Difference between revisions of "American Journal of Anatomy 5 (1906)"

From Embryology
Tag: Replaced
Line 2: Line 2:
{{Historic Disclaimer}}
{{Historic Disclaimer}}
{{Amer. J Anat. Volumes}}
{{Amer. J Anat. Volumes}}
[[Category:Journal]][[Category:Historic Embryology]][[Category:1900's]][[Category:USA]]
Universitij nf Wisco>ni)i.
Wisiar Institute of Aiiatotiiy .
Harvard Univemitij.
University of California.
Cornell Universitij.
University of Jlichiyan.
Columbia University.
Johns Hopkins University .
University of Jfic/iiyaii.
Harvard University.
University of Pennsylvania.
HENRY McE. KNOWER, Secuetaky,
Johns Hopkins University.
J 906
Zf)t fviiUnvoaii Company
/a ^»
No. 1. December 1, 1905.
I. John Warken. The Developnicnt of the Paraj^hysis and
the Pineal liegion in Nectiinis Maculatus .... 1
With 23 text figures.
II. E. T. Bell. The Development of the Thymus ... 39
With 3 plates and 5 text figures.
III. Jeremiah 8. Ferguson. The A^eins of the Adrenal. . G3
With 3 text figures.
IV. George Walker. The Blood Vessels of the Prostate
Gland 73
With 3 colored plates.
V. Bennet M. Allen. The Emhryonic Development of the
Pete-Cords and Sex-Cords of Chrysemys 79
Witli 1 doul)]e plate and G text figures.
VI. Frederic T. Lewis. The Development of the Lymphatic
System in Pabbits 95
With 8 text fi,gures.
VII. Frederic T. Lewis. The Development of the Veins in
the Limbs of Rabbit Embryos 113
With 1 text figure.
No. 2. May 31, 1906.
A'lTI. Poss Granvtllk IIarhison. Further Experiments on the
Development of Peripheral Nerves 131
With 5 figures.
iv Contciils
IX. Ai.i!i:i;t C. Eyclesiiymi:!; .-iiid James Meredith Wilson.
The (lastnilation and I'iinbrvo Formation in Amia
C'alva 13;i
With I (lonl.lc plalcs.
X. Charles F. W. McClure. A Contribution to the Anatomy
and Development of the Venous System of Didelpliys
Marsiii)ialis (L.)-— I'iH-t II. Development .... lOo
Willi 1 single and 4 double plates and 27 text fi2:ures.
No. 3. July 25, 190G.
XL Franklin P. Mall. A Study of the Structural Unit of
the Liver ^ 227
With 74 fio-ures and 7 tables.
XII. Albert C. Eycleshy'mer. The Development of Clnoma
tophores in Necturus 309
With 7 figures.
XIII. Sidney Klein, S. M., ]\L D. On the X^ature of the
Granule Cells of Paneth in the Intestinal Glands of
Mammals 315
AVith .5 figures.
XIV. Charles L. Edwards and Clarence W. LTahn. Some
Phases of the Gastrulation of the Horned Toad, Phrynosoma Cornutum Harlan 331
With 15 text figures.
No. 4. September 1, 1906.
XV. Robert Bennett Bean. Some Racial Peculiarities of
the Negro Brain ... 353
With 16 fig-ures, 12 charts, and 7 tables.
Contents ' v
XVT. Franklin P. Mall. On Ossification Centers in Human
Embryos Less Than One Hnndred Days Old 433
With (i text figures and 7 tables.
XVII. J. L. Bremer. Description of a 4-inm. Human Embryo. 459
With 16 text figures.
XVIII. Charles R. Stockard. The Development of tlio Mouth
and Gills in Bdellostoma Stouti 481-517
With 36 figures.
Decemher 27, 28, and 29, 1905 I-XX
Demonstrator of Anatomy, the Anatomical Laboratory, Harvard Medical
With 23 Text Figures.
The presence of the paraphysis in Necturns was noted by Prof. C. S.
Minot in his article " On the Morphology of the Pineal Region, based on
its Development in x'Vcanthias " (38), and a brief description of certain
stages given. C. L. Herrick (15, PI. YIII, Fig. 1, 3, 4) gives a brief
account of the adult paraphj'sis, and shows it in the above figures, vvlicre
it is named " Preparaphysis." Osbom (31, PI. IV) shows th-e paraphysis in an adult brain in comparison with the brains of other amphibia. Kingsbury (21) describes briefly the adult paraphysis as well
as a few of the earlier stages, and also gives an account of the epi))hysis
and the plexuses. I have found, however, no detailed account of all the
stages in the development of the paraphysis and the pineal region. This
term is used here in the same sense as in Minot's article, quoted above.
The greater part of the specimens studied for this article were taken
from the Embryological Collection of the Harvard Medical School, and
the numbers of each section used are given. Other specimens were prepared specially for this purpose. In some cases where the plane of section was uneven, two or more sections were used in drawing the figures
ill order to show all the structures, which should appear in the uK^dian
Fig. 1 is a median sagittal section through the brain of an einl)ryo of
8-9 mm. I am indebted to a colleague for the drawing of this section,
as this stage is wanting in the collection. In the roof of the fore brain
three arches are seen. From before backward these are the j)aia|)liysal
arch, P. A., the post- velar arch, P. V. A., and the epiphysal arch, Ep. A.
The first two are separated by the velum transversum. V, which marks
the limit between the two subdivisions .of the fore brain. Hence the
paraphysal arch belongs to the telencephalon, the other two to the
2 Paraphysis and the Pineal Ees^ion in Necturns Maenlatus
dieneephalon. The epiphysal are]i is homided l)y two angles, which represent the position of the future supra and posterior commissures. The
velum transversum is a simple infolding of the brain roof, and consists
of two distinct layers, one caudad and one cephalad, the space between
them being filled by a loose mesenchymal tissue, which later contains
numerous blood vessels. This figure is practically identical with Minot's
figure of acanthias of the same stage (28, Fig. 1), and is, therefore, of
great importance in showing the homologies of these parts in elasmobranchs and amphibians. It is probable, as Minot states, that these
arches occur in most of the vertebrate series.
The term post-velar arch, introduced by Minot (28), is much better
for purposes of description than the terms " zirbelpolster " of German
writers, and the " dorsal sack " or " postparaphysis " of American authors.
Fig. 1.
Fig. 2.
Fig. 1. Embryo of 8-9 mm. Sagittal section, X 63 diams.
Fig. 2. Embryo of 10 mm. Harvard Embryological Collection, Sagittal
Series, No. 269, Section 39, X 63 diams.
Fig. 2 represents the roof of the dieneephalon and telencephalon of
an embryo of 10 mm. The two layers of the velum are nearer together
and in the region of the epiphysal arch are seen the first signs of the
epiphysis, E. This structure is a small rounded diverticulum, which
arises from the cephalic end of the arch. It is hollow and opens into
the cavity of the fore brain.
Fig. 3 is a similar section of an embryo of 12 mm. The velum is a
trifle longer and the epiphysis a little larger than in the preceding figure.
Immediately cephalad to the velum a very small evagination in the
paraphysal arch can be seen, P. This is the first sign of the paraphysis,
and it appears distinctly later than the epiphysis. The latter overlaps
its short stalk both caudad and cephalad, and at this stage the stalk is
still hollow, though its cavity was obliterated in this section.
John Warren 3
Fig. 4 is a section of an embryo of 13 mm. The velum is again a
little longer and its caudal layer is now distinctly thinner than its cephalic
layer. The paraphysis is now a well-marked narrow diverticulum extending dorsad from the paraphysal arch parallel to the velum. The
paraphysal arch just cephalad to the opening of the paraphysis has been
Fig. 3.
Pig. 4.
Fig. 3. Embryo of 12 mm. Harvard Embryological Collection, Sagittal
Series, No. 49, Section 58, X 63 diams.
Fig. 4. Embryo of 13 mm. Harvard Embryological Collection, Sagittal
Series, No. 598, Sections 71 and 75, X 63 diams.
forced downward to a slight degree, as there is relatively more space
between it and the ectoderm than in the previous figures. The epiphysis
is about the same size as in Fig. 3, and its opening into the brain is
clearly seen.
Fig. 5 is a section of an embryo of 12A mm., which is, however, further
advanced than that of Fig. 4. The velum, the post-velar arch, and the
epiphy.sis are about the same, but the paraphysis is distinctly longer, and
Fig. 5. Embryo of 12.4 mm. Harvard Embryological Collection, Sagittal
Series, No. 675, Section 57, X 63 diams.
Fig. 6. Same as Fig. 5, X about 120 diams.
has become a narrow tube. The brain roof cephalad to it has descended
still more into the cavity of the telencephalon and the opening of the
paraphysis is much nearer the tip of the velum. Fig. 6 is the same
section as Fig. 5, only drawn on a higher scale to show the histological
details. The walls of the paraphysis and "velum consist of a single layer
of cells, with large oval nuclei and without very distinct cell boundaries.
4 rai;i])li\si> and the Pineal Kegion in Necturus Maculatus
These cells are, of course, continuous with tliose which form the brain
wall in this region. The same is true of the epiphysis, but the walls seem
thicker, as the organ has been cut somewhat obliquely. Close to the
paraphysis two vessels can be seen, a larger one cephalad and a much
smaller one caudad, _ F<?6'. The vessels lie in intimate relation to this
structure, and it is important to note their relation at this early stage,
because as development progresses the relation between paraphysis and
blood vessels becomes more and more intimate.
Fig. 7 is a section of an embryo of 15 mm. The most striking feature
here is the increase in size of the paraphysis. which has become a long
tube with a lumen extending its entire length, and at its distal end a lateral diverticulum has appeared. The roof of the fore brain has now
descended to such a degree that the opening of the paraphysis is on a level
with the tip of the velum. The velum itself has lost its cephalic layer,
and consists of one .layer only, which, however, is much longer than the
velum in Fig. 5. If Figs 4, 5, and 7 are compared it will be seen that
Fig. 7. Embryo of 15 mm. Harvard Embryological Collection, Sagittal
Series. No. 79, Sections 85 and 89, X 63 diams.
the distal end of the paraphysis is practically at the same distance from
the ectoderm in each case. As the paraphysis has developed during those
stages into a long tube, its growth must have occurred by a downward
extension of the neighboring parts into the cavity of the fore brain. This
is practically the same process described by Minot in Acanthias. It is
also shown by the great increase in distance between the roof of the
telencephalon and the ectoderm from Fig. 4 to Fig. 7. The opening of
the paraphysis in Fig. 3 is nearly on a level with tlie base of the velum,
and as the down growth of the parts takes place the opening of the paraphysis and the paraphysal arch descend, apparently pushing the cephalic
layer of the velum ahead of them. Therefore the single layer of the
velum in Fig. 7 really corresponds to the original caudal layer, plus the
cephalic layer, which has been forced down ahead of the opening of the
In stud5''ing Fig. 7 it might seem as if the posterior wall of the paraphysis corresponded to the cephalic layer of the velum. This, however, is
John Warren 5
not tlie case, as can be seen in a wax reconstruction of the parts, Fig. S.
This is a reconstruction of the brain of an embryo of 14.5 mm. The tops
of the hemispheres, H, have been removed to give a clear view of the
paraphysis, P, which otherwise would be more or less covered in by them.
The paraphysis appears as a straight tube in the median line and caudad
to it is seen a broad partition, V , extending the whole width of the diencephalon. This is the velum, consisting of one layer only, which represents the tw^o originally distinct cephalic and caudal layers. The down
grow^th of the parts in order to provide room for the development of the
paraphysis has formed a deep angle in the roof of the fore brain. This
Fig. 8. Wax model of brain of embryo of 14.5 mm. Harvard Embryological Collection, Sagittal Series, X 120 times.
angle is bounded caudad by the velum and cephalad or ventrad by the
narrow roof of the telencephalon (paraphysal arch) immediately
cephalad to the paraphysis. As the hemispheres develop, they grow at
first in a dorsal direction and occupy the space left by the formation of
this angle, so that the paraphysis is practically buried between the hemispheres in front and the velum behind. Fig. 10. The growth of the
paraph3^sis must, therefore, be regarded as liaving an important effect on
the development of the fore brain at this stage.
Up to this stage the development of the velum has been in a ventral
direction towards the floor of the fore brain, but now it begins to grow in
quite a ditferent direction. In Fig. 7 a distinct bulging of the velum is
6 Paraphysis and tlio Pineal Pogion in Xeclurus 'Maculatus
seen, which is extending candad at ncarl}- a right angle to its previous
line of growth. If the roof of the telencephalon be closely examined a
slight bulging will be seen just cephalad to the opening of the paraphysis.
These two outgrowths into the fore brain mark the beginning of the
choroid plexuses, which, therefore, have in their origin a very intimate
and definite relation to the opening of the paraphysis, one arising caudad
and the other cephalad to it. The epiphysis at this stage has increased
considerably in size, and the cavity in its stalk is now permanently obliterated. The body of the organ overlaps the stalk a little behind, and is
beginning to grow well forward of it. The posterior commissure, P. C,
appears here for the first time, a distinct interval in the roof of the brain
lying between it and the stalk of the epiphysis.
Fig. 9. Embryo of 17.5 mm. Harvard Embryological Collection,
Series, No. 540, Sections 113-115, X 63 diams.
Fig. 9 is a section through the brain of an embryo of 17.5 mm. The
paraphysis has increased in length, and from its distal end, which is
somewhat enlarged, small tubules are given off. The whole tube is tipped
somewhat forward. The choroid plexus is now well developed, and consists of two distinct parts, one dorsal and one ventral. The dorsal part
corresponds to the velum, which has grown caudad as far as the mid brain
and has absorbed a large part of the post-velar arch. The ventral part is
developed from the original paraphysal arch, and is growing towards the
floor of the fore brain. Burckhardt (3) refers to these plexuses as
" plexus medius " and " plexus inferioris," respectively, and Mrs. Gage
(13), who studied them in Diemyctylus, where the anatomical conditions
closely resemble those of JSTecturus, names them the " diaplexus " and
" prosoplexus." Prof. Minot has suggested the terms diencephalic
John Warren 7
plexus for the dorsal part, and tel encephalic plexus for the ventral part,
and I shall use these terms, as they express more clearly the exact origin
of each plexus.
The diencephalic plexus, V. Plx., appears as a large wedge-shaped mass
covered by a thin layer of cells, and consisting of a loose connective tissue
in the interstices of which numerous blood corpuscles can be seen. The
telencephalic plexus, Tel. Plx., has the same general characteristics as
the diencephalic. The epiphysis has become flattened and more elongated, and is attached by a narrow stalk to the brain wall. The supra
commissure, S. C, is seen just cephalad to the stalk of the epiphysis,
which is prolonged forward above it. I was unable to obtain any sagittal
series between 15 and 17.5 mm., but in a transverse series of 16.5 mm. the
first traces of this commissure can just be made out. and therefore it
Fig. 10. Wax model of brain of embryo of 18 mm. Harvard Embryological Collection, Frontal Series, No. 850, X about 75 diams.
probably appears between 16 mm. and 17 mm. as a rule, but at these early
stages there is a good deal of variation in the development of all these
parts. The posterior commissure is rather larger than in the previous
Fig. 10 is the drawing of the model of the brain of an embryo of
18 ram. This model is intended to show the circulation of the paraphysis
at this stage. The distal end of the paraphysis, P, is surrounded by a
venous circle, from either side of which veins, Ves., run outward and backward just caudad to the hemispheres, H, to terminate in the internal
jugular vein, I. J. V. This vein is passing backward external to the fifth,
V, and seventh, VII, cranial nerves. Fig. 6 showed the intimate relation
of the paraphysis to these vessels at 12.4 mm., and when the sections of
this series were followed out it was found that here the vessels surrounded the tip of the paraphysis. It seems that as the paraphysis devel
S Parapliv.sis and the Pineal Region in Nectnms Maculatus
ops it forces its way into the veins lying over this part of the fore brain,
and the tnbnles, as they are given off at la.ter stages, force their way into
these veins, Fig. 15, forming the sinusoidal type of circulation described
by Minot (29) and Lewis (25). From the venous circle shown in Fig. 10
smaller vessels run down along the sides of the paraphysis and anastomose with the vessels of the choroid plexuses. A vessel also runs back to
the epiphysis, and a larger one forward between the hemispheres. The
circulation of this region appears at this stage to be mostly venous, as I
could trace the arteries only to their point of entrance in the anlage of
the skull, and the return circulation probably occurs by means of a
minute capillary network over the surface of tlie brain.
Fig. 11 is a section through the brain of an embryo of 26 mm. The
Fig. 11. Embryo of 26 mm. Harvard Embryological Collection, Sagittal
Series, No. 377, Sections 125 and 126, X 63 diams.
paraphysis here is much more developed. It inclines somewliat forward,
and from its wide central lumen a number of tubules are given oif in
every direction. The epiphysis and the commissures show but little
change. The striking feature of this figure is the great development of
the plexuses. The diencephalic plexus, D. Pl.r.. has grown through the
mid-brain nearly to the hind-brain, and the telencephalic plexus, Tel.
Fix., has grown downwards into the depths of the cavity of the fore brain
towards the infundibular recess.
Fig. 12 is a transverse section of an embryo of 26 mm., corresponding
approximately to the line A-B, Fig. 11. The section passes through the
epiphysis, E, and the supra commissure, S. C, just beneath it. Then
through the diencephalic plexus, D. Fix., and that part of the cavity of
the diencephalon between this plexus and the roof, T>ien. The section
John Warrcu 9
then passes through the paraphysis at a point where two small tubules are
given off, then through the telencephalic plexus, Te2. Plx.. the telen
FiG. 12. Embryo of 26 mm. Harvard Embryological Collection, Transverse Series, No. 376, Section 89, X 63 diams. (See line A-B, Fig. 11.)
cephaloii, Teh. the lateral ventricles, L. V., and the foramina of Munro,
F. M. In this section the plexuses of tlie hemispheres, L. Plx., are seen.
They arise on either side of the origin of the telencephalic plexus, and
Fig. 13. Embryo of 26 mm. Harvard Embryological Collection, Frontal
Series, No. 378, Section 138, X 63 diams. (See line C-D, Fig. 11.)
pass outward at right angles to it through the foramina o£ Munro into the
lateral ventricles.
Fig. 13 is a frontal section through an embryo of 2Q mm., corresponding
Paraphysis aud the Pineal Kegion in Necturus Maculatus
closely to the line C-D, Fig. 11. The section passes through the paraphyses, P, and a large lateral tubule, and then through the entire length
Fig. 14. Same Series as Fig. 13. Section 108. (See line E-F, Fig. 11.)
of the diencephalic plexus, D. Fix., the distal end of which is here enlarged and has reached to the hind brain, H. B. Fig. 14 is of the same
^ m
Tel.Plx. .^1-^^
Fig. 15. Same as Fig. 11. X about 150 diams.
series as Fig. 13, and corresponds approximately to the line E-F, Fig. 11.
It passes through the telencephalic plexus, Tel. Fix., the plexus of the
John Warren
hemispheres. L. Plx., and the lateral ventricles, L. V. It shows clearly
how the plexuses of the hemispheres arise from the telencephalic plexus
and pass at first outward and then forward through the foramina of
Munro towards the cephalic extremity of the lateral ventricles.
Fig.- 15 is a high power drawing of Fig. 11, magnified 150 diams.
The wall of the paraphysis consists of a single layer of cells with large
oval nuclei, and these cells are continuous with the cells covering the
choroid plexuses, but the latter are flatter and form a thinner layer. On
either side of the paraphysis two large vessels are seen, ves., the epithelial
cells of which lie directly against the wall of the paraphysis. The little
tubules seem to be forcing their way into these vessels, which are branches
of the vessels seen in Fig. 6 and Fig. 10. The vessels also pass down into
the choroid plexuses. Fig. 16 is a section of an embryo of 31.4 mm.
Fig. 16. Embryo of 31.4 mm. Harvard Embryological Collection, Sagittal
Series, No. 537, Sections 119-122, X 63 dlams.
The general arrangement is practically the same as in Fig. 11, except that
all the parts have progressed somewhat in their development. The distal
end of the paraphysis has begun to grow distinctly more cephalad, and the
whole structure is much larger than at 26 mm. The choroid plexuses are
more extensive, and from the diencephalic plexus a prolongation is extending downwards towards the telencephalic plexus. This latter has
pretty well filled up the depths of the third ventricle, and from it prolongations dip down into the recesses in the floor of the fore brain. The
two commissures are practically the same as they were at 26 mm., and
12 Paraph vsis and the I'iiieal Heo-ion in Noetiiriis Maeulatus
though the epiphysis is a little larger it has been displaced considerably
candad, as this part of these sections was unluckily somewhat injured.
Fig. IT is a section through the
brain of an adult necturus. This
drawing is magnified 38 diams. only,
as it was too large to draw on the
same scale as the preceding figures.
The paraphysis, P. forms a very
complex structure extending far forward above and between the hemispheres. It consists in a general Avay
of a proximal and a distal part. The
former is broad and thick, and oxtends forward and upward. It then
turns forward at quite a marked
angle to form the distal part, which
is narrow and. tapering. The central
canal in the proximal part is very
large and irregiilar, but in the distal
portion much narrower. From all
parts of this canal a large number of
tubules are given off, which extend
in every direction, and between
which lies a confused mass of bloodvessels. One sees here a large vessel
ventrad to the organ, and a smaller
dorsad to it. the same relations as
appear at 12.4 mm., Fig. 6. From
these vessels branches pass into the
choroid plexuses.
Fig. 18 is a transverse . section of
an adult brain corresponding approximately to the line A-B, Fig. 17.
This is drawn on the same scale as
most of the preceding figures, 63
diams. The paraphysis, P, is seen in
the median line between the hemispheres. It shows a distinct central
eavit}', with many tubules running out in every direction, between which
is a mass of blood-vessels of all sizes. Below a portion of the telencephalic plexus aiu1 the ])l(\\-us of the hemispheres are seen.
Via. 17. Brain of adult necturus.
Sag-ittal Section, x 38 diams.
John Warroii
Fig. is. Brain of adult necturus. Transverse section, X 63 diams. (See
line A-B. Fig. 17.)
14 Paraphysis and the Pineal Eegion in Nectunis Maciilatus
Fig. 19 is from a wax recoustruction of tlie adult paraphysis on a
scale of 120 diams., made from the same series as Fig. 18. The angle
between the proximal and distal parts is quite striking, and is much more
marked in Ichthyophis (Burckhardt, 4, Fig. 1), but of course this division
into proximal and distal parts is really a purely arbitrary one. This
model gives a good idea of the complex structure of the organ. The
tubules are of all shapes and sizes, often convoluted and anastomosing
with each other. The spaces between them, which the vessels occupy, are
quite large and striking.
Fig. 19. Wax model of paraphysis of adult necturus, same series as Fig.
18, X about 120 diams.
Pig, 20 represents a small portion of the paraphysis of Fig. 17, magnified 560 diams., and shows clearly the relation of the tubules to the
vessels. In the centre of the figure is a tubule, T, dividing into two
branches, T^ T-. Surrounding these tubules on every side are sinusoids,
si., whose flat endothelial cells are seen lying directly against the epithelial
wall of the tubules with no connective tissue between them. We find
here in order first a sinusoid, then a tubulcj then another sinusoid and
another tubule, and finally a sinusoid. The wall of the tubules consists
of a single layer of cells with large oval nuclei and very indistinct cell
boundaries. The nuclei contain masses of granules arranged very irregu
Joljn Warren
larly. There can be no question abont the glandular nature of the paraphysis, and its circulation is evidently sinusoidal.
The choroid plexus, Ch. Fix., Fig. 17, appears as a confused mass of
Fig. 20. Small portion of adult paraphysls, same section as Fig. 17, X 560
vessels covered by a thin layer of cells. This mass completely fills up
the cavity of the fore and mid brains, and may in some cases appear in
the hind brain, Fig. 23, though there seems to be considerable variation
Fig. 21. Wax model of epiphysis of adult necturus. X 280 diams.
in the caudad development of this part of the ple.xus. Tlie two parts
of the plexus overlap each other, and are also closely interlaced.
The epiphysis, E. is still attached to the brain by a very narrow stalk.
The body overlaps the stalk somewhat behind, and then is prolonged
forward as an oval flattened body above the roof of the diencephalon, and
PiiTajiln'sis and tlic Pineal Rogig^i in Xccturus Maculatns
its cavity seems to be divided more or less into compartments. Fig. :H
represents a wax reconstrnction of an adult epiphysis seen from ^above.
It is irregularly triangular in shape with a broad base and a blunt apex.
Its surface is grooved more or less by vessels which lie against its Avails.
Fig. 22 is the same model with the top removed. The interior is more
or less subdivided by incomplete septa. At its apex there is a small cavity
Fig. 22. Same as Fig. 21, with top of epiphysis removed.
1)0unded behind by a partial septum, then comes a large chamber, which
divides into two passages running back towards the angles at the base.
Between these two passages appears a comparatively solid area, interrupted, hoAvever. to some extent by small spaces, aa^IucIi communicaio Avith
each other and the larger chambers. This solid area lies over tbe stalk
of the organ. The sujDra commissure appears to be comparatively small,
Fig. 23. Brain of adult necturus. Viewed from above. X 7 diams.
while the posterior is large and forms a deep groove in the roof of tlie
brain. Fig. 23 is a vioAv of the brain of an adult necturus shoAving the
relative positions of paraphysis and epiphysis. The tufted extremity
of the diencephalic plexus can be seen in the fourth ventricle.
If Fig. 1, the embryo of 8-9 mm., is compared Avith Fig. 17, the adult,
one sees that the paraphysal arch has been Avholly taken up in the
formation of the telencephalic plexus, the plexus of the hemispheres, and
John Warren 17
the paraphysis. The velum and the greater part of the post-velar arch
have been absorbed in the formation of the diencephalic plexus. A
portion of this arch, however, persists and forms that part of the roof
of the diencephalon between the diencephalic plexus and the supra commissure. The epiphysal arch has formed the epiphysis.
The paraphysis is a structure common to all vertebrates either in the
adult or embryonic condition (Selenka, 34, Francotte, 11), but previous
observations on mammals leave much to be desired. It always arises from
the telencephalon cephalad to the velum transversum, and its opening
is placed between and dorsad to the foramina of Munro as emphasized
by Dexter (5).
In the cyclostomes, Ammocoetes (Kupffer, 24), and Petromyzon
(Burckhardt, 3), the paraphysis appears as a small sac-like diverticulum
lying ventrad and close to the enlarged distal end of the epiphysis. In
elasmobranchs, Minot (28) and Locy (27) found that the paraphysis
in Acanthias appears at quite a late stage as a small outgrowth from
the paraphysal arch and, owing to the small size of the post-velar arch
and the compression of the velum, it comes to lie immediately cephalad to
the epiphysis. In ganoids, Kupffer found in Accipenser that the paraphysis appears first as a small outgrowth which later becomes a somewhat
sacculated vesicle (23, Fig. 19). Hill (18) and Eycleshymer and Davis
(9) studied the paraphysis in Amia. Here it begins as a simple vesicle,
which increases rapidly in size and gives off diverticuli from its central
cavity. In teleosts (Burckhardt, 3) the paraphysis appears late and
remains in a rudimentary condition. In the dipnoi, Burckhardt (3) describes the paraphysis in Protopterus as a wide outgrowth giving off small .
