Talk:Paper - On the vascularization of the spinal cord of the pig (1914)

From Embryology


Hoskins EK. On the vascularization of the spinal cord of the pig. (1914) Anat. Rec. 8(7): 371 - 392.

E. R. HOSKINS From the Inslilule of Anatomy, University of Minnesota



Until 1904 little work had been done upon the development of the blood-vessels of the spinal cord, except that of His ( '86) who undertook to follow the growth of these vessels in human embryos. The observations of this author have been largeh- disputed by Sterzi ( '04) and Evans ( '09), the two writers who have done most of the work in this field.

By far the most comprehensive publication upon the development of the vessels of the spinal cord is that of Sterzi ('04j. In this he discusses the development of the vessels in the five higher classes of vertebrates. As a tj'pe of the ^Mammalia he uses the sheep. He brings out the following points:

The blood-vessels first approach the cord at the ventro-lateral border and spread over the ventral surface, then over the lateral, and finally over the dorsal surface. Each vertebro-medullary artery as it approaches the cord divides into a ventral and a dorsal ramus, the ventral and dorsal radical arteries. The ventral radicals from either side halt at the lateral edges of the floor-plate, and each divides into a cranial and a caudal branch. These anastomose with those of adjacent segments and form two longitudinal arteries on tlu^ ventral surface of the cord, the "tractus arteriosus primitivus. " Later they send out medial rami and through these become connected. Still later, alternate parts of the two tracts degenerate while other parts continue to develop. These enlarged segments are joined together through their medial rami and form a single ventral artery which Sterzi terms the tractus arteriosus ventralis, and which is tiie anterior spinal artery of most authors. From the primitive tract, dorsal rami enter the substance of the cord. Each dorsal ramus forms a loop and gives rise to a vein, which courses ventrally and enters the primitive sulcus. Later other vessels extend into the cord from the lateral, and still later from the dorsal surface.

The dorsal radical arteries, where they di\ide, form many small longitudinal capillaries just ventral to the i:)()ints of emergence of the tlorsal ner\'e roots. From these capillaries there is formed later a longitudinal artery on either side of the cord, in this plane (tractus arteriosus lateralis). The vessels entering the cord are f'rst solid and later become hollow.

Evans ( '09) shows by a series of injected pigs the early de\'elopment of anterior spinal artery In his figures the mid-ventral surface of the cord is shown to be free from vessels until the embryos are 8.5 mm. in length, and the mid-posterior surface until after the pigs are between 8 and 10.5 mm. in length. He does not take uj) the later stages,


For a study of this nature, injected embryos are indispensable, and they are best injected while living, with warm india ink diluted one-half with weak ammonia water. It is pi-ef(M'al)le to inject through the umbilical arter}^ rather than through the umbilical vein, because the arteries are less readily ruptured and because the route between them and the vessels of the spinal cord is much more direct.

Eml)ryos used for thin serial sections are better if they are congested instead of being injected. This congestion is accomplished as follows: The uml)ilical cord is tied while the embryo is yet living, thus causing an increase in the blood pressure in the aorta. One of the most direct outlets for this increased pressure is the system of segmental arteries, and through these the bloodvessels in and around the spinal cord soon l)ecome engorged. When this condition is i-eached, as evidenced by the increased redness of the dorsal region of the embryo, the li\'e embryo is droi;)]Ded into a fixing fluid which penetrates rapidl}^ so that the capillaries are fixed before they collapse. Bouin's picro-formo-acetic mixture serves this purpose very well. P^mbryos treated in this manner show the smaller vessels much more plainly than those fixed in the usual way.

Small injected embryos which have been cleared in oil may be dissected under the binocular microscope and all the external vessels of the cord demonstrated, or they may be sectioned in celloidin to show the internal vessels.

From pigs larger than 25 or 30 mm. the cord with its membranes may be dissected out and embedded in celloidin and cleared, or for temporary preparations may be cleared directly.

Serial sections of the cleared embryo or spinal cord can be kept permanently between two pieces of paper soaked in oil, or can be transferred to slides and mounted in damar gum. To make slide preparations, the sections are cut in oil and placed on a piece of thin paper in the same order they are to have on the slide. Another piece of oiled paper is laid down upon them and the whole inverted. The first paper is now peeled off and the other paper holding the sections is inverted upon the slide. This paper is then peeled off, leaving the sections on the slide in their proper order. They may then be washed carefully with xylol, and covered.

