Z 8vLO0600 I9ZI € 


ll il 


NIV. OF 


RONTO 
RBRARY 


Digitized by the Internet Archive 
in 2008 with funding from 
Microsoft Corporation 


http://www.archive.org/details/earlystagesofvasO3schuuoft 


ar 


‘fi 
jt 1 , 


( 


y 


} 


h 


Nii OM 
j Tih ti 
1 he Uk we 
al Pilly i Fn, 
Ree We 
wits 
x i 
+ ' 
' Pi 
y 1 1} 
v4! Lyn ‘ <4 
y { . 
| iy ae 
{ 
+ 
” 
4 
7 
F 
’ 
\ 
as 
’ 
BP a 
ee 
fi 
tay 
JN 
4 


ag 


MEMOIRS 
OF 
THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY 


No. 3 


EARLY STAGES OF VASCULOGENESIS 
IN THE CAT (FELIS DOMESTICA) 
WITH ESPECIAL REFERENCE TO 
THE MESENCHYMAL ORIGIN 
OF ENDOTHELIUM 


H. vON W. SCHULTE 


FROM THE ANATOMICAL LABORATORY OF COLUMBIA UNIVERSITY 


PHILADELPHIA, PA. 
1914 


VASCULOGENESIS IN THE CAT 


Of the many problems that engage the attention of students of 
the vascular system, none is more fundamental, and hardly any 
more perplexing and confused today, than that of the origin of 
the simple squamous epithelium (endothelium) which is the com- 
mon lining of all vascular channels, whether haemal or lymphatic. 
It has been held, and the opinion of late has been zealously propa- 
gated, that so far as the vessels of the body are concerned, their 
endothelium arises solely from the already formed endothelium 
of the extra-embryonic vessels, or if there is any production of 
endothelium within the body of the embryo, it is for so brief a 
period and so inconsiderable in amount as to be theoretically neg- 
ligible.! For the supporters of this view Rabl’s dictum, “Endothel 
stammt nur von Endothel” is still a satisfactory summary of the 
facts. A logical sequence of this opinion is the doctrine of the spec- 
ificity of endothelium, obviously a much broader generalization, 
which takes endothelium out of the series of mesenchyme deriva- 
tives and separates it absolutely from the blood, should that prove 
to be of mesodermal origin. Consequent thinkers who support this 
view have seen the importance of assigning an entodermal origin 
to both blood and endothelium. Could this be established, it would 
lend a degree of antecedent probability to the doctrine of the 
angioblast but, strictly speaking, it is not a necessary postulate, 
for a tissue arising diffusely from the mesenchyme, may as de- 
velopment proceeds, be confined to definite localities or even 
ultimately be restricted to homoplastic proliferation. Such a 
tissue would be specific if it yielded no heterogeneous products, 
and I feel myself that this latter property enters more largely 

1 Minot, Evans, and Sabin in Keibel and Mall, Handbuch d. Entwickelungs- 
gesch. d. Menschen, Leipzig, 1912. These writers have made a clear and emphatic 
statement of the doctrines of the Angioblast and of the specificity of endothelium, 
and presented the evidence which may be adduced in favor of these views very 
completely. That they have not considered all the available evidence, however, 
is apparent from their all but complete omission of the results of recent haemo- 


tology (since 1909), which would indicate the wisdom of some reserve in accepting 
as final generalizations which neglect such extensive and important data. 


3 


AS VASCULOGENESIS IN THE CAT 


into our idea of specificity as applied to tissues than the commu- 
nity of origin of their elements. It would seem, therefore, that 
the advocates of specificity have charged themselves rather to 
prove that endothelium, once formed, produces only endothelium 
and never any other elements, as say blood or connective tissue, 
than that it is peculiar in its origin. 

On the other hand, it has been held that endothelium is simply 
a derivative of the mesenchyme from which it arises diffusely 
by a multitude of separate anlages. This view does not under- 
take a general interpretation of the mesoderm and mesenchyme 
and the apparent variations of their mode of origin as determined 
by the yolk-content of the ovum, but merely states that endo- 
thelium arises out of elements which have once formed a part 
of the general mesenchyme complex. A development of this 
point of view is the adaptive theory of endothelium, which holds 
that endothelium is a modification of mesenchyme, and by no 
means a stable one at that, but owes its presence to hydrody- 
namic forces, and when these cease to operate, endothelium dis- 
appears as endothelium and reverts to mesenchyme, or gives rise 
to products indistinguishable from those of mesenchyme. Endo- 
thelium is taken to be a form which mesenchyme assumes in 
certain positions which can not be retained with the loss of position 
and is then analogous to the biaiometamorphoses of plants grown 
under unnatural conditions, which are lost when they are returned 
to their normal environment. 

It is conceivable that some supporters of the adaptive theory 
of the origin of endothelium may not feel themselves obliged 
to maintain a variety of products on the part of endothelium, 
just as it is not to be required of those who hold the specific 
doctrine that they should demonstrate an entodermal origin. 
The former might then hold endothelium to be a derivative rather 
than a modification of mesenchyme; it would satisfy them to 
show the multitude of its separate anlages, their wide diffusion 
through the mesenchyme, in the somatopleure as well as in the 
splanchnopleure, within the body of the embryo as well as in 
the yolk-sac, and their continued formation in the embryo dur- 
ing an appreciable period of development, and along the major 


VASCULOGENESIS IN THE CAT 5 


vascular lines. This done, they could allow endothelium to 
manifest such specificity as it might acquire, as in the case of 
other mesenchymal derivatives, cartilage for example, or cross- 
striated muscle, which increase by accretion of new elements 
as well as by the proliferation of those already formed, and yet 
once formed, are specific to the degree that they probabably do 
not give rise to other types of connective tissue. 

In thus comparing the angioblastic with the mesenchymal, the 
specific with the adaptive views of endothelium, I have alluded 
to certain differences of opinion regarding the supposed mode 
of its wide dissemination through the connective tissue of the 
body, which are fundamental and may now be contrasted. The 
doctrine of specificity holds this process to be essentially one of 
growth. The period of development of endothelium is amaz- 
ingly short, and the site, for most entertainers of this view, is the 
yolk-sac at some distance from the embryo. For others there is 
an abstract possibility that certain parts of certain embryonic 
vessels may develop in situ, but it is not unfair to say that they 
consistently tend to minimize and curtail the developmental 
period of endothelium. The important part of their doctrine, 
the part which has engaged their chief attention, and is presented 
with most conviction, is the growth of the endothelium by solid 
sprouts from the ends of hollow injectible vessels. This is the 
sole source of endothelium (the case of a few embryonic vessels 
apart), and so far as I am aware, absolutely the only evidence 
of its specificity. 

The adaptive theory finds support in the presence, through 
a considerable period of ontogeny, of numerous separate sacs or 
vesicles of endothelium, definitely localized in positions subse- 
quently occupied by continuous vessels, and in the gradual en- 
largement and ultimate fusion of these vesicles to form the ves- 
sels. Further, it notes that the vesicles are preceded by minute 
clefts or spaces in the mesenchyme bounded by ordinary mesen- 
chyme cells, which subsequently flatten and eventually form a 
continuous layer as the vesicles enlarge. This view thus extends 
the period of development, and enlarges the area in which it is 
active to practically the whole body of the embryo. It does not 


6 VASCULOGENESIS IN THE CAT 


deny that endothelial cells undergo mitosis, that vessels after 
formation lengthen, or that they may branch and send out sprouts; 
for it by no,means implies that endothelium does not grow, but 
it relegates these phenomena to a later period in the life of the 
embryo, and endeavors to distinguish between a cytomorphosis 
terminating in a definable type of cell (development, genesis, 
anlage, das Werden) and a proliferation and extension of a fully 
formed tissue (growth, das Wachsen). 

One further point, the great vascularity of the embryo as com- 
pared with the adult makes the fate of the temporary vessels 
and their endothelium of some importance to a general theory 
of this tissue. The adaptive view considers that they revert to 
mesenchyme, the doctrine of specificity that they disappear. 

Divergent principles of interpretation rather than differences 
in method and technique underlie, I believe, these discrepant 
theories, although I am aware that the contrary has been em- 
phatically proclaimed, and in particular that the method of. in- 
jection has been held to be the ultimate criterion in these matters. 
Without discussion at this point of what I believe to be the claims 
of enthusiasm, it is still possible to admit the objective reality of 
the positive data obtained by this method, without attaching 
finality to its negations. One may admit that where the injec- 
tion passes, there is a channel, without denying that beyond the 
limit of injection there are forming channels, or considering it 
proved that injected channels are always lined by a complete 
endothelium. One may admit that the blind terminals of ves- 
sels are actual terminals, without assuming that they are always 
sprouts, for they may be regressive. And above all, one may 
admit the frequency of extravasation, without holding it to be 
ohne weileres an artifact. But if these reservations are allowed, 
or even conceded to be still matters of free and unbiased inquiry, 
the positive results of the injection method will form but a slender 
basis for such imposing doctrines as those of the specificity of 
endothelium, and of the angioblast. 

From the time of the establishment of the circulation, the body 
of the embryo and some of its membranes are supplied with con- 
tinuous channels which distribute the circulating medium, and 


VASCULOGENESIS IN THE CAT of 


return it to the heart. From this period throughout life, such 
continuous vessels are present and functionally active. In the 
adult body it has long been customary to inject them for conven- 
ience in demonstrating their course. In the foetus also, such 
procedures are familiar and have been held to demonstrate the 
plan of organization of the circulatory channels at the period of 
examination. Now it is obvious that the injection at earlier 
periods, even in young embryos, is a fact of the same order, how- 
ever admirable the technique or delicate the procedure—it sim- 
ply demonstrates continuous channels, which it may be presumed 
are functionally active. It reveals the extent and conformation 
of the formed and functional vascular system and enables the 
observer to follow its successive stages from the period at which 
injection first becomes practicable. In such studies, however, 
injection has no monopoly, but it shares its data with the method 
of direct observation of living tissue when the object is small and 
transparent, and with the exact but laborious method of recon- 
struction. On the other hand, it is idle to pretend that the in- 
jection method throws any light on areas beyond the limit of the 
actual injection, or affords the slightest presumptive evidence 
of their histological condition, in particular, as regards the pres- 
ence or absence of small discrete vesicles of endothelium. 

The dilemma is simple; we may recognize or refuse to recog- 
nize the uninjectible periphery. If we recognize it, the method 
of injection becomes not only partial and incomplete, but subor- 
dinate to the methods which reveal all of its findings and in addi- 
tion enlarge our field of observation. It ceases to be a critical 
method. If we refuse to recognize the periphery beyond our 
injection, we have begged the question at issue, our vessels are 
growing not only in a foreign tissue, as the angioblast doctrine 
believes, but in a theoretical vacuum and the only possible source 
of increment is the proliferation of their own elements. We 
reach the only conclusion the exclusive study of organized ves- 
sels by any method would allow, the formation of the new out of 
the old by sprouting. A comparison may perhaps serve to il- 
lustrate the inconvenience of taking continuity of lumen as the 
critical index of structure. Had the metanephros chanced to 


8 VASCULOGENESIS IN THE CAT 


be first studied by the method of injection, we should have a his- 
tory of the ureteric bud from the standpoint of its lumen. It 
would have sprouted, giving branches of many orders, and by 
growth in continuity its arborizations would in time form their 
definitive terminals, the Bowman’s capsules. Should an inves- 
tigator, with a belated confidence in slides and reconstruction, 
have ventured to observe the renal vesicles, he might have been 
told he had been misled by artifacts. Had he objected that they 
were constant, had definite structural characters, and were local- 
ized with exquisite precision, he might at length have obtained 
their tardy recognition. But it is probable, for doctrines as well 
as masses have momentum and inertia, that he would then have 
been informed, that his vesicles were but separated sprouts, be- 
cause nephrothelium comes only from nephrothelium. And 
these arguments would be as justifiable, nay as necessary, in the 
one case as in the other, if ultimate reliance is placed on the in- 
jection method which can only reveal continuity of lumina. The 
logic is good, the argument the typical scholastic deduction of the 
special case from the general principle, but the principle is intu- 
itive and arbitrary, not inductive. 

The blind terminals of vessels in the mammalian foetus are a 
favorite topic for study by injection, but so far as I am aware, 
the possiblilty of their regressive nature has never been seriously 
considered; it has simply been assumed that they were always 
growing sprouts. I desire here rather to note the gap in the evi- 
dence than to question this mode of growth as a factor in the ex- 
tension of formed vessels in late embryonic stages and in the foetus. 

W. G. MacCallum? (’02) made a careful and excellent histo- 
logical study of lymphatic terminals in the skin of pig foetuses, 
from 6 cm. to 15 em. in length, and reached the conclusion that 
the lymphatics did not communicate with tissue spaces. From 
their extremities project strands of cells in a single or double row 
which run for some distance in the embryonal connective tissue, 
or join another lymphatic. He thought that few of these cells 
were in syncytial relation with adjacent fibroblasts. Into the 

24902. Arch. f. Anat. u. Entwickelungsgesch., p. 273. 


3 Exponents of the doctrine of specificity seem inclined to lay stress on this 
point as indicating the independence of endothelium. In view of the interdermal 


VASCULOGENESIS IN THE CAT 9 


bases of the cords, the injection entered in a thin stream for a 
short distance revealing a cleftlike diminishing lumen. 

Bartels (’09)4 has given figures which closely correspond to 
those of MacCallum, but he considers the interpretation of the 
cell-cords as sprouts debatable. On the other hand,.Miss Sabin’s 
figure of lymphatic terminals in a 5 em. pig (02, fig. 2)° represents 
very different conditions; here are blind, blunt terminals and no 
sprouts. She states, however, that occasionally from one of these 
blunt ends, the “injecting fluid can be forced out into a long 
thread-like process’”’ which is a sprout and represents the process 
of growth. 

She has plainly described a somewhat different structure from 
the solid uninjectible strands of MacCallum and Bartels, though 
not improbably a phase of the same process. A number of in- 
terpretations suggest themselves: 

1. Miss Sabin’s, that the slender injectible processes are sprouts 
growing from formed vessels and that this is the only mode of 
their extension at this period. 

2. That in view of their infrequency, we are dealing with the 
inception of autonomy on the part of the endothelium, which in a 
few places is sprouting, while in most it is advancing by accre- 
tion, as has been definitely proved for the thoracic duct andthe 
systemic lymphatics in general in embryos.° 


cytodesmata described by v. Szily, Studniéka and others, it is of course void of 
theoretical importance in this sense, for if in general, elements of different germ 
layers may stand in syncytial relation to one another, the specificity of endothelium 
is in no way impaired by a like continuity with the mesenchyme, which in fact it 
does possesss in the embryo. If in the foetus a separation does actually occur, 
it ought to be interpreted in the light of the general tendency of the organism to 
resolve its synctia into independent cells, and not as a peculiar sign of the speci- 
ficity of endothelium. 

41909. Das Lymphgefiisssystem. Jena. p. 12. 

51902, Am. Jour. Anat. vol. 1, p. 367. 

6 Huntington & McClure, 1907, Am. Journ. of Anat., vol. 6, Abstr. Anat. Rec., 
vol. 1, pp. 36-41. Huntington, 1908, Anat. Rec. vol. 2, pp. 19-45; 1910, Anat. 
Rece., vol. 4, pp. 339-423; Compte Rendu 16 Congrés Internat. de Med. Sect. 1, 
fase. 2, pp. 127-142; Anat. Anz. Ergiinz. z. Bd. 37, pp. 76-94; 1911, Wistar Mem: 
No. 1; Anat. Rec., vol. 5, pp. 261-276. Stromsen, 1912, Anat. Rec., vol. 6, pp. 
343-356. Kampmeyer, 1912, Am. Jour. Anat., vol. 13, p. 401, and Anat. Rec., 
vol. 6, p. 223; McClure, 1912, Anat. Rec., vol. 6, p. 238. Tilney, 1912, Amer. 


10 VASCULOGENESIS IN THE CAT 


3. That the sprouts are regressive and that here, as in other 
tissues (bone, striated muscle), progressive and retrogressive proc- 
esses co-exist. 