In amphibia the organ becomes highly differentiated and its appearance in the adult brain is very striking. It appears as an elongated body
lying above and between the hemispheres, and extending cephalad for
a varjdng distance in various forms. Fig. 23. Osborn (31, PI. 4) shows
a view of the brains of Siredon, Necturus, Proteus, and Siren. The
paraphysis has the same general form in each of these, but it is somewhat
larger in Necturus. The paraphysis of Triton and Ichthyophis (Burckhardt, 4) has the same characteristics. In the latter the paraphysis appears in sagittal section as a hammer-shaped organ extending forward
above the hemispheres. (4, Fig. 1). In Rana the paraphysis has the
same position as in Necturus, but is smaller. On removing the top of the
skull in Necturus the paraphysis is seen lying beneath the pia surrounded
by the blood-vessels which cover this part of the brain. It appears to the
naked eye so vascular and also in sections so intimately related to the
18 Paraphysis and the Pineal Pegion in Necturns Maculatus
choroid plexus that it is not astonishing that it was at first regarded as
a portion of this plexus. According to Minot (28) the paraphysis of
Eana is characterized by the character of its epithelium, its tubular structure and its apparently sinusoidal circulation. This is practically similar
to the conditions found in Necturus.
In ^lenopoma Sorenscn (36) describes the paraph3'sis as a solid vascular mass, and in Ichthyophis Burckhardt (3) describes it as an
elaborately folded structure of a glandular character. In Amblystoma,
Eycleshymer (8) shows that the organ gives off tubules and has a digitated appearance. In Diemyctylus (Mrs. Gage, 13), in an embryo of
10 mm. the paraphysis closely resembles that of Necturus at 18 mm.
(Minot, 28, Fig. 13), and in the adult it is a long tube giving off many
tubules in close relation to vessels.
Herrick (15) calls the paraphysis of an adult Necturus the " preparaphysis " and says it consists of an irregular central chamber with
complicated diverticuli in close relation to vessels. This description
corresponds closely with Fig. 17. The model of the adult paraphysis,
Fig. 19, shows the complicated arrangement of the tubules, many of
which anastamose with each other, and the spaces between the tubules
are filled by blood-vessels.
Fig. 6 shows how the paraphysis at 12.5 mm. is beginning to invade a
large vessel lying over it on the surface of the brain. This vessel at this
point is much enlarged. Fig. 10 shows this relation much clearer and
also that these vessels in relation to the paraphysis are tributaries of the
internal jugular vein. Fig. 15 shows the relation of the paraphysis to
these vessels at 26 mm. and that the vessels pass into the choroid plexuses
both dorsad and ventrad to the organ. The little tubules can be seen growing out into the vessels. Fig. 17 shows how these relations between vessels
and tubules in the adult become much more intimate, and the vessels
corresponding to those in Fig. 15 are seen passing into the choroid plexuses dorsad as well as ventrad to the paraphysis. Schobel (33), who
studied the circulation in the brain of certain amphibia, of which
Necturus was one, shows that in the adult two large vessels pass outward
just caudad to the hemispheres to empt}'^ eventually in the internal
jugular vein. These vessels surround the paraphysis and anastomose
with two or three large vessels running forward between the hemispheres.
This is practically the same arrangement shown in the model in Fig. 10.
He does not mention the paraphysis but refers to it as a large venous
plexus. Eex (32) also has studied the veins in the amphibian brain,
and his preparations are practically the same as Schobel's. He refers
to the paraphysis as the " nodus chorioideus " and says that it is a sort
John Warren 19
of meeting point for the veins of the fore and mid brains. He shows
beautifully in his injections how veins pass both dorsad and ventrad to
the paraphj'sis to enter the plexuses, and how closely these vessels are
related to the tubules of this structure. Mrs. Gage shows practically the
same arrangement in Diemyctylus
As regards the arteries Schobel shows that they are much smaller than
the veins, and describes a small vessel passing caudad to the hemispheres
to pass eventually into the plexuses. The intercrescence of the tubules
of the adult paraphysis and the veins is shown clearly in Fig. 30, each
vessel and tubule lying back to back with no connective tissue between
them. In view of all these facts it seems evident that the circulation of
the paraphysis is sinusoidal. According to the above descriptions, the
development of the paraphysis into a complicated, glandular organ, which
is also very vascular, seems to be a striking characteristic of the amphibia.
In lacertilia the paraphysis of Anguis fragilis has been studied by
Francotte and of Lacerta vivipara by Francotte (10, 11, 12) and
Burckhardt (3). The latter shows the paraphysis in an embryo of
13 mm. as a narrow tube with a slightly expanded distal extremity,
much as that of ISTecturus of 15 mm. Francotte describes the paraphysis
of Lacerta vivipara as a long tube giving off a mass of tubules which
lies under the parietal eye, and resembles the epiphysis of birds (11, Fig.
14; 12, Fig. 24). In Anguis (10, Figs. 15 and 19) the paraphysis forms
a long narrow sack, with somewhat convoluted walls, which curves back
over the post-velar arch to end in close relation to the parietal eye.
The conditions in the lizard are essentially the same (10, Fig, 31).
In Phrynosoma coronata Sorensen (35, Fig. 2) describes the paraphysis
as a long, narrow tube, immediately cephalad to the epiphysis.
In the ophidia Leydig (26, Fig. 6) shows the paraphysis of an embryo
of Vivipara urcini near birth as a large, wide tube with no convolutions
or diverticuli and practically the same conditons in a young " Eingelnatter" and Tropidonatus natrix (26, Figs. 5 and 2).
Among the chelonia Voelzkow (39) has described the paraphysis in
Chelone imbricata as at first a wide tube much convoluted, which later
decreases somewhat in size. (Figs. 21 and 22.) Its distal end inclines
caudad close to the epiphysis. In Chelone mydas Humphrey (20, Fig. 7,
PI. II) shows the paraphysis as a long tubular structure, giving off small
tubules, and in an embryo of Chelydra it appears as a large, wide sack,
from which tubules arise. It is in closer relation to the epiphysis than
in Chelone mydas or imbricata.
In Cistudo Herrick shows a model of Sorensen (15, Fig. 5, PI. VI)
of the paraphysis, which is a wide tube with convoluted walls and tubules.
20 Paraphysis and the Pineal Region in Necturus Maculatus
In the crocodilia Voelzkow (39) has described the paraphysis of
Crocodilus madagascarensis grand and Caiman niger spix. In the former
the paraphysis is at first a wide tube which becomes convoluted and much
longer and narrower. In the latter the paraphysis forms a larger tube
and the convolutions and tubules are more complicated. In both cases
the organ reaches its greatest development in embryonic life and retrogrades later, though more so in the crocodile. He was unable, however,
to follow the development in the caiman as far as in the crocodile.
Owing to the thickenings in the brain wall the organ is crowded somewhat caudad against the post-velar arch.
In birds the paraphysis is relatively rudimentary. Burckhardt (3)
shows the paraphysis in an embryo of the crow as a small diverticulum
not unlike that of Petromyzon. Dexter (5) worked out in detail the
development of the organ in the common fowl, and showed that it appeared at first as a small diverticulum. The walls become much thickened and in a chicken of 10 days it is a small, oval structure, about 150 fx
in its greatest diameter, with very thick walls (5, Fig. 5). Selenka (34)
has described the paraphysis in the oppossum, but as far as I am aware
little is known of the development of the paraphysis in mammals, though
Francotte (11) has observed it in human embryo of twelve weeks.
From the cyclostomes to the amphibia the paraphysis shows a steadily
progressive development, and the various forms through which it passes,
from the simple diverticulum of Petromyzon to the elaborate gland of
the urodela, are illustrated in a general way by the stages of its development in Necturus. In the vertebrates above the amphibia the paraphysis
retrogrades and practically retraces its steps through the reptilia and
birds to mammals, reaching in the chick essentially the same form in
which it started in Petromyzon. Its development, therefore, may be
indicated by a curve, which ascends steadily from the cyclostomes, reaches
its height in urodela, and descends through the reptilia and birds to
The epiphysis is present in nearly all vertebrates. It is stated to be
absent in the alligator (Sorensen, 36 and 37), and in the caiman and
crocodile (Voelzkow, 39) and in Torpedo (d'Erchia, 7).
The epiphysis of Necturus as compared with the paraphysis is relatively
poorly developed and in this respect resembles the epiphysis of other
urodela (Mrs. Gage, 13). In Diemyctylus Mrs. Gage found that the
epiphysis was entirely cut off from the brain and that its cavity was
nearly obliterated. In Ichthyophis (Burckhardt, 1 and 4) the epiphysis
is a small, pear-shaped organ attached to the brain by a narrow solid
stalk. Herrick (15) in Menopoma, describes the epiphysis as an irreg
John Warren 21
ular number of vesicles attached to the brain by a narrow opening. According to Kingsbury (21) the structure in Necturus consists of an
aggregation of closed vesicles, forming an oval, flattened body, and there
is no connection with the brain. The cavity of the epiphysis communicates through its stalk with the cavity of the diencephalon up to 15 mm.,
when the cavity in the stalk becomes obliterated. The stalk persists and
was present in all the adult brains which I examined, but in some cases
it was so small that it could easily be overlooked. The reconstruction
of an adult epiphysis. Fig. 22, shows that the cavity of the organ forms
a large chamber subdivided to a certain extent by incomplete septa. A
much more solid area is seen towards the caudal extremity, which is
placed just over the stalk. The same characteristics I have observed in
another model made from a different brain. One gets the idea that the
epiphysis consists of a series of vesicles in studying sagittal sections a
little to one side of the median line, as for instance in Fig. 17, where the
epiphysis was displaced a little to one side.
. There has been such a vast amount written on the origin of the
epiphysis and the pineal or parietal eye and their homologies that it seems
superfluous to add anything more here. In a very general way, however,
there seems to be some sort of proportion in the relative development of
the paraphysis, epiphysis, and the parietal eye. In urodela where there
is no parietal eye and a small epiphysis, the paraphysis reaches its highest
degree of development. In those forms where the paraphysis is rudimentary or relatively slightly developed the parietal eye is present or
else the epiphysis is relatively highly developed. Compare, for example,
the figures of Burckhardt (3) of Petromyzon, Minot (28) of Acanthias,
Burckhardt (3) of Trout, Leydig (26) and Voelzkow (38) of reptilia,
and Dexter (5) of the fowl. Rana, however, seems to be a marked exception to this statement, as there the paraphysis, epiphysis, and pineal
eye are all present and well developed, and the same may be said for
Lacerta (Francotte, 10 and 11) and Sphenodon (Dendy, 6). As the
paraphysis and epiphysis are glandular structures they have probably
some sort of compensatory function and where one is highly developed
the other is relatively rudimentary or even absent. Compare in this
respect also Torpedo with Acanthias and the crocodile and alligator with
the chick.
As a rule the stalk of the epiphysis is placed immediately caudad to the
supra commissure, in all cases I believe, except in the toad, where there
is a distinct interval between it and this commissure (Sorenson, 36).
In Necturus there is an interval in the roof of the brain between the stalk
and the posterior commissure. This portion of the roof of the brain was
22 Paraphysis and the Pineal Eegion in Necturus Macnlatus
described by Kupffer (23) as the " schaltstlick," and according to him it
is best developed in amphibia. Burckhardt (3) maintains that it occurs
in all vertebrates from Pctromyzon np, but according to Kupffer it is
absent in Accipenser (23, Fig. 19), and it is also wanting in Acanthias
(Minot, 28, Pig. 10) and in the fowl (Dexter, 5, Fig. 9).
The velum transversum is probably characteristic of all vertebrates.
Minot (28). In Petromyzon the velum appears as a small transverse
fold, and the post-velar arch is well marked. The plexus development
is, however, very slight. In elasmobranchs the velum of Acanthias
forms a long, narrow, transverse fold, and the post-velar arch is so small
that the origin of the velum seems to be close to the supra commissure.
The caudal layer of the velum is distinctly thinner than the cephalic
(Minot, 28, Fig. 6). This is also seen in Torpedo (d'Erchia, 7, Fig.
12), and in Necturus, Fig. 6. The velum later on has the character of
a choroid plexus, but the plexus of the hemispheres is very rudimentary
(Minot, 28). In Notidanus Burckhardt (3) shows a long, narrow velum,
a short post-velar arch, and a small telencephalic plexus. The plexus
of the hemispheres, however, is absent.
In Accipenser (Kupffer, 23, Fig. 19) the velum is long, well developed
and folded to a certain extent, and the post-velar arch is quite extensive.
In ganoids (Studnicka, 38) the membranous roof of the brain serves as
the tela choroidea of higher types. In this class of vertebrates according
to Burckhardt (3) the plexus of the hemispheres is lacking, but the
telencephalic plexus is well developed, and in teleosts the former is also
wanting, but the latter present in a reduced form.
In amphibia all the plexuses are highly developed, and in J^ecturus
they are of marked extent (Kingsbury, 21). The velum in Necturus
appears at first as a transverse fold in the roof of the brain separating
the diencephalon from the telencephalon. This fold develops at first
ventrad and then caudad through the mid brain as far as the hind brain.
This great growth of the velum forms the diencephalic plexus. The
post-velar arch, which at first is wide and well marked, is practically
absorbed by the overgrowth of the velum, and a small portion only
persists in the roof of the diencephalon between the origin of the diencephalic plexus and the supra commissure. Fig. 17. The telencephalic
plexus develops from the paraphysal arch immediately cephalad to the
paraphysis, the opening of which therefore is surrounded by these two
plexuses. They fill up the cavity of the third ventricle and mid brain,
and the diencephalic plexus may appear in the hind brain (Osl>orn, 29).
This seems to vary in different cases, and in the majority of brains which
I was able to examine the extremity of this plexus did not actually extend
John Warren 23
into the hind brain. In- Fig. 33, however, this extremity appears as a
marked tuft in the fourth ventricle. The plexuses of the hemispheres
arise on either side from the origin of the telencephalie plexus and pass
into the lateral ventricles, extending nearly to their cephalic extremities.
In Lacerta vivipara (Francotte, 12, Fig. 24) the post- velar arch has
been much com})ressed from before backward so as to form a deep narrow
angle. At the apex of the angle the folds of the diencephalic plexus are
seen. The velum is smooth and apparently is not included in the
formation of the plexus. In Anguis fragilis (Francotte, 10, Figs. 19
and 15), the post-velar arch does not seem to be so much compressed,
and the plexus formation somewhat greater. As he says, however, the
development of those parts in Lacerta is practically the same as in Anguis.
According to Burckhardt the telencephalie plexus is much reduced
in size, consisting merely of small folds, but the plexus of the hemispheres is well developed (Burckhardt, 3).
In the turtles Humphrey (20) found that the velum of Chelydra is
but slightly developed, and no diencephalic plexus is formed. All the
other plexuses are telencephalie in origin.
Herrick (15, Fig. 5, PI. VI), shows in Cistudo a well-developed telencephalie plexus and a diencephalic plexus represented by many folds in
the caudal layer of the velum and the post-velar arch. In Chelone imbricata Yoelzkow (39, Figs. 19 and 22) shows at first a well marked
velum and a wide post-velar arch. In later stages the velum and practically all the arch are thrown into folds to form the diencephalic plexus.
The telencephalie plexus is also well developed. In the serpents much
the same arrangement can be seen. The velum (Leydig, 26, Figs.
2, 5, and 6) forms a prominent fold, and it and the post-velar arch
form a very vascular plexus. In the crocodilia (Voelzkow, 39, Figs.
7, 11, 13, 15), the velum and the post-velar arch are at first well
marked, but the parts later become so compressed from before backward that the arch forms a deep acute angle in the depths of which plexus
foldings are seen. The caudal layers of the velum, however, takes no part
in the plexus formation. In birds, Dexter (5) found that the velum of
the fowl is small and the post-velar arch broad at first. This becomes
compressed so as to form an acute angle much as in the crocodilia. The
cephalic limb of this angle and all the velum is converted into the choroid
plexus. In birds the plexus of the hemispheres is very well developed,
but the telencephalie plexus is practically absent.
In fishes the plexus development is quite simple, in many cases being
merely the thin membranous roof of the third ventricle; in others, however, this is much folded and vascular (Sorensen, 35). In certain forms
24 Paraphysis and tfio Pineal Eegion in Necturus Maculatus
there is a telenccphalic plexus, but the plexus of the hemispheres is
absent or rudimentary (Burckhardt, 3). In amphibia there is a great
overgrowth of all the plexuses, especially of the diencephalic plexus,
which here reaches its highest development. In reptilia the plexus of the
hemispheres is well developed, but the telencephalic plexus is reduced
in size (Burckhardt, 3), and the diencephalic much more so. In birds
the plexus of the hemispheres is highly developed, the telencephalic
plexus practically absent, and the diencephalic plexus, while very similar
to that of reptilia, approaches nearer to the tela choroidea of higher
Osborn first named the supra-commissure and worked out its homologies. According to him (30) the urodela are distinguished from the
anura by the frequent extensive development of this commissure, which
is large in Amphiuma, smaller in Necturus, and much reduced in Eana.
It appears in Necturus a little later than the posterior commissure, as
is usual in most cases, as far as I am aware, except in Ammocoetes, where
it appears shortly before the posterior commissure (Kupffer, 24, Fig. 5).
It is found in all the chief types of vertebrates, and is usually smaller
than the posterior (Minot, 28). It is developed from the diencephalon,
while the posterior belongs to the cephalic limit of the mid brain.
1. The paraphysis appears first in an embryo of 12 mm. It is developed from the telencephalon immediately cephalad to the velum transversum as a small diverticulum, which becomes eventually a complicated
gland with anastomosing tubules. The gland is very vascular, and has a
sinusoidal circulation.
2. The epiphysis appears first in an embryo of 9-10 mm., and is developed from the diencephalon. It is always attached to the brain by a
small solid stalk, and the cavity is partially subdivided by incomplete
3. The velum transversum grows at first ventrad and then caudad as
far as the hind brain, forming in this way the diencephalic portion of
the choroid plexus. The post-velar arch, which is at first quite extensive,
is almost entirely absorbed in this extensive growth of the v.elum.
4. The telencephalic plexus arises from the roof of the telencephalon,
and fills up the depths of the cavity of the third ventricle. The opening
of the paraphysis is surrounded by these two plexuses.
5. The plexus of the hemispheres arises at a right angle from the
telencephalic plexiis just cephalad and ventrad to the opening of the
John Warren 25
6. The supra-commissure appears first at 16-17 mm. It lies immediately cephalad to the stalk of the epiphysis and is comparatively small.
7. The posterior commissure appears first at 15 mm., and there is a
marked interval in the roof of the diencephalon between it and the
I wish in conclusion to express my acknowledgments to Prof. Minot
for his kind advise and interest in the preparation of this article.
The following are the principal articles consulted, but of course do not
form a complete bibliography of this subject:
1. BuRCKHAKDT, R. — Die Zirbel von Ichthyophis Glutinosus und Protopterus
Annectens. Anat. Anz., Bd. VI.
2. Die Homologien des Zwischenhirndaches bei Reptilien und Vogeln.
Anat. Anz., Bd. IX, 320-324.
3. Der Bauplan des Wirbeltiergehirns. Morpholog. Arbeiten, IV
Bd., 2 Heft, 131.
4. Untersuchungen am Gehirn und Geruchsorgan von Triton und
Ichthyophis. Zeitschr. f. Wiss. Zoologie, Bd. 52.
5. Dexter, F. — The Development of the Paraphysis in the Common Fowl.
American Journ. Anat., Vol. II, No. 1, 13-24.
6. Dendy, a. — On the Development of the Pineal Eye and Adjacent Organs
in Sphenodon (Hatteria). Quart. Journal Micros. Soc, Vol. 42, 111.
7. D'Ebchia, F. — Contributo alio studio della volta del cervello intermedio
e della regione parafisaria in embrioni di Pesci e di Mammiferi.
Monitore Zoologico, VII, 118 e 201.
8. Eycleshymer, a. C. — Paraphysis and Epiphysis in Amblystoma. Anat.
Anz., Bd. VII.
9. Eycleshymer, A. C, and Davis, B. M. — The Early Development of the
Paraphysis and Epiphysis in Amia. Journal of Comp. Neurology,
Vol. 7.
10. Francotte, p. — Recherches sur le developpement de L'epiphyse. Arch.
de biologie, T. VIII.
11. Note sur I'ceil parietal, l'epiphyse, la paraphyse et les plexus
choroides du troisieme Ventricule. Bull, de I'acad, royale, etc., d.
Belg., 3 Serie, T. 27.
12. Contribution a I'tHude de I'oeil parietal, de l'epiphyse chez les
13. Gage, S. P. — The Brain of Diemyctylus Viridescens. Wilder Quart. Cent.
Book, 1898.
14. Gaupp, E. — Zirbel, Parietalorgan und Paraphysis. Ergebn. Anat. Entwick.
Ges., VII, 208-285.
15. Hebbick, C. L. — Topography and Histology of the Brain of certain Rep
tiles. Journ. of Comp. Neurology, Vol. I, 37; Vol. Ill, 77-104, 119-138.
16. Topography and Histology of certain Ganoid Fishes. Journ. of
Comp. Neurology, Vol. I, 162.
26 Paraphysis and the Pineal Kegion in Nectun;s Macnlatus
17. Hkkuick, C. L. — Embryological Notes on the Brain of a Snake. Journ.
Neurology, Vol. I, lGO-176.
IS. Hill, C. L. — The Epiphysis in Teleosts and Amia. Journ. of Morphology.
Vol. IX, 237-268.
19. His, W. — Zur allgemeinen Morphologie des Gehirns. His. Archiv, 1892,
20. Humphrey, O. D. — On the Brain of the Snapping Turtle. Journ. of Comp.
Neurology, Vol. IV, 73-108.
21. Kingsbury, B. F. — The Brain of Necturus Maculatus. Journ. of Comp.
Neurology, Vol. V.
22. The Encephalic Evaginations in Ganoids. Journ. of Comp. Neurology, Vol. VII.
23. KxiPFFER, C. V. — Studien zur vergleichenden Entwicklungsgeschichte des
Kopfes der Kranioten. Hefte I.
24. Derselbe. Hefte II.
25. Lewis, F. T. — The Question of Sinusoids. Anat. Anx., Bd. XXV, No. 11.
26. Leydig, F. — Zirbel und Jacobsonsche Organe einiger Reptilien. Archiv f.
Mikrosk. Anatomie, Bd. 50.
27. LocY, W. A. — Contribution to the Structure of the Vertebrate Head.
Journ. of Morphology, XI.
28. MiNOT, C. S. — On the Morphology of the Pineal Region, based on its
Development In Acanthias. American Journ. of Anatomy, Vol. I,
No. 1, 81-98.
29. On a Hitherto Unrecognized Form of Blood Circulation without
Capillaries in Organs of Vertebrates. Pro. Boston Soc. Nat. Hist.,
Vol. 29, No. 10, S. 185-215.
30. OsBORN, H. F. — Preliminary Observations on the Brain of Menopoma.
Proceed. Phil. Acad., 1884.
31. Contribution to the Internal Structure of the Amphibian Brain.
Journ. of Morphology, Vol. II, 51-86.
32. Rex, H. — Beitrage zur Morphologie der Hirnvenen der Amphibien. Morph.
Jahrb., XIX, 295-311.
33. ScHOBEL, Jos. — Ueber die Blutgefasse des Cerebrospinalen Nervensystems
der Urodelen. Archiv f. Wissen. Mikros., Bd. XX, 87-91.
34. Selenka, E. — Das Stirnorgan der Wirbelthiere. Biolog. Centralbl., Bd. X,
35. Sorensen, a. D. — The Roof of the Diencephalon. Journ. of Comp. Neu
rology, III, 50-53.
36. Comparative Study of the Epiphysis and the Roof of the Diencephalon. Journal Comp. Neurology, IV, 153-170.
37. Continuation of above. Vol. IV, 153-170.
38. Studenicka, F. Cii. — Beitrage zur Anatomie und Entwicklungsgeschichte
des Vorderhirns der Cranioten.
39. VoELZKOw, A. — Epiphysis und Paraphysis bei Krokodilien und Schild
krot>en. Abhand. der SenchenbiXrgischen Naturforschenden Gesellschaft, Bd. XXVII, Heft. II.
John Warren
A. C. — Anterior commissure. L. Y.
Ch. Fix. — Choroid plexus. M. B
Dien. — Diencephalon. 0. C
D. Flex. — Diencephalic plexus. Tel. Plx.
i?.— Epiphysis. Tel.
Ep. A. — Epiphysal arch. F. C
F. B. — Fore-brain. P. T- A. —
F. M. — Foramen of Munro. ■ P. A.
H. — Hemisphere. F.
H-JB.— Hind brain. S.C.
Hyp. — Hypophysis. Si.
I. J. V. — Internal jugular vein. T.
L. Fix. — Choroid plexus of lateral Tes.
ventricle. ^'
Lateral ventricle.
Optic commissure.
-Telencephalic plexus.
-Posterior commissure.
Post-velar arch.
Paraphysal arch.
-Superior commissure.
-Velum transversum.
E. T. BELL, B. S., M. D.
Instructor in Anatomy, University of Missouri.
With 3 Plates and 5 Text Figures.
This paper is intended mainly as a contribution to our knowledge of
the histogenesis of the thymus in mammals. Special attention is given
to the origin and development of the corpuscles of Hassall, since their
mode of formation has never been satisfactorily described in mammals
and their significance in all forms is in dispute. An attempt is also
made to show in detail the changes that occur during the transformation
of the thymus from the epithelial to the lymphoid condition.
This work was begun at the suggestion of Dr. D. D. Lewis at the
University of Chicago. The greater part of it has been done at the
University of Missouri. Special acknowledgments are due Dr. C. M.
Jackson for valuable criticism and suggestions. T wish also to thank
Mr. Charles H. Miller of the University of Chicago for his kindness in
sending me material.
Material and Methods.
As material for the greater part of my work, I. have used pig embryos
from 8 mm. to full term (26 cm. to 30 cm.). These are especially suitable for such work since they may be procured in abundance from the
large packing houses at almost any stage of development. For special
purposes I have studied a few specimens from human foetuses, and from
the cat, rat, and guinea pig. The smaller pigs used (8 mm. to 27 mm.')
belong to the collection in the anatomical laboratory at the University of
Missouri. These were stained in bulk with alum-cochineal and mounted
in serial sections. In the later stages, which were prepared specially for
this work, tbe ventral half of the cervical and anterior thoracic regions
was usually out out and embedded from pigs from 3 cm. to 8 cm. On
specimens from 8 cm. to 30 cm.. I dissected out the thymus and used sucb
parts as were desired.
^ The crown-rump measurement is used in all cases.
30 Tlie Development of the Thymus
All pig material was fixed in Zenker's fluid, embedded in paraffin, and
mounted in serial sections from 3 ju, to 10 /x thick. Except those of the
young stages (8 mm. to 27 mm.) the sections were stained on the slide.
Most of them were stained with hsematoxylin or iron-hgematoxylin and
counterstained with Congo red. For special purposes many other stains
were used.
To demonstrate the delicate protoplasmic threads of the syncytium
during the later stages of the lymphoid transformation, I stained by the
iron-haematoxylin method but omitted the final decolorization in ironalum. Protoplasm is stained deep black; nuclear structure is poorly
shown, but the finest cytoplasmic processes may be seen.
For the demonstration of connective tissue fibrillae in the syncytium,
I found the method recommended by Jackson (13, S. 39) most satisfactory.
Text Fig. 2.
Text Figure 1. Cranial view of third gill pouch (thymic anlage) ; X 33;
pig embryo, 11 mm.; ec, ectoderm; I, lumen; nt, nodulus thymicus; ph, connection to pharynx; s p, sinus prsecervicalis.
Text Figure 2. Ventral view of thymic anlage; X 33; pig embryo, 15 mm.;
ec, ectoderm; I, lumen; n t, nodulus thymicus; ph, connection to pharynx;
s p, sinus prsecervicalis.
