Book - Contributions to Embryology Carnegie Institution No.18

From Embryology
Embryology - 22 Aug 2019    Facebook link Pinterest link Twitter link  Expand to Translate  
Google Translate - select your language from the list shown below (this will open a new external page)

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

Sabin FR. Origin and development of the primitive vessels of the chick and of the pig. (1917) Contrib. Embryol., Carnegie Inst. Wash. 6: 61–124.

Online Editor  
Mark Hill.jpg
This 1917 paper by Florence Rena Sabin (1871 - 1953) describes early vascular development in 2 well characterised animal models, chicken and pig. Florence Sabin was a key historic researcher in early 1900's establishing our early understanding of both vascular and lymphatic development.



Cardiovascular Links: cardiovascular | Heart Tutorial | Lecture - Early Vascular | Lecture - Heart | Movies | 2016 Cardiac Review | heart | coronary circulation | heart valve | heart rate | Circulation | blood | blood vessel | blood vessel histology | heart histology | Lymphatic | ductus venosus | spleen | Stage 22 | cardiovascular abnormalities | OMIM | 2012 ECHO Meeting | Category:Cardiovascular
Historic Embryology - Cardiovascular 
1902 Vena cava inferior | 1905 Brain Blood Vessels | 1909 Cervical Veins | 1909 Dorsal aorta and umbilical veins | 1912 Heart | 1912 Human Heart | 1914 Earliest Blood-Vessels | 1915 Congenital Cardiac Disease | 1915 Dura Venous Sinuses | 1916 Blood cell origin | 1916 Pars Membranacea Septi | 1919 Lower Limb Arteries | 1921 Human Brain Vascular | 1921 Spleen | 1922 Aortic-Arch System | 1922 Pig Forelimb Arteries | 1922 Chicken Pulmonary | 1923 Head Subcutaneous Plexus | 1923 Ductus Venosus | 1925 Venous Development | 1927 Stage 11 Heart | 1928 Heart Blood Flow | 1935 Aorta | 1935 Venous valves | 1938 Pars Membranacea Septi | 1938 Foramen Ovale | 1939 Atrio-Ventricular Valves | 1940 Vena cava inferior | 1940 Early Hematopoiesis | 1941 Blood Formation | 1942 Truncus and Conus Partitioning | Ziegler Heart Models | 1951 Heart Movie | 1954 Week 9 Heart | 1957 Cranial venous system | 1959 Brain Arterial Anastomoses | Historic Embryology Papers | 2012 ECHO Meeting | 2016 Cardiac Review | Historic Disclaimer
Immune Links: immune | blood | spleen | thymus | Lymphatic | lymph node | Antibody | Med Lecture - Lymphatic Structure | Med Practical | Immune Movies | vaccination | bacterial infection | Abnormalities | Category:Immune
Historic Embryology  
1909 Lymph glands | 1912 Development of the Lymphatic System | 1918 Gray's Lymphatic Images | 1916 Pig Lymphatics | 1919 Chicken Lymphatic | 1921 Spleen | 1922 Pig Stomach Lymphatics | 1932 Cat Pharyngeal Tonsil | Historic Disclaimer

chicken Pig Development

Historic Disclaimer - information about historic embryology pages 
Mark Hill.jpg
Pages where the terms "Historic Textbook" and "Historic Embryology" appear on this site, and sections within pages where this disclaimer appears, indicate that the content and scientific understanding are specific to the time of publication. This means that while some scientific descriptions are still accurate, the terminology and interpretation of the developmental mechanisms reflect the understanding at the time of original publication and those of the preceding periods, these terms and interpretations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)

Origin and Development of the Primitive Vessels of the Chick and of the Pig

Florence Rena Sabin (1871 - 1953)
Florence Rena Sabin (1871-1953)

Contributions to Embryology

By Florence R. Sabin


With seven photos and eight figures in the text.

Related Links: Cardiovascular System Development | Chicken Development | Pig Development

Contents

  • Introduction
  • Methods
  • Vascular system of the chick
  • General account of the va.scular .system of the chick up to stage of 14 somites
  • Origin of the cardinal veins
  • Vascular system of the brain and the primary head-vein
  • Development of the spinal arteries
  • The vascular system in young pig embryos
  • The form of the heart
  • Ventral branches of the aorta, including the allantoic arteries and the sub intestinal artery
  • The umbilical vessels
  • Neural branches of the aorta and the primary head-vein
  • The cardinal veins in the pig
  • Nephritic vessels in the pig

Introduction

In this paper is given an account of the primitive vessels of the chick and of the pig, as made out by injecting Uving embryos, and, in the case of the chick, as seen growing in the embryo. Such studies must nccessarilj- be accompanied by the study of sections. In the case of the marmnalian embryo I have made injections in earlier stages than had been done heretofore and in the case of the chick I have carried the method of injection to the earliest stage in which it is possible. Below the stage at which they can be injected, the vessels of the chick can be studied in the living blastoderm by a technique which has developed out of the method of tissue-culture introduced by Harrison. The chick thus offers unusually valuable material for the study of vascular problems, as it is possible to use both the method of injection and that of direct observation of the living embryo in the same stage.


In the course of this study two fundamental ideas have been under consideration. The first concerns the most essential question in connection with the vascular system, namely, the relation of differentiation and growth of endothehum. According to one theory there is a limited period for the differentiation of angioblasts out of undifferentiated mesenchyme, and after this period aU new blood-vessels arise from the growth or proliferation of older angioblasts. This theory seems to me to have the weight of evidence. The second theory is that angioblasts continue to differentiate out of mesenchj-me indefinitely If the former theory is correct and the period of differentiation of endothelium is a limited one, the fundamental problem concerning the early blood-vessels is to determine which differentiate and which are formed from preceding vessels. In practicalh' every embryo chick observed, up to a certain stage new angioblasts can be seen differentiating and joining the older angioblasts, but the phenomenon becomes less and less frequent as older stages are studied. In the living embryo the aorta itself can be seen to differentiate out of mesenchyme, and at the stage when the heart begins to beat every chick shows a few isolated angioblasts along the mesial border of the aorta, which will be seen to join the aorta if the specimen be watched for a short time. I have some evidence also that some of the iirimitive vessels along the neural tube differentiate out of mesenchyme, the process being observed in the living embryo. On the other hand, one can watch the growih of the entire wall of a vessel by cell-division in the living embryo and the formation of new vessels from the walls of old vessels; so that the study of the early blood-vessels is gradually becoming a more exact problem, namely, the determination for each vessel, whether it differentiates in situ or develops from preceding vessels. My present material is not adequate for the solution of this question, but throws some hght upon it.


The second question— which has proved of great interest — is the definition of the terms artery, vein, and capillary as thej- are used for the embryo. In the study of the vessels of the embryo particular stress should be laid on the time when circulation begins. That there is a very extensi\-e d(n'elo])ment of blood-vessels before there is any circulation of the blood due to the beat of the heart was well known to the earlier embryologists; for exani])le, to von Baer and later to His. Moreover, the heart beats for a considerable time before it starts any circulation. It is known that the blood-vessels spread over the body in definite and constant sheets of capillaries, and in these primitive vessels, after the circulation has begun, a vessel may serve as an artery for a time and then be reduced to a capillary plexus, in which the direction of the circulation is entirely different from that of the circulation of the original artery. 8uch a vessel, for example, is the subintestinal artery of the pig, which arises in a capillary plexus around the caudal end of the primitive gut and carries blood out to the arteries of the yolk-sac, where it must again pass through a capillary bed before it returns to the heart. This artery becomes broken into a capillary plexus in the wall of the gut, which makes new connections with branches of the omphalo-mesenteric veins within the wall of the mesentery, so that its blood, instead of flowing away from the embr3'o to the membranes, flows within the embryo toward the heart.


Again, a vessel may serve for a time as a vein in the return of blood to the heart and may subsequently receive new arterial connections and become an arterial plexus, with the direction of the flow of blood entirely changed. Such a vessel is the so-called vena capitis medialis. This is a primitive vessel along the hindbrain, which in the chick in the second and third days of incubation serves as a vein for the forebrain and midbrain, but as an arterial capillary trunk for the hindbrain; that is, it carries mixed blood and is the only vessel of the hindbrain, representing its entire capillary bed. Early in the fourth day it receives new arterial connections, a new vein develops to carry the venous blood for the forebrain and midbrain, and the primitive vascular channel of the hindbrain breaks into a capillary ])lexus in which the direction of the current of blood is at right angles to the direction of the original current. From these two examples it must be clear that in the study of the primitive vascular system it is very imjwrtant to understand the function of the vessels at each stage of development, and any presentation of the vascular system which overlooks this point and is dominated wholly by the pattern of the ves.sels of the adult becomes difficult to follow and may be misleading. In the cjuestion of nomenclature a decision has to lie made between two theories — that is, whether the vessels arc to be named according to the function they i)erform at any given stage or whether they are to be named according to the vessels for which they form the primordia. If the latter method is chosen it must be remembered that a given vessel of an embryo often disappears entirely in giving rise to new vessels — for example, the primitive vessel of the hindbrain.


In this study I shall use terms as consistently as possible, in the following manner: By the term artery, in reference to an organ, I mean a vessel which brings blood to that organ but does not form any part of its capillary bed, and I have colored such vessels red. By the term vein 1 mean a vessel which carries blood from an organ to the heart, provided it does not serve as the capillary bed for that organ or break up into another capillary plexus. Such vessels I have colored blue. All other vessels I have indicated in gray, and shall try to trace the complicated changes which such vessels undergo, serving at times as arteries, veins, or capillaries, or as vessels with a double function. This method of nomenclature is therefore ba.sed on the function of the vessel at the stage when it exists. It takes into account the very shifting course of the blood, as the vessels develop in the embryo better than does the method of making a too early identification of the adult vessels in those of the embryo. This usage of terms serves to restrict especially the term vein as applied to the embryo. The meaning of the terms artery and vein, as applied to the embryo, will become more clear in the discussion of individual vessels.

Methods

As has been stated, in this stud}' I have followed two methods: first, that of the injection of embryos, and second, the method of studying the living blastoderm in the case of the chick in a hanging-drop preparation. A general account of the methods of injecting embryos will be found in my paper on the ayzgos veins pubhshed as "Contribution to Embryologj', Xo. 7," by the Carnegie Institution of Washington in 1913. All of the injections of young embryos are made by blowing ink into the vessels through a very fine canula.


To inject the young chick the shell is opened and the embryo exposed to a strong light under a binocular micro,scope. A few drops of warm Locke's solution are placed on the blastoderm, and the vitelline membrane is removed. By the time the chick has 14 somites the sinus terminalis, or marginahs, is well developed in the edge of the area vasculosa and can easily be punctured with a fine canula. Nevertheless it is not easy to obtain complete injections of the blood-vessels of the chick through the veins until the embryo has about 16 somites. In stages between 9 and 16 somites more complete injections can be made by puncturing the aorta directly. This is a very interesting point in connection with the time of the beginning of the circulation. I shall show that, though the heart commences to beat about the time the tenth somite is forming, the circulation does not begin until about the stage of 16 somites. From the time the circulation begins it is easy to get complete injections by blowing a little ink into the ^'iteUine veins and allo^-ing the heart to pump the ink through the vessels of the embryo. If total specimens are desired it is well to dilute the ink one-half, so that the superficial vessels will not become so dense as to obscure the deep ones. I shall discuss in this paper the effects of injection in the embryo before the circulation has begun.


The earhest cliick embryo which I have injected was one of 9 somites; and I beheve this stage to be about the youngest to which the method is appUcable. At the stage of 6 somites the dorsal aorta is in the stage of a plexus of angioblasts, many of which are still sohd cells. This plexus of cells gradually acquires a lumen and becomes the aorta during the stages of from 6 to 9 somites. All stages up to 18 or 19 somites can be studied to great advantage in hanging-drop preparations.


Direct injections into the aorta can be made in the following manner: When the embryo is placed in a strong light the myotomes are of course very plainly visible, and along their lateral border can be seen a faint opaque streak, which is the intermediate cell-mass or nephrotome. Between the nephrotome and the lateral border of the myotomes is a thinner line which is more transparent. The canula is then introduced between the lateral border of the myotomes and the intermediate cell-mass, with the point toward the head of the embryo. If the canula enters the aorta, and only a slight pressure is used, there need be no extravasation at the point of puncture. After the embryo has been injected it is fixed bj' dropping Bouin's mixture on the specimen while it is still on the yolk. It is kept flooded with the fixing agent, and is not removed from the yolk until it is well hardened.


In regard to the injection of young mammalian embryos there are a few special iwints in technique which are of interest. In order to identify the young embryo pig the mucosa is examined carefully for long strings of chorion, which are so inconspicuous that they are more readily found by running the finger over the mucosa than by sight. These strings of chorion are then very carefully coiled on a glass sUde or piece of filter paper until the embryo is found. In the case of the pig, I have never succeeded in juncturing the veins on the yolk-sac or the umbilical vein. The latter is so large that it might be punctured if it contained enough corpuscles to render it visible. The aorta, on the other hand, is readily punctured opposite the mid-body region. Here in early stages the two aorta3 seem slightly dilated, or later are fused into a single vessel. The needle is introduced ventral to the myotomes. The injection mass in every case was india ink. Silver nitrate seems to damage the tissue much more markedly in very young embryos than in later stages. After injection the pigs were fixed in Carnoy's mixture of absolute alcohol 6 parts, chloroform 3 parts, and glacial acetic acid 1 part. They were then placed in 80 per cent alcohol, dehydrated in graded alcohols, and cleared by the Spalteholz method of benzine followed by oil of wintergreen.


The study of the blastoderm of the Uving chick embryo in a hanging-drop preparation depends on the methods originated by Harrison and developed by a large group of workers — Burrows, Carrel, M. R. Lewis and W. H. Lewis, and others. In 1912 AlcWhorter and ^^'hipple ajiplied the method to the study of the growing blastoderm of the chick, which was its first application to the entire embryo, so far as I am aware. These investigators mounted the blastoderm in clotted plasma and used the method to test the question as to whether blood-vessels arise from fusion of isolated vesicles. In 1913 Brachet pubUshed a study on the growth of a mammalian embryo in a hanging-drop preparation.


In studying the blastoderm of the chick by the method of the hanging-drop I have followed the technique of Margaret Reed Lewis, of growing the embryo in Locke's solution. In this way the embryo can ])e kei)t growing for several hours, and the cells which are nearest the cover-slip can be seen with great clearness and followed with an oil-immersion lens. The embryo is removed from the yolk and placed in a dish of warm Locke's solution and the vitelline membrane and most of the yolk removed. It is more difficult to study the blastoderm wath the dorsal surface against the cover-slip than from the ventral aspect, because the ectoderm does not adhere to the glass as well as does the endoderm, and it is necessary to have the embryo very flat in order to see the cells with higher powers. In regard to the blood-vessels, it is of course preferable to study the embryo from the ventral aspect, since the blood-vessels are nearer the endoderm than the ectoderm. In the study of the developing blood-islands it would be very advantageous to have the dorsal surface of the embryo against the cover-slip, as the blood-islands are farther dorsal and because the cells of the ectoderm over the area opaqua are much thinner than the cells of the endoderm over the same area. The cells of the endoderm of the area opaqua are so thick and so filled with globules of yolk that one can seldom focus through them in the Uving embrj^o. From the ventral aspect there is also an area of the axis of the embrjo that it is very difficult to study with high-power lenses, namely, the portion of the axis just caudal to the head-fold, because the heart lifts the embryo from the cover-slip and the cells can then be studied only with dry lenses.


AU of the chicks have been fixed in Bouin's mixture of saturated aqueous picric acid 75 parts, formalin (40 per cent) 20 parts, and glacial acetic acid 5 parts, for about 12 hours. Thej' are then placed directly in 60 per cent alcohol. In the case of the chicks which have been growing on the cover-slip it is very necessary to have the embryo stick to the cover-slip throughout the fixation and dehydration. If the specimens are to be mounted in toto, the}- are mounted on the same cover-shp on which they were growing. If they are to be embedded and cut they can be removed from the cover-slip after they have been cleared in the oil of wantergreen. The specimens do not become as brittle in the oil as in xylol. In the blastoderms which are kept on the cover-slips it is possible to watch the effects of dehydration much more accurately than with free specimens. The edge of the specimen around the outer margin of the area opacjua clings very closely to the cover-sUp; in fact, in mounting the embryo strands of tissue are pulled out which dry sUghtly and help the specimen to remain fixed. If the specimens are put into alcohol weaker than 60 per cent this outer margin wiU stick to the cover-slip, but the entire area pellucida will become free from the cover-slip and swell into a bleb. The space beneath fills in with fluid, and in the subsequent dehydration there is an uneven shrinkage which distorts the tissues. Thus, weak alcohol or water, or a dilute stain, macerates and swells the tissues and the subsequent shrinkage distorts the cells. On the other hand, if the specimens are placed directly in alcohol as strong as 60 per cent, a plexus of cells, which has been studied in the living specimen, can be readily identified in the fixed specimens. If the specimens are to be stained in toto thej' must be placed in a stain sufficiently strong so that the tissues will not macerate before they react to the stain. From such an experience one should avoid washing embryos in water and should also avoid the use of the lower grades of alcohol. I have used several changes of 60 and 65 per cent alcohol to eliminate the excess of picric acid. The dehydration can be done by changes of 5 per cent. If it is carried out too rapidlj' the specimens will crack, but the shrinkage with the higher grades of alcohol i.s b>- no means as marked as in the case of oltler embryos. The early embryos are injured much more by maceration due to weak alcohol than by shrinkage due to too rapid dehydration in the stronger grades. The specimens are all cleared by the Spalteholz method of benzine followed by oil of wintergreen. Specimens can be embedded from oil of wintergreen if they are passed directly from the oil to a mixture of the oil and paraffin. The tissue does not become too brittle to cut even after remaining in the oil for a year or two.


Vascular System of the Chick

General Account Of The Vascular System Of The Chick Up To Stage Of 14 Somites

In the study of the origin of the vessels of the chick I shall begin the detailed accoimt with the stage of 6 somites. The study of the bla.stoderm in a hanging-drop preparation offers a valuable method for a study of the early stages. In the stages of the early somites there is a i)le.\us in the area oi)a(iua which, by the older embryologists, Pander, von Baer, Remak, and later by His, was identified as the forerunner of the blood-vessels. Basing his studies on those of von Baer and Remak, His gave a description of the origin of blood-vessels which remains the foundation of our knowledge upon this subject (1868, pages 95 to 100). He described the first appearance of blood-vessels, or, as he later termed them, angioblasts, as occurring just before the appearance of the somites. He stated that the vessels began as a plexus of angular or spindle-shaped solid cells in the area opaqua. These cells from the beginning were in the form of a plexus (Cileich von Anfang an ein geschlossenes Mosaik, page 98). The plexus was at first made up of solid cells without a lumen, and grew by processes of solid spindle-shaped cells, exactly similar to those which formed the original network. This plexus was in a definite layer — the vascular layer (das Gefassblatt) of Pander.


The vascular layer, His said, consisted not only of solid angular cells, l)ut also of elements having a yellow color, or, in other words, it gave rise not only to bloodvessels but also to blood-cells. He regarded it as of the greatest importance that the first appearance of vessels was in the area vasculosa before the heart formed, and that these vessels arose entirely indeixMidently of any circulation. He then noted that the plexus of solid cells became^ transformed into vessels, the exact method of the transformation l)eing ini])ossible to determine; but that as the solid cells became the walls of vessels, their cytoplasm became less granular and their nuclei flatter. He then described the ajiproach of (he blood-vessels toward the axis of the embryo by means of the same type of st)li(l processes which formed the original plexus and found that they approached the axis in two zones: first, opposite the myotomes, and secondly, along the splanchnopleure in the region opposite the future emphalo-mesenteric veins that is, over the ventral surface of the two amnio-cardiac vesicles.


His noted that over the region of these anuiio-cardiac vesicles there was a double sheet of vessels which approached the axis of t he embryo, a more scanty sheet in the somatopleure, and a more abundant sheet in the splanchnopleure. Because of the development of the head-fold and the heart, it was impossible that the approach of the vascular plexus should be uniform over the entire length of the embryo. For example, the most cephalic part of the head became cut off from direct lateral connection with the embryonic membranes, and the vessels which approach the heart gradually rotated from the direct transverse direction to an oblique angle. He then noted that the dorsal aorta developed in the mesial edge of the plexus of angioblasts of the area vasculosa along the line of the lateral edge of the myotomes. From these observations he concluded that the vessels of the embryo are derived from the vessels of the membranes, and that the portion of the axis which can not be seen to receive the plexus of primitive angioblasts from the membranes receives its vessels by a growth of the plexus which has already invaded the embryo at other places.


In following the differentiation of the vascular area by improved methods wherebj' one can watch the living cells growing under an oil-immersion lens, it is astonishing how accurate is this description of His, which must have been made by far cruder methods. To his description must be added that with finer methods it is seen not only that the plexus out of which the aorta develops is the border of the common plexus of the entire area vasculosa, but that new cells differentiate along the axis of the embryo as well, so that angioblasts differentiate over the entire zone from the outer edge of the area ojxuiua to the margin of the future aorta along the lateral border of myotomes. Thus His's description must be extended to include a differentiation of new angioblasts in the axis of the embryo itself.


In the Uving blastoderm over the area opaqua, the endoderm-cells are so thick and so filled with yolk that the development of the blood-vessels and the blood-cells beneath them can be followed only with great difficulty. In the area pellucida, on the other hand, the endoderm is thin, and during the periods when the endoderm cells are not dividing they are so clear that it is easy to focus through them.


At the stage of 6 somites the head-fold is well formed and the amnio-cardiac vesicles have met in the mid-ventral line. Along the axis of the embryo there is a zone of dense tissue radiating from the primitive streak and from the embryo cephahc to the streak. This denser mass of tissue divides the area pellucida into an mner thicker zone containing the axis of the embryo and an outer thinner zone. In sections it is clearly seen that tliis denser zone is due to the further development of the ccelom nearer the embryo. Over the cephahc part of the denser zone the coelom has a wide lumen, and both its ventral mesoderm and the endoderm are thicker than the same membranes farther lateralward. This is very plainly shown in Duval's Atlas, plate xiv, figure 218, in the zone extending outward from his letter b. Farther caudalward in this dense band, opposite the undifferentiated myotomes and the primitive streak, there is no cavity of the ccelom, and its doisal and ventral mesoderm are fused and form a dense mass of cells. This entire thicker zone is difficult to study in the living chick, but the whole outer margin of the area pellucida is clear and the cells are so thin that one can readily focus through them from the endoderm to the ectoderm, and see every cell of the entire zone. Tliis is true, however, only when the endoderm is not dividing. The endoderm cells divide as a whole, and during the entire phase of cell division they are so opaque that it is impossible to focus through them. The phase requires about an hour, and to study the vessels beneath it is necessary to await until the cells become clear again.