4. That the ‘sprouts’ are the last stages in the annexation of 
vasoformative cells which have built themselves into cords, 
joined the lymphatics and are now establishing their lumen. 
Only in this last stage would they be accessible to injection. 
This interpretation is suggestd by Bartels (’09) and is of consider- 
able import. Referring to certain unicellular sprouts of MacCal- 
lum, he says: 

Gerade, dass es einzelne Zellen sind, welche die Striinge bilden, 
scheint mir eher fiir den Modus der Aufreihung dieser Zellen durch den 
zentralwirts gerichteten Saftstrom als fur eine zentrifugal gerichtete 
Sprossung zu sprechen. Doch will ich darauf, als auf eine nach dem 
bisher Vorliegenden nicht sicher zu entscheidende Sache, ebendsowenig 
eingehen, wie auf die Beurteilung des Verhaltens der benachbarten 
Gewebsbestandteile zu den Gefissen. Beides scheint mir weiterer ein- 
gehender Untersuchungen zu bediirfen.’ 

None of these observations, important as are the data they have 
collected, have given us the means of interpretation, once the 
increment of vessels is conceived as a minute process which must 
be treated in terms of cells and cell modifications. For this it 
is necessary to work in confined areas of well understood topog- 
raphy and with a minutely graded and continuous series of em- 
bryos. If cords of single or double rows of cells are to be studied 
developmentally, we must at least give ourselves the chance 
of seeing cell add itself to cell, or the sprout advance from the uni- 
cellular to the multicellular condition, and not work at wide in- 
tervals of time or with material from various regions where only 
gross comparisons are possible and minute processes may 
altogether escape us. 

From the standpoint of possible retrogression, the work of 
Spuler’ and of Fuchs? is interesting and their aloofness from this 
particular problem gives their testimony additional importance. 
Journ. of Anat., vol. 13, pp. 193-220. Miller, A. M., 1913, Amer. Journ. of Anat., 
vol. 15, pp. 131-168. Allen, W. F., 1913, Quart. Journ. Micr. Sci., vol. 59, pp. 
309-360. 


7 Loc. cit., p. 45. 81892, Arch. f. mikr. Anat., Bd. 40, p. 530. 
91903, Anat. Hefte. Band. 22, p. 97. 


* 


VASCULOGENESIS IN THE CAT 15 


They were writing in refutation of Schaffer’s derivation of ery- 
throcytes from vasoformative cells and of Minot’s!? expansion 
of this observation into an absolute distinction between nucleated 
and non-nucleated red corpuscles, and his classification of blood 
in the vertebrates on this basis. 

Spuler showed clearly that the erythrocytes in question and 
the vasoformative cells containing them were degenerative. 
Fuchs, working with the omentum of the guinea pig in the first 
few days after birth, confirmed these observations. The vaso- 
formative cells of Schaffer" and Ranvier” he found usually con- 
nected with functional vessels by narrow cell-strands, with or 
without lumina. In a few cases the strands were interrupted. 
This he attributed to rupture occasioned by unequal growth 
in the omentum. He subscribes to the production of vessels 
from vasoformative cells but makes a distinction between these 
cells and the erythrocyte containing cells which he found, as 
Spuler had done before him, usually connected with vessels. 
These, in view of their. degenerating contents, may fairly be 
deemed retrogressive. Their occasional occurrence free in the 
connective tissue then becomes a late and not an early phase in 
their history. The order of events in a degenerating capillary 
would appear to be a collapse of its lumen at some intermediate 
point or points, between or beyond which the vessel and content 
persist for a time; the solidification and the reduction of the col- 
lapsed regions to double or single rows of cells and eventually 
long protoplasmic strands, which finally give way, and the 
remnants of lumina with their contents of blood cells and their 
wall of endothelium become isolated in the connective tissue. 
Fuch’s excellent illustrations are astonishingly like the sprouts 
of MacCallum and the cell cords of Bartels, differing only in 
the cellular contents of the vessels (cf. Fuchs, fig. 2, with Bartel’s 
’09, fig. 12, and Fuch’s, figs. 4, and 3, with MacCallum, figs. 6 
and 7, respectively). 

Identical appearances are observed in definite regions of young- 
er embryos of the cat incident to vascular processes which are 


101890, Anat. Anz., Bd. 5, p. 801. 11874, Mo. Mic. Journ., vol. 11, p. 97. 
121874, Arch. de Physiol. 


12 : VASCULOGENESIS IN THE CAT 


known to be regressive, because they are preceded by blood con- 
taining plexuses and functional vessels and accompany their 
resolution. Huntington™ has noted the condition in connection 
with the perivenous plexuses: 


Some of the elements . . . . develop into permanent tributaries 
of the main veins. Others undergo a process of separation from the per- 
manent functional channels and degenerate. In many cases their 
blood cell contents break down and are eliminated, while their endothelial 
lining appears to revert to the indifferent type of the embryonic meso- 
dermal cell. 

Thus in embryos between 13.5 and 16 mm. many striking instances 


of this reversion are to be observed. [And further] . . . . these 
detached and retrograding channels . . . . constitute lines around 


which the most active primary lymphatic organization of the mamma- 
lian embryo centers. 

I am far from wishing to claim that ‘sprouts’ are never grow- 
ing points, I would only instance cases where they certainly are 
regressive, and express the doubt that their injectibility or lack 
of it affords a proof of their origin from endothelium. The sub- 
ject is a promising one and would repay unprejudiced observation 
under conditions where the direction of the process might be 
definitely ascertained—in a circumscribed area and with a closely 
graded series of foetuses. It may be that their formation in the 
lymphatic system will prove to be a late phenomenon and inci- 
dent to the ultimate sealing off of the lymphatics from the tissue 
spaces, if a closed system of lymphatic capillaries actually exists. 

The question of extravasation is perhaps the most difficult of 
the many problems the injection method has raised, and is giving 
us a literature of as much subtilty in connection with develop- 
ing vessels as it has done in regard to the circulation in the adult 
spleen. The point at issue is not the permeability of the vessel 
wall which may be taken as proved beyond controversy by the 
facts of metabolism, and the passage of erythrocytes and leuco- 
cytes, as well as by the results of injection. The real difficulty is 
as to how we may accept these indisputable facts and yet main- 
tain that when extravasations occur during injection they prove 
a discontinuity between the lumen of the vessel and the tissue 
space In mesenchyme or in connective tissue. 


131911, Memoirs of the Wistar Inst. of Anat. and Biol. No. 1., Philadelphia. 


VASCULOGENESIS IN THE CAT 13 


The positive conclusion that extravasation means communi- 
cation between vessel and tissue space seems justified when the 
injection is controlled by a manometer and practised at physio- 
logical or subnormal pressure, as in Herring and Simpson’s" in- 
jections of the liver or Pustoiwoitow’s" of the spleen. If, as in 
these cases, the injected substances escape under less than normal 
pressure it is permissable to infer a real continuity, though strictly 
speaking it is necessary to confirm this by direct observation of 
the tissue (vide infra). 

But if the pressure is not thus controlled, it will always be pos- 
sible to interpret the extravasation as an artifact, for it can always 
be argued that rupture occurred under excessive pressure. Cen- 
trifugal injections of lymphatics are obviously subject to this 
doubt, so that the presence here of extravasation is only suggest- 
ive and not demonstrative of communication between these ter- 
minals and the tissue spaces. 

Other methods of control than that of manometric measure- 
ment of the pressure can not be freed from subjectivity. To 
observe, for example, the passage of the injection with a binocu- 
lar and discontinue when it begins to extravasate is arbitrary; 
to succeed in stopping at the moment when extravasation has 
just begun, when the terminals have still but a mossy appear- 
ance, shows only the point to which the observer’s dexterity 
may be developed. One may, with practice, heat a thermometer 
rapidly till the mercury rises to a chosen point, and become able 
to discontinue just as the point is reached. The repetition of 
the act may increase exactitude, but will not prove the absence 
of luminal continuity beyond, however we might argue, when 
we had just passed our selected point, that we had ruptured a 
delicate and invisible membrane. 

Nor is even the “explosive” (MacCallum) rupture of the vessel 
wall at some intermediate portion of its course a proof that its 
terminals are closed. A very moderate experience with the in- 
jection of adult material will convince anyone that large veins 
may rupture before their radicles are filled, and the rupture may 


41906, Proc. Roy. Soc. B., vol. 78. 
16 Arch. f. Anat. u. Entwickelungsgesch., p. 219. 


14 VASCULOGENESIS IN THE CAT 


be attended with sensations, even emotions, on the part of the 
observer. Yet it cannot thereby be proved that the vessels end 
blindly; and we ought not to assume, because of the delicacy of 
embryonic structures that their physical properties are different 
in kind, or obey other general laws of stress and strain than larger 
tubes which are more accessible to examination. 

But while the results of injection demonstrate the permeabil- 
ity of the vessel wall, as do in life the passage of cells and fluids, 
they do not afford us the opportunity of ascertaining the structure 
of the wall or the nature of its orifices. They simply give us the 
concept of permeability, which it is not permissible, without fur- 
ther grounds, to translate into porosity. Every fact of extrava- 
sation can be admitted and the integrity of a completely closed 
vessel wall still be maintained if we introduce the concept of a 
structureless (invisible) membrane, the resistance of which to 
pressure is practically nil. 

Helly’® has shown this with admirable clearness in his interest- 
ing work upon the vessels of the spleen. He confirmed previous 
‘ observations of the passage of injections through the vessel walls, 
whether solutions, suspensions, colloidal masses, or foreign blood 
corpuscles were used. He further washed out the red blood cor- 
puscles from the pulp spaces by saline irrigation, and caused their 
numbers in the pulp to increase by retrograde injection (obstruc- 
tion of the splenic vein) and finally in sections of 5 « observed 
both erythrocytes and leucocytes in passage through the walls of 
the sinuses. Yet he considered the wall complete. 

Was zunichst den Widerstand anlangt, welchen die Gefiaisswand dem 
Durchtritte fester und fliissiger Bestandteile entgegenzusetzen vermag, 
so ist ersichtlich, dass derselbe bei den venésen Capillaren nur sehr gering, 
an gewissen Stellen der Wand iiberhaupt fast gleich null ist; gleicht sie 
doch, von der Fliche betrachtet, sehr einem Gitter, dessen Liicken viel- 
fach gross genug sind, um ein rotes Blutkérperchen ohne jede merk- 
liche Formverinderung durchtreten zulassen. Dem zwischen beiden 
Bestandteilen des Gitters,—den inneren, parallel zur Lingsachse des 
Gefiisses angeordneten, stabformigen Endothelzellen und den dusseren, 
quer um dasselbe verlaufenden Kreisfasern,—befindliche unmessbar 
diinnen strukturlose Hiutchen kann wohl kein irgend nennenswerter 
Finfluss im Sinne einer Behinderung der Diapedese zugeschrieben wer- 
den und dies umso weniger, als das gedachte Hautchen sehr hinfallig ist 


161903, Arch. f. mikr. Anat., Bd. 61, p. 245. 


VASCULOGENESIS IN THE CAT 15 


und ungemein leicht zerstért wird. (p.292). . . - Die Milz hat ein, 
tiberall von einer regelmassigen Endothelschichte ausgekleidetes, daher 
geschlossenes Gefasssystem mit sehr durchlassigen Wandungen. 

Den ich habe einerseits gezeigt, dass man unter Anwendung geeigneter 
Methoden sich davon iiberzeugen kann, dass der kreisende Blutstrom 
iiberall seinen Weg durch Capillargefiisse nimmt, deren wand keine 
Oeffnungen in Gestalt bestindiger Liicken fiir den Durchtritt rother Blut- 
korperchen zeigt, welcher dort, wo er der unmittelbaren Beobachtung 
zugiinglich ist, als wahre Diapedese aufgefasst werden muss; anderer- 
seits habe ich den Durchtritt von Leukozyten durch die geschlossenen 
und unversehtren Gefadsswande hindurch als thatsichlich vorhandene 
Erscheinung nachgewiesen.!’ 


It will thus be seen that Mall’ fails to appreciate the subtilty 
of the question and at the same time misrepresents Helly in citing 
him as a believer in pores in spite of his emphatic statement to 
the contrary: 


recent authors are of one opinion regarding the large 
pores in the capillary walls. It appears that these openings are smallest 
according to Thoma, larger according to Helly, still larger according to 
Mall, and so large that they communicate most freely with the pulp- 
spaces according to Weidenreich. 


Helly very carefully distinguishes between permeability and 
porosity, refuses to deduce the details of structure from physical 
and physiological properties, and accepts the structureless mem- 
brane of v. Ebner because he believes he has seen it. Mollier'® 
correctly represents Helly’s position and criticises it with acumen: 


Ob es freilich noch praktischen Zweck hat, von einer geschlossenen 
Rohre zu sprechen, wenn sie fast’ohne Widerstand alle zelligen Elemente 
in unbegrenzter Menge hindurchlisst, kénnte zweifelhaft erscheinen. 
Weidenreich hilt jedenfalls eine solche Annahme fiir wertlos, wie ich 
aus seiner temperamenty ollen Ausserung entnehme: ‘Ein Sieb ist doch 
kein Topf.’ 

Ich meine aber, er vergisst hier, dass ein wesentlicher Teil unserer 
Vorstellung iiber den V organg der Diapedese der ist, dass Zellen durch 
die Capillarwand den Kreislauf verlassen, ohne dass dabei nennenswerte 
Mengen des Plasmas mit austreten. Der Kreislauf der Blutfliissigkeit 
wird also kaum beeinflusst. 

_ Von diesem Gesichtspunkt aus hat die Annahme Hellys nicht nur 
ihre Berechtigung, sondern auch einen ganz bestimmten technischen 
Wert. Ob die Bezeichnung Diapedese auch fiir den Durchtritt roter 


471902, Arch. f. mikr. Anat., Bd. 59, p. 93. 
181903, Am. Jour. Anat., vol. 2, p. 315. 
191911, Arch. mikr. Anat., Bd. 76, p. 608. 


16 VASCULOGENESIS IN THE CAT 


Blutkérperchen ohne lokomotorische Eigenbewegung gebraucht werden 
kann, dariiber will ich hier nicht weiter sprechen. Wir haben bloss 
damit zu rechnen, dass Helly den Durchtritt roter Blutkérperchen (auch 
Froschblutkérperchen) durch die unversehrte Wand der venésen Milz- 
capillaren gesehen hat und dass er hierin em Hauptargument fiir die 
Lehre vom geschlossenen Kreislauf in der Milz erblickt. 


To summarize, the injection method gives us extravasation. 
This has a positive value when physiologic or subnormal pressures 
are used. It proves the permeability of the vessel wall. For 
the structure of that wall, whether complete or incomplete ana- 
tomically, and especially for the ultimate morphologic fact, the 
form and relation of its constituent elements, it is valueless and 
these questions must be settled by direct observation, and not 
by indirect conjecture founded upon physical or physiological 
experiment. To quote Mollier again, referring to the question 
of open or closed circulation in the spleen: 

Welcher Art diese Untersuchungen sein miissen, ergibt sich einmal 
aus dem Resultat der Injektionsversuche, welche lehren, das dieselben 
immer wieder sowohl zugunsten der einen wie der anderen Anschau- 
ung Beweismaterial geliefert haben. Sie sind also wohl nicht imstande, 


diese Frage zu entscheiden, und es ist besser, von ihnen zundchst ganz 
abziisehen.” 


His own brilliant solution of the problem of the structure of 
the sinus wall, is a ground of confidence to those who hold that 
structural problems are best approached by anatomic methods. 
In view of the complexity of structure and function in living 
beings, it will probably continue to be held the part of wisdom 
to determine their relation empirically and not attempt to deduce 
structure from the physical or even the physiological properties 
of organs and tissues. 

The positive findings of the injection method, that there are 
in the embryo continuous vascular channels, that these channels 
have ends, and their walls are very pervious, can therefore be 
accepted for they are corroborated by other methods of investi- 
gation. The negative propositions that have been advanced 
by users of this method, that the embryo has only continuous 
endothelium, that its only source is from sprouts, that the ends 
of vessels do not communicate with tissue spaces, are not proved 


20 Loc. cit. 


VASCULOGENESIS IN THE CAT il? 


by the evidence which is alleged to support them. The method 
is partial in that it excludes a vasculogenetic periphery beyond 
its limit of exploration, uncritical in that it can only reveal 
permeability and not the structure of the vessel wall, and its 
advocates are inconsistent in that without just grounds they in- 
terpret its only peculiar datum, the extravasation, as an artifact. 

Nor does the doctrine of specificity receive more support from 
observations intra vitam, if we distinguish between the growth 
and the formation of a tissue, and refrain from transferring proc- 
esses recorded of a given stage, perhaps a late one, to earlier 
periods in ontogeny. 