To determine the relation of the blood-vessels to the corpuscles of
Hassall, I put a young kitten under deep anesthesia and injected a large
quantity of a strong aqueous solution of Prussian blue into the aorta
through the common carotid artery. The heart continues to beat even
after an amount of fluid twice as great as the total volume of the blood
has been injected. An injection made in this way is under a slightly
increased blood pressure and easily reaches the finest capillaries. There
is therefore a thorough injection with little danger of rupturing delicate
My observations on this phase of development are not sufficiently complete to warrant a full discussion. A brief survey may however prepare
the way for a better understanding of the histogenesis. The text figures
show the shape in outline of the third gill pouch and thymic anlage
E. T. Bell 31
from 11 mm. to 27 mm. . They are graphic reconstructions. Since this
pouch is nearly all converted into thymus it may be regarded as the
thymic anlage from a very early stage.
At 11 mm. (Text Figure 1), the pouch is a hollow epithelial tube directed from without ventrally and mesially. The lumen (l) is large
and communicates freely with the pharynx. On the dorso-lateral aspect
of the pouch is a solid epithelial mass (n t) distinctly different in structure from the rest of the pouch. This is the nodulus thymicus
(Kastschenko, 14) and will be referred to by that term. This structure
has been described by Stieda (26), Prenant ' (22), and others as the
anlage of the carotid gland.
It was evidently mistaken by Minot ' in a 12-mm. pig for the anlage of
the entire thymus. It may be seen as early as the 8 mm. stage budding
off from the cranio-lateral aspect of the pouch. Immediately behind the
nodulus thymicus, but not connected to it at this stage is the inner blind
extremity of the sinus prsecervicalis (s p). These become fused at
12 mm. or 13 mm.
At 15 mm. (Text Figure 2) the thymic anlage is more elongated. It
now projects ventrally and medianwards, its free end lying immediately
caudad to the median thyroid anlage and just craniad to the pericardium.
Its lumen (Z) is still in communication with the pharynx. The sinus
prsecervicalis {s p) is drawn out, its lumen being smaller and longer.
It is now fused to the outer two-thirds of the posterior aspect of the
nodulus thymicus (n t).
At 18 mm. (Text Figure 3) the anlage is growing rapidly in a caudal
direction and just entering the thoracic cavity. It is connected to the
pharynx by a delicate epithelial cord. There is still a lumen in its caudal
part. The sinus praecervicalis has lost its connection to the nodulus
thymicus." The outer part of its lumen has disappeared and it seems
about to lose its connection with the ectoderm.
At 20 mm. (Text Figure 4) the thymus extends well into the thoracic
cavity. Its thoracic segment (t h) has increased considerably in size and
is united to the gland of the opposite side.
The nodulus thymicus still forms the greater part of the head. The
anlage has no connection with the pharynx or the epidermis. There is
"Prenant is said to have since abandoned this idea and accepted Kastschenko's view. (v. Ebner in Kolliker's Gewebelehre des Menschen. Aufl.
6, Bd. 3, 1, S. 340.)
^Laboratory Text of Embryology, p. 191; also p. 209 and Fig. 124.
* On the opposite side in this specimen, .these structures were fused over a
very small area.
The Development of the Thymus
now an elongated mass of thymic tissue extending upwards behind the
nodiilus thymicus, and fused with it. Its upper pointed extremity curves
outwards around the hypoglossal nerve. This is the " thymus superficialis " (t s) of Kastschenko and is regarded by him as being formed
from the sinus praecervicalis. Kastschenko describes this elongated
portion as being always separate from the rest of the head, being connected only by connective tissue. In my preparations it is clearly continuous with the rest of the anlage, and seems to have formed by growing
out from it. The presence of a lumen in its lower end favors Kast
Text Fig. 3.
Text Fig. 4.
Text Figure 3. Ventral view of thymic anlage; X 33; pig embryo, 18 mm.;
ec, ectoderm; I, lumen; nt, nodulus thymicus; ph, connection to pharynx;
s p, sinus praecervicalis.
Text Figure 4. Ventral view of thymic anlage; X 33; pig embryo, 20 mm.;
f, area fused with gland of opposite side; I, lumen; nt, nodulus thymicus;
th, thoracic segment; t s, thymus superficialis.
schenko's view, for there is no lumen in the head at 18 mm. in my preparations. On the other hand the separation from the head at the 18 mm.
stage favors the idea that the sinus prsecervicalis degenerates. I have not
studied a sufficient number of specimens at this transition stage to
enable me to decide this point, though I believe the ectoderm takes no
part in the formation of the thymic anlage. Kastschenko's results are
opposed by nearly all other students of this problem, but his work should
not be discarded before the development of the thymus superficialis in
the pig has been accurately determined.
At 27 mm. (Text Figure 5) the thymus is much longer and extends
E. T. Bell
well down into the thoracic cavity in relation to the base of the heart
where it has fused with the gland of the opposite side. Buds are now
beginning to form through the greater part of its extent. Traces of the
original lumen (I) may still be seen in several places. The "thymus
superficialis " is bent around the twelfth nerve. A delicate cord of thymic
Text Fig. 5.
Text Figxjre 5. Ventral view of thymic anlage; X 33; pig embryo, 27 mm.;
/. area fused with gland of opposite side; 1, lumen; nt, nodulus thymicus;
th, thoracic segment; t s, thymus superficialis.
tissue fused with the l)ack of the nodulus thymicus connects the thymus
superficialis to the rest of the head.
From about 3 cm. until toward the end of fcetal life the thymus shows
the two constrictions, described by Prenant (22) (for the sheep) as the
intermediary and the cervico-thoracic cords. These cords connect three
3-i Tlic Development of the Tliynius
enlargements which we may call the head, the luid-t-ervical segment, and
the thoracic segment. I will consider each part separately. The thoracic
segment develops rapidly, spreading out ahove and in front of the heart.
The glands of the two sides fuse completely in this region. The lymphoid
transformation is noticeable at 3.5 cm. and well advanced at 4.5 cm. The
medulla begins to form at 8 cm. The cervico-thoracic cord is at first
very narrow but soon thickens and joins the cord of the opposite side.
At full term they form a sharp constriction, 3 mm. to 4 mm. wide and
5 mm. to 6 mm. long, situated at the superior aperture of the thorax,
and connecting the mid-cervical and thoracic segments. The histological
changes take place here later than in the enlargements.
The mid-cervical segment develops like the thoracic segment but somewhat more slowly. Budding, lymphoid transformation, and formation
of the medulla all begin here a little later than in the head and thoracic
segment. Its caudal end is slightly in advance of its cranial end. The
intermediary cord is well marked at 4 cm. It soon becomes very attenuated, having at 6 cm. in many places a diameter of only 15 fx. Prenant
suggests that this drawing out of the gland is caused by the rapid growth
of the neck. Later it increases in size and at full term is noticeable only
as a very slight constriction between the head and the mid-cervical
segment. The histological changes are much later here than in other
parts of the gland.
The greater part of the head in the early stages is formed by the
nodulus thymicus. This body grows slowly, attaining a diameter in crosssection of .5 mm. at 8 cm. Its cross-sectional area at 8 cm. is about onethird that of the rest of the head of the thymus. At this stage the nodulus
thymicus is a rounded body lying on the inner aspect of the head in relation to the carotid artery. A small area of its outer surface is fused
with the lymphoid tissue of the thymus. Its histological structure has
been fully described by Prenant (22). From the earliest stages, it
consists of cords of epithelial cells separated by blood capillaries. At
8 cm. the thymus superficialis (Kastschenko) is a large body lying craniad
and dorsal to the rest of the head but connected to it at its caudal
I have no observations on the head of the thymus between 8 cm. and
full term, having overlooked it in collecting my material. In a full term
pig (30 cm.), the cervical part of the thymus is in two distinct parts.
The postero-ventral part, representing part of the head, the intermediary
cord, and the mid-cervical segment, is about 3 cm. long and 1 cm. wide.
It extends from the upper border of the thyroid cartilage to the thorax,
and with the corresponding part of the opposite gland encloses the
E. T. Bell 35
trachea, thyroid, and lower ]iait of the larynx. A slight narrowing at
the junction of the upper and middle thirds indicates the position of the
intermediary cord. The antero-dorsal part^ the thymus superficialis, is
rounded in cross-section, tapei-rrfg- to a point behind. Its anterior end is
about 7 mm. in diameter and loops around the twelfth nerve as in the
earliest stages. A lobule hangs over ventral to the nerve, a thin cord
being dorsal to it. The posterior end of the thymus superficialis extends
to the cricoid cartilage dorsal to the postero-ventral part of the gland.
It is united to this part of the gland by a very delicate band of thymic
tissue. I did not find the nodulus thymicus at full term. It has either
moved away from the head or degenerated.
The duct of the thymus is the lumen of the third gill pouch. A
glance at the Text Figures will show its development. It is broken up
into segments and finally obliterated. I could find no traces of it at
3.7 cm. or later. On the theory of the exclusively endodermal origin of
the thymus. I cannot explain the absence of a lumen in the head at
IS mm. and its presence at 20 mm. and 27 mm. unless it be due to
individual variation in different specimens.
It appears from the foregoing that in the pig the exclusively endodermal origin of the thymus from the third gill [loueh is probable, but
a slight participation by the ectoderm has not been satisfactorily excluded.
Kastschenko's conclusions, however, as to the ectodermal origin of the
thymus are imwarranted by his recorded observations. Prenant, after
his careful work (on the sheep), was not sure that a small mass of ectoderm did not enter into the formation of the head. Practically all other
investigators of this problem maintain that tlie ectoderm takes no part
in the formation of the thymus. The epithelial l)ody (nodulus thymicus)
developing in connection witli tlie head of the thymus from the third
gill pouch does not form the carotid gland. Kastschenko's description of
the origin of the carotid gland in mammals from the adventitia of the
internal carotid is now accepted by the majority of anatomists, and it
therefore has nothing to do with the thymus in oi'igin.
TiiK Hist<)L()(;y of tiik FrLLY-F()i;.Mi:i) Thymus.
Before taking u\) tlic histogenesis, I shall bi'iclly (.-(Uisider the histology
of the gland as shown in a 24-cni. embryo. At this stage ihe gland nuiy
be regarded as fully formed. As is well known, tlie thymic lobule consists of a cortical and a medidlary portion, — the medulla of all the
lobules being united by the medullary cord. The cortex consists of a
delicate reticulum with its spaces well filled by cells, usually lymphocytes.
The reticulum may be regarded as composed of small branched anastom
36 The Development of the Thymus
osing cells, though of eourse no cell houndaries are distinguishable. The
nuclei are poor in chromatin, rounded, and usually 4.5 /a to 5.5 /x in
diameter. The amount of cytoplasm around the nuclei and connecting
them is usually very small. At some nodes there is a greater amount of
cytoplasm giving the appearance of a large reticulum cell. In connection
with the lilood-vessels, which are numerous in the cortex, are often found
branched cells with pale nuclei and cytoplasm that stains more intensely
than that of the rest of the reticulum. By a modification of Mallory's
method used by Jackson (13, S. 39), I have been able to demonstrate
numerous fibrillEe in the reticulum. In some cortical areas at this
stage there are a great many erythroblasts. Masses of free erythrocytes
are often found, usually near a comparatively large vessel, but such cells
occur singly anywhere in the cortex.
In the medulla, the syncytial character of the stroma is much more
pronounced. The cytoplasm is much more abundant than in the cortex,
and the spaces are smaller and not so numerous. There is not so much
room for lymphocytes as in the cortex, hence the lighter color of the
medulla in stained preparations. The nuclei of the syncytium are either
pale or dark, both types showing wide variations in size. By Jackson's
method (13, S. 39), fibrilla3 may be readily demonstrated in the syncytium. In Plate I, Fig. 6, is shown the arrangement of these fibrillse
(s f) in the medulla at 24 cm. They often may be traced into the
areas which are forming the concentric corpuscles. In some parts of the
medulla the fibrillae are very numerous ; in a few places, entirely absent.
In both cortex and medulla eosinophile cells are often found. These
occur in groups in the interlobular tissue around the blood-vessels, around
some of the corpuscles of Hassall, and singly in the reticulum. These
have been described by Schaffer (24). Free erythrocytes a.re rarely found
in the medulla. In the medulla are also found the corpuscles of Hassall.
Since the structure of these bodies depends largely upon their age, it may
be better understood from the consideration of their development.
The Lymphoid Transformation.'
Kolliker, in 79, first advanced the idea that the leucocytes are formed
directly from the epithelial cells of the thymic anlage. According to
Minot (in human embryology, p. 878), "he records for the rabbit that
between the twentieth and twenty-third days the cells of the thymus become smaller and their outlines disappear, so that the organ appears to
° This term will be used to include those changes that occur in the thymus
during its passage from an epithelial to a characteristic lymphoid structure.
,E. T. Bell 37
bo ail atiuiiiulaticni of small round nuclei. At about the same period
blood-vessels and eonneetive tissue grow into the epithelial anlage."
His {I'i 1)). 80, and Stieda (26), 81, claimed that the corpuscles of
Hassall are the only remnants of the epithelial anlage, that the lymphocytes, reticulum, etc., are of mesenchymal origin.
Maurer (17 a), 86, described the leucocytes as arising directly from the
cells of the epithelial anlage in the thymus of teleosts. In the amphibian
thymus (17 b), 88, he thinks that the leucocytes are probably of mesenchymal origin. He was unwilling to conclude that they arose from the
epithelium because he could not find transition forms. In lizards (17 c),
99, he records that even before the separation of the epithelial anlage of
the thymus from the pharynx, changes begin. The peripheral cells are
closely crowded together and show many mitoses. There arises between
the central cells, or is formed in vacuoles in their protoplasm, a fluid which
accumulates until the nuclei surrounded by a thin zone of protoplasm are
connected only by protoplasmic threads. A loose medulla is thus formed
which looks like a cellular reticulum. The cortex is still solid. The
lymphocytes are formed from the epithelial cells; none come from without. Later, blood-vessels and connective tissue grow in. He believes
that the reticulum is of mesenchymal origin in all forms. Maurer
(17 d), 02, still holds that in amphibians the lymphocytes are probably
of mesenchymal origin.
Plermann et Tourneux (11), 87, find that in man and other mammals
the epithelial anlage of the thymus is gradually converted into leucocytes
and reticulum cells. Vacuoles appear during the transformation which
seem to be formed by the absorption of large cells. In a sheep embryo
of 130 mm., the clear epithelial cells have all disappeared, giving rise
to small round cells and reticulum cells. Prolongations of connective
tissue, each containing a blood-vessel, grow into the anlage during the
transformation. They are not sure that all the thymic elements are
epithelial in origin, being especially in doubt about the origin of some of
the reticulum cells.
Gulland (8), 91, describes the development of the tonsil in the rabbit.
Leucocytes first appear in the connective tissue around the thymus. Later
they appear in the connective tissue around the tonsil. They infiltrate
the tonsillar epithelium. Fo leucocytes are of epithelial origin. After
studying the tonsil he examined the thymus in a few specimens and
concluded that the same process of leucocyte infiltration occurred there.
He does not give the details of their infiltration, and did not see any of
the transition forms of nuclei in the thymus at that period.
Prenant (22), 94, made a careful study of the development of the
38 The Drvi'lopincnt ol' the 'riiynius
thynuis in sheep embrvos. His results are as follows. At 25 mm. the
gland is composed of distinct polyhedral cells with nuclei of only one
kind. A few mitoses and an occasional direct division are to be seen. At
20 mm. mitoses are numerous (oi^e nucleus in fifty). jSTuclei are regularly rounded or elliptical and some small nuclei occur juxtaposed to
large nuclei. At 28 mm. many mitoses are present and irregular spaces
have appeared. These spaces are not blood-vessels nor parts of the
thymic duet but vacuoles. Some nuclei, noticeably small and darkly
colored, lie close to the large, clear nuclei and seem to be budded off from
them. Some nuclei (rare) are broken into three or four chromatic
bodies. Embryos of 40 mm. have undergone in great part the lymphoid
transformation. All transitions are found between the large, pale elliptical nuclei of clear reticular structure and the small, deeply colored
rounded nuclei whose sap is strongly stained. These last are certainly
lymphocytes and constitute an immense majority of the cellular elements.
Large, clear nuclei are found Joined to small dark ones — nuclear couples.
At 85 mm. the medulla appears; the cortex corresponds to the entire
thymic mass of preceding stages. The cortex contains a great many
lymphocytes separated by islands and rows of pale nuclei. There are
about thirty lymphocytes to one pale nucleus. Mitoses are numerous in
the cortex. In the medulla at this stage, the large clear, and small dark,
nuclei are about equal in number, and mitoses are rarer than in the
cortex. In later embryonic stages a clear peripheral zone is present
where cell proliferation takes place. Mitoses are now more numerous in
the medulla than in the cortex. It is probable that a certain number of
the epithelial cells persist as reticulum cells in the fully-formed organ.
J. Beard (3 a), 94, (3 b), 99, thinks that the function of the thymus
is to form the first leucocytes. He finds that in the skate the epithelial
cells are converted early into lymphocytes which emigrate into the blood.
There are many breaks in the gland where the lymphocytes escape in
masses. The thymus is the only source of leucocytes until the other
lymphoid organs are formed.
Ver Eecke (28), 99, finds that in the frog the epithelial thymus is
invaded by lymphocytes and connective tissue. The epithelial cells are
not destroyed but merely dispersed by the mesenchymal elements. He
calls the resulting tissue lympho-epithelial. This idea of the commingling of the two tissues had already been advanced by Eettcrer.
Xusbaum and Prymak (20), 01, on teleosts, agree with j\Iaurer that
the lymphocytes are of epithelial origin but disagree on the details of their
formation. They find that the epithelial anlage is at first composed of
cells with distinct lioundaries. It is not different from the epithelium
E. T. Bell 39
of the pharynx. Before any blood-vessels or connective tissue have invaded the organ, changes begin in. the central part. These changes consist in the breaking up of the cytoplasm so that the cells become
branched and connected by delicate processes. These processes finally
break apart leaving a nucleus surrounded by a thin layer of protoplasm —
a lymphocyte. The peripheral epithelial layer multiplies rapidly, forming
nuclei somewhat smaller and darker than their own. These nuclei become gradually changed into the nuclei of lymphocytes and break away
from the other cells. All transitions are present between the large, clear
epithelial nuclei and the lymphocytes. Blood-vessels and connective
tissue grow in from the outside.
It appears from a survey of the literature that, of those who have
studied the origin of lymphocytes in the mamalian thymus. His, Stieda,
and Gulland have advocated the idea that they invade the gland from
without, and that the original epithelial anlage persists only as remnants,
the corpuscles of Hassall. They also consider the stroma of mesenchymal
origin. On the other hand, Kolliker, Hermann and Tourneux, and Prenant, have described the lymphocytes as derived directly from the epithelial
cells of the anlage. Hermann and Tourneux and Prenant ascribed a
similar origin to part of the reticulum.
Xeither His, Stieda, nor Gulland made a detailed histological study
of the changes that take place in the thymus during the transformation.
They did not see the vacuolization of the cytoplasm, the changes in the
epithelial nuclei, etc. — processes wdiich undoubtedly occur. Gulland
made nearly all his observations on the tonsil and then from a superficial
examination of the thymus concluded that the process is the same there.
The conclusions of these men are therefore not to be compared on this
point with those obtained by the thorough and careful work of Prenant.
On amphibians, Maurer hesitatingly agrees with His, and Yer Eecke
accepts the mesenchymal origin of the leucocytes ; while on fishes Maurer,
Beard, and Nusbaum and Prymak believe in the epithelial origin of
lymphocytes. Maurer's work on reptiles is in agreement with his work
on teleosts.
As to the origin of leucocytes in the lymphoid organs of tlie alimentary
canal, opinion is divided. Eetterer, v. Davidoff, Rudinger, Klaatsch,
and others have described the leucocytes as arising from epithelium and
being invaded by mesenchymal elements forming adenoid tissue. Stohr,
Gulland, Tomarkin, and others describe them as penetrating the epithelium from without.
I shall now discuss my own observations on the lymphoid transformation in the thymus of the pig. From a very early stage (11 mm.), the
40 The Development of the Thymus
t'pithelium of the third gill pouch is a syncytium. No cell boundaries
exist. The nuclei, large and irregular in shape, are embedded in a common mass of cytoplasm. In the thymus at 20 mm. I find a syncytium
of dense cytoplasm embedded in which are large nuclei of irregular shape
and size. No distinct types of nuclei are present yet; all stain with
medium intensity. A few mitoses are to be seen.
In a section of the mid-cervical segment at 3.7 cm. (Plate I, Fig. 1),
I find evidence that the lymphoid transformation has begun. The syncytium is composed of coarsely reticulated cytoplasm much looser in
texture than that of the preceding stage. It contains a few irregular
spaces (s s) which are evidently of the nature of vacuoles. These may
be formed, as Maurer suggests, by liquefaction of the cytoplasm. There
is no reason to suppose that cells degenerate and form them as Hermann
and Tourneux believed. Three types of nuclei may be distinguished;
large pale nuclei (Ipn) large dark nuclei (I dn), and small dark nuclei
(lymphoblasts) (Ih). Transition forms occur between these types.
The large dark nuclei are intermediate forms between the pale nuclei
and the lymphoblasts. A few mitoses (m) occur. No blood-vessels are
present inside the anlage but they may be seen between the buds just
outside. At this stage, the head and the thoracic segment have areas
that are somewhat farther advanced than this. The intermediary and
cervico-thoracic cords show no changes.
At a later stage than the above (Plate I, Fig. 2), in the thoracic
segment of a 4.5-cm. pig, jthe spaces of the syncytium (s s) have increased greatly in number and size. The anlage is now a cellular reticulum. The large pale nuclei are somewhat less numerous than the
dark nuclei and many have become angular, adapting themselves to the
nodes of the syncytium. They contain less chromatin than in the
preceding stage. Large dark nuclei and lymphoblasts are present; the
lymphoblasts are much more numerous than in the preceding stage. A
very few small dark nuclei are completely separated from the syncytium.
These are lymphocytes. There are no lymphocytes in the connective
tissue around the thymus or in the blood at this stage. I did not
examine the tonsil or spleen at any stage. A few small blood-vessels are
to be seen; their walls consist of endothelium only. There are more
mitoses than at the preceding stage, but none happened to be present in
the area shown in the figure. During mitosis, at all stages of development, except the early epithelial condition, the chromosomes are so
closely packed that it is very difficult to distinguish them individually.
In a section through the mid-cervical segment of a 7-cm. pig
(Plate I, Fig. 5), we see a stage somewhat later than the one shown in
E. T. Bell 41
Fig. 2. In various jiarts of the section lymphocytes (I) are completely
formed. Tlie great majority of the small round nuclei are in the lymphoblast (l h) condition, i. e.,thcy are not yet completely separated from the
syncytium. There are a few lymphocytes outside the thymus in the
interlobular tissue in this region; around the head and the thoracic
segment at 7 cm. they are numerous, these parts being in a later stage
of transformation. I have never seen lymphocytes outside the thyfnus,
where there were none inside it ; but they appear outside shortly after they
are formed here. Those formed next the interlobular septa seem to pass
out very early. Of course the lymphoblasts, which are distinguishable
from the lymphocytes only by being imbedded in the syncytium, are to
be seen in the thymus long before any appear outside.
At the stage shown in Fig. 5, a great many nuclei are in mitosis. I
have not seen at any stage, the amitoses and nuclear couples described
by Prenant for the sheep. In some parts of the section comparatively
large solid epithelial areas occur. These are found as often in the
central as in the peripheral part. Many of the pale nuclei are smaller
than those shown in Fig. 2. The blood-vessels are somewhat larger and
more numerous than those at 4.5 cm.
It is to be noted that the epithelial anlage does not at any stage become
converted entirely into small round cells as many observers have stated.
Distinctly pale angular reticular nuclei can always be seen.
In the mid-cervical segment at 8.5 cm. (Plate I, Fig. 4), a great many
lymphocytes (/) are formed. These lie between the persisting epithelial
cells which are now arranged in irregular cords and islands. In these
epithelial masses, lymphoblasts may still be seen indicating that the
formation of lymphocytes is still in progTCss. Many of the pale nuclei
are now small. The heavy hsematoxylin stain in this case makes the
nuclei darker than they would appear with an ordinary stain. A few
nuclei are in mitosis.
This figure shows also the first appearance of the medulla {ind). The
medulla is formed directly, as shown in the figure, from persisting parts
of the epithelial syncytium. Certain centrally situated masses of this
syncytium undergo changes of such a nature that they stain readily with
cytoplasmic stains such as Congo red. In sections stained with hematoxylin and Congo red, the medulla is first recognized as a l)rightly
colored area situated usually about the center of the lobule. Tlicse
epithelial masses that give rise to the medulla seem to increase in size
about the time of the change in staining capacity. The first differentiation of the medulla is chemical rather than morphological, for there are
other persisting epithelial masses even larger than it in the same section
42 The Development of the Thymus
that do not react in the same way with the cytophismic stains. The
meduna appears in the head and the thoracic segment at 7.5 cm. to 8 cm.
All the gland except this small central area forms the cortex. Bloodvessels now reach 'all parts of the gland, but are still few in number. I
cannot distinguish any wall except the endothelium on those actually
inside the gland.
In a 9.5-cm. pig, the medulla is larger. It contains pale nuclei of
various sizes, large dark nuclei, and lymphoblasts. Its spaces are smaller
than those of the cortex. The early stages in the formation of the corpuscles of Hassall appear as soon as the medulla begins to form. The
epithelial cords in the cortex have become less conspicuous, but are still
forming lymphocytes. A few nuclei are in mitosis. Many blood capillaries may now be seen penetrating the gland from the periphery.
These vessels run in the epithelial masses and have a wall of large
endothelial cells which gives them the appearance of radiating cords.
When these vessels first appear, as at this stage, they have only an
endothelial wall. The blood-vessels grow in as small capillaries which,
after their entrance, increase in size and branch ; they do not break in as
large vessels surrounded by mesenchymal tissue. I am fairly sure that
aside from the endothelial cells few or no mesenchymal cells come into
the thymus. Around the greater part of the periphery of the gland is
a solid zone of syncytium two or three nuclei deep which is in transformation like the epithelial cords inside. This zone, described by Prenant as a zone of proliferation, grows rapidly, as the frequent mitoses
indicate. Its inner boundary is forming lymphocytes and reticulum cells.
In a 14-cm. pig, the lymphoid transformation is practically at an end
except in the medulla. The peripheral zone of proliferation has disappeared. The cortex has about the same structure as at 24 cm., as
previously described. In the medulla, lymphoblasts, large pale nuclei,
and the large dark intermediate types are still present. There are a few
mitoses here. It is very probable therefore that the formation of lymphocytes is still in progress in the medulla. The medulla at 2-i cm.
shows a similar structure except that there are fewer lymphoblasts. These
facts persuade me to regard the medulla as a center for lymphocyte
formation at least as late as birth. Connective tissue fibrillae begin to
appear in the gland along the large blood-vessels and the interlolmlar
septa as early as 10.5 cm. They are only a little farther in at 16 cm. ;
but near full term they are present in nearly all parts of the stroma.
(See Plate I, Fig. 6.)
