Me

ete

JOURNAL

OF

MORPHOLOGY

FounDEpD By C. O. WHITMAN

EDITED BY

dhe oSig ) JEGILINI( GSH Fe
University of Illinois
Urbana, Ill.

WITH THE COLLABORATION OF

Gary N. CALKINS Epwin G. CoNKLIN C. E. McCuiunea
Columbia University Princeton University University of Pennsylvanis

W. M. WHEELER WILLIAM PATTEN
Bussey Institution, Harvard University Dartmouth College

VOLUME 32

MARCH, JUNE, SEPTEMBER
1919

THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY

PHILADELPHIA

CONTENTS

No. 1. MARCH

Cart G. Hartman. Studies in the development of the opossum (Didelphys
virginiana L.). III. Description of new material on maturation, cleavage,
and entoderm formation. IV. The bilaminar blastocyst. Bight text fig-
ures and twenty-two plates. . So Spats rcaneio aio OEE NOI oi

Epwarp PHeEeLps ALLIS, JR. The nee and ne nasal Pee tes in ie a
OSHA SiS NAHM syb-qR OANA DIKAS ARES ae ro ccia a ae Geer e Ao Gin aebioe atin dane c

No. 2. JUNE

Epwarpb Puetps Auuis, JR. The myodome and trigemino-facialis chamber of
fishes and the corresponding cavities in higher vertebrates. Four plates
Ghwemuyan ime lfielITes) iio feed 2 apres Sy em ereatee aye! Cia) days mal cee ae ne. ek ae

ArtHur WILLIAM Meyer. On the nature, occurrence, and identity of the
plasmeance lls7Ot HORACE i nee ecco a yrcadat ist clsa comets tt gla oes ree a ater eens

ApoupH R. RincoeN. The development of the gastric glands in Squalus acan-
thias. Three plates (seven figures).. fs eee etic Geel ce caeerentarars 6

GitMAN A. Drew. Sexual activities of Ne squid, ative peal (Les. ) “The
spermatophore; its structure, ejaculation, and formation. Six plates (forty-
(OMAR CAE DSTSD | eee NEP ane Cay SR ste ta us age oye eae SHU) ll a eI Le tS

No. 3. SEPTEMBER

Witiiam M. GotpsmitH. A comparative study of the chromosomes of the tiger
beetles (Cicindelidae). One hundred twenty-seven figures (ten plates). ...
BENNET M. ALLEN. The development of the thyreoid glands of Bufo and their
normal relation to metamorphosis. One plate (six figures) and one text
JORA DHS) 5 Bs teu aes ce Ata opener Nass RRS Let ba een 1 OS noe ce eee cele
Waro Nakanara. A study of the chromosomes in the spermatogenesis of the
stonefly, Perla immarginata Say, with special reference to the question of
synapsis. Three plates (fifty-one figures) .. ot ones STA :
Stipnney I. Kornwavser. The sexual cheteererenes of the twee Thelia
bimaculata (Fabr.). I. External changes induced s eats theliae
(Gahan). Fifty-four text figures .. 2 SOE COONS Sr REE RN eaais css ope
Toxuyasu Kupo. The facial oe Ge et abe Tiaaeae: ae text fics
and three plates. . niece Tne Nem Sica Beeege :
CLARENCE L. TURNER. “The Scone ne in ate spermary oe he ee epee
foumpapure se. 30. sew tas a tyne cre

ili

145

207

327

. dol

379

437

. 489

. 509

. ddl

. 637

. 681

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

STUDIES IN THE DEVELOPMENT OF THE OPOSSUM

DIDELPHYS VIRGINIANA L.