In the entire outer margin of the area pellucida, at the stage of 6 somites, there are two plexuses. The dorsal plexus, which is the developing coelom of this area, appears to be composed of very large and flat vessels. Distinctly ventral to this plexus of the coelom is another plexus, much less abundant and made up of sohd bands of cells, which are angioblasts. An exceedingly important point, which can be determined with great distinctness in the li\dng specimen, is that the plexus of angioblasts connects by many tiny filaments with the plexus of the mesoderm of the coelom, but never connects by filaments with the endoderm. In sections the angioblasts of the vascular layer often touch the endoderm, but in the Uving embryo they are always separate. The living specimens also bring out very sharply the fact that the entire layer of angioblasts is distinctly ventral to the plexus of the mesoderm; in other words, the term vascular layer of Pander is an appropriate one, for the filaments of the angioblasts can be seen to dip down from the vascular layer to the mesoderm beneath. In the flat living specimen, and in sections which have been made from a specimen which was growing out flat on a cover-slip, there is no intermingling of the mesoderm and the vascular layer, such as is seen in Duval's plate xvi, figure 264. Such an apparent intermingling of the two layers is due to shrinkage. In other words, the angioblasts differentiate out of the mesoderm and form a new layer, which is throughout ventral to the mesoderm. These two plexuses were well known to His, who recognized them in their relations to the coelom on the one hand and to the angioblasts on the other in his work published in 18G8, but described them more fully in the Lecithoblasl und Angiohlad published in 1900. His stated that the two plexuses were at times very hard to analyze.


The plexus of angioblasts is, then, distinguished first by its more ventral position, and secondly by the fact that the cytoplasm of the angioblasts is slightly more granular and reacts slightly more intensely to basic dyes than does the mesoderm. The following criterion, however, is the one which I ha\e found most useful. In the living specimens there seems to be a sort of rhythm in cell division. I have already referred to the fact that the entire endoderm may divide and become so opaque that none of the cells beneath can be seen. At other times the entire plexus of angioblasts over a very extensive zone will ]>ass into the phase of cell division. In this condition the cytojilasm of the plexus of angioblasts becomes very highly refractiU" and opa(iue, so that it can l)e distinguished from the i)lexus of the coelom with great ease, even with low ]iowers of the microscope. The protoplasm shows this change for about an hour before the chromosomes pass on to the spindle; so, in order to obtain the nuclear figures characteristic of cell division, one must watch until a few areas in the plexus begin to clear and then fix the specimen. The time required for the nuclear changes is much less than the time taken for the cytoplasmic changes. According to M. R. Lewis, the nuclear ]ihase lasts about 5 minutes, while the cytoplasmic change takes about an hour. The facts that not every nucleus divides at the same moment and that the cytoplasmic changes have not been recognized explain the failure to note the rhythm of cell division.


Using the criteria for distinguishing angioblasts which have just been indicated, I will now describe what has been made out concerning the vascular system at the stage of G somites, both in the living specimen and in sections which have been made from a blastoderm in which the cells had been charted m the total specimen before the sections were cut. For this description the axis of the embryo may be divided into four zones: (1) that part of the head which is covered by the head-fold, as seen from the ventral aspect; (2) the head between the headfold and the first myotome; (3) the zone of the myotomes; (4) the zone caudal to the myotomes. As has been described, there is a dense band of tissue on either side of the axis of the embryo which divides the area pellucida into an inner dense zone and an outer thinner zone. The area opaqua, on the other hand, is denser along its outer margin. Beginning with the area opaqua, in its outer margin there is a large marginal plexup of vessels partly filled with blood-cells which cUng in large masses to the dorsal wall of the vessels. The blood-cells can be distinguished from the angioblasts by the fact that in the edges of the masses they tend to separate from the mass and have a definitely round contour. Angioblasts never have a round contour. In this marginal zone the ccelom is clearly seen, with its dorsal and ventral mesoderm, and the ventral wall of the bloodvessels is very plainly distinguished from the endoderm; but the dorsal wall of the blood-vessels is closely attached to the ventral mesoderm, and in places can not be distinguished from it.


The inner margin of the area opaqua and the outer margin of the area pellucida have two definite plexuses: the dorsal plexus of the coelom and the scantier ventral plexus of solid angioblasts. Over the dense area on either side of the myotomes the coelom is no longer in the form of a plexus, but has a complete lumen; for there the body-cavity is well formed. The plexus of angioblasts covering this area is continuous with a plexus of angioblasts along the lateral margin of the myotomes. Caudal to the sixth mj'otome, the plexus extends for a short distance along the undifferentiated mesoderm, curving a little to the side, very interesting appearances are to be made out near the first myotome. Extending forward from the lateral border of the first myotome, the chain of angioblasts representing the aorta can be seen up to the margin of the head-fold, when it disappears under the fold. Opposite the first myotome, and extending forward from its mesial border, there is also a chain of angioblasts along the hindbrain, and this chain of angioblasts connects with the aorta above the first and between the first and second myotomes. The chain of cells along the margm of the hindbrain I should not recognize as angioblasts in sections; but in the living blastoderm the}' have exacth^ the appearance of the angioblasts of the aorta and connect with them by slender filaments.


In till' region of tlie head, which can not l)e analyzed in tlie living blastoderm, the angioblasts representing the heart are well known and easily identified. The two cardiac primordia have met in the mid-ventral line and can be followed a short distance into a ventral cephalic aorta, which gradually becomes too indefinite for recognition. The dorsal cephahc aorta is very clear opposite the region of the heart, gradually disappearing farther forward. Thus, within the embryo there are chains of angioblasts representing the heart, most of the ventral cephalic aorta, and a part of the dorsal cephalic aortir. ()))])osite the region of the heart the two dorsal aortae are definite, tiny vessels whi(!h emerge from under the head-fold and are continued partly as a plexus of solid angioblasts and partly as a vessel along the ventro-lateral border of the myotomes. The entire plexus which is exposed on the ventral aspect connects with the plexus of angioblasts of the area pellucida. In this account I wish to emphasize the very early appearance of angioblasts along the hindbrain — the forerunner of the so-called vena capitis medialis, which I prefer to call the primitive vessel of the hindbrain. I have not yet a sufficient number of observations to ])rove, first, whether the transitory vessel of the hindbrain does differentiate in situ while the aorta is differentiating, and secondly, whether it is established earlier than the vessels of the forebrain; but both of these propositions seem to me to be very probable.


In this study I have found Williams's (1910) very careful description of the vascular system of the early chick embryo of great value. His specimen of 6 somites is clearly a httle farther advanced than mine. He found that at 6 somites the aortae were established, but were still small and irregular. He then observed a vessel along the neural tube (hindbrain) connected with the aorta in the first and second interspaces, the vessel in the first interspace being nearly as large as the aorta itself.


It is now important to consider how the plexus of angioblasts increases. This occurs first by cell division and secondh^ by the differentiation of new angioblasts. Cell division in the plexus of angioblasts is very extensive, for in watching the living specimens it is seen that large areas of the plexus divide at the same time, and in these cj'cles of cell division every cell of the plexus divides. Besides this very extensive cell division new angioblasts differentiate and join the plexus. This process can best be observed along the mesial border of the dorsal aorta itself, near the lowest myotome. Here practically every blastoderm between to 10 somites will show one or two isolated angioblasts which are very readily marked from the dense mesoderm V)eneath. Out in the zone of the develojjing ccrloni the distinction is by no means so easy. These angioblasts are either single, spindle-slia])ed cells or clumi)s of two or three cells. When observed they are seen to put out tiny filaments toward the wall of the aorta, which at once responds by putting out a filament toward the young cells. These tinj* filaments meet halfway, and the new angioblasts thus join the wall of the aorta. They gradually ap])roach the wall and b(>come incorporated into the vessel. As the new cells become a part of the wall their protoplasm Ix'comes less granular and they actiuirc a lumen. The exact process by which angiol)lasts acfjuire a lumen is extremely difficult to determine, and concerning this \w\n\ nothing has yet been added to the original description of His.


These observations on tlie origin of the aorta, as well as the observations indicating that the transitory vessel of the hindbrain differentiates from angioblasts in situ, at once lead to the general question of the origin of the vascular system All are agreed— on the foundation of the work of \oii Baer, Remak, and His— that certain cells of the embryo differentiate to form angioblasts or vasoformative cells in the early stages of embryonic life, and that these angioblasts increase by cell division. There has, however, })een a wide divergence of ojjinion as to whether the differentiation of new angioblasts continues throughout life or whether there is a limit to the period of differentiation, after which all the new angioblasts must come from the growth of preceding endotheUum.


It is in relation to these two theories that I am making these studies on the living blastoderm. It is, I think, clear that the study of blood-vessels in the stages of their differentiation does not prove that they continue to differentiate out of mesoderm throughout life, any more than the finding of several primordia for the thymus proves that new thymus glands continue to arise throughout life. The question of the origin of the blood-vessels is now an exact one — namely, which vessels arise in the embryo (as does the aorta, at least, in part) by differentiation of angioblasts, and which grow from previous vessels. In other words, how long does the period of differentiation of angioblasts continue? His formulated the theory that the embryo itself is invaded by angioblasts from the yolk-sac. This theory was based on the following observations: First, that along the myotomes in the earh^ stages angioblasts can be seen streaming toward the axis of the embryo from the outer margin of the area pellucida; second, that he observed no such streaming of angioblasts toward the axis of the embryo in the zone between the head-fold and the first myotome (here, as a matter of fact, a few angioblasts can be found in early stages, but are much scantier in number than lower down); and third, that the most cephalic part of the head does not receive angioblasts from the membranes. From these observations he concluded that the vessels of the axis of the embryo must arise from a growth of the angioblasts which could be seen to enter the embryo at certain places.


Although these observations of His are for the most part correct, that a differentiation of new angioblasts does take ])lace along the axis of the embryo was shown by two series of experiments. First, those of Hahn, who cut out the membranes of one side of a chick in the stage of the primitive streak and obtained a few specimens in which the membranes were entirely lacking, but the aorta was formed on the injured side. Second, the experiments of Reagan, in which he cut off a part of the head of the chick in the stages just before and just after the head-fold is visible, and allowed the isolated parts to remain in the egg and develop. In these isolated fragments he obtained vessels.


The fact that angioblasts do differentiate in the axis of the embryo is conclusively proved by my observations, having watched certain cells differentiate and join the aorta in the living blastoderm. In what I have called the second zone of the axis of the embryo — that is, the zone between the head-fold and the first myotome — the process can not be followed with such minute detail as is possible opposite the myotomes, because in the former case the heart lifts this zone of the embryo from the cover-slip; but every specimen shows chains of angioblasts which arc apparently differentiating in situ in this area. None of the experiments or observations here recorded take into account the ultimate point of origin of the cells which differentiate into angioblasts.


By the time the chick has 9 somites the dorsal aorta is readily seen behind the head-fold as a complete vessel if the living embryo be viewed from the ventral aspect. Opposite the upper myotomes the aorta is directly ventral to the myotomes, but it gradually curves outward, so that opposite the ninth myotome and the undifferentiated mesenchyme it Ues along the lateral border of the myotomes. Along its entire lateral border it is corrected with the plexus of the area pellucida. As has been shown by Evans, the entire caudal portion of the aorta is a part of the capillary plexus of the area pellucida. In summing up the question of the origin of the aorta it may be said that it differentiates as a part of the plexus of angioblasts, extending over the entire area vasculosa, and is increased by the addition of new angioblasts along the axial line of the embryo.


By the time the chick has 9 somites the aorta can be injected; it forms from the plexus of angioblasts while the seventh, eighth, and ninth somites are forming. While the dorsal aorta of the region of the myotomes is best seen in the living chick, the cephalic aorta is best observed in an injection. As can be seen in plate 1, figure 3, the heart is a simple tube. In some specimens, even with 10 somites, it is in the exact mid-ventral line; in others, as in plate 1, figure 3, it is slightly to the right. In some of my injections the ventral aorta has numerous mesial and lateral sprouts; in this particular specimen these sprouts are more numerous along the dorsal cephalic aorta.


In one of my injections the heart itself shows a little of the primitive plexus. The dorsal cephalic aorta shown in plate 1, figure 3, is still in the form of a plexus; from the arch of the aorta two very constant sprouts extend to the ventral surface of the forebrain. The development of these sprouts is well shown in a figure by Evans from a duck embryo with 13 somites (fig. 398) in the "Manual of Human Embryology" (Keibel and Mall).


The mesial sprouts do not form permanent vessels; but in one very interesting abnormal embryo which I injected these mesial sprouts had formed anastomoses across the mid-line. They are thus, in this specimen, analogous to the vessels which cause the fusing of the two aortae lower down. Opposite the region of the heart some of the lateral sprouts extend out in the somatopleure, as shown in plate 2, figure 1, and in sections in Duval's plate xvii, figure 276. Opposite the midbrain some of these lateral sprouts may connect with the superficial plexus.


Fig. 1. — Transverse section of an injected chick of 12 somites, passing through the middle of the mesencephalon, to show the vascular plexus on the mesencephalon. On the left, the dotted area shows how far the ink passed through a dorsal artery from the aorta into the plexus on the midbrain. The section is from the same series as those iu figures 2 and 3, and it is to be compared with the total preparations shown on plate 1, figure 2, and on plate 2, figure 1. The section is 50 ;« thick and is unstained. X133. A. me., artery to the plexus on the mesencephalon; Ao. d. c, aorta dorsalis cephalica; Ao. V. c, aorta ventralis cephalica; Me., mesencephalon; P, pharynx; PI. mc, plexus on the mesencephalon.


The dorsal cephalic aorta itself, as seen in plate 1, figure 3, is very large. From the dorsal aspect it is broad and flat and is placed in a nearly exact transverse axis instead of in the oblique position which it subsequently assumes.


The next stages in development, including the relations of the primitive cerebral vessels and the cardinal system of veins up to the stage of 14 somites, I shall describe wdth the aid of two total preparations from chicks of 12 and 14 somites and three sections from the stage of 12 somites (figure 2 of plate 1 , figure 1 of plate 2, and text-figs. 1, 2, and 3).


At the stage of 12 somites the aorta is verj- readily injected. The vessels to the brain, however, though they connect with the aorta, are difficult to inject. In plate 1, figure 2, is shown the usual result of injecting a small quantity of ink into the omphalo-mesenteric veins at the stage of 12 somites; the ink passes through the heart and the aorta into the capillaries, which are the forerunners of the omphalo-mesenteric Fig. 2.— Transverse section of an mjectcd chick of 12 somites, passing vf y'cx Tl i i +1-11Q o » +1 V. through the first interspace to show the relations of the primitive arteries. XniS is true, even tnOUgn vessel of the hlndbrain, the transverse vein of the first Interspace, vessels to the enthe brain— that is, the same series as those of figures 1 and 3, and is to be compared with the total preparations shown on plate I, figure 2, and on plate 2, figure 1. The tO the forebrain, midbrain (text-fig.

section IS 50 M thick and is unstained. X133. .4o. d. c, aorta dor 1), and hmdbrain (text-ng. 2) are saUs cephaUca; C, coelomji', pharn.\-x; F. c. a., v.'cardinaUa ante. omphalo-mescnterica ; V. so., vein of the somato present and connect with the aorta, pleure; v. /., v. transversa of the first interspace: Va. p. r.. vasa although the common cardinal vein rhombcneephali; Ven. c, ventnculus cordis.


is present down to the twelfth interspace (text-fig. 3) and the entire lateral border of the aorta opposite the myotomes is connected with the plexus of the area vasculosa. In plate 1, figure 2, the only branch of the aorta injected is an unusual dorsal branch opposite the tenth somite, passing out into the somatopleure. In order to fill these different branches of the aorta in the stages shown in figure 2 of plate 1 and figure 1 of plate 2 before the circulation has begun, it is necessary to introduce the needle into the aorta and inject, as it were, backwards. In this way the pressure in the aorta is probably raised, the heart being sufficiently stimulated by the ink to force the injection mass into the tiny channels that would other-n-ise remain empty. Indeed, after the circulation has begun, if only a very small quantity of ink enters the heart it will return to the area vasculosa without injecting the branches of the aorta within the embryo. These fill up only as the injection is continued and the heart becomes well filled with ink.


That there are vessels within the embryo at the stage of 12 somites which can be injected from the aorta is in-oved by three sections from an injected chick of this stage (text-figs. 1 to 3). These sections are best followed by comparing them with the specimen shown in plate 2, figure 1, from a chick with 14 somites. The section shown in text-figure 1 passes through the midbrain and shows a plexus of vessels on the midbrain fully as large as the aorta itself. On the left side of the section (right side of the embryo) is a slender artery containing ink, connecting this plexus with the aorta. This plexus of large vessels on the midbrain, shown in text-figure 1 , also connects with a single longitudinal vessel along the hindbrain at this stage.


These neural vessels, which at this stage are connected with the aorta, have no vent, which probably explains the great difficulty in injecting them. They are full of fluid, and though the ink enters them from the aorta, it does not penetrate far (text-fig. 1). This point is, I think, interesting in connection with the time of the beginning of circulation. As is well known, the heart begins to beat early in the second day. 1 have made a number of observations which show that it beats at the stage of 10 somites. In one instance I injected an embryo of 10 somites in which the heart was not beating, and when a small amount of the ink entered the heart it was stimulated to beat. In another instance I had been watching

Fig. 3. — Transverse section ol an injected chicK of somites

an isolated blastoderm ot 9 somites lor passing through tiio twelfth interspace to show the recover an hour when the heart began to tion of the posterior cardinal veins to the aorta. The uAi iiv^v.» o section is from the same series as figures 1 and 2, and is beat. This occurred just as the tenth to be compared with the total preparations shown on . . ... plate 1, figure 2, and on plate 2, figure 1. The section is somite was begmmng to appear. It is 50 it thiclc and is unstained. XISS. Ao.. aorta-, .V, therefore quite certain that the stage of -ph 10 somites marks the beginning of the heart-beat.


At the time the heart begins to beat its venous end connects with the extensive capillary plexus of the area pellucida in which the omphalo-mesenteric veins arise, and the entire aorta oijposite the myotomes is connected with the capillary plexus in which the omphalo-mesenteric arteries arise. In other words, there is a plexus of vessels covering the entire area opaqua and area pellucida which connects with the venous end of the heart and with the entire dorsal aorta of the embryo o])i)osite the zone of the myotomes. In the area pellucida this plexus of vessels is filled with fluid, but there are very few free cells in the vessels. After the heart begins to beat most of the isolated blastoderms show occasional wandering cells of various types that float into the vessels of the area pellucida, slu)wiiig that these vessels are full of fluid; and when one of these cells approaches the heart in the omi:)halo-mesenteric veins it oscillates back and forth with each beat. It is thus very strikingly apparent that the circulation does not begin for a considerable time after the heart begins to beat. It is difficult to note the exact time of the beginning of the circulation while the chick is on the yolk, for the few red blood-corpuscles that are forced into the aorta are inconspicuous with the powers of the microscope that can be u.sed. In the isolated blastoderms the earliest chick in which I have seen the circulation hciiiii was one of 17 somites. ,\t the beginning of circiilaf ion a few corpustles are siiol into flu* aorta with each boat of the heart. The mounting of the blastoderm on a cover-slip, however, interferes with the circulation much more than with the heart-beat, because the flattening of the blastoderm tends to flatten the vessels and therebj^ impede the circulation. This is often strikingly shown when through mechanical difficulties the circulation is entirely cut off on one side of an isolated blastoderm and not on the other. It is therefore i^robable that the circulation begins in the chick about at the stage of 15 or 16 somites. It is interesting to note that it is at this stage that the duct of Cuvier breaks through into the omphalo-mesenteric veins, whereby the dorsal aorta and the veins of the embryo become connected with the venous end of the heart.


It is thus clear that at the stage of 12 somites, when the head of the embryo contains a complete aorta and a neural system of vessels which consists of a plexus of large vessels on the forebrain and midbrain, and a single channel on the hindbrain, there is no circulation through these vessels due to the beat of the heart.


The connections of the vessels of the brain with the aorta are of importance. The arteries connecting the vessels of the forebrain with the aorta consist of a group of vessels just at the primitive arch of the aorta. These are shown in plate 2, figure 1, and have been thoroughly demonstrated by Evans. These arteries connect wdth the neural vessels at the base of the optic cup, in the groove representing the line between the telencephalon and the dicncephalon. .Subsequent^ this group of vessels divides into two arteries, one of which encircles the optic stalk and the other extends caudalward along the ventral border of the thalamus and the midbrain (plate 6). Opposite the midbrain there is a group of tiny arteries connecting the plexus with the aorta, one of which is shown injected in a chick of 12 somites in text-figure 1.


It is very clear (in the section of text-figure 1) that the vessels to the neiu-al plexus are direct, dorsal branches of the aorta. The vessel along the hindbrain connects with the aorta by tw^o groups of tiny branches, one cephaUc and the other caudal to the otic ^■esicle. These branches are also for the most part direct dorsal branches. In one of my sections, however, two arteries to the vessel of the hindbrain are placed with reference to the aorta, as are the vessels on the left side in text-figure 2 — that is, one is dorsal and the other dorso lateral. These connections between the aorta and the primitive vessel of the hindbrain are shown, injected, by Evans, in his figure 393 in the "Manual of Human Embryology" (Keibel and Mall).


As far as the vessels which connect the \ascular channel of the hindbrain with the aorta are concerned, it has been shown that they differentiate as angioblasts at the stage of 6 somites, while the aorta and the neural vessels are differentiating. The origin of the primary plexus of deep vessels on the surface of the forebrain and midbrain requires more careful study during the stages of from 6 to 12 somites. It is probable that these vessels differentiate, and that their connections witn the aorta differentiate, as does the preliminary vascular channel of the hindbrain. The development of the deep neural vessels and the origin of the superficial plexus of vessels opposite the brain, as well as the origin of the primary head-vein, will be taken up subsequently.

Origin Of The Cardinal Veins

It is now importaiit to consider the cardinal veins — how they arise, how they become related to the primitive neural vessels, and how they become connected with the heart through the duct of Cuvier. The general relations of the cardinal veins are best shown in plate 2, figure 1, from a chick of 14 somites, but their origin can be traced back to the stage of 9 somites. They form as a longitudinal anastomosis which connects diverticula of the aorta that project dorsalward between the somites. In 1906 these dorsal diverticula were described by Grafe, who stated that the cardinal veins arose from sinus-like projections from the aorta. That the cardinal veins arise from dorsal intersegmental branches of the aorta was show-n by Rabl in 1892 and by Hoffmann in selachians in 1893.


The condition of the aorta just before the diverticula arise is of importance. Up to the stage of 9 somites it is clear that the entire aorta which can be seen from the ventral aspect in the living blastoderm is connected w'ith the plexus of the area vasculosa through so-called ventral branches which extend lateralward. Even cephalic to the first myotome a few^ chains of angioblasts connect the aorta with the plexus of the area vasculosa. However, these tiny branches all along the lateral border of the aorta are seldom injected, except opposite the caudal end of the aorta (plate 1, fig. 2).