The above account may be summarized as follows: In the pig the
epithelial syncytium of the thymic anlage becomes loosened up by the
E. T. Bell 43
formation of vacuoles in it. These vacuoles increase in number and
size, converting the anlage into a cellular reticulum. While this vacuolization is in progress, the nuclei, which at first are of one kind with a
medium amount of chromatin, ditferentiate into large clear, large dark,
and small dark (lymphoblast) forms. The large dark nuclei probably
divide by mitosis and form the lymphoblasts. The lymphoblasts gradually break loose from the syncytium, passing into its spaces and becoming
lymphocytes. Shortly after lymphocytes begin to be formed, some of
them pass out of the gland into the surrounding connective tissue. The
lymphoid transformation begins in embryos of 2.5 cm. to 3 cm. and continues in the cortex until 13 cm. or 13 cm. In the medulla it is not complete at birth. Since the thymus increases greatly in size during this
period the epithelial syncytium must grow rapidly. Lymphocytes are
constantly being formed at the expense of the growing syncytium. A
peripheral zone of proliferation is present from about 8 cm. to 13 cm.
The medulla is formed as a chemical differentiation of certain centrally
situated areas of the epithelial syncytium. The histological changes
occur earlier in the head and thoracic segment than in the mid-cervical
segment and very much earlier than in the cords. The reticulum of both
cortex and medulla is practically all of e])ithelial origin. Some branched
cells around the blood-vessels in the cortex may be of mesenchymal
My reasons for regarding tlie lymphocytes as of epithelial origin are
as follows :
A. The lympholilasts are true epithelial nuclei, because (1) there are
numerous transition forms between them and the large dark nuclei which
later cannot l)e regarded as invading lymphocytes; (3) they are closely
embedded in the syncytium and show no evidence of having eaten their
way through the protoplasm; (3) they are present from a very early
stage and increase in number as development proceeds; (4) they are
present before blood-vessels invade the gland and have no constant relation to blood-vessels or to the surface of the gland that indicates an invasion from either of these directions; (5) they are present before lymphocytes appear in the connective-tissue around the thymus.
B. Some observers admit tliat the small dark nuclei (lymphoblasts)
are of epithelial origin but do not admit that tlicy form lym])hocytes.
The considerations that lead me to bciicNc tliat the lymphoblasts do form
the lymphocytes are: (1) the small dark nuclei (lymphoblasts) show
every possible relation to the syncytium from being completely embedded
in it to lying free in the syncytial s]")aces. A comparison with later
stages shows tb.at tliis appearance is not due to poor fixation oi- to the
44 'V\\v ncvclopnioiit of tlio M'hymns
adherence of the nuclei to the reticuhini ; (?) the first free niiclci often
appear in the center of the gland whoi tlicrc are no other free nuclei in
the periphery at that level; (3) there is good evidence that lymphocytes
emigrate from the thymns in large numbers. If we examine the thymus
of a 7-cm. pig in serial sections we find that the lymphoid transformation
is less advanced in the mid-cervical segment than in the head. In the
mid-cervical segment there are a few lymphocytes in the interlobular
tissue. In the lower end of the head where there are more lymphocytes
inside the gland, lymphocytes pack the interlobular tissue and form a
thin zone around the periphery of the gland. In the middle of the head
where the transformation is far advanced, lymphocytes pack the interlobular tissue and form a thick zone around the entire gland. Indeed,
in some sections, there are more l3T3iphocytes in the zone outside than
are present inside the gland. If this zone of lymphocytes be passing into
the gland, it is not easy to understand why it is formed from Avithin outwards, and why it is thickest where the greatest number of lymphoc3ies
are already present inside. No satisfactory suggestion has yet been made
as to why lymphocytes should thus suddenly pour into the thymus at a
time when if present at all elsewhere they are rare. They do not come
to break up the thymic epithelium, for that is already a reticulum before
free cells are present (Fig. 3, Plate I). Where lymphocytes invade
intestinal epithelium as in the tonsil they eat paths through it leaving
spaces. The epithelial reticulum of the thymus is not formed in that
way. On the other hand it is not difficult to believe that this zone of
lymphocytes is formed l^y cells passing out the periphery of the thymus
and that the gland thus contributes a great number of lymphocytes to the
organism; (4) I have not been able to find lymphocytes in the connectivetissue around the th}Tuus before they are present inside. An invasion by
way of the blood-vessels may be excluded, since the thick zone of lymphocytes formed around the gland shows that these cells either enter or leave
it through the preiphery.
The Corpuscles of Hassall.
These bodies were first mentioned by Hassall ( 10 ) in 46, He speaks
of them as being composed of mother cells which enclose the newlyformed daughter cells and nuclei. He thought the central mass was
formed by the outer enclosing layers. He found bodies which lie regarded
of the same nature in fibrous coagulations in the heart.
Virchow (29), 51, in a discussion of endogenous cell formation, compares Hassall's corpuscles to carcinoma pearls. He had about the same
E. T. Bell 45
conception of the nature of the corpuscles as Hassall. This oft-quoted
comparison was therefore not based upon a deep insight into their nature.
Giinzburg (9), 57, did not advance beyond Hassall's conception that
the central mass is formed by the peripheral layers.
Paulitzky (21), 63, described the center of the corpuscles as homogeneous or granular. They sometimes contain an elliptical nucleus, sometimes fat droplets. The larger ones have in the center several nuclei' or
cell-like forms. The central part is formed from masses of epithelial
cells. Connective tissue cells grow around them and are transformed into
epithelial cells forming the peripheral part of the corpuscle.
The term "concentric corpuscles" was introduced by Ecker (6), who
described them as arising directly from gland cells by fatty metamorphosis. He distinguished (1) simple corpuscles, round vesicles with
thick concentric hulls, containing inside a fatty opalescent mass, and (2)
compound corpuscles, which consist of several vesicles with a common
hull. The peripheral layers of a corpuscle consist of flattened cells.
His (13 a, 12 b), 60, 80, described the corpuscles as consisting of an
outer striated shell, probably composed of nucleated cells, and containing lymphocyte-like cells inside. He supposed them to be the original
cells of the epithelial anlage which become entangled in the reticulum in
some way. Their rapid growth in their narrow confines causes the concentric form.
Cornil et Eanvier (5), 69, considered the corpuscles as arising from
the endothelium of blood-vessels and compared them to the spheres of their
" Sarcome angiolithique."
This suggestion of a vascular origin, made by Cornil et Eanvier, was
elaborated by Afanassiew (la), 77.
Afanassiew held that the corpuscles of Hassall arise from the endothelium of the smaller veins and capillaries. The endothelial cells increase in size, become cubical, and later fill the lumen of the vessel.
During the proliferation of the endothelium, the blood-vessels break up
into segments which are now nearly solid cords. These cords are at first
connected to each other and to blood-vessels, but they soon break apart.
The surest proof that the corpuscles are of vascular origin is that erythrocytes may be found inside them. Vascular injections, however, do
not go into a corpuscle except in a very early stage, since the lumen is
soon obliterated by the endothelial plugs. The corpuscles are formed
entirely by the endothelial cells. Afanassiew worked on embryos of man,
the rabbit, and the calf.
Stieda (26), 81, in sheep embryos, describes the epithelial mass of the
4G Tlie Dcv('lf)|)tn(Mit of the Tli\nms
thymic aulage as being broken up by ingrowing adenoid tissue. From
50 mm. to 60 mm., there arc no hirge epithelial ceils; but later at 100 mm.
ho iinds in the adenoid tissue large cells 9 /a to 15 /x in diameter, isolated
or united in groups, Avhose protoplasm colors light-red with carmine.
These large cells have a concentric structure. Some of them are enclosed
bv large cells whose cytoplasm does not color with carmine, giving rise to
a yellowish mass of irregular form and stratified appearance. In older
embryos (250 mm.), the cellular masses are numerous but the large
colored cells are rare. The yellowish masses are groups of the large cells
which have undergone a transformation like that of the stratum corneum
of the epidermis. Stieda considers the large colored cells which form
the corpuscles as remnants of the epithelial anlage, although he admits
that for a long period during development he found no trace of them.
He explains the formation of the corpuscles in accordance with Cohnheim's hypothesis that most tumors arise from unused tissue remnants.
Ammann (2) 82, made most of his observations on human foetuses.
He describes the corpuscles as arising from connective tissue. The corpuscles are cellular in structure and are formed of one, two, or three central cells around which a variable number of cells, increasing with age,
are arranged like the coats of an onion. The corpuscles are formed from
reticulum cells and leucocytes. Growth consists in the apposition of cells
from without. The life of a corpuscle consists usually of four stages :
(1) Stadium der Transparenz ; (3) Stadium der colloidcn Entartung;
(3) Stadium der Verkalkung; (4) Stadium des Zerfalls. The nucleus
of a reticulum cell or leucocyte increases in size at the expense of the
cell body. Its increase in size establishes the concentric form. The
corpuscle undergoes colloid and usually calcareous degeneration. Fat
droplets, cholesterin crystals, and colloid granules are found together in
the degenerating corpuscles. Breaking up in this way makes absorption
possible. No epithelial remnants are to be observed. Xo erythrocytes
are found in the corpuscles.
In four cases of atrophic thymus gland wliich yet contained lymphoid
tissue Ammann found corpuscles in all stages of development. He also
found that the corpuscles are formed most rapidly when the thymus is
at the height of its development. From these facts he concluded that
they are not connected with the involution of the thymus as Afanassiew
thought. He thought that their formation is due to a physiological decrease in the intensity of growth of the medulla, due to the rapid growth
of the cortex.
Watney (31), 83, agreed with Ammann that the corpuscles arise from
connective tissue cells.
E. T. Bell 47
Monguidi (18), 85, distinguished true and false concentric corpuscles — •
the latter being onl}- sections of blood-vessels.
Hermann et Tourneux (11), 87, gave a description of the structure
and formation of the concentric corpuscles about like that given by
Ammann except that they regard the reticulum cells from which the
corpuscles develop as of epithelial origin.
Gulland (8), 91, regarded the corpuscles as epithelial remnants and
compared them to the epithelial pearls of the tonsil.
Maurer (17 c), 99, described the corpuscles as epithelial in origin.
His description of their formation is however different from that of His.
All the cells of the epithelial anlage at first assume a lymphoid character.
Later, some of these cells reassume their epithelial nature and then form
the corpuscles. His conclusions for teleosts and amphibians are similar
to the above results which he obtained from the lizard.
Yer Eecke (28), 99, for the frog, describes the leucocytes and connective tissue cells as invading the thymic anlage and separating the epithelial cells. The epithelial cells, separated by the mesenchymal elements,
lie at first in groups or singly. They go through a cycle of two phases,
a stage of growth, and a stage of involution. In the former stage, they
increase to three or four times their original size and their cytoplasm
differentiates into circular layers like the coats of an onion. The majority
are monocellular. Some cells grow together making a more complex
multicellular type. There are some intermediate forms, cells with a dense
dark protoplasmic body, indistinct striations, and a nucleus partly or
completely hidden in a precocious degeneration. In the stage of involution, which sets in early, the cytoplasm degenerates by the formation
of vacuoles containing a hyaline liquid. The liquefaction may be in the
form of a diffuse vacuolization, a large central vacuole, or a peripheral
vacuole circular in section. The nucleus loses its affinity for stains, becomes deformed, breaks up, and finally disappears. The corpuscles are
finally absorbed. They never contain erythrocytes. The cells do not degenerate to form a corpuscle. The liquefaction forms an internal secretion which is forced out by the muscle tissue in the reticulum.
Entirely different results on amphibians are reported by Nusbaum and
Machowski (19), 02. These investigators revive the old idea of Afanassiew, accepting his results except that they think the adventitia as well
as the endothelium of the blood-vessels takes part in the formation of the
corpuscles. They find erythrocytes in the corpuscles. These erythrocytes
either gradually shrivel and disappear, or they are absorbed by leucocytes
or endothelial cells. The leucocytes after digesting the hemoglobin of the
erythrocytes become eosinophile cells which are numerous in the thymus.
48 The Development of the Thymus
Wallisch (30), 03, measured the volume of the human thymus and of
the corpuscles of Hassali at various stages. He finds that the total volume
of the corpuscles of a 7-mo. embryo is 4.6 mm./ and of those of a 6-mo.
child, 174.6 mm." The total volume of the thymus of a 78-mm. embryo,
when it has already been partly transformed into adenoid tissue is only
6.8 mm.^ Since there is no evidence that the cells of the corpuscles
multiply, he concludes that they cannot be regarded merely as remnants
of the original epithelial anlage.
Disregarding the crude observations of the earliest investigators, there
remain three distinct theories of the formation of the corpuscles of
1. The epithelial anlage of the thymus is broken up by the invading
mesenchymal elements. The separated masses of epithelial cells undergo
further changes mainly of a degenerative nature to form the corpuscles.
This was the belief of His and KoUiker. According to this interpretation,
the corpuscles are to be regarded as remnants that have nothing further to
do with the gland. Stieda, Maurer, and Ver Eecke held this view in a
modified form. Stieda regarded the cells forming the corpuscles as
epithelial remnants but admitted that they go through a stage in which,
for a time, they lose their epithelial character. This is substantially the
same as ]\Iaurer's view. He thinks that the cells of the epithelial anlage
all become lymphoid, and that some of them afterwards reassume their
epithelial nature and form the corpuscles. Yer Eecke regards the corpuscles as epithelial remnants but thinks that they are glandular in nature,
not mere useless remains.
2. The corpuscles are formed from the proliferating walls of bloodvessels. This idea was suggested by Cornil and Eanvier and elaborated by
Afanassiew. Nusbaum and Machowski accept Afanassiew's view except
that they believe the adventitia of the blood-vessels as well as their endothelium takes part in the formation of a corpuscle. These investigators
thought that the formation of the corpuscles is connected with the involution of the thymus.
3. The corpuscles are formed from reticulum cells of the medulla
and grow by apposition of the surrounding cells. This view was advanced
by Ammann. Ammann thought that the reticulum is of connective tissue
origin. He also believed that leucocytes formed the central part at least
of some corpuscles. Hermann and Tourneux accepted Ammann's results,
except that they ascribed an epithelial origin to the reticulum. (I do not
know whether they accepted the origin from leucocytes described by
Ammann.) Ammann thought that the corpuscles formed because of a
physiological decrease in the rate of growth in the medulla.
E. T. Bell 49
l\Iy own observations on tlie development of tlie eorpuseles of Hassall
in pi_o- embrvos. will now be considered. The medulla, as previouslv described, begins to form from the epithelial syncvtium usually near the
center of the lobule. It is first recognized by its more marked reaction
with cytoplasmic stains such as Congo red. Shortly after the medulla
begins to form, the earliest stages of the corpuscles may be observed. - A
few corpuscles have appeared at 9.5 cm. I did not find them earlier.
They are all formed from the epithelial syncytium of the medulla.
Before beginning this discussion I will explain the use of ray terms.
By a corpuscle of Hassall, I mean a modified area of the epithelial syncytium of the medulla, containing at some period of its development, one
or more nuclei, and whose cytoplasm has been in part or entirely transformed into colloid. The term colloid is applied to various substances
probably of widely different chemical nature, but is fairly adapted to our
imperfect knowledge. I shall use the term here in the restricted sense
employed by Ziegler," i. e., hyaline substances of epithelial origin, that do
not give the reactions of mucin.
Colloid in the corpuscles of Hassall does not usually appear as solid
masses in its early formation, but as fibers, granules, or sheets which are
separated by more or less cytoplasm that is not yet changed. This stage
I have called, "colloid in formation" (c f). It later assumes a more
solid homogeneous appearance which I call solid colloid (c s). Often
the solid colloid stains intensely with cyto])lasmic stains. I call this
kind solid deeply-staining colloid (c .*? d). In later stages, the colloid
often loses its affinity for cytoplasmic stains, staining a very pale color or
not staining at all. I call this variety old colloid (o c).
According to their mode of development, the corpuscles of Hassall may
be classified as follows :
A. Concentric Corpuscles.
a. Simple.
1. Ordinary type.
2. Epithelioid type.
3. Cystic ty.pe.
b. Compound.
B. Irregular Corpuscles.
a. Compact type.
b. Eeticular type.
*Gen. Pathology, 10th ed., Warthin's translation, p. 205.
50 Tlie Development of the Thymus
A. The concentric corpuscles inchule those that from their earliest
appearance are concentric in structure. Adopting Ecker's classification,
I distinguis^i simple concentric corpuscles and compound concentric
(a) Three types of simple concentric corpuscles are to be considered.
(1) The ordinary type is far more numerous than any other. The earliest recognizable stage is shown in Plate II, Fig. 11. A nucleus (n)
of the syncytium of the medulla has enlarged to perhaps twice its ordinary
volume and has lost the ability to stain in the characteristic way with
ha?matoxylin. Its sap is clear and a few reddish stained granules represent its chromatin. x\round it in the cytoplasm is an indistinct uneven
layer of colloid (c /). The colloid is not yet solid and is being formed
in concentric fibers or sheets. A slightly later stage is shown in Plate II,
Fig. 14 and Fig. 15 (left side of figure). Some of the colloid (c s)
next to the nucleus is now solid. The next stage is shown in Plate II,
Fig. 15 (right side of figure). These corpuscles show a thick layer of
colloid (c s d) that stains intensely with Congo red. Just outside the
deeply staining colloid, colloid in formation may be seen. The nuclei
are clear, and have become smaller and irregular in outline. The colloid
seems to be pressing upon them and obliterating them. The colloid
transformation gradually involves the adjacent c3^toplasm of the syncytium until other nuclei are involved. The corpuscle has now reached
the condition shown in Plate II, Fig. 12. The central area (o c) is
solid, the nucleus having disappeared entirely. Another (n') is nearly obliterated by the colloid. Part of the central area {o c) no longer stains
intensely, and it is breaking loose by the formation of a concentric space.
Several nuclei are surrounded by colloid in formation. Their long axes
are nearly in a tangential direction. These nuclei are clear but only
moderately swollen.
In the further development of the corpuscle (Plate III, Fig. 17 and
Plate II, Fig. 7), the central area {c s d) increases in size. The nuclei
involved in this area become obliterated probably by the pressure of the
colloid and are no longer distinguishable. This central area usually splits
off and may break up into many smaller masses. The peripheral part of
the corpuscle increases by extension of the colloid formation into the
adjacent part of the syncytium. This extension takes place in the early
stages by direct progressive involvement of the immediately adjacent cytoplasm; in later stages (Fig. 7), by the formation of concentric lamellae
which cut ofl' unchanged areas of cytoplasm. The lamelhp increase in
size and number, the cytoplasm included between them is changed into
colloid. They finally become closely packed, giving the characteristic and
E. T. Bell 51
well-kncnvn onion-like structure found in the fully-formed eorpusele.
The nuclei that are enclosed between the lamells! gradually lose their
chromatin and become flattened out. They do not swell and are not
obliterated. It seems that swelling occurs only in nuclei that are surrounded by deeply staining colloid, and that this change is preparatory to
their obliteration by or transformation into colloid. The amount of the
corpuscle that breaks up to form the softer center is very variable. The
size of the center usually seems to increase with the age of the corpuscle.
Plate III, Fig. 21, shows a variation from the ordinary concentric
type. The central nucleus (n) stains reddish but is not enlarged. Most
of the other nuclei are unchanged. All the colloid (c /") is in the early
fibrous and granular stage.
From 20 cm. to full term nmny corpuscles show masses of calcareous
material in or near the center. This material rarely appears in younger
corpuscles (Plate III, Fig. 17, cl). It stains a violet blue with Delafield's ha?matoxylin.
The majority of the corpuscles of Hassall belong to the ordinary type
of simple concentric cor])uscles described above. It is very clear that
they have nothing to do with blood-vessels. They never contain erythrocytes nor anything resembling them. Earely a lymphoblast or leucocyte is found inside the corpuscle. These seem to be usually involved in
the corpuscle like ordinary stronui nuclei during the formation of the
lamellae. (Their occurrence in other types will be discussed later.) It
is also clear that these corpuscles arise from the syncytium of the
medulla. They are epithelial in origin, since the entire stroma of
the gland is derived from epithelium, but they are certainly not remnants
of the original epithelial anlage. Xeither are they forjned from lymphoidlike elements that reassume their epithelial nature as Maurer described
for the lizard.
Some of Ammann's observations are in accord with my results. The
swelling of the nucleus was noted by Ammann as tlu; first step toward the
formation of the corpuscle. It should he noted, however, that rarely a
corpuscle begins to form as a mass of colloid out in the cytoplasm and
involves nuclei secondarily. I cannot distiiiguish his " Stadium der
Transparenz " for I cannot be sure that a corpuscle is beginning to form
until some colloid is present. The formation of the colloid is associated
with the swelling of the nucleus. His other three stages, "Stadium der
coUoiden Entartung," " Stadium der Yerkalkung," and " Stadium des
Zerfalls " are easily seen. I have never seen corpuscles begin in leucocytes
as x\mmann described. His statement thnt the corpuscle grows by apposition of reticulum cells is true in a modified sense. He thought that
52 Tln' I )L'V('lo[)iiu'iit of []]{' 'I'liymus
the outer ])art of a {•or])nscle is formed of reticuhim cells that have moved
up and thittened themselves out around it. The description just given
shows ihat tlie coi'puscles are iieNcr composed of distinct cells, and that
the increase in size is due to an extension outward of tlie colloid foi'mation
and not to a moving in of the adjacent tissue.
The concentric form of this type of corpuscle is due at first to its l)eing
formed around a spherical or ellipsoidal nucleus. The swelling of this
nucleus creates a centrifugal pressure in the adjacent cytoplasm. Before
or during its transformation into colloid, the cytoplasm also increases in
quantity\ That the cytoplasm increases in quantity is shown hy the fact
that the nuclei are fewer in the corpuscle than in any adjacent area of
the syncytium of equal size. This centrifugal pressure presses the newly
formed colloid into concentric lamellae. It at first turns the long axes of
the nuclei tangentially. and later flattens them and makes them concave
toward the center.
3. The epithelioid type of corpuscle is characterized by large areas of
cytoplasm so marked off by colloid lamella? as to give the appearance of a
mass of large epithelial cells. They may contain only one nucleus embedded in a well-defined area of cytoplasm (Plate III, Figs. 18 and 20).
These correspond to the monocellular corpuscles that have been described
for lizards and amphibians. They are rare in the pig. I have not been
able to trace these very far, as they soon become indistinguishable from
other forms. The only difference I have noted is that the outer colloid
lamellae begin to form early, causing the peculiar appearance of a large
epithelial cell. Again the epithelioid type may present an appearance
such as shown in Plate I, Fig. 3. These do not seem to be formed around
any special nucleus. The outer colloid lamella? form before any center
has been established, niarking off large cytoplasmic areas that may look
like large cells. The centrifugal pressure of expansion caused by the
great increase of cytoplasm in this area determines the concentric form
in these corpuscles. Pure epithelioid corpuscles are very rare, but epithelioid areas in other corpuscles are not uncommon. The occurrence of
epithelioid areas in corpuscles of the ordinary type shows that it is due
to variations in a fundamentally similar process.
3. In the cystic type of corpuscle, the central part, instead of becoming
transformed into colloid, undergoes early liquefaction, forming vacuoles.
The central nucleus does not increase in size as in the ordinary type, but
shrivels up and disappears. The corpuscle begins by the formation of
outer colloid lamellae-^the central mass is not changed into colloid.
In Plate II, Fig. 10, the central area (/; m) is undergoing a diffuse
liquefaction. The nucleus (iv) is colorless and shrunken. In Plate TI.
E. T. Bell 53
Fig. 8 (right side of figure), the central area has fonned two lai-ge vacuoles (v). On the left side of the same figure, a concentric vacuole (v)
has formed, separating off a central spherical nucleated mass of protoplasm. The nucleus of this mass of protoplasm is shrunken and the
cytoplasm shows many small vacuoles. The corpuscle shown in Plate II,
Fig. 9, is probably a later stage of the form just described. The central
protoplasmic mass has become converted into an ellipsoidal pale bo'dy
(pm). The small circular body in this shriveled mass is probably the
nucleus. Some corpuscles like the one shown in Fig. 9 are found in
wdiich the central mass has entirely disappeared. The further growth
of corpuscles of this type seems to be by formation of colloid lamella? as
in the ordinary type. They soon become indistinguishable from other
The cystic type of corpuscle is rare in the pig. This evidently corresponds to the form in amphibia that misled Nusbaum and Machowski
into reviving Afanassiew's theory. The central masses, in Figs. 8 and 9,
might readily be mistaken for red corpuscles in animals in which these
cells are nucleated. But the red cells of the blood of the pig are not
nucleated at this stage. I have traced a number of these corpuscles (as
well as those of other types) in serial sections and have never seen any
indications of a connection to blood-vessels. Nusbaum and Machowski (19), (Fig. 1, c, S. 116) show a corpuscle which is similar to jny
Fig. 9. It will 1)0 noted that the central space in neither of these figures
is lined by endothelium. The early form of corpuscle shown by Nusbaum and Machowski (Fig. 1, d, S. 116) is very probably a normal bloodvessel with cubical endothelium. I have often found such vessels with
cubical endothelium in the interlobular tissue of the pig's thymus at 10
cm. to 12 cm. They probably may be found at other stages also. In the
thymus of a kitten, injected l)y the intra-vitam Prussian blue method previously described, the majority of the corpuscles were found to be in
early stages. The injection did not penetrate any corpuscle. I had a
somewhat better opportunity to study the relations of the corpuscles to
the blood-vessels in a 14 cm. human embryo. Here the vessels of the
thymus were all very much distended with blood and the corpuscles were
in early stages. No blood cells were found in the corpuscles.
(b) Compound concentric corpuscles are formed whenever two or more
simple concentric corpuscles begin to form so close together that they
come in contact during their later growth. An early stage of such a
corpuscle is shown in Plate II, Fig. 15. The colloid lamellae are formed
around each center until they come in contact; they are tlien formed
around both centers. In Plate III, Fig. 22, a compound concentric cor
5-t Tlio D(>\('l()|)in('iit of the Tliyiinis
piisclc is shown. 'I'lici'c aiT three siiiiph' coiu'cntric corpuscles in it —
one of them (the h>\vcst in the liyiii-c) in a very early stage. Several
lamella' ai'c common lo the older cor|)uscles, and one is common to all
three. This ari'an<i'ement of the lamella' is a mechanical efTect of the
tension in the cytoplasm, due to the ccntrifug^al pressure from the two
centers. The size of the separate centers in a compound corpuscle depends upon the stage they have reached when they come in contact. If
a compound coi'puscle he formed hy the union of two simple corpuscles
in an early stage, as in Plate 111, Fig. 19, all indications of its compound nature are soon lost. A corpuscle originally compound may, then,
in later stages, hecome indistinguishal)Ie fi'om simple corjniscles. The
simple corpuscles uniting to form a compound concentric cor]niscle may
be of any of the types previously described.
B. Irregular Corpuscles.
This grou]) includes those corpuscles which are not at first concentric.
Concentric areas may appear later. According to the classification previously given, I distinguish a compact type and a reticular type.
(a) The compact type (Plate III, Fig. 16) first appears as a compact
area of syncytium of irregular shape. It is recognizable l)y the colloid
it contains. The nuclei are not noticeably increased in size and have no
regular arrangement. Their chromatin still stains dark with nuclear
stains. The colloid (c f) is not yet solid. The corpuscle has no distinct
center. These corpuscles grow by direct colloid transformation of the
adjacent syncytium. No distinct lamella^ are formed. The colloid may
remain in the fibrous condition shown in the figure (c f) or it may
become solid, but it never reaches the deeply staining condition unless a
concentric area be established.
A later stage of this tyi)e is shown in Plate III, Fig. 23. The corpuscle is sharply marked off from the syncytium. Some of its colloid is
solid. A concentric area (cs) is beginning to form. The nuclei are
not markedly different from those of the adjacent syncytium. These
corpuscles may become large and branched. Often one or more concentric areas are developed after the corpuscle has attained considerable
size. By the growth of these concentric areas, irregidar corpuscles may
become converted into concentric corpuscles.
