Book - An experimental analysis of the origin of blood and vascular endothelium (1915) 2

From Embryology
Embryology - 19 Apr 2024    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)

Stockard CR. An experimental analysis of the origin of blood and vascular endothelium. (1915) Memoirs of the Wistar Institute of Anatomy and Biology No. 7.

Historic Disclaimer - information about historic embryology pages 
Mark Hill.jpg
Pages where the terms "Historic" (textbooks, papers, people, recommendations) 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, interpretations and recommendations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)

II. A Study of Wandering Mesenchymal Cells on the Living Yolk-Sac and their Developmental Products: Chromatophores, Vascular Endothelium and Blood Cells

Charles Rupert Stockard (1879-1939)

  • This second part of the Memoir is a continuation and expansion of the foregoing study on "The origin of blood and vascular endothelium in embryos without a circulation of the blood and in the normal embryo." The first part is referred to in the following pages as the 'previous paper.'


The aim of the present consideration is an analysis of the histogenetic changes passed through by the mesenchymal cells in the living yolk-sac. A study of the origin and development of the blood and vascular endothelium in normal teleost embryos, and in other specimens in which the circulation of the blood had been experimentally prevented, made it evident that a detailed investigation of the development of the living yolk-sac would be most instructive for a comprehensive knowledge of the behavior of mesenchyme in forming the blood cells and vessels. The yolksac had been thoroughly investigated in sections and the appearance and location of the earliest blood islands and vascular formations were already familiar.

The egg of the Teleost is particularly adapted to the investigation of such a problem, since its yolk-sac has no definite mesenchymal layer and the freely wandering mesenchyme cells may be clearly seen between the ectoderm and yolk periblast. The remarkable extent to which the cells migrate and the great numbers of such wandering cells impress one with the importance of this cellular movement in embryonic development. The appreciation of this phenomenon also emphasizes the great danger of interpreting developmental processes from a mere study of serial sections. Sections fail to produce a correct impression of what is actually taking place in an area vasculosa. The study of living embryos is absolutely necessary and through it one quickly becomes acquainted with the remarkable role played by wandering cells in the formation of the heart and vessels, as well as the production of the future blood cells.

These facts have been pointed out long ago, but with little effect, as is indicated by the enormous literature containing the endless interpretations and guesses of numerous authors after studies of fixed and sectioned material. It is not intended to under rate the importance of the study of sections. However, such a study to be fully comprehended must be accompanied by observations on the living material as far as is practical. These observations are further greatly clarif:ed by an experimental modification of the normal developmental processes where such is possible.

Almost thirty years ago Wenckebach ('86), at that time a young medical student in Holland, described his observations on the living embryos of the developing bony-fish. In this contribution he lamented the fact that the knowledge of the embryology of the bony-fish, as well as of other vertebrates, was based almost entirely on studies of sections of the embryos. The germ layers were described actually as growing layers and from this layer formation the different organs were built or produced by foldings. Each cell was thought of as being passive and only through its division was the formation of the organs brought about. This initial investigation, and the only one so far as I know, by young Wenckebach gave him an entirely different view point regarding the processes of embryonic development.

Wenckebach readily observed the cells of the mesoblast independently wandering in amoeboid fashion, often with extraordinarily long protoplasmic processes, within the body of the embryo as well as upon the hypoblast-free yolk. The wandering cells move as with aim and purpose to form certain definite organs. In the formation of the anlage and further development of the heart, as well as the vessels and other structures, this independent wandering of the mesoblast cells performs a most important part.

Unfortunately, Wenckebach's scientific efforts ceased and his early study was unappreciated since on only one occasion was it considered by an investigator, Raffaele ('92), who studied similar material. Today the student gathers from text- and hand-books as well as descriptive embryological contributions much the same orthodox conception of the layers and foldings as all important factors in the origin and development of embryonic organs. No doubt the growth of layers and folds does contribute its part, but this part is almost negligible in a study of the development of the vascular system and blood.

Wenckebach and Raffaele, in their observations of the wandering cells on the yolk-sac failed to recognize the erythroblast. They were also unable to follow the processes in the differentiation of pigment and endothelial cells in the way at present possible with improved microscopes. The experimental embrj^os without a circulation of the blood are also most instructive for comparison in such a research.

The wandering mesenchymal cells on the yolk-sac differentiate into four distinct types of cells. The chromatophores of two varieties, one black and one of a reddish brown color, endothelial lining cells of the yolk vessels and islands of erythroblasts may all be seen to form in the living specimens. The association of the four types is also clearly determined. In addition to these cells the large periblast nuclei are conspicuously seen. The periblast never gives rise to any type of tissue or cells, but finally the nuclei become swollen and distorted and degenerate in the manner Wenckebach long ago described.

Material and Methods of Study

The material used for these observations and experiments was the embryo of the Teleost, Fundulus heteroclitus. The eggs of this fish are very transparent and may be readily observed by transmitted light with a high power microscope. A compound binocular microscope made by E. Leitz has been found to serve splendidly for such observations since with this instrument an oil immersion lens may be used to study the embryo suspended in a hanging drop from a thin cover glass over a hollow slide. A double or triple lens condenser facilitates the regulation of light and in a darkened field the almost transparent cells may be seen while granular or pigmented cells are distinctly outlined.

The mesenchyme cells are of sufficient size to be readily followed with a Zeiss DD objective while the eggs are grouped on the bottom of a watch glass. An ordinary microscope serves almost equally well for these observations but the Leitz binocular has the advantage of producing an apparently stereoscopic effect, since it permits the observer to look with both eyes at the same time. One is also enabled to see much better and look much more continuously than with one eye. An ordinary binocular microscope is unfit for the finer observations on account of the poor arrangements for condensing the light, and the magnification is insufficient for details of structure.

Observations have been made on the normal embryo at all developmental stages. Specimens in which the circulation of the blood was never established were also used, since in these the behavior and development of the cells of the yolk-sac are in no way contaminated b}^ the introduction of additional cells brought in the circulating current. Such specimens enable one to hold the yolk-sac in its original condition so far as cellular elements go, and further give an opportunity to test the influence of the circulation on the mode of differentiation and function of the mesenchymal cells.

The prevention of the circulation has been accomplished in the same manner as employed in the previous investigation and fully described in the September number of this journal. The eggs shortly after being fertilized are placed in solutions of alcohol in sea-water. The series of solutions most advantageously used is prepared as follows: 1.5 cc, 2 cc, 2.2 cc, 2.4 cc,, 2.8 cc, and 3 cc. of 95 per cent alcohol is added to 50 cc. of sea-water. These solutions are renewed after twenty-four hours and after another twenty-four hours the eggs are placed in pure sea-water.

Such a series gives, of course, gradiations of the effect. Eggs in the weaker solutions develop normally in many cases, while other individuals develop slightly slower than the normal and have the circulation of their blood arrested to different extents. Many individuals in all of the solutions fail entirely to establish a blood circulation and although the heart pulsates feebly it does not propel the plasma for one or another reason which has been previously discussed. In spite of the failure of the blood to circulate, the development of the cells on the yolk-sac progresses in an ahnost normal fashion and vessels and blood corpuscles arise in this region and may be carefully observed throughout the life of the embryo.

The observations on the living yolk-sac have been supplemented by a study of fixed and cleared specimens. Embryos at different stages of development are fixed in a saturated solution of corrosive sublimate to which 5 per cent of glacial acetic has been added. Eggs are left in this mixture for 4 or 5 minutes, then rinsed in tap water and placed in 10 per cent formalin. The formalin is changed after about one-half hour when it has become slightly cloudy. This method if carefully handled brings out in a most beautiful way the cell outlines of the ectoderm of the yolk-sac (fig. 34) . The yolk remains rather transparent and the mesenchymal cells may be observed beneath the clear cut net-work formed by the ectodermal cell borders. This method has been frequently employed by other workers and I have used it myself for ten years on Fundulus eggs, but have never before succeeded in getting this hea\"y outhne of the cells. It seems scarcely possible that so striking an appearance could have been overlooked, yet it is perfectly simple to obtain. Fundulus yolk-sacs fixed in this way are equally as beautiful as silver preparations of cell boundaries.

In addition to the above, I have used for the first time another solution which renders the specimen still more transparent. This is a mixture of strong formalin 5 parts, glacial acetic 4 parts, glycerine 6 parts, and distilled water 85 parts. Eggs are placed directly into this and left for two days, and then transferred to 10 per cent formalin for pemianent preser\^ation. The fluid mixture causes the egg to swell to some extent but it leaves the yolk as clear as in life and by fixing the cells causes them to stand out in beautiful contrast (figs. 5, 6, etc.). The mixture of glycerine and glacial acetic has been used for a long time in preparing transparent specimens of invertebrate eggs. Wilson in 1892 used it with Nereis eggs. The proportions here employed have been used by several students at Woods Hole and are not original with me, except that others leave the eggs permanently in the mixture while it seems better to put them in formalin after two days. The eggs remain equally transparent in formalin.

The cleared specimens are most valuable for use in connection with the studies of the living. But the remarkably beautiful filamentous processes of the wandering mesenchyme cells and the endothelial lining cells of early vessels are not so extensive as during life. Some shrinkage or contraction of these processes always accompanies fixation. The movement of the processes in life also gives one a much better conception of their form and structure.

III. The early wandering cells

About two hours after fertilization the eggs of Fundulus heteroclitus have undergone the first division and are in the two-cell stage. The cleavages then continue m a more or less regular fashion to form a discoidal mass of cells as a cap on the yolk. At eighteen to twenty hours the germinal disc is beginning to flatten or thin out in order to begin its expansion to cover the yolk sphere. After the fourth or fifth cleavage some of the peripheral cells of the germ disc are somewhat fused with the yolk mass and do not present a clearly formed distal cell wall. The nuclei of such cells continue to divide and begin to wander or are pushed out into the superficial yolk material. In this way are formed the so-called periblast nuclei, or more correctly periblast syncytimn, of the teleost. This periblast syncytium precedes the germ disc in its descent over the yolk, so that one observes loosely scattered nuclei of unusually large size forming an advance border around the periphery of the germinal disc. The nuclei multiply and finally lie scattered over the entire yolk surface by the time the germ ring or blastodisc has completely covered the yolk (figs. 5, 7 and 8).

These periblast nuclei are of interest to us in the present consideration only on account of the fact that they are located in a superficial syncytium covering the yolk. It is over this syncytium that the mesenchymal cells wander. The periblast of the hypoblast-free yolk-sac of the teleost, in so far as position is concerned, may be compared to the endodermal covering of the yolk-sac in other meroblastic eggs.

The outer cover of the yolk-sac in Fundulus is formed by the germinal disc as it grows over the yolk. This constitutes the ectoderm of the sac which is its only true or typical layer. Thus the yolk-sac consists of an outer-continuous ectodermal layer beneath which are freely wandering mesenchymal cells and below these the periblastic syncytium fuses into the yolk material itself.

The periblast nuclei were interpreted by Agassiz to represent the survival of the nuclei which had at one time in phylogeny controlled the segmentation of the yolk. These were the nuclei of the former yolk laden cells in the holoblastic cleavage of the ancestral teleost. Others have thought that they played some part in the formation of the ventral wall of the gut, etc. In Fundulus, however, they take no part in the formation of the body tissues or organs, but may be observed to degenerate in tho late embryo. The periblast nuclei become very much vacuolated,


irregular in shape and huge in size before their final degeneration. After the embryo has hatched the remains of the yolk contain a protoplasmic mass in which the periblast nuclei are packed together and the whole is finally absorbed.

The blastodisc is separated from the yolk by a space which arises during the early hours of development. This space between the ectoderm and periblast has been interpreted by Agassiz and Whitman ('84), Ryder ('87), Wilson ('90) and others to represent the blastocoel or segmentation cavity. It is actually into this space that the wandering mesenchymal cells migrate and we shall later recall this fact as of importance in interpreting the natme of the vascular lumen in connection with other body cavities or spaces.

