which by this time have all migrated into this layer, are



carried along in the lateral surfaces of the primitive intestine. There seems to be no evidence whatsoever, in the armadillo, of germ cells ever entering into the mesoderm or its forming blood vessels, as described by Swift in the chick. Since Swift's work is not only able but convincing, it merely remains to be said that the paths of migration in birds and in this mammal differ. It seems certain that the germ cells of the armadillo, passing

Text fig. B A reconstruction of specimen 290 (taken from Patterson, '13), showing the location of germ cells at this period. The dotted lines indicate the plane of the sections, in which germ cells were found at points (x).

along the blastocyst entoderm into the embryonic entoderm, become immediately incorporated in the intestinal wall without ever being seen to pass through the mesoderm at all. Of course this is no new thing in mammalian work, since Fuss ('12) and Rubaschkin ('08) have described similar conditions in the rabbit and pig. My observations seem merely to confirm those of Fuss on this point.

From the time of the late primitive streak up until a stage where the embryo shows a well-developed cervical flexure (figs. 6a, 7a, 8a), the germ cells remain in the intestinal entoderm.


During this time they seem to be traveling ventrally in the intestine and are distinctly amoeboid in shape (fig. 8). They are elongated and appear to be slightly smaller than before, due no doubt to the crowding among large entodermal cells. At no time durmg the history and development of the armadillo have germ cells been found in the blood vessels. Figure 6 shows the position of the germ cells in the seven, ten, and fourteen somite stages. A drawing of the three somite stage was not made, as the position of the germ cells here was almost identical with their position in the primitive streak stage before somite formation. Further description of this period of germ cell history is unnecessary. In the 4-mm. embryo, however, the germ cells begin to leave the intestinal entoderm, as shown in figure 9a, Superficially, the 4-mm. embryo is characterized by the acute cervical bend and prominent heart regions, but as yet shows no external signs of limb buds. It is at this stage that germ cells are first seen to be massing along the ventral wall of the now-closed intestine. Certain it is that germ cells are still to be seen in the lateral walls of the intestine, but their number is small (observe fig. 9). It will be noted from this same figure that a couple of germ cells are in the process of passing out of the intestinal entoderm, while one cell is already visible within ■ the loose surrounding mesenchyme, which shortly will go to form a part of the permanent mesentery. .This is a critical stage, and interesting because it so clearly presents the passage of the germ cells from the entoderm into the mesoderm. The next stage in my possession is the 5.5 mm. embryo which externally shows well developed Umb-buds. The examination of sections reveals the presence of germ cells in a well-developed mesentery. What course is followed by the germ cells in reaching this location cannot definitely be stated in the absence of an intervening stage. However, with ones knowledge of the formation of the mesentery it is not difficult to conceive of how this might happen. It is probably not amiss to say that, as the intestine continues to round up, the germ cells which have migrated into the loose mesenchyme around the intestine pass up and forward, through a process of growth and shifting of the


tissues, and also through their independent amoeboid movement into the forming mesentery. The germinal epithelium is present on either side of the mesentery (fig. 10a), but as yet no thickening has occurred to form the lateral ridge. The germ cells at this time are found in equal numbers in three places. They can be seen located between the blood vessels of the mesentery as seen in (fig. 106), but are never found at this time in. the mesentery below the level of these vessels. The germ cells may also be found at the angle of the mesentery and the germinal epithelium (fig. 10). The cell seen in figure 10 is unusually large and therefore not quite typical. A number of germ cells seem to pass dorsal, above the root of the mesentery and of the region of the germinal epithelium, into the loose mesenchyme beneath the aorta. Strangely enough, at this period the cells are not particularly amoeboid in shape (fig. 10). It is noticeable, also, that they are larger than usual. In addition, the nucleus instead of being granular has become more or less reticular in appearance. While the germ cells are traveling into the germinal epithelium, the latter thickens and germ cells become embedded in it. The germ cells .are very easily distinguished from the peritoneal cells among which they lie (fig. 11), so that it is impossible to believe that they could ever be derived from these cells.

By the time the embryo is 10 mm. long, the germ cells have ail migrated into the well-developed indifferent gonad (fig. 11). At this time as seen from the drawing they are very conspicuous for their size. The apparent increase in size is due, no doubt, to the fact that the cells are preparing for division.


The migration of the germ cells froni the entoderm to the sex anlage is unmistakable. Throughout, the germ cells can easily be followed. But the question as to the origin of these same cells remains somewhat doubtful, although the writer is of the opinion that the conclusions reached in this paper are of rather a convincing nature. The examination of stages now at hand has


broiig-ht out several interesting facts which suggest reasonable conclusions as to the origin of the germ cells in the armadillo.

It was pointed out in the first part of this papers that no germ cells of the character of those seen in later stages could be found in the wall of the early ectodermic vesicle, a stage which long precedes the laying down of any embryonic primordia. It was thought that possibly such a stage might reveal a definite point along the vesicle wall, where germ cells might be seen to be localized, previous to scattering and migrating into the future embryos. In this sense, one might attribute, in a polyembryonic form, a common origin to the germ cells of all the embryos of one vesicle. This, however, not proving to be the case, an examination of the vesicle next in order of development — that is, a stage where the very beginnings of lateral plates can be discerned — revealed the following fact: that there exist several germ cells lying close along the entoderm wall of the vesicle outside the region of the primary buds. The cytological resemblance of these cells to adjacent entodermal cells, and the presence of a dividing cell at once suggests the possibility that here, for the first time, germ cells are being proliferated. But in the two primary bud stages examined, there were found present in each blastocyst no more than two germ cells. This condition is in contrast to that of the secondary bud stage when the germ cells are relatively numerous in the region of each of the newly forming embryos. The germ cells of each quadruplet all at once become visible in the respective embryonic areas, without having been seen to migrate there — except for the few cells seen traveling between the embryonic areas of specimen 290 (text fig. 2) . As was mentioned earlier in the paper dividing cells are not infrequent during this period. A consideration of all observations would point to the fact that active germ cells do not arise, at least in any numbers, until the secondary bud stage is reached. The few germ cells appearing before this time may be said to have arisen more or less accidentally in anticipation of the later stage. Some of these cells doubtless migrate towards the embryonic areas; others, however, probably degenerate. Certainly their number is too few to warrant the belief that all the


germ cells of the future embryo arise before the appearance of embryonic primordia. Indeed, I believe that these early germ cells play but a feeble role in the origin and future history of the germ cells. Therefore, while recognizing that stray germ cells may be found as early as the young primary bud stage, the writer beUeves that the active germ cells of embryonic life arise for the most part at the very early primitive streak stage of the embryos. Such an origin for germ cells is in general, similar to Swift's findings in the chick, both as regards place, method and time. The entoderm of the mammalian blastocyst is analogous to the yolk sac entoderm of lower vertebrates. It is not unreasonable to suppose that in the armadillo the germ cells arise during the secondary bud stage in the embryonic areas through the influence of the ectodermic vesicle upon the blastocyst entoderm at the point where the two layers come in contact. Observation seems to confirm this. That the germ cells have not arisen in numbers any earlier may be due to the fact that there exists previous to the early primitive streak stages no incident, such as the coming in contact of ectodermic and entodermic layers, to favor the proliferation of germ cells.


1. The germ cells of the armadillo are conspicuously large, and first discernible along the entodermic wall of the blastocyst, just preceding the primary bud stages. They are extremely few in number. The active, embryonic germ cells, however, probably do not arise untrl the time of the secondary bud stage appearing in the vicinity of each of the four embryonic areas.

2. During early primitive streak stages germ cells are seen dividing, pre\'ious to pushing a way into the entoderm of the future gut region.

3. After gaining entrance into the gut entoderm, the germ cells are carried in the thickening intestinal wall as, during the somite stages, it rounds up to form a closed tube.

4. By the time the embryo has attained a length of 4 mm. and has a pronounced cervical bend, the germ cells may be seen


in the act of leaving the ventral, intestinal wall to enter the surrounding mesenchyme tissue. They are amoeboid in shape.

5. In the 5- and 6-mm. embryos, the germ cells appear at the base of the well-developed mesentery, usually not below the level of the three blood vessels of that region. They are also present in the loose mesenchyme under the aorta, and en route to the germinal epithelium, which has not yet thickened.

6. In the 10-mm. embryo, the germ cells are established in the indifferent gonad. They are slightly enlarged, preparatory to division.

7. A study of early stages suggests that germ cells may arise from certain cells of the blastocyst entoderm (yolk-sac entoderm) during secondary bud formation.

8. The path of migration is from the embryonic entoderm into the intestinal wall, thence into the surrounding mesenchyme to the mesentery, and onward into the germinal epithelium. No germ cells are found at any stage in the blood vessels.

9. It may be concluded that the germ cells of the four embryos of one vesicle do not have a common origin, in the sense of having arisen from a prelocalized region of the early plastocyst.


Allen, B. M. 1906 The origin of the sex-cells of Chrysemis. Anat. Anz.,

Bd. 29.

1907 An important period in the history of the sex-cells of Rana

pipiens. Anat. Anz., Bd. 31.

1911 Origin of sex-cells of Amia and Sepidosteus. Jour. Morph.,

vol. 22. Beard, J. 1900 The morphological continuity of the germ cells of Raja bates.

Anat Anz., Bd. 18.

1902 The germ cells of Pristuirus. Ibid. 21. Berenberg-Gossler, H. von 1912 Die Urgeschlechtszellen des Hiihner embryos u. s. w. Arch. Mikr. Anat., Bd. 81. DoDDS, G. 1912 Segregation of the germ, cells of the teleost Sophius. Jour.

Morph., vol. 21. EiGENMANN, C. 1896 Sex dinerentiation in the viviparous teleost Cymato gaster. Arch Entwich., Bd. 4. Federow, V. 1907 Ueber die Wanderung der Genital-Zellen bei Salmo Fario.

Anat. Anz., 31. Fuss, A. 1912 Ueber der Geschlechtszellen der Menschen und der Saugetiere.

Arch. Mikr. Anat., Bd. 81.


Hegner, R. W. 1909 The origin and early history, of the germ cells in some chrysomelid beetles. Jour. Morph., vol. 20.

1914 Studies on germ cells. Jour. Morph., vol. 25.

1915 Studies on germ cells (continued). Jour. Morph., vol. 26. Jarvis, May 1908 Segregation of germ cells of Phrynosoma cornutum. Biol.

Bull., vol. 15.

Kellicott, W. S. 1913 A Text-Book of General Embryology.

Montgomery, T. H. 1911 Differentiation of the human cells of Sertoli. Biol. Bull., vol. 21.

Newman, H. H. and Patterson, J. T. 1910 Development of the nine-banded armadillo from primitive streak stage to birth; with especial reference to the question of specific polyembryony. Jour. Morph., vol. 21, no. 3.

Patterson, J. T. 1913 Polyembryonic development in Tatusia novemcincta. Jour. Morph., vol. 24, no. 4.

Reagen, F. p. 1916 Some results and possibilities of early embryonic castration. Anat. Rec, vol. 11, no. 5.

RuBAscHLiN, W. 1908 Zur Frage von der Entstehung der Keimzellen bei Saugetieren Embryonen. Anat. Anz., Bd. 31.

Schapitz, R. 1912 Die Urgeschlechtszellen von Amblystoma. Arch. Mikr. Anat., Bd. 79.

Swift, C. H. 1914 Origin and early history of the primordial germ cells in the chick. Am. Jour. Anat., vol. 15.

1916 Organ of sex-cords and definitive spermatogonia in male chick. Am. Jour. Anat., vol. 20, no. 3.

TsCHASCHiN, S. 1910 Ueber die Chondriosomen der Urgeschlechtszellen bei

Vogelembryonen. Anat. Anz., Bd. 37. Wilke, G. 1912 Zur Frage nach der Herkunft der Mitochondrien in den

Geschlechtszellen. Winiwarten, H. von 1901 Richerches sur I'ovogenise et I'organogenese des

Mammiferes. Arch. Biol. T., 17. WiEMAN, H. L. 1910 A study of the germ cells of Septinotarsa signaticoUis.

Jour. Morph., vol. 21. Woods, F. A. 1902 Origin and migration of the germ cells in Acanthias. Am.

Jour. Anat., vol. 1.


The drawings were made at table level with a Spencer camera lucida. For the detailed drawings ocular 6 and Spencer 1.5 mm. apochromat N. A. 1.30 objective were used. For figures la to 11a ocular 2 and 16 mm. apochromat N. A. 0.30 objective were employed. For the most part the drawings were made from sections 8 to 10 micra thick.



1 Section through specimen 256 (early primary bud stage) showing a germ cell lying between entodermic and ectodermic layers. The rectangle represented as figure la indicates the area from which this drawing was made.

2 Detailed drawing from the area indicated in figure 2a (primary bud stage) showing a germ cell and a large dividing cell joined to the entoderm by a thin strand of tissue.

3 Detailed study from specimen 290 (early secondary bud stage) showing germ cells lying close to the entoderm among stray mesoderm cells of the primitive streak area.

4 Later secondary bud stage showing a section through one of the buds (fig. 4a).









'US ^>-T^







i^FSsV '

r.- -'







5 Section through specimen 276 to show the position and detail of germ cells in the area indicated in figure 5a.

6 Detailed study of a section through the posterior third of a seven somite embryo showing germ cells in the gut entoderm. The condition in the three somite stage is similar.

7 Section through the primitive intestine of specimen 449 (10 somite stage) depicting the position of the germ cells, and the comparative size of germ cells and erythrocytes. Refer to figure 7a for orientation.

8 Section through the gut region of specimen 365 (14 somite stage) showing the amoeboid shape of the germ cells at this time. The rectangle in figure So roughly indicates the area from which the drawing was made.






A ^•M;/.



i-y (5




. .^

>: 5

■ >






.- ^C^:? *



D ,.

-,■ ,. i<t,:_

■; /I



) :V-1_ V..'

.;. ■'H

. j^^-:-;




9 Section through the closed intestine of a 4 mm. embryo. Note that two germ cells have already left the intestinal entoderm. See corresponding sketch (fig. 9a).

10 and 106 Representing two detailed studies from the area indicated by the rectangle in figure lOo. (Transverse cut through 6 mm. embryo.) Figure 10 shows a section through the germinal epithelium. Figure 10b represents a section through the base of the mesentery. Note the enormous size of the germ cells.

11 Representing a section through the well developed indifferent gonad of the 10 mm. embryo. The portion of the gonad is indicated by the letter g in the corresponding sketch (fig. 11a).













^'.-•-.^r ->iV iH^S>.<^ .f).' ^ sD r>,. '■•v:


d '^'




^T^ 'ft (?D'"^.sv, ■'^•.W"■





\ # \ /

II a




Dcparhncul of Annlomjj, Washington, Unircrsiti/ Medical School, Sf. Louis



111 1873 von Ebiier described the acinal glands in the base of the tongue. Since then contributions to our knowledge of these glands have been mainly of a topographical or comparative nature. Surprisingly little work has been done on the development of the lingual glands. Graberg's ('98) figures of sections showing the early origin of the serous glands of the vallate papillae in man form the basis of the description of the development of the serous glands in Keibel and Mall's Embryology. He figured them as early lateral outgrowths from the epithelial walls of the papillae which in later development branch considerably but are not fully developed in a 56 cm. child. This work, together with Oppel's ('99) excellent figure of the topography of the lingual glands in various mammals, is particularly enlightening. Oppel's study of the arrangement of the lingual glands in man is based upon a single specimen. Maziarski ('01) gives a brief illustrated descrii^tion of a model of a small portion of the glands of a child of fourteen years.

Since there has been so little work done on the development of these glands and on their adult condition, an intensive study of the glands in various stages of development as well as a further iiKiuiry into their topographical distribution in late fetuses and the newborn, may be of interest. The to]3ographical distribution in the newborn, the arrangement of ducts and gland

1 I wish to thank Prof. J. Playfair McMurricli for the privileges of his hiboratory during a part of tlie time while this study was in progress.



masses, the distribution of blood vessels and the topographical relations of ducts, glands and blood vessels are considered in this study.

The greater part of the material used was preserved in formalin. Serial sections of the entire caudal region of the tongue of the younger fetuses form the basis for the study of the glands. Wax reconstructions of various stages of glandular development were made according to Born's method.


As indicated by Graberg, the vallate papillae first appear as solid epithelial do^\^lgrowths. A surface view of the tongue of a 6.5 cm. fetus shows no papillae under a low power lens although the foramen caecum is well developed. However, sections show seven papillae, well defined by epithelial downgrowths. Their arrangement is characteristic. No glandular outgrowths are distinguishable.

The earliest glandular outgrowths from the epithehum appear in an 8.5 cm. fetus. There are nine well developed papillae, the fifth lying in the foramen caecum and forming the apex of the 'V.' As yet there are no grooves outlining the papillae. A reconstruction of the anterior papilla of the right side shows a slight indication of a groove surrounding it and the papilla slightly raised above the surface of the tongue. The anlagen of four glands project from the lower border of the epithelial wall, two occupying a lateral position and two a medial (fig. 2). The caudo-lateral anlage is ridge-like wdth two protruding ends; the antero-lateral and the caudo-medial are rounded elevations; and the antero-medial has two slightly extended ends.

Several bulb-like outgrowths are found on the lower border of the epithehal wall of a vallate papilla from a 9 cm. fetus.

Of the eight papillae in the tongue of a 10 cm. fetus the one in the foramen caecum is the largest. This one and the two adjoining it on the right side were reconstructed. Deep furrows separate the papillae from each other. Of the three papillae modeled, the anterior one is di\dded into four parts by epithelial partitions. Both the partitions and the surrounding


wall bear glandular outgrowths. The second papilla is simple. Five glands project from the lower border of the epithelial wall, two of which have enlarged ends and slightly constricted necks. The third papilla modeled, that in the foramen caecum (fig. 3) shows many glands at its lower border, some having the same appearance as the two above described. The lateral walls of this papilla also bear gland anlagen resembling in some cases folds of epithelium.

Nine papillae are present in an 11.5 cm. fetus. One papilla on the left side, near the foramen caecum, was reconstructed. Most of the glands extend downward and slightly caudal ward. Three glands (fig. 4) are longer than the others, the constricted necks having apparently elongated. Gland anlagen are to be found on the outer surface as well as on the lower border of the epithelial wall.

The tongue in a 12.5 cm. fetus shows twelve papillae. In sections, a very slight furrow is present indicating the site of the developing groove. This groove is only indistinctly indicated in a model of the two right anterior papillae. A model of one of these papillae, with its glands, is shown in figure 5. The glands are elongated greatly, their ends are enlarged and the stalks constricted. The stalks and occasionally the bulblike ends have lumens. The walls of the ducts are formed by two rows of epithelium but the walls of the bulbous ends contain four or five rows. In both of the papillae modeled the greater number of glands are found at the anterior and posterior ends. These glands are longer than those at the sides. Five of the nineteen glands arise from the outer wall, one from the inner, and the remaining thirteen from the lower border of the epithehal wall of the papilla. Two of the caudal glands have arisen so close to each other that they give the appearance of branches from a single outgrowth. The larger of these glands divides almost immediately, one branch extending caudalward almost in a horizontal plane, the other extending downward and caudalward for a short distance, then again dividing. At the pomt of the latter division, the duct is somewhat enlarged and has a well-defined lumen. The two subdivisions project straight


down ward, one sc^iuliiig off n short caudal branch. Fi'oni the origin of the caudal branch, the duct enlarges gradually up to the end piece.

The condition just described seems to be true of all of the longer glands. That the end pieces are distinctly enlarged is apparent in sections as well as in reconstructions (fig. 5). The glands of the papilla occupying the foramen caecum are more highly developed than those of the other papillae, as evidenced by more branching and the greater length of the ducts.

Nine vallate papillae are present in the tongue of a 15 cm. fetus. Reconstructions were made of the two anterior papillae on the right side. In one a very long gland extends deeply into the tongue (fig. 6) . At its origin from the lower border of the wall of the papilla two short glands are found. The long gland extends downward about 0.5 cm. then divides into two branches. Both of these subdivide, a subdivision of each branch going lateralward. All carry enlarged knob-like masses at their ends. These show beginning subdivision into several parts. About 0.15 mm. from its origin, several branches are given off from the long gland. The latter show small rounded masses constricting from the end bulb of each (fig. 6).

Other glands of the same papilla are short and show occasional anlagen of lateral branches on the stalks. Some small glands, \\dthout terminal enlargements, are present on the lateral walls of the papilla.

A tongue from, a specimen slightly smaller than the previous one (14.5 cm.) also has nine papillae. In this and other specimens, some papiUae, when examined under low power lenses, appear to consist of several small, closely crowded papillae enclosed by one furrow. The condition described in the 10 cm. fetus, viz., epithelial downgrowths subdi\iding the papilla into smaller, closely associated ones, is foimd here. Four of the nine papillae in a 14.5 cm. specimen were of this compound form, and for the first time, a well-formed surrounding groove is present. The papillae are somewhat raised above the level of the dorsum of the tongue. A reconstruction of the right anterior papilla shows fourteen glands in various stages of development.


They are more branehccl than those of the 15 cm. specimen studied, this being particularly true of the terminal portions of the ducts. The gland ducts frequently divide dichotomously, although occasionally they resolve into three or four branches, ^^onie of the glands extend down into the muscular tissue as far as the transverse muscle layer, where they spread into terminal branches. The terminal branches, as a rule, ran horizontally, sometimes with many turns. A few, however, are so situated that their secretions are emptied into the main duct against its stream. The three largest glands of the papilla occupy a medial position. One of these shows especially well a terminal arborization similar to that seen in figure 6, as well as beginning alveolar subdivision of the end masses. In the shorter glands the end pieces do not show^ as yet this formation of alveoli. Only two glands arise from the outer wall of the papilla. Another, possibly the anlage of a mucous gland, has its origin from an epithehal fold lateral to the papilla. It shows, howe^'er, the same terminal enlargement as is characteristic of the glands of the vallate papilla.

One papilla of a specimen 19 cm. long was reconstructed. This specimen was singular in that it presented so many gland ducts to each papillae. The papilla chosen for reconstruction lies on the left side near the foramen caecum. With this papilla seventy-seven ducts are associated, w^hile with another papilla on the left side, one hundred and four are present. In the one reconstructed, tW'O glands arise from the inner w^all of the papilla, the others coming from the lower border and outer wall. Some ducts are 2 mm. long, others very short. The short glands are characterized by short side branches and enlarged end pieces (fig. 7) The breaking up of the end pieces is advanced far beyond that in younger specimens and is apparently a constricting of parts to form small round, or ridge-like alveoli, the latter connected by a long narrow base. Older stages demonstrate that these may first sejiarate in the middle, having the ends attached, and thus form anastomosing alveoli. The serous glands of the vallate papillae can, therefore, in later fetal stages, be considered as branched alveolar glands. Both the end


pieces and the stalks have lumens. In another specimen 19 cm. long, a spherical thyroid-like mass is present in the base of the tongue above the hyoid bone. It is made up of follicles which contain a colloid-like substance. No connection between this mass and any duct system is apparent.

Sections of a tongue of a 23.5 cm. fetus show twenty-six and forty ducts respectively in connection with the first and third papilla on the right side. The second papilla on the left side is provided with forty-five ducts. One of the ducts in this specimen has a greatly dilated end from which small ducts radiate in all directions. This cystic enlargement, irregular in shape, is lined by a layer of flattened epithelium and measures 0.6 by 0.4 by 0.15 mm. in its greatest diameters.

The vallate papillae of a fetus 25 cm. in length, of a new born, and of a nine months old child, all show the characteristic short and long glands sending off many branches wdth terminal arborization of alveolar-like glands with occasional anastomoses. The longest ducts extend into the upper strata of the transverse muscle, the gland masses being broken up by the vertical and the longitudinal muscle fibers. A small group of glands in the tongue of the nine months old child were modeled (fig. 8). Some of the glands are irregular in shape, showing constrictions, outpouchings and anastomoses. The method of formation apparently is as pre\'iously described. The main ducts frequently present many small, solid outpouchings, the anlagen of other gland groups. Two such anlagen attached to the large duct near the gland appear in the model (fig. 8).

Two papillae in a newborn have associated with them thirtyeight and forty-three ducts respectively. In another specimen three papillae have thirty-two, thirty-three, and thirty-eight ducts resj>ectively. One of the latter with its gland gi'oups, from the caudal end of a papilla, was reconstructed. Twentyone groups of glands are attached to this main duct by means of small lateral or terminal branches (fig. 1). The gland masses extend beyond all sides of the papilla, spreading antero-posteriorly 2.2 mm., laterally 1.3 mm. and projecting into the tongue tissue about 1.2 mm. Since the antero-posterior diameter of


this papilla is only 0.5 mm. and the lateral diameter 0.56 mm., one can readily see that there must be great intermingling of glands when there are thirty or forty such ducts, or one hundred as was found in a younger specimen (19 cm.).

A graphic reconstruction from the newborn to show the position of the papillae and the distribution of the serous glands resembles in all respects the reconstruction of Oppel ('99).

Since the glands of the newborn are not of the branched, tubular type of the child of fourteen years as modeled and described by Maziarski, reconstructions of glands from specimens

Fig. 1 Drawing made from a photograph of a reconstruction of a papilla and one duct with its gland groups from a newborn. X 65.

of intermediate ages were made in order to determine the character of the transition between these forms. From a reconstruction of a small group of glands of a nine months old child it appears that these resemble closely those of the newborn (fig. 8). The alveolar masses occasionally anastomose although this may not be apparent from a surface view. As stated above the method of development readily accounts for this anastomosing of end-pieces. In the newborn and nine months old child, the larger ducts frequently present small irregular outpouchings connected by short, constricted stalks. Some of these show the beginnings of secondary alveolar-like sacs. These groups may develop into glands similar to those already formed, or remain in a more or less undeveloped state.


In the specimen from a child of five years, as also in specimens from older individuals, mucous glands, either as single alveoli or in groups, join the ducts of serous glands. Occasionally a part of an alveolar group is formed of serous cells, which are succeeded by cells distinctly mucous in type.

A reconstruction of the glands from a twenty-two year old specimen shows anastomoses between the closely crowded alveoli. Some of the glands are somewhat tubular, although they generally appear to be more of the alveolar type (figs, 9 and 10). The main ducts are distinctly different in structure from the end pieces, but the terminal ducts, breaking up within a group or lobule, may be similar to the glandular end-pieces. Anastomoses occur between alveoli of two terminal ducts as well as between those from one duct (fig. 9). Irregular outgrowths of the main ducts noted in younger specimens are present also in this specimen; these outgrowths extend in every direction and some are just beginning to break up into end-pieces (fig. 9). Figure 9 shows a gland from one terminal duct anastomosing with one of these gland anlagen. The lumen in the specimen could not be traced from one duct to the other through the gland mass. However, the lumens are sometimes very minute, even in larger end-pieces. Figure 10a shows an alveolar-like endpiece connected with the terminal duct; from the former three alveolar-like end-pieces project in various directions. Figure 10 shows a group of glands from which a portion has been removed in order to show the terminal duct with its various end-pieces.

The histological structure of the serous glands from the twenty-two year old specimen varies greatly. In some of the glands the secreting cells are large and deeply-stained; in others the lumens are large, whereas the cells appear flattened. These differences are probably due to different stages of functional activity. .

Serial sections of a vallate papilla from a man fifty years old presented very closely crowded glands.

Although this work is mainly a study of the development of the serous glands and their topographical relations, some at


tention is given to their histological structui'e, especially in the better preserved material from older individuals.

An attempt to include the taste buds in the models was met with only partial success. Few taste buds appear in stages before that of the 19 cm. fetus, but in this specimen the number is relatively large. In all cases taste buds are more numerous on the sides of the papillae although some are present on the summit. They are also found on the dorsal surface in the newborn. In none of the specimens is there any definite arrangement of the taste buds in rows and tiers as has been described in sheep and pig by Schwalbe ('68).

In several specimens the lingual artery was injected and the materia) sectioned and studied. A number of rather large arteries ascend obliquely toward the serous glands about the vallate papillae. Smaller vessels enter a gToup of glands and then subdivide. Some of the latter vessels leave the glandular tissue and supply the surrounding musculature. The arteries do not follow the main ducts, or the terminal ducts of the lobules.


As has been stated, the earliest glands are downgrowths of the lower border of the papilla. Graberg ('98) figures the first outgrowths from the lateral wall of the papilla. My models show that lateral outgrowths are not infrequent but that the gland anlagen which first appear in about 8.5 cm. fetuses are on the lower border (fig. 2) of the papilla. Oppel ('99) gives us an excellent figure showing the topography of the serous glands. The conditions there sho^\^l are confirmed in the present study of the serous glands in a newborn. These glands extend 3-5 mm. on all sides of the vallate papillae as has been observed by Oppel and by von Ebner ('73). The extent of the area occupied by the group of serous glands about a papilla can be estimated by reference to figure 1 which shows a single duct with its gland groups. With thirty to fifty such ducts associated with a papilla it is apparent that the glandular tissue must be crowded and extend considerably beyond the surrounding fur THE AMERICAN JOURNAL OF ANATOMY, VOL. 22, NO. 3


row. I have not "found glandular tissue within the connective tissue of any papilla although von Ebner and others have noted this.

The ducts open into the bottom of the furrow or along the lateral side of the epithehal wall, and occasionally into the medial wall. No ducts are observed opening into the dorsum of the papilla or into other grooves in the surface of the tongue (except those belonging to the folliate papillae). Nor are any ducts lined by ciliated epithelium as described by Schwalbe ('68), von Ebner ('73) and'Gmellin ('92) found in my material. Von Ebner stated that small alveoli either poorly developed or not fully formed, and separated by considerable connective tissue are not infrequently present. It is apparently such a group of glands which Maziarski ('01) has reconstructed and figured. In my material I have seen such a condition in only one group of glands from a five year old child, and in this it is rather that considerable connective tissue separated the more or less tubuloalveolar components of the glands than that the glands themselves are small. This group of glands approaches the alveolar type more nearly than does the one figured by Maziarski and as in his, no anastomoses are present. For the most part, although the lobules may be closely crowded or scattered, the individual end-pieces are usually very much crowded, and as a result are often rounded or irregular in shape. This fact may account for the occasional anastomosing of glands observed in my material. It is possible that the anastomoses are not permanent and that anastomosing end pieces may separate in older specimens. Absence of lumens in these anastomoses may indicate beginning separation. However, as stated, the lumens are often very minute and may appear discontinuously in other places.

Simple outgrowths from larger ducts are present in several of the specimens studied (figs. 8 and 9). That these are gland anlagen seems probable since various stages from the earliest outpouchings to those showing beginning glandular division are present. The serous glands therefore are not fully developed


at birth nor even at five years. At twenty-two years simple outgrowths and anastomoses are still present (fig. 9).