The observations upon the tails of amphibia, especially the 
tadpole, a classical object since the days of Schwann, have accu- 
mulated a rich fund of observations regarding the sprouting both 
of blood vessels and of lymphatics. Kdlliker,?! Rouget” and 
Golubew** followed the process, step by step, describing the hol- 
low-pointed outgrowth, its acquisition of a lumen and its fusion 
with adjacent sprouts. The sprouts were found to arise, not 
only in the vicinity of nuclei but also at a distance from them 
(K6llicker, Flemming); in the latter event nuclei were found 
subsequently to move into the sprouts (Fleming, Bobritski).* 
v. Ebner, in his edition of K6lliker’s Handbuch, summarizes these 
findings and interprets them correctly as phenomena of growth: 

Die Sprossung muss demgemiiss als ein Wachsthumsvorgang des 
Protoplasma angesehen werden, denn auch in den Fallen, wo der Kern 
an der Sprossungsstelle liegt, lisst sich zunichst nichts von einer Kern- 
theilung sehen. Man muss wohl die primiren Sprosse als sich aushohl- 
ende Protoplasmaknospen einer Endothelzelle ansehen. Erst sekundir 
erfolgt dann eine Mitose in der Zelle des Muttergefiisses und der eine 
Tochterkern tritt nachtraglich und nachdem er bereits wieder zu einem 
ruhenden Kerne geworden, in den Gefissspross, der unterdessen weiter 
gewachsen ist, ohne dass eine Zelltheilung erfolgt, da man, wie bereits 


erwahnt wurde, in jungen eben sich bildenden Kapillaren keine Zell- 
grenzen nachweisen kann.” 


*1 1846, Ann. S. se. natur., T. v. p. 91. 1886, Zeitschr. f. wiss. Zool., Bd. 48, p. 1. 
* 1873, Arch. de Physiol. 

231869, Arch. f. mikr. Anat., Bd. 5, p. 49. 

741890, Arch. f. mikr. Anat., Bd. 35, p. 283. 

25 1885, Central bl. f. mikr. Anat. 

*8oc. cit, p: 672: 


MEMOIR NO. 3 


18 VASCULOGENESIS IN THE CAT 


Schwann had believed that the capillaries were formed by the 
fusion of stellate cells, and Koélliker held at first the same view, 
but later abandoned it and considered the new formation of ves- 
sels in the tadpole’s tail wholly due to sprouting. 

Sigmund Mayer?? made an important addition to our knowl- 
edge of the vascular system in his thorough study of its regres- 
sive processes. He distinguished between blood containing and 
bloodless vessels in the tadpole’s tail, and pointed out that not all 
of the latter were to be taken as lymphatic, for many of them were 
degenerating blood vessels. The changes in the circulation in- 
cident to their collapse supports this view. His figures show many 
stages of their reduction and present striking resemblances to 
those of Fuchs, MacCallum and Bartels, before mentioned. Some 
confusion has been introduced into the terminology of these 
structures by Miss Sabin?’ who has invented the term ‘“‘Mayer- 
_ Lewis anlage”’ to include these degenerating vessels and the 
endothelial sacs described by Lewis in the course of veins in 
the mammal. It will be remembered by the readers of F. T. 
Lewis’? valuable paper that he found the jugular lymph sae 
composed of several cavities, at first connected but later separated 
from the veins. Similar sacs were found adjacent to many of 
the veins, for which he assumed a like origin although he is most 
careful to state that in no single instance, other than at the jugu- 
lo-subclavian junction, was there any evidence of connection 
with veins. 

Miss Sabin has further extended this concept to include the 
discrete mesenchymatous anlages of the lymphatics, described 
by Huntington and MeClure, which have at least a totally dif- 
ferent structure from the formations described by Mayer. Their 
walls present every degree of transition from mesenchyme to 
fully formed endothelium and every degree of communication 
between their cavities and the spacés in the adjacent mesenchyme. 
They do not run off into long tapering processes, and they do 
not contain degenerating erythrocytes. It is highly probable 

27 1885, Sitz-Ber. Akad. wiss. Wien. Bd. 91, p. 204. 


281911, Anat. Rec., vol. 5, p. 417. 
291906, Amer. Jour. Anat., vol. 5, p. 95. 


VASCULOGENESIS IN THE CAT 19 


that they are identical with the structures observed by Lewis, 
absolutely certain that they are not the same as the degenerating 
vessels of Mayer and Fuchs. The degenerating vessels described 
by Huntington in the cat have a wholly different appearance 
and to one familiar with the study of sections, the danger of 
confusing the two seems inconsiderable. With the aid of recon- 
struction and the comparison of closely graded embryos to ascer- 
tain the direction of the process, it is difficult to conceive of a 
possibility of so gross an error. In passing, I would note that the 
term anlage has hitherto been applied to the developmental 
process as indicating the earliest appearance of an organ. It is 
quite improperly used in connection with regressive structures. 

More recently Clark*® has published painstaking and elaborate 
studies of growing lymphatics in the tail of the tadpole, contain- 
ing many interesting details and in major matters confirmatory 
of the work of his predecessors. We may therefore conclude that 
the vessels of the tail fin are formed by sprouting and that there 
is no evidence for the annexation of mesenchyme cells at this stage. 

The debate is again not as to the fact, but solely in regard to 
the construction that shall be put upon it, and in particular the 
degree to which such observations may be extended to endothe- 
lium in general. 

Goette*#! described the first formation of vessels in amphibia 
from clefts in the mesoderm, which secondarily acquire a wall 
from adjacent cells and from intravascular wandering cells. 
Marshall and Morgan* have reached similar conclusions, and 
recently Studnicka* has described the formation of the capillary 
as a channel in the mesostroma, into the wall of which cells mi- 
grate eventually forming an endothelium: 

Ganz ibnlich verhalt es sich auch mit den Blutkapillaren, welche 
jedenfalls sehr frith, wenn das Mesostromanetz noch ziemlich locker ist, 


zum Vorschein kommen. Die Bilder, die ich an meinen Objekten finde, 
sprechen durchaus nicht dafiir, dass es ausschliesslich nur die Zellen 


891909, Anat. Rec., vol. 3, and 1912, Am. Jour. Anat., vol. 13. 

11875, Die Entwickelungsgeschichte der Unke. Leipzig. 

* 1890, Stud. Biolog. Lab. Owens College 2: 1893. Vertebrate embryology. 
881897, The development of the frog’s egg. New York. 

441911, Anat. Anz., Bd. 40, p. 33. 


20 VASCULOGENESIS IN THE CAT 


wiiren, welche, wie man est meist annimmt, sich rohrenf6rmig zu Kapil- 
laren umbilden wiirden; alles spricht dafiir, dass die Kapillaren im 
Mesostromanetz auch dort, wo keine Zellen vorhanden sind, also ohne 
jede Beihilfe von Zellen weiterwachsen, und ihre Wand besteht da aus 
einer einfachen, zuerst plasmatischen Schicht, welche jener vollkommen 
entspricht, aus der sich z. B. das Corium bildet. Die Zellen sind in der 
Wand der jungen Kapillaren zuerst weit voneinander entfernt, und nur 
so, dass sie sich spiter vermehren und die innere Oberfliche der friiher 
schon vorhandenen Kapillaren sekundir bedecken, kommt die Endothe- 
lauskleidung derselben zustande. 


It is evident therefore that the processes in the tadpole’s tail 
are not characteristic of all stages even in amphibia, and that 
much caution is requisite in attempting to generalize from such 
data. In the light of our present knowledge, far from extending 
these findings into a principle of vascular development, it would 
seem more prudent to seek to ascertain the peculiar conditions 
in larval organs which may serve to explain a deviation on the 
part of the mesenchyme cell from its usual behavior to develop- 
ing vessels. 

The transparent embryos of teleosts also afford a favorable 
opportunity for the study of living vessels. Wenkebach* has 
recorded important observations, especially upon Belone longi- 
rostris, and finds that mesenchyme cells are an important factor 
in the formation of vessels and sprouts. 


Man sieht deutlich wie die Zellen, namentlich die des Mesoblasts, 
selbstiindig mittelst amoeboider Bewegungen und oft ausserordentlich 
langer protoplasmatischer Fortsitze sich im Kérper des Embryo und 
auf dem nicht mit Hypoblast umkleideten Dotter bewegen und nach 
bestimmten Stellen kriechen, als handelten sie mit Wille und Bewusst- 
sein. Bei der Anlage und weiteren Ausbildung des Herzens, so wie bei 
der Bildung der Gefiisse und andrer Organe spielt diese selbstiindige 
Bewegung der Mesoblastzellen eine grosse Rolle. 

Nicht alle Zellen, die von dem Embryonalsanm auswandern bilden 
Pigment. Hine sehr grosse Zahl derselben zerstreut sich uber die Ober- 
fliiche des Dotters und diese sind als Material fiir die spater sich dort 
bildenen Blutgefisse zu betrachten. 

Aus diesen Beobachtungen geht hervor, dass Mesoblastzellen durch 
selbstiindige amoeboide Bewegungen die Wande der Blutgefiisse des 
Dotters bilden. 

Endlich werden die ersten circulirenden Zellen auch benutzt bei der 
Sprossenbildung der Dottergefasse. Die drei auf dem Dotter vorhan- 


351886, Arch. mikr. Anat., Bd. 28, p. 225. 


VASCULOGENESIS IN THE CAT 21 


denen Gefisse fangen namlich schon bald an kleine Seitenzweigchen 
zu bilden (Fig. 12 u. 13). Eine solche Sprossung wird eingeleitet 
von einer kleinen Ausbiegung der Gefiisswand (verg. Fig. 25 a, b, e, d, e. 
und f.), an deren Spitze sich immer eine Zelle befindet. In dem Maasse 
als die Sprosse wiichst, lagern sich immer neue, aber nicht sehr zahlreiche 
Zellen an dieselbe an, welche sich auf die schon oben beschriebene 
Weise ausbreiten, sich mit langen Fortsitzen an einander haften und 
so die Wand herstellen. Hierfiir werden sowohl die vom Blutplasma 
herangefuhrten, als die sich schon auf dem Dotter befindenden Zellen 
verwendet. Diese letzteren Zellen scheinen eine grosse Neigung zu 
haben, Gefiisse zu bilden, denn einige Male beobachtete ich, wie sich in 
diesem Stadium der Entwicklung freie, sich auf dem Dotter befindende 
Zellen, unabhiingig von schon vorhandenen Geféssen, zusammenlag- 
erten und kleine Rohrchen bildeten, welche spaiter in das System der 
Dotterblutbahn eingereiht worden. 

In diesem prachtvollen klaren Embryo sah ich auch die Bildung der 
kleineren Gefiisse im Embryo selbst. Wie aus Figur 24 hervorgeht, 
bilden sich hier die kleinen um das Medullarrohr ziehenden Gefiisschen 
ebenfalls aus Zellen, welche sich mittelst amoeboider Bewegungen 
zwischen den Theilen des Embryo hinziehen. 


In their general features these observations were confirmed by 
Raffaele :*° 

Le osservazioni fatte dal Wenckebach e gia tante volte citate, e che 
io ho potuto in varie occasioni ripetere con risultati del tutto concordanti 
su varie uova di Teleostei, dimostrano chiaramente la migrazione di cel- 
lule mesenchimatiche in siti molto lontani dal punto di origine; ed é con 
questo processo che si formano tutti i vasi del vitello, compreso il cuore, 
il quale in taluni pesci ossei (Belone) si forma innanzi alla testa dell’ em- 
brione e in parte sul vitello. 


From the foregoing, it would appear that in teleosts also there 
is the closest affinity between mesenchyme and endothelium, 
and that here sprouting from a formed and specific endothelium 
hardly plays a réle in early stages of vascular formations. 

A few other intra vitam observations may be noted briefly. 
MeWhorter and Whipple were able to show and record photo- 
graphically the concrescence of separate vascular anlages in 
the area pellucida of the chick’s blastoderm in vitro. This 
confirms Afanassiev’s*? and Klein’s*’ observation of endothelial 
vesicles and seems an adequate refutation of Evan’s opinion: 

36 Mitt. Zool. Stat. Neapol. Bd. 10, p. 441. 

37 1869, Bul. de l’Acad. des Se. d. St. Petersbourgh, T. 13, p. 321. 


38 1871, Sitz.-Ber. Akad. wiss. Wien, Bd. 63, p. 339. 
: 


22 VASCULOGENESIS IN THE CAT 


Selbst auf dem Dottersack sind solche Gefisse, welche in loco ent- 
standen sind, nicht zahlreich, denn sie entstehen nur an den Stellen der 
sogenannten ‘‘Anlagen,” und die Hauptmasse des Kapillarplexus des 
Dottersacks entsteht durch Ausbreitung und reichliche Anastomosen- 
bildung von den primiiren Gefdssen aus.*! 


Further, the description by McWhorter and Whipple*’ of the 
to and fro movement of red corpuscles after the heart had begun 
to beat and prior to the establishment of the circulation points 
to their confinement in separate compartments. Their subse- 
quent abrupt entry into the blood stream, must mean the giving 
way of the wall of the vesicles and their confluence to form 
vessels—all which is consonant with the data obtained from sec- 
tions and reconstructions, and seems to afford a basis of inter- 
pretation for an interesting observation of Loeb’s:*! 


It turned out in my experiments that the heterogeneous hybrids 
between bony fishes formed eyes, brains, ears, fins and pulsating hearts, 
blood and blood vessels, but could live only a limited time because no 
blood circulation was established at all—in spite of the fact that the 
heart beat for weeks—or that the circulation, if it was established at all, 
did not last long. 

, I succeeded in producing the same type of faulty embryos 
in the pure breeds of a bony fish (Fundulus heteroclitus) by raising the 
eggs in 50 ce. of sea water to which was added 2 ce. one one-hundredth 
per cent, NaCN. The latter substance retards the velocity of oxida- 
tions and I obtained embryos which were in all details identical with 
the embryos produced by crossing the eggs of the same fish with the 
sperm of remote teleosts, e.g., Otenolabrus or Menidia. These embryos, 
which lived about a month, showed the peculiarity of possessing a beating 
heart and blood, but no circulation. 


These findings of Loeb can hardly be reconciled with the doe- 
trine that the vessels of the embryo have a primitive continuity 
of lumen with those of the yolk-sac, for it is inconceivable that in 
such circumstances, a beating heart could fail to effect a circu- 
lation. They are however corroborative of the observations of 
MeWhorter and Whipple which would lead us to interpret the 
condition as a Hemmungsbildung due to failure of concrescence 


*° 1912, In Keibel u. Mall. Entwickelungsgesch. d. Menschen, Bd. 2, p. 552. 
491912, Anat..Rec., vol. 6, p. 121. 
11912, Pop. Sci. Mo., vol. 80, p. 5. 


VASCULOGENESIS IN THE CAT eS 


between the separate anlages of the vessels in the yolk-sac and in 
the embryo. 

Very recently Miller and McWhorter have demonstrated the 
formation of the embryonic vessels in situ and their absolute gene- 
tic independence of those of the area vasculosa in the chick, by 
effecting a separation on one side between the body of the embryo 
and its membranes at a very early period, before vessels have 
appeared in the area pellucida. In their experiments, the blasto- 
derm is incised longitudinally near the primitive streak. The 
wound gapes widely and the:area vasculosa is separated from 
the embryonic body by a wide interval, which no sprouts cross. 
Notwithstanding which, the embryonic vessels are formed typi- 
cally, and differ from those of the normal side only in size and 
rate of development, both of which details may be correlated to, 
their reduced drainage area and the in consequence diminished 
quantity of circulatory fluid. 

It would seem therefore that the limitations of injection and 
the data of intra vitam methods are hardly favorable enough to 
the doctrine of specificity, hardly discordant enough with the 
findings by the method of slides and reconstruction,‘ to attribute 
differences of opinion in these matters wholly to the methods 


” Abstract, Proc. Am. Ass. Anat., Anat. Rec., vol. 8, no. 2. 

43 EB. R. Clark has endeavored (Anat. Rec., vol. 3, 1909) to show that fixation 
gives interruptions in the course of vessels which have previously been ascertained 
to be continuous by direct observation intra vitam. “‘A drawing was made of 
lymphatic and blood vessels during life. The tail was then cut into serial sec- 
tions ten micra thick, stained in hemotoxylin and Congo-red. In attempting to 
reconstruct it was found while blood capillaries, could often be fairly well con- 
structed, it was impossible to reconstruct the lymphatics beyond the muscular 
margin.’’ In a second paper (Anat. Rec., vol. 5, 1911) Clark shows evidence of 
retrograde progression in this line of technique, for the blood vessels in his 3, 6, 
and 7 figures can hardly be considered “‘fairly well reconstructed.’ There are no 
technical procedures that will not miscarry and such tests as Dr. Clark’s are 
critical of personal accomplishment and judgment not of method in the abstract. 
In view of the technical difficulties offered by amphibian material, and the fact 
that none of the data on which the theory of concrescense is based are derived 
from this form, Dr. Clark’s criticism is both unfair and far from the point. With 
the same plausibility he might have proved that good sections could never be 
cut, because the yolk of frog’s eggs is very likely to tear in paraffin. There is a 
relation between method and material, to disregard which for controversial pur- 
poses shows more of zeal than of sound reason. 