There are several advantages in a study of this kind, in section ing celloidin-emljcddcd embryos free-hand in oil. The cord can be turned so that sections may be cut through any plane. Sectioning is done more rapidly than with a microtome, and much time is saved. It is easy to make transverse, sagittal and frontal sections from any region of tlie same cord. The block can be examined witli a Umis. (hiring tlie sectioning, and any particular vessel or vessels included, in a section. Also, sections may be made of different shapes.


In order to determine as near as j>ossible the arrangement of the blood-vessels of the s|)inal cord in the adult comlition, a iHnnl)er of f(>tal pigs, neai' full-term, were injected and their spinal cords dissected out for study. Alt liough these showed some variation in the blood-vessels, as is to be expected, a general


plan was to be olisorvod. The following descrii)ti()n i.s of the tj'pical condition found in the blood-vessels in and around the spinal cord of pigs of about 240 nun. in length:

The terminology used by Kadyi ('89) for the blood-vessels of the adult human cord will he rt^ferred to frequently.

On the surface of the cord at this stage are four main longitudinal arterial systems which are located, one, median on the ventral surface; one, on each dorso-lateral surface; and one median on the dorsal surface.

The vertebro-meduUary branches of the dorsal segmental arteries approach the cord laterally and each divides into a dorsal and a ventral branch termed the posterior and ventral radical arteries respectively (figs. 1 and 2). The latter reach the cord along the cranial surface of the spinal nerves and ganglia, and, on the cord, run cranially. The}^ may be equally distributed to the two sides of the cord and to the different regions of it, or some regions may ha\'e more than others. They average eighteen in number. The \Tntral and dorsal radical arteries have lost their connection wdth the vertel)r()-medullary artery in places and extend between the artery on the ventral surface of the cord and one of those on the dorso-lateral surfaces.

Soon after reaching the cord each ventral radical artery branches, giving off one ramus which courses cranially, and another which extends caudally. These divisions anastomose with those of neighboring ventral radicals, on the same or opposite side

Fig. 1' Ventral svirfarc of the iimldlr thoracic- regiou of the spinal eonl of a fetal pig, 240 mm. in length. .4..S..4., anterior spinal artery; A.R.W, ventral radical artery; T.A.P., accessory anterior spinal artery (remains of primitive arterial tract); V.R.V., ventral radical vein; V^..S..4., anterior s])inal vein; C.P., capillary plexus. X 0.

Fig. 2 Dorsal surface of I he same part of t he eoid shown in figure I. A .M .D., median dorsal artery; A.P.L.. dorso-iatcral artery; A.R.I)., dorsal radical artery; V.^f.I)., median dorsal vein; V .I'.L.. dorso-lateral vein and plexus; V.R.D., dorsal radical vein; \'.fj.. dorso-laleral venous plexus. X 10.

' The figures in this pajx')-, exc('i)t nunilKMs '.] and I, were drawn witji a camera lucida. The size of the eml)ryf)s is the greatest length, as iiieasured in 7.") per cent alcohol.



TA P-1




>, '


37(3 E. R. HOSKINS

of the cord, and form in this way the anterior spinal artery, which Ues in or near the median ventral line. Occasionally a radical artery, instead of dividing, goes across the cord and joins the cranial or caudal ramus of the one on the opposite side. The anterior spinal artery has a winding course, bending laterally to meet the vessels which form it, and making many smaller irregular bends to one side or the other. In some places it may lie to one side of the mid-line for several segments and where this occurs there are found numerous longitudinal rami of the anterior spinal artery, or the cranial antl caudal tlivisions of the ventral radicals. These may be called accessory anterior spinal arteries, and are sometimes present even where the anterior spinal artery lies in the mid-line. Here they are located between this vessel and cord, at the lip of the ventral median fissure. They are the remnants of the "tractus arteriosus primitivus" of Sterzi.