Schmidt ('96) and Erdheim ('04) have described several cases in which cystic glands are found associated with the thyreoglossal duct, or, as isolated structures containing no ducts. In the present study, the occurrence has been noted of a spherical, thyroid-like mass, made up of follicles and containing a colloid-Hke substance, situated in the base of the tongue of a 19 cm. specimen. In another specimen dilatations of the ducts of some of the serous glands of the anterior papillae have been observed. A homogenous mass filled the cystic parts, and a duct connected the largest cyst to the groove surrounding the papilla. It appears, therefore, that besides the cystic glands associated with the thyreoglossal duct, cystic enlargements of serous glands of the tongue might also occur.

Frequently taste buds are found on the dorsal surfaces of the vallate papillae in the newborn. In other specimens they are noted only incidentally and occasionally are reconstructed mth the papillae. It has been stated above that in some of the newborn and older specimens, mucous glands intermingle with serous glands and join with the ducts of the latter. This condition has also been observed by Maziarski and others. It is possible therefore that the ducts and glands of the vallate papillae although usually serous in type are capable of developing mucous cells or alveoli or of being transformed into mucous alveoli, or that secondary connections are established between mucous end-pieces and the ducts of serous glands. It is noteworthy that mucous alveoli joined to ducts of the serous glands are observed only in adult material.

From the distribution of the blood vessels in the gland groups, it does not appear that the latter conform to our usual conception of lobules or histological units of organs.


Serous glands first appear in 8.5 cm. fetuses as outgrowths, originating usually from the lower border, but sometimes from the outer wall of the vallate papilla.


The first outgrowth is knob-hke. Soon a stalk develops givmg rise to lateral branches with enlarged end pieces. In a 19 cm. fetus, these enlargements present bulgings of the surface and beginnings of alveoli. These retain various connections with the ducts and with each other, so that in the newborn the serous gland is of the alveolar type with some anastomoses between the alveoli. In the adult (twenty-two years) some of the glands are of the tubular type with, some anastomoses between end-pieces of the same and separate ducts.

In the newborn, many knob-like outgrowths appear on the large ducts; in older specimens the number is less. These outgrowths are probably the anlagen of future glands, or at least potential anlagen.

Cystic dilatations of the serous ducts may occur. ■ Mucous end-pieces occasionally open into the ducts of serous glands of the vallate papillae.

The serous glands of the vallate papillae of man belong therefore to the branching tubulo-alveolar and not to the branched tubular type as stated by Maziarski.


Von Ebner, V. 1873 Die acinosen Drlisen der Zunge und ihre Beziehungen zu

den Geschmacksorganen. Graz. Erdheim, J. 1904 I. Ueber Schilddriisenaplasie. II. Geschwiilste des ductus

Thyreoglossus. III. Ueber einige menschliche Kiemenderivate.

Beitr. z. path. Anat. und allg. Path., Bd. 35. Gmelin, a. 1892 Zur Morphologie der Papilla vallata und foliata. Arch.

mikr. Anat., Bd. 40. Graberg, J. 1898 Beitrage zur Genese der Geschmackknospen des Menschen.

Morph. Arb., Bd. 8. Maziarski, S. 1901 Ueber den Bau und die Einteilung der Drlisen. Anat.

Hefte, Bd. 18. McMuRRiCH, J. p. 1912 In Keibel and Mall, Manual of Human Embryology.

Lippincott, Philadelphia. Oppel, a. 1899 Zur Topographic der Zungendriisen des Menschen und einiger

Saligethiere. Festsch. z. 70. Geburtstag von C. v. Kuptfer. PoDWisoTZKY, V. 1878 Anatomomische Untersuchungen iiber die Saligethiere.

Inaug. Diss. Dorpat. ScHMiTT, M. B. 1896 Ueber Flimmercysten der Zungenwurzcl und die drlisigcn

Anhjinge des Ductus Thyreo-glossus. Fesctsch. f. B. Schmitt. Jena. ScHWALBE, G. 1868 Ueber die Geschmacksorgane der Siiugethiere und des

Menschen. Arch. mikr. Anat., Bd. 4.





2 Ventral view of a reconstruction of the right anterior j^apilla of an 8.5 cm. fetus showing gland anlagen from the lower border of the papilla. X 135.

3 Same view of a reconstruction of the papilla from the foramen caecum of a 10.0 cm. fetus. X 135. '*: ".^. \

4 Same view of a model of a papilla of the left side from an 11.5 cm. fetus, showing enlarged end-pieces and constricted stalks. X 135.





J -i^



OJ ^ ^j CI o

2 "^ m

«<-i -s ts ,^ fl d

^ =^ — ^ CO 'bC iS « 2

a. 2 a ■c s °

o ^ .


>^ CO

<« ^ *=* o o «S

<u 03 t^


a ^ - s


a -g

oi '^

Pi <u

bc a





hJ «  < ^

CO ■<

Q •

< o






7 Lateral view of a reconstruction of some short glands from a 19 cm. fetus. X 135.

8 Reconstruction of a small group of glands connected by a terminal duct to a larger collecting duct from a nine months old child. X 135.

9 Reconstruction of a small group of glands, a larger collecting duct and a small simple outgrowth anastomosing with the glands of the group, from an adult specimen of twenty-two years. X 166.

10 Reconstruction of another group of glands showing alveolar-like endpieces from a specimen twenty-two years of age. X 166. a, alveolar end-pieces opening through an alveolus into the terminal duct; D, main duct; d, terminal duct to gland group; ^, early gland anlagen from main duct.





^' '/'


author's abstract of this paper issued bt the bibliographic service october 20.



Department of Anatomy, Universitrj of Toronto



Introduction 385

Preservation 386

Parental history 386

External appearance 386

Radiographs 391

Dissection of left arm ' 395

Muscles 395

Nerves 409

Vessels 411

Embryological and general considerations 413

Diaphragmatic hernia 425

Bibliography 428


The foetus forming the subject of description in this paper exhibits the rare deformity of complete absence of the ulna in each arm, accompanied by the still much rarer condition of monodactyly (figs. 1 to 4). This latter condition is not to be confused with the relatively common condition of syndactyly, where more than one digit is present, but they are united by a web of skin and other tissues. Monodactyly, the presence of only one digit, is very uncommon, and in a search through the literature only one case was found that resembles the present one, and that was presented as a freak exhibit at a medical society, no anatomical investigation of it having been made.



I have, therefore, undertaken to work out the special anatomical details of muscles, vessels and nerves in one of the deformed limbs, in the hope that light might be thrown on some of the primitive conditions of these parts, and also with the purpose of adding a definite and exact contribution to the present inadequate knowledge of this abnormal condition. '^ Indeed the inquiry into several types of malformation and structural anomaly has repeatedly thrown light not only on the malformation or anomaly itself but also upon the normal process of development the disturbance of which it represents." — (Ballantyne) .


This specimen was not obtained until about one week after its birth, and in the meantime had been kept immersed by the undertaker who sent it to us, in an embalming solution which, as far as can be ascertained, was practically a 10 per cent formalin solution. In the laboratory it has been kept in 80 per cent alcohol. No injection of the blood vessels was attempted, and though this has added somewhat to the difficulty of dissection, good results have been obtained.


The parental history, as far as could be ascertained, is practically negative concerning the deformity iri this foetus. The parents are about twenty-five years of age, in comfortable circumstances, have good mentality and are free from venereal diseases as far as known. There have been two miscarriages previous to this one, with no deformities.


The body of the foetus (figs. 1 and 2) is that of a well developed child born at the end of the seventh calendar month of pregnancy. It is well formed, healthy looking, and apart from the upper limbs has no superficial evidence of abnormality. The sex is male, and no aberrant development of the external genitals is present. The back is strongly curved, the head bent forward.


and the legs strongly flexed and drawn up against the abdomen. On following the line of the vertebral column, a slight scoliosis is observed in the thoracic region convex to the right.

The whole body is covered with a well developed lanugo moderately dark in color, and on the head is abundant fine black hair about 2 cm. in length. Nails are present on all the digits of both upper and lower limbs, but are yet some distance from the extreme ends.

The weight of the child is 1280 grams, and the length from the vertex of the skull to the ischial tuberosity, measured over the back, is 325 mm. These measurements correspond fairly well with figures given by Keibel and Mall ('10) and by McMurrich ('15) for the seventh month.

The deformed upper extremities show an upper arm segment with the forearm flexed upon it and united to it by a web of skin, a narrow carpal region and a single digit. On the right arm there is also a single digit located at the inner side of the elbow. The general resemblance to the wing of a chicken plucked for cooking is strong, and led to the assertion that the mother's fondness for visiting the zoological gardens and watching the birds was responsible for this deformity, because she had spent much time in this way during the spring and summer months of her pregnancy. JVIaternal impressions have been credited with many strange and miraculous powers without any rational basis, and this is surely an example where a credulous imagination has been led far astray. A mere coincidence has been used to work out a sequence of cause and effect, and, like much circumstantial evidence, there is here no basis for the assumption that the two facts have in truth any association whatever. Only a very sfight knowledge of human embryology is necessary to shatter the theory in this case. The bird impl-ession, if it may be so called, seized the mother during the spring and summer when she had a strong desire to be out of doors. It may be assumed that the deformity in the limbs was an accomphshed fact when the limb skeleton was laid down and so was present at the time of the appearance of ossification in the limbs in the seventh week of development. Indeed it may even be


assumed that the deformity was already estabhshed at the time when chondrification began and its origin is thus carried back to at least the fifth week and to a time when the mother would just begin to suspect that she had become pregnant, as her expected menstrual period would then be a week overdue. No visits to the zoo were yet thought of, as this was in midwinter, and yet the deformity was even then an accomplished fact which future development could not alter, but only make more clear and accentuated.

The deformed limbs will now be described in more details In each arm (figs. 3 and 4) the shoulder and scapular regions appear normal, but slightly flattened, as though from pressure from the body lying on its side. The upper arm segment lies parallel to the long axis of the body, close in at the side, and appears flattened from side to side so that its mediolateral transverse diameter is only two-thirds that of the dorso ventral. It is gently tapering in outline, narrowing as the elbow is approached. The elbow is fairly well rounded, and from it the forearm runs forward in the same plane as the upper arm and flexed on it at an acute angle, being maintained in the position by a thick web of skin extending across the interval between arm and forearm. The part of the forearm beyond the attachment of the web is rounded, with its transverse diameters about equal, and tapers gradually distally. The carpus, metacarpus and the single digit also taper continuously distally, and are all in a position of partial flexion, showing marked creases or folds on the volar surface at the line of the joints. There is a well developed nail on the digit, but it does not yet reach to or project beyond the end of the finger, as is the case in a child born at full term.

The left forearm and hand (fig. 3) are in the same plane as the upper arm and in a position of complete pronation. The hand lies against the side of the cheek, the palm facing directly ventrally. Flexion in this hand is gradual.

The right forearm and hand (fig. 4) are in a position midway between pronation and supination, a position identical with that normally assumed when the limb skeleton is first defined (Lewis, Keibel and Mall's Human Embryology). The distal end of the



forearm curves somewhat inward and the carpus is sharply flexed upon it and the hand thus comes to he across the body under the chin, with the palm facing caudally.

The left arm has no accessory appendages or indications of any of the missing parts, but on the right one (fig. 4) there is a flattened appendage attached by a very short narrow circular stalk to the medial surface of the forearm almost at the elbow. This structure widens immediately Ijeyond its attachment, being much compressed and running back applied against the surface of the arm, and from the distal part of this broader portion a narrow finger-like process extends at right angles up in the line of the limb, pointing toward the hand. This appendage strongly resembles another digit arising at the elbow.

Measurements of the foetus, and especially of the deformed limbs are here appended in tabular form:

Weight of foetus 1280 grams

Length from vertex to ischial tuberosity 325 mm.

Ischial tuberosity to bend of knee 80 mm.

Bend of knee to tip of heel 75 mm.

Ischial tuberosity to tip of heel 155 mm.

Tip of heel to distal end of digit 1 63 mm.

Deformed upper extremities




Acromion process to point of elbow

Point of elbow to distal end of radius ... .

82 mm. 57 mm. 35 mm.

35 mm.

28 degrees 50 degrees

12 mm. 18 mm. 20 mm.

74 mm. 52 mm.

End of radius to finger tip

Skin web

From point of elbow to free edge

25 mm. 37 mm.

Angle of divergence of axis of arm and forearm At rest

24 degrees 45 degrees

Extended to utmost

Extra digit

From pedicle to outer edge of broad portion

Outer edge of broad portion to tip of digit ...

From pedicle straight to tip of digit




Arm at axilla

Arm at free edge of skin fold

Arm one centimetre above elbow. . . .

Forearm where first free


Digit at middle phalanx

Extra digit


Proximal third (metacarpal region)

Middle third (proximal phalanx) . .

Distal third (distal phalanx)


33 29 27 14 12 6



mm. 19 16 18 15 12 5




30 25 23 14 10 6


20 17 17 13

The only case recorded that closely resembles this one is one reported by Barabo ('00). The complete description as given by him follows:

Ferner berichtet Hcrr Barabo fiber eine eigenartige Missbildungen an den Armen unci Hiinden eines nicht vollstandig ausgetragenen Kindes. Das Kind, 2800 gni. schwer, 46 cm. lang, war mit Wolfsrachen behaftet. Der rechte Oberarm von der Scliulter bis zur Ellbogenspitze war 7 cm. lang; der Vorderarm bis zum Handgelenk 4 cm. Am Vorderarm war nur ein Vorderarmknocken vorhanden. Von der Beugeseite des rechten Oberarmes ausgehend lief eine Hautfalte auf den Vorderarm, die 1 cm. unterhalb des Ellbogengelenks inserirte nnd den Vorderarm in spitzwinkeliger Beugestelkmg hielt. An der rechten Hand war nur ein vollkommen entwickelter Finger und ]\Iittelhandknocken vorhanden. Die librigen Finger imd Mittelhandknocken fehlten.

Der linke Oberarm zeigt ebenfalls ein Liingenmaass von 7 cm.; der Vorderarm war 5 cm. lang; die beiden ^^orderarmknocken waren normal entwickelt. Es fehlt ebenfalls die ganze mittelhand. Der Daumen, riidimentar entwickelt, sass direct auf dem Handgelenk auf und war 1 cm. lang und mit dem 4 cm. langen Zeigefinger durch Syndaktylie verbunden. Mittel- und Ringfinger fehlten. Der Kleinfinger war vorhanden und 4 cm. lang.

Der Vortragende liisst die Frage offen, ob die Missbildunge auf abschniirung durch amniotische Faden oder Hypoplasie zuruckzufiihren sei.

From the above account it will be seen that the right arm in Barabo's case shows exactly the same condition of a webbed


elbow, single bone in the forearm and monodactyly, as is shown in both arms of the foetus described by me. This is an important point as it leads to the assumption that this condition is a ver}^ definite one, which although very rare is not purely a chance occurrence but may have some definite cause. Thus it would be the concrete indication of the previous working at a certain particular period of development of some definite vicious or teratogenic influence.


Four radiographs were made of the foetus in the X-ray department of the Toronto General Hospital. Plates were made of the whole body from the front and from the side, and also special ones of each arm from the side. The definition of structures in the plates was excellent and identification of various parts was an easy task. Prints made from these plates, however, were unsatisfactory, since heavy prints intended to show structures with light shadows made heavier parts a solid mass of shadow without detail, while light prints did not bring out distinctly the lighter parts. Three prints of each plate were made, a heavy, a medium, and a light, and from these and the plates, the following description has been pieced together. The illustrations are from actual tracings from the plates and are designed to show only essential structures.

The radiograph of the left arm (text fig. A and fig. 6) shows a well-developed scapula of normal proportions, and articulating with it the humerus, which is fairly heavy and of typical shape. The upper end is well expanded as is also the lower, but as might be expected no ossification is yet present in the epiphyses. The lower end extends almost to the end of the bend of the elbow, and coming off in front of it is a single bone lying in the forearm. Owing to the cartilaginous condition of the epiphyses, no articulation can be demonstrated, only the osseous tissues showing. That this bone in the forearm is the radius is quite evident from its shape, the upper end being narrow and the shaft round above and gradually broadening as it proceeds distally, the entire bone being also slightly curved in



its length. Beyond the radius is a considerable clear interval including all the carpal and metacarpal region where there is yet no ossification, but in the single digit a small rectangular ossification is seen proximally and another occurs distally, these representing the shafts of the proximal and distal phalanges. Between the two is a clear space where the still unossified cartilage of the middle phalanx lies.

The right arm (text fig. B) presents a few differences from the left. The scapula and humerus are both typical. The humerus does not, however, reach as near the point of the elbow as does

Proxnnal phalanx

Distal phalanx


Text fig. A Sketch from a radiograph of the left arm showing the ossified portions of the skeleton.

Text fig. B Sketch from a radiograph of the right arm showing the ossified portions of the skeleton. Note the extra digit at the elbow.

that of the left arm, the end of the radius lying under it instead of in front of it. The radius is more curved than in the left arm. No carpal bones yet appear, but in the metacarpal region there is a small ossification representing the shaft of a single bone. As on the left side ossifications for the proximal and distal phalanges are present in the digit, while no middle phalanx yet shows. The proximal phalanx is not as well developed as on the left side.

The appendage at the elbow on the right limb (text fig. B) is interesting. Its pedicle appears in the interval between the humerus and radius and running dorsally in its broad part


is a well marked metacarpal ossification, and at right angles to this and lying in the narrow digital part of the appendage is the ossification representing the first phalanx. In the region of the second phalanx there is yet no bone, while the distal phalanx is represented by an extremely small centre of ossification.

Some delay is thus evident in the processes of ossification in these limbs since the appearance of the primary center in a metacarpal is usually in the ninth week and for a middle phalanx about the twelfth week, (Keibel and Mall.)

The skeleton of the lower limb (fig. 5) appears to be normal except that no middle phalanges yet show ossification. Metatarsals, proximal and distal phalanges are all ossified as are also the talus and calcaneus. The long bones are normal. Delay in ossification in the middle phalanges is again evident in these limbs.

The skull shows no abnormalities, although ossification is very heavy in the base, especially in the petrous regions and body of the sphenoid, but the vertebral column and ribs show some interesting features. The vertebral body (fig. 6) shows as a transversely oval patch with a small clear spot in the center, indicating the position of the notochord. The appearance of the body indicates the occurrence of ossification from bilateral centers or else from a center indicating a bilateral origin. The ossified part of the neural arch is still divided into its two halves, no fusion having yet occurred either with the bodies or dorsal to the spinal cord. The center in each half of the arch (fig. 5) is quite distinctly seen lying to the side of the body and on the thoracic vertebrae well marked transverse processes can also be seen. In the sacral region the centers for the neural arches are very insignificant and none are to be seen for the coccyx. The first three sacral vertebrae show a well marked center of ossification (fig. 6) on each side in the lateral mass. There are seven well marked cervical vertebrae, thirteen thoracic, five lumbar, five sacral and one coccygeal. The first sacral may be identified by the presence of the centers in its lateral masses, so that it is evident that the presacral vertebrae are twenty-five in number instead of the normal twenty-four. That the supernumary


vertebra is a thoracic one is assumed from the fact that there are the normal number of lumbars and cervicals, all typical of their region in appearance, and all free from ribs, while between these regions lie thirteen vertebrae, all of which bear ribs.

All the thirteen ribs (figs. 5 and 6) are well marked, though the first and the last are very short. It is unlikely that the rib at the upper end is cervical, or the lower one lumbar in origin in view of the fact that these regions have their full number of vertebrae without ribs.

The cause of the scoliosis mentioned previously is shown in the radiograph. The body of the third thoracic vertebra (fig. 6) is imperfect on the left of the mid-line. It shows ossification but is only half the size of the right half, and this center of ossification has remained separate from its fellow on the right side. The fourth body is slightly tilted up on the left to make up for the deficiency. The seventh thoracic vertebra on the left side of its body again exhibits the same deformity, with lack of fusion of the two centers of ossification in the body, and in this case the eighth, ninth, tenth and eleventh vertebrae, lying below it, are all tilted up to compensate for the deformity. Both defective vertebrae show good neural arches with well developed ribs articulating with them.

The only points in regard to the skeleton, which are not brought out by the radiographs, but become evident on dissection, are that eight costal cartilages articulate with the sternum, and that there are only two carpal bones. The carpal bones are not yet ossified, and so do not show in the radiographs. The proximal one is long and cylindrical, with a convex head proximally articulating with the lower end of the radius, and a concave facet distally for the other carpal. The second carpal is an irregular wedge, broad dorsally, narrow ventrally, with a proximal convex articulation for the other carpal, and a concavoconvex facet distally for the metacarpal. It is impossible to identify either of these bones with any one of the normal carpal bones, but they resemble the navicular and lesser multangular more closely than any others.



In describing the muscular system in this hmb frequent reference to variations and to comparative anatomy are made, where it would be tiresome to keep repeating the authority for such statements. In such cases it is to be considered that Le Double's book Variations du Systeme Musculaire de I'Homme" has been followed.

Where no comments are offered regarding the variations of origin or insertion, or additional attachments of any muscle noted here, it is to be inferred that such departures from normal have been frequently noted before by others, and are not of great significance.

As is to be expected, there is little change and abnormality in the muscles belonging to the upper part of the limb, but great structural differences become increasingly evident as one proceeds distally.



All the following muscles are present and exhibit normal origins and insertions (figs. 7 to 10).

Sternocleidomastoid. Subclavius.

Trapezius. Muscle fibers end at level of ninth thoracic vertebra, below this point there is only a thin aponeurosis. Rhomboidei, minor et major. Levator scapulae. Serratus anterior.

Latissimus dorsi — with an accessory head from the lower angle o' the scapula. The two pectoral muscles exhibit some variations from the normal.

Pectoralis major (figs. 7 and 8, P. Ma)

Origin. Normal. ,

Insertion. Into the outer lip of the bicipital sulcus by a heavy sheet of tendon. From the deep surface of this tendon two ab


normal accessory heads of origin of the biceps brachii are given off.

From the lower free edge of the muscle and from the main tendon there arises an aponeurotic strip which gradually narrows as it passes down the arm and forms a band arching over the biceps muscle and inserting into the medial epicondyle and the medial epicondylar ridge of the humerus. This band is the chondroepitrochlearis muscle, and is not an uncommon structure, being frequently found in the adult (8 times in 64 subjects, Le Double). Tt is much more frequent in females than in males. It is a normal part of the musculature of many of the lower animals, being known under various other names in cheiroptera, bears, foxes, Dasypus, Echidna, Batrachia and Cetacea, and is believed to be homologous with the tensor plicae alaris of birds (Le Double).

Pectoralis minor (fig. 8, P. Mi)

Origin. Statements differ in various textbooks as to the extent of origin of this muscle, some (e.g., Piersol) say the third to fifth ribs, others (e.g., Morris) include the second rib also. In this instance the more extensive origin occurs.

Insertion. The insertion is into the upper surface of the coracoid process and the outer part of the costocoracoid membrane is so intimately blended with this part of the muscle that I have debated whether or not to call it a second insertion into the middle third of the clavicle, an attachment which is occasionally exhibited. The lowest fibers are attached to the medial surface of the coracobrachialis muscle, an insertion which has been noted in other cases by Winslow (vide Le Double).


The deltoid, supraspinatus, infraspinatus, teres minor, teres major, and subscapularis are all present, and normal in extent.


Coracobrachialis (fig. 8, C)

Origin. From the coracoid process, and capsule of the shoulder joint, by a common tendon with the short head of the biceps. The capsular origin is uncommon. The muscle in its upper part receives fibers from the pectoralis minor as mentioned above.

Insertion. Into the medial side of the humerus from the level of the lesser tuberosity klmost down to the medial epicondyle. What are here present are thus all three divisions of the muscle, namely, superior, middle and inferior portions.

The superior portion here exhibited is rarely found in man though normal to some of the lower animals. The coracobrachialis superior, when present, inserts into the lesser tuberosity, surgical neck, and medial bicipital ridge of the humerus, also frequently into the capsule of the shoulder joint. It occurs only very rarely in the Anthropoidea but as a normal structure in the Quadrumana. It is also present in the elephant, giraffe, bear, cat, hyena, opossum. Echidna and several other animals.

The coracobrachialis medius is inserted into the middle portion of the humerus and forms the main mass of the normal human muscle, the remainder being constituted of the upper part of the coracobrachialis inferior. The medius is the only portion of the coracobrachialis present in the aye-aye, the bat, and the sloth, while it is absent in the kangaroo, otter, and seal.

The coracobrachialis inferior has an extremely variable insertion, extending in different cases from an attachment a couple of centimeters long on the shaft of the humerus below the medius, to an insertion on the inner edge of the whole lower half of the shaft of the bone and the inner epicondyle. In the latter case it bridges the supracondylar foramen in animals where this is present and so is perforated by the median nerve and brachial artery. This muscle is found in the cetacea, the hedgehog, the bear, great anteater and others. The inferior portion is much more developed here than is normal in man, but similar development has been frequently found before.


Between the upper and middle portions runs the musculocutaneous nerve, but there is no perforation of the lower part of the muscle by the brachial artery and median nerve, as occurs when the muscle extends as far as the medial epicondyle of the humerus. The medial edge of the upper third of the muscle is connected with the deep surface of the pectoralis major by a muscular band.


Biceps hrachii (figs. 7 and 8, Bi)

Origin. The long head arises normally from the supraglenoid tubercle of the scapula. Its tendon is very thin and narrow.

The short head is fleshy and heavy, arising by a broad tendon from the coracoid process and the capsule of the shoulder joint, the muscle formed by this head overlapping that of the long head.

In addition to these two heads two accessory heads are present on the lateral side, arising from the deep surface of the tendon of the pectoralis major and joining the long head at the level of the bicipital groove. On the lateral surface of this united bundle comes in a tough short tendon from the deltoid tubercle and under the long head there is also a distinct bundle arising from the shaft of the humerus to join the long head. There are thus seven distinct origins for this muscle. All these abnormalities have been noted by Le Double though some of them are extremely rare.

Insertion. The greater part of the muscle passes into a tough cylindrical tendon passing to the bicipital tubercle on the radius.

This is a second tendon, however, passing from the superficial and medial aspect of the muscle, as a broad fiat band with diverging crescentic edges. It is attached to the anterior surface of the medial epicondyle of the humerus, and to the shaft of the radius in front of and beyond the bicipital tubercle. Between these two points the inferior border of this aponeurosis


presents a free crescentic border under which are visible the other tendon of the biceps and the tendon of the brachiahs muscle. There is some fusion of the deep fascia of the arm to the muscle at the beginning of this superficial tendon, which might be interpreted as a rudimentary semilunar fascia.

The attachment to the humerus must be extremely rare as it has not been noted by such an authority as Le Double and no explanation of such an attachment can be drawn from comparative anatomy. The only plausible theory to be entertained is that this is possibly an extremely well developed semilunar fascia which has obtained a bony attachment by following the intermuscular septa to the bones. ,

The median nerve passes on the superficial surface of this broad tendon while the brachial artery and vein pass deep to it, and also behind the round tendon.

The biceps muscle is responsible for the position of partial supination of the radius, though the hand is pronated. It is to be remembered that one action of the biceps normally is rotation of the radius to produce supination, accomplishing this by a forward pull on the bicipital tubercle which lies posterior to the long axis of the bone in pronation. In this case the radius has been rotated until the bicipital tubercle lies facing the anterior surface of the humerus. There are no muscles attached to the radius capable of opposing the biceps in this action and so the position of supination will be permanently retained.


This muscle is divided longitudinally into two portions.

Medial portion (fig. 8, Br.)

Origin. Normal in extent from the lower half of the front of the shaft of the humerus.

Insertion. The muscle passes down on the humerus almost to the articulation with tlie radius. It is inserted along a continuous line on the back of the neck and head of the radius, the joint capsule and the medial epicondyle of the humerus dis


tally and deep to that part of the biceps tendon inserted here, and deep to the origin of the muscles of the forearm.

This portion of the muscle is supplied by the musculocutaneous nerve, which is normal, as this portion of the muscle develops from the ventral musculature of the arm.

The insertion of the brachialis on the radius is to be expected here, as the ulna is absent, and because it is a frequent abnormality to have accessory insertion on the radius in addition to its ulnar insertion. Indeed, in addition to the ulnar insertion in some of the lower animals, such as the horse, the ruminants and the rodents, a radial attachment is normal and in a few species, such as the platypus the radial insertion is the only one found.

Lateral portion (figs. 7, 9 and 10, Br.)

This portion is so distinct from the medial portion as to be practically a separate muscle. It is also divided longitudinally into two completely separate bundles.

Origin. The two bundles of this muscle arises alongside of each other, following the lower half of the circumference of the deltoid tubercle.

Insertion. They pass down the arm as parallel fasciculi and are inserted on the lateral border of the radius in line with each other, the most lateral fasciculus being at least a third the distance down the shaft of the radius. This portion of the muscle is supplied by the radial nerve and represents the portion of the muscle developed from the dorsal musculature of the arm and has, in this instance, separated from the rest of the muscle formed from the ventral elements. The radial nerve normally supplies a small portion of the human brachialis muscle on the lateral side, thus indicating the normal composition of the muscle, which always has a small portion of the dorsal musculature included in it. Le Double cites cases where the brachialis muscle has been found divided into two distinct heads, as found in this case, either one of which may be subdivided again. He does not state the nerve supph% but it is probable the primary separation is between the dorsal and ventral elements of the muscle.


This lateral portion forms a sharp fold projectiiis between the humerus and radius and occupies the deeper portion of the skin web previously described as binding the arm in flexion at the elbow. This muscle is very tight and prevents all extension of the radius on the humerus. It is the muscle so placed as to most thoroughly prevent this movement, and the part responsible for this is the lateral portion, due to its insertions down the shaft of the radius. There is no opposition to this force as the triceps is not attached to the radius.

Although this muscle occupies only about half the projecting extent of the skin web here, it is probably the cause of the web, forcing the skin out in a sharp fold ahead of it. The fold has developed beyond the extent of the muscle later on.

The lateral portion of the brachialis is responsible for another displacement of the radius. As its insertion is far down on the shaft of the radius, and its pull is all to the one side, it has swung the radius around laterally until the long axis of this bone lies in a plane parallel instead of perpendicular to the line joining the two epicondyles of the humerus. This latter relation is not at first sight apparent, for the forearm appears to be ventral, not lateral to the upper arm. The reason for this is that the scapula, carrying the humerus with it is rotated through a right angle forward and inward on the flattened chest wall. The scapula has medial and lateral surfaces respectively, instead of ventral and dorsal. The humerus similarly has medial and lateral surfaces instead of ventral and dorsal, and the axis at the lower extremity passing through the epicondyles is not mediolateral in direction, but dorsoventral. The forearm thus lies in a dorsoventral plane although actually rotated laterally through a right angle.

Triceps brachii (figs. 9 and 10, T^, T.„ 1\)

Origin. The long head is very large and arises from part of the axillary border of the scapula as well as the infraglenoid tubercle.