Twenty-four hours after fertilization the germinal disc is from one-quarter to one-third way over the yolk-sphere. Its wall has thinned out centrally and around the periphery is seen a thickened border, the so-called germinal ring. After about 48 hours the germ-ring has traveled almost completely over the yolk-sphere and now surrounds the small remaining uncovered pole of the yolk which may be considered a yolk-plug. A shieldshaped thickening, beginning about the twenty-fourth hour, extends from one region of the germ-ring towards the animal pole. This is the embryonic shield and along its median line a second thickening begins to appear which is the first indication of the embryonic body.

From the edges of the embryonic shield and from the germring as it finally encloses the yolk, an early migration of mesodermal cells takes place. The cells apparently do not wander far and some of them may again be included in the embryonic body. After the germ-ring has enclosed the yolk, from 45 to 50 hours usually, a very active migration of mesenchymal cells begins from the caudal and posterior lateral portions of the embryo. Figure 1 shows outlines of a few such cells wandering out from the side of an embryo of 45 hours. This figure is a camera sketch from life and even at this early tune there are some cells inclined to be more or less spindle or stellate in shape with long delicate filamentous projections, while other cells are



of more irregular shape with short amoeboid processes. Figures 3 and 4 show the cells highly magnified and in active movement; figure 3 the spindle-shaped delicate process type and figure 4 the heavier more amoeboid cell. These cells are actively wandering and changing in shape as figure 4 shows. From A to E are the outlines presented by a single cell at five minute intervals.

Fig. 1 A group of mesenchjane cells indicated in outline, camera lucida sketch, wandering out from the side of an early embryo, 45 hours old. Two kinds of cells are seen, one with delicate filamentous processes and another amoeba-like cell.

Fig. 2 A similar group in one microscopic field from an embryo 48 hours old, again showing the two types. The elongate spindle cells are future endothelial cells.




It is, therefore, difficult to state that the spindle-shaped cells may not change to the heavier amoeboid pattern and vice versa. But we shall see that these two forms are the probable if not actual forerunners of two groups of cells later, at any rate, with permanent shapes and structures differing in much the same way as the early appearances now differ. Figure 2 again shows a sketch from life of wandering cells on the yolk-sac of a 48 hour embrj^o, and here as usual the spindle cells are contrasted with the amoeboid ones. The spindle cells are assumed to be future vascular endothelial cells and the amoeboid cells are probably the future chromatophores of the yolk-sac.

The places from which cells wander out most actively are the borders of the tail, and particularly from that mass of cells representing the obliterated germ-ring. Figure 5 shows the cellular arrangement around the tail end of a 48 hour embr^^o fixed and cleared. The cells have withdrawn their processes. The open space in the cell group at the tip of the tail, yk, is the place still remaining between the borders of the germ-ring. Later, the tail of the embryo grows over this cell group so that it is less conspicuous. Figure 8, the tail end of a 56 hour embryo, shows such a condition.

The important fact which we shall later consider is that these cells form a mass continuous with the mesenchymal cells within the tail end of the embryo. The wandering cells may be interpreted to grow out from the end-bud or blastopore lip. They are a scant rudiment of the peripheral or ventral mesenchyme usually growing away from the blastopore lip over the yolk mass in the reptile and the bird. It will be presently shown that such an interpretation is upheld by the nature of the products to which these wandering cells give rise.

Figure 6 illustrates the wandering away of cells from the lateral mesoblast of an embiyo with two pairs of somites, 48 hours old. Figure 7 shows the head end of a 56 hour embryo. Scarcely any

Fig. 3 Camera outlines of wandering mesenchyme cells 48 hours old, all of the future endothelial type, highly magnified. A and B are two outlines of the same cell at a 6 minute interval.

Fig. 4 Camera outlines of one cell drawn at 5 minute intervals A to E. The cell is a migrating future chromatophore in an embryo 50 hours old (3b. DD ob.)




"=0 0^

009?^ o ^

aO o


Fig. 5 Camera lucida outline of the tail end and caudal yolk region of an embryo 48 hours old, fixed and cleared in gh^cerine-formalin. The germ ring just closing over the yolk pole, numerous mesenchjTne cells beginning to wander away from the caudal end, huge periblast nuclei indicated in outline and stipple over yolk. Yk, polar bit of yolk just being covered by union of germ-ring border.

Fig. 6 Portion of the caudal half of an embryo of 48 hours showing the first two pairs of somites recently formed, cells of the lateral plate mesoderm extend out upon the yolk as shown in outline, some beginning to wander away. Periblast nuclei outlined and stippled. Glycerine-formalin specimen. '


cell migration is taking place from this region. A few mesenchyme cells are found along the border of the head; these cells later take part in either the formation of the heart or pericardial wall. The tail end of the same embryo, figure 8, shows a remarkable contrast; here there is an enormous wandering out of cells from the mesoblast of the embrj^o. The two figures show the huge periblast nuclei to be widely distributed throughout the surface of the yolk sphere. These drawings are from cleared specimens and the cell outlines are more or less circular without the beautiful processes characteristic of the living.

The tail end of a living embryo 72 hours old, some time before the blood began to circulate, is illustrated by figure 9. The circle beneath the tail represents Kupffer's vesicle. The various shaped mesenchymal cells are represented in the act of wandering out over the nearby surface of the yolk. The embryo and yolk are beautifully transparent in life and the cells ^e clearly seen as they move upon the sm-face of the periblast.

An entire embryo, except the anterior portion of the head which extends beyond the curve of the yolk, is shown in figure 10 at a lower magnification. This specimen was 76 hours old when drawn. The heart had begun to contract slowly and feebly but no circulation of fluid had begun. Groups of mesenchymal cells are seen wandering away from the lateral and particularly the caudal regions of the embryo and are now scattered broadly over the yolk surface; there being very few, however, in the anterior region. The lateral plates of the mesoderm are seen at the sides of the head, and a circle at the caudal end indicates the Kupffer's vesicle which is always clearly shown at this stage.

In embryos of 72 hours, and somewhat earlier, there are wandering out from the tail region a nimiber of cells slightly smaller than the two types mentioned above. These small cells tend to be more or less circular in outline but show slow amoeboid movement as they send out short blunt processes. They group themselves into small clumps and are to give rise to erythroblasts or future red blood corpuscles in the yolk-sac as shall be discussed beyond. Figure 31, page 569, shows six such cells from the living yolk-sac of an embryo 90 hours old;


a circulation was partially established in this specimen but these cells had not yet been taken into the vessels.

The various wandering cells then represent, the mesodermal layer of the yolk-sac in the teleost. They never assume a membranous layer-like arrangement, but finally differentiate into the characteristic structures of the yolk-sac. As is shown by the illustrations, these cells are very numerous and during their earlier stages are actively changing their shape and moving over the yolk surface.

We may now consider the further development of such cells the living embryos.


O o O


Fig. 7 Outline of the head end of a 56 hour embryo, scarcely any wandering me.sench3Tnal cells in this region. Large periblast nuclei scattered over yolk surface.

Fig. 8 The caudal end of the same embryo; note the great contrast in the abundance of out-wandering mesenchymal cells. Glycerine-formalin specimen.

Fig._ 9 The caudal end of a living normal embryo of 72 hours, with beautifully delicate mesenchjmtial cells wandering away from the body; Kv., Kupffer's vesicle (3b. 2/.3 ob.).



Fig. 10 A camera sketch of an entire embryo of 76 hours except the anterior end of the head. The mesenchymal cells wandering away from the tail region and to a less degree from the sides of the body.

IV. Development and differentiation of the wandering cells

1. Chromatophores

a. The black type of chromatophore

The first to be considered of the four types of cells which develop on the yolk-sac are the black chromatophores. These are the largest and most conspicuous cells of the yolk-sac. In the early stages discussed above, one notes even in embryos of 2 days that certain" cells of the yolk mesenchyme are considerably larger than others.

These large cells may be followed through their development and the}^ will be fomid to differentiate into one or the other of the two types of chromatophores. The amoeboid cell shown in different stages of movement already referred to as figure 4, from a 52 hour embryo, is of this large type, and concluding from my observations on great numbers of embryos, this is an early condition of the future black chromatophore before any pigment granules are deposited.

Slightly older stages, figures 22 and 26, show the same cells containing a light amount of pigment granules. Between the second and third days the pigment granules appear and in an embryo 72 hours old, end of the third day, the chromatophores are already well differentiated freely moving huge cells.

A black chromatophore from an embryo 72 hours old is shown in figure 11 with one of its processes overlying the body of a brown chromatophore, the type to be considered in the following section. The black cell is loaded with coarse granules. The nucleus occupies a central position and is clearly shown on account of the displacement of the pigment granules by its transparent body. Several pseud opod-like processes project from the chromatophore which is actively moving. The clear cytoplasmic tip of the pseudopod extends beyond the granular mass.

Figures 12 and 13 are two other illustrations of the same cell after 15 minute intervals. Its shape is constantly changing and it is slowly moving in a direction towards the right side of the page. The brown pigment cell is also moving and their rates of progress are indicated by the increasing distance between them.

This movement of the chromatophores continues until about the end of the fourth or middle of the fifth day in the normal embryos. By this time all of the black pigment cells of the yolk-sac with few exceptions have taken up more or less permanent positions along the walls of the blood vessels or around the surface of the pericardial space. The individual chromatophores have increased enormously in size as is seen by comparing figures 11, 12 and 13 with figure 14, all drawn at the same magnification, though figure 14 is one-third more reduced in reproduction.

It must be appreciated, however, that some of the difference in extent is due to the flattening of the cells in figure 14.

Figure 14 shows two huge pigment cells on the yolk-sac of a 5 day embryo in the act of arranging themselves along a vessel wall. The granules are not so densely arranged as in the younger stages, since the cell body is greatly thinned out in pressing around the vessel. A number of granules are often arranged in solid black lines and masses as indicated in the figure.

The two cells are close together and a very peculiar phenomenon is taking place. Each cell sends out short processes to meet similar processes from its neighbor. The processes fuse, and finally the two cell bodies melt into one thus forming a pigmented syncytium about the vessels of the 3^olk-sac. The syncytia continue to expand along the vessels as enclosing sheaths (fig. 15) . The dense black of the young chromatophores becomes a steel grey as the granules are more thinly spread along the vessels.

In order to test whether the cells had actually joined or fused to fonii a true syncytium, I attempted to contract them, thinking that this should pull them apart unless they were actually united. The various solutions of KCl which Dr. Spaeth has found to contract the chromatophores within the embryo's body failed entirely to produce any change in the chromatophores of the yolk-sac. Solutions of adrenalin of one to 1000, one to 10,000 and one to 100,000, which Dr. Spaeth so kindly supplied me, were then tried. These solutions contract the pigment cells on the brain of the embryo until they appear as small black dots, but neither the black nor brown chromatophores on the yolk-sac respond in the slightest degree. Such specimens were preserved to show the extreme contraction of the chromatophores over the brain of the embiyo in contrast to the unchanged pigment cells of the yolk-sac.

Fig. 11 A black and brown chromatophore lying in contact on a j'olk-sac of 72 hours. The black cell is much the larger with broader pseudopod-like processes; both are in active movement as shown by comparing figure 12, of the same two cells 15 minutes later and figure 13, the same cells 20 minutes after figure 12. (3b. DD. ob.).

From this it would seem as though the material of the chromatophore had lost its contractile or wandering power after once becoming arranged around the yolk vessels. Those black chromatophores which retain their cellular individuality along the borders of the pericardial space also fail to contract when treated with KCl or adrenalin.

Although this physiological test failed to serve the purpose for which it was used I feel certain, after many observations, that the black chromatophores actually do form true sj^ncytial masses as they surround the vessels.

b. The brown type chromatophore

The brown chromatophores differentiate on the yolk-sac at about the same time as the black. They are always somewhat smaller and more delicately formed cells than the black, and react in a slightly different manner. Figures 22 and 26 show several brown chromatophores before the end of the third day. They are paler in appearance and more elongate in shape than the black cells.

The two types of cells are well contrasted in figures 11, 12 and 13 referred to above. The brown cell is smaller, with more delicate processes and is the more rapidly moving of the two. The three figures indicate its condition in embryos of 72 hours.