When the chick has 9 somites a new set of aortic branches begins to form, which are very distinct from the lateral vessels. In the living blastoderm of from 9 to 12 somites it can be seen that diverticula of the aorta project dorsalward into the interspaces. The more cephalic of these diverticula are dorso-lateral, as shown on the right side in text-figure 2, from a section through the first interspace; the more caudal ones are distinctly dorsal, as seen for the twelfth interspace in text-figure 3. This is due to the fact that at the stage of 12 somites the aorta is obli(iuely placed wuth reference to the lateral margin of the myotomes. As shown in text-figure 2, in the first interspace the lateral margin of the aorta is in the lateral fine, while in the twelfth interspace, as shown in text-figure 3, the aorta is directly under the lateral line. The first two of these diverticula have been seen at the stage of 9 somites; and they are present in all of the interspaces at the stage of 12 somites. In a total preparation of a chick of 12 somites the ink lodges in these dorsal diverticula and forms dark streaks ticross the aorta from the dorsal aspect; these streaks are very characteristic, but arc difficult to indicate in a drawing. The specimen of plate 1, figure 2, shows such streaks across the aorta in the interspaces.


The diverticula begin at the time when the first two somites lie within the arch formed by the two omphalo-niestMitoric veins where they join the heart. In this connection I have tried to determine whether there is a constant relation in regard to the time when the cardiac or head fold reaches tli(> level of the first somite; and in this regard the figures in His's "Untersuchungeii ueber die erste Anlage des Wirbelthierleibes" (1868), plate xii, niid tliosc in Lilly's "Development of the Chick" (1908) are the most helpful. In general, at the stage of 9 somites the position of the first somite is about as shown in Lilly's figure 61, page 106; but I have chicks of 10 somites in which the first somite is fartlier from the head-fold than in the usual specimen of 9 somites. As a rule the head-fold is along the cephahc border of the first somite when the embryo has 12 somites; but in some specimens, such as in His's plate xii, figure 20, there is at this stage an interval between the first somite and the head-fold. After the cephalic curve of the midbrain has formed, as shown in plate 2, figure 2, the embryos are not as flat from the direct dorsal aspect and the point can not be tested with the same definiteness.


The longitudinal vessel which connects these diverticula in the lateral line of the embryo is the common cardinal vein (plate 2, fig. 1). The cardinal vein has two fundamental relations — on the one hand to the primitive vascular channel of the hindbrain and on the other hand to the venous end of the heart. As is shown m plate 2, figure 1, and in the section in text-figure 2, the cardinal vein becomes connected with the neural vessel by two cross-anastomoses in the first and second interspaces. Of these vessels the one in the first interspace is the larger and more important. The cardinal vein itself is not shown on the left side in text-figure 2 (right side of the embryo), since the transverse vein of the first interspace is slightly oblique, as is plainly seen in plate 2, figure 1.


The transverse vein of the fixst interspace has been described and illustrated by Evans; and has been traced back to the stage of 6 somites by Williams. In the chick it is an important channel in the second and third days of incubation, for it is the channel by which all of the blood for the brain drains into the cardinal vein and thence to the heart. The transverse vein of the first interspace is characteristic of the chick. It does not form in the pig where the transitory A-essel of the hindbrain connects with the cardinal vein in front of the first somite instead of in the first interspace.


At the stage of 12 somites the dorsal diverticula of the aorta are present in all the interspaces, but there is not yet a continuous vein connecting them opposite the lower interspaces. The cardinal veins begin to form at a very early stage, when the zone along which they form is close to the aorta (text-fig. 3), so that the primitive common cardinal vein is an accompanying vein to the aorta. It is this accompanying vein of the aorta which connects with the venous end of the heart, forming the duct of Cuvier. So close is its relation to the aorta that the duct of Cuvier may be regarded as a direct connection between the dorsal aorta and the omphalo-mesenteric veins.


The position of the duct of Cuvier is well known, and is shown in plate 2. figure 1. At the stage of 14 somites, as sho\sTi in this figure, the common cardinal vein opposite the second, tliird, and fourth somites is in the form of a plexus; and it ^^■ill be noted that there is a vessel extending lateralwards from this plexus opposite the cephaHc border of the omphalo-mesenteric veins, and a similar vessel opposite the caudal border of the vein. These two vessels are in the somatopleure dorsal to Ihe omphalo-mesenteric veins. This is very clearly showTi in the section in text-figure 2. The more cephahc of these two vessels (V. so.) develops, as I shall show for the pig, into veins which drain the body-wall over the region of the heart cephahc to tlie (hict of Cuvier. They receive their l)l()od from lateral branches of the aorta (of which the lateral artery ojiposite the heart, shown in plate 2, fig. 1, may be one) and are analogous to the branches of the umbilical veins below the duct of Cuvier.


Of the veins of the somatopleure, those which are opposite the caudal border of the omphalo-mesenteric veins join the omi)halo-mesenteric vein in the septum transversum of His, us shown on the right side of text-figure 2. The connection, which has not taken place in the specimen in plate 2, figure 1, at the stage of 14 somites, occurs at the stage of 15 somites, as was shown by Evans. In making injections at the stage of 15 or 16 somites it sometimes happens that the ink first injected does not fill the neural vessels, but runs from the aorta into the duct of Cuvier through direct aortic branches, such as that shown in the third interspace in plate 2, figure 2.


One of the most interesting points in connection with the duct of Cuvier is that it forms just about the time or just before the time when the circulation begins, which is probably of great imjiortance from the standpoint of the physiology of the embryo. Thus plate 1, figures 2 and 3, and plate 2, figure 1, represent the blood- vessels of the chick before the circulation has begun, while figure 2 of plate 2 and figure 1 of plate 3 represent a series of chicks in which the circulation has commenced. Inasmuch as the duct of Cuvier has not connected with the omphalo-mesenteric veins (sinus venosus) at the stage shown in plate 2, figure 1, the longitudinal plexus and vessel of the lateral line at this stage is a common cardinal vein which will be divided into an anterior and a posterior division by the position of the duct of Cuvier. From a comparison of figures 1 and 2 of plate 2, figure 1 of plate 3, and plate 6 it is clear that the anterior cardinal vein must increase in length at the expense of the posterior cardinal vein as the heart shifts caudalward. In these figures it is plainly shown that the cardinal system opposite the duct of Cuvier continues in the form of an extensive plexus (see also plate 1, fig. 1 , of the i)ig) and that the plexus ultimately gives rise to the umbilical veins.


This completes the general account of the blood-vessels of the chick before the circulation has begun; that is, up to the stage of ])late 2, figure 1. I shall take up, under two headings, \hv study of the further develoi)ment of the primitive blood-vessels in the stages in which the blood is circulating; first, the ve.s.sels of the brain and their relation to the primary head-vein; second, the vessels of the sjiinal cord. The primitive vessels of the nephrotomcs will be taken u]) in connection with the pig embryos. It is of course evident that the two divisions overla]), for the ves.sels of the ])raiii begin in the period Ix-forc the circulation commences.


Vascular System Of The Brain And The Primary Head Vein

As has been shown, the neural vessels begin to form very early, before there is any circulation, and indeed before the heart has begun to beat. The primitive vessel of the hindbrain differentiates at the stage of 6 somites as a chain of angioblasts along the border of the hindbrain. and at the time it is differentiating connects with the aorta by chains of angioblasts which are forerunners of direct dorsal branches of the aorta.


Exactly when the angioblasts along the forebrain and the midbrain can be identified has not been determined, but at the stage of 12 somites there is a plexus of large vessels along the lateral surface of the forebrain and midbrain extending to the ventral surface of the forebrain at the base of the optic vesicle and anastomosing with the primary vascular channel of the hindbrain. This plexus connects with the aorta just at the base of the optic vesicle, as was shown by Evans in his figure 398 for a duck embrj^o of 13 somites in the "Manual of Human Embryology" (Keibel and Mall) At the stage of 12 somites this plexus also connects with the aorta opposite the midbrain, as shown in text-figure 1.


This deep primary ]ilexus, which I have uniformly represented in a gray tone, soon gives rise to a superficial plexus opposite the region of the forebrain and the midbrain, as shown in text-figure 1. In this superficial plexus there develops a venous channel for the forebrain and the midbrain, as will be seen in plate 2, figure 2, from a chick of 16 somites, which is the stage when the blood begins to circulate. The su]3erficial plexus opposite the forebrain and the midbrain arises, for the most i:)art, from the deep plexus (text-fig. 1), but I have also injected a few tiny connections between the superficial plexus and the aorta itself in early stages. These, however, disappear and the suj^erficial plexus drains only the deep plexus.


The vein which develops within the superficial plexus is characteristically placed, and is very adequately shown by Evans for the stages of 17 to 25 somites (Anatomical Record, 1909, III, figs. 3 to 6). At the stage of 29 somites this primitive cerebral vein is clearly shown in plate 6. Owing to the flexure of the midbrain, the primitive cerebral vein (v. cap. p. 1) runs directly across the thalamus; and it receives a very interesting series of branches. It is obvious that a very large number of the primitive veins opposite the cerebrum drain the eye. Beginning with the position of the Gasserian ganglion, as seen in plate 6, there is a plexus of veins which I have called the primitive maxillary veins (v. m. p.), which drain the inferior part of the eye and the mo.^t anterior border of the cerebrum. These veins have usually' been called the primitive inferior ophthalmic veins, and, according to the function which the}' actualh' perform at the stage of plate 6, this would be perhaps a more logical name. However, the stage when this plexus drains mainly the eye is very transitory. Soon the capillaries of the maxillary arch develop and the plexus of veins which, at the stage of plate 6, clearly lies in the maxillary arch, drains all of the structures of that arch, the roof of the mouth, and the nose. The position of the maxillary vein and its corresponding artery in the maxilla is shown for the ])ig in plate 7. In the chick of the fourth and fifth days of incubation this group of veins clearly drains the entire maxilla and receives branches from the most anterior part of the cerebrum and a group of inferior ophthalmic veins, of which one of the most important runs in the optic stalk. Therefore I have preferred to limit the name primitive inferior ophthalmic veins to the branches of the primitive maxillary vein instead of calling the entire trunk the ophthalmic veins. The emphasis on the fact that this group of veins belongs in the maxilla, bringing it into line with the veins of the mandibular arch and with the veins from the rest of the aortic arches, is interesting in comiection with the origin of the middle segment of the prhnary head- vein.


CephaUc to this plexus of maxillarj' veins is an extensive series of veins from the marginal vein of the optic cup. The vein in the margin of the optic cvip is verj'^ characteristic. Above these is a smaller, but very important, group of veins which drain the ccrebrimi proper. As can be seen in plate 6, they tap the deep plexus of the cerebrum at their tips; they gradually creep dorsalward on the deep plexus until they meet with those of the opposite side in the mid-dorsal line. The anastomosis of these veins in the mid-dorsal line will ultimately give rise to the superior sagittal sinus, as has been shown by INIall and Streeter. On account of the relation of the primitive veins of the neural tube to the ultimate formation of the dural sinuses, this process of the creeping of the primitive veins toward the mid-dorsal line on the deep plexus is very important.


Over the thalamus at this stage is the main root of the primitive cerebral vein and one large accessory root. For the midbrain the superficial branches of the primitive cerebral brain have not yet appeared, and all the blood of the midbrain drains through the deep plexus toward a characteristic deep vessel along the cerebellar ridge, which joins the primitive vessel of the hindbrain. In plate 2, figure 2, it is clearly shown that at the time when the circulation begins all the venous blood of the forebrain and midbrain must pass through the deep channel of the hindbrain and the transverse vein of the first interspace in order to reach the heart. This figure (plate 2, fig. 2) shows that the vessel of the hindbrain is mesial in position both to the primitive cerebral vein and to the anterior cardinal vein. Plate 6 shows that it is also farther dorsal than either of these veins; and also, what is well known, that the primitive vessel of the hindbrain is mesial to the Gasserian ganglion, the acoustic complex, the otic vesicle, and the ganglion of the glosso-pharyngeus. The cephaUc end of the ganglion of the vagus, on the other hand, is mesial to the transverse vein of the first interspace; that is, the primitive vessel of the hindbrain runs down to the region of the cephalic end of the ganglion of the vagus.


The primitive vessel of the hind-brain serves as a transitory vein for the brain of the chick during the second and third days of incubation, as is very evident in any living chick. On the other hand, it serves as the only channel for the blood to the hindbrain, and it can receive arterial blood directly from the aorta through tiny l)ranches. These branches are so small and are so seldom fully injected that it is probable that only a small amount of blood actuallj^ passes through them in the living chick into the primary channel of the hin(l))rain. The permanent arteries for the hindl)rain develop later, as will be shown.


Plate 6 shows how the primary vascular channel of the hindbrain ceases to serve as a vein for the forebrain and midbrain, and how the true head-vein, the vena capitis prima, develops. The specimen here shown also indicates the fate of the primary vessel of the hindbrain. The deep plexus of the cerebrum, the thalamus, and the midbrain has now made an ahnost complete covering for that part of the neural tube. Over the midbrain the plexus is farthest developed and has anastomosed across the mid-dorsal line with the plexus of the opposite side; over the thalamus and the cerebrum the deep plexus has almost reached the mid-line. The primary artery of the brain which suppUes this extensive plexus divides into two branches — -first, into a large arterial plexus which curves around the dorsal margin of the optic stalk and leads to the plexus around the eye and to the plexus of the cerebrum, and partly supplies the plexus of the thalamus; second, into an artery which curves along the ventro-lateral border of the thalamus and the midbrain, and is approaching the hindbrain. This arterj' will soon meet the ascending artery seen along the ventro-lateral border of the rhombencephalon.


Opposite the hindbrain the development of the vessels, both the arteries and the veins, is most interesting. As is shown in plate 6, there is now a most important new vein. This is as yet a tiny, irregular vessel, hardly larger than a capillary, which connects the veins of the maxillary, the mandibular, and the second aortic arch with the anterior cardinal vein. The primitive vessel of the hindbrain is a vein for the brain only; this new capillary develops out of the capillaries of the visceral arches and by means of the relation of the maxillary veins to the primitive cerebral vein it receives the blood of the primitive cerebral vein and hence it becomes a true head-vein. We shall call this new vein, which is usually called the vena capitis lateraUs, the middle segment of the vena capitis prima {v. cap. p. 2), and will say that as soon as this anastomosis between the primitive maxillary veins and the anterior cardinal veins takes place we can speak of a primary head-vein which extends from the region of the thalamus to the duct of Cuvier and drains the structures of the head, namely, the brain and the tissues of the visceral arches.


The specimen in the drawing of plate 6 is not shown from an exactly lateral aspect, but is tilted slightly to show the ventro-lateral surface of the hindbrain; but even with this tilting it is clear that the general position of the superficial vessel is such that it can become a direct line between the primiti^•e cerebral vein and the anterior cardinal vein. This direct line is very plain in plate 2, figure 2. In other words, it is a more favorable vessel for the drainage of the large primitive cerebral vein than is the primitive vessel along the hindbrain.


The exact course of this tiny chain of new capillaries is most interesting, because it conforms so closely to the structures that are present before it develops. In this connection the relation of this new capillary to the Gasserian ganglion is important to note, because it has been so Uttle understood. As is well known, the ganglion arises from the wall of the pons at the point shown in plate 6, grows v. ntralward, and becomes adherent to the skin, making the placode of the trigeminus. If sections from specimens at the stage of plate 6 are studied, it will be seen that it is tliis attachment of the Gasserian ganglion to the skin, occurring at the stage when the tiny capillaries that give rise to this superficial vein begin, that renders it impossible for the new capillaries to pass lateral to the ganglion; hence they grow mesial to it. The primitive vessel of the hindbrain is mesial to the Gasserian ganglion (trigeminal ganglion), t)ut lies against the hindbraiii; the primary head-vein is also mesial to the ganglion, but lies ventral to the hindbrain.


On the other hand, the new capillaries pass dorsal to the placode of the acoustic complex, and the sUght dorsal curve of the primary head-vein (which is very evident in plate 6) indicates this adjustment of the vein. The placode of the acoustic comjilex is indicated in plate 6 by a film over the jirimitive vessel of the hindbrain ()i)posite the root of the eighth nerves. The vena cajjitis prima passes ventro-lateral to the otic vesicle and again curves slightly dorsalward opposite the ganglion of the glosso-pharj'ngeus.


In another injected specimen of this stage the superficial vein is a slender capillary plexus spanning the gap between the second aortic arch and the anterior cardinal vein, and not yet connecting with the veins of the maxillary arch. Thus this middle segment of the vena capitis prima (the so-called vena cai:)itis lateralis) begins as an irregular cai;)illary plexus between the aortic arches and the anterior cardinal vein. It becomes a true head-vein, in the sense that it drains the entire head, whereas the primitive vascular channel of the hindbrain (vena capitis medialis) is a true neural vessel draining the brain only and not the entire head.


Up to the stage when the capillaries of the visceral arches develop, the primitive channel of the hindbrain serves as the only drainage channel in the head, and this means practically for the brain alone; but as more structures in the head differentiate, a new vascular channel develops to drain these structures. This new chain of capillaries which receives the blood of the primitive cerebral vein by means of the relations of the maxillary veins is so direct and so favorable a channel for the blood of the primitive cerebral vein that the vena capitis prima develops very rapidly at the expense of the primitive vessel along the hindbrain.


By far the most interesting way to follow this transformation is by watching the living cliick. As is very clearly shown in plate 6, there is a stage when there are two venous channels for the head of the embryo — a large, deep channel ak)ng the hindbrain and a superficial tiny capillary chain farther ventral and farther lateral. While this more lateral channel is very tiny, it is hard to see it in the living chick, because there are few if any blood-corpuscles in it, and it is by the injection of blood, as it were, that one sees the vessels. In one chick opened toward the close of the third day and kept in a warm box, the two veins were of equal size when first observed, but in the course of about 2 hours the ventral channel had become by far the larger. This important change can be followed in the living chick either by opening a number of eggs at the close of the third day of incubation and observing the veins by the blood within them or ])y ke(>i)ing a single chick of the right stage under observation for 4 or 5 hours.


In such living specimens it can be seen that the deep vessel of the hindbrain, which remains as a single vessel for 2 days, becomes a cai)illary plexus as soon as the mass of venous blood from the forebrain and midbrain becomes shunted through the superficial vein. In an injection many interesting details of this process can be made out which are not so clearly seen in the living chick. In a stage still earlier than that shown in plate 6, the deep vessel of the hindbrain begins to show very characteristic dorsal branches which conform to the surface of the hindbrain and to the roots of its nerves. In fact, the first of these branches, as can be seen in plate 6, tend to surround the root of the trigeminus, the root of the eighth nerves, and the otic vesicle. While these branches of the jjrimitive vessel of the hindbrain are ft)rmiiig, the vessel itself also becomes a plexus. I have injections which show how this takes place. At first the single channel gives rise to a plexus of very large vessels which tend to run longitudinally, following the pattern of the original vessel. Gradually the vessels of this plexus become smaller and the longitudinal pattern is lost. I have not illustrated the develop)ment of the plexus on the hindbrain for the chick, but this point is well shown for the pig in j^late 7, and the j^rinciples are the same in both forms. The plexus on the hindbrain ultimately covers the hindbrain as completely as the plexus on the midbrain shown in plate G, but the pattern of the plexus is modified by the structures of the hindbrain: (1) by the roots of the nerves and their sensory gangUa; (2) bj^ the otic vesicle, which for a time hes close to the hindbrain; (3) by the special vascular structure of the roof of the fourth ventricle. As has been said, the plexus into which the primary vessel of the hindbrain first breaks up tends to have a longitudinal pattern; the ultimate plexus over the hindbrain, on the other hand, tends, hke the rest of the neural plexus, to show indistinct transverse lines. This is, I think, plain in plate 7, and it leads to the subject of the new arterial supplj^ for the vessels of the hindbrain.


A most important point in the history of the transformation of the primitive vessel of the hindbrain concerns its relation to the neural arteries, and this point is well brought out in plate 6. Taking into consideration the entire neural tube, it is originally supplied by a series of arteries from the aorta: (1) a group of vessels to the forebrain, that is, to the cerebrum and the thalamus, at the base of the optic vesicle from the primitive arch of the aorta; (2) a few small arteries opposite the midbrain; (3) a series of small arteries to the primitive vessel of the hindbrain; (4) a series of intersegmental arteries, of which the most cephaUc is in the first interspace. In plate 6 an artery is shown on the right side from the primary arch of the aorta, which is growing along the ventro-lateral surface of the thalamus and the midbrain, and this artery is approaching a new artery, which is at the same time growing forward along the hindbrain. This new arter}- is very important; it starts as a longitudinal anastomosis along the neural tube between the segmental arteries. In plate 6 it connects the first, second, and third segmental arteries, which are occipital vessels, and is growing forward, making more and more new connections with the deep vessel of the hindbrain. Plate 6 shows none of the primitive arteries which connect the primitive vessel of the hindbrain dhectly with the aorta; but in plate 7, from a pig of a still older stage, it is very interesting to note that two of these original arteries still persist and take part in the formation of this new longitudinal artery. This longitudinal artery grows rapidly forward until it joins the corresponding descending artery opposite the midbrain. It is very clear in plate 6 that the longitudinal neural artery along the hindbrain is originally along the ventro-latoral border of the hindbrain. and thus that there is one on each side. In plate 6 this vessel is labeled the basilar artery {a.b.), which is an illustration of the fact that the relations of the arteries of the adult may have too great an influence on the naming of the embryonic vessels. This vessel is not even a capillary which will become the basilar artery, because it is not in the mid-ventral line; it is rather a vessel which will become a part of a capillary plexus that will gradually reach the mid-ventral line, where the basilar artery will form. At the stage of plate 6 there are bilateral longitudinal arteries along the thalamus, the mid-brain, and the hindbrain, as can be proved by a direct ventral view of the specimen. The relations and the importance of this vessel would be emphasized by calUng it a part of the primary longitudinal neural artery. On the other hand, the vessel shown in plate 7 from a pig embryo of an older stage is in the mid-ventral line and is thus the true basilar artery.


During the fourth day of incubation the longitudinal arterj- seen opposite the first, second, and thu-d somites in plate 6 grows caudalward along the ventrolateral surface of the spinal cord on either side, to the caudal end of the neural tube. These ventro-lateral arteries develop as a longitudinal anastomosis between all the segmental arteries of the spinal cord. At the stage of the fourth day of incubation it is clear that the vascular plexus along the entii-e surface of the neural tube is suppUed with blood by bilateral ventro-lateral arteries which extend from the groove between the cerebrum and the thalamus to the caudal tip of the tube. These two longitudinal arteries are originally in the form of a plexus on either side of the subthalamus, as is still better shown in plate 7 for the pig, and are more definitely a single channel along the rest of the course.


This longitudinal neural artery receives its blood from the forerunner of the carotid arteries on either side and from the segmental arteries. It is easy to see that it is these important longitudinal arteries which will ultimately give rise to the circle of Willis, the basilar artery, and the anterior spinal artery.


The development of the anterior spinal artery has been worked out in the pig by Evans (1909 and 1912). In the chick the anterior spinal artery does not form until the fifth day of incubation. During the fourth day there are two ventro-lateral arteries along the spinal cord which are placed on either side of the notochord and are not connected excei)t by an occasional cai>illary across the mid-ventral line; they make a sharj) ventral boundary for the lateral plexus on the spinal cord. These two longitudinal arteries are just mesial to the point where the spinal arteries meet the sjiinal cord, as can be seen in Evans's figure 437c in the "Manual of Human Embryology" (Keibel and Mall). They give rise to the characteristic anterior arteries which ])enetrate the si)inal cord. During the fifth day of inculcation these two longitudinal arteries l)ecome connected with each other across the mid-ventral line, which is the ])eginning of the formation of the anterior spinal artery. The stage of the fourth day of incubation for the chick in which there are bilateral longitudinal arteries along the ventro-lateral border of the entire neural tube from the jioint of origin of the carotid artery to the tip of the spinal cord is an important stage for understanding the blood-supply of the nervous system.