The lateral head arises from the upper third of the posterior surface of the shaft if the humerus above the groove for the


radial nerve, and is quite large. Its border blends with that of the long head throughout its extent.

The medial head lies on the back of the middle third of the humerus, below the groove for the radial nerve. It is overlapped largely by the long head and blends with the deep surface and medial border of the latter.

The lower two-thirds of the muscle exhibit a tendon running lengthwise, at the line of junction of the long and lateral heads. Towards this tendon fibers converge in the upper part muscle, and in the lower part they diverge again to their insertion on the bone.

Insertion. Owing to the absence of the ulna no normal insertion is possible, and the whole lower attachment of this muscle is transferred to the humerus. The insertion is into the whole of the lower third of the posterior surface of the shaft of the humerus and to the back of both epicondyles. The radius receives no attachment whatever from this muscle, so extension of the forearm is an impossibility. This explains the early fixation of the forearm in extreme flexion, allowing thus of the development of the skin web and shortening of the brachialis muscle to make this deformity a fixed one. Migration of the attachment of the brachialis down the shaft of the radius is thus permitted by the permanent flexion of the forearm. In this position the further the muscle passes down the radius the shorter it becomes, as its insertion approaches the level of its origin.

It might be asked why, in absence of the ulna the brachialis muscle becomes attached extensively to the radius but the triceps all ends on the humerus. Why does not the triceps also reach the radius? The difference seems reasonable in view of the following circumstances, comparative anatomy furnishing the answer to the problem. The brachialis is attached to the radius occasionally in man, and as before mentioned, normally in certain lower animals in addition to its ulnar insertion, while in a few species the radial insertion is the only one. In the case of the triceps, insertion on the radius is not normal in the


lower animals even where the ulna is of small importance in the forearm.

It is to be noted that although the two humeral heads of the triceps can produce no movement, as they both arise and insert on the humerus, yet they are both well developed muscle masses.


There has been great disturbance of the muscles in the forearm, due to the absence of the ulna and reduction of the hand, but it is still possible to homologise some of them with those of the normal type. The others however are difficult to define and the homologies given for them are more in the nature of probabilities than of definite facts. The extensors seem to be more reduced and more atypical than the flexors.


Mostly members of the superficial group are here present as all of the deep group with one exception are absent. There are four muscles to consider on this surface.

1. Brachioradialis muscle (figs. 9 and 10, B.)

Origin. High on the lateral epicondylar ridge of the humerus.

Insertion. A very short cylindrical muscle running across the bend of elbow to insert on the shaft of the radius at about its middle point, and just to the side of the insertion of the lateral portion of the brachialis muscle.

This muscle is probably the brachioradialis and its shortening is not extreme, having been noted in other cases, while in one of the anthropoids, the gibbon, its insertion is normally high up on the shaft of the radius.

2. Common superficial extensor mass (figs. 9 and 10, C.E.M.)

Origin. Lower part of lateral epicondylar ridge and outer surface of lateral epicondyle of the humerus.


Inserlioti. Runs directly parallel to radius and inserts at the middle of the shaft of that bone, just medial (owing to pronation apparently lateral) to the brachioradialis.

This muscle probably represents the undifferentiated remainder of the superficial extensor mass, except the extensor carpi ulnaris which is separate. It will thus include the extensors carpi radialis longus and brevis, digitorum communis and digiti quinti proprius. In some reptilia and amphibia these muscles are in a common supinato-extensor mass.

Why none of this mass reaches the carpus or digit cannot be explained, but the fact that none of it does so explains why the hand is carried in a position of permanent flexion, because there is a flexor muscle attached to the digit and it is thus without an opponent to its pull.

3. Supinator (figs. 8 and 10, A.)

Origin. Covered by the common extensor mass it comes from the anterior surface of the lateral condyle of the humerus. This represents the superficial or humeral portion only of the normal human muscle.

Insertion. It courses parallel and deep to the common extensor mass and is inserted into the capsule of the radio-humeral joint, head, neck and upper third of the shaft of the radius, right down to the insertion of the common extensor mass.

This muscle, it seems to me, is quite evidently the supinator, and so is the single representative here of the deep muscles of the extensor series in the forearm.

Extensor carpi ulnaris (figs. 7, 8, 9 and 10, E.C.U.)

Origin. Below the preceding muscle from the lowest part of the lateral epicondyle of the humerus. This is the last of the extensor group and lies in contact with the flexors. It is the longest of the extensors, being over double the length of any of the others.

Insertion. By a long slender tendon which is one-third the length of the muscle, into the middle of the dorsal surface at the


lower extremity of the radius and into the carpus. At the origin of the long tendon from the belly of the muscle there comes off also a very short tendon which courses obliquely toward the flexor surface of the radius and is inserted right alongside of and practically blended with a part of the flexor digitorum profundus, about three-quarters of the distance down the bone.

This muscle is named the extensor carpi ulnaris because of its superficial origin from the humerus and its insertion into the carpus, and because it is the most medial of the extensor muscles here found, and is in contact with the flexors. All the muscles inserting into the carpus also show attachment to the lower end of the radius, this attachment seeming to be due to a spreading out of the tendon at its insertion, and so I do not think the radial attachment here offers a serious obstacle to calling the muscle the extensor carpi ulnaris.


This group of muscles exhibits members of both the superficial and deep layers and although badly disorganized it still retains a somewhat closer homology to the normal divisions of this group than is to be found in the extensors.


First layer 1. Flexor carpi radialis (figs. 7 and 8, F.C.R.)

Origin. By a broad fleshy head from the upper part of the medial epicondyle of the humerus.

Insertion. This muscle is fleshy in the upper half of the forearm and has a long thin tendon coursing through the lower half to be inserted into the lower end of the radius and into the carpus.

The position of this muscle is along the lateral border of the radius on its volar sui'face, although it appears to be dorsal due to the rotation of the bone.

From its attachments and position it can be quite safely identified as the flexor carpi radialis muscle.



2. Flexor carpi ulnaris (figs. 7 and 8, F.C.U.)

Origin. A broad, flat, fleshy origin from the front of the medial epicondyle of the humerus and from the surface of the bone in front of and below this.

Insertion. This muscle is by far the largest of all those yet described in the forearm. It is fleshy to about two-thirds the distance down the radius where it narrows into a heavy tendon" which inserts at the lower end of the radius and into the carpus.

Second layer

S. Flexor digitorum sublimis (superficial portion) (fig. 8, F.D.S.)

Origin. Under the origin of the flexor carpi radialis as a thin flat fleshy muscle which courses obliquely to join one of the deep muscles arising on the radius, which will be described later.

This I would homologise with the flexor digitorum sublimis due to its position as the second layer of muscles from the medial humeral epicondyle. There is a possibility of this muscle being the humeral portion of the pronator radii teres. Against this latter view, are the facts that the muscle is entirely covered by the two carpal flexors, and that it is not inserted into the shaft of the radius but joins a muscle arising here to be inserted into the carpus.


Third layer 4. Flexor digitorum profundus (figs. 7 and 8, F.D.P.)

Origin. A thick fleshy muscle arising from the lower two-thirds of the volar aspect of the radius on its lateral (apparently dorsal) portion.

Insertion. This muscle passes as a compact fleshy bundle as far as the metacarpal region where it condenses into its tendon which is single and runs on the volar aspect of the single digit


to be inserted into the terminal phalanx. In its course it passes under the digital portion of the median nerve which divides on the digit, allowing the tendon to pass out under it in a manner similar to that usually shown by the tendons of the flexor digitorum sublimis muscle.

The flexor poUicis longus muscle is apparently entirely absent or much more probably its muscle mass is indistinguishably fused with that of the flexor digitorum profundus, since the primitive condition of the deep flexors is a single muscle mass giving tendons to the thumb and other digits. Man is one of the very few mammals possessing a flexor pollicis longus muscle and McMurrich ('03) has shown that in the other mammals its absence is not due to a lack of the muscle but to the fact that it has not differentiated out from the common deep flexor mass to the digits. It is thus present as the most radial portion of the flexor digitorum profundus in these forms.

5. Flexor digitorum sublimis (Deep origin) (fig. 8, F.D.S.)

Origin. From middle third of volar aspect of radius just medial (apparently ventral) to flexor digitorum profundus.

Joining the proximal part of this muscle is the superficial origin described above.

Insertion. The common mass so formed passes into a slender tendon inserted at the lower end of the radius and beginning of the carpus.

The reason of the failure of the tendon of this muscle to reach the digit I think must be sought in the failure of the palmar aponeurosis to which it is attached, to differentiate into a tendon. McMurrich ('03) has shown that primitively the sublimis muscle ends at the wrist inserting into the palmar aponeurosis. Muscles developed in this aponeurosis later fuse end to end with the flexor sublimis thus producing its tendons in the mammalia. The palmar structures included in the sublimis have evidently failed to form here, leaving the sublimis to end at the wrist.


6. Flexor Digitorum Profundus (detached portion) (fig. 8, F. D. P.)

Origin. From the neck of the radius and the shaft of the bone near this on the medial (apparently ventral) border.

Insertion. This muscle is long and slender. As it is followed distally into its tendons it divides into a superficial and deep layer which insert separately. The superficial tendon passes down to the lower end of the radius and to the carpus. The short deeper tendon ends almost immediately on the shaft of the radius a short distance above the lower extremity, and is fused with the short deeper tendon of the extensor carpi ulnaris already described.

This muscle I interpret as the ulnar part of the flexor digitorum profundus, which has differentiated during the muscle development of the limb and become attached to the nearest part of the radius. The flexor digitorum sublimis by the extension of its deep, radial origin, comes between it and the radial portion of the profundus layer and so may have prevented their fusion. On the contrary if the lack of fusion was primary this would allow of the sublimis layer becoming attached down the radius between the two parts of the profundus. There is no possibility of this being the flexor pollicis longus as it lies medial and not lateral to the rest of the flexor digitorum profundus.

Fourth layer

7. Pronator quadraius

A thin film of transversely disposed muscle fibers lying over the lower end of the radius represents the pronator quadratus muscle. It is very poorly developed and small in extent.

It is to be noted that by means of the muscle in the forearm voluntary flexion of the digit is possible but voluntary and active extension is impossible, as all extensors fail to reach the finger. A singular and interesting parallel to this case is found in a case cited by Schultze ('04). In a training school he observed a nineteen year old lad who had only one digit on each of all four limbs. Voluntary flexion of these digits was easily accomplished but he


had no power of extension. The probable exi)hination is that there was a condition such as present in the case I have dissected. The fact that in both these cases the flexors are evidently better developed than the extensors is significant and seems to point to certain definite conditions in the muscles being associated with the deformity.

MUSCLES OF THE HAND (figs. 7 and 8, L.)

Only one muscle is present here. It is a lumbrical, arising in the metacarpal region from the lateral side of the flexor digitorum profundus as this latter muscle passes into its tendon. The lumbrical passes in a spiral direction distally and laterally on to the dorsal surface of the digit where it inserts into the dense fibrous tissue over the phalanges.


The whole brachial plexus was dissected out as shown in figure 8 and conformed in all its arrangement and branches to the typical formation. Therefore it is only necessary to describe the course and distribution of its main terminal branches.

From the 'posterior cord

1. Axillary nerve. Normal course and distribution to skin, deltoid and teres minor muscle, and to shoulder joint (figs. 8 and 10, A. N).

2. Radial nerve. Runs ventral to the latissimus dorsi tendon, then winds behind the humerus (figs. 8 and 10, R. N) in the musculospiral groove, here giving branches to the three heads of the triceps muscle, and then enters the space between the triceps and postaxial portion of the brachialis muscle, where it supplies this part of the brachialis and gives off the dorsal antibrachial cutaneous nerve.

A short distance further on the nerve divides into

a. The superficial radial (figs. 9 and 10, S. R. N) which runs a

cutaneous course on the lateral side of the whole length of the

forearm and hand.


b. The deep radial nerve, which Ues under the three superficial extensor muscles (fig. 10) and on the surface of the supinator which is covered in by the others. The nerve supplies all these muscles.

From the lateral cord

S. Musculocutaneous nerve. Supplies the coracobrachialis muscle and penetrates it (fig. 8, Mc. N.) between its upper and middle portions to pass between the biceps and the preaxial portion of the brachialis, supplying both the latter muscles and ending cutaneously in the forearm.

4. Outer head of the median nerve. The median nerve is described under the inner cord.

From the inner cord

5. Inner head of the median nerve. Unites with the lateral head over the axillary artery.

The median nerve (fig. S, M. N . ) courses ventral and medial to the axillary and brachial arteries in the groove medial to the biceps muscle. It enters the forearm deep to the flexor carpi radialis and superficial head of the flexor digitorum sublimis, and in front of the biceps tendon and is accompanied by the medial vena comes of the brachial artery, while the artery and the lateral vein lie under the two biceps tendons. As it passes the elbow, it gives branches to the flexor carpi radialis, flexor digitorum sublimis and flexor carpi ulnaris and then divides into a superficial and a deep branch.

The deep branch evidently is the volar interosseous nerve of the normal arm, and it supplies the three deep muscles arising on the shaft of the radius.

The superficial branch of the median nerve (figs. 7 and 8, M. N.) comes immediately from under cover of the flexor carpi radialis and courses subcutaneously down the ventral surface of the lower two-thirds of the forearm and over the carpus. In the distal third of the forearm it gives ofT a large cutaneous branch on the medial side.


At the carpus a strong cutaneous branch is given off on each side and on the lateral side also a muscular twig to the lumbrical muscle. The rest of the nerve runs on the ventral surface of the single digit, finally forking to each side of the digit about the level of the second phalanx to let the underlying flexor digitorum profundus tendon pass through it. This nerve was at first mistaken for the tendon of the flexor digitorum sublirais muscle, so typical in appearance was it to this latter structure, when only its course in the forearm and hand was uncovered.

6. Ulnar nerve. Runs down the arm under the deep fascia (figs. 7 and 8, U. N .) in company with the basilic vein, pierces the deep fascia a little above the elbow, and divides into two branches, a volar and a dorsal, both running subcutaneously on the medial border of the forearm.

No muscular branches whatever were found on this nerve, its whole distribution being as a sensory nerve to the forearm.

All other nerves of the brachial plexus which are not specially described here are normal in their extent and distribution.


The vessels of the arm were not dissected above the axilla as it did not seem that any noteworthy changes from the normal would be likely to occur. No injection was employed as it was feared that if a vessel wall ruptured structures around the break might be so stained as to obscure valuable results. Small vessels were thus hard to follow, and arteries to the hand could not be identified. *

ARTERIES (figs. 8 and 10)

The axillary artery and all its branches were normal in extent and position.

The brachial artery lay in the groove medial to the biceps muscle, with the median nerve on its medial side throughout its course, so that there is no crossing of nerve and artery.

The brachial artery gave origin to numerous muscular branches and also to three larger branches, the profunda brachii, coursing


with the radial nerve through the musculospiral groove, and the superior and the inferior ulnar collaterals, running medially alongside the ulnar nerve.

At the elbow the brachial artery (fig. 8) took the astonishing course of passing behind both the biceps tendons and lying on the surface of the brachialis muscle. Just beyond this point the arter}^ bifurcated into two branches which passed down the arm, one on each side of the flexor digitorum sublimis. The lateral branch, the radial artery, lay under the flexor carpi radialis muscle, while the medial, the ulnar artery lay under the flexor carpi ulnaris. Both arteries became lost in the dissection before the wrist and hand were reached.


Superficial veins (figs. 7 and 9)

The cephalic vein (C. F.) is present here, starting in the hand and running on the lateral (apparently dorsal) border of the dorsal surface of the forearm, across the skin web at the elbow, ' up the lateral side of the arm, dividing into two channels. These turn ventrally below the insertion of the deltoid, reuniting here, then pass between the deltoid and pectoralis major muscles to terminate deeply in the thoracoacromial vein.

The basilic vein (B. V.) starts also at the wrist, and runs up on the medial border of the dorsal surface, turning medially to the ventral surface just above the medial epicondyle of the humerus. Here it passes under the deep fascia of the arm, running in the groove medial to the biceps as far up as the axilla where it unites with the common trunk formed by the union of the brachial venae comites to form the axillarj^ vein.

Across the back of the elbow a large vein connects the basilic and cephalic veins transversely.

The median vein (M. V.) courses up the middle of the ventral surface of the forearm as far as the bend of the elbow where it divides into two large branches, the median basilic and median cephalic.


The median cephalic {M. C. V.) runs vertically upward on the ventral surface of the postaxial part of the brachialis, receiving as it goes the deep cubital vein from the cubital fossa. The median cephalic joins the lower half of the cephalic and the common trunk joins the upper half of the cephalic.

The median basilic runs (M. B. V.) back over the medial epicondyle of the humerus then turns up to join the basilic. It is double in most of its course.

Deep veins {fig. 8)

The radia and ulnar veins coursing alongside the corresponding arteries unite to form the vena comes lying medial to the brachial artery, and passing behind the biceps tendons.

Another vein runs back alongside the median nerve in front of the biceps tendons and half way from the elbow to the axilla the brachial vein leaves the side of the artery, crosses in front of the median nerve, and unites with the vein accompanying the nerve. This common trunk ascends to the axilla and unites with the basilic to form the axillary.

The axillary vein lies medial and deep to the ulnar nerve and medial cord of the brachial plexus and receives the usual normal tributaries.


The first questions that naturally arise in connection with this case are as to the causative agent and time of production of the monstrous condition here exhibited. There are several different possibilities to be considered and as the time and the cause are closely related they will be taken up together.

This deformity may be hereditary and so transmitted in the germ cells. In the case referred to previously, which was described by Schultze ('04), there was only one digit on each hand and foot and this same identical condition was found in the mother and the mother's father, while a brother had monodactylous hands, and other deformities of the feet. It is a well known fact that monstrosities affecting the limbs show more


tendency to be hereditary than many other kinds. Adami ('08) gives certain good examples of hereditary transmission of such deformities. There is, however, in the case studied here no evidence that heredity plays any part in the production of the abnormality and the cause must be sought for elsewhere.

Again it is possible for a monstrosity to be produced by deficiency in either germ cell, which will produce a deficient fertilized ovum. A normal fertilized ovum may also be injured and Conklin ('05) has shown that even in the ovum there is. a differentiation and specific localization of organ forming substances, one of which could be damaged thus leading to the production' of abnormal embryos and monstrosities. This has been done by many workers, only one or two of whom, such as Werber ('15) and Stockard ('09-10) need be mentioned. In this case, however, damage to either of the germ cells and also to the fertilized ovum is improbable as there is no history of either of the parents suffering from venereal disease, alcoholism or drug habits and neither of them work in noxious surroundings where poisoning would be possible with lead, arsenic, phosphorus or other agents.

The period of the production of this deformity is thus excluded from the germinal stage and must be either in the embryonic or foetal stages. The foetal stage also can be excluded, for as pointed out by Ballantyne ('04) in his excellent book on antenatal pathology, foetal physiology is, if not identical, at least similar and parallel to that of the individual after birth, and thus, foetal pathology is mainly concerned with disease and disordered metabohsm. On the other hand the embryonic period is a period whose physiology is not that of functional activity of organs, but of organ formation and differentiation. Pathological conditions in the embryonic period, therefore, lead to malformations and so if severe to the production of monsters. The deformity in this case is thus limited in its production to a period between the first and seventh weeks of intra uterine life. During this period the limb buds appear and bones and muscles differentiate in them.


Schwalbe ('Oo) has pointed out that there is a definite termination period for the production of any deformity. Before the end of this period practically all deformities of that particular type must appear, and any produced later than this are to be regarded in the light of accidental occurrences injuring originally perfect parts and so simulating abnormalities produced as errors of development before this termination period. The termination period in each case marks that special time in which organogenesis ceases and functional activity begins in any particular organ or part and marks the limit in time beyond which a given deformity rarely if ever has its origin. This reckoning also places the latest period for the production of the limb deformity in this case at the seventh or eighth week, when the limb is fully differentiated and ossification in the limb skeleton begins.

Mall ('08) after a critical study of one hundred and sixty-three pathological embryos, has concluded that most monsters are produced by the faulty development of normal ova due to external influences, usually a vice of nutrition due to faulty implantation which in turn is generally due to an abnormal condition of the uterine mucosa. Such a condition for instance would be a mild, chronic endometritis which would not prevent the occurrence of a pregnancy but would be enough to catise faulty development. This might well be the cause here, as there is in this case a history of two miscarriages previous to the birth of this monster, without any apparent toxic agent or disease leading to their production, thus giving presumptive evidence of an abnormal condition of the uterus, which would cause faulty implantation and eventual death and expulsion of the products of conception.

Mall has estimated from statistics from various sources that in 100,000 pregnancies there are 80,572 normal births, 11,765 abortions of normal embryos, 7048 abortions of abnormal embryos and early monsters, and 615 monsters born at term. In view of the great prevalence of uterine disorders, superadded to the unsuitable conditions in which many pregnancies occur, the pathological development of approximately 7.5 per cent does not appear unduly high. It will be noted that one monster is


born at term in approximately every one hundred and thirty births.

For a full discussion of the many teratological theories the reader is referred to Ballantyne's text book on antenatal pathology. It is sufficient to mention briefly any other likely causes of the present deformity. Maternal impressions still possess many firm believers, but I think as a cause their utter powerlessness in this case is clearly demonstrated. The impressions were received later in pregnancy, the deformity, as shown above, must ha^'e been established very early, so the relation of the two as cause and effect was absolutely impossible. (See page 387).

Foetal diseases do not appear as a rational cause of this deformed condition and neither do amniotic diseases. Amniotic bands and adhesions have been ascribed almost universal teratological influences by devotees of this theory, and when they could not be demonstrated, their previous existence and later disappearance has been postulated. There is no cicatrix or other evidence of any band connected to the extremities here, and the symmetry of the deformity argues against its production thus. The accompanying defects in the vertebral column are evidently not due to such bands.

There'is one cause in the production of monstrosities and of pathological embryos that it seems to me is perhaps a fruitful one and which I have not found mentioned by other authors. I refer to attempts in the production of criminal abortion, which as every physician knows, are so prevalent amongst the women of this age. These attempts are not always immediately successful but sometimes the pregnancy is terminated by the death of the injured child at some later date and in some cases pregnancy goes* on to full term in spite of the injury. Is it not extremely possible that in these instances where the child continues to live for some time after the attempt to destroy it, that it should exhibit some monstrous condition, especially when the attempt is made in the first two months? Both the use of mechanical means and of drugs would result in these pathological conditions, the instrument by direct injury to the child or to the amnion, the drugs by affecting the implantation in the


uterus, and so being one cause of the condition to which Mall ascribes most pathological embryos. To show that attempts at abortion form a cause not to be neglected in this regard I quote from the Secretary of the Indiana State Board of Health, Dr. J. N. Hurty ('17) who says "It has been estimated that about one-third of pregnancies end in induced abortions, that at least 200,000 volitional abortions occur every year in the United States and that not less than 12,000 women die annually from the direct effects thereof." (This is quoted from another article as I regret I have been unable to obtain the journal with Dr. Hurty's original article in it.) Surely the arguments I have used above are sound in view of such conditions as Hurty states to exist and attempted abortions which are not immediately successful ought to be ranked amongst the causes of pathological embryos and monstrosities.

Some of the abnormal conditions found in this foetus can be correlated with interesting embryological stages of growth which it seems to me throw considerable light on what are otherwise obscure isolated facts. Statements as to normal skeletal and muscular development are taken from the accounts by Bardeen and Lew^is in Keibel and- Mall's Human Embryology. ('10).

In the early development of the vertebra, as the scleroblastema becomes chondrified, this process in the bodies of the vertebrae is brought about by two centers, one on each side of the notochord. At first there is no fusion of these two centers of chondrification dorsally or ventrally around the notochord, as there is present in the mid line a membranous perichordal septum (Keibel and Mall). Normally this septum is soon broken through both dorsally and ventrally and the notochord is completely surrounded by cartilage by about the fifth or sixth w^eek.

Ossification then occurs from a center which is usually single, but may divide or even arise paired.

The early presence of the perichordal septum appears significant in view of the fact that in this foetus are found two vertebrae with divided bodies, each half growing independently, and one-half growing less rapidly than normal. This septum was present at the period of embryonic life when that vice of develop


ment occurred which produced the monstrosity of the Umbs, Is it not very probable that the chondrification process in these two abnormal vertebrae was hindered so that the perichordal septum was not broken down, but remained intact, thus producing a vertebra with a divided body?

Ossification as mentioned above tends to occur in the body from one center, which may be divided. Under such conditions, with the perichordal septum intact it is possible that niore of the ossifying center should be in one half than the other, thus accounting for the unequal rate of growth in the two separated halves.

There are some other points of interest in the vertebral column. The lateral masses of the sacral vertebrae ossify as follows: the first at the fifth month of intrauterine Ufe, the second at the sixth month, the third at the seventh month, the fourth and fifth after birth about three months. In this foetus, the age was given as seven months and the third lateral mass center is just appearing, thus showing a normal rate of growth.

The first coccygeal vertebra in this foetus has a center of ossification in its body, while normally it appears in the first year after birth, so in this region there is an actual acceleration of ossification, in direct opposition to the retardation or suppression shown in the abnormal portions of the skeleton.

The core of the limbs at the third week is. filled with vascular mesenchyme which at the fourth week becomes a scleroblastemal condensation which then becomes successively chondrified and ossified. The primary failure of the digits and ulna of this foetus can thus be placed as far back at least as the fourth or fifth week of development, at the time when the differentiation of the skeletal parts should have occurred. This would correspond with the time of production of the defect in the abnormal vertebrae. These facts would seem to indicate that at this particular period was exerted the strongest and most active influence of the agent producing the deformities.

• Absence of the ulna is a much rarer condition in the forearm than absence of the radius. Kiimmel ('95) has collected a series of cases of defect in the bones of the forearm. Unfortunately


I could not secure the journal containing his original article but Ballantyne ('04) in his text l^ook and Schenk ('07) in an article on a case of defect of the ulna agree in their accounts of Kiimmel's cases which can be taken as correct. He found SO instances of defect in the bones of the forearm of which 67 were of the radius, 13 of the ulna. In the case of the ulna it was defective in 5, totally absent in 8 instances. In some of these cases there was associated absence of the ulnar side of the carpus and one or more fingers on the ulnar side of the hand.

The muscles of the limb definitely appear first proximally and differentiation proceeds distally. It might be expected that the muscles of the shoulder girdle and upper arm, being the first to appear after the skeletal deformities were produced, might show some anomalies. They do exhibit anomalies, but peculiarly not anomalies of defect, but of excess, such as supernumary heads and increased insertions. Of course, in the forearm and hand grave defects are associated with the loss of the skeletal structures.

The question naturally arises as to whether the muscle anomalies are a consequence of the skeletal defects or were independently produced by the same vice of development or nutrition to \\^hich the absence of the bones is due. In this connection it is to be noted that the suppression of muscles in the forearm is not confined to the ulnar border of the arm but affects also the radial side, so that more than mere absence of the skeleton underlies the anomalies. This can be proved by the fact that muscle is independent and self-differentiating. Muscles develop independently of functional activity as shown here by the two humeral heads of the triceps, inserted also on the humerus, incapable of movement, yet well developed. Harrison ('04) also proved that muscles develop independently of the nervous system, for he removed the spinal cord in early frog embryos, before the muscles had differentiated or received any nervous connection and yet the normal process of muscle development and grouping occurred. This power of self-differentiation goes right back to the ovum where Conklin ('05) has


demonstrated the presence of a myoplasm or muscle forming substance.

In the forearm the extensor and supinator group differentiate before the flexor and pronator set. As the muscle formation follows closely upon the definition of the skeleton, if the growth suppressing influence which acted on the skeleton lasted long enough to influence the muscles it is to be expected that the extensor group would exhibit the greatest amount of damage. Such is actually the case. Only four extensor muscles are present as against seven flexors and pronators plus one palmar muscle. Only one extensor muscle reaches as far as the lower end of the radius, nearly all the flexors reach that level. No extensor tendon reaches the digit, a flexor tendon passes right out to the terminal phalanx, in addition to bearing a lumbrical muscle to the digit. It is to be noted that in the members of the extensor group here present the muscle masses are of about normal proportion, covering half of the length of the radius but in only one case is a long tendon developed, the other muscles inserting at once on the middle of the shaft of the radius. This failure of the long tendons to differentiate out after the appearance of these muscles is a further example of the greater suppression of growth in this region. Grafenberg ('11) describes the musculature in a case of absence of the radius and the thumb. Here the radial musculature is present as a common mass high up in the forearm, possessing no tendons, and so appearing very much like the extensor muscles I have described. The other muscles both flexors and extensors, in Griifenberg's case are present and normal in extent.

Regarding the muscle that I have called the common superficial extensor mass, as separation into separate portions begins at the carpus after the appearance of the tendons, it is not possible here to have such a division into its component muscles, because its tendon is entirely absent.

Absence of the thumb is not enough to cause disappearance of the abductor pollicis longus and extensor pollicis brevis, the radial members of the deep extensors, for there is still opportunity for the muscles to develop over the radius. The triceps


did not fail when the uhia disappeared. The same is true of the ulnar members of this group, the extensor pollicis longus, and extensor indicis proprius. All this group have been obliterated by a specific suppressing agent during myogenesis.

In the flexor muscles it seems strange that the pronator teres is not present when so many of the other muscles are. Its complete absence has never been noted as an anomaly although its coronoid head has often been lacking. In lower vertebrates this muscle is a part of a common muscular layer known as the pronatoflexor mass. In this foetus it may be present in the superficial layer, included with the mass of the flexor carpi radialis, having failed to obtain an insertion at the usual level on the radius.

It is interesting to note that in this foetus a definite tendency in one direction is showai by all muscles, which are properly developed and which show anomalies. This tendency, for instance is shown by all the muscles on the front of the upper arm and is a regression or atavistic change, the anomalies resembling normal muscles of the low^er animals. Changes due wholly to loss of normal skeletal parts lead to anomalous attachments which of course cannot be properly included in this class as they are in the nature of monstrosities.

The question naturally arises as to what single digit it is that has persisted in this hand, and also what carpal bones are present.

It may be taken as a plausible working hypothesis that with loss of the ulna would be associated loss of the ulnar side of the carpus, with the fourth and fifth digits.

This hypothesis is supported by the fact that the main cutaneous digital nerves ventrally are two strong branches from the median while dorsally the radial reaches the base of the digit. The ulnar nerve has no digital distribution, and as it normally goes to the fourth and fifth digits while the median and radial supply the other three, the digit here present certainly ought to be one of the three on the radial side of the hand.

This would leave three digits still to decide between. This number can be further reduced to two as the thumb is certainly absent, for the persistent digit has a metacarpal and three pha THE AMERICAN JOURNAL OF ANATOMY, VOL. 22, NO. 3


langes, and a lumbrical muscle is also found attached to it. The median nerve normally supplies the lumbrical muscle to the second and third digits, the ulnar those to the fourth and fifth. The single lumbrical here present is supplied by the median, a further proof that the digit is the second or third.