(b) The reticular type is produced by colloid formation in the ordinary reticulum of the medulla. In the types previously described, the
spaces of the reticulum are \isually obliterated as the colloid formation
advances; but in this form the spaces persist as a part of the corpuscle.
Pure reticular corpuscles vary greatly in size, sometimes involving only
E. 'W Bell ■ 55
one node of the syncytium. Thev are never concentric, and never form
lamellae. Eeticnlar areas often occur in other forms of corpuscles. In
tiiis way leucocytes are often involved in tlic corpuscle, since they lie in
the spaces of the reticulum. Lymphocytes often get into a corpuscle in
the lymphohlast condition, heing cut off hy tlio formation of lamellae
outside tliem (Plate III, Fig. 22). The Icui'Dcytes shut in the cor])uscle in this way during development uuiy not degenerate. They probably persist and help to remove the corpuscle in its final stages of degeneration.
The amount of expansion of the cyt()i)lasm before or during the colloid
transformation is probably small in the irregular reticular corpuscles,
since it does not obliterate the spaces of the syncytium. In the compact
type, the spaces of the syncytium are obliterated and there is evidence
of some expansive force (note the arrangenunit of the nuclei in the
upper part of Fig. 23, Plate III). In the figure referred to, the number
of nuclei in any part of the corpuscle is less than in an equal area of the
adjacent reticulum. These facts indicate that there is an expansion of
the cytoplasm. That this expansive force does not produce a concentric
form is due primarily to the fact that there is no expansion of a nucleus
and distinct center of formation as is present in concentric corpuscles of
the ordinary type. The absence of the onion-like structure in irregular
corpuscles is due to the fact that the colloid is not laid down in lanielUw
Significance of the corpuscles of Hassall. It has been shown in the
preceding pages that the corpuscles of Hassall in the pig are not epithelial remnants, and also that they are not formed from blood-vessels.
There is no evidence connecting their development with the involution
of the thymus, for they begin to form before the lymphoid transformation is complete and are most numerous when the thymus is at the height
of its development. I have not been al)le to see the decrease in the rate
of growth of the medulla described by Amnumn, and even if such did
occur it is difficult to understand how it could cause the formation of a
The above theories are, therefore, inconsistent with the facts of development in the pig. It seems to me that the formation of a corpuscle is
not to be regarded as a ]u-ocess of degeneration. The fact that the formation of colloid is an essential feature in the development of every
corpuscle is a strong argument that it is a form of secretion such as
occurs in its neighl)oring branchial derivative, the thyroid. The fact
that the corpuscles differentiate in an aj^parently uniform syncytium is
further evidence airainst a theory of degeneration. Since the lymphocyte-forming fuiu-tion of tbc tliymus is probaI)ly secondary, it is not
56 The Dov(^lopin(<Tit of tlio Thymus
unreasonable to suppose that its primitive fimetion was the formation
of a colloid secretion such as occurs in the thyroid, and that the corpuscles
are abortive expressions of this primitive function/
Giant Cells.
Polykaryocytes may often 1)0 seen in the medulla. Tliese bodies develop froui the syncylium of the uieilulla. They are first noticeable as
groups of small s])herieal nuclei in a solid area of the syncytium. Tliese
nuclei stain with merlium intensity and are all very similar in size and
color. The area containing this group of nuclei becomes a well-defined
node of the reticulum and persists as such. A polykaryocyte is, therefore, a large node of the reticulum containing a number of small nuclei
very similar in appearance. These cells often occur in groups. They
are entirely distinct from the corpuscles. They are evidently similar
to the polykaryocytes found in bone marrow and other lymphoid tissues.
The following is a resume of the development of the thymus in the
pig :
The thymus of the pig is probably developed entirely from the endoderm of the third gill pouch.
By a gradual process of vacuolization and liquefaction of the cytoplasm, the epithelial syncytium of the thymic anlage is converted into a
cellular reticulum.
From the first appearance of vacuolization, three types of nuclei are
present: large pale nuclei; small dark nuclei (lymphoblasts), and large
dark intermediate forms.
The lymphoblasts gradually break loose from the cellular reticulum,
moving into its spaces and forming lymphocytes. Mitoses are most numerous at the period of the most rapid formation of lymphocytes. The
medulla continues to form lymphocytes at least as late as birth.
Lymphocytes appear in the connective tissue aromul tlu> th.ynuis
shortly after they are formed ; and lymphoblasts, wdiich are distinguishable from lymphocytes only by being embedded in the syncytium, arc
present in the thymus a long period before lymphocytes are found anywhere in the neighborhood of the thymus.
The celhdar reticulum of the earlier stages persists in a modified form
as the reticulum of both cortex and medulla. It retains more cytoplasm
' Ver Eecke (28) believes that the corpuscles in amphibians are of a
glandular nature.
E. T. Bell 57
in tlio inedulla. Practically all the reticnhira of both cortex and nie(liilla. as well as the lymphocytes, are, therefore, of epithelial origin.
The {■oi-])iisclos of Hassall develop from the syncytium and arc, therefore, epithelial in origin. They are, however, not to be considered as
remnants of the original epithelial anlage.
In development various types of corpuscles are distinguished. Tlie
ordinary typo of concentric corpuscles first appears as an enlarged clear
nucleus around which colloid is being formed. Before or during the
formation of colloid, the cytoplasm increases in quantity, filling the
spaces of the reticulum and producing a centrifugal pressure which
shapes the newly-formed colloid into concentric lamellse and flattens the
neighboring nuclei, making them concave toward the center. The central nuclei usually become obliterated.
The epithelioid type is distinguished by its resemblance to large epithelial cells, this appearance being due to the formation of colloid
lamella" around largo areas of clear cytoplasm. The central part of the
corpuscle usually remains unchanged until after some of the colloid
lamellfE are formed.
The cystic type differs from the ordinary type only in that the central
part undergoes vacuolization instead of colloid transformation. Those
with concentric vacuoles may simulate blood-vessels containing nucleated
red cells. Corpuscles never contain erythrocytes; neither can they be
injected at any stage of development. Serial sections also show tliat
there is no connection to blood-vessels at any stage.
Compound concentric corpuscles are formed by the union of two or
more simple concentric corpuscles during development.
Irregular corpuscles are not concentric ' in arrangement, and are
formed in the syncytium in an irregular manner. In the compact type
of irregular corpuscles, concentric areas may form.
The formation of colloid is an essential feature in the development of
every corpuscle, and is not to be considered as a process of degeneration.
Since the conclusion of my work and after my manuscript was given to the
publishers, two articles dealing with the thymus have appeared.
Ph. Stohr (Ueber die Thymus, Sitzungsberichte der phys.-med. Gesellschaft
zu Wiirzburg, June 8, 1905) believes that the thymus first epithelial in nature
becomes converted entirely into small cells of lymphoid appearance. Later
the large reticulum cells are formed from these by enlargement. The corpuscles of Hassall are formed by the massing together and enlargement of
these lymphoid-like cells. The small round cells of the gland are epithelial
in origin but are to be regarded not as lymphocytes but as epithelial cells.
The thymus is not a source of lymphocytes.
58 ''J'lu^ Devolopmont nf the Tliymiis
The author apparently believes that none of the small rounrt cells leave
the gland though he admits that lymphocytes enter. But as mentioned above
the zone of connective tissue immediately around the head at 7 cm. may
contain even more lymphocytes than are present inside the gland at that time
If these are all entering the gland then it is probable that most of the small
round cells are really lymphocytes. This conception then does not simplify
the problem but is only a theoretical compromise between the two views a'i
to the origin of the lymphocytes.
J. Aug. Hammar (Zur Histogenese und Involution der Thymusdriise, Anat.
Anz. Bd. XXVII, June 17, 1905) regards the reticulum as formed from the
epithelial anlage but thinks the evidence at hand insufficient to decide the
question as to the origin of the lymphocytes. He finds lymphocytes outside
the thymus in many animals (man, cat, chick, frog) before any are present
inside the gland. The corpuscles of Hassall develop from the epithelial
reticulum and undergo hyaline (colloid?) degeneration.
My description of the formation of the corpuscles of Hassall differs essentially from Hammar's, in that I believe the formation of the corpuscle consists
in the expansion of the cytoplasm of the syncytium and its conversion into
colloid. Hammar did not recognize " colloid in formation," though he speaks
of the coarse fibrillar structure of the protoplasm. He -did not describe such
corpuscles as are shown in Fig. 7, Plate II.
The considerations presented above in favor of the epithelial origin of the
lymphocytes seem to me much stronger than those given by Hammar. His
statements as to the presence of lymphocytes around the thymus before they
are present inside are to be taken with some reservation inasmuch as he
mentions small round cells separate from the syncytium earlier, but regards
ihem as degenerating epithelial cells (S. 65). His figure from the human
foetus (Fig. 18, S. 66) does not seem to be strong support for his statement.
Certainly many lymphocytes are present in the pig thymus when the reticulum is broken up as much as shown in the figure referred to. It is also to
be borne in mind that the different parts of the thymus undergo the lymphoid
transformation at different times and that a single section may therefore be
la. Afanassiew, B. — Ueber die concentrischen Korper der Thymus. Archiv
f. mikr. Anat., Bd. XIV, 1877.
lb. Weitere Untersuchungen iiber den Bau und die Entwickelung
der Thymus und der Wintcrschlafdriise der Saugethiere. Archiv f.
mikr. Anat., Bd. XIV, 1877.
2. Ammann, a. — Beitrage zur Anatomic der Thymusdriise. Basel, 1882.
3a. Beard. — The development and probable function of the thymus. Anat.
Anz., Bd. IX, 1894.
3b. The true function of the thymus. Lancet, 1899.
4. Born, G.— Ueber die Derivate der embryonalen Schlundbogen und
Schlundspalten bei Saugethieren. Archiv f. mikr. Anat., Bd. XXII,
5. CoRNiL et Ranvier. — Manuel d'histologie pathologique. Paris, 1869, p.
135 (cited from Ammann).
E. T. Bell 59
6. EcKER. — Art. " Blutgefassdrusen," Wagner's Handw. der Phys., Ill (cited
from Ammann).
7. Priedleben, a. — Die Physiol, der Thymusdriise. Frankfurt, 1858.
8. GULLAND. — The Development of adenoid tissue with special reference
to the tonsil and thymus. Laboratory Reports issued by the Royal
College Phys., Edinburgh, Vol. Ill, 1891.
9. GiJNZBURG. — Ueber die geschichteten Korper der Thymus. Zeitschr. f.
klin. Med., Bd. VI, 1857, S. 456 (cited from Henle und Meissner.
Bericht iiber die Fortschritte der Anat. u. Physiol.).
10. Hassall. — The microscropical anatomy of the human body in health and
disease. London, 1846 (cited from Ammann).
11. Hermann et Tourneux. — Article thymus, Diet, encycl. des Sciences Medi
cales. Troisieme Serie, 17, 1887.
12a. His, W.— Zeitschrift f. wiss. Zoologie, Bd. X, S. 348. Leipzig, 1860.
12b. Anatomic menschlicher Embryonen. Leipzig, 1880, S. 56.
13. Jackson, C. M. — Zur Histologie und Histogenese des Knochenmarkes.
Archiv f. Anat. und Physiol., Anat. Abth., 1904.
14. Kastschenko. — Das Schicksal der embryonalen Schlundspalten bei
Saugethieren. Archiv f. mikr. Anat., Bd. XXX, 1887.
15. Klein. — Neuere Arbeiten iiber die Glandula Thymus. Centralbl. f. allg.
Pathol, u. pathol. Anat., 1898.
16. Langerhans und Savei.iew. — Beitrage zur Physiologic der Brustdriise.
Virchow's Archiv, Bd. 134, 1S93.
17a. Maurer. — Schilddriise und Thymus der Teleostier. Morph. Jahrb.. Bd.
XI, 1886.
17b. Schilddriise, Thymus, und Kiemenreste bei Amphibien. Morph.
Jahrb., Bd. XIII, 1888.
17c. Schilddriise, Thymus, und andere Schlundspaltenderivate bei der
Eidechse. Morph. Jahrb., Bd. XXVII, 1899.
17d. In Hertwig's Handbuch der Entwickelungslehre der Wirbelthiere,
Lief. 6-8, S. 131 ff., 1902.
18. MoNGUiui. — Sulla glandula timo. Parma, 1885 (cited from Prenant).
19. NusBAUM, J., und Maciiowski. — Die Bildung der concentrischen Korper
chen und die phagocytotischen Vorgange bei der Involution der
Amphibienthymus, etc. Anat. Anz., Bd. XXI, 1902.
20. NusBAUM, J., und Prymak, T.— Zur Entwickelungsgeschichte der lym
phoiden Elemente der Thymus bei den Knochenfischen. Anat. Anz.,
Bd. XIX, 1901.
21. Paulitzky. — Disquis. de stratis glandulse thymi corpusculis. Habilita
tionsschr., Halis, 1S63 (cited from Henle und Meissner's Bericht iiber
die Fortschritte der Anat. und Physiol.).
22. Prenant. — Developpement organique et histologique du thymus, de la
glande thyroide, et de la glande carotidienne. La Cellule, Tome X,
23. Prymak. T. — Beitrage zur Kenntnis des feineren Baues und der Involu
tion der Thymusdriise bei den Teleostieren. Anat. Anz., Bd. XXI,
60 'I'lio Dovclnpmont of the T'hynms
24. SciiAKFER, J. — Ueber den feirieren Bau der Thymus und deren Beziehung^
zur Blutbildung. Sitzungsber. d. K. Acad. d. Wissensch. Math.naturw. Kl. Wien., Bd. CII, Abt. Ill, 1893.
25. SciiEnEL, J. — Zellvermehrung in der Thymusdriise. Archiv f. mikr.
Anat., Bd. XXIV.
26. Stieda, L. — Untersuchungen uber die Entwickelung der Glandula Thy
mus, Glandula Thyroidea, und Glandula Carotica. Leipzig, 1881
(cited from Hermann et Tourneux).
27. Sx'LTAN.^Beitrag zur Involution der Thymusdriise. Virchow's Archiv,
Bd. 144, 189G.
28. Vek Eecke. — Structure et modifications fonctionelles du thymus de la
grenouille. Bulletin de I'Acadc'mie royale de MC'dicine de Belgique,
29. ViKciiow, R. — Kritisches iiber den oberschlesischen Typhus. Archiv, Bd.
3, 1851, S. 222.
30. Wallisch. — Zur Bedeutung der Hassall'schen Korperchen. Archiv f.
mikr. Anat., 1903.
31. Watxey. — The minute anatomy of the thymus. Philos. Transact, of the
Royal Society of London, Vol. 173, Part III, 1883 (cited from Prenant).
All the figures were drawn with Leitz obj. 1/12, oc. 4, and camera lucida.
The magnification after the reduction of the plates is about 1060 diameters.
All drawings were made from transverse sections of the mid-cervical segment of the thymus unless they are otherwise indicated.
The following abbreviations designate the structures indicated in all the
figures :
c f — colloid in formation. I p n — -large pale nucleus.
cl — -calcareous deposit. m — nucleus in mitosis,
c s — solid colloid. vid — beginning of medulla,
c s d — solid colloid that stains n — nucleus.
deeply. o c — old colloid.
e — erythrocyte. p m — protoplasmic mass.
end — -endothelial nucleus. sf — fibril in syncytium.
I — lymphocyte. ss — space in syncytium.
Z& — lymphoblast. v — vacuole.
I d n — large dark nucleus.
Plate I.
Fig. 1. From a 3.7-cm. pig embryo. Stained with iron-lijematoxylin and
Congo red. Vacuolization of the cytoplasm and differentiation of the nuclei
have begun.
Fig. 2. From the thoracic segment of a 4.5-cm. pig embryo. Stained with
iron-haematoxylin and Congo red. A cellular reticulum is now formed. Large
pale nuclei, lymphoblasts, and the large dark intermediate forms are present.
Fig. 3. Epithelioid type of concentric corpuscle. From a 16-cm. pig embryo. Stained with hajmatoxylin and Congo red. Colloid lamellae (c s d)
E. T. Bell 61
separate large areas of cleai- cytoplasm, causing the appearance of large
epithelial cells. Colloid is being formed between the lamellae and around
several nuclei.
Fit;. 4. From a 8.5-cm. pig embryo. Stained with iron-hsematoxylin (not
decolorized). The medulla has appeared. Lymphocytes are present between
the epithelial cords.
Fig. 5. From a 7-cm. pig embryo. Stained with iron-htematoxylin and
Congo red. A few lymphocytes have been formed. In the cellular reticulum
are large pale nuclei, lymphoblasts, and large dark intermediate nuclei. The
nuclei in mitosis are very compact.
Fig. 6. From the medulla of a 24-cm. pig embryo. Stained with Jackson's
modification of Mallory's method (ref. in text). Many fibrilte are seen in
the syncytium.
Plate II.
Fig. 7. Ordinary type of simple concentric corpuscle. From a 16.5-cm.
pig embryo. Stained with hsematoxylin and Congo red. The corpuscle is
well advanced in development. Concentric lamellag of colloid have been
formed. The cytoplasm between the lamellse is in an early stage of colloid
transformation. Colloid fibers cut transversely appear as dots. The nuclei
are becoming flattened by the pressure of expansion. The central mass stains
irregularly and all traces of the nuclei in that region are gone.
Fig. 8. Two cystic concentric corpuscles. From a 16-cm. pig embryo.
Stained with iron-hgematoxylin and Congo red. On the left, a nucleated
mass of protoplasm has been separated off by the formation of a vacuole
annular in section. This might be mistaken for a blood-vessel containing a
nucleated red cell. In this central protoplasmic mass the nucleus is shrunken
and the cytoplasm vacuolated. In the small corpuscle on the right, two large
vacuoles have formed.
Fig. 9. Cystic concentric corpuscle. From a 14-cm. pig embryo. Stained
with hsematoxylin and Congo red. The central protoplasmic mass is pale
and shrunken. The small circular body in it probably is the remains of
the nucleus. Colloid lamellae are forming. Colloid fibers cut transversely
appear as dots.
Fig. 10. Cystic concentric corpuscle. From a 10.5-cm. pig embryo.
Stained with hsematoxylin and Congo red. The center contains no colloid
and seems to be softening. The nucleus is shrunken.
Fig. 11. Ordinary concentric corpuscle in a very early stage. From a
l{j.5-cm. pig embryo. Stained with iron-ha'matoxylin and Congo red. The
nucleus is enlarged and colloid is forming around it. A few colloid fibers
may be seen in the cytoplasm for some distance from the central nucleus.
Fig. 12. Ordinary concentric corpuscle. Several nuclei are involved.
From a 10.5-cm. pig embryo. Stained with hsematoxylin and Congo red. The
deeply-staining colloid has completely obliterated the central nucleus (in the
region o c), and nearly obliterated another («'). Some of the colloid now
stains pale (o c).
Fig. 13. Ordinary concentric corpuscle. . From a 10.5-cm. pig. Stained
with hiematoxylin and Congo red. The central nucleus is being obliterated
(52 The I)i'V('l()})iiHiit nf the Tlivnius
by the deeply-staiuing colloid. The neighboring nuclei are beginning to
show the effect of the centrifugal pressure.
Fig. 14. Ordinary concentric corpuscle in an early stage. From a 10.5-cm.
pig. Stained with haematoxylin and Congo red. A band of deeply-staining
colloid has been formed. Just outside this is colloid in formation.
Fig. 15. Two simple concentric corpuscles which would have formed a
compound concentric corpuscle. From a 10.5-cm. pig. Stained with haematoxylin and Congo red. The left corpuscle is a little more advanced than
Fig. 11. The right corpuscle shows a large area of deeply-staining colloid
which has pressed the nucleus into a small Irregular shape.
Plate III.
Fig. 16. Compact irregular corpuscle In an early stage. From a 14-cm.
pig embryo. Stained with haematoxylin and Congo red. The colloid is not
yet solid. The nuclei are not essentially different from those of the adjacent
Fig. 17. Ordinary concentric corpuscle. From a 12-cm. pig embryo.
Stained with haematoxylin and Congo red. There is a large, central, deeplystaining colloid mass in which calcareous deposits (c?) have been made.
The neighboring nuclei show the effects of the centrifugal pressui-e.
Fig. is. Epithelioid concentric corpuscle in an early stage. From a 10.5cm. pig embryo. Stained with haematoxylin and Congo red. The outer colloid lamella marks off a nucleated mass of cytoplasm resembling a large cell.
The nucleus is undergoing the same changes as occur in the central nucleus
of an ordinary concentric corpuscle.
Fig. 19. Compound concentric corpuscle. From a 10.5-cm. pig embryo.
Stained with haematoxylin and Congo red. This would have soon lost all
evidence of its compound nature.
Fig. 20. Epithelioid concentric corpuscle. From a 10.5-cm. pig embryo.
Stained with hematoxylin and Congo red. Some colloid is forming outside
the circular area. Solid deeply-staining colloid is forming.
Fig. 21. Ordinary concentric corpuscle, showing a variation from the
usual type. From a 16.5-cm. pig embryo. Stained with iron-haematoxylin
and Congo red. The central nucleus is reddish but not enlarged. No solid
colloid has been formed.
Fig. 22. Compound concentric corpuscle. From a lG.5-cm. pig embryo.
Three centers are present. The pale colloid in the upper part is probably
older than the deeply-staining variety. In the lower part of the figure, a
young corpuscle is shown.
Fig. 23. Compact irregular corpuscle. From a 16-cm. pig embryo. Stained
with haematoxylin and Congo red. Some of the colloid is solid. No definite
center is present but one is beginning to form (c s) . The nuclei are not
markedly different from those of the adjacent syncytium.
^^^ Idn
J^'\ i
1 ss
3 ?rt
5 ' ^P>i
rsil .,
From the Histological Laboratory of Cornell University Medical College.
Neio York, N. Y.
With 3 Text Figures.
Within the past decade our knowledge of the functions of the adrenal
glands, and of their relations to the rest of the economy, has been greatly
enhanced l)y many careful cliomical and physiological researches. Tlie
recent studies of Aichel (1), Wiesel (2, 3), Soulie (4), and others have
placed the early development of the organ upon a fairly certain basis.
These advances in the physiology and embryology of the organ have not
as yet been accompanied by corresponding advances in our appreciation
of its minute anatomy. Hence this branch of the subject is, at the
present time, one of unusual interest.
The intimate relation of the parenchyma of the adrenal to its bloodvessels, as shown by the general tendency to regard the organ as a true
gland whose secretion enters its blood-vessels and leaves the organ
through its efferent veins, makes it specially important that these vessels
should be carefully studied and their structure and distribution accurately
The exhaustive study of Flint (5), on the course of the adrenal vessels,
based as it was upon carefully prepared reconstructions, leaves little to
be desired along this line. The writer is, however, unable to find in the
literature any reference to the minute structure of the veins of the adrenal,
with the notable exception of Minot's (6) article on sinusoids.
To be sure Pfaundler (7) mentioned the occurrence, in the medulla of
the adrenal, of venous vessels whose only wall consisted of endothelium.
Gottschau (8) also, though omitting their description, has figured similar
vessels in his Plate XVIII, Fig. 1. But as to the structure of the larger
blood-vessels of the adrenal glands the literature seems to be entirely
The architecture of the arterial walls does not appear to offer any
distinctive peculiarities, the tissues of which they consist being arranged
in a manner precisely similar to that which characterizes the arteries of
American Journal of Anatomy. — Vol. V.
64 The N'ein.s oi' the Adrenal
similar size occnrriiiii- in olliei' or<ians. 'I'lie veins, liowever, j^resent
distinct and remarkable peeidiarilies which it is the iiur))()se oi' the
present |>a|)ci- to describe.
Methods (iiid Mafcr'ud. — The tissue used for this stud}^ has included
specimens of the adrenal from twenty-one human adults, together with the
casual examination of fetal adrenals of the pig and of man. The
adrenals of other mammals, e. g., nionke}^ dog, cat, rabbit, and guineapig, have also been more or less carefiilly studied.
These tissues have been fixed and hardened with many reagents, among
which are Zenker's solution, formol. Miiller-formol, alcohol, corrosive
acetic mixture, Tellyesniczky's fluid, and Flemming's solution. The
stains used were hematein by various methods, acid hematein, iron
hematein, etc., and for counter stains eosin, orange. Van Clieson's picrofuchsin, Weigert's elastic tissue stain, Mann's methyl blue-eosin mixture,
Congo-red, and Ehrlich's triacid mixture. A combination of Mann's
hematein, Weigert's elastic tissue stain and A'an Gieson's picro-fuchsin,
gave the best results for the differentiation of the muscular and connective tissues. This method was applied as follows, and may be used
after any of the above fixatives.
1. Stain 10-12 minutes in Mann's hematein or in Bohmer's hematoxylin, until somewhat overstained. 2. Wash well in water. 3. Stain
10-20 minutes in recently prepared resorcin-fuchsin solution after the
method of Weigert (9). -1. Wash in water. 5. Stain 1-3 minutes in
the freshly prepared picric acid-acid fuchsin solution of A^an Gieson
(10). 6. Wash and dehydrate in 95 per cent, or in absolute alcohol.
7. Clear, and mount.
Types of Adrenal Veins. — The efferent veins of the adrenals arise in
the medulla of the organ by the union of the broad capillaries of the
medulla and the adjacent zone of the cortex. These capillaries form
broad thin-walled vessels which have been described by Minot (6) as
sinnsoids. They converge toward the middle of the medulla, where they
pass into somewhat larger vessels, which, for convenience, may be termed
small central veins. These veins tend toward the hilum, are relatively
short, and by union with one another soon form thicker-walled vessels
which may be described as large central veins. These large veins pass
toward the hilum, near which, they unite to form a large efferent vessel,
the suprarenal vein. This last vein makes its exit from the hilum of the
organ and enters either the vena cava inferior, as is the rule on the right
side, or the renal vein, as frequently occurs on the left.'
From this brief review of the course of these vessels it will l)e seen thai
four distinct venous types have been enumerated, and it is the purpose
Jeremiah S. Ferguson 65
of the writer to show that these types exhibit well-defined structural
Observations. — The sinusoids, after the careful description by Minot
(6), will require but brief mention. These vessels possess the wall of a
capillary and the lumen of a venule. A number of such vessels may be
seen in Fig. 1, in the central portion of the medulla, on either side of the
group of central veins. Their wall consists of nucleated endothelial
plates which rest directly upon the parenchymal cells. Their lumen is
several times the diameter of the- medullary capillaries. They are dis
> >
Fig. 1. A group of vessels from the central portion of tlie medulla of the
human suprarenal gland, a, sinusoids; h. small central veins. Fixation, 5
per cent formalin; stain, Mayer's hematein; thickness, 8// ; photomicrograph,
X 100.
tinguished from the small central veins by the absence of connective tissue
from the wall of the sinusoids.
The small central veins are of the type shown in Fig. 1. The wall of
these vessels consists of two coats, endothelial and connective tissue. The
latter is always relatively thin, though the vessels possess a very considerable lumen. Venules of this type of structure. Fig. 1, collect the blood
from the sinusoids of the medulla. Frequently, however, the sinusoids
open directly into the small central veins and venules, the connective
tissue of the venous wall being occasionally continued for a very short
distance upon the endothelium of the sinusoid.
The Veins of the Adrenal
The connective tissue of the small central veins is richly supplied with
elastic fibers, which are disposed in oblique nnd circular directions,
Fig. 2. The medulla ol a suprarenal gland of man, showing a group of
large central veins. The middle and lowermost veins are in transection, the
uppermost vessel in longitudinal section. The series of sections shows that
this last vessel is a branch of that in the middle of the figure. Fixation,
Zenker's fluid; stain, hematein and methyl-blue, Mann's method; thickness,
10 // ; photomicrograph, X GO.
occasional elastic fibers are also longitudinal. The typical small central
veins contain no muscle. As they ap]iroach tlieir termination in the
JciTiniah S. i^\'rgiL-'on 67
large central veins a few smooth nnisole fibers are found, but these are
always disposed in a longitudinal direction. As soon as longitudinal
muscle fibers appear in apprcciahh' uiiinl)ors the venous wall acquires the
type of the succeeding variety, ihc large central vein.