The dorsal radicular rami of the vertebro-medullary arteries are much more numerous than the \'entral ones. They course dorsally along the cord and in a slightly cranial direction, to a plane just ventral to the emergence of the dorsal roots of the spinal nerves, where they divide into two rami, one extending cranially and one caudally. Each of these rami anastomoses with the one of the adjacent segment, and thus there is formed on either side an irregular longitudinal vessel, the dorso-lateral artery (fig. 2; tractus arteriosus postero-lateralis of Kadyi). From this artery recurrent rami supply the dorsal nerve roots and si:)inal ganglia, and the lateral surface of the cord. Other rami, two or three in each segment, and much larger than the above, run dorsally and by longitudinal anastomoses with each other, and with similar rami from the opposite side, form an artery in the mid-dorsal line which may be termed the median dorsal artery.

Wry small rami from the dorso-lateral arteries run ventrally along the cord and unite with oth(M"s from the anterior spinal artery, forming a plexus on the ventral and lateral sides. These ventral rami of the dorso-laterals anastomose freel_y in a longitudinal direction and form one or more small longitudinal arteries between the ventral and dorsal nerve roots in some parts of the cord. These are the tracti arteriosi laterales, and ventro-later


ales, of Kadyi. Still other small rami from the dorso-lateral and median dorsal arteries form a capillary plexus on the dorsal and dorso-lateral surfaces of the cord. In a few places along the cord the dorso-lateral arteries are double, one divisif)n lying dorsal to the dorsal nerve roots, and corresponding perhaps to the '"tractua arteriosus posterior" of Kadyi.

The median dorsal artery is a very irregular longitudinal vessel formed by the dorsal rami of the dorso-lateral arteries, as described above. In places it is double or ma\' show a longitudinal capillary arrangement. Many of its lateral rami anastomose longitudinally forming small arteries parallel with the median dorsal artery (fig. 2). This is also true of the dorsal rami of the dorso-laterals. By the anastomoses of the rami of the various arteries just described, the entire cord is surrounded by an arterial vascular system, and from all parts of this network smaller arteries penetrate its substance.

The veins of the spinal cord are in three princijial longitudinal systems, and other smaller ones. Of the three, two are dorsal and one ventral. All three show evidence of their capillary origin. The anterior spinal vein is the smallest of the three (tig. 1). It lies between the cord and the arteries, in the median ventral line. It is larger than the accessory anterior spinal arteries, but never attains the size of the anterior spinal artery i)r()i)er. It is very irregular and in some regions is entirely replaced by a narrow network of capillaries.

On either side of the median ventral sulcus, the cord is covered with large venules, some of which lie between the cortl and the arteries, and some of which are external to the latter. They are often two or three times as large as the arterioles to which they correspond. They anastomose freely with the anteri«)r spinal vein and empty laterally into the ventral radical veins which are in close relation with. the ventral nerve roots and ventral radical arteries, but which are nuicli nu^v numerous tiian the latter, one being present on nearly every nerve root. They ilrain blood also from the lateral sm-face of the cord. Their ventral and dorsal rami often form short, small, longitudinal veins by anastomoses, some of which in otluM- animals have been named, antero-lateral.


etc. The 1)1()()(1 from tho anterior spinal veins and venous capillaries of the gc'neral ventral surface form, in places, transverse channels which are perhaps large enough to be called veins.

On either side of the dorsal surface of the cord there extends longitudinall}' a large irregular vein, the dorso-lateral, about half way between the artery of the same name, and the median dorsal artery. Some parts of these vessels and their rami, like the ventral venous capillaries, lie external to the arteries and some internal to them. They are the largest vessels on the cord with the exception of the anterior spinal artery (compare figs. 1 and 2). Dorsally these veins are united through large capillaries, and blood leaving the cord in the median line may flow either to the right or left. Half way between two consecutive nerve roots the dorsolateral veins usually break up into many divisions so that each may be seen to drain blood from adjacent halves of two segments. Laterally' they empty through one or more divisions into the large dorsal radical veins, one of which lies upon each dorsal nerve root (fig. 2).

A fourth longitudinal venous system, smaller than the three described, lies in the median dorsal sulcus. It resembles the dorso-lateral veins except that it is more irregular, and in places it may be entireh^ lacking. Its lateral rami empty in the dorsal venous capillary plexus or directly through larger vessels into the dorso-lateral veins. It may be termed median dorsal venous system.

Some of the venous capillaries of the lateral surface drain into the dorso radical veins, some into the dorso-lateral veins, and some into ventre radical veins, and all these vessels together with the anastomoses of the veins on the ventral and dorsal surfaces already mentioned above, completely surround the cord with a venous system, corresponding to the system described for the arteries.

