Rasta iet
reeihenyy
Hives itches fare pet
+
TTSGEET
: s onsite
oth ; paepe Sataaesihoe ace:
‘i peStior ; f r " ? renin rye) me es : \
aousessecasrsont nienaienicetensnatbdteprandsoaece he Memmesr ns Catnacc aba denendy : “oth 12 alt WF : : s
Tynendensteaane ise Meds akie EL aSATEADoee MIRC Pe Mone aOr eaacaceaacash: kas rderurndarn cs S b 4 chek 5 pees NN
FELISLORC MLA arthur tee tab ot lara satty t : f REEL fergerasy er rscnee 5 >,
. : te Arse ge EAR %
x
\
~
*
rita : i ; : . , Sera Wy ¥
esdénsentses ; 2 : : ; : epeae , ~
SAsaes
: ~} Sa \
etipass rectaers, :
itecdarcnase ates ar aint ck des
oy
2
"6 +
NR
TEINS ; ay Ms ; : vatt ; i . : a
nncunend sen ae " 5 ? + . aS
it f cet Sr : pees eatars ; ‘ recreth é
Crem Sete SEER ERE NTSAT LN : ; i : Y i 5 ; : { bath
; tevrcercernriereet ns ‘ meth
rare
reas
eoreresrecretcess
et fry ata sa nices see N ep seneaan saa teeta
AEE BES :
Capretecieepecist
coarenernscntinaecdgen set3773 : ? mS steerertantee
nds Risen esosenaens Scien tec csgrsec ards ¥ ; f y % . : : ras Salter etrete. 6. pepe
perrceteprrsisrieenererrer ete brite, ¢ wa3 z ; Tae : ? Spy : ; ; SE eee Tras
pecenirnctin rs OF. TENT T * qsras g 5 ¢ < : 2 j ag VHS
THREEILEN LE TREC aseagcieataess 3 ; renee ¢ i zr : pcccntrr yy
adasesddenekste : 5 ; = errr are
renee
PaCEECERRE RITA ree ret eee "
Porinamet ent Se t3anentond Sige a ets eat oa ee : : etYs ’ esrsherseeet
errereneretrrcreceerretiet iti niteetkepe ste ~ ; Enaceaenasceareariens C : c ? : wae
a Saree lessee sels ret i : Sencanee 3 ,
spr
Ssscerss sessasaagnarssnisia Carey: : : * roesrerrspratentcs
erasiesntsan ; . ;
yeast arate prerecriarens : es ‘3
‘ ¥ f fs z f « STiSlsLine Ried
nest eastsiesise cigs : Z - "i 703
preree
perctreareaperere tpt
SHNLOT LC
cs
: =r
aaah oe
oot eee : rere seeaty ye Cl Peete
iatiarte chon
oe
ecxattap areecteet
. vi : LIME CET
epee E 7 e
Chars
PROS
SR
Seesate = aang, 5 ER RTEEINET
~.
a
Fiat Poe ae
i aes
a; "
Pie es
: nates
ata
auto
THE AMERICAN JOURNAL
CHARLES R. BARDEEN
University of Wisconsin
Henry H. Donaupson
The Wistar Institute
Srmon H. Gage
Cornell University
OF
HD DTOR lA L BOARD
G. Cart HUBER
University of Michigan
GrorGge 8. HunTINGTON
Columbia Untversity
Henry McE. KNower,
Secretary
University of Cincinnati
VOLUME 20
1916
ANATOMY
FRANKLIN P. Mau
Johns Hopkins University
J. Puayrain McMorricu
University of Toronto
GrorGE A. PIERSOL
University of Pennsylvania
THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY
PHILADELPHIA, PA,
tay
Aare
hi ‘
one _
i
(
be,
A
n
Pr
ela
‘ 9 &
I st
=
a ‘
we
we i
vn
\
> ‘
i
ui ‘
*)
.
@
* a ee
, i
os ,
ae
& :
: * i —~
4 ;
,
COMPOSED AND PRINTED AT THE
WAVERLY PRESS
By rap WILLIAMS & WILKINS Som1
Batimore, Mp., U.S. A
‘ bs
i
i
=
i
f
‘ ‘, ie
: ae? ill
CONTENTS
Nov to ULY
Wiuuram A. Locy anp Otor Larsety. The embryology of the bird’s lung. Based on
observations of the domestic fowl. Part II. Twenty-two figures.................. 1
H. von W. Scuutte. The fusion of the cardiac anlages and the formation of the cardiac
loop in: the ;cat) (elis-domestica): (Sixteen figures... 60.2.5. 0200 4c. ose cheese cee 45
VY. E. Emmet. Concerning certain cellular elements in the coelomic cavities and mesen-
chyma of the mammalian embryo. Four plates (forty-four figures)................ 73
J. Parsons ScHAmFFER. The genesis, development, and adult anatomy of the naso-
froOMialereeiOneimMemany euehinbeent MO UneSsukyerys ce ce era aa here ete ies see cn 125
No. 2. SEPTEMBER
Joan SunpDwALL. The lachrymal gland. Twenty figures................:.0....0002- 147
W. J. M. Scorr. Experimental mitochondrial changes in the pancreas in phosphorus
pononine. psevenwigures (one tplate))<.4. 43ers eee Mao alia Gy row ele aie lll 237
No. 3. NOVEMBER
VerA Dancuaxorr. Equivalence of different hematopoietic anlages (by method of
stimulation of their stem cells). I. Spleen. Two text figures and nine plates..... 255
NorMAN Cuive Nicuotson. Morphological and microchemical variations in mitochon-
dria in the nerve cells of the central nervous systems. Two plates............... 329
C. V. Morrizt. On the development of the atrial septum and the valvular apparatus
in the right atrium of the pig embryo, with a note on the fenestration of the anterior
CALCING Svein SemNTN Cot CURCS Meta. eee scr A biep eee ease ea cere ae eee ce ee OT
Cuarues H. Swirr. Origin of the sex-cords and definitive spermatogonia in the male
bee phe, HULSE nae eae aera AERIS, TESS gee ail wae NE NAA MR em et! 2PG Se ee 375
Epear H. Norris. The morphogenesis of the follicles in the human thyroid gland.
DEMETUCEDMNOUPCH a5 eine NN AY ct hoary | AO UR Renes a ONO ite ue le meu Ee 411
Wo. E. Kewuicorr. The effects of low temperature upon the development of Fundulus.. 449
Pau E. Linepack. The development of the spiral coil in the large intestine of the pig.
USHER TTB OR SIC ets ee a 2 ee a Pa a SRR a Ss Ae Se 483
THE EMBRYOLOGY OF THE BIRD’S LUNG
BASED ON OBSERVATIONS OF THE DOMESTIC FOWL
WILLIAM A. LOCY AND OLOF LARSELL
TWENTY-TWO FIGURES
PARE Ee
3. THE AIR-SACS AND THE RECURRENT BRONCHI
Morphologically considered, the air-sacs and recurrent bron-
chi are parts of the bronchial tree, but on account of their impor-
tance in the avian lung and their unusual interest they are sepa-
rately considered in this section. This plan also promotes
clearness of description, since, at best, the bronchial tree is very
complex. The recurrent bronchi, in particular, should receive
special notice, because they have been recently recognized and are
of capital importance in the physiological anatomy of the lungs.
The name ‘recurrent bronchi’ has been given to certain bron-
chial tubes that grow from the air-sacs into the lungs of birds to
connect with the other air passages. In this sense they are
‘recurrent.’ They are outgrowths from the air-sacs, rather than
extensions of the bronchial tree from within the lung, and the
alr-sacs and recurrent bronchi are so intimately related in their
development that the two structures should be considered to-
gether. In the course of development they unite with twigs of
the bronchial tree and thus establish complete circuits with the
air passages within the lungs. In the adult lung the air passes
from the air-sacs through these recurrent bronchi, entering the
lung by a returning current, and, in this sense, the air circuit
through these bronchi is a recurrent one.
The credit for the recognition of the morphological arrange-
ment as well as for the part which recurrent bronchi play in the
respiration of birds should be divided between Schulze (’09 and
‘Part I of this paper appeared in the American Journal of Anatomy, vol. 19,
no. 3, May, 1916.
1
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 1
JuLy, 1916
2 WILLIAM A. LOCY AND OLOF LARSELL
10) and Juillet (12) who, working independently, and without
knowledge of each others observations, grasped the essential
features of these very important and characteristic structures of
the avian lung. Although they were figured by earlier observers
(Campana, 775, Fischer, ’05), Schulze was the first to observe
carefully their arrangement and relation to the air-sacs and to
the other bronchi in a number of different birds, and to appreci-
ate their physiological réle, while Juillet added some morpho-
logical facts and made observations on their development.
Review of the extensive literature on the air-sacs seems to us
unnecessary, since the bulk of it relates to their position, size and
anatomical relationships in the adult. In this respect the papers
of Campana, ’75, Bruno Miller, 707, and Schulze, ’11, are espe-
cially good. Asregards their development, except for the paper
of Bertilli, little has been added to the embryology of the air-
sacs since Selenka’s paper of 1866 in which he described and fig-
ured their development in the chick.
The recurrent bronchi, however, have come into notice more
recently anda brief account of the published observations on
these structures should be given.
Campana (’75) in his extensive memoir dealing with the res-
piratory apparatus of birds and confined chiefly to a description
of the adult structures, described the air-saes and bronchial tree.
He also made a careful analysis of the orifices connecting lung
and air-sacs. Examination of his figures shows that some of the
recurrent bronchi also stand out quite distinetly, but Campana
considered them as the result of a reconstitution into a single
trunk of several tertiary bronchi, without recognizing their true
nature. He apparently considered them as merely a part of
the network of air passages with no special significance attached
to them.
On the other hand Campana used the term ‘Bronche recurrente’
in an entirely different connection, applying it to the curvilinear
branch of the first entobranchus. This statement should be
made to prevent confusion.
Guido Fischer (’05) likewise figures the recurrent bronchi of
several of the air-sacs. The only reference he makes to these
THE EMBRYOLOGY OF THE BIRD’S LUNG 3
features of his celloidin corrosion preparations, however, is in a
note of explanation of one of his figures in which he calls atten-
tion to a bronchial trunk larger than the others in the network
of air-tubes which extends to the dorsal surface of the lung on
its lateral side. This bronchial trunk, he says, ‘directs itself
toward’ the abdominal air-sac, but since it is the nature of re-
eurrent bronchi to grow inward from the air-sacs, the language
employed by Fischer shows that he had a wrong conception of
these important structures.
F. E. Schulze in 1909, 710 and ’11 recognized these bronchi as
coming from the air-sacs and designated them both ‘Riichliu-
figen Bronchen’ and ‘Bronchi recurrentes sue Saccobronchi.’
With sketches he describes their origin from basal pockets on
the four posterior air-sacs, variations in the number of their
branches as well as the nature oftheir connections with para-
bronchi. His comparative observations embraced a variety of
birds including the chick, duck, goose, pigeon, Rhea, ostrich,
Cassowary, etc. In the Cassowary he noted that recurrent
bronchi are lacking on the abdominal air-sac. Schulze also points
out that the recurrent bronchi carry air from the air-sacs into the
lung parenchyma and play an important part in respiration.
Juillet (12) made an extensive study of the recurrent bronchi,
and since he was unacquainted with the observations of Schulze
he claims rank as the discoverer of the true anatomical relations
of these structures and of the part they play in the respiration of
birds. In all this however he was preceded by Schulze, and to a
limited extent he engaged in the study of the embryology of the
recurrent bronchi, which was not touched upon by Schulze.
He found recurrent bronchi in all the twenty-four species of
birds which he examined. By a study of sections he traced some
stages of their development in the embryonic lungs of the chick,
and although he does not give an extended account of their de-
velopmental history, he arrived at a true conception of their ori-
gin and of their nature.
No more important advance in the knowledge of the avian
lung has been made since William Harvey, in 1651, discovered
the perforations of the bronchi into the air-saes and found them
4 WILLIAM A. LOCY AND OLOF LARSELL
“sufficiently conspicuous in the ostrich to admit the points of
my fingers.”’
It will be advantageous to describe the development of air-
sacs and recurrent bronchi by stages beginning with the seventh
day.
The seventh day stage. There are five air-sacs in the lung of
the adult fowl, and as will be shown later, one of these (the in-
terclavicular) is the result of the fusion of four moieties, two from
each lung, that arise independently. The names employed in
the following descriptions are: cervical, interclavicular, anterior
intermediate, posterior intermediate, and abdominal air-sacs.
All the air-saes, except the interclavicular of the adult, are paired.
The cervical and interclavicular arise anteriorly, the other three
upon the ventral and caudal surface of the lung.
The youngest embryo in which any of the air-sacs appear as
projections beyond the lung wall are of about six days six hours
incubation. As shown in figure 30, the abdominal air-sac of this
stage projects as an extension from the lung proper. The pri-
mordium of this sac is the slightly expanded distal portion of the
mesobronchus lying beyond the bend of the central lung tube.
In the same embryo may be seen the first indication of the
cervical air-sac in the form of a bud projecting from the distal
extremity of the first entobronchus. In its later development the
entobronchus becomes much branched, and the orifice of the
air-sac is not terminal, as in the embryo. but on the body of the
cranial branch of the entobronchus.
The anterior intermediate air-sac, with the mesial moiety of the
interclavicular united to it, is also foreshadowed in this specimen
as a bud of the third entobronchus. The third entobronchus
shows at this stage. The beginning of an unequal bifurcation
which shortly (figs. 834 and 37) becomes well differentiated. The
more caudad, and longer, branch of the bifurcation develops into
the foliate division of the entobronchus, and the forward project-
ing bud becomes eventually differentiated into the anterior in-
termediate air-sac and the mesial moiety of the interclavicular
sac. To avoid confusion, one should constantly keep in mind
that the interclavicular air-sac arises from two moieties on each
THE EMBRYOLOGY OF THE BIRD’S LUNG 9)
side, and in subsequent references we should follow with care the
development of a mesial and of a lateral moiety from different
sources.
The ninth day stage. In the interval between the seventh and
the ninth day the entire bronchial tree grows rapidly and the
air-sacs enlarge.
Early on the ninth day of incubation carefully prepared air in-
jections show important advances. ‘The primordia of all five air-
sacs now project beyond the lung surface.
The cervical sac (fig. 36, Cerv.sc.) is the forward prolongation
of the cranial lobe of the first entobronchus and is little changed
from the former stage.
In the meantime the first entobronchus has divided into sev-
eral branches and from one of them (the transverse branch) may
now be seen the beginning of the lateral moiety (Lat. mov.) of the
interclavicular sac. At this stage it is small and does not pro-
ject beyond the lung wall. As shown in subsequent develop-
ment this lateral moiety fuses with the mesial moiety to form a
part of the interclavicular sac of the adult.
The mesial moiety is well developed at this stage. It arises
on an anterior branch of the third entobronchus. This branch
bifureates early on the seventh day of development (not figured).
The smaller, and more cephalad, division becomes the mesial
moiety of the interclavicular air-sac, the larger, and more cau-
dad, division the anterior intermediate air-sac. As shown in fig-
ure 37, the mesial moiety, although very slender on the ninth
day, is nevertheless sufficiently elongated to project beyond the
lung wall.
Exceptionally the mesial moiety arises on a branch of the sec-
ond entobronchus, in which ease the third entobronchus gives
origin only to the anterior intermediate sac. This condition is
illustrated in figure 38 which represents a slightly earlier stage
than the one sketched in figure 37.
When the development follows the usual rule the mesial moiety
of the interclavicular sac is an offshoot of the anterior intermedi-
ate air-sac and the two are connected with the third entobron-
chus by a single orifice (the interclavicular canal).
6 WILLIAM A. LOCY AND OLOF LARSELL
By unequal growth the anterior intermediate air-sac (fig. 37)
has increased relatively much faster than the mesial moiety of
the interclavicular and forms, at this stage, a prominent landmark
on the ventro-mesial part of the lung. It remains until the
eleventh day of incubation the largest of the embryonic air-sacs.
Extending forward from its ventral anterior part may be seen
three small papilla-like buds (Rec.br.) connected with the sac by
a short stem. These buds are the beginnings of the recurrent
bronchi of the anterior intermediate air-sac. They make their
first appearance (not figured) as a single bud during the latter
part of the seventh day of incubation, and by division of the
distal end of this bud the three papille are formed. The proxi-
mal end remains as the stem and probably forms the basal pocket
of Schulze.
The posterior intermediate air-sac (fig. 36, P.int.sc.) also makes
its first appearance as a projection beyong the lung wall on the
ninth day. It is the distal continuation of the third laterobron-
chus, and at this stage is but slightly distended and shows no
indication of recurrent bronchi.
The abdominal air-sac, is on the ninth day of incubation greatly
elongated. From its anterior end and point of union with the
mesobronchus a pouch is developed which represents the begin-
ning of recurrent bronchi of this sac. The distal end of the sac
is but slightly more inflated than it was on the seventh day, but
about two-thirds of it now project beyond the lung proper.
From the position of the pouch of the recurrent bronchi one
would infer that the anterior limit of the abdominal air-sac of the
embryo is more cephalad than is usually recognized. This also
changes our idea of the position of the morphological tip of the
mesobronchus.
The tenth day stage. From the beginning of the ninth day of
incubation to the close of the tenth there is a steady growth of
the various air-sacs and of their recurrent bronchi. It is not
necessary to follow in detail the various steps between the two
stages, since a description of the conditions found in the later
stage will sufficiently indicate the changes through which the
various structures have passed.
THE EMBRYOLOGY OF THE BIRD’S LUNG
~
In an embryo of 93 days incubation (figs. 839 and 40) the cer-
vical sac shows only an increase in size proportioned to the gen-
eral growth of the lung.
The lateral moiety of the interclavicular sae (fig. 40, Lat.moz.)
now shows a well defined extension outside the lung wall. The
subsequent history of this sac indicates that some of the intra-
pulmonary part of the transverse branch of the first entobron-
chus, from which the sac arises, must be considered as a part of
the air-sac primordium. ‘The line of separation betwen the ento-
bronchus and the lateral moiety is where the more dorsal buds of
recurrent bronchi arise. Two groups of recurrent bronchi be-
longing to this division of the interclavicular have appeared.
The bud which represents the more ventral group of these bronchi
extends caudally and ventrally and is as yet undivided. The
more dorsal group is represented by an already bifurcated bud
projecting caudally and somewhat mesially. These buds develop
into the only recurrent bronchi arising from the interclavicular
air-sac and it is to be noted that they arise only on its lateral
moiety.
The mesial moiety of this stage, a part of which is shown in
figure 40, is scarcely changed from the condition described in the
eight day embryo, except that it has increased in size.
The anterior intermediate air-sac (A.7nt.sc.) at the close of the
tenth day of incubation has enlarged considerably, as compared
with the preceding stage. The recurrent bronchi have not
greatly changed, but the stem thereof has elongated to some
extent.
The posterior intermediate sac now projects beyond the lung,
and its distal end forms a flask-like swelling. The proximal part
remains still constricted and lies within the lung. From this
part two buds are seen projecting dorsad and cephalad. The
more anterior of these buds is already divided at its tip. These
branches indicate the beginnings of the recurrent bronchi of the
posterior intermediate air-sac.
The abdominal sac (figs. 39 and 40) has expanded greatly since
the eighth day stage, and now lies almost entirely outside the
lung. Its recurrent bronchi, of which there are two sets, have
8 WILLIAM A. LOCY AND OLOF LARSELL
also made an obvious growth. The bud (fig. 39) previously de-
scribed has divided into two main branches each of which has in
turn bifurcated, as represented in the figure.
A second group of recurrent bronchi, also belonging to the ab-
dominal sac, begins early in the tenth day of incubation. The
<=,
= ------ Tra
= eee OL
EZ fry). -- -.Cerv.Sc.
ZZ eee Se eatmor:
a ee =¥-\-- Rec.Br.
Sex,- \--Mes.moi.
a
af
4 is Vio —
as
==> Rec.Br. \
\
oe A. bse:
Fig. 47 Ventral view of the uninjected lungs of a ten and one-ha f day em-
bryo. Shows the air-sacs and their recurrent bronchi. The relations of the
mesial and lateral moieties of the interclavicular air-sac are well exhibited.
Rec. Br., recurrent bronchi, other reference letters as under figure 34.
bud for this group starts at a point just dorsal to the base of the
first-formed group. At the close of the tenth day this outgrowth
(fig. 39) has bifurcated so that there is an anterior and a lateral
limb. The anterior limb curves gently upward. The lateral
branch makes a more sudden turn posteriorly.
The eleventh day stage. The eleventh day stage is of some es-
pecial interest since it corresponds with Selenka’s figure (78) in
THE EMBRYOLOGY OF THE BIRD’S LUNG +9]
which two moieties of the interclavicular sac are shown on the
left hand of the figure (right lung). The air sacs are all clearly
outlined at this stage and the recurrent bronchi have started.
The air-sacs as shown in figures 47 and 48 are relatively much
larger than in the earlier stages.
AbdSe at
Fig. 48 Lateral view of the left lung of the same specimen. Showing the
early condition of the recurrent bronchi of the four air-sacs possessing them.
The mesial moiety of the interclavicular sae is hidden from view.
The cervical, the least modified of all the air-sacs, exhibits an
expansion of the distal end, but no recurrent bronchi are devel-
oped from it.
The lateral moiety (fig. 47, Lat.moz.) of the interclavicular sac
now projects well beyond the lung surface. The beginnings of
the recurrent bronchi of this moiety (first seen in the tenth day
stage) have elongated and undergone division. They are not
10 WILLIAM A. LOCY AND OLOF. LARSELL
symmetrically developed, those of the right lung showing more
branches. Schulze has pointed out that, after the union of the
parts to form a single median sac, there are, commonly, recurrent
bronchi only on one side.
The mesial moiety (Mes.moz.) of the interclavicular sac is
forked at its extremity into two lobe-like branches, the more
mesial of which extends towards the median plane and comes
nearly into contact with the corresponding branch from the op-
posite lung. The laterally extending branch, passes ventral to
the extra pulmonary bronchus, and partly engirdles the bronchus
on the left hand of the figure.
The anterior intermediate air-saec (A.int.Sc.) has increased in
size and its recurrent bronchi are much further developed. They
occupy the antero-lateral border very close to those of the lateral
moiety of the interclavicular sac.
Figure 47 shows also the connection between the mesial moiety
of the interclavicular, and the anterior intermediate—both aris-
ing on a common canal (interclavicular canal) that opens into
the third entobronchus (entobronchus not shown).
The posterior intermediate and abdominal air-sacs le at the
caudal extremity of the lung and have obviously increased in
size. In figure 47, only the proximal ends of the recurrent bron-
chi have been sketched since these show on the surface. They
exhibit the same relationships as in the tenth day stages.
Figure 48, which is a dorso-lateral view of the same specimen
shows to greater advantage the recurrent bronchi from the lat-
eral moiety and the anterior intermediate air-sac at ten and one-
half days. The specimen in this position also shows the recur-
rent bronchi from the posterior intermediate and abdominal sacs.
The enlargement at the base of the recurrent bronchi well exhib-
ited in the posterior intermediate probably corresponds to the
basal pocket of Schulze.
Although the general appearance of the air-saecs on the tenth
day of development (figs. 36 and 37) have been described, it will
be advantageous for comparison to insert at this point a separate
sketch of the anterior intermediate air-sac and the mesial moiety
of the interclavicular sac. Figure 49 represents these air-sacs as
THE EMBRYOLOGY OF THE BIRD’S LUNG 1g
removed from the left lung of an embryo of nine and one-half
days incubation. (A) represents a view upon that surface of the
anterior intermediate air-sac which is in contact with the lung.
It is notable for showing the primordia of recurrent bronchi
budding from the anterior lateral border of the air-sac. By
comparison with figure 47 it will be observed that the recurrent
bronchi of this sac lie close to those from the lateral moiety of the
interclavicular sac.
Figure 49 shows further the mesial moiety of the interclavicular
sac spinging from a common canal (the interclavicular canal)
into which the anterior intermediate sac also opens. The inter-
elavicular in turn opens into the third entobronchus. No recur-
rent bronchi arise from the mesial moiety. (B) shows a view
upon the cephalad surface of the anterior intermediate air-sac
and brings into prominence the forked extremity of the mesial
moiety of the interclavicular air-sac, and, also shows the common
origin from the third entobronchus of the mesial moiety and the
anterior intermediate sac.
The twelve and fifteen day stages. The subsequent history of
the air-sacs presents little difficulties except as regards the forma-
tion of the azygous condition of the interclavicular sac of the
adult from the union of four parts which arise separately in the
embryo. The changes in this air-sac are so unusual that they
will be described more in detail.
The condition of the air-sacs at the close of the twelfth day of
incubation is represented in figure 50. This is the camera out-
line of a dissection and in finishing is made only slightly diagram-
matic. The external aspects of lungs and air-sacs are represented
but the recurrent bronchi have not been sketched. By com-
parison with figure 47 (the eleventh day stage) it will be seen
that the medially directed forks of the mesial moieties of the
interclavicular sac have expanded and approach each other more
closely in the interbronchial region. The cranially directed
prongs have also lengthened and extend forward nearly parallel
to the trachea.
In this figure the connection between the mesial moiety of the
interclavicular and the anterior intermediate air-sac is clearly
2 WILLIAM A. LOCY AND OLOF LARSELL
shown as well as the single orifice by which they open into the
lung.
The lateral moiety of the interclavicular is relatively larger
than in earlier stages but otherwise shows no marked change.
Between the twelfth and fifteenth days occur relatively rapid
expansions of the moieties of the interclavicular sac. The me-
sial moieties have fused with each other along the median line
A i B
Fig. 49 Anterior intermediate air-sac of the left lung of an embryo, nine and
one-half days incubation. (A) View upon the surface that is in contact with the
lung. Notable for showing the primordia of the recurrent bronchi springing from
the anterior intermediate air-sac The mesial moiety of the interclavicular air-
sac is also shown. Ent. 3, opening into the third entobronchus; Bd., buds of re-
current bronchi; Mes.moi., mesial moiety of the anterior intermediate air-sac.
(B) the same as seen from the cephalic end. Illustrates the connection between
the anterior intermed ate andthe mesial moiety of the interclavicular air-sacs,
and also the forked extremity of the mesial moiety.
between the two lungs. They have apparently also united with
the greatly expanded lateral moieties of the interclavicular (fig.
51). The dividing membranes remain for several days subse-
quent to the fusion of these different parts. The wall between
the mesial moieties does not disappear until the first day after
hatching. The septum between the mesial and lateral moieties
is less persistent and, so far as dissections indicate, breaks down
during the eighteenth day. In attempting to designate the time
THE EMBRYOLOGY OF THE BIRD’S LUNG 13
at which particular morphological changes occur one must, as
previously indicated, take into account that individual variation
is very common in embryonic development.
The lateral moieties (fig. 51) have greatly expanded so as to
unite in the median line. At this period they are the most promi-
nent part of the interclavicular sac. The stalks connecting this
portion of the interclavicular with the first entobronchus are
/ : ; --—P |nt.Sc.
Fig. 50 Ventral view of the lungs and air-saes of a twelve day embryo.
clearly indicated but the recurrent bronchi have not been shown.
At the extreme antero-lateral margin of this division of the sac
there is a narrow neck opening into the axillary sac.
The method of formation of the single interclavicular sac of
the adult is now clearly foreshadowed. The four parts from
which it is formed (two moieties from each lung) are in contact
but still separated by partition walls.
The sixteenth day stage. The recurrent bronchi are now suf-
ficiently advanced to observe the main features of their distribu-
tion. Figure 52 represents a partly diagrammatic camera trac-
14 WILLIAM A. LOCY AND OLOF LARSELL
ing of a dorso-lateral view of the lung and shows the relations of
the recurrent bronchi of the two posterior air-sacs at the close of
the fifteenth day of incubation. The recurrent bronchi are rep-
resented black and the other air passages in stipple.
The distal tips of the longest recurrents of the abdominal air-
sac have anastomosed with the latero-ventral parabronchi of
=P invse:
ae Abd.Sc.
Fig. 51 Latero-ventral view of the lungs and air-saes of a fifteen day embryo.
Figures 48, 49 and 50 should be compared to show the progressive development
of the air-sacs and the way in which the four moieties of the interclavicular be-
come united.
the first entobronchus. The more dorsal branches unite with
parabronchi of laterobronchi. It is worthy of notice that the
main stems of the first and second groups of recurrent bronchi
of the abdominal sac have united so that the second group appears
to be a branch of the first. By reference to figures 39 and 40 it
will be seen that the two groups were originally separate. By
THE EMBRYOLOGY OF THE BIRD'S LUNG 15
uniting into a single trunk there is a single orifice (do.) opening
from the abdominal sae into the recurrent bronchi.
The recurrent bronchi of the posterior intermediate air-sae do
not extend so far forward as do the branches of the preceding
eroup. They occupy the extreme ventral part of the lung.
Their anastomoses (not sketched) are principally with parabron-
chi of the first and second laterobronchus.
The recurrent bronchi of the two other air-sacs anastomose
during the sixteenth day with parabronchi in parts of the lung
adjacent to them. The distal ends of the recurrents from the
interclavicular anastomose chiefly with the more ventral para-
bronchi of the first entobronchus and the recurrents of the an-
terior intermediate sac unite with parabronchi of the latero-
ventral part of the lung. It results that the anterior intermedi-
ate sac comes into communication with the air circuits in the
latero-ventral lung region and the interclavicular sac comes into
communication with passages in the anterior part of the lung.
As already pointed out, the anastomosing twigs are at first
very slender, but, by the eighteenth day of incubation have in-
creased in diameter so as to be practically the same size as the
branches which they connect. The recurrent bronchi have by
the eighteenth day of development assumed the relations to other
parts of the bronchial tree which they bear in the adult lung.
Transition to the adult. In showing how the adult condition
is reached it will be advantageousto summarize the principal
changes subsequent to the eleventh day after which period the
sacs grow more rapidly.
The abdominal sacs expand so as to fill the abdominal cavity,
and partly surround the viscera therein contained. About the
fourteenth day the walls of these sacs begin to fuse with the peri-
toneum and this fusion is apparently completed sometime be-
fore the eighteenth day of development. The left abdominal sac
is somewhat larger than is the right.
The history of the posterior intermediate sacs after the eleventh
day is closely parallel to that of the abdominals and does not
require detailed description.
16 WILLIAM A. LOCY AND OLOF LARSELL
The same general course is followed by the anterior intermedi-
ate sacs. Their walls fuse with the lining of the thoracic cavity.
The cervical and interclavicular sacs attain their most rapid
growth after the twelfth day of development. The cervical sacs
grow forward toward the neck of the chick and between the
fifteenth and nineteenth days of incubation their walls fuse to
some degree with the pleura. They give rise to several subdi-
visions in the cervical and axillary regions. .
The later stages of the interclavicular sacs require a more ex-
tended description than the others because of marked differences
in their formation.
Returning to figure 47, the representation of the condition in
the ten and one-half day embryo, we note again that the mesial
moiety is bifurcated at its distal extremity. The more mesial
lobe thus produced expands in such a manner that its walls come
into contact with the walls of the corresponding lobe of the in-
terclavicular sac of the opposite lung. This phase is reached on
the fifteenth day of incubation (fig. 51). By the nineteenth
day fusion of the walls has taken place, but there appears to be
no breaking down of the septum thus formed. This appears also
to be the case with the fused walls of the more anterior portion
of the sac. This condition was demonstrated both by dissections
and by Wood’s metal casts of the adult lungs and air-sacs.
On the sixteenth day of development portions of the mesial
moiety of both sides grow ventrally over the bronchus, and come
into contact with the lateral moiety of the interclavicular sac.
The membranous walls subsequently begin to fuse, and on the
eighteenth day union is approximately completed. The single
septum thus formed disappears sometime between the nineteenth
day and the end of the first day after hatching, so that the two
hitherto independent moieties coalesce to form one sac (fig. 53).
Thus the single large interclavicular sac of the adult is the
result of the union of two moieties on each side which arise from
different entobronchi. As diagrammatically illustrated in figure
53, the union of the sacs and the disappearance of the septum is
completed by the close of the first day after hatching.
THE EMBRYOLOGY OF THE BIRD’S LUNG 17
So far as we are aware a similar history of the formation of
the interclavicular sac has not been given. It has been known
since the time of Sappey (47) that the single interclavicular sac
of the adult was produced by the union of two parts, but the
formation of two moieties on each side from independent sources
©} p. Int.
Rec.Br. Lat. 3
Fig. 52 Diagram of the lateral surface of the right lung of a fifteen day em-
bryo, to show the recurrent bronchi of the abdominal and of the posterior inter-
mediate air-sacs. The outlines of the recurrent bronchi (in black) so far as rep-
resented were traced with the aid of the camera lucida from an air injected
preparation. Dotted portions diagrammatic. O.abd., orifice of the abdominal
sac; O.P.Int., orifice of the posterior intermediate; An., anastomosis of recur-
rent bronchi. Other reference letters as before.
and the details of their union had not, we believe, been antici-
pated.
We turn now from the air-saes to the recurrent bronchi, the
embryonic history of which has been outlined from their earliest
appearance to the time when they have established their union
with other branches of the bronchial circuits.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. |
18 WILLIAM A. LOCY AND OLOF LARSELL
Since the orifices of the recurrent bronchi are usually close to
the direct opening into the air-sacs, attention should first be
called to Campana’s (’75) analysis of orifices of the air-sacs: He
divides them into two groups, the monobronchial and the poly-
bronchial, according to the number of openings which they ex-
hibit. The polybronchial orifices are further separated into two
categories, the simple and the mixed polybronchials, the simple
polybronchial being composed of several openings of one grade
(parabronchi), and the mixed polybronchial embracing openings
of two grades, or, as in the case of the interclavicular and the
anterior intermediate, uniting two neighboring sacs. Juillet ac-
cepts Campana’s terminology but substitutes a clear basis of
distinction for his confused account of polybronchials. This dis-
tinction consists in recognizing direct and recurrent orifices as the
components of the polybronchials. A monobronchial orifice must
necessarily be direct, but a polybronchial group may be com-
posed of both direct and recurrent orifices (mixed polybronchial
or of recurrent orifices only (simple polybronchial).
According to this analysis, which is in harmony with the de-
velopmental history, the cervical sac has a direct monobronchial
orifice only, since it does not possess recurrent bronchi. A num-
ber of branches of the first entobronchus ramify posteriorly into
the lung from a point nearly opposite the direct orifice of the
cervical sac. These possibly serve the same purpose as the re-
current bronchi of the other air sacs, but do not appear to be
developed from the sac itself as are the tubes which have been
called recurrent bronchi.
The interclavicular air-sac is connected with the lung by two
groups of orifices. The more mesial orifice, which is of the mono-
bronchial direct type, opens into a short tube, the interclavicular
canal, which also receives the direct orifice of the anterior inter-
mediate sac. This short tube in turn communicates with the
third entobronchus.
The more lateral group of interclavicular orifices is mixed poly-
bronchial, having both a direct and several recurrent orifices.
The direct orifice is the opening of the ventral tip of the lateral
branch of the first entobronchus, from which as already deseribed
THE EMBRYOLOGY OF THE BIRD’S LUNG 19
the lateral moiety of this sac has its origin. The recurrent ori-
fices are three or four in number, and are the proximal openings
of the recurrent bronchi. Reference to figure 47 will show that
on the eleventh day of incubation there were but two recurrent
openings connected with the sac. The change to the adult con-
dition is brought about by the extension of the proximal ends of
the original recurrent bronchial buds.
Juillet insists that there is no direct connection of the inter-
clavicular with the first entobroncus, and accordingly, classifies
this lateral group of orifices as simple polybronchial. We con-
clude, however, that the group is a mixed polybronchial and will
return to the question under our general considerations.
The anterior intermediate air-sac also has two groups of ori-
fices. The monobronchial direct orifice (figs. 47, 54) has already
been mentioned in connection with the corresponding orifice of
the interclavicular air sac. The other and more ventral group is
of the polybronchial recurrent type. The number of orifices
varies, but is usuall five or six.
The posterior intermediate sac communicates with the lung
by a polybronchial group of orifices made up of both direct and
recurrent tubes (polybronchial mixed). There is one direct ori-
fice (figs. 53, 54), which is the opening of the third laterobronchus
from which this sac has its origin. The recurrent orifices (figs.
53) represent the lungs of a one-day chick, but the relation of
the orifices is the same as in the adult lung), three or four in
number, as the openings of the recurrent bronchi and have a
history very similar to that already described in connection with
the two preceding sacs. The anterior ends of the posterior in-
termediate recurrent bronchi have for the most part anastomosed
with the first and second laterobronchi.
The orifices of the abdominal air-sac are also of two kinds, di-
rect and recurrent, which are so arranged as to form a mixed
polybronchial group. The direct orifice (figs. 53 and 54) is the
opening of the mesobronchus into the sac. The recurrent ori-
fices arising as previously described are four or five in number.
Schulze has shown that variations as to number exist in different
birds being six to nine and sometimes reduced to one.
20 WILLIAM A. LOCY AND OLOF LARSELL
Figure 45 shows the surface aspects in the adult of the recur-
rent bronchi of the abdominal and posterior intermediate air-sacs
to the bronchial tree and the way in which their stems connect
with the air-sacs represented. In this metallic cast the abdomi-
nal air-sac was only partly injected so that the proximal end
only is shown. It will be seen that the main recurrent bronchi
of the abdominal air-sac extend more than one-third of the way
O.Lat.mol.
Rec. Br=~=
O.Mes.moi. -
OAMintat 4
htepeeao
ato = 2
O-P..Int.-—
Fig. 53 Diagram of a ventral view of the lungs on the first day after hatching.
The moieties of the interclavicular sac have united on each side and those of the
right and left lung have come into contact, being separated only by a temporary
partition wall. Shows also the nature of the orifices into the air-sacs. Cervical
sac not represented.
toward the ventral anterior border of the lung before branching
to any marked extend. The rami into which these recurrent
bronchi finally break up anastomose with the numerous air-
passages in the lateral facet of the lung.
Summary. Summarizing our observations we conclude that:
1. The recurrent bronchi are offshoots from the air-sacs lead-
ing into the lung parenchyma, and the air-sacs in turn, except
the abdominal (from the mesobronchus), are the expanded ter-
THE EMBRYOLOGY OF THE BIRD’S LUNG 21
minal portions of secondary branches of the bronchial tree.
Thus the recurrent bronchi sustain the same relation to the air-
sacs that the parabronchi do to the respective secondary branches
from which they have their origin.
2. By means of the recurrent bronchi and their anastomoses
with other branches of the bronchial circuits the air-sacs are
brought into communication with all parts of the lung. They
Rec.Or. _ y
A.lnt.Sc. 3s
Fig. 54 Diagram of the ventral aspect of the adult lung constructed from
studies of the prepared lung and from Wood’s metal casts. Shows the nature of
the seven orifices into the air-sacs as described in the text. DO., direct orifices,
Rec.O., orifices of recurrent bronchi.
have direct communication with bronchus and central lung tube
through their ‘direct orifices’ and a recurrent communication
through the recurrent bronchi.
3. The unpaired interclavicular air-sac of the adult fowl is the
result of fusion of four embryonic outgrowths: two moieties from
each lung which first unite and then undergo fusion across the
median line to form the single interclavicular air-sac.
22 WILLIAM A. LOCY AND OLOF LARSELL
4. THE DEVELOPMENT OF THE PULMONARY ARTERY
In describing the external appearance of the embryonic lung
the pulmonary artery and pulmonary vein were noted as a part
of the surface view. We shall now give a more detailed account
of the method of origin of the pulmonary blood vessels and of
the embryonic changes that they undergo.
The pulmonary artery is formed by the union of two parts, one
of which, the proximal end, sprouts from the sixth aortic arch,
and the other, the distal end, begins in the lung wall and grows
towards the sprout from the aortic arch.
The vascularization of the walls of the lung precedes the forma-
tion of the pulmonary artery and the distal extremity is the first
formed. Examination of sections of the fifty-two hour stage
shows the presence of rounded vascular spaces in the mesenchyma
of the lung primordium. When first formed these vascular
spaces are of small extent and can seldom be traced through
more than two sections, but the examination of numerous speci-
mens of this age shows that they are fairly constant in appearance
and as to their position in the median and dorsal portions of the
lung parenchyma. ‘These represent the rudimentary condition
of the vascular area of the lungs.
As development proceeds, the vascular spaces assume greater
definiteness, and by extension come together, and fuse forming
an incipient network of capillary-like canals. Between the sev-
enty-fifth and the eighty-second hour, in particular, these spaces
show with increasing definiteness, and by the eighty-second hour
of development the longer vascular spaces can be traced through
twelve or more sections. These are intermingled with other
spaces of less extent, all occupying the area of the lung in which,
at a little later period, the pulmonary artery is formed. In the
early formed vascular spaces it is not obvious which are des-
tined to give rise to the artery and which to the pulmonary vein.
This account agrees in essential particulars with Evans’ obser-
vations of the development of blood vessels from capillary spaces.
These vascular changes in the mesenchyma of the lung wall
begin before the formation of the sixth aortic arch. In many of
THE EMBRYOLOGY, OF THE BIRD’S LUNG 23
our injected specimens the pulmonary vein takes the injection
before the artery as shown in figure 6, part I. This injected
specimen gives a view of the pulmonary veins as seen from the
ventral aspect of the lung. A single well defined blood vessel
Fig. 55 Diagram of the dorso-lateral aspect of the adult lung. Exhibits
openings into the mesobronchus of the dorsobronchi (Dor.) as seen when their
stems are severed.
/
|
| f |
! J , t
Ent.4 Ent.3 Ent. 2Ent. 1
Fig. 56 Diagram of the mesial face of the adult lung to show parabronchi
connecting ento- and ectobronchi.
24 WILLIAM A. LOCY AND OLOF LARSELL
passes on the ventral surface of each lung; these two unite into
a trunk vessel situated in the median plane, and this, in turn,
passes to the left atrium of the heart. Anteriorly is situated
another vein that runs along the median line of the laryngo-tra-
cheal groove and also unites with the trunk vessel that leads into
the left atrium. In this specimen the pulmonary artery was
not seen.
There is considerable individual variation in the time of forma-
tion of the blood vessels, so that the precise time and the degree
of development is not identical in corresponding specimens.
Some of this observed variation may be owing to imperfect in-
jection. Nevertheless, the method of formation of the pulmon-
ary artery is sufficiently definite to leave little room for doubt.
By the beginning of the fifth day there is a stem vessel in the
lung wall and a short spur from the ventral end of the sixth aor-
tic arch. These are directed towards each other but they are
separated by a very obvious interval, they constitute the proxi-
mal and distal ends of the future pulmonary artery.
Sections of the 96-hour stage (fig. 58) show on each side a
short spur from the ventral part of the sixth arch. About twelve
hours later (45 days) we find in surface views (fig. 57) an almost
completed pulmonary artery. There is, however, satisfactory
evidence in the injected specimens ths the spur from the sixth
arch is not the only growing point in ie formation of the artery.
On the contrary, it meets a forward growing vessel from the
lung, which has been formed through the medium of the vascu-
lar spaces already described, and which precede its appearance.
Fourteen injected specimens of the middle of the fifth day (43
days) were dissected. All showed the complete outline of the
pulmonary artery, but in every specimen it was noticeable that
the pulmonary artery was not of the same calibre throughout
its course. Both ends were well developed, but about midway
between the two ends the diameter was reduced, so that it pre-
sented the appearance of a slender tread. This is well shown in
figure 57 which represents the dissection of an injected specimen
of four and one-half days development. The distal division of
the artery, from the sixth arch, is shorter than the proximal di-
THE EMBRYOLOGY OF THE BIRD’S LUNG 20
vision from the lung. ‘The slender thread-like portions shows the
region of junction between the two ends. It would appear, there-
fore, that the first rudiment of the pulmonary artery begins in
the mesenchyma of the lung walls in the form of vascular spaces,
which by extension and union form a network of connecting pas-
sages, and from these arise the distal end of the pulmonary ar-
tery. The proximal division arises slightly later, springing from
Fig. 57 Dissection of the lung territory of an embryo incubated four and
one-half days. Shows the pulmonary artery not yet completed in its middle
course. Drawn by G. H. A. Rech.
the sixth aortic arch, and then growing to meet the distal divi-
sion from the lungs. When their union is effected the pulmonary
artery is established. Figures 12 and 13, part I, show the sur-
face appearance at five and one-half days after union of the two
parts of the pulmonary artery.
Marshall (’92) was, we believe, the first to maintain that the
‘pulmonary arteries appear in the walls of the lung about the
26 WILLIAM A. LOCY AND OLOF LARSELL
middle of the third day before the two hinder pairs of aortic arches
are formed. On the appearance of the fifth (sixth!) pair of aor-
tic arches the pulmonary arteries become connected with their
ventral ends.”’
The spur from the sixth aortic arch arises before the arch is
completed and at its first appearance it is on the ventral part of
the arch near the truncus arteriosis. The sixth arch, in common
with the others, is formed by a dorsal moiety from the dorsal
aorta (fig. 5), and a ventral moiety from the truncus arteriosus.
| & GE =
/ a. i | | S&S \
/ 2 \ w /
} 2 } ey 1 | i ey
jaa { } i|
| (I of £
{y fe Se % |
Ali x %
(iz Ob eA
} \ \
Fig. 58 Cross section of an embryo of the 96-hour stage to show the distal
part of the pulmonary artery arising from the sixth aortie arch.
The ventral moiety is the longer and as it grows the relative posi-
tion of the arterial spur changes. At the 43 day stage the posi-
tion of the pulmonary artery is nearer the truncus arteriosus
than the dorsal aorta. On the second half of the sixth day (fig.
12) it emerges at about the middle point of the sixth arch. The
change in position is continued until, in later stages (fig. 22) the
base of the pulmonary artery is nearer the dorsal aorta.
While the figures just described convey a good idea of the ap-
pearance of the pulmonary artery from surface views, the study
of injected specimens as transparencies gives an idea of the in-
ternal distribution of vascular loops. Such a specimen of five
days nine hours incubation is represented in figure 59. This is
THE EMBRYOLOGY OF THE BIRD’S LUNG Pat
the stage at which the first entobronchus (nt. /) is given off and
the internal changes are to go on rapidly, accordingly, we may ex-
pect, a good development of blood vessels. The artery enters
the lung substance and passes dorsally as well as nearly parallel
to the lung tube. Loops of blood vessels pass from the dorsal
side around the lung tube and unite with the vein below. These
capillary loops are more abundant near the bud of the first ento-
bronchus and the expanded portion of the lung tube.
At five days, twenty hours, the pulmonary artery is seen (fig.
60) running nearly parallel to the course of the extra-pulmonary
bronchus and entering the lung dorsal to the bronchus. The
pulmonary vein runs through the ventral region below the cen-
tral lung tube.
As shown in figure 60 A, a sketch from the ventral aspect, and
figure 60 B, from a lateral view, the artery branches and divides
into capillary networks around the entobronchi, and, more cau-
dally, around the embryonic vestibulum from which the ectobron-
chi are soon to arise. The capillary network of the dorsal side,
having surrounded the lung tube and its outgrowths, passes ven-
trally, and comes into communication with the capillary net-
work of the pulmonary vein.
In the course of a few hours the branching of vessels within
the lungs has obviously increased. Figure 61 A is the sketch of
the left lung and 61 B of the right lung of an embryo incubated
five days, twenty-two hours. In the left lung are represented
the chief branches of the pulmonary vein and in the right lung
both veins, arteries and capillaries are sketched, but the capil-
lary network has been simplified. The pulmonary vein is split
into two great divisions that form a fork over the bronchus (fig.
61 B, fig. 21). One arises from venules situated deep within
the lung substance, dorsal to the bronchus, and is designated the
internal branch, the other is the external branch. In examining
figures 60 and 61 it is to be understood that the capillary network
is more complex than shown in the sketches, at the same time,
the network was sketched with the aid of a camera, and embraces
those minute vessels that carried the injection and which could
be readily made out.
28 WILLIAM A. LOCY AND OLOF LARSELL
Pul.Vn. 59
p B
A A Abd Se.
60
ay f = afl th ‘« 1 rs . wtr 7 ay =< = 1
Fig. 59 Transparency of the lung of an embryo of 5 days 9 hours incubation
to illustrate capillary loops surrounding the mesobronchus and connecting the
pulmonary artery and the pulmonary vein.
Fig. 60 Lungs of an embryo incubated 5 days 20 hours to show the capillary
network connecting pulmonary artery and vein. The four entobronchi have
been established. The bronchus is occluded in its anterior portion. (A) dor-
sal aspect, (B) lateral aspect.
THE EMBRYOLOGY OF THE BIRD’S LUNG 29
wes a
‘e
oe Jai
Bee
Fig. 61 Transparency of the lungs of an embryo of 5 days 22 hours incuba-
tion. (A) illustrates the chief branches of the pulmonary vein of the left lung.
(B) exhibits the pulmonary artery and the pulmonary vein of the right lung with
parts of the capillary network connecting the two. Buds of ectobronchi are
formed at this stage.
Fig. 62 Dissection of the lung territory of an embryo of 63 days incubation
to show especially the capillary plexus connecting the pulmonary artery and
the laryngo-tracheal vein. In other specimens a small vessel leaves the ventral
wall of the pulmonary artery and breaks into a capillary network (figs. 12 and
14). Shows alsothe connections of the pulmonary veins and the laryngo-tracheal
veins with the trunk vessel leading into the left atrium of the heart.
30 WILLIAM A. LOCY AND OLOF LARSELL
Cross-sections of these stages were also studied under the mi-
croscope and in so far as the territory of distribution is concerned,
bear out the observation on the injected specimens studied as
transparencies.
The surface study of a dissected specimen of 65 days incubation
(fig. 62) shows very well the connection between the pulmonary
artery and the laryngo-tracheal vein. In part I, in a number of
instances attention was called to a perpendicular branch (figs.
11, 12 and 14) emerging from the pulmonary artery upon its
ventral border. Figure 62 shows that this short artery breaks
into a network that is recombined into the vein running along
the ventral border of the laryngo-tracheal region. This speci-
men was imperfectly injected so that the network of blood ves-
sels on the anterior dorsal region of the lung did not show as in
figure 14, but the connections of the arterial branch and of the
laryngo-tracheal vein were well exhibited. The veins from the
two sides join into a median vessel which, in turn, unites with
the trunk vessel opening into the left atrium of the heart. The
branches of the pulmonary vein also unite with this trunk.
On the eighth day of development, as indicated in part I (figs.
14 and 17), surface views of the injected lung show a capillary net-
work occupying the antero-dorsal region of the lung, and, by the
ten day stage the entire latero-dorsal surface is covered by a net-
work of blood vessels. In addition to this there is a well defined
denser network of capillaries upon the dorsal surface. Figure 63
from a specimen of the ninth day, shows this denser area of capil-
laries from the dorsal aspect. In this figure the central portion
of the dorsal region is occupied by a distinctly limited network of
capillaries extending in the form of a longitudinal stripe from the
cranial to the caudal part of the lung. This vascular develop-
ment corresponds in position to the lanc-like area (before men-
tioned) between the ends of the ento- and ectobronchi as they
curve towards each other. It is a characteristic anatomical
landmark of all later stages.
In this sketch an oblique view of the lateral face is exhibited
and, on that surface, the capillary network is obviously more
scattered than on the dorsal surface.
THE EMBRYOLOGY OF THE BIRD’S LUNG 3l
We shall now give attention to the transformations of the sixth
aortie arch by means of which the pulmonary artery becomes
separated from the aorta so as to form an independent blood
vessel coming from the right ventricle of the heart.
In the embryonic bird, after the sixth day, each pulmonary
artery arises from one-half of the sixth arch. In the adult bud
this has changed to the condition of right and left pulmonary ar-
teries connected with a single trunk issuing from the right ven-
tricle. Not only has there been a complete separation of sys-
temic and pulmonic circulation, but, also, the distal extremities
of the sixth arch have atrophied. The way in which this is
brought about may now be followed.
About the middle of the fifth day (45 day stage) the fourth,
or systemic arch, the rudimentary fifth, and the sixth, or pul-
monic arch are present. The fifth arch frequently fails to appear,
and when it is present, it is extremely transitory. The disap-
pearance of the rudimentary fifth leaves the fourth and sixth as
the two posterior arches.
The left half of the fourth, or systemic, arch atrophies, the
right half alone remaining, which becomes much enlarged.
In the meantime a septum develops that separates the right
systemic from the base of the pulmonic, the systemic arch becom-
ing connected with the left ventricle and the stem of the pul-
monic with the right ventricle. When this has been accomplished
the right and left halves of the sixth arch are still present and
undiminished in size.
Figure 64 represents the condition at the close of the fourteenth
day of incubation. The lung of the chick does not become func-
tional as an organ of respiration until shortly before hatching.
Accordingly, most of the blood from the right ventricle passes
through both the left and the right divisions of the sixth arch to
join the aorta with which these are connected. The pulmonary
arteries springing from these divisions remain small (fig. 64).
Near the heart the fourth and sixth arches are separated but the
persistent right half of the fourth, and both right and left halves
of the sixth arch, join the aorta. The fourth arch issues from
the left ventricle, as the aortic arch, and the two halves of the
a2 WILLIAM A. LOCY AND OLOF LARSELL
Fig. 63 Dorsal aspect of the lung of an embryo of the ninth day, showing the
lane-like area of capillaries extending from the cranial to the caudal part of the
lung.
Fig. 64 Sketch from an embryo at the close of the fourteenth day of incuba-
tion to show the way in which the persistent right half of the fourth aortic arch
and the two halves of the sixth arch join the aorta. The pulmonary artery is
small. Drawn by G. H. A. Rech.
THE EMBRYOLOGY OF THE BIRD’S LUNG 33
sixth united into a common trunk are connected with the cavity
of the right ventricle.
Although the pulmonary arteries are relatively small, the right
and left halves of the sixth arch are large, and, since they join the
aorta, the principal function of the sixth arch at this time is in
connection with the systemic rather than with the pulmonic
circulation.
The pulmonary arteries remain relatively small up to the time
of hatching, but those portions of the sixth arch lying behind
them become constricted. This condition is represented on the
nineteenth day of development in figure 65. At this time the
pulmonary arteries are somewhat larger and the portions of the
sixth arch behind them has undergone an obvious constriction.
The parts of the sixth arch behind the pulmonary arteries are
designated each ductus Botalli or ductus arteriosus. In the
sketch the carotids and the pulmonics have been cut trans-
versely so as to show in each case the cut ends of two vessels
instead of single trunks into which they are united nearer the
heart.
The subsequent steps in the formation of the adult pulmonary
arteries involve the disappearance of the ductus Botalli on each
side. When respiration begins, the increase of blood flow
through the lungs reacts on the growing tissue so that the pul-
monary arteries become enlarged, and the ducti Botalli being
deprived of blood rapidly diminish in size and soon become re-
duced to strands of connective tissue which usually disappear
in the adult, although in some species of birds they are persistent
but not functional.
The progress of affairs is represented in figure 66 sketched from
a chick newly hatched. The pulmonary arteries have increased
in size and the ducti Botalli are much reduced.
After hatching the ducti Botalli become occluded and reduced
to slender cords of connective tissue. Figure 67 represents the
condition as seen from the right side, three days after hatching
and figure 68 shows the heart and arteries of a young chick of the
same age. In figure 68 the connections of the arterial trunks
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 1
34 WILLIAM A. LOCY AND OLOF LARSELL
Fig. 65 The arterial trunks on the nineteenth day of development. The
sixth aortic arches are constructed behind the pulmonary arteries into the ducti
Botalli. Drawn by G. H. A. Rech.
Fig. 66 Condition of the pulmonary artery and of the ductus Botalli of the
newly hatched fowl. Drawn by G. H. A. Rech.
Fig. 67 The atrophied ductus Botalli and the enlarged pulmonary artery on
each side, on the third day after hatching. Drawn by G. H. A: Rech.
Fig. 68 The heart and the arterial trunks on the third day after hatching.
The ductus Botalli of each side is reduced to a slender cord and the pulmonary
arteries have attained the diameter of the sixth arch of which they are contin-
uations. Drawn by G. H. A. Rech.
THE EMBRYOLOGY OF THE BIRD’S LUNG 35
with the heart is exhibited. The pulmonary arteries are now
large and the ducti Botalli are atrophied.
The effect of the obliteration of the ducti Botalli is that the
blood from the right ventricle can no longer pass through the
sixth arch to the aorta but is sent entirely through the pulmonary
arteries to the lungs.
COMMENTS
The morphology of the avian lung can be made clear only by
observations of its development. It is through this channel alone
that one becomes acquainted with the nature of the modifica-
tions of the bird’s lung that place it in a class by itself. It can be
safely said that the facts of morphology separate the avian lung
from the lungs of all vertebrates, with the possible exception of
the reptiles. The excellent papers of Milani (94) and of Hesser
(05) on the embryology of the reptilian lung are classics, but the
embryology of the air-sacs is insufficiently known. On account
of the development of air-sacs, of recurrent bronchi and laby-
rinthine connections between all air passages, the avian lung ex-
hibits a special architecture and upon our understanding of this
architecture will depend our conception of its physiology. While
we have described what we conceive to be its physiological anat-
omy, we have not engaged in experiments that entitle us to make
any special comment on its physiology.
The steps in arriving at a conception of the bird’s lung have
been gradual and dependent upon increasing definiteness in the
knowledge of its internal architecture. Since the observations of
the early investigators have been a factor in molding this concep-
tion, and since ideas of lung anatomy have been so dominated by
their results, they may be briefly summarized. Moreover, the
sketches of Selenka (’66) are still used in text-books to illustrate
the development of air-sacs (vide Lillie, Development of the
Chick, ’08) Hertwig-Kingsley, Manual of Zoology, Revised Edi-
tion, 712), accordingly a summary will not deal with obsolete
matters.
36 WILLIAM A. LOCY AND OLOF LARSELL
The traditional view, that somehow the bird’s lung should be
swung into line with the others, and ought to be compared part
by part with the mammalian lung, persisted for a long time and
created difficulties of interpretation. The idea that there are
culs-de-sac on the ultimate twigs of the bronchioles, correspond-
ing to the alveoli of the mammalian lung, has so often been
tacitly assumed in the descriptions of the anatomy of the bird’s
lung that confusion has resulted. The conception, so funda-
mentally different from this, of labyrinthine passages, all inter-
communicating, and forming bronchial circuits instead of a
bronchial tree, has been a matter of gradual growth.
Inasmuch as the pioneer observers examined the bird’s lung
with great care, by transmitted as well as by reflected light, it is
pertinent to inquire to what extent the structure of the bronchial
tree was anticipated.
Rathke’s (’28) figures of the embryonic lung of the chick show
an attempt to represent the internal anatomy of the lung. In
his figure 15 of the seven day stage, the main bronchus is shown
with hernia-like growths (entobronchi) coming from it. In one
of his five figures of the eleventh day (his fig. 16) he shows the
air-sacs and a more profuse branching of the bronchial tree. In
his figures 11 and 21 he sketches details of the terminal twigs
which he illustrates as ending in grape-like clusters. These he
compares directly with the alveoli of the sheep’s lung of which he
gives a similar picture.
Von Baer (28) gives no picture of the embryonic lung but his
descriptions show that his observations were carefully made.
While Rathke’s are tinctured with a subjective bias, Von Baer’s
are objective.
The next figures of importance for our review are those of
Remak (755). He was an excellent observer and his sketches are
of high quality. His illustrations show clearly two entobron-
chial buds on the fourth day of development (his fig. 78). This
shows well the mesenchymic sheath and the endodermal tube.
His pictures of the lung at 33 days (fig. 72), at the end of the
fourth day (fig. 74), and on the fifth day (fig. 75), show two en-
tobronchial buds. His sketch of the early sixth day (fig. 88)
THE EMBRYOLOGY OF THE BIRD’S LUNG Bi
shows three such buds and of the seventh day (fig. 79) shows four.
These figures are evidently not drawn with the camera since
although the number represented is correct, the buds are too
widely separated and cover too much territory of the central
lung tube. The sketch of the condition on the eighth or ninth
day (fig. 80) shows both ento- and ectobronchi as well as the ex-
panded end of the mesobronchus. It should be said that Re-
mak’s figures represent the essential features of the anatomy of the
embryonic lung. They are somewhat out of proportion and they
show the great difficulty of study by transmitted light without
the aid of some especial method of injection.
Selenka’s studies (’66) of the development of the air-saes of the
chick added a number of points to the anatomy of the avian lung.
His figure of the 35 day stage shows the occluded bronchus (fig.
2). In the fifth day stage he illustrates for the first time, the be-
ginning of the embryonic vestibulum (his fig. 3). In his sketch
of the sixth day stage he represents the bud of the first ento-
bronchus, not quite in position, however, in reference to the ex-
pansion of the lung tube, being in his figure posterior instead of
anterior to the expanded part. His sketch of the condition on
the seventh day (his fig. 5) shows seven or eight outgrowths
(buds of ento- and eetobronchi) of the central lung tube.
His figure of the lungs and air-saes of the eleven day embryo is
very interesting and, as already stated, has been extensively
copied in text-books. On the right lung is sketched (2’) the sac
that we have designated the lateral moiety of the interclavicular
sae and which he designates ‘‘cellula infra laryngeus, on the left,
not yet united’’—‘‘stetz sich spiiter in die cellula axillaris fort.”
Although this figure needs attention, and some correction, it is
for the methods at his command a good figure. This figure also
shows on the right lung the mesial moiety of the interclavicular
sac, but it is represented on the wrong side of the bronchus. As
shown in our figures 21 and 47 it arises on the mesial side and
extends across the bronchus ventrally to the lateral side.
The sketches of Selenka and of Remak, made by talented ob-
servers show the limitations of observing the internal structures
of the lung by transmitted light without the help of injections.
38 WILLIAM A. LOCY AND OLOF LARSELL
The reconstruction method has the advantage of giving relatively
large models, but it is difficult and protracted. Injection with
air, described under ‘technique,’ is simpler, 1t can be repeated
indefinitely with the same specimen and gives clear pictures of the
minuter details.
The researches of Sappey (’47) laid the foundation for those of
Campana (’75) which, in particular, mark the next advance in our
knowledge of the bronchial tree. These investigations were car-
ried out on the adult lung and resulted in accurate figures and ex-
tended descriptions. The quality of Campana’s work has been
spoken of before, but too great emphasis can not be laid on the
thoroughness of his anatomical analysis. It is now fifty years
in the past and on that account, coupled with the fact that his
memoir is not easily obtainable, it is likely to be slighted. The
memoir is complete and critical for the adult stage embracing
the intra- and extrapulmonary features, the bronchial tree, the
alr-sacs, their orifices and the bronchial circuits. Campana was
apparently the first to fully grasp the idea of bronchial circuits.
The earlier observers referred to, indicated the beginning of bron-
chial branches but their conception was apparently that of a true
bronchial tree comparable to that of mammals. Schulze (’71),
in giving a picture of the histology of the bird lung showing among
other features the air-capillaries evidently interprets them more
as alveoli than as a network of connecting passages.
Fischer (05) studied the bronchial tree by injections and pub-
lished many figures of wax casts. His descriptions are terse,
more general than critical and are somewhat burdened with a
new terminology.
It is through the investigations of Schulze (11) and Juillet (12)
that we arrive finally at our present conception of the architec-
ture of the avian lung. Their discovery of the recurrent bronchi
has been sufficiently commented upon in previous pages. Juil-
let’s investigations introduced some innovations besides recurrent
bronchi, as sketches of the embryonic tree of six and eight day
stages. He was also the first to show the method of develop-
ment of the recurrent bronchi, and after Schulze, to recognize
their significance.
THE EMBRYOLOGY OF THE BIRD’S LUNG 39
Some particulars in which our results differ from those of pre-
vious observers may now be mentioned.
Comparison of embryonic stages. On account of the method
employed, of air injection, we were able to carry our studies of the
embryonic bronchial tree to later stages of development than
those previously sketched. One who compares our figures of the
embryonic tree with those of Juillet will note several discrepancies.
The development of the bronchial tree in his reconstruction of the
six day stage (his fig. 4) is more advanced than in our sketch of
five day twenty hours which is well along in the last half of the
sixth day. His reconstruction of the bronchial tree of the eighth
day (his fig. 5) is also in advance of our sketches of the same
structure in the early part of the ninth day. There is substantial
agreement as to the number of main branches but differences in
detail including one of importance, viz., the relation of the trans-
verse branch of the first entobronchus to the lateral moiety of the
subbronchial sac.
It is also to be noted that Juillet omits the laterobronchi in
his reconstruction of the eight day stage. His plastic reconstruc-
tion of that stage embraces, besides the mesobronchus, only ento-
and ecto-bronchi. In his text, as well as in his sketches, he gives
little consideration to the laterobronchi, and sets to one side the
dorsobronchi. He states clearly his reasons for so doing. Never-
theless, after observing them carefully in embryonic stages as
well as in the metalic casts of the adult lung, we are inclined to
attach considerable importance to the dorsobronchi, and also to
the laterobronchi, on account of the part they play in helping
form the network of the bronchial circuits. As Campana pointed
out, the dorsobronchi (approximately twenty-five in number)
form a fine network in the middle of the dorsal face of the lung
that can be detected by surface studied of untreated specimens.
The laterobronchi, besides giving rise to an air-sac, form many
anastomoses through their branches in the ventral part of the
lateral ahd caudal regions of the lungs and also with the recur-
rent bronchi of the two posterior air-sacs.
The interclavicular air-sac. It is on the question of the em-
bryonic development of the interclavicular sac that our observa-
40 WILLIAM A. LOCY AND OLOF LARSELL
tions are more at variance with previous results than on any
other point. The method of development of this sac is very re-
markable. In late embryonic stage, as in the adult, it is an un-
paired medial sac but it arises from four separate moieties, two
from each lung. The lateral moiety springing from the transverse
branch of the first entobronchus and the mesial moiety coming
(with the exception noted above) from the third entobronchus.
The chief anticipation of the simultaneous existence of lateral
and mesial moieties of this sac is found in Selenka’s sketch of two
separate sacs on the right lung (see reproduction of the sketch in
figure 191 of Lillie’s development of the chick. He says that the
sae which we have designated the mesial moiety later unites with
the cellula axillaris. Only one moiety of the interclavicular (the
lateral moiety) is sketched on the opposite side. As indicated
above his sketch of the mesial moiety shows its origin on the
wrong side of the bronchus.
On account of its position on the lung, the lateral moiety is
the one that has usually been sketched in the published drawings
of embryonic stages of the lung and the mesial moiety has com-
monly escaped notice. However, when the moieties have united,
as in late embryonic stages and after hatching, the more obvious
opening of the interclavicular, which is the mesial, has been
correctly identified and that part of the sac that is derived from
the lateral moieties has been regarded merely as an extension of
the mesial moieties. Thus Juillet’s diagram of the lung of the
embryonic chick represents only the mesial moiety and this as
expanded and bearing recurrent bronchi on its lateral border.
It is only by following the embryonic development from the
eleventh to the sixteenth days that the complex relations of this
sac are cleared up.
Guido Fischer (05) maintains that the lateral branch of the
first entobronchus of the adult opens into the interclavicular sac.
He gives a figure of a plastic cast to show this but does not name
the bird in which it is found. While Campana does not mention
it, his figure 11 is suggestive in showing the termination of the
branch in question close to the lateral orifices of the interclavicular
sac.
THE EMBRYOLOGY OF THE BIRD’S LUNG 4]
Juillet (12) is very drastic in his criticism of several of Fischer’s
observations and of this one in particular he says: ‘‘G. Fischer
(05) a donné des bronches diaphragmatiques une discription
assez confuse et & laquelle il y’a plusieurs reproches 4 faire
Le sac interelaviculaire s’ouvre d’aprés lui sur ce
rameau’”’ (lateral or transverse branch) ‘“‘bronchique: orifice
qu'il signale ici est trés certainement lorifice récurrent de ce sac
dont il a mal saisi les rapports exacts,” ete.
However defective Fischer’s observations may be on other
structural matters (as Juillet claims), on this point, in particular,
our observations indicate that Fischer is probably correct. At
any rate, we have found the orifice in question in the embryo
chick and also in the adult. In our preparations there are several
that show both direct orifices, the lateral and the medial, of the
interelavicular sac as in the Wood’s metal cast of which figure 45
is a photograph. Moreover, the existence of a separate lateral
moiety implies, at least in the embryo, a direct opening from that
moiety into the lung. In this criticism Juillet overlooks the
fact that the direct connection between the curvilinear branch of
the first entobronchus and the interclavicular sac has been
indicated by a number of observers, as Huxley (’82), Baer (’96),
Lilie (08), Schulze (10). Accordingly, the claim of Fischer was
neither novel or unique. A further claim of Fischer, as Juillet
points out in the same paragraph, is that the curvilinear branch
of the first entobronchus has also a direct opening into the ante-
rior intermediate sac. On this point of the criticism our obser-
vations lead us to agree with Juillet that such an orifice is lacking.
Juillet’s Schema (fig. X, p. 313), to show the relations of the
alr-sacs and their recurrent bronchi with the ventral face of the
right lung of a chick embryo of the tenth day is faulty in show-
ing the interelavicular sae with a wide lateral extension crossing
the lung transversely and giving rise to recurrent bronchi. The
interclavicular sac at this stage of development (and for several
days thereafter) has separate lateral and mesial moieties as
shown in figures 47 and 49. The recurrent bronchi of the inter-
clavicular sac sprout from the lateral moiety, while, so far as we
have been able to determine, the mesial moiety never has any.
42 WILLIAM A. LOCY AND OLOF LARSELL
Excepting the analysis of Campana and of Juillet, there is
much confusion among authors regarding the orifices of the air-
sacs. In the chick we find seven groups of orifices, agreeing
with Juillet except in regard to the nature of the lateral orifice of
the interclavicular sac. This we find to be mixed polybronchial
instead of simple polybronchial as claimed by Juillet. On the
basis of our observations the seven groups of orifices are: one di-
rect monobronchial orifice for the cervical sac;one medial direct
monobronchial, and one, laterally placed, mixed polybronchial
for the interclavicular sac, one direct monobronchial and one
simple polybronchial for the anterior intermediate sac; one mixed
polybronchial for the posterior intermediate and one mixed
polybronchial for the abdominal sac.
As Juillet (12) has shown in admirable comparative studies,
much variation ‘exists as to numbers (6 to 9) and arrangement of
orifices in the twenty-five species of birds which he studied.
Since our observations are limited to the chick, the reader is re-
ferred to Juillet’s analysis of the different types of orifices (pp.
340-351) which can not be satisfactorily abbreviated.
As regards the method of growth and the type of branching
within the lungs, we shall limit ourselves to the brief remarks—
that the excellent observations of Moser indicate the general
method of growth and on the question of branching by mono-
podial or dichotomous formation, our observations incline us to
adopt the view of an unequal dichotomy.
Recurrent bronchi. The recurrent bronchi are the most im-
portant recently recognized structures connected with the lungs
of birds. As already indicated the credit for the recognition of
their structure, development and physiological function is shared
by Schulze (11) and Juillet (12). These new structures are of
especial interest. There is no doubt that they spring from the
air-sacs and grow into the lungs where they establish numerous
connections with the bronchial branches. This gives a new view
of air sacs: They are expanded parts of the bronchial circuits;
branches from the main bronchus lead into them, but they are not
terminal sacs, they are air reservoirs on the course of the bron-
THE EMBRYOLOGY OF THE BIRD’S LUNG 43
chial circuits. Conduits developed from them turn back into the
lung and join the network of air passages, so that, the air, little
changed, and warmed to body temperature, is carried back into
the lung for aération of the blood. The lung, although relatively
small, is highly vascular and a very efficient organ of respiration.
Rapid respiratory changes are favored by the structures de-
scribed and by the very intimate relations between air capillaries
and blood capillaries. Taken together they constitute a felt-
work of vascular and air capillaries mingled together.
It is clear from Campana’s text that he noticed the recurrent
bronchi of the adult on the four air-sacs from which they have
their origin. He makes comment upon them in each case. His
reference to those of the abdominal sac shows that he thought of
them as combinations of parabronchi. On page 54 of his memoir
he says: “‘On voit 4 la face dorsale du poumon des grosses ter-
tiaries, on pourrait presque dire des secondaires reconstituées
par la reunion des tertiaries plus fines, aboutir sur les parois du
septieme infundibulum (fig. 13, C) ou, ce qui revient au méme,
s’ouvrir dans la termination des dernieres secondaires externes.”
On the whole, Campana’s observations afford a sort of pro-
phetic anticipation of the full recognition of recurrent bronchi.
The discovery of the recurrent bronchi brings a new point of
interest into the study of the lungs of Sauropsida, and it is much
to be desired that extensive comparative studies may be entered
into that will embrace a careful consideration of the air-sacs and
their relations in reptiles.
* The variations in chronology of chick embryos represented in the illustra-
tions of the standard references is so great that an additional comment is
appropriate. Comparison of the figures of Duval and of Keibel and Abraham
shows variations sometimes exceeding twenty-four hours (ef. Duval, fig. 142,
140 hrs.; and Keibel and Abraham, fig. 27, 1144 hrs.). While it would be a satis-
faction to embryologists to have the chronology standardized, the essential point
is a correct analysis of anatomical conditions and the sequence of events.
44 WILLIAM A. LOCY AND OLOF LARSELL
BIBLIOGRAPHY
In order to limit as far as practicable the bibliographical references to small
compass, only those are listed that were of especial service in the preparation of
the paper. The readily accessible papers of Flint, Fischer, Moser and Juillet
contain more comprehensive lists which makes unnecessary the repetition by
title of the other articles consulted, and referred to in our Comments on the
Literature.
Campana 1875 Physiologie de la respiration chez les Oiseaux. Anatomie de
Vappareil pneumatique pulmonaire, etc., chez le Poulet. Masson,
Paris.
Fiscuer, Gurpo 1905 Vergleichend-anatomische Untersuchungen tiber der
Bronchialbaum der Vogel. Zoologica, Bd. 19.
Furnr, J. M. 1906 The development of the lungs. Am. Jour. Anat., vol. 6.
Huxiey, THos. H. 1882 On the respiratory organs of Apteryx. Proceed.
Zool. Soc. Lond.
Jutmttet, M. A. 1911 (a) Rapports des sacs aériens et des bronches chez les
Oiseaux; (b) Observations comparatives sur les rapports du poumon
et des sacs aériens chez les oiseaux. C. R. Acad. Sei. Paris, T. 152.
(c) Phases avancées du developpement du poumon chez e Poulet.
C. R. Soc. Biol., T. 70. (d) Face ventrale du poumon des oiseaux et
diaphragme. Jbid., T. 71.
1912. Recherches Anatomiques, Embryologiques, Histologiques et
Comparatives sur le Poumon des Oiseaux. Archives de Zool.
Expériment. et Gén. T. 9, pp. 207-371.
Larsetui, OLor 1914 The development of recurrent bronchi and of air-sacs of
the lung of the chick. Anat. Anz., Bd. 47.
Minter, W.S. 1893 The structure of the lung. Jour. Morph. vol. 8.
Moser, Fanny 1902 Beitrage zur Vergleichenden Entwicklungsgeschichte der
Wirbeltierlunge. Archiv. fiir Mikros. Anat. Bd. 60.
Saprrey, P.C. 1847 Recherches sur lappareil respiratoire des Oiseaux. Ger-
mer-Bailliére, Paris.
Scuuuze F.E. 1872 The lungs in Stricker’s Manual of Histology.
1908 Die Lungen der African Strausses. Sitzungsber. d. Kgl. Preuss.
Akad. Wiss.
1909 Uber die Functionen der Luftsiicke bei den Vogeln. First men-
tion of recurrent bronchi. Sitzungsber. d. Kgl. Preuss. Akad. Wiss.
6. Mai.
1910 Uber die Bronchi saccales und den Mechanisms der Atmung
bei den V6geln. Jbid. 2 Juni.
1911 Uber die Luftsiicke der Vogel. Verhandl. d. VIII Internat.
Zool.-Kong. zu Graz. Aug. 1910.
SELENKA, E. 1866 Beitrag Zur Entwickelungsgeschichte der Luftsiicke des
Huhns, Zeitsch. fiir wissenschft. Zool. Bd. F. 16.
Zumstein, J. 1900 Uber den Bronchialbaum der Ssuger und Vogel. Sitzungsber.
Ges. Z. Beford. d. Ges. Naturwiss. s. 39-48.
THE FUSION OF THE CARDIAC ANLAGES AND THE
FORMATION OF THE CARDIAC LOOP IN
THE CAT (FELIS DOMESTICA)
H. VON W. SCHULTE
~~ From the Anatomical Laboratory of Columbia University
\
SIXTEEN FIGURES
In the transformation of the bilateral anlages of the heart
into a single median organ, two processes are to be distinguished,
the fusion of the myoepicardial mantles and the fusion of the
endothelial tubes. For while the union of the mantles is a nec-
essary condition of the coalescence of the tubes and its char-
acter determines the general features of the endothelial fusion,
yet the latter process is relatively much retarded and evinces a
considerable degree of independence in many details. For these
reasons it will be convenient to follow the two processes sepa-
rately.
This study is based upon the now numerous early embryos
of the eat in the Columbia Embryological Collection. These
were cut in transverse serial sections of 13.3 » and variously
stained. In selected embryos reconstructions by the Born
method were made of the myoepicardium, of the endothelium
and mesenchyme, and finally of the lumina of the heart tubes
and angiocysts. By a comparison of the last two it is possible
to ascertain the precise limits of the solid and hollow parts of
the anlages.
The processes of approximation, of fusion, and of loop-forma-
tion take place in a short period of development as measured by
the rate of formation of the mesodermic somites, the whole
duration of these changes in the cat falling between the stages
8 and 21 somites. At 8 somites (fig. 1) ,the foregut in the car-
diac region is widely open and the splanchnopleure is spread out
45
46 H. VON W. SCHULTE
flat; the cardiac anlages lie in the same plane as the neural tube
and are separated by a wide interval. In the embryo of 11
somites (fig. 3) they are ventral to the closed foregut and their
proximal or bulbar ends are close together. At 12 somites
2
Fig. 1. Reconstruction of compact mesoderm of cat embryo of eight pairs of
somites. Columbia Collection, No.588. 180. Reduced one-half. Dorsal view
the parietal mesoderm has been exsected to show topography of parietal cavity.
1, Parietal cavity; 2, pericephalic cavity; 3, parietal recess; 4, myoepicardial
mantle; 5, oblique groove dividing mantle into mesal and lateral protions; 6,
retrocardiac plate; 7, precardiac plate.
—
=
FUSION OF CARDIAC ANLAGES IN THE CAT 47
fusion of the mantles is complete (fig. 4); at 14 the formation of
the loop is initiated (fig. 7); and at 16 it is completed. The
fusion of the endothelial tubes begins in the embryo of 12 som-
ites, in that of 14 it is all but complete, and the embryo of 20
somites is the oldest that shows a remnant of septum between
the endothelial tubes (figs. 11, 13, and 16).
THE MYOEPICARDIAL MANTLES
There are many brief statements and illustrations of the myo-
epicardial mantles in the literature recording conditions in these
structures immediately antecedent to and during fusion, from
which nevertheless it would be impossible to compile a history
of the process. The observations of Ko6lliker, His, Turstig,
Schultze, Spee, Bonnet, Fleischmann, Selenka and Heape, are
as familiar as they are important and have been adequately
summarized by Mollier,) who has also included the less well
known and less accessible work of Martin,? which is valuable
for its excellent illustrations. Strahl and Carius* and Parker,?
alone have given continuous and detailed accounts of the pro-
cesses here considered. The former have illustrated by a series
of diagrammatic cross sections two different types of fusion, one
in which the formation of the ventral mesocardium antecedes
that of the dorsal, the other in which the converse obtains.
Two factors here come into play, first the width of the foregut
and second the primitive position of the bilateral cardiac an-
lages. The area of visceral mesoderm which intervenes be-
tween the myoepicardial mantle and the mesal angle of the pari-
etal cavity, they designate the retrocardiac plate, similarly that
between the mantle and the lateral angle the precardiac piate
1 Mollier, S. Die erste Anlage des Herzen bei den Wirbeltieren in Hertwig’s
Handbuch der vergl. und experiment. Entwickel. der Wirbeltiere, Jena, 1906,
Bd. 1, which see for literature.
2 Martin. Lehrbuch der Anatomie der Hausteire. Bd. 1, Stuttgart, 1902.
3 Strahl and Carius. Beitriige zur Entwickelungsgeschichte des Herzens und
der K6rperhohlen. Arch. f. Anat. und Physiol., Bd. 15, 1899.
4 Parker, Katherine M. The early development of the heart and anterior
vessels in Marsupials, with special reference to Perameles. Proc. Zool. Soc.,
London, 1915, Pt. III.
48 H. VON W. SCHULTE
(fig. 1). If the retrocardiac plate is narrow relatively to the
foregut, the myocardial mantles will be widely separated upon
the closure of the gut and the ventral mesocardium will precede
the dorsal in its formation. This type obtains in the rabbit,
and to it the cat conforms with some peculiarities in detail which
are recorded below. If, however, the converse obtains, if the
retrocardiac plate is broad relative to the width of the foregut,
the formation of the dorsal mesocardium will be accelerated.
Of this modification of the process the guinea-pig is an example.
This in brief is the analysis of Strahl and Carius. Moller com-
ments justly that is it ‘not quite correct to say that the closure
of the foregut is the cause of the fusion of the anlages of the
heart, because this occurs independently and later, a point which
K6lliker had already made.’ The possibility of a third type of
approximation of the mantles must be admitted, such for ex-
ample as obtains in the chick, where the retro- and pre-cardiac
plates are of such proportions that ventral and dorsal mesocar-
diac are formed at approximately the same time. Further it is
not necessary to assume that the retrocardiac plate is of the same
breadth in its whole extent, and a difference in this respect might
reach such a degree as to produce a mixed type of mesocardial
formation.
In marsupials Parker describes a developmental process re-
sembling but not identical with the conditions present in the
eat. Like that species and the rabbit there is no ventral meso-
cardium. On closure of the foregut the mantles are widely sepa-
rated, the intervening remnants of the precardiac plates forming
a median plate between them. Associated witb this middle
cardiac plate are mesenchyme cells, later capillaries. Subse-
quently the middle cardiac plate is inconspicuous in marsupials
not forming a thicked ridge as in the cat (Cf. Parker, fig. 17,
with fig. 12 of this paper), nor are the angiocysts of the region
assigned a role in effecting the fusion between the endothelial
anlage. These first fuse-in the region of the bulb. In embryos
of 15-16 somites a constriction between bulb and ventricle is
present only on the right side, which as a whole exceeds the left
side in size. Caudad the ventricle is limited by a constriction,
FUSION OF CARDIAC ANLAGES IN THE CAT 49
and the following dilatation where endothelium and myocardium
laterale are closely approximated is interpreted as auricle. The
actual stages of fusion other than in the bulbs and the early
stages of the loops are not represented in Miss Parker’s material.
In the cat from the first the myoepicardial mantle has an ob-
liquely sagittal direction in consequence of which the retrocar-
diac plate broadens somewhat as it is followed caudad, and
tapers in the opposite direction, (embryos of 4-7 somites). The
topography of the parietal cavity in an embryo of 8 somites is
shown in figure 1, which agrees closely with Fleischmann’s
figure of a total view of a cat embryo of this stage.®
Before the infolding of the splanchnopleure is begun, the
topography of the mantles is such as would seem to entail their
earlier approximation caudad where the retrocardiae plates are
broadest. The direct contrary is the case; cephalad where these
plates are narrow the bulbar segments are quickly brought into
apposition, while the ventricular portions diverge and are wide-
ly separated caudad, nor is there sufficient difference in the
width of the gut opposite these two segments of the heart to ac-
count for their difference in position. The problem thus pre-
sented is not easy of solution and its difficulty is increased by
the rapidity of the process, for the heart passes in the period re-
quired for the development of three somites from the position
shown in figure 1 to that in figure 3. Thus in the interval be-
tween the appearance of the eighth and eleventh somites the
formation of the foregut 1s completed as far as the atrial ex-
tremity of the heart, and the anlages of that organ have been
moved through a dorso-ventral arcof nc arly 180° and become
approximated ventral to the pharynx.
It is not possible on the basis of the material in hand to at-
tempt a complete solution of this problem for which several
processes extrinsic as well as intrinsic require minute investiga-
tion. Primarily there are the changes incident at this period
in the general shape of the region, notably the shortening, asso-
ciated with the beginning ventral flexion of the forebrain, which
° Fleischmann, A. Embryologische Untersuchungen. I, 1889.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 29, No. 1
50 H. VON W. SCHULTE
together result in such a remodelling as is hardly to be attrib-
uted to unequal growth. In this shortening is to be found an
explanation of the kinking of the heart tubes prior to their fu-
sion. In the second place the acceleration of infolding of the
splanchnopleure cephalad tends to bring the cardiac anlages into
an oblique position with their arterial ends near together, their
venous extremities divergent. The supposition of accelerated
growth in the cephalic portion of the retrocardiac plate is not sup-
ported by a conspicuous increase of mitotic figures, but the per-
ceptible diminution of its thickness and its rectilinear position
at the end of the process suggest the action of a moulding force.
Finally Mollier’s comment must be borne in mind—the closure
of the gut is not the cause of fusion between the cardiac anlages.
It determines their approximation, but when this is complete the
mantles are separated by an interval, which is bridged by a plate
of mesoderm derived from the precardiac plates. By compres-
sion and ultimate absorption of this into the mantles fusion is
accomplished, and the process as a whole is evidently one of re-
modelling which cannot be explained simply as a cessation of
growth in this district. Similar forces may therefore come into
play in effecting the early and close approximation of the arterial
ends of the mantles notwithstanding the narrowness of the
cephalic portion of the retrocardiac plates.
In an embryo of 9 pairs of somites the cardiac anlages have a
position intermediate between those shown in figures 1 and 3.
The mantles are obliquely placed, their arterial ends close to-
gether though separated by a deep cleft, their venous ends wide-
ly divergent. Each is indented in its lateral contour by an an-
gular incisure which marks the junction of their approximated
and divergent portions. The differentiation of the heart tube
into segments prior to fusion was observed by His and the por-
tions have been variously designated bulbar and atrial or ven-
tricular and atrial. Their subsequent history in the cat show
the first segment to be the bulb continuing forward into the
truncus, the second the ventricle later expanding at its caudal
end into the atrium, but this only after the loop is initiated and
fusion well under way. The angulation is therefore the bulbo-
FUSION OF CARDIAC ANLAGES IN THE CAT 51
ventricular suleus. Caudad the mantles diverge and pass into
the lateral mesocardia with a gentle curve. <A distinet angula-
tion at this point is not apparent until the stage of 11 somites.
On account of the obliquity of the myoepicardial mantles, a
triangular interval with its base at the anterior intestinal portal
is left between them. This is bridged by the residue of the
Fig. 2 Photomicrograph of a cat embryo of nine pairs of somites. Columbia
Collection, No. 532. X 125. Reduced one-half. 1, Ventricular segment of
mantle; 2, middle cardiac plate.
precardiae plates after the closure of the foregut and the dis-
appearance of the ventral mesocardium. It may be designated
the middle cardiac plate.
Between the bulbar segments it is narrow in this differing
from the rabbit (Strahl and Carius), but caudad attains a con-
siderable width (fig. 2). It is seen to agree in thickness and in
the character of its cells with the myoepicardial mantles and
like them to project entadin numerous longitudinal ridges, here
shown in cross-section. These ridges as elsewhere in the splanch-
nopleure are intimately related to the formation of mesenchyme
t
iy H. VON W. SCHULTE
and differ in no important character here in the middle cardiac
plate and in the myoepicardial mantles from those observed in
the splanchnopleure at large. I shall recur to this field of mes-
enchyme and endothelium in connection with the description
of the endocardium. The middle plate is demarcated from the
mantle on each side by a suleus which gradually becomes shal-
low and is effaced as it is followed caudad.
Fig. 3. Myoepicardial mantle of a eat embryo of eleven pairs of mesodermic
somites. Columbia Collection, No. 534. > 300. Reduced one-half. 1, Mid-
dle, cardiac plate; 2, bulbus; 3, ventricle; 4, bulbo-ventricular sulcus; 5, atrio-
venous angle.
The mantles in their ventricular segments are dorso-ventrally
flattened and of great transverse extent. Between them in
ventral view the middle field forms a flat depression. The bul-
bar segments are close together but separated by a deep cleft,
the roof of which is continuous with the middle plate. They
are much narrower and less flattened than the ventricles.
Kes
FUSION OF CARDIAC ANLAGES IN THE CAT rorD)
EXmbryos of 11 pairs of somites show some variation in devel-
opment. Some have hardly progressed beyond the one of 9
somites, but in the most advanced of this stage the mantles are
approximated in the greater part of their length and the middle
plate is considerably reduced. Fusion has not yet occurred
(fig. 3).
The mantles present two angulations in thier ectal contour.
The first bend is the bulbo-ventricular sulcus already described,
which becomes accentuated on the left side in the later stages
as the flexion of the tube develops. In this embryo it is near-
ly rectangular and is approximately symmetrical on the two
sides. I take the deep incisure in Martin’s cat embryo of 4mm.
to be its equivalent. It is also to be recognized in Duval’s fig-
ure® of a chick of 8 somites, though obscured by the outline of
the amnion. The second bend is a more gradual change of
direction at the junction of the heart tube with the omphalo-
mesenteric vein and may accordingly be designated the cardio-
venous angle. It is not to be distinguished in Martin’s figure,
but in the heart of the chick as shown by Duval it is a deep
cleft. These angulations are important in that they determine
the earliest points of fusion in later stages between the endothe-
hal tubes.
The middle plate is much reduced. Its cephalic portion is
concealed in the deep cleft between the bulbs; caudad it forms a
convex triangle exposed in its whole extent. Throughout the
mantles are markedly dorso-ventrally flattened.
The fusion between the mantles and the formation of a dorsal
mesocardium is effected in embryos of 12 and 13 pairs of somites.
The dorsal mesocardium is very short, set off by a sulcus from
the mesoderm covering the foregut and ventrally by a deeper
sulcus from the mantles. Cephalad its leaves separate to give
passage to the forming aortic roots and caudad it expands and
is continuous with the lateral mesocardia.
® Both these cuts are given by Moller. Op. cit. figures 713 and 715. Allen
Thompson illustrates this form of the heart in the chick in his article ‘On the
development of the vascular system in the foetus of vertebrated animals.’ Edin-
burgh new philosophical journal, 1830, Pl. 2, fig. 12.
54 H. VON W. SCHULTE
The mantles are clearly demarcated from the lateral meso-
cardia by deep sulci so that the junction between omphalo-
mesenteric vein and heart tube is deeply indented on each side.
From this point the mantles gradually expand again to be in-
Fig. 4. Myoepicardial mantle of a cat embryo of twelve pairs of somites.
Columbia Collection, No. 547. X 300. Reduced one-half. 1, Middle cardiac
plate; 2, bulbus; 3, ventricle; 4, bulbo-ventricular sulcus; 5, atrio-venous angle;
6, shoulder of mantle.
dented by the deep oblique bulbo-ventricular clefts. Beyond
them the truncus has the form of a cylinder, slightly flattened
_dorso-ventrally, emerging between the high shoulders of the
mantles (fig. 4), In this stage the heart is approximately sym-
metrical and the bulbo-ventricular sulci of the two sides are in
FUSION OF CARDIAC ANLAGES IN THE CAT ‘319)
all important respects identical. The ental projections which
they occasion are shown in figure 5.
The mid-region of the fused mantles requires some comment.
Here the middle-plate has been compressed to a longitudinal
ridge bounded by well defined sulci. These are expressed entally
by ridges, which in an embryo of 13 somites extend far into the
bulbus. In the heart of this 12 somite embryo they are re-
duced in this segment and the middle region of the bulbus is
slightly concave. The fusion is here complete. As the bulbo-
1
Fig. 5 Cephalic portion of same model as figure 4, ental view. 1, Middle
cardiac plate; 2, fundus of right bulbo-ventricular sulcus; 3, fundus of left bulbo-
ventricular sulcus; 4, interior of bulbus; 5, dorsal mesoeardium; 6, right ventri-
cle; 7, left ventricle.
ventricular fissures are approached the ridges begin as low ele-
vations which increase in height to about the middle of the ven-
tricular segment fading out towards the terminal constriction
of the heart. Between the ridges projecting from the middle
plate entally are occasional small processes of mesoderm, evi-
dently remnents of the more numerous projections of earlier
stages.
As compared with the heart of the 11 somite embryo this
heart has lengthened somewhat, but its striking changes in con-
tour are associated with the deepening of the two pairs of in-
cisures at its lateral margin. The bulbar segment has dimin-
ished absolutely in breath and so in less degree has the ventric-
ular especially at its caudal end where it joins the lateral meso-
56 H. VON W. SCHULTE
cardia and septum transversum. The dorso-ventral increase
in diameter is also marked (fig. 12).
Upon this condition follows so rapidly the development. of
the cardiac loop that in only one embryo of 14 somites was an
Fig. 6 Myocardium of a cat embryo of fourteen pairs of somites. Columbia
Collection, No. 188. 300. Reduced one-half, ventral view. 1, Interven-
tricular sulcus; 2, shoulder of left mantle; 3, left bulbo-ventricular sulcus; 4,
right bulbo-ventricular sulcus; 5, lateral mesocardia.
intermediate stage observed. The model of this myocardium
is shown in figures 6 to 8. In figure 6 in which the reconstruc-
tion is viewed from in front it presents resemblances to Mall’s?
7 Mall, F. P. On the development of the human heart. Am. Jour. Anat.,
vol. 13, 1912, fig. 1. Cf. Dandy, W. E. A human embryo with seven pairs of
somites, measuring about 2 mm. in length. Id., vol. 10, 1910. Also Evans,
Keibel and Mall, Manual human embryology, vol. 2, figs. 408-9, and Mall, Ibid.,
vol. 1, figs. 382-6.
FUSION OF CARDIAC ANLAGES IN THE CAT ae
freely-treated model of the heart of a human embryo of 7 to 8
pairs of somites. In both there is a bulbo-ventricular cleft on
the left and there is little in this view of either model to suggest
a mode of formation of the loop different from that given by
Mall, which is simply the deepening of this suleus between bulb
Fig. 7 The same model as in figure 6 viewed somewhat from the right and
above. 1, Bulbus; 2, shoulder of left mantle; 3, shoulder of right mantle; 4,
bulbo-ventricular sulcus; 5, right bulbo-ventricular sulcus; 6, right lateral meso-
eardium.
and ventricle. The heart of the cat is rather more plump, its
contour more convex, but this may well be due to a greater dis-
tension with blood. Two details small in size but not in mor-
phologic significance are present in the cat, which are not shown
in the figure of the human heart. The right margin has a small
indentation which is the remnant of the right bulbo-ventrical
58 H. VON W. SCHULTE
cleft as is readily seen in the view from the right and above
shown in figure 7, and caudad there is a slight nearly transverse
depression furrowing the apex of the cardiac loop, associated
with an ental ridge which marks the beginning of the septum
ventriculorum.
Fig. 8 Cephalic portion of same model as in figures 6 and 7, ental view. 1,
Ridge representing middle cardiac plate; 2, fundus of right bulbo-ventricular
sulcus; 3, fundus of left bulbo-ventricular sulcus; 4, interior of bulbus; J, atrial
region; 6, dorsal mesocardium.
The changes that have supervened to transform this heart in
the period between appearance of the twelfth and fourteen pairs
of somites are easily appreciated on the comparison of the fig-
ures (figs. 4 and 7). The most striking changes affect the shoul-
ders of the mantles, that of the left side is greatly elevated, that
of the right correspondingly depressed. This entails a deep-
ening of the left bulbo-ventricular suleus and an opening out on
FUSION OF CARDIAC ANLAGES IN THE CAT 59
the part of the rght. There is also axial rotation. The shoul-
der of the left mantle thrusts ventrad displacing the bulbus to
the right, and the accompanying rotation displaces the depressed
shoulder of the right mantle dorsad so that it is concealed in ven-
tral view. The loop is directed to the right, ventrad and slightly
eaudad, and the left lateral mesocardium comes to occupy a
slightly more cephalic position than the right. As a whole, the
dorso-ventral depth of the myocardium is increased at the ex-
pense of its transverse breadth (fig. 15) and its extremities are
a little approximated, the sagittal distance between the end of
the bulbus and the junction of the mantles with the lateral
mesocardia being absolutely diminished. The dorsal meso-
cardium is retained unbroken in its whole extent.
The middle plate is undergoing reduction by being absorbed
into the mantles. It is represented by a ridge projecting entad
and at the sides gradually diminishing in thickness as it fades
into the mantles. It can be followed into the beginning of the
bulb where it hes opposite the partition between the endothelial
tubes (fig. 14). It then runs along the convexity of the loop
occupying the same position relative to the fusing endocardial
anlages as in the bulb (fig. 15). On reaching the caudal con-
tour of the loop it becomes continuous with the ental eleva-
tion produced by the interventricular sulcus. Thus by the
ridge and interventricular suleus the primitive median line of
the heart is represented, a conclusion which is borne out by their
location in their whole course opposite the line of fusion, as yet
incomplete, between the endothelial tubes.
In the heart of an embryo of 16 pairs of somites the loop has
increased and projects more strongly. The left bulbo-ventric-
ular sulcus is nearly horizontal and the right has been reduced,
appearing only as a slight furrow ectally and a slight angle with-
in the myocardium. The middle plate is again represented by
a low ridge extending from the end of the bulb to the beginning
of the interventricular septum. This now has an obliquely
transverse direction. In an embryo of 18 to 19 pairs of somites
the ridge of the middle plate disappears. The interventric-
60 H. VON W. SCHULTE
ular sulcus is still oblique. It is only as the venous end of the
heart begins to move towards the right that the septum assumes
a dorso-ventral direction. Its transverse direction in early
stages is rendered possible by the primary displacement of this
extremity to the left as will be demonstrated in the considera-
tion of the endothelial analges.
To summarize, the fusion of the myoepicardial mantles is
accomplished with the aid of a middle cardiac plate, which sub-
sequently becomes reduced to a ridge marking the primitive
median line during the formation of the loop. It is continued
eaudad as the interventricular septum, which thus forms in
the line of original fusion of the heart anlages and by its ap-
pearance separates again, so far as the ventricles are concerned,
the primitive bilateral anlages. The myocardium in the early
stages of fusion is bilaterally symmetrical with a well marked
bulbo-ventricular sulcus on each side. In the formation of the
loop the left suleus deepens and the right opens up and gradual-
ly is obliterated. It is possible that the middle plate now re-
duced to a ridge and located at the convexity of the forming
loop, is less plastic than the thinner portions of the mantles and
failing to lengthen to a sufficient degree exerts a traction which
occasions the appearance of the interventricular sulcus.
THE ENDOTHELIAL TUBES
The origin of the endocardium differs in nowise from the ori-
gin of endothelium elsewhere in the cat. It develops from mes-
enchyme which is formed in loco, first by migration of cells from
the compact visceral mesoderm, second by delamination of
groups of cells from the same source. In this process ridges
and projections are formed by the mesoderm from which the
mesenchyme loosens itself. A means of migration is afforded
the amoeboid cells by the presence of interdermal cytodesmata,
delicate protoplasmic bridges stretching between the mesoderm
end entoderm. ‘The early mesenchyme consists of single cells
and scattered groups which are arranged in plates or even
cords. Within these groups vacuoles appear and enlarging flat-
FUSION OF CARDIAC ANLAGES IN THE CAT 61
ten the containing cells to endothelium.’ The resulting vesicles
Bremer? has termed angiocysts.
In the heart the conformation of the myoepicardial mantle
confines the mesenchyme and angiocysts lodged within its con-
eavity and entails their transformation into a longitudinal
channel. In the embryo of 4 pairs of somites the projecting
ridges of the mantle still intervene between the angiocysts and
delay the formation of a continuous lumen. In embryos, of 7
to 8 somites this is nearly complete, and is so from omphalo-
mesenteric vein to ventral aorta in the embryo of 9 somites.
In all of these however, and to a less degree in still older embryos
there are present unannexed angiocysts and abundant mesen-
chyme about and especially between the endothelial tubes.
The mesenchyme is so gradually transformed into endothelium
that it is not easy to define the lmit between the two stages,
but the endothelium is the dominant tissue by the time fusion
begins. The cat thus conforms to Mollier’s!® observation of
the late fusion relative to condition of tissue in amniotes, the
fusion occurring when the heart is mesenchymatous in saurop-
sids, when it has become endothelial in mammals. His recog-
nition of a stage of solid cords antecedent to the mesenchyma-
tous stage, in which the accelerated fusion of anamniotes occurs,
is theoretically and terminologically not altogether fortunate,
for the undifferented cells first moving into the mesostroma
and subsequently giving rise to a variety of products are prop-
erly termed mesenchyme on grounds of morphology. The term
does not necessarily connote any theoretical prepossessions.
It simply designates a position and arrangement of cells other
than that obtaining in the three germ-layers. In immediately
subsequent stages the descendants of some of these cells retain
this character, while other become flattened in response to the
8 Cf. for interdermal cytodesmata, v. Szily. Anat. Anz., Bd. 24, 1903, p.
417; and Studniéka, Ibid., Bd. 40, 1910, p. 33. For origin of endothelium in the
eat. Fleischmann, A. Embryologische Untersuchungen I, 1889; Schulte, Mem.
Wistar Inst. Anat. and Biol., no. 3, 1914.
9 Am. Jour. Anat., vol. 13, 1912.
10 Mollier, Op cit., p. 1051.
62 H. VON W. SCHULTE
collection of fluid and are modified to endothelium. I should
prefer, therefore, to term Mollier’s first stage of solid cords, so
far as the mammal is concerned, simply mesenchyme. His sec-
ond ‘mesenchymatous’ stage 1s really a mixed condition of mes-
enchyme and endothelial vesicles; it might be termed the stage
Fig. 9 Reconstruction of endothelial heart tubes and adjacent mesenchyme
of a cat embryo of eleven pairs of somites. Columbia Collection, No. 534. X
300. Reduced one-half. 1, Bulb; 2, isthmus, corresponding to bulbo-ventric-
ular sulcus; 3, ventricle; 4, omphalo-mesenteric vein; 5, atrio-venous angle; 6,
mesenchyme between tubes; 7, anterior intestinal portal; 8, oral plate.
FUSION OF CARDIAC ANLAGES IN THE CAT 63
of angiocysts. The third is well designated as that of the endo-
thelial tube with a continuous lumen but in this stage there may
also be separate angiocysts and mesenchyme adjacent to the
tube.
The heart of the embryo of 9 pairs of somites is an ex-
ample of this last mentioned state. In it the endothelial tube
on each side has a continuous lumen not of equal diameter
throughout, it is true, for it has two constrictions and its walls
still show traces of their component angiocysts. The con-
strictions are located at the bulbo-ventricular sulcus and at the
eardio-venous angle. The former reduces the lumen to a nar-
row dorso-ventral cleft, the latter produces a smaller but quite
perceptible diminution of the lumen. The bulb and ventricle
are on the contrary dilated, and markedly so in their transverse
diameter.
Between the bulbs corresponding to the narrow middle cardiac
plate there is room for but little mesenchyme. Between the
divergent ventricles, however, it is more abundant and stretches
across between the endothelial tubes in plates and anastomosing
strands, among which are scattered small angiocysts (fig. 2).
Conditions in the embryo of 11 somites (figs. 9 and 10) differ
only in that the tubes are approximated and nearly parallel in
their whole length. The slight asymmetry of the two sides is
due mainly to the collapse of the shoulder of the left ventricle
and in the cast of the lumen to a similar collapse of the right
omphalo-mesenteric vein, which in other embryos of about this
stage is actually a little larger than the left.
The bulbs are dilated and merge into the ventral aortic roots
and the plexus forming about the foregut. As yet fusion has
taken place across the median line only at one point situated well
forward towards the oral plate. Elsewhere between the bulbs
are strands of mesenchyme in which are two small angiocysts.
Corresponding to the bulbo-ventricular sulcus on each side is
a narrow isthmus compressed laterally.
The ventricles are wide, prolonged into the shoulders of the
mantles and diminishing caudad to their junction with the
omphalo-mesenteric veins. The right in four places shows rem-
64 H. VON W. SCHULTE
nants of the partitions separating its component angiocysts.
On the left but one minute one is present. The separated
lumina at the shoulder of the left ventricle are due to collapse
of the tube. Between the two ventricles is a net work of mesen-
chyme strands which have for the most part a longitudinal direc-
tion. In the model of the lumina but two angiocysts are found
in this area, one median in position, one close to the left endo-
Fig. 10 Lumina of endothelial tubes of embryo shown in figure 9, same scale.