In the large central veins, as in the small, but two coats can be readily
distino-uished. The inner coat, or intima, in these vessels consists of a
lining endothelium, which rests upon a very thin membrane of delicate
connective tissue, containing numy elastic fibers.
The outer coat, or adventitia, is also a vei-y thin m(!ml)rane of fibroelastic tissue, but its fibrous Ijiindlos are coarser than those of the intima,
and its clastic fibers form a very close network. The outer portion of
this coat contains a few longitudinal smooth muscle tibers. The great
majority of these fibers, however, ai'c arranged in the form of longitudinal
ridges which project into the adjacent medullary tissue. From one to live
of these muscular ridges occur in the circumference of the vein (Fig. 2
and 3). Except at those points at which the muscle occurs, the venous
wall is extremely thin (Fig. 3). The muscular ridges are frequently so
large as to materially obstruct the lumen of the vessel (e. g., the middle
vessel in Fig. 2, also the uppermost vessel, which is cut in v(!ry neirly
longitudinal section), and they form so noticcuible a peculiarity that
their presence may be considered charactei-istic of this type of vessel.
The writer has never failed to find these peculiar muscular ridges more or
less highly developed in each of the human adrenals which he has examined : he believes them to be constantly present. They are less highly
developed in the suprarenal vessels of the lower mammals, but even there
they may frequently be demonstrated.
The muscle fibers of the larger ridges are an-anged in bundh^s which
are enveloped in fibro-elastic septa of connective tissue. All of the
muscle fibers in these bundles are longitudinally disposed. This arrangement is well shown in Fig. 3, in which a hii'ge central vein is seen in
transection at a point near the entrance of a large branch. p]xamination
of sections somewhat higher in the series shoAvs the union of these two
In the section photographed, tlu; l)raiicli has been longitudinally cut.
The fine dai'k lines shown in the figure are bauds of elastic fibers which
are enveloped in delicate white fibrous tissue inclosing the cut ends of
the bundles of smooth muscle. The tendency to form longitudinal ridges
is shown in this figure by the irregular disti-ibufion of the muscle, one
side of the vessel, in both the parent stem and the branch, beirig almost
devoid of muscle fibers. The muscular character' of these ridges is
beyond doubt.
The Veins of the Adrenal
Fig 3 Large central veins from the medulla of the human suprarenal
gland The figure shows the distribution of the elastic tissue and the bundles
of smooth muscle which are seen in transection in the larger vein and m
longitudinal section in the smaller ones below. The series shows thes.
latter vessels to be branches of the former, the section being selected to
show a plane near the point of division. The smaller vessels are jeiy obliquely cut and the muscle is distinctly longitudinal. Fixation Zfnker s
fluid; stain, Mann's hematein, Weigert's elastic tissue, and Van Gieson s
picro-fuchsin; thickness, 10 a; photomicrograph, X 37.
Jeremiah S, Eerguson 69
The writer has observed that the formation of such heavy ridges as
those shown in Fig. 3, nearly always occurs at those points where the
vessel branches. It is possible that, as in the case of the somewhat similar
ridges in the veins of the erectile tissues (see Kolliker's Handbuch der
Gewebelehre, 6te Aufl., 1902, pages 486 and 487), these muscular protuberances may to some extent serve the purpose of valves.
As the large central veins approach the hilum of the organ they form
still larger vessels which partake of the structure of the suprarenal vein.
The point of transition from the one type to the other is variable, occasionally the type of the large central veins is continued to the exit of the
suprarenal vein at the hilum of the organ. More frequently the primary
branches of the suprarenal vein may be traced for a considerable distance
into the medulla of the organ, still retaining the type of structure found
in the larger vessel.
The suprarenal vein presents three coats, intima, media, and adventitia.
The tunica intima, in addition to its endothelial lining, possesses a thin
membrane of very delicate connective tissue in which occasional branched
connective tissue cells may be distinguished; such cells are, however, very
scanty. This coat also contains a delicate network of elastic fibers.
The tunica media of the suprarenal vein is extremely thin, rarely ever
does it exceed in thickness the tunica intima. It consists chiefly of
fibro-elastic tissue, the elastic fibers forming quite a dense network.
Few muscle fibers occur in this coat, nowhere are they found in sufficient
numbers to form a definite layer, as in veins of similar size in other
organs. Some of the muscle fibers are circularly disposed, but many of
them are longitudinal.
The tunica adventitia is by far the thickest of. the three coats and
forms two-thirds to five-sixths of the entire vascular wall. It consists
chiefly of smooth muscle fibers, all of which are longitudinally disposed.
These smooth muscle fibers form characteristic coarse bundles which are
distributed around the entire circumference of the vessel. The largest
of these bundles may occasionally form projecting ridges as in the smaller
veins, but as a rule the muscular tissue is more evenly distributed than in
the central veins. Each of the muscle bundles is enveloped in a perimysial sheath of connective tissue, which blends with that of the tunica
media. These adventitial sheaths possess a dense network of elastic
fibers, in fact the greater part of the elastic tissue in the vascular wall
is frequently found in the adventitia. On its outer surface the tunica
adventitia is continuous with the capsule of the adrenal or with the
adjacent connective tissue.
This peculiar type of vessel is not strictly confined to the suprarenal
•70 'JMu' W'ins of tlie Adiviial
gland, but occurs, more or less typically developed, in many of the large
abdominal veins, notably in the renal veins and vena cava, into which the
suprarenal veins empty. But UQwhere is this peculiar venous type more
strikingly developed, nowhere is the adventitia relatively so much thicker
than the media, nowhere is a greater proportion of the smooth muscle of
the venous wall longitudinally disposed, nowhere is there relatively less
circular muscle, than in the suprarenal vein. Eealizing the intimate
relation of the parenchyma of the organ to its blood-vessels, and adopting, if we may, the accepted physiological function of the adrenal — the
formation of an internal secretion, a powerful vaso-constrictor which is
poured into the blood within the capillaries and veins of the organ — ^the
peculiar longitudinal arrangement of the muscular tissue, the valve-like
protuberances at the junctions of the venous vessels, the absence of circular
muscle from the walls of the veins of all sizes, and the general appearance
of these vessels which are so remarkably different from the veins of most
other organs, become, to say the least, extremely significant of a close
structural relation, physiologically speaking, to the presence of an
astringent secretion in the outflowing blood current.
In this connection, one further observation is of importance. In the
periadrenal connective tissue are numbers of small veins which return the
abundant blood supply of the tissues of this region, most of them
emptying into the phrenic veins. Many of these veins do not differ from
the similar veins of other parts, but in many others the writer has observed that the muscle tissue is almost entirely disposed in a longitudinal
direction, a condition which is quite the reverse of that found in the
adipose and areolar tissues of other portions of the body.
The writer also finds that many of the small veins of the adrenal,
instead of opening into the central veins as is usually the case, pursue a
less frequent course, penetrating the cortex and capsule of the organ, and
emptying into the small veins of the surrounding connective tissue. The
frequency with which this condition was associated with the occurrence
of longitudinal muscle fibers in the periadrenal veins, suggests a more
than casual relationship between the two conditions.
In conclusion, the above facts may be summarized as follows:
1. The efferent blood-vessels of the adrenals form four successive vascular types, the sinusoids, the small central vein, the large central vein,
and the suprarenal vein.
2. Each of these types presents distinctive characteristics.
Jeremiah S. Ferguson 71
3. In all four tA^pes circular muscle is either absent or noticably
4. In the large central veins prominent and characteristic muscular
ridges are constantly present, and are frequently in relation with those
points at which the branches of these vessels enter.
5. These peculiarities of structure may possibly bear a close physiological relation to the function of the adrenal as a gland that forms an
internal secretion which has been shown to be a powerful vaso-constrictor
and stimulant of smooth or involuntary muscle.
1. AicHEL.— Munch, med. Wochenschr., 1900, XLVII, 1228; and Arch. f. mik.
Anat, 1900, LVI, 1.
2. WiESEL.— Anat. Hefte. 1901, XVI, 115.
3. Ibid., 1902, XIX, 481.
4. SouLiE.— J. de I'anat. et de la physiol., 1903, XXXIX, 197, 390, 634.
5. Flint. — Contrib. -dedicated to W. H. Welch, Baltimore, 1900, 153; also in
Johns Hop. Hosp. Rep., 1900, IX, 153.
6. MiNOT.— Proc. Post. Soc. Nat. Hist., 1900, XXIX, 185.
7. Pfaundler. — Sitz. d. Akad. d. Wissensch., Wien, 1892, CI, 515.
8. GoTTSCHAU.— Arch. f. Anat., 1883, 412.
9. Weigert.— Centralbl. f. allg. Path. u. path. Anat., 1898, IX, 289.
10. Freeborn.— Proc. N. Y. Path. Soc, 1893, 73.
Associate m Surgery, Johns Hopkins University.
From the Anatomical Laboratory, Johns Hopkins University.
With 2 Colored Plates.
As the structures of the body are being more and more carefully investigated it is found that organs are composed of like structural units,
which when repeated a number of times form the whole organ. In general these units are formed by the glandular structures, the blood-vessels, or by both, as is the case in the liver.
Some eight years ago, at the suggestion of Dr. Mall, I undertook the
study of the prostate gland, with the hope of finding structural units in
it. In this search I was successful. Since then my work has been
continued in the laboratories of Professor Born ^ of Breslau and Professor Spalteholz ^ of Leipzig, and although this communication is several years late in appearing, it should in reality have preceded those that
were published in 1899.
In the present study for the most part the prostate glands of dogs were
used. Several cadavers were injected and the gross blood supply was
studied in part from these. After the animals had been killed by chloroform, the aorta was exposed just above the bifurcation and injected with
various substances. A preliminary washing out of the blood-vessels with
salt solution was practised in a few of the first injections, but this was
soon discarded as it did not seem to enhance the value of the method.
Carmine gelatine, followed by ultramarine-blue gelatine, as an injecting
mass, gave the most satisfactory results. About 250 cc. of the carmine
gelatine were injected first, the injection being stopped as soon as all
of the tissues had acquired a maximum carmine hue. This was followed by the injection of ultramarine-blue gelatine, which was kept up
until no more of the material would pass in. The carmine gelatine
^Walker. George: Ueber die Lymphgefasse der Prostata beim Hunde.
Arch, fiir Anatomie, 1899.
= Walker, George: Beitrag zur Kenntnlss der Anatomie und Physiologie
der Prostata, etc. Ibid., 1899.
American Joukxal of Anatomy. — Vor-. V.
74 Tlif lUood Vessels of the Prostate Cilaiicl
filled the arteries, capillaries, and veins ; the blue passed into the arteries
and arterioles, displacing the gelatine and filling them, but was stopped
at the capillaries because the nltramarine-blue granules were too large
to enter them. In a specimen thus prepared the arteries appear blue, and
the capillaries and veins red. This is shown in Figure 1, with colors
reversed, in order to present the conventional appearance. As it was
impossible to get a perfectly complete injection in one specimen, several
of the best were selected and the gaps filled in, with the results as shown
in Figure 1. One section, however, is remarkably beautiful and presents
a picture very closely resembling that seen in this figure.
In order to map out the complete network of arteries surrounding a
separate lobule, I injected them with Prussian blue, then opened the
urethra, and injected carmine gelatine into a prostatic duct through a
very fine blunt hypodermic needle. A specimen made in this way is shown
in Figure 2 where the ducts are represented in brown. The capillaries
were studied in a specimen which had been completely injected with carmine gelatine. A very thin section of this was stained with iron hematoxylin, and is shown in Figure 3. The basement membrane is artificially tinted with yellow so as to make it visible.
The technique of the injecting is rendered difficult by the fact that the
situation of the gland in the pelvis is somewhat remote. In all, about
75 dogs were used before a complete circulation cycle could be seen.
Cinnabar, lampblack, and various other substances were tried, but they
did not prove as good as the combination of carmine gelatine followed by
ultramarine-blue gelatine.
When the ordinary directions for preparing carmine gelatine were followed, it always proved difficult to get a perfectly transparent substance.
The trouble is connected with the neutralization of the ammonia by the
acetic acid. The gelatine should be rendered practically neutral, but
if the reaction is carried the least bit too far, the solution becomes
cloudy. Sometimes two drops of the acetic acid are sufficient to make
turbid a whole litre of the prepared carmine. After a good many
trials, the following method was adopted : Take 10 cc. of the ordinary laboratory ammonia and dilute with 40 cc. of distilled water,
then determine by titration the exact amount of the laboratory acetic
acid which will neutralize it. After this determination has been made,
10 grms. of pure carmine are rubbed up with 50 cc. of distilled water;
then 25 cc. of the ordinarj^ ammonia are measured, and a few drops at a
time are poured into the carmine mixture which is kept constantly rubbed
up. This process is very closely watched, and the ammonia is gradually
added until the carmine is completely dissolved, and the mixture becomes
George Walker 75
translucent and assumes .a dark red color. Tlie amount of ammonia
used is determined by referring to the vessel in which the 25 cc. have been
measured. The gelatine in whatever proportion it is required — according
as a thin or thick solution is desired — is dissolved in the distilled water,
and the carmine solution is added to it. We then calculate how mucli
acetic acid will be required for the amount of ammonia w^hich has been
used; this is measured and added, drop l)y drop, to the mixture which is
constantly stirred. A sufficient quantity of water is then added to bring
the amount up to a litre. I found that in this way I could always obtain
a beautifully clear gelatine and was never annoyod by the failures and
uncertainties belonging to tlie other method.
Akteries. .
The prostate gland derives its arterial supply from the internal iliac
arteries by means of four branches; the superior vesical, the inferior
vesical, a small branch from the inferior hemorrhoidal, and a small terminal branch from the internal pudic artery. These vessels will be found
illustrated in my paper published in 1899. The superior vesical, a
branch from the internal iliac, which is given off high up, divides before
reaching the bladder, into two fair-sized branches ; the lower and smaller
branch extends downward and supplies the vesical third of the prostate ;
this branch is sometimes called the middle vesical artery. The inferior
vesical, which is a large branch, is practically the main blood vessel of the
prostate gland, and should be called the prostatic artery for, in the
majority of instances, it does not send any branches to the bladder. The
major part of the gland is supplied by this vessel; it courses along the
vesicorectal fascia and meets the prostate at its lower border, where it
usually divides into seven branches, four of these enveloping the anterior,
and three the posterior surface. The posterior are about one-half the
size of the anterior lu'anchcs. These vessels are situated in the capsule of
the gland and envelop it as the fingers of one's hand would do in clasping
a round object. From these trunks a numljer of smaller ones are given
off, so that a very close arterial network is formed over the surface of the
gland. The ])ranch from the inferior liemorrhoidal is not constant; in
fact, it appears to be more often absent than present. When it is seen,
it occurs as one or two small branches wbicli ]uQci llic prostate in its
urethral half, and extend over the surface as line vessels which anastomose
with the vesica] artery. The internal ])iulic branch is fairly constant.
It extends along Ibc membranous urotlira and ])lunges directly into the
prostatic substance usually without giving off any l)ranclies to the surface.
7G The Blood Vessels of the Prostate Gland
A slight anastomosis is occasionally seen. The vessels supplying the two
sides of the gland are distinct. The only anastomosis across the median
line is by way of the venous channels around the urethra.
From the large superficial branches above described, smaller ones are
given off at right angles, and pierce the gland in places corresponding to
the divisions of the lobules (Art. Fig. 1). Here they penetrate the
fibrous-tissue septa, and extend to the urethra, becoming smaller and
smaller, however, as they approach it, so that in this region they are seen
as very delicate terminal vessels. As they pass down, they give off
branches which penetrate into the lobule and finally divide into myriads
of capillaries which pass around the alveoli, and come in very close
relationship with the secreting cells. From these cells they are separated
simply by a delicate basement membrane composed of fine fibrils. From
the superficial vessels branches are given off which enter the lobule directly, that is, they do not pass first into the fibrous-tissue septa {8up. Br.
Fig. 1). On the anterior surface there are usually two branches which
do not give off as many smaller ones as the rest, and consequently remain
larger and extend over to the middle line, where they dip into the median
fissure and supply the median side of the lobules (Med. Br. Fig. 1).
The arrangement on the posterior surface corresponds to that seen on
the anterior surface, in so far as the supply of the lobules is concerned.
On the posterior surface toward the bladder one vessel penetrates the substance of the gland and runs directly to the caput gallinaceum (Art. Col.
Sem. Fig. 1). Here it divides into a fine network and supplies the erectile tissue of the organ. Before this vessel reaches the eminence a small
trunk is given off which extends to the ejaculatory duct (Ai-t. duct. ej.
Fig. 1). The branch supplying the caput gallinaceum is usually derived
from the pudic; sometimes it comes from the inferior vesical.
The arterial supply in the connective tissue toward the urethra is much
poorer than in the secreting portion. Here the vessels terminate in fine
branches, relatively somewhat sparsely scattered. The arterial arrangement is shown on the red side of Figure 1.
The capillaries form a very complete and elaborate network around the
alveoli of the lobule. Here, as is seen in Figure 3, they surround an
alveolus in a more or less circular manner, and upon these vessels the
cells rest almost directly, being separated only by the very delicate connective-tissue basement membrane. From this outside capillary, a folding in is seen, which forms a definite loop (Cap. L. Fig. 3.) This at
George Walker 77
first sight might appear to end blindly, but a more careful study reveals
the two branches, which sometimes appear winding around each otlier,
and presenting enlarged club-shaped ends. The cells rest on these as
they do on the circular portion. Under the low power, the epithelial cells
appear to be in direct contact with the capillaries, and it is only by the
aid of the oil immersion that a very delicate connective-tissue basement
membrane is seen. This is shown artificially colored as B. M. Fig, 3.
This membrane contains a few elastic fibers.
On the surface of the gland are veins corresponding to the arteries
which lie in the capsule. As a rule they merge into two main trunks
corresponding to the vesical arteries; occasionally several small branches
pass off into the middle hemorrhoidal vein.
The superficial veins do not drain the blood from the whole gland, but
only from the outer fourth, as is shown in Fig. 1. From the inner threefourths of the gland the blood passes towards the centre, and into the
large venous sinuses which are a continuation of the corpora spongiosa.
(Co. 8p. Fig. 1). These immediately surround the urethra. The large
venous trunks which collect the blood from the gland do not lie on the
same plane as the arteries, but are situated in the fibrous septa some
little distance removed from them. These run, as do the arteries, on the
outside of the lobule, and are interlobular, not intralobular. For the
venous return from the caput gallinaceum there is no distinct vessel corresponding with the artery, but there are anastemoses with the spongy
The venous plexus around the urethra is, as before stated, a continuation of the corpora spongiosa. The blood from this region passes away
into the internal pudic vein. Occasionally two or three small veins
drain the tissues from this region, pass out of the prostate and run along
the membranous urethra and off into the vesicorectal fascia.
There is an anastomosis of the veins in the prostate and bladder where
these organs come together, and also on the outside through the superior
vesical veins. There is, of course, an anastomosis of the urethral veins
through the corpora spongiosa plexus.
The prostate gland is supplied with blood by branches of the internal
iliac arteries, viz., the superior vesicals, inferior vesicals, inferior hemorrlioidals, and internal pudics; the main blood supply comes from the
inferior vesicals.
78 The Jilood Vessels of I lie Tr. .slate 11 land
Branelies of llieso envelop iho siirfaec! of tlie <ilnii(l and f^ivo ofT smallcMon(>s, wliieli iieiielratc between tlie lohnles in (he fihrons-tissuc septa.
The caiiilhii-ies are separated iVom I he ciiilhclial cells only by a very
lliiii haseiiieiil iiieiiibraiie.
'I'here are siipei-llcial vcins correspond iiiii' with the arteries.
I'\)i- the Older superficial fourth of the gland the return flow is towards
(lie siirfaei'. 'I'he inner three-fourths are drained by veins wliieb enijjiy
into the venous plexus immediately aronnd the nretlira.
The lobule is formed primarily by the individual gland dnets as shown
in Figure 2. The main arteries surround this lobnle wliicb they penetrate at many points. The veins leave the lobule mainly at its peripheral
5ind central (urethral) ends as shown in Fig. ].
Fid. 1 is Irom a section of a prostate gland of a dog injected with carmine
gelatine and ultramarine-blue gelatine. The arteries in the section were
blue, the veins and capillaries red. The section was cut free hand, about
r.O// in thickness, and cleared both in glycerine and in creosote. In the
figure this artery is red and this vein blue. Ar^., Arteries; Art. Col. Ficm..
Artery of the colliculus seminal is; Art. duct, cj.. Artery of the ejeculatory
duct; Col. Scni., Colliculus s(>niinalis; V. J'L, Venous plexus around the
Fio. 2. Lobule of prostate from a gland which had been injected with
uHramarine-golatine blue into the artery, and with carmine gelatine into
the prostatic duct. Pr. duct., Opening of the prostatic duct into the urethra;
Gl. Tis., Gland tissues distended with carmine gelatine; Art., Surrounding
artery. In this figure the artery is represented in red and the ducts in brown.
Fig. 3. Very thin section from Ihe i)rostate gland of a dog. Capillaries in
red, injected with carmine gelatine. Section stained with iron h;rmatoxylin,
with artificial yellow tinting of basement membrane. Oil immersion with
one inch eye-piece amplification. Cap., Capillaries; B. M., Basement membrane; 07. Ep., Glandular epithelium; Cap. L., Capillary loop.
Fig. 2
Krcj. 3
Instructor in Yertehrate Anatomy, University of Wisconsin.
With 1 Double Plate akd 6 Text Figures.
A glance at the diagrams on the next page will at once serve to show the
great difference of opinion that has prevailed in regard to the origin of
the sex-cords and rete-cords of the Sauropsida. In fact, it is hard to
conceive of any possible manner of origin that has not been held to be
correct by some well-known embryologist.
The Chelonia have remained almost untouched in the study of this
problem. Only one work has appeared upon the rete-cords (Von Moller,
98), while no work has been published upon the subject of the sex-cords.
Von ]\Ioller studied two turtles, one a specimen of Emys lutaria of
2.5 cm. plastron length, and the other Clemmys leprosa of 4.9 cm.
plastron length. He observed no connection between the testis and
Wolffian body. This caused him to remark :
" Dieser Befund wird hochst auffallig, wenn mann bedenkt dass die
Beobachtungen an Amphibien zeigen dass die Verbindungen zwischen
Hoden und Wolffschen Gauge schon dann angelegt und vollendet werden,
wenn die iibrigen Organe sich noch in der Entwickelung befinden, und
wenn das Junge in der Eischale, respective in Uterus eingeschlossen ist.
Die zwei von mir untersuchten Schildkroten hatten dagegen schon seit
Monaten die Eischale verloren, und doch war bei ihnen noch kein einzige
Verbindung zwischen Hoden und Wolffschen Gauge vorhanden, obwohl
Anlagen dieser Verbindungen sich bereits vorfanden."
It is quite unfortunate that he considered these stages to be early
enough for his purpose, since my work has shown the rete-cords to be
formed at a relatively early stage of development in Chrysemys.
Von Moller sums up his results as follows : " Ich finde also bei diesem
Thiere zwischen Hoden und Wolffschen Korper noch keine Verbindungen, dagegen im Mesorchium und im oberflachlichen Bindegewebe der
Umiere solide Zellenstrange, fiir welche ich genotigt bin einen UrAmeeican Journal op Anatomy. — Vol. V.
The Eete-Cords and Sex-Cords of Chrysemys
Text Fig. A.
Text Figs. A-D. Diagrams ilhistrating various views held by authors
whose Avritings are reviewed in this
^D.— Fundament of the adrenal
GER.— Germinal epithelium.
il/.— Mesentery.
M. P.— Malp'ghian corpuscle.
J?.— Rete-cord.
S. C.-Sex-cord.
U. r.— TJriniferous tubule.
W. I).— Wolffian duct.
spningsort anzunehmen, der wedcr in
den Geweben der Urniere noch in denen
dcs Hodens liegt denn ieh woder im
Stande bin einen Ziisammenhang mit
den gewundenen Kanalchen des Hodens
nachznweisen, noch einem solchen mit
deni Epithel Bowmanscher Kapseln
oder sonst mit Theilen der Urniere.
Ich nehme daher an, dass sie vom Peritoneum stammen." No further allusion need be made to this article.
Turning to the other groups of the
Sauropsida, we find a large mass of literature. To intelligently discuss this,
we must use precise terms. The sexcords are those masses or cords of cells
which eventually become the seminiferous tubules of the testis or the medullary cords of the ovary. The rete-cords
are those structures which eventually
give rise to the canals which unite the
seminiferous tubules or medullary cords
of the sex-glands with the ducts of the
It is not necessary to enter into a
lengthy review of the literature upon
this subject. That has been thoroughly
done by Born, 94, MihalkovicS;, 85,
Janosik, 85, Coert, 98, Winiwarter, 00,
and others. A few diagrams will suffice to show, in a sufficiently vivid manner, the wide differences between the
many views upon this subject as expressed in the papers most worthy of
The names associated with the different diagrams, Text Figures A-D. are
those of the authors who have held
views represented by the diagrams so
indicated. After the name of each
Bennet M. Allen 81
author are placed the names of the forms which he studied in arriving
at his conclusions.
A. Tubules arise from the Wolffian duct and grow into the sex-gland
fundament. Their distal portions form the sex-cords while their proximal portions form the rete-tubules.
70, Waldeyer — Chick (Gallus).
B. According to this view, evaginations grow out from the capsule
of Bowman. Distal branches from these stems pass down into the sexgland fundament to form sex-cords, while the more proximal portions of
the evaginations remain attached to the capsules of Bowman and serve
as rete-tubules.
Brann, 77, Platydactylus, Tropidonotus.
Weldon, 85, Lacerta.
Hoffmann, 8g and 92, Lacerta, Hjematopsis, Sterna, Gallinula.
Semon. 87, Gallus.
Peter, 04, Lacerta.
Brann, 77, considers the rete-sex-cords to be, in the strictest sense,
segmental in arrangement. He expressly denies that the cells that contribute to the formation of the adrenal body are derived from branches
of the evaginations from the capsules of Bowman, as asserted by Weldon,
85 and Hoffmann. 89 and 92. These two last named authors asserted
that each evagination divides into a dorsal and a ventral branch, the
former suppljdng the cells of the cortical portion of the adrenal body,
and the latter forming the sex-cords. Semon, 87, was not so clear upon
the question. He merely stated that the anastomosing cords arising from
the capsule of Bowman pass into the adrenal and sex-gland fundaments,
— the more dorsal to the former, the more ventral to the latter.
C. Large numbers of cells migrate from the germinal epithelium into
the underlying stroma. From this unorganized blastema, the sex-cords
are formed, suddenly crystallized as it were. The rete-cords are formed
of evaginations from the capsule of Bowman.
Schmiegelow, 82, Gallus.
Mihalkovics, 85, Lacerta, Gallus.
Laulanie, 86, Gallus.
D. According to Janosik, the sex-cords arise as direct ingrowths from
the germinal epithelium. Cords of cells grow from their distal ends to
the capsules of Bowman, thus forming the rete-cords. Cords of cells
grow in from the peritoneum between the sex-gland fundament and the
mesentery to form the cortical portion of the adrenal body.
Janosik. 90, Gallus. .
