Book - An experimental analysis of the origin of blood and vascular endothelium (1915) 1

From Embryology
Embryology - 17 May 2021    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)

I. The Origin of Blood and Vascular Endothelium in Embryos Without a Circulation of the Blood and in the Normal Embryo

Charles Rupert Stockard (1879-1939)


The origin of blood presents an almost unique problem in embryology. First, on account of the fact that the initial blood anlage in many animals is contributed to by wandering cells. Second, owing to the establishment of an early flow or circulation of embryonic fluids before the blood corpuscles have arisen.

Soon after the cells and corpuscles are formed they are swept into this circulating current and carried to all 'parts of the body. In this way the blood cells become associated and mixed with numerous other types of cells, and it is difficult, if not impossible, to estabhsh their true relationship mth their surroundings. For the above reasons one is often ready to believe that many of the even careful and long thought out contributions to the development of blood are, after all, largely a matter of the author's own interpretation rather than a record of the actual processes.

The general current of opinion at the present time would seem to indicate that all blood cells arise from a mesenchymal tj'pe of cell. A number of very competent workers have described the change of tliis mesenchymal cell into a stem cell or mother cell. On one side from this mother cell are developed various leucocytes, which it is important to note always occur in an interstitial position, while on the other hand, this same general type of mother cell gives rise to other cells which later differentiate into tjT^ical erythroblasts, and finally eiythrocytes which are always found to be located within the vessels. These socalled indifferent mesenchymal cells probably, from the evidence contained in the Uterature, do form blood cells, but to the discriminating reader the evidence is not at all convincing that both white blood cells and red blood cells really arise from one common mother cell or common embryonic anlage.

The possibihty, and even probability, is certainly present that these so-called stem mother cells may in reahty not all belong to one type, but are different and may already be destined to form either red cells or white cells. Yet on account of their wandering capacities as well as on account of the fact that the earliest blood cells are swept around in the circulating current they have become so mixed and confused that it is almost impossible to separate the cell groups or differentiate between them. One must in this connection, remember the fact that almost all authors have concluded that only red blood cells are formed in the blood islands of the yolk-sac in most vertebrates. All authors who have studied the development of the blood in Teleosts have invariably described only red blood cells as arising in the intermediate cell mass. No one has ever mentioned the presence of white blood cells in the true blood forming anlage of the Teleosts.

The monophyletic school really goes still further and not only claims that all tjrpes of blood cells arise from a common mother mesenchymal cell, but also that the vascular endothehal cell is hkewise capable of giving rise to the various types of blood cells and is originally a cell of the same type as the stem mother cell. There are numerous descriptions and illustrations of the origin of blood cells from the vessel hnings in the literature of the past twenty-five years, since Schmidt in 1892 described the transformation of individual endothelial cells into white and red blood corpuscles. Yet again, I beheve that the really skeptical reader will not be at all convinced that such a thing ever takes place from the evidence presented in the literature, certainly not from any of the illustrations that have been made of this process.

No real vascular endothehal cell has been actually observed to metamorphose into a blood cell or to divide off another cell which forms a blood cell, and until such a dhect observation is forthcoming one can only question the accm'acj^ of the interpretation of the various observations up to now recorded.

The mesenchyme is a very generalized embryonic tissue and from it arise the various kinds of blood cells, endothehal cells, connective tissue cells, etc. There can be no doubt of the great genetic difference between blood cells and connective tissue cells, yet their parent cells are with our present methods indistinguishable. We may wdth equal justification go further and hold Hkewise that the cells from which the vascular endothelium, red blood cells and white blood cells arise are mesenchymal cells really differing in nature according to whether they will give rise to one or the other of the three cell types. Yet they may not differ from one another in smy way by which we can at present distinguish them. If this proposition be true, or even if the weight of evidence lean in this direction, it is scarcely more justifiable to derive these completely different cells from a common mother cell than it would be to derive connective tissue and blood cells from a common mother cell.

Of course, we are only considering the mesenchymal cell just before its differentiation is to begin. Carried back further, no doubt, all the cells become more and more ahke and possess more and more complex potentiahties as is so thoroughly demonstrated by the numerous studies of cell lineage. In the beginning, of coiuse, all cells arose from one single egg cell capable of giving rise to every tissue of the body, but after tendencies in differentiation have proceeded sufficiently far in the various cells some then form real mesenchymal cells. Later individual mesenchymal cells incline in certain directions and finally become incapable of giving rise to any other than the definite type of tissue or cells towards which their particular tendencies have directed just as certain endodermal cells become specialized to form the liver while others near by and at first indistinguishable from these give rise to the ducts and acini of the pancreas.

All of the vertebrate classes present these many questions of blood origin, etc., but the forms upon which this investigation has been conducted, the Teleosts, possess in addition many extremely interesting special problems. In all other meroblastic embryos the majority of the earliest blood cells arise in yolk-sac blood islands. Yet in many of the Teleosts there are apparently no early blood islands on the yolk, but all of the blood forming cells are contained within the embryonic body.

This intra-embryonal blood anlage has been frequently described by many authors as the "intermediate cell mass." The intermediate cell mass as has been suggested by Marcus ('05), Mollier ('06), and others, is really the homologue of the blood forming yolk-sac mesoderm in the other meroblastic types.

The bony fish is important as an object of study on account of the fact that so many of its organs and tissues arise in a way pecuhar to the group and differing from the other vertebrate classes. The sohd gastrular invagination described by Sumner ('00), the original solid condition of the central nervous system, the solid optic knob which changes into the optic vesicle, and in the present connection, the very particularly interesting solid cord of cells, the intermediate cell mass, which is to give rise to the red blood corpuscles of the individual make the Teleosts a group of great embryological interest.

The complexity of the problem concerning the origin of the various types of blood cells is then largely due to the migration and mixture of the cells involved. It is strange that up to now no investigator has attempted in an expermiental way to analyze the situation. It would seem to be one of the most favorable problems for an experimental analysis, and in the end it is certainly an analytical problem.

If it were possible by any means to separate the anlage of the red blood cells from that of the white blood cells and prevent the flow of fluid in the embryonic body so that these cells would not frequently become intermixed, then it would seem possible to determine clearly the entire genesis of the various type ceUs. If all the types of blood corpuscles did arise from a common mother mesenchymal cell they should then be found in intimate association throughout all blood forming regions. Further, if the vascular endothelium really has blood forming power, it should be found that blood cells arise in any region of the embryo which possesses vessels lined by such endothehmn.

There have been various experiments performed which have interfered more or less with the circulation of the body fluids of the embryo, but none of these experiments where aimed at a solution of the genesis of blood cells or have been used for such a purpose. Knower ('07) removed the heart anlage from early frog embryos and the}^ continued to develop in some cases with almost no circulation. In other specimens there was a very feeble sluggish circulation due to the pulsation of the lymph hearts or of remnants of the heart which remained after the operation. The embryos were not particularly adapted for the study of the blood questions since some circulation always took place, and this no doubt was sufficient to contaminate the original sources of blood cells and so confuse the situation. Loeb ('12) has reported experiments on bony fish hybrids and embr^^os treated with certain chemicals in which there was a heart beat but no circulation. These embryos were, however, not studied for either blood or vascular genesis.

The first demonstration of the fact that the embryo could develop without the circulation of the blood was given by Loeb in 1893. He showed that Fundulus eggs developing in solutions of KCl had no heart beat and no circulation of the blood, yet some vessels formed. In 1906 the writer repeated this experiment and confirmed Loeb's results entirely, but found that the vascular system and general development of the embryo was extremely abnormal and was hardly reliable for conclusive studies on the origin of special tissues.

With these experiments in mind, and appreciating the problems indicated above regarding the origin of blood as weU as vascular endothelium, I have undertaken an extensive experimental analysis of this subject in conjunction wdth a careful systematic study of the histogenesis of the blood and vessels in normal embryos. The results of the experimental study which has been carefully followed during the past three years are presented in the following pages of this paper.

Methods of Experiment and Material

Six years ago, while studying the influence of alcohol and various anaesthetics on the development of Fundalus embryos, I noticed that many of these embryos had a feeble heart beat, but no circulation of the blood. At that time, particular attention was given to a study of the defects of the central nervous system and of the organs of special sense and no attempt was made to investigate thoroughly the conditions present in other tissues and organs of the body.

During the last few years, special attention has been devoted to the study of these embryos without a circulation of their body fluids with the main object of analyzing as completely as possible, the origin and subsequent development of the heart, vessels and blood. All of the experiments have been repeated through three siunmers in the Marine Biological Laboratory at Woods Hole. It has been possible to produce embryos that were almost normal in all particulars yet in which the blood failed to circulate on account of the fact that the heart was blind at one or both ends or disconnected at the venous end or finally was a completely soUd cord of tissue.

The embryos were studied in life and the particular individuals which had been so observed were fixed and finally studied in microscopic sections. The observations made on the experimental embryos have in all instances been completely checked and controlled by careful detailed study of the blood and vessels in normal embryos.

The experiments have been performed on two species of Fundulus, heterocHtus and majalis, and the results are practically identical for both. The eggs were stripped from the female into small dishes containing no water and fertilized by milt pressed from the male. About fifteen minutes after the application of the sperm water was added to the eggs. In this way one gets a very high percentage of fertihzed eggs, while fertilizing the eggs under water or adding spemi to a dish of water gives a much poorer result. The fertilized eggs are then di\Tided into groups so that the experiment and control are all from similar sources. The eggs just before dividing into the two-cell stage were introduced into solutions of alcohol.

The solutions which gave the most favorable results were prepared in the following way: 50 cc. of sea-water was placed in each dish and to this was added respectively, 1.5 cc, 2 cc. 2.3 cc, 2.6 cc, 2.8 cc, 3 cc, 3.5 cc, of 95 per cent commercial alcohol. The eggs remained in these solutions for twenty-four hours, after which time the solution was renewed. After forty-eight hours the eggs were removed from the alcohol solutions and placed in pure sea-water.

To give a general idea of the way in which the embryos developed in these different grade solutions of alcohol, I may cite some of the details of one experiment. After forty-eight hours many eggs are dead in all the solutions. The dead eggs are thrown out. WTien seventy-two hours old many others are dead in the solutions with 1.5 cc, 2 cc, 2.3 cc, and 2.6 cc, while all but one individual had died in 2.8 cc. . It should be added here that about seventj-five eggs are placed in each dish. From the 1.5 cc. solution sixteen are alive at seventy- two hours, several with various eye defects, the hearts are beating but contain colorless fluid and the only trace of red blood color is in the intermediate cell mass and caudal vein. Several others of this lot have a full circulation with corpuscles in the current. From the 2 cc. solution one has a feeble heart beat but no blood is visible and there is no circulation; some embryos have no circulation but blood is present in the posterior end of the body, while many have blood circulating with a strong heart beat.

This brief reference to the notes of the experiment show that such doses are just on the border Une of effectiveness since some indi\dduals in the solutions are not able to develop a circulation, while others with a higher degree of resistence do develop a more or less normal circulation. It is very important in such experiments to use these threshold doses since they are the least injurious possible to give the desired result. In this way one gets individuals which have no circulation of the blood, and, therefore, in which the blood anlage develops and remains in its permanent position, without having any serious defects or abnormalities in the general body tissues of the embryo.

From a study of such specimens carefully controlled by a study of normal individuals, one is fully justified, I believe, in drawing final conclusions as to the significance of the developmental processes taking place. It cannot be argued so far as the blood anlage is concerned that the conditions recorded are pathological or other than those which would occur in a normal genesis of the blood except that it never circulates.

Embryos which are intended for microscopic study have been prepared in the following manner: The eggs are placed in picro-acetic (saturated aqueous solution of picric acid and 5 per cent glacial acetic) from thirty to forty minutes, then put into 70 per cent alcohol. This is frequently changed in order to wash out the picric acid. After they have been about one-half hour in the 70 per cent alcohol, the egg membrane is removed with fine dissecting needles. This is the most favorable time for removing the membrane. If the eggs have been left for a long time in the alcohol, the membrane is more difficult to remove and the embryo is brittle and more liable to injury.

After removing the entire membrane, the yolk-sac is then punctured at its ventral pole and the yolk mass very slowly and cautiously removed from the sac. To remove the yolk mass, requires a great deal of practice and extreme care in every case. It should be done with the use of a binocular microscope so that the operator can be certain not to tear or destroy the yolk-sac or injure the delicate heart lying close above the yolk. After a great deal of practice it is possible to remove the yolk from a number of embryos and leave the yolk-sac in perfect condition with the heart and pericardium practically undisturbed. In the great majority of cases, however, it is generally impossible to completely remove the yolk. It is usually necessarj^ to remove the yolk on account of the fact that when the eggs are imbedded in either paraffin or celloidin, the yolk becomes so hard that it often breaks the sections and makes it very difficult to get a complete or perfect series. "WTien the yolk-sac is punctured onehalf hour after having been in the 70 per cent alcohol, the yolk material is in a giumny or viscid condition and is more easily removed than at any other period tried.

After having removed the egg membrane and the yolk, the embryos are then allowed to stand twenty-four hours in 70 per cent alcohol when they are changed to 80 per cent to be kept until the time of imbedding. The embryos are imbedded in paraffin and cut in serial sections from five to ten micra thick. They are stained in hematoxylin and eosin and extracted or carefully differentiated so as to bring out a clear stain of the blood cells and tissues. A complete series of these embryos have been made from a time before the appearance of blood up to sixteen days old, the normal embryo hatches and becomes free swimming at about twelve days.

As mentioned above, a similar series of normal embryos have been prepared and used for comparison with these non-circulating individuals.

In order to be certain of the final developmental product of the blood cells of the fish, mmierous smears have been made from various tissues and from heart blood taken from the adult Fundulus. In these smears one finds the various types of white blood cells and the ordinary- red blood corpuscles of Teleosts.

III. The Study of Living Embryos with and without the Circulation of the Blood

1. Normal development up to the establishment of a circulation

The rate of development of Fundulus embryos is very variable, differing at different periods of the breeding season, and also differing in groups of eggs from different individuals.

When twenty-four hours old, the germ ring has descended almost to the equator in the most rapidh' developing indi\dduals. In others the ring is only one-third way over the yolk sphere. The embrj'onic shield and the first line indicating the position of the embryo's body is now to be made out. At forty-eight hours, the yolk sphere is completely covered by the ectoderm, the embryonic body is well shown with the optic knobs projecting prominently and several somites easily distinguishable. The heart has not begun to pulsate and no blood cells or blood anlage are distinguishable in the living specimen. Very soon after this time, or at least by sixty-eight hours, there are about ten somites present and collections of cells on the yolk-sac are the first indication of blood islands. No pigment cells have formed up to sixty hours.

At seventy-one hours pigment cells are recognizable but the blood islands are not yet colored and are sparsely arranged over all the yolk region except the anterior half. Near the lateral borders of the embryo and on the posterior yolk surface the islands are most abundant. At this tune there is still no visible heart-beat or circulation to be seen with the high power microscope. At seventy-five houi's the pigment particles are just beginning to show in the chromatophores. A well formed vesicle is clearly seen at the posterior end of the embryo from forty-eight to seventy-two hours and older. This is the so-called Kupffer's vesicle, and it, like the pericardiimi, becomes greatly distended by an accumulation of plasma in those individuals which have no circulation.

Embrj^os with fourteen somites may still have no heart beat and no circulation of the blood. Any later than this, however, all normal embryos estabhsh a heart-beat and a circulation of colorless fluid, there being no blood cells present in the initial circulating medium. Very soon after the circulation is established at first a few but quickly many blood cells are added to the stream. This is merely an abbreviated sunmiary of the development of the Fundulus embrj^o up to the time of the establishment of the heart-beat and circulation, but as stated above, the rate is variable and it often happens that an embryo of seventy-two hours has already established a vigorous circulation and the plasma is loaded with well formed blood cells.

2. History of experimental embryos to the time when a circulation should begin

In the experimental embryos development proceeds more slowly than in the normal. The plasma which should circulate in the vessels accumulates in the sinuses over the yolk and finally seems to collect in great amount in the pericardium, the lateral coelomic cavities and in Kupffer's vesicle, so that these spaces become hugely distended and appear as great sacs or vesicles of colorless watery fluid. The excessive presence of this fluid in the pericardium seems to exert a mechanical effect which tends to separate the head of the embryo an unusually great distance from the yolk mass and thus stretches the heart out into a long stringhke cord, passing from the embryo to the surface of the 3'olk.

This pushing away of the head from the yolk is very well indicated in figures 15 to 20, which show various types of hearts found in these embryos during later stages. The stretching or pulling out of the heart may possibly be the cause of the failure to develop its proper connections with the veins, or in some cases to establish and maintain its lumen. On account of these mechanical deficiencies in the heart, we find that it is incapable of propelling the body fluids and establishing the circulation of the blood. The fluids thus accumulate in the large sinuses or spaces, in most cases the coelomic spaces and in the case of Kupffer's vesicle in an endodermic or endo-mesodermic cavity.

The red blood cells become evident after the fluid accumulation has partially taken place. They are always seen to originate in definite localities and are never found in any place very distantly removed from these localized regions unless a partial circulation or accident of some kind has occurred. Although red blood cells in many places arise from wandering cells the blood cells themselves have little capcity to wander.

3. Early formation of blood cells in living embryos

a. Intra-embryonic blood cells: The chief place of blood cell formation is the intermediate cell mass which extends from about the level of the anterior portion of the kidney back posteriorly to behind the anus and well into the tail of the embryo. This is the principal blood mass, but in addition to this, there are present in all of these non-circulating individuals small blood islands over the posterior and ventral yolk regions. These blood islands are also present on the yolk of normal individuals but in these the islands are swept away when the circulation begins.

b. Yolk-sac blood islands. A number of observations were made on the embryos which failed to develop a circulation and also on normal embryos to determine the significance and relationships of the blood islands on the yolk as far as was possible in the living individuals.

In one experiment when the embryos were seventy-two hours old, or just about the time that the heart-beat was beginning in many, it was found that although the plasma was circulating blood islands were present on the posterior yolk region. Ten embryos were selected which showed these posterior yolk blood islands and isolated to determine whether the blood islands would subsequently enter the circulation or what their fate would be. Ten other embryos mth a feeble heart-beat but with no circulation yet also showing small posterior yolk-sac blood islands were isolated for comparison.

On the following day, nine out of ten individuals which had established a vigorous circulation had no blood islands remaining on the yolk. All of the islands had been taken into the circulation or vascularized, so that instead of blood islands there, was now a network of vessels over that portion of the yolk and the blood cells had entered the current. One of the ten embryos exhibited an abnormal arrangement of the blood vessels that was particularly instructive. On one side there was a large vein running from the embryo out onto the yolk and on this side all of the blood islands had disappeared forming a network of vessels which conducted the blood to the venous end of the heart. On the other side of the specimen there seemed to be a suppression in the development of vessels near the embryo. The islands were still in the same condition they had been on the previous day except that the cells composing them had become much redder so that there was now no doubt whatever that they contained erythroblasts and early erythrocytes.