1, Bulb; 2, isthmus; 3, ventricle; 4, omphalo-mesenteric vein; 5, atrio-venous
angle; 6, angiocysts.
thelial tube. The irregular contours of both tubes mesad sug-
gest the addition of imperfectly assimilated angiocysts.
In embryos of 12 and 13 somites the mesenchyme of the middle
cardiac plate is condensed into a single strand which occupies
the concavity of the plate between the ridges which demarcate
it from the mantles (figs. 11 and 12). In the model of the en-
dothelium (not illustrated) this strand extends from the con-
stricted regions or isthmi between bulbs and ventricles as far as
FUSION OF CARDIAC ANLAGES IN THE CAT 65
the anastomosis between the omphalo-mesenteric veins. In it
are three angiocysts. ‘Two are very small; the more caudal of
these is in process of annexation to the right ventricle, the other
is attached by its wall to the left ventricle but does not commun-
icate with its lumen. The largest angiocyst is elongated and
has important connections. It communicates cephalad with
the isthmian region of each bulbus and so establishes the first
Fig. 11 Reconstruction of the lumina of the endothelial heart tubes and adja-
cent vessels of a cat embyro of twelve pairs of somites. Columbia Collection,
No. 547. XX 300. Reduced one-half. 1, Median angiocyst; 2, bulb; 3, ventricle;
4, atrio-venous angle; 5, omphalo-mesenteric vein; 6, forming aortic arches.
continuity of lumen between the endothelial tubes. In addi-
tion it has a small connection with the right ventricle. Thus
the median angiocysts play a role in the coalescence of the en-
dothelial anlages analogous to that of the middle plate in the
fusion of the myoepicardial mantles. For the rest the changes
accomplished in this stage are easily appreciated on comparison
with the heart of the embryo of 11 somites (fig. 10). The cardio-
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 1
66 H. VON W. SCHULTE
venous angles are greatly deepened and the omphalo-mesenteric
veins are brought close together. Between their approximated
portions a small communication has formed. As a whole the
right tube is larger, perhaps more distended than the left. The
right omphalo-mesenteric vein is certainly larger than that of
the opposite side.
Fig. 12 Photomicrograph of section of heart shown in figure 11. X 125.
teduced one-half. 7, Middle cardiac plate; 2, median angiocyst; 3, ventricle.
In the heart of the embryo of 14 somites as marked changes
have supervened in the endothelial tubes as in the myocardium.
The loop is well formed but in its entire length is double, being
composed of two parallel tubes as yet only at the beginning of
fusion. The general configuration is shown in figure 13, in
which the model is seen from behind and slightly from above,
bringing its caudal aspeet prominently into view and to a less
Lied
FUSION OF CARDIAC ANLAGES IN THE CAT 07
degree its ventral surface. The drawing is greatly foreshort-
ened. The bulbar segments are approximated but their lumina
are still completely separated by an endothelial partition. The
right bulb has a slightly greater cross-section than the left (fig.
14) and in its terminal segment the left is slightly irregular in
contour as can be seen in the illustration of the model. These
changes, though small, foreshadow the reduction of the left bulb
Fig. 13 Reconstruction of the lumen of the heart in acat embryo of fourteen
pairs* of somites. Columbia Collection, No. 188. X 300. Reduced one-half.
1, Right bulb; 2, left bulb; 3, left ventricle; 4, right ventricle; 5, fusion between
isthmi; 6, fusion between ventricle; 7, fusion at cardio-venous angles; 8, left
omphalo-mesenteric vein; 9, 1ight omphalo-mesenteric vein.
in later stages. The isthmus of the left side is strongly kinked
by the deepening left bulbo-ventricular sulcus, while the open-
ing out of the sulcus of the right side places that bulb under more
favorable conditions as regards flow. The area of fusion be-
tween the isthmi has greatly increased.
The ventricles are pyramidal in form, tapering toward the
isthmus and caudad and prolonged in strong angular projec-
68 H. VON W. SCHULTE
tions into the shoulders of the mantles. The projection on the
left side is long and pointed, that of the right blunt and short
conformably to the alteration in shape of the mantles at this
ot POE z —
Cems y
age fy
— .
tn
* Seen *
ae Pe
Fig. 14 Photomicrograph of section through the bulbs of embryo shown in
figure 13. > 125. Reduced one-half. 1, right bulb; 2, left bulb; 3, shoulder of
left ventricle myocardium.
period. Coalescence has been effected at about the middle of
the ventricles. Immediately ventrad of this area is a region
where the endothelium of the tubes is not yet in contact and a
perforation extends through the heart from the convexity to the
FUSION OF CARDIAC ANLAGES IN THE CAT 69
eoneavity of the loop (fig. 15). Elsewhere only a partition of
endothelium separates the ventricles.
Fig. 15 Section through ventricles of the same embyro. X 125. Reduced
one-half. 1, Area of coalescence of ventricles; 2, interval between endothelial
tubes; 3, middle cardiac plate; 4, atrium.
Caudad the area of union between the omphalo-mesenteric
veins has increased and extends upon the terminal portion of
70 H. VON W. SCHULTE
the heart tubes, which here are dilated. This enlargement rep-
resents the atrium. It is displaced well to the left of the median
line. This preliminary excursion in a direction opposite to that
occurring in later stages probably depends on the larger size of
the right omphalo-mesenteric vein and seems capable of play-
Fig. 16 Reconstruction of the lumen of the heart in a cat embryo of sixteen
pairs of somites. Columbia Collection, No. 551. X 300. Reduced one-half.
1, Right bulb; 2, remnants of left bulb; 3, left ventricle; 4, right ventricle; 4,
interruption of right tube; 6, atrium; 7, left omphalo-mesenteric vein; 8, right
omphalo-mesenteric vein.
ing a decisive réle in determining the direction which the car-
diac loop will take. For the movement of the venous conflu-
ence to the left favors flow into the left ventricle as against the
right. This becoming engorged thrusts itself strongly against
the shoulder of the left mantle, and the direction of the flow be-
FUSION OF CARDIAC ANLAGES IN THE CAT 71
ing ventrad as well as cephalad tends to throw the left mantle
ventrad and so initiates the axial rotation begun in this stage.
These changes of position on the part of the left heart would
seem to entail as consequences the observed displacements of
the right, the opening out of the right bulbo-ventricular sulcus,
the reduction of the shoulder of the right mantle, and its
rotation dorsad.
But there is an additional factor to be considered. The ac-
centuation of the left bulbo-ventricular sulcus increases the com-
pression of the isthmus of that side and in so far impedes cir-
culation through it. This condition favors the engorgement of
the left ventricle and so participates in producing the effects
enumerated above. However, prior to these events a connec-
tion has been formed between the two isthmi close to their Junc-
tion with the ventricles. When the left isthmus is compressed
this communication serves as a collateral channel leading the
blood stream into the right bulb which from now on exceeds the
left in development. The distension of the right ventricle is
maintained by the interventricular communication already
described. There is need of some arrangement ofthis sort for
the communication of the atrium with the right ventricle is of
smaller caliber than on the left side. This also depends upon
the shift of the venous end of the heart to the left with a conse-
quent marked accentuation of the right cardio-venous angle
and a diminution of angularity on the left.
The excess of the right omphalo-mesenteric vein over the left
seems then the efficient cause of the displacement of the atrium
to the left and this joined with the configuration of the tubes
and mantles at the beginning of the process is capable of afford-
ing a mechanical explanation of the formation and direction of
the loop.
Conditions in the embryo of 16 pairs of somites are corrobo-
rative of the findings in the embryo just described. The lumen
of the heart is shown in figure 16. The right omphalo-mesen-
teric is the larger, the atrium is strongly displaced to the left and
joins the atrio-ventricular canal almost at right angles. Here
an important change has been effected for the right tube, atten-
a2 H. VON W. SCHULTE
uated in the embryo of 14 somites is now interrupted and repre-
sented in the model of the lumen only by a pointed projection
of the atrium. In the model of the endothelium, which was
made of this embryo, there was also a solution of continuity at
this point and in addition to the atrial protrusion shown by the
lumen there was also a small projection of collapsed endothelium
from the ventricle. The ends of the two processes were sepa-
rated by a small but perfectly definite gap.
The fusion between the ventricles has progressed, especially
caudad, and the two gaps in the line of luminal coalescence are
filled with epithelium. There is no longer a foramen leading
between the ventricles as in the 14 somite stage. At the isthmus
also the communication between the two tubes is greatly in-
creased, and from this point on the functional bulb is that of
the right side. The left is interrupted in its continuity and
represented only by irregular projection of the lumen and
patches of cells adherent to the wall of the right tube.
In later stages there is no trace either of the left bulb or the
right atrio-ventricular canal and nothing in these slightly older
embryos as far as they have come under observation is indica-
tive of the history of these regions. The remnants of the inter-
ventricular partition persist awhile but are entirely effaced in
an embryo of 21 somites.
The coalescence of the endothelial tubes in the cat is thus seen
to be delayed until the cardiac loop has been completed. It is
begun just prior to the initiation of the loop at points where
the tubes are bent mesad. In the actual coalescence angio-
cysts developed in relation to the middle cardiac plate are in-
volved. Two segments of the primitive tube are in part sac-
rificed—the left bulb and the right atrio-ventricular canal.
The factor seemingly responsible for the direction of the loop is
the larger size of the right omphalo-mesenteric vein with the
consequent displacement to the left of the region of venous
confluence.
CONCERNING CERTAIN CELLULAR ELEMENTS IN
THE COELOMIC CAVITIES AND MESENCHYMA OF
THE MAMMALIAN EMBRYO
V. E. EMMEL
From the Anatomical Department, Washington University Medical School and the
—= —
rs
ID fe
DV:
Nil.
Vit:
Ae
Institute of Anatomy, University of Strassburg
FOUR PLATES (FORTY-FOUR FIGURES)
PP EOC CLIOMS ee aye cae vee reso ae en a TE ze ake ae tN 74
. Occurrence and distribution of the free cellular elements in the em-
breyani co Olam ssiaree 5 1s cetunren eee ee ese oop, otal gal. ner 74
The coelomic macrophags
i eCytological characueristics sass occa teins ees os eee 76
2. Evidence as to their origin from the coelomic mesothelium ...... 81
a. Certain characteristics of the coelomic mesothelium in
ETOH Rca c ster tes cane Seater: A Na So Ae etal Bema bow 81
b. Free mesothelial cell masses in the coelomic cavities........ 84
c. The mesothelium of the pleuro-pericardial and _pleuro-
peritonealimrembramesier weeny tater eet iene ieieraes 85
Erythrocytic elements in the coelomic cavities...................... 87
is fsrmaeyllll Gorse Sheers lO Ecos ces coon sone obons cobs ueoaabeueac 85
Dee NucleatedneryjubnocwticEcell sees oene per ae sa ae 90
. Erythrocytic disintegration in the mesenchyma..................... 93
1. Degenerative changes in erythrocytic nuclei with reference to... 93
a. The question of the mesenchymal secretion of erythrocytic
ClEMTEM GS ase a OR CET Re ORs teeen eke 93
lore, © ab Oty Semin e sierae mares socke phat ctor fe ody sce ee tee ets ce td 97
2. Degenerating erythroblasts and the so-called eosinophilic leu-
cocytes.of theembryonie mesenchyma,. .......4.....-22.s0.e5 ms 99
Concerning the present status of the question as to the origin of mac-
rophags trom, the coclomic mesothelium... 20. 0.32006 0. s oe 103
ieSmnbryological and comparabivies 2.00 oti. 4: <4 dade eadeen came ee 103
Moe Juae Ko MEM gee ar nee) IS he ae ar reas Se a Pe Pe Ser rt te 104
| ESSIEN, Sak ne Pol SE ae he Ree Se a A RANE ee 109
NESE L ATIC MCLLCE tee rtrays riety NIE eal acs orci cic a ahS See oe Choe ete oe ecie 113
74 V. E. EMMEL
I. INTRODUCTION
The following study is concerned with the free cellular ele-
ments in the pericardial, pleural and peritoneal body cavities.
The results of recent morphological as well as experimental and
clinical investigations regarding the nature and origin of these
structures as they occur in the serous cavities of the adult mam-
mal are of a divergent character. In view of this fact together
with the apparently entire absence as yet of embryological data
bearing on the problem, it has appeared desirable to obtain if
possible more definite information as to the cytological condi-
tions in the coelomic cavities of the embryo. The nature of the
present subject has also necessitated the extension of the study
to include certain cellular structures occurring in the embryonic
mesenchyma.
Part of the following study was made while at the University
of Strassburg during a leave of absence from the Washington
University Medical School. I wish to express my _ indebt-
edness to Prof. G. Schwalbe for the generosity with which the
facilities of the Anatomical Institute were placed at my disposal
and the encouraging interest taken in the work by Prof. Franz
Weidenreich. It ts with regret that in consequence of the pres-
ent disrupted political condition in Europe it has been neces-
sary to forego the pleasure of Professor Weidenreich’s valued
criticism of the final results of the research.
II. OCCURRENCE AND DISTRIBUTION OF THE FREE CELLULAR
ELEMENTS IN THE EMBRYONIC COELOM
The presence of free cellular elements in the embryonic body
cavities were first noted in a 7.4 mm. pig embryo. In order to
ascertain to what extent these elements were normally present,
the observations were subsequently extended to include the fol-
lowing material: 5 to 12 mm. pig embryos, 9 mm. rabbit em-
bryos, 5 to 9 mm. mouse embryos and several 9 mm. opposum
embryos. The specimens were fixed in Zenker-formalin (Helly’s
modification), embedded in celloidin or paraffin and the serial
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO FD)
sections stained with Giemsa or Azur-Eosin in accordance with
the technique developed by Maximow.
In the study of these cellular elements as they occur in the peri-
cardial, pleural and peritoneal cavities, the following precautions
are to be taken into account. In the ease of early embryos in
which the peritoneal cavity js stillin communication with the extra
embryonic coelom, if the umbilical cord has been cut previous
to fixation there is present the possibility of an accidental en-
trance of external elements into the peritoneal cavity through
the cut cord during the fixation and removal of the embryo from
the uterus During the necessary manipulation of staining
and mounting, extra coelomic blood and tissue cells may become
accidentally detached and transferred to the coelomic areas of
the section. In the present study in the case of a doubt as to
confusion with such dislocated cells the data was either discarded
or recorded with a question mark. In the majority of cases,
however, erroneous data arising from such sources can be satis-
factorily eliminated by confining the evidence to such cells or
groups of cells which can be demonstrated to extend through
two or more successive sections in the series, or are clearly em-
bedded in the coagulum of the serous fluid. The results of
such a critical study seem to leave no doubt as to the normal
and constant occurrence of a considerable number of free cel-
lular elements (somewhat variable perhaps in quantity) in the
body cavities of the mammalian embryo.
These free cells are irregularly distributed throughout the
serous fluid. At certain stages of development, i.e., previous
to the closure of the pleuro-pericardial and pleuro-peritoneal
canals, they may be especially abundant in the region of the
developing pleuro-pericardial and pleuro-peritoneal membranes
In general they are not infrequently found relatively more
numerous and aggregated at one side or another of the cavity,
a condition no doubt due in part to the settling of the coelomic
fluids and their cellular content toward one side of the body
during fixation, as is not infrequently observed in the case of the
blood in the heart and blood vessels.
76 Vv. E. EMMEL
In the following account the coelomic cellular elements found
in the present material may, on the basis of their cytological
and functional characteristics, be conveniently described as fall-
ing into two groups: first, the basophilic staining and usually
phagocytically active cells, which, as will become more evident
in the ensuing description, may be designated as the coelomic
macrophags; and second, cellular elements characterized by their
eosinophilic staining qualities and nonphagocytie activity.
Ill. THE COELOMIC MACROPHAGS
1. Cytological characteristics
The majority of the free coelomic elements are embraced in
the first of the above indicated groups. These basophilic cells
may be further roughly subdivided into three types which may
be described as follows:
The first of these types is illustrated in figures 1, 2, 9 and 14
icm. These cells are more or less spherical in form and some-
what smaller in size than those of the other two types. The
nucleus varies from a central to an eccentric position within the
cell, and may be either round or more or less indented on one
side, so as to approximate a kidney shape. The cytoplasm is
typically basophilic in staining reaction, is without any specific
granular structure and may occasionally contain several small
vacuoles. In form, size, nuclear and cytoplasmic structure these
cells appear comparable to certain phagocytic cells occurring
in the embryonic circulation of the same embryos. The cells of
the second type are illustrated in figures 4, 5, 6, 8b, 10, 14 pem.
They are as a rule larger in size and more oval or irregular in
contour. The nuclei are quite eccentric in position and as a
rule more flattened and kidney shaped in form. The cytoplasm
usually takes a much lighter basophilic stain. A distinguishing
characteristic is the phagocytic inclusions contained in the cyto-
plasm. These inclusions consist almost entirely of nuclei and
cell bodies at various stages of intracellular digestion. It is of
interest to note that these inclusions consist largely of red stain-
ing or apparently erythrocytic elements. The majority of the
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 77
free cells of the coelomic cavities belong to this second class.
Cells of the third type are illustrated in figures 14 vem. and 13.
The chief characteristic is that of the highly vacuolated condi-
tion of the cytoplasm. The rounded or flattened nucleus is
eccentric in position. Not infrequently cytoplasmic knobs or
buds are observed projecting from the surface of the cell (figs.
14 and 44). These cytoplasmic buds may vary considerably in
intensity of stain but in all the cases observed they were
basophilic in their stain reaction. The relation of some of
these buds as partially if not entirely detached from the cell
(fig. 446), together with the occasional occurrence of apparently
similar basophilic bodies free in the coelomic fluid, indicates
the possible constriction off of these cytoplasmic processes from
the parent cell. If such is the case the phenomenon appears
comparable to the liberation of detached portions of the cyto-
plasm as described by Weidenreich, ’12, p. 2602 (also Downey
and Weidenreich, ’12, and Downey, ’12).
The question now arises as to whether these cell types repre-
sent genetically distinct kinds of cells or whether these cells are
more or less closely interrelated structurally and functionally.
In attempting to answer this question from sectioned material,
we are necessarily largely dependent upon such evidence as can
be obtained by a comparative cytological study, and endeavor-
ing to ascertain the occurrence or absence of data indicative of
structural intergradations between the cells in question. In
the course of such a study the following results were attained.
Attention has already been directed to the fact that a chief dis-
tinguishing characteristic of the second type of cells is that of
the phagocytic ingestion of other cell bodies. In comparing
cells tem and pem in figure 14 as representative of the first two
groups just described, a considerable size difference is at once
evident. It may be questioned, however, whether this is not
largely a result of functional activities. The ingestion of cell
bodies may be expected to increase the cell size and this is what
is actually found, for the coelomic macrophags containing
three, four or more cellular inclusions are larger than those hay-
ing only a single inclusion, as may be observed for example in
78 V. E. EMMEL
figures 4, 5 and 6. Cells with a single inclusion usually do not
differ greatly in size from those of the first group. Reddish
staining bodies, corresponding apparently to centrospheres, may
be found in both types (figs. 1 and 8a) although less frequently
observed in the more highly phagocytically active cells. The
flattened or more kidney shaped form and more eccentric posi-
tion of the nucleus in the cells of second type is evidently largely
due to the bulging out of the cytoplasm, as the result of the in-
gestion of the cellular inclusions and a correlated eccentric dis-
placement of the nucleus toward the opposite side or pole of
the cell body. Between these two cell types a striking difference
in the intensity of basophilic stain may also be frequently noted.
The present data, however, warrants the conclusion that these
staining differences are correlated to an important degree with
functional activities. All gradations are to be observed between
the strongly basophilic cells, figures 1, 2,9, 14 tcm and the pale
staining cells in figures 6, 10, 14 pem. Furthermore, it appears
of significance that the macrophags containing the larger number
of cellular inclusions are almost always the paler in cytoplasmic
stain. After the study of a large number of these cells one is im-
pressed with the suggestion that this may, in part at least, be a
direct result of intracellular digestive functions. Certainly there
can be no question but that the digestive chemical interactions
between the phagocyte and cellular inclusion, whatever their
exact nature may be, are of such a character as to change the
ingested erythrocyte from a bright red staining cell to a practi-
cally colorless non-staining mass (figs. 5,6, and 8). In case of
an ingested erythroblast the erythrocytic nucleus may first be-
come fragmented (figs. 4 and 5) or else manifest progressive
changes from center toward its periphery (figs. 6n and 33e, also
p. 95), but in either case it eventually experiences a practically
complete loss of its staining properties. On the other hand, how-
ever, may not this digestion also involve correlated changes in
the staining substances of the phagocytic cell? The frequent
association of a pale staining cytoplasm with maximum phago-
eytic activity strongly indicates that this is the case. If this
conclusion is correct it offers an explanation for the basophilic
differences between the two types of cells under consideration.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 79
No evidence was obtained as to the exact nature of this change
in the staining qualities of the cytoplasm of the phagocyte
whether it is the result of an internal absorption of digestion
products or whether it is due to a modification or reduction of
cytoplasmic elements of the phagocyte itself as they participate
in the digestive processes. In either case it is remarkable how
constant and sharp is the demarkation between the cytoplasm
of the phagocyte and that of the ingested cell, and that one is
able to detect little if any difference in the cytoplasm of the
phagocyte in the immediate vicinity of the inclusion as com-
pared with that in the more remote parts of the same cell body.
It is to be observed that these color changes in the phagocyte
are apparently not confined exclusively to its cytoplasm for the
nucleus also may become much lighter in the highly active
macrophags, a change involving apparently the nucleoplasm
rather than the chromatin (ef. figs. 1, 2, 9 and 14 icm, with 5,
6, 10, and 14 pem). These cytoplasmic and nuclear differences
appear especially well demonstrated in figure 8. These two
cells frcm the pericardial cavity of a 9 mm. pig embryo were ly-
ing side by side in the manner drawn. Consequently there can
be no question of variation in fixing or staining technique as factors
in their staining differences. In (a) both nucleoplasm and cyto-
plasm are quite basophilic, while in (6) whichis at a comparatively
much higher stage of phagocytic activity, nucleoplasm and cyto-
plasm are both much paler in color. The same observation also
applies to figures 9 and 10 which were taken from two consecu-
tive sections of a second 9 mm. pig embryo. In examining the
literature bearing upon this subject it is of interest to note that
similar changes in the basophilic character of the cytoplasm is
described by Downey (710) in the phagocytes of the lympho-
renal tissue of the ganoid fish, Polyodon spatula, as indicated in
the statement that ‘as seen in Polyodon these cells are strongly
basophilic right after phagocytosis (fig. 2). As the phagocytosed
erythrocyte breaks down the cell gradually becomes pale and
often metachromatic” p. 85.!
1 Kyes (15, p. 546) in a recent paper also describes in the liver of birds varia-
tions in nuclear structure and cytoplasmic staining reaction of endothelial cells
as correlated with the intra-cellular digestion of ingested erythrocytes.
SO Vv. E. EMMEL
Turning finally to the cells of the third group it will be ob-
served that their most distinguishing characteristic is the pres-
ence of cytoplasmic vacuoles. The nuclei may be round or flat-
tened, are usually more or less eccentric in position and do not
differ materially from the nuclei of the phagocytically active
macrophags. The vacuoles vary greatly in number and size,
so that while in some cases only one or two may be observed, in
other instances they are sufficiently numerous to fill almost the
entire cell body (figs. 14 vem. and 13). While many of these
vacuoles are filled with an apparently homogeneous material
which is non-staining with Giemsa, others contain remnants of
hemoglobin and nuclear staining elements. In comparing such
cells as figure 5 with its almost entirely digested erythrocytic
inclusion, figure 6 and 14 pem., with large vacuoles containing
just a trace of ingested material, and figure 8b in which one of
the vacuoles contains a cellular remnant while the two others
are practically clear, there can hardly be any question but that
many if not the majority of these vacuoles have arisen in con-
nection with intracellular digestive processes. Upon the com-
plete transformation of the chromatic elements of the ingested
erythrocyte there may thus still remain, for a time at least,
a non-staining vacuole-like structure in the cytoplasm of the
phagocyte.
In resume it appears, therefore, that the size, form, nuclear
and cytoplasmic difference between these three types of cells
are to be regarded as correlated with variations in the degree
of differentiation and functional activity rather than as indica-
tive of differences of a more fundamental character. The most
prominent function of these cells being that of phagocytosis,
they may not inadequately be designated as the coelomic macro-
phags of which the cells of the first type present the least differ-
entiated and least active stages and the cells of third type end
stages in functional activity. That these coelomic cells are not
only manifesting normal functional changes and cytological
differentiation but are also undergoing cell multiplication is
positively demonstrated by the not infrequent occurrence of
mitosis (figs. 3 and 7). Indeed in some instances it appears that
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 81
the macrophags may undergo mitosis even while still retaining
incompletely transformed remnants of previously ingested mate-
rial in its cytoplasm (fig. 11).
2. Evidence as to their origin from the coelomic mesothelium
a. Certain characteristics of the mesothelium in general. Are
these macrophags of the coelomic cavities cells which have mi-
grated into these body spaces from the neighboring blood vessels
or are they tissue cells which have become detached or liberated
from the tissue walls surrounding these cavities where they have
undergone further differentiation and assumed phagocytic and
possibly other functional activities in the serous fluids? As to
the first view, while no conclusive evidence of such a migration
of macrophags from the blood vessels and adjacent tissues was
obtained, it is to be recognized as not improbable that such cells
may enter the embryonic coelom in this manner as has been
maintained by Maximow to occur in the serous cavities of even
the adult mammal. But the crucial question still remains as
to whether this is to be regarded as the only source of the coe-
lomic macrophags in the embryonic body cavities.
The embryonic mesothelium consists of cells which are rather
flattened in form. The nuclei also have a correlated flattened
oval shape and cell walls are not clearly evident. Typically
these cells form a single epithelial layer lining the coelomic cavi-
ties (mes in figs. 15 to 18 and 41 to 44). Sucha layer is, however,
by no means always sharply defined, for in various regions the
surface cells are in such an intimate syncytial association with
the deeper lying cells that characteristic structural differences
between them are not readily evident, indeed in certain regions
the conditions are such as to suggest that the surface meso-
thelial cells may have given rise to many deeper lying cells
comparable to the endothelial growths described by Mall (12,
pp. 258, 261) in certain endocardial regions of the heart. In a
careful study of the mesothelium as seen in serial sections, it
may be noted that the form and structure of its component
cells are not always constant throughout the body cavities. In
various regions the flattened mesothelial cell body, as well as
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 1
82 Vv. E. EMMEL
its nucleus are observed to have assumed a more rounded form
and not infrequently a more basophilic stain reaction. Instances
of such changes involving only single cells or small groups of
cells are especially evident in the visceral epicardium. Cells
may be observed in which the nuclei instead of being oval or
flattened are more rounded or spherical in shape and not in-
frequently indented in a kidney shaped manner (figs. 17 and 42).
The cytoplasm also is rounded up, the cell as a whole projects
from the mesothelial surface and in some instances is attached
by only a slender basal cytoplasmic process (fig. 41). It may be
observed that as a rule the cytoplasm of such cells also presents
a more basophilic stain. Many of these cells have every ap-
pearance of being destined to become subsequently detached
from the coelomic wall, and it appears evident that when liber-
ated into the body cavity they would be practically indistin-
guishable in either nuclear or cytoplasmic structure from the
macrophags already present in the serous fluid. The important
point that the mesothelial cell may function in a phagocytic
manner seems clearly demonstrated in figure 42 showing eryth-
rocytic inclusions in the mesothelial cytoplasm. Figure 44
represents a section through a region of the visceral epicardium
in which the proliferative activities involve a larger number of
cells. Many of the apparently recently liberated cells are pha-
gocytically active, present vacuoles and bud-like cytoplasmic
processes and appear identical with the typical coelomic
macrophags.
Are we to conclude from such data that the coelomic meso-
thelium is really giving rise to coelomic macrophags? In view
of the close approximation of many of the mesothelial cells to
the macrophags in both their cytological characteristics and
potential phagocytic functions, as observed with the present
technique, it would not appear a great step to their differentia-
tion into such cells. But a substantial proof of such a process
is a more difficult matter especially from fixed material where
all the intermediate stages in a given case cannot of course be
directly observed. In evaluating the present data it seems
clear, however, that the form and surface relations of the cells
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 83
just described are at least not to be discarded as being merely
misleading appearances due to tangential planes of section as
ean be ascertained from the serial sections. The possibility
was also considered as to whether such cell forms may not be
due to a shrinkage or folding of the coelomic walls or artificial
ruptures of the mesothelial surface. All of these artificial con-
ditions may of course occasionally occur in consequence of the
histological technique employed but the rounded projecting
cells do not have the appearance of elements artificially torn
from the living mesothelium. If they are the products of his-
tological shrinkage, it is not easy to understand how single iso-
lated cells could be made to assume the present forms, for such
cells may be found in regions of the coelomic wall where the reg-
ular surface curvature is not indicative of any artificial irregu-
larity in its reaction to fixing agents. When found in regions
where the surface of the wall appears more irregular, the cell
types in question may occur on both convex areas of the slight
surface elevations as well as in the concave areas of the meso-
thelial depressions. What appears a criticism of a more seri-
ous character, however, is the possibility that the cell forms
under consideration are either instances of cells from the adja-
cent mesenchyma and blood vessels migrating through the meso-
thelium to reach the coelomic cavities or else coelomic macro-
phags themselves merely resting upon or in close proximity to
the mesothelial surface. That a coelomic macrophag may occa-
sionally be caught in the coagulum of the serous fluid and fixed in
close contact with the coelomic surface must be granted. Fur-
thermore since one is apparently forced to admit that erythro-
cytic elements, as will be subsequently described, must enter
the coelom from extra coelomic regions it seems necessary to
admit the possible migration of other cellular elements into the
embryonic coelom. Much of the evidence also tends to be viti-
ated by the absence of decisive cytological differences between
many of the mesothelial cells and either the coelomic macro-
phags or similar cells occurring in the vascular channels.
In view of the above difficulties it becomes necessary to ascer-
tain whether there exists any other sources of evidence condu-
84 Vv. E. EMMEL
cive to a more convincing conclusion regarding the problem.
The results of such a further study may be presented in the form
of two groups of data: the first referring to the character of cer-
tain cellular masses found free in the coelomic cavities and the
second to the structure of the mesothelium in certain regions of
the coelomic walls. .
b. Free cell masses in the coelomic cavities. In addition to
single free coelomic cells there are also found certain interest-
ing groups or masses of cells (figs. 15 and 16). Such masses are
of fairly constant occurrence. Some of them may consist of
only two or three component cells (fig. 15) or of a much larger
number of cells as in figure 16 which is a section of a mass
extending through as many as five sections.
These masses cannot be said to represent merely a loose aggre-
gation or agglutination of otherwise free coelomic macrophags,
for upon closer study it may be observed that as a rule their
component cells are organized into a definite peripheral or epi-
thelial layer with an occasional cell more centrally situated.
Consequently as seen in section, the larger masses present the
appearance of epithelial rings surrounding a lighter and less
cellular core. The majority of the cells in these masses, espe-
cially the more peripheral ones, are indistinguishable in both
nuclear and cytoplasmic structure from the mesothelial of the
adjacent coelomic walls. In other words there is every indica-
tion that these masses are groups of mesothelial cells in which
the mesothelial character of the cells are still clearly evident.
"Upon the careful examination of serial sections it can also be defi-
nitely established that the majority of these masses (such as
shown in figs. 15 and 16) are entirely free in the coelomic cavity,
nor do they present the appearance of having been artificially
separated from the coelomic walls. Occasionally such masses
are, however, still attached to the coelomic wall and their rela-
tions in this case are such as to indicate that the mesothelium
has grown out into the lumen of the coelom in the form of a
papillary projection which may subsequently become detached.
We have therefore, what appears to be a clear case of the actual
separation of mesothelial cells from the coelomic wall and lying
free in the coelomic cavity.
s
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 85
The important question next arises as to the fate of these
mesothelial cell masses. How long they may persist as defi-
nitely organized structures the present data does not indicate.
While mitotic figures were not observed neither was there any
evidence of degeneration noted. That their component cells
may assume phagocytic activities is indicated in the smaller
of the two masses in figure 15. But the point especially perti-
nent to our present purpose is the fact that some of these masses
show evidence of further modification and subsequent disinte-
gration into isolated free cells. In the smaller of the two groups
shown in figure 15 the nuclei are more rounded in shape and the
epithelial character of the component cells is no longer so clearly
evident. The single isolated cell in the same figure is appar-
ently of the same type except that the nucleus has assumed a
kidney shape and the cytoplasm a deeper basophilic stain. These
changes seem strikingly shown in figure 16 in which several of
the peripheral cells have become more spherical in form and the
cytoplasm takes a darker stain. Of the two approximately
spherical cells in the lower part of the figure, one is partially and
_ the other almost entirely free from the main mass. Such cells
appear quite comparable to the basophilic cells or macrophags
found free in the coelomic cavities.
Granting the correctness of the above conclusions, these dis-
integrating mesothelial cell masses, therefore, furnish a type of
evidence indicative of .the cytological transformation of meso-
thelial cells into coelomic macrophags which appears to obviate
the possible objections previously noted with reference to simi-
lar changes at the surface of the coelomic wall. For it is evident
that the rounded cells at the periphery of these masses can cer-
tainly not be regarded as migratory cells from mesenchymal or
vascular sources and it is highly improbable that they represent
free coelomic macrophags incidentally resting or fixed at the
surface of the mass.
c. The mesothelium of the plewro-pericardial and pleuro-peri-
toneal membranes. ‘The second of the two sources of evidence
referred to on page 84 is found in the regions concerned with the
subdivision of the embryonic coelom by the development of the
86 Vv. E. EMMEL
pleuro-pericardial and _ pleuro-peritoneal membranes. ‘These
membranes as can be directly observed, are localized centers of
increased mitosis and cellular growth. Without taking into ac-
count here the deeper lying causes it appears evident that this
increased cellular proliferation is an important factor in the
gradual outward extension of the coelomic walls or membranes,
the final fusion of the juxtaposed surfaces of which is destined
to effect a closure of the pleuro-pericardial and pleuro-peritoneal
canals. In such regions of fusion the free mesothelial surfaces
necessarily disappear. Consequently, if mesothelial cells can
differentiate into coelomic macrophags such regions might be
expected to furnish valuable evidence of such a process.
Figure 43 is one of fourteen sections of such a region in a 7 mm.
pig embryo, showing the embryonic pleural cavity (pc) and the
pleuro-pericardial canal (pplc). At the center of the figure is
seen a section through the distal border of the pleuro-pericar-
dial membrane (pp) for the left pleuro-pericardial canal (the
embryo having been cut in the sagittal plane). Figure 14 shows
the same central mass drawn at a higher magnification. If this
cellular mass is traced back through the fourteen sections in
which this membrane is present to its attachment to the parietal
wall, its component cells are found to merge and become con-
tinuous with the mesothelial and mesenchymal elements at the
juncture of the parieto-pleural and parieto-pericardial walls.
Directing. attention more especially to the present object of in-.
quiry it is important to note that toward the more peripheral or
distal margins of these advancing membranes the superficial cells
have a more nearly spherical shape and that a definite flattened
mesothelium is no longer evident. In the more central portion
of the section shown in figure 14 there may be recognized a some-
what more definite layer of mesothelial cells (mes) surrounding a
central core (c). Toward the periphery of the section, however,
‘the cells are no longer so intimately united and many of them
are partly if not completely detached as free rounded cells.
The structural characteristics of such an area is not indicative
of degenerative changes. On the contrary the frequent mitotic
figures (m) furnish ample proof of active cell multiplication and
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 87
there is abundant evidence of phagocytic activities (pem), intra
cellular digestion and vacuole formation (vem). It may be fur-
ther observed in these regions that the more superficial cells are
‘not only modified in form, but that the cytoplasm also tends to
take a deeper basophilic stain. Such cells as icm and pcm are
’ certainly identical in structure with the coelomic macrophags.
At the same time there are present all transitional stages be-
tween these cells and the original mesothelial elements. Fi-
nally beginning at these proliferating areas, detached free cells
of the same type may be traced through successive sections out
into the adjacent body spaces as they become scattered through-
out the pericardial, pleural and peritoneal cavities where they
appear indistinguishable from the free cells normally present in
these regions. Evidently here again, just as in the case of the
free mesothelial cell masses previously described, the proliferat-
ing and disintegrating mesothelium is giving rise to free cells
functioning as macrophags in the coelomic cavities.
In resume it may be stated, therefore, that the data derived
from both the pleuro-pericardial and pleuro-peritoneal mem-
branes and the free mesothelial cell masses support in a sub-
stantial manner the conclusion suggested by the cytological con-
‘ ditions observed at the surface of the coelomic walls. Namely,
that the coelomic mesothelium is an important source of the
phagocytic cells found in the embryonic coelom, and that as
these mesothelial cells round up and become detached from
various regions of the coelomic wall, they assume structural and
functional characteristics identical with that of the coelomic
macrophags typical of these embryonic body cavities.
IV. ERYTHROCYTIC ELEMENTS IN THE COELOMIC CAVITIES
As already indicated (p. 76) the second group of cellular coe-
lomic elements are characterized by their eosin staining quali-
ties and non-phagocytic activity. These eosin staining ele-
ments again fall into two sub-groups, the one consisting of small
non-nucleated bodies and the other of larger nucleated cells.
88 V. E. EMMEL
1. Small eosin staining bodies
In figures 14 and 20 several small round red bodies (e) are to
be observed about a third or less than a third the size of an aver-
age erythrocyte. Such bodies were found more or less constant
in young pig, mouse and rabbit embryos although they may .
vary considerably in number in different specimens. These
structures may be observed lying free anywhere in the coagu-
lated coelomic fluid or, as is more frequently the case, in con-
tact with the coelomic walls. As a rule they are quite round,
sharply defined and take a brilliant red stain with Giemsa and
Azur-eosin. Upon careful focus they sometimes present the
appearance of a slightly clearer central area. They are non-
granular in structure. Occasionally similar bodies are also found
in the circulating blood.
As to the nature of these bodies, the first suggestion to pre-
sent itself is that of cytoplasmic fragments of disintegrated
erythrocytes. Upon closer examination, however, it may be
observed that a narrow basophilic rim can in many cases be
detected at the periphery of structures in question (figs. 21,
24 to 26). Consequently, without excluding the possibility of
their partial or even entire cytoplasmic character in some in- -
stances, the latter observation necessitates the identification of
the majority of these bodies as elements other than merely eryth-
rocytic cytoplasmic fragments, as will be presently more fully
discussed. A second possibility to be considered is that of their
identification: with cytoplasmic buds constricted off from coe-
lomic macrophags. Downey (13 p. 42) in a description of the
detachment of cytoplasmic buds from lymphocytes and large
mononuclear cells in the lymph gland of the rabbit, found that
these detached bodies may vary greatly in their basophilia and
concluded ‘‘that after separation from the cell the irregular
masses assume a spherical shape and that they gradually lose
their basophilia.”” It was thought that possibly similar changes
might account for the eosinophilic bodies in question, but no
satisfactory evidence was obtained demonstrating transitional
stages between such basophilic bodies and these intensely eosin-
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 89
stained structures. Some of the bodies may be much paler than
others, but the sharply defined basophilic rim which can still be
observed in many cases seems to render them structurally dif-
ferent from the hyalin bodies derived from lymphocytes and
mononuclear cells.
While the present data is consequently negative as to the
derivation of these eosinophilic bodies from the cytoplasm of
either erythrocytes or macrophags, it does, however, furnish
evidence of a more positive character as to another conclusion
concerning their origin. Concerning the basophilic material
already described it appears significant to note that aside from
the form of a narrrow peripheral ring it may present pronounced
additional accumulations in the form of either irregular masses
(figs. 22, 23), one or more delicate crescentic masses (fig. 21)
or small, compact and more or less centrally situated spherules
(figs. 25, 26). In studying erythrocytic nuclei (p. 597) condi-
tions are met with which are strongly suggestive with reference
to the present question. Not infrequently both free and in-
gested erythrocytic nuclei are observed undergoing changes in
which instead of becoming more or less compact in a single
pyknotie body or broken up into several compact nuclear frag-
ments the nucleus becomes lighter colored at its center and has
a dark staining periphery (figs. 34 and 36). In such cases it
appears that the dissolution or chemical modification of the
chromatin proceeds from the center towards the periphery in
such a manner that a stage may be reached in which the baso-
philic staining material remains as only a very thin peripheral
envelope, the interior of which may take an eosin stain of such
a character as to render it practically indistinguishable from the
hemoglobin containing cytoplasm of an erythrocyte. Jolly
(07, p. 245) has further shown that under certain circumstances
these changes may proceed to a complete tinctorial transforma-
tion of the entire chromatin content to oxychromatic staining
material. The close approximation of the cytological appear-
ance of such highly modified nuclei to that of the eosin staining
bodies in the coelomic cavities appears to justify the conclusion
that the latter are also of a similar character. Concerning the
90 Vv. E. EMMEL
occurrence of such degenerating erythrocytic nuclei in the
coelomic cavity two possible sources of origin may be noted:
first, through the disintegration of the erythroblasts occurring
in the coelom itself as will be presently described, and second,
through the occasional passage or elimination into the coelomic
cavities of the degenerating erythrocytic nuclei frequently found
in various regions of the coelomic walls. The latter possibility
would be in accord with the fact that these eosin staining bodies,
as already noted, are frequently found in intimate contact with
the surface of the coelomic walls.
2. Nucleated erythrocytic cells
Figure 18 illustrates the second sub-group of coelomic cellular
elements. In contrast to the coelomic macrophags, these cells
have an eosinophilic instead of a basophilic cytoplasm. The nu-
cleus also may be more irregular in form, lobulated or even sub-
divided into two or more almost wholly if not entirely separated
segments (fig. 19). The cytoplasm is frequently vacuolated.
The cell as a whole may be either round or more irregular in
shape with the peripheral cytoplasm presenting a fragmented
appearance. In the material studied such cells were most fre- -
quently observed in rabbit embryos although they were also
present in both mouse and pig embryos. They may occur as
isolated elements or in groups consisting of two or three to a
dozen cells.
Upon first examination the polymorphonuclear character of
some of these cells is suggestive of leucocytic elements. On the
other hand in no instance was there any special leucocytic gran-
ulation detected. On the contrary the cytoplasm is of a homo-
geneous structure and in many instances (figs. 18, 19) the stain-
ing reaction of both cytoplasm and nucleus is apparently iden-
tical with that of the typical erythrocytes in the same embryo.
Concerning the origin of these cells the possibility was consid-
ered as to their representing a partial or abortive tendency
toward erythrocytic differentiation on the part of the coelomic
macrophags. There can be no question but that such cells as
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 91
(a) in figures 18 and 19 present nuclear and cytoplasmic char-
acteristics readily comparable to the erythrocytes to be found in
the adjacent blood vessels, although this is by no means so clear-
ly evident in such cells as (c), figure 18. In view of the recent
observations of Haff (14) and Bremer (’14) indicative of a par-
ticipation of mesothelial cells in the origin of blood islands and
blood cells it would indeed not appear so remarkable if the same
cells liberated into the coelomic cavity would here be found to
manifest a potentiality for erythrocytic differentiation. How-
ever, it must be admitted that no convincing evidence of such a
differentiation was obtained. The irregular form and structural
character of the nuclei of these cells is not typical of develop-
ing erythroblasts, the nuclei of which during the earlier phases
of their differentiation are normally more or less spherical in
shape (Emmel 714). No evidence of phagocytic activity as a
characteristic which might justify associating these cells with
the coelomic macrophags was observed, nor is the vacuolated
and fragmented condition of the cytoplasm in many of these
cells suggestive of progressive erythrocytic differentiation.
On the contrary the following considerations support a dif-
ferent conclusion. As already indicated wherever these eosin
staining cells are found in groups, some of the cells in such a
group can almost always be clearly identified as erythrocytic.
It is equally evident that in the case of erythrocytic degenera-
tion, the nuclei (i.e. of erythroblasts) may undergo form changes
identical with those to be observed in these eosin staining cells.
Such a lobulation and subdivision of erythrocytic nuclei can not
infrequently be found even in the circulating blood (fig. 37 and
the subsequent description on p. 100). Weidenreich (’03, p.
420) has fully described degenerative changes in which the eryth-
rocytic nucleus becomes irregular in form, indented, bilobed,
dumb-bell and clover leafed shaped, and finally constricted into
two or more parts connected by a small thread-like strand or
entirely separated from each other and thus give rise to a so-
called double nucleated cell. In some cases such degenerating
nuclei may become smaller, more compact and take a much
darker stain, in other instances, however, the nucleus may main-
92 . Vv. E. ‘-EMMEL
tain a comparatively open chromatin network, as is well shown
in figure 33 of Maximow’s (’09) work. Such modified nuclei are
indistinguishable from the irregularly lobulated nuclei of the
eosin staining cells in the coelomic cavity. In other words on
the basis of the present data the cells in question are evidently
correctly interpreted as degenerating nucleated erythrocytes, in
which the nuclei are greatly changed in form, the cytoplasm
having become deficient in hemoglobin, stains a paler color with
eosin and presents the vacuolated condition described by Minot
(12, p. 511) as preliminary to further changes in certain types of
erythrocytic disintegration. Here and there fragments of such
disintegrated cells can be readily found. Occasionally a cell is
observed with a striking peripheral fringe of eosin staining mate-
rial. In some cases this material has the appearance of frag-
ments of disintegrated cells incidentally resting against or adher-
ing to the cell in question. In other instances the union with
the cell body is so complete that a question arises whether it
may not represent a phase in the degeneration of the hemoglobin
containing cytoplasm of the erythroblast (fig. 12).
How these degenerating erythrocytes come to be situated in
the coelomic cavities is more difficult to determine. That cells
with erythrocytic characteristics are normally present constant-
ly in the embryonic serous cavities appears positively demon-
strated by the character of the cellular inclusions in the macro-
phags. In the absence of conclusive data as to their differentia-
tion in situ there remains the alternative assumption that under
various conditions they may escape from the blood vessels and
pass through the coelomic mesothelium into the coelom. Very
young erythroblasts may possibly ‘do this by an active migration
although evidence of such a migration was not obtained. It may
be noted that erythroblasts may also be found within such extra
vascular spaces as the lumen of the Wolffian tubules, occasionally
in the lumen of the developing lung buds, and in mesenchymal
spaces (p. 603) throughout the embryo.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 93
V. ERYTHROCYTIC DISINTEGRATION IN THE MESENCHYMA
1. Degenerative changes in erythrocytic nuclei with reference to:
a. The question of the mesenchymal secretion of erythrocytic
elements. The fact that degenerating erythroblasts, especially
their nuclei, may undergo changes resulting in the production
of small red spheres or masses with either central or peripheral
accumulations or remnants of basophilic material as observed in
the coelomic cavities, merits further consideration with refer-
ence: first, to the bodies interpreted by Maximow as mesenchy-
mal secretions, and second, to the ring bodies of Cabot. 3
Concerning the possible relation of these eosin bodies to mes-
enchymal activities it may be noted that Maximow (09, p. 513)
describes the observation in the blood of the embryonic rabbit
of apparently similar eosin-basophilic-droplets which he states
soon disintegrate in the circulating plasma. Maximow inclines
to the conclusion that these bodies are secretion products of
mesenchymal cells and represent an abortive or precocoius dif-
ferentiation of erythrocytic or hemoglobin containing elements.
Evidence is advanced for the occurrence of such a secretion in
various regions of the mesenchyma, such as that of ~septum
transversum, in the head region and adjacent to the distal ends
of growing blood vessels. In these regions he records the ob-
servation of many large and small red and blue stained spherical
or angular bodies generally embedded in clear vacuoles in mes-
enchymal cells. These bodies are described as consisting of
red spheres containing one or more central blue spherules, red
bodies furnished with one or more deep blue peripheral cres-
cents or caps, or blue rings filled with a clear eosin stained con-
tent (p. 500). After the consideration of several possibilities
as to their origin it is decided that they are probably elaborated
in situ in the mesenchymal cytoplasm and it is stated that one
ean observe how the inclusion in the cytoplasm of the mesen-
chymal cell develops from a few initial small erythrocytic gran-
ules and how it grows in size, and the bee substance
appears within it or on its surface.
94 V. E. EMMEL
Since these structures are so closely similar to the small eosin-
ophilic elements occurring in the coelomic cavities it became
necessary to reexamine the evidence for an intra-cellular origin
of such bodies. For the writer the subject had also an addi-
tional interest in consequence of a previous study of the cyto-
logical differentiation of erythrocytes in which there was occa-
sion to consider the possible origin of erythrocytic or hemo-
globin containing elements in the cytoplasm of the mesenchymal
cell (Emmel 714).
As already indicated, the preceding results of the present in-
vestigation with reference to the eosinophilic bodies in the
coelom were negative as to their origin as intra-cellular secre-
tions. Furthermore after careful study I have been unable on the
following grounds to convince myself that the bodies described
by Maximow in the mesenchyma necessarily represent intra-
cellular secretions of mesenchymal cells. In the first place
these eosin staining bodies are found equally as abundant and
indeed frequently even more so within the ectodermal tissue of
the brain wall (fig. 33e) cranial and spinal ganglia (fig. 28) and the
entoderm of the growing lung buds and digestive tube, in situ-
ations where they would be least expected if they are derivatives
of mesenchymal cells. They also occur in inter- as well as intra-
cellular situations.
Figure 31 is from the mesenchyma in the ventral thoracic
wall of a 9 mm. pig embryo and figure 32 from the mesenchyma
of the septum transversum of a 7 mm. pig embryo. It will be
observed that in both cases practically all the eosinophilic
bodies in question are clearly situated in inter-cellular mesen-
chymal spaces. Similar relations can also be readily demon-
strated in the brain wall and the cerebrospinal nerve ganglia
(fig. 28). On the other hand similar bodies can also be found
which appear unquestionably situated within the cytoplasm
of cells in the mesenchyma. Since these bodies are both inter-
and intra-cellular in position it seems clear from this aspect of
the subject that they may be as adequately interpreted as either
extra-cellular elements, some of which may have become phag-
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 95
ocytically ingested, or originally intra-cellular elements, some
of which had been subsequently extruded from the parent cell.
‘ The point then to be established is the correctness of one or
the other of these two alternatives. The abundant occurrence
of these elements in such tissues as the brain wall and nerve
ganglia which are relatively deficient in mesenchymal cells is a
fact in itself sufficient to raise a question as to their intra-cellular
origin. Again upon examining the bodies in question it will
be observed that they may present a variety of structural forms,
among which may be noted single small basophilic spherules
surrounded by just a trace of eosin staining material (fig. 31a),
several basophilic spherules situated in a larger red staining
body (fig. 316), a crescent, or less frequently, a spherule of baso-
‘philic substance peripherally located (figs. 31, c, and d and 382),
or a peripheral basophilic ring completely surrounding the red
stained material (figs. 3le, 33e, and 28). Now if one turns to
the circulating blood of these same embryos, degenerating nu-
cleated erythrocytes are occasionally found, as is well known, in
which the disintegrating nuclei have become broken up into sev-
eral small more or less rounded fragments. Again in other in-
stances such degenerating nuclei, especially in the mouse em-
bryo, present the appearance of a red stained central area sur-
rounded by a peripheral basophilic ring as shown in figures 34a,
35 and 36. It may also be observed in the same figures that
portions of these rings may be very thin while other areas are
much thicker and present the form of basophilic crescents.
Similar degenerating nuclear changes can also be demonstrated
in phagocytically ingested erythrocytic nuclei as partially shown
in figures 4, 6 and 8b. Now it is well known and can be readily
verified that erythrocytes not infrequently escape from the
blood vessels and become isolated in the mesenchymal and other
tissue spaces of the embryo where they may undergo various
types of disintegration (Minot, 712, p. 509). From the data
derived from the erythrocytes occasionally degenerating in the
circulating blood it becomes evident that these degenerating
corpuscles may assume structural appearances identical in char-
96 Vv. E. EMMEL
acter with the bodies under consideration. Perhaps the most
constant difference between the bodies in the tissue spaces and
the degenerating corpuscles in the vascular channels is the pres-
ence of only a small amount or frequently the entire absence of
any cytoplasm peripheral to the basophilic rings or spherules in
the case of the tissue spaces as compared with the conditions in
the degenerating corpuscles of the blood. But this appears
readily accounted for on the basis of a more rapid and earlier
disappearance of the peripheral cytoplasm of the erythrocytes
degenerating in the environment of the inter-cellular fluids.
Indeed evidence of such peripheral cytoplasmic changes may
be encountered even in the vascular channels as illustrated in
figure 36 from a 9 mm. pig embryo in which one of the eryth-
rocytes shown contains only a relatively narrow rim of cyto-
plasm peripheral to the nuclear ring, whereas it is entirely ab-
sent in the remaining two cells. In figure 34a the peripheral
cytoplasm of the degenerating erythrocyte is much paler than
that of the adjacent normal corpuscle. Instances in the cir-
culating corpuscles of nuclear rings without any evident periph-
eral cytoplasm is demonstrated in figure 29 from the heart
blood of a 7 mm. pig embryo. It appears evident, therefore,
that in pig, rabbit and mouse embryos all transitional stages
ean be found between degenerating nucleated erythrocytes and
the eosin. staining bodies in the embryonic tissue spaces.
A further possible source of origin of many of these bodies
which may be noted, especially in older embryos in which non-
nucleated erythrocytes are beginning to appear, is in connection
with the formation of non-nucleated erythrocytes. In a pre-
vious study of the pig embryo (Emmel 714) evidence was ad-
vanced indicating the origin of non-nucleated red blood cor-
puscles by a process of cytoplasmic constriction resulting in the
separation of the original erythroblast into a non-nucleated re-
mainder consisting of the erythrocytic nucleus together, not 1n-
frequently, with a small amount of cytoplasm remaining from
the parent cell. This nucleated remainder may present an ap-
pearance practically identical with that of many of the eosino-
philic bodies under discussion.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 97
On the basis therefore, of their distribution, structure and close
correspondence cytologically to both disintegrating erythro-
cytes as observed in the circulating blood, inter-cellular tissue
spaces and phagocytic inclusions and to erythrocytic nuclei
persisting after the formation of non-nucleated erythrocytes,
the conclusion is drawn that the small eosin staining bodies in
the embryonic mesenchyma are correctly interpreted as con-
sisting primarily of disintegrating, and in many cases phagocyti-
cally ingested erythroblasts which have escaped into the em-
bryonic tissue spaces ‘and second, especially in older embryos,
of nucleated erythrocytic bodies resulting from the cytogenetic
processes involved in the formation of erythro-plastids, rather
than the products of secretory or other cytological activities in
mesenchymal cells.
b. Cabot’s rings. The second phase of the present subject is
concerned with the question of Cabot’s rings. The erythro-
eytic ring-like structures, first observed by Cabot (’03) in anemic
blood and which are now known to also occur under other ab-
normal conditions such as obtain in leukaemia and lead poison-
ing, are usually described as staining red or reddish violet with
Giesma. Both Cabot (p. 455) and Naegeli (12, p. 152), how-
ever, also record the occurrence of blue stained rings. These
rings which have been interpreted as nuclear elements, possibly
in part nuclear membranes, (Schliep, 707, p. 455) are regarded
as occurring only in the pathological blood of the adult organism
and never in either the human or mammalian embryo (Naegeli,
p. 152, Griiner, ’13, p. 83). But in view of the present data it
may be questioned whether analogous structures are not, how-
ever, also encountered in the embryo as well as in the adult.
The nuclear rings already described in the ingested erythro-
blasts of the coelomic macrophags, the basophilic periphery of
some of the eosin bodies in the coelomic cavity (figs. 21 to 26),
the nuclear ring-like structures arising in the degenerating eryth-
roblasts in the mesenchymal (figs. 27 and 3le) and other tis-
sue spaces (figs. 29, 30, 34 to 36), appear closely related if not
identical with the ring bodies of Cabot. In the embryo these
nuclear rings, especially in the mesenchyma are typically blue
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 1
98 Vv. E. EMMEL
with Giesma’s stain in contrast to the reddish tone of the rings
in pathological blood of the adult, although as already noted
blue rings are not entirely lacking even in the latter case. This
may be in part due to the different conditions under which they
are formed for it is to be observed that in certain regions such
as that of the brain wall and nerve ganglia many of the ring
bodies stain a reddish rather than a blue tone (figs. 28 and 38e).
In the embryo these rings are more frequently found without
any surrounding peripheral cytoplasm, that they may, however,
also occur within the still intact erythrocyte is well illustrated
in figures 34, 35 and 36 (also p. 96). This may be due in part
to an earlier disintegration of the cytoplasm in relation to the
nucleus in the case of the embryo as contrasted with the adult.
Although it is not to be overlooked that Gabriel (08, p. 604)
records the observation even in the adult of ring bodies a par-
ently lacking a peripheral rim of cytoplasm. As to the normal
or abnormal character of these structures there appears no doubt
but that in the embryo just as in the adult (Naegeli) their pro-
duction is a phenomena of abnormal nuclear disintegration and
not a normal mode of erythrocytic eytomorphosis. Concern-
ing the conditions under which the degenerating erythrocytic
‘nucleus will present the form of a ring or that of small compact
spherules, the possibility is suggested that this may be asso-
ciated in part at least with the stage of cytomorphosis at which
the degenerative processes are initiated. In the younger eryth-
roblasts, as is well known, the nucleus is both relatively larger
and the chromatin granules are more loosely distributed through-
out the nucleoplasm, whereas in later stages of differentiation
the nuclei become not only smaller but also much denser and more
compact in chromatin structure. Degerierative changes initi-
ated at these different stages may consequently be expected to
manifest correlated differences in nuclear disintegration. It is
possible that a longer persistence of the thickened reticulum
described by Cupp (15) at the periphery of the erythrocytic
nucleus may also be associated in part with the formation of
nuclear rings.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 99
2. Degenerating erythroblasts and the so-called eosinophilic
leucocytes in embryonic mesenchyma
Attention has already been directed to the fact that upon
first impression certain characteristics of the eosin staining coe-
lomic cells were suggestive of leucocytiec elements but that the
results of subsequent investigation indicated the nature of these
cells to be that of degenerating erythrocytes. In the case of the
rabbit eosinophilic leucocytes are also absent in the serous cavi-
ties of the adult (Weidenreich, 712, p. 127). In connection with
this conclusion it is to be taken into account, however, that cells
which appear practically identical in both nuclear and cyto-
plasmic structure with these degenerating erythrocytes in the
coelomic cavities are also found in the intercellular spaces of
the mesenchyma of the same embryos (figs. 38 to 40), concern-
ing the nature of which Maximow (’09) in his description of the
7 mm. cat embryo reaches the theoretically important conclu-
sion that they are eosinophilic leucocytes differentiating in situ
from mesenchymal cells.
The apparent identity of these cellular elements of the mesen=
chyma, both as described by Maximow and as observed in the
present material, with the cells interpreted as degenerating
erythrocytes in the coelomic cavities has rendered it necessary to
reexamine the evidence concerning the nature of the cellular
structures in the mesencyhma. In presenting the results of
such a study it may be stated that the following considerations
have led to a negative conclusion as to the leucocytic character
of the cellular elements in question in the mesenchyma. In
the first instance it is to be observed that no special eosinophilic
granulation can be demonstrated in these cells (figs. 38 to 40).
Indeed Maximow himself, although maintaining that in the
mesenchymal wandering cell the nucleus becomes lobulated into
a number of subdivisions held together in some cases by only
fine connecting strands, is nevertheless obliged to admit that no
leucocytic granules can be recognized in the cytoplasm of the
cells here in question (p. 525). In explanation of this, Maxi-
mow points out that it is likewise also very difficult to demon-
100 V. E. EMMEL
strate such granules in the granular leucocytes of even the adult
cat and that consequently failure to demonstrate these granules
in the embryonic cells does not constitute evidence of a neces-
sarily negative character. With reference to this point, how-
ever, it may be noted that the eosinophilic, non-granular cells
in question are not limited to the embryo of the cat for appar-
ently the same cells can be also demonstrated in the embryonic
mesenchyma of the rabbit, mouse and pig, mammals in the adults
of which a corresponding difficulty in staining the granules in
the granular leucocytes cannot be said to be encountered. Sec-
ond it can be demonstrated that degenerating erythroblasts,
occasionally found in the embryonic circulation may undergo
cytological changes apparently identical with the structural
characteristics to be found in the mesenchymal cells in question.
Erythrocytic nuclei as is well known, may become very irregular
in shape: bilobed or even constricted into several subdivisions.
Early stages in such nuclear changes are indicated in figure 37
and reference has already been made to Weidenreich’s account
on this subject (p. 595). In man lobulation of erythrocytic
nuclei has also been described under pathological conditions in
the circulation of the adult (Jiinger, ’00, p. 109). Maximow
(09, p. 478) also recognizes the occurrence of such nuclear lob-
ulation even in the embryonic circulation but described the
nuclei in such instances as becoming smaller, more compact and
taking a darker stain. It may be questioned, however, whether
this is necessarily always the case. For even in the circulating
blood, degenerating erythroblasts may be observed with lobu-
lated nuclei which cannot be said to present an especially more
compact structure (fig. 37). A similar comparison can be made
in figure 33 of Maximow’s work in which lobulated erythrocytic
nuclei are shown which do not appear either essentially darker
in stain nor more compact in structure than the unchanged nu-
clei of the adjacent erythroblasts or the nuclei of the co-called
leucocytes in figure 27 of his monograph. As for the relative
size of the cells no conclusive distinction can be clearly drawn
on this basis between the cells in question in the mesenchyma
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 101
and the erythroblasts in the circulating blood (cf. figs. 37 and 38,
drawn with the same magnification).
That the cytoplasm of the degenerating erythrocyte may be-
come vacuolated and assume a lighter stain is illustrated in
figures 37d and 34a. Similar degenerative erythrocytic changes
have also been described by Minot (12) and may be observed
in tissue cultures (Emmel 714). Finally, attention has already
been directed to the fact that erythrocytes frequently escape
from the vascular channels into the adjacent embryonic tissue
spaces where they may eventually disintegrate, be ingested by
phagocytes, or possibly in some cases eliminated through the
lymphatic vessels as suggested by Minot (12) for mammals and
as observed by Clark (’09) in the living tadpole. Cells which
can be unquestionably identified as such degenerating erythro-
cytes can be readily observed in the mesenchyma. ‘The nuclei
may vary from round to highly irregular and lobulated forms
and the cytoplasm may be paler in stain reaction, vaculoated
and under certain conditions may even disappear. Among these
cells are such forms as are illustrated in figure 40 the nucleus of
which may still be identified as erythrocytic as seems clearly
indicated by the persistence of a thickened peripheral accumula-
tion of chromatin in a manner comparable to that of the circu-
latory erythrocytes shown in figures 37 and 34. Such a nuclear
structure is in decided contrast to that of the adjacent mesen-
chymal cells as may be seen in the same figures. Practically
all transitional stages can be found between such cells and those
shown in figure 38 representative of the so-called eosinophilic
leucocytes under discussion. It appears evident that such par-
tial hemolysis, cytoplasmic vacuolation and consequent periph-
eral disintegration in the erythrocytes which have escaped into
the mesenchyma may result in the production of cytoplasmic
processes of such a character as not to be readily distinguishable
from cellular processes of the adjacent mesenchymal cells with
which they appear to fuse. Such cells may present the decep-
tive appearance of differentiation in situ from mesenchymal
cells as indicated in figures 39 and 40.
102 Vv. E. EMMEL
In conclusion, therefore it may be stated that in view of the
fact that the cellular elements under consideration in the mesen-
chyma are not only deficient in any definite leucocytic granules
but that nucleated erythrocytes degenerating in the embryonic
tissue spaces may assume cytological characteristics apparently
identical with those of the non-granular eosin staining cells with
lobulated nuclei in the mesenchyma, it appears difficult to es-
cape the conclusion that many if not the majority of the latter
just as in the case of the coelomic elements, are degenerating
erythroblasts rather than granular leucocytes developing in situ
from mesenchymal cells.
Concerning the distribution of such degenerating erythro-
blasts in the embryonic tissues, it is of interest to note that the
corpuscles in which the nuclei assume the lobulated condition
without becoming noticeably more compact in structure or
darker in staining reaction occur typically in the looser tissues
with larger intercellular spaces. On the other hand degenerat-
ing erythroblasts assuming the more compact forms containing
either nuclear rings or dark homogenous nuclear spherules, as
previously described, and presenting a great reduction if not
entire absence of peripheral cytoplasm occur typically in the
denser tissue regions (ef. figs. 89 and 40 with 31-383). Presum-
ably the structural characteristics which may be presented by
the degenerating erythroblasts are in part determined by the
nature of the environment in which such degeneration takes
place (cf. however also p. 98).
In concluding the present subject it may be stated with ref-
erence to Maximow’s most interesting and stimulating work
eoncerning the participation of the mesenchyma in the formation
of blood cells, that while the present results do not in themselves
necessarily constitute a conclusive argument against such a
possible role of the mesenchyma, they are presented as contrib-
uting toward an evaluation of some of the evidence which has
been advanced toward the establishment of such a conclusion.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 103
VI. CONCERNING THE PRESENT STATUS OF THE QUESTION AS TO
THE ORIGIN OF MACROPHAGS FROM THE COELOMIC
MESOTHELIUM
1. Embryological and comparative
As already intimated, a study of the cellular elements in the
embryonic coelom of mammals has not been previously made.
At the same time it may be observed that the present conclu-
sion that the coelomic epithelium may give rise to free functional
elements in the coelomic cavities is indirectly supported in an
interesting manner by the results of a number of recent inves-
tigations. Reference may be made to Bremer’s (’14) work in
which he finds anlages of the earliest blood vessels in man to
arise from the surface mesothelium. His ‘‘observations point
to the ingrowths of the mesothelial layer covering the yolk-sac
and body-sae as the anlages of the blood vessel endothelium
and of a lesser extent of the blood corpuscles”’ (p. 459), and the
conclusion is drawn that “True blood islands may occasionally
arise by the multiplication of the cells of the mesothelial in-
growths, or scattered blood corpuscles may arise singly’ within
these ingrowths” (p. 464). For the eat embryo Schulte (14)
records the occurrence of ‘‘funnel-like diverticuli of the coelom,
the walls of which are intimately united to the blood vessels,”
and which he suggests may be of morphological significance
with reference to the development of the embryonic blood ves-
sels (p. 80). Haff (14, pp. 346 and 333) states the conclusion
that the peritoneum covering the embryonic liver of the chick
may give rise to cells within the liver differentiating into eryth-
rocytes. Scammon (715), in the histogenesis of the Selachian
liver, records the occurrence of mesothelial tubules ‘‘the walls
of which are continuous with the splanchnic mesothelium and
the lumen with the coelomic cavity” (p. 276). Although un-
able to find that the lumen of these tubules connected with that
of the blood spaces, it is stated that the tubules break up into
mesenchymal strands and that ‘“‘the mesenchymal and endothe-
lial cells form free anastomosis” (p. 280). Phylogenetically ref-
erence may also be made to primitive vascular conditions in
104 Vv. E. EMMEL
some of the lower invertebrates. Lang (’04, p. 152) states that
in the body cavities of annelids there occur not only sex cells
but also amoebocytes (lymphocytes) and coelomocytes, some of
which as the result of formation of hemoglobin are designated
haemocytes, and that these elements arise from the coelomic
epithelium. Abbot (713, p. 6) describes the observation of
hemoglobin containing cells or ‘haematids’ in the body cavity
of the Echiurid worm, Thallasema mellita, arising from the
“living membranes of the general body cavity which buds off
masses of cells, usually eight to twenty-four in number, which
ultimately break up into individual haematids’”’ (p. 6). Data
of a similar character could be greatly extended not only among
annelids, but also in the Echinoderms and Coelenterates. The
views of Biitschli (83), His (00) and Arnold (’04) concerning
the phylogenetic origin of the circulatory system from the body
cavities are well known.
2. In adult mammals
In connection with these embryological results the question
arises as to whether these mesothelial activities are confined to
the embryo, or whether such a potentiality may be retained
even in the adult animal. Without entering into a detailed ac-
count of the extensive hematological literature bearing upon the
much debated question as to the nature and origin of the free
cells in the adult serous cavities (for a discussion of which cf.
Weidenreich (11, pp. 126-138), certain aspects of the problem
may briefly be considered in the light of the more recent inves-
tigations.
Schott (09), from cytological and experimental studies reaches
the important conclusion that the surface lining cells of the adult
body cavities in the guinea pig and rabbit are not highly special-
ized and fixed passive structures, without potentiality for fur-
ther differentiation, but that on the contrary they may assume
phagocytic activities, become detached from the serous mem-
brane and be liberated as free, active, living cells identical with
the lymphocytes and macrophags of these cavities. He states:
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 105
‘‘Wir miissen aber andrerseits auf Grund unserer Flaichenpri-
parate vom Netz . . . . unbedingt an der Fahigkeit der
Deckzellen sowohl zur aktiven Phagozytose wie zur Isolierung
und Loslésung aus dem Zellverbande festhalten” (201), and in
conclusion finds: ‘‘dass die grossen ungranulierten Exudat-
zellen und nicht nur diese, sondern auch die grossen Elemente
des normalen Transudates, Abkémmlinge fixer oder sessiler
Gewebsbestandteile sind. Urspriinglich fixe Gewebselemente
lésen sich aus dem Zellverbande, runden sich ab und werden
zur freien Zellen der serésen Hohlen’”’ (p. 208). Among subse-
quent investigators confirming this conclusion may be noted
Szecsi (712) who states: “‘Aus der Endothelzellen wird zuerst
durch Wachstum im Plasmaabrunding der Lymphoidozyte und
Macrophage” (p. 18). This is again reafirmed in the later work
of Szeesi and Ewald (13, p. 182). Lippman and Plesch (’13)
in a summary of their Thorium experiments, relative to the exu-
date cells of the serous cavities further support this conclusion:
“Somit kénnen die ’kleinen lymphozyten’ weder Hiimatogen
sein (aleukozytires Blut.’), noch von den Adventitialzellen,
noch von den taches laiteuses stammen—sie sind Abkémmlinge
des serosaendothels” (p. 1396). Weidenreich (11) writes with
no uncertainty concerning the origin of the large phagocytic
cells of the serous cavities: ‘‘sie sind losgeléste Netzelemente,
und somit sowohl Abkémmlinge von Deckzellen als auch Binde-
gewebszellen, . . . . die nicht degenerieren, sondern in
Gegenteil sehr lebenskriftig und mitotischer Teilung fihig
- sind” (p. 133). In opposition to the above conclusion Pappen-
heim-Fukushi (’13) maintain that the exudate cells of the serous
cavities are not mesothelial derivatives: “‘Sie sind Abkémmlinge
nicht der Deckzellen’”’ (p. 305). . . . . Allerdings leiten
Lippman und Plesch die entziindungezellen ebenso wie Wie-
denreich-Schott, auch von den serosen noch nicht angenommen
werden darf”’ (p. 289). Recently through the employment of
vital staining methods new evidence has been introduced into
the discussion of the problem. Thus Goldman (712), on the basis
of his studies with vital stains, draws a distinction between the
macrophags and serous mesothelium on the ground that the
106 Vv. E. EMMEL
latter in eontrast to the former does not stain with pyrolblue
(p. 45, 49) and the inference consequently arises that the macro-
phags of the serous cavities are not derived from the mesothe-
lium. Tschaschin (13, p. 350) not having succeeded in obtain-
ing a vital stain for the peritoneal endothelium, also draws a
similar conclusion. In experiments with celloidin plates in-
serted into the peritoneal cavity, Tschaschin (13) furthermore
failed to find that the mesothelial cells manifested any special
reaction or potentiality for transformation, but that on the con-
trary they quickly disquamated and under the conditions of
the experiments took no part in the formation of macrophags
(pp. 271, 285; 289).
Unanimity cannot, therefore, be said to have as yet been at-
tained in the solution of the problem. As already indicated
with reference to negative evidence, perhaps the strongest data
which has been more recently advanced is that derived from such
results as that of Goldmann and Tschaschin with vital stains.
It remains to be seen what is to be the final evaluation of the
data derived from this method: At the present stage of such
investigations it may not be without value to note the following
points which do not appear to render some of the results so far
attained as of a necessarily conelusive character with reference
to the question in hand. ;
It may be observed that if the reaction of the given tissue
(whether endothelial or mesothelial) to the vital stain is nega-
tive and that of the macrophags positive, it appears, in some
cases at least, that this is taken as evidence that the macrophags
could not be derivatives of the tissue in question. But with
reference to such a conclusion the question arises as to whether
it has been clearly demonstrated, first, that there is such a sharp
difference in the reaction of the mesothelium and macrophags
and second, that even a material difference in vital stain reac-
tion in itself establishes an entire absence of any genetic rela-
lationship between the tissues under consideration. In the first
place it is not to be overlooked that in some instances at least
both mesothelium and macrophags may react alike, for Schule-
mann (712) after injection of Trypanblue in rabbits did not ob-
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 107
tain a stain reaction for either the macrophags of the peritoneal
eavity or the peritoneal epithelium, a result which, as he him-
self notes (p. 289) consequently substantiates Schott’s conclu-
sion. On the other hand it has been shown that in cases where
the macrophags do manifest a typical vital stain reaction the
reaction of the mesothelium is by no means necessarily entirely
negative, for Evans (14) records the observation that whenever
certain cells such as the clasmatocytes (resting wandering cells)
of the connective tissues and macrophags of the great serous
cavities ‘“‘react in a typical intense manner to the vital stain”
other cells ‘“‘are normally found with much smaller often very
minute granules of the stain” in which latter class, it is import-
ant to observe, is included the mesothelium ‘‘lining the peri-
toneum and covering its organs” (p. 100). Again it does not
appear that the fact that the reaction of a tissue to the vital
stain in a given case is negative or stains only slightly in con-
trast to the macrophags necessarily leads to the conclusion that
such a tissue cannot give rise to the macrophag elements. <A
case in point is that of the endothelium of some of the larger
blood vessels in the liver. It has been shown for example by
Taschaschin (18) that whereas after ‘Kallargol’ injection black
silver granules are specifically deposited in the large mononu-
clear elements or macrophags of the blood, they are not found
in the endothelial cells of the portal vessels (p. 353). On the
other hand Batchelor (714) ,while he finds that injections of try-
pan blue just as in the case of ‘Kallorgol’ ,do not normally stain
the endothelium of the larger portal vessel of the liver in con-
trast to the positive reaction of some of the phagocytic endothe-
hal giant cells in the same organ, nevertheless endothelial pro-
liferations experimentally produced in the same vessels by means
of albumen emboli do react to the stain: thus newly formed en-
dothelial tissue at the site of the embolus and in its immediate
neighborhood it is stated, ‘“‘is stained vitally, a phenomenon
never seen with the normal endothelial cells of larger vessels,
and showing that the vital stain is adequate for the detection of
endothelial growths although the parent tissue does not show this —
property” (p. 139). MacCurdy and Evans (’12, p. 1695— and
108 Vv. E. EMMEL
also Tschaschin (’13, p. 370), record the observation that in the
case of blood vessels in the vicinity of wounds or other irritants,
the vascular endothelium which normally does not stain, may
under these changed conditions now react to vital stain. Fi-
nally it may be noted that the macrophags themselves do not all
react alike. Thus Tschaschin (’13b) recognizes the occurrence
of distinct variations in the intensity of the vital stain in the
macrophags: ‘‘Es muss jedoch hervorgehoben werden, dass die
freien Macrophagen der Bauchbohle sich vital bei weitem nicht
immer gleich intensiv fiirben” (p. 351). Tschaschin associates
this with variations in different types of stain and methods of
injection, rather than as furnishing any ground for identifying
these lighter stained cells with detached cells from the peri-
toneal endothelium, but in view of the above referred to results
by MacCurdy, Evans and Batchlor, it does not appear that the
latter possibility can as yet be said to have been successfully elimi-
nated. Variations in the vital stain of such detached cells may
well be correlated with different degrees of differentiation as
has indeed been emphasized even by Tschaschin (p. 382) in con-
nection with difference in the vital stain reaction of blood cells
so that undifferentiated cells reacting negatively with a given
vital stain may with further differentiation give a positive re-
action with the same stain. Consequently on the basis of the
data so far at hand, the ground does not appear clear on which
it can be stated with entire assurance that the mesothelial cells
which are able to take up a small number of the granules of the
vital stain, may not under given conditions just as in the case of
the vascular endothelium, come to manifest an increased ex-
pression of the same function such that as detached cells would
identify them with true macrophags.?
In conclusion, therefore, it appears that a convincing case can
hardly as yet be said to have been made against the possible per-
2 In connection with the question as to the degree to which mesothelial and
endothelial tissue may manifest common morphological and functional poten-
tailities it is of interest to note the results of Hooper and Whipple (715) indicat-
ing that mesothelium as well as endothelium may participate in the formation
of bile pigments.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 109
sistence in the adult mammal of the potentiality on the part of
the mesothelium of contributing, under certain stimulating en-
vironmental conditions, liberated cellular elements to the body
cavities in a manner comparable to the processes taking place
in the embryonic coelom.
VII. RESUME
1. A considerable number of free cellular elements were found
to be more or less constantly present in the coelomic cavities of
pig, rabbit and mouse embryos.
2. These coelomic elements may be described as falling into
two groups, the one consisting of basophilic staining and usually
phagocytically active cells and the other of cellular elements
characterized by their eosinophilic staining qualities and non-
phagocytic activity.
3. The coelomic macrophags
a. The basophilic cells may be conveniently further subdivided
into the following types: 1) cells relatively smaller and more
spherical in form, containing an occasional small cytoplasmic
vacuole and a rather dark staining, round or kidney shaped
nucleus; 2) cells usually larger in size, more oval or irregular in
form, containing one or more phagocytic inclusions, having a
round or kidney shaped and more or less eccentrically situated
nucleus and the cytoplasm and nucleoplasm of which take a
considerably lighter basophilic stain; 3) cells characterized by
the vaculoated condition of the cytoplasm and the not infre-
quent occurrence of cytoplasmic processes or buds projecting
from the surface of the cell.
b. The transitional stages which may be found between these
different cells are of such a character as to justify correlating
their size, form, nuclear, and cytoplasmic differences with varia-
tions in differentiation and function. Consequently practically
all of these cells are here regarded as belonging to a common
group which in view of their evident phagocytic functions may
be designated as coelomic macrophags.
c. As to the origin of these coelomic macrophags, some of
them may no doubt have entered the coelomic cavities from
110 Vv. E. EMMEL
extra coelomic regions, but the present results do not indicate
this to be the only source of origin of these cells. On the con-
trary the observation that at the surface of the coelomic walls,
in the free mesothelial cell masses, and the pleuro-pericardial
and pleuro-peritoneal membranes, mesothelial cells are found
which are rounded in form, manifest phagocytic characteristics
apparently identical with that of the typical macrophags, and
the evidence advanced that these cells may become detached as
free cells, support the conclusion that the coelomic mesothelium
is an important source of origin for the phagocytic cells found in
the embryonic coelom. |
4. Erythrocytic elements in the coelomic cavities.
a. The second of the two groups of coelomic cellular elements
above indicated again fall into two sub-groups, the one consist-
ing of small non-nucleated bodies and the other of larger nucleated
cells.
b. The present data indicate these structures to be erythro-
eytic in nature; the larger nucleated cells representing degen-
erating nucleated red blood corpuscles and the smaller eosin
staining bodies consisting chiefly of degenerating erythrocytic
nuclei.
5. Erythrocytic disintegration in the mesenchyma.
a. In the embryonic mesenchyma there occur small eosin
staining bodies which have been interpreted as hemoglobin
containing secretion products of mesenchymal cells (Maximow).
But on the basis of the present evidence concerning their dis-
tribution, structure and close cytological correspondence to disin-
tegrating erythrocytes as observed in the circulating blood, inter-
cellular tissue spaces and phagocytic inclusions and to eryth-
rocytic nuclei persisting after the formation of non-nucleated
erythrocytes, the conclusion is drawn that the bodies in question
represent chiefly erythrocytic elements consisting of degenerat-
ing and in many cases phagocytically ingested erythroblasts
and, in older embryos, including nucleated erythrocytic bodies
arising in connection with the formation of erythro-plastids or
non-nucleated erythrocytes, rather than the products of mesen-
chymal secretory activities.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO 111
b. The ring-like configuration presented by many of these
degenerating erythrocytic nuclei appear identical in many re-
spects with the nuclear structures occurring in the erythrocytes
of pathological blood known as Cabot’s rings and they conse-
quently furnish evidence that the latter are not limited to the
blood of the adult animal as has been previously assumed
(Naegeli).
c. Certain cellular elements in the mesenchyma present in
some respects the appearance of eosinophilic leucocytes and are
suggestive of a possible mesenchymal origin as maintained by
Maximow. In view of the fact, however, that all of these are
deficient in any definite leucocytic granules and that nucleated
red blood corpuscles escape into the embryonic tissue spaces
where they may present degenerative nuclear and cytoplasmic
characteristics apparently identical with those of the non-gran-
ular eosin staining cells in the mesenchyma, it appears difficult
to escape the conclusion that the majority of the latter, just as
in the case of the corresponding coelomic elements, are degen-
erating erythrocytes rather than granular leucocytes develop-
ing in situ from mesenchymal cells.
6. Concerning the problem of mesothelial origin of macro-
phags.
a. In connection with present conclusion that the coelomic
epithelium may give rise to free functional elements in the coe-
lomic cavities, it appears not without significance on embryo-
logical and comparative grounds to note the recent work of Brem-
er, Haff and Schulte concerning the participation of coelomic
epithelium in vasculogenesis and the formation of blood cor-
puscles in vertebrate embryos and the conditions in certain in-
vertebrates where the coelomic epithelium gives rise to cellular
elements functioning as respiratory and phagocytic cells in the
body cavities (Lang, His, Arnold).
b. Regarding the origin of the macrophags in the serous cavi-
ties of adult mammals unanimity still remains to be attained in
the solution of the problem. On the basis of the data so far at
hand, it does not appear, however, that a convincing case has
as yet been made against the possible persistence of a potential-
a2 Vv. E. BMMEL
ity in the mesothelium of the adult organism of contributing
under certain conditions free cell elements to the body cavities
in a manner comparable to the process here described as taking
place in the embryonic coelom.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO is
VIII LITERATURE CITED
Asport, F. J. 1913 The blood cells of Thallasema Mellita. Washington Uni-
versity Studies, vol. I, Part I, p. 1.
BatscHetor, R. P. 1914 Demonstration of preparations showing the behavior
of endothelium after the introduction of emboli in the portal vein.
Anat. Rec., vol. 8, p. 149.
Bremer, J. L. 1914 The earliest blood vessels inman. Am. Jour. Anat., vol.
16, p. 447.
Biscuit, O. 1883 Uber eine Hypothese beziiglich der phylogenetischen Her-
leitung des Blutgefassapparates eines Theils der Metozéen. Morph.
Jahrb., Bd. 8, p. 474.
Capor, R. C. 1903 Ring bodies (nuclear remnants?) in anemic blood. Journ.
Med. Research, vol. 9, p. 15.
Crark, E. R. 1909 Observations on living, growing lymphatics in the tail of
the frog larva. Anat. Rec., vol. 3.
Curp,.C. D. 1915 On the structure of the erythrocyte. Anat. Rec., vol. 9,
p. 259.
Downey, H. 1910 Phagocytosis of erythrocytes in the lymphorenal tissue of
Ployodon spatula. Folio Haematologica, Bd. 9, p. 81, Archiv.
1913. The origin of blood platelets. Folia Haematologica Bd. 15,
p. 25, Archiv.
Downey AND Weipenreicu, R. 1912 Uber die Bildung der Lymphozyten im
Lymphdrusen und Milz. Archiv. f. Mikr. Anat., Bd. 80, Heft. 2.
Emmet, V. E. 1914 Concerning certain cytological characteristics of the ery-
throblasts in the pig embryo, and the origin of non-nucleated ery-
throcytes by a process of cytoplasmic constriction. Am. Jour. Anat.,
Volz16; p: 127.
1914 Ibid. Anat. Record, vol. 8, p. 101.
1916 The cell clusters in the dorsal aorta of the mammalian embryo.
Am. Jour. of Anat., vol. 19, p. 401.
Evans, H. M. 1914 The physiology of the endothelium. Anat. Rec., vol. 8.
p. 99.
GasrigEL, S. 1908 Uber Ringkérper im Blute Animisches. Deutsch. Archiv.
Klin., Med. Bd. 92, p. 602.
GotpMANN 1912 Neue Untersuchungen iiber die dussere und innere Sekretion,
etc. Tibingen.
Goopricu, E. 1899 On the Nephridia of Polychaeta. Part II. Quart. Journ.
Micr. Sce., vol. 41, p. 4389.
Gruner, O. C. 1913 The biology of the blood cells. Bristol.
Harr, R. 1914 Bindegewebe und Blutbildungsprozesse in der Embryonalen
Leber des Huhns. Archiv. f. Mikr., Anat. Bd. 84, p. 321.
His, W. 1901 Lecithoblast und Angioblast der Wirbelthiere. Histogenetische
studien. Abhandl. der Math.-Phys. Classe Gesellsch. der Wissen-
schaften. Bd. 26, p. 171.
Hooper, C. W. anp Wuippte, G. H. 1915 Icterus. A rapid change of hemo-
globin to bile pigment in the pleural and peritoneal cavities. Proc.
Soe. for Exper. Biol. and Med., vol. 15, p. 22.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 1
114 V. E. EMMEL
Jotyy, J. 1907 Recherches sur la formation des globules rouges des mammif-
eres. Archives d’Anatomie, T. 9, p. 133.
Jiéncer, 1900 Uber Kernhaltige rothe Blukérperchen im Str6menden Men-
schlichen Blute. Deutsch. Archiv. Klin. Med. Bd. 67, p. 109.
Kyes, P. 1915 The physiological destruction of erythrocytes in birds. In-
ternat. Monatsschrift J. Anat. u. Physiol., Bd. 31, p. 543.
Lane, A. 1904 Beitriige zu einer Trophocéltheorie. Jenaische Zeitschr. f.
Naturwissench., Bd. 388.
LippMANN-PLEscH 1913 Uber die Genese Der ‘Lymphozyten’ in den Exsuda-
tion serdser Héhlen von aleukozytiren Tieren. Deutsche Med.
Wochenschr. No. 29.
MacCurpy anp Evans 1912 Experimentelle Lisionen des zentralnerven-
systems, untersucht mit Hilfe der vitalen Farbung. Berl. Klin.
Wochenschr. No. 36.
Maun, F. P. 1912 On the development of the human heart. Am. Jour. Anat.,
vol. 13, p. 249.
Maximow, A. 1909 Untersuchungen iiber Blut und Bindegewebe. I. Die
friihesten Entwicklungsstadien der Blut-und Bindegewebezellen beim
Siugetierembryo, bis zum Aufang der Blutbildung in der Leber.
Archiv. f. Mikr. Anat. Bd. 73, p. 444.
Minor, C. 8. 1912 The origin of the angioblast and the development of the
blood. Keibal-Mall Human Embryology, vol. 2.
Narceur,O 1912 Blutkrankheiten und Blutdiagnostik. Leipzig.
PAPPENHEIM-I'uKusHI 1913 Neue Exaudatstudien und wietere Ausfiihrungen
iiber die Natur der lymphoiden peritonealen Entziindungszellen.
Folia Haemat. Bd. 17, p. 257.
Scammon, R. E. 1915 The histogenesis of the selachian Liver. Am. Jour.
Anat., vol. 17, p. 245.
Scutere, K. 1907 Uber Ringkérper im Blute Aniimischer. Deutsch. Archiv
f. Klin. Med. Bd. 91, p. 455.
Scuorr uND WEIDENREICH 1909 Morphologische und experimentelle Unter-
suchungen iiber Bedeutung und Herkunft der Zellen der Serdsen
Hohlen. Arch. f. Mikr. Anat. Bd. 74.
Scuuutte, H.W. 1914 Early stages in the vasculogenesis in the cat with espe-
cial reference to the mesenchymal origin of endothelium. Anat. Rec.,
vol. 8, p. 78.
Szecst, Sr. 1912 Experimentelle studien iiber serosa-exsudatzellen. Fol.
Haemat. Bd: 13, \p: 1.
Szecsi, St. anp Ewatp, O. 1913 Zur Kenntnis der Peritoneal exsudatzellen
des Meerschweinchens. Folia Haemat. Bd. 17, p. 167.
Tscuascuin, S. 1913a Uber die Herkunft und Entstehungsweise der lym-
phozytoiden (leukozytoiden) zellen, der ‘Polyblasten’ bei der Ent-
zindung. Fol. Haemat. Bd. 16, Archiv.
1913b Uber die ‘ruhenden wanderzellen’ und ihre Bezeihungen zur
cen anderen Zellformen des Bindegewebes und zur den Lymphozy-
ten. Folia Haemat. Bd. 17, p. 317. Archiv.
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO TS
WerpEnreEIcH, F. 1903 Die roten Blutkérperchen. I. Ergebn. d. Anat. u.
Entwicklungsgesch. Bd. 13.
1911 Blutkérperchen und Wanderzellen. Sammlung Anatomischer
und physiologischen Vortrige und Aufsitze. Bd. 2, Helft 2.
1912. Die Thymus der Erwachsen Menschen als Bildungsstitte un-
granulierter und granulierter Leukozyten. Miinchner Med. Wochen-
schrift. No. 48, p. 2601.
PLATE 1
EXPLANATION OF FIGURES
The following figures are from camera lucida drawings in all of which, with
the exception of the low power drawing in 41 and in 1 to 7, 11 to 13, which were
outlined with a No. 8 ocular, the details were drawn from observations with a
Zeiss appochromatic immersion objective and No. 4 and 6 compensation oculars.
Fifteen of the drawings were made by Herr Kretz, artist at the Anatomical In-
stitute of the University of Strassburg and twenty-one by Mr. Jarrett and Miss
Ehinger at the Anatomical Laboratory of the Washington University Medical
School. In the reproduction plates 2 and 4 were reduced by one-fifth from the
original drawings.
1, 2. Cells belonging to the first of the three types described under coelomic
macrophags. The cytoplasm is decidedly basophilic, contains a number of small
cytoplasmic vacuoles and the nuclei are either round or kidney shaped. What
appears to be a centrosphere is seen at the left of the nucleus in figure 1. These
cells are interpreted as coclomic macrophags in a stage of inactivity with refer-
ence to phagocytic functions. From the pericardial cavity of a 9 mm. rabbit
embryo (compare with fig. 9).
3. Mitosis in the type of cells shown in figures 1 and 2. From the pericardial
cavity of the same embryo as above.
4,6. Macrophags at a stage of active phagocytosis. The inclusions appear
to consist chiefly of erythrocytic elements. The lighter cytoplasmic and nucleo-
plasmic stain as compared with 1 and 2 appears correlated with an advanced
stage of phagocytic activities. mn, is nuclear inclusion with a lighter stained
central area. From the pericardial cavity of a9 mm. rabbit embryo.
7 Mitosis in a coelomic macrophag containing two large cytoplasmic vacu-
oles (ef. fig. 3). Pericardial cavity 9 mm. rabbit embryo.
8 Two coelomic macrophags lying side by side and consequently subject to
identically the same technique. They demonstrate the lighter nuclear and cyto-
plasmic stain in the cells at stage of greater phagocytic activity b as compared
with the less active cellsa. In cell} one of the inclusions still retains an unmodi-
fied remnant, in the form of a crescent, of the original basophilic material of the
ingested nucleus. Pericardial cavity, 9mm. pig embryo.
9, 10 Furnishes a striking contrast in the cytological characteristics of a
phagocytically active (10) anda phagocytically inactive cell (9). From the same
source as figure 8.
11 Coelomie cell (macrophag?) showing a highly vacuolated condition of
the cytoplasm. Peritoneal cavity, 9 mm. rabbit embryo.
12 Cells occasionally found in the coelomic cavities of rabbit embryos, show-
ing a peripheral border of eosin staining material. Similar structures are also
found in the mesenchyma where they undoubtedly represent degenerating hemo-
globin containing elements. Pleural cavity, 9mm. rabbit embryo.
13. A macrophag undergoing mitosis in which the large digestive vacuole
still contains a visible undigested remnant of the phagocytic inclusion. Peri-
rardial cavity, 9mm. pig embryo.
116
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO
Vv. E. EMMEL
os
‘
iy 12
lp gy
PLATE 1
7
PLATE 2
EXPLANATION OF FIGURES
14 One of fourteen sections through the pleuro-pericardial membrane of a
7mm. pigembryo. c, represents a central issue core surrounded by mesothelial
cells (mes). Many of these cells have assumed a rounded form and are partially
if not entirely detached from the main mass. These modified cells are phago-
cytically active (pem), contain cytoplasmic vacuoles (vem) and present struc-
tural and tinctorial characters practically identical with that of typical macro-
phags. For general relations see figure 43. e, small eosin staining bodies (ef.
figs. 20 to 26).
15 Free mesothelial cell masses in the pericardial cavity of a 9 mm. rabbit
embryo. The visceral and parietal pericardial walls, respectively, are at the
right and left sides of the figure (mes).
16 Free mesothelial cell mass in the ventral region of the pericardial cavity
ofa 7mm. pigembryo. It extends through five sections, but in both this case
as well as in figure 15 these cellular masses are at no point attached to the coelomic
wall. Some of the mesothelial cells are partially detached, present a rounded
form and a slightly more basophilic stain reaction (7).
17 A region of the mesothelial surface showing a rounded, partially detached
basophilic cell (m) which appears to be a modified mesothelial cell. Visceral
pericardium of a9 mm. pig embryo (ef. figs. 41, 48, 44).
18 A group of free cells in the peritoneal cavity of a9 mm. rabbit embryo,
interpreted in the text as degenerating erythrocytes. In some of the cells indi-
cations of hemoglobin are still evident (a), in others the cytoplasm is paler in
color show cytoplasmic vacuolation and degenerative nuclear changes (0, c).
19 Other instances of degenerative nuclear and cytoplasmic changes in
erythrocytes (cf. fig. 18). Peritoneal cavity, 9mm. rabbit embryo.
ABBREVIATIONS
cm, coelomic macrophags mc, mesenchymal cells
d, mitotic figures mes, mesothelium
icm, coelomic macrophags, apparently mm, mesothelial cell masses found free
in a stage of inactivity with refer- in the coelomic cavities
ence to phagocytic functions pev, pericardial cavity
in, phagocytic inclusions pem, coelomic macrophags in a stage
m, rounded cells interpreted as meso- of active phagocytosis
thelial cells in the process of becom- vem, vacuolated coelomic macrophags
ing detached as free cells in the coe- —_w, vessel wall
lomic activities
118
PLATE 2°
119
eure
8 OF THE MAMMALIAN EMBRYO
Vv. E. EMMEL
ELLULAR ELEMENT
“
*
PLATE 3
EXPLANATION OF FIGURES
20 to 26 Small eosin staining bodies (degenerating erythrocytic nuclei) in
the coelomic cavities, some of which contain variable quantities of basophilic
material in the form of small spherules and peripheral rings. 20, 22, 24, 26 are
from the pleural cavities of 7 mm. pig embryos; 21, 23 from the pericardial cavity
of a 13 day mouse embryo, and 25 from the pleural cavity of a9 mm. pig embryo.
27, 28 Ring form structures (degenerating erythrocytic nuclei) in the tissue
spaces. 27 is from the mesenchyma of the ventral thoracic wall of 9 mm. rabbit
embryo and 28 from the gasserian ganglion of a 13 day mouse embryo.
29, 30 Degenerating erythrocytic nuclei in the heart cavity of a7 mm. pig
embryo (29) and of a 13 day mouse embryo.
31, 82,33 Groups of small bodies, e (degenerating erythrocytic elements con-
sisting chiefly of nuclear material) in the embryonic tissues. 31 is from the
mesenchyma of the ventral wall of a 9 mm. rabbit; 382 from the mesenchyma the
septum transversum of a7 mm. pig embryo; and 33 from the ventral wall of the
fore-brain of a 13 day mouse embryo. Shows their variation in size, structure,
stain reaction and intercellular relations.
34, 35, 86 Ring form nuclear structures (Cabot’s rings) observed in degen-
erating erythrocytes found in the embryonic circulation. 34 and 35 respective-
ly, are from a small blood vessel in the mesenchyma and a sinusoid in the liver
of a 13 day mouse embryo; and 36 from a blood vessel (or posssibly lymphatic) in
the mesenchyma of a9 mm. pig embyro.
37 Erythrocytes in the hepatic sinusoid of a 13 day rabbit embryo showing
earlier stages in degenerative nuclear and cytoplasmic changes including lobula-
tion of nuclei, hemolysis and cytoplasmic vacuolation.
38 Cells found in the intercellular tissue spaces of the mesenchyma of the
same embryo as for figure 37, concerning which grounds were advanced in the
text indicative of their degenerating erythrocytic nature rather than leucocy-
tic elements differentiating in situ from mesenchymal cells.
39, 40 Cells from the mesenchyma which appear in nuclear and cytoplas-
mic structure clearly intermediate between the cells shown in figure 38 and the
degenerating erythrocytes in figure 37. From the same source as figure 38.
120
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO
V. BE. BMMEL
_@ : 2) MC © a
e&D ® @ @ © \ , =
A
\
P21 22 23 \
F . xe ;
20 . i. : d
oe i } "kh
17 4
i
©) 6) (> ha
ys - Tees 2 4 \
aa »”* n cf
24 25 c
26
> se
©) G fe \
27 28 29 20
a
34 ,
FLATE 3
eo ‘e
Ce 2
b ety
aot
mec net
w
S12
ar os a
QO 4
af
4
7 @». GS
36 & > :
3 3
38
40
A?
“Se Seen
PLATE 4
EXPLANATION OF FIGURES
41 Showing a rounded cell (m) projecting into the pericardial cavity from
the visceral pericardium, but still attached by a slender cytoplasmic pedicle to
the mesothelial surface. 7mm. pig embryo (cf. fig. 17, 42, 44).
42 Demonstrating the potentiality for phagocytic activity on the part of
the mesothelial cell. The mesothelial character of the cell can hardly be ques-
tioned, at the same time it is partially raised above the level of the mesothelial
surface and the nucleus approximates a kidney-shaped form. From the parietal
pericardium of a9 mm. pig embryo.
43 Showing the position and general relations of the section of the pleuro-
pericardial membrane (pp) drawn at a larger magnification in figure 14. It may
be observed that the cellular mass in question hes in the pleural cavity (pev)
and that the latter is still in communication with the pericardial cavity through
the pleuro-pericardial canal (pplc). The membrane continues through fourteen
sections and is found to connect with the parietal wall at the junction of parietal-
pleural (ppl) and parietal-pericardial (ppc) walls where its cells become contin-
uous with the mesothelium lining the coelomic walls.
44 Showing a region of the visceral pericardium in which the mesothelial
cells present the appearance of transformation into free cellular elements. This
region continues through several successive sections. Many of the cells assume
a more basophilic stain than that of typical mesothelial cells, and phagocytic
activities, cytoplasmic vacuolation and peripheral processes or buds (b) appar-
ently identical with that of the coelomie macrophags. 7mm. pig embryo.
*
——
CELLULAR ELEMENTS OF THE MAMMALIAN EMBRYO PLATE 4
V. E. EMMEL
‘ GED
Z1
Pe Dip PPL
RES et
MES.
riugt SE
ie
be
Cae
ig
be
Ge
Oe,
We
oe
Ge
iD
oF
D.
Re athe
SOT band ps ‘’.
os One S
SOB Le re
ee)
‘ © Pr r pe ond | YY,
~ a a | - 5 Ps Wits Ke
5 : Y
Se oe a) f -
ss" \
, : 1 AT @ tr; _™ se ? a a
~ 1 i We a 8 :
x 1 Mf ot oe 7 rus
co ’ F mi
4 ere
F ' 5
‘ ‘
4
- 4
ae ry
*
*
ij
_ :
.
THE GENESIS, DEVELOPMENT, AND ADULT ANAT-
OMY OF. THE NASOFRONTAL REGION IN MAN
J. PARSONS SCHAEFFER
Daniel Baugh Institute of Anatomy and Biology of the Jefferson Medical College
of Philadelphia
THIRTEEN FIGURES
Owing to the contradictory and often but general statements
extant in the literature on the nasofrontal connections, the writer
deemed it important to make a more detailed study of the em-
bryology and adult anatomy of this region. The present com-
munication will in a sense supplement previous studies on the
embryology of the nose by the writer. Special attention is. here
given to an analysis of the adult anatomy, and an effort is made
at an intelligent interpretation of the complicated region by refer-
ring to the genesis and development of the parts involved. The
anatomy of the nasofrontal connections is of considerable im-
portance clinically, since the sinus frontalis is now frequently
approached from the nasal cavity in operative procedures.
With the latter thought in mind, important anatomic types of
the region commonly encountered are illustrated by drawings
from actual personal dissections. The embryology is dealt
with but briefly; the reader is referred to previous papers by the
author for more detailed discussions.
EMBRYOLOGY
The nasofrontal region is genetically an outgrowth from the
ventral and cephalic end of the meatus nasi medius, operculated
by the concha nasalis media (middle turbinated bone). The
mucosa of this part of the meatus nasi medius is, therefore, the
proton of what subsequently becomes the recessus frontalis of
the meatus nasi medius (early in evidence) and derivatives there-
125
126 J. PARSONS SCHAEFFER
from. The recessus frontalis in turn is the anlage of the sinus
frontalis and certain of the anterior group of cellulae ethmoi-
dales (also called cellulae frontales by Killian, Onodi and others).
As early as the end of the third or beginning of the fourth
month of embryologic life, one sees evidence of a beginning ex-
tension of the meatus nasi medius in a ventrocephalic direction.
This early extension is the anlage of the recessus frontalis and
is, strictly speaking, the first step in the formation of the sinus
frontalis and certain of the anterior group of cellulae ethmoi-
dales. For some time the lateral wall of the recessus frontalis is
ei ; Lrontal tuarrows
1-4
bis my
Concha nas.med (cut) / a Litindib. eth
eee Bulla eth.
Pocuncinutus. ue =
Fig. 1 From aterm fetus. The recessus frontalis is exposed for study by the
removal of the operculating concha nasalis media. Note the frontal furrows
and the relations of the infundibulum ethmoidale. Compare with figure 6.
The most ventral of the frontal furrows or pits are referred to throughout this
paper as the first (1), the next in order as the second (2), ete.
even and unbroken and gives no evidence of the later configura-
tion and complexity which characterizes the region in the adult
nose. Coronal sections and transections of the recessus frontalis
of a 4-month fetus show the lateral nasal plate of cartilage
thickened at certain points. These thickened cartilaginous
areas—-the forerunners of the folds or accessory conchae which
later configure the lateral wall of the recessus frontalis—vary
in number and are for a period low and inconspicuous and do
not throw the nasal mucosa into relief.
Upon examining the recessus frontalis in the late fetus, one
finds a variable number of low accessory conchae on its lateral
wall (figs. 1 to 4). The folds with the cartilaginous skeleton,
NASOFRONTAL REGION IN MAN 127
now partly ossified, are at this time sufficiently developed to
throw the nasal mucosa into relief. Between the folds are
found pits or furrows, the positive growth or outpouching of
which aids materially in making more prominent the folds. It
is appropriate to speak of the latter as accessory or hidden
frontal folds or conchae and the pits as frontal furrows, of the
meatus nasi medius. As mentioned above, there is no con-
stancy in the degree of differentiation and development of the
frontal folds and furrows. The number varies from a complete
absence to four or five. In some instances, therefore, the re-
eessus frontalis remains a simple blind outgrowth from the meatus
nasi medius without configuration of its lateral wall (fig. 5).
The processus uncinatus and the folds composing the bulla eth-
moidalis likewise should be considered as accessory conchae of
the meatus nasi medius (analogues and homologues of the
frontal conchae), and the infundibulum ethmoidale andthe supra-
bullar furrow as accessory meatuses or furrows of the meatus
nasi medius (analogues and homologues of the frontal furrows).
The accessory furrows of the meatus nasi medius are fore-
runners of certain of the sinus paranasales, 1.e., the sinus frontalis,
the sinus maxillaris, and the anterior group of cellulae eth-
moidales (by anterior group is meant all those ethmoidal cells
which communicate with the nasal fossa caudal to the attached
border of the concha nasalis media, including both the ante-
rior and middle group according to another classification).
The frontal furrows or pits early evaginate and form certain
of the anterior group of cellulae ethmoidales or cellulae fron-
tales. Semi-coronal sections through the recessus frontalis show
these early cells. When these cells are followed in serial sec-
tions toward the recessus frontalis, they are shown to be ex-
tensions or outpouchings of the frontal furrows and in com-
munication with the recessus frontalis. Some of the cellulae
ethmoidales having their genesis in frontal pits remained di-
minutive and ethmoidal in topography, while others grow to
considerable size and often develop beyond the confines of the
ethmoidal bone.