83 The Kete-Cords and Sex-Cords of Chrysemys
The following table will show the great difference of opinion held by
authors working upon the same identical species. The view held is indicated in the same manner as above.
Lacerta agilis — Weldon (B) ; Hoffmann (B) ; Mihalkovics (C).
Chick (Gallus)— Waldeyer (A); Semon (B) ; Mihalkovics (C) ;
Laulanie (C) ; Schmiegelow (C) ; Janosik (D) ; Weldon (?),
We cannot close an account of the literature upon the subject without
refeiring to the work of Semon, 91, upon Ichthyophis, one of the Gymnophiona, and a paper by Semper, 75, upon the Sex-glands of the
Semon^ 91, considers the nephrotome to be the ventral portion of the
mesoblastic somite. This view, by the way, is also held by Brauer, 02.
Semon states that after the nephrotome breaks away from the myotome
and sclerotome, it still remains attached to the peritoneum (unsegmented
mesoderm) by means of two bridges of cells — a lateral and a medial.
The major part of each nephrotome forms a Malpighian corpuscle of the
mesonephros. The lateral of the two bridges connecting it with the
peritoneum becomes its peritoneal funnel (nephrostome), while the
medial bridge sends out a process which divides into a dorsal branch passing to the adrenal body, and irre|gular branches (sex-cords) non-segmental in character, that pass to the sex-glands, there to come in
contact, in the case of the male, with the seminal vesicles, which, are
derived from the germinal epithelium. He holds a theory that the
pronephros extends in rudiment, at least, along the entire length of the
mesonephros, and that this pronephric rudiment develops into the
adrenal body. He considers the dorsal branches spoken of above, to be
these vestiges of the pronephros.
Semper, 75, gives the most interesting account of the rete in the male
of Acanthias. According to him, each of the 34 primary Malpighian
corpuscles of the kidney is connected with the body cavity by a peritoneal
funnel. Seven of the most anterior of these funnels lose their union
with the peritoneum and take on the form of vesicles. Three or four
of them now fuse together to form the " central canal," which lies at the
base of the testis and parallel with it. From this central canal there arise
a number of irregular anastomosing canals which extend into the testis
and come in contact with the true sex-structures (Vorkeimketten) that
have arisen from the germinal epithelium. This net-Avork of rete-cords
he calls the rete-vasculosum.
In other forms there exists a somewhat modified condition of considerable interest. In comparing Acanthias and Mustelus, Semper said:
" Trotzdem scheint ein grosser Untershied in Bezug auf die Entstehung
Bonnet j\r. Allen 83
des Centralcanals des Ilodens /wisclien ]\[usleliis ;ind Acantliias zu bestehen. Bei dieser Gattung wird or seiner ganzer Liingc nach gebiklet durcli
die Verwachsimg der seitlicli vom Segmantalgang nach vorn sich wendenden Trichterblasen. Seitliche Aiisbuchtiingen der letztoren bildeii den
basalem Theil der rete vascolosum. Bei Mnstehis dagegen ist es nur der
vorderste iiber die Hodenfalte hinaiis vorgreifende Abschnitt des Centralcanals den mann entstanden ansebcn konnte, denn nur an diesen setzen
sich 2 (oder 3) Segmentalgange an. Der ganze iibrige viel langere
Theil des Centralcanals entsteht aiis den in das Stroma der Epithelfalte
eingestiilpten Keimepithel Zellen."
Balfour, 78, shows that in the forms which he studied, the anterior
end of the sex-gland only, was directly united to the mesonephros by
means of the rete-canals.
The condition in the lizard Platydactylus is, according to Braun, 77,
quite similar. He considers the union to be formed in adult life by two
or three rete-cords joining the anterior ends of mesonephros and testis;
although he states that they are connected along the entire length of the
testis in early stages. Ploffmann, 89, finds the union of rete-cords to be
complete and intact along the entire length of the testis in Lacerta at the
end of the first year. He did not study older specimens. Semon, 87,
claims that there is a degeneration of the rete-cords at both the anterior
and posterior ends of the sex-gland of the chick; but Janosik, go, denies
Material axd Technique.
Our lakes in the vicinity of Madison abound in the little painted
tortoise, Chrysemys marginata. The number of embryos to be gathered
in the season is limited only by ones patience in the work of preserving
them. I have prepared a large number of serial sections of the mesonephros and sex-gland, as well as of entire embryos, comprising an unbroken chain of stages from gastrulation to adult life.
As a fixative, TQllyesnitzky's Bichromate-acetic fluid was almost exclusively used, as it gave most excellent results. Haidenhein's iron-ahnn
hsematoxylin stain proved unsatisfactory for early stages of the embryos
under 7 mm. length. For later stages than this it gave excellent results
and was used almost exclusively. A counter-stain of Congo red was also
employed. The sections were cut at a thickness of 7 fi.
Measurements Avere made of the distance between the cervical bend
and the tail bend (C-T). In the later stages the length of carapace was
also given.
To more clearly understand the origin of the sex-cords, it will be
The Eete-C'ords and 8ex-Cords of Chrysemys
necessary to first understand certain features in the development of the
niesonephros. Reference to these features will be made only in so far as
they concern the subject of this paper. In an early stage of development (C-T. 3.5 mm.), a section through the posterior part of the sexgland fundament shows the mesoblastic somites to be attached to the
lateral plates by the unmodified middle plate (Text Figure E). The
cells of the latter are arranged in two rows, in such a manner as to leave
a line of weakness between, which may be considered as a rudimentary
lumen, connecting the body-cavity on the one hand with the cavity of the
Text Fig. E. Transverse section through the middle of the mesonephros
fundament of an embryo of 3.5 mm. C-T. length.
AO. — Aorta. NO. — Notochord.
EC. — Ectoderm. SO. — Somatopleiire.
MY. — Myotome. 8P. — Splanchnopleure.
N. — Neural canal. WD.— Wolffian duct.
NEP. — Nephrotome.
mesoblastic somite on the other. In the region posterior to this, these
relations become even more marked.
More anteriorly, just behind the interesting region which forms a
transition between the pronephros and mesonephros, the middle plate is
found to be wholly broken away from the mesoblastic somites, and to be
divided by transverse intervals into nephrotomes which occur in the
number of three to four per somite.
I found no evidence of a primary metamerism of these nephrotomes.
So soon as the middle piece appeared to be broken up at all, the number
of nephrotomes here recorded appeared. Special investigation along this
Bennet M. Allen 85
line, however, might show a primary metamerism, from which the above
described condition was derived by further secondary splitting of the
nephrogenous tissue. Each nephrotome becomes vesicular to within a
sliort distance of the peritoneum thus forming the primary Malpighian
corpuscles. The remaining portion of the nephrotome uniting it with
the peritoneum becomes, in later stages, the peritoneal funnel or nephrostome, while the uriniferous tubule arises as an outgrowth from the distal
end of the nephrotome. The mesonephric peritoneal funnels are vestigial
structures from the time of their origin.
In later stages {C-T. 6 mm.), two sharply defined regions of the
mesonephros may be distinguished from one another. In the anterior
part of the sex-gland, only the primary Malpighian corpuscles are
formed. Each is well developed, the glomerular invagination having
already taken place. The 11th to 21st Malpighian corpuscles are connected with the peritoneum by peritoneal funnels (Plate I, Fig. 5), some
of which are much better developed than others, there being great variation among them. In the best developed among them, the end attached
to the peritoneum flares open to form an actual funnel-like mouth, yet
this opening is never continuous with that of the Malpighian corpuscles.
The greater part of the peritoneal funnel is merely a cord of cells. In
some cases even, it has lost its continuity with the capsule of Bowman.
At this stage the first ten Malpighian corpuscles are without peritoneal
Caudad of the 21st Malpighian corpuscle, each nephrotome shows two
or three rudimentary vesicular enlargements. Each enlargement is destined to form a Malpighian corpuscle. The most ventral of these we
shall consider as the primary Malpighian corpuscle. It is still rather
broadly connected with the peritoneum. This place of union we shall
consider as a rudimentary peritoneal funnel, although it has no flaring
In later stages, secondary and tertiary Malpighian corpuscles appear
in the anterior region described above, thus making the total number per
somite approximately equal to that in the posterior region. Eoughly
speaking, from nine to twelve Malpighian corpuscles in all, appear in
each somite.
Eeference to Plate I, Fig. 1, will show certain of the points mentioned
above. Furthermore, one can see an elongated mass of tissue that extends from each peritoneal funnel dorso-mediad and which lies just laterad of the V. renalis revehens (vena cava). This we shall term the funnel-cord. They appear in both the anterior and posterior regions of the
mesonephros as described above and are co-extensive with the sex-gland
86 'I'lic li'i'ti'-('<)i-(ls and Sex-Cords of Chrysemys
I'liiidaiiiciit. in fact (licy ai'c found foi' a slioi't distance anterior to it
Naturally each I'unnel-iord lies ()|)])ositc a ])i'iniarv Malpit^'liiaii eorpnscle,
and likewise to the series of secondary, tertiary, etc., corpuscles formed in
a vci'tical row above it. Each cord is made np of rather loosely arranged
cells that bear a rather close resemblance to the mesenchyme cells. In
fact the nuclei of these cells, the funnel cells, and the cells of the" peritoneum are not to be distinguished from one another. Cytoplasmic differences alone appear and these depend upon the density of the tissue. In
some cases a slight evagination of the capsule of Bowman is found at
the point where it joins the peritoneal funnel. This evagination may
take various forms and in many cases is -wholly absent. Such an appearance may have led to the view held by some authors that these cords
arise as outgrowths from the capsules of Bowman. This view would be
still further justified if the peritoneal funnel were to break away from
the peritoneum at a stage prior to that observed. There can be no
question, however, but that the funnel-cords are outgrowths from the
peritoneal funnels ; in fact their bases are the funnels themselves.
The distal portions of the funnel-cords lie above the vena cava in the
fundament of the adrenal body, contributing the greater part of the tissue
that in later stages constitutes the cortical substance of that gland. Peritoneal ingrowths may also be seen extending dorso-laterad from the
peritoneum at a point near the base of the mesentery to the adrenal fundament. These also contribute to the cortical tissue of the adrenal body.
They are of less regular occurrence than the funnel cords, and in later
stages lose their connection with the peritoneum, although they are easily
distinguishable in the stage of 7 mm. G-T. length.
The sex-gland can be clearly distinguished in the embryo of 6.8 mm.
C-T. length. It extends through six somites, although the last ^ of it
remains in a rudimentary condition. Even in this stage it consists
merely of thickened peritoneum containing scattered primitive sex-cells
The sex-gland develops from a portion of the germinal epithelium lying
between the bases of the funnel-cords and the base of the mesentery. In
the embryo of 6 mm. C-T. length, a few primitive sex-cells were already
beginning to appear in this region. At this time, the V. renalis revehens
(vena cava) lies close above the germinal epithelium which has not yet
begun to thicken to form the sex-cords. In an embrvo of 6.8 mm. C-T.
length the germinal epithelium has sent out masses of cells towards the
Y. renalis revehens, and has at the same time bent outward in such a
manner as to form in transverse section, the periphery of a semi-circle,
the interior of which is occupied by the sex-cords. The tips of the sex
Bennet M. Allen 87
cords remain stationary and almost, if not quite, in contact with the
wall of the V. renalis revehens, while their bases grow peripherally with
the germinal epithelium. Mesenchyme cells between the sex-cords are
few and far between.
At some points, the tips of the sex-cords penetrate to one side or the
other of the Y. renalis revehens, and penetrate to the adrenal fundament
to which they contribute.
Plate I, Fig. 3 shows a wax plate reconstruction of a large part of the
sex-gland of the 7 mm. C-T. stage. In this stage the carapace has just
formed. The prominent funnel-cords afford the most striking feature of
the model. Their bases are attached to the peritoneum at the lateral
boundary of the sex-gland. They extend in a dorso-medial direction.
It will be noticed that each is connected with a primary Malpighian
corpuscle. The other Malpighian corpuscles are not shown in the model.
Mediad of the funnel-cords the peritoneum is greatly thickened, forming numerous irregular elevations and ridges between which are deep
clefts and pits. These thickenings are the sex-cords. They are solid
and their cells show no evidence of a radial arrangement to form a
lumen. The peritoneum is far more cut up than would appear from the
model. Many slight fissures separating adjacent sex-cords do not appear.
In any case many of these rudimentary sex-cords are from the first,
united with the funnel-cords while others anastomose freely with one
another, so that all are either directly or indirectly connected with the
Primitive sex-cells are frequently met with in the germinal epithelium,
as well as in the distal parts of the funnel-cords. Aside from the
scattered primitive sex-cells, these tissues are composed of ordinary peritoneal cells. The cells of the germinal epithelium are so crowded as
to make it stain very deeply. The sex-cords are less dense, their cells
being distinct and having clear, sharp outlines, thus differing from
those of the sex-cords of the pig and rabbit, in which a syncytium is
formed among the pure peritoneal cells. The cells of all but the most
proximal parts of the funnel-cord are elongated in the direction in which
the cords extend. This elongation of the cells is so marked that they
resemble the surrounding mesenchyme save for the fact that their cytoplasm is more dense thasi that of the latter. The cells are so closely
associated that these funnel-cords stand out quite clearly from the surrounding mesenchyme.
The proximal part of each funnel-cord is met by one, or sometimes two,
evaginations from the capsule of Bowman of the adjoining primary
Malpighian corpuscle. These evaginations are very clearly distinguish
88 The Eete-Cords and Sex-Cords of Chrysemys
able in this stage from the tissue of the funnel-cords but are in close
contact with tliem.
In earlier stages the funnel-cords are not even in contact with the
capsules of Bowman, although they lie close to them. In these stages
there are no evaginations from the capsules of Bowman, although a
thickening of the cells of the medio-dorsal portions of them indicates
the general region where these evaginations will take place. In the much
earlier stages described above, 6 mm. C-T., the primary union of Malpighian corpuscle, peritoneal funnel and funnel-cord has already been
described. The later union of Malpighian corpuscle and funnel-cord
is a secondary one, and has nothing to do with the temporary primary
union. The breaking away and reuniting of these elements seems to be a
useless process which I confess I am at a loss to explain. I can merely
describe it. It is, however, a most easily demonstrated fact.
In later stages, the evaginations from the Malpighian corpuscles closely
fuse with the funnel-cords, and are not to be distinguished from them.
As development proceeds, the primary Malpighian corpuscles are often
drawn some distance laterad of the sex-gland, at the same time pulling
the funnel-cords laterad and causing them to stretch. In these cases
each funnel-cord becomes sharply bent at the point where the evagination
from the capsule of Bowman meets it; it is then continued in a dorsomedial direction to the adrenal body. As shown above, each primary
Malpighian corpuscle is connected with the sex-cord by a cord of tissue,
formed by an evagination from the capsule of Bowman plus the basal
portion of a funnel-cord. These strands uniting the mesonephros with
the sex-gland are the rete-cords and constitute the rete-testis or rete-ovarii,
as the case may be. In these later stages the funnel-cords are more
elongated and slender, but far more compact than in the early stages.
Plate I, Fig. 4 shows the rete-cords and the relation that they bear to
the sex-cords and primary Malpighian corpuscles. Here the base of the
funnel-cord lies within the sex-gland and forms one of the sex-cords.
This has been observed in many cases. In very many instances, however,
the funnel-cords lie wholly outside the sex-gland, their bases being still
attached at a greater or less distance from the sex-gland to the peritoneum
covering the mesonephros.
It will be noticed that the two rete-cords shown in this model are
united to one another by a thickening of each in the direction of the
long axis of the sex-gland. This represents a tendency to form a longitudinal canal uniting the rete-cords as in the Amphibia and to a certain
extent in the Elasmobranchs, and in the lizard (Braun, 77). This
Bennet M. Allen " 89
longitudinal canal remains incomplete, however, although it may unite
several rcte-eords in the manner shown.
Young males taken at the time of hatching, show many of the retccords to have already acquired a lumen in places. The rete-cords of
females at this age do not show a lumen, nor do they at any time, because
they have already paused in development. They are, however, still
recognizable. Up to this point no distinction of sex has been noted
although well marked differences had begun to appear in the stage of
13 mm. C-T. length. Close study has yet to be made to determine the
earliest evidences of sex differentiation.
It is not our aim to follow the later development of the rete-cords or
sex-cords. In its general features, the further development of the sexglands of the turtle shows many points of similarity to that in the mammals. The sex-cords degenerate in the female forming the medullary
cords while the " cords of Pfiiiger " arise as a later thickening of the
germinal epithelium. In the males the sex-cords lengthen, assuming a
more regular form and arrangement. Their thorough anastomosis with
one another allows the semen to be poured from several into a common
rete-cord. The mesonephros degenerates leaving a number of the uriniferous tubules to function as vasa efferentia. In the adult male the retecords are found to be reduced in number, there being nine in the specimen
studied while sixteen were counted on the right side of an embryo of
C-T. 8 mm. length. No attempt was made to determine how or when
this reduction was brought about. It is quite probable that some retecords are weak and become broken by shifting of the organs in the
process of growth. In any case there is no systematic degeneration of
the rete-cords in any particular region or regions along the sex-gland.
Summary and Conclusions.
The sex-cords are formed from irregular ingrowths of the germinal
epithelium. It is not until relatively late in development that they take
on the semblance of cords. They are made up of ordinary peritoneal cells,
together with primitive sex-cells which are also found in the peritoneum
at this stage.
The rete-testis and rete-ovarii are formed by the union of funnel-cords
with evaginations from the capsules of Bowman. The funnel-cords are
derived from the peritoneal funnels of the Malpighian corpuscles. They
occupy a region lying along the lateral edge of the sex-gland, and not only
co-extensive with the latter, but extending a short distance anterior to
it. The bases of the funnel-cords may, or may not, be included in the
sex-gland to form a part of the seminiferous tubules of the testis or
90 The Eete-Cords and Sex-Cords of Chrysemys
iiRHluUary cords; of the ovary, as the case may he. The ])roxiinal portions
of these fi)nnel-cords go to form a large part of the rete-testis-ovarii,
while the more distal portions join the adi'onal fundament and contril)nte
the major portion of the cortical suhstance of that organ.
This leads me to briefly consider the adrenal body, although this was
not within the original plan of the present work. Soulie, 02, finds that
in Lacerta and the chick, the cortical substance arises wholly from cords
of cells proliferated from the peritoneum mediad of the sex-gland and
at the base of the mesentery. He states, however, that these cords become
closely applied to the capsule of Bowman of the Malpighian corpuscle.
It is difficult to understand how, arising from the base of the mesentery,
they could reach the Malpighian corpuscle without growing dorsad along
the medial side of the Y. renalis revehens to the adrenal body fundament,
and thence laterad and ventrad to the Malpighian corpuscle. It is difficult to understand how they could take this course, without passing
through and beyond the fundament of the adrenal body. There certainly
are, in the turtle, cords of cells that arise as Soulie and others claim, near
the base of the mesentery, and these contribute to the formation of the
adrenal body; but certain sex-cords and the funnel cords contribute to
it as well, and in even greater measure. Brauer, 02, also holds a view
similar to that of Soulie as regards Hypogeophis one of the Gymnophiona.
Poll, 03, reached similar results with the Elasmobranchs, Acanthias and
Spinax. Be this as it may, I feel quite sure of my ground in the case of
Chrysemys, and the work of Weldon, 85, and Hoffmann, 89, would lend
color to this view, though they hold views in some points radically different from mine.
. In this connection it may be well to state that several of Hoffmann's,
89, figures of the " Sexual Strange " would serve fairly well to represent
the funnel-cords as I have seen them. They certainly do not prove
his contention that the cords in question, sex-cords and adrenal-cords,
arise from the capsule of Bowman ; although he has so interpreted them.
Those who held view C, probably used insufficient material and lacked
the intermediate stages between the period just before the formation of
the sex-cords and those subsequent to their separation from the germinal
Janosik, 85, D, worked upon the chick. It is quite possible that future
work may in large part substantiate his results for that form. My results
agree with his as regards the origin of the sex-cords,- but differ from his
upon the origin of the rete-tissue, although even here there may be a
reconciliation between our views.
In the literature upon the morphological significance of the uro-genital
Bennct M. Allen
system we have some melancholy examples of the futility of making
rash hypotheses unsupported by a sufficient array of facts. Still it is
of interest to consider the possible interpretation that may be placed
upon these structures when they are viewed from the standpoint of
I am inclined to consider tlie funnel-cords as modified sex-cords. The
fact that their distal extremities contribute to the formation of tbe
adrenal bodies does not conflict with this interpretation, because that is
also true of undoubted sex-cords. The funnel-cords arise just laterad of
the true sex-cords and in a very similar manner. The fact that they
arise from the peritoneal funnel would not be contrary to this view if the
Text Fig. F. Diagram to show essential structures of the mammalian
M. — Mesonephros.
MP. — Malpighian corpuscles.
R. — Rete-region.
R. C. — Rete-cord.
S. — Sex-gland region.
8. C— Sex-cord.
y. — Vestigial portion of genital
W.D.— Wolffian duct.
funnels could be shown to be mere recesses of the peritoneum, and similar
to the latter in histological character. A more careful study of the
origin of the sex-glands in the Amphibia is much to be desired as it might
throw new light upon this question. It will be of interest to compare
the results of this paper with those of a previous paper upon the same
structures in the pig and rabbit. Allen, 04.
The very schematic diagram of the testis of the pig (Text Figure F),
shows the following points seen in a sagittal section passing through the
genital ridge and the mesonephros. The genital ridge may be divided
into three regions : (1) rete, (3) sex-gland, (3) rudimentary sex-gland
ridge. The rete-cords arc liomodynamous with the sex-cords, being formed
at the same time and in tlie same luannei" as the latter, and occupying
92 The Eete-Cords and Sex-Cords of Chrysemys
the anterior third of the genital ridge, whose middle portion is occupied
by the sex-gland. As the rete-cords develop, they come in contact with
slight evaginations from the Malpighian corpuscles in that part of the
mesonephros which lies nearest the rete region. They then grow back
to the anterior portion of the sex-gland and at a relatively late ]ierio(l
of development advance along its entire length, giving off numerous
branches (tubuli recti) which fuse with the tips of the seminiferous
tubules. The rete-cords of the mammals are the peritoneal ingrowths of
the anterior part of the genital ridge. Speaking in terms of phylogeny
they are the sex-cords of the anterior part of the sex-gland. The analogous structures of the turtle, the funnel-cords, appear at intervals along
the entire lateral margin of the sex-gland.
It is quite probable that the mammalian sex-gland was derived from
that of some reptilian group and that some now existing groups of reptiles
may show sex-gland conditions from which those of the mammals were
derived. ]S[o existing group is more likely to show mammalian affinities
than that of the Chelonia.
jSTothing exactly corresponding to the funnel-cord has ever been found
in the embryonic development of the mammals. It is true that Aichel,
00, has found that the cortical portion of the adrenal body of the rabbit
(Lepus) arises from funnel-like invaginations of the peritoneum near
the base of the mesentery. He is very positive in his claim that these
are the peritoneal funnels of the mesonephros. Nevertheless, he does not
claim to have followed these funnels back to stages in which they were
actually connected with the Malpighian corpuscles. The rete-tubules that
may have directly united the sex-gland proper along its entire length with
the adjacent Malpighian corpuscles of the mesonephros have disappeared
without leaving a recognized vestige, in any of the mammals thus far
studied. The rete-region of these mammals has been evolved from that
part of the genital ridge which was primitively the anterior part of the
sex-gland in the ancestors of the mammals.^
It is scarcely possible to be more specific as regards the nature of the
rete-region of the mammals. Two assumptions are possible: one, that
^ It will be well to note that in Chrysemys, several funnel-cords occur in
a well-marked region, anterior to the sex-gland, in which the sex-cords remain
vestigial. Upon closer study of some sagittal sections of the sex-gland and
mesonephros of Chrysemys I have been struck with the resemblance that
this region bears to the rete-region of the pig and rabbit as seen in similar
sections. In Chrysemys the funnel-cords of this anterior region together with
those of the sex-gland region are joined to form the central canal. This
shows some points of resemblance to the portions of the rete-cords lying
parallel to the peritoneum anterior to mammalian sex-gland.
Bennet M. Allen ' 93
the sex-cords have disappeared leaving only the funnel-cords, the other,
that the sex-cords which primitively existed in this region have taken
on the character and function of funnel-cords. It is difficult to decide
this question, I can merely say that the latter assumption seems the
more probable one, because often two or more rete-cords can be seen in
a single transverse section to arise from more than one point of the
peritoneum covering the rete ridge. In fact the strongest and most numerous rete-cords arise from the portion of it that lies nearest the mes'entery. This question might be solved with certainty by a study of the
conditions in the Monotremes or even in other less primitive groups of
AiCHEL, Otto, oo. — Vergleichende Entwickelungsgeschichte unci Stammes
geschichte der Nebennieren. Arch. f. mikr. Anat., Bd. LVI, 1900.
Allex, B. M., 04. — The Embryonic Development of the Ovary and Testis of
the Mammals. American Journal of Anat., Vol. Ill, 1904.
Balfour, P. M., 78. — A Monograph of the development of the Elasmobranch
Fishes. Works of F. M. Balfour, 1878.
Braux, M., 77. — Das Urogenitalsystem der einheimischen Reptilien. Arb.
Zool.-zoot. Institut, Wiirzburg, Bd. IV, 1877.
HoFFMAxx, C. K., 89. — Zur Entwickelungsgeschichte der Urogenitalorgane bei
den Reptilien. Zeitschr. f. wiss. Zool., Bd. XLVIII, 1889.
92. — Sur le developpement de I'appareil. uro-genital des oiseaux. Verb.
d. Koninklyke Akademie v. Wetenschappen te Amsterdam. Sectie 2,
Deel I, No. 4, 1892.
Jaxosik, J., 85. — Histologisch embryologische Untersuchungen iiber das Urogenitalsystem. Sitz. Ber. Akad. Wien, 3 Abth., Bd. XCI, 1885.
90. — Bemerkungen iiber die Entwickelung des Genital Systems. Sitz.
Ber. Akad. Wien, 3. Abth., Bd. XCIX, 1890.
Laulaxie. F., 86. — Sur le mode d'evolution et la valeur de I'epithelium germi
natif dans le testicule embryonnaire du Poulet. C. R. Soc. de
Biologie, T. 3, 1886.
Mihalkovics, v., 85. — Untersuchungen iiber die Entwickelung des Harn- und
Geschlechtsapparates der Amnioten. Inter. Monatschr. f. Anat. Hist.,
Bd. II, 1885.
MoLLER, F. v., 99. — Ueber das Urogenitalsystem einiger Schildkrbten. Zeitschr.
f. wiss. Zool., Bd. LXV, 1899.
The most plausible theory is that the rete-region of the mammals has not
been directly derived from a condition like that in Chrysemys; but that the
genital ridges of both have been derived from a type in which the anlage of
the sex-cords was co-extensive with that of the funnel-cords.
To be more exact then, the rete-region of the mammals corresponds to the
anterior end of the sex-gland of the turtle plus the modified region of funnelcords anterior to it.
94 The Rete-Cords iind Sex-Cords of Chrysemys
Peter, K., 04. — Normeiitafel ziir Eutwickelungsgeschichte der Zauiieidechse
(Lacerta muralis). Normentafeln z. Entw. gesch. d. Wirbelthiere,
Heft IV, 1904.
Poll, H., 03. — Die Anlage der Zwischenniere bei den Haifischen. Arch. f.
mikr. Anat., Bd. LXII, 1903.
SCHMIEGELOW, E., 82. — Studien iiber die Entwickelung des Hodens und Neben
hodens. His u. Brawne Archiv f. Anat. u. Physiol., 1882.