The digit is thus either the index or middle finger, but to decide upon which of these two it is, is much more difficult as there is nothing in the disposition of the muscles to help solve the problem. The distribution of the cutaneous branches of the median nerve seem to offer the only key to the solution. In text figures C and D is given side by side the cutaneous distribution of the median nerve in this foetus and in the normal hand. As the cutaneous distribution of the median is wholly digital it is assumed that branches found from the trunk of the median running into the hand were intended for those digits which did not appear. By checking these off against the branches in the normal hand it is found that the digit here present ought to be the index finger.

There is a palmar cutaneous branch from the median arising in the lower half of the forearm and ending in the palm. It is not to be mistaken here for one of the digital nerves, these latter arising in the palm. There are three such nerves, only the middle one passing out on the digit, where it forks to supply each border, while the flexor profundus tendon passes on under it. The other two nerves end at the base of the digit on its medial and lateral borders. To save a long description the reader is referred to the figure explaining the distribution of these nerves. Here at a glance it can be seen that the part of the nerve found on the digit in this foetus, is the portion to the index finger from the first and second common volar digital branches. From this distribution it seems fairly definite that the sole remaining digit on this hand is the index finger.

On the arm which was not dissected it will be remembered that in addition to the single finger carried at the end of the limb there was a well developed digit found on the medial side of the elbow. Radiographs showed this to contain a metacarpal and three phalanges. Of course, this digit can be logically as



sumed to be one of the ulnar members which has differentiated in spite of the total suppression of the ulna and part of the carpus. Its appearance at the elbow and not the carpal region lends color to the view that the ulnar anlage of the limb skeleton never appeared at all even in the early mesenchyme, so that

Flexor digitoru sublimiS

Tendon of Flexor digitoru prof und vs

Text fig. C Outline of the cutaneous distribution of the median nerve in the normal human hand.

. Text fig. D Outline of the cutaneous distribution of the median nerve in the left hand of this monodactylous foetus.

I^The part of the nerve shown in solid black in the two figures, is reckoned as identical in the two^^hands, and is used to determine what single digit is present in the foetus.

the primary reason for nonappearance of the ulna was not a lack of chondrification and ossification.

There is another view in regard to this digit, and that is that the digit is really the representative of all five normal ones, being the result of development of the original undivided digital


anlage in the earliest stage of the Hmb skeleton as the distal end of the condensed scleroblastemal core.

In view of the facts already expounded it seems to me that this latter view is not likely to be correct. The ulnar nerve ought to have a digital cutaneous distribution if the ulnar fingers of the hand are represented in this common finger, but the ulnar does not pass out on the digit, thus supplying one argument against this hypothesis.

The presence of one digit at the elbow joint on the right arm postulates the separation of one digital rudiment from the common mass. If it separated then clearly the tendency to division of the skeleton of the hand into rays was present and it is just as tenable to suppose that the five-rayed condition of the hand was provided for, but growth suppressed in four, as it is to suppose all five rays of one hand and four in the other to be included in a common mass.

The fingers here present, both in the hand and at the elbow, as will be seen from the table of measurements, are normal in size for a single digit. The development of an undivided common digital mass might be expected to produce a condition of macrodactyly, which is not found here. Considering all the facts, the view that the digit as found on the hand here represents only one of the five of the normal hand seems to be the correct view in this case.

What carpals are present is not capable of definite answer. There are only two present, a proximal one articulating with the radius and bearing beyond it a distal one which carries the digit. These two in their shape as previously described resemble the navicular and lesser multangular more than any of the other carpals. Their absolute identification, however, as these two, is hardly to be warranted from these facts alone. If it be true that these are the two carpals present it adds another proof for the digit being the index finger as these two particular carpals are in the direct line of the radius and the second digit.

In the mechanism of the production of the deformity in the limb several different conditions have to be considered. First, in the early limb bud the ulnar segments may not have been


carried out in the distal part of the evagination from the trunk of the body, being drawn out later only in the proximal part of the limb, so that a complete upper arm is formed but only the radial half of the rest of the limb. Secondly, these segments may have been drawn out, the limb bud being normal, but further differentiation not occurring, so that what is seen in the limb represents a fused radius and ulna in the forearm, fused carpals and digits in the hand. The arguments against the digit really representing all five have already been reviewed, and against the view of the ulna being included in the forearm is the absolutely typical shape and size of the radius, the distribution of nerves and muscles, and the appearance on the right arm of a digit at the elbow, as if this point represented the distal end of the ulnar portion of the arm. Thirdly, the limb bud again may have been normal, without fusion of the radial and ulnar anlagen in the skeleton, only the radial half going on with its development, the ulnar half failing entirely, except for the digit at the right elbow. The presence of this digit lends color to this third view.


After the rest of this paper was written, out of curiosity aroused by the flatness of the abdomen, I opened the body cavity to examine the viscera, and was surprised to discover a diaphragmatic hernia with a large proportion of the abdominal viscera situated in the left pleural cavity. The right half of the diaphragm was intact and perfect, but the left half was almost entirely absent. The sternal and vertebral regions were present and joined in the central tendon, forming a free edge to the diaphragm in the midsagittal plane. The left costal origin was indicated in front by a muscular ridge 2 to 3 mm. high following the costal margin as far back as the axillary line and the whole of the left half of the diaphragm except this narrow peripheral band was absent, leaving a wide open communication between the pleural and peritoneal cavities. The left mediastinal pleura passed over the medial free edge of the opening to become diaphragmatic peritoneum under the right half of the diaphragm,


the costal pleura passed on down as parietal peritoneum on the abdominal wall.

The hernia is thus of the variety known as hernia diaphragmatica spuria. Cases of hernia diaphragmatica vera have a hernial sac formed of diaphragmatic peritoneum and pleura invaginated into the pleural sac, so that the abdominal viscera are not in reality in the pleural sac. In this case however, there is no hernial sac, but a complete hole through the diaphragm and its coverings. The genesis of this condition I would interpret as a persistence of the embryonic pleuroperitoneal passage, the original communication between the pleural and peritoneal cavities, which has not been shut off, due to the failure of the septum transversum to grow back on this side. The left side normally closes a little later than the right (Keibel and Mall, '10) and this may be one factor in the greater prevalence of hernias on the left side.

This defect in the diaphragm must have had its origin during the development of the structure, and so occurred between the fourth and eighth weeks of intrauterine life, probably, on account of its size, in the first half of this period, say the fifth week, which synchronises exactly with the production of the defects in the limbs and vertebral column.

The heart has been pushed over entirely to the right side by the other viscera, but apart from its position is quite normal. The left lung shows two lobes, but is extremely small and flattened against the mediastinal wall just above the heart. The abdominal viscera are all fairly normal in relation to each other and seem to have been rotated en masse up and over toward the right. The left lobe of the liver is thus vertical, and against the mediastinal wall. The oesophagus comes from behind the upper end of the heart into the stomach and the latter is vertical, the pylorus being in the abdomen. The duodenum lies over the vertebral column and the small intestine runs from it into the pleural cavity, successive coils being piled continuously above the previous loops up to the apex of the cavity, where the gut is reflected down medially. Opposite the lung occurs the junction with the caecum and appendix. The colon descends


as far as the duodenum, then turns suddenly back on itself and ascends in the great omentum against the stomach to its upper end, then turns sharply down on the body wall, loses its mesentery and runs on the wall to the brim of the pelvis, where it turns suddenly into a large loop extending up again as high as the liver before turning to come down into the rectum.

Diaphragmatic hernia seems to be a fairly common condition as Ballantyne ('04) collected one hundred cases in the literature from 1888 to 1900. It is a peculiar coincidence, that in one of those cases, just as in this present one, there was also absence of the ulna. This is all the more interesting because Ballantyne states that associated malformations occur less frequently in conjunction with ulnar defects than with defects of other bones in the limbs.

In bringing this study to a close I wish to very cordially thank Prof. J. Playfair McMurrich for providing the material for the work and also for his valuable, kindly criticism of this paper during its preparation.

May 1st, 1917.



Ad AMI, J. G. 1908 The Principles of Pathology, vol. 1. Lea and P^ebiger,

Philadelphia and New York, 1908. Ballantyne, J. W. 1904 Manual of Antenatal Pathology and Hygiene, vol.

I. The Foetus; vol. 2. The Embryo. Green, Edinburgh, 1904. Barabo 1900 Eine Missbildung. Offic. Protok. der Niirnberger Med. Ges.

und Polikl. Munch. Med. Wochensch, 1900, S. 713. CoNKLiN, E. G. 1905 Organ-forming substances in the eggs of Ascidians.

Biological Bulletin, vol. 8, p. 205. Grafenberg, E. 1911 Die Muskulatur in Extremitatenmissbildungen Anat.

Hefte, Bd. 42, S. 195-250. Harrison, R. G. 1904 An experimental study of the relation of the nervous

system to the developing musculature in the embryo of the pig. Am.

Jour. Anat., vol. 3. Hurty, J. N. 1917 Indianapolis Medical Journal, January. Keibel, F., and Mall, F. P. 1910 Manual of Human Embryology, vols. I and

II. Lippincott, Philadelphia and London, 1910.

KtJMMEL, E. 1895 Die Missbildungen der Extremitaten durch Defect, Ver wachsung und Uberzahl. Biblioth. Med., Abt. E., Heft 3. Le Double, A. F. 1897 Traite des Variations du Systeme Musculaire de

I'Homme. Schleicher, Paris 1897. Mall, F. S. 1908 A study of the causes underlying the origin of human monsters. Jour. Morph., vol. 19. McMuRRiCH, J. P. 1903 The phylogeny of the forearm flexors. Am. Jour.

Anat., vol. 2.

1915 The development of the Human Body, Fifth Edition. Blakiston,

Philadelphia. ScHENK, E. 1907 Ueber zwei Falle typischer Extremitaten Missbildungen

(Ulnadefekt, Fibuladefekt). Frankfurter Zeitsch. fur Pathol., Bd. 1,

H. 3 und 4, S. 544-62. Schultze, E. 1904 Familiare Symmetrische Monodactylie. Neurol. Central blatt, S. 704. Schwalbe, E. 1906 Ueber Extremitatenmissbildungen (Spalthand, Spaltfuss,

Syndaktylie, Adactylie, Polydactylie.) Munch. Med. Wochensch.,

S. 493.

1906 Die Morphology der Missbildungen des Menschen und der

Tiere, T. 1 und 2. Fischer, Jena, 1906-7. Stockard, C. R. 1909 The development of artificially produced Cyclopean

fish — "The Magnesium Embryo." Jour. Exp. Zool., vol. 0.

1910 The influence of alcohol and other anaesthetics on embryonic

development. Am. Jour. Anat., vol. 10. Werber, E. I. 1915 Is defective and monstrous development due to parental

metabolic toxaemia. Abstract in Anat. Rec, vol. 9, p. 133.

1915 Experimental studies aiming at the control of defective and

monstrous development. A survey of recorded monstrosities with

special attention to the ophthalmic defects. Anat. Rec, vol. 9, p.

529. Standard recent textbooks of Human Anatomy.




A, supinator muscle A. A., axillary artery Ac, acromion process A.N .y axillary nerve

A.T.N. , lateral and medial anterior

thoracic nerves ^.T'., axillary vein

B, brajcjiioradialis muscle B.A., brachial arterj' Bi, biceps muscle

Br, brachialis muscle B.V., basilic vein

C, coracobrachialis muscle

C.E.M., common superficial extensor

muscle mass Ch, chondroepitrochlearis muscle CI, clavicle Cu.V., cubital vein C.V., cephalic vein

D, deltoid muscle

Dx, cut edge of deltoid muscle

E, lateral epicondyle of humerus E.C.U., extensor carpi ulnaris muscle F.C.R., flexor carpi radialis muscle F.C.U., flexor carpi ulnaris muscle F.D.P., flexor digitorum profundus

muscle F.D.S., flexor digitorum sublim'is

muscle H, head of humerus H.R., head of radius I., medial epicondyle of humerus I.B.N., intercostobrachial nerve Inf., infraspinatus muscle L., lumbrical muscle L.C., lateral cord of brachial plexus L.D., latissimus dorsi muscle L.S., levator scapulae muscle L.T.N. , lateral thoracic nerve

M. A.C.N. , medial antibrachial cutaneous nerve M.B.C.N., medial brachial cutaneous

nerve M.B.V., median basilic vein M.C., medial cord of brachial plexus Mc.N., musculocutaneous nerve M.C.V., median cephalic vein M.N., median nerve M.V., median vein P. A., profunda brachii artery P.C., posterior cord of brachial plexus P.C.A., posterior humeral circumflex

artery P. Ma., pectoralis major muscle P. Mi., pectoralis minor mufecle R., rib

Rh, rhomboid muscles R.N., radial nerve S.A., serratus anterior muscle S.C.M., sternocleidomastoid muscle S.N., suprascapular nerve Sp., spine of scapula S.P.I., serratus posterior inferior

muscle Spl, splenius cervicis et capitis muscle S.R.N., superficial radial nerve Sup, supraspinatus muscle Ti, long head of triceps muscle To, lateral head of triceps muscle Ti, medial head of triceps muscle T.Ma, teres major muscle T.Mi, teres minor muscle Tr, trapezius muscle Trx, cut edge of trapezius muscle U.N., ulnar nerve

X, depression in back over defective vertebrae



1 Deformed foetus seen from in front.

2 Deformed foetus seen from left side.

3 Left arm, viewed laterally, showing monodactyly and webbed elbow.

4 Right arm, viewed ventromedially, showing monodactylous hand and extra

digit located at elbow.





fjj0W ::




5 Radiograph of foetus from right side. Thirteen thoracic vertebrae and ribs are shown.





Cervi Vert

TKoracvc VerTeWaXlii':


LuTTibar Vertebv

Sacra Vertebra I

Sacral Vertebra V,

Ver lebta l' ^

Proximal Phalanges Distal Phalanges





6 Radiograph of foetus from ventral surface to show thirteen thoracic vertebrae and ribs. Two defective vertebrae are seen in the thoracic region.






Level of

Csf vica I

Vet-tebra I .

Thoracic. "^"VerCebrd 1.

Leva] oi-_j




7 Superficial dissection of the ventral surface of the left arm.

8 Deep dissection of the ventral surface of the left arm.

9 Superficial dissection of the dorsal surface of the left arm. 10 Deep dissection of the dorsal surface of the left arm.








adthor's abstract of this paper issued by the bibliographic service october 6.



From the Anatomical Laboratory of the Northwestern University Medical School^



A. Introduction 439

B. Historical 440

1 . Results from drawn blood 440

2. Results from circulating blood 447

3. Results from fixed material 449

C. Observations 451

1. Experimentation with drawn blood 451

2. Examination of circulating blood 460

3. Action of fixatives 463

D. Discussion 465

E. Summary '. 470

F. Bibliography 471


It has been a classical teaching that the normal shape of the mammahan red blood corpuscle is that of a biconcave disc. Within the last decade and a half, however, a few workers have vigorously assailed this view and have asserted that intravitally the erythroplastid is concavo-convex, i.e., has the form of a cup or bell, and that the biconcave disc first appears after blood is drawn from the vessels. According to this latter view, the cup is the normal form, the disc the derived one.

That under certain conditions cup-shaped corpuscles can actually be found in ordinary preparations of drawn blood, in fixed tissues, and even in circulating blood no one will deny;

1 Contribution No. 49; May 25, 1917.



such forms have been seen and described since the days of the pioneer microscopists. The issue, therefore, hinges entirely on the determination of what is the normal intravital condition, and what the modification or artefact.

Wagner ('33, p. 4) was the first to appreciate and definitely formulate this, our present contention: "Ob die menschlichen Blutkornchen auf beiden Flachen platt oder konvex oder gar konkav sind, oder konvex-konkav, wie wohl behauptet worden ist liisst sich schwer ausmitteln ....

Evidence as to the shape of the erythroplastid has been derived from three sources: (1) drawn blood; (2) circulating blood; (3) fixed tissues or smears. The results obtained previously in each of these fields will first be considered separately.


1. Results from drawn hlood

The desultory microscopical observations of Leeuwenhoek (1719) included an examination of mammalian blood, he apparently being the first to observe this tissue attentively. Blood drawn from the finger was mixed with an aqueous decoction of pareira brava, the resulting dilution facilitating its study. Translated, his description (epistola 44, p. 422) is as follows: " . . . . most of the corpuscles have- a certain concavity or sinus receding into them, as if we have a vesicle full of water and by pressure of the finger should indent the middle of the vesicle as a pit or depression."

Muys (1738), Fontana (1787), and Dujardin ('42) essentially substantiated Leeuwenhoek's conclusion. It is evident, however, that the uncontrolled type of observation recorded by Leeuwenhoek retains an historical interest only, for the action of water in altering the shape of these corpuscles is a matter of common knowledge, dating back to the time of Muys (1751).

Schultze ('65) was the first to record the occurrence of some spherical red corpuscles in drawn blood. Later, in 1877, Litten examined the blood of severely anemic individuals and found cup-like corpuscles which he described as 'pessary' forms; these


he believed existed also .in limited numbers in normal blood. Quincke ('77) and Grawitz ('99) saw cup shapes among the poikilocytosed corpuscles of severe anemias; Grawitz ('02) further recorded having observed a tendency toward crenation in fever patients.

At about this time Ranvier ('75) demonstrated that increased temperature produced cups or spheres according to the degree of elevation. This heating effect has been emphasized by v. Ebner ('02), Fuchs ('03), Albrecht ('04), and Zoth (cited by Lohner '10). Here might be mentioned the probably erroneous contention of Hamburger ('02) that a correlation exists between the oxygen-carbon dioxide content of the blood and the shape of the corpuscles; in blood rich in oxygen (carotid artery) discs were described, whereas the corpuscles of blood having a high carbon dioxide content (jugular vein) were believed to be cupshaped.

During the two hundred years in which the foregoing data as to the existence of cups were being collected, observations of another sort were recorded.

de Senac (1749) referred to red corpuscles as having a lenticular shape, plus approchantes d'une sphere appatie au veritablement lenticulaires."

The corpuscles were studied with considerable care by Hewson (1777) who says (p. 214) : These particles of the blood, improperly called globules, are in reality flat bodies . . . ." He diluted the blood of animals with blood serum^ and records his observations (p. 215) : " . . . . these particles of blood were as flat as a guinea."

J. Miiller ('32) stated that the corpuscles in side view resemble coins; Schultz ('36), Prevost and Dumas ('21) and other contemporaneous writers made essentially the same comparison.

In 1838 Wagner decided in favor of the normality of the biconcave disc (cf. p. 440). Henle ('41), the atlases of Funke ('53) and Ecker ('51-'59), and many more recent works figure the familiar biconcave shape.

2 Although Muys (1751) mentioned the difference between the action of serum and water, it was not until 1813 that Young proved water^not actually to dissolve the corpuscles.


The effect of water, first noted by Muys (1751), was studied by Malassez ('90) who showed that shapes intermediate between the disc and sphere are obtainable in examining media of different concentrations and that hypertonic solutions induce crenation. Crenation does not necessitate permanent injury, for a return to weaker solutions allows recovery (Heinz '90; Weidenreich '02).

Previous to the year 1902 observations on the shape of the red blood corpuscle for the most part had been of a casual nature. That the cup might be the true normal form was not considered seriously. Standard texts and atlases continued to describe the classical disc, although in a few cases (e.g., the atlas of Brass, '97) cups w^ere also figured.

The renewal of interest in the cup form resulted from a series of detailed investigations by Franz Weidenreich who vigorously assailed the common teaching and contended for the normality of the cup, supporting his contention by exhaustive experimentation and by ingenious argument. As might be expected these conclusions did not pass unchallenged; heated controversies followed in which arguments and counterarguments, rebuttals and rejoiners held sway; technical methods were attacked; interpretations of results were impugned. A few converts^ to the new school were made, but the majority of workers were unconvinced and the anatomical world at large, to say the least, has remained highly skeptical.

In order to bring out the various disputed points it will be necessary to present in some detail the main features of the contributions of this period.

Incited by observations of Schwalbe on the porcupine and on man, Weidenreich began an investigation, the first report of which comprised his notable contribution of 1902. When a moist chamber was used in examining fresh human blood (on a warm-stage at 37.5°C.) he observed first a rapid streaming; as the movements decreased rouleaux formed, but isolated corpuscles appeared as 'bells' (p. 464) : .... ist die wahre

^ Weidenrich ('10) listed the following as having accepted the normality of the cup: Fuchs, Lewis, Radasch, Bonnet, Minot, Schleip, Schridde, and Stohr.


gestalt der roten Bliitkurperchen die einer Glocke mit ziemlich dicken Wandungen . . . ." This shape he compared to a cup (Becker), bowl (Napf), medusa, or gastrula. Identical results were reported from a series of mammals including a macacus monkey.

Weidenreich next asked (p. 468): Wie kommt es aber nun, dass man sich bisher liber die wahre Form so tauschen konnte?" Examining blood in 0.9 per cent sodium chloride solution he found discs exclusively; in 0.6 per cent solution cups were seen almost without exception. He concluded that saline solutions exert a pure osmotic effect on the blood corpuscles; in higher concentrations the corpuscles give up water thereby becoming discoidal or crenated, in lower concentrations they imbibe water and change to the cup or spheroidal condition. Since typical cups are found in 0.65 per cent saline solution he considers this to be isotonic with the blood of man and mammals.

Weidenreich therefore believed he was in a position to account for the alleged popular misconception concerning the true shape of these elements (p. 469) :

Nun wird aber verstandlich, wie die Tauschung iiber die wahre Form moglich ist. Operirt man nicht sehr rasch bei der Untersuchimg des unverdiinnten Bliites und benutzt man namentlich kaltere Objekttrager und Deckglaser, so geniigt die erhobte Verdimstung der warmen Blutfllissigkeit, um eine starkere Konzentration des Serums herbeizufi'ihren ; aber schon eine Schwankung des Kochsalzgehaltes um 1 00 reicht wiederum hin, um eine Gestaltsveranderung der Blutkorperchen auszulosen. Verhindert man also die Wasserabgabe des Blutes durch Verdunstung in der oben geschilderten Weise, dann erhalt man auch die richtige Glockenform.

. In 1903 .Weidenreich learned of the recent careful freezing point determinations which showed conclusively that the blood is isotonic with ca. 0.9 per cent sodium chloride solution instead of; 0.6 per cent as he believed. Hence he was forced to retract his former view and conclude that the shape of the corpuscle is not exclusively dependent on the osmotic pressure of the examining medium.

In the same contribution he presented an hypothesis designed to harmonize his previous conclusions with certain well estab


lished facts. Freezing point determinations have shown that human blood plasma is isotonic with a sodium chloride solution of 0.85 per cent to 0.9 per cent in strength (Hamburger, '02; Hober, '02; Dekhuyzen, '01), and, according to Hamburger, a 0.99 per cent solution is isotonic with rabbit's blood. If, therefore, cups exist normally in the blood, why should discs be found exclusively in isotonic saline solutions, whereas cups are first obtained in hypotonic solutions of about 0.6 per cent? Weidenreich explains away this discrepancy by assuming that there is an elastic corpuscular membrane which varies in elasticity in salt solution and in plasma.^ There was postulated a decreased elasticity, due to a swelling of the membrane in salt solution, which opposes the entrance of liquid, thereby preventing the imbibition of as much water as should enter to bring the contents of the corpuscle and the surrounding medium into equilibrium. In other words, the internal pressure of a corpuscle in a 0.6 per cent sodium chloride solution is greater than the pressure of the plasma by an amount corresponding to a 0.3 per cent saline solution, and this is a measure of the tension exerted by the decreased elasticity of the' corpuscular membrane.

Additional evidence was presented in this 1903 paper. Among other things it was stated that if blood, as it issues from a cut, is drawn directly between two cover slips, and the preparation rung with oil to prevent evaporation, isolated corpuscles appear as typical cups.

In 1905 a Weidenreich recommended another method for demonstrating cups. Blood, obtained from an animal by decapitation or by blood letting, was defibrinated and centrifuged, and in this serum blood was examined. Cups, not .discs, were • observed.

Heidenhain ('04) referring to Quincke, held that the effect of colloids on the Molecular kraft" of a medium is important. Starting from this clue Weidenreich ('05 a) reasoned that if the

^ Weidenreich was influenced by the work of Koeppe ('99) who had shown by hematocrit methods that the swelling of corpuscles in dilute salt solutions was not as great as it should be if osmosis alone were responsible — due, he said, to the elasticity of the membrane.


molecular force'^ of a medium opposes the expansion of a plastic body, then by diminishing the molecular force an expansion should result; from Quincke's results it seemed probable that the difference in the effect of isotonic salt solution and serum lay in the presence of albumen in the latter. Gelatin was tried, three per cent in 0.85 per cent sodium chloride solution. In this medium bells were obtained, although excessive rouleaux formation and agglutination into masses occurred, making the result, as frankly admitted, unsatisfactory.

Hence Weidenreich gradually attained the view that the shape of a corpuscle depends on: (1) osmotic pressure, i.e., salt content of medium; (2) 'Molecularkraft,' i.e., colloid content of medium; (3) probably the elasticity of the membrane.

Hamburger ('02) found that corpuscles in lymph lost their disc form and became cups.*^

Lewis ('04; '13) was the first ardent advocate of Weidenreich's contention. He reported that human blood on a warm slide shows cup-shaped corpuscles but when the slide cools and the corpuscles come to rest the conventional discs appear. Weidenreich's early view that a 0.65 per cent saline solution is a suitable examining fluid is apparently accepted and with identical results.

Stohr ('06, p. 115) makes the following non-committal statement: Sie haben beim Menschen und bei den Saugetieren die Gestalt einer bikonkaven Scheibe auch eingedellten Blase ('Glockenform') oder eines flachen kreisrunden Napfchens."

Heidenhain ('04), on the contrary, rejected the general cup thesis as unproven.

By pricking his finger through a drop of vaseline Triolo ('04 a, '04 b) obtained an embedded droplet of blood, the coagulation of which was said to be retarded. Examination showed spheres, which, he states, (p. 309) were 8-10 n in diameter (cf . p. 457) ; " . . . . mais, jaimais dans le sang examine par ce procede de la lubrification, on ne voit le figure classique du globule rouge : le disc biconcave."

^ The vagueness of this conception of the action of a 'molecular force' has been justly criticized by Jolly ('05). '^ Cited by Weidenreich ('02).


Jolly ('04), repeating Triolo's experiment with the blood of the guinea-pig and man, constantly obtained discs; occasionally, and especially at the periphery spheres or crenated forms were seen. Weidenreich ('05 b) likewise pointed out that vaseline is not an indifferent medium and that in ordinary preparations rung with oil, the adjacent corpuscles also ultimately become spheres.

Jolly ('05; '06 a) discredits the cup shape on the grounds that the separating lines in rouleaux are transverse and the terminal corpuscles usually present a plane face, as do free corpuscles.

By using oblique illumination and by observing rotating corpuscles David ('08) became convinced that the 'cup' is an optical illusion which high magnifications increase. He constructed enlarged glass models of biconcave discs and filled them with aurantia ; photographs of these taken at various angles apparently depicted cups. True concavo-convex cups, as resting forms in blood preparations prepared as quickly as possible, were not seen.

Ors6s ('09) was able to induce temporary mechanical distortion in corpuscles but the return was to the biconcave disc which he regards as the equilibrium form in isotonic plasma.

Lohner ('10) decided that to avoid criticism such as had been interposed drawn blood, should be examined vmder conditions which eliminated the evaporation. Accordingly he constructed a cabinet of sufficient size to contain a microscope and into which he could insert his hands through arm-holes; this was heated to a constant temperature of 38°C. and the air saturated with moisture. When the apparatus had reached a state of equilibrium as regards moisture and temperature, blood was drawn from a finger and examined (p. 418): "Wurden nun unter den angegebenen und jedenfalls ziemlich einwandfreien Bedingungen Blutpraparate untersucht, so wurden stets und ausschliesslich nur Erythrocyten in der Gestalt von bikonkaven Scheiben wahrgenommen." To this experiment Weidenreich ('10) replied with his famiUar objection — slowness of observation, due to the awkwardness of working in a cabinet ; he further


records that red corpuscles appear as cups when examined in Ij^mph from the thoracic duct.

In a later paper Lohner ('11) asserted that the cabinet did not hamper his movements in the least, and rejoins concerning the alleged role of evaporation (p. 102): Hier kommt man jedenfalls mit dem Schlageworte Verdunstung nicht aus."

Jordan ('09) examined drawn blood in ordinary preparations and by means of sealed hanging drops on warmed slides. He reports (p. 411) that the biconcave disc form preponderates in numbers very generally in the hanging drop and that all variations from this shape can be interpreted in terms of pressure, contact, or contraction."

Recently Jordan ('15) has communicated his results with the use of Hogan's ('15) normal salt gelatin mixture, it being claimed that this solution simulates the colloidal constitution of blood plasma. Blood was drawn into a drop of the mixture, and a sealed depression slide preparation made in which air was excluded (p. 168): A rapid preliminary examination revealed not a single indubitable cup form. Careful searching may discover a few cups in most preparations." The same technique applied to Tyrode's, Ringer's, and a 0.9 per cent salt solution gave essentially identical results. Cup forms were observed most abundantly in ordinary preparations with Ringer's fluid, the cover glass being supported by a hair (p. 169) : "The explanation that immediately suggests itself is that the floating discs become altered into cups through adjustment to the narrow confines between slide and cover glass."

2. Results from circulating blood

Since blood may be observed intravitally in the transparent parts of animals many extraneous complicating factors are eliminated.

Weidenreich ('02) records his observations (p. 468) :

Ich wahlte zur Priifung am Icbenden Tiere ein Kaninchen; . . . . Wenn die Stromung recht lel:)haft ist, gelingt as allerdings nur schwer, ein einzelnes Korperchen schiirfer in's Auge zu fassen, bei Verlangsamung des Stromes aber oder bei eingetretener Stagnation erkennt man


dagegen leicht, dass auch hier irn Profil die Korperchen die schonste Gloekenform zoisen. Damit diirfte also wohl die Beobachtung gegen jeden Einwand gosicliert scin.

In his next contribution ('03) additional evidence was presented. A rat was killed by decapitation and a thin slice of muscle observed between slide and cover. Cup-shaped corpuscles were seen in circulation. Weidenreich further recommends for study the wing of the hibernating bat.

Lewis ('04, p. 516) reported that, in the omentum of the guniea-pig, "The flowing bodies were seen to be flexible bodies, somewhat variable in their proportions, some deeper, some flatter but all that could be clearly observed were cup shaped." A demonstration was made to Professor Minot who became convinced of the correctness of the view ('12). In his text ('13) Lewis incorporates these conclusions and figures circulating cups.

Triolo ('05) stated that the corpuscles examined by him in the mesentery of the guinea-pig were complete spheres.