These pigment cells also wander to the vessel walls and yolk spaces and take on their permanent condition about the fifth day Figure 16 illustrates one of the exquisite brown pigment cells in a yolk-sac of 5 days. The nucleus is still distinguishable in life while it is not in the black cells of this age. The mossy branched processes projecting from all sides give to this cell a most fascinating form.

Fig. 14 A camera lucida drawing of two huge black chromatophores h'ing upon a yolk vessel of a 5 day embryo. The adjacent sides of the chromatophores are beginning to fuse to form a sync3'tium. The direction of blood flow is indicated by the arrows.

Fig. 15 A syncytial mass of black chromatophores forming a sheath about the vitelline vessels. The chromatophores become so thin that the pigment granules are spread apart giving a less intense color. The individual cells are completely lost in the syncytium (3b. DD. ob.).

Fig. 16 A brown chromatophore on the yolk of a 5 day embryo. The cell is coming in contact with two vessels shown in outline. The moss-like processes extend from all sides of the cell.

Fig. 17 A similar cell 12 days old surrounding a yolk vessel. The complex processes from this cell are quite in contrast to the almost smooth border of the black chromatophores of figures 14 and 15. The brown cells never fuse to form syncytia.

Finally, in older embryos the cell body often surrounds a vessel, as shown in figure 17, but the processes persist and project from it in all directions, forming a striking contrast to the more or less smooth outlines f nally assumed by the black cells, figures 14 and 15, as they surround the vessels.

The brown chromatophores do not group themselves together or form a syncytial mass as the black pigment cells are prone to do. They remain individually separated and many really never become associated with vessel walls, but lie scattered on the yolk surface.

In early embryos, from 72 to 90 hours, the brown pigment cells may sometimes, though rarely, get into the blood stream. I have never been able to observe one in the act of entering the current. Yet in a quiescent state they might become surrounded by endothelial cells along with the erythroblasts, and finally be swept away. They might, on the other hand, actually migrate through the porous wall of an early vessel.

The enormous brown pigment cell presents a smooth circular outline as it is carried along in the blood current. On account of its size the chromatophore often meets with difficulties in passing narrow portions of the vascular system. Several such cells were seen in the blood circulation of different embryos during the course of the observations, and when once located in the current the same cell could be seen periodically for a long time as it came around again and again through the vesssel within the field of study. There is no question of the identity of these cells, as their characteristic reddish brown color and coarse granular structure is readily recognized. It is most improbable to think of them as becoming changed into any type of blood corpuscle, and it is doubtless entirely by accident that they occasionally become entrapped within the vessel wall and washed away by the current.

c. Behavior of the chromatophores in specimens with no circulation

The behavior of both the black and brown types of pigmented cell is distinctly different in embryos without a circulation of the blood from that described in the two previous sections for normal embryos.

During the early stages, up to the beginning of the fourth day, the cells wander in amoeboid fashion much the same as in ordinary specimens. In other words, at this time the condition is the same in all embryos since the blood has not begun to circulate in any. At about 72 hours the blood circulation begins in the normal embryos and the pigment cells seem to be attracted to the vessel walls, as already pointed out. If the circulation does not begin at this age, the plasma accumulates in various spaces, chiefly the pericardial sac and Kupffer's vesicle at the caudal end of the embryo. The excessive accumulation of plasma in these spaces causes them to be in many cases hugely distended. The heart in such specimens also becomes a sacular structure filled with plasma which it is unable to pump on account of one or another deficiency in the vascular system.

Large numbers of chromatophores of both types tend to aggregate about these plasma filled spaces and partially cover their walls. The spaces are thus rendered more conspicuous. In some specimens this coating of the distended plasma sacs by pigment cells is most remarkable, but such an arrangement is not invariable and in a number of individuals the pigment cells are irregularly scattered over the yolk-sac with no recognizable pattern or system.

The heart of embryos in which there is no blood circulation is almost without exception covered with chromatophores. These cells often form a perfect sheath about such hearts whether the heart is a plasma filled sac or a mere string. The patterns of these arrangements are illustrated by numerous figures, particularly figures 15 to 20 in the previous paper.

A point of much interest in this connection is the fact that the heart of the normal embryo is entirely free of pigment cells.

The behavior of the chromatophores of the yolk-sac in normal individuals where they tend so decidedly to arrange themselves along the blood vessel walls along with their affinity for the plasma filled spaces in the non-circulating condition would seem to indicate that the chromatophore was attracted by the plasma itself, or some element which it contains. The distended plasma filled heart in the non-circulating cases is covered with pigment as would be expected, yet the solid string-like heart present in many such specimens is also covered with pigment though it of course is entirely empty of plasma. In this last case, however, the string-like heart actually stretches as an axis through the pericardial space which is distended with fluid. The cells arrange themselves around the wall of the pericardium and on reaching the venous end of the heart migrate along it and so cover the heart string or tube in their effort to come in close proximity to the plasma. The distended condition of the pericardium may in this way account for the pigment arrangement along the heart in the cases with experimentally arrested circulation.

The normal heart is constantly pumping the plasma through itself, yet pigment cells are never present in its wall since they are all arranged along the vessels of the yolk-sac. In non-circulating cases many vessels form on the yolk-sac and some become quite well developed, while others actually surround the blood corpuscles of the yolk islands. Such vessels are at times covered with pigment but probably through accident as the pigment is irregularly scattered over the entire yolk-sac. Yet the pigment cells on such vessels never arrange themselves in the definite sheath-like fashion characteristic of the vessel pigment in the normal embryo.

Figure 18 illustrates the lack of arrangement of the chromatophores on the yolk-sac of a 16 day embryo without a circulation; compare figures 15 and 17 of pigment in the normal embryo. Figure 35, see beyond, also illustrates in a striking way the irregular grouping of black chromatophores in the neighborhood of a collection of stagnant blood islands in an embryo of 14 days that never had a circulation of its blood.

All of these reactions cause one to wonder what is the actual function of the pigment cells upon the yolk-sac. The entire egg is rather transparent and their function might be to protect the vessels from the light, yet the vessels are never completely covered and the development of the eggs in the dark is normal but not in any way supernormal.

The pigment cells form such a complete sheath about the vessels in some cases that one might be led to imagine that their expansion and contraction would serve as vaso dilator and contractor. Yet when they are along the vessel wall I have failed to see them contract or expand even when treated with substances such as KCl and adrenahn that violently contract the chromatophores within the embryonic body. These experiments have not been carried sufficiently far since they were directed towards another point, still they indicate at least that the pigment sheath of the vessel wall does not respond as a delicate vaso-motor apparatus.

Fig. 18 A group of brown, indicated in outline only, and black chromatophores on the yolk-sac of a 16 day embryo in which the blood has never circulated. There is no arrangement of the pigment cells on vessels and no real syncytium of black chromatophores as compared with the conditions in the normal embryo.

Wenckebach ('86) found in certain pelagic eggs containing a number of oil drops which invariably floated up that pigment cells often completely surrounded the oil globules, and as he thought prevented these globules from focussing heat or light on parts of the embryonic body. The oil drops in the demersal Fundulus yolk do not particularly attract the pigment cells and they are rarely found to lie against the oil globule.

The function of the pigment in the yolk-sac of Fundulus, if it has any function other than its own existence, is most difficult to determine. The same is true of the abundant pigment in the coelomic wall and other internal structures of many animals.

d. Relationship of chromaiophores to blood and endothelial cells

There has been much discussion in the hterature regarding the relationship of the chromatophores to blood corpuscles and to endothelial cells. The actual relationship of these cells is clearly brought out b}^ a careful study of the living yolk-sac in Fundulus. The cells are completely different and their structures when once established are consistent in their particular type.

The black chromatophores, the brown chromatophores, the endothelial cells and blood corpuslces are all derived from mesenchymal cells which wander away mainly from the caudal region of the embryo during early stages of development. These cells come to lie in the primary segmentation cavity of the yolk-sac beneath the ectoderm and over the periblast syncytium. The mesenchymal cells very soon begin to show certain differential characters in structure and behavior. When certain ones of the cells incline in a definite direction their development progresses continuously along this line.

Observations on the normal living embryos and comparison with those individuals without a yolk-sac circulation lead one to conclude as follows regarding the wandering mesenchymal cells. At the time these cells leave the embryo proper to wander over the yolk, differentiation has proceeded to some extent in the embryo and the mesenchyme cells are probably somewhat limited in their potentialities. Certain of them are derived from the same portion of mesenchyme that gives rise to the intermediate cell mass or future red blood cell forming mass within the embryo. This mass is located towards the caudal end of the embryo and the wandering cells derived from it finally come to lie on the posterior and ventral surfaces of the yolk-sac and form islands of red blood corpuscles. Few if any of these cells reach the anterior regions of the yolk-sac before the circulation of the blood begins. In embryos that never have a circulation the blood islands lie beneath the tail of the embryo and on the ventral yolk surface. The future endothelial cells wander out from the caudal end and side of the embryonic body and finally line up to form vessels in a manner to be described beyond. The pigment cells also wander out from the lateral regions and differentiate into chromatophores of either the black or brown variety.

It would seem that these cells must have some potential differences at the time they come to lie in the yolk-sac, since from that time on they all appear to be in an identical environment. Two cells lying side by side in the yolk-sac above the periblast and beneath the ectoderm would be expected to develop and grow in similar fashion unless there were some internal difference between them. I have thus concluded that the mesenchymal cells which wander in the yolk-sac of the Fundulus embryo must be potentially of four different classes when they first wander out, although all have the ordinary appearance of embryonic mesenchymal cells. Otherwise, it is difficult to conceive why they should develop into four distinct types of cells while all are surrounded by an identical environment so far as is possible to discover. Differentiation in various directions must be due either in the first place to similar cells developing in different chemical or physical surroundings, or in the second place it may result from potentially different cells developing under identical conditions.

The four types of cells are all derived from mesenchyme, just as the thyroid follicles and pulmonary epithelium are derived from endoderm but from different endodermal anlagen, and further than this there is no relationship. Pigment cells and blood corpuscles are perfectly separate and distinct types derived from different mesenchymal analgen and are not in any way transmutable.

2. History of the endothelial cells

The endothehal cells on the living yolk-sac of Fundulus embryos are readily recognized. Their entire behavior in the formation of the earliest yolk vessels may be traced in a manner to fully repay the patient observations necessary in order to follow through the processes.

Among the early wandering cells migrating away from the lateral borders and caudal end of the embryo it is noted that certain ones assume a delicate spindle shape with filamentous processesx^extending from their ends and occasionally projecting from their long sides. In a 48 hour embryo these cells alreadypresent the appearance of individual endothelial cells. They migrate indefinitely for a few hours and then tend to group themselves in more or less irregular collections.

Up to this time no one from mere observation could be absolutely certain that the cells of this rather characteristic appearance are actually to become vascular endothelial cells in all cases. The possibility, of course, exists that the elongate spindle cells may at times round up and then assume the more amoeboid shape of the probable future chromatophore. Yet since the shape of these cells is so characteristic and such shapes are so constantly present, one is inclined to believe that the same cell may actually retain this character until it really becomes a component part of a vascular endothelial arrangement.

Figures 19, 20 and 21 illustrate the region along the side of the embryo's head at 48 hours old. It is in this region that the first large yolk vessel develops. This vessel carries blood from the body of the embryo around a short circuit to reach the venous end of the heart and thus in a way relieves the flow that otherwise would force itself through the small poorly formed vessels in the embryonic body. This vessel is, therefore, of necessity one of the earhest to develop. The three figures, 19, 20 and 21, show variations in the arrangement of the wandering mesenchymal cells in the region of the future vessel.

In figure 19 there is really no definite cell aggregation except along the edges of the head mesoblast as it spreads somewhat over the yolk, yet a few of the cells show the typical spindle shape. Figure 20 indicates a tendency of the mesenchj'mal cells to line themselves in a group exactly along the course of the coming vessel. ]\Iany of the cells in this group give the actual appearance of an endothelial cell after it is fully developed and forming one of the units in a vessel wall. The embryo illustrated by figure 21 shows much the same condition. Very few mesenchymal cells occur between this cell aggregation and the side wall of the head. Lateral of the vessel group the cells are also not numerous and have no system of arrangement.