The ten embryos which had no circulation on the previous day were now found to present the following conditions: Two had established a perfectly free circulation and all of the blood islands had been swept away. In two other indi\dduals circulations were established in an abnormal manner so that many blood islands still remained and presented a bright red appearance. Six of the ten specimens had no circuTation whatever and the entire arrangement of blood islands over the yolk was in exactly the same condition as on the day previous, except that the blood cells were now much redder in color.

During the course of the experiments similar tests and observations have been repeated four times. In each instance isolated groups of embryos showing yolk-sac blood island were selected and examined in order to ascertain on the following day the fate of these islands. In every case my experience was identically similar to that just described. The islands on the yolksac are often very far distant from the embryonic body as is readily seen by reference to figures 21 to 24. There is no doubt that wandering cells migrate out on to the yolk surface at a very early period and here give rise to blood islands comparable to those formed in other meroblastic embryos. It must be recognized, as I shall bring out in a consideration of the microscopic study of these embryos, that the yolk-sac of the fish is not entirely comparable to that of all other vertebrate types, yet there are many observations on the early living embryos which have convinced me that mesenchymal cells do wander from the embryo to various parts of the yolk-sac. These cells occupy a position between the ectoderm and the periblast (periblastic endoderm ?) just as the peripheral mesoderm would in other yolk-sacs. Some of the wandering cells are future pigment cells of either the red or black variety to be mentioned later, others future endothelial cells, but many at least are to give rise to future blood cells.

Therefore, in embryos at about the beginning of the circulation one finds two distinct blood regions: The major region and most evident is the intermediate cell mass of former investigators, and the second position in which the blood cells are seen is the yolk-sac blood islands. The earliest yolk-sac blood islands are very easily overlooked. The writer had examined these embrj^os in great numbers and studied them for sometimes before finally discovering the existence of the early islands. With a high power single objective binocular, however, after their location is known the observer is readily able to see the nuclei of these cells in the posterior region of the yolk-sac, and they may then be followed from time to time. After these observations there is finally no doubt that blood islands do form on the yolk-sac of the Teleost embryo but these islands are probably to be regarded as disconnected portions of the intermediate cell mass or blood anlage.

It is freely admitted that this yolk-sac island blood formation may not take place in all Teleosts. The early wandering cells that here give rise to blood islands may in reahty be compared to the growing out onto the yolk of masses of cells from the intermediate cell mass as described by Swaen and Brachet ('99, '01). These cells in the species here studied may wander earlier and more freely than in the trout.

At any rate, as shall be brought out later, the ventral mesoderm of the yolk-sac in other vertebrates and the intermediate cell mass of the Teleosts are very closely related homologous portions of the mesoderm, if not one and the same thing.

4. The five-day embryos

The conditions which the embryos have attained five days after fertilization are illustrated in figures 1 to 4. In figure 1 a normal individual of this age is shown. The heart is slightly twisted but still more or less tubular. The vascular network on the yolk-sac is well established. The pigment cells are numerous but not yet fully developed and have not assumed an alignment along the blood vessels or taken on the usual embryonic pattern. The heart, of course, is pulsating vigorously and the blood current is easily seen both within the embryo and on the yolk-sac.

The three other figures in this group show different conditions of arrested development in individuals without a blood circulation. In figure 2 the pericardium is seen to be hugely distended so that the head is pushed or raised away from the 3'olk surface and the heart is stretched out into a long narrow tube extending from the ventral surface of the head to the sheer anterior surface of the yolk. This heart pulsates feebly and can be seen to contain a small amount of fluid which is churned up and down by the pulsations. None of this fluid, however, is ever pumped away from the heart. The pigment is much less plentiful than in the normal embryo of the same age, and the individual chromatophores are smaller in size than those of the normal embryo and have not sent out processes of any great length. No blood vessels at all are seen within the yolk-sac but very small scarcely noticeable blood islands are present on the posterior yolk region though not indicated in the sketch.

Figure 3 illustrates a more or less similar condition seen from a dorsal view. A small portion of the heart slightly projects beyond the anterior end of the head. If this egg were placed in lateral view, as was the case in figure 2, then the heart would be seen to show a similar condition, since it also was stretched out into a long narrow tube.

Figure 4 represents a very defective embryo. Specimens similar to this occur in great munbers in the stronger alcohol solutions. The bodies are very short since the descent of the germ ring over the yolk sphere is slow and at times incomplete, and the tail end of the embryo is often bifid or split giving a condition of cauda-bifida. At the anterior end, the upper right side of the figure, is shown the distended pericardial vesicle, and at the posterior end another large distended vesicle is in most cases the Kupffer's vesicle, but in some instances this is possibly distended spaces in the yolk mass just below the Kupffer's vesicle. These two sacs or spaces at the opposite ends of the body seem to be the places in which the non-circulating plasma most often accumulates to a great degree, in fact the accumulation of plasma is the actual cause of the exagerated condition of the spaces. In the individual illustrated by figure 4 the chromatophores are extremely small, but have arranged themselves to some extend so that they are very abundantly accumulated around the periphery of the Kupffer's vesicle while others have collected in the region of the pericardium. In the lateral yolk regions there are scarcely any pigmented cells present. All of these chromatophores, however, are small and contracted with very few processes of any extent.

In Fundulus embryos there are readily seen two distinct types of chromatophores. The one is a large dense perfectly black body with short broad processes. While the second is of a reddish color at first small and without processes, but later sending out very long graceful radiations which grow at the expense of the central mass until finally the whole chromatophore assume a mossUke branched structure, figure 6 shows both types well expanded.

Loeb ('93) at one time pointed out that these chromatophores migrate to the blood vessel walls and thus map the circulation on the yolk-sac of a normal embryo. While in embryos treated with KCl in which there was no circulation, the pigment cells failed to asume any definite pattern. They remained more or less indefinitely scattered over the surface of the yolk and the body of the embryo in no way tending to align themselves along the vessel walls. From this fact, Loeb concluded that it was probably due to some chemotactic reaction that the pigment cells lined up along the blood vessels when the blood began to circulate and the attracting substance was possibly the oxygen contained within the blood corpuscles. Observing the various indi\aduals without a circulation which we shall here consider, it will be seen that the pigment cells have a strong tendency to migrate to any cavity filled with plasma or fluid, and it is not probable that this plasma or fluid contains any more oxygen than is present in the other portions of the yolk-sac or body. It would, therefore, seem more hkely that some constituent of the plasma itself and not the oxygen contained within the blood cells was the stimulating principle which caused the migration of the pigment cells to the vessel walls.

Fig. 1 The head end of a normal Fundulus heteroclitus embiyo from life, five days old; ht, the heart, showing its S-shape; e, eye; o, ear.

Fig. 2 From a living embryo of the same age, developed in a weak solution of alcohol; the blood does not circulate and the body is small; ht, the heart stretched into a straight tube surrounded by a much dilated pericardium; ch, small and unexpanded chromatophores.

Fig. 3 The head end of a five-day embrj'o without a circulation, from a weak alcohol solution; ht, the heart slightly projecting from beneath the head.

Fig. 4 A five-day embryo from a stronger alcohol solution; the eye is Cyclopean, the posterior end of the body split, cauda-bifida, and no blood circulation. Pigment cells are collecting about the sinuses; Pr, distended pericardium; Kv, Kupffer's vesicle, also hugely distended.

5. The eight- and ten-day embryos

The next group of figures illustrates the advanced condition which the embryo has reached by the eighth day. Figure 7 represents a normal Fundulus embryo of this age. The body is seen to be well developed, the fins are already capable of mo\'ement, and the brain and spinal cord are well shown and co\'ered with the black type of chromatophores. The heart is seen to be more twisted than in the younger embryos and now occupies a position further under the head of the specimen. The network of vessels on the yolk-sac is beautifully mapped out by the arrangement of the pigment cells, largely the red type of chromatophore. It is to be especially noticed that pigment cells are never present on the heart of the nonnal embryo.

Fig. 5 An embryo eight days old, without a circulation; the heart, Jd, poorly developed, beats feebly twenty-eight times per minute, about one-quarter the usual rate; Pc, pericardium greath' distended with fluid; from a "1.5 cc. alcohol solution."

Fig. 6 Eight-day embrj'o without a circulation, ht; the heart dilated with plasma pulsates ninety-five times per minute; RCh, the red chromatophores beautifully expanded, but no vessels present on the yolk. Coe, the lateral coelomic cavity dilated with fluid; icm, the intermediate cell mass now a great string of red blood corpuscles.

Fig. 7 A normal eight-day embryo, the heart, ht; pulsating rapidly and the network of yolk vessels mapped out by the chromatophores.

Figs. 8 and 9 Two eight-day embryos without blood circulation; chromatophores unexpanded but collected on the heart, ht; the normal heart has no pigment cells on it; Pc, the dilated pericardium; icm, median mass of erythroblasts: CV, cardinal vein containing erythroblasts.

The other four figures of this group show individuals in which there was no circulation of the blood, although the hearts pulsated in a more or less feeble manner. In figure 5 the greatly distended pericardium is again shown, the heart is stretched from the head to the anterior surface of the yolk, and the lower part of the heart is completely sheathed with pigment. All of the pigment cells, however, are small and unexpanded.

In figure 6, the heart is very greatly distended and filled with plasma, yet it is apparently closed at one end since the plasma is churned up and down and never pumped out of the heart. In this case, there were several cells or particles suspended in the plasma contained within the heart, and these particles could be watched for long periods of time constantly moving up and down but never going out of their confined position. The pigment cells in this individual are greatly expanded, the red type chromatophores showing beautiful mossy processes. The lateral body ca\dties, Coe, or the coelomic spaces formed between the layers of the lateral plates of the mesoderm are greatly distended with plasma. A condition particularh^ noticeable in many such individuals. Red blood corpuscles are distinctly seen throughout the entire extent of the intermediate cell mass as indicated in the figure by the stippling in the posterior region of the body and the tail. The heart of this specimen is also richly covered with pigment and thus presents a striking contrast to the normal heart in figure 7.

In figure 8 much the same condition is presented except that here again the pigment cells are still contracted. The pericardium, however, is distended and the heart is covered with pigment. The anterior end of the intermediate cell mass showing erythroblasts is just seen where the body of the embryo turns over the yolk sphere.

Figure 9 illustrates a lateral view of a small embryo. In all of these embryos in which the blood fails to circulate the fins are much smaller and less well developed than in the control, the entire body of the embryo is smaller and the whole appearance is that of a general developmental arrest, the rate of development being behind the normal. Yet such individuals have rather perfectly formed bodies, are capable of movement and seem in general to be very well developed, their only defect, so far as can be determined in many cases, is the absence of the circulation of the blood.

In figure 9 the heart is again sheathed with pigment cells, the blood cells in the intermediate cell mass are very distinctly present in the posterior body region, and in this individual a lateral vein in the position of the posterior cardinal also contains blood corpuscles. This appearance is seen in a number of individuals and may merely result from the fact that in these the intermediate cell mass is bilateral or split rather than entirely median in position. Such an explanation will seem probable, I think, after a consideration of the embryos in section.

Figures 12, 13 and 14 show three individuals of ten days old. These happen to be more or less abnormal. Figure 12 has veiy small eyes but the general body structure and shape are fairly normal. The pericardium is dilated, and the heart is small and pulsating feebly with a little pigment towards its aortic end. There is a great accumulation of pigment cells around the posterior region of the yolk sphere and near the distended Kupffer's vesicle in this embryo. Here again the cardinal veins are seen to be loaded with blood and only in the posterior body region do the two lateral masses come to unite into a median cell mass. Figure 13 shows much the same conditions, the heart is a mere filament indicated b}-^ the chromatophore along it.

Figure 14 gives a dorsal view of an embryo of ten days. The pigment spots are very few in number and the embryo has a pale

Fig. 10 An eight-day embryo of Fundulus majalis from 2 cc. alcohol solution, showing a condition similar to the heteroclitus embryos; Ce, Cyclopean eye; Pc, distended pericardium; ht, straight heart and no circulation.

Fig. 11 A normal majalis embryo of eight days with S-shaped heart, ht, and yolk vessels forming a net. Same magnification as the smaller heteroclitus embryos.

Fig. 12 A ten-day heteroclitus embryo, no blood circulation, chromatoi)hores expanded and accumulated on posterior yolk region; ht, heart; Pc, distended pericardium; CV, cardinal vein filled with erythroblasts; SV, stem vein also full of red blood corpuscles.

Fig. 13 Embryo ten days old; hi, the heart a mere string covered with pia:ment; /, an independent crystalline lens; other lettering as in figure 12.

Fig. 14 View of head of ten-day embryo, no circulation, distended pericarddium, PC; I, free lens.

appearance when compared with a normal individual, such as the one of eight days shown in figure 7.

Figures 10 and 11 illustrates two specimens of Fundulus majalis drawn to the same scale as the previous figures. The egg of this species is considerably larger than that of heteroclitus, but its response to the experimental treatment is the same.

Figure 11 represents a normal embryo eight days old. It is seen not to be comparatively so far advanced at this period as is the heteroclitus, since its development is much slower and it requires from five to ten days longer to hatch. Very little pigment is present, yet the vessel net is well formed on the yolksac and the heart is distinctly seen to be more or less S-shaped.

Figure 10 is an embryo of the same age that had been subjected for forty-eight hours to a solution of 2 cc. of 95 per cent alcohol in 50 cc. of sea-water. Scarcely any pigment is present, the pericardium is typically distended and the heart is stretched into a long conical shape. No vessels are seen upon the yolk. The posterior end of the embryo is not shown in the figure, but in it could be seen early blood cells in the intermediate cell mass while a few small blood islands were present on the yolk-sac near the posterior end of the embryo.

After these eight- or ten-day stages very few changes of interest take place. The normal embryos hatch at from eleven to fourteen days as a rule, and become free swimming. The individuals without a circulation of the blood never succeed in breaking out of the egg membrane but may remain alive for twenty-five or thirty days in some cases, and almost all of them will live at least sixteen to twenty days. The red blood corpuscles are very distinctly noticeable in these older individuals and can be seen to remain permanently in their original places of origin. The intermediate cell mass may cease to be distinguishable, as was evident in the old specimens. This, however, is not due

Figs. 15 to 20 Living embryos sixteen days old, without blood circulation, showing variations in the pericardial distension, the position of the heads, and the perculiar heart conditions. These hearts, ht, all pulsate feebly and in several figures the slight lifting of the anterior yolk membrane is shown; this small membr:mous cone is rythmically raised with the pulsations.

to the blood cells having wandered away from the mass smce they have largely degenerated in situ probably on account of lack of aeration. The blood islands on the yolk-sac maintain their red color for a much longer period of time, and they continue to present a pattern closeh' identical with that seen in the same individual during its earlier stages.

6. Condition of the heart in old embryos without a circulation

The conditions of the heart in some of these old embryos is well shown in the series, figures 15 to 20. These figures are from embryos of sixteen days old. The control specimens at this time would as a rule have hatched. In the sketches the peculiarly distended pericardium is strikingly shown. This great distention of the pericardium seems to have exerted pressure in such a way as to have straightened the anterior end of the embryo and lifted it well away from the yolk surface. The mechanical pull caused by the separation of the head from the yolk would seem to be largely responsible for the fact that the heart becomes stretched into a very much attenuated tube or string.

In the upper left hand figure 15, the heart is not so greatly stretched and the pericardium in this case is not distended so much as in the others. The upper right hand figure 16, and the two central figiures, 17 and 18, show the pericardium distended to its utmost, and in these specimens the heart is pulled out into a mere string. Pigment cells seem invariably to wander along these string-like hearts, and they, therefore, stand out in the embryos as a black cord just as is indicated in the figures. The venous end of the heart which is connected with the yolk-sac is seen at each pulsation to lift slightly the yolk membrane in a cone-like projection from the surface of the yolk. As these hearts pulsate in their feeble fashion, one thus observes the j'olk membrane as it is puUed up and down.

The two lower figures, 19 and 20, show other somewhat different conditions of the heart. The hearts are small and do not reach so far towards the ventral portion of the yolk.

There is an ahnost limitless variety of peculiarl}^ abnormal hearts in these embrj^os and the six figures convey but, a slight idea of the many very strange conditions which are presented.

7. Development of the yolk-sac hlood islands in life

The blood islands in the living embryos, as was mentioned before, are quite difficult to see in the early stages. But a few hours before the heart begins to pulsate and the circulation bebecomes estabhshed they are very evident in the posterior ventral yolk regions. The arrangement of the blood islands display various patterns in different individuals in some being inconspicuous while in others an extensive network is present. These blood islands probably arise largely from wandering mesenchymal cells since the yolk-sac of the Fundulus embryo consists at first only of the yolk periblast with the ectoderm immediately above it. There is no true mesodermal layer to the yolk-sac and this mesenchj-mal blood formation on the 3'olk can, in all cases, be traced to the early wandering cells.

In figure 21 the posterior end of an embryo of ninety-six hours is shown. The Kupffer's vesicle, Kv, is dilated and pigment cells have accumulated around it. Immediately posterior to this are a number of blood islands indicated by stippling.

Figure 22 shows an embryo eight days old with the posterior ventral surface of the yolk well covered with blood cells. Erythrocytes are also seen in the intermediate cell mass. The blood in this embryo has never circulated and one can scarcely conceive that the blood islands on the extreme ventral surface of the yolk are due to the crowding out or pushing away of cells from the intermediate cell mass within the embryo. These cells are rather to be regarded as true yolk-sac blood islands which have arisen from early wandering mesenchymal cells probably in the beginning derived from same source as the intermediate cell mass.

Figures 23 and 24 show the caudal ends of two normal embryos of seventy-two hours. In these the heart has not begun to pulsate nor the blood to circulate, yet a distinct group of erythroblasts or early blood cells are seen already arranged on the yolk-sac in this posterior region.

In figure 24, it would look as though these cells had wandered out from and grouped themselves around the tail end of the embryo. At this period, seventy-two hours, the intermediate cell mass within the embrvo is not visible in life.

Fig. 21 Caudal end of an embryo of ninety-six hours, without a circulation; Kv, the distended KupfTer's vesicle with pigment cells collected around it. Bi, yolk-sac blood islands on the posterior yolk region.

Fig. 22 Lateral view of eight-day embryo, without a circulation, showing red blood corpuscles in the stem vein, SV, and also masses of blood islands, Bi, on the posterior ventral yolk regions.

Figs. 23 and 24 Posterior veiws of two seventy-two hour embr^-os, without blood circulation. Cells are seen wandering out from the tail, T, region into the position of the peripheral mesoderm in most meroblastic eggs; these cells collect into groups and form the blood islands, Bi.

It may then be concluded from a study of the hving embryos with a circulation and others without a circulation, that in the normal ordinary individuals as well as in those having their blood flow prevented, the origin and formation of blood in the bony fish occurs as follows: The chief source of origin of the erythroblasts is that so fully described by previous investigators as the intermediate cell mass, la masse intermediare. This mass, according to Felix ('97), Swaen and Brachet ('99, '01) and others, arises from the median portions of the two lateral mesodermal plates, primary seiten-platten. These bi-lateral masses migrate towards the middle line and there fuse to form the intermediate cell mass or blood string. In the living embryo this very mportant mass of blood cells is readil}^ demonstrated. It is usually median in position, but in many cases, as illustrated above, it may be double or bi-lateral, at least in its anterior portion. This bilateral arrangement may possibly be the result of a failure of the blood forming portions of the two lateral plates to move to the middle line and fuse to form a Stammvene, in other words, a type of arrest.