128 J. PARSONS SCHAEFFER
It is a well established fact that the sinus frontalis develops
variously by a direct extension of the whole recessus frontalis;
from one or other of the anterior group of cellulae ethmoidales
which have their point of origin in frontal furrows; and occa-
sionally from the ventral extremity of the infundibulum eth-
moidale, either by direct extension or from one of its cellular
outgrowths. Indeed, the sinus frontalis may be unilaterally or
bilaterally present in duplicate or triplicate, indicating a genesis
from more than one of the aforementioned areas. The sinus
frontalis is in many instances, embryologically speaking, a
a) as Tiontal furrows
lrontal folds
(accessory conchac) -«
_ Loc. uncrn ct us
Concha nes.med.-— -
(
\\
Bulla Hes 2 e-2
Fig. 2 From a term fetus. Recessus frontalis exposed. Note frontal fur-
rows and folds. Especially note the continuity of the suprabullar furrow and the
fourth (most dorsal) frontal furrow (see reference in text). The infundibulum
ethmoidale is in line with the first frontal furrow, but not directly continuous
with it. After Schaeffer.
cellula ethmoidalis anterior which has grown sufficiently far
into the frontal region to be topographically a sinus frontalis.
The first evidence of the sinus frontalis must not be sought
in the frontal bone, but in the recessus frontalis of the meatus
nasi medius. Lack of observance of this rule has led to such
statements as: ‘“‘in the newborn infant no trace of a frontal sinus
is visible,” ‘‘the earliest sign of a frontal sinus is seen about the
end of the first year in the form of a shallow depression,” ‘‘the
frontal sinus is completely absent in the newborn infant.’? Poi-
rier states that the frontal sinus is first seen about the end of the
second year. Tillaux puts it as late as the twelfth year. Onodi,
Schaeffer, Davis and others recognize the sinus frontalis as such
NASOFRONTAL REGION IN MAN 129
in some instances early in extrauterine life. Kilhan operated
upon a diseased sinus frontalis in a child fifteen months old. As
stated before, the recessus frontalis of the meatus nasi medius
is demonstrable as early as the fourth fetal month. During late
fetal life the recessus frontalis becomes complex by the formation
of frontal furrows or pits, ete. One is not justified at this time
to hazard an opinion as to the specific point in the recessus
frontalis from which the sinus frontalis will ultimately develop.
\ _. Frontal told or concha
ee, bey, an neat Say eee ee Bulla eth.
Geenee . Concha nas, suprema
“>. Contha nas.supranal
Concha nas. sup.----[- ;
Concha nas. 70a.---
Concha nas. tnt ---- ~~ ya
Fig. 3 From a term fetus. Here a single frontal fold or concha presents,
bordered by a dorsal and a ventral frontal furrow. The infundibulum ethmoidale
and the frontal furrows are continuous channels. Whether this is a secondary
condition due to growth, 1.e., whether the frontal furrows and the infundibulum
ethmoidale were discontinuous anlagen, cannot be said. According to my
series the condition is not common.
There are exceptions to this rule. Occasionally at birth the
genetic point for the sinus frontalis is obvious. Again, one
cannot be certain until the second or the third year.
From the suprabullar furrow develop most of those cellulae
ethmoidales anterior which in time honeycomb the bulla ethmoi-
dalis. Rarely the suprabullar furrow seems to be the gen-
etic point for the sinus frontalis. This may be apparent only
and not the actual condition. The most dorsal of the frontal
pits and the suprabullar furrow are at times continuous channels
(fig. 2). This might lead to the interpretation that the sinus
frontalis developed from the suprabullar furrow, when in re-
ality it developed from a frontal pit (early anterior ethmoidal
cell). At times some of the bullar cells develop from occasions.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 1
130 J. PARSONS SCHAEFFER
furrows on the medial surface of the bulla ethmoidalis. The
infundibulum ethmoidale dorsally and caudally gives rise to the
sinus maxillaris and ventrally it usually ends blindly by forming
a cellula ethmoidalis anterior of variable size, lateral to the
recessus frontalis. Various aberrant cellulae ethmoidales anterior
(agger, conchal, infundibular, ete.) also frequently develop
from the aforementioned points. The posterior or dorsal group
of the cellulae ethmoidales do not concern us here.
_.. Kontul furrows or pits
Kecessus trontalis —~---|- +" 22
Sinus sphcnoidalis
s 7 ==5. Bulla cue
UD
Fig. 4 From a term fetus. Recessus frontalis exposed by partial removal of
concha nasalis media. Note intimate relationship between ventral extremity of
the infundibulum ethmoidale and the second or most dorsal frontal furrow or
pit. Should the sinus frontalis develop from the second frontal furrow, the
adult relationship between the former and the infundibulum ethmoidale would
be very intimate. Indeed, there might be direct continuity, allowing for further
changes in development. The conclusion would be wrongly drawn, however, to
say that the sinus frontalis is a derivative of the infundibulum ethmoidale.
The processus uncinatus and the folds composing the bulla
ethmoidalis are often in direct continuity with one or more of
the frontal folds or conchae (fig. 1). Again, the bulla ethmoi-
dalis and the processus uncinatus are fused across the ventral
extremity of the infundibulum ethmoidale (figs. 1 and 5). Like-
wise in many instances the dorsal extremity of the processus
uncinatus divides, in a sense, into two roots, one of which turns
cephalad and fuses with the bulla ethmoidalis, thus causing the
infundibulum ethmoidale to end in a blind pouch dorsally (fig. 8).
The infundibulum ethmoidale and one or other of the frontal
furrows or pits are in the same axis in the fetus and are at times
contiguous (fig. 1). It must, however, here be pointed out that
NASOFRONTAL REGION IN MAN 131
it is unusual for the infundibulum ethmoidale to be directly con-
tinuous with a frontal furrow or pit (fig 3). The latter embryo-
logical fact is significant when one recalls the careless state-
ment frequently made without qualification, that in the adult the
“infundibulum ethmoidale is continued upwards as the naso-
frontal duct into the sinus frontalis.”’
Because of the intimate relations, in the adult, of the infun-
dibulum ethmoidale and the ductus nasofrontalis or the sinus
frontalis directly, the infundibulum ethmoidale serves, in many
instances, as a channel to convey secretion from the sinus fron-
talis to the sinus maxillaris. This is enhanced in those cases in
which the infundibulum ethmoidale is deep and ends dorsally
in a blind pouch, thus directing drainage through the ostium
maxillare into the sinus maxillaris (figs. 8 and 12). In other
words the sinus maxillaris is often a cesspool for infectious ma-
terial from the sinus frontalis and certain of the anterior group
of cellulae ethmoidales
The above well known clinical fact has doubtless led to the
erroneous belief that the infundibulum ethmoidale is, in the
majority of instances, directly continuous anatomically with the
nasofrontal duct or, in the absence of the latter directly with
the sinus frontalis. The inference is also wrongly drawn that
in many cases the sinus frontalis is embryologically a direct
outgrowth of the ventral and cephalic end of the infundibulum
ethmoidale. From what has been said previously on the em-
bryology, it is needless to enter further into a discussion here.
Suffice it to say that it is not a common adult anatomic condi-
tion to find the infundibulum ethmoidale directly continuous
with the ductus nasofrontalis, or in the absence of the latter
with the ostium frontale. There are in many instances close
relationships established, even a contiguity, but a direct contin-
uity is an occasional occurrence only. According to the series
worked it is, likewise, not common for the sinus frontalis to have
its genesis in the infundibulum ethmoidale.
It should, however, be pointed out that in probably as many
as 50 per cent of adult cases the relationship is so intimate be-
tween the infundibulum ethmoidale and the sinus frontalis or
v32 J. PARSONS SCHAEFFER
its duct (ductus nasofrontalis) and certain of the cellulae eth-
moidales anterior that drainage from these paranasal chambers
finds its way in whole or in part into the infundibulum eth-
moidale, thence via the latter into the sinus mavxillaris.
ADULT ANATOMY
In order to properly interpret points in adult anatomy it is
frequently necessary to resort to the embryology of the part or
parts involved. ‘This, indeed, is true of the nasofrontal region.
Doubtless many of the erroneous statements extant in the
ficcessus frontalis - 4 (Pe ,
Bulla eth... oon
litundib, cth
Fig. 5 From a child aged fourteen months. Note the apparent absence of
frontal furrows and folds. The whole recessus frontalis is expanding or growing
frontalward in the establishment of the sinus frontalis. After Schaeffer.
literature on the nasofrontal connections are the result of draw-
ing conclusions from a study of too few specimens, of studying
adult material alone, and of errors in interpretation due to the
fact that embryologic and adult studies were not carried on
simultaneously.
The adult nasofrontal region presents a varied anatomy, a
fact in accord with the varied genesis of the parts involved. In
the adult one usually finds evidence of the previous embryologic
condition that must have obtained in the particular case. Careful
analysis of the nasofrontal region reveals, as a rule, the deriva-
tives of the frontal furrows or pits and of the frontal folds or
NASOFRONTAL REGION IN MAN 133
conchae; provided, of course, these structures were differentiated.
As stated before, there are instances in which the lateral wall
of the recessus frontalis does not become configured by pits and
folds (fig. 5). In some specimens the adult anatomy is so altered
that interpretation is very difficult, even impossible.
It may be well here to refer to specific dissections of the
region for study and analysis. In figure 6, for example, we
have represented an adult nasofrontal region exposed for study
by the removal of the operculating concha nasalis media. There
Sinus troutalis__..... (Ww Peer sere Ccllulae eth. ant.
. WN Bt
‘ _ \ \ i!
Ductus nasotrontalis..... > \ ye Pui on SE ee rt, Bulla th.
‘
Concha ras.ied. oe _--- Sinus sphenvidali's
Litiundib. eth... Hypophysis cercbri
a
A
SS
Froc.uncinatis...___ 4 Ss Concha nas. suprema IT
Ostinm mux. ace._ £ m6 Concha nas.supremal
sine nee
ANN
AW
s Wy
Fig. 6 From an adult. Recessus frontalis and nasofrontal connections ex-
posed for study by the removal of part of the concha nasalis media. See text
for a discussion of this dissection.
is positive evidence of four embryological frontal furrows or pits.
The first or most ventral of the latter differentiated into a cellula
ethmoidalis anterior of small dimensions communicating directly
with the meatus nasi medius, medial to the processus uncina-
tus. The third and fourth frontal furrows or pits likewise de-
veloped into cellulae ethmoidales anterior both of which com-
municate with the meatus nasi medius cephalic to the hiatus
semilunaris of the infundibulum ethmoidale.
134 J. PARSONS SCHAEFFER
The second frontal furrow or pit after first developing into a
cellula ethmoidalis anterior continued to extend its boundaries
until it became topographically the sinus frontalis. It should
be noted that the duct of the sinus frontalis (ductus nasofron-
talis) is in the position of the embryonic second frontal furrow
or pit and that it is in the same axis as the infundibulum eth-
moidale and the hiatus semilunaris,! but not in direct continuity
with them. The sinus frontalis in this instance (fig. 6) com-
municates, therefore, with the recessus frontalis directly via the
ductus nasofrontalis. On the other hand, the infundibulum
ethmoidale ends blindly as a cellula ethmoidalis anterior (in-
fundibular cell) lateral to the recessus frontalis and the ductus
nasofrontalis.
The anatomy represented in figure 6 is that found in a cer-
tain number of adult specimens, and is illustrative of one of the
anatomic types of the region. It should be noted that the in-
fundibulum ethmoidale is not directly continuous with the ductus
nasofrontalis, but that it bears an intimate and important rela-
tion to it. The relation is, in a sense, a contiguous and not a
continuous one. Drainage from the sinus frontalis would find its
way partly into the meatus nasi medius directly. An explora-
tory probe passed towards the frontal region via the infundibulum
ethmoidale would, of course, find its way into the ventral, blind
end of the latter and not into the sinus frontalis. To probe the
sinus frontalis in this case it would be necessary to pass through
the proximal ostium of the ductus nasofrontalis located in the
recessus frontalis.
It is interesting and instructive to compare the embryologic
anatomy of the recessus frontalis illustrated in figure 1 with the
adult anatomy illustrated in figure 6. In the former the third
frontal furrow and the infundibulum ethmoidale are in the
same axis; in the latter, the second frontal furrow (now the
nasofrontal duct) is in the same axis as the infundibulum eth-
‘The term hiatus semilunaris should be applied to the lunate cleft which
establishes a communication between the infundibulum ethmoidale and the
meatus nasi medius, i.e., the slit between the free border of the processus unci-
natus and the bulla ethmoidalis.
NASOFRONTAL REGION IN MAN hes}
moidale. If in figure 1 the-sinus frontalis had developed from
the same frontal furrow as in figure 6, the relation between the
ductus nasofrontalis and the infundibulum ethmoidale would
have been less intimate.
The dissection of the adult nasofrontal region illustrated in
figure 9 gives evidence of the early embryologic frontal furrows
or pits. The adult derivatives of the latter are readily identi-
Cellulae cth. ant. _.-----
Cllihae eth.post. .--.. ~
Sinus sphenoidalis .
Hypophysts cerebré
Concha nas. supromal.
a /
7? So.
{
Concha nas.ned..-ff2
/ Oo:
Fig. 7 From an adult. Recessus frontals and nasofrontal connections ex-
posed forstudy. Especially note the derivatives of the frontal pits, the tortuous
and narrow ductus nasofrontalis, and the termination of the ventral extremity
of the infundibulum ethmoidale. See text.
fied. The first frontal pit developed into a small cellula eth-
moidalis anterior which is in direct communication with the re-
cessus frontalis by means of its ostium. The second and the third
frontal pits developed into sinus frontales (sinus frontalis in du-
plicate). Both of the latter communicate directly by means of
independent ostia with the recessus frontalis—no ductus naso-
frontalis being present. A study of the dissection shown in
figure 9 clearly points out that the infundibulum ethmoidale ter-
136 J. PARSONS SCHAEFFER
minates blindly (indicated by a probe) as a cellula ethmoidalis
anterior (infundibular cell) lateral to the recessus frontalis.
Loose interpretation of the anatomy of this region in this in-
stance, might lead to the erroneous statement that the sinus
frontalis developed as an extension of the infundibulum eth-
moidale. One sees even a channel-like depression on the lateral
wall of the recessus frontalis connecting in a sense the sinus
(i EEN
DAC MCE
2
\ 4
\
Fig. 8 From an adult. Dissection shows the nasofrontal connections and the
ethmoidal labyrinth exposed. Especially note the derivatives of the frontal pits
and the direct continuity of the sinus frontalis with the infundibulum ethmoidale.
The dorsal blind end of the infundibulum ethmoidale, due to a mucosal fold (X )
passing from the free border of the processus uncinatus to the bulla ethmoidalis,
should also be noted. In this specimen practically all secretion from the sinus
frontalis would find its way into the sinus maxillaris. See text.
frontales with the infundibulum ethmoidale. It is obvious that
drainage from the sinus frontales would in part find its way into
the infundibulum ethmoidale, thence via the latter to the ostium
maxillare and into the sinus maxillaris (antrum of Highmore).
In figure 7 we have evidence of four embryologie frontal pits.
The derivatives of these pits are two cellulae ethmoidales anterior
and two sinus frontales, all in communication with the recessus
NASOFRONTAL REGION IN MAN MN S3/
frontalis of the meatus nasi medius. The first (most ventral) and
fourth (most dorsal) frontal pits developed into two small cells.
The second frontal pit developed sufficiently far to be topograph-
ically a sinus frontalis (indicated in drawing as an anterior eth-
moidal cell). The sinus frontalis proper took its origin from the
cellula ethmoidalis anterior which had its genetic point in the
third frontal pit. The result of the encroachment of the cell
from the second frontal pit is a narrow channel (ductus naso-
frontalis) communicating between the sinus frontalis and the
recessus frontalis. As in figures 6 and 9, in figure 7 the infun-
dibulum ethmoidale ends blindly lateral to the recessus frontalis.
As stated in previous paragraphs, fewer frontal pits and folds
are at times differentiated in the fetus. This changes the pic-
ture of the adult anatomy of the recessus frontalis. In figure 12
there is evidence of but two embryologic frontal pits. The first or
most ventral of the latter developed into the sinus frontalis. It
should be noted that the duct of the sinus frontalis is in the same
axis as the hiatus semilunaris and the infundibulum ethmoidale.
The latter terminates lateral to the recessus frontalis as a cellula
ethmoidalis anterior (infundibular cell). The second or dorsal
pit (fig. 12) developed into a small cellula ethmoidalis anterior
in line with the suprabullar furrow (now cellula ethmoidalis ante-
rior, honeycombing the bulla ethmoidalis). This same dissec-
tion shows a well developed dorsal limb of the processus uncina-
tus (x). This causes the infundibulum ethmoidale to end in a
deep, blind pocket just over the ostium maxillare.
In an earlier paragraph mention was made of occasional adult
specimens in which the ductus nasofrontalis is in direct con-
tinuity with the infundibulum ethmoidale. In figure 8 is repre-
sented a dissection of an adult nasofrontal region in which the
ventral extremity of the infundibulum ethmoidale is directly
continuous with the ductus nasofrontalis and secondarily with
the sinus frontalis. In this dissection one notes a plate of tissue
intervening between the free border of the processus uncinatus
and the bulla ethmoidalis, thus bridging over the ventral ex-
tremity of the infundibulum ethmoidale and, in a sense, re-
placing the hiatus semilunaris in this position. One encounters
€
138 J. PARSONS SCHAEFFER
difficulty in interpreting the anatomy of the nasofrontal connec-
tions in this specimen. Did the sinus frontalis develop from the
infundibulum ethmoidale (by a direct extension or from an in-
fundibular cell) or from the second frontal pit (early cellula
ethmoidalis anterior)?
The infundibulum by its ventral and cephalic extension usually
comes into topographic relationship with some of the cellulae
\ _ Sinus trontulis
Clllula eth.ane. (Frontal topography) .
Cllulae eth. ant. (med)... .-----_ i _Cllula cth. ant.
tll ml
mW R
Fifi (UU
| oN \
a
Lellulae hh. posts ~ a <i i : \\ ___ Mecessus trontalis
( ©
~2\ Concha nas. med.
ef pS ee unc alas
\ |
BE) \
)
\ )
a
‘ t
Fig. 9 Dissection from an adult. Note the two sinus frontales, the absence
of ductus nasofrontalis, and the ventral termination of the infundibulum eth-
moidale. See text.
ethmoidales anterior which arise from the frontal pits. In this
instance (fig. 8), a relationship may early have been established
with the second frontal pit (there is evidence in support of this
belief).
Resorption of the intervening barrier would, of course, bring the
infundibulum ethmoidale in direct continuity with the cellula eth-
moidalis anterior arising from the second frontal pit, likewise
with the sinus frontalis. The dissection gives positive evidence
of three frontal pits (now cellulae ethmoidales anterior). Whether
NASOFRONTAL REGION IN MAN 139
an additional frontal pit which gave rise to the sinus frontalis
was present in the position of the ductus nasofrontalis is, of course,
impossible to say. ‘Two of the cellulae ethmoidales anterior are
separated by a considerable interval. This space may have been
the second frontal pit. Again, the two frontal pits in question
(cellulae ethmoidales anterior) may have been crowded apart by :
Sinus trontalis.---- een Mh _-Cellulae eth. ant.
Probein ductus.nasotrontalis t Bee _Hypoephy sis corebri
Sinus trontalis--------~ (Sinus sphonoidalis
Y Cone jaca.
Latundib. hh... 22... Zag _Contha nas. Hit
Je 7 <x iG
PBocuncinatus -. yy, S| Ostium pharyngeun
// AO
// fubac auditivac
/
Bulla cth.--- “Tf y =
Fig. 10 Dissection from an adult. Note the sinus frontalis present in duph-
cate, the proximal ostia frontales in relation to the recessus frontalis, and the
ventral termination of the infundibulum ethmoidale. See text.
bullous-like ventral and cephalic growth of the infundibulum eth-
moidale in the establishment of the sinus frontalis, My experi-
ence has been that it is unusual for the sinus frontalisto arise
from the infundibulum ethmoidale.
Drainage from the sinus frontalis in such instances (fig. 8)
would almost wholly pass into the infundibulum ethmoidale, and
via the latter to the ostium maxillare, thence into the sinus maxil-
laris. Should the floor of the infundibulum ethmoidale in such
eases be largely replaced by an elongated ostium maxillare (a
140 J. PARSONS SCHAEFFER
rather common occurrence), the sinus frontalis and the sinus
maxillaris would from a practical viewpoint be in direct communi-
cation. It should be recalled that the sinus maxillaris is geneti-
cally an outgrowth from the floor of the infundibulum ethmoi-
dale. The initial area of the outgrowth varies considerably in
extent, thus accounting for the varied size of the adult ostium
maxillare.?
The sinus frontalis is occasionally present unilaterally or bi-
laterally in duplicate or in triplicate. In these cases each sinus
frontalis is absolutely independent of others and possesses an
individual ostium frontale. The condition of duplicity or tri-
plicity of the sinus frontalis is readily explained when one re-
calls the potentiality of development referred to in previous
paragraphs (figs. 1 to 5). In figures 9 and 10 are represented
dissections of adult nasofrontal regions in which two frontal pits
(early ventral or anterior cellulae ethmoidales) developed
sufficiently far to be topographically sinus frontales. Dupli-
cate sinus frontales are either side by side in the sagittal plane
(fig. 10) or are ventral and dorsal in relation, in the coronal plane
(fig. 9). Intermediate relations are, of course, encountered. In
figure 11 the first and second frontal pits developed into sinus
frontales; in figure 10 the second and third. In both instances
the sinuses communicate independently with the recessus fron-
talis of the meatus nasi medius. At times when the sinus
frontalis exists in duplicate (or triplicate) one sinus may en-
eroach bullous-like on the other. The name bulla frontalis was,
however, applied by Turner to infundibular cells which encroach
upon the dorso-caudal boundary of the sinus frontalis.
The ductus nasofrontalis is a very variable channel. One en-
counters very many specimens in which no true duct is present.
2 “Tn my series of 90 cases it (the ostium maxillare) has a great range of di-
mensions; varying from 1 to 20 mm. in length and from 1 to 6 mm. in width. In
some instances where the ostium has reached considerable size, it almost entirely
replaces the caudo-lateral wall of the infundibulum ethmoidale, thus forming a
long, slit-like communication between the sinus maxillaris and the infundibulum
ethmoidale.”’ J. P. Schaeffer, Am. Jour. Anat., vol. 10, 1910, p. 351.
NASOFRONTAL REGION IN MAN 141
Witness, for example, the dissection represented in figure 11.
Here the sinus frontalis is very small. In fact, partly eth-
moidal in topography. The interesting thing about this case
is that the sinus frontalis was bilaterally very diminutive in size.
Its communication (fig. 11) with the recessus frontalis of the
meatus nasi medius is established by means of a large ostium
frontale in the same axis as the infundibulum ethmoidale.
Ostium frontale
Kecessus trontalis - :
_.-- Sinus trontalis
bes Litundib. eth,
_ Loc.uncinatus
Fig. 11 Dissection from an adult. Recessus frontalis exposed. Note the
diminutive sinus frontalis and the absence of a ductus nasofrontalis. ‘The
views held on the presence and absence of the sinus frontalis are, doubtless, largely
due to differences of opinion as to what should be called a sinus frontalis, and how
far the development must have progressed into the frontal region before the cell
has reached the dignity of a sinus frontalis.’”’ (J. P. Schaeffer).
The dissection shown in figure 9 likewise presents the sinus
frontalis in duplicate in which no ductus nasofrontalis is pres-
ent. Each sinus communicates with the recessus frontalis by
means of a large ostium frontale. On the other hand, one en-
counters specimens with true ductus nasofrontales. Some of
these ducts are straight and short (fig. 6), others straight and
long (figs. 12 and 13). Again, the ductus nasofrontalis may be
long and more or less serpentine. Witness, for example, the
specimen shown in figure 7. Here is a sinus frontalis with a long,
147, J. PARSONS SCHAEFFER
narrow and curved ductus nasofrontals. The duct communi-
‘ates with the recessus frontalis. There are very definitely two
ostia frontalia to the duct, one proximal and the other distal in
position. The duct is encroached upon by a cellula ethmoidalis
anterior (really a second sinus frontalis) which developed from
the second frontal furrow. The slightest swelling of the mucosa
of such narrow and tortuous ductus nasofrontales (figs. 7 and 10)
would, of course, occlude its lumen. In some instances the
_- Cellula eth. ant.
Cellula chh. ant {med
Duetus nasotrontalis . / oe
Littandib.eth.
< emia hig tls (2. -Ostium sphenoidale
Concha nas.7med..___ BSN a Nts {2 AN Dp)
4 z We S5e, ( .
\ XS ne ? by i : 1
Ostiummaxillare - {6-eO G BEMIS as -..Balla cth.
6 eee S Ostinin pharyn-
* \genne tibae
Ale y aUathivae
hee
e |
\
Fig. 12 Dissection from an adult. Recessus frontalis exposed. Note the
discontinuous channels, i.e., the ductus nasofrontalis and the infundibulum
ethmoidale.
ductus nasofrontalis is roomy, possessing large proximal and
distal ostia, thus affording a better drainage channel for the sinus
frontalis (fig. 13).
The ductus nasofrontalis with its proximal ostium or in the
absence of a true duct, the proximal ostium frontale (in the
latter the distal ostium frontale is wanting), bears a varied rela-
tion to the ventral extremity of the infundibulum ethmoidale.
The latter usually ends blindly lateral to the terminal portion of the
ductus nasofrontalis. The infundibulum ethmoidale and the duc-
NASOFRONTAL REGION IN MAN 143
tus nasofrontalis are at times in the same axis (figs. 6 and 12).
Again the ductus nasofrontalis with its proximal ostium is not in
line with the infundibulum ethmoidale. Witness, for example,
figure 10: Here the proximal ostium frontale is located medial
to the cephalic extremity of the processus uncinatus. Drainage
in such cases would in a large measure be diverted directly into
Ccllula eth. ant. (mea)
Sonus trovtalis.._-.
_..Cllula eth. post.
: ae. yan tase) we ‘
Ductas nasotrontilis va v Recessus spheno—eth.
Cllala ath.ant... {i> __ NOptivns
loncha nas.mea. i sats
Se
Willa Cl aes ae
Intundib. eth. ---- | Af . 4
\
f/ S
/ },
fonchanas.med. /-/4
Fig. 13 From an adult. Note the roomy ductus nasofrontalis, discontinuous
with the infundibulum ethmoidale. The intimate relation between the sinus
sphenoidalis and the A. carotis interna is also indicated in the dissection. The
section is to the right of the mid-sagittal plane, hence the absence of the
hypophysis cerebri.
the meatus nasi medius. In figure 7 the relation is not one of
alignment. In figure 9, passing from the proximal ostia fron-
talia, on the lateral wall of the recessus frontalis towards the
infundibulum ethmoidale, is a shallow, gutter-like channel.
Drainage from the sinus frontales here would largely find its way
into the cephalic end of the infundibulum ethmoidale.
In the specimen shown in figure 8 in which the infundibulum
ethmoidale and the ductus nasofrontalis are in direct continuity,
144 J. PARSONS SCHAEFFER
drainage from the sinus frontalis would, of course, readily find
its way into the sinus maxillaris. The infundibulum in these
‘ases acts In every sense as a gutter of communication between
the sinus frontalis and the sinus maxillaris. The same is true
in a large number of cases in which the relations are intimate but
not continuous between the ventral extremity of the infundibu-
lum ethmoidale and the derivatives of the recessus frontalis.
The sinus maxillaris in turn becomes a reservoir for drainage
from the sinus frontalis and certain cellulae ethmoidales anterior
(some infundibular and others frontal in position). If in those
‘ases In which the sinus frontalis or its duct is directly continu-
ous with the ventral extremity of the infundibulum ethmoidale
(fig. 8) the ostium maxillare should occupy a goodly or greater
portion of the floor of the infundibulum ethmoidale (a condition
encountered), the sinus frontalis and the cellulae infundibulares
would, from a practical viewpoint, be in direct communication
with the sinus maxillaris. This close relationship is, however,
secondary and one must not infer that the frontal and maxillary
sinuses and the infundibular cells arise from the same point.
CONCLUSIONS
The materials studied for this paper seem to justify the fol-
lowing conclusions:
1. The sinus frontalis is in the vast majority of cases a deriva-
tive (a) of the recessus frontalis directly, (b) of one or more
of the cellulae ethmoidales anterior which have their genesis
in frontal pits, or (c) of both, when present in duplicate or
tripheate.
2. The sinus frontalis appears occasionally to arise from the
ventral extremity of the infundibulum ethmoidale. This rela-
tionship, however, is in some instances secondary owing to de-
velopment. It is questionable whether the sinus frontalis ever
develops from the suprabullar pit or furrow.
3. The ductus nasofrontalis (or the sinus frontalis directly in
the absence of a ductus) and the infundibulum ethmoidale are
in the vast majority of instances, discontinuous channels in the
NASOFRONTAL REGION IN MAN 145
adult. The topographic relationships may, however, be very
intimate, i.e., they may be contiguous.
4. The ductus nasofrontalis (or the sinus frontalis) and the
infundibulum ethmoidale are oceasionally directly continuous in
the adult.
5. The infundibulum ethmoidale in approximately as many
as 50 per cent of adult bodies acts as a channel for the carriage
of secretion or infection from the sinus frontalis and certain of
the cellulae ethmoidales anterior to the sinus maxillaris: In-
cluded in this group are (a) those cases in which the sinus fron-
talis and the infundibulum ethmoidale are continuous, and (b)
those cases in which the sinus frontalis and the infundibulum
ethmoidale are discontinuous, but intimately and vitally related
topographically.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 1
146 J. PARSONS SCHAEFFER
LITERATURE CITED
Cutari, O. 1902 Die Krankheiten der Nase. Vienna.
Davis, W. B. 1914 Development and anatomy of the nasal accessory sinuses
inman. Philadelphia, Saunders.
Kiny1an, G. 1896 Anatomie der Nase menschlicher Embryonen. Archiv f.
Laryngologie, Bd. 3, Bd. 4.
1900 Heymann’s Handbuch der Laryngologie, vol. 3.
1908 Die Erkrankungen der Nebenhéhlen der Nase bei Scharlach.
Ztschr. f. Ohrenh., Weisb., Bd. 56.
LanceEeR, C. unp Tonpt, C. 1902 Anatomie.
Onopi, A. 1910 The accessory sinuses of the nose in children, New York,
Wood.
PorriEeR, P. 1903 Traité d’ anatomie. Paris.
ScHAEFFER, J. Parsons 1909 Some practical considerations on the sinus maxil-
laris. Univ. of Pennsylvania Medical Bulletin, vol. 22, no. 8.
1910 The sinus maxillaris and its relations in the embryo, child, and
adult man. Am. Jour. Anat., vol. 10, no. 2.
1910 On the genesis of air cells in the conchae nasales. Anat. Rec.,
vol. 4, no. 4.
1910 The lateral wall of the cavum nasi in man, with especial refer-
ence to the various developmental stages. Jour. Morph., vol. 22,
no. 4.
1912 The genesis and development of the nasolacrimal passages in
man. Am. Jour. Anat., vol. 13, no. 1.
TiLuaAux, P. J. 1882 Anatomie. Paris.
Turner, A. L. 1901 The accessory sinuses of the nose. Edinburgh.
ZUCKERKANDL, E. 1893 Normale und pathologische Anatomie der Nasen-
hohle und ihrer pneumatischen Anhinge. Bd. 1, Bd. 2, Wien und
Leipzig.
THE LACHRYMAL GLAND
JOHN SUNDWALL
From the Hull Anatomical Laboratory, University of Chicago and the Department
of Anatomy, the University of Kansas
TWENTY FIGURES
CONTENTS
em ELOUUCLION Sere ee 427s ele © shies 2 > SRO ten sere obra be ake 148
NEG TOSSICHATACTENISOLCS! staan stisig veestas 5 c\Gia een neNanee eR ale sb eua arene oete est 152 /
Gropspstructiner and relations -yy....¢ S475 Sake accel eateede dsp 152
WiasCul anes Up pl vans weed itbramiecis snot phe ed neat a a etoeueee emus 153
ION VCS RR ace inact arco NLL SILT ile) LAIR RONRE naeatse ENC elt sich th 2h 153
ID AWVa ALU Ley goes Kol ah eins deve chia icabieet etn ae neto rns cleo © cache enOicereRe al omreatin taint 153
TES SW POLGIN CwbISSIE see Pra sir. Site ctaaitetesee Secs vane PeEeING Hud mamttleeds ate 155
SED) CYST MEks 33 aps oh Ite eae eV eM nA Ea ORR ee CEE: EOC LAREN 155
MME MIMGET OM ULAR IS pibaMe a Maet re mclcrskirtie rey To CET ete Sree 156
Maley shat He) Voy OLMIS OLA econ prom amos am p.c.cu coue oad otics 6 pido Gomer 157
TEMIEWE NFS 11 OV StS ea eR eA Sn Se RC a Ea heal! Meters Bieri ray eer oe aS 157
Peritubular, connective tissue, ete. <i. cca ad pels Rasbaee atte sisie:lsela « 158
Blasi ance llsisaeyece mee eset tree Seyi Nios 25 Lh i ae eR ela ae ci ee ea it ul 161
iV Duct systemman qGulo mle sigur. 4 stkos cts-< 05 ava cneasis Cnetkacenerae a caste. 6 5 aller os 162
Injection-corrosion specimens of duct system................0.2055 163
Dienst Tiss Geb Tata oe ee ene Set pore, yt sass. Nan a Mise are yey teres eee at 166
Histological structure of ducts and tubules:
IN IITTg CUI CIES ease eae eee cua teaste staid teers erate Se estas Ree es Fae 167
BEIM Ary, GUC tS eet ce AIO eee PA iene ee lea cine aoe 169
anderlioh ular GUuUcésts cst myst ok tatinke Spin caries ea miata lene ois ighe = 171
Inatral ool ar tu CGS a ctevsasc eens ce ieee er eee eer oie saat 173
Opera a eyant ii 0) pe RR Se ee ee ie oe Al me ORO a cot 174
TENGE CAary HCC US ham i to cud Acetone tes See ere ee tee else oherasel sts 175
UU OTC Rees o pe eC EIb COPRSIS. aTeccREAL SG Stee! sae PRES Set ea Cee tn 179
ee SORE Rirnie pra s e ts oy Aira te eeepc cs Seal Fo NE fa Sy sched) cps eigha we oul g} «= 180
SSiROVS Es) alli 2 S25 Saal eh ces Seas Bate ie ORCI a ion oo band aired tein 180
PUULEEIS PCOS cr perenne PU as Wes, Pa ee ac tac ye ss SRM EA EOS tte: hate) ois 181
Similarities of mucous and serous. cells. 06.550 ).0.0es See ee ene es 182
Granules in lachrymal gland:
GES MGISSU Ch eee ies Mena Se ree Tes. aici) oss. caBactolctegtecaer eked ster eatiahov = 184
imeixed) and: stained Preparauions.. . ca... cemeade sees as os a se LOS
Piscussionton Granilesye es fe. ke 0k, eR ae ak 193
147
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 2
SEPTEMBER, 1916.
148 JOHN SUNDWALL
ViesMitochond rian sachs: 5 Aes ae She ee Ore Ce eee oe Cee 202
Tab il CS aes als te deaths tons io RIE ae A Ch OREO ee nn ein oe 202
Intercalary duct. 2: .cen:s phenpoc tees eee et tet oasis cane eae 205
War eer Guctse ssh <ad..iec nee ee PO a oe ne ee ee ee 205
Vile Basalostriations: .226 2. 2.3505 See ae oa. kee 208
16) BG Gee oh: eee on ate RN CAN eo oes, heen Ce RMS Oo oe so 2 211
IX. Secretion capillaries and cement substance................-+.+s020-- 213
XCrCanahicularapparatus: > seme sk oes Oc eee 219
Mol see chnirqMe sia calles cen he A eed te cae foot ahs ence dsicanes ene eke Oe cee eee 221
XII, Miscellaneous—Other fixations and stains..............2..:0:.-.20- 224
D.C ieare bb rnisith oh eera ein woPEseS om ooo 5 aa oo RGD TADS Omics coos boos 226
MV. “Bibliographiy-%: 5. sccm ese sae Sissel eid sine sal scan: qe eee ea 230
I. INTRODUCTION
The investigations on which this paper is based began several
years ago (’05) when it was first learned that lachrymal glands
as well as those of the nictitating membrane, plica semilunaris, of
many animals stain with certain dyes which are generally re-
garded as specific to mucous secreting cells. It was found that
these glands of the ox, pig, sheep, goat, and horse stained readily
with either mucicarmin or muchaematein, while those of the cat,
dog, rabbit, and guinea pig did not. Preparations of the lachry-
mal glands of these animals were shown at the twenty-second
session of the Association of American Anatomists held at the
University of Wisconsin, and a preliminary report on the stain-
ing reactions of the lachrymal gland tissue found in the nictit-
ating membrane of-the ox was published in 1906.
It was the intention at that time to publish in detail the results
of the investigations on the staining characteristics of these
glands. There was a delay, however, and in the meantime a
paper by Hornickel (’06) on this particular subject appeared.
In this he claimed that by the use of mucicarmin a positive
stain was obtained on the lachrymal glands of the pig, sheep,
goat, and dog, while similar glands of the horse, ox, ass, and cat
were negative. His results in some instances were opposite
to mine; namely, I obtained a positive stain in these glands of
the ox and horse while the glands of the dog were negative. The
difference in the results obtained in the two laboratories is inter-
esting. Hornickel’s paper served as added impetus to further
THE LACHRYMAL GLAND 149
work on my part, and I decided that a detailed study of the
structure of this gland in one species might well precede a general
comparative study.
In view of the work already begun on the Harderian gland in
Bos and the availability of the lachrymal gland in this species
for both structural and chemical study, it was chosen for this
purpose.
Numerous contributions have appeared dealing with various
phases of the lachrymal gland of different animals—such as its
phylogeny, embryology, innervation, gross and microscopic
structures, etc. Chief efforts, however, have been directed
toward the study of the structure of the cells and the secretory
changes they undergo during the stages of rest and activity.
The origin and distribution of the gland in vertebrates have
been described by Wiedersheim (’76) and Sardemann (’87).
Wiedersheim (’07) states:
The first attempt of a vertebrate to exchange an aquatic for an
aerial existence necessitated the development of a secretory apparatus
in connection with the eye. Thus in Urodeles a glandular organ is
developed from the conjunctival epithelium along the whole length of
the lower eyelid; in Anurans and Reptiles this becomes more developed
in the region of the anterior, and in many Reptiles also of the posterior
angle of the eye, the original connecting bridge gradually disappearing:
thus two glands are developed from the primitively single one, each of
which becomes further differentiated both histologically and physio-
logically. From one is formed the Harderian gland—anterior angle
of the eye—while the other, posterior angle, gives rise to the Lachrymal
gland. In Crocodiles, Snakes, and Hatteria, the lachrymal gland is
wanting while in the Chelone it is extremely large.
Further contributions regarding the comparative anatomy of
the lachrymal gland are given by Van Trotzenburg (’01) who
studied it in both the old and new world monkeys and its phy-
logenetic relation to man.
Contributions to the embryology of this gland and its accesory
structures have appeared from time to time. Falchi (’05)
studied the development of the gland in the rabbit, guinea pig,
sheep, andhuman. Monesi (’03), Matys (’05), (06), Kiisel (’06),
and Lang (11), also, have contributed to this phase of the sub-
ject. Their efforts, however, have been concerned chiefly with
150 JOHN SUNDWALL
the lachrymal sae and canal. Speciale-Cirincione (’08) states
that in man—‘‘Die Anlage der Trénendriise beginnt mit dem
2. Monat, zu einer Zeit, wo die Lidspalte noch weit offen steht.
Die Entwicklung geht alsdann sehr rasch vor sich, so dass von
einem Tage zum anderen eine erhebliche Aenderung zu bemerken
ist.”’
The innervation of the lachrymal gland has been studied by
Dogiel (’93), Jendrassik (’94), Tepliachine (’94), Laffaye (’97),
W. Klapp (97), P. Klapp (97), Landolt (’00), Puglisi-Allegra
(03) (04), and Schirmer (’04). The works of Dogiel and
Puglisi-Allegra are of particular interest here as the others are
chiefly contributions from a clinical standpoint.
Dogiel used the methylene-blue and Golgi’s methods on the
lachrymal glands of the guinea pig and rabbit. According to
him ‘Die Thranendriise empfingt fast ausschlieslich marklose
Nervenfasern, solche, die Blutgefaisse und die Ausfiihrungsgiinge
umflechtend, . . . und bilden ein Geflecht in dem auf der
memb. Propria derselben sich lagern.’’ <A_ peritubular and
intercellular net work of these fibers is described.
Puglisi-Allegra (03), using Golgi’s method and the vital
staining with methylene-blue, describes peritubular, intraperi-
cellular, intracellular fibres, and sympathetic ganglia cells situ-
ated on the course of the sympathetic fibres.
Schirmer describes both medulated and nonmedulated nerve
fibres in this gland.
Clinical observations seem to favor the view that the secre-
tory fibres are derived from the facial nerve through the Nervus
petrosus superficialis major by way of the Ganglion Spheno-
palatinum (Klapp).
The vascular supply and lymphatics have been studied by
Puglisi-Allegra (’04).
The position and gross characteristics of the gland have been
described by Ellenberger (06) and by Ellenberger and Baum
(08).
Regarding the finer structures of the gland there have been
many contributions. Not all are agreed, however, as to the
form of the secreting elements (whether tubules or alveoli).
THE LACHRYMAL GLAND 151
Leydig (’57), Frey (59), Henle (’73), Ellenberger (’88), Toldt
(01), V. Ebner (02), Bohm and Davidoff (03), and others
have described the lachrymal gland of man and domestic animal
as being in general tubulo-acinous in form. Flemming (’88),
Stohr (91), Zimmermann (’98), Sobotta (02), and Schirmer,
among others, describe it as a compound tubular gland, while
Franck (’83), Langer (’90), and Leiserung-Mueller-Ellenberger
(90) describe it as being of the acinous type.
Boll’s (68) research on the lachrymal glands of the pig, sheep,
calf, and dog was one of the first of importance. His work was
confined to the star-shaped supporting cells surrounding the
alveol. Later (71) he compares the structure of this gland to
that of the salivary gland.
Schwalbe (87) divides the gland into “‘eine gréssere compacte
obere, die obere Thranendriise (Glandula lachrymalis superior s.
innominata Galeni; Portio superior s. orbitalis [Sappey] und eine
kleinere aus locker geordneten Lippchen gebildete untere, die
untere Thrénendriise (Glandula lachrymalis inferior s. glandulae
congregatae Monroi; Portio inferior s. palpebralis [Sappey]).”’
Much of the later investigation has been confined to the study
of secretion granules and the secretory changes in the lachrymal
gland during rest and activity. Among the most important of
these are the contributions of Langley (’79), Reichel (’80),
Nicolas (’92), Solger (96), Kolossow (’98), Lor (98), Noll (01),
and Puglisi-Allegra (’04).
Maziarski (02), using the method of Born, constructed a
model of the human lachrymal gland. ‘‘Es ist eine deutlich
tubulése Driise; es wire also ganz unrichtig, sie mit den Speichel-
driisen zu vergleichen.” ;
Fleischer (’04) follows with an extensive article on the structure
of the lachrymal gland of the ox.
Numerous other contributions have appeared dealing with
special phases of the lachrymal gland, among which may be
named Merkel (’83), Kirchstein (’94), Stanculéanu and Théor-
hari (98), Garnier (’00), Axenfeld (’00), Alt (’00), Dubreuil
(07), Gotz (08), and Riquier (’11).
152, JOHN SUNDWALL
II. GROSS CHARACTERISTICS
The following description of the lachrymal gland in the ox is
made from careful dissections after embalming heads and in-
jecting the arteries. :
Gross structure and relations
The gland in general is a flattened oval or almond shaped
structure situated on the superior contour of the bulb, with an
auricular appendage which descends on the outer or posterior
contour of the bulb. The gland is more or less moulded to
conform with the bulb and the bony orbit so that the superior
surface of the oval mass as well as the outer surface of the ap-
pendage is convex in contour while the surface directed towards
the bulb is concave. Both surfaces possess the characteristic
glandular lobulations.
The anterior or outer margin of the gland including the ap-
pendage measures on an average from 5 to 5.5 em. in length.
The long axis of the superior mass is from 3 to 4 em. in length,
the average width of the mass being 3 cm. and its average thick-
ness 1 em. The superior mass is generally termed the Pars
superior while the appendage is known as the Pars inferior or
accessorius. The Pars inferior or appendage generally measures
about 2 em. in length, 1 em. in width, and 4 mm. in thickness,
but this portion of the gland is subject to much variation.
The weight of the entire gland mass averages in the adult
ox from 6 to 7 grams. The weight of the appendage (Pars
inferior) is from 1 to 1.5 grams. It will be seen from a com-
parison of these weights that the larger mass of the gland is
confined to the Pars superior.
The Pars superior is situated on the superior and posterior or
temporal half of the Bulbus oculi while the long axis of this oval
shaped mass is directed obliquely posteriorly and medianward.
The anterior pole is located immediately above the insertion of
the M. levator palp. sup. and M. obliquus sup. This corresponds
to a point immediately behind the center of the Margo supra
orbitalis. From this point, the lateral margin of the Pars super-
THE LACHRYMAL GLAND ae
ior extends to the posterior or temporal canthus just under and
slightly within the Margo supra orbitalis. At this point the
Pars inferior or appendage takes origin and extends downward
and behind the bulb, just within the bony orbital wall. It
terminates at the level equal to the insertion of the M. obliquus
inf. and the inferior margin of the insertion of the M. rectus
lateralis.
The gland is completely enclosed within the orbit by the peri-
orbital fascia, a thick fascia which surrounds and encloses all
the structures of the orbit and lines the bony orbital wall. The
gland itself is embedded and surrounded by a condensed and
tough mass of areolar tissue and fat which is a part of this general
tissue which fills the interstices between the muscles, nerves, and
bulb. Bands of dense: tissue derived from the inner surface
of the periorbital fascia are fused with the capsule of the gland.
Vascular supply
The gland derives its blood supply from the A. lacrimalis
which is a branch of the A. ophthalmica. The artery enters
the gland from the interior or bulbar surface by numerous
branches. A branch of this artery continues and enters the
superior palpebral fascia. Corresponding veins leave the gland
and are collected into the V. lacrimalis which empties into the
V. ophthalmica.
Nerves
The innervation is from the N. lacrimalis. Two distinct
lachrymal nerves take origin from the N. ophthalmicus near
its origin. These nerves follow the general course of the blood
vessels (fig. 1.) .
Ductult excretori
From six to eight ducts carry away the secretion of the gland.
These ducts leave the lateral margin of the gland, enter the
palpebra superior and terminate by piercing the superior con-
junctiva about 1.5 em. internal to the free margin of the superior
154 JOHN SUNDWALL
Glandula lacrimalis a -
oo a cous M.reclus
ieee SU
R ‘ Vas a, ;
V. lacrimalis <M Y M. levator
, Palp.
4 y = Sup.
A. lacrimalis YL UY -
fy, ZB ca
N lacrimalis yf fy YE
Y Z —_
N fronialis
G. semilunare
(Gasseri)
N- opthalmicus
Fig. 1 Gross structure of gland showing relations.
I. Concave surface with vessels and nerves are shown.
From dissection, Tech.
lid. These terminals are arranged in a straight line on the con-
junctiva extending from its center to the external or posterior
canthus. ‘The distance between these terminals averages about
0.5em. The openings readily admit a porcupine quill (fig. 2).
Quills in
Terminals of
ducts
i
a s 2 ‘3 £ Ducfuli
ye S He Pan excrelorii
Angulus oculi Vf: ot Lp \ (ferminals)
med. (@nt)
ADB Avie dain cel
Fig.2 Terminals of Ductuli excretorii with quills inserted. From tissue
‘prepared by Tech. I.
THE LACHRYMAL GLAND 155
The ducts range from 1.5 em. to 2 em. in length. They are
more or less tortuous in their course as it is impossible to obtain
a duct in its entirety even in thick sections. There are seen on
the conjunctiva surrounding these terminals aggregations of
lymph nodes. From four to six ducts leave the Pars superior
while from one to two leave the Pars inferior (fig. 3).
M rectus sup.
M levator palp. sup.
Fars superior
gland /ach
es ee Li EE $/ Mrectos lat
¥ oT vias 3 EL fZ Fars inferior
Ifill | i ia YA gland lach.
palp. sup. Hi By: Zz
BB AFAZZ .
Mihtiatdaix ae. B Zs Duetul excreforii
Fig. 3 Pars. superior; Pars inferior; Ductuli excretorii. Specimen prepared
by Tech. I.
III. SUPPORTING TISSUE
For the study of the capsule and connective tissue framework
of the gland the methods and stains used were: Flint and Spalte-
holz, Mall’s method and a modification of his method for the
demonstration of reticulum, Mallory’s, Van Gieson’s, Weigert’s,
and Herxheimer’s. (See Technique.)
Capsule
The capsule is fused with the general areolar and adipose
tissue of the orbit. Strands of tough connective tissue fibres
derived from the periorbital fascia are also intimately fused with
the capsule. It is bound down to the gland by numerous inter-
lobular septa which take origin from the capsule and ramify
throughout the gland thus separating it into numerous lobules.
It varies in thickness. In general that portion covering the
156 JOHN SUNDWALL
convex surface is thinnest, averaging 0.5 m. Where the septa
take origin in this area it is much denser. The thickest portion
is found on the concave surface.
It is composed of areolar tissue in general. Both collagenic
and elastic fibres are present in abundance. Numerous globules
of fat are seen singly and in groups irregularly distributed through
out, being more abundant around the large vessels. Small
lobules of gland tissue are also included within the capsule.
These are more numerous in the neighborhood of the exits' of the
main ducts. Smooth muscle fibres are also present irregularly
distributed. On the concave surface the capsule contains the
large blood vessels and nerves while on the outer margin of the
gland the large ducts are surrounded by the capsular tissue.
The elastic fibres are long and fine and are arranged parallel
to the surface of the capsule, as described by Fumagalli (97),
Riquier, and Schirmer. They are most numerous on the con-
cave surface where they surround the larger vessels. In the
region of the exits of ducts the elastic fibres are abundant also
and interlace more than elsewhere, forming a network around
the ducts.
The interlobular septa
The interlobular septa, which are derived from the capsule
and divide the gland into numerous irregularly shaped and sized
lobules, are composed of practically the same tissues described
in the capsule. Secondary septa take origin from two or three
large primary septa and enter the various lobules, forming the
supporting structures of the lobules and surrounding the various
vessels, nerves, and ducts, which are in close proximity to each
other.
Observations inthis laboratory regarding the capsule and septa
agree with those of Schirmer. However, I have failed so far
to see in the connective tissue framework of the gland—septa or
intralobular tissue—of either young or old Bovidae a constant
accumulation of lymphoid cells such as described by Schirmer
for man and Fleischer for the lachrymal gland of the ox. Axen-
feld and Gotz have also referred to lymphoid infiltrations in
THE LACHRYMAL GLAND 157
man which increase from childhood to old age. According to
Bensley these are probably plasma cells.
The supporting elements finally ramify and surround each
acinus, or tubule, forming a network or reticulum. The larger
septa contain fat and small glandular lobules in addition to the
vessels, nerves, and ducts. The septa are composed chiefly of
eollagenic fibres (Van Gieson’s). Elastic fibres are present to
some extent in all the interlobular framework, especially around
the ducts and vessels. Cells suggesting smooth muscle cells are
sometimes seen, in addition to those seen’ in the vessel walls,
occurring very sparsely and as a rule singly in the neighborhood
of the ducts.
The intralobular septa
The intralobular connective tissue septa are practically com-
posed of collagenic fibres and reticulum. In the larger of these,
only, are elastic fibres seen independent of the vascular walls.
In the smaller septa well within the lobules elastic fibres are
seen only in connection with the blood vessels. No elastic
fibrils are seen surrounding the alveoli of the gland.
Elastic fibres
My observations regarding the final distribution of elastic
fibres do not entirely agree with Boll (71), Schirmer, and Fuma-
galli, who claim that elastic fibres surround the alveoli, or tubules.
Regarding these fibres Riquier remarks—‘‘nei sepimenti con-
nettivali attorno ai vasi ed ai condotti escretori, sono invece
difficilimente dimostrabili intorno ai tubuli secernenti.”’
My investigations indicate that elastic fibres do not enter the
lobules and surround each tubule. Repeated examinations of
tissues fixed in various solutions and stained both by Weigert’s
and the Unna-Taenzer method—the latter was used by the
Italian investigators—did not show the presence of elastic
fibres surrounding the acini.
In using these stains, it is most essential to fully decolorize
the section. Otherwise the usual collagenic fibres and reticulum
158 JOHN SUNDWALL
may assume the staining characteristics of elastic fibres. This
is especially true of the Unna-Taenzer orcein method which
stains elastic fibres brown. It was my practice to place the
sections from the stain in 70 per cent alcohol and then decolorize
under the microscope with acid alcohol until the section became
for the most part colorless. The characteristic elastic fibres of
the blood vessels remain deeply stained and serve as controls
for the degree of decolorization. In the use of Weigert’s it is
also necessary to decolorize in alcohol until the section becomes
yellowish or light gray. The elastic fibres alone should be stained
black.
With ‘the proper degree of differentiation, using the elastic
fibres of arterial wall as the criterion, it will be found that elastic
fibres are present, as described, in the capsule; to a less degree
in the interlobular septasurrounding the blood vessels; and in the
basement membrane of the main, primary, interlobular, and
larger intralobular ducts. Only occasionally are they seen in
the walls of the small intralobular ducts. They are not present
as forming the walls of the acini except in rare instances, and
then when lobules are in contact with capsule.
Peritubular connective tissue, ‘Korbzellen’, basement membrane
The finer structural tissue surrounding the terminal tubules or
acini of glands in general is composed of several elements such as
fixed connective tissue cells, the so called ‘Korbzellen’ or basket
cells, and the basement membrane. Much confusion exists in
the literature respecting the terms used to designate these vari-
ous tissues. Some refer to all these elements as the basement
membrane. Shafer (’12) states:
In most glands the secreting cells of the alveoli and also the cells
which line the ducts are bounded. . . by a thin membrane, which
is sometimes continuous, sometimes interrupted, and which has nuclei
here and there scattered upon it. This is the basement membrane
and as the presence of nuclei indicates, it is composed of more or less
fused flattened cells of connective tissue nature which are sometimes
united edge to edge, sometimes connected only by branch processes
so as to form a sort of flattened basket-work around the alveoli. Even
THE LACHRYMAL GLAND 159
in this case the meshes of the basket-work are not quite empty, being
occupied by a delicate filmy membrane which is a condensation of the
reticular connective tissue.
Flint and others refer to this latter alone as the basement
membrane.
Boll (68) was the first to describe ‘Korbzellen’ in the lachrymal
glands of the pig, sheep, calf, and dog. Owing to the resemblance
of these cells, in transverse sections, to demilune cells, he regarded
them as similar in nature to those cells described by Gianuzzi
in salivary and mucous glands. Cells similar in structure had
been described in other glands by Krause; Henle, who regarded
them as nerve cells; Pflueger; and von Kolliker. Others held the
same view as Boll regarding the function of these cells. Noll
(01) regarded these ‘Korbzellen’ as having completed a stage of
secretion and having been pushed back against the wall by neigh-
boring cells filled with secretion. In a later communication Boll
(71) refers to these cells as the basement membrane.
Kollosow refers to the cells described by Boll as muscular
epithelium, as did Zimmermann (’98). Schirmer and Puglisi-
Allegra described similar cells in lachrymal glands, to which
they ascribe the function of contractility.
These ‘Korbzellen’ are readily seen in my Zenker-Van Gieson
preparations. They are more prominent in the lachrymal glands
of younger animals. Frequently these irregular anastomosing
cells are seen between the deep red staining reticular membrane
and the epithelial cells of the tubules, often obscuring the latter.
I am inclined to regard them as of connective tissue origin rather
than contractil musculo-epithelium. The anastomosing proc-
esses of the cells stain deeply red in Van Gieson’s. Further
work, especially of an embryological nature, will be essential
to determine the positive nature of these cells.
Other connective tissue cells are seen in addition to these
large cells—Korbe. These manifest themselves as elongated
nuclei and lie alongside of the basement membrane. Occasion-
ally nuclei appear at the intersection of the fibrils of the reticulum
and as a consequence appear as a part of it. Nuclei of endothel-
ial cells, lymph cells, and plasma cells are also seen.
160 JOHN SUNDWALL
The basement membrane proper has been described by Zimmer-
mann and Fleischer in the lachrymal gland.
According to my observations, the final ramification of the
supporting tisstie is in the form of a reticulum enclosing within
its meshes the acini, intercalary and intralobular ducts, and
vessels. This reticulum forms the membranae propriae of the
acini or tubules and consists of a delicate network of interlacing
fibrils which intimately surround the epithelial cells of the acini
and small ducts. From this basement membrane secondary
Fig. 4a Reticulum, Tech. III, 2, 3, Oc. 6, obj. 8, X 300. Parenchymatous
and other tissue digested away in pancreatin. Reticulum is readily seen showing
the outlines of the original tubules.
Fig. 4b Reticulum, Oc. 2, obj. 2mm. oil, X 770. Portionin figure 4a enclosed
within the circle is shown here enlarged.
fibres apparently take origin and pass radially into the acini
between the individual gland cells. Others may appear to enter
the base of the cell, thus simulating Holmgren’s Trophospongium.
However, in reality this probably does not occur.
The basement membrane and its relation to the tubules, as
described, can readily be seen in sections stained by Mallory’s
method, whereby it is stained dark blue. As a rule it is closely
apposed to the epithelium, and the interepithelial processes are
seen with ease. In Van Gieson’s stain it is red. This network
THE LACHRYMAL GLAND 161
can be considered as reticular connective tissue in the sense of
Mall, and is almost identical in appearance to the reticular
framework of the submaxillary gland as shown by Flint. Accord-
ing to the Flint, Spalteholz method (Technique II, 1) and a
modification of Mall’s method (Technique III, 2) it was shown
that this basement membrane is not digested in pancreatin while
the parenchyma entirely disappears. Satisfactory specimens
for study may be made by these methods and then staining by
either Mall’s or Mallory’s method (figs. 4a and4b). Van
Gieson’s stain also may be used. The presence of this basement
membrane after prolonged digestion shows that elastic fibres
do not compose it.
Plasma cells
Numerous plasma cells are seen in both the interlobular and
intralobular tissue, principally around the ducts and between
the acini. These cells are typical in appearance—round, more
or less eccentric nuclei with peripheral arrangement of the chro-
matin, and fine basophilic granules in the protoplasm. ‘These
cells are differently shaped—some spheroidal, others elongated ,
and irregular in outline. Sections stained in neutral safranin
show these cells very clearly. They may be found singly or in
clusters (fig. 5).
Since Waldeyer (’75) first designated certain cells as plasma
cells and Unna (’92) described these cells as purely pathological
in the skin, followed by other observers who stated that they are
present in normal tissues, there have been numerous contributions
regarding plasma cells and their distribution in glands. Klein
(79, ’82) and Cajal (96) describe them as being present in the
submaxillary gland of man. Zimmermann shows in the lachry-
mal gland of man ‘‘Granulirte Zellen (Plasmazellen?) mit je
zwel stapchenf6rmigen Centralkérpern innerhalb eines granula
freien Hofes.’”’ Plasma cells have been described also by Ioan-
novices (’99) in glands of the tongue, and by Maximow (’01)
in submaxillary and retrolingual of the dog. According to the
latter they are comparatively few in the submaxillary and
numerous in the retrolingual. Krause (’98) calls attention to
162 JOHN SUNDWALL
the relation of plasma cells to the secretory activity of the retro-
lingual gland of the hedge hog—they being fewer in number in
the stimulated gland than in the resting. Dantchakoff (’05)
observed the relation of plasma cells to the secretory activity
of the submaxillary gland of the rabbit, and Hannes (711) de-
scribes them in the lachrymal gland of man.
Katharine Hill.
Fig. 5 Interstitial plasma cells. Oc. 4, obj. 2mm. oil.
IV. DUCT SYSTEM AND TUBULES
For purposes of description I have adopted the following
classification of the duct system; (1) main ducts, Ductuli excre-
tori; (2) primary ducts; (3) interlobular ducts; (4) intralobular
ducts; (5) intercalary ducts; and (6) acini, alveoli, or tubules.
I see no reason for the inclusion of separate sublobular ducts in
the lachrymal gland.
THE LACHRYMAL GLAND 163
The methods used for the study of the duct system have been
as follows: (1) injection methods, (2) vital staining with pyronin,
and (38) histological study of sections.
The injection-corrosion specimens of the duct system as prepared in
this laboratory (Technique IT, 2
The main duct follows a more or less tortuous course
through the palpebral fascia and enters into the gland substance
at the outer margin. At this point it is seen to branch. The
branching is somewhat variable—generally two different types
are seen: (a) a dichotomous branching wherein the main duct
divides into two ducts, which by the same method immediately
form four ducts, ete. (This method of branching is described
for the submaxillary glands by Flint) and (b) the main duct
may continue for some distance into the gland substance
somewhat similar to the trunk of a poplar tree, while the primary
ducts take origin from this main trunk at various levels. In both
instances, however, before the primary ducts take origin, sev-
eral small branches are seen to leave the main duct at right
angles. These are the ducts of small accessory lobules located
on the outer or lateral margin of the gland.
The primary ducts vary in length, depending upon their
distribution. They divide, as a rule, dichotomously into either
equal or unequal branches. In the former case, where the
branches are equal, these immediately undergo dichotomous
branching again to form the interlobular ducts. Where the
branches are unequal, the smaller one may directly form an
interlobular duct, while the larger one may continue as a pri-
mary duct for some distance farther and then branch to form
interlobular ducts. Frequently very small branches take origin
directly from the primary ducts at right angles as in the case of
the main ducts. . These small branches are intralobular ducts
which empty directly into the primary ducts from neighboring
lobules.
An extensive ramification of the interlobular ducts is seen.
The branching is much similar to that described for the primary
THE AMERICAN JOURNAL OF ANATOMY, VOL, 20, NO. 2
164 JOHN SUNDWALL
duets—(a) dichotomously or (b) the main interlobular ducts
continuing as such for some distance with smaller intralobular
ducts taking origin at various levels.
The intralobular ducts have an extensive ramification as well.
The branching is chiefly of the dichotomous type although
trichotomous branching is seen. The branches are as a rule
very unequal in calibre. Each lobule as a rule contains many
of these ducts of unequal sizes. At various levels in the course
of these ducts, nodular enlargements are seen. Frequently
the ducts are seen to terminate in nodular enlargements. It is
somewhat difficult to force the injection mass beyond these
nodules. However, after careful and repeated trials it is possible
to foree it (celloidin is preferable) to the alveoli or tubules and
i PM ;
Fig. 6a Corrosion cast of main duct with branches. Drawing, binocular,
somewhat diagrammatic. MM, main duct; P, primary duct; J, Interlobular duct;
I’, intralobular duct; /’’, intercalary duct; 7, tubules.
when this is accomplished it is seen that these nodules mark the
exits of the interealary or junctional duct.
The lumina of the intercalary ducts are exceedingly fine and
threadlike as represented by the celloidin or celluloid cast. As
a rule two or three of these ducts leave each nodule at right angles
to the intralobular duets and terminate in the tubules. Fre-
quently they are seen to undergo dichotomous and in some
instances trichotomous branching before so terminating. From
two to three tubules mark the termination of the intercalary
duct.
The entire duct system belonging to one main duct forms an
elongated gland (fig.6a). Generally the primary duct leaves
the main duct at an acute angle. The branching of the secondary
THE LACHRYMAL GLAND 165
duct system—interlobular and intralobular duets—is similar.
The interealary ducts, on the other hand, as a rule branch off at
almost right angles or even greater angles from the intralobular
duets from which they have origin (fig. 6b). Of course there
are numerous exceptions to this generalization, but it is not
by any means imaginative to compare these corrosion casts of
the lachrymal gland to six or eight trees varying in length and
placed in a line so closely that the smaller branches and leaves
intermingle. The main trunks are then comparable to the
Ductuli excretorii; and the primary, secondary, and tertiary, ete.
branches to the primary, interlobular, and intralobular ducts.
The leaves represent the acini or tubules and their stems the
intercalary ducts.
Fig. 6b Corrosion cast. High power drawing, projection apparatus. A,
intralobular duct; B, intercalary duct; C, tubule.
In preparations where the mass has not passed beyond the
nodular enlargements of the intralobular ducts the branching
is readily observed through the binocular microscope. If the
mass has passed into the tubules the duct system is greatly
obscured. For the study of the intercalary ducts and tubules,
small pieces were teased off and mounted on slides, then studied
by high power.
It must be appreciated that in the study of corrosion casts, it
is impossible to classify with certainty the various ducts; 1.e., one
cannot differentiate between the larger intralobular and the
smaller interlobular ducts. As an auxiliary to this study carmin
gelatin was injected into the ducts of other glands and sections
of these tissues were prepared. In the latter preparations the
166 JOHN SUNDWALL
diameters of various ducts were ascertained and then compared
with the former. Even with this aid it was impossible (and not
essential) to classify all ducts.
Vital staining
Vital staining with pyronin (‘Technique VI, 1) greatly facili-
tates the study of the smaller ducts and their distribution. Bens-
ley found that the entire duct svstem of the pancreas was stained
Fig. 6¢ Lobule of gland, fresh, after vital staining with pyronin. Drawing,
binocular. The lumina of ducts are deeply stained red. A, intralobular duct;
B, interlobular duct.
by this method. In application of his methods to the lachrymal
gland it was found that its duct system stains a deep red while
the tubules are only faintly stained (fig.6¢). In the larger ducts
the stain is limited to the periphery of the lumen, while in the
smaller ducts the entire lumen stains. The smaller intralobular
and interealary ducts are especially prominent, and when fresh
sections 0.5 mm. thick are examined through the binocular
microscope the ramifications are clearly seen. The deep red
ducts can be readily traced, in thin sections, to the acini. Very
THE LACHRYMAL GLAND 167
thin sections of this tissue previously fixed in ammonium molyb-
date (Technique VI, 1) show that all the intercellular secretion
capillaries are stained. See Secretion capillaries.
Histological structure of ducts and tubules
Main ducts. The excretory ducts in the superior lid are so
tortuous in their course that it is impossible to obtain a continu-
ous duct in longitudinal sections—only portions of one duct can
be obtained. The ducts lie close to the conjunctival surface.
As the terminal of the duct is approached there is seen a gradual
increase of lymphoid tissue which surrounds the terminal open-
ings as true lymph nodes.
-
Fig. 7 Epithelium of main duct. Drawing, Zeiss, oc. 4, obj. 8, as seen in
cross section of duct near its terminal in lid. A, goblet cells; B, epithelial cells.
The large clear goblet cells are readily seen surrounded by smaller epithelial
cells.
A cross section of the duct near its terminal shows the follow-
ing characteristics: The lumen is either slit-like or very much
corrugated, the epithelium being thrown up in folds with crypts
between. Two types of cells line the duet—goblet cells and more
or less irregularly elongated or cuboidal epithelial cells arranged
in layers (fig. 7). The former cells occur in great numbers.
They average as a rule 40u by 20y although there is much vari-
ation. Sections stained in iron haematoxylin and counterstained
with mucicarmin show the surface epithelium to the best ad-
vantage. Here the goblet cells are stained red and can readily
be differentiated from the epithelial cells in general. Many of
168 JOHN SUNDWALL
these goblet cells reach the lumen and from the ends secretion
masses can be seen projecting into the lumen. Others again
do not reach the surface but are entirely surrounded by the
general epithelium. As a rule a flattened nucleus is seen in the
base of these cells and a definite network is seen in the cytoplasm.
(fig. 8).
The epithelial cells show great irregularity as to form and size.
They occur in several layers. In the crypts usually two layers
are seen while in the folds three or even six layers of cells are seen
arranged irregularly. Those bordering on the surface (super-
ficial cells) are usually elongated and the base is drawn out into a
wedge shaped process which projects downward between the
¢ ARC ce
Se ig
yy
Ry)
Fig. 8 Epithelium of main duct. Zeiss, oc. 4, obj. 2mm. oil. A, epithelium;
B, goblet cells.
deeper layers of cells. In the erypts these wedge-like pro-
jections may reach the basement membrane. The deeper layers
of cells are smaller as a rule than the superficial cells and are
irregularly cuboidal or polygonal in shape. The deepest layer
of cells borders on a basement membrane which is readily seen
in sections stained by Mallory’s method. The epithelial cells
in general stain deeply and homogeneously in the usual stains.
Reference is made to mitochondria later on. No secretory
granules can be made out in any of these cells. The nuclei are
prominent, oval or spheroidal, and show irregular clumps of
chromatin. As a rule no definite nucleolus is seen.
The basement membrane is directly surrounded by a circular
coat of connective tissue which averages from 100 to 150u in
THE LACHRYMAL GLAND 169
thickness. The inner portion is very cellular containing numer-
ous lymphoid cells, while the outer boundary ismore or less fibrous.
This outer boundary is immediately surrounded by the general
connective tissue of the palpebral fascia. Sections stained by
Van Gieson’s method show the presence of numerous small ves-
sels accompanying the duct. Smooth muscle cells are also seen
in this connective tissue coat. They occur as a rule singly and
do not constitute a coat. Collagenous fibres greatly predominate.
Weigert’s stain shows the presence of numerous interlacing
elastic tissue fibres which form a more or less indefinite layer
outside of the basement membrane.
Primary ducts. These ducts represent the first branches of the
main duct and are very irregular in size. As a rule the lumina
are round or oval and have an average diameter of 0.2 mm.
The primary ducts are embedded and surrounded by a large
amount of connective tissue which is derived from the septa.
The epithelial wall averages 20u in thickness and is as a rule
made up of several layers of cells—varying from two to three
layers, the former predominating. In many instances, however,
a single layer of long columnar cells forms the entire epithelial
wall. Where two layers are present there is an interdigitation
seen between the long columnar cells and the basal more or less
irregularly cuboidal cells.
The cytoplasm of the surface layer in the usual fixatives stains
homogeneously. The nuclei are either oval or round and have
an average diameter of from 7 to 10u. Clumps of chromatin
irregularly distributed are readily made out within the nuclei.
In many stains a definite cell boundary cannot be made out.
However, it is plainly seen in thin sections and especially after
the use of iron haematoxylin.
The cells of the outer layer, or layers, are irregularly cuboidal
in form. The nuclei and cytoplasm show the same staining
characteristics as the inner layer. The nuclei of the outer layer
are frequently so arranged that their long axes are parallel to
the circumference of the duct, while those of the inner layer are
radially arranged to the duct. This condition, however, is
much more pronounced in the smaller ducts.
170 JOHN SUNDWALL
Goblet cells are also seen in the primary ducts. They are
not so numerous by any means as in the main ducts. Many -
transverse sections of the former may show none of these cells
while others again may show from one to three. Their staining
characteristics are similar to those deseribed for these cells in
the main ducts (figs. 9 and 10).
The epithelial cells of primary ducts border on a more or less
indefinite basal membrane, which stains a deep blue in Mallory’s
stain, and is derived from and is a condensation of the circular
layer of collagenie fibres which surround the epithelium and
form the duct wall. The average thickness of this connective
Fig. 9 Epithelium of primary duct, cross section. Zeiss, oc. 4, obj. 8 mm. A,
goblet cell.
Fig. 10 Epithelium of primary duct. Enlarged drawing, Zeiss, oc. 4, obj.
2mm. oil. A, cement lines.
tissue wall is 50 to 100u. Its outer borders are gradually lost
in the general connective tissue septa of the gland, which are
always more abundant where these ducts are found. Long fine
interlacing elastic fibres are present throughout the entire thick-
ness of this circular connective tissue coat. Immediately sur-
rounding the basement membrane these fibres form a closely
interlacing layer similar to that described for the large main
ducts. As the duets become smaller, through ramification,
there is a gradual decrease of these fibres and in some of the
smaller primary ducts they are difficult to make out.
The chief component of the circular connective tissue wall is
the collagenic or white fibrous tissue. The collagenic fibres
THE LACHRYMAL GLAND wal
stain deeply red in Van Gieson’s. Within the interstices of
these fibres many nuclear elements are seen which are (a) the
nuclei of the endothelium, a rich plexus of arterioles, venules,
and capillaries which are found in the walls; (b) lymphoid cells,
which in very rare instances are so numerous that they obscure
the collagenic fibres (I do not consider these lymph cell ac-
cumulations normal as they occur so infrequently) ; (¢) oecasion-
ally solitary nuclei are seen in a cytoplasm which stains yellow
in Van Gieson’s. The structure of these cells, together with the
staining characteristics, suggests that they are smooth muscle
cells; (d) plasma cells are also seen.
In addition to the very small and numerous vessels which
accompany and supply the ducts, large vessels and nerves are
seen in close proximity to the ducts and following their ramifi-
cations to the gland substance. There is no regular distribution
of these vessels, so far as numbers are concerned, in relation to
the primary ducts. Sometimes only one of these larger arteries
is seen in close proximity to a groupof primary ducts. Frequently
two are seen. Generally one large vein is seen to accompany
these ducts. This is true also of the non-medullated nerves.
However, much variation is the case with these as well.
Interlobular ducts. The primary ducts ramify to form the
interlobular ducts. These further subdivide to form the intra-
lobular ducts. It must be borne in mind, however, that intra-
lobular ducts may originate directly from the main or primary
duct. Again practically all of the primary ducts and even the
main duets for some distance are interlobular. Consequently
the artificial subdivision of the duct system into primary, inter-
lobular, and intralobular ducts must be accepted in a general
sense only. In the following description of the interlobular
ducts reference is made to those duets which as a rule result
from the ramification of the primary ducts and lie between the
lobules.
The lumina of these interlobular ducts are wide and average
as a rule about one-half to one-fourth the diameter of those of
the primary ducts. Here again an explanation is necessary, for
the diameters of the former are much larger at their origin than
1 JOHN SUNDWALL
later on after ramification. The epithelial wall, which averages
from 12 to 18u in thickness, is composed of from one to two
layers of cells similar in arrangement to those described for the
primary ducts. The larger interlobular ducts generally possess
two complete layers while in the smaller ones there is a notice-
able reduction of the cells which form the outer layer. Conse-
quently in these smaller ducts many of the epithelial cells
extend down to the basement membrane.
The single layer cells are cylindrical in form. The cytoplasm
stains deeply and evenly in the various cytoplasmic stains.
Study of these cells by means of the oil immersion shows the
cytoplasm to be very finely granular. No basal filaments or
secretory granules can be seen in the general stains. I did not
see In my preparations the central bodies described by Zimmer-
mann. ‘Tissues fixed in acetic osmic bichromate and stained
in acid fuchsin-methyl green show numerous mitochondria.
These are described under ‘Mitochondria.’ The cell boundaries
are indefinite and only in the thinnest sections can they be made
out clearly. As in the other ducts where a second layer of cells
is present the inner layer sends basal processes between them
which reach the basal membrane. The nuclei, which average
from 6 to 8u, are round or oval and show the same staining char-
acteristics as those of the primary ducts.
The cells in the outer layer, which, as already stated, are fewer
in number, are more or less elongated—this elongation being
parallel to the circumference of the duct. Here also the cell
boundaries are indefinite. The cytoplasm stains similarly to
that of the cells of the inner layer. In very thin sections both
stain with the same degree of intensity. In thick sections, how-
ever, the inner layer appears more densely stained. This obser-
vation was made by Fleischer, who states, as a consequence,
that the cytoplasm of the cells of the inner layer stains more
intensely than does that of the outer layer. In my opinion this
apparently deeper stain is due to the fact that these inner cells
are more numerous and more compact, while those of the outer
layer are fewer in number and more loosely surrounded by other
cells. The nuclei of these cells are either oval or elongated, the
THE LACHRYMAL GLAND eae
long axes being parallel to the cirumference of the tubules.
Here is seen a more pronounced example of the statement that
the long axes of the nuclei of the outer layer are at right angles
to the long axes of the nuclei of the inner layer cells. The
cells of the outer layer are frequently so elongated that they
resemble connective tissue cells. The nuclei of both layers are
similar in staining characteristics.
The connective tissue wall surrounding the epithelium varies
in amount depending upon the thickness of the septa in which
the ducts are situated. A well defined basal membrane upon
which the epithelial cells rest is plainly seen with Mallory’s
stain. Practically all the elements described for this wall
under the caption of the primary ducts are seen in these walls—
with a reduction in quantity, however. There is a marked
decrease of elastic fibres. It is only necessary to contrast the
smaller interlobular ducts with the larger primary ducts to
appreciate this statement. In the wall of the primary ducts
numerous elastic fibres are seen throughout the connective tissue
wall while in the smaller interlobular ducts the elastic fibres
occur sparsely or may not be seen. The accompanying arteries
and veins, however, show these deep staining fibres in their
walls and may be used for comparison. The decrease and dis-
appearance of the elastic fibres take place between the origin
of the primary ducts and the smaller interlobular ducts. The
quantity of elastic fibres present is proportional to the thickness
of the walls and septa in which the ducts are embedded. Where
the connective tissue wall is much reduced and the septum is
thin elastic fibres are only sparsely present or are not seen at all.
The walls are composed almost entirely of collagenic fibres as
revealed by Van Gieson’s stain. Numerous cellular elements
are present which may be classed similar to those of the primary
ducts.
Intralobular ducts. Intralobular ducts vary much in the size
of the lumina and the quantity of the surrounding connective
tissue. Some ducts appear greatly dilated with lumina 70.
in diameter. These dilated ducts are not frequently found and
are doubtless abnormal. Many again are seen which correspond
174 JOHN SUNDWALL
in size and structure to the interlobular duets. The intralobular
duets present as a rule the following characteristics. There is
a reduction in the cells of the outer layer so that the smaller
ducts appear to possess but a single layer of cells which show the
same characteristics as those described for the interlobular
ducts. The nuclei are elliptical. Frequently, however, one
sees ducts with two layers of cells wherein the cells of the outer
layer are as prominent as are the cells of the inner layer. This
is true even in some of the smallest of the intralobular ducts.
The connective tissue surrounding these ducts is much less
abundant than that of the interlobular structures sometimes
forming only a thin layer or again it may be considerable in quan-
tity. It is derived from the intralobular connective tissue.
Malloryv’s connective tissue stain shows a definite basal mem-
brane. The wall is made up almost entirely of collagenic fibres.
Only occasionally are elastic fibres seen. Capillaries are seen
in these walls.
No secretory granules or secretion capillaries are present in
these ducts. I did not observe by using the ordinary stains the
basal filaments described by Hornikel in the cells of these ducts
in the lachrymal gland cof the ass and the more or less indefinite
pencil like structures deseribed by Fleischer in the ox. How-
ever these are seen in special preparations (see Basal striations).
Merkel saw none in the lachrymal gland of the dog.
Cement lines. These are seen in connection with the surface
epithelium of all the collecting ducts—primary, interlobular,
intralobular. On the surface they outline the polygonal margins
of the cells and at the various angles appear as dots. In longi-
tudinal sections this cement substance forms a well defined
point which projects some distance basalward between the cells
sometimes half the length of the cell. As they near the base
these intercellular cement structures become finer and are finally
lost in the cell membrane (fig. 20). Fleischer failed to observe
this characteristic of the cement lines in the lachrymal gland
of the ox. Kolossow saw no cement lines in his studies on glands.
Zimmermann, on the other hand, noted this peculiar arrange-
ment of the cement lines in the human lachrymal gland. He
THE LACHRYMAL GLAND io
describes these cement structures as bands projecting down
between the cell boundaries completely surrounding the proxi-
mal end of the cell, and states that this structure of the cement
is peculiar to the lachrymal gland. I am inelined to think that
the basal projecting intercellular cement lines are more or less
limited to the angles of the hexagonal cell margins and that they
do not form bands or caps surrounding each cell. They are
always seen as intercellular continuations of the fine black dots
seen at these angles. Cross sections of these ducts show their
complete absence between many cells. This would not be the
case if they formed true bands. The entire proximal half of the
cells would appear in stained sections much darker than the
basal half, which is not the case.
Intercalary ducts. These have been described by many
histologists among whom may be named Schwalbe, Béhn and
Davidoff, Merkel, and Henle. This duct has been compared
to the Speichelréhren (Schaltstiicke) of the salivary glands,
especially the parotid. V. Ebner describes these ducts in the
submaxillary gland as short blue tubes. Their deep staining
characteristics have also been described by Merkel in the lachry-
mal gland of the dog; and by Nussbaum and Langley in the
submaxillary of the rabbit. The latter also refers to the presence
in them of ‘large copious granules.’ Zimmermann compares this
duct in the lachrymal gland in man with the parotid gland.
However, he saw no basal striations in the ducts. According
to him the intercalary duct in man is not so well defined as that
in the ox but the transition from the intralobular duct to the
tubules is much more gradual. Merkel, in his observations on
the lachrymal gland of the dog, also describes intercalary ducts
but “‘fand hier keine Gange mit Stabchenepithlien.”” Ac-
cording to him these ducts stain more deeply than the other
structures. Regarding the intercalary duct or schaltstiicke,
Fleischer states—‘‘Die Form der Zellen, ihre einschichtigkeit,
das enge Lumen, ihr haufiger pl6tzlicher Uebergang in die
Ausfithrungsgiinge entspricht ganz dem, was zuerst V. Ebner
bei Speicheldriisen als Schaltstiicke bezeichnet hat. Sehr char-
akteristich. . . . ist auch ihre diffuse intensive Fiarbung.”
176 JOHN SUNDWALL
Secretion granules in the cellsof the intercalary duct are described
by him. The chief characteristics then of the intercalary duct
according to these observers are the more intense staining re-
action and the presence of granules.
The deeper staining of these ducts is readily seen in sections
of tissue fixed in the ordinary solutions and stained with iron
haematoxylin and neutral stains.
For the study of the intercalary duct I found that the most
interesting results were obtained from tissues fixed in Zenker’s,
embedded in celloidin, and stained in muchaematein. This
stain was prepared and used according to Bensley’s method
Fig. 11 Interealary duct and tubules. Stipple board drawing projection.
The granules of the intercalary duct are stained specifically with muchaematein,
Tech. IV, 1. A, unstained intralobular duct; B, intercalary duct with granules;
C, tubules; D, transitional cells.
(Technique IV, 1.) The sections were cut from 10—-20u thick.
It was found that by following this technique the granules of
the intercalary duct are stained specifically and definitely blue
while the remaining tissue is unstained (fig. 11). This
method therefore is of great value in the study of the intercalary
duct and its relations to the intralobular duct on the one hand
and the alveoli or tubules on the other. In the preparation of
these sections it was found that after fixation in Zenker’s the
tissue should be thoroughly washed in order to remove as much
of the salts as possible, for if iodine is used for this purpose
the distinct granular stain cannot be obtained. Furthermore,
THE LACHRYMAL GLAND LEC
if in thick sections the stain affects the tubules this can be re-
moved by the use of acid alcohol. ‘The time required to obtain
the intense blue stain is brief, as less than one minute was neces-
sary. Sectionsof the salivary gland andof the mucous membrane
of the intestinal tract were used for controls. I found that the
muchaematein stained the granules of the intercalary ducts of
the lachrymal gland in the same time that was required to obtain
a corresponding stain in the control sections. The similarity
of this gland in its staining characteristics to the mucous glands
and cells is interesting as it suggests the possibility of the presence
of mucous secreting cells in the lachrymal gland. [| shall discuss
this phase later on. These muchaematein stained sections show
the interlobular and the intralobular ducts unstained. The
tubules generally take a light diffuse stain, and connecting the
tubules with the intralobular ducts are the deep blue granular
stained interecalary ducts—the stain is confined to the granules
only.
The intercalary ducts vary greatly in length. The longest
are from 200 to 250u in length and the shortest are about 70
to SOu. The lumen of the interealary duct is very narrow when
contrasted with that of the other ducts. It averages about 6u
in diameter although much variation exists not only inits width
in different ducts but also in a single duct. As a rule the lumen
becomes wider as it approaches the intralobular duct. Primary
intercalary ducts frequently give rise to a number of secondary
branches which in turn terminate in one or more tubules. The
cells of these ducts have an average height of 10u. The nuclei
are not stained in muchaematein, they are seen, however, as
large oval structures in the basal end of the cell. They never
appear angular or compressed against the base. The granules
are confined to that portion of the cell forming the lumen (the
proximal end) and oecupy about one half of the cell mass (fg 12).
This arrangement of granules is rather constant, rarely is a cell
seen where the granules extend to the base. They are proximal
to the nuclei and in many eases obscure their rounded proximal
surface.
178 JOHN SUNDWALL
In cross sections of the intercalary duct the cells are pyramidal
in form, the apices being in contact with the lumina which are
seen to be very narrow when contrasted with the wide lumina of
the intralobular and interlobular ducts. The lumina are formed
as a rule by an average of seven triangular shaped cells. (These
same cells in longitudinal sections appear cuboidal in form.)
The granules are grouped in the apices of the cells and consequently
these masses of granules are triangular on outline. A narrow
es - eo a, S Sine eee X Hilt 07.
Mig. 12 Secretion granules in intercalary duct, Zenker’s muchaematein.
Zciss, oc. 4, obj. 8. Granules are specifically stained. A, intralobula duct; B,
intercalary duct with granules; C, cross section of intercalary duct; D, tubules.
margin of unstained non-granular cell substance is seen between
these granular masses and the cell membrane.
In longitudinal sections these cells appear cuboidal in form and
rest on a well defined basal membrane which represents practi-
cally all of the supporting tissue.
Not all of the cells of the intercalary duct are stained by the
muchaematein. Many cells are present which possess no granules
and consequently are not stained. Occasionally ducts are seen
in which in longitudinal sections the majority of cells of one side
show no granular stain while the cells of the other side possess the
THE LACHRYMAL GLAND 179
characteristic granular stain. The presence of nongranular
cells in the intercalary duct is plainly seen in cross sections of
those ducts, the number of unstained cells varying in different
sections. In some ducts only one cell may be present which
does not possess the muchaematein stained granules, again such
ducts are seen where the majority of the cells in cross sections
are not stained. As arule, however, most of these cells possess
specifically stained granules.
At the origin of the intercalary ducts from the intralobular
ducts the granules as a rule make their appearance in the first
cells as a narrow layer in the lumen surface of the cell, having
the appearance of a theca or cuticula. As the distance from
the intralobular duct increases this stratum of granules widens
until it occupies the entire proximal half of the cell. The com-
plete transition from the normal amount to the complete dis-
appearance of granules takes place in from four to seven cells.
Frequently one sees an abrupt transition at the junction of the
intercalary and intralobular ducts. In this type the granules
in the cells of the intercalary duct remain normal and constant
in quantity until the intralobular duct is reached when they
disappear completely. Thus two neighboring cells may be seen
one belonging to the intercalary duct and having the characteris-
tic granules occupying the proximal half of the cell while the
bordering cell belonging to the intralobular duct possesses no
granules whatsoever. The granules in the intercalary duct are
seen to extend to the tubules. They are readily seen in sections
prepared by Weigert’s method (Zenker’s fixation) for the dem-
onstration of elastic fibres. In the alcohol osmic bichromate,
and methyl green anilin fuchsin preparation, the granules are
well preserved and stain green.
Tubules. The cells of terminal tubules vary in size and form
depending upon the stage of secretion they are in. Thus they
may be columnar, cuboidal, oval, polyhedral, pyramidal, spheri-
eal, or flattened. When the cells composing these elements are
in the height of the granular stage the cells are large and bulging
and the secreting end may then appear as an alveolus or acinus.
Again when the cells are observed in a partial granular stage
THE AMERICAN JOURNAL OF ANATOMY VOL. 20, No. 2.
180 JOHN SUNDWALL
they are smaller and more cuboidal and the secreting end has a
tubular form. This doubtless explains in a large measure the
controversy regarding the structure of the secreting terminals—
whether tubules or acini. The finer structure of the cells of
tubules is considered in the Section on granules.
V. SECRETION GRANULES
Serous cells!
For some time there was much discussion as to whether the
granules seen in fixed preparations of glands were natural or the
product of fixation. <A. Fischer (’99) regarded them as artefacts—
Fallungsprodukte. E. Miller, (96) on the other hand, found
that certain granules seen in fresh cells became more distinct
when fixed with sublimate solution. The process of fixation
was directly observed through the microscope. Held (’99)
and many others have substantiated Miiller’s observations, and
now secretion granules are no longer regarded as artefacts.
Milawsky and Smirnow: (’93) studied, in the parotid and sub-
maxillary glands of dogs, the secretory changes resulting from
electrical stimulation of the cerebral and sympathetic nerves
supplying these glands.
Bensley (’96) among others (see his various papers for bibliog-
raphy) has contributed much to our knowlege of the anteced-
ent substance which goes to make up the granules of serous
cells. He has observed that
During digestion a substance similar in chemical properties to the
chromatin of the nucleus makes its appearance in the outer clear zone
of the chief cells of the fundus glands. This substance, which may be
called prozymogen, stains deeply and readily in haematoxylin, and
presents a characteristic fibrillated appearance. (He suggests that in
some cells this is dissolved in the nuclear substance and that sometimes
it is collected in masses—plasmosomata.) During rest this proxymogen
is used in some way giving rise to zymogen granules.
Speaking of the chief cells he states that the fibrillae in the
outer zones of the cells are not so regular or distinct as in the
? Metzner (’07) in Nagel’s Handbuch der Physiologie des Menschen gives an
excellent discussion of secretion granules in general.
THE LACHRYMAL GLAND 181
salivary ducts and that they remind one of the nebenkerne in the
pancreatic cells of the amphibians as desertbed by Macallum,
Erberth, and Miiller. In a later contribution (08) he suggests
the probability of serous cells being made up of various groups.
By using the following technique (1) examination of fresh mate-
rial in blood serum, teased preparations or sections cut with
Valentine knife, (2) experimental fixation, by which a fixative
medium was obtained which would preserve the secretion ante-
cedents in a form which is present in living cell, (3) differential
microchemical staining, Bensley has obtained results “which
seem to justify the subdivision of the serous class of cells into a
number of subordinate groups” such as (1) ‘tropochrome cells’
which stain metachromatically under certain fixations and stains,
and (2) ‘homochrome cells’ which do not stain metachromatically
and in all liklihood form a heterogeneous group. Ina still later
contribution (’11) he speaks of prozymogen granules ‘‘zymogen
granules in process of formation. . . . These must not be
confused with the basophile substance of the cell which has
elsewhere been called ‘prozymogen’ by Macallum and myself.”
These prozymogen granules stain deeply in neutral red when used
as a vital stain.
Mucous cells
Regarding the main differences between mucous and serous
cells as revealed by more perfect fixations and staining processes,
Bensley’s (’03) conclusions may well be inserted here. The
following reasons are given for his conclusion that the cells of
the glands of Brunner, in eighteen out of nineteen genera exam-
ined, are of the pure mucous type—(1) granules (droplets?)
have a low refractile index which corresponds closely to the
mounting media; (2) no basal filaments, ‘prozymogen,’ in fixed
preparations (the microchemical test for organic iron shows
only relatively small amount of cytoplasmic nucleoproteid) ;
(3) the granules do not stain in iron haematoxylin or neutral
gentian; (4) mucous cell stain in mucicarmin and muchaematein ;
(5) granules are soluble in weak alkaline solutions, insoluble
in 5 per cent solution of hydrochloric acid and in artificial gastric
182 JOHN SUNDWALL
juice containing 0.2 per cent of hydrochloric acid; (6) structure
of cells. However ‘‘It is obvious that no absolute proof of
the mucous character of the glands of Brunner can be brought
forward until a positive microchemical test for the various mucins
is devised.”” He suggests that mucous glands may contain small
amounts of ferment. See also Bensley’s (02) summary of
results in the histology of cardiac glands.
Similarities of mucous and serous cells
Regarding serous and mucous cells investigators agree that in
both types the process of secretion 1s very similar and that the
secretory substance is present in the form of very small globules
generally termed granules. These granules are seen in fresh
tissues with a few exceptions. They vary in size and in degree
of light refractibility. The small granules have a greater re-
fractive index than do the larger ones. In the case of the latter
this index may be very nearly the same as that of the mounting
medium so that they appear very dim, ‘matte,’ or can not be
seen at all—this is true also of the granules of certain cells through-
out their secretory activity. In the resting cell before secretion
has begun the cell is full of granules, and the nucleus which may
appear flattened or angular lies at the base of the cell. After
stimulation the granules are greatly decreased or disappear, the
nucleus becomes round or oval and approaches the center of the
cell. In most instances these granules can be fixed and stained.
Usually the smaller ones stain more intensely than do the larger.
In many glands the larger granules disappear in the process of
fixation and an intergranular protoplasm remains.
Granules in lachrymal gland
According to Langley (lachrymal gland of the rabbit)—
In the resting gland the alveoli are throughout unevenly stained.
The nucleus is irregular and lies in the peripheral portion of the cell.
During activity the outer portion of each alveolus begins to stain
evenly at first without much alteration in the nuclei or in the inner
portions of the cells, later the nuclei become larger and travel towards
THE LACHRYMAL GLAND 183
the center of the cells, they are then much less distinct, at the same
time the lumen becomes more obvious.
The observations of Reichel, Kolossow, and Lor were also con-
fined to the secretory changes. Puglisi-Allegra (’04) contrib-
utes an extensive article.
Studio della glandola lagrimale. . . . In una glandola normal-
mente funzionante si possona distinguere due specie de cellule. :
Dopo una eccitazione prolungata é facile notare numerose cellule
pluttosta piccole con protoplasma oscura é fortemente granuloso,
totalmente prive di secrezione.
Nicolas saw numerous granules in the lachrymal gland of
man after fixation in sublimate or Flemming’s solution and stain-
ing with aniline-safranin, Biondi-Ehrlich and Altmann’s methods.
Noll saw in the nonstimulated fresh lachrymal gland of the
cat “‘die meisten Zellen deutlich granulirt. Die Cellgrenzen
treten nicht immer gut hervor.’’ The nucleus—‘‘der. Basis
nahe gelegen, als rund oder ovalen.’’ According to him these
granules had remarkable differences of refractibility; they dis-
appeared when water was drawn under the cover-glass and re-
appeared when 2 per cent NaCl was added; other cells were seen
with no apparent granules except numerous ‘Protoplasmak6rn-
chen (these cells he termed ‘matte Zellen’); the tissue when
fixed and stained by Altmann’s method showed faintly and
deeply stained cells—‘‘hellere und dunklere Zellen.’”? Transi-
tional cells were also seen which showed both the characteristics
of faintly stained and deeply stained cells. Noll fixed fresh
tissue with Altmann’s fluid and observed the process by means of
the microscope. The strongly refracting granules in fresh tissue
became the deeply stained granules in prepared sections while
the less refracting (‘matteren’) granules became the faintly stain-
ing cells with the network. In the latter case the granules were
not conserved but only meshes remained. The deeply and
faintly staining cells then represent different stages in the secre-
tory activities of the gland, the latter representing the higher
stage of granule formation. After electrical stimulation fresh
tissue showed ‘‘das die Alveolen bei Weiten nicht so viel Granula-
Zellen enthalten. In der Hauptsache sieht man Zellen mit
184 JOHN SUNDWALL
matter Grundsubstanz.’’ The nuclei were round, granules were
present in the neighborhood of the nuclei.
Fleischer, in the lachrymal gland of the ox, describes peculiar
eranular forms—ring, crescent, or demilune granules—which
had previously been observed by M. Heidenhain (’90) (See
other references to him in Bibliography), Nicolas, and Held.
Held, in the submaxillary gland of the rabbit, saw this crescent
or ringlike type in addition to the highly and slightly refracting
granules. Nicolas failed to observe them in the parotid gland.
Heidenhain describes these Halbmondkérperchen in the ac-
cessory sex glands of the Triton. The tissue was taken during
the heat period. According to him these demilune granules
represent a phase in the secretion of the cell; they are formed
from small structureless primary granules, and in turn pass into
large round secondary granules which are excreted from the
cell. . Fleischer found these types of granules in the lachrymal
gland of the ox. In sections fixed in 10 per cent trichloracetic
acid or picric acid with brillantschwarz-toluidin-blausafranin,
he describes (1) round deeply stained small granules, (2) granules
of the same size and form which take on a much lighter stain
(3) granules with two zones, a dark crescent shaped cap (kapuze)
which partly surrounds a lighter stained trager, (5) oval forms
of the type just named, (6) granules in the form of crescents or
demilunes without any indications of a trager or lighter stained
area, (7) granules with ringlike borders surrounding a lighter
stained center. These various types were seen without any
special arrangement or distribution. Some cells possessed only
the deeply stained round forms, others contained only the demi-
lunes, and others again were seen in which all the various forms
were present. [Fleischer verified his observations by at least
four different kinds of fixations in order to escape the possi-
bilities of artifacts. He saw also these variously formed granules
in the fresh gland. Hence he was positive of their presence. No
Halbmondformen granules were seen in the parotid or sub-
maxillary gland of the calf.
Granules in Fresh Tissue. My observations show that gran-
ules are readily seen in all fresh lachrymal gland tissue when
THE LACHRYMAL GLAND 185
mounted in either serum or physiological salt solution. The
refractive index of these granules is much greater than that of
the mounting media and consequently the granules stand out
clearly. In the great majority of tissues examined, practically
all cells of both the tubules and the intercalary ducts appear to
possess granules. Only rarely are any of these cells seen free
from granules. This condition is due to the fact that the few
eranule-free cells are masked by granular cells which may be
above or below them as it is impossible to procure fresh tissue
sections one cell layer thick. Fresh lachrymal gland tissue
consequently appears as a rule to be in an extremely granular
stage. It is possible, however, after examinations of numerous
glands to make out several phases in the secretory activities of
glands although this is much better seen in permanent prepar-
ations. These various phases observed in fresh tissue mounted
in ox serum may be generally classified into three groups or
types. However, every gradation exists between these groups.
A. The first type may be termed the granular stage—in the
height of granule formation. All cells of both tubules and inter-
ealary ducts are filled with granules. The alveolus or tubule is
easily recognized surrounded by the basal membrane and inter-
stitial tissue. The cell boundaries are readily seen. The
majority of cells forming the tubules are completely filled with
secretory granules. Consequently no clear basal zone is present
as is the case in the pancreas. These granules extend from the
base of the cell to the summit. So numerous are they that most
cells appear to bulge and as a consequence no lumina are seen.
These granules obscure for the most part the nucleus of the cell.
In many cells, however, the distal half of the nucleus can be
seen lying on the base of the cell. The granules vary in size
from almost imperceptibly fine structures to granules several
micra in diameter. They vary also in their powers of light
refractivity—the smaller ones refract light to a much greater
degree than do the larger. There is no regularity in the dis-
tribution of large and small granules within the cell. They are
scattered throughout the cell, appearing at both the base and
the summit. Frequently large granules are seen which have
186 JOHN SUNDWALL
dark ring-like contours. ‘This phenomenon is undoubtedly due
to variations in the refraction of light as it passes through the
granules and does not represent any peculiar constituent of the
granules as held by Fleischer.
The granules in the intercalary duct show the same char-
acteristics as do those of the tubules so far as size and the powers
of refraction are concerned.
B. In the second group or type of fresh glands examined,
one sees considerable variation in the form and size of the cells
of the tubules as well as in the quantity of granules—large bulg-
ing cells described under A and smaller cells which have various
forms (irregularly cylindrical, pyramidal, or appearing as demi-
lunes wedged in between the bulging cells). These smaller
cells may be filled with granules or may possess but few. Their
nuclei are always round or oval and are separated from the base
of the cell by a narrow zone of cytoplasm.
C. Glands are occasionally seen in which the tubules for the
most part are made up of cells which are regular in form—low
cylindrical, pyramidal or cuboidal in outline. The granules
are confined to the proximal half of the cells. The nuclei are
round and oval and do not touch the bases. The lumina of the
tubules are wide and open. ‘This condition represents a stage
of partial exhaustion of granules.
At the margin of these fresh tissue preparations, cells are seen
which have been ruptured by teasing, cutting, ete. Here the
granules are seen passing from these cells into the serum and
still retaining the same continuity as they possessed within the
cell. In fact the entire margin of the tissue is characterized by
the presence of these free granules. In order to exclude the
possibility of the more finely formed elements in the blood being
responsible for this loose granular mass the serum was carefully
filtered before using as a mounting medium. Microscopic
examination of the serum, after filtering, revealed no such formed
elements.
When sections of fresh tissue are mounted in physiological
salt solution, the cells show the same characteristics as those
described for the serum mounted tissue. The loose granules
THE LACHRYMAL GLAND 187
do not preserve their continuity as long, however, in the salt
solution as in the serum.
Fresh tissues mounted in distilled water soon lose (in five
minutes) their granules. As a consequence the nuclei which
are round or oval become very prominent. They are seen for
the most part in the basal portion of the cells and are opaque
and homogeneous. No nuclear granules are seen. In no in-
stance are the nuclei seen to be angular and flattened as in many
fixed and stained preparations. They have in all likelihood
taken up enough water to be swollen when observed under
these conditions. The water is absorbed before the granules
have sufficiently disappeared for the process to be observed.
If 2 per cent salt solution is drawn under the slide after the
granules have disappeared in distilled water they reappear again.
This phenomenon was observed by Noll in the lachrymal gland
of the cat and by Langley and others in other serous glands.
When fresh tissues are mounted in five per cent solution of
hydrochloric acid or sodium bicarbonate the granules become
indistinct—disappearing much more rapidly in the former solu-
tion than in the latter. An intracellular network remains after
treatment with these solutions, the meshes of which were formerly
occupied by the granules. This net work should be interpreted
in the sense of Bensley ‘‘optical sections of the continuous cyto-
plasmic partitions which separate the granule holding spaces
from one another.”’
Thin sections of fresh lachrymal gland were mounted in the
following vital stains in physiological salt solution with these
results:
1) naphthol blue—granules stained a purplish blue.
2) neutral red—granules were very faintly stained. I did not
observe here prozymogen granules more deeply stained as ob-
served by Bensley in the pancreas of the guinea pig.
3) janus green—the mitochondrial elements were stained.
This will be discussed under that caption.
4) trypan blue—a diffuse light blue stain, granules only
faintly stained.
188 JOHN SUNDWALL
5) isamin blue—a very faint blue stain limited chiefly to the
interstitial tissue. Certain cells in the alveolus also appeared
more deeply stained than did others.
6) trypan red—a diffuse light red stain.
7) sulforhodamin—a diffuse faint light red stain.
8) pyronin—intercellular canals and lumena of ducts stained.
Granules in fixed and stained preparations. Zenker’s. At-
tention has already been called to the fact that the granules of
the intercalary duct when fixed in Zenker’s solution stain deeply
in muchaematein and mucucarmin while the other granule
stains do not affect them. The tubules, on the other hand, show
no deeply stained granules but appear under low power only
diffusely and faintly stained. Careful examination of these
tubules with the oil emersion, however, shows various intensities
of staining reactions on the part of the cells—which may be
generally classed as follows: a) Large rounded bulging cells as
seen in the fresh sections. Here a faintly stained cytoplasmic
network is seen throughout the cell. The meshes of this net-
work are round and appear to hold large unstained granules.
The nucleus is angular or flattened and lies against the base of
the cell. In fresh tissues the granules of these cells are dis-
tinctly seen and the intergranular network is not seen, while in
Zenker’s muchaematein preparations the granules are not
stained but the intergranular network is distinctly seen. b)
Cells more cylindrical in form frequently showing indented
sides as if they had been pushed in by the round bulging cells
(also seen in fresh preparations). They possess very faintly
stained granules. The nuclei as arule are round or oval. c) So
called transitional cells. In the use of the term transitional
cells reference is made to those cells that mark as a rule the
junction of the intercalary duct and tubule. They possess
granules similar in structure and staining reaction to those of
the interealary duct and are always preserved in Zenker’s solu-
tion (figs. 11 and 12).
Bensley’s solutions. For the fixation of granules inthe tubules,
it was necessary to use other solutions than Zenker’s. The
granules in tubules are best preserved by Bensley’s sublimate
THE LACHRYMAL GLAND 189
alcohol bichromate solution (Technique IV, 2 a) and by formalin
bichromate sublimate solution (Technique IV, 2 ¢). Tissues
must be small as the penetration of these solutions is not great.
In these solutions the granules of both the tubules and the
interealary ducts are preserved. These are readily stained with
muchaematein (Technique V, 1), mucicarmin (Technique V, 2)
iron haematoxylin, iron haematoxylin counterstained with
mucicarmin, copper chrome haematoxylin (Technique V, 5) neu-
tral gentian (Technique V, 6) and safranin-acid violet (Tech-—
nique V, 7).