24 VASCULOGENESIS IN THE CAT 


employed. I have already expressed a doubt that this was the 
case and have suggested that different principles of interpretation 
were at the root of the difficulty, and I believe that the fault lies 
with the upholders of the doctrine of specificity, in their too great 
confidence in the uniformity and simplicity of ontogeny, in their 
failure to distinguish between genesis and growth, and in the pre- 
maturity of their assertion that endothelium is specific, before 
they have canvassed its potentialities and seriously studied its 
regressive phases. 

The tendency to minimize the ‘period of development, while 
the natural outcome of reliance upon injections which reveal 
only formed endothelium and continuity of lumina, has given 
the theory of specificity a strangely preformist character. Con- 
sistently carried out, it would reduce the history of endothelium 
to the expansion and ramification through the body of a single 
anlage, produced at a distance from the embryo in the mesoderm 
or from the entoderm of the yolk-sac. The innumerable separate 
anlages here, and their gradual conerescence are supported by a 
weight of evidence impossible to set aside, but the facts seem easier 
to be borne when this multiplicity is veiled as far as possible by 
using the term ‘anlage’ in the singular number and the collec- 
tive sense. Yet even this ‘anlage’ proves unable to account for 
all the vessels of the embryo by its expansion, though Evans feels 
that most of those in the splanchno-pleure are formed by sprout- 
ing. Exception must of course be made of the early ones in the 
area opaca, the ‘anlage’ itself, and those in the area pellucida 
which MeWhorter and Whipple have shown are formed by 
concrescence. 

Still there are further difficulties, certain vessels of the embryo 
arise in situ—the aorta in the head, the first arch, parts of the 
dorsal aorta, perhaps the heart; it is possible to attach too much 
importance to the early connection of the caudal portion of the 
aorta with the splanchnopleurice plexus (in chicks of 20 somites, 
Evans), and in the face of very general evidence of in situ origin 
it can only be maintained that any vessels which have not been 
individually proved to arise in loco are derived by ingrowth from 
the ‘anlage,’ with the lumen of which their lumina are shown by 


VASCULOGENESIS IN THE CAT 25 


injection to be continuous, when injection becomes practicable. 
From this moment (20 somites in the chick is the earliest recorded 
‘perfect’ injection of Evans‘) there is a sudden and abrupt end 
to the production of discrete anlages, contrary to what one would 
expect from the general gradual transitions in natural processes. 
All of the endothelium is now continuous, development in the 
sense of genesis has ceased, and growth alone takes place from this 
time forward. The moment at which injectibility is acquired 
is then the critical point in the whole history of endothelium. 

But in view of the many separate anlages, admittedly present 
in the splanchno-pleure and in the embryo and of the transitions, 
almost universally observed in natural processes, the possibility 
must be conceded that a belated vesicle or two of endothelium 
might conceivably escape a sudden annexation to the injectible 
system. It becomes, therefore, a matter of interest to determine 
its status after injection has been practised and it has failed, 
hypothetically, to be injected. A moment before, had it been 
observed in vitro, or had the blastoderm been sectioned, it would 
have appeared like any of the other discrete anlages in the ‘an- 
lage.’ But the test of injection made, it at once ceases to have 
a future, it has become an artifact, or if pertinaciously insisted 
upon, a ‘Mayer-Lewis anlage.’ Because by a hair’s breadth, by 
the veriest fraction of a second, it failed of conerescence prior to 
the moment of injection, it is absolutely devoid of vascular poten- 
tialities, and could never, had the injection been omitted, have 
joined its fellows and participated in the formation of the vascular 
system. 

But our belated vesicle loses more than its future, it can hardly 
be said to have a real present existence. When the injection is 
made, it stands a chance of never being seen by the injector. 
The injection is then ‘perfect.’ If it is seen, it is an ‘artifact,’ 
or the injection is not ‘complete.’ If it is pointed out by another 
observer, then again the injection is incomplete, or it is a ‘Mayer- 
Lewis anlage’ and has lost a previous continuity with the injected 
vessels. All possibilities have been thoughtfully canvassed by 
the exponents of injection, and it can be confidently asserted that 


441909, Amer. Jour. Anat., vol. 9, p. 281. 


26 VASCULOGENESIS IN THE CAT 


a perfect injection injects all the endothelium present; if more is 
found, then the injection is a failure but the principle remains 
unchanged. 

There is, however, no reason to assume that injection has reach- 
ed the acme of delicacy, great as is the skill which has been de- 
veloped in its use. It is conceivable that some day it will be 
possible to inject a portion of the ‘anlage’ before its admittedly 
discrete vessels have fused—even a small fragment might be in- 
jected alone, the aorta in the head say, or the umbilical vein. 
This would then become to the injectionist the only source of the 
whole vascular system. The uninjected parts of the anlage would 
be demonstratedly not vascular amlages at all, but artifacts or 
‘Mayer-Lewis anlages’ and no study by slide or section could 
establish their right to serious consideration; they would be in the 
same status as the separate anlages of the thoracic duct, or any 
of the mesenchymatous vesicles of Huntington and McClure, 
and no degree of care in topographical localization, no minute- 
ness in the gradation of reconstructed stages, no matter how fully 
their enlargement, gradual approximation and fusion were shown, 
would suffice to prove their participation in the formation of 
vessels. If not injectible they can never become endothelium. 
But the discrete vesicles and blood islands of the splanchno-pleure 
are as an actual fact not injectible from the first and their struc- 
ture is hardly known apart from sections. It would seem to be 
a reasonable and consistent conclusion from the injectionist stand- 
point, that the ‘anlage,’ in the splanchno-pleure is an artifact 
or a ‘Mayer-Lewis anlage,’ and has absolutely nothing to do with 
the vascular system,which should consistently be held to come into 
being ex abrupto with the first moment of injectibility, for the 
chick, at present, the stage of 20 somites. 

But some may not feel themselves obliged to confine their 
knowledge of endothelium to such facts only as can be corrobo- 
rated by injection, for there is a stage of vascular genesis admit- 
tedly inaccessible to injection, and here the method of section 
and reconstruction has acknowledged validity and further is 
entirely consonant in its results with the data of intra vilam 
observation. 


VASCULOGENESIS IN THE CAT 27, 


This granted, it is arbitrary to assert, that injection once it 
becomes practicable, dispenses us from the examination of the 
periphery beyond the terminus of the injection, where it is pos- 
sible, where there is even antecedent probability that processes 
analagous to those in the ‘anlage’ continue and play an impor- 
tant rdle in the increase of vessels within the embryo. Whether 
this is true or not must be decided by the methods which do not 
prejudge the question. As in the case of the ‘anlage’ the results 
of the study of sections are in agreement with intra vitam find- 
ings, it is but reasonable in later stages and in the body of the 
embryo, where intra vitam methods are as yet inapplicable, to 
credit the method of sections and reconstruction with the same 
validity it is proved to possess in early stages. 

To deny this, and insist that the limited field explored by in- 
jections is alone accessible to investigation, is not to assert the 
primacy of the injection method, but to fall back upon an ancient 
logical device known as the petitio principi. 

The final objection and absolute disproof of the doctrine of 
specificity is the heterogeneity of the products to which endo- 
thelium gives rise.*® I have already mentioned Huntington’s de- 
scription of the conversion to mesenchyme of regressive vessels 
in embryos of the cat (wide ante, p. 12). 

Aurel von Szily*® has described and illustrated the production 
of mesenchyme from the endothelial tube of the heart. And to 
pass over other investigators who have glanced the subject, 
Mall’ has recently followed the process in detail and noted its 
theoretical importance. 

The great importance of this distinction is at once apparent for it 
shows that connective tissue arises also from endothelial cells and that 
the intima of the entire vascular sytem including the valves of the heart 
has a like origin. 


Further study will probably show that endothelial connective tis- 
sues are by no means of rare occurrence. 


These observations of vy. Szily, Huntington and Mall are of 
decisive importance and seem finally to dispose of the doctrine 

4° The idea that endothelium gives rise to conne:tive tissue was clearly ex- 
pressed as early as 1872 by Franz Boll. Arch. f. Mikro. Anat., Bd. 8, p. 53. 


46 1903, Anat. Anz., Bd. 24, p. 417, cf. especially figure 2. 
471912, Amer. Jour. Anat. vol. 13, p. 249. 


28 VASCULOGENESIS IN THE CAT 


of the specificity of endothelium, for it is fatuous to predicate 
specificity when products deviate in form and function from the 
type of the parent tissue. 

Of even greater import, because evincing an identity of poten- 
tialities on the part of mesenchyme and endothelium, are the 
recent investigations in haematopoesis, especially those of Maxi- 
mow,'® Dantschakoff,4? Mollier®® and Weidenreich.*! The last 
(11) has published a comprehensive critical review of the mod- 
ern literature of the white blood cells. The extent and variety 
of the evidence and the number of observers render impressive 
their concordant testimony to the mesodermal origin of blood 
cells, in the yolk-sac and the mesenchyme of the embryo, in the 
liver, spleen and lymphnodes, together with their frequent deri- 
vation from endothelium, and in a word the equivalence and per- 
mutability of mesenchyme, endothelium and blood cell. A few 
quotations from Weidenreich may serve to substantiate these 
statements (the italics are his): ° 


Jedenfalls entstehen also schon in den friihesten Entwickelungsperioden 
durch Abrundung und Loslésung aus dem Verband der Mesenchymzellen 
gleichfalls ‘primitive Leucocyten,’ die teils den innerhalb der Blutbahn 
circulierenden vollig gleichen, teils in Grésse und Kernform differieren. 
Der Mutterboden dieser extravasculir gebildeten Elemente ist aber 
seiner Art nach durchaus derselbe, wie der der intravasculir entstehenden; 
denn sie entstammen beide mesenchymatésen Zellen, sei es nun, dass sie 
sich direkt aus den Blutinselelementen absondern oder sekundir aus 
dem Endothel der Blutbahnen hervorgehen. . . . . Hs ist ferner 
von besonderer Wichtigkeit dass tibereinstimmend angegeben wird, 
dass neben den wngranulierten primitiven Leucocyten verhdltnismassig 
frithe auch schon granulierte Formen angetroffen werden, aber und das 
ist wesentlich nicht innerhalb, sondern ausserhalb der Blutbahn; es gehen 


$8 1906, Arch. f. mikr. Anat., Bd. 67, p. 608, 1907, Fol. haemot. Bd. 4, p. 611; 
and Beit. z. Path. Anat. u. allg. Pathol., Bd., 41, p. 122. 1909, Arch. f. mikr. 
Anat., Bd. 73, p. 444; and idem, Bd. 74, p. 525 ane Fol. haemotol. Bd. 8, p. 125; 
and Centralbl. f. allg. Pathol. u. path. Anat., Bd. 20, p. 145. 1910, Arch. f. rile 
Anat., Bd. 76, p. 1; and Verh. Anat. Ges. Brussel, 8. 64. 

49 1908, Anat. Hefte, Bd. 37, p. 473; Arch. f. mikr. Anat. Bd., 78, p. 117. Vehr. 
Anat. Ges. Ber., 1908, p. 72. 1909, Arch. f. mikr. Anat., Bd. 74, p. 855. Vehr. Anat. 
Ges. Brussel. 1910, p. 17. 

50 Loc. cit. 

Die Leucocyten und verwandte Zellformen. Weisbaden 1911. Sonderausgabe 
aus Merkel u. Bonnet, Ergeb. d. Anat. Bd. 19, Abt. 2 


VASCULOGENESIS IN THE CAT 29 


also hier die granulierten Elemente zunichst aus leucocytiiren Wander- 
zellen d. h. freigewordenen Mesenchymzellen hervor.®? 


In the Liver: 

Die ‘primitive Blutzelle,’ die sich abrundende Mesenchym-oder Reti- 
culumzelle, die undifferenzierte ‘Gefasswandzelle’ ist viberall dasselbe Ele- 
ment, es erscheint bald als Begrenzung einer Reticuluwmmasche, bald als 
Wand einer schon etwas kanalisierten Blutbahn, bald als freie Blutzelle; 
genau die gleichen Verhaltnisse kehren also in der Leber wieder, wie 
man sie in den Blutinseln und Blutstriingen und den ersten Gefissan- 
lagen des Mesoblasts findet, nur mit dem Unterschiede, dass hier Ento- 
dermzellen in Form der Leberanlage einwachsen und das Ganze mehr 
zerteilen. Als wesentlichstes Ergebnis ist sonach festzuhalten, dass 
rote Blutkérperchen and Leucocyten von einer Zelle mit mesenchymati- 
schem Charakter gebildet werden. Die Frage, ob sie beide oder nur die 
eine extra, oder intravasculir entstehen, ist nach der eben erérterten 
Auffassung ziemlich belanglos, da ja die Zelle, welche die Abgrenzung 
der primitiven Blutbahn nach innen und aussen bildet, eigentlich die 
Fahigkeit der Produktion beider Arten besitzt, eme Fahigkeit, die auch 
zuniichst noch ihre freien Abkémmlinge bewahren.** 


In Bone Marrow: 

Nach dieser auf genaue Detailbeschreibung und zahlreiche Ab- 
bildungen gestiitzten Auffassung entsteht also auch im Knochenmark das 
blutbildende Gewebe durch Differenzierung aus losgelésten Mesenchyme- 
zellen vom Charakter kleiner und grosser ungranulierter basophiler Leu- 
cocyten, die sich sowohl in der Richtung der roten Blutkérperchen, wie auch 
der weissen weiterdifferenzieren; noch ausgesprochener wie in der Leber 
erschient dabei der ganze Prozess als ein extravascularer Vorgang.*! 


In the Spleen: 

So unbestimmt auch vorerst noch die Angaben iiber “die Herkunft 
der Blutbildunsgzellen der Milz lauten, so schient mir es doch im héch- 
sten Grade wahrscheinlich, dass in der Milz wie in Leber und Knochen- 
mark die Entwickelung ausgeht von indifferenten basophilen grésseren 
und kleineren Zellen, die als abgerundete Mesenchymzellen oder Wander- 
zellen an Ort und Stelle sich bilden und sich dann zundchst sowohl in der 
Richtung der Erythrocyten wie der Leucocytenrethe (granulierte Leucocyten) 
differenzieren. Dieses erste Elemente, die ‘primitive Milzzelle,’ wire 
also auch hier wieder die primitive Blutzelle bezw. der primitive Leucocyt 
oder Molliers Hiimogonie. Die Umwandlung dieser Bildungszellen 
in die kleinen Lymphocytenformen im Gebiete der Malpighischen 
K6rperchen setzt als besondere Gewebsdifferenzierung erst zu einen spit- 
eren Zeitpunkte ein; ich werde zeigen, warum das fiir die Gesamt- 
auffassung der Lymphocytenfrage vollig gleichgiiltig ist.” 

In Lymphnodes: 


52 Ibid, p. 188. 3 Ibid, p. 196. 54Tbid, p. 199. 55 Ibid, p. 201. 


30 VASCULOGENESIS IN THE CAT 


Aus allen diesen Angaben geht also hervor, dass die freien Elemente 
der Lymphknoten an Ort und Stelle enistehen und zwar wie iiberall aus 
abgerundeten und losgelésten Mesenchymzellen, die eine bestimmte Zeit- 
lang auch hier die Fahigkeit haben, sich in der Richtung der roten und 
weissen Blutkérperchenrethe zu differenzieren. . . . . % 

In Connective Tissue: 

Als wesentliches Resultat aller dieser Untersuchungen ergibt sich 
also, dass dem embryonalen Bindegewebe zundchst im weitesten Umfange 
die Fdahigkeit zukommt, aus seinen Mesenchymzellen freie bewegliche 
Blutzellen zu bilden, die den Typus der primitiven Blutelemente aufweisen 
und sich anfinglich sowohl in der Richtung der Erythrocyten wie der 
Leucocyten differenzieren kénnen.5* 


The doctrine of specificity has found its support in phenomena 
of growth, chiefly in the solid terminations of lymphatics which 
are observed in rather late periods of ontogeny, and the avoidance 
of these sprouts by mesenchyme cells in the tail of the tadpole. 
The former fact has been by implication projected into earlier 
stages of development, and into peripheral regions which are not 
accessible to the method of injection. The latter has been erect- 
ed into a general principle of vascular development, in spite of 
facts of diametrically oppposite import, both in the frog (Goette, 
Marshall, Morgan, Studniéka) and in other transparent embryos 
(Wenkenbach, Raffaele). 