Lohner ('10) viewed the capillaries in bits of excised mesentery of the mouse and in muscle fragments. In great part cups were observed but a suspicion that the cups seen were not real arose when the corpuscles emerging from the capillaries appeared as discs; also the corpuscles within vessels viewed strictly in profile showed constantly a disc shape. To test further this deception Lohner constructed a model. Colorless and colored biconcave glass discs 5 mm. in diameter were made. These were placed in a correspondingly large glass tube filled with fluid and the tube laid horizontally in a fluid-filled receptacle having a glass top and bottom. By properly choosing the hquids (alcohol in the tube, xylol or glycerine in the outer receptacle) the effect was said to be startling, one receiving the impression of observing typical cups.

In a series of contributions Jolly ('05, '06 a; '06 b; '09) presented the results of his studies on the circulating blood in the wings of bats prematurely brought out of hibernation. He describes long chains of rouleaux which fill the capillaries and break up within the larger vessels into short segments; this phenomenon


he considers normal (cf. Weber and Soiichard, '80). Jolly emphasizes that the separating lines of rouleaux are transverse, the terminal corpuscles flat, and the free corpuscles discs; exceptionally a cup was seen at the end of rouleaux or isolated. Spherical corpuscles were never observed in the bat, but were seen in the rat and guinea-pig as Weidenreich and especially Triolo ('05) had reported.

Jordon ('09) examined the omenta of two anesthetized cats and reports that both cup or saucer shapes and discs were observable in equal numbers.

According to Schafer ('12, p. 366) the cup view " . . . . can not be accepted for, on examining the circulating blood in the mesenterj^ and other transparent parts of mammals, it is easy to observe that, with few exceptions, the erythrocytes are biconcave."

Of interest to the present discussion is the conclusion of Gage ('88) concerning the red corpuscles of the lamprey. These bodies, described as cup-shaped by several workers (e.g., GiglioTos, '99) are said by Gage to be biconcave discs within the circulation.

3. Results from fixed material

The loss of the nucleus was believed by Rindfleisch ('80) to be responsible for the early bell shape of the erythroplastids, the subsequent assumption of the adult biconcave form resulting from mutual impact. With a variety of fixatives, however, he obtained cup-shaped corpuscles in adult blood.

Howell ('00) considered Rindfleisch's hypothesis erroneous (p. 103): "It seems to me very natural to suppose that the biconcavity of the mammalian corpuscle is directly caused by the loss of the nucleus from the interior."

Malassez ('96) found 2 per cent osmic acid produced cups, and complete spheres.

Cup-shaped corpuscles were described by Dekhuyzen ('99) as a transient developmental stage, yet he records that his assistant, Blote, obtained bells when blood was drawn into osmic acid. Heinz ('01) likewise held cups to be immature forms and also described nucleated cups.


Fuchs ('03) decided that Zenker fixation preserved the original cup shape, whereas he had formerly thought the cup to be an artefact.

Osmic acid was found to produce cups and spheres by Jolly ('05), who, however, questioned the significance of the result ('06 a) because he believed swelling occurred. Later ('09), having first observed circulating discs in the bat's wing, he fixed an area in situ with 1 per cent osmic; after fixation the corpuscles were found to be spheres.

Lewis ('04) showed that pricking through a drop of osmic acid produced many shrunken corpuscles and cups. Zenker's fluid acts violently on drawn blood. From the study of the tissues of various mammals he concludes (p. 516) : "In preserved mammalian blood the typical red blood corpuscle is cup shaped. The biconcave disc is but one of several forms of shrunken cups" In 1913 he again says (p. 192): .... in well preserved tissues of all sorts, and with all fixatives such as are relied upon to reveal the structure of other tissues the mammalian erythrocytes are typically cup-shaped. . . . where, the tissues in general are excellently preserved the corpuscles appear as cups. The biconcave discs are flattened cups."

Weidenreich ('02) considers the evidence gained from the use of fixatives (e.g., 1 per cent osmic acid) alone sufficient to establish the cup shape. In 1905 b he stated that osmic acid preserves corpuscles which have become discs as discs. Later ('06 a) he recommended a rapid method for preserving smears by osmic acid vapor; the form was said to be fixed in three to five seconds, this being less time than is required for the corpusles to change shape osmotically.

Human material, several hours old, was preserved by Radasch ('06). Several organs of the body were examined, including the placenta with its maternal and fetal blood. A large majority of cups were observed.

Jordan ('09) records his experience with fixatives although he discounts the value of evidence obtained by such methods. With various reagents cups, irregular forms, and a few irregular discs were found whereas in one Zenker-fixed preparation discs


preponderated. He concludes that fixation causes contraction, which is probably unequal at the center and rim, thereby producing cups.

Lohner ('11) elaborated this latter view. The coagulation of fixation involves a diminution in diameter. By experiment he showed that when blood issues from a puncture of the skin into a drop of fixative, or when it is drawn by capillarity into osmic acid between two cover glasses, the conditions are present for an unequal action of the fixative with a resulting distortion of the corpuscles. Corpuscles meeting the fluid edge foremost become wedge-shaped; when the flat surface is first fixed a cup results. Discs are obtainable provided the fixation is uniform.

Wiedenreich ('10) accused Lohner ('10) of being inconsistent (p. 448), for in living animals he held cups to be illusions and in drawn blood he considered them artefacts. Having observed cups in a portion of excised mesentery, Weidenreich added fixative and still saw cups, which, he said, were not illusions, for if squeezed from the vessels they retained the cup appearance. Lohner ('11) interpreted this experiment as follows. In the vessels there are discs which may give a deceptive appearance of cups, as well as temporary cups due to distortions, and real cups. The apparent cups were changed to artificial cups by the fixative; hence this form was seen when the corpuscles were pressed from the vessels.

It is evident from the foregoing resume that the normal shape of the erythroplastid remains undetermined. The hope of obtaining new evidence on this fundamental question has induced me to undertake the pr<&sent \Vork, concerning which a preliminary communication has already been published (Arey, '16).


1. Expeiimentation with drawn blood

In ordinary preparations of undiluted blood, made as quickly as possible and examined between warm slides and covers, I have usually observed a few cups intermingled with large numbers of discs. The transformation of cups into discs, which


according to Weidenreich and Lewis occurs largely prior to microscopic inspection, I have never seen, and, so far as I am aware, no one asserts to have actually traced this transformation in individually scrutinized corpuscles. With the formation of rouleaux the apparent number of discs is, of course, increased, for more are seen on edge.

The factors which might be suspected as responsible for this alleged alteration are decreased temperature and increased concentration of the plasma. Weidenreich, in particular, has insisted on these factors as causing the widespread 'deception' concerning the true shape of the blood corpuscle.

The statement that the evaporation and consequent concentration of a blood droplet, before the preparation can be made and examined, is sufficient to inaugurate these modifications can not be arbitrarily dismissed no matter how improbable it may seem in the light of the readily observed effects of dilution and concentration upon the shape of red corpuscles ; when, however, a large drop of blood is used the concentration change must be slight. But that it is necessary to maintain an elevated temperature to prevent rapid evaporation (Weidenreich, '02) is as astonishing as a statement as it is embarrassing to defend as a doctrine (cf. p. 443).

The momentary exposure to air necessitated in making ordinary preparations may practically be eliminated by using the following procedure. Superimposed cover glasses, separated by a hair, are fused at one point by heat ; if an edge be now applied to a needle prick in the finger the issuing blood is drawn by capillarity between the two surfaces. Such preparations, examined quickly with or without the aid of the warm stage, have never jdelded evidence for the general existence of the cup shape. A few cups may usually be found whereas scores of indisputable discs appear.

Since only corpuscles viewed edgewise furnish reliable data as to shape, any pressure effect from the approximation of the slide and cover glass in the foregoing experiments will tend to increase the number of cups seen.


The experiments in which Weidenreich ('05 ar) added gelatine to 0.85 per cent sahne solution in order to reduce the 'Molecularkraft' have been described (p. 445). Admitting the results to be unsatisfactor}^ he, nevertheless, features them prominently and emphasizes their significance ('05 a; '05 b). If the recommended three per cent gelatin (purified and dialized) be added to 0.85 per cent sodium chloride, or to Tyrode's solution, a medium is obtained which is a dense gel at room temperature. This obviously does not simulate blood plasma, but it was chosen, we are told, because it is the optimum concentration and gives the best results.

When the finger is pricked through a drop of this gel and the resulting mixture is^ examined in a hanging drop, abundant rouleaux form, and the corpuscles commonly agglutinate into amorphous masses or become distorted and tailed. Cups of various shapes and som,e discs are also to be found. If the foregoing experiment be duplicated, except that an assistant by means of a needle mix the small droplet of issuing blood evenly throughout the gelatin, the resulting preparation more closely approximates the normal. Furthermore, the number of discs seen is increased. The distorted, agglutinated, and ruptured corpuscles in the first case are apparently referable to the resistance of the dense gel ; the issuing blood as it breaks up into tiny streamlets w^hich dart along irregular paths in the gel following the line of least resistance testifies strikingly in favor of this probability.

When the 3 per cent gelatin mixture is warmed it changes to the sol condition. If a small droplet of blood be drawn into a drop of the gelatin solution at body temperature, care being taken as before that the mixing is even, rouleaux formation need hardly exist. Moreover, large numbers of typical biconcave discs are observable in edge view^; a few cups may also be found. I therefore conclude that with certain precautions this experiment proves the precise converse of that which Weidenreich designed it to show. Confirmatory results are found in the recent work of Jordan ('15) who decides that human red blood



corpuscles are biconcave discs in Hogan's normal salt-gelatin mixture which contains 2.5 per cent gelatin.

But the foregoing evidence is not crucial. Blood is not a gel at room temperature; neither does it contain gelatin; nor is the arbitrarily chosen 3 per cent solution rational. If colloids are to be added it is highly desirable to use the proper amount of the protein normally present in blood.

Beside fibrinogen these proteins are serum albumin and serum globulin. Due to the conditions imposed by the world war I was unable to obtain serum globulin but did procure a purified sample of the closely similar serum albumin (Merck). To Tyrode's solution, which is claimed to duplicate accuratelj^ the inorganic composition of blood, was added enough serum albumin (re-dialized to make certain of its purity) to correspond to the amount of both albumin and globulin normally present in plasma. Blood corpuscles examined in this diluting medium proved to be almost exclusively discs.

Experimentation with undiluted blood is, at its best, unsatisfactory. The crowded conditions, the tendency toward rouleaux formation and coagulation which make such preparations unfavorable are obviated by the use of diluting media. If, however, artificial 'physiological solutions' be used, the results may ever, though perhaps unjustly, be subjected to criticism. At best these are artificial media, the tonicity and colloidal constitution of which may or may not simulate blood plasma. To preclude such criticisms natural serum must be used. Accordingly I had 20 cc. of blood drawn from my basilic vein. This was defibrinated by whipping and centrifuged quickly; thus an examining medium was obtained, identical with blood plasma except for the loss of one of its minor protein constituents — fibrin.

By utilizing an electrically heated warm-stage a hollow-centered life slide, cover glass, and the air of the cell itself may all be maintained constantly at body temperature. A drop of serum ^ was placed on a finger previously cleaned with alcohol, and the finger pricked through the drop. The diluted droplet

' Previous microscopic examination had made certain that the serum was free from blood corpuscles.


of blood, thus obtained without direct contact with the air, was touched to a cover and suspended, as a hanging drop, in the Hfe cell,^ Vasehne served to seal the cell, the air in which could be kept saturated with moisture by previously introducing a drop of water and sealing. The entire procedure demands no more time than in making ordinary preparations; if a large drop of serum be used, the evaporation prior to sealing is inconsiderable whereas further evaporation within the life cell can not occur.

A microscopic examination of blood prepared according to this technique reveals numerous isolated corpus,cles. A favorable place for scrutiny is near the center of the drop. Here sinking corpuscles revolve slowly, showing alternately their two faces. Usually a few cups can he found, whereas quantities of biconcave discs are seen in every field.^ Streaming movements initiated by rolling the suspended drop towards the edge of the cover also allow many corpuscles to be viewed from both surfaces.

Another technical procedure, used by Jordan ('15) in his experiments with physiological solutions, consists in filling shallow concave sUdes with serum into which the drop of diluted blood, prepared as before, is introduced. Evaporation is prevented by immediate sealing with a cover glass and vaseline. Body temperature is maintained by the aid of the electrical warm stage. The conclusions drawn from the study of many such preparations substantiate those already reached with the hanging drop.

The necessity or desirability of observing blood which has not been allowed to cool has been emphasized by those who uphold the normality of the cup shape. Weidenreich ('02) contended that the use of cold slides and covers is largely responsible for the widespread 'deception' as to the true shape of the red corpuscles; in his contributions of 1903, 1905 and 1910, however, he apparently abandons this contention for he urges only the necessity of rapid manipulation. Lewis ('04) states that as the

Room temperature 27°C. ' The disc shape is retained in properly sealed preparations 48 hours old.


preparation cools the cups become biconcave discs arranged in rouleaux. From my personal experience I do not believe that temperatures between 0°+ and 40°C. directly condition the shape of the erythroplastids. Hanging drop preparations, cooled for several minutes on pulverized ice, precautions being observed to prevent dilution of the drop by condensation of moisture, show no essential difference from those at body temperature examined immediately; if retained for 10 minutes the free corpuscles are also typical discs. Subnormal temperature of itself induces neither crenation nor rouleaux formation. These tests merely show that cooling does not modify the shape of the disc; those who defend the cup shape would maintain that an almost instantaneous change from cup to disc had already occurred while the preparation was being made.

Experimentation with other samples of human serum has been possible through the kindness of three of my colleagues. ^"^ The results obtained both when corpuscles were examined in their own serum and in each of the other three sera were identical with those already described. More cogent proof concerning the primary shape of the human erythrocyte to be derived from the study of drawn blood, I can not imagine. Similar extensive tests have likewise been made with 0.85 per cent and 0.9 per cent saline solutions and with Tyrode's solution;" the latter is claimed to simulate blood plasma more closely than other physiological solutions with the possible exception of Hogan's mixture.

It is a familiar fact that the dilution of drawn blood with water causes the red corpuscles to assume a spherical shape, whence they 'lake' and become colorless spheres; on the contrary media stronger than normal plasma crenate the corpuscles. Various dilutions of human serum in distilled water were next prepared and used as dilution media for hanging drop preparations. The ultimate concentration of any mixture obtained by

10 Prof. S. W. Ranson, Mr. L. H. Kornder, and Mr. M. R. Waltz. For operative assistance I am indebted to Dr. Joseph Jaros.

" For formula see Rona, P. und Neiikirch, P., 1912. Archiv f. d. gesam. Physiol., Bd. 148, pp. 273-284.


adding a droplet of blood to a diluted serum obviously depends on the relative amounts of each used. It is believed that the percentages stated in the following paragraph are sufficiently accurate for the purpose at hand.

WTien a droplet of blood is mixed with human serum containing ca. 25 per cent water some erythroplastids assume the shallow cup shape shown in figure I, B; most, however, remain as biconcave discs (A). When ca. 40 per cent water is present there is a great preponderance of typical cups, with here and there an unchanged disc. These cups appear somewhat like figure 1, D in ca. 50 per cent mixtures. In dilutions containing ca. 60 per cent water, the walls of the cups become swollen and the concavity is reduced (E); this imbibition is so marked in mixtures of ca. 65 per cent water that the appearance is that of deeply dimpled spheroids (F). Perfect spheres result when the water content of the mixture is ca. 70 per cent.^'- In concentrated serum erythroplastids crenate.

It is evident, therefore, that the shape of a corpuscle is, at least in part, a function of the concentration of the medium, the changes being referable to the action of osmotic pressure. In progressively hypotonic solutions the corpuscles imbibe increasing amounts of water, ultimately becoming spheres and laking. In hypertonic media, water is given up and crenation results. All corpuscles, however, are not affected similarly by the same concentration. This is strikingly shown by crenation experiments and especially by dilution phenomena. When the percentage of water present is 25, only part of the corpuscles are clearly affected. Such a result might conceivably be due to the unequal elasticities of the corpuscular membranes which oppose differently imbibitory swelling.

Analogous series were obtained by diluting Tyrode's solution and 0.85 per cent saline with distilled water. The shapes of the red corpuscles at the various dilutions approximated closely

'^ The following measurements hold:

Diameter of disc 7.5 Diameter of cup 7.0=1= Diameter of sphere 5.0



those already described when human serum was used. This result supports the standard freezing point determinations of Hamburger ('02), Hober, ('02) and Dekhuyzen ('02) which find human plasma isotonic with an 0.85 to 0.9 per cent saline solution. On the contrary it militates directly against the conclusion of Weidenreich and Lewis, who, notwithstanding the determinations just alluded to, hold that a 0.6 to 0.65 per ctnt^f-alt solution is isotonic with human plasma; these workers have

Fig. 1 Profile sketches illustrating the shape assumed by the human erythroplastid in various dilutions of human serum with water. A, in undiluted serum > B, ca. 25 per cent water; C, ca. 40 per cent water; D, ca. 50 per cent water; E, ca. 60 per cent water; F, ca. 65 per cent water; G, ca., 70 per cent water.

arrived at this point of view in the following way: blood issuing from the veins is supposed to change suddenly from the cup to the disc shape, due presumably to a slight concentration through evaporation (or to loss of heat; Lewis?); on introducing drawn blood or blood diluted in 0.9 per cent salt solution saline discs become cups, hence the cup shape is normal and a 0.6 per cent salt solution is isotonic with plasma.

It was found that erythroplastids not only assume shapes which are correlated with the concentration of the medium but that these changes are also repeatedly reversible. Typical experiments will illustrate this behavior.


Experiment l.G.l. A drop of human blood, diluted in human serum, on examination showed the erythroplastids to be typical discs. When a droplet was transferred to a hanging drop, composed of half serum and half water, the corpuscles assumed cup shapes (fig. 1, -S). The disc shape was recovered by retransferring to normal serum but in the 50 per cent serum-mixture swollen cups (like fig. 1, F) were again obtained. Following a return to discs in normal serum, crenation was effected by transference to somewhat concentrated (evaporated) serum.

Experiment 1.6.2. Disc-shaped corpuscles in normal serum crenated in hypertonic serum. Transference to serum diluted one-half with distilled water induced a return to discs for some corpuscles and to cups for others. In hypertonic serum crenation again occurred.

As the size of the droplet used in the transfer affects greatly the ultimate concentration of the mixture, the figures given in these experiments have no quantitative value. Considerable variability was found in the responses of individual corpuscles. It is perhaps significant that in experiment 1.6.1 the second transfer to dilute serum produced more highly swollen cups (fig. 1, F) than did the first transfer (fig. 1, -B) ; that it is indicative of an increased elasticity of the corpuscular membrane through injury is not impossible. Corpuscles in which crenation has proceeded too far seem to be permanently injured and incapable of a return to the normal; similarly figure 1, F, marks the approximate stage beyond which corpuscles become irreversibly altered. In terms of saline solutions these limits correspond to concentrations between 0.9 -f- per cent and 0.3=*= per cent. The possibility of the action of a toxic time factor was not investigated.

The importance of these diverse dilution phenomena on the question of the normal shape of the human erythroplastid seems to me paramount. Since within wide limits the form of a corpuscle depends on the concentration of its medium, how can the cup shape he normal when human serum must he diluted at least one-third to produce this type?

Experimentation with the serum of the cat and dog, both as regards their own corpuscles and those of other individuals and of man, has confirmed the conclusions already reached concerning the normaUty of the biconcave disc.


With the guinea-pig, rat, and rabbit I have obtained variable results. The corpuscles of one guinea-pig examined in their own serum were constantly discs; in another specimen, tested twice in four days, it was impossible to find corpuscles except in crenated condition; in a rabbit it was difficult to obtain preparations which did not show extensive crenation in their own serum; although human blood in this serum showed discs almost exclusively. The blood of a white rat examined in its own serum had one-half or more of the corpuscles strongly cupped; the blood of two other rats examined in the serum of the first also showed a majority of cups, although human blood corpuscles remained discs. When serum from one of the last mentioned rats was prepared its own corpuscles were discs, as were also the corpuscles of the third individual. I have not worked on many individuals of these species which have been used so extensively by other experimenters (Weidenreich, Lewis, et al.), but the variable results just cited do not inspire confidence in the employment of this class of animals; from my experience they are untrustworthy and unfavorable material. The untested suggestion presents itself that in those rodents which usually do not drink water, but depend on green vegetables for their

supply, the concentration of the plasma may vary; external temperature (the above tests were made in June) and the rations lof the average animal house perhaps play no inconsiderable role.

2. Examination of circulating blood

Blood observed circulating in the transparent parts of mammals should furnish extremely reliable data concerning the question at hand. There are, however, certain technical difficulties to be overcome, as well as the infeasibility of observing rapidly moving corpuscles under high magnifications.

It is conceivable that the pressure on the deUcate vessels, caused by the ordinary use of a cover glass and oil immersion objective applied to the omentum (Jordan, Lewis, et al.) might induce the assumption of the cup forms "through narrowing the confines to which the delicate discs must adjust themselves"


(Jordan). To avoid this pressure I employed an immersion objective with Tyrode's sohition, as in the water-immersion lenses of former days ; in this way the spread omentum was observed directly without the aid of a cover glass. A Leitz no. 4 dry objective and a no. 12 compensating ocular, with the draw tube of the microscope set at 190 mm. also gave a very satisfactory magnification and was used extensively as a check on the wet method.

The omenta of eight cats and two dogs were studied continuously for periods from one to four hours. The animals used were purposely in a state of deep surgical shock resulting from previous laminectomies', in each case the anesthetic had been stopped two to four hours prior to the experiment.

Capillaries of sijialler calibre than the diameter of an erythroplastid are obviously unsuited for observational purposes. Both the vessels figured by Lewis ('13) are open to this objection; of the eight cup-shaped corpuscles shown in edge view, six if flattened to discs would exceed considerably the diameter of their containing vessel.

Regions of the omentum where temporary stases have caused corpuscles to adhere in clumps or agglutinated masses I do not consider favorable; when the flow is resumed many cups are seen, the cup form apparently being in some way referable to the former massed condition.^* The rapidity of normally circulating blood makes it impossible to observe satisfactorily the individual corpuscles which pass across the field as an ill-defined blur;^^ in the rhythmical release from stasis which sometimes occurs in a pulsating fashion the corpuscles are mutually compressed to an unfavorable degree.

Since ordinary circulation is much too rapid to enable accurate observation, I believe that the most reliable data are obtainable under the following conditions. It is sometimes possible to find a bifurcation of precapillary or larger vessels in which

" It is to be noted that Weidenreich ('02) made his observations when the current had slowed to the point of incipient stasis.

• Such illustrations of corpuscles within vessels, as figured by Lewis ('13) could not have been drawn from normally circulating blood as the legend implies.


the flow selects one limb almost exclusively, separate corpuscles, nevertheless, being intermittently 'kicked off' into the slowly moving plasma of the other limb.^^ Such a situation, where the current in the main vessel is rapid and normal (to find which often necessitates considerable diligent search), I regard as most favorable for study. Criticisms of pressure, agglutination, and of observing vessels so small that the corpuscles must necessarily adjust themselves to their exiguous confines are obviated.

Erythroplastids emerging from the main stream one or two at a time in the manner indicated were found to be discs; most of these corpuscles are revolving when first seen and it is easy to be certain of their biconcavity. In such situations I have observed hundreds of discs with only an occasional cup- or saucer-form;'^ this observation has been corroborated by several of my colleagues.

In anesthetized guinea-pigs and rabbits, cups were very common, and in a dog under ether anesthesia a great preponderance of cup shapes was observed. The query immediately presents itself whether under these conditions the anesthetic is responsible for the cup shape. The following experiment is highly suggestive:

Experiment 1.4.1. Hanging drop preparations of human blood and the blood of the cat, dog, guinea-pig, rabbit, and rat, diluted with serum, were made. When a drop of ether or chloroform was now introduced into the bottom of the life cell the drop took on the vapor and the discs were seen to change rapidly through the various cupshapes to spheres, finally laking and becoming shadows. ^'^

I believe that my observations indicate that the erythroplastids of normal circulating mammalian blood are biconcave discs ;

^* For making these observations I can particularly recommend the dog.

^^ Perhaps the number of cups is somewhat increased by the presence of corpuscles brought by the capillary net from regions of the omentum in stasis.

1' A curious surface tension effect was obtained when corpuscles were removed from a hanging drop of Tyrode's solution before the effect of the anesthetic had proceeded far. On transference of these cupped forms to a drop of pure Tyrode they became discs whereupon some moved edge foremost across the field with a wobbling motion for longer or shorter distances then turned abruptly and continued at an angle. This would be repeated for some time.


the burden of proof rests on those who have used anesthetized animals (and apparently most previous workers have done so) to show that the anesthetic held in the blood is not responsible for the preponderance of cups observed.

3. Action of fixatives

Many workers have recorded that mammalian tissues preserved in various standard fixatives contain cup-shaped erythroplastids. Those who uphold the normality of the cup have laid great stress on this line of evidence particularly in view of the fact that '^in well preserved tissues of all sorts, and with all fixatives such as are relied upon to reveal the structure of other tissues,' the mammalian erythrocytes are typically cup-shaped" (Lewis) .

Smear preparations, such as advocated by Weidenreich, would hardly seem to furnish reliable data concerning the moderately flexed saucer shapes which are held by these workers to be the normal intravital form. It is necessary that the corpuscles be observed more ot less on edge.

An examination of what is ordinarily called well-preserved mammalian tissues demonstrates convincingly that the cup shapes are indeed preponderatingly abundant within the vessels. ^^ With Jordan ('09), however, I must deny the universality of this statement. Occasionally tissues have been observed in which the corpuscles seen were discs almost exclusively. Not only have cups and discs been observed within the same vessel, but rarely vessels, have been found side by side, one containing cups, the other discs.

I have fixed small pieces of human vascular fat, obtained fresh from operations, in 1 per cent osmic acid, in saturated sublimate in 0.75 per cent sodium chloride, and in the fluids of Zenker, Orth and Helly. In each case only a few moments elapsed between the removal of the tissue and its immersion in the fixative. Celloidin sections showed constantly a great preponderance of cups.

^* Sites must of course be chosen where the corpuscles are well separated.


In view of the rapid action of hypo- or hypertonic salt solutions in changing the shape of corpuscles an attempt was made to discover whether the concentration of the fixative could influence the shape of the corpuscles before fixation occurred. Zenker's fluid was practically saturated with cane sugar or with sodium chloride and fresh human tissue fixed as before. The result w^as unchanged; the corpuscles became cupped.

When to drawn human blood, diluted in human serum, is added the fluids of Zenker, Helly, or Orth, cups, discs, and distorted forms are seen. The action is more violent than when blocks of tissue are preserved in the same fluids. With Zenker an especially curdy coagulum forms, w^hereas in Orth there is only a fine granular coagulum; in the first named fixatiye cups are abundant, in the latter many fine discs may also be obtained. In Perenyi's fluid the corpuscles assume a peculiar pitted appearance.

If the finger be pricked through a drop of 1 per cent osmic acid solution, the fixed corpuscles show many cups as well as discs, wedge shapes, and distorted forms. If a drop of blood be first exposed to the air and the osmic acid then added, a greatly increased number of discs are seen, although discs are by no means exclusively present as Weidenreich ('05 b; '10) would have us believe. These facts have been advanced in support of the normality of the cup shape, for it is argued (Weidenreich '08 c) that osmic acid must give faithful preservation since it fixes not only the cup but also discs which have been formed from cups after exposure to air.

The query immediately arises as to the weight which should properly be given to evidence derived from the action of fixatives. Weidenreich ('02), for instance, considers this evidence alone sufficient to establish the cup form (cf. also Lewis '13).

The fact, however, must not be lost sight of that these corpuscles are plastic structures of extreme delicacy, mere contact with adjacent corpuscles or obstacles sufficing, when gentle streaming is induced, to cause excessive and varied temporary distortions. Fixation is essentially a coagulation process and it has been shown (Weidenreich '06 b) that the so-called best fixa


lives actually diminish (i.e., shrink) the diameter of the corpuscle. If this shrinkage were unequal at the thin center and thick rim a disc might conceivably become a cup, as Jordan maintains. Furthermore, if the reagent does not act on all sides of a corpuscle simultaneously, is not a buckling of the more contracted side on which the reagent first acts to be expected? Indeed, the preexistence of biconcavities would favor this alteration. It seems plausible that the delicatelj^ constructed erythroplastid is more easily subject to distortion, through the action of reagents, than are ordinary tissue cells, for it is neither supported by contiguous cells nor by intercellular products.

If this reasoning be sound the variable action of osmic acid on drawn blood allows of another interpretation. When blood enters a drop of fixative directly from a minute needle prick in the finger the conditions for unequal fixation would appear to be present (Lohner). Besides cups, numerous wedge-shaped corpuscles are seen; according to the conception of uneven fixation such forms are easily explained. The presence of more discs in blood that has first come into contact with air need not be interpreted as due to a rapid change from the cup to the disc with a subsequent fixation of the latter form; assuming that such corpuscles have not been exposed to air sufficiently to induce incipient crenation, which conceivably could affect the physical condition of the corpuscular membrane (without, however, necessitating a change in form) , the result is explainable on the basis of a more even intermixing of blood and the added fixative.

The following experiment of Lohner ('11) which I have often corroborated is instructive from this viewpoint:

Experiment 2.1.3. If a droplet of blood be drawn by capillarity between two cover slips, separated by a hair and fused at one point, discs are observed. (Blood should occupy part of the capillary space only.) If 1 per cent osmic acid be now drawn in cautiously from one side only, the conditions for uneven fixation are present and many cups, some wedge-shaped discs, discs, and distorted forms are seen.


Only certain aspects of the problem of a more or less general nature will be considered here, critiques of individual results and


methods having for the most part been introduced in connection with the previous section.

We have seen that the examination of undiluted drawn blood has led various workers to diametrically opposed conclusions. Those who champion the cup form believe that a rapid transformation of cups into discs, before preparation are made and examined, is responsible for the finding of discs by slower motioned workers. The whole cup hypothesis, therefore, hangs on the sudden secondary transformation of cups into discs when in contact with the air. The reproach of slowness, which has been repeated so frequently is, nevertheless, not incontestable but is open to scrutiny and analysis.