It must be made very clear indeed that the longitudinal artery seen along the hindbrain in plate 6 is a neural artery and is not the vertebral artery. This is a specimen of the third day of incubation and the artery shown in this specimen forms along the neural tube at a stage when the occipital arteries supply only the neural tube. On the fifth day of incubation, on the other hand, these same arteries also supply the corresponding myotomes with vessels, and there then forms a second longitudinal anastomosis on either side along the upper segmental arteries which is nearer the aorta than the neural vessel. These second longitudinal vessels become the vertebral arteries. These arteries form at the stage of the fifth day of incubation in the chick and are present in a pig measuring 15 mm., a very much older stage than the one shown in plate 7, which measured 6.5 mm. The vertebral arteries form as the heart is shifting farther caudalward; and indeed it is clear that the basilar and anterior spinal arteries together, as well as the vertebral arteries, provide for the arterial supply of the hindbrain when the shifting relations in the neck interfere with the direct arteries from the aorta. The fundamental relations of the neural arteries to the plexus on the surface of the neural tube has now become clear. This plexus is fed with arterial blood from bilateral longitudinal arteries which are along the ventro-lateral border of the plexus and eventually come to Ue for the most part in the mid-ventral line. Over the surface of the subthalamus the vessels remain bilateral.


It is now necessary to consider how the neural plexus becomes related to the veins. In the studj^ of the development of the veins of the brain as distinct from those of the spinal cord, it is of primary importance to study how the deep plexus of vessels becomes related to branches of the primary head- vein. This point I have worked out more in detail in the pig and shall therefore take up its consideration later. The fundamental points are, however, (1) that the branches of the primary head-vein opposite the forebrain and midbrain are transverse veins sujierficial to the deep plexus which constantly tap the deep plexus at their tips and grow toward the mid-dorsal line; (2) that the transverse veins of the hindbrain are profoundly influenced by the presence of the gangUa of the hindbrain and by the otic vesicle. The sensory gangha become as completely surrounded by a capillary plexus as the neural tube itseK, and each of these plexuses gives rise to a vein or group of veins. Moreover, the same is true for the spinal gangUa.


In tliis account of the origin of the neural vessels great stress has been laid on the development of the vessels of the hindbrain, on account of the pecuhar relations of the primitive vessel of the hindbrain to the drainage of the forebrain. In the course of the development of the vessels of the hindbrain the direction of the circulation of the blood is ultimately exactly at right angles to its original course. This change takes place, (1) by the completion of the true head- vein, by which the pial vessel is relie^•ed of a great volume of venous blood from the brain; (2) by the development of a new longitudinal arterial channel, by which it can receive a much greater arterial supply. By these changes the blood over the hindbrain soon runs from the ventral toward the dorsal border, at right angles to its original course from the cephaUc to the caudal border.


In the transition from tlie stage in which the primitive channel of the hindbrain serves as the vein of the brain to the stage when the new lateral su])erticial vessel — the true primary head-vein — is complete, it is clear that the primitive transverse vein of the first interspace is cut out of the main line of drainage for the head. It does not form a part of the ])iiniary head and neck vein of the embryo. Thus the primary head-vein, from the standpoint of development, consists of three parts: an anterior division, which is the primary cerebral vein; a second portion, which is a true head-vein draining the entire brain, forebrain, midbrain, and hindbrain, as well as the visceral arches; and thirdly', the anterior cardinal vein. The transverse vein of the chick persists as a root of a characteristic vein of the hindbrain — namely, a vein which arches caudalward along the lateral surface of the medulla. This vein of the medulla will be followed farther in the pig. It was called the posterior cerebral vein by Mall. The position of the transverse vein of the chick embryo in the first interspace is also just opposite the cephalic end of the ganglia of the vagus nerve. As soon as the superficial vein — the prunary head-vein — is formed, the vascular channel of the neck straightens out, and there is then no longer any way of distinguishing the exact place where the second segment of the primary head- vein joins the third segment or the anterior cardinal vein, for the two become a single, continuous channel. From now on, the place of transition can be indicated only in a general way by the root of that vein of the medulla which follows the roots of the vagus nerve along the medulla; and it is well known that veins are shiftmg landmarks. Stated in other words, the anterior cardinal vein extends along the entire zone of the occipital myotomes, and as the occipital muscles develop these myotomes become indistinct landmarks.


The first interspace is thus a transitory landmark, and in later stages and as soon as the superficial head-vein connects directly with the anterior cardinal vein and eliminates the transverse vein of the first interspace from the direct line of drainage for the blood of the brain, the distinction between the head-vein and the anterior cardinal vein becomes less obvious. The cephalic portion of the head- vein develops to drain the forebrain and midbrain; the middle portion develops to drain the brain and the gill-arches. The vein of the anterior part of a chick of the fourth day of incubation is therefore a composite structure, so far as development is concerned. However, at the fourth and fifth day of incubation there is a single long vein extending from the groove between the cerebrum and the thalanms down to the duct of Cuvier. This vein receives branches from all the various structures of the head. The neural branches come from the cerebrum and the eye from the thalamus, the midbrain, and the hindbrain. The branches from the hindbrain are especially modified by the ganglia of the hindbrain and the otic vesicle. On the ventral as])ect this vein receives branches from the developing visceral arches and from the somalo])leui(' opi)osite the heart. The entire vein may thus l)e called the embryonic lunid-vein, or the vena capitis prima.


As far as the relations of the vena cai)itis prima to the vessels of the adult are concerned, it has been shown by Mall and Streeter that only a very small portion of the primary head-vein persists within the skull cavity — namely, the segment just mesial to the Gasscrian ganglion which becomes the cavernous sinus. The neural branches of the primary head-vein ultimately give rise to the other dural sinuses.


In regard to the relations of the anterior cardinal vein of the embryo to the internal jugular vein, it is interesting to note, in plate 6, that the entire anterior cardinal vein is opposite occijiital myotomes; that is, it is entirely within the head. The caudal jjart of the anterior cardinal vein will become a vein of the neck when the duct of Cuvier shifts into the zone of the cervical myotomes. The cephahc end of the anterior cardinal vein of the embryo is opposite the upper zone of the medulla. The cardinal system of veins in general covers the entire zone of the myotomes, which includes a part of the head as well as the entire body of the embryo.


In closing this account of the origin of the primary head-vein, it is important to emphasize again the relation of the new vessel, the middle jiortion of the vena capitis prima, to the various stiuctures related to the hindbrain — that is, to the otic capsule and to the gangha of the hindbrain. The middle portion of the head-vein develops after these structures are formed and must conform to their position. It grows in as straight a line as possible, and passes mesial to the placode of the trigeminus, lateral to the acoustic complex, to the otic capsule, and to the ganglion of the glosso-pharyngeus. It is entirelj- a new vessel, and has no remnants whatever of the prehminary vascular channel of the hindbrain which arises and runs along the neural tube. As is seen in plate 6, there are two entirely distinct vessels in the head of a chick of the early part of the fourth day — the so-called vena capitis mesiaUs, a neural vessel, and the so-called vena capitis lateralis, a true head-vein.


After following this account of the origin of the primary head-vein of the chick, it will be of value to consider the long series of previous studies upon which it has been based. The observations which seem to me to lead to a clear understanding of this subject are those of Salzer, Mall, Grosser, Evans, WiUiams, and Streeter. The view first held in regard to the development of the veins of the head was that the external jugular vein was the primary vein of this region. This view, which was incorrect, was based on the work of Rathke. In 1887 Kastschenko described a remarkable relationship between the jugular vein and the cranial nerves in the cliick. He stated that up to the end of the third day the cranial nerves were lateral to the jugular vein (primitive vessel of the liindbrain), and noted that this vein was not in the form of a jilexus. At the end of the third day the facial and glosso-pharj'ngeal nerves became mesial to the vein, and on the sixth day the v.igus became mesial. He thought that the nerves cut through the veins, as it were, without the latter losing their continuity.


In 1895 ^^alzer piil)lished an article on the development of the veins of the head in the guinea-pig, which forms the basis of the correct interpretation of this difficult subject. He described the head-vein, in an embryo guinea-pig 2.5 mm. long, as a vessel running from the region of the optic cup, close to the neural tube (primitive vessel of the hind-brain), to the level of the first vertebra, where it turned lateralward and lay lateral to the aorta (anterior cardinal vein), ending in the duct of Cuvier. This vein (the primitive vessel of the hindbrain), was mesial to the cranial nerves, and Salzer called it the anterior cardinal vein. It was a transitory vein, for by the time the embryo was 2.8 nam. long he found a second vein, lateral to the nerves, from the region of the acoustico-facial complex forward. This second vein he called the vena capitis lateralis, and concluded that, not only in the guinea-pig but in vertebrates in general, the anterior cardinal vein (deep channel of the hindbrain) is the first vein to develop in the head, and that it is replaced by a vena capitis lateralis, which as the neck develops is continued into the neck as the internal jugular vein. This description of the veins of the early embryo by Salzer is nearly correct, and was a great step in advance, though more complete studies give a different interpretation and naming of the veins.


The next step was made in 1907, by Dr. Mall, who studied the cerebral sinuses in the human embryo and, on the basis of this work of Balzer, demonstrated that the first drainage canal for the head (primary head-vein including the anterior cardinal) gives rise to the cerebral sinuses and the internal jugular vein. This drainage canal (the vena capitis prima) he called the anterior cardinal vein, using the term in its generally accepted sense as applying to the entire headvein and neck-vein of the embryo.


In the same year Grosser made it clear that the first vascular channel for the head (deep vessel of the hindbrain and the anterior cardinal) can be analyzed into two parts: a cephalic part which Ues close to the neural tube, and a caudal part which has an entirely different position — namely, ventral to the myotomes and lateral to the aorta, in the same position as the posterior cardinal vein. He Umited the term "anterior cardinal vein" to this caudal portion, and analyzed the cerebral portion into a primary vessel (the vena capitis medialis) and a secondary vein (the vena capitis lateralis).


At this point Evans gave his beautiful injections of early blood-vessels, published in 1909. He showed the form of the primitive vascular plexus of the brain and also how this plexus covered the surface of the forebmin, encircling the large optic vesicle with a chain of capillaries and spreading over the surface of the thalamus and midbrain. He described liow this extensive plexus became a single slender channel along the wall of the hindl)rain, leading down to the transverse vein and the duct of Cuvier; and also deinon.st rated the connections of the plexus of the forebrain and the single vessel of the hindbraui with the aorta.


Streeter has recently published a study deaUng with the later stages of the vena capitis prima. It was from the branches of this vein that Dr. Mall had shown that the dural sinuses were derived. Streeter has worked out the development of the dural sinuses more in detail and has sliown that the only jiart of the vena capitis prima to persist is the part mesial to the Gajsserian ganglion which becomes the cavernous sinus. The rest of the dural sinuses come from branches of the primary head-vein. The sinuses of the mid-dorsal line arise from the anastomoses of the veins of the two sides; the basal sinuses arise for the most part from veins which border the Gasserian ganglion and the otic capsule.


In 1916 Stracher pubUshed an article on the veins of the head of the chick, in which he deals with the fate of the vena capitis medialis and the origin of the vena capitis lateraUs. . In this work he uses the method of reconstruction in preference to the method of injection in a form in which it is easy to obtain abundant injected material, on the ground that with reconstructions the relations to the surrounding tissues can be better analyzed. Stracher's owti work, however, suffers from the limitations of his method. His reconstructions show the larger trunks, which are not always the most important ones, and do not show certain tiny channels which are essential to an understanding of the relations of the vessels. He shows the stage at which the primitive vessel of the hindbrain (vena capitis medialis) and the primarj- head-vein are both present in the same specimen and equal in size. This had not been done previously, and is an important point. He also shows in part how the middle segment of the primary head-vein arises, but misses several points that are essential to an understanding of this vein. In his text-figure 2, from a cliick of 30 somites, he shows a short branch from the anterior cardinal vein and a branch from the maxillary (ophthalmic) veins, and recognizes that these two branches become connected and form the vena capitis lateraUs. He speaks of the branch from the inferior orbital vein (my maxiUarj' vein) as arising from a swelling on the vena capitis medialis, not reahzing that it is a new outlet, not for the blood of the vena capitis medialis but for the blood of the primitive cerebral vein, as is plainly shown in plate 6. In discussing the origin of the vena capitis lateralis from the lower border of the Gasserian ganglion to the anterior cardinal vein he says (page 55) : Kastschenko gibt keine Abbildung, die ilu'e Entstehung zeigeu \s-urde, seine Tafel slellt sie da, nachdem ihre Ausbildung vollendet ist. Nach seiner Schilderung "durchsclineiden" die Nerven die Yene. Demgegeniiber ist zu betonen, dass der eben geschilderte Tail der Vena capitis lateralis — es folgt spater noch die Ausbildung vveiterer caudal and cranial davon gelegener Strecken — frei im Gewebe, ziemlich entfernt von der Vena capitis medialis entsteht.


Thus he reaUzed a part of the method of origin of the vena capitis lateralis, but missed entirely its relation to the capillaries of the visceral arches. In regard to the relations of the portion of the primary head-vein in the region of the Gasserian ganglion, Stracher's models are better than his interpretations. The essential facts are that the vena capitis medialis is a vessel on the hindbrain, the vena capitis lateraUs is a more superficial vein which Ues ventral to the hindbrain; both are present in the same specimen at a given stage; both are mesial to the Gasserian gangli-^u, one as a part of the sj^stem of vessels of the pia mater and the other as a part of the primarj^ head-vein.


Stracher shows both the vena capitis mediaUs and the vena capitis lateraUs in their correct position mesial to the Gasserian ganglion, and then concludes that the medial \ein of this area becomes transformed to make the lateral vein. His own figures do not warrant this conclusion, which was formed through not following the fate of the primitive vessel of the hindbrain. Thus his diagram (page 68), which should show the primarj' h(vid-vein coming, embryologically, from three segments — namely, from (in his nomenclature) the vena cerebralis anterior, the vena capitis lateralis, and the vena cardinalis anterior, shows it coming from five: (1) the vena cerebralis anterior; (2) a short stretch of the vena capitis lateralis; (3) the vena capitis medialis; (4) the vena capitis lateralis again; (5) the vena cardinalis anterior. He observed the beginning of the breaking of the vena capitis medialis into a ple.xus, and then missed the plexus as it became finer, so that he lost the very important ]:)oint of the fate of the primitive channel of the hindbrain. These points are covered in his summary (page 67) : "Sodann entwnckelt sich eine neue Venenbahn (Vena capitis lateralis), die parallel zur medialeii Kopfvene, aber lateral voni Ner\ais acustico-facialis, glosso-pharyngeus and dem Horblaschen verliiuft. Sie verbindet das Stiick der Vena capitis medialis, das sich medial von der Trigemiuusanlage findet, niit der vorderen Kardinalvene. Zur selben Zeit weiten sich Ciefiisse des ^'enennetzes am Hinterhiru zu eiuer Balm aus, die bedekt \-on der Trigemiuusanlage beginnt, an der Seite des Hinterhh-ns dorsal von Horblaschen im Bogen verliiuft und in der Gegend des Nervus glosso-pharyngeus wieder zur medialen Kopfvene zuriickkehrt (Vena capitis dorsalis). Sie tritt mit Beginn der ObUteration der Vena capitis medialis auf und verschwindet weder, sobald die ^'ena capitis lateralis voUstiindig ausgebildet ist. Die Vena capitis medialis vcrodet zuerst im Bereich des Horblaschen, dann caudal davon in der Ciegend des Nervus glosso-pharyngeus. Weiterhin entwickelt sich dadurch um den ersten Ast des NerAiis trigeminus em Veneming, dass lateral vom Nerven eine \eue entsteht, die rostral vom Nerven aus dem Stamm austritt und sich caudal wieder mit ihm vereinigt. Der mediale Schenkel dieses Ringes verschwindet alsb;ild. In almlicher Weise entwickelt sich auch um den Nervus vagus ein Ring mit Beihilfe der dorsal einmiindenden Zweige. Auch hier verodet die alte medial vom Nerven gelegene Balm. Damit is die Vena capitis lateralis vollstjindig ausgcbildet, und die Kopfvene iindert ihren \'erlauf, was ihre Lage zu den Nerven anlangt, nicht mehr, da die Nervi accessorius and hypoglossus beim Huhn auch im ausgebildeten Zustande lateral von Vena jugularis interna ziehen." It is, I think, clear that the primary blood-vessels which arise in the head are neural vessels. These neural vessels form a continuous plexus of capillaries which closely invests the brain. Along the hindbrain angioblasts probably appear first, but here they form a single, characteristic long channel which serves temporarily as a vein and does not take the form of a plexus characteristic of the neural vessels until relatively late. This single, large, primitive vessel does not extend the full length of the rhombencephalon, however, but at the zone of the cephaUc roots of the vagus nerve, or, in other words, opposite the first occipital myotome, becomes a plexus on the side of the medulla which gradually extends the full length of the cord and connects with every intersegmental arterj' and vein. Th(! neural system of vessels becomes cotmected with the venous end of the heart by means of the two cartlinal veins. These connections are very characteristic; the most cephalic, which is either in front of the first occipital myotome (as in the pig) or between the first two occii^ital myotomes (as in the chick), is always the largest and drains the entire brain. All the other connections are small intersegmental veins. Thus it may be said that two organs determine the early blood-vessels, the neural tube, and the nephrotome. i-^oon a third set of organs (the visceral clefts) develop and give rise to capillaries, ^vJlich connect on the one hand with the anterior veins of the brain and on the other with the cardinal veins; and in this manner the head-vein of the embryo is completed.


The method of nomenclature of the primitive vessels of the head is certainly open to discussion. The primitive vessel of the hind-brain, of which I have shown the origin, the relations, and the fate, is the vessel seen by Kastschenko in 1887 and more clearly by Salzer in 1895, and recognized by all who have since worked on this subject as being the first long vein in the head region and as lying along the wall of the hindbrain. Grosser gave it the name of "vena capitis medialis," a name which has been universally accepted.


It may be argued that it is a mistake to attempt to change a name of this type which has been generally adopted; but on the other hand a name w'liich w^ould emphasize the essential point in regard to this vessel — namely, that it is a neural vessel and that it develops into neural vessels — rather than the accessory fact that it serves temporarily as a vein for the head, would, I am convinced, clear up much of the confusion in regard to the primitive veins of the head. It does serve for tw^o daj^s in the chick as a vein for the forebrain and the midbrain, but at the same time it is the entire capillary bed of the hindbrain and ceases to be a single long channel as soon as the cerebral blood is shunted through another channel. It then develops, as have the rest of the neural vessels, into an extensive capillary plexus in the position of the pia mater. I therefore wish to avoid the use of the term vein in connection with it and to reserve the term vein for the vena capitis lateralis, for which I shall use the term vena capitis prima, because this is the first vascular channel of the head which is purelj^ a vein and because it is the first vessel which drains the head and not the brain alone. I have therefore called the vena capitis medialis the primitive vessel of the hindbrain. The term vessel is more indefinite than the term vein, but for that very reason it applies better to a channel which serves both as a vein and as a capillary at the same time, and ultimately becomes a capillary plexus, out of which both arteries and veins will arise. I propose to call the long vein of plate 6, extending from the region of the thalamus to the duct of Cuvier, the primary head-vein. This primary head-vein develops in three segments — a cephahc segment which is the primitive cerebral vein, a middle segment opposite the hindbrain, and a caudal segment which is the vena cardinalis anterior. At the stage of plate 6 this vein is entirely within the head, because the duct of Cuvier is still opposite occipital myotomes. As soon as the duct of Cuvier shifts into the neck region this vein will become the primary head and neck vein. The terms vena capitis medialis and lateralis have the sanction of usage; but it seems to me that the terms primitive vessel of the hindbrain and prijnary head-vein better express the function of these vessels.


Development Of The Spinal Arteries

It will now be necessary to go back and follow the development of the spinal arteries. This can be done in a series of injections of the stages from 15 or 16 somites ujiward. or the entire process can be followed in any one chick up to the stage of about 30 somites. The process is easier to illustrate after the embryo has rotated so that the lateral instead of the dorsal surface is presented. The entii-e process can thus be readily followed in plate 3, figure 1, from a chick of 25 somites, with six sections from two different series of chicks of 25 and 30 somites. The general stage of development of the vessels of the head of the embryo at the stage of 25 somites can be seen in Evans's figure 6 (Anat. Record, 1909, III, p. 505); or can be estimated from my plate 6, the stage of 25 somites being just before the superficial capillaries which make the middle segment of the primary head-vein begin. The deep vessel of the hindbrain is still the vein for the brain, and is shown in its relation to the capillary plexus on the lateral surface of the spinal cord in plate 3, figure 1.


The general development of the area vasculosa at this stage is also of interest m following the vessels in sections. The roots of the omphalo-mesenteric arteries at the stage of 25 somites are opposite the twentieth and twenty-first somites. As was indicated above, in the earlier stages the entire lateral border of the aorta oi^posite the somites was originally connected by direct lateral (that is, ventrolateral) branches with an arterial plexus of the area vasculosa. In this plexus, on either side of the embryo, the omphalo-mesenteric veins gradually extend caudalward from the region of the sinus venosus and thus are formed two veins, or a plexus of veins, with direct short connections w'ith the aorta. This process explains the large veins of the si)lanchnopleure shown in figure 3, plate 2, and figures 2 and 3, plate 3.


In the chick the spinal arteries do not arise as direct dorsal arteries from the aorta to the cord, but the direct dorsal arteries make a primary arch to the dorsal border of the nephrotome, where they give rise to the cardinal veins. The spinal arteries then arise from these arches instead of from the wall of the aorta itself. In following the development of the spinal arteries I shall begin with the more caudal segments in plate 3, figure 1, because they show the earlier stages. As has been described in connection with the origin of the cardinal veins, the first dorsal branches of the aorta are direct dorsal diverticula of the wall of the aorta into the interspaces, as is shown best in text-figure 3 for the stage of 12 somites. Figure 2 on plate 4 and figure 4 on plate 2 are both from the lower segments of a chick of 30 somites. They are both taken below the origin of the omphalo-mesenteric arteries in the zone where the arteries of the j^osterior limb-buds are forming, 'i'he cardinal veins are also developing in this area. Figure 2 on plate 4 i)asses through the twenty-fifth interspace, and plate 2, figure4, is still lower down and passes through the twenty-seventh interspace. In iilate 4, figure 2, it can be seen that even in later stages the dorsal branches start as direct diverticula of the aorta. These diverticula soon arch lateralward and, as can be seen in plate 2, figure 4, dilate slightly just dorsal to the nephrotome. These dilated portions of the arches become connected with similar dilatations in the other interspaces and make the cardinal veins. These observations indicate that the cardinal veins begin with dilatations of the dorsal branches of the aorta— that is, that they start as an outgrowth from the wall of the aorta in the different interspaces and that these intersegmental vessels become connected along the lateral line.