Mucous stains. Sections fixed in Bensley’s sublimate alcohol
bichromate solution or in formalin bichromate sublimate solu-
tion when stained with muchaematein or mucicarmin show
that the selective staining is confined entirely to the granules
of the tubules and ducts. The various types of cells as seen in
these sections depend upon the secretory state of the gland at
the time of fixation. The various stages described in the tubules
of fresh tissue are seen in these preparations and may be grouped
as follows: .
(1) Large rounded or bulging cells completely filled with
intensely blue stained granules—type 1. All other structures
within these cells are obscured as a rule. Frequently, however,
flattened nuclei are seen at the base. These cells correspond to
those described in both fresh tissues and in Zenker’s muchae-
matein sections. In the latter case the granules were not stained
but the intergranular network was prominent as a consequence
of a light blue stain.
(2) and (8) Cells more or less cylindrical or pyramidal in
outline, or appearing as demilunes in the periphery of the tubules
completely shut off from the lumen by the large bulging cells—
frequently the cylindrical cells are constricted in their vertical
axis by the pressure of the bulging cells and thus simulate in
outline hour glasses. The nucleus is generally round or oval and
does not lie directly on the base of the cell but is separated from
it by a narrow zone of cytoplasm. These cells can be divided
into two classes. The one contains relatively few or, rarely,
no granules. The granules are. seen in the proximal end of the
190 JOHN SUNDWALL
cell or throughout the cytoplasm and stain with different de-
grees of intensity—many are but faintly stained. These will
be referred to as type 2. The cells of the other group (type 3)
are similar to those of type 2 in form but are full of deeply stained
granules.
Generally there is no regularity of distribution of these various
types. Tubules are seen wherein all the cells composing them
are of the first type. Others are seen made up of all three types.
Occasionally glands are obtained wherein practically all the
cells within the tubules are cuboidal in form and show the granules
proximal to the nuclei as described under fresh tissue (Fresh
tissue C). Here the cells are not large or bulging and the lumina
are open.
In other sections, which are rare, one sees numerous granules
with deeply stained rings around them. ‘These were described
by Fleischer who considered them as a stage in lachrymal secre-
tion. I do not think such is the case, however (see Discussion
of granules.)
The cells of the intercalary duct show the same characteristics
as those described in Zenker’s celloidin preparations.
Iron haematoxylin. In iron haematoxylin stained sections
one sees the same types of cells described under the mucous
stains. The large bulging type 1 cells are filled with black
granules of various sizes, the nuclei angular or flattened against
the bases of the cells. The type 2 cells have grayish granules
and their round or oval nuclei show as a rule one or two large
chromatin masses. The type 3 cells are filled with deeply stained
granules or a mixture of both black and gray granules.
The different intensities of the staining reaction on the part
of granules is a striking feature of the tubules. Large and small
granules may be seen side by side in the cell some staining deeply
black while others are gray or faintly stained. The granules of
the interealary duct vary also in the intensities of staining re-
actions but not to such a degree as those of the tubules.
Iron haematoxylin and mucicarmin. Sections stained with
iron haematoxylin and counterstained with mucicarmin show
some very interesting features. All the granules of certain
THE LACHRYMAL GLAND 19]
cells stain red in mucicarmin and all the granules in other cells
stain black in iron haematoxylin while in still others both red
and black granules are seen in the same cell (fig. 13.)
Copper chrome haematoxylin. Sections stained in copper
chrome haematoxylin show the granules to be deeply stained.
The cells are readily classified into the three types already
described. The large bulging cells are filled with black stained
granules of various sizes. The granules in the type 2 cells are
very lightly stained—in many they are so faintly stained that
13
Fig. 13 Secretion cells with granules, Tech. IV, 2, a; V. 4. Leitz. oc. 2. obj.
fs oil. A. intralobular duct; B, cells possessing granules all of which stain
black with iron haematoxylin; C, cells with the entire granular content stained
red with mucicarmin D, cell with granules some of which are stained with iron
haematoxylin and others with mucicarmin; /, secretory cells with no granules.
it is difficult to make out any granules whatsoever. The type 3
cells are filled with black granules.
Sections fixed in the sublimate alcohol bichromate solution
and stained in copper chrome haematoxylin show, in the larger
ducts, mitochondria either as irregularly distributed granules
or as filaments (see Mitochondria. )
Neutral gentian and safranin-acid violet. These proved to be
excellent stains for granules in the lachrymal gland. Tissues
fixed in either the sublimate alcohol bichromate or the formalin
bichromate sublimate solution show the characteristics already
192 JOHN SUNDWALL
described. In fresh safranin-acid violet the nuclei stain red
while the granules stain a dark greenish blue. The advantage
of this stain lies in this contrast. In the large bulging cells
the red flattened basal nuclei are generally seen while in the
other stained cells they are made out with difficulty sinee the
colors of nuclei and granules are similar.
The granules in the cells of the intercalary duct stain similarly
to those of the tubules. Lumina are always seen but vary in
diameter. The cells are generally uniform in size and shape.
No bulging cells are seen. In each cell there is a narrow zone of
non-granular cytoplasm surrounding the granules. No vari-
ations are seen in the size, shape, and position of the nuclei
whether the cell contains the maximum of granules, only a
narrow zone at the proximal end, or no granules whatsoever.
They are similar in structure and staining characteristics to
the round and oval nuctei seen in the tubules.
In thin sections stained with neutral gentian the outer layer
of cells surrounding the secretory cells of the interealary duct
are readily made out. Near the origin of the duct these cells
are still more or less cuboidal but towards the tubular end the
cells with their nuclei become more and more elongated. These
cells stain more faintly than do those of the inner layer. They
surround the tubule as elongated cells between the secreting
elements and the basal membrane (see Connective Tissue. )
As in the case of the iron haematoxylin stained sections one
is especially struck by the various intensities which these granules
stain with neutral gentian and safranin-acid violet. In well
fixed tissues the majority of granules stain deeply in both of
these dyes. However, one frequently sees in the same cell, in
both the tubules and the interealary ducts, granules lying in
close proximity to each other some of which are stained deeply
blue, others only faintly yellow or brown, and still others barely
stained at all. This variation in intensity of staining is in no
way related to the position of the granule or its size. Cells
are frequently seen bordering each other, one possessing deeply
stained blue granules throughout, the other possessing yellow-
ish brown or faintly stained granules. The probable expla-
THE LACHRYMAL GLAND 193
nation of this phenomenon is, as suggested by Bensley—‘‘the
granule substance has:a primary difference of density due to
difference in water content or a change in dispersion grade of
the colloid.’”’ Certainly some granules retain the stains much
more tenaciously than do others. I have not observed these
extreme variations in the staining reaction of granules in other
glands such as the pancreas and submaxillary.
The secretion granules of the lachrymal gland are not stained
in neutral red or in any other of the vital stains employed.
I saw no free granules in the lumina of either the tubules or
the ducts. Frequently such appears to be the case however,
for one sees, especially in the intercalary duct, elongated groups
of granules which seem to be in the lumen. Careful study,
however, reveals the fact that these granules mark the proximal
end of the other cells which contribute to the formation of the
lumen.
No cuticle, striated or otherwise, was observed in the epithelium.
Discussion of granules
' Secretory changes in cells. Unfortunately the lachrymal gland
of the ox does not lend itself for experimental study of a stage
of rest, a stage of activity, and a stage of exhaustion. Studies
based upon microscopic changes in the cells as induced by rest
and stimulation must be accomplished in the more accessible
laboratory animals. My study of granules was based chiefly
on the methods of their conservation and their staining char-
acteristics. However, a sufficient number of glands was exam-
ined to gain some light upon the changes that do occur in the
normal secretory activity of the gland. According to the tissues
fixed in Bensley’s sublimate alcohol bichromate solution and in
formalin bichromate solution supplemented by a study of fresh
tissues the secretory phases of the gland may be classified as
follows:
A. A maximum granular stage in which most of the cells of
the tubules and interealary ducts are filled with granules. The
majority of cells in the tubules belong to those described as type
194 JOHN SUNDWALL
1 under the various stains—completely filled with granules and
bulging, lumina of the tubules not visible.
B. A medium granular stage in which the tubule possesses for
the most part cells of the types described as types 2 and 3.
Rarely does one find an entire tubule made up of these types
alone, as bulging cells of type 1 are generally present. It will
be recalled that type 3 cells are those described as cylindrical,
pyramidal, or crescent cells filled with deeply stained granules.
The nuclei of these cells are still round or oval and separated
from the base of the cell. This type doubtless represents an
earlier phase to that of type 1 in the stage of granule formation.
Granules have formed in sufficient quantities to fill the cells but
not to such an extent as to enlarge them—cause them to bulge.
Cells described as type 2 in all likelihood represent two phases
in secretion. a) Those which possess few granules at their
proximal ends suggest the terminal phase of secretion wherein
the granules for the most part have passed out of the cells.
b) Those in which comparatively few granules are seen scattered
throughout the cells suggest cells in the beginning phase of granule
formation. These granules stain with varying intensities. Some
are large and pale, others small. There is no regularity to their
distribution. The few nongranular cells may be interpreted as
occupying a position between the terminal of granule expulsion
and the beginning of the formation of new granules. The very
few nongranule cells seen and the fact that the type 2 cells
usually possess granules both at the lumen end of the cell (term-
inal secretion) and also a few granules throughout the cytoplasm
(beginning granule formation) strongly suggest that the extrusion
* of granules and the new formation of granules go hand in hand
in the same cell during the normal secretory activity of the gland.
We have then in this gland a continual secretion as indicated by
the cells showing all phases of secretory activity.
C. In the third group of glands the secreting cells were much
reduced in size, the lumina wide and open, the few granules
present confined to the proximal end of the cell, and the nuclei
round and oval. This condition was uniform throughout the
gland and simulated preparations obtained by other observers
THE LACHRYMAL GLAND 195
after stimulation with pilocarpin. In my opinion these glands
represent a condition of intense secretion before the animals
were killed. Many factors could be responsible for this exces-
sive activity.
The results of my study of the various phases of secretion
agree in a general way with those of other observers. Detailed
study of the cellular changes during secretory activity, must be,
as stated, on the laboratory animals. Preliminary work on cer-
tain species of Anura and Urodela gives much promise, and I
trust that a paper will soon appearonthissubject. In thelachry-
mal glands of the ox I have failed to observe the secretory ele-
ments of the cell divided into two zones by a strand of reticular
protoplasm stretching across the cells as noted by Bensley (’02)
in the cardiac glands of mammals and in the glands of Brunner
(Bensley ’03). Nor was an intermediate stage present as is
seen in the Pancreas (Bensley, (11) which contained small gran-
ules (prozymogen granules) that stain with neutral red, intra
vitam.
I have failed to observe the paranucleus described for secreting
cells, including the lachrymal gland, by Gaule, Ogata, Nuss-
baum, Garnier, and others. As stated elsewhere, the three
zones described by Zimmermann were not observed. No
definite light has been thrown as a consequence of this study
upon the origin of these granules. They seem to make their
appearance anywhere within the cytoplasm, independent of
position or of any marked demonstrable antecedents such as a
basophil substance (toludin blue), iron possessing substance
(Macallum reaction), or prozymogen (Bensley’s vital neutral
red). Exception to this statement should be made if we are to
consider mitochondria as an intermediate substance. This
will be discussed under mitochondria. Perhaps the presence of
such antecedents should not be anticipated, as the lachrymal
gland possesses so far as I know no demonstrable specific enzyme.
Whether the nucleus plays any part in the formation of granules,
I am unable to say. It is true that in the tubules when in a
stage of maximum granule formation (cell type 1), the nucleus
is more or less flattened and compressed against the base while
196 JOHN SUNDWALL
in cell types 2 and 3 it is round or oval and separated from the
base of the cell by a narrow zone of cytoplasm. However, no
distinct changes are seen in the chromatin. In both instances
one or frequently two large chromatin masses are surrounded by
numerous finer granules of chromatin. The change of position
and form of the nucleus may be due entirely to the quantity of
granules. The nucleus of the intercalary duct, on the other hand,
shows no appreciable change in form or position whether the
cell is at its maximum stage of granule formation or its minimum.
The fixation and staining of granules. When the tissues are
properly fixed in either Bensley’s solution or formalin bichloride
bichromate solution all granules are conserved. Such is not
the case when alcohol, Zenker’s or numerous other fixatives are
used. Zenker’s and Altmann’s fix the granules in the intercalary
duct and frequently in some few cells in the tubules. The
granules in the large bulging cells are not preserved in alcohol,
Altmann’s, or Zenker’s but, instead, the intergranular network
remains. ,
Neutral gentian, safranin acid violet, copper chrome haemato-
xylin, iron haematoxylin, muchaematein, and mucicarmin
proved to be the most satisfactory stains for the granules.
Are secretory cells of tubules similar in function? Notwith-
standing the variable staining reactions on the part of granules
and the various sizes and shapes of cells, cells of the tubules
are doubtless similar in function. We have no evidence that
more than one functional type of cell is present. Proof of this
is seen in the granular stage when all the cells of the tubules
stain similarly. Furthermore other structures in these cells
are alike so far as I have been able to make out. The changes
in position and form of the nuclei only represent different secre-
tory phases.
Staining reaction of granules, selective action. In order to
prove that the cells in both the tubules and the intercalary
ducts are capable of taking various granular stains, serial sec- ~
tions three micra thick were cut and fastened to different slides—
one section for each slide. The first section was stained with
neutral gentian, the second with muchaematein, the third with
THE LACHRYMAL GLAND 197
iron haematoxylin, the fourth with mucicarmin, and so on until
all the granule stains were used. It was found that all granules
in these sections were stained with the various dyes. This
demonstrated that the granules of the same tubule and inter-
ealary duct are capable of taking various stains. Furthermore
the same granule is apparently capable of taking any one of
these stains. This is shown in the counter staining. When
sections stained in either iron haematoxylin or copper chrome
haematoxylin are differentiated to such an extent that the gran-
ules are pale, then stained with mucicarmin they appear red.
This condition is further substantiated by sections from those
rare tissues which contain the peculiar ring-like or crescent
granules. In many of these granules the caps or crescents and
the rings stain black in iron haematoxylin while the bodies of
the granules stain red after counterstaining in mucicarmin.
That many granules, however, show an affinity for certain stains
seems very probable, as cells are present in the tubules and
interecalary ducts the entire granular contents of which stain
black with the iron haematoxylin while all the granules of the
neighboring cells stain red after counterstaining with mucicarmin,
and again a mixed condition of red and black granules is seen
in other cells. However, the distribution of these red and black
granules within one cell as well as the distribution of cells con-
taining either all black or all red granules in the tubules and
interealary ducts is very irregular and the ratio of these cells
to each other is very inconstant.
Various explanations for this selective action on the part of
some granules for iron haematoxylin, others for mucicarmin
may be advanced as follows:
1) The granules at different stages of their development have
different selective action for various stains. Against this view,
however, is the fact that this selective action on the part of
granules bears no relation to their size or distribution in the
cell. Large and small granules, either at base or summit, may
take one or the other stain. The large bulging granular cells
of the tubules seem to have a greater affinity for mucicarmin,
however, many exceptions to this observation were noted.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 2
198 JOHN SUNDWALL
2) Even the granules within one cell may be subdivided into
a number of groups, each group representing a specific chemical
structure and contributing some element to the sum total of
secretion. If such be the case one may readily assume a selec-
tive action on the part of each group for specific stains. Of course
we have no proof for such an assertion.
3) A possible explanation les in the fact that in the differen-
tiation of sections stained in either iron haematoxylin or copper
chrome haematoxylin some granules hold these stains much more
tenaciously than do others. This may be due to varying degrees
of fixation of the granules. It is the faintly stained granules,
those differentiated most, that are stained again in mucicarmin.
This observation had been made.
However, as Bensley has pointed out:
The difficulty of all these discussions of the different staining prop-
erties of granules in the same cell arises from our ignorance of the
ehanges which would be produced in the absorptive properties of these
colloids with variations in their water content and so in their dispersion
grade. The dyes which we use for this staining are all dyes which
may be absorbed. I think under the circumstances it would be rash
to assume a difference in the fundamental composition of the granules
on the basis of the different staining.
Are the cells of the tubules and intercalary ducts of the same
functional type? Attention has been called to the fact that the
intercalary duct stains more deeply in the ordinary dyes than do
the tubules (v. Ebner, Nussbaum, and Langley, submaxillary
gland; Merkel and Fleischer, lachrymal gland). Fleischer con-
cludes that this duct has a specific secretion differing from that
of the tubules. He gives as his reason the presence of such
large granules.
I too have observed that the interealary duct as a rule stains
more deeply in the ordinary dyes. This can be readily explained
by the fact that the cells are smaller and more compact and
that the granules are conserved in such general fixatives as
Zenker’s solution, while in the majority of cells of the tubules the
granules are not fixed but an intergranular cytoplasmic network
remains.
THE LACHRYMAL GLAND 199
Fleischer holds that a specific secretion is produced in the
interecalary duct:
Das Yorhandenseim derartiger grosser Granula in den Schaltstiick-
zellen, wie ich sie auch in der Thrinendriise gefunden habe, spricht
mit grosser Wahrscheinlichkeit fiir eine besondere sekretorische Bedeu-
tung dieser Zellen und zwar muss es sich um eine andere Art von Sekret
handeln, als dasjenige, das die Zellen der End-abschnitte auscheidet.
While I found that the granules in the cells of the tubules—
cell type (1)—in well fixed preparations are as large or even
larger than those in the intercalary ducts, still certain marked
differences do exist between the cells of the intercalary duct and
the tubules. The granules in the intercalary ducts are fixed
in Zenker’s solution while those of the tubules are for the most
part not preserved in this solution. When fixed in this solution
the granules of the intercalary ducts stain deeply and readily
-in mucicarmin and muchaematein while the other granule dyes
do not affect them. In tissues fixed in Bensley’s acetic, bi-
chromate osmic solution and stained with anilin fuchsin-methyl-
green these granules are stained green while they are not pre-
served in the bulging cells of the tubules.
A narrow band of non-granular cytoplasm is always seen
between the cell membrane and the granules. The granules
never reach the base of the cells. The cells do not undergo any
marked variation in size or shape during their secretory activity.
The nuclei remain practically constant in form and position
whether the cells are empty or possess numerous granules. No
secretory capillaries are present. Frequently, however, inter-
cellular indentations are seen. These are some of the features
that characterize the cells of the intercalary duct.
In the tubules, on the other hand, the granules in the majority
of the cells are not fixed in Zenker’s solution, only those that mark
the transition from the duct to the tubule are fixed. The
granules when fixed in the height of the secretory stage fill the
entire cell and cause it to bulge out. Marked changes conse-
quently appear in the form and position of the cell which is true
also in the case of the nucleus. Secretion capillaries are also
present.
200 JOHN SUNDWALL
Nature of granules. Ellenberger states that the lachrymal
gland of the pig is a mucous secreting gland, while Boll states—
regarding pig, sheep, calf, and dog—‘‘ Es ist hieraus mit Sicher-
heit zu erschliessen dass das Secret der Thréanendriise nie mucin
enthilt.’”’ Which is true in the case of the lachrymal gland of the
ox? Although the specific reaction of these granules to the
mucous stains might indicate that this gland may be mucous in
character, in my opinion it cannot be classed as mucous, since:
1) If one takes fresh pieces of the pancreas, submaxillary, and
lachrymal glands and compares the secretions which can be
pressed out one is impressed with the similarity of certain physi-
eal properties of the secretions of the pancreas and lachrymal
glands. ‘The fluids of both are thin and watery with no adhesive
qualities. The secretion from the submaxillary gland, on the
other hand, is viscid, thick and sticky to such an extent that
pieces of tissue are readily supported by it. 2) A careful analysis
of numerous lachrymal glands from the standpoint of their
physiological chemical structure showed that they are in no
wise similar to mucous glands. 3) The structure of the cells of
the tubules and interealary ducts does not simulate those of the
submaxillary gland and other mucous glands. 4) One finds that
the lachrymal gland has the following characteristics which
are true of serous glands in general: a) The granules have a
high refractive index. b) The granules are insoluble in weak
alkaline solution. Following Bensley’s procedure I found that
sections fixed in his bichromate alcohol sublimate solution
showed no change after having stood for twelve, twenty-four
and forty-eight hours in 5 per cent potassium carbonate solution
at a temperature of 38°, while in similarly treated sections
of the submaxillary gland the granules of the mucous cells
had completely disappeared in twenty-four hours. The granule
stain in the lachrymal gland is as definite and as intense as Is
seen in the stained normal sections. Similarly treated sections
in five per cent hydrochloric acid for the same length of time
did not affect the granule stain in the least. ¢) The granules
stain readily in all serous granule stains. d) The presence of
secretory capillaries. On the other hand certain characteristics
THE LACHRYMAL GLAND 201
are present also which are generally true of mucous secreting
cells: a) The flattened nucleus seen compressed against the
base of the cell in the large bulging cells. b) The absence of
basal striations as seen in many serous cells. ¢) The presence
of but minute traces of prozymogen (toludin blue stain) and
organic iron—the latter demonstrated by Macallum’s method
d) The specific reaction to the mucous stains.
5) Examination of lachrymal secretion in the human shows
only a small amount of mucin present which is readily accounted
for by the presence of goblet cells both in the ducts and con-
junctiva.
Analyses of human lachrymal secretions show:
FRERICHS (46) ARLT LERCH? | MAGAARD (82)
Wither casera ie Nett eee eras SOONG 98 .70 98 .233 98 . 1200
Mprblreliumimenvessrssns crs cc se ae tee 0.14 0.32
PAU OUITMTE eee sen ne mee Be as ae 0.08 0. 0.504 1.4638
IMRTOUES BING CIB oes hob aee bose se 0.03 0.34 Trace
Salt NaCl eR AE Pri
Ritosphateeaerern coaratetee fen: \ 0.43 0.54 1.257
Othersaltsapysncnca ceca aes: j 0.016 ey
No positive proof can be advanced that these cells are either
serous or mucous in character as we have no specific methods for
determining this. The lachrymal gland may be considered as
not highly specialized in function when compared with such
glands as the pancreas. It appears to occupy an intermediate
position between the more highly specialized serous and mucous
glands and possesses many characteristics of both. My results
plainly show that great care must be taken in determining the
nature and function of glands from the standpoint of micro-
chemical staining.
Rings and demilune granules. Attention has been called to the
fact that these peculiarly shaped granules have been described
by Held, Nicholas, and Heidenhain in other glands, and espe-
cially by Fleischer in the lachrymal gland of the ox, who described
* Nagel’s handbuch.
202 JOHN SUNDWALL
them as being present in all lachrymal glands and stated that
they represent constant and distinct phases in the evolution of
the granule. My observations do not agree with those of Fleisch-
er. While these peculiar forms of granules were seen occa-
sionally in the numerous preparations of lachrymal gland which
I made, their occurences were relatively so rare that they must be
considered as exceptional. I have also seen these peculiar gran-
ules in the accessory lachrymal gland tissue of the third eyelid
of the ox as well as in the orbital glands of frogs. Frequently
ring granules are seen in fresh preparations of these glands, but
the explanation les in the effect that a large granule may have
on light passing through it. Light may pass through the center
of the granule with but little or no refraction while it may be
ereatly refracted at the contour. I think that these peculiar
granule structures in fixed preparations can be explained from
the standpoint of fixation. It is the large granules which show
these peculiarities when present at all. It can be readily con-
ceived that the surfaces of these granules may be acted upon
more thoroughly by the fixing fluids than the centers and con-
sequently stain more deeply. Small tissues well fixed in Bensley’s
solution or formalin bichromate sublimate solution seldom show
these peculiar granule structures. It is difficult to conceive of
these granules of fluid or semifluid consistency taking the form of
rings, demilunes or crescents. Our knowledge of the physical
properties of matter in such condition will not permit of such a
conclusion.
VI. MITOCHONDRIA
I found that lachrymal gland tissues prepared according to
Bensley’s method for the demonstration of mitochondria—
acetic osmic bichromate, anilin fuchsin methyl green—show these
elements very clearly.
Tubules. In the tubules two chief cell types are made out.
somewhat similar to those described by Noll in the lachrymal
gland of the cat after fixing with Altmann’s solution—lght and
dark cells.
THE LACHRYMAL GLAND 203
Light cells: The light cells correspond to the large bulging
cells (type 1) heretofore described. The secretory granules are
not preserved in this solution but an intergranular cytoplasmic
network remains. The meshes, which originally possessed
granules, now appear empty. This condition is seen also in
tissues fixed in Zenker’s solution. Irregularly | distributed
throughout this dark stained cytoplasmic network are seen these
minute fuchsinophil granules or mitochondria. The majority
are slightly elongated simulating somewhat bacilli. Others
again are round, cocci like. They are seen in every portion of
the cell from the base to the summit. (fig. 14). In no instance
Xatharine Hill.!
Fig. 14 Mitochondria in cells of tubule, Tech. IV, 2, e; V. 8. Zeiss, oc. 6,
obi. 2mm. oil. Mitochondria are readily seen irregularly distributed through-
out the cells.
have I seen in these large bulging cells mitochondria arranged
in the form of filaments at the base of the cell. Nor is this to
be anticipated, for these cells in other fixations (formalin Zenker
and alcohol bichromate sublimate solution) are filled with gran-
ules from the base to the summit. Occasionally larger fuchsino-
phil granules are seen which resemble the secretion granules in
size and shape, and are seen surrounded by the darker cyto-
plasmic network. These secretion granules when present show
as much affinity for fuchsin as do the mitochondria. I have
204 JOHN SUNDWALL
frequently observed these fuchsinophil secretion? granules in
the orbital glands of frogs. Just what the significance of these
granules is, I am unable to state. The explanation that they
represent an intermediate stage between the secretion granules
and mitochondria is unwarranted in view of the fact that so few
are seen.
In some cells, though rarely, fuchsinophil globules are seen
which are much larger than the secretion granules. It will be
well to add here that red blood cells possess a marked affinity for
fuchsin, although I was unable to make out mitochondria within
them, confirming Cowdry (’14 ¢).
Other structures are seen in the cell. These are stained black
with the osmic acid. The larger spherical ones undoubtedly
are neutral fats. These can be demonstrated with Herxheimer’s
stain. The finer structures are liposomes. (See discussion
under Fat.)
Dark cells: These cells are so termed because the darkly
stained cytoplasm shows no network with empty meshes but
instead is either filled with deep green stained secretion granules
or is continuous. These cells correspond to types 2 and 3 here-
tofore described. In form they are irregularly cylindrical,
pyramidal, elongated, or hourglass shaped due to indentation
caused by the bulging type 1 cells. Many appear as demilunes.
The position and form of the nucleus are similar here to that
already described for these cells. The transitional cells of Noll
are seen also wherein the central half of the cell possesses the
characteristics of the light cells while the basal half shows the
structures described in the dark cells. Throughout the cytoplasm
mitochondria are readily observed. In the cells filled with
secretion granules they are very plainly seen between the gran-
ules. The contrast is marked. In those cells which possess few
or no granules, the red stained mitochondria arereadily observed
in the dark stained cytoplasm.
There is no regularity of distribution of mitochondria in the
cells of the tubules in any of the types. Frequently one sees
them in clumps at the central end, frequently at the basal end.
THE LACHRYMAL GLAND 205
More often, however, they are irregularly distributed through-
out the cell. In all the cells of the tubules the mitochondria
appear as very short irregular rods or as minute spheroids which
may be slightly irregular in outline. In no instance have I
seen long rods, threads, or loops as described especially for grow-
ing cells. Further, in the tubule the mitochondria do not form
rows in the base of the cell and perpendicular to its base (fila-
ments) such as found in the larger duct cells. Careful study
failed to reveal any general irregularity in the quantity of these
fuchsinophil granules. While some cells appear to possess more
than do others yet the hight cells (type 1) which represent the
maximum of the granular stage apparently possess approxi-
mately the same number as do the dark cells (types 2 and 3)
which represent either earlier or later stages of secretion granule
formation. The mitochondria in these cells are not numerous
when contrasted with those seen in the cells of the intralobular
ducts (Cf. figs. 14 and 15). Consequently the granules do not
appear to be used up in the formation of secretion granules.
Intercalary duct. The secretion granules in the intercalary
duct are well preserved in Bensley’s acetic osmic bichromate
solution. In fact, so far as the preservation of secretion gran-
ules is concerned the effect of this solution is similar to Zenker’s.
The granules here have a marked affinity for the methyl-green.
Between these green granules red stained mitochondria are
readily observed. They simulate those in tubules so far as
size, shape and distribution are concerned.
Larger ducts. In the intralobular and interlobular ducts the
mitochondria are especially abundant. In the smaller intra-
lobular ducts they may be so numerous that the entire cell, with
the exception of the nucleus, appears to be composed entirely of
these fuchsinophil granules. In these cells there exist marked
irregularities so far as their distribution and arrangement are
concerned. Some cells may be full of these granules while
others again possess but relatively few. Generally the granules
are arranged in rows—these rows being parallel to each other,
perpendicular to the base of the cell, and extending through-
206 JOHN SUNDWALL
out the entire length of the cell (fig. 15.) Cells are frequently
seen, however, in which the mitochondria appear in heaps with
no definite arrangement whatsoever.
The mitochondria of the ducts are readily seen when preserved
in Bensley’s sublimate alcohol bichromate solution (Technique
IV, 2) and stained in copper chrome haematoxylin (Technique V,
5). Here they have an arrangement similar to that already
described. Whether the mitochondria in the tubules and inter-
Fig. 15 Mitochondria in interlobular duct, Tech. same as figure 14. Here
they are arranged as filaments.
calary duct are preserved I can not say. It will be recalled that
this fixative preserves the secretion granules and these stain
deeply black in this copper stain. Consequently the mitochon-
dria if present are masked. I did not see them in tissues of this
fixation when stained with iron haematoxylin or the neutral
stains—neutral gentian, and neutral safranin. In all likelihood
THE LACHRYMAL GLAND 207
mitochondria are affected by these latter stains but in the differ-
entiation do not hold them so tenaciously as do the secretion
eranules. The presence of demonstrable mitochondria in tissues
fixed in alcohol sublimate bichromate solution is interesting in
view of the observations of others that mereuric chloride and
aleohol fixatives do not preserve them.
The mitochondria were also stained by the intra vitam method
of Michaelis (00) (Technique VI, 1, 2). Here they appear as
deep blue structures being the only stained constituents of the
cell. In size, form and distribution they appear as already
described.
Function of mitochondria in gland cells. Regarding the nature
and function of mitochondria in general many hypotheses have
been advanced which may be briefly summarized as follows: 1)
the mitochondrial theory—that these elements are specific ele-
ments of the cytoplasm just as the chromosomes are fixed ele-
ments of the nucleus and like the latter arise from preformed
elements in sex cells and are carried over in all mitosis—this
view was championed by Benda and Meves. It has many sup-
porters and is gaining ground. 2) Others do not accept the
mitochondrial theory but claim that mitochondria arise from
nuclear material, from ferment products of the centriole, or
that they represent other phases in the metabolic activity of the
cell. Many of those who accept the mitochondrial theory hold
that these structures later develop into the fixed specifie struc-
tures of cells; 1.e., neurofibrils (Meves and Hoven 710), muscle
fibrils, ete. Cowdry (’14 a) has demonstrated that in the case of
nerve cells this is not true.
Cowdry (14b) assumes that they have to do with the metab-
olism of the cell since they are almost coexistent with all active
protoplasm. The fact that they do not occur in red blood cells
(Cowdry), in superficial layers of epithelial cells (Firket), and
in the terminal stages of the cycles of development of certain
grains and legumes (Guillermond), according to Cowdry, partially
substantiates this view because these cells are in the terminal
stages of metamorphosis. Mitochondria, according to him,
208 JOHN SUNDWALL
is in all hklhood a lipoid albumen complex. In nerve cells if
they are numerous the lipoid granules are few and vice versa.
In my opinion Cowdry’s conclusions are correct.
Regarding their function in secretory cells, Altmann held the
view that they were capable of forming secretion granules and
also related to the formation and absorption of fat. Noll held
views similar to Altmann regarding their relation to secretion.
Schirmer could not say. Bensley (11) states that certain types
may possibly represent secretion antecedents. Champy (11)
makes the bold assertion that secretion granules originate from
mitochondria. Hoven (11, 712) also hold that the different
products of the mammary gland originate from these granules
and according to him this is true of the secretion granules for all
glands. Arnold (11) lkewise holds the same view.
My studies on the lachrymal gland have not revealed any
positive evidence that secretion granules have their origin from
mitochondria. Certain facts may suggest that such a hypothesis
is tenable. These are, 1) The absence of demonstrable ante-
cedent substances such as prozymogen—hbasophile substance
(toluidin blue), prozymogen granules (intra vitam, neutral red),
nuclear material in cytoplasm (Macallum reaction). 2) The
secretion granules seem to make their appearance in any part
of the cell independent of the nucleus. 3) The small amount of
mitochondria in the secreting cells as contrasted with that in the
cells of the ducts suggests that it may be used up in the formation
of secretion granules.
On the other hand as valid objections can be advanced against
this theory—1) Demonstrable antecedents for secretion granules
are not found in many other serous and mucous cells. 2) The
secretion may arise directly from other cytoplasmic structures
independent of mitochondria. 3) If secretion granules originate
from mitochondria one would expect to find variations in quan-
tity depending upon the secretory stages of the cells. 4) The
universal distribution of mitochondria in all cells speaks against
a specificity in gland cells. Much light on this subject would
result no doubt from both embryological and comparative study
of glands aided by certain pharmacodynamic reactions.
THE LACHRYMAL GLAND 209
VII. BASAL STRIATIONS
Whether such structures occur in the lachrymal gland as were
early described by R. Heidenhain in fresh gland tissue and by J.
Miiller and Pfliiger especially in the salivary ducts, has occasioned
much discussion. Boll described “Tréinenréhren’ in the lachrymal
glands studied by him. Maziarski found none in man. Merkel
found none in the lachrymal gland of the dog. Garnier, on the
other hand, describes them in this gland in the dog and cat
while Zimmermann claims that basal striations in the cells of the
human lachrymal gland correspond to a lamellar structure.
Fleischer, Hornickel, and Puglisi-Allegra have also recorded
indications of the presence of such striatious or lamellar
structure.
Thus we find much difference of opinion regarding the nature
and presence of these structures. Even investigators working
on the same gland have disagreed regarding them. In some con-
tributions it is often difficult to determine just what particular
cells are said to possess them. ‘To Bensley (711) we are especially
indebted for clearing up the situation. He has shown that the
so termed striations are due to two distinctly different substances
—a) mitochondrial filaments of Altmann and Michaelis, and b)
basal filaments of Solger and others.’ The former are seen in
fresh tissue, are stained vitally by Janus green, can be demon-
strated by fixing tissues in.acetic osmic bichromate solution and
staining by Bensley’s acid fuchsin methyl green method in which
these filaments are fuchsinophilic in reaction, and are readily
destroyed in solutions containing much acetic acid. The basal
filaments, on the other hand, are due ‘‘to the fact that there
are in the cell (speaking of his pancreatic acini fixed in chrome
sublimate and stained in toluidin blue) unstained areas shaped
like the filaments observed in the fresh cell after staining with
janus green. These are the spaces originally occupied by the
mitochondrial filaments.’’ The basal filaments, then, are inter-
mitochondrial basophile substance. In preparations fixed in
solutions containing sufficient amounts of acetic acid to destroy
the mitochondria this basophile substance is broken up into a
feltwork of fine filaments. ‘‘These are the familiar basal fila-
210 JOHN SUNDWALL
ments of Solger or the ergastoplasmic filaments of Prenant,
Garnier, Bouin.’’ Bensley is inclined to the opinion that these
basophile filaments are fixation artifacts. At least he did not
see them in the living cells.
Unfortunately many investigators have failed to take cogni-
zance of this valuable contribution of Bensley’s. Since its publi-
cation many still. are laboring under the old confusion. Even
Hoven (712) speaks of ‘‘vegetativen Ergastidien und Chondrio-
men” as the same structures. Champy likewise fails to ap-
preciate their difference—‘‘ Mitochondria und Ergastoplasma
scheinen eine einzige und die gleiche Formation darzustellen.”’
Basal striations then are due to (a) Mitochondria which are
either rod- or thread-like structures or granules so arranged as
to form rows (Chondriomiten). The long axis of the former
and the rows of the latter are generally arranged parallel to each
other and perpendicular to the base of the cells. (b) An inter-
mitochondrial cytoplasm which may be basophile in reaction
and depends upon the existence of the former (a). When the
mitochdonria are not preserved or remain unstained the latter
(b) is prominent.
The presence, shape, and arrangement of mitochondria, then,
should determine to a great extent whether basal striations are
present. It will be recalled that in the discussion of mitochondria
these structures in the cells of the secreting tubules and inter-
calary ducts are more or less spherical or very short rods and
irregularly distributed throughout the cytoplasm. No parallel
rods or rows were seen. Further, no deeply staining basophil
substance was observed. Again the secretion granules filled
the entire cell. These conditions exclude the possibility of basal
striations being present. I have not observed any semblance
of such in the tubules or intercalary ducts in any of my prep-
arations fixed in at least ten different fixing solutions.
The intralobular and interlobular ducts on the other hand
possess an arrangement of mitochondria necessary to form basal
filaments. However, no intermitochondrial basophil substance
is present. It was not until I had availed myself of Bensley’s
method that I was convinced that basal striations of any kind
THE LACHRYMAL GLAND Zt
are present in these cells. In most fixations the basal half of
the cell appears in no wise different from the proximal half.
I could not account then for the claims of Fleischer who states
that he saw ‘pinselartige Aufaserung’ in the cells of these ducts,
ox—5 per cent ammonium chromate solution; Hornickel, beim
Esel, and others. There is no doubt in my mind now but that
the structures seen by these observers were imperfectly fixed
mitochondria, which are readily seen in the acetic osmic bichro-
mate-anilin fuchsin methyl green, and alcohol sublimate chro-
mate-copper chrome haematoxylin preparations.
VE Ad
Axenfeld states that the presence of fat in the secreting ele-
ments of the human lachrymal gland is a normal condition and
that apparently it is related to the secretory processes. Many
globules of fat are found in the small dark cells. Kirschstein
speaks of the large amount of fat in the interstitial connective
tissue of the human lachrymal gland of old people. Ellenberger
states that globules of fat are normal constituents of the lachrymal
glands of all domestic animals except the pig, as does Lutz
(99). Staneuléanu and Théorhari describe much fat in the
human lachrymal gland after epiphora. This condition prob-
ably was one of fatty degeneration. Schirmer—‘‘Jedenfalls
ist es von Wichtigkeit zu betonnen, dass in ganz normalen Driisen
grossere mengen Fett sich finden kénnen.’’ Hornickel—‘‘ Das
Fett tritt als konstanter Zelleinschluss bei allen Tieren auf.”’
According to him, in the horse and pig, the droplets of fat are
small and in the periphery of the cell. In ruminants the dis-
tribution is regular, in the dog there is much variation, and in the
eat the least amount.
More recent studies regarding the distribution of fat and Lipoid
substances in other tissues have been made by Bell (10) and
Kingsbury (11). The former shows that in fresh tissues Herx-
heimer’s stain affects all fat globules as well as the liposomes.
According to him tissues fixed in formalin have a tendency to
lose the liposomes while the neutral fat globules are unaffected.
He shows that the granules obtained by Albrecht (’02) in the
ZAZ JOHN SUNDWALL
kidney after staining with neutral red are not lipoids. Kings-
bury used osmic acid and Weigert myelin sheath method.
For my purposes I found that Herxheimer’s stain (Technique
III, 5) served best notwithstanding that it stains lecithin, cho-
lesterin, and myelin. This however did not materially affect
the results. Osmic acid was also used and proved satisfactory
but with this fixation it is very difficult to differentiate between
the finer lipoids and other structures within the cell. Numerous
glands were examined and it was found that the fat content of
the cells is subject to much variation. As was pointed out by
Bell, fresh glands show more of the finer liposomes than do those
after preservation for some time in formalin.
In some glands numerous droplets of fat were present in all
the epithelial elements. The sizes of these varied from some as
large as the nucleus to others so fine that they could barely be
made out. In these glands the larger fat goblets were generally
seen in the base of the cell while the finer ones were located more
in the body and proximal portions. There was, however, much
irregularity in the distribution. In other glands again the
amount of fat present within the epithelial cells was limited to
very fine globules irregularly distributed throughout the epithel-
ial elements. An abundance of interstitial fat is always seen
both in the capsule and the trabeculae.
In comparing these preparations with similar ones of other
glands (pancreas and submaxillary gland of the same animals)
the lachrymal gland in the majority of those examined showed
very little more fat, if any, in the epithelial cells than did the
others. My conclusions then are that frequently the lachrymal
gland cells show considerable fat globules but that as a rule the
cells in this gland do not possess more than are ordinarily seen
in other secreting glands, and that where numerous globules are
seen it should be considered in the nature of pathological
degeneration.
Frequently the technique of preparing tissues by frozen
method involves, by many, the infiltration with gum arabic.
The latter stains red in Herxheimer’s solution and unless it is
completely removed it may simulate fat globules. I was misled
by this at first.
THE LACHRYMAL GLAND pees}
IX. SECRETION CAPILLARIES AND CEMENT SUBSTANCE
Injection of various masses into the duct system of glands
was an early method employed for the demonstration of secretion
capillaries. Langerhans, Saviotti, Gianiozzi, Pfliiger, Ewald,
and Boll, among others, utilized this means. Objections, how-
ever, to this method were advanced by some observers in that
the injection mass under pressure produced these capillaries.
It was found that Golgi’s silver impregnation method demons-
strated the secretion capillaries very clearly, as a black deposit
is formed within them. Cajal, Retzius, E. Miller, Langendorf,
Laserstein (94), employed this means.
While intercellular secretion capillaries for serous glands in
general are readily demonstrated by these methods as well as
with certain stains much controversy has existed regarding their
terminations. Some claim that they enter the cell, intracellular,
as in the ease of bile canaliculi, while others maintain that they
remain intercellular throughout their course.
Noll described intercellular secretion capillaries in the lachry-
mal gland of the tat and Fleischer describes them in the tubules
of the lachrymal gland of the ox, of varying lengths—some mere
depressions, others almost touching the basal membrane, branch-
ing. Hornickel agrees with Fleischer regarding the presence
and relations of these capillaries in the lachrymal glands of
domestic animals. He further states that none are present in
this gland in the pig and dog and explains this condition as being
due to the fact that these glands are mucous in character as
they react to mucous stains. Puglisi-Allegra demonstrated these
structures by Golgi’s method.
For the study of secretion capillaries I used the following:
Kopsch-Golgi method (Technique II, 4); vital staining with
pyronin (VI, 1); and tissues fixed in sublimate alcohol bichromate
solution (IV, 2, a) or formalin bichromate sublimate solution
(IV, 2, c) stained with iron haematoxylin, copper chrome haema-
toxylin, neutral gentian, and neutral safranin (V, 3, 5, 6, 7).
My results agree with those of Fleischer and Hornickel.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 2
214 JOHN SUNDWALL
Kopsch-Golgi
In Kopsch-Golgi preparations the lumina of the intercalary
ducts and tubules and the secretion capillaries are definitely
marked by the heavy black deposit, which gives an exact cast of
them (fig. 16.)
In the tubules numerous secretory capillaries are seen radiat-
ing from the central irregular lumina. These vary in length,
some forming only slight indentations between the cells while
others almost touch the basement membrane. The ends of
Fig. 16 Ikopsch-Golgi silver impregnation, Tech. II, 4. Zeiss comp. 18, obj.
AA—showing secretion capillaries, lumina of tubules and ducts.
these capillaries are rounded and in many instances terminate
in sight bulb-like swellings or knobs. Many branch at irregular
distances from their origin. Some branch immediately upon
leaving the lumina—others, after coursing some distance be-
tween two cells, branch to straddle the small demilune eells.
In many instances the terminal branches are so numerous that
they form irregular rosettes with many short club-like rays pro-
jecting in every direction. These rays frequently appear to be
THE LACHRYMAL GLAND PANNE)
intracellular although it is impossible to demonstrate by this
method that such is the case.
The lumina of the interecalary ducts appear either smooth and
even or many short knot-like projections are seen on either side
showing that the surface cells which form the lumina are either
in close apposition to each other or have indentations between
them. That both conditions may be present has been considered
in ‘* Duet System and Tubules.”’
Fig. 17 Tubule and intercalary duct after vital staining with pyronin.
Tech. VI, 1. A, secretion capillaries; B, lumen of tubule; C, lumen of interecalary
duct.
Pyronin
The most satisfactory method for the study of secretion
capillaries is the vital staining with pyronin. The capillaries
as well as the lumina of the tubules and ducts stain a deep red
while the cells are unstained (fig. 17). The advantage of this
method lies in the fact that very thin sections from 2 to 5u can
be obtained and these can be further stained with either iron
haematoxylin or copper chrome haematoxylin. Thereby the re-
lation of the capillaries to the other structures of the cells can
be readily made out.
216 JOHN SUNDWALL
In sections thus prepared the red stained secretion capillaries
are seen between the cells. Frequently they become finer and
finer and gradually disappear as they approach the basal ends
rie. IS) Tubule showing secretion capillaries and cement lines. Secretion
granules are not preserved. The intergranular cytoplasmic network is seen.
of the cells. Others terminate in bulb-hke swellings. In no
instance was I convinced that they had an intracellular
termination.
bo
—
ca |
THE LACHRYMAL GLAND
Fixed and stained preparations
Secretion capillaries are readily seen in tissues stained with
the serous granule stains already named (figs. 18 and 19). Here
they are seen as intercellular structures and are outlined bya
thin deeply stained cement substance. These sections plainly
show that the secretion capillaries are much modified by the
stages of granule formation within the cells. If bordering cells
are in a Maximum stage the capillaries are not seen, only the
eement line projects downward between these cells. On the
other hand, if the cells are in a medium or minimum granular
stage the capillaries are wide and open and outlined by the cement
substance.
In none of my preparations have I observed true intracellular
secretion capillaries, notwithstanding they have been described
by others. It might be well to add here that the literature on
this particular subject is very confusing. Under the caption
of intracellular canaliculi, one finds at least three separate and
distinet types of these canals described without any particular
discrimination as to their nature—(a) intracellular secretion
‘canaliculi as described by E. Miller and R. Krause; (b) intra-
cellular blood capillaries, in liver cells, as shown by Schifer;
and (c) the canalicular apparatus (reticular apparatus) (Holm-
gren’s canals) which is found in most animal and vegetable
cells. The adoption of a uniform nomenclature is certainly
desirable in this instance. It is interesting to observe that
similar methods (silver impregnation) have been used for the
demonstration of all of these and is it not possible that in many
instances intracellular secretion capillaries have been described
when the canalicular apparatus of Holmgren alone was present ?
The latter is discussed under X.
The cement substance of glands was first described by Heiden-
hain and later by Zimmermann, Bonnet, Cohn, Solger, Carlier,
Meyer, Oppel, Bensley, and others. Kolossow, on the other
hand, doubted its existence.
I have already described the cement structures in the larger
ducts (VI. The Duct System and Tubules.) Cement lines are
218 JOHN SUNDWALL
readily demonstrated with iron haematoxylin, copper chrome
haematoxylin, neutral gentian, and neutral safranin. In the
tubules and interecalary ducts they appear either outlining the
lumina and the secretion capillaries or as lines varying in length
projecting basalwards between the cells (fig. 19). As already
Fig. 19 Tubules, Tech. IV,2,a;V,6. Zeiss, oc. 2, Leitz. obj. 7s oil—showing
secretion capillaries and cement substance. Note varying intensities of stain-
ing of the granules. A, deeply stained granules; B, faintly stained granules;
C, secretion capillaries.
Fig. 20 Intralobular duct, cross section, tech. same as figure 19. Cement
lines projecting towards the base of the cells and between them—A.
stated, the cement substance is much modified in appearance
by the secretory stage of the cells. If the cells are in a maximum
stage of granule formation the secretion capillaries may be
completely obliterated, in which case the cement substance
likewise may be completely obscured or may be seen as fine
lines extending between the cells towards their basal membrane.
THE LACHRYMAL GLAND 219
As Carlier ('99) and Bensley (02) have shown in gastric epithe-
lium, the cement substance frequently appears on the free sur-
face of the cell as a fine irregular network.
X. CANALICULAR APPARATUS
It has been shown that many types of cells possess within
their cytoplasm very fine anastomosing canaliculi. These were
first described by Golgi (98) in the nerve cells. Later his
students, Negri (00) among others, demonstrated them in
many gland cells. Golgi’s silver impregnation method was
used for the demonstration of these minute canals.
Kopsch (02) showed that they could be demonstrated by
long immersion of the tissue in 2 per cent osmiec acid. By this
method von Bergen (04) demonstrated the canaliculi in a great
many animal cells.
Holmgren (02), in a number of papers beginning with 1899,
has contributed much to this particular subject, using picric
acid-sublimate, toluidin blue and erythrosin method, and later
trichloracetic acid and fresh Weigert’s resorein fuchsin. At
first Holmgren thought that the canals were lymphaticin nature
as he had demonstrated to his own satisfaction that they com-
municated with the exterior. Later, mainly as a result of his
second technique, he held that these canals Gin nerve cells)
possessed a network of fibres which had their origin from the
nerve capsule. To this network of anastomosing fibres he
applied the term spongioplasma and from this developed the
well known ‘“Trophospongium theory.’
Bensley (10 b) has shown the analogy of these canaliculi to
the vacuoles of plant cells. In the latter (root tip of onion)
using formalin bichromate sublimate solution and Kopsch
solution as fixatives, he has demonstrated that in the youngest
cells the well known vacuoles appear as a canalicular system
similar to that in the characteristic animal cell. As the plant
cell becomes older the canals enlarge and finally form vacuoles.
These were seen by him in the living plant cells, as well. In no
instance did he find these canals communicating with the exte-
rior. Bensley (’11) describes the canaliculi in both the acinus
220 JOHN SUNDWALL
cells and islet cells of the pancreas. In the former, after stimu-
lation in order to rid the cell of the secretion granules which hide
the canalicull, they are seen—‘‘the apparatus is located in the
portion of the cell between the nucleus and the lumen but
branches of the canal may appear basalwards along the sides of
the nucleus.’ In the islet cells they resemble those in the
acinus both in topography and relation. In neither were com-
munications with the exterior seen.
Cowdry (12 b) deseribes them in the spinal ganglion cells of
the pigeon.
That these canaliculi communicate with the exterior of the
cell has been claimed by Holmgren and Retzius, among others;
that they do not, by Golgi, Kopseh, Misch, Studnicka, and
Bensley.
The canalicular apparatus in the cells of the tubules of the
lachrymal gland of the ox can not be easily studied owing to the
general presence of secretion granules. However, in many
glands cells with few or no granules were seen in which this
apparatus could be made out readily. It was seen for the most
part in that part of the cytoplasm proximal to the nucleus. It
appears as a network of open spaces or canals with branches
which frequently terminate in shght nodular enlargements. A
layer of eytoplasm was always seen between these terminations
and the cell membrane. The apparatus at one end comes into
close proximity to the nucleus, and processes are frequently
seen extending around the nucleus towards the base of the cell.
This apparatus is seen best in sections preserved in formalin
Zenker’s solution, where it is seen as clear spacesin the cytoplasm.
In the tissue stained by the vital pyronin method, which demon-
strates the intercellular secretion capillaries, the canalicular
apparatus of Homlgren can be made out as well. Here also it
appears unstained, in contrast to the red secretion capillaries,
and is seen as clear spaces in the cell. Consequently it is readily
differentiated from the intercellular structures. Since the
pyronin stains the secretion substance, one can readily conclude
that the apparatus is In no manner concerned with the secretion
substance of the cell and is independent of the secretion
capillaries.
THE LACHRYMAL GLAND 221
XI. TECHNIQUE
I. For the study of the gross characteristics of the gland,
calves’ heads were embalmed by injecting, under pressure, into
the carotid arteries equal parts of glycerine, 95 per cent alcohol,
and ecarbolie acid full strength. A suspension of red lead, starch
and hot water was then injected, shortly after which dissection
of the orbit was begun.
II. For the study of the ducts and their ramifications the
following methods were used:
1. Flint (02), Spalteholz (97).
2. Injection method, see Hiiber’s technique given in American
Journal of Anatomy, vol. 6, 1907, The Arteriolae Rectae of the
Mammalhan Kidney. The celluloid mass withstood the corrosion
much better than the celloidin. Total peptic digestion required
on the average two weeks.
3. Vital staining method with pyronin (See VI, 1).
4. Kopseh (710)—Golgi chrome silver method (somewhat
modified) for the demonstration of Jumina of tubules and secre-
tion capillaries.
Ill. For the study of the connective tissue framework.
1. Flint, Spalteholz—After complete digestion, tissues were
imbedded in paraffin and thin sections made.
2. Mall’s method (96) for the demonstration of reticulum with
frozen sections (25pz).
3. Tissues fixed in 70 per cent alcohol, imbedded in celloidin,
sections 10 to 20u thick, fastened on slides, and digested and
stained by Mall’s method were especially useful for the study
of the framework. The method (Jeffrey’s) was as follows:
Sections were placed in a mixture of equal parts of glycerine and
95 per cent alcohol for some time before mounting. A thin
coating of Mayer’s albumin was applied to clean slide, to which
section was transferred. A piece of smooth writing paper was
placed on the section and over this several layers of blotting
paper. A second slide was prepared in the same manner and
the two were bound tightly together with the blotting paper in
the middle. This was placed ina thermostat 60° C. for one-half
hour after which the slides freed from the paper. The slides
222 JOHN SUNDWALL
(to which the sections adhered) were then placed in aleohol and
ether until the celloidin was dissolved (15 to 20 minutes) ; trans-
ferred to absolute alcohol, 95 per cent alcohol, 70 per cent alco-
hol, and finally to water; and digested in pancreatin, which
required from one to three days.
Tissues fixed in any of the chrome salts were not affected by
pancreatin in one week’s time.
4. For further study of the connective tissue, tissues were
fixed in Zenker’s solution, embedded in celloidin and stained as
follows: (a) for collagenic fibres and smooth muscle—Van
Gieson’s, Mallory’s and haematoxylin and eosin; (b) for elastic
fibres—Weigert’s and Unna-Taenzer (710).
5. For the study of fat in connective tissue, pieces of gland
were fixed in 10 per cent formalin twenty-four hours, washed,
cut by frozen method, and stained by Herxheimer’s (’10) method
—absolute alcohol 70 ec., sodium hydroxide (10 per cent solution)
20 cc., water 10 cc., Sharlach R. to saturation. After staining
for from five to ten minutes sections were washed in 70 per cent
alcohol, washed in water, and mounted in glycerin.
IV. For the study of secretion granules and the finer histologi-
eal characteristics of the cell.
1. Granules in the intercalary duects—Fixed in Zenker’s,
embedded in celloidin, fastened to slides (see III, 3); celloidin
dissolved off; and section stained in muchaematein or mucicarmin
prepared according to Bensley’s method (see Stains).
2. Granules in the tubules as well as those in the intercalary
ducts and the minute cell structures—Small pieces of tissues
were fixed in the following solutions, embedded in paraffin, and
cut from 2 to 5 u thick.
(a) Bensley’s (96) sublimate alcohol bichromate solution—
equal parts of saturated solution HgCl. in 95 per cent alcohol and
3 per cent aqueous solution of K,Cr.O;. Small pieces of the
tissue were placed in this fixative for about three hours. The
solutions were not mixed until time of using. Fresh mixtures
were used every thirty minutes during fixation.
(b) Modification of Kopsch’s formalin bichromate solution—
THE LACHRYMAL GLAND Papi
Formalin 40 per cent, 1 part...
heCrO;S percent Agosol. 3 parts “9 one Uae
Nab.-seloot HgGisan'95 per cent alcohol. .. 2.0.6... coe ee one part
Dial eC nWAet ee eRe Poe Sia aly ec Beak eM eR ae Portes te two parts
Fixative was made at time of using. Small pieces of tissue were
placed in it for about three hours to insure proper fixation.
(ec) Formalin bichromate sublimate method—Fixed for twenty-
four hours in the following solution: neutral formalin 10 ec. and
Zenker’s solution without the acetic acid 90 ee.
(d) 70 per cent alcohol.
Tissues fixed in (a), (b), (¢), and (d) were stained in muchae-
matein, mucicarmin, iron haematoxylin counterstained with
mucicarmin, Bensley’s (11) neutral gentian, Bensley’s (’11)
safranin-acid violet, and copper chrome haematoxylin used
singly or with mucicarmin (see Stains).
(e) Acetic osmic bichromate method (Bensley ’11, p. 308).
V. Staining methods.
1. Muchaematein (Bensley ’03)—haematein 1 gram, aluminum
chloride 0.5 gram, 70 per cent alcohol 100 ce. Haematein and
chloride rubbed together, dissolved in alcohol and allowed to
stand for a week to insure ripeness. (If aleohol is made by
diluting absolute alcohol with tap water stain can be used im-
mediately.) Sections were flooded with stain, placed on stage
under microscope and watched until deep color, then rapidly
washed in 95 per cent alcohol, dehydrated, cleared and mounted
in xylol balsam. The granules in the interealary duct stained
definitely and intensely blue.
2. Mucicarmin—1 gram of carmin and 0.5 gram aluminum
chloride ground in porcelain evaporating dish after which small
amount of water was added; heated over a Bunsen burner,
grinding process continued during heating, till mass became very
‘dark red (almost black) then dissolved in absolute alcohol and
filtered. Stain must be used while fresh. Technique same as
Nid.
3. Iron haematoxylin.
4. Iron haematoxylin counterstained with mucicarmin.
5. Copper chrome haematoxylin (Bensley 711, p. 310).
224 JOHN SUNDWALL
6. Neutral gentian—solution of gentian-violet (erystal violet)
precipitated by its equivalent of orange-G solution (Bensley
14, ps S08):
7. Safranin-acid violet—precipitate of a saturated solution of
safranin O with solution of acid violet (Bensley’s 11, p. 309).
8. Acid fuchsin methyl green—(a) Altmann’s acid fuchsin
anilin solution: acid fuchsin 20 grams, anilin water 100 ec. and
(b) 1 per cent solution methyl green (Bensley 7°11, p. 309).
9. Macallum’s (’95) iron reaction.
VI. Vital staining methods.
1. Pyronin. (Bensley 711, p. 305.) About eight liters of
1—1000 solution in isotoni¢ salt solution injected into the carotid
arteries of heads of freshly killed calves; cut arteries in neck
clamped off to prevent leakage. Lachrymal gland was deeply
stained. By means of the Valentine knife sections 0.5 to 1 mm.
were cut and studied with binocular. The acini or tubules
stained only lightly while the ducts, as well as the lumina of the
acini including the intercellular secretion capillaries, were deeply
stained. The capillaries were best studied by fixing small
pieces of this pyronin stained gland in 8 per cent solution of ice
cold ammonium molybdate for twelve to twenty-four hours,
after which they were placed in ice cold 95 per cent alcohol
one hour, absolute alcohol one hour, toluol one hour, and paraf-
fin one-half hour. Seetions were made and capillaries studied
with the microscope.
2. Janus green—one gram in 15,000 ec. of isotonic salt solution.
(Bensley 711, p. 305.) Technique same as VI, 1.
3. Methylene blue—one gram in 10,000 ce. of isotonic salt
solution. Technique same as VI, 1.
4. Neutral red—one gram in 15,000 ce. of isotonic salt solution.
Technique same as VI, 1. This stain cannot be fixed.
For detailed consideration of these stains see Bensley (711).
XII. MISCELLANEOUS-OTHER FIXATIONS AND STAINS
In addition to the technique outlined and referred to in the
general discussion other methods of fixation and staining were
employed, the results of which may be briefly summarized as
follows:
THE LACHRYMAL GLAND 225
1) Fixations with alcohol failed to preserve the secretory
granules and the mitochondria. The cells of the tubules show
only the intergranular cytoplasm which appears as a network.
2) When 5 per cent glacial acetic acid was added to the 70
per cent alcohol it was found that the granules were partially
preserved. In these preparations many of the ring and crescent
granules of Fleischer were seen.
3) The same results outlined in (1) are obtained when Car-
noy’s fixation is used.
4) The secretion granules are fairly well preserved in picric
acid. When stained in the neutral stains of Bensley, one is
impressed with the varying intensities to which granules stain.
One cell may be filled with the characteristic dark blue granules
while the neighboring cell may be full of lightly stained yellow-
ish granules. The granules in cell type 1 generally show in this
fixation greater affinity for the neutral stains than do those in
other cells.
5) Tissues fixed in trichloracetic acid preserved the secretion
granules fairly well. However, they did not show great affinity
for the neutral stains, for after slhght differentiation they are
but faintly stained.
It will be recalled that fixations (4) and (5) were used by
Fleischer in the demonstration of ring and crescent granules.
While these were frequently seen in these fixations they were
by no means constant.
6) Using Kolossow’s method, I was unable to demonstrate the
epithelial intercellular bridges as described by him. Much
shrinkage of tissue was in evidence.
My experiments with neutral red, vital staining, were not
very satisfactory. Four different calves heads were stained by
this method and in each instance the gland remained either
uncolored or appeared slightly pink. This was found to be in
marked contrast to the results obtained in other animals where
the gland stained deeply red (monkey). Microscopic exami-
nations show that all cells are diffusely but faintly stained. The
granules of the tubules and intercalary ducts also appear
very faintly stained. As elsewhere stated, careful examination
226 JOHN SUNDWALL
failed to reveal prozymogen granules as described by Bensley
for the pancreas.
Vital staining with methylin blue showed the sympathetic
nerve fibres of the gland, as demonstrated by Dogiel. The
granules are not stained.
XIII. SUMMARY
1. The lachrymal gland in Bovidae is made up of two parts—
a Pars superior which comprises by far the greater bulk of the
gland and an. auricular appendage which extends downwards
between the Bulbus oculi and the outer bony orbital wall—the
Pars inferior.
2. The gland is composed of a series of from six to eight com-
pound tubular glands serially arranged and in close apposition
to each other. Each gland may be subdivided, beginning with
the terminal opening and proceeding to the secretory elements,
into the following structures: main duct, primary duct, inter-
lobular duct, intralobular duet, interealary duct, tubule or
alveolus.
3. Elastic fibres are abundant in the capsule and in the larger
interlobular septa. They appear to some extent in the more
prominent intralobular septa but in the finer they are not seen.
These fibres do not surround the individual secreting tubules
as claimed by Boll, Schirmer, Fumagalli, and others, in various
animals. Lymph cell infiltrations of the septa are not normal
conditions though so held by Schirmer for this gland in man
and by Fleischer in the ox. The basement membrane is formed
by a reticular connective tissue similar to that seen in the sali-
vary gland. In addition to this membrane, irregular, anasto-
mosing connective tissue cells are seen between and surrounding
the tubules together with plasma cells, endothelial cells, and
lymph cells. While smooth muscle cells are frequently seen
occurring singly in the capsule and in the largest interlobular
septa, they are not found within the lobule surrounding the
individual tubules as held by Kollosow, Zimmermann and others.
4, The epithelial cells of the main duct are irregular in outline
and arranged in several layers. Numerous goblet cells are seen.
THE LACHRYMAL GLAND 227
In the primary ducts a gradual reduction of both goblet cells
and the layers of epithelial cells takes place. The interlobular
ducts possess no goblet cells, and two layers of epithelial cells
form the lumina. In the intralobular ducts one or two layers
of epithelial cells are seen. The outer layer gradually disappears
in the smaller intralobular ducts and in the intercalary ducts.
5. Secretion granules are present in both the intercalary duct
and in the tubules. They are not found in the other ducts.
The granules are readily seen in all fresh glands when examined
in serum or isotonic salt solution. They disappear when the
fresh cells are placed in distilled water but reappear upon the
addition of 2 per cent sodium chloride solution.
6. The granules in the interecalary ducts are preserved when
the tissue is fixed in Zenker’s solution and stain specifically in
muchaematein and mucicarmin. They are not stained—when
fixed in Zenker’s —with the serous granule stains. The granules
in the tubules are not preserved as a rule in Zenker’s solution.
This phenomenon suggests that the cells or the tubules may
differ in function from those of the intercalary duct. I am not
prepared, however, to make this claim.
7. The granules of both the interealary duct and the tubules
are fixed in Bensley’s alcohol sublimate bichromate solution and
in formalin bichromate sublimate solution. When tissues are
fixed in these solutions the granules in both the tubules and
intercalary ducts stain in the well known mucous stain—muc-
haematein and mucicarmin as well as in the serious granule stains
iron haematoxylin, copper chrome haematoxylin, neutral gen-
tian, neutral safranin. After staining tissues fixed in the former
solution with iron haematoxylin and then counterstaining with
mucicarmin, all the granules in certain cells stain black with the
former stain and all the granules in other cells stain red in the
latter, while in other cells both black and red granules are seen—
thus showing that even within the same cell some of the secretion
granules are affected by serous stains while others are affected
by mucous stains. Notwithstanding this double staining re-
action, there is not sufficient evidence to claim that the cells
forming the tubules are heterogeneous in character.
228 JOHN SUNDWALL
8. The cells constituting the tubules or acini present different
pictures depending upon the secretory stage they are in. In the
maximum granular stage the cells are large spherical or oval
(bulging), the granules fill the entire cell and the nucleus is
flattened against the base. When the secreting cavity is made
up of these cells it may have the form of an alveolus or acinus.
Cells in a medium or minimum granular stage are cylindrical,
pyramidal, hour-glass shaped, or may appear as crescents. As
a rule they are seen compressed between the large bulging cells
in each tubule. They are frequently seen, however, making
up the entire secreting cavity. These cells are much reduced
in size when compared with the former. Granules may fill the
entire cell or only few may be present. Occasionally these cells
show no granules. The nucleus is round or oval and is always
separated from the base of the cell by a zone of cytoplasm.
9. The cells of the interecalary duct do not show such dis-
turbances. Whether these cells are in a maximum or minimum
granular stage, the size and shape remain practically the same.
No changes are seen in the form or position of the nucleus.
10. The lachrymal gland of the ox is not a mucous secreting
gland notwithstanding that the granules stain specifically in the
mucous stains. I do not agree with other observers, namely
Fleischer, that the secretion granules in the process of formation
regularly assume various peculiar forms such as rings and demi-
lunes. Rings and demilunes are sometimes seen in fixed prep-
arations but these peculiar forms are no doubt due to the action
of the fixation fluid.
11. No light has been gained as a result of my studies on the
origin of secretion granules. The absence or at least the pres-
ence of very small amounts of demonstrable secretion granule
antecedent substances—prozymogen, nuclear derivatives, ete.—
does not admit of the hypothesis that the nucleus is the sole
originator of secretion granules. On the other hand the abun-
dance of granules, their variation in size without regard to position
in the cell (the largest and smallest granules are seen side by
side in any portion of the cell), their apparent origin from the
evtoplasm—all strongly suggest that the cytoplasm plays a
THE LACHRYMAL GLAND 229
very unportant role in the formation of granules. I have seen
no evidence that mitochondria are diretly concerned in the pro-
duction of secretion granules.
12. Mitochondria are abundantly present in the cells of the
larger duets—intralobular, interlobular, ete.—where they are
either arranged as filaments or appear irregularly distributed
without any special arrangement whatsoever. In the cells of
the interealary duct and secreting tubules they are much less
numerous and are distributed irregularly throughout the eyto-
plasm. The number of mitochondria for these cells appear
approximately equal—whether the cells are a maximum granular
stage or contain the minimum of granules.
13. I did not find that an abundance of fat globules in the
secreting cells is characteristic of the lachrymal gland as claimed
by other investigators. While glands were studied in which
numerous fat globules were observed, in the vast majority the
quantity of fat globules was found to be no more than that
observed in the submaxillary gland and the pancreas.
14. Intercellular secretion capillaries varying in width and
length—depending upon the secretion stages of the cells—and
bounded by cement substance are prominently seen in the
tubules. In the interealary duct these capillaries may not be
seen or they may appear as slight indentations between the
cells. Cement substance is found throughout the secretion
passages. Pyronin when used intra vitam selectively stains the
secretion capillaries.
15. The canalicular apparatus is readily seen in the cells of
the lachrymal glands. In the cells of the tubules they are only
seen when few or no granules are present.
16. I have failed to observe the constant presence of centro-
somes deseribed by Zimmermann and Fleischer. Likewise I
have failed to find the constant presence of the paranucleus
described by other investigators.
I am especially indebted to Professor R. R. Bensley of the
University of Chicago for numerous suggestions. This work
was begun in his laboratory under his inspiring direction. Since
230 JOHN SUNDWALL
then I have taken advantage of his numerous contributions to
the knowledge of gland structures, and the new methods of tech-
nique devised by him have been of greatest value in my work.
BIBLIOGRAPHY
ALBRECHT, FE. 1902 Neue Beitrige zur Pathologie der Zelle. Deutche Path.
Gesell.
Aut, ApotpH 1900 Trans. Acad. Se., St. Louis, vol. 10, pp. 185-207. Quoted
from Schwalbe’s Jahresb.
ALTMANN 1890 Die Elementarorganismen und ihre Beziehungen zu den Zellen.
_ Leipsig.
ARNOLD, G. 1911 The role of chondriosomes in the cells of the guinea pig’s
pancreas. Arch. f. Zellforsch., Bd. 8, pp. 252-271. Quoted from
Schmidt’s Jahresb., 47 Jahrgang, B. f. d. 1912, Berlin, 1913, p. 67.
AXENFELD, TH. 1900 Ueber die feinere Histologie der Trinendriise. besonders
tuber das Vorkommen von Fett in den Epithelien. Bericht tiber die
28 Vers der Ophth. Gesell., Heidelburg, 8. 160. Quoted from Ellen-
berger.
See also—Bemerkungen zur Physiologie und Histologie der Trinen-
driise. Bericht d. Ophth. Gesell. zu Heidelberg.
Betu, E. T. 1910 The staining of fat in epithelium and muscle fibres. Anat.
Recs, vol. 4, noy5, 5p. 199:
Brenstey, R. R. 1896 The histology and physiology of the gastric glands.
Proceed. of the Canadian Institute, p. 11.
1902 The cardiac glands of mammals. Am. Jour. Anat., vol. 2, No. 1,
pp. 105-156.
1903. The structure of the glands of Brunner. The Dee. Pub. Uni-
versity of Chicago, vol. 10, pp. 1-40.
1908 Observations on the salivary glands of mammals. Anat. Rec.,
vol. 2, pp. 105-107.
1910 a On the so-called Altmann granules in normal and pathological
tissues. Trans. Chicago Path. Soc.
1910 b On the nature of the canalicular apparatus of animal cells.
Biol. Bull., vol. 19, 3, p. 179.
1911 Studies on the pancreas of the guinea pig. Am. Jour. Anat.,
voli 12) no. 3, p: 298:
vON BercEN, F. 1904 Zur Kenntniss gewisser Strukturbilder (Netz-apparate,
Saftkanilechen, Trophospongien) im Potoplasma verschiedener Zel-
lenarten. Arch. f. Mik., Anat. Bonn, vol. 64,'p. 498.
Bernarp 1856 Mémoire sur le pancréas et le réle du sac pancréatique. Paris.
BIEDERMANN 1886 Zur Histologie und Physiologie der Schleimsecretion.
Wiener Sitzungs Berichte, 3, Abth.
Boum AND Daviporr 1903 Lehrbuch der Histologie des Menschen.
1900 ibid, Huber’s edition, p. 482.
Bout, Franz 1868 Ueber den Bau der Tranendriise. Arch. f. Mik. Anat.,
Bd. 4, pp. 146-153.
THE LACHRYMAL GLAND 231
Boru, Franz 1871 Die Trinendriise. Strieker’s Handbuch der Lehre von den
Geweben, Leipsig, 2, p. 1161.
Casat, Ramon y 1896 Estudios histologicos sobre los tumores epiteliales.
Rev. trim. micr., T. 1, (Schifer).
Caruier, E. W. 1899 Changes that occur in some cells of the Newt’s stomach
during digestion. La Cellule, Louvain, T. 16, pp. 250-255.
Cuampy, C. 1911 Recherches sur l’absorption intestinale et le rédle des mito-
chondries dans l’absorption et la sécrétion. Arch. d’Anat. Mier., T.
12, F. 1, pp. 55-170. Quoted from Schmidt’s Jahresb. 46, Jahr 1911-12.
Cowpry, E. V. 1914a The development of the cytoplasmic constituents of the
nerve cells of the chick. JI. Mitochondria and neurofibrils. Am.
Journ. Anat., vol. 15, no. 4.
1914b The comparative distribution of mitochondria in spinal gang-
lion cells of vertebrates. Am. Jour. Anat., vol. 17, no. 1.
DantcHakorr 1905 Les cellules plasmatiques dans la glande sousmaxillaire
du lapin. C. R. de l’assve des Anat. 7, Reun Genéve.
Doatet, A. 8. 1893 Die Nervenendigen in der Thrinendriise der Siugetiere.
Arch. f. mik. Anat., Bd. 42, S. 632-647.
DrascH 1889 Beobachtungen an lebenden Driisen mit und ohne Reizung der
Nerven derselben. Arb. aus d. Physiol. Inst. in Leipsig.
Dusprevit 1907 Les glandes lacrymales des mammiferes et de l’homme. A.
Rey et Cie, Lyons.
v. EpneR 1873 Die acinésen Driisen der Zunge und ihre Beziehungen zu den
Geschmacksorganen. Graz.
1902 Kolliker’s Handbuch der Gewebelehre des Menschen, Bd. 3,
6 Auflage, Leipsig, p. 881.
ELLENBERGER, W. 1906 Handbuch der Vergleichende mikroskopischen Ana-
tomie der Haustiere, Erste Bd., p. 542; p. 545.
ELENBERGER, W. AND Baum, H. 1908 Anatomie der Haustiere, Zwélfte Auflage,
p. 935.
ELLENBERGER, W. AND GUNTHER 1888 Grundriss der Vergleichende Histologie
der Haussiugetiere, pp. 187-88.
Fatcut, F. 1905 Sur le developpement de la glande lacrymale. Arch. ital.
Biol., vol. 44, pp. 412-415.
Fiscuer, A. 1899 Fixierung, Farbung und Bau des Protoplasmas. Jena.
FietscHer, B. 1904 Beitrage zur Histologie der Trainendriise und zur Lehre
von den Secretgranule. Anat. Hefte., Bd. 26, pp. 103-166.
Fiemmine, W. 1888 Ueber den Bau und Einteilung der Driisen. Arch. f.
Anat. u. Physiol., Anat. Abt. pp. 287-302.
Furint, J. M. 1902 A new method for demonstration of the framework of glands.
Johns Hopkins Bulletin, vol. 18, p. 48.
Franck, L. 1883 Anatomie der Haustiere.
Frericus 1846 Trinensecretion. R. Wagner’s Handworterbuch, 3, pp. 617-—
631.
Frey, H. 1859 Histologie und Histochemie des Menschen. Leipsig.
FumaGatui1, ARNALDT 1897 II tessuto elastico nella glandola lacrymale dell’
uoma. Monit. Zool. Ital., Anno 8, pp. 167-69.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 2
232 JOHN SUNDWALL
GaRNIER, Co. 1900 De la Structure et du Fonctionnement des Cellules glandu-
lares Sereuses. Jour. de l’Anat. et de la Physiol., Vol. 36, p. 23.
Goual, C. 1908 Diun metodo per la facile e pronta dimostrazione dell’ apparato
reticolare interno delle cellule nervose. Boll della Societa medico-
chirugica, Pavia, Anno 22, della Societa, No. 2.
Gotz, Tu. 1908 Untersuchung von Trénendriisen aus Verschied Lebenaltern.
Inaug. Diss. Tiib.
Hannes, BertHotp 1911 Ueber das Vorkommen und die Herkunft von Plas-
mazellen der menschlichen Trinendriisen. Arch. Path. Anat. Physiol.,
Bd. 205, S. 410-417. :
Hemwennain, M. 1890 Beitrige zur Kenntnis der Topographie und Histologie
der Kloake und ihre driisigen Adnexa bei den einhaemischen Tritonen.
Arch. f. Mik. Anat., Bd. 35, pp. 173-74.
HEIDENHAIN, R. 1868 Beitrage zur Lehre von des Speichelabsonderung.
Studien d. Physiol. Inst. zu Breslau, Heft. 4, S. 1.
1883 Physiologie der Absonderungsvorginge. Hermann’s MHand-
buch der Physiologie, Bd. 5, T. 1.
Hetp, Hans 1899 Beobachtungen am tierischen Protoplasma I. Driisen
granula und Driisenprotoplasma. Arch. f. Anat. u. Physiol., Anat.
Abt,, pp. 284-311.
Hentz, J. 1873 Handbuch der sytematischen Anatomie des Menschen, Bd.
II, Handbuch der Eigeweidelehre, 2. Auf., pp. 734-745.
HERXHEIMER 1910 Enzyklopidie der Mikroskopischen . Technique Ehrlich,
Krause, Mosse, Rosin. Bd. I, p. 451.
HoumeGren, E. 1902 Neue Beitrige zur Morphologie der Zelle. Merkel and
Bonnet, Ergeb. d. Anat., Bd. 11, p. 274.
HornickEL 1906 Vergleichende Untersuchungen iiber den _histologischen
Bau der Trinendriise unserer Haussiugetiere. Internat. Monatschr.
f. Anat. u. Physiol., Bd. 23, p. 361.
Hoven, H. 1910 Sur l’histogenése du systéme nerveux peripherique chez le
poulet et sur le réle des chondriosomes dans la neuro fibrillation.
Arch. de Biol., T. 25, pp. 426-492.
191i Du rédle du chondriome dans l|’elaboration des produits de
sécrétion de la glande mammaire. Anat. Anz., Bd. 39, Nos. 11 u. 12,
S. 321-826. Quoted from Schmidt’s Jahresb. 46, 1911.
1912 Contribution a l’étude du fonctionnement des cellules glandu-
laires. Du role du chondriome dans lasécrétion. Arch. f. Zellforsch.
Rd. 8. 8. 555-611. Quoted from Schmidt’s Jahresb. 47, 1912, p. 67.
Ioannovics, G. 1899 Ueber das Vorkommen die Bedeutung und Herkunft
der Unnaschen Plasmazellen bei Verschiedenen pathologischen Pro-
zessen. Zeitschr. f. Heilk., Bd. 20.
JenDRASSIK, H. 1894 Neues Beitrag zur Frage nach der Innervation der Thrin-
endriise. Orvose. hetl. Budapest, Jg. 38, S. 209 (Ungarisch).
Kinesspury, B. F. 1911 The histological demonstration of lipoids. Anat.
Rec., vol. 5, no. 6, p. 3138.
KrrcustTEINn, F. 1894 Ueber der Trinendriise der Neugebornen und die Unter-
scheidung derselben von den Erwachsenen. Inaug. Diss., Berlin.
Quoted after Fleischer.
THE LACHRYMAL GLAND 233
Kuapp, P. 1897 Beitrige zu den Untersuchungen tiber die Innervation der
Thrinendriise. Inaug. Diss., Greifeswald.
Kxie1n, 1879 Observations on the structure of cells and nuclei. Quart. Jour.
Mic. Sc., vol. 19, p. 125.
1882 ibid., vol. 22.
Ko.tossow 1898 Eine Untersuchungsmethode des Epithelgewebes besonders
der Driisenepithelium, ete. Arch. f. Mik. Anat., Bd. 52, pp. 1-33.
Korscu, F. 1902 Die Darstellungen des Binnennetzes in Spinalen Ganglien-
zellen und andern Korperzellen mitts. Osmiumsiure. Sitz.—Ber.d. K.
Preuss. Akad. d. Wiss. Math. K1. 40, p. 929.
1910 Golgi method, Enzyplopidie der Mikroskopischen Technique,
Ehrlich, Krause, Mosse, Rosin, Bd. 1, p. 564.
Krause, R. 1898 Zur Histologie der Speicheldriisen. Die Speicheldriisen
des Igels. Arch. f. Mik. Anat., Bd., 45, p. 93.
Kiser, W. 1906 Zur Entwickelungsgeschichte der Tré#nenrohrchen. Zeit.
Augenheilk., Bd. 16, pp. 54-57.
LarrayE 1897 Recherches sur les glandes lacrymales et leur innervation.
Thése de doctorat en médecine Paris.
Lanpout, H. 1900 Ueber die Innervation der Thranendriise. Habitations-
schrift, Strassburg, 8. 34.
Lana, Paut 1911 Zur Entwickelung des Trainenausfuhrapparates beim Men-
schen. Anat. Anz., Bd. 38, pp. 561-569.
Lancer, C. V. 1890 Lehrbuch der Systematischen und Topographischen
Anatomie. Vierte Auflage, p. 647.
Lancuey, J. M. 1879 On the changes in serous glands during secretion. Jour.
of Physiol., vol. 2, pp. 261-281.
1889 On histology of mucous salivary glands. Jour. of Physiol., p.
433.
LASERSTEIN 1894 Ueber die Anfinge der Absonderungswege in den Speichel-
driisen und im Pankreus. Arch. f. die ges. Physiol., Bd. 55, pp. 417-
433.
Lavpowsky 1877 Zur feineren Anatomie und Physiologie der Speicheldriisen,
inbesonders der Orbitaldriise. Arch. f. Mik. Anat., Bd. 13, p. 281.
Lewis, M.R., anp Lewis. W.H. 1914 Mitochondria in tissue culture. Science,
vol. 39, p. 330.
Leypic, F. 1857 Lehrbuch der Histologie des Menschen und der Tiere. Frank-
furt a. M., p. 228.
Lor, L. 1988 Notes Anatomiques, Sur les Glandes de l’Orbite mécconue chez
le Lapin. Jour. de |’Anat. et de le Physiol., vol. 34, p. 463.
Lutz 1899 Beitrige zur Kenntnis der Driisen des dritten Augenlides. Deut-
sche Zeitschr. f. Tiermedizin. N. F.
Macatuium, A. B. 1895 On the distribution of assimilated compounds of iron
other than haemaglobin and haematin in animal and vegetable cells.
Quart. Jour. Mic. Se., London, vol. 38, pp. 175-274.
Maaaarp 1882 Ueber das Secret und Secretion der Menschlichen Trinendriise.
Virchow Archiv., 89, pp. 258-270.
Matz, F. P. 1896 Reticulated connective tissue and its relation to the con-
nective tissue fibrils. Johns Hopkins Hospital Reports, vol. 1, p.171.
234 JOHN SUNDWALL
Marys, V. 1905-06 Die Entwickelung der Tranenableitungswege. Zeit.
Augenheil., Bd. 14, pp. 222-246; Bd. 16, pp. 303-314.
Maximow, A. 1901 Beitriige zur Histologie und Physiologie der Speicheldriisen.
Arch. f. Mik. Anat., Bd. 58, pp. 1-34.
Maziarski, 8S. 1902 Ueber den Bau und die Einteilung der Driisen. Anat.
Hefte., Bd. 18, pp. 171-237. .
MerKEL, F. 1883 Die Speichelréhren. Rektoratsprogramm, Rostock. Quo-
ted from Fleischer.
Metzner 1907 Die histologischen Verinderungen der Driisen bei ihrer Tatig-
heit. Nagel’s Physiol. des Menschen, vol. 2, p. 977.
Micuaeuis, L. 1899 Die Vitale Farbung, eine Darstellungsmethode der Zell-
granula. Arch. f. Mik. Anat., Bd. 55, pp. 558-575.
MILAWSKY UND SmirNow 1893 Zur Lehre von der Speichelabsonderung. Du
Bois Arch. Suppl., p. 29.
Monest, Luet 1903 Die Morphologie der Fétalen Trénenwebe beim Menschen.
Klin. Monat. Augenheilk., Jahrg. 4 , pp. 1-387.
Mier, E. 1895 Ueber Secretcapillare. Arch. f. Mik. Anat., Bd. 45, p. 463.
1896 Driisenstudien I. His. Archiv. p. 463.
See also Driisenstudien II. Zeitschr. f. wiss. Zool., Bd. 64.
Niconas 1892 Contributions 4 Vétude des cellules glandulares. Arch. de
physiol. norm. et path., Vingtquatrriéme année, T. 4., p. 193-208.
Nout, ALFRED 1901 Morphologische Verinderung der Trinendriise bei der
Secretion. Arch. f. Mik. Anat., Bd. 58, pp. 487-558.
Nusspaum, M. Ueber den Bau und Tatigheit der Driisen. Arch. f. Mik. Anat.,
Bd, 21
Priiiger Artikel. Speicheldriisen, in Stricker’s Handbuch der Lehre von den
Geweben des Menschen und der Thiere, Bd. 1. Quoted from Noll.
PuauisI-ALLEGRA, STEFANO 1903 Sui nervi della glandula lacrymale. Anat.
Anz., Bd., 23, pp. 392-3938.
1904 Studio della glandula lacrymale. Arch. Ital. Anat. e. Embryol.,
Firenze, vol. 8, Fasc. 2, pp. 298-340.
REICHEL, Paunt 1880 Ueber die Morphologischen Verinderungen der Traén-
endriisen bei ihrer Titigkeit. Arch. f. Mik. Anat., Bd. 17, pp. 12-13.
RiquIER, GuISEPPE C. 1911 Contributo allo studio della glando lacrimale
umana. Monit. Zool. Ital., Anno 22, M. 8, pp. 56-65.
SARDEMANN, FE. 1887 Beitrige zur Anatomie der Trinendriise. Inaug. Diss.
Freiburg Preisschrift. Quoted from Hoffmann Schwalbe Jahresb.
16, 1, 1888, S. 465. :
ScHdrer, E. A. 1912 Textbook of Mikr. Anat., Quain’s Anatomy, vol. 2, pt.
I, p. 431.
SCHIEFFERDECKER 1884 Zur Kenntnis des Baues der Schleimdriisen. Arch. f.
Mik. Anat., Bd. 23, p. 382.
ScHIRMER, O. 1904 Mikroskopische Anatomie und Physiologie der Trinenorgane
Im Handbuch der Gesamten Augenheilkunde von*Graefe Saemisch.
2eAuilss 1 aD Kap:
Scumipt, Kurr 1882 Kernverinderungen in den Secretionszellen. Inaug.
Diss. Breslau.
THE LACHRYMAL GLAND 205
ScHWALBE 1872 Beitrige zur Kenntnis der Driisen in den Darmwandungen,
inbesondere der Brunner’schen Driisen. Arch. f. Mik. Anat., Bd. 8,
p. 92.
1887 Lehrbuch der Anatomie der Sinnesorgane. pp. 253-255.
Soporra, J. 1902 Atlas und Grundriss der Histologie und Mikroskopischen
Anatomie des Menschen. Lehman, Miinchen, pp. 238-39.
Souerr, B. 1896 Ueber den feineren Bau der Glandula Submaxillaris des
Menschen mit besondere Beriicksichtigung der Driisengranula. Fest-
schrift f. Gegenbaur. Leipsig. Quoted from Fleischer.
SpaLtTEnouz, W. 1897 Das Bindgewebesgeriist der Diindarmschleimhaut des
Hundes. Arch. f. Anat. u. Physiol., Anat. Abt. Supl., Bd. 1, p. 273.
SPECIALE-CIRINCIONE 1908 Ueber die Entwickelung der Trinendriise beim
Menschen. Arch. Ophthalm., Bd. 69, pp. 193-230. Quoted from
Schwalbe’s Jahresb.
STANCULEANU UND THHoRHARI 1898 Etat dela glande lachrimale dans Larmoie-
ment chronique. Arch. f. Ophthalm., 18, 8. 737.
Stour 1887 Ueber Schleimdriisen. Festschrift f. v. Kélliker.
1891 Lehrbuch der Histologie des Menschen, 4 Auf., Jena, pp. 249-250.
SuNDWALL, J. 1906 The Harderian Gland of the Ox. Anat. Rec., vol. 1, no. 4,
p:. 2.
TEPTIACHINE 1894 Recherches sur les nerfs sécretoires de la glande lacrymale.
A. d. ophthalmol., Paris, Année 14, pp. 401-413.
Unna 1892 Ueber die Bedeutung der Plasmazellen etc., Berl. Klin. Wschr.
Unna-TaEnzeER 1910 Enzyk. der Mik. Tech., vol. 1, p. 394.
Van TrorzenBurRG 1901 Petrus Camper D. 208-227. Reviewed from Schwal-
be’s Jahresb.
WaupEyYER 1875 Reference in Enzyk. Mik. Tech., 1910, vol. 2, p. 409.
WIEDERSHEIM, R. 1876 Die Kopfdriisen der geschwinzten Amphibien und die
Glandula intermaxillaris der Anuren. Zeitschr. f. wsch. Zool., Bd. 27.
1907 Comparative Anatomy of Vertebrates, Parker’s Translation,
p. 288.
ZIMMERMANN 1898 Beitrage zur Kenntnis einiger Driisen und Epithelium.
Arch. f. Mak. Anat., Bd 52; p: 552.
All drawings are made from preparations of the lachrymal gland obtained from
Bovidae.
i
i]
7
‘i
| j
.
5 A
ihe: o
Vn ees
r
ee,
ae ri
tine EMA,
*
i
a ue
"y
'
1
?
“ .
4
4
.
1
‘
Rc
or
‘
{
EXPERIMENTAL MITOCHONDRIAL CHANGES IN THE
PANCREAS IN PHOSPHORUS POISONING
W. J. M. SCOTT
Anatomical Laboratory, Johns Hopkins University
SEVEN FIGURES (ONE PLATE)
The interest which has been aroused in mitochondria in recent
years and their possible importance in the cell economy is evident
when Champy (11, p. 154) is willing to make the statement
that ‘‘I would not regard as living a cytoplasm which does not
contain mitochondria” and when Cowdry (716 b) remarks that
‘‘mitochondria are as characteristic of the cytoplasm as chro-
matin is of the nucleus.”’
There is one cardinal fact which must be emphasized and given
its true value in any discussion of the literature on mitochondria
in pathological conditions. This important point, so frequently
ignored, is that these structures were brought into prominence
over a quarter of a century ago by the brilliant investigations
of Altmann, but they were seen and described by many investi-
gators before Altmann, notably by Flemming. It is indisput-
able that many of his granules, particularly those styled vegeta-
tive filaments, are identical with mitochondria. His investi-
gations had a very stimulating effect upon pathologists with the
result that several important and truly classical papers on Alt-
mann’s granules in various pathological conditions appeared in
rapid succession. Reference need only be made to the early
numbers of Lubarsch and Ostertag’s Ergebnisse and to the ex-
tensive bibliography given by Galeotti (’95, p. 544).
Now Altmann’s work fell into disrepute and was actually
scoffed at on account of the bizarre theoretical interpretations
in which he indulged. He thought, for instance, that the gran-
ules were elementary organisms which existed in the form of
colonies in all cells. This very naturally prejudiced investi-
237
238 WwW. J. M. SCOTT
gators against him and his work and they speedily lost sight of
his very valuable objective findings. Soon his researches lost
all their novelty and were forgotten together with the patho-
logical studies based on them. 3
Very recently interest has been revived in these granules of
Altmann, now called mitochondria, but unfortunately we are
rather inclined to regard their study as something new and novel.
We are altogether too apt to think that anything relating to them
in pathology is new. On the other hand, there is a danger in
accepting Altmann’s work unreservedly because the technique
which he used was far from specific and brought to ight many
granules which are certainly not mitochondria. So that we
have to deal with two mitochondrial literatures in pathology,
separated by a gap of twenty years or more, the first was stimu-
lated by Altmann, and the second is to be regarded as one of the
manifestations of the recent revival of interest in protoplasm, as
contrasted with the nucleus.
So far as I have been able to ascertain there are no observa-
tions on mitochondrial changes in phosphorus poisoning in the
older literature but several contributions of importance have
been made in the last few years. The first of these is a paper by
Ciaccio (13, p. 725) who has made use of phosphorus, along with
other poisons, to bring about experimental changes in mito-
chondria. His one brief reference to the pancreas is the only
observation which I have been able to find concerning mitochon-
drial changes in that tissue as the result of phosphorus poisoning.
He has two figures illustrating the phenomena which he observed.
They show simply a diffuse knotty swelling of the mitochondria
to which he applies the term ‘Praeplastorhexis.’ Indeed he does
not mention these changes in the pancreas in the text at all,
apart from the general discussion of the changes in other organs.
We owe by far the most detailed information regarding the
mitochondrial changes in phosphorus poisoning, as well as in a
large variety of other intoxications, to Mayer, Rathery and
Schaeffer (14, p. 607) who worked on the liver. They found
that the mitochondria respond by a change in their form and
staining reactions. On the basis of their observations they divide
CHANGES IN PANCREAS, PHOSPHORUS POISONING 239
the reactions of liver cells into two types. The first is character-
ized by cytolysis and chondriolysis. The cells increase in size,
clear spaces appear in their cytoplasm, the mitochondria de-
crease in number and the nuclei change in appearance. In the
second there is a ‘homogéneisation protoplasmique”’ and a “‘chon-
driomegalie,”’ the protoplasm staining diffusely with fuchsin,
the mitochondria. increasing in size, and the cells themselves
decreasing in volume.
These investigators also showed that, when they were able to
increase the amount of mitochondrial substance in the liver
experimentally, the same liver, on chemical analysis, showed an
increase in the content of phospholipin. In this way they con-
nected up their histological with their chemical findings, a result
which might be expected in view of the evidence that the mito-
chondria are themselves, at least in part, composed of phospholi-
pin. Three investigators, Regaud, Fauré-Fremiet, and Léwschin,
working independently on mitochondria in mammals, inverte-
brates and plants, each came to the conclusion that they were
made up of a combination of lipoid and albumin.
The possible significance of mitochondria in this connection
is brought home to us when we reflect upon their lipoid nature
and upon the increasing importance which investigators are now
inclined to attach to the réle of lipoids in cell processes. This
point of view is aptly stated by Mathews (’16, p. 88) who regards
phospholipins as the most importance substances in living matter:
For they are found in all eells, and it is undoubtedly their function
to produce, with cholesterol, the peculiar semifluid, semisolid state of
protoplasm. This physical substratum of phospholipin differs in
different cells and probably in the same type of cells in different animals,
but everywhere, from the lowest plants to the highly differentiated
brain cells of mammals and of man himself, it possesses certain funda-
mental chemical and physical properties.
Workers in this field, however, have been slow to realize the
importance of the relation which may exist between the histolog-
ical study of mitochondria and the new analyece of the lipoidal
content of cells.
The idea which underlies most of the modern work on mito-
chondria in pathological conditions is that they constitute a
240 WwW. J. M. SCOTT
new and delicate cytoplasmic criterion of cell activity which is
indeed supported by recent observations. Thus Dubreuil (138,
p. 188) found that they increased definitely in inflammatory
processes just as Romeis (13, p. 12) observed an increase in
regenerating tissues. Homans (715, p. 12) was able to relate
them to an increased demand upon the activity of islet cells of
the pancreas in experimetal diabetes; and Goetsch (16) dis-
covered that an increase in their numbers was correlated with
an increase in the activity of the thyroid epithelium and with
the severity of the clinical symptoms of hyperthyroidism, in man.
MATERIAL AND METHODS
White mice were used. The phosphorus was administered
by subcutaneous injections of an olive oil solution. In the first
experiments the concentration of phosphorus in the solution was
0.0125 per cent which was later increased to 0.05 per cent.
Several injections of various amounts of 0.1 to 0.2 cc. were given
to each mouse at intervals of a day or more—for a longer or a
shorter time depending upon whether it was desired to bring
about a severe or a slight reaction. The animals usually showed
pronounced symptoms within a few days, so that the intoxi-
cations must be regarded as rather acute.
Animals in the desired stage of poisoning were killed with
chloroform. Pieces of the pancreas were first removed to
physiological salt solution. A part of it was then examined in
the fresh condition. Other portions were vitally stained by
immersion in solutions of Janus green as well as with the other
dyes in general use. Pieces were also fixed in a variety of mix-
tures, chief among which may be mentioned formalin and bi-
chromate (Regaud) and acetic osmic bichromate (Bensley) for
mitochondria; osmic acid, commercial formalin, neutral and
alkaline formalin, arnong others, for the demonstration of fat
and other substances. Many different methods of staining
were employed; the acid fuchsin methyl green and the iron
hematoxylin methods for mitochondria; Sudan III, Scharlach
R, Mallory’s stain, hematoxylin and eosin and a variety of
others were used.
CHANGES IN PANCREAS, PHOSPHORUS POISONING 241
The best preparations of mitochondria were obtained in pan-
creases fixed by injection through the blood vessels with a mix-
ture of neutral formalin and potassium bichromate in accordance
with instructions given by Cowdry (16a). Fixation by in-
jection in this way obviates the factor of poor penetration as
well as certain other undesirable accompaniments of the more
crude method of fixation by immersion only. This is even more
necessary in investigations on the central nervous system. I
venture to quote this method of technique in detail:
Fixation. Chloroform the animal. Inject warmed 0.85 per cent
NaCl solution into the aorta through the ventricle. Clamp the vessels
to the part which it is not desired to fix. Continue the injection until
the salt solution is returned uncolored through the veins. Gravity
pressure of not more than 6 feet may be employed.
Follow the salt solution with the formalin and bichromate mixture:
3-per cent potassium bichromate, 4 parts; neutral formalin, 1 part.
The potassium bichromate acts best when freshly prepared. Neutral
formalin is made from the commercial variety by the addition of mag-
nesium carbonate, a deposit of which should always remain at the
bottom of the formalin bottle. Itisimportant that the pressure should
be at the maximum when the mixture is first injected, so that the blood
vessels may be fixed in a state of dilation. If the pressure is low when
the fixative comes in contact with the vessel walls they will be fixed
in a condition of collapse. It will then be difficult, or even impossible,
to obtain a complete injection. The injection should be continued
for about an hour.
The organ is then dissected out and immersed in the fluid. It
should be cut into pieces not more than 1 em. cube. The fixative
must be changed every day for 4 to 5 days, otherwise it undergoes a
change evidenced by a darkening in color. This change is accelerated
by light and by heat, so that the tissue should be kept in the dark and
in a cool place. Fixation may also be effected by simple immersion
of the tissue in the fixative, instead of by injection, but this procedure
is not recommended.
After this prolonged fixation the tissue is mordanted in a fresh 3
per cent solution of potassium bichromate in which it remains for 8
or 9 days, changing every second day.
Wash in running water for 24 hours. The object of this careful
washing is to remove most of the formalin and bichromate, for other-
wise the tissue will be extremely brittle and hard to. cut.
Dehydration and embedding. 50 per cent alcohol 12 hours; 70 per
cent and 95 per cent alcohol 24 hours each; absolute alcohol 6 to 12
hours; half absolute and xylol 6 hours; xylol 3 hours;: paraffin 60°C.
3 hours; cut in 4 serial sections.
Staining. 1) Pass the sections, mounted on slides, down through
toluol, absolute, 95, 70 and 50 per cent alcohol to distilled water.
242 W. J. M. SCOTT
2) 1 per cent aqueous solution of potassium permanganate 30 seconds;
but this time must be determined experimentally.
3) 5 per cent aqueous solution of oxalic acid also about 30 seconds.
4) Rinse in several changes of distilled water about a minute. In-
complete washing prevents the staining with fuchsin.
5) Stain in Altmann’s anilin fuchsin, which is to be made up as
follows: Make a saturated solution of anilin oil in distilled water by
shaking the two together (anilin oil goes into solution in water in about
1 per cent). Filter and add 20 grams of acid fuchsin to 100 ce. of the
filtrate. The stain should be ready to use in about 24 hours. It goes
bad in about a month. To stain, dry the slide with a towel, except
the small area to which the sections are attached. Cover the sections
on the slide with a small amount of the stain and heat over a spirit
lamp until fumes, smelling strongly of anilin oil, come off. Allow to
cool. Let the stain remain on the sections for about 6 minutes. Re-
turn the stain to the bottle.
6) Dry off most of the stain with a towel and rinse in distilled water,
so that the only stain remaining is in the sections. If a large amount
of the free stain remains it will form a troublesome precipitate with the
methyl green; on the other hand, if too much stain is removed the
coloration of the mitochondria will be impaired.
7) Again dry the slide with a towel, except for the area covered
by sections. Allow a little 1 per cent methyl green, added with a
pipette, to flow over the sections, holding the slide over a piece of white
paper so that the colors may be seen. Apply the methyl green for
about 5 seconds at first and then modify the time to suit the needs of
the tissue.
8) Drain off excess of stain and plunge the slide into 95 per cent
alcohol for a second or two, then rinse in absolute for the same time,
clear in toluol, and mount in balsam.
Several difficulties may be met with: 1) The methyl green may
remove all the fuchsin, even when it is only applied for a short time.
This is due to incomplete mordanting of the mitochondria by the chrome
salts in the fixative. It may often be avoided, either by omitting the
treatment with permanganate and oxalic acid, or by treating the sec-
tions with a 2 per cent solution of potassium bichromate for a few
minutes immediately before staining (as advised by Benseley). The
action of the permanganate and oxalic is to remove the excess of bichro-
mate. 2) Or the fuchsin may stain so intensely that the methyl green
removes it very slowly or not at all. This, on the other hand, is due
to too much mordanting. It may be corrected by prolonging the
action of the permanganate and osalic. 3) Sometimes, after obtaining
a good differentiation, the methyl green is washed out before the slide
is placed in toluol. This may be avoided by omitting the 95 per cent
alcohol, by passing from the methyl green to the absolute direct. 4)
Unfortunately the stain is not very permanent. Under favorable con-
ditions it will last for 3 or 4 years. The fading in color is hastened by
light and by heat, and it proceeds very rapidly in a damp atmosphere.
CHANGES IN PANCREAS, PHOSPHORUS POISONING 243
OBSERVATIONS
The mitochondria in the normal pancreas of the white mouse
differ in no noteworthy particular from the mitochondria in the
pancreas of other animals, which, indeed, have been described
again and again (for reference see Bensley ’11, p. 361). At-
tention, however, may be called to the fact that the mitochondria
are filamentous, that they occupy the basal zone of the cell and
that they possess bleb-like swellings all of which can easily be
distinguished in the living cells. This may be seen by reference
to figure 1.
The first change resulting from a very mild degree of phos-
phorus poisoning consists of the loss of the swellings on the
mitochondrial flaments (fig. 2). At the same time the mito-
chondria become shorter and thicker. Some of them appear
spherical, others ovoid. These altered mitochondria can be
clearly distinguished from the zymogen granules by their stain-
ing reactions. The zymogen granules stain a purple or an olive
brown color depending upon the degree of differentiation, and
the mitochondria bright crimson with the fuchsin methyl green
stain. This alteration in the mitochondria precedes any notice-
able change in the other cell constituents. The nucleus still
stands out sharply, the nuclear membrane being quite distinct,
the cytoplasm stains a homogeneous bright green color, and
the zymogen granules occupy the distal zone of the cell.
The second stage of the process is represented in figure 3.
It can readily be seen that the mitochondria exhibit a remarkable
tendency to clump together like agglutinating bacilli. The
clumps occur most frequently in the basal portions of the cells.
They are of variable size. Some of them are composed of only
three or four mitochondria. In others, however, it is possible
to count eighty or even more individual granules. The clumps
vary in consistency, some of them being loose and others quite
compact. The mitochondria are more closely crowded in the
center. But even here the individual granules may be dis-
tinguished from one another by the reflection from their curved
surfaces. The peripheral part of such a clump contains mito-
chondria less densely packed together. Strands, three or four
244 W. J. M.°SCorTr
mitochondria thick, often radiate from such a clump. These
usually extend parallel to the nearest margin of the cell, and the
mitochondria in them are oriented so that their axes correspond
in a general way with the direction of the strand. The cell on
the right shows two such strands running out from the large
mass of agglutinated mitochondria. In some cells almost
all the mitochondria are agglutinated in a single large mass, but
in others many scattered clumps of small size are formed. Where
there are large clumps of mitochondria those of them which are
usually distributed throughout the cytoplasm are greatly di-
minished in number. There seems to be a complete transition
between the condition of unclumped mitochondria (shown in
fig. 2) and these masses of agglutinated ones. Parallel with
agglutination of mitochondria there are other evidences of cel-
lular damage: the nuclei, instead of being sharp and clear, are
hard to define; frequently the nuclear membrane cannot be dis-
tinguished, the nucleoli are apparently absent; and the cytoplasm
stains intensely as well as unevenly.
In the next stage, that shown in figure 4, a noticeable change
is observed in the clumps above described. The agglutinated
mitochondria fuse together and lose their individuality. Figure
4 is the representation of a single cell which shows the transition
from agglutination to fusion. The triangular shaped mass just
above the nucleus is made up of closely packed, yet discrete
granules. It is a typical clump of agglutinated mitochondria.