The second seat of differentiation of red blood cells which is distinctly shown in living embryos is to be found on the yolksac in the posterior and ventral region where numerous typical blood islands form and develop. All recent investigators of the development of the blood in Teleosts have denied the development of blood on the yolk-sac. Most of their investigations have been on the eggs of the trout, and it may be that in this group there are no blood islands. But in Fundulus we seem to have a transitional condition in which the yolk-sac islands have not been firmly incorporated within the intermediate cell mass but still remain out or wander out upon the yolk. At any rate, we must conclude that there is a secondary seat of red blood formation in Fundulus embrj^os, and that in life it presents the typical appearance of yolk-sac blood islands.

From a study of the living embryos, it is apparently impossible to determine whether all cells of these blood islands are only erythroblasts or of mixed types. This is, however, readily ascertained by a careful study of sections.

IV. The origin and histogenesis of vascular endothelium and blood corpuscles as determined by study of microscopic sections

1. The structure of the heart in ertibryos without a circulation

The hearts of the embryos in which there is no blood flow have been described in the h\dng in the preceding consideration, but when they are studied in section an additional number of very instructive points are brought out.

In the first place, the heart wall is usually very thin and not well developed. This is particularly true, in the long stringlike hearts that are present in those individuals in which the pericardium is so greatly distended. In the group of figures 25 to 28, one sees sections of these hearts taken through various regions.

Figures 27 and 28 show sections through a long narrow heart. In figure 28 the myocardium is seen to be practically one layer of cells, and within this the endothelial lining is distinctly formed. No noticeable structural difference beyond sUght variations in shape can be determined between the nuclei of the mycocardium and those in the endocardium. The myocardial layer is a thick more or less structureless cell mass while the endothelium is well differentiated into a thin single cell layer lining. This condition is found in a non-circulating embryo of four days old. Tracing the series towards the conus end of the heart, we find the arrangement indicated in figure 27. The myocardium is here also a thick layer of cells enclosing a distinct endothelial tube.

Fig. 25 Section through the heart of a four-day embryo without a circulation; Experiment 11, 1912, Embryo 6. Heart wall poorly formed; large chromatophore, Ch, in wall; 'ph., pharynx.

Fig. 26 Section of a similar heart; Experiment 11, 1912, Embryo 2. The guide outline gives the general relationships of the heart. MC, mj'ocardium; EC, endocardium; Br, brain; p6, periblast nuclei, and, p6.?, periblastic material filling the heart cavity, c; red staining cell.

Figs. 27 and 28 Through the aortic end and figure 28 through the tube-like body of a similar heart; Experiment 11, 1912, Embryo 7. Br, brain; p/?, pharynx; £JC,definitely formed endocardium, endothelium; MC, myocardium. The nuclei of the endocardium and myocardium are indistinguishable except for slight differences in shape.

Figure 25 is a section through another heart of the same age. In this a huge pigment cell is shown within the heart cavity. It is recalled that pigment cells were frequently seen to lie along these abnormal hearts while chromatophores were never present on the normal heart. The endothehum in figure 25 is more or less broken and the general condition of the heart is poorly developed.

In figure 26, a section is illustrated through a heart as it leads into the aortic arches. Here also large pigment cells are present. The endothelimn is indicated in several places and within the cavity of this heart is a mass of periblastic material. It would look as though the periblast had been sucked from the surface of the yolk into the heart cavity. Several large periblast nuclei, pb, are indicated and are easily recognized on account of their amorphous shape and huge size.

In figures 26 and 28 there are several cells, c, of a questionably degenerate type, the cytoplasm of which stains an extremely red color while the nucleus is small and pycnotic in appearance. These cells might in cases be looked upon as some type of wandering cell, l^ut in most instances they are very degenerate in appearance.

It must be distinctly noticed that in none of the figures are erythroblasts shown. Throughout these heart regions at all stages the observer is impressed by the entire absence of any form of red blood cells in embryos that have absolutely had no circulation. One must constantly guard against the possibility of a slight circulation having existed for a short time and then having ceased. Another reason for blood movement may be the twisting or twitching reactions of the embryonic body. Conclusions regarding the permanent position of blood must be based only on embiyos that have ))een carefully observed throughout their existence.

Figs. 29 and 30 Two sections through difTerent parts of the heart in an embryo sixteen days old, without a blood circulation; Embryo 314.

Fig. 29 The aortic end of the heart, an almost solid mass with endocardial, EC, cells near the center.

Fig. 30 Through the string-like portion of the same heart; jih, periblastic nuclei and material completely fill the heart cavity; EC, endocardial cells; Ch, chromatophores surrounding the heart wall. The upper part of the section is cut slightly oblique.

Figures 29 and 30 show sections through different parts of a solid heart string from an embryo of sixteen days old. In figure 29, the aortic end of the heart is shown to be almost a solid mass, and only near the center of the figure is a slight endothelial-like cavity or formation.

Figure 30 is a cross-section through the long string-like portion of this heart. It is seen to be completely solid, the central portion or core consisting of periblastic material containing large amorphous periblast nuclei, ph. Chromatophores have almost ensheathed the structure and present in the figure a dense black border. In one part of the section, however, a distinct endothelial-like formation is shown surrounding the periblastic core, and this heart again would seem to have sucked itseK full of periblastic material from the surface of the yolk. As stated, this heart was from an embryo sixteen days old in which the blood had never circulated, and it is quite evident that at the time the embryo was fixed, it would have been impossible to have had a circulation of blood through such a solid heart. In this specimen, however, numerous blood islands on the yolksac and well formed blood cells in the intermediate cell mass were to be seen.

The endothelial lining of these hearts has certainly arisen in loco, and has emphatically not grown into the heart from the yolksac vessels since the heart is not connected with such vessels, and further, no typical vessels are present on the anterior portion of the yolk-sac. In all cases, the intra-embryonal vessels are much better developed than the vessels of the yolk-sac. A general survey of these embryos would quickly convince one that the vessels within the embryo are in no case derived from ingrowths. This fact is peculiarly emphasized in a study of bony fish embryos, and is so convincing that it led Sobotta ('02) to develop a theory of vascular outgrowth from intra-embryonic vessels in contrast to the older parablast notion of His ('75), but I must agree with Mollier ('06) in his view that both theories are equally untenable.

The hearts in these experimental embryos as a rule lead directly into a more or less well formed aorta which, in all cases shows an endothehal Ihiing. The arches arising from this ventral aorta are very variable in the different embryos, yet in some cases are formed in an almost normal fashion. These arches also show a beautifully formed endothehal Uning, but here again one is impressed by the absolute absence of erythroblasts in any stage of development within the neighborhood of the heart or aortic endothelia.

2. The "intermediate cell mass" its origin, position and significance as an intra-embryonic blood anlage

On tracing the sections posteriorly one finds the intermediate cell mass to begin caudad of the pectoral fins and in the region of the anterior portion of the kidney duct. In studying a progressive series of very young stages forty-eight, sixty-six and seventy-two hours, the intermediate cell mass may readily be demonstrated to arise from the lateral mesodermic plates in the manner so clearly described by Swaen and Brachet ('01, '04). Felix ('97) previously pointed out chat the primary lateral mesodermic plates extend away from the somite and later become divided into the following three parts. The median cells lying close to the somites separate away to form a continuous longitudinal string, the string from each side forming one lateral half of the future intermediate cell mass. The intermediate cells of the primary lateral plates just lateral to the above median group give rise to a second cord of cells which later forms the primary nephric duct. The remaining lateral layer of cells now constitutes the secondary lateral plates which split to form two lamellae.

The primary lateral plate mesoderm thus gives rise to the intermediate cell mass, the primary nephric ducts and the somatic and splanchnic mesodermic layers of the lateral body wall. Between these two lateral mesodermic layers arises the portion of the coelomic cavity which we have seen in the living embryos without a circulation of the blood to be greatly distended with fluid.

The later development of the intermediate cell mass is found to proceed in almost exactly the manner described by Swaen and Brachet ('01, '04). This mesenchymal mass of cells is at first of an indefinite type lying between the notochord above and the intestine below and being flanked on either side by the primary nephric ducts. The first notable differentiation of the intermediate mass in the normal embryo begins shortly previous to the establishment of a heart beat. In an experimental embryo of seventy-two hours old, that is one in which the heart was just about ready to begin beating, figures 31 and 32 show the condition of the intermediate cell mass in cross section.

In figure 31, which is the extreme anterior end of the mass and, therefore, less well differentiated than the more posterior regions, the cells are seen to possess large round nuclei differing but slightly from the nuclei of the surrounding cells and those of the epithelium of the Wolffin ducts. The mass of cells is completely unsurrounded by endothelium, and I agree entirely with Swaen and Brachet that the central cells go to form the red blood corpuscles while the cells about the periphery of this mass form the vascular endothelium.

Figure 32 illustrates a section through a more posterior region of the same embryo, the intermediate mass is seen to be much further differentiated. The cells are here typical early erythroblasts and many are observed to be in active mitosis. The cells in the mass are becoming dissociated so that they are no longer so densely packed as in the section through the anterior region. This section is posterior to the ends of the Wolffian ducts, as well as the closed intestine and beneath the cell mass is shown the periblast over the yolk.

On tracing the series still further caudad, we find the indefinite cell mass end Knospe described by Marcus ('05), figure 33. This is a ventral cellular mass into which leads the notochord, intermediate cell mass and end of the endoderm. In other words, this mass would seem to represent the end bud at the dorsal blastopore lip, as if it were the point from which differentiation had taken place or from which the layers had grown forward.

Fig. 31 Section through the trunk region of a seventy-two hour embryo without a circulation; Experiment A, 1913. The extreme anterior end of the intermediate cell mass, icm, is represented by deeper staining cells between the notocord and intestine, Int; the primary kidney ducts, ll'D, are lateral to the mass.

Fig. 32 Represents a more posterior section through the same embryo; in this region the intermediate cell mass, icm, is more extensive in cross-section and its cells are further differentiated than those in the more anterior region, Ph, periblastic material and large nuclei between the embryo and the yolk.

Fig. 33 A still more posterior section through the same seventy-two hour embrj-o as figures 31 and 32. This section is posterior to the place where the intermediate cell mass and gut endoderm fades out into the indifferent cell mass, EB, which may be considered to represent the end-bud mass; Pb, the yolk periblast.

The condition in the seventy-two hour embryo is, of course, quite early and the cells are not yet as a rule taken into the circulation. The appearance shown in these figures is exact h' that of ver}^ slightly younger normal embryos in which a circulation would later be established. The figures, however, were made from sections of an embryo that had no heart beat at the time of its fixation, and, therefore, there is no chance that any of these cells could have become misplaced by having been circulated or carried about.

The most careful study with the highest power of the microscope has failed to reveal any type of cell in the intermediate cell mass other than the early erj^throblast, and in later conditions one finds here only red blood corpuscles. In other words, this chief stem blood anlage of the bony fish seems to be a specific red blood cell forming mass. This mass was first discovered by Oellacher in 1873 and has been shown by numerous investigators, Zeigler ('87), Winckebach ('86), Henneguy ('88), Sobotta ('94), FeUx ('97), Swaen and Brachet ('99-'01) and others, to be peculiar to the Teleosts.

It is important to know that none of these investigators have 3"et recorded any type of blood cell arising from this mass other than the erythroblast. Of course, it may be argued that no special study was made of this particular point. Yet it is certainly true that if lymphocj^tes or leucocytes had been present to any extent, they should have been observed by many of these very capable and careful workers. It has recently been claimed by Maximow r'09), Dantschakoff ('07, '08) and others, that the popular opinion that leucocytes are very late in arising is erroneous since they actually arise just as early as the erythroblast. Then it seems all the more probable that if lymphocytes or leucocytes had been present in this intermediate cell mass such cells would have been discovered, since the mass has been carefully investigated right up to the moment at which it becomes swept into the circulating plasma.

Various mvestigators have differed as to the vascular products derived from the intermediate cell mass. Some ha^'e claimed that it forms only blood cells and no vascular endothelium, Sobotta ('02), while others have attributed the production of cardinal veins or venous endothelimn as well as the blood cells to this mass, FeUx ('97), and finally others, Swaen and Brachet ('01, '04) in particular, have considered this to be the source of both the aorta and carchnal veins as well as the red blood cells. I have spent considerable time in a study of this question and am inclined to believe that the endothelium of the cardinal veins and aorta arises from the mesenchymal cells surrounding the intermediate cell mass, which are different in nature from the cells actually constituting the mass. Yet it must be admitted that up to the presenf moment a complete demonstration of the origin of aortic endothelium from the cells about the periphery of the intermediate cell mass has not been satisfactorily shown. This question will be more fully considered in another section.

In the embryos in which the red blood cells remain confined within the median region throughout life these cells develop in a normal manner and become completely differentiated into typical ichthyoid erythrocytes and exist as such for some time. Finally, however, for reasons at present impossible to state but Hkely associated in some way with an insufficient supply of oxygen, these erythrocytes begin to degenerate and in old embryos of sixteen to twenty days only a very few or in some cases none are left in the large intermediate vessel. Mesenchymal cells seem to wander into the mass of erythrocytes and may take part in their destruction.

Figure 41 is a section through the intermediate cell mass of a sixteen-day old embryo and presents this degenerate condition. The erythrocytes are all small and necrotic and many mesenchymal cells are scattered among them.

As we shall see below, the power of existence of the erythrocytes is very much stronger in the blood islands where aeration is no doubt considerably better than in the intermediate cell mass.

3. Blood islands of the yolk-sac, their origin and development

The question of origin of blood cells on the yolk-sac of the Teleostian embryo has been a much debated topic. Almost all of the earliest workers claimed that blood arose in the yolk-sac islands of the bony fish just as in other meroblastic eggs. The later workers, however, have denied this statement and hold that the bony fish forms an exception to the rule, and is the only type of meroblastic embryo in which blood cells do not occur in islands on the yolk-sac.

It has been frequently admitted by several recent workers that certain wandering mesenchymal cells do migrate to the yolksac from the embryo and there form isolated blood cells or small cell groups, but that this blood formation is insignificant in amount as compared with the great blood forming intermediate cell mass.

The yolk-sac of the bony fish is pecuHar in this connection. In most meroblastic embryos there is a definite mesodermic layer or membrane between the ectoderm and entoderm of the yolk-sac, and it is in this mesodermal layer that the blood islands arise. When one examines the yolk-sac of the Teleost embryo, the mesodermic layer is found to be largely, if not entirely, absent. Thus, the ectoderm hes directly over the yolk periblast which may be considered to represent the primary entoderm. Between these two layers many long spindle-shaped mesenchjrmal cells are noticed on careful examination, but these cells in the specimens examined are never arranged in a definite continuous layer.

Goodall ('07) has recently stated that in the sheep embryo, the yolk-sac mesenchyme is not to be considered a continuous layer, but consists merely of diffusely scattered wandering mesenchymal cells. These mesenchymal cells in the sheep as in the fish finally collect into groups and such groups ultimately give rise to the blood islands. In the fish it would seem as though the entire ventral or yolk-sac mesoderm, the chief source of blood formation, had been in its phylogenetic development incorporated or drawn into the body of the embryo as the intermediate cell mass, and only a few cells lag behind or later wander out to form the collections of mesenchymal cells upon the yolk. Ontogenetically there is no longer any indication of a mechanical drawing-in process but the wandering out of cells may be readily observed. It is also easily conceivable that this condition probably differs in different species of Teleosts. Therefore, some species may really form no blood cells in the yolk-sac, while again others might have an almost complete mesenchymal layer in the sac and in such a case would probably give a typical blood island arrangement. Whereas, an intermediate condition would be well represented in the species of Fundulus here studied in which there are numerous disconnected wandermg cells later grouping themselves to form the blood islands on the yolk-sac.

The appearance of the wandering cells as they radiate out from the caudal end of the embryo on to the yolk-sac is strikingly similar to that shown by the cells wandering away from the central tissue mass in a living tissue culture. The cells are elongated spindle-form and all are moving straight away from their seat of central origin. This phenomenon is well illustrated by the numerous figures of tissues growing in culture media and I shall give illustrations of it in a special study of this subject now in preparation.

In all of the non-pelagic bony fish eggs investigated up to now, the chief blood forming cells are without exception the intra-embryonic intermediate cell mass, and this mass is claimed to form both vessels and blood. While in the pelagic type of bony fish egg the mass is usually concerned in the formation of vascular endotheUum, and the blood cells only arise after the embryo is hatched and free swimming.

This peculiar specialization in intra-embryonic blood formation which seems typical for the bony fish has caused the yolksac formation of blood to be almost completely neglected or overlooked by recent investigators. Yet in the species upon which I have experimented there is no doubt whatever that blood islands do arise on the yolk and their origin is from the wandering mesenchymal cells. The wandering cells may be connected in some manner with the intermediate cell mass, yet the presence of the islands cannot be explained in the way Swaen and Brachet COl) have attempted to account for the yolk-sac blood. They assume that the islands are pushed out laterally as branches or portions of the intermediate cell mass. In many cases no direct continuation of cells is traceable between the yolk islands and the intermediate cell mass and even in extremely young embryos yolk islands may appear on the ventral yolk surface at a great distance away from the intermediate cell mass.

The group of four figures, 36 to 39, indicate the progressive patterns assumed in life by these yolk islands. In the very early stage, figure 36 shows separate collections of cells here indicated by stippling. These groups then become confluent as in figure 37, then more or less net-like in appearance with certain nodes or portions thicker and darker than the general net. In these nodes cell proliferation or blood formation is more active. Finally a typical vascular network arises which goes to make up the capillary yolk circulation of the embryo.

These appearances, as stated above, are not readily distinguishable in the very young embryos, yet with a little experience and a high power microscope any one may convince himself that the blood island formation proceeds to a very definite and considerable degree in these embryos.

Figure 34 represents a cross-section through the yolk-sac of an embryo of seventy-two hours old. The ectoderm of the yolk-sac now becomes two-layered, this continues to thicken as age advances until finally in old embryos the yolk-sac ectoderm is many cells thick and often folded and complex in arrangement sometimes showing villus-hke processes. Beneath the ectoderm a group of early erythroblasts or blood cells is illustrated. These cells lie immediately upon the yolk mass here indicated by the heavy dark granules. The appearance of the cells in this blood island anlage are closely similar to those shown in cross sections of the intermediate cell mass in figures 31 and 32. The cell nuclei and general cellular arrangements of the two tissues are seen to correspond in appearance, and the manner of differentiation followed in both cases is identical.

In figure 35, a group of five early erythroblasts are shown which were present in a neighboring blood island. They had loosened themselves from the general island mass and appear very much, if not exactly, similar to the early erythroblast seen separating themselves from the compact mass, the intermediate cell mass (fig. 32). The nuclei in all cases are typically those of early red blood cells and the cytoplasm just begins to stain a very pale pink color characteristic of the halo seen around the young erythroblast. All of the cells shown in these yolk islands, both in the earliest condition of the island and in the late old yolk vessels of embryos without a circulation are invariably of the erythroblast or erythrocyte type. In no case has any type of lymphocyte or leucocyte been present in these yolk islands except as late wandering cells.