Figs. 19, 20 and 21 Outlines of the head regions of three living embryos from 48 to 50 hours old, showing different conditions in the grouping of mesenchymal cells on the yolk which later give rise to the large vessel that short circuits blood from the side of the embryo around over the yolk to the venous end of the heart. The future vessel wall is now separate spindle shaped mesenchjTne cells.

This cellular aggregation inay then he regarded as the actual anlage of the vascular endotheliuin of the future vessel. The anlage consists merely of a group of separate wandering mesenchyme cells, and not of a capillary net in any sense.

A slightly older embryo shows a still more definite alignment of the mesenchymal cells and still later presents the appearance of cellular strings or cords as illustrated in an embryo of 67 hours by figure 22. Here the wandering mesenchyme cells have differentiated to such an extent that they are readily distinguishable as black and red chromatophores and elongate endothelial cells.

Fig. 22 A sketch of a 67 hour embryo showing the stage in the origin of yolk vessels in which the mesenchymal cells have a linear arrangement. Early black and brown chromatophores are also shown in the yolk-sac. Od, oil drops.

Early erythroblasts are also seen on the caudal region of the yolksac in such an embryo, but are not shown in the aspect here illustrated.

The endotheUal cells are strung out in various directions and several linear groups are more or less isolated from the rest. The string to be the future large vitelline vessel is not clearly continuous posteriorly, but anteriorly it is well outlined extending towards the venous end of the now forming heart which has not yet begun to pulsate. Here there is no further doubt that these elongate spindle cells are the elements which will make up the endothelial lining of the vessel wall.

There is considerable variation in the rate of development of the yolk vessels in embryos of the same number of hours. Some individuals may be in the condition just described, while others of this age may have already begun to establish a circulation of the blood. Figure 23 shows the yolk region lateral of the head in another embrj^o at 67 hours. In this specimen an incipient circulation has begun and the cord of cells illustrated in figure 22 has now become a small hollow tube sufficiently open to allow the passage of a single file of corpuscles from the side of the embryo around to the venous end of the heart. The individual cells composing the vessel are distinctly seen and their nature is clearly made out with a higher magnification. They retain the same general appearance presented before entering into the vascular arrangement.

Near this vessel is shown in figure 23 a partially formed vascular plexus which is broken in several places and entirely disconnected from the large vein through which the blood is flowing. There is of course no circulation of fluid in this partially formed plexus.

Figure 23 was sketched with a camera lucida at 12 m. and about three hours later at 2.45 p.m., figure 24 was sketched from the same field. The main vessel in accord with Thoma's ('93 and '96) first law of vessel growth has increased in calibre on account of the increased flow and pressure of the circulation. It now permits the passage of three or more corpuscles abreast and is a strongly developed vessel. The former disconnected vascular plexus has grown towards the large vessel and two of the projections shown in figure 23 have now met the wall of the vein and joined with it. One of the first corpuscles from the circulation to enter the plexus is shown in the figure tightly held in the small vessel. Immediately opposite this vessel a sprout is seen growing away from the wall of the large vein.

Figure 25 illustrates the state of arrangement at 6.00 p.m., three hours older than figure 24. The third process from the plexus has here joined the vein and corpuscles are freely passing into the vessels. The sprout from the vein wall is still seen opposite the entrance of the middle vessel. The small plexus arose in loco entirely independently and subsequently became connected with the larger vessel which also arose as we have seen from a group of mesenchjTiial cells.

Fig. 23 The large vitelline vein in an embryo of 67 hours just beginning to permit the passage of blood through its lumen. Corpuscles moving in single file. This becomes the largest vessel of the embryo and arose from the arrangement of the freely wandering mesenchjanal cells of figures 19, 20 and 2L In the lower part of the camera sketch is shown an independent capillary plexus not yet connected with the vein. Ht, heart.

Figure 26 shows another lateral view of the head region of an embryo of 67 hours in which the large vein is filled with circulating corpuscles and the begmning of the same plexus followed in figures 23, 24 and 25 is seen lateral of the vein. At this stage the plexus is entirely disconnected and separated from the vein.

In the formation of the large yolk-sac vein, as well as all other vessels arising upon the yolk, there is nothing to be seen of the forerunning capillary plexus so strongly emphasized by Thoma in the yolk-sac of the chick. There is no selection and dilation of certain channels in the capillary plexus of the teleost's yolksac to form the veins. Here the veins seem to arise in rather definite localities and soon expand into their full form after the circulation has become established.

This method of the formation of vessels was beautifully brought out by AVenckebach ('86) in the early study already referred to so often. He concludes: "Aus diesen Beobachtungen geht hervor, dass Alesoblastzellen durch selbstandige amoeboide Bewegungen die Wande der Blutgefasse des Dotters bilden." Raffaele ('92) later confirmed this observation and further was strongly of the opinion that in selachians and other vertebrates a similar process of vessel building from wandering cells also takes place.

From my present studies on the normal and abnormal Fundulus embryos, I can see no way to doubt that the endothelial wall of the primary yolk vessels in the bonj^-fish is formed by arrangements of wandering mesenchymial cells.

Fig. 2-4 The same vessels 2 hours and 40 minutes later. The main vessel has increased in caliber and two branches of the capillary net have joined the vessel.

Fig. 25 The same vessel three hours later, a sprout is given off opposite the union with the middle capillary and corpuscles now enter all three capillaries. The arrows indicate the direction of blood flow.

Wenckebach's description cannot be fully agreed with in all detail. He thinks, for instance, that cells forming part of the vessel wall are brought by the blood stream. These cells have small protoplasmic processes but they are not in any way to be confounded with the '^definitiven Blutkorperchen. " Such cells have never been observed in the Fundulus embryos and if they exist, which is very improbable, their part in vascular formation is extremely insignificant.

Wenckebach observed the three primarj^ vessels on the yolk to bud and give off sprouts forming other vessels. The wandering cells also formed small separate tubes which later became connected to form a portion of the vascular net. By these methods the complex vascular net of the yolk-sac was finally formed. This agrees closely with what may actually be seen to occur in the embrj^os of Fundulus.

Considerable variation occurs in the position of vessels and a number of actual abnormalities are found in which the bilateral arrangement is completely disturbed. These abnormalities are frequently very instructive for a thorough understanding of the origin and development of the yolk-vessels. Occasionally, a group of cells will form a completely isolated endothelial space which may fail to associate itself with a vessel. Figure 27 shows such an isolated space in a yolk-sac of 90 hours old; solid endothelial tips project from the space, yet it is completely isolated so far as can be determined with the highest power, and at the same time every part of it is clearly and distinctly seen.

\\Tien the early yolk vessels are studied under high magnification, the individual cells may be clearly observed and they are strikingly the same as before they became associated to form the vessel. The cells are not closely arranged but distinct intervals and spaces exist between them and filamentous processes often project far into the lumen and may actually at times fuse with a similar process from a cell on the opposite side of the wall. These filaments thus stretch across the vessel and may even persist after the blood has begun to flow. They are well seen by focussing so as to get an optical section through the cavity. Corpuscles often strike against the cell processes and cause them to wave back and forth as the current flows past.

Fig. 26 Outline of the yolk vascular condition in the head region of a 67 hour embrj'o. Blood is circulating freely in the large vessel but the yolk net of vessels is not j'et connected with the current. Early black and brown chromatophores are also shown.

Fig. 27 An isolated endothelial cavity with solid projecting tips, no connection can be seen with any other vascular spaces. From an embryo of 90 hours.

The cells of the vessel wall thus maintain much of their individuality and may actually separate themselves or loosen away from the small growing tip of a vessel. The tips of the vascular sprouts probably break up or disassociate in this manner to include small groups of corpuscles which may be seen to enter the vessels frcm the yolk surface.

A most instructive specimen for a study of the cellular elements of the vascular endothelium is one in which the circulation has just begun. The vessels in such an embryo are still growing in length and sprouting off branches rather actively. Figure 28 illustrates such a vessel with its incipient branches. Corpuscles are passing through the vessel in the plasma current; one of these, X, is seen harbored behind an endothelial cell at the base of the outgrowth to the right. This corpuscle remained in position for more than one hour, being protected from the current by the projecting endothelial process. Such a condition is frequently seen and conveys some idea of the actual irregularity of the vascular wall.

The cells constituting the walls of the outgrowths from the vessel are changeable in shape and doubtless move their positions to some extent. The cells at the tip of the growing sprout may be seen to send out processes as if they were actually creeping along. The behavior of these vessel walls is strikingly similar to that which Clark ('09 and '12) has so clearly described for the growing lymphatics in the tail of the tad})ole.

Figure 29 shows a sunilar vessel with a projecting bud. The cells of this bud are seen to exhibit a most indefinite arrangement; their processes project across the space and join the cells of the opposite side and none are completeh' elongated or flattened as are the cells of the main vessel wall. The tip cells might still be described as stellate mesenchjTnal cells. The appearance shown in figure 30 is much the same. The walls of these early vascular buds are extremely loose membranous arrangements with irregularities in their surfaces and openings and spaces between the cellular components.

Figs. 28, 29 and 30. Portions of vessels from three yolk-sacs of 6 day embryos. The vessels show blind endothelial sacs projecting from their walls. The constituent cells of these sacs are distinctly seen, and still retain their wandering mesenchymal characters. Filamentous processes from these cells may extend entirely across the lumen and fuse with processes from the cells on the opposite side of the wall. Corpuscles are often entangled in the filaments as well as the spaces between the endothelial cells. X, a resting corpuscle harbored behind an endothelial projection.

These porous or incomplete endothelial walls permit the blood cells to occasionally escape from the vessel cavity and become free within the space of the yolk-sac; or, on the other hand, a growing vascular tip may be observed at certain stages to come in contact with a group of erythroblasts, or actually a blood island unsurrounded by vascular endothelium. The tip of the vessel seems to disorganize to some extent and its cellular elements slowly surround the group of corpuscles which are later taken into the circulation as the current becomes established in the including vessel.

Unfortunately, I have never been able to observe the consecutive steps in any one case of this kind, so that an absolutely positive statement cannot be made at present. Yet numerous observations of the contact of vascular sprouts with groups of uncovered corpuscles and the ends of such sprouts containing corpuscles, as well as other arrangements, would indicate that the behavior of the endothelium in surrounding the naked groups of erythroblasts on the yolk-sac probably takes place about in the manner just outlined.

Figure 32, page 568, illustrates a highly magnified field on the yolk-sac of a 90 hour embryo. This field shows a very interesting composite of the vascular condition at such an age. The rapidly flowing blood current is freely passing through the vessel on the right. A short circuit is forming across below the curve of an arch in the vessel. This small vessel permits only a single line of corpuscles to pass. At this time only one corpuscle has entered and it is caught in the narrow tube. This corpuscle remained fixed for a long time and so enabled a comparison of its structure and appearance with the erythroblasts forming the group just below the huge black chromatophore. The cells of this group are uncovered by endothelium. On the left of the figure a portion of a vascular net not yet connected with the circulating current presents the typical appearance of an early blood vessel formation. The individual cells are loosely associated and the tip projecting towards the right slowly changes position. This tip later approached the group of erythroblasts and finally these cells were all included within the vessel by a process which, as stated above, I was not able to follow definitely.

After closely studying these early vessels in a large number of living yolk-sacs, the observer is able to establish very clearly the actual relationship between the vascular endothelial cells and the erythroblast or early blood corpuscles. The corpuscles on the yolk are always of a distinctly different shape and size, and lie, as described below, in small groups with originally no endothelial cells around them. The groups are later either surrounded by endothelium or taken into the early vessels as already indicated. Nothing has been observed during a long study of these cells on the living-yolk sac of nonnal embryos with a free circulation, or on the yolk-sacs of embiyos experimentally prevented from establishing a circulation, or finally in sections of embryos of various ages, that would indicate even a tendency of endothelial cells to change into any type of blood corpuscle. The endothelial cell, whether in the free and wandering mesenchymal state or constituting a part of the vessel wall, presents a typical shape and clear appearance entirely distinct from the wandering erythroblasts.

