Py

SF pnt a Ae neem SEP

s

fF

Pato tere cee Ctr et erp ipeea gee

Ea
<

UEAEaNee ed saa eed

e

mae) beriat) Ets)

rVECG Teta Ghs
Vi Grty tenes }
TD Ee GT Ee « 4

AAD Ta } notin
i Ain

pets
in

9}
Bt f

rede

ena
ppheeaes ye ‘ hd revit
yar elegedgactpny , gp eee ert babe ny

rrviih
Hidiried teres

ca
S
aS

OSS

SS

Vibiter
Pian
ty Geet te :
Veet s
ii WG oars
eV eyfegeageteey

SS

oo
Sea

a bpeeTyy Tet!
Hien eerers ser
terlestiaseseenyy
Sapey
1

| Wiel!

: '
PEEPS Eer ey Ree

TSS
Sane

we

he
{yit?
ie ir if

aha
Lass use EECA

in

seyeey
eteedd

eg Sad van hep
SPR SS

YiFrige

SP eine
IS EV SEATS
t

STAsyev asec itpaeee tay else etsy
PAUUFLUPACaL tats et aas tae estas ina a
PEPESSESEVOSAISES Lesa A Pasa RRS
THPtPAP i eenrsy OPEL DER Ep tati ey

ff TSI TEST

WH

s(Uitet
Puarheepyy

iertypypet STE Ld
SECEDCrot pec testea reesei teers i pegee leery
Spetasdboarri Oba byeadeeneeaeseeleeerer bey ee
Prpeeepgaeberapeapaeces nage eter {
TELESLORYOSC PEDROS STSN eRe teed ey
Letseapeuepetep ety rages eeeeietey ceed ii

Pe aber

LPUECOTSUSSeaPeeP poser inatyap sagas rp eeceel

Copp peer CpeReapese steeped isasadean leg phy)
eUSTISTERESEI OT USES TAs ee MTT Lea

tredpepereyes Pecapeapererogpesasiegagtestac

fetieriyes POR ee
PPIRITABIARSTER SETI UET EAT LTRS Cea
bhapeee say iteagieadpeeiettiespeir egy

Speecpap prey easter eeaa ls teeee

IVS e eed

THPbE Read rece ebb ere bee
\ ERASSESSEREAAR EROS RRA I
Crivabanpaatoeds Dea)
CPU Ce SUE aera eae
ShUSeSee ead eaeaeepayaehesae seas area
SRPRTED CC HOGO PED a Peed ta tage eee besdagayay

LTRS Pattee
Tareeet enna
vesteeipegpels
dhereteelp ats

SOT TED Op Pa

PI TAAIT SS RPLREES LAIR UP TAD O STEED 00000) 2 ETT

conta

nif)
hiiytt
ya.

a,

suet a
iy etn)
PP PVTUHSLITITTTAATTLETELI AAT ELEY ESET LAA TIE pefiahicarg tt
TCA Baa OO BOUT ae
AEUEVPOOTELDAE RE DeR ED R OSTEO ELE Ey deed
PRESRATUIESS CPAP er tet ir
Wyte te Chiddeeceeabeal peed,
Tpesjup abaya ragee ibasarae sti }
DHSTIr nd PEST TESTER ei a jrai } Fis)
ATAU gaaeieroasssquctdorspiatpal lid saeahiad Wests)
TENT LT Hoe eae
STDP Rae Spey adbiste
GHGNUT EU LUE j ii
PINUS STTT ETS PRT EP TE ES)
HDD WADE 3
pipet bbdbae pagpeaganrgd yd |
Pel ebeecabey ies
in uf. :
dy UU
. THI
FEY)

Wiis
Husa
eye

Rte
Sollee eee rbestanrg ate
f ya

irs
itt ws

Sets
Lindh
Hath

onrecoane
epee

facie etna
SEES

eo.

i :
ait

sitet

Heeger apron tr
Bihar eae

pote

+
TATE TIA
ace
Ne ee arate
7) Wein teaitit,

=

aS

2.

Ag

ane
aes

>

iv
iss
ayes

iT

ie SE

peat tide
iia

fi

rit tai 4 Tater si

Meet ry aie, Parente?

it yay fi) it

ath ats

on Bs nae
pets

anaes
Sa

at

(rate

rersiah it
Higabie!

iatabla ‘
Dita SIRES LOPES Fore rssese

f pieblibas tate iebdys aa
Hf Pees eennenee tee tele eaves vem sesonte
pasrenryers aersitieo Waiak paid is
PEST SIOM ES eeceeEruet oy ronty

Paty Abbe Le iA

ts
ya SPusEess te

prey veserpesasarerss
beFeveree)

Addis
S135 aka
Lu. aeeAghs
28)

am
pene

=
tate
Te

THE AMERICAN JOURNAL

OF

ANATOMY

EDITORIAL BOARD

CHARLES R. BARDEEN

University of Wisconsin

Henry H. DoNaALpson

G. Cart HUBER J. Puayrain McMourrica
University of Michigan

University of Toronto
GerorGE S. HUNTINGTON GEORGE A. PIERSOL
The Wistar Institute Columbia University University of Pennsylvania
Simon H. GaaGe

Henry McE. KnNower, SECRETARY
Cornell University

University of Cincinnati

VOLUME 25
PAIN UA YY
1919

THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY
PHILADELPHIA, PA.

CONTENTS
No. 1. JANUARY

GerorGE L. STREETER. Factors involved in the formation of the filum terminale. Three

ENGST OUI CS Sarge Panta te ENED Pree mente treed cB 2 Hat scch adie plea otis RTE ME OED Sree ft
J. A. BapertscHer. The ultimobranchial bodies in postnatal pigs (Sus scrofa). Four
HOU CSeeee eee BO Greboiatg a hia ke Ut Gas Gord OER CE eae ORE oie ecto eae ic Orne C.
FRANK haees edema “The origin of the phagocytic mononuclear cells of the
peripheral blood. Eleven figures. . Ba 27
H. D. Sentor. The development of the sieibes a tite ipa apes eaeuien "Bleven
THEW INS Sh oe ctor rg Etna EN get oo Se PURGE CR i Se a Do ME 19)

No. 2. MARCH

C. H. DanrortH. The ane relations of brachydactly in the domestic fowl.
Five figures. oe o We

Dexa Dries. eraciee on inet ovary Os “fhe pperamonie (Gomaneniite eielioe eudecont
lineatus) with special reference to the corpus luteum. Twenty-nine figures........ 117

Royat Norton CHapMan. A study of the correlation of the pelvic structure and the
habits of certain burrowing mammals. Five plates (twenty-six figures)............ 185

No. 3. MAY

C.M. Jackson. The postnatal development of the suprarenal gland and the effects of

inanition upon its growth and structure in the albino rat. Ten figures. sae PPall
Joun C. Donatpson. The relative volumes of the cortex and medulla of ang eee
gland in the albino rat. Four charts.. Ae. Me eee. 20
FRANKLIN PARADISE JOHNSON. The eooleamiont of the jopule an ie pig’s silva ne
eight figures. . : .. 299
Lucite WITTE. Pie eoeeneais a ae fees mmmeele ae the pig in reiaiee fa the poueinemnes
and development of the intercalated discs. Two plates (eighteen figures).......... 333
NozA JULY

Epwarp Puecps Auuis, Jr. The homologies of the maxillary and vomer bones of Polyp-
terus. Three hee (eighteen figures)... pee . 349

J.A.Mynrs. Studieson themammary Hemel, “iy, The fieealoey a hed mammary ation in
male and female albino rats from birth to ten weeks of age. Nineteen figures (eleven
text figures and two plates).. ; . 395

H. E. Jorpan. The histology of fhe blood one fis réd bone -marrow oe ihe aed frc og,
Rana pipiens. Two plates (seventy-three figures) >...........:..0ccccecccsccccvsecs 437

il

é Aver ais : ‘ _ at . i be i
) Cee ne ea ty 2) eee Lae ae
pings shies, iL ta Sy ke
< Ave . J sd d Mi ah aie q
i aa” 7 — wey Nl 8 2. Oe

- =~ el fat ue aie |
«a 7s.) sees | >
- ee i Nae 7
- : : ‘ ; Lt vy) ‘

‘ ; .

7 '
é
Ae, baad
— q 7
h
hy
D -_—
t
- ~
* a
it =o
’

‘ S

‘

“
= " :
.
r’ = :
ol
4
- a toy wi)
he a7
A jae

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JANUARY 6

FACTORS INVOLVED IN THE FORMATION OF THE
FILUM TERMINALE

GEORGE L. STREETER

From the Department of Embryology, Carnegie Institution of Washington,
Baltimore, Maryland

THREE TEXT FIGURES

In a study recently published by the writer! on the develop-
ment of the cartilaginous capsule of the ear in human embryos
it was pointed out that the changes in size and form which the
capsule undergoes during its development are accomplished not
only by a progressive differentiation, but also in part by a retro-
gressive differentiation of its constituent tissues. The margins
of the cartilaginous cavities are in a continual state of change;
they exhibit an unstable equilibrium between two opposing ten-
dencies: on one hand, toward the deposit of new cartilage, and
on the other, toward the excavation of the old. The margins
thereby are always advancing or receding, and as a result of
this there is provided a suitable suite of chambers for the
contained membranous labyrinth in all stages of its development.

It is the feature of retrogressive differentiation or dedifferen-
tiation that I wish particularly to recall here. The fact that
certain areas of cartilaginous tissue revert to an earlier em-
bryonic type and are subsequently redifferentiated into a tissue
of a widely different histological character, is very clearly shown
in the case of the otic capsule, and is a factor of great embryo-
logical significance. Such a process of retrogressive change,
combined with redifferentiation of the same tissue, greatly in-
creases the facilities for and the range of certain structural

1 Streeter, G. L., 1917. The factors involved in the excavation of the cavities
in the cartilaginous capsule of the ear in the human embryo. Amer. Jour. Anat.,
vol. 22.

1

ys GEORGE L. STREETER

adjustments that occur in many regions in the development of
the human embryo.

Another instance of dedifferentiation has recently been pointed
out by Kunitomo.? This writer has published the results of a
careful study of the tail region in alargenumber of human embryos,
representing the period of greatest development of the caudal
appendage, and also the later period of its gradual reduction.
He shows that in very young specimens the spinal cord reaches
the extreme tip of the tail and throughout its length is quite
uniform in structure. Somewhat later (11 to 15-mm. stage) it
can be divided at about the level of the thirty-second vertebra
into two parts—a cranial or main part, having a wide central canal
and thick walls in which can be recognized well-developed mantle
and marginal zones, and a caudal slender part, having a narrow
canal with walls consisting only of an ependymal zone. Kunitomo
shows that it is this caudal atrophic portion that eventually
forms the filum terminale. The main part lying cranial to the
thirty-second vertebra undergoes uninterrupted and progressive
differentiation, whereas the portion caudal to this undergoes re-
gressive changes and, with the exception of the extreme tip,
finally becomes converted into a fibrous strand, the tip forming
the coccygeal medullary vestige. This, therefore, is another
instance in which an absorptive adjustment is brought about by
the reversion of the tissue to an earlier embryonic type with a
certain amount of subsequent redifferentiation.

Kunitomo further calls attention to the fact that in the forma-
tion of the filum terminale, in addition to the dedifferentiation
of the caudal end of the medullary tube, there is also the
mechanical disproportion between the growth of the medullary
tube and that of the vertebral column. How much of one and
how much of the other of these two factors is responsible for the
further development of the filum terminale was not determined
by him. It has occurred to the writer that this question could
be answered by the determination of the elongation of the nerve

2 Kunitomo, K., 1918. The development and reduction of the tail and of the
caudal.end of the spinal cord in the human embryo. Contributions to Embry-
ology, vol. 8, Publication No. 271, Carnegie Inst. of Wash.

FORMATION OF THE FILUM TERMINALE 3

roots. In the younger stages the spinal cord and the vertebral
column lie alongside of each other in a metameric manner, cor-
responding in position segment for segment. Owing to their
disproportion in growth, there occurs a relative displacement of
their segment levels, so that, for instance, the thirtieth segment
of the cord comes to lie opposite the twentieth segment of the
vertebral column. The segment levels of the vertebral column
are, of course, evident; in the spinal cord they are just as plainly
marked by the attachment of the nerve roots, for these become
attached to the cord before the displacement begins, and thus
permanently mark the various segmental levels. In the case of
each segment of the spinal cord there are two fixed topographical
points: the spinal ganglion, which is held in the intervertebral
foramen and registers the original position of the segment relative
to the vertebral column, and the place at which the dorsal root
is attached to the cord and which moves as the cord moves. By
locating those points for the different stages one can determine
the exact elongation of the nerve roots, and this in turn is the
index of the relative displacement of the spinal cord as regards
the vertebral column. Conversely, it will be seen that the
alteration not explained by mechanical displacement must be
attributed to the retrogressive changes referred to above. The
determination of the amount of displacement was made by
comparison of selected stages by means of profile reconstructions
of the smaller specimens and actual dissection of the older ones.
I was assisted in this by Mr. James F. Didusch, of the Carnegie
Embryological Laboratory, who made careful dissections of these
structures in several older fetuses, two of which will be used for
illustration. The results of this determination are given in the
following note as a matter of interest to those who have read
the paper by Kunitomo, and also because it offers an opportunity
to emphasize the significance of dedifferentiation of tissues in
the processes of development in the human embryo.

The part played by dedifferentiation in the caudal region of
the spinal cord is more apparent in the younger stages of develop-
ment, as pointed out by Kunitomo. The so-called ‘absorption’
of the tail is completed before the embryo reaches a length of

4 GEORGE L. STREETER

30 mm. It is also well known that the remodeling which takes
place in the gill region completes the obliteration of the gill bars
before the embryo is 20 mm. long. One well might expect these
processes of dedifferentiation and redifferentiation to be more
active in the earlier stages. They are not confined, however, to
- this period, for in the case of the ear capsule they were found to
be very active throughout fetal life. In the case of the spinal
cord dedifferentiation is well demonstrated in the period repre-
sented by embryos between 11 and 30 mm. long. A comparison
of these two stages can be made in figure 1. It will be noted in
the first place that the spinal ganglia show a regression varying
from arrest in development to complete disappearance. All but
two of the coceygeal ganglia have disappeared in the 30-mm.
specimen, and the remaining two are of about the same size as —
the same two ganglia in the 11.5-mm. specimen.’

As for the cord itself, the changes are equally marked. In the
younger stage (embryo 11.5 mm. long) the extreme caudal end
of the spinal cord, the part belonging to the non-vertebrated tail,
shows little differentiation, consisting only of indifferent cells
resembling embryonic ependyma. In the coccygeal region, how-
ever, the development is more advanced. Opposite the five
coccygeal ganglia the wall of the cord is differentiated into dis-
tinct ependymal, mantle, and marginal zones, with well-developed
rootlets entering into it from the first two ganglia. Sections
through it show nothing to indicate that this region is not going
on to complete its differentiation into the adult condition. When,
for comparison, one examines the very same region in the older
specimens (fig. 1, embryo 30 mm. long) it is found that its con+
dition, relative to the remainder of the cord, has undergone a
marked change. While the precoccygeal cord has continued to
increase in the thickness of its walls and in the elaboration of the
mantle and marginal zones, the coccygeal region is less advanced

3 Throughout this paper the twenty-fifth to the twenty-ninth segments have
been uniformly regarded as sacral. The slight variation which is known to exist
in this respect is too small to be taken into account in our genéral conclusions,
and for convenience the regional terms, lumbar, sacral, and coccygeal, will be
used, upon the assumption that the specimen concerned has the usual regional
distribution of its segments.

FORMATION OF THE FILUM TERMINALE 5

in its development than it was in the younger stage. Whereas in
the 11-mm. embryo there existed a distinct elaboration- into
ependymal, mantle and marginal zones, the mantle zone is com-
pletely missing in the 30-mm. embryo, and we find thin walls
consisting only of ependymal cells covered by a thin marginal
zone. The coccygeal spinal cord in the 30-mm. embryo is in an
earlier embryonic state than that of the 11.5-mm. embryo; that

0080U)

ve.

11.5 mm (x25)

30mm (x 12.5)

Fig. 1 Profile reconstructions showing the spinal ganglia and their dorsal
roots in the tail region of the human embryo. The last two lumbar ganglia are
shown in white, the sacral ganglia are stippled, and the coccygeal ganglia are
solid black. It will be noted that in the period included between these two
stages marked regressive changes have affected the entire coccygeal region of the
spinal cord, with complete disappearance of the last three coccygeal ganglia, in
sharp contrast to the sacral region of the cord, which undergoes uninterrupted
development. The reconstructions are taken from embryos No. 544, 11.5 mm.
long, and No. 75, 30 mm. long, belonging to the Carnegie Collection.

is, it has undergone dedifferentiation. In later stages the proc-
ess goes still farther and, as has been pointed out by Kunitomo,
this ependymal tube eventually becomes converted or redif-
ferentiated into a fibrous strand.

How much of the spinal cord is involved in this retrograde
process can be seen by comparing the two stages shown in
figure 1. In the region of the attachment of the fifth sacral
nerve the wall of the cord remains thick and develops a well-

6 GEORGE L. STREETER

differentiated mantle zone. About opposite the first coccygeal
nerve in the 30-mm. embryo the mantle zone abruptly disappears,
and there is a corresponding enlargement of the lumen of the ~
cord, thereby producing the thin-walled ventriculus terminalis.
There is some variation in different embryos as to the segmental
level caudal to which the mantle zone has dedifferentiated and
also in the manner of transition from the well-developed sacral
cord into the atrophic coccygeal cord, including sometimes the
doubling or partial obliteration of the central canal. The tran-
sition is quite abrupt, involving only one segment. In the
30-mm. embryo in figure 1 the cord at the level of the first
coccygeal nerve shows some decrease in the size of its mantle-
zone area. Opposite the second coccygeal nerve the mantle
zone is entirely gone. The second coccygeal ganglion present
in this case would probably soon have disappeared.

The ventriculus terminalis at this stage tapers caudally and
may be said to extend to the third coccygeal segment. Caudal to
this the differentiation of the cord is more complete and results
in the gradual obliteration of the lumen and the replacement of
the ependymal substance by a fibrous strand, embedded in
which can be found isolated groups of persistent ependymal
cells. At its extreme tip there is often found a more or less
detached group of such cells which undergoes cystic enlargement
and constitutes the coccygeal medullary vestige. The interval
of cord lying between this and the ventriculus terminalis con-
stitutes what is later known as the filum terminale. Thus far
its formation is based upon the process of dedifferentiation; its
subsequent growth and elongation is accomplished by an inter-
stitial increase of its constituent fibres, and not by the further
invasion of the process of dedifferentiation into the sacral region
of the cord. -This will become evident on examination of
figure 2

It has been pointed out that in embryos 30 mm. long a ven-
triculus terminalis is formed opposite the second and third
coccygeal vertebrae, owing to a retrogressive thinning out of
the walls of the spinal cord, with a consequent irregular enlarge-
ment of the central canal. In fetuses with a crown-rump

FORMATION OF THE FILUM TERMINALE

“SUOTIOOSSIP WOLJ OM} YSBT OY} ‘SUOTJONAYSUODaL a[Yord WOy
O1B OM) YSIY OT, “UU TZZ “CPET “ON UU TTT “E291 “ON ‘uu 29 “QCOT “ON ‘WUE Og ‘Gy “ON :SMOT[OJZ SB UOTJDETIOD
osoUIvA) OY} UL poyst] o1v Sosnjoy oy, ‘oTddiys ur uMOoYS sv ‘UIOJ JO OATJOOdSOIIT ‘OVIGQOJAOA [BIOVS SV PopavBoI OOM
OBIGOIIOA YBUTU-AJUIM 07 YZJJ-AQUOMY OYT, “Soul, AAvoy Aq UMOYS SI SOAIOU [BIOVS IoY}O oY} JO syood [esop yy jo
jue MYPIBIYR oY} PUB “UL UMBAP ST OAIOU [BIOS SI OY} JO JOO [VSLop oY} aSvO YOR UT ‘paod ywuIds oYy UO syuIOd poxy
PUB OFTUYOP 9}NITSUOD YOIYM ‘SOOT DAIOU [BAIS OY} JO JUD MYOR}IV oY} PUB O[OIIZUOA [VUTtC10} OY} 07 UOTPBIOA UT ‘WuNJoo
[eAgo}I0A oY} JO JUdUOOR]dstp [Vpnvo Jo yuo}xXo pu oye OY] SMOYS soBv4s osoy} Jo uostaRdWoD “oom YIJy-AjyuaMy

oy} OF YIYSto oy} WrOAZ SHyoF UBINY oY} UT psoo [vuTds oY} Jo puso [VpNvo oY} Jo suoTyejor [woTyduasodoy, Z “Biy

(g'1x) wu 47z (9€ x) eau | ||

19

re)

NS

ih

mh LT

8 GEORGE L. STREETER

length of 100 to 200 mm. the ventriculus terminalis can be
recognized in gross dissection with the naked eye, as a piriform,
translucent area at the tip of the conus medullaris. Apparently
in the natural condition it presents smooth outlines, but in pre-
pared sections its thin walls are thrown into what are evidently
shrinkage folds. This ventricle bears a permanent relation to
the rootlets of the fifth sacral nerve at their entrance into the
cord, as is shown in figure 2. In the four stages there represented
it lies just caudal to the entrance of the fifth sacral nerve. The
rootlets of the first coccygeal nerve in these specimens were so
delicate that they could not be traced with certainty and were
therefore omitted. The roots of the five sacral nerves, however,
could be very accurately followed, and are all indicated in the
figures. Their entrance into the substance of the cord con-
stitutes in each case a firm attachment and remains a fixed topo-
graphical point. By comparing the four stages from 30 mm. to
221 mm. it will be seen that the ventriculus terminalis and the
points of attachment of the sacral roots maintain the same
relative positions, there being no further encroachment of the
former into the territory of the more cephalic lying spinal cord.
In other words, there is no further dedifferentiation of the sacral
region of the cord after the embryo has attained a length of 30
mm. The cephalic migration that is subsequently experienced
by the ventriculus terminalis and points of attachment of the
sacral nerve roots, relative to the bodies of the vertebrae, is
clearly a result of the fact that the vertebral column gradually
extends farther caudalward than the spinal cord, and since the
nerve roots and the filum terminale are attached at both ends
they are correspondingly elongated. The latter process is not a
simple stretching, for, as these structures lengthen they actually
become thicker. In other words, there is a compensatory inter-
stitial growth. This increase in thickness is not apparent in
figure 2, as the older stages are shown at a progressively de-
creasing scale of enlargement.

The rapidity and extent of the caudal thrust of the vertebral
column—that is, its caudal displacement in relation to the
terminal ventricle—can be seen in figure 2. This covers a little

FORMATION OF THE FILUM TERMINALE 9

over the first half of fetal life (twenty-five weeks). In the
adult the corresponding points fall at the interval between the
bodies of the first and second lumbar vertebrae. Thus in the
first twenty-five weeks there is an ascent of the ventriculus
terminalis from the level of the second coccygeal to the third
lumbar vertebra, or a distance of nine segments, and there
remain but two segments before the adult position is reached.
One may say that the principal part of the migration is accom-
plished during the first half of fetal life.

The dura mater and its relations can be plainly recognized in
the 67-mm. fetus, where it can be seen to reach and adhere to
the filum terminale at the lower border of the fourth sacral
vertebra, thereby sealing off the lower end of the subdural
space. It is of interest to note that it undergoes very little
change from its position here and that which it occupies in the
adult. In the 111-mm. fetus it extends to about the same level
and ends in the same manner. In the adult it terminates about
two segments higher up. Thus the dural sac conforms more to
its bony environment than does the spinal cord and shows very
little of the migratory adjustment of position that is noted in
the latter. We therefore find the ventriculus terminalis
gradually receding cranialward from the caudal end of the sac.

In figure 2 the specimens are enlarged upon a decreasing
scale of magnification according to age, so that the segments of
the different stages are brought to about the same size. This
has been done in order to facilitate the comparison of segment
levels. The actual elongation of the spinal root of a given nerve
is greater, therefore, than would appear from the figure.
Measurements of the dorsal root of the first sacral nerve from
the margin of the ganglion to the point of entrance into the
cord yield the following figures: 30-mm. fetus, 0.65 mm. long;
67-mm. fetus, 4.75 mm. long; 111-mm. fetus, 12.25 mm. long;
221-mm. fetus, 32 mm. long.

The actual elongation of the first sacral root is indicated for
the first three of these stages in figure 3, in which the topography
of the spinal cord and the vertebral column is drawn on the same
scale of enlargement. The dorsal root of the first sacral nerve

10 GEORGE L. STREETER

Fig. 3 Topography of the spinal cord and the dorsal root of the first sacral
nerve in three fetal stages. 'These are taken from three of the same specimens
shown in figure 2, but here they are shown on one scale of enlargement in order

to indicate the actual changes in size. The 30-mm. specimen is shown both in
figure 2 and figure 1. y

FORMATION OF THE FILUM TERMINALE Lt

is indicated by a heavy black line; the first thoracic, first
lumbar, first sacral, and first coccygeal vertebrae are marked
by small circles. Comparison of the stages, as shown in this
figure, gives perhaps a better representation of the actual topo-
graphical changes that occur in this apparent ascent than does
figure 2.

From these results we may conclude that in the human
embryo the greater part of the coccygeal and post-coccygeal
cord—that is, the part caudal to the thirtieth segment—under-
goes dedifferentiation, the more cephalic part of it persisting as
the ventriculus terminalis and the more caudal part redif-
ferentiating into a fibrous strand—the fi!um terminale, with the
coccygeal medullary vestige at the tip. The first twenty-nine -
segments of the spinal cord are not affected by this process of
dedifferentiation, but continue in a progressive development.
When the embryo reaches 30 mm. in length there begins a dis-
‘proportion in the rate of growth as between the vertebral column
and the spinal cord, the former elongating more rapidly than
the latter. This results in a relative displacement of the two,
the ventriculus terminalis in the 221-mm. fetus (twenty-five
weeks) lying nine segments higher than it did originally, and by
the time the adult form is attained two more segments have been
added to this disproportion. We may say, therefore, that the
filum terminale represents that portion of the spinal cord caudal
to the second coccygeal segment (thirty-first segment), which
has undergone dedifferentiation and has finally become con-
verted into a fibrous strand. This strand, like the sacral nerve
roots, elongates by interstitial growth in adaptation to the
ascending displacement of the spinal cord. The caudal tip of
the dural sac maintains its relation to the vertebrae rather than
to the spinal cord and remains attached to the filum terminale
in the sacral region at a more or less fixed point.

Resumido por el autor, J. A. Badertscher.

Los cuerpos tiltimobranquiales después del nacimiento en el
cerdo (Sus scrofa).

Una porcién de los cuerpos tltimobranquiales puede persistir
durante largo tiempo en la glindula tiroides del cerdo, después
del nacimiento. En cerdos hasta de 56 dias de edad, estan
formados por areas de cordones sincitiales nucleados y masas en
las cuales est’ comenzando a formarse materia coloide, o por
dreas en las cuales los foliculos son pequenos pero con la estruc-
tura tipica de la glindula tirofdes. La porcién central de
algunas de las dreas sincitiales esta desprovista de coloide. Un
' drea alargada de pequefios foliculos que contienen coloide, la
cual representa probablemente un cuerpo tltimobranquial, fué
encontrada por el autor en dos entre tres glindulas tiroides de
cerdos subadultos. Los cuerpos tltimobranquiales estan colo-
cados generalmente en la mitad posterior de la glandula tiroides,
cerca de su borde dorsal o dorso-lateral. Es imposible deter-
minar exactamente la proporcién relativa en que los cuerpos
ultimobranquiales y el esbozo tiroideo medio contribuyen a la
formacion de la glandula tiroides. A causa del comportamiento
variable durante el desarrollo, por parte de los cuerpos tltimo-
branquiales, la proporcién relativa en que dichos cuerpos con-
tribuyen a la constitucién de la glindula tiroidea es indudable-
mente variable en diferentes cerdos. Es sin enibargo evidente
que solo una pequefia porcién de la glandula deriva de los cuerpos
ultimobranquiales. Folfeulos excepcionalmente grandes (cis-
toideos) abundan en algunas de las tiroides examinadas. De un
modo general puede decirse que la situacién de estos foliculos
esta limitada a la porcién de la tirofdes en la cual se encuentran
generalmente los cuerpos ultimobranquiales, es decir, en la mitad
posterior de la glandula.

Translation by Dr. José F. Nonidez,
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, DECEMBER 23
|

THE ULTIMOBRANCHIAL BODIES IN POSTNATAL
PIGS (SUS SCROFA)

J. A. BADERTSCHER

From the Department of Anatomy, Indiana University, Bloomington, Indiana

FOUR FIGURES

In a recent study (’18) of the ultimobranchial bodies in a wide
range of successively older developmental stages of pig embryos
(before their fusion with the median thyroid anlage to full
term), the writer was convinced that these bodies contribute to
the structural elements of the thyroid gland. The time at
which they are completely transformed into typical thyroid
structures, that is, when they can no longer be recognized
structurally from the median thyroid anlage, varies greatly.
Even in a full-term embryo a portion of the ultimobranchial
bodies may be free from colloid. It thus became evident that in
order to follow out a more complete developmental history of these
structures, the thyroid gland of postnatal pigs must be examined.

The material used for this investigation was obtained from a
litter of pigs and from three young adult hogs (age unknown).
The pigs were killed at the following ages: one a few hours
after birth, one 7.5 days old, one 15 days old, one 28 days old,
one 42 days old, and one 56 days old. The thyroid and a portion
of the trachea were removed from the pigs, while only the thyroid
was removed from the adults. All the material was fixed in
Zenker’s fluid and imbedded in paraffin. The thyroid glands of
the pigs were cut transversely into sections 15y in thickness, all
the sections were mounted in serial order, and stained with
eosin and Unna’s alkaline methylene-blue solution. The thyroid
gland of the adults were cut transversely into sections 20y
thick, only every fifth section was mounted, and stained with
hematoxylin and eosin.

13

14 J. A. BADERTSCHER

In the description of the following stages special attention will
be given: 1) to the structure of the ultimobranchial bodies and
to their location in the thyroid gland, and 2) to the location and
extent of areas of unusually large (cystoid) follicles. Since in
the embryonic material it was found that cystoid follicles may
develop in the ultimobranchial bodies, the latter consideration
is of importance.

Pig at birth. The thyroid gland is 9.3 mm. long and its
greatest width is 4.8 mm. Caudally it terminates in a rather
blunt point, while the anterior portion is drawn out into a slender
streamer. The greater portion of the bulk of the gland is thus
located in its posterior half. The more bulky part of the gland
is crescent in shape in transverse sections. In the middle third
of the left lateral half of the thyroid gland the ultimobranchial
body is represented by three small areas which are composed of
tortuous syncytial cords and masses quite closely packed to-
gether. These areas are embedded beneath the dorsal surface
of the gland lateral to its medial plane, a position usually
occupied by the ultimobranchial bodies in the later embryonic
- stages. Anterocaudally, they extend through eight, six, and
twelve consecutive sections, respectively. The most anteriorly
located of these areas is free from colloid and lies in a field of
follicles that are on an average smaller than the average size of
the majority of follicles present in the thyroid gland. In the
central and caudal areas the colloid is just beginning to form.
On account of the absence of colloid in one and its scanty amount in
the other two of these areas, they stand out sharply from the thy-
roid follicles immediately surrounding them. No difference could
be observed between the structure of the nuclei in these areas and
the nuclei in the cells composing the follicles. In the right lateral ,
half of the thyroid gland the ultimobranchial body is absent.

1 As this work is practically a continuation of a previous investigation (18)
by the author of the ultimobranchial bodies in pig embryos, it was deemed
unnecessary to repeat an historical sketch of this subject in this article. Also
the bibliography includes only those references to articles in which may be’
found more orless definite statements concerning the fate of the ultimobranchial
bodies. If an extensive bibliography on this subject is desired, reference should
be made to the, works of Verdun (’98) and Grosser (12).

ULTIMOBRANCHIAL BODIES IN POSTNATAL PIGS 15

Unusually large follicles are present in both lateral halves of the
caudal fourth of the thyroid gland. These are located mainly
near the dorsal and dorsolateral surface of the gland. A few
very large follicles are located near the most caudally located
area of the ultimobranchial body. Also an area of very large
follicles (extending through a series of thirty-four sections) is
present in the caudal portion of the middle third of the right
lateral half of the thyroid gland just below its dorsal surface.

Pig 7.5 days old (fig. 1). The thyroid gland is 15 mm. long
and its greatest width is 6.8 mm. The greater portion of the
anterior half of the thyroid gland is in form a slender band, so
that by far the greater portion of its bulk les in the posterior
half of the gland. The more bulky portion of the gland is
crescent in shape in transverse section. Traces of both ultimo-
branchial bodies are present in the caudal fourth of the thyroid
gland. The left one is represented by three oblong areas com-
posed of closely packed syncytial cords and masses which are
almost free from colloid. The nuclei in the syncytium have a
structure identical to the structure of the nuclei in the cells
composing the follicles. Anterocaudally, these areas extend
through nineteen, seven, and thirty consecutive sections, re-
spectively. The most anteriorly (fig. 1, U.) and caudally lo-
cated of these areas are partially exposed to the free surface
on the dorsal border of the thyroid gland, while the centrally
located one is embedded only a short distance below the dorsal
surface of the gland. The ultimobranchial body on the right
side has a structure similar to the left one. It extends through
a series of ten sections and is entirely embedded below the
dorsal surface of the thyroid gland.

In the left lateral half of the thyroid gland and just anterior
to the ultimobranchial body represented in figure 1 is the caudal
termination of an elongated area containing many large (cys-
toid) follicles. This area of large follicles is located just below
the dorsal border of the thyroid ‘gland and extends through
seventy consecutive sections. Very large follicles are also found
in the immediate neighborhood of the centrally located area of
the ultimobranchial body. Two rather large follicles lie near

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 1

16 J. A. BADERTSCHER

the right ultimobranchial body. A few large follicles are found
near the caudal end of the thyroid gland.

Pig 15 days old (fig. 2). The thyroid gland is 10.5 mm. long
and its greatest width is 6.3 mm. It tapers to a blunt point at
each end. Only the right ultimobranchial body is present. It
is embedded deeply below the dorsal surface in the posterior
half of the thyroid gland. It is represented by two areas which
are composed of syncytial cords and masses in which follicles
containing colloid are quite numerous, but on an average much
smaller than the majority of follicles in the immediate neigh-
borhood of these areas. The more anteriorly located and larger
of these areas (fig. 2, U.) extends through a series of eighty-five
sections, while the more caudally located one extends through
sixteen consecutive sections. Thirty-six sections intervene be-
tween the two areas.. The structure of the nuclei in these areas
is identical to that of the nuclei in the cells composing the
follicles. In the caudal fourth of the thyroid gland there are
many very large follicles which are located chiefly in the dorso-
lateral margin of the gland.

Pig 28 days old. The thyroid gland is 11.1 mm. long and its
greatest width is 7.7 mm. It tapers to a blunt point at each
end and its more bulky portion is crescent in shape in cross-
section. Both ultimobranchial bodies are present. They are
located just below the dorsal surface in the caudal portion of the
middle third of the thyroid gland, lateral to its medial plane.
The right one is similar in structure to the ultimobranchial body
in the thyroid gland of the 15-day-old pig. It extends through
a series of sixteen sections. The left ultimobranchial body is
composed largely of an attenuated syncytial mass in which are
found a few small follicles. It extends through a series of
thirty-eight sections.

In the immediate neighborhood of the left ultimobranchial
body are found a few very large follicles. Many very large
follicles are present in the taudal fourth of the thyroid gland.
These are located chiefly near the dorsal and dorsolateral border
of the gland, excepting near its caudal end where they are
scattered throughout the entire thickness of the gland.

ULTIMOBRANCHIAL BODIES IN POSTNATAL PIGS 17

Pig 42 days old (fig. 3). The thyroid gland is 14.6 mm. long
and its greatest width is 7.3 mm. It tapers to a point at each
end and, excepting near its ends, is crescent in shape in trans-
verse section. The ultimobranchial body on the right side is
located midway between the two ends of the thyroid gland and is
embedded just below the dorsal surface of the gland lateral to
its medial plane. It extends through a series of twenty-six
sections and is composed of syncytial cords and masses. In
both ends and in the peripheral portion of this structure the
follicles are quite numerous but comparatively small, while in
places its center is free from colloid. The ultimobranchial body
on the left side is located in the anterior portion of the caudal
fourth of the thyroid gland and is embedded just below the dorsal
surface of the gland lateral to its medial plane. It extends
through fifty-four consecutive sections and has a structure
(fig. 3, U.) similar to the right one. In place it is almost
separated from the rest of the thyroid gland by connective
tissue.

An area of large follicles in the right lateral half of the thyroid
gland extends anteriorly from the ultimobranchial body. This
area of large follicles is located chiefly just below the dorsal
border of the gland, but in places it extends to its dorsalateral
margin. In the dorsolateral portion of the middle third of the
left lateral half of the thyroid gland is an area, variable in width,
of very large (cystoid) follicles. Near the caudal portion of the
anterior third of the thyroid gland these areas of large follicles
become continuous just below its dorsal surface and gradually
extend more deeply into the gland so that large follicles are
found throughout its extreme anterior portion. No follicles of
an unusually large size are found in the extreme caudal portion
of the thyroid gland.

Pig 56 days old. The thyroid gland is 13.5 mm. fae and its
greatest width is 10.2 mm. It tapers to a blunt point at both
ends. The left lateral half of the gland is considerably more
bulky than the right lateral half. The only traces of an ulti-
mobranchial body is an area of very small follicles near the
dorsolateral margin in the right lateral half of the thyroid gland.

18 J. A. BADERTSCHER

This area extends through a series of ten consecutive sections
and is located nearly midway between the two ends of the
gland. A feature very noticeable in the thyroid of this pig is
the presence of only a small number of large follicles. These
are located chiefly near the dorsal border in the caudal fourth of
the gland.

Young adult hog no. 1 (fig. 4). The thyroid gland is 26 mm.
long and its greatest width is 18.5 mm. Its anterior end termi-
nates in a single blunt point, while its posterior end terminates
in two blunt processes each about 3.5 mm. long.. The only
structural feature representing a possible derivative of an ulti-
mobranchial body is an elongated area of small follicles (fig. 4,
U.) which extends through 125 consecutive sections. This area
is located along the dorsal border in the posterior fourth of the
right lateral half of the thyroid gland and extends for a short
distance into its right terminal process. More interfollicular con-
nective tissue is present in this area than in other parts of the
gland. The anterior two-thirds of the thyroid is free from
unusually large follicles. Many are found in the posterior third of
the gland. Two large follicles (C.F.) are represented in figure 4.

Young adult hog no. 2. The thyroid gland is 29 mm. long
and its greatest width is 20 mm. The anterior end terminates
in two blunt processes each 6.5 mm. long, while the posterior end
terminates in a single blunt point. The only structural feature
representing a possible remnant of an ultimobranchial body is
an area of small follicles in the posterior fourth of the left lateral
half of the thyroid gland. This area extends through a series
of forty sections and has a structure similar to that of the
ultimobranchial body represented in figure 4.

Two areas of large follicles are present. In the interior and
middle thirds of the thyroid gland these areas extend from the
tip end of the two anterior processes along the dorsolateral-
margin of the thyroid. In the caudal third of the gland these
areas gradually become larger, so that at the extreme caudal
end they are found throughout the entire thickness of the gland.

Young adult hog no. 3. The thyroid gland is 26.5 mm. long
and its greatest width is 20mm. At each end it tapers to a blunt

ULTIMOBRANCHIAL BODIES IN POSTNATAL PIGS 19

point. No remnants of the ultimobranchial bodies are present.
Only a few follicles of an unusually large size are present in the
extreme anterior and posterior ends of the gland.

SUMMARY

In pig embryos? it was found that the peripheral portion of
the ultimobranchial bodies generally develops into typical thyroid
structures before its more central portion. It thus becomes
evident that the ultimobranchial bodies which can be recognized
structurally as such in the thyroid gland of postnatal pigs are
mere remnants of these structures that have not fully developed
into typical thyroid structures. Also, the structure of the
ultimobranchial bodies in the late developmental stages of pig
embryos and in the postnatal pigs are similar, namely, areas
composed of nucleated syncytial cords and masses the central
portion of which may be free frum colloids. Areas of small
follicles (developmentally young) are not so marked in the
immediate neighborhood of the ultimobranchial bodies in post-
nata pigs as in the immediate vicinity of these structures in
most of the late embryonic developmental stages.

One of the ultimobranchial bodies in the thyroid gland of
three pigs is not a continuous structure, but is broken up into
segments. For example, the one on the left side in the 7.5-day-
old pig is composed of three oblong areas which vary in length.
It seems that a satisfactory explanation for this condition is a
more rapid transformation into typical thyroid structures of
some parts of the more central portion or core than of other
parts, thus dividing it into segments which are separated from
each other by follicles. The length of the portion of an ultimo-
branchial body broken up into segments is obtained by measur-
ing the distance between the anterior and posterior parts of the
most anteriorly and most posteriorly located segments, re-
spectively. Thus the left ultimobranchial body in the 7.5-day-
old pig extends through a series of seventy-five sections. The

2 In the summary it will be necessary to refer quite frequently to the writer’s

previous work on the ultimobranchial bodies in pig embryos. This will be done
without calling attention to the bibliographic reference of that article.

20 J. A. BADERTSCHER

right ultimobranchial body in the same pig is represented by a
single area which extends through a series of ten sections. Judg-
ing from the so variable developmental behavior of these
structures in pig embryos, this single area (and similar single
areas in the thyroid gland of other pigs) undoubtedly does not
represent the greatest length of the central core of this ultimo-
branchial body, but only what is left of it at the time the pig
was killed. It underwent a more rapid transformation than the
left one.

A considerable lapse of time intervenes between the age of the
pigs and the young adult hogs, thus producing an undesirable
break in the continuity of the postnatal developmental history
of the ultimobranchial bodies. Although the conclusion in re-
gard to these structures in the young adult hogs is thus rendered
somewhat uncertain, I have a strong feeling that the areas of
small follicles in the thyroid of the young adult hogs nos. 1 and 2
represent the ultimobranchial bodies in an advanced stage of
development. Their structure, location in the thyroid gland,
and their proximity to unusually large follicles strengthen this
- interpretation.

In some of the later developmental stages of pig embryos it
was found that unusually large (cystoid) follicles develop in
connect’on with the ultimobranchial bodies. The extent and
location of areas of this type of follicles in the thyroid gland of
postnatal pigs need therefore to be considered.

The existence of a possible interrelationship between the large
follicles and the ultimobranchial bodies is exemplified in a single
thyroid gland in scme of the postnatal pigs. Thus, in the pig
at birth the ultimobranchial body that is present is located in
the middle third of the left lateral half of the thyroid gland,
while an area of large follicles occupies an almost corresponding
position in the right lateral half of the gland. Also a few very
large follicles are found near the most caudally located segment
of the ultimobranchial body. In the thyroid gland in pig 7.5
days old large follicles are found in the immediate neighborhood
of the right ultimobranchial body and near the central segment
of the left one. Also cephalad to the most anteriorly located

ULTIMOBRANCHIAL BODIES IN POSTNATAL PIGS 21

area of the left ultimobranchial body is an elongated area, of
large follicles that extends into the middle third of the thyroid
gland. <A few very large follicles are found in the immediate
neighborhood of the left ultimobranchial body in the thyroid
gland in pig 28 days old. Large follicles are present in the
immediate neighborhood of the ultimobranchial bodies (?) in
the young adult hogs nos. 1 and 2.

In a general way, it can be stated that the unusually large
follicles are most numerous in the majority of postnatal stages
in the portion of the thyroid gland in which the ultimobranchial
bodies are generally found, that is, in the posterior half of the
gland. ‘This feature is in accord with the fact that in embryos
from about 50 mm. in length to full term the ultimobranchial
bodies are also usually located in the posterior half of the
thyroid gland, although in some they are located in the middle
third or in the middle two-fourths of the gland. Exceptions to
the usual location of the majority of large follicles in a single
gland are not wanting. The. most striking example of this
exception is found in pig 42 days old, in which the great majority
of large follicles are found in the anterior half of the thyroid
gland. In consideration of the variable developmental behavior
of the ultimobranchial bodies* in pig embryos, it is therefore,
not surprising (granting an interrelationship between the large
follicles and the ultimobranchial bodies) to occasionally find
large follicles out of their usual place (pig 42 days old) or the
almost entire absence of large follicles, as is the case in the
thyroid gland in pig 56 days old and in the young adult hog
no. 3.

Another significant feature of the location of the very large
follicles is the depth at which they are embedded in the thyroid
gland. In the more bulky portion of the gland they are located
near the dorsal or dorsolateral border of the gland, that is, in
the region in which the ultimobranchial bodies may be found,
and not in the ventral or ventrolateral region which is derived

3 Variations as to their size, the time of their complete transformation into

typical thyroid structures, their location in the thyroid gland, and the presence
or absence of large (cystoid) follicles in connection with them.

Pip J. A. BADERTSCHER

from the median thyroid anlage. The caudal end of the thyroid
gland in some of the postnatal pigs (pig at birth, pig 15 days old,
and young adult hogs no. 1 and 2) contains many large follicles.
This feature apparently bears with it a significance when it is
correlated with the fact that in some of the comparatively early
embryonic stages the caudal portion of the tripartite complex is
largely composed of ultimobranchial bodeis.

Although the evidence of an interrelationship between the
large follicles and the ultimobranchial bodies in postnatal pigs
is circumstantial, yet when this evidence is correlated with their
development in connection with the ultimobranchial bodies in
some of the later embryonic developmental stages, it appears
that the ultimobranchial bodies are largely responsible for the
large (cystoid) follicles.

In conclusion, it can be said that since the ultimobranchial
bodies fuse with the thyroid gland and also form colloid the
boundary between these structures and the gland becomes ob-
literated, so that it is impossible to determine the exact relative
proportion that is contributed to the thyroid gland by the
ultimobranchial bodies and the median thyroid anlage. Owing
to the variable developmental behavior of the ultimobranchial
bodies, the relative proportion they contribute to the thyro‘d
gland undoubtedly varies in different pigs. It is, however,
quite evident that only a relatively small portion of the gland is
derived from the ultimobranchial bodies.

ULTIMOBRANCHIAL BODIES IN POSTNATAL PIGS Dex

BIBLIOGRAPHY

BaperRTSCHER, J. A. 1918 The fate of the ultimobranchial bodies ‘n the pig
(Sus serofa). Amer. Journ. Anat. vol. 23.

Born, G. 1883 Uber die Derivate der embryonalen Schlundbogen und Schlund-
spalten bei Siugetieren. Arch. f. mikr. Anat., Bd. 22.

Grosser, O. 1910 Zur WKenntnis des ultimobranchialen Kérpers beim Menschen,
Anat. Anz., Bd. 37.

1912 The development of the pharynx and of the organs of respira-
tion. Manual of Human Embryology, edited by F. Keibel and F. P.
Mall, vol. 2.

HrrRMANN, G., AND VERDUN. 1899 Persistance des corps post-branchiaux chez
Vhomme. Remarques sur l’anatomie comparée des corps post-
branchiaux. Comptes Rend. Soc. Biol. Paris.

1900 Note sur les corps post-branchiaux des Caméliens.—Les corps
post-branchiaux et la thyroide; vestiges kystiques. Comptes Rend.
Soc. Biol. Paris.

iXrnaspury, B. F. 1914 On the so-called ultimobranchial body of the mamma-
lianembryo. Man. Anat. Anz., Bd. 47.

Maurer, F. 1899 Die Schilddriise, Thymus und andere Schlundspaltenderi-
vate bei der Eidechse. Morph. Jahrb., Bd. 27.

1899 Die Schlundspalten—Derivate von Echidna. Anat. Anz. Ergiin-
zungsheft. Bd. 16.

Moopy, R. M. 1912 Some features of the histogenesis of the thyroid gland in
the pig. Reprints of papers from the Dept. of Anat. of the Univ. of
Cal., vol. 4.

Prenant, A. 1894. Contribution 4 l'étude organique et histologique du
thymus, de la glande thyroide et de la glande carotidienne. La Cellule,
Ar, lO.

Rasy, H. 1913 Die Entwicklung der Derivate des Kiemendarms beim Meer-
schweinchen. Arch. f. mikr. Anat., Bd. 82.

Simon, Cu. 1896 Thyroide latérale et glandule thyroidienne chez les mammi-
feéres. Thése de Nancy.

VerpuN, P. 1898. Contribution A l’étude des dérivés Branchiaux chez les
vertébrés supérieurs. Thése, Toulouse.

PLATE 1
EXPLANATION OF FIGURES

1 Froma photograph of a portion of a transverse section of the caudal fourth
of the thyroid gland showing the left ultimobranchial body (U.). The colloid
in the thyroid gland is represented by black dots. Froma pig7.5 days old. X 60.

2 From a photograph of a portion of a transverse section of the caudal half
of the thyroid gland showing the more anteriorly located segment of the right
ultimobranchial body (U.). The white spots in the thyroid gland represent
follicles out of which the colloid has dropped. From a pig 15 days old. X 60.

3. From a photograph of a portion of a transverse section through the caudal
fourth of the thyroid gland showing the left ultimobranchial body (U.). From
pig 42 days old. X 60.

4 From a photograph of a portion of a transverse section of the posterior
fourth of the thyroid gland showing the right ultimobranchial body (U.?) and
two very large (cystoid) follicles (C.F.) which are filled with colloid. From
young adult hog no. 1. X 45.

ABBREVIATIONS

C.F., eystoid follicle; 7., thyroid; U., ultimobranchial body

PLATE 1

ULTIMOBRANCHIAL BODIES IN POSTNATAL PIGS

J. A. BADERTSCHER

Resumido por el autor, Frank Adam McJunkin.

El origen de las células mononucleares fagocitarias de la
sangre periférica.

Introduccién: La fagocitosis como una funcidn especializada
de ciertas células; naturaleza del carbon ingerido por fagocitosis
bajo condiciones experimentales; condiciones bajo las cuales
tiene lugar la fagocitosis del carb6n in vivo. Experimentos que
demuestran el origen de las células mononucleares fagocitarias
de la sangre, a expensas del endotelio de los vasos sanguineos.
Localizacién del carbon ingerido en los leucocitos mononucleares
fagocitarios de la sangre y células endoteliales de los vasos
sanguineos de pequeno tamafio; mitosis de células que contienen
carbon, en las paredes de los vasos sanguineos que no estan
alargandose; actividad fagocitica de las células mieloblasticas
jOvenes; probable origen de cierto ntimero de células mononu-
cleares fagocitarias de la sangre a expensas de la ttinica endotelial
de los vasos linfaticos. Conclusiones. Mediante una inyeccién
intravenosa simultanea de negro de humo y citrato sédico en
perros y conejos, el carbén es solamente ingerido por ciertas
células mononucleares de la sangre y por las células endoteliales
de los vasos sanguineos pequenhos. En la mitosis de las células
endoteliales con carbén ingerido, de los capilares de los érganos
solidos (sinusoides del higado), se encuentra una nueva prueba
del origen de dichas células sanguineas a expensas del endotelio
y de su penetracion en el torrente circulatorio. Aunque el 3 por
ciento de las células mononucleares que han ingerido carb6én se
encuentran en la sangre periférica unas pocas horas despues de la
inyeccion, tales células no se encuentran pasadas 24 horas. Este
hecho explica perfectamente el fracaso de los que han trabajado
con coloraciones intravitales para demostrar la existencia de
células tehidas por este método, en la sangre periférica.

Translation by Dr. José F. Nonidez,
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY
THE BIBLIOGRAPHIC SERVICE, DECEMBER 23

THE ORIGIN OF THE PHAGOCYTIC MONONUCLEAR
CELLS OF THE PERIPHERAL BLOOD!

FRANK A. McJUNKIN

Department of Pathology, Marquette University School of Medicine, Milwaukee,
Wisconsin

ELEVEN FIGURES (THREE PLATES)

CONTENTS
dig MINIS TIGGER SU Ran enna ne 27
a. Phagocytosis as a specialized function of certain cells............. 27
b. The nature of carbon ingested under experimental conditions by
FINA COC MU OST SPI ce ete aes taint SA cue ee che a RUE a oy uss o's. Sle 31

ce. Conditions under which Rititoc fore of carbon takes place in vivo 34
B. Experiments that show the origin of the phagocytic mononuclear cells of
the blood from the endothelium of blood-vessels................. 36
a. Limitation of ingested carbon to the phagocytic mononuclear
leucocytes of the blood and the endothelial cells of small b!ood-

WOES! IS Bis 6 A ci ED he ee bil de Ry ithe ti a ot fe ea ie oe ge 36
b. Mitosis of carbon-containing endothelial cells in the walls of

blood-vessels that are not lengthening.................. Ste 40
ce. The phagocytic activity of immature myeloblastic cells.......... 41

d. The possible origin of a certain number of the phagocytic mono-
nuclear cells of the blood from the endothelial lining of lymph-

OSES Soin ory htt Oa 0 co GO Dace Oe OTL CID ORE CILIA ET en a eee 42

Ch NORGE ONE: AES Ea LE ne ee ee a 45

A. INTRODUCTION
a. Phagocytosis as a specialized function of certain cells

In an earlier report by the writer (’18) about five per cent of
phagocytic mononuclear cells are shown to be present in normal
human blood. The purpose of the present paper is to trace
the origin of the phagocytic mononuclear cells found in the
blood from fixed tissue cells. Since the technic of the experi-
ments employed for this purpose as well as that of the earlier

1 A second report of Studies on the Mononuclear Cells of the Blood.
27

28 FRANK ADAM McJUNKIN

method used to demonstrate a non-lymphocytic group of mono-
nuclear blood-cells is based on the ingestion by the cells of
microscopic particles, a consideration of the phenomenon of
phagocytosis so far as the ingestion of carbon is concerned
becomes desirable.

Although Langhans (’70) called attention to the ingestion of
red-b'ood-corpuscles by the leucocytes in the tissues about
hemorrhages, it was not until about 1883 that general attention
was focused on the phagocytic activity of the cells of metazoa,
at which time Metschnikoff (’83) began his publications on the
subject. This activity of cells is often discussed in connection
with such questions as immunity, the physiological atrophy of
certain organs, removal from the tissues of mechanical irritants
such as coal-dust, and the solution of dead cells and cellular
debris. The ingestion of extraneous material is intimately
related to the phenomena of chemotaxis ahd intracellular
digestion.

-Among the protozoa as among the differentiated cells of
metazoa the ability to incorporate’ microscopic particles is
present in some (amebae) and absent n others (trypanosomes).
The determination of the phagocytic property of the cells present.
in higher animals is less simple than it is of protozo6én cells, but
experimentally microscopic particles may be brought into con-
tact with many different living tissue cells. In distinction to
true phagocytosis substances, such as bile pigment, hemoglobin,
melanin, lead, and many other products, present normally or
pathologically in the tissue may pass into the cells in solution
and there be deposited as microscopic matter. The entrance in
this manner of molecular and larger ultramicroscopic particles
(colloids) into cells does not come under the term of phagocy-
tosis which is a microscopic phenomenon of the cells. During
phagocytosis the cytoplasm flows about the foreign body probabi.
as the result of an alteration of surface tension (Wells, ’14) at
the point of contact, while molecules and small colloids pass
through the cell membrane by diffusion without any change in
the cell contour, and it is likely that the two processes have a
different explanation. The demonstration of microscopic foreign

- ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS 29

particles inside cells is not, therefore, certain evidence of phago-
cytosis. It appears further that foreign bodies may come to
lie within the cytoplasm of cells other than by the diffusion of
substances in molecular or colloidal solution and by phagocytosis.
The introduction of a foreign body into cells without the activity
of the latter appears to take place when leprosy bacilli pro-
liferate in epithelial structures such as the coil glands of the skin
(Mallory, 714). The writer has found rather large particles of
carbon in living cells twenty-four hours after the injection of
lampblack suspensions directly into the liver parenchyma. Here
as in the growth of leprosy bacilli in epithelial cells it appears
that the microscopic particles are mechanically forced into the
cell protoplasm. .

In considering phagocytosis in human tissue or in the tissue
of any metazoa the question of the differentiation and the
specificity of cells arises, because specific activities cannot be
assigned to cells that are not themselves specific and that may
change into entirely different cells under a varying environment.
Proof of any such transition of one type of cell to another, how-
ever, is lacking except in the case of a limited number of closely
related cells, and the very function of organs and tissues is
dependent upon the specificity of the cells which compose them
and which frequently undergo mitosis to replace dead cells.
Such unfailing reproduction of the same variety of cells can
result only from the existence of a living protoplasm specific for
that kind of cell, and once a definite physical and chemical unit
characterizing a cell has arisen in embryologic development it
certainly rarely changes. In the adult body many examples of
cell differentiation may be observed. The neutrophile dif-
ferentiates from a younger cell through well-defined changes,
but once differential characters have been acquired the neutro-
phile does not change to an eosinophile or a basophile.

A point of equal importance with the specificity of differen-
tiated cells is the identification of such cells by morphologic or
other characters. It is idle talk to speak of phagocytosis by a
fibroblast when the preparation on which the statement is hased
reveals none of the ‘earmarks’ which are known to characterize

30 FRANK ADAM McJUNKIN

this cell. In this particular case it has been stated by some that
only the younger cells manifest such activities, yet here the
characteristic fibrils must be present, for they appear in this cell
within twenty-four to forty-eight hours after mitosis, in which
time’ ingested particles could scarcely be disposed of. Ingested
foreign particles, such as carbon, or other cells, such as red
blood corpuscles, neutrophiles, or lymphocytes, have not been
demonstrated in fibroblasts stained in such a way as to bring
out the characteristic fibrils either in human or experimental
tissue.

In the blood the identification of cells for obvious reasons is
carried out under greater difficulties than elsewhere and tenta-
tive non-histological terms, such as transitional leucocyte, large
mononuclear leucocyte, etc., have been introduced. Had mor-
phological characters been more closely defined these terms
would have served a more useful purpose until such time as the
origin of the cells could be established, but no uniformity exists
in regard to the cells which are to be included under these terms.
The number of observations on normal and pathological bone-
marrow appears to warrant the statement as a fact that the
neutrophiles, eosinophiles, and basophiles (polymorphonuclear
leucocytes) differentiate from younger myeloblastic cells nor-
mally present only in the bone-marrow. A _ parallel state-
ment holds for the lymphocytes, and that the smallest cells of
the blood come from the lymphoid tissue is not seriously
questioned. There are present, however, non-lymphocytic mono-
nuclear cells of a different type and these are larger than the
lymphocytes and frequently have irregularities in the nuclear
contour. It is about these that the greatest doubt exists. The
significance of the experiments tabulated below calculated to
show the origin of phagocytic non-lymphocytic mononuclear
cells does not rest on the acceptance of any prescribed origin of
lymphocytes or polymorphonuclear leucocytes (neutrophiles,
eosinophiles, and basophiles), but is based rather on the speci-
ficity of the function of phagocytosis by a given cell under given
conditions. The only point urged here is that lymphocytes,
eosinophiles, basophiles, fibroblasts, most epithelial cells, etc., do

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS 31

not commonly ingest microscopic particles, and it does not seem
necessary to review the great mass of literature to prove that
phagocytosis is not a facultative function of most differentiated
cells of adult human tissue because such cells may readily be
brought into contact with foreign substances in vitro and in,
Vivo.

b. The nature of carbon ingested under experimental conditions
by phagocytosis

Although a considerable number of microscopic substances in
suspension, such as neutral oils, India ink, staphylococcus,
bacillus subtilis, and bacillus tuberculosis, were employed to
determine the phagocytic properties of cells in vivo, lampblack
was chosen as the most suitable for this purpose. All the
suspensions of this substance used show on examination with the
high-dry lens great numbers of microscopic particles which cor-
respond in size and shape to those in the cytoplasm of phago-
cytic cells placed in the suspensions. Since the particles of carbon
are small, the question of colloid solution and intravital staining
is raised.

Bouchard (’06) was the first to demonstrate the possibility of
staining living tissue intra vitam by the administration of certain
benzidine dyes in colloid solution. Recently many dyes, pyrrhol
blue, trypan blue, Janus green, and carmine solutions, some in
molecular solution and others in colloid solution, have been
injected into animals in normal and pathologic conditions to
determine the behavior of living cells toward them. Many
conflicting opinions in regard to the results obtained by intra
vitam staining have been advanced. It was found by H. M.
Evans (’15) that the endothelial cells lining small blood-vessels
were uniformly stained by the intra vitam method. His ob-
servations on the peripheral blood failed to show appreciable
numbers of cells containing the dye granules, and he says,
‘‘Macrophages do not occur in the peripheral blood stream except
when produced in numbers so great that the condition may be
called. pathological.” He defines a macrophage as a vitally
staining mononuclear cell and applied the term phagocytosis to

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 1

32 FRANK ADAM McJUNKIN

the ingestion of microscopic particles as well as to the ‘drinking
in’ of colloids (ultramicroscopic). In a small series of rabbits
injected with trypan blue and with a carmine solution used by
Kiyono (’13) practically the same results have been obtained by
the writer. The probable explanation of the absence of vitally
stained cells in the peripheral blood is given below. Conflicting
opinions have arisen about the specificity of the cells stained by
this method rather than in regard to the presence of the dye in
the endothelial lining of blood-vessels. Evans (’15) speaks of
macrophages other than the vascular ones which are more or
less fixed and probably of endothelial origin. This view of
tissue spaces lined with endothelium is scarcely in accord with
most work which shows that not only the blood-vessels, but the
lymph-vessels as well are closed endothelial lined. tubes (Mac-
Callum, ’03). y

In the first experiments with lampblack it was found that
this substance is removed from lampblack-citrate suspensions by
filtration through muslin, but that it passes through if blood is
added to the suspension before filtration. This corresponds in
general to what is known as protective colloid action, and later
it was found that gelatin has the same effect on the lampblack
as the protein of the blood and that lampblack-gelatin-citrate
suspensions not only pass through muslin, but even through
filter-paper. Microscopic examination with the oil-immersion
lens of the muslin filtrate reveals innumerable rather coarse
particles together with many minute particles at the limit of
vision. In the filter-paper filtrate there are no coarse particles,
but on examination with dark-field illumination innumerable
ultramicroseopic particles are visible. The lampblack-gelatin-
citrate suspensious after filtration through the finer filter-papers
do not separate even after standing a number of days and they
correspond in all respects to the typical colloid sols. The exact
way in which gelatin, albumin, and other protective substances
increase the permanency of colloid solutions is not agreed upon
by physical chemists. If such a colloid carbon suspension made
up largely of ultramicroscopic but partly of microscopic particles
is incubated in vitro with phagocytic cells instead of a suspension

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS oe

of much coarser carbon particles, the carbon taken up by the
phagocytes is invisible or only a few particles (those visible
microscopically in the sol) are seen in the cells. There is no
absolute limit on the maximum size of colloid particles, but
masses or lumps which settle out in a few minutes would hardly
be called colloid particles. Many at least of the dyes used for
staining intra vitam appear to fall in the class of colloid sols,
but the term colloid in this connection has been rather loosely
applied because, although there is no definite minimum size for
colloids and no clear separation between colloid and molecular
sol, no one should call a readily diffusible substance such as an
alkaline carmine solution a colloid. The point of greatest
interest is an explanation for the failure of those working with
intra vitam stains to demonstrate the presence in the blood of
numerous cells containing the dye granules, since cells which
have ingested microscopic particles of lampblack are readily
-demonstrated here in vivo and in vitro. In the animals receiving
intravenous injection of lampblack-citrate suspensions carbon-
containing cells appear in the peripheral blood within thirty
minutes, but at the end of five hours few remain in the blood
obtained in the usual way and after twenty-four hours no
carbon-containing cells are demonstrable. Examination of all
the tissues shows that owing to some change brought about in
the leucocytes by the presence of carbon particles in them they
are held in the capillaries of the liver, spleen, lungs, and other
organs, and that after twenty-four hours many are migratng
though the walls of the capillaries into the extravascular tissue.
There is every reason to think that vitally stained cells may
behave in the same way, and since the process of intra vitam
staining requires twenty-four hours or more granule-containing
cells do not appear in the peripheral blood. In the capillaries
even in the thinnest and most perfectly stained paraffin sections
it is not easy to differentiate between the cells present in the
capillaries and the tissue about them, and it would seem that
little accurate information could be obtained by the study of
frozen sections of such tissue.

34 FRANK ADAM McJUNKIN

Further, it was found that large quantities of lampblack must
be injected else most of it is removed from the plasma by
endothelial cells rather than by free cells in the blood. Again
as many as 90 per cent of the neutrophiles of the peripheral
blood contain lampblack after its injection, and in rabbits
especially the number of mononuclear carbon-containing cells
is very small after the intravenous injection of non-citrated'
carbon suspensions. By injecting sodium citrate at the same
time that the lampblack is injected the percentage of neutro-
philes with carbon is reduced to below one per cent and as many
as 3.03 per cent of phagocytic mononuclear cells have been
counted. It appears, therefore, to be necessary for the satis-
factory intra vitam demonstration of so-called macrophages in
blood not only to inject large amounts of microscopic particles,
but also to inhibit the active phagocytosis of the neutrophiles by
the use of sodium citrate.

c. Conditions under which phagocytosis of carbon takes place
in vivo

In a previous report (McJunkin, 718) it is shown that the
ingest on of carbon by certain cells in vitro may be controlled
by the concentration of sodium citrate, and since both polymor-
phonuclear leucocytes and the mononuclear ones ingest the
carbon it became a matter of importance to employ certain con-
centrations of sodium citrate in order to inhibit the activity of
the neutrophiles and to obtain a clear picture. It was found
that in vitro 15 mg. of this salt (Merck highest purity) per
cubic centimeter of blood almost completely inhibits the ingestion
by all cells under the conditions prescribed, while with 5 mg.
per cubic centimeter of blood not only the phagocytic mono-
nuclear cells, but more than half of the neutrophiles ingest the
carbon. By using 7.7 mg. sodium citrate per cubic centimeter
of blood the best differentiation of cells is obtained, since a small
percentage of neutrophiles ingested the carbon and practically
all of the phagocytic mononuclear leucocytes ingest it.

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS ah

Since publication of the first report it has been found that
even 40 mg. sodium citrate per cubic centimeter of blood does
not prevent the ingestion of large numbers of staphylococcus
pyogenes aureus when this organism in suspension is substituted
for the lampblack. By using strengths of citrate varying
between 5 and 15 mg. and substituting tubercle bacilli for the
lampblack, 2.2 per cent is the maximum number of macro-
phages containing the bacilli, while as many as 20 per cent of
neutrophiles have ingested the organisms. In a small number
of experiments with bacteria the limitation of phagocytosis to
the mononuclear cells has not been successful. Leucocytes
(polymorphonuclears and mononuclears) ingest the carbon in a
plasma-free medium, provided the carbon has previously been
brought into contact with fresh plasma or serum, although it has
subsequently been washed (four times) with saline. Washed
leucocytes, however, do not ingest untreated carbon.

In preliminary experiments heavy suspensions of lampblack in
1 per cent gelatin were injected intravenously into rabbits and
dogs with discouraging results so far as the blood is concerned,
for although polymorphonuclear leucocytes contained the carbon
in abundance, very rarely could a mononuclear leucocyte be
found which had ingested it. Only after repeated injections of
large amounts did any mononuclear carbon-containing leuco-
cytes appear in the peripheral blood of dogs.

The next step was to inhibit by the simultaneous ‘njection of
citrate the phagocytic activity of the polymorphonuclear leuco-
cytes just as it had been successfully done in the test-tube. The
amount of citrate to be injected was determined by estimating
the weight of blood at one-twentieth of the body weight and
injecting 7.7 mg. citrate per cubic centimeter of calculated
blood.

The exact way in which the citrate acts in the blood stream to
prevent phagocytosis is problematical, but there are indications
that it acts by diminishing the protective colloid action of the
albumin of the plasma and in this way allows flocculation of the
carbon so that it may be ingested by the leucocytes.

36 FRANK ADAM McJUNKIN

B. EXPERIMENTS THAT SHOW THE ORIGIN OF THE PHAGOCYTIC
MONONUCLEAR CELLS OF THE BLOOD FROM THE
ENDOTHELIUM OF BLOOD VESSELS

The points to be established by experiment are, first, the
limitation of the phagocytosis of carbon to certain mononuclear
cells of the blood and the endothelial cells lining the small blood-
vessels and, second, absolute proof of the latter’s becoming free
in the blood stream. There are other points discussed, such as
a diminished power of immature neutrophiles (metamyelocytes)
to ingest carbon, since they have a direct bearing on the question.

a. Limitation of ingested carbon to the phagocytic mononuclear
leucocytes of the blood and the endothelial cells
of small blood-vessels

The carbon (commercial lampblack ground for thirty minutes
in a mortar) Was given intravenously to the animals (table 1) as
a 5 per cent unfiltered suspension in | per cent gelatin. In both
rabbits and dogs the suspension was injected into a vein of the
left ear at the same time that the citrate was injected as a 10
per cent solution into a vein of the right ear. The lampblack
must not be thrown into the circulation too rapidly and the rate
of the citrate injection must be controlled by the dyspnea that
develops. Cover-glass films of blood were taken at varying
intervals from the ear employed for the citrate injection and
stained with polychrome blood stain. The films are made from
drops of fresh blood of the exact size required. In the peripheral:
blood of the rabbits (121, 138, 139, and 213) that received lamp-
black alone very few carbon-containing mononuclear cells are
present, although numerous polymorphonuclear leucocytes con-
tain it. Thus in rabbit 139 one hour after the sixth injection
less than 0.5 per cent carbon-containing mononuclear cells are
present and on previous days none are found. In the blood of
dog 151 four hours after the seventh injection of lampblack
alone 2.15 per cent mononuclear carbon-containing cells along
with 9.1 per cent carbon-containing polymorphonuclear leuco-
cytes are present. Thirty minutes after the fifth injection there

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS BY A

are present 2.66 per cent mononuclear cells with carbon and 8.85
per cent of polymorphonuclears containing it. Twenty-two and
one-half hours after the fifth injection no carbon containing
cells are present. One hour after the third injection no carbon-
containing mononuclear cells are found.

Rabbit 164 was the first to receive both citrate and lampblack,
and for the first time a distinct number of carbon-containing
cells are present in the peripheral blood of a rabbit after a single
injection. After application of the method used for determining
the number of phagocytic mononuclear cells in human blood

TABLE 1

DURATION DOSE OF TIME KILLED
NO. , ANIMAL Peeve oe DOSE OF CARBON CITRATE eT aee
1 Rabbit 1 da. 2.25 ec. once 0 24 hr.
138 Rabbit 3 da. 1.5 ec. three times 0 10 min.
139 Rabbit 7 da. 1.5 ce. six times 0 i aie:
151 - Dog 9 da. 10.0 ec. seven times 0 5 hr.
164 Rabbit 30 min. 5.5 ec. once 5.5 ec. | Not killed
176 Dog 1 da. 9.0 ce. once 9.0 ce. 24 hr,
200 Rabbit lene. 2.0 ec. once 4.25 ce. 1 hr.
201 Rabbit 30 min. 3.0 ce. once Gafolices 30 min,
205 Rabbit PD \ oie: 2.5 ec. once 5.0 ce. PB Sour
208 Rabbit 24 hrs. 1.0 ee. once 3.5 ec.| Dead 24 hr,
209 Dog 1 da. 3.0 ce. once S0mees 24 hr.
212 Rabbit 1 da. 1.75 ce. once 4.5 ce. 24 hr.
213 Rabbit 1 da. 1.5 ec. once. 0 24 hr.

1 Citrate contains opalescent suspension typhoid bacilli.

(McJunkin, 718) variations between 1.09 and 4.7 per cent of
these cells are found in the peripheral blood of rabbits. The
next animal to receive the lampblack and citrate was a dog
(176). The smears made at the end of one hour are not satis-
factory and an accurate count cannot be made.

A count of 595 cells in the peripheral blood of rabbit 200 taken
one hour after injection shows 3.03 per cent mononuclear cells
with carbon together with 2.02 per cent carbon-containing poly-
morphonuclear leucocytes. Of 1026 cells in the blood of rabbit
201 thirty minutes after injection 2.63 per cent are mononuclear
cells containing carbon with only 0.97 per cent polymorpho-.
nuclear carbon-containing ones. In rabbit 205 two hours after

38 FRANK ADAM McJUNKIN

injection there are 0.49 per cent carbon-containing mononuclear
leucocytes and 3.21 per cent polymorphonuclear ones. In rabbit
208 there are present five hours after injection 6.9 per cent
polymorphonuclear leucocytes with carbon, but no carbon-con-
taining mononuclear cells. In the blood of dog 209 no earbon-
containing cells are present twenty-four hours after injection.
Again in rabbit 212 no cells with carbon are present after
twenty-four hours.

The intravascular phagocytosis of carbon is readily produced.
After large injection of non-citrated carbon suspensions repeated
a number of times, as many as 2.15 per cent of the mononuclear
cells (dog) may contain carbon, and after single injections of
smaller amounts of carbon together with large doses of citrate
3.03 per cent mononuclear leucocytes may contain carbon two
hours after injection with the reduction of carbon-containing
polymorphonuclear ones to a minimum (less than 1 per cent).
The number of phagocytic cells in the peripheral blood varies in
different animals, reaching a maximum in one-half to two hours
with a very noticeable diminution in five hours and a total
disappearance of cells with carbon within twenty-four hours.

The disappearance within twenty-four hours from the periph-
eral circulation of all cells that have incorporated foreign
particles appears.to explain perfectly the failure of Evans (15)
and others to find vitally stained cells in the peripheral blood,
since the cells stained by the intra vitam methods contain
microscopic particles and considerable time (more than twenty-
four hours) and frequently multiple injections of the dye are
required.

The limitation of the ingested carbon to the. mononuclear .
cells of the blood (3 per cent in some animals) and to endothelial
cells lining the capillaries of various organs is of the nature of
a highly specific stain. In dog 209 (fig. 6) two or three and
less commonly more perfectly round carbon-containing mono-
nuclear cells may be found in the sinusoids of the liver which are
lined with elongated carbon-containing endothelial cells. The
» round contour of the former and an elongation due to ameboid
motion (fig. 5) is a differential point between the two. There

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS 39

are very few polymorphonuclear leucocytes (fig. 6) with ingested
carbon. At this time only an occasional carbon-containing cell
has migrated into the extravascular tissue, but the accumulation of
these cells in the sinusoids explains their disappearance from the
peripheral blood. After several days, as in dog 151, numerous
carbon-containing cells are present in the periportal tissue and
tubercle-like foci made up of carbon-containing mononuclear
cells about which there is a small amount of fibrin (fig. 7) have
formed in the sinusoids. In the liver of dog 209 and of other
one-day animals an occasional small focus of carbon-containing
mononuclear cells resembling the large foci of animal 151 may
be seen in the sinusoids. Wide areas of surrounding liver
parenchyma reveal no extravascular leucocytes, and the only
possible source of these cells is the blood. In the capillaries of
the spleen, bone-marrow, and lymph nodes and to a less extent
in the capillaries of other organs there is a similar accumulation
of leucocytes. In the one-day animals the reaction to the
carbon is entirely intravascular and is limited to free and attached
cells similar morphologically.

That most normal tissue contains practically no extravascular
phagocytic cells may readily be proved by injecting lampblack
into organs and fixing the tissue after the cells of the tissue have
had ample opportunity to ingest carbon, but before leucocytes
have had time to pass from the blood stream into the tissue
spaces. <A tissue or organ, such as the bone-marrow or spleen,
is not suitable for such an experiment, for it is difficult’ to de-
termine what cells are intravascular and what ones extravas-
cular. In the intestinal wall, the subcutaneous tissue beneath
the nipples and the kidney of animals 211 and 212 practically no
extravascular carbon-containing cells are present thirty minutes
and two hours after injection. There is a tendency to recognize
a close relationship or identity of cells to which such terms as
adventitial cells, gland connective tissue cells, areolar connective
tissue cells, reticulum cells, ete., and the fibroblast. Certainly
proof that. is positive and easily obtained shows that such
extravascular cells do not ingest by phagocytosis and do not
give rise to phagocytes. |

40 FRANK ADAM McJUNKIN

b. Mitosis of carbon-containing cells in the walls of blood
vessels that are not lengthening

The identical behavior to carbon of the phagocytic mono-
nuclear leucocytes of the blood and the phagocytic endothelial
cells of the capillaries is evident and observations to show that
most normal extravascular tissue contains practically no phago-
cytic cells may be readily made. That mononuclear phagocytes
of the blood arise from the endothelial cells lining the capillaries
may be demonstrated by a careful histologic technic. A desqua-
mation into the lumen of a vessel is frequently spoken of, but
there seems to be little evidence of such a process aside from the
formation of a new cell by mitosis. In routine autopsy exami-
nations an occasional mitotic figure may be found after prolonged
search in the endothelial cells of capillaries in the liver, pancreas,
and other organs. In the animals that received no typhoid
bacilli, such as rabbit 188 (fig. 3), prolonged search is required
to find a karyokinetic figure, and even the accumulation of
enormous amounts of carbon in the cells (dog 151) does not
increase the proliferation of the endothelial cells to any ap-
preciable extent. In animals receiving typhoid bacilli (dog 209)
mitoses (figs. 1, 2, and 4) may be demonstrated with much
greater ease. A careful histologic technic is required and demon-
stration of the achromatic spindle is necessary for certain in-
dentification, since the dense polymorphous nuclei of neutro-
philes, pyknotie nuclei of endothelial cells and even masses of
carbon may be confused with imperfect division figures, and
judging from the comparatively small number of the mitoses in
many of the animals receiving intravenous injections, such
mistakes have not been uncommon. ‘The tissues were removed
from the body immediately after the animal was killed and
pieces 2 mm. thick placed in Zenker’s fluid. After embedding in
paraffin, sections not more than 7 » thick were cut and stained
by the eosin-methylene blue method of Mallory.

The endothelial cells of the liver contain more carbon than
those of other organs and is well adapted for a search for mitoses.
In the liver, heart, kidney, and other solid organs in which the

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS Al

parenchymal cells are normal, cells resulting from mitoses neces-
sarily enter the blood stream since the capillaries cannot lengthen.

It is true that the dividing endothelial cells in the lengthening of
vessels in granulation tissue often remain more or less elongated
during division, but the demonstration of mitroses in vessels
that are not lengthening is more certain proof of the entrance of
these cells into the blood stream. A careful search in the liver
of animals one to three days after a lampblack-citrate-typhoid
injection shows the presence of mitoses in endothelial cells with
particles of lampblack in them (figs. 1 and 2). In an organ such
as the heart (fig. 3) in which the capillary wall is somewhat
thicker than that of the sinusoids of the liver the relationship
of the dividing cell is better defined.

c. The phagocytic activity of immature myeloblastic cells

The identification of cells in sections of bone-marrow stained
with eosin-methylene blue is less certain than in smears,
although the position of the cells in the tissue is helpful. The
relationship of bone-marrow structures is difficult to determine
and the exact location of the myeloblastic-cells as regards the
endothelial lin‘ng of vessels is not easy to make out even in the
best paraffin sections. In the animals receiving single doses of
both lampblack and citrate examination of sections of bone-
marrow shows little if any carbon in any cells except the endo-
thelial cells lining the blood-vessels and an occasional mono-
nuclear phagocyte in the lumen of the vessels. The only other
cells in which carbon may be found are small numbers of mature
polymorphonuclear leucocytes. The younger myeloblastic cells
(myelocytes and metamyelocytes) contain little carbon. On the
other hand in animals that have received lampblack alone the
carbon is present not only in the endothelial cells of the vessels,
but also in many of the polymorphonuclear leucocytes. In
these animals myelocytes are almost free of carbon. In the
cells that have become older and smaller with shrinkage of the
nucleus (so-called metamyelocyte) there is an occasional one
that has ingested the carbon.

In smears properly prepared the cells have the appearance

42 FRANK ADAM McJUNKIN

seen in cover-gloss films of the peripheral blood. To make the
smears about 5 gm. red bone-marrow crushed with forceps is
placed in a 15 cc. graduated centrifuge tube with 15 cc. saline
and shaken vigorously for five minutes, when the suspension is
filtered through thin muslin moistened with saline and the
. filtrate centrifuged. The supernatant fluid and as much fat as
possible is siphoned off. A small drop of the sediment is placed
on a cover-glass by means of a capillary pipette with rubber
nipple attached and cover-glass smears made in the usual way.
The mature polymorphonuclear leucoctyes in smears ‘made in
this way from dog 151 contain much carbon, but the immature
ones (metamyelocytes) and myelocytes contain little. The
comparison of the metamyelocytes with the phagocytic mono-
nuclear leucocytes present in smears of blood shows that the
cytoplasm of the former is more acidophilic with a marked
granulation, while its nucleus is more pyknotic and has the shape
of a three-quarter segment of a doughnut rather than a slight
to moderate indentation (kidney nucleus).

An attempt to prove diminished power of immature _poly-
morphonuclear leucocytes to ingest carbon by phagocytosis was
attempted by substituting the suspension of bone-marrow cells
for blood and performing the test in vitro in the usual way
(McJunkin, ’18) except 0.2 mg. lampblack is used per cubic
centimeter of the suspension and the citrate omitted. The
technic does not give entirely satisfactory results, but serves
to demonstrate that the phagocytic property of many polymor-
phonuclear leucoctyes is greater than that of the myelocytes
and metamyelocytes.

d. The possible origin of a certain number of the phagocytic
mononuclear cells of the blood from the endothelial
lining of the lymph-vessels.

Two methods of determining whether mononuclear phago-
cytes enter the blood from the lymp-vessels were utilized. The
first was to induce in the lymphatics changes that supposedly
accentuate the activities of the endothelial cells lining the

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS 43

lymphatic vessels and the second was to determine whether or
not phagocytic mononuclear cells are present in the lymph
entering the great veins of the neck. The first technic was
carried out by injecting a 5 per cent suspension of lampblack
in 1 per cent gelatin subcutaneously over large areas on the
lower portion of the abdomen, in the groins, and in the flexor
surfaces of the thighs. In guinea-pigs 211 and 212 (table 2) the
lampblack was introduced into the subcutaneous tissue, the
substance of the kidney, and the intestinal wall.

After repeated injections (dog 152) a large anthracotic iliac
node is present on either side of the vertebral column. Exami-
nation of these shows carbon in the endothelial cells lining the
lymph sinuses, but most of it is within large mononuclear cells
in the sinuses and in the extravascular spaces near them
(fig. 10). In an animal killed twenty-four hours after a single
injection (dogs 177 and 210) an occasional cell with carbon is
found in the sinuses, but none elsewhere (fig. 9). In the regional
nodes in the one-day animals there is no macroscopic blackening,
but with the lapse of a few days a sufficient number of carbon-
containing cells have arrived from the site of injection to cause
a blackening. Dilated lymphatic vessels in the subcutaneous
tissue into which carbon has been forced are lined with carbon-
containing cells (fig. 8), and there is little question but what these
divide to form free phagocytic cells. That mobilization of phago-
cytic mononuclear cells at the point of injection requires about
twenty-four hours (dogs 177 and 210) indicates that these cells
at first at least come from free cells in the blood stream, since
cells do not arise by mitosis in less than this time. The attrac-
tion of leucoctyes from the blood stream requires a number of
hours for very few have appeared after two hours.

Although the impossibility of phagocytosis by the lymphocytes
has not been seriously questioned, the resemblance of some
phagocytic cells to lymphocytes morphologically and the many
conflicting statements on the subject have led to a close study
of this phase of the subject. The distinct separation of lymph-
vessels and lymphoblastic tissue has been referred to in fore-
going paragraphs. The lymph-vessels like the blood-vessels are

44 FRANK ADAM McJUNKIN

closed channels lined with endothelial cells which have nothing
in common with the lymphoblastic cells. In the lymph nodes,
therefore, there are present two varieties of endothelial cells
(those lining blood-vessels and those lining lymph-sinuses) and
the lymphoblastic cells. Twenty-four hours after a single intra-
venous injection (dog 176) the endothelial cells lining the capil-
laries are the only cells that contain carbon. Even in the
germinal centers made up of large proliferating lymphoblasts
carbon is present in the endothelial cells of the capillaries
(fig. 11). Later (dog 151) carbon-containing cells migrate into
the lymphoblastic tissue, and then there is an admixture of
carbon-containing cells and lymphoblasts. This origin of phago-
cytic cells from blood-vessels in the germinal centers has been
lost sight of by some.

TABLE 2

el HOURS KILLED
NO. ANIMAL oe eae, DOSE OF CARBON AFTER

r more LAST INJECTION

130 Rabbit 6 da. 5 ec. four times 24 hours
152 Dog 10 da. 10 cc. five times * 24 hours
177 Dog 1 da. 40 ce. once 24 hours
210 Dog 1 da. 30 ce. once 24 hours
211 Guinea-pig 30 min. 5 ec. once 30 min.
212 Guinea-pig 2 hrs. 5 ec. once 2 hours

In (dog 151) a deeply pigmented node was found in the retro-
peritoneal tissue below the liver. The structure of this gland
strongly suggests a hemolymph gland. With this exception, no
markedly blackened nodes have been found in animals receiving
intravenous injections.

In one etherized animal (dog 152) lymph was obtained from
the thoracic duct and no cells (count of 500) containing carbon
are present. Owing to irregularities in the course of the lymph-
vessels, some difficulty has been encountered in freeing the
vessels near their entrance into the jugular or subclavian, but
once dissected free collection of the lymph is easy. Coagu-
lation is prevented by 5 mg. citrate per cubic centimeter and the
fluid centrifuged. Smears from the sediment are fixed in
alecohol-ether to free them from fat and stained in the usual way.

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS 45

This observation scarcely proves that phagocytic cells do not
enter the venous circulation by way of thoracic duct, but ceér-
tainly large numbers do not travel this route for the reaction in
the lymph glands of this animal is marked. It seems more likely
that a certain number of phagocytic cells arising from the endo-
thelial lining of lymph-vessels pass directly into the blood
stream by migration.

C. CONCLUSIONS

For studying the phagocytic activity of cells lampblack has an
advantage over intra vitam stains, since the reaction is almost
immediate. By showing that cells which have ingested carbon
(3 per cent mononuclear ones) completely disappear from
the peripheral circulation within twenty-four hours, the fail-
ure of those employing intra vitam stains to demonstrate
any considerable number of phagocytic cells in the blood is
explained. Another advantage of lampblack over intra vitam
stains is that thin paraffin sections may be made of the tissue
fixed in a way best calculated to bring out the histologic struc-
ture. ‘In such sections the limitation of the ingested carbon to
the endothelial cells and mononuclear leucocytes assumes the
character of a specific reaction for cells of endothelial origin.
The demonstration of mitoses in endothelial carbon-containing
cells lining capillaries which are not lengthening is proof of the
origin of phagocytic mononuclear blood-cells from the endo-
thelial cells of vessels.

That the five per cent of phagocytic mononuclear cells present
in human blood consists entirely of endothelial leucoctyes seems
likely in view of the fact that there is positive proof of this
origin and negative evidence showing that they are not of
myeloblastic or lymphoblastic origin and that few phagocytic
cells of any variety are present in normal extravascular tissue.

For the incentive to undertake these experiments, it affords
me great pleasure to acknowledge my indebtedness to Prof.
F. B. Mallory, who has for a great many years taught his
students the important rédle of the endothelial leucocytes. I
am indebted to Miss Alice Charlton for valuable technical
assistance and to Mr. L. Massopust for the illustrations.

46 FRANK ADAM McJUNKIN

BIBLIOGRAPHY

Boucuarp 1906 Ann. de l’Inst. Pasteur, vol. 20.

Evans, H. M. 1915 Am. Jour. Physiol., vol. 37.

Kryono, K. 1914 Folia Hemtalog., vol. 18.

Lanauans, TH. 1870 Virchow’s Archiv, vol. 49.

MacCatuum, W. G. 1903 Johns Hopkins Hosp. Bull., vol. 14.

Mattory, F. B. 1914 Principles of pathologic histology. W.B. Saunders Co.,
Phila.

McJunkin, F. A. 1918 Archives of Internal Medicine, vols. 21 and 22.

Mertscunikorr, EH. 1883 Arb. a. d. Zool. Inst. Wien, Bd. 5.

We tts, H. G. 1914 Chemical pathology. W. B. Saunders Co., Phila.

PLATES

47

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 1

The sections from which the drawings were made are paraffin sections of
Zenker-fixed tissue stained by the eosin-methylene-blue method of Mallory.
The drawings were made with the aid of a camera lucida.

PLATE 1
EXPLANATION OF FIGURES
Oil-immersion objective and no. 4 ocular (Leitz).

1 Monaster showing achromatic spindle in an endothelial cell containing
carbon in a sinusoid of the liver (dog 209, table 1).

2 ‘Two monasters (liver dog 209, table 1).

3 Diaster in capillary of heart muscle (rabbit 138, table 1). Note carbon
particles in cytoplasm.

4 Monaster in an endothelial cell of a vessel (portal vein?) in a portal tract
at a point where the vessel narrows and enters the liver parenchyma (dog 209,
table 1).

48

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS PLATE 1
FRANK ADAM MC JUNKIN

EZ LTASS UST 28

49

PLATE 2
EXPLANATION OF FIGURES
Oil-immersion objective and no. 4 ocular (Leitz).

5 Migrating phagocytic mononuclear leucocyte containing carbon in a
sinusoid of the liver (dog 209, table 1).

6 Four phagocytic mononuclear leucoctytes (three with carbon) and two
polymorphonuclear leucocytes (no carbon) in a sinusoid of the liver (dog 209,
table 1).

7 Tubercle-like focus consisting of carbon-containing mononuclear leuco-
cytes in sinusoid of liver (dog 151, table 1). A distinct fibrin network between
the leucocytes.

50

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS PLATE 2
FRANK ADAM MCJUNKIN

* e {“g.)
apg: a \
‘S eoP ee WY
aly, ew fi }
e 7

ui
Oo

oy . > r9
a (ah EO

LMASSOPUST 78.

a

PLATE 3
EXPLANATION OF FIGURES

Oil-immersion objective and no. 4 ocular (Leitz) except fig. 8 which was made
with the high-dry lens (no. 7, Leitz) and fig. 10 which was made with the low-
power objective and no. 2 ocular (Leitz).

8 Carbon-containing endothelial cells lining a dilated lymph-vessel in the
subcutaneous tissue ten days after first subcutaneous injection (dog 152,
table 2).

9 Carbon-containing phagocytic mononuclear leucocyte and one containing
a polymorphonuclear leucocyte in the peripheral sinus of a retroperitoneal
lymph node one day after subcutaneous injection (dog 177, table 2).

10 Lymph nodule from a retroperitoneal lymph node ten days after subcu-
taneous injection (dog 152, table 2). Most of the carbon is in phagocytic mono-
nuclear leucoctyes in and near the sinuses.

11 Blood-vessel at the periphery of a germinal center (retroperitoneal lymph
node, dog 151, table 1). The endothelial lining shows carbon. <A mitotic figure
and two large lymphoblasts at the right of the vessel.

PLATE 3

53

»

FRANK ADAM MC JUNKIN

°PUST: 18.
10

LITASS

ORIGIN OF PHAGOCYTIC MONONUCLEAR CELLS

Resumido por el autor, Harold Dickinson Senior.

Sobre el desarrollo de las arterias de la extremidad inferior
del hombre.

La arteria axial primitiva nace de la raiz dorsal (secundaria) de
la arteria umbilical; perfora el plexo lumbosacro y sigue el
trayecto del nervio tibial a lo largo del muslo. Por debajo de la
rodilla pasa primero entre el musculo popliteo y la tibia y de-
spués entre el mtsculo tibial posterior y la membrana interésea.
Entra la planta del pié pasando por detras del maleolo medio y
se ramifica para constituir la red plantar y ramas que atraviesan
el seno tarsal para formar la red dorsal. Algunas partes de la
arteria axial persisten para formar en el adulto la arteria glitea
inferior, anastomosis postfemoral, porcién proximal de la arteria
poplitea, raiz de la arteria articular infero-media, una pequefia
parte del trayecto de la arteria tibial anterior, toda la tibial
recurrente posterior y una pequena porcién de la arteria peronea.
En el miembro se desarrollan muchas arterias secundarias. La
arteria iliaca externa nace de la umbilical y se divide en la
epigdstrica inferior y la femoral. Dos de las arterias secundarias
de la pierna se combinan con una porcién de la axial para formar
la arteria peronea y otras tres arterias se combinan para formar
la arteria tibial anterior. Porciones de dos arterias se mezclan
para formar la parte distal de la arteria poplitea y partes de
estas mismas arterias entran en la planta del pié para transfor-
marse en las arterias plantares. Las otras arterias del pié se
derivan de las redes embrionarias.

Translation by Dr. José F. Nonidez,
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 23

THE DEVELOPMENT OF THE ARTERIES OF THE
HUMAN LOWER EXTREMITY

H. D. SENIOR

Department of Anatomy, New York University

ELEVEN FIGURES
I. INTRODUCTION

a. General considerations

It is well known that considerable differences exist between
the various types of arterial distribution which are normally
encountered in the pelvic limb of the different representatives
of the mammalian series. In all mammals in which the de-
velopmental history of the arteries of the limb has been investi-
gated, however, it has been found that the adult arterial system
of the part results from the elaboration of two embryonic vessels.

Both of the embryonic arteries in question take origin from
the dorsal (secondary) root of the a. umbilicalis. One of them,
the primitive artery of the limb, traverses the axis of the thigh
and leg and divides into a number of branches for the supply
of the foot. This vessel is present before the limb forms a
distinct prominence upon the surface of the body and is known
as the axial or ischiadic artery.

The other is an artery of later development which traverses
the pelvis and the ventral region of the thigh. It joins the
axial artery a short distance above the knee and gives rise to the
a. iliaca externa, the a. femoralis, and the a. epigastrica inferior
and to the branches which arise from these vessels. All other
arteries of the limb arise from the axial artery itself or from the
branches to which it gives origin.

Whether the relations of the embryonic axial artery to the
other constituents of the limb are identical in the embryos of

55

56 H. D. SENIOR

all mammalian forms is a question to which direct observation
of the vessel has not yet furnished an answer. Much indirect
evidence bearing upon this subject is furnished by Zuckerkandl’s
comparative study of the arteries of the leg which appeared in
1895.

The study in question, although it is mainly concerned with
the relations of the adult arteries, is influenced throughout b4
embryological considerations. It includes, in fact, a study of
the arteries in a series of vertebrate embryos in which the
mammalia are represented by the rabbit and cat. Zucker-
kandl’s work is obviously based upon the conception of the
identity of the course of the axial artery throughout the mamma-
lian series. It may be said that the general results of the
investigation are greatly in favor of the correctness of the author’s
conception.

The principal relations of the axial artery of the human
embryo have been established by the present study. An ade-
quate account of the relations of this and other arteries of the
developing limb of one of the quadruped mammals would be of
great value for purposes of comparison.

The only consecutive account which has been given of the
development of the arteries of the pelvic hmb in any mamma-
lian form is that of DeVriese, which appeared in 1902. It deals
with the human embryo. The other papers contained in the
literature of the subject are concerned with isolated stages in the
development of one or more forms rather than with a complete
history of the arteries of any particular mammal.

Hochstetter showed for the first time, in 1890, that the primi-
tive artery of the mammalian thigh, which he termed the a.
ischiadica, follows the course of the n. ischiadicus in the embryos
of both the cat and the rabbit. He also described the develop-
ment of the a. femoralis, which, appearing at a later stage,
supersedes the proximal part of the a. ischiadica as the chief
artery of the thigh. The secondary assumption of the original
function of the a. ischiadica by the a. femoralis was shown to
oecur in man by DeVriese in 1902.

=

ARTERIES OF HUMAN LOWER EXTREMITY of

Hochstetter did not succeed in following the continuation of
the a. ischiadica through the leg. Leboueq, however, described
it in 1893, as pursuing, in the human embryo, an axial course
between the tibia and fibula and finally perforating the tarsus
to reach the dorsum of the foot.

In 1894, Zuckerkandl described the continuation of the a.
ischiadica of the rabbit as traversing the flexor region of the leg
and dividing into a number of branches for the supply of the
sole. He also described two branches of the artery which supply
the extensor aspects of the leg and foot, respectively. In 1895
the same author described the perforating artery of the tarsus in
eat and rabbit embryos and made considerable progress in
the nomenclature to the primitive artery of the limb.

In Zuckerkandl’s second paper a distinction was made between
the terms axial and ischiadic which formerly had been used more
or less interchangeably to denote the primitive artery of the
limb. For the entire artery Zuckerkandl employed the name
axial. He restricted the use of the term ischiadic to the part of
it which traverses the thigh and used the term a. interossea for
the remainder of the vessel. The same paper contained the
first accurate description which had been given of the distal part
of the axial artery of any mammal. The a. interossea was de-
scribed in the rabbit as lying between the interosseous membrane
and m. tibialis posterior.

Grosser, in 1901, and DeVriese, in 1902, described the a.
interossea (the a. nervi interosseil cruris of the latter author)
in bat and human embryos, respectively, as pursuing the course
it had been described by Zuckerkandl as following in the rabbit.
Both of these observers also recorded the presence, in the
subjects of their respective studies, of the r. perforans tarsi.

During the course of the present investigation it has become
apparent that the existing literature contains no definite state-
ment regarding the course taken by the part of the axial artery
which traverses the popliteal fossa. It is questionable whether
the popliteal section of the axial artery has been tacitly included
as a part of the a. ischiadica or not, since the distal limit of the
latter artery does not appear to have been definitely fixed. In

58 H. D. SENIOR

the following description the term a. ischiadica has been re-
stricted to the part of the axial artery proximal to the site at
which it is subsequently joined by the a. femoralis. This
restriction has necessitated the use of a new term for the part
of the axial artery which extends from the hiatus tendineus
(the approximate site of the femoro-ischiadic junction) to the
point (in the neighborhood of the distal border of the m.
popliteus) at which the a. interossea begins.

An examination of the popliteal portion of the human axial
artery shows that its distal part does not lie upon the posterior
surface of the m. popliteus as does the distal part of the adult
a. poplitea, but upon the anterior surface of that muscle. It
seems clear, therefore, that the distal part of the a. poplitea of
the majority of adult mammals, which lies upon the anterior
surface of the m. popliteus represents a persisting portion of the
embryonic axial artery. On this account the name a. poplitea
profunda has been used in the following pages to designate
entire popliteal section of the human axial artery.

The solution of Zuckerkandl’s difficulty regarding the differ-
ence of the mutual relations between the a. poplitea and the m.
popliteus in man on the one hand and’in the majority of other
mammals upon the other has thus been furnished by determining
the course of the popliteal portion of the axial artery of the
human embryo.!

The part of the adult human a. poplitea which extends from
the hiatus tendineus to the origin of the a. genu inferior
medialis is a direct survival of the embryonic a. poplitea pro-
funda. The part of that artery which lies upon the posterior
surface of the m. popliteus is derived from an embryonic vessel
of later formation referred to in the following pages as the a.
poplitea superficialis.

1 Arteria poplitea. Ein Stiick dieser Arterie lagert bei den Halbaffen und den
Primaten auf der freien (dorsalen) Fliiche des Musculus popliteus, bei den iibrigen
Siugetieren auf der Gelenkskapsel, bedekt von dem oben erwaihnten Muskel.
Diese Verschiedenheit im Verlaufe der Poplitea kann nur auf die weise erklirt
werden, dass entweder die den Muskel querenden Stiicke der Poplitea nicht
homolog oder die beiden Muskeln nicht dieselben sind. Nach meinen bisherigen

Erfahrungen scheint ersteres wahrscheinlicher zu sein. (Zuckerkandl (’95), p.
255.)

ARTERIES OF HUMAN LOWER EXTREMITY 59

The literature dealing with the arteries of the mammalian
pelvic limb shows a tendency towards the perpetuation of a
conception regarding the relation of the adult human a. peronaea
to the embryonic a. interossea which Zuckerkandl has already
shown to be erroneous. Stieda stated, in 18938, that the study
of the variations of the arteries of the human leg had led him to
the conclusion that the a. peronaea represents a persisting portion
of the embryonic axial artery. A similar opinion regarding the
relationship between these two arteries was expressed by
Zuckerkandl in 1894. In 1895 the recognition of the course of
the a. interossea led Zuckerkandl to a modification of the views
he had previously expressed upon the subject. He drew atten-
tion to the fact that it would be impossible for the part of the
a. peronaea which is separated from the interosseous membrane
by the m. tibialis posterior to be a derivative of the part of the
a. interossea which lies between the membrane and the muscle.
In 1902 DeVriese revived the conception of the identity of the
aa. interossea and peronaea, notwithstanding the obvious justice
of Zuckerkandl’s contention.

A careful examination of the embryos which form the basis of
the present study and of a large number of others belonging to
the collection of the Carnegie Institution has been made in the
search for a r. saphenus of the a. femoralis comparable in extent
to that of the a. saphena which occurs very rarely in adult man
and invariably in most of the other mammals. The result has
been entirely negative.

That the a. saphena occurs occasionally in the human embryo
is indicated by the fact that its presence has been noted at least
five times in the adult. In the embryos examined in connection
with the present study, however, the r. saphenus has not been
found to extend in a single instance beyond the middle of the
leg.

I take this opportunity of expressing my gratitude to Professor
Thyng for the trouble he has taken in the revision of this
manuscript.

60 H. D. SENIOR

b. Material

In the present investigation the lower extremities of embryos
have been studied in seven stages of development, one extremity
or both having been reconstructed in wax. The selection of
stages depended upon the accessibility of well-preserved material
rather than upon a preconceived plan regarding the most in-
structive stages to use.”

The following embryos have been studied, the right limb
having been reconstructed in all cases. The embryos of which
both lower limbs have been reconstructed are marked with an
asterisk :

6.0 mm.* Carnegie Institution, Embryological Collection (C.I.E.C.) No.
1075.

8.5 mm.* Cornell University, Embryological Collection (C.E.C.) No. 9.

12.0 mm.* Cornell University, Embryological Collection (C.E.C.) No. 3.

12.0 mm.* Minnesota, Embryological Collection (M.E.C.) No. H. 16.

14.0 mm. Cornell University, Embryological Collection (C.E.C.) No. 5.

17.8 (?) Harvard University, Embryological Collection (H.E.C.) No. 839.

18.0mm. Carnegie Institution, Embryological Collection (C.I.E.C.) No. 409.
22.0 mm. Cornell University, Embryological Collection (C.E.C.) No. 1.

For their generosity in placing their material at my disposal,
I wish to express my great obligation to Profs. C. M. Jackson,
F. T. Lewis, F. P. Mall, C. R. Stockard, and G. L. Streeter.

Toward the end of the investigation, there were a number of
difficulties, for the solution of which the examination of rather
close intermediate stages was required. Such stages were found
in the collection of the Carnegie Institution at Baltimore.

In the formation of this valuable collection, so ably cared for
and sympathetically administered, Professor Mall has made
available to his fellow-workers a wealth of material adequate for
the solution of any ordinary problem in human embryology.

The reconstruction of vascular plexuses in wax, using every
second or fourth section as the case may be, is somewhat difficult.
The practice followed has been to unite the parts of the adjacent

2 In the case of the embryos from the Carnegie Institution and Minnesota
University the measurement is crown-rump. The Harvard and Cornell measure-
ments represent the greatest total length.

ARTERIES OF HUMAN LOWER EXTREMITY 61

plates which fit after careful adjustment and to remove those
which do not join. In this way the plexus represented in the
reconstruction is probably less dense than that occurring in the
embryo. The reconstruction, in fact, reproduces the spirit rather
than the letter of the original.

In reconstructing the stage of 6 mm. consecutive sections
were used, and these were comparatively thick (20 »). In this
case the parts usually fitted so accurately that the plexuses,
as reproduced, must represent, as nearly as possible, the actual
conditions in the original.

c. Nomenclature

As already noted, the term a. poplitea profunda is used in the
following account to denote the popliteal section of the em-
bryonic axial artery. The term a. ischiadica and a. interossea
have been retained for the proximal distal parts of that artery,
respectively.

To the artery which normally perforates the tarsus of adult
ungulates, and which has been recognized in all mammalian
embryos hitherto observed, numerous terms have been applied.?
The name ramus perforans tarsi is used here.

For two of the embryonic arteries the names used by Hyrtl in
1864 have been retained. They are the r. coronarius (of the
medial malleolus) and the a. peronaea posterior superficialis.

The description of a number of embryonic vessels, the existence
of which has not been noted heretofore, has necessitated the use
of several new terms. ‘These conform, as nearly as may be, with
current usage.

In referring to the relative positions of the structures of the
limb, the adult terms of orientation have been used throughout.
This course has been adopted in order to avoid the confusion
which might arise from the alternative use of two sets of terms
in making comparisons between the relative positions of struc-
tures in the adult and embryonic limb, respectively.

’ Ramus and sinum tarsi, Hyrtl (’64); Arteria tarsea perforans, Siissdorf
(89); perforans tarsi, DeVriese (’92) Arteria anastomotica tarsi, Salvi (’99).

62 H. D. SENIOR

The limb preserves its primitive position which, with the
exception of progressively increasing flexion of the knee, remains
unchanged throughout the period of development under con-
sideration. The flexor aspect of the embryonic thigh and leg
and the plantar aspect of the sole are directed medially; the
great toe is preaxial or cephalic. The term posterior for the
adult has, therefore, the same significance as the term medial for
the embryo; so have, respectively, the terms anterior and lateral,
medial and pre-axial, lateral, and post-axial. For the parts
above the hip-joint there can be no possibility of confusion, since
like terms of orientation serve equally well for the postnatal and
embryonic periods.

Il. THE ARTERIAL SYSTEM OF THE LOWER LIMB IN PROGRESSIVE
STAGES OF DEVELOPMENT

a. stage of 6mm. C. 1. EK. Cy no. 1075, figs, and OA:

In the lumbar region the nerve roots are not recognizable. The
ganglia of the lumbar and sacral regions appear as segmentally
disposed swellings upon the continuous neural crest. Distal to
the second lumbar segment the postcardinal vein becomes plexi-
form. The medial part of the plexus receives the segmental
veins, the lateral part represents the still indefinite marginal
vein.

The dorsal segmental arteries, with the exception of the fifth
lumbar and second sacral, pass directly to the spinal cord with-
out branching. The umbilical arteries, the ventral roots of
which are still very large, arise opposite the intervals between
the third and fourth lumbar segmentals.

The secondary, or dorsal, roots of the umbilical arteries are
present, but are smaller than the original or ventral roots of that
vessel. Each dorsal root, in the embryo under consideration,
arises from the union of two arteries.