Not all of the wandering cells which are found on the yolksac go to form blood cells since many of them are future chromatophores or future endothelial vessel cells. The types, however, are distinguishable in rather early stages and do not seem to be in any way related except that all are of mesenchymal origin. The chromatophores as before mentioned, often come to lie along the walls of the blood vessels.

The early yolk-sac is non-vascular, the blood masses being completely uncovered bj^ endothelium. Later endothelial walls are formed around the blood cell masses and a vascular network is established in the yolk-sac of the normal embryo though poorly formed in the individuals without a circulation. All of these yolk vessels seem to arise by arrangement of wandering mesenchymal cells. Certain of these cells elongate and group themselves in such a way as to form vessel tubes. After the vessels are formed they may then be seen to send off buds and sprouts in the manner Clark ('09) has described in amphibians. The difference between the cells giving rise to the vascular endothelium and those forming the blood cells is not distinguishable in early stages. Yet after considerable study and careful observation, nothing has been observed that would indicate that these vascular endotheUal cells possess the power to change into the blood cell type, nor is there any evidence to indicate that cells having once assumed even the earliest blood cell type are capable of metamorphosis to form endothehal cells. It is impossible to state emphatically that the vascular endothelium of the yolksac in all Teleosts arises in the same way as that described here for Fundulus embryos. But any one familiar with the very complex yolk circulation of the trout family, in the light of the above knowledge is scarcely justified in assuming that this network of vessels is completely derived from outgrowths from the aorta and cardinal veins within the embryo as Sobotta

Fig. 34 Section through the yolk-sac of an embryo seventj^-two hours old, without a blood circulation. A group of cells forming a blood island are distinguished by a slight condensation of cytoplasm about their nuclei; Experiment A, 1913; Ec, the ectoderm several cells thick; Bi, the cells of the blood island; FA;, granular yolk.

Fig. 35 Young erythroblasts just isolating themselves in another island on same yolk as figure 34. Compare the early blood cells with those of figures 31 and 32, in the intermediate cell mass.

Figs. 36 to 39 Illustrate the progressive steps in the development of the network of yolk-sac blood islands.

('02) would have one believe. It is probably true that in the trout family also, wandering mesenchymal cells are of great importance in the formation of extra-embryonic vascular endothelium. There is a strong possibility, as admitted by Swaen and Brachet COl) that some blood cells are also formed on the yolk-sac of the trout from wandering mesenchymal cells.

The blood cells in the yolk islands increase by mitotic division and soon become very prominent features in those embryos without a circulation, so that in old individuals of eight or ten days the entire posterior and ventral regions of the yolk are almost completely covered with red blood islands. The corpuscles in these blood islands persist in a more or less normal condition for a considerable length of time.

4. Fate of the blood corpuscles in embryos without a circulation

Figure 42 shows the corpuscles in a yolk-vessel of an embryo of sixteen days old in which the blood had never circulated. The vascular endothelium is well formed about the corpuscles and proliferation or multiplication of blood cells has completely ceased, the nuclei are very densely stained and somewhat pycnotic in appearance suggesting a more or less atrophic condition of these erythrocytes.

Figure 40 is a cross-section of a vessel from the yolk-sac of a normal embryo of seven days old. In this the vascular endothelium is also fully developed, large chromatophores have spread themselves along the vessel wall and the erythrocytes are in a vigorous physiological state. The nuclei are lightly staining alveolar structures quite different in appearance from those of the erythrocytes in the older embryo of sixteen days that has never had its blood to circulate. Yet the erythrocytes in the old non-circulating embryo are ladened with haemoglobin and certainly function to some degree.

Figure 41 in the same group is a section through the intermediate cell mass of a sixteen-day embryo without a blood circulation. This is the only intra-embryonic vessel which contains blood cells in this individual. The vascular endothelium is completely disintegrated and has disappeared, and the very degenerate small erythrocytes are now intermixed with mesenchymal cells. In a slightly older embryo, all of these blood cells have disappeared within the tissue as if the invading mesenchymal cells had really assimilated or destroyed the old blood cells.

It is thus seen that in these non-circulating individuals, although red blood cells arise in a perfectly normal fashion and differentiate as completely as in the control embryos, yet they are not capable of maintaining their fully developed condition. Sooner or later they undergo degeneration and finally are completely absent from the body of the embryo.

It is noticed in all cases that very soon after the erythroblasts become completely surrounded by endothelium, they gradually lose their power of multiplication and then differentiate into typical erythrocytes. Before the vascular wall has completely enclosed the eiythroblasts, all groups often show many cells in active mitosis, and as I shall bring out below those spaces in which blood cells multiply both in the embryo and in the adult are spaces not completely surrounded by vascular endothelium.

In examining figures 40 and 42, it may be of interest to note that the erythrocytes in figure 40 are the typical ichthj^oid type of Minot ('11), while those in figure 42 are what Minot would describe or term, the sauroid type; that is, erythrocytes in which the nucleus has become sUghtly more degenerate or more densely staining than in the ichthyoid type and in which the cell body is smaller. This sauroid type of corpuscles Minot has designated as being characteristic of reptiles and amphibians, and the condition in these embryos without a circulation indicates the very artificial nature of the proposed classification of Minot. The cells are, of course, ichthyoid but are degenerate and, therefore, assume the 'sauroid type.'

It is difficult for one to believe that all of the functioning erythrocytes in the amphibians and reptiles really have a degenerate nucleus for the simple fact that mammals have blood corpuscles which have completely lost or discarded their nuclei. It must here be remembered that birds are as truly derived from reptiles as are the mammals, in fact, the connection between the reptiles and birds is even closer. Yet the degenerating nuclei of the reptilian blood corpuscle is able to maintain itself in the corpuscles of the bird, although it is according to Minot, so far gone as to degenerate entirely in the corpuscles of the mammal. Such classifications are extremely misleading as they convey to the mind the impression that there is a continuous developmental or evolutionary chain of events illustrated in the blood cells of the different vertebrate groups and actually repeated in the development of the blood in the mammals. The "biogenetic law" is scarcely vigorous enough at present to be submitted to such a strain.

Finally in considering these yolk-sac blood-corpuscles, one must mention the possibihty of origin from the yolk periblast or endoderm. It is often stated even in modern text-books and contributions that blood-cells may arise from endoderm and that the primary blood forming layer was actually the endoderm. It is very positively certain that none of the blood cells on the yolk islands of the fish arise from the periblast, but all are derived from wandering mesenchymal cells. The sharp distinction between endoderm and mesoderm is not a thing of any great or definite importance, since everyone recognizes the primary association and origin of mesoderm from the endoderm and the ectoderm. When the mesoderm is once formed, however, it contains within itself a blood forming anlage. It must be further remembered by speculators on the phylogeny of the vertebrate blood that the invertebrate animals, many of which possess highly functionating; white blood cells, amoebocytes, as well as oxygen carrying corpuscles, are thought to derive these cells and the vascular endothehuni from mesenchyme and not from endoderm.

Fig. 40 A highly magnified section through a yolk-sac vessel in a normal embryo of seven days; Et, the vascular endothelium with chromatophores along it. The large beautifully developed erythrocytes are seen in the lumen.

Fig. 41 An equally magnified section through the intermediate cell mass in an embryo without a circulation when sixteen days old; Embryo 413. This is the only intraembryonic blood present, the vascular endothelium, which probably at one time surrounded the erythrocytes, has broken down and mesenchymal cells, Men, are now intermixed with the small degenerate erythrocytes, Ery, which should be compared with the normal ones in figure 40.

Fig. 42 Shows erythrocytes in a yolk-sac vessel also in Embryo 413, at the same magnification as in the two preceding figures. These erythrocytes are in a better condition than the intra-embryonic ones, j'et they are very degenerate as compared with those of figure 40; all, however, still contain haemoglobin.

5. Has vascular endothelium a haematopoetic power?

It has been mentioned in describing the origin of blood in various parts of these embryos that no observation could be interpreted to indicate that blood corpuscles ever arise from vascular endothelium. The endothelium of vessels containing blood never presents any cell in a transitional stage. These experiments, I think, furnish a crucial answer to persistent claims that vascular endotheUum has the power to change into various types of blood corpuscles. If vascular endothelium had such a power, then one might expect that this power would show itself in cases where it was most needed, for example, in these embryos in which the blood has never circulated. The blood cells are conf ned entirely to the intermediate cell mass and to the blood islands on the yolk.

The heart and aorta and numerous vessels in the head and anterior portion of the body are lined with typical vascular endothelium, yet in no instance has it been found that one of these vessels contains a single red blood cell in any stage of development. From these experiments, one is warranted in making the bold assertion that the endothelial Uning of the heart and aorta is perfectly incapable of gi\'ing rise to any type of blood cell. This fact has been mentioned in considering the endothelium of the heart. When we now refer to figure 43, a section through the anterior region of a four-day-old embryo without a circulation, two dorsal aortae are shown. These vessels are lined by typical embryonic endothehum but are completely empty so far as cellular elements are concerned. This is true of the dorsal aortae of all embryos from the earliest to the latest stages when the circulation of the blood has been prevented. FeUx ('97) has also noted the fact that the aortae in early normal Teleost embryos are invariably free of blood cells.

Fig. 43 Section throush can anterior body region of a four-day embryo without a circulation; P:xperiment 11, 1912, Embryo 6. The dorsal aortae, DA, are seen lined by typical embrj^onic endothelium, yet they are throughout completely empty, never containing blood cells.

Fig. 44 A section through the cardial vein of a similar four-day embryo without a circulation; Experiment 11, 1912, Embryo 4. Again the embryonic endothelium, Et, is well formed but the lumen of the vessel is parked with erythroblasts; Ec, ectoderm; Mes, mesenchyme; En, entoderm.

Figure 44 accompanying figure 43 shows a striking contrast in the contents of the cardinal vein. This section is through a more posterior region of the same embr^^o. The vascular endothelium is here also well differentiated and the vessel is completely packed with early erythroblast some of which are still dividing. None of these erythroblasts, however, have been derived from the vascular endothelium and were actually present before the endothelium was differentiated.

The only source of intra-embryonic blood is from the blood anlage which is contained within the intermediate cell mass, as a rule. But in the splitting away of this mass from between the lateral plate and the somites, it is, of course, conceivable that some future blood forming cells might be left either in the somitic portion or the lateral plate portion. In such cases all those organs arising from regions which had been in contact with the intermediate cell mass either medially or laterally might be contaminated with blood forming cells. If the separation of the blood cell anlage takes place in a clean and complete manner in the indi\'idual embryo, then I believe the statement is true that all the intraembryonic blood will be contained in the intermediate cell mass or cardial veins which amount to the same thing.

6. The origin of lymphocytes and leucocytes or so-called white blood corpuscles

We may now turn to a consideration of the origin of lymphocytes and leucocytes or cells other than red blood corpuscles. Many authors have claimed from observations on various embryos that these cells are entirely distinct in their origin from the origin of the red blood corpuscles. Both types, however, arise from the same germ layer or mesenchyme. It has been repeatedly pointed out and seems to be thoroughly substantiated by fact that the lymphocytes and leucocytes in their first appearance are always interstitial in position and are only later contained within the vessels. Whereas, the erythroblasts are invariably formed or divided off into the vessels. In other words, the red blood cell formation tends to be towards or into the vessels and the formation of white blood cells seems to be extra-vascular or interstitial.

It is recently claimed by Goodall ('07) that in the haematopoetic organs of the sheep embryo such as the liver, there are definite groups of proUferating cells forming the various types of white blood corpuscles, and these are distinctly isolated from other groups of proliferating erythroblasts. In the bone marrow this same state of affairs has been described, and in a number of diseased conditions of human marrow I have observed that certain nests or groups of cells were giving rise to leucocytes while other separate groups consisted of erythrocytes. This observational e\'idence might seem to indicate that white and red blood cells were arising from different parent cells. Yet in normal embrj^os it is very difficult to obtain material which will conclusively establish such a position, since both types of cells are swept around by the circulation and are intimately inter-mixed in all of the haematopoetic organs.

It would seem that in these experimental specimens in which the blood was prevented from circulating that there might possibly be some way to distinguish completely the source of origin of the white blood cells from the red blood corpuscles if these sources were really different. Should the two types of cells arise from the same common stem cell or parent cell, then the white and red blood cells should be invariably found in association in all embryos. If the two types of cells had different origins they might be found to occur in separate regions of the body and the various sources could thus be readily differentiated.

As frequentl}' stated, in the early intermediate cell mass and among the cells immediately developing out of this mass no leucocytes or lymphocytes are found. The yolk-sac blood islands also consist entirely of cells of the erythroblast type. These observations are in accord with those of all other investigators studying the development of the blood in the bony fish. They have invariably described the intermediate cell mass as being the source of red blood corpuscles and no one has ever recorded either Ij^mphocytes or leucocytes as arising from this mass.

The only cells within the embryo which resemble lymphocytes or leucocytes in their general structure and staining capacities have been found in the anterior portions of the body and in the head regioii, of the young embr^^os. In very young embryos of seventy-two hours, numerous isolated cells and occasionally small groups of cells are found within the mesenchyme which present a pecuUar appearance. The nuclei are more or less dense, the cytoplasm very small in amount in many and in others very extensive, and staining with a color quite different from that of other cells in the embryo.

Figiu-e 45 shows a section through the head just behind the optic stalk of an embrj^o of seventy-two hours. In this section there is seen a nest of the above-mentioned cells, several are poljmuclear and present various pecuUar appearances. The mesenchyme within this region is in active mitosis.

Figure 46 is also taken from the anterior end of an embryo and shows two large mesenchymal nuclei with numerous small leucocyte-like ceUs within the mesenchyme. Numerous pigment granules are also present in these mesenchyme cells. Some of the cells present nuclei of the polymorphonuclear type.

Figure 47 shows an enlarged binuclear cell and indicates the fine granular nature of the cytoplasm. Such cells resemble very closely the embryonic white blood cells.

Figure 48 represents a section through the anterior end of an embr>^o, and shows an endothehal artery which is entirely empty of blood cells. Within the mesenchyme, near the vessel, are two of the leucocyte-hke cells. Other cells in these embrj^os resemble very closely ordinary lymphoc}i:es and these

Fig. 45 A section immediately posterior to the optic stalk in an embryo without a circulation when sevent3^-two hours old; Experiment A. 1913. A nest of peculiar finely granular cells lies in the mesenchyme which contains many dividing cells; Br, brain; Ofv, optic vesicle.

Fig. 46 Cells from a four-day embryo; Experiment 11, 1912; Men, mesench3Tne nucleus; small leucocyte-like cells are grouped in the neighborhood of chromatophores.

Fig. 47 A binuclear leucocyte from a sixteen-day-old embryo.

Fig. 48 Section through one of the dorsal aortae in a four-day embryo; Experiment 11, Embrj-o 6; embryonic leucocytes in the mesenchyme.

also are within the tissue spaces and not in the vessels. None of these cells are found in associations with the early red blood cells.

I have examined a number of smears of heart blood, spleen pulp, bone marrow and peritoneal fluid from adult Funduli and have found within these specimens numerous coarse granular leucocytes, lymphocytes and various types of wandering cells. The embryonic cells above described all show a more or less degenerate appearance, but if they can be classed as any type of white blood cell their origin is definitely removed and entirely distinct from that of the red blood cell. In appearance they are as closely similar to embryonic leucocytes as are the cells designated by other investigators to be of that nature. It seems to me that the only possible method of differentiating between the origins of white corpuscles and red blood cells is to prevent their association in the circulating body fluids. These experiments along with numerous observations do show that the red blood cells arise in a region distinctly separated from those localities in which the white blood corpuscles are formed.

When one examines a specimen such as those in which Maximow, Dantschakoff and others have described the origin of white and red cells from a common stem cell, it is impossible to be absolutely certain that the two types of cells do arise from the same individual stem cell. The stem mother cell is shown within the mesenchyme, near this on the one side are various early lymphocytes or leucocyte-like cells, and on the other side are the various stages in erythroblast development. Each type of cell graduates directly back to the stem mother cell or to a mesenchymal cell. This much may be freely admitted, but to say more is merely a matter of guess or interpretation. Since it is absolutely impossible on fixed material to make the definite statement that both of these two types of cells have arisen from the one individual stem mother cell. The observer must actually witness the stem mother cell divide into two cells, and then observe one of these two cells either by differentiation or continued division give rise to white corpuscles, and the other either by differentiation or continued division give rise to erythroblasts before the existence of a common mother cell is proven. This very necessary observation has never been made and all of the evidence in the present literature seems insufficient to warrant the conclusion that such a thing does actually take place, whereas there is considerable evidence to indicate that white and red blood cells probably arise from two different mesenchymal cells. Of course, these two parent mesenchymal cells may be, so far as our powers of observation go, indistinguishable. Yet this would not indicate that they were not different in their potentialities. One of the two mesenchymal cells might be capable of giving rise to the various types of white blood cells depending upon the conditions of differentiation and function, while the other apparently similar mesenchymal cell could on account of its internal difference give rise to erythroblasts. It is a little strange at least that the white blood cells arise far interstitially, while the erythroblasts have such a decided tendency to proliferate into the sinusoids or vascular spaces if they both arise from a common stem cell. The two environments in which they develop could scarcely account for the differences between red and white corpuscles, since in the body of an embryo in which the blood circulates there are several places where the two types of cells develop side by side, as Maximow and others have described. The reasons for the differences are the internal differences between the mesenchymal cells from which the two types of corpuscles arise.

The white blood cells and red blood cells, although both are derived from mesenchyme, arise from mesenchymal cells which have already differentiated sufRcienctly far not to be interchangeable. This statement is probably true also of the vascular endothehum in its relationship to blood forming mesenchyme. The embryonic mesenchymal cell if taken in an early enough stage could no doubt give rise to other mesenchymal cells which would later form any of these different type cells. When, however, differentiation has proceeded to some degree, sufficiently far to form what is termed by embryologists an organ anlage, and yet not far enough to make it possible to distinguish between the appearance of various mesenchymal c^lls, they are then, nevertheless, different in their potentialities.

The mesenchymal cell with the power of formmg vascular endothelium is probably very diffusely distributed throughout the embryonic body as well as the yolk-sac. Numerous investigators have supplied evidence indicating this fact. On the other hand, the mesenchymal cells which are to form the erythroblasts are in the bony fish very definitely localized. The latter cells are chiefly confined to the intermediate cell mass, but in addition other erythroblast forming cells wander out probably from the same source of origin, the primary intermediate cell mass, to become distributed on the yolk-sac.

Finally, the mesenchymal cells which are to give rise to lymphocytes and leucocytes of various types seem in Fundulus embryos to be more or less localized in the head and anterior region of the body and do not seem to be particularly associated with vessels. For this reason in early embryos the first apparent lymphocytes and leucocytes are found in the head and anterior body regions, and even in older individuals such cells are more abundant here than in other portions of the body. Yet these cells are doubtless of a roving or wandering type and may finally become scattered throughout the embryo's body. While the non-migrating red blood corpuscles rarely if ever leave their original sites of differentiation.