Very soon the vascular arches which give rise to the cardinal veins give off sprouts which extend toward the spinal cord, as is shown from the tenth to the seventeenth interspaces m plate 3, figure 1. The position of these sprouts is shown in section on plate 2, figure 3. This section passes through the twentyfirst interspace of a chick of 30 somites. These neural sprouts soon reach the spinal cord, as is showm in section on plate 3, figure 4, which is to be compared with the seventh interspace of plate 3, figure 1. The section shown in plate 3, figure 4, is from a chick of 25 somites — from another series than that of all the other sections on the plates. It is from a series of nearly the same stage as that of plate 3, figure 1 ; it passes through the seventeenth interspace. It was selected because it shows so well the double dorsal arch to the posterior cardinal vein, the primary direct one and the secondary neural one. One has only to imagine the primary direct arch disappearing to obtain the well-known pattern of the spinal arteries shown in section on plate 3, figure 2, and in the upper interspaces of plate 3, figure 1.


From this studj^ it is, I think, clear that the spinal arteries of the chick arise from dorsal intersegmental vessels which give rise to the cardinal veins, and not directly from the aorta. In plate 3, figure 1, it is very evident that the capillary plexus which forms along the lateral surface of the sjnnal cord is a direct continuation of the primitive vessel of the hindbrain. The original simple chain of capillaries on the lateral surface of the cord, such as is shown in plate 3, figure 1, from a chick of 25 somites, very soon becomes a plexus on the neural tube, as indicated opposite the second somite in plate 6. By the fourth day of incubation this plexus covers the entire lateral surface of the spinal cord. The relation of this plexus to the spinal arteries on the one hand and to the spinal veins on the other is very regular and characteristic. Every spinal artery, on approaching the cord, bifurcates into a short ventral branch and a longer lateral branch (plate 3, fig. 2). The ventral branch leads to a longitudinal neural artery which at the stage of the fourth day Ues on the ventral surface of the cord just lateral to the notochord. In other words, there are sjTmmetrical ventral longitudinal arteries. These arteries form a ventral border for the plexus along the lateral surface of the cord. The lateral arteries run in the plexus on the surface of the cord; they lie just cephaUc to each spinal ganglion and extend nearly- to the dorsal border of the cord. The veins which accompany these transverse arteries, in contrast to the arteries, are lifted off from the surface of the cord, as it were. They also correspond to the cephalic border of each ganghon, and they are more superficial in ever}- case than the corresponding artery. I am emphasizing the fact that the arteries Ue in the plexus on the siu'face of the cord and that the veins are more superficial, because the same is true for the primary arteries and veins of the brain, as can be seen very clearly in plate 7.


The series of sections of injected chicks shown on plates 2, 3, and 4 allow an interesting comparison of the primitive branches of the aorta. It is clear that the branches of the aorta can be described best according to what organs they supply rather than by regarding their exact ]ioint of origin in the wall of the aorta. The primary branches extend to the splanchnopleure and are ]5rimitively directly lateral branches, as can be seen on the left side of text-figure 3. That they come to be ventro-lateral and then ventral l)ranches is well known and is shown in plate 2, figure 4, in which it is clear that there are three sets of arteries on the right side of the section. The first is a dorsal branch to the cardinal vein; the second is a lateral branch to the somatopleure, and the third is a ventral branch to the si)lanchnopleure. The branches that extend to the cardinal veins need careful attention. There are in the first place the original dorsal arches that giv(> rise to the cardinal veins, such as are shown in figure 4 of plate 2 and figure 4 of ])late 3. These branches are strictly intersegmental; moreover, the intersegmental branches are the first arteries related to the cardinal veins because they are the only ones present at the stage of 12 somites. In later stages (for example, at the stage of 25 or 30 somites) there develop a few direct dorso-lateral arteries to the cardinal veins, and these arteries may lie opposite the somites instead of between them. Such an artery, for example, is shown on the left side of figure 3 of plate 3 This section is the next one below that of figure 2 of plate 3 in the series and indeed the edge of this artery is shown in the latter section. Similar direct dorsolateral arteries to the cardinal veins are shown on both sides of figure 4 of plate 3. This latter figure demonstrates that these dorso-lateral arteries are new vessels and not remnants of the original dorsal arches. On the left side of figure 4 of plate 3 l)lood reaches the cardinal vein in three ways: from the aorta along the surface of the cord, from the aorta along the primary dorsal arch, and from the aorta through a dorso-lateral artery. It must also be brought out that these dorsolateral arteries to the cardinal vein are not the same as the direct lateral arteries to the tubules of the pronephros and the metanephros, which develop later and are quite differently placed, as can be seen in text-figure 8 from a chick of 35 somites. These dorso-lateral arteries to the cardinal veins are of imjiortance in connection with th(; extension of the cardinal veins caudalward and are verj^ important in comparing the chick with a form like the pig, where the dorso-lateral branches are more numerous.


It may be well to enumerate here the different types of branches of the aorta which may be found in the embryo from the standi)oint of the structures they supi)ly: first, there are the arteries to the si)lanchn()])leure; second, mesial branches which connect the two aorta?; third, lateral art(>ries to the somatoijleure l(>ading to the umbilical veins; fourth, dorso-lateral arteries to the cardinal veins; fifth, lateral arteries to the limb-buds; and sixth, lateral arteries to the nephritic tubules.


The Vascular System In Young Pig Embryos

In the study of the vascular sj-stem in a mammal it is not as cas.y to obtain young stages for injections, as in the case of the chick. The material, however, offers valuable ojiportunities for comparison with human embryos, and to obtain injections in much earher stages than have ever been injected in human speclinens. I shall follow the development of the vessels in the pig by the aid of six figures of injected embryos, and shall describe the specimens and follow the development of the vessels under six headings: First, the form of the heart ; second, the ventral branches of the aorta, including the allantoic arteries and the subintestinal artery; third, the umbilical veins and the vessels of the thoracic body-wall; fourth, the ^•ascula^ system of the nervous system and the formation of the prmiary head- vein; fifth, the cardinal veins; sixth, the vessels of the pronephros and the mesonephros.


The Form Of The Heart

The youngest pig which I have injected is shown on plate 4, figure 3. This is from a specimen which measures 4 nam. in oil and which has 14 somites. It corresponds in development with a human embryo, Xo. 470 of the Carnegie collection, which measures 3.3 mm. and is in the fourth week of development. In this embryo pig an injection was made into the aorta opposite the origin of the omphalo-mesenteric arteries. The point of injection was obscured by extravasation, so that it is not shown in the drawing. The stage of development of the specimen can be judged by the form of the brain, the otic vesicle, and the form of the heart.


The extensive venous plexus covering the anterior or cephalic wall of the 3^olk-sac converges on either side into large right and left omphalo-mesenteric veins, which meet in a conjoined tube, the sinus venosus. The sinusoids of the liver have not yet begun to form, so that the sinus venosus stands out clearly. The sinus has a marked diverticulum, which Tandler called the horn. The dorsal wall of the sinus shows a series of sprouts, representing the duct of Cuvier, which is probably developed at this stage, as indicated by the posterior cardinal vein, but is incompletely injected. The most caudal of the sprouts form a small plexus representing the umbilical vein in the somatopleure.


Above the sinus venosus is a well-marked groove between the sinus and the atrium. The atrio-ventricular canal, on the other hand, is only just indicated. The form of the heart corresponds closely with the description by Tandler (^Manual of Human Embryology, Keibel and MaU, page 536), which is based on the studies of Born, in which he says that the heart becomes a horizontal loop, the two hmbs of which are separated by an almost horizontal bulbo-ventricular cleft into two parts, a ventricular limb and a bulbar limb. In my specimen the bulbar limb consists of three parts: first, the bulbus cordis; second, a short constricted portion of the tube, the fretum HaUeri; third, the large truncus arteriosis, which gives off the two aortae. In the use of the term fretum Halleri I am following the usage of His (Anatomie menschlichen Embryonen, pages 131 and 140). He describes this portion of the tube as the portion which ultmiately gives rise to the semilunar valves.


In connection with the deveh)i)nient of the heart, figure 1 of plate ') and figure 1 of plate 1 are very interesting. The specimen from which the former was taken was one of a litter of five, all of which were injected. It measures 6 mm. in oil, that is, after fixation and dehydration, and has 20 somites. The specimen on plate 1, figure 1, was one of a litter of six embryos, all of which were injected. It measured 7 nam. when fresh and is 6.2 mm. long in oil. It has 23 somites. It should be noted that these embryos do not have a caudal flexure, so that these measurements must not be confused with the same measurements of older specimens after the flexure has formed. The number of somites gives more valuable data in regard to the stage of development in these stages than do measurements.


If comparisons are made wdth human embryos at the stage of 23 somites it will be noted that at this stage the human embryo has two very marked flexures, shown, for exami^le, in the R. Meyer embryo No. 300, represented in Felix's figure 531, in the "Manual of Human Embryology" (Keibel and Mall), and hence it is very much shorter.


In figure 1 of plate 5 the changes in the heart from the stage shown in plate 4, figure 3, are readily followed. The direction of the ventricular arch has changed from the horizontal to an oblique position. The atrio-ventricular canal has become the characteristic long, slender channel, and there is a marked constriction between the ventricle and the large bulbus cordis. The fretum Halleri is now a long, slender tube, and both the bulbus cordis and the truncus arteriosus are shown in maximum distension.


In plate 1, figure 1, the form of the sinus venosus is not clear, as it is concealed by the injection of the sinusoids of the liver. In all of the six specimens of this litter the sinusoids of the liver are farther developed on the left side than on the right. In all of the other sjjecimens, however, and on the right side of this specimen, there is a marked constriction between the liver and the sinus venosus just below the upper large opening of the umbilical vein. At this stage the umbilical vein connects with the liver and with the sinus venosus by large openings, and with the duct of Cuvier by an extensive capillary plexus in the somatopleure. There is a constriction between the sinus venosus and the atrium, and a well-marked atrio-ventricular canal. The bulbo-ventricular cleft gives the effect of an hour-glass constriction of the heart. This is true of all the si)ecimens of the litter, but in one the contraction of the bulbar portion is particularly marked. The difTerenccs in the form of the heart in figure 1, plate 5, and figure 1, plate 1, are partly due to the fact that the hearts in these specimens were fixed while beating and were caught at difTerent phases of the beat. For example, in plate 5, figure 1. and bulbu.s cordis and the truncus arteriosus show a maxinunn distension, while in i)late 1, figure 1, the l)ulbus cordis and the truncus arteriosus are contracted and there is a general distension of the cejjhalic aorta. On the other hand, in plate 1, figure 1, is shown the begiiming of a torsion df the ventricular loop, by means of which the beginning of the fretum Halleri will come to be oi)i)osite the ventricular cud of the atrio-ventricular canal.


This torsion is more clearly seen on figure 2 of plate 5 and figure 1 of plate 4. These two specimens are from the same litter. They measure 7.1 mm. in oil, and have 27 somites. Figure 1 of plate 4 is given because of an e.xtravasation in the vessels of the head in the specimen of figure 2 of plate 5. In this latter figure the sinusoids of the liver have markedly developed. The sinusoids of the left side anastomose across the ventral fine with those of the right side. The opening of the left umbilical vein into the Uver is directly mesial to the umbilical vein itself and is hidden b}^ it, while the opening into the duct of Cuvier is plainly visible. There is also a plexus from the umbilical vein in the somatopleure connecting it with the posterior cardinal vein and with the duct of Cuvier; but this is omitted in the drawing.


There is a well-marked constriction between the sinus venosus and the atrium. The change in the heart is due to the twisting of the obUquely placed ventricular arch, whereby the point which marks the beginning of the fretum Halleri comes to lie exactly opposite the opening of the atrio-ventricular canal into the ventricle. The bulbus cordis lies far to the right and its connection with the fretum Halleri is hidden bj' the ventricle, while the opening of the auricular canal is far to the left. These relations as seen from the other side are shown in plate 4, figure 1. From these two figures it is obvious that a still further twisting of the heart must take place before the arterial orifice comes to lie directly anterior.


Ventral Branche5 Of The Aorta, Including The Allantoic Arteries And The Subintestinal Artery

One of the most interesting subjects in connection with these injections has been the study of the ventral branches of the aorta, or the branches to the yolksac, the gut, and its derivatives.


The study of the early vessels of the embryo emphasizes the fact that the vessels should be considered in relation to the organs which they supply. The fundamental relations of the ventral branches of the aorta to the yolk-sac and to the allantois are shown m two total preparations of injected pig embryos (plate 5, fig. 1, and plate 1, fig. 1) and in two sections (text-figs. 5 and 6). Plate 5, figure 1, is from a specimen of approximately the same stage as in Evans's figure 394 in the "Manual of Human Embryology," which shows the state of development of vessels of the brain at this stage.


The position of the embryo should be carefully noted. The caudal half of the specimen is seen from the direct ventral aspect, while the cephaUc half is from a direct lateral view. The place of rotation is around the ninth somite.


Extending from the level of the eleventh somite to the caudal end of the embryo there is a series of tiny ventral arteries from the two aortre. These are of uniform size and are placed at regular intervals, approxmiately one opposite an interspace and one opposite a somite. In this particular embryo only a few of these ventral b/anches are injected; but other specimens show that the entire length of both aortse gives rise to branches Uke those shown opposite the twelfth, thirteenth, and fourteenth somites. From the region of the eleventh to the fourteenth or fifteenth somite these tiny branches from the two aorta' unite in a plexus of large arteries on either side of the stalk of the yolk-sac, which join and give rise to the omphalo-mesenteric arteries on the yolk-sac. The large arteries are seen only on one side in plate 5, figure 1, and plate 1, figure 1, but are shown on 1)oth sides in figure 2 of plate 5. From the fourteenth somite caudalward the ventral branches of the aorta are uninjected in this specimen (plate 5, fig. 1), but show in other specimens leading to a single artery which arises in the caudal end of the embryo. Opposite the caudal end of the embryo the ventral branches of the aortse form a sheet of capillaries on either side of the alimentary canal, which deserves careful consideration. These two sheets of capillaries form a plexus which completely surrounds the entire caudal end of the gut cej^halic to the allantois, the stalk of the allantois, and the blind end of the gut, caudal to the allantois. This capillary plexus gives rise to two arteries, the ])aired allantoic arteries and the single subintestinal arterJ^ Thus, we have here examples of arteries in the embrj^o which arise in a capillary plexus and end in a capillary plexus. The primitive allantoic arteries arise in a plexus around the stalk of the allantois and pass to the capillaries of the liody of the allantois; the subintestinal artery arises in a cajiillary plexus around the gut and runs to the capillaries of the yolk-sac.


The allantoic arteries, as seen in plate 1, figure 1, extend into a ple.xus on the ventral or cephalic surface of the allantois; this plexus arches around the dome of the allantois, though not completely shown in the drawing, and reaches the \'eins on the caudal surface. The two allantoic veins join the umbilical veins at the point where the stalk of the allantois is fused with the body-wall. A section through the allantoic arteries from an injected embryo of the same litter as the specimen of plate 1, figure 1, is shown in text-figure 6, and shows the allantoic arteries following the wall of the gut into the allantois. In the series from which text-figure 6 is taken there are a few tiny capillaries extending dorsalward from the allantoic arteries just at the point where these arteries pass ventral to the ccelom. These cai)illaries grow lateral to the ccelom, and when the ]iosterior limb-buds begin they will anastomose with the iliac arteries. These capillaries will become the umbilical arteries in the somatopleure.


These observations on the pig agree with the findings of Hochstetter in the labbit (1890) and show that in these forms the primary allantoic arteries are vitelline vessels, while the central ends of the umbilical arteries are vessels of the somatopleure, which appear later and anastomose with the primitive allantoic arteries.


In the study of the R. Meyer human embryo 300, Felix (1910) gives an exceedingly interesting reconstruction of the vascular system of a human embrj'^o which is of the same stage as my figure 1 of jilate 1. This reconstruction (fig. 7. Mf)ri)h. Jahrb, 1910, XLI, p. 590) shows that the primitive artery of the fetal membranes at the caudal end of the embryo arises in a capillary plexus aroimd the gut, just as is shown in my figure 1 of plate 5 and figure 1 of i^late 1. The position of this plexus in the wall of the gut is shown in section in Felix's figure 9, which is to be compMred with my text-figure G. The same relations are shown for the chick in Duval's Atlas, plate xxiii, figure 372. In the human embryo this artery has been traced back as a vitelline vessel to the stage of 5 somites by Felix (1910), and to the stage of 8 somites by Dandy (1910). This artery in the wall of the gut, which is the primitive allantoic artery in the pig, has been called the umbilical artery in the human embryo on account of the insignificance of the allantois and the earlier vascularization of the chorion. The relations of these two vessels in connection with the human embryo were summed up by Evans (1912, page 595) in the plii-ase that the umbilical artery is merely a modified vitelline vessel. The entire question of the relation of the arteries for the fetal membranes at the caudal end of the embryo has centered around the position of the central end of the arteries with reference to the coelom, as can be seen in text-figure 6; that is to saj^ whether the artery is mesial or lateral to the ccelom. In general, both in birds and in mammals there is a primitive artery mesial to the coelom; that is to say, a splanchnic vessel, and a secondary vessel, the umbilical artery, lat<?ral to the coelom running in the somatopleure. Thus the vessels develop in the same mamier in the different forms, for there is a primitive splanchnic artery followed later by an artery in the somatopleure, but there are variations in the relative importance of the allantois itself.


Besides the two arteries of the allantois, the two sheets of capillaries of the wall of the caudal end of the guc give rise to another artery. Extending forward from the stalk of the allantois, as seen in jilate 5, figure 1, the two plexuses meet in a capillar}^ plexus ventral to the caudal root of the yolk-sac. This jilexus is continued as a single, ventral, subintestinal artery which joins the omphalomesenteric plexus opposite the fourteenth or fifteenth somite. The point where the subintestinal artery joins the omphalo-mesenteric plexus is the well-known intestinal landmark where the stalk of the yolk-sac joins the gut. A figure which gives a verj' clear understanding of these relations is Tandler's figure 1 in the Anatomische Hefte, 1904, P\ page 192.


This subintestinal arterj' in the pig is the more interesting in view of the corresponding subintestinal vein in the chick, discovered by Hochstetter in 1888 and accurately described bj' liim. He described its relations not only to the omphalo-mesenteric veins, but also to the intestinal and the allantoic vessels, and noted that it disappeared and that the left vein was larger than the right. A complete understanding of the development of this vein in the chick can be gained from the figures of Popoff (1894). As was mentioned in connection with the chick, during the early hours of the third daj' of incubation the entire capillary plexus of the area vasculosa caudal to the omphalo-mesenteric arteries must be regarded as an arterial capillary plexus down to the marginal vein, as shown in Popoff's plate v. During the last hours of the third day, as seen in Popoff's plate VI, branches of the omphalo-mesenteric vein gradually extend caudalward on either side of the embryo in the wall of the yolk-sac, and arch around the posterior end of the embryo; the left vessel starts ahead of the right and is alwaj's larger than the right. As these veins gradually- extend backward mto the territory of the pre-existing arterial plexus, forming more and more new connections with the plexus, they change the direction of the current of the blood in the plexus (which lias l^een away from the heart) to a direction towards the heart. The vein on the left side (luickly extends to the marginal vein, making the single posterior vein of the yolk-sac of Popoff, which lies a little to the left of the mid-line, as shown in Popofif's plate viii. The two lateral veins form an arch around the posterior end of the embryo; this arch is just cephalic to the point where the stalk of the allantois will develop.


On the third day of incubation there is a very extensive capillary jilexus on either side of the posterior end of the gut, and beginning at the very caudal tip of the gut on either side are symmetrical ventral veins, which unite in a loop just cephalic to the base of the allantois and then run forward, at first as two veins and then as a single ventral vein in the ventral wall of the yolk-sac. The subintestinal vein is thus the primitive vein for the entire posterior end of the gut, for the caudal tip of the gut, the allantois, and the entire rectum and intestine up to the margin of the yolk-sac. Caudal to the allantois these vitelline veins receive the most caudal branches of the posterior limb-bud. This relation has been described by Evans (Anat. Record, 1909. iii). On the third day the umbilical artery develops around the somatopleure in connection with the posterior limb-bud and anastomoses with the primitive allantoic capillary plexus in the wall of the allantoic stalk. By the beginning of the fourth day the vessels in the stalk of the allantois show an exceedingly interesting relation. On either side there is one large artery coming from the aorta and now running in the somatopleure instead of in the splanchnoi)leure; but this artery is fed also from a cai)illary plexus in the wall of the si)lanchnopleure, which completely surrounds the stalk of the allantois and the caudal tip of the gut, and by a few capillaries of the somatopleure from the tail of the embryo, which capillaries, however, tend to drain more and more into the posterior cardinal veins.


These relations are clear in the light of the development of these vessels. There is at first a plexus of capillaries arising from the aorta and running in the stalk of the allantois, in w^hich arise the jirimitive allantoic arteries; and secondarily, a capillary plexus in the somatopleure of the caudal end of the embr>'o, in which an umbilical artery develops. The umbilical artery joins the original allantoic artery in the fused area of allantois, somatopleure, and amnion (see text-fig. 6), and then the jjrimitive allantoic arteries from the aorta become reduced again to a capillary jilexus. Thus the allantois has a double arterial suj^ply and a double venous drainage, the former in the wall of the gut and the latter in the somatopleure. The primitive allantoic arteries arise in a i)lexus of the splanchnopleure, and the corresponding venous return is through the subintestinal vein; the subintestinal vein anastomoses w'ith the allantoic veins, but the direct continuation of the allantoic veins is into the uml)ilical veins, which devcU)]) in the somatopleure. Finally the umbilical arteries develo]) in the somato])l('ur(', coimect with the allantoic arteries, and soon bring most of the blood to the allantois.


The fate of the submtestinal vein in the chick is very interesting. If an injected chick of the fourth and fifth daj^s be dissected so as to expose the caudal end of the gut and the straight posterior segment of the gut which leads up to the open bell of the yolk-sac, it will he seen that the entire wall of the gut is surrounded by a capillary plexus. At the caudal end of the gut and just cephalic to the stalk of the allantois the ventral vein has entirely disappeared in this capillary plexus, while farther forward it is still clear in the ventral wall of the gut, though freely connected with the plexus. It is clear also that this posterior segment of the gut is receiving new arterial and venous connections which grow in along the dorsal border at the cephalic end of the segment. The new artery is a branch of the omphalo-mesenteric artery given off just at the root of the yolk-sac; it extends caudalward along the dorsal wall of the gut and anastomoses with the aortic branches which are the forerunners of the inferior mesenteric arteries. The new veins are branches of the omphalo-mesenteric veins within the mesentery, the forerunners of the portal system. The entire subintestinal vein gradually disappears as a single channel by developing into the plexus of the wall of the gut. In this plexus it is clear that the direction of the flow of the blood in the wall of the gut is from the ventral toward the dorsal border, at ri;.';ht angles to the direction of the stream in the subintestinal vein.