III. DESCRIPTION OF NEW MATERIAL ON MATURATION, CLEAVAGE

AND ENTODERM FORMATION
IV. THE BILAMINAR BLASTOCYST

CARL G. HARTMAN
The Wistar Institute of Anatomy and Biology and the University of Texas

CONTENTS

PART III. DESCRIPTION OF NEW MATERIAL ON MATURATION, CLEAVAGE, AND

ENTODERM FORMATION

Er OC WE ENO Mixers: S50. |: Sta Ee Os Ca Re POS RR ye ES Ce aes
Gee TeravOny LEMALKS terest reer oe, so PIER oicmerceners Goo aire eked nicl ce oe
(Sy lB GST rrope Oren Moe tian PeWA tmietonr lorie Aare met icics Gibt Uaemne Eon aera Riche, cictovc sak Scola ame eee
Cu Viatenralvanditechmiquessseesnc..: poor secrete rite a ree a
d. External changes in the female opossum at ovulation

Maturation and cleavage to the formation of the blastocyst................
See AEN PERO MALI AMY COD emvatiane tac jac elect Me cP ceo oinpetn eases oi ROME:
eheb uly slain’. ifn abl sate sac we tecve Sasa 2 Sepa ten aia weer sictansese «ptt
CaphheryouneMberin ese resis ew eater oa se Mn ee, ceases, ES eee
edtheshirsty Clea aoe e ciens hey ected See eae teks renee arene exch ccna aks: sepa oe
en hersecondsclea Varese seme varus Metneh san Vif nepal deh auc gti ne Senet
f. On the origin of the crossed arrangement of the first four blastomeres. .
g. Comparison of the 4-celled egg of the opossum and of Dasyurus......
h;. Deutoplasmolysis‘or the elimination of yolk. .2 3.5.22. 6 ee 2 olan
i. Later cleavage to the formation of the blastocyst.....................
j2.On the: fate omthe timstrtwo blastomenress ess. cis os. ce nc nel ene ore

The formation of the entoderm

SAG ENE Lr alliak bases yay aera Me Pe ted revere i hat LR ES ee lia sc Sg ARS etn oh aN
a bhe youngest unilamnar! pla sbocysiseyact cas cee x: -/ «evoke cle olotele en epee
Pebhe rst ENCOGSLIN MAG UME EGOS 8st oho cusehe cca sjeus' x cas) sameeren urchaebdle
. The detachment of the entoderm mother cells.......................::
. The proliferation of entoderm confined to one pole of the egg.........
AeA CRO LM CAL Alc Creer am ey Meee eth: eer d tes 5 ctenc) ah = ,ere.chsuoieted epee os accchs keys
. Included: cells which may not be entodermal........ . cece vo csao >=
Hurthenpolanrrdiienrentiatlones ya. a. ei 4ci- <ccias clteteite teksts =) orto
, Thetemtbnyonte ares, supenicial in postion... ..-.....” sameprmeh acca ee

1

= Doo po Qo»

2 CARL G. HARTMAN

jARRHeyprimmbiviexento dena... savrdcie cise ac c/s ore SIO OIC en eee 65
ky Furtherierowth-of the blastocyst... i. 5i\ssetoa: «+ cece setae erie pees 65
1. The end of entoderm formation and the spreading of the entoderm.... 68
m. The maximum. attenuation of the trophoblast......................... 70
n. The cause of the spreading of the entoderm.............:....-2:ss-«¢+ 71
o. The changing shape and position of the blastocyst.................... 71
De Some, abnormal: egg: 72.251 ee a eee wee ee er nee ele) oof cog. eee ney Ree 72
PART IV. THE BILAMINAR BLASTOCYST
General description............... SELLER EEE ese Soh ie seals onlee eee 73
geekhemmatertali ss ac scene MEGAMI ined 5 ls 2 gee 73
b. The development as seen in the living eggs.........................- 73
he just completed bilamimar blastaeyste yess sere 6 85 &. os uns ee 76
Tia Dhe embryonic ectoderm 72-5 Seen eh eee: ee 76
ba Lhetrophoblasticiareay.c.04o eee COA Oe oe ee 78
¢: Phe entoderm: ..52. 25 cate ets Oa 1d oe) oo ch ee 78
‘The: t=1mm “blastocyst. o%, tistce ac. tee teres ee a eR ics, 3 tice 79
as General desoription: 2.3 ations. ko ne mers ok Oar Es mee 79
b. The bilaminar blastocyst according to Belen yi oy Sar a Ae eR 82
c. /DheAl=mm. blastocyst according ton Miainotw-7,...21ae pe eee See ee 82
‘Eheilatesbilaminar blastocyst parce 2044 eee eee Cer ee 84
a. General-deseription(-s: -cc1oe een ae Bore ee ohare so. ee 84°
by’ The: central light field in the embryonic area. .........2...-.62544.-) ee 85
e. Modifiedentodermalicells;ee. 4. eee ee eee ark Cotiasateeecdc 86
d. The ectoderm of late bilaminar blastocysts............ ae Buia EES 86
é€. Yolk spherules in eetodermpandsentoderm:....-<. 5. 55..04:6.e ete eeee 87
f,. Mesoderm tormationsiniguapedm. cen cesre ss a. eee ae ae ee 88
UMN AT YF 3.8 Po ieee a ey As Se eins EE ea 2 Ck ee ee 89
Tatera ture cited 2... on<4 oS Wee RTs OA One eon) oe She. ae een 97
Explanation of fistires c's. (eee cs & ee ie pees ee eee ee 98
Bates... Cecte btkcucae eee ee a Se ae Si Rae ee ne va See ee ee 99