7. Environmental conditions necessary for blood cell multiplication and differentiation

The above facts and interpretations lead us to a consideration of the later conditions of cell multiplication and differentiation. Are the so-called haematopoetic organs of the embryo and even the adult actually haematopoetic, or are they merely favorable localized environments in which various types of blood cells may multiply or reproduce themselves throughout the life of the embryo or individual? There is little doubt, from the recent suggestive studies on the cultivation of tissues in artificial media out of the body of the organism, that certain environmental surroundings are conducive to cell proliferation and growth while other environments inhibit these processes and tend to favor differentiation and functional activity.

Many points mentioned in the previous pages indicate that blood cells change their mode of behavior as the conditions of the embryonic vessels and body are changed during development.

A careful consideration of various embryos as well as the different regions of the same embryo suggests that erythroblasts only multiply in spaces unlined by endothelium, whether these be on the yolk-sac, or within the embryonic hver, spleen or bone marrow. . The unUned spaces thus afford an environment in which for physical or chemical reasons the erythroblasts are able to multiply and reproduce themselves. When, however, such spaces or channels become converted into endothehal hned tubes then the erythroblasts tend to differentiate into erythrocytes and very soon cease to reproduce themselves any further iii this location. In the intermediate cell mass of the fish embryo for instance, one notices in early stages many dividing erythroblasts. Just about the time that this mass becomes completely enclosed by vascular endothehum, this division process slows down and finally ceases although the confined erythrocytes may never be able to leave the vessel. Soon after being surrounded by vascular endothehum the erythrocytes assume a passive non-productive state and remain in this condition throughout their existence.

One of the earhest places of blood cell prohferation or haematopoesis is the sinuses on the yolk-sac. Almost as soon, however, as these sinuses become converted into the yolk vessels, blood cell production ceases in the yolk-sac and only blood circulation takes place.

The haematopoetic processes are then transferred to the embryonic hver and this in most vertebrate embryos is an important seat of blood cell multiplication. The multipljdng erythroblasts are never enclosed by endothelium. In other words, they are not within the vessels but their final products are invariably budded or divided off into the sinusoids. Finally, the spaces in which blood cells multiply in the liver become obhterated or converted into endothelial lined vessels and spaces and very soon after this takes place the haematopoetic processes cease in this organ. The hver cells themselves or the interstitial tissues of the liver are not blood forming cells, the only blood formed within the hver arises from existing blood cells which are carried there in the circulating cmrent.

In this manner the multiplication of blood cells is shifted from place to place and becomes more and more localized until ultimately in higher animals the red bone marrow is the only body tissue in which these open spaces furnishing the required environment exist. It is, therefore, the only body tissue in which erythroblasts Hve and continue to multiply and give rise to the entire stock of red blood corpuscles which circulate throughout the body.

Blood cells always multiplj- in the unhned spaces but normally never multiply to any extent within closed vessels. It might be possible that certain abnormal growth tendencies on the part of the endotheUiun of the vessels and sinusoids in the bone marrow might cause an inclusion or vascularization of the spaces in which blood cells multiply and this growth might indirectly result in the cessation of the production of red blood corpuscles. This might be experimentally tested should some method be devised by which the growth of vascular endothehum could be so stimulated as to close the spaces of the bone marrow.

8. Question of haematopoetic organs

In the fish embrj^o the haematopoetic function of the hver is not of great importance. Yet in the liver of normal individuals numbers of blood cells are alwaj^s found and numerous dividing blood cells are present. In the non-chculating embryos the blood is unable to reach the hver and in such cases there are no blood cells of any type to be seen in this organ.

Figure 49 represents a section through the gall bladder, bile duct and body of the liver in a sixteen-day old embryo. In

Fig. 49 A section through the liver of a sixteen-day-old embryo, without a circulation; Embryo 413, 1913. The gall-bladder, GB, and bile duct are seen connected with the intestine, hit; the liver, L, is a compact mass containing neither vessels nor any type of blood cell. A, the well developed dorsal aorta is lined by endothelium but its lumen is completely empty except for a slight coagulum near the center; icm, if followed posteriorly leads into the remains of the intermediate cell mass; WD., nephric duct; Xch, notocord.

this individual the heart is a sohd string and the blood had never circulated. The liver presents a dense appearance, no blood vessels are seen and blood corpuscles are entirely absent. The general differentiation and condition of the tissues are, however, fairly normal and not at all degenerate. The intestinal epithehum is typical in structure. .Above the intestine the well differentiated dorsal aorta is shown with connective tissue fibers abundantly present in its wall and a definite endothelial lining. The lumen of this aorta, however, has never contained any type of blood cells and the only sohd particles within it are a slight coagulum near the center of the vessel. Above the dorsal aorta are the two Wolffian ducts and between them under the notochord are a few mesenchymal cells which represent more posteriorly the remains of the intermediate cell mass. Almost all of the erythroc3'tes in this mass have .completely degenerated or have been destroyed by mesenchymal cells.

The embryos without a circulation thus furnish a definite means of estabhsliing the actual haematopoetic value of any organ. They demonstrate that unless the blood current reaches the organ and thereby introduces embryonic blood cells into it the organ iself is incapable of giving rise to blood cells.

This experiment also demonstrates with equal force the inability of vascular endothehum to form blood cells in the I can see no reason if vascular endothehum possesses a blood forming power whj' the aorta and other interior vessels of these embryos are invariably emptj' and never contain any type of blood cell. It cannot reasonably be claimed that this inabihty is due to the abnormal condition of the embryo having taken away the power of the endothelium to form blood cells, since it is so absolutely demonstrated that real blood forming material in other portions of the embryo possesses its perfectly normal capacity to produce blood and does produce it in a very abundant fashion.

These embryos furnish no evidence to indicate that there is any connection or association betw^een the mesenchymal cells which are to form the connective tissue and those destined to form blood cells. There is no instance of a tendency for connective tissue cells to change into blood cells or of blood cells to give rise to any type of connective tissue cells.

Finally, one may conclude that the blood cells like many other specific tissues and organs have a definite localized specific anlage and that this anlage is distinct and separate in most cases from that of the vessel linings. In some cases, however, the blood and endothelial anlagen may come into intimate association, yet even here the two are probably of different mesenchymal origins.

V. A consideration of the experimental study on the origin of blood in Teleosts in relation to the more recent studies on the origin and development of vessels and blood cells

1. Introduction

This expermiental study of the origin and development of blood and vessels relates itself to three more or less separate fields of investigation.

In the first place, the manner in which the blood anlage in Teleosts has separated itself as a unique intermediate cell mass has caused it to be studied as a special subject somewhat isolated from the more general literature on the development of blood in other vertebrates. Yet one very soon appreciates the mistake of this isolation since contributions such as those of Felix ('97) and Swaen and Brachet ('99, '01, '04), in particular, on the Teleosts are of more general importance than most investigations dealing with the broad subject of blood development in the vertebrates. The very fact that in this group the blood anlage is so pecuUaiiy localized in the embryo lends itself as a great aid to the solutions of many questions of haematopoesis or blood genesis.

Secondly, a consideration of the origin and formation of the heart lining or endocardium and the vascular endothelium, in these embryos which have developed without having had plasma or fluids to circulate within their vessels, may furnish much important data towards a final solution of the origin and significance of endothelial lining cells, and the manner of spread and distribution of such cells through the embryonic body and the yolk-sac.

Lastly, such an experimental study bears closely upon the general questions of relationship between different blood cell types. The time and place of origin of the different cells, and the developmental relationship and powers of transmutabihty existing between various sorts of blood corpuscles as well as the endothehal lining cells of the vessel walls are all problems upon which the experimental results discussed above may throw light.

Each of these three divisions of the problem embraces an extensive and often cumbersome literature which it would be quite out of place to consider in detail at the present time. We shall, therefore, only consider the bearing of the facts recorded in the previous pages upon the opinions and positions maintained by the more recent investigators of the origin and development of blood and endothelium.

2. The specific problems of blood and vessel formation m the bony fish

It may be well to review first the special problems and questions involved in the development of the blood in Teleosts as a group. According to Swaen and Brachet ('01), the mesoblast in the middle and posterior regions of the trout embryo is arranged in two parts, a median primary somite portion and an outer primary lateral plate part. The lateral plate in the midbody region then divides off a portion immediately adjacent to the somites to constitute the intermediate cell mass. Lateral to this a second part of the lateral plate is separated off to form the primary nephric duct. In the mid-region of the body the somites become separated from the primary lateral plate and the lateral plate pushes or grows towards the median plane and gives off a keel shaped mass between the somites and hypoblast. This mass unites in the median plane with a similar mass from the other side and here forms a large cell group triangular in cross-section, the intermediate cell mass. In the posterior region of the embryo a similar mass pinches away from the primary lateral plate and becomes the posterior continuation of the intermediate cell mass.

Anterior of the first somite in the unsegmented mesoblast of the head this division or pinching away also takes place. Thus the intermediate cell mass of the body becomes continuous with a definite lamella of the head. This well definied topographical portion of the embryonic mesoblast, the intermediate cell mass and cell lamella, is, according to Swaen and Brachet, the only material wliich gives rise to the heart, the chief vessels and the blood in the embryo. This description by Swaen and Brachet ('01) agrees very closely with that formerly given by Felix ('97), except that Felix disagrees in not deriving the aorta from the intermediate cell mass but from the sclerotoms.

The observations made upon the intermediate cell mass in Fundulus are in close accord with this summary. But no attempt has been made to solve the detailed question as to whether the aorta is derived' from the intermediate mass or from the sclerotoms. It would seem that this vessel might arise from either source and still be formed from practically identical cells. Since in the separation of the primary lateral plate from the somite it is easily conceivable that some cells which generally accompany the primary lateral plate might be left as part of the lateral portion of the somite. This lateral portion of the somite is the part which later separates as the sclerotom so that the cells destined to form the aortic endothelium might occur equally well within the intermediate cell mass or within the sclerotom. Their location might vary among different species or even among individuals, and yet these aortic cells would be. derived from the same genetic source.

Swaen and Brachet also indicate the head mesoblast as separated into three portions: the intermediate cell mass close to the top of the pharynx, the lateral plate split into two lamellae and the general head mesoblast close around the brain. The intermediate cell mass is more intimately connected with the splanchnic layer of the lateral plate. The pharynx widens in forming the gill pouches which continue to grow dorsally and finally separate the intermediate cell mass into two portions, one part thus comes to lie ventral of the pharynx and the other part dorsal. The ventral portions, at first solid masses below either side of the pharynx, begin to migrate towards the middle Une. The two masses fuse into one, spaces are developed in the mass and finally the endothehal Uning of the heart is differentiated out of this group of cells. The lamellae of the side plate become separated and the space between them gives rise to the pericardial cavity. Oellacher (73), Wenckebach ('86), Henneguy ('88), and Sobotta ('94) have all described the origin of the heart in Teleosts in much the same way.

Several of these investigators, Wenckebach, Swaen and Brachet and others, have called attention to a small mass of cells derived from the heart anlage which comes to lie beneath and outside the heart endothehum. This mass of cells has been claimed to wander away from below the pericardium and later to give rise to vessels and blood on the yolk-sac. In the noncirculating Fundulus embryos, however, neither vessels nor blood are formed on the extreme anterior portions of the yolk-sac. I have seen nothing in my studies which would indicate that any cells left over from the heart formation had wandered upon the yolk or given rise to blood cells or vascular endothelium.

Swaen and Brachet are alone in showing that the heart cells are definitely continuous with the intermediate cell mass of the the trunk mesoderm.

Many early workers on the fish embryo have claimed, as has been done for most vertebrae classes, that the heart fining arises from endoderm. The weight of evidence at the present time is so overwhelmingly against such a view that it warrants only a passing mention. Again, however, it must be realized that in the separation of the mesoderm from the endoderm it is possible that some future mesoderm cells may be left behind not cleanly separated. These cells might later isolate themselves from the endoderm to form vessels or blood. It nevertheless seems generally true that all blood forming cells are at one time in development contained within the mesodermal portion of the embryo. Gregory ('02) came to the conclusion that the endoderm and mesoderm could be traced to an indifferent cell mass mesentoderm in certain Teleosts, and according to his view, there is no way to say from which germ layer the heart endothelium actually arises. A mixture of endoderm and mesoderm cells gave rise to endocardium.

The later development of the heart of the bony fish proceeds much as in the case of other vertebrates, as has been carefully described in detail by Senior ('09). The only point of interest in the present discussion is the origin and significance of its endotheUal portions, and here Senior after a very thorough investigation confirms in all general points the previous findings of Swaen and Brachet.

In Fundulus as in other Teleosts the heart endothehum partially forms in loco but is also added to by wandering cells or ingrowths of mesenchymal cells adjacently located. The venous end of the heart leads directly down upon the yolk periblast, and as was shown in several figures, this periblastic material with huge amorphous nuclei may be at times drawn up into the cavity of the heart. This would indicate that the venous end was entirely free or not connected with any other vascular endothelium. This condition is, no doubt, due to the absence of the anterior yolk vessels which should in ordinary cases unite or fuse with the end of the heart so as to establish a closed circulation.

According to Swaen and Brachet in the region of the third somite the intermediate cell mass forms only the aorta, while caudad the aorta arises from the dorsal cells of the mass and the great part of the mass forms the red blood corpuscles and the venae cardinales. The endothelium of the cardinal veins finally surrounds the blood cells, but before these cells are fully developed or free, plasma has begun to flow in the aorta and other arteries.

In pelagic forms in which the egg is extremely small and develops very rapidly, the intermediate cell mass in the forward body sections is very small, sometimes only seen between the somites. This portion gives rise to the aorta. The cells are somewhat more numerous in the middle and posterior sections, but they never form a mass to the extent found in the larger demersal eggs. At the time of hatching the posterior cell strings form two lateral longitudinal vessels from the beginning of the mesonephros caudad to the anus. These two vessels, Swaen and Brachet consider to be homologous to the unpaired median stem vein of the trout and this is thought to represent the conjoined cardinal veins. We have noticed that in Fundulus the intermediate cell mass is sometimes divided forming two lateral cardinals loaded with, blood cells, while generally it exhibits a median unpaired condition. In the pelagic forms the vessels are all hollow at the time of hatching and the blood cells have not appeared.

Derjugin ('02) claims from a study of the pelagic egg of Lophius that the vessel cells of the aorta and cardinals are derived from the sclerotom. Fehx ('97) like Ziegler ('87) differs with Swaen and Brachet ('01, '04) in that he derives the aorta not from the "Venenstrang" but from the sclerotom which under the notochord forms a mesenchymal aortic string. Fehx states that no blood cells are to be seen in the aortic anlage, while the cardinal veins, of course, are loaded with the blood cells of the intermediate cell mass. Fehx, therefore, derives the two chief vessels of the embryo from two different parts of the mesoderm, the somites or sclerotoms and the lateral plates.

Sobotta ('02) terms the intermediate cell mass Blutstrange" and derives it from the lateral plate, though he had earher claimed it to arise from the somites. He described it in the trout embryo in the region from the eighth to the thirty-third somite. The 'Blutstrange' at first paired, are naked cellular strings without a true vessel covering. This they receive later as the cardinal vein anlagen. The endothelial cells of the cardinal veins he derives from the same source which produces the aorta, namely, the sclerotoms.

Finally, then, Swaen and Brachet derive the blood and vascular endotheUum of the aorta and venae cardinales from the intermediate cell mass which arises from the lateral plate. Felix derives only the blood and vascular endothehum of the cardinals from the intermediate cell mass which is separated originally from the lateral plate. The aortic endothelium arises from the sclerotoms. Sobotta considers the intermediate cell mass an exclusive blood forming material, while all vascular endothehum, including the heart, is derived from the sclerotoms which are budded off from the somite system. This disagreement, as we have pointed out before, is not of primary importance and may result merely from the fact of the intimate connection of the sclerotom and intermediate cell mass before their original separation.

The question now arises whether all the blood of the Teleost embryo is exclusively derived from the intermediate 'Blutstrange.' Felix admits that the endothelium of the glomerular vessels of the mesonephros arise in loco and at the same time blood cor-puscles often occur in this region. Sobotta claims that in the vascular network in the tail of the trout embryo some of the blood corpuscle anlage exists.

Both of these exotic positions of origin may be easily understood. In the first place, the nephric anlage is formed from cells in direct association with those constituting the early intermediate cell mass, and in the separation it probably happens that some future blood cells are held within the kidney anlage and these cells later develop in their proper fashion. The presence of blood corpuscles in the vascular network of the tail is due to the fact that the intermediate cell mass in many Teleosts, as Marcus ('05) has pointed out and as Senior ('09) particularly emphasized extends far back into the caudal region.

A similar consideration is the question of origin of vessels from material other than that of the intermediate cell mass and sclerotom. This is also important, and numerous observations would indicate that in the early bony fish embryo vessels unquestionably arise in loco and not solely as outgrowths or sprouts from a central vessel anlage. Sobotta ('02) on the contrary imagines a gradual growing away of the vascular system from its local origin, the sclerotom. The aorta is the primary vessel and, for example, the sub-intestinal vein arises from the aorta by vascular sprouts which grow around the gut, broaden out and fuse on its ventral side and finally give rise to the longitudinal vein. This theory of Sobotta is as unacceptable in the face of the great body of evidence to the contrary, Felix ('97), Ruckert ('88), Hahn ('09) and many others, as is the opposite ingrowth parablast theory of His ('75).

The consideration has been confined so far to the intra-embryonic blood vessels. We may now briefly discuss the development of vessels and blood upon the yolk. There are here two opposed or different views. The first derives the yolk vessels and blood directly from the yolk syncytium or periblast. The second derives blood cells exclusively from the intermediate cell mass in the embrj^o, but admits that cells may secondarily come to lie on the yolk by being pushed out from the intermediate cell mass with which, however, they maintam a definite continuity. The vascular endothelial cells are also derived from the embiyo as mesoblastic wandering cells, but these are not to be compared directly with blood cells since their parent cells have a separate place of origin. Most of the earlier workers thought that the blood in the Teleosts arose on the yolk-sac, as it does in other meroblastic embryos. The more recent workers have gone to the other extreme and deny the presence of blood islands upon the j^olk-sac as separate from the mtermediate cell mass.

As mentioned in describing the heart formation, nmiierous investigators have recorded wandering mesenchymal cells upon the yolk-sac, but from a study of the literature no clear conception can be formed as to the origin of blood cells or the vas. cular endothelial cells upon the yolk from these wandering cells. Some authors claim that the majority of wandering cells become pigment cells, while the remainder form the yolk vessels. In Fundulus the pigment cells very soon present a different appea,rance from the mesenchymal cells which are to form the vascular endotheUum, Both types of cells may be readily seen wandering over the yolk between the ectoderm and periblast. Before the yolk vessels are completely formed, the circulation of a cell free plasma has begun. The extent of the spaces in which this circulation takes place is verj^ variable. The arrangement of the veins of the yolk circulation is also extremely different in the different groups of Teleosts. One must agree with Hochstetter ('93) in stating that the yolk circulation in different forms is so different from the start that it is not possible clearly to summarize the condition in order to give satisfactory comparisons with the same vessels in Selachians, and Amphibians.