It may seem curious that the pig should have a subintestinal artery in place of the well-established subintestinal vein of the chick. As has been shown, the subintestinal vein in the chick develops as a part of the process by which the primitive circulation of the yolk-sac, with arteries and veins as far apart as possible, becomes changed so that everj^ zone of the area vasculosa is invaded by veins. The pig of plate 5, figure 1, rejiresents the more primitive condition for comparison with Popoff's plate iv, in which the caudal part of the yolk-sac is still arterial.


The subintestinal artery of the pig can be seen in section in text-figure 5 from a pig of the same Utter as the one shown in figure 1, plate 1; it receives numerous ventral arteries from the aorta, as does the corresponding vein in the chick; but it joins the om])halo-mesenteric arteries at the point of loop of the mesenteric arteries instead of the veins. This same artery is still present in the pig measuring 9 or 10 mm. after the caudal flexure has formed, at which stage it is breaking up into the capillary plexus within the wall of the gut. By the time the pig is 15 to 17 mm. long there is a new longitudinal artery in the dorsal wall of the gut, extending from the superior mesenteric artery caudalward and anastomosing with all the ventral aortic branches which represent the interior mesenteric artery. At the same time the accompanying venous plexus from the omphalomesenteric vein extends along the dorsal border of the gut. As this new bloodsupply for the lower half of the intestine develops, the ventral vein of the earUer stages of the chick, or the ventral arterj' of the pig, becomes reduced to a part of the capillary plexus in the wall of the gut. It is interesting to note that in a pig of 9.5 mm. the ventral artery of the gut is also accompanied by a plexus of ventral veins, which correspond to the single ventral vein in the chick. Thus the difference in the two forms becomes readily understandable, for the invasion of that part of the gut by the vems is merely relatively later in the pig, and the veins are thus much more transitory.


Branches of the omphalo-mesenteric veins growing down the mesentery begin early in the pig. These are shown in plate 4, figure 3. They are not seen in plate 1, figure 1, because uninjected in the specimen. In other specimens from the same fitter there is a vein in the mesentery underneath the umljilical vein, as seen from the side, and joining the main omjihalo-mesenteric vein at the lower margin of the liver. These l^ranches are shown in plate 5, figure 2. The veins in the root of the mesentery anastomose with the mesial cardinal (subcardinal) vein as soon as it develops. This anastomosis was described by Hochstetter.


The ventral subintestinal artery here described was discovered by Ravn in 1894 in the rat and mouse. The vessel was also described by Evans in the pig (Manual of Human Embr_yology, Keibel and INIall, foot-note 56 on page 656). Ravn's description can be readily followed in my i)late 5, figure 1, as he described the main omphalo-mesenteric artery arising in the caudal end of the embryo. Both Ravn and Evans describe this subintestinal arterj'^ as arising from the umbilical artery. My specimens, however, are from still earlier stages, and prove that this vessel arises, as does all the rest of the omphalo-mesenteric system, in the wall of the yolk-sac or gut; that it is a true vitelline vessel. Its anastomosis with the umbilical arteries in the somatopleure occurs later. Thus the subintestinal artery in the pig and the subintestinal vein in the chick are vitelline vessels. They disappear as single channels and help in the formation of the primitive plexus in the wall of the gut in connection with the changes by which the gut receives its permanent blood-supply and in connection with the gradual reduction of the yolk-sac.


The study of the ventral brandies of the aorta in the human embiyo is ba.sed on the work of Mall, who in 1891 published an account of a human embryo 7 mm. long, in which he described two main ventral branches, a coeliac axis and an omphalo-mesenteric artery, and a series of small ventral branches in the lumbar region, making a capillary network in the mesentery. He noted that the position of both the ca>liac axis and the omphalo-mesenteric arterj' was farther forward than in the adult, and analyzed all the available material in a study of the shifting of the arteries caudalward along the aorta. In this study he recorded human embryos with the coeliac axis opposite the first, second, fourth, and sixth dorsal nerves, as compared with the position in the adult opposite the twelfth nerve. In 1897 he made a further study of the ventral arteries, especially in a human embrj^o 2.1 mm. long. In this specimen he showed a series of ventral branches extending from the seventh somite to the caudal end of the aorta. These vessels he groujK'd together as the omi)halo-mesenteric arteries. In the reconstruction he showed that the u])per arteries tended to be ()])])osite the middle of the .somites rather than between the somites, as are the dorsal intersegmental vessels. In a second analysis of the ventral aortic branches he showed that there is a constant shifting of the coeliac axis and om))halo-mesenteric arteries caudalward. A doubl(> origin of the omi)halo-mesenteric arteries in one emljryo suggested the method of the wandering of the vessels.


The same idea of the shifting of the arteries caudalward was further developed by Tandler in two papers in 1904 and by Broman in 1908. These workers extended their observations over a long series of embryos, Broman giving a study of 41 specimens. In one of the youngest specimens in his series, an embryo measuring 3.4 mm., the uj^pcr ventral branch was between the sixth and seventh interspaces. He described the branches as tending to occur between the interspaces, there being two or three to a somite. He found that the cceUac axis and the superior mesenteric artery are not segmental vessels (that is, opposite the interspaces), while the inferior mesenteric artery is sometimes opposite and sometimes between the somites. Broman gives an analysis of the hterature and an extensive discussion of the methods by which the shifting of the ventral arteries may take place.


In the human embrj-o ventral branches of the aorta have been described from about the seventh segment caudalward. In the pig these ventral branches are very numerous — approximately one to a segment and one to an interspace. They are originally of uniform size and about equidistant apart. They unite into an extensive plexus of larger vessels in the more cephalic region and into a long artery in the caudal region. It is easy to follow the method of the shifting of arteries from such a primitive pattern; that is, any of the vessels of the original system could easily- enlarge and the blood-stream be increased or decreased according to the development of the region of the organ supplied. The entire wandering of the arteries can be understood without presupposing the development of any new vessels, but rather through the shunting of the blood through different channels already present in response to the varying de\elopment of the parts supplied b}^ these arteries. Moreover, it is plain that the point brought out by Evans is of importance — nameh^, that the so-called wandering of arteries takes place while the vessels have the structure of capillaries; that is, while their wall consists of endothelium alone. From the position of the primitive ventral arteries it is also easily seen that there might be variations as to whether the ultimate ventral arteries of the older embryo came oi)posite an interspace on the same level as the dorsal arteries or opposite a somite.


The Umbilical Vessels

My series is not very complete in regard to the umbilical veins, but it shows a few interesting points. In plate 4, figure 3, the relation of the somatopleure to the fold of the amnion is very plain. In the somatoi)leure is the beginning of a capillary plexus representing the umbilical veins. In plate 5, figure 1, the umbilical veins are not injected, but they are well shown in plate 1, figiire 1, in which it is clear that the retiu-n flow of the blood from the caudal end of the embryo is in part through the subintestinal artery in the splanchnopleure and in part through the umbihcal veins in the somatopleure. At the stage of plate 1, figure 1, the umbilical veins nave estabhshed their connections with the Uver, though they still connect with the sinus venosus. In figure 1 of plate 1 and figure 2 of plate 5 it is clear that cephalic to the duct of C'uvier there is also a capillary plexus in exactly tlie same position as tlie umbilical veins; that is, in the 8omato])leure. In the specimen of plate 1, figure 1, this venous ca]iillary plexus connects with a tiny lateral aortic branch shown just oj^posite the zone of the second aortic arch. This lateral artery is not the second aortic arch, which arises from the ventral rather than from the lateral surface of the aorta. P>om this tiny lateral artery a straggling chain of capillaries is injected within the somatopleure, out over the heart, and down to the duct of Cuvier; they are omitted in the drawing. It is clear that they are vessels for the body-wall analogous to the vessels which drain into the umbilical veins; but they are cephalic to the duct of Cuvier. The venous end of the plexus is injected in plate 5, figure 2. It was shown in the chick that the corresponding vessel of the somatopleure over the heart develops very early. In later stages these vessels in the somatopleure over the heart anastomose freely with a plexus of capillaries lateral to the occipital myotomes, as shown in textfigure 5 in my article on the Origin and Development of the Lymphatic System, 1913.


Neural Branches Of The Aorta And The Primary Head-Vein

In connection with the neural vessels, I have no specimens of embryo pigs corresponding to the chicks of 6 somites in which to trace their beginning. I have one litter of very young pigs, measuring 3 mm., in which the heart and aorta are present; the neural folds are open at the cephalic end, and I can find no angioblasts along the clo.sed hindbrain.


At the stage of figure 3, plate 4, the vessels to the forebrain can be injected; and the vessel of the hindbrain must be present, for it is seen in a human embryo of the same stage of development. Opposite the third and fourth somites the lateral plexus of the neural tube has been injected from the aorta. I found only one specimen of the htter of figure 3, plate 4; but the fact that the posterior cardinal vein is almost completely injected indicates that the anterior cardinal vein is present and that it connects with the deep vein of the hindbrain. At the stage of plate 5, figure 1, the vessels of the head are in about the stage of development of those of plate 1, figure 1, as is proved by the injections of the same litter. In one specimen of the same litter as plate 5, figure 1, the anastomosis of the capillaries around the optic stalk is complete, just as was shown by Evans for the later stage of three aortic arches in his figure 395 (Keibel and Mall, Manual of Human Embryology', II, p. 579).


The best view of the earl>- neural \essels in my series is given in plate 1, figure 1. In order to analyze the relations of the vessels of the head, I have used gray to indicate all of the capillaries which are true neural vessels, in the sense of lying close to the wall of the neural tube and giving rise to the vessels of the subsequent pia mater.


As can be seen in plate 1, figure 1, the deep capillary plexus of the forebrain and midbrain is covering the wall of the brain, and the form of this i)lexus indicates the form of the brain. The vascular arch which surrounds the large peduncle of tlu; oi)tic vesicle (see Evans's figure 395) is incom])letely injected in plate 1, ligurc 1. It shows the relative size of the ojjtic vesicle and the forebrain at this stage. The side of the thalamus and the midbrain is nearly covered by a plexus extending toward the dorsal wall of the neural tube. Along the cephahc part of the hindbrain is a wide vessel connected with the aorta by two arteries. It already shows sprouts along its dorsal border, two of which bound the otic vesicle. This deep single channel becomes a plexus along the side of the neural tube at a point just in front of the first myotome. This is the point where the cephalic end of the anterior cardinal vein joins the neural vessels, and, in terms of the neural tube, it is at the cephalic end of the origin of the roots of the vagus nerve. The transverse vessel of the first interspace which is so prominent in the chick is but a small vein in the pig Uke the other intersegmental veins, and does not become an important vessel, as in the chick. As is well known, the upper myotomes are occipital mj^otomes, so that it is clear that the point of transition between the deep vessel of the hindbrain and the primitive plexus, as shown in plate 1, figure 1, is not between the hindbrain and cord, but is near the upper part of the medulla. The lateral plexus along the cord is injected in the specimen of plate 1, figure 1, down to the fourteenth somite, which is opposite the lowest transver.se artery injected, and the .spinal arteries are injected down to the twentieth interspace. These lower vessels are omitted in the drawing.


In plate 1, figure 1, can be traced very clearly the origin of the cephalic part of the primary head- vein; that is, the primitive cerebral vein. Extending from the groove between the telencephalon and the diencephalon (as Evans showed in his figure 395 in 1912), is a superficial capillary plexus, indicated in blue, which receives its blood from the deep plexus of the forebrain and midbrain and drains into the deep vessel of the hindbrain. In this plexus will develop the primitive cerebral vein; at this stage it is entirely a plexus without any definite longitudinal channels. The specimen is just at the stage of the second vascular arch, which is probably present and uninjected, as shown in Evans's figure 394 from an earUer stage. Opposite the lower end of the primitive vessel of the hindbrain is a plexus of exceedingly tiny vessels spanning the gap between the deep vessel of the hindbrain and the anterior cardinal vein on the one hand, and reaching toward the second aortic arch on the other. These tinj' capillaries form the origin of the lateral vein of the region, that is, the middle segment of the primary head-vein, just as has been shown for the chick. This plexus will span the gap between the second and third aortic arches as they form, and the cephahc end of the primary head-vein, until there is a double vascular channel from the head, as shown on plate 4, figure 1.


Figure 1 of plate 4 is from a specimen of the same Utter as that of figure 2 of plate 5, and is given because of the extravasation in the head region in the latter figure. At the stage of three aortic arches the primary head-vein is complete. The primitive veins which pass ventral to the eye are not injected in the specimen of plate 4, figure 1, except just where they join the primary head- vein in front of the ganglion of the trigeminus. The primary head-vem starts opposite the thalamus and extends in a double curve down to the anterior cardinal vein. It lies mesial to the Gasserian ganglion and lateral to the otic vesicle. The pattern of the deep and the sujierficial vessels in plate 4, figure 1, deserves careful study. The place of origin of the roots of the Gasserian gangHon is marked by a plexus of the deep vessels which are growing around it, leaving a non-^'ascular area where the nerves emerge. The deep plexus is also forming a dorsal arch around the otic vesicle which now Ues between the deep and the superficial vessels. The pattern of the vessels also indicates the position of the acoustic complex of gangha and the glosso-pharyngeal ganglion, both of which lie between the deep capillaries and the superficial veins, one cephahc to the otic vesicle and the other just caudal to it. It is clear that the relations of the primar}- head-vein to the Gasserian ganglion and to the acoustic complex are the same in the pig as in the chick, and are due to the fact that this vessel forms while the gangUa are attached to the skin in their respective placodes. The primary head-vein develops mesial to the placode of the Gasserian ganglion, but curves dorsal ward opposite the acoustic gangha and opposite the ganglion of the glosso-pharyngeus. Sections of an embryo slightly older than that of i)late 4, figure 1, cut so that a long stretch of the primitive head-vein is included in one section, show that the lateral border of the acoustic ganglion is in a straight line with the mesial border of the Gasserian ganglion, so that the superficial vein takes the shortest course in passing mesial to the ganglion of the trigeminus and lateral to the ganglia of the acoustic complex.


The relations of the branches of the vena capitis i)rima are very important at the stage of plate 4, figure 1. The branches from the aortic arches are not injected, nor are the primitive maxillary veins. The lateral veins from the cerebrum have hardly begun. The superficial veins opposite the midbrain have a very characteristic pattern; they are, as it were, creeping along on the deep plexus toward the mid-dorsal line. It will be noted that the deep plexus itself has not yet reached the mid-dorsal line at this stage, but it is in advance of the superficial veins. This gradual extension of the branches of the vena capitis prima to the mid-dorsal line characterizes the branches of this vein over the entire brain. When the superficial veins meet in the mid-dorsal line they will give rise to all of the sinuses and veins of this line, as has been shown by Mall and Strecter.


Opposite the hindbrain the branches of the vena capitis pruna have the same fundamental relation to the deep plexus. It is true that caudal to the otic capsule there are a few veins from the deep plexus draining into the ventral border of the vena cajntis i^rima, which may be forerunners of the small \entral veins of the medulla in the adult, but almost all of the veins of the hindbrain drain into the dorsal border of the primary head-vein. These veins have the same characteristics as the rest of the neural veins; that is, they gradually creep dorsalward on the deep plexus. Over the hindbrain, however, the i)attern of the veins is not as simple as over the midbrain, because here they are iirofoundly affected by the ganglia of the hindl)rain and by the otic cai)sule.


It has been shown that the deep i)lexus makes an arch of capillaries around the roots of the nerves, as seen around the root of the trigeminus in i)late 4, figure 1. The superhcial veins also curve around the roots of the nerves. Their beginning is shown in plate 4, figure 1. Here it is clear that branches of the primary head-vein are tapping the deep neural plexus around the root of the trigeminus.


I.ateral to the acoustic complex and to the ganglion of the glosso-pharyngeus, the branches of the primary head-vein make a very extensive plexus. The superficial venous arch around the otic vesicle is just beginning in plate 4, figure 1. The veins around the trigeminus and around the otic capsule are exceedingly important, because of their ultimate relations to the basal sinuses of the dura. These vessels are shown in plate 4, figure 1, at the stage when they are scarcely more than capillary sprouts. They will be traced farther in the next figure.


The specimen of plate 7, from a pig which measures 6..5 mm. in oil, is given to emphasize the fate of the primitive vein of the hindbrain, to bring out the ventral artery that now extends the full length of the nervous system from the base of the optic cup to the tip of the tail, and to show the characteristic relations of the veins to the neural tube and its gangUa.


The injection of the specimen of plate 4, figure 1, did not bring out the ascending neural arteries as did a corresponding injection of the chick (plate 6), but the specimen of plate 7 shows that there is now a longitudinal arterj- which extends from the primary aortic branch to the brain opposite the subthalamus, along the ventral or ventro-latcral border of the neural tube to its caudal tip. This artery is an anastomosis between all of the neural arteries, both cerebral and spinal. As can be seen in plate 7, the carotid artery leads to an arterial plexus which covers the lateral surface of the subthalamus and gives rise to a cerebral artery passing dorsal to the eye. The plexus on the subthalamus anastomoses with the plexus of the opjiosite side in the mid- ventral line; it is tajijied by a vein leading to the primary head-vein just cephalic to the maxillary vein. Opposite the groove between the thalamus and the midbrain the two jjlexuses on either side of the subthalamus gives rise to a single ventral artery which curves along the ventral border of the neural tube down to the level of the third occipital interspace, where the single median artery becomes an arterial plexus. From this point to the caudal end of the spinal cord there is a double line of capillaries, such as was shown by Evans in his figure 440 (1912). As Evans showed, this double capillary chain will give rise to the anterior spinal arter^^ The importance of this longitudinal neural arter}-, which gives rise to the circle of Willis, the basilar artery, and the anterior spinal artery, is obvious. The anastomosis of the arterial plexus of the subthalamus and the ventral surface of the cerebrum vnth the corresponding jjlexus of the other side across the mid-ventral line accounts for the anterior communicating artery of the circle of Willis. At the stage of plate 7 the longitudinal neural artery is supplied by the two carotid arteries, by direct arteries opposite the hindbrain, of which two are shown on the right side of plate 7, and by all the intersegmental arteries on either side. This artery is not supplied as yet by the vertebral a/teries, which form later as an anastomosis between the upper intersegmental arteries.


The arterial plexus over the subthalamus leads into a finely meshed plexus which covers the entire cerebrum except a small area in the mid-dorsal line near the thalamus. This i)lexus is not shown in the drawing, but it has the same character as the i)lexus over the midbrain. The cerel)ral plexus comijletely surrounds the optic stalk; in this plexus the only vessel larger than the rest is the cerebral artery, which is seen dorsal to the eye in plate 7. The longitudinal neural artery along the ventral border of the midbrain and the hindbrain gives off a series of nearly ecjual, regular, small arteries which lead into the cajiillary plexus on either side of the neural tube.


The capillary jilexus on the neural tube is very characteristic. As has been said, it is finely meshed over the cerebrum, the thalamus, and the midbrain; it is more coarselj' meshed over the hindbrain, where the plexus has developed later, especially around the roof of the fourth ventricle, which has not yet been invaded by the Aessels. The plexus on the hindbrain in jilate 7 demonstrates the fate of the primitive vessel of the hindbrain, the beginning of this plexus as coming from the jnimitive vessel of the hindbrain having been seen in the living chick. The primitive vessel of the hindbrain disappears only in giving rise to the capillary plexus of the hindbrain. If the pattern of the neural plexus in plate 7 is observed carefully it will be seen that there is just a suggestion of transverse lines in the plexus, indicating that the direction of the flow of the blood is from the ventral to the dorsal border of the neural tube. In this plexus will ultimately come transverse arteries. Opposite the first somite will be noticed the beginning of three layers of vessels, a deep layer of very fine capillaries, a second layer of larger vessels also shown in gray, and a third layer of more superficial veins. This is the very beginning of the next stage in the development of the neural vessels.


The most imi)ortaiit jioint about the form of the deep i)lexus on the neural tube is the way it conforms exactlj' to the neural tube and its nerves. Over the midbrain the plexus is very uniform, but over the hindbrain the character of the plexus indicates very clearly the position of the nerves. At the stage of plate 7 there are bare spots, that is, places with no blood-vessels, on the hindbrain corresponding to each nerve root; in later stages the vessels penetrate between the small bundles of the fibers of each root and then an injection of the deep plexus does not show the position of the nerves so clearly. As seen in jilate 7, the positions of the roots of the trigeminus nerve and of the acoustic grou]) of nerves are very clear. The otic capsule now lies just lateral to the deep cai)illarv ])lexus. and thus its position is indicated only by the sui)erficial veins. ()p])osite the ganglion of the glosso-pharyngeus is a bare sjiot in the deep i)lexus, which is nearly hidden by a very extensive group of superficial veins. The position of the roots of the vagus and the spinal accessory roots along the line of the posterior cerebral vein is very important. It is clear that the deep i)lexus outlines this long line of nerve roots, and the same is true along the more ventral line of the medulla, wluM-e the pattern of the ve.s.sels indicates the position of the roots of the hypoglossal nerves.


Along the spinal cord the pattern of the capillary plexus shows the position of the ventral nerve roots in the same manner.


The veins which form the branches of the vena cajiitis puma must now be followed. The veins from the visceral arches, still largeh' in the form of capillaries, are completely injected in the specimen, but are indicated in the drawing only at the point where they join the middle segment of the primitive head-vein. In plate 7 the cerebral and the cardinal segments of the vena capitis prima are shown in plastic form, but the middle segment is shown merely in outline in order to make more plain the relations of the neural artery beneath. Beginning with the maxillary vein, the entire maxilla is filled with a capillary plexus which leads to the maxillary vein. This capillary plexus anastomoses with the plexus of the mandibular arch. Besides these capillaries the vein receives a large group of tiny superficial veins which arise in the deep plexus that covers the entire olfactory area of the cerebrum, together with primiti\'e ophthalmic veins which arise in the marginal vein of the optic cuji, as in plate 6. One of these subophthalmic veins runs in the groove of the optic stalk. These cerebral veins from the rhinencephalon and from the inferior part of the eye are very important in the earh' drainage of the brain, but it is well known that the main permanent ophthalmic veins develop dorsal to the eye.


In the zone dorsal to the eye at the stage of plate 7 is a group of tiny superficial veins opposite the cerebrum which are like the small veins over the midbrain. They were omitted in plate 7, but are adequately shown for the chick in plate 6, and they are aUke in both forms. These are the primitive cerebral veins. Over the midbrain the veins are characteristic. It is plain that thej- are lifted off from the surface of the neural tube, that they are all superficial to the deep plexus; they spread out like a fan from the primary head-vein and clearly extend along the deep plexus, which they tap at their tips, and approach the mid-dorsal line.


Opposite the hindbrain the veins are exceedingly interesting; they follow exactly the same general course of development as the rest of the neural veins; that is, they he superficial to the deep plexus, are transverse to the long axis of the neural tube, and graduallj- extend toward the mid-dorsal line, constantly tapping the deep plexus at their tips. On the other hand, they are profoundly modified in their development of the gangha of the hindbrain and by the otic vesicle, so that their pattern is much more complex than the pattern of those opposite the midbrain.