Two factors have been emphasized as responsible for the alleged sudden mutation of shape. Weidenreich ('02) and Lewis ('05; by implication) urged the necessity of maintaining normal temperature if cups are to be seen: Weidenreich' s position was obviously untenable (p. 452) and in his papers of '03, '05 a, and '10 he abandoned his insistence on temperature as a causative modifying agent. The second factor is that of evaporation resulting in an effective concentration of the plasma, before drawn blood is observed." By drawing in blood, as it issued from the cut, between two parallel cover glasses fused at one point, the exposure to air before examination was reduced to a rhinimum (p. 452) yet I am not able to conclude that these experiments support the cup theory. Lohner's ('10) tests w^ith his constant-temperature and moisture-saturated cabinet (p. 446) precluded evaporation, yet he obtained '^stets und ausschliesslich" biconcave discs. In the experiments in which I diluted a droplet of blood in a large drop of human serum, without the blood first coming in contact with air, the increased concentration of the mixture during the few seconds before the preparation was sealed must have been negligible; when it is further discovered (p. 457) that human serum must be diluted one-third to obtain cups, the futility of the evaporation argument becomes apparent. For these reasons I am unwilling to admit that the evidence derived from undiluted drawn blood either establishes or supports the normality of the cup.

From the work of Ranvier in 1875 it has been known thai;


graded temperatures can alter disc-shaped corpuscles to deep cups, thick- walled cups, or even to spheres, e.g., typical cups are found exclusively when blood is warmed to 55° (Zoth). It is impossible that some investigators who advocate the cup shape have unduly heated their slides and covers (perhaps in attempting to allow for cooling when warm stages were not available) in overzealous attempts to maintain normal (!) conditions.

We have seen (p. 444) that Weidenreich, at a loss to reconcile the cup shape of corpuscles in 0.6 per cent salt solution with the disc shape in the well established isotonic 0.9 per cent first held as responsible a decreased elasticity of the corpuscular membrane in saline solutions; later he shifted the emphasis to a hypothetical influence of a changed 'Molecularkraft' in the solution due to the presence of colloids. Both the results of Jordan ('15) with Hogan's normal salt-gelatin mixture and my own in repeating Weidenreich's experiment are not in agreement with the latter's conclusion; hence I believe that Weidenreich is still confronted with his original dilemma. This conviction is strengthened by the fact that when serum albumen was added to normal saline and Tyrode's solutions in an amount which duplicated the protein content of blood plasma, I obtained an examining medium in which the corpuscles were unquestionably discs.

Since the results of freezing point determinations are not accepted by those who champion the cup shape as giving reliable information regarding the isotonicity of physiological salt solution (p. 443), it is evident that the use of artificial media alone serves only to incite controversy. For these reasons much of the work of Weidenreich and that of Lewis ('05) and Jordan ('15) with respect to this point is in itself not crucial. If one believes in the cup shape and is able to obtain this form only in 0.6 per cent instead of the accepted isotonic 0.9 per cent saline solution, he of course can ever invoke the aid of extraneous factors to explain the discrepancy. The escape from this quandary lies in using serum as the diluent. I have already given my reasons (p. 460) for distrusting the data obtained from the use of the rat, guinea-pig, and rabbit. Hence I must for the pres


ent remain skeptical concerning the value of data obtained from these animals both as regards the corpuscles of their undiluted blood or from the use of their blood sera and lymph as diluting media. On the other hand in my experiments (p. 454) in which the blood of the cat, dog, and man (four individuals) was examined both in their own sera and each of the other sera, I constantly obtained biconcave discs almost exclusively. Furthermore, the necessity of diluting human sera one-third with water to obtain the cup-shape is to my mind incompatible both with the doctrine that the cup shape is normal and with the view that a 0.6 per cent salt solution is isotonic with human plasma.

In studying the circulating blood of living mammals the results recorded in this paper were obtained without involving the possible distortion of corpuscles through pressure from a cover glass and oil immersion objective as has formerly been the case. Although I am not altogether certain that this is a real danger, as Jordan ('15) believes, it is, nevertheless, easily and properly avoided. If capillaries of too small calibre to possibly allow the assumption of the disc shape (p. 461) be not chosen and if blood which is not, and has not been in stasis be observed in nonanesthetized cats or dogs, I feel sure that my observation of a great preponderance of discs can be verified, Here, again, the animals formerly used have largely been either guinea pigs or rabbits whose appropriateness is questionable. Since ether and chloroform visibly change discs to cups or spheres (p. 462) those who make use of anesthesia must disaprove its effect intra vitam.

Weidenreich ('03) reported cups in the wing of the living bat and asserts that the discs seen by Jolly ('05; '06 a; '06 b; '09) in the same location represent cups which had previously formed rouleaux, these being again resolved into their constituent elements and then appearing as discs. One would like to know more about the details of these experiments. Were anesthetics used? Jolly animals were brought out of hibernation and showed excessive rouleaux formation which he considers a normal intravital condition but which is more likely referable to the recent hibernating condition or to partial stasis. No details of his observations are given by Weidenreich except that he used


a hibernating bat. For several reasons the bat might be expected to furnish valuable evidence on this problem and arrangements are under way by the writer for the further study of these animals.

David ('08) first called attention to the resemblance which a biconcave disc, viewed obliquely, bears to a cup, a deception which is intensified by hi^h magnifications and which he illustrated by photographing glass models. Lohner ('10) developed this idea and constructed an elaborate model (p. 448) which was said to corroborate his view. That a biconcave corpuscle strikingly simulates a cup when viewed obliquely is true, but that this illusion alone has influenced a decision on the part of other observers favoring the cup shape seems improbable. Cups and discs viewed in profile are unmistakable and it is from profile views that crucial evidence must be derived.

But little need be added regarding the action of fixatives. Serious doubt has been cast on the trustworthiness of standard fixatives in preserving the original shape of red blood corpuscles (p. 464). Both Radasch and Lewis regard discs as representing collapsed cups; It may be thought that the depression which makes the cup is itself due to shrinkage, or due to vacuole formation. The only proof to the contrary is to be had from the circulating blood of a living mammal"- (Lewis '04, p. 516). A critique of the methods and results of this 'only' source of proof to the contrary has been sufficiently dealt with in the foregoing pages.

It is interesting that agents such as heat (Ranvier, '75), electricity (Lohner, '07), and ether or chloroform produce cups from discs. It is perhaps significant that these so-called destructive agents in each case alter discs to cups, not the reverse.

A limited and probably inconstant number of cup-shaped erythroplastids undoubtedly are present in normal blood. Possibly they represent corpuscles, whose structure is such that unequal tensions with respect to themselves or to the osmotic balance exist; perhaps they are old (or young?) corpuscles. In anemias the presence of many cups has been reported (Quincke, '77; Grawitz, '99) and in fevers it is said crenation may occur



(Grawitz, '02). Experimentation with diluting media at the critical concentrations which first produce the cups from discs, or which cause laking and crenation, makes it certain that there is considerable variability in the responsiveness of individual corpuscles. Weidenreich ('02) further notes that there is a limited variation in different individuals, and Lackschewitz ('92) and Hamburger ('02) have compared the unequal resistance of the corpuscles in certain of the lower animals.

May it be that the blood of certain individuals contains normally excessive numbers of cup-shaped corpuscles? Is it possible that this explains why some of our most careful workers have been led to describe this form as normal?

The teachings of comparative histology do not support the cup shape; but it may be objected that the loss of the nucleus among mammals is in itself directly or indirectly responsible for the assumption of the cup shape (cf. Rindfleisch, et al.). In this connection might be mentioned Howell's ('00) statement that biconcavity is a physical advantage because the absorbtive surface is increased, and the conclusion of Rice ('14) that the biconcave form is physically the 'best' since it is one having less surface energy than any surface obtained from it by a small deformation consistent with constant volume.

It is conceivable that the action of hypotonic solutions in swelling corpuscles assymetrically is associated with the loss of the nucleus. Whether the side through which the nucleus is expelled becomes more elastic (weakened) or less elastic (i.e., like scar tissue) is, however, pure conjecture.


The shape of the mammahan red blood corpuscle depends largely on the osmotic pressure of the examining medium. In solutions corresponding to ca. 0.9 per cent sodium chloride the erythroplastid possesses a biconcave form. In progressively less concentrated (hypotonic) solutions water is imbibed and the corpuscles swell to thin-walled cups, thick-walled cups, dimpled spheres, and finally lake formjng 'shadows.' In hypertonic media crenation results.


Between the limits of form induced by a 0.3 per cent sodium chloride solution and by mild crenation the shape of the red corpuscles is repeatedly reversible.

Indi^'idual variability exists in the response of erythroplastids to diluting media; this is perhaps referable to diverse elasticities of the corpuscular membranes.

Undiluted drawn blood, and blood diluted with human serum, show the red corpuscles to be biconcave discs. Human serum must be diluted about one-third with water before ,the cup form predominates.

Freezing point determinations which show that mammalian plasma is isotonic with a 0.9 per cent saline solution (instead of 0.6 per cent) are roughly substantiated by such dilution experiments.

The study of circulatitig blood in non-anesthetized living mammals corroborates the normality of the disc.

The results gained by the use of fixatives, although seemingly adverse to the disc view, may be satisfactorily interpreted in terms of unequal fixation; this is supported by experiment.

The several lines of evidence here presented seem to justify the conclusion that the biconcave disc represents the normal shape of the mammalian erythroplastid — the concavo-convex cup being merely an occasional modification.


Albrecht, E. 1904 Cytopathologische Mitteilungen. Verhandl. d. deutsch.

Pathol. Gesell. Jahrg., Heft 1, pp. 88-94. Arey, L. B. 1916 The mammalian erythrocyte — a biconcave disc. Science,

n.s., vol. 44, no. 1133, pp. 392-395. Brass, A. 1897 Atlas of human histology. Trans, by R. A. Young. Wm.

Wood and Company, New York, 160 pp. V. David, C. 1908 Ueber optische Einstellungsbilder kreisscheibenformiger

Erythrozyten. Arch. f. mikr. Anat., Bd. 71, pp. 159-163. Dekhuyzen, M. C. 1899 Becherformige rote Blutkorperchen ('Chromokra teren'). Anat. Anz., Bd. 15, no. 11 und 12, pp. 206-212.

1901 tJber die Thrombocyten (Blutplattchen). Anat. Anz., Bd. 19,

no. 21, pp. 529-540. DujARDiN, F. 1842 Nouveau manuel de I'observateur au microscope. Li brairie encyclopidique de Roret, Paris. 330 pp. vox Ebner, V. 1902 Vom Gefasssystem. Koelliker's Handb. d. Gewebelehre

d. Menschen. Englemann, Leipsig. Bd. 3, 1020 pp.


EcKER, A. 1851-99 Icones physiologica>. Voss, Leipzig. 48 pp.

FoNTANA, F. 1787 Abhandlung liber das Viperngift u. s. w. Ch. Himburg,

Berlin, 11+ 500 pp. FucHS, H. 1903 tJber die sogenannte "intracellular" Entstehung der roten

Blutkorperchen junger und erwachsener Siiuger. Anat. Hefte, Bd. 22,

Heft 1, pp. 95-136. FuNKE, O. 1853 Atlas of physiological chemistry. Cavendish Soc. Works,

London, 29 pp. Gage, S. H. 1888 The form and size of the red blood corpuscles of the adult

and larval lamprey eels of Cayuga Lake. Proc. Amer. Soc. of Micros copists, vol. 10, pp. 77-83. GiGLio-Tos, E. 1899 A proposito dei Cromocrateri nel sangue della Lam preda. Anat. Anz. Bd. 15, no. 16, pp. 298-300. Grawitz, E. 1899 Ueber kornige Degeneration der roten Blutkorperchen.

Deutsche med. Wochenschr. Jahrg. 25, no. 36, pp. 585-587.

1902 Klinische Pathologie des Blutes u. s. w. G. Thieme, Leipsig,

2te Auflage, 796 pp. Hamburger, J. H. 1902 Osmotischer Druck und lonenlehre in den medizin ischen Wissenschaften. Bergmann,« Wiesbaden, Bd. I. 13+ 539 pp. Heidenhain, M. 1904 Ueber die Oberflachenkrafte als Ursache der sogen.

"GeldroUenform" der roten Blutkorperchen und verwandte Erschei nungen. Folia haematol., Jahrg., no. 8, pp. 461-475. Heinz, R. 1890 Arbeiten aus dem pharmakolog. Institut der Univ. Breslau.

1. Die Wirkung konzentr. Satzlosungen. 2. Morphologische Veran derungen der roten Blutkorperchen durch Gifte. Arch. f. path.

Anat. u. Physiol, u. fur klin. Med., Bd. 22, Heft 1, pp. 100-116.

1901 tiber Blut-Degeneration und Regeneration. Beitrage z. path.

Anat. u. zur allgem. Path., Bd. 29, pp. 299-404. Henle, J. 1841 AUgemeine Anatomic. Voss, Leipsig. 23 + 1048 pp. Hewson, W. 1777 The Works of William Hewson. London (1846). 56+ 360

pp. Hober, R. 1902 Physikalische Chemie. Der Zelle und der Gewebe. Engle mann, Leipsig, 12 + 344 pp. Howell, W. H. 1900 The life-history of the formed elements of the blood,

especially the red blood corpuscles. Jour. Morph., vol. 4, no. 1, pp.

57-130. HoGAN, J. J. 1915 The intravenous use of colloidal (gelatin) solutions in

shock. Jour. Amer. Med. Assn., vol. 64, pp. 721-726. Jolly, J. 1904 Sur la forme des globules rouges ^ propos des communications

de M. Triolo. Comp. rend. soc. Biol., T. 57, no. 2, pp. 339-342.

1905 Sur la forme des globules rouges des Mammiferes. Comp. rend. soc. Biol., T. 58, pp. 481-483.

1906 a Quelques remarques k propos de la forme, de la structure et de la fixation des globules rouges des Mammiferes. Folia haematol., Jahrg. 3, no. 4, pp. 183-186.

1906 b Courte response k la note preccdente de M. Weidenreich. Folia haematol., Jahrg. 3, no. 5, p. 244.


Jolly, J. 1909 Sur quelques points de la inorphologie du sang 6tudi^s par

I'observation de la circulation dans I'aile de la chauve-souris. Arch.

d'Anat. microscopique. T. 11, pp. 9-1-109. Jordan, H. E. 1909 The shape of the red blood corpuscles. Anat. Anz., Bd.

34, no. 16 u. 17, pp. 406-412.

1915 The shape of the human red blood corpuscles. Proc. Soc. Exp.

Biol, and Med., vol. 12, no. 7, pp. 167-169. KoEPPE, H. 1899 Die Volumensanderungen rother Blutscheiben in Salzlos ungen. Arch. f. Anat. und Physiol., Physiol. Abt., Jahrg., pp.

504-517. Lackschewitz, T. 1892 tJber die wasseraufnahmefjihigkeit der roten Blut korperchen nebst einigin Analysen pathologsichen Blutes. Laak mann, Dorpat, 43 pp. Leeuwenhoek, A. 1719 Epistolae physiologicse. Beman, Delphis, 48 + 466

pp. Lewis, F. T. 1904 The shape of mammalian red blood corpuscles. Jour. Med.

Research, vol. 10 (n. s. 5), no. 4, pp. 513-517.

1913 A text-book of histology. Blakiston, Phila., 11 + 539 pp. LiTTEX, M. 1877 Ueber einige Veriinderungen rother Blutkorperchen. Berliner klin. Wochenschr. Jahrg. 14, no. 1, pp. 1-5. LoHNER, L. 1907 Ueber einige neue Beobachtungen am Blute nach Einwir kung des elekrischen Entladungsschlages. Arch. f. d. gesam. Physiol.,

Bd. 120, pp. 193-204.

1910 tJber die Glockenformen von Saugererythrocyten und ihre Ursachen. Arch. f. d. gesam. Physiol., Bd. 131, pp. 408-124.

1911 Ueber das Osmium tetroxyd als Blutfixationsmittel und die Form der Saugererythrocyten. Arch. f. d. gesam. Physiol., Bd. 140, pp. 92-108.

MAL.A.SSEZ, M. L. 1896 Sur les pretendus liquides conservateurs ou fixateurs des globules rouges, et les erreurs qu'ils peuvent causer dans les mensurations et evaluations de volume de ces elements. Compt. rend, soc. Biol., T. 48, p. 511-514.

MiNOT, C. S. 1912 Keibel and Mall, Manual of Human Embryology, vol. 2, Lippincott, Phila., 8 -|- 1032 pp.

MtJLLER, J. 1832 Beobachtungen zur Analyse der Lymphe, des Blutes und des Chylus. Annalen d. Physik u. Chemie, Bd. 25, pp. 513-590.

MuYS, W. W. 1738 De carnis musculosaj, etc. Langera, Bat., 431 pp.

1851 Musculorum artificosa fabrica etc. Bonk et Becker, Bat., 431 pp.

Ors6s, F. 1909 Ueber die Form und die FormverJinderungen der bikonkaven roten Blutkorperchen. Folia haematol., Bd. 7, Heft 1, pp- 1-18.

Prevost, J. L. ET Dumas, J. A. 1821 Examen du sang et de son action dans les divers phenomenes de la vie. Biblioth. univers. des sc. b. 1. et a. Geneve, 24 pp.

Quincke, X. 1877 Weitere Beobachtungen iiber perniziose Anamie. Deutsch. Arch. f. klin. Mediz. Bd. 20, pp. 1-31.

Radasch, H. E. 1906 Ein Beitrag zur Gestalt des roten Blutkorperchens beim Menschen. Anat. Anz., Bd. 28, no. 23, pp. 600-604.

Ranvier, L. 1875 Traite technique d'Histologie. 2d ed., Paris, 871 pp.


Rice, J. 1914 Note on tho form assumed by the red blood corpuscles of the

blood, or by the suspended particles in a lecithin emulsion. Philosoph.

Mag., vol. 28, pp. 664^670. RiNDFLEiscH, G. E. 1880 Ueber Knochenmark und Bhitbildung, II, III.

Arch. f. mikr. Anat., Bd. 17, pp. 21-42. ScHAFER, E. A. 1912 Quain's Anatomy, Vol. 2, Pt. 1; 11th ed., Longmans

Green & Co., London, 11 + 739 pp. ScHULTZ, C. H. 1836 Das System der Circulation, u. s. w. Cotta, Stuttgart

u. Tubingen, 368 pp. ScHULTZE, M. 1865 Ein heizbarer Objekttisch uhd seiner Verwendung bei

Untersuchungen des Blutes. Arch. f. mikr. Anat., Bd. 1, pp. 1-42. DE Senac, J. B. 1749 Traite de la structure du Coeur, etc. Vincent, Paris, T.

2, 504 pp. Stohr, p. 1906 Lehrbuch der Histology und der mikroscopischen Anatomie des

Menschen. 12th ed., Fischer, Jena, 15 + 464 pp. Triolo, M. 1904 a Nouvelle recherches experimentales sur la morphologic des

elements figures du sang. Comp. rend. soc. Biol., T. 57, no. 145, pp.


1904 b Examen du Sang humain in vitro par le methode de la 'lubrification' (Methode a I'huile de vaseline). Comp. rend. soc. biol., T. 57, pp. 307-309.

1905 Nuove ricerche sperimentali sulla morphologia degli elementi figurati del sangue. Gazzetta d. Osped. Milano., vol. 26, pp. 15341536.

Wagner, R. 1833 Zur vergleichenden Physiologie des Blutes. Voss, Leipzig,

8+80 pp.

1838 Nachtrage zur vergleichenden Physiologie des Blutes. Weber, E. et Souchard 1880 De la disposition en piles qu'afifectent les globules

rouges du sang. Arch, de Physiol., Annee 1880, pp. 521-531. Weidenreich, F. 1902 Studien liber das Blut und die blutbildenen und -zer storenden Organe. Arch. f. mikr. Anat. Bd. 6.1, pp. 459-507.

1903 Die roten Blutkorperchen. Ergeb. d. Anat. und Entwickl.,

Bd. 13, pp. 1-94.

1905 a Ueber die Form der Sougererythrocj-ten und die formbestim menden Ursachen. Folia haematol., Jahrg. 2, no. 2, pp. 95-104.

1905 b Einige Bemerkungen iiber die roten Blutkorperchen. Anat. Anz., Bd. 27, no. 24, pp. 583-596.

1906 a Eine neue einfache Methode zur Darstellung vom Blut — Trockenpriiparaten. Folia haematol., Jahrg. 3, no. 1, pp. 1-7.

1906 b Studien iiber das Blut und die blutbildenen und -zerstorenden Organe. IV. Weitere Mitteilungen liber rote Blutkorperchen. Arch, f. mikr. Aftat., Bd. 69, pp. 389-438.

1906 c Einige Bemerkungen zu dem Aufsatze J. Jolly's iiber die Form, Strukture, und Fixation der roten Blutkorperchen der Saugethiere. Folia haematol., Jahrg. 3, no. 5, pp. 241-244. 1910 Uber die Form der Saugererythrocyten. Eine Erwiderung an Dr. L. Lchner. Arch. f. d. gesam. Physiol., Bd. 132, pp. 143-146. Young, T. . 1813 An Introduction to Medical Literature, etc.. Philips, London, 26 + 659 pp.

author's abstract of this paper issued by the bibliographic service september 28.



From the De partment of Histology and Embryology, Cornell University, Ithaca,

Neio York



It is generally accepted that there is a close relationship between the testis of male vertebrates and the development of the genital tract and of the secondary sexual characters. There is a further tendency to regard the testis as one of the endocrine organs which elaborates some internal secretion which governs the above mentioned growth processes and which plays some role in connection with the sexual instincts. Naturally many workers along this line have attempted to ascertain what tissue in the testis is responsible for the production of the 'autacoid,' if such there be. The extensive and comparatively recent reviews of the literature by Biedl (Innere Sekretion, II Aufl., '13) and by Tandler and Grosz (Die biologischen Grundlagen der sekundaren Geschlechtscharaktere, Berlin, '13), as well as researches reported since thesfe works appeared, seem very convincing that the interstitial cells (cells of Leydig), and not the germinal cells nor Sertoli cells, are the producers of the internal secretion. Hence the term 'interstitial gland of the testis,' first used by Bouin and Ancel in 1903, occurs quite frequently in the literature of the present day. It must be admitted, however, that the Sertoli cells, especially, have not been adequatelj ruled out as a possible factor in this endosecretory function.



Adding still more interest to this subject is the serious consideration which it is receiving by clinicians.' While during the last few years, surgeons have performed a number of grafting operations on men with some success, a rather unexpected result has been reported by Morris ('16), who found that testicular grafts caused an undeveloped testis to enlarge and become apparently normal. This is mentioned here only because it indicates the practical importance of advancing our knowledge concerning the endocrine function of the testis.

Since all the evidence in favor of the secretory function of the interstitial cells of the testis, as well as of the ovary, is indirect, there are those who hesitate (c.f. Kingsbury, '14, W. Blair Bell, '16) in ascribing to these cells the function of producing a specific substance of 'hormonic' or 'chalonic' action in the organism. Bell is especially emphatic as may be seen from the following quotation (p. 145) :

Again, it is extremely interesting to note how erroneous has been the view, generally held, that the interstitial cells of the ovary and testis are responsible for the secondary sex characteristics. For many years I have contended that the gonads play but a subservient part; and this is emphatically demonstrated .... by the fact that in the testes of tubular partial hermaphrodites with feminine secondary characteristics, . . . . , the interstitial cells are always developed to a remarkable extent to a degree which is rarely seen even in the undescended testis and never in the normal .testis. These cells cannot, therefore, be responsible for the secondary characteristics.

The evidence supporting the secretory function of the interstitial cells has been derived from various sources, the most important of which are: the general histological character of the interstitial cells (epithelioid) ; various pathological conditions of the testis and abnormal sex development, such as the various forms of hermaphrodism and cryptochism, with the corresponding histological character of the tissues of the testis; effects of castration, vasectomy and transplantation in man and laboratory animals; exposure of the testis to the influence of X-rays; injection of testicular extracts; degree of de ^ See, for example, On abnormalities of the endocrine function of the gonads in the male, Lewellys F. Barker, Amer. Journ. Med. Sci., 1915, vol. 149, p. 1.


velopment of the interstitial cells at different periods of life and the periodic changes which they undergo in adult animals as related to the sexual cycle. This last line of evidence is somewhat conflicting. It is, of course, difficult in most of the higher animals, epecially domesticated ones, to correlate the various phases of the spermatogenic cycle with changes in the interstitial cells because the various progressive and regressive changes in the tubuli contorti are going on more or less side by side at the same time. For this reason Tandler and Grosz ('11) selected as material for a study of this question an animal (mole) in which the interval between rutting periods was sufficiently great to separate the stages in the spermatogenic cycle, since this animal is sexually active only in the spring. The woodchuck, which is abundant in many parts of the United States and Canada and even Alaska (Howell, '15), similarly should furnish facts of interest in this connection. This animal also is sexually active only in the spring. The female gives birth to but one litter a year, the young being born about the last part of April or the first of IVIay, according to Merriam('84) . These dates agree well with the general life history of these animals as observed in this vicinit}^ Merriam further states that along the western border of the Adirondacks they go into hibernation late in September and remain till the middle or last of March. In this region in their normal habitat they do not retire till nearly a month later, and those kept in captivity usually remain awake till the last of November. Probably those that retire in natural burrows do not become actually dormant till several weeks afterwards.

Incidentally the study reported here is of interest also in connection with the subject of histological changes during inani-' tion — a subject having many important bearings upo'n growth,^ metabolism and physiological adaptation. All species of American marmots hibernate profoundly. They store up no food, ex 2 This point is especially discussed by Sergius Morgulis, Arch. f. Entw. d. Organ., 1911, vol. 32, p. 169. Here is also reviewed the literature on the effect of experimental inanition on histological changes in the testis. This deals, however, with spermatogenesis and liot with the interstitial cells. The data reported in this connection does not show very uniform results.


cept within their own bodies, and hence are deprived of food, in the ordinary sense, for about four months out of each year. This may thus be considered a long period of physiological inanition. During much of this time they are very dormant and have the usual low body temperature (a few degrees above 0°C.), slow circulation and respiration, etc., characteristic of hibernation.


While the interstitial cells were discovered in 1850 by Leydig, the first report on changes in these cells either in connection with the seasons of the year or with the sexual cycle of the adult, did not appear till many years later when Hansemann ('95) reported that he had observed the testis of the marmot and found that the testis of the hibernating animal, in which there is no spermatogenesis, contains practically no interstitial cells, there being only a few spindle-shaped cells between the tubules. After the animal has been awake for two months, however, and spermatogenesis is going on, the interstitial cells are very numerous, so much so that they give the appearance of a sarcoma. He considered that these cells probably constitute an organ with some specific function.

Friedmann ('98) followed with an extensive study of the more or less parallel development of the interstitial tissue and the progress of the spermatogenic cycle in frogs (Rana fusca, Rana viridis, Hyla arborea) and the toad (Bufo vulgaris). In frogs he found an increase in the interstitial cells during the progress of spermatogenesis as autumn approaches. Beginning with the end of October with the cessation of spermatogenesis, the interstitial cells almost disappear and remain minimal till about' May. An important point mentioned by Friedmann in connection with Rana viridis is the observation that in the same testis there may be a difference in the amount of interstitial tissue. Where the spermatogenic process is most active, there the interstitial cells are most developed. In the tree frog (Hyla arborea) the interstitial cells seem to be about a month behind in development in comparison with the brown and the green frog.


At the end of May the same condition prevails in the toad as in the frog, there being only a few interstitial cells and mostly spermatogonia in the tubules. As spermatogenesis advances, there is a more or less parallel growth in the interstitial cells. Active spermatogenesis continues on into the winter. Free spermatozoa are most numerous at the end of April at which time the interstitial cells are maximal and loaded with fat, but contain practicallj^ no pigment, as compared with the numerous pigment granules present in the interstitial cells later in summer. He also observed that the first fat .to appean was not interstitial but intratubular.

Ganfini ('03) found that in the hibernating marmot the interstitial cells are not fewer in number than during the active period, as was reported by Hansemann, but are only smaller in size and different in structure. They stain less readily and as a whole give the appearance of a structure which has ceased secreting. During winter-sleep they also assume a rounder form. He does not think these changes have anything to do with spermatogenesis, but are due rather to the same causes that arrest the processes going on in the other organs. In this animal he describes the interstitial cells as being arranged in lobes and cords bounded by endothelium, but some are also found isolated.

Regaud ('04) reported that spermatogenesis goes on in the mole (Talpa europaea) during autumn and winter. In December the tubules occupy most of the testis. The tubules, although latge, are separated from each other by wide spaces containing only a few interstitial cells. By February the testis has become more than 15 mm. in length. In June and July the interstitial cells are voluminous, closely packed together and occupy more space than the tubules. The cytoplasm of the interstitial cells at this stage is greatly vacuolated. By July the testis has decreased to only 3 or 4 mm. in length ; but the interstitial cells still persist, giving the adult organ the appearance of a foetal testis. Spermatogenesis has ceased and in the tubules there is only a syncytium of Sertoli cells and a few spermatogonia. Thus he considers these observations to be just the opposite


of those reported by Hansemann. Although the seasonal changes were not followed out any farther, he concluded that the interstitial cells do not degenerate parallel with the germinal epithelium during the retrogressive changes in the spermatogenic cycle.

It is to be noted that in the mole the testes are abdominal and situated beside the bladder in December; but with the increase in size which occurs from January on and culminates in March, the testes come to occupy two pouches beside the root of the tail. Periodic changes in the size of testis are well known in many species, having been observed as far back as the days of Aristotle (Marshall, '11). The tendency for the testes to enlarge and also to descend into a sessile scrotum during rut in most rodents was mentioned by Owen ('68) fifty years ago. The general subject of the descent of the testis cannot be discussed here. The excellent papers by Hart ('09) and Frankl ('00) present this subject most admirably.

Champy ('08) reported that in Rana esculenta spermatogenesis is at its highest in July and at this time the interstitial cells are at a minimum. In the autumn there is a great increase in the interstitial cells and spermatogenesis is at its lowest. This observation seems to be rather an exception to what Friedmann and later Mazzetti report in other species of frogs. However, in view of a lack of details, there may not be as much disagreement as the bare statement above would indicate at first sight, the breeding season being about two months later.

Lecaillon ('09) in general confirms the observations of Regaud on the mole both in regard to the change in size of the testis and the relation of interstitial cell development to spermatogenesis. He claims, however, that in July there is much degeneration in the interstitial cells and that this is responsible for a large part of the decrease in the size of the testis at this time ; but some of the interstitial cells persist throughout the entire year.

Mazzetti ('11) in working with frogs (Rana fusca and Rana viridis) found the seasonal changes in the interstitial cells to be essentially as had already been described by Friedmann. He Incideiitally states that interstitial cells are extraordinarily abundant in hibernating snakjes but not in the 'ghiro.'