When the plasma is flowing in a closed system within the embryo, it is still running as a wandering stream through lacunae and sinuses on the yolk. This probably explains why the blood cells reproduce for so long a time on the yolk-sac while no such reproduction is taking place in the well formed vessels of the embryo.

It is difficult to determine the exact moment, when, or place at which the first blood cells get into the circulation. This probably varies even among embryos of the same species. Ziegler ('88) thinks, however, that just beyond the lateral plates in the plasma filled spaces of the yolk-sac which lie between the periblast and ectoderm, the first blood cells project into the circulation. They are in the form of cell strings which later connect the cardinal veins with the vascular yolk net. Swaen and Brachet saw in trout embryos of eleven days in the region of the fourteenth somite and posterior that the intermediate cell mass spreads out laterally below the lateral plate and on to the yolk surface. The cells thus came to lie above the yolk syncytium and first attained their red color in this position. These authors thus claim that in the bony fish with a large yolksac the haemoglobin free early blood cells through continued contact with the yolk become transformed into erythrocytes.

The experimental embryos considered in the present paper demonstrate, however, that it is not at all necessary in such a Teleost to have the erj^throblasts reach the yolk-sac in order to acquire their red haemoblobin condition. The tightly packed erj^throblasts within the intermediate cell mass of the embrx'o develop perfectly and readily attain a normal red haemoglobin color.

Finally, comparing the processes of vessel and blood formation in Teleosts with these processes in other vertebrate embryos, we find no definite explanation for the formation of the intermediate cell mass. In other embryos the blood is largely formed upon the yolk. However, it must be recognized from recent contributions that the formation of intra-embryonal blood is much more extensive and important than has formerly been supposed. The relation of the blood anlage to the cardinal vein and the position of the blood forming cells dorsal of the gut are unique in the Teleosts. The late formation of the yolk vessels and their type of origin from wandering mesenchymal cells is also of special interest.

It would seem as though the peripheral mesoblast which in other vertebrate types grows and develops outside the embryo, had in the Teleosts been peculiarly concentrated and drawn into the embryo during its phylogenetic history. Yet in this intraembryonic position the peripheral mesoblast gives rise to the same cells which it would ordinarily produce on the yolk-sac. The different Teleosts probably show this drawing in of the peripheral mesoderm to various degrees so that in some cases only part of the mesoderm is incorporated in the intermediate cell mass, while the remaining part may still be outspread upon the yolk and there differentiates extra-embryonically. The intermediate cell mass is connected caudally with the end bud, just as the peripheral mesoblast of the Selachians is with the blastopore lip. In its genesis the intermediate cell mass is split off from the lateral plate and localized along its median border.

Marcus ('05) in his study on Gobius capito advanced the opinion that the intermediate cell mass in this embryo is comparable to the peripheral blood forming mesoderm of other meroblastic eggs. In an embryo of eleven somites, the intermediate cell mass passes without a break caudad to the end bud and there connects with both the ectoderm and entoderm, just as the peripheral mesoderm would meet the other two germ layers at the blastopore lip. He attempted to show by diagrams the relationship between the intermediate cell mass in Teleosts and the blood forming mesoderm of Selachians.

As the homologue of the peripheral mesoderm the intermediate cell mass has the power to form vessels and blood cells. Most authors admit this power and only Sobotta ('02) denies the vessel forming property, while others claim that only the cardinal veins arise from the intermediate cell mass, still others, as Swaen and Brachet, would derive the endocardium and aorta also from this common source.

The important fact is that in the small pelagic embryos, where no blood formation takes place before hatching, the intermediate cell mass forms the aorta and the cardinal veins and is also derived from the lateral plate. The lateral plate thus contains cells capable of forming vascular endothelium, and this is the case in all vertebrates.

At an early time in evolution the extra-embryonic blood forming mesoderm has been included within the body of the Teleost embryo and lies over the gut as the intermediate cell mass representing the yolk vascular layer. Here it is important to note that the yolk-sac of the Teleosts contains no real mesodermal layer, only separate wandering mesenchymal cells are found between the ectoderm and periblast, and these wandering cells have migrated out from the embryo.

Finally, as Mollier ('06) states in his review of this subject, it is not a question of the formation of the intermediate cell mass in the individual bony fish, but the wider question of the behavior of the blood forming peripheral mesoderm in the bony fish. All of the results must be considered in this light in their application to other animal classes.

The intra-embryonic blood formation in the bony fish does not represent the primitive type for vertebrates as Sobotta ('02) claims, but this is, no doubt, a modified secondary condition accompanying the various other modified and special developmental processes which bony fish embryo so frequentlj^ presents.

Wilson ('91) states of the mesoderm of the Teleost that: "The ventral subvitelline mesoderm, having in this way losts it function in the Teleost, must be regarded as a rudimentary organ of the gastrula. It always remains very small, and does not form any special organ or set of organs in the embryo." The real fact is that the subvitelline mesoderm is misplaced, being within the embryo as the intermediate cell mass and here forms the blood of the individual and, therefore, the yolk-sac of the bony fish has no mesodermic layer.

3. Vascular endotheliuvi, and vascidar growth and development

Mollier ('06) concludes in his review regarding the origin of vessels as follows.

As to the genesis of embryonal vessels we may pass the judgment that the theory of the local origin of the vascular endothelium is valuable. The notion of His (75) and Vialleton ('92) that the vessel strands of the embryo grow in as sprouts from the extra-embryonal anlage (vascular anlage) is not nearly so probable as that the individual vessel cells arise in loco and thus form the vascular nets.

This statement agrees in every way with the contentions so fully presented by Huntington ('10, '14), McClure ('10, '12) and others, regarding the origin of lymph vessels. Lately it receives additional substantiation from the experimental results recorded by Miller and McWhorter ('14) on the origin of blood vessels in the chick embryo. Such a position is further strengthened by the still more recent experimental evidence, presented by Reagan ('15) which shows the origin in loco of vessels in isolated parts of chick embryos. All of these experiments confirm the earlier results of Hahn ('09) on the origin of vessels in the chick.

In the Teleost embryos studied during the present investigation there can be no doubt that the heart endothelium and aortae arise in loco within the embryo, and here there are no vessels, or even mesoderm, present on the yolk-sac in the anterior portion. Certain vessels do partially grow from the embryo out on to the yolk-sac and other smaller vessels arise in many separate regions of the yolk-sac as the products of wandering mesenchyme cells which become arranged to form the tubular vessels. All of these vessels after they have arisen may grow by budding or sprouting off new vessels or may increase in length by a forward growth so well described in living embryos by E. R. Clark ('09, '12) in his careful studies of this subject.

Felix ('97) describes the origin of the aorta as follows:

The 'mesenchymaortenstrang' arises from the two lines of sclerotoms after they are finally pinched away from the somites. No fusion of cell material occurs between this and the 'venenstrang,' the intermediate cell mass. This 'mesenchymaortenstrang' comes from that part of the somites that was immediately in contact with the intermediate cell mass portion of the primary seitenplatte. As the forward somites bud off sclerotoms, these also are added to the 'mesenchymaortenstrang.'

The median part of the 'strang' forms the aorta, 'aortenstrang/ the lateral the 'mesenchymgewebe' (mesenchymestrang) . The 'aorlenstrang' is at first solid and does not obtain a continuous lumen to begin with, hut here and there develops a space, and these spaces become confluent to form the tubes and build the paired aortae. Certain portions of the Strang remain solid much longer than others. The association of the paired aortae to form an unpaired single vessel soon follows. While the aorta is being so formed, one never finds blood cells within its lumen. Blood cells occur only in the 'venenstrang' and in certain vessels of the nephric glomeruli. Occasionally certain of the glomerular vessels contain blood corpuscles at a time when the blood circulation is not yet closed.

Felix ('97) cites the observation of P. Mayer ('94) on very young Selachian embryos in which the medulary tube was still open. It was found in such embryos that the aorta is segmental and derived from the somites and subsequently the longitudinal tube is formed by the fusion of these isolated points. Felix agrees with P. Meyer's observations from his study on the Teleost.

There has been great diversity of opinion regarding the germ layer from which the vascular endothelium and blood corpuscles arise. In the hterature it may be found that certain competent investigators have in each vertebrate class claimed the vascular endothelium and blood cells to be derived from the endoderm, while other workers of equal authority have found the vessels and corpuscles to arise from the mesoderm. The consistency of the disagreement which one finds in a review of this literature is most pecuHar. These disagreements have their foundation in the extreme difficulty of the problem on fixed material.

It is interesting to note that in no case has the same author derived the blood and vessel endothelium from different germ layers. Each author always takes the position that blood and vascular endothelium arise from either the mesoderm or the endoderm.

We have here much to do with wandering cells which become lost from their epitheUal layer, and it is impossible to state their •origin. This is left to the imagination of the individual investigator and further possibihties of error are open.

Wenckebach's ('86) observations of living embryos are most important in this connection. He noted .that not only the layers but that independent mesoblast cells with amoeboid processes wander out of the embryo and over the yolk. These wandering cells play a great part in the formation of the anlage of the heart endothelium and great vessels. In the Teleost embrj'o one may readily observe these wandering cells in the yolk-sac, and they doubtless give rise to the yolk vessels and blood islands as well as the pigment cells so abundantly present.

Ziegler ('87) has suggested that it may be that the blood anlage in phylogeny has been passed to the mesoderm from the endoderm, and for this reason the endodermal origin may sometimes occur in coenogenetic development. Goette ('90) also held that the endodermal origin of the blood was the more primitive one. This point of view overlooks the fact that in the invertebrates generally the blood and vessel walls are derived from the mesoderm.

In discussing the question of the place of origin of the vessels, Felix ('97) points out that Rlickert ('88) claimed in Selachians, that the aorta arose in loco. P. Mayer and Strahl ('95), have also stated that the great vessels are late in appearing and arise in loco in the embrj^o's body. Felix states that the glomerulus of the bird mesonephros originates in loco independently of the aorta. Further that the stammvene, venenplexus of the mesonephros, certain vessels of the glomerulus, and also the mesenteric artery along with the aorta in the Salmoniden arise in loco. Regarding the anlage of the heart and vena sub-intestinales, Felix is not certain but thinks that these likewise arise in loco. All of these observations are directly opposed to the theory of ingrowth of vessels from the yolk-sac, the parablast theory of His ('75) as well as the outgrowth of vessels in the sense advocated by Sobotta ('02).

Ziegler ('89) and Felix ('97) have both speculated considerably as to the relationship of the cavity of the circulatory system with the primary body cavity and the coelom. Ziegler pointed out that in the phylogenetic origin of the blood vascular system we have the following changes: The primitive condition is represented by the development of a space between the body wall, the ectoderm, and the gut wall, the endoderm, that is, the primary body cavity or protocoel. Embrj^ologically the blastocoel of the blastular or after gastrulation, the space between the invaginated endoderm and the ectoderm, the schizocoel, represents the primitive vascular space. The body cavity in rotifers, nematodes, brj'ozoa and arthropods is a primaiy body cavity and is filled with a fluid, the haemolymph. In the arthropods on the dorsal side of the body is the pulsating heart which sets the fluid in circulation and this fluid contains corpuscles similar to the white blood corpuscles of vertebrates.

In the arthropods the vessels and heart are often highly developed but all communicate with lacunae and spaces betv/een the gut wall and body wall. The heart is surrounded by a pericardial space (not truly coelomic) which is full of haemolymph, and as the heart pulsates this haemolymph is drawn in through ostia along its walls and then propelled out through the aorta and its arches to the vessels and spaces of the body. These body spaces, or the haemocoele, are thought by some to be a secondary or specialized cavity. Yet it is not coelomic and has no definite lining and resembles veiy closely the primary body cavity of the rotifers, nematodes, and other invertebrate forms which it most probably represents. In some of the higher Crustacea a secondary body cavity or coelomic space of limited extent is present enclosing the ophthalmic artery in Paelamonetes. The cavities surrounding the gonads are also coelomic, and since these are well developed species the coelomic space here probably represents a progressive rather than a regressive condition.

The second step in Ziegler's evolution of blood vessels is illustrated by the conditions in the molluscs. In these animals between the gut and body wall lacunae and interstitial spaces exist which occupy the position of the primary body cavity and these are filled with blood. Vessels lead into the lacunae and the cavities of these vessels as well as the cavity of the well formed heart are also considered to be part of the primary body cavity with which they are continuous. The pericardial cavity in the molluscs is true coelom and not a part of the primary body cavity and contains no blood. In almost all of the molluscs the pericardium is in conmaunication Tvdth the nephridia and the nephricduct usually leads from the pericardium to the outer bodywall. The pericardial cavity in contrast to the primary" body cavity is designated as secondary body cavity or true coelom.

The final step in the phylogeny of the blood vascular system is characterized by an important expansion of the secondary body cavity or coelom as is the case in the echinoderms, anneUds and vertebrates. As a result of the expansion of the secondarj^ body cavity, the primary cavity is reduced merely to a system of channels or vessels and small interstitial lacunae. In the vertebrates, therefore, according to Ziegler, the blood and lymph vascular system represents the persistent part of the primary body ca\'ity. Ziegler considers the blood vascular system and lymph vascular system to have had a common origin. The blood vessel endothelium is closely similar in all respects to the lymphatic endothehum. He thus agrees with Biitschli ('82) that in all metazoa the blood vascular system has its origin from the blastocoel.

FeUx ('97) holds that his studies on the Salmoniden will not fit into Ziegler's scheme. He claims that the origin of the stammvene in the cranial portion is the same as that of the primary mesenephros in the caudal region, and is also of the same origin as that of the primary nephric duct. Cells of the splanchnic as well as cells of the parietal layer of the mesoderm enter into the structural material of the stammvene. The cavity of the venenstrang is the same as the cavity between the lamellae of the secondary lateral plate, that is, true coelomic cavity. The three structures referred to are all portions of the same base, the lateral plate mesoderm, the primarj^ seitenplatte. Felix states, as there is httle doubt that the cavity of the primarjnephric duct is homologous with the secondarj^ body cavity, so there is Uttle doubt that the cavity of the venenstrang is also. The development of the aorta shows similar relations. It arises, according to Felix, from the sclerotomes which come from the somites and contains both the somatic and splanchnic layers of mesoderm. The origin of the aorta from the 'mesenchymaortenstrang' is from the same cell material as the mesodermal layers. The cavity of the myotom is secondary body cavity, coelom, and so also is the aortic cavity. Neither is in any way primary body cavity. The formation of the aortic cavity is a similar process to the canalization of the stammvene. Felix in this way arrives at a conclusion diametrically opposed to Ziegler.

These conclusions he recognizes are not facts but are based on facts obtained from a study of Teleosts which are a side branch of the vertebrate stem, but from which one may still generalize to some extent. Felix calls attention to the fact that in the selachians Zeigler ('92), and in the reptiles Strahl ('83, '85), and in the bu-ds KoUiker ('84) and Ziegler ('92), and in the mammals Kolliker ('84), all claim that the first vessel anlagen are found in the mesoderm and not between the mesoderm and endoderm. Only in the mesoderm the secondary body cavity arises by splitting, and since the sohd vascular anlagen are formed within the mesoderm their cavities should not be considered primary body cavity. The wTiter is entirely unable to agree with such an analysis of the origin of vessels, particularly yolk vessels, as well as of the primary and secondary body cavities for reasons given below.

Fehx ('97) now goes further and assumes that the lymph vessels arise in mesenchyme and their cavity is primary body cavity and their wall cells are modified connective tissue cells. This position is difficult to appreciate since it must be admitted that mesenchyme is a direct product of the mesoderm, and, according to FeHx, any definitely formed cavity arising between such cells would seem to be coelom. I question, however, whether any other morphologist would put the same interpretation on all the spaces cited by FeUx as being in the coelomic category. Felix states, for example, that the aorta arises from a mesenchymaortenstrang derived from the sclerotom. The sclerotom is more or less mesenchymal in nature and certainly contains many cells which will later give rise to types of connective tissue. If the aorta did arise from this group of cells its cavity is scarcely of an origin comparable to that of the coelom. Its endothehal wall is certainly much the same as that of the lymph vessels.

The cavities of the nephric duct, ovarian duct, kidney tubules and other tubules derived from the mesoderm are not usually considered to be parts of the coelomic cavity. The blood and lymph vessels do arise from the mesoderm but not in such a way that their cavity can be readily homologized with the coelomic space originating between the lamellae of the mesoderm. The vessels on the yolk-sac of the Teleosts are formed from disconnected wandering mesenchyme cells which are easily demonstrated. The cavity of these vessels surely cannot be interpreted to arise between mesenchyme cells some of which are derived from the somatic and some from the splanchnic mesodermal layers. The yolk vessels in Teleosts arise by arrangement of mesenchyme cells and so apparently do other vessels within the embryo. Thus these blood vessels are similar in origin to the lymphatics according to FeUx's notion of the mesenchymal origin of lymphatics. The numerous recent investigators of the origin of the lymphatics, athough to some extent divided into two schools, all treat the lymph vessels and blood vessels as being of the same general genetic type Sabin ('13) and Huntington C14).

Finally, the most damaging evidence against Felix's notion that the blood vascular spaces are derived from the coelom, and that these spaces are actually now comparable to the coelomic space is the following: Before a true coelom, such as that to which Felix refers in the vertebrates, has arisen in the animal series blood vessels are already present and these vessels often communicate with or are actually a part of the primary body cavity. When the true coelom does arise in the invertebrate series blood vessels never open into its cavity or conamunicate with it. Fehx has therefore derived an older and more generalized animal system from a newer or later formation. This of course is contrary to any principle of phylogenetic calculations.

The weight of evidence at the present time is then in favor of the earher notion of Ziegler. The blood vascular system if it is associated with, or phylogenetically derived from any other body cavity, that cavity is really the primary body cavity or embryologically the blastocoel.

4. Haematopoesis, the monophyletic and polyphyletic views, etc.

The experiments recorded above are of particular value in the solution of that very complex problem, the origin and relationship of the different types of blood corpuscles. We may here then briefly consider the evidence they furnish in connection with the various theories and points of view recently advanced in explanation of the origin of blood cells.

The vertebrate animals present two entirely different types of cells floating in their blood fluid. The white blood corpuscles are cells of primitive type and are not only found within the vessels but they also wander through the interstitial spaces of all the tissues of the body. These wandering white blood cells, amoebocytes, are almost universally distributed throughout the animal kingdom being found in all the invertebrate groups above the one or two very lowest as well as in all the vertebrate classes. In no animal do these cells contain haemoglobin, haemocyanin or any compound that would particularly qualify them as oxygen carriers, or give to them any function as an organ of respiration. These white blood cells found outside of the blood currents as well as in the blood are to be looked upon as cells which are not particularly associated with any specific blood function. They merely find the blood current a ready or rapid means of being carried from place to place within the body.

The red blood corpuscles, erythrocytes, are in contrast to the white cells a very highly specialized type of cell and specifically a blood cell. In fact, this is one of the most specialized cells within the body. In mammals, for example, it is specialized to such a degree that its functional perfection is actually accompanied by the loss of its nucleus and necessarily, therefore, the loss of its own future existence after a short period of time.