Of the arteries entering the cord, the largest are those in the ventral fissure, the ventral central arteries, which form two nearly parallel rows, but which are not paired. They arise from the anterior spinal arteries, or the accessory anterior spinal arteries. They show evidence of the capillary origin in longitudinal anas


tomoses found between vessels of the same side. These anastomoses are numerous in the fissure, particularly near the vessels from which the ventral central arteries arise.

The ventral central arteries vary considerably in size, some being as large as the vessels they arise from and others much smaller (fig. 3) . The course of the smaller vessels is usually more irregular than that of the larger. They pierce the substance of the cord at different levels, some entering near their origin and others extending some distance into the fissure. Their general course is dorso-lateral, but those entering near the mouth of the fissure may bend very sharply to the side and enter the ventral horn of the gray substance. The others course more dorsally nearly to the level of the central canal where they make a decided lateral bend, and divide into two or more rami, although sometimes they give off rami more ventrally than this (figs. 3 and 4i. The principal divisions of these arteries extend in a longitudinal plane, and anastomose with similar rami of adjacent vessels. They also give off smaller arteries and capillaries which ramify through the gray matter in all directions, helping to form a dense plexus. The longitudinal arteries tend to form loops after they have coursed in one direction for a short distance, as thej' do in young embryos (fig. 5). One artery may form several such loops, producing as many longitudinal vessels, each succeeding vessel lying dorsal or lateral to the last, and of a lesser cahber. These smaller longitudinal vessels anastomose with each other ventro-doi*sally and laterally by rami which usually leave them at right angles, and also anastomose with rami from vessels other than the ventral central arteries, as will be described later.

Besides the rami of the central arteries just described, other rami extend farther laterally into the gray substance before l)ranching. Some of these, instead of ft)rming longitudinal vessels, form small irregular ones which ramify through the gray matter in all directions, anastomosing with similar vessels from other arteries in this region and forming a dense capillary plexus in the ventral and dorsal horns.

Other arteries, smaller than the ventral central arteries, enter the c(ird from the dorsal median sulcus and course ventrallv and


laterally to tlio dorsal horns of tho gray substance. Here they form still smaller vessels resembling to some extent those formed by tlie \'entral central arteries, but most of their rami are short and do not extend longitudinally. These may be called the dorsal central arteries, but they are more similar to the peripheral arteries from other surfaces of the cord, than to the ventral centrals, and perhaps should be called dorsal peripheral arteries. They give ofT many small lateral rami in the white substance and in the outer part of the gray substance.

In addition to these vessels, other small arteries enter the cord from all sides, from the arteries and arterioles which surround it. These are the peripheral arteries referred to above. The}^ are very numerous and in a single thick cross section as man}^ as fifty or sixt}' of them ma}^ be counted. They give off short rami in the white layer of the cord and extend into the gray layer. These rami branch and anastomose and form a loose capillary network. The vessels entering the gray substance enter into the longitudinal plexus already described and give oflf lateral rami which branch freely and anastomose.

The longitudinal vessels which arise from the ventral central arteries are quite large, but the other vessels formed from these trunks, and those formed from the dorsal central and peripheral arteries, are much smaller. A very thick section presents a l)icture of an inner core of longitudinal vessels with other vessels extending into it at right angles from all points on the periphery of the cord (figs. 3 and 4).

Via. 3 Saf!;ittal section from tlic lower thoracic region of the .si)inal coni of a 240 mm. fetal pig. A.S.A., anterior spinal artery; .l.f. /I ., ventral central artery; ,1./'., peripheral artery; A.C.P., dorsal central artery; r..S..4., anterior spinal vein; V.C.A., ventral central vein; \'.('.l'., dorsal central vein; l'./'.. jx-ripheral vein. X 3.5.

Fig. 4 Transverse section through the lower thoracic region of the spinal cord of a 240 mm. fetal pig. A.S.A., anterior s|)inal artery; A. P., peripheral artery; A., artery; A.C.A., ventral central artery; S.A.S.A., accessory anterior spinal artery; A.I'.L., dorso-Iateral artery; A.M.D., median dorsal artery; .t.N.r., anterior spinal vein; V'., vein; V.P., peripheral vein; V.P.L., dorso-latcral vein; 1'.. !/./>., median dorsal vein. X 35.