The theory of adaptation rests upon the observation of the 
transformation of mesenchyme cells into endothelium in a large 
number of forms at various periods of ontogeny, ascertained by 
a variety of methods and by observers with very different prob- 
lems in view; second on the reversion of endothelium to mesen- 
chyme (vy. Szily, Huntington, Mall): and third, on the commu- 
nity of products and the permutability of endothelium, blood 
cells and mesenchyme (Maximow, Dantschakoff, Mollier, Weiden- 
reich et al.). . 

The doctrine of the angioblast is so closely entwined with that 
of the specificity of endothelium that there is little promise in 
an attempt to consider it separately, especially as this conception 
of vasculogenesis loses its chief raison d’etre with the recognition 
of the plasticity (non-specificity) of endothelium. As presented 


56 Ibid, p. 203. 57 Tbid, p. 205. 


. 


VASCULOGENESIS IN THE CAT 31 


by Minot* the doctrine of the angioblast rests on two general 
propositions. 

1. The early appearance of the angioblast and its possible 
entodermal origin: 


Die vergleichende Embryologie lehrt uns, dass die ersten Gefiisse 
gleichzeitig auf dem Dottersack entstehen. Sie bilden eine einheit- 
liche Anlage, die wir dem Vorschlage von His folgend kurzweg Angio- 
blast nennen. Es ist aber gleich hervorzuheben, dass mehrere For- 
scher, wie Mazximow in seiner neuesten Schrift, Blutgefiisse aus dem 
mesoderm des Embryo direkt entstehen lassen. In der Tat koénnen 
wir die vollkommene frithzeitige Unhabingigkeit des Angioblastes von 
dem eigentlichen Mesoderm nur als héchst wahrscheinlich behaupten. 
Der Angioblast liegt ursprunglich unmittelbar auf dem Dotter und 
bildet ein Netzwerk, das fast so fort nach dem ersten Erscheinen der 
Anlage sich zu erkennen gibt. Das Mesoderm, sensu strictu, bildet 
eine zusammenhangende Schicht, die oberhalb der Gefiissanlagen liegt, 
und nur in den Liicken des Gefiissnetzes unmittelbar mit dem Dotter 
in Berithrung kommt. Der Angioblast scheint, der grossen Mehrzahl 
der Angaben nach, bei allen Wirbeltieren sich direkt vom Dotter abzu- 
spalten. Es ist also sehr schwer zu entscheiden, ob der Angioblast gene- 
tisch als dem mittleren Keimblatt gehérig oder als Abkémmling des 
Entoderms aufzufassen ist. Selbstverstiindlich gehen die Anschichten 
hieruber weit auseinander. 


2. Its probable complete independence (specificity of endothe- 
lium). 


Es bleibt aber die Tatsache, dass der Angioblast sehr friihzeitig selhst- 
stiindig wird, und diiss er das erste Gewebe des Embryo ist, welches 
eine unzweifelhafte Differenzierung und scharfe Abgrenzung aufweist. 

Der Angioblast behilt aller Wahrscheinlichkeit nach zeitlebens seine 
vollstandige Unabhangigkeit. Mit anderen Worten, es ist wahrsch- 
einlich, dass die Endothelien der Blutgefiisse (und der Lymphgefiisse) 
und die Blutzellen in jedem Stadium simtlich direkte Abkémmlinge 
des primitiven Angioblastes sind. Leider aber erlauben uns unsere 
gegenwirtigen Kenntnisse nicht, mit absoluter Sicherheit eine Meinung 
hieritber auszusprechen. 


It will be seen that the extremely problematic nature of both 
of these postulates is perceived by Minot and that his view of 
the derivation of blood and endothelium from the angioblast 
is hardly to be reconciled with the results of the careful and 
thorough investigations summarized by Weidenreich. One may 


588 Keibel a. Mall. Handbuch d. Entwickgeschicht. d. Menschen, Bd. 2, p. 483. 


32 VASCULOGENESIS IN THE CAT 


share Dr. Minot’s regret that the early closure (’09) of his chap- 
ter on the blood prevented him from availing himself of the fun- 
damentally important data of recent haematology, but it is more 
seriously to be deplored that such far-reaching doctrines as those 
of the angioblast and of specificity should have been promul- 
gated upon a narrow factual basis in a text-book that is sure to 
fall into the hands of the inexperienced when, at the time of pub- 
lication, (12), so much evidence was at hand which would seem 
to invalidate these hypotheses. 

The statement (vide supra) that the testimony of the majority 
of findings supports a direct origin of the angioblast from the yolk 
must not be taken to imply an entodermal origin, for in this sense 
it is not substantiated by the literature as a few citiations from 
the careful articles of Riickert and Mollier®? may serve to show. 


Amphibia, Urodeles: 

Trotz vieler Bemuhiingen habe ich auch nicht zu entscheiden ver- 
mocht, ob an der Lieferung dieser Gefiisszellen fur das Dotterfiissnetz 
(Fig. 692 gfz) ausser dem Mesoblast auch der Entoblast beteiligt ist. 
Ich glaube bestimmter nur sagen zukénnen, dass im Bereiche der epi- 
thelialen Splanchnopleura keine Gefisszellen vom Entoderm abgegeben 
werden, wiihrend viele Beobachtungen dafiir sprechen, dass seitlich 
am Dotter dies vorkommt.*” 

Anura: 

Bei Anuren ist folglich die Blutzellenbildung vom histogenetischen 
Standpunkte entschieden eine vorwiegend entodermale gegentiber der 
vorwiegend mesodermalen bei Urodelen. Will man, um diesen Unter- 
schied zu beseitigen, die Annahme von Schwink gelten lassen, dass es 
sich bei Anuren um eine verspiitete Ablésung von eigentlich mesoder- 
malen Zellen vom Entoderm handelt, so ist nichts dagegen einzuwenden 
und es mag dieselbe fiir manchen vielleicht eine gewisse Beruhigung sein."' 

Cytostomata: 

Durch diese liickenhafte Untersuchung einiger Entwickelungsstadien 
von Petromyzon sind wir doch zu dem Resultat gelangt, dass die Ge- 
fass- und Blutbildung bei Cyclostomen in einfacherer Weise sich ab- 
spielt als bei Amphibien. Es kommt das vor allem darin zum Ausdruck, 
dass bei Cyclostomen die Aufgabe, Gefisse auf dem Dotter absugrenzen, 
fast ausschliesslich noch der diinnen einschichtigen Mesoblastlage selbst 
zufillt, da nur stellenweise vereinzelte Gefisszellen auf dem Dotter sicht- 
bar werden. 


59 1906, Hertwig’s Handb. d. vergl. u. exp. Entwick. d. Wirbelth., Bd.1, Th. 1. 
60 Mollier, loc. eit., p. 1068. 
61 Mollier, loc. cit., p. 1074. 


VASCULOGENESIS IN THE CAT 33 


Die Entstehung der Blutzellen ist aber abhingig von der Art und 
Weise der Mesodermbildung, so dass der alte Streit, ob das Blut aus 
dem Mesoderm oder dem Entoderm sich bildet, gar nich generell fiir 
die Wirbeltiere entschieden werden kann." 

Dipnoi: 

Es ‘ihnelt also die Entwicklung mehr den Urodelen als den Anuren, 
da der Mesoblast frithzeitig vom Dotter sich trennt und die Gefiissan- 
lagen zuniichst also ein mesodermales Produkt sind. 

Selachians: 

In welchem Umfang diese entodermale Gefiissbildung stattfindet 
lasst sich nicht leicht abschitzen. Die Zahl der Sprossen an einer 
Keimscheibe ist eine missige und an den verschiedenen Keimscheiben 
zudem wechselnde, aber andererseits zieht sich ihre Bildung iiber eine 
verhaltnismissig lange Entwickelungszeit hinaus. Es ist auch diese 
Frage weniger von prinzipieller Bedeutung gegenuber der Konstatierung 
der Thatsache, dass tiberhaupt die mesodermalen Gefdssanlagen von Tor- 
pedo sich durch entodermalen Nachschub ergdnzen, eine Ansicht (Riickert 
1887), die so allgemein bezweifelt worden ist, dass ich selbst mit einen 
gewissen Vorurteil gegen meine eigenen vor 18 Jahren angestellten 
Beobachtungen an eine erneunte Untersuchung des Gegenstandes her- 
angetreten bin.* 

Man sieht hieraus, wie wenig Wert der viel diskutierten Frage ob der 
Meso- oder der Entoblast die Blutanlagen der Selachier liefert, zukommt. 
Es ist im Grunde nur Sache des Uebereinkommens welchen Namen man 
der betreffenden Zellenschicht geben will. Entscheidet man sich zu 
Lehrzwecken fiir eines der beiden Blatter so wird man dem mittleren 
den vorzug geben diirfen, weil ein Teil der Inseln im selbstdndigen Meso- 
blast auftritt und die tibrigen in einer Schicht die bald darauf zum Meso- 
blast wird.°° 

Teleosts: 

Als Produkt des peripheren Mesoderme wire der intermediiiren Zell- 
masse zuniichst die Fahigkeit zuzusprechen, Gefiss- und Blutzellen 
liefern zu kénnen, und es wire eine derartige Differenzierung derselben 
zu erwarten.”® 

Es lasst sich zundchst also wohl auch die gefdssbildende Kraft der in- 
termedidren Zellmasse als peripheres Mesoderm aufrecht erhalten, und es 
miissten diese Gefisse den Dottergefassen echter Meroblastier in der blut- 
bildenden Zone verglichen werden.* 

Reptilia: 

Was die genetischen Beziehungen der Blutinseln zu den Keimblattern 
anlangt, so leiten alle neueren Forscher diese Gefassanlagen aus dem 
Mesoblast ab, wie dies zuerst Strahl fiir Lacerta gethan hat. Das ist 
im Grunde auch die Meinung von Mehnert fiir die Schildkréte und von 


62 Mollier, loc. cit., pp. 1088-9. 6 Ruckert, loc. cit., p. 1095. 
83 Mollier, loc. cit., p. 1079. 66 Mollier, loc. cit., p. 1145. 
64 Ruckert, loc. cit., p. 1111. 67 Mollier, loc. cit., p. 1146. 


MEMOIR NO. 3 


34 VASCULOGENESIS IN THE CAT 


Vélzkow fur das Krokodil, wenngleich beide Autoren den betreffenden 
Mesoblastabschnitt als Entoblast bezeichnet wissen wollen.®8 

Birds: 

Fassen wir das Gesagte zusammen, so hat sich ergeben, das die 
soliden Anlagen der Dottergefiisse des Huhnes ein verschiedenes Verhalten 
zum mittleren Keimblatt darbieten, je nachdem sie hinten oder vorn, 
aussen oder innen in der A. vasculosa auftreten. Diese Differenzen 
lassen sich aber alle leicht verstiindlich machen, wenn man von dem 
allerersten Entwickelungszustand derselben ausgeht und zugleich das 
jeweilige Verhalten des in ihrem Bereich befindlichen Mesoblast mitbe- 
riicksichtigt; Die Gefdssanlagen der A. opaca bilden sich, sie mégen hinten 
oder vorn auftreten, als Verdichtungen und Verdickungen eines vom Keim- 
wall sich delaminierenden Primitivstreifenmesoblast. Sie scheiden dann 
an threr Oberfldche eine Mesoblastdecke ab, wodurch die unterbrochene 
Kontinuitat des Blattes wiederhergestellt und sie selbst aus demselben aus- 
geschaltet werden.. Die vorderen Anlagen unterscheiden sich von den hin- 
teren genetisch darin, dass der Mesoblast sich von ihnen frihzeitiger differ- 
enziert und abtrent, weil er vorn viel rascher der Célombildung zustrebt. Die 
zu innerst in der A. opaca befindlichen Anlagen differieren von den tibrigen, 
weiter peripher gelegenen, darin, dass sie infolge ihres spdteren Auftretens 
in einem Mesoblast sich bilden, der seine Verbindung mit dem Keimwall 
bereits gelést hat. Sie erscheinen daher als rein mesodermale Bildungen.® 

Mammals: 

Nach dieser ausfiirlichen Behandlung der peripheren Mesoblast- 
bildung kénnen die Vorgiinge, welche direkt zur Herstellung des Blutes 
und der Gefisse auf dem Dottersack der Siugetiere fiihren, kiirzer dar- 
gestellt werden um so mehr, als dieselben bisher nur von wenig Forschern 
genauer untersucht worden sind. Hier ist wieder an erster Stelle Kol- 
liker (1875 u. 1879) zu nennen, der beim Kaninchen zu dem gleichen 
Befund einer mesodermalen Blutbildung gelangt ist, wie beim Huhn. 
Seine Angaben wurden durch Heape (1887) fiir den Maulwurf, Robin- 
son (1892) fiir die \V aus, Janosik (1902) fiir Schwein und Ziesel, Fleisch- 
mann (1889) fiir die Katze und Keibel fiir das Meerschweinchen (1889) 
und den Menschen (1890) kurz bestitigt und vor allen in jiingster Zeit 
durch die ausfiirlichen Darstellungen Van der Stricht’s fiir das Kanin- 
chen (1895) und die Fledermaus (1899) erweitert.” 

Eine Beteiligung des Entoblast bei der Blutbildung durch Abgabe 
vom Zellen, schliesst Van der Stricht sowohl fur das Kaninchen wie 
fiir die Fledermaus mit Entschiedenheit aus.”! 

An den gennanten Orten gehen beim Schaf nach Bonnet die Gefiisse 
aus vorgebildeten Riumen hervor, aus Liicken, die auf der Nabelblase 
zwischen dem Entoblast und dem visceralen Mesoblast, auf dem Amnios 
zwischen Entoblast und parietalem Mesoblast ‘‘ausgespart’? und von 
den Mesoblastzellen allmihlich ‘umscheidet” werden. Auf diese weise 
entstehen blutleere, d. h. erythrocytenfreie Endothelréhren.” 


68 Ruckert, loc. cit., p. 1193. 7 Ruckert, p. 1253. 
59 Ruckert, loc. cit., p. 1219. 72 Ruckert, p. 1254. 
7 Ruckert, loc. cit., p. 1251. 


VASCULOGENESIS IN THE CAT 35 


To return to Minot’s conception of an angioblast, his description 
of its formation permits a somewhat more definite examination 
of his position. 


Kurzgefasst findet die Differenzierung des Angioblastes bei Amnioten 
folgendermassen statt: Er bildet ein Netzwerk, das anfinglich aus Zell- 
stringen besteht, diese werden bald hohl; nach vielen Beobachtern kann 
die Héhlung auf der unteren Seite zuerst nur vom echten Dotter begrenzt 
sein. Die Angioblastzellen bilden sich zum Teil in Endothelzellen, 
zum Teil in junge Blutzellen (Mesaméboide) um. Das Endothel ent- 
steht aus der peripherischen Lage der Striinge, die Blutelemente dagegen 
aus den mehr central gelegenen Zellen. Nur die Endothelien bilden 
ein ununterbrochenes Netz, die Blutzellen dagegen stellen zerstreute 
Haufen, die sogennten Blutinseln dar.” 


In this passage a stratum is described interposed between the 
entoderm and the mesoderm (sensu strictu), which is evidently 
ythe Gefissblatt of many authors, a structure eminently vasculo- 
genetic but not exclusively so as was accurately stated by Bon- 
net. Its non-vascular products have generally been lost sight 
of, but its proendothelial stage is clearly recognized by Felix:’® 


Die Darmarterien des Menschen entwickeln sich aus einem Ge- 
faissnetz, welches Darm-und Dottersack umspinnt (rete periintes- 
tinale). Fig. 539 zeigt die Rekonstruktion der Gefiisse eines Embryo 
mit 5 bis 6 Ursegmentpaaren. Die Aorta ist bis zum zweiten Ursegment 
als Rohr mit allseitig geschlossener Wandung entwickelt, von da ab 
caudalwirts ist sie erst in der Anlage begriffen; sie erscheint hier auf 
dem querschnitt bald als geschlossener Ring, bald aus vereinzelten Zel- 
len bestehend. Vom zukiinftigen 7. Ursegment bis zur Schwanz- 
spitze ist zwischen Dottersack bzw. Enddarm einerseits und visceralem 
Mesoblast anderseits ein Gefissnetz eingeschaltet, das rete periintes- 
tinale (Fig.521 und 450). Im Bereiche des Mitteldarms sind die Maschen 
dieses Netzes schon mit deutlicher Wand versehen (dunkelrot in Fig. 
539), im Bereiche des Enddarms haben wir noch gar keine Maschen, 
sondern nur ein einfaches Gefissblatt (Fig. 540), welches durch Delami- 
nation von dem visceralen Mesoblast abgelést wird und noch keine ein- 
zelen Gefiisse entwickelt hat (blass rot in Fig. 539); in ihm verliert sich 
die aorta (Fig. 539). Aus diesem Gefiissblatt entstehen also rete 
periintestinale und aorta. 