Contrasted with the ahnost universal distribution of the white cells within the animal kingdom the erythrocyte is confined to the vertebrates phylum and to certain particular cases among the invertebrates. The respiratory function of the invertebrate blood is often claimed to be confined to the fluid or plasma mass, and only among certain members of the higher groups is a cell developed with the function of carrying oxygen to the body tissues and even this cell can not be said to possess the regular typical characters of the vertebrate erythrocyte.

The vertebrate erythrocyte along with the typical vertebrate mouth, the pharyngeal gills, the dorsal nerve cord, the notochord, and bony skeleton and the many other possessions characterizing the vertebrate group, separates it in gulf-like fashion from the invertebrates. The white blood cells bridge this gulf but the red blood corpuscle differs from that of the invertebrate in a way comparable to the difference between the vertebrate mouth and that of the invertebrate, both serve the same function but are structurally unlike. Just as the mouth and pharyngeal gills and vertebral column have no invertebrate forerunner, so no cell within the invertebrate animals can at the present moment be sought out or designated as the certain ancestor of the red blood cell.

The cells of the vascular walls are closely similar in both vertebrates and invertebrates, as pointed out above. In both animal divisions they probably arise and develop in the same fashion. The white blood corpuscles probably do also. Yet the red cells, although they too originate from the mesenchyme in the vertebrates, are not in any way certainly descended from the invertebrate oxygen carrying cell or the wandering leucoblast or amoebocyte.

There is certainly no phylogenetic or comparative morphological evidence to warrant one in deriving vascular endothelium, leucocytes, and erythrocytes from a common cell ancestry except, of course, they are all derived from the mesenchyme or sam© germ-layer.

The fundamental histological study of the early developmental stages of the blood elements in vertebrates was contributed by Van der Stricht ('94). His studies were especially confined to the mammals. As has often been the case the conclusions reached from this pioneer study are largely correct in the light of recent investigations. Van der Stricht held that the first blood cells arising within the area vasculosa are entirely young red cells, erythroblasts. When one surveys the literature of this subject, it is found that all authors with three or four recent exceptions (Bryce ('05), Dantschakoff ('07) and Maximow '09), hold that the blood islands give rise exclusively to red, haemoglobin bearing corpuscles, erythroblasts or finally erythrocytes. This is true for the Fundulus embryos described in this paper and even though the cells are confined to their place of origin and never flow away, since there is no circulation, yet the groups always consistently contain only erythrocytes.

Van der Stricht holds that the leycoblasts and leucocytes are independent of the erythroblasts and arise extra-vascularly in the mesenchyme and later wander into the vessels.

Browning ('05) and Goodall ('07) have both recently claimed that the leucocytes have a different origin from the erythrocytes and arise at a later period. Goodall states:

When leucocyte proliferation in the liver has begun, the islands of erythroblasts and leucoblasts are definitely separate in position, and the distinctness of their identity is obvious, and no transitions between them can be seen. These facts argue strongly against the view that the erythroblasts are derived from the primitive leucoblasts.

Jolly and Acuna (^05) have pointed out that in early stages only red cells are found in the blood. The first lymphocytes occur very late and still later the granulocytes, so that the guinea-pig embryo has attained a, length of sixteen mm. before white blood corpuscles are present.

Again all authors with few exceptions seem entirely agreed that the leucoblasts arise much later than the erythroblast. All without exception also agree that the leucoblasts arise extravascularly while the erythroblasts arise partially within the sinuses, and that the island groups of erythroblasts soon become surrounded by vascular endothelium while no vessel walls have ever been described to form around the groups of leucoblasts. These facts are no doubt of much genetic importance.

The question involved is then: Which is correct, the monophyletic or polyphyletic theory of haematopoesis? It is recognized by all that both propositions are classed only as theory. It must further be recognized that both theories are based at the present time only upon the interpretations of various observers, these interpretations are not necessarily facts. I trust, therefore, that the experiments on Fundulus embryos may add a basis of unquestionable facts which may show the correctness of one or the other of these interpretations.

With this point of view, we may undertake a critical examination of the evidence so ably presented by Maximow ('09) in his study on the mammalian embryo. The observations he construes as strong argument in favor of the monophyletic origin of all types of blood cells and vascular endothelium. This contribution by Maximow ('09) has been accepted by many embryologists a bras ouverts, and has been largely incorporated into several recent chapters on the development of the blood, for example, by Schaefer ('12), and Minot ('12).

In the primitive streak stage of the rabbit embryo Maximow states that the peripheral mesoderm in which the blood islands will later occur has in no sense the character of a connected epitheUal layer, but consists merely of local accumulations of cells of mesenchymatous type. The cells of this mass are of long thin spindle shape or with star-like processes. These cells are probably much of the same type as the wandering cells seen on the yolk-sac of the Fundulus embryos. In this peripheral mesenchymatous mesoblast the first blood islands arise in the caudal part of the area opaca, as originally described by Van der Stricht. The blood islands are formed from the spindle or branched mesenchyme cells which become associated into groups.

Maximow states that the first endothelial cells like the primary blood cells are also derived from the mesoblast-mesenchyme. With this one may fully agree and several other tissues could be included in the statement as derived from mesenchyme. Maximow, however, goes further and thinks this general source a common specific source. Thus the endothehal cells and blood cells are closely related and arise from a common stem cell in the blood islands and may continue to arise from such a cell during later development.

Die ersten Endothelien und die ersten Blutzellen sind also beides Mesoblast- resp. Mesenchymzellen. In den Blutinseln sehen wir sie von unseren Augen aus einer geraeinsamen Quelle entstehen. Auch in der spateren Entwicklung werden wir oft Gelegenheit haben, die enge Verwandtschaft dieser beiden Arten von Mesenchymzellen zu beobachten.

This is merely a matter of interpretation and not at all a demonstrated fact. In reply to such a position we must call for an explanation of the demonstrated fact presented on previous pages showing that vascular endothelium forms in a perfectly normal fashion within the heart and head region of embryos without circulating blood, but in no case in early or late stages was the endotheUal hning of the aorta or other vessels capable of giving rise to any type of blood corpuscles. Yet the power to form blood corpuscles was abundantly present in the same embryos as shown by the huge numbers of blood cells within the blood forming regions, the intermediate cell mass and yolk islands. Why do not the mesenchyme cells within the liver and all vascular endothelium form blood when no circulating blood reaches them? (If ever, there should then be the stimulus to give rise to its formation).

The red blood cell anlage is a definite mesenchyme cell or group of cells and only members of this cell group possess the blood forming power. To cite a parallel case, the Uver cells are derived from the common endodermal cell stock yet not all early endodermal cells, in fact only a few, have the power to develop into a liver, or a pancreas or a lung as the case may be. The embryological argument is indeed rather loose that on account of the fact of vascular endotheUum and blood cells arising from mesenchyme would assume, therefore, that these very different cells had a common stem mother cell and later actually possessed some powers of transmutability.

Maximow advances the interpretation that the first blood cells in the area vasculosa are not all erythroblast or future red blood corpuscles. These cells he designates as 'primitive blood cells' since they may form either white or red corpuscles. Yet in the yolk islands of the Fundulus embryos without circulation only red blood cells, erythrocytes, are produced and they remain in this location to be observed throughout embryonic life. The evidence for Maximow's position seems to me somewhat insufficient.

During the summer of 1914, I had the privilege of examining Mme. Dantschakoff 's preparations which both she and Maximow cite in support of the monophyletic theory of blood cell origin. One so inclined might interpret these specimens as showing that the red and white corpuscles do arise from the common stem mother cell. The youngest lymphocytes were invariably scattered among the mesenchymal cells while the erythroblasts were budded off into more or less well defined vessels. No one could emphatically state that the two classes of blood corpuscles had ever actually divided off from any one single mother cell.

The more or less constant separation of the early leucoblasts and erythroblasts, as is also shown in Maximow's figures and those of other workers, would seem to indicate their origins from two different mother cells. If one mother cell only forms or divides off cells which develop into lymphocytes or leucocytes and another mother cell gives rise to only erythroblasts, then there is no reason to say that the two mother cells were the same although they appeared to be two similar mesenchymal cells. They were potentially different, and this potential difference is all that the diphyletic notion of blood cell origin demands. A careful study of the embryos without a blood circulation will demonstrate the fact of this different origin of white and red corpuscles.

Maximow then advocates the last clause in the monophyletic code, and states that the intravascular primitive blood cells are not only increased by mitosis but are also added to by the production of the same kind of cells from the fixed endothelial wall of the primitive vessels. Endothelial cells may wander away into the mesenchyme or may wander into the vessel lumen. One often sees according to Maximow a cell project into the vessel, its body assumes a rounded form and its protoplasm changes into that of an erythroblast. It must be distinctly remembered that these appearances are in dead stained specimens and man}^ possibiUties exist which might explain their occurrence.

Granting that such a phenomenon actually appears to occur there is one very probable explanation without assuming that true vascular endothelium may form blood corpuscles. Let it be supposed, for example, that in the formation of the vascular wall around the yolk-sac blood islands that some of the peripheral cells of the island might lag behind in their difTerentiation retaining their more or less mesenchymal type. Such a cell may come to be closely pressed against the vascular wall and really appear as though it were one of the vessel wall cells. This might readily happen, and probably does happen, and may account for the occasional appearance of 'vessel wall cell' forming a blood cell.

Why do not the endothelial cells in the experimental embryos possess the power to form blood cells when the vessel is totally empty of blood cells? Even though it is clearly shown that other cells of the embryo do possess the normal blood building power. These specimens are exactly such as should supply definite proof of blood cells arising from endothelium, but the evidence they furnish really disproves the proposition.

Schridde ('07, '08) according to Maximow has gone so far as to claim that in young human embryos endothelium can directly form primitive erythroblasts. Maximow does not agree with this since in his specimens the endothelium gives rise only to indifferent colorless cells. Shridde's claim is based upon misinterpretation and so, I believe, is any claim that blood cells arise from formed vascular endothelium.

Most authors find that in the very early embryonic blood there are no white corpuscles but only red cells present. Bryce ('05) however, describes in Lepidosiren the very early origin of leucoblasts from primitive blood cells, and later Dantschakoff and Maximow find lymphocytes not only in the vascular net of the area vasculosa but also, though at first very few, in the circulating blood. Maximow thinks that when these early lymphocytes are not seen it is due to poor technique or defective material.

Maximow believes the red blood cells may finally arise from lymphoblasts as erythroblasts, then erythrocytes. This mode of development of the definite erythroblasts continues throughout life and is accomplished in the same manner in all erythropoetic organs. Wherever such indifferent mesenchyme cells, lymphoblasts, are found this locahty is eo ipso a new place of origin of erythroblasts out of these colorless stem cells. If this be actually true, why then do not red cells, erythroblasts, finally form all through the body of non-circulating fish embryos, since the wandering lymphocytes surely have the power to reach many places other than the normal sites of erythroblasts formation, the intermediate cell mass and yolk islands?

Maximow claims that both types of blood cells red and white arise at one and the same period from one and the same source, the primitive blood cells in the area vasculosa. The experiments on Teleosts do not bear out such a position since the original or first blood cells from the intermediate cell mass are all erythroblasts and show a characteristic type at a very early time. The blood origin on the area vasculosa is not so extensive, but here also first form only erythroblasts.

Supporters of the polyphyletic origin of blood cells have been able to make equally strong observations in favor of their view on similar material to that studied by Maximow, Dantschakoff and other advocates of the monophyletic theory.

Maximow suggests that since the "primitive blood cell" has no haemoglobin it really stands nearer to the leucocyte than to the erythrocyte, and one might say that the leucocyte arises first in development and the haemoglobin cell later. This is most decidedly not the case in the Teleosts where the primitive mesenchymal blood cell passes directly into the erythroblast without ever showing a stage suggesting either lymphoblast or leucocyte.

With Weidenreich ('05), Maximow takes the position which he had earlier manitained that all non-granular leucocytes and also the wandering cells of the tissues constitute one great cell group. The position determines the direction of their development and only in certain places does one find all types being formed, for example, in the embryonic liver and the adult bone marrow.

In reply to the extreme monophyletic position it may be asked : Why are only erythrocytes present in the old blood islands on the yolk of non-circulating specimens? Why is no cellular blood element present in the aorta and other endothehal lined vessels in the anterior region of similar embryos? Why are wandering primitive blood cells" unable to form blood in the liver and other positions while blood forming power is present to a vigorous extent in certain regions of the same embryo but from a definite anlage? Wandering cells in the Teleost embryos have to do with yolk-sac blood origin, but these wandering cells are the equivalent of a part of the peripheral mesoderm and always wander out on to comparable regions of the yolk and never wander to other places within the embryo.

Maximow probably criticizes correctly the many artificial distinctions between various leucocytes which are pointed out by some advocates of the polyphyletic theory. My experiments do not bear on this point up to the present time.

Finally Maximow does not imply that a granular leucocyte or red corpuscle can change into anything else, they cannot undifferentiate. He states that before there is any development of granulation or haemoglobin in two different cells, there is really a difference in the cells though we are unable to distinguish it. This invisible difference determines the destiny of the cell to form either a leucocyte or erythrocyte. This is certainly true, but we cannot stop just at this point; these differences must be carried a step further or really back to their actual beginning. Then it is found that although two wandering mesenchyme cells on the yolk-sac of the fish embryo are indistinguishable so far as our powers of observation go, yet they are fundamentally different since one is destined to form only an erythroblast while the other possesses no such power and can form only an endothelial lining cell or pigment cell as the case may be.

This is all the diphyletic or polyphletic school would ask. That is, that certain definite mesenchymal cells are actually the red blood cell anlage and only from these particular mesenchymal cells do red blood cells arise. Here we logically stop, for this is what is conceived by embryologists to be an anlage, back of this we go to certain germ layers and still further back to the developmental potentialities of certain individual cells as followed in studies of cell-lineage and finally we reach the elementary proposition of deriving everything from the original egg cell. But to stop with the tissue anlage we find strong evidence to indicate that certain special mesenchymal cells are designated to form erythroblasts, others leucoblasts, and still others, and these are much more universally scattered throughout the embryo, give rise to vascular endothelium. These latter may really be admitted to form endothehum largely as a response to physical conditions.

Summary and Conclusions

The present contribution attempts an experimental analysis of the origin and development of blood cells and the endothehal lining cells of the vascular sj^stem. Studies on the origin of blood and endothelium in the normal embryo are rendered peculiarly difficult on account of the important role that wandering mesenchyme cells play in this process as well as the perplexing mixture of cells of different origin brought about by the early established circulation. The origin of no other tissue is so confused by mechanical and physical conditions.

The first difficulty has been met by a study of living Fundulus heteroclitus embryos with the high power binocular microscope. The wandering mesenchyme cells may in this way be followed to a great extent. The disadvantages due to the intermixture of cells in the blood current have been overcome by the investigation of embryos in which a circulation of blood is prevented from taking place.

When Fundulus eggs are treated during early developmental stages with weak solutions of alcohol the resulting embryos in many cases never establish a blood circulation. In other respects these embryos may be very nearly normal and the development and differentiation of their tissues and organs often proceeds in the usual manner though at a somewhat slower rate. The heart and chief vessels are formed and the blood cells arise and develop in a vigorous fashion. The heart pulsates rythmically but is unable to propel the body fluid since its venous end does not connect with the yolk vessels and in many cases its lumen is partially or completely oblitered by periblastic material and nuclei which seem to be sucked into the heart cavity from the surface of the yolk.

In these embryos without a circulation of the blood one is enabled to study the complete development of the different types of blood corpuscles in the particular regions in which they originate. There is no contamination of the products of a given region through the introduction of foreign cells normally carried in the blood current.

The actual haematopoetic value of the different organs and tissues may be determined in the experimental embryos, and clearly distinguished from the ordinary reproduction or multiplication of blood cells which in the normal embryo would reach these organs through the circulation.

The debated question as to the production of blood cells from vascular endotheUal cells may be conclusively answered, at least for the species here studied.

The results and conclusions derived from these experiments may be summarized as follows:

1. The Teleost embryo is capable of living and developing in an almost normal fashion without a circulation of its blood. Red blood cells may be seen to arise and differentiate in these living embryos in two definite localities, one within the posterior body region, and the other the blood islands on the yolk-sac.

The blood cells remain confined to their places of origin, yet they attain a typical red color and may persist in an apparently functional condition on the yolk-sac for as long as sixteen or twenty days. The normal embryo becomes free swimming at from twelve to fifteen days, but these individuals without a circulation never hatch although they may often live for more than thirty days.

All recent investigators have claimed that there are no blood islands present on the Teleostean yolk-sac. Yet the presence of such islands is readily demonstrated in living Fundulus embryos, in normal specimens as well as in those with, no circulation.

2. The plasma or fluid in the embryos failing to develop a circulation begins to collect at an early time in the body cavities. The pericardium becomes hugely distended with fluid as well as the lateral coelomic spaces and the Kupffer's vesicle at the posterior end of the embryo. The great distension of the pericardium due to this fluid accumulation pushes the head end of the embryo unusually far away from the surface of the j^olk. The heart is thus stretched into a long straight tube or string leading from the ventral surface of the head through the great pericardial ca^^ity to the anterior yolk surface (compare figures 15 to 20).

No blood vessels form on the extreme anterior portion of the yolk-sac, so that the venous end of the heart is never connected 'ttdth veins, and does not draw fluid into its cavity to be pumped away through the aorta. When the heart cavity does contain fluid it is unable to escape and small floating particles may often be observed rising and falUng with the feeble pulsations of the heart.

3. The hearts in embryos wdthout a circulation are lined by a definite endocardium, but the myocardium is poorly developed, sometimes consisting of only a single cell layer. Chi'omatophores are not present in the wall of the normal heart but in the experimental hearts these large cells ladened with pigment granules are invariably found. The cavity in many of the hearts is almost if not entirely obUterated by the presence of periblastic material and large amorphous periblast nuclei.

The conus end of such hearts leads directly to a more or less closed ventral aorta, portions of the aortic arches are seen in the sections as open spaces, and dorsal aortae are almost invaribly seen as typical spaces hned by characteristic embryonic endothehum.

A point of much importance is the fact that neither these hearts with their endothelial linings nor any portion of the aortae at any stage of development have ever been seen to contain an erythroblast or an erythrocyte. Cells of this type are completely absent from the anterior region of the embryo.

4. Pigment cells normally occur on the Fundulus yolk-sac and arrange themselves along the vascular net so as to map out the yolk-sac circulation in a striking manner. Loeb has thought that this arrangement along the vessel walls was due to the presence of oxygen carried by the corpuscles within the vessels. In the embryos without a yolk-sac circulation the pigment cells arise but rarely become fully expanded so that the usual long branched processes are represented only by short projections, the chromatophore consequently seems much smaller than usual.

The unexpanded pigment cells, however, wander over the yolk-sac and collect in numbers around the plasma filled spaces. The yolk surface of the pericardium and the periphery of the Kupffer's vesicle are often almost covered with pigment. The hearts are during early stages full of plasma and the pigment cells form a sheath around them, while pigment cells are never present on the normal hearts during the embryonic period.