The early development of the anterior spinal artery has been described by Evans ('09) and Sterzi ('04).

Some pig embryos of 12 mm. show a fairly well developed anterior spinal artery, while in others of 14 or 15 mm. it is just beginning to form. Although, after this vessel is once formed, it does not undergo marked changes, there is some modification. For example, the ventral radical arteries meet it at right angles or nearly so, until the embryo is 40 or 45 mm. in length. Thereafter, the growth of the cord and the fixed position of the radicals seem to cause the artery to be pulled laterally by the radicals, and a gradually decreasing angle is formed at the points where the radicals meet it (fig. 1).

As the embryo grows, the number of radical arteries continues to decrease even after the anterior spinal artery is well formed. This seems to be true until the embryo reaches the length of about 100 mm.

Some of the arterial capillaries on the ventral surface of the spinal cord are continuous with the anterior spinal artery directly, or indirectly through remains of the tractus arteriosus primitivus, and others are continuous with the capillaries of the lateral surfaces of the cord.

A dorso-lateral artery is formed in the capillary plexus on each of the lateral surfaces of the cord, just ventral to the point of emergence of the dorsal nerve roots. The dorsal radical arteries branch in this region and give oiT dorsal and lateral rami, which are continuous with the lateral capillaries just mentioned. A very irregular longitudinal vessel develops where certain of these

Fig. .5 Transverse section through the mid thoracic region of the spinal cord of an 11 mm. |)ig embryo. T.A.P., primitive arterial tract; R.D.T.A.P., R.L. T.A.I'., R.M.. dorsal, lateral, and medial rami of the primitive arterial tract; A.V..\I., vertehro-medullary artery; A.R.V., A.R.D., ventral and dorsal radical arteries; C.R., Cr.R., D.R.,, caudal, cranial, dorsal and ventral rami of the dorsal radical artery; I). P., D.L.P., dorsal and dorso-lateral capillary plexuses; D.L.C.G., V.L.C.G., dorso-lateral and ventro-lateral groups of peripheral capillaries; V.L.V.P., ventro-lateral venous plexus; V.R.V., V.R.D., ventral and dorsal radical veins; S.P.G., spinal ganglion; V.N.R., D.N.R., ventral and dorsal nerve roots; S.N ., spinal nerve. X 200




capillarios iiuToaso in sizc\ perhaps on account of the increased pressure from the dorsal radical arteries. This longitudinal vessel is indicated in embryos of 12 mm. and is quite stronglj' developed in embryos of 15 to IS nnn. In these stages it seems to dip ventrally to meet the approacliing radicals, as pointed out by Sterzi for the sheep ('04). -Vs the embryo grows, this dorso-lateral artery becomes more and more regular. It is still somewhat irregular in embryos of (30 mm. but quite regular in those of 75 mm. The dorso-lateral artery never attains the size of the anterior spinal nor is it ever so regular in its course. In places it may develop as two or more vessels, but these are always smaller than the single artery. The dorso-lateral arteries are each continuous with the capillaries of half the cord in the early stages, but as the cord increases in size they supply directh^ only the dorso-lateral surface.

The capillary network on the lateral surface of the cord is at first continuous with tliat extending through the mesenchyma of this region as far, laterally and dorsally, as the myotomes and body wall respectively. In later stages when the membranes of the cord begin to develop, the connections between the vessels of the cord and those in the mesenchyma around it are lost.

The median dorsal artery is the last of all the vessels on the cord to develop. In pigs of 30 nun. it is still very irregular and indefinite, and is entirely lacking in places, although the vessels which go to form it, the dorsal rami of the dorso-lateral arteries, may be seen in embryos of 20 mm. In pigs of 45 mm. it is quite definite, lying in or near the mid line of the dorsal surface, as described above for the 240 mm. embrj'o It never becomes very regular, and in pigs of 100 nnn. it resembles the condition seen in the pig oi 240 mm. It is continuous with the arterial caj)illaries of the dorsal surface of the cord and with the dorso-lateral arteries.