The chief share in the formation of the dorsal root of the a.
umbilicalis is taken by a vessel which arises from the fifth
lumbar segmental artery a short distance beyond its root. The
vessel in question is joined almost perpendicularly near its
origin by a smaller one which springs from the aorta in the
interval between the fourth and fifth lumbar segmental arteries.

ARTERIES OF HUMAN LOWER EXTREMITY 63

The dorsal root of the a. umbilicalis, which may be said -to
begin at the junction of the two vessels mentioned above, con-
tinues the transverse direction of the larger of the two, until it
reaches the dorsal aspect of the Wolffian duct of its own side. In
this situation it turns ventrally, passing upon the lateral side
of the duct, which is now enclosed between the two roots of the
umbilical artery, to join the ventral root.

As the dorsal root of the a. umbilicalis curves around the
lateral side of the Wolffian duct it gives origin to the axial artery
of the lower extremity and to the a. pudenda interna.

The axial artery, takes an almost transverse, slightly recurrent,
course towards the surface. It ends by dividing into two

plex. abdom. ee

g. umbil. : :
plex. ae —

ad. OxXx/s.

radix vent.

a. umbilic.
radix dorsalis.

Fidel

Fig. 1 Reconstruction showing the distal end of the aorta; also the arteries
of the right lower extremity and neighboring parts in a human embryo of 6
mm. C.J.E.C., 1075). Medial aspect. X 40 diams.

branches, each of which breaks up into a plexus which passes
over into the postcardinal venous plexus. The a. pudenda
interna follows the dorsolateral surface of the Wolffian duct
towards the urogenital sinus and enters the pelvic arterial plexus.

In addition to the arteries already described, there are two
extensive arterial plexuses, which may be called the abdominal
and pelvic plexuses, respectively. Of these the abdominal
arises by seven or eight stems from the concavity of the umbilical
artery somewhat distal to the junction of its two roots. It
invades the flexor region of the thigh, although it is uncertain
to what extent, for the lower limb bud is not distinctly circum-
scribed at this period.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 1

64 H. D. SENIOR

The abdominal plexus has no connection whatever with the
segmental arterial series. It is drained partly by postcardinal
plexus, but mainly by the umbilical vein. The pelvic plexus arises
by two roots from the convexity of the umbilical artery opposite
the roots of the abdominal plexus. It covers the cloaca and
receives the pudendal artery and a large branch from the second
sacral segmental. It is drained by the postcardinal plexus.
The abdominal and pelvic plexuses appear to be connected by
a few vessels passing lateral to the umbilical artery: It is
probable that all the visceral branches of the adult a. hypo-
gastrica (excepting the superior vesical) arise out of the primitive
pelvic plexus of the original umbilical artery. The parietal
branches (as far as they are present before the stage of 22 mm.)
arise from, or very near to, the axial artery. They are therefore
connected with the dorsal root.‘

This stage agrees in most essentials with the 5- -mm. embryo
(II) described by Tandler (’03). The inferior mesenteric artery
arises opposite the first lumbar segmental.

b. Stage of 8.5 mm., C. U. C., no. 9, figs. 2 and 9 B

The femoral, obturator, genitofemoral, and peroneal nerves
are readily recognized as short relatively unbranched trunks.
The extremity of the tibial is divided into what appear to be
the medial and lateral plantar nerves.

The aa. umbilicales have lost their original (ventral) roots
and now spring from the aorta in the intervals between the
fourth and fifth dorsal pairs of lumbar segmental arteries.» The

4 The distinction between the respective origins of the visceral and parietal
branches of the a. hypogastrica has already been pointed out by Wes Mle
(704).

5 It is questionable whether the eommon iliac artery should be regarded as
belonging to the fifth lumbar dorsal segmental or not. _ At the stage of 8.5 mm.
of this series the fifth lumbar segmental has regained its independence, and
springs from the aorta (on both sides) between the fourth and fifth. The work
of Levy (’02) shows, however, that absence of the a. lumbalis ima is the rule
rather than the exception. Whether this vessel is frequently retained by the
root of the a. iliaca communis or merely overshadowed (as it were) by it is
doubtful.

ARTERIES OF HUMAN LOWER EXTREMITY 65

fifth lumbar segmental arteries have thus regained their inde-
pendence, and, like the other segmental arteries in this region,
pass to the spinal cord without branching.

The axial artery passes distally into the lower limb, crossing
the n. tibialis posteriorly from the medial to the lateral side.
It then follows the posterior aspect of the skeletal mesenchyme
throughout the leg into the sole. In the latter situation it
breaks up into a flattened plexus which extends somewhat beyond

a. iliac. ext.
@. iliac. com.

a. hypog. _n genitofem.
rr perfor. tarsi. n. femor.
@. GXIS.

rete dorsal
rete plant~ n pe
a pudend int. a. femor. a. epigast. inf.
r. Saphen.
2 @. axis... Com. Sup—_~=5 7}:
r. perf, tarsi.

rete dorsal.

n. peron.

Fig. 2 Reconstruction showing the arteries of the right lower extremity in a
human embryo of 8.5 mm. (C.E.C., 9). Medial aspect. 20 diams.

Fig. 3 Reconstruction showing the arteries of the right side of the pelvis,
right thigh, leg, and dorsum of the foot in a human eubrye) of 12mm. (M.E.
C.,H. 14). Medial aspect. X 20 diams.

the extremities of the plantar nerves. Just as it is breaking up
into the plantar plexus the axial artery gives origin to two or
three branches which pierce the mesenchymal skeleton of the
foot to reach the dorsum. Upon the dorsum of the foot the
perforating branches produce another flattened plexus. The
two plexuses (or retia) of the foot are separated from one another
by the mesenchymal skeleton of the foot, the plantar rete lying
between the latter and the plantar nerves.

66 H. D. SENIOR

In passing from the medial to the lateral side of the sacro-
pudendal plexus, the axial artery runs between the main plexus
and a branch from its lower part. The branch referred to has
been removed from the reconstructions illustrated in figures 1
to 6, since it partially covers the artery when viewed from the
medial side.

From the lateral aspect of the concavity of the umbilical
artery, some distance proximal to the origin of the axial artery,
there now arises a new vessel, the a. iliaca externa. This vessel
takes a cephalic direction, nearly parallel to the aorta, coursing
medially to the origin of the obturator nerve. At this stage
the wall of the external iliac artery is thin and its course slightly
tortuous. The artery is not connected with the dorsal segmental
arterial series or with any other artery. The further history of
the external iliac is rather remarkable. It soon becomes quite
straight and acquires a wall of great thickness, but remains un-
branched until a stage of (approximately) 12 mm. The a. iliaca
externa appears as a thick-walled straight artery in the well-
known pig embryo of 12 mm.

The place of origin of the a. iliaca externa marks the permanent
subdivision of the ‘dorsal root of the umbilical artery into two
parts. The proximal part becomes the adult a. iliaca com-
munis while the distal part represents the a. hypogastrica and a
short proximal section of its umbilical branch.

At this stage the marginal vein is fully formed; its caudal
root, represented by the v. ischiadica, runs in close contact with
the proximal part of the corresponding artery.

c. Stage of 12 mm. M. E. C., no. H. 16, figs. 3 and 9 C

Condensation of the mesenchymal skeleton is now well ad-
vanced. It is not sufficiently definite in the tarsal region, how-
- ever, for the determination of the course taken by the connection
between the axial artery and the dorsal rete of the foot. The
surfaces of the growing nerves are considerably roughened by
the beginning outgrowth of numerous branches, many of which
can be identified.

ARTERIES OF HUMAN LOWER EXTREMITY 67

The course of the axial artery is somewhat less straight than
in the preceding stage. This is due chiefly to a sharp convexity
directed toward the growing extremity of the a.femoralis. The
convexity is surmounted by a short sprout indicating the point
at which the femoral is later to unite with the axial artery.
Although the axial artery is straighter in the preceding stage
than at the stage of 12 mm., an irregularity is noticeable at the
stage of 8.5 mm. in a similar situation. It seems to represent
an earlier indication of the convexity which is so pronounced at
the present stage of development.

Distal to the knee the axial artery lies in the narrow interval
between the tibia and fibula. The interosseous membrane and
individual muscles are still unrecognizable, but, although the
artery bulges toward the extensor region, its distal course
clearly indicates that it does not leave the flexor aspect of the
leg.

At what appears to be the proximal end of the very short
tibiofibular interspace a second bend occurs in the axial artery.
This bend is so pronounced as to practically amount to the bud
of the vessel which is shortly to grow into the extensor aspect of
the leg from this point.

The dorsal and plantar retia of the foot are richer than in the
preceding stage. The connection between the axial artery and
the dorsal rete is now reduced to a single vessel of large size, the
r. perforans tarsi.

The a. iliaca externa has bifurcated into the a. epigastrica
inferior and the a. femoralis.* The latter is contrasted sharply
from the a. iliac externa by the thinness of its walls, which con-
sist of endothelium only. Its structure resembles very closely
that of the femoral vein which accompanies it. The femoral
artery runs parallel with and upon the medial side of the n.
saphenus. Its extremity, now about half way along the femur,
is bifurcated into a lateral and a medial branch. The former is
short and will later join the axial artery; it may be called the

6 That the a. epigastrica inferior is an independent branch of the a. iliaca
externa, which considerably antedates it and the a. femoralis in development,
has already been pointed out by the writer (Senior, ’17).

68 H. D. SENIOR

ramus communicans superius. The latter is longer and is
recognizable as the ramus saphenus of the a. genu suprema.
The difficulty in finding developmental stages in which the a.
femoralis is present, but not yet united with the axial artery,
coupled with the thinness of the wall of the femoral artery at
this stage, indicates that the artery is one of extremely rapid
growth. In a 12-mm. embryo slightly younger than the speci-
men described (C. E. C., No. 3), reconstruction was begun under
the impression that the femoral was absent. After more thorough
study the artery was identified, bifurcated as in M. E. C., No.
H 16, but much shorter. At this stage the a. femoralis has no
branches, other than the terminal bifurcation already referred to.

d. Stage of 14 mm. C. E. C., no. 5, figs. 4 and 9 D

This stage is principally characterized by the participation of
the a. femoralis in the blood supply of the leg and by the
appearance of three branches of the axial artery.

The r. communicans superius of the a. femoralis has joined the
axial artery at the more proximal, of the two angular bends
noticed in the preceding stage. The r. saphenus remains free
and can be traced to the level of the knee joint.

At the more distal bend of the axial artery there is now a
branch, which passes to the extensor surface of the leg. This
artery passes through the proximal end of the tibiofibular inter-
space and takes a recurrent course toward the knee. The
proximal part of the artery takes part in the formation of the
adult a. tibialis anterior. It may be referred to as the ramus
perforans cruris. The recurrent vessel continuing from the r.
perforans cruris is the arteria recurrens tibialis anterior of the
adult.

The points marked upon the axial artery by means of the r.
communicans superius and by the origin of the r. perforans
cruris, respectively, may be used for the convenient subdivision
of the vessel into three parts. The part upon the proximal side
of the r. communicans superius will be referred to as the a.
ischiadica and that upon the distal side of the r. perforans cruris

rmusc.ortic a.femor _a epigast. inf
r. saphen. _Z ailiaca ext.
FarneGuin tad: 2 Ve <a
r. perf crur. ==

r perf tars l,

rete dorsal|.

rete plantar. FER
a. intero \ 3
a. popl prot’ /a peron. p. supert. £0 glutaea.
a. plant. med. (cut) Dp ~a.ischiod
a tib.p. supert
4
r. saphen. r. muse. artic.
rete plantar. a popl. prof. a. epigast. inf.
a. interos. g. iliac. ext.
a. iliac. corm.

@. peron. p. supert. a. tib.\p. superf.
a. plant. med. a. ischiad. FEP

5

10. tib.p. superf —_. saphen._ rr. musc. artic.
a, plant med. |a genu. inf med a circ. fem. lat.
@. Coronar | a. epigast. int.
N Lg iliac. ext.

aos

a. popl prof
a. peron.p. superf a. ischiad.

q glutaea. Sup.
6

Fig. 4 Reconstruction showing the arteries of the right side of the pelvis and
right thigh and leg in a human embyro, of 14 mm. (C.E.C., 5). Medial aspect.
< 20 diams.

Fig. 5 Reconstruction showing the arteries of the right side of the pelvis and
right lower extremity in a human embryo of 17.6 mm. (H.E.C., 839). Medial
aspect. > 20 diams.

Fig. 6 Reconstruction showing the arteries of the right side of the pelvis and
right lower extremity in a human embryo of 18 mm. (C.E.1.C., 409) xX 20
diams.

69

70 H. D. SENIOR

as the a. interossea. The intervening part will be called the a.
poplitea profunda.

At this stage the a. poplitea profunda has two branches which
course longitudinally through the posterior crural region. One
of these arteries resembles, in a general way, the a. tibialis
posterior of the adult, it may be called the a. tibialis posterior
superficialis. The other slightly more distal in its place of
origin is the a. peronaea posterior superficialis (of Hyrtl).7 The
relations borne by these arteries to the n. tibialis remain constant
throughout a prolonged period of development.

The a. tibialis posterior superficialis passes at first backward
upon the medial side of the n. tibialis to reach its posterior
aspect. Running upon the posterior aspect of the n. tibialis in
the leg, it passes into the sole upon the inferior aspect of the n.
plantaris medialis. The part of the artery which enters the
sole becomes the a. plantaris medialis of the adult.

The a. peronaea posterior superficialis passes from the medial
to the lateral aspect of the n. tibialis by skirting around the
proximal and lateral sides of a large muscular ramus of the
nerve.’ It follows the anterior and lateral aspect of the n.
tibialis as far as the place of origin of the nn. plantares. In this
situation the artery bifurcates into a lateral and a medial branch.
The branches are both short and end blindly.

Two branches now spring from the lateral side of the a.
femoralis. One of these is situated upon the proximal side of
the r. saphenus and represents the r. musculo-articularis of the
adult a. genu suprema. The other is the a. circumflexa femoris
lateralis of the adult a. profunda femoris.

The branches arising from the medial side of the a. femoralis
which will be discussed under stages E and F.

The a. glutaea superior is present and takes origin from the
root of the a. ischiadica.

7 According to the system of nomenclature in use at the present time, the
adjective ‘posterior’ as used by Hyrtl is redundant. In view of the essentially
fugitive nature of anatomical terminology it has seemed unnecessary to modify
Hyrtl’s original term, which adequately meets the requirements of the case.

8 This is resolved later into the rami musculares for the mm. popliteus,
tibialis posterior, flexor digitorum longus, and flexor longus hallucis.

ARTERIES OF HUMAN LOWER EXTREMITY 71

e. Stage of 17.8 mm. H. E. C., no. 839, figs. 5 and 9 E

During the preceding stages of development the arterial retia
of the foot have received their blood supply exclusively through
the a. interossea; the reta plantare directly, and the rete dorsale
by means of the r. perforans tarsi. The mesenchymal skeleton
of the foot is now definitely circumscribed, and it is plain that
the latter vessel reaches the dorsum by way of the tarsal sinus.
The arterial retia of the foot lie in close contact with the tarsus
and metatarsus and represent the following vessels of the adult
circulation—the arcus plantaris and its articular branches, the
aa. tarseae mediales and laterales, the a. arcuata, the aa. meta-
tarseae and digitales (both dorsal and plantar), and the rami
perforantes (including the ramus plantaris profundus).

The present stage is characterized by the presence of four
arteries which traverse the leg for the supply of the sole. The
a. interossea is still present, the aa. tibialis posterior superficialis
and peronaea posterior superficialis have formed their definitive
connections, and a new vessel, the a. tibialis anterior, pars
distalis, has arisen.

The a. tibialis posterior superficialis and the a. peronaea
posterior superficialis preserve in the leg the same relations to
the n. tibialis as obtained in the preceding stage. At the ankle
the medial terminal branch of the a. peronaea posterior super-
ficialis has traversed the fork formed by the diverging plantar
nerves to join the a. tibialis posterior superficialis. The lateral
terminal branch of the same artery has reached the lateral side
of the sole to join the plantar arterial rete. The connection thus
formed between the a. tibialis posterior superficialis and the
rete plantare becomes the a. plantaris lateralis of the adult foot.
At the present stage of development it receives blood from two
sources.

° The greatest total length of this embryo, measured in formalin, was 17.8 mm.
In 80 per cent alcohol it measured 13.6 mm. It has been described by Thyng as a
17.8 mm. embryo (Am. Jour. Anat., vol. 17, p. 31), and referred to by Thyng
(Am. Jour. Anat., vol. 7, p. 489) and by Thyng and Lewis (Am. Jour. Anat., vol. 7,
p. 505) as an embryo of 13.6 mm. The distribution of the arteries of the lower
extremity resembles that of other embryos of about 15 or 16 mm.

v2 H. D. SENIOR

The part of the a. tibialis posterior superficialis distal to the
newly formed origin of the a. plantaris lateralis now becomes
the a. plantaris medialis. The portion of the remainder of that
vessel which is not concerned in the formation of the a. poplitea.
becomes the a. tibialis posterior.

The branches of the medial plantar artery anastomose, upon
the lateral side of the foot, with the plantar rete, thus consti-
tuting a (transitory) superficial plantar arch.

The a. tibialis anterior, pars distalis, extends from the part of
r. perforans cruris which lies in the extensor region of the leg to.
the rete dorsale. The portion of the r. perforans which now lies
beyond the proximal end of the pars distalis corresponds to the
adult a. recurrens tibialis anterior. The portion which extends.
from the a. poplitea profunda to the pars distalis enters into.
the composition of the adult a. tibialis anterior. The a. recurrens.
tibialis anterior is connected by means of a plexiform anastomosis,
with the r. saphenus of the a. femoralis.

Upon the medial side of the a. femoralis there are at this.
stage three branches exclusive of the r. saphenus. Of these the
most distal, which has a longer individual course than the others,
takes a recurrent direction. The three branches break up to.
form an extensive plexus (not indicated in the figures) which
ramifies throughout the thigh and is particularly rich around.
the perichondrium. It is probable that this plexus is eventually
taken over by the adult a. profunda femoris and its branches.
Two branches arise from the lateral side of the a. femoralis as.
in the preceding stage.

The a. glutaea superior now takes direct origin from the a. hy-
pogastrica upon the proximal side of the origin of the a. ischiadica.

f. Stage of 18 mm. C. 1. E. C., no. 409, figs. 6 and 9 F

From the time of the junction between the r, communicans
superius and the axial artery the caliber of the a. femoralis has.
gradually exceeded that of the a. ischiadica, which has now
become exceedingly slender. After having traversed the lower
portion of the sacropudendal plexus, the a. ischiadica passes.

ARTERIES OF HUMAN LOWER EXTREMITY io

to the region of the hiatus tendineus upon the lateral side of the
n. tibialis. The n. peronaeus lies immediately upon its lateral
side.

In a general way the aa. poplitea profunda and interossea
and their branches have the same arrangement as in the pre-
ceding stage. The exact course taken by these vessels can now
be clearly recognized, since it is possible to identify the individual
muscles of the limb.

The a. poplitea profunda runs between the m. popliteus and
the tibia. At the proximal border of the muscle the artery
gives off the a. genu media and a short trunk which is being
formed by the progressive union of the proximal ends of the aa.
tibialis posterior superficialis and the a. peronaea posterior super-
ficialis. This short trunk, which lies upon the posterior surface
of the m. popliteus, may now be called the a. poplitea super-
ficialis. It forms the distal part of the a. poplitea of the adult.
A short distance below the origin of the a. poplitea superficialis
there arises the a. genu inferior medialis.

The a. interossea passes down the leg between the m. tibialis
posterior and the interosseous membrane. Just above the
medial malleolus it gives off a branch which passes around the
posterior to the medial side of the tibia, the ramus coronarius
of the medial malleolus (Hyrtl).1° At the malleolus the a.

10 In this vessel lies, according to Hyrtl, the key to the deep supramalleolar
anastomosis (i.e., the r. communicans) between the a. peronaea and the a.
tibialis posterior. The r. coronarius is regarded by Hyrtl as primarily passing
from the a. peronaea around the medial side of the tibia to unite with the a.
tibialis anterior. By means of a secondary connection occurring between the r.
coronarius and the a. tibialis posterior, the former vessel eventually becomes
converted into an anastomosing chain in which the aa. malleolaris anterior
medialis, malleolaris posterior medialis and r. communicans are the named
components. At thestage of 18 mm. of this series the r. coronarius arises from the
part of the a. interossea which persists to take part in the formation of the adult
a.peronaea. It runs, at this stage, as at the stage of 22 mm., toward the a. tibialis
anterior as described by Hyrtl, and would, no doubt, eventually join it. Since
in both these stages the a. tibialis posterior is situated upon a plane superficial
to that occupied by the r. coronarius, no junction of the vessels in question has
occurred. There can be little doubt, however, that Hyrtl has interpreted the
nature of the r. coronarius correctly. His views upon this subject, which were
derived from an extensive study of the normal and abnormal conditions found
in adult legs, have been of the greatest assistance in the interpretation of the
conditions obtaining during development.

74 H. D. SENIOR

interossea lies upon the lateral side of the tendon of the m.
tibialis posterior (i.e., between tendon and malleolus) and then
passes across the plantar end of the sinus tarsi to join the plantar
rete.

The a. peronaea posterior superficialis is placed anteriorly and
laterally to the n. tibialis and les upon the posterior surface of
the m. flexor hallucis longus. At the distal border of the muscle
the artery lies near the a. interossea and interosseous membrane.
In this situation it gives off a very large r. calcaneus lateralis
which runs upon the lateral side of the tuber calcanei. This
branch eventually forms the termination of the a. peronaea.

The a. tibialis posterior superficialis follows the posterior
surface of the n. tibialis until the latter bifurcates in the neigh-
borhood of the medial malleolus. Here the artery is placed
between the n. tibialis and the tendon of the flexor hallucis
longus. The a. tibialis anterior, pars distalis, has relations
identical with those of the adult.

The relations existing between the various structures of the
leg were ascertained by making a reconstruction which, from
the fact that it yields but little information when entirely built
up, is unsuitable for reproduction. The data gained from an
examination of its separate parts are embodied in two diagrams
(fig 7, A and B) which may serve as a rough guide to the inter-
pretation of anomalies. These diagrams show, approximately,
the course taken by the embryonic aa. poplitea profunda,
interossea, and peronaea posterior superficialis.

The order in which the various structures passing from the
leg to foot are arranged in the hollow between the medial mal-
leolus and tuber calcanei is as follows: m. tibialis anterior and
a. interosseus together; m. flexor digitorum longus; a. peronaea
posterior superficialis; n. tibialis, a. tibialis posterior super-
ficialis, and m. flexor hallucis longus.

Embryos of about this age are instructive mainly by reason
of the fortunate circumstance that the individual muscles are
differentiated before the continuity of the axial artery has
been broken. Only one important change has occurred at this
stage of development; this consists in the appearance of a com-

ARTERIES OF HUMAN LOWER EXTREMITY 15

municating branch which foreshadows the development of the
future a. peronaea. :

- The communicating branch in question leaves the a. peronaea
posterior superficialis at the proximal border of the m. flexor
hallucis longus. It passes distally between the mm. flexor
hallucis longus and tibialis posterior to join the a. interossea,

t.p., Mm. tp, m.

a |||, Fe)

5

Fig. 7 Indicates the course of three arteries of the embryonic lower extremity,
represented diagrammatically as they would appear if persisting in the adult.
The vessels which normally persist are indicated by shading. A, A. poplitea
profunda and a. interossea. B. A. peronaea posterior superficialis. d. 1. m.
flexor digitorum longus; h. l., m. flexor hallucis longus; t. p., m. tibialis posterior.

as that artery lies upon the interosseous membrane. It may be
called the r. communicans inferius.

The entire r. communicans inferius persists in the adult as
the portion of the a. peronaea which lies between the mm. flexor
hallucis longus and tibialis posterior. The portions of the aa.
peronaea posterior superficialis and interossea with which the

76 H. D. SENIOR

proximal and distal ends of the r. communicans inferior are
respectively connected also take part in the formation of the
a. peronaea. ‘

Of the branches of the a. femoralis, those arising from the
lateral side are identical with the branches of the preceding
stage. There are three branches upon the medial side. The
recurrent branch (R in figures) seems to be identical with the
similar branch of the preceding stage. It is doubtful whether
the two other medial branches correspond to those observed in
the preceding stage or not.

g. Stage of 22mm C.I.C no. 1, figs. 8 and 9 G

At this stage of development the continuity of the a. ischiadica
has been interrupted and the a. femoralis alone conveys blood
to the region beyond the knee.

The femoral artery pursues a more direct course than before
and gives origin to most of its important branches. In addition
to the a. circumflexa lateralis which appeared at the stage of
14 mm., it gives origin to the a. pudenda externa and the a.
profunda. The r. musculo-articularis has migrated to the em-
bryonic r. saphenus. The root of the latter branch, therefore,
now appears as the a. genu suprema and gives origin to the r.
saphenus and the rr. muscularis and articulares of the adult.

The a. profunda arises from the a. femoralis quite close to the
origin of the a. circumflexa lateralis. It seems probable that
the embryonic a. profunda sometimes ar:ses from the root of
the latter artery and sometimes directly from the femoral. In
the former case the adult a. circumflexa lateralis would
appear to arise from the a. profunda and in the latter from the
femoral. The great variation in the site of the origin of the
vessel is well known.

The a. profunda femoris gives origin to one perforating artery
which, since it pierces the adductores brevis and magus, repre-
sents the first or second of the adult series. The single a. per-
forans is connected, by means of an extensive plexus, with the
more distal section of the now interrupted a. ischiadica.

ARTERIES OF HUMAN LOWER EXTREMITY 77

_ The a. ischiadica passes through the sacropudendal plexus as
before and reaches the lateral side of the a. tibialis, which is now
contained in the same sheath as the n. peronaeus. After its
exit from the pelvis the artery gives origin to a few gluteral
branches and leaves the n. tibialis with the n. cutaneus femoris
posterior. In company with the latter nerve and the v. ischiadica,
it passes to the posterior surface of the thigh and ends by
dividing into several cutaneous branches. This part of the a.
ischiadica persists as the a. glutaea inferior.

The more distal portion of the a. ischiadica is traceable as a
definite path through a plexus upon the posterior surface of the
m. adductor magnus, but has lost its continuity with the a.
poplitea. The other vessel participating in the formation of
the plexus is the perforating branch of the a. profunda femoris.
This plexus evidently represents the terminal anastomosis
between the perforating arteries of the adult. It probably takes
a large share in the formation of the perforating arteries
themselves.

The distal part of the anterior tibial artery has received blood
since the time of its first appearance by way of the a. poplitea
profunda and r. perforans cruris. An alternative path has now
been provided by the appearance of a vessel which, passing
around the distal border of the m. popliteus, connects the
developing a. peronaea with a more distal part of the a. poplitea
profunda. It is the r.communicans medius which will later become
the definitive proximal end of the a. tibialis anterior. The
significance of this vessel in relation to the formation of the
adult aa. profunda and tibialis anterior will be dealt with in the
succeeding section.

The principal changes which have occurred in the embryonic
arteries of the posterior crural region since the preceding stage
of development involve the vessels which participate in the
formation of the adult a. peronaea. These changes cons st in
the disappearance of considerable parts of the aa. interossea and
peronaea posterior superficialis.

The portion of the a. interossea which formerly extended from
the r. perforans cruris.to the distal end of the r. communicans

78 H. D. SENIOR

a. iliaca. com.
a. iliaca. ext.
q. hypogast.

a. epigast. inf.
@CiiGal| 3 p,0.

a. pudend. ext. F
gd. circ. fem. lat.—_«g
a. prof. femor. \
dg. genu. suprem.
rf muse. et artic:

‘

q. perforans.
a. ischiad.

a.genu med.

a. popl. prof.

a. popl. superf.

r. com. med.

a. peron. p. superf.

lL 6. peronsea:

a. interos.

Sik ele peron. p. superf.

___ — ram. calean. lat.
_@. plant. lat.
arcus plant.

r. saph sete

a. genu. inf. med.

a. tb. p. superf.
GICOLONnGi: aaa We

r. pert. ag ae se
rete calc.

a. plant. med. (cut)

f.. profund.

FEPRiOn

Fig. 8 Reconstruction showing the arteries of the right side of the pelvis
and right lower extremity in a human embryo of 22 mm. (C.E.I.C., 1.). Medial
aspect. X 20 diams.

ARTERIES OF HUMAN LOWER EXTREMITY 79

inferius has practically disappeared. ‘The portion immediately
beyond the distal end of the latter vessel lies upon the posterior
surface of the interosseus membrane and gives origin to two
branches. The remainder of the vessel is involved in the meshes
of the rete calcaneum.

The portion of the a. peronaea posterior superficialis which
extends from the present termination of the a. poplitea super-
ficialis to the proximal end of the r. communicans inferius per-
sists to become a part of the adult a. peronaea. A large part of
the remainder of the artery has. been lost, but the distal portion
of the vessel can still be recognized as forming a projection upon
the proximal aspect of the rete caleaneum. This portion of the
artery can be traced through the meshes of the rete as far as
the a. plantaris lateralis.

The rete calcaneum is very extensive at this stage of develop-
ment and occupies the concavity upon the medial side of the
caleaneum and surrounds the deep flexor tendons in the
malleolar region of the leg.

The component parts of the a. peronaea are now recognizable.
The proximal part is derived from the a. peronaea posterior super-
ficialis. The part between the aa. flexor hallucis longus and
tibialis posterior represents the entire r. communicans inferius.
The immediately succeeding part, which rests upon the mem-
brana interossea, is derived from the a. interossea. The terminal
part of the artery consists mainly of the r. calcaneus lateralis of
the a. peronaea posterior superficialis. The connection between
the a. interossea and the part of the latter artery which gives
origin to the caleanean branch is SugerE through the agency of
the plexus caleaneum.

The part of the a. interossea which takes part in the formation
of the a. peronaea has been noted as giving rise to a branch, the
r. coronarius, at the stage of 18 mm. At the present stage
another branch arises from it, namely, the r. perforans. Both
of them become branches of the adult artery. At the stage of
22 mm. the r. coronarius can be traced around the posterior
and medial aspects of the tibia almost as far as the a. tibialis
anterior. The views expressed by Hyrtl upon the conversion

!

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 1

n. gen. fem. a.iliac.ext.

oa. mes. inf. ; g n. fem. eS

a. umbilic. g.hypog SS
{red ventr n. per. SS

radix dors. a

L3 plex.
a axis. dors
a. pudend. int i q. Iliac. com.
plex. pelv. SNa. pudenel. int.
-LS. gd. OXIS.

q. glutaea. sup.
a. glut. inf.

a. circ. il. prof.
a. epigast. inf.

a. circ. fem. lat.

a. prof. femor.
n.cut. fem. post.

a. glut. inf.

a. perforans.

gd. genu. suprem.

a. popl. superf.

a.tib.psupert .

_-@. peron. p. supert.
rete calcan.

my =-q, plant. med.

im
RET}.

MM

qd. peron. p. superf. :
a. peronaed. / a. plant.lat. + refund.
q.dorsal. ped. F EP

9

Fig. 9 Reconstructions showing the arteries of the right lower extremity in
the seven human embryos shown in the preceding figures. The cephalic (em-
bryonic pre-axial, adult medial) aspect is shown in all cases. A, X 40, the others
X 20 diams. The skeleton is‘shown, in part, in E, F and G. Parts of the tibia

80

G. 22 mm.

qa. fernor
a. epigast. inf.
r. saphen.

g.iliaoco ext.
@ epigast. inf.
oa. circ. fem. lat.

n. saphen. n. femor. ?
Ff. Com. sup. Sr
mob
a. axis Q a
> r muse.
gee , artic.
rete plantar.
r.saphen,/ £ Q
g. pudend. int. nN. peron: E ~ a. tib p. supert.

n. tibial.

a. ilioc. ext.
a. epigast int.
a circ. fem. lat. pw AE

a. genu. inf. med.
a. recur. t.@. / «=m
a. tib. ant. /
past

\e pudend. inf.

. ischiad.

. tih. p. super.

a. peron. p. supert

r. calc. lat.
r. com. int
Pak

“a. plant. lat.

a. peron.p. supert.
a. pop!. prot.
a.recur.t.a.

r pert crur.
@.interos. |

r pert. tarsi.
rete. plantar.

14 mm.

rete.

a. epigast. inf.
a. circ.fern, lq
r. musc.-artic.

. ischiad.
= \% . popl. prof.

rper B AGE —«. tib. p-superf.

. peron. p. superf

go.

Zs

. @.tib. ant . plant. med.
r pert. tarsi. :
rete plantar p. distal. . plant. lat.
. perf. tarsi.

and tarsus are omitted in E, the tibia in F and the tibia and medial femoral
condyle in G. The fibula is omitted in all. The medial plantar artery has
been almost entirely removed from E and F. In C, a. femor. should read a.
iliaca ext.

81

82 H. D. SENIOR

of the r. coronarius into the communicating branch and medial
malleolar arteries of the adult have been noted above (footnote
WO; th G3):

A small portion of the a. peronaea posterior superficialis
appears at the present stage as a branch of the a. peronaea. It
runs upon the posterior surface of the m. flexor hallucis longus.
The disintegrating a. interossea has left a similar vestige of its
proximal end upon the posterior surface of its interosseous
membrane. This vestigial branch occasionally persists as a
branch of the adult a. tibialis anterior.11. The terminal part of
the main trunk of the a. peronaea posterior superficialis is now
involved in the rete caleaneum. It usually leaves its mark in
the adult as a recurrent element among the leash of branches —
known as the r. calcaneus medialis of the a. plantaris lateralis.

The aa. iliolumbalis and sacralis lateralis now arise from the
a. glutaea superior. The root of the latter artery has conse-
quently become the posterior division of the adult a. hypo-
gastrica. All the visceral arteries of the pelvis, with the excep-
tion of the a. vesicalis superior, have assumed their adult
relations.

h. Changes occurring in the principal arieries subsequent to the
stage of 22 mm.

A. Formation of the adult a. poplitea and a. tibialis anterior.
Fig. 10, AS Co The.) 22: B,C EXCrAl Cr Cr ls tne eae

The r. communicans medius, the presence of which was noted
at the stage of 22 mm., seems to arise at about that stage of
development. In C. E. C., no 1, it is larger upon the right side
than upon the left and in C. E. C, no. 6, which also measures
22 mm., it is absent altogether.

11 Mr. Kimbrig has kindly dissected six adult legs for evidence of persistence
of this vessel. In two instances he found a branch accompanying the n. interos-
seus cruris for a short distance. In both cases the branch arose from the root of
the ramus fibularis.

It is very questionable whether the B. N. A. is correct in classifying the r.
fibularis as belonging to the a. tibialis posterior. It has usually been regarded,
in English-speaking countries at all events, as a branch of the a. tibialis anterior.

ARTERIES OF HUMAN LOWER EXTREMITY 83

At the stage of 20 mm. (fig. 10, A) blood traverses the pop-
liteal region to reach the r. perforans cruris through the original
channel, namely the a. poplitea profunda. The a. poplitea
superficialis is longer than at the stage of 18 mm., having in-
creased in length at the expense of the aa. tibialis posterior
superficialis and peronaea posterior superficialis.

At the stage of 22 mm. (fig. 10, B), a junction between the
developing a. peronaea and the distal part of the a. poplitea

Fig. 10 Three reconstructions, each showing a sagittal segment (0.25 mm.
thick) from the left leg of an embryo of the length indicated. Lateral aspect.
xX 19. A, 20 mm.; B, 22 mm.; C, 24.8mm. The tibia appears in all cases. P.
indicates the position of the m. popliteus. c.m., a. communicans media; g. 7. m.,
a. genu inferior medialis; g. m., a. genu media; 7., a. interossea; 7.m., interosseous
‘membrane; p., a. poplitea; p.c., a. perforans cruris; p. d., a. tibialis anterior, pars
distalis; p. p., a. poplitea profunda; p. s., a. poplitea superficialis; per., a.
peronaea; 7. a., a. recurrens tibialis anterior; r. p., a. recurrens tibialis posterior;
t. a., a. tibialis anterior; ¢. p. s., a. tibialis posterior superficialis.

profunda has been effected by the development of the r. com-
municans medius. Blood may now pass from the proximal part
of the a. poplitea profunda to the a. tibialis anterior, pars dis-
talis by two routes: 1) anterior to the m. popliteus, by way of
the distal part of the a. poplitea profunda and r. perforans cruris,
or 2) posterior to the m. popliteus, by way of the a. poplitea
superficialis, developing a. peronaea, r. communicans medius, a.
poplitea profunda, and r. perforans cruris.

84 H. D. SENIOR

At the stage of 24.5 mm. (fig. 10, C) the direct arterial route
upon the anterior surface of the m. popliteus has ceased to
exist. The part of the a. poplitea profunda not incorporated in
the definitive a. poplitea or tibialis anterior is represented:
1) by the root and a small branch of the a. genu inferior medialis
of the former and 2) by the a. recurrens tibialis posterior of the
latter. The main arterial channel through the popliteal region
has thus been transferred from the anterior to the posterior
surface of the popliteus muscle. Both the a. poplitea and the
a. tibialis anterior have assumed their adult arrangement.

The formation of the adult a. poplitea results from the com-
bination of two embryonic elements. The part of the vessel
which extends from the hiatus tendineus to the origin of the a.
genu inferior medialis is derived from the embryonic a. poplitea
profunda. The remainder of the artery corresponds to the
embryonic a poplitea superficialis.

The embryonic components of the a. tibialis anterior consist
of the r. communicans medius, a short section of the distal end
of the a. poplitea profunda, the proximal part of the r. perforans
cruris, and the entire tibialis anterior, pars distalis.

The a. poplitea superficialis reaches completion at the stage
of 24.5 mm. by extending as far as the r. communicans medius
which then becomes one of its terminal branches. By this
means the a. peronaea relinquishes its temporary participation
in the formation of the a. tibialis anterior.

B. The conversion of the embryonic a. tibialis posterior super-
ficialis into the adult a. tibialis posterior.

The a. tibialis posterior superficialis is originally a branch of
the a. poplitea profunda and extends into the sole. The distal
portion of the artery, marked off from the remainder of the
vessel upon the development of the a. plantaris lateralis, becomes
the a. plantaris medialis. The proximal portion of the a. tibialis
posterior superficialis has blended with the corresponding section
of the a. peronaea posterior superficialis at the stage of 24.5 mm.
as far as the termination of the popliteal artery. The further
blending between these two arteries which is to occur during
the later stages of development will cause the migration of the

ARTERIES OF HUMAN LOWER EXTREMITY '85

origin of the a. peronaea from the termination of the a. poplitea
to a point upon the a. tibialis posterior. :

After the formation of the a. tibialis posterior has been com-
pleted by the occurrence of these changes the vessel still retains
its original relation to the n. tibialis. The relations between
the a. poplitea and the n. tibialis also differ from those of the
adult.

The entire a. poplitea lies upon the medial side of the n.
tibialis. The a. tibialis posterior crosses the medial side of the
nerve to gain its posterior aspect upon which it runs until it
ends behind the medial malleolus. The relations of the aa.
poplitea and tibialis posterior to the n. tibialis have been
thoroughly studied up to the stage of 22 mm. They seem to
remain unaltered in the oldest serially cut embryo which has
been examined in this regard. It is C. I. E. C., no. 1134, which
measures 33.5 mm. .

It is not difficult to see how an agency capable of moving
the termination of the a. poplitea across the anterior aspect of
the n. tibialis from the medial to the lateral side would produce
a condition differing little from that normally encountered in
the adult. Since the knee is flexed throughout intrauterine life
it would seem that the modification of the relative positions of
the arteries and nerve might depend upon the straightening of
the n. tibialis which occurs at birth. In order to put the matter
to the test, a foetus of 7.9 cm. was dissected.!2 The rela‘ions,
however, were found to be identical with those of the adult.

Apart from the study of a few special points, this investigation
has not extended beyond the stage of 22 mm. At that period
all the vessels of the adult limb are present with the exception
of two of the three aa. perforantes, which usually arise from the
adult a. profunda femoris, the a. circumflexa femoris medialis,
and the a. obturatoria. At the stage of 22 mm. two obturator
veins are present in C. E. C., no. 1, one of these is tributary
to the v. hypogastrica, the other to the v. femoralis.

The leading features in the process of the arterial development
of the limb are indicated in figure 11.

122 For this dissection and for many others made upon various mammals, I
wish to express my indebtedness to Mr. B. Spector.

86 H. D. SENIOR

III. BRIEF DEVELOPMENTAL HISTORY OF THE INDIVIDUAL
ARTERIES

A. Axis. This artery is a branch of the dorsal root of the a.
umbilicalis, its presence has been noted by Tandler as early as
the stage of 5mm. ’03). At the stage of 6 mm. it terminates by
giving origin to the r. perforans tarsi and to the rete plantare.

The axial artery has three rami communicantes, and gives
origin to three rami perforantes as follows:

The r. communicans superius, from the a. femoralis, joins it
near the hiatus tendineus shortly before the stage of 14 mm.

The r. communicans medius, from its own branch the a.
peronaea posterior superficialis, has joined it near the distal
border of the m. popliteus at the stage of 22 mm.

The r. communicans inferius, from the a. peronaea posterior
superficialis, joins it, near the distal border of the m. tibialis
posterior, between the stages of 17 and 18 mm.

The r. perforans cruris arises between the stages of 13 and
14 mm.

The r. perforans artertae peronaea arises between the stages
of 18 and 22 mm.

The r. perforans tarsio is present at the earliest stages of de-
velopment at which the axial artery has been observed.

Two points have been marked upon the axial artery at the
stage of 14 mm. which may be used for the convenient sub-
division of the axial artery into three parts. These points
correspond to the termination of the r. communicans superius
and to the origin of the r. perforans cruris, respectively.

The respective parts of the artery are defined and named as
follows: The part upon the proximal side of the termination of
the r. communicans superius is the a. ischiadica, while that
upon the distal side, of the origin of the r. perforans cruris is
the a. interossea. That part intervening between the other two
is the a. poplitea profunda. Further information regarding the
axial artery may be found under the headings devoted to the
description of its respective parts.

A. dorsalis pedis. ‘This artery is a channel through the em-
bryonic rete dorsale. Its importance dates from the appearance

a. iliac. ext.

a. epig. inf inf" _,

‘
i
&
&

a. prof. fem.

ones

g.femor.

is saphenu

a. genu Sup.

a. popl. prof.
a. recur. tib. post

F. perf.cruris:

i}

" Ste

Bee)
SA
vs

\:

ETA

, WY
Bose

a.tib.ant.,pars dista/_—+

a. interosseq——

r. perf. PUTER 4 a a
rete plant. te

a. dors. ped. pe
arcus plant. a,

Fig. 11 Diagram to illustrate the general development of the arteries of the
human lower extremity. Adult arteries are stippled and their names underlined.
The chief embryonic channels are outlined in black. The black line is
continuous only in the case of the axial artery, otherwise it is broken.
letter P indicates the position of the m. popliteus; 7, that of the m. tibialis

posterior, and H that of the m. flexor hallucis

87

a. iliac. com. dext.

a. hypog.
a. glut. sup.

a. glut. inf.
a. pudend. int.

a. ischiadica.

r. commun. sup.

r. commun, med.

a.tib. a.tib. poster

<—a. tib. post. superf
-r. commun. int.

+#—a. peron. p. super.

@. peronaea.

a. plant. med.

longus.

ie ad

88 H. D. SENIOR

of the distal part of the a. tibialis anterior which is developed
between the stages of 15 and 16 mm.

Branches. One of the adult rr. tarseae laterales represents
the dorsal end of the r. perforans tarsi, all other branches are
derived from the rete dorsale.

A. femoralis. This artery arises from the a. iliaca externa
between the stages of 11 and 12 mm. From the time of the
first appearance of the femoral artery its growing end is bifur-
cated into the r. communicans superius and the r. saphenus.
The r. communicans superius traverses the hiatus tendineus and
joins the a. axis between the stages of 12 and 14 mm., while the
r. saphenus terminates below the region of the knee-joint.

Branches. The A. profunda femoris seems to be derived from
an arterial plexus which envelops the femur as early as the stage
of 14mm. The proximal end of the artery is distinguishable at
the stage of 22 mm. The a. circumflexa lateralis arises from the
femoral trunk shortly before the stage of 14mm. Althoughthis
artery is commonly regarded as a branch of the a. profunda, its
original connection with the a. femoralis is frequently retained
in the adult. The a. circumflexa medialis is not present at the
stage of 22 mm. One of the a. perforantes, the first or second,
is present at the stage of 22 mm., and is involved at that period
of development in an extensive arterial plexus connected with
the remainder of the distal part of the a. ischiadica. The other
perforating arteries are not present. at the stage of 22mm. The
a. pudenda externa is the only superficial inguinal branch present
at the stage of 22 mm. The a. genu suprema corresponds to
the part of the r. saphenus upon the proximal side of the origin
of the rr. musculares and articulares. These branches arise
directly from the a. femoralis, but have migrated to the em-
bryonic r. saphenus by the stage of 22mm. The distal portion
of the latter vessel corresponds to the r. saphenus of the adult.

A. hypogastrica. The main stem of this artery represents the
part of the dorsal root of the a. umbilicalis which lies beyond
the place of origin of the a. iliaca externa.

Branches. The anterior division of the a. hypogastrica is de-
rived from the proximal part of the a. axis, which becomes the

ARTERIES OF HUMAN LOWER EXTREMITY 89

a. glutaea inferior, from the root of the a. pudenda interna, and
from the part of the a. umbilicalis beyond the distal end of its
dorsal root. The portion of the original a. umbilicalis which
contributes to the formation of the a. hypogastrica is the source,
in all probability, of all the visceral branches of the artery. The
a. obtoratoria is not present at the stage of 22 mm. In conse-
quence of the transference of the origins of the aa. iliolumbalis
and sacralis lateralis to the a. glutaea superior, the proximal
part of that artery becomes the posterior division of the a.
hypogastrica.

A. iliaca communis. This artery is derived from the portion
of the dorsal root of the a. umbilicalis which lies upon. the
proximal side of the origin of the a. iliaca externa.

A. iliaca externa. This artery already springs from the con-
vexity of the a. iliaca communis at the stage of 8.5 mm. Its
termination is unbranched until a period shortly before the stage
of 12 mm., when it divides into the a. epigastrica inferior and
the a. femoralis. The a. circumflexa ileum profunda arise
from the a. iliaca externa between the stages of 18 and 22 mm.

A. interossea. The course of this part of the axial artery is
described on page 73.

The a. interossea is joined before the stage of 18 mm. by the
distal end of the r. communicans inferius. The junction occurs
at a point immediately beyond the inferior margin of the m.
tibialis posterior. This artery has given origin to the r. coro-
narius at the stage of 18 mm. and to the r. perforans and to
the branch of communication with the a. peronaea posterior
superficialis at the stage of 22 mm.

The part of the a. interossea extending from the termination
of the a. poplitea profunda to the distal end of the r. communicans
inferius disappears between the stages of 18 and 22 mm. A
vestige of the proximal end of this part of the artery sometimes
persists in the form of a small branch of the a. tibialis anterior
or of the r. fibularis.

The small portion of the a. interossea which extends from the
distal end of the r. communicans inferius to that of the membrana
interossea persists as the third part of the a. peronaea (p. 79).

90 H. D. SENIOR

The branches which arise from this portion of the a. interossea
are transferred to the a. peronaea. They are the r. coronarius
malleolaris, part of which becomes the adult r. communicans,
and the r. perforans. The remainder of the a. interossea has.
lost its identity at the stage of 22 mm. The r. perforans tarsi
disappears at about the same period.

A. ischiadica. The course taken by this part of the axial
artery is described on page 73.

The continuity of the a. ischiadica has been broken in the
region of the glutaeal fold, at the stage of 22 mm. The proximal
part of the vessel, which has now become the a. glutaea inferior,
follows the v. ischiadica and the n. cutaneus femoris to the
surface of the limb where it ends by dividing into cutaneous
branches. The more distal portion of the artery is involyed,
with the single perforating artery present at the stage of 22
mm., in the formation of an extensive plexus upon the posterior
surface of the m. adductor magnus. _

A. peronaea. This artery is clearly recognizable at the stage
of 22 mm. (p. 79). It may be divided into four parts:

The first part of the a. peronaea, which extends from the
origin of the artery to the proximal margin of the m. flexor
hallucis longus, is a persisting portion of the a. peronaea posterior
superficialis (p. 74).

The second part, which lies between the m. tibialis posterior
and the m. flexor hallucis longus, represents the entire embryonic
r. communicans inferius (p. 76).

The third part, which is short and rests upon the interosseus
membrane near the distal border of the m. tibialis posterior, is
a persisting portion of the a. interossea (p. 73).

The fourth part, which ramifies upon the fibula and caleaneum
in the neighborhood of the lateral malleolus, represents a sur-
vival of the short portion of the a. peronaea posterior super-
ficialis and of the entire r. calcaneus lateralis of that vessel
(p. 74).

Branches. The r. perforans and r. communicans, belong
orig nally to the a. interossea. The rr. calcanei laterales are.
branches of the r. calcaneus lateralis of the a. peronaea posterior
superficialis.

ARTERIES OF HUMAN LOWER EXTREMITY 91

A. peronaea posterior superficialis. The course of this em-
bryonic artery is described on pages 70 and 74. The main stem
of the vessel is present at the stage of 14 mm., its terminal
branches form their definitive connections at a stage of 15 or 16
mm.

The proximal part of the a. peronaea posterior superficialis
unites with the a. tibialis posterior superficialis to form the .a.
poplitea superficialis and the part of the a. tibialis posterior
above the root of the a. peronaea. A more distal portion persists
as the first part of the a. peronaea, and a still more distal portion
enters slightly into the formation of the fourth part of that
vessel.

The terminal branches of the a. peronaea posterior super-
ficialis form the adult a. plantaris lateralis. The r. calcaneus
medialis of the latter artery represents the end of the stem of
the parent vessel. The r. calcaneus lateralis, which arises from
the a. peronaea posterior superficialis shortly before the stage
of 18 mm., persists to form the major portion of the fourth part
of the a. peronaea (p. 79).

A. plantaris lateralis. The proximal part of this artery
represents the medial terminal branch of the a. peronaea pos-
terior superficialis (p. 71), which unites with the a. tibialis
posterior superficialis. The distal part represents the lateral
terminal part of the same vessel.

Branches. The r. caleaneus medialis represents the terminal
portion of the main trunk of the embryonic a. peronaea posterior
superficialis. The arcus plantaris and all its branches are de-
rived from the embryonic rete plantare.

A. plantaris medialis. This artery is the distal portion of the
embryonic a. tibialis posterior superficialis. It is marked off
- from the remainder of that vessel by the medial terminal branch
of the embryonic a. peronaea posterior superficialis.

A. poplitea. The part of this artery above the origin of the a.
genu inferior medialis represents a surviving portion of the
embryonic a. poplitea profunda (p. 73). The part lying upon
the posterior surface of that muscle represents the embryonic a.
poplitea superficialis. The proximal portion of the a. genu
inferior medialis is derived from the a. poplitea profunda.

92 H. D. SENIOR

Branches. The a. genu media is present at the stage of 18
mm. The other branches appear at a later stage of the
development.

A. poplitea profunda. The course of this vessel is described
on page 73. The part of the a. poplitea profunda which extends
from the hiatus tendineus to the origin of the a. genu inferior
medialis becomes the proximal part of the adult a. poplitea.
The remainder of the artery is represented in part by the root
of the a. genu inferior medialis. It is represented also by the a.
recurrens tibialis posterior and by the second part of the a.
tibialis anterior (p. 84).

A. poplitea superficialis. The formation of this artery results
from the gradual union which takes place between the proximal
part of the embryonic a. tibialis posterior superficialis and that
of the a. peronaea posterior superficialis. The vessel lies upon
the posterior surface of the m. popliteus and has nearly reached
completion at the stage of 22 mm. The a. poplitea superficialis
persists as the distal portion of the adult a.. poplitea (p. 84).

A. tibialis anterior. This vessel may be divided into four
parts which correspond to the four embryonic components of
the adult artery.

The first part extends from the origin of the vessel to the root
of the a. recurrens tibialis posterior. It corresponds to the
whole of the r. communicans medius (p. 84).

The second part occupies the immediate neighborhood of the
origin of the a. recurrens tibialis posterior. It represents, like
that artery, a persisting portion of the embryonic a. poplitea
profunda (p. 84).

The third part extends from the a. re¢currens tibialis posterior
to a point immediately beyond the root of the a. recurrens
tibialis anterior. It is derived from the proximal portion of
embryonic r. perforans cruris (p. 68).

The fourth part of the artery extends from the termination of
the third part to the proximal end of the a. dorsalis pedis. It
represents the entire a. tibialis anterior, pars distalis, of the
embryo (p. 72).

ARTERIES OF HUMAN LOWER EXTREMITY 93

Branches. The a. recurrens tibialis anterior represents NG
terminal portion of the a. perforans cruris (p. 68).

The a. malleolaris anterior medialis is derived from the r.
coronarius of the a. interossea which arises shortly before the
stage of 18 mm.

A. tibialis aniertor, pars distalis. This embryonic artery
arises at about the stage of 15 or 16 mm. Its course is identical
with that of the fourth part of the a. tibialis anterior of the
adult.

A. tibialis posterior. The part of this artery proximal to the
origin of the a. peronaea is the product of the union between the
a. tibialis posterior superficialis and the a. peronaea posterior
superficialis. The remainder of the vessel is a survival of the
part of the stem of the a. tibialis posterior superficialis proximal
to the point at which it is joined by the medial terminal branch
of the a. peronaea posterior superficialis.

Branches. The a. peronaea is described on page 90. The
a. malleolaris posterior medialis and the r. communicans are
derived from the embryonic r. coronarius malleolaris medialis of
the a. interossea (p. 73).

A. tibialis posterior superficialis. The course of this artery,
which arises shortly before the stage of 14 mm., is described on
page 70.

The whole of the a. tibialis posterior superficialis persists in
the adult life. In combination with the a. peronaea posterior
superficialis, the more proximal part of the artery forms the
part of the a. poplitea in contact with the m. popliteus and the
part of the a. tibialis posterior proximal to the origin of the a.
peronaea.- The more’ distal part of the artery persists as the
portion of the a. tibialis posterior beyond the origin of the a.
peronaea and as the a. plantaria medialis.

The relation of the part of the a. tibialis posterior super-
ficialis which forms the a. tibialis posterior of the adult to the
n. tibialis is discussed on page 84.

Rete dorsale. The dorsal rete of the embryonic arises from
the r. perforans tarsi. It forms all the arteries which are
distributed upon the dorsum of the adult foot.

94 H. D. SENIOR

Rete plantare. This embryonic plexus arises from the terminal
branches of the a. interossea. It furnishes the arcus plantaris
and all the other arteries of the adult sole excepting the aa.
plantares.

BIBLIOGRAPHY

CaILLARD 1832 Proportions de Méd. et dela Chir. Thése inaug., Paris.

CRUVEILHIER, J. 1843 Traité d’Anotomie descriptive. Paris.

DeVrigEsE, BertHa 1902 Récherches sur I’évolution des vaisseux sanguins
chez ’homme. Archiv. de Biologie, T. 18, p. 665.

DusrveEIL, J. M. 1847 Des anomalies artérielles, Paris.

Exits, G. V. 1853 An account of an instance of remarkable deformity of the
lower limbs. Medico-Chir. Trans., vol. 36, p. 489.

Faace, C. H. 1864 Case of aneurism nenied on an abnormal main artery ie
the lower limb. Guy’s Hosp. Rep., vol. 10, p. 15.

Froriep, L. Fr. V. 1832. Notizen aus dem Gebiete der Natur u. Heilkunde,
B. 34, S. 45.

GREEN, P. H. 1832 Ona new variety of the femoral artery, with observations.
Lancet for 1831, vol. 1, p. 731.

Grosser, O. 1901 Zur Anat. u. Entwichlungsges. des Gefiss-systems der
Chiropteren. Anat. Hefte, B. 17, H. 2, S. 203.

Hocustetrrer, F. 1890 Uber die urspriingliche hauptschlagader der hinterén
Gliedmasse des Menschen, u.s.w. Morph. Jahrb., B. 16, 8. 300.

Hyrrt, J. 1864 Uber normal u. abnorme Verhiltnisse der Schlagadern des
Unterschenkels. Wien.

Lespoucg 1893 Verhandl. d. Anatom. Gesellsch., auf. d. siebenten Versamml.
Anat. Anz., Erginzungsh. zum. B. 8.

Levy, G. 1902 Morphologia delle arteriae iliache, Parte 2. Archiv. Ital. di
Anat. e di Embriol., V. 295.

McMoraricn, J. P. 1904 The development of the human body, 2nd ed., Phila-
delphia.

PorowskI, J. 1893 Uberbleibsel der Arteria saphena beim Menschen. Anat.
Anz., B. 8, S. 580.
1894 Das Arteriensystem der unteren Extremititen bei den Primaten.
Anat. Anz., B. 10, 8S. 55 u. 99.

QuaINn, R. 1844 The anatomy of the arteries of the human body. Lenore

Rue, C. 1863 Anomalie der Arteria cruralis. Wiirzb. mediz. Zeits., B. 4,
S. 344.

Ruce, G. 1894 Varietiiten im Gebiete der Arteria femoralis des Menschen.
Morph. Jahrb., B. 22, 8S. 161.

Satyr 1898 Arteria dorsalis pedis. Pisa.

Senior, H. D. 1917 The development of the external iliac artery in man.
Ann. N. Y. Acad. Sci., vol. 27.

Srrepa, H. 1893 Verhandl d. Anatom. Gesellsch., auf d. siebenten Versamml.
Anat., Anz., Erginzungsh. zum B. 8.

ARTERIES OF HUMAN LOWER EXTREMITY 95

Suressporr, M. 1895 Lehrbuch der vergleichenden Anatomie der Haustiere.
Stiittgart. i

TANDLER, J. 1903 Zur Entwickelungsges. der menschlichen Darmarterien.
Anat. Hefte, B: 23, Hi, S: 187.

VELPEAU 1839 Eléments de Médicine Opératoire.

ZAGORSKI 1809 Mém. de l’Académ. des Sci. de Petersb.

ZUCKERKANDL 1894 Zur Anat. u. Entwickelungsges. der Arterien des Vor-
derarms (1. theil). Anat. Hefte, B. 4, H. 1, 8S. 1.
1895 Zur Anat. u. Entwickelungsges. der Arterien des Unterschenkels
u. des Fusses. Anat. Hefte, B. 5, H. 2, S. 207.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 1

¥ cae ;
mie ne ng fata

aes. Pal
ia Si 4 nga Piginh,
it ayn in) icevcatlt

a i : F i vt iat at Mal ks oe aa
Oe CARNE er. en fi nile: AaB Ae

Py? . My ‘
ae Man aan Prigee uds, Ky . bP Gon SAA
Shing 7 . * Wie, :
i : 3 ; aa
Pas My ’ ‘ ‘ A Sn, yy Le Re ' i“
. AY) trp i 4 ¢
he ; : ; , ay
Ey F ‘ a t 54 eek apa. ; f
OL : , ‘ b "4
et, iy : Ae RCA AA ie GMA. 8: > aah
A) ? ‘ *% . ¥ ~
7 f > Live prsay ‘ abel . e 4 eo.
4i4 Aiton ¥ a ‘
; 4 Wi! ¢ : et)
Ale i;
’ Wi E
. va . it Pal | #) a , ' * a
; +
ye j
o~ a A .
4 =
: kage - ; %
<
*
hy a4 , % . :
. -
Tie
“A *
‘ . _-
i 7 sie ‘
hy '
, aly ' 4 , i
‘
a /
\ - ‘
4 1
\ bs = .
.
‘J :
‘ oy
. } ;
+ bd =
rn ee
, 4 cer
' AW
iy
; e. > f : gt

3 c pant 7) iyhe' i : abe ae Padi a)
Ny sls oor Mee WU “pela pO E a 8

| ' ; “ ie ‘ Shes j

ie

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 24

THE DEVELOPMENTAL RELATIONS OF BRACHY-
DACTYLY IN THE DOMESTIC FOWL

C. H. DANFORTH

Department of Anatomy, Washington University School of Medicine

FIVE [FIGURES

The factors which exercise determining influences in the
ontogeny of animals have proved difficult of recognition and
evaluation. To what extent any given structure is the direct
expression of some more or less specific ‘determiner’ in the
germ cell and to what extent it represents the product. of
reactions to the influence of other parts of the developing
organism cannot often be estimated in any reliable manner.
Nevertheless, some knowledge of the relative weight of the
two kinds of factors involved is essential to a satisfactory con-
ception of the processes .of embryology and morphology. The
observations recorded in the present paper are reported because
they seem to throw some light on this problem with reference
to a rather special case; namely, the correlation between skeletal
peculiarities of the fourth toe and the presence of feathers on
the tarsi in the common fowl.

The data to be presented were collected as the result of the
following observations: In July, 1915, an egg incubated in the
laboratory hatched out a chick (no. 12) which showed an almost
complete absence of the fourth toe on each foot. This particular
chick was helped from the shell and the condition of its toes
noted at that time, so there is no possibility of the malformation
having been due to injury received after hatching. Subsequent
dissection showed that the three terminal phalanges and the nail
were lacking on both sides (fig. 1, B). During growth the
lateral rudimentary toes turned under the balls of the feet
causing the bird some inconvenience, especially in perching.

97

98 Cc. H. DANFORTH

The specimen, a male, had moderately booted tarsi and was
also strongly polydactyl. It met an accidental death at about
nine months of age before any offspring had been obtained.

The parents of this chick were full brother and sister (no. 8,
#, and no. 7, 2). Both were polydactyl and both had booted
tarsi. On closer inspection it was found that their fourth toes
were somewhat shorter than normal. It was at first thought
possible that the brachydactyly and polydactyly might be cor-

rq

A B

Fig. 1. Bones of the left foot from (A) normal and (B) brachydactyl cockerel.

related, that the excessive toe development on the medial side
of the foot might be more or less at the expense of the tissues
on the lateral side. The following data show that such was not
the case.

From the mating of nos. 7 and 8 only a few chicks were ob-
tained. These included specimens with four and five digits,
bare and feathered tarsi, brachydactyl and normal fourth toes.
Some additional specimens secured by crossing no. 8 with a
barred Plymouth Rock hen gave essentially the same distri-

BRACHYDACTYLY IN THE FOWL 99

bution of characteristics. The association of characters in these
chicks was such as to suggest that the brachydactyly might be
correlated with booting but the numbers were not adequate to
warrant definite conclusions.

Through the courtesy of the officials in charge, it was possible
to make a superficial examination of the feet in a considerable
series of booted fowls exhibited at poultry shows in St. Louis.
Such examinations sufficed to show that brachydactyly occurs in
several races with booted tarsi. One highly valued buff Cochin
fowl was found to present the characteristic in as pronounced a
form as did the specimen which first attracted our attention to
the peculiarity. Owners and judges are inclined to attribute
such cases to accident, but the fact that the standard of require-
ments for some of these breeds specifically calls for well formed
toes suggests that the tendency to brachydactyly has been
vaguely recognized by breeders. No instance was found in
which a short lateral toe occurred on a smooth-shanked bird.

These observations .were sufficient to indicate that brachy-
dactyly is of rather common occurrence in the domestic fowl
and seemed to justify a more careful investigation of the
peculiarity. Consequently a further breeding experiment was
undertaken in order to furnish material for study of the heredity
and embryology of the condition. -

MATERIAL AND METHODS

To obtain this material male no. 8 was mated to a flock of 13
white Leghorn hens. All eggs laid between February 15 and
April 18, 1917, were incubated, except when, as occasionally
happened, the number of eggs remained for several days in
excess of the capacity of the incubator. At such times the
oldest eggs were discarded. It is not probable that this method
could have resulted in any differential selection. A total of 300
eggs were used. Of these, 108 yielded living embryos and 86
were allowed to hatch. Two of the remaining eggs were infected
with bacteria, 34 were non-fertile or developed only slightly,
and 70 contained embryos that died before hatching.

100 Cc. H. DANFORTH

The living embryos were fixed in Zenker’s fluid or in 10 per
cent formalin. They were subsequently examined under a
binocular microscope and parts of many of them were either
sectioned or stained by the van Wijhe method and studied in toto.

The contents of eggs found to contain dead chicks were put
in formalin and after a short period of hardening the embryos
were examined, their age estimated and characteristics recorded.

All the toes on both feet of each of the 86 living chicks were
measured, the grade of polydactyly, the grade of booting and
also the form of the comb were noted. These data likewise
were recorded so far as possible for the older embryos. For
each of the living chicks an index of brachydactyly was obtained
by dividing the sum of the lengths of the two lateral toes (digits
IV) by the sum of the lengths of the two medial toes (digits II).
This gives a value varying above and below 1. The value
obtained is multiplied by 100 in each case to eliminate fractions.
The index is thought to be fairly reliable, since careful measure-
ments have shown that the fourth toe is the one chiefly affected
while the third is only slightly so and the second possibly not at
all. The toe measurements represent the distance, obtained by
the use of dividers, from the base of the nail to the metatarso-
phalangeal articulation.

DESCRIPTION OF STOCK USED FOR BREEDING

The females employed in the breeding experiment were all pure
bred white Leghorns of a well established strain. Their shanks
were free from feathers and there was no indication of poly-
dactyly or brachydactyly. The fourth toe index ranged from
106 to 121 with an average of 112. All were pure white with
large single combs. Eight of these and a male were obtained
in 1916 from a poultryman who devotes himself exclusively to
this breed. The other five were raised at the laboratory from
this original flock. There can be little doubt as to the purity
of the stock especially in regard to the characteristics under
investigation. At the time of the experiment five of the hens
were less than ten months old, the rest were two years or more.

BRACHYDACTYLY IN THE FOWL . 101

The male (no. 8) differed from these hens in a considerable
number of characteristics, for most of which he seems to have
been heterozygous. His color was partly that of a barred
Plymouth Rock, although hackle, saddle feathers, and_ tail
contained much white and the breast was somewhat spotted.
The comb was broad and long with irregular elevations and a
flattened, truncated and slightly trifid posterior prolongation.
It did not seem to be either typically ‘rose’ or ‘walnut’ although
perhaps more closely approaching the latter. The tarsi were
moderately booted (grade 2 in an arbitrary scheme adopted for
purposes of this paper). Digit I on the left foot was partially
doubled (grade 2) while on the right foot it was normal. Digits
IV were short, giving an index of 89. At the time of the
experiment this bird was in the last half of his third year of life-

HEREDITY OF CHARACTERISTICS

The heredity of the five outstanding characteristics, color.
comb form, polydactyly, booting and brachydactyly may be
briefly summarized. The first four of these have been subjected
to careful study by Davenport (’06, ’09) and others.

A. Color. All. chicks that hatched and all embryos that
reached the stage for developing down were white, or white with
occasional small dark spots. White was to have been expected
in the offspring since the white of the Leghorn is a well known
dominant character.

B. Comb. According to the prevailing view as to the unit
factors involved (Bateson, ’09) four types of comb should have
occurred in equal numbers: walnut, rose, pea and single. These
four types did occur, but since in the embryo and chick it is
often difficult to differentiate with certainty between rose and
walnut (of the type met in this experiment), pea and single,
the four categories were grouped into two; broad combs and
narrow combs. Of 172 chicks and late embryos, 88 had broad
combs and 84 narrow, the expectation being 86 of each. This
is very close to the Mendelian ratio.

2 -. Cc. H. DANFORTH

C. Polydactyly. For the sake of convenience three grades of
polydactyly, corresponding essentially to those designated by
Anthony (’99), were employed. Grade 1 includes cases in which
only the distal phalanx is involved; grade 2 includes cases
where two phalanges are involved but with the toes usually
more or less united and equal; grade 3 includes all cases of a
more pronounced form. In grade 3 the metatarsal is also
affected and the accessory toe generally has three phalanges.
Only one instance of six toes occurred. Occasional specimens in
which there were really only four toes but in which the first
digit showed three phalanges instead of two were classed as
polydactyl, grade 1. Such individuals were included because
the evidence seems to the writer to indicate that polydactyly
in the fowl is not so much a matter of added elements as of
readjustments following early hypertrophy of a single anlage.
In possibly a third of the cases the two sides were not strictly
symmetrical. Of the 28 polydactyly chicks that hatched, 8
were of grade 1 and 10 each of grades 2 and 3. The father
‘was of grade 2 (on the left side only).

As shown by Kufmann-W olf (’08), some cases of polyadactyly
can be recognized as early as the fifth day of incubation. Prob-
ably all individuals which would develop info grade-3 speci- .
mens can be recognized as polydactyl at this age. But for
purposes of the present record no individual under seven
days is included. Of 220 chicks and embryos of 7 days and
over, 147 were normal, 73, or 33 per cent, were polydactyl.
If polydactyly behaved as a strictly dominant character the
expectation in this case would be 50 per cent. The failure of
polydactyly in poultry to give typical Mendelian ratios has
been discussed by several authors, especially Davenport (’09).

D. Booting. In chicks that are to be booted, feather (down)
rudiments generally become apparent on the tarsi during the
eleventh day of incubation (fig. 2) or considerably later than on
other parts of the body. After the twelfth day, booting is
easily recognized. Thirty-six living chicks which were booted
were classified into three grades, based upon the amount of down
on the tarsi. Grade 1 represents individuals with from one

BRACHYDACTYLY IN. THE FOWL 103

to several feathers on the shank. Grade 3 stands for speci-
mens with about twelve feathers on the shank, three or four
on the lateral toe, and occasionally one or two on the middle
toe. Grade 2 includes specimens intermediate between 1 and
3. The distribution among the three grades was as follows:
grade 1, 14; grade 2, 10; grade 3, 12. The father was grade 2.

Out of a total of 129 chicks and embryos of thirteen days and
over, 80 were normal and 49 booted,—38 per cent. Booting,

EXIT.

FLEX.

Fig. 2 Sketches of cross-sections of comparable regions of the right shank
from (A) nine-day and (B) eleven-day embryos, both of which were strongly.
brachydactyl. Dr., down rudiment; Fzt., extensor group of tendons; Flez.,
flexor group of tendons; Mz., metatarsal cartilages.

like polydactyly, does not give strictly Mendelian proportions
(Hurst, 705, Davenport, ’09).

E. Brachydactyly. In chicks with normal feet the lateral toe
is about 10 per cent longer than the medial; in brachydactyly
chicks it is of equal length or shorter. While the differences are
only slight and in embryos difficult to measure a certain number
of obvious cases can be recognized in nine-day specimens. From
the tenth day on they are clearly distinguishable, probably in
nearly all cases. It is possible, however, that a small number

104 Cc. H. DANFORTH

of brachydactyl specimens, having a relatively high index, may
be overlooked even among late embryos. Among 90 embryos
of ten days and over, 57 had long fourth toes; 33, short. Of 86
young chicks, 55 had an index of more than 100 while 31 showed
an index of 100 or less (brachydactyl). The totals for embryos
and chicks are: out of 186 individuals, 112 normal, 64 brachy-
dactyl. This is 36.4 per cent, which is intermediate between
the percentages for booting and polydactyly. These character-
istics behave in heredity quite differently from color and comb
form which give results corresponding closely to the Mendelian
expectation.

EMBRYOLOGY OF BRACHYDACTYLY

Since brachydactyly in poultry has not seemed to attract
much attention (Anthony, ’99, states that he has seen specimens
which showed four and six phalanges in their fourth toes), a
fuller account of the condition may be entered upon at this
point. Inasmuch as the shortening of the other toes is relatively
slight and difficult to measure the fourth only will be considered.
The skeleton of a normal foot is shown in figure 1, B. Digit IV
has five phalanges, the terminal one bearing a nail. Among
brachydactyl specimens, the fourth toe varies from a condition
in which nail and all five phalanges are present with a total
length equalling that of the second toe (but not exceeding it as
in normal feet) to a condition showing complete absence of the
nail bearing phalanx and only two remaining phalanges. Inter-
mediate conditions are represented by toes with three or four
well formed phalanges and by toes with the full number of
joints but of considerably reduced length. The brachydactyl
toes also varied from the normal in other characteristics such
for example as the number of rows of scales. The two sides of
the same individual were found to be very closely correlated.
Among the 31 brachydactyl chicks that hatched, the index varies
from 70 to 100, with an average of 89, which is identical with the
index of the father. All of the specimens had the nail reduced
and seven of them had only four bones in the toe as shown
either by dissection or surface configuration (fig. 3).

BRACHYDACTYLY IN THE FOWL 105

Tracing the condition back in the development of the chick,
we find, as already stated, that brachydactyl can be recognized,
at least in many cases, as early as. the ninth day, when the
fourth toe in certain individuals can be seen to be conspicuously
shortened. It has not been possible to determine whether or
not the shortening is actually present at an earlier date. The
cartilage of the third phalanx of the fourth toe is laid down
during the 8th day. The fourth and sometimes the fifth car-
tilage are formed on the 9th day. By the end of the 10th day
all of the phalanges to develop are present in cartilages. Definite
reduction in the number of cartilages cannot be recognized with
certainty earlier than the tenth day.

e @

Fig. 3 Foot-prints of (A) brachydactyl and (B) normal left feet. In A the
absence of one phalanx in the fourth toe is revealed.

The fact that brachydactyly is apparent at the time when
cartilages are formed, if not actually before then, would seem to
indicate that the reduction of the skeleton takes place in re-
sponse to conditions already manifest in the toe, and not as a
result of any factor acting specifically on cartilage or bone forma-
tion. In other words the cartilage anlagen form while subjected to
a kind of compression resulting from the shortening of the toe.
The consequent reduction in the cartilages, and the correlated
changes in ligaments, tendons, etc., would therefore seem to be

106 Cc. H. DANFORTH

due not to any inherent peculiarity of the structures themselves,
but rather to the conditions under which they develop. The
shortening of the soft parts, seemingly secondary is more
probably the primary factor in causing brachydactyly. In the
adult foot, however, the bony framework provides an accurate
index of toe length.

As already indicated, the reduction in the skeletal elements
is both qualitative and quantitative. It may be that this
reduction does not take place in the same way in all short-toed
specimens, but a series of grades may be described which
seem to represent successive steps in the process. 7

The first degree of shortening represented by the least pro-
nounced type is characteristic of the majority of brachydactyl
specimens. Here all of the phalanges are present, but the
fourth, or sometimes the third, is more or less shortened. The
amount of shortening in these phalanges varies greatly and can
often be estimated in the living chick. It has not been possible
to determine accurately whether or not the other phalanges are
also shortened, although it is probable that they are, especially —
- the fifth since the nail is almost always abnormally small. The
fact that the fourth, or occasionally the third, phalanx is the
one first affected suggest that the reducing factor is most effective
at the time these phalanges are being laid down, viz: on the
ninth, or occasionally the eighth day.

The second degree, found in slightly over a fifth of the cases
studied, shows the first numerical reduction in the phalanges.
Here the third and fourth segments are replaced by a single
element. The resulting cartilage is commonly a slender rod
slightly longer than a normal fourth which it otherwise re-
sembles. This cartilage also varies in length so that there is
some fluctuation within this grade. It has not been possible to’
determine with certainty whether the coalescence of the two
elements is due to one cartilage forming from the substance
that normally gives rise to two, or whether both elements form
separately and then fuse. The specimen shown in figure 4 seems to
favor the latter alternative, but where the adult organ presents
an unlimited series of gradations it is impossible to say in the

BRACHYDACTYLY IN THE FOWL 107

case of any individual embryo whether the observed condition
is transient or definitive.

The third degree is represented by one twenty day embryo
and possibly one or two younger embryos. In this case there
seems to be a general reduction in which the terminal phalanx
suffers most. The third and fourth phalanges are fused. The
claw is represented only by a small flat scale and the fifth bone

Fig. 4 Feet of a brachydactyl embryo on the tenth day of incubation. In
the fourth toe of the right foot, at B, is shown the completely fused third and
fourth phalanges. In the left foot, at A, these two phalanges are seen to be
only partially fused.

by a little nodule into which the terminal slip of the flexor
profundus tendon is inserted. Since the flexor profundus is
also inserted into the plantar surface of the normal fourth
phalanx or the fused third and fourth phalanges, a stage showing
a complete absence of the fifth segment would differ very slightly
from the case here described. It is interesting to note in this
connection, however, that the basal and terminal phalanges are

108 Cc. H. DANFORTH

the most stable, the intermediate elements being the ones first
affected.

The fourth degree was found in no. 12 (half brother of these
chicks), referred to at the beginning of the paper, and in several
chicks obtained later. Here there are only two bones present
in the fourth toe (fig. 1, B). There is no nail. The tendon of
the flexor profundus is inserted on the plantar surface of the
terminal bone and the superficial flexér into the lateral aspects
of both segments. In view of this arrangement, the conforma-
tion of the two bones, and the condition observed in other
specimens it would seem probable that the proximal element
represent the normal basal phalanx while the distal one repre-
sents a fusion (in effect, if not in reality) of phalanges II, III
and IV; phalanx V being entirely absent.

Owing to the brief period between the first indication of
clearly demonstrable brachydactyly and the appearance of
phalangeal cartilages, the possibility that the precartilage cells
are the ones first affected by the shortening factor can not be
entirely excluded. If such be the case, it is the skeletal elements
that from the earliest period determine the form and proportions
of the toe; but as already indicated, the evidence seems to point
quite as strongly to the alternative possibility, namely, that the
form of the skeletal elements as first laid down is the result of a
more or less passive response of cartilage forming cells to the
influence of surrounding conditions. The relation of brachy-
dactyly to booting (to be discussed in a later paragraph) adds
considerable support to the latter interpretation. If this inter-
pretation is correct, it follows that the early shortening of the
toe as a whole results in a reduction of the skeletal elements,
the degree of reduction being directly correlated with the degree
of shortening, and not improbably passing successively through
the above described stage till the definitive condition is reached
on the tenth day or shortly thereafter.

INTERRELATION OF CHARACTERISTICS

The relations of brachydactyly and the other hereditary
characteristics observed in this experiment are indicated in the

BRACHYDACTYLY IN THE FOWL 109

TABLE 1

A summary of the data arranged to show the relation between brachydactyly
polydactyly, booting and comb form in the eighty-six
living chicks

A. BRACHYDACTYLY AND D. BRACHYDACTYLY AND
COMB FORM POLYDACTYLY
Broad comb Narrow comb Five toes Four toes

Obs. Exp. Obs. Exp. Obs. Exp. Obs. Exp.

Shontmounth toe... 2-54. 15 15e5 16 15e5 10 10 21 21
Monestourth tev... +... .4.4)) 28 DAL ® Dal 1 ies) 18 17.9 37 347/01
B. POLYDACTYLY AND E. POLYDACTYLY AND
COMB FORM BOOTING

Broad comb Narrow comb | Booted shank | Smooth shank

Obs. | Exp. | Obs. | Exp. | Obs. | Exp. | Obs. | Exp.

TRINVEATE Ss Oo ete See eee 16 14 12 14 13 pes 15 | 16.2
HOMERGOCS Hearne ao. teas Pal 29 31 29 23 24.2 35 | 33.8
Cc. BOOTING AND COMB F. BOOTING AND BRACHY-
FORM DACTYLISM

Broad comb Narrow comb |Short fourth toe|Long fourth toe

Obs. | Exp. | Obs. | Exp. | Obs. | Exp. | Obs. Exp.

DOoLed Shakes a .2c)225...| AT 18 19 18 dl 12.9 iy | aye
Smooth shank.............| 26 25 24 25 Osieis 50 | 32

accompanying table (table 1, A—F), which is based on the 86
chicks hatched. Since the male used was, in Mendelian terms,
heterozygous for all four of the peculiarities studied, the de-
terminers for each (with the partial exception of comb form)
having been supplied by his paternal gamete alone, and since
the females were homozygous for the absence of all these
characters, an especially favorable opportunity was afforded for
testing linkage relations. If each character behaved in a
strictly Mendelian fashion and segregated independently, the
expected value for each combination would be 21.5, and the
sum of each pair added either vertically or horizontally would
be 43. While, as already indicated comb form agrees exactly
with the Mendelian expectation (43: 43 for those that hatched)

110 Cc. H. DANFORTH

the other three characters do not, so that linking can only be
tested by comparing the proportional distribution of one pair of
contrasted characters with reference to other pairs. The table
shows the observed values for all possible combinations and the
expected values for each of these based on the behavior of the
several characters considered separately.

It will at once be apparent that the observed and expected
values agree very closely in the first five sets of combinations.
The slight apparent deviations in favor of a correlation between
polydactyly and broad comb and between polydactyly and
booting are too small to be of signficance. So far as this part
of the data goes it is in full accord with Davenport’s (06) con-
clusion that correlation of characteristics in poultry is very rare.

But, on the other hand, between brachydactyly and booting
there is found to be a close correlation. Here the values are
clearly significant. Of thirty-one brachydactyl chicks, all had
booted tarsi; and of thirty-six chicks with booted tarsi, thirty-
one were brachydactyl. The distribution is shown graphically in
figure 5 where it also appears that the length of the fourth toe is
rather more variable in specimens with booted shanks than in
those with smooth shanks.

There is, then, a distinct relation, which becomes apparent in
embryonic stages, between brachydactyly and booting, but
there is no evidence of a significant connection between either
of these characters and comb form, color or polydactyly.

SIGNIFICANCE OF THE CORRELATION BETWEEN BRACHY-
DACTYLY AND BOOTING

The explanation for the relation between short toes and
feathered tarsi is not obvious. Several possibilities suggest
themselves.

In the first place, it might be that we have here a case of true
linkage, such, for example, as Morgan (715) and others have
successfully demonstrated in fruit flies. Morgan believes that
in such cases the determiners for both of the linked characters
are located in the same chromosome. If this were the true
explanation, ‘crossing over’ might possibly be expected to occur

BRACHYDACTYLY IN THE FOWL 112

in a certain percentage of cases, the frequency depending on the
distances of the two determiners from each other. There “is
some evidence that could be regarded as favoring this inter-
pretation. Five chicks with more or less booted tarsi (grades 1,
1, 1, 1, 2) were not brachydactyl. If these represented true
cases of ‘crossing over’ there should have been an equal number
of brachydactyl chicks. which were not booted. One specimen
was obtained. To fully substantiate this hypothesis it would

Fig. 5 Curve showing the relation between brachydactyly and booting in
86 living chicks. Each square represents an individual, the shaded ones being
booted. The indices of brachydactyly are indicated below.

be necessary by further breeding experiments to derive from the
same hybrid stock one strain with feathered feet and long toes
and another strain with unfeathered feet and short toes.

In this connection it was thought desirable to investigate
the relation of these characters in other species. The writer is
greatly indebted to Mr. Louis Agassiz Fuertes for carefully
examining for him a large number of booted grouse, ptarmigans,

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 2

ae Cc. H. DANFORTH

and owls. Mr. Fuertes was unable to find any indication
whatever of brachydactyly in any of these birds.

In the case of the domestic pigeons on the other hand, the
writer has found somewhat the same condition that occurs in
hens. ‘Swallow’ and ‘pouters’ both of which are booted, often
show short lateral toes. Among ‘muffed tumblers’ also heavily
booted, several specimens were found to have the third and
fourth toes fully webbed, and I have been informed that the
same condition sometimes obtains in ‘pouters.’ Davenport (’09,
pp. 29 to 36) describes a similar condition in a strain of fowl.
It is not clear whether all of his syndactylous specimens were
also booted but the fact that the first cases arose in breeds that
are normally booted, is probably significant. In another flock
of fowl, related to those here described and having so far as
known no syndactyly in the ancestry, a pair of brachydactyl
birds produced two syndactylous chicks of the type described by
Davenport. These cases suggest that syndactyly like brachy-
dactyly may represent an arrest of development, occurring in
the former case a little earlier than in the latter, and that the
two conditions may be dependent upon a single factor.

While the suggestion of linked determiners cannot be entirely
ignored as a possible explanation of the results obtained,
there is still further evidence against such an interpretation.
This rests chiefly on the fact that there is a correlation between
the degree of booting and the index of brachydactyly. Despite
wide fluctuations, this fact is brought out by averaging the
indices for the different grades. After excluding the two ex-
treme cases with indices of 70 and 120 the average index for
each group is as follows, grade 1, 93; grade 2, 91; grade 3, 89.
If the two extremes are included, they tend to raise the value
of grade 1 and lower the value of grade 3, thus accentuating
the differences. Such a result is not easily explained on the
assumption of two independent determiners but, on the con-
trary, strongly suggest that the association is due to some-
thing more than a chance proximity of the determiners in a
chromosome.

BRACHYDACTYLY IN THE FOWL fale

If this is the case the question presents itself as to whether the
relation is causal, one condition producing the other or whether
the two conditions are both dependent on some one determining
factor. It was at first thought probable that the abnormal
development of feathers or down on the feet overtaxes the
nutritional supply of the limb with the result that the fourth
toe becomes somewhat stunted in its development. The em-
bryology of the condition, however, does not support this view.
On the contrary, in light of the sequence in which the character-
istics make their appearance in ontogeny, brachydactyly would
seem to be the cause of feathered tarsi, an interpretation which,
a priori, does not seem plausible.

Since, on the whole, the evidence does not favor the idea
that either condition is really the cuase of the other the only
alternative that would seem to be left is that they are both
caused by the same factor. The close association of the two
characteristics and the fact that they tend to fluctuate together,
points strongly toward this conclusion. What the factor may
be or precisely how it acts cannot at present be answered. Nor
can the number of characteristics which it affects be stated
since it is quite possible that some of them have been overlooked
in the present study.

If the connection between brachydactyly and booting had
been a little less obvious the former might have been described
as an independent unit character and a separate determiner in
the germ plasm postulated for it, as has already been done for
booting by the several authors who have studied that character.
In general the relations here described would seem to be such
as tend to support the contention of Morgan (loc. cit.) that the
determiners in the germ plasm are not strictly specific, but tend
to influence the character of the organism as a whole.

In the particular case under consideration it is quite possible
that the determining factor may actually be effective for only
a short time, possibly a few days. After two cartilages have
once fused it is doubtful if the removal of the exciting cause
would in itself induce them to separate. It is also possible that
a feather germ once formed would need no further stimulus than

114 Cc. H. DANFORTH

proper nutritional conditions to continue its development to the
end. If these suppositions are correct it is quite conceivable
that the whole complex is due to a slight irregularity (delay or
acceleration) in the beginning functioning of some one or other
of the endocrine glands. The character of the gland would of
course be determined by other factors which would have an
ultimate basis in the nature of the germ plasm itself. An ex-
planation of this sort, both for these characters and for poly-
dactyly would much more readily account for the aberrant
nature of the ratios obtained in breeding experiments.

SUMMARY

The observation reported in this paper establish the fact
that, at least in certain strains of poultry (and probably in
pigeons) there is a close correlation between brachydactyly
involving the size and number of bones in the feet, and the
presence of feathers on the tarsi. The embryology of both
conditions has been studied, the evidence from this source
indicating that while the size and number of skeletal elements
are determined by the length of the embryonic toe, there is no
causal relation between toe length and feathering. The data
cited seem to indicate, on the contrary, that brachydactyly,
feathering of the tarsi, and probably syndactyly are all de-
pendent on one and the same factor. No attempt is made to
postulate the nature of this factor but it is suggested that a
study of the early functioning of the endocrine glands in normal
and abnormal embryos might throw some light on the question.

No. correlation could be detected between either of these
characters and polydactyly or comb form.

BRACHYDACTYLY IN THE FOWL its

LITERATURE CITED

Antuony, R. 1899 Etude sur la Polydactylie chez les Gallinacés. Jour. de
l’Ant. et de la Physol., T. 35, pp. 711-750.

Bateson, W. 1909 Mendel’s principles of heredity. Cambridge University
Press.

Bateson, W., AND PunNETT, R.C. 1905 Poultry. Jn Report to the Evolution
Committee of the Royal Society. Report II, pp. 99-131.

Bepparp, Frank E. 1898 The structure and classification of birds. Long-
mans, Green & Co., London, New York, and Bombay.

Davenrort, C. B. 1906. Inheritance in poultry. Publications of the Carnegie
Institution of Washington, no. 52.
1909 Inheritance of characteristics in domestic fowl. Carnegie In-
stitution of Washington, Publication no. 121.

Horst, C. C. 1905 Experiments with poultry. In Report to the Evolution
Committee of the Royal Society. Report II, pp. 131-154.

KAvuFMANN-Wo Lr, Marie 1908 Embryologische und anatomische Beitrige zur
Hyperdactylie (Houdanhuhn). Morph. Jahrb., Bd. 38, 8S. 471-531.

Lituin, Frank R. 1908 The deyelopment of the chick. Henry Holt & Co.,
New York. ;

Moraan, T. H., Sturtevant, A. H., er AL 1915 The mechanism of Mendelian
heredity. Henry Holt & Co., New York.

Resumido por la autora, Della Drips.

Estudios sobre el ovario del esperméfilo, con especial menci6én del
cuerpo amarillo.

En el presente trabajo se considera el ciclo de cambios que
ocurren anualmente en los ovarios del espermO6filo, dando detal-
ladas descripciones histol6gicas del cuerpo amarillo en cada uno
de los estados de su desarrollo. Se han empleado coloraciones
especificas para poner de manifiesto los caracteres nucleares y
protopldsmicos de las células amarillas. En el ciclo vital del cuerpo
amarillo se reconocen tres fases: Primera, una fase que se carac-
teriza por la presencia de un gran nimero de granulos rojos, in-
dudablemente grénulos de secreccién, en el protoplasma de las
células amarillas. Esta fase comprende, practicamente, todo el
periodo de la prefiez. Segunda, la fase lipoide, asi, llamada por
la abundancia de gotitas de substancia lipoide en el protoplasma
de dichas células. Esta fase comienza algin tiempo antes del
parto y dura pr6éximamente unas seis semanas después de este,
que es también pr6éximamente el tiempo que requiere para com-
pletarse la involucién normal del utero. Tercera, la fase de re-
eresion. Se dan a conocer también ciertos estudios experimen-
‘tales, tales como los efectos de la ovariotomia sencilla y doble,
practicada sobre animales prefiados y no prefiados. La ovario-
tomia sencilla produce resultados negativos. La ovariotomia
doble en animales no prefiados causa una atrofia funcional del
utero, que se manifiesta muy gradualmente. En las hembras
prefiadas esta operacién da lugar a abortos, excepto cuando se
practica cuando la prefiez est’ muy avanzada. Como resultado
de los estudios histol6gicos y experimentales, la autora llega a la
conclusién de que los cuerpos amarillos producen dos secrecciones
internas que presiden sobre los cambios que tienen lugar en el
utero a consecuencia de la prefiez. La primera secreccién pro-
duce la implantaci6n normal y desarrollo del embrién, y la se-
creccién lipoide ulterior ayuda a la involucién normal del utero.

Translation by Dr. José F. Nonidez
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 24

STUDIES ON THE OVARY OF THE SPERMOPHILE
(SPERMOPHILUS CITELLUS TRIDECEMLINEATUS)
WITH SPECIAL REFERENCE TO THE CORPUS
LUTEUM!

DELLA DRIPS

Mayo Foundation, Rochester, Minnesota
Division of Experimental Surgery

TWENTY-NINE FIGURES

CONTENTS

LLNS IRS s RSS os SEG AN Sy les ao ke. al gala ce oe lane te te ge eae ae Se Pe 117
The ovarian cycle. A. Material and methods. B. Histologic descriptions
of ovaries at consecutive intervals of the annual cycle. 1. The ovary of
the early fall. 2. The ovary of the late fall. 3. The ovary of the early
spring. 4. The ovary of early pregnancy with corpora lutea of the red
granule stage. 5. The ovary of the middle of pregnancy with its char-
acteristic corpora lutea. 6. The ovary at parturition with its char-
acteristic corpora lutea. 7. The ovary of July with corpora lutea of
the lipoid stage. 8. The ovary of the late summer with corpora lutea
TEE, RETEST 011 | LN pala eG + Coe 2 ie oe a SOY a Oe eer eS 126
Experimental studies. A. Effects of the removal of the uterus on the
ovaries. 1. In non-pregnant spermophiles. 2. In pregnant sper-
mophiles. B. Effects of the removal of both ovaries on the uterus. 1,
In non-pregnant spermophiles. 2. In pregnant spermophiles. C,
Effects of the removal of one ovary. 1. In non-pregnant spermophiles.
2. In pregnant spermophiles. D. Effects on the remaining ovary of the
removal of one ovary and the uterus in pregnant spermophiles. E.
Studies on the normal bursting of follicles and the formation of corpora.
lutea. F. Studies on ovulation, fertilization and corpora lutea forma-

Summary and Discussion. A. Summary of the results obtained from the
histological and experimental studies. B. Conclusions as to the func-
tions of the corpora lutea in the ovaries of spermophiles............... 160

LITERATURE

The first papers written on the corpus luteum of the ovary had
to do largely with its formation, particularly with the origin of
the luteal cells. Were these cells connective-tissue elements

1 Thesis for degree of Master of Science, University of Minnesota (Mayo
Foundation, 1917).
1ly/

118 DELLA DRIPS

from the internal theca of the follicle according to the hypoth-
’ esis of von Baer, or were they epithelial in nature and derived
from the membrana granulosa of the follicle as advocated by
Bischoff? Marshall, in his book, ‘‘The Physiology of Repro-
duction” (’10), carefully reviews this early literature.

It was not until Sobotta (96) published the first of his series
of papers on the corpus luteum that the discussion over the origin
of the luteal cells began to come to an end. Sobotta’s study on
the corpus luteum of the mouse was the first systematic record
of the transformation of the follicle into a corpus luteum and
the latter’s subsequent development. He describes, first, the
follicle about to burst; then, one just after bursting; one, one-half
hour afterward, and others at succeeding short intervals up to
seventy-two hours after bursting. All the descriptions are very
detailed, every change in the structure as it proceeds in its de-
velopment being noted. Sobotta traces the origin of the luteal
cells to the epithelial cells of the granulosa of the follicle. The
internal theca cells, he says, expend themselves utterly in the
formation of connective tissue and blood-vessels. The external
theca remains as it was.

‘Sobotta believed the function of the corpora lutea was to
maintain a constant equal tension in the ovaries, which shows
how little thought he gave to this phase of the problem. He was
all absorbed in the formation of the luteal structure. This paper
of Sobotta’s was followed shortly by a very similar one on the
rabbit, in which he confirmed all of his findings in the mouse.
These papers started a real investigation of the corpus luteum,
and for the following ten or fifteen years there were any number
of papers written regarding it. A few writers continued the
discussion as to the origin of the luteal cells. Some criticised
Sobotta, trying to disprove his statements. Among these was
Clarke (’98), who contributed an account of the formation of the
corpus luteum in thesow. His studies were made on serial sections
of pig’s corpora lutea and follicles in different but not subsequent
stages of development. The sections in each set were treated
alternately to a process of tryptic digestion and a picro-fuchsin
stain. From a study of these sections he concludes that the

THE OVARY OF THE SPERMOPHILE 119

luteal cells are of a connective-tissue origin and that the function
of the corpus luteum is to preserve the ovarian circulation.

Sobotta (99) published his third paper on the corpus luteum
of mammals. Honore (’00) was one of the first to confirm
Sobotta’s findings. He wrote concerning the corpus luteum in
the rabbit. Marshall (01) published a paper on the corpus
luteum in the sheep, coming to practically the same conclusions
as Sobotta. Cohn (’03) further confirmed Sobotta’s work on
the rabbit. Jankowski (’04) published a paper in which he
came to entirely different conclusions. They are as follows:
‘““He (Sobotta) simply lets the internal theca vanish. If one
layer must vanish, it would be the one for whom the conditions
after the follicle bursting are very unfavorable and that is the
case with the epithelium of the follicle. The corpus luteum is
not an epithelial but a connective tissue structure.” Sobotta
(06) published a fourth paper on the formation of the corpus
luteum in the guinea-pig, again confirming his former work.

Such an amount of histologic investigation over the formation
of the corpus luteum could not very well go on without arousing
much interest in regard to the physiologic function of the struc-
ture. The men who upheld the connective-tissue origin held
many curious ideas as to its function, all of which tended toward
making its action more or less mechanical. Several thought the
only function of the luteal structures was to prevent ovulation.
Clarke thought its function was to preserve the ovarian
circulation. —,

On the the other hand, those who accepted Sobotta’s con-
clusions began looking for a much more important function for the
corpus luteum. If the luteal cells were epithelial and each was so
intimately in contact with the blood stream, and the whole organ
had so much the appearance of a gland, why couldn’t it be a
gland of internal secretion? Several histologists began studying
the cells for evidences of a secretory product. Regaud and
Policard in 1901 were the first to publish any results. They
stained sections of ovaries of the dog, that had been fixed in
acetic potassium bichromate, with a copper-hematoxylin method
of Weigert. They described some black secretion droplets in

120 DELLA DRIPS

the luteal cells. In Cohn’s paper previously mentioned, he
describes some granules which he thinks may be the same as
those described by Regaud and Policard. He also gives his
reasons for believing that the fatty osmic-stained droplets,
seen in the luteal cells in greatest abundance when the cells
reach their maximum hypertrophy, are real secretion droplets
and not evidences of fatty degeneration as had been previously
contended.

Along with the attempt to discover by a special histologic
technic, evidences of secretory products in the luteal cells, a
number of men sought to discover, by animal experimentation,
proofs of the glandular action of the corpus luteum. The uterus
and the mammary glands, in the minds of all, were the most
closely associated of all the organs of the body, with the ovaries.
Ovulation had been observed to be closely related to menstrua-
tion in the human and to the heat periods in animals. Complete
double ovariectomy was known to stop. menstruation, cause an
atrophy of the uterus, and, in young women, to bring on meno-
pause symptoms. In the very young castration had prevented
the development of the uterus and mammary glands. Double
ovariectomy in pregnant women, especially in the first months,
was known to be followed by abortion.

Various attempts had been made to overcome the bad effects
of ovariectomy by the transplantation of ovaries and by ovarian
medication. These were reported to have given good results,
at least partially.

All these facts seemed to speak for the production of an
internal secretion in the ovaries which affected the uterus and
mammary glands.

Fraenkel (’06) was the first to attempt to prove by a series of
experimental studies on rabbits and cats that the corpora lutea
in the ovaries were responsible for the various effects produced
on these organs.

He begins his papers as follows: ‘‘The corpus luteum must,
from its structure and development, be a gland of internal
secretion, made to insure the implanting and development of
the fertilized egg in the uterus. The corpus luteum maintains

THE OVARY OF THE SPERMOPHILE a

the state of nutrition of the uterus during the years of sexual
activity.”” He thought it ruled over the phenomena of rut as
well as pregnancy.

About this same time Marshall and Jolly published a series of
experiments very similar to Fraenkel’s. They used dogs and
rabbits, and their conclusions were much like those of Fraenkel.

Daels (08) published a paper in which he gives his several
objections to Fraenkel’s theories and records a series of experi-
ments on guinea-pigs. In his first series of experiments he tried
to determine the influence of bilateral ovariectomy on the
pregnant animal, concluding that this operation in the pregnant
animal always interrupts pregnancy during more than the first
half of its duration. He also tried giving Fraenkel’s lutein
tablets, with no results.. He had better results from a product
of the whole ovary.

Ancel and Bouin (’08-’09) contributed several papers to the
literature on the corpus luteum. They believed with Fraenkel
that the corpus luteum produced rut and the other changes
incident to pregnancy. They performed a series of experiments
on rabbits in which they produced an unfertile coitus either
between a normal female and a male in which a part of the vas
deferens had been resected, and a female in which a part of the
uterus had been resected and a normal male. They wished by
these experiments to eliminate any action of the egg and the
placenta on any changes taking place in the uterus incident to
ovulation. They describe structural changes in the uterus and
mammary glands for a period equal to the period of activity
of the yellow body.

In this same year (’09) two other Frenchmen, Regaud and
Dubreuil, published several articles. ‘They were particularly in-
terested in the cause of rut and ovulation. They made asystem-
atic study of a large number of uteri and ovaries in different
phases of the genital cycle and concluded that rut is independent
of the corpora lutea, and that it is improbable that the corpora
lutea plays a réle in originating the pregestative changes in the
uterus, for the graphic curve of their development is much later
chronologically than the curve of its changes. They claim that

122 - DELLA DRIPS

coitus only will bring about ovulation. The congestive phenom-
ena which they notice in the ovaries during rut will not produce
the rupture of a single follicle without coitus.

Niskoubina, the same year (’09), published a series of studies
which confirmed the observations of Ancel and Bouin. He first
made a histologic study of a series of ovaries removed at varying
intervals after coitus and then did some experiments similar to
Fraenkel’s to determine the period during which the ovaries
seem to exert an influence on the pregnant uterus. From his
experimental studies, he concludes: ‘‘The corpus luteum exer-
cises an obvious action on the physiology of pregancy. It puts
the uterus in a condition necessary to assume the development
of the fertilized egg. This action lasts during the first half of
pregnancy, after which it ceases to act.”

Loeb (’08) published a paper in which he states that deciduo-
mata can be produced experimentally in the uterine mucosa of
guinea-pigs by making a number of transverse and longitudinal
cuts so as to break the continuity of the tissue. He states that
this can happen only during a certain definite period after
copulation, between the fourth and the eighth days. This is
the time when freshly formed corpora lutea are present in the
ovaries. These changes were not excited by the presence of
ova, since they took place when the uterus was ligated and the
passage of ova prevented. If the ovaries were removed, de-
ciduomata could not be produced. He concluded, then, that
the ovaries at certain periods after ovulation elaborate a pre-
disposing substance, in the presence of which indifferent stimuli
may produce deciduomata.

Parhon, Dumitresco and Nissipesco (’09) published a paper
on the lipoids of the ovary. From various staining reactions
on sections of ovaries and from chemical reactions of a powder
made from ovaries, they conclude that in the interstitial cells
of the ovary and in those of the yellow body fats are found which
differ in many characteristics from the fat of the adipose tissue.

Mulon (’10) reported that the fat in the corpus luteum, which
stains only faintly with osmic acid, was similar to other fats
found in the adrenal and other organs of the body which form a

THE OVARY OF THE SPERMOPHILE 123

class of fats different from the ordinary body fat. He thought
these so-called glandular fats had to do with the neutralization
of the glandular excretions or the ordinary poisons formed in
the cellular activity. The specific action of the fat of the corpus
luteum was to neutralize the poisonous products formed by the
developing embryo.

Miller (’10, ’14) published some studies on human corpora
lutea. He tried out many fat stains on fresh corpora lutea and
claimed they contain no fat. He says the negative result of
the fat reaction on fresh corpora lutea makes it possible to tell
the difference easily between these and other ovarian structures.
When the involution of the yellow body begins, the neutral fat
reaction begins. The peripheral parts show the fat reaction
first. In the corpus luteum of pregnancy the reaction to neutral
fat remains negative to the end of pregnancy. He says that in
a corpus luteum of five days the fat reaction was negative, in
one of six to eight days there was a little, in one of eleven days
more, and in one of twelve to sixteen days the cells were rich in
fat (not neutral fat).

Meyer (’11, 713) wrote on the human corpus luteum. In his
first paper he described the development and regression of the
human corpus luteum of pregnancy, which confirms Sobotta’s
and Cohn’s work.

Van der Stricht (’12) published the results of his studies on the
corpora lutea and the interstitial cells of the ovary of the bat.
This is one of the most valuable studies ever published on the
corpus luteum, because the ovary of the bat, with its contained
structures, is one of the simplest ever studied. No confusion
arises from old corpora lutea; these are gone before the new ones
are formed, as the periods of ovulation are so far apart, occurring
each spring only. The author describes two secretions in the
luteal cells. The first is a serous secretion which is very like the
liquor of the follicle. This is secreted by the cells from the time
of bursting until about the time the egg enters the uterine horns.
The second is a lipoid secretion. Beginning some time after the
bursting, there is a slow elaboration of fatty granules from the
depth of the cytoplasm of the luteal cells, the amount of which
increases as the cells increase in size.

124 DELLA DRIPS

He says: ‘‘Far be it from our idea of admitting two abso-
lutely distinct phases for serous secretion and lipoid secretion.
On the contrary, at a moment in the development of the corpus
luteum these two processes coexist, but the first is especially
marked at the beginning of gestation and the second exists alone
during the following period.’”’ He concludes that: ‘‘the serous
secretion exercises its influence on the transformations of the
uterine mucosa of the first phase of gestation during the dis-
placement of the egg and that the lipoid secretion intervenes
principally to provoke the arrest and fixation of the blastocyst
and the formation of the placenta.” Van der Stricht has not
been able to demonstrate this serous secretion in the luteal cells,
but because it is present in the central cavity of the young corpus
luteum and in the lymphatics when they are first formed, he
thinks the cells must be secreting it. Both the serous secretion
and the lipoid secretion are carried away, he says, by the
lymphatics.

Corner (’16) published a paper on the corpus luteum of preg-
nancy in swine. He claims to have found in the corpora lutea
of pregnancy, beside the cells which are descendants of the
granulosa cells and the cells which are descendants of the theca
cells, two more types of cells which can be found at all stages of
pregnancy.

Livon (’09) published the results of the effect of injecting
luteal extract into guinea-pigs. He writes as follows: “We have
employed an extract of the corpora lutea of the sow and the
cow, a product that I have today called the Product A. In-
jected into the peritoneal cavity of guinea-pigs, we find a toxic
action varying with the rapidity of the absorption and with the
individual. The toxic.dose obtained generally is 20 to 30 centi-
grams per kilogram of animal. The animals die presenting
general tremors,dyspnea, convulsions and uttering weak cries.”’

Champy and Gley (11) are said to have been the first to show
that the corpus luteum from pregnant cows was exceedingly
active, whereas that from non-pregnant animals possessed little
physiologic action.

Hare (712) reported very favorably on the clinical use of
corpus luteum extract.

THE OVARY OF THE SPERMOPHILE 125

Frank and Rosenbloom (’15) published the results of some
experimental work done on rabbits with extracts of the placenta
and the corpus luteum. They claim to have gotten better results
than former investigators because they used a more concentrated
and an alcoholic (fat solvent) solution of the active substance
of the luteal extract. This active substance, they say, “‘is not a
lipoid but is carried along the lipoids.’”’ They state that the
only corpus luteum substance extracted, which was found to be
active, was derived from pregnant animals.

Dannreuther (’14) reported his results obtained clinically from
the use of corpus luteum extract. He calls attention to the
necessity of using the extract of pregnant animals only.

Up to date, the most noteworthy publication concerning the
clinical value of the extract of the corpus luteum is that of
Culbertson (’16), entitled, ‘‘A study of the menopause.” He
regards the climacteric as a’ ‘‘functional disarrangement on the
part of the endocrine glands, the ovarian secretion having
ceased.” His theory concerning the value of luteal extract in
the treatment of menopause disorders in as follows:

“Thus the chief characteristic stamping the vasomotor dis-
turbances of the climacteric seems to be a disarrangement of the
systolic-diastolic relation producing elevation in the pulse
pressure. In blood-pressure estimations, then, we find a fairly
reliable measure of the vasomotor disturbances of the menopause,
as will be shown, a satisfactory method of treatment.