The vessels around the ganglion of the trigeminus deserve careful study. At this stage the entire lateral surface of the Gasserian ganglion is covered by a capillary plexus which was omitted in the drawing. This capillary plexus extends along the second and third divisions of the nerve and becomes continuous with the capillary plexus of the maxillary and mandibular processes. Besides this sheet of capilh*ries which covers the lateral surface of the ganglion, there are two transverse veins above and below the ganglion which outline the root of the trigeminus nerve. These veins are very characteristic, and mark the position of the Gasserian ganglion in any injected specimen up to the stage measuring 20 mm., when the transformation of the veins into tlie (hu-al sinuses is well advanced, as can be seen in Streeter's figure 3 (Amer. Journ. of Anat., 1915, XMII, page 150). The su])erficial vessels around the ganglia of tlie eighth nerve are still in the form of capillaries in jjlate 7. Opposite the otic vesicle the deep jilexus has completely covered the surface of the hindbrain ; there are a few superficial veins across the lateral surface of the vesicle, which are shown cut off in the drawing close to the primary head-vein. Two of these transverse veins make a border for the otic vesicle exactly as do those above and below the Gasserian ganglion. In other words, the veins of the hindbrain can be most simply described as a series of transverse vessels, some of which are forced to curve by the Gasserian ganglion and the otic vesicle. Opposite the ganglion of the glossol^harj'ngeal nerve is a series of transverse veins draining into the primary head-vein.


The veins opposite the vagus nerve are also very interesting. It is clear that the largest vein of the medulla at this stage is one which in a general way follows the roots of the vagus nerve. This vein was called the jiosterior cerebral vein by Mall. In general, the place where it joins the vena capitis prima marks the cephalic end of the anterior cardinal vein; it may be a single vein at its roots or a group of veins. In the pig the vagus nerve curves around its cephaUc border, l)assing in the angle between this vein and the primary head-vein. Some of the injections show the nerve passing through a venous loop in this angle. Stracher descri})es the vagus nerve just caudal to the vein in the chick. The relations of the vagus nerve to the i>rimarv head-v(>in formed the basis of Kastchenko's original study of the primitive veins of the head.


As will be seen in plate 7, the main vein of the medulla primaril}' follows the course of the roots of the vagus nerve. It arches caudalward along the dorsolateral surface of the medulla in the line of the spinal accessory nerve and roots of the vagus. The line of the vein on the medulla can be well seen by following the vagus roots in Streeter's plate 11 (Amer. Jour, of Anat., 1905, IV).


While it is clear that this vein and its tributaries originally follow the path of the vagus nerve, if its development is followed it will be seen that it becomes a very important vein of the embryo, not even limited to the diainage of the neural tube. At the stage of plate 7 it anastomoses witli the lateral venous plexus of the lower medulla, and the first and second occijiital veins are correspondingly small. Sub.seciuently it gives rise to an extensive group of dorsal b.ranches that grow over the caudal ])art of the roof of the fourth ventricle and largely drain the devclojiing choroid ])lexus. The posterior cerebral vein next develojis an exceedingly interesting relation to the \ascular system of the occipital myotomes. This relation was illustrated in two figures from injected embryo jiigs in my article on the origin and development of the lymphatic system (1913, figs. 4 and 5). Opjjosite the entire zone of the myotomes a jilexus of cai)illaries develops, forming the third vascular .sheet of this region. Primarily there is a i)lexus of c:ii)illaries on the surface of the neural tube; secondly, a more lateral i)lexus of cai)iilari('s and veins especially relatt'd to the gangli;i: tiiirdly, this sheet of caijillarics lateral to the myotomes. Oj)posite the occipital myotomes the capillary plexus drains, by a series of veins on the one hand into the main vein of the medulla, on the other hand into the anterior cardinal vein. The history of the neural branches of this vein of the medulla involves the entire subject of the circulation of the medulla. The relation of the branches from the occipital myotomes involves the subject of the development of the external jugular vein and its branches. The main stem of the vein was shown by Mall, in 1905. to become a part of the great transverse sinus. For this vein I am using the term primitive 'posterior cerebral vein. It might also be termed the primitive vein of the medulla.


The stage of plate 7 shows the beginning of the veins of the hindbrain. It will be seen that the primitive branches of the primary' head-vein draining the hindbrain are greatly modified by the ganglia of the hindbrain and the otic capsule. Opposite the midbrain these veins are regular and nearly eciuidistant; opposite the hindbrain they are grouped according to the gangha. Of these veins of the hindbrain, the group caudal to the Oasserian ganglion and the stem of the posterior cerebral vein bear the most important relations to the future cerebral sinuses at the base of the brain.


In this account of the early blood-vessels of the neural tube three facts have been brought out which are essential to an understanding of the development of the neural vessels. First, there forms a ventral neural artery, originalh' paired, which extends along the ventral surface of the entire neural tube from the base of the optic cup to the caudal end of the spinal cord, which is an anastomosis of all the direct neural arteries from the aorta; second, this artery leads to a capillarj^ plexus which completely invests the neural tube and all its ganglia; third, the primary veins of the neural tube are all transverse vessels superficial to this ])rimary plexus, and they gradually extend toward the mid-dorsal line and are profoundly modified by the ganglia, both cerebral and spinal. All of the veins of the brain drain into the primary head-vein. As has been shown by ^lall and Streeter, the only segment of the vena capitis prima which remains as a part of the dural sinuses becomes the cavernous sinus, which is that portion of the primary head-vein medial to the Gasserian ganglion. All other dural sinuses develop from the branches of the vena capitis prima.


It has been shown that the middle segment of the vena capitis prima develops in the pig, as in the chick, as a chain of capillaries between the aortic arches and the anterior cardinal vein; it becomes very large, because it makes a more direct outlet for the primitive cerebral vein. The vena capitis prima develops from three segments and is the first true vein for the head; the primitive vessel of the hindbrain serves temporarily as a vein for the brain and then gives rise to the capillary plexus of the upper part of the hindbrain.


Cardinal Veins In The Pig

It was .shown in the chick that the carcUnal veins begin from dorsal diverticula of the aorta which project into the interspaces and dilate just opposite the dorsal border of the nephrotome. In the line of the nephrotome these separate dilatations become connected, making a common cardinal vein which, at the stage of 12 somites, is oi)posite every interspace. I have not the corresponding early stages of the cardinal veins in the pig. In my earliest stage in the pig, the cardinal veins are related to the aorta and to the spinal veins, as is shown for the cliick in the section on plate 3, figure 2; that is, there are direct spinal arteries from the aorta to the cord and spinal \'eins leading to the cardinal vein. At the stage of plate 4, figure 3, the posterior cardinal vein is injected, extending from the zone of the ninth intersegmental artery almost to the duct of Cuvier. The anterior cardinal vein is not injected, but must be present in the specimen. The pig embryo showai in figure 1, plate 1, gives the best view of the cardinal veins in my series. In this specimen it is clear that the anterior cardinal vein joins the neural plexus cephalic to the first somite, so that the vein of the first interspace which was so important in the chick is hke all of the rest of the intersegmental veins in the pig. Opposite the first nine somites in the pig, as shown in plate 1, figure 1 , the cardinal veins appear to be an accomjianying vein to the aorta. Just below^ the ninth intersegmental artery in the pig there are the lateral arteries to the nephrotomes, and over all of the rest of the course of the posterior cardinal veins the lateral cardinal vein must also be considered. Opposite the first nine somites I have not been able to find any direct connections between the cardinal veins and the aorta, such as were shown for the earlier stages in the chick. In other words, the cardinal veins are w^ell formed rather than just beginning in all of my specimens. One embryo, of the same litter as the one in plate 1, figure 1, showed some tiny sprouts of the anterior cardinal vein opposite the second somite extending toward the aorta; sections, however, did not demonstrate any connections, and I could not prove that they were not the beginning of tiny veins that soon drain the pharynx.


The series of the ])ig embryos also does not show the origin of the duct of CAivier, but the fact that it is made up of an extensive plexus is well shown in ])late 1, figure 1, as well as its relation to the umbilical veins. Below the zone of the ninth somite the cardinal veins will be considered with relation to the vessels of the pronephros.


Nephritic Vessels In The Pig

The nephritic tubules in the pig receive an early and characteristic blootlsui)i)ly. For the limit for the chick between the pronephros and the mesonephros I have followed Lilly, who regards the tubules as belonging to the pronephros down to the fifteenth or sixteenth somite (page 190). For the pig I have arbitrarily followed Felix's estimation for the hiunan embryo (1912, ]>age 7(52). He places the limit of the ])rone])liros at the fourteenth .somite. It will be .seen in figure 3, i)late 4, that just below the ninth inter.si'ginental artery a series of lateral arteries gives rise to a plexus which is ventral to the posterior cardinal vein, but which connects with it. The Wolffian duct intervenes between this plexus and the posterior cardinal vein. In figure 1 of plate 5, and figure 1 of plate 1, are given very characteristic views of these lateral arteries to the pronephros, and in figure 5, plate 3, is shown a ventral view of the arteries of the jjronephros from a specimen of the same litter as that of ])late 5, figure 1, but a httle farther developed. Figures 1 of plate o and 1 of plate 1 show a series of tiny lateral arteries beginning just below the ninth dorsal segmental artery. These arteries are about four to a somite, corresponding to the number of the nephritic tubules, and are connected by a tiny longitudinal artery close to the aorta and by a tiny lateral vein. In textfigure 4 is shown a section from a specimen of the litter of plate 1, figure 1, passing through about the fourteenth somite, showing an injection of one of these lateral arteries. Its exact position with reference to the developing tubule is, I think, important. This is most clearly recognized from the diagram given by FeUx of the development of the nephritic tubules (fig. 561, Keibel and Mall, Manual of Human Embryology, page 804) . The stage corresponds with diagram d of Felix's figure, and the artery passes directh' across the curved bowl which makes the neck of the future Malpighian corpuscle. This is the earliest stage of the vessels of the nephritic tubules I have injected. As is seen in text-figure 4, the lateral vein, the vena cardinalis lateralis, lies ventral to the Wolffian duct, while the vena cardinalis posterior lies directly dorsal to the duct. The posterior cardinal vein is plainly shown in text-figure 4, but was not injected so far caudalward in any of my series.


Text-figure 5 gives a -very interesting section from the same series as textfigure 4. The level of the section is shown in plate 1, figure 1; it is about halfway betw-een the level of the low'est transverse artery injected and the allantoic arteries. At this level the nephritic tubule is in the stage of Fehx's figure 5616, consisting of a Wolffian duct and a mass of nephrogenic epithelium. Here, instead of an artery which caa be injected, the section shows a chain of angioblasts running ventral to the nephritic tissue to the lateral cardinal vein, and other sections show similar chains of angioblasts connecting the aorta and the posterior cardinal vein.


Fig. 4. — Transverse section : , . 23 somites, passing through one of the lateral arteries ui ihe pruuephros. The section is from a specimen of the same litter as the one shown on plate 1. figure 1, and from the same series as figures 5 and 6. The level of the section is shown by a line on plate I, figure I. The section is 20 II thick and is stained with hematoxylin and counterstained with orange G, eosin, and aurantia. XII.t. ^. om., a. omphalo-mesenterica; A. pr., a. of the pronephros which was injected from the aorta; V. om., v. omphalo-mesenterica; V. c. I., v. cardinalis lateralis; y. c. p., v. cardinalis posterior: \V. d.. Wolffian duct.



Fig. 5. — Traosverse section of an embn.. cm ..1 — > ^imites, passing through one of the lower myotomes and the mesonephros, to show the position of the subintestinal artery and a chain of angioblasts which will form an artery of the mesonephros. The section is from a specimen of the same litter as the one shown on plate 1, figure 1, and from the same scries :ts figures 4 and 6. The level of the section is shown by a line on plate 1, figure 1. The section is 20 ii thick and is stained with hematoxylin and counterstaincd with orange G, cosin, and aurantia. XUo. A. mes., a chain of angioblasts which connect the aorta with the v. cardinalis lateralis and which will form an artery of the mesencephalon but were not injected because they are still solid; A. si., a. subintestinalis; V. c. I., v. cardinalis lateralis; V. c. p., v. cardinalis posterior.



These angioblasts are mesial to the nephritic tissue. This section is, 1 think, similar to the .^^ection in Evans's hgurt^ 416 from a human embryo of the same stage, namely, with 23 somites, which shows the jiosterior cardinal ^•ein dorsal to the Wolffian duct Mild the lateral cardinal vein ventral to the duct.


A comparison of text-figures 4 and 5 seems to me to indicate that the primary arteries of the nejihrogenic tissue are ventral to the nephrotome, but when the tubules are farther developed the artery crosses the neck of the tubule; in other words, the tubules grow ventral to the arteries.


The study of the eml^ryo pig at the stage of 23 somites (as shown in plate 1, figure 1, and in the sections of text-figures 4 and 5), seems to me to indicate that the posterior and lateral cardinal vessels extend caudalward in connection with chains of angioblasts from the aorta which pass dorsal and ventral to the nephritic tulniles in lines which are very plain in figure 5.


In figure 5 of plate 3 is shown a ventral view of the pronephritic vessels in a pig of 20 somites, in which it is clear that there is a tendency toward a grouping of the transverse arteries of the prone])hritic tubules around segment ;il lateral arteries. For example, Ijetween the ninth and tenth sjiinal arteries there is one lateral artery giving off four branches ; between the tenth and eleventh spinal arteries are two lateral arteries from the aorta, \\ith three transverse branches.


The longitudinal artery shown in plate 3, figure o, persists for some time in the pig and connects the glonuiiilar arterites even after tlie arterial tutts of the glomeruli are well formed. As seen in plate 3. ligure ">, tiie tran.>*verse arteries lead directly to a lateral vein, which in turn connects wit li the posterior cardinal vein. Moreover, as is shown opposite the tenth somite, the posterior cardinal v(>in lias many direct connections with the aorta.


Fig. 6. — Transverse section of an embryo ])!« of 23 somites, passing through the allantoic arteries to show that the primitive allantoic arteries are in the splanchnopleurc. Tiie section is from a specimen of the same litter as the one shown on plate 1, figure 1, and from the same scries as figures 4 and 5. The level of the figure is shown by a line on plate 1 . figure 1 . The section is 20 /i thick and is stained with hematoxylin and counterstaincd with orange CI, eosin, and nurantia. X-'J.'l. A- al.. artery of the allantois; C, coclom: V. al., vein of the allantois in the zone where the splanchnopleurc, the somntopleure, and the amnion ari-' fused.


The next stage in the development of the circulation of the Wolffian bodies is the formation of the mesial cardinal vein. I have illustrated the position of this vein in two sections, one from an injected pig embryo of 30 somites, measuring 7 mm. before the caudal flexure has formed, and the other from an injected chick of 69 hours' incubation (text-figs. 7 and 8). The mesial cardinal vein lies ventral to the nephritic arteries, close to the aorta, in the angle between the root of the mesentery and the Wolffian ridge. The course of the mesial cardinal vein can be readily imagined in plate 3, figure 5, wherein it is noted that opposite the tenth and eleventh somites the posterior cardinal vein is in the form of a plexus, dorsal



Fig. 7. — Transver3e section of an injected embrj-o pig of 30 somites, to show a typical cross-section of the vessels of the pronephros of the pig after the v. cardinalis mcsialis has formed — that is, to show the pronephros with a central artery and three peripheral veins. The embr>-o measured 7 mm. after fixation and dehydration; it had no caudal flexure and was a little farther developed than the one on plate 5, fifjure 2. All the vessels shown were injected. The arteries are represented in black, the veins in white. The section is 50 n thick and is unstained. X53. A. pr., artery of the pronephros which gives off capillaries to the tubules and extends to the v. cardinaUs lateralis; V. c. I., v. cardinalis lateralis; V. c. m., v. cardinalis mesialis; V. c. p., v. cardinalis posterior; W. d.. Wolffian duct.




Fig. S. — Transverse section of an injected chick of 35 somites, after 69 hours of incubation, passing through the fifteenth somite. The section shows a tj-pical crosssection of the vessels of the pronephros in the chick after the V. cardinalis mesialis has formed — that is, it shows the pronephros with a central artery and three peripheral veins. All the vessels were injected. The aorta is shown with a black rim, the artery is black, and the veins are white. The section is .50 /t thick and is unstained. Xo3. -4. p., artery of the pronephros; V". c. /., v. cardinalis lateralis; V. c. m., v. cardinalis mesialis; V. c. p., v. cardinalis posterior; W. d.. Wolffian duct.


to the nephritic tubules (text-fig. 7). At the stage of 30 somites a vein from this plexus passes ventral to the nephritic artery opposite the eleventh somite and grows caudalward just ventral to the ncplxritic arteries, between the aorta and the longitudinal artery of plate 3, figure 5. This is the medial cardinal vein, the subcardinal vein of F. T. Lewis. There is thus formed the primitive pattern of the circulation of the Wolffian body, as shown in text-figures 7 and 8, consisting of a central artery and three longitudinal superficial veins — the posterior cardinal vein dorsal to the Wolffian duct, the lateral cardinal vein just ventral to the duct, and finally the mesial carcUnal vein near t\w root of the mesentery. The mesial cardinal forms the comiection with the vessels of the liver and (as shown by Hochstetter) also anastomoses with branches of the omphalo-mesenteric vein along the mesentery.


I emphasize the lateral cardinal vein because it has not been adequately recognized in the literature. In the pig it is very obvious in total preparations, such as are shown in i)late 3, fip;ure 5. It develops early and is very straifi;lit. In the chick it is not straight and therefore is much less striking in total jm'parations. Its primary connections with the i)osterior cardinal vein are lateral to the Wolffian duct, as seen in text-figure o for the pig. That this is also true for the chick is shown by Graefe's figure 6 (1900), which shows the pronejihros of the chick at the stage of 2 days and 15 hours. Later, in both the pig and chick, these two veins are connected by branches which are mesial to the duct, as shown in text-figures 7 and 8.


The failure to take into account the lateral cardinal veins has led to some confusion in the Uterature; for exami)le, in the studj^ of the pronephros, Graefe (in his figure 11) has labeled the lateral vein close to the Wolffian duct the subcardinal, while in figure 13 he has labeled the true subcardinal vein ventral to the nei)hritic artery the subcardinal, but has not labeled the lateral vein at all, though it is shown in the section.


In Keibel and Mali's Embryology, Felix gives some extremely interesting sections from the R. Meyer human embryo No. 300. This embryo had 23 somites and was 2.5 mm. long. It is to be compared with my plate 1, figure 1. In figure 532 o Felix shows solid angioblasts, both dorsal and ventral to the Wolffian duct; he does not label the dorsal angioblasts which rej^resent the posterior cardinal vein, but on the other hand calls the ventral angioblasts the posterior cardinal vein. Again, in figure 559 he calls angioblasts which are ventral to the duct the posterior cardinal vein. These sections show that in the human embryo there are angioblasts both dorsal and ventral to the duct and bring out the value of the two names for the veins, the posterior and lateral cardinal veins. They also showthat the posterior and lateral cardinal veins extend as solid angioblasts and so l)ring u]) the question as to whether these veins may not differentiate as chains of angio])lasts connected with the aorta by chains of angioblasts.


Conclusion

In this study it seems clear to me that the chick afifords verj^ valuable material for the study of the most fundamental jioint in connection with the vascular system that is still at issue, namely, how long in the Hfe of the embryo do new angioblasts continue to differentiate from nu>senchyme antl join tlie blood-vessels? The answer to this ((uestion involves more extensive observations on the living blastoderm than I have yet made. It has ])een shown that l>]o()d-vessels first arise not only in the membranes but also in the embryo by a differentiation of cells into angioblasts, by the process which His had described, and not from a dilatation of s])aces in the mesenchyme and a fiattening-out of cells to form their border.


It has l)een i)roved that the aorta at least in part dilTcnMitiates i» silii. lOvidence has been given tliat a part at least of the neural vessels and their connections with the aorta differentiate in sittt. On the other hand, tlie cardinal veins begin as a growth from the wall of the aorta. They are a longitudinal anastomosis between direct branches of the aorta. A more detailed study of the later stages of the cardinal veins is necessary tf) determine if any partof tliein differentiates in situ.


I tliink that it is important to emphasize the extent of the development of the blood-vessels both of the membranes and of the embryo at the time when the circulation begins. This has been done for the chick, and it would be of great value to obtain the same observations for the mammal.


This study gives a more complete account of the primitive vessel of the hindbrain than is to be found in the literature. I have followed its origin, its relations, and its fate. The fate of this vessel is a very important point. This primitive vessel of the hindbrain differentiates earlj'-, opposite the first part of the neural tube to develop. It has been shown why it remains so long a single channel, namely, because it serves temporarily as a vein for the forebrain and midbrain before it takes the characteristic form of a plexus like the other early vessels on the surface of the neural tube. As the vena capitis j^rima becomes complete, so that the blood of the forebrain and midbrain is shunted out of the primitive channel of the hindbrain, this channel receives new arterial connections and breaks down into the verj' important capillary plexus of the rhombencephalon.


It has been shown that the first true vein of the head, the vena capitis prima, as contrasted with veins which drain only the brain, develops in three segments. The anterior segment is a i)urely cerebral vein which drams the forebram and midbrain and originally emjities into the primitive vessel of the hindbrain; the posterior .segment is the anterior cardinal vein; the middle segment develops last, as a capillary chain between the capillaries of the maxillary, the mandibular and the other visceral arches, and the anterior cardinal vein. This middle segment anastomoses with the primitive cerebral vein from the forebrain and midbrain and forms, a much more direct and favoralile channel for draining the brain, and so rapidly supjjlants the more indirect channel along the hindbrain. It drains the other structures of the head in addition to the neural tube. The embrv'onic vein extending from the region of the thalamus to the duct of Cuvier is the first true vein of the head, in the sense of draining the entire head, that is, the brain and the visceral arches, and may thus be termed the vena capitis prima.


In connection vnth. the vascular system of the nervous system, it has been shown that the early pattern of the blood-vessels is very uniform for the entire tube. There is a capillary plexus which completely invests the tube and all of its ganglia. It is fed bj' bilateral longitudinal arteries, which form as an anastomosis between all of the neural arteries from the aorta and extends from the carotid arteries at the base of the optic stalk to the tip of the spinal cord. The bilateral character of these arteries persists onlj^ around the subthalamus, where the circle of ^^'illis is formed ; elsewhere the two arteries become a single ventral arterj'^ — the basilar artery and its primarj' continuation, the anterior spinal artery. I have thus brought out the origin and the significance of the basilar and anterior spinal arteries and have shown that they precede the vertebral arteries.


The first neural veins are all transverse superficial vessels, which tap the deep plexus and gradually extend dorsal ward on the deep plexus. They are profoundly modified by the eye, the ear, and by all the sensory gangUa. Opposite the brain they all drain into the primarj' head- vein; all the rest of the neural veins are intersegmental branches of the cardinal veins. It is thus clear that the general direction of the blood to the neural tube is from the ^■entral to the dorsal border and that the direction of the flow of blood from the neural tube is the reverse.