In addition to these main arterial trunks there develop on various parts of the cord, especially on the lateral surfaces, short longitudinal arteries. These are never large or regular. The}' have been desci-ibcd in connection with adult human coi-d under the terms "tractus arteriosus; ventro-lateralis, posterioris, and


lateralis" (Kadyij. Of these, the "tractus arteriosus posterior" is the most prominent and corresponds to the description in this paper of parts of the dorso-lateral artery, where it sometimes has two divisions, one of which runs dorsal to the dorsal nerve roots and the other ventral to them. The dorsal divisions are evidently the same as this ' tractus. '

The veins on the cord develop in much the same way as do the arteries. The ventro-lateral surface of the cord in very young embryos is covered with capillaries, and these are continuous laterally with the capillaries in the mesenchyma round the neural tube. Medially they become continuous with the lateral rami of the primitive arterial tract. When this tract becomes separated from the cord by the ingrowth of mesench>niia, these capillaries send medial outgrowths between the tract and the cord, as seen in embryos of 12 to 15 mm. Dorsall}' the}' grow along the cord and spread over the dorso-lateral surface (pigs of 6.2 mm.) and later over the dorsal surface (pigs of 7.5 mm.). Laterally they spread over the ganglia.

From the ventral surface, the blood draining away through the capillaries soon establishes segmental vessels, the ventral radical veins, which course laterally along the nerve roots. Each radical vein on one side drains adjacent halves of two segments. These receive blood from the capillaries of the ventral, lateral, and ventro-lateral surfaces. Lying in the ventral median fissure in young embryos, small longitudinal veins may be seen in different regions of the cord, and in embryos of 25 to 30 mm. a fairly detinite longitudinal vessel may be found here. This vessel in still older embryos becomes a more definite trunk and may be called the anterior spinal vein. It never attains the size of the anterior spinal artery. Laterally it drains into the ventral radical veins.

Some of the ventral and lateral capillaries of the younger embryos, early become dilYcrentiated into veins. This is especially true of the dorsal vessels. From these, some of the bUxxl drains laterally out through vessels in the mesenchyma to the myotomes. .\ pig of (').2 mm. shows three planes in whicii this occurs, one on a le\el with tlu^ dorsal surface, one just al)ove the level of the ven


tral surface, and one about half way between the other two. At the myotomes the blood drains Aent rally into the intersegmental veins. Some of the capillaries of the lowest of these three planes, which tlrain the blood from the lateral surface of the cord and from the ganglia, soon become large and are called the vertebromeduUary veins, one pair of which is formed for each segment. In older embryos they course along the spinal nerves with the vertebro-meduUary arteries. They recei\e the blood from the ventral and dorsal radical veins. The former have been described. The latter develop along the sides of the gangUa in the capillaries already mentioned. At firsi they carry only a part of the blood from the dorsal surface of the cord, but later (pigs of 25 mm.) they carry practically all of it. They are more numerous than the corresponding ventral radicals, and are found in every segment.

The venous capillaries of the dorsal-lateral surface on either side draining toward the nerve roots early establish longitudinal veins. These are onlj' about half as long as a segment of the cord. Figure 5 of an 11 mm. pig, shows an indifferent plexus on this surface, but in 15 to 17 mm. embryos, fairly definite vessels may be seen. These become more and more regular as the animal develops, and as embryos of 50 to 60 mm. show, they form a venous system on either side of the cord just dorsal to the dorsal nerve roots, much like that described for the 240 nun. stage. These systems constitute the dorso-lateral veins (fig. 2).

The first blood vessels entering the cord grow in as capillaries from the ventral surface. Sterzi ('04) reports vessels in the cord of a sheep of 5.5 nmi., but they were not apparent in the cord of pig embryos of less than 7.5 nmi. These vessels are the dorsal rami of the primitive arterial tracts, of the lateral rami of these tracts, and of the other capillaries near the median line. They are the first indications of the central arteries and veins. They form two nearly parallel rows, one on either side of the epend>nnal layer, f)r some of them may lie in this layer. They grow dorsally about half way to the dorsal surface of the cord. They exhil)it numerous longitudinal anastomoses and form a plexus along the ateral side of the ependymal layer in each half of the cord. These are true capillaries at first, but soon differentiate into arteries and veins.


Those coming directly from the primitive arterial tracts all become arteries, while those coming from the vessels lateral to the tract may become either veins or arteries.

In embryos of 9.5 mm. another group of capillaries may be seen to have entered the cord. These come from tjie lateral surface, extending medially nearly to the central canal. Later they anastomose ventro-dorsally and longitudinally, among themselves and with the vascular sprouts from the ventral surface.