“Tf we accept the propositions thus far laid down, that the
cessation of ovarian activity leads to a functional over-efficiency
on the part of the pituitary and adrenal glands and that this, in
turn, produces an arterial hypertension, the corollary is that by
the administration of corpus luteum extract, the pressor sub-
stances will be neutralized and the tension will decrease.”’

In summing up, it may be stated that up to 1906 practically
all the literature on corpus luteum had to do with the histologic
origin and consequent structure at various succeeding periods in
its life cycle. The writings of Sobotta and Cohn practically
established the epithelial’ nature of the luteal cells and the
glandular character of the structure as a whole. From 1906 to

126 DELLA DRIPS

1912 most of the papers written were reports of experimental
studies undertaken to prove that the corpus luteum is the gland
of internal secretion in the ovaries, and that, through this
secretion, the luteal structure produces specific effects on other
organs, particularly the uterus. In spite of the many criticisms .
directed against his work, Fraenkel still stands preéminent
among the experimental workers who established beyond a doubt
the foregoing hypothesis.

Since 1910, efforts have been put forth by investigators actually
to demonstrate this secretion in the corpus luteum. Van der
Stricht comes nearer the goal than any others. The greater
part of recent literature, however, concerns the extract of the
corpus luteum, its chemical constituency, its physiologic action,
and its clinical value. .

THE OVARIAN CYCLE

In the summer of 1914, while studying microscopic sections of
the various tissues and organs of the spermophile, the relatively
immense size of the ovaries, compared with those observed the
previous spring, strikingly presented itself. On further compari-
son, it was very evident that this great increase in size had been
brought about by a growth in the corpora lutea only. One
ovary contained eleven of these bodies; there remained only
a framework of ovarian stroma with a few atretic follicles.

With the particular: stain the luteal cells bore a marked re-
semblance to the cells of the cortex of the adrenal. The former
- were much larger, but the shape of both, their arrangement in
columns, the position and appearance of the nuclei, and the
presence of lipoid droplets in the protoplasm accounted for the
likenesses. In fact, the luteal cells resemble secreting cells.

From observations that had been going on, it was known that
these spermophiles had given birth to young about a month
before. According to most writers, degeneration of the luteal
cells begins not later than birth. Here were what looked like
actively secreting cells a month after birth. ‘Thereupon it was
decided to try out some differential stains on these luteal cells
at every stage in their life history, and to study the complete

THE OVARY OF THE SPERMOPHILE Wea

ovarian cycle in the spermophile with a view to gaining some
accurate information of the origin, development, life history,
and function of the corpora lutea of the ovary.

In the spring of i915, numbers of spermophiles were captured.
But not having realized how very soon the animals become im-
pregnated after coming out of hibernation, no strenuous efforts
were made to obtain them until they were quite numerous in
the fields, and consequently they were found to be either in
advanced stages of pregnancy or lactating. However, all the
animals that could be gotten were used for a study of the ovarian
cycle through the summer and fall, until hibernation began.

Several animals were sacrificed each week. They were killed
quickly with ether and bleeding. The ovaries were immediately
placed in one of several fixatives, 10 per cent formalin, Zenker’s
fluid with acetic acid, Bensley’s formalin Zenker, and Bensley’s
acetic acid bichromate. Many stains were experimented with.
After considerable study of the fixed and stained sections, it was
decided that for the problem in hand, two fixatives seemed
best, Bensley’s formalin Zenker and Bensley’s acetic osmic
bichromate—the former particularly for the nuclear structures
and the latter for the elements in the protoplasm. In all the
work of the past spring (’16), one ovary of each animal sacrificed
was routinely placed in formalin Zenker and the other in acetic
osmic bichromate.

Of the sections fixed in zenker, the best results were obtained
with a modified Weigert stain (copper-chrome-hematoxylin),
Ehrlich’s hematoxylin and eosin, Mallory’s connective-tissue
stain, and Bensley’s acid fuchsin and methyl green. A few
sections of each series were prepared with these stains.

Of the sections fixed in acetic osmic bichromate, one of each
series was stained with the Weigart stain and several with the
acid fuchsin and methyl ‘green of Bensley. Complete paraffin
serial sections were made of all the ovaries studied.

The spermophiles went into hibernation about the middle of
October although many of them became partially torpid earlier
than this. The next spring it was determined to get them early
enough. The frost was not out of the ground until about

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, Nu. 2

128 DELLA DRIPS

April 15. As a result of strenuous efforts, the females were
obtained while in rut and every day through the period of
pregnancy, which was found to be twenty-eight days. Two or
more were sacrificed each day.

The ovarian cycle was now completed. Ovaries were at hand
for every week of the year except during the hibernation period
when only a sufficient number of animals was sacrificed to make
sure there were no changes taking place in these organs. For
the period of pregnancy, a time which is especially related to
ovarian activity, there were ovaries for even fractions of a day.

The period of rut evidently follows immediately on the awaken-
ing of the spermophiles in the spring. Ovulation follows on
coitus. Most of the females are impregnated in a very short
time. The period of pregnancy follows. From the time of
ovulation until about September 1 the ovaries contain corpora
lutea. These approach their greatest size about July 1. From
September 1 to 15 the large corpora lutea have disappeared.
With a disappearance of the corpora lutea, there is a very notice-
able rapid growth of the follicles together with a noticeable
decrease in size. The, ovaries which in July consisted almost
entirely of large corpora lutea with a small amount of ovarian
stroma containing a few atretic follicles, by September 15 con-
tained no corpora lutea, but instead, many medium and good-sized
growing follicles containing very little liquor folliculi, but filled
with mitotic figures.

In this paper descriptions will be presented, gross and micro-
scopic, of the ovaries which, as the year goes around, show the
characteristic changes of their cycle.

A typical ovary of the early fall will be presented first; second,
one of the late fall just before hibernation, with its inactivity,
and, third, one of the early spring, showing the characteristics
of the rutting season. For the fourth period, the period of
pregnancy, several ovaries will be described, marking the suc-
cessive changes occurring in the corpora lutea principally.
Finally, will be presented descriptions of a number of ovaries
of the summer months, which show the final stages in the life
history of the corpora lutea and the corresponding notable
differences in the rest of the ovary.

THE OVARY OF THE SPERMOPHILE 129

Experiment 379-15 (spermophile 200). Captured during the spring
of 1915. The ovaries were removed surgically September 14, 1915.
Weight, 150 grams.

Microscopic observations of an ovary. Fixative, formalin zenker.
Stain, hematoxylin and eosin (fig. 10). There are nine good-sized
follicles, the larger of which measure about 0.4 x 0.4 mm. There are
twice as many follicles half as large. There is a goodly number of
small hyalinized follicles. Around the periphery of the ovary are
numbers of primordial ova. The larger follicles are approaching
maturity, but are still growing. They contain many mitotic figures
and very little liquor folliculi. They are located through the cortex
of the ovary, only three being near the surface. None of these larger
follicles appear to be atretic. Through the medullary portion of the
ovary are conspicuous clumps of interstitial cells.

The spermophiles begin to become torpid about September 15, but
they are active by spells for some time after this, depending on weather
conditions. In the laboratory, some are active until November 15.

Experiment 503-15 (spermophile 254). Captured during the spring
of 1915. Sacrificed November 15, 1915. The animal had been hiber-
nating six days.

Microscopic observations of an ovary. Fixative, formalin zenker.
Stain, hematoxylin and eosin. This ovary appears very similar to
that of spermophile 200. The larger follicles are about the same in
number, size, and location. There are about the same number of
smaller and hyalinized follicles as well as primordial ova. The size
of the ovaries as a whole, however, has decreased. This is probably
due to a marked decrease in the size of the blood-vessels and sinuses.
The clumps of interstitial cells are much less conspicuous. The
ovaries appear to have prepared themselves for their long period of
functional inactivity.

Experiment 246-16 (spermophile 291). Captured April 25, 1916.
Was injured in being caught, so was sacrificed immediately. Weight,
170 grams.

Gross observations of the uterus. The animal is in rut. The
rutting season evidently lasts about two weeks. During this time
practically all the females become impregnated. It may be stated
here that the laboratory animals come out of hibernation much earlier.
Those killed from the 1st to the 15th of March appeared to be in rut
and one killed April 11 was found to be pregnant, showing that weather
conditions set the time of the commencement of their sexual activities.
The large size of the uterus is immediately noted (fig. 7). There has
been an increase in length and breadth. It is twice as large as the
inactive uterus. The cervix shows the greatest increase in size and it
is filled with a thick mucoid substance. The walls of the vagina are
swollen. It is very evident that the great increase in the size of the
whole organ is due to a marked increase in the fluid content, which
gives it a pale appearance. (For measurements see table.)

130 DELLA DRIPS

Gross observations of the ovaries. The ovaries of rut show several
changes from those of the fall. They are larger and seem to be slightly
congested. On the surface of each ovary may be seen several slightly
raised, tiny, colorless, cyst-like bodies which are the mature follicles.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stain, hematoxylin and eosin. Serial sections show ten larger follicles,
all of which lie at the surface of the ovary and several of which are
projecting slightly from the surface. One of these follicles measures
0.5 mm. x 0.4 mm., and another 0.5 mm. x 0.6 mm. None of them
shows atretic changes and all are apparently mature. There is con-
siderable liquor folliculi, and few mitotic figures among the granulosa
cells (fig. 12).

The internal and external theca of these large follicles are distinct
and comparatively thick layers. There is a well-marked membrane
between the stratum granulosum and the internal theca. Acetic osmic
bichromate sections stained with acid fuchsin and methyl green show
a red secretion along this membrane and between the cells of the gran-
ulosum near it. The cells of the granulosum seem to have an increased
amount of protoplasm which makes them larger.

Of the smaller follicles, a few are in good condition, but the greater
number have become hyalinized. The hyalinization of so many of
the smaller follicles gives a characteristic appearance to the ovaries of
the spring (fig. 11). Indeed, it would seem that a few mature follicles
have been produced at the expense of many. The primordial ova are
few in number. ‘The interstitial cells are not at all conspicuous.

Toward the end of the rutting season many uteri present a different
picture. They are smaller and instead of appearing edematous appear
congested.

Experiment 256-16 (spermophile 298). Date of capture April 27,
1916. Sacrificed April 28, 1916. Weight, 105 grams. (For measure-
ments of uterus see table.)

Gross observations of the uterus. The blood-vessels to the uterus
are all much congested and the organ itself shows some congestion
throughout. There is one especially congested area in each horn
about 1 em. from their point of union. There is another specially con-
gested area in the body near the point of union of the horns.

As there are no corpora lutea in the ovaries of this spermophile, the
uterus is still one of rut, not pregnancy. What brings about this
change in the uterus is not evident. The congested condition is, how-
ever, without doubt preparatory to the reception of the fertilized ova.

This brings us in the life cycle to the ovaries of pregnancy. It
has been shown that the ovaries of the fall, winter, and early
spring contain no corpora lutea. Occasionally some remains of
these bodies of the previous year may be found, but this is very
unusual. Thus when coitus takes place during rut and the

THE OVARY OF THE SPERMOPHILE 131

follicles burst and become transformed into corpora lutea, these
new bodies are the only corpora lutea in the ovary. They are
all produced simultaneously and they also develop simul-
taneously if they are not abnormal in some way.

The picture of the ovary changes, then, when fertilization
brings on pregnancy. From this time on until the period of the
growing follicles is reached in the late summer, the ovaries con-
tain corpora lutea. These are the predominating structures in
the ovaries of the sprmg and summer. The changes which take
place in these organs during the period have to do with the
corpora lutea principally. The following descriptions of ovaries
will be attempts at picturing them with corpora lutea of various
ages. As this study is very largely concerned with these luteal
structures, they will be described in considerable detail. Prob-
ably the ovaries which contained the very youngest corpora
lutea seen were those in which ovulation took place in the
laboratory, the animals being sacrificed very shortly afterward.

Experiment 268-13 (spermophile 303). Captured May 3, 1916:
Sacrificed the same day. Weight, 112 grams.

Gross observations. There are no visible signs of pregnancy in the
uterus except congestion. The ovaries contain several slightly raised,
small spherical bodies which resemble mature follicles except that they
are red or pink instead of colorless.

Microscopie observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. This ovary con-
tains five luteal bodies, three of which are normal and two of which
are not. Beside these, there are several growing medium-sized fol-
licles, no large ones, quite a number of small atretic follicles and a few
primordial ova. The interstitial cells are not as conspicuous in this
ovary as they were in some of the ovaries of early pregnancy. All the
blood-vessels and sinuses in the medullary portion of the ovary are
very much dilated. Most of the ovary is made up of the five corpora
lutea. Serial sections of the three normal ones show them to be of
different sizes from 0.7 mm. x 0.8 mm. in diameter to 1.3 mm. x
0.1 mm., depending on the amount of blood which they contain, for
practically all the young corpora lutea contain blood in their centers
(figs. 13 and 17). A hemorrhage from a blood-vessel in the wall of the
follicle must occur as the follicle bursts. The exact point of this burst-
ing cannot be made sure of in any of these structures. Each luteal
body is surrounded by a very thin connective-tissue capsule, no doubt
the same theca externa which surrounded the follicle. From this thin
capsule to the central core of blood are massed the luteal cells. They

1382 . DELLA DRIPS

seem to have no definite arrangement. Among them may be seen
numerous fibroblasts and endothelial cells. These are most numerous
about the periphery of the central mass of blood. Some are already
making their way intoit. The cells all seem to have their axes directed
radially as if they were approaching the central mass of blood from the
periphery of the luteal body. The luteal cells are of various sizes and
shapes. Some are spindle-form and some polygonal, but the majority
are spherical or oval. There is one specific characteristic of all young

luteal cells and that is the existence of spherical granules in their pro- —-

toplasm. In the sections fixed in Bensley’s acetic osmic bichromate
and stained with acid fuchsin and methyl green, these granules are
strikingly brought out (fig. 26). They are colored a brilliant red.
Their sizes vary somewhat, but they are all spherical. The protoplasm
of some cells is so full of these granules that it resembles a homogeneous
red secretion, but on examination with very high power, the separate
granules may be seen. In many cells where the granules do not fill
the protoplasm, they are grouped about the nucleus, leaving a narrow
clear zone about the periphery of the cell. The nuclei take the green
stain and are strikingly brought out against the red granules. Each
nucleus contains one or two good-sized bright, red staining nucleoli.
The chromatin threads stain green and do not show very well with
this stain. For the nuclear characteristics, another ovary stained
differently will be described.

The two abnormal luteal structures of this ovary are very interesting.
One of them is a luteal cyst (fig. 23). There is only a single layer of
luteal cells next to the thin capsule. No fibroblasts or endothelial
cells are present. The whole body is filled with a transparent greenish-
colored fluid which has every appearance of being of the same compo-
sition as the liquor folliculi.

The other structure (fig. 21) is much more normal. The center con-
tains what resembles the above green-tinged fluid containing numer-
ous red blood corpuscles. Masses of normal luteal cells are found most
of the way around the body. But on one side is a mass of follicle or
granulosa cells persisting untransformed. In several places, as in
figure 22, some granulosa cells are found among the luteal cells. This
peculiar luteal structure appears to give striking evidence to the theory
that luteal cells are simply transformed granulosa cells. The differ-
ences between them are well brought out in the picture. |

It may be well to state here that other fixatives and stains bring
out these same specific luteal-cell characteristics, particularly the
granules in the protoplasm. With a formalin zenker fixative and the
acid fuchsin and methyl green stain, the granules appear the same in
every way except in color. They are pink instead of red. With this
same fixative and a copper-chrome hematoxylin stain, the granules
appear brownish-black. With an acetic osmic bichromate fixative and
the copper-chrome hematoxylin stain, the granules appear bluish-black.
The nuclear characteristics of the early luteal cells are best brought
out with a formalin zenker fixative and a hematoxylin and eosin stain.

THE OVARY OF THE SPERMOPHILE 133

Experiment 275-16 (spermophile 310). Captured May 1, 1916.
One ovary was removed and the uterus ligated May 4, 1916. -

Gross observations. There are no recognizable signs of pregnancy
in the uterus, but the ovaries contain what resemble young corpora
lutea.

Microscopic observations of thé ovary removed. Fixative, formalin
zenker. Stain, hematoxylin and eosin (fig. 13).

The nuclei of these early luteal cells are strikingly like those of
the follicle cells. The nuclei of the latter have several small nucleoli
with quite conspicuous chromatin strands. The nuclei of most of
these early luteal cells also have several small nucleoli with numerous
chromatin strands. But in some cells, the nucleus is much larger and
contains only one or two larger and darker nucleoli, while the chro-
matin strands are finer and fewer.

A very few of the luteal cells in the early luteal structures show
mitotic figures. There were none at all in those of spermophile 303.
Of all the early corpora lutea studied, mitotic figures were found in
these structures in the ovaries of only three spermophiles. These
were apparently the earliest luteal bodies found. If mitosis occurs, as
a rule it occurs just after the bursting of the follicle. It may be that
the luteal cells which show mitotic figures are the transformed follicle
cells undergoing mitosis as the bursting occurred.

There is one abnormal early luteal structure in the ovary (spermo-
phile 310) which deserves mention, as it seems quite common and
furnishes further proof that the luteal cells are simply transformed
granulosa cells (fig. 24). The structure with this staining appears at
first glance like a mature follicle. The liquor folliculi is present and
the ovum lies over at one side of the central cavity against the sur-
rounding cells, which resemble the granulosa cells of the follicle.
On closer observation, however, it will be seen that the cells which
were thought to be the epithelial cells of the follicle are larger, richer
in protoplasm, and more irregular in shape and size. Scattered through
them are numerous fibroblasts and endothelial cells. These fibroblasts
are quite numerous about the ovum, as if they were attempting to wall
it off. The ovum has been stripped of its own rim of granulosa cells
and appears to be undergoing degeneration. The internal theca is
missing. In fact, what we have here is a corpus luteum formed in a
follicle which, if it burst, did not throw out enough of its contents to
get rid of the ovum. Practically all-of the epithelial cells must then
have been retained. Where are they if they are not the luteal cells?
The only missing cells are those of the internal theca and the only
new cells are the fibroblasts and endothelial cells. Does it not appear
reasonable that the internal theca cells which are of the same origin
as the connective-tissue cells expend themselves in the formation of
the new fibroblasts and endothelial cells? It is hard to account for
these abnormal structures. Several of them showed blood in the
central cavity, as if the normal hemorrhage had occurred into them.
Possibly they are formed in the mature follicles that are ready to burst,

134 DELLA DRIPS

and undergo the same changes incident to this phenomenon as the
others except that, on account of not occupying a position close enough
to the surface of the ovary, they are not able to discharge their contents.

Experiment 293-16 (spermophile 328). Captured May 4, 1916.
Both ovaries were removed May 6, 1916. Weight, 119 grams.

Gross observations. The fetuses-in the uterus measure about 2
mm. in length, which makes the luteal bodies in the ovaries older than
those previously described.

Microscopic observations of the right ovary. Fixative, formalin
zenker. Stain, acid fuchsin and methyl green. This ovary contained
six corpora lutea. These luteal bodies appear differently, due prin-
cipally to the rapid growth which has been going on among the fibro-
blasts and endothelial cells. These ovaries suffered some congestion
through the manipulation of removal, and this helps to show the great
numbers of capillaries and blood-vessels that have been formed in a

_short time. The central mass of blood is undergoing rapid organiza-
tion. No doubt the presence of this blood with its serum and fibrin is
the great attractive force which aids in the complete formation of the
luteal body. Fibroblasts and endothelial cells are always attracted by
serum and fibrin. As soon as the hemorrhage occurs in the follicle,
they start in to organize it. This is evident from the radial direction
which the axis of the fibroblasts all take very early. As they go into
the center, the transformed epithelial cells of the follicle are carried in
by them. Endothelial cells grow in, and so very early there is formed
in the corpus luteum a complex system of blood-vessels and capillaries,
as is seen in sections of this ovary.

In one of the corpora lutea in this ovary, the hemorrhage was so
extensive that instead of trying to organize it, the fibroblasts have.
formed a wall around it. Since the fibroblasts have not penetrated
very far, there is only a narrow rim of luteal cells. This structure is
what is ordinarily called a hemorrhagic luteal cyst (fig. 25). Beside
the corpora lutea in a section of this ovary, one notices readily the large
clumps of interstitial cells through the medullary portion. These are,
as a rule, conspicuous in the ovary of early pregnancy.

Microscopie observations of the left ovary. Fixative, formalin
zenker. Stain, hematoxylin and eosin. This ovary contains only two
corpora lutea. There are present in it several large, apparently mature
follicles. The number of such follicles in an ovary evidently depends
on the number of corpora lutea. Where there are a good many of the
latter, the follicles evidently cannot grow. When there are only a few
corpora lutea in an ovary, one or two follicles may reach the size of
0.5 mm. x 0.5 mm., or 0.56 mm. x 0.7 mm. These will, of course,
degenerate as ovulation takes place only once a year, during the rut-
ting season which has just gone by. And as there are practically no
large follicles ever seen in the ovaries of July which contain the largest
luteal bodies, they must degenerate before this time. Perhaps the
pressure of even one large, growing corpus luteum is enough to bring
this about.

THE OVARY OF THE SPERMOPHILE 135

Experiment 264-16 (spermophile 302). Captured May 1, 1916, and
sacrificed the same day.

Gross observations. This animal is definitely pregnant, the fetuses
measuring 7 mm. in length. The blood-vessels going to the uterus
and the ovary are very much congested. They stand out sharply,
showing plainly the blood supply to the two organs. The blood supply
to the ovary is practically separate from that to the uterus, there being
only one small anastomosing branch close to the ovary. There are
what look like corpora lutea in the ovaries, but they cannot be counted
with any certainty. They resemble little reddish-pink cysts sticking
out from the surface of the ovary. Naturally, from the size of the
fetuses, the corpora lutea in these ovaries are older than those of sper-
mophile 328.

Microscopic observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. There are four
luteal structures in this ovary. They show some changes over the
younger ones previously described. They are slightly larger, measur-
ing about 0.7 mm. x0.8mm. The luteal cells have become larger and
more regular in shape. Many more have assumed an oval form and
all seem to be approaching this. They seem to be tending toward a
radial arrangement also. This is being effected evidently by the ar-
rangement of the connective-tissue strands. The latter are running
from the capsule to the central core of blood, which is almost organ-
ized. The system of capillaries and blood-vessels is even more com-
plex than that of the luteal body of spermophile 328. In the luteal
cells themselves the red granules have increased in number and vary
slightly in size. They occupy the same position in the cell. The
nuclei of all the cells are larger and contain one or two large bright
nucleoli with numbers of very fine chromatin strands.

Experiment 296-16 (spermophile 331). Captured and _ sacrificed
May 6, 1916. Weight, 146 grams.

Gross observations. The fetuses in the uterus are 1 ecm. in length,
which lead us to expect to find changes in the corpora lutea of the
ovaries.

Microscopie observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The size ofthe
luteal structures has increased. There are two in this ovary, one
measuring 0.9 mm. x 0.9 mm., and the other, 1 em. x 0.8 mm. The
most noticeable feature of this later luteal body is the absence of any
blood in the center. Instead, there is a core of connective tissue. The
size of this connective-tissue core depends evidently on the amount of
hemorrhagic material there is to organize. In some bodies the core is
much larger than in others. Very perceptible strands of connective
tissue run from this central mass to the capsule, enclosing columns of
luteal cells. Gross strands have developed also so that connective-
tissue strands seem to be enveloping each cell. There is a complete
capillary network following the arrangement of the strands of connec-
tive tissue. Good-sized blood-vessels are located about the periphery

136 DELLA DRIPS

of the structure. Several smaller ones are present in the central con-
nective-tissue core. There are several sinuses about the periphery
lined with endothelium which appear to contain lymph. The lutea
cells themselves have increased in size. Many more have taken on an
oval shape and are lying with their long axes perpendicular to the
capsule. The same red granules are still present in the protoplasm.
These do not seem to be quite as numerous in the cells, which fact is
in part due, no doubt, to the increased size of the latter. The granules
are now found scattered throughout the protoplasm, the clear zone
about the periphery of the cell having disappeared. The granules
show more difference in size than formerly, but they are still all spher-
ical. The nuclei have not changed. In a few cells there is a clear
space in the protoplasm on one side of the nucleus.

Microscopic observations of the right ovary. Fixative, formalin
zenker. Stain, hematoxylin and eosin. There are seven corpora lutea
which make this ovary larger than the other one. Beside these bodies,
this ovary contains approximately five good-sized, growing follicles
about 0.4mm.x 0.4mm. Four follicles nearly the same size are under-
going atretic changes and there are about twenty small follicles, some
of which are degenerating. Around the edge of the ovary are a few
primordial ova. Through the medullary portion are some small
groups of interstitial cells. They are not nearly as conspicuous as
they were earlier in pregnancy (fig. 14).

Experiment 355-16 (spermophile 375). Captured May 20, 1916.
Both ovaries were removed May 22, 1916. Weight, 128 grams.

Gross observations. The fetuses in the uterus are 1.5 cm. in length.

Microscopic observations of the right ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. There are six
corpora lutea in the right ovary, measuring about 0.8 mm. x 0.9 mm. in
diameter. Something is.noted in the luteal structures in this ovary which
has not been seen before, namely, that there are present in the cells close
to the periphery some osmic-stained droplets. The size of the corpus
luteum and the size of the individual cells are about the same as that
of spermophile 331. There is a slight increase in the number* of red
granules in the cells. The clear space next to the nucleus is present
in many more cells. The osmic-stained droplets are located at the
periphery of the cell. They are very large compared with the red
granules. They vary somewhat in size, but not in shape; all are
spherical. Of course, in sections fixed with formalin zenker and stained
with the various stains which were used, these lipoid droplets appeared
as vacuoles. But they could be easily recognized by their correspond-
ing size and location in the cells. For convenience, these droplets
will be called lipoid droplets, because they certainly are a lipoid prod-
uct. They do not appear in the luteal cells before the fetus is about
1.5 cm. in length or about fourteen days old, that is, until the period
of pregnancy is half over. These droplets make up the ‘lutein’ of the
corpus luteum. which has been described for many years and which
has given the corpus luteum its name.

THE OVARY OF THE SPERMOPHILE © 137

When the lipoid droplets appear in the cells, the period of the red
granules is waning. The latter seem to reach their crisis of abundance
when the fetus measures about 8 mm. in length. But the granules are
still very abundant in the cells until the lipoid droplets begin to appear. .
From this time on, the former grow fewer and fewer and the latter
increase in number, as will be shown, until they, too, reach a crisis of
abundance and then decline.

Experiment 363-16 Garenonhale 383). Captured and _ sacrificed
May 24, 1916. Weight 140 grams.

Gross observations. The animal was in labor when killed. The
uterus still contains two live fetuses, four having already been born.
The crown-rump measurement of a fetus is from 4 to 5 em.

Microscopic observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The luteal struc-
tures in the ovaries of this animal show some changes over those pre-
viously described (fig. 18). There are four of them. They have
increased slightly in size, measuring 0.9 mm. x 0.9 mm. and the indi-
vidual cells have increased correspondingly. Still more noticeable
than their increase in size is the regularity of their oval form and the
uniformity with which all their axes point in a radial direction. This
seems to be due to an increased amount of connective-tissue frame-
work, which, from the first, has seemed to govern the position and
shape of the cells. The increase of connective-tissue framework has
been accompanied by an increase in the size of the blood-vessels and
capillaries. In the luteal cells (fig. 27) the red granules have decreased
still more than in the cells last described, and this is very general
throughout the structure. The lipoid droplets are much more numer-
ous in all the cells. In fact, it seems hard to tell which is the predom-
inant product of the cells, the red granules or the black droplets. The
nuclei of these cells are slightly larger than the ones of the preceding
,description. Otherwise, they are the same. Most writers have agreed
“that degeneration of the corpus luteum begins about the time of birth.
None is in evidence here. Several writers have stated that the prin-
cipal reason for their belief was the entrance into the cells of osmic
staining droplets which they considered to be evidences of fatty degen-
eration in the cells. It would not seem from the nuclear character-
istics of the cell nor from the color, shape, and the regular size of the
droplets, that they could be fatty degeneration products, especially
when compared with the true fatty degeneration which occurs much
later in the life history of the corpus luteum and which will be described
accordingly.

Microscopic observations of the right ovary. Fixative, formalin
zenker. Stain, hematoxylin and eosin. There are two corpora lutea
in this ovary. A striking feature is the number of good-sized atretic
follicles. There are at least twelve. This seems to be a noticeable
feature of other ovaries about this same time. In fact, the follicles
which were growing when ovulation took place, evidently go on and
develop if there are not too many corpora lutea in the ovary, but as

138 DELLA DRIPS

the latter structures begin to increase in size quite rapidly about the
time of parturition, there seems to be a degeneration of all the mature
follicles even where there are only one or two luteal structures present.
There are five medium-sized and about twenty small growing normal
follicles, only a few of which show atretic changes. The primordial
ova are very few in number, and the interstitial cells can scarcely be
distinguished from the connective-tissue cells of the stroma.

Experiment 368-16 (spermophile 387). Captured May 24, 1916.
Sacrificed May 25, 1916, twelve hours after having given birth to
young. Weight, 115 grams.

Microscopie observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The three luteal
structures in the left ovary of the animal measure either 0.9 mm. x
1.1 cm. or 0.9 mm. x 1 em. The appearance of the luteal structure
as a whole and of the individual cells is very similar to that of sper-
mophile 383.

Experiment 369-16 (spermophile 388). Captured May 24, 1916, and
sacrificed May 26, 1916, twenty-four hours after having given birth
to young. Weight, 115 grams.

Microscopic observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The corpora lutea
of the ovaries of this animal measure 0.8 mm. x 0.9 mm. There is
scarcely any new noticeable difference unless it is a slight increase in
the number of lipoid droplets.

Experiment 370-16 (spermophile 389). Captured May 24, 1916.
It gave birth to young on May 25, 1916. It was with the young until
May 26, 1916, when it killed them. The next day, May 27, 1916, the
animal was sacrificed. Weight, 135 grams.

Microscopic observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The corpora
lutea in this ovary show a marked increase in the amount of lipoid in
the luteal cells. The red granules have correspondingly decreased in
number. There are no other differences except a slight increase in
size of the luteal structure and the cells.

Experiment 372-16 (spermophile 391). Captured May 18, 1916,and
gave birth to normal young May 22, 1916. The young were destroyed
May 26, and the animal was sacrificed May 31, 1916.

Microscopic observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The corpora lutea
in the ovaries show the lipoid content to be still more increased in
amount and the red granules to be very scarce. There are no other
differences. It might be stated here that there is some variation in
the time when this lipoid change comes on. For instance, the corpora
lutea of the ovaries of spermophile 390, ten days after parturition, do
not contain as much lipoid as those of spermophile 391, nine days after
parturition. But the majority of luteal bodies for any given time
before or after parturition appear very similar.

oe

THE OVARY OF THE SPERMOPHILE 139

Experiment 351-16 (spermophile 371). Captured May 20, 1916.
The left ovary was removed on May 22, 1916. Weight, 135.5 grams.

Gross observations. The fetuses in the uterus are 2 em. in length.
This animal gave birth to young May 31, 1916, and was sacrificed
June 17, 1916.

Microscopic observations of the right ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The individual
luteal cells are of about the same size and shape. The red granules
are still quite abundant, but there seems to be a marked difference in
the cells as to their individual content of red granules. This appears
to vary with the number of lipoid droplets in the cell. In a few cells
where the latter are very abundant, the red granules are found only
in a rim about the periphery. Where the lipoid droplets are still few
the cell protoplasm still contains many red granules with these few
droplets scattered among them or occupying a peripheral position.
Many cells have not changed at all; they have no lipoid droplets.
Some cells have a peculiar appearance. Their protoplasm appears
honeycombed. Since none of the cells had this appearance in the
unstained sections, it was concluded that the cells must have been filled
with lipoid which was dissolved out in the staining process and the
result was this honey-combed appearance to the protoplasm. This
was later proved to be the case. The lipoid droplets, when they aré
numerous, show much less variance in size than when they are few in
number in the cell. The nuclei of these cells appear very similar to
those of earlier cells. In some cells there is a slight change in their
position. Instead of being directly in the center, in the cells filled
with lipoid, the nucleus lies a little to one side of the center. The Ist
of July, about thirty-five days after parturition, the corpora lutea in
the ovaries were larger than at any other time.

Experiment 412-16 (spermophile 415). Captured and sacrificed
July 1, 1916. Weight, 125 grams.

Gross observations. The ovaries are the largest yet seen, owing
to the comparatively immense size of the corpora lutea. These are
now of a yellowish-cream color and stand out prominently so that the
ovary looks as if it were made up of several spherical bodies 1.5 mm. in
diameter. Any other ovarian tissue except that holding the spherical
bodies together can scarcely be distinguished. There are three corpora
lutea in the right ovary and four in the left, which numbers correspond
to the tiny white spots marking the former placental sites in the uterus.

Experiment 413-16 (spermophile 416). Captured and sacrificed on
July 3, 1916. Weight, 157.3 grams.

Gross observations. The right ovary appeared grossly just like that
of spermophile 415 (fig. 15).

Microscopic observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The corpora
lutea are much larger than those previously described (fig. 19). They
measure 1.1 mm. x 1.5 mm. in diameter (1.7 mm. x 1.3 mm., grossly
some shrinkage). The luteal cells are correspondingly larger, and their

140 DELLA DRIPS

protoplasm is absolutely full of a mass of lipoid droplets. This is seen
to advantage in the unstained sections. In the stained sections the
cells appear more or less honeycombed, according to the amount of
lipoid which has been dissolved out. These lipoid droplets are very
uniform in size and are dark brown, quite a different color from the
black fat droplets of fatty degeneration. The red granules are gone.
Where there is any protoplasm visible, it appears granular and pinkish-
gray in color (fig. 28). The nuclei of the cells appear slightly smaller
than formerly, but this apparent decrease in size is evidently due to
the increase in the size of the cells, for the nuclei are no smaller by
measurement. The location of the nucleus in the cell is the same as
described under spermophile 371, either in the center or to one side of
the center. The nucleolus stands out large and bright and the chro-
matin strands appear as they did. There are no apparent degenerative
changes. Beside the three corpora lutea in this ovary, there are six
or seven medium-sized normal growing and three atretic follicles.
There are no large follicles or even any of good size. A few primordial
ova are present, fewer than in any of the ovaries described thus far,
and no interstitial cells can be distinguished as such (fig. 15).

Experiment 439-16 (spermophile 436). Captured June 23, 1916.
Sacrificed July 15, 1916. Weight, 194.2 grams.

Gross observations. The uterus still shows several tiny white spots
marking the placental sites. It is otherwise normal. The ovaries
appear to contain corpora lutea, but these latter are certainly much
reduced in size over those of spermophile 416. They appear congested
or of a reddish-yellow color.

Microscopie observations of the left ovary. Fixative, acetic osmic
bichromate. Stain, acid fuchsin and methyl green. The corpora
lutea are much smaller (fig. 20). They measure 0.7 mm. x 0.8 mm.
and 0.9 mm. x 0.9mm. The cells are smaller. The protoplasm of
the cells contains no red granules and very few lipoid droplets or any
honeycombing suggestive of these. It has a grayish, granular appear-
ance. Something is present, however, which has not been seen before,
and that is fat. Scattered here and there thoughout the luteal struc-
ture, fat globules, characteristic of fatty degeneration, are present in
the protoplasm of the cells. They are of various sizes and take on a
characteristic black color with the osmic acid in the acetic osmic bichro-
mate fixative. The nuclei of the cells show degenerating changes.
The nucleolus has disappeared in some cells and in others appears
pale and fringed. The chromatin strands are fewer and appear clumped
in some cells. One striking new feature in the luteal structure is the
great increase in the size of the blood-vessels and capillaries. The con-
gestion of blood is not common to the whole ovary, but is only in the
corpus luteum. The vascular change is apparently one factor in the
disappearance of the luteal body.

Experiment 515-16 (spermophile 458). Captured August 18,1916.
Sacrificed August 21, 1916. Weight, 190 grams.

Gross observations. There are no evidences of placental sites in the
uterus. There are no signs of corpora lutea in the ovaries.

THE OVARY OF THE SPERMOPHILE 141

Microscopic observations of the left ovary. Fixative, acetic osmic ®
bichromate. Stain, acid fuchsin and methyl green. The left ovary
shows three luteal stuctures (fig. 16). The largest measures 0.8 mm.
x 0.8mm. No red granules or lipoid droplets are discernible. The
greenish-gray granular protoplasm is everywhere filled with various
sized fat droplets. In some cells these fat droplets are so large that
they occupy nearly the whole cell, squeezing the degenerated nucleus
out to one side of the cell (fig. 29). The nuclei are so degenerated,
no. chromatin strands or nucleolus are recognizable as such. What
is left of the nucleus takes the acid fuchsin rather than the methyl
green stain—so it is red. All the blood-vessels and capillaries are
markedly dilated. There is a marked increase of connective tissue
throughout the body. The thin capsule of connective tissue which
before surrounded the luteal structure seems to have disappeared in
places, making it appear as if the connective tissue of the body were
continuous with that of the ovary around it. This connective-tissue
invasion is evidently another factor in the disappearance of the corpus
luteum. Three factors, then, are associated with the disappearance of
the corpora lutea in the ovaries, cellular degeneration, vascular dilata-
tion, and connective-tissue Invasion.

A word may be added concerning the changes in the ovary
outside the luteal bodies. There were in this organ six or seven
good-sized, normal, growing follicles and four and five atretic
ones. The most noticeable feature is the number of small
hyalinized follicles. This is a noticeable feature of all the
ovaries of this date which contain old corpora lutea. It would
appear that as long as there are still luteal bodies in the ovary,
there is very little growth in the follicles. There are a few
primordial ova. The interstitial cells are not recognizable as
such. By September 1 the corpora have disappeared from the
ovaries and the organs again have assumed the appearance
described for September 15.

There are occasional exceptions to the normal cycle. An
ovary of spermophile 462, sacrificed October 30, 1916, was
found with a little structure in it which appeared in every way
to be a young corpus luteum. The cells contained red granules.
Perhaps impregnation had occurred in the fall. This instance is
mentioned because one function of the corpus luteum in spermo-
philes has been suggested by it. What keeps the follicles from
becoming mature and ovulation from taking place at other
times in the year from the spring? Evidently, the corpora

142 DELLA DRIPS

v lutea help in regulating the periods between oestrus. No new
follicles become mature in the fall as long as there are luteal
bodies in the ovary and by the time the latter have disappeared
it is too late for the follicles to become mature before the period
of hibernation comes on.

A word concerning the interstitial cells may be added. These
appear very prominent at two periods of the cycle, during the
period of early pregnancy when the corpora lutea are young
and apparently very active, and in the early fall when the fol-
licles are developing and growing rapidly. Their significance is
not evident. Several writers have claimed that the cells of the
corpus luteum become the interstitial cells of the ovary. There
is absolutely no basis for such a supposition as far as the spermo-
philes are concerned. |

EXPERIMENTAL STUDIES

These experiments were made with a view to determining
whether or not the results of certain experimental studies on the
spermophile would substantiate the same work done on other
animals.

The spermophile stand experimental surgery well. They are
very satisfactory to work on as they are not nearly as susceptible
to infectious diseases and are more resistant to local infections
than most small laboratory animals. They are easily anes-
thetized. The surgical technic must be aseptic, and the animals
must be placed in separate cages after the operation or they will
chew one another’s wounds open. They recover from the effects
of the operation quickly, as a rule, and the wounds usually heal
by first intention.

To all workers who have been interested in the functional
relationship of the ovaries and the uterus and especially to those
trying to isolate a particular function or functions for the corpora
lutea, two problems have seemed of vital importance: First,
what are the effects on the uterus of the removal of one or both
of the ovaries? And, second, what are the effects on the
ovaries of the extirpation of the uterus? ‘These two problems

THE OVARY OF THE SPERMOPHILE 143

have a special bearing on the function of the corpora rhe
when they dea! with pregnant animals.

Another much studied problem closely related to the {unction
of the luteal bodies attempts to expla'n the means by wh ch
ovulation s brought about and the results on the development
of the corpora lutea i fertilization is prevented.

Following are protocols and results of experiments performed
in connection with these various problems. Only those experi-
ments which proved operative successes and in which reliable
data were obtained wil! be included.

Series 1. Effects of the removal of the uterus on the ovaries of non-
pregnant spermophiles

Experiment 230-15 (spermophile 146). Captured in the spring of
1915. Weight, 128 grams. Operated on June 11, 1915. Complete
removal of the uterus.

Gross observations. There are good-sized corpora lutea in the
ovaries.

Sacrificed July 20, 1915.

Gross observations. The ovaries appeared very much smaller.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stain, hematoxylin and eosin. The ovaries show no degenerative
changes except in the corpora lutea. These are undergoing regression.
Connective tissue has heavily invaded every part of the luteal struc-
ture. The blood-vessels are numerous and good sized. There are
many fat vacuoles present. The decrease in size of these ovaries is
no doubt due to the decrease in the size of the corpora lutea.

Experiment 443-15 (spermophile 247). Captured in the spring of
1915. Weight, 207 grams. Operated on October 1, 1915. Complete
removal of the uterus. '

Died October 8, 1915.

Gross observations. Death was due to peritonitis.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stain, hematoxylin and eosin. There is a marked congestion through-
out the ovary. Many of the larger follicles are undergoing degenera-
tion. Practically all the smaller follicles are markedly degenerated.
This, no doubt, is a pathologie condition.

Experiment 444-15 (spermophile 248). Captured in the spring of
- 1915. Weight, 196 grams. Operated on October 1, 1915.

Died January 13, 1916.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stain, hematoxylin and eosin. The great number of small hyalinized
follicles are immediately apparent. They are so numerous that the

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 2

144 DELLA DRIPS

ovary has a lacy appearance. There are a few good-sized, apparently
normal follicles and on the periphery are numbers of primordial ova.

Experiment 445-15 (spermophile 249). Captured in the spring of
1915. Weight, 110 grams. Operated on October 4, 1915. Complete
removal of the uterus.

Sacrificed April 22, 1916.

Gross observations. The blood supply to the ovaries is intact.
The ovaries are very small; the left so small as to leave doubt as to its
identity. The right ovary appears to contain several tiny cysts.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stain, hematoxylin and eosin. Tissue removed for left ovary proves
to be a bit of granulation tissue. The right ovary appears to be nor-
mal. The several cysts noted grossly are large mature follicles which
normally occur on the surfaces of the ovaries at this time of the year.
No effects from the removal of the uterus are noted.

Discussionand summary of results. There seemsto be no marked
effect on the ovaries from the removal of the uterus They seem
able to repeat their life cycle as far as the follicles are concerned.
It was thought that the great number of hyalinized follicles noted
in the ovaries of spermophile 248 was a sign of degeneration due
to removal of the uterus. However, on studying the ovaries of
normal spermophiles killed about the same time for controls,
there were found in the ovaries of two a great number of these
hyalinized follicles (fig. 11). Evidently the hyalin degeneration
occurs normally at this time of the year. It is, no doubt, part
of the attempt to produce a few large mature follicles at the
expense 0 many smaller ones.

Series 2. Effects of removal of the uterus on ovariés in pregnant
animals

Experiment 270-16 (spermophile 305). Captured May 1, 1916.
Weight, 98 grams. Operated on May 4, 1916. The uterus was
‘removed to the cervix.

Gross observations. There are no signs of pregnancy in the uterus,
but the right ovary contains what appears to be a hemorrhagic cyst.

Died May 9, 1916.

Gross observations. There was no apparent cause for death.

Microscopic observations of one ovary. Fixative, formalin zenker. -
Stain, hematoxylin and eosin. The ovary contains several corpora
lutea of an early stage. One is markedly hemorrhagic. There is some
degeneration apparent in the luteal bodies. Some cells are undergoing -
chromatolysis. There is an increased amount of connective tissue.
The rest of the ovary appears to be in a very normal condition.

THE OVARY OF THE SPERMOPHILE 145

Experiment 294-16 (spermophile 329). Captured May 4, 1916.
Weight, 117 grams. Operated on May 6, 1916. The uterus was
removed to the cervix.

Gross observations. The animal appears to be in early pregnancy.

Sacrificed May 20, 1916. Weight, 130 grams.

Gross observations. There are no adhesions about the ovaries;
they appear to contain several white, transparent corpora lutea.

Microscopic observations of the right ovary. Fixative, formalin
zenker. Stains, hematoxylin and eosin, and acid fuchsin and methyl
green. There is no apparent pathologic condition of this ovary. Even
the corpora lutea appear normal. They have proceeded in their
development without the uterus and now have the same appearance
as others of the same age. The only detectable difference might be a
smaller number of red granules in the luteal cells.

Experiment 251-16 (spermophile 348). Captured May 6, 1916.
Weight, 108 grams. Operated on May 9, 1916. The uterus was
removed to the cervix.

Gross observations. The animal is pregnant. There are five
placental swellings in each horn, measuring 6.5 mm. in length.

Sacrificed May 19, 1916. Weight, 102 grams.

Gross observations. Everything in the abdominal cavity is in
excellent condition.

Microscopic observations of the ovaries. Right ovary: fixative
acetic osmic bichromate; stain, acid fuchsin and methyl green. Left
ovary: fixative, formalin zenker; stain, Weigert’s copper-chrome
hematoxylin. The ovaries are apparently normal. The corpora lutea
appear like those of the controls except that the luteal cells contain a
larger number of lipoid droplets and fewer red granules. There are no
signs of degeneration in the nuclei of the cells.

Experiment 314-16 (spermophile 349). Captured May 6, 1916.
Weight, 122 grams. Operated on May 9, 1916. The uterus was
removed to within 3 mm. of the cervix.

Gross observations. The animal is pregnant. There are four
placental swellings in the right horn and five in the left about 4 mm.
in length.

Sacrificed June 5, 1916. Weight, 135 grams.

Gross observations. Everything in the abdominal cavity is in
excellent condition. Both ovaries were readily found and appear to
contain large corpora lutea.

Microscopic observations of the ovaries. Right ovary: Fixative,
formalin zenker; stain, Weigert’s copper-chrome hematoxylin. Left
ovary: fixative, acetic osmic bichromate; stain, acid fuchsin and
methyl green and Weigert’s copper-chrome hematoxylin. There is no
apparent abnormality in the ovaries. The corpora lutea have devel-
oped normally. They are still in the red-granule stage. The cells
contain some lipoid droplets, but are full of red granules. The lipoid
droplets in these ovaries are not as numerous as would be expected,
but normal ovaries vary somewhat as to the time when the lipoid
droplets begin to appear in the cells.

146 DELLA DRIPS

Experiment 317-16 (spermophile 353). Captured April 25, 1916.
Weight, 116 grams. Operated on May 9, 1916. The uterus was
removed to the cervix.

Gross observations. The animal is pregnant and, having been kept
by itself since capture, the period of pregnancy must be at least fourteen
days. The placental swellings measure 13 mm. in length and 10 mm.
in width.

Sacrificed May 20, 1916. Weight, 130 grams.

Microscopic observations of the ovaries. Right ovary: fixative,
formalin zenker; stains, hematoxylin and eosin, and Weigert’s copper-
chrome hematoxylin. Left ovary: fixative, acetic osmic bichromate;
stain, acid fuchsin and methyl green. There is no pathologic condi-
tion apparent in the ovaries. The corpora lutea have gone on in their
normal development. They appear very ‘similar to those of spermo-
philes whose fetuses are nearing parturition.

Experiment 318-16 (spermophile 353). Captured April 28, 1916.
Weight, 127 grams. Operated on May 9, 1916. The uterus was
removed to the cervix.

Gross observations. The animal is pregnant and, since it has been
kept by itself since capture, it must have been pregnant at least eleven
days. The placental swellings measure 7 mm. in length.

Sacrificed May 31, 1916. Weight, 120 grams.

Gross observations. The ovaries appear very small. The blood
supply is intact and there are no adhesions. There is a mass of fat
around each ovary.

Microscopic observations of the ovaries. The right ovary was
carefully studied. The other was lost. Fixative, formalin zenker.
Stains, hematoxylin and eosin, and acid fuchsin and methyl green.
Many degeneration changes are present all through the ovary. Prac-
tically all the follicles are degenerating. The corpora lutea show many
degeneration changes also. There are practically no red granules in
the cells and many fatty degeneration vacuoles are present. The
nuclei of the cells show degenerative changes. Their margins are
indented and their nucleoli are faded and fringed. The strands of
chromatin are clumped. The capillary sinuses seem dilated and
filled with blood. Evidently the ovaries have been injured by the
operation.

Experiment 331-16 (spermophile 361). Captured May 4, 1916.
Weight, 141 grams. Operated on May 12, 1916. The uterus was
removed to the cervix.

Gross observations. The animal is pregnant, the placental swell-
ings being 9 mm. in length.

Sacrificed May 20, 1916. Weight, 120 grams.

Microscopic observations of the ovaries. Right ovary: fixative,
formalin zenker; stains, acid fuchsin and methyl green, and hema-
toxylin and eosin. The left ovary was lost. The right ovary shows no
abnormality. The corpora lutea have gone on in their development.
They are just entering the lipoid stage.

THE OVARY OF THE SPERMOPHILE 147

Experiment 61-16 (spermophile 381) Captured May 20, 1916.
Weight, 145.5 grams. Operated on May 22, 1916. The uterus was
removed to the cervix.

Gross observations. The animal has given birth to young, probably
a few days previously, as the uterus appears to be in an early stage of
involution. There are good-sized corpora lutea in the ovaries.

Sacrificed October 20, 1916. Weight, 170 grams.

Gross observations. The ovaries are so small they are hard to find.

Microscopic observations of the ovaries. Right ovary: fixative,
formalin zenker; stain, hematoxylin and eosin. Left ovary; fixative,
acetic osmic bichromate; stain, acid fuchsin and methyl green. There
is nothing abnormal about the ovaries. They are very similar to other
ovaries removed at this time. The corpora lutea have disappeared
normally.

Discussion and summary of results. Removal of the uterus
has no apparent pathologic effect on the ovaries conta ning the
corpora lutea of pregnancy. The latter pass through their
norma’ cycle of development and regression. There seems to
be a slight irregularity about the time at which the various
changes in the life cycle come on; for instance, the corpora lutea
in the ovaries of spermophile 348 seem to lose the red granules
and take on the lipoid droplets sooner than normally. Then, in
the ovaries of spermophile 349, the corpora lutea have retained
their red granules longer than usual. However, this may not
be significant since there is some irregularity about the cycle
normally.

Series 3. Effects of the removal of both ovaries on the uterus of
non-pregnant animals

Experiment 200-15 (spermophile 189). Captured in the spring of
1915. Weight, 130 grams. Operated on May 28, 1915. Both ovaries
were removed.

Gross observations. The uterus is in a stage of early involution.

Sacrificed April 22, 1916.

Gross and microscopic observations of the uterus. Fixative, for-
malin zenker. Stain, hematoxylin and eosin. The uterus is very
small. It shows no signs of rut of having been in rut, either grossly
or microscopically. This is very abnormal for this time of the year.
A comparison of this uterus with a normal uterus in rut at this time of
the year brings out the differences. (Compare figs. 1 and 3; also
2and 4.) A comparison of the dimensions of the two uteri brings out
the marked differences in size (table).

148 DELLA DRIPS

Microscopic observations of the uterus. Fixative, formalin zenker.
Stain, hematoxylin and eosin. Cross-sections through the horns and
body of the uterus show much fibrosis of all the layers. There is
much less muscular and glandular tissue than in a normal inactive
uterus (compare figs. 3 and 5, 4 and 6). One striking feature is the
number of obliterated blood-vessels with hyalinized walls. The lumen
of the uterus is closed and no mucous secretion is present. The cervix
shows these same changes to an even greater degree than the rest of
the organ.

Experiment 201-15 (spermophile 140). Captured in the spring of
1915. Weight, 101 grams. Operated on May 28, 1915. The ovaries
were completely removed.

Gross observations. The uterus is in the condition of early involu-
tion.

Postoperative notes. This animal became very fat and went into
a torpid condition at times. On July 24 it was changed to a cage with
two other animals who killed it. The body was badly traumatized
and specimens were not saved.

Experiment 202-15 (spermophile 141). Captured in the spring of
1915. Weight, 108 grams. Operated on May 28, 1915. Both ovaries
were removed.

Gross observations. The uterus is undergoing involution.

Died September 21, 1915.

Gross and microscopic observations of the uterus. Fixative, for-
malin zenker. Stain, hematoxylin and eosin. This uterus appears
very similar to the uterus of a normal animal killed the same day.

Experiment 218-15 (spermophile 142). Captured in the spring of
1915. Weight, 135 grams. Operated on June 3, 1915. Both ovaries
were removed.

Gross observations. The uterus is undergoing involution.

Sacrificed November 10, 1915.

Gross and microscopic observations. Fixative, formalin zenker.
Stain, hematoxylin and eosin. The uterus is very similar to the uteri
of the controls.

Experiment 220-15 (spermophile 151). Captured in the spring of
1915. Operated on June 28, 1915. Both ovaries were removed.

Gross observations. The uterus has become completely involuted.

The uterus was removed September 28, 1915.

Gross and microscopic observations. Fixative, formalin zenker.
Stain, hematoxylin and eosin. Compared with a normal control, this
uterus shows no abnormalities. .

Experiment 423-15 (spermophile 231). Captured in the spring of
1915. Weight, 200 grams. Operated on September 22, 1915. Both
ovaries were removed.

Sacrificed March 18, 1916. Weight, 100 grams.

Gross and microscopic observations of the uterus. Fixative, for-
malin zenker. Stain, hematoxylin and eosin. This uterus shows the
‘same changes as that of spermophile 139.

THE OVARY OF THE SPERMOPHILE 149

Experiment 424-15 (spermophile 232).. Captured in the spring of
1915. Weight, 205 grams. Operated on September 22, 1915. Both
ovaries were removed.

Gross observations. The uterus is normal.

Sacrificed May 8, 1916. Weight, 117 grams.

Gross and microscopic observations. Fixative, formalin zenker.
Stain, hematoxylin and eosin. This uterus shows the same changes as
that of spermophile 139. (Photograph of this uterus with a normal
uterus of rut removed the same day (fig. 7). Table for measurements
of these uteri).

Experiment 435-15 (spermophile 242.) Captured in the spring of
1915. Weight, 210 grams. Operated on September 28, 1916. Both
ovaries were removed.

Sacrificed May 9, 1916. Weight, 155 grams. The same day a
control animal, spermophile 233, which was captured the same time as
spermophile 242, and had lived in the laboratory under the same con-
ditions, was sacrificed also and the uteri of these two animals were
photographed together (fig. 8).

Gross and microscopic observations. Fixative, formalin zenker,
Stain, hematoxylin and eosin. The uterus of spermophile 242 shows
all the changes noted in that of spermophile 139. (For dimensions of
uterus see table.) The control uterus (table) does not show as marked
enlargement as the uteri of animals brought in from the fields in the
condition of rut, but aside from this variation in size, the uterus ap-
pears in every way like a typical one of rut.

Experiment 436-15 (spermophile 246). Captured in the spring of
1914. Weight, 95 grams. Operated on July 10, 1914. Both ovaries
were completely removed.

Sacrificed September 29, 1915.

Gross and microscopic observations. Fixatve, formalin zenker.
Stain, hematoxylin and eosin. Compared with the uteri of the con-
trols, this uterus shows atrophic changes of a similar nature to those of
the uterus of spermophile 139, except that the atrophy must be even
more marked to be noticed in a comparison with the controls of this
time of the year.

Experiment 354-16 (spermophile 874). Captured May 20, 1916.
Weight 128 grams. Operated on May 22, 1916. Both ovaries were
removed.

Gross observations. The uterus is undergoing involution.

Sacrificed October 30,1916. Weight, 120 grams. This uterus was
photographed with that of a control animal and that of an animal
Co eens and sacrificed on the same date—spermophile 378
fig. 9).

Experiment 358-16 (spermophile 378). Captured May 18, 1916.
Weight, 115 grams. Operated on May 22, 1916. Both ovaries were
removed.

Sacrificed October 30, 1916. Weight, 135 grams. This ae was
photographed with that of spermophile 374 and a control.

150 DELLA DRIPS

Gross and microscopic observations. Some slight differences can be
noted between the uteri of spermophiles 374 and 378 and their control.
The uteri of the doubly ovariectomized spermophiles are both smaller
than their control. This decrease in size is more noticeable in the
cervix than in the rest of the uterus. The cervices of these two uteri
are much firmer also and contain very little mucus. It is true, the
differences are only slight and are scarcely recognizable microscop-
ically, especially after fixation. The main microscopic difference is a
decrease in size of the blood-vessels in the uteri of the ovariectomized
spermophiles. Perhaps this accounts for the fact that the uterus of
the control has a healthier appearance grossly.

Discussion and summary of results. The uteri of animals
doubly ovariectomized in the spring of the year show some
changes over their controls by the fall of the same year. These
are slight and are all of the nature of a functional atrophy. The
cervix is affected the most. This atrophy increases so as to be
quite noticeable by the fall of the next year. The very striking
effect of double ovariectomy is the discontinuation of the changes
in the uterus incident to the phenomena of rut.

Series 4. Effects of removal of both ovaries on the pregnant uterus

Experiment 293-16 (spermophile 328). Captured May 4, 1916.
Weight, 119 grams. Operated on May 6, 1916. Both ovaries were
removed.

Gross observations. The animal is in an early stage of pregnancy,
the placental swellings being just visible grossly in the uterus.

Sacrificed May 20, 1916. Weight, 115 grams.

Gross observations. There are no signs of placental sites in the
uterus. To be sure the animal had been pregnant, the ovaries were
studied carefully. They contained early corpora lutea. Evidently
the regression changes began in. the uterus immediately after the re-
moval of the ovaries.

Experiment 300-16 (spermophile 335). Captured in the spring of
1916. Weight,101 grams. Operated on May 8,1916. Both ovaries
were removed.

Gross observations. The animal was in an early stage of pregnancy,
the placental swellings being just large enough to be recognizable.

Died May 12, 1916.

Gross observations. The external wound is in bad condition, per-
haps the fault of too much iodin. The inside of the abdominal cavity
appears normal. The uterus is in good condition save on the ends
where the blood supply has been injured. The congestion is much
reduced. There are placental swellings, hard and dark red; five in
the right horn and three in the left, very hard to see. The placentas
and fetuses are apparently undergoing degeneration.

THE OVARY OF THE SPERMOPHILE Lt

Experiment 302-16 (spermophile 337). Captured in the spring of
1916. Weight, 120 grams. Operated on May 8, 1916. Both ovaries
were removed. ;

Gross observations. The animal is pregnant, the placental swellings
being 6 mm. in diameter.

Sacrificed May 18, 1916. Weight, 142 grams.

Gross observations. The uterus is in a very unnatural condition.
It appears dark red in color all over. (The color is not due to conges-
tion as can be seen on microscopic section. Instead, it must be due to
the presence of old clotted blood in the lumen.) There is one hard
swelling still palpable. The pregnancy was interrupted and the
placentas with the fetuses have been undergoing degenerative changes.

Experiment 303-16 (spermophile 338). Captured April 23, 1916.
Weight, 221.5 grams. Operated on May 8, 1916. Both ovaries were
removed.

Gross observations. The animal is pregnant, the placental swellings
being of good size.

Sacrificed May 19, 1916. Weight, 140 grams.

Gross observations. The uterus is very dark-red colored, dead-
looking and contains several dark-colored swellings. The blood supply
to the uterus is intact as tested by Dr. Mann. It appears as though
degeneration of the placentas with the fetuses has been going on.

Experiments 305-16 (spermophile 340). Captured May 6, 1916.
Weight, 124 grams. Operated on May 8, 1916. Both ovaries were
removed.

Gross observations. The animal is in an advanced condition of
pregnancy, the placental swellings measuring 2.2 cm. in length and
1.3 em. in breadth.

Died May 11, 1916.

Gross observations. The cause of death could not be determined.
The uterus is very unhealthy appearing, dark red in color. The
remaining placental swellings are of various sizes. They are very dark
red and are hard. Apparently degeneration changes have been going
on.

Experiment 307-16 (spermophile 342). Captured May 6, 1916.
Weight, 142 grams. Operated on May 8, 1916. Both ovaries were
removed.

Gross observations. The animal.is pregnant, the placental swell-
ings being 6.5 mm. in length. There are nine swellings in the left horn
and three in the right.

Sacrificed May 20,1916. Weight, 152 grams.

Gross observations. The uterus is very dark in color. There are
some adhesions on the left end. In the left horn are four swellings,
the first and fourth measuring 4 mm. in length and breadth, the second
4mm. by 5 mm., and the third, 2 mm. x 2 mm. In the right horn
are three swellings—the first measuring 4 mm. x 4 mm., the second

L52Z DELLA DRIPS

5mm.x5mm., and the third, 3mm.x2mm. Each swelling con-
sists of a light-colored band surrounding a hard dark red mass. The
same condition was found in all the other cases. Some of the pla-
cental swellings with the fetuses have disappeared entirely, the others
are degenerating.

Experiment 333-16 (spermophile 362). Captured April 28, 1916.
Weight, 168 grams. Operated on May 12, 1916. Both ovaries were
removed. ,

Gross observations. The animal is in advanced pregnancy. The
operation was difficult as there were so many large fetuses. No at-
tempt was made to count the number for fear of disturbing them.

Sacrificed May 20, 1916. Weight, 125 grams.

Gross observations. The uterus appears about normal in color.
There are three swellings in the left horn, 4mm. by 5 mm., and ten in
the right horn, same size, dark red and hard.

Experiment 355-16 (spermophile 375). Captured May 20, 1916.
Weight, 143 grams. Operated on May 22, 1916. . Both ovaries were
removed.

Gross observations. The animal is in advanced pregnancy. The
operation was performed with very little trauma, the fetuses not being
removed from the abdominal cavity. The placental swellings were
1.5 em. or more in length.

Sacrificed May 26, 1916. Weight, 130 grams.

Gross observations. The animal had aborted some time previously,
perhaps twenty-four hours. There are seven placental sites in the
right horn and two in the left, all 5 mm. x 6 mm. The uterus is
normal in color. The placental swellings appear congested but not
hard. They look very different from those previously described.

Experiment 357-16 (spermophile 377) Captured May 18, 1916.
Weight, 116 grams. Operated on May 22, 1916. Both ovaries were
removed.

Gross observations. The animal is in advanced pregnancy, the
placental swellings measuring 2 cm. in length. The operation was
performed with very little trauma, the fetuses not being removed from
the abdominal cavity.

Sacrificed May 25, 1916. Weight, 110 grams.

Gross observations. The animal must have aborted shortly after
the operation. There are three placental sites in the right horn and
four in the left, 5 mm. by 4 mm.

Experiment 309-16 (spermophile 344). Captured May 6, 1916.
Weight, 120 grams. Operated on May 8, 1916. Instead of removing
both ovaries, the left ovary was removed and the right tube ligated on
the opposite side.

Gross observations. The animal is pregnant, the placental swellings
measuring 1.4 cm. in length. There are many of the latter in the left
horn and only one in the right.

Sacrificed May 20, 1916. Weight, 120 grams.

THE OVARY OF THE SPERMOPHILE 153

Gross observations. The uterus appears healthy, a normal pale
pink color, but it is very much swollen (table). There are three small
palpable swellings i in the left horn LS only from the increased
size of the uterus at these points. Placental sites in the right horn
not definitely established. Evidently the animal had aborted. The
cause of the swollen condition of the uterus is not evident.

Microscopic observations of the ovaries. Left ovary: fixative,
formalin zenker; stain, hematoxylin and eosin. Rfght ovary: fixative,
formalin zenker; stain, hematoxylin and eosin. The left ovary re-
moved at operation May 8, 1916, is normal. It contains eight corpora
lutea, 0.8 mm. x 0.9 mm. in size. The right ovary obtained at au-
topsy May 20, 1916, is very interesting. It contains only one corpus
luteum, 0.4 mm. x 0.5 mm. in size. This is in the final stage of its
life cycle, the degenerative stage described in the first part of this
work. The rest of the ovary is normal. Through its cortex are eight
large follicles, apparently mature, not showing any atretic changes.
One of these measures 0.8 mm. x 0.6 mm. There are many smaller
growing follicles also, and quite a number of primordial ova. Almost
in the middle of the ovary, just inside the cortex, is the one degenerated
corpus luteum. It is full of large fatty degeneratiye vacuoles and
enlarged blood-vessels and capillaries. The connective tissue has made
great inroads. What made this corpus luteum degenerate is not evi-
dent. It will be shown later that after the removal of one ovary the
other showsno pathologic effects. Andevenafter removal of one ovary
and a uterus containing fetuses, the other ovary with its corpora lutea
shows no pathology. Evidently the ligation of the tube caused the
corpus luteum in the right ovary to degenerate and abortion occurred,
the same as on removal of both ovaries. This would make it appear as
if the corpus luteum was the part of the ovary necessary to the develop-
ment of the fetuses.

Discussion and summary of results. Ten animals were operated
on to get results from the removal of both ovaries containing
corpora lutea on the uterus containing fetuses. One of the
animals died from too much ether. The others lived, but none
of them came to term. Those operated on early in pregnancy
did not abort. The placentas and fetuses degenerated. Those
operated on when the pregnancy was well advanced, aborted.
That this was not the effect of operative trauma will be shown by
another series of experiments on the removal of one ovary in
pregnant spermophiles.

154 DELLA DRIPS

Series 5. Effects of the removal of one ovary on the one remaining
in non-pregnant spermophiles

Experiment 219-15 (spermophile 143). Captured in the spring of
1915. Operated on June 3, 1915. The left ovary was removed.

Gross observations. There are spots marking placental sites in the
uterus. ‘

Right ovary removed September 20, 1915.

Microscopic observations of the ovaries. Both ovaries: fixative,
formalin zenker; stain, hematoxylin and eosin. The left ovary contains
several large corpora lutea of the late lipoid stage. The right ovary
contains no corpora lutea. It is a typical ovary of the fall of the year.
There are many medium-sized growing follicles and many primordial
ova. There is no demonstrable pathology. The corpora lutea must
have disappeared normally.

Experiment 221-15 (spermophile 159). Captured in the spring of
1915. Operated on June 29, 1915. The left ovary was removed.

Gross observations. The uterus is normal and inactive.

The right ovary was removed August 5, 1915.

Microscopic observations of the ovaries. Both ovaries; fixative,
formalin zenker; stain, hematoxylin and eosin. The left ovary con-
tains six very large corpora lutea of the lipoid stage. The right ovary
contains two small corpora lutea in the final stage of their life cycle,
evidenced by their size, the large amount of fatty degeneration, the
numerous large blood-vessels, and the great invasion of connective
tissue. This ovary contains many medium-sized, ripening follicles,
also many small follicles and primordial ova. No pathology is evident.
It is a typical ovary of August 5. The results of these two successful
experiments were so evidently negative that it seemed unnecessary to
repeat them the next year.

Series 6. Effects of the removal of one ovary on the other ovary and
the uterus in pregnant spermophiles

These experiments were controls for the experiments under
series 4.

Experiment 352-16 (spermophile 372). Captured May 20, 1916.
Weight, 148 grams. Operated on May 22, 1916. The left ovary was
removed.

Gross observations. The animal is pregnant, the fetuses measuring
1.5 em. in length. There are not many fetuses.

Sacrificed May 25, 1916. Weight, 135 grams.

Gross observations. All the fetuses are alive and no abnormalities
are apparent. There are five fetuses in the right horn and two in the
left. They measure 2.7 cm. by 1.5 cm.

THE OVARY OF THE SPERMOPHILE | 155

Microscopic observations of the ovaries. Right ovary: fixative,
formalin zenker; stains, hematoxylin and eosin, and Weigert’s copper-
chrome-hematoxylin. The left ovary was lost. The right ovary
shows no abnormality. It contains five corpora lutea. They measure
0.8 mm. x 0.9 mm. in diameter and are in the early lipoid stage, the
condition which exists about the time of the birth of the fetuses. They
still contain numbers of red granules, but the lipoid droplets are very
conspicuous in the cells. In fact, some cells are so full of the lipoid
that the red granules are in evidence only at the periphery of the cell.

Experiment 351-16 (spermophile 371). Captured May 20, 1916.
Weight, 135.5 grams. Operated on May 22, 1916. The left ovary
was removed.

Gross observations. The animal is pregnant, the fetuses measuring
2 cm. in length. The animal gave birth to young, May 31, 1916.

Sacrificed June 17, 1916. Weight, 130.2 grams.

Gross observations. The remaining ovary and uterus appear nor-
mal. There are three spots marking the placental site in the right
horn and three in the left.

Microscopie observations of ovaries. Left ovary: fixative, formalin
zenker, stains hematoxylin and eosin, and Weigert’s copper-chrome-
hematoxylin. Right: ovary: fixative, acetic osmic bichromate; stain,
acid fuchsin and methyl green. The left ovary contains three
corpora lutea. These are in the late red-granule stage. The red
granules are still quite numerous, but some cells contain lipoid drop-
lets about their periphery. The right ovary contains three corpora
lutea of the lipoid stage. It appears very similar to other ovaries
containing luteal bodies of this stage. There is no demonstrable
pathology in the ovary.

Experiment 359-16 (spermophile 379). Captured May 20, 1916.
Weight, 154 grams. Operated on May 22, 1916. The left ovary was
removed.

Gross observations. The animal isin advanced pregnancy. There
was more trauma experienced in this operation than in the one per-
formed on spermophile 377, in which both ovaries were removed. The
animal gave birth to young on May 28, 1916, and kept them until
June 8, 1916.

Sacrificed June 24, 1916. Weight, 136.5 grams.

Gross observations. The placental sites in the uterus are scarcely
visible.

Microscopic observations of the ovaries. Both ovaries: fixative,
formalin zenker; stain, hematoxylin and eosin. The left ovary con-
tains six corpora lutea of the same stage as the left ovary of spermo-
phile 371, the late red granule stage. There are a few lipoid droplets
located through the cells. The right ovary contains four corpora
lutea of the lipoid stage. There is nothing abnormal about it.

156. - DELLA DRIPS

Discussion and summary of results. Three spermophiles were
operated on in this series. The results were positive. The
removal of one ovary during the second half of pregnancy does
not affect in any way the normal development of the fetuses.
They go on to term. There is no effect produced on the re-
maining ovary and the development of rts corpora lutea.

rv

Series 7. Effects on the remaining ovary of the removal of one
ovary and the uterus in pregnant animals

Experiment 273-16 (spermophile 308). Captured May 1, 1916.
Operated on May 4, 1916. The right ovary and the uterus were
removed.

Gross observations. Three are four placental swellings in the right
horn of the uterus and three in the left about 4.5 mm. in length. Both
ovaries show hemorrhagic areas which look like tiny hemorrhagic
cysts. They are more noticeable in the right ovary.

Died May 11, 1916.

Gross observations. Absolutely no cause for death is apparent in
the abdomen. The blood-vessels to the remaining ovary are very
much congested.

Microscopic observations of the ovaries. Both ovaries: fixative,
formalin zenker; stains, hematoxylin and eosin, acid fuchsin and
methyl green, and Weigert’s copper-chrome-hematoxylin. The right
ovary contains five corpora lutea of a very early stage, measuring 0.7
mm.x 0.9mm. The left ovary contains four corpora lutea, measur-
ing 0.8mm.x0.5mm. This organ shows no pathology except in the
corpora lutea. These bodies show many degenerative changes. The
cells have lost their red granules and their nuclei are undergoing chro-
matolysis. The connective tissue has increased greatly.

Experiment 290-16 (spermophile. 325). Captured May 4, 1916.
Weight, 111 grams. Operated on May 6, 1916. The right ovary and
uterus were removed.

Gross observations. The animal is pregnant, the fetuses being just
recognizable.

Sacrificed June 8, 1916.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stain, acid fuchsin and methyl green. The right ovary contains five
corpora lutea of a very early stage, about 0.5 mm. x 0.6 mm. in diam-
eter. The luteal cells are filled with the red granules. There is a mass
of blood in the center and practically no organization of connective
tissue or blood-vessels. The left ovary contains six corpora lutea of
the late red-granule stage. They have increased in size to 0.9 mm. x
1mm. The normal development of the corpora lutea has apparently
gone on in this ovary, as it compares very favorably with other ovaries
of this date. There is no demonstrable pathology.

THE OVARY OF THE SPERMOPHILE 157

Experiment 279-16 (spermophile 314). Captured May 1, 1916.
Weight, 100 grams. Operated on May 5, 1916. The left ovary and
the uterus were removed.

Gross observations. It is impossible to recognize any fetuses in the
uterus.

Sacrificed May 18, 1916. Weight, 112 grams.

Gross observations. The right ovary appears to contain corpora
lutea. Everything in the abdomen is in excellent condition. There
are no adhesions.

Microscopic observations of the ovaries. Right ovary: fixative,
formalin zenker; stains, acid fuchsin and methyl green, Weigert’s cop-
per-chrome-hematoxylin, and hematoxylin and eosin. The left ovary
was lost. The right ovary shows no pathology. It contains two
corpora lutea which measure about 0.8 mm. x 0.8 mm. They seem
to have reached their normal development for eighteen days. The
number of red granules might have been slightly less than normal, but
they are much more abundant still than the lipoid droplets. There is
a well-developed connective-tissue framework and network of blood-
vessels throughout the luteal body. As there are only two corpora
lutea in this ovary, several follicles have had room to mature. Two of
these show atretic changes.

Experiment 292-16 (spermophile 327). Captured May 4, 1916.
Weight, 130 grams. Operated on May 6, 1916. The right ovary and
uterus were removed.

Gross observations. There are six swellings in the right horn of the
uterus and four in the left, 8.5 mm. in length.

Sacrificed May 31, 1916. Weight, 170 grams.

Gross observations. The left ovary is found embedded in a mass of
fat, but the blood supply to the ovary seems intact.

Microscopic observations of the ovaries. Right ovary; fixative,
formalin zenker; stains, hematoxylin and eosin, Weigert’s copper-
chrome-hematoxylin. Left ovary: fixative, acetic osmic bichromate;
stain, acid fuchsin and methyl green. The right ovary contains six
corpora lutea, measuring 0.6 mm. x 0.7 mm. They are of the red-
granule stage. There is no lipoid. The mass of blood in the center is
fairly well organized. The connective-tissue framework and network
of capillaries are fairly well developed. The left ovary contains six
corpora lutea which measure 0.7 mm. x 1.1 mm. in diameter. They
have developed very normally, although not quite as rapidly as their
controls. The red granules are still very numerous and there is scarcely
any lipoid present. The connective-tissue framework and the blood-
vessel network are very well developed.

Experiment 295-16 (spermophile 330). Captured May 4, 1916.
Weight, 116.5 grams. Operated’‘on May 6, 1916. The right ovary
and uterus were removed.

Gross observations. There are five swellings in each horn, all about
5.5 mm. in length.

Sacrificed May 26, 1916. Weight, 130 grams.

158 DELLA DRIPS

Microscopic observations of the ovaries.. Left ovary: fixative, acetic
osmic bichromate; stain, acid fuchsin and methyl green. The right
ovary was lost. The left ovary contains five corpora lutea, measuring
0.7 mm. x 0.8 mm., and showing many degeneration changes. These
are most apparent in the corpora lutea and the clumps of interstitial
cells of the stroma. The latter cells appear swollen and their nuclei
are undergoing chromatolysis. The cells of the corpora lutea are most
degenerated in the center of the structure. Some cells seem to have
disappeared entirely here, leaving holes. Many are undergoing fatty
changes. Toward the outside of the structure there are cells which are
still in fairly good condition, still containing red granules, though their
nuclei have not taken the stains well, have indented margins and
clumped chromatin strands.

Experiment 308-16 (spermophile 343). Captured May 6, 1916.
Weight, 110 grams. Operated on May 8, 1916. The ovary and the
uterus were removed.

Gross observations: The animal is pregnant. There are eight pla-
cental swellings in the left horn and one in the right, all 6.5 mm. in
length.

Sacrificed May 31, 1916. Weight, 153 grams.

Microscopic observations of the ovaries. Right ovary: fixative,
formalin zenker; stains, hematoxylin and eosin, Weigert’s copper-
chrome-hematoxylin, Mallory’s connective-tissue stain. Left ovary:
fixative, formalin zenker; stains, hematoxylin and eosin, Weigert’s
copper-chrome-hematoxylin. The right ovary contains one corpus
luteum, measuring 0.8 mm. x 0.8 mm., and is of the red-granule stage.
The left ovary shows no apparent pathology. The corpora lutea meas-
ure 1 cm. x 0.9 mm. and are of the early lipoid stage. They have
developed normally.

Discussion and summary of results. In four of these animals
the remaining ovary with its corpora lutea was not at all affected
by the operation. In two of the spermophiles operated on, the
second ovary showed changes. In spermophile 330 the whole
ovary was affected, and in spermophile 308, only the corpora
lutea. There was evidently some injury to the blood supply.

Series 8. Studies on the producion of the-normal bursting of follicles
and the formation of corpora lutea

Experiment 311-16 (spermophile 346). Captured in the spring of
1916. Weight 88 grams. This animal appeared to be in rut on May
7, 1916. She was placed with a male at 10:10 a.m. on this date and
removed at 3 p.m. on May 8, 1916.

Sacrificed May 8, 1916.

~

THE OVARY OF THE SPERMOPHILE 159

Gross observations. The uterus shows no signs of pregnancy.
The ovaries show very small hemorrhagic areas, the smallest noted so
far. In all probability, the animal has ovulated during the time if
was with the male.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stain, hematoxylin and eosin. There are three corpora lutea in the
right ovary and five in the left. They are in a very early stage of
development, appearing just like other very early ones. There were a
few mitotic figures, one very sure proof of their very recent develop-
ment.

Experiment 323-16 (spermophile 357). Captured April 27, 1916.
Weight, 125 grams. This animal was kept separate from the time
of her capture until May 8, 1916. At 3 p.m. on May 8 a male was
placed with her and kept there until 10:45 a.m. on May 10, 1916.

Sacrificed May 10, 1916, 10:45 a.m.

Gross observations. Gross evidences did not show that the animal
had ovulated.

Microscopie observations of the ovaries. Fixative, formalin zenker.
Stain, hematoxylin and eosin. Both ovaries showed beautiful early
corpora lutea, three in the right ovary and four in the left. These
could not have been more than twenty hours old.

Discussion and summary of results. The results obtained in
these two experiments prove that ovulation follows on coitus
during rut.

Series 9. To determine if ovulation is followed by the formation of
corpora lutea when fertilization is prevented by the resection of
1 cm. of the uterus. One ovary was removed, also, to determine
absolutely whether the animal was pregnant or not.”

Experiment 280-16 (spermophile 315). Captured May 3, 1916.
Weight, 100 grams. Operated on May 5, 1916. The right ovary
was removed and 1 cm. of the uterus just above the body was ligated
and resected.

Gross observations. The animal is not pregnant. She was placed
with a male on May 20, 1916, and remained with him until sacrificed.

Sacrificed June 5, 1916. Weight, 120 grams.

Gross observations. The remaining ovary appears to contain
corpora lutea. There is a good deal of pus about the ligature on the
uterus, with adhesions to the intestines.

Microscopic observations of the ovaries. Fixative, formalin zenker.
Stains, hematoxylin and eosin and Weigert’s copper-chrome-hema-
toxylin. There are no corpora lutea in the right ovary. Several

* This series of experiments is at present being continued. No conclusions
can be drawn from the results obtained so far.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 2

160

large mature follicles are present.
in the left ovary.
mm. x 0.7 mm. or 0.7 mm. x 0.7 mm.

DELLA DRIPS

There are six corpora lutea present
They are of a very early stage, and measure 0.6

The cells are full of granules.
There is a core of blood in the center which is undergoing organization.
It certainly appears as if ovulation had taken place shortly before the
animal was sacrificed.

Comparative sizes of uteri

LENGTH | LENGTH SIZE OF WIDTH WIDTH

EXPERI- | SPERMO- DATE OF OF BODY SIZE OF OF OF
MENT PHILE RIGHT LEFT OF CERVIX RIGHT LEFT
HORN HORN UTERUS HORN HORN

cm. cm. mm. mm. mm. mm.

20(-15 | 139 | April 22, 1916 AS 4.5. | 92a |) 2206") ama eaten
42’-15 | 232 | May 8, 1916 5.6 | 5.6 | 4.0 | 1.8 4) O:75)\euee
42°-15 | 233 | May 9, 1916 7.0 | 6.5. | 8.0 9) 430) |) e0/a ieee
437-15 | 242 | May 9, 1916 6.3: | 6.3.) BION e250) soca anon
246-16 | 291 | April 25, 1916 12.4 || 10:31, 2895 1 Sy 0.) . Aa Oana
2°€-16 | 298 | April 29, 1916 7.2 VO 704) 900. |) sO. |) oe0nr aoe
354-16 | 374 | October 20,1916| 6.8 | 6.0 | 35 | 2.0 | 1.0) a0
35-16 | 378 | October 30, 1916 | 6.3 | 6.3 | 45 | 26 | 1.0 eee
657-16 | 462 | October 30,1916] 6.7 | 6.5 | 5.5 | 2.0 | 1.5 | 1.5
30-16 | 344 | May 20, 1916 6.8.9) 7.2 WeSs0). 18:0) 7) eos ae

SUMMARY AND DISCUSSION

The results obtained from the histologic and experimental
investigations carried on may be summarized as follows:

1. In the spermophiles, ovulation occurs only once a year,
during the rutting season in the early spring. Ovulation is
dependent on the stimulus of coitus, for no corpora lutea were
found in the ovaries of animals which were kept from the males.

2. The corpora lutea cannot be responsible for the phenomena
of rut for they are not present in the ovaries at this time.

3. The corpora lutea develop and pass through their normal
cycle after ovulation whether fertilization follows or not (foot-
note 2, p. 159).

4. While the corpora lutea are present in the ovaries, especially
during the two months following parturition, the process of
developing and ripening the follicles is at a standstill.

5. If the uterus is removed after conception, the corpora lutea
do not begin to degenerate, but pass through their normal cycle.
No effects are noted in the ovaries.

THE OVARY OF THE SPERMOPHILE 161

6. Removal of the uterus at any time does not produce
noticeable effects on the ovaries even after a year’s time. ;

7. Double ovariectomy performed at any time during the
period of pregnancy interrupts gestation. If the operation is
performed after a little more than the first half of pregnancy,
the placentas with the fetuses simply degenerate. The invo-
lution of the uterus in these cases is very abnormal. If both
ovaries are removed late in pregnancy, the animal aborts and
the uterus undergoes a much more normal involution. The
removal of only one ovary does not interrupt the pregnancy.

8. Double ovariectomy at any time prevents the recurrence
of the cyclic changes in the uterus and produces an atrophy of
the organ scarcely noticeable within a year.

9. The corpora lutea apparently do not influence the develop-
ment of the mammary gland. When the uterus was removed
very soon after conception, before any signs of pregnancy could
be noted grossly in the uterus, and before any development of
the mammary glands could be noted grossly, the developing
corpora lutea in the ovaries produced no development in the
mammary glands. This would seem to substantiate the work
of Lane-Claypon and Starling who attribute to the fertilized egg
the stimulus for the development of the mammary glands.

10. The corpus luteum of the spermophile derives its elements
from the follicle just as Sobotta states occurs in the mouse. The
luteal cells are the transformed granulosa cells of the follicle.
The connective tissue and vascular network are derived from
the cells of the internal theca which spends itself entirely in
their formation. The capsule of connective tissue surrounding
the luteal structure is the same external theca which surrounded
the follicle. ‘The microscopic pictures of the corpora lutea in
the succeeding stages of their development correspond also to
Sobotta’s descriptions. It may be well to emphasize the com-
plexity of the vascular network throughout the luteal structure
which, when it is completed, brings every luteal cell in intimate
contact with the blood stream.

11. The life cycle of the corpus luteum is made up of three
distinct phases. First, the phase characterized by the presence

162 DELLA DRIPS

of great numbers of red granules in the protoplasm of the luteal
cells. This phase embraces a period dating from the bursting
of the follicle and covering the entire period of pregnancy. From
a point of time very shortly after the bursting, the protoplasm
of the luteal cells shows these red granules which become more
and more abundant until they seem to reach a crisis of abundance
when the organization of the luteal structure is about perfected,
which is not until the placental swellings have reached a length
of about 8.5 cm. From this time on the granules seem very
gradually to decrease in number in the cells until parturition,
when there is a sudden considerable reduction in their number.
Some are found in the cells, however, even as late as the fourth
week after parturition. Second, the phase characterized by the
presence of many lipoid droplets in the protoplasm of the luteal
cells. This phase begins sometime before parturition and lasts
for about six weeks afterward. About the fourteenth day of
pregnancy when the placental swellings in the uterus measure
1.5 em. to 2 em. in length, the lipoid droplets usually begin to
make their appearance at the periphery of the luteal cells next
to the capsule of the corpus luteum. ‘They increase in number
until at the time of parturition, they are quite noticeable in the
cells, being found scattered all through the protoplasm among
the red granules. After parturition, there seems to be a more
rapid increase in the number of lipoid droplets, which coincides
with the sudden decrease in the number of red granules pre-
viously noted. With this increase in lipoid content the cells
which, from the beginning, have been growing constantly larger,
seem to begin to hypertrophy more rapidly. The luteal cells
are largest and contain the greatest amount of lipoid about six
weeks after parturition. In two more weeks practically all the
lipoid has disappeared from the cells and they are beginning
to show evidences of degeneration. Third, the phase of regres-
sion. This period begins about eight weeks after parturition and
lasts for four weeks. By the last of August the corpora lutea
have disappeared from the ovaries. This phase is characterized
by a fatty degeneration of the luteal cells by an increased
vascularization and a connective-tissue invasion.

THE OVARY OF THE SPERMOPHILE 163

From these results, the following conclusions were drawn as to
the functions of the corpora lutea in the ovaries of spermophiles:

The corpora lutea fix the period of estrus by preventing the
development and the ripening of follicles until the time for the
next rutting season is at hand.

The corpus luteum is a gland with two internal secretions,
both of which have specific effects on the uterus, one bringing
about the changes incident to pregnancy and the other effecting
the normal involution of the organ. The first internal secretion
is represented in the luteal cells during the period of pregnancy
by granules which are very similar in their location and staining
reactions to the granules in the A cells of the islands of Langer-
hans, the glands of internal secretion of the pancreas, described
by Bensley. The granules of the luteal cells, however, are
much larger than those of the A cells, being very easily seen
with high powers of the microscope. No mitochondrial granules
or filaments could be observed, perhaps because of the abun-
dance of the granulations in the protoplasm. These luteal cell
eranules are very much like other secretion granules described
by various writers as occurring in the secreting serous cells of
several glands of the body.

The majority of writers have agreed that there is no fatty
product demonstrable in the corpus luteum of several species of
animals and man in the very early stages. They all seem to
have been of the same opinion that the activity of the ovarian
gland of internal secretion begins with the appearance of the
lipoid droplets in the cells. These lipoid droplets were con-
sidered by them to be the evidence of the secretory activity of
the corpus luteum. Its period of activity would then begin
when these droplets begin to appear in the cells, which time
varies with different species, but in all seems to be about the
time of the fixation of the blastocyst. This activity lasts, they
consider, for varying periods in different species. In the rabbit,
Cohn, Fraenkel, and Niskoubina consider that it lasts for nine
or ten days, when regression sets in about the fifteenth day.
Van der Stricht says that in the bat the lipoid droplets are in
much greater abundance during the second half of the period

164 DELLA DRIPS

of pregnancy and that regressive changes do not begin until the
period of pregnancy is over. Miller says there is no neutral fat
in the human corpus luteum until regression sets in at birth.
Because the first-mentioned group of men found that double
ovariectomy did not cause abortion in rabbits after the fifteenth
day, and did so earlier than this, they considered this lipoid
secretion related to changes in the uterus occurring between the
fourth and fifteenth days after coitus, or between the time of
the fixation of the blastocyst and the middle of the period of
pregnancy.

Van der Stricht seems to have been the first to conceive of the
presence of a secretion in the luteal cells prior to the appearance
of the lipoid droplets which coexists with them for some time
after their appearance. He judges of the presence of this
secretion in the cells from its presence in the near-by inter-
cellular spaces and lymphatics. The latter, according to van
der Stricht, are the avenues of excretion of both the serous and
the lipoid secretions.

In spermophiles, the lipoid product does not begin to appear
in the luteal cells until the period of pregnancy is half over and
is not very abundant until after birth. As far as these animals
are concerned, then, the lipoid product is not the active sub-
stance of the corpus luteum which has specific effects on the
uterus during pregnancy. This active substance is rather a
secretion represented in the cells by the secretory granules
previously noted, which are of a very different nature from the
lipoid droplets.

The second internal secretion which is Rea eee in the
luteal cells by lipoid droplets and which formerly has been
considered the secretion which is responsible for the changes
occurring in the uterus incident to pregnancy, must be con-
sidered, as far as the spermophiles are concerned at least, as
having another function. There seems to be some relationship
in these animals between the period of greatest abundance of
the lipoid product in the cells and the period of regression and
atrophy in the uterus. The uterus of the spermophile atrophies
very slowly, much more so than in animals that bear several

THE OVARY OF THE SPERMOPHILE 165

litters of young every year. The atrophy is not completed until
six or seven weeks after parturition, about the time when the
lipoid product reaches its crisis of abundance in the cells and
begins to disappear. Another result which substantiates the
theory that the lipoid secretion brings about the normal invo-
lution of the uterus in the very abnormal, even pathologic process
which goes on in the uterus following the removal of both ovaries
during all but the more advanced stages of pregnancy. During
the first half of the period of pregnancy there is no lipoid in the
corpora lutea, which, according to this theory, would account
for the pathology in the uterus following double ovariectomy.
If the ovariectomy is performed late in pregnancy after the
lipoid droplets have become quite abundant in the luteal cells,
the animal aborts and the uterus undergoes an involution more
nearly like the normal, due to the specific effect of the lipoid
secretion which is already present in the circulation. Mulon
thought the lipoid of the corpus luteum had an antitoxic action
toward the poisons elaborated in the development of the fetuses.
It would seem more reasonable to suppose that it neutralizes
the toxic products produced in normal involution, which would
be only a part of its function as a specific agent in effecting
this normal involution of the uterus.

It may be added, in closing, that the two luteal secretions are
undoubtedly emptied into the blood stream in these animals.
An observation of the elaborate capillary network of these
structures could lead to no other conclusion. Lymphatic sinuses
are demonstrable in the corpora lutea, but they are found only
near the capsule in the proximity of the larger blood-vessels.
There is no anatomical evidence for concluding that the secre-
tions are carried away by the lymphatics.

The writer wishes to express her appreciation for the valuable
aid given her by Dr. Frank C. Mann, head of the Division of
Experimental Surgery and Pathology of the Mayo Clinic, under
whose direction this work was done.

166 DELLA DRIPS

LITERATURE CITED

Atquier, L., AND THEUVENY, L. 1909 Etat de l’ovaire de chiennes ayant
subi l’extirpation partielle ou totale de l’appareil thyro-parathy-
roidien. Compt. rend. Soc. de biol., T. 66, pp. 217-219. .

ANCEL, P., AND Bourn, P. 1909 Sur la fonction du corps jaune; action du
corps jaune vrai sur l’uterus. Compt. rend. Soc. de biol., T. 66,
pp. 505-507.
1908 Sur le follicule de De Graaf mur et la formation du corps jaune
chez la chienne. Compt. rend. Soc. de biol., T. 65, pp. 314-316.
1909 Sur les homologies et la signification des glandes a secretion
interne de l’ovaire. Compt. rend. Soc. de biol., T. 67, pp. 497, 464-466.

Beacu, R. 1916 The management of ovarian tumors complicating pregnancy,
labor and the puerperium. Am. Jour. Obst., vol. 73, pp. 1029-1041.

Beit, W. B. 1916 The sex-complex. New-York, Wood, 1916.

BensutEy, R. R. 1911-12 Studies on the pancreas of the guinea-pig. Am.
Jour. Anat., vol. 12, pp. 297-388.
1916 The normal mode of secretion in the thyroid gland. Am. Jour.
Anat., vol. 19, pp. 37-57.

Bonp, C. J. 1906 An inquiry into some points in uterine and ovarian phys-
iology and pathology. Brit. Med. Jour., vol. 2, pp. 121-128.

Burnam, C. F. 1912 Corpus luteum extract, with suggestions as to its use in
gynecologic practice. Jour. Am. Med. Assn., vol. 59, pp. 698-703.

CaRMICHEL, E. 8., anp Marsuatu, F. H. 1907 The correlation of the ovarian
and uterine functions. Proc. Roy. Soc., Lond., Series B, vol. 79, pp.
387-395. 5

CHALFANT, S. A. 1915 Subcutaneous transplantation of ovarian tissue, report
of thirty-two cases with special reference to its effects on the meno-
pause. Surg., Gynec., and Obst., vol. 21, pp. 579-590.

CuHApPpELLIER, A. 1909 Follicles pluriovulaires et degenerescence ovulaire
chez la souris blanche. Compt. rend. Soc. de biol., T. 66, pp. 543-545.

CHAUFFARD, A., LARocHE, G., AND Gricaut, ‘A. 1912 Function cholesterini-
genique du corps jaune. Compt. rend. Soc. de biol., T. 72, pp. 223-
225, 265-267.

Cuark, J.G. 1898-99 The origin, growth and fate of the corpus luteum. John
Hopkins Hospital Reports, vol. 7, pp. 181-221.

Coun, F. 1903 Zur Histologie und Histogenese des Corpus luteum und des
interstitiellen Ovarialgewebes. Arch.f. mikr. Anat., Bd. 62, 8, 745-
Tie.

Corner, G. W. 1915 The corpus luteum of pregnancy in swine. Carnegie
Institution, Washington, Publication no. 222, pp. 69-94.

CuLBRETSON, C. 1916 A study of the menopause. Surg., Gynec, and Obst.,
vol. 23, pp. 667-685.

Darts, F. 1908 On the relations between the ovaries and the uterus. Surg.,
Gynec. and Obst., vol. 6, pp. 153-160.

DannreEvTHER, W. T. 1914 Corpus luteum organotherapy in clinical practice.
Jour. Am. Med. Assn., vol. 62, pp. 359-362.

THE OVARY OF THE SPERMOPHILE 167

De Lez, J.B. 1916 Auto-transplantation of the corpusluteum. Surg., Gynec.

and Obst., vol. 22, pp. 228-231.
FRAENKEL, L. 1903 Die Function des Corpus luteum. Arch. f. Gynak., Bd.
68, S. 438-545.

Frank, R. T., Aanp RosenBiLooM, J. 1915 Physiologically active substances
contained in the placenta and in the corpus luteum. Surg., Gynec.
and Obst., vol. 21, pp. 646-649.

Graves, W. P. 1913 Influence of the ovary as an organ of internal secretion.
Am. Jour. Obst., vol. 67, pp. 649-665, 779-785.

Hare, H. A. 1912 Therapeutic application of ductless glands. Am. Jour.
Obst., vol. 64, pp. 514-518.

Harrower, H.R. 1916 Relation of the internal secretions to neurasthenia in
women. Am. Jour. Obst., vol. 73, pp. 630-637.

Hearst, W. 1906 The source of the stimulus which causes the development of
the mammary gland and the secretion of milk. Am. Jour. Physiol.,
vol. 34, pp. 499-500.

Hirst, J.C. 1916 The control of disagreeable symptoms of the surgical meno-
pause by the hypodermic intramuscular administration of corpus
luteum extract. Am. Jour. Obst., vol. 73, pp. 648-650.

Iscovesco, H. 1912 Les lipoides de l’ovaire. Compt. rend. de la Soc. biol.,
T. 73, pp. 16-18, 189-191.

JANKowskKI, J. 1904 Beitrag zur Entstehung des Corpus luteum der Siuge-
tiere. Arch. f. mikr. Anat., Bd. 64, S. 361-389.

Lane-Criaypon, J. E., AND StaruinG, E. H. 1906 An experimental inquiry
into the factors which determine the growth and activity of the mam-
mary glands. Proc. Roy. Soc. Lond., Series B, vol. 77, pp. 505-523.

Lane-Crayron, J. E. 1906 On the origin and life history of the interstitial
cells of the ovary in the rabbit. Proc. Roy. Soc. Lond., Series B,
vol. 77, pp. 32-58.

LetcutTon, A. P. 1915 The use of corpus luteum extract in the treatment of
menopause disorders. Am. Jour. Obst., vol. 72, pp. 878-888.

Lewis, M. R., ann WarreN, H. 1914-15 Mitochondria in tissue cultures.
Am. Jour. Anat., vol. 17, pp. 339-402.

Livon, J., Jr. 1909 Sur l’action des extracts du corps jaune de l’ovaire.
Compt. rend. de la Soe. biol., T. 66, p. 549.

Logs, L. 1906 The formation of the corpus luteum in the guinea-pig. Jour.
Am. Med. Assn., vol. 46, pp. 416-423.

MarsHatu, F.H. A. 1910 The physiology of reproduction. Lond., Longmans.

MarsHat., F. H., anp Runciman, J.G. 1915 On the ovarian factor concerned
in the recurrence of the oestrous cycle. Jour. Physiol., vol. 49, pp..
17-22.

Martin, F. H. 1915 Ovarian transplantation. Surg., Gynec. and Obst.,
vol. 20, pp. 568-579.

Meyer, R. 1911 Uber Corpus-luteum-Bildung beim Menschen. Arch. f.
Gynik., Bd. 98, S. 354-404.
1913 Uber die Beziehung der Eizelle und des befruchteten Eies zum
Follikelapparat, sowie des Corpus luteum zur Menstruation. Arch.
f. Gynaik., Bd. 100, S. 1-19.

168 DELLA DRIPS

Mruter, J. W. 1900 Die Riickbildung des Corpus luteum. Arch. f. Gynak.,
Bd. 61, S. 263-286.
1914 Corpus luteum, Menstruation und Graviditaét. Arch. f. Gynak.,
Bd. 101, 8. 568-619.

Mouton, P. 1910 Sur l’existence de graisses antitoxiques. Compt. rend. Soc.
biol., T. 69, pp. 389-391.

Nisxousina, N. 1908 Sur la structure du corps jaune pendant et apres la
gestation. Compt. rend. Soc. biol., T. 65, pp. 767-769.

Paruon, C., Dumrrresco, G., AND Niserpgsco, C. 1909 Note sur les lipoides
des ovaires. Compt. rend. Soc. biol., T. 66, pp. 650-652.

PEart, R., AND Surrace, F. 1914 On the effect of corpus luteum substance
‘upon ovulation in the fowl. Jour. Biol. Chem., vol. 19, pp. 263-274.

Reaaup, C., aNp DusreviL, C. Sur. les relations fonctionnells des corps jaunes
avec. l’uterus non gravide:
1909 1. Etat de la question et methodes de recherches. Compt.
rend. Soc. de biol., T. 66, pp. 257-259.
1909 2. Statistiques de variations de volume de |’uterus par rapport
a Vetat des ovaires. Compt. rend. Soc. de biol., T. 66, pp. 299-301.
1909 3. Etats successifs de l’uterus chez le meme sujet aux diverses
phases de la periode pregravidique. Compt. rend. Soc. de biol.,
T. 66, pp. 413-415. ji
1909 Sur les follicules ovariens hemorrhagiques et sur le mecanisme
de la dehiscense des follicles. Compt. rend Soc. de biol., T. 66, pp.
828-830.
1908 Action du male sur le rut et l’ovulation chez la lapine. 1. Le
voisinage prolonge, sans accouplement, est insuffsant pour provoquer
Vovulation. Compt. rend. Soc. de biol., T. 65, pp. 501-503.

Recavup, C., ann Pouicarp, A. 1901 Fonction glandulaire de l’epithelium
ovarique et de ses diverticules tubuliformes chez la chienne. Compt.
rend. Soc. de biol., T. 53, pp. 615-616.

Soporra, J. 1896 Uber die Bildung des Corpus luteum dei der Maus. Arch. f.
mikr. Anat., Bd. 47, S. 261-309.
1897 Uber die Bildung des Corpus luteum beim Kaninchen nebst
einigen Bemerkungen tiber den sprungreifen Follikel und die Rich-
tungsspindeln des Kaninchens. Anat. Hefte, Bd. 8, 8. 469-521.

Sunpwatu, J. 1916 The lachrymal gland. Am. Jour. Anat., vol. 20, pp. 137-
247.

Turrier, T. 1915 Transplantation of ovaries. Surg., Gynec. and Obst.,
vol. 20, pp. 30-34.

VAN DER Srricut, O. 1912 Sur le processus de l’excretion des glandes endo-
crines; le corps jaune et la glande interstitielle de l’ovaire. Archiv.
Biol., T. 27, pp. 585-722.

Vincent, S 1912 Internal secretion and the ductless glands. Lond., Arnold.

VINEBERG, H. 1915 Fate of ovaries after hysterectomy. Am. Jour. Obst.
vol. 22, pp. 144-147.

‘Or X ‘SolmvAOo Y}OqG
jO [BAOWOL oY} JoJJe sndojn B JO UIOY 9Y} Ysnoryy UOTZIOS-SsOIQ) “UTSOO pur
ulyAxoyeUoYy “UTeyS fayuez UT[VULIO, ‘OATPEXTY “(GT-O0Z JUOWTIOdx9) F “SI

‘OF X = ‘SolIvAO YJOq Jo
[eAowol oy} JoJje snsoyn v Jo Apoqg oy} YSno1y} UWOTZDOS-SssOIQ, “UISOD pUv
ul]AXOyVUOYy ‘UIBIS f1oyUoZ UT[VUIIOJS ‘BATZVXIY (GT-00Z JUoWTIedxe) Eg “SI
ZI X “4nd JO sndoyn [VuOU B JO UIOY oY} YsSno1y} UOTZIOS-SSOIQ, “UISOa pUv
uljAxoyemMoy ‘UIeys fdoyuoz UI]BVUIIOJ ‘AATYVXTY “(QT-SFZ JUoUWTIedxe) Z “SI
‘SIX «(yma Jo sniojn jeursou & Jo Apog oy} Ysnory} UOTJDES-Sso1g, “UISOD puUv
ul[AXOJVUIOY ‘UTBYS {1oyUeZ UI[VULIOJ ‘OATYVXTY *(9T-SFZ JUoUTTIOdx9) J “SI

170

171

‘0g & ‘Sutids Aj1vo JO AIVAO [VULIOU B JO UOTJOOS-SSOIZ) “UISOD pu
UT[|AXO}BUOY ‘ULEIS f1oYUOZ UI[VUIIOJ ‘OATZVXTY *(OT-G1Z JUOUTIOdxe) [][ ‘SI
‘>% X “AOUIUINS 909¥B] OY} JO AIVAO UB JO UOTJDOS-SSOIQ, “N’[GIOSSBA
pue ulIsvaq ‘UTeyS {1oyUeZ UITBUIAOJ ‘OATJVXTIY *(GT-9FPE JUOUTIOdX9) OT “Sly
‘OP X = “SI6I ‘ZZ Toquiey
-dog JO Sniojn oATJOVUT [VUILOU B JO ULOY VY} YSNOIY} UOTIOS-SsoIg, “UIsOD pUuB
ul[Axo}BMOY ‘UTBIS f1oyUOZ UT[BUIIOJ ‘OATYVXIY *(G[-OPP JUOUWTIOdxXe) 9 “ST
‘OP X = “ET6T ‘ZZ toquioydag
JO Sndojn oATJOVUT [VUTIOU B JO ApOg oY} Ysnory} UOTJDES-SSOIZ) “UISOD PUB
UI[AXOJVUWIOY ‘UTBIS ‘toyUAZ UT[VUTIOJ “OATJVXTIY “(C[-GpPpY JUEUITIEdxX9) G “STi

172

ae

“<u
os

173

“AGLOJDIIIVAO o[qnop JoJfe sy~eurtue Jo Tlojn ole

Q pu’ q = “[o1}uU090 yeur1ou & fo Snojn of} ST V

‘(9I-7GE JUOUITIOdxo) § “SI

“S]BUTUG ALOVVIOGV] OM YOY “Jud fo snaoJN [eurlou

B YJIM SOLIBAO TOC JO [BAOWOL OY} to} J SILIOJN V

BYJIM SOlIGAO YO JO [BAOU OL OY} 10} JB SN.OIN V

'(G[-GeF JUOWTIOdXO) 8 “SITY
“4nd JO snaeyn [BuLOU
‘(G[-OFF JUOUITIOdXO) J “STyT

174

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 2

‘OT & “uoTyLIMyed
Joye SsyoomM JySIo ynoqe AIVAO UB JO UOTPOS-SSOT,) ‘usot3d [AYJOU puB UISyony
PIO’ ‘UIGIS So}BUIOAYOIG OTWSO YT}oB ‘QATIEXTY *(QI-FFP PUe witodxo) GT Sty
OZ X "poztuvS10 AjoxoTdu09 9 LOdeq VAVY
Boy] VLOAIOD OY} IdJJB Aouvusoad ut Ap{va ATBAO UB JO WOTPDOS-SSOL) “UTSOO PUB
UI[AXOPVUNOY ‘UTILS ‘oxUOZ UT[VULLOJ ‘QAEXTY “(QT-L1Z PUouttodxe) FT rol |
02 -Koueugoid ur Ayavo ATaA AIVAO UB JO WOTPOS-SSOID) “UISOD PUL
G ; : 9) !
ULT[AXOYVIOY ‘ULEYS + toxWoZ UL[BULLOF ‘QATYCXTT ~(91-GLSZ qyuoulLtodxe) ¢] “oly
0Z ‘and Jo ATRAO UB JO UOTJDOS-SSOIg, ‘“UooId [AYJoU pu
G } 5 ‘
UISYONJ plo’ ‘ULL}S ‘1oyUoZ UL[BULLOJ ‘QATJVXT *(QT-9FS JUO uTtodxo) CT ‘Sly

176

: FO X “uOTpLIMgaed

JoVJe SYOoM XIS wo] sndioo9 B JO UOTPOS-ssOI “UdIId [AYJOUL puw UISPONJ
plow ‘ures foyRUOALYoT OTLUSO OTPOOR “OATYRXTT *(OT-ETF yuowtIodXxXo) GT “STL

‘¥9 X

uotanjzaed 4yv umnoyny sndaoo wv Jo UOTJOS-ssot) “Uoold [AYJOU puB UISyoNy
plow ‘UTeYS foyRULOAYOIG oTUISO OTZOOR ‘OATIEXTY *(OT-ENE qyuourttodxo) ST “ST

‘£9 X ‘Uojn] snd1z09 AjAvo ATOA B JO UOTPOS-ssOA “UooIS [AYJoW puB UISyONJ
plow UTEZS foJVUIOIYOTG OLVISO DTJOOV “OATPVXTY *(YT-89T quowuliodxe) JT “STyL

‘0G X

HOTPLUNGAVA IoJJW SYOoM OATOMY JNOge AIVAO UL JO WOTDS-SsOL) “UTSOO PUB
UIA[XOJwUOY ‘UTEIS {AoYUoZ UTPBULIOJ ‘OATPVXTY “(OT-GTG JUouttodxo) 9] “BMT

178

179

$9 X ‘gsKO windy] snda0d B Jo UOTJOS-SsOIQ, ‘“UdoIs [AYJOU puw UISyoNJ
plow ‘ureys foyRUIOAYOIG OTUISO OTJOOR ‘OATYVXTIY *(9T-S9Z JUoUITAIOdX) EZ “SIT
‘OSF X ‘ST[e0 [woyny] AjAvo oy} Suowe Surystsaoed
apPITJOF oy} JO UNsSO|NUBIS UNYVIYS oY} JO S][oo powAossuUBIyUN SUTMOYS WM]
sndio9 yeudouqe UB JO ][VM oy} YSno1y} uoTpo9g “Usold [AYJou puB UTSyONyJ
plow ‘ureys foJVUIOIYOIG OTUISO OTZOOR ‘OATJVXTY *(9[-S9T JUOWTIodxa) ZZ “ST Y
‘$9 X ‘WMoyn] sndio9 [vuUAOUG’ UB JO UOTJoS-SSsOl) “Usods [AYJOU puB UISyoNJ
plow ‘uTeys foyVUIOAYOIG OTUISO OTJOOR ‘OATJVXIY *(Q[T-S9Z JUourTIodxo) [Z “Sly
FO XX ‘uoTpUngavd 10qye
SYooM JYSto ynogR uNoyzN] sndiood vB JO UOTJDES-SSOL{) “Udetsd [AYJOUL puv UTS
plow ‘uTeys fo}PVUIOAYOIG OTUISO DTJOOV ‘OATZVXIY. “(O[-GEP JUoUTIodxa) OZ “Sy

180

181

‘F9 X “ysho OISvYIIOWIOY WNeyN] SNd40d B Jo UOTPeS-SsoIT “UTSOO PUB
urpAO}xvUoYy ‘UTG4S !foyWoZ UTPBULLIOF “OATPEXTT ‘(QT-86z JUoUTIOdx9) 6% “STW

‘$9 X ‘UMAO poUTezal B YIM UMN] SNd10d B JO UOTPS-SSOI- “UTSO9 PUB
uT]AXoOZVUDOY ‘UTeYS foyUoZ UTPVUAOJ “OATPEXTY *“(9T- ELE quoultiadxe) FZ ‘SI

182

"096 X “TOM

-OUIBAY ONSSTJ-OATPOOUUOD POULLO J-[JOM VST oo, “SoPNUBIT pod oy} JO FVYY YFIM

aZIs Hoy} olvduoy ‘s{joo ywoyny oyy Jo Atoydtsod oy} 4B syo[doup prod youlq oyy

OJON “WOTPLANZAed ye wunoyny sndsoo vB jo UoTPIOg “Uoold [AYJOUL puB UTsyony
plow UTEZS foPVUIOAYOIG OTWUSO OTJOOR “OATPVXIY (QT-E9E JUoWTIOdXO) 2% “BIT

‘096 X

YAOMOWIVAF ONSST]-OATJPIOUUOD OY} BULULIOJ o1V S|SV[G-O1qI, “pod SnOouUdsOUIOY, &

savodde uisvdozyoad toy} YVYyy SoNUBAS AUBUT OS YFIA S[[oo [BoyNT OJON (‘osvyS

oynueis-poy) “wmnoyny sndaoo AjAvo AJoA JO UOTPAOG “Woods [AYJOUL puB UTSYyony
plow ‘uIeZS foyVULOAYOTG OTUISO OTZOOV ‘OATPVXIY *(9T-89Z JUOUIIedxo) 9Z “BIT

183

096 X “WOT?

-e19uddop JO VOUSPTAD UB ST YOIYA UIVIS UISYON plo’ oy} Uoye} oAVY TOPONU oy],

‘UOTyRAoUdBop A}ZBF PUL UWOTPEZTIV[NOSVA OJON *(05L4S UWOTSso190.1) uorytangaed

IO}JV SYOOM OATOM} Jog’ Wyn] snd109 v Jo UOT}IOG “Wools [Aqjour pue urlsyony
plow ‘Ureys foyRUIOATOIG OTWUSO dTJOOB “OATPEXTY *(OT-CTG quoulltodx9) 6% ‘STi

‘096 X ‘sjo[doap prodiy, jo souvpunqe

oy} puv ‘Toponu Suryoor-oaTzow oy} ‘S{[oo oy} JO ozIs oy} OJON *(08BYS ptodry) wor

-1IN}AV 19}ZB SYOoA XIs uNeyN] sndso9 v Jo UOT}IOg “Wools [AYJour puv ursyong
ple ‘ureZs foyBUIOAYOTG oTwUsO TOD" “OATPEXTY *(OI-1EP yuowtdtadxa) 8% “ST

184

* 7 ,
aay

Resumido por el autor, Royal Norton Chapman.