In observing the early yolk vessels certain things may be seen which are most miportant in interpreting sections supposedly showing the transition of vascular endothelial cells into erythroblasts or primitive blood cells. Frequently, one or more corpuscles become entangled within the spaces and filaments existing between the cells of the vessel wall. Other corpuscles flowing in the current strike these entangled ones and beat against them sometimes for hours before they become disentangled from the vascular pits and holes, to flow again in the current. This is an extremely common sight during the early hours of the circulation of blood in an^^ of the yolk vessels.

It may now readily be imagined that if the embryo was killed and fixed while the corpuscles were tangled in the spaces of the vascular endothelium, a study of sections might produce the impression that the cells of the vessel wall were "protruding into the lumen and assuming the typical characters of primitive blood cells," according to the description of many that imagine the occurrence of such things. I have previously offered another explanation of these appearances and many phenomena observed on the living yolk-sac bear out the point of view. It may sometimes happen when the vascular endothelium encloses a group of primitive corpuscles that one or more of these future blood corpuscles come to lie in the plane of the vessel wall, or may actually seem to form one of the cellular components of the wall.

When the circulation begins, however, this cell becomes loosened away from the wall for mechanical reasons, the lack of long processes, etc., and projects into the Imnen to be finally washed away. Any one may readily observe such occurrences who will study the living yolk-sac of Fundulus with a high power microscope and a strong condenser so as to use a darkened field.

All of these observations lead one to conclude that the only connection between vascular endothelium and primitive blood cells is one of association. The endothelial cells never metamorphose into blood cells. It is important here to recall the fact previously emphasized by the author that in those specimens in which there is never a circulation of the blood or plasma the vascular endothelium develops in a perfectly normal manner in the aorta and other intra-embryonic vessels, as well as in the vessels on the yolk-sac, yet in none of these does one find any appearance indicating a tendency of the lining cells of the vessel wall to change into any type of blood cell. Numerous other details from my notes might be enumerated which would bear on this question, but sufficient care has been taken to definitely establish the above crucial points as facts. This may not of course hold for all types of animals but it does for those studied.

I have seen a number of sections on which other investigators have based their claim that vascular endothelial cells do change into primitive blood cells and although inclined towards the acceptance of such a view from a mere acquaintance with the literature, a study of such material has convinced me that the negative interpretation is equally plausible in all cases.

3. Blood corpuscles on the yolk-sac of teleost emhryos

In all meroblastic eggs except those of the teleost a great sheet of mesoblast is found extending over the yolk as the socalled peripheral or ventral mesoderm, or subvitelline mesoderm. It is this peripheral mesoderm that gives rise to the blood islands of the yolk-sac in selachians, reptiles, birds and mammals The teleost, however, presents a unique case in that the ventral mesoderm does not spread out over the yolk but is included within the embryonic body. The differentiation of this mesoderm within the embryo is much the same as that of the peripheral ventral mesoderm in the yolk-sac of other groups. Thus in the bony-fish the great bulk of blood formation takes place within the so-called intermediate cell mass, the probable homologue of a portion of the yolk-sac mesoderm of other vertebrates.

The intermediate cell mass first described by Oellacher ('73) and later fully studied by Ziegler ('87), Swaen and Brachet ('99, '01), and numerous others, is derived from the primary lateral plate mesoderm, separating away from the median border of this plate. The cell masses of the two sides remain apart in some species and form the future cardinal veins and red blood corpuscles, wiiile in others the two masses unite in the median line to form the conjoined cardinals or stem vein, which is loaded with the primitive erythroblasts — the red blood anlage.

All recent workers on the development of the blood in the bony-fish have considered this intra-embryonic blood formation as being the only source of blood cells in these animals. Several authors, however, Swaen and Brachet among them, have recognized that the cells of the stem vein may become so packed and crowded within the embryo that masses of them are directly pushed out laterally upon the yolk. They have also thought it possible that a very few cells might wander upon the yolksac and there form blood, but this has been considered questionable in all cases. No one has recognized the actual occurrence of blood islands upon the teleostean yolk-sac. Even Wenckebach in his study of living embryos, although he made so many important observations on the formation of the periblast and yolk vessels, entirely overlooked the early or primitive yolk-sac blood cells. This is probably due to the fact that he studied only normal embryos. In embryos without a circulation of the blood one observes the yolk islands much more readily as the cells finally become filled with haemoglobin and present a bright red color. When they are once located in these experimental embryos it becomes much easier to trace them in the younger normal individuals, and finally the observer readily locates these cells and may follow completely their migrations and association to form the yolk-sac blood islands.

The arrangement of these wandering cells in the yolk-sac, figures 7, 8 and 10, suggests in a way the regions of growth of the yolk-sac mesoderm in other meroblastic eggs. About the caudal region there is an 'area opaca' formed by the great number of wandering cells, while around the head end the scarcity of mesenchimal cells might be considered an area pallucida.

All of the yolk-sac blood islands in the Fundulus embryos are formed from certain of the early wandering mesenchymal cells on the yolk. During the early wandering stages when the future endothelial cells have a spindle shape and the future chromatophores are large amoeboid cells, other mesenchjTiial cells on the yolk are small and more or less circular in outline. These small circular cells move slower than the other types and throw out short thick pseudopod-like processes.

"Whereas the spindle shape cells wander away from the embryo along its entire lateral border as well as the caudal end, and the large amoeboid future chromatophores have almost an equally extensive place of origin, the small round cells wander out only from a limited region. The earhest ones of these to be seen are near the caudal end of the embrv^o before the tail fold has completely separated the caudal end from the yolk surface. As the tail is moulded free from the yolk-sac, the point of out wandering of the circular cells follows the place of union between the ventral wall of the tail and the yolk-sac. Just at this place the mesenchyme of the embryonic body extends itself out on to the yolk as free wandering cells.

In a study of sections this mesenchyme is found to lead directly to the end-bud, Endknospe, which may be considered to represent the blastopore lip. The cord of mesoblast which has been designated as intermediate cell mass leads caudally to the end-bud which is well out in the tail. The ventral cells from this mass wander away into the yolk-sac from the extreme caudal position and other cells also wander away laterally from the intermediate cell mass. Figure 8, the tail end of an embryo 56 hours old, illustrates very well the place of outwandering of the round cells. Few, if any such cells wander out from the lateral borders of the embrv^o in regions more cephalad than this.

This confined local origin of the round cells and the period at which they wander out, along with their general appearance, lead one to believe that these cells are actually derived from the same general mass or group of cells which goes to form the intermediate cell mass or red blood anlage within the embryo. All of these slowly wandering circular cells finally differentiate into red blood corpuscles, as described below, just as cells of the intermediate cell mass will finally do.

In this connection a most instructive defect is frequently found among embiyos developing in the stronger solutions of alcohol. JVIany such embryos are of the common short type with their tails split, cauda-bifida, figure 4 of the previous paper. This defect is due to the fact that the germ-ring descends over the yolk in a slow or arrested fashion and may never succeed in fully enclosing it. The caudal end of the embryo is thus divided and the two tail moieties remain spread apart laterally along the line of the germ-ring. This condition renders the caudal portion incapable of including all of its usual median tissues and such cells extend past the angle of the spht and lie between the two parts of the bifid tail. The interesting thing is that the cells constituting the blood-forming intermediate cell mass lie in just this position.

Figure 33, a diagram, illustrates the embryonic body with a bifid caudal end. The great lake of blood corpuscles is situated beyond the angle of the split tail. Such an abnormality liberates the future blood forming mass from the body of the embryo and the mass spreads posteriorly over the yolk surface, yet not in a diffuse manner since it maintains its densel}' packed cellular structure. We might consider that here the evolutionary events are reversed. The blood anlage in the prunitive fishes, the selachians, is spread over the yolk in the area opaca. In the normal teleost this primary yolk-sac blood anlage has been included within the embryonic body and localized in the intermediate cell mass, ^\^lile in the abnormality here considered the intermediate cell mass is again outspread upon the yolk somewhat suggestive of the old ancestral selachian arrangement. This abnormality, in other words, may give some notion of the actual incorporation of the primitive blood-forming mesoderm of the yolk-sac into the body of the teleost embryo.

Fig. 31 A group of six early erythroblasts unsurrounded by endothelium on the yolk-sac of a 90 hour embryo. Short amoeboid processes project and the cells move very slowly

Fig. 32 A camera lucida sketch of a microscopic field on the yolk-sac of a 90 hour embryo. All four derivatives of the wandering mesench\Tnal cells are shown. The circulation is established in the vessel to the right and the current follows the direction of the arrows. To the left is a small vessel not yet connected with the current; its wandering endothelial tip is approaching a group of erythroblasts still uninclosed by endothelium as they lie near the huge black chromatophore. A brown chromatophore is seen on the large vessel.

Fig. 33 An outline of a short 6 day embryo from a strong alcohol solution. The embryo presents a cauda-bifid condition and the normally intra-embryonic blood-forming mass is represented by a densely packed expanse of corpuscles outside the body of the embryo and spread upon the yolk.

The situation may be pictured as follows. In the reptiles and birds, for example, the peripheral mesoderm is outspread over the yolk and in it differentiates the blood islands of the area vasculosa. The peripheral mesoderm in Fundulus and teleosts generally does not become outspread over the yolk, but is concentrated into a median cord or mass within the caudal half of the embryo. Yet even here there is actually a tendency for the cells of this mass to be attracted to the regions of the yolksac, and during the early stages of development many cells separate from the mass and wander freely on the yolk. The extent of such wandering is probably variable in different species. Yet in all eggs with an extensive vitelline circulation the wandering of these future red blood cells probably takes place to a considerable extent in spite of the fact that the cells have been so generally overlooked. A fact easily accounted for when one realizes that most of the studies in blood origin in teleosts have been confined to sections of the embryos of the salmon and trout. Sections are extremely slow in revealing the significance or even existence of wandering cells in development.

We may now consider the individual wandering cells and their subsequent differentiation. Figure 31 shows a group of six such cells. They are small when compared with the huge chromatophores but are about as large as the endothelial cells, though completely different in shape, texture and behavior. Figure 32, a camera lucida sketch, serves well to illustrate the relative sizes of the different type cells on the yolk-sac. In this figure are shown all four t;^npes, the enormous black chromatophore, the very large brown chromatophore, the delicate elongate endothelial cells of the vascular wall with their filamentous processes, and the almost circular erythroblast, small when compared with the first two types, but as large or even larger than the endothelial cell.

Figure 34 was drawoi from an embryo that had been fixed in such a way as to render conspicuous the cell outlines of the ectodermal layer of the yolk-sac. Below the ectoderm groups of erythroblasts forming blood islands are shown, and the extremely small size of the erythroblasts in comparison with the enormous dimensions of the ectodermal cells is most striking.

As mentioned before, the erythroblasts tend towards a circular shape but send out short processes, while they move in a sluggish amoeboid fashion. The cytoplasm of these cells is slightly greyish and not so perfectly transparent as that of the spindle cells; this difference between the two types is not so marked during the early stages but is readih' noticeable in embryos of 80 or 90 hours.

The future erythroblasts very slowly wander away from the tail region of the eml3ryo and down the posterior surface of the yolk to reach its ventral surface. At various places on the posterior and ventral yolk surfaces the cells collect into groups and become less active, although their movement does not entirely cease. These groups constitute early blood islands. They are at first not surrounded by endothelial cells but later become enclosed or taken into the ends of the incipient vessels as briefly described above.

Fig. 34 A camera lucida outline of the caudal region of the yolk-sac in a 96 hour embryo with the ectodermal cell borders made visible by mercury fixation. An island of blood corpuscles is indicated by small circles which give some idea of the minute size of the erythrocytes as compared with the huge ectodermal cells.