III. DESCRIPTION OF NEW MATERIAL ON MATURATION, CLEAVAGE,
AND ENTODERM FORMATION

INTRODUCTION

a. Prefatory remarks

The writer’s work on the development of the opossum began
in 1912-13, when a preliminary study of the problem was made
and the approximate breeding season determined for Austin,
Texas. Active collecting was done in January and February,
1914, and again in 1915, and the results of the study of the 415
eggs secured from twenty females were published in March,
1916. A considerable number of eggs, including several missing

DEVELOPMENT OF THE OPOSSUM 3

stages, were also collected during 1916, and at this time many
more eggs and embryos were sacrificed for a series of physio-
logical experiments on the female opossum. As a result of
these experiments I learned a simple and comparatively certain
means of diagnosing a’ female opossum in the earliest stages of
pregnancy and in early oestrus. Since it was felt that this ex-
perience would greatly facilitate collecting in 1917, plans were
made to secure a complete series of eggs, embryos, and pouch
young of this species. ‘The more than hoped for success of the
effort was due to the active interest of Dr. M. J. Greenman,
Director of The Wistar Institute, for it was through the generosity
of the Institute that I was enabled to secure and care for the requi-
site number of animals and also to have the advantage of the able
services of Dr. C. H. Heuser, embryologist at the Institute, who,
with the assistant of Miss Aimée Vanneman, technician in the
School of Zoology, the University of Texas, saw to the proper
fixation and after-treatment of the specimens. Entire credit
also belongs to Doctor Heuser for the unique series of photo-
graphs of living eggs, some of which are herein reproduced. To
Dr. J. T. Patterson is due the initiation of the work on this
interesting marsupial, and his scientific zeal and keen interest in
mammalian embryology have been a constant inspiration to the
writer. I am indebted for indispensable assistance in the
operations on the animals during the last two years’ collecting to
a number of premedical students of zoology, notably to Mr.
Victor Tucker, who stood ready to help at any hour of the day
or night and who performed many of the operations; also to
Miss Janoch and Messrs. Goff, Stiefel, and Kaliski.

During the year 1917-18 I have enjoyed the privileges of a
fellowship at The Wistar Institute and have been its guest while
engaged in a study of the material collected. I am further
indebted through the Institute to Dr. C. H. Heuser for making
some of my best preparations of serial sections and to Mr. T.
H. Bleakney, artist at the Institute, for drawing plate 12 and for
shading and finishing the figures drawn for this paper.

The new material collected since the publication of my former
paper covers many stages there described, and in addition

4 CARL G. HARTMAN

thereto transitional stages not secured before. Among the latter
are litters Nos. 194’, 344, 349, 356, 175’, 339, and 347, which
show the process of entoderm formation in an unbroken series.
Since this phase of the problem had to be entirely rewritten,
and since I now have new material on the early stages, besides
a series of photographs of the living egg in all stages, it has
seemed desirable and profitable to give a complete account of
the development of the opossum from the beginning. ‘This has
been done in the present paper; but the reader is referred to the
writer’s former publication for certain details.