Estudio sobre la correlacién entre la estructura de la pelvis y las
costumbres de algunos mamiferos cavadores.

En el presente trabajo se describen y estudian las pelvis de los
mamfferos cavadores y segtin la estructura de dichos huesos se
clasifican en ‘cavadoras tipicas’ (topos, musaranas, Geomys, y
algunas especies de Microtus) en las cuales los bordes ventrales
de los huesos pubianos son horizontales y divergen posterior-
mente y la sinfisis falta en absoluto 0, si exis te, esta formada por
una barra Osea transversal; las ‘semi-cavadoras’ (muchas espe-
cies de ratones, algunos insectivoros), en las cuales los huesos
pubianos estan inclinados caudo-ventralmente y convergen para
formar una sinfisis en sus puntos de contacto. Las pelvis ‘cava-
doras tipicas’ se han dividido en las de ‘tipo estrecho’ (topos y
musaranas) y las de ‘tipo ancho’ (Geomys, topos marsupiales y
otros). La sinfisis pubiana puede existir o faltar en las hembras
de Geomys pertenecientes a la misma especie. Todas estas pel-
vis son horizontales y coosificadas con la columna vertebral, y de
este modo dirigen la fuerza producida por los miembros posteri-
ores al cavar, a lo largo de una linea recta paralela ala columna
vertebral. La capacidad de la pelvis es, por consiguiente, tan
limitada que es necesaria la desaparicién de la sinfisis pubiana
para permitir el paso del feto durante el parto. La falta de sin-
fisis pubiana esté compensada en el soporte de las visceras por
el cruzamiento de los mtsculos rectos del abdomen—el musculo
derecho se inserta en el hueso pubiano izquierdo y el mtisculo de
este lado en el mismo hueso del lado derecho. Estas estructuras
paralelas que se presentan en los diferentes grupos de animales de
costumbres semejantes, indican la importancia de fuerzas simi-
lares, que actudn de un modo semejante, como factores de la
modificacién de la estructura.

Translation by Dr. José F. Nonidez
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 24

A STUDY OF THE CORRELATION OF THE PELVIC
STRUCTURE AND THE HABITS OF CERTAIN
BURROWING MAMMALS

ROYAL NORTON CHAPMAN

From the Zoological Laboratory, Cornell University

FIVE PLATES (TWENTY-SIX FIGURES)

CONTENTS
Teg NRCG ITA isa ee a EE: be Aa 2 MRS eS en Re 185
MRE STEUCEHTE OL thesPel Vises. seamen ioe + «cele eee se Beles Aargele SS ws 187
i; Lhe pelvie bones: of the Rodentia..25 25... ccus cine st ose site nes = 2 187
Zeeinespelvicsbomes-or theudnsectiwvGnan.. 7.25. 2658s. ceca. eee es 192
3. The pelvic bones of the marsupial mole (Notoryctes)..........--+++++- 194
4. A classification of the pelves of burrowing mammals..............-+-- 194
Therstatus of the rectus, abdominis muscles: ........--+-.+s +2227 s 005s se ss = 195
1. The relations of the rectus abdominis muscles of the Rodentia........ 196
2. The relations of the rectus abdominis muscles of the Insectivora...... 197
The relation of the pelvic estates and the habits.................-+--+2055 198
it Bhe attainment of the horizontal pelvis: 2.2.6. .56.-.0<2 082s ese ees oe 198
Pehiculoss, OF he Symp MySis: PUDIS: a sete. - gas siesew sss aje aoe mes ee oles 199
3. The persistence of the symphysis in certain caseS............+-+++++++ 203
4. The function of crossed rectus abdominis muscles..............++++++5 205
SHTECTECT 15 GSS NN ee SEP RI et Str Wor Eke Aa A RE a a Sr a 206
INTRODUCTION

The results embodied in this paper are from data accumulated
during several years of observation. This line of work first
suggested itself during the examination of specimens in the
field. It was noticed that the pocket gophers (Geomys bur-
sarius) have two types of pelvis, one with a symphysis and the
other without. This led to a more extended investigation of
the pelvis of burrowing mammals, resulting in the revelation
that they possess certain structural characteristics In common.

Huxley (’75), in a consideration of the mammalian pelvis,
employed certain angles to show the progress of its phylogenetic
development. A longitudinal line drawn through the centra of

185

186 ROYAL NORTON CHAPMAN

the sacral vertebrae he termed the sacral axis (fig. 3, Sa.). A
line passing through the junctions of the ischium and the pubis
on both sides of the obturator foramen designates the obturator
axis (Os). Another line, drawn along the ilum through the
middle of the sacral articulation and the center of the acetabu-
lum bears the designation of the iliac axis (Ja). The iliopectin-
eal axis ([/p) is determined by a line traversing the union of
the ilium with the pubis in front and with the ischium behind.
This author concludes that in the phylogeny of the Mammalia
the iliac axis (Ja) has tended to form a more acute angle with
the sacral axis (Sa), that the angle formed between the iliopectin-
eal axis (//p) and the sacral axis (Sa) has approached a right
angle, and that the angle between the obturator axis (Oa) and
the sacral axis has become more and more acute. ‘That is, the
pelvis, which in the Reptilia (fig. 1) was nearly perpendicular to
the vertebral column, has approached a horizontal position.
The symphysis is said to move posteriorly, until it is formed
largely by the ischial bones rather than the pubic bones.

G. E. Dobson (82) incidentally included descriptions of the
estate of the pelvis in his study of the Insectivora. The nature
of the symphysis pubis in the families treated is mentioned, but
no generalizations are given.

W. Lecke (’84, ’92) considered the pelvis and the rectus
abdominis muscles of the Insectivora from both the vhylo-
genetic and ontogenetic points of view. His conclusions will be
referred to later in this paper.

H. L. Osborn (94) concluded that the symphysis proper of
the pocket gophers (Geomys bursarius) was absent and stated
his belief that the bone present in the male, forming a ‘pseudo-
symphysis,’ was an ossification of a tendon and not a part of
the pubic (or ischial) bone.

Broek (14) and others have made studies of the pelvis of
man and the apes in which correlations with the habits of loco-
motion have been found and described. There is a vast amount
of literature dealing with the human pelvis but, since it does
not bear directly upon the subject at hand, it has not been in-
cluded in this discussion.

THE PELVIS OF BURROWING MAMMALS 187

Thanks are due Dr. H. D. Reed, under whose direction this
work was done, for many suggestions and for assistance while
the work was in progress. Acknowledgments are also due the
authorities of the Smithsonian Institution for placing at the
disposal of the author the material in the National Museum,
especially Drs. Leonhard Stejneger and Gerrit S. Miller and
Mr. Ned Holister for kindnesses and assistance while examining
material in the Smithsonian collection.

THE STRUCTURE OF THE PELVIS
1. The pelvic bones of the Rodentia

The two types of pelvis which occur among the pocket gophers
(family Geomydae) are well illustrated by Geomys bursarius
(figs. 4, 5, 6, and 7). It will be noticed that these two pelves
differ not only in the presence or absence of the symphysis, but
in that a triangular portion of bone lying at the sides of the
symphysis in the pelvis of the closed type is absent in that of
the open type. These two types of pelvis occur not only in the
same species, but also in the same sex, as shown by the females
where both types occur. All the males examined possess the
symphysis.

No intermediate conditions have been found; the bones either
meet to form a symphysis or are widely separated. Very young
females have been found possessing a symphysis, while old
females without a symphysis are not uncommon. It seems,
therefore, that the presence or absence of the symphysis cannot
be a matter of the ossification of the bones due to age.

The following table shows the proportion of the open and
closed pelves found in the species of the pocket gophers which
have been examined. Since the symphysis was present in all
of the males examined, only the females have been included in
the table.

188 ROYAL NORTON CHAPMAN

SPECIES : pol raat OPEN TYPE pees
Geomys bUTSATIUS.< 5. 2.8 nares Steet em oe 23 9 14
Geomiys Cuz fae ees eg ake ean a 6 2 +
Geomiys hispidGs:. xyes seen eae cee a oe ee eee 1 0 1
Geomys: (species not known): .2...5..4.te0eee4e- 5 1 4
Thomomiys talpowdesmufescens:!... 2-24 .2e- ase. 2: 3 3 0
Mhomonmiysiboniaeyoouidelecy 5 see ea eee 3 1 2
‘Thomomys' bogiae miericans.«.1.---e ome eee. 1 0 1
Thomonitys ‘bottade pasenlis. 2. aes ace a: oo 3 0 3
Mhomomiyssbottaeyanicae yee ae ee wee 3 1 2
Mhomomiysicolumblanussseee eee eee eee 1 1 0
Thomomys fulvus intermedius.................... 1 0 1
Rhomomiysmmonticolameazamiaeeee ee eee eee 5 0 5
Thomomiys quadratusiquaduatisa. «2s ase ae 1 0 i

These data are based upon specimens in the flesh which were
dissected to determine the condition of the pelvis and the sex
of the specimens.

It is obvious that no conclusions as to the percentage of the
open and closed types of pelvis in the various species can be
drawn without the examination of a much more extended series
of material. This table shows, however, that the closed type of
pelvis predominates in the females. But the point to be em-
phasized by a study of the table is, that both the open and the
closed types of pelvis occur among the females of nearly all the
species. In every species where only one type of pelvis was
found, so few specimens were examined that the occurrence of
the other type is not precluded.

Aside from these variations with regard to the presence or
absence of the symphysis, the pelves of the pocket gophers
present certain characteristics which are constant. The pelvis
of Geomys bursarius has again been used, in comparison with
figures from Huxley (’75). The first two caudal vertebrae are
coossified with the sacral vetebrae and with each other forming
an extension of the sacrum which arches dorsad in such a way
that the second caudal vertebra is on a level with the dorsal
margin of the ischium with which it is coéssified. For this
reaon the line designated as the sacral axis (Sa) has been drawn
through the articulation of the sacrum with the ilium and the

THE PELVIS OF BURROWING MAMMALS 189

fusion of the second caudal vertebra and the ischium rather
than through the centra of the sacral vertebrae which are
arched dorsad.

The iliac and sacral axes (fig. 6, Ja and Sa) are very nearly
parallel. This is an advanced step in the attainment of a
horizontal pelvis, in which direction Huxley (’75) believed
mammals to be tending. In the case of the pocket gophers,
the symphysis has moved caudad and is evidently formed by the
ischial bones, as shown by the study of immature specimens.
The pubic bones are nearly parallel and their ventral margins
diverge posteriorly (fig. 6).

This condition is undoubtedly a high specialization. By
referring to the figure of the Monotreme pelvis (fig. 2), it will
be noticed that this structure in the primitive mammals pos-
sesses a2 symphysis which is very long and is formed by both
the ischium and the pubis. In still more primitive vertebrates,
amphibians and reptiles, morphologically the symphysis is even
longer.

A study of the pelvis of the rodents, in general, contributes
instructively to the understanding of specializations such as are
found in the pocket gophers. Apodontia of the family Apo-
dontiidae has a relatively short symphysis. Among the ‘mole-
rats’ (Bathyergidae), Bathyergus miratimus and Heterocephalus
glaber were examined in the flesh. The symphysis was found to
be very short in both cases, being a rod-like structure very much
like that found among the pocket gophers. Since only one
specimen of Bathyergus and four specimens of Heterocephalus
were examined, it is not known whether there is any variation
of the symphysis in this family. The position of the pelvic
axes are also unknown.

The beavers (family Castoridae) possess a long symphysis,
and the caudal vertebrae are not codssified as they are in the
pocket gophers. The condition in the guinea-pigs (Cavidae) is
very similar to that of the beavers and must be considered as a
generalized state, where the caudal vertebrae are free and a
well-developed symphysis is present. In a morphologic way,
the chinchillas (Chinchillidae) might also be included with

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 2

190 ROYAL NORTON CHAPMAN

these families, since the condition of the pelvis is a very general-
ized one.

Of the Jerboas (Dipodidae), the pelvis of Allactaga longior
was examined. The symphysis is very long, one-sixth the
length of the entire pelvis, and the pelvis is so inclined that the
iliac axis forms a large angle with the sacral axis. The caudal
vertebrae are free, save for a slight codéssification of the trans-
verse processes of the first caudal vertebra with the sacrum.

The ‘pocket rats’ (Heteromyidae), although in their sys-
tematic relations placed very close to the pocket gophers
(Geomydae), have a very well-developed symphysis. In the
pocket mouse (Perognathus flavus) the symphysis is over one-
sixth the total length of the pelvis, and in Dipodomys deserti it
is about one-fourth the length of the pelvis. No other repre-
sentatives of this family have been studied.

The porcupines (Hystricidae), as found from the study of
Hystrix, have the symphysis one-fifth the length of the pelvis,
and the iliac axis forms a large angle with the sacral axis of the
pelvis. The transverse processes of the first caudal vertebra
are fused to the sacrum, but it is important to note that the
caudal vertebrae are free from the ischium and, because of the
pelvic inclination, dorsal to it.

The mice (Muridae), with numerous genera with varying
habits, present some very interesting conditions. The deer
mouse (Peromyscus) and mice belonging to several other closely
related genera (Phyllotis, Akodon hirtus) have been examined,
and in all cases the symphysis is short and the pelvis is slightly
less inclined than it is in some of the other rodents; that is, the
pelvis approaches the horizontal position. ‘The sacral and first
two caudal vertebrae are codssified along their transverse proc-
esses, but the caudal vertebrae are free from the ischium and
the pubic bones converge posteriorly.

The meadow mice (voles, subfamily Microtinae) illustrate
another step in the attainment of a horizontal pelvis and the
loss of the symphysis. However, it is only a part of this sub-
family which is so greatly modified. The musk rat (Ondatra)
possesses a symphysis and, although the pelvis is nearly hor-

THE PELVIS OF BURROWING MAMMALS _ 19%

izontal, the sacrum, along with the caudal vertebrae, is arched
‘dorsally so that the pelvis is far from being a closed box. The
red-backed mouse (Hvotomys grapperi) and the meadow mouse
(Microtus pennsylvanicus) have no’ symphysis and the ventral
margins of the pubic bones are nearly parallel in the adult
meadow mouse (figs. 8 and 9). The pelvis is very nearly hori-
zontal, but the direction of the pubic bones is quite oblique
ventrocaudad. In the young meadow mouse the pubic bones
converge posteriorly and form the symphysis at their points of
contact.

The house mouse (subfamily Murinae) has a short symphysis
formed by the point of contact of the two converging pubic
bones as in the young meadow mouse. The pelvis is also less
horizontal than it is in the meadow mouse (figs. 10 and 11).

Throughout, the mouse family (Muridae), with the exception
of some of the meadow mice, the estate of the pelvis is quite
different from that of the pocket gophers (Geomydae). In the
former, the pubic bones converge posteriorly, forming a sym-
physis at their point of contact (fig. 10), while in the latter the
pubic bones diverge posteriorly and the symphysis, when it is
present, is formed by the union of bony processes from each
side (fig. 4).

The pelvis of the ‘Quette vole’ (Myospalax), the only member
of the family Spalacidae which has been studied, was examined
in the flesh. The symphysis is very short (one-twenty-fourth
of the length of the pelvis) and appears to be formed by the
fusion of two long bony processes, such as that prevailing
among the pocket gophers. The ventral margins of the pubic
bones also suggest the condition among the pocket gophers in
that they are quite parallel.

The squirrels (Sciuridae), in so far as they have been ex-
amined, seem to be quite uniform in the possession of a well-
developed symphysis and having the pelvis inclined to form a
large angle with the sacrum. They have, therefore, been con-
sidered to be a group in which the pelvis is quite generalized and
not of special interest in the present consideration.

192 ROYAL NORTON CHAPMAN
2. The pelvic bones of the Insectivora

The pelvis of the Insectivora, as described by Lecke (’84), is
found to exhibit a series of changes similar to those described
above for the rodents. The material which has been examined
in the course of this investigation agrees with the descriptions
given by him.

The pelvis of the tree shrew (Tupaia) is very generalized, the
symphysis is long and the angle of inclination formed with the
sacral axis is large. Elephatulus also has a similar pelvis.

The hedgehogs (family Erinaceidae) possess a short symphysis
which Lecke (’84) describes as cartilaginous. However this
may be, the pubic bones converge posteriorly in all the specimens
which have been examined and the pelvis is more like that of
the mice than like that of the pocket gophers in that the sym-
physis is formed at the point of contact between the two
converging pubic bones.

The shrews (family Soricidae) represent a high specialization
with regard to the pelvis, and Lecke (’84) speaks of them,
together with the moles (Talpidae), as differing from all other
mammals in the possession of diverging pubic bones. The
‘mole shrew’ (Blarina talpoides), the ‘short-tailed shrew’ (Blarina
brevicauda), Cryptotis parva, ‘musk shrew’ (Crocidura hal-
conus), Pachyura luzoniensis, and the ‘wood shrew’ (Sylvisorex
gemmeus) have been examined and in every case, regardless of
sex, the symphysis has been found to be absent. The pubic
bones not only diverge posteriorly, but at the anterior end tend
to converge until, in the ‘mole shrew’ (Blarina talpoides), they
are less than a millimeter apart at the iliopectineal prominence
(fig.12 and 13). There is thus formed a virtual secondary symphy-
sis, but, unlike the original symphysis, it is at the anterior end
of the pubic bones and dorsal to the digestive tract and the urino-
genital ducts. In all the specimens examined the pelvis is
horizontal and very much narrowed, except in the ‘musk
shrew’ (Crocidura).

Of the moles (family Talpidae) the ‘star-nosed mole’ (Con-
dylura cristata), Scapanus townsendii, Scalopus aquaticus aquat-

THE PELVIS OF BURROWING MAMMALS 193

icus, the European mole (Talpa europea), Scaptochirus sp.,
Urotichus sp., and Galemys sp., were examined. In all but
Galemys the symphysis is absent, and even in this case the pubic
bones diverge posteriorly and the symphysis appears as a rod
of bone uniting the two sides. It was found that in Scaptochirus
the pubic bones converge anteriorly and form a complete
secondary symphysis dorsal to the viscera. Dobson (’82) found
the same condition in Mogera wogura (fig. 14). In all cases
the pelvis is horizontal (fig. 15) and the transverse processes
of the sacral and caudal vertebrae are codssified and the second
caudal vertebra is fused to the ischium in all but the star-nosed
mole (Condylura) in which the transverse processes of the
caudal vertebrae are greatly reduced. The pelvis is very long
and narrow in all cases.

Solenodon, the only member of the family Solenodontidae
examined, possesses the symphysis well developed. Among the
Tenrecidae, Ericulus setosis, Tenrec ecaudatus, Hemicentetes
semispinosus, Ericulus setosus, and Oryzorictes hova have been
studied. Here the pubic bones converge posteriorly and the
symphysis is formed by the line of contact between the two
bones rather than by a rod of bone from the two sides as in the
two preceding families. The bones are slightly separated in
Tenree and Ericulus with a heavy ligament uniting them, but
in others the symphysis is present. The inclination of the
pelvis is greater than among the moles and the shrews.

The South African golden mole (Chrysochloris leucorhina of
the family Chrysochloridae) was examined in the flesh. It was
found to resemble other moles (Talpidae) in that the pubic
bones diverge posteriorly and are widely separated. The angle
of the pelvis was not determined.

Lecke (’84) reports that the ontogeny of the insectivors agrees
with the phylogeny, that the pelvis represents a high degree of
specialization and that the symphysis appears in the embryo of
all the members of the group and is lost later.

194 ROYAL NORTON CHAPMAN

3. The pelvic bones of the marsupial mole

Attention must be called to the marsupial mole (Notoryctes)
in this connection (fig. 16). Huxley (75) considered the mar-
supial pelvis to be inclined to form a large angle with the sa-
crum, as is the case among marsupials in general, but, in the
case of the marsupial mole, the pelvis is horizontal and is firmly
coéssified with the sacrum. Furthermore, the symphysis is
greatly reduced (fig. 17), which is another departure from the
usual condition among the marsupials and an approach to the
condition found among other burrowing mammals.

From the above consideration of the pelvis of burrowing
mammals it is clear that the pelvis of such a form, regardless of
the taxonomic group to which it may belong, may be distin-
guished from the pelvis of mammals in general by -its more
horizontal position, its more complete ossification with the
vertebral column, and by the great reduction or the entire absence
of the symphysis. In the absence of the symphysis, a pubic
ligament extends between the two innominate bones.

Among both the rodents and the insectivors there are forms
which possess the generalized pelvis with the marked inclination
and the long symphysis, while, on the other hand, there are
burrowing forms in both groups which possess the typical
‘burrowing pelvis.’

4. A classification of the pelves of burrowing mammals

The pelves of the burrowing forms may be classified as: the
typical burrowing (moles, shrews, pocket gophers, and certain
meadow mice), in which the ventral margins of the pubic bones
are horizontal and diverging posteriorly, and the symphysis is
absent or, when present, formed of a transverse rod of bone
(figs. 4 to 9 and 12 to 15); the semi-burrowing (mice in general
and certain insectivors), in which the pubic bones are inclined
caudoventrad and converge posteriorly to form a symphysis at
their points of contact (figs. 10 and 11). The typical burrowing
pelvis may be further divided into the narrow type (moles,

THE PELVIS OF BURROWING MAMMALS 195

shrews, and meadow mice), in which the symphysis is always
absent and the pelvis is very narrow, and the broad type (pocket
gophers and a few others), in which the symphysis may or may
not be present. This classification becomes clearer when placed
in synoptic form as follows:

Typical burrowing forms ( Narrow type of pelvis
1. Ventral margins of pubic 1. Symphysis always absent.
bones horizontal and di- 2. Pelvis narrow.
verging caudad. 3. Includes: moles, shrews,
2. Symphysis absent or, if pres- and meadow mice.
ent, formed of a trans-
~ verse rod of bone. Broad type of pelvis
3. Includes: moles, shrews, 1. Symphysis may or may
pocket gophers and cer- not be present.
Pelvis of bur-} tain meadow mice. 2. If present, in form of a
rowing mam-| 4. Figures 4 to 9 and 12 to 15. transverse rod.
mals | Semi-burrowing forms 3. Includes: pocket gophers
1. Pubic bones inclined caudo- | and a few others.

ventrad, converging pos-
teriorly to form a sym-
physis at points of con-
tact.
2. Includes: mice in general
and certain Insectivors.
3. Figures 10 and 11.

THE STATUS OF THE RECTUS ABDOMINIS MUSCLES

Having considered the variations of the symphysis pubis
and having found that in many cases it is absent, a consideration
of the muscles which are typically attached to the bones forming
the symphysis, the M. rectus abdominis, is very instructive.
Typically, the rectus abdominis nruscle is paired, the right and
left muscles lying on either side of the midventral line, with the
origin on the pubic bones at the symphysis and the insertion
on the first two or three ribs near their costal cartilages. The
rabbit (fig. 18) illustrates very well the usual condition of these
muscles near their points of origin.

When the pelvis is modified by the loss of the symphysis, it is
obvious that these muscles must also become modified, for the
bone to which they normally attach is absent. It has been

196 ROYAL NORTON CHAPMAN

found that the rectus abdominis muscles show modification to
be directly correlated with the variations of the symphysis.

In the absence of the symphysis, the usual condition of the
rectus abdominis muscles is a crossed one; the rectus muscle of
the right side originating on the left pubic bone and the left
muscle originating on the right. The two muscles cross a short
distance anterior to their origin and proceed to their insertion
in the usual manner (fig. 22).

1. The relations of the rectus abdominis muscles of Rodentia

Among the pocket gophers (Geomyidae), the rectus abdominis
muscles show variations to be correlated directly with those of
the pelvis. In the common pocket gopher of the Middle West
(Geomys bursarius), the muscles spread out into a tough
aponeurosis Just anterior to the pelvis when the symphysis is
absent (fig. 19), while, in case the symphysis is present, the
muscle fibers may be traced to their origin on the bones forming
the symphysis (fig. 20).

In the one specimen of Geomys hispidus which has been ex-
amined, the muscles attached to the bones forming the sym-
physis and are, in the main, uncrossed, although some of the fibres
do appear to cross over to the opposite side.

In Thomomys bottae bottae the muscles are crossed and
attached to the bones forming the symphysis when the symphysis
is present (fig. 21) or to the pubic ligament when it is absent.
Thomomys bottae anitae possesses the same relation of the
muscles as the preceding subspecies. In Thomomys monticola
mazama the muscles are crossed and attached to the pubic
ligament. Thomomys bottae pascalis and Thomomys fulvus
intermedius possess crossed muscles which are attached to the
bones at the sides of the pelvis in the absence of the symphysis.

The pocket gophers, therefore, possess rectus abdominis
muscles which, like the pelvis, are varying in their relations. In
some cases the muscles exhibit only the rudiments of a crossed
condition, the majority of the fibers composing the right and
left muscles originating on the right and left sides, respectively,

THE PELVIS OF BURROWING MAMMALS 197

and only a few of the fibers of the ventral layers of the muscles
crossing over to the opposite side. In other cases the crossed
condition is completely developed and the muscles originate on
the bones at the sides of the pubic opening. Even the various
subspecies of Thomomys bottae which have been examined
differ as to whether the muscles originate on the bones forming
the symphysis—in the absence of the symphysis on the pubic
ligament—or on the bones at the sides of the pubic opening
(fig. 22).

Throughout the rodent group the tendency of the muscles to
become crossed seems to accompany the loss of the symphysis.
In the ‘mole rats’ (Bathyergidae) the muscles are only slightly
crossed in the few specimens which have been examined. Among
the mice (Muridae) the muscles are crossed in all the specimens
examined which have no symphysis (fig. 23). Even the mice
with the short symphysis formed by the converging pubic bones
possess crossed rectus abdominis muscles which originate on the
posterior border of the ischium (fig. 24).

2. The relations of the rectus abdominis muscles of Insectivora

The insectivors present a series of muscle variations as com-
plete as the symphyseal variations themselves. The crossing of
the rectus abdominis muscles accompanies the loss of the sym-
physis in all the cases which have been examined (figs. 25 and 26)
except the South African golden mole (Chrysochloris), in which
case there is a specialized condition of the gracilis muscles,
whereby they meet on the midventral line and form a firm
union with the rectus abdominis muscles.

Lecke (92) has studied the modification of the rectus ab-
dominis muscles in the Insectivora, both as to their phylogeny
and their ontogeny, and has concluded that their crossing is to
be correlated with the loss of the pubic symphysis. He states
that the beginning of the crossed condition is found in the
‘hedgehogs’ (Erinaceidae) and the Gymnura and that it is
mostly highly developed in the moles (Talpidae) and the shrews
(Soricidae). Thus, the burrowing insectivors agree with the

198 ROYAL NORTON CHAPMAN

other burrowing mammals in the possession of a pelvis which
lacks a symphysis and in the peculiar eae condition of the
rectus abdominis muscles.

THE RELATION OF THE PELVIC ESTATES AND THE HABITS
1. The attainment of a horizontal pelvis

If the evidence offered by the probable ancestors of mammals
is accepted, it must be considered that the pelvis was originally
- confined to the ventral side of the body and served to articulate
the hind limbs which pushed the body along over a parallel
surface. Upon the attainment of a strictly terrestrial habit,
the body was elevated above the substratum and the limbs
necessarily functioned for support as well as for propulsion.
This necessitated a firm union between the limbs and the frame
work of the body. This was accomplished by the union of the
pelvis and the vertebral column.

Huxley (’75) has followed the series of changes from a condition
in which the pelvis was nearly perpendicular to the vertebral
column (fig. 1) to a condition in which it is nearly parallel (fig. 3).
In the first instance the locomotive force exerted by the limbs
must pass along a line extending dorsally through the pelvis (the
iliac axis, Ja) and then anteriorly, almost at a right angle, along the
vertebral column (the sacral axis, Sa). In the case of the mammal
with the nearly horizontal pelvis, the force exerted passes along
a nearly straight line, for the pelvis and the vertebral column
are in nearly the same plane (the iliac and sacral axis nearly
coincide). The mechanical advantage of such a structure is at
once evident in the more direct application of the locomotive
force to the framework of the body.

The series of pelves of burrowing mammals presented in this
paper agree in exhibiting the pelvis in every case more nearly
parallel to the vertebral column than in any case described by
Huxley (75). <A further modification is seen in the codossifi-
cation of the sacral and caudal vertebrae and, in many cases,
the codssification of these posteriorly with the ischium (fig. 6).
This provides a still more firm union between the pelvis and the

THE PELVIS OF BURROWING MAMMALS 199

vertebral column, and the line of locomotive force exerted by:
the hind limbs passes directly along the vertebral column.

The pocket gophers, moles, and shrews burrow continually in
search of their food, and in so doing depend upon their fore
limbs for propulsion. A pocket gopher, for example, digs the
dirt loose and uses the hind limbs as a brace while so doing.
The pelvis, parallel to the vertebral column and firmly co-
ossified with it, offers the greatest advantage in directing the
locomotive force along a straight line to the anterior part of the
body.

The firm union of the pelvis and the vertebral column is well
illustrated by the marsupial mole (fig. 16) in which the co-
ossification is complete and the structure is widely different
from ‘that of other marsupials. Among the moles, Scaptochirus
and Mogura (fig. 14) are also good examples of a rigid pelvis
and sacrum. In these forms the iliopectineal prominences of
the two sides have united to form a secondary symphysis dorsal
to the viscera and just anterior to the articulation of the hind
limbs. .

Having seen that the pelvis of the mammalian ancestors was
perpendicular to the vertebral column and that it became more
nearly horizontal when the terrestrial habit was attained and
finally that, in the mammals which had the habit of continual
burrowing, the pelvis became horizontal, it seems most logical
to conclude that these modifications are to be correlated with |
the increased locomotive force exerted along the line of the
pelvis and the vertebral column. The attainment of the hori-
zontal position by the pelvis must be further considered in
connection with the loss of the symphysis pubis.

2. The loss of the symphysis pubis

As the pelvis attains the horizontal position, the sacrum and
the caudal vertebrae become more completely codssified and the
symphysis moves posteriorly, probably due to the enlarged
flexor muscles of the hind limbs which attach to the posterior
portion of the pelvis, until it is formed entirely by the ischial

200 ROYAL NORTON CHAPMAN

bones. In the ordinary terrestrial habitat where the animals
are free-ranging, an optimum condition is evidently reached, in
which the force exerted by the hind limbs in propelling the body
will be directed along the spinal column through the ilium which
is placed at an efficient angle. But there are certain limitations
in the attainment of a horizontal position of the pelvis. The
symphysis may be brought so near the vertebrae that there will
not be sufficient room to allow the passage of the fetus at the
time of birth. When the caudal and sacral vertebrae are not
coossified, the caudal vertebrae may be elevated in providing
room, but in the event of their codssification the bones dorsal
to the pelvic opening are rigid and the pubis, obviously, must
provide the extra space needed for the passage of the fetus. The
horizontal pelvis and the absence of the symphysis are therefore
conditions which are interlinked, the latter dependent upon the
former.

There are two types of pelves among the burrowing mammals
with regard to width, as has already been stated, and, in the
broad pelvis of the pocket gopher, the symphysis sometimes
persists, while in the narrow pelvis of the moles and shrews it
is always absent. These two types of pelvis are to be correlated
with two methods of burrowing.

The pocket gophers have the habit of standing with the final
feet wide apart while digging, so that the dirt may be thrown
out between them. The moles, on the other hand, press the dirt
to the sides and push it upward with the back—literally crawling
between the earth particles.

Seton (’09), in describing the habits of the western pocket
gopher, wrote:

‘‘TIts method of burrowing, as observed in a captive specimen,
is to loosen the earth with the powerful front claws, as it stands
with the hind feet advanced and wide spread, then throw it
backward between the hind legs, to be further passed on by the
hind feet; and when a sufficent pile is ready, the gopher turns
round and pushes with its broad head and powerful front feet,
forcing the pile ahead of it to the first gallery, up that and out,
usually without exposing itself.”

THE PELVIS OF BURROWING MAMMALS 201

Such action undoubtedly calls for strenuous exertion on the
part of the hind legs while loosening the earth and again while
the earth is being pushed out of the tunnel. It also requires a
broad pelvis in order that the hind legs may be placed well
apart. This broad pelvis possesses a large pelvic aperture,
allows more room for the passage of the fetus, and hence the
loss of the symphysis is less imperative than it would be if the
pelvis were narrower.

Seton (09) describes the star-nosed mole (Condylura cristata)
as piling up the earth as the pocket gopher does, while working
in moist soil, and as tunneling without throwing up the earth
in the dryer soil. Anyone who has seen the ridges of earth
raised up by moles in a garden will realize that they force
their way through the soil in a manner quite different from that
of the pocket gopher. And from the length of the tunnels
which they are capable of making, even in a single night, one
can well believe that a great deal of force is being continually
' exerted. This crawling between the earth and pushing it aside
has a continual tendency to compress the pelvis laterally rather
than to broaden it, as the action of the pocket gopher tends to
do. Such a narrow pelvis so restricts the pelvic aperture that
the loss of the symphysis is imperative to allow the passage of
the fetus.

Concerning the shrews, it may be said that their pelvis, so
similar to that of the moles, though less highly specialized, is a
result of similar habits. The shrews burrow, though they are
less strenuous in their operations, confining themselves to the
surface of the ground and the litter which covers it. Seton refers
to the habits of the short-tailed shrew (Blarina brevicauda) as
follows:

‘The furrowed—sometimes tunneled—track that this animal
leaves in the snow is the exact expression of its methods and of
its summer life beneath the leaves and rubbish in the woods.”’

From this author’s notes the following quotation is taken
describing an observation:

‘“‘Free as a mole in the soil, he drove his sub-leaf gangway
where he would, and doubtless lived on the country as he went.

202 ROYAL NORTON CHAPMAN

This then was his way of life—this little inter world between
the floor and carpet was for him; and thus I learned why he had
bartered his eyesight for keener powers of smell and touch.”

Shull (07) concluded from his observations in the field and
laboratory that shrews, like moles, push the earth aside as they
burrow. We may therefore class the moles and shrews together
with respect to the forces which have developed their narrow
horizontal pelves lacking the symphysis.

Brief attention should be given to those forms in which the
conditions are less specialized. Some of the meadow mice
(Microtinae) possess habits similar to those of the shrews which
live in the litter over the surface of the ground. Seton (’09)
may again be quoted with regard to the habits of some of the
mice:

“Tf we make of our six common mice a ladder to show their
chosen elevations, we shall put Peromyscus arcticus (the deer
mouse) at the top, far above the ground, next Evotomys (the
red-backed wood-mouse) very near the ground, next Microtus
minor (the small meadow mouse or vole) a little below, and
lowest of all, much of the time below the surface, we find the
present subterranean group (the common meadow mouse,
Microtus pennsylvanicus).”’

The modification of the pelves in these mice conforms in a
remarkable way to the elevations as given in this account of
their habits. The deermouse (Peromyscus) possesses a sym-
physis and the pubic bones converge posteriorly, while the
meadow mouse (Microtus pennsylvanicus) is without the
symphysis and the ventral margins of the pubic bones are
parallel. Other forms with less specialized pelves have not been
studied sufficiently to determine all the correlations between
their habits and structure. Mice in general live more under
cover, where they crowd under and between obstacles, than in
the open where their action is free and unimpeded, and their
progress, between and under the litter and under the vegetation
which covers the ground, may have to do with the approach of
the pelvis to the position parallel to the sacrum and the shortening
of the symphysis pubis, as in the typical burrowers.

THE PELVIS OF BURROWING MAMMALS 203

What has been said of the rodents which have a tendency to
a reduction of the symphysis may also apply to the insectivors,
such as the hedgehogs (Erinaceidae), which have a similar
pelvis. Here again there is a correlation between the habits
and structure which is parallel to the condition among the
rodents, but a more extended study is necessary before making
final statements.

3. The persistence of the symphysis in certain cases

There are certain cases. which seem, upon superficial exami- °
nation, to present exceptions to the explanations offered in this
paper. Among the squirrel family (Sciuridae), the so-called
ground squirrels, woodchucks, and a few others are considered
fossorial mammals, and in all of these the symphysis is well de-
veloped. However, all of these animals burrow only to make
a shelter for themselves. Burrowing is a secondary matter
with them and but a very small portion of their time is devoted
to it. On the other hand, they are adapted to rapid running,
for sitting erect, and for other activities which call for a pelvis
very different from that found in pocket gophers and moles
which are so restricted to continual burrowing.

The mole rats (Bathyergidae) have many characteristics of
well-adapted burrowers. Thomas (’85) considers them to have
been specialized for a purely subterranean life. They are said
to burrow in the sand and throw up the loose earth in piles with-
out exposing themselves. Two of these forms were examined
in the flesh (Bathyergus and Heterocephalus glaber) and the
symphysis was found to be present, but, as explained above, it
is very short and appears like a rod reaching between the two
sides of the pelvis, a condition very similar to that found in the
pocket gophers.

Since the habits and structure are so similar to those of the
pocket gophers, the presence of the symphysis may be explained
in ‘the same way as that which is sometimes present in the
pocket gopher is explained. The pelvis is‘broad and has allowed

204 ROYAL NORTON CHAPMAN

sufficient room for the passage of the fetus at the time of birth,
and there has, therefore, been no necessity for its loss.

The pocket gophers evidently represent a stage in which the
room within the pelvis is greatly restricted, and the symphysis
is in the process of being lost. The pelvis has attained a position
nearly parallel to the sacrum and the caudal vertebrae, the
sacrum and the ischium are codssified, forming a closed box.
At the same time, the pubic bones have diverged caudally, and
the symphysis, when it is present, is maintained by a rod of
bone which seems to be unnecessary otherwise, for when the
symphysis is absent, a portion of the bone on either side is also
absent (compare figs. 4 and 6).

Having arrived at this condition, it would seem to te a distinct
advantage for the female to lose the symphysis and thus pro-
vide for the easier passage of the fetus at the time of birth. No
advantage can be seen in the retention of the symphysis for the
attachment of the rectus abdominis muscles, since they have
been. found to accomodate themselves to all of the modifications
of the pelvis. There seems to be no reason, therefore, why the
female should retain the symphysis.

The males, as far as can be seen, have no call for the loss of
the symphysis. On the other hand, they have well-developed
levator muscles of the penis attaching to the bones forming the
symphysis which, in a mechanical way, would seem to justify
its retention.

From this consideration of the pocket gophers, it is concluded
that, when the mechanical necessities of propulsion have brought
the pelvis to a horizontal position, caused it to fuse with the
sacrum, and the pubic bones to diverge posteriorly in such a
way that a portion of bone, otherwise unnecessary, must be
present in order to maintain a symphysis, other forces appear
which are different in the two sexes. Thus, when the symphysis
has been greatly reduced and is on the verge of being lost, the
step seems to be hastened in the female associated with the
necessity of more room in the pelvis and retarded in the male,
an association with the necessity of a support for the levator
muscles of the penis.

THE PELVIS OF BURROWING MAMMALS 205

The occurrence of both types of pelvis in the females of the
same species cannot be explained unless it is the case of a
structure in the course of reduction, the pocket gophers repre-
senting a stage in which the first step is being taken, while the
mole-rats (Bathyergidae) have not yet reached this stage, the
moles and shrews having passed on to the completed stage as
shown by the absence of the symphysis in both sexes.

4. The function of the crossed rectus abdominis muscles

The relations of the rectus abdominis muscles show a general
conformity to the condition of the pelvis. Lecke (’92) believed
that the crossed conditions of these muscles was for the purpose
of compensating the loss of the symphysis, giving support to
the viscera. As stated above, he found the beginning of the
crossing of the muscles where there was the first tendency toward
the loss of the symphysis in the hedgehogs (Erinaceidae) and
found the crossing of the muscles most complicated in the
moles (Talpidae) where ‘tthe symphysis is entirely absent. The
South African golden mole (Chrysochloris), constituting an
exception to this, but in this form the gracilis muscles meet in
the midline and, together with the rectus muscles, form an
efficient support to the viscera.

The rectus abdominis muscles of the pocket gophers have
been shown to vary in their attachments, but the crossing of the
muscles has accompanied the loss of the symphysis. In the
pocket gopher (Geomys bursarius), a strong aponeurosis is
formed which affords an efficient support to the viscera. Even
though there is a variation in the muscles throughout the family,
the provisions for the support of the viscera, by the crossing of
the muscles or some other device, is constant. The fact that
the muscles are crossed in some of the mice (Muridae) in which
the symphysis is present would substantiate Lecke’s findings,
that the muscles have a tendency to cross over in forms in which
the symphysis shows the beginnings of reduction.

Here, then, is another parallelism between the Rodentia and
the Insectivora which may be referred to the similar necessity

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 2

206 ROYAL NORTON CHAPMAN

for the support of the viscera in the absence of the symphysis.
In both groups there are cases, certain pocket gophers and the
South African golden mole, where the symphysis is absent and
the rectus abdominis muscles are not crossed, but in both groups
provisions for the support of the viscera are constant.

SUMMARY

This paper is based upon a study of structures in animals
belonging to different systematic groups but which have similar
habits. All of the forms are found to agree in the general
structure of the pelvis which is correlated with the habits, and
the minor differences in the structure are correlated with specitic
differences in the habits.

The moles (Talpidae), so well adapted to the burrowing habit,
possess horizontal pelves which are firmly codssified to the
vertebral column, the ventral margins of the pubic bones are
horizontal and diverge from each other posteriorly, and the
symphysis pubis is absent. In some cases (Scaptochirus and
Mogura) a second symphysis is formed dorsal to the viscera,
contributing to the greater strength of the pelvis.

The pocket gophers (Geomydae), also well adapted to the
burrowing habit, possess a horizontal pelvis firmly codéssified to
the vertebral column, and have the ventral margins of the pubic
bones horizontal and diverging from each other posteriorly.
The symphysis, however, is present in the males and in some of
the females, while it is absent in others even of the same species.

There are other rodents (meadow mice, Microtinae) and in-
sectivors (shrews, Soricidae) which possess similar pelves with
various degrees of specialization. These forms live in runways
in the debris which covers the ground, and the degree of their
specialization is correlated with the depth to which they drive
their tunnels.

The marsupial mole (Notoryctes) has a habit similar to that
of other moles and has departed from the marsupial type of
pelvis, as is shown by the firm codssification of the pelvic bones
with the vertebral column, the greatly reduced symphysis, and
the horizontal pelvis.

THE PELVIS OF BURROWING MAMMALS 207

The parallelism exhibited by these forms in the attainment
of a horizontal pelvis which is firmly fused with the vertebral
column and in which the symphysis is greatly reduced or absent
is correlated with the mechanical force exerted by the hind
limbs in propelling the body during continuous burrowing. This
position of the pelvis transmits the locomotive force exerted by
the hind limbs along a straight line from the articulation of the
limbs to the anterior part of the body. The action of the flexor
muscles is much greater than that of the extensor muscles, and
consequently their attachments, posterior to the articulation of
the hind limbs, have been developed at the.expense of the
anterior portion of the pelvis. When this condition is attained,
it is believed that the symphysis is a hindrance to the passage
of the fetus at the time of birth, for the pelvis has become a
closed box. The bones are so firmly codéssified dorsally and so
necessary for the attachment of the large flexor muscles of the
hind limb that it follows that the symphysis is lost in orerd to
provide the necessary room for the passage of the fetus.

The retention of the symphysis in the pocket gophers and a
few others is attributed to the fact that in these forms the pelvis
is very broad, while in the moles, shrews, and the meadow mice
the pelvis is so narrow that the elimination of the symphysis
has been more imperative.

The broad and narrow pelves are correlated with two methods
of digging. The pocket gophers stand with the hind legs wide
apart and throw the earth back between them, while the moles,
shrews, and the others with the narrow pelves push the earth
to the sides and crowd their way through the soil. In the
former case the pelvis is constantly drawn out laterally by the
action of the muscles, while in the latter case the pelvis is
compressed by the action of the hind limbs.

A parallelism is also found in the crossing of the rectus ab-
dominis muscles in the forms in which the symphysis is reduced
or absent. This is believed to function as a support for the
viscera in the absence of the symphysis. Exception to this cross-
ing of the rectus abdominis muscles have been found among both
the rodents and the insectivors which have lost their symphysis,

208 ROYAL NORTON CHAPMAN

but in such cases some other means of support for the viscera
have been supplied.

This series of structural modifications so well correlated with
the habits of the animals may justify one or more of the following
conclusions: (1) that the correlation is a coincidence due to chance
and that by the same chance other types of pelvis do not exist
among the forms studied; (2) that from a number of chance varia-
tions in each group of burrowing mammals, the most efficient
burrowing structure has been preserved by natural selection;
(3) that similar forces acting in a similar way have been factors
in causing similar modifications of structure in animals belonging
to widely different systematic groups, just as tuberosities develop
on bones at the points of attachment of muscles of strenuous
action. A combination of the second and third conclusions seems
to be in entire harmony with the findings in this study and
ascribes a dominant influence to the environment in the produc-
tion of the parallelisms in the forms studied.

LITERATURE CITED

Broek, A. J. P. v. d. 1914 Studien zur Morphologie des Primatenbeckers
Morphologisches Jahrbuch. 1914-15; Bd. 49; ppl-118.

Dosson, C. E. 1882 A monograph of the Insectivo_a, systematic and ana-
tomical. Par. I. London, 1892.

Huxuey, Tuomas Henry 1875 On the characters of the pelvis in the Mam-
malia, and the conclusions respecting the origin of the Mammalia
which may be based on them. Proceedings of the Royal Society of
London, 1878-79, pp. 395-405.

Lecxre, W. 1884 Mammalia, pelvis. Bronn, Klassen und Ordnung des Tier-
reichs. 1892, Bd. 6, s. 571.

1892 Mammalia, Myology. Bronn, Klassen und Ordnung des Tier-
reichs, Bd. 6, s. 784.

OsporNn, Henry L. 1894 Notes on Geomys bursarius. Science, vol. 23, p. 102.

Seton, Ernest THompson 1909 Life; Histories of northern animals. New
York, 2 vols.

Suinner, H.W. 1903 Adaptations to aquatic, arboreal, fossorial, and cursorial
habits in mammals. III. Fossorial adaptations. American Natural-
ist, vol. 37, pp. 819-825.

SHutu, A. FRANKLIN 1907 Habits of the short-tailed shrew, (Blarina brevi-
cauda). American Naturalist, vol. 41, pp. 495-522.

Tuomas, P. 1885 Proceedings of the Zoological Society of London, 1885,,p. 611.

Drawings by Helen Sanborn Chapman, Department of Animal oleae The

University of Minnesota.

PLATES

ABBREVIATIONS
I1., ilium Il.a., iliac axis
Pb., pubis Ip.a., iliopectineal axis
TIs., ischium Ob.a., obturator axis
M.b., marsupial bone S.a., sacral axis

Pbl., pubic ligament

209

PLATE 1

EXPLANATION OF FIGURES

1 Pelvis of crocodile (Crocodilus). Lateral view. After Huxley.

2 Pelvis of duck bill (Ornithorhyncus). Lateral view. After Huxley.

3 Pelvis of rabbit (Lepus). Lateral view. After Huxley.

4 Pelvis of pocket gopher (Geomys bursarius), closed type. Ventral view.

5 Pelvis of pocket gopher (Geomys bursarius), closed type. Lateral view.
1 : -

6 Pelvis of pocket gopher (Geomys bursarius), open type. Ventral view.

THE PELVIS OF BURROWING MAMMALS PLATE 1
ROYAL NORTON CHAPMAN

TATE

a8

Lp @.

5

PLATE 2

EXPLANATION OF FIGURES

Pelvis of pocket gopher (Geomys bursarius), open type. Lateral view

Pelvis of meadow mouse (Microtus pennsylvanicus). Ventral view. X 3.
Pelvis of meadow mouse (Microtus pennsylvanicus). Lateral view. X 3.
Pelvis of house mouse (Mus musculus). Ventral view. X 4.

Pelvis of house mouse (Mus musculus). Lateral view. X 4.

Pelvis of shrew (Sorex sp.). Ventral view. X 52.

212

‘THE PELVIS OF BURROWING MAMMALS PLATE 2
ROYAL NORTON CHAPMAN

213

13
14
15
16
17

PLATE 3
EXPLANATION OF FIGURES

Pelvis of shrew (Sorex sp.). Lateral view. X54.

Pelvis of Mogera wogura. Ventral view. After Dobson. X 23.
Pelvis of star-nosed mole (Condylura cristata). Lateral view. X 2.
Pelvis of Marsupial mole (Notoryctes sp.). Ventral view. X 3.
Pelvis of Marsupial mole (Notoryctes sp.). Lateral view. X 3.

214

THE PELVIS OF BURROWING MAMMALS PLATE 3
ROYAL NORTON CHAPMAN

PLATE 4

EXPLANATION OF FIGURES

18 Rectus abdominis muscles of rabbit (Lepus). After Huxley.

19 Rectus abdominis muscles of pocket gopher (Geomys bursarius), open
type of pelvis. ’

20 Rectus abdominis muscles of pocket gopher (Geomys bursarius), closed
type of pelvis.

21 Rectus abdominis muscles of Western pocket gopher (Thomomys bottae
bottae), closed type of. pelvis.

216

THE PELVIS OF BURROWING MAMMALS PLATE 4
ROYAL NORTON CHAPMAN

bo
pa
~J

PLATE 5 :

EXPLANATION OF FIGURES

22 Rectus abdominis muscles of Western pocket gopher (Thomomys bottae
pascalis), open type of pelvis.

23 Rectus abdominis muscles of meadow mouse (Microtus pennsylvanicus).

24 Rectus abdominis muscles of house mouse (Mus musculus).

25 Rectus abdominis muscles of star-nosed mole (Condylura cristata).

26 Rectus abdominis muscles of shrew (Sorex sp.).

218

THE PELVIS OF BURROWING MAMMALS PLATE 5
ROYAL NORTON CHAPMAN

219

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MARCH 17

THE POSTNATAL DEVELOPMENT OF THE SUPRA-
RENAL GLAND—AND THE EFFECTS OF INANITION
UPON ITS GROWTH AND STRUCTURE IN THE AL-
BINO RAT

C. M. JACKSON

Institute of Anatomy, University of Minnesota, Minneapolis

TEN FIGURES

CONTENTS
USE SETS GN ive Tir ght 03 Fos Be Ren sr OS ee cr eae 222
Wolumes ofsuprarenal cortex and umedullan ska. os 2 cto sous es deans os aa 228
Volumes of suprarenal parenchyma and stroma.....:............0.....000- 232
pase al. parenenyima cells and nuclei. .2.00, 4259.05). let sods woo bs ee 233
Cell division—frequency of mitosis.............. ia treet Sts WE NEY ts hey 239
Histological changes in the suprarenal gland................0.......000000 244
PE NGriMal MOS Natal MIS LOPCHOSIS. 2°. 2450)... 5 5.cae-ols iisaatae sate da clas ape ae 45,0 244
2. Changes in young rats stunted by underfeeding...................... 261
3. Changes in rats refed after stunting by underfeeding................ 263
4. Changes in adult rats after acute or chronic inanition............... 2A4
NMorpnopenesis of the suprarenal gland’. 2.5... 224406. 0. s ese deeds cee tele 269
"SUS ads hi oe Weare GR gl id ph A a a 2 aR in eine OR One a 270
PReTAUEECICILOO es, Ave t,t oy eked ta ek ee he a eae eek. Ep. 277

In a previous paper (Jackson, ’13) the relative growth curve of
the suprarenal gland in the albino rat was described, including
the characteristic difference in weight according to sex. In
later papers (Jackson, ’15 a, ’15 b) the changes in the weights of
the suprarenal in underfed young rats, and in adult rats during
acute and chronic inanition were considered. The object of
the present study is to extend these observations so as to include
the volumetric analysis and histological structure of the supra-
renal cortex and medulla during normal postnatal development
and as affected by inanition and by refeeding.

This paper forms the fifth of a series of studies upon the effects
of inanition in the albino rat, the investigation being supported
by a special grant from the research funds of the Graduate School
of the University of Minnesota.

99]

am

bo
oN
iN

Cc. M. JACKSON

MATERIAL AND METHODS

The material used included the suprarenals from 108 albino
rats (Mus norvegicus albinus), obtained partly in connection
with my previous studies (15 a and 715 b), partly, from material
collected by Hoskins (’16) and Stewart (’16), and partly from new
material. The 108 rats include 52 normal (control) rats of both »
sexes, varying from new-born to about one year of age; 16 young
rats held nearly at maintenance (constant body weight) or se-
verely stunted by underfeeding for various periods; 10 rats refed
after such underfeeding and 1 refed after acute ananition (adult) ;
13 adult rats subjected to acute inanition; and 6 subjected to
chronic inanition (table 1).

The diet in all cases was whole-wheat (graham) bread soaked
in whole milk, the amount being reduced during the maintenance
and chronic inanition experiments and cut off entirely in acute
inanition. Water ad libitum was supplied in all cases. In the
chronic-inanition experiments the amount of food was gradually
reduced through a period of about five weeks. The percentage
losses and final body weights are given in tables 1 D and 1 E.

In order to keep the animals alive during the longer experi-
mental periods, especially in the case of the young rats underfed
for long periods, it was found necessary to keep them in a warm
room. ‘The cages were provided with wire-net bottoms to pre-
vent the rats from eating their feces.

The general data for the individual rats used are Sica in
table 1. In the first column the letters indicate the series to
which the rats belong. ‘The number preceding the decimal point
is the litter number and the number following designates the indi-
vidual rat. In several of the rats used for adult acute- and
chronic-inanition tests, however, the litter records were not avail-
able. The sex, age (where known), nose-anus length, gross body
weight (in some cases also net, without intestinal contents), and
the fresh weight of both suprarenals are recorded in the table.

In the subsequent tables 2, 3, and 5, the data are usually
grouped for economy of space. The individual data will be filed
in The Wistar Institute of Anatomy, Philadelphia, where they
may be consulted by those interested.

SUPRARENAL GLAND—EFFECTS OF INANITION 223

The suprarenals were removed immediately after the animals
had been killed. The rats were killed by chloroform, excepting
those used for the chromaffin tests (series F), which were
killed by cutting their throats. One rat (F 6.2) died from acute _
inanition.

In most cases the glands were fixed in Zenker’s fluid, embedded
in paraffin, cut in serial sections (3 » to 5 uw), and stained with
haematoxylin and eosin. Ina few cases other fixatives (formalin,
Zenker-formol, Flemming’s) and other stains (iron-haematoxy-
lin, Mallory’s anilin-blue connective-tissue stain, etc.) were em-
ployed. In the series F and a few others one gland was immedi-
ately placed in Miiller’s fluid for one week, then sectioned with
the freezing microtome. The sections were mounted in glycerin
for the study of the chromaffin reaction. The other gland was
first fixed one or two hours in 10 per cent aqueous formalin solu-
tion (neutralized with lithium carbonate). In a few cases the
glands were left in 5 per cent formalin from eighteen to twenty-
four hours, without apparent effect upon the liposomes. Bell
(10), however, found in some cases a partial disappearance of
the liposomes in muscle and epithelium after even a brief exposure
to formalin. The formalin-hardened gland was cut with the
freezing microtome into sections (at 10 y).

Some of these sections were placed for about twenty hours in
1 per cent aqueous osmic acid solution, and the others stained
with scarlet red and mounted in glycerin for special study of the
lipoids and fat. Herxheimer’s formula (saturated solution of
scarlet red in 70 per cent alcohol containing 2 per cent of sodium
hydroxide) was used. The sections were first placed in 60 per
cent alcohol (one minute), then in Herxheimer’s fluid (three min-
utes), then rinsed in 60 per cent alcohol (one-half minute),
washed in distilled water a few minutes, and mounted in glycerin.

Frozen sections were also studied fresh, mounted in normal
salt solution, and also treated with 1 per cent potassium hydroxide,
1 per cent acetic acid, alcohol, ete. The frozen sections were cut
by means of a Spencer automatic freezing microtome, using
liquid carbon dioxide for refrigeration.

224 Cc. M. JACKSON

The method used for the volumetric determinations is a modi-
fication of Hammar’s paper method, as described in my work on
the hypophysis (Jackson,’17). The outlines of the magnified sec-
tions are projected and drawn upon ‘American linen record’ paper
(sheets 18 x 23 inches, 36 lbs. per ream) by means of an Edinger
projection apparatus. Four samples, each 5 cm. square, are
weighed from each sheet, and the area corresponding to each gram
of paper determined. The structures whose volume is to be
determined are then cut out and weighed, and the corresponding
area calculated. This magnified area is then reduced to actual
area (dividing by the square of the magnification). Multiply-
ing by the actual thickness of the sections now gives the actual
volume of the parts concerned. For determining the volumes of
cortex and medulla (table 2) it was not considered necessary to
draw every section. Enough were taken to make a total of at
least fifty drawings for each suprarenal. This required every
fourth section for some of the smaller glands, but fewer for the
larger glands, some of the largest requiring only every sixteenth
section. A magnification of 75 diameters was used for the cortex
and medulla.

As was found in the hypophysis, the volumes obtained (in
cubic centimeters) are usually considerably less than half the
corresponding fresh weight of the gland (in grams). The differ-
ence is due: 1) to the density of the gland; 2) to the fibrous cap-
sule and attached extracapsular tissue, weighed but not measured
(forming about 11 per cent of the gland in the dog, according to
Flint’s data), and 3) to the great shrinkage due to the process of
fixation, dehydration, and embedding in paraffin.

For determining the relative volumes of parenchyma and stroma
(including blood-vessels), a higher magnification was required.
A Leitz ocular no. 4 and objective no. 7a were used, with large ©
Spencer camera lucida, giving at table level a magnification of
about 500 diameters. As it is impracticable to measure the en-
tire gland in this way, typical areas are selected and drawn from
the various regions of the cortex and medulla, outlining merely
the areas of parenchyma and vascular stroma in a given plane.
These areas are then cut out of the paper and weighed. Since

SUPRARENAL GLAND—EFFECTS OF INANITION 225

it is desired merely to determine the relative proportions of
parenchyma versus stroma, it is necessary merely to compare the
weights of the paper representing the corresponding areas. In
this way the results summarized in table 3 were obtained. These
results are of course merely approximate, based upon the as-
sumption that the areas selected represent the typical or average
condition for the entire gland.

For the volumetric determinations on parenchyma cells and
nuclei (table 4) a similar plan was used. In this case a Zeiss 2-
mm. apochromatic objective with ocular no. 4 was used, giving
at table level a magnification of about 1600 diameters. Typical
fields were selected and drawn with the camera lucida, outlining
parenchyma cells and nuclei (stroma and vessels omitted) in the
various zones of the cortex and medulla. No change of focus is
permissible while the drawing is being made, since the volumetric
calculations are based upon the assumption that it represents a
true optical plane. ;

The further procedure is as follows: The paper representing
the areas drawn from the various zones is weighed (to the milli-
gram). Then the nuclei are cut out with a sharp-pointed scalpel,
the paper resting upon a wax plate. The paper representing the
nuclear areas and the remainder (cytoplasm) are then weighed
separately. From the weight of the nuclear areas, the magnified
area in square centimeters is calculated. This divided by the
square of the magnification and by the number of nuclei involved
gives the average actual area per nucleus. If we assume that the
nuclei are spherical and not greatly different in size, the average
area per nucleus in a given optical plane should, according to the
rules of solid geometry, represent two-thirds the area of the cor-
responding great circle.

This result is derived as follows: The volume of any sphere
equals two-thirds the volume of the circumscribed cylinder. The
diameter of the sphere equals the height of this circumscribed
cylinder, and the area of the great circle of the sphere equals the
cross-sectional area of the cylinder. The volume of the cylinder
is obtained by the product of its cross-sectional area x height of
cylinder. The volume of the sphere may be obtained by the

226 - @. M. JACKSON

product of the average cross-sectional area (i.e., the average of a
series of parallel planes, perpendicular to any diameter) of the
sphere x the diameter of the sphere. Since the height of the
cylinder equals the diameter of the sphere, the average cross-sec-
tional area of the sphere must equal two-thirds of the cross-sec-
tional area of the cylinder (or of the great circle area of the
sphere), since the corresponding products (the volumes of the
sphere and the circumscribed cylinder) bear that ratio.

Hence in an optical plane, as seen in a histological section, as-
suming the nuclei to be approximately spherical and similar in
size, the observed average nuclear area should correspond approx-
imately to two-thirds the corresponding great circle area of the
nucleus. From the great circle area (= 7z r*), the average nu-
clear diameter (= 2 r) and volume (= 4/3 7 r') are readily
calculated.

The average nuclear dimensions being known, the average
total cell dimensions may also be calculated, assuming that the
cytoplasm surrounds the nucleus in a uniform layer, making the
whole cell approximately spherical. “

Another relation, which may not appear evident at first glance,
is that in the optical plane of any section the sum total of the
nuclear area bears the same ratio to the sum total of the cell
areas as does the volume of the average nucleus to the volume of
the corresponding average cell. This relation is proved as fol-
lows, assuming the cells to be fairly uniform: The volume of any
given region is made up of its parenchyma cells (stroma being
excluded). The ratio of this volume, the total cell volume, to the
total nuclear volume is the same as that of the average cell volume
to the average nuclear volume (since the average volume is de-
rived from the total by dividing by the total number of cells (or
nuclei) in the region). But the total cell volume and total nu-
clear volume for a region may be derived in another way. The
region may be considered as made up of parallel sections of mini-
mal thickness, as found in histological preparations. In each
section, the total cell volume is equal to the total area of the
region measured (stroma excluded) x the thickness of the sec-
tion; and the total nuclear volume likewise corresponds to the

SUPRARENAL GLAND—EFFECTS OF INANITION 227

sum total of the nuclear areas x the thickness of the section.
The total cell and nuclear volumes for the region would be ob-
tained by multiplying by the total number of sections. Thus
it is evident that the ratio of total cell volume to total nuclear
volume of a region (or of average cell volume to average nuclear
volume) is the same as the ratio of the total area to the total nu-
clear area in a typical optical plane. The total area of a section
is represented by the total paper weight for that section and the
total nuclear area by the paper weight of the nuclear areas.
Moreover, the ratio of the average cell diameter to the correspond-
ing nuclear diameter is the same as the ratio of the cube roots of
the average cell volume and nuclear volume. Therefore, for a
given region, assuming that the cells are fairly uniform spheres,
the average

*! Total paper weight —
weight of nuclear areas
From the cell diameter, the cell volume is readily obtained, and

the volume of cytoplasm in this average cell will be the difference

between the cell volume and the nuclear volume. The results
are given in table 4. Errors in computation! may be checked by
comparing the ratio of nuclear volume and cell volume with the
corresponding ratio of the cube of the nuclear diameter and the
cube of the cell diameter. Crelle’s Rechentafeln were found

“helpful in the calculations.

The mitoses were counted in each case in an entire section
through the middle of the gland. A 2-mm. Zeiss apochromatic
objective and a mechanical stage were used. ‘The results are
summarized in table 5.

Cell diameter = Nuclear diameter x

1In my previous paper on the hypophysis (Jackson, 717) an unfortunate
error was made in the calculation of the parenchyma cell diameters listed in the
last column of table 3, p. 333. Instead of 10.1, 11.9, 13.6, 9.7, 10.2, 10.0, 10.8,
and 11.0, this column should read: 8.4, 9.3, 10.4, 7.8, 8.0, 8.1, 8.6, and 8.7. A
corresponding correction is necessary for cell diameters on the lower part of
page 335, and in the Summary, paragraph 7, page 355, of the same paper. A
similar correction for the diameter of the parenchyma cells should be made in
the abstract of that paper, published in the Anatomical Record, vol. 11, p. 370,
lines 5 to 8.

228 Cc. M. JACKSON

VOLUMES OF SUPRARENAL CORTEX AND MEDULLA

Volumetric data for the relative volumes of the suprarenal cor-
tex and medulla are given in table 2. These will be considered
successively, 1) during normal postnatal development; 2) in
young rats stunted by underfeeding; 3) in such rats refed after
stunting, and 4) in adult rats subjected to acute or chronic
inanition.-

1. During normal postnatal development

The data summarized in table 2 A include the relative volumes
of cortex and medulla in groups at various ages, beginning at
seven days. (In the new-born the cortex and medulla are still
intermingled and were not measured.) If the data for the abso-
lute volumes (not given in the table) are considered, it appears
that both cortex and medulla continue to grow from birth to one
year of age. Although there are marked individual variations,
it is on the whole clearly evident that the cortex and medulla
grow at different rates, so that their relative sizes change. Dur-
ing the second and third weeks, the cortex grows more rapidly,
increasing from 75 or 80 per cent to about 90 per cent of the en-
tire gland, by volume. It apparently continues to increase rela-
tively, averaging about 93 per cent at ten weeks of age. In
older rats, there are marked individual variations, but the cortex -
appears to decrease slightly to 90 or 91 per cent in average relative
volume.

The medulla increases more slowly in absolute size, so that its
relative volume decreases (inversely with the cortex). The me-
dulla is relatively small in the new-born, though difficult to meas-
ure on account of the intermingled cortex. In connection with
the process of confluence, the medulla expands rapidly, being
relatively largest at one week, forming 20 to 25 per cent of the
total volume. It decreases to about 15 per cent at two weeks, 10
per cent at three weeks, and to about 7 per cent at ten weeks, in-
creasing again slightly to an average of 9 or 10 per cent of the
entire suprarenal in the older rats. In a few cases in which the
whole suprarenal gland is relatively small, the medulla was found

SUPRARENAL GLAND—EFFECTS OF INANITION 229

relatively large (as in the female St 5.1 at 56 days). In very large
glands, on the contrary, the medulla is usually relatively small.
This suggests that the variability in the size of the suprarenal may
be due mostly to variability in the volume of the cortex.

On account of the known sexual difference in the weight of the
suprarenals (which become relatively heavier in the female from
the age of about six weeks),? the question naturally arises concern-
ing the relative sizes of cortex and medulla in the two sexes. As
shown by table 2 A, however, the present data grouped by sexes
reveal no constant or important differences in this respect. A
study of the individual data likewise reveals no constant sexual
differences in the absolute or relative sizes of the suprarenal cor-
tex and medulla. It is possible that a larger number of obser-
vations would reveal differences obscured by large individual
variations (or experimental error) in the present small series of
data. Provisionally, however, on the basis of the present data, I
would conclude that the relatively larger suprarenal gland in the
female is due to an increase in both cortex and medulla, with no
important change in their relative sizes as compared with those
of the smaller gland in the male. Thus the conditions in the
suprarenal would appear to be different from those in the hypo-
physis, where the relatively larger gland in the female is appar-
ently due to a larger pars anterior (Jackson, 717).

In one female (J 1.7) which had just given birth to a litter
the suprarenal glands were not hypertrophied (table 1), being on
the contrary below the normal weight for corresponding body
weight or length. The volumetric relations of cortex and medulla
appear normal, the latter forming 7.2 per cent of the total vol-

2 Elliott and Tuckett (’06) noted in several cases that in the rat, cat, guinea-
pig, and rabbit of reproductive age the suprarenal glands were larger in the
female. They considered this a temporary hypertrophy, however, concluding
that ‘“Gestation accelerates the growth of the gland, so that for a time that of
the female outstrips the male. In such increase the cortex is concerned, but
there is undoubtedly growth of the medulla also. Our analyses are too few to
trace the changes satisfactorily.’? That the larger suprarenal in the female rat
is a regular sexual characteristic, independent of gestation, was first established
by Hatai (13) and Jackson (13).

230 Cc. M. JACKSON

ume. Hypertrophy of the suprarenal gland (especially of the
cortex) during pregnancy has been described by various authors.

Elliott and Tuckett (06) found great variation in the relative
volumes of cortex and medulla in the suprarenal gland of adult
animals. Thus the medulla varies from 1.6 per cent of the gland
in the guinea-pig to 50 per cent in the fowl. In the adult rat the
medulla forms about 5 per cent of the gland. In very young
mammals the medulla is relatively larger, the growth of the su-
prarenal after early youth being apparently in the cortex alone.
They also noted a relatively larger suprarenal in the female,
ascribing this to pregnancy.*

Bager (’17) found the suprarenal glands of the rabbit relatively
heavier in the female after the age of puberty.

A relative increase in the amount of cortex, with corresponding
decrease in the relative size of the medulla in the suprarenal gland
according to age, was previously noted by Canalis (’87) in the
rabbit, and by Hultgren and Anderson (’99) in various mam-
mals, especially cat and rabbit. Soulié (’03), however, describes
the suprarenal medulla as relatively small in the human fetus,
forming hardly 10 per cent in the new-born, the proportion be-
ing a little higher in the adult and relatively greater in other
mammals. Scheel (’08), Starkel and Wegrzynowski (10),
Thomas (’11), and others describe a relative increase in the post-
natal human medulla (which will be considered later under
‘Morphogenesis of the suprarenal gland’). Gottschau (’83), by
comparison of the relative widths of cortex and medulla in sec-
tions of the suprarenal, concluded that there is great variability
in different mammals. Pfaundler (’92) found the ratio of cortex
to medulla in the suprarenal of the horse to be variable, the aver-
age being about 4:1, which is nearly in agreement with the data
of Flint (’00) for the dog. In the rabbit, Bager (’17) found the
medulla to decrease in relative size from 20.8 per cent of the
gland in the new-born to 1.84 per cent at one year. No differ-
ence was noted according to sex in the ratio of cortex and medulla.

On the whole, therefore, the evidence would appear to support
the conclusion that in general the suprarenal medulla in mammals

3 See footnote on page 229,

SUPRARENAL GLAND—EFFECTS OF INANITION 231

forms but a small part of the gland. It usually appears rela-
tively large in the new-born, however, decreasing later in relative
size on account of more rapid growth in the cortex. In the rat,
although the suprarenal becomes relatively larger in the female,
there is apparently no constant sexual difference in the relative
size of cortex and medulla. The whole question of the changes
in relative size of cortex and medulla according to age and sex in
various species deserves a more careful and extensive study.

2. In young rats stunted by underfeeding

As previously shown (Jackson, 715 b), in young-rats held at
maintenance (nearly constant body weight) by underfeeding for
various periods there is a definite tendency to increase in the
weight of the suprarenal gland, especially in the female. The
observations in table 2 B indicate no marked change in the rela-

tive volumes of the cortex and medulla as a result of such experi-.

ments, however. While during normal growth up to ten weeks
the suprarenal medulla tends to lag behind (cortex becoming
relatively larger), in the maintenance experiments it apparently
remains nearly constant in relative size. Lewis and Pappen-
heimer (716) likewise found no constant change in the relative
size of cortex and medulla in the suprarenal glands of children
emaciated through malnutrition.

3. In rats refed after stunting

As shown by table 2 B, in rats refed one or two weeks (after
being held at maintenance from three to twelve weeks of age),
the relative sizes of the suprarenal cortex and medulla appear
nearly normal. Ina permanently stunted female (St 33.120, re-
fed to one year after maintenance from three to twenty-one weeks
of age), the medulla appears relatively large, while in a male
(S 33.118) similarly refed the ratio appears nearly normal. Stew-
art (16) found that in rats refed four weeks or less, after main-
tenance from three to twelve weeks of age, the suprarenals
seem to lag behind somewhat in weight, although they appeared
normal in such rats refed to the age of one year.

Zoe Cc. M. JACKSON

4. In adult rats during acute or chronic inanition

In adult rats subjected to acute inanition seven to nine days,
with loss of about one-third in body weight, the relative volumes
of cortex and medulla remain nearly normal (table 2 B), indicat-
ing that the loss (if any) has been nearly proportionately distrib-
uted. In rats after chronic inanition due to gradual decrease
of food for five weeks, with loss of about 36 per cent in body
weight, the medulla averages a little higher. This may indicate
a relatively greater loss of the cortex, but the difference is small
and of doubtful significance.

VOLUMES OF SUPRARENAL PARENCHYMA AND STROMA

The averages (and ranges) of the percentages by volume formed
by the stroma, including the vessels, are summarized by groups
in table 3. The variations are due chiefly to the differences in
blood content. On account of the great variability shown, no
conclusions can be drawn with certainty. Neither the groups
nor the individual data reveal any marked change in the normal
vascularity of the gland from new-born to adult. There is in
general, however, an increased vascularity evident upon passing
from the outer zone of the cortex toward the medulla, with the
exception of a decrease in the outer part of the middle zone.
More than one-fourth of the medulla by volume is composed of
stroma and included vessels.

In the young rats held at maintenance the vascular stroma ap-
pears moderately decreased in relative volume in the outer and
middle zones and greatly increased in the inner zone, with but
little change in the medulla. The hyperemia of the inner zone
is usually strikingly apparent in the sections under the micro-
scope (low power).

In the rats refed one or two weeks after maintenance experi-
ments, there appears in general a reduction in the relative volume
of the vascular stroma in the most hyperemic regions—medulla
and inner cortical zones—and an increase in the vascular stroma
of the middle zone. The outer cortical zone, however, appears
even more anemic than before.

SUPRARENAL GLAND—EFFECTS OF INANITION 233

In adult rats subjected to either acute or chronic inanition,
there appears to be an increased vascularity of the middle and
inner cortical zones, with a decreased vascularity in the medulla.
Marked hyperemia of the suprarenal cortex has also been noted
as a result of inanition in the guinea-pig (Martinotti, ’92; Ron-
doni and Montagnani, ’15).

The most constant effect of inanition upon the suprarenal
stroma is therefore an increased hyperemia of the inner cortical
zone. The changes in the vaseularity of the remainder of the
gland appear variable in different individuals and according to
the type of inanition. The variations are due chiefly to the
amount of blood remaining in the vessels, which of course may
vary greatly under different circumstances. Moreover, the
amount retained in the vessels after death may be quite different
from that during life. In pathological histology, however, it is
customary to judge as to relative anemia or hyperemia by the
distention of the vessels in prepared sections of organs. So an
attempt to measure the amount appears justified, even though
the results are only roughly approximate.

SIZE OF PARENCHYMA CELLS AND NUCLEI

The average sizes of the parenchyma cells and nuclei were not
determined by direct measurement, but were calculated by the
method previously explained. The data are given in table 4.
The cell and nuclear dimensions will be considered, 1) in normal
postnatal development; 2) in young rats stunted by underfeeding;
3) in such rats refed after stunting, and 4) in adult rats with acute
or chronic inanition.

1. During normal postnatal development

As noted by Canalis (’87) in the rabbit, there is in the rat a
general increase in the size of the cells from birth to maturity
(table 4). In average diameter, the cells of the outer cortical
zone (zona glomerulosa) increase gradually from about 7 uw at
birth to about 8 yw at three weeks and to about 9 uw from ten
weeks on.

234 C..M. JACKSON

The middle zone (zona fasciculata) cells during the first week
average about 9 » in diameter, being of approximately the same
size throughout the zone. In the outer portion of this zone they
increase to about 9.6 » at two and three weeks, increasing to 14 or
15 » from ten weeks on. The cells in the inner portion of the
middle zone appear slightly larger (10.3-10.7 u) at two and three
weeks, but lag behind thereafter, reaching about 12 » in diameter
from ten weeks on. .

The inner zone (zona reticularis) cells during the first three
weeks are similar in size to those in the adjacent outer part of
the middle zone. Later, however, they lag behind, and at ten
weeks are about equal in average diameter to those of the outer
zone (8-10 u). In the adult rat, they are in average size the
smallest cells of the suprarenal cortex. They do not vary greatly
from 9. in average diameter throughout postnatal life. The
small size of the cells is associated with the atrophic condition of
the inner zone, which undergoes continual absorption (as will
discussed later).

The cells of the suprarenal medulla increase steadily from an
average diameter of about 8 » at birth to 10 » or 11 » during the
second and third weeks. At ten weeks, they average about 13 yu,
and in the adult reach an average diameter of nearly 16 uz.

The nuclear growth of the suprarenal parenchyma cells is in
general less rapid than that of the cytoplasm, so that with the in-
creasing volume of the cell the nucleus decreases in relative size,
forming a steadily decreasing percentage of the average cell vol-
ume (table 4). The outer zone nuclei apparently decrease in
average diameter from 5.4 » at birth to 4.6 » at one week. They
increase to about 5.5 4 at three weeks, thereafter remaining
nearly constant in absolute size. In relative volume, however,
they decrease from 44 per cent of the average cell volume in the
new-born to about 23 per cent in the adult rat.

The middle-zone nuclei likewise appear to decrease slightly dur-
ing the first two weeks (from average diameter of 6.2 u to 5.6 y),
but increase to 6.2 uw at three weeks. Thereafter the nuclei of
the inner part of the zone (like the corresponding cells) increase
but little in average diameter, while those of the outer part of

SUPRARENAL GLAND—EFFECTS OF INANITION Daw

the middle zone increase to an average diameter of about 7u.
In relative volume, the nuclei of the outer part of the zone de-
crease from about 33 per cent of the average cell volume at birth
to about 10 per cent in the adult, while the nuclei of the inner
part of the zone decrease from about 28 per cent to about 15 per
cent in the adult.

The nuclei of the inner cortical zone apparently decrease in
average diameter from 6.1 u in the new-born to from 5 to 5.8 u
thereafter. In relative volume, they decrease from about 28
per cent of the average cell volume at birth to about 20 per cent
thereafter.

The nuclei of the medulla parenchyma cells increase in average
diameter from about 6 uw at birth to about 6.7 u at three weeks,
7.2 w at ten weeks, and 7.3 uw in theadult. In relative volume,
they decrease from about 46 per cent of the average cell volume in
the new-born to less than 10 per cent in the adult.

Da Costa (713), in the suprarenal gland of the rat (Mus decu-
manus), gives the following measurements for the nuclei: outer
cortical zone, 4 xX 6 uw; middle zone, spherical, 6 «; inner zone,
spherical, smaller. My data (previously cited) indicate a some-
what larger average size for the‘nuclei of the middle zone.

Canalis (’87) noted that, in general, the suprarenal parenchyma
cells of both cortex and medulla are larger in the adult than in the
young (rabbit, dog, guinea-pig). In the mouse Inaba (91)
found the nuclei of medulla cells increasing in diameter from
5.2 w in the new-born to 6 w at twenty-nine days and later. The
cortical nuclei (near the corticomedullary boundary) showed a
decrease in diameter from an average of 6.5 » in the new-born to
5 » at twenty-nine days and later. This would indicate that
there is an atrophic process of the inner cortical zone in the mouse
corresponding to that found in the rat and other forms (to be
discussed later).

Mann (16) noted a seasonal change in the suprarenal cells
(both cortex and medulla) of the gopher, which increase in size
in the spring, with decrease later and during hibernation.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 3

236 Cc. M. JACKSON

2. In young rats stunted by underfeeding

Table 4 includes the individual results in four young rats held
at maintenance from three weeks to ten or twelve weeks, and
one in which the experiment continued to 139 days. Since in the
latter case the cells and nuclei are similar in size to those of the
other cases, the average is also given for the entire group. In the
following discussion the data are compared with those of normal
rats at the same body weight (that is, at three weeks, when the
experiment began).

The cells of the outer zone have in some cases apparently in-
creased in average diameter, but the average for the group is but
slightly increased, if at all. The cells of the outer part of the middle
zone show an increased average diameter in all cases, the average
for the group being 11.1 uw. This is greater than the normal at
three weeks (9.7 «), though less than the normal at ten weeks
(14.5 wu). The cells of the inner part of the middle zone, on the
contrary, shows a definite decrease to 8.9 u (from normal 10.5 1).
The inner-zone cells likewise decrease in average diameter to 8.1 yu
(from 9.5 uw), but the cells of the medulla show but little if any
change in size.

In nuclear size the outer zone in the suprarenal cortex of the
maintenance rats is but slightly changed. The average nuclear
diameter is slightly lowered, with a corresponding decrease in the
relative volume of the nucleus to 25.7 per cent (from normal 34.5
per cent). The nuclei in the outer part of the middle zone re-
main nearly unchanged in average diameter, but on account of
the increase in cytoplasm and cell volume the relative size of the
nucleus decreases to about 17 per cent (from normal 27 per cent)
of the cell volume. The nuclei in the inner part of the middle
zone show a definite decrease in diameter to 5.5 uw (from normal
6.2 uw), but as the reduction in cell diameter (cytoplasm) is still
greater, the relative volume of the nucleus averages slightly
higher at 23 per cent of the cell volume (normal 21 per cent).
The inner cortical zone nuclei decrease in average diameter to
4.9 uw (from 5.7 uw), but since the cytoplasm decreases similarly,
the relative volume of the nucleus is unchanged. The nuclei

SUPRARENAL GLAND—EFFECTS OF INANITION Zal

of the medulla parenchyma cells show a slight decrease in average
diameter to 6.4 uw (normal 6.7 uw) and their relative volume de-
creases to 17.2 per cent (normal 23.5 per cent).

3. In young rats refed after stunting

Table 4 shows the results in four individuals, two refed one
week and two refed two weeks after being held at maintenance
(nearly constant body weight). In general, the dimensions are
slightly greater in those refed two weeks, but as the differences
are usually slight, the average is given for the whole group.
Comparison is made with the data for the maintenance group
and for the normal.

In cell diameter the cells of the outer cortical zone have in-
creased to about normal size (from 8.1 » at maintenance to 8.7 y).
The cells of the outer part of the middle zone have increased
materially ( rom 11.1 » maintenance to 11.8 » after one week of
refeeding and to 13.1 uw after two weeks), but are still below the
normal. The inner-zone cells have increased to about normal
size (9.2 u). The medulla cells have increased in diameter (from
11.4 at maintenance to 12.2 uv), though still below normal
average (13.2 yu).

In nuclear size the outer-zone nuclei remain unchanged in di-
ameter, though decreasing slightly in relative volume (percentage
of cell volume). The middle-zone nuclei have not changed much
in diameter, but on account of the increase in cell volume their
relative volume has decreased to nearly normal ratio. The in-
ner-zone nuclei have increased somewhat in diameter, so the
nucleus-plasma ratio is now nearly normal. In the medulla the
nuclei have apparently decreased in average diameter (to 6.1 u
as compared with 6.4 uw at maintenance), and this reduces the
nuclear percentage of the cell volume from 17.2 per cent to 12.9
per cent, which is considerably below the normal (about 16 per
cent). This nuclear change is difficult to understand.

238 Cc. M. JACKSON

4. In adult acute inanition

As shown by table 4, a comparison of the average data from the
four adults with those of the normal adult indicates that the
cells of the outer cortical zone have decreased little or none in
diameter. The middle-zone cells, however, have decreased in di-
ameter very markedly—from 14.8 » (normal) to 12 yin the outer
part of the zone, and from about 11.5 » to 9.2 » in the inner part
of the zone. The inner zone cells have decreased but slightly—
from 8.7 uw to 8.5 u. The medulla parenchyma cells have de-
creased markedly—from 15.9 u to 12.7 » in average diameter.

The nuclei of the outer zone indicate no shrinkage, but rather
an insignificant increase in average diameter from 5.5 uv to 5.8 yu,
with a corresponding slight increase in relative volume (from
23.2 per cent to 27.1 per cent of the average cell volume). The
middle-zone nuclei have decreased slightly in diameter (6.9 yu
to 6.5 w in the outer part and 6.4 » to 5.5 w in the inner part
of the zone). But since the loss in cytoplasm is greater, the
relative nuclear volume has increased—from 10.1 per cent to
16.2 per cent in the outer part and from 15.5 per cent to 21.7
per cent the inner part of the zone. The inner-zone nuclei
remain nearly constant in size and relative volume. The
medulla nuclei appear slightly decreased in average diameter
(7.3 » to 7.0 uw). But on account of greater loss in the cyto-
plasm, the relative size of the nuclei has increased from 9.6
per cent to 16.8 per cent of the cell volume.

Barbéra and Bicci (’00), in rabbits starved seven days with
‘oss o° 34 per cent in body weight, found the nuclei in the su-
prarenal cortex to decrease from an average diameter of 7.238 yu
to 4.817 y, and in the suprarenal medulla from 7.773 y» to 5.587 yu.
The loss in cytoplasm appeared relatively greater than in the
nucleus. Similar results were obtained in dogs. These de-
creases in cell dimensions would appear to be much greater
than those observed by me in the rat. Traina (’04) noted a
decrease in the size of cortical cells in the suprarenal of the
rabbit during starvation, but gives no exact data.

SUPRARENAL GLAND—EFFECTS OF INANITION 239

5. In adult chronic tnanition

The data from the two individuals (table 4) indicate that the
suprarenal cells of the outer and middle cortical zones have
lost in diameter somewhat more than during acute inanition;
the inner zone and medulla cells about the same as during acute
inanition (or slightly less).

The nuclei in all the regions during chronic inanition have
apparently lost in average diameter somewhat more than dur-
ing acute inanition. The relative (percentage) volume of the
nucleus is therefore usually lower than during acute inanition.
In the outer zone and the medulla, the nucleus-plasma ratio
is not greatly different from that of the normal animal.

CELL DIVISION—FREQUENCY OF MITOSIS

No conclusive evidence of amitosis was found in any of the
suprarenal glands studied, although irregular or bilobed nuclei,
which might easily be so interpreted, were occasionally observed
in the various zones of the cortex and medulla. If amitosis
occurs at all (in the rat), it is probably a degenerative phenom-
enon of no especial significance or importance in the normal
development of the gland.

The frequency of mitosis in the suprarenal, under normal and
abnormal conditions at various ages, is given in table 5. The
average number of mitoses counted in entire sections through
the center of the gland is shown for the various zones in the
rats grouped according to age or condition.

1. Mitosis during normal development

As shown in table 5 A, the average number of mitoses in an
entire section of the suprarenal is somewhat variable in in-
dividual cases, but the general trend according to age is evident.
The average number per section at birth is 20, decreasing to
an average of 10 at one week, and increasing again during the
second and third weeks. At twelve days (one case only) there
were 40 mitoses per section, which was the maximum number

240 Cc. M. JACKSON

observed, excepting one of the nine cases at three weeks. The
average number of mitoses again decreases somewhat slowly
from 27 at 14 days to 24 at 21 days, 11 at 56 days, 4 at 67 to
94 days, 2 at 112 to 138 days, and 1 at 340 days. In other
adults no mitoses were observed in the suprarenal.

The relative rate of mitosis (‘mitotic index’ of Minot) de-
creases more rapidly than is indicated by the figures for the
absolute frequency per section shown in table 5A. This is
because the size of the gland, and consequently the total number
of cells in a section, is progressively increasing. Thus between
birth and three weeks of age the suprarenal gland of the albino
rat has increased about four times in weight (and volume).
This corresponds to an increase in diameter of about 1.6, or
of about 2.6 times in the area of a cross-section. Since the
size of the cells has not greatly increased during this time (table
4), the number of cells in a corresponding section of the gland
is probably twice as great at three weeks as in the new-born.
Thus the relative rate of mitosis would be obtained by com-
paring the number of mitoses in a section at birth (20) with
half the number in a section at three weeks (half of 24 = 12).
Therefore, although the absolute number of mitoses per section
increases in the second and third weeks, it is doubtful whether
the relative rate of mitosis exceeds that of the new-born.

The distribution of the mitoses in the various zones of the
cortex and the medulla is also shown in table 5A. The rate
in the medulla is in general nearly parallel with that for the
whole gland, decreasing from an average of 6 per section in the
new-born to 3 at one week, and gradually increasing again to
6 at fourteen days. Thereafter the number of mitoses decreases,
and none are found in the medulla after eight weeks, though
continuing to occur in the cortex. The later growth of the
medulla must therefore be due entirely to the growth in the
preéxisting cells, which continue to increase in average size
(table 4).

As to localization in the various regions of the cortex, the
number of mitoses in the outer zone (glomerulosa) is strikingly
large, although this zone is the narrowest of the three. When

SUPRARENAL GLAND—EFFECTS OF INANITION 241

it is further considered that the majority of the middle-zone
mitoses occur in the outer part of the zone, it is evident that
in general the outer portion of the cortex is the chief cortical
germinative region. Some mitoses are found scattered through-
out the cortex, however, an occasional one being found even in the
inner zone (reticularis). In general, mitoses disappear first
in the medulla, and the inner cortical zone, and later in the
middle zone, persisting longest in the outer cortical zone.

The occurrence of amitosis in the suprarenal has been: {aimed
by various investigators. Mulon (03 b, ’05a) in particular
even maintains that in the guinea-pig amitosis is the predom-
inant type of cell division in the suprarenal, mitosis being much
less frequent. Kolmer (12 b), on the contrary, found mitoses
more frequent (occasional amitosis) in the suprarenal cortex
of the younger guinea-pigs, though in older, non-pregnant fe-
males amitosis was observed with no mitosis. Most authors
have considered amitosis in the suprarenal (as elsewhere) to
be atypical and degenerative in character. Thus Poll (’99)
observed amitosis (more rarely mitosis) in the degenerating
cortical cells in transplanted suprarenal glands of the rat. Del-
amare (’03) observed occasional amitosis in the senile human
suprarenal cortex, producing multinucleated masses. Bonna-
mour (’05 a) believes that the figures which have been inter-
preted as amitosis (direct division), especially in the outer cor-
tical zone and in the medulla, are merely deformed nuclei, without
significance for cell division.

In relation to age, cell division (especially mitosis) in the
suprarenal has generally been found more frequent in young
animals, as might be expected, and especially in fetal stages
(Canalis). It has also been observed in adults, especially in
the outer cortical region, by Canalis (’87) in the dog, guinea-
pig, rabbit, and mule, Mulon (’03 b) (guinea-pig), Kolmer (’12 b)
(guinea-pig), and others. Da Costa(’13) has observed mitoses
in the cortex of both young and adult dog and guinea-pig.

It is interesting to note that the rate of mitosis in the young
rat corresponds rather closely to the growth of the gland in
absolute and relative weight. As previously shown (Jackson,

242 Cc. M. JACKSON

’13), the suprarenal gland in the albino rat increases but slightly
in average absolute weight (from 0.0019 gram to 0.0023 gram)
during the first week after birth, thereby decreasing in relative
weight (percentage of the entire body). This is in agreement
with the sma ler size of the cortical cells (already mentioned)
and the corresponding marked decrease in the rate of mitosis
observed in the present study (table 5). The weight of the
gland increases rapidly during the second and third weeks, reach
ing its maximun relative size corresponding to the period when
an increase in the number of mitoses appears. The formula
for the growth of the suprarenal in the albino rat as derived
by Hatai (13) similarly indicates a maximum relative weight
for suprarenal at about 10 to 15 grams body weight. His data,
however, do not show the arrested growth during the first week.
The suprarenal weights in Donaldson’s (715) tables (which are
based upon Hatai’s formulas) are therefore apparently too high
for this early period. According to these tables, at a body
weight of 10 grams, the suprarenals should weigh about 0.0045
gram, whereas in fifty-seven cases (30 m., 27 f.) I found an aver-
age of only about 0.0023 gram.

As to regional distribution, cell division has in general been
found to occur throughout the gland in the earlier stages, espe-
cially prenatal (Canalis, ’87, in dog, guinea-pig, and rabbit),
with a progressive tendency to be restricted later to the outer
cortical region (Canalis, ’87; Mulon, ’03 b, and others). The
occurrence of mitosis in the deeper layers of the suprarenal
cortex (occasionally even in the zona reticularis) in growing
or adult animals has also been observed by Bonnamour (’05 a),
Kolmer (’12 b), and Da Costa (’13) (dog, rat, rabbit, guinea-pig).
An extraordinary increase in mitosis in the outer part of the
zona fasciculata was noted by Kolmer (12 a) at the end of preg-
nancy in the guinea-pig. That the outer cortical region of the
suprarenal is the chief ‘germinative zone’ from which by cell
division the remainder of the cortex is derived (especially in
postnatal development), is indicated by the observations of
Canalis (’87), Roud (’02), Cristiani (’02), Soulié (03), Mulon
(03 b, 05 a), Kern (11), Kolmer (12b), Cottentot, Mulon

SUPRARENAL GLAND—EFFECTS OF INANITION . 243

and Zimmerman (’12), and others. Gottschau (’83) considered
the germinative zone to lie between the zona glomerulosa and
the zona fasciculata, while Wiesel (’01) (in the fetal pig) and
Da Costa (13) placed it in the middle zone (fasciculata), and
Dsershinsay (cited by Da Costa, 713) even in the inner zone.
Graham (’16) found that after toxic lesions regeneration of the
suprarenal cortex proceeds from the zona glomerulosa and outer
part of thefasciculata. In the child, mitoses may occur through-
out the cortex.

2. Mitosis during inanition and refeeding

From the data given in table 5 B it is evident that the effect
of underfeeding in young rats held at maintenance (nearly
constant body weight) is in general to diminish very greatly
the number of mitoses, and ultimately to prevent them entirely.
In rat St 80.9 underfed from birth to twelve days (body weight
increased from about 5 grams to 8.9 grams), the total number
of mitoses found in a section of the suprarenal was 4, whereas
normally (at corresponding age or body weight) it would be
between 10 and 20. In ten rats held at maintenance from three
to ten weeks of age, the average number of mitoses was reduced
- to 1, the maximun being 3, and none present in six of the ten
cases. In three cases in which the maintenance was continued
to the age of twelve to twenty weeks, no mitoses were found
in the suprarenal.

The distribution of the mitoses in the various cortical zones
of the suprarenal in the stunted rats appears similar to that
in the normal glands. That is, the larger number occurs in
the outer zone and the outer part of the middle zone. None
was noted in the inner zone and (excepting the youngest rat)
none in the medulla.

In the rats refed fully for one week (after maintenance from
three to twelve weeks) mitosis in the suprarenal has recom-
menced. ‘The rate is still below normal, however, the average
number being only 2 per section. In rats refed two weeks,
the number of mitoses has increased to 10 per section, which

244 Cc. M. JACKSON

is nearly the same as that found in the normal rat at fifty-six
days. Since the body weight (77 to 79 grams) and suprarenal
weight are also nearly the same as in the normal at fifty-six
days, we may conclude that the normal rate of mitosis has been
reéstablished in the suprarenal glands of stunted rats after two
weeks of refeeding. As might be expected, no mitoses were
found in the rat refed one year, which had reached maturity.

Martinotti (’92) briefly states that in guinea-pigs starved
three or four days an abnormal increase in the number of mitoses
(20 to 25) in the suprarenal cortex was always noted. An in-
crease in mitosis was also noted upon refeeding after several
days of inanition. Unfortunately, further details are omitted,
including the age and number of animals used. Bonnamour
(05 b), on the contrary, found no mitoses in the suprarenal
glands of the rat, guinea-pig, rabbit, and cat starved to death.
Rondoni and Montagnani (’15) likewise observed no mitosis
in the suprarenals in starved guinea-pigs. This would agree
with my results on the albino rat.

HISTOLOGICAL CHANGES IN THE SUPRARENAL GLAND

In addition to the volumetric data previously given, the gen-
eral changes in histological structure were observed. These
include: 1) normal postnatal histogenesis in the albino rat; 2)
changes in young rats stunted by underfeeding; 3) changes in
rats refed after stunting by underfeeding, and 4) changes in
adult rats after acute or chronic inanition.

1. Normal postnatal histogenesis

A brief account of the histology and histogenesis of the supra-
renal will be given, with special emphasis upon the structural
features of importance in connection with the changes pro-
duced by inanition.

New-born and first week. In sections fixed in Zenker’s
fluid and stained with hematoxylin-eosin, the three cortical
zones of the suprarenal in the new-born rat are fairly distinct.
Under low magnification they appear as, 1) an outer zone

SUPRARENAL GLAND—EFFECTS OF INANITION 245

(glomerulosa), narrow and deeply stained; 2) a middle zone (zona
fasciculata), broader, with irregular, radial columns, and 3) an
inner zone (zona reticularis), less distinct an more irregular,
and (in the new-born) intermingled with the medulla. The
medulla, which at birth is not clearly differentiated and sepa-
rated from the cortex, is more vascular. It is composed of
irregular cell groups or cell cords and contains a few scattered,
deeply stained cell masses. Similar masses also appear occa-
sionally in the cortex, especially in the region of the hilus.

Under higher magnification, the outer cortical zone appears
narrow, but variable in width, about six to ten cells deep. The
cells are somewhat variable in size and structure. The cyto-
plasm is scanty, more or less granular, and presents a few fine
vacuoles (lipoids.) The nuclei are spherical or ellipsoidal in form,
somewhat deeply chromatic and (as previously shown) present
frequent mitoses.

The middle-zone cells are larger. The cytoplasm is more
abundant, with distinct eosinophile granules and a few (vari-
able number) of lipoidal vacuoles, somewhat more prominent
in cells toward the middle of the zone. The nuclei are spheri-
cal and larger than those of the outer zone. They are less deeply
chromatic, through occasionally pycnotic.

The inner-zone cells resemble those of the adjacent part of
the middle zone, with distinctly granular cytoplasm and a few
fine vacuoles. Many of the cells are degenerative in appear-
ance, with indistinct cytoplasm and karyolytic or pycnotic
nuclei, sometimes fragmented (karyorrhexis). The inner zone
cells of the cortex extend into the medulla, and are intermingled
closely with groups of the medulla cells, from which they are
frequently difficult to distinguish (in the usual preparations).
Inaba (91) likewise found an intermingling of cortex and medulla
in the new-born mouse.

The medulla cells of the suprarenal in the rat are polymorphic,
in irregular cords and clumps, separated by relatively large
blood sinuses. The cytoplasm is very scanty and the cell
boundaries are usually indistinguishable (as also in the cortex).
The cytoplasmic (chromaffin?) granules usually appear baso-

246 Cc. M. JACKSON

philic, with a bluish-violet stain, in contrast with the typical
reddish (eosinophile) color in the cytoplasm of the cortical cells.
The nuclei appear similar to those of the adjacent cortex. ‘The
stroma is not clearly differentiated. The deeply staining cell
clusters above referred to are apparently in most cases remnants
of the embryonic sympatho-chromaffin cells, consisting of
deeply staining nuclei with scanty cytoplasm. They are dis-
tinct from the intermingled cortical cell strands, which are
especially evident in preparations stained for lipoids or .chro-
maffin reaction.

The chromaffin reaction is present, but faint, in the new-born
rat and increases but slightly in intensity during the first week.
The medulla cells assume a pale brownish color, by which they
may be clearly distinguished from the cortex and from the in-
termingled cortical cord present in the medulla. In mammals
generally the chromaffin reaction of the medullary cells appears
during the fetal period (compare Poll, ’05) and was noted by
Soulié (’03) well marked in the new-born rat and guinea-pig.

Lipoids. In unstained frozen sections of the suprarenal (for-
malin-fixed a few hours) in the new-born rat, the cortex appears
opaque, due to the emulsion formed by the lipoidal granules.
Following the terminology of Albrecht and Bell, these lipoidal
granules and droplets will be designated as ‘liposomes.’ They
are soluble in absolute alcohol and xylol, but insoluble in 1 per
cent aqueous potassium hydroxide and in 1 per cent acetic
acid. With the exception of an indistinct clear streak between
the outer and middle zones, they appear somewhat uniformly
distributed throughout the cortex and in the cortical cell strands
extending throughout the medulla.

These sections, when stained with Herxheimer’s scarlet red
or with 1 per cent osmic acid and mounted in glycerin, reveal
still more clearly the liposomes (fig. 1). Although not so abun-
dant as later, the lipoids appear in much larger amount than
would be expected from the corresponding lipoidal vacuoles
as seen in the usual stained paraffin sections. Their staining
reactions already appear distinct from those of the ordinary
fat droplets outside of the fibrous capsule, even when these

SUPRARENAL GLAND—EFFECTS OF INANITION 247

droplets are as small as some of the liposomes. With Herx-
heimer’s scarlet red, the liposomes stain a deep scarlet-reddish
color, while the extracapsular ordinary fat droplets assume a
characteristic lighter reddish color. With osmic acid, the or-
dinary extracapsular fat stains jet-black, the suprarenal liposomes
stain in varying shades of brown (mostly light brown).

In distribution the liposomes, as noted in the fresh sections,
are nearly uniformly present throughout the cortex, although
relatively few occur in the narrow clear band between the outer
and middle zones. In size they vary from extremely fine to
coarse. The largest, however, are much smaller than later,
now rarely reaching half the nuclear diameter. They form in
each cell a circumnuclear zone which appears very distinct in
the scarlet red or osmic-stained sections. The liposomes in
the cortical cell strands through the medulla of the newborn
rat usually appear similar in size and number to those in the
cortical cells elsewhere. No liposomes are present in the cells
of the medulla proper, which appear as clear and unstained
masses, intermingled with the cortical strands.

In a rat of the second day (F. 10.2), the cortical liposomes
in frozen sections stained with scarlet red appear in general
as in the new-born, possibly slightly more abundant in the middle
zone. The cortical cell strands in the medulla, however, are
apparently undergoing absorption, and their liposomes appear
fewer and finer than elsewhere in the cortical cells. At three
days (fig. 2), the cortical cell strands throughout the medulla
are very inconspicuous, though still distinguishable under high
power in many places by their content of very fine (rarely coarser)
liposomes. In some places the corticomedullary border is now
clearly defined, though in other places still irregular. The
exact date at which the cortical cell strands in the medulla are
absorbed, leaving a clean-cut corticomedullary boundary, is
subject to individual variation.

Second week. In rats one week of age, in sections of the su-
prarenal stained with hematoxylin-eosin, under low magnifica-
tion, the cortical zones appear as in the new-born, but the me-
dulla is much lighter in appearance and now appears separated
from the cortex by a sharply defined border.

248 Cc. M. JACKSON

SUPRARENAL GLAND—EFFECTS OF INANITION 249

Under higher magnification the cortical zone cells appear in
general similar to those of the new-born. The inner-zone cells,
however, no longer intermingle with those of the medulla, but
(aside from a few scattered islands of cortical cells) present a
fairly even and very distinct line of demarcation. Next to
this border, the cortical cells frequently appear flattened and
more or less atrophic, probably due chiefly to absorption by
the expanding medulla.