The second large mass in the cell stains crimson with the acid
fuchsin just as the other does. Its outline, however, is roughly
spherical, but still somewhat irregular in shape, and it is impos-
sible to resolve it into separate mitochondria. It is not perfectly
homogeneous, though, for it shows evidence of its original granu-
lar nature. These two masses represent practically the entire
mitochondrial content of the cell although occasional scattered
mitochondria remain. It is possible to see in one section many
cells which illustrate well the agglutination as well as others
which show complete fusion.
Figure 5 shows the results of a more severe intoxication with
phosphorus. The cells contain globular masses which bear a
CHANGES IN PANCREAS, PHOSPHORUS POISONING 245
remarkable resemblance to those formed in the preceding stage
through the fusion of the agglutinated mitochondria. They
differ, however, from these in that they are more spherical and show
not the least evidence of a granular structure. Some are quite
small, but the largest may actually exceed the nucleus in size.
There is a variable number of themin each cell. Their properties,
so far as I have observed, are as follows: 1) They are drop-like in
form; 2) They stain faintly bluish with hematoxylin and eosin;
3) They do not stain specifically with Sudan III or Scharlach R;
4) They are insoluble in absolute alcohol and toluol after chro-
matization (3 per cent solution K,Cr,O0; eleven days, 8°C.); 5)
They stain with acid fuchsin, in fuchsin methyl green, and with
Orange G in Mallory’s triple stain. I assume that they are
lipoid.
Within these droplets are to be observed little spherical
vacuoles which do not stain at all. These clear spaces are some-
times quite numerous especially in the largest droplets. Some
of the droplets, indeed, appear as only a rim of stainable sub-
stance about such a vacuole. That this phenomenon is not due
to the imperfect penetration of the chromatizing solution is
proven by the fact that the vacuoles may be arranged along the
periphery as is seen in figure 5 in the largest droplet. It must
be due to a difference in the solubility, to a loss of the staining
capacity, or to an alteration in the composition of this part of
the lpoid droplet. The protoplasm surrounding these droplets
stains very intensely with the methyl green. Further away this
dark color shades gradually off into the ight green which the
cytoplasm usually stains. In this more deeply staining zone
mitochondria are seldom present; but in that part of the cyto-
plasm possessing the more normal staining reaction they are
to be observed. It is to be especially noted that they show no
tendency to agglutinate and possess no bleb-like swellings.
There are many nuclei which seem to be quite normal in appear-
ance but there are others which clearly bear evidence of excessive
cell damage.
In the same specimen which shows these large lipoid droplets
cells are to be found riddled with clear vacuoles (fig. 6). Such
246 WwW. J. M. SCOTT
cells are relatively scarce. The vacuoles range in size from 0.1u
to 3u in diameter. They have every appearance of being the
empty spaces left after neutral fat has been dissolved out. It
seems quite probable that this is the case for the chromatization
(eleven days at about 8°C.) would be quite insufficient to render
neutral fats insoluble. It is to be particularly noted that sucha |
cell contains the lipoidal droplets, and that its cytoplasm stains
deeply. A few mitochondria are still present in cells of this sort
visible between the vacuoles. This cell is of interest since
within it are shown the results of two processes, fatty infiltration,
and the formation of the lipoid droplets from mitochondria.
This great accumulation of neutral fat within the cell is not
associated in any demonstrable way with the formation of the
lipoid droplets, and is not a part of any stage of this process, as is
evidenced by the fact that only a few scattered cells show it and
that it isnot aregular accompaniment of any stage in the aggluti-
nation or fusion of mitochondria.
A very interesting condition, quite unlike any described above,
is observed in a number of the cells in the same sections from
which figures 5 and 6 were drawn. The whole cell is densely
packed with zymogen granules of varying sizes, which are
usually confined in the normal acinus cell, to the distal zone
(fig. 7). It is possible to distinguish clearly mitochondria as
brightly stained rods between the zymogen granules. These
mitochondria are short and thick and have no blebs. They
certainly do not seem to be increased in number; and, while it is
impossible to estimate their number exactly, because the zymo-
gen granules are so densely packed, the impression is gained
that they are somewhat diminished in number. The mito-
chondria and secretion granules can be clearly distinguished
here also by their difference in morphology and staining reaction:
the zymogen granules being always perfectly spherical and
staining an olive-brown to a purplish red (depending on the
differentiation), while the mitochondria are scarcely ever per-
fectly spherical and stain a brilliant crimson. Indeed it is
possible in all our preparations to distinguish between mito-
chondria and zymogen granules by their staining reaction in the
CHANGES IN PANCREAS, PHOSPHORUS POISONING 247
acid fuchsin-methyl green preparations after neutral formalin-
bichromate fixation. In the living cell the distinction is easy,
too, by reason of the relatively high refractive index of the zymo-
gen. These secretion granules are in no discoverable way differ-
ent from those in the normal cell. They stain specifically with
neutral gentian, and their size and form is unchanged. There
are no droplets of lipoid and no agglutinated masses of mito-
chondria in these cells. The only possible interpretation of this
condition is that it be either a retention of secretion or else an
excessive formation of it. Curiously enough such cells packed
with secretion and exhibiting mitochondrial changes of only the
first stage are found near cells containing the large lipoid drop-
lets and very few, if any, zymogen granules. Cells of this
variety are really quite numerous. They frequently occur
together in acini but they may also be seen in acini with other
cells which show the typical lipoid droplets, and scarcely any
zymogen granules. The reason for changes seemingly so opposite
in nature occurring in neighboring cells is not at all apparent.
It is true, however, that the reaction of individual cells to the
poisoning in the other types of change varied greatly in degree.
It may be that this heaping up of secretion in the cell is merely
an evidence of an altered metabolism, possibly an abnormal
stimulation, in these cells showing the least evidence of damage.
This type of cell, whatever be its cause, emphasizes the fact that
whenever the poisoning has affected a change in another con-
stituent of the cell the mitochondria are found to be altered.
Among the large number of pancreases examined certain of
them showed intracellular infiltration of hyaline substance in
both acinus cells and islet cells, while others contained small
areas of necrosis in which the mitochondria had practically dis-
appeared. Even some of the control animals presented minute
foci of fatty infiltration in the pancreas.
Other tissues of the animals poisoned with phosphorus were
examined as a check upon the changes in the pancreas. The
alterations are of course most pronounced in the liver and my
findings in this organ are in a measure confirmatory of the ob-
servations of Mayer, Rathery and Schaeffer (14, p. 608), but I
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 2
248 W. J..M. SCOTT
have not seen their stage of “homogenéisation protoplasmique”’
in which the mitochondria are supposed to go into solution and
the whole cell to stain intensely with acid dyes. Briefly I find
that the mitochondria lose their filamentous or rodlike outlines,
become spherical and progressively lose their staining reactions
so that they finally are not seen at all or merely as shadowy
structures. There is also, as is to be expected, a very pronounced
fatty infiltration. The alterations in the kidney are those of a
typical cloudy swelling in which the mitochondrial changes
have been carefully described, under the heading of Altmann’s
granules by several authors since Lubarsch’s original publication.
The cells of the central nervous system show no outspoken
changes in their mitochondrial content. Finally it must be
noted that the pancreas alone showed agglutination and fusion
of mitochondria and that no tendency of this sort was observed
in any other tissue.
DISCUSSION
The points of chief interest in this work are (1) the immedi-
ate loss of the bleb-like swellings on the mitochondria as the
first evidence of pathological change, (2) the phenomenon of
agglutination as a type of mitochondrial reaction and (3) the
source of the lipoid droplets. ;
The swellings in the course of the mitochondrial filaments are
considered by many investigators to be the precursors of secre-
tion. The crucial point is the similarity or the dissimilarity
between the mitochondria, these blebs and the zymogen granules.
Key (16, p. 216) is opposed to this view since he finds that, on
increasing the secretory activity of the cell, through the adminis-
tration of secretion and pilocarpin, the mitochondria show no
change indicative of participation in the formation of zymogen.
Moreover Cowdry (’16b) has been able to stain the blebs in
quite a different way from the zymogen granules as well as from
the remainder of the mitochondrial filaments. He fixed mouse
pancreas by injection of neutral formalin-bichromate through
the blood vessels. The tissue was then mordanted in bichromate
and embedded in the usual way. Sections stained with iron
CHANGES IN PANCREAS, PHOSPHORUS POISONING 249
hematoxylin and counterstained with safranin and light green
showed the mitochondria and the zymogen granules blue-black,
the nuclei bright red and the blebs light red against a green
background.
My observations seem to bear upon this problem in two ways.
In the first place I have found that the blebs in the normal
pancreas, with the method of technique employed, always stain
differently from the zymogen granules. In the second place
the changes which I have observed in phosphorus poisoning
seem to be still more suggestive. The fact that the blebs are
the very first structures to disappear with a very mild degree of
phosphorus poisoning (fig. 2) and that zymogen granules are
heaped up in enormous numbers in some cells of animals more
severely affected (fig. 7) would seem to indicate that the production
of zymogen granules does not necessarily cease with the disap-
pearance of the blebs, which is rather at variance with the hy-
pothesis that the mitochondria participate, through their bleb
like swellings, in the formation of zymogen granules.
The agglutination observed during the course of phosphorus
poisoning is a new reaction on the part of mitochondria. It is
interesting to compare this clumping of mitochondria (which
are thought to be lipoidal in nature) with that of bacilli and red
blood cells. Jobling and Peterson (14, p. 453) make the state-
ment that ‘‘with or without a morphologically distinct limiting
membrane we can reasonably assume that the external surface
of the bacterial cell is potentially lipoidal”’ and the lipoid nature
of red blood cells is well recognized. In each of these three cases
the fact that this phenomenon of agglutination is a reaction to
pathological conditions should not be lost sight of. Moreover
agglutination is a phenomenon which always occurs in a fluid
medium which fact is not without significance from the point of
view of cell structure because it indicates the fluid nature of
the protoplasm of the pancreas cell, and militates against the
doctrine of a cytoplasmic reticulum.
It has long been known that there is a deposition, or perhaps
more correctly speaking a formation, of fatty lipoid droplets
in the cells of the pancreas as the result of phosphorus poisoning,
250 WwW. J. M. SCOTT
but the source of the material has been much debated. There
are several possible, though unlikely, sources of these lpoid
droplets which must be considered. Normal pancreases often
contain, in their acinus cells, small droplets of neutral fat and of
lipoid in variable quantities. I have studied their distribution by
staining with Sudan III and Scharlach R, after formalin fixation,
and by vitally staining with dilute solutions of nile blue B extra,
Meldola’s blue (Sandoz, same as naphthol blue) and brilliant
cresyl blue. Fat of this sort does not stain with Janus green. It
is also well known (Bensley 711, p. 363) that the mitochondria,
themselves, occasionally contain droplets of fat, though I have
searched diligently, but without success for traces of it in the
mitochondria of acinus cells of white mice. No relation could
be found between the formation of the lipoid droplets in the
poisoned pancreas and this neutral fat occurring free in the
cell or embedded in the filaments.
Pieces of the pancreas, fixed in neutral formalin and bichro-
mate and stained in the routine manner with hematoxylin and
eosin show these droplets but do not indicate their source. The
whole process is made clear, however, when sections prepared in
this way are stained by the fuchsin methyl green technique
which is here advocated. The mitochondria are brought to light
and the changes of agglutination (figs. 3 and 4) and of fusion
(figs. 4 and 5) to form the lipoid droplets, which they undergo,
are at once revealed. It seems that the mitochondria are the
actual source of the droplets.
I wish to take this opportunity to express my appreciation
for the continuous advice and interest of Dr. E. V. Cowdry in
this research.
CONCLUSIONS
Mitochondria are the first constituents of the acinus cell of the
pancreas to show pathological change in phosphorus poisoning.
They lose their filamentous form, become shorter and thicker,
and their bleb-like swellings which are so characteristic of the
normal pancreas completely disappear (figs. 1 and 2). Then
CHANGES IN PANCREAS, PHOSPHORUS POISONING 251
follows the stage of agglutination in which the mitochondria col-
lect in large compact clumps (figs. 3 and 4). The mitochondria
in these agglutinated masses fuse to form droplets possessing
the characteristic properties of lipoid (figs. 4 and 5).
BIBLIOGRAPHY
Benstey, R. R. 1911 Studies on the pancreas of the guinea pig. Am. Jour.
Anat. vol. 12, pp. 297-388.
Cuampy, C. 1911 Recherches sur l’absorption intestinale. Arch. d’anat. miecr.,
t. 13, pp. 55-170.
Craccio, C. 1913 Zur Physiopathologie der Zelle. I. Entartungsbilder der
Plastosomen. Centralbl. f. Allg. Path. u. Path. Anat. Bd. 24, 8. 721-
727.
Cowpry, E. V. 1916a The structure of the chromophile cells of the nervous
system. Contributions to Embryology, Carnegie Institution of
Washington, No. 11.
1916 b General functional significance of mitochondria. Am. Jour.
Anat., vol. 19, p. 423.
DupreviL, G. 1913 Le chondriome et le dispositif de ’activité secretoire ete.
Arch. a’Anat. Micr., t. 15, pp. 538-151.
FaurE-FREMIET, MAYER AND ScHAEFFER 1910 Sur la microchemie des corps
gras. Application a l’étude des mitochondries. Arch. d’Anat.Miecr.,
t. 12; pp. 12-102.
Gateorti, G. 1895 Ueber die Granulationen in den Zellen. Intern. Monatschr.
f. Anat. u. Phys., Bd. 12, pp. 440-459.
Gortscu, E. 1916 Mitochondrial changes in toxic adenomata of the thyroid
gland. Johns Hopkins Hosp. Bull., vol. 27.
Homans, J. 1915 <A study of experimental diabetes in the canine and its re-
lation to human diabetes. Jour. of Med’ Res., vol. 33, pp. 1-51.
JOBLING, J., AND PrererseNnN, W. 1914 Bacterial antiferments. Jour. Exp.
Med., vol. 20, pp. 452-467.
Key, J. A. 1916 On the relation of mitochondria to zymogen granules. Anat.
Rec., vol. 10, pp. 215-216.
Lowscuin, A. M. 1913 ‘Myelinformen’ und Chondriosomen. Ber, d. Deut.
Bot. Ges., Bd. 31, pp. 203-209.
Marnews, A. P. 1915 Physiological Chemistry. New York. William Wood
and Company. 1039 pp.
Mayer, RarHery AND SCHAEFFER 1914 Les granulations ou mitochondries de
la cellule hepatique. Jour. de phys. et de Path. gen., t. 16, pp. 607-622.
Romets, B. 1913 Das Verhalten der Plastosomen bei der Regeneration. Anat.
Anz., Bd. 45, pp. 1-19.
DESCRIPTION OF FIGURES
All the illustrations have been drawn from pancreas cells of white mice, all
except the control showing varying degrees of poisoning with phosphorus, fixed
by injection of a mixture of neutral formalin and potassium bichromate through
the blood vessels, according to the instructions given by Cowdry (16a). See-
tions, 4u thick, were stained with fuchsin and methyl green. In the original
preparations the mitochondria are bright crimson, the zymogen granules purple
and the ground substance green. Zeiss apochromatic objective 1.5 mm., com-
pensating ocular 6 and camera lucida were used in making the drawings. They
were not reduced in reproduction so that they represent a magnification of 1640
diameters as they now appear on the plate.
PEAR:
EXPLANATION OF FIGURES
1 A portion of an acinus of the pancreas of a normal female white mouse
weighing 36 grams illustrated for control. Note particularly the long filament-
ous mitochondria with their bleb-like swellings stretching from the basement
membrane toward the lumen. The zymogen granules are present, in moderate
amount, in the distal parts of the cells.
2. A group of acinus cells of a female white mouse, weighing 17 grams, which
was injected subcutaneously with 0.2 ec. of an 0.0125 per cent solution of phos-
phorus in olive oil. Five days later the animal became comatose, and was killed
and examined. ‘The cells show very nicely the first stage in phosphorus poison-
ing in which a change in the mitochondria is alone noticeable. -They have lost
their bleb-like swellings and have become shorter and more rounded. Compare
with figure 1.
3 Two acinus cells showing a rather more pronounced change. The mito-
chondria are seen in clumps in the basal parts of the cells, the zymogen granules
are decreased in amount and the nuclei are hard to define.
4 A single acinus cell, also from the same pancreas, which is of interest be-
cause it shows on the left hand side a triangular agglutinated mass of mitochondria
and, on the right, such a clump which has undergone partial fusion in the process
of formation of a lipoid droplet.
5 Portion of an acinus of a female white mouse weighing 16 grams which was
more severely poisoned with phosphorus. It received 0.2 ec. of a 0.05 per cent
solutions subcutaneously. On the third day the mouse was comatose. It was
killed and examined in the usual way. All the agglutinating mitochondria have,
in this stage, fused, and have formed numerous spherical lipoid droplets. These
droplets contain clear vacuoles and the cytoplasm surrounding them stains
intensely.
6 A cell from the same pancreas showing a pronounced fatty infiltration as
well as the formation of lipoid droplets.
7 Another cell from the same pancreas showing a great increase in the num-
ber of zymogen granules. One or two mitochondria may be seen scattered among
them.
CHANGES IN PANCREAS, PHOSPHORUS POISONING
W. J. M. SCOTT
PLATE 1
EQUIVALENCE OF DIFFERENT HEMATOPOIETIC
ANLAGES. (BY METHOD OF STIMULATION OF
THEIR STEM CELLS)
I. SPLEEN
VERA DANCHAKOFF
The Wistar Institute of Anatomy and Biology
TWO TEXT FIGURES AND NINE PLATES
. 5
1. Introduction. Statement of the problems of hematology. Different
methods of attacking the problems. Myeloid metaplasis. Statement
ep-the problemol sthe; present paper... - S-aetseepteaees vee si Seosek (200
De Vethodssoniumvestimationsandenexminolopye ae eee ceiece elt cires 267
3. Histogenesis in the spleen in relation to structural environment........ 274
Ae Histozenesis) in thelnormal chick spleen sAess-ee eee 274
B. Histogenesis in the chick spleen after stimulation ............... 285
a. Changes after stimulation in early stages.................. 287
6. Changes obtained by stimulation in the stage with definite
Spleen vasculaxcization....2..... 2: \6ceeeeeeeetos scien s sske 294
C. Conclusions. Data concerning: a) the histogenesis of spleen cells,
b) the conception of cell differentiation, histogenesis, c) the
general meaning of myeloid metaplasis........................ 296
1. INTRODUCTION
Statement of the problems of hematology. Different methods of
attacking the problems. Myeloid metaplasis
The special hematological literature has been rapidly and
constantly growing during the last decade, and there is enough
ground to anticipate a further specialization in this partial
domain of biology. Surveying the voluminous hematological
literature, one is astonished by the discrepancy between the
simplicity of the fundamental problems of hematology and the
entanglement which they have undergone during the last
255
THE AMERICAN JOURNAL OF ANATOMY, VOL, 20, No. 3
NOVEMBER, 1916
256 VERA DANCHAKOFF
decade. One is naturally inclined to doubt, whether the small
variations in the innumerable schemes are as important, as
they are represented in the deductions of different workers.
The fundamental problems of hematology have a close bear-
ing upon practical medicine as well as on the general problems
of biology. Wilson and Conklin (6) especially, and many other
biologists established as a fact the gradual segregation of differ-
ent chemical materials collected in the egg. This is accomplished
by a range of cleavages beginning in certain eggs at the first
cleavage, in others delayed to later stages. As a result there
appear groups of cells, different in their constitution and there-
fore possessing different potencies for differentiation. These
cell groups proliferate and give tissues, which in their own differ-
entiation may exhibit a great complexity. The diversity in the
products of differentiation may be due either to differences in
the physico-chemical constitution of the cells, or to differences
in environmental conditions. One of the important questions
of biology consists in determining the limit to which each of the
agents cited has an active role. Does the segregation lead to
a production of a definite number of uninterchangeable blood
stem cells, of which the differences in the chemical structure imply
a definite metabolism and their further specific development?
Or does the process of segregation lead to a production of one
group of numerous homogeneous primitive blood cells, which
under different conditions in the embryo, as well as in the adult
organism, split off variously differentiated cells? This problem
was differently conceived by the monophyletic and polyphyletic
schools.
The importance of a true conception of this fundamental
problem of hematology for practical medicine is obvious. The
uninterrupted, everyday destruction of the blood elements is
too often accompanied by a failure of regeneration in the organ-
ism. A whole range of stimulating agents for blood regener-
ation was found empirically. Still the success of interventions
in similar cases as well as of interventions in various other
deviations from the normal course of hematopoiesis (leukaemias)
depends greatly upon a clear understanding of the action of the
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 257
agent on the processes underlying a definite symptom complex.
An expenditure implies a complete knowledge of the reserve
stock, the reserve being represented in the particular case by
the group of cells, which by their uninterrupted reproduction
and differentiation may supply the deficiencies.
The reserve stock is differently appraised by the monophyletic
and polyphyletic schools. According to the polyphyletic school
these reserves are formed by a number of specifically differenti-
ated cell groups, which may multiply and ripen into still more
specialized cells. The monophyletic school, on the contrary,
finds in the organism along the various intermediate stages of
differentiation also the young, undifferentiated polyvalent cells,
which in the embryo have been the source of the various lines of
blood cell differentiation. 'The monophyletic school admits a
greater amplitude in the regeneration, and hence a wider sphere
of influence upon the blood tissue. These common stem cells
were identified by Pappenheim (31) as the large lymphocytes,
by Dominici (10) as the small lymphocytes, by Weidenreich (44)
and Downey (11, 12) as reticular cells.
However great the differences between the mono- and poly-
phyletic schools may seem, both groups of hematologists admit
in the adult organism the existence of younger and riper cells
and of their gradual differentiation products. The difference
between them consists chiefly in the value of the amplitude of
the differentiation process in the adult organism, in other words
in the admission of different degrees of segregation, which may
preserve or dissolve the common stem cells for various blood
elements.
Various methods were applied for the elucidation of the funda-
mental problems of hematology, as well as for the understanding
of the mutual relationship of the various blood-cells and of the
structure of the hematopoietic organs. As expected, the his-
tological study of normal preparations of hematopoietic organs
in adult organism was unable to solve the problem. The mere
statement of coexisting cells in the hematopoietic organs offers
a wide field for personal interpretation. Therefore the most
258 VERA DANCHAKOFF
contradictory conclusions regarding the structure of the blood
tissue were made on the basis of this method.
The embryo-genetic method may give more definite results
if used exhaustively. The gradual appearance of the different
blood cells becomes a reliable criterion for the judgment of the
mutual relationship of the blood cells. The discovery of young-
er cells, which first develop in the embryo and the study of
their gradual differentiation may be of great help in the identifi-
cation of the blood cells in the more complex structure of the
hematopoietic organs in the adult organism. However, the
demand for strict exhaustiveness is often disregarded and leads
to gaps, which are filled by a number of more or less keen inter-
pretations of the investigator. The omission of the study of
the first stages in the development of blood cells led investigators
to an admission of specific stem cells for various differentiation
products—Denys, (8), Bizzezero (38),- and others. How-
ever, the recent histogenetic studies made by different investi-
gators on various animals gave similar results. Bryce (2),
Danchakoff (9), Maximow (25), Mollier (28), Haff (18), all
admitted that the different blood cells are derived from mesen-
chyme of mesodermal origin, and that the various blood-forming
organs developed autochthonously at the expense of mesen-
chymal cells of which the various differentiation may depend
upon external physico-chemical agents.
Studies of pathological changes in hematopoietic organs did
not contribute greatly to the solution of the problems of hemat-
ology. The complexity in the mutual relationship of the blood
cells was recognized and led many pathologists to take re-
course to embryo-genetic studies, (Fischer (16,) Schridde (38),
Schmidt (89) and others). The conclusions drawn from studies
made by these methods were based upon the existence of
morphological similarities and differences between various blood
cells. They lead to more or less plausible probabilities but do
not determine irrefutably the functional dynamics of the various
blood cells in different hematopoietic organs.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 259
The recourse to the experimental method may contribute
more definite information about mutual relationship of blood-
cells as well as about the source of their regeneration and the
function of different blood-forming organs. The study of in-
tense general destruction of blood tissue through bleedings and
of its partial destruction by X-ray application, followed by the
study of its regeneration, have already contributed undeniable
data. The use of a method known as lympho and myelotoxic
intoxication, (I would better say stimulation) may give in this
respect most conclusive results. This method was indicated in
1901 by Dr. Flexner (17). Specific leucolytic, spleno- lympho-
marrow-lytic sera are injected into animals. The consecutive
circulation of antibodies present in these sera find adequate
receptors in certain cell groups and a proliferation follows as a
result of molecular changes in these cells.
One of the most striking results obtained in experimental
pathology of the hematopoietic organs is the myeloid metaplasis
of lymphoid tissue. First described by Fraenkel it was experi-
mentally obtained by Dominici (10) by producing traumatic
anaemias and typhoid infection. Later the myeloid metaplasis
of the lymphoid tissue was repeatedly observed during various
pathologieal processes, (anaemia, leukemia, intoxications, infec-
tions, tumors). According to the statements of various investi-
gators, the myeloid metaplasis invariably consists in a simul-
taneous development of erythro- and leucopoietic or granulo-
blastic tissues. This was accounted for by most of the patho-
logists as being a proof of a close relationship between these
tissues. The common origin of leucocytes and erythrocytes
was therefore admitted by many pathologists who formed the
so-called dualistic school, which distinctly separated the myelo-
blastic (erythro-granuloblastic) tissue from the lymphatic tissue.
The present paper has for its main subject the study of a very
extensive chiefly granuloblastic (or myeloid) metaplasis of the
embryonic mesenchyme, therefore a survey of different opinions
concerning the subject may be permitted.
The myeloid metaplasis was observed chiefly in the spleen,
and occasionally mentioned in the lymph-glands and other
260 VERA DANCHAKOFF
organs of the adult organism. Dominici (10) admitted in his
paper (’01) that cells with the structure of small lymphocytes
give rise to the myeloid tissue. ‘‘Les petits mononucléaires
en question (in another part of his paper he uses the term of
small lymphocytes) ont des dimensions égales ou inférieures 4
celles des hématies: un noyau rond, une bordure protoplasmique
mince. Un élément figuré, offrant de tels attributs, n’est ce pas
ce que l’on a dénommé communément une cellule embryonnaire?”’
A closer study of the embryogenesis of the blood tissue would
certainly not have allowed Dominici to consider the small lym-
phocytes as being young embryonic cells, for the cells bearing the
structure of small lymphocytes appear in the organism consider-
ably later than other lymphatic cells, at least in mammals, birds
and reptiles. Though the stem cells of the myeloid tissue,
according to Dominici, bear a morphological structure identical
with cells, which differentiate into lymphatic tissue, he still
attributes to the myeloid stem cells specific potencies of differ-
entiation, which they preserve from early embryonic period.
Later, in 1909, Dominici (10) changed his opinion and attributed
to the lymphatic tissue itself the faculty of myeloid transform-
ation. Dominici was thus first to stand for the conception of
autochthonous development of the myeloid tissue in the lymph-
atic organs.
According to Ehrlich (13) the myeloid tissue, which appears
under certain pathological conditions in different parts of the
adult organism, derived from the central myeloid organ, namely
from the bone-marrow. ‘This opinion was expressed by Ehrlich
at a time when the differentiation of every blood-cell seemed to
be predestined generations back by the specific constitution of
their ancestral cells. Ziegler (47) and Helly (19) still adhere to
this view. They base their opinion upon the occurrence of bone-
marrow Ausschwemmungen, which follow venous injections of
parenchymal mash, also after bone-marrow traumas. Helly
finds support to his view also in the fact that after
Beeinflussung des Knochen-marks mittels Bakterien schon nach
kurzer Zeit in der Milz des Kaninchens, welche unter normalen Ver-
haltnissen so gut wie gar keine specifischen Markelemente jiingerer
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 261
Generation wit etwa Myelozyten, enthalt, eine derart hochgradige und
herdweise auftretende Einlagerung solcher Zellen vorkommt, welche
sich nicht nur im Blute sondern auch in den Kapillaren anderer Organe
reichlich finden, dass die Erklirung mit Hilfe der gedachten Ver-
schleppung bei weitem die nahe liegendste ist.
However, the existence of myeloid metaplasis without any
changes in the circulating blood, and further the discovery of
diffuse autochthonous development of granuloblastic tissue in
early stages of embryonic development, led most of the investi-
gators to oppose strongly Ehrlich’s conception of the myeloid
metaplasis as being a metastasis from the central myeloid
organ. If most of the investigators are inclined to consider the
myeloid metaplasis as a local autochthonous process, their
opinions concerning the source of the myeloid tissue, which
develops in lymphatic organs, differ widely.
The fact that the myeloid metaplasis is chiefly localized in
the pulpa of the spleen, became one of the main arguments of
the dualists for denying any relationship between the lymphatic
and the myeloid tissue in the adult organism. According to
the dualists, cells from remote embryonic periods remained in the
regions of the spleen-pulpa. Though their differentiation has
not fully been accomplished, yet it reached stages at which these
young cells respond to a stimulation by development of only
myeloid tissue, (containing erythroblastic, leucoblastic tissue
and megakaryocytes). According to them, the myeloid tissue,
which in embryonic life is more diffusely spread out, does not
disappear completely, but persists in a sort of latent stage.
Bezangon et Labbey (1), Fischer (16), Heinecke (26), Lobenhofer
(24), Meyer (27), Nageli (81), Schridde (88), Schmidt (39),
Sternberg (40), Tiirek (42), and others uphold this view. Most
of them attribute the development of myeloid tissue under
pathological conditions to a “sudden wakening”’ of embryonic
potencies in various cell-groups. These cell-groups are differ-
ently identified by different investigators. Preexisting myeloid
tissue is the source of the myeloid metaplasis for Sternberg (40),
indifferent lymphocyte like pulpa-cells for Meyer and Heinecke
(25, 27), Schmidt (39), Lobenhofer (24), Fischer (16), connective
262 VERA DANCHAKOFF
tissue cells for Fischer (16) and Klein (22) and finally reticulum
cells for Klein (22). An attempt was made to explain the mye-
loid metaplasis, not only by attributing potencies, characteris-
tie for embryonic cells to cell-groups in adult hematopoietic
organs, but also by assuming the existence of a mysterious proc-
ess of dedifferentiation of already differentiated cells, Schridde
(38), Fischer (16), Nageli (31).
Among the investigators who studied the myeloid metaplasis
only a few admitted the existence of a common stem-cell for
various blood-cells in the adult organism. Pappenheim (82),
Werzberg (46), later Dominici (10) and Blumenthal (4) became
supporters of the monophyletic interpretation of the myeloid
metaplasis. They pointed out that the localization of the new
myeloid tissue during myeloid metaplasis is not as strict as the
dualists admit. Moreover, the separate localization of the mye-
loid and the lymphatic tissues, where it exists, may become the
differentiating factor for a common stem cell, which of course
could not develop under equal conditions into different products.
The monophyletic interpretation of the myeloid metaplasis
was corroborated to a great extent by embryogenetic studies,
Bryce (2), Danchakoff (9), Maximow (25), Mollier (28), Haff (18)
and by histological studies Se ceierch (44), Downey (11, 12),
Ferrata (383).
The study of myeloid metaplasis of lymphatic tissue has been
mede in adult organs, in which at least a temporary embryonic
or even a permanent partial granuloblastic differentiation existed.
Therefore a possibility of persistence of specific granuloblastic
stem cells in such organs could not be denied and a differenti-
ation of specific cells could be explained by the specific consti-
tution of their stem cells. ‘Under definite stimulation, these
specific stem cells might intensely proliferate. On the other
hand, the hematopoietic organs could also preserve young
specifically undifferentiated cells, polyvalent in their potencies.
Their partial and local differentiation into myeloid tissue might
have been caused by specific conditions of their environment.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 263
It was repeatedly pointed out that the stem cells, which in the
pulpa and in the follicle of the spleen give different products of
development, seem to bear a perfectly similar morphological
structure, Dominici (10), Hirschfeld (21), Meyer u. Heinecke
(26), Weidenreich u. Downey (12), Butterfield (5), and others.
However, many recent data seem to indicate that the isomorph-
ism does not imply either isogenesis nor especially isodynamics.
The study of myeloid metaplasis in such organs, in which both
directions of differentiation coexist permanently in the adult
organism or temporarily in the embryo, does not evidently offer
favorable opportunity of solving the problem: whether the adult
organism does preserve a stock of embryonic undifferentiated
cells capable of various differentiation, or not. Neither do
these studies solve the question: whether the morphological
structure so characteristic of and common to the young cells
both of the lymphatic and myeloid tissue is a result of definite
physico-chemical constitution of which further changes are im-
plied by differences in environmental conditions; or whether a
group of morphologically identical cell units may have different
physico-chemical constitution which would imply their further
different development.
My (9) personal studies on the normal histogenesis of the
*blood-cells and of the hematopoietic organs in birds and reptiles
led me to a monogenetic conception of their origin. The study
of regeneration of hematopoietic tissue after bleedings as well
as that of changes undergone by this tissue during starvation
(9) seemed to corroborate the monogenetic interpretation.
The admission of the existence of common stem cells in differ-
ent hematopoietic organs implies therewith the admission of
identical reaction of these stem cells to a stimulating agent as
far as these stem cells are submitted to the same conditions.
The structural environment in the full-grown organism is,
however, highly differentiated in the various organs. The
stem cells in different hematopoietic organs are of course expected
to respond to stimulation by simultaneous proliferation; but their
differentiation will be specific according to environmental condi-
tions, for their differentiation depends upon conditions, usually
264 VERA DANCHAKOFF
not experimentally controlled, upon the localization of the stem
cell in the pulpa or in the follicle, in the case of the spleen. The
possibility of identical reaction of the stem cells to definite stimuli
is, however, not completely excluded. There may be certain
kinds of stimuli, which may cause deviation from the normal
differentiation of stem cells. A similarity of the reaction of stem
cells in different organs may be expected also in case the differ-
ent environmental conditions are made alike, or in case the
environment is not fully differentiated, as for example in an
embryo.
Under the influence of these premises an observation made by
Dr. Murphy of the Rockefeller Institute attracted my attention.
He observed 2-3 years ago an enlargement of the spleen in the
host embryo after grafts of various tissues.' A closer study of
this process led me to conclude, that the considerable enlarge-
ment of the embryonic spleen is induced by an intense prolif-
eration of the young stem cells. This fact seemed to enable a
1 This observation was not published by Dr. Murphy at the time. While this
paper was in press a brief note by Dr. Murphy regarding the general effect of
_the spleen grafts on the organism of the host embryo appeared in the Journ. Exp.
Med., July, 1916. It appeared after a number of,my papers and communica-
tions (Meeting at New Haven, Staff Meeting at the Rockefeller Institute) and
after numerous demonstrations to Dr. Murphy of mypreparations. Dr. Murphy
states in the above quoted note, that at his suggestion I undertook ‘‘the working ”
out of the finer histological details of the process,’’ discovered by him. This,
in view of the above mentioned facts, I venture to consider unwarranted.
Dr. Murphy writes the brief note “‘for completeness and record,‘‘ and gives a
reference of his previous work, in which “‘observations were made on the effects
of certain organ grafts on the embryo itself. Murphy, Jas. B., Journ. Exp.
Med. 1913, vol. 17, p. 482.’’ This paper however does not contain any obser-
vation on this subject. The only passage in this paper, referring to the effect
of the graft upon the embryo, reads: “‘Apart from the thin continuation of
the chick membrane, which covers the tumor and the ingrowth of vessels with
their scant accompanying stroma, there is no histological evidence of reaction on
the part of the embryo to the invasion of foreign tissue.’’ Nor is it possible to
find the slightest indication about the effect of the grafts in the body ‘tissues
of the host in any of Dr. Murphy’s previous papers. Through personal com-
munication from Dr. Murphy I knew about the enlargement of the embryo
spleen. By deduction from my previous hematopoietic work I reached the con-
clusion regarding the necessary coexistence of analogous changes in other hema-
topoietic organs, a conclusion which the results of the experiments undertaken
proved to be correct.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 265
thorough revision of the fundamental problems of hematology.
Indeed, if the various anlages of the hematopoietic organs were
equivalent and contained identical stem cells, hematopoietic
organs other than the spleen must have reacted also. Since
the structural peculiarities in the different organs of the embryo
are not fully differentiated, the reaction of the stem cells in
different hematopoietic organs could be expected to be more
homogenous. As shown further, the specific intervention was
indeed followed by changes both in the spleen and in all the
other hematopoietic organs and the reaction of these different
organs was substantially similar. These facts were mentioned
and demonstrated by me at the New Haven Meeting.’
The grafting of adult spleen on the allantois of the embryo is
a complex intervention. The grafted tissue contains small and
large lymphocytes, together with the so-called reticular tissue,
and with the vessels and their different layers, all belonging to
a full-grown organism. Which of these elements has to be re-
garded as the source of the stimulation could not yet be de-
fined conclusively. However, there is no doubt the intervention
applied introduces in the embryo heterogeneous substances.
It is known that the organism reacts to the introduction of
heterogeneous substances by a production of antibodies and to
the introduction of heterogeneous cells by production of the
so-called lysins. The erythrolysins which are developed by the
immunized animal have a dissolving power on the red blood
corpuscles against which the animal is immunized, the leuco-
2 Abstr. Proc. Anat. Record, Jan. 1916.
In this connection I desire to call attention to the substitution of a reference
to Dr. Murphy’s last communication (21) on p. 96 of No. 1, vol. 24, Journ. Exp.
Med. July, 1916, in my paper on “‘Differentiation . . . ” fora footnote, in
which I stated: ‘“‘I wish to express my indebtedness to Dr. J. B. Murphy, who
kindly demonstrated to me the method of grafting described in this paper.
Murphy, J. B. and Rous, P., 1912. The behavior of chicksarcoma
Journ. Exp. Med. vol. 15.” My paper, which was received for publication in
February 716, and published in July ’16 could certainly not contain a reference
to Dr. Murphy’s communication, which received for publication in May,
appeared also in July. This alteration seems to me the much the more inap-
propriate, because it has been made to apply to the demonstrations given by
me at the Anatomical Meeting, 1915.
266 VERA DANCHAKOFF
lysins stimulate the hematopoietic organs and induce intense
proliferation of their cells.
Antibodies are regarded as being specific and may influence
only cells in which they find adequate receptors. Dr. Flexner (17)
has shown an apparent lack of specificity both of certain kinds
of leucolytic sera to different hematopoietic organs and of differ-
ent leucolytic sera to one definite hematopoietic organ. The
spleno- lympho and marrow-lytic toxins, each of them acted in
a stimulating manner upon all the hematopoietic organs,—hence
the antibodies of the leucolytie sera evidently found cells with
adequate receptors in all hematopoietic organs. These cells
under the influence of certain amboceptors responded by com-
mon proliferation in different hematopoietic organs.
The results of the experiments cited offer a further corro-
boration of the monogenetic conception of blood development.
The histological studies established in all the hematopoietic
organs—the presence of cells, of which the morphological struc-
ture seemed to indicate a great potency for differentiation and
proliferation. The embryogenetic studies pointed out these
cells, as the true stem cells, common to all the hematopoietic
organs and endowed with faculty to intense polyvalent differen-
tiation. Finally, studies on stimulation of the hematopoietic
organs by agents, which were supposed to be specific, seemed
to indicate in all the hematopoietic organs the presence of
cells, which respond to each of these stimuli by a common
proliferation.
The experiments used in the present work are closely related
to the experiments referred to above. The present study,
however, is connected exclusively with the stage of antibody
production. The requirement of heterogenicity of tissues might
be found in the differences of the tissues in the adult organism
and the embryo. (Even the morphological structure of the same
kind of cells, for example the hemocytoblasts, changes some-
what with age, the cells undergoing an ontogenetic development.)
The intense proliferation, exhibited after the appearance
in the embryo of heterogeneous substances by all the embryonic
hematopoietic organs seems to indicate that different hemato-
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 267
poietic organs in the embryo all contain elements, which react
in analogous manner to the appearance of these substances.
The fact that the proliferating cells exhibit a more homogene-
ous differentiation in the various embryonic organs than in
adult organism may be accounted for by absence of specific
structural environmental conditions in the embryonic organs.
Though chiefly concerned in histogenesis of the blood cells, the
present paper may have a close bearing on the immunity
problem in establishing a connection between introduction of
heterogeneous substances and reactions exhibited by hemato-
poietic tissue.
The grafting of an adult spleen on the allantois of the embryo
produces changes similar to those described in myeloid meta-
plasis, and in principle are similar in different hematopoietic
organs. However, these changes seem to be differently ex-
hibited in hematopoietic organs at different stages of embryonic
development. The stimulating agent may differently influence
various cell groups in the hematopoietic organs at different
stages of their development and leads often to the appearance
of peculiar pathological processes. There appear in different
organs characteristic changes depending upon their structural
peculiarities which will be studied and described consecu-
tively. As the most conspicuous phenomenon is the appear-
ance of an enormous hypertrophy of the spleen, therefore a
study of this organ will give a basis for further investigation
of the changes undergone by other hematopoietic organs, in-
cluding the mesenchyme of different parenchymal organs. <A
study of changes in the circulating blood of the embryo, as well
as a study of the growth and differentiation of the grafts them-
selves will follow.
2. METHODS OF INVESTIGATION AND TERMINOLOGY
It has been shown in the introduction that the usual methods
of investigations applied to the problem of the origin and the
mutual relationship of blood cells failed in the attempt of solving
them. There must be found new methods for the elucidation
268 VERA DANCHAKOFF
of the problems so unsuccessfully discussed year after year.
Such a new method for study of hematological and other biologi-
cal problems may be offered by the study of transplantations of
tissues on the allantois of an embryo and also in the study of the
changes, which occur in the tissues of the host after grafting.
The influence of identical environment upon different hemato-
poietic organs may be easily tested by this method. Trans-
plantations on the allantois of embryos were used by Murphy
and Rous (30) in their studies of transplantability of tissues to
the embryo and by Murphy (29) in his study of the factors of
resistance to heteroplastic tissue-grafting. Transplantations of
adult spleen and bone-marrow seemed to supply the embryo
with a refractory mechanism against heteroplastic grafting,
which in a normal embryo is lacking. These transplantations
as told are followed by a considerable enlargement of the host
spleen.
Every theory is a deduction of a lmited number of facts,
but if the theory is true, it must apply to all analogous. cases.
The enlargement of the embryonic spleen, mentioned above,
which soon was discovered to be a true hypertrophy, could not
be explained from the standpoint of the monophyletic school
as an isolated process. The monophyletic school, if true, had
to assume that changes in the embryonic spleen were accompanied
by analogous changes in other hematopoietic organs. Since
the hyperthrophy of the embryonic spleen has been involved
by a considerable proliferation of the stem cells, stem cells
in other hematopoietic organs admitted by the monophy-
letic school equal for all, must have been affected also and must
have proliferated. Since the structural environmental con-
ditions in the embryo are less differentiated than in the adult
organism, the reactions in the different embryonic hematopoietic
organs may be expected to be more homogeneous. As will be
seen later, this assumption, resulting from the premises of the
monogenetic conception of the blood origin, has been fully con-
firmed by the results of the experiments.
In principle similar to the tissue cultures the grafting method
has a great advantage over them. The allantois offers for the
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 269
transplanted tissue ideal conditions in supplying the growing
tissue with nutritive material and also in withdrawing the pro-
ducts of the metabolism of the developing and differentiating
cells. Though the tissue growth and differentiation cannot
be observed right under the microscope, this is compensated by
the facility of obtaining abundant experimental material in
different development stages.
As Murphy and Rous (80), I used the method developed
by Peebles for studies in experimental embryology. Since I had
to overcome many difficulties, and failures were due sometimes
to apparently minute details, I will undertake a thorough de-
scription of the culture-method on the allantois of the chick
embryo, as I used it in my experiments. I am indebted to the
kind efforts of Mr. Ebeling of the Rockefeller Institute for being
supplied through the whole winter with fertilized eggs, which
developed in the incubator in the number of 60 to 70 per cent.
The cultures succeed on the allantois easily from the beginning
of the 7th day of incubation. The cultures of the hematopoietic
organs grow if in contact with the mesodermal surface of. the
allantois. The ectoderm, which overlies the area vasculosa
and the yolk does not offer favorable conditions for cultures of
hematopoietic tissue. From the 7th day of incubation, however,
the allantois appears as a well-sized sac, flattened under the egg-
shell.
It is easy to determine by means of illumination, which of
the incubated eggs started to develop. The localization of the
embryo body as well as that of the allantois with its vessels
comes thereby clearly out. Good places to choose for grafting
are regions between the junction of two vessels at a distance of
1 to 2 em. from the embryo body. The illumination and the
provisional marking of the eggs must be done quickly and the
eggs immediately returned to the incubator. A sawing out of
small windows in the region marked on the eggshell follows. It
is advisable to saw the windows in the form of a trapezium,
which form allows to orientate it easily, when the window has
to be closed. A great care must be observed in sawing the win-
dows out of the eggshell. The pressure of the instrument has
2
270 VERA DANCHAKOFF
to be slight, otherwise the shell cracks easily outside the region
marked. Splits may be stopped by paraffine and the egg still
used. As instruments for sawing, small scalpels used in ophtal-
mologie may be recommended, a few jaggs on their edge are
often useful. An experienced hand will easily determine the time
when the eggshell is passed through. Then the egg is again
returned to the incubator. After all the eggs are opened instru-
ments for aseptic extraction of organs, (a few scalpels, scissors,
forceps and bone cutter), and instruments for grafting, (tissue
crusher—it is a syringe with a bolt-bottom, another syringe
with divisions of 0.1 gram, and a needle, a fine forceps and two
pairs of scissors), are sterilized.
All the next work has to be done quickly. The organ, asepti-
cally extracted, is cut in small pieces and passed through the
tissue crusher, the tissue mash is then pulled into the syringe
and is ready for grafting. Now egg after egg is taken out of the
incubator, the shell window is removed, the shell membrane is
lifted by the small forceps and an opening cut out with scissors.
The allantois becomes visible and often in earlier stages falls off
somewhat fromthe eggshell. Through the needle 0.1 to 0.2 of the
tissue is then pushed in, introducing the tissue if possible under
the eggshell at least under the shell membrane. Great care must
be taken in grafting eggs in advanced stages, because the allan-
tois then bleeds intensely, and the contact of the transplanted
tissue with the allantois is removed by the extravasat. No
graft usually takes, if the grafted tissue is introduced into the
cavity of the allantois or deeper. In these cases the tissue is
found floating in the form of a greyish mass, containing numer-
ous small grains.
After the tissue has been implanted on the surface of the al-
lantois, the window is closed by the piece of the eggshell with-
drawn and the splits are covered by paraffine. The paraffine
of a higher melting degree is preferable, in order to prevent its
melting during the following incubation of the eggs. If a local
graft is desired any mechanical disturbance should be avoided,
otherwise a more diffuse distribution of the grafted tissue is
easily obtained. The graft takes more successfully if the egg
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 271
during the first twelve hours is put in the incubator, the window
down. This secures a closer contact of the tissue introduced
with the surface of the allantois. Next day I usually changed
the position of the egg in directing the window above. The
strict observance of the directions given secures usually 80
per cent of successful cultures. Interesting results in the form
of diffuse growth of transplanted tissue were obtained by intro-
ducing an emulsion of the tissue mash between the allantois and
the chorion under the eggshell. The whole allantois appeared
in a few days covered by innumerable small grafts. The same
results were occasionally observed also after applying the usual
method of grafting.
Twelve hours after transplantation, the tissue is usually in
firm contact with the allantois. Later it is attached by numer-
ous vessels growing from the allantois into the tissue. For fixation
of the material, it is advisable to free and fix first the embryo
after a ligature of the vasa umbilicales is done; then to cut out a
large piece of the shell in the region of the culture. The allan-
tois which covers the shell is transported in the fixing fluid
together with the shell and is removed from the shell 5 to 10
minutes later. The fixation of the allantois and of the graft is
completed in 4 to 1 hour in Zenk.- formol. The celloidin was
successfully substituted by the parloidin Dupont, and I may
recommend this product as being in no way inferior to the
celloidin Schering. The imbedding in ecelloidin or parloidin
remains a sine qua non for hematological work, what easily
was deduced by a study of a few specimens imbedded in paraf-
fine last autumn (1915) when no more celloidin was available.
Since the staining of preparations attached to the slides and
freed from parloidin is more effective, it may -be permitted to
recall the method of Rubashkin (36), somewhat modified by
Danchakoff (9, ’08d).
The greatest part of the material was stained by eosin-azur
and some of the preparations by Dominici and Pappenheim.
For the staining of fibrous tissue the iron-hematoxylin and
subsequently van Gieson were used. Most of the illustrations
are given in black on account of the special temporary conditions,
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 3
22 VERA DANCHAKOFF
and one colored plate is added, as example of the preparations,
from which the ink drawings were made. The photographs on
plates 1 and 2 were kindly made by Mr. Schmidt of the Illus-
tration Department of the Rockefeller Institute.
Since the nomenclature in the hematology has become nowa-
days extremely complex and often under one name different
cell units are understood, or even oftener one cell unit is termed
by different names, it is useful to state in advance what termin-
ology will be used in this paper.
The stem cells of the different blood elements, which first
appear after isolation of the blood-islands, have the structure
of the well-known large lymphocytes. The term of large lymph-
ocytes was applied for these cells by Pappenheim (33), Dan-
chakoff (9) and Maximow (25). Studies of hematopioetic tissue
led to recognize everywhere cells of the structure of large lym-
phocytes and to identify them as stem cells for the blood tissue.
Only few differentiation potencies were first assigned to the large
lymphocytes. This however corresponds but little to the various
differentiation potencies, exhibited under different conditions
by this cell. So the name of large lymphocyte seemed to cor-
respond little to the given cell in its new conception. Since
personal studies did not give me any data, bestowing all the
lymphatic cells with equal potencies, the less appropriate seemed
to me the name of large lymphocyte in connection with the
stem cell for different blood elements. In my last papers I
called the stem cell, which in itself is a good name, lymphoid
hemocytoblast—lymphoid in order to take into consideration
its morphological structure, hemocytoblast on account of its
potencies to differentiate into various blood cells. The same
name will be used throughout the paper. The names of erythro-
blasts and erythrocytes do not require any explanation.
The names of myeloblast and myelocyte seem to me unfitted
for the purpose used. This name is wrongly applied to cells
which neither appear first in the bone-marrow, nor are cell
units exclusively characteristic of this organ at any time of its
existence. ‘These terms will be substituted, as in my previous
papers, by granulocytoblasts and granulocytes or leukocytes.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 273
However, the term myeloid tissue, or myeloid metaplasis, is
used throughout the paper as a collective name, under which
all the characteristic cell elements of the bone-marrow are under-
stood, it is the erythroblastic, granuloblastic tissues, and re-
spectively the megakaryocytes. The terms promyelocytes,
metamyelocytes, mikromyelocytes, which correspond to inter-
mediate stages between a lymphoid hemocytoblast and a leuko-
cyte are omitted. These stages are characterized by unessential
features and are often overstepped in embryonic life during
intensive regeneration. ‘The use of so many terms for expressing
small differences between development stages in a cell lineage
seems more to confuse than to help. For this reason I did not
introduce them lately in the scheme given in the Anatomical
Record, and in the present paper, only terms which designate
definite morphologically well defined stages will be used, and
these are lymphoid hemocytoblast, granulocytoblast, and granu-
locyte or leucocyte.
The reciprocal relations in the lymphatic cell group are some-
what. more obscure. The different lymphatic cells are looked
upon by Maximow (25) and Weidenreich (44) as being merely
temporary appearances of young undifferentiated cells, all
characterized by the same differentiation potentialities. Though
these authors admit a specific morphological structure for the
large and the small lymphocytes and the histogene wander cell,
yet they assume that these cells may easily change reciprocally
their structure according to the environmental conditions. In
birds and reptiles (Danchakoff (9)) as well as in mammals
(Maximow (25)) it is easily demonstrated that small lympho-
cytes may both proliferate and differentiate further, but their
lines and products of differentiation are not identical with those
of the large lymphocytes (Pappenheim (32), Danchakoff (9) ).
Neither is it proved definitely that the small lymphocytes may
grow into the large. Nor is a possibility of erythrocyte develop-
ment at the expense of small lymphocytes shown to exist in
birds and reptiles, as Freidsohn (14) admits lately for amphib-
ians and Venzlaff (48) for birds. Therefore, under the name
of small lymphocyte, cells characterized both by a definite
274. VERA DANCHAKOFF
morphological structure and by well defined differentiation
potencies will be understood.
The histogene wander cells, which term I substituted by
histiotopic wander cells seem to be in close relation to the hemo-
cytoblasts and often are merely an intermediate but morphologi-
eally well defined stage of development between a mesenchymal
cell and a hemocytoblast.
I wish to express my thanks to the director of The Wistar
Institute of Anatomy and Biology, Dr. M. J. Greenman, and to
the Staff of the Institute, for the generous hospitality shown to
me. I am indebted to Dr. 8. Flexner for the kind admission
to the laboratories of the Rockefeller Institute for Medical
Research where the work has been partly done; to Dr. C. E.
McClung for the great interest shown in my work and for the
revision of a number of my preparations; to Dr. F. P. Mall for
the revision of the text and the proofs during my absence.
3. HISTOGENESIS IN THE SPLEEN IN RELATION TO STRUCTURAL
ENVIRONMENT
A. HISTOGENESIS IN A NORMAL CHICK SPLEEN
The study of histogenesis in the spleen by the method of
stimulation of the stem cells in the spleen anlage requires a
thorough knowledge of the histogenetic processes, which nor-
mally take place in the spleen. The early stages of spleen devel-
opment in birds were studied by Tonkoff (41). Yet no investi-
gation of embryogenesis of the characteristic spleen elements
was made by modern technique. The general lines of differenti-
ation of the spleen elements were studied by Danchakoff (’16a)
in Tropidonotus natrix, but the question—what are the con-
ditions which determine the differentiation of the polyvalent
stem cell—remained unsolved. It seems, therefore, necessary
to make first a study of normal spleen development in the chick,
and to attempt to determine the conditions which imply the
various differentiation of the stem cells.
The appearance of the spleen anlage in the chick embryo and
its first development corresponds to the description given by
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 275
Tonkoff (41) in 1900. The spleen anlage appears in an embryo
nearly 4 days old—in the mesenterium dorsale duodeni in the
region of pancreas dorsale. It is distinetly separated from the
coelomic epithelium, which surrounds it. In the I stage of its
appearance the spleen anlage is purely mesenchymal. It is
distinguished from the surrounding mesenchymal tissue by the
denser appearance of the tissue. At this time mesenchymal
cells of the anlage present short ramifications, which soon are
lost, the cellsmultiplying intensely and joining finally in a common
syncytium (fig. 5 and 6). The spleen anlage at early stages is
identified rather by its localization than by the character of its
cells. Similar agglomerations of syneytium-like mesenchyme are
encountered in many other places, and their further development
exhibits a great analogy with that of the spleen anlage, resulting
in a differentiation into lymphatic tissue. The more intense
proliferation of the mesenchyme in the region of the spleen is
evidently due to local favorable conditions. The development
of the spleen at the expense of mesenchymal cells without any
relation to the endoderm nor to the coelomic epithelium may be
regarded as a well-founded fact.
The appearance in the spleen anlage of numerous ameboid
cells, the lymphoid hemacytoblasts, and their differentiation
into granulocytoblasts and into small lymphocytes was described
by Danchakoff (9) (16a), in Tropidonotus natrix. The fact
that the small lymphocytes develop in the spleen in later stages
was also noticed. However, it is improbable that the different
stages in themselves should be accounted for as differentiating
factors. If the small lymphocytes do appear later, the granulo-
cytoblasts nevertheless continue to differentiate in the spleen, at
leastin embryonic life. Conditions for both lines of differentiation
must therefore coexist in later stages. The differentiating
factors should be sought rather in different structural conditions,
appearing at definite stages, and determining from the time of
their appearance the lines of differentiation of the polyvalent
stem cells. .
Since the structural peculiarities are exhibited in a more
striking manner in the spleen of an adult or a young chick, it
276 VERA DANCHAKOFF
may be advantageous to demonstrate them by a study of a
fully developed spleen, and then to attempt to find out whether
the gradual development of the peculiarities influence the his-
togenesis of the hematopoietic tissue. A distinctive feature in
the spleen structure is given by its special vascularization. The
regions with veinous and arterial vascularization, though they
penetrate each other, remain nevertheless independent, and
communicate together only in places, where the white pulpa ©
passes into the red. The most characteristic cell element of the
spleen—the small lymphocyte—belongs to the white pulpa and
accumulates here in the form of follicles and follicular strings.
The granulocytoblasts, though not numerous in the adult spleen,
are chiefly localized in the pulpa. Other ameboid elements, the
basophilic large lymphocytes (lymphoid hemacytoblasts), and
mononuclear leucocytes, often in the form of macrophages, are
common to all the regions of the spleen. The syncytial cell
reticulum is also ubiquitous; it forms in the red pulpa wide
meshes. In the white pulpa the cells of the reticulum appear
denser, and the meshes formed by their ramifications are smaller.
As a further study of the spleen development will show, the
chief characteristic feature, primarily, determining different
regions of the spleen as red or white pulpa, consists in the type
of vessels by which a region of the spleen is supplied rather than
by the presence of certain kind of ameboid cells. The wide
veinous capillary net together with sinuses and lacunae forms the
red pulpa, the bunches of arteries resolving themselves into a net
of narrower branches—belong to the white pulpa. The question,
whether the specific differentiation of the ameboid elements
depends upon the peculiar vascularization of the spleen may be
decided after a study of the spleen development. The chick
spleen is a favorable subject for elucidation of this question,
for the identification of the vessels is easy in the spleen anlage
from the time of their appearance. .
As mentioned above, the development of the spleen in the
earliest stages is characterized by its loose mesenchymal structure.
The intense cell proliferation leads soon to a transformation
of the loose mesenchymal anlage into a denser syncytium.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 277
Numerous basophilic cells differentiate at the expense of the
syncytium and become ameboid (figs. 5 and 6, LZ. Hbl.). Acido-
phylic granules appear in the cytoplasm of a part of these cells
and characterize them as granulocytoblasts (fig. 8, Grbl.). The
development of these cells (ymphoid hemocytoblasts and granu-
locytoblasts) is not specific for the spleen mesenchyme and is
observed in other regions of the embryo body also. At the same
time the spleen anlage becomes vascularized, and its vasculariza-
tion is at this stage of its development exclusively veinous. In
the peripheral layers of the anlage, later on through the whole
organ, appear splits, which evidently are filled by a liquid,
which separates the cells. These splits are at first surrounded by
the irregular surface of the mesenchymal cells. Some of the cells
show still their processes, projected in the lumen of the sinuses.
(fig. 6, S.). These sinuses soon join together and form a net.
From the other hand a communication with branches of the
intestinal veins is established. The whole mesenchymal anlage
exhibits at this second stage of its development a spongious
structure. Whether the appearance of the splits in the anlage
is due to a secretion of the surrounding cells, or to a transudation
of a liquid through the vessels growing from outside is difficult
to determine. There is, however, no doubt that the splits
mentioned are of local origin. This has been shown by Laguesse
(23) in fishes.
The appearance of the splits in the tissue of the spleen anlage
- is accompanied by more intense isolation of ameboid cells
(fig. 6, L.Hbl’’). Some of them are surrounded by a developing
sinus and become situated in its lumen. These cells have
invariably first the structure of lymphoid hemocytoblasts (large
lymphocytes). Similar cell groups are not seldom encountered
in the larger sinuses of the peripheral layers in the spleen. (fig. 5,
S). As soon, however, as these lacunae unite with the veinous
vessels, what is indicated by the sudden appearance of differ-
entiated erythrocytes within the lacunae, the lymphoid hemo-
cytoblasts begin here their differentiation into erythroblasts (fig.
5, Hrbl.). The plasma of the blood must evidently contain
factors for differentiation of lymphoid hemocytoblasts into ery-
278 VERA DANCHAKOFF
throcytes. The slowness of the blood current in the large
veinous capillaries and sinuses offers moreover favorable condi-
tions for this line of differentiation.
The vascularization of a normal spleen proceeds, however,
gradually. The cells, surrounding the splits, become gradually
flattened and finally form an even endothelial surface (fig. 8).
The sinuses of a normal spleen contain usually merely a few
young cells undergoing an erythroblastic differentiation. I do not
think it right therefore to consider the normal embryonic chick
spleen as an active erythropoietic organ, though potentially it
must be considered as such. It may be noticed that spleens of
embryos at the same stage may offer in this respect well pro-
nounced individual differences. New sinuses continue to appear
with the growth of the spleen. In later stages, at the 11th
day of incubation, as the fig. 7 shows, in newly formed
sinuses there may be. found large groups of lymphoid
hemocytoblasts, which soon undergo an erythroblastic differ-
entiation. Between the new formed vessels the mesenchyme
continues to proliferate and to split off lymphoid hemocytoblasts.
They multiply also and partly differentiate into granulocytoblasts,
which increase also their number by mitosis and partly differen-
tiate further into granulocytes (figs. 8 and 9).
The processes of growth and differentiation proceed slowly in a
normal spleen. In an embryo of 9 days—5 days after the begin-
ning of its development—the size of the spleen reaches 1—1.5mm
the long diameter. The abundance of nuclei however may serve
as index of: intense proliferation processes taking place in the
syncytial tissue of the embryonic spleen. The spleen during its
II stage of development is characterized by a development of a
net of wide veinous capillaries developed in the mesenchymal
syncytium, by an intense granulopoiesis outside the vessels and
by a potential erythropoiesis within the vessels. These lines
of differentiation, as known, are also characteristic of the mye-
loid metaplasis in the spleen pulpa. The study of the normal
spleen development leads to the conclusion that the first differ-
entiation processes of the mesenchymal spleen anlage transform
it into a pulpa-like organ (figs. 7 and 8). The spleen remains
pulpa-like until the 12 to 13 day.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 279
The intense development of the arterial vascularization begins
at this time and the spleen enters in the III phase of its develop-
ment, characterized by the appearance of follicles. A cursory
glance on spleens during development of arteries brings forth a
striking difference between the development of the veinous and
of the arterial vascularization. The arteries and their smaller
branches never appear as irregular splits limited by mesenchy-
mal cells. These vessels on the contrary always develop as reg-
ular narrow tubes (fig. 9, Art.c.) I did not intend to undertake a
special study of the vascularization of the spleen, therefore I did
not apply special methods of investigation for this purpose.
However, a thorough study of preparations may give some infor-
mation upon the development of the veinous and arterial vascu-
larization. The veinous sinuses and capillaries are local forma-
tions, the arteries and their branches seem to grow into the
spleen from outside and here ramify by budding (fig. 9).
The arteries, growing into the pulpa like tissue of the spleen,
divide it in regions, which are soon further subdivided by smaller
arterial ramifications. The arteries lie first in the mesenchyme,
where numerous granulocytoblasts are present (fig. 9). The
mesenchymal cells continue to proliferate around the arteries.
The process of splitting off of lymphoid hemocytoblasts persists
in these regions, but their differentiation into granulocytoblasts
is suspended under conditions in which the mesenchymal cells
develop around the narrow arteries. These conditions do not
correspond to those which prevail around the thin walled veinous
sinuses. The arteries and their smaller branches finally become
surrounded by clear zones of mesenchymal cells (fig. 10), which
markedly contrast with the granuloblastic tissue. These zones
appear in preparations in the form of islands of mesenchymal
tissue, which fill up all the interstices between the arterial
vessels and the pulpa like tissue of the spleen anlage. These
islands are anlages of follicles.
At the time of the intense development of the mesenchyme
around the arteries a new line of cell differentiation may be
traced in the spleen which soon will become predominant; this
is the differentiation of small lymphoctyes. The figures 12
280 VERA DANCHAKOFF
and 13 show the development of small lymphocytes in the
follicles. From right to left the figure 12 represents the tissue of
the follicle from the periphery to its center. In the peripheral
parts of the follicle an intense isolation and proliferation of
lymphoid hemocytoblasts takes place (5 mitoses in a part of the
microscopical field) which leads to the formation of dwarf
hemocytoblasts (fig. 12 S.L.H6l.). A process of differentiation,
starting in the groups of small hemocytoblasts (fig. 12) soon
transforms them into true small lymphocytes (S.Lmc.). The
cytoplasm around the nucleus of these cells becomes smaller, the
typical nucleolus of the hemocytoblast is replaced by chromatin
particles which may be now permanently discovered in the nucleus
(fig. 12, S.Lme.). A similar development of small lymphocytes is
shown in the figure 13 in the region adjacent to the red pulpa
(right edge of the drawing). Both in the red and in the white
pulpa numerous larger and smaller hemocytoblasts are present
and offer in both regions similar morphological structures.
In both regions they sometimes appear as groups of cells joined
together,—probably an index of their syncytial origin. The
further differentiation of the lymphoid hemocytoblasts is shown
to be different, aecording to their localization. In the red pulpa
they are in close contact with the larger veinous capillaries and
differentiate into granulocytoblasts. In regions with scarce ar-
terial vascularization they undergo a differentiation in, small
lymphocytes.
As in the thymus, the differentiation of small lymphocytes
is preceded in the spleen also by development of generations of
small-sized hemocytoblasts. Their appearance in the thymus
seemed to depend upon an intense proliferation of the cells in a
limited space. If the cells in the spleen are not heaped up, as
they are in the thymus, yet the analogy of the conditions for
their differentiation in both regions may be easily traced. Both
in the thymus and in the spleen the small lymphocytes develop
under conditions of poor nutrition. They appear in regions of
the spleen where the swift blood current passes by the narrow
arterious vessels.
The first groups of small lymphocytes differentiate between
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 28]
the 15 to 17 days of incubation. They increase their number by
differentiation at the expense of lymphoid hemocytoblasts and
by their own proliferation. ‘They gradually infiltrate the regions
with arterial vascularization; finally they accumulate in dense
masses around the arteries. One may observe amongst them a
few degenerated cells, even in an embryonic spleen; they are
phagocytosed by mesenchymal and endothelial cells and some-
times even by hemocytoblasts. A number of mesenchymal and
endothelial cells are gradually transformed into typical macro-
phages (Evans) (15). Similar macrophages are observed also
in the lumen of the sinuses, where they lie free, or form a part
of the surrounding mesenchymal tissue (fig. 17, 18, 19). In the
pulpa, however, the activity of the macrophages is directed chiefly
toward the degenerated erythrocytes. Though at this time the
connection between the arterial and the veinous vessels is com-
pleted, these regions continue to be very distinct.
The appearance of the follicles and of the new line of cell differ-
entiation does not seem to influence the life of the cells in the
pulpa. Both the granuloblastic and the lymphoblastic processes
of differentiation coexist now in the spleen, and are displayed by
lymphoid hemocytoblasts in different regions under different
structural conditions. The granulo- or leukopoiesis develops
around the large veinous sinuses under conditions identical to
those under which they develop in the yolk sae and in the bone-
marrow. ‘The leukopoiesis in the spleen is much reduced after
the embryo is hatched, and the spleen becomes chiefly a lympho-
poietic and erythrolytic organ. .
The structural peculiarities which determined the various lines
of differentiation of the polyvalent stem cells remain in the
adult spleen unchanged. Different stimulating agents may cause
a proliferation of the stem cells. Their differentiation, however,
will correspond to the structural environmental conditions to
which they are submitted. It is therefore only natural that
the myeloid metaplasis in the adult spleen should develop
in the pulpa, where the structural conditions determine
normally a granuloblastic differentiation of the hemocytoblast.
A differentiation of small lymphocytes in the pulpa, or vice
282 VERA DANCHAKOFF
versa, could be expected only in the case if the structural con-
ditions of the pulpa be changed and corresponded to those in
which the small lymphocytes normally develop.
The study of the normal development of the spleen in the
chick allows the following conclusions:
1. The anlage of the spleen develops from the mesenchyme of
mesodermal origin. It appears at the end of the 4th day of
incubation. It consists first of a loose mesenchymal tissue,
which gradually becomes denser and finally is transformed into
a syncytium.
2. The development of a veinous vascularization transforms
the spleen anlage into a homogeneous pulpa-like tissue. This
stage is characterized by a well developed granulopoiesis and a
potential erythropoiesis. ‘The stem cell for both directions of
differentiation is the lymphoid hemocytoblast, which develops
at the expense of mesenchymal cells. The line of the stemcell
differentiation depends upon the conditions to which the stem-
cells are submitted.
_ 3. The development of the follicles and of the lymphoid tissue
coincides with the development of the arterial vascularization.
The small lymphocytes develop at the expense of the lymphoid
hemocytoblasts and their differentiation is preceded by appearance
of generations of small-sized hemocytoblasts. The small lymph-
ocytes in the spleen, as well as in the thymus, appear in regions,
characterized by special structural conditions of the environment.
Different kinds of cells appear, as final results of differentiation
processes, observed in the spleen. Reticular tissue cells, and
endothelial cells, erythrocytes, granulocytes, small lymphocytes
and macrophages develop in the same organ. Can anything
definite be told on the basis of the study of normal spleen devel-
opment about the origin of these cells? Are all the cells of the
mesenchymal spleen anlage identical and polyvalent, their further
development being determined by the structural environmental
conditions, or though apparently identical in morphological
structure their potencies to different development are predeter-
mined by intrinsic imperceptible differences between them?
The general outlines of the development of the spleen anlage
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 283
under normal conditions should be regarded as _ necessarily
determined. The organ develops in a complex environment of
other organs which proliferate and differentiate at the same time.
Each of these factors, which necessarily ensues from the collective
action of all the surrounding conditions, becomes itself one of
the factors which partly impels the results of the general differen-
tiation. The necessity and mutual dependence of the develop-
ment processes result from reciprocal influence of the factors and
is characterized by striking purposefulness. The general out-
lines of the development may therefore be changed, as will be
shown later, only in the case if the regular interaction of factors
has been disturbed. :
Variations in the intensity of differentiation of various cell
groups may be observed, however, in normal spleens. Numerous
or merely scant lympho- and granulocytoblasts may develop at
the expense of mesenchymal cells. The proliferation of granu-
locytoblasts may also be more or less intensive. But this differ-
entiation process is extended in the spleen anlage more or less
uniformly at a time when the anlage presents a homogeneous
structure and is lacking special conditions of vascularization.
The differences in the extent of the normal granulopoiesis are,
however, not excessive and deviation in other lines of differen-
tiation may compensate them. A new line of differentiation
necessarily starts with the appearance of new structural
characters, namely with the development of the arteries. After
this new factor has been established, it will permanently
influence the cells, which stand under its control. The existence
of regular unchangeable relations between definite structural
conditions and differentiation of certain kind of cells seems to
explain sufficiently the necessity which appears in a group of |
identical stem cells to diverge in their further development. Is
there a need of recurring to invisible differences between cells,
lest it should be necessarily required by deduction from the
results of development?
One phase in the spleen development could, however, be inter-
preted by the dualists in their favor—this is the apparently late
ingrowth of the arteries in the pulpa-like spleen anlage. If both
284 VERA DANCHAKOFF
arteries as well as veins developed in loco, there would remain
no doubt in the common origin of the lymphatic and the myeloid
tissues. The ingrowth of the arteries may be regarded as a mere
stimulus for a new intense proliferation of the mesenchyme and
its subsequent differentiation into small lymphocytes. On the
other hand, the ingrowing vessels might have brought new cell
material for a new line of differentiation, which starts at this
period. It is difficult to establish with certainty whether the
mesenchymal cells of the follicle anlages develop at the expense
of the local mesenchyme, which remains undifferentiated, or
whether they are derived from the tissue brought by the
growing arteries. No matter how this question will be solved,
both in the pulpa and in the follicles, the mesenchymal cells
bear the same morphological structure. They present moreover
in their first differentiation stages undeniable analogies. In
the pulpa as well as in the follicles they partly differentiate into
lymphoid hemocytoblasts and partly form the so-called reticular
tissue. The special differentiation according to environmental
conditions is then undergone by the lymphoid hemocytoblasts.
The dualists admit that the development of the myeloid
metaplasis of the spleen, which under definite pathological
conditions is localized in the pulpa is due to proliferation and
differentiation of adventitial or endothelial vascular cells. Endo-
thelial cells are present both in the pulpa and in the follicles,
the adventitial cells are even more numerous in the arteries of
the follicles. If the strain in the further differentiation of the
lymphoid hemocytoblast has to be laid in the cell itself and not
in the physico-chemical conditions, given by structural peculiar-
ities of the environment, the myeloid metaplasis should have ap-
peared as a diffuse process, for endothelial and adventitial cells
are found in the pulpa as well as in the follicles. In order to
remain consequential, the dualists ought to recognize various
kinds of adventitial and endothelial cells—the ones being capa-
ble of a myeloid differentiation, the others not.
The striking dependence of the various differentiation of
morphologically identical cells upon a change in the environ-
mental conditions may greatly strengthen the monogenetic
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 285
conception of the blood origin. It cannot, however, be con-
sidered as definite proof, because various differentiation products
could under different conditions derive from isomorphic but
heteropotential cells. The study of the spleen development
under influence of stimulation of its stem cells will give further
information about causal relation between structural] environ-
mental conditions and development of the polyvalent stemcells.
B. HISTOGENESIS IN A CHICK SPLEEN AFTER STIMULATION
An intense stimulation of the mesenchymal spleen anlage is
obtained by means of grafting a tissue mesh of an adult spleen
on the allantois of a chick embryo. The photographs on figures
1 and 2 demonstrate the enlargement of the host spleen. Both
photographs belong to an embryo of 18 days. The spleen is
enlarged in both embryos. While the spleen on figure 1 as
compared with a normal is enlarged approximately 4 times the
spleen on figure 2 shows a much more intense enlargement. The
long diameter of the spleen at this stage is about 1-2 mm. After
the stimulation it may exceed 1 em. It is remarkable that the
enlargement of the spleen is usually in a visible relation to the
intensity of the graft growth. The spleen a corresponds to the
graft on figure 3, the spleen b to the graft on figure 4.
The macroscopical appearance of the enlarged spleen does not
give any indication whether the spleen enlargement is due to a
local proliferation of the embryonic tissue. This proliferation
could be incited by heterogeneous products of metabolism, which
from the transplanted and growing cells could penetrate in the
vessels and be transported by the blood current into the embryo
body. On the other hand, cells from the graft could be trans-
ported in the spleen and here under favorable conditions pro-
_iliferate. Tumor like accumulations of tissue would develop
in the last instance and enlarge the spleen. The surface of the
spleen after stimulation is often covered by protuberancies,
numerous small whitish points may be seen on its section. ‘These
facts at first suggested rather the idea that the enlargement of the
spleen was caused by tumor-like metastases brought by the blood
286 VERA DANCHAKOFF
current from the growing graft. However, the systematical study
of the gradual enlargement of the spleen urged to finally admit a
growth of the embryonic spleen tissue in loco. This local growth,
at least in the beginning, seemed to be caused by an intense uni-
form stimulation of the mesenchymal cells and of the lymphoid
hemocytoblasts and their further differentiation. After the
hypertrophic character of the enlargement of the spleen had been
established, it was proposed on the basis of this fact to test
the validity of the monogenetic conception of the blood origin.
If true, the monophyletic interpretation implied analogous
changes in other hematopoietic organs as seen in the spleen.
The chief character of the changes observed in an embryo,after
grafting of an adult spleen, consists indeed in a stimulation of the
whole hematopoietic tissue. The hematopoiesis, stimulated ex-
perimentally follows strictly the fundamental principles, estab-
lished for birds and reptiles during embryonic and adult life. The
hematopoiesis develops however in a peculiar way; first in that
groups of cells, which normally would slumber indefinitely,
become involved in the process; secondly, that an outburst of
hematopoiesis is incited at a time when the normal hematopoiesis
is conveyed in definite channels; and finally, that various direc-
tions of differentiation are displayed by hemocytoblasts in places
where they are usually absent. The study of the changes in the
spleen after grafting will form a basis for a comparative study
of the changes, incited by the same intervention in other hem-
atopoietic organs.
Since the stimulation leads to different changes according to
the stage at which it has been applied, the results of the stimu-
lation in early and later stages will be described separately.
As mentioned, grafts of hematopoietic tissue on the chick al-
lantois take easily, if grafted from the end of the 6th day of
incubation and on. The grafts, made before the establishment _
of arterial vascularization in the spleen, incite in the spleen
anlage uniform proliferation, which may considerably vary in
intensity. The changes are different if the stimulation has
been applied after the arterial vascularization has taken place.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 287
a. Changes after stimulation in early stages
Transformation of the mesenchymal spleen anlage into a uni-
form granulobastic tissue. The study of the normal spleen
development has shown that the spleen anlage at the stage of
6 to 8 days consists of a more or less dense mesenchymal tissue
in the form of syncytium. The oblong nuclei, characterized
by the presence of well pronounced nucleoli and scant chro-
matic particles lie closely together. At the periphery of the
organ the nuclel are more scattered and the tissue looser.
The limits of the cells are here also undefinable, for the cells
all are united together by numerous short processes. In the
peripheral layers of the organ sinuses are already developed,
their connection with the veinous circulation is effected and the
stimulating substances find easy access into the organ. _
The first effect of the stimulation is exhibited by an intense
proliferation of the mesenchymal cells and numerous mitoses.
The sinuses spread swiftly over the whole organ and form a net
with large meshes. The protoplasm of the mesenchymal
syneytium loses soon its uniform structure and becomes less
dense; numerous vacuoles appear. At the same time many of
the cells become isolated and lose their connection with the
plasmodial cell mass. These cells appear intensely basophylic
and are very ameboid in their character.’
An intense development of lymphoid hemocytoblasts is seen
on figure 6. The differentiation of lymphoid hemocytoblasts
can assume such proportions, that the greatest part of the mes-
enchymal anlage may be converted into free ameboid cells and
merely a few mesenchymal cells may remain between them
(fig. 20). No regular development of sinuses can take place in
such cases. The lymphoid hemocytoblasts multiply intensely,
many of them show a beginning differentiation into granu-
locytoblasts and acidophylic small granules appear in their
’ The changes which occur immediately after stimulation at early stages of
embryo development merely accelerate and intensify the normal histogenetic
processes in the spleen. Therefore illustrations given on plate 3, figures 5 and 6,
refer at the same time to normal development as well as to development after
stimulation.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
288 VERA DANCHAKOFF
cytoplasm. Numerous spleens of embryos, which were grafted
in early stages were transformed 4 to 5 days after grafting into
a tissue, similar to that represented on the figure 20.
The ameboid cells appear here pressed closely together, are
often of polygonal form and fill up all the interstices between
the few sinuses. This uniform wide-spread conversion of the
spleen anlage into a hemogranuloblastic tissue offers some
points of general interest. These changes show first, that the
action of the stimulus applied is intense and imperative. Prod-
ucts of metabolism appear in the blood current from the growing
graft cells. These substances incite an intense proliferation
of the mesenchymal cells in the spleen anlage and swiftly trans-
form the syncytial plasmodium of the anlage into an accumu-
lation of ameboid cells, which proliferate and differentiate fur-
ther. The results may remind one of the experiments of Rous,‘
in which connective tissue cells became spherical under the action
of trypsin.
As mentioned above the transformation of the mesenchymal
cells into lymphoid hemocytoblasts in the spleen anlage may be
universal, if stimulation is applied in early stages and merely a
few mesenchymal cells may remain undifferentiated, while in
normal development numerous mesenchymal cells persist and give
in course of time the reticulum tissue. Mollier (28, 711) describes
this tissue as a cellular syncytium with a fibrous net adjusted
to the cells. The changes in the experiments reported are so
intense (fig. 20), that it may be right to assume that numerous
cells, which under normal conditions ‘would have passed into
the reticulum tissue, now become hemocytoblasts and differenti-
ate further into granulocytoblasts. This result can be taken
as a corroborative proof for the polyvalency of the mesen-
chymal cells, which may either differentiate into.granular leu-
cocytes or become a constituent part of the reticular tissue.
The general granuloblastie differentiation of the mesenchymal
spleen anlage is observed in embryos to which the stimulation was
applied before a development of vascularization has taken place.
4 Personal communication.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 289
Development of intense erythropoiesis after stimulation.—The
stimulation of the lymphoid hemocytoblasts to intense prolifera-
tion may influence the development of the vessels and of their
contents in the mesenchymal spleen anlage. As above stated,
the stimulation in early stages leads often to defective develop-
ment of veinous sinuses and large parts of the spleen anlage are
transformed into granuloblastic tissue. If stimulation is ap-
plied a little later, at the ninth day, when a number of sinuses
are already formed, others still appear, a development of intense
erythropoiesis may be observed in the spleen anlage. The nor-
mal spleen tissue at this stage consists of a mesenchymal tissue,
in which numerous hemocytoblasts continue to develop. The
stimulation of the lymphoid hemocytoblasts and the simul-
taneous opening of splits in the mesenchymal tissue leads to the
appearance of numerous hemocytoblasts within the vessels.
Larger groups of lymphoid hemocytoblasts may occupy the
lumina of the sinuses and here undergo an erythroblastic dif-
ferentiation. Figure 15 shows a similar split developed locally
in the spleen mesenchyme. It is of irregular form and sur-
rounded by mesenchymal cells. These cells enter as a constit-
uent part into the general spleen tissue and send some of their
processes within the lumen of the sinus. A development of
numerous lymphoid hemocytoblasts is observed in this tissue
(fig. 15 H.L.HObl.) and large groups of lymphoid hemocytoblasts
are discharged in the lumen of the sinuses (fig. 15, 7.L.HOl.).
Outside the vessels the lymphoid hemocytoblasts differentiate
into granulocytoblasts (fig. 15, Grbl.), as elsewhere, within the
vessels they develop into erythroblasts (fig. 20, Erbl.) and
erythrocytes (Hrc). ‘The process of the normal differentiation
of a lymphoid hemocytoblast into an erythrocyte was studied
in birds by Danchakoff (9) and corresponds to what is seen
under the condition of an experimental stimulation. Therefore
I refer in this respect to my previous papers on the develop-
ment of different hematopoietic organs in birds and reptiles.
Figure 16 shows the spleen tissue in later stages (four days
after stimulation of an eight day embryo). The veinous sinuses
appear surrounded by flattened mesenchymal cells. The inner
290 VERA DANCHAKOFF
surface of the sinuses appears even and regular and is covered
by cells which resemble endothelial -cells. The vessels still con-
tain a large number of young undifferentiated blood cells which
continue to proliferate and to differentiate. The tissue between
the vessels begins to assume the character of a reticular tissue, in
the meshes of which numerous free ameboid cells are lying.
They chiefly consist of lymphoid hemocytoblasts, granulocyto-
blasts, and granulocytes (fig. 16, #.L.Hbl., Grol, Gre) together
with cells at intermediate stages of differentiation. However a
few macrophages may occasionally be seen.
The intense development of erythropoiesis and the activation
of granulopoiesis convert the spleen at this stage into a true
myeloid organ. As above stated, an occasional differentiation of
lymphoid hemocytoblasts into erythroblasts may be observed
during normal development of the spleen. Under stimulation
large groups of lymphoid hemocytoblasts fall into the developing
sinuses. At this time the vascularization in the spleen appear as
a net of large veinous capillaries annexed to the portal system and
the blood current must be here very slow. Situated in the sinus
lumen the lymphoid hemocytoblast is not conveyed away by
the blood current, but it remains in the sinus, proliferates and
differentiates into a red blood eell. The conditions of the de-
velopment of the erythropoiesis in the spleen correspond to those
known in the yolk sac annexes and in the bone-marrow. A de-
velopment of large veinous capillaries with a slow blood current
and a close connection between erythro- and granulopoiesis
is seen in all these regions.
However, the erythropoiesis incited by the experimental stimu-
lation does not persist in the spleen permanently. When the
arterious vascularization develops and the connection with the
sinuses is effected, the blood current becomes swifter, the-younger
cells are withdrawn and gradually replaced by differentiated
erythrocytes. In other cases in which the arterial vasculari-
zation develops defectively and no regular food supply is estab-
lished a lack of nutritive material finally may determine the
suspense of erythropoiesis. Large macrophages develop in the
sinuses and an intense phagocytosis of erythrocytes may be
observed (figs. 17, 18, 19).
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 291
The intense development of erythropoiesis at the expense of
lymphoid hemocytoblasts, which develop locally is another
striking evidence for the truth of the monogenetic conception
of the blood development and consequently for the polyvalency
of the mesenchymal cells in the spleen anlage. The same
mesenchymal cells, which normally would remain between the
vessels and either become a part of the reticular tissue or differ-
entiate into granular leucocytes now differentiate into red blood
cells after they are discharged into the vessels.
Deficiencies in the development of vascularization and thewr effects.
—The deficiencies in the vascularization of the spleen anlage
occur as results of the changes in the spleen incited by stimu-
lation of the stem cells. They can affect either the development
of the veinous net or the ingrowth and the distribution of arteries.
The stimulation applied at the 7 to 8 day of incubation may
convert the whole spleen anlage or considerab’e parts of it into
granuloblastic tissue. The sinuses develop thereby defectively
and considerable accumulations of granuloblastic tissue are often
provided merely with scant and narrow capillaries. The scarce
development of veinous sinuses is the natural sequence of the
excessive differentiation of granuloblastic tissue. The study of
the normal spleen development has shown, that the sinuses de-
velop locally as splits amidst the mesenchymal tissue. Since the
greatest part of mesenchyme may be transformed into accumu-
lations of free ameboid cells (fig. 20), the net of sinuses cannot
locally develop between the free cells.
Defects in the development of veinous sinuses become them-
selves the source of interesting alterations in the spleen tissue.
The normal histogenesis leads to a gradual differentiation of
granular leucocytes at the expense of granulocytoblasts. The
granulocytoblasts, though ameboid, do not migrate normally in
the vessels which evidently do not contain adequate chemiotactic
substances. The granular leucocytes, however, penetrate easily
into the veinous sinuses through thin walls (fig, 11, Z) and
are drifted by the blood current away. Though the differentiation
of granular leucocytes is in a normal embryonic spleen continuous,
292 VERA DANCHAKOFF
they never accumulate in the tissue. The conditions develop
quite differently after stimulation, and lead to an excessive
granulocytspoiesis (figs. 20 and 21). In this case two factors
work together in one direction—first: the granular leucocytes
are formed in excessive numbers and secondly vessels develop
defectively. Considerable accumulations of granular leucocytes
remain therefore in the spleen tissue. They may densely
infiltrate the mesenchyme, if there is mesenchyme tissue left
(fig 21).
On the other hand, they may form enormous agglomerations
in the form of spherical masses of semi liquid tissue and finally
perish (fig. 22, Gre.’’’). Centers of necrotic tissue appear in the
spleen as a result of the excessive production and stagnation of
granular leucocytes. The development of large accumulations
of granular leucocytes are interesting in so far as it seems to
indicate that in the particular case cells cannot stop in their
development. The reactions displayed by these living cells nec-
essarily lead the cell to the last stage of its differentiation.
The universal and imperative conversion of the spleen mesen-
chyme into granuloblastic tissue also offers an example of tissue
reaction, which being a response to the stimulation, seems to
lack the character of purposefulness. The stimulation breaks
up the reciprocal normal proportions of the development proc-
esses in the spleen. The exclusive development of granu-
lopoiesis in the spleen-anlage leads to formation of considerable
centers of completely differentiated cells which finally suecumb
in large masses. The embryos, in which such wide spread
changes are observed, do not hatch, and usually die after 16 to
19 days of incubation.
Around the necrotic centers, if they are not too considerable
in size and numbers, a characteristic reaction of mesenchymal
tissue may be observed. ‘The mesenchymal cells proliferate and
form plasmodial masses in the form of giant cells around the
necrotic center (fig. 22, Gt.c). The specific appearance of the
nuclei accumulations may suggest here also the idea of occurrence
of amitosis (fig. 22, Y). These giant cells are perfectly similar
to those encountered around the foreign bodies. Though their
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. . I. SPLEEN 293
appearance seems to be purposeful, I do not feel right to take
them for such. It seems more correct to attribute their appear-
ance to a reaction similar to that which characterizes the develop-
ment of macrophages—reaction to particulate matter, Evans (15).
The considerable hypertrophy of their cytoplasm is brought by
digestion of phagocytozed material. Smaller necrotic centers can
undergo a complete resorption by giant cells. The formation of
giant cells around necrotic centers can serve as an example of a
new differentiation potency assigned to mesenchymal embryonic
cells.
Besides deficiencies in the development of veinous vasculariza-
tion, irregularities in the development of arterious vasculariza-
tion may also be observed. A partial lack of development of
arterious ramifications in districts of the spleen is a frequent
occurrence under the conditions of the experiments. <A partial
deficiency in the veinous vascularization usually coincides and
this leads to a nearly complete lack of vascularization in parts
of the spleen. In such regions mesenchymal cells are still capable
of proliferation, but they soon develop very differently than they
do normally, namely, they become fibroblasts. Large regions of
the spleen can be transformed into typical connective tissue (fig.
23). The elongated cell bodies may appear pressed closely
together, in other instances they are separated by accumulations
of collagenous fibers. A few mitoses are usually observed in
these cells. Sometimes the presence of a few lymphoid hemocyto-
blasts or granular leucocytes may indicate a more intense
hematopoietic differentiation which previously has taken place
in such regions.
The development of fibrous tissue in the embryonic spleen
completes a whole range of transformations to which the young
mesenchymal cells are capable. The leucocyte, the erythrocyte
and perhaps the fibroblast and the macrophage are under normal
conditions final non interchangeable differentiation stages; the
endothelial, the giant and reticulum cells are different morpho-
logical structures, which a mesenchymal cell may assume. All
these cells derivefrom one stemceell, and its differentiation depends
upon the various conditions to which the polyvalent cell is
submitted.
294 VERA DANCHAKOFF
b. Changes obtained by stimulation in the stage with definite spleen
vascularization
Studying the myeloid metaplasis, Hertz (20) found that besides
an intense development of granuloblastic tissue in the pulpa,
well pronounced changes in the follicles appeared also. The
tissue of the whole follicle might have undergone a differentiation
into large lymphocytes (lymphoid hemocytoblasts). The changes
observed in the spleen after stimulation in later stages (at 14
to 15 days of incubation) in which the structural peculiarities
of the organ are developed, correspond closely to those described
by Hertz. The large lymphocytes (lymphoid hemocytoblasts)
develop, according to Hertz, in the follicles partly at the expense
of the reticulum cells, partly at the expense of small lymphocytes.
The development of lymphoid hemocytoblasts at the expense of
the mesenchymal reticulum is most evident in the embryonic
spleen after stimulation. However, the present experiments offer
no evidence that they develop at the expense of small lymph-
ocytes, because they appear at a time when the small lymph-
ocytes are either still very scant or have not yet developed at all.
Figure 24 represents the tissue of a follicle 3 days after stimu-
lation, applied to a 12 days embryo. The largest part of the
cells consists of lymphoid hemocytoblasts, which multiply in-
tensely in a hetero- and homoplastic way. The artery walls
themselves are usually infiltrated by large basophylic cells,
though normally they are formed at this time by a loose tissue,
in which free cells are absent. The splitting off of lymphoid
hemocytoblasts by the mesenchymal reticulum in such follicles
is easily discernible. Though this splitting leads to the
development of numerous hemocytoblasts, it however never
attains the intensity observed in the pulpa-like spleen (fig. 20)
after stimulation at early stages. Between the ameboid cells
a distinct net of mesenchymal cells is apparent and they pro-
liferate and continue to split off hemocytoblasts. Endothelial
cells of the arteries and their smaller branches seem to be
exempt from the stimulating action. Whether the lack of a re-
action in the endothelial cells is caused by the final specialization
of these cells, or whether the endothelial cells are merely
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 295
slow in their response, cannot be definitely determined on the
basis of the available experimental material. A similar intense
development of lymphoid hemocytoblasts at the expense of
reticulum cells is observed in the pulpa. Thus the stimulation
applied at a stage with developed spleen vascularization incites
equal reactions through the whole organ. Both in the pulpa and
in the follicles this reaction is manifested by an intense splitting
off of lymphoid hemocytoblasts and their intense proliferation.
A study of further development of such spleens shows, how-
ever, that the differentiation of the lymphoid hemocytoblasts
develops differently according to their localization in the pulpa
or in the follicle. The study of the normal histogenesis has
demonstrated that lymphoid hemocytoblasts, situated in the
cavernous tissue of the pulpa, differentiate finally into granular
leucocytes; in the reticular tissue of the follicle they develop
into small lymphocytes. The same strong dependence of
differentiation of lymphoid hemocytoblasts wpon the environ-
mental structural conditions is observed in a spleen after
stimulation. The stimulation stirs up the less differentiated
cells, which are the mesenchymal cells and the lymphoid hemo-
eytoblasts. The further differentiation is effected in compliance
with the conditions met and granular leucocytes are differenti-
ated in the pulpa, and small lymphocytes in the follicles.°
>It is difficult to harmonize with the results of my present paper, the data
given by Dr. Murphy in his recent note regarding the effect of adult spleen
erafts on the organism of the hostembryo (Journ. of Exp. Med. July, 1916) in
which he states that, ‘‘while the spleen of a normal embryo of this age (18 days)
presents only a beginning differentiation of cells, after grafting this process is
well advanced and numerous cells of both the granular and non-granular type
are found.”
As seen in figures 7, 8, 9, 10, 12 and 13 of the present paper taken at the
llth, 13th and 15th days of incubation a normal spleen presents a highly ad-
vanced differentiation of granular cells from the 11th day and a good start of
development of small lymphocytes from nearly the 15th day of incubation.
The changes in the host spleen after grafts at the 17th and 18th day of incu-
bation, 10-11 days after grafting (the only stages to which Dr. Murphy refers
in the above quoted note) present the ultimate results of intense modifications,
which are very different from the advanced stage of differentiation mentioned
by Dr. Murphy.
A remarkable stimulation of granulopoiesis can be seen however in early
stages,—at the 2nd, 3rd, and 4th day after grafting. A stimulation of develop-
ment of small lymphocytes has not been observed as yet.
296 VERA DANCHAKOFF
There were repeatedly described cases of myeloid metaplasis,
in which the chief character was represented by a development of
granuloblastice tissue and a consequent suppressing of lymphoid
tissue. For the true understanding of similar conditions one
may remember that granulocytoblasts only might have under-
gone a stimulation. Granulocytoblasts are specifically differ-
entiated cells, the proliferation faculty of which may be aroused
by specific agents, which could leave inert the other offspring
of the common stem cells and the stem cells themselves.
The general proliferative reaction of the younger undifferen-
tiated cells to the stimulation observedin the present experiments,
no matter where these stem cells are localized—whether in the
pulpa or in the follicle—point again to a similarity of these cells
in their simultaneous reaction to the same factor. It seems,
therefore, that to morphological and histogenetic data, new
data of uniform biological reaction may be added as evidence for
the uniformity of the hemocytoblastic cell group. This bio-
logical reaction consists in an intense heteroplastic and homo-
plastic multiplication as response to the stimulus. Isomorphism,
isogenesis and isodynamism under equal conditions, evidently
associate in the existence of the lymphoid hemocytoblasts.
C. CONCLUSIONS
On the basis of observations and experiments described in the
present paper some data may be won concerning (1) the histo-
genesis of the blood cells in the spleen, (2) the cell differentiation,
and (3) the general meaning of the myeloid metaplasis.
The data concerning the histogenesis of the spleen cells were
recapitulated at the end of every section, therefore, I may here
merely outline the general conclusions.
The chief results of the study of normal spleen development
is the statement of a regular and unalterable relation between
differentiation of certain kinds of blood cells and structural
environmental conditions.
These results have been confrmed by the experimental part
of the work. Moreover it has been shown experimentally that
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 297
the development of mesenchyme and lymphoid hemocytoblasts
‘an be intensely stimulated and that the respective boundaries,
in which the differentiation of certain cell groups usually takes
place, can be shifted (displaced), in other words, that the pro-
spective potency (after Driesch’s terminology) of the blood stem-
cells is greater than their prospective value.
The fact that under certain conditions nearly the whole amount
of the mesenchymal cells of the spleen anlage may undergo a
granuloblastie differentiation; that under other conditions they
show a fibroblastic, or an erythroblastic differentiation, and so
forth, is the natural sequence of the polyvalenecy of the mesen-
chymal cells in the spleen anlage. The early process of segre-
gation must have led in the spleen anlage to a production of one
group of numerous homogeneous stemeells, which under different
conditions split off variously differentiated cells.
Data concerning the conception of cell differentiation
The histogenetical studies of the spleen under normal con-
ditions and after stimulation have shown that a mesenchymal
cell, which normally would contribute to the formation of retic-
ulum tissue, can develop into a hemocytoblast or into a fibro-
blast or a giant cell, that the lymphoid hemocytoblasts can
develop into a granular leucocyte, or into an erythrocyte or a
small lymphocyte. The study of hematopoiesis in birds and
reptiles shows very definite conditions for each of the lines of
differentiation. Though Haff (18) has lately described the
presence of extravascular erythropoiesis in the liver of hens,
yet I was not able to confirm his data. My personal obser-
vations concerning the existence of small centers of erythropoiesis
in the connective tissue of the hen do not seem to contradict
the general conditions. These centers, scattered irregularly
and finally phagocytosed may be interpreted as locally developed
blood and vessel anlages, of which the connection with the general
circulation has not been fully effectuated.
It is known, however, that both erythropoiesis and granulopoi-
esis develop in mammals extravascularly and, according to
Maximow (25) (’11)—‘‘zwei neben einander liegende ganz gleiche
298 VERA DANCHAKOFF
Vera Danchakoff
Text fig. 1. Scheme of cell differentiation and multiplication by equal division
wives as result a range of differentiated cells.
4unthad lamceuleblat
q
Vera Danchakoft
Text fig. 2. Scheme of cell differentiation and multiplication by unequal division
ives as result a range of differentiated cells, a number of stem cells and intermediate
wm UG
tages of differentiation.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 299
Lymphocyten, die sich ja sicherlich auch in gleichen 4ussern
Existenzbedingungen befinden, doch zu verschiedenen Endpro-
dukten entwickeln.’?’ Maximow rightly points out, that the
problem of the differentiation of blood cells forms merely a par-
tial question of the general problem of differentiation of an egg -
in a multicellular organism. The observation that cells begin
their granuloblastic differentiation soon after completion of
mitosis, led Maximow to connect differentiation with mitosis.
The differentiation factor is conceived by Maximow as ‘“‘eine
tiefe Gleichgewichtsstérung’’ which occurs during cell division
is not reversible, and concerns both daughter cells. It leads to
specific qualitative changes,—to the development in the cytoplasm
of hemoglobine or granules—‘‘ Warum aus einem sich teilenden
Lymphocyt in dem einen Fall ein Paar junger Myelozyten, in
einem anderen ein Paar junger Erythroblasten u.s. w. hervorgeht,
hingt wahrscheinlich vom Zufall ab.” It is, however, not deter-
mined how the chance can lead to definite and well pronounced
differences in the development of two identical cells which seem
to be submitted to equal surrounding conditions. Maximow’s
hypothesis attributes the differentiation to ‘Gleichgewichtsstor-
ung,’ which depends upon chance. It seems to me that the factor
of differentiation which so evidently appears in birds and reptiles,
and consists here in definite structural conditions, cannot be es-
sentially different in mammals. However, these conditions are
more difficult to trace In mammals.
The proliferation, of the mesenchyme and lymphoid hemo-
cytoblasts—and perhaps the development of lymphoid hemo-
cytoblasts at the expense of mesenchymal cells—seems to depend
upon stimulation by enzyme-like substances, which appear in
the organism of the embryo from the growing graft cells.
The circulating substances call forth an enormous production
of lymphoid hemocytoblasts through the whole spleen tissue.
In early stages, when the spleen tissue is homogenous and the
conditions correspond to those which are required for granu-
lopoiesis, only granular leucocytes are developed, and the dif-
ferentiation of small lymphocytes does not appear until struc-
tural conditions develop in the spleen under which normally
300 VERA DANCHAKOFF
small lymphocytes are differentiated. According to new envi-
ronmental conditions the mesenchymal cells may develop into
typical connective tissue. Again giant cells appear as a reaction
of the mesenchymal cells around the necrotic centers. Thus the
study of the spleen development in normal conditions and after
stimulation adds strong evidence for the conception that at least
one of the factors for the differentiation of the polyvalent stem
cells consists in the physicochemical conditions to which the
cell is submitted.
If the environmental conditions which determine the vari-
ous differentiation of the stem cells are easily traced, it is much
more difficult to understand how differentiation of the stem cells
takes place simultaneously with their uninterrupted multipl-
cation. The process of differentiation affects both the cyto-
plasm and the nucleus. Specific substances are developed in
the cytoplasm. The nucleus during differentiation process
loses gradually its typical nucleolus and accumulates chromatin,
which permanently remains in the form of intensely basophylic
particles. Maximow’s recourse to the ‘Gleichgewichtsstérung”’
during mitosis cannot explain the persistence of the stock of the
young stem cells. May the persistence of the young stem cells
be explained perhaps by a higher rate of cell multiplication in
comparison with their differentiation? I do not think so. If
the differentiation is a result of the influence of certain condi-
tions upon the cell, whatever rate of cell proliferation may be
admitted, certain kinds of cells under definite conditions will
all differentiate simultaneously. The simultaneous and _ per-
manent differentiation and multiplication of the lymphoid
hemocytoblasts cannot be explained by a high rate of cell pro-
liferation.
A group of hemocytoblasts (let us say a group of similar A
cells) develops in the loose connective tissue. Some of these cells
differentiate into B cells, or granulocytoblasts; others continue
to multiply as such. If the environmental conditions, as they
seem to appear, are similar for all these cells, how is the differ-
ence in their behaviour to be explained? Is a difference in the
constitution of different cells involved and does the group of
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 301
differentiating cells, A, consist of a number of Aa, and of Ab
cells? The Aa cells may be supposed to be excluded from the
differentiating influence of the environmental conditions and
multiply as such. The Ab cells may possess the faculty of de-
veloping acidophylic granules and become granulocytoblasts.
What will then happen? The Aa cell will divide and give two
Aa’s, the Ab cell will differentiate into a B cell. At what time,
then, does the differentiation of an Aa cell into an Ab cell take
place? If between two mitosis, then it is to be expected that under
definite conditions at a given time all the Aa cells will differentiate
into Ab cells and further into B and C cells (text fig. 1.) Yet the
stock of A cells does not show any signs of exhaustion and the
hemocytoblasts preserve their existence as such.
The simultaneous differentiation and proliferation of young
cells could hardly be explained otherwise than by a specific
process of differentiation during mitosis (text fig. 2.) The
division of a cell A must lead to the development of a cell A
and a cell A'. The cell A will continue to give rise to cells A
and A! and will truly become the inexhaustible source of the
young undifferentiated cells. The cell A! will undergo further
differentiation and will develop finally into a B cell, or a granu-
locytoblast, which again wili divide into B and B'.. The cell B'
will differentiate into a C cell, or a granular leucocyte. The
differentiation of the hemocytoblasts into granulocytes outside
the vessels as well as their development into erythroblasts and
erythrocytes within the vessels and their simultaneous inexhausti-
ble proliferation could hardly be explained on other grounds.
I have not been able yet to trace any difference between two
daughter cells. However, numerous examples of similar un-
equal cell division may be found in the life history of other cells.
Boveri represents the differentiation of germ and somatic cells
in Ascaris as due to cleavages, which result in formation of cells,
of which a part conserves the whole chromatine and another
loses a considerable amount of it. Again the differentiation of
various cell ranges with persistence of the stem cells in Clepsines
depends upon unequal division of the stem cells.6 Conklin
® This example was kindly given to me by Prof. H. H. Donaldson.
302 VERA DANCHAKOFF
sees in the cytoplasm division itself a differentiation factor,
which accomplishes the segregation of different substances in the
cytoplasm. Lately interesting observations were made in the
laboratory of Prof. C. McClung by Dr. Wenrich and Miss Car-
others, on formation of heteromorphic chromosomes in the
spermatogenesis of grasshoppers. The admission of similar
unequal chromosome-division in somatic cells may explain the
gradual development of different mitosis figures characteristic
of different tissues. The small sizes of the somatic cells do not
allow to gain direct data on the inequality ofthe daughter cells.
However, the examples cited above sufficiently explain the
possibility of simultaneous persistence of young stem cells and
their further differentiation.
In connection with the problem of cell differentiation, the
existence of a large number of amitotic cell divisions, may be
mentioned. Lately Patterson has described in the Keimblittern
of the pigeon, the presence of numerous amitoses. In regions of
intense proliferation the cells seemed to undergo a full amitotical
division, and give rise to apparently normal daughter cells,
which could themselves multiply mitotically. Maximow (25)
(08) found similar conditions in certain regions of embryos in
mammals. In early stages of spleen development in chicks,
when the nuclei proliferate intensely in the mesenchymal spleen
anlage, numerous pictures of amitotical nucleas division may be
encountered. Figure 14 illustrates four cells in a stage of amito-
tic nuclei division. Asa result of the division of the nucleus two
daughter nuclei arise, which can have both the same dimensions,
or differ considerably in size. Both daughter nuclei receive
always a part of the nucleolar substance. In most cases the nuclei
divisions are not followed by cell divisions, for the very fact that
their localization in the syncytium mass of the mesenchyme does
not allow a division of the cytoplasm. However, sometimes free
‘cells may be observed undergoing amitotic nuclei division (fig.