When the circulation first begins on the yolk-sac of a normal embryo a great many of these islands are present, but are more or less isolated or out of the channels through which th^ fluid is flowing. The taking up of the islands by the circulation is most interesting to observe. At first those cells near the tail of the embryo, which are enclosed by endothelium, are taken. A few of the corpuscles are shaken by the current and these strike against the other members of the group mitil all become loosened and move sUghtly to and fro; finally one or two are suddenly washed away, then others follow — few at first, mitil the entire group loses its stand and is swept away by the current. One and then another of the islands may be seen to perish in this manner before the irresistible force of the tiny stream.

Yet even after the yolk circulation is fairly well established a number of islands of the romid cells maj" still exist unsurrounded by the endotheUal vessels. Figure 32 shows such a case. A well established current flows through the vessel to the right, while to the left is an incipient vessel not yet connected with the current. In the center of the figure is a group of corpuscles below a huge black chromatophore. These round cells constitute a blood island still unenclosed by vessel endothelium; in the course of a few hours, however, they too will become enclosed by a vessel and subsequently be included among the circulating blood cells.

The early erythroblasts which are in this manner included within the circulation assume a circular outline as they float in the current. In the previous paper, figures 31 and 32 of cross sections through the intermediate cell mass, and figures 34 and 35 of cells m a yolk island, all from a young 72 hour embryo, illustrate the definite circular outUne of these cells. In life they may be carefully observed, in the small vessels where a single cell passes with difficulty, and arc here seen to be globular in shape and to retain their slight amoeboid movement. The form of the early erythroblast is readily changeable for the first one or two days after entering the current. After two or three days, that is, in embryos five or six days old, the cells in the blood current assume a typical erythrocyte appearance, becoming elongate and elliptical in shape when seen from one position, while they are thin in profile view. The}' are now ellipsoidal nucleated red blood corpuscles. At about the time they begin to change from the globular to the elipsoidal shape, the accumulation of haemoglobin takes place and the cells begin to show a typical straw color.

It may then actually be seen in the living that the early or primitive erythroblast is really a more or less globular amoeboid cell without haemoglobin and resembling more closely a lymphocyte or early leucocyte than a fully formed erythrocyte. This is probably true of the early stages in many cases of cytomorphosis, yet the globular amoeboid cells of the yolk islands are not indifferent 'primitive blood cells' in the sense Maximow ('09) has concluded, but they are definitely future erythrocytes.

This point is established by a study of these cell groups in the normal as well as in embryos without a circulation of the blood. In the latter individuals the isla4ids arise by the formation of local aggregations of the early wandering cells in an exactly similar manner to that described for the normal embryo. In fact, the observer cannot distinguish between the two specimens in many cases, yet one fails to establish a circulation and the islands are thus enabled to retain their positions on the yolk-sac.

The constituent cells of such a permanent island may be observed from time to time or continuously, and will be found to pass through changes exactly identical with those which take place in the island cells that become swept into the blood stream of the normal embryo. They are for a few days globular in shape, but capable of slightly changing their form and sending out short pseudopod-like projections. When the embryos are five or six days old the cells in these islands then become flattened ellipsoidal corpuscles and attain a haemoglobin content exactly as in the normal embryo. The blood islands now appear bright red in color and are quite conspicuous on the yolk-sac where they permanently remain. The globular colorless cells are thus seen to differentiate directly into the typical ictheoid erythrocyte.

From a study of the living embryos alone one could not, of course, be certain that all of the cells of these islands had differentiated into red blood corpuscles. However in the previous study of the non-circulating embryos, I have examined a large number of yolk islands completely and have never seen any type of blood cells other than erythrocytes in such a position. The same is true of the cellular 'products of the intermediate cell mass; in hundreds of sections studied with the oil immersion not one case has been found of a lymphocyte or leucocyte arising from a cell of the original intermediate cell mass. It is thus concluded that this mass is the red-blood anlage of the teleost and the wandering cells of the yolk-sac which are very probably derived from the same mesodermal stem that forms. the intermediate cell mass are likewise a portion of the red blood cell anlage.

The embryos that fail to develop a circulation are most important material for the study of these questions of relationship among the different tj^Des of blood cells. The observations on the living show definitely the qualities of the early blood forming mesenchymal cell in its assumption of the globular slowly wandering type, which would fully satisfy the descriptions of Maximow's 'primitive blood cell' as being closer in appearance to a lymphocyte than to a red corpuscle. When one follows these supposedly indifferent primitive blood cells which are claimed to possess the power of differentiating into either a leucocyte or an erythroblast, he invariably observes them to differentiate only into erythrocytes. Although all cells of the early islands are distinctly visible, they are not mixed in type, but are of one class. A long study of these early islands in sections also fails to reveal any other than red blood cells. The leucocytes are not very numerous in the blood, yet they should be seen if present in these islands, as they are found in other positions and are well represented in the blood of the adult Fundulus.

The further history of the cells in the stagnant masses on the yolk-sac of an embryo in which the blood has never circulated is instructive in several ways. In the first place, all of these islands seem to become surrounded by endothelium, yet the endothelial arrangement cannot be completely traced in every case and it rarely extends much beyond the island mass so far as one can observe in life.

Pigment cells are irregularly scattered in the neighborhood of the islands, as they are throughout the yolk region. The chromatophores do not, however, assume any arrangement with reference to the islands of cells and as mentioned above they retain their original condition as separate cells instead of fusing into s^Ticytial masses, as is the case in specimens with a free blood circulation.

Fig. 35 The arrangement of red blood corpuscles in more or less connected endothelial sacs on the ventro-lateral surface of the yolk in an embryo 14 days old that had never had a circulation of fluid in its vessels. All of these blood cells wandered away from the caudal body regions of the embryo during early stages. The chromatophores are irregularly scattered among the old bloodislands. Od, oil drop.

Figure 35 illustrates a group of blood masses on the yolk-sac of an embryo 14 days old. In this specimen there had been absolutely no circulation or movement of the blood fluids within the vessels at any time. This is most important to know in the case of all specimens without a circulation at the period they chance to be examined. I shall return to this point below. The cell groups in the specimen without a circulation are arranged somewhat like a vascular net in the region illustrated by figure 35, yet they present a deadly still appearance as contrasted with the lively movement of the coipuscles within the j^olk vessels of a normal individual. The erythrocytes forming these islands are brilliantly red in color and their shape and size are apparentlynormal.

The blood corpuscles are thus found to differentiate in a typical fashion and to retain their haemoglobin reaction for a long period without having circulated in the vessels. The function of an erythrocyte would thus seeyn to be entirely independent of its circulation so far as its capacity to form oxyhaemoglobin goes. These cells are also able to accumulate oxygen within the intermediate cell mass in its central position in the embryonic body. The embryo from which figure 35 was taken had lived 14 days without its blood having circulated, which is about the period required for the young fish to hatch and become free swimming.

In older embryos the erythrocytes begin to degenerate and in many they lose their red color, the haemoglobin probably breaking down. The islands on the yolk then become pale in color and finally almost white, as if the cells were dead. The color of the blood cells seems to fade within the embryo earlier than on the yolk-sac as a rule, probably due to the better chances of obtaining oxygen on the thin yolk-sac than in the thicker embryonic body. The non-circulating specimens often continue to live for a long time even after the blood cells have lost their color. Some such specimens may exist for more than 40 days, which is a very long time considering that the normal embryo may hatch when from 11 to 20 days old. The specimens without a circulation are always weak and delayed in development and of course never succeed in hatching from the egg membrane.

In a study of the embryos treated with weak alcohol solutions, one very frequently finds cases in which the circulation of the blood may start almost normally and finally stop permanently, although the embryo continues to live. Other embryos may fail to establish a circulation at the proper time and yet may develop a freely flowing circulation of their blood some days later. Again, an embryo may have a fairly normal circulation and for some reason lose it for a few hours, or for one or two days, and then regain it, at first slowly and finally in a fairly strong fashion. All three of these phenomena have also been observed in eggs developing in ordinary sea-water when they were not properly separated so as to allow free respiration. The egg membrane is covered with long hair-hke filaments which become entangled with those of neighboring eggs and in this way masses become closely packed. The central eggs of such a mass develop in a poor atmosphere and go more slowly than their neighbors on the outside of the mass. These centrally located eggs show many abnormal and arrested conditions of much the same type as may be obtained by treating the eggs with various injurious solutions.

The changeable states in the circulation offer many pitfalls for one attempting to determine the sites of origin of blood cells in the non-circulating embryos. Old embryos are seen in which there are beautiful blood islands on the yolk-sac and great clots of blood in the head or other unusual position. The heart is very frequently completely loaded with corpuscles, and yet there is not the slightest movement of the blood cells or any sign of a circulation at this time. The heart itself may be pulsating feebly or even practically stopped.

Another source of blood movement which is slight, yet to be guarded against, is that due to the muscular twitching of the embryo's body. This movement may frequently serve to push cells from the intermediate cell mass out on to the j^olk-sac, but usually by way of the vessels. These dangers are to be taken seriously in experiments of this kind. Since one is able to be absolutely certain that the blood never circulates in a great number of embryos, only such embryos should be considered in a study of blood origin. During a study of this exact problem now extending over four spawning seasons, I have seen blood in almost every conceivable position in embryos without a circulation at the time of the observation. The accumulation of blood is more frequent in certain positions and regions than in others. The venous end of the heart is a most common place for a clot, the sides of the head, the large vessels of the yolk just lateral to the body, and various places on the anterior and lateral yolk surfaces.

When, however, the experimenter collects a number of embryos that have really never experienced the slightest flow of their blood, the case is very definite. No blood clots ever occur in regions other than the 'intermediate cell mass/ within the embryo, and the islands on the caudal and ventral yolk surfaces which have been formed as described by the early wandering cells that rnigrate away from the caudal region of the embryo. All embryos lohose history for lack of circulation throughout their existence is actually known show the blood pattern most consistently, there being of course a certain amount of variation in the extent and position of the yolk-islands but not enough in any case to confuse the problem.

These observations may readily be made by any observer, but can only be made in a reliable fashion with the high power microscope and strong condensers so that the Ught may be sufficiently regulated to observe the most transparent cells. The movements and differentiation of these cells should be carefully followed through every step in a number of cases, in order to fully appreciate the significance of their position and behavior.

The cells may be seen even with an ordinary binocular microscope to some extent, but the arrangements for light regulation and the magnification are insufficient for determining the important details. After the red blood cells have formed, they are readily located even with a low power yet such an examination could only determine their places of origin provided the embryo has been carefully watched with a high power magnification to make certain that it has had no blood flow.

The condition of the yolk-sac mesenchyme must be fully understood and must always be considered in interpreting the origin of blood-islands and clots. For example, clots seen at the venous end of the heart or on the extreme anterior surface of the yolk must be most cautiously considered, remembering the scarcity or even absence of the wandering mesenchyme in these regions. Clots in such places probably always result from a partial circulation of short duration and there is abundant evidence to support such a view.

Although the future red blood cells migrate upon the yolk in their early mesenchymal stages, after they once group themselves and differentiate into erythrocytes their powers of wandering become very much limited if they exist at all. I have never seen anything to indicate that a fully formed erythrocyte was capable of automatic migration. Yolk-sac blood islands of all ages have been examined in great numbers, but never has an erj^throcyte appeared wandering away from such an island into neighboring regions. This fact is most important in the study of the blood-islands in the non-circulating specimens.

When the sUghtest flow does exist for any length of time, there is a definite tendency, as mentioned above, for the blood to accumulate in certain sinuses and vessels. The positions of accumulation vary somewhat with the stages at which the circulation ceased, as well as the manner of stoppage of the flow, whether it was gradual or sudden.

When the circulation stops during early stages, there is a great accumulation or massing of the blood over almost the entire ventral surface of the yolk. In other words, there is a hemorrhage or bleeding into these spaces or vessels until no more blood is left in other regions of the embryo, the heart gradually becomes empty of corpuscles and no longer passes them along. The packing of the yolk vessels probably clogs or blocks the circulation so that it ceases. Again, the circulation may stop more suddenly and the venous end of the heart or the entire heart may be seen packed with corpuscles while the vessels immediately entering and leaving it are comparatively or entirely empty. In older embryos there is the tendency to accumulate red cells in the vessels of the head so that brilliant red clots are frequently seen in these positions.