The medulla has become confluent as a central mass, which
may also reach the surface at the hilus (compare fig. 10). It
has undergone a marked change in structure and appearance.
The stroma is well differentiated, forming an irregular syncytial
net work containing elongated nuclei and fine fibrillae. The
parenchyma cells are likewise syncytial. The cytoplasm is
more abundant and presents small vacuolated (non-lipoidal)
spaces, variable in size and number. ‘The fine violet cytoplasmic
granules are present, but variable and usually scanty. The
nuclei are somewhat variable in form, but usually spheroidal
and vesicular, very slightly chromatic. Some are smaller and
more deeply staining, occasionally even pycnotic. Bilobed
dumb-bell-shaped nuclei are occasionally seen in the medulla
(as also in the cortex), but these are comparatively rare. It
is doubtful whether they are to be considered as evidence of
amitotie cell division. The deeply staining masses (sympatho-
chromaffin cells) mentioned in the new-born occur rarely in
the suprarenal medulla at one week.

In a rat at eight days (F 11.3) frozen sections unstained or
stained with scarlet red or osmic acid show the liposomes some-

Fig. 1 Longitudinal section of the suprarenal gland of a new-born albino
rat (F 8.1). Formalin fixation; frozen section stained with Herxheimer’s scarlet
red. Liposomes visible in the cortex and in the cortical strands throughout the
medulla. The light band, relatively lipoid-free, which separates the outer and
middle cortical zones is already evident. X 80.

Fig. 2. Longitudinal section of the suprarenal gland in an albino rat three
days old (F 11.1). Formalin fixation; frozen section stained with Herxheimer’s
scarlet red. The medulla is confluent, and the intermixed cortical strands are
undergoing absorption, with the establishment of a definite corticomedullary
boundary. X 80.

250 Cc. M. JACKSON

SUPRARENAL GLAND—EFFECTS OF INANITION Pas w |

what similar to those of the first week. Those of the outer
zone are relatively most abundant, varying from fine to coarse,
the largest approaching nuclear size. The subjacent narrow
light band is relatively free from lipoids, with only a few fine
granules. In the middle zone the liposomes are less uniformly
distributed than heretofore, becoming more abundant in the
outer half of the zone. They are less abundant in the inner
zone. Here occur a few coarsely granular cells, which also
appear as individual cells or small groups sparsely scattered
through the medulla. These islands represent unabsorbed
remnants of the earlier cortical strands.

The chromaffin reaction of the medulla, though distinct, is
still comparatively weak, as during the first week.

At ten to fourteen days, the suprarenal in general structure
and appearance, in sections stained with hematoxylin-eosin, is
similar to that at one week. Lipoidal vacuoles are becoming
more abundant in the outer and especially the middle cortical
zone. The reddish (eosinophile) cytoplasmic granules of the
cortical cells are in contrast with the pale violet (faintly baso-
phile) granules of the medulla cells. The cortical cells of the
inner zone are usually normal in appearance, but frequently
atrophic and degenerated, especially in certain areas. Large
sympathetic ganglion cells appear in the medulla.

Fig. 3 Portion of a section of the suprarenal gland in a normal adult albino
rat (F 3.1). Formalin fixation; frozen section stained with Herxheimer’s scarlet
red. Liposomes most abundant in the outer cortical zone and the outer half of
the middle zone. fF, fibrous capsule, with ordinary fat droplets in the tissue
outside; O, outer zone (glomerulosa); 7’, transition band, relatively lipoid-free;
Mo, outer part of middle zone (fasciculata); Mz, inner part of middle zone; J,
inner zone (reticularis); 7, medulla. > 90.

Fig. 4 Portion of a section of the suprarenal gland of an adult albino rat
(F 3.2), after nine days of acute inanition. Formalin fixation; frozen section
stained with Herxheimer’s scarlet red. For explanations, see figure 3. No
apparent decrease in the liposomes, which appear more uniformly distributed
throughout the cortex. The light band at the transition between outer and
middle zones is obscured. X 90.

Fig. 5 Portion of a section of the suprarenal gland of an adult albino rat
(F 4.2) after twelve days of acute inanition. Formalin fixation; frozen section
stained with Herxheimer’s scarlet red. The liposomes persist in the outer zone
(O), but have nearly all disappeared elsewhere. A few are barely visible in the
outer half of the middle zone (Mo), and in the inner zone (J). X 90.

Fig. 6 Portion of a section of the suprarenal gland of a normal albino rat
(J 1.1) three weeks old. Zenker fixation, paraffin section stained with hema-
toxylin-eosin. Most of the middle cortical zone is omitted. F, fibrous capsule;
O, outer zone (glomerulosa), one cell in mitosis; 7’, transition band, nearly lipoid-
free; Mo, outer part, and M7, inner part of middle zone (fasciculata); 7, inner
zone (reticularis); M, medulla. X 300.

Fig. 7 Portion of a section of the suprarenal gland of an albino rat (S 7.31)
held at constant body weight by underfeeding from three to ten weeks of age.
For explanations, see figure 6. Progressive differentiation of lipoidal vacuoles
in the outer zone (O) and of pigment (P) in the inner zone (J). Some cell
atrophy, with hyperemia and increased degeneration in the inner zone. X 300.

Fig. 8 Portion of a section of the suprarenal gland of a normal albino rat
(St 7.45) ten weeks of age. For explanations, see figure 6. This represents the
normal adult structure. Lipoidal vacuoles well marked in the outer zone (OQ)
and outer part of the middle zone (Mo), but rare in the transition band (7).
One pigment mass (P) visible. > 300.

Fig. 9 Portion of a section of the suprarenal gland of an adult albino rat
(J 1.5) subjected to chronic inanition for five weeks. For explanations, see
figure 6. The lipoidal vacuoles have disappeared, except in the outer zone
(O). Marked atrophy with nuclear and cytoplasmic degeneration in the inner
cortical zone (J). Three pigment masses (P) visible. X 300.

Fig. 10 Longitudinal section of the suprarenal gland in an adult albino rat
(F 9.3) after ten days of acute inanition. Fixation in Miiller’s fluid; frozen
section (unstained). Normal chromaffin reaction in the medulla, which in this
case extends to the surface at the hilus. X 80.

253

254 Cc. M. JACKSON

In a rat at seventeen days (F 7.1) the fresh frozen sections
of the formalin-hardened gland show the characteristic cor-
tical opacity due to the liposomes. The clear line between the
outer and middle zones is distinct. The opacity decreases
greatly in the inner half of the middle zone and in the inner
zone, the medulla being perfectly clear.

This irregular distribution of the liposomes is still more ap-
parent when the sections are stained with scarlet red or osmic
acid. The liposomes are most abundant in the outer zone and
the outer half of the middle zone, where the largest droplets
now approach nuclear size. Toward the inner cortical zone
they become fewer and finer. Some of the inner zone cells are
entirely free from liposomes, others occasionally appear well
filled. Aside from a few scattered cortical cell islands, no lipo-
somes appear in the medulla.

The other supararenal gland from this rat was hardened in
Miiller’s fluid. The chromaffin reaction of the medulla is more
definite than in the earlier stages, though not so intense as later.
All of the medulla parenchyma cells are stained light brown,
the nucleus darker than the cytoplasm. The vacuoles of va-
rious size in the medulla cells are unstained, not giving the chro-
maffin reaction. The blood in the vascular spaces of the medulla
occasionally presents a reddish-brown color to a variable extent
(probably from absorbed epinephrin).

At three weeks. The paraffin sections were stained as usual.
The outer cortical zone (fig. 6, O) is narrow, usually 6 to 8 (rarely
12) cells deep. The cells are in irregular masses or columns,
separated by blood capillaries. Cell boundaries are ill defined.
The cytoplasm is scanty, contains fine eosinophile granules
and more numerous lipoidal vacuoles. The nuclei are hyper-
chromatic, some almost pyenotic. The deepest cells (fig. 6, T)
are transitional to the middle zone, and form a narrow zone
nearly free from lipoidal vacuoles.

The middle cortical zone is broadest. The cells (fig. 6, Mo,
M71) are arranged in very distinct cell columns, radially arranged,
and usually but one cell wide. The radial cell columns are
separated by blood capillaries, which become wider toward the

SUPRARENAL GLAND—EFFECTS OF INANITION 255

center of the gland. Cell boundaries are more distinct than
heretofore. The relatively abundant cyoplasm is filled with
characteristic eosinophile granules and a variable number of
lipoidal vacuoles. In some cases these vacuoles appear no
more numerous than in the second week. In others they are
more abundant especially in the outer half of the middle zone,
whereby these cells become much larger than the cells of the
inner half of the middle zone. The nuclei are typical in struc-
ture, spherical, and moderately rich in chromatin. Atrophic
cells occasionally occur in various stages of degeneration.

The inner zone (zona reticularis, fig. 6, 7) is rather narrow,
but of variable width, representing the area of irregular cell
columns next to the medulla. The irregularity of structure is
probably associated with the process of absorption accompany-
ing the expansion of the medulla, although only occasionally
are the border cells flattened as though atrophic from pressure.
Most of the cells in the inner zone are similar to those of the
adjacent middle zone, the cytoplasm containing eosinophile
granules and a few small lipoidal vacuoles. Some scattered
cells show various stages of degeneration and disintegration. Oc-
casional islets of such cortial cells occur also in the adjacent
medulla, but rarely deeper, toward the center of the medulla,
as in the earlier stages.

The medulla in stained sections (fig. 6, M) appears very light,
in’ strong contrast with the darker cortex. (The converse is
true if the suprarenal gland has been fixed in Zenker-formol
instead of Zenker’s fluid, the sections being stained with hema-
toxylin.) The parenchyma forms irregular cell masses, sepa-
rated by delicate fibrous stroma (with elongated nuclei) enclos- ‘
ing wide, sinusoidal blood-vessels. The cytoplasm of the
parenchyma cells is abundant, containing pale violet (chromaffin?)
granules and numerous non-lipoidal vacuoles, variable in size
and number. The nuclei are typically vesicular in form and
only moderately chromatic. A few are smaller and more deeply
staining, sometimes pycnotic. Degenerative cells with karyo-
lytic nuclei are rarely seen. Occasional large spherical sym-
pathetic ganglion cells and bundles of non-medullated fibers
appear.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 3

256 Cc. M. JACKSON

At eight weeks. The appearance of hematoxylin-stained sec-
tions under low power is similar to that at three weeks. The
cell structure in the outer and middle zones is also similar, with
some increase in the size and number of lipoidal vacuoles, es-
pecially in the outer half of the middle zone. The inner zone
shows in general a more atrophic appearance, and occasional
pigment cells occur for the first time. These are parenchyma
cells containing one (sometimes two) irregularly spheroidal,
light yellowish-brown, coarsely granular masses. As will be
shown later, this pigment is probably a lipochrome. The
nuclei of these pigment cells are irregular, sometimes central,
sometimes crowded to the side of the cell and flattened. The
inner zone cells bordering on the medulla are usually more or
less flattened and atrophic in appearance. The medulla ap-
pears similar to that at three weeks.

At ten weeks to adult. At ten weeks, the histological struc-
ture of the suprarenal (fig. 8) has reached practically the adult
condition. In fact, although the gland increases in size, but
few changes in structure are noticeable after the eight weeks’
stage previously described. In the usual hematoxylin-stained
sections the structure appears as follows (fig. 8):

The outer cortical zone (fig. 8. O) remains narrow, 6 to 12
cells deep. It is somewhat variable in structure, with irregular
cell cords separated by blood capillaries. The cells remain
relatively small, with nuclei of variable form and structure.
The cytoplasm is relatively scanty and somewhat granular. It
contains a variable amount of lipoidal vacuoles, usually giving
the cytoplasm a honey-comb appearance. The deepest cells,
on the border between the outer and middle zones (fig. 8, 7),
are relatively free from lipoidal vacuoles as found in the earlier
stages.

In the middle zone the only noteworthy change is in the
amount of lipoids, as shown by the vacuoles. These increase,
but to a variable extent. They are larger and more numerous
in the outer half of the zone (fig. 8, Mo), but smaller and fewer
in the inner half. In the larger cells of the outer part, the cyto-
plasm presents a reticulated (honey-comb) appearance, like

SUPRARENAL GLAND—EFFECTS OF INANITION 257

that of sebaceous gland cells. In cells with fewer lipoidal vac-
uoles, the eosinophile granules are more conspicuous. Occa-
sionally basophilic granules also occur. These are especially
evident in preparations stained with iron-hematoxylin, and prob-
ably in part correspond to the ‘corps sidérophiles’ or the mito-
chondria described by various French investigators. The nuclei
are usually central in position, spherical, and typical in structure.
A few degenerative cells occur as heretofore.

The inner zone (fig. 8, J) appears in general similar to that at
eight weeks, the outer cells of the zone being transitional to the
middle zone and the inner cells (next to the medulla) being typi-
cally more or less atrophic in structure. Atrophic or degenerative
changes have likewise been described in the inner cortical zone
of various animals (including the rat) by Gottschau (’83), Pfaund-
ler (92) Soulié (’03) and others. According to Kolmer (12 a,
12 b) these degenerative changes are increased by pregnancy
(guinea-pig).

Pigment cells in the inner cortical zone of the rat at ten w pole
occur more frequently than at eight weeks, but in variable de-
gree. They are somewhat variable in number, often numerous,
especially in the older rats, and may even extend somewhat into
the middle zone. The nuclei of the pigment cells may be central
in position, surrounded by the light yellowish-brown or greenish
yellow granular masses. In this case the nucleus is irregular in
form, but it is often pushed aside and flattened. The pigment
is still visible, though inconspicuous, in stained paraffin sections.
In thin sections the pigmented substance frequently appears
vacuolated (fig. 8, P). In fresh frozen sections the pigment cells
are clearly visible, and they are deeply stained by osmic acid or
scarlet red. This would indicate that the pigment is a lipochrome,
probably derived by a transformation of lipoids or other fatty
substances. The vacuoles probably represent untransformed
lipoids, observed by Ewald (’02) and Da Costa (’13).

The cortical fats (lipoids) in general are best studied in frozen
sections of suprarenals, either fresh or hardened a few hours in
formalin. In the unstained sections the distribution of the lipo-
somes is clearly evident on account of the varying degrees of

258 Cc. M. JACKSON

opacity caused by them. The extracapsular fat droplets, on the
other hand, appear clear and light. The liposomes present the
same reactions mentioned for the new-born.

The amount and distribution of the liposomes are most élearly

shown in the frozen sections stained with scarlet red (fig. 4). In
the narrow outer zone, they are usually very abundant and closely
packed, the largest droplets somewhat exceeding the average
nuclear size. The subjacent clear line or narrow band is rela-
tively (but not absolutely) free from liposomes. They are most
abundant and closely packed in the outer half of the middle zone,
where the largest may reach double the average nuclear diameter.
In the inner half of the middle zone they become more or less
reduced in amount, and are usually scanty in the inner cortical
zone (aside from those associated with the pigment cells). No
liposomes are present in the medulla, aside from occasional groups
which appear in the cortical islands near the boundary.
. The distribution of liposomes in sections stained with osmic
acid corresponds closely with that in sections stained with scarlet
red. Thestriking contrast in the staining reactions of the cortical
lipoids and the extracapsular ordinary fat was mentioned in the
new-born suprarenal, and is constant. The pigment cells (lipo-
chrome) of the inner cortical zone frequently react more like the
ordinary fat, however, staining a lighter reddish than the lipo-
somes with the scarlet red and darker than the liposomes in prep-
arations stained with osmic acid.

The amount and distribution of lipoids, as is well known, vary
much according to age and species. In general, the formation of
lipoids begins during the prenatal period (Poll, ’05; Starkel and
Weegrzynowski, 710), and increases during postnatal life, though a
decrease in the human suprarenal fat after childhood is noted by
Napp (’05). As to species the lipoidal content appears greatest
in man, carnivora, and rodents, less in ruminants and pachy-
derms (Ciaccio, ’10, and others). The lipoids are usually, as in
the rat, most abundant in the middle zone, with less in the outer
and inner zones. In some cases, however, a greater abundance
of lipoids of the outer zone has been observed (Hultgren and An-
derson, 99; Ewald ’02; Napp, ’05; Bonnamour, ’05 b; Starkel and

SUPRARENAL GLAND—EFFECTS OF INANITION 259

Weerzynowski,’10) ; and in the sheep, according to Mayer, André,
Mulon and Schaeffer (’12), the sparse lipoidal granules are lo-
cated exclusively in the outer zone. The characteristic fat-free
band between the zona glomerulosa and the fasciculata was
noted in the rat by Bonnamour (’05 b). Babes and Joneseo (’08)
likewise noted a scarcity in fat between these zones in the dog,
but Mulon (712), on the contrary, finds larger droplets at this
level. The composition of the suprarenal liposomes is generally
considered to be chiefly lecithin or cholesterin with a variable
admixture of ordinary fats (compare Ciaccio, 710, and Kawa-
mura, 711).

Pigment has often been described in the inner cortical zone of
the guinea-pig (even in the albino, according to Mulon) and man,
but less frequently in other forms. It was noted by Gottschau
(83) and Pfaundler (92) in most mammals, Baroncini and
Beretta (’01) in bats, and by Elliott and Tuckett (’06) sparsely
in the duckbill and pigeon. Bonnamour (’05 a) found it variable
in the dog and horse, and absent in the cat, rabbit, and marmot.
He found it rarely inthe rat, with no difference in pigment be-
tween white and black rats. DaCosta (713) also noted a few
pigment cells in the zona reticularis of Mus decumanus. Dewit-
ky (12) likewise noted brownish pigment in the cortex of the
rat at seven weeks. - Delamere (’03) and others have noted that
the amount of pigment usually increases with age, being rare or
absent in the young. Mulon (’02, ’03 a), Ciaccio (’05), Kolmer
(712 b), and others believe that the pigment formation may rep-
resent a secretion with physiological significance, and an in-
crease during pregnancy (guinea-pig) is claimed by Kolmer (12 a).
As to its composition, this pigment has usually been considered
as related to fat (lipochrome), although Starkel and Wegrzynow-
ski (710) and Thomas (’11) consider that the pigment appearing
in the degenerating inner cortical zone of the suprarenal in the
human new-born is of hemal origin and different from the pig-
ment in the adult gland.

The suprarenal medulla in the rat at ten weeks and older may
extend to the surface at the hilus (fig. 10), as found by Poll
(99). It is similar in structure to that described in earlier stages.

260 Cc. M. JACKSON

The stroma (also that in the inner cortical zone) stains indistinctly
bluish with Mallory’s anilin-blue connective-tissue stain. The
parenchyma cells (fig. 8, M) are large. With Zenker’s fixation
and hematoxylin-eosin stain, the cytoplasm contains the usual
faintly basophilic (chromaffin?) granules and occasionally a few
spherical eosinophile bodies of variable size. The cell periphery
frequently contains irregular vacuoles, non-lipoidal and of un-
known significance. They are variable in size and are somewhat
more prominent than in earlier stages. According to Ciaccio
(05), they are acidophile in reaction, like true nucleoli, and
are comparable to ‘plasmosomes.’ The structures described by
Ciaccio probably correspond to the eosinophile bodies referred
to above, and not to the characteristic vacuoles. The nuclei, as
heretofore, are typically vesicular, though a few of them are
small and hyperchromatic (sometimes pycnotic). The blood sin-
uses, sympathetic ganglion cells, and occasional atrophic cortical
islands occur as heretofore.

In preparations fixed in Miiller’s fluid, the chromaffin reaction
of the medulla is always well marked (as in fig. 10). The paren-
chyma cells appear as brownish masses separated by the un-
stained vascular areas. There is some variation in the intensity
of the reaction in different individuals, although the variation is
not great in sections of the same thickness. It also usually ap-
pears fairly uniform throughout the medulla, all of the paren-
chyma cells being somewhat similar in their reaction. Some
masses or clumps of cells may stain more deeply, however. The
brownish color appears in the cytoplasm, which may appear
homogeneous or granular (the granules being most distinct in
very thin sections). The cytoplasmic vacuoles of various size
remain perfectly clear and unstained. Whatever their nature,
they evidently do not contain epinephrin. The nucleus gives the
chromaffin reaction, being stained slightly darker than the cyto-
plasm. This was noted by Dostoiewsky (’86) confirming Henle
(versus v. Brunn). According to Ciaccio (’05), Diamare claims
that the chromaffin substance fills the whole cell, while Grynfeld,
localizes it in the cytoplasmic granules. The reaction is weak-
ened or lost a few hours post mortem (Dostoiewsky, ’86; Ciaccio.

SUPRARENAL GLAND—EFFECTS OF INANITION 261

705, and others), and is also affected by anesthetics (Schur and
Wiesel, 08; Hornowski, 09). The chromaffin reaction sometimes
appears also to a variable extent within the blood-vessels of the
medulla of the rat, as has often been noted in other animals by
various observers.

2. Changes in young rats stunted by underfeeding

Underfed from birth. In arat (St 80.9) underfed from birth to
twelve days and reaching a body weight of 8.9 grams, the supra-
renal gland has increased in weight to 0.0026 gram, which is about
normal for seven days of age. The normal differentiation of
the suprarenal cortex and medulla has occurred, and the struc-
ture (in sections stained with hematoxylin) is essentially similar
to that of the normal rat at seven days. Thus the histological
differentiation has continued, as in a norma! gland of correspond-
ing weight, although( as previously shown) the rate of mitosis has
been greatly diminished.

In a rat (St 247.5) underfed from birth to seven days, there is
apparently no increase in the weight of the suprarenal. Frozen
sections stained with scarlet red and osmic acid show the cortical
liposomes somewhat similar in appearance to those in the new-
born, but progressive absorption in the cortical cell strands of the
medulla is evident. The chromaffin reaction is normal.

In another rat (St 228.4) underfed from birth to fifty-eight
days, the suprarenal gland has increased markedly in weight and
shows a progressive differentiation of liposomes similar to those
in a normal gland of similar weight (F 7.1). The chromaffin re-
action in the medulla is variable. Some cell masses show a very
definite reaction. In others it is entirely absent, giving the
medulla a spotted appearance. This irregularity of the chro-
maffin reaction is probably due to the inanition, as it is not ob-
served in the normal animals.

Rats at maintenance from three weeks to eight, ten, or twelve weeks
of age. In these underfed young rats the body weight and supra-
renal weight (table 1) are but little above the initial weight at
three weeks. In general, the stained sections (Zenker’s fixation,

262 Cc. M. JACKSON

hematoxylin-eosin stain) of the suprarenal appear similar to
those normal at three weeks, though certain changes are clearly
evident (fig. 7).

The vacuoles in the outer-zone cells usually indicate a progres-
sive lipoidal differentiation (fig. 7, O). While in a few cases the
lipoidal vacuoles appear scanty (as normally at three weeks), in
others the outer lipoid zone was as distinct and clearly differen-
tiated as normally at ten weeks. In most cases the lipoidal
content appears somewhere between these two extremes. The
eosinophile cytoplasmic granules are somewhat indistinct.

The lipoidal vacuoles of the middle zone were also found in-
creased in size and number in most cases, though more variably
and not so definitely as in the outer zone. As in normal differen-
tiation, the lipoids usually accumulate to a greater extent in the
outer half of the middle zone. Otherwise there is no constant
change to be noted, excepting apparently a somewhat greater
number of degenerating cells than occurs normally.

The inner zone (fig. 7, 7), however, when compared with the
normal shows the most conspicuous changes. This zone is con-
stantly hyperemic and very markedly atrophic. The cells in
general appear greatly decreased in size, and many show various
stages of nuclear and cytoplasmic degeneration. More cells are
flattened at the medullary border than occurs normally. The
nuclei are frequently pycnotic or karyolytic. Pigment masses
(not present at three weeks) appear (fig. 7, P) and are even more
abundant than in the normal gland at ten weeks.

The medulla (fig. 7, 1) shows relatively less change than any
other part of the suprarenal. In most cases it resembles closely
the normal at three weeks, though in some cases there is an evi-
dent increase in the proportion of smaller, hyperchromatic (oc-
casionally pycnotic or karyolytic) nuclei. The cytoplasmic gran-
ules are usually indistinct. No chromaffin tests were made in
these cases.

Maintenance from three to fifteen or twenty weeks of age. In
these two young rats (St 33.1 and St 38.8) held at maintenance
for unusually long periods, the changes in the suprarenal are
more pronounced. The glands were fixed in Zenker’s fluid, sec-

SUPRARENAL GLAND—EFFECTS OF INANITION 263

tioned and stained as usual. In general, the nuclei appear more
hyperchromatic, though not much more than in the normal
gland at three weeks. Although the fat in the fibrous capsule
around the gland has apparently almost entirely disappeared,
the lipoidal vacuoles of the outer zone of the cortex are abundant.
In the middle zone, however, they are apparently not more num-
erous than in the normal at three weeks. In the inner part of
the middle zone many cells are degenerating some in various
stages of disintegration. The atrophic degeneration and pig-
ment formation in the inner cortical zone is very marked. The
greenish-yellow, vacuolated pigment masses in many cases have
entirely filled the cells and are quite numerous. The medulla
cells show increased vacuolization, sparser granulation, and hy-
perchromatic (frequently pyenotic, some karyolytic) nuclei. In
general, however, the structure is less changed than in the cortex.

3. Changes in young rats refed after stunting by underfeeding

In the rats refed fully after maintenance from three to twelve
weeks of age, the suprarenal gland rapidly returns to approxi-
mately normal structure. Even after one week of refeeding the
cortex appears distinctly lighter in stained sections, due to a de-
crease in the chromatin content of the nuclei. The lipoids ac-
cumulate more abundantly (especially in the outer half of the
middle zone) and the inner cortical zone becomes somewhat less
atrophic in appearance. By the end of two weeks of refeeding,
the gland has reached nearly normal structure, although some
areas of degeneration may persist for longer periods. Even those
held at maintenance up to twenty weeks of age (S 33.118 and $
33.120), with permanently stunted body weight, show practi-
cally normal structure in the suprarenals after being refed fully up
to about one year of age.

4. Changes in adult rats after acute or chronic inanition

Adult acute inanition. Adult rats were given water only for
seven to twelve days, with loss in body weight of 29 to 45 per
cent (table 1 D). The suprarenal glands, as previously shown

264. Cc. M. JACKSON

(Jackson, 715 a), lose but little in absolute weight during adult
inanition. The cells (especially those of the middle cortical zone
and the medulla) during acute inanition apparently diminish in
size, however, as shown in table 4. The decrease in cell size is
counterbalanced by hyperemia of the cortex, though apparently
the reverse (decrease in proportion of vascular stroma) occurs in
the medulla.

Sections of the suprarenal, stained with hematoxylin-eosin,
show changes in histological structure. The cytoplasm is re-
duced in amount and the eosinophile granules usually indistinct.
The nuclei are variable, frequently hypochromatic. More fre-
quently, however, they appear hyperchromatic, though less so in
the outer half of the middle zone. Pycnosis and deformity of
the nuclei are frequent, as described by Bonnamour (’05b) in
the outer cortical zone of the starved rat. The lipoidal vacuoles
are variable. Usually, however, they are absent or considerably
reduced in number and size in the middle zone, though persisting
nearly unchanged in the outer zone. The eosinophile granules
of the cortical cells, though sometimes indistinct, are often well
preserved and become more evident with the reduction in the
amount of lipoids present.

The inner cortical zone shows the hyperemia and atrophic con-
dition usually more pronounced than in the normal animal.
Pigment masses are frequent, but it is somewhat doubtful whether
they are increased in number. Degenerating cells with pycnotic
or karyolytic nuclei in various stages of disintegration are
numerous.

Scarlet red or osmic staining of fresh frozen sections reveals
_the cortical liposomes much more abundant than would be sus-
pected from the ordinary stained paraffin sections. In some
cases the usual fat-free boundary line between the outer and mid-
dle zones is nearly obliterated by an increased development of
liposomes, and they may appear more uniformly . scattered
through the middle zone, and even the inner zone (fig. 4).

This tendency toward a more uniform distribution of the lipo-
somes throughout the cortex was noted in three of the eight cases
of adult acute inanition stained especially for lipoids, though the

SUPRARENAL GLAND—EFFECTS OF INANITION 265

total amount of the lipoids appears somewhat reduced in one of
them (F 1.2). A similar ‘Verbreitung der lipoidhaltigen Schich-
ten’ was observed by Landau (713 b) in fasting guinea-pigs.

In five of the eight rats in my series, however, the change con-
sists In a very marked reduction in the liposomes of the middle
and inner zones. Under the low power of the microscope only a
few liposomes may remain visible in the outer half of the middle
zone, the remainder of the middle and inner zones being appar-
ently free from them (fig. 5). Under higher power, however,
numerous extremely fine granules are often still visible in all these
cortical cells. The effect in such cases is therefore apparently
merely to reduce the liposomes very muchinsize. In other cases,
the liposomes have entirely disappeared. The pigment cells
near the corticomedullary zone apparently persist nearly un-
changed, contrary to Rondoni and Montagnani (’15), who found
a decrease in fasting guinea-pigs.

In striking contrast with the remainder of the cortex is the
narrow outer zone, in which the liposomes persist with great
tenacity (fig. 5). With the scarlet stain, this zone therefore ap-
pears nearly unchanged as a deep reddish band, while the re-
mainder of the cortex appears pale, nearly colorless, or with a
pale, diffuse reddish color suggesting some fat-like substance in
solution. Osmic preparations give corresponding pictures. In
only one case (F 9.3) the outer zone in places appears somewhat
broken and irregular. The tenacity of the outer lipoidal zone is
not due to any visible difference in the structure or composition
of the liposomes in this zone. Landau (13 b), however, finds
that in the fasting guinea-pig, although there is no decrease in
the total lipoid content, the cholesterin decreases in all but the
outer cortical zone. It may be dependent upon the vascular
arrangement, as all of the blood to the cortex passes first through
the capillaries of this zone (Flint, ’00).

The amount of decrease in the liposomes of the suprarenal cor-
tex is usually, but not always, somewhat proportional to the
length of the fasting period or to the loss in body weight. In
general, the lipoids appear much more resistant to inanition than
does the ordinary fat just outside the suprarenal capsule.

266 Cc. M. JACKSON

The suprarenal medulla is often considerably affected during
acute inanition, though usually less so than is the cortex. The
‘eytoplasm frequently becomes more vacuolated, although the
characteristic (chromaffin?) granules persist in the parenchyma
cells. The nuclei may become more hyperchromatic, and a
larger proportion are pycnotic. In many cases the nuclei are
hypochromatic, undergoing chromatolysis. Some areas of
marked cellular degeneration occur.

The chromaffin reaction is apparently but slightly Gf at all)
affected by the inanition experiments. The slight variations in
the intensity of the brown color are apparently no greater than
appear in the normal controls. Even in the rat starved twelve
days with loss of 45 per cent in body weight (F 5.2) the chromaffin
reaction of the medulla appears fully as intense as in any of the
controls. This rat was still active when killed, but the reaction
persists in another (F 9.3) which was killed while very weak and
near death with a loss of 33 per cent in body weight after ten
days of acute inanition (fig. 10). .

In only one case (F 6.2) is there a marked decrease, only traces
of the chromaffin reaction being present. This was probably a
postmortem change, as the rat was found dead. It had lost only
34 per cent in body weight after seven days of inanition. It
therefore appears that in adult rats acute inanition produces no
appreciable decrease in the chromaffin reaction of the suprarenal
medulla, even (aside from postmortem changes) in those starved
to death.

Adult chronic inanition. In six adult rats underfed thirty to
thirty-five days with gradual loss in body weight amounting to
33 to 38 per cent, the histological changes (in the ordinary prepa-
rations) were found in general very similar to those after acute
inanition (fig. 9). These include a general atrophy of cells and
nuclei with a variable reduction in the amount of lipoids in the
middle (but not in the outer) zone (fig. 9, O). There is similarly
a pronounced cellular atrophy in the inner zone, and in one case
(J 1.5, fig. 9, J) the vacuolated cytoplasm in a few places had ap-
parently disintegrated to form extensive intercellular spaces, sim-
ilar to the condition described in the human suprarenal by Meyer

CZ):

SUPRARENAL GLAND—EFFECTS OF INANITION 267

The cells usually undergo simple atrophy, however. Many of
the nuclei appear chromatolytic, some pyecnotic. There is ap-
parently no change in the number and appearance of the pigment
cells, as a rule, although sometimes they appear increased in
number (fig. 9, P). The changes in the medulla are similar to
those noted under acute inanition. The special lipoid and
chromaffin stains were not employed in the chronic-inanition
series.

From the foregoing it appears that during inanition in the al-
bino rat the behavior of the suprarenal lipoids is somewhat vari-
able according to ircumstances. In young rats severely stunted
by underfeeding there is no apparent loss, the liposomes continu-
ing to differentiate as in the normal gland of corresponding
weight. Likewise in older rats there is sometimes no evident
decrease, but instead a tendency to more uniform distribution of
the liposomes throughout the cortex. This appears characteris-
tic where the inanition has not been carried to extremes. No de-
crease in suprarenal fat (lipoids) during inanition has also been
observed by Frederici (’03) in the guinea-pig, Traina (’04) in
man and rabbit, Napp (’05) in man, Kawamura (’11) in man,
Landau (713 b) in man, cat and guinea-pig. Bonnamour (’05 b)
even finds an increase in the starved rat, cat, rabbit, and guinea-
pig. Ciaccio (according to Landau, 713 b) during inanition found
first an increase, with subsequent decrease, in the suprarenal fat
content.

In most cases of acute and chronic inanition in the rat, espe-
cially in those carried to extremes, there is awell-marked decrease
in the suprarenal lipoids, although they are retained in the outer
cortical zone with remarkable tenacity in all cases. A decrease
in the fat (lipoids) of the suprarenal during inanition has been
observed by Orth (’93) in man, Beneke in man and animals,
Ewald (02) in the rabbit, Herman (’05) in man, Ponomarew
(14) in the mouse, and Rondoni and Montagnani (715) in the
guinea-pig. In the latter two investigations, more or less reten-
tion of fat in the zona glomerulosa (of mice and guinea-pigs) is
mentioned, which would be in agreement with my observations
on the rat. A variable decrease in suprarenal fat (lipoids) as a

268 Cc. M. JACKSON

result of various toxic conditions has been noted by various in-
vestigators in man and lower animals. The apparently contra-
dictory results of different observers as to the effect of inanition
upon the suprarenal fat (lipoids) is doubtless due to variations in.
the age, species, and individual animals, as well as to variations
in the extent and character of the inanition.

In hibernating gophers, Mann (716) found no change in the
lipoid content of the suprarenal. Frederici (’03) found no de-
crease in hibernating bats, but a progressive decrease is noted by
Baroncini and Beretta (’01) and Ciaccio (10). Bonnamour
(05 b) and Ciaccio (710) found a decrease in hibernating hedge-
hogs.

As to the effect of inanition upon the chromaffin reaction of
the suprarenal medulla, my results for the rat are more uniform.
They indicate that (with rare exceptions in extreme cases, espe-
cially in the younger stunted rats) there is no appreciable decrease
in the reaction, except as a result of postmortem changes. This
is in agreement with the results of Luksch (’05, ’11) and Kuri-
yama (718), who found no decrease in the epinephrin content of
starved rabbits, but contrary to Venulet and Dmitrowsky (710)
in the rabbit, Borberg (’12) in the cat and guinea-pig, Rondoni
and Montagnani (’15) in the guinea-pig, and Pellegrini (716) in
the later stages of fasting. In most cases, these positive results
were obtained on animals starved to death, and are therefore
probably explainable as due to postmortem changes. Effect of
anesthetics and acid in the fixative (for example, in the formalin
used for Wiesel’s mixture) are also possible sources of error, as I
have learned by experience.

MORPHOGENESIS OF THE SUPRARENAL GLAND

Some general features in the process of morphogenesis of the
suprarenal gland may now be discussed. As is well known (for
details in various species compare Poll, (’05), the medulla arises
in the embryo in connection with the sympathetic system, and
later migrates, usually in the form of multiple, small sympatho-
chromaffin cords or masses, which (in mammals) pass through

SUPRARENAL GLAND—EFFECTS OF INANITION 269

the cortical anlage and finally collect in the center of the gland.
Here they later become confluent, a few cells (‘sympathoblasts’)
forming the sympathetic ganglion cells, the majority ‘phiochro-
moblasts’ or ‘chromaffinoblasts’) ripening into the characteristic
parenchyma of the medulla.

The time at which this confluence of the medulla is finally
completed varies in different mammals, but is usually during the
late fetal period. In Echidna, however, the immigration of the
sympatho-chromaffin anlage is greatly delayed, and the supra-
renal medulla is not formed until long after birth (Keibel, ’04).
Also in the new-born mouse (Inaba, ’91), cat, dog, and guinea-pig
(Soulié, 703) and occasionally even in the human new-born
(Starkel and Wegrzynowski, ’10; Zuckerlandl, ’12), the process is
still incomplete and admixture of cortex and medulla persists to
a variable extent after birth. In the mouse, Inaba (’91) found
the corticomedullary boundary line still indefinite at ten days,
but distinct at thirty days.

Soulié (’03) described the confluence of the medulla in the
suprarenal of the rat (Mus decumanus) as appearing in the 25-mm.
fetus. Dewitzky (’12), however, found the suprarenal medulla in
the rat very ill defined at birth, becoming distinct at three days
and thereafter. This is in general agreement with my observa-
tions, although I find the confluence of the medulla in the albino
rat to be a gradual process and subject to some individual varia-
tion. Itis usually completed by the end of the first week. Even
in the adult, however, short cords of cortical tissue may occa-
sionally extend into the medulla, as observed by Bonnamour
(05 a) in the rat and rabbit, and small cortical islands occur
near the margin of the medulla.

As to the exact manner in which the confluence occurs, but
few definite statements appear in the literature. Flint (00)
mentions.appearances of pressure atrophy in the cortical strands
in the medulla of the suprarenal in the fetal pig, and Soulié (’03)
states that the cells of the medulla anlage ‘‘etouffent peu a peu
les cordons corticaux emprisonnés au stade de pénétration.”’ It
seems to be generally assumed, however, as is stated definitely
by Inaba (’91) for the mouse, that these cortical cell strands and

270 Cc. M. JACKSON

masses intermingled with the medulla are finally displaced and
- squeezed out of the medullary mass. In the rat, at least, there is
no evidence that such a displacement occurs. It appears rather
that the medulla becomes confluent through degeneration and
absorption of the intermingled cortical remnants. This process
is difficult to observe in preparations fixed and stained in the
usual manner, on which account it is easily overlooked, but it is
clearly evident in frozen sections stained with Herxheimer’s scar-
let red. In such preparations, the stained liposomes reveal the
atrophic cortical cells undergoing gradual atrophy and absorp-
tion. A careful study of similar preparations would probably
show that in other forms the primitive cortical strands in the
medulla likewise undergo degeneration and absorption.

The process of absorption of the cortical tissue continues during
the postnatal growth and development of the suprarenal gland,
associated with the expansion of the medulla. The continued
postnatal growth of the medulla has been noted in the rat, cat,
rabbit, and,guinea-pig by Elliott and Tuckett (06) and in man
by Scheel (’08), Starkel and Wegrzynowski (10), Thomas (’11).
Kern (711), and others. The continued postnatal increase in the
absolute volume of the medulla in the rat is confirmed by the
extensive data in the present study.

This expansion of the medulla necessarily involves an encroach-
ment upon the space formerly occupied by cortex. A priori, this
might happen in three ways: 1. There might be a corresponding
interstitial growth and expansion of the adjacent cortex. This,
however, would require continued multiplication and growth of
the cells in the inner cortical zone, and numerous observers (as
previously shown) agree that postnatal cell division in the inner
cortical zone during postnatal growth rarely or never occurs.

2. In the absence of cell division with interstitial growth, the
inner cortical zone might remain passive and be mechanically
displaced by the expansion of the medulla. Such a displacement,
however, would inevitably result in a very marked flattening of
the cortical cells on the adjacent surface of the expanding medulla
It cannot be denied that occasionally such a flattening does ap-
pear, but it is irregular and inconstant. Indeed the characteristic

SUPRARENAL GLAND—EFFECTS OF INANITION 271

irregularity of arrangement of the cell cords in the zona reticularis
is perhaps in part due to the pressure of the expanding medulla.
In the rat, however, and apparently in other forms, the histo-
logical structure of the inner (reticular) zone at the cortico-
medullary border does not, in general, support the idea of a
mechanical displacement by pressure.

3. The remaining possibility is that there is an actual absorp-
tion and removal of the cortex at the corticomedullary border.
This theory is strongly supported by the available evidence. Hy-
peremia of the inner cortical zone, absence of cell division and a
more or less well-marked cell atrophy with degeneration and pig-
ment formation are (as has been shown) characteristic not only
for the rat, but for mammals in general. In most cases, this proc-
ess of absorption is comparatively slow and inconspicuous, as in
the rat, but in the human infant it is more prominent. The ex-
tensive degenerative atrophy of the inner cortical zone of the
suprarenal in the human new-born, as described by Starkel and
Wegrzynowski (’10), Thomas C11), Kern (’11), Elliott and Ar-
mour (’11), Landau (’13 a) and Lewis and Pappenheimer (’16), is
therefore not a unique phenomenon, as heretofore supposed. It
appears to be merely an exaggeration of the same fundamental
process found in the development of the suprarenal in other
mammals. It is thus incorrect to claim that nothing similar
occurs in the lower animals (Kern, ’11 ; Dewitzky, ’12; Landau,
13a). The erosion of the inner cortical zone is evidently a con-
tinuation of the same process of degeneration and absorption of
the cortex which removes the cortical strands at the time of the
original confluence of the medulla. Minot (97), on the other
hand, believed that the cells of the primitive medulla anlage dis-
appear in the fetus, and agreed with the view of Gottschau (’83)
that the permanent medulla is derived by transformation of the
suprarenal cortex.

The process of cortical erosion by the medulla may be com-
pared with that of the absorption of the cartilage by the osteo-
genic tissue in the zone of enchondral ossification. In both
cases, small islands of the invaded tissue may persist for variable
periods. It is significant that such cortical islands in the medulla

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 3

272 Cc. M. JACKSON

of the suprarenal in animals after the earlier postnatal stages
usually occur only in the immediate neighborhood of the cortico-
medullary border. Those originally in the central portion of the
medulla have usually undergone complete absorption. Flint
(00), Pellegrino (04), and others have described cortical islands
in the medulla of the adult suprarenal, and they occur also in the
rat; but these probably represent the results of later cortical ero-
sion, rather than persistent embryonic remnants.

The exact nature of this absorptive process is uncertain. It
does not appear to be a mere pressure atrophy, due to the expan-
sion of the cortex, although this may be a factor of subordinate
importance in the process. . It would appear to be rather a chem-
ical process of obscure nature, possibly a result of the contact
between the essentially alien cortical and medullary tissues, not
yet fully adapted to each other in their comparatively recent
phylogenetic association. The observation by Lewis and Pap-
penheimer (10) of similar involutional changes in accessory Su-
prarenals composed of cortical tissue only would seem to invali-
date this theory, but this point needs further investigation. It
does not appear probable that the absorption of the cortex at the
corticomedullary border is of functional significance, as claimed
by Gottschau (793) and Mulon (02, ’03. a, 03 b, ’05 a, 705 b,
1D):

If the inner zone of the cortex is subject to continued erosion
and absorption during the morphogenesis of the suprarenal, it is
evident that the zone must be constantly regenerated. As has
been previously shown, the abundant evidence in various animals
clearly establishes the fact that although during embryonic and
fetal periods cell division occurs throughout the cortex, during
postnatal development it becomes progressively restricted to the
outer region of the cortex in general, and to the zona glomerulosa
in particular. There is therefore during the postnatal growth
period a continued renewal of the suprarenal cortex, the cyto-
morphosis of the cells including an origin in or near the outer
zone (glomerulosa), a descent through the middle zone (fascicu-
lata), and a final atrophy, absorption and removal in the inner
zone (reticularis) at the corticomedullary border. The process

SUPRARENAL GLAND—EFFECTS OF INANITION Zhe

may be compared with the cytomorphosis of the cells of dhe epi-
dermis, where, however, the process is centroperipheral instead
of peripherocentral in direction.

The weight or volume of the suprarenal gland as a whole, to-
gether with the relative volumes of the cortical and medullary
constituents, will therefore vary according to the rate of expan-
sion of the medulla, the rate of erosion of the cortex at the inner
zone, and the rate of regeneration from the outer zone. Thus
the postnatal involution of the inner zone in the human suprare-
nal is so rapid that it is not fully compensated by regeneration
from the outer zone, or by expansion of the medulla; so the gland
during the first year actually decreases in weight, as shown by
the data of Scheel (’08) Starkel and Wegrzynowski (710), and
others. In the rat the retardarion in the growth of the gland
during the first week is perhaps explainable upon the same basis.
Subsequent changes in the absolute and relative volume of the
cortex in the rat and other forms are evidently subject to much
variation in different individuals and species. As a rule, how-
ever, as in the rat, the medulla appears relatively small in vol-
ume in the earlier prenatal stages (during immigration), expands
rapidly to a maximum relative size following its confluence (early
postnatal stages), and thereafter decreases relatively correspond-
ing to the later more vigorous growth of the cortex (Canalis, ’87;
Hultgren and Anderson, ’99; Soulié, ’03; Elliott and Tuckett, ’06;
Scheel, ’08; Starkel and Wegrzynowski, 710; Thomas, ’11, and
others).

SUMMARY

1. In the new-born rat, the suprarenal cortex and medulla are
not yet distinctly separated, the cortical cell strands in the
medulla being absorbed during the confluence of the medulla
in the first, week. During the second and third weeks after
birth, the cortex increases from 75 or 80 per cent to about 90 per
cent of the entire gland, by volume. It apparently continues to
increase relatively to about 93 per cent at ten weeks of age, de-
creasing slightly in the adult. The medulla increases more
slowly in absolute volume, thereby decreasing in relative volume

274 Cc. M. JACKSON

from 20 or 25 per cent of the gland at one week to 15 per cent
at two weeks, 10 per cent at three weeks, and 7 per cent at ten
weeks, increasing again slightly in the adult.

2. The relative volumes of cortex and medulla are subject to
considerable individual variation. Their ratio shows no distinct
difference according to sex and is not materially changed in
young rats stunted by underfeeding or in adults subjected to
acute or chronic inanition.

3. The vascular stroma (in comparison with parenchyma) nor-
mally shows considerable individual variabilty in relative volume,
due chiefly to the varying distention of the blood-vessels. In
general, however, there is evidently an increased vascularity upon
passing from the outer zone of the cortex (average 13 per cent)
toward the medulla (average 28 per cent), with no constant
change according to age.

4. The changes in the volume of the vascular stroma during
inanition are variable. In general, the stunted young rats show
a marked hyperemia in the inner cortical zone, with relative
anemia in the middle and outer ones; medulla unchanged. In
stunted rats refed one to two weeks, the relative volume of the
stroma in general returns toward normal, though the outer zone
remains anemic. In adult rats with acute or chronic inanition
the inner and middle cortical zones appear hyperemic, the medulla
relatively anemic.

5. With the exception of the first week, there is in general an
increase in the size of the suprarenal parenchyma cells from birth
to ten weeks of age, with little or no increase thereafter. The
average change in cell diameter for the various zones between
birth and maturity is as follows: outer zone, 7 » to 9 »; middle
zone (outer portion), 9 u to 15 w; middle zone (inner portion), 9 u
to 12 w; inner zone remains about 9 »; medulla, 8 » to 16 un.

6. The nuclei increase slightly in diameter during postnatal
life, excepting those of the inner cortical zone. The cytoplasmic
increase is much greater, however, so the nuclei in general lag
behind in relative size (nucleus-plasma ratio). Thus the relative
nuclear volume decreases, in the outer cortical zone, from about
44 per cent of the cell volume in the new-born to 23 per cent in

SUPRARENAL GLAND—EFFECTS OF INANITION 200

the adult; in the middle zone (outer part) from 33 per cent to
10 per cent; in the middle zone (inner part) from 28 per cent to
16 per cent; in the inner cortical zone from 28 per cent to 21
per cent; in the medulla from 46 per cent to 10 per cent.

7. In the underfed young rats stunted three to ten weeks or
more, the suprarenal cells may increase in size (outer part of
middle zone), or decrease (inner cortical zone and inner part of
middle zone), or remain nearly unchanged. The nuclei are
similarly variable, with slight changes in the relative volume.
In the stunted rats refed one or two weeks, the cells and nuclei
in general approach, but have not yet reached, their normal size.

8. In adult acute inanition there is but little change in the size
of the cells and nuclei in the outer and inner cortical zones. In
the middle zone and medulla, however, there is a marked loss in
size of the cells, though but slight loss in the nuclei (with corre-
sponding increase in relative nuclear volume). In chronic adult
inanition there is a greater decrease in the size of the outer- and
middle-zone cells, but about the same as during acute inanition
in the inner zone and medulla. The nuclear loss is somewhat
greater than during acute inanition, with relative nuclear volume
not very different from normal, except in the middle zone (where
it is high).

9, Amitosis in the suprarenal of the rat is infrequent and of
doubtful significance. Mitoses are frequent in the new-born (av-
erage about 20 per section), but fewer at the end of the first week
(10 per section). They increase to a maximum rate during the
second and third weeks, slowly decreasing in number thereafter,
although occasionally found even in the adult. Mlitoses are
most frequent in the outer zone and outermost part of the middle
zone (these forming the ‘germinative zone,) less frequent in the
medulla, and rare in the inner cortical zone.

10. In young rats stunted by underfeeding, mitosis is more or
less completely suppressed in the suprarenal. On refeeding one
week, mitosis begins again, and in two weeks the normal rate is
reestablished. Cell division in the suprarenal is therefore con-
trolled by the amount of nutrition.

276 Cc. M. JACKSON

11. The three cortical zones of the suprarenal are distinguish-
able from birth and well differentiated at three weeks. Lipo-
somes (lipoidal granules or droplets) are present in considerable
amount at birth, being rather uniformly distributed through
the cortex and the cortical strands through the medulla. They
increase slowly rather and the distribution changes. They be-
come more abundant in the outer zone and outer half of the
middle zone, decreasing in amount toward the inner zone, with
none in the medulla. The inner cortical zone is atrophic in char-
acter, with cells in various stages of degeneration and absorption.
Pigment (lipochrome) appears in these cells after eight weeks of
age. The degeneration and absorption of this zone is associated
with the expansion of the medulla in the morphogenesis of the
suprarenal gland. The absorption of the inner cortical zone dur-
ing the growth of the suprarenal is characteristic for mammals,
though greatly exaggerated in the human infant. The chromaf-
fin reaction of the medulla is weak at birth, becoming well marked
in the rat after the age of three weeks.

12. In the young rats stunted by underfeeding certain cell
changes occur in the suprarenal. The cortex tends to undergo
the normal differentiation of liposomes, but their amount is
variable. In some cases many of the cells appear atrophic and
degenerative, especially toward the inner cortical zone, where
the process is extreme. Pigment appears in unusual abundance.
The medulla is usually less affected, although pyenotic nuclei
frequently occur. Only in extreme cases is there any appreciable
decrease in the chromaffin reaction. In stunted rats refed two
weeks the structure is already gaining a normal appearance,
though some areas of degeneration persist.

13. In adult rats subjected to acute or chronic inanition the
suprarenal cells present a simple atrophy, together with a vari-
able amount of degeneration. The nuclei may be either hypo-
chromatic or hyperchromatic, with frequent pyenosis or karyoly-
sis. The liopsomes are retained tenaciously in the outer zone,
though usually decreased to a considerable extent in the rest of
the cortex, especially in extreme inanition. The pigment re-
mains unchanged. The medulla cells show degenerative changes,

SUPRARENAL GLAND—EFFECTS OF INANITION 277

though usually less marked than those of the cortex. The chro-
maffin reaction is apparently retained undiminished up to the
point of death, though occurring as a postmortem change.

LITERATURE CITED

Bases, V., ET Jonesco, V. 1908 Distribution de la graisse dan. ~+ capsules
surrénales. C. R. Soc. Biol. Paris, T. 65, Ann. (1908, T. 2,

Bager, BerteL, 1917 Bidrag till binjurarnas aldersanatomi hos kai orn.
(Zur Altersanatomie der Nebennieren des Kaninchens.) Up. 2
Lakareférenings Férhandlingar. Ny. fdljd., Bd. 23, H. 1-2, S. 48-
116.

Barspra, A. G., £ Biccrt, D. 1900 Contributo istologico alla conoscenza
delle modificazioni che il digiuno apporta negli elementi anatomici dei
vari organi e tessuti del economia animal. Prima nota: Capsule
soprarenali. Bollettino delle scienze mediche. SocietaMedico-Chi-
rurgica e della Scuola Medica di Bologna. Anno 71, Serie 7, vol. 11,
pp. 679-682.

Baroncini, L., £ Beretta, A. 1901 Ricerche istologiche sulle modificazioni
degli organi mammiferi ibernanti. IV. Capsule surrenali (Nota pre-
ventiva). Lo-Riforma Medica, Anno 16, pp. 162-163; Anno 17, pp.
76-78.

Betz, E. T. 1910 The staining of fats in epithelium and muscle fibers. Anat.
Rec., vol. 4, no. 5, pp. 199-212.

BENEKE —— Pathologische Anatomie der Nebennieren, in Ziilzer’s Handbuch
der Harn- und Geschlechtsorgane. (Cited by Ewald, ’02.)

Brigepu, A. 1913 Innere Sekretion. 2 Aufl., 2 Abth. (S. 16).

Bonnamovur, 8. 1905a Etude histologique des phénoménes de secrétion de la

capsule surrénale chez les mammiféres. Thése, Lyon.
1905 b Modifications histologiques de la capsule surrénale dans cer-
tains états physiologiques (hibernation, inanition) et pathologiques
expérimentaux (diphthérie, rage). C. R. de l’Assoe. des Anatomistes
(Int. Congr. Anat. Geneva, 1905), vol. 7, pp. 87-93.

Borsperae 1912 Das chromaffine Gewebe. Nebennierenuntersuchungen. II.
Skand. Archiv f. Physiol., Bd. 28, 8.91. (Cited by Biedl, ’13.)

Canauis, P. 1887 Contribution 4 l’étude du développement et de la pathologie
des capsules surrénales. Jour. internat. mens. d’anatomie et de
physiol. (Int. Monatschr.), T. 4, pp. 312-333.

Craccio, C. 1903 Ricerche sui processi di secrezione cellulare nelle capsule
surrenali dei vertebrati. Anat. Anz., Bd. 23, S. 401-424.

1905 Sur la fine structure et sur les fom oiieas des capsules BHETEUAIeS
des vertébrés. Arch. ital. de biol. T., 43.

1910 Contributo alla distribuzione ad alla fisio-patologia cellulare
dei lipoidi. Arch. f. Zellforschung, Bd. 5, 8S. 235-363.

278 Cc. M. JACKSON

CotTtTeNntToT, Muton, Et ZIMMERN 1912 Action des rayons X sur la corticale
surrénale. C.R. Soc. Biol. Paris, T. 73, pp. 717-720.

Cristrant, M. er Mug. 1902 Histologie pathologique des greffes de capsules
surrénales. C.R. Soc. Biol. Paris, T. 54, pp. 811-814.

Da Costa, A. C. 1913 Recherches sur l’histo-physiologie des glandes sur-
rénales. Archives de Biol., T. 28, p. 111-196.

DeLAMARE, G. 1903 Recherches sur la sénescence de la glande surrénale.
C. R. Soc. Biol. Paris, T. 55, pp. 1152-1154.

DewiTzky, W. 1912 Beitrige zur Histologie der Nebennieren. Ziegler’s Bei-
trige, Bd. 52, S. 431-443.

Donatpson, H. H. 1915 The rat. Reference tables and data. Memoirs of
The Wistar Institute, no. 6, Philadelphia.

DostotEwsky, A. 1886 Ein Beitrag zur mikroskopischen Anatomie der Neben-
nieren bei Séugethieren. Archiv f. mikr. Anat., Bd. 27, S. 272-296.

ExutiottT, T. R., anp Armour, R. G. 1911 The development of the cortex in
the human suprarenal gland and its condition in hemicephaly. Jour.
of Path. and Bact., vol. 15, pp. 481-488.

Exutott, T. R., anp Tucxertt, I. 1906 Cortex and medulla in the suprarenal
glands. Jour. Physiol., vol. 34, no. 4-5.

Ewatp, Pavt 1902 Uber Fettgehalt und multiple Adenombildung in der
Nebenniere. Inaug.-Dissert. Miinchen.

Fuint, J. M. 1900 The blood-vessels, angiogenesis, organgogenesis, reticulum
and histology of the adrenal. Johns Hopkins Hosp. Rep., vol. 9,
pp. 153-230.

Freperic1 1903 Lo sperimentale (cited, without title, by Ciaccio, ’10).

Gortscuav, M. 1883 Structur und embryonale Entwickelung der Neben-
nieren bei Séugethieren. Archiv f. Anat. u. Physiol., Anat. Abth.,
S. 412-458.

GraHamM, G.S. 1916 Toxic lesions of the adrenal gland and their repair. Jour.
Med. Research, vol. 34, pp. 241-261.

Hartalr, 8S. 1913 On the weights of the abdominal and the thoracic viscera, the
sex glands, ductless glands and the eyeballs of the albino rat (Mus
norvegicus albinus) according to body weight. Am. Jour. Anat.,
vole 15) nosl-

Herman 1905 Uber Vorkommen und Verinderungen von Myelinsubstanz in
der Nebenniere. Inaug.-Dissert. Tiibingen. (Cited by Kawamura,
211)

Hornowsk!I, J. 1909 Verainderungen im Chromaffinsystem bei unaufgeklarten
postoperativen Todesfillen. Virchow’s Archiv, Bd. 198.

Hoskins, E. R. 1916 The growth of the body and organs of the albino rat as
affected by feeding various ductless glands (thyroid, thymus, hypoph-
ysis and pineal). Jour. Exper. Zool., vol. 21, no. 3.

HULTGREN UND ANDERSON 1899 Studien iiber die Physiologie und Anatomie
der Nebennieren. Skand. Arch. f. Physiol., Bd. 9, S. 73-311.

InaBa, M. 1891 Notes on the development of the suprarenal bodies in the
mouse. Jour. College of Science, Imp. Univ. Japan, vol. 4, part 1,
pp. 215-235.

SUPRARENAL GLAND—EFFECTS OF INANITION 279

Jackson, C. M. 1913 Postnatal growth and variability of the body and of the
various organs in the albino rat. Am. Jour. Anat., vol. 15, pp. 1-68.
1915 a Effects of acute and chronic inanition upon the relative weights
of the various organs and systems of adult albino rats. Am. Jour.
Anat., vol. 18, pp. 75-116.

1915 b Changes in the relative weights of the various parts, systems
and organs of young albino rats held at constant body weight by
underfeeding for various periods. Jour. Exper. Zool., vol. 19, pp.
99-156.

1917 Effects of inanition and refeeding upon the growth and structure
of the hypophysis in the albino rat. Am. Jour. Anat., vol. 21, pp.
321-358.

Kawamura, R. 1911 Die Cholesterinesterverfettung (Cholesterinsteatose).
Ein differentialdiagnostische morphologische Studie iiber die in den
menschlichen und tierischen Geweben vorkommenden Lipoide. Ver-
lag G. Fischer, Jena.

KeIBeEL, F. 1904 Zoologische Forschungsreisen in Australien und dem Malay-
ischen Archipel von Richard Semon, Bd. 3. Monotremen und Mar-
supialer, II. 2. Teil, S. 151-206. (Cited by Poll, ’05.)

Kern, H. 1911 Uber den Umbau der Nebenniere im extrauterinen Leben.
Deutsche med. Wochenschr., Bd. 37, S. 971-974; 1180; 1318-1319.

Kotmer, W. 1912a Uber gewisse physiologisch-histologische Vorginge in der
Nebenniere und deren Beziehung zum Genitalapparat. Zentralbl. f.
Physiol., Bd. 25, S. 1009.

1912 b Beziehung von Nebenniere und Geschlechtsfunktion. Pflii-
ger’s Archiv, Bd. 144, S. 361-395.

Kuriyama, 8. 1918 The adrenals in relation to Ee holeaeare metabolism.
III. The epinephrin content of the adrenals in various experimental
conditions. Jour. Biol. Chem., vol. 34, pp. 299-319.

Lanpav, M. 1913a Zur Entwicklung der Nebennierenrinde. Deutsche med.
Wochenschr., Bd. 39, no. 7, Feb. 13.

1913 b Nebenniere und Fettstoffwechsel. Deutsche med. Wochen-

. schr., Bd. 39, no. 12, May 20.

Lewis, R. W., AND PAPPENHEIMER, A. M. 1916 A study of the involutional
changes which occur in the adrenal cortex during infancy. Jour. Med.
Research, vol. 34, pp. 81-93.

Luxscg, F. 1911 Uber das histologische und funktionelle Verhalten der Neben-
nieren beim hungernden Kaninchen. Arch. f. exper. Path. u. Pharm.,
Bd. 65, S. 161-163.

Mann, F.C. 1916 The ductless glands and hibernation. Amer. Jour. Physiol.
vol. 41, no. 2.

Martinotti, C. 1892 Contributo allo studio delle capsule surrenale. Giornale
della R. Accad. di Med. di Torino. Serie III., vol. 40, pp. 299-301.

Mayer, Muton, eT ScHAEFFER 1912 Contribution A la microchemie des sur-
rénales. I. Recherches sur les surrénales du cheval. C. R. Soe.
Biol. Paris, T. 73, pp. 313-315. II. Recherches sur les surrénales de
mouton. Ibid., pp. 315-318.

280 Cc. M. JACKSON

Meyer, A. W. 1917 Some morphological effects of prolonged inanition. Jour.
“A Med. Research, vol. 36, pp. 51-78.

Minot, C.S. 1897 Human embryology. New York.

Muton, P. 1902 Excretion des capsules surrénales du cobaye dans les vais-
seaux sanguins. C. R. Soc. Biol. Paris, T. 54, pp. 1540-1542.

1903 a Sur le pigment des capsules surrénales chez le cobaye. C. R.
de l’Assoc. des Anat., T. 5, pp. 148-151.

1903 b Divisions nucléaires et réle germinatif de la couche glomé-
rulaire des capsules surrénales du cobaye. C. R. Soc. Biol., Paris, T.
55, pp. 592-595.

1905 a Evolution de la corticale surrénale du cobaye aves l’age de
animal. C. R. Soc. Biol. Paris, T. 59 (1905, T. 2), pp. 837-339. (Cf.
also pp. 592-593.)

1905 b Sur le pigment des capsules surrénales (Cobaye). Bibl. Anat.
T. 14, pp. 177-182.

1912 La corticale surrénale du chien. C. R. Soc. Biol. Paris, T. 73,
pp. 714-716.

Napp, O. 1905 Uber den Fettgehalt der Nebenniere. Virchow’s Archiv, Bd.
182, S. 314-326.

OrtH, J. 1893 Lehrbuch der path. Anat., Bd. 2.

PELLEGRINI, R. 1916 (Effect of fasting on the suprarenals.) Atti r. Inst.
Veneto, vol. 72, p. 781. (Cited from ‘Chemical Abstracts’ in Endo-
erinology, vol. 1, no. 3, 1917, p. 354.)

PELLEGRINO, M. 1904 Sopra una particulare disposizione della sostanze midol-
lare nelle capsule surrenali (mammiferi). Bollet. d. Soc. di Nat. in
Napoli, Anno 18, Série I, pp. 139-142. (Cited by Poll, ’05, and Da
Costa, 713.)

PFAUNDLER, M. 1892 Zur Anatomie der Nebenniere. Sitzungsber. d. k.
Akad. d. Wiss. in Wien. Math.-Naturw. Classe, Bd. 101, Abth. 3,
S. 515-533.

Pott, H. 1899 Verinderungen der Nebenniere bei Transplantation. Archiv f.
mikr. Anat., Bd. 54, p. 440-481.

1905 Die Entwickelung der Nebennierensysteme, in Hertwig’s Hand- |
buch der Entwickelungslehre der Wirbeltiere. Bd. 3, Teil 1.

PonoMAREw, A. 1914 Uber den Ursprung der Fettsubstanzen in der Neben-
nierenrinde. Ziegler’s Beitriige, Bd. 59, H. 2. (Reviewed in Zentral-
blatt f. norm. Anat. u. Mikrotechnik, Bd. 11, H. 12, 1914.)

Ronpont, P. © Monraanant, M. 1915 Lesioni istologiche nel maidismo, nel
digiuno e nello scorbuto sperimentale. Lo sperimentale, Anno 69,
pp. 659-696.

Roup, Aug. 1902 Contribution A l’étude du développement de la capsule
surrénale de la souris. Bull. de la soc. vaudoise des se. nat., T. 38,
pp. 187-258. (Cited from Jahresbericht f. Anat., volume for 1903,
and Poll, ’05.)

Scuret, O. 1908 Uber Nebennieren. Sekretkérnchen-Oedem-Gewicht. Vir-
chow’s Archiv, Bd. 192, 8. 494-513.

Scuur, H., unp Wirset, J. 1908 Uber das Verhalten des chromaffinen Gewebes
bei der Narcose. Wiener klin. Wochenschr., Jahrg. 21, No. 8.

SUPRARENAL GLAND—EFFECTS OF INANITION 281

Soutié, A. H. 1903 Recherches sur le développement des capsules surrénales
chez les vertébrés supérieurs. Jour. de l’anat. et de la physiol., T. 39,
pp. 197-293; 492-533; 634-662.

STARKEL, S., UND WEGRzYNOwSKI, L. 1910 Beitrag zur Histologie der Neben-
niere bei Feten und Kindern. Arch. f. Anat. u. Physiol., Anat. Abth.,
S. 214-326. (Also (Polish) Medycyna Warschau, reviewed by Hoyer
in Jahresbericht f. Anat. for 1910, III, p. 461-462.)

Stewart, C. A. 1916 Growth of the body and of the various organs of young
albino rats after inanition for various periods. Biol. Bull., vol. 31,
pp. 16-51.

Tuomas, E. 1911 Uber die Nebenniere des Kindes und ihre Veriinderungen bei
Infektionskrankheiten. Ziegler’s Beitriige, Bd. 50, S. 283-316.

Traina, R. 1904 Uber das Verhalten des Fettes und der Zellgranula bei chroni-
schem Marasmus und akuten Hungerszustiinden. Ziegler’s Beitrige,
Bd. 35, 8. 1-92.

VENULET, F., UND Durrrowsxky, G. 1910 Uber das Verhalten der chromaffinen
Substanz der Nebennieren beim Hungern und unter dem Einfluss von
Iodkali. Archiv. f. exper. Path. u. Pharm., Bd. 63, 5S. 460-464.

WieseL, J. 1901 Uber die Entwicklung der Nebenniere des Schweines, be-
sonders der Marksubstanz. Anat. Hefte, Bd. 16, S. 117-150.

1902 Zur Entwickelung der menschlichen Nebenniere. Centralbl. f.

Phys., Bd. 15, Verh. d. Morphol.-Physiol. Gesellsch. zu Wien (1901),
S. 614-615.

ZUCKERKANDL, E. 1912 The development of the chromaffin organs and of the
suprarenal glands. In Kiebel and Mall’s Human Embryology, vol. 2.

TABLE 1

Individual number, age and condition, sex, body length, body weight and weight of
the suprarenal glands tn the albino rats used for histological study

NOSE- BODY WEIGHT bee SEE
RAT NO. AGE, ETC. SEX ANUS GROSS ~— SUPRA- NOTE
LENGTH (OR NET) seas
A. Normal rats
mm. grams grams
F 10.1 New-born m. 46 4.2 0.0014 ss
F 10.2 1 day m. 50 4.8 | 0.0014 =
J1.7a New-born m. 4.9
J 1.7b New-born m. 4.9
F 8.1 New-born m. 50 il 0.0018 =
ype Ube 3 days m. 57 6.5 0.0020
Ret. 2 5 days tie 60 8.3 0.0020 *
halles 8 days m. 64 9.6 0.0018 cd
St 72.5 7 days ie 66 10.8 (10.1) | 0.0024
St 72.2 7 days m. 66 10.8 (10.0) | 0.0024
V 18.1 10 days m. 12.0 0.0030
St 80.5 12 days rs aD 15.5 0.0040
Vals 14 days m. 75 15.4
Be 17 days 1; 86 26.3 0.0054 =
V1.4 21 days m. 87 20.4
§ 5.2 21 days i 95 22.9 (21.1) | 0.0100
pils® 21 days ie 100 28.2 0.0078
J1.1 21 days m. 100 29.0 0.0078
St 47.2 21 days m. 102 34.3 0.0134
St 5.1 56 days ie 63.0 0.0116
St 5.2 56 days m. 77.0 0.0182
St 228.2 58 days m. 156 96.0 0.0183 a
St 47.6 67 days it 169 124.0 0.0364
St 47.5 67 days m. 191 196.0 0.0326
H 70.3 : 70 days m. 194 208 .0 0.0357
H 68.11 72 days m. 192 171.0 (164) 0.0228
H 68.8 74 days ie 184 142.0 (137) 0.0339
S 5.4 74 days fe 168 126.0 (118) 0.0319
S$ 5.3 74 days m. 180 172 0 (167) 0.0289
M 1.2 74 days m. 190 181.0 (173) 0.0264
piles 94 days m. 183 177.0 0.0249
H 64.3 101 days m. 191 185.0 (179) 0.0264
H 60.7 103 days te 203 (182) 0.0302
Leas 105 days m. 188 192.0 0.0234 .
H 58.3 106 days m. 225 (258) 0.0404
Apilire 112 days li 185 161.0 0.0365 | (1)
10) oil 116 days m. 176 149.0 0.0268 -
Mo. 9 120 days fe 180 157.0
H 36.3 138 days m. 202 (202) 0.0406
H 50.3 141 days 0 ee ee (222) | 0.0406

i)
90
bo

TABLE 1—Continued

WEIGHT

NOSE- BODY WEIGHT OF SEE

4 RAT NO. AGE, ETC. SEX ee See (on yan) SUFRA- NOTE
. mn grams grams
| aa | 141 days m. 195 200.0 0.0234
F 9.1 170 days m. 185 173.0 0.0246
F 6.1 150 days m. 188 178.0 0.0290
$14 (adult) m. 205 252.0 (247) 0.0414
21.3 202 days m. 216 (232) 0.0304
H 34.3 224 days i 194 (173) 0.0453
H 34.6 225 days E; 205 (188) 0.0604
L 3.7 234 days m. 190 241.0 (238) 0.0310
H 27.3 253 days ie 195 (166) 0.0475
H 27.6 254 days m. 214 (222) 0.0298
S 33.116 340 days Fi: 195 194.0 (188) 0.0414

S 33.117 346 days m

5 228 302.0 (294) 0.0330

B. Stunted young rats
!

53 4.5 0.0016 | *(2)

S 11.65 Maint. 21-73 days 100 23.8 (22.5) | 0.0118 | (2)

St 247.5 Maint. 1-7 days £:

St 80.9 Underfed 1-12 days 1 64 8.9 0.0026 (2)
St 228.4 Underfed 1-58 days m. 74 10.9 0.0054 | *(2)
St 9.36 Maint. 21-51 days m. 113 30.5 0.0124 | (2)
S 12.69 Maint. 21-66 days f. 100 24.5 (22.7) | 0.0090 | (2)
St 47.4 Maint. 21-66 days m. 113 32.3 0.0140 | (2)
St 47.3 Maint. 21-66 days m. 120 34.0 0.0136 | (2)
he oe ar Maint. 21-67 days m. 95 23.3 (21.2) | 0.0086 | (2)
S 7.34 Maint. 21-70 days m. P25 30.5 (29.0) | 0.0116 | (2)
8 7.31 Maint. 21-70 days m. 120 34.8 (31.4) | 0.0126 | (2)
S 7.32 Maint. 21-71 days m, 120 30.0 (28.8) | 0.0100 | (2)
9) 7.00 Maint. 21-71 days i 117 35.3 (31.0) | 0.0120 | (2)

m.

St 12.50 | Maint. 21-82 days m. | 123 45.0 (41.2) (2)
Sig Gayl Maint. 22-104 days it 89 19.1 (18.2) | 0.0086 (2)
St 38.8 Maint. 21-139 days m. 118 30.0 0.0161 | (2)

C. Refed after inanition

St 12.48 | 88 days Refed 7 days
St 11.42 | 88 days Refed 7 days
St.11.45 | 88 days Refed 7 days
St 10.27 | 89 days Refed 7 days
St 12.51 | 95 days Refed 14 days
St 11.43 | 95 days Refed 14 days
St 11.40 | 95 days Refed 14 days
St 10.26 | 96 days Refed 14 days
F 2.3 122 days Refed 6 days
S 33.120 | 339 days Refed 189 days
S 33.118 | 346 days Refed 206 days

125 50.2 (46.5) | 0.0105 | (3)
125 55.0 (50.3) | 0.0160 | (3)
130 67.6 (57.7) | 0.0156 | (3)
127 55.0 (51.4) | 0.0115 | (3)
137 77.2 (70.7) | 0.0150 | (3)
143 79.0 (74.7) | 0.0196 | (3)
142 84.5 (76.5) | 0.0158 | (3)
150 91.0 (86.3) | 0.0180] (3)
176 144.0 0.0350 | -*(3)
181 162.0 (156) 0.0396 | (3)
204 229.0 (219) 0.0348 | (3)

Bear Side) meee eee wee

TABLE 1—Concluded

NOSE- BODY WEIGHT | yaa | SEE
RAT NO. AGE, ETC. SEX ANUS GROSS SUPRA- NOTE
LENGTH (OR NET) Fenn e
D. Adult acute inanition
| mm. grams grams
22 116 days (loss 29%) m. 173 104.0 | 0.0316 | *(4)
ie 147 days (loss 30%) m. 188 162.0 0.0374 | *(4)
S 27 (loss 30%) m. 215 223.0 (219) 0.0424 | (4)
M 2 (loss 338%) m. 185 (?)| 170.0 (167) 0.0338 (4)
F 9.3 170 days (loss 338%) m. 179 129.0 0.0334 | *(4)
Rubee 151 days (loss 34%) m. 183 114.0 0.0422 | *(4)
Beane 130 days (loss 34%) m. 171 115.0 0.0280 | *(4)
J 1.4 94 days (loss 35% m. 180 107.0 0.0302 | (4)
S 25 (loss 35%) m. 205 202.0 (198) 0.0458 | (4)
F 9.2 170 days (loss 36%) m. 191 127.0 0.0374 | *(4)
S 26 (loss 39%) m. 205 174.0 (171) 0.0233 (4)
F 4.2 153 days (loss 43%) m. ie 105.0 0.0332 | *(4)
Heda 105 days (loss 45%) m. 185 110.0 - 0.0318 | *(4)
E. Adult chronic inanition

Jie5 117 days (loss 34%) m. 175 97.0 0.0240 (5)
M 3 (loss 34%) m. 175 125.0 (122) 0.0252 | (5)
M 12 (loss 36%) m. 173 128.0 (125) 0.0306 (5)
M5 : (loss 37%) m. 190 129.0 (127) 0.0320 (5)
M 6 (loss 37%) m. 175 138.0 (134) 0.0270 (5)
M 11 (loss 38% m. 190 163.0 (159) 0.0322 (5)

* The ‘F’ series, and a few others as indicated, were cut by frozen sections
for study in the fresh condition, or for lipoids, chromaffin reaction, ete. All
of the remaining glands were embedded and cut in paraffin sections.

1Rat J 1.7 had just given birth to a (first) litter. The remaining females
were virgins.

2 Of the stunted rats (table 1 B), St 80.9 and St 228.4 had practically doubled
their initial weight during the period of underfeeding. The others were held
nearly at maintenance (constant body weight) during the period indicated.

3 Of the rats refed after inanition (table 1 C), all excepting the last three
were refed after having been held nearly at maintenance from three to about twelve
weeks of age. S 33.118 and S 33.120 had been held with but slight increase in
body weight from 21 to 140 and 150 days of age, respectively, and were apparently
permanently stunted. Rat F 2.3 decreased from 151 grams to 104 grams (loss
of 31 per cent) during seven days of acute inanition, and was then refed as
indicated.

4 The rats subjected to acute inanition (table 1 D) had been given water only
for periods varying from seven to twelve days, with loss in body weight as indi-
cated. In four cases the age was unknown. The final body weights are given
in the table.

5 The rats with chronic inanition (table 1 E) had been fed gradually decreasing
amounts of food during a period of about five weeks, with losses in body weight
as indicated. The final body weights are given. The age was unknown in all

but the first rat.
284

‘
TABLE 2

Postnatal growth of volume in suprarenal cortex and medulla in the iho rat

SUPRA- |pERCENTAGE FORMED BY THE

eee supRa- | RENALS
2 1 : PER CENT
cata s | “AND SEX be WEIGHT none Cor tex! Medulla, average

ace |“VERAGE) \ picur Be (and range)

AVERAGE| ~~

A. Normal
I

grams grams | per cent oe per cent

20:

NOMA aY Ss... oo y= > oth 2 3 m | 13.7 |(0.0031)|(0.019)| 80.9) 19.1 (16. 8-22.3)
TIDE WIG NG ke (ee eae Df 15.5 | 0.0040 | 0.026 | 84.3] 15.7 (15.3-16.4)
UGE: Se cee ge ae 5 m | 27.9 | 0.0106 | 0.033 | 90.6} 9.4 (8.5-10.8)
ZAG Ey (stoi Gear ari See 2 £ 28.2 | 0.0078 | 0.028 | 89.5] 10.5 (10.0—11.0)
GR OLY GEER ora «ic svoles * 2 m | 77.0 | 0.0182 | 0.024 | 92.3) 7.7 (7.48.1)
Cia. G RNS. eee ee one 1 63.0 | 0.0116 | 0.018 | 82.7) 17.3
GYSVERO ES a eoaey hep ae 3 m |180.3 | 0.0218 | 0.012 | 92.8} 7.2 (5.8-8.0)
BTA IOMV Ss 5 neue f Sep30 3 f |132.6 | 0.0414 | 0.032 | 93.5} 6.5 (5.3-7.2)
94-346 days............- 8 m |220.1 | 0.0322 | 0.015 | 89.9) 10.1 (6-16)
112-253 pee ar eke: Sineeeie ee tenia wee A pte ee eA 4 f£ |161.1 |.0.0426 | 0.027 | 90.9} 9-1 (7.0-11.9)

B. det sk ee he be ee A Oe ie One Bebe 2 ee ae ee experiments

79.3} 20.6 (18.6-22. ert lis ed. | 2 2) foes heonasl| otnay| gala 206 cis.6-22.7)
90.2) 9.8

89.8) 10. 2 (8.7-12.5)

88.2} 11.8 (11.4-12.2)
89.3} 10.7

Underfed 1-12 days..... 8.9 | 0.0026 | 0.029

30.5 | 0.0124 | 0.040

f
Maintenance 21-51 days. m
a) 34.1 | 0.0131 | 0.041
m
m

2
1
Maintenance 3-10 weeks.| 7
1
a
1

Maintenance 21-82 days.

Maintenance 21-139 days 30.0 | 0.0161 | 0.054

mptiy Pires Ges Se wd. Tt; 1 m | 10.8 | 0.0024 | 0.025 | 74.5] 25.5 .
7 days Pep 10.8 | 0.0024 Oe ees 20.4 (18.6-21.7)

ated bweeks 1H) ps 6 £ | 57.6 | 0.0125 | 0.024 | 89.4) 10.6 (9.6-11.6)
2 s

See as as : oh 82.9 | 0.0171 | 0.622 | 90.4) 9.6 (8.8-11.1)

Refed 189 days..........| 2 £ |162.0 | 0.0396 | 0.024 | 85.7] 14.3 (13.4-15.3)

Refed 206 days.......... 2 m |229.0 | 0.0348 | 0.016 | 90.8} 9.2 (9.0-9.4)

Adult acute inanition...| 6 m |177.3 | 0.0393 | 0.024 | 90.1) 9.9 (5.8-12.8)

Adult chronic inanition.| 9 m|123.4 | 0.0278 }\0.023 | 88.7) 11.3 (8.1-13.1)

RES er ee
1'The ‘No.’ in the second column refers to number of glands, not animals.
The suprarenal weights were missing in one of the three normal males at ten
to fourteen days.
285

Cc. M. JACKSON

286

poinsvetl yOu o1aM 9UOzZ o[pprul ay} JO suOTyI0d JoUUT pue Ja4no 9y4

(00-¥F ZI) Z'9T
(9E-G'9T) ¢° Sz
(Z'$2-G ZZ) F SZ
(F'22-F 21) 6 '6T
(I 'FF-Z'9T) 2°62
(G°$8-8°61) 2°22

(9° 28-"2Z) 8°Ze

quao wad

Spon

(9" 2-8" 12) 2S]
(S'O8-O'FI) 2°22

(8°F2-81) T 1
(9° 8I-G°ZI) 9°ST
(6°68-L°91) 0°82
(€°$2-F OT) T6T

(2:21 SL) PS

quao sad

(Z 92-82) 9° FZ
(F'92-T SI) 6°9T
(8'OI-¥'8) 9°6
(9°S1-Z'L) F OT
(F'OI-S'°2) 0°6
(L°9T-O1) $°2T

(F'6-9) F'8

quad wad

auoz JoUUT

Ged Jouur) e[PPIA

(I°LT-O1) ¢°ST
(I’8I-¥'8) 9°ZI
(€°02-F'8) £°F1

(8°6-9'8) 2'6

(9°8-4'F) F'9
(G°ST-8'°¢) 9°8

(F'6-9) F'8

quaa wad

(J1B8Y 109NO) eTPPI

“Ajoyeredas

skep uw9e4Inoj 0} ouO Jo dnoas jeuso0u uy ,

(O°ST-2°S1) P21
(Z°SI-#'9) FSI
(8'OI-8'2) £°6

(PF L-1'°¢) 2°9
(8°ZI-6'8) 9°01
(O'ST-I' TT) Z'et

Co S1-L 2p) FoF

qu99 wad

9uoz I9jNQO

X@LYU00 IVNAUVUdOS JO SANOZ NI VWOULS UVTINOSVA JO ADVINGOUAd ADVUAAV

NHN

N

uw

UOTPIUBUT OTUOIYD YNpPYy
‘“ UOTZIUBUT ayn 4yNpYy
ee QOUBUOUTVUT
IoyJB SYooM OM} Pofoy
she Sas s0URvUO} UTR
Iayye yooM J] pejoy
Sot baat angi MON syaoa.
€ WOlJ ddUBUOZUTe PT

0} SyooM ¢ [BUIION
‘''“sKep pI-[ [BULON

SaNVID
40 Wau
-WON

NOILIGNOO GNV HOV

1.1 OUngyD 9Y}

fo pun)6 qouapudns ay} Ur DIINpIaU puv xa}109 ay) fo sauoz SNOLIDA OY} U1 S]aSSaa puD DUOLIS 94) fo sawnjoa 9a1VD)94 DBDLIAW

€ WTAViL

SUPRARENAL GLAND—EFFECTS OF INANITION 287-

TABLE 5

Mitoses in the suprarenal gland of the albino rat

NORE AVERAGE NUMBER OF MITOSES PER SECTION IN

AGE AND CONDITION BER OF
GLANDS| Outer
zone

Inner | Me- | Total average
zone dulla |; (and range)

Middle
zone

A. In normal rats

ISG SL OVET E.G ge le i ee a 3 11 2 1 6 | 20 (12-29)
ODEON pena mee ep oat es ae ameemetr on 3 5 2 4 3 | 10 (7-18)
LU Sa a Se Be 6 5 3 4- | 16 (14-18)

LES ee An eed 1 18 17 0 5 | 40

ives. (OEE S FS che. hehe ie a 2 10 10 3 6 | 27 (22-381)
=D) ee a ae 9 9 12 z 2 | 24 (10-41)
SPOS 72 dg ae 1 3 5 0 S | 11

i 94 dances aie n Sh see. sort es. 4 2 2 a 0 4 (2-5)
CAE ICr hs 2 1 1 0 0 2 (1-3)
CUE ee, fe a tee Seat 3 oe as a el 1 1 0 0 0 1

B. In test rats

Maintenance birth to 12 days...... 1 1 1 0 mS) 14.
Maintenance 3-10 weeks...........| 10! 3 4 0 0 1 (0-8)
Maintenance 3-12 to 20 weeks..... 3 0 0 0 0 0

Refed 1 week........... Bs 23 =, 2 SS: 3 3 1 0 3 2 (1-3)
ELCONAtWEEKS este ot ccc en 2 3 5 0 2 LO“ (712)
PEI suG MUVGHT Ree ceteris Ss 3 553 1 0 0 0 0 0

1 Six of the ten held at maintenance from three to ten weeks of age showed no
mitoses in the sections examined.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 3

JACKSON

M.

C.

288

6'Z116 61S 6LI6 IIS 2% 19 $C P'S 19 eG Z‘SLlis 6 r'6 Cre TLR a op Ss wre bp k eee apa ee © Bele bale WA one 5 BAe Be dnoid OJ aSBIIAY
OT Scl2 CTS ST SalGaoy ea 2S (Sr Oe |S SaiGeeiG S Se GrSselio® le a ee SSSA Payoy SP 1148
GealiG a0 Feel Omen OscG GeOn ance PeGe|OROe GaGa Omnn | Ge GulZvGiel Fark) Situimecma 6) ke 8c ages gas Syo9M GZ pofoy 93° OT 48
ZAOUSROL OSE Seen Setaizeas |OncaCeGs OVOe Pes Orel OnOslzeGeSeGhiGne | 9-0 6 op 8 4 8 er oa 90M T pofoy 16°01 48
GVCT GEST Sale| Seow Geach eon ec alike Ge |S Cale GaIGe GL Gree Ges soa LCi (emt ey ce eek ees as Y9oM T pojoy cy Il 48
re Died T' £216 912 Salt '9 6'P Cae 19 eG FILLS 68 Tet Tess tb, ete a eis sutmathes O- 5 6 md) ae) See evele ais wiele, sisle ere dnoid IOJ IsBIOAV
8° LTE 1219 02/2 S1I9'ZZ|6'S IZ'P I'S |8°S IFS |P OLB 2 |6'8 I OTI8' 8 | SABP GE[-1Z doUBUOZUTBI, 8°88 48
CHOICES CIS ROGSecnlOsPraocONG: Piz GeiG ZaiSeralmelL2 2jOeea\Oeenis 2 | 0 9. = 6 skvp Z8-1Z eouvuUszUIBI, OS ZI 48
SECTIO SGlOeZclOeZLOLOS|S OFlheS |9eG |8°S S°G 9 TL6 8 9 Se-ORS 44 skup QJ-1Z eouvuszUIB], LEsZeS8
Peet |Saetl tani GeGal Zeon el OeGelG: GalPaGalG Gillie GuselO OLS 8) limes See es skevp QJ-[Z vouvUeJUTIBIT CEL Ss
O'LZT/9 OZIL §Z/6 FLIS SZiP'9 12'S |Z°S |8'°S |S'S |S TII8'8 |F 6 16 OTI9'8 Je sABp 99-1Z ddUBUSZUTBI] €' Ly 48
ORGE aL GC GaG ROL GeGchorsel Gea GiOalGcQuiGeG GsGT 2 SaiGeTTSePTiOhG |) es he = 8 0 wok [ [BULLION LIT'S&$
CSCTGAOn Gar Zep SeoG real) (P29: (GeO. (ra |QaekicnOliGrcklOmeicusei: > °° © We.) get Gere es SYOOM (QT [BULLION 9°17 48
CE OLIGELCIGE LIS Aa Sead Pam ORG he Onl Ge AnSaG ISucLiGns: (Ge DEMePtiOsGal ete) 28 ey SOOM OT [BULIO Ny C1748
OETGIGRCC ROG ORG |SEGe Sno Ouse GON CaO Gi Sa Li GaGa OD QiaG |S) ile 8 ee Be a ee SyooM § [BUION Galet:
GEGEN GUL G| GEOG Ga oe ORM Sea a GeOn GiOuGEGel PaO TiswOnlemOUiZOnlZ snl "m0 0 8 1 yee 205 iene SyooM § [BUILON fe
DREZOSOC Swe TONG SecGen Go lsecn|9 GiGi lo OLig Gels OL Ga Zclee= 5 S88 SY9OM G [BUILON Sele
BaeTie Cri pEie dale colo ose 29 210 Ono (2 Oli 8 10 Gt GZ Or) ee eee YooM [T [BULLION 6 GL48
OOF Se CI CeRG ees Oral on i On ek GeOelpeGe iG. 2 2-6. |Go6 |Geelkedalees. 6 (coho ag a ULOGMOT. [BULIO KT GL Te
ya pa vee ea pa df W W nf nf rf rf n rT] nf
“glee let-olo lel oled-o-| of & lel sel oto.
e | § ial |B |eBi la .
® 9uoz ® & 9u0z ® 5 9uoz 5 aDV ANY NOILIGNOO ON Lyu
eIPPIN 9e[PPIN 9eIPPUN 3 j
40 duntox Eaves RGECHAN EO RHW AY CH AMEM RAISE AGI TER HEREIN

1D 0U29Q)0 ay? fo Dwhyouasnd poussvudns ay) UL Lajanu puD 7/20 fo azIs aarjD)aL pun aynjosqy

: yA TAVL

289

EFFECTS OF INANITION

SUPRARENAL GLAND

€’OLIS SliF 6LIF ZLI9 ZzlS 9 6'P Te e°9 PS L°e119'8 8°8 G'IL|9'8 S, Oe, DSM ee a6 ms) 876 e's, Fine Gkenn s © 0, elas ere eee ewer dnoi3d IO ISVIIAYV

TO1[e9T/4'E1l0'ET/9'SI|F'9 |8"F |2'F [8°¢ Jo's jO'SI|L"8 |L'6 |G T1/8°6 | Sone ae UOTPUBUT OTUOLYD 4NPY € W
b'O1/0'0z|0°G|L"TZ|g9e|¢"9 [6h |h°S [2'9 [T'S |8"ET|h'S [G°S |PITI6 2 | Tore MOTYTUBUT OTOL NPY Gift
8°91 C°Z2\2 121% 91 T'2210°2 Ga, GAG G’9 Q°¢ ray C'3 Z'6 0 Z116'8 COCO Chr a} Ae O eud Pere {e © 6, 5..0nR enmOnN © \@0e "a> Sa) (e * dnoas3 OJ VSVIIAY

E812 OZ|S 6I/L STI6"E|Z"9 T'S 2's le'9 [eS |Z T1Z°8 \6°6 [2TTh's fo Rane 3x “UOTTUBUT ONT 4NPY 9% S
0°FT/8 0Z/6 0g|z'ZT/8 'Sziz"2 |P's |g°¢ jo9 |o'g |S-er|E'6 [eG |o'etl4's fo UOTJTUBUT OyNOv 4INPY ZW
8°Zolh FZ|T €Z|6 OzlO"GElz"2 \o'S e's |2°9 [oo [2110'S |4°8 leITA's fo uOTUBUT 9ynow 4INPY PIL
6°I1|9°F2/9'ez|0'9T|Tezl2'9 |Z°¢ los |P'9 |T'9 [Z'eTIe's |O'G 8II|6'6 | uOTPTUBUL OyNOT 4/NPY ee

Resumido por el autor, John C. Donaldson. |

El volumen relativo de la corteza y médula en la glandula adrenal
de la rata albina.

Las medidas que han servido de base al presente estudio han
sido tomadas en las glindulas adrenales de diez y siete ratas
albinas, fijadas en solucién de Bouin y después reconstruidas por
medio de cortes seriados. La adrenal izquierda es, en general,
mas grande que la del lado derecho. El volumen relativo que
ocupa el tejido medular decrece desde préximamente el 12 por
ciento en los machos y el 10 por ciento en las hembras, ambos
recién nacidos, hasta llegar a ser el 7.5 por ciento y 6.5 por ciento,
respectivamente, hacia el periodo de la pubertad. Desde este
momento en adelante hay, relativamente, un cambio muy pe-
quenho. Cuando se comparan las glandulas de la hembra con las
del macho, en ejemplares del mismo peso, las de la primera con-
tienen relativamente menos médula.

Translation by José F. Nonidez
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MARCH 17

THE RELATIVE VOLUMES OF THE CORTEX AND
MEDULLA OF THE ADRENAL GLAND IN
THE ALBINO RAT

JOHN C. DONALDSON

Department of Anatomy, School of Medicine, University of Cincinnati

FOUR CHARTS

This study was undertaken at the suggestion of Dr. H. H.
Donaldson to determine the relations of the cortical and medul-
lary portions of the adrenal gland in the albino rat, M. nor-
vegicus albinus, and to demonstrate the changes in their volume
which take place with age.

In the albino rat at birth there is a fairly well-marked cortex
and medulla in the adrenals, though there are numerous groups
of medullary cells still scattered through the cortex. Within a
few days most of these have disappeared and the cortex and
medulla are sharply marked off from one another. There are
present in the albino rat, in addition to the tissues in the adrenal,
microscopic masses of chromaffin cells in the retroperitoneal
tissue, Fulk and Macleod (16), and a small but constant mass
of cortical cells in the epididymis of the testis, Swale Vincent
(12). No attempt was made to include these extra-adrenal
masses in this study.

The materials used were the adrenal glands from seventeen
rats, nine males and eight females, of Wistar Institute stock.
The glands were fixed in Bouin’s solution. After having been
stored in cedar oil, they were passed through xylol, embedded
in paraffin, and cut into serial sections 10 » thick. The sections
were stained with haematoxylin and eosin. This gives a very
sharp contrast between the cortex and the medulla, the latter
showing as a blue mass in the surrounding pink-stained cortex.
Every tenth section was projected with an enlargement of fifty

291

292 JOHN C. DONALDSON

diameters, and the outlines of the -entire section and of the
medulla were traced. The area of each of these figures was
measured with a planimeter. From these areas and the known
thickness of the section the actual volume of each section
traced and of its medulla was calculated. The nine sections
between each of those traced were considered as each having
the same volume as the nearest traced section. By adding
the volumes determined together the total volumes of the
gland and of the medulla were found. Blood sinuses and
blood-vessels were considered as a part of the tissues in which
they were found. In other words, no attempt was made to
subtract their volume from that of the gland. The thin capsule
was included in the area of most of the sections. In those
toward the end of the gland, where the capsule appeared thick
owing to the plane of section passing tangentially to the surface
of the gland, the capsule was omitted from the areas.

Somewhat similar methods were used by Elliot and Tuckett
(06) in their study of the adrenals of a number of mammals.
They examined the cat, the guinea-pig, and the rabbit, and found
in these animals that the relative volume of the medulla, com-
pared with the volume of the whole gland, decreased with age.
There were marked differences in the relative amount of med-
ullary tissue in the gland, the maximum being 14 per cent of
the entire gland, in a young guinea-pig, and the minimum 1 per
cent in an old one. The values for the cat and the rabbit fell
between these extremes. These authors give the value of 9 per
cent for the amount of medullary tissue in the gland of a male
rat of 120 grams, and 6 per cent for that in one of 190 grams.
They do not state the variety of rat. They admit that the method
they used would not give accurate results with so small a gland
as that of the rat. In his study of the adrenal of the rabbit,
Bertel Bager (’17) finds that the medulla forms 20 per cent of
the total volume of the gland at birth. At twelve months it
makes up a little less than 2 per cent. The ratio rises to 3 per
cent as the animal gets older. The medulla is relatively smaller
in the females.

VOLUMES, CORTEX AND MEDULLA, ADRENAL GLAND 293

TABLE 1

AGE |VOLUME OF BOTH GLANDS|THE RATIO OF THE PEPE RENCE OX
ESTIMATED VOLUME or |YOLUME BETWEEN

104.0 71 =: {17.60 | 1.00 |16.60
192.0 112. _|17.90 | 1.50 |16.40

OR etaiax: Bobs THOT DE eR THM MED ULNA ros Pee oe
BER WEIGHT eee LEFT ADRENAL.
Leyes Total dulla Cortex THE a ee H eae
en RRNA OO za60 (gn || pe ominellh’ percent Ont eter tent
1 M 4.6 1 0.825} 0.095} 0.730 33 4.0
2 F 5.9 2 1.042} 0.114} 0.928 10.2 20.0
3 M 17.0 14 2.010} 0.260} 1.750 13.0 4.0
4 F a0 16 2.130) 0.200) 1.930 9.1 11.0
5 M 26.0 29 ~ | 5.370} 0.580} 4.840 9.8 12.0
6 F 28.0 29 5.49 | 0.39 | 5.10 G2, 20.0
1h F BY (al) 35 7.35 |0.53' | 6.82 iad 15.0
8 F 39.0 35 7.55 0.61 | 6.94 8.0 8.0
9 M 39.0 38 7.46 | 0.60 | 6.86 8.0 —0.2
10 M 50.0 45 7.06 | 0.50 | 6.56 7.0 2.0
11 M 7A \0) 55 10.68 | 0.87 | 9.81 14.0
12 F 79.0 64 16.70 | 1.20 |15.50 11.0
13 M
F
M

8.2
6.9
103.0 69 {11.51 | 0.81 {10.70 7.0 —17.0
6.0
8.2

16 | Damaged in prepa-

ration
AT El e20SeOhN, 202) 1830) |) 2.2: | 20:8 6.6 12
18 Tumor of the adrenal

19 M | 300.0 | 2SO ye 2b 2 | S2e 2° O20 10.2 18.0

The values for the volume of the cortex were obtained by subtracting the
volume of the medulla from that of the total gland.

The results of my observations are shown in the accom-
panying table 1 and the four charts. All figures are for the sum
of the two adrenals of each rat. In this connection it is worth
noting that in fifteen out of the seventeen rats the left adrenal
was larger than the right; exceeding it in volume by an average
of about 10 per cent and in one case by as much as 20 percent.

Chart 1 shows the change in the volume of the cortical and
medullary material in the gland of the rat from birth to nine
months—5 to 300 grams. The values were gotten by finding
out what fraction of the whole gland was composed of medulla
and reducing that fraction to its percentage value. The values

294 JOHN C. DONALDSON

for the males and females have been plotted separately. It
will be seen that the curves for the two sexes are essentially
parallel, but that the one for the females, marked by crosses,
indicates a smaller relative value for the medulla, and at the end
of the female curve the relative volume becomes stationary.
Both curves show a rapid decrease in the relative volume of the
medulla until about the age of puberty, then in the males an in-

50 100 150 260 250 300

Chart 1 The change in volume of the medulla expressed as a percentage of
the whole adrenal gland. Dots, e, males, individual values. Crosses, +, females
individual values.

crease and in the females no further loss. Dr. Jackson, in some
recently completed work (Jackson, ’18), has found essentially
the same relative decrease in the chromaffin tissue in the rat. In
his rats, however, he did not find any differences between the
sexes. His results, as he has kindly told me, were not arranged.
to bring out any such differences which might exist.

The recent work of Elliot and Armour (11), and Lewis and
Pappenheimer (716) shows that in the human adrenal just after
birth there is a rapid degeneration and absorption of the inner

VOLUMES, CORTEX AND MEDULLA, ADRENAL GLAND 295

portion of the cortex. The resulting rapid reduction in the
relative size of the enormous cortex is much greater than any-
thing shown in chart 1, and is in the direction of a relative in-
crease of medullary tissue. This degeneration has never been
noted except in man. There is nothing resembling it in my
specimens.

VOLUME
CUBIC M.M.

BODY WEIGHT
GRAMS

30 !06 150 200 250 300

Chart 2 The growth in volume of the male adrenal cortex and medulla. The
lowest solid line represents the medulla. The broken line marked by crosses
represents the vakues for the medulla multiplied by ten, so that its growth can
more conveniently be compared with that of the cortex.

Charts 2 and 3 show, in solid lines, the growth curve, in volume,
for the cortex and for the medulla, chart 2 for the males and
chart 3 for the females. The broken line represents the values
for the medulla multiplied by ten; so that the relative growths
of the two parts can be more conveniently compared. The
values for both the cortex and medulla are expressed in cubic
millimeters.

Chart 4 shows the values for the whole of the adrenal in cubic
millimeters. When compared with the values for the weight of

296 JOHN C. DONALDSON

the adrenal found in table 71 of the rat (Donaldson,’15) and
shown in chart 20 of the same, it will be seen that the curves are
similar in form and relation, but that the values in chart 4 are
about one-half what they should be to satisfy the weights given
for the respective ages in table 71. Jackson (717) found a similar

VOLUME
CUBIC M.M

50 100 150 200 250 300

Chart 3 The growth of the female adrenal cortex and medulla. The lines
have the same significance as those in chart 2.

discrepancy between the volumes and the weights, when recon-
structing the hypophysis from serial sections. He suggests
that shrinkage by reagents and the weighing of a certain amount
of tissue which is not included when reconstructing the organ
explain this. Recent work at The Wistar Institute mdicates

VOLUMES, CORTEX AND MEDULLA, ADRENAL GLAND 297

that when specimens are embedded in paraffin directly from
cedar oil they may show a loss of as much as 41 per cent of their
fresh volume.

VOLUME
CUBIC M.M.

BODY WEIGHT
GRAMS

50 100 150 200 250 300

Chart 4 The growth of the whole gland expressed in cubic millimeters. The
upper line gives the values for the females, the lower line that for the males.
Dots, *, males, individual values. Crosses, +, females, individual values.

To sum up: The left adrenal in the albino rat is usually dis-
tinctly heavier than the right. The relative volume occupied
by the medullary tissue decreases rapidly from birth to puberty,
50 to 100 grams, and then remains stationary or increases a little

298 JOHN C. DONALDSON

with age. When compared with those of the male, the glands
of the female, body weight for body weight, contain relatively
less medulla. The difference, however, is slight, less than 2

per cent (chart 1).
Results found by the reconstruction method are comparable
with results of growth expressed by weight (chart 4).

LITERATURE CITED

Bacer, Bertret. 1917 Bidrag till binjurarnas 4ldersanatomi hos kaninen
(The anatomy of the adrenal in the rabbit as modified by age). Up-
sala Likareférenings Férhandlingar. Ny foljd. Bd. 23. H. I-2.

CrowrE AND Wistockr 1914 Experimental observations on the suprarenal
gland with special reference to the functions of their interrenal por-
tions. The Johns Hopkins Bull., no. 284, pp. 287-304

Donatpson H. H. 1915. The rat. Reference tables and data. Memoirs of
The Wistar Institute, no. 6, Philadelphia.

Exuiot, T. R., anp Armour, R. G. 1911 The development of the cortex in
the human suprarenal gland and its condition in hemicephaly. J.
Pathology and Bacteriology, vol 15, pp. 481-488.

Exuiot, T. R. anp Tuckett, I. 1906 The cortex and medulla of the adrenal.
J. Physiology, vol. 34, pp. 333-369.

Fuux, M. E., anp Macteop, J. J. 1916 Evidence that the active principle
of the retroperitoneal chromaphil tissue has the same physiological
action as the active principle of the suprarenal gland. Amer. J. of
Physiology, vol. 40, no. 1, pp. 21-29.

Jackson, C. M. 1917 Effects of inanition and refeeding upon the growth
and structure of the hypophysis in the albino rat. Amer. J. Anatomy,
vol» 21, mo. 2.

1918 The postnatal development of the suprarenal gland and the
effects of inanition upon its growth and structure in the albino rat.
(In the course of publication.)

Lewis, R. W., AND Parpenneimer, A. M. 1916 A study of the involutional
changes which occur in the adrenal cortex during infancy. J. of Med.
Research, vol. 34, pp. 81-93.

VINCENT, SwALE 1912 Internal secretion and fie ductless glands. Edward
Arnold. London.

1917 Experimental and clinical evidence as to the influence exerted
by the adrenal bodies upon the genital system. J. Surgery, Gynae-
cology, and Obstetrics, vol. 25, pp. 294-299.

b) rar aS
aN i euk i
ole re

fiver
a »
wi he | § t

oF
NG ‘ 4
‘
\

ss + ea dieee ete A wo i mio on ohana MELO ae

Bede! Pl) ig ke Cu A ba Dal ef 4 Rey ad
sige Aaah beni ‘Sea aah 1 bldanacets | cia
ae ; way anes | Fes 4! 7 a ea 4 I “ me a ‘ta Lae 5 et Fe
Aiea AA F ohpidis “ht Eada weet th Ba
thn 5) i ‘wht We] lab th ney a i f al) oebeigt otis : pts he m4 ; ty

riche iis i POUT ERY pyr + RANT

a ol a sib eke yy yt wagal BB ah » hs teat hil: i ‘ Paths)
wl eset: ib, enor Og awry ng Oy ODES.

. grat ty le 2 Be a uy a ey Bat wh ORIN ia get

7 es 4

“a Ligs
im P igac ‘ ‘ 5 1% | his :
pike ih Paci Tewata SPER Bes ase ahi ANTES rity, Tome it OG

a

et dat sbkthinty “vb ied “ols Siiwelel Sab sod: Kea OMIM, eh

‘ . or i Li ae et a ee
“ohh daateene igi bine dias toate 14 (0 ee eenhs

DAM ay Aven ec ons Tait {e otay een vine aa
fe ey epee 4 or 9 yr Oe y Eyat , dees ye hr oo
grind aa rs J BI Said Pe Tf PE Cree bk? tay eat Ghat Be TS ¥ eh: usr

r : a : r . fs ww aie
ae eye , a peaett | REWER Y) ih NUE IE dvsteeaesd AS) 44). CEES et ae) rf y al

4 : Dorn be : ove ‘ 5 / ) a)
LS Cre PERT 4 ‘esa MUM NP METER yy i) tr
pacts: oe ay Up BAI EE PRR OTS (CRED OE ae aT eee EES
fe yea : ; / :
Nae CATON PAT) Here Ae ees igtts EVEL MUTVTLATEE TD es C
‘ ve . ’
a -_ > ‘ae ; ’ A = ass ' ee ' aa eS ea c r o i Pew
ABT EN vs Bas 4 Ato AMOR: ; us) Res ict bun | ax hi 4 { rt, ¥ 3
‘ : , 7 * f } be
RAY sft Su, OD LIE (gt 1 Ray AS he RT ty gaa
LA Bee i] edey igi on CH Wore ky SH CaF; 28 as va ik if 5
i : ) J 4 : , ¥ i ” ‘ y , i" "
ihn aS ee Ae 4 tu, we 5 f Donate vi
at. ed ae { "he tne nel PH SAN es ft voir. Cane Lou 8 ; : ‘ i 4
‘ van us + .) <; ‘ Fy
A e ee ees : a ma ¥ be fa *
tab Seta WRC Re Or OR aeiea at OMT) SY ;
p wedhetgas' te) taba ih itt arp) SVT bind ALN iy ae bi
4 } 7 A res) ya ,, i .
: 7 ite FR 1 ' : a, % Sn Vee) ee ahi } .
Ht aia Ck Vp nat Leo) Ville SR Paks heya dath
! ‘ F ‘i bs — J i Pee 7 Me ee ‘ e; « i
Var eect i) aii np dale! it Sing or ib CONE) 20 rs | it :
¥ S :

a i. ide eteidlad cit ama Wi jena” abr Aa Mime aes DMI Ares tig 18
mempatlte Prenat «cbbtaviaha nny Kt ABE hp} sii ATI CTT bs iia
Tae. ity; wed Ng

Se iyi aia abet Hens) ree is MURR RAH 20! 08 gx > ec
‘ win Wik olan ise. me ‘aha sali vo Barbe et ay |

ih ul’

hs he : “f if " ie 2, ett LG ,
CL he Lise J Beit A D ive hn nha 7 ee om i
} ia ihe i,

reat i os ;

ERR TD. Brel ti

J
i
J A
‘ \
i
; \
i ;
{ ,
¥ 4
t , ,
ra i oa F
bia eae OT
‘ A
oy *
f , eS
= 1 7 a
x ) het i) a7
; Pr
- et, i
Wn ; f il’,
i vid) '
{
‘ , Ue oa eat
j - LJ
i Wy (anes ie

Resumido por el autor, Franklin P. Johnson.