13, 14d). The further destiny of such nuclei cannot be followed
directly. The frequence of occurrence of amitotic nuclei di-
vision must, however, lead in the spleen mesenchyme to a
production of a larger number of nuclei, derived through
amitosis. These nuclei contribute in a great part to the growth
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 303
of the apparently homogeneous mesenchymal anlage, at the
expense of which many different cells are developed. I cannot
enter now into a detailed consideration of this problem. How-
ever, the question whether the amitosis may not be looked upon
as one of the factors in the differentiation of the living substance,
ean find its justification.
Data concerning the general meaning of the myeloid metaplasis
The results obtained in the embryonic spleen at different
stages of its development have a close bearing upon the myeloid
metaplasis, well known to the pathologists. What is the general
meaning of these changes, and what are their relations to the
stimulus applied and to other reactions displayed at the same time
by the organism? The specific conditions under which I had
to work last winter did not allow me so far to obtain definite
data in this respect. However, the study of the experiments
made may give a few suggestions concerning the problem.
There is no doubt the intervention applied introduces in the
organism heterogeneous substances. The connection between
the introduction of these substances and the changes described
above may be conceived in two different ways. Either these
substances have general stimulating action on the mesenchyme
and on the blood stem cells (as for example the thyroid sub-
stance in the experiments of Gudernatsch’ on the develop-
ment of tad-pole limbs); or the action of these substances may be
similar to that of specific antigenes, which introduced in the
organism incite the production of antibodies. In the latter case
the proliferative reaction exhibited by the blood stem cells may be
looked upon as a material basis for the phenomena of immunity.
The proliferation in this case would have been brought about
by the interaction between specific antigenes and their cell
receptors. Only results of new series of experiments will decide
which of these two conceptions has to be accepted.
The analogy between the myeloid metaplasis and the results
of the experiments described is obvious. It is important to
7Gudernatsch, F. Feeding Experiments on Tadpoles. II, A further con-
tribution to the knowledge of organs with internal secretion, Am. Jour. Anat.,
1914, vol. 15.
304 VERA DANCHAKOFF
notice that the myeloid metaplasis is produced by different
causes. The toxines of various bacteria, the specific products
of metabolism of malignant tumors, finally, inorganic chronic
intoxications may incite an extensive myeloid metaplasis. It
is difficult to conceive in such qualitatively different agents a
specific stimulating influence on the stem cells. The response
to the action of these factors is specific in so far as it is exhib-
ited by a certain kind of tissue (even not of cells). The stimu-
lus itself may largely vary. The cell, being understood as a
complex group of receptors, may offer adequate receptors to
different stimuli or antigenes. <A similar example of stimulation
to proliferation by different agents may be found in the phe-
nomena of fertilization. The specific or usual stimulus in the
form of the spermatozoon may be replaced by other chemical
stimuli, which may find adequate receptors in the egg cell and
incite therefore molecular changes, which are followed by pro-
liferation.
If more than occasional coincidence is to be seen in the regular
connection of the appearance of different antibodies and the
myeloid metaplasis after infections, the specific antibodies may
consist of substances derived from the proliferation and differ-
entiation activities of the hematopoietic tissue. If so, a mere
stimulation of the hematopoietic tissue would suffice for strength-
ening or developing immunity; the production of a large amount
of antibodies would follow as a result of stimulation plus speci-
fic action of the antigene. Otherwise how could be explained the
development of immunity against heteroplastic grafting by intro-
duction in the organism of an emulsion of various tissue (Da
Fano)? This intervention, similarly to the experiments de-
scribed, must have also stimulated the hematopoietic tissue.
It is too early now to attempt to draw more definite conclusions
concerning the specific functions of the hematopoietic tissue.
Confronted with the simultaneous development of widespread
changes in the mesenchyme, which occur after the appearance in
the organism of. antigen-like substances and usually followed
by a production of specific antibodies, one may find it natural
to think of the mesenchyme and its differentiation products as of
an organ in close relation to the production of immune bodies.
EQUIVALENCE OF HEMATOPOIBRTIC ANLAGES. I. SPLEEN 305
LITERATURE CITED
(1) BrezaAncon Ev LABBe 1904 Traité d’Hématologie.
(2) Bryce, F. 1904 The histology of the blood of the larva of Lepidosiren
Paradoxa.
1904 Part I Structure of the resting and dividing corpuscles.
1905 Part II. Haematogenesis. Transactions of the Royal Society
of Edinburgh.
(3) BizzozErRo 1890 Neue Untersuchungen iiber den Bau des Knochenmarkes
bei Vé6geln. Arch. fiir Mikroskopische Anatomie, Bd. 35.
(4) BuumentTHaAL, R. 1904 Recherches exp. sur la genése des cellules sanguines,
ete. Travaux du Laborat. de Physiologie de l’ Inst. Solvay, T. 6.
1908 Le sang et ses territoires d’ origines. Bull. de la Societé Roy.
des Se. med. et nat. de Bruxelles, N. I.
1909 La morphogenése des cellules h6molymphatiques. Folia Hemat,
Bd. 7.
1910 Comment faut-il envisager la parenté des cellules sanguines
chez l’adulte, Folia Hemat., Bd. 9.
(5) Burrerrirevp, BE. E. 1908 Uber die ungranulierten Vorstufen der Mye-
lozyten und ihre Bildung in Milz, Leber und Lymphdriisen. Deut.
Arch. f. Klin. Med., Bd. 92.
(6) Conxiin, E.G. 1898 Cleavage and differentiation. Woods-Hole Lectures
1905 Organization and cell lineage of the ascidian egg. Journ.
Acad. Nat. Sci. Phila., vol. 138.
1912 Exper. studies on nuclear and cell division. Journ. Acad. Nat.
Sci. Phila., vol. 15.
(7) Da Fano C. 1910 Zellulire Analyse der Geschwilstimmunititstreaction.
Zeitschrift fiir _Immunititsforschung, Bd. 5.
(8) Denys Sur la structure de la moelle osseuse des os et la genése du sang ©
chez les oiseaux. La Cellule, T. 4.
(9) Dancuaxorr, V. 1907 Uber das Auftreten der Blutelemente im Hiihner-
embryo. Folia Haematologica, IV Jahrg., Suppl., Heft 2.
1908a Die erste Entstehung der Blutzellen beim Hithnerembryo und
der Dottersack als blutbildendes Organ. Anat. Hefte, Bd. 37.
1908b Das lockere Bindegewebe des Hiihnchens im fétalen Leben.
Arch fiir mikr. Anat., Bd. 73. ;
1908e Uber die Blutbildung im Dottersack des Hiihnchens. Ver-
handl. d Anat. Ges. Berlin.
1908d Zur Herstellung der Celloidinserien. Zeitschr. fiir wiss. Mikr.
1909 Uber die Entwickl. des Knochenmarks bei den Végeln und iiber
dessen Verinderungen bei Blutentziehungen und Ernihrungsstéh-
rungen. Arch. fiir mikr. Anatomie, Bd. 74.
1910 Uber die Entwicklung der embr. Blutbildung bei Reptilien.
Verh. Anat. Ges., Briissel.
1916a Uber die Entwicklung des Blutes in den Blutbildungsorganen
bei Tropidonotus Natrix. Arch. fiir mikr. Anat., Bd. 87.
1916b Origin of the blood cells, ete. Anatom. Record, vol. 10, no. 5.
306 VERA DANCHAKOFF
(9) DancHakorr, V. 1916c¢ Cell differentiation as criterion for cell identifica-
tion. Journ. of Exp. Med., vol. 24.
1916d The wandering cells in the loose connective tissue of the bird
and their origin. Anat. Rec. vol. 10, no. 7.
(10) Domrnicr, H. 1900 Sur l’histologie de la rate normale. Arch. de Méd.
Expset dAn: <Pate i412:
1900 Sur Vhistologie de la rate au cours des états infectieux. Arch.
de Méd. Exp. et d’Anat. Pat. T. 12.
1901 Sur le plan de structure du systéme hematop. des mammiféres.
Arch. de Méd. Exp. et d’Anat. Pat. T. 13.
1901 Sur Vhistologie de la rate 4 l’état normal et pathologique.
Arch. de Méd. Exp. et d’Anat. Pat. T. 13.
1909 De l’origine lymphatique ou myéloide des polynucléaires. Folia
Haemat. Bd. 8.
(11) Downey, H. 1915 The origin and development of eosin, leucocytes, etc.
Folia Haemat. Bd. 15.
(12) Downey, H. anp WeIpEnreEIcuH, F; 1912 Uber die Bildung der Lymphocy-
ten in Lymphdriisen und Milz. Arch. fiir Mikr. Anat. Bd. 80.
(13) Enrutcu-Lazarus 1909 Die Animie. 2 Auflage.
(14) Frerpsonn, A. 1910 Zur Morphologie des Amphibienblutes ete. Arch.
fiir Mikr. Anat. Bd. 75.
(15) Evans, H. M. 1915 The Macrophages of mammals. Amer. Jour. of
Physiol., vol. 37.
(16) Fiscomr, H. 1909 Myeoloische Metaplasie und fétale Blutbildung und
deren Histogenese. Berlin.
(17) Fiuexner, 8. 1902 The pathologie of lymphotoxie and myelotoxic intoxi-
cation. Univ. of Penna. Med. Bulletin.
(18) Harr, R. 1914 Bindegewebe und Blutbildungsprocesse in der embr.
Leber des Huhnes. Arch. fiir Mikr. Anat. Bd. 84.
(19) Hextiy, K. 1903 Die Blutbahnen der Milz und deren functionelle Bedeut-
tung. Arch. fir Mikr. Anat. Bd. 61.
1906 Die haematop. Organe in ihren Beziehungen zur Pathologie
des Blutes. Nothnagels Handbuch der Pathologie und Therapie.
Bd. 8; 1 Abt. sil Merl:
1910 Animische Degenerationen und Erythrogonien. Ziegler’s Beitr.
Bd. 49. |
(20) Hertz, R. 1910 Zur Frage der experim. myeloischen Milz-Metaplasie.
Zeitschr fiir Klinische Med. Bd. 71.
(21) Hirscuretp, H. 1902 Uber myeloide Umwandlung der Milz und der
Lymphdriisen. Berl. Klin. Wochenschr. Bd. 39.
1906 Weiteres zur Kenntniss der myeloiden Umwandlung. Berl.
Klin. Wochenschr. Bd.
(22) Kuern, St. 1914 Die Myelogonie. Berlin, Julius Springer.
(23) Lacurnssn, E. 1890 Recherches sur le développement de la rate chez les
poissons. Journ. de l’Anat. et de la Physiologie, 26 Année.
(24) LoBENHOFFER, W. 1908 Uber extravaskulire Erythropoese in der Leber
unter path. u. normalen Verhiltnissen. Ziegler’s Beitrige, Bd. 43.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES. I. SPLEEN 307
(25) Maximow, A. 1907 Uber die Entwicklung der Blut- und Bindergewebs-
zellen beim Saugetierembryo. Folia Haemat. Bd. 4, N. 5.
1908 Amitosis.-Anatom. Anz. Bd. 33.
1909 Untersuchungen tiber Blut und Bindegewebe. I- Die friiheste
Entwicklungsstadien ete. Arch. fiir Mikr. Anat. Bd. 73.
1911 Untersuchungen tiber Blut und Bindegewebe. III- Die embr.
Histogenese des Knochenmarkes der Saugetiere. Arch. f. Mikr.
An. Bd. 76.
(26) Meyer unp Hernecker, A. 1907 Uber Blutbildung bei schweren Aniimien
und Leukiimien. Deut. Arch. fur Klin. Med. Bd. 88.
(27) Mreyrr, E. 1908 Weitere Untersuchungen itiber extra-uterine Blut-
bildung. Miinch. Med. Wochenschr. N. 22.
28) Monier, 8. 1909 Die Blutbildung in der embr. Leber des Menschen und
der Siugetiere. Arch. fiir Mikr. Anat. Bd. 74.
1911 Uber den Bau der Kapillaren der Milzvenen. Arch. f. Mik.
An. Bd. 76.
(29) Murpuy, Jas. B. 1913 Transplantability of tissues to the embryo of
foreign species Jour. of Exp. Med. vol. 17.
(30) Murpuy, Jas. B., AND Rous, P. 1912 The behaviour of chicken sarcoma
implanted in the developing Embryo.-Journ. of Exp. Med. vol. 15.
(31) NxXeeur, O. 1906 Beitraige zur Embriologie der blutbildendenden Organe.
Verh. Kong Inn. Med.
1912. Blutkrankheiten und Blutdiagnostik. 2 Auflage. Leipzig.
1913 Blut. Aschoff’s Pathologische Anatomie. Bd. 2.
(32) Pappennerm, A. 1907 Uber die Stellung der akuten grosszellig-lymphozyt.
Leukimie Folia Haemat. Bd. 74.
1908 Uber lymphoide basophile Vorstufen der Erythroblasten.
Folha Haemat. Bd. 5.
1909 Pathologie und Therapie der Leukimien. Prolegomena. Fol.
Ham. Bd. 7.
1909 Zur vorstehenden Mitteilung Dominici’s Fol. Haem. Bd. 8.
1911 Kurze technologische Zusammenstellung der Farbungsforschrift
mit Panchrom. Folia Haemat. Bd. 11.
1913 Prolegomena. Folia Haematol. Bd. 14.
(33) PAPENHEIM AND Ferrata 1910 Uber die verschiedenen lymphoiden
Zellformen des normalen und pathologischen Blutes. Folia Haemat.
Bd. 10.
(34) Paremusorr, J. 1911 Zur Kenntniss der Zellen der Milzpulpa. Fol.
Haemat. Bd. 12.
(35) Patrerson. 1908 Amitosis in the Pigeon Egg. .Anat. Anz. Bd. 382.
(36) Rupascukin, W. E. 1907 Eine Methode zur Herstellung von Celloidin-
serien. Anat. Anz. Bd. 81.
(37) Scuatttorr, P. 1908 Uber die histologischen Verinderungen der blut-
bildenden Organe bei pernicioser Animie. Miinch. Med. Wochenschr.
Bd.
(88) ScarippE 1907 Myeloblasten, Lymphoblasten und lymphblastische Plas-
mazellen Ziegler’s Beitrige. Bd. 41.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
308 VERA DANCHAKOFF
(38) Scurippe 1907 Die Entstehungdererstenembr. Blutzellen des Menschen.
Fol. Haemat., Suppl. Bd.
1908 Uber Regeneration des Blutes unter normalen und krankh.
Verhaltnissen. Centralbl. fiir allg. Path.. Bd. 19.
1908 Uber die Histogenese der myeloischen Leukiimie. Miinch. Med.
Wochenschr., N. 20.
1913 Die blutbildenden Organe. Aschoff’s Patholog. Anatomie. Bd. 2.
(40) SternBerG, C. 1905 Pathologie der Primiirerkrankungen der lymph.
und haematopoet Apparates. Wiesbaden.
1911 Uber die akute myeloische Leukiimie. Wien. I<lin. Wochenschr,
N. 47.
(41) Tonxorr, I. 1900 Die Entwicklung der Milz bei den Amnioten. Arch.
f. Mikr. Anat., B. 56.
(42) Turck, W. 1912 Vorlesungen uber klinische Hamatologie. Wien.
(43) Venzuarr 1911 Uber Genesis und Morphologie der roten Blutkérperchen
der Vogel. Arch. fur Mikr. Anatom., Bd. 77.
(44) Werpenrercu, F. 1901 Das Gefiisssystem der menschlichen Milz. Arch.
fur Mikr. Anat., Bd. 58.
1905 Bau und morphologische Stellung der Lymphdriisen. Arch.
fur Mikr. Anat., Bd. 65.
1909 Zur Morphologie und morpholog. Stellung der ungranuliert.
Leukozyten-Lymphozyten des Blutes und der Lymphdriisen. Arch. fiir
Mikr. Anat., Bd. 73.
1911 Die Leukozyten und verwandte Zellformen. Wiesbaden.
(45) Wenricu, D. H. 1916 The spermatogenesis of Phrynotettix magnus,
ete. Bull. of Com. Zoology, Vol.
(46) WerzBerGc, A. 1911 Neue experim. Beitrige zur Frage der myeloiden
Metaplasie. Virchow’s Archiv fiir Path. Anat., Bd. 204.
(47) Zieater, K. 1906 Experim. und klinische Untersuchungen itiber die
Histogenese der myeloiden Leukimie. Jena.
1908 Uber die Beziehungen zwischen myeloiden Umwandlung und
myeloiden Leukiimie u.s. w. Folia Haemat., Bd. 6.
1910 Uber Morphologie bei Blutbereitung bei pernicioser Anamie.
Deutsch. Arch. fur Klin. Med., Bd. 99.
(48) Zora, L. 1910 Uber die Bedeutung und den klinischen Wert des Verhalt-
nisse zwischen Erythro- und Leukozytose, etc. Folia Haemat., Bd. 10.
Uy
~
-
hee
EXPLANATION OF PLATES
All the figures were drawn with the camera lucida at stage level with Zeiss
Apochromat 2 mm. oil immersion obj.
The compensatory ocular 4 was used
for the figure 10, the oc. 6 for the figures 7, 8, 9, 11, 12, 13, 23 and 24, the oc. 8
for the figures 15, 16, 17, 18, 19, 21 and 22 and the oc. 12 for the figures 5, 6, 14 and
20.
ABBREVIATIONS
Art., artery
Art.c., arterial capillary
Erbl., erythroblast
Erc., erythrocyte
Erc.,’’’ degenerated erythrocyte
Fbr., fibrous tissue
Grbl., granulocytoblast (myelocyte)
Gre., granulocyte (granular leucocyte)
Gre.,’"’ degenerated granulocyte
Gtc., giant cell
L.Hbl., lymphoid hemocytoblast, with
a denomination—mitosis of the corre-
sponding cell
L.HObl.,’’ lymphoid hemocytoblast in the
stage of its isolation from the mesen-
chymal syneytium
Msc., mesenchymal cells
Ms.St., mesenchymal syncytium
S., sinus
S.L.Hbl., small lymphoid hemocyto-
blast
S.Lme., small lymphocyte
I.L.Hbl., intra-vascular lymphoid he-
mocytoblast
E.L.Hbl., extra-vascular lymphoid he-
mocytoblast
PLATE 1
EXPLANATION OF FIGURES
1 Shght hypertrophy of the spleen, corresponding to the graft, reproduced
on figure 3. Eighteenth day of incubation.
2 Enormous hypertrophy of the spleen, corresponding to the graft, repro-
duced on figure 4.
Eighteenth day of incubation.
310
EQUIVALENCE OF HEMATOPOIETIC ANLAGES PLATE 1
VERA DANCHAKOPI
PLATE 2
EXPLANATION OF FIGURES
3 Spleen graft of adult tissue, growing on the surface of the allantois of a
chick embryo. The culture gave merely a slight growth.
4 Thesame. Graft intensely growing. The tumor-like graft is well provided
with vessels. The culture was made on the 7th day of incubation. The graft
is fixed on the 18th day of incubation, being 9 days old.
EQUIVALENCE OF HEMATOPOIETIC ANLAGES
VERA DANCHAKOFI
PLATE 2
PLATE 3
EXPLANATION OF FIGURES
5 Part of the spleen, adjacent to the surrounding loose mesenchyme. Large
group of hemocytoblasts within the sinus, beginning their differentiation into
erythrocytes. Eighth day of incubation.
6 Part of the spleen from the center of the organ. S.Anl., sinus anlage.
Ninth day of incubation.
314
EQUIVALENCE OF HEMATOPOIETIC ANLAGES PLATE 3
VERA DANCHAKOFF
Vera Danchakoft
PLATE 4
EXPLANATION OF FIGURES
7 and 8 Parts of a normal pulpa-like spleen. Eleventh day of incubation.
Figure 7 shows newly formed sinuses with particularly numerous lymphoid
hemocytoblasts inside. Figure 8 shows a part of the spleen with definitely
formed veinous capillaries and well developed granulopoiesis between the vessels.
9 Ingrowth of arterial capillaries into the pulpa-like spleen. Thirteenth
day of incubation.
316
EQUIVALENCE OF HEMATOPOIETIC ANLAGES PLATE 4
VERA DANCHAKOFF
Vera Danchakoff .
PLATE 5
EXPLANATION OF FIGURES
10 Part of a normal spleen, in which the follicle is adjacent to the pulpa.
11 Infiltration of the spleen tissue with granulocytes and immigration of
granulocytes into the vessels. Z, immigration of a leucocyte in the vessel. Eight
days after stimulation of an 8 day embryo.
318
PLATE 5
EQUIVALENCE OF HEMATOPOIETIC ANLAGES
VERA DANCHAKOFF
Vera Danchakoft
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
PLATE 6
EXPLANATION OF FIGURES
12 and 13 Differentiation of small lymphocytes in the follicle anlages. Fig-
ure 12 shows the center of a follicle, figure 13 a part of a follicle adjacent to the
pulpa. M.Rtc., mesenchymal reticulum; St.L.Hbl., group of lymphoid hemo-
cytoblasts forming a kind of syncytium; Y, amitotic division of a nucleus in a
reticulum cell; XY, intermediate stage between a lymphoid hemocytoblast and a
small lymphocyte. Fifteenth day of incubation.
14 Different stages of amitotical division of the nucleus; a, b, c, in mesenchy-
mal cells, d in a free cell.
320
EQUIVALENCE OF HEMATOPOIETIC ANLAGES PLATE 6
VERA DANCHAKOFF _23MRte
Vera Danchakoff IS le}
321
PLATE 7
EXPLANATION OF FIGURES
15 Intense erythropoiesis within developing sinuses. Granulopoiesis out-
side the vessels. Two days after stimulation of an 8 day embryo.
16 The same. Four days after stimulation of an 8 day embryo.
17, 18 and 19 Intense phagocytosis of erythrocytes by reticulum and endoth.
cells in figure 17, by a polynuclear macrophage in figure 18 and a lymphoid hemo-
eytoblast in figure 19.
322
EQUIVALENCE OF HEMATOPOIBTIC ANLAGES PLATE 7
VERA DANCHAKOFF
End.
Ea Verb. Gren
Vera Danchakoff g
323
4
PLATE 8
EXPLANATION OF FIGURES
20 Granuloblastic transformation of the spleen anlage 5 days after stimu-
lation of a6 day embryo. ‘ . a
21 Leucocytic infiltration of large regions in the spleen 6 days after stimu-
lation of an 8 day embryo. z
22 Reaction of the spleen tissue around large necrotic accumulations of
granulocytic tissue. Eight days after stimulation of an 8 day embryo. acl
-
Tcpracck AM
See 324
PA Stet oy
EQUIVALENCE OF HEMATOPOIETIC ANLAGES PLATE 8
VERA DANCHAKOFF
f°
¥
-
ine 325
°4e
PLATE 9 —
EXPLANATION OF FIGURES
23 Transformation of the spleen tissue into a fibrous tissue. Ten oe
after stimulation of a 9 day embryo.
24 Lymphocytoblastic transformation of a follicle 3 da ays a afte ae
of a 15 day embryo. ;
*
*
326
EQUIVALENCE OF HEMATOPOIETIC ANLAGES PLATE 9
VERA DANCHAKOFF
Vera Danchakofi an
Nar
MORPHOLOGICAL AND MICROCHEMICAL VARIA-
TIONS IN MITOCHONDRIA IN THE NERVE CELLS
OF THE CENTRAL NERVOUS SYSTEM
NORMAN CLIVE NICHOLSON
Anatomical Laboratory, Johns Hopkins University
TWO PLATES
One of the landmarks in neurology is the demonstration
by Nissl in 1885 that different types of nerve cells may be dis-
tinguished from one another by the arrangement of the baso-
philic material within them. By common consent this baso-
philic material came to be called the Nissl substance. This
departure in neurology, new at that time, proved extremely
fruitful. Investigators flocked to the study of the Nissl sub-
stance and through their investigations brought to light many
facts of fundamental importance. Since the arrangement of the
Nissl substance is more or less specific in different types of cells
it was thought that the cells might well be functionally as well as
structurally different, just as muscle cells of different structure,
gland cells and blood cells of different appearance are assumed
to function differently. Thus arose the doctrine of neurone spe-
cificity according to which it is supposed that the nervous impulse
varies in character with different cell types.
Through the recent discovery of mitochondria in the cells of
the central nervous system it has become possible to attack this
old problem from a new point of view and with greatly improved
methods of technique. The object of this investigation is to
ascertain whether the morphology and microchemical reactions
of the mitochondria vary in different types of nerve cells. It is
essentially a study of qualitative mitochondrial variation, and as
such it is supplementary to the investigation of Thurlow (’16,
p. 253) on quantitative variations. The Nissl substance as we
329
THE AMERICAN JOURNAL OF ANATOMY, VON. 20, No. 3
330 NORMAN CLIVE NICHOLSON
see it in fixed and stained preparations is undoubtedly an arte-
fact (Mott 715, p. 68, who unfortunately extends his artefact
idea to the mitochondria also), caused by the action of the fixa-
tive upon a material present as a diffuse deposit in living cells,
which fact does not, however, detract in any way from the value
of the numerous detailed and careful observations on Nissl
bodies in fixed tissues, because obviously, changes in the char-
acter of the coagulum (or precipitate) must be the visible mani-
festation of either quantitative or qualitative variations in the
diffuse deposit occurring in the living cell. But it is important
to note that the morphology of mitochondria in fixed prepa-
rations of nerve cells is, so far as can be ascertained, identical
with that seen in the living condition, which of course gives
additional value to observations on mitochondria. Then again
mitochondria are bodies quite different chemically from the
Nissl substance, playing in all probability an altogether different
role in the cell economy.
It is clear then, that although this study of their morphology
will be complementary and supplementary to the older work, it
should enable us to push our investigations much further than
is possible with studies upon the Nissl substance alone.
Comparatively little has been done thus far by other workers,
who have, for the most part, confined themselves to determi-
nations as to whether or not mitochondria do occur in adult
nerve cells and to the relations of mitochondria to other constitu-
ents of nerve cells. The only paper on the mitochondria in the
different. kinds of nerve cells, apart from several preliminary
notices of a page or more, is one by Busacea Archimede (18,
p. 322) on Testudo Graeca. No precise information has been
gathered regarding qualitative variations in mitochondria in
the nerve cells of the brain of any animals above the reptiles, or
below them, for that matter. The following investigation: was
undertaken with the hope of being able to fill up this gap to a
certain extent by careful study of a mammal.
MITOCHONDRIA VARIATIONS IN NERVE CELLS 331
MATERIAL AND METHODS
White mice of known age were used. Care was taken that they
were in good condition. They were killed with chloroform, thus
reducing the possible factor of fright to a minimum. ‘They were
then fixed by injection through the blood vessels of a mixture
of formalin and bichromate so as to guard against the production
of mechanical injury on removal of the brain as well as to insure
a good penetration of the fixative. Sections were stained ac-
cording to the method recommended by Cowdry (16 b, p. 30),
the essentials of which follow: (1) a fixation by injection, through
the blood vessels, of a mixture of neutral formalin and potas-
sium bichromate; (2) a mordanting in bichromate, followed by
dehydration and imbedding in the usual manner; and (3) a
staining of the sections, cut 4 microns in thickness, with fuchsin
and methyl green. The fuchsin stains the mitochondria a
bright crimson color and the methyl green colors the Nissl
substance in the same cell green, thus giving a good color con-
trast between the two. Specimens were also stained with iron
hematoxylin and by the Benda method, for control. They
gave results confirmatory in every way.
The observations are based upon actual measurements of mito-
chondria which were made by using one of the new Spencer wheel
‘ocular micremeters. Each measurement was made five times
and the average taken so as to reduce the experimental error as
far as possible.
OBSERVATIONS
Morphological variations. A general survey of the central
nervous system was made and it became at once evident that
the mitochondria did present considerable variation in form in
different types of nerve cells. Nerve cells of the same kind, in
the same nuciei, generally contained the same form of mito-
chondria. Somewhat more individual variation appeared in
cells of the spinal cord, spinal ganglion and Gasserian ganglion,
due perhaps to mechanical injury on removal. Mechanical
manipulation and faulty technique will bring about great varia-
Son NORMAN CLIVE NICHOLSON
tion in the morphology of mitochondria occurring in the same
type of cells, in a single center, which actually possess similar
mitochondria. In fact one comes to suspect those preparations
which exhibit 2, great dimorphism of mitochondria in cells of the
same type. Poor penetration of the fixative generally results
in the filamentous mitochondria breaking up into granules, or
else forming spherules; a phenomenon well known in other
tissues. There is no evidence that filaments are ever formed as
a result of bad technique. In all my observations a careful
check has been kept upon the technique.
In some nerve cells, like the anterior horn cells of the spinal
cord (fig. 1), and the large cells of the reticular formation of the
midbrain (fig. 11), the mitochondria are present in the form of
filaments of variable length, which frequently attain a length of
4.66 microns in the former and 6.49 microns in the latter. They
are much longer in the processes than they are in the cell body.
They are more granular in the immediate vicinity of the nucleus
where they are also more numerous. ‘This is a striking feature
of all the cells observed. It is well to remember at this point
that filamentous mitochondria occur in other types of cells also;
indeed, filamentous mitochondria occur more frequently than
any other form in the central nervous system.
Again we find cells with mitochondria in the form of granules
or short plum» rods. The cells of the mesencephalic nucleus of
the trigeminal nerve (fig. 5) are crowded with mitochondria of
this description. The granules are about 0.29 microns in di-
ameter and the filaments about 0.62 microns long. Mitochon-
dria of intermediate size and shape are invariably to be seen.
It is interesting to note that the same sort of mitochondria are
encountered in the large cells of the Gasserian ganglion, because
the nature of these mesencephalic cells has been long in dispute
and this new point of similarity between them and cells known
to be sensory constitutes additional evidence that they themselves
are sensory, and as Thurlow! also has emphasized, supports the
view that they are in reality neural crest cells which have been
enclosed in the neural tube in the course of development. This
1 Thurlow, personal communication.
MITOCHONDRIA VARIATIONS IN NERVE CELLS one
similarity is the more striking when we bear in mind that granu-
lar mitochondria are not at all common in the cells of the cen-
tral nervous system. The small cells of the Gasserian ganglion
contain exceedingly minute granular mitochondria (fig. 6) which
are usually clumped together in the vicinity of the nucleus leav-
ing the peripheral cytoplasm free. This last fact considered in
connection with Cowdry’s (’14, p. 27, fig. 13) demonstration of
mitochondria of like nature and distributed in the small cells
of the spinal ganglion, which Ranson (’14, p. 123) believes to be
concerned in conduction of pain and temperature sensations,
is significant. One might expect similar functions to be the
property of these cells, and we should be on the lookout for
changes in them in cases of trifacial neuralgia.
All nerve cells, even those with otherwise granular mitochon-
dria invariably contain filamentous mitochondria in their proc-
esses, whether they be dendrites or axones, from which it fol-
lows that there is greater variation in the mitochondria in cell
bodies than in cell processes. It may be recalled that the mito-
chondria are all filamentous in the cells of the neural tube of
the developing embryo. In other words, mitochondria retain
their embryonic form in the processes and become specialized in
the cell bodies. The mitochondria are usually filamentous in
the axone hillock.
The cells of the nucleus of the corpus trapezoideum present as
peculiar a picture as is found in any other part of the nervous
system (fig. 2). Large block-like mitochondria are found in the
peripheral layer of the cytoplasm. The large mitochondria are
frequently oblong. Some of those illustrated in the figure are as
much as 1.74 microns long by 0.63 microns in breadth. This
suggests a possible relation to the unique synaptic connection of
these cells (Collin ’05, p. 313). The peculiarity of the connec-
tion lies in the very large pericellular fibres which arborize
about the cell circumference and encompass it in a sort of cone.
Nowhere else in the nervous system do we find such large fibres
making connections of this kind. Further examination of the
cell shows, more centrally, a distinct diminution in the size
of the mitochondria which here occur as small grains or filaments,
334 NORMAN CLIVE NICHOLSON
almost beyond the limits of accurate measurement. Compara-
tively large areas of cytoplasm seem to be devoid of mitochon-
dria. An abrupt change occurs in the form of mitochondria as
we pass to the neighboring cells of the pontile nucleus, in which
the block-like mitochondria, as well as the peculiar synapses,
are absent.
It has been generally assumed (Busacea Archimede 713) that
the mitochondria occur between the Nissl bodies and not within
them. So far as I have been able to ascertain, with a method
of staining which permits of observation of Nissl substance and
mitochondria at the same time, this is not the case. Indeed one
would not expect to find the mitochondria between the Nissl
bodies in view of the fact which Cowdry (14, p. 20) emphasizes,
that the Nissl substance is present as a homogeneous diffuse
deposit in the living cell and that the Nissl bodies as seen in the
fixed preparations are produced by a process of coagulation or
precipitation.
It may be mentioned in passing that cells with the typically
filamentous variety of mitochondria (reticular formation cells,
fig. 11), the granular or rod-like mitochondria (mesencephalic
cells, fig. 5) and the blocklike mitochondria (cells of trapezoid
nucleus, fig. 2) all occur in the same section, proving that the
differences in form of mitochondria cannot be due to variations
in technique. Furthermore the variations in morphology were
found to occur constantly in all members of the species which
were examined.
Another fact worthy of note is that cells of quite different type
like the mitral cells of the olfactory bulb (fig. 3), the Purkinje
cells of the cerebellum (fig. 9), and the cells of the septum (fig.
10) all contain mitochondria of the same kind. It is clear that
variations in the form of mitochondria cannot be used to dif-
ferentiate between sensory and motor cells, nor can quantita-
tive variations be so used, according to Thurlow (16, p. 253).
This is in marked contrast to variations in the Nissl substance,
which can be used for such differentiation (Malone 718, p. 129).
Again, the general assumption that the morphology of the
mitochondria is related to the shape of the cell containing them
MITOCHONDRIA VARIATIONS IN NERVE CELLS B95)
does not seem to hold, for instance, the shape of the cells of the
corpus trapezoideum (fig. 2) and the cells of the corpus striatum
(fig. 8) is somewhat similar, yet the mitochondria in them are
quite different. An inspection of the cells of the mesencephalic
nucleus of the trigeminal nerve (fig. 5) and of the large and small
cells of the Gasserian ganglion (figs. 6 and 7) shows that the mito-
chondria are rod-like or granular when present in great abun-
dance (which would agree well with Dubreuil’s hypothesis to be
mentioned subsequently). I have not observed any change in
the nucleus, or in the nucleoli, or-in the Nissl substance, or in-
deed in any other cell structure which runs parallel with the
above mentioned variations in mitochondria and which might
offer a possible explanation of them. I have observed mito-
chondria in the surrounding cell processes but am unwilling to
state that they actually occur between the cells themselves
(that they are intercellular).
I have not found mitochondria with the bleb-like swellings,
which are so common in secreting cells; or in networks. (All
the net-like formations observed are illusory, being due to
superposition of individual filaments, which can usually be re-
solved by careful focussing.) Neither have I seen them swell
up to form vacuoles with clear centers, and there is no evidence
of an agglutination of mitochondria as occurs in pathological
conditions (Scott ’16, p. 249). I should therefore feel that the
occurrence of such mitochondria was evidence of pathological
change. !
Microchemical variations. In the inspection of a large amount
of material a certain number of brains were studied, which were
for some reason poorly fixed, and it was noted that in these,
certain groups of cells contained mitochondria while others did
not. An instance in point is that of a brain in which it was ob-
served that, while all the cells of the mesencephalic nucleus of
the trigeminal nerve contained their normal complement of
mitochondria, the neighboring cells of the locus cceruleus, scat-
tered among them were found to be devoid of mitochondria.
In order to ascertain whether there were differences in the
solubility of mitochondria beyond chance variations, the follow-
ing experiment was carried out.
336 NORMAN CLIVE NICHOLSON
Brains of mice were fixed by injection in the regular manner,
with the formalin bichromate mixture to which in one case 0.5
per cent, in another 1 per cent, in a third 2.5 per cent, in a
fourth 5 per cent and finally 10 per cent of acetic acid had been
added. ‘They were carried through and stained in the usual
way.
No mitochondria were found after using the mixture contain-
ing 10 per cent acetic acid except in the cells of the hypophysis.
This was also true in the case of the 5 per cent mixture. With
increase in concentration of aeetic acid the sections became more
and more difficult to stain and required longer and longer treat-
ment with permanganate and oxalic. The fluid containing 2.5
per cent acetic acid gave apparently the same fixation as the 5
per cent mixture, but the 1 per cent acetic mixture preserved the
mitochondria in the Purkinje cells of the cerebellum and de-
stroyed the mitochondria in the nerve cells of the medulla.
With regard to variations in staining reactions it need only
be said that we do observe mitochondria taking the stain more
intensely in certain parts of the cell. This is often the case in
the region of the axon hillock. As the staining reaction does not
occur regularly and inasmuch as it may be due to differences in
the degree of mordanting with the bichromate I am not in-
clined to attach much significance to it. Moreover when mito-
chondria are very abundant they sometimes stain more intensely,
which may be due to the presence of the stain in greater mass and
consequently washing out more slowly than where only a few
mitochondria occur. Careful search has not revealed any definite
difference in the staining reactions of the different forms of
mitochondria, although one might expect this, if differences in
morphology were assumed to be related to differences in density.
DISCUSSION
Significance of morphological variations in mitochondria. ‘The
true significance of the morphological variations in mitochondria
is unknown. Yet the demand for information is very insistent
as it is highly desirable that we should in some measure under-
stand the variations which unquestionably do occur both in
MITOCHONDRIA VARIATIONS IN NERVE CELLS 337
normal states and in pathological conditions. Two important
interpretations have been advanced.
Rubaschkin (710, p. 428) found, in the study of guinea pig
embryos, that the mitochondria were granular in the primordial
germ cells and filamentous in other more specialized epithelial
cells. He arrived at the general conclusion that the primitive
granular form of mitochondria is peculiar to undifferentiated
cells and that the process of differentiation shows itself. by a
change of the primitive granular type into chain-like and fila-
mentous forms. This view has been much criticised. It is in-
consistent with the investigations of Swift (14, p. 495) who
found that in the primordial germ cells of the chick the mito-
chondria are rodlike and do not differ from those in somatic
cells. It is sufficient merely to state that my observations, that
granular mitochondria occur constantly in some types of nerve
cells and filamentous ones in others, are also at variance with
Rubaschkin’s hypothesis, because both the types of nerve cells
in question (large cells of the Gasserian ganglion (fig. 7) and
anterior horn cells (fig. 1) ), are undoubtedly highly differen-
tiated.
Dubreuil (13, p. 137), on the other hand, is of the opinion that
granular mitochondria are in a state of rapid multiplication by
division and are characteristic of active stages in the life of the
cell and that filamentous ones are indicative of rest. He bases
this belief upon his study of the changes which the mitochon-
dria undergo in the development of fat cells from fixed connect-
ive tissue cells. He found that when the cells are most active
the mitochondria are most numerous and are granular; when
the cells are less active the mitochondria are filamentous and less
abundant. He adds to this the observation that when inflam-
mation sets up, the mitochondria immediately increase greatly
in number and are granular. The observations recorded in this
paper would, at first sight, seem to support this view, for an in-
spection of the plates reveals at once that where the mitochon-
dria are most abundant, that is to say in cells of the mesencepha-
lic nucleus of the fifth nerve (fig. 5) and in the large cells of the
Gasserian ganglion (fig. 7) they are also granular. The conten-
338 NORMAN CLIVE NICHOLSON
tion is, however, ruled out, by the fact that in other tissues the
mitochondria may still be filamentous even though they be in-
creased greatly in amount (Policard, 710, p. 284).
The variations in the form of mitochondria must be due to dif-
ferences in themselves or in their environment or in both.
There is evidence that the chemical constitution of mitochon-
dria is different in different cells. Regaud (10, p. 301) has
shown that there is a progressive increase in the resistance of
mitochondria to acetic acid in the course of spermatogenesis.
My experiments have shown that the mitochondria in the nerv-
ous system also differ in their susceptibility to acetic acid. In
the nervous system this difference in chemical behavior does
not seem to be related to a difference in morphology, as mito-
chondria of quite different form exhibit similar solubilities. So
much for chemical composition. Now with regard to density,
the only indicator which I have is the difference of intensity in
staining with fuchsin and as I have said this is not uniform and
is of uncertain meaning. That the form of mitochondria is, in
a measure dependent upon their own organization is evident
when we remember that if the long filamentous mitochondria in
the acinus cells of the pancreas are squeezed out of the cell
into the surrounding fluid they maintain their original form,
unaltered, for a surprisingly long time.
As to the differences in the cytoplasm in which the mito-
chondria are embedded I have observed that the cells of the
mesencephalic nucleus of the trigeminal nerve are more notice-
ably shrunken in some preparations than the other cells in the
vicinity, which may be accounted for on the assumption of a
higher water content. The difference in form of mitochondria
and of water content of their surroundings may not be unrelated.
As has been noted the mitochondria are invariably filamentous
in the processes, though they may sometimes be granular in
the cell bodies. This led to the belief, that there might be
a difference in water content in gray and white substance of the
brain, which curiously enough, was found, on looking up the
literature, to be actually the case. The possible influence of the
water content seems the more likely since Loéwschin (13, p..203)
MITOCHONDRIA VARIATIONS IN NERVE CELLS 339
has been able to alter the form of his artificial mitochondria,
made out of lecithin, by varying the physico-chemical proper-
ties of their environment. Since there is a general concensus of
opinion in favor of the view that mitochondria are a combina-
tion of lipoid and albumin it is possible that alkalinity or acid-
ity would affect their form. The acidity acting upon the pro-
tein fraction might cause it to become hygroscopic and to swell
- (Cowdry ’16b, p. 440). The fact that the Nissl bodies, as well
as the mitochondria, are often larger in the peripheral cyto-
plasm than-they are in the immediate vicinity of the nucleus,
would seem to indicate that some common environmental fac-
tor may be operating in the case of both, notwithstanding the
fact that the Nissl bodies are probably a coagulum or a pre-
cipitate resulting from fixation.
In a general discussion of this kind the mechanical factors
which sometimes operate, in the surrounding fluid, in shaping
the morphology of mitochondria must not be lost sight of.
Thus N. H. Cowdry has observed the changes in the form of
mitochondria in the streaming protoplasm of living plant cells.
He has seen straight filaments assume the form of loops and
spirals in response to currents and eddies in the stream, indi-
eating clearly that they are flexible and that their form is in a
measure determined by their environment. Conditions of pro-
toplasmic stress and strain, occurring especially in the course of
development, probably influence the form of mitochondria also.
In the outgrowing nerve fibers, for instance, the mitochondria
are generally filamentous. Whether mechanical factors of this
sort may play any considerable role in determining the form of
mitochondria in adult nerve cells is unknown. It is highly prob-
able that some combination of the two factors of variation in
internal composition and of changes in the surroundings are
mutually responsible for the variations in form observed.
Bearing upon doctrine of neurone specificity. These observa-
tions on variations in the morphology of mitochondria bring to
light another specific difference between the internal structure
of nerve cells of different categories; for it has already been
pointed out that these differences in the form of mitochondria
340 NORMAN CLIVE NICHOLSON
are in all likelihood associated, in some obscure way, with
changes in their environment; that is, in the cytoplasm. The
intimate bearing of such differences in the cytoplasm upon the
doctrine of neurone specificity is apparent, inasmuch as any dif-
ference in the specialized activity of a cell, ike conduction, is in
all probability related to some difference in the cytoplasm rather
than in the nucleus. This is not at variance with the results
which others have obtained.
Since Cowdry (14, p. 21) has found that there is a surprising
constancy in the mitochondria in the spinal ganglion cells of
different vertebrates, including man, it is altogether likely that
the variations in morphology which I have described in the
mitochondria in different types of cells in the brain of the white
mouse may hold in other mammals also. This seems to indicate
that when differences in the morphology of the mitochondria
occur they are not chance variations but are fundamental
differences ingrained in the very organization of the cell.
Moreover Thurlow (16, p. 253) has found, by a detailed quan-
titative study of mitochondria in the cells of the different cranial
nerves, that the actual number of mitochondria per unit vol-
ume of cytoplasm varies considerably. The largest amount of
mitochondria was found by her in the cells of the mesencephalic
nucleus of the trigeminal nerve and the least in the visceral
motor nucleus of the seventh. Even if we adopt the ultraconser-
vative view that the mitochondria are purely deutoplastic ele-
ments and do not play an active role in cellular physiology
(which I would be loath to do) it is perfectly obvious that their
presence in great amount in some cells and in very small numbers
in others, must indicate either a qualitative or else merely a
quantitative difference in the nature of the activity of the cells
in question. For it is impossible to regard the activity of a cell
as being entirely uninfluenced by the heaping up in it of inert
substances.
Pathological bearing. The practical value of this work lies in
its possible pathological application. The standardization of
material for experiment, the enumeration of the qualitative
variations which normally occur in mitochondria in the differ-
MITOCHONDRIA VARIATIONS IN NERVE CELLS 341
ent types of nerve cells of the central nervous system of a mam-
mal and the mention of the forms of mitochondria which should,
in any experimental study of the nervous system, be regarded
as pathognomonic (i.e., net-works, bleb-like swellings, aggluti-
nations and vacuolations), are all of interest from a pathologi-
cal point of view. The necessity of taking cognizance of mito-
chondrial changes is brought home to the clinician through the
recent investigations of Goetsch (’16, p. 132) who found that an
increase in the mitochondria in the thyroid epithelium was asso-
ciated with an increase in the activity of the epithelial cells and
with the severity of the clinical symptoms of hyperthyroidism
in man.
As nothing has yet been done on the changes in the mitochon-
dria in the nerve cells of man in pathological conditions it is quite
clear that the following are merely suggestions.
If Regaud (711, p. 20) is at all justified in his statement that
“by a still unknown physico-chemical mechanism, the mito-
chondria retain a great variety of substances which come in con-
tact with the protoplasm, normally as well as accidentally
(medicines, poisons, toxins, ete.)’’ and if the conclusion is war-
ranted that the point of action of tetanus toxin on the nerve cell
is lipoidal, it is quite possible that a study of mitochondria in
the nervous system in tetanus might yield interesting results.
At any rate the fact that tetanus toxin is rendered innocuous when
mixed with an emulsion of brain pulp (Wassermann and Takaki
98) is evidence that the toxin combines with some component
of nerve tissue. Leathes (’10, p. 123) believes that it acts upon
a fat which he calls ‘cerebrone.’
There are up to the present no observations on mitochondria
in nerve regeneration, although there is evidence from many
sources that the mitochondria are delicate indicators of cell
activity, and may reveal interesting facts with regard to the
formation of myelin; etc.
Pathologists have frequently noted that the mitochondria
are the first structures in the cell to respond to disturbances in
function, which suggests the possibility that a study of them in
the nervous system may serve to localize brain lesions which
342 NORMAN CLIVE NICHOLSON
heretofore have eluded the grasp of neurologists. One would
naturally inquire first into those diseases in which chemical
analysis has revealed a disturbance in the lipoid content, particu-
larly general paralysis since Koch and Mann (’09) have detected
a destruction of the brain phosphatides in this condition and there
is a good deal of evidence that the mitochondria are, themselves,
closely related to phosphatides.
Before concluding I want to express my deep appreciation to
Dr. E. V. Cowdry for his friendly interest and encouragement.
SUMMARY
There are qualitative differences in the mitochondrial content
‘of certain types of nerve cells in the brains of white mice. The
variation in morphology between cells of different variety is often
quite pronounced. Filamentous mitochondria are the most
common form met with in the cells of the central nervous sys-
tem. They are particularly apparent in large anterior horn
cells (fig. 1) and in the large cells of the reticular formation (fig.
11). Rod-like and granular mitochondria are rarer. They are
characteristic, however, of the cells of the mesencephalic nu-
cleus of the fifth nerve (fig. 5) as well as of the cells of the Gas-
serian ganglion (figs. 6 and 7). The cells of the nucleus of the
corpus trapezoideum (fig. 2) may be distinguished by their large,
swollen block-like mitochondria.
There is also, in the majority of cases, a variation in the form
of mitochondria in different parts of the same cell. For instance,
they are usually more granular in the vicinity of the nucleus
than in the peripheral parts of the cytoplasm, and in the pro-
cesses. In the processes they are invariably rod-like or fila-
mentous. This is shown in most of the drawings but it is par-
ticularly well illustrated in figures 1, 5, 9, 10, 11 and 12. - The
cells of the nucleus of the trapezoid body constitute a special
case because in them the mitochondria always occur in the form
of long blocks in the peripheral cytoplasm in sharp contrast to
the minute granular and rod-like mitochondria in the imme-
diate neighborhood of the nucleus. The mitochondria occur not
MITOCHONDRIA VARIATIONS IN NERVE CELLS 343
only between the Nissl bodies (as is generally believed) but also
embedded in them.
They also vary in different kinds of nerve cells microchemi-
cally. The most striking instance is the greater resistance of
the mitochondria in some cells to fluids containing acetic acid
as compared with other cells of different types. The mitochon-
dria in the different parts of the cytoplasm of the same cell
react in the same manner to solvents.
BIBLIOGRAPHY
Busacca ARCHIMEDE 1912 L’apparato mitocondriale nelle cellule nervose
adulte. Anat. Anz., Bd., 42, pp. 620-622.
1913 L’apparato mitocondriale nelle cellule nervose adulte. Arch.
Zellforsch., Bd. 111, pp. 327-339.
Couuin, R. 1905 Sur les Arborisations Pericellulaires dans le Noyau du Corps
Trapezoide. Biblio. Anat., t. 14, pp. 311-315.
Cownry, E. V. 1914 The comparative distribution of mitochondria in spinal
ganglion cells of vertebrates. Am. Jour. Anat., vol. 17, pp. 1-29.
1916 a The general functional significance of mitochondria. Am.
Jour. Anat., vol. 19, pp. 423-446.
1916 b The structure of chromophile cells of the nervous system,
Contributions to Embryology, Carnegie Institution of Washington,
No. 11, pp. 27-48.
DusreviL, G. 1912 Le chondriome et le dispositif de l’activité séerétoire aux
differents stades du développement des éléments cellulaires de la
lignée connective, descendants du lymphocyte. Arch. d’Anat. Micr.,
t. 15, pp. 53-151.
GortscH, EH. 1916 Mitochondrial changes in toxic adenomata of the thyroid
gland. Johns Hopkins Hospital Bull., vol. 27, p. 129-133.
Leatues, J. B. 1910 The fats. Monographs in Biochemistry, New York, pp.
138. -
Lowschin, A. M. 1913 ‘Myelinformen’ und Chondriosomen. Ber. d. Deut.
Bot. Ges., Bd. 31, pp. 203-209.
Luna, E. 1913 I condriosomi nelle cellule nervose adulte. Fol. Neur. Bio.,
’ Bd. 7, pp. 505-511.
Matonr, Epwarp F. 1913 Recognition of members of the somatic motor
chain of nerve cells by means of a fundamental type of cell structure
and the distribution of such cells in certain regions of the mammalian
brain. Anat. Rec., vol. 7, pp. 67-82.
Morr, F. W. 1915 Microscopic examination of the central nervous system in
three cases of spontaneous hypothyroidism in relation to a type of in-
sanity. Proc. Roy. Soc. Med., vol. 8 (Sec. of Psych.), pp. 58-70.
Poticarp 1910 Contribution a l’étude du méconisme, ete. Arch. d’Anat.
Mier., t. 12, pp. 177-288.
344 NORMAN CLIVE NICHOLSON
Ranson, 8S. W. 1914 The tract of Lissauer and the Substantia gelatinosa
Rolandi. Am. Jour. Anat., vol. 16, pp. 97-126.
Recaup, Cr. 1910 Etude sur le structure des tubes séminiféres et sur le sper-
matogénese chez les mammiféres. Arch. d’Anat. mier., t. 11, pp. 291-
431.
1911 Les mitochondries. Rev. de Med., t. 31, pp. 681-699.
RUBASCHKIN, 1910 Chondriosomen und Differenzierungsprozesse bei Séuge-
tierembryonen. Anat. Hefte, Bd. 41, pp. 401-437.
Scorr, W. J. M. 1916 Mitochondrial changes in the pancreas in experimental
phosphorus poisoning. Am. Jour. Anat., vol. 20.
Swirt, C. H. 1914 Origin and early history of the primordial germ cells in
the chick. Am. Jour. Anat., vol. 15, p. 483.
Tuurtow, M. DeG. 1916 Observations on the mitochondrial content of the
cells of the nuclei of the cranial nerves. Anat. Rec., vol. 10, p. 253.
WASSERMANN and Taxkaxt, T. 1898 Tetanusantitoxische Eigenschaften des
Zentral-Nervensystems. Ber. Klin. Woch., Bd. 35, pp. 5-6.
MITOCHONDRIA VARIATIONS IN NERVE CELLS 345
‘DESCRIPTION OF PLATES
The figures have been drawn from cells in the central nervous system ofa
female white mouse, thirty-five days old, weighing 9 grams, body length 6.8 em,
and tail length 6.1 em., fixed in neutral formalin and bichromate and stained with |
fuchsin and methyl green. By this method the mitochondria are stained bright
crimson against a green background of Nissl substance. The mitochondria are.
represented in black in the illustrations which were drawn with Zeiss apochro-
matie objective 1.5 mm., compensatjng ocular 6 and camera lucida. The figures
are reproduced without reduction so that they represent an actual magnification
of 1660 diameters as they now appear on the plates. They are intended to illus-
trate differences in the morphology and in the arrangement of mitochondria, —
which are peculiar to and which are constantly met with, in certain types of
nerve cells in the brains of white mice.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, Nu. 3
PLATE 1
EXPLANATION OF FIGURES
1 Large anterior horn cell of the spinal cord in which the mitochondria are
typically filamentous. They are longer in the processes than they are in the
neighborhood of the nucleus. ‘They may be seen imbedded in the flake-like Nissl
substance.
2 Cells from the nucleus of the corpus trapezoideum. ‘The striking feature
is the presence of relatively enormous block-like mitochondria in the peripheral
cytoplasm, leaving certain areas devoid of mitochondria. It is to be noted fur-
thermore that the mitochondria are quite minute in the immediate vicinity of the
nucleus. <A large pericellular arborization is visible between the two cells.
3 A mitral cell from the olfactory bulb. Filamentous mitochondria are to
be seen embedded in a homogeneous background of Nissl substance.
4 Large pyramidal cell from the hippocampus showing practically the same
arrangement of mitochondria as in the preceding.
5 A large cell of the mesencephalic nucleus of the fifth nerve witha cell of
the locus coeruleus immediately adjacent. This large cell is among the most
remarkable seen in the whole nervous system. It contains large numbers of
small rod-like mitochondria and the striking resemblance which it bears to the
large cells of the Gasserian ganglion (fig. 7) is at once apparent.
6 Small cell of Gasserian ganglion with minute granular mitochondria
clumped about the nucleus.
7 Large cell of Gasserian ganglion containing an abundance of granular
mitochondria in sharp contrast with the filamentous mitochondria which occur
in the majority of other nerve cells.
346
MITOCHONDRIA VARIATIONS IN NERVE CELLS PLATE 1
NORMAN CLIVE NICHOLSON
S :
Py pc a
FS A ane
i i : \w S s
wos Wek tt
( Pa
y M,
ie AN {
} » SA
Api eas |
a Pa ages ae Z
Al 4 14 1 SOx. ok : v4 v 4\
,- } ») = . : y nse ? x)
is //® ’
a“ Z ‘ \
IG fd “)
¥ 3 - )
\ @ 5 os
Se /
j =
saan
PLATE 2
EXPLANATION OF FIGURES
S ‘Two cells of the corpus striatum. They contain spherical and rod-shaped
mitochondria which, in fact, resemble closely the fine medullated fibres cut in
section in the neurophil about them. To the left two red blood cells are to be
seen in a capillary.
9 Large pyramidal cell of the cortex cerebri with its contained mitochondria
which are quite filamentous. This cell also is embedded in a mass of fibers which
are cut in section and which resemble mitochondria very closely.
10 A cell from the septum.
11 Large cell from the formatio reticularis of the mid brain. The mito-
chondria are thread-like and remind one of those which occur constantly in an-
terior horn cells. They are, as is invariably the case in all the cells studied,
more filamentous in the processes than in the cell body.
12 Purkinje cell of the cerebellum with neighboring granule cells Just be-
neath it. The mitochondria look like minute bacilli near the nucleus but they
become elongated as one proceeds toward the great dendrite. The granule cells
also contain them. The black dots in the molecular layer are cross sections of
fibres.
MITOCHONDRIA VARIATIONS IN NERVE CELLS PLATE 2
NORMAN CLIVE NICHOLSON
7S gears
PSF ESA TS on) + BP: —
4s reais 1% ieee On ees ee oe
oF Ae SES, ame
»'
Fé
349
’
= te
= 7
" ; =
; - - :
i oa 2 os Sel
. = 7 , - « ne -
s a ’ = - u a
é ‘ = ee
i = t = - ve Pay) - 2
= ” f 7 os
2 ' = a
“ =
, ad ma .
~ FR —
7 .
’ ~—_e
= 4 > - v
ON THE DEVELOPMENT OF THE ATRIAL SEPTUM
AND THE VALVULAR APPARATUS IN THE RIGHT
ATRIUM OF THE PIG EMBRYO, WITH A NOTE ON
THE FENESTRATION OF THE ANTERIOR CAR-
DINAL VEINS
C. V. MORRILL
Department of Anatomy, Cornell University Medical College, New York City
NINE FIGURES
The method of formation of the atrial septum and nearly re-
lated parts in the mammalian heart has been known since the
careful investigations of Rose (’88 and ’89), Born (’89), and
more recently Favaro (713). These writers are in substantial
agreement regarding the structures concerned in this process,
though differing in some respect as to detail. The now classi-
eal work of Born on the rabbit has been used extensively as a
basis for descriptions of cardiac development in some of the
more recent text-books of embryology, notably by Hoch-
stetter (01-03) and Tandler (12 and 713). Born in his ac-
count, which included some observations on human embryos,
corrected the error of His (’85) regarding the formation of the
atrial septum and its relation to the foramen ovale. His de-
seription differed from the earlier one of Rése (’88) in several
important respects but later Rose (89) accepted Born’s corree-
tions. Rése and Favaro extended their observations to a num-
ber of different mammals, but not, as far as I am aware, to the
pig. Retzer (’08) published a brief note on the development of
the heart in which he claimed that Born’s account of the atrial
septum in the rabbit could not be applied to the pig. Since pig
embryos are extensively used for studyin American laboratories,
it seemed advisable to re-examine the development of the septal
351
3O2 Cc. V. MORRILL
and valvular apparatus in this form in order to clear up any
uncertainty which may exist regarding the processes involved.!
This study is based on series of pig embryos of 6.8, 7.9, 8.5,
12.3, 15.2 and 21 mm. total lengths, supplemented by dissec-
tions of the hearts of foetuses from the 45 mm. stage to birth.
The hearts of the 7.9, 15.2 and 21 mm. embryos were recon-
structed by the wax-plate method. In addition several human
embryos from 7.6 to 22 mm. were examined for comparison.
The terminology of Born which has been largely adopted by
other writers will be used, with some modifications, in this
account.
In the stage represented by a 6.8 mm. embryo, septum I
forms an incomplete interatrial curtain. Ostium I is. still
widely open and ostium II has begun to form. In the recon-
struction of a 7.9 mm. embryo (fig. 5), the relation of these
three structures is shown. Septum I (S./) though fused for the
most part with the atrial walls, still presents a short, free bor-
der facing the now very narrow ostium I (0.7). In the pos-
terior superior corner of the septum, the new opening or ostium
II (O.IT) is well advanced and much larger at this time than
ostium I. It is apparent from the figure that ostium II is
formed by fenestration in the pig and is not at first a single
opening as Born found in the rabbit. Born thought that a
primitively single opening was normal for mammals generally,
but Roése found a fenestrated septum in three cow embryos, a
mole embryo and two rabbit embryos and states that Bruch’s
observations on sheep, cow and horse and Rokitansky’s? on a
human embryo support the conclusion that fenestration is the
more usual condition. Favaro found numerous orifices, repre-
senting ostium IT in the sheep, but only a single opening in the
guinea-pig. As a rule, Rése believes, the larger openings coa-
lesce to form a single ostium II while the smaller ones close up.
‘This investigation was begun in the anatomical department of the Univer-
sity and Bellevue Hospital Medical College and completed at Cornell Univer-
-sity Medical College, New York.
2 The observations of Bruch and Rokitansky are known to me only through
the brief mention made by Rose (’88).
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO one
{xceptionally the Jatter may persist until birth. In the pig
later stages always show a single opening.
The closure of ostium I in the pig is effected in the same
manner as described for the rabbit, guinea-pig, man and other
mammals. This process is illustrated in figures 1, 2, 3, and 5.
Septum I is composed for the most part of developing muscle
which is continuous with the musculature of the roof, posterior
wall and floor of the atrium. Its free border, however, is capped
by a prominent endocardial thickening. As the septum grows
downward and forward toward the atrial canal, the endocardial
thickening develops two extensions or horns which fuse with the
endocardial cushions of the atrial canal. The upper one blends
with the upper cushion (fig. 1), the lower one with the lower
cushion (fig. 3). Ostium I (fig. 2) is thus entirely surrounded
by thickened endocardium except where it is continuous with
the narrow slit (transverse fissure) between the endocardial
cushions of the atrial canal. At a slightly later stage (8.5 mm.)
the blending of the separate portions of endocardium is com-
plete and ostium I together with the transverse fissure is closed.
In the 15.2 mm. stage (fig. 6) ostium IT (O./7) is much enlarged
and forms a single large opening between the atria. Septum I
(S.J) presents a very irregular free border which faces upward
and slightly posteriorly toward ostium II. (The irregularity of
the free border could not be represented in wax.) The septum
as a whole bends toward the cavity of the left atrium.
The mode of formation of septum II (the later limbus Vieus-
seni) is first clearly indicated in the 7.9 mm. embryo though
a faint trace is sometimes distinguishable in an earlier stage.
Concerning the origin of this septum there has been some dif-
ference of opinion. Indeed Retzer (’08) has denied its existence
in the pig, though Rose (88 and ’89), Born (’89) and Favaro
(13) have found it in all the forms they studied, Rése expressly
stating that it is characteristic of all placental mammals. In
the present study it will be shown that septum II is as definite
and well developed in the pig as in any form so far examined.
The model of the 7.9 mm. embryo (fig. 5) shows a very dis-
tinct spur-like thickening (S.J7) which projects into the cavity
354 Cc. V. MORRILL
Figs. 1 to 3. Transverse sections of the heart region of a 7.9 mm. embryo
(histological structure semi-diagrammatic). Figures 1 and 3, above and below
the level of ostium I respectively; figure 2, through ostium I. A., dorsal aorta;
A.d., right atrium; A.s., left atrium; Hn.s., upper endocardial cushion; En.7.,
lower endocardial cushion; 0.7, ostium I; S.in., interventricular septum; S.J.,
septum I; S.J/J, septum II; Sn.d., right sinus horn; Sp.in., spatium intersepto-
valvulare; V.car.c.d., and V.car.c.s., right and left common cardinal veins;
V.v.d., right sinus valve; Vv.v., sinus valves. Magnified about 25 diameters.
Vt
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO Be
of the right atrium in the angle between the floor and septum
I, and just posterior to ostium I (0.7). It is continuous with the
left sinus valve. ‘Transverse sections through this region (figs.
1, 2 and 3) show that at first this spur (S.J/) is composed of
connective tissue alone and is sharply marked off, histologically,
from the developing musculature of the sinus valves, septum I
and the atrial floor. It appears to be nothing more than a pro-
jection, toward the right, of the endocardial thickening of sep-
tum I (S.J) and is continuous through the latter with the endo-
eardial cushions of the atrial canal (Hn.s. and Hn.i.). This
structure which is the anlage of septum II, is undoubtedly what
His (’85) called the spina vestibuli and, taken together with the
endocardial cushions, would constitute his septum intermedium.
The lower end of the left sinus valve blends with it, while the
corresponding end of the right valve flattens down in the atrial
floor close to it (figs. 2 and 5, Vv.v.). At this stage, the anlage
of septum II is not in relation with ostium II.
In later stages, when ostium I has completely closed, the
spur lengthens out into a definite ridge. This is well shown in
the medel of a 15.2 mm. embryo (fig. 6, S.JZ). When traced
from its point of origin in the anterior inferior corner of the
atrium, this ridge extends upwards then bends backward along
the roof toward the posterior wall where it flattens down and
may become continuous with two or three muscular trabeculae
which develop in this region, 1.e., the spattum intersepto-valvu-
lare. These trabeculae are not constant in number. The por-
tion of the ridge in relation to the roof of the atrium? is broad
and low contrasted with its lower anterior end or root which is
sharp and well-defined. Both sinus valves (V.v.s. and V.v.d.)
now blend with it, the right, however, slightly covering its
lateral side.
The internal structure of septum II is shown in a transverse
section taken at about the middle of its anterior vertical por-
tion (fig. 4, S.J7). The endocardial thickenings which entered
into its formation in an earlier stage, have fused together into
3 Owing to the position of the model this part of septum II is hidden in
figure 6.
356 Cc. V. MORRILL
a solid quadrilateral mass. A cap of developing muscle covers
its posterior border (upper in the figure) and extends anteriorly
for some distance along its lateral surfaces. This muscular
mass is continuous with the musculature of septum I and, at a
lower level (not shown in the figure), with that of the two sinus
valves which, as mentioned, blend with the root of the septum.
A eareful study of serial transverse sections shows that the
musculature of this part of septum II is derived from that of
Fig. 4. Transverse section of the heart region of a 15.2 mm. embryo at the
level of the root of septum II (histological structure semi-diagrammatic). A.,
dorsal aorta; A.d., right atrium; A.s., left atrium; Hn.c., fused endocardial cush-
ions; S.in., interventricular septum; Sp.in., spatium intersepto-valvulare; S./.,
septum I; S.JJ, septum IT; V.car.c.d., and V.car.c.s., right and left common
eardinal veins. Magnified 25 diameters.
~
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO Balt
the sinus valves and septum I. The developing muscle elements
appear to invade the connective tissue and gradually to replace
it. This invasion has not yet begun in the 8.5 mm. embryo but
is well advanced in the 12.8 mm. and s'‘ill more so in the 15.2
mm. described above. The anterior portion of the connective
tissue Mass or septum intermedium? (fig. 4 Hn.c.) broadens out
in a lateral direction between the two atrio-ventricular orifices
and is joined by the interventricular septum, the line of fusion
being toward the right. The fusion is not complete, however,
for the interventricular ostium (fig. 6, O.in.) is still widely open
at this stage. If septum II is now traced upward and back-
ward into the roof of the atrium, the connective tissue elements
gradually disappear and it appears as a low, broad thickening
of the atrial musculature which follows the external depression
caused by the bulbus in this region. It is not, however, due
to an infolding of the atrial wall, but to a local increase in the
developing muscle elements. This upper part of septum II
now comes into relation with ostium II (fig. 6, O.J7) which
meanwhile has become much larger.
Passing now to the 21 mm. embryo, the further development
of septum ITI (fig. 7, S.J7) is clearly shown. It appears as a
erescentic ridge extending from the lower inferior corner of the
atrium, upward and backward toward the posterior wall. Its
lower segment or root which is thickest and most sharply de-
fined is partly hidden by the high right sinus valve (V.v.d.)
which blends with its lateral surface. Traced upward it curves
over the roof of the atrium bordering ostium II (O.J7) and
reaching the posterior wall, bends downward somewhat in the
region of the spatium intersepto-valvulare (not shown in the
* The invasion of the connective tissue mass (septum intermedium) by the
musculature of the sinus venosus was described by Retzer (’08) as the source of
the Purkinje fibres in the ventricles. The present investigation leads to the
conclusion that a part of this muscle, at least, thickens to form the root of
septum IT. '
° The term ‘septum intermedium’ originally used by His (’85), may be con-
veniently retained for this structure as suggested by Favaro (713), although
Born and some others discarded it.
° A portion of the septum intermedium thus separates the right atrium from
the left ventricle (conus arteriosus of the aorta) forming the septum atrioven-
triculare of Hochstetter.
358 Cc. V. MORRILL
figure). This upper segment is broad and low where it faces
ostium II but becomes narrow and sharp in the posterior wall.
Septum I meanwhile has changed its position relative to septum
II, so that its free border now faces upward and forward, and
ostium IT has become an oblique cleft, the definitive foramen
ovale, between the septa.
Before considering the later changes in septum II it will be
convenient to compare the conditions found in the pig with
those described for other forms. According to Born septum II
first appears (rabbit and man) in the upper part of the posterior
wall of right atrium a little to the mght of septum I. It is a
crescentic spur which encroaches upon the spatium interseptale
and forms the principal part of the limbus Vieussenii. In the
pig irregular muscular ridges develop in the same locality but
they vary greatly in form, size and number. In the model of
the 15.2 mm. embryo, two such ridges can be distinguished
(fig. 6). When septum IT has extended backward to the posterior
wall, one or more of these ridges is incorporated in it, but no one
of them is sufficiently well-marked to be taken as a starting
point for the developing septum. Further, at the time when
the anlage of septum II is first distinguishable in the lower
anterior wall (7.9 mm. embryo) the spatium intersepto-valvu-
lare is relatively very wide and numerous small irregular ridges
appear in its upper posterior wall. Along with the narrowing
of the spatium, the ridges decrease in number until finally they
are reduced to one or two which, as stated, are incorporated in
the septum.
An examination of several human embryos seemed to indicate
that much the same sort of process takes place there. In an
embryo of about 12 mm., total length, at least three muscular
ridges could be distinguished in the upper, posterior wall of the
spatium intersepto-valvulare, while in one of about 14 mm.
only one such ridge appeared which, however, was quite promi-
nent.’ In the lower anterior part of the right atrium of both
7 Through the kindness of Dr. Thyng I have had the opportunity of examin-
ing the human embryo which formed the subject of his paper (14). In this
specimen (17.8 mm.), the region of the spatium intersepto-valvulare was entirely
devoid of muscular ridges.
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO 359
these embryos there was a thick connective tissue ridge or promi-
nence continuous with the fused endocardial cushions and
partly overlaid and invaded by developing muscle. As in the
pig this musculature was continuous with that of septum I
and the sinus valves. The resemblance between this structure
and that described for the pig in the corresponding place is so
close that I am inclined to think it represents the origin of
septum II in man. Indeed Thyng (14) in his description of a
human embryo notes the presence of a ‘‘ridge or tubercle’ in
the same region and states that ‘‘from the relation which the
tubercle bears to the septum primum and the left sinus valve
it can scarcely be doubted that it would eventually form part
of the adult limbus fossae ovalis.”’
Born seems to have overlooked the early appearnace of this
ridge entirely, though he states that in man the bay of the
crescent (septum II) swings further downward in later stages
and finally unites with the lower end of the left sinus valve.
He thus includes the region in which the ridge is found in the
pig. In the rabbit, however, septum II remains as a narrow
short crescent in the anterior upper wall of the atrium. One
may conclude from this that the ridge if present at all is poorly
developed in this animal.
Rose (89) in summarizing the results of his earlier work (88)
speaks of septum II (limbus or annulus ovalis) as follows:
On the anterior and upper atrial wall there appears a ridge-like
infolding, the septum musculare which, together with the septum inter-
medium . . . . . is formed into a closed ring diaphragm, the
annulus ovalis . . . . . The septum intermedium arises from a
connective tissue spina vestibuli overlaid, however, on its upper surface
by a thin continuous muscle layer, which unites with the broad atrio-
ventricular cushions, the latter fused together in the middle.
It is apparent that this description corresponds very closely
with that given for septum II in the pig with the exception that
the upper part of this septum corresponding to Rése’s septum
musculare is not formed by an infolding of the atrial wall
but by a local thickening. In his later paper (89) following
Born’s work, R6ése modified his description of the atrial septa
360 CV.’ MORRIGI
extensively. He rejected the idea that the septum intermedium
is concerned in the formation of the limbus or annulus ovalis.
He also discarded the term ‘septum musculare,’ calling it simply
the limbus (Vieussenii) and described it as arising in the anterior
upper atrial wall as an infolding caused by the truncus arterio-
sus (bulbus) imbedding itself between the two atria. From here
it spreads along the upper and then along the lower wall of the
right atrium. It contains a connective tissue core which was
pinched off from the surface of the truncus during the process
of infolding. The presence of this core, Rése thinks, is a proof
that infolding has occurred.
In the pig the upper part of septum II or limbus which cor-
responds to the curve of the bulbus, is at first a broad thicken-
ing of the musculature of the atrial wall which in later stages
becomes more prominent. Only in its lower anterior part or
root are connective tissue elements found. These are the re-
mains of the original thickened endocardium which was shown
to form the basis for the septum. Retzer (08) in denying the
presence of a septum IT in the pig says: ‘‘ This supposed septum
which His correctly described as a ‘muskulése Leiste’ is nothing
but a fold in the atrial wall at that place.” It is formed by
the atrial growing around the conus arteriosus as a fixed point,
thus causing a bulging inwards of the atrial wall.
The descriptions of Rése and Retzer thus correspond quite
closely as far as the origin of the muscular fold is concerned.
Retzer, however, thinks it never attains sufficient size to be
called a septum. His (’85) described the same structure, call-
ing it the anterior septum or ‘sickle’ which later gives rise to
the limbus Vieussenii. He thought that it was formed by the
septum intermedium growing up on the anterior atrial wall to
meet the anterior end of the septum superius (septum I). This
upgrowth of the septum intermedium, or a portion of it over-
laid by muscle, is practically what takes place in the pig, but
as pointed out previously, there is no evidence of an infolding
of the atrial wall at a higher level.
Favaro’s (’13) recent account of the septum secundum in the
euinea-pig and sheep differs very little from that given above
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO 361
for the pig. He finds that the myocardium which invades the
connective tissue of the septum intermedium becomes more
compact and raised up into a distinct prominence on the right
side of the insertion of the atrial septum (septum I). This, he
says, represents the site of the inferior segment of the limbus
fossae ovalis. The prominence is continued upward as a thick-
ening of the myocardium of the ventral wall of the atrium on the
right of septum I. It corresponds to a broad sulcus externally
but there is no infolding as described by Rése and Retzer. In
the guinea-pig and sheep, however, Favaro finds that the septum
spurium or tensor valvulae bends over the upper wall to join
the newly-found limbus or septum II. This does not occur in
the pig as reference to figures 5, 6 and 7 will show. In this
form the tensor valvulae’ (7.vv.) extends upward to the roof
of the atrium where it is lost sight of among the developing
trabeculae in that region. In the human embryos examined
the course and relations of the tensor seemed to be more like
those in the guinea-pig and sheep as described by Favaro.
From the first the lower end of the left sinus valve blends
with septum II as mentioned previously. The right valve is
at first independent of the septum (fig. 5) but later when the
septum has increased in height, the valve becomes attached to its
right side (figs. 6 and 7). In later stages the right valve in-
creases rapidly in height so that when seen from the right side,
it completely covers the lower end of the left valve and the root
of septum II (fig. 7). When a portion of the right valve is cut
away (fig. 8) the connection between the left valve and septum IT
can be seen. Separating the orifices of the inferior vena cava
(V.c.z) and the coronary sinus’ (Sn.cor.) is a thick ridge, the
sinus septum (S.sn.), which extends from the root of septum II
downward and to the right toward the attached border of the
5’ The term ‘tensor valvulae,’ first proposed by Rése (’89), seems more ap-
propriate than the older ‘septum spurium’ of His. This structure probably
helps to approximate the valves and render them tense during systole.
° When the model was sawn through, the cut opened the right extremity of
the coronary sinus which forms a deep bay as it turns backward and upward to
its orifice (figs. 7 and 8).
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
362 Cc. V. MORRILL
right sinus valve, where it blends with the floor of atrium.
Owing to the extreme height of the right valve the inferior vena
cava and coronary sinus appear to open into a common chamber
flanked by the valve.
The further changes in the right atrium may now be consid-
ered. From the 21 mm. stage (figs. 7 and 8) up to about 58
mm. the relations of the parts in question change very little.
Septum I or the valvula foraminis ovalis, as it may now be
called, becomes thinner and somewhat folded. Its free border
is fimbriated. The left sinus valve is still well developed, its
lower end blending with the root of septum II. It now les so
close to the septal wall of the atrium that the spatium inter-
septo-valvulare is reduced to a narrow cleft. The right sinus
valve remains high, guarding the orifices of the inferior cava
and coronary sinus. The sinus septum between the two orifices
is behind and medial to the valve but not fused with it. In
two specimens (58 to 62 mm.) and possibly a third, a new open-
ing appeared in the lower anterior part of the right sinus valve
close to its attached border. The course of the blood-stream
from. the coronary sinus in this case was directly into the lower
anterior part of the atrial cavity rather than upward, back-
ward and medially toward the foramen ovale. The usual path-
way, however, was not shut off. His (’85) described a secondary
opening of the coronary sinus in human embryos and assumed
that it represented the normal condition. Other writers have
not found this new opening, Born expressly denying its exist-
ence. It probably occurs exceptionally in man as in the pig
and may represent a premature degeneration of this part of the
valve.
The tensor valvulae (septum spurium) in the older foetuses
becomes flattened down considerably in its upper part, though
still fairly well marked near the sinus valves. Meanwhile a
new muscular ridge, the crista terminalis has appeared in the
atrial wall just to the right of the right valve. Its upper part
blends with the trabeculae in the roof of the atrium near the
tensor valvulae. Traced downward it passes first close to the
attached border of the valve and then diverges from it more
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO 363
and more until it is lost among the trabeculae of the lateral
wall of the atrium. The right sinus valve does not seem to
take any part in its formation, though both Rose and Tandler
(13) consider that the upper part of this valve is involved.
The final disposition of the various parts of the septal and
valvular apparatus is shown in figure 9 which represents a dis-
section of the heart of a foetus of about 85 mm. Owing to the
preponderating growth of the upper part of the atria, the region
of the septa appears sunken into the entrance of the inferior
cava, while the orifice of the superior cava (V.c.s.) remains high
up on the posterior wall. Septum I, now the valvula foraminis
ovalis (V.f.o.) lies in the bottom of a slight depression or fossa
whose margin is formed for the most part by the limbus fossae
ovalis (Z.V.). The free, fimbriated border of the valvula
extends obliquely into the cavity of the left atrium and overlaps
the limbus on that side. The oblique cleft between them is the
foramen ovale. Above the fossa and extending downward on
each side of it, the limbus is thick and muscular. This part
is formed by septum II.
Continuous with the anterior end of septum II is the left
sinus valve (V.v.s.) which in this place is muscular like the
septum and may be said to form part of the limbus. When
traced backward and upward around the fossa, the valve be-
comes reduced to a thin, pale streak which finally crosses the
posterior part of the limbus proper and then extends along the
left margin of superior caval orifice where it gradually fades out.
This is practically in agreement with the descriptions given by
Born, Rése and others.
The right sinus valve (V.v.d.) is still well marked at this
stage. Its lower anterior part bounds the orifice of the inferior
cava laterally and extending forward and medially partially
covers the orifice of the coronary sinus. The first portion rep-
resents the valvula Eustachii, the last (X in fig. 9) the valvula
Thebesii of other mammals. In the pig the right sinus valve
never becomes divided structurally. There is always a nar-
row cleft between it and the sinus septum (S.sn.). The upper
part of the right valve extends as a thin narrow membrane
364 Ci “Ve MORRILE
along the right border of the superior caval orifice almost to its
upper end. The tensor valvulae has entirely disappeared.
At this stage the crista terminalis (Cr.ter.) forms a thick ridge
extending downward from the roof, in the angle between the
posterior and lateral walls of the atrium. It was pointed out
that when the crista first appears it is entirely independent of
the right sinus valve, but as it becomes broader and more fully
developed, the extreme upper part of the valve comes to le
right on its root. There is usually, however, a narrow cleft
separating it from the remains of the valve below. For this
reason, the crista terminalis of the pig cannot be said to mark
exactly the boundary between the primitive sinus cavity and
the atrium proper as has been maintained by His, Rése and
Tandler for man.
A glance at figure 9 will show that there is a broad, rounded
ridge in the angle between the orifices of the superior and inferior
venae cavae (the line marked V.v.s. passes across it). It is
formed by a thickening of the musculature of the atrial wall
in this region and represents the tuberculum intervenosum
(Loweri) of the human heart. There has been some doubt
about the occurrence of this structure but recently Tandler (713)
came to the conclusion that it is always present in the adult
human heart if hardened in situ, ie., with the pericardium
intact. He suggests the term ‘torus Lowerl’ as more appro-
priate. In other mammals he finds it more strongly developed
than in man, especially in the horse and still more so in the seal
where it forms a ‘veritable septum.’ In the pig this ridge
occurs constantly in older foetal stages. Owing to its position
it would tend to direct the blood-stream from the superior cava
toward the right atrio-ventricular orifice.
The chief results of this investigation may be summarized
briefly as follows:
In the pig, the method of formation of septum I and ostia I
and II is essentially the same as described in other mammals.
The present account of septum II, however, differs consid-
erably from those of other writers with the exception of Favaro.
The anlage of this septum in the pig appears in the lower anterior
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO 365
corner of the right atrium. It is a spur-like thickening devel-
oped on the right side of the line of insertion of septum I and is
composed, at first, of connective tissue. Later it lengthens out
and becomes overlaid and invaded by developing muscle. It
then extends upward and backward along the roof of the atrium
as a low, broad thickening of the musculature in this region and
finally reaches the posterior wall where it bends downward as a
narrow, sharp ridge. For the reasons stated in the text, I am
inclined to believe that septum II is formed in the same manner
in the human embryo.
The right sinus valve is well-developed in the pig. Its lower
part guards the orifices of the inferior vena cava and coronary
sinus in later stages but does not become divided by the sinus
septum into Eustachian and Thebesian valves as in other forms.
The crista terminalis develops independently of the tensor
valvulae (septum spurium) and right sinus valve. It does not
accurately mark off the primitive sinus cavity from the atrium
proper.
A tubereulum or torus Loweri is always present in older foetal
stages.
One further point which is illustrated by the models may be
mentioned although it is not directly related to the subject of
this paper, namely, the fenestration of the proximal part of the
anterior cardinal vein. In the 7.9 mm. embryo (fig. 5, V.car.a.)
this process is well advanced and many of the fenestrae have
coalesced, thus tending to separate off a dorsal portion of the
cardinal which would receive the intersegmental veins in this
region. In the 15.2 mm. embryo (fig. 6, V.car.a.) the separation
is complete up to the point where the subclavian vein (V.scl.)
enters. On the left side (not shown in the figure) where the sub-
clavian appears to enter one segment higher (cephalad) than
on the right, the separation is carried upward correspondingly.
These observations support the opinion of Thyng (14) that the
proximal part of the vertebral vein is segregated from the main
venous channel (anterior cardinal) in the manner described.
In an earler paper Thyng (711) figured this condition in a pig
366 Cc. V. MORRILL
embryo of 7.8 mm. Lewis (’03) mentions a splitting of the
anterior cardinal vein near its entrance into the duct of Cuvier,
the subclavian vein arising from its outer part. A comparison
of his Plate IV and Thyng’s (711) figure 2, with figures 5 and 6
of the present paper seems to indicate that this process is normal
for the pig embryo.
In conclusion I wish to express my thanks to Prof. H. D.
Senior who kindly suggested a re-investigation of the atrial
septum and to Prof. F. W. Thyng for advice and criticism.
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO 367
LITERATURE
Born, G. 1888 Uber die Bildung der Klappen, Ostien und Scheidewiinde in
Siugerthierherzens. Anat. Anz., Bd. 3,
1889 Beitrige zur Entwickelungsgeschichte des Saéugerthierherzens.
Arch. f. mikr. Anat., Bd. 33.
Favaro, G. 1913 Ricerche embriolog. ed. anatom. intorno al cuore dei verte-
brati, Parte I. Padova.
His, W. 1885 Anatomie menschlicher Embryonen. III. Zur Geschichte der
Organe. Leipzig.
HocustettTer, F. 1901-1903 Die Entwickelung des Blutgefisssystems. Hert-
wig’s Handbuch der vergleich. und experim. Entwickelungslehre der
Wirbeltiere. Bd. 3, 1906.
Lewis, F. T. 1903 The gross anatomy of a 12-mm. pig. Am. Jour. Anat.,
vol. 3.
MorriLut, C. V. 1915 A preliminary note on the septum secundum in the pig.
Anat. Rec., vol. 9, p. 111.
Rerzer, R. 1908 Some results of recent investigations on the mammalian
heart. Anat. Rec., vol. 2, p. 149.
Rose, C. 1888 Beitrige zur Entwicklungsgeschichte des Herzens. Dissert.
inaug. Heidelberg.
1889 Zur Entwicklungsgeschichte des Siéugerthierherzens Morph.
Jahrb., Bd. 15.
1890 Zur Vergleichenden Anatomie des Herzens der Wirbelthiere.
Morph. Jahrb., Bd. 16.
TANDLER, J. 1912 The development of the heart. Keibel and Mall’s Manual
of Human Embryology.
1913 Anatomie des Herzens. Bardeleben’s Handbuch der Anatomie
des Menschen Bd. 3, Abt. 1.
Tuyne, F. W. 1911 The antomy of a 7.8 mm. pigembryo. Anat. Rec., vol. 5.
1914 The anatomy of a 17.83 mm. human embryo. Am. Jour. Anat.,
vol. 17.
PLATE I
EXPLANATION
5 Reconstruction of the heart of a 7.9 mm. embryo.
OF FIGURES
The right atrium and
right ventricle have been opened from the right side, the cut passing through
the junction of the transverse part of the sinus venosus (Sn.t.) and the right
sinus horn (Sn.d.).
Magnified 25 diameters.
6 Reconstruction of the heart of a 15.2 mm. embryo, thé right atrium and
right ventricle opened from the right side.
Magnified 25 diameters.
ABBREVIATIONS
A., Aorta dorsalis
Ao., Aorta (ventral part)
A.d., Atrium dextrum
A.s., Atrium sinistrum
Bulb., Bulbus cordis
Cr.ter., Crista terminalis
En.s., upper endocardial cushion
Enai., lower endocardial cushion
En.c., fused endocardial cushions
L.V., Limbus fossae ovalis (Vieussenii)
OT, Ostium primum
OTT, Ostium secundum
O.in., Ostium interventriculare
Per., Pericardium
S.I, Septum primum
SIT, Septum secundum
S.in., Septum interventriculare
Sn.t., Sinus venosus, trans. part
Sn.d., Sinus venosus, right horn
S.sn., sinus septum
Sn.cor., Sinus coronarius
Sp.in., Spatium intersepto-valvulare
Sul.in., Sulcus interventricularis
T.vv., Tensor valvulae = Septum spu-
rium
Vov.v., Valvulae venosae
V.v.d., right valve
V.v.s., left valve
V.f.o., Valvula foram. ovalis
V.car.c.d., right com. card. vein
V.car.c.s., left com. card. vein
V.car.a.d., right ant. eard. vein
V.car.p.d., right post, eard. vein
V.c.s., Vena cava superior
V.c.z., Vena cava inferior
V.az., Vena azygos
V.scl., Vena subclavia
V.h., Vena hepatica
X, (fig. 9)= part of right sinus valve
guarding coronary orifice
368
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO PLATE 1
Cc. V. MORRILL
H. Murayama de].
369
PLATE 2
EXPLANATION OF FIGURES
7 Reconstruction of the heart of a 21.0 mm. embryo. The right atrium and
right ventricle have been opened from the right side, the cut passing through a
part of the coronary sinus (Sn.cor.). Magnified about 25 diameters.
8 Same as figure 7 but in slightly different position. The right sinus valve
(V.v.d.) has been partly cut away, exposing the orifices of the coronary sinus
and inferior vena cava. Probes have been inserted in the superior and inferior
cavae and the coronary sinus. The probe in the coronary sinus passes across
the sinus septum (S.sn.) under the right valve. Magnified about 25 diameters.
370
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO PLAT, 2
Cc. V. MORRILL
Sul.in
/ \ \
V.v.d NAA LSI) p\M Spal olol a
H. Murayama del.
371
PLATE 3
EXPLANATION OF FIGURES
9 Dissection of the heart of a foetus of about 85 mm., the right atrium and
right ventricle opened from the right side. The lateral wall of the inferior
vena cava (V.c.7.) and the right sinus valve (V.v.d.) have been divided by a
vertical cut and the parts somewhat separated. Magnified about 10 diameters.
PLATE 3
DEVELOPMENT OF ATRIAL SEPTUM-PIG EMBRYO
c. V. MORRILL
t
H. Murayama del.
373
ORIGIN OF THE SEX-CORDS AND DEFINITIVE
SPERMATOGONIA IN THE MALE CHICK
CHARLES H. SWIFT
From the Hull Laboratory of Anatomy, University of Chicago
SIX FIGURES
CONTENTS
baie hen gavel HENAN Y OM INIA egos ba boo Gos n0on oe Sab voascadsdouodec 375
IME Neen suave! camernlaveyel Gui UWChiA, Gobi oe owe oboe ea oe cc kscccacsadbeudocboadode 383
SRE RINE EET e Mitiee OLA ee sOr cee ee ere ed oko 4k REDS cles fe yee tOOD
Differentiation of sex and origin-of the sexual cords.......:...............- 387
The evolution of the sexual or seminiferous cords of the embryonic testis... 390
The interstitial cells and the mitochondrial crescent of the spermatogonium. 403
S\UIGTT COPA oy ay ape Seen a en ES, SN een 407
Syl} EICOyC ay 0) int Anns aneian ee he aR conn erate CL. 2. di am eae 409
INTRODUCTION AND REVIEW OF LITERATURE
The year 1870, in which appeared Waldeyer’s famous article,
Eierstock und Ei, really marks the beginning of modern research
relative to the sex glands, their origin and products.
According to a majority of the investigators who have studied
the question since 1870, there are two possible sources of origin
of the definitive male cells—either from cells of the germinal
epithelium or from primordial germ-cells.
These terms require some elucidation and will be explained
before taking up the more important articles which bear on the
subject.
In the early stages of all vertebrate embryos, the epithelium,
or mesothelium, lining the coelomic cavity is flat and a single
layer of cells in thickness. After the appearance of the Wolffian
body the coelomic epithelium on its ventro-medial surface
begins to change. The cells begin to elongate and become co-
lumnar or cylindrical and they may finally be several layers in
thickness. This change is strictly limited; viewed in section
370
376 CHARLES H. SWIFT
the elongated cells protrude out from the general level of coelomic
epithelium like a small narrow hillock and if looked at from
above, the area has the appearance of a light narrow streak or
stria on the medial and anterior face of the pink mesonephros.
This region of differentiated coelomic epithelium is the germinal
streak of Kélliker (’61) and the germinal epithelium of Wal-
deyer (70).
The primordial germ-cells (ureier, keimzellen, gonocytes,
ovules males, ovules primordiaux, urgeschlechtszellen, and germ-
cells of the various authors) are found among the cells of the
germinal epithelium from its very commencement. They are
easily recognized, for they are quite different from the surround-
ing cylindrical cells. They are large, oval or round, have a
large excentric nucleus, which appears vesicular and does not
have much chromatin. The cytoplasm also stains faintly.
These cells were discerned by Bornhaupt (’67) and were clearly
described by Waldeyer (70), both investigators noticing them
in the germinal epithelium of the chick. Since Bornhaupt many
investigators in every vertebrate phylum have seen the cells,
and yet there is great uncertainty, even at the present time,
in regard to their origin and fate.
The earlier students, as a rule, believed that they arose in
situ, in the germinal epithelium from the epithelial cells by a
process of differentiation (Waldeyer, ’70; Balfour, ’78; and
Semon, ’87). Of late, however, there is an increasing belief
that they antedate the germinal epithelium; that they origi-
nate in some region of the embryo at a distance from the site
of the future gonads, and only migrate into the Wolffian region
at about the time of the appearance of the germinal epithelium
(Niissbaum, ’80; Hoffman, ’92; Eigenmann, 792; Beard, ’04;
Rubaschkin, ’07; and Swift, 714).
Having given these short explanations of terms which will be
frequently employed we are now in a position to review the
more important papers which relate to the origin of the definite
sex-cells in the male.
Waldeyer (’70), although he described so exactly the germinal
epithelium and primordial germ-cells, did not believe that either
SEX-CORDS AND SPERMATOGONIA IN CHICK an
played any role in the histogenesis of the male sex gland. He
believed that the sex cords arose in the middle portion of the
Wolffian body and that they grew into the stroma under the
germinal epithelium. The primordial germ-cells degenerated
and the definitive sex elements arose from the Wolffian tissue.
Braun (’77), like Waldeyer (’70) and Bornhaupt (’67), de-
scribed two kinds of cells in the germinal epithelium, the cylin-
drical, differentiated, peritoneal or coelomic cells and the large
primordial germ-cells. Braun foundin reptiles (Lacerta) that
the walls of the Wolffian tubules produced cell buds, which grew
in height, anastomosed one with the other, and ended by send-
ing up cord-like, cellular processes which fused with the germinal
epithelium. He called these processes segmental cords, and
believed that they gave rise to the seminiferous tubules of the
testis.
In reality, Braun did not describe the origin of the true sexual
cords, but rete cords, or cords of uro-genital union. Up to the
time of Mihalkowics (’85) the true sex cords were constantly
confused, both as to origin and function with the rete cords.
The rete cords, or cords of uro-genital union, arise prior to the
sex cords in the male and female. The rete cords, according to
Mihalkowies (’85) and Sainmont (’05), are derived from the Wolff-
ian tubules and the epithelium of Bowman’s capsule. Accord-
ing to Allen (704, ’05 and ’06), they appear in reptiles and mam-
mals as ingrowths of the germinal epithelium. Bouin (’00), in
the frog, Firket (14), and Swift (15), in the chick, believed them
to arise as condensations of the mesenchyme between the ger-
minal epithelium and the Wolffian body. Whatever their ori-
gin they do not produce the definitive sex-cells, but are devel-
oped into the tubuli recti and rete testis; interposed between the
tubuli contorti, which are of true sex cord origin, which pro-
duce the definitive sex-cells, and the vasa efferentia, which are
modified Wolffian tubules.
Mihalkowies (’85) described the origin of the true sex cords
from the germinal epithelium and replaced Braun’s name seg-
mental cord with sex cord. He did not get the details exactly
correct—that was left for Janosik—but his idea was correct.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
378 CHARLES H. SWIFT
He believed that cells of the germinal epithelium migrated into
the subjacent mesenchyme and grouped themselves into cords
which later united with the overlying epithelium, while Janosik
described the sex cords arising as invaginations or processes
of the germinal epithelium. For a very complete’ account of the
origin of rete cords and the two series of sex cords in the female
ana the single series in the male, the reader is referred to Allen
(04) and Firket (14).
Niissbaum (’80) in Rana and in Teleosts, agreed with most of
investigators of the period prior to Mihalkowies, that the seg-
mental or medullary cords arose as outgrowths of the Malphigian
capsules. However, he disagreed with them as to the origin
of the large clear, primordial germ-cells which were present in
the cords and germinal epithelium. He did not believe them
to be differentiated coelmic cells because he found them away
from the site of the gonads, in the mesentery, because there
were no transition forms and because they contained a great
deal of vitellus, which was not present in the neighboring cells
of the germinal epithelium. He believed them to be primitive
cells and the direct ancestors of the adult sexual cells.
In passing it should be said that Niissbaum was the first to
hint at a distant site of origin and migration for the primordial
germ-cells. He also formulated the hypothesis that ‘the sex-
ual cells do not come from any cells that have given up their |
embryonic character or have gone into building any part of the
body, nor do sexual cells ever go into body formation.” Here,
then, is a suggestion of the germ-plasm idea.
Laulanié (’86), in the chick, thought that the primordial sex-
ual gland was bi-sexual; that the germinal epithelium produced
the female elements and that the male elements or cells were
present in the network of cords, which were developed from the
stroma in the midst of the gonad. The cortical ovules, in the
germinal epithelium were female and the medullary ovules were
male. According to his idea, the gonad was not indifferent but
hermaphroditic. Later on in development one of the sexual
elements degenerated and the gland became male or female as
SEX-CORDS AND SPERMATOGONIA IN CHICK 379
the case might be. This idea is of interest only because of its
novelty and historic interest.
We now come to two important papers, which will always
stand out as landmarks, and which, together with those of Wal-
deyer, Niissbaum and Mihalkowics, will always be studied in
connection with any work on the origin of the sexual cells. I
refer to the work of Semon (’87) and Hoffman (’92).
Semon studied the indifferent gonad and the testis of the
chick. Like Balfour (78), Braun (’77), and Niissbaum (’80), he
believed the sex cords to be outgrowths of the capsules of the
Wolffian bodies. He saw, as Waldeyer did, two kinds of cells
in the germinal epithelium, the columnar epithelial coelomic
cells, and the large clear primordial germ-cells. He described
the various stages which the columnar cells passed through in
becoming the primordial germ-cells. In fact, his opinion’in re-
gard to the primordial germ-cells is exactly that of Waldeyer.
He found that the sexual cords, growing out from the Malphigian
bodies, reached the germinal epithelium about the sixth day and
that the primordial germ-cells passedinto them. The cords
became separated from the epithelium by a connective tissue
layer, differentiated from the stroma, called the albuginea.
The cords anastomosed freely with one another, became tubu-
lar and in this way developed the tubuli contorti seminiferi.
The cavity in the cords began to appear during the third week
of incubation and at hatching had reached a great size. These
tubes contained two kinds of cells, like the germinal epithelium,
the small peritoneal cells and the large clear cells which resem-
bled and were descendants of the primordial germ-cells. In re-
gard to the products of these two varieties of cells itis better
to quote Semon: “Es kann natiirlich keinen Zweifel unter-
liegen, das die kleinen Zellen der Segmentalstringe die soge-
nannten Stiitzzellen der Samenkanilchen, die Ureier aber die
grossen, runden Hodenzellen reprisentiernen.”’
Semon and Niissbaum (’80) were the first to show a conti-
nuity between the primordial germ-cells and the male sexual cells
of the adult. It must be remembered, however, that Semon
obtained his ureier from the cells of the germinal epithelium by
380 CHARLES H. SWIFT
a process of differentiation, while Niissbaum noticed the pri-
mordial germ-cells at a distance from the sexual primordium.
Hoffman (’92) employed in his researches about a dozen species
of birds, mainly taken from the group of waders, and in three of
them, Haematopus ostralegus, Sterna paradisea, and Gallinula
chloropus, there was sufficient evidence brought out to show that
some, if not all the primordial germ-cells, did not arise in the
modified coelomic epithelium. In the three species mentioned
above, he found at the proper time, numbers of the primordial
ova in the germinal epithelium, but, in addition, he found cells,
supposedly primordial ova, because of their resemblance to those
found later in the germinal epithelium in embryosof 23 somites.
An embryo of 23 somites does not possess the so-called germinal
epithelium, the coelomic epithelium over the Wolffian body not
having been modified at this age, yet in these he found primor-
dial germ-cells far removed from the site of the future sex gland,
in the splanchnic plate of mesoderm, in the region between
splanchnic mesoderm and entoderm and in the entoderm itself.
To quote Hoffman:
Maintenant qu’il est évident que les ovules primitifs ne dérivent pas
des cellules péritonéales prévilégiées, mais qu’ils se rencontrent déja
dans trés jeunes périodes de développement, bien que leur premiére
ébauche chez les Vertébrés, soit encore entiérement inconnue il sera
nécessaire de laisser tomber le mot ‘“épithélium germinatif.” Voila
pourquoi j’appellerai desormais, cette partie de l’épithélium périto-
néal, qui se transforme en un assise, dont les cellules sont disposées en
pousieurs rangées et entre lesquelles sont placées les ovules primitifs,
couche des ovules primitifs.
This ‘‘couche des ovules primitifs’”’ gave origin to the true
sexual cords in the male, and the medullary cords and cortical
cords in the female.
The true sexual cords in the male and the medullary cords in
the female are homologous—the cords of first proliferation ac-
cording to Firket (14). The cortical cords—cords of second
proliferation—occur only in the female and appear later. The
true sexual cords in the male produce the sexual elements of the
male. The medullary cords in the female—their equivalents—
go partly into the formation of ovarian stroma, but in greater
part disappear, while the second crop in the female—the cor-
SEX-CORDS AND SPERMATOGONIA IN CHICK 381
tical cords—give rise to the sexual elements in the female and
the follicular epithelium (Swift, 715).
Hoffman found in Totanus calidrus, Vanellus cristatus, and
Limosa ergocephala that the true sexual cords, just after hatch-
ing, developed a lumen and became the seminiferous tubule.
To quote Hoffman again:
Les tubes seminiféres renferment deux espéces de cellules; des cel-
lules grandes et rondes et des cellules plus petites d’une forme conique
ou cylindrique qui les premiére enveloppent. Les grandes cellules
ont un noyau arrondi et volumineux, qui ne contient que peu de chro-
matine et qui ne se colore que trés faiblement par les reactifs; son
diamétre est de 9-ll yu. Le corps protoplasmatique est également
trés pale, les contours sont ordinairement trés indistincts. Le noyau
des petites cellules est ovale, il a une longueur de 6-7 yu, une largeur
de 4-5 uy et est tienté trés fortement. Les deux espéces de cellules
se trouvent mélées sans aucun ordre, ne formant ordinairement qu’une
seule assise, comme cela se voit dans les sections, qui ont coupées le
tube ou exactement dans son axe longitudinale ou transversale. Entre
les grandes cellules des tubes seminiferés et les ovules primordiaux de
la couche des ovules primitifs, il n’y a pas de moindre différence. I
n’est pas douteux que les grandes cellules soient des ovules primor-
diaux émigrés et qu’elles représentent les propres cellules du testi-
cule—les spermatogonies—tandis que les petites cellules forment les
cellules de soutient (Stiitzzellen des auteurs allemands).
Hoffman’s work is of extreme importance because of the con-
tinuity which he established. The primordial germ-cell was
first seen away from the germinal epithelium; it migrated into
that structure, then passed into the sexual cord, which the ger-
minal epithelium produced, became a spermatogonium, and
later developed into the mature sexual cell.
According to Semon the peritoneal mesoblastic cell is the
ancestor of the sperm; according to Hoffman the primordial
germ-cell of whose ancestry he is ignorant, is the direct forebear
of the sperm.
Allen (’04) made a very thorough study of the development of
the testis and ovary of the pig and rabbit and found the sex-
and rete-cords to be invaginations ofthe peritoneum.
In regard to the primordial germ-cells he said:
Primitive sex cells are being formed in the rete cords from the synci-
tial cells that have retained the primitive character exhibited by the
382 CHARLES H. SWIFT
peritoneal cells, from which these cords arise. This development of
undifferentiated peritoneal derivatives to form primitive sex cells is
probably homologous with the process by which germinative cells
(descendants of the peritoneal cells of the germinal epithelium) of the
seminiferous tubules of the testis and the cells of the cords of Pfliiger of
the ovary are being transformed into primitive sex-cells.
And again:
Clear transition forms are found to connect the primitive sex-cells
with the germinative cells of the seminiferous tubules. Transition
forms connecting the germinative cells with the sex cells occur in the
pig embryos between 2.5 em. and 13 cm. length; later SHIGE show no
transition forms.
As regards the formation of the definitive sexual cells he said:
It has been shown in the pig and rabbit, however, that these sex
cells, appearing in the indifferent stages, do not contribute to the for-
mation of functional sex products in the ovary. The same is probably
true of the testis, it being at least certain that the great bulk of the
spermatogonia are formed from the germinative cells—cells derived
originally from the peritoneum and maintaining, first, their indifferent
character, at least so far as our technique is able to show.
Dustin (’07, 710) studied the question of germ-cells origin,
migration and evolution in Amphibia—Triton and Rana—and
in Reptilia—Chrysemys.
In the first paper, dealing with the Amphibia, he reported
that the primordial germ-cells were of mesoblastic origin; from a
part of the primitive coelom which he called gonocle. He called
the germ cells gonocytes of the first line and found that they
reached the sites of the definitive sex glands by active migra-
tion. He found that a second line of gonocytes was produced
by transformation of small germinative cells, which were in
their turn the result of peritoneal epithelium proliferation.
Both lines of gonocytes were able to produce definitive sexual
cells.
In his second paper relating to Chrysemys, he found that
there were also two lines of gonocytes, with this exception, that
in the reptile the primary gonocyte came from the entoderm.