The vessels in the cord of a pig of 11 mm. present the following characteristics, as shown in figure 5. Rami from the primitive arterial tract may anastomose with those from the ventral capillaries. Neighboring vessels of the same kind anastomose freely and give off lateral rami into the anlagen of the ventral horns of gray substance. These rami branch and anastomose with each other and form loops which anastomose with the central vessels from which they arise, or with neighboring vessels. In a plane just above the anlagen of the ventral horns each of the central vessels ends blindly, or divides into a caudal and a cranial ramus, which anastomose with adjacent similar rami and form irregular longitudinal vessels. By other anastomoses among the central vessels, a longitudinal plexus is formed, which covers very coinpletely the lower half of the lateral side of the ependymal layer.

A comparison of figures 3 and 5 shows how closely the form and arrangement of these capillaries corresponds to that of the future central arteries and veins. Besides these main capillaries two smaller lateral groups are present at this stage. These may be called the ventro- and dorso-lateral groups, and later fonn peripheral arteries and veins. Both groups enter the cord from the capillaries on the lateral surface between the dorsal and ventral nerve roots. The ventro-lateral group enters at the level of the dorsal extremities of the central vessels, and courses medially and anastomoses with them. Occasionally the ventro-lateral group gives off rami which extend into the anlagen of the ventral horns. The capillaries of the dorso-lateral group are confined to the dorsal two-fifths of the cord, and although they anastomose with each other at this stage, they do not anastomose with the central or ventro-lateral capillaries. They course medially and



dorsally along the ependjana, ending blindly or forming loops, but do not reach the dorsal surface.

As development proceeds, the lateral groups of capillaries shown in figure 5 spread dorsally and ventrally and capillaries enter the cord from the periphery. With the exception of the above-mentioned dorso-lateral group of capillaries, all the vessels entering the sides of the cord grow toward a common center, namely, an area on the lateral border of the ependyma about half way between the dorsal and ventral surfaces. The dorso-lateral group of capillaries which are shown in the same figure send rami toward this center after the embryo attains the length of 14 mm.

The vessels from the dorsal surface grow ventrally along the ependyma and unite with the dorsal rami of the primitive arterial tract. This union continues the plexus on the lower part of the epend>Tna dorsally so that the ependyma except below the floorplate and above the roof-plate, is entirely surrounded by a capillary plexus. A thick transverse section of the cord of an embryo of 25 mm. shows this plexus with numerous vessels extending from it laterally at right angles. These lateral vessels are joined together by dorso-ventral rami. This picture is characteristic of the cord until the embryo reaches the length of 30 or 35 mm. when it is changed by other peripheral vessels meeting the ependymal plexus obliquely and by the branching of the vessels in the anlage of the gray substance.

By this time the central arteries from both the ventral surface (ventral central arteries) and from the dorsal surface (dorsal central or dorsal peripheral arteries) have become quite large, although the latter do not nearly equal the size of the former. The ventral central arteries have formed more longitudinal loops similar to those shown in figure 5. They are separated more and more from each other, owing to the growth of the cord, and as this separation continues the longitudinal vessels grow in length.

In embryos of 35 to 40 mm. in length the peripheral arteries from all sides together with the lateral rami of the central arteries have formed a dense plexus in the gray substance, although the white substance contains only the peripheral arteries running through it, and the short branching rami given off at right angles


from them. By the time the embryo reaches a length of 50 mm. the capillaries in the white layer have much the same appearance as those of the full term fetus, except that in the latter they branch and anastomose more freely and the growth of the cord tends to separate both the peripheral vessels and the central vessels. Embryos of 75 to 100 mm. in length show the arteries in the cord quite as completely developed as in the 240 mm. embryo.

The posterior rami of the primitive arterial tract in the ventral part of the cord of embryos of 12 to 15 mm. are more numerous than the central arteries in the 240 mm. embryo which are formed from them.

The veins within the cord develop in the same planes as the arteries, and from the same plexus of capillaries that form the latter. They may be called the central and peripheral veins corresponding to the similarly named arteries. They are shown in figures 3 and 4 in a fully developed condition.