The original notes and the preparations upon which this work
is based, together with alcoholic specimens, will be deposited in
the archives of The Wistar Institute, where they will be easy
of access, and anyone who wishes to do so may examine the
material and test the validity of the conclusions at which I have
arrived in this paper.

b. Historical

In my former publication I reviewed in some detail the work
of Selenka (’87) on the opossum and that of Hill (10) on Das-
yurus. Mention was also made of Caldwell’s discovery (’87) of
the shell membrane enveloping the marsupial egg (Phasco-
larctus), and of a short paper by Professor Minot (’11) on the
bilaminar blastocyst of the opossum. Simultaneous with the
publication of my article, a paper by Spurgeon and Brooks
appeared, giving a description of two litters of opossum eggs
in cleavage (2 to 8 cells). I wish at this point to recur briefly
only to the work of Selenka, leaving the other articles to be
discussed under appropriate headings in the body of the paper.

Salenka’s work on the cleavage and blastocyst formation is
based on 26 eggs secured from two females. One animal yielded
one 2-celled, one 20-celled, and nine unfertilized eggs, all badly
shrunken. I suspect that the ‘2-celled’ and the ‘20-celled’ eggs
are probably specimens in different stages of fragmentation.
The other animal furnished two unfertilized eggs, one4-celled and
one 8-celled egg, two blastocysts of 42 and 68 cells, respectively,
two slightly older blastocysts with a mass of entodermal cells,

DEVELOPMENT OF THE OPOSSUM Da)

and six normal vesicles with thin, partly bilaminar walls. Of
his unsegmented ova I think all were unfertilized: Hence the
42-celled and the 68-celled blastocysts which Selenka describes
are the youngest of his specimens which approach a normal
opossum egg. These two are practically normal except for the
shrinkage of the vesicle from the vitelline membrane and for the
regular gradation in size of the blastomeres from one pole to
the other—a condition entirely accidental and not at all char-
acteristic for this or any other stage in the development of the
opossum. His pear-shaped, thick-walled vesicle with spreading
entoderm (Selenka, ’87, Fig. 1 and 2, Taf. XVIII) is clearly a
degenerating specimen, as I judge by comparison with num-
bers of similar preparations from my collection. Whenever,
in any batch of eggs, there are very retarded specimens, these
are to be regarded with suspicion. Many such abnormal eggs
can be seen in my photographs of living eggs, as, for example
in figure 4, plate 1, and figures 3 and 4, plate 11. Selenka’s in-
terpretation of certain gaps in the walls of his young blastocysts
as the ‘blastopore’ must be rejected for the reason that these gaps ~
are not to be found in completed blastocysts, of which I have a
hundred specimens. Where openings in the blastocyst wall do
exist In young specimens, they are easily explained when the
method of blastocyst formation is understood.

On the origin of the entoderm in the opossum Selenka is not
clear. I must support one of his suggestions, however, for his
designation ‘Urentodermzelle’ as applied to the large cell in his
42- and 68-celled eggs expresses its true function. I previously
described the rather constant occurrence of such cells, all in an
excellent state of preservation; but in the absence of the suc-
ceeding transitional stages, I rejected the view that these are
true entoderm mother cells and considered them of ‘nomor-
phological importance.’ I am now enabled to give a complete
account of the most interesting behavior and the destiny of these
cells.

On the time relations in the development of the opossum my
data substantiates Selenka’s account only in regard to the time
between copulation and parturition, which is thirteen days.

26: CARL G. HARTMAN

But the ages given for all his early stages are far too low, because
the author greatly overestimated the postoestrous period, that
is the interval between copulation and ovulation, which he
states to be five days. The time of beginning of cleavage he
fixes at ‘exactly five times 24 hours,’ a period which he ap-
parently determines on the basis of one experiment in which he
secured what I regard as fragmenting eggs in a condition that
accords very well with eggs about three days old. Again, his
10-hour vesicle is nearer three days old and his 32-hour vesicle
nearer four days; hence the interval of twenty-two hours between
these two stages is substantially correct. In a subsequent paper
I shall discuss these time relations from the abundant, though by
no means simple and harmonious data on hand.