73 loc. cit., 484. 

741889, Arch. f. Anat. u. Entwickelungsgesch. p. 1. 

75 1912, Keibel and Mall Handb. d. Entwickschich. d. Mensch. 
Bd. 2, pp. 757-8. 


36 VASCULOGENESIS IN THE CAT 


More recently Bremer” has reconstructed what is essentially 
a prevascular stage of this layer. 

I would emphasize the early stage and the cells which do not 
become endothelium, because I believe that in them we find an 
exit from the embarrassment of recognizing an independent angio- 
blast which yet in its products is identical with mesenchyme, 
for we have too little evidence of convergence in ontogeny to make 
the idea of such an assimilation a priori attractive. It would seem, 
therefore, simpler to accept an origin of mesenchyme in the 
splanchno-pleure at an earlier period than has been universally 
recognized, by a migration of elements from the germ layers into 
the mesostroma as has been described by v. Szily,77 Studniéka7s 
and others. The peculiarity of this district would then consist 
in its great degree of vascular productivity which might be oeca- 
sioned physiologically by its relation to the yolk,” rather than 
determined genetically by a hypothetical derivation from the’ 
entoderm. We here approach the standpoint of Raffaele’°— 
“Te cellule endotheliali provengono dal mesenchime’’—and are encour- 
aged to hope that eventually the formally genetic views embodi- 
ed in the doctrine of the angioblast will give way to a more bio- 
logical standpoint of interpretation, viz., that endothelium is the 
result of mechanical, blood of biochemical factors, both conse- 
quent on position in the broad sense of the term, and are neither 
of them rigidly predetermined to a special cytomorphosis by 
obscure causes associated with a peculiar provenience of their 
elements. 

The mesenchyme has long been held to be a complex of various- 
ly derived elements, which owe their resemblance and solidarity 
to a common position, not to likeness of origin. That cells free 
themselves from germ-layers and migrate into the mesostroma 
between them is well established, and that vessels make an early 
appearance in these strata is indisputable, in the somato-pleure 
as well as in the splanchno-pleure. But that in the latter situ- 

751912, Amer. Jour. Anat., vol. 138. p. 111. Vide addendum. 

77 Loc. cit., and Anat. Hefte, Bd. 33. 

78 Loc. cit. 

79 Cf. Ruckert, loc. cit., pp. 1180-81. 

80 Loc. cit., p. 446. 


VASCULOGENESIS IN THE CAT fh 


ation we have two independent subsidiary germ layers, mesen- 
chyme and angioblast, which are absolutely distinct, though in- 
separably united, and are in no sense phases or modifications of 
one. structural basis, is not an obvious conclusion and is insuper- 
ably difficult in the view of the later permutability of endothelium 
and mesenchymal derivatives. 

The entodermal origin of some of these elements, the mesoder- 
mal of others, the great proportional variation ef the two com- 
ponents in different ova, would seem to point to position in the 
broad sense of the term, and not to derivation as the determin- 
ing factor in their behavior. The comprehensive and detailed 
resumé of Rtickert and Mollier, together with their individual 
contributions seem to necessitate a view of this sort, for they 
definitively dispose of an entodermal or a mesodermal origin of 
vessels as a general principle of vasculogenesis. 

Clearly therefore, though the causative factors of vasculogene- 
sis in the blastoderm are but imperfectly understood, yet in 
the light of present knowledge the source of the vasofactive cells 
must be conceived as of secondary rather than of primary im- 
portance. The doctrine of the angioblast can therefore find no 
support in the facts of comparative embryology in so far asit 
hints an entodermic origin for endothelium and blood. But if 
angioblast and mesenchyme are thus conterminous, arising 
from the same sources, and possessing like potentialities, and have 
further the most intimate topographical relations, it is difficult 
to see how they have come to be separated conceptually by some 
investigators, unless it be from their too intensive confinement 
of attention to the splanchnopleure, were the excess and _pre- 
cocity of vasculogenesis somewhat masks the presence of 
mesenchyme. 

This is the phase of the problem which I have endeavored to 
approach directly in the light of observations upon a series of 
blastoderms of the domestic cat, ranging from the period of the 
three-layered blastoderm, to the stage of fourteen somites. They 
were fixed in Zenker’s solution, embedded in paraffin, cut into 
serial transverse sections 134 y» thick, and variously stained— 
the most satisfactory results on the whole being gained with 


38 VASCULOGENESIS IN THE CAT 


Delafield’s haematoxylin and eosin, or Heidenhain’s haematoxy- 
lin. Reconstructions were made by the Born method in four 
cases; for others, in small easily defined areas, where conditions 
were simple, the quicker but less satisfactory graphic method 
was employed. 

For convenience of analysis, the data are presented under two 
headings: 

1. The early formation of vessels beneath the ectoderm, with 

special reference to the umbilical vein. 
"2. The early formation of vessels above the entoderm, es- 
pecially the omphalo-mesenteric plexus and the aorta. 


THE UMBILICAL VEIN 


The early stages in the development of this vessel afford an 
opportunity to study the formation of endothelium in a cir- 
cumscribed region and under comparatively simple conditions. 
It appears first near the lateral margin of the incipient intra- 
embryonic coelom and is separated from the entoderm by the 
whole thickness of the as yet imperfectly cleft mesoderm. The 
inner germ layer patently is excluded from direct participation 
in the formation of this endothelium, and its appearance, ante- 
dating the establishment of continuity with any derivative of 
the splanchno-pleuric plexus, must entail the acceptance of its 
origin in situ. Tam not aware that the independence of the um- 
bilical vein and the associated hypectodermal vessels has ever 
been impugned. It was recognized by His and more recently 
has been considered by Mollier. In the cat it retains its iso- 
lation to the stage of 14 somites. It would seem, however, that 
its importance from the theoretical standpoint has not been uni- 
versally appreciated. The position of the umbilical vein at the 
amnio-embryonic angle suggests that the vascular potentiali- 
ties of intra-and extra-embryonic portions of the germ layers 
are not, of necessity, totally disparate. The extent of the vessel, 
its comprehensive drainage area, its high physiological import- 
ance, seem to require that its mode of formation be respected 
in general interpretations of vascular development. <A doctrine 
such as that of the origin of endothelium from entoderm, or by 
the crowthfoffanlagesformed in the splanchnopleure alone, neg- 


VASCULOGENESIS IN THE CAT 39 


lects the case of the umbilical vein, and is to that large degree 
incomplete and partial. 

The youngest embryo which requires consideration here is 
No. 550 of the Columbia Collection. The somites are forming 
but as yet there is no complete intersomitic cleft; the coelom is 
little more than a line of beginning cleavage; in a few places the 
layers of mesoderm separate and narrow discontinuous clefts 
appear. 

The parietal mesoderm is, in general, compact. It is separated 
from the ectoderm by a space which must be considered in large 
part due to shrinkage. In general, the outline of the mesoderm 
is clear cut, but lateral to the neural plate and dorsal to the 
coelom-clefts short protoplasmic processes extend from its periph- 
eral cells into the hypectodermal space; a few are also to be 
seen on the basal surface of the ectoderm. These are evidently 
the fibers of Aurel v. Szily,*! the interdermal cytodesmata of 
Studniéka,* which collectively constitute the mesostroma of the 
latter investigator. Subsequently, these processes form a felt 
work between ectoderm and mesoderm, into which cells from the 
mesoderm migrate. The latter process, the formation of the 
somatopleuric mesenchyme, is in its inception in this embryo. 
We may distinguish first a loosening of the cells of the mesoderm 
characterized by an increased distance between the nuclei, the 
appearance of intercellular clefts, and the projection of an oc- 
easional nucleus beyond its fellows; second, the appearance of 
cells, or more strictly energids, connected by broad protoplasmic 
bridges with the mesoderm; in some cases the bridges are narrow, 
the energids are distinctly pedunculated, and finally there are 
a very few elements which are joined to the mesoderm by tenuous 
eytodesmata only. 

In embryos in which the first intersomitic cleft is complete 
(Columbia Collection Nos. 449. 554, 555) the mesenchyme has 
increased in amount and the area involved is larger, now extend- 
ing mesad along the parietal mesoderm as far as the somites 
(fig. 1). The mesostroma is more abundant and in places forms 
a feltwork of protoplasmic strands. The dorsal contour of the 


#11903, Anat. Anz., Bd., 24, p. 417. 
81911, Anat. Anz. Bd. 40, p. 33. 


40 VASCULOGENESIS IN THE CAT 


mesoderm is irregular: many of its elements project into the hy- 
pectodermal space, some broadly, some, narrowly pedunculate, 
others connected to the underlying mesoderm only by fine cy- 
todesmata. The cytoplasm of these energids’ is very scanty; 
the nuclei are large, vesicular, with one or sometimes a second, 
large spherical karyosome. In view of their scanty cytoplasm 
and the protoplasmic nature of the mesostroma there is no serious 
error in describing the process as a wandering ,of the nuclei, as 
Malls does in the case of a similar phenomenon in the endocar- 
dium. In both cases mesenchyme is forming, here from mesoderm, 
there from endothelium; the latter therefore has not with its 
change of form lost the potentialities of mesoderm in this respect. 

An embryo* (Columbia Collection No. 539), in which two 
complete intersomitic clefts are present shows a marked increase 
in the hypectodermal mesenchyme. Its distribution is shown 
in figures 2 to 4. In this case as in the following models, the com- 
pact mesoderm is reconstructed in white, in contradistinetion to 
the mesenchyme which is green. The surface appears more 
uniform than is actually the case, for it was found impossible 
to retain the innumerable small strands of mesostroma and all 
the minute projections and irregularities that characterize the 
contour of the mesoderm in section (fig. 5). The model, therefore, 
shows only the topography of the mesenchyme. This is most 
abundant in the head region where the axial mesoderm is less 
compact than elsewhere, and along the lateral margin of the 
coelom in the plane of the somites and for some distance caudad. 
To a less degree, mesenchyme has formed from the parietal 
mesoderm and the dorsum of the somites. 

In section, the somites have regular contours, but elsewhere 
the mesoderm gives rise to innumerable strands of mesostroma 
joining it to ectoderm. The mesenchyme elements have in- 
creased, and in addition to individually projecting cells, which 
show every degree of emergence from the mesoderm, larger proc- 
esses of this layer extend into the hypectodermal space. These 


831912, Amer. Jour. Anat., vol. 13, p. 249. 
81] desire to express my appreciation and thanks to Prof. J. H. McGregor for 
the gift of this beautifully preserved embryo. 


VASCULOGENESIS IN THE CAT 41 


changes are at a maximum along the lateral margin of the coelom 

_where the parietal as well as the visceral layers are fairly resolv- 
ing themselves into mesenchyme, to an extent that threatens 
to interrupt their continuity—a conformation which character- 
izes this region until the coelom is fully formed. In individual 
sections there are many cells which appear free of the mesoderm 
save for cytodesmata. The study of serial sections shows that 
many of these are actually free (save for ceytodesmata), but that 
others are the tips of the processes of mesoderm just mentioned. 
While the mesenchyme continues to be recruited by the migra- 
tion of single cells, from this period on there occurs a more inten- 
sive process, in which projections are formed and separate en 
masse, and as these projections are inclined to the mesoderm at an 
acute angle and often extend through several sections, this phase 
of mesenchyme production may be described as delamination. 

In the embryo of four somites (Columbia Collection No. 409) 
there is an interval between the mesenchyme of the lateral sheet, 
and that in the head and in the region of the nephrotomes, occa- 
sioned by the relative unproductiveness of the parietal layer of 
the cardiac coelom, which is thin and compact. The two areas 
become continuous opposite the last somite (fig. 6). 

Our next embryo is of eight somites (Columbia Collection No. 
530). The two areas of mesenchyme are now closer together and 
the parietal layer of the cardiac coelom is active in the produc- 
tion of this tissue (fig. 7). 

In both of these embryos we find the same intimate relation 
between mesoderm and mesenchyme that has been described. 

There are evidences of active migration of single energids into 
the mesostroma assisted by a coarser process of delamination. 
In addition, there are evidences of the inception of endothelium. 
Small clefts or cavities have appeared here and there amid the 
cells of the mesenchyme, which are thus coming to enclose minute 
vesicles. As these cells are in a syncytium, the question of the 
locality of the cleft, whether inter or intra-cellular is of secondary 
importance. The situation of these incipient vesicles (vaso- 
factive cells) is indicated by the yellow color in figure 7. 


42 VASCULOGENESIS IN THE CAT 


In figures 8 and 9 are shown sections of a vesicle in relation to 
the second nephrotome of the four somite embryo, which it will 
be seen has some superficial resemblance to a pronephric anlage. 
There is also a smaller one opposite the incipient fifth somite 
and a solid mass opposite the first somite. In the eight somite 
embryo, there is a similar sac-like anlage at the interval between 
the first and second somites; elsewhere along the nephrotomes 
are scattered mesenchyme cells. The formation of clefts has also 
been initiated in the lateral portion of the mesenchyme at several 
points. 

The comparison between these embryos does not suffice to 
give a local individuality to the several endothelial anlages, it 
merely shows the inception of the latter tissue diffusely from 
mesenchyme. In older embryos the condition is found repeated- 
ly and appears to be an intermediate stage between mesenchyme 
and endothelium, as it certainly is intermediate in form between 
the mesenchyme mass and the endothelial sacs; these cells of 
intermediate form are here designated vasofactive. 

In the umbilical line, the formation of endothelium is acceler- 
ated cephalad; almost none is found caudal of the fifth somite in 
the ten somite embryo (Columbia Collection No. 532). Its 
distribution is shown in figure 10. There is abundant mesen- 
chyme in the hypectodermal space as far caudad as the third 
somite. In its midst, but tending to a deep position near its 
junction with the compact mesoderm, are groups of cells with 
more abundant protoplasm which stains a little faintly. They 
form small masses, in intimate syneytial connection with the 
ordinary mesenchyme. In their interior are frequently small 
clefts. In some, the cavity is larger and the enclosing cells are flat- 
tened, eventually becoming thin and endothelioid (fig. 11). These 
last we may call angiocysts, using Bremer’s term. The presence 
of several minute clefts or vacuoles in a single mass of vasofac- 
tive cells, suggests their formation by the running together of 
these small cavities. In figure 10 the ordinary mesenchyme is 
omitted, the clefts and cavities are shown in red, the protoplasm 
of their bounding cells in black. Opposite the fourth and fifth 
somites are scattered cells and cell-groups of a similar appearance 
in which cavities have not yet appeared. These are readily dis- 


VASCULOGENESIS IN THE CAT 43 


tinguished from the other mesenchyme cells by their transverse 
position, in addition to their slightly more abundant protoplasm 
and slightly paler stain. It is possible that the latter two char- 
acteristics may be due to the imbibition of fluid. The anlages 
of the umbilical line are disposed along the lateral margin of the 
coelom. A mesal line of similar anlages is forming above the 
nephrotomes, while a few scattered elements have an intermediate 
position. In all these situations the process is discontinuous, 
the masses of vasofactive cells as well as the cavities are separated 
by appreciable intervals in which ordinary mesenchyme is present. 
They can be made into a continuum only by the inclusion of this 
tissue. The result would be a sheet of mesenchyme studded with 
scattered endothelial anlages as in the model of the twelve so- 
mite embryo (figs. 12, 13). 

A second embryo of twelve somites (Columbia Collection No. 
540) is shown in figure 14 for comparison with the embryo just 
described. The discontinuity of the anlages is again apparent. 
In addition to those before mentioned in the umbilical line, in . 
the cardinal line (along the nephrotomes) and intermediate 
between the two (somato-pleuric plexus), an angiocyst has formed 
just in front of the first somite. The transverse vasofactive 
cells have increased in number and are effecting junctions with 
one another and with the angiocysts. In a third embryo of 12 
(Columbia Collection No. 536) somites, there were rather more 
numerous vasofactive cells, rather less developed endothelium 
but the general arrangement was very similar; the angiocyst in 
front of the first somite was also present. 