El desarrollo de los l6bulos del higado en el cerdo.

En el presente trabajo el autor se ocupa principalmente del
desarrollo tardio del higado del cerdo, considerando especial-
mente la formacién de los tabiques interlobulares formados por
tejido conectivo, el modo de formacién de nuevos lébulos, el
crecimiento de las venas hepatica y porta y el desarrollo del
higado en conjunto. Los resultados obtenidos pueden resumirse
del siguiente modo: los tabiques de tejido conectivo se distin-
guen ya en el animal recién nacido, pero no se hacen bien patentes -
hasta que el animal tiene dos meses de edad. Se originan como
extensiones del tejido conectivo que rodea a la vena porta. La
excisiOn de los lébulos origina otros nuevos. Durante este pro-
ceso la vena central del l6bulo o bien se bifurca o produce una
nueva rama en uno de sus lados, y siguiendo su trayecto aparece
un nuevo tabique intralobular de tejido conectivo que se extiende
formando angulo recto con la superficie del l6bulo, entre las dos
nuevas venas centrales. Cuando se completa la excisién resultan
dos nuevos lébulos provisto cada uno de ellos de una vena cen-
tral, mientras que el tabique de tejido conectivo y la antigua vena
central adoptan posicién interlobular. Los tabiques de tejido
conectivo desempefan tan solo un papel pasivo en la formacién
de nuevos lébulos. El crecimiento del higado durante todos los
estados de su desarrollo tiene lugar: 1) por el aumento del nit-
mero de lébulos; 2) por el aumento de tamajio de estos ultimos.
En general todas las partes del higado crecen simultaneamente y
de una manera semejante. Hay un cambio constante en la posi-
cién de los lébulos en la periferia, pero este cambio se lleva a
cabo de tal modo que las relaciones de un l6bulo con otro sufren
una alteracién minima.

Translation by José F. Nonidez
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, APRIL 7

THE DEVELOPMENT OF THE LOBULE OF THE PIG’S
LIVER

FRANKLIN PARADISE JOHNSON

Department of Anatomy, University of Missouri

TWENTY-EIGHT FIGURES

CONTENTS
LOSS TL Pe ghe B a SRS Sane een A ee ON Ree, rere ETO Tee a eee eee 299
Development of the connective tissue septa...................0 2 eee eee ee 301
CHEESES VES el See get a ae Neat nse dnd « bh DAEs Seay iat Ape ae at 2 lied 308
HOLM GION OMmneWPLODULES Hs cine o.cteoeee aie ook oe eee elo ae ACs cae Sete 308
Rilevof the/ connective tissue septa: :.:... 224 anh elas d. sie. ate 2 oda rinen 318
PmeeaRMLRMRE TEI UOT Sikes oh ete sade Ova a Was dei nines DEA Ro alngn are 5 cue Sty ae 319
UD TL SVE Cis ree ii inde as fe ec eee mete Mon op 1 bee onl cae), eda 325
INTRODUCTION

The development of the liver lobules offers a difficult problem
owing to the fact that in most animals the lobules have no defi-
nite boundaries, one running directly into the other without
demarcation. In section, therefore, the livers appear to con-
sist of solid masses of parenchyma, pierced at more or less reg-
ular and alternating intervals by various-sized branches of the
portal and hepatic veins. This lack of lobule definition is so
great that one readily appreciates why Weber, in 1842, denied
the presence of true lobules in the human liver, and one may
himself doubt the appropriateness of the term ‘lobule’ for either
the portal or hepatic units.

Early recognizing this difficulty, I chose for material the
liver of the pig, for I hoped that by using an animal in which
the liver lobules are definitely marked out, this difficulty would
be greatly overcome. I soon found that the pig’s liver does
not show indications of dividing septa until about birth, and
that the connective tissue septa are not definite until about
the second month of postnatal life. Contrary to the state-

299

300 FRANKLIN PARADISE JOHNSON

ment of Mall (06), lobule formation is not complete at this
time, for I find with Lewis (712) that ‘‘the multiplication of
lobules continues long after birth.”’ The few late stages of the
pig which I have been able to procure, consequently, have
been of great value in furnishing evidence concerning the de-
velopment of the hepatic lobules.

Because of the importance of the connective tissue septa,
I have found it advantageous to divide the development of the
pig’s liver into two definite periods—one before and one after the
septa are indicated. I shall speak of the former simply as the
‘early stages,’ of the latter as the ‘late stages.’ The early stages
include those up to but exclusive of an embryo of 254-mm.
in length; the later stages include the 254-mm. embryo and
extend to the adult.

In attempting to determine the manner in which the units
of the liver multiply, I first gathered together a series of se-
lected stages of developing livers. But so far as lobule forma-
tion is concerned, this was unnecessary. I agree with Mall,
(06) when he says, ‘‘the great difficulty is to recognize the
same thing from step to step,’”’ but I find the greatest difficulty
is to recognize the limits of a lobule in its three dimensions in
any single early stage. In the later stages, however, because
of the presence of the connective tissue septa, the lobules are
definitely bounded. Since in any developing liver the lobules
present numerous instances of every stage of development,
it is possible by a little study to arrange the various stages in
their proper sequences. In this manner the development of
the liver lobules may be easily and most satisfactorily deter-
mined from any single late stage.

Inasmuch as I have worked out the development of the liver
lobules from stages in which the connective tissue septa are
present and continually growing and delimiting new lobules, I
found it necessary to study first the origin of the septa. This
was the more essential, for an understanding of the different de-
velopmental stages of the septa makes it possible not only to
recognize dividing lobules, but to distinguish between the newly
formed and the old lobules.

DEVELOPMENT OF LOBULE OF PIG’S LIVER 301

DEVELOPMENT OF THE CONNECTIVE TISSUE SEPTA

In stages until nearly birth the liver of the pig shows no indica-
tions of connective tissue septa. The parenchyma is made up of
cells not greatly different from those of the adult. The sinu-
solids, which appear proportionately large, are lined with en-
dothelial cells. The ‘Gitterfasern’ or ‘reticulum of Mall’ (’96)
is demonstratable following Bielschowsky’s silver-impregnation
method, and is also slightly discernible after staining with
Mallory’s triple connective tissue stain. The branches of the
hepatic and portal veins interdigitate with one another and
furnish a means by which hepatic and portal lobules may in
certain areas be roughly outlined. The large branches of the
portal veins are readily distinguished from the hepatic, for they
are accompanied, as in the adult, by branches of the hepatic
artery and of the bile duct; some of the smaller veins are deter-
minable only with the aid of serial sections.

In figure 1 is shown a longitudinal section through a branch
of the portal vein, taken from the liver of a pig 229-mm. in
~length. The interdigitation of its branches with those of the
hepatic vein is clearly shown. Surrounding the larger branches
of the portal veins, bile ducts, and hepatic arteries is connective
tissue. Staining with Mallory’s triple connective tissue stain
clearly demostrates in it the presence of collagen fibrils. Where
the collagen fibrils are in contact with the parenchyma they
often send short fibrils into its reticulum. Surrounding the
hepatic veins there is but a thin layer of connective tissue,
which, moreover, does not extend out to as small branches as
does that surrounding the portal veins. The hepatic cells
are polyhedral in shape, with slightly granular protoplasm and
distinct chromatic nuclei. They are grouped in strands, but,
as stated by Theopold, a radial arrangement about the central
vein is not yet to be found. Here and there are to be seen small
clusters of nucleated round cells. These cells, which we know
to be developing blood-cells, were first described by Ko6lliker
C79), but were thought by Toldt and Zuckerkandl (’75) to be
young hepatic cells. That Kolliker’s interpretation is correct,

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 3

Do

°
a

FRANKLIN PARADISE JOHNSON

302

Mallory’s triple connective-

. in length.

ig 229 mm

ion of liver of p

1 Sect

o
>

1

2 3
_
fe) 3
2S
m5 §
Nee,
=
on —_
a ae
aoe
ple) J}
ae 2
ae
ace
Oo S @
(Slips
ee
“nm
See
Soe
n a
BS
oo :
> ei {|
Fane
So
- Se
Sy free
Sieh
io)
~ of
A, o0°S
ere
pees S
oe ee
P Rex
Sy ios!
ile ae
~~
Sox .
oo Ue
oO
5 § 0
~ ed
Mao x
seus Eace
geet
Ea )ten ae
nis aa
2 -~.e
oe (cl gS
25 oe
Pome) °

DEVELOPMENT OF LOBULE OF PIG’S LIVER 303

however, has been abundantly proved by the researches of Van
der Stricht (’91); Kostanecki (92), Engel (99), Lobenhoffer
(08), and others. Large mononuclear giant-cells, also de-
scribed by Kdlliker, are to be found sparsely scattered through
sections of the liver. In the later fetal stages both the group
of embryonic blood-cells and the giant-cells become fewer and
fewer, as is agreed by all investigators. I find with Theopold,
however, that they have not all disappeared at birth and that
a few may be found for several weeks after birth.

The first indication of a segmentation of the liver paren-
chyma into hepatic lobules which I have found has been in an
embryo of 254-mm. in length. According to Engel (’99), this
pig would be one at about full term, since he states that a pig
at birth measures 25 em. Theopold (10), however, shows that
pigs at full term vary considerably in length; that they even
attain a length of 35 em. before birth. In the 254-mm. speci-
men, the liver cells appear more vesicular, their protoplasm
stains but faintly. The boundaries of the hepatic lobules as
seen in figure 2 are in many places definitely marked. A close
analysis of the dividing medium, however, shows that it con-
tains no collagen fibrils, but rather its appearance is due to a
change in the parenchyma. A higher-power drawing of this
stage is shown in figure 3. To the right and left of the draw-
ing are shown terminal hepatic on central veins; above and
below are terminal portal veins. Extending from portal vein
to portal vein is shown the first evidence of an interlobular
septum. It consists of hepatic cells which are more coarsely
granular and stain more deeply than the remaining cells; similar
cells are found surrounding the portal veins. However, the
thing which makes the septum most apparent is the arrange-
ment of the cells. As seen in sections, they form two more
or less regular parallel rows, an arrangement which becomes
greatly accentuated in slightly older stages. Another factor
which in places intensifies the distinctness of the septa is the
presence of small branches of the portal veins. Nucleated
blood-cells found in the sinusoids along the septa also aid in
making it more distinct. I have studied carefully the retic-

304 FRANKLIN PARADISE JOHNSON

<a fo 2,

aon

a8

5
ee
0
eb
b Oe

me

Bt

Fig. 3 Higher magnifications of portion of figure 2, showing interlobular
septum. H, hepatic vein; P, portal vein. XX 260. ,

Fig. 4 Section of liver of a pig four days old, showing arrangment of lobule
border cells in parallel layers. H, hepatic veins; P, portal veins. X 50.

DEVELOPMENT OF LOBULE OF PIG’S LIVER 305

ulum of the septum stained by the Bielsechowsky method and
believe that a thickening of it at this stage is very doubtful.

In a slightly longer pig measuring 265-mm. (also unborn)
a somewhat similar condition is found, although it is not quite
so far advanced as the 254-mm. specimen. This variation
in the degree of development I have found to be especially
marked in stages around birth. Thus, in specimens of two
hours and twenty-four hours after birth, the septa appear no
more distinct than those of the 265-mm. stage, while a speci-
men of three days is no further advanced than a 254-mm.
embryo. A similar varation in the degree of development is
noted by Theopold (10). Again a radial arrangement of the
hepatic cells is absent.

In a pig of four days the liver shows a more advanced con-
dition (fig. 4). The cells are coarsely granular and stain deeply.
The lobules are in most places definitely marked out by the
arrangement of their border cells. This consists in their form-
ing continuous sheets of cells about each lobule; thus, when
seen in cross-section, they form two parallel rows. Between
these rows of cells is found reticulum which is slightly thick-
ened, but contains no collagen fibrils. A Berlin-blue-gelatin
injection of the liver of a second pig of four days also shows
definite evidence of the lobular nature of the liver. As shown
in figure 5, the injection mass has passed through the sinu-
soids into the central veins. The sinusoids between the lobules,
although present and forming free anastomoses between one
lobule and another, are smaller and somewhat less numerous
than those within the middle of the lobules. This condition
results in the appearance shown in the figure, the lobules being
separated from one another by a clear translucent zone—an
appearance which is more strikingly brought out in the thick
sections.

For the first time can a radial arrangement of the liver cells
be seen. This point is shown in the injected specimen, since
the direction of the trabeculae of liver cells corresponds with
the direction of the sinusoids. Although the radial arrange-
ment is not well marked at this stage, it becomes increasingly

306 FRANKLIN PARADISE JOHNSON

more definite in livers of pigs of one, two, and three weeks. At
two months it is as strongly marked as in the adult. These
observations are in accord with those of Theopold (10), who
states that the radial arrangement of the liver cells is first seen
in pigs in the second half of the first week of postnatal life, and
that the arrangement gradually increases in definition, although
at the sixth week it is not strongly marked in all lobules.

In pigs of three and four weeks old the hepatic lobules are
still more strongly marked. The parallel rows of cells with the
intervening layer of reticulum are still in evidence. In addi-
tion, however, the reticulum in many places shows the addi-

4

4

iS)

Fig. 5 Portal injection of liver of a pig four days old, showing fewer and
smaller sinusoids along interlobular septa. X 40.

tion of delicate strands of collagen fibers (fig. 6), as demonstrated
by staining by both Mallory’s and Bielschowsky’s methods.
In certain places these fibers extend from portal vein to portal
vein;.in other places they extend from one portal vein toward
another, leaving, however, the nodal-point region devoid of
them. From various degrees of this condition it is safe to con-
clude that the connective tissue of the interlobular septa has
its orgin in the connective tissue of the portal canals and grows
outward from them, pushing its way into the reticulum already
separating the lobules. According to Theopold (710), com-
plete septa arising from connective tissue already present are
in evidence, in certain places in the liver of a pig eight days

DEVELOPMENT OF LOBULE OF PIG’S LIVER 307

old. He does not state, however, whether or not this con-
nective tissue contains true collagen fibrils.

Fig. 6 Interlobular septa of liver of a pig three weeks old. Mallory’s triple
connective-tissue stain. 250.
Fig. 7 Same, two months old pig.

The exact manner in which the connective tissue of the septa
is laid down, that is, whether the collagen fibrils have their
origin from fibroblasts situated in the portal canals and push
their way out into the interlobular reticulum or whether the
fibroblasts first migrate and give rise to the collagen fibrils

308 FRANKLIN PARADISE JOHNSON

in situ, I have not determined. This point, it seems to me, is
of little moment, inasmuch as in either case the origin is from the
connective tissue of the portal canals. It is interesting to note
that the connective tissue surrounding the sublobular veins
does not take part in the formation of the septa to nearly so
great an extent as does that of the portal canals. Around
some of the larger sublobular veins a few out-pushing collagen
fibrils may be found, but it is evident that the bulk of the con-
nective tissue forming the septa springs from the portal canals.

Capsula fibrosa (Glissoni). Collagen fibrils are differentiated
in the pig’s liver as early as the 80-mm. stage. They are found
principally in the portal canals, and to a less extent around the
larger sublobular veins, being especially abundant in the region
of the porta hepatis. Extending out in either direction from
this region, as seen in the cross-section, there are thin strands
of collagen fibrils which spread around, but do not entirely
encircle the liver. These strands represent the forming external
capsule of Glisson. A similar section of the liver of a pig 111
mm. in length, shows a complete capsule. It consists, however,
of a very delicate layer of collagen fibrils which lies close to the
liver parenchyma. In all stages up to birth the capsule has
this appearance, the collagen fibrils showing no appreciable
thickening; on the other hand, in certain stages it appears to
be less distinct than in the younger ones; in some its presence
as a distinct layer is extremely doubtful. After birth it rapidly
becomes thicker and continues to augment in strength and den-
sity until the adult condition is reached.

FORMATION OF NEW LOBULES

Mall (’06), in his study of the structural unit of the liver,
concluded that Thoma’s laws concerning the formation of new
blood-vessels explain the formation of new lobules. He states:

As the vessels grow the liver tissue increases In quantity, but the
liver lobules do not increase in size indefinitely, because Thoma’s
first law is constantly at work and will soon break up the larger lobules
into a number of smaller ones. In all cases the length of the capil-
laries remains constant, and when they appear to be too long and too

DEVELOPMENT OF LOBULE OF PIG’S LIVER 309

numerous, it is always found that some of them have already turned
into small veins and thus mark the beginning of new lobules or of new
portal units.

While in general the idea expressed in the above statement
may be correct, nevertheless it should be emphasized that the
sinusoids of the liver are not all of equal length. Mall, him-
self, shows that within a single lobule they are of various lengths,
that. there are short paths and longer ones between the portal
and hepatic veins. Yet he believes that all the capillaries of
the lobule and of the liver as a whole are equally favored by
the circulation. In a former paper (’18 a) I have shown that
in the adult pig the lobules vary greatly in diameter; the lengths
and numbers of their capillaries, consequently, must also vary.
Again, in my study of the development of the lobules, I find
that the lobules do not reach a certain constant size before
they begin to divide into new ones; a large lobule and a smaller
one placed side by side may be in a similar stage of division.

In his recent studies on the growth of the blood-vessels in
the tail of the frog larva, Clark (’18) has shown that as certain
areas of tissue increase in size, the surrounding capillaries send
into the area endothelial sprouts which develop into additional
capillaries. When these new capillaries become functional, that
is, when a flow of blood through them is inaugurated, there
results an increased flow through and an increase in the size of
certain ones of them which are favorably located. According to
Clark, it is the increased amount of blood flowing through these
vessels which causes them to enlarge. Conversely, Clark has
shown that when the flow through capillaries is diminished, the
capillaries become smaller, when the flow stops altogether, they
close, break in two, and are gradually retracted into the vessels
with which they were originally connected.

The application of these observations to the growing liver
may be made as follows: As the hepatic cells increase in number,
new capillaries are formed, bringing about an increased flow
of blood in, and a subsequent increase in size of, certain capil-
laries leading to and from the new ones. These enlarged cap-
illaries become new branches of the portal and hepatic veins.

310 FRANKLIN PARADISE JOHNSON

As the portal and hepatic veins increase in length, the blood
flow through certain capillaries along their sides is diminished
and finally stopped, with the result that these capillaries dis-
appear through retraction. Thus it is that capillaries arise
from only the terminal branches of the hepatic and portal veins.

It was Mall’s idea (06) that new lobules form by the split-
ting up of old ones, yet he does not make clear the exact manner
in which the splitting takes place. Apparently, he believed
that the lobules split and become fragmented into a number
of parts, each of which is capable of giving rise to a new lobule.
Furthermore, he believed that the fragmented parts of several
different lobules are capable of uniting into a single lobule, for
on page 278 he describes a lobule as arising from three adjoin-
ing ones. Yet his diagrams illustrating the formation of new
lobules (figs. 41 to 43) are not suggestive of any such ‘shattering’
and reuniting of parts of lobules, but rather of an accumula-
tion of orderly binary fissions; if we should add other figures
to his diagrams to fill in the wide gaps which he has left, we
should have similar pictures to those which I have shown in
figures 11 to 15.

As stated above, the manner in which the hepatic lobules of
the pig develop may be understood from the study of any single
stage of development. However, it is only in the later stages
where lobule boundaries are made definite through the ingrowth
of connective tissue that the process can be determined with
any degree of accuracy. I have accordingly used for the
following description a liver from a pig two months old. The
observations recorded have been checked with other stages,
both before and after the connective tissue septa have made
their appearance; so far as I can determine, the formation of
lobules is the same in all.

The development of new hepatic lobules can be best under-
stood by watching the changes which take place in the grow-
ing hepatic veins. Each of their new branches marks the begin-
ning of a new hepatic lobule. The portal veins spread over the
surfaces of or between the lobules, dividing repeatedly, their
terminal branches being more numerous than those of the

DEVELOPMENT OF LOBULE OF PIG’S LIVER pied

hepatic vein, as pointed out by Mall. In their growth they
bear a certain constant relation to the growing hepatic veins,
through which relation a constant flow of blood is maintained.

In figures 8 to 10 are shown hepatic lobules in successive
stages of development. Figure 8 is a section of a slightly elon-
gated lobule with a bifurcating vein. Although in all other
respects this lobule is similar to other single lobules, it can be
definitely stated, because of the branching of its hepatic vein,
that it has begun to divide. In figure 9 the new branches of
the central vein are longer, having kept pace with the length-
ening lobule. The new veins grow, as described by Mall, by
the enlargement of certain sinusoids. In addition to the new
veins is to be seen the beginning of a cleavage of the lobule.
This is represented by the arrangement of certain hepatic cells
into parallel rows, between which collagen fibrils later appear.
It is to be noted that the septum occupies the plane of the orig-
inal central vein and bisects approximately the angle formed
by its two new branches. The free edges of the septum lie
just above the fork of the central vein, which position (fig. 9)
it maintains for some time.

That the new central veins lengthen in many instances at
the expense of the old one, that is, by cleavage of the latter,
there is little doubt. In those lobules in which the central
vein is just beginning to bifurcate it is relatively long; in those
in which the branches are long the old central vein is usually
shorter. This apparent splitting of the old central vein seems
to be caused by the growth of the septum into the fork of
the central vein. In many instances where the cleavage of a
lobule is nearly complete the old central vein is split down
almost to the point where it enters the lobule.

The further separation of the lobule into two new ones takes
place slowly by the continued growth of the new septum from
the sides of the lobule toward the old central vein. When the
septum is complete, the remaining portion of the old central
vein is found to he between the two new lobules; it may then
be spoken of as a small sublobular vein.

ae FRANKLIN PARADISE JOHNSON

Fig. 8 Dividing lobule from liver of pig two months old. H, hepatic vein;

P, portal:vein. X 50.
Fig. 9 Dividing lobule from liver of pig two months old. Ci and C2, branches

of central vein; S, new interlobular septum. X 50.
Fig. 10 Dividing lobule from liver of pig two months old. (Ci, old central

vein; Cz, new central vein; S, new interlobular septum. X 50.

DEVELOPMENT OF LOBULE OF PIG’S LIVER ol3

Usually, however, before the time the connective tissue sep-
tum is complete, the new lobules themselves have begun to
divide, thus giving rise to what Kiernan (’33) has described
as compound lobules. That the compound lobules of the adult
are, as I have stated before (’18 a), incompletely divided lobules
due to incompletely developed connective tissue septa is quite
apparent; they are similar in every detail to those found in
the developmental stages.

Compound lobules have also been described and modeled
by Debeyre (10) who states that in the pig as well as in other
mammals they are more numerous than the single lobules.

“Le petit lobule classique, isolable, existe, mais il est presque
exceptionnel.”’

While I am able to confirm Debeyre in that the compound
lobules of the pig’s liver are numerous, I have not found the
single lobule to be exceptional. If blocks of pig’s liver are treated
with 50 per cent to 75 per cent hydrochlorie acid (Johnson,’18),
the connective-tissue septa are destroyed and the lobules fall
apart. The compound lobules, joined together by liver paren-
chyma, are preserved intact. Examination of the lobules
under the binocular microscope shows that many of the lobules
are single. It is true that the compound lobules may be torn
into their component parts, but this does not happen if they
are handled carefully. If the maceration is allowed to proceed
to the proper degree, the lobules separate from one another
on a very slight amount of shaking. The compound lobules
ean be divided only upon rough handling. From a number
of such preparations I have observed that the compound lobules
vary greatly in number, size, and number of component parts
in the livers of different adult pigs.

In addition to the above-described process of division,
there is another which frequently takes place. Instead of the
distal end of the hepatic vein bifurcating, a new branch is given
off from its side. The new branch, as shown in figure 10, enters
a portion of the lobule. As it gradually becomes larger and
longer a cleavage takes place in the lobule in a-similar manner
to that described above. The cleavage divides the original

314 FRANKLIN PARADISE JOHNSON

lobule into two new ones, one of which is supplied by the old
central vein, the other by the new one. That new branches
of both the hepatic and portal veins arise as outgrowths from
the main trunks is stated by Mall.

Surface lobules often divide by the combination of the two
processes just described. The long axes of such lobules are
placed at right angles to the surface of the liver. When a sur-
face lobule divides its central vein bifurcates, the new branches
spread apart and then bend to become perpendicular to the
surface (fig. 23). A septum then starts from the surface, mid-
way between the two new branches, and grows down to the
fork of the central vein. Thus two elongated lobules are marked
off. These then become cut up transversely, new branches
from either the new or old central veins extending in to the
lower ends of the lobules. In some instances, however, the
new hepatic veins which supply the lower ends of the lobules
are formed earlier by a bifurcation of the new central vein.

Slight variations in development are often met with in cer-
tain groups of new lobules. Figure 24 is from a wax recon-
struction of three almost completely formed lobules which have
developed from a single one. The original central vein H has
bifureated into the two branches h and h. The original septum
is shown between A and H; it is almost complete. A new cleav-
age, B, however, has grown in from the side and has joined the
original septum, thus completely cutting off the lobule VN. A
new vein, h’, has grown out from the original central vein into
the new lobule O, which is not yet entirely separated from the
lobule M. In the process of splitting, the original central vein
H has become interlobular in position.

Very seldom in the pig’s liver is found a lobule through which
the central vein extends to become the central vein of another
lobule. One of the few which I have found in looking over
a large number of lobules is shown in figure 25. Judging by
the stage of development of its connective tissue septa, I have
explained its existence in the following manner: The three
lobules, A, B, and C, were originally one: its central vein H
bifureated into h and h, and the longitudinal septum grew

DEVELOPMENT OF LOBULE OF PIG’S LIVER a

down from the surface, bisecting the angle formed by the new
branches of the central vein. This septum instead of following
the central vein, either turned off to one side or was met by
another septum from the side, thus leaving the old central vein
completely surrounded by liver cells. That such central veins
do not long remain within the lobules is evident, since none of
the larger sublobular veins are found passing through single
lobules. The manner in which such a vein leaves its lobule,
however, I have not fully determined. It may be that they
later become interlobular in position by the ingrowth of addi-
tional septa. It seems more probable, however, that they grad-
ually become shifted to one side, since in several other similar
lobules, I have found them eccentrically placed. In these
lobules new central veins were found arising at right angles
from the eccentrically placed ones, seemingly marking the
path along which the old veins must have shifted.

From the above description, it is evident that lobule forma-
tion takes place by the binary fission of an elongated lobule.
This is preceded by either a bifurcation of the central vein or
by the sprouting out of a new vein from it. In either case, the
two veins extend into the halves of the old lobule. By the time
the cleavage is complete, each half of the old lobule represents
a fully formed new lobule. The self-explanatory figures 11
to 15 show diagrammatically the manner in which a single lob-
ule may give rise to a larger group.

In figure 28 is shown a wax reconstruction of a group of lob-
ules, all of which are supplied by a single sublobular vein. This
vein and its accompanying branches are shown in figure 27.
From the above description, it follows that all the lobules of
this group must have arisen from a single lobule, and, con-
versely, all the lobules which have arisen from this original
lobule belong to this group. Each branch of the hepatic tree
goes to an individual lobule and represents an independent
cleavage of the hepatic parenchyma. With the exception
of a few of the terminal branches, it is impossible to tell which
have arisen through a dichotomous bifurcation of terminal cen-
tral veins and which as new outgrowths from existing stalks.

316 FRANKLIN PARADISE JOHNSON

It is interesting to note, however, that all the branchings of
the hepatic vein shown, with possibly one exception, are dichot-
omous. This condition which has been mentioned by Roux
(95) obtains in the adult and is easily verified by the examina-
tion of celloidin corrosions of either the hepatic or portal trees.

14

Figs. 11 to 15 Diagrams representing manner of formation of hepatic lob-
ules. The numbers represent the number of times the hepatic vein has
bifurcated. Vessels marked 1’, 2’, etc., indicate new growths, not terminal
bifurcations.

DEVELOPMENT OF LOBULE OF PIG’S LIVER a7,

Thus far I have shown that the terminal hepatic veins always
keep within the lobules; that each of its new branches is indicative
of the formation of a new lobule. The portal veins in growing
spread between the lobules, tending always to keep within a
certain distance of the central veins. It should not be inferred,
however, that the growth of the portal veins follows in point
of time that of the hepatic veins. Both undoubtably grow hand
in hand, increasing in length and diameter, and branching as
the increasing parenchyma demands. With the increase in
the size of lobules, there is ever the tendency to spread the two
sets of veins further apart and this tendency is at all times being
counteracted by the formation of new veins.

The manner in which the new veins develop, that is by a
widening out of certain sinusoids, has already been described.
These new veins, according to Mall, grow into those regions
on the surface of the lobules which are most distally situated
from the branches of the portal veins, areas which Mall has
termed ‘nodal points.’ The whole matter of nodal points can
only be clearly understood from a study of the lobule in its three
dimensions; they are uncircumscribed areas on the surfaces of
the lobules which lie between the terminal branches of the por-
tal veins and between the central vein of one lobule and that of
an adjacent lobule. In the dividing lobule shown in figure 26
as many as ten nodal points may be counted, one of which is
just forming coincidentally with the formation of the new divid-
ing septum. The statement of Mall’s (06) that the portal and
hepatic veins alternately grow toward the nodal points and
break them into fragments to form new nodal points holds good,
according to my observations on the pig’s liver, for only the por-
tal veins. In figure 26 the whole upper surface of the lobule
was originally represented by a single nodal point; the new
branch of the portal vein has broken it up into two new ones.
But the terminal hepatic veins always lie within the lobules, they
grow toward the nodal points only as the parechyma of the
lobule increases, but they never reach them. In the lobule
represented in figure 26 the single central vein H was originally
directed toward the single nodal point of the top surface, but

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 3

318 FRANKLIN PARADISE JOHNSON

with the splitting of the lobule two new nodal points are formed,
and the two new central veins are again found to be directed
toward these new nodal points. That the hepatic veins do not
grow into the nodal points and break them into fragments is
still more evident when it is considered that the terminal cen-
tral veins are always found within the lobules, while the nodal
points are always interlobular in position.

The role of the connective tissue septa in lobule formation. In
order to understand clearly the rdle played by the connective-
tissue septum in the splitting of one hepatic lobule into two, it
will be necessary to review in their proper sequence the events
which take place in the formation of new lobules. The first
unmistakable evidence of splitting in a lobule is the formation
of an additional branch of the hepatic vein, either by the bifurea-
tion of a terminal vein or by its outward growth from the side of
an existing vein. At the same time, however, new branches of
the portal vein are similarly forming, although these latter
branches are not indicative of the formation of new lobules.
Next follows the beginning of the cleavage, which consists at first
of the parallel arrangement of liver cells and is afterward followed
by an ingrowth of connective tissue. The growth of the new
branches of the portal vein takes place along the fissures of the
new septum, if they have not preceded it (fig. 26). The new
septum comes to possess in this way a new nodal point. Branches
of the portal vein, however, do not actually enter the new septum
to break up its nodal point until both new lobules are fully
formed, and not until the new lobules have increased sufficiently
in size to warrant this addition to the vascular system.

It is evident, therefore, that the real and essential phase of
lobule formation is that concerned with the vascular system.
New lobules are actually in existence when certain relations be-
tween the new branches of the portal veins and those of the he-
patic veins become established. The new branches of the veins
determine the position of the new septa, just as they do shortly
before birth when the first evidences of dividing septa appear. In
the early stages, before traces of the septa are at all apparent,
similar branches of both systems of veins are formed, establish-

DEVELOPMENT OF LOBULE OF PIG’S LIVER 319

ing similar relations to one another. Thus they form undefined
lobules, which are only different from those which are separated
from one another by connective tissue septa, in that their vascu-
lar systems are In more open communication with each other.
We may unhesitatingly conclude, therefore, that the connec-
tive tissue septa play no active part in the formation of lobules in
the pig’s liver. The boundaries of the lobules are determined by
the vascular system; these in turn become apparent by the paral-
lel arrangement of the border cells; finally the connective tissue
fibers invade the reticulum between the parallel layers of cells
and the septum becomes established. Just what may be the
stimulus which causes the growth of the connective-tissue fibers,
I have not determined; it may or it may not be due to certain
influences of the circulatory system. When this problem is
solved we may understand why it is that in certain animals the
septa are complete, in others partially, and in still others entirely
absent; and further, why in the seal (Mall, ’06) the connective
tissue septa bound the portal rather than the hepatic lobules.

THE GROWTH OF THE LIVER

I have thus far considered the formation of new lobules with-
out regard to the changes which they produce in the liver as a
whole. The growth of the liver is accomplished by two means:
first, by an actual increase in size of the lobules; second, by a
multiplication of lobules. I have shown in a previous paper the
approximate rate of growth of the hepatic lobule of the pig. In
the table below, it will be seen that the size of the larger lobules
remains constant between the 80-mm. and the 154-mm. stages.
From the 229-mm. stage on they show a gradual increase. The
growth of the lobule is shown graphically in figures 16 to 19, all
of which are camera-lucida drawings of equal magnification.

It was believed by Illing (05) and Theopold (710) that after
the connective-tissue septa are laid down the growth of the liver
takes place entirely by an enlargement of its lobules. Theopold,
however, states that the septa are not completely formed until

320 FRANKLIN PARADISE JOHNSON

the first few weeks of -postnatal life and that during this time
innumerable new lobules are forming. He further states that
in the early stages of the liver the hepatic branches outnumber the
portal, and that lobule formation ceases when a single portal

Figs. 16 to 19 Outline camera-lucida drawings of lobules of four different
stages to show increase in size of hepatic lobules. Fig. 16. 254 mm.; fig. 17, four
weeks old; fig. 18, two months old; fig. 19, adult. X 25.

DEVELOPMENT OF LOBULE OF PIG’S LIVER 321

TABLE 1

Table showing the average diameters of some of the larger lobules of the pig’s liver
at various stages of the development

AGE DIAMETER AGE DIAMETER
2.1) CHT A ee a eae 0.33 S GRYGE. 4i3é:. Jon Gee 0.54
fe mm Ee ike eh. nie. 0.33 3 weekstai hs see . aie 0.49
DAMM ese phen eteye oe tic ole 0.33 AC WEEKSER Ne cotich oranda ee 0.51
DOORN Ae ea sheets os ac 0.35 Per seVGualrlovere ay Seer Pees aber 0.59
NT ee Meee San, ee 0.43 DNC (LGU rat Ae eS be eM ee a 1.2)

1 The diameter of the lobules of the adult pig as given here (1.2 mm.) is the
same as that given previously (Johnson, 7184). Mall (’06) gives it as 1.4 mm.
There can be no question that Mall obtained his figures as I did, by averaging
only the diameters of large lobules as seen in sections of the liver. The actual
‘average diameter’ of the liver lobule is an altogether different thing, one difficult
to calculate. It must be pointed out, as I have recently shown (’18a), that
the hepatic lobules of the pig vary greatly in size within a single liver, and also
that the average size varies for different adult livers. Reference to the table
given in my former paper will show how I obtained the ‘average weights per
lobule’ given in the table. From an average of these ‘average weights per
lobule,’ I have attempted to calculate the average diameter, assuming a spherical
form for the lobule. This gives 0.8 mm. Since, however, the lobules are not
spherical, the diameter would in reality be less than0.8 mm. With regard to the
relatively average diameter of the lobules in the young stages, I am confident that
there is less variation in size than in the adult, and that the average diameter
of the lobules of the stages 80 to 154 mm. is only slightly less than 0.33 mm.

branch has formed for each hepatic branch present. These
observations, however, I have been unable to confirm.

The growth of the lobules likewise takes place in two ways:
first, by an increase in the numbers of the hepatic cells and sinu-
soids; second, as has been pointed out by Toldt and Zuckerkandl
(75) and by Illing (’05), by an actual increase in the size of the
hepatic cells. The latter, however, is counterbalanced to a cer-
tain degree by a decrease in the diameter of the sinusoids. The
rate of increase in size of the cells and sinusoids, and numbers of
the cells, is shown in table 2.

The formation of lobules begins at an early stage (Mall’s one-
lobule stage), and, I believe, continues throughout all stages to the
adult. This is in agreement with the statement of Lewis (’12),

322 FRANKLIN PARADISE JOHNSON

that ‘‘the multiplication of lobules continues long after birth,
and partly divided, compound forms were recognized in the adult
by Kiernan.”’ At just what time lobule formation is fully com-
pleted I have not determined; there are some evidences that a few
lobules are undergoing division in the so-called ‘adult’ stages I
have studied.

In studying the growth of a solid mass of tissue such as the
liver, we must consider the possibility of both peripheral and cen-
tral growth; in the former the increase takes place by a laying
down of successive layers on the surface; in the latter the growth
takes place by an increase in the size and number of the units
within the organ. The growth of the liver may be described as

TABLE 2
ESTIMATED
SIZE OF DIAMETER NUMBER OF
STAGE HEPATIC CELLS |OF SINUSOIDS IN|CELLS IN AVER-
IN MICRA MICRA AGED-SIZED
LOBULE
CO: natal yi ee RRRSIS S! . 250d EL I0. HOYER. 13.9 22.9
152 mms. MEER CM Se ee 12.5 13.6
Beds riipy. HAAS: ee Se aie, vee 16.0! 8.0 21,000
4cweekstoldiy Hk, AIRE SAS Lk, Ly 7/ 9.8 79,000
2 month sioladwain |. nes ns eyes y. Mis es 13.1 12.1 97,900
Alult.e SPO eee eee: be. ue se 19.8 8.0 465,000

1 The cells in this stage were unusually large and vesicular.

central, that is, taking place more or less evenly throughout.
After a study of sections of developing livers, I was at first led
to believe that mitosis proceeded somewhat more rapidly in the
surface lobules than in the deeper seated ones (718 b) but the
further study of this point does not confirm this view. Mitotic
figures even in favorable sections of well-preserved livers are not
abundant,! making it extremely difficult to note any actual varia-
tion in their distribution. I am inclined to believe that in general
all parts of the liver are growing at the same rate, since the liver

1 Tlling (’05), who has studied the liver cells of growing and adult animals,
states: ‘‘In hiesigen Institute (Tierarztlichen Hochschule zu Dresden) sind im
Laufe der Jahre Tausende von Leberpriparaten von allen Haustieren unter-
sucht, aber niemals mitotische Kernfiguren gefunden worden.”’

DEVELOPMENT OF LOBULE OF PIG’S LIVER 323

tissue is everywhere of the same type, since all its parts are equally

favored by the circulation, and since, so far as can be determined,

dividing cells and lobules are uniformly distributed in sections.
Concerning the growth of the liver, Mall (’06) states:

In development the liver structure shifts distalwards, successively
tearing off its capillary connections with the main veins, gradually
rearranging the architecture of the lobules, often fracturing and
scattering them.

Although I agree with Mall that the liver parenchyma shifts
distalward as the liver grows, I do not believe that the lobules
become fractured and scattered or that capillary connections are
torn away. I find no instances of fractured lobules other than
those undergoing normal binary fission and no instances of the
scattering of lobules since the lobules maintain their connections
with the hepatic veins throughout. As for the tearing away of
capillary connections along the beginning of sublobular veins, it
seems more probable that such capillaries retract (Clark, ’18), be-
cause of the decreased flow of blood through them. I believe
that the shifting of the liver substance peripherally produces no
destructive changes in the arrangement of its units other than
that the form of the lobules is altered by the pressure of one lob-
ule against another due to their manner of growth.

That the shifting of the liver tissue from the center towards
the periphery is taking place constantly in the growing liver,
there can be no doubt. This does not mean, however, that cer-
tain lobules or groups of lobules are shifting or slipping past one
another or that they shift along the walls of the blood-vessels.
In fact, the shifting takes place in such a way that the general
relations are not at all disturbed. The blood-vessels, ducts, con-
nective tissue, etc., grow and shift correspondingly with the lob-
‘ules, and while the liver tissue is shifting from the center toward
the periphery, it is not to be inferred that the deeper-seated lob-
ules will eventually come to lie on the surface or even nearer to
it. On the contrary, they are continuously getting further and
further away from the periphery for the simple reason that the
more peripheral lobules are likewise constantly increasing in
number.

324 FRANKLIN PARADISE JOHNSON

These points concerning the growth of the liver can best be
exemplified by the following diagrams. In figure 20, let the circle
A represent a circumscribed mass of liver tissue surrounding a
large vein of either the hepatic or portal systems. The zones B
and C represent concentric zones of liver substance about A.
When A increases in size (fig. 21), it might at first seem that it
would press against the zone B and this in turn against C, tend-
ing to produce changes in them. But the zones B and C are
likewise growing at the same rate as A. In their growth they
expand just as metal rings expand when they are heated. The

20 21

Figs. 20 and 21 Diagrams to show manner of growth of liver as a whole.

fact that the center is filled by A alters in no way the shape the
zone B will assume. The expansion of C similarly makes room
for B. The process is exactly similar to that which takes place
in the expansion of any solid ball of metal when heated evenly.
A lobule in figure 21 at the point m would be further away from
both center and periphery than at its former point m in figure 20.
Similarly, the vein increases in length as the liver expands. Its
increase is due in part to the growth at the tips of its branches
and in part to an increase in length of its trunk.

The above hypothesis explains in a general way the growth
of the liver and is based on the assumption that the liver tissue

DEVELOPMENT OF LOBULE OF PIG’S LIVER o25

is everywhere growing at an equal rate. While this assumption
is probably true for groups of lobules taken collectively, it is not
always true within individual lobules. I have shown above that
when a lobule grows, it may increase more in one diameter than
in another. It is probable that this type of unequal growth
gives rise to local pressures here and there and that these in turn
determine to a certain degree the shapes of the lobules.

CONCLUSIONS

1. The first evidence of connective tissue septa found was in
an embryo of 254 mm. in length. The growth of the septa is
gradual and they are not fully formed until about two months
after birth. Their origin is first indicated by the arrangement
of the border cells of the lobules in parallel layers. The con-
nective tissue of the portal canals sends sprouts into the reticulum
between the layers of border cells. These sprouts, coming from
opposite directions, meet in the region of the nodal points thus
completing the septa.- Additional septa are formed with the de-
velopment of new lobules. Their paths are similarly marked out
by the parallel arrangement of border cells. The collagen fibrils
of these septa sprout out from the connective tissue about the
portal veins and from that of those septa already present.

2. The collagen fibrils of the capsula fibrosa (Glissoni) appear
first in the region of the porta hepatis and spread over the sur-
face of the liver. They completely cover the liver in a pig of 111
mm. ‘The capsule remains as an extremely thin and delicate
layer until birth, after which time it gradually thickens.

3. The formation of new lobules may be described as a neces-
sary readjustment on the part of the parenchyma to circulatory
difficulties. It is accomplished by a binary fission of lobules al-
ready present, the cleavages taking place only in lobules which
have developed two central veins. The new central veins may
arise, 1) by a bifurcation of the tip of a growing central vein, or 2)
by a sprout from the side of an existing vein. The plane of cleav-
age bisects the angle formed by two new veins. The cleavage is
evidenced by the arrangement of the hepatic cells into parallel

326 FRANKLIN PARADISE JOHNSON

layers, into the reticulum between which collagen fibrils push
from the surface of the lobule. In the case of 1) the septum com-
pletes itself gradually in the plane of the old central vein, which,
when the septum is complete, becomes a sublobular vein. Usu-
ally before the septum is fully completed the newly formed
lobules show evidences of further segmentation.

4. The growth of the portal veins takes place by an increase in
length and by the formation of new branches which spread them-
selves between the lobules. They grow into the nodal points
and split them up into additional ones.

5. The central veins grow by an increase in length and by the
formation of new branches which are always intralobular in posi-
tion. They are directed and grow toward certain nodal points,
but never reach them. .

6. In their growth both the terminal hepatic and portal veins
never approach nearer one another than one-half the diameter of
the lobules, as pointed out by Mall.

7. The connective tissue septa of the pig’s liver play only a
passive role in the formation of new lobules. They grow in
between the new lobules after they have really formed.

8. The growth of the liver takes place throughout all stages,
1) by an increase in the number of its lobules; 2) by an increase
in size of its lobules.

9. The growth of the lobules likewise takes place, 1) by an in-
crease in the numbers of its hepatic cells and sinusoids; 2) by an
increase in the size of the hepatic cells. The latter is counter-
balanced to a certain degree by a decrease in the diameters of
the sinusoids.

10. In general, all parts of the liver grow simultaneously and
equally. There is a constant shifting of lobules peripherally,
but this takes place in such a way as to produce a minimum
amount of change in the relation of one lobule to another.

11. It is probable that the lengthening of lobules in their
growth (not an equal swelling in all directions) produces local
disturbances which manifest themselves by giving rise to the
great variety in shapes of the lobules.

DEVELOPMENT OF LOBULE OF PIG’S LIVER SWAT

BIBLIOGRAPHY

Cuark, E. R. 1918 Studies on the growth of blood-vessels in the tail of the
frog larva—by observation and experiment on the living animal. Am.
Jour. Anat., vol. 23, pp. 37-88.

Depeyre, A. 1910 Morphologie du lobule hepatique. Bibl. Anat., T. 19, pp.

249-263.

ENGEL, C. S. 1899 Die Blutkérperchen des Schweines in der ersten Hialfte

des embryonalen Lebens. Arch. f. mikros. Anat., Bd. 54, 8.24-59.

Fevix, W. 1892 Zur Leber- und Pankreasentwickelung. Arch. f. Anat. und
Entw., S. 281-328.

Inuinc, G. 1905 Vergleichende histologische Untersuchungen iiber die Leber
der Haussiugetiere. Anat. Anz., Bd. 26, S. 179-193.

JoHnson, E. . P. 1917 The later development of the lobule of the pig’s liver.
Anat. Rec., vol. 11, pp. 371-372.
1918 a The isolation, shape, size, and number of the lobules in the
pig’s liver. Amer. Jour. Anat., vol. 28, pp. 273-283.
1918 b Additional notes concerning the development of the pig’s
liver. Anat. Rec., vol. 14, p. 40.

KIERNAN, F. 1833 The anatomy and physiology of the liver. Trans. Roy.
Soc. of London, pp. 711-770.

K6uuiker, A. 1879 Entwickelungsgeschichte des Menschen und der héheren
Wirbeltiere. II Aufl., Leipzig.

Kostanecki, K. V. 1892 Die embryonale Leber in ihrer Beziehung zur
Blutbildung. Anat. Heft., Bd. 1, S. 303-822.

Lewis, F. T. 1912 The development of the liver. In Keibel and Mall’s
Human Embryology, vol 2, pp. 403-428. Philadelphia and London.

Lospennorrer, W. 1908 Uber extravaskulire Erythropoése in der Leber
unter pathologischen und normalen Verhaltnissen. Beitr. zur. path.
Anat. und zur Allg. Path., Bd. 48, S. 124-146.

Matt, F. P. 1896 Reticulated tissue and its relation to connective tissue
fibrils. Johns Hopkins Hospital Reports, vol. 1, pp. 171-208.
1906 A study of the structural unit of the liver. Am. Jour. Anat.,
vol. 5. pp. 227-308.

Minot, C. S. 1900 On a hitherto unrecognized form of blood circulation
without capillaries in the organs of vertebrata. Proc. of the Bos-
ton Soc. of Nat. Hist., vol. 29, pp. 185-215.

Roux, W. 1895 Gesammelte Abhandlungen iiber Entwickelungsmechanik der
Organismus. I. Functionelle Anpassung. Leipzig.

TuHeoprotp, J. 1910 Untersuchungen iiber die Entwickelung der Leberlepp-
chen des Schweines. Bern. 32 pp.

Toupt, C., unp ZucKERKANDL, E. 1875 Uber die form und a
gen der menschlichen Leber wihrend des Wachstums. Sitz.-Ber.
der Kais. Akad. der Wiss. Wien, Bd. 72, Abt. 3, S. 241-295.

VAN DER Srricut, O. 1892 Nouvelles recherches sur la genése des globules
rouges et des globules blanes du sang. Arch. di Biol., T. 12, pp.
199-344.

PLATE 1
EXPLANATION OF FIGURES

22 Dividing lobule of the liver of a two-months-old pig. X 55.

23 Dividing surface lobule of same liver. X 55.

24 Dividing lobule of same liver; for description see page 314. X 55.

25 Dividing surface lobule of same; for description see page 314. X 55.

26 Dividing lobule of same, showing both portal and hepatic veins; for
description see page 315. X 55.

328

PLATE 1

DEVELOPMENT OF LOBULE OF PIG’S LIVER

FRANKLIN PARADISE JOHNSON

SPASM:

.

329

PLATE 2
EXPLANATION OF FIGURES

97 and 28 Wax reconstructions of hepatic tree with the group of lobules
which it drains. The central veins and the lobules which they drain have
corresponding letters. Those central veins not lettered belong to lobules not

shown in figure 28.

300

DEVELOPMENT OF LOBULE OF PIG’S LIVER PLATE 2
FRANKLIN PARADISE JOHNSON

331

Resumido por la autora, Lucile Witte.

Histogenésis del musculo cardiaco del cerdo con relacién a la
aparicién y desarrollo de los discos intercalados.

1. El tejido cardiaco del cerdo tiene al principio estructura
celular y est& compuesto de células fusiformes que mas tarde se
anastomosan terminal y lateralmente para formar una red de
fibras. 2. La estriacién aparece mas temprano en los discos que
en las otras partes, pero en focos diseminados en toda la extensién
del tejido. 3. Los discos aparecen en el estado de embrién de
76 mm. de longitud, mucho mds pronto que en cualquiera de los
animales estudiados hasta el presente, con la excepcién del em-
briédn del. gato de cuatro dias. 4. Los discos aparecen primero
como puntos o bandas inicompletas que comienzan en la periferia
de la fibra y crecen hacia su interior. Avanzando el desarrollo
se transforman en discos planos que atraviesan toda la fibra ex-:
tendiéndose después a dos o mas fibras y finalmente adoptan el
tipo mas complejo de discos. 5. Los discos no son més numerosos
en las dreas contraidas que en las relajadas. 6. Los discos no se
presentan en los extremos de una serie de nticleos, constituyendo
de este modo una célula, porque se presentan invariablemente
en fntimo contacto unos con otros, apareciendo en manchas, y
raras veces existe un nucleo entre ellos. 7. La autora propone
la teorfa que supone a los discos como bandas encargadas de re-
forzar las fibras musculares, puesto que aparecen préximamente
durante la transformaci6n de las células en fibras y aumentan en
numero y complejidad con el crecimiento y actividad del corazon.

Translation by José F. Nonidez
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MARCH 17

HISTOGENESIS OF THE HEART MUSCLE OF THE PIG
IN RELATION TO THE APPEARANCE AND DEVEL-
OPMENT OF THE INTERCALATED DISCS

LUCILE WITTE
Department of Zoology, University of Kansas

EIGHTEEN FIGURES

INTRODUCTION

It is the purpose of this paper to discuss the histogenesis of
the heart muscle of the pig with special emphasis upon the time
of appearance and the development of the intercalated discs. It
is a much-disputed question whether the intercalated discs are
true cell boundaries of muscle cells or whether they are only
peripheral, darker staining substances in the regular form of
bands. <A further aim of this paper is, if possible, to throw
some light on the question stated above, by tracing carefully
the formation of the discs with regard to their probable function.

METHODS

In making this study, pig embryo and young pig hearts were
used in the following series, respectively, 25 mm., 38 mm., 76
mm., 89 mm., 95 mm., 102 mm., 115 mm., 120 mm., 140 mm.,
165 mm., 176 mm., 182 mm., 201 mm., 238 mm., 251
mm., 277 mm., and 303 mm., and four months, five months,
eight months, and one year. Only the ventricular muscles
were examined. The method of technic found most satisfactory
for demonstrating the discs clearly was that employed by Zim-
merman! and his students Palezewska? and Werner.? The
tissues were fixed in a solution of absolute alcohol and 25 per
cent nitric acid and stained ‘in toto’ in Griibler’s haemalum,
after which they were embedded according to the usual paraf-
fin method. Sections were cut at 5 w and 8 un.

333

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 3

334 LUCILE WITTE

Other methods of staining were used with no success at dif-
ferentiating the discs. The iron-hematoxylin method was used
on the earlier tissues with no results and the Delafield’s hema-
toxylin with an eosin counterstain was used on the older tissue
where the dises are known to exist, with equally unsuccessful
results. The striations, however, were very evident.

HISTORICAL REVIEW

Zimmerman,! Palezewska,? and Werner® in 1910 worked out
very complete studies of adult human and mammalian heart
muscle with regard to the function of the intercalated discs
and they all concluded definitely that the dises did form distinct
cell boundaries. Palezewska, in her work on the human heart, |
evolved some diagrams (figs. 2, 9, and 10) which at first glance
would seem very convincing. However, on closer study, they
cannot be accepted. H. E. Jordan‘ has reviewed both Pal-
ezewska’s and Werner’s papers in his study of the heart muscle
of the humming-bird, and the reader is referred to his paper
for a more detailed review. Werner, in making her study of
the mammals, worked out the cardiac muscle of the adult pig
quite thoroughly, and she noticed a striking peculiarity which
was that the nuclei of the ‘muskelterritorien’ tended to range
themselves in long rows, numbering in multiples of from 2 to
32, though the most usual number was 8. The intercalated
dises seemed to divide these series of nuclei into distinct cells,
as her figures 1, 2, 3, and 4 show. She compared the ventric-
ular muscles with those of the auricles and found that the nuclei
were not nearly so numerous in the auricles ranging usually
+n number from 1 to 2 and 4. The discs occurred in zigzag lines
in both ventricle and auricle.

Concerning the structure of heart tissue, Jordan* stated briefly
that in the humming-bird the muscle ‘‘is syncytial in character,
the fibers anastomosing laterally and apically.” A distinct
membrane or sarcolemma existed and seemed to cover the en-
tire fiber.

J. B. MacCallum,® in his study of the histogenesis of heart
tissue, used the pig embryo in a series of from 10 to 100 mm.

DEVELOPMENT OF DISCS, HEART MUSCLE OF PIG 335

He found that there were four different types of cells in the heart
of the 10-mm. pig. One presented a network of irregular meshes
in which there was a clear, unstained substance; another, a
regular network; a third, a network in which some of the meshes
were broken up into smaller discs by radial division, and
fourth, a network in which a single row of fibrils began to ap-
pear. The later stages of development showed an increased
number of fibril bundles. The cells of 55-mm. embryo hearts
were still spindle-shaped, but very much lengthened, and in the
72-mm. and 100-mm. stages the cells had lost the spindle form
and had taken almost the form of the adult muscle. The fibrils
were found to be all through the cell rather than just at the
periphery. MacCallum described the adult muscle of the
human heart as being composed of rhomboidal cells two or three
times wider than long. They sometimes broke up into branches
which united with branches from other cells. The lines of union
were at acute angles to the length of the fibers and each fibril
of the cells sent out two or three processes which ran through
these lines of union and met each other. He represented by his
figure 1 a very complex structure and stated that the same
structure was not to be found in other animals, but that the
structures commonly known as protoplasmic bridges corre-
sponded to his lines of union in the human heart. The cells
were composed of fibrils which were stained dark and sur-
rounded by unstained sarcoplasm. He described the fibrils with
their surrounding sarcoplasm as made up of dises separated from
each other by a narrow line known as Krause’s membrane. He
found the sarcoplasmic discs also to be divided somewhat radi-
ally and the lines of separation were continuous with Krause’s
membrane.

In the embryonic tissue MacCallum noticed that the cells at
the periphery of the heart were farther developed than those at
the anterior, indicating that the cells grow on the inside of the
muscle. He considered that the development of the cells from
an irregular network to cells of the adult form with numerous
fibril bundles increased the heart’s capacity for work and that
the network of the earlier stages must be contractile.

336 LUCILE WITTE

To return to Jordan’s paper on the humming-bird, he gave
fourteen very conclusive points as to why he thought the inter-
calated discs could not be cell boundaries. One point which
seemed very readily to contradict the cell-boundary theory was
that he had found discs which lay over the nucleus. They
usually were peripheral in the fiber, and they also varied to a
great extent in coarseness—a condition which is not common to
cell walls. As has been stated, Jordan concluded from his four-
teen observations that intercalated discs were not cell walls or
cement lines. As to the function of the discs, he only conjec-
tured. They might have something to do with contraction,
since they were found in patches and much more numerous in
contracted than in relaxed fibers. His figures, however, failed
to prove his statement that the dises were more numerous in
contracted than in relaxed areas. Figures 2 and 4 represented
discs in contracted fibers, while figures 1 and 3 represented
normal areas of the fiber. According to these, there were not
as many discs in contraction (fig. 4) as in relaxation (fig. 1 and 3).
Figures 1 and 2 might easily be explained as discs occurring in
normal relaxed fibers.

Up to the time of writing his paper, Jordan had not found
discs in fetal hearts. This would seem to disprove any idea
that the discs were related to the rhythmic beat of the heart.
His final conclusion was that the discs ‘‘were of the same nature
as the anisotropic bands, were closely related to them in position,
and might represent a definite physiologic or functional state.”

H. E. Jordan and K. B. Steele* in 1912 worked out a com-
parative study of vertebrate hearts in which they attempted to
show that intercalated discs were to be found in animals lower
than birds and also in fetal material. In this they succeeded,
for they were able to demonstrate discs in amphibians (frogs
and toads), reptiles (turtles and lizards), and fishes (trout), and
in the fetal guinea-pig hearts during the last week of gestation.

The technic used was Zimmermann’s. The discs in the fetal
hearts appeared simultaneously with the striations and increased
in number with the growth of the animal after birth. The discs
were also found to be present in a cat embryo of four days.

DEVELOPMENT OF DISCS, HEART MUSCLE OF PIG 337

There is some doubt here, whether this statement refers to the
embryo four days after conception or to the animal four days
after birth.

The lower vertebrates presented dises which were much less
numerous and complex as the animals went down the scale of
complexity. ‘These discs presented no evidence in favor of the
cell-boundary theory, since they were superficial in position,
often lay over a nucleus, were situated at random with
relation to the nucleus, and did not appear earlier than the
striations. They seemed to be a part of, or closely related to,
the anisotropic lines, since they shaded into these and were
parallel to them in all cases.

H. E. Jordan and J. B. Banks’ in 1917 worked out a study of
the intercalated discs in the heart of the beef. In this paper
the fetal heart was used after a study of the adult heart had been
completed. The youngest fetal heart was one of approximately
two and one-half or three months of age. In this, the cells were
found to be fusiform, showing some signs of anastomosing later-
ally and terminally. Slight striations were visible and the
intercalated discs appeared as large dots, shading into the tel-
ophragma. They were situated at random in the cells, but did
not occur at the point of terminal fusion of two cells. None of
them were more than peripheral. With the increasing age of
the embryo, the discs became more distinct and more complex
until in the adult heart they took the form of step-like formations
very much complicated in structure.

No investigator, as far as I have been able to find, has taken
up the study of fetal heart tissue in any animal with the thought
of especial investigation as to the time of appearance of the discs,
their development, and their function. Whatever work has
been done on fetal material concerning the discs has been more
or less superficial, rather than very detailed.

DESCRIPTION

In this description the early stages of the series of pig embryos
will be considered briefly with regard to the histogenesis of heart
muscle. The later stages will be considered more in detail
since they show the origin and development of the discs.

338 LUCILE WITTE

In the 25-mm. embryo heart the cells are spindle-shaped in
longitudinal section with a large central nucleus. The cytoplasm
of the cells remains clear and here and there very faint striations
appear. In cross-section the cells present the same condition
that MacCallum describes. The cells seem broken up into
smaller sarcoplasmic discs by walls which are continuous with
the bounding membrane of the sarcoplasm. Some fibrils have
appeared at the centers of these discs.

At 38 mm. the cells have developed into a more fibrous struc-
ture. There are still some of the spindle-shaped cells to be found
near the center of the ventricle, but the cells of the auricle and
peripheral part of the ventricle have elongated into fibers in
which the cross-striations appear more distinctly, especially at
the periphery of the heart. The fibers are finer and the nuclei
are still large and centrally located.

At 76 mm. the cells seem to have lost all eanilatice to spindle
cells, practically all through the heart. They have elongated
and anastomosed until there is a complete network or syncytium
of fine fibers, all of which are definitely striated. At this stage,
darker portions of the striations occur at intervals, beginning at
the periphery of the fiber and proceeding across at right angles
to the length of the fiber. The discs, as these dark portions
very clearly seem to be, do not extend across the entire width of
the fiber, but gradually shade into the striations or telophragma
(fig. 1). In a few instances, there are faint discs extending
entirely across a fiber (fig. 2).

At 89 mm. the general structure of the tissue is the same as
that of the preceding stage. The discs, however, are more dis-
tinct and almost invariably extend across the fiber. They are
no more numerous than in the 76-mm. stage (fig. 3). Figure 4
represents a type of dise which occurs only very rarely at this
stage.

The 115-mm. embryo heart presents much more development
in the discs. They are much more numerous and distinct and
in every case extend entirely across the fiber. In this tissue
some of the fibers are wide while others are still narrow (fig. 5).
There is no change in the 120- and 140-mm. stages. The discs

DEVELOPMENT OF DISCS, HEART MUSCLE OF PIG 339

are numerous, straight, and occur in patches in areas devoid of
nuclei. The general arrangement of the striations is different in
that some of the bands present a greater density than others.
These light and dark bands occur alternately and are considered
by most observers as areas of contraction. I have been unable to
find any difference in the distribution of discs in this region and
in other relaxed areas. There is also no marked change in the
165-mm. stage.

At 182 mm. we have the first example of a disc extending over
more than two fibers. In figure 6 the fibers anastomose ter-
minally and the disc runs across the four fibers.

The 201-mm. embryo heart shows the colorless cytoplasm
which exists invariably as a streak through the center of the
fiber in older tissue, and surrounds the nucleus. This has not
been noticed in the earlier tissues nor is it constant at this stage.
The dises run to this cytoplasm, but not across it (fig. 7).

In the periphery of the heart muscle of the 238mm. stage,
the polynuclear fibers are first noticed. The nuclei tend to
arrange themselves in rows consisting of two or four, rarely
_ three. The mononuclear state exists farther toward the center
of the ventricle. The fibers are very compact and give the ap-
pearance of broad fibers. The discs take the step formation
more regularly than heretofore (figs. 8 and 9).

Figure 10 represents the discs at the 251-mm. stage. They
are quite frequent in occurrence and many of them, as here rep-
resented, are thick.

Figures 11 and 12 represent the condition of the discs in the
277-mm. and 303-mm. stages, respectively. There is no marked
change. The risers occur more frequently and the discs often
extend over two or more fibers.

In all of the tissue studied so far the discs have increased in
number by means of new and incomplete discs developing and
gradually growing across the fiber. Figure 7 is an example of
such discs.

There is a gap between the embryonic tissue and the postnatal,
since no stages were studied between the 303-mm. pig, which is
very close to the age for birth, and the four-month-old tissue.
There are, however, no very marked changes between the two.

340 LUCILE WITTE

At five months a new feature in the discs appears. Instead of
the discs’ extending across the fiber in a compact formation, they
seem to be made up of coarser granules on each separate fibril,
as is shown in figure 13. Figure 14 shows three fibers which are
parallel and are separated by wide, clear spaces. The dises are
at the same level in the three fibers, but instead of their being
straight bands, they are zigzag, as though the separate fibrils
had been pulled back and forth till the regions of coarse granules
were out of line with each other. This type of disc occurs quite
frequently throughout this stage.

The year-old tissue shows almost entirely the zigzag bands
running across several fibers. They cannot be described as exact
step formations, for they do not lie parallel to the telophragma
(figs. 15 and 16). Those dises which are straight are short. The
type of dises and the arrangement of the nuclei in the tissue are
just as described by Miss Werner. However, I cannot verify her
statement or her drawing showing that the nuclei are bounded at
either end of the series by discs. If the discs were visible at one
end of the series of nuclei, they were not at the other end.

The year-old tissue, then, represents the limit reached in this
study, where the discs extend over several fibers at nearly the
same level, or go up and down in a series of ‘steps’ and ‘risers.’
These last represent the most complex of the discs found.

DISCUSSION

As regards the early embryonic structure of the heart muscle, I
have been able to verify J. B. MacCallum’s statements. The early
tissue exists as spindle-shaped cells, and it is very evident how the
individual cells gradually anastomose and form the fibers of the
adult heart tissue. MacCallum contends that these spindle-
shaped cells gradually lengthen and branch and unite with each
other at the ends of the branches by definite lines ef demarcation.
He adds that these lines of demarcation correspond to the proto-
plasmic bridges found in tissues of other animals described by
some investigators.

Thus MacCallum places himself on the side with Zimmermann
and his students and contends that adult heart muscle does not

DEVELOPMENT OF DISCS, HEART MUSCLE OF PIG 341

become truly syncytial, but remains cellular with the lines of de-
marcation, or discs, as I shall call them, representing the walls
between the cells.

In my study of this change from cellular to fibrous tissue, I
cannot show that the discs form cell walls. They are too irregu-
lar in position and too infrequent in occurrence at such an early
stage.

Observations show that striations and dises do not appear si-
multaneously in the heart tissue of the embryonic pig. ‘The stri-
ations appear first. These facts need not contradict what Jordan
and Steele said, for they studied the embryonic heart of the guinea
pig and cat. Nor should they contradict Jordan and Banks’
statement that the discs appeared in the heart of beef simultane-
ously with the striations. The other work done on fetal mate-
tial with regard to the discs has been more or less superficial.
However, it may be true that the time of appearance of the dises
varies widely among the different animals.

It cannot be said that the dises in the embryonic pig’s heart
are cell walls or cement lines. In the first place, they are only
peripheral, for they pass out of focus very readily. Nowhere in
the tissue is it possible to keep the discs in focus through the en-
tire thickness of the fiber. Also, they occur so many times in
rather close proximity with no nucleus between them, as-is shown
in figures 5,7, 9,10,and11. The discs do not present a different
structure in the older material except that they are more complex
and occur more frequently in step-like forms. They are still
peripheral in the five-month tissue and they are very distinctly
darkened, portions of each fibril arranged in the form of a band,
rather than extending across the fiber as a connected disc (figs.
13 and 14). These facts would disprove what Zimmermann,
Palezewska, and Werner concluded concerning the intercellular
nature of the discs. |

/ Jordan presents the theory that the discs may be functional in
connection with the contraction of the heart. From the facts
observed in this study, it seems*improbable that the discs would
be caused by the contractions of the heart muscle or would result
from the rhythmic beat of the heart. The striations do not ap-

342 LUCILE WITTE

pear till the heart has attained definite form, some time after the
beating of the heart begins. The discs do not appear till some
time after the striations, so that any theory that discs were formed
at contraction or with the beginning of contraction of the heart
would be disproved.

It may be that the dises develop physiologically for the purpose
of strengthening the muscle fibers. They appear shortly after
the striations and at about the time the muscle cells begin to
anastomose and form fibers. The fact that the discs become more
numerous and complex with the increasing growth and activity
of the heart seems to strengthen the theory mentioned above.

SUMMARY

1. The early heart tissue of the pig is cellular in structure,
composed of spindle-shaped cells which later anastomose termi-
nally and laterally to form a network of fibers.

2. The striations appear earlier than the discs, but only here
and there throughout the tissue.

3. The discs appear at the 76-mm. stage, much earlier than in
any other animal studied, with the exception of the cat embryo
of four days.

4. The dises are at first dots or incomplete bands beginning at
the periphery of the fiber and growing across. With the advance
in development, they become straight discs across the entire
fiber, then across two or more, and finally assume the more com-
plex type of discs and ‘risers.’

5. The dises do not appear more numerous in contracted areas
than in relaxed areas.

6. The discs are not to be found at either end of a series of
nuclei, thus forming a cell, for they are almost invariably in
close proximity to each other, occurring in patches, and seldom
is a nucleus to be found between them.

7. The theory is put forth that the discs serve as strengthening
bands in the muscle fibers, since they appear at about the time
of the change from cells to fibers, and increase in number and
complexity with the growth and activity of the heart.

DEVELOPMENT OF DISCS, HEART MUSCLE OF PIG 343

In conclusion, I wish to acknowledge my indebtedness to Prof.
W. J. Baumgartner for his aid in obtaining material and for his
very helpful suggestions and criticisms.

BIBLIOGRAPHY

1 Zimmermann, K. W. 1910 Uber den Bau der Herzmuskulatur. Arch. f.
Mik. Anat., Bd. 75, S. 40.

2 PanczewsKa, IRENE von 1910 Uber die Structur der menschlichen Herz-
muskelfasern. Arch. f. Mik. Anat., Bd. 75, S. 41.

3 WERNER, Marie 1910 Besteht die Herzmuskulatur der Sidugetiere aus
alseits scharf begrenzten Zellen oder nicht? Arch. f. Mik. Anat.,

Bdat75,.5., 101.

4 Jorpan, H. E. 1911 The structure of the heart muscles of the humming-
bird, with special reference to the intercalated discs. Anat. Rec.,
Vole Ssap. old.

5 MacCauium, J. B. 1897 On the histology and histogenesis of the heart
muscle cell. Anat. Anz., Bd. 18, S. 609.

6 Jorpan, H. E., anp Steete, K.B. 1912 A comparative microscopic study
of the intercalated discs of vertebrate heart muscle. Am. Jour.
Anat, vol: 43, p. 151.

7 Jorpan, H. E., anp Banks, J. B. 1917 A study of the intercalated discs
of the heart of beef. Am. Jour. Anat., vol. 22, p. 285.

DESCRIPTION OF PLATES

The illustrations were made with the aid of the Bausch & Lomb camera lucida.

All figures are of tissue fixed in the alcohol nitric-acid mixture and stained
in hemalum according to Zimmermann’s technic, and are magnified 3500 diam-
eters, except 17 and 18, which are 2000. The original magnification is reduced
one-fifth in reproduction.

PLATE 1

EXPLANATION OF FIGURES
*

1 Fibers from ventricular tissue of 76-mm. pig embryo, in which the dises
first make their appearance. The bands here are short or merely granular dots
in line with the anisotropic bands.

2 Portion of fiber from 76-mm. pig embryo, showing a completely developed
disc.

3 Type of disc commonly found in an embryo of 89-mm.

4 A dise taking the step formation, only rarely found in the 89-mm. stage.

5 Fibers from heart muscle of the 115-mm. pig, showing the variable width
and the larger number of complete discs.

6 Fibers showing terminal anastomosing and a disc running across all four
fibers (182-mm. embryo).

7 Fibers showing the definite unstained cytoplasm which surrounds the
nuclei in older heart tissue almost invariably. Also developing and completed
discs are shown which explains how discs become more numerous (201 mm.).

8 A definite example’of the step and riser found in 238-mm. heart muscle.

j44

DEVELOPMENT OF DISCS, HEART MUSCLE OF PIG
: LUCILE WITTE

To
TTA

TI
TLL
fuse

wu Rit
wu
mo

we |
WW
TU

ty nn

\ ne

\\

NOME ggg LHL
iw

avant
wind

ii
\
hh |

Ht

345

CAT

UD ae
LA

PLATE 1

PLATE 2

EXPLANATION OF FIGURES

9 Another example of the ‘step’ and ‘riser’ forms of discs found in 238-mm.
heart muscle.

10 Dises from the 251-mm. stage. They are thicker at this stage than in
the earlier stages.

11 Fibers from the 277-mm. stage, showing a typical grouping of discs.

12 Fibers from the 303-mm. stage, showing a more developed ‘step’ and ‘riser. y

13 Fibers from five-month tissue in which the discs appear as granules laid
down on the separate fibrils.

14 Three parallel fibers from five-month tissue in which a peculiar type of
zigzag dise occurs.

15 and 16 Zigzag discs as found in year-old tissue. These are very charac-
teristic of this tissue.

17 and 18 These figures were drawn from 277-mm. tissue at a magnification
of 2000 diameters, to show a larger area and thus show the comparison between
the number of discs in a contracted area (17) and a relaxed area (18).

346

VELOPMENT OF DISCS, HEART MUSCLE OF PIG
LUCILE WITTE

WAKA iinet

TY AT i
Lint HEA AN anne set alan
we a My

oases
Suit

I
invA HH 7 7
10
eae
Ptr
Avani
TC

A ee
UT +
iis ‘i HETUTTN]

Y
yy Vg
Wi
L] EL)
dy

347

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JUNE 30.

THE HOMOLOGIES OF THE MAXILLARY AND
VOMER BONES OF POLYPTERUS

EDWARD PHELPS ALLIS, JR.

Menton, France

EIGHTEEN FIGURES (THREE PLATES)

In a work published in 1900, on the maxillary and premaxillary
bones of Polypterus, I came to the conclusion that those bones
were each formed by the fusion of two components usually found
separate and distinct in the Holostei and Teleostei, one of those
components being developed in relation to the teeth that the
bone in question bears, and the other in relation to adjacent
portions of the laterosensory canals. The teeth on the maxillary
were said to not be the homologues of the maxillary teeth of
Amia and the Teleostei, and to be, in all probability, dermo-
palatine ones and the homologues of the maxillary teeth of the
Mammalia. The so-called vomer was considered to be de-
veloped in the maxillary breathing-valve of the fish.

In a later work (Allis, 714), I described in certain of the
Selachii a fold of the mucous lining of the buccal cavity that
enclosed the mesial portion of the palatine process of the
palatoquadrate. This fold had so strikingly the position of the
maxillary breathing-valve of the Teleostei that I suggested that,
in the latter fishes, the cartilage enclosed in the fold had been
resorbed, thus leaving the fold itself as the breathing-valve of
the fish. The so-called vomers of Polpyterus, considered to have
been developed in this fold, were then mesial dermopalatines,
and the maxillary bone a lateral dermopalatine.

In a still later work (Allis, ’17, ’18), I described the lips and
labial folds in certain fishes, and came to the conclusion that the
definitive lips of the Crossopterygii, Holostei, and Teleostei were
secondary ones that had been developed external to the primary
lips, and that that part of the buccal cavity that lies between

349
\

350 EDWARD PHELPS ALLIS, JR.

these two lips was primarily a part of the external surface of the
head. The teeth on the maxillary and premaxillary bones were
said to have been developed in relation to the upper one of these
two secondary lips, and hence to lie external to the primary lip.
The upper labial cartilages of the Selachii were said to occupy
a similar position in relation to the secondary lip of those fishes,
and, whenever a labial fold had been differentiated, to both lie
in that fold.

The maxillary and premaxillary bones of fishes, like the
tooth-bearing bones of all vertebrates, are generally considered
to have been primarily formed by the fusion of the bases of the
teeth they bear. Gaupp (05), however, says that while these
tooth-bearing bones of fishes were undoubtedly thus primarily
developed, they frequently later become so completely emanci-
pated from the teeth to which they owe their origin that they
develop wholly independently of them, never, in certain cases,
having, ontogenetically, any relation to teeth of any kind, while
in other cases the teeth and bone may develop independently of
each other, and the teeth later become implanted on the bone.
Gaupp then further says that the maxillary-and premaxillary
bones of vertebrates in general consist of two components, a
tooth-bearing one derived from fusion of the basal plates of the
teeth they bear, and a non-toothbearing, or facial portion, of
independent integumental origin.

The tooth-bearing bones of the buccal cavity are said by
Gaupp to probably all have been primarily developed in topo-
graphical relations to underlying parts of the preexisting car-
tilaginous skeleton, for there would be no sense in their being
developed where there was no firm support beneath them. _ The
underlying structures in relation to which the maxillary and
premaxillary were developed are said to have probably been,
primarily, cartilages belonging to the category of the labial
cartilages, but that, later, secondary relations to the ethmoidal
cartilage were acquired. Gegenbaur (’98) apparently held a
similar view, and he considered the underlying cartilages in
relation to which the maxillary and premaxillary were re-
spectively developed to be the anterior and posterior upper

MAXILLARY AND VOMER BONES OF POLYPTERUS 351

labials of the Selachii. Teeth could, however, evidently have
been developed in relation to any other firm underlying support
found in the region included between the primary and secondary
upper lips, for, that whole surface having been primarily a part
of the external surface of the head, no one part of it was probably
better suited than another to the production of teeth. The
maxillary teeth of Polypterus would then not necessarily belong
to the primary dental arcade, and hence be dermopalatine ones,
as I formerly concluded, simply because they occupy a position
definitely internal to that of the teeth of Amia.

This view of the subject has led me to reexamine my Polypterus
material, and as I find that the labial folds and furrows of that
fish were not properly identified in my earlier work, they will be
fully redescribed after first describing the condition in Acanthias,
specimens of which I have received since my recent work (Allis,
18) was sent to press, and also those in Amia. The conditions
in Lepidosteus, Polyodon, and certain of the Teleostei will then
be considered, and the maxillary bones of these several fishes
then compared with those in certain of the early fossil fishes.

ACANTHIAS BLAINVILLII

In this fish (figs. 1 and 2) the labial fold of either side extends
dorso-anteriorly beneath the supralabial fold, and the latter
fold, pressing upon the labial fold, separates from its oral edge a
smaller fold which is all that is seen in external views. This
smaller fold encloses the anterior upper labial, and as it is repre-
sented, in Polypterus and certain of the Teleostei, by a flap-like
structure, it may be called the maxillary labial flap. The slight
furrow that, in Acanthias, separates this fold from the remainder
of the labial fold is then the maxillary labial-flap furrow, and it
runs anteriorly into the supralabial furrow. The supralabial
furrow has, as stated in my earlier work (Allies, ’18), supra-
labial and postlabial portions, and my present work shows that
these two parts of the furrow are, in certain fishes, distinctly
different morphological structures.

In the lower jaw there is a mandibular labial flap and a labial-
flap furrow similar to those in the upper jaw, but much more

abe ‘EDWARD PHELPS ALLIS, JR.

strongly developed. This mandibular labial flap and labial-flap
furrow both lie external to the mandibular labial cartilage, and,
because of this, they were called, in my descriptions of other
Selachii (Allis, 718), the supramandibular fold and furrow, but
the names above given to them seem more appropriate. Another
deep furrow lies internal to the mandibular labial cartilage, and
is the mandibular furrow of my recent work, but sublabial fur-
row seems a better name for it. This sublabial furrow is con-
tinuous, posteriorly, with the postlabial furrow, a labial fold with
both maxillary and mandibular portions thus being formed.

A deep furrow extends upward between the hind ends of the
anterior and posterior upper labial cartilages, and extending
laterally (morphologically posteriorly) from the secondary angle
of the gape across the labial fold, falls into the postlabial furrow.
This furrow thus separates the lateral (morphologically posterior)
portion of the labial fold into two parts, a superficial one which
encloses the anterior upper labial and a deeper one which en-
closes the posterior upper and mandibular labials. This furrow
may be called the submaxillary one, because in the Holostei
and certain of the Teleostei, it lies beneath the hind end of the
maxillary bone. In the fishes described in my earlier work
(Allis, ’18) this furrow was but slightly developed, this being due
to the fact that the hind end of the anterior upper labial there
articulated with, or was bound by ligament to, either the
posterior upper or mandibular labials.

The primary upper lip of Acanthias is strongly developed, and
between it and the pterygopalatine teeth there is a deep sulcus,
which may be called the primary superior alveololabial sulcus.
A maxillary preangular crease, such as I described in my recent
work, runs symphysially and aborally from the angle of the
primary gape, and gradually vanishes on the oral surface of the
labial fold, this crease lying external to the primary upper lip,
but internal to the posterior upper labial. If teeth were to be
developed internal (oral) to this preangular crease, they would
evidently still be secondary teeth, but they would not be the
homologues of teeth developed along the edge of the labial fold.
The crease is continuous, posteriorly, with the anterior end of

MAXILLARY AND VOMER BONES OF POLYPTERUS 353

a deep furrow that lies along the lateral edge of the posterior
portion of the palatoquadrate, and which may be called, for
reasons that will later appear, the dorsolateral diverticulum of
the buccal cavity. In the other Selachii that I have examined
this diverticulum was not so pronounced and evident as in
Acanthias.

Internal to the primary teeth of Acanthias, there is, as in
Mustelus (Allis, 714, p. 355), a relatively large groove, from
the bottom of which a thick fold of tissue arises and projects
posteroventrally. I formerly considered this entire groove of
Mustelus to represent the suprapalatine recess of my earlier
descriptions of Chlamydoselachus, but a reexamination of my
material shows that it is the furrow that lies postero-dorsal to the
fold that alone represents that recess, and this furrow is in Acan-
thias, as in Mustelus, simply a postpalatine groove and not a
suprapalatine recess. Which of these two conditions is the more
primitive, I am unable to determine, but the fold of Acanthias,
which may be called the palatine fold, is certainly the homologue
of the mucous fold that forms the posterior portion of the palatine
shelf of my descriptions of Chlamydoselachus. This fold is
found in all of the several species of the Selachii that I have
examined, but it varies greatly in different species. In Scyllium
it is a simple flap which strikingly resembles the maxillary breath-
ing-valve of the Teleostei, but is related to the primary dental
arcade instead of to the seconary one. In Lamna and Triakis
it is as in Mustelus and Acanthias. In Centrina it is, as in
Scyllium, a simple flap, but it projects externoventrally instead
of posteriorly, and there is no furrow along its posteromesial
edge. ~ |

AMIA

In this fish the conditions were examined in specimens varying
from 10 mm. to 45 mm. in length.

The supralabial furrow first appears as a pit-like depression
that lies dorsal to the angle of the gape and immediately ventral
to the suborbital laterosensory line, as shown in my figure of
a 10-mm. specimen of this fish (Allis, ’89, fig. 4, pl. 31). In

354 EDWARD PHELPS ALLIS, JR.

a 114-mm. specimen (l.c., fig. 6, pl. 31), this pit-like depression
has become a long slit-like furrow which extends posteriorly
beyond the angle of the gape and there vanishes on the outer
surface of the cheek. The posterior portion of the upper lip,
which in the 10-mm. specimen was nearly straight, has, in the
114-mm. one, grown ventrally so that it overlaps externally the
dorsal edge of the mandible. This overlapping portion of the
lip forms the ventral half of the future labial fold, and its in-
ternal surface is formed by part of the internal surface of the
secondary lip. The maxillary preangular crease of the Selachii
should accordingly, if present, be found on this surface. In the
lower lip a furrow has been developed which, for reasons to be
later given, would seem to be a labial-flap furrow and not a
sublabial one.

In later stages the hind edge of that part of the upper lip that
overlaps externally the dorsal edge of the mandible is prolonged
posteriorly, and becomes separated from the side of the head by
a crease, rather than furrow, which extends a short distance
posterior to the angle of the gape and corresponds to the sub-
maxillary furrow of Acanthias. A postlabial furrow is then
formed, this furrow extending downward from the supralabial
furrow to the hind end of the submaxillary crease. The hind
end of the entire labial fold then grows posteriorly, and the fold
is completely differentiated, this apparently taking place at
different ages in different specimens, for the completed fold is
shown in my figure of an 18-mm. specimen, but not yet com-
pletely differentiated in a 3l-mm. one (l.c., figs. 12 and 13,
pl. 34). In these older specimens the supralabial furrow extends
anteriorly to the anterior end of the lachrymal bone, and there
ends immediately posterior to the articular head of the maxillary.
It is, however, the relation to the lachrymal, and not that to the
articular head of the maxillary, that determines the length of
the furrow, for when the lachrymal has an anterior extension, as
in Gadus, the furrow continues to its anterior end.

On the roof of the mouth (figs. 15 to 18), immediately external
to, and concentric with, the pterygo-palato-vomerine teeth, there
is a sulcus which is the homologue of the primary superior

*

MAXILLARY AND VOMER BONES OF POLYPTERUS 355

alveololabial suleus of Acanthias. This sulcus .begins, pos-
teriorly, somewhat anterior to the transverse plane of the angle
of the gape, and is there a’slight furrow along the lateral edge
of the ectopterygoid, but farther forward it becomes a small but
well-developed groove. Still farther forward, in the region of
the anterior end of the dermopalatine, the internal edge of this
groove begins to flatten out, the external edge at the same time
increasing in height and becoming the slightly developed maxil-
lary breathing-valve. That valve is thus formed by the an-
terior portion of the primary lip of the fish, the posterior portion
of that lip being represented in the lateral edge of the slightly
developed primary alveololabial sulcus. In a 45-mm. specimen,
the line of origin of this lip lies wholly internal to that of the
labial fold. In one of two adult specimens examined, its external
surface had partly coalesced with the internal surface of the
fold in the region of the anterior end of the maxillary, this
undoubtedly being an initial step in that attachment of the base
of the breathing-valve to the internal surface of either that
bone or the premaxillary that is probably found in most, but
certainly not in all, of the Teleostei, for in small specimens of
Ameiurus this fusion does not take place, the primary lip
(breathing-valve) having strictly the course and position that it
has in Amia, and everywhere projecting as a definite lip.

Lateral to the posterior portion of the primary upper lip,
_there is a maxillary preangular crease, which extends upward
lateral to the bottom of the supralabial furrow, these two furrows
enclosing between them a thin sheet of integumental tissue which
connects the dorsal end of the labial fold with the side of the
head slightly dorsal to the lateral edge of the roof of the mouth.
This crease is directly continuous, posterior to the angle of the
gape, with a dorsally directed diverticulum of the buccal cavity
which lies along the lateral edge of the palatoquadrate, and, con-
tinuing posteriorly, vanishes in the region of the angle of the
primary gape. This diverticulum is thus the homologue of the
deep furrow that lies in Acanthias, along the lateral edge of
the posterior portion of the palatoquadrate of that fish, and in
both fishes it lies posterior to the secondary angle of the gape.

356 EDWARD PHELPS ALLIS, JR.

The dorsal end of the preangular crease forms the mesial
boundary of a slight ridge, or fold, of the tissues at the dorsal
edge of the internal surface of the labial fold, this ridge enclosing
a ligament which extends from the internal surface of the maxil-
lary bone, near its anterior end, to the top of the coronoid
process of the mandible (Allis, ’97, p. 548).

The secondary upper lip, represented in the ventral edge of
the labial fold, is continued forward to the anterior end of the
snout, where it is continuous with its fellow of the opposite side.

The labial fold of Amia is thus a strictly maxillary fold, con-
taining no mandibular component and differing markedly in this
from the fold of Acanthias. The maxillary labial flap of the
latter fish is also wanting in Amia. It nevertheless seems un-
questionable that the supralabial and postlabial furrows of the
two fishes are homologous, and, this being so, the deeper part
of the maxillary portion of the fold of Acanthias, the part that
encloses the posterior upper labial, must be contained in the
fold of Amia, and it is probably represented in that little ridge
on the internal surface of the fold of Amia that encloses the
ligament just above referred to. The mandibular portion of the
fold of Acanthias must then be represented either in some part
of the mandible itself of Amia or in the fold of tissue that lies
dorso-external to the little furrow that I have described in the
mandible. That fold has decidedly the position and appearance
of a labial flap, and it contains no tissues that would seem to
represent a mandibular labial; and, furthermore, there is, in
Amia, a tall coronoid process of Meckel’s cartilage that may,
perhaps, represent a mandibular labial. That process is not
found, according to Pollard (’95, p. 413), in any of the Selachii,
and in most, if not all, of those fishes there is a mandibular labial.
The process lies external to, or gives insertion to, all parts of the
adductor muscles of the mandible; and its dorsal end lies pos-
terior to, and approximately in the line of, the secondary angle
of the gape. This process and the stout ligament that arises
from its dorsal end and runs forward in the little ridge at the
dorsal edge of the internal surface of the labial fold, thus have
to each other and to the adjacent parts quite closely the relations

MAXILLARY AND VOMER BONES OF POLYPTERUS BY A

of a mandibular and posterior upper labial, and it is quite
probable that they represent those labials.

The maxillary bone of Amia lies in the labial fold, and al-
though it is found in 12-mm. specimens, the teeth related to it
do not appear until much later, when the fish is about 40 mm.
in length, and they are implanted only along its ventral edge
(Allis, ’00, p. 273). This bone is thus certainly not developed,
ontogenetically, in any relation whatever to those teeth, and, in
my opinion, it is a dermal bone strictly similar, in origin, to the
bones that cover the cheek of the fish. That it may have been
primarily developed in some relation to underlying labial car-
tilages is possible, but it seems to me improbable. The teeth
that later become implanted upon it were, however, quite
probably primarily developed in relation to an anterior upper
labial, for the primordium of that labial must lie in the ventral
portion of the labial fold. That these teeth and the dermal plate
on which they are actually implanted were primarily independent
structures would seem to be further shown by the conditions
found in most of the Teleostei, for, in those fishes, the dental
component of the bone of Amia has become attached to the hind
end of the premaxillary, and forms a posterior prolongation of
that bone, the dermal component of the bone of Amia becoming
an independent and always non-dentigerous maxillary.

The maxillary and premaxillary teeth all issue from the
secondary lip slightly internal to its ventral edge, and there is
no sulcus formed external to them. The maxillary teeth are
much smaller than the premaxillary ones, and seem to be a
later and wholly independent acquisition. The premaxillary
teeth have the appearance of completing the pterygopalatine
arcade; as if they had primarily occupied the existing interval
between the dermopalatine teeth of opposite sides, and had
then been carried bodily forward onto the ventral edge of the
secondary upper lip, as seen in my figures of this fish (Allis,
98). But if this had been the case, the fold that I consider
to be the primary upper lip would evidently be formed by the
palatine fold of Acanthias, and that cannot be, for the fold of
Amia is continued posteriorly external to the external edge of

308 EDWARD PHELPS ALLIS, JR.

the palatoquadrate, while the fold of Acanthias lies, throughout
its entire length, posteromesial to that cartilage. There ap-
parently is in Amia no fold that corresponds to the palatine
fold of Acanthias.

The vomer bones of Amia le posterior to the primary upper
lip, and hence are ossifications in the roof of the buccal cavity.
The dermopalatine is a lateral dermopalatine.

POLYPTERUS

In Polypterus there are, as is well known, thick fleshy upper
and lower lips, each of which has the appearance of having been
folded back upon the remainder of the lip. In my earlier work
on this fish (Allis, 700), I considered these apparently folded-
back portions of the lips to, alone, represent the maxillary and
mandibular labial folds, but it was said that there were, internal
to each of these folds, another smaller one. I now find that the
so-called labial folds of these earlier description are simply labial
flaps, apparently strictly similar to those above described in
Acanthias, and that they and the so-called smaller folds that lie
internal to them together form the entire labial fold. The folds
and furrows of this fish must accordingly be fully redescribed,
and for this purpose I have made use of serial sections of a 75-mm.
specimen of Polypterus senegalus, certain of the sections being
' reproduced in the accompanying figures 5 to 14. Figure 3 gives
a lateral view of the head of a small adult Polypterus, with the
mouth slightly opened, and figure 4 a similar view with the
mouth forced widely open.

The maxillary labial-flap furrow, which separates the maxil-
lary labial flap, above referred to, from the remainder of the
labial fold, begins immediately anteromesial to the nasal tube,
encircles the Jateral edge of that tube and then continues pos-
teriorly to the ventral corner of the labial cartilage, where it
ends external to that cartilage. The maxillary labial flap,
which lies external to this furrow, thus encloses no portion of
the labial cartilage, and there is no other cartilage or bone en-
closed in any part of it. Otherwise, the flap and furrow ap-
parently correspond strictly to those in Acanthias, excepting in

MAXILLARY AND VOMER BONES OF POLYPTERUS 359

that they have a greater anterior extension, which is doubtless
due simply to the greater extension of the secondary lip.
Slightly posterior to the nasal tube, the labial-flap furrow
crosses the anterior section of the main infraorbital laterosensory
canal, passing across the primary tube that issues from that
eanal between the lachrymal bone and the antorbital process of
the premaxillary. If the furrow were to be here deepened, it
would evidently interfere with the development of the latero-
sensory canal, and this is apparently what has taken place in
many of the Teleostei, for a furrow is there found that is either
the homologue of this labial-flap furrow or is a closely adjacent,
but independent furrow, and it extends across the dorsal surface
of the anterior end of the snout to fall into its fellow of the
opposite side. In all such fishes, so far as I know, the antorbital
section of the infraorbital canal is short, or even wholly want-
ing, and associated with this absence of the canal there is no
independent antorbital bone. Furthermore, it is, so far as my
material permits me to judge, only in these fishes that the pre-
maxillary is prolonged more or less posteriorly into the labial
fold, and that the upper jaw becomes more or less protrusive.
The supralabial furrow of Polypterus begins near the anterior
edge of the eyeball, along the ventral edge of the lachrymal
bone, as a slight furrow which lies internal to the dorsal edge of
the maxillary labial flap. Proceeding posteriorly from there,
this furrow deepens and soon becomes a deep cleft which arches
upward beneath the laterosensory component of the maxillary
bone and then downward to a line that lies, in its anterior por-
tion, lateral to the bases of the maxillary teeth, but, in its
posterior portion, along the lateral edge of the ectopterygoid.
On the roof of the mouth, internal to the labial fold and im-
mediately external to, and concentric with, the premaxillo-
maxillary dental arcade, there is a furrow which, because of its
relations to that arcade, may be called the secondary superior
alveololabial sulcus. The bottom of this sulcus is, in the region
of the anterior end of the maxillary labial-flap furrow, directed
toward the bottom of that furrow, the bottoms of the two fur-
rows being separated from each other by a narrow band of in-

360 EDWARD PHELPS ALLIS, JR.

tegumental tissue which connects the labial fold with the lateral
surface of the head immediately dorsal to the bases of the maxil-
lary teeth. Proceeding posteriorly in the sections, these two
furrows deepen, the bottom of the maxillary labial-flap furrow
passing downward lateral to that of the secondary alveololabial
sulcus and diverging gradually from it. The narrow band of
integumental tissue that connects the anterior end of the labial
fold with the side of the head thus becomes considerably widened,
and, at the same time gradually thickening, becomes the deeper
and larger part of the maxillary portion of the entire labial fold.
Farther posteriorly, where the anterior end of the supralabial
furrow is first cut in the sections, the bottom of that furrow is
directed toward the bottom of the secondary alveololabial sulcus,
the bottoms of these two furrows there enclosing between them
a thin and narrow band of integumental tissue which connects
the labial fold with the side of the head. Proceeding posteriorly
from there, the supralabial furrow deepens, its bottom passing
downward lateral to that of the alveololabial sulcus, and the
anterior end of another furrow is soon cut in the sections. This
latter furrow lies lateral to the secondary alevololabial sulcus
and lateral also to the bottom of the supralabial furrow, and as
its course and position indicate that it is quite certainly the
homologue of the maxillary preangular crease of Acanthias, it
may be so designated. The supralabial furrow now gradually
becomes arched in transverse section, running at first upward
and then downward beneath the canal-bearing component of the
maxillary bone, and when the region of the lateral process of the
ectopterygoid is reached, the dorsal portion of the labial fold is
wholly concealed beneath the overhanging, canal-bearing com-
ponent of the maxillary. The anterior portion of the labial car-
tilage is here cut in the sections, and the sheet of integumental
tissues that connects the dorsal end of the labial fold with the
side of the head is, when the mouth is closed, markedly U-shaped
in transverse section. When the mouth is opened and the labial
fold extended, this ‘U’ is pulled out into a flat sheet, but its
proximal and distal (labial) edges are still marked, respectively,
by the bottoms of the secondary alveololabial sulcus and the
maxillary preangular crease.

MAXILLARY AND VOMER BONES OF POLYPTERUS 361

Internal (posteromesial) to the maxillary and premaxillary
teeth, between them and the external edges of the ectopterygoids
and so-called vomers, there is a well-developed sulcus, and
internal (posteromesial) to the vomers there is a depressed region
bounded on either side by a suleus which passes slightly be-
neath (dorsal to) the vomer and ends, posteriorly, along the
mesial edge of the anterior end of the ectopterygoid. This latter
sulcus thus certainly corresponds to the postpalatine furrow of
Acanthias, and the vomers certainly lie, in part, in a palatine
fold. The only question then is, does the furrow that, in Po-
lypterus, lies between the vomer and the premaxillary teeth
correspond to the furrow that, in Acanthias, les between the
palatine fold and the palatine teeth, or to the sulcus that lies ex-
ternal to the latter teeth; and there seems no question that it
corresponds to the latter sulcus, and hence is a primary superior
alveololabial sulcus, for its posterior continuation lies definitely
along the lateral edge of the ectopterygoid.

The premaxillomaxillary dental arcade of Polypterus is thus a
secondary one, but it apparently lies directly along the line of
the primary lip, instead of anterior to that lip. A slight sulcus,
only, separates the premaxillary teeth from the ventral edge of
the secondary lip, this suleus becoming, opposite the maxillary
teeth, the well-developed secondary alveololabial sulcus already
described. Posterior to the most posterior maxillary tooth, this
secondary alveololabial sulcus and the primary one are sepa-
rated, for a certain distance, by a ridge of the mucous lining
membrane of the roof of the mouth, but this ridge gradually
diminishes posteriorly, and vanishes at the anterior edge of the
lateralprocess of the ectopterygoid. The primary and secondary
alveololabial sulci are then represented, for a certain distance, by
a single sulcus, which passes ventral to the lateral process of the
ectopterygoid and vanishes posterior to the angle of the gape.

The fold of the secondary upper lip has thus been carried
forward until it falls into its fellow of the opposite side at the
anterior end of the snout, but in its anterior portion it passes so
close to the primary lip that it is there but slightly differentiated
from it. In accord with this, and also in confirmation of it,

362 EDWARD PHELPS ALLIS, JR.

there is, in this fish, no maxillary breathing-valve, that valve
being formed, as already explained, by the anterior portion of
the primary lip.

In the lower jaw there are two furrows that correspond, in
their relations to the lower lip, to the supralabial and maxillary
labial-flap furrows in the upper lip. One of these furrows
separates the dermal flap of the lower lip from the remainder of
that lip, and is the mandibular labial-flap furrow. Like its
fellow in the upper lip, it begins far forward and extends pos-
teriorly to the ventra end of the labial cartilage, being deep in
its middle portion and vanishing at either end. The other
furrow is a sublabial one, and separates the mandibular portion
of the labial fold from the external surface of the mandible. It
begins anteriorly in about the transverse plane o! the anterior
end of the supralabial furrow, and there lies, like that furrow,
internal to the free edge of the related labial flap. Posteriorly
it becomes deeper, and at its hind end passes upward beneath the
hind end of the labial fold and is confluent with the hind end of
the supralabial furrow. These two mandibular furrows lie, the
one directly beneath the lateral edge of the oral surface of the
lower lip and the other beneath the mesial edge of that surface
(figs. 8 to 11). The lip here presents, in transverse sections,
three surfaces, one presented mesially, one dorsolaterally, and
the other ventrolaterally. The surface that is presented dorso-
laterally increases in width posteriorly, and at the angle of the
gape coalesces with the internal surface of that posterior portion
of the maxillary portion of the entire labial fo'd that encloses the
labial cartilage. Along the median line, approximately of this
surface of the lower lip, there is a furrow formed by a fold in the
lip that is related to its attachment to the top of the ascending
process of the splenial, this furrow corresponding approximately
to the mandibular preangular crease of Acanthias. Internal to
the lip, between it and the dentary teeth, there is a secondary
inferior alveololabial sulcus, and between the dentary and
splenial teeth a primary inferior alveololabial sulcus.

When the upper and lower lips have coalesced at the angle of
the gape, as above explained, the labial fold projects ventro-

MAXILLARY AND VOMER BONES OF POLYPTERUS 363

laterally, its ventral end being free and its dorsal end attached
to the side of the head by two sheets of integumental tissue, one
of which extends to the lateral edge of the ectopterygoid and the
other to the dorsal edge of the mandible. These two sheets of
tissue enclose between them a dorsolateral diverticulum of the
buccal cavity similar to that in Amia, and it corresponds, in
position, to what Greil (713) calls, in Ceratodus, the preman-
dibular furrow (Praemandibularfalte). This furrow is a direct
posterior continuation of the maxillary preangular crease, and,
diminishing in depth posteriorly, vanishes near the quadrato-
mandibular articulation. Posterior to the angle of the gape,
the ascending process of the splenial rises in the integumental
sheet that connects the dorsal end of the labial cartilage with the
dorsal edge of the mandible, and the dorsal end of the labial now
becomes attached to the dorsal end of this process. This
process of the splenial is called by Traquair (’70) the opercular
process, but it is the homologue of that process of reptiles,
formed by the so-called complementary, and not of the opercular
process of Amia, which is a process of Meckel’s cartilage and
wholly wanting in Polypterus.

The labial cartilage is, as described in my earlier work, some-
what triangular in shape, with a long straight dorsoposterior
edge, and anterior and anteroventral edges that are somewhat
rounded and separated by a rounded angle. The ventral corner
of the cartilage is also rounded, but a pad of tough ligamentous
tissue lies against the posterior surface of this rounded angle
and gives to the entire structure a right-angled postero-
ventral corner, the posterior edge of this angle lying in the line
of the straight posterior edge of the labial cartilage, while the
ventral edge forms a short straight ventral edge to that cartilage.
The anterior end of this so-formed short ventral edge of the
labial cartilage lies directly posterior to the secondary angle
of the gape, its hind end forming the ventral end of the posterior
edge of the labial fold. There is no furrow running from the
angie of the gape across the labial fold and corresponding to the
submaxillary furrow of Acanthias. When the mouth is widely
opened (fig. 4), the labial cartilage lies in a nearly horizontal

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 4

364 EDWARD PHELPS ALLIS, JR.

position, the hind end of the labial fold, in which the cartilage
lies, forming a narrow horizontal shelf which projects posteriorly
from the angle of the gape. Anterior to this cartilage, the oral
edges of the maxillary and mandibular portions. of the labial
fold lie, at first, at right angles to the cartilage, and then each
curves gradually forward toward the anterior end of the jaw
to which it is related.

The maxillary and dentary bones are both somewhat scooped
out, on their external surfaces, to lodge the related portions of
the labial fold, this being particularly marked on the maxillary.
No bone is found in any portion of the labial fold. The labial
cartilage is, in my 75-mm. specimen, apparently of cartilage that
is strictly similar to the other cartilages of the head.

Posterior to the hind end of the labial fold, the supralabial
and sublabial folds coalesce, excepting as they are in part sepa-
rated by a thin web of integumental tissue, and the dorsal edge
of the one furrow and the ventral edge of the other meet at an
acute angle which forms what I called, in Ceratodus (Allis, 718),
the tertiary angle of the gape. The lines that bound this angle
in Ceratodus I considered to represent tertiary upper and lower
lips, the upper lip alone becoming functional and forming the
definitive upper lip of the fish. The secondary upper lip was
said to be hardly recognizable in my specimen of Ceratodus,
and to be found at the angle of the gape. My present work
tends to confirm this conclusion, and I now definitely identify
the secondary lip in the short fold shown lying immediately
anterior to the secondary angle of the gape in the figure given
in my earlier work (l.c., fig. 9). I, however, now find that
the exposed lateral edge of the labial fold of Ceratodus corre-
sponds to the ventral edge of. that part of the fold of Polypterus
that lies posterior to the secondary angle of the gape, and not,
as I formerly concluded (Allis, ’00), to the dorsal edge of the
maxillary labial-flap. The labial fold of Ceratodus has accord-
ingly not been turned upward and inward upon the roof of the
buceal cavity, as I formerly concluded. It has simply acquired
a permanently nearly horizontal position, similar to that in
Polypterus when the mouth is widely opened, and has then

MAXILLARY AND VOMER BONES OF POLYPTERUS 365
coalesced with, or never been completely separated from, the
tissues of the roof of the mouth.

The lips and labial folds and furrows of Polypterus are thus
strictly comparable to those of Acanthias, and the labial fold is,
as in that fish, a maxillomandibular one. The labial cartilage,
which lies in that fold, must then be the homologue of one or
more of the labial cartilages of Acanthias. If the labial flaps of
Polypterus are the homologues of those of Acanthias, as seems
so probable, the anterior upper labial must be greatly reduced,
if not wholly wanting, in Polypterus; and as the labial cartilage
of Polypterus seems definitely related to the upper lip, the
mandibular labial must also be greatly reduced, or wholly want-
ing. The labial cartilage of this fish is thus certainly in large
part a posterior upper one, remnants only, at the most, of the
other labials persisting. Pollard’s conclusion (’95, p. 407), that
this labial is a coronoid one, is then certainly in error, if, as the
name coronoid would seem to imply, the labial was considered
by him to be a mandibular structure.

The premaxillary teeth of my 75-mm. specimen are implanted
upon a bone that lies in part along the ventral surface of the
anterior end of the ethmoidal cartilage, and in part encloses the
ethmoidal section of the infraorbital laterosensory canal, and
they are apparently, as already explained, secondary teeth and
the homologues of the corresponding teeth of the Holostei and
Teleostei.

The anterior maxillary teeth are implanted upon a plate of
bone that corresponds to the palatal process of the maxillary
bone of the adult, and that lies in a layer of fibrous tissue that lies
slightly internal to the epithelial lining of the buccal cavity.
This plate lies, in its anterior portion, beneath, but not directly
against, the cartilage of the nasal capsule, and in its posterior
portion directly beneath the anterior end of the palatoquadrate
cartilage. It extends mesially beyond the primary alveolo-
labial sulcus, to the mesial edge of the vomer, and the anterior
maxillary teeth arise from its lateral edge and issue, as already
stated, along the line of the primary lip. The bone is somewhat
thickened along its lateral edge, where the teeth are implanted

366 EDWARD PHELPS ALLIS, JR.

upon it, but it is, in this part of its length, wholly separate from,
and independent of, the bone that encloses the overlying section
of the laterosensory canal. Slghtly posterior to the anterior
end of the palatoquadrate cartilage, this dental component of
the maxillary bone loses its mesial, palatal extension, and then
fuses with the overlying canal bone, the teeth now being implanted
upon the so-formed bone, and that bone being separated from
the lateral edge of the ectopterygoid by the primary alveolo-
labial sulcus. The line of the maxillary continues to follow the
line of the primary lip, and when the primary and secondary
sulci fuse with each other at the anterior edge of the lateral
process of the ectopterygoid, the line of the maxillary teeth
comes into contact with the lateral edge of the ectopterygoid,
as described in my earlier work (Allis, 00). These teeth lie
wholly internal to the labial fold, and hence cannot have been
developed in any relation to either of the upper labial cartilages.
The maxillary and ‘premaxillary bones of Polypterus thus both
have dental and facial components, which Gaupp says is char-
acteristic of higher vertebrates, and they seem to certainly be
the homologues of those bones. The maxillary bone of this fish
is thus properly so-called, and as it and the teeth it bears are not
the homologues of the bone and teeth of Amia, the bone and
teeth of the latter fish should be otherwise designated, and I
shall refer to them as. the holostean maxillary bone and teeth,
and to the dental arcade of the fish as a premaxillolabial one.
The so-called vomers occupy a position that corresponds, as
already explained, to that of the palatine processes of the
Selachii plus the palatine fold, and they are accordingly mesial
dermopalatines. These bones and the teeth they bear were
accordingly not developed, primarily, in any direct relation to
the base of the neurocranium, differing radically in this from
the parasphenoid. The anterior end of the basal plate of the
latter bone of Polypterus lies in the roof of the depressed region
between the vomers of opposite sides, that is, in the roof of
what corresponds to the suprapalatine recess of the Selachii.
The vomers of Polypterus lie, morphologically, in the floor of
that recess, and they are, in my 75-mm. specimen, nowhere in

MAXILLARY AND VOMER BONES OF POLYPTERUS 367

contact with the parasphenoid, and their hind ends lie, even in
the adult, ventral to the anterior ends of the ectopterygoids.