These facts would seem to indicate that the plasma rather than the erythrocytes contain the substance which attracts the chromatophores and initiates their arrangement along the normal vascular net of the yolk-sac.

5. A definite mass of cells characteristic of the Teleost embryo is located in the posterior half of the body between the notochord and the gut and extends well into the tail region. This socalled 'intermediate cell mass' is the intra-embryonic red blood cell anlage in many of the species.

The peripheral cells of the mass as claimed by Swaen and Brachet or the mesenchyme about the mass, Sobotta, forms a vascular endotheUum which encloses the central early erythroblasts.

In individuals without a circulation the erythroblasts arise in a normal manner in this centrally located position, and become erythrocytes filled with haemoglobin. Typical vascular endothehum completely surrounds the erythrocytes which instead of being swept away as usual by the circulating current remain in their place of origin. All of the early blood forming cells of this intermediate mass give rise only to erythroblasts.

6. Contrary to the opinion of most recent observers on blood development in Teleosts, the Fundulus embryos both with and without a circulation possess blood islands on the posterior and ventral portions of the yolk-sac. These blood islands are formed by wandering mesenchymal cells which migrate out from the posterior region of the embryo. They represent all that remains of the peripheral yolk-sac mesoderm in the Teleosts and probably wander way from mesoderm related to that of the intermediate cell mass. The intermediate cell mass may possibly represent the bulk of the peripheral mesoderm which is here included within the embryonic body, while in other meroblastic eggs it is spread out over the yolk. The only mesodermal portion of the yolksac in Fundulus is made up of the disconnected wandering mesenchyme cells some of which group themselves to form the blood islands, while others give rise to the yolk vessel endothelium, and still other wandering cells develop into the chromatophores.

7. The non-circulating red-blood corpuscles within the embryo remain in a fully developed condition for eight or ten days and then undergo degeneration. In an old embryo of sixteen days it is sometimes found that very few of the corpuscles in the intermediate mass are still present and these are degenerate. The vascular endothehum has been lost and numerous mesenchymal cells have wandered in and lie among the corpuscles.

On the yolk-sac the corpuscles no doubt have a better oxygen supply and here they maintain their color longer but finally also present a degenerate appearance with small densely staining nuclei and cell bodies much reduced in size.

8. Vascular endothelium arises in loco in many parts of the embryonic body in which blood cell anlagen are not present. This endothelium is in all cases utterly incapable of giving rise to any type of blood cell. This incapacity cannot be attributed to the abnormal condition of the embryo as true blood cell anlagen in the same specimen produce blood corpuscles in abundance.

Vascular endothelium in the fish embryo has no haematopoetic function.

9. Neither lymphocytes nor leucocytes have been found to arise in the yolk-sac blood islands nor within the intermediate cell mass.

The embryonic v/hite blood cells are most abundant in the anterior body and head regions, and these cells occupy extravascular positions usually lying among the mesenchymal cells.

The sources of origin of the white and red blood corpuscles in Fundulus einhryos are distinct, and these two different types of cells cannot he considered to have a monophyletic origin except in so far as both arise froin mesenchymal cells.

The adult blood of Fundulus contains lymphocytes and several varieties of granular leucocytes.

10. There is evidence to indicate that definite environmental conditions are necessary for blood cell proliferation or multiplication. Blood cells do not normally divide when completely enclosed by vascular endothelium. This is the key to the shifting series of so-called haematopoetic organs found during embryonic development.

Erythroblasts lying about spaces unenclosed by vascular endothelium proliferate steadily and give off their products into the spaces from which they find their way into the embryonic vessels. Should such an erythroblast be carried by the circulation to another unlined space it may become arrested there and again undergo a series of divisions giving rise to other erythroblasts. When, however, these spaces become lined by endothelium the blood cell reproduction stops.

It most embryos the earliest blood cell formation occurs in the yolk-sac blood islands. The cells in these islands continue to divide until they become surrounded by endothelium, then the yolk-sac blood islands lose their haematopoetic function and become a vascular net through which the blood circulates. The liver now takes up the role of harboring dividing blood cells within its tissue spaces, when these spaces become vascularized by endothelium, here again the blood cells no longer multiply but merely circulate.

Finally, in the mammalian embryo, one organ after another ceases to offer the necessary harbor for dividing blood cells until the red bone marrow is the only tissue presenting the proper relationship of spaces and vessels, and here alone the erythropoetic function exists to supply the red blood cells for the entire body circulation. The red blood corpuscles are always produced so as to be delivered into the vessels and thus very soon occupy an intra-vascular position, while the white blood cells arise and remain for some time among the mesenchymal tissue cells in an extra-vascular position.

11. Lymphocytes and leucocytes along with the invertebrate amoebocytes are all generaUzed more or less primitive wandering cells, and are almost universally distributed throughout the metazoa.

Erythrocytes are very highly speciaUzed cells with a peculiar oxygen carrying function due to their haemoglobin content. In contrast to the universal distribution of the leucocytes the erythrocytes are only found in the vertebrate phylum, except for a few cases existing in some of the higher invertebrate groups. Yet even in these particular cases the oxygen carrying blood cell never presents the typically uniform appearance of the vertebrate erythrocyte. The oxygen carrying function in invertebrates is usually confined to the liquid plasma.

Typical vascular endothehum is widely distributed in the animal kingdom and appears to be formed from a simple slightly modified mesenchymal cell.

These three very different types of cells all seem to arise from mesoderm — the mesenchyme. Yet the present investigation would indicate that each arises from a distinctly different mesenchymal anlage.

The erythrocyte anlage is localized and perfectly consistent in the quality of its production.

The lymphocyte and leucocyte anlage is more diffusely arranged and not definitely locahzed in any particular cell group.

The vascular endothelium appears to be formed in loco in almost all parts of the embryonic body, and its formation is absolutely independent of a circulating fluid or the presence of blood cells.

The facts presented seem to indicate that vascular endothelium, erythrocytes and leucocytes although all arise fro?n 7nesenchyme are really polyphyletic in origin: that is, each has a different mesenchymal anlage. To make the meaning absolutely clear, I consider the origin of the liver and pancreas cells a parallel case both arise from endoderm but each is formed by a distinctly different endodermal anlage, and if one of these two anlagen is destroyed, the other is powerless to replace its product.


Browning, C. H. 1905 Observations on the development of the granular leucocytes in the hvunan fetus. Jour. Path, and Bacter., vol. 10.

Bryce, T. H. 1905 The histology of the blood of the larva of Lepidosiren paradoxa. Part II. Haematogenesis. Transact. Roy. Soc. Edinburgh, vol. 41, Part II, no. 19.

BtJTSCHLi, O. 1882 tJber eine Hypothese beziiglich der Phylogenetischen Herleitung des Blutgefiissapparates eines Theil der Metazoon. Morphol. Jahrbuch, Bd. 8.

Clark, E. R. 1909 Observations on living growing lymphatics in the tail of the frog larva. Anat. Rec, vol. 3.

1912 Further observations on living growing lymphatics: their relation to the mesenchyme cells. Am. Jour. Anat., vol. 13.

Clark, E. R.,andE. L. 1912 Observations on the development of the earliest lymphatics in the region of the posterior lymph heart in living chick embryos. Anat. Rec, vol. 6.

Dantschakoff, W. 1907 "Qber das erste Auftreten der Blutelemente im Hiihnerembryo. Vorl. Mitt. Folia haematologica, 4. Jahrg., Suppl. no. 2.

1908 Untersuchungen iiber die Entwicklung des Blutes und Bindegewebes bei den Vogeln. I. Die erste Entstehung der Blutzellen beim Hiihnerembryo, etc. Anatomische Hefte, Bd. 37.

Derjugin, K. 1902 Ueber einige Stadien in der En*wickelung von Lophius piscatorius. Travaux de la Societe Imp. des Naturalistes de St. Petersburg, tom. 33.

Felix, W 1897 Beitrage zur Entwickelungsgeschichte der Salmoniden. Anat. Hefte, Bd. 8.

GoETTE, A. 1874 Beitrage zur Entwickelungsgeschichte der Wirbeltiere. II. Die Bildung der Keimblatter und des Blutes im Hiihnerei. Arch, f. mikr. Anat., Bd. 10.

1890 Abhandlungen zur Entwickelungsgeschichte der Tiere. Hefg 5. Entwickelungsgeschichte des Flubneunauges. Hamburg und Leipzig.

GooDALL, A. 1907 Haematogenesis in fetal sheep. Jour. Path, and Bacter., vol. 12, nos. 2and3.

Gregory, E. H. 1902 Beitrage zur Entwickelungsgeschichte der Knochenfische. Anat. Hefte, Bd. 20.

IIahn, H. 1909 Experimentelle Studien liber die Entstehung des Blutes und der ersten Gefasse beim Hlihnchen. Archiv fiir Entwicklungsmechanik der Organismen, Bd. 27.

Hennegut, L. F. 1888 Recherches sur le developpement des poisons osseux. Journ. de I'Anat. et de la Physiol., T. 24.

His, W. 1875 Der Keimwall des Hlihnereies und die Entstehung der parablastischen Zeller. Zeitschr. f. Anat. u. Entwickelungsgesch. 1900 Lecithoblast und Angiolbast der Wirbeltiere. Abhandl. d. math, -naturw. Kl. d. K. sachs. Ges. d. Wiss., Bd. 26.

HocHSTETTER, F. 1893 Entwickelung des Venensystems der Wirtelbiere. Ergebnisse der Anatomie und Entwickelungsgeschichte von Merkel u. Bonnet, Bd. 3.

1905 Ueber die Entwickelung der Dottersackcirkulation bei Scyllium stellare. Arch. f. mikr. Anat., Bd. 66.

Huntington, G. S. 1910 The phylogenetic relations of the lymphatic and blood vascular system in vertebrates, and the genetic principles of the development of the systemic Ij^mphatic vessels in the mammalian embryo, p. 339. Anat. Rec, vol. 4.

1914 The development of the mammalian jugular lymphsac, of the tributary primitive ulnar lymphatic and of the thoracic ducts from the view point of recent investigations of vertebrate lymphatic ontogeny together with a consideration of the genetic relations of lymphatic and haemal vascular channels in the embryos of Amniotes. Am. Jour. Anat., vol. 16.

Jolly, J. 1905 Sur revolution des globules roughes dans le sang des embryons des Mammiferes. Compt. rend. Soc. Biol., T. 57.

Jolly and Acuna 1905 Les leucocytes du sang chez les embryons des Mammiferes. Archives d'Anat. Microscopique, T. 7.

Knower, H. McE. 1907 Effects of early removal of the heart and arrest of the circulation on the development of frog embryos. Anat. Rec, 1.

KoLLiKER, A. 1884 Die embryonalen Keimblatter und die Gewebe. Zeitschr. fur wiss. Zoologie, Bd. 34.

VON Ku'PFFER, C. 1868 Beobachtungen liber die Entwickelung der Knochen fische. Arch. f. mikr. Anat., Bd. 4. LoEB, J. 1912 Heredity in heterogeneous hybrids. Jour. Morph., vol. 23.

1893 Ueber die Entwicklung von Fischembryonen ohne Kreislauf.

Pflliger's Archiv., Bd. 54. Marcus, H. 1905 Ein Beitrag zur Kenntniss der Blutbildung bei Knochen fischen. Arch. f. mikr. Anat., Bd. 66.

Maximow, a. 1909 Untersuchungen iiber Blut und Bindegewebe. Arch. f. mikr. Anat., Bd. 73.

Mayer, P. 1887 Mitteil, der zoolog. Station Neapel, torn. 7.

1894 Uber die esrten Stadien der Gefasse bei Selachiern. Anat. Anz., Bd. 9.

McClure, C. F. W. 1910 The extra-intimal theory and the development of the mesenteric lymphatics in the domestic cat (Felis domestica). Verhandl. der Anat. Gesellsch.

1912 A few remarks relative to Mr. Kampmeier's paper on the value of the injection method in the study of lymphatic development. Anat. Rec, vol. 6.

Miller, A. M., and McWhorter, J. E. 1914 Experiments on the development of blood vessels in the area pellucida and embryonic body of the chick. Anat. Rec, vol. 8.

MiNOT, C. S. 1911 Die Entwicklung des Blutes. Handbuch der Entwicklungsgeschichte des Menschen. Keibel und Mall, Bd. 2, Leipzig.

MoLLiER, S. 1906 Die erste Entwickelung des Herzens, der Gefasse und des Blutes der Teleostier. Handbuch der Vergleich. u. Experimen. Entwick. der Wirbeltiere. Jena.

Oellacher, J. 1872 Beitrage zur Entwickelungsgeschichte der Knochenfische, I. Zeitschr. fur wissenschaftl. Zoologie, Bd. 22.

1873 Beitrage zur Entwickelung der Knochenfische II Zeitschr. f. wiss. Zoologie, Bd. 23.

Rabl, C. 1887 Uber ^ die Bildung des Heriiens bei Amphibien. Morphol. Jahrbuch 13. 1890 Theorie des mesoderms. Morphol. Jahr., Bd. 15.

Raffaele, F. 1892 Ricerche sullo sviluppo del sistema vascolare nei Selacei. Mitth. a.d. Zoolog. Station z. Neapel, Bd. 10.

Reagan, F. P. 1915 Vascularization phenomena in fragments of embryonic bodies completely isolated from yolk-sac blastoderm. Anat. Rec, vol. 9.

RtJCKERT, J. 1887 Ueber die Anlage des mittleren Keimblattes die erste Blutbildung bei Torpedo. Anat. Hefte. Bd. 2.

1888 Ueber die Entstehung der endothelialen Anlagen des Herzens und der ersten Gefassstamme bei Selachierembryonen. Biol. Centralbl Bd. 8, nos. 13, 14.

1903 Ueber die Abstammung der bluthaltigen Gefabanlagen beim Huhn und fiber die Entstehung des Randsinus beim Huhn und bei Torpedo. Sitzungsber. der K. bayr. Akad. d. Wiss., Bd. 32.

Sabin, F. R. 1913 The origin and development of the lymphatic system. Johns Hopkins Hospital Reports. Monographs, new series, no. 5.

Schaefer, E. a. 1912 Textbook of microscoi)ic anatomy. Longmans, Green and Co., New York.

Schmidt, M. B. 1892 tjber Blutzellenbildung in Leber und Milz, etc. Zeiglers Beitriige, Bd. 11.

ScHRiDDE, 1906 tJber Myeloblasten und LjTnphoblasten. Verhandig. der Kongresses f. innere Medizin, Bd. 23, Vers., Mtinchen. 1907-1908 Die Entstehung der ersten embryonalen Blutzellen des Menschen. Verhandl. d. deutsch. path. Gesellsch., 11. Tag. Dresden, 16-19.

ScHULTE, H. VON W. 1914 Early stages of vascologenesis in the cat (Felis domestica) with especial reference to the mesenchymal origin of endothelium. Mem. Wistar Inst, of Anat., no. 3.

Senior, H. D. 1909 The development of the heart in shad (Alosa sapadissima) Am. Jour. Anat., vol. 9.

SoBOTTA, J. 1894 Ueber IMesoderm-, Herz-, Gefab- und Blutbildung bei Salmoniden. Verh. d. Anat. Ges. auf d. 8. Vers, zu Strassburg. 1902 Ueber die Entwickelung des Blutes, des Herzens und der grossen Gefassstamme der Salmoniden nebst Mitteilungen iiber die Austildung der Herzform. Anat. Heft 63 (Bd. 19, Heft 3.)

Stockard, C. R. 1907 The influence of external factors, chemical and physical, on the development of Fundulus heteroclitus. Jour. Exp. Zool., vol. 4. 1915 An experimental study of the origin of blood and vascular endothelium in the Teleost embryo. Proc. Amer. Ass'n Anat., Anat. Rec, vol. 9.

Strahl, H. 1883 Ueber die Anlage des Gefasssystems in der Keimscheibe von Lacerta agilus. Marb. Sitzungsber. 1885 Der Parablast der Eidechse. Marb. Sifzungaber. 1888 Die Dottersackwand und der Parablast der Eidechse. Zeitschr. f. wiss. Zool., Bd. 45.

Sumner, F. B. 1900 Kupffer's vesicle and its relation to gastrulation and concrescence. Mem. X. Y. Acad. Sci., vol. 11.

SwAEN, A., and Bracket, A. 1899 and 1901 Etude sur les premieres phases du developpement des organs devises du mesoblaste chez les poissons Teleosteens. Arch, de Biol., T. 16 and 17.

1904 Etude sur la formation des feuillets et des organs dans le bourgeon terminal et dans la queue des embryons des poissons teleosteens. Arch, de Biol., T. 20.

Van der Stricht, O. 1892 Nouvelles recherches sur la genese des globules rouges et des globules blancs du sang. Arch, de Biol. T. 12. 1893 Sur I'existence d'ilots cellulaires a la peripherie du blastoderme de poulet. Anat. Anz. T. 8.

1895 De la premiere origine du sang et des capillaires dans I'aire vasculaire du lapin. Comptes rend, de la Soc. de biol. S. 10. T. 2. "1896 Origine des globules sanguins, de I'aorte et de I'endocarde chez les embrvons de Selachiens. Ibidem.

Van der Stricht. O. 1899 L'origine des premieres cellules sanguines et des premiers vaisseaux sanguins dans I'aire vasculaire, de chauves-souris. Bull, de I'Acad. Roy. de med. Belgique. S. 4. T. 13.

ViALLETON, L. 1892 Sur l'origine des germes vasculaires dans Tembryon de poulet. Anat. Anz., Bd. 7.

Weidenreich, F. 1904r-05 Die roten Blutkorperchen II. Ergebnisse der Anatomie und Entwicklunggeschichte von Merkel u. Bonnet, Bd. 14. 1905 tJber die Enstehung der weissen Blutkorperchen im postfetalen Leben. Verb. d. Anat. Gessellch., Bd. 19. Vers., Genf.

Wenckebach, K. F. 1885 The development of the blood corpuscles in the embryo of Perca fluviat. Jour. Anat. and Physiol., vol. 19.

1886 Beitrage zur Entwickelungsgeschichte der Knochenfische. Arch. f. mik. Anat., Bd. 28.

Wilson, H. V. 1891 The embryology of the sea-bass (Seranus atrarius). Bull. United States Fish Comm., vol. 9.

Zeigler, H. E. 1882 Die Embryonalentwickelung von Salmo salar. Diss, inaug. Freiburg i. B.

1887 Die Entstehung des Blutes dei Knochenfischenbryonem. Arch, f. mik. Anat., 30.

1888 Der Urspring der mesenchymatischen Gewebe bei Selachiern. Arch. f. mikr. Anatomie, Bd. 32.

1889 Die Entstehung des Blutes bei Wirbeltieren. Berichte d. naturforsch Gesselschaft Freiburg i. B. Bd. 4.

1892 Uber die embryonale Anlage des Blutes dei Wirbeltieren. Verhandl. der deutschen Zoolog. Gessellschaft.

Ziegler, H. E., and Ziegler, E. 1892 Beitrage zur Entwickelungsgeschichte von Torpedo. Arch. f. mik. Anat., Bd. 39.