The transverse position of the vasofactive cells requires ex- 
planation. They stand out clearly against the other mesenchyme 
cells which have their long axis in general radial, and are further 
distinguished by their rather deep position, near or in the almost 
cell-free interval between the bulk of the mesenchyme and the 
compact layer of parietal mesoderm. The nephrotomes are now 
resolving rapidly into mesenchyme, they are segmented, and are 
separating from the coelomic mesoderm; through the spaces thus 
formed, which contain only a loose mesostroma, the hypecto- 
dermal and hyperentodermal regions communicate. In the latter 


44 VASCULOGENESIS IN THE CAT 


the aorta, in the former the umbilical angiocysts are enlarging. 
The distending fluid must be derived from the adjacent region 
and its determination to these reservoirs must set up currents in 
the neighborhood. It is suggested that the conformation of the 
region (fig. 15) is favorable to the direction of such a current 
along the dorsal surface of the lateral mesoderm in the narrow 
interval between it and the bulk of the somatopleuric mesenchyme 
and that to this current the vasofactive cells owe their trans- 
verse orientation. Mesal to the coelom where the nephrotomes 
are resolving and where the current, if there be one, is directed 
ventro-mesad towards the aorta, the vasofactive cells have an 
oblique position. I have already referred to Bartels’ view that 
the addition of cells to the tips of lymphatics is occasioned by 
currents; it would be satisfactory to think the same principles 
at work in both late and early stages of vascular formation and to 
find that the current, the sine qua non of the persistence of endo- 
thelium, is also its determining cause. 

The vasofactive cells are dorso-ventrally flattened, they occur 
singly, in groups of two or three, continuous with, or wholly sepa- 
rate from the angiocysts. Where joined to the latter they could 
of course, in section be taken for sprouts, in spite of their flat 
plate-like form. But if more caudal regions of the same embryo 
are examined, or a corresponding region in a younger embryo, 
isolated flattened cells are found constantly, which I take to be 
evidence that the continuous plates, or if you will, cords are 
built up by the accretion of separate cells. When it is borne in 
mind that the distance from the umbilical to the cardinal line 
can be spanned by a half dozen of these cells, the opportunity of 
finding the angiocysts of these lines, continuous by means of strands 
of ‘angioblast,’ is not wanting. In one case (an embryo of 16 
somites, Columbia Collection No. 551) I found such a cord ex- 
tending from the umbilical vein to the aorta, which could have 
been taken to prove the origin of the aorta from the umbilical 
vein or vice versa by one who was apt at interpretation and had 
refrained from examining earlier stages. 

In the 14 somite embryo (Columbia Collection No. 188) the 
umbilical vein has become organized in its whole length; there 


VASCULOGENESIS IN THE CAT 45 


are scattered but not very numerous anlages in the cardinal line, 
and in the interval between the two. Many in the last position 
are now coalescing with the umbilical vein, which is assuming 
its function as a somatopleuric drainage channel. Candally it 
turns dorsad in the wall of the amnio-embryonie cavity and termi- 
nates among vasofactive cell-masses. Conditions opposite the 
first somites are shown in figure 16. The dilated cephalic ex- 
tremity of the umbilical vein shows evidence of its origin by 
coalescence in the presence of several perforations and septa. 
At the level of the third somite it has effected a connection with 
the omphalo-mesenteric vein which crosses it ventrad at right 
angles. The intervening compact mesoderm lateral to the coe- 
lom has given way over a small area (fig. 17) through which the 
anastomosis between the veins is established. Near the cephalic 
margin of the third somite a transverse vessel is forming and is 
continued forward obliquely to the front of the first somite where 
it is continuous with the channel along the neural tube. A com- 
parison of these vessels with the angiocysts of figures 10 and 14 
makes evident their correspondence in position. There is further 
a small communication between the transverse vessel and a branch 
of the aorta. Several small and as yet unannexed vesicles are 
present. 

The whole length of the umbilical line in a twelve somite em- 
bryo (Columbia Collection No. 547) is shown in figures 12 and 
13. The somatopleuric mesenchyme is seen from below, the 
mesoderm having been omitted as far mesad as the nephrotomes. 
Cephalad the vein is represented by a large multilocular cyst 
which extends from the limb bud, in which is a blood island, to 
the level of the fifth somite. Then follow a series of angiocysts, 
which are coalescing opposite the sixth and seventh somites but 
farther caudad are scattered and often widely separated. Vaso- 
factive cells amid which clefts have appeared are shown in yel- 
low. There are numerous discrete vesicles in the cardinal line, 
which being dorsal to the nephrotomes are not well seen in the 
figure. Opposite the last somites the mesenchyme forms a net- 
work of cords, the solidity of which is necessarily exaggerated 
by the reconstruction. The resemblance to Bremer’s angioblast 
cords is striking. 


46 VASCULOGENESIS IN THE CAT 


Caudad the anlages of the umbilical line are in the wall of the 
amnio-embryonic cavity (figure 18). The amnionic position 
of vascular rudiments has been described by Bonnet, Spee, and 
Keibel.** Although the position of the umbilical in the cat is 
really at the amnio-embryonic junction, and appears to be un- 
usual only because of the moderate entypy of the blastoderm 
and the early closure of the amnion caudad, there are in addition 
endothelial sacs actually in the amnion itself as may be seen in 
figure 19 from a 16 somite embryo. 

The foregoing data seem fully to warrant the conclusions that 
endothelium is formed directly from mesenchyme in the soma- 
topleure and that vessels result from the conerescence of discrete 
anlages. With these facts in mind we may approach the more 
difficult region of the splanchnopleure. 


THE EARLY FORMATION OF VESSELS ABOVE THE ENTODERM 


The formation of mesenchyme and endothelium in the splanch- 
nopleure is essentially the same process as in the somatopleure; 
the differences are mainly in the proportions of the two tissues 
for here endothelium predominates and is associated with the 
extresive production of blood cells. In brief, there is a separation 
of cells from the mesoderm, by individual migration and exten- 
sively in groups and masses; these elements are interposed between 
the entoderm and the compact mesoderm, that is they come to 
occupy and finally obliterate the hyperentodermal space. The 
major part give rise to endothelium, and in the ease of the larger 
masses (blood islands) to blood cells. Of the remainder, a con- 
siderable proportion subsequently become endothelium, but there 
is always a residuum of mesenchyme present. Later this becomes 
conspicuous and surrounds and embeds the vessels. 

The process is initiated by the formation of blood islands. 
These appear as spindle shaped discrete thickenings of the meso- 
derm at a considerable distance lateral and caudal to the embryonic 
axis, in regions where the mesoderm is free and separated by an 
interval, probably much increased by fixation, from the entoderm. 


8° Quoted from Ruckert, loc. cit., pp. 1253-4. 


VASCULOGENESIS IN THE CAT 47 


The thickenings soon come to project from the ventral surface 
of the mesoderm, and begin to separate from it. At about the 
time when the first somites are forming and before definite inter- 
somitie clefts have appeared, the migration of energids becomes 
extensive and desquamation is also well marked. The uncleft 
mesoderm, the visceral layer of the incipient coelom, and the in- 
termediate cell mass are all affected (fig. 20). Many strands 
of mesostroma cross the hyperentodermal space and among these 
the mesenchyme cells are scattered, singly or in flat plate-like 
groups. They show every degree of connection with the com- 
pact mesoderm, the plates in particular being often broadly con- 
nected. It is a repetition on a larger scale of the processes ob- 
served in the somatopleure, earlier in time, and more active. As 
yet no endothelium or blood cells have been formed. 

In the model of the embryo of two somites (figs. 21 to 23) the 
distribution of the tissue in the hyperentodermal space is shown 
in relation to the compact mesoderm (white)—the mesenchyme 
is green, endothelium red, and the intermediate forms of tissue 
(vasofactive cells) yellow. Cephalad the line of the heart, ven- 
tral aorta and first arch is occupied by scattered mesen hyme cells 
and cell groups. There are also a few isolated mesenchyme ele- 
ments of more mesal position. The line of the aorta is vacant 
to within a short distance of the first somite. Here scattered 
cells and plates are present. Ventral to the somites and imme- 
diately caudal to them an irregular continuous plate or strand 
has appeared. As yet this is connected with the more lateral 
mesenchyme in which the omphalomesenteric plexus is forming 
only at one point. There are, however, numerous patches of 
mesenchyme scattered in the interval. Caudad in the region of 
the primitive streak there is no axial mesenchyme. There is 
much mesenchyme, associated with vasofactive cells and endo- 
thelium, along the lateral margin of the coelom and extending 
mesad below the visceral mesoderm. This is most abundant 
caudad of the heart and the transverse plane of the somites, ap- 
proximately in the region which Schwink* regarded as the point 
from which endothelium sprouts. 


8 1891, Morph. Jahr., Bd. 17, p. 288. 


48 VASCULOGENESIS IN THE CAT 


The embryo of four somites (fig. 24) shows a very considerable 
advance beyond this condition and illustrates the rapidity, al- 
most abruptness, of the changes in vasculogenesis which have been 
described by Huntington.s’? The heart which is now concealed 
in the myocardium is composed of a series of endothelial vesi- 
cles the walls of which are in contact and are often connected 
with processes of the mesoderm. In general structure it is 
not dissimilar to the segment of the omphalomesenteric vein 
with which it is continuous caudad. Cephalad the line of the 
ventral aorta is represented by plates of mesenchyme, vasofac- 
tive cells, and small endothelial sacs. -A large multilocular cyst, 
showing evidences of concrescence in its septa, forms the cephalic 
element of the dorsal aorta. Laterad it is connected with the 
first arch which is already endothelial, and a similar connec- 
tion is established with the juxta-neural line, i.e., the angiocysts 
are in contact, but their lumina are still divided by numerous deli- 
cate septa (fig. 25). A second endothelial sac in the aortic line 
is situated just cephalad of the somites, ending near the first so- 
mite in continuity with a flattened plate of mesenchyme cells. 

Corresponding to the other somites are plates of mesenchyme, 
vasofactive cell masses with cleft-like lumina, and a few angio- 
cysts. The omphalomesenteric plexus is composed of larger 
vesicles which are now coalescing, but their lumina are not as 
yet continuous with either the heart or the aorta. Between the 
mesal vasculogenetic line (the dorsal aorta) and the lateral, which 
comprises the ventral aorta, the heart and the omphalomesenteric 
vein and plexus, are scattered mesenchyme and vasofactive 
cells; cephalad these are more abundant and are now coalescing 
to bridge the interval between the dorsal and the ventral aorta. 
The first arch is endothelial and situated somewhat in front of the 
point at which connection is established with the juxta-neural 
line. Two other transverse strands are present, one largely 
mesenchyme, one of vasofactive cells. 


% Developmental relations between lymphatic amd haemal vascular channels. 
Paper read at the XVII Internat. Med. Congress. London, August, 1913. To be 
published in the Am. Jour. Anat. 


VASCULOGENESIS IN THE CAT 49 


In the embryo of eight somites (fig. 26) a moderate advance has 
_ been made in organization and considerable changes in topog- 
raphy have been effected. The most striking at first glance is 
the increase in the internal limiting sulcus incident to the for- 
mation of the foregut. But it will also be observed that the heart 
has been carried forward and is now much nearer the blind ex- 
tremity of the foregut than in the four somite embryo. The first 
somite is also at a position cephalad to that of the younger embryo. 
This might be taken to imply an addition of somites at the cranial 
end of the series, a most difficult point to establish in the absence 
of fixed landmarks and in the shifting topography of this region. 
But this assumption would not explain the advance of the heart. 
The increase of the mesoderm dorsad, and the formation of the 
foregut seem to call for a rearrangement of this layer and may 
possibly account for the shortening actually observed, though 
this naturally does not exclude the possibility of the production 
of somites from the head mesoderm. 

The aortic angiocysts are confluent from the first arch to the 
fourth somite, and mesal to them are mesenchyme cells, many 
but not all of which are adherent to their walls. The heart is 
multilocular. The omphalomesenteric vein now has a continu- 
ous lumen and has effected several small connections with the 
omphalomesenteric plexus, which is now extensive. Amid its 
channels are scattered mesenchyme and vasofactive cells and 
numerous projections from the mesoderm, the tips of which on 
reaching the entoderm spread out, form horizontal processes, 
and partially enclose the endothelial channels (fig. 27). Sub- 
sequently these projections are resolved into mesenchyme in 
which the vessels are imbedded. 

In the embryo of twelve somites (Columbia Collection No. 
547) the dorsal aorta has acquired a continuous lumen from the 
first arch to the region of the primitive streak. Caudad it di- 
minishes in diameter and terminates in connection, with scattered 
angiocysts and vasofactive cells, which farther caudad give place 
to mesenchyme (fig. 13). The first arch is double, one arm pass- 
ing laterad, one cephalad to the lateral cul de sac of the foregut. 

“The ventral aorta is endothelial and at the sides of the gut is 


MEMOIR NO. 3 


50 VASCULOGENESIS IN THE CAT 


connected with plates and strands of mesenchyme, which are the 
presumptive anlages of the aortic arches and the adjacent mesen- 
chyme. Amidst this tissue are a few blood islands. 

The precise source of the cells which form the endothelium of 
the aorta is somewhat difficult to determine. The clear cut con- 
tour of the somites until the formation of the sclerotomes makes 
it difficult to believe that cells are here given off to the mesen- 
chyme. Mollier has held that its material is derived from the 
intermediate cell masses, and has shown sections in which they 
are connected to the aorta by solid processes, similar to those 
which elsewhere give rise to mesenchyme and endothelium. 

Such connections also are found in the eat. At the stage of 
eight somites, and more clearly in that of twelve (figs. 12 and 13) 
the aorta becomes joined to a more lateral point of the visceral 
mesoderm. From the convexity which this layer forms below 
the coelomic angle a ridge of cells projects ventro-mesad. This 
is separated in many places from the mesoderm sensu strictu 
by a narrow cleft, and thereupon comes into apposition with the 
aorta(figs. 29 to 33); in places the cells of this ridge become endo- 
thelial and contain a lumen which eventually communicates with 
that of the aorta. It is posssible that the connection with the 
intermediate cell mass described by Mollier may be of a similar 
nature, and that in both cases we are dealing with an increment 
to already formed endothelium. The anlages of the aorta both 
in front and behind the somites are freely connected with the over- 
lying paraxial mesoderm and it is probable that they are directly 
derived from it prior to its segmentation. Later when the scler- 
otomes are formed the aorta enters into intimate relation to 
their loosely arranged cells (figs. 29 to 33). The addition of 
elements from this source to the aortic endothelium has been de- 
scribed by Mollier,’* and his results are abundantly confirmed 
by conditions in the cat. Some of these cell masses acquire a 
lumen before addition to the aorta. In other sections the vessel 
loses its well defined outline mesad and its wall is completed by 
typical mesenchyme, with the interstices of which its lumen com- 


8 Loc. cit., p. 1267. 


VASCULOGENESIS IN THE CAT 5 


municates. Finally the mesal angle of the aorta is often occupied 
by masses of adherent bood cells, as though the sclerotome had 
given rise not only to endothelium but blood cells as well. 

To summarize: the formation of endothelium in the splanchno- 
pleure is precisely analogous to that in the somatopleure. Mesen- 
chyme is formed in situ and secondarily in large part is transform- 
ed into endothelium. There are few solid strands of mesenchyme 
(‘angioblast’) and these are not primary but arise from the con- 
crescence of scattered plates of mesenchyme. They are rarely 
transverse, as though they had grown in, but conform to the 
lines of future vessels. More important than their frequent 
interruptions as evidence of in situ origin, are their innumerable 
connections with the compact mesoderm, every degree of emer- 
gence from which, both by migration of single energids and by 
delamination, is conspicuous in every embryo examined. 