Semox, R., 87. — Die indifferente Anlage der Keimdriisen beim Hiihnchen und
ihre Differenzierung zum Hoden. Jena Zeitschr. f. Naturwiss., Bd.
XXI, 1887.
90. — Ueber die morphologische Bedeutung der Urniere in ihrem Ver
haltniss zur Vorniere und Nebenniere und iiber ihre Verbindung mit
dem Genitalsystem. Anat. Anz., Bd. V, 1890.
91. — Studien iiber dem Bauplan des Urogenitalsystems der Wirbel
thiere. Jenaische Zeitschr. f. Naturwiss., Bd. XXVI.
Semper, C, 75. — Das Urogenitalsystem der Plagiostomen. Arb. Zool.-zoot.
Institut, Wiirzburg, Bd. II, 1875.
SovLiE, 04. — Recherches sur le developpement des capsules Surrenales chez
les vertebres Superieurs. Journ. de I'Anat. et de la Phys., T. 39.
Waldeyer, W., 70. — Eierstock und Ei. Leipzig, 1870.
Weldon, W. F. R., 85. — On the Suprarenal Bodies of the Vertebrata. Quart.
Journ. of Micr. Sci., Vol. XXV, 1885.
Ao. — Aorta. P. — Peritoneum.
Art. — Arterial branch passing to P. F. — Peritoneal funnel.
the Malpighian corpuscle. P. C. — Posterior cardinal vein
FC. — Funnel-cord. 8C. — Sex-cord.
M. — Mesentery. Y. R. R. — V. renalis revehens.
M.C. — Malpighian corpuscle. W.D. — Wolffian duct.
Fig. 1. Transverse section of the mesonephros and sex-gland fundament
of an embryo of 6 mm. C-T. length. X 190.
Fig. 2. Transverse section of the sex-gland fundament of an embryo of
7 mm. C-T. length (carapace 5 mm. long). X 190.
Fig. 3. Wax plate reconstruction of the indifferent sex-gland of an embryo of 7 mm. C-T. length (carapace 5 mm. long). This includes as much of
the sex-gland as lies within a little more than two somites. X 190.
Fig. 4. Reconstruction of a small part of the sex-gland of an embryo of
13 mm. C-T. length (carapace 12 mm. long). X 190.
Fig. 5. Drawing of a part of a section adjacent to that shown in Fig. 1.
The proximal portion of the peritoneal funnel is here better shown than in
Fig. 1.
%tt 4 "% '
^ »> J
From the EmbryoJogical Laboratory, Harvard Medical School.
With 8 Text Figures.^
In following the transformations of the subcardinal veins in rabbits,
the writer observed that a portion of those veins seemed to become detached from the venous system, and to be transformed into l3'mphatic
vessels (02, p. 238). This supposition is not identical with the theory
that the lymphatic system is a gland-like outgrowth of venous endothelium, always connected with the veins by means of the lymphatic ducts.
It differs also from the older idea that lymphatic vessels are excavations
in mesenchyma.
In favor of this mesenchymal origin, the work of Sala, 00, is the most
convincing. He observed in the chick that both the posterior lymph heart
and the thoracic duct arose independently of the veins or of other lymphatics, and that their permanent openings into the veins were acquired
subsequently. In the rabbit, as will be shown presently, there are many
disconnected lymphatic spaces, but to their origin from mesenchyma there
are four objections : 1st. The lymjjhatic spaces do not resemble mesenchyma even when it is cedematous, but on the contrary, are scarcely distinguishable from blood-vessels (Langer). 2d. After being formed, the
lymphatics increase like blood-vessels, by means of blind endothelial
sprouts, and not by connecting with intercellular spaces (Langer, Eanvier,
MacCallum, Sabin). 3d. In early embryos, detached blood-vessels may
be seen without proving that blood-vessels are mesenchymal spaces. These
detached vessels are not far from the main trunks, from which they may
have arisen by slender endothelial strands, yet often the connecting
strands cannot be demonstrated. A similar supposition would account
for detached lymphatic vessels. 4th. The endothelium of the embryonic
lymphatics is sometimes seen to be continuous with that of the veins.
^ This investigation, and the one which follows, were accomplished with
the aid of a Bullard Fellowship, established in memory of John Ware.
Ameeican Journal of Anatomy. — Vol. V.
96 The Development of the Lymphatic System in Eabbits
The second thcoiy, that of the giand-like origin of the lymphatic system,
is supported by the remarkable injections of pig embryos, made by
Prof. Sabin.' She considers that in mammals, this system buds from the
venous endothelium at four points, forming four lymphatic ducts. The
ducts are dilated to form four lymph hearts, which, though destitute of
muscles, correspond with the four lymph hearts of amphibia. Starting
from these hearts, lymphatic outgrowths invade the body, and those from
the anterior pair unite with those from the posterior pair. Then the
posterior hearts lose their original openings into the veins, but those of
the anterior hearts persist as the outlets for the thoracic and right lymphatic ducts respectively. The lymph hearts themselves are said to become transformed into lymph nodes (05, p. 355).
According to this idea, the lymphatic vessels are true lymphatics from
their earliest inception. They differ from other branches of the veins
by their very oblique angle of entrance, and by failing to anastomose with
arteries or veins. Anastomoses with other lymphatics are abundant, due
to absorption of contiguous walls (Ranvier, 97, p. 74).
The supposition suggested by the study of the subcardinal veins is
intermediate between those of Sabin and Sala. The endothelium of
the lymphatics is considered to be a derivative of that which lines the
veins, since the lymphatics are at first a part of the venous system; but
by becoming detached from their origins these lymphatics form closed
sacs in the mesenchyma. Later they acquire permanent openings into
the veins, and many connections with other lymphatics.
In studying the development of the lymphatic vessels, several methods
have been employed. Sala used serial sections, generally of injected
embryos, and made wax reconstructions of the posterior hearts. Sabin
perfected the method of injection whicli had been employed by Ranvier
for pigs of 100 mm., so that it was applicable to those of 20 mm. By
this means she studied the large jugular hmpli sacs, or " anterior hearts,"
which, as Saxer discovered (p. 370), are the earliest lymphatic vessels
to appear. On the basis of injections she was enabled to present the
first connected account of the development of the mammalian Ij^mphatic
system. This was illustrated by a series of conventional diagrams, in
which the blood-vessels are shown without details. Thus the internal
^ Ranvier described the interesting analogies, botli functional and embryological, between typical glands and the lymphatic system. Sabin does not
adopt the idea that the whole lymphatic system represents a few large
glands. She does, however, describe it as arising from four blind epithelial
(endothelial) outpocketings which ramify in the connective tissue, and this
origin may be designated, after Ranvier, as " gland-like."
Frcdoi'ic T. Lewis
and t'xtonial juuular veins are merged in an '" anterior cardinal vein,"
the subcardinals are omitted, the renal and iliac anastomoses are made
continuous with one another, and the sciatic and femoral veins are
Fig. 1. Rabbit, 13 days, 9.5 mm.. Harvard Embryological Collection, Series
498, X 13 diams. 3, 4, and 5 indicate the position of the corresponding
cervical nerves in this, as in the following figures. The veins shown are
those of the left side: D. C. duct of Cuvier; Ex. M.. external mammary;
In. J., internal jugular; Pr. VI., primitive ulnar.
It was thought that more accurate figures might bo obtained bv tlie
graphic reconstruction of uninjected embryos. The possibility of overlooking minute orifices guarded by valves, and the limitation of this
method to small embryos are obvious disadvantages, but these are offset
Tlie Dovclojiiiu'iit of tlio Lyniitliatic Svf^tein in Kabbits
l)v the avoidance of rnptnre of very thin-walled vessels and by the opportunity of seeing lymphatics too small for injection. The method has
been employed ^Yith the following results.
Fig. 2. Rabbit, 14 days, 10 mm., H. E. C, Series 155, X 13 diams. The
lymphatic vessels are lieavily shaded, as in all the following figures. The
veins are those of the left side: An. T., anterior tibial; C, caudal: c. b..
" connecting branch "; D. C. duct of Cuvier; Ex. J., external jugular; Ex. M.,
external mammary; In. J., internal jugular; P.O., posterior cardinal; Pr.Fi.,
primitive fibular; Pr. VI., primitive ulnar.
In a rabbit of 13 days, 9.5 mm., no lymphatics could be found. The
reconstruction, Fig. 1, shows the veins along which the first lymphatics are
soon to appear. The internal jugular vein receives a great many small
Frederic T. Lewis 99
branches. One of these, nearly parallel with the dorsal border of the vein
and wider than the others, opens into the vein at either end. It is in
relation with the third cervical nerve. From its position and appearance
it is believed that this branch of the vein becomes a lymphatic vessel.
The second reconstruction is a 10 mm. embryo of 14 days. In this
specimen a chain of lymphatic spaces has appeared along the internal
jugular and the dorsal root of the primitive ulnar veins. The most
anterior segment of the chain extends back to the third cervical nerve. It
sends out short blind sprouts like a vein and contains many blood
corpuscles. The partition between it and the jugular vein is very thin,
and at one point there is a suggestion of communication between the
two, as shown in the figure. No opening into the vein can be demonstrated, however. The second segment of the chain, proceeding posteriorly, extends to the fifth nerve. It equals the internal jugular vein in
diameter, and is closely applied to its wall. Behind the third nerve it
sends a blind diverticulum around the ventral end of the dorsal body
muscles, into the deep subcutaneous tissue of the back. This diverticulum, not matched on the opposite side of the embryo, contains blood
which apparently entered it from rough treatment in preserving the
specimen. The third segment of the chain, between the fifth and sixth
nerves, seems to connect with the root of the ulnar vein. This connection,
however, lies in the plane of section, and a thin intervening wall may have
been carried away in the process of cutting. A detached lymph space
follows the dorsal root of the ulnar vein. A small and somewhat questionable one, not matched on the opposite side, rests against the superior
vena cava, between the roots of the ulnar vein. The most significant
structure found in this embryo is a space filled with blood, which opens
into the external jugular vein near its junction with the internal jugular.
This space lies quite near the third segment of the lymphatic chain. On
the opposite side of this embryo, and in the following one, this blood-filled
sac connecting with the vein appears to . be replaced by a lymphatic
space, detached from the vein, but connecting with the chain.
Fig. 3. from an embryo of 1-1 days, 11 mm., shows the fusion of all the
lymphatics of the previous stage into one large sac which encircles the
external jugular vein. On neither side could this sac be seen to communicate with the veins. No lymphatic vessels were found which did not
connect with the jugular sacs. The dorsal subcutaneous extension, described in the preceding stage, occurred on both sides. In the posterior
part of the embryo, no lymphatics were found. The reconstruction of
the cardinal veins is that already figured in this journal. Vol. I, Plate
2, Fig. 5 (following p. 244).
100 The Development of the Lymphatic System in Rabbits
The cardinal veins of the 14.5 mm. rabbit, Fig. 4, were also shown in
the earlier paper (Plate 3, Fig. 7). In the plate, the lower portions of
the subcardinal veins are detached from the rest, and, though colored blue
Fig. 3. Rabbit, 14 days, 11 mm., X 13 diams. The structures drawn are
the same as in Fig. 2, except that in the trunlt of the embryo the following^
veins, belonging to the median plane and to the right side, have been added:
Az., azygos; G., gastric; R. A., renal anastomosis of the subcardinal veins;
Sc, subcardinal; /S. M., superior mesenteric; V., vitelline; V. G. I., vena cava
Frederic T. Lewis 101
■ like the veins, they are deseribed and figured as " spaces in the mesentery "
suggesting the lymph hearts of the chick (p. 238). It is stated that
these spaces " may be siibcardinal derivatives." Ee-examination of this
embrj'O has yielded no more definite information. The spaces which are
midoubtedly lymphatic, as shown by their later development, seem to replace veins of the preceding stage. In the same way the lymphatic
vessels in the mesentery, accompanying the superior mesenteric ancVthe
gastric veins may have arisen as the branches of those vessels seen in
Fig. 3. They extend around the superior mesenteric artery, which the
corresponding vein accompanies. The fused vitelline vein is destitute of
small branches, and is not provided with lymphatics.
The jugular lymph sac in Fig. -t has completely surrounded the third
and fourth cervical nerves. It envelops two-thirds of the circumference
of the internal jugular vein. On the right side of the embryo, in one
section (No. -176), a miniite orifice connected the sac and the vein.
It was not in the position of the adult opening between these structures,
and was not matched on the opposite side. The deep subcutaneous outgrowth from the jugular sac has become greatly dilated in its distal portion.
Near the beginning of the external mammary vein, a large lymph space
is found wedged between two converging venous branches. This space is
not connected with the veins. It may be a remnant of the lymphatic
vessels which in the preceding stage accompanied the dorsal root of the
ulnar vein. A few slender detached lymphatics follow the external
mammary vein. Finally there are two lymphatics which appear to have
arisen from branches of the azygos vein, one near the vagus nerve (Fig.
4, x) and the other along the aorta (Fig. 4, y). The former connects
with a small vein, the latter ends blindly not far from one. Obviously
when a connection with a vein is well preserved the structure in question
would be considered a venous branch; and after becoming detached, were
it not for its endothelial wall, it might be called a mesenchymal excavation. The study of this and the following specimens seems to show that
the lymphatics along the aorta (thoracic ducts) are derived in part from
the azygos veins; below, from the subcardinals; and above, from the
jugular sacs.
In order to determine whether the lymphatic system of the rablut
differed materially from that of other mammals, reconstructions were
made of a 21 mm. pig, and a 15 mm. cat. The former is of special
interest as a basis of comparison between the present work and > that of
Prof. Sabin. The lymphatics in the pig (Fig. 5) consist of a pair of
jugular lymph sacs, a pair of subcardinal sacs which fuse with one
another irregularly and are variously subdivided by thin septa, and
Fic. 4. Rabbit. 14 days 18 hours, 14.5 mm., H. E^ C Sen^^l^^'^^,/,":
diams. X designates a lymphatic vessel '-^^^^^P^!^^"^?^ f ^/^ Jf ceiSc;
.,. a lymphati^c along ^^e aorta. The veans of th -^^,^%^,_ ,;
Pr. Ul.. primitive ulnar. Those ot tne leg axe. ^-i". ^ ••
"connecting branch"; Fe., femoral; Pr. Fu, primitive fibular.
Frederic T. Lewis ' 103
finally some irregular spaces behind the aorta, probably derived from the
azygos veins. These spaces also fnse across the median line at several
The jngnlar sac is shaped like a D of which the chief portion is vertical
and closely applied to the internal jugular vein. Through the aperture
in the D pass the third, fourth, and fifth cervical nerves, and from its
dorsal arch several deeply subcutaneous sprouts pass off, corresponding
with the single large sac of the rabbit. ISTo connection between the
jugular sac and the veins could be detected. Except for this point, the
reconstruction agrees with, and combines, the figure and diagram presented by Prof. Sabin in this journal. Vol. 3, p. 184, and Vol. -1, p. 359.
It does not agree so well with the diagram on p. 380 of A^ol. I. In the
latter the subcardinal l3anph spaces are not shown. The posterior portion
of the body contains instead two " lymph hearts " arising from tlie posterior cardinal veins " below the Wolffian body "" but anterior to the
femoral vein. In later stages, outgrowths from these hearts invade the
skin of the back, and nltimately, as has already been noted. Prof. Sabin
considers that the hearts become transformed into lymph nodes. From
this description, it appears that the posterior lymph hearts are in the
position of the ilio-lumbar veins. In the pig embryo represented in Fig.
5, however, no lymphatics were found in relation with the ilio-lumbar
Considering its lymphatic development the pig of 21 mm. is less advanced than the rabbit of 14.5 mm., since there are no lymphatic vessels
along the external mammary vein nor in the mesentery. The cat of
15 mm. is more advanced than either. In this embryo the D formed by
the jugular sac is almost bisected diagonally. The second, third, and
fourth nerves pass through its aperture, but the fifth penetrates the
posterior section of the sac by a separate opening. There are two deep
subcutaneous diverticula corresponding with the single one in the rabbit
and several in the pig. In one section (266) a branch of the jugular
sac may enter the innominate vein a little anterior to the subclavian, but
it is not clear that an actual opening exists and none can be found on
the opposite side.
"Where the external mammary vein joins the brachial there is a large
sac, and the question arises M'hether or not the detached lymphatics following the mammai-y vein are independent formations, or are outgrowths
from that sac. The occurrence of the lymphatics especially near the
places where the veins l)ranch suggests that they may have budded at such
points. On the other hand, as in the rabbit, their order of appearance
is from the proximal ])art of the vein distally. Similarly there are
vf w r Series 59, X 10 diams. The veins are:
Fig. 5. Pig, 20 mm.. H. E. C., series oy, ^ Cuvier; Ea-. J., exter
A^., azygos; Br., brachial: Ce cephalic D. ^^^^^^ f ^^'^^ _ ^^stric; In. J.,
nal' jugular: i... M -^--^^ ^.^^^rn'osifof ^b'ardinals- ..n.. sciatic;
;rS!tp^e^S?;ies?nte^riV;?^^viSlfne; V. C. I. vena cava inferior.
Fig. 6. Cat. 15 mm., H. E. C, Series 436, X 13 diams. The lettering Is
the same as in Fig. 5.
10() The Development of the Lymphatic System in Eabbits
obscure spaces which ap})ear to ho lymphatic, along the aorta, and in relation with the azygos veins. An occasional apparent connection with
the vein suggests their venous origin in situ. The mesenteric and subcardinal plexuses have united with one another. They do not empty into
the veins. The subcardinal sacs extend from the renal anastomosis
almost to the sciatic vein, connecting with one another across the median
line, as in the pig. No lymphatic vessels follow tlie ilio-lumbar veins
into the posterior body wall.
Eeturning to the rabbit embryos it will be seen that Fig. 7 from a
21 mm. rabbit differs from Fig. 4, the 14.5 mm. embryo, chiefly in regard
to the thoracic duct. The duct is represented by a pair of vessels which
connect with one another and pass on to the left jugular sac. Sometimes
in the adult rabbit, as figured by Gage (02, p. 650), and occasionally in
man, the thoracic duct bifurcates anteriorly and passes to the jugular
sacs on either side. This did not occur in the 21 mm. embryo, which
exhibited the relations figured by Sabin, Vol. I, p. 383.
In Fig. 7 scattered lymphatics are shown along the external jugular
vein and its branches. One much larger than the rest occurs wdiere the
anterior and posterior facial veins unite. From its isolation it probably
arose independently of the large jugular sac. Other and more isolated
lymphatic centers are seen in the oldest rabbit studied, one of 20 days,
29 mm., Fig. 8, notably along the pudic and the sciatic veins. They arise
near the junction of several venous branches, with which, however, they
are not in communication.
In the oldest embryo the lymphatic system has invaded the skin to such
an extent that it is impracticable to represent more than a small part of
it. In entering the skin the lymph vessels accompany the veins, those
of the head following chiefly the external jugular vein. The jugular sac
has become relatively less important, and persists as the lymphatic sheath
of the internal jugular vein. The deep subcutaneous extension has
becom^ covered by a thin layer of muscle, presumably the panni cuius,
and does not appear to connect with the more superficial vessels of the
skin. There are no lymphatics in the distal part of the arm, Init the subcutaneous vessels of the shoulder are attended by rich netw^orks. These
veins are the external mammary, and another w^hich is ventral to the
scapula and posterior to the shoulder joint, — a subscapular vein. The
lymphatics along this large subscapular vein do not connect with the
jugular sac. At the point L. N., indicated in the figure, a small but
very distinct lymph node has developed in relation to these subscapular
lymphatics. A corresponding node is found on the opposite side of the
Fig. 7. Rabbit, 17 clays, 21 mm., H. E. C, Series 738, X 10 diams. The
veirs not previously lettered in the rabbit figures are: 7?., ilio-lumbar; Ss.,
subscapular; R., radial; Sci., sciatic.
i^Z u ^^^'Y- ^^ ^f^^' ^^ "'"'•• ^- ^- C- Series 170, X 6.9 diams. The
first lymph nodes develop at L. lY., along the subscapular vein. ^^s. ; and at
J. «., along the iho-lumbar vein, II. The veins of the arm are: Br., brachialte cephalic: J. Ce.. jugulo-cephalic: R.. radial. Those of the legs are- A« T
anterior tibial; Sci., sciatic: Po. T.. posterior tibial: Fe.. femoVal; c. &., connecting branch between femoral and sciatic. P. marks the pudic vein
Frederic T. Lewis 109
The jugular sac on the leCt side, except for an extensive rupture, does
not connect with the vein. On the right, a pore is found leading from
the sac to the internal jugular vein near its union with the external, but
this also may he artificial. Thus in all the series of rabbits no bilateral
communication of the lymphatics and veins, in the position of the adult
openings, could be found. The pores, sometimes detected in various
positions, are not adequate to empty the large sacs, and may indeed be
artifacts. Communication with the veins in these stages must be by
osmosis, therefore, and the permanent outlets of the lymphatic system
must develop later.
The left jugular sac in Fig. 8 connects with the thoracic duct, which
arises from a plexus of lymphatics surrounding the aorta. Ventral to the
aorta these vessels receive the lymphatics from the mesentery. There are
none in the leg. The body wall is supplied by those which follow the
external mammary A'ein in its anastomosis with the superficial epigastric,
and by vessels accompanying the ilio-lumbar vein. The ilio-luml)ar vein
of Krause, whicli Hochstetter named the posterior transverse lumbar,
supplies the subcutaneous tissue of the back, and seems to be inversely
homologous with the much larger subscapular vein. At the position
I. n., indicated in Fig. 8, a node is found among the lymphatics accompanying this vein. A similar node exists on the opposite side, and the pair
was identified in a duplicate series of a 20-day rabbit. These superior
inguinal nodes (Krause) develop almost simultaneously with the subscapular nodes already described. The early appearance of the inguinal
nodes further identifies the lymphatics of the ilio-lumbar vein with the
" posterior lymph heart " of Prof. Sabiu. It is my opinion that an
identification of this structure with the amphibian or avian lymph heart
is, at present, not justified. The posterior heart of the bird empties into
the coccygeal veins (Sala), and that of the frog into the transveree iliac
vein, a vessel connecting the femoral with the sciatic vein (Gaupp). The
ilio-lumbar vein is more anterior than either. Its lymphatics do not differ
in form, from those accompanying other veins, and they are presumably
non-contractile. If the first lymph nodes can be utilized in making comparisons, then this " posterior heart " of the rabbit should be compared
with the lymphatics of the subscapular vein, and not with the jugular sac.
The jugular sac itself does not empty into the vertebral vein, like the
anterior heart of the frog. It is non-contractile, so far as known. 1 f it
shall l)e found that the anterior heart of the frog develops from the first
]ym})hatics which are formed in that animal, a comparison between the
jugular sac and a lymph heart may be possible. At present it is not
evident that mammals possess any lymph hearts.
110 'I'lio Dovi'lopinoiit of tlic Lympliatic Systom in liabbits
The lymphatic system of rabbits begins along the internal jugular vein
as a detached sac formed by the coalescence of several venous outgrowths.
Similar though smaller sacs arise from the subcardinal and mesenteric
veins at a slightly later date.
Subsequently lymphatic vessels develop along the courses of the azygos
and cutaneous veins, apparently from independent venous outgrowths.
All of these vessels unite with one another to form a continuous system,
which acquires new and permanent openings into the veins near the
subclavian termination.
The first lymph nodes observed are two pairs, one beside the subscapular vessels, and the other beside the ilio-lumbar vessels.
In order to facilitate comparison with Prof. Sabin's work, the following
conclusions may be added :
The lymphatic system does not arise from the venous system by four
outgrowths, but by several. It is not always in communication with the
veins. The outlets of the thoracic and right lymphatic ducts are not persistent primary openings. An identification of mammalian lymph hearts,
comparable with those of the amphibia, should not be made, on the evidence now available. Judged by their relation to the early lymph nodes,
the jugular sac is not comparable with the lymphatics along the iliolumbar vein. However, the study of rabbit embryos confirms the chief
conclusion established by Prof. Sabin, that the lymphatic system is a
derivative of the venous system.
Gage, Simon H., 02. — A Reference Handbook of the Medical Sciences. Edited
by Albert H. Buck. 2d ed., Vol. 5, pp. 624-659, New York.
Gaupp, Ernst, gg. — Anatomic des Frosches. 2d ed.. Part 2, Braunschweig.
HocHSTETTER, FERDINAND, g^. — Beitragc zur Entwicklungsgeschichte des
Venensystems der Amnioten, III. Morph. Jahrb., Vol. 20, pp. 543-648.
Krause, W., 68. — Die Anatomic des Kaninchens. Leipzig.
Langeb, C, 68. — Ueber das Lymphgefasssystems des Frosches, III. Sitz.-Ber.
d. Akad. d. Wiss., Wien, Vol. 58, pp. 198-210.
Lewis, Frederic T., 02, — The development of the vena cava inferior. Amer.
Journ. of Anat. Vol, 1, pp. 229-244.
MacCallum, W. G., 02. — Die Beziehung der Lymphgefasse zum Bindegewebe.
Arch. f. Anat. u. Phys., Anat. Abth., pp. 273-291.
Ranvier, L., gj. — Morphologic et developpement des vaisseaux lymphatiques
Chez les mammiferes. Arch. d'Anat. mic. Vol. I, pp. 69-81.
Sabin, Florence R., 02. — On the origin of the lymphatic system .from the
veins and the development of the lymph hearts and thoracic duct in
the pig. Amer. Journ. of Anat., Vol. 1, pp. 367-389.
Frederic T. Lewis 111
Saiun, Florence R., 04. — Onthe development of the superficial lymphatics in
the skin of the pig. Amer. Journ. of Anat., Vol. 3, pp. 183-195.
05. — The development of the lymphatic nodes in the pig, and
theii- relation to the lymph hearts. Amer. Journ. of Anat., Vol. 4,
pp. 355-389.
Sala, LuKii, 00. — Sullo svillupo dei cuori linfatici e dei dotti toracici nell'
embrione di polio. Ric. fatte nel hab. di Anat. norm. d. R. Univ.
di Roma, Vol. 7, pp. 263-296.
Saxer, Fr., 96. — Ueber die Entwickelung und den Bau der normalen Lymph
driisen und die Entstehung der roten und weissen Blutkorperchen.
Anat. Hefte, Abt. 1, Vol. 6, pp. 349-532.
From the Emhryological Laboratory, Harvard Medical School.
With 1 Text Figure.
In connection with the preceding study of the lymphatic system it was
necessary to reconstruct the veins of the shoulder and hip in a series of
rabbit embrj^os. The reconstructions were then extended to include the
distal portions of these vessels, complete figures of which had never been
published. Hochstetter, in 1891, had observed the veins in the limbs of
living rabbit embryos, and had studied them in serial sections. His
drawings, however, show only detached portions of the veins such as
could be seen under most favorable conditions, in living embryos. Ten
years later Grosser described but did not reconstruct, the developing veins
in the extremities of bats. To these two investigators embryology is indebted for the present knowledge of the veins in mammalian limbs. It
is proposed to review their work, while describing the reconstructions,
considering first the veins of the anterior extremity, then those of the
posterior extremity, and finally the homologies which exist between the
two sets.
In the youngest rabbit figured, an embryo of 13 days. Fig. 1, p. 97,
the small vessels along the radial or anterior border of the arm unite to
form a vein which follows the periphery of the limb to its posterior or
ulnar border, and then ascends behind the brachial plexus to terminate
near the junction of the anterior and posterior cardinal veins. It receives
a branch which at this stage is not well defined, ascending in the body
wall. This is the Seitenrumpfvene of Hochstetter, and becomes the <