Deux espéces de cellules sexuelles se succédent done au cours de
V’ontogenése; l’une privient de l’éndoderme; elle est extraordinaire-
SEX-CORDS AND SPERMATOGONIA IN CHICK 383
ment précoce et subit une série de migrations qui lui assurent trois
localisations successives (gl. paires primaries, glande impaire, gl.
paires définitives); l’autre est d’origine mésodermique et se développe
in situ, au niveau des glandes définitives.
Chez Chrysemys marginata, les gonocytes de la premiére espéce ne
présentent de karyokinesés qu’au début de leur différenciation; ul-
térieurement ils ne se multiplient plus; beaucoup d’entre eux dégé-
nérent méme.
He believed that more of the primary gonocytes functioned, in
producing definitive sexual cells, in Amphibia than in Reptilia.
Rubaschkin (712), working with the guinea-pig, stated that
the definitive sex cells in both testis and ovary are descendants
of the primordial germ-cells. Rubaschkin used as his criterion
for distinguishing the primordial germ-cells the granular type
of mitochondria. He believed that the primitive cells, or blas-
tomeres, possessed this type of mitochondria, and that as the
somatic tissues were differentiated from them, the cells of the
tissues acquired the rod-shaped mitochondria. In this way he
was able to recognize the germ-cell at any stage in the young
embryo and differentiate it from the ordinary somatic cell. By
this means he came to the conclusion that the definitive sex
cells were descended from the primordial germ-cells.
For a more detailed account of the literature on the subject
of the origin of the definitive sexual cells, the reader is referred
to. the ample bibliographies of Bouin (’00), Allen (’04), and
Firket (14).
MATERIALS AND METHOD OF STUDY
In the course of this investigation two fixatives were employed-
namely, Meves’ modification of Flemming’s fluid and the acetic,
osmic-bichromate mixture. In a previous investigation on the
primordial germ-cells of the chick (14) the osmic acid content -°
of these fluids was found to be a serious disadvantage owing
to their staining of the numerous vitellus granules, which, when
blackened, obscured the cytological details. However, that
drawback was not present in the embryos and young chicks
used in this research for the germ-cells had lost their vitellus
384 CHARLES H. SWIFT
content, and the preservation of the mitochondrial granules and
attractions sphere was a decided asset.
In this work only the testes were fixed and not the whole
embryo, as was the custom, when dealing with the younger
stages.
In the case of the youngest stages used in this investigation,
the testes were allowed to remain in the killing fluid 24 hours
but as there were signs of over-fixation in the sections, this time
was reduced to 10 hours with much better results.
All the sections were cut 4u in thickness, and following the
acetic-osmic-bichromate fluid stained by Bensley’s ‘anilin-acid
fuchsin—Wright’s blood stain method and the anilin-acid-fuch-
sin-methyl green method. There is no need of entering into
detail in regard to these staining methods since they have been
described by Bensley (711), Cowdry (12), and Swift (715).
Following Meves’ fluid the iron hematoxylin stain was
employed.
The following table will show at a glance the number of stages
employed, their age, and the methods used in fixation and
staining.
TABLE 1
a
4 Q
=) % a
oO ° =
fe Es
METHOD On| 7] 8 | 9 | 11 | 43 | 15 | 17) 20 |<) ya
np
aE >
a D
: g
A 5
Meves’ fixation and iron-hematoxylin
SENN one eRe oS cits oo lsc 1g ie
| Bensley’s anilin acid
Ne atic fuchsin and methyl-
Pee eer ae green stain..... se Teal:
ae ea Bensley’s anilin
ERTS acid fuchsin and
: Wright’s blood
Stainaiit seekers 2)2)2/2)|3)2);2/2)2)2)2;)1
SEX-CORDS AND SPERMATOGONIA IN CHICK 385
THE INDIFFERENT GONAD
If the viscera are removed from the abdominal cavity of a
5-day chick embryo, the Wolffian bodies will be seen on the
posterior wall on either side of the mid-line. On the ventro-
medial surface of each mesonephros, beginning at the cephalad
extremity and extending caudad about two-thirds of the length
of each Wolffian body, a narrow, white, rounded ridge will be
observed. These narrow elongated elevations are the indiffer-
ent gonads. At this time they have a length of about 2 mm.
and there is a noficeable difference in size in favor of the left
gonad.
On examining a section through the anterior portion of the
Wolffian body, the gonads will be seen to resemble two small
abrupt hillocks on the surface of the mesonephros turned
towards the mesentery, and, it will be noticed also that they
have a broad connection with the mesonephros.
The microscope reveals, farther, that there are three distinct
tissues in the gonads.
Clothing the free surface of each genital anlagen, and extend-
ing for some distance over the root of the mesentery, is an epithe-
lium made up of two to three layers of tall columnar cells. This
tissue is the germinal epithelium of Bornhaupt (’67) and Wal-
deyer (’70) and all later investigators. This epithelium is made
up of columnar or rather cylindrical cells, which have oval or
round deep staining nuclei and possesses a distinct basement
membrane. This last structure is very noticeable and at this
stage—5 days—can be followed over the germinal hillock as a
regular curved line without any irregularities of any kind.
Under the germinal epithelium, which is thicker and more ex-
_ tensive over the left gonad, is another tissue—the stroma of the
gonad. This tissue is nothing but embryonic mesenchyme, and
as such is directly continuous with the same tissue in the Wolff-
ian body and in the root of the developing mesentery. ‘The
cells of this tissue are loosely packed and have faintly staining
nuclei. The cytoplasmic part of the cells appears as indistinctly
bounded processes which anastomose with the processes of other
386 CHARLES H. SWIFT
neighboring cells. The cells, then, are apt to be stellate in
appearance and the whole tissue to resemble a loose syncytium.
This stroma, more voluminous in the case of the left gonad,
fills up the region between the concave germinal epithelium and
the Wolffian body and, in fact, forms most of the area of the
gonad.
This stroma is denser Just under the epithelium, and from
this region, if the sections happens to be right, a narrow cord
may be seen to extend obliquely towards the Wolffian body.
This cord, the result of a condensation of mesenchyme tissue,
is known as a rete cord or cord of uro-genital union and must
not be confused with the sexual cord, which arises later and in a
different way. According to Firket (14) there are 16 of these
cords. It is not my intention to enter into the origin of these
cords or to engage in the controversy as to their origin, for an
excellent review of all the facts and literature will be found in
Firket’s (14): article.
The third tissue present in the gonad is made up of primordial
germ-cells.
The cells are easily seen, for their great size, large nucleus,
and clear cytoplasm make them conspicuous in the chick as in
all the other forms in which they have been observed. They
merit, however, a more extended description.
The primordial germ-cells do not form a compact or continu-
ous tissue in the indifferent gonad of 5 days, but are present,
usually, as isolated cells or groups of several cells in the germinal
epithelium and subjacent stroma (Swift, 715). They may be
even seen in the root of the mesentery subject to the same
arrangement. :
These primordial germ-cells are, at a glance, seen to be dif-
ferent from the other tissues of the gonad. Although present
in the epithelium and stroma of the genital hillock yet I have
never seen a stage which could be called a transition between
them and the surrounding cells, for in addition to their large
size, great nucleus and clear cytoplasm, they possess even at this
late date (5 days) many droplets of vitellus in the cytoplasm
and a conspicuous attraction-sphere (Swift 714 and ’15). The
SEX-CORDS AND SPERMATOGONIA IN CHICK 387
cells of the germinal epithelium and stroma have none of the
vitellus and inconspicuous spheres.
In fact the primordial germ-cells do not originate in the ger-
minal epithelium as was believed by many of the older investiga-
tors, but at a distance from the site of the gonad. They arise
in the germ wall entoderm (Swift, 714) and after an extended
migration through the blood vessels (Swift ’14 and von Beren-
berg-Gossler, 714) they reachthe splanchnopleure entering into
the formation of the mesentery, where they leave the vessels
and then migrate in an amoeboid manner into the forming
germinal epithelium.
During all their migration period and up to this time they
divide very infrequently. That they do divide is proven by the
increase of numbers from stage to stage (Firket, 714, and Swift,
714 and 715) and by the fact that they are frequently seen in
small groups, which are probably the resultof several successive
divisions. .
As was indicated in the short historical review of this article,
it is still an open question as to the réle played by the primor-
dial germ-cells in the formation of the definitive sex-cells in the.
male.
DIFFERENTIATION OF SEX AND ORIGIN OF THE SEXUAL CORDS
The period of embryonic development in the chick from 53
to 64 days is of extreme interest and importance, for during that
period the sexual cords appear and at its termination it is pos-
sible to tell the sex of the individual.
The sexual cords, which begin to appear at about the 132d
hour of development, are the true sexual cords in the male or
seminiferous cords, and in the female the medullary cords or
cords of first proliferation. As the names imply, there is only
one series of cords in the male while in the female there are two
—the last series being known as the cortical cords or cords of
second proliferation (Mihalkowies, ’85, Hoffmann, ’92, Firket,
14 and Swift, ’15).
The cords of first proliferation are first noticed as buds or
protrusions of the germinal epithelium into the underlying
388 CHARLES H. SWIFT
stroma. The basement membrane appears wavy but is continu-
ous around the swelling. The buds enlarge and elongate but
remain attached to the germinal epithelium for some time (fig.
5, Swift, 715). These cords, which are at first small, are evi-
dently of germinal epithelium origin, for the basement mem-
brane of the latter structure is, for some time, continuous around
the cord. Neither are they formed by infolding or invagination
of the germinal epithelium, for in that case a tubular lumen
would be present in each cord, and in some cases could be
seen connected with the coelomic cavity. They are simply
the result of increased local mitotic activity in the germinal
epithelium. The primordial germ-cells which are present in the
cords do not play any evident role in their formation, for none
are ever found dividing just at the time of cord formation,
while the peritoneal cells show signs of increased activity.
Although the cords begin to appear at 53 days, yet it is not
until 12 hours later that the greatest activity of the germinal
epithelium in their production occurs. At 63 days they have
about ceased to appear.
At the end of the period of cord production the interior of
the gonad seems to be made up nearly entirely of sexual cords.
They are close together, separated from one another by a little
stroma; a majority of the cords are no longer attached to the
germinal epithelium but are still surrounded by a definite base-
ment membrane.
The cords have become autonomous, for, although no longer
attached to the germinal epithelium, they are growing rapidly as
is indicated by the rapid increase in diameter of the cord and
size of the gonads.
The cords have a definite orientation—they are straight and
extend from the germinal epithelium down towards the Wolffian
body.
The primordial germ-cells are during this period found in
three situations in the gonad, in the germinal epithelium, in the
sexual cord, and in the small amount of stroma between the
cords.
SEX-CORDS AND SPERMATOGONIA IN CHICK 389
When the embryo chick has reached the 156th hour of devel-
opment (63 days), the formation of cords of first proliferation
ceases rather abruptly, and about this time it is possible to
determine the sex of the individual.
In determining the sex of the embryo there are several cri-
teria which are of value.
1. The relative size of the gonads. It has been known for
a long time that the left gonad in the female grows rapidly while
the right does not, so that in a short time after sexual cord for-
mation, the left gonad or ovary has far outgrown the right. In
the case of the male individual both gonads continue to grow.
This criterion is of value, but too much reliance can be placed
upon it especially during the 6th day, for in all cases, male as
well as female, the left gonad is the larger. The left gonad is the
larger during the indifferent stage and in the male is the larger
even in the adult (Etzold, 791).
2. Thickness of the germinal epithelium. This and the fol-
lowing criterion I believe to be the most important. In the
male the germinal epithelium is thin at the end of sexual cord
formation and continues so, while in the female the germinal
epithelium over the left gonad or ovary still consists of the sev-
eral layers of columnar cells. In the male the epithelial cells
become very quickly a single layer in thickness and the indi-
vidual cells cuboid in character. This is an excellent test.
3. Number of primordial germ-cells in the germinal epithe-
lium. In the germinal epithelium of the left female gonad the
number of primordial germ-cells appears undiminished, while in
the epithelium covering the male gonads there is hardly a single
one to be seen (Hoffman, ’92 and Swift, 715). In the case of
the male they all appear to have gone into the sexual cords, and
in the female to be remaining quiescent until the formation of
» the cortical cords, in the evolution of which they play so im-
portant a role (Swift, 715).
4, Attachment and growth of the sexual cords. The cords
remain thin and attached to the germinal epithelium for a long
time in the male gonad, while in the female they increase rap-
idly in diameter and become detached early in their evolution.
390 CHARLES H. SWIFT
This rapid increase in size of the cords accounts for the more
rapid growth of the embryonic ovary.
This last criterion is of less weight than the preceding, and
is of importance only when used in connection with the others.
THE EVOLUTION OF THE SEXUAL OR SEMINIFEROUS CORDS OF
THE EMBRYONIC TESTIS
The next three embryonic stages studied, namely, 7, 8, and
9 days, may be described together, since they are characterized
in common by the same facts, and the changes aithough oc-
curring, are not abrupt or great.
During this period the sex of the individual can be easily
ascertained, either with or without the aid of the microscope.
In the latter case the first criterion mentioned above, relative
size of the gonads, is amply sufficient, and with the microscope
a single glance at the germinal epithelium, the sexual cords,
and the number of primordial germ-cells in the epithelium is all
that is needed.
Both testes are slowly increasing in size but in all cases the
left is found to be the larger. They are becoming rounder and
in the process the broad connection with the Wolffian body is
being narrowed; in other words, they are being constricted or
pinched off of the Wolffian body.
The germinal epithelium is reduced to a single layer of cu-
boidal cells, and by the end of this period of 7 to 9 days of de-
velopment, all connection with seminiferous cords is severed.
This is brought about by a condensation of the mesenchyme
under the epithelium to form the tunica albuginea.
The seminiferous cords, which will in the future claim most
of our attention, make up the greater part of the testis. They
have a definite orientation from the epithelium obliquely toward
the narrowing attachment of the testis with the Wolffian body.”
Each cord is, of course, not entirely straight, but wavy. They .-
are separated one from another by a very thin layer, one might
almost say film, of stroma and a number can be followed at one
time across the field of the microscope.
SEX-CORDS AND SPERMATOGONIA IN CHICK 391
The cells of the seminiferous cords are of two kinds—the
ordinary peritoneal cells, issue of the germinal epithelium, and
the primordial germ-cells. The former are the more numerous
and, at this period, can be easily seen to have definite mem-
branes. They still preserve their cylindrical form, acquired in
the formation of the germinal epithelium, and the dark staining
round or oval nucleus.
The primordial germ-cells, although not numerous, are as
conspicuous as when observed in the capillary of a 21 somite
chick or the root of the forming mesentery (Swift, 14). They
are a little smaller and have lost the characteristic vitellus but
in all other ways are unchanged. They are still large, have the
same clear cytoplasm, large round nucleus and conspicuous
attraction sphere. From 2 to 8 of the primordial germ-cells are
present in a single cord and there are no signs of division. In 9
embryos of this age studied I was not able to find a single pri-
mordial germ-cell in mitosis. This last fact is of interest when
a comparison is made with the primordial germ-cells inthe ger-
minal epithelium of the ovary of the same age. It will be re-
called (Swift, 15) that during the 8th and 9th day in the fe-
male the germ-cells are in an extremely active condition and that
the formation of the cortical cords is well under way.
As regards the stroma only a few lines are necessary. Only
under the germinal epithelium is the connective tissue develop-
ment considerable. In that region there is a broad band of
tissue, the albuginea. In the rest of the gonad, between the
seminiferous cords, the small amount of stroma still remains
embryonic in nature, forming a kind of mesenchymal syncy-
tium. It resembles the mesenchyme of the mesonephros from
which it was originally derived and with which it is still con-
nected.
Four embryos aged 11 days were studied. Considerable
change has taken place. The testes have increased greatly in
size, are rounder, but still have a narrow attachment to the
Wolffian body. On examining a section through the testis it
will be noticed immediately that the stroma has increased in
amount between the seminiferous cords (fig. 1), and that the
392 CHARLES H. SWIFT
layer under the epithelium has thickened. In fact most of the
increase in the size of the testes is due to the increase in amount
of connective tissue. The cells of this tissue are still mesen-
chymal in character; cell boundaries are hard to determine and
the cytoplasmic processes of the cells unite forming a kind of
syneytium. The nuclei are round or oval (fig. 1).
The seminiferous cords no longer run in a regular way as
described above, but form a kind of network. On studying
them one receives the impression that they have grown consid-
erably in length and as a result have had to become convo-
luted in order to adapt themselves to a slower growing space.
In all four embryos studied several peculiarities were found to
be common in regard to the seminiferous cords. In the first
place a cord runs only a short distance across the field before
being lost. This indicates folding (fig. 1). In the second
place there are more cords present in that part of the testis
next the Wolffian body, and here the net formed by the cords
is more compact. In the region towards the germinal epithe-
lium there is always a single cord which tends to run parallel
to the germinal epithelium separated from it by the developing
albuginea and from the other cords by a thick layer of stroma.
The seminiferous cords are surrounded by a definite basement
membrane (fig. 1), which they acquired when given off from the
germinal epithelium.
The cells present in the cords are principally of the perito-
neal type, although an occasional primordial germ-cell can be
seen (fig. 1). The cell walls of the peritoneal cells can be seen
but are not so distinct as formerly. Peritoneal cells are occa-
sionally seen dividing but the primordial germ-cells are still
quiescent. In this stage the increase in the cords is one of
length but not of diameter.
In the 13 day embryo several important changes must be
recorded.
There has been a great increase in the amount of stroma (fig.
2) and as a result of this, in the size of the testis, which is now
almost separated from the mesonephros. In an embryo four
days younger the stroma was present simply as a thin film be-
SEX-CORDS AND SPERMATOGONIA IN CHICK 393
tween the seminiferous cords; two days later there had been some
increase in the connective tissue, especially in the region under
the germinal epithelium, but in these embryos of 13 days de-,
velopment the area of the mesenchyme far exceeds that of the
Fig. 1 Portion of a transverse section through the left testis of an 11 day
chick embryo. This section shows the seminiferous cords, composed of perito-
neal cells and containing several primordial germ-cells. The cords are sepa-
rated from one another by a mesenchyme-like connective tissue.
The figures illustrating this article were drawn by Mr. A. B. Streedain. Zeiss
apochromatic objective 1.5 mm., and compensating ocular 6 were employed for
all the figures. The camera lucida was used in making all the drawings and
magnification calculated at table level in all cases. The figures were reduced by
one-fourth in reproduction giving a magnification of 1125 diameters for all the
illustrations. The figures were drawn from preparations fixed in Bensley’s
acetic-osmi¢ bichromate mixture and stained with Bensley’s anilin acid fuchsin
—Wright’s blood stain. All the sections were cut 4 u in thickness.
ABBREVIATIONS
int.C., interstitial cells S., cell of support
L., lumen of seminiferous cord S.C., seminiferous cord
M.C., mitochondrial crescent Sp., spermatogonium
p.c., peritoneal cell Sp’., spermatogonium in mitosis
pr.o., primordial germ-cell Str., stroma
pr’.o’., primordial germ-cell in mitosis.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
394 CHARLES H. SWIFT
cords (fig. 2). In addition a few interstitial cells are present
in the stroma between the cords. These interstitial cells are
very numerous in later stages, so that an account of them and
their development will be deferred until those are described.
The seminiferous cords have a larger diameter, and now form
an open net with large interstices (fig. 2). The cords anasto-
mose with each other in every conceivable plane and direction,
so that their original orientation is completely lost. This tangle
of cords now occupies the entire testis, the area of mesenchyme
under the germinal epithelium, described in a previous stage,
fs “ud Bar ¢ ‘
Se)
Fig. 2 Portion of a transverse section through the left testis of a 13 day
chick embryo. This section shows a seminiferous cord containing many primor-
dial germ-cells. Two of the germ-cells are dividing.
SEX-CORDS AND SPERMATOGONIA IN CHICK 395
having disappeared. The cords stain more intensely now, stand
out more sharply against the light connective-tissue back-
ground and contain peritoneal cells and primordial germ-cells.
There has been an important change, however, for the latter
are much more numerous than in any preceding age (fig. 2).
This increase in number of the germ-cells is evident even when
examining the section under the low power microscope, for
they are seen as clear spots thickly sprinkled throughout the
cords. On using the higher powers this is confirmed, and, in
addition, it is seen that the primordial germ-cells are actively
dividing (fig. 2). In a single section through the testis from
4-10 are found in mitosis. The increase in the size of the semi-
niferous cord is due to the increase in the number of the germ-
cells for the peritoneal cells are quiescent.
Semon (’87) in the chick and Popoff (09) in the chick and
a number of other forms, described the ‘ureier’ and ‘les ovules
males’ (primordial germ-cells) as having a definite position near
the center of the seminiferous cords, while the peritoneal cells
occupied the periphery.. I have not found that to be the case
(figs. 2 and 3). In all stages up to and including the 15 day
stages examined, I find them to be scattered evenly throughout
the cord—some at its center and some at the periphery. Later
on they are all found at the periphery of the cord.
A few words are necessary in regard to the structure of the
primordial germ-cells. As will be seen from all the figures illus-
trating this article there is little change in the primordial germ-
cells or their tissue. In fact, there is little change to record in
all their history from origin until they become oogonia or sper-
matogonia, unless it is a gradual loss of vitellus and a progres-
sive slight decrease in size (Swift, 714 and 715). However, with
the beginning of division in the female line of germ-cells, which
resulted in the oogonia, there was a change in the arrangement
of the mitochondria (Swift, 715). While the primordial germ-
cells were quiescent in the germinal epithelium of the embryonic
ovary, up to the 8 day stage, the mitochondria were evenly
scattered in the cytoplasm (Swift, 715). At 8 days the germ
cells began to divide, and in all the later stages the mitochon-
396 CHARLES H. SWIFT
dria were found grouped around the attraction-sphere in a
characteristic manner. It is a striking and suggestive fact that
the same change occurs in the male line. In all the primordial
germ-cells in the sexual cords of the male chick, up to and in-
cluding the 11 day stage, the rod shaped and granular mitochon-
dria are evenly distributed (fig. 1). This arrangement also
ofe
*, *e
0" “e$e
Fig. 3 Portion of a transverse section through the right testis of a 15 day
chick embryo. This section shows a massive seminiferous cord, composed of
primordial germ-cells, spermatogonia and peritoneal cells. One of the germ-
cells is dividing. The spermatogonia, at the left, contain the characteristic
mitochondrial crescent which serves to identify them.
holds for the germ-cells of the 13 day embryo and for the few
present in the 15 day stage.
Thus, beginning with the 13 day embryo, a new mitochondrial
arrangement is noticed; the mitochondria are grouped around
the attraction-sphere, so that they form a granular crescent or
cap on the nucleus, depending upon the plane of the section
(figs. 3, 4, 5 and 6). As was stated above and in a former
SEX-CORDS AND SPERMATOGONIA IN CHICK 397
article (Swift, 15) this same arrangement was observed in the
oogonia of the chick and it is my belief that any cell in the male
which has this character is a spermatogonium. Certainly all
the spermatogonia of much later periods possess this distinguish-
ing character. I believe, then, that spermatogonia formation
Fig. 4 Portion of a transverse section through the left testis of a 17 day
chick embryo. This section of a massive seminiferous cord contains peritoneal
cells and spermatogonia. All the spermatogonia contain the mitochondrial
crescent. At this age practically all the primordial germ-cells have given way
to the spermatogonia.
398 CHARLES H. SWIFT
begins in the male chick embryo at 13 days as a result of pri-
mordial germ-cell division, and that the peculiar arrangement
of mitochondria around the attraction-sphere, which will be
taken up at greater length in a later chapter, distinguishes the
spermatogonium from the primordial germ-cell.
At this time, also, it is necessary to say something about the
continuity of the mitochondria in the germ line.
In this regard, I can say that they are always present in the
primordial germ-cells, oogonia and spermatogonia, either as
short rods or granules, depending somewhat upon the fixation.
About the same amount seems to be present in the germ-cells
at all stages, from origin in the primitive streak stage to oogonia
and spermatogonia formation in the 8 and 13 day embryo
respectively. Even in the oogonia and spermatogonia the num-
ber of mitochondria remains about as in the germ-cells, for the
mass around the attraction-sphere is formed at the expense of
the rest of the cytoplasm (figs. 4, 5 and 6).
In the 15 day chick embryo there are an immense number of
interstitial cells in the stroma. They appear grouped in masses
and cords in the stroma between the seminiferous cords. So
numerous are they that it is possible to identify them in the
section before removal of the paraffin. This can be easily done
since they appear black, owing to the staining of their many
fatty granules by the osmic acid.
The seminiferous cords branch and anastomose to form a net
as described under the 13 day embryo. They contain, however,
relatively and absolutely, many more spermatogonia, for the
increase has been in that element owing to continued division
of the germ-cells (fig. 3). The number of spermatogonia, com-
pared with the preceding stage, is enormous. They are seen
several hundred in a field, when the low power microscope is
used. Numbers of mitoses in the primordial germ-cells and
spermatogonia are present (fig. 3), and this is the stage in which
they are found dividing most actively.
In regard to the 17 day chick embryo not a great deal need
be said, for, in it, the findings closely resemble those described
in the 15 day embryo.
SEX-CORDS AND SPERMATOGONIA IN CHICK 399
The interstitial cells are present in greater quantity, and
there has been some increase in the size of the cords; the amount
of cord tissue, too, as compared with the stroma, has increased
(fig. 4). Practically all the primordial germ-cells have been
changed by division, giving rise to the spermatogonia (fig. 4).
There is no evidence of division in the peritoneal cells but there
number has not decreased. In certain regions of the seminif-
erous cord complex, it is evident that the spermatogonia are
beginning to assume a position at the periphery of the cord, next
to the basement membrane. ‘This is very evident in the next
stage. There is no evidence of a Jumen beginning to appear
in the cords. They still are solid, although it was at this age
that Semon (’87), in the chick, described one as being present.
In the next embryo, that of 20 days development, there are
some important changes to record.
The seminiferous cords are very massive and seem to be grow-
ing at the expense of the stroma, which is considerably reduced
in quantity (fig. 5). The interstitial cells (fig. 5), too, are not
as numerous as in the two preceding stages. The greatest
changes, however, have taken place in the structure and orien-
tation of the cords.
The spermatogonia, which hitherto have had no definite ar-
rangement, being scattered helter-skelter throughout the cord,
now begin to be arranged in a definite manner. They are
placed against the basement membrane of the seminiferous
cord in such a way that their long axes are at right angles to the
long axis of the cord (fig. 5). Another striking fact is to be
noted, that the attraction-sphere, with its mitochondrial body,
is next to the basement membrane, while the nucleus is turned
toward the axis of the cord (fig. 5). That is to say, the vege-
tative pole of the spermatogonium is next the basement mem-
brane. The nucleus is excentrically placed, and always the
greatest mass of cytoplasm, in which is placed the attraction-
sphere, is towards the basement membrane (fig. 5).
The cords, also, are beginning to have a cavity. This cavity
is far from being continuous and can be seen here and there
only, in the central axis of the seminiferous cord (fig. 5). This
400 CHARLES H. SWIFT
lumen is formed, not as one would suppose, by a fissure appear-
ing between cells and then enlarging, but by a liquefaction of
the cells in the central axis of the cord. When this destruction
which involves chiefly the peritoneal cells, is complete, a slit
appears in the debris and the lumen is formed (fig. 5). The peri-
Fig. 5 Portion of a transverse section through the right testis of a 20 day
chick embryo. This section shows a seminiferous cord in which a lumen is be-
ginning to develop. The spermatogonia are next to the basement membrane,
with elongating peritoneal cells between them. Notice that the mitochondrial
crescents of the spermatogonia are next to the basement membrane of the cord.
A small mass of interstitial cells is present in the stroma.
toneal cells are the principal sufferers because the spermato-
gonia have placed themselves along the wall. .
_.1 do not know positively through what agency they are
enabled to place themselves against the basement membrane,
-but suppose that it is through their ability to move like an
SEX-CORDS AND SPERMATOGONIA IN CHICK 401
amoeba. They or rather their ancestors, the primordial germ-
cells, had this power in their early history, and it may have been
in abeyance during their sojourn in the germinal epithelium and
sex cords, only to be assumed again at this time. Color is lent
to this theory by the fact that frequently, at this stage, the end
of the spermatogonium turned toward the basement mem-
Fig 6 Portion of a transverse section through the left testis of a chick 3
days old. This section shows a seminiferous tubule containing spermatogonia
and supporting cells. The spermatogonia are next the basement membrane and
have a definite polarity—nucleus near the central portion of the tubule and
mitochondrial crescent at the basement membrane. The supporting cells are
descendants of the peritoneal cells.
brane is pointed, as if it were forcing its way in that direction
between the intervening cells (fig. 5).
The cords, although forming a network, are again beginning
to have a definite orientation, that is, they run obliquely from
the germinal epithelium towards the remains of the Wolffian
body.
402 CHARLES H. SWIFT
In this, and the next stage to be described, the position of
the peritoneal cells is also of great interest. In general there
are one to three of them between adjacent spermatogonia and -
their long axes, too, are at right angles to the long axis of the
cord (figs. 5 and 6). Their nuclei are placed next the basement
membrane, while most of the cytoplasm is towards the central
axis of the cord or tubule (figs. 5 and 6). This is the reverse
of the condition in the spermatogonia.
In this stage and in the next we begin to have an idea of what
obtains in the adults which have been studied. The spermato-
gonia at the basement membrane of the tube and between them
the supporting cells—cells of Setroli—which are derived from
the peritoneal cells.
In the next stage—a 3 day chick—the stroma is greatly re-
duced in amount as are the interstitial cells also. Most of the
testis is taken up by the seminiferous cords and tubules, which
are arranged in the way described in the 20 day chick embryo.
Most of the seminiferous cords have by this time acquired a
cavity and may now be called tubules. In these seminiferous
tubules nearly all the spermatogonia are against the basement
membranes, as described before (fig. 6). The peritoneal cells
have lengthened, or been pressed out, so that they begin to look
like the supporting cells of an adult seminiferous tubule (fig. 6).
In this as well as in the last stage, the spermatogonia are occa-
sionally seen dividing (fig. 6). In these rare cases the mitotic
figure is arranged, not like that which gives origin to the sper-
matocyte, but so that the line of cleavage between the daughter
cells is at right angles to the long axis of the seminiferous tubule
(fig. 6). In this way both daughter spermatogonia are kept
in their proper position, next the membrane, and the number of
peritoneal cells, or cells of support, which is still too large, is
reduced by pressure (fig. 6).
In the 6 and 10 day chicks there are no changes of any mo-
ment to record, except that there is a progressive increase in the
size of the lumen of the seminiferous cords and a reduction in
the amount of stroma and in the number of interstitial cells.
The cavity in the seminiferous cord, even at 10 days is slit-like
SEX-CORDS AND SPERMATOGONIA IN CHICK 403
and is very far from resembling the large cavity pictured by
Semon (’87) in stages of about this age.
THE INTERSTITIAL CELLS AND THE MITOCHONDRIAL CRESCENT
OF THE SPERMATOGONIUM
It is not possible to enter into any account of the various
opinions which have been advanced and held as regards the ori-
gin of the interstitial cells of the testis. For the various theories
on this subject the reader is referred to Sainmont’s article which
appeared in 1905. However, to give a slight historical back-
ground a few leading articles will be cited.
Tourneux (’79) identified the interstitial cells in the testes of
various animals and described them as being differentiated con-
nective tissue cells.
Niissbaum (’80) thought that the interstitial cells were de-
rived from cell columns, which were given off by the germinal
epithelium in early embryonic history.
Plato (97) described the different stages which exist between
the connective tissue cells and the differentiated interstitial
cells. He studied the question in the testes of the cat and vari-
ous other animals. In connection with this work he brought
forward a rather novel theory which ascribed a nourishing func-
tion to the interstitial cells. He believed that they manufac-
tured fat, which was passed through the basement membrane
of the seminiferous tubule into the cells of Sertoli and used as
food in spermatogenesis.
Allen (03) and Whitehead (’04) believed that the connective
tissue elements of both testis and ovary of the pig and rabbit,
from which the interstitial cells were differentiated were de-
rived from the peritoneum. Allen says: ‘“‘In early stages they
(the connective tissue elements) are not distinguishable from the
cells which make up the sex-cords, except that the latter are
marked off from the stroma by their membrana propria.”
Whitehead ‘‘found himself in accord with the conclusion of
Allen, that the interstitial tissue of the testis is derived from
the peritoneum, meaning thereby the mesothelium of the genital
ridge.”
404 CHARLES H. SWIFT
There are, then, two main opinions held as to the origin of the
interstitial cells; the one school, including Kolliker, Tourneux
(79) and Plato (97), hold that they arise from the connective
tissue cells; the other group, to which belong Niissbaum (’80)
and von Bardeleben, believe that they come from the general
epithelium. Allen (03) and Whitehead (04) must be classed
as holding the latter, and Sainmont (’05) the former opinion.
In the male chick embryo the interstitial cells are first evi-
dent in the 13 day testis. They are not numerous and are
evenly scattered throughout the testis. Usually they appear as
single cells but at times 2-6 may be found together. The cells
themselves have various forms, sometimes cubical, sometimes
polyhedral and frequently fusiform. The nucleus is round or
oval, stains deeply and usually has a central position (fig. 5).
The cytoplasm of these cells also stains strongly and appears
dense (fig. 5). The interstitial cells are remarkable, however,
because of the immense amount of fat which they contain.
This fat in the stained specimen has been dissolved out, but the
racuoles, which contained it are present, and occupy no small
part of the cytoplasm (fig. 5). If the section be examined care-
fully intermediate stages between the connective tissue cells
and the interstitial cells can be seen. That is, certain of the
connective tissue cells will be seen containing small vacuoles and
arranged at times in small groups.
In the 15 and 17 day embryonic testes there is an immense
amount of interstitial cell tissue. It is arranged in masses and
cords of varying sizes between the seminiferous cords. The in-
terstitial cells of these and later stages can be easily seen before
the section is stained and while it is still in paraffin. The inter-
stitial cell cords and masses, owing to the contained fat, stain
black with osmie acid and hence show up dark egainst the
lighter background. Sainmont (’05) described a rich plexus of
capillaries in relation to these masses of interstitial cells in the
eat, but in the chick there seems to be no increase in vascularity
around them.
Beginning with hatching and continuing until the chick is 10
days old there is a progressive decrease in the amount of inter-
SEX-CORDS AND SPERMATOGONIA IN CHICK 405
stitial tissue as compared with the true connective tissue and
the seminiferous cords. I do not know whether there is an
actual decrease, or only apparent, due to the great increase in
the size of the testis. In any event I was not able to find any
signs of degeneration in the interstitial cells.
It is an interesting fact that the first appearance of the inter-
stitial cells in the testis of the 13 day embryo is synchronous
with the appearance of the spermatogonia, and that they only
reach their maximum development when all the primordial
germ-cells have become spermatogonia.
There is no doubt, in the chick, that they arise from the
connective tissue of the testis, which in its turn is derived from
the general mesenchyme of the mesonephros. That they arise
from the connective tissue elements by simple differentiation is
proven, first, by the presence of transitional forms at the time
of their origin, and secondly, by the very rapid increase in the
numbers of the interstitial cells in the absence of any evidence
of division.
In a previous paper (Swift, 715), and in a preceding page of
this article, a short account was given of the arrangement of the
mitochondria around the attraction-sphere of the oogonia and
spermatogonia respectively. I shall hereafter call this body,
made up of sphere and mitochondria, the mitochondrial crescent.
As has been previously stated, the mitochondrial crescent
consists of attraction-sphere and mitochondria, and occupies
that part of the cell in which the cytoplasm is most voluminous;
in other words, it occupies a part of the vegetative pole of the
cell. The whole body in one plane, appears like a crescent
fixed on, or capping the nucleus, the attraction-sphere in the
middle enclosing the centrosomes, around it a thin clear area,
and extending down on either side the arms of the crescent
composed of mitochondria (figs. 3, 4, 5 and 6). In case the
knife does not section the nucleus, but passes through the sphere,
then the mitochondrial crescent appears as a circular mass,
composed of attraction-sphere, clear area and surrounding circle
of mitochondria (fig. 4). In nearly all cases the cytoplasm which
suspends the mitochondrial portion of the mitochondrial cres-
406 CHARLES H. SWIFT
cent appears denser and stains darker than that of the rest of
the cell (figs. 4, 5 and 6).
This mitochondrial crescent appears in cells which are the
result of primordial germ-cell division, in other words, in the
oogonia and spermatogonia. It appears in a few cells in the
female at 8 days and in the male at 13 days and this marks the
ending of the primordial germ-cell line and the beginning of the
oogonial and spermatogonial generations. By 17 days in the
male (fig. 4) every cel] of primordial germ-cell lineage possesses
it, thus indicating that all are spermatogonia. When the sper-
matogonia place themselves against the basement membrane,
when the cavity begins to appear in the cords, they do so in
such a manner that the mitochondrial crescent is against the
basement membrane (figs. 5 and 6).
This mitochondrial crescent, whose appearance and formation
I have just described, is the same structure which D’ Hollander
(04) described in the oocytes of the female chick. He, how-
ever, called the central sphere and its surrounding clear area
the yolk nucleus of Balbiani and the mitochondrial portion the
‘couche vitellogéne’ or ‘couche palléale.’ His methods did not
demonstrate the mitochondria, but in the region of the mito-
chondrial crescent which they occupy his text described, and his
figures showed a denser zone of cytoplasm. This denser zone
may be dissolved mitochondria or it may be more condensed
cytoplasm, which seems also to be present in my preparations
(fig. 4).
There is, of course, a great deal of dispute as to the nature
of the yolk nucleus—its cytological structure and its réle in the
cell.
Mertens (’94), in his study of the yolk nucleus of Balbiani in
the oocytes of birds and mammals, came to the conclusion that
two very different entities are described under the one name,
first certain elements found in the cytoplasm which originate
in the nucleus and, secondly, the attraction-sphere, which van
Bambeke called the ‘couche palléale’ and Van der Stricht the
‘couche vitellogéne.’ At the present moment [I am not. pre-
pared to say anything positively as to the origin of the denser
SEX-CORDS AND SPERMATOGONIA IN CHICK 407
material in the mitochondrial portion of the mitochondrial
crescent, but I am very certain that the attraction-sphere and
its surrounding body of mitochondria ought not to be called the
yolk nucleus of Balbiani for it does not have a nuclear origin.
The denser ground substance in the mitochondrial portion of
the mitochondrial crescent may be derived from the nucleus
and fall into the class of a true yolk nucleus or a chromidial sub-
stance in the sense of R. Hertwig.
It may be that the mitochondrial portion of the mitochondrial
crescent plays a role in the oocyte in the formation of vitellus, as
has been suggested, acts as a true ‘couche vitellogéne,’ but as to
its function in the male, I cannot even surmise.
SUMMARY
1. In the male chick the true sexual cords or seminiferous
cords originate from the germinal epithelium during the sixth
and seventh days of development and are the result of localized
activity of the epithelium. Nearly all the primordial germ-cel]s
present in the germinal epithelium are carried down into the
seminiferous cords but they play only a passive réle for at this
time they show no evidences of cell division.
2. The sexual cords remain attached to the germinal epithe-
lium for only a short time, and continue to grow, after formation
of the albuginea, as a result of division of the peritoneal cells.
3. At the end of the seventh day of development the sex of
the individual can be easily told, for in the male the gonads are
of nearly equal size, while in the female the left gonad is much
the Jarger. In the male the germinal epithelium remains thin
after the formation of the sexual cords and contains very few
primordial germ-cells, while in the female the epithelium of the
left gonad or ovary continues to be thick and contains many
primordial germ-cells.
4. During the eighth and ninth days of development the
gonads increase slowly in size and the thin sexual cords make up
most of the volume of the testes. They are separated from one
another by a thin layer of stroma and have a definite orienta-
tion from germinal epithelium obliquely down towards the
408 CHARLES H. SWIFT
Wolfhan body. When the embryo is 11 days old the stroma
begins to increase in quantity and the seminiferous cords com-
mence to meander and to anastomose with each other.
5. Up to the thirteenth day of development the primordial
germ-cells in the sexual cords do not divide, their numbers re-
maining about the same as when they left the germinal epithe-
lium. Beginning at this stage they divide actively and continue
to do so for the next four days. The primordial germ-cells
give rise to the spermatogonia, which are from now on very
numerous in the sexual cords. The spermatogonia differ from
the primordial germ-cells in possessing the mitochondrial cres-
cent. This cecupies part of the vegetative pole and consists of
attraction-sphere containing centrosomes, and mitochondria,
which in section are seen to extend over the nuclear mem-
brane, capping the nucleus. The arms of the mitochondrial
body may embrace more than one-half of the nucleus. This
mitochondrial body is suspended in dense cytoplasm and _ is
exactly comparable to the yolk nucleus as described by D’Hol-
lander and Van der Stricht in the oocyte. I have also described
it in the young oogonia of the chick. The interstitial cells
appear in the stroma on the thirteenth day but reach their
greatest development during the seventeenth day. They are
simply differentiated stroma cells, since they do not divide and
transitional forms can be seen.
6. Cavities begin to appear in the network of seminiferous
cords during the twentieth day. These cavities are formed, not
by fissures between the cells, but by liquefaction of the cells in
the central axis of the cord. At the twentieth day the sperma-
togonia are found against the basement membrane, with the
nucleus towards the central axis of the cord and the mitochon-
drial crescent near the basement membrane. They probably
reach this position by amoeboid migration. The elongated
cells between the spermatogonia are derived from the peritoneal
cells of the seminiferous cords.
7. The primordial germ-cells give rise to the spermatogonia
and the coelomic cells of the germinal epithelium produce the
supporting cells of the seminiferous tubule.
SEX-CORDS AND SPERMATOGONIA IN CHICK . 409
- BIBLIOGRAPHY
Auten, B. M. 1903 The embryonic development of the ovary and testis of
the Mammalia (preliminary account). Biol. Bull., vol. 5.
1904 The embryonic development of the ovary and testis of the
Mammalia, Am. Jour. Anat., vol. 3, p. 86.
1905 The embryonic development of the rete cords and sex-cords of
Chrysemis, Am. Jour. Anat., vol. 5
1906 The origin of the sex-cells of Chrysemis, Anat. Anz., Bd. 29.
Batrour, F. M. 1878 On the structure and development of the vertebrate
ovary, Quart. Jour. of Mic. Sci., vol. 18.
Brarp, J. 1904 The germ-cells, Part I, Jour. Anat. and Phys., vol. 38.
Bens.tey, R. R. 1911 Studies on the pancreas of the guinea-pig, Am. Jour.
Anat., vol. 12.
von BrreBERG-GossLER, H. 1912 Die Urgeschlechtszellen des Hi hnerem-
bryos aus 3 und 4 Bebritungstate., Arch. f. mikt. Anat. Bd., 81.
1914 Uber Herkunft und Wesen der sogenannten primiren Urge-
schlechtszellen der Amnioten., Anat. Anz., Bd. 47.
Bornuaupt, TH. 1867 Untersuchung iiber die Entwicklung des Urogental-
systems bein Hiihnchen, Inaug. Diss. Dorpat.
Bourn, M. 1900 Histogenése de la glande génitale femelle chez Rana tem-
poraria, Arch. de Biol., T. 17.
Braun, Max 1877 Das Tipe enitalayeters der einheimschen Reptilien Arch.
aus d. Zool., Zool. Institut in Wurzburg, T. 4.
Cownry, E. V. 1912 The relations of mitochondria and other ea pe
constituents in spinal ganglion cells of the pigeon, Intern. Monatschr.
f. Anat. u. Phys., Bd. 29.
D’Houuanper, F. 1904 Recherches sur l’oogenése et sur la structure et la
signification du noyau vitellin de Balbiani chez les oiseaux, Arch. d.
Anat. micr., TL. 7.
Dustin, A. P. 1907 Recherches sur l’origine des Gonocytes chez les Amphi-
biens, Arch. d. Biol., T. 23.
1910 L’origine et evolution des Gonocytes chez les Reptiles, Arch.
da Biolt yi. 25.
EIGENMANN, C. H. 1892 On the precocious segregation of the sex-cells of
Cymatogaster, Jour. Morph., vol. 5.
Erzoup, F. 1891 Die Entwicklung des Testikel von Fringilla domestica, etc.,
Zeit. fiir wiss. Zool., Bd. 52.
FirKET, JEAN 1914 Recherches sur l’organogénése des glandes sexuelles chez
les oiseaux, Arch. d. Biol., T. 29.
Horrmann, C. K. 1892 Etude sur le développment de l’appareil urogenital
des oiseaux, Verhand. der Koninklyke Akad. van Wetenschappen,
Amsterdam, Tweedie Sectie, vol. 1.
K6uiikmr, A. 1861 Entwicklungsgeschichte des Menschen und der hodheren
Tiere, Leipzig.
LAULANIE, F. 1886 Sur le mode d’évolution et la valeur de |’épithelium ger-
minatif dans le testicule embryonnaire du Poulet. C. R. Soc. Biol.,
Paris, T. 3
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, No. 3
410 CHARLES H. SWIFT
LotseL, G. 1900 Etudes sur la spermatogénese chez le moineau domestique,
J. de l’Anat. et de la Physiol., T. 36.
Mertens, H. 1894 Recherches sur la signification du corps vitellin de Bal-
biani dans l’ovule des Mammifers et des Oiseaux. Arch. d. Biol.,
ied:
von Mruarkowics, G. V. 1885 Untersuchungen tiber die Entwicklung des
Harn- und Geschlechtsapparates des Amnioten Intern. Monatchr. f.
Anat. u. Phys., Bd. 2.
Ntsspaum, M. 1880 Zur differenzierung des Geschlechts im Tierreich, Arch.
f. mikr. Anat., Bd. 18.
1901 Zur Entwicklung des Geschlechts beim Huhn. Anat. Verhand.
Puiato, 1897 Zur Kenntniss der Anatomie und Physiologie der Geschlecht-
sorgane. Arch. f. mikr. Anat., Bd. 48.
Poporr, N. 1908 L’ovule mAale et le tissu interstitiel du testicule chez les
animaux et chez homme, Arch. d. Biol., T. 24.
Ruspascukin, W. 1907 Ueber das erste Auftreten und Migration der Keim-
zellen bei Vogelembryonen, Anat. -Hefte, Bd. 35.
1912 Zur Lehre von der Keimbahn bei Siugetieren, Anat. Hefte,
Bd. 46.
Sarnmont, G. 1905 Recherches relatives a l’organogenése du testicule et de
Vovaire chez le chat., Arch. d. Biol., T. 22.
Semon, R. 1887 Die indifferente Anlage der Keimdriisen bein Hiihnchen und
ihre Differenzierung zum Hoden, Jen. Zeitschr. Naturwiss., Bd. 21.
Swirt, C. H. 1914 Origin and early history of the primordial germ-cells in
the chick, Am. Jour. Anat., vol. 15.
1915 Origin of the definitive sex-cells in the female chick and their
relation to the primordial germ-cells, Am. Jour. Anat., vol. 18.
TourRNEUX. 1879 Des cellules interstitielles du testicule. Jour. d. Anat. et
de la Physiol.
Wa.pEyer, W. 1870 Eierstock und Ei. Leipzig.
WuitreHead, R. H. 1904 The embryonic development of the interstitial cells
of Leydig, Am. Jour. Anat., vol. 3.
THE MORPHOGENESIS OF THE FOLLICLES IN THE
HUMAN THYROID GLAND
EDGAR H. NORRIS
Institute of Anatomy, University of Minnesota
SEVENTEEN FIGURES
CONTENTS
Vhs CLM YATROXS LOLOL AUC ere we aR SC NORM ENT ue 411
UG ALPEN Tee os oe Bits trim 2.0 tle er aaa Rete 412
Me Ma terralvandsmethodSaesas.<: << <6 os fos cca ee aka on 417\
Ve orphogenesis.ofthe thyroid follicles... :.. ae. tools oe 420
Dees ReLoliGular METIOMs cei. csc sss 5 oe apes eee 420
paeBollicularmertodt fre sso. osha oA ge Sean ea 423
NM eDISeISSionM ancl) CONCMISIONS: Li. .oiais.6 clo. ct. ee Re secrete etc 437
NLS SUED FOTW A ae Ae a eg 2 eee re a Cf. deh ey an eles te 445
Mele Pantera Gures C1LE sm ewstnichiin soe vs ces «yards 2 u)o oaks a I he icone 446
I. INTRODUCTION
In spite of numerous investigations, many questions concern-
ing the development of the thyroid gland are still unsettled.
This applies particularly to the morphogenesis of the thyroid
follicles. The various and contradictory views in the literature
on the development of the follicles are doubtless due in part to
lack of adequate series of successive embryonic stages (especially
of human embryos) available for study. The greatest difficulty,
however, has arisen from the use of inadequate methods of
investigation. ‘The method hitherto almost exclusively used,
that of direct observation of the microscopic sections, is insuffi-
cient. By the use of reconstruction methods, however, it has
been found possible in the present study to reach a satisfactory
solution of this difficult and important problem, at least regard-
ing some of the more fundamental features.
This study was undertaken in the Anatomical Laboratory of
the University of Minnesota at the suggestion of Prof. C. M.
Jackson, under whose supervision the work was conducted. I
wish to thank Dr. Jackson for his valuable aid and criticism.
411
412 EDGAR H. NORRIS
II. LITERATURE
The literature concerning the thyroid follicle will be consid-
ered in chronological order. First, it is desirable to mention
briefly the various views which have been held concerning the
morphology of the adult (human) thyroid follicle. Then fol-
lows a brief statement of the conclusions concerning follicle de-
velopment (including also prefollicular stages), which have been
arrived at by the observers who have worked upon human ma-
terial. A few observations, made on lower forms, which have
seemed especially pertinent to the present problem, are also in-
cluded. Unless otherwise indicated, however, all statements
tefer to human material.
The follicular structure of the adult human thyroid gland has
long been known. According to Boéchat (’73), Lalouette (1750),
who was the first to describe the minute structure of the thyroid
gland, found vesicles which seemed to communicate with each
other. Bardeleben (’41) is said by Zeiss (’77) to have been the
first to describe the adult thyroid follicles as isolated structures.
Five years earlier, however, Jones (’36) described the thyroid
follicles in considerable detail as completely closed vesicles.
Although there has been considerable disagreement concerning
the structure of the adult thyroid follicle, the majority of the
later observers have, like Jones, described the vesicles of the
adult gland as closed, spheroidal bodies. Cruveilhier (’43),
Virchow (’63), and more recently Boéchat (’73), Zeiss (77), and
Hitzig (94), however, have followed Lalouette in describing
the follicles as forming a system of branched and communicat-
ing cavities within the gland. Still others, like Streiff (97),
have maintained that both branching forms and isolated vesicles
occur in the adult gland.
Jones (’36), who was perhaps the first to describe the micro-
scopic structure of the human fetal thyroid, found that in a fetus
of four and one-half months the cells of the gland had become
partially arranged into solid, globular masses; but no vesicles
were observed at this stage.
MORPHOGENESIS OF THE FOLLICLES 413
Remak (’55) described in chick embryos in the wall of the
primitive saccular, epithelial thyroid anlage the formation of
thickenings which become separated and later give rise to the
thyroid follicles. He also thought that the original saccular
anlage might persist for some time and form new secondary
vesicles by a process of constriction. He described similarly the
origin of secondary follicles, both by constrictions and by solid
budding, in the thyroid of pig fetuses four inches and above in
length.
Peremeschko (’67) described the division of primary into sec-
ondary follicles in mammalian fetuses. Colloid is described as
arising partly by secretion and partly by colloid metamorphosis
of epithelial cells.
W. Miller (71) described a developmental stage in which
the thyroid consists of a network of cylindrical tubes. Such
tubes were found in a 24 mm. fetus and in decreasing numbers
in later fetuses and even in a three year old child. These tubes
arise from solid epithelial cords by the development of a central
lumen. The segmentation of the tubes with the formation of
the gland-vesicles is produced by ingrowth from the mesoblast.
Horcicka (’80) found the thyroid gland of a four months’ fetus
to be made up for the most part of solid cell masses with a be-
ginning of lumen formation in the central cells of these masses.
Typical gland structure is found after the fifth fetal month.
Wolfler (80) described the formation of follicles from solid
masses of epithelial cells. Toward the end of the fetal period
and after birth the peripheral cells of the groups dispose them-
selves in a circle. The central cells become at first granular,
then degenerate and disappear in the pale, granular mass
which fills the lumen of the vesicle thus formed.
L. Stieda (81) noted that the anastomosing epithelial cords
(‘Epithelstrange’) of the embryonic, mammalian thyroid are at
first always solid, but that in the ends of these cords lumina
appear, and the resultant vesicles are gradually constricted off.
Baber (’81) described the fully formed follicle as spheroidal
in form, but observed also branching follicles which are prob-
ably giving rise to secondary follicles by a process of division.
414 EDGAR H. NORRIS
Wolfler (83) described the process of later development in
the human thyroid gland as centrifugal. He distinguished a
cortical and a medullary portion, which are respectively young-
est and oldest, least developed and most developed portions.
His (’85) described in the thyroid gland of an embryo (Zw)
between 16.5 and 22 mm. in length cells grouped to form acini
or tubes. The inner ends of the cells have a light, colloidal
appearance.
Biondi (’89) found that the (postnatal) thyroid vesicle dis-
charges its contents, collapses and finally rearranges itself in the
form of a number of small acini which repeat the process. He
held that the colloid arises by cell secretion, and not by cell
degeneration.
Ribbert (’89) described a centrifugal growth of the thyroid in
embryos and newborn. Follicles are formed by the outgrowth
of solid buds or sprouts from the old follicles.
Lustig (’91), who studied the thyroid gland in the pig and
other animals, affirmed that colloid and follicles appear syn-
chronously as the result of the degeneration of the central cells
of the preéxisting solid masses.
Podack (’92) found well formed follicles in a fetus of five
months. In some parts of the gland the follicular structure is
only suggested and many cell-masses and cell-cords are present.
Marshall (93) found that the thyroid in chick and frog em-
bryos presents a stage in which the gland is made up of com-
municating, epithelial tubes. In the rabbit he described the
presence of out-growths, some solid and some hollow, from the
primitive epithelial anlage. In the human embryo: ‘‘At an
early stage the lobes are excavated by a number of detached
cavities, which become the vesicles of the adult thyroid.”
Zielinska (’94) found the structure of the thyroid in newborn
children variable both in size and numberof the follicles, and
also in the amount of solid cell masses. The relations ‘‘errinern
an acindse Driisen und erwecken den Gedanken, dass hier ein
sich verdistelnder Driisenkanal vorliegt, als dessen Endblaschen
die solide Zellhaufen gelten kénnen.”’
MORPHOGENESIS OF THE FOLLICLES 415
Hurthle (94) described in the thyroid of young dogs scattered
masses of interfollicular epithelium, in which new (primary)
follicles arise by the secretion of colloid into the angles between
adjacent cells.
Anderson (’94) described secondary follicles (postnatal) aris-
ing from the collapsed epithelium of emptied follicles in various
mammals. The new lumina are formed by cell-secretion of
chromophile spherules.
According to L. R. Miiller (96), the origin of small secondary
follicles from the larger follicles, as deseribed by Ribbert (’89)
is clearly evident, even in the human adult.
Tourneux and Verdun (’97) ‘in a careful study of the branchial
derivatives in the human embryo described the transformation
of the (median) thyroid plate (and later of the lateral thyroid
anlage) into a richly anastomosing network of solid epithelial
cords by ingrowth of vascular connective tissue in a 14 mm.
embryo. This network was likewise observed in embryos of
19 mm. to 37 mm. in length. At 37 mm., the cords become
varicose, and follicles develop by the formation of a cavity
within each of the enlargements. A similar process of morpho-
genesis is described in the rabbit embryo by Soulié and Verdun
(97).
Streiff (97) made wax reconstructions of normal, adult
human thyroid tissue, and found it to be made up of closed fol-
licles, ovoidal or spindle shaped. Branched forms due_ to
budding or to secondary fusion were also described; some of these
more complex forms he thought may, represent persistent branch-
ing, a continuation of the embryonic process. He concluded
that the thyroid arises as a branched tubular gland, the follicles
being formed by constriction of the tubes.
Schreiber (98), in a fetus of three months, found the thyroid
gland for the most part arranged into follicles which contained
much colloid.
Kiirsteiner (’99) in fetuses from 8 to 30 em. in length found
the thyroid lobules made up of round or elongated, solid or
hollow follicles. The lumina are few in number up to about 20
em., but in the older fetuses they are numerous and evenly
416 EDGAR H. NORRIS
distributed throughout the Band Some branching vesicles
were also noted at 17.5 cm.
Prenant (01) (p. 13) stated that in the embryonic thyroid the
solid epithelial cords are transformed into a network of tubes
from which the follicles arise by a process of constriction.
Von Ebner -(’02) found numerous well developed follicles in
older fetuses and newborn. Between the follicles are found,
even in the adult, frequent solid strings and nests of epithelial
cells, which are in the majority during development.
Elkes (03), who studied the thyroid in fetuses from four and
one-half to six and one-half months in age, found that it pre-
sents both solid cords and well developed follicles in variable
number. In the newborn the earlier follicles have largely
disappeared, leaving only a few at the periphery of the gland.
Hertwig (10) (pp. 444-446) described in the embryonic thy-
roid anastomosing epithelial cylinders. These become tubular;
varicose dilations are by ingrowth of the adjacent connective
tissue cut off to form the permanant follicles.
Isenschmid (’10) found that in the thyroid of children the
gland grows not only by the increase in the size of the follicles,
but by the formation of new follicles by two methods: budding
and division. He found no evidence that follicles are formed
from solid cell-masses (interfollicular epithelium) retained from
the embryonal period.
According to Hesselberg (’10): ‘‘Die Ausbildung der Thy-
reoidea in der fétalen Periode erfolgt durch Zerfall der urspriing-
lich soliden Zellplatte in solide Zellstriinge. Diese schniiren
sich zu Bliischen ab, die zuerst am kaudalen Pol auftreten.”’
The normal structure of the thyroid is established from the
fourth fetal month on. Desquamation of epithelial cells was
found in about half of the cases from the seventh to the ninth
fetal month, and the follicles are almost entirely obliterated in
the newborn. During the first week of postnatal life the fol-
licles are reformed and increase in number by a process of
budding.
Prenant and Bouin (’11) give an account of the development
of the median thyroid anlage similar to that given by Pre-
nant CO).
MORPHOGENESIS OF THE FOLLICLES 417
Broman (’11) described in the differentiation of the thyroid
anlage a tubular stage transformed by constrictions into beaded
chains and finally into separate follicles.
According to Grosser (12), the thyroid anlage begins to sepa-
rate into solid cords in the human embryo of 8 mm. _ In the 50
mm. fetus the cords, especially in the periphery, appear beaded.
The beaded cords become divided into separate cell-masses, the
anlages of the follicles. The lumina may appear as independent
cavities (no tubular stage), before the follicles are detached, or
they may arise later even in early postnatal life.
Simpson (’12) referred to the tubular structure of the thyroid
and describes the gland of a seven months old child as tubular
in character.
Aschoff (13) stated that in the developing thyroid gland con-
nective tissue separates round epithelial balls from the anasto-
mosing cords, and it is in these ‘balls’ that the follicular lumina
develop.
Sobotta (15) described the first lumina as appearing in the
peripheral parts of the lateral thyroid lobes in a fetus 50 mm. in
length. The final breaking up of the cell cords into single
groups, which will later form follicles, progresses very gradually,
so that the final structure of the gland is arrived at only after
birth. Interfollicular epithelium persists, which may later give
rise to follicles. .
Kingsbury (15) in describing the early development of the
thyroid states that lumina (follicular cavities) appear within the
cell cords in fetuses of 32 mm., but colloid is not demonstrable
until 40 mm.
III. MATERIAL AND METHODS
This study is based upon the collection of human embryos in
the Anatomical Laboratory of the University of Minnesota.
Several of the series used are in excellent condition for histo-
logical study. The embryos of the collection have been vari-
ously fixed and stained. The additional glands specially pre-
pared have, for the most part, been fixed in Formol-Zenker;
embedded in paraffin, sectioned at 10 4; mounted serially and
418 EDGAR H. NORRIS
stained with alum-haematoxylin and eosin, or iron-haematoxy-
lin and eosin. -
Besides the specimens mounted in the collection, otherglands
from newborn children and from children in the early years of
life were obtained at autopsy. These were used merely for pur-
poses of comparison, and have neither been listed below, nor
discussed in this paper.
The following table shows the materials used in this work.
The embryos and fetuses used are arranged in the order of their
crown-rump lengths. ‘Minn. E. C.’ refers to the Minnesota
Embryological Collection. An asterisk (*) following the num-
ber signifies that the thyroid gland alone was sectioned. Other-
wise the entire embryo was available in serial sections.
The ordinary reconstruction methods, both plastic (Born’s
wax-plate method) and graphic, were utilized in the present
study. In all cases where a determination of the follicular form
or structure was attempted, special precautions were observed
in making the reconstructions as accurate as possible.
The drawings for reconstruction were made with the camera
lucida on transparent paper. After the drawings were com-
pleted, those of successive sections were superimposed upona
tracing-table, and each epithelial structure in the section given
a letter or number. The drawings were controlled by careful
microscopic observations, to determine the frequently compli-
cated relations of neighboring follicles. By this method it was
possible to determine with certainty what the limits of any par-
ticular mass or follicle might be.
SECTION
- R. LE
SERIAL NO. MINN. E. C. NO. oN FIXATION THICKNESS IN
IN MM. MICRONS
1 H6 6.0 Zenker 10
2 Ist 1s eo Alcohol 15
3 H 60 11.0 Aleohol 20
4 H 68 11.0 Bouin 15
5 H 134 TRO Aleohol 20
6 H1 15.0 Alcohol 12
7 H 23 15.0 Alcohol 10
8 H 18 15.5 Formalin 10
9 H 28 16.0 Alcohol 10
10 H 62 16.0 Formalin 20
ll H 58 ivan) Formalin 20
12 H. 260 18.0 Formalin 5)
13 H 24 19.0 Zenker 12
14 H 2 20.0 Picro-sulphuric 15
15 late 20.0 10
16 H 265 21.0 Formalin 20
17 H 3 22.0 Zenker 12
18 H 15 22.0 Alcohol : 12
19 H 64 23 (0) Alcohol-formalin 20
20 H 304 24.0 Zenker 20
21 H 56 24.0 Alcohol 20
22 Ele 25.0 Zenker 15
23 H 21 26.0 Alcohol 15
24 H 29 26.0 Alcohol 12
25 H 99 26.0 Alcohol-formalin 20
26 H 48 27.0 Bouin 12
27 H 10 29.0 Alcohol 15
28 lel ayes 29.0 Formalin 10
29 H 259 30.0 Formalin 20
30 H 108 30.0 Zenker 10
31 H 57 31.0 Formalin 20
32 H 16 33.0 Zenker 15
33 H 313 35.0 Formalin 20
34 Hy 122 39.0 Formalin 20
35 H8 41.0 Formalin 12
36 ez, 41.0 Zenker 15
37 H 121 46.0 Formalin 60
38 H 115 50.0 Formalin 40
39* H 290 56.0 Formalin 10
40* H 85 60.0 Formalin 10
41 H 26 65.0 Formalin 30
42* H 285 66.0 Formalin 10
ae H 81 83.0 Alcohol 10
44* H 286 86.0 Zenker ‘ 10
Abe lal (x 90.0 Formalin 10
46* H 267 115.0 Formalin 10
47* H 34 120.0 Formalin 10
48* H 49 126.0 Bouin 10
49* H 187 158.0 Formalin 10
50* H 381 163.0 Formol-Zenker 10
419
420 EDGAR H. NORRIS
IV. MORPHOGENESIS OF THE THYROID FOLLICLE
a. Prefollicular period
A brief consideration of the thyroid gland in the prefollicular
period is essential to an understanding of the follicular develop-
ment. For it is during this prefollicular period that the anlages
of the primitive follicles are derived from the original epithelial
mass; and, to a certain extent at least, the size, form, and ar-
rangement of the earlier follicles are thereby predetermined.
The earlier well-known stages in the development of the
(median) thyroid anlage in the human embryo are not consid-
ered in the present paper. The original thyroid diverticulum
becomes detached and transformed into a small epithelial plate,
well shown in the 6 mm. embryo (No. 1 of the present series).
This and several succeeding prefollicular stages of the thyroid
were carefully reconstructed by Born’s wax-plate method, but
for the present purposes it is unnecessary to figure or describe
these models.
As is well known, the (median) thyroid epithelial plate soon
presents irregularities, as shown in the 7.5 mm. embryo (No. 2),
and rapidly becomes transformed into what appears in cross-
sections to be (as heretofore almost universally described) a
network of anastomosing epithelial cords.
With the details of this process of transformation the present
study is not concerned. One feature of the end result, however,
which comes out clearly in the reconstructed models, is that the
cord-like appearance seen in the sections is largely an illusion.
Fundamentally the plate-like structure of the thyroid anlage
persists for a considerable time, although somewhat modified by
a complicated process of fenestration, splitting, and budding
during the growth of the primitive epithelial plate. The re-
sultant structure consists essentially in a mass of irregular,
branching and fenestrated plates, for the most part longitudinally
arranged (parallel to the long axis of the body), so that in cross-
section they appear as ‘cords’ of epithelial cells (fig. 1). This
type of structure, with varying degrees of complexity is found
in the various prefollicular stages of embryos from about 10 mm.
MORPHOGENESIS OF THE FOLLICLES 421
to 22 mm. in length. (Numbers 3, 4, 5, 6, 7, 8, 9, 10, 11, 12,
13, 14, 15, 16, 17, 18 of the present series.)
Such a description, however, does not apply to the so-called
lateral thyroid anlage (ultimobranchial body). As described by
Tourneaux and Verdun (’97) and others, this body long remains
Fig. 1 Semi-diagrammatie drawing (camera lucida outlines) of a trans-
verse section through the left lateral lobe of the thyroid gland and neighboring
.Structures from a human fetus 22 mm. long (No. 18). This shows the complex
interrelation of the smooth epithelial plates (Pl.) (which appear as ‘cords’ in
the section) and the vascular mesenchyme (Mes.) found in the interspaces be-
tween them. The apparently detached epithelial masses (Ma.) shown are sec-
tions through projections from plates located in the upper part of the isthmus
of the gland. B.V., blood vessel; M., muscle; L., larynx; C., cartilage. X 133.
as a compact, deeply staining mass of epithelium (the ‘inner
condensation’ of Kingsbury 714) on the dorso-medial aspect of
each lateral lobe. Whether in the human thyroid the lateral
anlage later atrophies or finally becomes transformed into per-
422 EDGAR H. NORRIS
manent thyroid tissue is still uncertain. This region is there-
fore purposely excluded from the following descriptions, which
apply primarily to the process of morphogenesis in the struc-
tures derived from the median thyroid anlage.
The thyroid gland in the final stage of the prefollicular period °
is shown in a fetus of 22 mm. (No. 18). This stage will be de-
scribed in some detail, in order to make clear the subsequent
process of morphogenesis of the thyroid follicles. When studied
in sections, the thyroid at this stage is seen to be made up appar-
ently of epithelial ‘cords’ which are interrelated in a complex
manner (fig. 1). A loose, anastomosing network is thus formed,
the interspaces of which are filled with a vascular mesenchymal
tissue (fig. 1). The ‘cords’ are, in general, only two cells in
width, a feature characteristic of the thyroid plate and its
derivatives at various stages.
Upon reconstruction (figs. 11 and 12) it is found that the epi-
thelial network seen in sections, and described by so many ob-
servers as consisting of ‘rods’ or ‘cords,’ is merely a section of
flat, slab-like plates, or bands, representing portions of fenes-
trated plates. It is true that some few of the epithelial masses
are actually cord-like in form, but by far the greater number
are better described as bands or plates. <A definition of terms
is necessary at this point. The term ‘plate’ will be used to sig-
nify a structure of relatively slight thickness, presenting expan-
sive surfaces which are more or less smooth, and which may or
may not be perforated (fenestrated). A band is a narrow
plate. Therefore a fenestrated plate may be considered as
made up of a number of anastomosing bands. There is great .
variability in the way in which these plates and bands are
arranged. Some are mere slabs, which may or may not be per-
forated; others form irregular prisms or rounded cylinders,
which open at both ends into the surrounding mesenchyme.
The length and width of these plates is also quite variable. In
three respects, however, they are in general agreement. ‘They
are two cells thick; present fairly smooth surfaces; and are
longitudinally placed. Within these epithelial bands (fenes-
trated plates) the primitive follicles of the thyroid gland develop.
MORPHOGENESIS OF THE FOLLICLES 423
b. Follicular period
A thyroid follicle will be defined as a completely closed sac
whose wall is usually made up of only a single layer of epithelial
cells. This definition includes all the features of the follicle
which may be regarded as absolute and constant. The size and
shape of the follicle may vary, and great differences are found in
these respects in follicles of the same gland as well as in follicles
of fetuses at different stages. Typically the thyroid follicle may
be considered spherical or spheroidal in shape; but, as will appear
later, this type is subject to considerable variation. The term
primary follicle will be used to include those follicles develop-
ing independently of preexisting follicles. The follicles derived
from preexisting follicles, by budding or otherwise, are termed
secondary follicles.
In the present series, the first primary thyroid follicles appear
in a fetus 24 mm. in crown-rump length (No. 20). In this fetus
the thyroid gland has essentially the same structure as has that
of the last (No. 18) described in the prefollicular period, i.e. it
is made up chiefly of longitudinally placed epithelial plates or
bands, only two cells in thickness. But in this stage, the plates,
which have in previous stages been characterized by compara-
tively smooth surfaces, now present surfaces which are more or
less roughened by the appearance of scattered hillocks or
mounds (figs. 13 and 14). They are placed very irregularly
with respect to one another, and may appear for the first time
in any part of a plate, at its periphery or in a more central
region.
When studied in cross sections (figs. 2 and 3) it is found that
these hillocks are the immediate anlages of the early thyroid
follicles. It is further seen that the hillocks are apparently pro-
duced by the concurrence of four different processes in the epi-
thelium. The first process is that of cell rearrangement, the
second that of cell proliferation, the third that of absolute cell
growth, and the fourth that of lumen formation. These proc-
esses, although described separately, may occur simultaneously.
The first departure from the two-celled plate arrangement, in
the process. of follicle formation, is found in a rearrangement of
424 EDGAR H. NORRIS
the cells of the plate (fig. 3). The cell outlines can be made
out only with difficulty in most cases. But in those places
where they can be seen, they bound cells which are more or less
columnar in form. The nuclei are ovoidal or elliptical in out-
line and are placed with their long axes perpendicular to the sur-
face of the plate. Here and there along the course of the plate
(fig. 3) it will be noticed that some of the nuclei have shifted
their axes and have changed their relative positions. Certain of
the nuclei have rotated through an are of 90 degrees so that their
CTR O71 Wt
Fig. 2 Semi-diagrammatie drawing (camera lucida outlines) of a section
through the right lateral lobe of the thyroid gland and neighboring structures
of a human fetus 30 mm. long (No. 30). Shows epithelial plates developing
rough surfaces; also early follicles in various stages of separation from the plates.
Note the general increase in width of the plates as compared with those shown
in figure 1, and the swellings and constrictions, giving ‘beaded chain’ appearance.
Cap., capillary; Fol.L., follicle lumen; S.#p.Pl., smooth epithelial plate ;
R.Ep.Pl., rough epithelial plate; Mes., mesenchyme; C., cartilage; L., larynx.
x 133.
MORPHOGENESIS OF THE FOLLICLES 425
Fig. 3 Small portion of a cross section of the thyroid gland in a human
fetus 30 mm. long (No. 30), magnified to show the cell structure. The location
of the section is indicated (a-b) in figure 14. Note the appearance of lumina
(L.) in buds (B.) from the surface of the plate as well as in swellings along its
course (L’). In the upper left hand corner a section through one of the hollow
epithelial prisms is seen (Hp.P.). Mes., mesenchyme. X 400.
Fig. 4 Drawing of a section through follicle ‘d’ in figure 16. Note the irreg-
ular form of the follicle, both the lumen (L.) and the wall. Mes., mesenchyme;
S.B., solid bud. X 400.
Fig. 5 Drawing of a section through an epithelial plate taken at the level
marked a—b in figure 7. Mes., mesenchyme; L., lumen; Cap., capillary; Kp.T.,
epithelial tag. > 400.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
426 EDGAR H. NORRIS
long axes, in their final position, are at right angles to their origi-
nal position in the plate. Asa result of this shifting process, little
circlets (really spheres) of nuclei are formed in the plate.
This shifting of nuclei is but the visible expression of the
changed position of the cell. For while it is impossible to ob-
serve the cell boundaries in most cases, it is hardly probable
that the nuclei shift their axes independent of the cytoplasm;
moreover, the few faint cell-boundaries which may be made out
show the same changes in position as do the nuclei. Further, it
is usually found that at the point from which a nucleus has
shifted toward the center of the plate a shght depression appears
on the surface of the plate, indicating that the cytoplasm has
shared equally with the nucleus in the movement. From these
three facts it may be concluded that the first process manifested in
follicle formation is the shifting of the axes of certain cells of the
epithelial plate through an arc of 90degrees. There is no evidence
that the depressions on the surface of the epithelial plate are due
to invasions or activity of the adjacent mesenchyme (fig. 3).
This process results in the transformation of the smooth sur-
faces of the bands (fenestrated plates) into surfaces which are
somewhat roughened. The irregularities are apparently not, at
first, due to swellings on the plates, but rather to the shght in-
dentations produced by the shifting of certain cells toward the
center of the plate as above described. When studied in cross
sections (fig. 2) such a plate appears as a sort of beaded chain,
with alternate swellings and constrictions. But, as noted
above, the initial swellings due to this process are only apparent
and are actually not greater in thickness than is the plate in
other parts of its extent where indentations have not yet
occurred.
The extraordinary cellular activity of the epithelium at this
stage is clearly manifested by the large number of mitotic figures
to be seen. There is no section of the gland in the 24 mm.
fetus (No. 20) which does not present several cells in process of
mitosis. But the localization of these mitoses is even more sig-
nificant than is their number. It will be noticed that the
nuclear figures are usually found only in those places in the
MORPHOGENESIS OF THE FOLLICLES 427
epithelial plates where actual thickenings on the plates are
being formed. Therefore the little mounds which appear on
the plates, as the immediate anlages of the early follicles, may
be formed not only by the rearrangement ofthe already existing
cells of the epithelial plates, but also by the formation of new
cells as well. Consequently, it can easily be seen how the ap-
parent swellings on the plates, produced by the rearrangement
of the existing cells, may be transformed into actual swellings,
by the absolute increase in number of the cells found in a local-
ized area. These swellings become roughly spheroidal in form.
The third process referred to above is the absolute increase in
size of the cells. While the cells are shifting their axes and pro-
liferating, they are also growing in size. This fact results in
the appearance seen in figure 2, where the solid, two-celled
plates are found in some cases to be no greater in cross section
than the one-celled plate which surrounds the follicle lumen. It
might be thought that the cells do not actually increase in size,
but only increase in height by a closer crowding together. But
a study of figure 3 will show that such is not the case. For the
cells are not more closely packed together in the newly formed
follicles than they are in the two-celled plate.
This progressive increase in the height of the cells corre-
sponds to the progressive stages in the differentiation of the two-
celled plate into newly formed follicles. So that the thyroid
gland of a 30 mm. fetus (No. 30) presents in different regions
epithelial cells varying greatly in height. The lowest cells are
found at the beginning of the process, in the two-celled plate;
the highest being found at the other extreme, in the completely
formed follicle.
Three of the four processes above mentioned as apparently
involved in the evolution of the follicle from the epithelial plate
have now been reviewed in detail. The formation of the lumen
remains to be considered. Just preceding the appearance of the
lumen, the spherule (in which it is about to develop) appears in
cross section as a circlet of very tall columnar cells, whose
nuclei are peripherally placed. This arrangement results in the
formation of a striking picture. The nuclei are regularly placed
428 EDGAR H. NORRIS
at the periphery of the circle and form a dark band, which
surrounds an expansive, clear, central cytoplasmic portion.
The magnitude of this cytoplasmic area and the sharp contrast
between the two portions (in the stained preparations) are
usually striking features (fig. 3). It is in the center of this cyto-
plasmic area that the lumen makes its appearance as a tiny
spherical space outlined by a definite and regular margin. It is
as though the cells had but drawn a little apart, so that their
central ends, instead of remaining in contact one with another,
might be separated by an interval. The relation of the early
lumina to one another is well shown in figure 6, which is a
graphic reconstruction of a plate. It is important to note that
no tubular stage is found in the process of lumen formation.
The lumina appear as absolutely independent spaces.
As the lumina first appear they apparently have no content;
but undoubtedly they contain some substance which is not
stained with the ordinary methods, and which increases in
amount with the size of the follicular cavity. Certain of the
larger lumina (not all of them), which are perhaps older, are
found to contain a hazy, granular substance. Typical colloid
does not appear until later, in the 60 mm. stage (No. 40).
The various possibilities as to methods by which the follicular
lumen may arise will be considered later in the discussion.
Fig. 6 A graphic reconstruction (surface view) of parts of two fenestrated
epithelial plates from the thyroid gland of a human fetus 30 mm. long (No. 30)
to show the relative position of the lumina (Fol.L.) as they appear in the plate.
Lumina indicated by stippled areas. Note that in all cases these early lumina
are quite distinct and never connected with one another. Perf., perforation.
x 267.
Fig. 7 A graphic reconstruction of a number of epithelial masses from the
thyroid gland of a human fetus 60 mm. long (No. 40). Note the various degrees
in the breaking up of the plates and the relative positions of the lumina. a-b
indicates level of section in figure 5. X 267.
Fig. 8 A graphic reconstruction of a follicular complex shown in the lower
part of figure 17. Lumina indicated by stippled areas. Note the solid buds.
x 26%.
Fig. 9 Model (reconstructed by Born’s wax-plate method) of a large follicle
from the peripheral part of the lateral lobe of the thyroid gland in a human
fetus 60 mm. long (No. 40). The three buds shown have lumina communicat-
ing with the lumen of the main follicle. > 267.
MORPHOGENESIS OF THE FOLLICLES 429
430 EDGAR H. NORRIS
From the time of their first appearance, considerable varia-
bility in the size of the lumina found in any particular gland is
to be noted. The first isolated follicles are found in the more
peripheral parts of the thyroid gland, and it is in these regions
that they first attain large size and considerable complexity.
For some time after the formation of follicles has begun, all
stages previously described may be found in different parts of
the same gland, a considerable portion of which retains the
irregular plate-like type of structure characteristic of the pre-
follicular stages.
In the foregoing account, the cell masses in which the lumina
develop have been described as spherules whose cross section is
circular in outline. While this is true for typical follicles and in
most cases, some variation within comparatively narrow limits
is found. Ovoidal or somewhat irregular follicles occur, but
these are not more numerous than would be expected in a rapidly
erowing tissue.
The foregoing descriptions of the early primary follicles have
been taken in large part from observations made on two fetuses,
one of 24 mm. (No. 20) and one of 30 mm. (No. 30). The ex-
cellent condition of these specimens has made possible studies
of considerable detail. The members of the series (Nos. 21, 22,
23, 24, 25, 26, 27, 28, 29) intervening between these two, al-
though not favorable for such intensive studies, show substan-
tially similar structure. These stages may be summarized very
briefly. The comparatively smooth epithelial plates of the pre-
follicular period have been transformed into plates with rough
surfaces. The roughenings on the plates are the early indica-
tion of the follicles about to be formed. With the progressively
increasing number of follicles the plates are transformed into
irregular bands, which in turn give rise to groups of solid or hol-
low masses of cells. In the 30 mm. stage, however, the thyroid
gland is still largely made up of anastomosing bands (fenes-
trated plates) (figs. 3, 6, 138, 14), although some entirely isolated
follicles are found.
While the fetus is increasing in length from 30 mm. to 50 mm.
the thyroid gland presents merely a continuation of the process
MORPHOGENESIS OF THE FOLLICLES 431
above described; so that by the time the 50 mm. stage is reached
the band or plate formations are relatively insignificant, while
the isolated epithelial masses make up the greater part of the
organ. All stages in the breaking up of the bands or plates may
still be found, however. At this stage (50 mm.) the gland is
therefore characterized by the presence of few epithelial plates
or bands, very many isolated, solid cell-masses (anlages of fol-
licles) and relatively few well developed follicles.
At 56 mm. (No. 39) the thyroid gland presents slightly more
follicles than at 50 mm. Not only are the follicles increased in
number, but they also show some changes in both size and form.
In the previous stages the follicles have ranged from 16 to 55 u
in diameter (average diameter about 40 uw), and have been typi-
cally spherical in form. But in this stage many of the follicles
at the periphery of the gland have enlarged considerably and
have departed from the more usual spherical form of previous
stages. This condition is more strikingly seen in the next fetus
of the series (60 mm., No. 40) where both increase in size and
change in form of the follicles are very evident (fig. 15). The
number of follicles at this stage appears about equal to the num-
ber of solid epithelial masses. Figure 7 is a graphic recon-
struction of a number of epithelial masses from which follicles
are being formed and of certain isolated follicles. A study of
sections at this stage will show follicles situated at the periph-
eral portions of the lobe whose diameters are three or even four
or five times as large as those of greatest magnitude located
more centrally. The average diameter of the centrally placed
follicles is about 15 microns, while the average diameter of the
follicles in the peripheral regions is approximately 55 microns.
There is much less absolute variation in the size of the central
follicles than in those of the periphery. Centrally placed fol-
licles range from 13 to 17 microns, while peripherally placed
follicles range from 20 to 103 microns in diameter. It will
further be seen that these same large peripheral follicles are
quite irregular in form. Instead of being circular in outline,
they have a more or less regularly elliptical or ovoidal form, and
may present a number of solid or hollow buds. Figure 9 rep-
432 EDGAR H. NORRIS
resents a reconstruction of a typical one of these large budding
follicles. Figure 7 is a graphic reconstruction of a number of
epithelial masses found in the thyroid gland of the 60 mm. fetus
(No. 40) to demonstrate stages in the breaking up of the plates
and the formation of isolated follicles, which are present at this
time.
The 65 mm. fetus (No. 41) is the first member of the series in
which practically no epithelial plates or bands are found. From
this stage on through the remainder of the series the thyroid
Fig. 10 Microphotograph of a portion of a cross section of the thyroid gland
in a human fetus 86 mm. long (No. 44). Note the numerous branching follicles,
also the elongated follicles in various stages of division by constriction. This
section is in the region of the reconstruction shown in figure 15. 100.
gland is made up almost entirely of independent, solid or hol-
low, more or less irregular epithelial cell-masses. The stages
(Nos. 42 and 43) included between this embryo and the one of
86 mm. (No. 44) may be passed over briefly with the statement
that the processes of budding and formation of new follicles as
previously described are advancing rapidly parallel with the
increasing length of the fetus.
The 86 mm. fetus (No. 44) deserves special attention for it
exhibits what appears to be the culmination of the budding proc-
esses referred to above. <A study of figure 10 reveals the ex-
MORPHOGENESIS OF THE FOLLICLES 433
treme complexity of the gland at this stage. The hollow and
solid epithelial masses are about equal in number. The ex-
treme variability in the form of the follicles is the most striking
feature (figs. 4, 10 and 16), and the reconstructions reveal more
clearly the complexity of these structures than is apparent in
sections. Elongated follicles with numerous hollow or solid
buds, as well as numerous varieties of other more or less com-
plex arrangements are to be seen. The whole picture is one of
active growth. There is great variability in the size of the fol-
licles, the largest being for the most part located in the periph-
eral zone; but the extreme variation in the follicular form makes
absolute measurements for comparison of follicles of little value.
This condition of the gland prevails in the remaining members
of the series, up to and including the 158 mm. fetus (Nos. 45, 46,
47, 48, 49). There are but two differences to be noted. In the
first place, the relative number of follicles is increasing from
stage to stage, so that by the time the fetus is 158 mm. in
length, the follicles in number are far in excess over the solid,
interfollicular masses. Secondly, the number of irregular and
complex forms, although still occurring, is becoming fewer (fig.
17), and the spheroidal follicles are relatively numerous.
The thyroid of the 163 mm. fetus (No. 50) has quite a different
structure from that described for the members of the series just
preceding (Nos. 44, 45, 46, 47, 48, 49). The irregular and
branching follicular complexes are comparatively few in num-
ber, while small, spherical follicles make up almost the whole
of the tissue. There are, however, a few solid, interfollicular
epithelial masses still present at this stage. The picture has
changed from one in which the epithelial structures, instead of
being complex in their form, have become relatively simpler in
character and for the most part are organized into small
follicles.
My material from this stage on up to and including newborn
children shows to a greater or less extent the process of epithe-
lial desquamation in the thyroid follicles as described by Elkes
(03), Hesselberg (710), Isensehmid (’10) and others. Whether
this process is physiological or pathological is as yet undeter-
434 EDGAR H. NORRIS
ne Gy % re
> Nae re
MORPHOGENESIS OF THE FOLLICLES 435
mined. Because the material at hand is not sufficient to war-
rant the drawing of any conclusions in this matter, it is not
listed with the materials nor discussed in this paper.
In four of the fetuses studied (Nos. 40, 42, 43, and 44) a
number of cyst-like follicles located in all cases in the lower and
posterior (dorsal) part of the lateral lobe, were observed. Some
of these follicles measure as much as 200 microns in diameter.
The size of these structures is not much greater than that of some
of the larger normally appearing follicles. But in structure they
are quite different, having walls made up of very much flattened
epithelial cells, whose nuclei cause the cells to bulge and are
separated from one another by much greater distances than is
the case with the nuclei in the more usual follicles. The lumina
of these cysts are quite regularly circular in outline and in many
cases contain a granular substance. It is as though a follicle
had been greatly distended, the cells of the wall being stretched
and flattened.
Fig. 11 Model (reconstructed by Born’s wax-plate method) of the left lat-
eral lobe of the thyroid gland from a human fetus 22 mm. long (No. 18). Antero-
lateral surface view to show the perforated (fenestrated) plates, which present
a relatively smooth surface. Line a—b indicates the level of the section shown in
figure’1, and the level at which the model was divided to show the relations
pictured in figure 12. The carotid arteries are shown on the right hand-side of
the figure. X 150.
Fig. 12 The lower section of the model pictured in figure 11. The upper
portion has been removed in order to demonstrate the relations to the gland
mass of the appearances found in cross section. The separation of the two
portions is made at the level of the section shown in figure 1 and indicated (a—b)
in figure 11. X 150.
Fig. 13 Model (reconstructed by Born’s wax-plate method) of the upper
portion of the left lateral lobe of the thyroid gland from a human fetus 30 mm.
long (No. 30). Plates are more broken up than in figure 11 and present surfaces
which are relatively rough and irregular. The upper parathyroid (P.) is shown.
Lateral view. X 100.
Fig. 14 Model (reconstructed by Born’s wax-plate method) of a part of a
plate from the thyroid gland of a human fetus 30 mm. long (No. 30) at a higher
magnification, to demonstrate more exactly than shown in figure 13 the appear-
ance and relations of the mounds forming on the surface of the plate. The
plane a—b indicates the position of the section shown in figure 3. Lateral view.
6 Pie
y
MORPHOGENESIS OF THE FOLLICLES 437
V. DISCUSSION AND CONCLUSIONS
Remak’s theory of the derivation of the thyroid follicles di-
rectly from the primitive saccular thyroid anlage has not been
confirmed. In the prefollicular stages, the thyroid is by recent
investigators quite generally described as assuming the form of
irregular, anastomosing ‘cords’ or masses of epithelium. This
undoubtedly appears to be the case when sections of the gland
are observed (figs. 1 and 2). But the reconstruction methods
used in the present investigation reveal a surprisingly different
condition. It is found that, as a matter of fact, in the great
majority of cases the cords are illusions and in reality are
merely sections of fenestrated epithelial plates longitudinally
arranged.
As to the further steps in the process of morphogenesis of the
follicles from these anastomosing ‘cords’, widely divergent views
have been held, as noted in the section on ‘Literature.’ While
differing considerably in detail, these views may be classified
according to their principal features. The morphogenesis of the
primary follicles and of the secondary follicles will be considered
in their order.
According to the apparently most widely accepted v-ew, based
perhaps ‘argely upon preconceived ideas or theories concerning
the evolution and morphogenesis of the thyroid gland, there are
two d'stinct stages in the transformation of the epithelial ‘cords’
into follicles. First the anastomosing ‘cords’ acquire lumina, so
that the gland becomes a more or less definite network of hol-
low epithelia! tubes. The tubes then become constricted (by
Fig. 15 Wax reconstruction of a portion of a peripheral lobule of the thy-
roid gland of a human fetus 60 mm. long (No. 40). This model is from the region
shown in figure 10. The structure placed across the top of the figure is a blood
vessel. The oblique lines designate cut surfaces. Viewed from above. X 270.
Fig. 16 A series of follicles (a, b, c, d, e) reconstructed by Born’s wax-plate
method to show the varying degrees of complexity. All are taken from the
thyroid gland of a human fetus 86 mm. long (No. 44). x 135.
Fig. 17 Wax reconstruction (Born’s method) of a peripheral lobule of the
thyroid gland in a human fetus 158 mm. long (No. 49). The variable form and
mutual relations of the follicles are evident. The structure at the left of the
figure is a blood vessel. The oblique lines designate cut surfaces. 130.
438 EDGAR H. NORRIS
ingrowth of vascular connective tissue) into spheroidal seg-
ments, each of which becomes a small sac or follicle whose cavity
represents a portion of the originally continuous lumen of the
tube. This view has been advocated by W. Miiller (71), Mar-
shall (93 (in chick and frog), Streiff (97), Prenant (01), Hert-
wig (10), Prenant, Bouin and Maillard (11), Broman (11) and
others. Some like Streiff (97) and Simpson (712) have de-
scribed this branching tubular condition as persisting in part
throughout fetal life and even in postnatal life.
Other investigators, however, have described the lumina of
the primary thyroid follicles as appearing directly and _ ide-
pendently, with no preceding tubular stage. The anastomosing
solid cell-cords are usually described as becoming varicose, with
successive enlargements and constrictions, so as to present an
irregular beaded chain appearance. Sooner or later each of
these spheroidal swollen masses acquires a lumen and becomes
separated so as to form an independent follicle. This method
of follicle formation (with no tubular stage) has been described
by Tourneux and Verdun (97), Soulié and Verdun (97), Gros-
ser (12), Aschoff (13), Sobotta (15) and Kingsbury (715).
It is impossible to decide from direct observations of sections -
which of the preceding theories is correct. By reconstruction
methods, however, both graphic reconstruction and wax-plate
models, evidence has been secured in the present investigation
which definitely disproves the tubular theory and establishes in
the human thyroid the independent origin of the lumina of the
thyroid follicles. The follicles, however, appear not in epithe-
lial ‘cords’ as described by earlier observers, but in the fenes-
trated epithelial plates above mentioned. The view that the
thyroid is a modified branching tubular gland (Zielinska, Streiff,
Simpson, and others) therefore obtains no support from its
morphogenesis, aside from the initial stage of the primitive
diverticulum.
In the glands studied in the present series the first follicles
appear in a fetus of 24 mm. in length (No. 20). This is earlier
than the time of appearance described by most observers. His
(85), however, described follicles in a fetus (Zw) whose absolute
MORPHOGENESIS OF THE FOLLICLES 439
length is not recorded, but is placed in series between fetuses
of 16.5 and 22.0 mm. in length. Kingsbury (’15) described fol-
licles in a fetus of 832 mm.; Tourneux and Verdun (’97) in one of
32.4 mm.; and Grosser (712) and Sobotta (15) in fetuses of 50
mm. ‘These are the only cases found in the literature where
the presence of early follicles has been noted in fetuses of defi-
nite length. Several observers refer to the age of fetuses in
which the thyroid follicles appear, but in terms too indefinite
to be of value for comparison.
Although by definition the prefollicular period ends abruptly
with the appearance of the first follicles, it is not true that the
structure of the gland undergoes any corresponding sudden
change with the advent of the follicular period. The epithelial
bands (fenestrated plates) are only gradually replaced by the
primary follicles and structures characteristic of the prefollicular
period may be present through a considerable part of the fol-
licular period, at least until the fetus has attained a length of
65 mm.
Concerning the first three processes (rearrangement of the
cells, cell proliferation, and increase in the size of the cells) in-
volved in the development of the follicles from the epithelial
plate, no further discussion is necessary. The fourth process,
however, that of lumen formation, calls for further considera-
tion. The follicular lumen might arise in various ways, which
have been suggested by earlier investigators. Most of the
workers, however, do not mention the process by which the
lumen of the follicle is formed.
Wolfler (80) and Lustig (91) have described the formation
of lumina in the solid cell masses by a degeneration of the more
centrally placed cells. The present investigation does not sup-
port this view, however. In the first place none of the so-called
central cells have been found; and secondly no evidences of
degeneration have been observed in the primary follicles.
It might be supposed, as Hiirthle (94) and Anderson (’94)
have suggested, that the lumen is formed by the accumulation
of colloid between the angles of the cells which compose the solid
mass. Such a process would leave a colloid-containing space
440 EDGAR H. NORRIS
surrounded by epithelial cells. But in no case in the present
observations was ‘colloid’ found within the very early follicles;
although the accumulation of some other (precolloidal?) secre-
tory product between the angles of the cells, might result in
lumen formation. The fact is that the smaller lumina, which
are probably the most recent in formation, have been univer-
sally found to be devoid of any demonstrable content, and that
some of the larger and supposedly older lumina do contain a
stainable substance. As to the nature of the precolloidal sub-
stance or substances, nothing definite is known.
One might suggest (as thought by His?) that the lumen could
be formed by a degeneration or liquefaction of the central ends
of the cells which later form its outline. In this case it would
be expected that the early follicles would present a lumen out-
lined by an irregular or ragged margin. Without exception,
however, the lumina of the early follicles are clearly outlined
and marked off by a very sharp margin.
Having studied the way in which the follicle forms within the
plate, it is of further interest to determine the method by which
the follicle frees itself from the plate and comes to take up an
isolated existence. Harlier observers, in describing the forma-
tion of follicles from anastomosing rods or tubes, have laid em-
phasis upon the activity of the adjacent mesenchymal tissue as
the factor operating in the separation of the follicles. In the
present study no definite morphological evidence appears in
favor of this view. For as shown in figure 3 there is no special
differentiation of the mesenchyme or increased vascularity in
the regions in which follicles are being separated off from the
plate. The evidence seems rather to indicate that the follicles
themselves are the active agents in their separation. Thus as
certain of the cells leave their original positions to assume a
position more nearly in the center of the plate, the indentations
previously described appear on the plate. These may be con-
sidered as weak points. And as the cells increase in number
between these indentations it is not difficult to see how the in-
creased pressure due to the increased mass might force certain
follicles out from the row in which they formed and thus iso-
MORPHOGENESIS OF THE FOLLICLES 441
late them from the parent plate. The vascular mesenchyme
doubtless takes some slight part in the process, however.
As previously pointed out, the primary follicles may not in
every case be at once separated completely from the plate.
Instead of being in all cases sharply outlined and spheroidal in
form, small portions of the p'ate (epithelial tags) may be !eft
hanging to the follicle wall. The significance of such epithelial
tags is readily understood when it is considered how easily they
might be mistaken for epithelial buds arising from the primary
follicle in which secondary follicles were about to form. These
structures will be mentioned later in the discussion of the sec-
ondary follicles. It has been seen that there is no sharp line of
demarcation between the prefollicular and follicular periods.
Similarly, in the origin of follicles, the period of primary follicle
formation is not sharply marked off from that of secondary fol-
licle formation.
The number of thyroid follicles is apparently not absolutely
established or finally limited at any stage of development. (as is
the case, for instance, with the glomeruli of the kidneys). In
the earlier stages, the number of follicles increases by the forma-
tion of additional primary follicles. Soon, however, these pri-
mary follicles begin to give rise to secondary follicles (at 56 mm.).
In later stages the formation of primary follicles apparently
ceases, although their occurrence even in the adult has been
claimed, e.g., by Hirthle (94) and Sobotta (15), the new-
formed follicles being all secondary in character. Various
methods of secondary follicle formation have been described.
1. Origin from buds or sprouts. Ribbert (89), L. R. Miiller
(96), Streiff (97), Isenschmid (10) and others have described
this process. The bud is usually described as a local thickening
of the follicle wall, which continues to increase in size by the pro-
liferation of cells, until a solid bud, projecting into the stroma,
is formed. Directly, through the concentric rearrangement of
the cells, the form of the lumen can be made out, even while the
young follicle is still in contact with the mother-follicle.
2. Origin from collapsed follicles. Biondi (’89) Anderson (94),
and others have described the process as follows. After filling
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
442 EDGAR H. NORRIS
the vesicle discharges its contents, collapses and finally re-
arranges itself in the form of a number of small acini which
repeat the process.
3. Origin by fusion of follicles. Streiff (’97), v. Ebner (’02),
Isenschmid (710) and others have observed follicles which are
apparently formed by the secondary fusion of two or more
preexisting follicles.
4. Origin by division of follicles. Isenschmid (710) has de-
seribed the formation of daughter-follicles by the growth of an
epithelial spur across the lumen of the mother follicle; and
Peremeschko (’67) has noted the formation of secondary fol-
licles by the constriction of the parent follicle.
In discussing my observations concerning the formation of
secondary follicles in comparison with those of earlier observers,
it may be said at the outset that nothing to support either the
second or the third methods just outlined has been noted.
These, however, have been described chiefly by previous ob-
servers upon postnatal material. The other two methods
(budding and division of follicles), however, are in general
agreement with the findings of the present investigation.
There appear to be three general methods by which secondary
follicles arise in the fetal thyroid, the third of which might be
regarded as a modification of the second. But each of these
types is subject to a wide degree of variation, so that many
intermediate and modified forms are found.
1. Solid epithelial buds may develop on the follicle wall (figs.
4, 8 and 16d). These may become separated from the parent
follicle while in the solid state, or they may develop lumina while
connected with the wall of the mother-follicle, and subsequently
be constricted off. This method is essentially that advanced by
Ribbert (89), L. R. Miller (96), Streiff (97), and Isenschmid
(10). It is difficult, especially in the earlier stages, to distin-
guish these solid buds from the ‘epithelial tags’ representing
persistent portions of the original epithelial plates remaining
attached to the earlier primary follicles.
2. Hollow buds whose cavities are continuous with that of the
mother lumen are also found (figs. 4, 9, 10). It might be sug-
MORPHOGENESIS OF THE FOLLICLES 443
gested that these were originally solid buds whose lumen ap-
peared independently and later established a secondary con-
nection with the lumen of the parent follicle. While it is diffi-
cult to disprove such an occurrence, it would tend to reduce
-rather than to increase the number of follicles and seems im-
probable. The parent follicle apparently sends off extensions
or branches, which represent both the wall and the lumen and
eradually become constricted off to form new follicles. All pos-
sible stages in such a process, from the slightest outpouching of
the wall to the finally separated follicle are easily observable.
' 3. The third method of secondary follicle formation is the
simple division of the parent follicle somewhat as described by
Peremeschko. The process appears similar to that of hollow
bud formation. The follicle first takes on the form of an elong-
ated ellipse which becomes constricted about its center. No
cases of division by the ingrowth of epithelial spurs as de-
scribed by Isenschmid (for postnatal stages) have been observed
in my fetal material, though such may possibly occur.
Having considered the primary and secondary follicles sepa-
rately it remains to consider them in their relations to each
other and to the gland as a whole.
The formation of the secondary thyroid follicles begins when
the fetus has reached a length of about 56 mm. From the 70
mm. stage on the formation of secondary follicles progresses
with such rapidity that the total number of follicles is very
greatly increased within a short period of time, and the relative
number of primary follicles becomes progressively smaller.
It might be thought that such elongated, irregular, branched
and budding forms as appear in large numbers in certain of the
older members of the series (especially in the 86 mm. fetus, No.
43) are derived from irregular, branching bands (fenestrated
plates) retained from the prefollicular period. But the evi-
dence disproves that view. After the appearance of the first
follicles, the bands and plates are being progressively broken
up until the gland parenchyma in a fetus of 65 mm. is practi-
cally devoid of such structures, and is almost entirely made up
of isolated solid or hollow masses of cells. So closely does this
444 EDGAR H. NORRIS
resolution of the plates parallel the increasing length of the
embryo up to 65 mm. that it is very improbable that individual
variations could explain the presence of these structures in such
great numbers at the relatively late stage of 86 mm. Further,
as shown in figure 16, forms of all degrees of complexity may be.
found from simple spheroidal follicles to those of extreme com-
plexity. The conclusion is therefore reached that these branch-
ing structures, which are better described as follicular complexes
than as follicles, are developmentally only follicles which have
grown excessively and attained a high degree of complexity.
Such a rapid increase in the number of thyroid follicles as
occurs in fetuses between 65 and 158 mm. in length might be
expected to produce a marked increase in the size of the gland.
But according to Jackson (09) the growth curve for the pre-
natal thyroid gland shows no remarkable increase in the size
(weight) of the gland during this period. These two obser-
vations, which at first may appear contradictory, are readily
explained when the size of the follicles is taken into account.
The secondary follicles formed by budding and division of the
primary follicles are very small and arise from follicles which
are In most cases relatively of much greater size. So that while
the number of follicles is greatly increased during this period,
the gland mass is not correspondingly larger. As previously de-
scribed, the formation of secondary follicles becomes less rapid
before the fetus has reached 163 mm. in length. From this
point on, the number of thyroid follicles apparently increases
but slowly, the subsequent growth of the gland being due rather
to the increase in the size of the individual follicles than to a
further increase in their number.
The significance of the large cyst-like follicles described in
four of the fetuses is uncertain. Kiirsteiner (99) has described
the presence of similar follicles in four fetuses. The remarkable
regularity with which they were found, in his cases as well as in
those of the present series, located in the lower and posterior
(dorsal) part of the lateral lobe of the gland, is a striking fact.
Possibly they may be related to the cysts of the thyroid gland,
frequently met in pathological conditions of postnatal life.
MORPHOGENESIS OF THE FOLLICLES 445
VI. SUMMARY
By methods of reconstruction (both graphic and wax-plate),
the complicated process of morphogenesis of the follicles of the
prenatal human thyroid gland has been worked out and several
mooted points definitely established.
1. The so-called ‘cords’ forming the anastomosing network in
sections of the thyroid (median anlage) in the later prefollicular
stages represent chiefly sections of epithelial bands, two cells in
thickness, and forming irregular, fenestrated plates.
2. The frequently described stage in which the ‘cords’ are
transformed into an anastomosing set of epithelial tubes from
which the follicles are derived does not exist. The process of
follicle formation gives no evidence or indication that the thyroid
has been derived from a branching tubular gland.
3. The primary thyroid follicles arise directly as isolated and
independent structures from the epithelial plates of the pre-
follicular period, by the rearrangement of cells, cell proliferation,
increase in the size of the cells, and lumen formation.
4. The primary follicles appear in fetuses about 24 mm. in
length. The epithelial bands (fenestrated plates) have prac-
tically disappeared in a fetus of 65 mm., but a few solid inter-
follicular epithelial masses are still present in fetuses 163 mm.
in crown-rump length.
5. Secondary thyroid follicles are formed from preexisting
follicles apparently by three methods: by solid buds; by hollow
buds; and by constriction of the parent follicle.
6. The first secondary follicles appear in fetuses about 56 mm.
in length, but are formed most rapidly in stages when the fetus
is between 80 mm. and 158 mm. long. After 163 mm., the
growth of the gland probably takes place largely by the increase
in size of the individual follicles, rather than by increase in
their number.
7. Large cystic follicles were observed in the lower and pos-
terior (dorsal) parts of four glands from the older fetuses. Their
significance is uncertain, as is likewise the apparent involution
of the follicles with desquamation of epithelium observed in the
later fetal and newborn stages.
446 EDGAR H. NORRIS
‘VII. LITERATURE CITED
Anperson, O. A... 1894 Zur Kenntnis der Morphologie der Schilddriise.
Archiv f. Anat. u. Entw. S. 177.
Ascuorr, L. 1913 Schilddriise—Normale Anatomie. In Pathologische Ana-
tomie. 3 Aufl., Bd. 2, 8. 928-9.
Baser, C. E. 1881 Researches on the minute structure of the thyroid gland.
Philosophical Transactions, vol. 2.
BARDELEBEN, 8S. 1841 Observationes microscopicae de glandularum ductu
excretorio carentium structura deque earundem functionibus experi-
menta. Diss. inaug. Berlin. (Cited by Zeiss ’77.)
Bronp1, D. 1892 Contribution A l'étude de la glande thyroide. Arch. ital.
debiol:, “L. 17:
Bofcuat, P. A. 1873 Structure normale du corps thyroide. These, Paris.