IMarshall ('11), from a study of fourteen hedgehogs, found no spermatogenesis in this animal during winter, at which time the sexual organs are small. Beginning about the end of March, shorth^ after the close of the hibernating period and at the approach of the rutting season, these organs enlarge, reaching complete development in May. Spermatogenesis is going on simultaneously and by the end of April free sperms are recognizable. The testis does not, however, enlarge to the same extent as in the mole. While the increase in the size of the testis is due in part to an increase in the spermatogenic tissue, a greater factor in the growth of this organ is the proliferation of interstitial cells, which leaves the tubules widely separated £rom each other, especially in the central part of the testis. Large blood vessels apparetnly also develop in this interstitial tissue. This condition persists till October, after which retrogression sets in, the interstitial cells largely disappear and with them the blood vessels just mentioned. The tubules are thus brought almost into contact and spermatogenesis is in abeyance. There is little or no change till after hibernation when the next rutting season begins. The female produces one litter in May or June and another again in August or September. During the period from about April till as late as October, the testes are descended into sac-like continuations of the abdominal cavity in the neighborhood of the perineum, where they may be detected from the exterior. The author concludes that the simultaneous growth in the accessory generative organs, especially noticeable in the case of the seminal vesicles, is probably due to an internal secretion elaborated by the interstitial cells during their period of increase.

The most detailed description so far encountered in the literature on this subject is that given by Tandler and Grosz ('11) who examined the testes of moles (Talpa europaea) sacrificed at various times during more than two years so that every month was represented in the series. They concluded that rutting goes on during the month of March at which time the testis has enlarged to about three times the usual diameter and the tubules contain active spermatozoa. Similar changes occur


in the epididymis and other genitals. The increase in the size of the testis is due to development of the generative part of the organ. Interstitial cells are present only as small islands. The cytoplasm of the individual cell is homogeneous. At the end of March when most of the sperms are gone, the interstitial cells are still only minimal in quantity. As summer approaches the whole testis and the tubules decrease, reaching their lowest in July. The interstitial cells, however, increase till their individual boundaries are nearly obliterated. The cytoplasm stains more brilliantly with eosin. The nucleus, on the other hand, remains about the same size; but the chromatin is less evident. By July the tubules are separated from each other by masses of interstitial cells whose cytoplasm is greatly vacuolated and whos^ nuclei have lost their definite nuclear structure. From the end of September and on, spermatogenesis advances again and the interstitial cells slowly decrease so that by February, when spermatozoa are nearly free, they are limited to a few scattered individual cells. Thus when spermatogenesis is at its highest, the interstitial cells are at their lowest, and vice versa. This latter condition the authors compare to the foetal type as did Regaud. They think that the increased interstitial cell growth which occurs at the low ebb of spermatogenesis, has something to do with the coming spermatogenic cycle.

A few incidental statements made by Gushing and Goetsch ('15) may be referred to here, since the observations concern the woodchuck. The authors examined the testis of a woodchuck killed March 22, having been captured a few days previously (March 17, 1913). Judging from the weight of the animal (2360 grams), the testis (1 cm. by 1.75 cm.) was enlarged as compared with the testis during the autumn and winter. Histologically they found no spermatozoa but a few spermatids and an abundance of interstitial cells. In another animal killed January 12, 1914 — an animal which they were not positive had hibernated — the testis showed active stages of spermatogenesis but no spermatozoa. The interstitial cells were abundant, but no statement is made as to how the interstitial cells compared with those of the first animal examined.


The great variability in the correspondence between the progressive stages in the sexual cycle and increased interstitial cell growth evidently calls for more observations if this line of evidence is to be utilized in the interpretation of the internal secretory function of the interstitial cells of Leydig.


This report is based upon a study of the testes of thirty-five male woodchucks which have been killed at various times during the past four years. The series includes twenty-three adult animals, one or more years old, some of which have been sacrificed in each month of the year; six younger animals (several months to nearly one year old), some of which were sexually mature when killed, and six animals which were from about five weeks to several months old.

Most of the testes were removed before death while the animal was under an anaesthetic (usually ether). Some were removed immediately after death and, in a few cases, several hours after the animal was shot. In the early spring at the critical period of the year when the rutting season commences and when it is difficult to capture the animal alive, a number were shot and brought to the laboratory and the testes taken out. Such material was used as control for that taken from animals kept in captivity in artificial burrows. While both males and females were always together in captivity and while the artificial burrows, as described elsewhere,^ are such that practically normal conditions should prevail, it was found that females kept as late as April were not pregnant. Since they did not breed in captivity, normal controls taken directly from their usual habitat were necessary. Such controls showed, however, that the male sexual cycle was not interrupted, at least as far as the histological picture and the descent and growth of the testis are concerned.

Except in the first few cases, the gross weight of the animal, the weight of the gastro-intestinal contents and urine and the

^ Rasmussen, A. T. 1915. Amer. Journ. Physiol., vol. 39, p. 23.


weight of each testis, were obtained. From this data the weight of the testis as per cent of the reduced body weight (gross weight minus gastro-intestinal contents and urine) was calculated. The position of the testis was also recorded. In every case the weight of each testis of the same animal was practically the same.

From about one-fourth to one-eighth of one testis (depending upon its size) was fixed for eight hours in Zenker's fluid with the acetic acid reduced to only four drops per 100 cc. A few whole testes were also fixed in this solution and sectioned longitudinally in order to see if the interstitial cells were equally distributed in the various regions of the testis at the different seasons of the year. After washing in running water for an hour, such issue was placed in 2 per cent potassium dichromate for four and one-half days. After a second washing of two hours in running water, it was dehydrated in the usual grades of alcohol containing iodine, two hours with several changes being allowed for each grade till 98 per cent was reached. Here only one hour was allowed. From 98 per cent alcohol the tissue was transferred to chloroform for one hour with several changes and then to chloroform-paraffin for one hour, and finally to paraffin (melting at 54°C.) for two hours or longer. The aim of this rapid embedding was to preserve the lipoids. This technique as well as the subsequent Weigert staining was recommended by Kingsbury who has found it very useful in demonstrating lipoids (Kingsbury, '11). The fixer was not sufficiently washed out by this method so that it was necessary to leave the sections, after having been fixed on the slide, several hours in the lower grades of alcohol containing iodine before they were free from precipitates.

This material was stained in ordinary hematoxylin and eosin, iron hematoxylin, acid fuchsin and methyl green according to the technique employed by Bensley ('11), but especially with copper hematoxylin (Weigert's), the older technique of differentiating in the potassium ferricyanide and borax mixture (diluted 3 to 10 times) being employed.


Another part of a testis was fixed in Carnoy's fluid (6:3: 1) and stained with iron hematoxylin and also with Mayer's haemalum and eosin.

A third and smaller section was fixed in Meves' ('08) modification of Benda's fixer for four and one-half days both with and without the subsequent pyrohgneous-chromic acid and potassium dichromate mordantage. This material was also dehydrated, cleared and embedded rapidly according to the schedule described •for the modified Zenker's material. It was stained for mitochondria with sodium alizarinsulphonate and crystal violet according to the technique employed by Meves and Duesberg ('08) and also as used by Wildman ('13). Sections were also mounted unstained in balsam without any cover glass to demonstrate the fatty globules which develop in the peripheral cytoplasm of the interstitial cells and which are blackened by the osmic acid of the fixer.

All sections were cut 4 /x and 6 m thick. Those cut 4 ^i thick were used almost entirely and comparable figures in the accompanying plates are from such sections.

On account of the large size of the testis, especially at certain seasons of the year, and the soft consistency of the structures within the tough tunica albuginea, it is difficult to get small pieces, even with the sharpest razor, for fixers that penetrate poorly, without disturbing the relationship of the interstitial tissue to the tubules. As a consequence the osmic acid stained preparations used to demonstrate the fatty globules in the peripheral cytoplasm of the interstitial cells, had to be taken near the cut surface of the block where sucb disturbance had occurred to a greater or less degree.

There is a strong tendency for the interstitial tissue to draw away from the tubules in material fixed in both Carnoy's and Zenker's fluid. This is especially true at certain seasons when the interstitial tissue is very loose. The heat of the paraffin bath may also have contributed to these artifacts. These defects, however, serve to bring out even more clearly the relative interstitial cell development at the various seasons and do not invalidate the cytological details.




In late August, September and October the testes of the adult woodchuck are minimal in size, being only 0.015 per cent to 0.020 per cent of the reduced body weight. They are abdominal, occupying a variable position in the dorsal portion of the abdominal cavity on a level usually with the upper two sacral vertebrae. They are a chocolate brown in color, due to a dark yellow pigment found in the interstitial cells. The relative amount of interstitial tissue may be judged from figure 1, which, however," is taken from an animal killed much later in the year. This tissue is so loose that if an entire testis is placed in Carnoy's fluid the organ collapses from the absorption of the lymph faster than the fixer enters the dense connective tissue tunica. It consists of a fine connective tissue framework in which, besides the usual fixed connective tissue nuclei, blood and lymph vessels, etc., there are found a number of interstitial cells of Leydig varying in size from small cells about 5 ix in diameter, with practically no cytoplasm and a more or less vesicular nucleus, to larger cells about 10 ix in average diameter, with a fair quantity of cytoplasm usually containing pigment granules, which are slightly more numerous early in this period than later. The nuclei of some of the smaller cells may be somewhat irregular or slightly spindle shaped and resemble the fixed connective tissue nuclei. Indeed it is practically certain from the investigations upon the origin of these cells that they are of connective tissue origin, and it is not improbable that they may return to this type again. The nucleus, however, of most of these interstitial cells has the oft-described spherical and vesicular appearance containing one nucleolus and a chromatin network which' is especially coarse next to the nuclear membrane. The cytoplasm varies greatly in amount, although it is always scanty at this stage and gives a variety of irregular shapes to the cells, as may be seen from figure 2 and the smaller cells in figure 3.

In nearly all these cells there are a variable number of the small brown or dark yellow pigment granules. These, while undoubtedly derived from the very labile lipoid or fatty glob


ules which once developed in the peripheral cytoplasm, as will be indicated later, are preserved even in Carnoy's fluid and after such fixation are stained black with iron hematoxyhn. This pigment is darkened with osmic acid, but is not stained with ordinary hematoxylin or eosin. The granules are usually aggregated at one side of the nucleus, which is then somewhat eccentric, as may be seen in figure 2.

In the material fixed in the modified Zenker's fluid and stained with copper hematoxylin, there appear to be a few other granules in the cytoplasm. These vary greatly in size from very small ones about the size of mitochondria to larger ones of the size of the pigment granules. From their staining reaction and the fact that they are not always preserved by the above method but show as small vacuoles in the dense cytoplasm at certain stages, and especially so after Carnoy's fixer, they undoubtedly contain a lipoid moiety. The only thing of a mitochondrial character that could be demonstrated in the interstitial cells at any period were these fine granules which were most clearly shown with copper hematoxylin. However, this point needs further investigation with different and more carefully conducted technique, as a particular tissue may be refractory to a particular technique even in the hands of experienced workers (c.f., e.g., Duesberg, '17). Whitehead ('04, '05, '08) has described granules in the cytoplasm of these cells in the testis of the pig and other animals, which he finally concludes ('12) are a combination of protein and fatty material and somewhat of the nature of mitochondria though differing from the latter in certain staining reactions and in size, generally being larger. He questions the presence of mitochondria as specific granules in the interstitial cells of the testis. Granules, probably of a similar nature, are mentioned by other authors (Regaud, '01, Hanes, '11). The correspondence or identity of these granules with the ones described here in the woodchuck can not be stated at this time, since this would involve special work of a microchemicai nature, which was not anticipated when this work was commenced.


A few of the larger cells contain nearer the periphery of the cell other fatty globules which are blackened by osmic acid, and which are very much more soluble, disappearing even more readily than the fat of ordinary adipose cells. These are the fatty granules so often found in these cells in animals generally.

In addition to the above described interstitial cells, which for convenience will be spoken of hereafter as the ordinary type of interstitial cells, there are scattered here and there, sometimes singly and sometimes in small groups, in the interstitial tissue and quite frequently adhering to the basement membrane of the tubules, a second type of Leydig cells. These are much fewer in number but larger in size and contain large pigmented granules to such an extent that the nucleus may be crowded into an irregular space between the granules or to one side of the cell. The nucleus is frequently irregular in shape, in conformity to cytoplasmic pressure, and occasionally gives the appearance of being in the process of degeneration. A group of such cells is shown in figure 3 in connection with a few interstitial cells of the ordinary type, which, as described above, also contain pigment but in much smaller quantities and as smaller granules. Occasionally the pigment granules in these large cells, which may be spoken of as pigment cells, are nearly as large as the nucleus (fig. 5). The presence of so many large pigment granules may make the cell outline very irregular, as may be seen in the upper left hand corner of figure 2 where two cells of this kind are shown with the pigment granules stained black thus obscuring the nucleus. These larger pigment granules are evidently of the same composition as the smaller ones, staining black with copper hematoxylin and iron hematoxylin, at least on the surface, darkened with osmic acid, also mostly on the surface, but not stained with ordinary hematoxylin or eosin. Acid fuchsin may stain some of them red.

Whitehead ('08 a) in reporting on the cryptorchid testis of a horse, describes cells of a similar character filled with a lipochrome which had pushed to one side the nucleus, which was small and pyknotic. He concluded that since there was a


normal scrotal testis in this particular case, the abdominal one was fmictionless as far as internal secretion is concerned, and hence the interstitial cells of the cryptorchid testis had undergone a pigmentary degeneration. As will be disclosed later, the origin of these cells in the woodchuck is undoubtedly of this order, having resulted from the degeneration of some of the ordinarj^ interstitial cells when they undergo retrogression in mid-summer, at which time the very soluble fat of the peripheral cytoplasm undergoes a radical change such that from these fatty globules there is evolved a pigmented substance much less soluble. Sehrt ('04) considered the pigment usually found in the interstitial cells as a lipochrome. These pigment cells as well as the ordinary interstitial cells are distributed at all stages about equally in all parts of the testis.

At this stage there are also a number of ordinary adipose cells scattered through the interstitial tissue. Small fatty granules are also found in the tubules, in fact at no time are the tubules free from demonstrable fat which blackens with osmic acid (figs. 14, 22, and 26).

Within the tubules early stages of spermatogenesis are in progress. The lumen is filling up with spermatocytes, which are enlarging. No karyokinetic figures appear in them, however, till about October.

The interstitial cell picture undergoes but little change till after hibernation, or until early in March in these particular animals. There is a slight gradual decrease in pigmentation due to a disappearance of some of the pigment granules in the ordinary interstitial cells and probably a slight decrease in the number of pigment cells, which usually become smaller and somewhat more irregular. The weight of the testis by November or December has increased to about 0.034 per cent of the reduced body weight. This increase is evidently due to a filling up of the tubules with spermatocytes, for as seen in figure 1, which especially represents this stage, the tubules are gorged with spermatocytes showing open karyokinetic figures.

There is no sudden change in the testis with the onset of hibernation as might be expected if the testis is an organ of


internal secretion, since at this time the bodily functions are greatly reduced and metabolism profoundly modified. However, since the internal secretion of the testis has to do with the reproductive side primarily, and not with the vegetative — phases which may be more or less independent — it may not necessarily follow that any marked observable change need be registered in this organ at the beginning of or during this dormant state. Certainly the interstitial cells in the woodchuck remain practically unchanged during the hibernating period — ■ December, January, February. In the ground squirrel (Citellus tridecemlineatus) which also hibernates, Mann ('16) reports that the testis, while undergoing definite seasonal variations, does not show any specific change due to the torpid condition.

At the beginning of March when the animal begins to awaken, the tubules are -ready for a sudden and rapid production of spermatids and spermatozoa. The woodchuck may be semidormant during this waking up period for several days and even longer. During this sluggish period the changes in the testis commence, so that before the animal has attained what may be termed its homoiothermal temperature, changes in the testis have already occurred. Figures 6 and 7 show these beginning changes in an animal that is just waking up. In the peripheral cytoplasm of the ordinary interstitial cells large fatty globules, which may be blackened with osmic acid, make their appearance and the cells begin to round out. The testis now represents from 0.040 per cent to 0.050 per cent of the reduced body weight. However, the animal has lost about one-third of its body weight during the preceding months of inanition, consequently this relative weight of the testis is much exaggerated.

In the newly awakened and active animal the testis increases very rapidly. The interstitial spaces become crowded with enlarging interstitial cells of the ordinary type as is seen in figure 11. In cells which have been fixed in Carnoy's fluid, the cytoplasm is greatly vacuolated (fig. 12) due to the fixer having dissolved the fat. Figure 14 gives an idea of the relative amount of fatty material at this stage. The fat is here blackened with osmic acid.


The denser cytoplasm around and to one side of the nucleus expands, spreading out the pigment granules (fig. 12). Various stages of this expansion are well shown in figure 13. This figure in comparison with figure 12 shows also that there is in this newly formed dense cytoplasm a number of the fine non-pigment granules, characteristically associated with the central cytoplasm (endoplasm), as mentioned above.

During this interstitial cell development there is no direct evidence of mitotic or of amitotic cell division. Whitehead ('04) in describing the rapid growth in the size of these cells, as it occurs in pigs from 20 to 28 cm. in size, remarks that there is no evidence of cell division after the 7 cm. stage. This early cessation of signs of division in the interstitial cells of the testis seems to have drawn the attention of many investigators. Allen ('04) found, for example, no evidence of cell division after the 7.5 cm. stage in the pig, nor after eight days after birth of the rabbit. Plato ('96), Finotti ('97) and Kasai ('08) comment on the absence of mitosis in the interstitial tissue of the > human testis. Kasai in 130 human testes, representing the wide range from the four months foetus to eighty-four years, saw only one mitotic figure in the interstitial cells. However, in the woodchuck at this stage of rapid interstitial cell growth, such , stages as are shown in figure 8 are not uncommon. Later when the cells reach their maximal size, one occasionally finds cells evidently containing two nuclei as shown in figure 9. In 4 out of 8 cases where the cells had reached their largest size, or nearly so, the cells were arranged more or less in groups within what appears to be a common membrane, such as is seen in figure 10. This would suggest that there is cell division and that each group of cells represents the daughter cells of a single parent cell. Von Bardeleben ('97), while not seeing any mitotic figures in the interstitial cells of executed criminals, frequently saw evidences of direct cell division. Von Hansemann ('95),. Reinke ('96), von Lenhossek ('97) and Pick ('05), however^ have reported mitotic figures in human materials, including adult.


A comparison of the number of nuclei appearing in a cross section of the entire testis before enlargement sets in, with nearly twice the number of nuclei seen when these cells are at their maximal development (i.e., when the testis is fully twice its former diameter) indicates that there is an increase in the number of interstitial cells at this time in the adult woodchuck. It is necessary to make the comparison in this way since the testis, having doubled in diameter, will during the highly developed stage give twice as many sections as when small. If the cells have not changed in number, only half as many nuclei will appear in a section of a given thickness in case of the large testis as in the case of the small one, provided the nucleus has not also changed in size, since the same number of nuclei would in the enlarged testis be distributed in twice as many sections. But in reality the nucleus has increased about 1.5 ix, or 30 per cent, in diameter. While this is much less in proportion to the increase (more than 100 per cent) in the diameter of the whole testis, it is sufficient to call for some allowance. By counting the number of nuclei in numerous groups of interstitial cells resulting from the arrangement of the tubules in the reduced testis, and comparing this with the number of nuclei found in the same number of groups in the enlarged testis, it appears that there are at least as many nuclei in a whole crosssection of the hypertrophied testis as in a cross-section of the testis before growth takes place. Making due allowance for the increase in the size of the nucleus, the indications are that there is a distinct increase in the interstitial cells of the hypertrophied testis. The assumption is that since there are as many tubules intersected in the enlarged testis as in the small one (as will be shown later in this paper), ther^- will be as many cell groups confined between them.

Further evidence also appears from the number of nuclei seen in a section of the testis that has just undergone retrogression such as will be described shortly. In such a testis the number of nuclei appears to be distinctly greater than obtains in the testis just before it enlarges (figs. 23 and 24). A larger number of animal at each stage with an actual count of the number


of cells from serial sections of the testes at th? various periods of the year would, of course, be necessary to rule out individual variations and to definitely prove that there is an increase in the number of cells and what that increase in number amounts to. The facts cited above, however, make it very probable that in the adult woodchuck there is a considerable increase also in the number of interstitial cells after waking up from hibernation although there is no direct evidence of either mitosis of or amitosis.

The pigment cells do not undergo any growth or increase in number. On the contrary they appear very inert and gradually decrease in prominence.

Going hand in hand with this interstitial cell hypertrophy, there is still further increase in the size of the testis and a renewed activity in the tubules. The spermatocytes rapidly change to spermatids and free spermatozoa are seen by the last of March. The most active stage (when most spermatozoa appear to be set free) in the spermatogenic process is reached early in April, by which time the testes have descended into sessile scrotal pouches beside the penis. Thus the renewed activity in the testis anticipated by Gushing and Goetsch ('15) actually takes place; but their assumption that this might be attributed to the influence of the functionally reactivated pars anterior of the pituitary body, does not necessarily follow, since there is nothing to show that the testes or other organs of the body — all of which show this renewed activity upon awakening of the animal from hibernation — are influenced through the pituitary rather than that the pituitary in common with the other organs is influenced by the factors responsible for the general awakening; that is, the pituitary and the testes may have been influenced by the same factors rather than the latter by the former. Furthermore, it is not even certain that the pituitary does always undergo the change reported by these' authors and by Gemeli, for Mann ('16) found that in the thirteen-lined groundsquirrel such changes while occurring in some animals did not in others, although the testis underwent a seasonal change. Jackson ('17) thinks it highly probable that


these changes described in the hypophysis during hibernation are simply the effects of the chronic inanition involved since he finds similar changes in the hypophysis of the albino rat subjected to inanition and ref ceding.

The interstitial cells do not reach their maximum development until the last of April when spermatogenesis is at its lowest. Although many free spermatozoa may remain in the tubules as late as this, the epithelium has been reduced to a single layer of cells and thus a wide empty lumen results. The new spermatogenic cycle may be considered to date from the last of April or early May, since spermatogonia begin to increase from this time on.

The interstitial cells remain at their height of development until as late as July, or for at least two months after the end of the corresponding spermatogenic cycle. These dates will, of course, vary somewhat from year to year.

During this time when the interstitial cells are enormously enlarged, from about April to June — a period during which the testes usually are scrotal — the testis represents from 0.078 per cent to 0.132 per cent of the reduced body weight. The tubules are forced far apart as will be seen in figure 15, or still better in figure 19, which is at a lower magnification and is intended to show an especially large compact node of highly vacuolated interstitial cells at the point marked with a +. In the center of this mass the boundaries of the individual cells are not evident and so gives the appearance of numerous nuclei entangled in an open network. Such nodules were found in two of the eight cases representing this stage. Several smaller areas of this sort may be encountered in a single cross section. The vacuoles are filled in life with fatty globules.

As mentioned above, half of the woodchucks killed during this stage showed the 'nest' arrangement of many and in one Case practically all of the interstitial cells as seen in figure 10. This grouping was first described by Nussbaum ('80) as the typical arrangement. Each group of cells seems to be surrounded by an epithehal sheath of flat cells and the individual cell boundaries are very indistinct. Ganfini ('03) states that


in the European marmot many of the cells of Leydig are arranged in ceil masses and in cords thus bounded by 'endothelium'; but some are also found isolated. Such an arrangement as shown in figure 10 may or may not be present in the case of the woodchuck. The columnar arrangement described by Ganfini ('02) is not found here, though some of the cell groups in the cases just described are somewhat elongated. Whitehead ('08) does not find this group arrangement typical. In the cat he found columns of cells among the tubuli recti.

Due to mutual compression, the interstitial cells at this stage vary in shape, with an average diameter in general of 20 ^ to 25 11. A few cells may be found that are as small as 14 /x in average diameter. The nucleus is more eccentric and slightly larger, being now 6 ^ to 7 ^ in diameter as compared with about 5 M when the cells are minimal in size. The pigment granules in these ordinary interstitial cells ' have decreased greatly and only now and then is a granule found as may be seen from figure 16. This recalls the findings of Friedmann ('98) who noted that when the interstitial cells of the toad contained much fatty material the pigment was greatly reduced and when the fat disappeared, the pigment was abundant again.

The dense central mass of cytoplasm is conspicuous and contains a vast number of the fine lipoid granules, some of which are very small (fig. 17). As stated above, these granules are not preserved in Carnoy's fluid and hence 'appear as small vacuoles in the dense central cytoplasm in figure 16 (seen best in the cell marked with a +)• The peripheral cytoplasm is densely packed with the large fatty granules as indicated in figure 18, especially in the insert where three cells are photographed under higher magnification.

A few of the pigment cells are seen here and there. Some are very much reduced in size and are very irregular in outline, but others are well preserved. In one case the testis was practically free from pigment either as sma41 granules in the ordinary interstitial cells or as larger spherules filling pigment cells. As a reslilt the testis at this stage is much lighter in color than at any other time.


In mid-summer the testis again decreases. Figures 20, 21 and 22 show the conditions during the earUer stages of the retrogression. In the particular case from which these figures were taken, the testis represented only 0.038 per cent of the reduced body weight, much of the fat already having been absorbed, as is evident from figure 22 which is to be compared directly .with figure 18. The fat which is left is found as rather large globules only a few of which are found in one cell as compared with the numerous globules present earlier. During this atrophy the interstitial cells are undergoing profound modifications. Pigment granules are increasing due to a chemical transformation of some of the fatty material into a pigmented compound. A number of the ordinary interstitial cells do not decrease much in size but all the fatty globules within them become changed to this pigment.

Spermatogenesis is steadily advancing. Spermatocytes are increasing in number and are filling up the lumen of the tubules. The testis becomes abdominal.

By August the testis is minimal, being only 0.015 per cent of the reduced body weight. The tubules are much closer together (fig. 23). Most of the interstitial cells are reduced to Httle more than the nuclei, which have also become smaller, many being under 5 ^ in diameter. Some of the smaller nuclei tend to stain more solidly, due undoubtedly, as suggested by Whitehead ('08), to a lack of decolorization. Plato ('96) and Ganfini ('02) observed that the nuclei of the nonvacuolated cells appear to stain more intensely than do those of the vacuolated ones. Surrounding one-half of the nucleus there is in the scanty cytoplasm a dense cap of pigment granules as shown in figure 24. If there are any of the finer lipoid granules, such as occupied the dense central mass of cytoplasm before the atrophy occurred, they are masked by the numerous pigment granules. The fatty globules which filled the peripheral cytoplasm of the enlarged interstitial cells have disappeared entirely (fig. 26).

A number of the cells have not undergone much change in size. In figure 26 they are seen as the larger c^lls with coarse spherical granules ot various sizes within them. A group of


such cells under higher magnification and as affected by the osmic acid of Meves' fixer, is shown in figure 25. The large pigment globules have evidently been derived from the fatty material with which these cells previously were filled. These large pigment cells are most numerous just at the close of these retrogressive changes. The pressure having been relieved by the enormous decrease in the size of the other cells, these pigment cells are more or less spherical at this stage. Osmic acid still darkens the granules, at least on the surface; but they are very insoluble, being preserved fairly well even in Carnoy's fluid as will be seen in figure 4, which also shows that from the very first the nucleus may be irregular, which is most often the case, though it may in some instances be apparently normal and vesicular as in figure 5. Figure 5 is taken from a section which passes through the cell near the middle plane and indicates that the now-pigmented globules may still retain their peripheral arrangement, leaving a less pigmented area in the center. This possibly represents somewhat of an intermediate stage in the formation of the more solid and irregular pigment cells. Here then .we evidently see the source of the pigment cells that have been followed through the preceding stages. The interpretation that they originate from ordinary interstitial cells which undergo a special pigmentarj' degeneration at the time when the rest of the cells lose their fat and become small, is borne out by the absence of these pigment cells in w^oodchucks that are less than one year old and have not passed through this adult retrogression of the interstitial cells. No pigment granules are found in any of the interstitial cells of the twelve animals less than a year old. Fatty granules which are blackened with osmic acid are however present in small quantities in the scanty cytoplasm of some of the interstitial cells of these young animals. Thus a new crop of pigment is produced once a year as fine granules within the ordinary interstitial cells and as larger granules which fill more or less completely certain other interstitial cells, which as a result do not at this time decrease much in size but remain as large pigmented cells for many months or even a year and perhaps a few survive even longer, though ap


parently most of them disappear by the end of the next period of hypertrophy.

The origin of the pigment in the interstitial cells has been debated. Von Hansemann ('95) considered that' it was not due to pigmentary degeneration, but rather that it is an infiltration from some other source, since the pigmented cells are the larger. Kasai ('08), however, states that in the human testis the pigmented cells are not especially larger and that it is the younger cells that are not pigmented. Pigmentation, according to the latter author, commences in the human testis first at 21 years of age. With the advance of years and especially in old age the pigmentation increases. These facts Kasai took to indicate that it is a pigmentary degeneration — a view fully supported by this work on the woodchuck. Whitehead ('08) believes that the large pigment-laden cells which he observed in a case of cryptorchism are due to pigmentary degeneration of ordinary interstitial cells which in the retained testis have become useless — there being a normal scrotal testis to supply the necessary internal secretion.

During this disintegration of the interstitial cells there is a great increase in the number of ordinary adipose cells, which are present in the testis at all times of the year. Fatty degeneration within the tubules is also seen at this time; in fact, fatty globules are demonstrable in the germinal epithelium at all seasons.

During all this time spermatogenesis is slowly progressing. The diameter of the tubules changes but little during the entire year. They are probably somewhat smaller later in July, August, and September and larger in November and December. However, the extremes in size may be encountered at other periods of the year and the limited number of animals representing any one period makes a- definite statement impossible. The rather marked variation in the diameter of the tubules included in the general views of the testis in the figures accompanying this paper, is not representative. It happened that in selecting places to show the relative amount of interstitial tissue the tubules were not carefully observed as to size, as was evident from a compari


son of the pictures side by side. To check this point the diameter of a large number of tubules were measured with a filar ocular micrometer with the results stated above. However, there is a great difference in the degree to which the tubules are filled with germinative cells. The open lumen seen during April and May is now filling up with spermatogonia.

Cross and longitudinal sections of the testis rather indicate that the tubules are merely separated farther from each other when the interstitial cells are greatly developed. No data is available as to the length of the tubules. Reconstructions from serial sections would undoubtedly be necessary to show what changes, if any, occur in this regard. There is no evidence of any atrophy of the tubules as a whole at this or any other period. The number of cross sections of tubules encountered in a complete cross section of the entire testis near the middle, just after the testis has been reduced in size, was about 850 as compared with about 1000 when the testis is maximal in size. In six animals with testes at a minimum the number of times tubules were cut in a complete cross section varied from about 750 to about 900 with an average of about 850; while in six others with testes maximal in size the number varied from 900 to 1100 with an average of about 1000. Since the testis enlarges from growth within, which undoubtedly increases the pressure, the testis is more nearly round at this time than when it is minimal. The transverse diameter is therefore relatively greater and this may explain this difference. On account of the shape of the tubules these figures are probably of very little value. Since there is no sudden change in the tubules at the time when the testis rapidly diminishes in size, the decrease must be due to the atrophy of the interstitial cells, first in size and then more slowly in number.