The position of these bones of Polypterus thus, in itself, indi-
cates that they are quite probably not the homologues of the
unpaired vomer of the Teleostei, which bone lies in part
directly upon the base of the chondrocranium, and in part
directly upon the anterior end of the parasphenoid, and con-
clusive proof of the non-homology of these bones is given by
the conditions in Macrodon. In the latter fish Sagemehl (’84)
describes an accessory palatine which is said to be but loosely
bound to the other bones of the palatine arch, and to be un-
known in other fishes. I find these bones of Macrodon having
exactly the relations to the other bones of the palatine arch
that the so-called vomers of Polypterus have, and I have twice
suggested (’00, p. 278, and ’09, p. 26) that they might be the
homologues of the latter bones. I, however, then considered
these bones of both these fishes to have been developed in the
maxillary breathing-valve of the fish, which, as my present work
shows, is incorrect in so far as Polypterus is concerned, and ref-
erence to Miiller and Troschel’s (’45) figure of the widely opened
mouth of Macrodon shows that it is incorrect for that fish also,
for what is apparently a well-developed maxillary breathing-valve
is there shown, and the accessory palatine lies posterior to it.

It is thus certain that the so-called vomers of Polypterus do
not contain the homologue of the basal plate of the unpaired
vomer of the Teleostei. The teeth implanted upon that basal
plate may, however, be the homologues of those of Polypterus,
for with the disappearance of the suprapalatine recess and the
disintegration of the mesial portions of the palatine processes of
the palatoquadrates, these teeth of Polypterus would lie directly
superficial to a vomer such as is found in Macrodon, and if they
persisted they would necessarily become implanted upon it. And
in accord with this view, the vomer of Macrodon is non-
dentigerous (Sagemehl, ’84).

The vomer of the Teleostei is thus quite probably formed of
two primarily independent components, one dermal and the
other dental, but this may possibly not apply to the paired

368 EDWARD PHELPS ALLIS, JR.

vomers of Amia. In any event, Sagemehl’s suggestion (’83,
p. 187), that the vomer bone of fishes might have been primarily
the most anterior one of the investing bones of the palatine arch,
is quite certainly true in so far as the teeth on the vomers of

Amia are concerned.
LEPIDOSTEUS

In Lepidosteus there is a small labial fold, similar to that in
Amia, and it encloses an holostean maxillary bone which Parker
(82) refers to as the edentulous mystaceum of an Acanthopter-
ous Teleostean. That part of this bone that lies in the labial
fold is, as Parker states, edentulous, but in both 40-mm. and
80-mm. specimens that I have examined in serial transverse
sections, I find two or more teeth related to an anterior portion
of the bone that extends anteriorly beyond the root of the labial
fold. These teeth lie in the line prolonged of a series of teeth
that are implanted upon those laterosensory canal bones that
Parker refers to as the maxillary chain, but the holostean maxil-
lary is not, as Parker states, one of that chain of bones, for its
anterior end lies definitely ventral to them, and ventral to the
line prolonged of the supralabial furrow, while the canal bones lie
dorsal to that line prolonged, and hence, as they normally
should, in the line of an anterior prolongation of the supralabial
fold. The teeth that are implanted upon these canal bones were,
however, quite certainly not developed in relation to the supra-
labial fold, for the posterior ones pierce the ectoderm ventral
to the anterior end of that fold. These teeth are, as above
stated, implanted upon the canal bones, and they form the row
of small sharp teeth said by Parker (l.c., p. 478) to be implanted
along the outer edges of those bones of the adult. Internal to
these teeth there is, in my embryos, a row of teeth that are not
implanted upon any bone whatever, their bases not reaching
the overlying portions of the canal bones. In the adult most
of these teeth are firmly implanted upon those bones, some
of them, however, still remaining unattached, and, being held
only by the surrounding tissues, are loose and movable. They
form the row of large sharp teeth said by Parker to lie in the

MAXILLARY AND VOMER BONES OF POLYPTERUS 369

long valley between the palatine splints and the maxillary chain
of canal bones. In embryos this long valley is a large and low
rounded ridge of the tissues of the roof of the mouth, with a
slight furrow at either edge, and this ridge probably represents
the primary upper lip. On the palatine splint there are two
rows of teeth, the teeth of the inner row being slightly larger
than those of the outer row.

In my embryos the outer row of teeth related to the canal .
bones are somewhat larger than those of the inner row. In the
adult, on the contrary, the teeth of the inner row are much the
larger. The smaller teeth of the outer row are, in the adult,
directly continuous with the premaxillary teeth, the large teeth
of the inner row ending posterior to the premaxillary teeth, in
practically the positional relation to them that the dermopala-
tine teeth of Amia have to the premaxillary ones. External to
the main row of premaxillary teeth there may be a second, but
shorter row of teeth. These premaxillary teeth are all large
ones, similar to those of the inner row of the canal bones.

There is a well-developed dorsolateral diverticulum of the
buccal cavity, but it is not prolonged anteriorly as a recognizable
preangular crease. ‘There is, as in Amia, a tall coronoid process
to Meckel’s cartilage, and from this process a ligament arises
and has its insertion on the holostean maxillary. There is no
maxillary breathing-valve. Itis therefore quite certain that both
of the rows of teeth related to the canal bones of this fish are
secondary ones, and those of the outer row quite certainly corre-
spond to the maxillary teeth of Amia. The teeth of the inner
row are perhaps similar to those of Polypterus, but it seems
much more probable that they lie external to the primary lip,
along a ridge that corresponds to that ridge on the internal
surface of the labial fold of Amia that encloses the ligament that
I consider to represent the posterior upper labial cartilage.

370 EDWARD PHELPS ALLIS, JR.

POLYODON

Bridge (’79) considered the cartilages in the upper jaw of this
fish to be similar to those in the upper jaw of the Selachii, and
he says that these cartilages of all these fishes are pterygoquad-
rates, without ‘‘even rudiments of recurrent palatine out-
growths,” the palatine being considered by him to be a preoral
_ structure. It nevertheless seems best to refer to the upper jaw
of this fish as a palatoquadrate, that being the term usually
applied to it in the Selachii.

In the one adult specimen of Polyodon that I have examined,
and also in three young specimens 154 to 300 mm. in length, I
find the external surface of the palatoquadrate looking laterally
in its posterior portion, as Bridge states, and anterodorsally in
its anterior portion. Its quadrate portion has the leaf-like out-
growth that Bridge describes and calls its orbitar process, this
process projecting forward lateral to the musculus adductor
mandibulae. The hind end of the so-called maxillary splint is
closely applied to the external surface of the orbitar process, the
surface of the process being slightly excavated to receive it.
For a certain distance anterior to the anterior end of the orbitar
process, the maxillary splint is not in contact with the palato-
quadrate, but farther forward its dorsal edge is V-shaped, fits
closely upon the ventrolateral edge of the palatoquadrate, and
so extends to the point where that cartilage articulates, in the
median line, with its fellow of the opposite side of the head.
The anterior portion of the ventral edge of the splint is furnished
with small teeth, and, extending approximately the same dis-
tance along the dorsomesial edge of the palatoquadrate, there is
a row of somewhat stronger teeth. The bone to which these
latter teeth are attached fits closely upon the related edge of
the palatoquadrate, and although fused anteriorly with the
pterygoid splint of Bridge’s descriptions, it has markedly the
appearance of being of independent origin. It is called by
Gegenbaur (’98, p. 342) the palatine, but it is to be noted that
it is, even in my quite young specimens, definitely of membrane
origin, and that it lies along the dorsomesial, instead of the

MAXILLARY AND VOMER BONES OF POLYPTERUS 371

ventrolateral, edge of the palatoquadrate cartilage. The ptery-
goid bone is also of membrane origin, and I find, in my speci-
mens, no indications of the so-called ectosteal scales described
by Bridge and said by him to probably represent mesopterygoid
elements.

The maxillary splint lies in the upper lip of the fish, and the
posterior portion of this lip overlaps externally the dorsal edge
of the mandible, this overlapping portion of the lip enclosing
the ventral corner of the orbitar process of the palatoquadrate,
but no portion of the maxillary splint. A slight labial sulcus lies
along the external surface of the ventral edge of the anterior
portion of the maxillary splint and separates the teeth on that
splint from the definitive and quite small upper lip. There is
no slightest indication of a supralabial furrow, but that furrow
and the overlying supralabial fold are apparently well de-
veloped in embryos of Acipenser, as shown in my recent work
(Allis, ’18), the fold and furrow there being called the supra-
maxillary ones. In the mandible there is a short furrow
which extends upward a short distance immediately posterior to
the angle of the secondary gape, and it is apparently the homo-
logue of the sublabial furrow of Polypterus and the Selachii,
and in apparent correlation to this there is no coronoid process
to Meckel’s cartilage.

The conditions in this fish are thus similar to those in 114-mm.
embryos of Amia, excepting in that the furrow in the mandible
is probably a sublabial one instead of a mandibular labial-flap
furrow. ‘The posterior portion of the upper lip of Polyodon is
then certainly a secondary lip, and the slightly differentiated
labial fold is probably, as in Polypterus, a maxillomandibular
one. The orbitar process of the palatoquadrate, which lies in
this fold, must then be a labial cartilage, as Gegenbaur (’98
p. 342) concluded, but there seems nothing to indicate that it
is a posterior upper labial alone, as Gegenbaur concluded, rather
than a cartilage formed by the fusion of both upper labials, for
the labial fold certainly encloses the primordia of both upper
labials. This upper labial, or labials, of Polyodon having fused
with the palatoquadrate, supralabial, and submaxillary furrows
could naturally not be developed.

a2 EDWARD PHELPS ALLIS, JR.

The fold of the secondary upper lip, thus well developed in its
posterior portion, probably extended forward far enough to fall
into its fellow of the opposite side on the anterior end of the
snout, but it was probably not there completely differentiated
from the primary lip. Because of this failure to there be com-
pletely differentiated, no premaxillary bone has been developed,
and the maxillary splint has been prolonged anteriorly to replace
it. This latter bone is a superficial one, and, as it extends into
the oral edge of the secondary lip, it is quite certainly an holos-
tean maxillary. The attachment of a maxillary bone to the
ventrolateral edge of the palatoquadrate thus does not, in itself,
necessarily imply that the bone is the homologue of the maxillary
of Polypterus.

A so-called vomer is found in this fish, but it les wholly an-
terior to the buccal cavity, on the ventral surface of the long
spatula-shaped snout. This bone thus actually lies external to
the primary lip, and if it actually be a vomer, its anomolous
position needs explanation.

TELEOSTEI

In these fishes the labial folds and furrows vary greatly, some-
times resembling those in Amia and sometimes those in Polyp-
terus, but usually differing, in certain details, from those in
either of those fishes. In Esox, for example, there is a maxillary
labial fold and a mandibular labial flap that are apparently
similar to those in Amia, but the supralabial furrow lies ata
certain distance dorsal to the labial fold and is definitely a sub-
lachrymal furrow, the dorsal edge of the labial fold being bounded
by an anterior prolongation of the postlabial furrow. This thus
suggests, as does the development of these furrows in Amia, that
the supralabial furrow is primarily a sublachrymal one, and only
secondarily comes to bound the dorsal edge of the labial fold.
In Gadus the labial fold is a maxillomandibular one, as in
Polypterus and Acanthias, but there is a suprapremaxillary fur-
row which lies between the premaxillary and maxillary bones
and is continued anteriorly until it falls into the anterior end of
its fellow of the opposite side on the anterior end of the snout.

MAXILLARY AND VOMER BONES OF POLYPTERUS 373

This furrow is apparently found only in those fishes in which
the premaxillary has that posterior prolongation in the labial
fold, ventral and parallel to the maxillary, that has already
been referred to, and it is associated with a premaxillary that is
more or less protrusive. This suprapremaxillary furrow has a
position markedly similar to that of the maxillary labial-flap
furrow of Polypterus and Acanthias, and it is probably its
homologue, but there may, nevertheless, be two distinctly dif-
ferent furrows here. This suprapremaxillary furrow, the pos-
terior prolongation of the premaxillary ventral and parallel to
the maxillary, the transference of the teeth along the ventral
edge of the labial fold from the maxillary to the premaxillary,
and the acquisition of a protrusive premaxillary are apparently
all correlated features and are distinctly characteristic of the
higher Teleostei, for Sagemehl says (’84, p. 101) that this type
of premaxillary is found in all of the non-physostomous fishes
and in many of the physostomous ones, while the holostean type
of maxillary is found only in the remaining physostomous fishes.

There are thus, in the Teleostei, at least two distinctly dif-
ferent secondary dental arcades in the upper jaw, the teeth in
these two arcades being homologous, but the bones on which
they are implanted totally different. Whether or not a third
and still different arcade is also found in certain of these fishes,
the Apodes, for example, I am unable as yet to definitely de-
termine. These latter fishes are classed with the Physostomi,
and Boulanger (’04) says that the premaxillaries are absent in
all of them. Gegenbaur (’98, p. 357) says that these bones are
present in the Muraenidae, but reduced, and fused with each
other and with the vomer to form the anterior end of the snout.
Smith Woodward (’01) also says that the premaxillaries are
present in the Muraenidae, but he considers them to have there
fused only with the ethmoidal rostrum. Maxillary bones are
said by Boulanger to be present in the Anguillidae and certain
other families of the suborder, but absent in the Muraenidae,
and there replaced by the palatopterygoids. Smith Woodward
considers these so-called palatopterygoids of the Muraenidae to
be maxillaries, and J, in an earlier work (Allis, ’03), concluded

374 EDWARD PHELPS ALLIS, JR.

that they must be palatopremaxillaries, the so-called premaxil-
laries of certain descriptions being maxillary breathing-valve
bones.

Because of these differing opinions regarding these bones of
the Apodes, I have reexamined my old material of Conger, and
I find that the anterior portion of the maxillary bone of that
fish has the relations to the lips and labial folds either of the
maxillary bone of Polypterus or of the inner row of teeth on the
so-called maxillary chain of canal bones of Lepidosteus, while
the posterior portion of the bone apparently has the position
of that ligament of Amia that I consider to apparently represent
the posterior upper labial cartilage of the Selachii. This maxil-
lary bone has, however, fused with some element of the palato-
quadrate, that element being either a dermopalatine or a
dermopalato-ectopterygoid, and bearing, along its ventrolateral
edge, a row of small teeth that lie immediately internal to the
maxillary teeth and are separated from them by a slight sulcus.
This sulcus is continued forward between two groups of teeth
on the ventral surface of the anterior end of the neurocranium,
one of those groups evidently being vomerine teeth and the
other premaxillary ones. The latter teeth are implanted upon
a bone that encloses the two anterior organs of the ethmoidal
laterosensory canal, and the bone so formed has fused with the
vomer and with the anterior end of the neurocranium. There
is in this fish, as in Polypterus and Polyodon, no trace of a
maxillary breathing-valve.

The conditions in the Teleostei thus vary greatly and need
special investigation that my material does not at present permit
me to undertake. They are accordingly here left out of further
consideration.

FOSSIL FISHES

Of these fishes I have only considered certain of the earliest
known Chondrostei and Crossopterygii, the Chondrostei being
the order to which the Holostei, Teleostei, and the recent
Polyodon all belong.

A premaxillary is said by Smith Woodward (’95) to be found
in all of the Palaeoniscidae, the earliest known family of the

MAXILLARY AND VOMER BONES OF POLYPTERUS S10

Chondrostei, and although it is said by him to be comparatively
small and insignificant, it was nevertheless well developed in
certain of these fishes. This indicates that a complete secondary
upper lip had already been developed in these fishes, but pos-
sibly, in certain of them, not yet markedly differentiated from
the primary lip. In Nematoptychius greenocki, the premaxil-
lary is shown well developed by Traquair (’77), and there in
articular contact, posteriorly, with the anterior edges of a large
median supraethmoid and a lateral bone on either side that is
called both a prefrontal and an anterior frontal. This anterior
frontal is traversed by what is apparently the anterior portion
of the supraorbital laterosensory canal, that canal ending im-
mediately lateral to the single nasal opening of these fishes,
approximately in the relation to that opening that it has in
Amia (Allis, ’89) to the anterior nasal opening. The anterior
frontal of this fish is thus apparently a canal bone, and hence, as
Traquair states, cannot be the homologue of the prefrontal
(ectethmoid) of the Holostei and Teleostei. The relations of
the bone to the supraorbital canal would seem to show that it is
the homologue of the nasal of recent fishes, but the bone lies
lateral to the single nasal opening, while the nasal bone of
recent fishes lies mesial to both nasal openings. ‘There is no
indication, in any of the Palaeoniscidae, of a posterior nasal
aperture either mesial or lateral to this bone.

The maxillary of the Palaeoniscidae forms a direct posterior
continuation of the premaxillary, and hence, like the maxillaries
of Amia and Polypterus, is definitely posterior in its relations to
that bone. It has a narrow anterior portion and a broad pos-
terior one. Its dorsal edge is said by Traquair (’77) to be cut
away so as to follow exactly the contours of the suborbital and
postorbital bones, and to be in articular contact with them and
also with the preoperculum. It is not said that these bones
were firmly attached to each other, but this seems implied in
the descriptions, and they are apparently shown so attached in
a figure of the head of Oxygnathus. In that figure the head of
the fish has been flattened out in such a manner that the mouth
is shown widely opened, but the maxillary nevertheless retains

376 EDWARD PHELPS ALLIS, JR.

its contact with the suborbitals, postorbitals, and preoperculum,
excepting only along the ventral end of the latter bone. There
was thus apparently no supralabial furrow in these fishes.

The ventroposterior corner of the maxillary projects ventro-
posteriorly and forms a process-like portion which overlaps
externally the dorsal edge of the mandible, there covering the
quadratomandibular articulation. The ventral margin of the
bone is furnished, throughout its entire length, with teeth. In
Palaeoniscus these teeth are said to be very small and to be
closely set, the more internally placed ones being larger than the
external ones. In certain other genera of the family there is
said to be a set of strong and powerful conical or laniary teeth
placed in a row internal to an outer series of closely and some-
what irregularly placed smaller ones. The ventral portion of
the bone is said by Traquair to be slightly reflected inward so as
to form a sort of narrow ledge, and some part of this ledge is
said to have apparently come into contact with the ventro-
lateral edge of a somewhat narrow and elongated bony lamina
said to represent the palatoquadrate arch. This bony lamina is
said to be apparently composed of two parts, one of which repre-
sents the quadrate while the other is probably the equivalent
of the ectopterygoid of the osseous fishes. This latter part of
the lamina is nevertheless called the palatopterygoid, and it is
with it that the reflected ventral edge of the maxillary is in
contact.

The maxillary of these fishes would thus seem to be a super-
ficial bone that had, as in Polyodon, acquired attachment to the
ventrolateral edge of the palatoquadrate, and the two rows of
teeth on the ventro-internal surface of the bone would seem to
be, one a row of holostean maxillary teeth, and the other a row
corresponding to the inner row of teeth on the bones of the so-
called maxillary chain of Lepidosteus or to those on the maxil-
lary of Conger. With the disappearance of the inner row, the
acquisition of articular connections of the maxillary with the
neurocranium, and the development of a supralabial furrow, the
bone of Amia would arise, while the retention of the inner row
alone might give rise to the bone of Conger.

MAXILLARY AND VOMER BONES OF POLYPTERUS er

The hyomandibula of the Palaeoniscidae is said by Traquair
to be an elongated and slender bone which descends downward
and backward from the squamosal region to the neighborhood
of the quadrate articulation, thus apparently articulating with
the neurocranium dorsal to the vena jugularis, as it does in all
recent Teleostomi; this showing that these early Chondrostei
had already departed markedly from the Plagiostoman type.

In the Platysomidae, a family of the Chondrostei closely
related to the Palaeoniscidae, the teeth were feeble or wholly
wanting (Traquair, ’79). The feeding habits of these fishes
could not then have required a fixed and rigid support for their
teeth, and doubtless in correlation to this the maxillary is not
fixed in position by attachment either to the palatoquadrate or
to the suborbital and cheek bones. That it was not attached
to the cheek bones is evident from the fact that its dorsal edge
is no longer cut away to fit against those bones, as it is in the
Palaeoniscidae, and in Platysomus it even overlaps them ex-
ternally. There must then have been a more or less developed
supralabial furrow in these fishes, and quite possibly it existed
in the condition that I have described in 113-mm. embryos of
Amia. These fishes thus presented conditions favorable to the
development of a maxillary bone such as is found in Amia.

In the Crossopterygii there are apparently two distinctly
different types of maxillary, one resembling somewhat that found
in the Palaeoniscidae and the other that found in the recent
Polypterus. The former type is found in fishes of the suborder
Rhipidistia of Smith Woodward’s classification, and the other in
those of the suborder Actinistia.

In the Rhipidistia there are, according to Smith Woodward
(91), dentigerous premaxillaries which are usually fused with
each other, and also, more or less, with small plates which intervene
between them and the frontals. The maxillary is said to be
_ dentigerous, and to be bounded above by the suborbital and
cheek bones; and in restored figures of these fishes it is ap-
parently shown in articular contact with the latter bones and
frequently tuberculated. In restored figures of Rhizodopus, its
hind end is shown overlapping externally to a considerable

378 EDWARD PHELPS ALLIS, JR.

extent the dorsal edge of the mandible. Smith Woodward says
that there are some traces of an inward palatal extension both of
the maxillary and the premaxillary, and that there are paired
vomers, each of which bears a formidable tooth.

Aside from these very general description of these fishes, I
only have at my disposal Huxley’s descriptions of Glyptolaemus
kinnairdi, and Traquair’s descriptions of Tristichopterus, the
one belonging to the family Osteolepidae and the other to the
family Rhizodontidae.

In Glyptolaemus, called by Smith Woodward Glyptopomus,
the maxillary is not particularly described by Huxley (’61), but
he shows it, in one of his figures, lying along the lateral edge of
the anterior end of the palatosuspensorium and overlapped ex-
ternally by a large tooth-like structure which apparently either
issues from beneath the suborbital bones, or has its origin on the
pa’atosuspensorium. If this be the ‘very strong tooth’ mentioned
by him in the text, and there said to be implanted on the anterior
portion of the palatosuspensorium, it must be greatly displaced
in the fossil. It somewhat strikingly resembles, in general posi-
tion and appearance, the labial cartilage of Polypterus, and if it
be that labial, here ossified, as it is in Macropoma, the maxillary
of this fish cannot be an holostean one. The maxillary is, how-
ever, shown strongly tuberculated in restored figures of certain
of these fishes, particularly so in Osteolepis, and if it was a super-
ficial bone, as this would indicate, it cannot be the homologue of
the bone of Polypterus.

In Tristichopterus, the maxillary is said by Traquair (’75) to
be a long and narrow bone, the oral edge of which is distinct in
the fossil, but the dorsal edge not. Internally, along its anterior
two-thirds, it is said to be certainly very firmly united to the
outer margin of the palatoquadrate arch, an interval being left
posteriorly for the passage of the muscles of the lower jaw.
The bone is shown, in a restored figure, in contact with the _
suborbital and preopercular bones, and slightly overlapping ex-
ternally the dorsal edge of the mandible. The palatosuspen-
sorium is shown extending forward approximately to the an-
terior end of the maxillary, and its posterior border, which

MAXILLARY AND VOMER BONES OF POLYPTERUS 379

contains the hyomandibula, gives articulation, externally, to the
hind edge of the preoperculum. The palatosuspensorium is
said to be immovably fixed to the maxillary along the anterior
two-thirds of its outer margin, and at what appears to be the
hind end of this region of attachment—that is, at the posterior
third of the length of the palatosuspensorium—there is a laterally
directed process on its lateral edge which somewhat resembles
the process on the lateral edge of the ectopterygoid of Polyp-
terus; this process coinciding, in the latter fish, with the hind
end of the tooth-bearing component of the maxillary. In two
of Traquair’s figures the maxillary of Tristichopterus is shown
ending at this process, and hence having a much less extensive
posterior prolongation than that shown in the restored figure.

The maxillary of Tristichopterus was thus probably more or
less firmly attached both to the suborbital bones and the pre-
operculum, and firmly attached to the outer margin of the
palatoquadrate. It was therefore certainly incapable of inde-
pendent movement. In this it agreed with the maxillary of
both Conger and Polyodon, and if the bone were tuberculated
it would seem as if it must be the homologue of the bone of one
of these two fishes, and probably of that of Polyodon.

In the suborder Actinistia there is but one family known, that
of the Coelacanthidae, and there has been marked disagreement
as to the homologies of the bones related to the upper jaw of
these fishes.

Macropoma mantelli, one of this family, was described by
Huxley in 1866. The prodtic is said by him to be “a large
plate of bone, rising perpendicularly towards the roof of the
skull, which it nearly reaches in front. Further back it sends
out two great processes, one superior and the other inferior, at
right angles to its own plane.’”’? These two processes are said
to be separated by a deep oval fossa, and comparison with
Polypterus would indicate that this fossa occupies the position
‘of the interspace of cartilage that lies between the sphenotic and
parietopterotic bones above, and the dorsal edge of the as-
cending process of the parasphenoid below, this cartilage forming
part of the lateral wall of the trigemino-facialis chamber. ‘The

‘

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 4

380 EDWARD PHELPS ALLIS, JR.

fossa therefore probably represents that chamber, the superior
process of the prodtic of Macropoma then being represented in
the sphenotic and pterotic of Polypterus. The inferior process
of Macropoma gives off certain smaller processes, one of which,
called the process ‘h,’ evidently corresponds to that posterior
portion of the ascending process of the parasphenoid of Polyp-
terus that passes dorsal to the foramen faciale (posterior opening
of the trigemino-facialis chamber). Another process, called the
process ‘i,’ corresponds to the lateral point of the base of the
ascending process of the parasphenoid of Polypterus. The deep
fossa said to lie posterior to the process ‘h’ would seem to be the
large posterior opening of the canal that traverses the ascending
process of the parasphenoid of Polypterus and which I have
lately referred to as the posterior opening of the canalis para-
basalis (Allis, ’18).

The lateral surface of this portion of the neurocranium of
Macropoma thus evidently quite closely resembled that of
Polypterus, and this would naturally suggest that there should
also be some resemblance in the bones related to its ventral
surface, and if Huxley’s figure (l.c., fig. 3, pl. 8) of that surface
of the skull of Macropoma be compared with my figure (’00,
text fig. 1), giving a similar view of the skull of Polypterus, there
is seen to be a marked similarity. The anterior portions of the
parasphenoids of the two fishes are strikingly similar, and on
either side of that bone there is, in Macropoma, a single tuber-
culated bone, and in Polypterus two such bones, both related
to the palatoquadrate. Along the lateral edge of the single
bone of Macropoma there is a dentigerous bone that Huxley
considered to be the maxillary, the posterior portion of the
maxillary of Polypterus having similar relation to the lateral one
of the two bones of that fish. Anterior to the lateral one of
these two bones of Polypterus, and overlapped ventrally by the
anterior end of that bone, there is a palatal process of the max-
illary, and in Macropoma, in strictly similar position, there is
a bone that Huxley calls the palatine. No teeth were found on
this bone in the specimen of Macropoma used by Huxley for the
figure given, but teeth apparently belonging to it were found by

MAXILLARY AND VOMER BONES OF POLYPTERUS 381

him in other specimens; and Smith Woodward (’91) says, in his
list of the characteristic features of the Coelacanthidae, that this
bone, also called by him the palatine, is furnished with more or
less formidable teeth. These teeth were quite certainly arranged
in line along the lateral edge of the bone, and hence formed an
anterior continuation of the teeth on the maxillary of Huxley’s
descriptions, these two bones together thus corresponding ex-
actly to the tooth-bearing component of the maxillary of Polyp-
terus, but here not yet fused with each other and with the
overlapping suborbital bones. On the anterior end of the snout
there is a triangular bone, beset with small cylindrical teeth,
which Huxley considered to be either the fused premaxillaries or
the vomer.

Macropoma thus probably had a maxillary bone similar to
the tooth-bearing component of the maxillary of Polypterus,
but the single bone of Polypterus was represented by two sepa-
rate bones. The anterior one of these two bones of Macropoma
might be part of an holostean maxillary, but the other one
certainly could not be part of such a bone. Fritsch (’78), in a
specimen of Macropoma speciosum shows these two bones of
Huxley’s descriptions fused to form a single bone, and imme-
diately in front of, and in line with it, he shows a well-developed
premaxillary. .

Posterior to the maxillary, there is a bone that Huxley calls
the postmaxillary. It is said by him to be an elongated tri-
angular bone which fills up the interval between the suborbitals
opercula, and mandible, and covers the quadrate articulation.
Its anterior corner abuts against the hind end of the maxillary,
and it is in contact dorsally with the suborbitals. Reis (’88,
pp. 74 to 79) says that the upper one of its three corners lies
upon the outer surface of the pterygoid, that it projects laterally
(seitlich) into the buccal cavity, and that its oral surface is covered
with tooth-like structures (Zahnstreifen oder Zahnkérnchen). It
is said to lie internal to the quadratojugal and jugal, and in a
figure (l.c., fig. 11, pl. 5) giving a diagrammatic transverse
section through this part of the head, it is shown extending up-
ward internal to the suborbital bones, along the lateral surfaces

382 EDWARD PHELPS ALLIS, JR.

of Huxley’s maxillary and the pterygoid. Its ventral edge is
in articular contact with a bone that Reis calls the postsplenial,
and it is said that the arrangement of the tubercules on the
outer surfaces of the dentary and articular show that the fish
had thick lips (eine starke Lippenbildung), and that the angle
of contact of the postmaxillary and postsplenial lay immediately
posterior to the angle of the gape of those lips. Reis accordingly
compares these two bones of Macropoma with the posterior
upper and mandibular labial cartilage of the Selachii, and with
the single labial of Polypterus. Zittel (87-90) accepts the
homologies thus proposed for these bones, and my work strongly
favors them in so far as they refer to the labial cartilages of the
Selachii. Reis further considers these bones to be rudiments of
a preoral visceral arch; but if they be such structures, one re-
lated to the upper and the other to the lower Jaw, the mouth
could not represent the coalesced mandibular visceral clefts of
opposite sides of the head, for those clefts would le posterior to
the labial cartilages, between them and the mandibular cartilages.

The conditions in Macropoma were thus strikingly similar to
those in the recent Polypterus, the only noteworthy differences
being that the labial cartilage of Polypterus is replaced by bone
in Macropoma and that there is a mandibular labial bone as well
as a maxillary one. The retention of the maxillary one of these
two bones, the so-called postmaxillary, and its fusion with the
hind end of the maxillary would apparently give rise to the
conditions found in the recent Apodes, while the development
eof a labial fold, and the persistence of a cartilaginous labial,
would give rise to the conditions in Polypterus, and if this be
the origin of the bones in these fishes; then the maxillary
bone of the Apodes must be the homologue of that bone of
Polypterus.

AMNIOTA

In my work on Polypterus (Allis, 00), I concluded, as already
several times stated, that the maxillary teeth of that fish were
probably the homologues of the maxillary teeth of mammals, —
the teeth of the latter animals then not being the homologues of
the maxillary teeth of the Holostei and most of the Teleostei.

MAXILLARY AND VOMER BONES OF POLYPTERUS 383

To control this statement I have now endeavored to trace the
labial folds and furrows in the Amniota, and for this purpose I
have used exclusively Fuchs’s (’07, ’08) very careful descriptions
of the development of the primary and secondary palates in
these animals.

In all of the Amniota the crest of the secondary upper lip passes
either between the nasal apertures of either side or across the
oral nasal aperture, the fold thus either simply traversing the
nasal bridge, without overlapping its oral edge, or slightly
overlapping that edge and so adding somewhat to the width of
the bridge (Allis, 718). The lateral portions of the secondary
upper lip are formed by the so-called maxillary processes of
embryos, and Fuchs says (’08, p. 221) that those processes take
part in the formation of the primary palate in mammals, but
not in the Sauropsida. In early embryos of Emys he, in his
figures, shows the process running forward along the lateral
edge of the oral (internal) nasal aperture and then forward along
the external surface of the lateral nasal process nearly to the
oral end of the aboral (external) nasal aperture. The ridge of
the process is bounded on either side by a furrow, and another
furrow overlies the line of fusion of the nasal processes above
the nasal groove. This latter furrow is joined, somewhat
anterior to the oral nasal aperture, by the furrow that bounds
mesially the maxillary process, and Fuchs considers this fusion
of ‘these two furrows to mark the oral end of the nasal process
and the beginning of the fusion of the maxillary process with
the lateral edge of the ventral end of the nasal septum (Vomer-
polster). Posterior to this point, the nasal bridge is said to no
longer form a primary palate, but a secondary one, and the oral
nasal apertures, lying posterior to this secondary palate, are
choanae reliquae, and represent remnants only of the pri-
mary choanae. The maxillary process bounds laterally the
choana of either side, and the bottom of the nasal groove opens
directly into the angle formed by the inner surface of the maxil-
lary process and the roof of the mouth. The furrow that forms
the lateral boundary of the maxillary process has, in this trans-
verse plane, become the bounding line between the maxillary

384 EDWARD PHELPS ALLIS, JR.

process (secondary upper lip) and the side of the head, and on
the external surface of this process (secondary lip) there is, in
slightly older embryos (l.c., pl. 23), a shallow longitudinal de-
pression. In still older embryos of Emys and also in those of
Chelone (l.c., pl. 24), this depression becomes a furrow which
lies dorsal to the ventrolateral edge of the maxillary bone, and
hence in the position of the supralabial furrow of fishes.

The oral nasal aperture has thus been enclosed in the buccal
cavity without having to cut through the premaxillo-maxillary
dental arcade, as I formerly suggested must be the case (Allis,
00), and it even lay, primarily, aboral to the primary lip.
Later, it cut through that lip and opened into the primary
superior alveololabial sulcus, a median remnant of the hp ap-
parently being left between the apertures of opposite sides and
being represented in the choanate papilla, or papilla palatina, of
Fuchs’s descriptions, this papilla being said by him to always
lie, excepting in certain of the Sauria, immediately posterior to
the primary palate, between the anterior ends of the primary
choanae (oral nasal apertures). Having cut through this pri-
mary upper lip and reached the primary alveololabial sulcus,
the nasal groove encountered, in its posterior progression, the
primary alveolar ridge, and was apparently there at first de-
flected laterally along the external surface of that ridge, for it is
so shown by Fuchs (’08, figs. la and 1b, pl. 61) in Hatteria, where
the conditions are considered by him to be the most primitive
found in the Amniota. The nasal groove, thus prolonged, evi-
dently interfered somewhat with the free development of the
bones in the primary dental arcade, and the groove eventually
acquired a passage between the vomerine and palatine portions
of that arcade and then had a sagittal course along the lateral
edge of the body of the vomer. The continued coalescence of
the lips of the nasal groove and the invasion of the bridge so
formed by processes of the vomer, palatine, and maxillary, then
gave rise to the conditions described by Fuchs in Chelone. In
the sections of Hatteria that are given, the supralabial furrow is
not shown, but it is shown, well developed, in the sections of
Lacerta, and the lachrymal duct leads from this furrow to the
primary choana.

MAXILLARY AND VOMER BONES OF POLYPTERUS 385

The primary upper lip is apparently represented in the so-
called median lateral-fold of Fuchs’s descriptions. This fold is
not described by him in embryos of Chelone, but it is in em-
bryos of certain others of the Sauria, there lying lateral to
the lateral lip of the nasal groove, the so-called choanate fold,
between that fold and the maxillary teeth, and where there are
palatopterygoid teeth, as in Tropidonotus, they apparently de-
velop immediately internal to this fold. This fold thus certainly
represents the primary upper lip,and is the homologue of the
maxillary breathing-valve of fishes. In mammals it is said by
Fuchs to become the secondary palate, that palate thus being
represented in the maxillary breathing-valve of fishes, as |
suggested in my earlier work on Polypterus (Allis, ’00).

SUMMARY

There are thus, in recent teleostoman fishes, several distinctly
different types of secondary dental arcade in the upper jaw.

In one of these types the maxillary and premaxillary each has
two components, one dental and the other dermal, the latter
component being formed by bones developed in relation to the
laterosensory canals. The teeth on each of these bones ap-
parently issue approximately along the line of the primary upper
lip, the maxillary teeth lying internal to the labial fold and
hence definitely internal to both upper labial cartilages. This
dental arcade is found in Polypterus, and also in Macropoma,
one of the fossil Crossopterygii, and it apparently corresponds
to the dental arcade of the Mammalia.

A second type is found in Conger, the premaxillary here being
represented by teeth that have fused with the dermal ethmoid,
and the maxillary corresponding either to the maxillary and post-
maxillary bones of the fossil Macropoma, or, more probably,
being a bone developed in relation to teeth that correspond to the
inner row of teeth on the so-called maxillary chain of canal-
bones of Lepidosteus, those teeth having apparently been pri-
marily developed in relation to tissues that represent the posterior
upper labial cartilage.

386 EDWARD PHELPS ALLIS, JR.

A third type is found in the Holostei, and probably also in
certain of the physostomous Teleostei, a premaxillary being
present, and the maxillary having dental and dermal com-
ponents. The dental component of the maxillary is formed by
the fusion of the basal plates of teeth developed in relation to
tissues that represent the anterior upper labial, the dermal
component being a bone similar in origin to the dermal plates
on the cheek of the fish.

A fourth type is found in Polyodon. Here there is no pre-
maxillary, and the maxillary resembles that of the Holostei, but,
because of the absence of a premaxillary, it has been prolonged
anteriorly until it meets its fellow of the opposite side in the
median line. | .

A fifth type is found in most of the Teleostei; the dental com-
ponent of the holostean maxillary here becoming part of the.
premaxillary, and forming a posterior prolongation of that bone
which lies ventral and parallel to the maxillary and excludes it
more or less from the upper margin of the mouth, while the
dermal component of the holostean maxillary becomes the
teleostean maxillary.

In Amia and most of the Teleostei, an anterior portion of the
primary upper lip persists as the maxillary breathing-valve.
In Polypterus, Polyodon and Conger this breathing-valve is
not found, the fold of the secondary upper lip apparently not
here passing sufficiently anterior to the primary one to permit
of the retention of the latter as an independent structure. In the
Mammalia the primary upper lip (maxillary breathing-valve)
becomes the secondary palate.

The so-called vomers of Polypterus are mesial dermopalatines,
lie morphologically in the floor of the suprapalatine recess, and
do not contain the homologue of: the basal plate of the unpaired
vomer of the Teleostei, which plate lies, morphologically, in the
roof of that recess. The teeth on the two bones may, however,
be homologous.

Palais de Carnolés, Menton
June 29, 1918

MAXILLARY AND VOMER BONES OF POLYPTERUS 387

LITERATURE CITED

Auuts, E. P., Jr. 1889 The anatomy and development of the lateral line system
in Amia calva. Jour. Morph., vol. 2, no. 3, Boston.

1897 The cranial muscles,and cranial and first spinal nerves in Amia
ealva. Jour. Morph., vol. 12, no. 3, Boston.

1900 The premaxillary and maxillary bones, and the maxillary and
mandibular breathing valves of Polypterus bichir. Anat. Anz., Bd.
18, no. 11/12, Jena.

1903 The lateral sensory system in the Muraenidae. Intern. Mo-
natsschr. f. Anat. u. Phys., Bd. 20, Hft. 4/6, Leipzig.

1904 The latero-sensory canals and related bones in fishes. Intern.
Monatsschr. f. Anat. u. Phys., Bd. 21, Hft. 9/12, Leipzig.

1909 The carnial anatomy of the mail-checked fishes. Zoologica,
Hft. 57 (Bd. 22), Stuttgart.

1914 Certain homologies of the palato-quadrate of Selachians. Anat.
Anz., Bd. 45, Jena.

1917 The lips and the nasal apertures in the gnathostome fishes, and
their homologues in the higher vertebrates. Proc. Nat. Acad. Sci.,
vol. 3, no. 2, Baltimore.

1919 The lips and the nasal apertures in the gnathostome fishes.
Jour. Morph., vol. 32, no. 1.

Bou.encer, G. A. 1904 Fishes (Systematic account of Teleostei). The Cam-
bridge Nat. History, vol. 7, Fishes, Ascidians, etc., London.

Bripce, T. W. 1879 The osteology of Polyodon folium. Phil. Trans. Royal
Soc., vol. 169, London.
1904 Fishes (Exclusive of the Systematic account of Teleostei). The
Cambridge Nat. History, vol. 7, Fishes, Ascidians, etc., London.

Fritscu, A. 1878 Die Reptilien und Fische der béhmischen Kreideformation.
Praz. '

Fucus, H. 1907 Untersuchungen iiber Ontogenie und Phylogenie der Gaumen-

- pbildungen bei den Wirbeltieren. Erste Mittheilung: Uber den Gau-

men der Schildkréten und seine Entwickelungsgeschichte. Zeit.
Morph. Anthrop., Bd. 10, Hft. 3, Stuttgart.
1908 Ibid. Zweite Mittheilung: Uber das Munddach der Rhyncho-
cephalen, Saurier, Schlangen, Krokodile und Sauger und den Zusam-
menhang zwischen Mund- und Nasenhohle bei diesen Tieren. Zeit.
Morph. Anthrop., Bd. 11, Hft. 2, Stuttgart.

Gaupp, E. 1905 Die Entwickelung des Kopfskelets. Handbuch d. vergl. u.
experim. Entwickelungslehre d. Wirbeltiere von Oskar Hertwig, Bd. 3,
Teil 2, Jena.

GEGENBAUR, C. 1898 Vergleichende Anatomie der Wirbelthiere mit Beriick-
sichtigung der Wirbellosen, Bd. 1, Leipzig.

GreIL, A. 1913 Entwickelungsgeschichte des Kopfes und des Blutgefasssys-
temes von Ceratodus forsteri. Zweiter Teil: Die epigenetischen
Erwerbungen wihrend der Stadien 39-48. Jenaische Denkschriften,
Bd. 4, Jena.

388 EDWARD PHELPS ALLIS, JR.

Huxtey, T. H. 1861 Glyptolaemus Kinnairdi. Memoirs of the Geological
Survey of the United Kingdom. Decade X., pp. 41/46. London.
1866 British fossils. Illustrations of the structure of the cross-
opterygian ganoids. Mem. Geol. Survey of the United Kingdom.
Decade XII. London.

Miuer, J., anpD TrRoscHeL, F.H. 1845 Horae Ichthyologicae. Beschreibung
und Abbildung neuer Fische. Die Familie der Characinen. Hft.
1/2, Berlin.

Parkur, W. K. 1882 On the development of the skull in Lepidosteus osseus.
Phil. Trans. Royal Soc., vol. 173, London.

PouuarD, H. B. 1895 The oral cirri of Siluroids and the origin of the head in
vertebrates. Zool. Jahrb. Abth. f. Anat., Bd. 8, Hft. 3, Jena.

Reis, O. M. 1888 Die Coelacanthinen, mit besonderer Beriicksichtigung der
im Weissen Jura Bayerns vorkommenden Gattungen. Palaeonto-
graphica, Bd. 35, Lief. 1, Stuttgart.

SAGEMEHL, M. 1883 Beitrige zur vergleichenden Anatomie der Fische. I.
Das Cranium von Amia calva (L). Morph. Jahrb., Bd. 9, Hft. 2,
Leipzig.

1884 Beitrige zur vergleichenden Anatomie der Fische. III. Das
Cranium der Characiniden nebst allgemeinen Bemerkungen iiber die.
mit einem Weber’schen Apparat versehenen Physostomenfamilien.
Morph. Jahrb., Bd. 10, Hft. 1, Leipzig.
1891 Beitrige zur vergleichenden Anatomie der Fische. IV. Das
Cranium der Cyprinoiden. Morph. Jahrb., Bd. 17, Hft. 4, Leipzig.

Traquair, R. H. 1870 On the cranial osteology of Polypterus. Journ. Anat.
and Phys., Ser. 2, vol. 5 (No. 7), London.

1875 Structure and affinities of Tristichopterus alatus. Trans. Royal

Soc. Edin., vol. 27, Pt. 3.

1877 The Ganoid fishes of the British Carboniferous formations.
’ Pt. 1. Palaeoniscidae. Palaeontographical Society, London.

1879 On the structure and affinities of the Platysomidae. Trans.

Royal Soc. Edin., vol. 29, Pt. 1.

Wise, J. W. van 1882 Uber das Visceralskelet und die Nerven des Kopfes der
Ganoiden und von Ceratodus. Niederl. Archiv. f. Zoologie, Bd. 5,
H. 3, Leiden.

Woopwarp, A.S. 1891-1901 Catalogue of the fossil fishes in the British Museum
(Natural History). Pts. 2, 3,4. London.

ZitTeEL, K. A. 1887-1890 Handbuch der Palaeontologie, Vertebrata (Pisces,
Amphibia, Reptilia, Aves), Bd. 3, Miinchen und Leipzig.

PLATES

ABBREVIATIONS

al, anterior upper labial

d, dentary

div, dorsolateral diverticulum of the
buccal cavity

ect, ectopterygoid

ent, entopterygoid

toc, infraorbital laterosensory
and the enclosing bone

Ic, labial cartilage

If, labial fold

lg, ligament from maxillary bone to
coronoid process of mandible

mdf, mandibular labial flap

mdff, mandibular labial-flap furrow

mdl, mandibular labial cartilage

mdpc, mandibular preangular crease

mz, maxillary bone, or its teeth

maf, maxillary labial flap

maff, maxillary labial-flap furrow

canal,

maxpc, maxillary preangular crease

ne, nasal cavity

nt, nasal tube

pas, superior or inferior primary al-
veololabial sulcus

pl, posterior upper labial cartilage

pmx, premaxillary bone, or its teeth

ps, parasphenoid

pul, primary upper lip, or maxillary
breathing-valve

sas, superior or inferior secondary al-
veolo-labial sulcus

sbf, sublabial furrow

slf, supralabial furrow

smf, submaxillary furrow

sp, splenial

spr, suprapalatine recess

vo, vomer

389

PLATE 1
EXPLANATION OF FIGURES

1 Ventral view of the head of Acanthias blainvillii, showing the labial folds
and furrows. X 2/5.

2 The same, dissected on one side to show the labial cartilages. X 2/5.

3 Lateral view of the head of a small adult Polypterus, the mouth slightly
open, ~X<.-2.

4 The same, the mouth forced widely open. X 2.

390

LLARY AND VOMER BONES OF POLYPTERUS PLATE J

EDWARD PHELPS ALLIS, JR.

391

PLATES 2 ann 3
5 to 14 Transverse sections of the head of a 75-mm. specimen of Polypterus

senegalus, showing the lips and labial folds and furrows.
15 to 18 Similar sections of the head of a 43-mm. specimen of Amia calva.

392

MAXILLARY AND VOMER BONES OF POLYPTERUS PLATE 2
EDWARD PHELPS ALLIS, JR.

393

MAXILLARY AND VOMER BONES OF POLYPTERUS PLATE 3
EDWARD PHELPS ALLIS, JR.

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JUNE 2

STUDIES ON THE MAMMARY GLAND

IV. THE HISTOLOGY OF THE MAMMARY GLAND IN MALE AND FEMALE
ALBINO RATS FROM BIRTH TO TEN WEEKS OF AGE

J. A. MYERS
Institute of Anatomy, University of Minnesota, Minneapolis

NINETEEN FIGURES (ELEVEN TEXT FIGURES, TWO PLATES)

In earlier works (Myers, ’16 and ’17b) the growth and dis-
tribution of the milk-ducts in male and female albino rats from
birth to ten weeks of age were studied. Owing to the fact that
the finer structure of the gland during these ages has not been
carefully studied and that a knowledge of such structure is
essential before proceeding with the changes in the gland during
pregnancy, the present work was undertaken. An abstract of
the results has already been published (Myers, ’18).

MATERIAL AND TECHNIQUE

In this work. approximately forty apparently healthy rats of
about normal weight were used. The age ranged from birth to
ten weeks. These individuals were fairly evenly distributed
among the described stages. For general histological work the
glands were fixed in Zenker’s fluid, paraffin sections cut 7 or 8 u
and stained with iron-hematoxylin or Mallory’s anilin-blue con-
nective-tissue stain. Other glands were stained with Dominici’s
combination stain. For the study of elastic fibers, glands were
fixed in 10 per cent formalin and stained in Weigert’s resorsin-
fuchsin stain or Unna’s orcein stain. Mayer’s mucicarmine was
also applied to some glands with the hope of determining more
accurately the nature of the contents of the intraepidermal
portion of the primary duct.

395

396 J. A. MYERS

The method described by Lane-Claypon and Starling (’06)
was used in making cleared preparations.

Born’s method was used in making the wax reconstructions.

For the study of the secretion in the new-born, animals were
killed by bleeding, and fresh glands were cut with the freezing
microtome. Others were fixed in 10 per cent formalin for an
hour or so, after which frozen sections were cut. Frozen sec-
tions were stained in 1 per cent osmic acid, plain scarlet red
(prepared by adding an excess of the stain to boiling 70 or 80
per cent alcohol) or Herxheimer’s alkaline scarlet red. Other
glands were fixed in Flemming’s fluid, paraffin sections cut at
5 » and stained with safranin. Still others were fixed, cut, and
stained according to the method described by Bell (14).

OBSERVATIONS
New-born

Microscopic sections show that the epidermis over each nipple
area of the new-born female albino rat is nearly two times as
thick as that in the skin adjacent to this area (fig. 1). Such a
thickening renders the skin more opaque, thus preventing the
blood in the deeper layers from giving a red appearance to the
surface. Therefore, the nipple area appears much lighter than
the immediately surrounding skin.

On the surface in the central portion of the light area is a very
slight elevation which represents the developing nipple, while
partially surrounding the nipple is a shallow depression (fig. 1)
which in most cases is partly filled with cornified cells. Periph-
eral to the depression the epidermis becomes continuous as a
much thinner layer surrounding the mammary gland area.

Deep to the depression is the epithelial hood described in
earlier papers (Myers, 716, ’17 a, and ’17b). In the new-born
the free projection of the hood extends slightly further into the
corium than in twenty-day fetuses. The outer and inner sur-
faces of this projection are smooth and are covered with the
stratum germinativum. The stratum corneum dips slightly
between the two surfaces of the hood.

STUDIES ON THE MAMMARY GLAND 397

ED. S. 1.

Fig.1. Drawn from a section through the left first inguinal nipple of a female
albino rat at birth. ep., epidermis adjacent to mammary gland area; ep. in.,
epithelial ingrowth or hood; /., lumen; n., nipple; p.d., primary duct. X 100.

Fig. 2. Drawn from a section through the left second thoracic nipple of a

female albino rat of one.day. ep., epidermis adjacent to mammary gland; ep. in.,

epithelial ingrowth; h., hair follicle; n., nipple; p., developing processes on epi-
thelial hood; p.d., primary duct. X 100.

Fig. 3. Drawn from a section through the right third thoracic nipple of a

female albino rat of five days. ep., epidermis adjacent to mammary gland; ep.

in., epithelial ingrowth; h., hair follicle; n., nipple; p., developing processes on
epithelial hood; s., sulcus partially surrounding nipple. X 100.

398 J. A. MYERS

Traced toward the nipple from the free end of the epithelial
projection the stratum germinativum extends toward the sur-
face of the skin thus lining the inner surface of the epithelial
hood. Traced away from the nipple from the free end of the
epithelial projection it forms the outer surface of the projection
and becomes continuous as the stratum germinativum of the
surrounding integument. Mitotic figures are numerous in the
stratum germinativum covering the epithelial projections.

The contents of the hood are densely arranged connective-
tissue cells, small blood-vessels, small nerves, white blood cor-
puscles, and the developing primary milk-duct. The connective-
tissue cells are densely placed as in other parts of the corium.
Near the walls of the epithelial hood these cells are arranged
somewhat parallel to the stratum germinativum. In other parts”
of the hood they are arranged in a more irregular manner except
around the developing primary duct, where they are concentri-
cally placed, thus forming the sheath of the duct.

In microscopic sections near the highest part of the developing
nipple a very shallow pit is being formed by the process of des-
quamation. This pit represents the future opening of the pri-
mary milk-duct—the milk-pore. As this pit deepens in the later
stages it becomes a part of the lumen of the primary milk-duct.
This part of the lumen becomes greater in extent than the
thickness of the epidermis covering the nipple, for the sake of
convenience in description, however, it shall hereafter in this
paper be known as the intra-epidermal part of the primary duct.
From this very shallow pit a solid cord of cells extends through
the thickened epidermis. This is the anlage of the intra-epider-
mal portion of the primary milk-duct. At birth there is no trace
of a lumen in this portion of the duct. That part of the duct
within the epithelial hood takes a course perpendicular to the
surface of the integument. Near the surface the duct is repre-
sented by a solid cord of cells, but as it is traced deeper an oc-
casional lacuna representing the first trace of a lumen is encoun-
tered. In some glands several lacunae have flown together, thus
representing a larger lumen with an irregular outline (fig. 1).
The cells in the walls of this part of the duct present no definite

STUDIES ON THE MAMMARY GLAND 399

arrangement into layers (fig. 7). As the primary duct leaves
the epithelial hood and passes into the tela subcutanea, nearly
to its bifurcation, the lumen presents a very irregular outline,
in some sections being absent entirely. The walls of the duct.
are from three to four cells thick.

The secondary ducts which result from the bifurcation of the
primary duct possess lumina with more regular outlines. The
cell walls appear somewhat thinner and there are indications
of the cells beginning to arrange themselves into layers. From
the secondary ducts to the ends of the terminal ducts there is,
in most cases, a fairly definite layer of low columnar or cuboidal
cells surrounding the lumen. External to this layer is a second
layer of somewhat irregular-shaped cells which is present in most
places, but may be partly or entirely lacking in some parts of the
terminal ducts. The free ends of the terminal end-buds are
composed of solid masses of epithelial cells.

The milk-ducts are usually spread out so as to occupy a single
plane (fig. 18) except in the case of the second inguinal gland
(Myers, 716), where the limited space for the distribution of ducts
forces them into two or more planes. Immediately surrounding
all segments of the ducts are a few concentrically placed layers
of connective-tissue which represents the future mantle layer.
The irregularly arranged connective-tissue between the adjacent
ducts and the mantle layer may be regarded as representing a part
of the true stroma found in later stages.

As previously shown (Myers, ’17 b), no trace of a mammary
gland appears on the surface of the skin of new-born male rats.
This is due to the fact that the epidermis in most cases presents
no thickening in the region of the mammary gland. The skin,
therefore, is as transparent over this region as in that immedi-
ately surrounding the gland, so the color of the blood in the
corium gives a uniform appearance to both areas.

The end of the primary duct attached to the epidermis in the
male is represented by a solid cord of cells. The free end of this
duct and the secondary ducts possess lumina in slightly more
advanced stages of development than in females of corresponding
age. The remaining parts of the gland correspond closely in
structure and development with the glands of the female.

400 J. A. MYERS
Twelve hours

In female rats twelve hours after birth the epidermis covering
the developing nipple appears slightly thinner than at birth.
There is no noticeable change in the epithelial hood. The at-
tached end of the primary duct is still without a lumen. The
remaining ducts possess quite well-developed lumina and their
walls are generally lined with two layers of epithelial cells, as
described in the preceding stage. However, in some cases the
outermost layer is almost entirely absent from a small part of
the ducts. These parts with a single layer of cells constituting
their epithelial lining are not confined to the terminal ducts only,
but may be found in any of the other ducts except the primary
duct.

There is a considerable extravasation of red blood-cells in the
stroma around the ducts in some specimens. A considerable
number of such cells are present in the walls of the milk-ducts,
while a few appear in the lumina.

In addition to some red blood-cells in the lumina of the ducts,
a few leucocytes are present. A somewhat homogeneous sub-
stance is also present in some segments of the ducts. In no case
have degenerating cells and nuclei been observed! in the lumina
of the milk-ducts. The cells and other contents present exist
in such small quantities that distention appears in no part of
any system of ducts examined. Special fat stains failed to reveal
any trace of fat in the walls or lumina of milk-ducts. In the
fetus it was found (Myers, 717 a) that the lumina of the milk-
ducts are formed by the process of rearrangement of the cells
and not, as was formerly believed, by the degeneration of the
central cells of the ducts. At birth and for some time after no
degenerating cells have been observed in the lumina; moreover,
the free borders of the cells immediately surrounding the lumina
are quite sharply indicated. It therefore is evident that the
substance contained in the lumina of the milk-ducts outside of
normal cells is not a detritus, but is probably a secretion pro-
duced by the cells surrounding the lumen.

STUDIES ON THE MAMMARY GLAND 401

The stroma very distinctly presents the anlages of its two
parts, the mantle layer immediately surrounding the ducts and
the true stroma intervening between the ducts of a system of
ducts. The true stroma is rather compact and contains many
developing fat cells located toward the free ends of the ducts.

Twenty-four hours

_ Twenty-four hours after birth the developing nipple presents
about the same appearance as in the preceding stage. The
epithelial projection of the hood now possesses a few very short
processes, some of which are located on the inner surface and
extend toward the primary duct, while others are on the outer
surface projecting away from the hood (fig. 2). Such processes
are formed by a thickening of the stratum germinativum. In
some cases the cells of this layer have only elongated, while in
others at least two layers of cells enter into the formation of the
processes.

The contents of the lumina are approximately the same in
quality and quantity as in the last-described stage. In the walls
of the ducts just outside the lumen an occasional vacuole is seen.

In the inguinal region laterodorsal to the inguinal and abdomi-
nal glands a long mass of fat is beginning to appear in the tela
subcutanea. This mass extends from the lateral surface of the
abdomen just anterior to the ilium to a position posterolateral
to the vagina. In the remainder of the present paper this mass
of fat will be designated as the inguinal fat pad. Adipose tissue
is also developing rapidly laterodorsal to the second and third
thoracic glands and cephalic to the first thoracic gland, but it
does not present such a definite mass as in the inguinal region.

Four and five days

In female rats of four and five days the nipple is somewhat
more elevated above the surface of the surrounding skin than at
the end of the first day of life (figs. 3 and 12). The nipple is now
represented by a somewhat oblong elevation (fig. 12). Starting
at a point slightly posterior to the first thoracic nipple and passing

402 J. A. MYERS

cephalad, we find a very gradual elevation until the highest part
of the nipple is reached. If we start a short distance cephalic
to the nipple, we enter a shallow groove or sulcus before reaching
the base of the nipple. From the bottom of this groove we find
a steep approach to the summit of the nipple elevation. The
groove appears cephalic to and passes slightly to the medial and
lateral sides of the nipple. In the case of the second and third
thoracic glands, the abdominal gland and the first inguinal gland,
the groove is on the medial side and extends slightly to the
cephalic and caudal parts of the base of the nipple (fig. 5). It
is absent on the lateral side of the nipple. In the second inguinal
nipple the relations of the groove to the nipple are just opposite
to those in the first thoracic. Near the highest part of the nipple
the lumen of the intra-epidermal portion of the primary duct
appears somewhat deeper than when it was observed in an
earlier stage.

The projection of the epithelial hood extends more deeply
into the corium than in the earlier stages (fig. 3). The pro-
cesses which appeared on the epithelial projection in rats of one
day are now longer and have the appearance of rather sharp
‘spines. A few of them possess secondary processes.

The lumen of the primary duct now extends nearer to the sur-
face, yet it is not in continuity with the lumen of the intra-
epidermal part of the duct. No degenerating cells have been
observed in the formation of the lumen of the primary duct
proper. This lumen apparently is formed by a process of re-
arrangement of cells as was shown to be the case (Myers, 717 a)
in the other milk-ducts of the fetus. From the secondary ducts
through the terminal ducts the amount of secretion present in
lumina has increased somewhat since the end of the first day.
While it is present in considerable quantities, yet the ducts are by
no means engorged. In sections stained with Weigert’s iron-
haematoxylin the secretion fails to take the stain, while in those
stained with Dominici’s mixture it takes a light pink color. In
fresh frozen sections very rarely a fat droplet was observed in the
walls of the milk-ducts. The infrequency of such droplets, their
relation to the cells, and their absence in sections fixed, sectioned

STUDIES ON THE MAMMARY GLAND 403

in paraffin, and stained in other fat stains, proves fairly conclu-
sively that they were carried by the razor from the fat in the
connective-tissue stroma to the epithelial walls of the ducts.
Leucocytes inside the lumina are very rare at this stage. A few
normal cells with nuclei resembling those of the epithelial wall are
‘present in the lumina. In a few instances normal nuclei with
apparently no accompanying cytoplasm were observed in the
lumina. Such elements are slightly more numerous toward the
free ends of a system of ducts. In only one or two instances have
the cells within the lumina shown any signs of degeneration.

Dominici’s combination stain reveals some leucocytes in the
stroma outside the ducts and in the epithelial hood, but no infil-
tration of such cells in the stroma around the ducts has been
observed. There apparently is no extravasation of red blood-
cells as was observed in rats of one day.

End of first week.

At the end of the first week the nipple is more elevated, the
lumen of the intra-epidermal part of the primary duct and the
projection of the epithelial hood extends deeper into the corium.
The lumen of this part of the primary duct is now nearly filled
with debris in which appear some desquamated epithelial cells.
The processes on the epithelial projection of the hood are very
conspicuous.

The milk-ducts in most places possess two layers of epithelium.
The inner layer is formed of very compact epithelial cells of the
low columnar or cuboid type. The cells of the outer layer are
less compact and in fact may be absent in some places. They are
irregular in size and shape. The amount and quality of the
secretion in the lumina is approximately the same as at four and
five days.

At this stage the primary duct of the male rat is attached to the
epidermis, there being no nipple at any’stage. Its attached end
does not possess a lumen. In some eases there is a very slight
excavation on the surface of the skin immediately over the at-
tached end of the primary duct which resembles the developing

404 J. A. MYERS

milk-pore. As the primary duct is traced toward the tela sub-
cutanea a distinct lumen appears and the wall is seen to be com-
posed of two layers of cells which are similar in shape and arrange-
ment to those of the walls of ducts of corresponding thickness in
the female. The remaining ducts in the male resemble in number,
distribution and structure the corresponding ducts in the female.
The secretion in the lumina appears the same in every respect as
in the female except that it appears present in slightly larger
quantities in the male.

In both sexes the free ends of the milk-ducts of the abdominal
and inguinal glands are imbedded in the medial margin of the
inguinal fat pad, but in most cases they have not penetrated this
pad very deeply. Besides the ducts growing toward the inguinal
fat pad, the pad itself has increased in size.

Two weeks

It has been pointed out by Myers (’16) that the nipple (in the
female rat) takes on a very rapid development during the second
week. The present study shows that the greater part of that
development occurs in the latter part of the second week. The
nipple becomes so large that it is very conspicuous to the naked
eye in living animals and has been used to determine the sex at
this age (Jackson, 712). Microscopic sections and wax recon-
structions (fig. 13) show that the nipple is much more elevated
than in. the preceding stage. It is at this time a somewhat
elongated structure and has not yet taken on the typical form of a
nipple. The shailow groove which was present around a part of
the base of the nipple in earlier stages is now somewhat deeper.
This groove now entirely surrounds the base of the first thoracic
nipple, except on the caudal side where the base of the nipple
becomes continuous with the surface of the surrounding skin
withouta groove. Incaseofthesecond and third thoracic glands,
the abdominal gland, and the first inguinal gland, the groove
extends around the base of the nipple except on the lateral side.
In the second inguinal gland the groove is absent on the cephalic
side. From the bottom of this groove, which lies immediately

STUDIES ON THE MAMMARY GLAND 405

superficial to the attached end of the epithelial hood, may be seen
short strands of epithelial tissue which react to the stains differ-
ently from the surrounding epithelium. Higher magnification
shows that the cells in such strands are undergoing degeneration
and are being cast out, thus deepening the groove.

The epithelial projection of the hood has taken on a marked
development since the fifth day. It still presents processes
which are now more numerous on the inner side of the projection.
Such processes are likewise present on the inner surface of the
epidermis covering the nipple.

As the layer of smooth muscle which lies immediately below the
hair follicles reaches the neighborhood of the nipple, it turns
toward the surface and enters the epithelial hood. Here many of
the fibers take a course parallel with the primary duct and appar-
ently enter into the formation of the sheath of the duct. Others
soon take an oblique direction and apparently attach to the pro-
cesses on the inner surface of the epithelial projection. Some
muscle fibers, after extending into the nipple, take an oblique
course and appear to attach to the processes on the inner aspect
of the surface epidermis of the nipple. The function of the
smooth muscle fibers will be dealt with in studies on the activity
of the mammary gland.

The lumen of the intra-epidermal part of the primary duct contain
debris apparently in sufficient amount to distend the walls of this
part of the duct. Near the milk-pore the wall is reduced to a
rather thin structure. As the bottom is approached, the wall
becomes thicker. It is here that desquamation is taking place at
present. The cord of epithelial cells, which in all of the previous
stages separated the lumen of the intra-epidermal part of the
duct from the lumen of the primary duct proper, is at this age
perforated by a very narrow canal. Thus for the first time a
communication is established between the lumina of the system
of milk-ducts and the exterior.

The lumen of that part of the primary duct within the epithe-
lial hood and nipple now has a fairly regular outline as far asthe
region intervening between it and the luman of the intra-
epidermal part of the duct. Here the walls consist of cells

406 J. A. MYERS

(three or four deep) that have not yet arranged themselves into
definite layers. As the primary duct is traced from the epithelial
hood into the tela subcutanea, where it turns at right angles and
courses parallel with the surface of the skin, the walls of the duct
become thinner. The cells are here arranged in layers, rarely
more than two layers present. The lumen has a well-defined
outline which is occasionally interrupted by the budding off of
short lateral buds. Many of these buds likewise possess lumina.
The remainder of the ducts has the same general appearance in
structure as at the end of one week. The epithelial walls are
composed of two layers of cells (fig. 8).

The contents of the milk-ducts have diminished in quantity
since the earlier-described stages.

Hair follicles are very numerous peripheral to the sulcus sur-
rounding the nipple superficial to the epithelial projection, but
they have in no specimen been observed medial to this groove.

Three, four, five, six, and seven weeks

From the end of the second to the beginning of the eighth week
the growth of the mammary gland, as shown in an earlier work
(Myers, 716), in the female albino rat is apparently only sufficient
to keep pace with the general body growth. Cleared prepara-
tions show that beyond question the ducts of the mammary gland
grow in length and, moreover, send out numerous new collateral
branches during this period. The nipple makes a gradual
increase in height. The groove has approximately the same
extent around the nipple as in rats of two weeks, but it becomes
deeper as age advances. The projection of the epithelial hood
retains its processes (fig. 4). The epithelium between the lumen
of the intra-epidermal part of the primary duct and the lumen of
the primary duct proper at the end of the second week is com-
pletely broken down about the fifth or sixth week. The lumen of
the primary duct proper therefore becomes directly continuous
with that of the intra-epidermal part of the duct and communi-
cates with the exterior through the milk-pore. The lumen of the
intra-epidermal portion of the duct is through the seventh week

Fig. 4. Drawn from a section through the right second inguinal nipple of a
female albino rat of six weeks. ep., epidermis adjacent to mammary gland area;
ep.in., epithelial ingrowth; m.po., milk-pore; n., nipple; s., sulcus surrounding
a part of base of nipple; se., secretion apparently solidified and dislodged from
milk-pore. X 66-3.

Fig. 5. Drawn from a section through the right first inguinal nipple of a fe-
male albino rat of nine weeks. ep., epidermis; ep.in., epithelial ingrowth; h.,
hair follicle; x., nipple; p., process on epithelial ingrowth; p.d., primary duct;
s., sulcus surrounding a part of base of nipple. At the bottom of this sulcus may
be seen indications of degeneration. X 66-3.

Fig. 6. Drawn from a longitudinal section through the primary duct of the
right first inguinal gland of a female albino rat of nine weeks. c., collateral duct;
l., lumen; p., primary duct proper. X 66}.

407

408 J. A. MYERS

still engorged with debris. That part of the debris which occu-
pies the milk-pore and is in contact with the exterior apparently
becomes solid, thus forming a plug. In some specimens this plug
has been removed from the milk-pore during the process of sec-
tioning and lies as a solid body outside of, and independent of the
milk-pore (fig. 4). Owing to the presence of such a plug, the
contents formed in the lumen of the intra-epidermal part of the
primary duct are unable to escape.

The contents of the milk-ducts gradually diminish in quantity
after about the second or third week. However, considerable
debris is present in the lumina at the beginning of the eighth
week.

In all specimens examined the smooth muscle fibers of the
nipple have the same arrangement and distribution as in the
earlier described stages.

The terminal end-buds of the ducts are prominent in the above-
described stages. The lumen of the terminal duct gradually
increases in size until it reaches a point near the middle part of
the end-bud (fig. 11). Here the lumen ends abruptly, and the
remainder of the end-bud is represented by a solid mass of
epidermal cells.

It has been pointed out (Myers, ’17 b) that the milk-ducts of
the male fail to proliferate as rapidly as those of the female after
about the fourth or fifth week. This difference together with the
absence of the nipple and epithelial hood constitute the only
characteristics that serve to distinguish the glands of the male
from those of the female from the second to the beginning of the
eighth week. As far as the structure of the remaining parts of the
gland is concerned, one is unable to distinguish between the sexes.
Figure 16 shows the milk-pore on the surface of the skin of a male
rat of six weeks. The intra-epidermal part of the primary duct
contains a shallow lumen. A solid mass of epithelial cells sepa-
rates this lumen from that of the primary duct proper.

The connective-tissue sheath or mantle layer surrounding the
milk-ducts becomes somewhat thicker as age advances. This
sheath is quite thick around the primary and secondary ducts,
but when traced toward the free end of a system of ducts gradu-

STUDIES ON THE MAMMARY GLAND 409

ally becomes thinner until it forms only an extremely thin layer
around the terminal ducts. Previous to the third week, no elastic
tissue fibers were observed in the mantle layer. At the end of the
third week, however, Weigert’s resorcin-fuchsin stain reveals a
few elastic tissue fibers sprinkled among the other fibers of the
mantle layer of the primary and secondary ducts. During the
fifth, sixth, and seventh weeks more elastic fibers are present in
the sheath of the primary and secondary ducts, and they are
found in the sheath nearer the free end of a system of ducts than
at three weeks. ‘The true stroma is formed of loose connective-
tissue in which various kinds of leucocytes are found. In no
instance has an abnormal number of leucocytes been observed.
The stroma also presents a great deal of adipose tissue. The
connective tissue which occupies the nipple and epithelial hood
is much more compact than in other parts of the gland.

About the third week the free ends of the milk-ducts of the
abdominal and inguinal glands are observed to extend some dis-
tance into the inguinal fat pad where they send off numerous
ramifications. As age advances the fat pad becomes more
voluminous. It is not only thicker and longer, but wider than
in the preceding stages. Hence its medial margin extends to the
level of the abdominal and inguinal nipples. A great deal of
adipose tissue has developed also in the stroma of the thoracic
glands, so that the ramifications of many of the ducts are sur-
rounded by fat cells.

Eight, nine, and ten weeks

It may be recalled that between the eighth and tenth week it
was found in an earlier work (Myers, ’16) that the ducts of the
mammary gland of the female grow and proliferate very rapidly.
Cleared preparations show that a tremendous number of new
ducts bud out from the already existing ducts. These ducts bud
off not only at the ends of the terminal ducts, but also as collateral
ducts from the sides of all other ducts represented, including that
part of the primary duct which lies parallel with the surface of the
skin. In some of the ten weeks’ stages the proliferation of ducts
is not as marked as in those of nine weeks.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 4

410 J. A. MYERS

The nipple has enlarged somewhat and now at ten weeks more
closely approaches the form of the nipple of the adult female rat.
The groove at the base of the nipple has deepened apparently by
the process of desquamation. In some individuals the tissue at
the base of the nipple caudal to the first thoracic, cephalic to the
second inguinal and lateral to the remaining nipples which in
earlier stages connected the base of the nipple with the adjacent
skin without the interruption of a furrow has now broken down
soas toforma groove. This groove therefore extends completely
around the base of the nipple. In fresh preparations the basal
half of the nipple rests in the pocket formed by the groove. Such
changes have not yet occurred, however, in all rats of nine and
ten weeks of age (fig. 17). This furrow lies immediately super-
ficial to the projection of the epithelial hood into which its cavity
slightly extends. From the bottom of the furrow in some speci-
mens may be seen the degenerating strands which were described
in animals of two weeks (fig. 5). This furrow may contain debris
which reacts to stains and appears under high magnification
similar to the contents of the lumen of the intra-epidermal part
of the primary duct.

The intra-epidermal part of the primary duct is still engorged
with debris. Its walls are lined with stratified squamous epi-
thelium of approximately the same thickness as that covering the
surface of the nipple. At the end of the intra-epidermal part of
the duct there is a gradual transition from the stratified squamous
epithelium of this part of the duct to the epithelium of the pri-
mary duct proper. The walls. of this part of the primary duct
are composed of two or three layers of cells. The innermost
layer is composed of low columnar cells, while the remaining
layers are composed of cells of more irregular form. As the pri-
mary duct leaves the nipple and enters the deeper layers of the
skin, its epithelium is arranged in two layers. The inner layer is
composed of low columnar or high cuboidal cells containing
elongated nuclei with their axes perpendicular to the lumen of the
duct. The cells of the outer layer are somewhat irregular in size,
shape, and position (fig. 9). In a longitudinal section of the pri-
mary duct, as it courses parallel with the surface of the skin,

STUDIES ON THE MAMMARY GLAND 411

numerous processes may be seen some of which have lumina
opening into the duct (fig. 6). These processes represent the
collateral ducts. Some possess lumina and are lined with the
same kind of epithelium as the primary duct, others are merely in
the bud stage of formation and consequently have no lumina.
Longitudinal sections of other ducts, as the secondary, tertiary,
etc., likewise show numerous collateral ducts in various stages of
formation. As the collateral ducts bud from the main ducts they
carry with them a thin sheath of the connective tissue surround-
ing the main duct. Mitotic figures are numerous in the epithelial
cells in these stages.

At this stage very little more than the ducts or excretory part
of the gland is developed. However, at the end of some of the
collateral ducts which branch from the secondary ducts one may
observe several outpouchings which somewhat resemble a small
cluster of grapes. Microscopic sections show that such out-
pouchings possess lumina which are surrounded by a layer of
cuboidal cells, external to which may be seen an occasional some-
what flattened nucleus (fig. 10). These outpouchings may con-
tain a slight amount of secretion, but they can hardly be called
alveoli in the true sense of the word. True lobules are not yet
present, although the above-mentioned clusters are doubtless an
early indication of lobulation. The ramifications of ducts have
become so numerous that they no longer occupy a single plane
(fig. 19) as in the earlier stage (fig. 18), but form a rather dense
arborization which in places reaches a thickness of 0.8 mm.

No true milk sinus, such as has been observed by various
authors in several animal forms, has appeared in any of the
described stages of the rat.

Elastic fibers are prominent in the mantle layer of the primary
and secondary ducts. They are present in this layer of all the
larger ducts, but tend to decrease in number as one passes toward
the free end of a system of ducts. No elastic fibers have been
observed in the sheath of the terminal ducts. The connective-
tissue stroma is similar to that described in earlier stages except
that it contains a greater amount of adipose tissue (fig. 19). The
inguinal fat pad is so extensive that a thin layer of it now extends

412 J. A. MYERS

somewhat medial to the abdominal and inguinal nipples. Just
under the nipple and partially occupying the epithelial hood is a
condensed mass of fibrous tissue which becomes continuous with
the less dense fibrous tissue of the nipple. Deep to and sur-
rounding this mass of tissue is the loose stroma of the gland.

During this stage the blood-vessels in the stroma become
enlarged so that in fresh specimens one notices a considerable
increase in the vascularity of the gland as compared with the
preceding stages.

As in the six weeks’ stage, the primary duct in the male does
not communicate with the exterior on account of a short solid
cord of cells existing between its lumen and the very shallow
lumen of the intra-epidermal duct. Moreover, it has been pointed
out Myers (717 b) that the marked growth of the female gland at
the time of puberty does not usually occur in the male. The
terminal enlargements appear similar to those in females of six
and seven weeks. The walls of the ducts are composed of two
layers of cells with the same arrangement as in the female.
However, observations have revealed no indication of lobulation
or developing alveoli in the male.

DISCUSSION AND CONCLUSIONS

In the following discussion, the nipple, the epithelial hood,
milk-ducts, secretion in the new-born, and gland stroma will be
successively considered.

The nipple pocket and epithelial hood

It will be recalled that in female fetuses of twenty days and six
hours the nipple appears as a small elevation lying at the bottom
of the mammary pit (Myers, 717). After the twenty days and
six hours’ stage the nipple grows so that at the time of birth in
the female albino rat it nearly fills the mammary pit and appears
as a somewhat round light area lying at the same level or very
slightly elevated above the surrounding epidermis. At the end of
the first week of postnatal life (fig. 12) the nipple is only a little
more elevated than in the new-born. During the second week

STUDIES ON THE MAMMARY GLAND 413

and especially in the last half of this week the nipple takes on a
rapid development (fig. 13). At this time the nipple has reached
a height of approximately 0.4mm. From the third to the eighth
week the nipple increases in size and gradually becomes some-
what conical in shape (fig. 15). During the eighth and ninth
weeks, the time when the first ovulation normally occurs, the
nipple begins to appear in size and form similar to the nipple of
the adult virgin rat. In some individuals it now closely ap-
proaches the shape of a cone with a height of 1 mm., while in
others it has not taken on its definitive conical form (fig. 17).

It is perhaps due to the fact that the nipple has no function to
perform in the early life of the individual that it remains in a
somewhat rudimentary state of development during this period.
The rapid elevation during the second and third week may be
closely related with the secretion that has already been produced
in the milk-ducts. After the second and third weeks there seems
to be nothing to stimulate a rapid growth of the nipple as it grows
very gradually until near the time of puberty. As the first
opportunity for pregnancy or the first ovulation approaches, the
nipple takes on the shape and size of that of the adult virgin
rat. This development of the nipp’e is apparently closely cor-
related with contemporaneous changes in the ovary as macro-
scopic observations show that Graafian follicles are rapidly
growing in the early part of this stage. Furthermore, the vas-
cular supply to the ovary very materially increases as the
period of puberty is approached.

Rein (’82) found in the mammary glands of the rabbit two
months after birth that the nipple presents a slight elevation in
the form of the head of a pin. This elevation is about 1 mm.
in height with a basal diameter of 1 to 1.5 mm. In rabbits of
six months Schil (’12) found the nipples are conical elevations
1 to 1.5 mm. in height and 4 mm. in diameter at the base. Schil
also states that the nipple is rarely indicated in the new-born
human. Herz (’00), however, found that the nipple may or
may not be elevated in new-born children.

While measurements on the nipples are lacking for many
forms of animals, the present work together with general state-

A414 J. A. MYERS

ments found in the literature are sufficient to substantiate the
conclusion that the nipple remains in a rudimentary state
during the early life of the individual. In the female it takes
on its adult virgin form during the prepubertal and pubertal
stages. In most animal forms a rudimentary nipple is present
in the male at birth; however, it usually remains rudimentary
throughout life. A more complete discussion on this subject
may be found in an earlier paper (Myers, ’17 b).

Epithelial hood

Immediately peripheral to the basal part of the developing
nipple of the new-born the basal layer of the epithelium dips
deeply to form the epithelial hood. Beginning about the second
week of postnatal life, degenerative changes appear in the central
part of the attached end of the epithelial hood. Such changes
apparently have for their purpose the deepening of the groove ~
which at first partially and later completely surrounds the base
of the nipple.

At the time of birth a very shallow furrow may be seen sur-
rounding the base of the nipple (fig. 1).. A few days after birth
this furrow in most cases has nearly disappeared. However,
on the fifth day an apparently new depression (fig. 12) appears
around only a part of the base of the nipple. As age advances,
this depression or groove becomes deeper and extends around
a greater part of the base of each nipple, until finally about the
tenth week in some specimens it completely surrounds the base,
thus forming a definite pocket in which approximately the basal
half of the nipple rests.

Hair follicles appear lateral to the furrow surrounding the
nipple (figs. 2 and 3), but have not been observed within this
furrow. :

Through the ten-weeks stage there is no indication of furrow
or epithelial hood in the male albino rat.

In 1876, Gegenbaur called attention to the fact that in Mus
decumanus and Mus musculus the basal part of the nipples lies
in a pocket-like depression of the integument. Gegenbaur be-

STUDIES ON THE MAMMARY GLAND 415

lieved the depression or pocket to be like the nipple pocket of
marsupials. Rein (’82) studied the gland in a few of its stages
in the white mouse and white rat. He concluded that the pocket
surrounding the nipple in the adult has nothing to do with the
nipple pocket of marsupials, but that it is identical with the
marsupial pouch. Bresslau (’02, ’10, and ’12), in a series of
studies, finally concluded that the sheath or pocket surrounding
the nipple of the adult rodent and insectivorous mammals de-
velops quite independently of the nipple pocket of marsupials
or the marsupial pouch. The groove or sulcus at the base of
the nipple in the albino rat was mentioned in my earlier work
(Myers, 716), but its formation was not discussed at that time.
The shallow groove which appears around the nipple in the new-
born (fig. 1) is undoubtedly a remnant of the prenatal mammary
pit which the growing nipple has not yet completely occupied.
Sections through nipples of twelve hours, one day, two and three
days after birth show that the sulcus around the new-born nipple
has disappeared; in other words, the nipple completely fills the
mammary pit. The very slight depression that surrounds only
a portion of the base of the nipple at five days and continues to
deepen and become more extensive until it forms a pocket con-
taining nearly the basal half of the nipple at the tenth week
apparently develops independently. We must therefore con-
clude that the pocket in which the base of the nipple of the adult
virgin mouse and rat rests is not, as Gegenbaur believed, homol-
ogous with the nipple pocket of marsupials, but, as Bresslau
suggested, it is an entirely independent structure which in the
albino rat at least develops after birth.

The epithelial projection or hood was described in Mus decu-
manus and Mus musculus by Gegenbaur (’76). Later Rein
(82) observed such projections in white mice and white rats
and described them.from sections of fetuses as two solid epi-
dermal thickenings which surround the mammary anlage. He
described them as being somewhat curved with concavities di-
rected toward each other. Rein found similar structures present
in the mole. He believed that they have to do with the for-
mation of the pocket around the nipple. Klaatsch (84) also

416 J. A. MYERS

called attention to such projections. Bresslau (’12) observed
these projections developing in Talpa europaea, white mouse,
Mus musculus, Mus rattus, and Mus decumanus, and concluded
that they develop into the secondary sheath or pocket of the
nipple. In the albino rat wax reconstructions have been used
by me to show that the epithelial projections seen in sections
form a continuous hood around the primary duct. The attached
end of this hood lies immediately peripheral to the base of the
nipple. In view of the fact that the present work shows that
the sulcus is deepened through degeneration and desquamation
of the attached end of this hood, we must conclude that a portion
of the sulcus is formed at the expense of the hood. In other
words, the lateral wall of the attached part of the hood becomes
the lateral wall of the deep part of the sulcus, while the medial
wall of the attached end of the hood becomes the medial wall
of the deep part of the sulcus. At present I am unable to say
whether the sulcus forming the pocket around the nipple con-
tinues to deepen in the same manner during pregnancy and
lactation until the free end of the hood is reached.

Milk-ducts

It will be recalled that the milk-ducts present several branches
in female fetuses of about twenty days (Myers, ’17 a). A con-
siderable growth and branching occurs between this stage and
the time of birth (compare fig. 12, p. 223, vol. 22, Am. Jour. Anat.,
with figs. 3 to 6, inclusive, pp. 361-365, vol. 19, Am. Jour. Anat.).

At the time of birth the lumina of the milk-ducts are not
completely formed. The attached end of the primary duct for
a short distance has no trace of a lumen. However, as one
passes toward the deeper part of the duct one observes isolated
lacunae. A little further along the duct one meets larger spaces
which result from the flowing together of several lacunae (fig. 7).
Finally, a slit-like lumen appears in the still deeper part of the
duct, which extends through the secondary ducts. In the re-
mainder of the ducts a fairly well-developed lumen is present.
As age advances the lumen of the primary duct becomes com-

8

Fig. 7 Drawn from a cross-section through the primary milk-duct of the left
first inguinal gland of a female albino rat at birth. Zenker’s fixation; Weigert’s
iron-hematoxylin stain. Drawn with the aid of a camera lucida. c.t., develop-
ing connective tissue forming sheath around the duct; ep.w., epithelial wall com-
posed of epithelial cells without definite arrangement into layers; /., developing
lumen. X 800.

Fig. 8 Drawn through a cross-section of a terminal duct (about 20u from its
termination) of a female albino rat of eleven days. Zenker’s fixation; Weigert’s
iron-hematoxylin stain. Drawn with the aid of a camera lucida. c.t., connec-
tive tissue; 7./., inner or glandular layer of epithelial cells; 0./., outer or myo-
epithelial layer of cells. X 800.

417

418 J. A. MYERS

pletely formed. Near the highest part of the nipple elevation
in the new-born female appears a shallow pit, the first indication
of the milk-pore and the beginning of the intra-epidermal part
of the primary duct. As age advances, this pit deepens and
enlarges until at five or six weeks it comes in contact with the
lumen of the primary duct proper. Thereafter the system of
ducts is in direct communication with the surface through the
milk-pore.

The walls of the ducts, especially the primary and secondary
ducts, are quite thick at birth. The free half of each terminal
enlargement is composed of a solid mass of epithelial cells. At
birth some of the ducts toward the free end of a system present
a wall composed of two definite layers of cells. After the first
week all of the milk-ducts except the attached end of the pri-
mary duct are surrounded by an inner layer of very compactly
placed low columnar or cuboidal cells and an outer layer of
irregularly shaped cells with no regular arrangement (figs. 8
and 9).

Ancel and Bouin (’09) found in rabbits as many as ten milk-
ducts radiating from the nipple. Each duct in animals before
the stage of puberty, however, does not exceed 2 mm. in length.
The diameter of the gland reaches about 3 or 4 mm. The fact
that there is a single duct leading from the nipple in the albino
rat perhaps accounts for the much greater length and branching
of the ducts in the prepuberty stage. Steinhaus (’93) found in
young guinea-pigs that the glands are composed only of excretory
ducts. Dulcert (’93) arrived at the same conclusion. They
did not observe many mitotic figures. In a heifer of one year
Dulcert (’93) reported only excretory canals and observed no
mitotic figures. O’Donoghue (712) found in the deeper parts
of the glands of Dasyurus a few structures of doubtful signifi-
cance which may or may not represent mitotic figures, but he
found that the actual formative growth of the gland does not
appear to start until after ovulation.

Langer (’51), Raubitschek (’04), and others have reported
true alveoli in the mammary gland of human after birth and
before puberty. Berka (’12), however, was unable to find true

STUDIES ON THE MAMMARY GLAND 419

alveoli present in these stages. Schil (712) states that from birth
until shortly before the puberty period the mammary gland of
the rabbit undergoes very little or no growth, that the slight
elongation of the ducts is due to a stretching produced by the
developing stroma. He states that there are two facts which
characterize this period: first, the gland maintains the develop-
ment already reached at birth; second, some.of the cells produce
a very slight secretion. These facts, he believes, prove that the
gland does not receive any impulse of growth or secretion. How-
ever, Schil thinks that during this entire period the glands are
under the influence of the ovaries, for after double ovariotomies
in young animals the free parts of the ducts not only fail to pro-
duce a slight secretion, but disappear so that the gland is re-
duced to some short ducts centering around the nipple. On
the other hand, he cites the experiments of Knauer and Foges
which show that after simple hysterectomies without ablation
of the ovaries the gland remains as in normal or control animals.
Schil concludes, therefore, that the presence of the ovary is
necessary to maintain the mammary gland in the stage it has
reached at the time of birth.

My previous work (Myers, ’16) shows beyond doubt that the
ducts do grow and proliferate a great deal after birth and before
the period of puberty is reached. According to the castration
experiments of Steinach (’12) and others, it seems probable that
this gradual growth in the milk-ducts of the albino rat from
birth to a short time before puberty is under the influence of the
ovaries. Further castration experiments are being attempted
to confirm this point. It has been pointed out, however, (Myers,
’17 b) that the milk ducts of the male albino rat undergo a similar
but less extensive growth and proliferation from the time of
birth to puberty. If, as Schil believes, the presence of the ova-
ries is necessary even to maintain the gland of the rabbit in its
infantile state until puberty, we are at present unable to locate
the factor which stimulates the growth of the glands of the male
during this period.

The rapid growth and proliferation of the milk-ducts at the
time of puberty has been mentioned by a number of investigators,

420 J. A. MYERS

but only a few have made a detailed study of this growth. Lan-
ger (751) called attention to the growth of ducts at this period
in human. Ancel and Bouin (’10 and ’11) showed that during
the period of puberty the mammary gland in the rabbit reaches
a diameter of 2 to 38 cm. Frank and Unger (’11) showed that a
very decided growth occurs in the mammary glands of rabbits
during puberty. Schil (12), in a careful series of observations
on rabbits, found that the glands gradually increase throughout
the first period of heat, after which they undergo slight regressive
changes until the next period of heat. If pregnancy does not
occur, the glands proliferate with each heat. The regression
following heat depends upon the interval between two periods.
Brouha (’05) has shown in Vespertilio murinus, in which there
appears a single ripe follicle each year, that between two preg-
nancy periods the gland may return to the infantile stage. The ©
fact that the glands of some rats at ten weeks of age show less
development than those in others at nine weeks may be due to
slight regressive changes since the first ovulation, as this ovu-
lation may occur earlier than nine or ten weeks (Lantz, 710;
Jackson, 712). However, owing to the fact that careful studies
on the oestrus cycle in the albino rat have not yet been published,
the question as to changes in the mammary gland between and
during successive periods of heat has been omitted in this study.

The present study, my previous work (Myers, ’16), and the
findings of the above-mentioned authors lead to the conclusion
that there is in the female of several animal forms a very marked
development of the mammary gland at the time of puberty.
Since such development does not ordinarily occur in the glands
of the male or in castrated females, it seems probable that these
changes are at least stimulated by simultaneous changes in the
ovaries.

The time and manner of formation of the lumina of the milk-
ducts were previously discussed (Myers, ’17 a). Suffice it to
say here that the method of formation of the lumen already
begun in the fetus is continued in the postnatal stages until
the lumina are completely developed. The lumen of the intra-
epidermal part of the primary duct forms an exception, however,
as it develops by the process of degeneration and desquamation.

STUDIES ON THE MAMMARY GLAND 431

The lining of the milk-ducts in the adult virgin human has
been described by von Ebner as consisting of a single layer of
epithelial cells. Starling and Lane-Claypon (’06) also reported
a single layer of epithelium surrounding the ducts in virgin rab-
bits. Rein (’82) divided the duct in rabbits of two months into

Fig. 9 Drawn from a cross-section of a secondary milk-duct of a female al-
bino rat of nine weeks. Zenker’s fixation; Weigert’s iron-hematoxylin stain.
Drawn with the aid of a camera lucida. c.t., connective tissue; 7./., inner or
glandular layer of epithelium; o./., outer or myo-epithelial layer. > 800.

Fig. 10 Drawn from a section through the terminal branches of a collateral
duct of a female albino rat of nine weeks. Zinker’s fixation; Weigert’s iron-
hematoxylin stain. Drawn with the aid of a camera lucida. c.t., connective
tissue. The cross-sections through the six terminal parts show in general a
single layer of epithelium. An occasional flattened nucleus represents the outer
or myo-epithelial layer. > 800.

422 J. A. MYERS

three parts: first, the short excretory duct which extends from
the milk-pore to a noticeable enlargement; second, the enlarge-
ment or milk-sinus, and, third, the milk-duct or canal. He
found the excretory duct lined with a stratified pavement epi-
thelium, while the sinus and milk-ducts possess two epithelial
layers. Benda (’94) found the epithelium near the attached
parts composed of two layers of cells, while in the deeper and
free parts of the ducts the epithelium was described as consisting
of a single layer. Raubitschek (’04) found in a human female
of seven days that the epithelium and the ducts is composed of
two layers. Brouha (’05) described this epithelium as consisting
of two layers throughout the entire system of ducts in virgin
Vespertilio murinus, the rabbit, and the domestic cat. Schil
(12) observed two distinct layers of cells around the ducts before
the stage of puberty. O’Donoghue (’12) confirms Brouha’s
finding in the domestic cat and writes concerning Dasyurus as
follows: ‘‘As in the foetus, so in the adult Dasyurus, all the
branching tubules of the mammary gland are lined by a double
layer of cells, at any rate in the resting animal, and for some
time after ovulation.”’

The fact that a few ducts were observed in the early postnatal
stages of the albino rat with small areas covered with a single
layer of epithelium probably has no significance as in later stages
all ducts observed were lined with a double layer of cells. The
outer layer of cells doubtless represents the so-called basket cells
or myo-epithelial cells of the active mammary gland, while the
inner layer represents the true glandular cells.

True alveoli lined with a single early of cells have been re-
ported in the glands of various animal species before puberty
by several writers. Such alveoli or acini have probably been
observed, as there is no reason to doubt that through individual
variation a few acini might develop just as a mature Graafian
follicle sometimes appears in the ovary of the human fetus or
new-born. However, Schil (’12) concluded that during this
period the mammary gland in the rabbit ordinarily presents
no glandular acini. O’Donoghue (’12), after studying the glands
of Dasyurus, made the following statement: ‘‘The true secre-

STUDIES ON THE MAMMARY GLAND 423

tory alveoli of the gland with their simple epithelial lining do
not make their appearance until the last few days of pregnancy
or until some time after ovulation when this has not been fol-
lowed by fertilization.” The work of these investigators, the
present work, and my previous work (Myers, 716) prove that in
certain animals true alveoli do not usually appear before puberty.

It has been proved quite conclusively that true secretory
alveoli do appear after the first ovulation. In fact, Ancel and
Bouin (711) and Schil (12) showed that in the rabbit the mam-

Fig. 11 Internal view of a wax model reconstructed from two terminal duct
and end-buds of the right first thoracic gland of a female albino rat of seven weeks.
e.b., end-bud; ep.w., epithelial wall; 7., lumen. X 100.

3

mary gland develops after each ovulation not followed by preg-
nancy to the same extent as it develops during the first half of
pregnancy. O’Donoghue (712) and Hartman (’18) reported
somewhat similar changes in the mammary glands of Dasyurus
and the opossum. Most of the investigators believe this develop-
ment is due to the corpus luteum. The enlarged terminations of
the ducts which resemble alveoli in some ten-week rats were
perhaps developed during the first period of heat or some time
after the first ovulation.

424 J. A. MYERS

Secretion in the new-born

It will. be recalled that twelve hours after birth a slight secre-
tion appears in the lumina of the milk-ducts. This secretion is
more abundant at the fourth and fifth days and remains present
in approximately the same amount during the first week. At
no time does it completely fill the lumina of the ducts. After
the second week the secretion gradually appears present in
smaller amounts; however, at the ninth and tenth weeks there
is still some secretion to be found in the lumina. There is only
a very occasional free epithelial cell present in the secretion.
During the first few days of postnatal life such cells manifest
no signs of degeneration, however, later a few were observed
undergoing degeneration. A few red blood corpuscles were ob-
served in the lumina of the ducts at twelve hours after birth,
but none appeared in any of the later stages. An occasional
leucocyte may be seen in the lumina at any stage from birth to
ten weeks. In none of the stages studied could fat be demon-
strated in the lumina or the epithelial cells lining the ducts.

The secretion in the new-born or witches’ milk has been ob-
served in human by a large number of investigators. According
to Brouha (’05), Aristotle called attention to it. The writers
of the second half of the nineteenth and the early part of the
twentieth centuries presented different theories concerning the
origin of this secretion. For example, De Sinety (’75), Rein
(82), Barfurth (82), Czerny (’90), Unger (’98), Schalachta (’04),
Brouha (’05), Berka (’11), Schil (12), and others regarded it
in the human new-born as a true milk secretion, while K6lliker
(54), Milne Edwards (’70), and Raubitschek (’04) believed it
is the product of necrobiotic changes which result in the for-
mation of the lumina. Still another view is that of Keiffer (02),
who described the formation of milk in the new-born as a patho-
logical process.

Several theories have been offered to account for the stimulus
which excites the secretion in the new-born human. The fact
that Brouha concluded that the new-born secretion in well-
developed human may completely precede the birth of the child

STUDIES ON THE MAMMARY GLAND 425

and that Schil observed true milk secretion in the ducts of a
human fetus at the beginning of the eighth month, proves fairly
conclusively that a secretion capable of stimulating the glands
to action is formed in the placenta or ovary of the mother and
transmitted to the blood of both mother and offspring, thus
bringing about the formation of milk in both maternal and fetal
mammary glands.

Berka (711) believes the contents of the lumina of the milk-
ducts through the virgin stage is a part of the witches’ milk
which has been retained in the ducts. On the other hand, Schil
(12) thinks this results from a constant slight secretory activity
of the epithelial cells. O’Donoghue (’12), after studying Dasyu-
rus agrees with Schil in the following:

A trace of a secretion somewhat resembling colostrum is always to be
found as a coagulum in the lumen of the tubules and ducts until it is
removed by the more active secretion of colostrum or milk. It would
appear, then, that the gland, quite apart from the proper milk flow, is
the seat of slow secretory activity, although this secretion is quite dif-
ferent in microscopic appearance from true milk.

Aristotle is said to have observed witches’ milk in the goat.
Creighton (’78) reported it in the guinea-pig in the following
manner:

The fluid expressed from the nipple of the new-born guinea-pig had
the appearance of a watery kind of milk; on microscopic examination,
the milk-globules of ordinary milk were not found, but a more uniform
fluid mass irregularly broken up under the cover-slip into large or small
drops, and without any mixture of cellular elements.

Barfurth noticed such a secretion in the guinea-pig, bitch, and
rabbit and regarded it as a coagulum resulting from simple
transudation. After studying it in Vespertilio murinus, rabbit,
and the domestic cat, Brouha (’05) came to the following con-
clusions in regard to these forms; “la sécrétion natale existe
également, mais elle n’aboutit pas, au moins dans les trois pre-
miéres semaines de le vie, 4 une véritable lactation.’ Schil
(12) observed secretion in the ducts of rabbit embryos as well
as postnatal stages. No secretion has been observed in the
ducts of rat fetuses by me. Brouha’s work and the present

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 4.

426 J. A. MYERS

study lead to the conclusion that in the rabbit, cat, bat, and
white rat the so-called new-born secretion does not contain fat
and hence is not a true lactation as is found in human. This
failure to produce true milk in some of the lower forms may be
due to the feeble development of the mammary gland in the
new-born. The lumina are not completely formed at birth in
the rat, cat, and rabbit, therefore the cells may be unable to
produce a true milk secretion when the stimulus is received.
On the other hand, Brouha found the lumina quite well repre-
sented in Vespertilio murinus at the time of birth. The question
then arises as to whether the stimulus which produces a true
lactation in the human new-born is actually present in the lower
forms of animals. Further experimental evidence is necessary
to decide this question.

Gland stroma

Attention was called to the fact that the stroma in the fetus
is divisible into the mantle layer closely surrounding the ducts
and the true stroma between the ducts. This arrangement of
the stroma persists through the ten weeks’ stage. The mantle
layer increases somewhat in thickness as age advances. ‘This
layer diminishes in thickness as one passes from the primary to
the terminal ducts. About the third week of postnatal life a few
elastic fibres appear in the mantle layer surrounding the primary
and secondary ducts. In later stages elastic fibers are more
numerous and extend further toward the free end of a system of
ducts. The true stroma is at first composed of loose connective
tissue, but soon after birth there appears such a tremendous
invasion of fat that the mantle layer of the ducts is almost en-
tirely surrounded by fat cells.

Free red corpuscles were observed in the stroma twelve hours
after birth, but were not seen in any of the later stages. White
blood-cells were observed in the true stroma in all stages studied,
but in no stage did they appear in greater numbers than they
are usually found in the neighboring tissues.

Berka (’11) called attention to the presence of elastic fibers
in the mantle layer of the virgin human ducts. Schil (12)

STUDIES ON THE MAMMARY GLAND 427

observed elastic fibers in rabbits six months of age and at the
age of puberty.

Red blood corpuscles were observed in the true stroma of the
glands of premature and mature infants by Schlachta (’04).
Brouha (’05) found capillary hemorrhage exists in the stroma of
all infants that died from abrupt asphyxia. He noticed, how-
ever, that in infants dying from other causes there may be a
considerable extravasation of red blood corpuscles. Both of
these authors concluded that the red corpuscles probably reached
the stroma through the process of diapedesis. Owing to the
presence of red blood-cells in the mammary-gland stroma of the
albino rat in comparatively small numbers, it is possible that
they reach the stroma through the process of diapedesis. How-
ever, it must be borne in mind that during parturition and the
removal of the skin there is considerable danger of trauma.

Several authors have reported the presence of an abnormally
large number of leucocytes in the true stroma of the mammary
gland at or soon after birth. Czerny (’90) believed they play
an important réle by entering the lumina of the ducts and carry-
ing the secretion into the lymphatics. Schlachta (04) was un-
able to confirm Czerny’s finding. Brouha (’05), however, found
the immigration of leucocytes in the gland stroma to be less
marked during the activity of the new-born gland than before
or after such activity. The fact that the secretion in the albino
rat is very slight and never becomes a true milk secretion proba-
bly accounts for the absence of an infiltration of leucocytes.

SUMMARY

The results of the present study on the mammary gland of
the male and female albino rat from birth to ten weeks of age
may be summarized briefly as follows:

1. Owing to the thickening of the epidermis over the develop-
ing nipple of the new-born female albino rat, the mammary-
gland area appears lighter than the surrounding epidermis. In
the male no epidermal thickening is present, hence the mammary-
gland area is not visible.

428 J. A. MYERS

The nipple elevation is oblong in shape and very slightly
elevated at the fourth or fifth day. At the end of the second
week it is still oblong, but much more elevated. After the second
week its growth is a gradual one until the time of puberty, when
it takes on the size and shape of the adult virgin nipple.

2. The sulcus around the nipple in the new-born female rat
is apparently a remnant of the prenatal mammary pit. During
the early days of postnatal life this sulcus disappears, but on
the fourth or fifth day an apparently new depression makes its
appearance. This depression deepens so that by the tenth week
there may be developed a definite pocket in which the basal part
of the nipple is located.

3. The epithelial ingrowth (hood) at birth presents a smooth
and quite regular outline. Two or three days after birth, how-
ever, the stratum germinativum becomes somewhat thickened
in some places, thus forming short processes. Such processes
appear on both the outer and inner surfaces of the hood through-
out the stages studied. The central portion of the proximal part
of the hood apparently degenerates, thus deepening the sulcus
around the base of the nipple.

4. Hair follicles are numerous in the integument peripheral’
to the hood and sulcus, but have not been observed nearer the
nipple.

5. Near the highest part of the nipple there appears, a few
days after birth, a slight excavation—the developing milk-pore,
communicating with the intra-epidermal part of the primary
duct. At two weeks there is a slight connection between this
cavity and the lumen of the primary duct. About the sixth
week a complete connection is established with the primary
duct. When the nipple becomes cone-shaped, the milk-pore
appears near the apex.

6. The lumina of the milk ducts are apparently formed by the
process of rearrangement of the cells. The lumen of the primary
duct is not completely formed until after the second week. The
other ducts possess fairly well-developed lumina at the time of
birth.

STUDIES ON THE MAMMARY GLAND 429

7. The walls of the milk-ducts are for the most part lined with
a two-layered epithelium. The inner or glandular layer being
composed of cuboidal to low columnar cells, while the outer (or
so-called myo-epithelial layer) is composed of cells irregular in
size, shape, and arrangement. At nine or ten weeks some of
the terminal processes show indications of developing into alveoli.
The walls of the intra-epidermal part of the primary duct are
lined with stratified epithelium similar to that covering the
surface of the nipple.

8. Masses of subcutaneous fat develop soon after birth. Such
masses Increase in size as age advances, and it is in this fat that
many of the milk-ducts ramify.

9. The stroma is divisible into its usual parts. The mantle
layer forms a thin sheath immediately surrounding the milk-
ducts. At the third week and thereafter elastic tissue fibers
were visible in this layer surrounding the larger ducts. The
true stroma is formed of a loose connective-tissue network in
the early stages, but later shows a very marked infiltration of
fat. Extending through the adipose tissue may be seen an
occasional fair-sized connective-tissue lamina.

10. In no stage was an infiltration of leucocytes observed in
the stroma.

11. A slight secretion appears in the lumina of the milk-ducts
soon after birth. In no case has such secretion appeared in
sufficient quantities to cause distention of the ducts. A trace
of secretion may be seen in the milk-ducts through the ten weeks’
stage. Specific fat stains fail to reveal fat in the secretion of
the milk-ducts or in the epithelial walls, notwithstanding the
fact that much fat was seen in the true stroma outside the walls
of the ducts.

BIBLIOGRAPHY

ANCEL, P., ET Bourn, P. 1909 Sur la fonction du corps jaune. Action du corps
jaune vrai sur la glande mammaire (troisiéme note préliminaire).
Compt. rend. Soc. Biol., T. 66, p. 605. 1911 Recherches sur les fone-
tions du corps jaune gestatif. 11. Sur le détermimisme du développe-
ment de la glande mammaire au cours de la gestation. Journ. dw phys.
et de path. générale, no. 1, p. 31.

430 J. A. MYERS |

BarFurtH, Dietrich 1882 Zur Entwickelung der Milchdriise. Inaug. Diss.,
Bonn.