In all cases it is interesting in these mdividuals in wliich the circulation has ceased at one or another period in development, and doubtless for different reasons in different specimens, to observe the way in which the blood sooner or later accumulates in one or another vascular space and does not remain uniformly distributed throughout the vascular system. Only when the heart is suddenly stopped and the blood quickly fixed by some strong killing fluid does one get a good pattern of the vascular system loaded with coipuscles throughout most of its extent. In rare cdses, three during the present summer in some hundreds of embryos examined, will a specimen without a circulation at the time observed show almost all of its vessels loaded with blood cells, and this is probably due to a slowing down gradually of the circulation on account of the heart itself which finally stops with the vessels in a balanced state.

The study of the yolk-sac in the living embryo enables one to observe every phase in the development of the red blood corpuscles from the early time when they wander as amoeboid 7nesenchyme cells to collect into groups of globular cells with short processes, the 'primitive blood cells^ of descriptive histologists , to be later surrounded by vascular endothelium, and then to change from the globular wandering cells into the flattened ellipsoidal erythrocyte loaded with haemoglobin, and finally freely floating in the current of the blood stream. The fully formed corpuscles apparently become incapable of independeiitl}^ migrating even when not carried by the circulation.

Discussion and Conclusions

In the previous paper on the origin of blood and endothelium, a somewhat full discussion of the problems of blood formation in the teleosts and other vertebrates was entered into. A consideration of the questions of origin and development of vascular endothehum was also undertaken in the light of the results there presented and the more recent general literature bearing on this subject. The experimental results then contributed seemed in the light of the past hterature to render highly probable, if not to actually prove, the polyphyletic origin of the various types of blood cells, as opposed to the now extrernely improbable monophyletic theorj^ of origin of blood cells and vascular endothelium. For a general consideration, the reader is referred back to these discussions.

A number of particularly significant points are brought out in the present study of the living normal and experimental embryos which bear directly on several of the past theories and speculations regarding the origin of vessels and blood. Only these special points will be briefly considered and analyzed at this time.

In the first place, the writer cannot resist the impulse to highly recommend that all students of haematogenesis and vascular origin spend some time at least in a study of living mesenchymal cells and their cytomorphosis. Such a study will soon convince one of the great disadvantages under which an investigator labors in attempting to solve the origin of blood from observations on dead material in serial sections. The problem becomes so simplified and devoid of laborious uninstructive technique that it seems almost superficial. One may learn as much from the living yolk-sac in an hour of careful study as in almost a week's perusal of sections. Most important is the fact that certain things may actually be seen to occur that sections could scarcely stimulate the mind to imagine. The only disadvantage is that the worker may be led to wonder whether so apparently simple a problem is actually of scientific importance. Fortunately this mental state is soon passed over on realizing the necessary care and precaution which must be taken in following the movement and changes in the Uving cells.

Each cell must be recognized as a living complex and the observer will realize the importance as well as the difficulties of thoroughly understanding and interpreting correctly its manifold changes and behavior. Material which to some extent allows such a study is often available. The Fundulus yolk-sac, however, is exceptionally adapted to this study on account of the beautiful simphcity of its structure, as well as the remarkable clearness with which each cell may be observed.

An investigation of the Fundulus yolk-sac readily supplies a crucial answer to the old question regarding the relation of the blood vessel lumen to other body cavities and spaces. Ryder ('84) was right in describing the blastocoel of the bony-fish as remaining an extensive cavity for some time. This is the space between the ectoderm and yolk and is identical and continuous with the cavity which arises very early beneath the blastoderm and above the yolk periblast. Agassiz and Whitman ('84), as well as Ryder ('84), Wilson ('90), and others, have identified this correctly as the cleavage cavity, the blastocoel. Later in development, the blastocoel extends over the yolk, forming the space into which we have seen the free mesenchyme cells wander, and finally within this space groups of these cells form the yolk-sac vessels.

Wenckebach ('86) described very clearly the origin of vessels from the free mesenchyme within the segmentation cavity. My study of a somewhat similar yolk-sac confirms the main points brought out by Wenckebach and all serve as crucial facts in support of the early theory advanced by Biitschli ('82) in his Die phylogenetischen Herleitung des Blutgefassapparates der Metazoen." Biitschli held that in the Metazoa the lumen of the blood vascular system was derived from the blastocoel. Later, Hubrecht ('86) supported the same standpoint from his studies on Nemertines. Hubrecht also found wandering cells playing an important role. Ziegler ('87) gives a most careful analysis of the continuity of the vascular lumen with the blastocoel in his studies on the development of the bony-fish.

The foregoing description and figures of the origin of vessels on the yolk-sac of Fundulus leaves no doubt that the vascular lumen in these animals, coenogenetically at any rate, is continuous with the blastocoel or primary body cavity and is in no way related to the coelom.

Almost twenty years ago, Felix ('97) advanced the opposing theory that the vascular lumen was really a localized or separate part of the secondary body cavity, or true coelom. The many decided objections to this theory from the standpoint of comparative anatomy, the presence of blood vessels before the acquisition of a true coelom in the animal kindgom, and the numerous embryological contradictions in its path were pointed out in the discussion of this matter in the previous paper.

Very recently Bremer ('14) has advocated the theory of the origin of vessels as parts separated from the coelomic cavity, or strands of cells from the coelomic epithelium. In the first place, the material on which his investigation was based, early human embryos, will scarcely permit such generalizations. At least more suitable material could be found for the analysis of this problem. Further than this, his consideration of the questions involved does not lead one to form a definite idea of the exact direction he considers his evidence to lead. He credits Biitschli ('82) with having originated the coelom theory that accords, so he thinks, with his evidence. This is entirely incorrect, at Blitschli's theory is exactly on the other side. The morphology of the yolk-sac of the chick, sheep and numerous other animals, as the Uterature of the subject readily shows, is entirely out of accord with such speculations. The yolk-sac of the bony-fish shows this view to be really impossible and there should be no longer any doubt that vessels arise from loose and wandering mesenchymal cells in many animal species, and certainly not from ingrowths from the coelomic epithehum in any species.

The formation of vessels on the yolk-sac of the teleost further limits the generalization of the origin of larger vessels from capillary nets. Thoma ('93) in his masterly study of the vasculogenesis of the yolk-sac of the bird, held that the first vascular spaces, the rudimentary capillaries, were formed by the secretory activity of the cells forming their wall." These capillaries formed an extensive net and the arteries and veins arose secondarily and differentiated from the capillaries on account of the flow of blood set in motion by the beat of the heart. The anlage of the vascular system was the capillary.

These principles of Thoma are not, however, applicable to the development of vessels in the embryonic bony-fish. The aorta arises as one or two vessels independent of any flow of blood or the existence of a capillar}^ net. The first vessels on the teleostean yolk-sac are the large vitelline veins, as described by Wenckebach, and the median vitelUne vein or the net of vessels in its place. These large important channels arise entirely independently and separated from the capillary net if such exists at the time. They also develop entirely independently of the blood flow, and not as a result of the pressure due to the heart beat. The capillaries and other vessels in many cases arise separately or away from these primary vessels and finally connect with them in a way similar to the connection formed between the Randvene and the venous end of the heart. Other capillaries and small vessels arise as buds or sprouts from the first formed veins on the yolk-sac.

More recently Evans ('09) with a very efficient and delicate method of injection has shown many of the larger intra-embryonic vessels of the chick embryo to develop from a foregoing capillary network. He found the same principles of development that Thoma had observed on the yolk-sac to hold for the development of certain vessels within the embryo. These principles of vascular development Evans thought applied to vertebrates generally, but such is certainly not the case, the large vessels of the teleost embryo arise directly from associated mesenchymal cells and are not preceded by a capillary net.

His evidence was derived from injected vessels and could not justify the statement, p. 512, of The presence always in the embryo of a united vascular sj^stem" — "sl single branched endothelial tree." Such a united vascular system is rather late in its establishment in the fish embryo and there is no "single branched endothelial tree" present when the first blood vessels are formed. These facts may readily be demonstrated on the living embryo bj^ direct observation.

The vessels of the yolk-sac and several of the larger vessels within the body of the teleostean embryo form independently of any foregoing capillary network. In the teleost, then, the anlage of the vascular system is not the capillary hut the mesenchymal cells which directly give rise to the chief arteries and veins, as well as to numerous groups of isolated capillaries. Other small vessels and capillaries grow as branches or sprouts from the arteries and veins.

Thoma advanced three laws for the formation and growth of vessels. The first law was considered the most important, but rather destructive evidence is thrown against it by the present study. The law may be stated as follows: "The increase in the size of the lumen of the vessel, or what is the same thing, the increase in the surface of the vessel wall, depends upon the rate of the blood current." The vessel increases in size when the rate is exceeded, becomes smaller when the rate is slowed, and disappears when the flow is finally arrested. Thoma ('96) states: "This law which brings the growth of the surface of the vessel into dependence upon the rate of the flow of the blood is, I consider, the first and most important histo-mechanical principle which determines the state of the lumen of the vessel under physiological and pathological conditions."

Thoma again states this principle thus: In development the vessels in which the blood stagnates degenerate, and in those in which the rapidity is too great the lumen in enlarged."

No one could fail to admire the splendid manner in which Thoma attacked the problem of vasculogenesis in the yolk-sac of the bird, or the ingenious way in which he attempted to analyze the problem and deduce his three laws of histo-mechanical processes. Yet the ' ' First and most important histo-mechanical principle" does not apply to the development of vessels in Fundulus embryos where there is no circulation of the blood. Many vessels grow in size or "what is the same thing, show an. increase in the surface of the vessel walV without any rate of the blood current. " The aorta in old embryos that never had their blood to circulate and in which the heart is actually a solid string of tissue, grows and attains a well developed lumen and a wall lined with endothelium and surrounded by concentric fibers of connective tissue as is shown in figure 49 in the previous paper, drawn from such a specimen. This vessel is very slow to degenerate, in fact, it shows no sign of degeneration and actually persists as long as the embryo is able to exist without a circulation, for 30 days or more. Vessels also develop upon the yolk-sac without ever having a fluid to circulate through their lumen. Other vessels are developed around the blood cells of the intermediate cell mass and the yolk-sac islands and in such vessels 'the blood stagnates' from the first yet the vessels degenerate very slowly, in some cases scarcely at all.

In still other cases the blood may have circulated for a while and then stopped for some time, but the vessels do not degenerate as is proven by the fact that the circulation through them may again be resumed. Such a sequence of events may be occasionally observed in the experimental embryos. The function of the vessel as a blood conductor, therefore, seems in these embryos of Fundulus, both the early normal and those without a circulation of the blood, to have little if anything to do with its early development and not much effect on its ability to survive.

On the other hand, when Thoma has a straight normal case, the lumen may readily be seen to increase in size with the rate of flow. Yet in the entire absence of this action the vessel is still capable of increasing in size and it becomes questionable whether the rate of flow is evet an actual cause of size increase bej^ond mechanical stretching.

These facts are most significant in a consideration of the influence of function on growth and development, auto-differentiation. Here it is seen that the structure both grows and develops in entire absence of its function. 'In normal cases the function of the vessel as a blood conductor exerts more likely a physical rather than a biological effect on development.

Thus the development of blood vessels on the yolk-sac of the living Fundulus embryo proves that the capillaries are not universally the anlage of the arteries and veins, but that these larger vessels may arise directly from wandering mesenchyme cells. Such arteries and veins may grow and persist without a circulation of the blood through their lumen and even though stagnant masses of corpuscles may crowd the vessel cavity.

The development of vessels in Fundulus also directly disproves the claims made by Sobotta ('02) that the vessels in the teleost grow over the yolk entirely as branches from those near the embryo and without the wandering cells taking part. This assumption is probably due to the difficulty of estimating the part played by wandering cells from a study of serial sections. From a study of the living yolk-sac there is no question of the major part played by the wandering cells* in the origin and formation of vessels on the yolk.