At no time was there any evidence of the crystalloids, first described by Reinke ('96) in the human testis as passing from the interstitial cells into the lymph vessels. While many subsequent writers have seen these crystals in the human testis, their passage into the lymph vessels has not been confirmed. It appears that these crystals have been found to any extent only


in the human testis. Neither is there any special relation between the blood vessels and the interstitial cells such as was first reported by Boll ('71) as being the typical arrangement in case of the rabbit, but which has not in general been supported by later investigators. When the interstitial cells have enlarged so as to crowd the intertubular space, the capillaries run between the interstitial cells, which of necessity must surround the vessels with little or no connective tissue between the capillary wall and the cells. But at other times when the interstitial tissue is loose, they are not especially arranged about the vessels.


In order to get at a single glance the essential points in regard to the question at issue, the principal facts with the authority for the same, has been placed together in the accompanying chart. Occasionally a few incidental facts which are included were not given by the specific authority. Thus the time of hibernation has in some cases been added, as well as other data deemed of interest. Only the approximate time intervals are, of course, possible in such a chart, since these will vary with the seasons and the localities. How.ever, they are believed to be sufficiently accurate to give a proper setting for the various phases of the cycle. The curves are not constructed upon any quantitative basis, since hot sufficient data are given to make that possible. The highest point of the curve is intended to indicate merely the approximate time when either spermatogenesis or the development of the interstitial cells, as the case may be, is at its highest, according to 'the author's statement of the case. When more than one investigator have reported on the same animal and only slight variations exist in the findings, the curves have been combined into one. Undoubtedly other observations on seasonal changes in the interstitial cells of adult animals are recorded in the enormous literature upon the testis, but these included in the table are all that Avere found to be directly to the point, after a reasonable search through available sources.



It is clear from this siuninary that interstitial cell development does not always run parallel with spermatogenesis. The notable exceptions are the mole (giving precedence to the most complete report by Tandler and Grosz) and the woodchuck. In the latter case the interstitial cells suddenly undergo retro

Curves of Seasonal Dimorphism of Testis

- 5permato^ene5i5


= Interstitial Cells ; IB-^l - hibernation



gressioii while spermatogenesis goes on uninterruptedly. The interstitial cells remain apparently inactive and greatly reduced for many months while spermatogenesis goes on progressively. It is only towards the last phases of spermatogenesis that the interstitial cells show increased activity. This seems to be the most striking correlation, namely, that the interstitial cells follow wth renewed growth somewhat behind the spermatogenic cycle. And in this sense it is conceivable that in the mole and Rena esculenta the interstitial cells are only somewhat more than usually behind.

The conclusion expressed by Allen ('04), that in the embryology of these cells they develop with reference to degeneration in the sex glands, is pertinent, although Whitehead ('04) did not confirm this point, claiming that the interstitial cells appear before there are any signs of degeneration. Mazzetti ('11) explains their decrease in the frog in October as being due to the fact that they have accomplished their purpose of absorbing the useless testicular elements after intensive spermatogenesis. Kingsbury ('14) emphasizes this point in connection wdth the homologous cells of the ovary, where they are most numerous about atretic follicles. In old men it appears well established, especially from the extensive work of Spangaro ('01) and the more recent examination of 130 human testes by Kasai ('08), that while there is usually a noticeable increase in the interstitial cells at puberty, the number seems to decrease again during the active sexual life of the individual to again increase with old age, a time when there is more or less atrophy in the tubules. It is certainly difficult to understand why this increased growth should take place late in the spermatogenic cycle or with a cessation of sexuality if this growth has anything to do with the production of an internal secretion that is of importance in the development of the genital tract, secondary sex characters or the sexual instincts. It appears that in frogs (Rana fusca and Rana viridis) spermatogenesis is initiated without any corresponding observable change in the interstitial cells and when they are least active. On the other hand in the woodchuck and mole the spermatogenic cycle commences when the interstitial cells are well developed or beginning to decline.


The relation of the interstitial cells to the breeding period, when copulation or fertilization takes place, is equally confusing, since this i)eriod of sexual activity (spoken of as rutting in inanj" animals) may occur either when the interstitial cells are showing evidence of increased activity (toad, hedgehog, marmot (?), woodchuck) or when not noticeably changing (frogs, mole). The descent of the testes may occur when the interstitial cells are minimal (mole) or after these cells have commenced to hypertrophy (woodchuck, hedgehog). In the woodchuck they tend to remain scrotal while the interstitial cells are enlarged. Here attention should be called to the experiments of Harmes ('13) in which he showed that the characteristic thickening of the skin on the hands of the male frog during the breeding season may be brought about through some influence exerted by Bidder's organ alone in the absence of the testis. This is true even when Bidder's organ has been transplanted into the dorsal lymph sac, and Bidder's organ contains no interstitial cells. However, the testis alone in the absence of Bidder's organ can also cause this periodic thickening of the skin.

The general functional reduction which takes place during hibernation also applies variously to the interstitial cells. In most of the hibernating animals they are quiescent during dormancy, but in the toad they are developing. Nor is there necessarily any marked change in their behavior at the onset of torpidity as is also shown in the case of the toad, where they go on increasing, and in the case of the woodchuck, where they remain in the minimal stage to which they have been reduced during the late summer preceding.

It thus appears that while the periodicity of the interstitial cells would suggest some important function for them, it is difficult to say what this function is specifically, because of the lack of uniformity in their behavior. Granting that in the main the ol)servations here discussed are correct, one would hesitate to use them as evidence of any weight in support of the generally accepted idea that the interstitial cells of the testis produce an internal secretion of specific importance to the sexual life of the organism.


The role of the large amount of fatty material which may accumulate in these cells in certain animals — though rather scarce in others, notably the pig — is of course equally obscure. The chemical nature of the large globules in the interstitial cells of the cat has been investigated especially by Whitehead ('12 b), who concluded that they are a mixture for the most part of phosphatid lipoid but that cholesterinesters and neutral fats are probably also present. The views that this lipoid is the material out of which the internal secretion is made (Loisel '02), that it is the internal scretion (Ganfini, '02), etc., need to be supported by many more facts. The effects of lipoid extracts of the testis, such as those reported by Iscovesco ('13), are, of course, very suggestive that the lipoid is an active agent.

In this connection we may recall the view of Plato ('97), that this fatty material passes through the walls of the tubules into the Sertoli cells to be used as food for spermatogenesis. For this reason he termed the interstitial cells 'Trophische Hiilfzellen.' Lenhossek ('97) entertained somewhat the same idea and Regaud ('01) presented evidence of such a passage of substance from the interstitial cells to the Sertoli cells. Plato's idea was strictly opposed by Beissner ('98) and in general this passage of material through the walls of the tubules from the interstitial cells has not been supported by the later investigators. Friedmann ('98), however, suggested .that the pigment in summer and the fat in winter, in the case of the toad, especially, serve as the chief sources of nutrition for the processes going on in the tubules. Herxheimer concluded that in individuals not yet sexually mature, the fat is mostl}^ found in the interstitial cells and is reserve material for the growing testis, while in the sexually mature the fat is mostly found in the tubuli contorti where it serves as reserve material for the development of spermatozoa. Hanes ('11), however, associates the fat storage with the Sertoli cells.

Of historical interest, at least, is also the conclusion of von Bardeleben ('97) that the interstitial cells of the testis are young Sertoli cells which can pass through the walls of the tubules and replace worn out Sertoli cells. Goldmann ('09),


111 fact, claims to have demonstrated such a passage of cells into the tubules. Hanes ('11), on the other hand, repeated (ioldmann's vital staining method and could find no suggestion of this migration of cells as described by Goldmann.


1. In the woodchuck (Marmota monax) during late summer and autumn, the interstitial cells of the testis are minimal in size and probably reduced in number. The scanty cytoplasm of these cells contains numerous pigment granules, some fine lipoid granules, but only a very few of the cells contain the coarser and more fat-like globules which are easily demonstrable with osmic acid. There are a number of large interstitial cells, which are gorged with prominent pigmented granules and which have resulted from a degeneration of some of the other common and more numerous type of interstitial cells. A new spermatogenic cycle is in progress. The testis is small, dark in color and abdominal in position.

2. There is no sudden change in the interstitial cells with the onset of hibernation and little or no change during dormancy, except that there is a slight gradual decrease in pigmentation. Spermatogenesis remains much the same during the torpid state as just before winter-sleep sets in. The tubules are filled with spermatocytes showing open maturation figures during the entire winter.

3. In the spring as the animal is waking up from hibernation, the interstitial cells rapidly enlarge and apparently increase in number. The nucleus increases only slightly. The great increase is primarily in the cytoplasm and is due to the development of a dense central mass of cytoplasm and the accumulation of fatty globules in the more peripheral portion. Fine lipoid granules are also abundant in the central cytoplasm. Ths great interstitial cell development forces apart the tubules and doubles the diameter of the testis, which descends into pouches essentially representing a scrotum remaining in communication with the abdominal cavity proper. Spermatogenesis suddenly shows renewed activity and free sperms are seen by the


last of Mtirc'h, or two to three weeks after waking up. The interstitial cells do not, however, reach their maximal size till the last of April. There is a distinct decrease in pigmentation.

4. Regressive spermatogenesis occurs during the last of April and a new cycle begins early in May, while the interstitial cells remain greatly developed for at least two months longer.

5. By July the testes have returned to their abdominal position, the interstitial cells begin to show^ signs of decrease and by August most of them are reduced to almost nothing more than naked nuclei. The cytoplasmic lipoids have been absorbed or transformed into a new^ crop of pigment w^hich remains as small brown granules in the scanty cytoplasm in the form of a cap covering about one-half of the nucleus. The nucleus is reduced again in size. A number of the interstitial cells do not thus decrease in size but remain large, the lipoids within them having been transformed bodily, so to speak, into a pigmented substance apparently of the same chemical nature as that of the smaller pigment granules in the ordinary interstitial cells. The nucleus of these special pigmented cells frequently is irregular or pyknotic and may be forced to the periphery of the cell. The testis as a whole is reduced to nearly one-eighth of its former size and is darker in color than at any other stage. Spermatogenesis is slowly progressing uninterruptedly.

6. A review of the literature on the correspondence between interstitial cell activity, spermatogenesis, breeding period and hibernation indicates great variability, with interstitial cell growth more uniformly related to the later and regressive stages of spermatogenesis than to the initial stages, though there seem to be exceptions even to this generalization.

It affords me great pleasure to acknowledge the guidance received from Dr. B. F. Kingsbury in this research, and the assistance of Mr. R. S. Outsell, in the retouching of the photographs.



Allen, B. M. 1!H)4 The embryonic development of tlio ovary and testis of the

mammals. Am. Jour. Anat., vol. 3, p. 89. vox Bakdelkben, K. 1897 Beitrpge zur Histologic des Hodens und zur

Spermatogenese beim Menschen. Arch. f. Anat. ii. Physiol., Anat.

Abth., Suppl. Bd., p. 193. Beissner, H. 1S98 Die Zwischenzellen des Hodens und ihre Bedeutung.

Arch. f. mikr. Anat., Bd. 51, p. 794. Bell, W. Blair 1916 The Sex Complex. New York. Bensley, R. R. 1911 Am. Jour. Anat., vol. 12, p. 309. Boll, F. 1871 Untersuchungen iiber den Bau und die Entwickelung der

Gewebe. Arch f. mikr. Anat., Bd. 7, p. 275 (footnote p. 324). BouiN ET Ancel 1903 Recherches sur les cellules interstitielles du testicule

des mammiferes. Arch. d. zool. exper., Ser. 4, T. 1, p. 437.

1904 Recherches sur la signification physiologique de la glands inter stitielle du testicule des mammiferes. Jour. d. physiol. et de path.

gen., T. 4, p. 1012.

(And many other articles which may be obtained from the extensive

bibliography given by Biedl and by Tandler and Grosz.) Champy, C. 1908 Note sur les cellules interstitielles du testicule ch^ez les

batraciens anoures. Compt. Rend. Soc. Biol., T. 64, p. 893. CusHiNG, H. and E. Goetsch 1915 Hibernation and the pituitary body. Jour.

Exper. Med., vol. 22, p. 25. DuESBERG, J. 1917 Chondriosomes in the cells of fish-embryos. Am. Jour.

Anat., vol. 21, p. 465. FiNOTTi, E. 1897 Zur Pathologie und Therapie des Leistenhodens, nebst

einigen Bemerkungen liber die grossen Zwischenzellen des Hodens.

Arch. f. klin. Chir., Bd. 35, p. 120. Fraxkl, L. 1900 Beitrage zur Lehre vom Descensus testiculorum. Sitzungs ber. d. Kais. Akad. d. Wiss., Wien, Math.-Naturw. KL, Abt. 3, Bd.

109, Heft 1. Friedmann, F. 1898 Beitriige zur Kenntniss der Anatomie und Physiologie

der miinnlichen Geschlechtsorgane. Arch. f. mikr. Anat., Bd. 52,

p. 856. Ganfini, C. 1902 Struttura e sviluppo delle cellule, interstiziali del testiculo.

Arch. Ital. di Anat. e di Embri., T. 1, p. 233 and Comi)t. Rend. Soc.

Biol., 1903, T. 39, p. 150.

1903 Les cellules interstitielles du testicule chez les animaux hiber nants. Arch. Ital. d. Biol., T. 40, p. 323. (Apparently the complete

report was made in Bollettino della R. Accademia Medica di Genova,

1903.) GoLD.MAXx, E. S. 1909 Die iiussere und innere Sekrction des gesunden und

kranken Organismus im Lichte der vitalen Farl)ung. Beitr. z. klin.

Chir., Bd. 64, p. 192. Haxes, F. M. 1911 The relations of the interstitial cells of Leydig to the production of an internal secretion by the manunalian testis. Jour.

Exper. Med., vol. 13, p. 338.


Hanes, F. INf. and J. Ivosenbloom 1911 A lusiol(),a;ic:il and chemical study of

the fatty mutter of normal and cry])t()rchid testis. Jour. Exper.

Med., vol. V.i, p. 355. Hansemann, D. 1S95 tJber die sogenanntcn Zwischenzellen des Hodens und

deren Bedeutung bei pathologischen Veriinderungen. Arch. f. path.

Anat. u. Physiol, u. f. klin. Med., Bd. 142, p. 538.

1896 liber die grossen Zwischenzellen des Hodens. Arch. f. Anat.

u. Physiol., Physiol. Abt., p. 176.

1912 Kurze Bemerkungen liber die Leydigschen Zwischenzellen des

Hodens. Arch. f. Entw. d. Organ., Bd. 34, p. 475. Harmes, AV. 1913 Die Brunstschwielen von Bufo vulgaris und die Frage ihrer

Abhiingigkeit von den Hoden oder dem Bidderschen Organ; zugleich

ein Beitrag zur der Beteutung des Interstitiums. Zool. Anz., Bd. 42,

p. 462. Hart, D. B. 1909 The nature and cause of the physiological descent of the

testis. Jour. Physiol., vol. 44, p. 4. Herxheimer, G. ilber Fett im Hoden bei Menschcn (und Thieren). Sonder Abdruck aus der Orth-Festschrift. Howell, A. H. 1915 North American Fauna, No. 37, Revision of the American Marmots, U. S. Dept. of Agri., Washington. Iscovesco, H. 1913 Action physiologique d'un lipoide (II B b) extrait du

testicule. Compt. Rend. Soc. Biol., T. 75, p. 445. Jackson, C. M. 1917 Effects of inanition and refeeding upon the growth and

structure of the hypophysis in the albino rat. Am. Jour. Anat., vol.

21, p. 321. Kingsbury, B. F. 1911 Histological demonstration of lipoids. Anat. Rec,

vol. 5, p. 313.

1914 The interstitial cells of the mammalian ovary: Felis domestica.

Am. Jour. Anat., vol. 16, p. 59. Lecaillon, a. 1909 Sur les cellules interstitielles du testicule de la taupe

(Talpa europaea) considere en dehors de la periode de reproduction.

Compt. Rend. Soc. Biol., T. 66, p. 599. Kasai, K. 1908 Uber did Zwischenzellen des Hodens. Arch. f. path. Anat. u.

f. klin. Med., Bd. 194, p. 1. Lenhossek, M. v. 1897 Beitrage zur Kenntniss der Zwischenzellen des Hodens.

Arch. f. Anat. u. Physiol., Anat. Abth., p. 65. LoiSEL, G. 1902 Sur I'origine du testicule et sur sa nature glandulaire. Compt.

Rend. Soc. Biol., T. 54, p. 57.

Sur le lieu d'origine, la nature et le role de la secretion interne du testicule. Compt. Rend. Soc. Biol., T. 54, p. 1034. Mann, F. C. 1916 The ductless glands and hibernation. Amer. Jour. Physiol., vol. 41, p. 173. Marshall, F. H. A. 1911 The male generative cycle in the hedgehog; with

experiments on the functional correlation between the essential and accessory sexual organs. Jour. Physiol., vol. 43, p. 247. Mazetti, L. 1911 I caratteri sessuali secondari c le cellule interstiziali del

testicolo. Anat. Anz., Bd. 38, p. 361. Merriam, C. H. 1884 The vertebrates of the Adirondack region, Northern New York, (Mammalia). Trans. Linnaean Soc. N. Y., vol. 2, p. 142.


IMeves, F. 19().S Arch. f. niikr. Anat., Bd. 72, p. S32.

Meves, F. uiul J. DuESBERG 1908 Arch. f. mikr. Aiiat.. Bd. 71, p. 573.

Morris, R. T. 1916 A case of testicle grafting with unexpected results. Jour.

Amer. Med. Assoc, vol 67, p. 741. NrssBAUM, M. 1880 Von der Bedeutung der Hodenzwischensubstanz. Arch.

f. mikr. Anat., Bd. 18, p. 85. Owen, R. 1808 Comparative Anatomy and Physiology of Vertebrates, London, vol. 3, p. 649. Pick, L. 1905 Uber Xeubildungen am Genitale beim Zwitter, nebst Beitriigen

zur Lehre von den Adenomen des Hodcns und Eierstocks. Arch. f.

Gyn., Bd. 76, p. 191. Plato, J. 1896 Die interstitiellen Zellen des Hodens und ihre physiologische

Bedeutung. Arch. f. mikr. Anat., Bd. 48, p. 280.

1897 Zur Kenntniss der Anatomie und Physiologic der Geschlechts organe. Arch. f. mikr. Anat., Bd. 50, p. 640. Regaud, C. 1901 Etudes sur les structure des tubules seminiferes et sur les

spermatogenese chez les mammiferes. Arch. d'Anat. Mikr., T. 4,

p. 101.

1904 Etat des cellules interstitielles du testicule chez la taupe pendant la periode de spermatogenese et pendant I'ctat de repos des canalicules seminaux. Compt. Rend. d. I'Assoc. d. Anat., p. 54.

Reinke, F. 1896 Beitrage zur Histologic des Menschen. I. tJber Krystal loidbildungen in den interstitiellen Zellen des menschlichen Hodens.

Arch. f. mikr. Anat., Bd. 47, p. 34. Sehrt, E. 1904 Zur Kenntniss der Fetthaltigen Pigmente. Arch. f. path.

Anat. u. Physiol, u. f. klin. Med., Bd. 177, p. 248. Spangaro, S. 1901 Uber die histologischen Veranderungen des Hodens, Neben hodens und Samenleiters von Geburt an bis zum Greisenalter. Anat.

Hefte, Bd. 18, p. 592. Tandler, J. UND S. Grosz 1911-1912 Uber den Saisondimorphismus des

Maul wurf hodens. Arch. f. Entw. der Organ., Bd. 33, p. 297 and

Bd. 35, p. 132. Whitehead, R. H. 1904 The embryonic development of the interstitial cells

of Leydig. Am. Jour. Anat., vol. 3, p. 167.

1905 Studies of the interstitial cells of Leydig. Am. Jour. Anat., vol. 4, p. 193.

1906 The presence of granules in the interstitial cells of the testis. Anat. Rec, vol. 1, p. 60.

1908 a A peculiar case of cryptorchism, and its bearing upon the problem of the function of the interstitial cells of the testis. Anat. Rec, vol. 2, p. 177.

1908 b Studies of the interstitial cells of Leydig. No. 3. Histology. Anat. Rec, vol. 1, p. 213.

1912 a On the chemical nature of certain granules in the interstitial cells of the testis. Am. Jour. Anat., vol. 14, p. 63. 1912 b A microchemical study of the fatty bodies in the interstitial cells of the testis. Anat. Rec, vol. 6, p. 65. WiLDMAN, E. E. 1913 Jour. :\Iorph., vol. 24, p. 426 (footnote).



1 Photogiapli of section of testis of adult woodchuck in early stage of hibernation; Dec. 1; rectal temperature 19°C. ; reduced body weight 2525 grams, of which one testis constituted 0.034 per cent. Carnoy's fluid; 4 ^ thick; iron hematoxylin. X 129. Testes abdominal. Shows the relative amount of interstitial tissue. Tubules filled with spermatocytes.

2 Retouched photograph of characteristic interstitial space of above section under higher magnification. X 527. Many small ordinary interstitial cells with pigment granules stained black. Three large interstitial cells so filled with large pigment granules that nucleus is obscured.

3 Retouched photograph of section of testis of above animal showing a characteristic group of large and especially pigmented interstitial cells. Carnoy's fixer; 4 ^t thick; hematoxylin and eosin. X 527. Nucleus is irregular and frequently displaced. A few ordinary interstitial cells, containing a few pigment granules, are present.

4, 5 Retouched photographs of newly foruLcd 'pigment cells' from same animal as in figure 23, under higher magnification. X 1053. In figure 5 the pigment granules appear to have a peripheral distribution, occupying the position of the fat globules froin which they were evidently formed.

6 Photograph of a section of the testis of adult woodchuck in last stage of hibernation; March 6; rectal temperature 12°C. ; reduced body weight 1S49 grams, about two-thirds of the weight before hibernation. Same technique and magnification as in figure 1. Testes abdominal. Animal beginning to wake up.

7 Retouched photograph of section of testis of same animal as in figure 6. Modified Zenker's fluid; 4/i thick; copper hematoxylin (Weigert). X 527. Many granules (lipoid?) in addition to pigment. One large h^avily pigmented cell present.

8 Camera luoida drawing of cell pictures not infrequently met with during the enlarging period in the interstitial cell cycle. Carnoy's fluid; 4 ^i thick; iron hematoxylin. X 527. A suggestion that there has been nuclear division.

9 Retouched photograph of interstitial cell at maximal stages of development containing two nuclei. Carnoy's fluid; 4 n thick; iron hematoxylin. X 1053.

10 Photograph of section of testis of adult woodchuck with interstitial cells maximal; May 8. Carnoy's fixer; 6 n thick; iron hematoxylin. X 200. Illustrates how the interstitial cells tend to be arranged in 'nests' in about half of the animals in which the interstitial cells were greatly developed.








11 Photograph of section of testis of adult woodchuck practically at full activity and been so for a week or more; March 21; rectal temperature 32°C.; reduced body weight 1650 (about two-thirds of original weight before hibernation). Testes abdominal. Carnoy's fluid; 4 /x thick; iron hematoxylin. X 129. Interstitial cells have greatly increased.

12 Retouched photograph of a mass of interstial cells from preceding section under higher magnification. X 527. Pigment scattered out in the abundant cytoplasm which is vacuolated from having fat dissolved out.

13 Retouched photograph of a group of interstitial cells from testis of same animal as in figures 11 and 12. Modified Zenker's; 4m thick; copper hematoxylin. X 527. Various stages in the expansion of the cells.

14 Photograph of section of testis of same animal as in preceding figures in this plate. Meves' fixer; 4 n thick; unstained and uncovered. X 129. Fatty globules blackened with the osmic acid. Tubules contain much fat. Five adipose cells are present (four in one group). Large blood vessels. Section too near the surface of block to be perfect.

15 Photograph of section of testis of adult woodchuck when interstitial cells are maximal; during full activity and after feeding at least six weeks; May 10; rectal temperature 37°C.; reduced body weight 2420, of which one testis constituted 0.111 per cent. Testes scrotal. Same technique and magnification as in figures 1, -6 and 11.

16 Retouched photograph of interstitial cells of preceding section under higher magnification. X 527. Pigment nearly absent. Central cytoplasm finely vacuolated (seen best in the cell marked with a +); peripheral cytoplasm coarsely vacuolated. Vacuoles due to dissolving out the lipoids. To be compared with figures 2 and 24.

17 Retouched photograph of interstitial cells of same. animal as in figures 15 and 16. Modified Zenker's, etc., as in figures 7 and 13. X 527. The fine vacuoles of figure 16 are here filled with granules (lipoid) stained black. Peripheral fat not stained.

18 Photograph of section of testis with interstitial cells maximal. Meves' fixer, etc., as in figure 14. X 129. Peripheral fat of interstitial cells and adipose cells black. Two blood vessels. Insert shows three cells more highly magnified. X 527. Nucleus of one cell shown as white spot. Section too near the periphery of block to be perfect. The poor penetrating power of osmic acid and the distortion of the testicular contents at the cut surface made perfect sections illustrating relative amount of fat unavailable.





'^i^LA i^"*' \'^

■h\ .

«' Hid '^T".;




1!) Pilot ()<;ra])li of a section of testis of adult woodchuck with interstitial cells at a maxinuim; April 29; reduced body weight 2850 grams, of which one testis constituted 0.106 per cent. Carnoy's fixer; 4 ix thick; iron hematoxylin. X 20. The + is placed on a large nodule of interstitial cells which are especially rich in fat. Lumen of tubules open.

20 Photograph of section of testis of adult woodchuck with interstitial cells beginning to atrophy; July 10; reduced body weight 2825 grams, of which one testis represents 0.038 per cent. Testes still scrotal. Carnoy's fixer; 4 /^ thick; iron hematoxylin. Tubules no longer open. To be compared directly with figure 15.

21 Retouched photograph of section of testis of same animal as in figure 20. Modified Zenker's fluid; 4 ^ thick; copper hematoxylin. X 527. Central cytoplasm still contains numerous fine granules (lipoid) in addition to pigment. To he cnmnared directly with figure 17.

22 Photograph of section of testis of same animal as in two preceding figiu-es. Meves' fixer; 4 m thick; unstained and uncovered. X 129. A number of large globules of fat are still left in the interstitial cells. To be compared directly with figures 18 and 14.

23 Photograph of section of testis of adult woodchuck just after atrophy of interstitial cells; August 5; reduced body weight 3190 grams, of which one testis represents 0.016 per cent. Testes abdominal. Carnoy's fixer; 4 ju thick; iron hematoxylin. X 129. Interstitial cells numerous but greatly reduced in size. Tubules contain many spermatocytes.

24 Retouched photograph of above section under higher magnification. X 527. In addition to numerous ordinary interstitial cells, now rich in pigment but greatly reduced in size, there are four newly formed 'pigment cells' easily distinguishable by their large size, darker color and numerous large pigment granules.

25 Photograph of a portion of the section immediately following under higher magnification. X 527. A group of newly formed 'pigment cells' as they appear after being acted upon by osmic acid. Only colored plates could do justice to their appearance as seen when the pigment is unstained such as occurs after Carnoy's fixer and hematoxylin and eosin. Carnoy's fixer tends to make them more irregular in outline as may be seen by comparing this figure with figures 4 and 5, all of which are from sections of the same animal.

26 Photograph of section of testis of this last animal; Meves' fixer; 4 m thick; unstained and uncovered. X 129. The osmic acid darkens the pigment granules of the interstitial cells and blackens the fatty globules in the tubules as well as ordinary adipose cells, which now are very numerous. The hirge newly formed 'pigment cells' stand out prominently. There is a striking absence of fat globules in the interstitial cells. Compare with figure 18.





•^1 •; « 5 jtk. L • • . '•'"'

'^^.^ l^^

i\*'jL- "*>'*• r,^--. i-"-^/*" „ - - ■














i' .








ALBINO rat. Studies on the mammary gland. II. The fetal development of the mammary gland in the female.. . . 195 Ammocoetes. The relationships and histogenesis of thymus-like structures in 127

Anomalies in the human embryos and infants

at birth. On the frequency of localized.. 49 Arey, Leslie B. The normal shape of the

mammalian red blood corpuscle 439

Armadillo, Tatusia novemcincta. The early history of the germ cells in the 341

BANKS, J. B., Jordan, H. E. and. A study of the intercalated discs of the heart of the beef 285

Batjmgartner, E. A. The development of the serous glands (von Ebner's) of the vallate papillae in man 365

Beef. A study of the intercalated discs of the heart of the 285

Behavior of chicken bone marrow in plasma

medium. Cj'tological observations on the. 73

Birth. On the frequency of localized anomalies in the human embryos and infants at 49

Blood corpuscle. The normal shape ,of the mammalian red 439

BoLK, L. On metopism 27

Bone marrow in plasma medium. Cytological observations on the behavior of chicken 73

CAPSULE of the ear in the human embryo. The factors involved in the excavation of the cavities in the cartilaginous 1

Cartilaginous capsule of the ear in the human embryo. The factors involved in the excavation of the cavities in the 1

The early history of the germ 341

Cells in the armadillo, Tatusia novemcincta. The early history of the germ 341

Cells of the testis in the woodchuck (Marmota monax). Seasonal changes in the interstitial 475

Changes in the interstitial cells of the testis in the woodchuck (Marmota monax). Seasonal 475

Chicken bone marrow in plasma medium. Cytological observations on the behavior of 73

Corpuscle. The normal shape of the mammalian red blood 439

Cross striated muscle in tissue cultures. Behavior of 169

Cultures. Behavior of cross striated muscle in tissue 169

Cycle in the guinea-pig — with a study of its histological and physiological changes. The existence of a typical oestrous 225

Cytological observations on the behavior of chicken bone marrow in plasma medium. 73


rSCS of the heart of the beef. A study of the intercalated 285

EAR in the human embryo. The factors involved in the excavation of the cavities in the cartilaginous capsule of

the 1

Embryos and infants at birth. On the frequency of localized anomalies in the

human • • • • 49

Erdmann, Rhoda. Cytological observations on the behavior of chicken bone marrow in plasma medium 73

FACTORS involved in the excavation of the cavities in the cartilaginous capsule of the ear in the human embryo. The. 1 Fetal development of the mammary gland in the female albino rat. Studios on the

mammary gland. II. The 195

Foetus exhibiting bilateral absence of the ulna accompanied by monodactyly (and also diaphragmatic hernia). Anatomy of a seven months' 385

GERM cells in the armadillo, Tatusia novemcincta. The early history of

the 341

Gland. II. The fetal development of the mammary gland in the female albino rat.

Studies on the mammary 195

Glands (von Ebner's) of the vallate papillae

in man. The development of the serous. 365 Guinea-pig— with a study of its histological and physiological changes. The existence of a typical oestrous cycle in the 225

HEART of the beef. A study of the intercalated discs of the 285

Hernia. Anatomy of a seven months' foetus exhibiting bilateral absence of the ulna accompanied by monodactyly (and also diaphragmatic) 385 <