The dorsal rami of the primitive arterial tract, and other rami from the capillaries in its immediate vicinity enter the cord, forming an undifferentiated capillary plexus (fig. 5) and this plexus later becomes differentiated into arteries and veins. It was not found, as stated by Sterzi for the sheep, that each dorsal ramus of the primitive arterial tract grows into the cord, and forms a loop, giving rise to a vein which grows back along the artery to the ventral surface.

The dorsal rami of the primitive arterial tract are more numerous than the ventral central arteries which develop from them.

Sterzi reports solid blood-vessels in the cord of sheep of 5.5 mm. and hollow ones in those of 6.6 mm. In pig embryos the bloodvessels within the cord seemed to appear first as hollow vessels. These are seen first in embryos of 7.5 nmi. in length.

The "tracti arteriosi laterales" of Sterzi, are the dorso-lateral arteries of this and postero-lateral of other papers, and are the posterior spinal arteries of human descriptive anatomy. Evans shows these two tracts first united by medial anastomoses in a


pig of 8.5 nun. in length, but many such anastomoses are to be found in embryos as small as 7.5 mm. in the cervical and thoracic regions, and one specimen of 6.2 mm. showed them in the cervical region.

The embryos described in this paper show the mid- ventral and mid-dorsal surfaces of the cord to be covered with blood-vessels at a somewhat earlier stage than has been described.

As reported by Sterzi ('04) and Evans ('09), blood-vessels first appear on the ventro-lateral surface of the cord, then on the ventral, then on the dorso-lateral, and finally on the dorsal surface.

The blood-vessels on the cord are continuous with those in the mesenchyma surrounding it until the membranes of the cord are formed.

It is generally stated in textbooks of human anatomy that the spinal artery arises from the vertebral arteries, and is reinforced by segmental spinal arteries. It il rather to be considered that this artery arises from the segmental spinal arteries, and anastomoses with, or is reinforced by, the vertebrals.

The term median dorsal is suggested for the artery present in places in the median dorsal line of the spinal cord.

My thanks are due to Dr. Richard E. Scammon for his constant interest in the progress of this work, and for his many helpful criticisms.



Adamkiewicz, a. 1881 Die Blutgefiisse des Menschlichen Riickenmarkes. I. Teil: Die Gefasse der Ruckenmarksubstanz. Sitzungsber. k. Akad. Wiss., Wien, Math.-Xaturwiss. Kl., Bd. 84, iii. Abt.

1882. II. Teil: Die Gefasse der Riickernmarksoberflache. Sitzungsber. d. k. Akad. d. Wiss., Wien, Math.-Xaturwiss. Kl., Bd. 84, lu. Abt.

DoRELLE, P. 1911 Rapporti tra encefalomeria e vascalarizzazione. Del cervello embrionale. Ricerche Lab. Anat. Xorm. R. L'niv. Roma, voL 15.

EvAXS, H. M. 1909 On the development of the aortae, cardinal and vunbilical veins, and other blood vessels of vertebrate embryos, from capillaries. Anat. Rec, vol. 3, p. 498.

1909 On the earliest blood vessels in the anterior limb buds of birds, and their relation to the primary subclavian arterj'. Am. Jour. Anat., vol. 9.

1912 Development of the vascular system. In 'Human embryology,' Keibel and Mall, vol. 2.

His, W. 1887 Zur Geschichte des Menschlichen Riickenmarkes und der Xervenwurzlen. Abhandl. der Konigl. Sachs. Gesellschaft, der Wissenschaften 22. Math.-Phys., Classe 13, Leipzig.

HocHE, A. 1899 Vergleichenden-anatomiches iiber die Blutversorgungen der Ruckenmarksubstanz. Zeitschr. Morph. u. Anthrop., Bd. 1.

HoF.MAX, M. 1900 Zur vergleichenden Anatomie der Gehirn und Riickenmarksarterien der Vertebraten. Zeitschr. Morph. u. Anthrop., Bd. 2.

Kadyi, H. 1889 Ueber die Blutgefasse des menschl. Ruckenmarks. Wien. (Also Denkschr. Math. Xatunv. Kl. Akad. Wissensch. Krakau).

Ross, J. 1880 Distribution of the arteries of the spinal cord. Brain, vol. 9.

Smith, H. W. 1909 On the development of the superficial veins of the body wall in the pig. Amer. Jour. Anat., vol. 9.

Sterzi, G. 1904 Die Blutgefasse des Ruckenmarks. Anat. Hefte, Bd. 24.