Sobotta also advances the opinion that the entire yolk vessels may sprout from the heart. This is much of the same nature as the ingrowth or parablast theory of His (75), and is obviously defeated by the same array of facts which long ago relegated the parablast theory to a place of mere historic interest, in spite of the fact that it is so often revived for literary reasons.

Finally, we may consider the study of the developmental products of the early wandering mesenchymal cells on the yolksac of the Fundulus embryo as a problem of cell lineage carried to its ultimate end. The primordial mesoderm cell or cells carry within their bodies all the potentialities of the mesoderm and may give rise to a series of cells which are capable of developing muscle, cartilage, bone, connective tissue proper, blood cells, vessels, etc. Yet after a few cell generations the individuals in the series derived from these early cells containing all the mesodermal potentialities no doubt become somewhat limited as to their potentialities. In a certain generation there may be definite cells more or less generally distributed which possess the capacity to give rise to muscle cells, but to no other type of mesodermal tissues. Still later in development these cells may come to be even more limited in their developmental capacities and thus have the power to produce only a certain type of muscle cell and no other type.

Collections of such cells would then be designated embryologically as the anlage of striated muscle, smooth muscle, or heart muscle as the case might be Yet it is not to be forgotten that at this stage there might be really no means of distinguishing between the several different types of mesodermal cells.

Limitization of potentialities in the individual cells has apparently reached a comparable stage just about the time when the mesenchymal cells begin to wander upon the yolk-sac of Fundulus. We have seen these cells as they wander out and have noted how very soon they may be separated into four distinctly different types, and following the development and behavior of these types it has seemed evident that they are entirely separate and do not intergrade or transmutate. The black chromatophore does not change its nature or divide off other cells which become different in type from the parent cell. Neither do the endothelial cells lining the vessel walls change into chromatophores or into erythroblasts, or vice versa.

From all the observations on these yolk-sacs we must conclude that the four types of cells described above have developed from four different anlagen, although these anlagen were not necessarily localized groups of cells, but were diffusely scattered mesenchymal cells capable of developing into a definite product, either normal or abnormal, depending upon the nature of the developmental environment. Therefore, the four distinct mesenchymal anlagen each give rise to a perfectly typical and distinct cell type although all develop in, as far as one can judge, an identical environment, the cavity of the yolk-sac between the ectoderm and the periblastic syncytium. The differences among the four cell types produced are from the standpoint of our present knowledge in all probability due to the potential differences among the apparently similar mesenchymal cells from which they arose. The four types including endothelial cells and erythrocytes we must consider from an embryological standpoint as arising from different mesenchymal anlagen.


The yolk-sac of the teleost egg is a most beautiful object for observing the movements and migrations of cells in the developing embryo. Such a yolk-sac has only one really definite continuous membranous cell layer, the ectoderm ; a true endodermal layer is absent, though a superficial syncytium, the periblast, fuses with the actual yolk surface. The mesodermal layer is represented by numerous separate wandering mesenchymal cells. These freely wandering mesenchymal cells may be clearly observed through the perfectly transparent ectoderm as they move over the surface of the periblast.

The present contribution attempts to give a full account of the movements of the mesenchyme cells and their manner of development and differentiation in the j^olk-sac. Observations have been made on the normal embryos from the earliest stages at which the mesenchyme wanders out upon the yolk up to the late embryo in which a complex vitelline circulation is fully established, and all of the products of the yolk mesenchjane completely differentiated. The study has been greatly facilitated by a comparison of the normal embryos with specimens in which the circulation of the blood was experimentally prevented from taking place. In such specimens the cells on the yolk-sac never became confused or contaminated with other cellular elements introduced by the circulating blood. The wandering cells may thus be completely followed through all stages in their isolated position.

The behavior of the migrating cells impresses one with the very important role of such elements in the formation of tissues and organs, particularly the blood vessels and certain blood cells. The observer is also struck by the fact that such phenomena are extremely difficult if not actually impossible to interpret from a mere study of dead specimens cut in serial sections. Of course the study of sections greatly aids the observations on the living, and but for the fact of a long acquaintance with the Fundulus yolk-sac in sections, the wiiter would have found it much more difficult to identify many of the cells in life.

The results of this investigation of wandering mesenchymal cells may be summarized as follows :

1. The wandering cells begin to migrate away from the embryonic shield or line of the embryonic body at an early period, when the embryo is about 40 hours old, the germ ring having almost completely passed over the yolk sphere to enclose its vegetal pole. The cells migrate away chiefly from the caudal end of the embryo, only a few wandering out from the head region. The regions of the yolk-sac thus suggest an area opaca about the tail end and an area pallucida around the neighborhood of the head.

All of the cells wander into the so-called subgerminal cavity, the space Wilson ('90) and others consider a late stage of the segmentation cavity, between the yolk-sac ectoderm and the periblast sjTicytimii.

AYhen the cells first appear they are all closely similar in shape and about the same size. Very soon, however, they begin to exhibit certain differences. Many become elongate spindle cells with delicate filamentous processes, sometimes producing a stellate appearance. Others are more amoeboid in shape with conical pseudopod-like processes which are constantly being thrown out at one place and withdrawn at another. Still a third class of cells appear somewhat later than the other two; these are more circular in outline with short thick pseudopods and are more slowly moving.

The movements of these extremely nmnerous cells and their changes of position may be readily followed \vith. a high magnification. In embryos of about 60 hours, still some time before the heart begins to beat or the blood to flow, fom* clearly distinct types of cells may be recognized among these originally similar mesenchymal cells, and the further history of the four types has been completely traced.

2. The amoeboid cells with conical pseudopod-like processes shortly after 60 hours begin to show an accumulation of pigment granules within their cytoplasm. Just at this time they are seen to be of two distinct varieties, one depositing a black and the other a brownish red pigment.

The black chromatophore increases rapidly in size and by the end of the third day becomes an enormous amoeboid body wandering over the yolk. These cells are attracted to the walls of blood vessels and plasma filled spaces, such as the pericardial cavity becomes in specimens without a blood circulation. When the embryo is five days old the chromatophores are abundantly arranged along the walls of the vitelUne vessels, but the pigmented cells are distinctly separate. After this time neighboring cells begin to fuse along their adjacent borders and large pigment syncytia are formed which completely surround and ensheath the vessels. A single syncytium is often of considerable extent, as shown in figure 15.

The brown chromatophores have a somewhat different history-. They never become so massive as the black, and their processes are more delicate and graceful in appearance. Yet these cells also attain a large size and in embryos of 72 hours are scattered over the entire yolk-surface. After the third day when the blood begins to flow in the yolk vessels, the brown chromatophores likewise become attracted to the vessel wall. These exquisitely branched cells apply themselves to the wall of the vessel and may often completely surround it, as shown in figure 17. This t;^T3e of chromatophore, however, always maintains its cellular individuality and never fuses with other cells to form a sjnicytium as is the case with the black type.

The function of the chromatophores on the yolk-sac is most difficult to decide, but one thing is certain, they never become changed into any type of blood cell. The brown chromatophore in early stages may accidentally reach the blood current; it then becomes spherical and may be readily observed for a long time on account of its huge size as compared with the blood cells. It never, however, changes in type.

In specimens without a circulation of the blood both types of chromatophores arise in a normal manner and differentiate normally. Their arrangement along the vessel walls fails to occur and they remain scattered over the yolk or collected about the plasma filled spaces. The heart in such embryos is sheathed with pigment, while the normal heart never has a chromatophore on it.

3. The elongate spindle cells with their delicate filamentous processes are small in comparison with the two chromatophore types. These spindle cells retain in general their original appearance, but their behavior is most important. In embryos of about 48 hours such cells aggregate into certain rather definite groups; later, these groups become more linear in shape and finally these lines of cells arrange themselves so as to form tubular vessels. Several of the larger vessels arise independently upon the yolk, and certain ones of them later become connected with the venous end of the heart, while in all cases capillary nets which also arise independently become connected with the larger vessels. These processes may actually be followed through every step in the living yolk-sac.

The wall of the early vessels is very irregular with spaces existing between the component cells. Corpuscles are often caught in these spaces or entangled in the filamentous processes of the endothelial cells. Such conditions in sections would appear as though the corpuscles actually formed a part of the endothelial wall and might incorrectly be interpreted as endothelial cells changing into blood cells. Nothing has been seen in the living embryos to indicate that an endothelial cell has the power to produce a blood cell or to change into a blood cell of any type, but much has been seen to the contrary.

The generalization particularly made by Thoma ('93) that larger vessels arise from a net-work of capillaries is not true for the large vitelhne vessels on the fish yolk-sac. It is also found in the specimens without a circulation of the blood that the vessels arise and increase in size and persist for a long time without ever experiencing any effect of the blood current upon their walls. In many embryos the circulation after having begun may stop for a time and then later be reestablished, the vessels having persisted in a normal condition. Thoma's so-called laws of vessel formation are, therefore, rudely violated by the development of the vascular system in these embryos.

The vessels arising from independent mesenchymal cells in the space of the blastocoel in the teleost yolk-sac entirely overthrow any notion that vessels arise ontogenetically as portions of the coelomic epithelium. The vascular lumen is originally continuous with the primary body cavity, the segmentation cavity, and never with the secondary body cavity or coelomic cavity.

4. The fourth class of cells wander out from the embryonic body somewhat later than the three former types. These are small globular cells with short pseudopod-like processes. They move very slowly, but finally collect into groups on the posterior and ventral regions of the yolk-sphere.

The round cells wander away only from the caudal region of the embryo and probably are derived from the so-called intermediate cell mass which is the anlage of the red blood corpuscles in the fish embryo.

The groups of round cells are slow in their differentiation but just before the circulation of the blood begins, they are seen to be circular erythroblasts. The observer may follow the disappearance of the islands of cells one by one as they are enclosed by the vessels and swept into the circulating stream. About the fifth day these circular erythroblasts become flattened ellipsoidal erythrocytes filled with haemoglobin, the typical red blood corpuscle. The complete change from wandering more or less spherical mesenchymal cells into typical haemoglobin bearing corpuscles may be followed in the living yolk-sac.

In several instances the entire body proper of the embryo failed to develop or else degenerated very early, yet the yolk-sac formed or persisted with numerous blood islands fully differentiated.

The embryos in which there has been no circulation of the blood form the blood islands from the wandering cells on the yolk-sac, and the constituent elements of these islands differentiate perfectly and may maintain their red color for many days. Yet they never leave the locality in which they have differentiated. The fully formed red blood corpuscles have Uttle if any power of migrating. When the observer can be positive that the blood has never circulated, and this requires very consistent watching, the islands of the yolk are always limited to certain regions and never occur so far anteriorly on the ventral surface of the yolk as to reach the venous end of the heart.

5. On the yolk-sac of Fundulus embryos one thus finds four distinctly different products differentiating from the apparently sunilar wandering mesenchjinal cells. The environment in which the four types differentiate is identical as far as is possible to detemime, and the only explanation of their various modes of differentiation is that the original mesenchymal cells that wandered out were already of four potentially different classes. These differences in potentiality within the early cells gave rise to the four different directions of cytomorphosis in one and the same environment. The four resulting types of cells are then in an embryological sense derived from different mesenchymal anlagen.

Literature Cited

(Only those titles not given in the previous paper) Agassiz, L. and Whitman, C. O. 1884 On the development of some pelagic fish eggs. Proc. Amer. Acad. Arts and Sci., xx. Bremer, J. L. 1914 The earliest blood-vessels in man. Amer. Journ. Anat.,

16. Evans, H. jNI. 1909 On the development of the aortae, cardinal and umbilical veins and the other blood vessels of vertebrate embryos from capillaries. Anat. Record, 3. Ryder, J. A. 1884 A contribution to the embryology of osseous fishes. Report U. S. Fish Comm. for 1882, Washington.

1887 On the development of osseous fishes. Report U. S. Fish Comm.

for 1885, Washington. Thoma, R. 1893 Untersuchungen ueber die Histogenese und Histomechanik des Gefasssystems. Enke, Stuttgart.

1896 Text-book of general pathology and pathological anatomy. Translated by Bruce, London. ,