c. Material and technique

1. Material. The opossum eggs on which the present study is
based represents collections made during four seasons. In 1914
- eighteen litters or batches of fertile eggs were secured; in 1915,
seventeen litters; in 1916, fifteen litters, and in 1917, 37 litters—
a total of 87 litters. These refer, of course, only to stages
coming within the field of this paper, for besides these many
litters of older stages were collected; and unfertilized eggs were
removed ad nauseam. The 87 fertile litters, which include eggs
through the bilaminar stage, contained 1009 eggs, of which 641,
or nearly two-thirds, are normal. Thus, about one-third of the
eggs secured from pregnant females are unfertilized or abnormal,
chiefly the former. My previous estimate of one-sixth is there-
fore too low. The average number per litter is 11.5, the ex-
tremes are 1 (No. 94) and 22 (No. 346’), not taking into con-
sideration No. 117’, which numbered 43 eggs by virtue of the
compensatory hypertrophy of the ovary. Table 1 summarizes
the number of eggs mentioned under ‘‘ History of the Animals’’
in the next section.

2. Animals used; reference to illustrations. In the following
summary a brief protocol is presented of each animal furnishing
eggs used in the present study. The data for Nos. 21 to 144

“IJ

DEVELOPMENT OF THE OPOSSUM

TABLE 1

Summary of eggs

Nn 4 f
i=] ia] °
B ae 3
| 3 Zoo | Bi
{Vee mor) ||| 8
= NUMBER OF NORMAL 5 A =
RAS 4 =)
2 EGGS = A < z
& BNE n
es] me & Ho
2 sino co)
Dp 5A 5 =]
a a I
A. Previously reported: litters Nos. 21 to 144, 21 different animals
1, From pregnant animals............... 35 248 130 | 378
2. From pseudopregnant animals........ 2 0 37 37

B. New material: litters Nos. 173 to 415, 45 different animals

3. From pregnant animals at first opera-
Grone (eftuberus)) ss sce ces eerie: ole 22 166 97 263
(63% normal)
4. From second operation (right uterus)
eggs used for this article............ 14 107 79 | 186
; (57.5% normal)
5. Later stages mentioned in summary,
Secondeoperatlomeemsaeeeiers ss caer 16 137 39 | 176
(78.4% normal)
6. From pregnant animals, proportion es-

Cin abe Cheeta no er OAR Aer on Ge 2 os ie 16 120 62 | 182
(Estimated)

7. From pseudopregnant animals........ 15 0 156 | 156

Motaleitemshleeoe 4snanduOwenne sea sees | Lod 641 368 |1009

(63.6% normal)

Total mentioned in summary.......... 120 778 600 |1378

are abstracted from the writer’s previous study (Hartman, ’16),
to which the reader is referred for further details.

In the system which I have employed for the identification
of the specimens each animal receives a number, and the litters
of eggs taken from that animal receive the same number. With-
out further designation, a number may represent all of the eggs
secured from both uteri when the animal is merely killed and
both uteri removed simultaneously; but when the animal was
used for two stages, the simple number represents the first batch

8 CARL G. HARTMAN

of eggs, that is, the contents of the left uterus, removed under
anesthesia and aseptic conditions. The litter of eggs removed
from the right uterus at a later period is designated by the
prime of the number given to the animal. Thus, figure 1 in
plate 1 shows the eggs No. 320 taken from the left uterus of
animal No. 320 at 9 p.m., Jan. 24; figure 2 shows the eggs yielded
by the right uterus of the same animal 53 days later, and these
are designated as No. 320’. The same system applies to 299
and 299’, 292 and 292’, ete. The first litter of eggs is invariably
from the left, the second from the right uterus.

For the reader’s convenience references are made to ‘he
figures illustrating the respective litters of eggs. An asterisk
(*) is placed after the figure or plate containing heliotype illus-
trations of eggs photographed in Ringer’s solution in the living
state.

Animals used inthe study. No. 21. Killed three days after attempted
copulation; mature ovarian eggs (fig. 1, pl. 14).

No. 28. Captured Aug. 23, when seven or eight months old; kept
in solitary confinement until Jan. 23, when she was placed with a
male; male almost killed by female Jan. 26, indicating that oestrus
had passed. Killed Jan. 27; large undischarged follicles with ripe
eggs (fig. 1, pl. 13).