ADDENDUM 


Since the foregomg paper has been in proof a communication made 
by Bremer at the thirtieth session of the A.A.A. in Philadelphia and 
published in abstract (Proc. Am. Ass. Anat., Anat. Rec., vol. 8, no. 2) 
makes a brief consideration of the results of this investigator desir- 
able. In his first paper (Am. Jour. Anat., vol. 13, 1912) Bremer 
describes the development of the intraembryonic blood vessels in the 
rabbit as resulting from the ingrowth of a continuous network of 
“angioblast cords” derived from the extraembryonic plexus of the 
yolk-sac. Although dealing in this paper with the development as 
found in rabbit embryos he examined other species, as chick, pig, 
sheep, etc., and felt “satisfied that in all essential points the story 
of the development of these primary vessels in other vertebrates will be 
found similar. .’ The evidence on which these conclusions rest 
is briefly the continuity of the aortic and cardiac angiocysts with 
the extraembryonic vessels of the splanchnopleure by solid cords of 
cells, which are interpreted as angioblast. The essential data are 
given succinctly and the principle of his interpretation made clear 
in the description and figure of the rabbit embryo of 5 segments. “Fig. 
1. a reconstruction of the angioblast cords of one side of a rabbit em- 
bryo of five segments, shows these cords, streaming in from the net- 
work over the yolk-sac. . . . . They have grown from left to 
right of the figure, occasionally anastomosing until near the median 
line, which lies at the left of the figure. The shape of the meshes of this 
net indicates it seems to me the direction of this growth and the rapidity 
with which it has occurred.” I question both the angioblastic nature 


52 VASCULOGENESIS IN THE CAT 


of these cords and the principle on which they are interpreted as in- 
growths. Continuity of structure at a given period of development can 
not be held to establish the presence of continuity at earlier stages, for 
it may be secondarily achieved, as is shown by the union of the meso- 
nephroic tubules with the pronephric duct, to instance but one of sev- 
eral similar junctions in the development of the genito-urinary tract. 
In this particular case I must consider both the assumption of primitive 
contimuity and of ingrowth wholly invalidated by the experiments of 
Hahn (Arch. f. Entw. Mech. I Organismen Bd. 27, 1909) and the 
recent very satisfactory confirmatory results of Miller and Me- 
Whorter, which after some amusing editorial adventures in the 
office of the American Journal ot Anatomy were presented at the 
thirtieth session of the A.A.A. (Proc. Am. Ass. Anat., Anat. Rec., vol. 
8, no. 2,p.91). As to the nature of the cellular cords designated angio- 
blast, it is to be noted that Bremer does not figure or describe either 
mesostroma or mesenchyme in his‘embryos. As he does not distinguish 
these formations, it is possible that they are included in his angioblast 
cords, which would then represent an early for the most part pre- 
vascular stage of the splanchnopleuric mesenchyme. I am inclined 
further to this view by the loss of cords in the embryo of 6 to 7 seg- 
ments as evidenced in the comparison of figures 1 and 3, and to which 
Bremer calls attention in his text. The most plausible explanation 
would be that they had resolved themselves into mesenchyme; that 
therefore the primitive network was mesenchyme, which in widely 
separated regions was partially transforming itself into endothelium. 

I am interested to note that in his second paper, a highly interesting 
study of early vessels in man (Proe. Am. Ass. Anat. 30th Sess. Anat. 
Rec., vol. 8, no. 2, p. 97) Bremer abandons his original position and 
accepts the development of discontinuous vascular anlages independent 
of the vessels of the yolk-sac, but nevertheless retains the principle 
of growth in continuity in the vascularization of the embryonic body. 
With the general principles of blood vessel origin from mesoderm I am 
in hearty agreement, as also with the multiplicity of anlages, and their 
intimate relation to the coelomic mesoderm. In these points there 
now seems to be a very general agreement between the students of 
mammalian vasculogenesis. But I must question the validity of a 
continuous angioblast network “composed of angiocysts and _ solid 
cords,” which appears superfluous in view of Bremer’s conclusion that 
vessels “originate from ingrowths of mesothelium as a number of sepa- 
rated cords, angiocysts, or blood islands, which are soon connected by 
sprouts of endothelium.” In the interpretation of the solid strands 
connecting angiocysts as ‘angioblast’ Bremer states that he relies on 
his earlier work upon the rabbit, and I would again express my dissent. 
The mesenchyme and mesostroma form a protoplasmic continuum in 
which are contained angiocysts. To designate such portions of it as 
intervene between two angiocysts “angioblast’’ is in my opinion to make 
a distinction which does not exist, and which would not suggest itself 
in regions where the mesenchyme preponderates over the blood vessels 
as in the somatopleure. 


PLATES AND FIGURES 


Fig. 1 Transverse section of an embryo with one pair of intersomitic clefts. 
Columbia collection No. 554, slide 4, row 5, section 8, X 300. Reduced +. 


1 Ectoderm 5 Hyperentedermal mesenchyme 
2 Mesoderm 6 Mesostroma 

3 Entoderm 7 Coelom 

4 Hypectodermal mesenchyme S Neural plate 


Figs. 2 to 4 Dorsal view of a model of the mesoderm and mesenchyme of 
an embryo of two somites. Columbia collection No. 539, X 300. Reduced 3. 
The mesoderm sensu stricto is white, the mesenchyme green. 


1 First intersomitie cleft ft From this point caudad the meso- 
2 Second intersomitic cleft derm is continuous across the me- 
3 From this point craniad the meso- dian line 


derm is continuous across the me- 
dian line 


Fig. 3 Dorsal view of a model of the mesoderm and mesenchyme of an embryo 
of two somites. 1, First intersomitic cleft; 2, second intersomitic cleft. 


Pig. 4 Dorsal view of a model of the mesoderm and mesenchyme of an embryo of 
two somites. 4, From this point caudad the mesoderm is continuous across the 
median line. 


5 


He Oo bo 


lig. 5 Transverse section of an embryo of two somites at the’ level indi- 
cated in Figure 3. Slide 5, row 1, section 6, & 300. Reduced +. 


Hetoderm 5 Hyperentodermal mesenchyme 
i Coelom 


7 Neural plate 


Mesoderm 


Entoderm 
Hypectodermal mesenchyme 


ig. 6 Dorsal view of model of mesoderm (white), mesenchyme (green) and 


vasofactive cells (yellow) of an embryo of four somites. Columbia collection No. 
409, X 300. Reduced }. 


bo 


Somite 1 4 Site of quintal ganglion 
Anlages of juxta-neural anastomosis 5 Endothelial anlage opposite the 
Connection between the juxta-neural second nephrotome; cf. figures 8 
anastomosis and the aorta, ef. fig- and 9 
ure 25 6 Region of cardiac coelom 


60 


Fig. 7 Dorsal view of model of mesoderm (white), mesenchyme (green), 
vasofactive cells (yellow) and endothelium (red) of an embryo of eight somites. 
Columbia collection No. 530, < 300. Reduced 3. 


1 Somite 1 4 Region of cardiac coelom 
2  Anlages of juxta-neural anastomosis 
3 Connection of juxta-neural anasto- 

mosis with the aorta 


8 


Figs. S and 9 Successive transverse sections of Embryo No. 409 passing 


through the second somite. Left side. Slide 4 row 6, section 12 and row 5, 
section 1, X 800. Reduced }. To show derivation of endothelium from interme- 


diate cell mass. 


1 Eetoderm 

2 Parietal mesoderm 
3. Visceral mesoderm 
4 Entoderm 

5 Coelom 

6 Mesostroma 

7 Dorsal aorta 


8 Intermediate cell mass 

9 Projection from intermediate cell 
mass 

10 Endothelial vesicle continuous with 
the projection from intermediate 
cell mass of the preceding section 


Fig. 10 Graphic reconstruction showing the distribution of endothelium 


and vasofactive cells in the somatopleure of an embryo of ten somites. The cavi- 


ties of angiocysts and clefts in vasofactive cells are shown in red. Columbia col- 
lection No. 532. Left side, dorsal view, X 300. Reduced 4}. 


1 Mesal margin of coelom 
2 Lateral margin of coelom 
3 Angiocysts of umbilical line 


t Angiocysts of cardinal line 
5 Somite 1 


i Somite 5 


SN a GT 


10 
bo 


11 


Fig. ll Transverse section of same embryo at level indicated in preceding 
figure, X 300. Reduced }. Slide 6, row 2, section 8. 


L Aorta { Angiocyst of cardinal line 
2 Omphalomesenteric vein 5 Angioeyst of intermediate position 
3 Angiocyst of umbilical line 6 Coelom 


Figs. 12 and 13. Ventral view of model of mesoderm, mesenchyme and vas- 


cular anlages of an embryo of twelve somites. Columbia Collection No. 547. The 
coelomic mesoderm from the nephrotomes laterad has been omitted, to allow the 
hypectodermal mesenchyme (green) to be seen from below, with its endothelium 
(red), vasofactive cells (yellow) and blood islands (blue), * 300. Reduced >. 
The umbilical vein extends into the incipient limb bud. 


L Dorsal aorta 6 Cells derived from visceral meso- 
2 Truncus arteriosus derm in process of accretion to 
3 Hypectodermal mesenchyme the aorta cf. figure 28 

{ Angiocysts of umbilical vein 7 Mesenchyme about foregut contain- 
5 Coelomic angle ing anlages of ventral aortic root 


and caudal aortic arches. 


66 


MEMOIR NO. 3 


lig. 13 Ventral view of model of mesoderm, mesenchyme and vascular anlages 
of an embryo of twelve somites. 1, Dorsal aorta; 3, hypectodermal mesenchyme; 
4, angiocysts of umbilical vein; 4, coelomic angle. 


68 


—FIG.15 


14 


Fig. 14 Graphic reconstruction showing the distribution of the endothelium 
and vasofactive cells in the somatopleure of an embryo of twelve somites. The 
cavities are shown in red. Columbia Collection No. 540, left side, dorsal view, 
x 3800. Reduced f. 


1 Mesal margin of coelom 4 Angiocysts of cardinal line 
2 Lateral margin of coelom 5 Somite 1 
3 Angioecysts of umbilical line 6 Somite 4 


Fig. 15 Transverse section of same embryo at level indicated in figure 14. 
Slide 38, row 2, section 10, *300. Reduced +. 


1 Vasofactive cells 3  Coelom 
2 Aorta 


Fig. 16 Graphic reconstruction of the vessels of the somatopleure and of a 
portion of the aorta from an embryo of fourteen somites. Columbia Collection 
No. 188, left side, dorsal view, * 300. Reduced }. 


1 Umbilical vein 5 Communication between the aorta 
2 Site of communication of the um- and the umbilical vein 

bilical with the omphalomesen- 6 Aorta 

teric vein 7 Isolated angiocysts 
3 Anterior jugular vein 8 Somite | 


4 Communications of the anterior 9 Somite 4 
jugular vein with the juxta-neural 


anastomosis 
70 


Fig. 17 


ceding 


Transverse section of the same embryo at level indicated in pre- 


figure, to show the communication between the umbilical and omphalo- 
mesenteric veins. Slide 4 


1 Aorta 


2 Dorsal segmental artery 


, row 4, section 6, * 300. Reduced }. 


5 Juxta-neural vessel 
6 Somite 4 


7 Coelom 


3 Omphalomesenteric vein 
4 Umbilical vein Section of vessel of cardinal position 


~I 
te 


show the 


lig 18 Transverse section of an embryo of twelve somites to 
amniotie position of the caudal anlages of the umbilical vein. Columbia Collec- 
tion No. 547, slide 5, row 2, section 7, X 300. Reduced ¢. 
Funnel-like diverticulum of coelom 


| Umbilical angiocyst | 
between vessels of the splanchno- 


2 Amnion 


3 Mesostroma pleure 


*} poonpeyy 
“€ MOI G apl[s ‘Tee “ON UOT}DOT[OD BIquINjor 


VOISOA [BIyJopuy T 
‘OES X “WOTULUE OY} UT OPIISOA [VIPOYJOpUS poyejost MoYs 04 ‘y UOTZV—aS 
‘SOPMMOS UD}XIS JO OAIGUIO UY JO UCI}I0S oSMoASUBIT, 6G] ‘“SIq 


74 


Transverse section from embryo in which the somites are forming, 


Fig. 20 
Columbia Collection No. 5: 


but no complete intersomitic cleft has appeared. 
slide 3, row 4, section 6, * 300. Reduced }. 


| Eetoderm { Hyperentodermal mesenchyme and 
2 Mesoderm incipient blood islands 
3 Entoderm 5 Coelom 


Figs. 21 to 23 Ventral view of a model of the mesoderm (white), splanchno- 
pleuric mesenchyme (green), vasofactive cells (yellow) and endothelium (red) 
of an embryo of two somites. Columbia Collection No. 539, X 300. Reduced 3. 


1 Anlages of ventral aorta and heart 5 From this point craniad the meso- 
2 Anilages of dorsal aorta derm is continuous across the 
3. First intersomitic cleft median line 

{ Second intersomitic cleft 6 rom this point caudad the meso- 


derm is continuous across the me- 
dian line 


76 


EE 


Fig. 22. Ventral view of a model of the mesoderm (white), splanchnopleuric 
mesenchyme (green), vasofactive cells (yellow) and endothelium (red) of an 
embryo of two somites. 2, Anlages of dorsal aorta; 3, Virst intersomitic cleft; 


4, second intersomitic cleft. 


78 


Fig. 23 Ventral view of a model of the mesoderm (white), splanchnopleuric 
mesenchyme (green), vasofactive cells (yellow) and endothelium (red) of an 
embryo of two somites. 6, From this point caudad the mesoderm is continuous 
across the median line, 


80 


Fig. 24 Ventral view of a model of the mesoderm (white), hyperentodermal 
mesenchyme (green) vasofactive cells (yellow) and endothelium (red) of an embryo 
of four somites. Columbia Collection No. 409, * 300. Reduced 4. 


1 Anlages of the dorsal aorta ! Position of heart 
2 Virst aortie arch 5 Omphalomesenteric vein 
3 Anlages of ventral aorta 6 Omphalomesenteric plexus 


83 


MEMOIR 


Vig. 25 Transverse section of an embryo of four somites at the level indicated 
5 2 
by leader 3 in figure 24. Columbia Collection No. 409, slide 5, row 6, section 3, 


< 300. Reduced }. The several angyocysts are in contact but their lumina are 


not continuous. 


1 Aorta { Vasofactive cells attached to juxta- 
2 Juxta-neural line neural vessel; this ecleft-like lu- 
3 Quintal anlage men is not continuous with that 


of the vessel 


ig. 26 Ventral view of a model of the mesoderm (white), mesenchyme (green), 
vasofactive cells (yellow) and endothelium (red) of an embryo of eight somites. 
Columbia Collection No. 5380, * 300. Reduced }. 


1 Dorsal aorta { Projections of visceral mesoderm 
2 Omphalomesenterice vein which later become connected 
3° Omphalomesenteric plexus with the aorta 


5 Position of heart 


S4 


4 3 4 Sia ; ao m2 1 


Fig. 27 Transverse section of an embryo of eight somites, Columbia Collec- 


tion 530, slide 7, row 4, section 4, * 300. Reduced 1. 


1 Aorta t Projections of mesoderm between 
2 Coelom blood vessels 
3 Omphalomesenteric plexus 


S6 


iv} 
| ——_____— 


bo 
= 


4 
28 


Figs. 28 to 33 Serial transverse sections of an embryo of twelve somites, Col- 


umbia Collection No. 547, slide 7, row 2, sections 7 to 12, & 300. Reduced +. 
1 Aorta t Notocord 
2 Cells derived from visceral meso- 5 Entoderm 
derm in process of accretion to 6 Coelom 
the aorta 7 Angiocyst added to aorta; a rem- 
3 Cells of sclerotome in syncytial con- nant of septum is still visible 


tinuity with endothelium of aorta 


=A 2 


Fig. 29 Serial transverse section of an embryo of twelve somites. 2, Cells 


derived from visceral mesoderm in process of accretion to the aorta. 


8S 


{ 
2 3 2 


lig. 80 Serial transverse section of an embryo of twelve somites. 2, Cells 
derived from visceral mesoderm in process of accretion to the aorta; 3, cells of 
sclerotome in syncytial continuity with endothelium; 7, angiocyst added to aorta; 
a remnant of septum is still visible. 


89 


oy) D2 
“ “ 


ig. 3L Serial transverse section of an embryo of twelve somites. 2, Cells 


derived from visceral mesoderm in process of accretion to aorta, 


90 


| 
| 
2 2 


Vig. 32. Serial transverse section of an embryo of twelve somites. 2, Cells 


derived from visceral mesoderm in process of accretion to aorta. 


91 


| 


pe =- i 
Wig. 33 Serial transverse section of an embryo of twelve somites. 2, Cells 
derived from visceral mesoderm in process of accretion to aorta. 


QL Schulte, Hermann von 

835 Wechlinger 

S35 Early stages of 
vasculogenesis in the cat 


PLEASE DO NOT REMOVE 
CARDS OR SLIPS FROM THIS POCKET 


UNIVERSITY OF TORONTO LIBRARY 


Sensei 
SSS