Broman, J. 1911 Entwicklung der Schilddriise. In Normale und abnormale
Entwicklung des Menschen. S. 288-9.
CruvEILHIER, J. 1843 Glande Thyroide. In Traité d’Anatomie Descriptive.
2 Hd: -T.3. «Paris:
v. Esner, V. 1902 Schilddriise. In K6lliker’s Handbuch d. Gewebelehre.
Bd. 3.
Evxes, C. 1903 Der Bau der Schilddriise um die Zeit der Geburt. Dissert.,
Konigsberg. 1. Pr.
Grosser, O. 1912 The Development of the pharynx and of the organs of res-
piration. In Keibel and Mall’s Human Embryology, vol. 2, pp. 453,
468, 469.
Hertwic, O. 1910 Die Schilddriise. In Hertwig’s Lehrbuch der Entwick-
lungsgeschichte des Menschen und der Wirbeltiere. 9 Aufl., pp.
444-446.
HesseLBerc, C. 1910 Die menschliche Schilddriise in der f6talen Periode
und in den ersten 6 Lebensmonaten. Frankf. Zeitschr. f. Path.,
Bd. 5.
His, W. 1885 Anatomie menschlicher Embryonen. Bd. 3.
Hirzic, 1894 Beitrige zur Histologie und Histogenese der Struma. Diss.
Ziirich. (Cited by Streiff ’97.)
Horcicxa, J. 1880 Beitrige zur Entwicklungs- und Wachstumsgeschichte der
Schilddriise. Prager Zeitschr. f. Heilk., Bd. 1. (Cited by Elkes ’03.)
Hirruute, K. 1894 Beitrige zur Kenntnis des Secretionsvorgangs in der
Schilddriise. Pfliiger’s Archiv. f. d. ges. Physiol., Bd. 56.
IsenscHmMip, R. 1910 Zur Kenntnis der menschlichen Schilddriise im Kinde-
salter, mit besonderer Beriicksichtigung der Herkunft aus verschie-
denen Gegenden im Hinblick auf die endemische Struma. Frankf.
Zeitschr. f. Path., Bd. 5.
Jackson, C. M. 1909 On the prenatal growth of the human body and the
relative growth of the various organs and parts. Am. Jour. Anat.,
vol. 9, No. 1.
Jones, C. H. 1836 Thyroid Gland. In Todd’s Cyclopaedia of Anatomy and
Physiology, vol. 4, pp. 1102-1118.
MORPHOGENESIS OF THE FOLLICLES 447
Kinessury, B. F. 1914 On the so-called ultimobranchial body of the. mam-
malian embryo: man. Anat. Anz., Bd. 47, pp. 609-627.
1915 The development of the human pharynx. Am. Jour. Anat...
vol. 18, p. 329.
Kirsterner, W. 1899 Die Epithelkérperchen des Menschen in ihre Bezie-
hung zur Thyreoidea und Thymus. Anatomische Hefte, vol, 2.
Latourette. 1750 Recherches anatomiques sur la glande thyroide. Mé-
moires de mathematiques et de physique présentés 4 l’academie royale
des sciences (savants etangers). T. 1., p. 159. (Cited by Boéchat
Gal
Lustre, A. 1891 Contribution a la connaissance de l’histogenése de la glande
thyroide. Arch. ital. de Biol. T. 15.
MarsuHatut, A. M. 1893 Vertebrate embryology.
Miuuer, L. R. 1896 Beitrige zur Histologie der normalen und der erkrinkten
Schilddriise. Ziegler’s Beitrige, Bd. 19.
Miitter, W. 1871 Uber die Entwicklung der Schilddriise. Jena. Zeitschr. f.
Med. u. Naturw., Bd. 6.
PEREMESCHKO. 1867 Ein Beitrag zum Bau der Schilddriise. Zeitschr. f. wis-
sensch. Zool., Bd. 17.
Popackx, M. 1892 Beitrige zur Histologie und Funktion der Schilddriise.
Dissert., Konigsberg. i. Pr.
Prenant, A. 1901 in Poirier et Charpy, ‘‘Traité d’anatomie humaine,’’ T. 4,
p. 13.
PRENANT, Bourn ET MartuarD 1911 Traité d’histologie, T. 2, Histologie, p.
971.
Remax, R. 1855 Untersuchungen iiber die Entwickelung der Wirbelthiere.
Berlin. S. 122-123.
Rrpsert. 1889 Ueber die Regeneration des Schilddriisengewebes. Virchow’s
Archiv, Bd. 117.
Scurerper, L. 1898 Beitriige zur Kenntnis der Entwick!ung und des Baues’
der Glandulae parathyreoideae des Menschen. Archiv f. mikr. Anat.,
Bd. 52.
Simpson, B. T. 1912 Growth centers of the benign blastomata with especial
reference to thyroid and prostatic adenomata. Jour. of Medical
Research, vol. 22. :
Soporra, J. 1915 Anatomie der Schilddriise (Glandula thyreoidea). 29 Lie-
ferung des ‘‘Handbuchs der Anatomie des Menschen’”’ (von Bardele-
ben’s). Bd. 6, Abth. 3, Teil 4.
Soutif£ ET VerDUN. 1897 Sur les premiers développements de la glande thy-
roide, du thymus et des glandules satellites de la thyroide chez le
lapin et chez la taupe. Jour. del’Anat. et dela Phys. T. 23.
Stiepa, L. 1881 Untersuchungen iiber die Entwickelung der Glandula Thy-
mus, Glandula thyroidea und Glandula carotica. Leipzig.
Streirr, J. J. 1897 Ueber die Form der Schilddriisen Follikel des Menschen.
Archiv f. mikr. Anat., Bd. 48, 8. 579.
TOURNEUX ET VERDUN 1897 Sur les premiers développements de la thyroide,
du thymus et les glandules parathyroidiennes chez l’homme. Jour.
de l’Anat. et de la Phys. T. 23.
448 EDGAR H. NORRIS
Vircuow, R. 1863 Die krankhaften Geschwiilste. Bd. 3, 1 Halfte.*
Wotrier, A. 1880 Ueber die Entwicklung und den Bau der Schilddriise mit
Riicksicht auf die Entwicklung der Krépfe. Berlin. (Cited by Elkes
03.)
1883 Ueber die Entwickelung und den Bau des Kropfes. Archiv f.
klin. Chirurgie. Bd. 29.
Zeiss, O. 1877 Mikroskopische Untersuchungen iiber den Bau der Schild-
" driise. Dissert., Strassburg, 1877.
ZIEvINsKA. M. 1894 Beitrige zur Kenntniss der normalen und strumésen
Schilddriise des Menschen und des Hundes. Virchow’s Archiy.
Bd. 136.
THE EFFECTS OF LOW TEMPERATURE UPON THE
DEVELOPMENT OF FUNDULUS
A CONTRIBUTION TO THE THEORY OF TERATOGENY
WM. E. KELLICOTT
From The Marine Biological Laboratory, Woods Hole, and the Biological Labora-
tory, Goucher College
There have recently appeared several embryological contribu-
tions of importance, based upon the development of the familiar
minnow, Fundulus heteroclitus. Stockard (’09, ’10, ’13) and
more recently Werber (715, a) have described a large number of
abnormal and monstrous embryos resulting from chemical
treatment of the fertilized eggs and have made several sugges-
tions regarding the modus operandi of the treatments used and
the causes of monstrous development. And in the works of
Stockard (15) and Reagan (’15) advantage is taken of the pos-
sibility of causing, also by chemical treatment, certain embry-
onic abnormalities, in attacking a group of mooted questions
in normal embryogeny and histogenesis.
During the summer of 1915, at The Marine Biological Labora-
tory, I was able to carry out a rather extended series of experi-
ments on the production of monstrosities in Fundulus by a
method which involved no direct chemical stimulation and which
thus permits me to test certain previously suggested causes of
abnormal development in the species studied. The method is
so simple and so markedly effective that it seems desirable to
make a brief report upon it and to point out its bearings upon
the general theories of teratogeny, although at this time I shall
not attempt to give an extended analysis either of the precise
results obtained or of the exact mechanism of the disturbance of
normal development.
449
450 WM. E. KELLICOTT
In a word the method consists simply in subjecting the fer-
tilized ova or young embryos to the temperatures of the ordinary
household refrigerator, namely, 8-13°C. for a few hours or days.
In connection with the work on Teleosts in the Embryology
Course at The Marine Biological Laboratory, we have for several
seasons made a practice of placing the eggs, at various stages of
development, in the refrigerator in order to slow their develop-
ment and thus secure certain stages at hours convenient for
study. It was noticed that a slight reduction in temperature
merely slowed development without leading to any serious dis- ’
turbance in morphogenesis, while a few hours at a temperature
below 12-14°C. were followed, upon returning the eggs to the
laboratory, by a considerable mortality and by very frequent
abnormality (the eggs of Tautogolabrus (syn. Ctenolabrus), not
Fundulus, were then used). In fact abnormalities in nuclear
and cytoplasmic arrangements were to be noted while the eggs
were still at the lower temperature. It was not until last sum-
mer, however, that the more precise examination of the results
of this treatment was undertaken. While these experiments
were in progress Loeb (715) reported certain observations on the
effects of much lower temperatures upon the development of
Fundulus. His results will be discussed and compared with my
own later in this paper. My apparent neglect to test certain
points of disagreement is due merely to the fact that his paper
did not come to my attention until some time after the spawn-
ing season of Fundulus was entirely past, which made such an
attempt impossible.
The eggs were inseminated in the usual manner, 1.e., dry,
with the addition of testis teased in only a few drops of sea-
water. They were placed in finger bowls with a half-inch or so
of water; the bowls were covered to prevent evaporation, and
placed on the shelves of the refrigerator. The range of tem-
peratures used was secured merely by using higher or lower
shelves, in compartments alongside or below the ice compart-
ment. The water was renewed daily, or in some instances
every other day, the fresh water being at the refrigerator tem-
perature. That covering the bowls with glass plates did not
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 451
materially affect the oxygenation of the eggs is sufficiently in-
dicated by the normal development of similarly handled con-
trols in the room temperature. The possible effect of dark-
ness in the refrigerator was controlled in only one instance,
when no effects were observed.
If the temperatures used be not too low (11—-13°C.) the only
marked immediate effect seems to be the retardation of normal
development, which is resumed at its usual rate when the ova are
returned to the ordinary temperatures of the laboratory (20—
— 24°C.). It is a simple matter thus to prolong the age of the
early blastula from the normal average of about five hours, to
more than two weeks, that is some sixty-five times, and prob-
ably it might be prolonged much more. Such blastulas, re-
placed in the room, may develop and hatch quite normally, or
they may form embryos showing marked abnormalities, indi-
eating that underlying their normal appearance earlier there was
an actual disturbance of some kind. One ean not.always say
by superficially inspecting the entire blastula, or later stage,
whether it is completely normal or not.
Subjection to a lower temperature (8—10°C.) for a few hours
or days, is usually followed either by death or by abnormal
development, normal development after such treatment being
found in only a small percentage of the total treated. It should
be noted, however, as a point of some importance, that in every
experiment where development continued at all, after removal
from 'the refrigerator, at least a small percentage of the embryos
developed normally. ,
Considerable variation was observed among different lots of
eggs, some showing few abnormalities after a treatment that re-
sulted in very few normally hatching larve from other lots of
eggs. It should also be noted that in one, and in only one, con-
trol, did the total proportion of abnormal embryos equal the
minimum observed after subjection to a low temperature for
more than five hours; and in no other case did the abnormalities
in the control even approach the minimum in treated lots.
Eggs placed in the refrigerator within two to five minutes after
imsemination usually proceed to form a.germ-dise which differs
452 WM. E. KELLICOTT
from the normal in being smaller, that is in containing less pro-
toplasm, and in being markedly more convex. ‘This increase in
convexity may simply indicate a decrease in surface tension di-
rectly due to the lower temperature. If the temperature be not
lower than about 12°C. cleavage continues slowly and with little
or no apparent abnormality, save that the blastomeres, like the
germ-disecs, are smaller and unusually convex, and therefore
more than normally separated from one another.‘ Stages of
eight to sixteen cells may be found twenty-four to thirty-six
hours after fertilization.
At a temperature of about 8-10°C. the effects are quite varied.
‘In the first place, in a considerable and widely variable propor-
tion of the eggs no true germ-disec is formed, cleavage rarely
occurs, and after a time the protoplasmic parts become wholly
vacuolated, giving no sign of living processes, either during con-
tinued treatment to cold or after their removal to a higher tem-
perature. Other eggs, however, are not killed by this treat-
ment; a few may form small germ-dises and cleave regularly
once or twice during twenty-seven to thirty hours. Or cleavage
may become very irregular, blastomeres later showing wide differ-
ences in size and extensive divergences from the typical arrange-
ment. Many instances were noted where the germ-disc was
very imperfectly formed, the thin protoplasmic cap remaining
spread over one-fourth to one-third of the egg, with occasional
mounds, cell-like though without cell-walls, scattered irregu-
larly over it. In still other instances there were no true cellular
structures whatever present, even after several days, and aggre-
gations of nuclear substance might be seen scattered promiscu-
ously through a cytoplasmic mass of irregular form. Several
eggs were noted in which the protoplasm had collected in sev-
eral distinct and widely separated regions. After several days
all these eggs showed at least two kinds of material derived from
the cytoplasmic part of the egg; one was dark and granular,
quite opaque but still apparently active, the other clear and
vacuolated and apparently no longer living.
It should be noted that throughout this paper the appearances
described are those given by microscopic examination of total,
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 453
unfixed eggs. The results of the detailed study of sections of
such eggs, which will perhaps enable me to determine more pre-
cisely the effects of the cold, will be reported separately at a
later time.
Upon removal to room temperature, after having been some
hours or days in the condition just described, many eggs failed to
develop, and after some hours longer died without undergoing
any apparent structural changes. On the other hand, most of
those eggs in which there were present considerable masses of
the darker granular material showed some processes of develop-
ment, exhibiting before their death every degree and form of
abnormality, from irregular protoplasmic masses, suggesting
cellular structures, to hatched larve, well-formed though usually
abnormal in some respect. In a very few instances (three were
noted) normal larve resulted from the development of these
dark, irregular, non-cellular masses.
When the eggs were allowed to develop normally for a few
minutes (fifteen to thirty) before refrigeration, the general re-
sults did not seem to be markedly different, although there are
some indications that fewer embryos developed normally.
Two lots of eggs were allowed to develop normally for twenty-
two to twenty-three hours after insemination before they were
placed in the refrigerator at about 11°C. At this age and at
room temperature, the germ-ring is formed and just commencing
its extension around the yolk. From a fourth to a third of these
were dead after twelve hours in the refrigerator. Most of the
remainder, after eleven days in the cold were alive, and upon
transference to the room developed, some into normally hatching
larvee, others into larvee with various defects and abnormalities,
some very pronounced. Upon removal after twenty-one days
many died and nearly all of the survivors developed very ab-
normally; but two hatched normally after ten to eleven days in
the laboratory, i.e., thirty-two to thirty-three days after fer-
tilization.
One lot of eggs was allowed to develop for forty-three hours
before refrigeration at about 10°C. At this age the germ-ring is
just closing, the embryo is well established and the optic vesicles
454 WM. E. KELLICOTT
in some instances just forming, in others clearly formed. After
eight and one-fourth days at 8-10°C. these were removed to the
room; nearly all were alive but development had continued very
slowly so that none showed beating hearts (normally the heart
begins to pulsate about twenty-four hours after the establish-
ment of the optic vesicles and the closure of the germ-ring).
Some were already abnormal at the time of their removal, many
others became abnormal as development in the room continued,
but about two-thirds hatched normally after eight to eleven days
longer.
It is interesting to compare certain of these observations with
those of Loeb (15) who found that treatment with a tempera-
ture somewhat lower than any I used, namely 7°C., was followed
by no abnormalities in development. At that temperature
Loeb found that ‘‘the newly fertilized eggs can live for weeks
without being injured” and that they ‘‘developed
very slowly but no abnormal embryos were observed, although
some of the eggs were kept at a temperature of 7°C. for four
weeks” (p. 62). Loeb also used temperatures much lower than
any with which I worked and found that even at 0—2°C., if the
treatment were not prolonged more than four to seven and one-
half hours, and if stages from the time of insemination to about
four cells were used, only 20-30 per cent became abnormal upon
transference to the room; longer treatment increased the mor-
tality considerably, none surviving forty-eight hours at this
temperature. If the eggs were first allowed to develop fifteen
hours at a normal temperature (stage given as about one hun-
dred and twenty-eight cells, although at the usual laboratory
temperature of 22°C. this stage is reached about five hours after
insemination) Loeb found that after two days at 0-2°C. normal
development was still possible, although he does not state that
abnormal development did not also occur in some individuals.
But he found that if the embryo is once formed before treat-
ment, it survives weeks of subjection to 0-2°C. “‘without any
injury.” ‘‘As soon as it is put back to room temperature it
continues to develop” (p. 59).
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 455
I am entirely at a loss to interpret the difference between
Loeb’s observations and my own on this point. It is just pos-
sible, referring to the latter instances mentioned above, that at
a temperature of only 0—2°C. the living processes of the embryo
are so completely stopped that when development is resumed at
a higher temperature no disturbing effects are to be seen; in
other words not even abnormal development may proceed at so
low a temperature. But this possibility seems largely negatived
by the observation that extremely harmful-results follow simi-
lar treatment of uncleaved or four cell stages. And no such
interpretation seems possible for his results at 7°C., which should
be more nearly comparable with my own at 8-10°C. I ob-
served no instance of normal development occurring after eggs
had been three weeks at 8-10°C., although several lots of eggs
were tested and among the same lots normal development some-
times followed briefer treatment; and abnormalities were very
common after treatment for only twenty-four hours, in some lots
even for only five and one-half hours. ‘Two lots of eggs, treated
in just the same manner as the others, showed no abnormalities
upon development after twelve hours at 8-10°C. It seems quite
unlikely, however, that all of Loeb’s material could have been
so unusually resistant, in view of the rarity with which such lots
came under my observation. J may add in passing that Wer-
ber (15, b) notes that he too was able to secure abnormal em-
bryos by treatment with temperatures much higher than those
~ used by Loeb, but he gives no details either of his experiments
or results.
It would not be profitable to attempt to describe in detail the
abnormalities observed, nor even to enumerate all of them. In
general it may be said that the embryos resulting from this
treatment showed every degree of abnormality, from irregular
masses of protoplasm, alive though exhibiting few of the char-
acteristics of organisms, up to completely normal hatched
larve. Among several hundred embryos examined there was
found almost every conceivable kind of disturbance. Every
characteristic that could be observed externally with the lower
powers of the microscope showed some degree of abnormality in
456 WM. E. KELLICOTT
some embryo. Certain organs, such as those of the circulatory
system—the heart, vessels and blood-cells—and the eyes, were
especially subject to abnormality, others such as the ears and
covering ectoderm were rarely affected; yet no structure was
found which was not affected in some degree in some embryo.
By way of general support for this statement. I may mention
just a few of the observed abnormalities, without mentioning
the details of the treatment in individual instances.
In those cases where no definite embryo might be said to
have formed, such conditions as the following were noted: form-
less, non-cellular but ‘living’ protoplasmic masses; protoplasmic
masses with varying degrees of cellular structure; irregular pro-
toplasmic masses (probably cellular) showing suggestions of
organs, such as brain fragments, lenses, portions of optic cups,
groups of somites, masses of erythrocytes, rhythmically contrac-
tile cells arranged either as flat sheets or as tubular ‘hearts,’
scattered pigment cells of the usual types, endothelial cells over
the surface of the yolk, fragments of notochordal tissue.
Among those cases where a more or less complete embryo was
formed (connected with the preceding condition by various gra-
dations) a few of the abnormalities noted were the following:
two separate and complete embryos on a single yolk; absence of
head; absence of tail; large malformed head; short stumpy tail;
sinuous body and tail; short deep body and tail; malformed
pectoral fins; localized ectodermal proliferations; some regions of
the brain not closed dorsally; absence or hypertrophy of various
regions of the brain; various degrees of anterior approximation
and fusion of the eyes; ventral fusion of the eyes; absence of one
eye, the other remaining normal in size and position; eyes of
unequal size; optic cups not closed; various degrees of albinism
never quite complete; absence of certain types of pigment cells; .
atypical concentrations of pigment cells; greatly dilated peri-
cardial cavity; heart abnormally placed, posterior or lateral to
its normal location; two hearts dissimilar in size and form,
asymmetrically located and with different rates of contraction;
absence of heart; heart short and ‘telescoped;’ heart elongated
and thread-like; heart long and dilated; heart flat and’ plate-
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 457
like; heart undifferentiated into chambers; only the sinus end of
heart contractile; well-marked rhythms in rate of contraction of
heart, sometimes occasionally stopping; no circulation, either
on account of absence, or proximal or distal closure of the heart;
few vessels over the yolk; yolk-vessels abnormally arranged;
heart thread-like but with a good and complete circulation;
very few erythrocytes but abundant plasma circulating freely;
erythrocytes in masses on postero-dorsal surface of yolk; large
masses of erythrocytes antero-ventral to heart or along the an-
terior margin of pericardial cavity (after hatching this latter
mass was in a median ventral position); dense mass of erythro-
cytes collected in tail, 7.e., caudal aorta and vein anastomosing
at base of tail. ;
Many more abnormalities might be mentioned and of course
innumerable minor details of abnormal character might be cited
from the material. What variety of conditions might be re-
vealed by the thorough study of sections of these embryos can
only be imagined.
One further point might be mentioned. Stockard observes
(15, p. 26) that embryos developing without a circulation are not
able actually to hatch. Asa rule I had a similar experience but
one exception was so remarkable that it seems worth noting.
After eleven days at 10-11°C., one lot of eggs which had devel-
oped in the laboratory for twenty-three hours before cooling,
showed many normal embryos, for the most part at a stage
when the optic vesicles were just forming. Others were grossly
abnormal or entirely dead. Among those appearing normal
upon removal from the cold, many abnormalities appeared later
and one of these, which was found hatched twelve days after
removal, or twenty-four after fertilization, had no circulation
whatever, although the heart was present and pulsating weakly.
This, however, was not its only defect. It was below average
size, had but a single median anterior eye and correspondingly
the nose and upper jaw region were, as usual in cases of cyclopia,
narrow and elongated, the tail was short and ended in an un-
differentiated mass, pigment cells were largely massed on the
ventral surface of the yolk. It seemed unable to direct its swim-
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
458 WM. E. KELLICOTT
ming movements (it was probably blind) and both in swimming
and in resting it turned on either side instead of maintaining a
normal position. This fish hatched sometime during the twenty-
eight hours preceding the notice of its having hatched, and after
living in this condition for thirty hours longer, its heart seemed
somewhat more feeble, its fin motions became slower and it was
killed.
That fish lacking so physiologically important a process ‘as
the circulation may develop to the point of hatching is alone a
most remarkable fact. But that such an organism can actually
hatch from its egg-membrane and, in spite of the other abnor-
malities noted, remain alive and active for between thirty and
fifty-eight hours, can not fail to arouse many questions regarding
the general problem of embryonic adaptations. [Hf a fish can
live for weeks, and then hatch and remain active for hours longer,
entirely lacking a whole physiological system commonly re-
garded as so essential, one may certainly be permitted to doubt
whether slight details in the arrangement of this or of other
parts, may have the functional importance often assumed.
In connection with the observation of the irregular, only par-
tially cellular protoplasmic masses mentioned above, it is inter-
esting to notice that we have here an illustration of the fact
that the protoplasmic substance of a highly specialized form may
still be capable of existence as protoplasm, though not showing
any of the normal morphogenetic processes characteristic of the
living substance at that age; and may remain able for days and
weeks to carry on some of its vital processes. Yet in spite of the
fact that these protoplasmic masses were so widely aberrant in
form and appearance, exhibiting none of the usual morphologi-
eal characteristics of the protoplasmic portions of fish-eggs, nor
of any other kind of vertebrate egg, save the primary differen-
tiation into nuclear and cytoplasmic materials, they were still
undoubtedly living. They remained free from bacterial infec-
tion, form and appearance slowly changed, coagulation did not
occur, 1 many instances for weeks. Moreover it is not true
that in all such cases death was the inevitable result, for when
this condition was not too long maintained, such irregular,
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 459
largely though not wholly non-cellular masses occasionally gave
rise to true embryos exhibiting the usual developmental proc-
esses. Such embryos usually showed various degrees and kinds
of abnormality, but some few ultimately produced normally
hatching larve. These masses were therefore more than merely
living; their later history shows that even in that state there
must have remained present, in many instances, some kind or
remnant of an underlying construction or organization that
determined either directly or through some regulatory process,
development of parts, at least, of Teleost embryos, and in a few
instances of essentially normal Fundulus embryos.
My chief object in describing, at this time, the effects of low
temperature upon the development of Fundulus is not to give
a morphological or histological description of the malforma-
tions and abnormalities produced, but to suggest the bearings of
these results upon certain current hypotheses as to the real
causes of such defects and the way in which the unusual physi-
cal conditions may have affected the morphogenetic properties of
the ovum. I shall not attempt to review extensively the various
suggestions as to the causes of monstrous development, but
some of the more recent hypotheses only may be examined from
the point of view of the results described here. For this pur-
pose I shall refer chiefly to the suggestions made by three recent
workers in this field. .
As a preliminary word I should note that such observations as
these show that ‘“‘the idea that the low temperature only re-
tards the chemical reactions underlying development,’ (Loeb
"15, p. 59) is, for the Fundulus egg, true only to a certain point.-
When temperatures are lowered below say 12—14° C. the orderly
developmental processes are not only slowed but may be actu-
ally modified so that some of the consequent processes are ren-
dered abnormal. This may, of course, be due primarily to the
fact that certain processes are slowed more than others, but the
result is a disturbance of normal development which is different
from a mere slowing down of the entire mechanism, the com-
ponent processes remaining normally related to one another.
Apparently the precise temperature when this effect is produced
460 WM. E. KELLICOTT
varies with same internal or fluctuating condition of the egg and
also with its stage of development.
We should also point cut here that Lillie and Knowlton (97)
found that eggs of the frog (Rana virescens) at 2-3° C. always
developed abnormally, if at all, the abnormalities usually appear-
ing in the region of the blastopore.
In his classic ‘“‘Study of the Causes Underlying the Origin of
Humen Monsters,’ Mall (08) included an extensive survey of
the more important results in the whole field of experimental
teratology, particularly among the vertebrates. He concludes
that even in such dissimilar instances of teratogeny as those in
the toad resulting from fertilization with X-rayed spermatozoa
(Bardeen, ’07), those in the Teleost following chemical treat-
ment during cleavage (Loeb, 798, Stockard, ’06, ’07) and the
human instances described by himself, “although the methods
employed are very different, the principle involved and the
results obtained are much the same”’ (p. 24). And, ‘‘In general
the methods employed by experimental teratologists is to sub-
ject the eggs to various insults which affect the nutrition and
impair the growth of the embryo” (p. 52). “A monster is due to
the influence of external substances which retard the growth of
the embryo, usually one portion more than the other” (p. 36).
In brief Mall’s conclusion regarding the causes of human mon-
sters is that ‘faulty implantation of the ovum, which naturally
affects the growth of the embryo” (p. 25), by interfering with
the normal nutritional relations leads to monstrous develop-
ment; ‘“‘that certain parts of the embryo are more susceptible to
insults than others” (p. 32), and consequently it is these that
are affected first (p. 16), though subsequently the faulty im-
plantation must be remedied so that the embryo may continue
to grow (p. 25). ‘But in order to produce a finished monster
the nutrition must not be impaired too much” (p. 31). Many
other quotations from the same work might be added to show
that the words nutrition and growth are used with their custom-
ary significance, implying only the supply of materials and en-
ergy necessary for the usual processes of extension of at least
partially differentiated’ stiuctures or anlagen. Disturbances of
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 461
the general chemical reactions underlying development but not
directly related to the income of the embryo, such as nuclear
differentiation, mitosis, distribution of nuclear and cytoplasmic
substances, specification of protoplasmic areas, et cetera, are
evidently not included under the terms nutrition and growth as
used here.
Kssentially the same idea is contained in Stockard’s hy-
pothesis of the causes of the defects observed in Fundulus after
chemical treatment, namely, that such treatment tends to lower
the developmental vigor of the embryo, this lack of vigor being
subsequently shown by the failure of certain structures to de-
velop normally. This hypothesis is formulated chiefly in con-
nection with the explanation of optic defects, particularly cy-
clopia. “A certain amount of energy is necessary for differen-
tiation of the eye to take place . . . . but,when the re-
quired energy for any reason is not available the eyes are inca-
pable of any differentiation” (13 b, p. 271). In the absence of
any other evidence for such a deficiency in available energy on
the part of the embryo, such an explanation comes to little
more than a statement that the eyes do not develop. I can not .
be certain whether or not it is intended that this explanation
should be applied to the other abnormalities observed by him,
regarding which, however, no other suggestion is made. But
as I shall show there are some difficulties in making a general
application of this idea in explaining all of the defects noted in
his experiments with various chemical substances.
Using a butyric acid or acetone treatment, Werber (’15 a,
16) secured not only abnormalities similar to those recorded by
Stockard, but also a wide range of the same general types, and
in many instances the same specific types, as those which I have
found to follow the action of a low temperature. Werber be-
lieves that there are some factors common to the morphogenesis
of all the diverse abnormalities and monstrosities observed in a
given experiment. He applies the term ‘blastolysis’ collectively
to the factors involved and believes that there occurs, as a con-
sequence, a destruction or dispersal of parts or all of the germ’s
substance.
462 WM. E. KELLICOTT
My own findings reported here certainly tend strongly to
confirm the idea that the causes underlying abnormal develop-
ment are to be sought in some derangement of the fundamental
developmental mechanism or relations of the diverse compo-
nents of the very early stages of the ofganism. In the first
place it should be noted that by using a low temperature as the
abnormal stimulus I have eliminated the necessity of supposing
that there are any specific chemical alterations, such as precipi-
tating or solvent effects, due to the use of chemical stimuli,
which is so important an element of Werber’s hypothesis.
There may be some osmotic changes in the cooled eggs and there
certainly is an increase in oxygen tension, but McClendon (712)
was unable to find that such conditions affected the frequency
of eyelopia in Fundulus. On the other hand some of the effects
of cold may be directly observed in the eggs during treatment,
so that the character of the disturbance is not wholly left to be
inferred from the study of later development.
As mentioned above, eggs which were placed in the refrigera-
tor within a few minutes after insemination were found, some
- days or weeks later, whether a few cleavages had appeared mean-
while or not, to contain at least three classes of materials. 1) Not
only nuclei of the usual appearance but also irregular, large and
small masses and fragments of nuclear material scattered indis-
criminately through the cytoplasmic parts of the egg. There
were ordinarily no cellular outlines corresponding with these
masses of nuclear substance. 2) Masses of granular cytoplas-
mic substance somewhat resembling the material of the greater
part of the normal uncleaved germ-dise. 3) Masses of clearer,
possibly protoplasmic substance usually in the form of vacuoles,
slightly resembling in physical appearance that mass of clear
cytoplasm that in normal development forms, for a brief period
preceding cleavage, a lens-shaped disc on the lower side of the
central part of the germ-disec. I am not yet in position to iden-
tify these forms of cytoplasmic substance with the two chief
forms observable in normal development; but the similarity is
suggestive. Eggs in this state are still capable of some sort of
developmental process when returned to ordinary temperatures,
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 463
but abnormalities are very frequent and show the remarkable
range of characters described above.
It seems, then, that the causes of the observed abnormalities
in development may be referred to these conditions: (a) abnor-
malities in nuclear composition and in the distribution of dif-
ferentiated nuclear substances; (b) abnormal distributions and as-
sociations of two more evidently differentiated cytoplasmic sub-
stances; (c) abnormal associations between nuclear and cyto-
plasmic substances, either or both of which may be abnormal
in its own composition as compared with the corresponding
materials in the regularly developing egg.
In a word this means that the organization of any part of the
early organism may be disturbed, and that such disturbances
are the causes of abnormal and monstrous development. That
development can occur at all under such circumstances indicates
remarkable regulatory properties of the egg substance. The
complete lack of specificity in the effects of the cold indicates
that the disturbance is profound and that it affects the funda-
mental organization of the ovum rather than any especially dif-
ferentiated representative substances or anlagen, which, more-
over, have been shown by other evidence not to exist in the
Teleost ovum (Morgan, 793, Sumner, ’04).
Going a step back of these observations to the query as to
just how the low temperature can produce such disturbances, we
can at this time make suggestions only in very general terms.
It is evident that not all of the processes of development or
parts of the developmental mechanism are affected similarly,
for that would lead to development normal in all save its rate.
Some physical or chemical processes or structures must be more
extensively interrupted or altered than others, as a result of the
lowered temperature. They are thrown out of their normal re-
lations to the other processes or parts of the mechanism and
sooner or later a whole train of consequences may become evi-
dent. It is very important to note, however, that the substances
or processes thus affected are not to be thought of as specific
tissue—or ‘organ-forming substances’ nor as differentiated, cellu-
lar or formed rudiments or anlagen, but as elements or factors
464 WM. E. KELLICOTT
in that whole complex mechanism which as an entire system
epigenetically gives rise to such substances or rudiments. It is
quite likely that, in part at least, these altered reactions are such
as may lead to disturbances of the mitotic processes of matura-
tion and fertilization, such as were described in such great
variety in Ascaris by Sala (95) or of nuclear division during
cleavage, as described by Conklin (12) in Crepidula, whereby
abnormal nuclear structures may be formed. It is also not
unlikely that the observed failure to form cell-walls in many
parts of the cytoplasm, or the abnormal location of cell-walls
when formed, may open the way to the possibility of abnormal
nuclear and cytoplasmic associations. Another possibility lies
in the physical slowing of the translocatory movements of the
eradually differentiating cytoplasmic materials. In eggs sub-
jected to cold immediately after fertilization the normal flow-
ing together of the cytoplasm to form the germ-dise and the sub-
sequent rearrangement and redistribution of different cyto-
plasmic substances, often appear superficially to be seriously
interrupted. But all such conditions are themselves to be re-
garded as consequences of antecedent modifications of some
more elementary chemical or physical organizational processes
es yet beyond analysis and descriptign.
The great variety in the results following treatment is Just
what would be expected on the basis of such a disorganizational
effect. Both nuclear and cytoplasmic materials are themselves
so complex, and the complete system of relations, both ma-
terial and energetic, between them and among the various
component parts of each, which as a whole we term the ‘organ-
ization’ of the ovum, is so extremely complicated, that it affords
almost unlimited possibilities for modification and disturbance.
Without knowing very much more than we do about the physics
and chemistry of this organization and about its regulatory
capacity, it would seem largely or wholly a matter of chance,
what would be the precise later results of any single modifica-
tion or simple group of disturbances in this system. The results
of such disturbance are entirely unpredictable in individual
instances at present.
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 465
We should recognize, as preliminary to much of the discussion
that follows, that either the morphological or the chemical ex-
tent of the initial disturbances may not be the only condition
correlated with the extent of the later derangements. A com-
paratively slight modification in an essential or highly impor-
tant organizational factor would have a more marked result in
later development, than would a more extensive alteration in
factors of lesser importance. The quality of the disturbance
rather than its extent is primarily involved. It is further likely
that no relation could be determined between the extent of the
initial disturbance and the final effect on account of the possi-
bility of regulatory action and because comparatively slight
initial alterations might give a wholly abnormal trend to a long
series of consequent processes finally resulting in very pro-
nounced abnormalities.
With this general statement of the essential nature of the
disorganization hypothesis we may turn to an examination of
the widely current nutrition hypothesis, based largely upon the
results following chemical treatment, in order to draw attention
to certain difficulties in its application to some of the observed
facts and to inquire whether the suggestions made here avoid
any of these difficulties without creating others.
If any justification seems necessary for the attempt to criti-
cize, from the viewpoint of the results of the action of low tem-
peratures, the hypothesis based upon chemically prod ced
effects, it is to be found in the essential identity of the conse-
quences of these different modes of treatment. Indeed I should
go further and bring under this same point of view the abnor-
mal types of development following certain other experimental
conditions, such as heterogeneous hybridization, and the sub-
jection of gametes or zygote to radium radiations, which will be
discussed later.
The nutrition hypothesis as stated by Mall, and to some extent
adopted by Stockard, has for some time now largely been held
to account for such illustrations of teratogenesis as those de-
scribed here, but it seems open to certain serious objections.
In the first place it might be pointed out that the work of Pack-
466 WM. E. KELLICOTT
ard (’14) on the effects of radium radiation upon the eggs and
sperm of Nereis, has largely rendered untenable the suggestion
of Mall that the effects’ of such treatment are nutritional in
character. Packard concludes upon very clear evidence that
there are strong reasons for believing that the radium radia-
tions act indirectly upon both the chromatin and cytoplasm of
either or both germ cells or the zygote, by bringing about in
these, destructive chemical processes. Development is thus ren-
dered abnormal both by the destruction of the normal chemical
and physical mechanism of early development, and also possibly
by the toxic presence of the abnormal substances thus formed and
present in the cleaving egg. That is, the abnormalities found in
later development may be referred to abnormal nuclear and ecy-
toplasmic behavior during the cleavage processes, before the
specific germ-layer or tissue differentiations are inaugurated.
This clearly removes the results from the category of nutri-
tional effects and affords an explanation of the same general
character as that which I have suggested above, with this dif-
ference, however, that I am not inclined to stress the possibility
that the effects are due to the presence or absence of specific
chemical-substances, but rather hold them to be due to unusual
combinations of differentiated materials both nuclear and cyto-
plasmic, in a word to a disturbance of the ‘organization’ of the
ovum or of certain parts of it; and this difference seems to me
of quite an essential character.
How such results as those of Lewis (’09) and Spemann (’03,
04) and some of Stockard’s (13 b) who caused eyelopia in Fun-
dulus and Amblystoma by the actual physical destruction or
even removal of the differentiated anterior end of the central
nervous system, can be interpreted as due to a nutritional effect
is even less clear than the possibility of such an interpretation
of the effects of radium upon the spermatozoa previous to fer-
tilization. ‘To compare and identify the cyclopia in Fundulus
resulting from such removals of already differentiating struc-
tures, with the cyclopia resulting from chemical treatment of
four to eight cell stages, involves the assumption that during
early cleavage in Fundulus there is already present some repre-
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 467
sentative, definitive, organized rudiment of the anterior tip of
the central nervous system, and further that it possesses a sensi-
tivity to general nutritional disturbances which is also specific,
i.e., not shared equally by other rudiments which must also be
assumed to be present at this time. But it is well known that
the Teleost ovum is of the indeterminate type, and injury or
removal of whole blastomeres of the normal cleavage group
causes no later defects or abnormalities in the developing em-
bryo. The assumption of specificity in the action of nutritional
effects in connection with cyclopia, is further negatived by the
great variety of the results following chemical treatment. I
have already pointed out the great variety of abnormalities and
monstrosities observed by Stockard, Werber, and others, appear-
ing under the same experimental conditions. ‘The only sugges-
tion of actual specificity is that of Stockard (10, p. 369) that
treatment with certain percentages of alcohol gives, among the
surviving embryos 90-98 per cent with abnormal eyes, generally
eyclopean. But he does not say that other defects may not also
be found in these or other embryos similarly treated, indeed on
the contrary he mentions a great variety of ear, brain and other
defects which may either accompany eye-defects or appear in-
dependently of them, under similar treatment. Moreover, the
expression ‘eye defects’ covers a number of different conditions,
not all of which can be referred to the cause which he assumes
for eyclopia, namely, the inhibition of the development of the
median anterior tip of the central nervous system. To find a
great variety of abnormalities among embryos subjected to the
same treatment leads definitely to giving up the idea of specific
reactions of the rudiments of such structures to the unusual con-
ditions. To relate the appearance of such a variety of abnor-
malities during later development of the embryo, to the effects
of the treatment upon some specific parts of the materials of the
cleavage group is, therefore, to assume the existence of differ-
entiations during cleavage which have been shown not to exist,
and then to require further that each or any.or all of these may
be specifically affected, which is to say that there is no speci-
ficity at all in the action or in the reaction.
468 WM. E. KELLICOTT
Stockard (13 b) explains “the fact that a number of egys
when subjected to the same solution do not all respond in a like
manner” merely by regarding this as ‘“‘a typical case of differ-
ences in individual resistance and vigor which is observed among
any one hundred individuals of any living species” (p. 282).
But the kinds of differences found here are not at all such that
they may be regarded as illustrations of that normal fluctuation
in all characteristics which represents the reactions of organisms
to the incidence of environing conditions. It is probably true
that eggs and embryos do differ in those complex conditions
which we summarize in the words ‘resistance and vigor.’ But
these qualities do not determine whether an embryo shall or shall
not have eyes, hearts, pigment cells, and so forth. We are
dealing with a wholly and fundamentally different phenomenon.
An important objection to this nutrition hypothesis, it seems
to me, is that there is little or no actual evidence given that the
nutritional conditions of the egg or embryo are directly affected.
The evidence offered is that: defective and monstrous embryos
result from various methods of chemical treatment; but it is
these defective and monstrous embryos that are to be explained.
In the few instances where there is evidence of abnormal rela-
tions between embryonic and vitelline portions of the egg, it is
more reasonable, in view of the great variety of other conditions
found, to regard this too as a result of a primary disturbance,
rather than as the cause of a variety of conditions which may
also occur in its absence. ,
The suggestion as to the responsibility of abnormal nutri-
tional relations as the cause of abnormal and monstrous develop-
ment was made primarily in connection with the human embryo
and then extended to other forms. It is quite possible that in
the human and other placental mammalian embryos such effects
may in some cases be found to be specific and to be related to
nutritional abnormalities due to disturbances in the parental
organism. However, the mammalian embryo develops for some
hours or days before implantation occurs and becomes effective
as a nutritional factor so that it is only subsequently to
that time that the effects of faulty nutrition might be exercised.
LOW TEMPERATURE
DEVELOPMENT OF FUNDULUS 469
Embryonic differentiations having been begun before that time,
the way might be opened to a specificity of action upon some
part requiring a large supply of energy and material. The later
the abnormal conditions act, the more likely are they to be
specific in their action. But in the absence of definite proof of
such a cause, it is on the whole much easier to interpret mon-
strosities, even among the Mammals, as resulting from organi-
zational disturbances produced by the presence or absence of
some definite chemical environment, whether that be nutritional
or not. And such a chemical stimulus might be operative not
only after, but also before implantation, when the extent of the
resulting abnormality might be very great, so great as to result
in the formation of a ‘complete monster.’ I should not, how-
ever, oppose the idea that mammalian monsters may be due,
in some cases to nutritional defects: in this connection I should
merely suggest that the observed facts do not exclude the pos-
sibility of a general organizational disturbance as the cause of
monstrous development in this group, and that even in the pla-
cental mammalian embryo, although with less likelihood than in
oviparous forms, the nutritional disturbance, when it is known
to exist, may itself be a result rather than a cause of the de-
ranged organization. :
But it seems, in the light of subsequent observations, that it
was a mistake to extend this interpretation of the causes of
abnormal development so generally to other classes of verte-
brates and to say that such abnormalities result from conditions
‘“‘which affect the nutrition and impair the growth of the em-
bryo” (Mall ’08, p. 52) or which ‘‘tend to lower the develop-
mental vigor of the embryo” (Stockard, ’13 a, p. 83), or that
‘‘a certain amount of energy is necessary for differentiation of
the eye to take place . . . . but when the required energy
for any reason is not available the eyes are incapable of any dif-
ferentiation” (Stockard, 713 b, p. 271). Such a statement seems
to leave unexplained such cases as have been observed both by
Werber and myself, where portions of eyes, fragments of optic
cups, lenses, or even fairly complete eyes, may be found either
without other true tissues or organs being differentiated, or
with scattered parts of other organs and bits of tissue.
470 WM. E.- KELLICOTT
The difficulty of explaining such an abnormality as monoph-
thalmia asymmetrica, or other asymmetrical abnormalities, on
this nutritional basis is also apparent. Stockard recognizes this
and merely suggests that ‘It might be that at some critical
point in-development one of the future eye centers is affected
after the growth centers had begun to localize in more or less
lateral positions” (713 b, p. 278). Such asymmetrical abnor-
malities offer no special difficulties of interpretation on the
disorganizational hypothesis.
Stockard (13 b, p. 281) has also noted that the earlier the
treatment is administered during the development of Amblys-
toma the more extensive are the resulting abnormalities: ‘‘the
developmental period of administration is of as high importance
in determining the result as is the nature of the stimulus used.”’
This is readily understood upon the hypothesis that the unusual
stimuli act by disturbing the normal organizational relations of
nucleus and cytoplasm, since a simple or localized disturbance
at an early stage would be followed by much more widespread
effects than would an equal disturbance at a later stage, when
many of the differentiations might be already determined and
the effects consequently more localized. After the essential dif-
ferentiations of the organism have been made, the effects of
external stimuli would be likely to have relatively slight mor-
phogenetic results. (Compare the morphogenetic results of
stimulation in embryo and in adult organisms. )
Stockard believes that ‘‘all of the eye conditions [in Fundulus]
may be interpreted as arising through developmental arrests
(13 a, p. 83); and throughout his papers the abnormalities ob-
served are continually referred to as ‘defects,’ an interpreta-
tion that is cited in support of the nutrition hypothesis. While
most of the abnormalities observed have the nature of defects,
if by defect we mean only failure to differentiate, yet not all of
the abnormalities noted in Fundulus, to mention but this case,
are of this nature. In my own observations I might mention
for example, the development of two complete and separate
embryos on a single yolk; the development of two separate
hearts, not paired but in different regions of the embryo; the
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 471
proliferation of ectoderm cells, the formation of large masses of
erythrocytes, et cetera. Apparently the only way to interpret
such conditions as due to some defect is to have recourse to
the interpretative method employed by some geneticists (with
the difference, however, that there it may be clearly justified)
and say that when abnormal development occurs it is due to a
defect in that which would have kept it normal. But that
would merely be to say that abnormal development occurs.
And it should be recalled just here that Loeb (’15) noticed
that eyeless embryos developing from a cross between Fundulus
and Menidia reached that condition after passing through
earlier stages in which the eyes appeared to be normal. In such
instances anophthalmia is certainly not due to any original
destruction of ‘ophthalmoblastic anlagen.’
Loeb occasionally inclines toward this nutrition hypothesis
in explanation of some abnormalities and points out that, since
in most of his observed instances of anophthalmia circulation
also is lacking, ‘‘the inference is possible that the anomalous
condition of the eye may be due to lack of circulation” (715, p.
67). But the large number of instances in which the eyes may
develop normally in the complete absence of circulation renders
such an inference untenable.
A rather significant test of the nutrition hypothesis, or at
least of certain phases of it, can be found in those cases where
the embryo develops without a heart or without a circulation.
Here is certainly a profound disturbance of the nutritive rela-
tions of the entire embryo. Whatever the primary cause of
such a lack, the conditions should, on this hypothesis, be accom-
panied or followed by marked and varied abnormalities. If
nutritional disturbances so easily affect development as the gen-
eral hypothesis requires, such embryos should certainly produce
‘complete monsters’ capable of but a brief existence. It is often
the case that embryos lacking these organs are also defective
in various other ways, but it not infrequently happens that such
embryos may develop with a high degree of normality in all other
respects and when removed from the egg membranes may con-
tinue to live and react almost normally for some time.
472 WM. E. KELLICOTT
The main objections to the nutrition hypothesis of the causes
of embryonic abnormalities and monsters as it has been stated,
are, then, the following: it does not afford an interpretation of
the results, such as those of Bardeen and Packard, following
treatment of the sperm with radium radiations; it does not
afford any suggestions as to the nature of the underlying dis-
turbances through which the abnormalities are produced by the
unusual conditions used or assumed to be present; it does not
explain why the action of the experimental conditions during
cleavage should not produce visible results until much later; it
does not explain why monsters of the same parentage are di-
verse (see Mall ’08, p. 12); it does not explain why the effects
of treatment are greater when applied during the earlier stages
of development; it does not explain the production of abnor-
malities which are not defects or developmental arrests; it does
not avoid the necessity of assuming a degree of differentiation
during cleavage and a specificity of the action of the external
conditions, both of which have been shown not to exist.
On the other hand, the hypothesis suggested earlier in this
paper, that the causes of abnormal and monstrous development
are to be found in the disturbance of the normal organization of
the fertilized ovum or cleavage group, as evidenced by the ab-
normal characters and distribution of the nuclear and cytoplas-
mic substances, avoids these objections and affords an easy
interpretation of the observations mentioned.
It seems to have been premature to have assumed (Mall ’08)
that all the classes of embryonic abnormalities and monstrosi-
ties described by Mall, Bardeen, Spemann, Lewis and Stockard,
are really of the same essential nature and due to similar
causes. I see no reason for not admitting that such embryos
may result from different causes in different cases, whether they
be (1) the abnormal characters of the gametes before fertiliza-
tion, or disturbances of the early cleavage processes as these
concern both nuclear and cytoplasmic constituents, (2) the me-
chanical removal of differentiated rudiments, or (3) the lack of
energy and materials ordinarily supplied through nutritional
pathways, including the pathways within and among the parts
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 473
of the egg and embryo themselves; and I believe that these are
not merely different phases of the action of a single principle,
certainly not all due to a lack of ‘developmental vigor’ although
that is a pretty general phrase and one that might cover a mul-
titude of varied conditions, individually unlike and due to varied
primary causes. —
Turning now to a much briefer consideration of Werber’s
hypothesis, as stated above, I shall refer only to his latest papers
(15, 716). While admitting that the nutrition hypothesis may
account for the production of certain types of human monsters,
Werber rejects its general validity in favor of the view that
modifications in the physical or chemical! environment, for
example in the blood of the parent organism, affect directly
either the germ cells before fertilization or the fertilized ovum or
later developmental stages (15 a, p. 5380). Thus he agrees with
Spemann in opposing the hypothesis of Stockard mentioned
above, regarding the cause of cyclopia, and believes that the
eye-defects, which form the main subject of disagreement here,
are really due to morphological defects of some kind, and not
to an inhibition resulting from lack of developmental vigor or
energy. He goes farther than this, however, and suggests -that
the substances used in his experiments, namely, butyric acid and
acetone, caused ‘‘an elimination of materials of the blastomeres
or of the germ-dise and probably also of the yolk-sae.’’ ‘‘ Blas-
tolysis either destroys part or all of the germ’s substance, or it
may split off and disperse parts of the latter” (16). And fur-
ther ‘This elimination of material may be due either to the
precipitating or solvent effect respectively of the chemicals
which were used” (15a, p. 559). From the examination of
very extensive material he concludes that “either the blasto-
meres or the germ-dise had been blastolytically fragmented
owing probably to both physica] and chemical factors”’ (p. 559),
and that whatever scattered parts survive this fragmenting and
blastolytic process ‘‘may go on developing into a whole defec-
tive [sic] or a meroplastic, embryo, or even into an isolated
organ” (p. 559). And in another place (15 b) he adds that an
increased imbibition of water following an increase in the per-
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
ATA WM. E. KELLICOTT
meability of the egg membrane is another factor in producing
blastolysis. Thus through the operation of such factors together,
some parts of the germinal substance may ‘‘be entirely destroyed
owing to the increase in osmotic pressure, while the remainder
mzey go on developing and eventually give rise to various
monstrosities” (p. 240).
The first part of Werber’s hypothesis, namely that the effects
are due to a physical and chemical modification of the germ
cells, cleavage group or germ-disc, is clearly in accordance with
the more general disorganization hypothesis. But in the ab-
sence of direct evidence, which it would be extremely difficult to
secure, of the destruction of certain specific materials and not of
verminal substance in general, it seems more nearly in accord
with the general conception of development to believe that what
is effected is a disturbance or disarrangement of the constituents
cf nuclei or cytoplasm, or both, or of the normal relations be-
tween end among these materials. Otherwise, and in accord-
ance with the latter part of Werber’s hypothesis, it is necessary
to assume in the germ cell or cleavage group, a whole series of
unlike substances whose differentiations are already specific and
definitely necessary, not as organizational factors, but as the
rudiments or anlagen of the later differentiating tissues and
organs of the embryo. It is here that I should take exception
to the hypothesis, for there is as yet no direct evidence for the
existence, in such early stages, of such differentiated rudiments.
It seems much more likely that it is just the mechanism of dif-
ferentiation that is disturbed by these abnormal environments,
and not even the earliest formed results of the operation of such
a mechanism, although it is quite possible that there may also
be an added actual destruction of specific materials that as such
are necessary to normal development. This difference between
Werber’s view and that stated here is not a minor one; it is the
difference between the predeterminational and the epigenetic
views of development.
The fact is an important one, that results exactly parallel to
those of Werber (end Stcekard) follow upon the mere lowering
of temperature, 2 condition which eliminates the possibility of a
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS A475
direct chemical action, such as precipitation or decomposition,
and which can also produce disturbances jn the arrangement of
nuclear and eytoplasmic substances that are actually visible,
but which may only be assumed to have destroyed any chemi-
cally and developmentally differentiated and specifically neces-
sary materials. In this respect my experiments afford, I think,
a valuable check on the results following chemical treatment
and indicate the more general validity of the disorganization
hypothesis.
And further, since I have observed several instances where
eggs exhibiting rather extensive fragmentation and dispersal of
protoplasmic parts, were still able to form normal embryos, I am
led to believe that such conditions in themselves need not lead
to abnormal development, but that some other condition must
be primarily responsible. If certain essential organizational
conditions may remain present or susceptible of restoration,
development may proceed normally even though some consider-
able parts of the egg protoplasm may actually have been de-
stroyed; a conclusion which is also indicated by the results of
removal and pricking experiments on single blastomeres of the
Teleost ovum (Morgan 793).
Most of Werber’s discussion of the action of his mode of
treatment centers about the actual mode of the production of
eye-defects and of the respective merits of the ‘fusion’ and the
‘inhibition’ hypotheses. But such a discussion seems concerned
rather with the after effects than with the real causes of the
abnormality. That is, while it is very important to know the
proximate causes of such a condition as cyclopia, neither of the
suggested causes is really fundamental, and either fusion or
inhibition may result from a primary organizational disturbance,
in the same way that varieties of nutritional abnormality may
also result from a similar underlying cause.
In one respect Werber seems to fall into the error made by
Stockard in assuming that such conditions as cyclopia result
from a specific effect upon a differentiated rudiment of the an-
terior end of the central nervous system, already differentiated
in the early cleavage group. This leads him (’15, pp. 557-8) to
476 WM. E. KELLICOTT
support Child’s ‘axial gradient’ theory in explanation of the -
frequency with which there occur defects of the tissues and
organs developing in the anterior end of the embryo. ‘When
the egg is acted upon by a toxic substance, a restricted area at
the anterior end of the embryo’s median body axis becomes so
altered chemically as to be eliminated from further development
or it may go on developing to a certain point beyond which it is
chemically. unable to proceed” (p. 557). ‘‘The size of the in-
jured area at the anterior end is probably subject to considerable
variation,’ and thus the effects may be limited to the future
interocular area, or they may include parts, varying in extent,
of the potential optic anlagen, or one optic anlage only, and so
on. In addition to these assumptions, it is further necessary to
assume (p. 558) not only the existence of differentiated ophthal-
moblastic anlagen in the very early cleavage group, but definite
and symmetrically placed double ophthalmoblastic anlagen with
different degrees of susceptibility to the chemical substances in
solution. By means of these and other assumptions (p. 558)
Werber is able to reconcile the two hypotheses of the causes of
cyclopia mentioned above, but in accomplishing this he runs
contrary to the demonstrated lack of specification or deter-
mination in the cleavage group of the Teleost. Eggs were
treated in. the one- to sixteen-cell stages, and as Werber himself
remarks (pp. 531-2) it is probable ‘“‘that it is mainly the initial
effect of the toxic solution on the ovum that causes it to develop
in an atypical manner.”’ There is no anterior end of an embryo
represented by any differentiated material in the sixteen-cell
stage of the Teleost, no ophthalmoblastie anlage; but there is an
organization or developmental mechanism capable of producing
these parts, much later, a mechanism that is interfered with
and upset, and there is no specificity in the result of the de-
rangement. This lack of specificity is directly opposed to the
application of the axial gradient hypothesis, for as a matter of
fact, any part, posterior as well as anterior, may become abnor-
mal following this or other modes of treatment.
The explanations for the observation that all of the organs of
the anterior end of the Teleost embryo—eyes, brain, heart, etc.,
DEVELOPMENT OF FUNDULUS 477
LOW TEMPERATURE
are the more likely to exhibit abnormality seem much less rec-
ondite. In the first place, from the mode of its formation by
confluence (coneresence) the head end of the Teleost embryo is
the first part to be formed and differentiated, and any disturb-
ances of the normal processes of tissue and organ differentiation
are much more likely to be exhibited first in the region earliest
differentiating. And in the second place, the organs of the
head region, especially the sense organs, brain and heart are in
general more highly and therefore more extensively differen-
tiated, their normal development involves the interaction of a
larger number of factors, than in most other parts of the embryo
where muscles and the connective tissues form the chief con-
stituents. Hence a slight initial disturbance would produce a
more frequent as well as more marked result in the anterior
part of the embryo, where a more precise arrangement of the
underlying conditions of differentiation is necessary, than in the
more posterior parts where the tissues and organs are simpler
and possess greater regulatory properties.
And it should not be forgotten that as a matter of fact every
part of the embryo is subject to abnormality following treat-
ment with chemicals or low temperature; no part has been found
to be wholly free from abnormality in every case. This is an
observation, by the way, which has a decided bearing upon the
use of experimentally treated material as an aid to the solution
of problems in the normal development of Fundulus. If any
part, organ or tissue, whatever, may be affected abnormally by
such treatment, an extreme degree of caution should be exer-
cised in applying to the interpretation of the events of normal
development, the evidence drawn from the histogenesis of
embryos developing from treated eggs.
In conclusion I should like to refer briefly to the bearing of
the hypothesis stated here, upon a group of observations of a
wholly different kind. I refer to certain results of Teleost
hybridization described by Moenkhaus (10), Loeb (12), and
Newman (714). Moenkhaus found that in such hybrids ‘“de-
velopment in its early stages proceeds normally, 7.e., when
superficially viewed the deleterious effects of the two strange
478 WM. E. KELLICOTT
sex products upon each other showing only at later cleavage
and subsequently; that gastrulation, 7.e., the formation of the
germ layers—the most marked of the earlier differentiations of
the embryo, was a period of high mortality; that numerous ab-
normalities appeared in the hybrid embryos surviving this
period. He interpreted this to signify that the sex cells exer-
cised a poisonous action upon each other, preventing normal
development, and suggested as an analogy, merely, the toxic
effects of transfused bloods. Loeb also noted that various
abnormalities were not infrequent among heterogeneous hy-
brids, especially in respect to the circulatory system and eyes,
and suggested that the small size of the embryos thus produced
might be due to their failure to digest the yolk as rapidly as
the pure bred embryos.
Newman similarly finds the period of gastrulation one of high
mortality, but he shows that the effect of the ‘foreign’ sperm
may frequently be detected even during the earher cleavage
period. He also refers to the variety of abnormalities that may
appear during the later development of those embryos surviv-
ing the gastrulation period, and partially relies upon the nutri-
tion hypothesis to explain them. However, since the problem
he was investigating was an entirely different one, he does not
attempt a careful analysis of the probable causes of abnormal
development, and it would be unfair seriously to eriticise his
suggestion that death or abnormality during gastrulation is due to
failure to establish nutritive relations with the yolk, which is
to be regarded rather as a passing suggestion than as a definite
opinion. I do not understand that he, or anyone, has demcn-
strated that this period is especially characterized by the estab-
lishment of such relations with the yolk.
It is of considerable importance from my point of view to
note that Newman finds that one of the common effects of
hybridization is a disturbance in the time relations of various
processes of development, both acceleration and retardation
being quite common consequences. ‘This means that not only is
there a disturbance of the normal morphological sequences in
such hybrid organisms, but thet the whole organizational, in-
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 479
tegrational mechanism is or may be affected. Several obscure
relations noted by Newman might be profitably discussed from
the viewpoint of the general hypothesis which I am suggesting,
for example, his statement (p. 469) that ‘It is difficult to imagine
what factors underlie this wide range of success of individual
hybrids of the same parentage,’ an explanation of which I think
these suggestions afford.
However, it is not my intention, at this time, to attempt a
general application of this disorganizational hypothesis to such
phenomena, and the work of Moenkhaus and Newman is men-
tioned in this connection chiefly to show that such observations
are not counter to this general explanation of the causes of
abnormal development. For it is, I take it, a strong point in
favor of any suggested cause of a restricted group of phenomena,
that it is not opposed by the facts concerning nearly related
phenomena. Such seems to be the case with these suggestions
as to the causes underlying the formation of abnormal and mon-
strous embryos in Fundulus; for it not only accounts for the
results following treatment with low temperatures and chemical
substances, but it is not opposed, to say the least, by observa-
tions on the causes of similar developmental phenomena, follow-
ing treatment of the germ cells by radium radiations and fol-
lowing hybridization, and further, it accords with what is known
of the early development (cleavage) of the Teleost and with the
current general conceptions of the developmental process.
SUMMARY
1. By subjecting the eggs of Fundulus, immediately after
fertilization, to the temperatures of the ordinary refrigerator,
many of them are caused to develop abnormally when returned
to the laboratory temperature.
2. The abnormalities observed after such treatment cover a
very wide range and no characteristic, externally observable,
is found not to be affected to some degree in some embryo.
3. Similar treatment after the embryo has become well-formed
also leads to similar results, though with lesser frequency.
480 WM. E. KELLICOTT
4. The effects of the low temperature which may be actually
observed in the treated individuals, take the form of irregulari-
ties in the distribution and combinations of both nuclear and
cytoplasmic substances, and in the formation of cell-walls.
5. The results obtained by this method are essentially similar
to those already known to follow chemical treatment and hetero-
geneous hybridization.
6. It is suggested that the cause of abnormal and monstrous
development here and in other simuar instances, is to be found
ina disturbance of the normal organization of the ovum, as
expressed by the unusual characters and distributions of the
differentiated materials of the egg protoplasm.
7. This ‘disorganization’ hypothesis seems to afford a better
explanation of many instances of abnormal and monstrous
development among the vertebrates, than does the current
‘nutrition’ hypothesis, which is jn many particulars opposed by
the results reported here.
LOW TEMPERATURE—DEVELOPMENT OF FUNDULUS 481
LITERATURE CITED
BarDEEN, C. R. 1907 Abnormal development of toad ova fertilized by sper-
matozoa exposed to the Roentgen rays. Jour. Exp. Zodl., vol. 4.
Conkuin, E.G. 1912 Experimental studies on nuclear and cell division in the
eggs of Crepidula. Jour. Acad. Nat. Sei., Philadelphia, Ser. 2, vol. 15.
Lewis, W. H. 1909 The experimental production of cyeclopia in the fish em-
bryo (Fundulus heteroclitus). Anat. Rec., vol. 3.
Lintig, F. R: anp KNowtTon, F. P. 1897 On the effect of temperature on the
development of animals. Zool. Bull., vol. 1.
Lors, J. 1893 Ueber die Entwicklung von Fischembryonen ohne Kreislauf.
Arch. f. d. gesammte Physiol., Bd. 54.
1912 Heredity in heterogeneous hybrids. Jour. Morph., vol. 23.
1915 The blindness of the cave fauna and the artificial production of
blind fish embryos by heterogeneous hybridization and by low tem-
peratures. Biol. Bull., vol. 29.
McCuienpon, J. F. 1912 An attempt toward the physical chemistry of the
production of one-eyed monstrosities. Amer. Jour. Physiol., vol. 29.
Mautu, F. P. 1908 A study of the causes underlying the origin of human mon-
sters. Jour. Morph., vol. 19.
Morenxkuaus, W. J. 1910 Cross fertilization among fishes. Proc. Indiana
Acad. Sei., 1910 (1911).
Morean, T. H. 1893 Experimental studies on the teleost eggs. Anat. Anz.,
Bd. 8.
Newman, H. H. 1914 Modes of inheritance in teleost hybrids. Jour. Exp.
Zool., vol. 16.
PackarbD, C. 1914 The effect of radium radiations on the fertilization of
Nereis. Jour. Exp. Zodl., vol. 16.
Reaaan, F, P. 1915 <A further study of the origin of the blood vascular tissues
in chemically treated teleost embryos, with especial reference to hae-
matopoesis in the anterior mesenchyme and in the heart. Anat. Rec.,
vol. 10.
Sana, L. 1895 Experimentelle Untersuchungen iiber die Reifung und Befruch-
tung der Hier bei Ascaris megalocephala. Arch. mikr. Anat., Bd. 44.
SpeMANN, H. 1904 Ueber experimentelle erzeugte Doppelbildungen mit cyclo-
pischem Defect. Zool. Jahrb., Supplement-Band 7.
1912 Zur Entwicklung des Wirbeltierauges. Zo6dl. Jahrb. Abt. f.
allgem. Zodl. u. Physiol., Bd. 32.
Srockarp, C. R. 1909 a The development of artificially produced cyclopean
* - fish—‘‘The magnesium embryo.’”’ Jour. Exp. Zodl., vol. 6.
1909 b The origin of certain types of monsters. Amer. Jour. Obstet-
rics, vol. 59.
1910 The influence of alcohol and other anaesthetics on embryonic
development. Am. Jour. Anat., vol. 10.
1913 a The artificial production of structural arrests and racial de-
generation. Proc. N. Y. Path. Soc., vol. 13, N.S.
1913 b An experimental study of the position of the optic anlage in
Amblystoma punctatum, with a discussion of certain eye defects.
Am. Jour. Anat., vol. 15.
WM. E. KELLICOTT
Srockarp, C. R. 1915 An experimental analysis of the origin of blood and
vascular endothelium. Am. Jour. Anat., vol. 18. Reprinted as Mem.
Wistar Inst. Anat..and Biol., No. 7. :
Sumner, F. B. 1904 A study of early fish development. Experimenta! and
morphological. Arch. Entw.-Mech., Bd. 17.
Werser, E. I. 1915 a Experimental studies aiming at the control of defective
and monstrous development. A survey of recorded monstrosities
With special attention to the ophthalmic defects. Anat. Rec., vol. 9.
1915 b Further experiments aiming at the control of defective and
monstrous development. Year Book Carnegie Inst., Washington,
No. 14.
1916 Blastolysis as a morphogenetic factor in the development of
monsters. Proc. Amer. Assoc. Anat. in Anat. Rec., vol. 10.
THE DEVELOPMENT OF THE SPIRAL COIL IN THE
LARGE INTESTINE OF THE PIG
PAUL E. LINEBACK
Harvard Medical School, Boston, Mass.
TWENTY-THREE FIGURES
The student of human anatomy who happens to examine the
viscera of the adult hog will be greatly impressed by the spiral
arrangement of the ascending colon. For the colon in the pig,
after arising from the caecum which is almost entirely on the
left of the mid-ventral line, passes at once to the left, and then
swings around the abdominal cavity in voluminous coils. The
small intestines are mostly hidden behind it, though they appear
in the right iliac fossa, and altogether the colon is the principal
object seen when the abdominal cavity is opened.
To describe the course of the pig’s colon in detail is a diffi-
cult undertaking, which John Hunter skillfully attempted in the
following passage (Hssays and Observations, 1861):
It makes five spiral turns like a screw, coming nearer the center; at
the end of which it is bent back upon itself, passing between the former
turns as far as the first: but in this retrograde course it gets nearer the
center of the screw, so that it 1s entirely hid at last, then makes a
quick turn upward, adhering to itself and to the left kidney, as high
as the first spiral turn; from thence it passes across and close to the
spine, and before the mesentery, adhering to the lower surface of the
pancreas, and, as it were, encloses the fore-part of the root of the
mesentery; it then passes down the right side before the duodenum,
gets behind the bladder, and forms the rectum.
Hunter’s description was cited by Owen who notes that ‘‘the
spiral turns of the colon, above described, form one of the char-
acteristics of the Artiodactyle order.’’ To a certain extent this
is true, but the cow and sheep, and doubtless other forms, pre-
sent considerable modifications of the arrangement found in the
483
484 PAUL E. LINEBACK
pig. These are all more or less adequately described in the
text-books of veterinary anatomy. The spiral coil in the pig is
clearly figured by Sisson, as seen both dissected and in situ, and
he adds a diagram of its course.
The musculature of the pig’s colon is also peculiar, in that it
possesses two taeniae instead of three throughout most of its
course, and it was while studying the development of these
taeniae that my attention was diverted to the coil itself by Dr.
F. T. Lewis. At his suggestion the following account of its
embryological history has been prepared, and it is a pleasure to
record my indebtedness to him for codperating throughout this
work. Such an extraordinary and conspicuous formation has
not escaped previous study, but the existing descriptions are so
meager that they should certainly be supplemented by further
investigation. This was Martin’s opinion when in 1889 he
published the first of his papers containing most of the informa-
tion now available.
In the Schweizer-Archiv fir Thierheilkunde Martin presents
a series of diagrammatic figures showing the probable evolution
of the spiral colon in the sheep, including hypothetical drawings
of some stages which he had not actually observed. They are
accompanied by a brief description, containing references to
corresponding stages in the cow, but little is said regarding the
pig. In 1891 Martin supplied a new set of diagrams of the de-
velopment in the sheep, and a fuller description. In brief, he
considers that the colon, which previously has been quite
straight, forms a loop, the apex of which soon becomes bent like
a hook. The loop continues to elongate spirally, ‘‘as one of its
limbs grows slower than the other, and thus we have the begin-
ning of the spiral coil.”’
In the same year (91) Bonnet published modifications of
Martin’s earlier diagrams which are clearer, but apparently
more arbitrary. He records the formation of a ‘‘primitive loop
of the colon” in embryos of the horse. This, he observed, has
become somewhat S-shaped in a specimen measuring 10 cm.,
but it never makes more than one revolution. In the pig, Bon-
net finds that a corresponding loop ‘winds itself spirally around
THE LARGE INTESTINE OF THE PIG 485
an jmaginary axis and forms the colon-labyrinth, later shaped
like a bee-hive, consisting of 353 concentric outer and 33 excentric
inner convolutions.”’
In 1901 MacCallum described the development of the pig’s
intestine, intending to show that its coils are measurably con-
stant. In a 32-mm. embryo, he states that the large intestine,
“in the region where it turns to form the réctum”’ is thrown into
‘irregular twists.’ The complexity of this rectal group of coils
is said to increase in later stages, but its further evolution is not
described in detail. However, it is clear that MacCallum failed
to find a primitive loop, produced from an otherwise straight
colon, which elongated and grew into a helicoid spiral, as de-
scribed by Martin and Bonnet; and in the following study it
will be shown that the course of development is more compli-
cated than these authors have represented.
Beginning with an embryo of 12 mm., it will be found, wpon
dissection, that the intestine has formed its primary loop ex-
tending into the umbilical cord, and that torsion has not yet
occurred. The intestine at this stage may therefore be com-
pared with that of human embryos of 7-10 mm. In both forms,
the large intestine occupies the greater portion of the posterior
limb of the loop, beginning at the small bulbus coli, which marks
the future caecum (fig. 1). The apex of the loop in the pig is
more persistently attached to the vitelline duct than in man,
and this duct has been cut across in figures 1 to 4. Moreover
the length of the primary loop in the pig is greater than in
human embryos. In the pig the distance from the base of the
loop to its point of attachment to the yolk-stalk is between 4 and
2 of the length of the entire embryo. (For example, in figure
‘1 it is 4, and in the reconstruction of a 12-mm. pig by Lewis it
is 4). In human embryos possessing a primary unrotated loop,
as reconstructed by His, Elze, and Lewis, the length of the loop
is between + and 3 of that of the whole embryo. It is not im-
probable that the distinctly longer and more slender loop in the
pig provides for the more extensive intestinal convolutions, char-
acteristic of the pig in the stages immediately following.
486 PAUL E. LINEBACK
In pig embryos of 24 mm. the torsion of the primary lcop
has begun, so that the large intestine passes across the left side
of the duodenum and becomes the anterior limb of the loop (fig.
2). It extends from the freely-projecting caecum in a remark-
ably straight course, free from all convolutions, to the right-
angled bend where it descends to the rectum. ‘The other limb
of the loop, which forms the small intestine, has become three
times as long as the colic limb and is thrown into many convolu-
tions, arranged in linear series. Together they form a striking
and characteristic figure quite unlike anything seen in human
development. By referring to Mall’s reconstructions of the
intestines of two 24-mm. human embryos (1897, Tafeln 21 u.
22) or to Johnson’s more comparable drawing of a 22.8-mm.
specimen (Lewis, ’12, p. 321) the differences will be apparent.
According to Mall (97) the convolutions of the human small
intestine are quite constant, and there are six which are pri-
mary. With many secondary subdivisions, he found that these
could be identified in the adult. Following Mall, MacCallum
studied the development of the coils in the pig and likewise
found that in ‘‘embryos of the same size the coils are constant
in their arrangement and definite in their position.’”’ But this
conclusion ought not to be accepted without further investiga-
tions. In MacCallum’s figures of embryos of 23 and 25 mm.,
there is a well-marked stretch of small intestine toward the apex
of the loop, which is quite free from coils. No such interval is
shown in figure 2, and having found it but once in many em-
bryos dissected, I must regard it as exceptional. The four
primary groups of MacCallum are not apparent in my speci-
mens, and the individual coils in figure 2 cannot be homologized
with those in MacCallum’s reconstructions.
The torsion of the primary intestinal loop is carried much
further in the pig than in man. The human intestine rotates
through an are of approximately 180°, so that the original pos-
terior limb becomes anterior and vice versa. That is, it accom-
plishes such a rotation as is nearly completed in figure 2 and
stops at that point. But the pig’s intestine goes further, per-
forming a complete revolution, as shown in figures 3 and 4.
THE LARGE INTESTINE OF THE PIG 487
After this rotation of 360°, the limb of the loop which was origi-
nally posterior has again become posterior (as seen by compar-
ing figs. 1 and 4), but the anterior limb now crosses it twice.
As duodenum, it passes down on the right side of the colie or
posterior limb, then bends to the left beneath the colon, and
finally passes upward crossing the colon a second time, but now
on its left side. Figure 3 is an interesting intermediate stage in
this process. The first rotation, of 180°, has been completed,
even to the apex of the primary loop (which is not the case in
figure 2), and the second rotation of 180° has occurred in the
proximal part of the loop, but not distally. The rotation ex-
tends from the proximal portion of the loop outward, and the
process has been completed in the distal part of the loop in
figure 4.
Martin observed a similarly complete rotation in the sheep and
described it as follows:
The more the small intestine forms coils, the more it crowds the re-
current colic limb dorsally and at the same time to the right and cau-
dally, until it is surrounded by a ring of coils of small intestines; and the
half axial rotation about the mesentery in a 56-day embryo is trans-
formed into a complete rotation. Thereby the relation of duodenum
to the large intestine becomes changed. Earlier only a simple crossing
took place, but now there is an encircling.
In the pig, MacCallum described the rotation in connection
with the various groups of coils which he recognized. Thus he
states that ‘Group D,’ which includes those coils of the small
intestine which are nearest the caecum, “has rotated posteriorly,
dorsally, and to :the mght: o)/.:. ov. ) Itethussmoves: past
Group C and carries the caecum with it, so that the beginning
of the large intestine lies dorsally and posterior to Group D.”
Correspondingly a group of coils of the large intestine (Group
E) is said to “rotate through three-quarters of a circle.” Al-
though it is apparent that the rotation was observed, its
description is unnecessarily involved, since it js based on
groups of coils of questionable distinctness, rather than on the
limbs of the primary loop.
A488 PAUL E. LINEBACK
The formation of coils in the large intestine begins in pig
embryos of about 30 mm. At 35 mm. (fig. 4), they are well
developed and are gathered in a knot between the caecum and
the splenic flexure. This flexure it should be noted, is the only
one which is found in the pig, so that the pig’s ascending colon
corresponds with both the ascending and transverse colons of
human anatomy. A sinuous condition, preceding the formation
of distinct coils, is seen in the 30-mm. embryo, figure 3. Simi-
larly MacCallum found coils of the colon in pigs of 30 and 32
mm., and at 40 mm. he states that they ‘form a conspicuous
mass.’
The transformation of the mass of coils into the well-arranged
spiral of the adult may be followed in special dissections, and
from a large number of such preparations ten have been chosen
for illustration (figs. 5 to. 14). The final condition is shown in
figure 14, representing the colon from an adult, and this is pre-
ceded by a figure of the spiral four weeks after birth (fig. 13).
The other drawings are from embryos ranging from 50-120 mm.
Throughout the series, the smal! intestines have been cut away
near the colic valve, but in figures 5 to 11 a short piece of the
ascending portion of the duodenum has been retained for orien- -
tation. Around this the colon makes a characteristic bend, and
then, at the splenic flexure, it becomes the descending colon,
which is free from coils or kinks even in the adult. In all the
dissections the rectum has been cut away, and the entire mesen-
tery has been removed. The coils of the colon have separated
from one another, but only so far as necessary to show the
course of the tube. This necessitates a slight displacement of
some of the flexures, but none has been omitted or radically
altered.
In order to understand the stages in embryonic development
included in this series, it is necessary to have in mind certain
features of the condition to be finally attained, which are pre-
sented in figures 15 to 19. The apical portion of the coil in. the
adult (fig. 14), and likewise in the more advanced embryos, pre-
sents the curious pattern shown in figure 15 (from an embryo of
200 mm.). The observer will be uncertain whether the point a
THE LARGE INTESTINE OF THE PIG 489
or the point b is the actual apex. If a is selected, the adhesions
of the mesentery and of the adjacent coils may be torn apart so
that the spiral may be unwound as shown in figure 16. It
would then consist of two parallel limbs, which, in this speci-
men, make 33 revolutions. If 6b is taken as the apex, the coil
may be unwound as shown in figure 17. Beginning at the valve
of the colon as before, there are now only three revolutions.
It is evident, however, that the unwinding shown in figure 16
is natural, and that the other is an ‘artificial’ dissection, for
the adhesions yield more readily in the former and there is Jess
tearing of the mesentery. Moreover, at the base of the spiral,
the proxima] and distal limbs of thecoil early become adherent
to one another and to the body-wall, establishing a fixed point.
The basal limbs retain this position in figure 16, but have been
separated in figure 17, so that, the former is clearly the correct
picture, and the true apex is at a.
Additional revolutions may take place without changing the
apical pattern. If half a turn is added to the coil shown in
figures 15 and 16, the bend a will be carried up between y and z
toward 6, and the conditions shown in figures 18 and 19 will
result. In figure 19 the bend ap, which before unwinding the
coil appears to correspond with 6 in figure 15, is clearly the
true apex. .
The number of revolutions actually produced varies, and
small fractions, generally neglected, appear quite as often as
whole or half turns. Hunter, in the passage cited, speaks of
‘five spiral turns,’ evidently referring to the five tiers shown in
figure 14. In this specimen, however, beginning at the valve of
the colon, there are but four revolutions, and this appears to be
the normal number. Bonnet’s statement that there are 33 can
be applied to this specimen only by regarding b as the apex
instead of a.
The length of the spiral portion of the colon in the young
adult (fig. 14) is 2.6 meters. The distance from the colic valve
to the apex is 1.4 meters, or 53 per cent of the total length.
Thus the apex is finally located just beyond the middle of that
part of the colon which forms the spiral.
THE AMERICAN JOURNAL OF ANATOMY, VOL. 20, NO. 3
490 PAUL E. LINEBACK
In the following description the terms proximal or outer limb
will be applied to the part of the colon leading from the caecum
to the apex, and distal or inner limb to the part from the apex
to the splenic flexure. The inner limb in the adult is of much
smaller diameter than the outer, except toward the apex; and
its revolutions, which closely accompany those of the outer limb,
are hidden within the dome-shaped mass. Having in mind
these relations and the descriptions of previous writers that a
primary loop, with its apex at the middle point, simply winds up
to produce the adult form, the conditions in the embryo may
be carefully examined.
At 50 mm. (fig. 5), as in'the adult, the colon may be divided
into two nearly equal parts. The proximal half (in fig. 5 and in
the following drawings) has been heavily stippled to contrast
with the distal half. Beginning at the caecum, which at this
stage points ventrally, the colon passes toward the dorsal body-
wall, near which it makes a rather sharp turn and doubles back
upon itself. After running ventrally it redoubles by a sharp
turn and goes dorsally. This folding continues back and forth
throughout the proximal half. Distally the colon consists of
several short coils, irregularly arranged, which become adher-
ent to the body-wall near the duodenum. Except at this fixed
point, the colon at this stage is freely movable.
The most notable feature of the following stage (fig. 6, from
an embryo of 55 mm.) is the elongation of the first loop in the
proximal half of the colon. The proximal or outer half of this
first loop is now clearly a portion of the basal convolution of
the permanent spiral. In figures 7 and 8 (from embryos of 64
and 75 mm. respectively) the first coil has further elongated,
accomplishing, in figure 8, one half of a revolution. In con-
nection with this development, the caecum has shifted toward
the left of the body where it becomes permanently located,
and the entire colon has become twisted upon itself, dupli-
cating the torsion of the primary intestinal loop of earlier stages.
In other words, the part of the colon toward the caecum has
come to cross the left side of the distal part of the colon, just as,
in the 24-mm. stage (fig. 2), the large intestine crosses the left
THE LARGE INTESTINE OF THE PIG 49]
side of the small intestine. In fact the arrangement shown in
figure 8 might well suggest to a student of human embryology a
large intestine surrounding coils of small intestine. Where the
crossing takes place, the proximal and distal portions of the colon
become adherent to one another and thus the basal ends of the
ascending and descending portions of the future spiral become
fixed. It is therefore from this basal portion outward that the
future spiral is to be established, clearly necessitating a re-
arrangement of the irregular coils existing in the 70 mm. stage.
From the stages which have now been considered, it is evident
that the method of development in the sheep described by Mar-
tin and accepted by Bonnet is not applicable to the pig. The
colon does not present a simple primary loop, but shows many
convolutions. It is true that among these the first or basal
loop has a definite bend or apex as seen in figures 5 to 8, but
this, unlike the apex in Martin’s primary loop, does not become
the apex of the spiral in the adult. If it did so, the proximal
tenth of the colon in the 50-mm. pig must produce the proxi-
mal half of the adult spiral, and the distal nine-tenths would
produce only the distal half; but there is no evidence of such
unequal growth. It is therefore reasonable to suppose that
the proximal half in the embryo, which has been heavily stippled,
will produce a corresponding proportion in the adult. Accord-
ingly the future apex may be approximately located at the transi-
tion between the dark and light stippling, at a point which in
these stages has not been definitely established.
The continued advance of the spiral arrangement of the outer
coil is clearly shown in figures 9 and 10 (embryos of 90 and 95
mm. respectively). No further torsion of the colon than that
already recorded has taken place, but the winding up of the
outer basal coil has advanced from 4 a revolution in figure 8,
through 1 revolution in figure 9, to 2 complete revolutions in
figure 10. The way in which the bends of earlier stages are
obliterated in this process is suggested in figure 9, where several
_ are evidently about to be taken up in a well-rounded curve.
The first of these, at a, has nearly disappeared. It occurs at
a point x of the distance from the ileo-colic junction to the
492 PAUL E. LINEBACK
place where the colon comes into relation with the duodenum.
In figure 8 the apex of the bend a is at # of this distance, and
accordingly the flexures labelled a in the two figures may be
regarded as homologous. But in figure 9, a no longer marks the
apex. The bend which it designates has been incorporated in
the outer coil, together with the reversed bend 6, and a new apex
appears at c (which may fairly be compared with c in figure 8).
Beyond this point, in figure 9, the proximal half of the colon
still pursues a zig-zag course as in earlier stages, but it swings
back and fourth through shorter ares, and the transfer of the
apex from c to g is already suggested. To accomplish this, the
flexures d, e, and f are destined to pass through the condition
at present exhibited by a and 6. In the 95-mm. embryo (fig.
10) this has taken place, and slight irregularities in the outer
coil are all that remain of the former to-and-fro oscillations.
The outer or ascending spiral develops in advance of the
inner, descending spiral, as may be seen in the figures already
examined. The coils in the distal half of the colon are quite
disorganized in figures 5 to 8. Beginning in figure 9 (at g) and
more extensively in figure 10, the apical portion of the ascend-
ing coil is accompanied by a descending coil, and thus the final
relation between the outer and inner coils is beginning to appear.
An apex is thus established which will not advance further by
taking up flexures in its path, but chiefly through elongation,
which it shares with the rest of the colon, and by becoming more
tightly wound about its axis. However, it will beshown that a
slight advance of the apex along the inner or descending limb
is yet to occur, at the time when the characteristic apical pat-
tern is produced.
The final stages are shown in figures 11 to 14. In the embryo
of 110 mm. (fig. 11) three revolutions have been completed, and
except for the loosely wound apex, the spiral appears finished.1
The flexure a is destined to turn to the right and upward into the
concavity of the flexure c, and thus it will form the apical pat-
! The counting of the revolutions in the figures will be facilitated by placing
a straight edge from the ileo-colic junction to the apex.
THE LARGE INTESTINE OF THE PIG 493
tern already discussed. At the same time the apex will advance
from a to 6. This has happened in the 120-mm. embryo (fig.
12), and accordingly 33 revolutions are there present. The
same is true of the coil from a pig four weeks after birth (fig.
13). Gradually the spiral becomes more compact and its coils
more adherent to one another, and at the same time the portions
of the inner coil which are visible on the exterior become buried.
These changes are clearly shown in the figures. In the adult
(fig. 14) the apex of the coil has rotated so that instead of point-
ing downward, it is directed toward the left, and thus four
revolutions are completed. Less of the inner coil is exposed
than at birth, and only half a turn can now be observed at the
apex. With these relatively slight changes in the constitution
of the coil, its general appearance has been transformed through
the development of the sacculations, which at birth are scarcely
indicated.
The principal morphological feature of the developing coil
which the figures fail to suggest, is its increase in size and much
greater increase in length. This can be shown by measurements;
and at the same time the position of the apex can be more ac-
curately located. The following table, therefore, includes the
total length of the spiral part of the colon (from the colic valve
to the contact with the duodenum), and also the distance from
the colic valve to the apex of the coil. The apex in the earlier
stages 1s temporary, and in the 95-mm. specimen (fig. 9) it
must be chosen somewhat arbitrarily. When the distance from
the colic valve to the apex of the coil has reached 50 per cent
of the length of the entire spiral, the permanent apex has pre-
sumably become established. Accordingly, in the table, the
distance to the apex is followed by the percentage, which it
represents, of the spiral part of the colon. Measurements of the
small coils are made with some difficulty, so that the results
are.only approximately correct; but even with these limitations,
the measurements are found instructive.
494 PAUL E. LINEBACK
Stage of develcepment Spiral part of colon From colic valve to the apex
50 mm. 17 mm. 1.7 mm., 10 per cent
55 mm. 20 mm. 2.5 mm., 12 per cent
* 64 mm. 28 mm. 2.8 mm., 10 per cent
75 mm. 38 mm. 6 mm., 15 per cent
90 mm. 56 mm. 14. mm., 25 per cent
95 mm. 72 mm. 28 mm., 38 per cent
110 mm.? 101 mm. 51 mm., 50 per cent
120 mm. 121 mm. 68 mm., 56 per cent
4 weeks 1200 mm. 600 mm., 50 per cent
6 weeks 1635 mm. 845 mm., 51 per cent
young adult 2670 mm. 1480 mm., 53 per cent
2 Average of two specimens.
The way in which the coil develops suggests the possibility
of several sorts of anomalies, some of which were observed in a
series of one hundred adults and one hundred embryos examined
for this purpose. Five of the adults had coils with an addi-
tional half-turn, just as the embryo of 180 mm. shown in figure
18 has half a turn more than is usual atthat stage (cf. fig. 16).
No adult showed less than four revolutions, and none showed
reversals or other malformations of the spiral. Among the
embryos, anomalies were more abundant. Figure 21 represents
the colon of an embryo of 55 mm. which is placed beside a nor-
mal one (fig. 20) for comparison. In the anomaly the basal
coil has begun to rotate dorsally and to the right, in the reverse
direction. Three other specimens of coils reversed from the
base were found among the one hundred examined. A second
type of anomaly was observed in an embryo of 108 mm. (fig. 23,
likewise placed beside a normal specimen, fig. 22). Here the
spiral began to wind in the normal direction, but evidently
encountered a sharp flexure which could not be taken up, so that
a reversal occurs at the point x. Presumably at a stage cor-
responding with that shown in figure 8, such a bend as is there
labelled a persisted, and the advancing apex took the direction
of the flexure b. Accordingly, beyond the point x in figure
22 the coil is reversed. The inner coil has adapted itself to the
outer throughout, and in its concealed portion it reverses its
course opposite z. Another very similar anomaly was found
in an embryo of 95 mm., in which the inner coil likewise re-
THE LARGE INTESTINE OF THE PIG 495 -
versed so as to accompany the outer in its abnormal course.
Since six cases of partial or complete reversal were found in one
hundred embryos and none among one hundred adults, the ques-
tion arises whether the condition may be ultimately corrected,
or whether like a volvulus, it may lead to fatal results. But
the number of specimens Sere s. is perhaps too small to be
significant in this respect.
SUMMARY
In a series of drawings which largely explain themselves, an
attempt has been made to present the development of the colon
in the pig in greater detail than heretofore.
The torsion of the primary intestinal loop, which in man stops
at 180°, proceeds to a complete revolution in the pig; in this it
corresponds with the development in the sheep as described by
* Martin.
But the spiral coil in the pig does not begin as a single loop
of the colon which simply winds up, as in the sheep, according
to Martin and Bonnet. On the contrary it arises as a knot of
kinks and coils. The first of these forms the basal portion of
the outer limb of the permanent spiral.
Within the limits of the colon there then appears a rotation
or torsion, so that the proximal part crosses the distal part, and
the basal coil encircles the other convolutions in the way that
the human colon encircles the small intestine.
The basal coil advances by taking up secondary flexures in
its path until it makes two revolutions. By that timethe apex
of the coil is about midway in the course of the spiral part of the
colon, and further growth of the spiral is chiefly by the coiling
of the apex. In establishing the characteristic apical pattern,
however, the apex advances half a turn further along the inner
or descending limb of the coil.
In case the basal loop is turned in the wrong direction, or if
having started normally it encounters bends which do not yield,
complete or partial reversals of the spiral occur, six cases of
which were found in embryos.
496 PAUL E. LINEBACK
LITERATURE CITED
Bonnet, R. 1891 Grundriss der Entwickelungsgeschichte der Haussiuge-
thiere. 272 pp. P. Parey. Berlin.
Exuze, C. 1907 Beschreibung eines menschlichen Embryo von zirka 7 mm.
grosster Lange. Anat. Hefte, Abth. I, Bd. 35, pp. 409-492.
His, W. 1885 Anatomie menschlicher Embryonen. III. (Eingeweiderohr,
pp. 12-25). F. Vogel, Leipzig.
Hunter, J. 1861 Hssays and observations on natural history (ete.). Posthu-
mous papers, edited by R. Owen., vol. 2, (pp. 120-121 cited.) Van
Voorst, London.
Lewis, F. T. 1903 The gross anatomy of a 12-mm. pig. Am. Jour. Anat.,
vol. 2, pp: 211-225,
1912 The early development of the entodermal tract. Human Em-
bryology, ed. by F. Keibel and F. P. Mall, vol. 2, pp. 295-334. Lip-
pincott, Philadelphia.
MacCatuum, J. B. 1901 Development of the pig’s intestine. Johns Hopkins
Hosp. Bull., vol. 12, pp. 102-108.
Matz, F. P. 1897 Ueber die Entwickelung des menschlichen Darmes. Arch.
f. Anat. u. Entw., Jahrg. 1897, Suppl.-Bd., pp. 403-484.
Martin, P. 1889 Die Entwickelung des Wiederkauermagens und -Darmes.
Schweizer-Arch. f. Thierheilkunde, Bd. 31, pp. 173-214.
1891 Die Entwicklung des Wiederkauermagens und -Darmes. Fest-
sehr. f. Nageli u. Kélliker, pp. 59-80.
Owen, R. 1868 On the anatomy of vertebrates, vol. 3 (p. 474 cited). Long-
mans, London.
Sisson, S. 1914 The anatomy of the domestic animals. 2d.edition. (pp. 483-
488 cited). Saunders, Philadelphia.
THE LARGE INTESTINE OF THE PIG 497
Fig. 1 Ne
Fig. 4
Fig. 3
All the figures represent dissections of the intestines of the pig.
Figs. 1-4 The stomach and the small and large intestines seen from the
left side. X 14 diam. Figure 1, embryo of 12 mm.; figure 2, 24 mm.; figure 3,
30 mm.; figure 4, 35 mm. _b.c., bulbus coli.
498
Figs. 5-7
followed, seen obliquely in left-ventral view.
may
mm.
PAUL E. LINEBACK
The coils of the colon,
be
: figure 6, 55 mm. ; figure 7, 60 mm.
ascending colon; Col. desc., descending
Fig. 7
slightly displaced so that their course
Figure 5, embryo of 50
-all X 14 diam. Cae., caecum; Col. asc.,
colon; Duo., duodenum; II., ileum.
THE LARGE INTESTINE OF THE PIG 499
Fig. 10
Figs. 8-10 Same dissection and view as figures 5-7. Fig. 8, 75 mm.; fig-
ure 9, 90 mm.; figure 10, 95 mm.: all X 10 diam.
500 PAUL E. LINEBACK
Fig. 12
Figs. 11-12 Late stages in the development of the spiral coil of the colon.
Figure 11, embryo of 110 mm.; figure 12, 120 mm.: both x 7 diam.
THE LARGE INTESTINE OF THE PIG 501
Fig. 14
natural size. Figure 14, from a
ole2
Figure 13, from a pig 4 weeks after birth;
young adult; 4 natural size.
502 PAUL E. LINEBACK
Fig. 17
Figs. 15-19 Sketches showing the arrangement of the spiral and its apex.
Figure 15, the apical pattern, and figures 16 and 17, two methods of unwinding
the coil, from an embryo of 200mm. Figure 18, the spiral and figure 19, its apex,
from an embryo of 180 mm.; the spiral here presents half a revolution more than
that shown in figures 15 to 17.. The letters mark points referred to in the text.
THE LARGE INTESTINE OF THE PIG 503
ig. 21
Fig. 20 a
Fig. 23
Fig. 22
Figs. 20-23 Anomalies of the spiral. Figure 21, abnormal coil from an
embryo of 55 mm., placed beside a normal specimen (fig. 20) for comparison:
< 14 diam. Figure 23, an abnormal coil from an embryo of 120 mm., placed
od
beside a normal coil (fig. 22) from an embryo of 100 mm.: X 7 diam.
; .
7a
=, ae ; mn
2A ET ‘
” 7.
> 7 ea , A
% ” is rye a
ead): - i
r
SUBJECT AND AUTHOR INDEX
Ag of the pig embryo, with a note on
the fenestration of the anterior cardinal
veins. On the development of the
atrial septum and the valvular apparatus
AMRUNOMNLLGe eee en ae os ce
Atrial septum and the valvular apparatus in
the right atrium of the pig embryo, with
a note on the fenestration of the anterior
ae veins. On the development of
Ora co Saari SORBENTS ea
IRD’S lung. Based on observations of
the domestic fowl. Part II. The em-
bEyOlorysoOlobhe- sce cn tee scene cden.
(eget ome anlages and the formation of
the eardiae loop in the cat (Felis do-
mestica). The fusion of the
Cardiac loop in the cat (Felis domestica). The
fusion of the cardiac anlages andthe for-
TALEH AO) It GIT AL NYS5.6.14 A OO EL IRR Ee oe eo
Cardinal veins. On the development of the
atrial septum and the valvular apparatus
in theright atrium of the pig embryo, with
a note on the fenestration of the anterior. .
Cat (Felis domestica). The fusion of the car-
diac anlages and the formation of the car-
diac loop in the
Cavities and mesenchyma of the mammalian
embryo. Concerning certain cellular ele-
ments in the coelomic....................
Cells of the central nervous system. Morpho-
logical and microchemical variations in
mitochondria in the nerve
Cells). I. Spleen. Equivalence of different
hematopoietic anlages (by method of stim-
ulation of their stem
Cellular elements in the coelomic cavities and
mesenchyma of the mammalian embryo.
Concerning certain
Central nervous system. Morphological and
microchemical variations in mitochondria
in the nerve cells of the
Changes in the pancreasin phosphorus poison-
ing. Experimental mitochondrial........ 2
Chick. Origin of the sex-cords and definitiv
spermatogonia in the male................
Coelomic cavities and mesenchyma of the
mammalian embryo. Concerning certain
cellular elements in the
Coil in the large intestine of the pig. The de-
velopment of the spiral
ANCHAKOFF, Vera. Equivalence of
different hematopoietic anlages (by
method of stimulation of their stem cells)
MER SMICE NIA) s).cteis atlals wisest steteasise cies ss
Development, and adult anatomy of the naso-
frontal region in man. Genesis
Development of Fundulus. The effects of
low temperature upon the................
Development of the atrial septum and the
valvular apparatus in the right atrium of
the pig embryo, with a note onthe fenestra-
tion of theanterior cardinal veins. Onthe
Development of the spiral coil in the large in-
testine of the pig. The
THE AMERICAN JOURNAL OF ANATOMY,
351
351
255
125
449
| eres in the coelomic cavities and
mesenchyma of the mammalian embryo.
Concerning certain cellular
Embryo. Concerning certain cellular ele-
ments in the coelomic cavities and mesen-
_ chyma of the mammalian
Embryo, with a note on the fenestration of the
anterior cardinal veins. On the develop-
ment of the atrialseptum and the valvular
apparatus in the right atrium of the pig...
Embryology of the bird’s lung. Based on
observations of the domestic fowl. Part
| (aed Bl oY ee eR aoe Os oc Oe
Emmet, V. E. Concerning certain cellular
elements in the coelomic cavities and mes-
enchyma of the mammalian embryo
ENESTRATION of the anterior cardinal
veins. On the development of the atrial
septum and the valvular apparatus in
the right atrium of the pig embryo, with
a note on the
Follicles in the human thyroid gland. The
morphogenesis of the
Fundulus. The effects of low temperature
upon the development of
Fusion of the cardiac anlages and the forma-
tion of the cardiac loop in the cat (Felis
domestica). The
ENESIS, development, and adult anat-
omy of the nasofrontal region in man..
Gland. The lachrymal 2
Gland. The morphogenesis of the follicles
in the human thyroid.
EMATOPOIBTIC anlages (by method of
stimulation of their stem cells). I.
Spleen. Equivalence of different......
NTESTINE of the pig. The develop-
ment of the spiral coil in the large
ELLICOTT, Wm. E. The effects of low
temperature upon the development of
Fundulus
| Fimo abystel; MUG A apap aneaaee oe
LARSELL, Otor, Locy, Wiut1aM A., and. The
embryology of the bird’s lung. Based on
observations of the domestic fowl. Part
Lrvepack, Paut B. The development of the
spiral coil in the large intestine of the pig
Locy, Wittr1aM A.,and Larsei, Ovor. The
embryology of the bird’s lung. Based on
observations of the domestic fowl. Part
Loop in the cat (Felis domestica). The fusion
of the cardiac anlages and the formation
of the cardiac
Lung. Based on observations of the domestic
fowl. Part Il. The embryology of the
JOVHAG Lisa ns oncran din GOD OOREInEEED Oc ooUIO CODD
vou. 20, No. 3
73
73
351
411
125
147
411
255
483
483
45
506 INDEX
AMMALIAN embryo. Concerning cer-
M tain cellular elements in the coelomic
cavities and mesenchyma of the....... 73
Man. Genesis, development, and adult an-
atomy of the nasofrontal region in....... 125
Mesenchyma of the mammalian embryo.
Concerning certain cellular elements in
the coelomic cavities and........-......-. 73
Mitochondria in the nerve cells of the central
nervous system. Morphological and mi-
crochemical variations in......... Dondaons 329
Mitochondrial changes in the pancreas in phos-
phorus poisoning. Experimental........ 237
Morphogenesis of the follicles in the human
thyroid gland. The.........-..---.+....+ 411
Morritt, C. V. On the development of the
atrial septum and the valvular apparatus
in the right atrium of the pig embryo,
with a note on the fenestration of the
anterior cardinal veins...........-.--++++- 351
ASOFRONTAL regionin man. Genesis,
N development, and adult anatomy of
Nerve cells of the central nervous system.
Morphological and microchemical vari-
ations in mitochondria in the............ 329
Nervous system. Morphological and micro-
chemical variations in mitochondria in the
nerve cells of the central.................: 329
Nicuotson, Norman Cuive. Morphological
and microchemical variations in mito-
chrondia in the nerve cells of the central
MEGVOUS SVStCM 1: -c.c @ ease ee aeeterenre 329
Norris, Epcar H. The morphogenesis of
the follicles in the human thyroid gland.. 411
| OS aan in phosphorus poisoning. Ex-
perimental mitochondrial changes in the 237
Phosphorus poisoning. Experimental mito-
chondrial changes in the pancreasin...... 237
Pig embryo, with a note on the fenestration
of the anterior cardinal veins. On the de-
velopment of the atrial septum and the
valvular apparatus in the right atrium of
TE Sch ci ects Sis ese race tel ONG ee eet 351
Pig. The development of the spiral coil in the
lange intestine of the...........ts.se.nces 483
Poisoning. Experimental mitochondrial
changes in the pancreas in phosphorus... 237
CHAEFFER, J. Parsons. The genesis,
development, and adult anatomy of the
nasofrontal region in man............... 125
Scuutts, H. von. The fusion of the cardiac
anlages and the formation of the cardiac
loop in the cat (Felis domestica)........ 45
Scorr, W. J. M. Experimental mitochondrial
changes in the pancreas in phosphorus
OLSON od) ayovd is ve ole Lae coenartetes cote Pee 237
Septum and the valvular apparatus in the
right atrium of the pig embryo, witha note
on the fenestration of the anterior cardinal
veins. Onthe development of the atrial.. 351
Sex-cords and definitive spermatogonia in the
male chick. Origin of the............ Seo
Spermatogonia in the male chick. Origin of
the sex-cords and definitive.............. 375
Spiral coilin the large intestine of the pig. The
development, ofithe: .......... s-.as51s aelanwaiecets 483
Spleen. Equivalence of different hematopoie-
tic anlages (by method of stimulation of
=) the gstemyacellls) ly teiecr..ssa-/elee eaters 255
Stem cells). I Spleen. Equivalence of dif-
ferent hematopoietic anlages (by method
Of stimulation) Of wher) .pee tess eee 255
Stimulation of their stem cells). I. Spleen.
Equivalence of different hematopoietic
anlages (by method of. ...¢:...25. 002. 255
SuNDWALL, JOHN. The lachrymal gland.... 147
Swirt, CHartes H. Origin of the sex-cords
and definitive spermatogonia in the male
CHICK: 3 os See ieee aie She eee ees 375
EMPERATURE upon the development
of Fundulus. The effects of low...... 449
Thyroid gland. The morphogenesis of the
folliclesinmithe human... sec ceeseee ee 411
ALVULAR apparatus in the right atrium
of the pig embryo, with a note on the
fenestration of the anterior cardinal
veins. On the development of the atrial
Sepbumian dubhexs ns ete eens ‘ye eS
Variations in mitochondria in the nerve cells
of the central nervous system. Morpho-
logical and microchemical.............. . 829
Veins. On the development of the atrial sep-
tum and the valvular apparatus in the
right atrium of the pig embryo, with a
note on the fenestration of the anterior
cardinal...) cache aet ceareuitece alana tenet 351
ran
\|
\\}1\
| |
open
aes
raerten
ea spare eas
posters!
rasp tte
preheat aan eatvaseay
Wssiseasea setae sy er teases
pasate sept
presents
Has
zeke
ebnbineltaee!
eta Teel
pie
ate
Beas :
yy
ALINE LT LE ET
yet
Keay oh
vets
Be apes
aba vennaeeaehiees
Phe adapt ical
vats
ensues
sha:
Se Ane!
poeta tice
ASeAay,
Maest yates
Lats
ee
Sayre
aeitelt
Ane
SSL pURe reo ent
4
bratuere estes
eystbertstunsnt
rE
SPA Sh92
pasate
C-Map - woe
Snes RD RET mel NNT nena
are faee
corte
ISISISS
are
ie
nye
<
genie
bier
petvtrssteremte tected) wenticrencbeeerctrs
wervetert bvereisereeresreettocrieerttr
sheet errprereteecert reer.
pebrysr t t fs ecbereret
eatete ; i ceehrwwcrtrercereer mbeeeertn ren
fanaesarareae: zee spatecererecereteerertect eee
ti
te
Bereta
beth
ureeeprsbest
SAMSTAG
Su
aieaistanea tee