In connection with the pig it has been shown that the large branches of the aorta near the caudal end of the embryo are primary allantoic arteries which run in the splanchnopleure, and that the umbilical arteries in the somatopleure develop later and anastomose with the primary allantoic arteries, exactly as in the chick. I have also given an analysis of the subintestinal vein of the chick and of the corresponding artery in the i)ig, and have shown that the fact that the vessel is an artery in the pig means that the primitive type of circulation of the yolk-sac persists longer in that form than in the chick. It has been shown that both the primitive allantoic arteries and the subintestinal arteries arise in a capillary plexus and end in a capillary plexus, so that in the case of these two vessels the blood must pass through two cajiillary plexuses in its return to the heart.


The study of the circulation of earlj^ embryos by means of injecting living embr3'os and watching the flow of the ink in them or by watching the circulation of the blood in the living specimen brings out some remarkable changes in the direction of the circulation; for example, the change in the direction of the circulation in the vessels of the area vasculosa in the chick when the veins invade a plexus which had been arterial. Again, in connection with the development of the i^rimitive vessel of the hindbrain into a capillary jilexus, the direction of the circulation is entirely changed. In the original vessel the blood flowed from the cephalic to the caudal border of the hindbrain, while when the new arterial connections bring blood to the entire ventral border of the vein the blood begins to flow from the ventral to the dorsal border of the hindbrain. In the case of the subintestinal artery is a third examjjle of a profound change in the direction of the circulation. The blood originally runs through this artery out to the j^olk-sac, but when the vessel becomes a capillary i^lexus in the wall of the gut, the blood flows toward the heart within the embryo in the new mesenteric veins.


From these studies it is clear that it is important to consider each vessel of the embryo from the standpoint of the function it performs throughout its de\'elopment and that the effort toward a precise usage of the terms artery, capillary plexus, and especially of the term veiji, is an effort to understand the circulation of the embryo.


Bibliography

Von Baer, K. E.: Ueber Entwicklungsgeschichtc der Thiere, Beobachtung und Reflexion. Konigsberg, Gebr. Borntragcr. 1 Theil, 1828; 2 Theil. 1837.


Brachet, a.: Recherches sur le diterminisme hirfeditaire de I'cEuf des mammifferes. Dfiveloppement "in vitro" de jelines vfoicules blastodemiiqucs du lapin. Arch, de Biol., Liege, 1913-14, XXVIII, 447-503, 2 pi.


Broman, I. Ueber die Entwieklung und "Wanderung" der Zweige der Aorta abdominalis beim Menschen nebst Bemerkungen uber Gefii-sswiirzelwanderungen im allgemeinen. Anat. Hcfte, 1 Abt., Wiesb., 1908. XXXVI, 405-550, 5 pi.


D.iNDT, AV. E. A human embrj-o with seven pairs of somites, measuring about 2 mm. in length. Amer. Jour. Anat., Phila., 1910, X, 85-108, 6 pi.


Duval, M. Atlas d'Embrjologie. Paris, G.Masson, 1889.


Ev.\N"S, H. M. On the earliest blood-ve.ssels in the anterior limb buds of birds and their relation to the primary subclavian arterj-. Amer. Jour. Anat., Phila., 1909, IX, 281-319, 9 pi.


— . On the development of the aortse, cardinal and umbilical veins and other blood-ves.sel.s of vertebrate embrjos from capillaries. Anat. Record, Phila., 1909, III, 498-518.


. Die Entwieklung des Blutgefasssystems. Handbuch d. Entwcklngsgesch. d. Menschen, (Keibel and Mall). Leipzig, 1911, II, 551-fi88. Aho, Manual of Human Embrvol. (Keibel and Mall), Phila. and Lond., 1912, II, 570-709.


Felix, W. Zur Eutwicklungsgeschichte der Rumpfarterien des menschlichen Embrj-o. Morphol. Jahrb., Leipz., 1910, XLI, 577-614.


. Die Entwieklung der Harn- und Geschlechtsor gane. Handbuch d. Entwcklngsgesch. d. Menschen, (Keibel and Mall), Leipz. 1911, II, 732-955. Also, Manual of Human Embrjol. (Keibel and Mall), Phila. and Lond., 1912. II, 7.52-979.


GR.tFE, E. Beitrage zur Entwieklung der L>niere und ihrer Gefasse beim Huhnchen. Arch. f. mikr. Anat., Bonn, 1906, LXVII, 143-230, 5 pi.


Gr-aper, L. Untcrsuchungen iiber die Herzbildung der Vogel. Arch. f. Entwcklngsmechn. d. Organ., Liepz., 1907, XXIV, 375-410. 4 pi.


Grosser, O. Zur Anatomie und Entwicklungsgeschichte des Gefasssystemes der Chiropteren. Anat. Hefte, 1 .\bt., Wiesb., 1901. XVII, 203-424, 13 pi. . Die Elemente des Kopfvenensystem der Wirbeltiere.


Anat. Anz., Verhandl. d. anat. Gesellsch., Jena, 1907, XXX. H.AHN, H. Experimentelle Studien uber die Entstehung des Blutes und der ersten Gefasse beim Hiihnchen. Arch.


f. Entwcklngsmechn. d. Organ., Leipz., 1909, XXVII, 337^33, 3 pi. His, W. Untcrsuchungen uljer die erste Anlage des Wir belthierleibes. Leipz., F. C. W. Vogel, 1868. . Anatomie menschlicher Embrj-onen. I. Embr>-o nen des ersteu Monats. Leipz., F. C. W. Vogel, 1880. . Lecithoblast und Angioblast der Wirbelthiere.


Abhandl. d. math.-phys. CI. d. k. sUchs. Gesellsch. d.


Wissensch.. Leipz., 1900, XXVI, 171-32S. HocHSTETTER, F. Beitrage zur Entwicklungsgeschichte des Venensystems der .-Vmnioten. Morphol. Jahrb., Leipz., ISSS, XIII. 575-585, 1 pi. . Ueber die urspriingliche Hauptschlagader der hinteren Gliedmasse des Menschen und der Saugethiere, nebst Bemerkungen Qber die Entwieklung der Endaste der Aorta abdomir.ilis. Morphol. Jahrb., Leipz., 1890, XVI, 300-318, 1 pi. HoFFJUXN, C. K. Zur Entwicklungsgeschichte des Venensystems bei den Selachiern. Morphol. Jahrb., Leipz., 1893, XX, 289-304, 1 pi.

Kastschenko, N. Das Schlundspaltengebiet des Huhn chens. Arch. f. Anat. u. Entwcklngsgesch., Leipz., 1887.


258-300. 3 pi. Lewis, F. T. The development of the vena cava inferior.


Amer. Jour. Anat., Bait., 1901-1902, I, 229-244, 2 pi. Lewis, M. R., and Lewis, W. H. The cultivation of tissues from chick cmbrj-os in solutions of NaCl, CaCIj. KCl, and NaHCOj. Anat. Record. Phila., 1911, V, 277-293. LiLLiE, F. R. The development of the chick. New York, Henn,' Holt 4 Co., 1908. Mall, F. P. A human embrj-o twenty-six davs old. Jour.


Morphol., Bost., 1891, V, 459-480, 2 pi. . Development of the human coelom. Jour. Morphol., Bost., 1897, XII, 395-453. . On the development of the blood-vessels of the brain in the human embrj-o. Amer. Jour. Anat., Bait., 1904, IV, 1-18, 3 pi. McWhorter, J. E., and Whipple, A. O. The development of the blastoderm of the chick in vitro. Anat. Record.


Phila., 1912, VI, 121-1.39. PopoFF, D. Die Dottersack-Gefasse des Hulines. Wiesb., C. W. Kreidel, 1894. R.VBL, C. Ueber die Entwieklung des Venensystems der Seiachier. Festschr. z. 70. Geburtstage Rud. Leuckarts, Leipz., W. Engelmann, 1892. Ravx, E. Ueber die Arteria omphalo-mesenterica der Ratten und Mause. Anat. Anz., Jena, 1894, IX, 420^24. Reagax, F. P. Vascularization phenomena in fragments of embryonic bodies completely isolated from yolk-sac blastoderm. Anat. Record, Phila., 1915, IX, 329-341. Remak, R. L'ntersuchimgcn uber die Entwieklung der Wirbelthiere. Berlin, G. Riemer, 1851 [1855]. Sabin-, F. R. Der Ursprung und die Entwieklung des Lj-mphgefasssystems Anat. Hefte, 2 Abt., Wiesb., 1913, XXI, 1-98. Also transl. The origin and development of the lymphatic systems. Johns Hopkins Hosp. Rep., Monographs, New Seiies, No. 5, Baltimore, 1913, 1-94. . On the fate of the posterior cardinal veins and their relation to the development of the vena cava and azygos in the embryo pig. Contributions to Embry ologj', No. 7, Carnegie Inst. Wash., Pub. No. 223, 1915. SAlzer, H. Ueber die Entwieklung der Kopfvenen des Meerschweinchens. Morphol. Jahrb., Leipz., 1895, XXIII, 232-255, 1 pi. Spalteholz, W. LTeber das Durchsichtigmachen von menschlichen und tierischen Praparaten und seine theoretischen Bedingungen. Leipz., S. Hirzel, 1911; 2 Aufl., 1914. Streeter, G. L. The development of the venous sinuses of the dura mater in the human embn-o. Amer. Jour.


Anat., Phila., 1915, XVIII, 145-178. " STR.4.CKER, O. Entwieklung der Kopfvenen beim Huhn bis zur Ausbildung der \'ena capitis lateralis. Morphol.


Jahrb., Leipz., 1916. L, 49-71, 2 pi. Taxdler, J. Zur Entn-icklungsgeschichte der Kopfarterien bei den Mammalia. Morphol. Jahrb., Leipz., 1902, XXX, 275-373, 3 pi. . Zur Entwicklungsgeschichte der menschlichen Darmarterien. Anat. Hefte. 1 Abt., Wiesb., 1903, XXIII, 187-210. . U'eber die Varietaten der Arteria coeliaca und deren Entwieklung. Anat. Hefte, 1 Abt., Wiesb., 1904, XXV. 473-500.


Die Entwicklungsgeschichte des Herzens. Handbuch d. Entwcklngsgesch. d. Menschen, (Keibel and Mall). Leipz., 1911, II, 517-551. Also, Manual of Human Embrjologj- (Keibel and Mall) , Phila. and Lond., 1912, II. 534-570. Williams, L. W. The somites of the .chick. Amer. Jour. .\nat., Phila., 1910-11, XI, 55-100.


Explanation of Plates

Plate 1.

1. Injection of the vascular system of an ombryo pi^ of 2;j somites, which measured 7 mm. when fresh and C.2 mm. after fixation and dehydration. The injection is nearly complete and shows especially the primitive relations of the vaaa primitiva rhombencephali at the stage when it .serves as a vein for the forebrain and midbrain and as a capillary plextis for the hindbrain. X3S. A. I., artery of the first interspace; at., allantois; a. si., a. subintestinalis; 6. c, bulbus cordis; I., liver; pL, plexits on the spinal cord; s. !'., sinus venosus; t. a., truncus arteriosus; c. c. a., v. cardinalis anterior; v. c. I., v. cardinalis lateralLs; v. c. p., v. cardinalis posterior; V. n., v. umbiUcalis; va. p. r., vasa primitiva rhombencephali; ven. c, ventriculus cordis.


2. Partial injection of the vessels of a chick of 12 somites. The needle \v!is introduced into one of the omphalomesenteric veins near the heart. The transverse lines show the position of the interspaces. The sections shown in text-figure,s 1, 2, and i are from a chick of the same stage which was completely injected. X54. Me., mesencephalon at the level of the section shown in figure 1; v. om., v. omphalo-mesentcrica.


3. Injection of the heart and the cephaUc aorta>, both dorsal and ventral, in a chick of 9 somites. The needle was introduced into the dorsal aorta opposite the somites. X04. Ao. il. c, aorta dorsalis cephalica; n«. v. c, norta vcntralis cephalica ; h., heart; ine., mesencephalon; v. om., v. omphalo-mesentcrica.


Plate 2.

1. Partial injection of the vessels of a chick of 14 somites. The needle was introfluced into the dorsal aorta opposite the somites. The vascular ])lexas on the mesencephalon is not injected, though it is present at this stage, as is shown in text-figiu-c 1. X 100. A . .s-o., artery of the soniatoi)leure; d. C, duct of Cuvier just before it has connect etl with the omphalo-mesenteric vein; me., mesencephalon; v. c, v. cardinalis communis, that is, before it has an anterior and a posterior division; ii. so., vein of the somatopleure; v. t., v. transversa of the first interspace; in. p. r., vasa primitiva rhombencephali.


2. Injection of the blood-vessels of a chick of 10 somites, to show the relation of the primitive vessel of the rhomben cephalon to the primitive cerebral vein, on the one hand, and to the anterior cardinal vein, on the other. This is the stage before the vena capitis prima is completed. X58. D. C, ductus Cuvieri; v. c. a., V. cardinalis anterior; v. c. p., v. cardinalis posterior; u. ci . p., v. cerebralis primitiva, which will become the cephalic division of the v. capitis i)rima; v. I., v. transversa of the first interspace; va. p. r., vasa primitiva rhombencephali.


3. Transverse section of an injected chick of 30 somites, after 52.J hours of incubation, passing through the twenty first interspace. The section is from the same series as figure 4 on the same i)late, figures 2 and 3 on i)late 3, and figure 2 on plate 4. Figure 4 on plate 3 is from another series. This section is to show the beginning of the spinal arteries as they show in the tenth to the seventeenth interspaces on plate 3, figure 1. The section is below the level of the omi)halo-niesenteric arteries. It is .i() ^ thick and is unstained. X 140. Ao., aorta; f. c. p., V. cardinaUs posterior; v. om. p., v. omphalo-mcsenterica posterior; l)'. </., Wolffian duct.


4. Transverse section of an injected chick of 30 somites, after .52J hoius of incubation, passing through the twenty seventh interspace. The section is to show the relative jiosition of the direct arteries to the jjosterior cardinal vein and the arteries of the somatoiileure. The section is below the level of the omphalo-mesenteric prtcries and is in the region of the posterior limb-bud. It is oO/i thick and is unstained. X140. .4o., aorta; v. c. p., v. cardinalis posterior; r. om. p., v. omphalo-mcsenterica posterior; W. d., Wolthan duct.


Plate 3.

1. The cardinal veins from an injected chick of 2.") somites, to show tlie method of origin of the spinal aiteri<\s. X106.


.1. .i and tt. 18, arteries of the third anil eighteenth interspaces; d ('., ductus Cuvieri; /*/., plexus on the spinal coni; u. c. a., v. cardinalis anterior; v. c. p., v. cardinalis posterior; r. /., \-. transversa of the first interspace; m. p. r., vasa primitiva rhombencephali.


2. Transverse section of an injecte<l chick of :U) somites, after 52} hours of incubation, passing through the fifteenth interspace. The section is to show a spinal artery like that of the seventh interspace of plate 3, figure 1; it is above the level of the onipludo-niraenteric arteries and shows the |)osterior omphalo-mesenteric veins on either side. The section is .")0 ii thi(^k and is unstained. X 140. Ao., ;iorta; v. c. p., v. cardinalis posterior; I', om. p., V. oniphalfwnesenterica posterior; 11'. (/., Wollhan duel.


3. Transverse section of an injected chick of 30 somites, after .52J hours of incubation, psissing through the sixteenth somite. This is the next section in the scries below that of plate !!, figure 2. It shows a ilirect dorso-lateral artery to the pfwterior cardinal vein. The section is M ^ thick and is imslained. X140. Ao., m)rta; V. c. p., V. I'ardinalis posterior; /■. om. p., v. omph:ilo-inesent erica posterior; M". d., WolfTian duct.


4. Tran.sverse section of an injected chick of 2.5 somitis aft«r .52 hours of incubation, passing through the seventeenth interspace. The section is to show the transition between the stage of figure 3 of plate 2 and figure 2 of plate 3, in the formation of a spinal artery. It is .50 m thick and is unstained. X 140. Ao., aorta; v. c. p., v. cardinalis posterior; ir. d., Wolffian duct.


.5. Injection of the aorta, the arteries of the pronephros, and the lateral anrl posterior cardinal veins in an embr>'0 pig of 20 somites, from a specimen of the same litter as the one on plate .5, figure 1. The specimen is shown for the ventral aspect. X140. A.9ioA. /.?, arteries to the spinal cord in the ninth to the twelfth interspaces; an., aorta; v. c. I., v. cardinalis lateralis; v. c. p., v. cardinalia posterior.


Plate 4.

1 . Injection of the vessels of the head of an embryo pig of 27 somites, measuring 7.1 mm. after fixation and dehydration.

The specimen is from the same Utter as the one on plate 5, figure 2, and is to show the completion of the vena capitis prima and its relation to the vasa primitiva rhombenccphali. It shows that the primitive vessel of the hindbrain does not atrophy when the vena capitis prima is completed, but rather develops into a plexus on the hindbrain. X94. ylt., atrium; b. c, bulbas cordis; /.//., fretum Halleri; /. n., truncus arteriosus; r. cap. p. I, V. capitis prima, first or cerebral segment, which drains the forebrain and midbrain; r. aip. p. 2, v. capitis prima, second segment, which drains the forebrain, the midbrain, and the visceral arches; r. cap. p. S, V. capitis prima, third segment, which is the anterior cardinal vein; va. p. r., vasa primitiva rhombencephali ; ven. c, ventriculiLS cordis; ves. a., vesicula auditiva; V, position of the root of the n. trigeminus; 17//, position of the roots of the nn. cochlearis et vestibularis; IX, position of the root of the n. glossopharyngeus.


2. Transverse .section of an injected chick of 30 somites, after .52j hours of incubation, passing through the twenty fifth interspace. The section is to show the diverticula of the dorsal aorta like those of the first and second interspaces in plate 1, figure 2, which give rise to the ctirdinal veins. The section is 50/i thick and is unstained. X140. ylo., aorta; IT'. 'f., Wolffian duct.


3. Partial injection of the vessels of an embryo jiig of 14 somites, measuring 4 mm. after fixation and dehydration.

The specimen was injected tlu-ough the dorsal aorta opposite the somites. X56. A. 9, artery to the spinal cord in the ninth interspace; at., atrium; h. c, bulbus cordis; /. //., fretum Halleri; s. 1, first somite; s. ('., sinus venosus; I. a., truncus arteriosus; v. c. p., v. cardinalis posterior; v. om., v. omphalo-mesenterica ; ven. c, ventriculus cordis; ves. a., vesicula auditiva.


Plate 5.

1. Partial injection of the vessels of an embryo pig of 20 somites, measuring 6 mm. after fixation and dehydration.

It shows the omphalo-mesenteric arteries, the subintestinal arterj- and the arteries of the pronephros; X41. A. 9, artery to the spinal cord in the ninth interspace; a. om., a. omphalo-mesenterica; a. si., a. subintestinalis ; al., allantois; h. c, bulbus cordis; s. al., stalk of the allantois; (. a., truncus arteriosus; v. c. l, cardinalis lateralis; ven. c, ventriculus cordis.


2. Partial injection of the vessels of an embryo pig of 27 somites, measuring 7.1 mm. after fixation and dehydration.

It shows the general development of the vascular system at the stage when the vena capitis prima is completed. The vessels oppo.site the hindbrain, both deep and superficial, are extniva-sateil in this embryo and hence they are shown on plate 4, figiire 1, from an embryo of the same Utter. X.51. -4. om. il., a. omphalo-mesenterica dextra, the other two omplialo-mesenteric arteries in the figure being on the left side; at., atrium; b. c, bulbus cordis; (/. C, ductus Cavieri; /., liver; t. a., tnmcus arteriosus; v. cap. p. 1,\\ capitis prima, first or cerebral segment, which drains the forebrain and midbrain; i'. cap. p. 3, v. capitis prima, third segment, which is the anterior cardinal vein; v. om., v. omphalo-mesenterica; v. u., v. umbiUcalis; ven. c, ventriculus cordis; X.. extra vasation involving both the vasa primitiva rhombencephali and the vena capitis prima, as can be seen on plate 4. figure 1.


Plate 6.

Injection of the ve.s.sels of the head of a chick of 29 somites, to show the origin of the vena capitis prima. The vein extends from the region of the diencephalon to the duct of Cuvier. The injection shows that the vein arises in throe segments; the first segment is a true primitive cerebral vein, which drains the forebrain and will soon drain t lie midbrain ; the second segment is an anastomosis between the maxillary, the mandibular, and the other visceral arches and the anterior cardinal vein, and it drains the forebrain, the midbrain, and the visceral arches; the third segment is the anterior cardinal vein, which drains the brain and the visceral arches. X128. .4. 6., artery on the rhombencephalon, which at this stage is bilateral and is part of a plexus which will give rise to the basilar arterv; a. S, arterj' to the medulla in the third interspace; d. C, ductus Cu\-ieri; v. c. p., V cardinalis posterior; v. cap. p. 1, v. capitis prima, first or cerebral segment which drain the forebram and midbrain; r. c<:p. p. 3, v. capitis prima, second segment which drains the forebrain, the midbram, and the visceral ajches; v. cap. p. 3, v. capitis prima, third segment, which is the anterior cardinal vein; v. m. p., V. ma-xiilarls primitiva; r. om., v. omphalo-mesenterica; v. I., v. transversa of the first interspace; c. u., plexus in which the v. umbiUcaUs wiU arise; va. p. r.. vasa primitiva rhombencephaU ; ves. a., vesicula auditiva; V position of the root of the n. trigeminus; 17//, position of the roots of the nn. cochlearis et vestibularis.


Plate 7.

Injection of the vessels of the brain of an embryo pig measuring 0.5 mm. in length after fi.xation and dehydration. The injection is a complete one, but the vessels of the ^■isc■eral arches and mo-st of the vessels of the cerebrum have been omitted in the drawing. The figure show.s, first, tlie longitudinal artery of the central nervous system, which Extends from the tip of the carotid artery to the caudal tip of the spinal eor<l; this artery is a plexus opposite the subf halanuL*, a single vessel down to the low^er part of the iiuMlulla, and again a plexus on the cord; second, a part of the capillary plexus which invests the entire neural tube; third, the relation of the primitive veins of the forebrain, the midbrain, and the hindbrain to tlie vena capitis prima. X73. .4. h., a. basilaris; a. c. 1,&. carotis interna; a. 1, artery to the medulla in the first interspace; a. m. p., a. maxillaris primitiva; v. cap. p. 1, v. capitis prima, first or cerebral segment, which drains the forebrain and midbrain; v. cap. p. ii, v. capitis prima, second segment, which is shown only in outline, and which drains the forebrain, the midbrain, the hindbrain, and the visceral arches; v. cap. p. S, v. cajjitis prima, third segment, which is the anterior cardinal vein and which drains the brain and the visceral arches; v. m. p., V. maxillaris primitiva; vcs. a., vesicula auditiva; .i, 4, "nd fl, third, fourth, and sixth aortic arches, which are coming from the heart and are leading to the descending aorta, which is concealed by the cardinal segment of the vena capitis prima; V, position of the root of the n. trigeminus; VIII, position of the roots of the ni\. cochlearis et vestibularis; 7.Y, position of the root of the n. glosso-pharyngeiLs; A', position of the root of the n. vagas; XII, position of the root of the n. hypoglossus.