Bei, E. T. 1914 On the differential staining of fats. Journal of Pathology
and Bacteriology, vol. 19, p. 105.

Benpa, L. 1894 Das Verhiltnis der Milchdriisen zu den Hautdriisen. Der-
matolog. Zeitschr., Bd. 1.

Berka, F. 1911 Die Brustdrsiie verschiedener Altersstufen und wihrend der
Schwangerschaft. Frankfurter Zeitschrift fir Pathologie, vol. 8.

Bressuau, E. 1902 Beitrige zur Entwickelungsgeschichte der Mammarorgane
bei den Beutelthieren. Zeitschrift f. Morphologie u. Anthropologie,
Bd. 4.

1910 Der Mammarapparat (Entwicklung und Stammesgeschichte).
Ergebn. d. Anat. u. Entw., Bd. 19.

1912 Die Entwickelung des Mammarapparates der Monotremen,
Marsupialier und einiger Placentalier. III. EHntwickelung des Mam-
marapparates der Marsupialier, Insectivoren, Nagethiere, Carnivoren
und Wiederkiiuer. Semon’s Zoolog. Forschungsreisen, Bd. 5.

Brounwa, Dr. 1905 Recherches sur les diverses phases du développement et de
l’activite de la mamelle. Archives de Biologie, T. 21.

CREIGHTON, CHARLES 1878 Physiology and Pathology of the breast. London.

Czmrny, A. 1890 Uber die Brustdriisensecretion beim Neugeborenen u. tiber
das Verhiltniss der sogenannten Colostrumkérperchen zur Milchse-
cretion. Festschr. d’Henoch.

De Srnety, 1875 Recherches sur la mamelle des enfants nouveau-nés. Arch.
de physiol. norm. et patholog., T. 2.

Dutcert 1893 Etude histologique de la séerétion lactée. Thése de Mont-
pellier.

Foares, Artuur 1905 Zur physiologischen Beziehungen von Mamma und
Genitalien. Zentralblatt f. Physiol., Bd. 19.

1908 Zur Physiologischen Beziehungen von Mamma und Genitale.
Wien. klin. Woch., Bd. 21.

Frank, R. T., anp UNcmr, A. 1911 An experimental study of the causes which
produce the growth of the mammary gland. Archives of internal
medicine, vol. 7.

GEGENBAUR, C. 1876 Zur genaueren Kenntniss der Zitzen der Siugethiere.
Morpholog. Jahrb., Bd. 1.

Hartman, Cart 1918 Personal communication. Work in progress at The
Wister Institute of Anatomy, Philadelphia.

Herz, P. 1900 Klinische Untersuchungen an 100 Neugeborenen. Inaug.
Diss. Freiburg i. B., 8. 1-61.

Jackson, C. M. 1912 On the recognition of sex through external characters
in the young rat. Biol. Bull., vol. 28.

Kerrrer, H. 1902 La glande mammaire chez le foetus et chez le nourrisson.
Bull. Soe. Belge de Gyn. et d’Obstet., T. 13.

Kuaatscu, H. 1884 Zur Morphologie der Siiugethierzitzen. Morph. Jahrb.,
Bd. 9.

Knauer, Emin 1900 Die Ovarientransplantation. Arch. f. Gyn., Bd. 60.

K6uurker, TH. 1850 Mikroskopische Anatomie, Bd. 2.

STUDIES ON THE MAMMARY GLAND 431

LANE-CLaAyPon, Miss J. E., AND Staryine, E. H. 1906 An experimental en-
quiry into the factors which determine the growth and activity of
the mammary glands. Proc. of the Royal Society, Series B, vol. 77.

Lancer, Cart 1851 Uber Bau und Entwicklung der Milchdriise. Denksch-
rift der Wiener Akad. d. Wiss., Bd. 3.

Lantz, Davin E. 1910 The natural history of the rat. In ‘‘The rat and its
relation to the public health,” by various authors. P.H. and M.H.
Service, Washington, D. C.

Lors, Leo, AND HESSELBERG, Cora 1917 The cyclic changes in the mammary
gland under normal and pathological conditions. Jour. of Experi-
mental Medicine, vol. 25.

Mitne-Epwarps 1870 Lecons sur la physiologie et l’anatomie comparée de
Vhomme et des animaux. Paris, T. 9.

Myers, J. A. 1916 Studies on the mammary gland. I. The growth and dis-
tribution of the milk-ducts and the development of the nipple in the
albino rat from birth to ten weeks of age. Am. Jour. Anat., vol. 19.
(Abstract also published in the Anat. Rec., 1916, vol. 10, p. 230.)

1917 a Studies on the mammary gland. II. The fetal development
of the mammary gland in the female albino rat. Am. Jour. Anat.,
vol. 22. (Abstract also published in the Anat. Rec., 1917, vol. 11,
p. 390.)

1917 b Studies on the mammary gland. III. A comparison of the
developing mammary glands in male and female albino rats from the
late fetal stages to ten weeks of age. Anat. Rec., vol. 13. (Abstract
also published in the Anat. Rec., 1917, vol. 11.)

1918 The histology of the mammary gland in the male and female
albino rat from birth to ten weeks of age. (Abstract.) Anat. Rec.,
vol. 14, p. 46.

O’Donoauve, Cuartes H. 1912 The growth-changes in the mammary appa-
ratus of Dasyurus and the relation of the corpora lutea thereto.
Quarterly Journal of Microscopical Science, vol. 57.

RavusirscueK, H. 1904 Uber die Brustdriisen menschlicher Neugeborenen.
Zeitsch. f. Heilkund., Abth. f. pathol. Anatomie, Heft 1.

Rein, G. 1882 Untersuchungen iiber die embryonale Entwicklungsgeschichte
der Milchdriise. Archiv fiir mikr. Anatomie, Bd. 20 and 21.
Scnuacuta 1904 Beitriige zur mikroskopischen Anatomie der Prostate u.

Mamma des Neugeborenen. Arch. f. mikros. Anat. u. Entwick.

Scuizt, L. 1912 Recherches sur la glande mammaire, sur les phases qu’elle
présente au cours de son évolution et leur determinisme. Thése,
Lyon.

Sternacu, E. 1912 Willkiirliche Umwandlung von Saiugetier-Minnchen in
Tiere mit ausgeprigt weiblichen Geschlechtscharakteren und wei-
blicher Psyche. Archiv fiir die gesammte Physiologie (Pfliiger’s).

SrernHAus, Juttus 1892 Die Morphologie der Milchabsonderung. Arch. f.
Ant. u. Physiol., Suppl. Bd.

Uncer, E. 1898 Beitriige zur Anatomie u. Physiologie der Milchdriise. Anat.
Heft.

vonEpner 1902 Handbuch der Geweblehren des Menschen von Kélliker, Bd.3.

PLATE 1
EXPLANATION OF FIGURES

12 External view of wax model reconstructed from the right third thoracic
gland of a female albino rat of five days. Reconstruction built by A. H. Nerad.
ep., epidermis; m., nipple eminence; s., sulcus. X 37}.

13 External view of a wax model reconstructed from the right second in-
guinal gland of a female albino rat of two weeks. Reconstruction built by A. F.
Branton. ep., epidermis; m.po., milk-pore; n., nipple; s., sulcus. X 37%.

14. Internal view of a wax model reconstructed from the right second inguinal
gland of a female albino rat of two weeks. Reconstruction built by A. F. Bran-
ton. ep., epidermis; ep.in., epithelial ingrowth; p.d., primary duct. X 373.

15 External view of a wax model reconstructed from the right second in-
guinal gland of a female albino rat of six weeks. Reconstruction built by Nellie
Pederson and Frieda Radusch. ep., epidermis; m.po., milk-pore n., nipple; s.,
sulcus. X 373.

16 External view of a wax model reconstructed from the right abdominal
gland of a male albino rat of six weeks. ep., epidermis; m.po., milk-pore. X
373.

17 External view of a wax model reconstructed from the right first inguinal
gland of a female albino rat of nine weeks. Reconstruction by M. H. Litman
and A. F. Branton. ep., epidermis; m.po., milk-pore; n., nipple; s., sulcus. X

OTs.

432

STUDIES ON THE MAMMARY GLAND PLATE 1
J. A. MYERS

433

PLATE 2
EXPLANATION OF FIGURES

18 Drawn from a section through the right first inguinal gland of a female
albino rat at birth. Zenker’s fixation; Mallory’s connective-tissue stain. ¢.,
corium; d.f., developing fat cells; ep., epidermis; h., hair follicles; m.b., muscle
bundle; m.d., milk-ducts; t.s., tela subcutanea. X 83}.

19 Drawn from a section through the right first inguinal gland of a female
albino rat of nine weeks. Zenker’s fixation; Mallory’s connective-tissue stain.
b., blood vessel; c., corium; c.t., connective-tissue lamina; ep., epidermis; f., fat
cells; h., hair follicles; m.l., mantle layer; m.d., milk-ducts. X 833.

434

STUDIES ON THE MAMMARY GLAND PLATE 2
J. A, MYERS

435

Resumen por el autor, Harvey Ernest Jordan.

Histologia de la sangre y médula 6sea roja en la rana leopardo,
Rana pipiens. .

El autor susministra pruebas de que el homélogo del megalo-
eariocito de los mamfferos no es la célula fusfforme de los an-
fibios sino una célula gigante mononucleada que se deriva, como
el megalocariocito de los mamfferos, de un hemoblasto. Todos
los tipos de leucocitos, incluso las células gigantes, pueden formar
pseudépodos, los cuales pueden segmentarse para formar cuerpos
semejantes a plaquetas, granulares o hialinos. Los que tienen
erdnulos metacromiaticos son idénticos a las plaquetas de los
mamiferos. Tales corpusculos se producen también por un
proceso de fragmentacién del protoplasma de un leucocito o un
trombocito. La formacién de plaquetas es, aparentemente, una
produccién secundaria de la actividad normal de los leuco-
citos, la cual se expresa en la formacién y segmentacién de pseu-
dépodos y la degeneracién de estas células. La hematopoiesis
tiene lugar segun el modo monofilético. Los leucocitos neutré-
filos de nucleo polimorfo contienen un sistema astral bien patente,
cuyo centrosoma adopta diversas formas y puede ser sencillo,
doble o multiple; pero estas células no se dividen por mitosis en
su condicién de nucleares polimorfas. El autor discute la im-
portancia de estos hechos sobre la significacién de la amitosis.

Translation by José F. Nonidez
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JUNE 2

THE HISTOLOGY OF THE BLOOD AND THE RED
BONE-MARROW OF THE LEOPARD FROG,
RANA PIPIENS

H. E. JORDAN
Laboratory of Histology and Embryology, University of Virginia

TWO PLATES (SEVENTY-THREE FIGURES)

CONTENTS

esa MOND oe tips 3 (sone eckcnuiin boos for hore ie a anor bale ashes Ge Be Rene Es eae aes ae 437
Cer ee TGLOOOR Nest We AEs lacie oa aodlecs Rue RE Be A 4 ete tectee 438
PRRs aN ROTM ae pac cLace's urs ole ace Seca eee eke ee TR Rte CL an 439
ai Lhe histology ‘of the circulating blood)! )..0)0...2.. 00/0850. Rts 439
Pee ELE EM URTOP VES. 444). tare) Sunfss ys dyatowe oa Pacheco aos Bete 439
PAL VATU CV LES 8 \02/5 tae by Abbe e. q.es.s de Ss cape Be oe 439

ae eae POsinopnilic: leucacytes.% zc ss<. =. + 445,50 «8.0 WU ROS hs ee 441

4- ihe. basophilic leucocytes. //2 6. 6o7 oA Le ie 442
cpne MCULFOpHilic IEUCOGYLES... 1. 5. « Leis. satan unten ed dase eon e. 442

Gop tiie. ChmOMmOCVLER sn. poeknte i auars'sy 5s eae rene SSO Hetrecih kee 443
beeitne histology of the red bone-marrow... ..... 2 ..,02.ccsa cade deen une de. 445
1, The development of the erythrocytes......................4 Petes oes 446

2: The development of the lymphocytes. :... 2. 29.0.0. 2h Jcci aonb wee 447

3. The development of the eosiniplfilic leucocytes.................0. 448

4, The development of the basophilic leucocytes.................... 449

5. The development of the neutrophilic leucocytes................... 450

®,, he development of the thrombocytes..;. 0.2... 2 seat. sss tee 451

7. The development of the plasma-cells and giant-cells............... 451
Pireminuiand Conclusions 3a <2 sqia ceo pili ee a ere als Ae eae oe 452
SSUNLETIAOTAS AAR Ve ee eM SEER Par RAR dete cent eae eee eRe S Hicecrs och ORs Aone ee ee 471

INTRODUCTION

This investigation was begun with the study of the bone-
marrow. Special interest in this tissue was first aroused through
a previous investigation” of the giant-cells of the red bone-
marrow of certain mammals. The primary object of this work
concerns the identification of the amphibian homologue of the
hemogenic giant-cell of mammals, the so-called megakaryocyte.

437

438 H. E. JORDAN

Since the latter cell, under certain conditions, becomes the source
of the blood-platelets, interest centers also on the amphibian
homologue of the mammalian platelet.

It became apparent early in this investigation that a confi-
dent interpretation of the developmental stages of the hemal
elements of the marrow demanded a precise knowledge of the
types of cells of the circulating blood. Accordingly, blood and
marrow were: studied coincidentally and with the same tech-
nics. As regards the blood, the cells of special importance in
this connection are the polymorphonuclear neutrophilic leuco-
cytes and the spindle cells, both containing similar metachro-
matic (azurophil) cytoplasmic granules.

In the red marrow the origin and development of these same
elements claim first attention. Incidentally must be con-
sidered also the developmental history of the eosinophilic and
basophilic leucocytes, the origin of the lymphocytes and of the
erythrocytes, the genetic relationships among the several series
of blood-cells and the bearing of these data on the prevalent
monophyletic theory of blood-cell origin.

MATERIAL AND METHODS

The species of frog used in this work is Rana pipiens. Smear
preparations of blood were stained according to Wright’s technic.
The frogs employed had been left over from the early fall, kept
in cages until the middle of January. Blood smears were made
in October, in January, and in April. The bone-marrow was
taken from the femurs of frogs killed in January, in a certain
number of which the shaft was well filled with red marrow.
The marrow was fixed in a mixture of 100 parts of a saturated
normal-salt solution of corrosive sublimate and 10 parts of forma-
lin. Paraffin sections were cut at 54 and stained according to
Wright’s® technic. This technic gave very excellent results.

BLOOD AND BONE-MARROW OF FROG 439

DESCRIPTION
a. The histology of the circulating blood

Though the study was begun in the reverse order, it seems
preferable to pass from a description of the blood to that of the
marrow. The cellular elements include the following types
which will be described in the order enumerated: erythrocytes,
lymphocytes, eosinophilic leucocytes, basophilic leucocytes,
neutrophilic leucocytes, and thrombocytes.

1. The erythrocytes. 'The typical elliposidal, centrally nucle-
ated, discoid erythrocytes of amphibia are well known and need
no special description for this species of frog. However, certain
atypical forms, comparable to those previously specified for cer-
tain turtle bloods, should again be noted, namely, larger, smaller,
and senile types and their transition forms. The largest type,
comprising a relatively small number of cells, is approximately
one and a quarter times the average size; the smallest type, like-
wise including only a few cells, may be less than half the average
size; these are generally stoutly oval with a spheroidal nucleus
situated nearer one pole; the senile type (fig. 3) is relatively
abundant in the frogs kept under laboratory conditions over
winter; this cell has a large, spheroidal, lilac-colored nucleus and
an expansive shell of very faintly-staining cytoplasm. The
typical erythrocyte (figs. 1 and 2) has a much smaller oval nu-
cleus which stains deeply blue, and a greenish-yellow cytoplasm.
Account must of course be taken of the fact that in smear prep-
arations the cells are spread out under varying pressures and
tensions which operate to produce definitive size variations;
nevertheless, the above-enumerated atypical forms occur in
some degree. ‘This classification agrees in general with that of
Werzberg” except that he makes no mention of the senile forms.

2. The lymphocytes. Under this head may be listed a relatively
large group of mononuclear leucocytes, varying greatly in size,
but with the same nuclear and cytoplasmic characteristics. The
smallest are but slightly larger than the nucleus of a thrombo-
cyte, with a scarcely perceptible shell of basophilic cytoplasm
(fig. 4); the largest (fig. 9) have a diameter approximately twice

44() H. E. JORDAN

that of the smallest, with a larger nucleus and a more expansive
shell of cytoplasm. All the various sizes of lymphocytes have a
common type of nucleus, which stains deeply as a whole, due to
a coarse close-meshed reticulum with many net-knots. It
assumes a deep lilac color. Grouped according to. size, the
lymphocytes may be classified as large and small, but abun-
dant transition forms occur (figs. 6 to 10). Their resemblance to
mammalian lymphocytes is striking.

The most characteristic features of these lymphocytes are
their apparently granular cytoplasmic content and their numer-
ous pseudopods (figs. 5 and 7). Various distortions, nuclear
as well as cytoplasmic, demonstrate that these cells are extra-
ordinarily delicate. The cytoplasm is extremely viscid. Some
of the ‘pseudopods’ are most probably artifacts. But that
at least many should not be so interpreted is sufficiently indi-
cated by the fact that these same cells, or their ancestors, in the
sections of bone-marrow likewise are covered with abundant
pseudopods. For the same reason certain isolated portions of
such pseudopods (fig. 7) must be regarded as actual constrictions
or segmentation products.

In the bone-marrow the lymphocytes have a homogeneous
basophilic cytoplasm (figs. 34 to 40); in the blood smears they
appear to have a coarsely granular basophilic cytoplasm; the
‘oranules’ are very irregular in form and size and frequently
appear massed into clumps (fig. 9). It seems possible that the
granules in question are in fact fixation artifacts, that is, proteid
coagula of the basophilic cytoplasm. However, certain of the
lymphocytes in the blood contain also metachromatic granules
scattered throughout the basophilic substratum. These would
seem to correspond with the ‘leucocytoid lymphocytes’ of Werz-
berg, which this author regards as a distinct class. The evi-
dence seems to indicate that the potentiality to form metachro-
matic granules is a common property of the lymphocytes. There
is no good reason, in my opinion, for classifying lymphocytes on
the basis of the presence of ‘azurophil’ granules. ‘Lymphocytes’
comprise a group of cells characterized by common nuclear and
fundamental cytoplasmic (basophilic) ¢haracteristics; they vary

BLOOD AND BONE-MARROW OF FROG 44]

greatly with respect of size, and the presence or abundance of
metachromatic granules. Moreover, certain of the pseudopods,
especially isolated portions, contain masses of this basophilic and
metachromatic granulation. Such bodies resemble very closely
the blood-platelets of mammals.

There is abundant evidence, which will be detailed below, to
show that the various blood-cells undergo further growth and
development in passing from the blood spaces of the marrow to
the general circulation; in view of which it may be inferred that
the non-granular lymphocytes of the marrow elaborate meta-
chromatic granules after entering the blood stream.

The nucleus of the lymphocyte likewise suffers a considerable

change during the passage of the cell from the marrow to the
circulation. ‘The nucleus of the medullary lymphocyte is vesic-
ular; it contains a distinct plasmosome, several deeply chromatic
karyosomes, many smaller chromatic granules ranged mainly
along the nuclear membrane, and a very delicate chromatic
reticulum (figs. 35 and 37). In general appearance it is much
clearer and lighter staining than the nucleus of the circulatory
lymphocyte.
The attraction sphere of the lymphocytes is usually masked
by the basophilic granulation. Occasionally it appears very
conspicuously as a spheroidal clear area containing a centrosome
or diplosome (fig. 10).

While the evidence at hand does not permit of a final conclu-
sion regarding the genetic relationship between lymphocyte and
platelet, the striking resemblance of the lymphocyte pseudopods
and their constriction products to mammalian blood-platelets is
unmistakable.

3. The eosinophilic leucocytes. This type of cell will be fully
discussed below in connection with the description of its myelo-
cyte ancestors. It will suffice here to state that in its adult con-
dition it contains a polymorph or, in rare instances, a multiple
nucleus, generally located at one pole of the spheroidal cell. Oc-
casionally a mononucleated form may occur. Figure 11 shows a
binucleated form. The spheroidal eosinophilic granules are of
approximately uniform size. They appear hollow, or annular,

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 4

442 H. E. JORDAN

due probably to difference in density between center and pe-
riphery. The granules are imbedded in a lightly basophilic
substratum. Eosinophils constitute only a very small propor-
tion of the leucocytic content of the frog’s blood. In the winter
specimens the eosinophils are rarer than the basophils; in the fall
specimens they are relatively more numerous.

4. The basophilic leucocytes. This type of cell is somewhat
smaller than the eosinophil. It is approximately the size of the .
medium-sized lymphocyte (fig. 12). Its nucleus is centrally
located and vesicular. The nucleus contains a very delicate
reticulum. The cytoplasm is closely packed with spherical baso-
philic, lilac-colored granules, larger on the average than the
eosinophilic granules, but varying somewhat in size and shape,
certain granules having an oval form. In the marrow these
cells have a deeply staining nucleus (figs. 51 to 53) and the cyto-
plasmic granules take a deep blue color with Wright’s stain.

5. The neutrophilic leucocytes. This is by far the most abun-
dant type of leucocyte (figs. 13 to 26). Its nucleus is generally
polymorphous. Mono- and binucleated forms also occur. The
cytoplasm consists of a homogeneous basophilic (blue-staining)
substratum, throughout which are scattered fine metachromatic
(neutrophilic) granules. The granules are always very small,
but vary in size from a scarcely perceptible dust to very definite
spheroidal granules of light lilac color. The resemblance to the
polymorphonuclear neutrophilic leucocytes of mammals (com-
pare figs. 22 and 23) is striking. In areas where the neutro-
philic granules are sparse, the basophilic (blue) substratum also
appears granular.

The most conspicuous feature of these cells, aside from their
characteristic polymorphous nucleus, is the attraction sphere.
‘This is generally situated within the concavity of the lobulated
nucleus (figs. 13, 16, 18, and 19). The neutrophilic granules are
arranged in lines radiating from the sphere (fig. 17). The attrac-
tion sphere may consist of a spherical mass of minute granules
imbedded in a denser protoplasm (figs. 16, 21, 23, and 25) or it
may occur as a Clear area (centrosphere) containing centrally -a
deep-staining granule or centrosome (centriole) (figs. 13, 17, 18,

BLOOD AND BONE-MARROW OF FROG 443

and 19). The latter may be simple, bilobed, or double (diplo-
some). These cells were never seen in mitosis in the circulating
blood. The persistence of such a definite and conspicuous at-
traction sphere in a cell whose nucleus undergoes successive
constrictions in the formation of a polylobular condition (figs. 13
to 21) is of much theoretical interest, a point which will be dis-
cussed below in connection with its myelocytic history.

The relationship between the two types of attraction spheres
is not perfectly clear. However, they correspond closely with
similar types occurring in other cells, e.g., maturing eggs of
Cummingia,? where the clear sphere with its central granule
-(centrosome) of earlier stages becomes converted into a granular
darker sphere through division of the centrosome. In these
leucocytes of the frog the monosome and diplosome occur more
commonly in those cells with the less complex nucleus, the finely
granular spheres in those with the more complex polymorphous
nucleus (compare fig’. 13 and 19 with fig. 25). Though figures
16 and 24 show that this relationship is not invariable, it may be
said that no cells with nuclei like that of figure 13 were seen with
the granular type of sphere. The inference seems justified that
here also the granular type of sphere arises from the clear type
with diplosome, by repeated division of the centrosome. The
granular type would thus represent a disintegration or partition
product. The bearing of this conclusion will be discussed in
connection with the description of these same features of the
medullary neutrophilic leucocytes, where also these same two
types of attraction spheres occur.

6. The thrombocytes. These cells typically are stout fusiform
elements with central, deeply staining (violet), dense, oval
nucleus (figs. 28 and 30). The cytoplasm stains only very faintly
(pink) and contains metachromatic spherical granules of uni-
form size. These granules appear identical with the larger
granules of the neutrophilic leucocytes. The granules are gen-
erally aggregated more or less closely into small groups. The
cytoplasm is apparently extremely viscid and is drawn out at
numerous points into pseudopod-like projections (figs. 27 and
31). These contain the majority of the granules. Certain of

444 H. E. JORDAN

these pseudopods appear to constrict off spheroidal . granular
fragments, resembling blood-platelets of mammals (figs. 30 and
31.) The extreme viscosity of the thrombocyte cytoplasm
would seem to account fully for the fact that these cells generally
occur in larger and smaller closely adherent groups. One line
of evidence casts doubt upon the interpretation of the projec-
tions of the thrombocytes as pseudopods and the isolated bodies
as segmentation products: The cytoplasm is unmistakably very
delicate and very viscid; in contrast with all of the leucocytes of
the marrow, the spindle cells of the marrow do not show pseudo-
pods. This would seem to indicate that the ‘pseudopods’ of the
thrombocytes are artifacts, formed through the operation of
mechanical factors in the process of making the smears. On the
other hand, numerous naked and disintegrating nuclei of throm-
bocytes occur in the blood (fig. 32); these recall the naked nuclei
of the megakaryocytes of mammalian marrows, and in a meas-
ure support Wright’s suggestion that the megakaryocytes of
mammals and the spindle cells of amphibia are homologous
elements.

The resemblance between the granular cytoplasm of the mega-
karyocytes of mammals and that of the thrombocytes of am-
phibia is very close; but it appears no more close than between
this cytoplasm and that of the neutrophilic leucocytes. More-
over, the constriction products of the pseudopods of thrombo-
cytes, neutrophilic leucocytes and certain lymphocytes of the
frog, and those of the pseudopods of the megakaryocytes of
rabbit and guinea-pig are very similar. These corpuscles consist
in common of a spheroidal viscid mass of faintly basophilic
cytoplasm containing a central group of fine metachromatic
spherical granules. The question of homology will be further
discussed below.

At this place must be noted the character of the nucleus. In
the smear preparations this appears dense and chromatic (deep
lilac) with numerous irregular vacuoles. In the marrow this
nucleus is more oval, it contains a chromatic reticulum with
numerous karyosomes, it stains much less deeply and it takes a
blue color (figs. 42 and 43). Moreover, it shows several deep,

BLOOD AND BONE-MARROW OF FROG 445

approximately longitudinal grooves; certain of these are so deep
as to appear to completely divide the nucleus. Similar condi-
tions were previously reported also in blood smears of certain
turtles.1! Downey* questions the accuracy of this interpretation
and inclines to regard all these phenomena as simply deep fur-
rows. One could hardly base final conclusions regarding this
point on smear preparations; the process of spreading the blood
on the slide might very readily divide the already deeply con-
stricted (grooved) nucleus. But similar appearances occur also
in the marrow sections. However, here it could always be
argued that in such cases the plane of section passed above the
bottom of the groove, thus giving the deceptive appearance of a
division of the nucleus. Against the interpretation that certain
of the thrombocyte nuclei are actually split must be cited the
fact that these cells in the circulating blood of the frog contain
generally a more compact, more nearly spherical and ungrooved
nucleus (figs. 27 to 32). If a certain groove of the nucleus of
the immature thrombocyte of the marrow is conceived to lead to
a direct division of the nucleus, then certain mature thrombo-
cytes should be binucleated, which is apparently never actually
the case in the frog. On the contrary, the mature nucleus seems
to have lost the grooves of its immature condition, apparently
through a process of dilatation of the nuclear vesicle, involving a
change from oval to spherical form and an obliteration of the
grooves by reason of a filling up of the nucleus.

b. The histology of the red bone-marrow

This description is based almost exclusively on the bone-mar-
row of the shaft of the femurs of an adult specimen killed the
middle of January. This particular specimen had an especially
abundant red marrow in its femurs. Practically the entire shaft
was packed with red marrow. Other specimens examined at
the same time and during the following month, as also specimens
from a new shipment in April, showed only small patches of red
marrow in the shaft of the femur. Sections were made of this
essentially yellow marrow and also stained according to Wright’s”®

446 H. E. JORDAN

technic. This tissue proved helpful for comparisons in the
course of this study, but the hemopoietic sequences were worked
out on the one specimen with the abundant red marrow. In
the frog the process of blood-corpuscle formation is said to occur
only in the early summer, at which time only the bones contain
red marrow (Starling). The specimen under consideration had
presumably simply anticipated his fellows in this process of con-
verting the yellow into red marrow.

The reticular stroma of the red marrow is packed centrally
with fat cells, blood-vessels, and hemogenic vascular spaces
(angiocysts), and covered peripherally by a layer of differentiat-
ing leucocytes. Erythrocytes are seen only centrally within the
vascular spaces.

The development of fat cells from the mesenchymal stroma
can be traced through a complete series of stages. The process,
however, presents nothing new except that a certain number of
practically adult fat cells contain two nuclei. Endothelium,
erythrocytes, lymphocytes, and leucocytes can likewise be traced
through a complete series of developmental stages back to the
marrow mesenchyma. Hemopoiesis is an essentially similar
process, involving the formation of ‘blood-islands,’ in yolk-sae
and red marrow. The spindle cells arise only intravascularly
as differentiation products from small lymphocytes and from
endothelium. Endothelium may also produce hemoblasts sec-
ondarily.

It seems preferable to describe the developmental history of
the blood-cells in the marrow in the following order: erythro-
cytes, lymphocytes, eosinophilic leucocytes, basophilic leuco-
cytes, neutrophilic leucocytes, and thrombocytes.

1. The development of the erythrocytes. This history can be
read in the smaller blood spaces of the central portions of the
red marrow. The larger spaces and blood-vessels contain only
more mature erythrocytes, mingled with the granular leucocytes
of extravascular origin. The erythrocytogenic process passes
from a stage of solid strands of hemoblasts (marrow ‘blood-
islands’) to one of forming vessels in which the central cells are
free mature erythrocytes and the peripheral cells represent

BLOOD AND BONE-MARROW OF FROG 447

various stages of the differentiation from original hemoblasts.
A certain number of the latter are in intimate connection with
the superficially differentiating endothelium. Endothelial cells
and blood-cells have differentiated from the same primitive cell
mass, and subsequently for a time young endothelium may give
origin to hemoblasts, including thromboblasts.

The original progenitor of the erythrocyte is a cell with a rela-
tively large vesicular nucleus and a narrow shell of slightly baso-
philic eytoplasm. The nucleus is granular in character (figs. 37
and 38). At first irregular in shape, i.e., polyhedral or even
fusiform, the hemoblast soon assumes a spheroidal form. These
cells proliferate mitotically. In becoming an erythrocyte the
hemoblast (erythroblast) again changes into a stoutly oval form
(fig. 41), the cytoplasm becomes oxyphilic as it elaborates hemo-
globin, and the nucleus becomes more compact and more
chromatic.

The great difference in size between the red cells of the marrow
and those of the circulating blood is surprising (compare figs. L
and 41). Those of the smear preparations are approximately
twice the size of those in the sections of the marrow. A certain
large amount of this difference may be accounted for on the basis
of a spreading out in the process of making the smear, but the
residue can only be explained in terms of continued growth after
leaving the marrow. The same phenomenon of postmedullary
growth is evident in some degree in the case of all of the other
types of blood-cells except the basophilic leucocytes (compare
figs. 12 and 452).

2. The development of the lymphocytes. As in the circulating
blood, so in the marrow two main types of lymphocytes can be
distinguished, differing in no essential nuclear or cytoplasmic
features, but only in size (figs. 34 to 41). Transition forms
occur abundantly. These intravascular lymphocytes are struc-
turally indistinguishable from similar extravascular cells from
which the granulocytes develop (compare figs. 88 and 41). Fur-
thermore, the smaller varieties correspond with the hemoblasts
from which erythrocytes, and in part thrombocytes, develop
intravascularly. Intravascularly, the original hemoblast is more

448 H. E. JORDAN

generally of the smaller lymphocyte type with granular vesicular,
nucleus; this may grow into the larger type. Extravascularly,
the original hemoblast is more generally of the large lymphocyte
type (figs. 34 and 35); this may divide mitotically to form small
lymphocytes. These marrow types correspond with the large
and small lymphocytes of the blood smears, except that in the
latter the originally basophilic homogeneous cytoplasm of the
marrow cells becomes granular‘and may in addition elaborate
metachromatic granules. Similarly, the small lymphocytes form
metachromatic granules in becoming thrombocytes within the
blood spaces of the red marrow.

3. The development of the eosinophilic leucocytes. ‘The eosino-
philic myelocytes arise in the extravascular connective tissue.
Their progenitor is indistinguishable from the large and medium-
sized medullary lymphocytes. The mature eosinophils second-
arily enter the vascular spaces, and continue their development
during their passage into the peripheral circulation (compare
fig. 49 with fig. 11). The first indication of the beginning of
differentiation of a ‘lymphocyte’ into an eosinophilic myelo-
cyte is the appearance of a few, very minute, purplish-red gran-
ules (figs. 36 and 44). The first granules are most commonly
aggregated in a certain restricted region, generally in the vicinity
of the centrosome, and only gradually appear in all portions of
the cytoplasm. They increase gradually in number and in size,
always maintain a fairly uniform size for any particular stage
(figs. 45 and 46), and during their later phases resemble those of
the circulating eosinophils in that their centers are less dense,
giving them the appearance of rings (figs. 47 to 50). The gran-
ules apparently grow by a process of swelling involving, beside
increase of diameter of the granules, a rarefaction especially of
_the core, and a slight general decrease in staining capacity.

The original eosinophilic granule has a purplish-red color, the
definitive form a slightly orange-red color. This material gives
no indication of the primitive basophilic granules described by
Downey® for the eosinophilic myelocytes of the guinea-pig.
Nor does either the developing or the definitive eosinophilic
leucocyte contain intranuclear eosinophilic granules, as claimed

BLOOD AND BONE-MARROW OF FROG 449

by Niegolewski!® for Rana esculenta. We are in like disagree-
ment with Niegolewski regarding the basophilic granules of the
mast-cells. Nor is there any evidence that the original eosino-
philic granules of the myelocytes have a nuclear origin, nor any
that they have an extracellular origin, as claimed by certain
investigators, e.g., Weidenreich.2t The eosinophilic granules
arise gradually within the cytoplasmic area, apparently as a
result of some specific activity of this protoplasm.

Coincident with the above-described changes in the granules,
the nucleus also passes through a series of structural and slight
tinctorial alterations. At first the nucleus is approximately
spherical, centrally located, contains a distinct plasmosome, and
is vesicular in character (figs. 44 to 48). It gradually moves
excentrically, meanwhile showing a coarser network and more
numerous karyosomes, and a modification in shape leading
through a reniform to a polylobular, and eventually, in some
cases a multiple, character. Only the younger forms with
spherical or stoutly reniform nucleus are seen in miototic divi-
sion. In common with all myelocytes, including the lympho-
cytes, these cells also form pseudopods which may constrict and
fragment to form free ‘hyaline bodies’ (fig. 36) or, very rarely,
globules with eosinophilic granules. Pseudopod formation and
segmentation seem to be a common property of leucocytes, both
circulatory and medullary.

4. The development of the basophilic leucocytes. 'The basophilic
leucocytes or mast-cells likewise take origin from the common
lymphocyte progenitor, of medium size, in the extravascular
tissue (figs. 38, 39 and 51). The nucleus is characteristically
deep-staining, apparently homogeneous except for a few, barely
visible, large nucleoli (fig. 52). The granules are at first small,
but always larger than the eosinophilic granules, and stain very
deeply blue (fig. 53). The granules become coarser, meanwhile
maintaining their deep coloration. These same cells in the
blood smears have a lighter-staining, centrally located nucleus,
and their basophilic granules show a violet or lilac tinge (fig. 12).
The nucleus of this cell in the marrow, due to its deep-staining
and apparently homogeneous character, suggests degeneration.

450 H. E. JORDAN

However, if this nucleus were properly interpreted thus, it could
scarcely attain a more normal appearance in the circulating
blood. It appears that these cells undergo a further differen-
tiation in passing from marrow to the circulation. They are the
rarest type of leucocyte, but are most probably to be inter-
preted as normal and specific blood elements, as maintained by
Maximow.'® There is no evidence that the granules are nuclear
extrusions, nor that they result from a mucoid degeneration of
the cytoplasm. As in the case of the eosinophilic granules, they
appear to represent the result of some metabolic activity of the
cytoplasm.

5. The development of the neutrophilic leucocytes. These cells
are in certain respects the most interesting among the blood
elements of the frog. It is only in the light of their origin in the
marrow that their true significance can be determined. Such
study shows that they correspond much more closely to the
polymorph neutrophils of mammals than to the leucocytes with
special eosinophilic granules of sauropsida, rabbit, and guinea-
pig. They are in fact the amphibian homologues of the neutro-
philic leucocytes of certain mammals (compare figs. 21 and 22).

These cells alse originate from the common lymphocyte an-
cestor (figs. 34, 35, and 54). The first indication of their differ-
entiation is the appearance of an oxyphilic halo about the
centrosome in the otherwise basophilic cytoplasm. This halo
becomes finely granular and spreads in radiating fashion toward
the periphery of the cell (fig. 54). This disposition of the neu-
‘trophilic granules in radii is, maintained in the definitive forms
(figs. 17 to 26). The granules vary somewhat in size and in the
degree of their lilac coloration, but are always smaller than the
eosinophilic granules. The basophilic substratum remains vis-
ible, and is variably conspicuous in different regions. Coin-
cident with the differentiation of the metachromatic granules,
the nucleus undergoes great morphologic changes, passing ulti-
mately into a polylobular (figs. 69 and 70), and occasionally a
polynuclear, condition (figs. 66 to 68). Only the earliest types,
in which the nucleus is still spheroidal or reniform (fig. 54),
divide by mitosis. The same statement can be made for the

BLOOD AND BONE-MARROW OF FROG 451

eosinophils; basophils were not seen in division. A conspicuous
and significant feature of these cells is the abundance and length
of their granular pseudopods, certain of which may be seen seg-
menting or fragmenting into faintly basophilic globules with
granular centers, simulating thus very closely the blood-platelets
of mammals in their structure and in their origin from mega-
karyocytes (figs. 62 and 68). However, no naked nuclei could
be found. The cells form common lyin the extra vascular tissue
(fig. 69), but a few may possibly arise also from hemoblasts
within the developing blood spaces of the marrow (fig. 41).

6. The development of the thrombocytes. ‘These cells only arise
intravascularly, from small lymphocytes, and in small part
directly from endothelium. As they take on their definitive
oval or fusiform shape they develop metachromatic granules.
In their passage into the circulation they undergo further nuclear
and slight dimensional changes as described above. They are
never seen to arise extravascularly, nor do they undergo divi-
sion; and the nucleus never assumes the crescentic lobulated
condition characteristic of certain leucocytes. In the marrow
they occur singly, in the blood smears generally in groups. The
blood smears show thrombocytes with pseudopods, certain of
which apparently constrict to produce platelets, as first described
by Wright?§ for Amblystoma, leaving eventually disintegrating
naked nuclei (figs. 31 and 32).

7. The development of the plasma-cells and giant-cells. Certain
large lymphocytes undergo a type of differentiation leading to
typical plasma-cells. These plasma-cells are characterized by
the coarse chromatic reticulum of their deep-staining nucleus,
their irregular shape, and their very faintly basophilic, exten-
sively vacuolated cytoplasm (fig. 71). This observation agrees
with Downey’s® conclusion regarding the chief source of origin of
the plasma-cells from lymphoid cells in the mesentery of Rana.
_ A small number of lymphocytes undergo also a hypertrophy
leading to mononuclear giant-cells. These cells contain a rela-
tively enormous nucleus very like that of the younger lympho-
cytes, and a variable shell of basophilic cytoplasm containing
many fine metachromatic (lilac) granules (figs. 72 and 73). Con-

452 H. E. JORDAN

stricting pseudopods of such cells produce bodies comparable
with the blood-platelets of megakaryocyte origin in mammalian
marrows.

DISCUSSION AND CONCLUSIONS

This search for the amphibian homologue of the mammalian
giant-cell from which the blood-platelets take origin has revealed
two types of cells which in a measure fulfill requirements, namely,
the polymorphonucleated neutrophilic leucocytes and the throm-
bocytes. That the so-called megakaryocytes of red bone-
marrow of mammals (e.g., of rabbit and guinea-pig) do actually
at certain stages liberate blood-platelets can be abundantly
demonstrated by the Wright” technic. These mammalian cells
are commonly polymorphonucleated, occasionally polynucleated.
Their faintly basophilic cytoplasm contains an abundance of fairly
uniform metachromatic granules. These cells are not phagocytic,
their occasional content of a leucocyte or two, generally eosinoph-
ilic, is probably to be interpreted as an invasion following early
stages of degeneration. Under certain conditions these cells are
erythrocytopoietic. ‘The mammalian megakaryocyte also traces
its origin to the common lymphocyte (hemoblast) ancestor of
the blood-cells, both in the yolk-sac and in the red marrow.
In both locations the nuclear characteristics are very similar,
and the apparently identical granular cytoplasm produces com-
parable blood-platelets in an identical manner (Jordan). More-
over, in the yolk-sac the smaller, usually binucleated, type of
these giant-cells may differentiate into erythrocytes (Jordan).”

Wright?’ has suggested that the thrombocytes of Amblystoma
represent the megakaryocytes of mammalian red marrow. But
the thrombocyte resembles the megakaryocyte only in respect of
its metachromatic granules scattered throughout a lightly baso-
philic cytoplasm, and its elimination within the blood of gran-
ulated globules similar to platelets, leaving eventually a naked
nucleus. Within the bone-marrow it apparently does not lib-
erate ‘platelets,’ in which respect it contrasts sharply with the
mammalian giant-cells. Moreover, these cells differ greatly in
regard to nuclear characteristics. The thrombocytes resemble

BLOOD AND BONE-MARROW OF FROG 453

platelets in respect of a very adhesive protoplasm, in consequence
of which they frequently become grouped into larger and smaller
masses. But this property of adhesiveness is characteristic also
especially of the lymphocytes, certain of which also have a
variable amount of metachromatic granules.

The polymorphonucleated neutrophilic leucocytes of the frog
seem at first to bear a much closer resemblance to mammalian
megakaryocytes. They are of course smaller than the mega-
karyocytes. While not of the extreme ‘basket’ form, the nu-
cleus is frequently extensively lobulated. The centrosome is
always conspicuous and frequently multiple, as described by
Heidenhain for megakaryocytes. These cells contain, moreover,
a very similar metachromatic granulation, which is likewise
scattered through a basophilic substratum, the latter forming
an exoplasmic layer of variable width. The granules of the
amphibian neutrophilic leucocytes stain less deeply than those of
the mammalian megakaryocytes, but otherwise they are very
similar. Moreover, these leucocytes protrude pseudopods, which
fragment to form platelet-like bodies, like those of megakaryo-
cytes (figs. 61 to 68). These pseudopods even project into
blood-vessels, as do those of the megakaryocytes (fig. 63). The
presence of these alleged giant-cell homologues in the amphibian
blood, as compared with the restriction of the megakaryocytes
of mammals to the bone-marrow, might be due simply to the
fact of the great size of the latter prohibiting entrance into the
capillary circulation. It would seem on the basis of the histo-
logic evidence that the neutrophilic leucocyte of amphibia meets
more nearly the requirements of a megakaryocyte homologue
than does the thrombocyte.

However, the fact that no naked nuclei of these neutrophilic
leucocytes occur, while such are numerous of thrombocyte origin,
contravenes in a measure the assumption of homology between
the polymorphonuclear neutrophilic leucocytes of the frog and
the megakaryocytes of the marrow of the rabbit. In addition to
this objection, there is the further contradiction that certain
mammals have a red marrow containing both typical megakary-
ocytes and typical polymorphonuclear neutrophilic leucocytes
(e.g., cat, dog).

454 H. E. JORDAN

On the other hand, the common property of pseudopod for-
mation and constriction possessed by all the types of leucocytes
within the marrow casts doubt upon the specific nature of the
process as restricted to megakaryocytes of mammalian marrow.
In the frog marrow, primitive lymphocytes, eosinophilic leu-
cocytes and hemoblasts produce ‘hyaline’ bodies by this
method of pseudopod ‘segmentation.’ The neutrophils produce
granulated bodies, resembling platelets. Basophilic leucocytes,
and to some extent eosinophiles, likewise produce granulated
globules. The evidence seems to indicate that this property is
common to both lymphocytes and granulocytes, and that plate-
let-like bodies are formed only incidentally. A thorough study
of the giant-cells of rabbit and guinea-pig also leads to the con-
clusion that these cells produce platelets to some extent by a
constriction of pseudopods, but chiefly by a process of fragmen-
tation of large cytoplasmic areas of degenerating giant cells.
The degeneration is indicated chiefly by the irregular and pyc-
notice character of the nucleus. Similarly, in the case of throm-
bocytes and leucocytes with metachromatic granules, degenera-
tion involves a fragmentation of the cytoplasm and the incidental
formation of platelet-like bodies.

In this connection must be considered the question whether
the cells above described as polymorphonuclear neutrophils are
actually such or only types of non-granular leucocytes, as main-
tained by Werzberg.2> Werzberg’s failure to interpret these
cells as granulocytes must be ascribed to his disregard of the
Wright or a similarly favorable technic. Studied with the
Wright technic, the metachromatic granular content is con-
spicuous and indubitable. This conclusion agrees with the
earlier one of Niegolewski!® and the later one of Downey.*°

Downey® made a special study of the polymorphonucleated
leucocytes of the amphibian Amblystoma. He employed only
smear preparations; these were stained with Wright’s blood
stain. He describes their granules as ‘azurophil,’ which he
regards as close to the ‘special’ granules of the higher animals,
i.e., the neutrophilic granules of the polymorphs of certain mam-
mals. He agrees with Werzberg®* that the lymphocytes of most

BLOOD AND BONE-MARROW OF FROG 455

other amphibia, including the frog, lose their azurophil (i.e.,
neutrophilic, metachromatic) granules when they differentiate
into polymorphonuclears, and maintains that the cytoplasm of
' the definitive polymorphs of these forms is oxyphilic. Neu-
mann!’ likewise denies the presence of granules in the poly-
morphonuclear leucocytes of the frog. However, my prepara-
tions of frog marrow very clearly show that the neutrophilic
myelocytes retain and increase their ‘azurophil’ granules as they
differentiate into the definitive forms of the circulating blood.
Downey* states further that all possible intermediate stages
between larger lymphocytes (with neutrophilic granules) and
neutrophilic polymorphs occur in the circulating blood of Am-
blystoma. This finding is at variance with that of Maximow"™
in the case of Axolotl.

In view of the fact that the complete developmental history
of the polymorphonucleated neutrophilic leucocytes from non-
granular basophilic primitive lymphocytes can be traced in the
sections of the red marrow, Downey’s interpretation of transition
forms between ‘definitive’ lymphocytes and neutrophilic granu-
locytes in the circulation of Amblystoma at first seems quite
improbable. The nucleus of the circulatory lymphocytes is
very different from that of the medullary lymphocytes (compare
figs. 4 dnd 37), a change which indicates progressive differentia-
tion. However, a careful study of the different types of circula-
tory lymphocytes in the frog forces the conclusion that these
lymphocytes do actually metamorphose into the neutrophilic
granulocytes, as urged by Downey* for Amblystoma, but denied
for the frog. Moreover, the nuclei of the medullary and circu-
latory neutrophils and of the circulatory lymphocytes are prac-
tically identical in their fundamental features. It would seem
that a lymphocyte with an already considerably differentiated
nucleus may develop neutrophilic granules abundantly and so
pass over into a neutrophilic granulocyte whose nucleus may
subsequently undergo lobulation. Such a developmental proc-
ess is illustrated in figures 8, 13, 16, and 19.

Figure 8 is a typical lymphocyte with a reniform nucleus.
The only perceptible cytoplasmic difference between it and a

456 H. E. JORDAN

lymphocyte like that illustrated in figure 6 pertains to the
relative proportion of neutrophilic and basophilic granules. The
designation ‘basophilic granules’ in connection with lymphocytes
is always used here with the reservation that the ‘granules’ may
be actually a coagulation phenomenon in an essentially homoge-
neous cytoplasm. Certain circulatory lymphocytes contain only
a few neutrophilic granules; these types lead through transition
forms to neutrophilic granulocytes which show many granules and
interspersed small areas of only basophilic granular material.
The conclusion seems inescapable that lymphocytes may, and
continually do, differentiate into polymorphonucleated neutro-
philic granulocytes within the circulation. This would seem to
dispose of Werzberg’s” classification of the lymphocytes of frog
as nongranular large and small lymphocytes, ‘leucocytoid lym-
phocytes with azurophil granulation,’ and ‘lympholeucocytes.’
In the frog the lymphocytes differ in their cytoplasmic features
only in respect of the relative abundance of neutrophilic (azuro-
phil) granules. This conclusion has an important bearing upon
the discussion regarding the validity of the monophyletic the-
ory of hemopoiesis. The circulatory lymphocyte, the slightly
modified persistent medullary lymphocyte which functions as
the common progenitor (hemoblast) of all types of blood-cells,
still maintains its capacity to differentiate at least into a neutro-
philic granulocyte.

The question then arises whether the Amphibian neutrophils
are actually the homologues of the mammalian megakaryocytes
or of the mammalian polymorph neutrophils. If these cells are
regarded as representing polymorph neutrophils of higher mam-
mals, then amphibian blood, except for thrombocytes, would
seem much closer to mammalian than to sauropsid bloods. This
would conflict with the accepted phylogenetic seriation. How-
ever, neither neutrophilic leucocytes nor recognized hemogenic
giant-cells occur in sauropsid marrows, while leucocytes with
‘special’ (eosinophilic, ellipsoidal) granules are abundant, facts
which favor Downey’s® interpretation. The evidence seems to
point to the homology of the amphibian and mammalian poly-
morph neutrophils. Whether both are the homologues of the

BLOOD AND BONE-MARROW OF FROG 457

‘special leucocytes’ with minute or ellipsoidal eosinophilic granules
characteristic of those forms which lack neutrophils, namely,
certain mammals (e.g., rabbit, guinea-pig) and sauropsids, respec-
tively, is a question of a different order, and one that need not
here be further discussed.

Having thus disposed of both these alleged amphibian homo-
logues of the mammalian megakaryocytes, we may analyze the
possibilities of still another type of cell. This cell occurs only
sparsely. It contains a relatively enormous nucleus, with one or
several plasmosomes, numerous irregular karyosomes, and a
delicately reticulated vesicular nucleoplasm (fig. 73). The
nucleus is enveloped with a variable shell of lightly basophilic
cytoplasm containing metachromatic, lilac-colored granules. It
may be assumed that this cell forms pseudopods which may con-
strict and become free corpuscles. This amphibian giant-cell
develops from a primitive lymphocyte and represents a hyper-
trophied hemoblast. In all these respects, then, it corresponds
with the mononucleated giant-cells of the red marrow of the
femur of the rabbit. It has been shown that the polymorpho-
nucleated giant-cells, the so-called megakaryocytes, develop
from mononucleated giant-cells through nuclear modifications
(Jordan). It seems that in the frog the marrow giant-cells
develop only to the mononucleated stage. Possibly at certain
periods this marrow also would show later polymorphonucleated
phases. The usual statement that giant-cells are not found in
the red marrow of forms below the mammals must therefore be
revised. Careful study of the marrow of sauropsida may pos-
sibly also reveal giant-cell homologues of even closer corre-
spondence.

Another important body of evidence supplied by this material
concerns the monophyletic theory of blood-cell origin. Maxi-
mow? has published a brief preliminary report on hemopoiesis
in Rana temporaria, and interprets his evidence in accord with
this theory. His material includes larvae and the adult bone-
marrow. As regards the process in the latter, his statements
are especially brief. However, the chief points are touched.
My results agree completely with his earlier findings except in

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, NO. 4

458 H. E. JORDAN

one important respect. Maximow claims that in Rana tempo-
raria lymphocytes arise only in the blood-vessels. He notes
that this is a significant departure from what occurs in the case °
of selachians, birds, and mammals, where the lymphocytes form
both extra and intravascularly, and he suggests that this differ-
ence may inhere in ontogenetic differences associated with holo-
blastic and meroblastic cleavage. My preparations of the
marrow of Rana pipiens show, however, that in this species
lymphocytes arise in both locations. If a difference actually
exists in the case of Rana temporaria it must be a specific dif-
ference or a chance variation in Maximow’s specimens and not
one characteristic of amphibia as a group. Maximow” further
calls attention to the fact that in amphibia the lymphoid and
myeloid tissues are not sharply separated topographically, there
being here no sharply defined lymphoid organs with the specific
function of lymphocyte production, like the lymph nodes of mam-
mals, and another tissue with specific granulopoietic function.
This mingling of lymphoid and myeloid tissues and functions in
the bone-marrow gives additional support to the monophyletic
theory.

The erythrocyte, lymphocyte, and leucocyte series in common,
can be traced through gradual steps of development from a type
of cell which is indistinguishable in the several series, namely, a
lymphocyte-like cell, the ‘hemoblast.’ The determining differ-
entiation factor seems to be exclusively environmental. ‘Lym-
phocytes’ that become enclosed by endothelium differentiate
into erythrocytes, or in small part they may proliferate or per-
sist as lymphocytes which undergo slight differentiation in pass-
ing into the blood stream, or some may differentiate into
thrombocytes. Furthermore, the already slightly differentiated
endothelium may to some extent during its younger stages further
differentiate into erythroblasts and into thromboblasts. Con-
sideration of the common origin of the endothelium and the
hemoblasts, from the original medullary ‘blood-islands,’ makes
such differentiation intelligible. Neutrophilic leucocytes may
also to some extent differentiate from intravascular hemoblasts
and from circulatory lymphocytes. But extravascular condi-

BLOOD AND BONE-MARROW OF FROG 459

tions seem more favorable for the neutrophilic differentiation in
the case of the majority of the lymphocytes. Eosinophilic and
basophilic granulocytes, apparenthky differentiate only extra-
vascularly. These enter the blood-channels secondarily through
ameboid activity.

Danchakoff? calls attention to an apparent contradiction in
the argument of the supporters of the monophyletic theory of
blood-cell origin: they note that morphologically identical mes-
enchymal cells and hemoblasts in the same limited regions
develop into both erythrocytes and granulocytes, and infer
from this fact the equipotentiality of these blood-cell ancestors.
Danchakoff argues that if these ‘stem-cells’ develop into differ-
ent products in the same region, then they must have had dis-
similar potentialities and were in fact originally distinct, as
claimed by the adherents of the polyphyletic theory. But this
argument must postulate an identity of the environmental
factors playing upon the hemoblasts in these restricted areas;
this involves an assumption which cannot be supported by tan-
- gible data. It is not at all inconceivable that two adjacent,
identically endowed, cells are nevertheless under environments
sufficiently dissimilar to determine erythropoiesis in one case
and granulopoiesis in the other case. It is no doubt generally
true, as abundant recorded observations show, that hemoblasts
enclosed by endothelium develop into erythrocytes, while ex-
travascular hemoblasts differentiate into granulocytes. But
Danchakoff‘ herself observed in regions in the allantois of the
chick embryo upon which had been grafted particles of adult
spleen, that extravascular hemoblasts could develop into eryth-
rocytes. She, however, interprets her sections to mean that
the extravascular hemoblasts in these instances had received an
erythropoietic bias while previously confined by endothelium,
which could not be reversed under the influence of the new
environment, or that the extravascular spaces containing these
displaced hemoblasts are actually in continuity with the orig-
inal lumen of the blood-vessel through breaks in the endothelial
wall. However, there are other instances in which such explana-
tions will not suffice. For example, in the area vasculosa of the

460 H. E. JORDAN

young turtle embryo I have seen an eosinophilic granulocyte
within an early blood-island, otherwise differentiating largely
into erythrocytes. Until it-can be actually demonstrated that
the environment is identical in such instances, these facts do not
contradict, but sustain the monophyletic view.

Unless we assume an identity of developmental potencies in
the case of the hemoblastic mesenchymal cells, one is forced to
the position that the mesenchymal ancestors of fat cells, pig-
ment cells, endothelium and smooth muscle cells also differ in
respect of specific developmental potencies. But such unquali-
fied position must ignore a large group of histogenetic data.
Mesenchyme cells are apparently of identical nature and en-
dowed with multiple potencies, that is, ‘equipotential’ and
‘polyvalent.’ The expression of any one type of mesenchymal
differentiation—whether as endothelium, erythrocytes, granu-
locytes, or fat—seems accordingly dependent upon extrinsic
factors of a type or degree not yet detectable or measurable by
our means of investigation. Indeed, Danchakoff describes endo-
thelial cells of the spleen graft separating from the wall of small
blood-vessels both centrally, where they pass into the lumen as
hemoblasts and differentiate into erythrocytes, and peripherally
where they become reincorporated with the mesenchyma and
may redifferentiate into granuloblasts. I haye observed com-
parable phenomena in the bone marrow and the body mesen-
chyma of the turtle embryo. These observations show that the
same cell, an endothelial hemoblast, may become either an
erythrocyte or a granulocyte according as the environmental
stimuli change ever so slightly. They furnish, moreover, the
very strongest support to the monophyletic theory. But they
show also that adjacent identically differentiated cells, namely,
young endothelial cells, suffer divergent further differentiation
in an apparently identical environment, namely, inclusion in an
endothelial wall with luminal and mesenchymal surfaces. But
since the cells are identical, i.e., mesenchymal cells slightly
differentiated into functional young endothelal cells, then the
environment of extrinsic stimuli determining the direction of
migration and the initial steps in the further differentiation into

BLOOD AND BONE-MARROW OF FROG 461

different varieties of blood corpuscles, must indeed have differed
to some extent, however slightly.

An interesting feature of this marrow concerns further the
fact of a more or less sharp grouping of the several types of cells
at the various phases of development, very much as is the case
in the red marrow of the femur of the pigeon and of the rabbit
during stages of intense hemopoiesis, that is, a certain group of
cells comprises predominantly small lymphocytes, another
large lymphocytes, another neutrophilic leucocytes, and another
eosinophilic leucocytes. Intravascularly, developing erythro-
blasts predominate; among these are intermingled lymphoblasts,
thromboblasts, and neutrophilic leucoblasts.

Here again it might be argued in favor of the polyphyletic
theory that since the general environment is apparently the
same, the progenitors of the several groups were cells with dif-
ferent and specific developmental potencies, in spite of their
apparent morphologic identity. But in view of the fact that
relatively slight environmental differences apparently determine
developmental differences in early ontogeny, as, for example, in
the developing gut of the mammalian embryo where smooth
muscle, connective tissue, and blood develop in the same re-
stricted regions, it seems more in accord with histogenetic data
to conclude that the several regions filled with different segre-
gated groups of cells were under the influence of different meta-
bolic (fundamentally perhaps only relational, both temporal
and spatial) factors, which determined the specific type of
development. -

This brings us to the question of the development of the
eosinophilic leucocytes. The red marrow of the frog offers an
especially favorable material for the study of the developmental
history of the eosinophilic granulocytes. Unequivocal histologic
evidence accrues with respect to the debated questions regarding
the origin of the cells, whether heteroplastic or homoplastic;
source of the eosinophilic granules, whether endogenous or exog-
enous; the alleged basophilic ‘unripe’ eosinophilic granules, and
the genetic relationship between mast-cells and eosinophilic
leucocytes. An enormous literature of conflicting opinion and

462 H. E. JORDAN

conclusions exists covering these disputed points; this need not
be here again reviewed; it is fully covered in a number of recent
works (e.g., Downey,’ Ringoen?!).

The superior feature of the frog material pertains primarily to
the absence of special eosinophilic leucocytes. The presence of
such types—with fine eosinophilic granules, as in guinea-pig and
rabbit, and ellipsoidal eosinophilic granules, as in sauropsida and
certain fishes—greatly confuses the picture in these forms. In
the marrow of the frog the neutrophilic ‘special’ granulocytes
and the eosinophils are clearly distinguishable from their earliest
stages of development from the common lymphocyte ancestor.
The neutrophilic granulocytes show at first an oxyphilic halo
about the centrosome, which early becomes granular, the granu-
lation spreading throughout the entire cell, the granules always
being of a lighter or darker lilac color. The eosinophils, on the
other hand, begin their differentiation from the parent lympho-
cyte by the elaboration of a few minute deep-staining (purple)
eosinophilic granules. These gradually increase in number and
size, and assume a more brilliantly red (or orange-red) color.
No basophilic granules (i.e., blue-staining) were seen among the
eosinophilic granules at any stage of the development of this
type of granulocytes. The earliest granules are, it is true, of
purple-red color and apparently less oxyphilic than the definitive
granules, but there is no evidence of a differentiation of definitely
basophil granules into eosinophil granules. The definitive gran-
ules, moreover, have a lighter-staining (less dense) center, giving
the appearance of a ring-shaped structure. This appearance is
due to a rarefaction of the center of the granules coincident with
its growth. The growth of the granules, and their change into
more intensively oxyphilic granules is apparently due largely to
a dilution of their substance, specially pronounced in the center.

Downey® comes to the conclusion, on the basis of his study of
the marrow of the guinea-pig, that the eosinophilic granules
develop from basophilic predecessors. His illustrations show
deep blue-staining granules among eosinophilic granules in the
eosinophilic myelocytes. Such blue-staining granules never
appear in the eosinophilic myelocytes of frog in my material;

BLOOD AND BONE-MARROW OF FROG 463

they are present only in the mast-cells. However, I find abun-
dant examples of just such cells as depicted by Downey in sec-
tions of young turtle embryos in the periaortic connective tissue.
It must be emphasized that the fixing and staining fluids in
these instances. (turtle embryo, Jordan; guinea-pig, Downey)
were the same, namely, Helly’s fluid followed by the Giemsa
stain (Downey employed also other similar stains but not
Wright’s combination). In the case of the frog’s marrow the
fixation was secured with a corrosive-sublimate-formalin mixture
and the staining was done with Wright’s stain. It seems prob-
able that a difference of appearance, as concerns presence or
absence. of blue-staining granules among the red granules in
eosinophilic myelocytes depends upon the type of fixing fluid and
staining combinations employed. In the frog material under
consideration, it can only be said that less oxyphilic granules
precede more oxyphilic granules, and that the former are always
smaller than the latter.

All the evidence in this material, moreover, clearly points to a
heteroplastic origin of these cells from lymphocyte ancestors,
and the endogenous origin of the eosinophilic granules. Only
the younger types of myelocyte, i.e., with non-polymorphous
nucleus, are capable of mitotic division and thus of forming new
eosinophilic myelocytes. The granules are in no case ingestion
products of hemoglobin-containing fragments of disintegrating
erythrocytes, but result from the specific activity of the myelo-
cyte protoplasm. Moreover, the nucleus is never invaded by
these cytoplasmic granules, as maintained by Niegolewski,'* nor
is there the slightest evidence in support of Weidenreich’s*
theory that eosinophilic granules are the ingested fragmenta-
tion products of erythrocytes. In this material the erythrocytes
develop only intravascularly, where they also fragment; the
eosinophils, on the other hand, develop only extravascularly.
All the evidence inclines towards the interpretation of these
granules as the product of a specific cytoplasmic activity.

The chief question regarding the basophilic leucocytes (mast-
cells) is whether they represent a specific type of normal blood-
cell or whether they represent degenerations of lymphocytes or

464 H. E. JORDAN

other leucocytes. This study does not touch the disputed ques-
tion of the genetic relationship between the so-called histogenous
and hemotogenous types of basophilic leucocytes. The litera-
ture on this subject has been very fully reviewed in the recent
papers of Maximow,'® Downey,’ and Ringoen.?®* This study is
concerned only with the hemotogeneous type of mast-cells. My
results are in accord with those of the above-mentioned authors,
who conclude that in the mammals investigated (including
guinea-pig and rabbit) the mast leucocytes represent a specific
‘and independent line of granulocytes, of heteroplastic origin,
with basophilic granules resulting from a specific activity of the
protoplasm.

In the marrow of the frog the first indication of mast-cells
‘appears in the nuclear modification of the lymphocyte parent,
namely, a relatively greater contraction and condensation of the
central spheroidal nucleus, giving a deep-blue coloration and a
more homogeneous character to this body. Coincident with
these early nuclear alterations, minute basophilic (deep blue-
‘staining) granules appear; these granules are, however, from the
beginning larger than the ancestors of the eosinophilic granules.
“They grow in size, exceeding that of the definitive eosinophilic
granules, and maintain their original deep blue coloration.
These cells also liberate granulated globules through segmenta-
tion of pseudopods. As found in the blood, these cells show a
centrally located, more vesicular nucleus, a somewhat larger
size, and cytoplasmic granules of deep lilac color. There is not
the slightest indication of a transformation of these basophilic
‘granules into eosinophils at any stage, nor of any degeneration
process connected with the formation of the granules. None of
these cells were seen by me in mitosis at any stage. The cells
apparently originate only heteroplastically from parent lym-
phocytes, and form their granules by an endogenous process.
These observations agree with those previously reported for the
basophilic granulocytes in the turtle embryo.!!' Not only can
both eosinophils and basophils be traced to the same type of
lymphocyte, but the smaller granules of the youngest basophils
are always considerably larger than the smallest eosinophilic

BLOOD AND BONE-MARROW OF FROG 465

granules, and the nucleus of the mast-cells is at all stages after
the earliest quite different from that of the eosinophils. The
presence of these cells in the smear preparations in considerable
numbers renders inadmissable their interpretation in terms of a
disintegration of other types of cells. The deep blue-staining
granules of the basophilic myelocytes change to a violet color as
seen in blood smears. The nucleus also changes from a deep-
staining homogeneous body to a vesicular body with a delicate
reticulum. It maintains an approximately central position in
this cell throughout its history.

The most perplexing matter regarding the thrombocytes con-
cerns their origin. In a study of hemopoiesis in the turtle,
Jordan and Flippin" have described their origin in part from the
endothelium of. the original vascular spaces of yolk-sac and red
marrow and in part from small lymphocytes. Danchakoff?
claims that they arise exclusively from lymphocytes (throm-
boblasts), while Werzberg?> maintains that they are a type sui
generis, having origin neither from endothelium nor lympho-
cytes. The evidence from this study of the red marrow of the
frog confirms our former conclusion regarding their origin, as
described in a study of turtle embryos. The thrombocytes
arise only intravascularly. They can be traced directly to
small lymphocyte-like cells. Secondarily, both lymphocytes
and thrombocytes may arise from endothelium. The close
relationship of small lymphocytes and thrombocytes appears
throughout their entire history. This is especially emphasized,
aside from the close similarity in structure of the nucleus, by the
presence in certain of the small lymphocytes of a variable quan-
tity of metachromatic granules.

The mode of multiplication of the constituent cells of this
marrow centers interest upon amitosis and throws additional
light on the question of this variety of cell division. The orig-
inal blood-cell progenitor (lymphocyte; hemoblast) arises from a
mesenchymal cell. Relatively little proliferative activity occurs
among these mesenchymal cells; the multiplication of hemoblasts
results mainly from a division of already differentiated mesen-
chymal cells. This differentiation process consists essentially in

466 H. E. JORDAN

the rounding up of an irregular, generally stellate cell (fig. 33),
and a condensation and increase in basophily of the cytoplasm
(fig. 34). The nuclei of the typical mesenchymal cells and the
derived hemoblasts are practically identical. These nuclei are
of a vesicular lightly-staining character, with a pale nucleolus,
several deeply chromatic karyosomes, and many very minute
chromatic granules scattered among a very delicate chromatic
reticulum. This description fits also the smaller types of hemo-
_blasts, both intra- and extravascular. In general, further dif-
ferentiation into the various types of blood-cells involves first an
increase in the number of larger, more regular karyosomes, so that
the nucleus has a coarsely granular appearance, many granules
lying peripherally upon the more robust, chromatic, nuclear
membrane. ‘The plasmosome meanwhile persists. Later stages
are characterized mainly by increase in the general chromaticity
of the nuclear sap, giving the entire nucleus, now with a coarser
reticulum, a deep blue coloration. Only the larger lympho-
cytes divide mitotically; the smaller lymphocytes do not divide
as such; they may grow to larger size and then divide mitotically.
Excessive growth of the large lymphocytes leads to giant-cells
(fig. 73). Both neutrophilic and eosinophilic leucocytes also
divide mitotically during their earlier stages, while the nucleus
remains of spheroidal shape. Similarly, young spheroidal eryth-
roblasts may divide mitotically within the blood-vessels.
Neither basophilic leucocytes nor thrombocytes were seen in
either mitosis or amitosis.

Besides proliferation by mitosis, large lymphocytes, young
erythroblasts, and young neutrophilic leucocytes also divide
amitotically. What determines whether a lymphocyte or eryth-
roblast shall divide directly or indirectly remains obscure;
possibly the two modes of division are determined by divergent
metabolic conditions as expressed in the nucleo-cytoplasmic
relationship. However, the amitotic division of the neutro-
philic leucocytes, of which the extreme lobulation of the later
nucleus must probably be reckoned a phase, is the more surpris-
ing, since here an astral system is maintained intact and is at all
stages conspicuous. Possibly here the mitotic incapacity of the

BLOOD AND BONE-MARROW OF FROG 467

centrosomes is related to the high stage of differentiation as
expressed fundamentally in the neutrophilic granules. This
suggestion is in harmony with the facts that the later eosino-
philic myelocytes, the basophilic myelocytes, the thrombocytes,
and the giant-cells also do not divide mitotically. All of these
cells express in their peculiar cytoplasmic condition a high degree
of differentiation. The reason for the failure of basophilic
myelocytes, thrombocytes, and giant-cells fot the most part,
to proliferate even amitotically is under this view not clear.
However, if we assume that the lobulated nucleus of the neutro-
philic myelocytes represents an unfinished amitosis, then pos-
sibly the bilobed nucleus of the basophilic myelocytes, the deeply
grooved condition of the thrombocyte nucleus, and the poly-
morphous character of the giant-cell nucleus (in mammals) may
legitimately be similarly interpreted.

The nuclear amitosis of the neutrophilic myelocytes and the
amitosis of the hemoblasts (large lymphocytes and erythroblasts)
are apparently similar phenomena resulting from quite different
causes— cytoplasmic specificity or high differentiation and intense
proliferative demands, respectively. These dissimilar causative
-factors may, however, be brought under a common head
as regards their effect on the potency and integrity of the
kinetic center, the centrosome. Both conditions may be con-
ceived to reduce relative nutritive conditions below the possi-
ble minimum for centrosomal activity. In essence, amitotic
proliferation, in contrast with mitotic proliferation, results
presumably when the metabolic demands of the protoplasmic
mechanism are such as to deprive the astral system of its
minimum nutritive requirements. This assumption can at
least harmonize the apparently contradictory facts that amitosis
occurs in the cells of rapidly growing tissues as well as in highly
specialized and degenerating cells, an idea first suggested by
Child.!

In this connection attention must again be directed to the two
main types of astral systems in the neutrophilic granulocytes:
one clear with a central monosome or diplosome, the other con-
sisting of a larger or smaller granular sphere. As described

468 H. E. JORDAN

above, these types correspond with those found in the earlier
and later stages respectively, of the first maturation spindle of
the Cummingia tellinoides egg, where the pluricorpuscular is
derived from the unicorpuscular sphere, and represents a dis-
integration or partition product.* This conclusion respecting
the significance of the pluricorpuscular centrosphere is sup-
ported by the facts that the second maturation spindle of Cum-
mingia has almost invariably this type of sphere and that the
segmentation spindles again show spheres of both types. It
would seem a legitimate inference that the two types of centro-
spheres of the neutrophilic leucocytes of frog bear to each other
the same genetic relationship, and that the pluricorpuscular
variety also signifies disintegration. The medullary neutrophils
support this inference; certain examples with lobulated nuclei
contain a granular sphere (fig. 56), but here the sphere is smaller,
the granules less numerous and more conspicuous. It seems
very probable that the mitotic incapacity on the part of these
cells results from an untoward influence upon the centrosome
due to a relative lack of nutritive materials following the main-
tenance of the high degree of specialization involved in the elab-
oration of metachromatic granules.

In a paper dealing with amitosis in the cells of the ioonee
efferentes of the testis of the mouse, 4° suggested that the ami-
totic division of these cells was a consequence of the loss of the
integrity of the centrosome through partition into the basal
granules which give rise to the cilia. While confirming the
observation that ciliated cells of vertebrates do not multiply by
mitosis, but may divide amitotically, Saguchi” claims to be
able to demonstrate the presence and integrity of the centrosome
in ciliated cells of vertebrates, and the origin of the basal gran-
ules and the cilia from mitochondria. Apart from the fact that
his illustrations are far from convincing, both as regards identi-
fication of the centrosome and the genetic relationship between
mitochondria and cilia, the divergence of such a process from the
known functional behavior of mitochondria in general renders
his claims dubious. The recent investigations on mitochondria
have demonstrated that these cytoplasmic elements have no

BLOOD AND BONE-MARROW OF FROG 469

direct genetic relationship to such structures as nerve, muscle
or connective-tissue fibrils, but are intimate cytoplasmic con-
stituents most probably subserving general cell metabolism, not
specific differentiations. On the other hand, we have the very
suggestive fact that the axial filament of the flagellum of the
sperm (comparable to a coarse cilium) does grow out from one
of the partition products of the centrosome of the spermatid.
The fact that a centrosome can still be detected in a ciliated cell
is not disproof that some of the original partition products served
as basal bodies for the development of the ciliary apparatus.
Not all of the partition products need have been thus employed;
several might have remained as discernible granules near the
central portion of the cell. Nevertheless, Saguchi’s” general
conclusion that the ‘‘occurrence of amitosis in ciliated cells is
not owing to the lack of centrosome,” but is ‘due essentially to
the degree of differentiation of the cell-plasm”’ (p. 262) is not in
contradiction with my earlier suggestion that the primary com-
mon cause of amitosis is some deleterious influence of variable
type upon the centrosome. This might be narcotization, rela-
tive lack of sufficient materials to meet metabolic demands as in
conditions of very rapid growth or differentiation, lack of suff-
cient supply of oxygen, or the presence of toxic substances as in
degenerating or pathologic tissues.

The theory that can in my opinion best harmonize the appar-
ently contradictory observations that amitosis occurs in rapidly
growing tissues, degenerating tissues, highly specialized cells
(e.g., secretory, ciliated, granulocytes) and in tissues grown
under experimental conditions (e.g., root tips grown in water
with ether) is one expressed in terms of primary influence upon
the centrosome, effecting either a loss of morphological integrity
as by partition in ciliated cells and certain leucocytes, or loss of
specific physiologic capacity as in narcotized, degenerate, or se-
eretory cells. The effective factor may in the latter case still be
fundamentally a disturbance of the optimum nucleo-cytoplasmic
relationship (producing a nutritive want, subsequently affecting
the centrosome), as recently suggested by Nakahara!’ for ‘se-
cretory or reserve-forming cells,’ where he concludes that amitosis

470 H. E. JORDAN

in adipose cells of insects ‘‘may be for the purpose of securing an
increase of the nuclear surface to meet the physiological neces-
sity due to the active metabolic interchange between the nucleus
and cytoplasm” (p. 509). Even in cases where the original
centrosome has fragmented, presumably under the influence of
cytoplasmic specialization, the effective factor in determining
amitotie division may still in part be the metabolic condition
following a certain degree of nucleo-cytoplasmic balance. That
the fragmentation of itself is not in all cases effective seems
proved by the case of the mitotically dividing blastomeres of
the Cummingia embryo in which pluricorpuscular centrosomes
abound. In the ciliated cells of vertebrates a partitioned cen-
trosome, under the metabolic conditions underlying the type of
differentiation characteristic of a cell developing cilia, is gen-
erally incapable of supporting mitotic division; in the more
vigorous blastomeric cells of Cummingia, characterized by pre-
sumably different metabolic conditions, such a centrosome may
still be able to function in indirect division.

Probably the most important result of this search for the am-
phibian homologue of the mammalian giant-cell turns out to be
the light shed upon the significance of the hemogenie giant-cells,
especially the so-called megakaryocytes of mammals, and upon
the morphologit and genetic significance of the blood-platelets.
Though both thrombocytes and polymorphonucleated neutro-
philic leucocytes show certain characteristics in common with
the mammalian megakaryocytes, the genuine amphibian homo-
logue is a very large mononucleated cell, comparable with a
similar giant-cell of mammalian marrow from which the poly-
morphonucleated megakaryocytes develop. These cells are the
homologues also of the mono- and polymorpho- and polynucle-
ated hemogenic giant-cells of the mammalian yolk-sac during its
hemopoietic phase.!2 These cells in all of these locations develop
from the primitive lymphocyte or hemoblast. This study of
frog’s marrow has shown that pseudopod formation and constric-
tion is a characteristic common to leucocytes at all stages of their
history. Moreover, it is well recognized that cytoplasmic frag-
mentation is a concomitant of degeneration. The processes by

BLOOD AND BONE-MARROW OF FROG 471

which blood-platelets arise from megakaryocytes in mammalian
marrow thus appear to be coincidences of these two cytoplasmic
properties of active and degenerating leucocytes respectively.
This conclusion is further strengthened by the fact that the hemo-
genic giant-cells of the yolk-sac (e.g., 12-mm. pig embryo) like-
wise produce typical platelets in abundance."

SUMMARY

1. The blood of the leopard frog contains the following cellular
elements; erythrocytes, thrombocytes, large and small lympho-
cytes, and neutrophilic, eosinophilic, and basophilic granulocytes.

2. Thrombocytes, neutrophilic granulocytes, and lympho-
cytes contain a variable quantity of metachromatic granules.
Both thrombocytes and lymphocytes show conspicuous granular
pseudopods which may constrict to form platelet-like bodies.

3. The polymorphonuclear neutrophilic leucocyte of the frog
resembles the cell of this designation in certain mammalian bloods.
It is characterized especially by its conspicuous astral system,
its extremely lobulated nucleus, and the presence of metachro-
matic granules arranged in lines radiating from the centrosphere.

4. The red bone-marrow contains the following types of
myelocytes; erythroblasts and thromboblasts, only intravas-
cularly: lymphoblasts, both intravascularly and extravascularly ;
granular myelocytes, including neutrophilic, eosinophilic, and
basophilic cells. All of the leucocytic series show pseudopods,
which may constrict to form ‘hyaline’ bodies and granular
platelet-like corpuscles. Certain lymphocytes differentiate also
into plasma-cells and giant-cells extravascularly.

5. All the types of myelocytes can be traced back to a similar,
apparently identical, progenitor, a lymphocyte-like cell or hemo-
blast arising from the mesenchyma. The evidence from this
material is wholly in accord with the monophyletic theory of
blood-cell origin. Environmental conditions are apparently the
chief factors which determine the line of differentiation a certain
hemoblast shall take. Cords of hemoblasts enclosed by endo-
thelium produce erythroblasts and thromboblasts, or persist in

472 H. E. JORDAN

part as lymphoblasts. The extravascular hemoblasts develop
into lymphocytes and granulocytes, which may enter the vas-
cular spaces secondarily. The lymphocytes of the circulating
blood are apparently only slightly modified marrow lympho-
blasts or hemoblasts, which occur both intra- and extravas-
cularly. A certain number of neutrophilic leucocytes also
originate intravascularly. Circulatory lymphocytes also may
differentiate further into neutrophilic leucocytes.

6. Pseudopod constriction and cytoplasmic fragmentation of
leucocytes are two fundamentally distinct processes leading to
similar results, namely, the production of free cytoplasmic
globules. Pseudopod formation and constriction is a common
property of leucocytes; fragmentation is a degeneration phe-
nomenon associated with nuclear pycnosis and subsequent dis-
integration. Lymphocytes and eosinophilic leucocytes produce
the hyaline bodies; neutrophilic granulocytes, thrombocytes,
and certain lymphocytes with metachromatic granules produce
platelet-like bodies. Globules with basophilic granules arise
from pseudopods of mast-cells. Platelet formation from mega-
karyocytes in mammalian red bone-marrow is apparently a by-
product of this common property of leucocytes and their deriva-
tives, and especially of the disintegration of senile types of these
cells.

7. The amphibian homologue of the mammalian hemogenic
giant-cell is a large mononucleated cell with a relatively large
nucleus, comparable to the similar mononucleated giant-cell of
mammalian marrow from which develop the polymorphonu-
cleated (‘megakaryocyte’ with ‘basket nucleus’) and multinu-
cleated older types. Both cells differentiate from a hyper-
trophied primitive lymphocyte or hemoblast.

8. The polymorphonucleated neutrophilic leucocytes contain
a conspicuous centrosphere which may include a simple, a bi-
lobed, a double, or a multiple centrosome. These cells do not
divide mitotically. It is suggested that mitotic incapacity on
the part of these cells is the result of a relative nutritive want in
consequence of the high degree of specialization involved in the
elaboration of the metachromatic granules. The underlying

BLOOD AND BONE-MARROW OF FROG 473

metabolic demands are conceived to effect an untoward influ-
ence upon the kinetic center, a morphologic aspect of which is
expressed in the pluricorpuscular variety of centrosome.

10

11

13

14

15

16

LITERATURE CITED

Cuitp, C. M. 1907 Studies on the relation between amitosis and mitosis
Biol. Bull., vol. 12, nos. 2, 3 and 4.

Dancuakorr, VERA 1910 Uber die Entwicklung der embryonalen Blut-
bildung bei Reptilien. Bd. 37 (Erginzungsheft).

1918. Cell potentialities and differential factors considered in relation
to erythropoiesis. Am. Jour. Anat., vol. 24, p. 1.

1918. Equivalence of different hematopoietic anlages (by method of
stimulation of their stem-cells). II. (Grafts of adult spleen on the
allantois and response of the allantoic tissues. Am. Jour. Anat., vol.
24, p. 127.

Downey, Hat 1911 The origin and structure of the plasma cells of normal
vertebrates, especially of the cold blooded vertebrates, and the eosin-
ophils of the lung of Amblystoma. Folia Haematologica, Bd. 11,
S. 25-
1913a The granules of the polymorphonuclear leucocytes of Ambly-
stoma, with a few notes on the spindle cells and erythrocytes of this
animal. Anat. Anz., Bd. 44, 8S. 309.

1913 b The development of histogenous mast cells of adult guinea-pig
and cat, and the structure of the histogenous mast cells of man,
Folia Haematologica, Bd. 16, S. 49.
1915 The origin and development of eosinophil leucocytes and of
haematogenous mast cells in the bone marrow of adult guinea-pig.
Folia haematologica, Bd. 19, 8. 148.

Jorpan, H. E. 1910 A cytological study of the egg of Cummingia with
special reference to the history of the chromosomes and the centro-
somes. Arch. f. Zellfors., Bd. 4, S. 248.

1913 Amitosis in the epididymis of the mouse. Anat. Anz., Bd. 43,
S. 589.

JorpAN, H. E., anp Furippin, J. C. 1913 Haematopoiesis in Chelonia.
Folia Haematologica, Bd. 15, 8. 1.

JorpAN, H. E. 1918 A contribution to the problems concerning the origin,
structure, genetic relationship and function of the giant-cells of
hemopoietic and osteolytic foci. Am. Jour. Anat., vol. 24, p. 225.
1918 The histogenesis of blood-platelets in the yolk-sac of the pig
embryo. Anat. Rec., vol. 15, p. 391.

Maximow, A. 1906 Uber entziindliche Bindegewebsneubildung beim Axo-
lotl. Zeigler’s Beitr. z. path. Anat. u. z. Allgem. Path., Bd. 39.

1910 Uber embryonale Entwicklung der Blutzellen bei Selachiern
und Amphibien. Anat. Anz., Bd. 37 (Ergiinzungsheft), 8. 64.

1913 Untersuchungen iiber Blut und Bindegewebe. VI. Uber Blut-
mastzellen. Arch. f. mikr. Anat., bd. 83, Abt. 1.

THE AMERICAN JOURNAL OF ANATOMY, VOL. 25, No. 4

474 H. E. JORDAN

17

18

19

20

NakaHara, W. 1918 Studies of amitosis: its physiological relations in
the adipose cells of insects and its probable significance. Jour. Morph.,
vol. 30, p. 483.

NeuMANN, E. 1903 Haematologische Studien. Virchow’s Archiv., Bd.
174.

NieGoutewskt, F. 1894 Die Ehrlische Granulation der weisen Blutkér-
perchen bei einigen Tierspezies. Inaug. Dissert., Miinchen (cited
from Werzberg).

Ringorn, A. R. 1915 a Observations on the origin of the mast leucocytes
of the adult rabbit. Anat. Rec., vol. 9, p. 233.

1915 b Observations on the differentiation of the granules in the
eosinophilic leucocytes of the bone-marrow of the adult rabbit. Anat.
Rec., vol. 9, p. 683.

Sacucu1, S. 1917 Studies on ciliated cells. Jour. Morph., vol. 29, p. 217.

Staruinc, E. H. 1915 Principles of human physiology, p. 833. Lea &
Febiger, Phila.

WerpenreicH, F. 19095 Zur Frage nach der Entstehung der eosinophilen
Leukozyten. Folia Haematologica, Bd. 2.

Werzsera, A. 1911 Studien zur vergleichenden Haemozytologie einiger
poikilothermer Vertebraten. Folia Haematologica, Bd. 11, 8. 17.

Wriacut, J. H. 1910 The histogenesis of the blood platelets. Jour. Morph.,
vol. 21, p. 263.

EXPLANATION OF PLATES

All figures were made by aid of ;'; Leitz oil-immersion lens and a drawing
camera at an original magnification of 2000 diameters, which was reduced one-
third in reproduction. Figures 1 to 32 are from smear preparations of blood,
stained with Wright’s stain; figures 33 to 73 are of sections of the red marrow of
the femur, fixed in a corrosive-sublimate-formalin solution and stained according
to Wright’s technic for demonstrating the origin of blood-platelets from mega-
karyocytes.

PLATE 1
EXPLANATION OF FIGURES
Blood

1 to3 Normal, older, and senile types, respectively, of erythrocytes. The
color of the nucleus changes from violet to lilac, the cytoplasm from bluish green
to yellow, in passing from 1 to 3.

4 to 6 Small, ameboid, and medium-sized lymphocytes, respectively. The
nucleus stains a deep lilac or violet color; the cytoplasm contains finer and
coarser irregular and clumped basophilic (blue) granules, among which may
appear a variable number of small spheroidal metachromatic (lilac-colored)
granules.

7 Similar medium-sized lymphocyte with numerous pseudopods, some of
which have liberated spheroidal granulated globules by a process of constriction.
Except for the blue color of their granules, these globules are apparently identical
with blood-platelets of mammals.

8 Large lymphocyte with reniform nucleus. The cytoplasm contains a
small amount of a metachromatic granulation scattered among the predomin-
ating basophilic (blue) granules.

9 Large lymphocyte, with nuclear and cytoplasmic characteristics like those
of small lymphocytes.

10 Similar large lymphocyte, with reniform nucleus and a conspicuous
diplosome opposite the nuclear concavity.

11 Binucleated eosinophilic leucocyte (mononucleated, polymorphonuclear,
and polynuclear types also occur). The nucleus has a deep lilac color. The
eosinophilic granules are imbedded in a homogeneous basophilic (blue) sub-
stratum. The granules maintain a fairly uniform size; they appear ring-shaped,
indicating a difference in condensation between center and periphery. The
nucleus almost invariably takes a polar position.

12 Basophilic granulocyte (mast-cell). The nucleus is located centrally; it
is relatively large and vesicular, staining a light blue color, and showing a delicate
reticulum. The granules have a deep lilac color, are fairly uniform in size, and
in general slightly larger than the granules of the eosinophilic leucocytes.

13 to 15 Mononucleated (young) types of neutrophilic granulocytes. The
nucleus has a lilac color. The fine neutrophilic granules are imbedded in a
lightly basophilic (blue) cytoplasm. Figure 13 shows a conspicuous dipiosome
about which the granules are disposed in radiating lines. Figure 15 shows two
pseudopods.

16 Neutrophilic granulocyte with bilobed (dividing) nucleus, and conspicuous,
finely granulated centrosphere. |

17 Binucleated neutrophilic leucocyte.

18 to21 Various types of polymorphonucleated neutrophilic leucocytes.

22 Polymorphonucleated neutrophilic leucocyte of blood of dog.

23 to 26 Various, more complex types of polymorphonucleated neutrophilic
granulocytes. Figure 23 shows a dense granular centrosphere; figure 24, a clear
sphere with a dumb-bell-shaped centrosome (centriole).

27 to 29 Types of thrombocytes. The nucleus is a dense granular body,
staining a violet color. The cytoplasm forms a meagre shell of viscid homo-
geneous character. This contains a variable amount of fine uniform, meta-
chromatic (lilac) granules. The general shape of the thrombocytes is fusiform,
but spheroidal and oval forms also occur (fig. 29).

30 and 31 Thrombocytes with pseudopods, some of which liberate granulated
stellate globules. The latter simulate very closely the blood-platelets of
mammalian bloods.

32 Naked disintegrating nucleus of a degenerating thrombocyte.
476

BLOOD AND BONE-MARROW OF FROG PLATE 1

H, E. JORDAN

477

PLATE 2
EXPLANATION OF FIGURES
Bone-marrow

33 A mesenchymal marrow cell, source from which hemoblasts develop.
The cytoplasm is only very slightly basophilic, staining very faintly blue, and
of homogeneous character. The nuclues is relatively large. It invariably con-
tains a faintly staining plasmosome, occasionally several, and a number of larger
karyosomes and innumerable minute chromioles scattered over a very delicate
chromatic reticulum. The nucleus is vesicular in character and takes only a
faintly blue stain.

34 Primitive large lymphocyte (hemoblast). The nucleus is practically
identical with that of its mesenchyme ancestor. It stains only slightly darker.
The cytoplasm likewise seems more condensed and slightly more basophilic, of
light blue color.

35 Slightly older extravascular lymphocyte, in ameboid activity. The
pseudopods may constrict off hyaline globules, similar to non-granular ‘platelets.’

36 Eosinophilic myelocyte; a slightly differentiated lymphocyte, with pseu-
dopods forming hyaline ‘platelets.’

37 Medium-sized primitive lymphocyte (hemoblast); a slightly differentiated
mesenchyme cell. Note the mesenchymal character of the nucleus. This cell
may likewise develop into a granulocyte.

38 Small extravascular primitive lymphocyte.

39 Slightly later stage in the development of an extravascular small lympho-
cyte, leading to a large lymphocyte or a granulocyte.

40 Similar small lymphocyte with numerous pseudopods, which may form
hyaline ‘platelets.’

41 Group of myelocytes from the periphery of a developing, small marrow
blood-space. The cells include in order from above an erythrocyte, a poly-
nuclear neutrophilic leucocyte, a large lymphocyte, an erythroblast, and a small
lymphocyte. From the latter type of cell develop intravascularly both erythro-
cytes and thrombocytes; from large lymphocytes develop intravascularly only
neutrophilic granulocytes. From these large and small intravascular lympho-
cytes develop also the various definitive lymphocytes of the circulating blood.

42 A medullary erythrocyte (above) and a thrombocyte. (Compare with
figs. 1 and 28 for demonstration of growth in passage from narrow to peripheral
blood stream.) The cytoplasm of the thrombocyte varies in color from a very
faint blue to a very light pink. It contains minute metachromatic (lilac) gran-
ules of uniform size, grouped at the poles, and ranged apparently in single file
along the lateral border. Thrombocytes apparently do not protrude pseudopods
within the marrow. Their nucleus has in general the features of a small or
medium-sized primitively mphocyte, from which cell the thrombocyte develops,
but it is characteristically furrowed by deep oblique and longitudinal grooves.

43 Similar medullary thrombocyte.

44 to 50 Successive steps in the development of an eosiniphilic myelocyte
from a primitive lymphocyte or hemoblast (compare figs. 50 and 11).

(Continued on page 480)
478

BLOOD AND BONE-MARROW OF FROG PLATE 2

H. E, JORDAN

(Continued from page 478)

51 and 52 Basophilic myelocytes (medullary mast-cells). (Compare with
ancestral cell, fig. 39, and circulatory mast-cell, fig. 12.)

53 Earlier stage of medullary mast-cell with finer basophilic (deep blue)
granules and a long pseudopod in process of constricting off a ‘platelet’ with
basophilic granules.

54 Young neutrophilic granulocyte. (Compare with the ancestral large
lymphocytes, figs. 34 and 35.) The centrosome is conspicuous in the concavity
of the reniform nucleus. The neutrophilic granules have not yet spread through-
out the entire cytoplasmic mass; a relatively wide non-granular hyaline area
appears at the left.

55 to 59 Suecessively later stages in the development of the neutrophilic
myelocytes. The polylobular nucleus may constrict to form a polynuclear cell
(fig. 57); this amitotic division of the nucleus may in some cases be followed by a
fission of the cytoplasm (fig. 59). These cells contain astralsy stems like those
of the circulatory cells.

60 Small polymorphonuclear neutrophilic myelocyte. (Compare with fig. 22,
a polymorphonuclear neutrophilic leucocyte of blood of dog.)

61 and 62 Young neutrophilic myelocytes with pseudopods.

63 A neutrophilic myelocyte which has protruded four pseudopods into a
capillary space of the marrow, a phenomenon duplicating that by which meg-
akaryocytes of mammals pass platelets directly into the blood stream.

64 to 68 Various forms of neutrophilic myelocytes with pseudopods which
constrict to form typical platelets (fig. 65).

69 Complex polymorphonuclear neutrophilic myelocyte arising directly from
mesenchyme among a group of fat cells, two of which are indicated at the right.

70 Similar neutrophilic myelocyte from the fatty mesenchyma. (Compare
with fig. 24.)

71 Plasma-cell from the medullary mesenchyma, a derivative from a large
lymphocyte. (Compare with fig. 35.)

72 and 73 Mononucleated giant-cells from the marrow mesenchyme. These
are strictly comparable to the mononucleated hemogenic giant-cells of mammals,
from which the polymorphonuclear giant-cells (so-called megakaryocytes), the
source of blood-platelets, develop. These amphibian homologues of mammalian
hhemogenic giant cells likewise possess a basophilic cytoplasm with metachro-
matic (lilac colored) granules, and arise from hypertrophying hemoblasts.
Segmenting pseudopods of such cells produce typical platelets.

480

SUBJECT AND AUTHOR INDEX

DRENAL gland in the albino rat. The
relative volumes of the cortex and me-
Gillaton thes sean eee reece 291

Albino rats from birth to ten weeks of age.

Studiesonthe mammary gland. IV. The

histology of the mammary gland in male

Cy aolliGicit Geeteeten SRR RMEE San ooae soDedce

Albino rat. The postnatal development of
the suprarenal gland and the effects of in-
pation upon its growth and structure in
the

Albino rat. The relative volumes of the cor-
i and medulla of the adrenal gland in
TLE epee, et Oak TRS DIRE ae Cir cer

Axuis, EpwarpD PHEtps, Jr. The homologies
of the maxillary and vomer bones of Po-

INO Sie gue secon OEE patos sn Ud nubeconer

Arteries of the human lower extremity. The
developmention ther....9-:t. ssn =

ADERTSCHER, J. A. The ultimobran-
chial bodies in postnatal pigs (Sus scrofa)
Blood and the red bone-marrow of the leopard
frog, Rana pipiens. The histology of the
Blood. The origin of the phagocytic mono-
nuclear cells of the peripheral............
Bone-marrow of the leopard frog, Rana pip-
aay The histology of the blood and the

EC eee ete Serine Ae cere cm eee oye a
Bones of Polypterus. The homologies of the
Maxi lanysanG wOMer s,s cee ee oe
Brachydactyly in the domestic fowl. The de-
velopmental relations of................-.
Burrowing mammals. Astudy of thecorrela-
tion of the pelvic structure and the habits

of certain

ELLS of the peripheral blood. The origin
of the phagocytic mononuclear..........

CHAPMAN, Royat Norton. A study of the
correlation of the pelvic structure and the
habits of certain burrowing mammals....

Corpus luteum. Studies on the ovary of the
spermophile (Spermophilus citellus tride-
cemlineatus) with special reference to the

Correlation of the pelvic structure and the
habits of certain burrowing mammals. A
SEUGVAOL tHE a5. hase ee ee OE ei dente

Cortex and medulla of the adrenal gland in

pe albino rat. The relative volumes of
WIG ond anon soctucdccdudoOocoucccononucuraL

ANFORTH, C. H. The developmental
relations of brachydactyly in the domes-

97

Development of the arteries of the human

KOWemextreniityn yer atte ieee ees 55
Development of the intercalated discs. His-

togenesis of the heart muscle of the pig in

relation to the appearance and............ 333
Development of the lobule of the pig’s liver.

21) TCU Eee aM en sent ee EOE EOE 299

Development of the suprarenal gland and the
effects of inanition upon its growth and
structure in the albinorat. The post natal 221

481

Discs. Histogenesis of the heart muscle of
the pig in relation to the appearance and
development of the intercalated..........

Domestic fowl. The developmental relations
of brachydactyly in the..................

DonaLpson, JOHN C. The relative volumes
of the cortex and medulla of the adrenal
gland in the albino rat

Drips, Detia. Studies on the ovary of the
sphermophile (Spermophilus citellus tri-
decemlineatus) with special reference to
the corpus luteum

XTREMITY. The development of
arteries of the human lower . . ....

333

the

ILUM terminale. Factors involved in
fie kormanomol theses ores esterase ne

Fowl. The developmental relations of brachy-
dactyly in the domestic...............--
Frog, Rana pipiens. The bistology of the
blood and the red bone-marrow of the
Jeopand ie sea. c as oss ase sees ce cee
LAND and the effects of inanition upon
its growth and structure in the albino
rat. The postnatal development of the
Suprarenalls. com. sc ach detest tems
Gland in the albino rat. The relative vol-
umes of the cortex and medulla of the ad-
renal
Gland. IV. The histology of the mammary
gland in male and female albino rats
from birth to ten weeks of age. Studies
One MamManys ose ee eee ice
Growth and structure in the albino rat. The
postnatal development of the suprarenal
gland and the effects of inanition upon its

395

221

ABITS of certain burrowing mammals.
A study of the correlation of the pel-
vic structure and the.............-..--
Heart muscle of the pig in relation to the ap-
pearance and development of the inter-
calated discs. Histogenesis of the........ 3
Histogenesis of the heart muscle of the pig in
relation to the appearance and develop-
ment of the intercalated dises............ 333
Histology of the blood and the red bone-mar-
row of the leopard frog, Rana pipiens.
MB aYS Yu rs Bh crc ca SOC TODO eee aicrtcicl
Histology of the mammary gland in maieand
female albino rats from birth to ten weeks
of age. Studies on the mammary gland.
TIW/, MB GH one): Seca ee ns HOME DOS eon ae
Homologies of the maxillary and vomer bones
of Polyptenuss he. 20 e622 ee oi - 4
Human lower extremity. The development
oltheartenies!Of thes:4<. <<< + sere
| Ey
development of the suprarenal gland and
tHeehectsOle s,s c-ee== Mess aes ee
Intercalated discs. Histogenesis of the heart

muscle of the pig in relation to the ap-
pearance and development of the

437

ANITION upon its growth and struc-
ture in the albino rat. The postnatal

482

ACKSON, C.M. The postnatal develop-
ment of the suprarenal gland and the ef-
fects of inanition upon its Bryn and
struture in the albino rat. : 221

JOHNSON, FRANKLIN PARADISE. ‘The ‘de-
velopment of the lobule of the pig’s liver 299

Jorpan, H. BE. The histology of the blood
andthe red bone-marrow of the leopard
frog. Ranaypiplenseessdsratemicer ce eeccl: 437

EOPARD frog, Ranapipiens. The histol-
ogy of the blood and the red bone-
marrow of the.. 437

Liv a ne development ‘of the lobule of the

Tebale of the pig’s liver. The dev clopment _

Gh the este aclacrrras panes wees eae ee 299
Lower extremity. The development of the
arteries of the human.. 55

Luteum. Studies on the ovary ‘of the spermo-
phile (Spermophilus citellus tridecemline-
atus) with special reference to the corpus 117

cJUNKIN, Frank Apam. Theorigin of
M the phagocyte mononuctlear cells of
the peripheral blood: e.s-- eee 27
Mammals. A study of the correlation of the
pelvic structure and the habits of certain
PULFO WINE, evs heen earl sees creer ees 185
Mammary gland. IV. The histology of the
mammary gland in male and female al-
bino rats from birth to ten weeks of age.

Studiesion GHel icc cs cen sleustestoe ee rors eichate 395
Maxillary and vomer bones of Polypterus.
Thevhomologiesiof the: ....5--+-s-2e-- <- 349

Medulla of thea ‘drenal gland in the albino rat.
The relative volumes of the cortex and... 291
Mononuclear cells of the peripheral blood.
The origin of the phagocytic............. 27
Muscle of the pig in relation to the appearance
and development of the intercalated discs.
Histogenesis of the heart................. 333
Myers, J. A. Studies on the mammary
gland. IV. The histology of the mam-
mary gland in male and female albino
rats from birth to ten weeks of age........ 395

VARY of thespermophile (Spermophilus
citellus tridecemlineatus) with special
reference tothe corpus luteum. Studies
OMMENEN eas gi mete cichetes see etee wabeereoanerase 117

ELVIC structure and the habits of certain
burrowing mammals. A study of the

COLLElAtIOMIOL thes eriemce eres cic ties eee 185
Peripheral blood. The origin of the phagocy-

tic mononuclear cellsof the............... 27
Phagocytic mononuclear cells of the periph-

eral blood. The origin of the............ 27

Pigs in relation to the appearance and devel-
opment of the intercalated discs. Histo-

genesis of the heart muscle of the......... 333
Pig’s liver. The development of the lobule
Gisbheres ast Pet eeehtce enema 299

an

INDEX

Pigs (Sus scrofa). The ultimobranchial bod-
Jes)in postnatal kes... =...ceoe eee eee 13
Pipiens. The histology of the blood and the
red bone-marrow of the leopard frog,
1 St hol: HAE a 5 eR eR eee aoac to tas Seinioac 437
Polypterus. The homologies of the maxillary
andtvomen bones ols teeeserece oe eee cae 349
Postnatal development of the suprarenal
gland and the effects of inanition upon its
growth and structure in the albino rat.

TCs. che a Ne eee OO eee oe God Clare 221
Postnatal pigs (Sus scrofa). The ultimo-
branchialibodiestineneeescst cee. s sees 13

R? ANA pipiens. The histology of the blood
eae the red bone-marrow of the leopard

Rats ae birth to ten weeks of age. * Studies
on the mammary gland. IV. The his-
tology of the mammary gland in male
anditemaleialbinosecacuese aoe ee een 395
Rat. The postnatal development of the su-
prarenal gland and the effects of inanition
upon its growth and structure in thealbino 221
Rat. The relative volumes of the cortex and
medulla of the adrenal gland in the albino 291
Red bone-marrow of the leopard frog, Rana
pipiens, The histology of the blood and res
Ne Bae ectnorecons UnUoR oo OQOope Codes aos 70°

ENIOR, H. D. The development of the
arteries of the human lower extremity... 55

Spermophile (Spermophilus citellus iiideoen
lineatus) with special reference to the cor-
pusluteum. Studies on the ovary of the. 117

SrrEETER, GeorGE L. Factors involved in
the formation of the filun terminale...... 1

Structure and the habits of certain burrowing
mammals. <A study of the correlation of
theypelwAGee ears camera reer Cee eee 185

Structure in the albino rat. The postnatal
development of the suprarenal gland and
the effects of inanition uponitsgrowthand 221

Suprarenal gland and the effects of inanition
upon its growth and structure in the al-
bino rat. The postnatal development of

ERMINA E. Factors involved in the
formation of the flum)......---e- sehr

LTIMOBRANCHIAL bodies in postnatal
pigs (Sus scrofa). “hess: sss... seer

OLUMES of the cortex and medulla of the
adrenal gland in the albino rat. The
TOLADIVE: Meee ata falaisl oe tyer elton sera eaten 291

Vomer bones of Polypterus. The homologies
of themaxillary and te as--a 2 eee erent

ITTE, Lucite. Histogenesis of the
heart muscle of the pig in relation to
the appearance and development of

the intercalated dises..... ..........-.- 333

ee

a
ae
A
UF

f

F fi

bs

rit geet?

Rixit ;
vf fe
a eat ik ba)
Ath ante
#

roth
recanyeri etna
Hrectet td penetrates fe
sree aceite
anette)
seererirt
tryst n

tyyrny

mypatcnrney
savaitnuts
e ys

bf
‘i

ay
Dee

Res
ae

thal

add
;

:
sity

ete
tant ne

PEAT
seat

relate

EL eA CUAA MERU HL

i
¢

i
¢ t' ‘
wT Sri taht
SEER AY
t ry ye

i
“

MOT

34

tie

ne yeh ah
se
i ; ;

; 4
au {vents

rect

yayey:

o coitis i
agian LEEALERAL ALAA SLATS PeURLIRELOAS LOSE % ier ' ; treregeeanl
VEE Biearey t ; ayeiee day ray i H ¢
a q

teetey
ray

aad
ids ated

repay,

eae ee

asada
a Rate

hE

sae
yh ik

+ Ee

Tay

= ar

att

=:
aS

ata cat
amet

ast =
SSasy

on
a,

sees

RO

Tati tacesete tans

ots
CRONE EEO PEEL OEE:
LAPEREDCE ECT CED EECE (AORN EE
Ps uM
PeCHease peerage dye? if
seasteziateslertectes
epee feeseecabrenye
ereeveLpentpe en ceaun cs tieed
seeTegcpcedaeeeecner
elie
i

, k (et
4) ANTeCe ata
TeeUN EHEC SeeeUbetitevee?y
RTC r eter eg ree reeee eye’
gerberneneigte egestas wed
TOU s ce tea ree tener ge
SCE CA EO EE CCETSE

i
thoy ne RCE a Nati i y tie reeds

PeSeeseec (ener cops ee eerste
reqgeadeera eats eee arte
TRETUETE Lee CORR Recent red ay
UT RSRRREAL ES LAS
me

The cede peecgegey
eres tedieeeereteee
Crete cee

Pee
f ferreeeerde
ay a Lecter ee
(eceteeeyee ees
teeyeet acer geet

Ale mao

rt
Dats
fig!

se,

Se
maaan:
ae 3

ais)
ne ae

eet
Tamim

manent
rs
Sees

x

{ided
pute
tet
t

=

=

=a

sa 5.

aa aes

eh
i

art

pad

i)

weape

cu

i
cepa
wee eelteae
Piste

t Hine
t Epc tha reels
‘ cuantsbarget
{

t

Seren tonaietee

ii
i
ti
VE

tis

Gi
waar!
teat
{

‘
4
1
i
1
‘

Hf

J i fiseaeeny
AGUea CEOs Cat seed aas elated
PetTeCreri se tessa

ESTEE EEE

ed hadenient

Feegin stag

3S;

Pepeese
beesbey hse
Laie

sigs
in

an

‘2S 32S

{
t
nes
WANG

Sy

ee
peeten:
eoabeey,

ASSO

WF anaes

re

aoe

=)

weno te