No. 40. Blastocysts near end of entoderm formation with greatly
attenuated non-formative area (figs. 3 and 4, pl. 18).

No. 43. Eggs 0.8 to 1 mm., blastocysts bilaminar throughout
(figs. 4 and 4A, pl. 20), except several like those of No. 40 (fig. 1,

No. 46. 2- to 5-celled eggs (text fig. 4, P).

No. 50. Gianna vesicles of about 50 to 70 cells with none or
with one to several entodermal mother cells (figs. 1 and 3, pl. 7; fig. 11,
pl. 13; figs. 2 and, spl: 16).

No. 52. From pronuclear to.4-celled stages, but mostly unseg-
mented eggs (fig. 21, pl. 14).

No. 54. About same as preceding.

No. 55. Numerous bilaminar blastocysts about 1 mm. in diam.
(figs: 2, 2A, and 6,spleZ)):

No. 56. Unfertilized tubal ova still devoid of albumen layer (fig. 3,
pl. 13; figs. 7 and 14, pl.-14).

No. 58. Undivided, unfertilized uterine eggs.

No. 76. Tubal ova with small trace of albumen (fig. 2, pl. 13; figs. 8,
1 13; and 15°to 177 ple):

No. 81. Fertile eggs, all 4-celled (text fig. 4, N).

DEVELOPMENT OF THE OPOSSUM 9

No. 82. Bilaminar blastocysts like those of No. 50 (figs. 1 and 1A,
pl. 20).

No. 88. Four 4-celled eggs (figs. 8, 11, and 12, pl. 15) and three
young blastocysts! (figs 12 and 19, pl. 16).

No. 85. Cleavage stages; one each of 6, 7, 9, 10, 12, 14, 15, 17,
and 18 cells; three of 8 cells; five of 16 cells (figs. 15 and 17, pl. 15).

No. 88. Of these eggs the collection contains twenty-seven excellent
preparations consisting of 50 to 70 cells and ranging up to 103 cells
each. Most of the eggs have from one to several entoderm mother
cells in their earliest proliferation (figs. 2, 4, and 6, pl. 7; figs. 3, 7
to 11, 18, 21, and 22, pl. 16; fig. 13, pl. 22; compare also page 36,
Hartman, 716).

No. 94. <A single bilaminar blastocyst about 1 mm. in diameter.

No. 112. Degenerating ova from known second oestrus period.

No. 117’. Forty-three eggs, mostly in cleavage, 2- to 16-celled, from
a single uterus, the organs on the opposite having been removed 33
days before; ovary hypertrophied; eggs subnormal in size (figs. 9
and 10, pl. 15).

No. 144. Blastocysts more advanced than those of No. 88; at-
tenuation of non-formative area well under way (figs. 1 to 3, pl. 17).

QOOOo®O

Fig.1. Three blastocysts and one unfertilized egg of litter No. 175’, sketched
with the aid of the camera lucida immediately upon immersion in the fixing
fluid (aceto-osmic-bichromate). X 8.

No. 173. Received Jan. 17. Left uterus and ovary removed 8
p.m., Jan. 18; about 12 eggs, of which 8 were sectioned: 7 are 4-celled
(fig. 5, pl. 3) and one is 3-celled (text fig. 4, L; fig. 8, pl. 15).

No. 173’. At 8 p.m., Jan. 19 (interval 24 hours) about 12 just com-
pleted blastocysts were secured from right uterus; no entodermal
mother cells present (fig. 5, pl. 7). Killed Feb. 9, when the completely
hysterectomized, semi-spayed animal was again coming into heat.

No. 175. Received Jan. 17; removed left ovary and uterus; pseudo-
pregnant; the degenerating eggs were not counted or preserved.

No. 175’. Removed male Feb. 9; killed Feb. 14 (interval 28 days
after operation); 14 eggs: 6 unfertilized, 8 very attenuated vesicles,
entoderm reaching almost to equator (fig. 8, pl. 18; figs. 7 and 7A,
pl. 19 and accompanying text fig. 1). The measurements of the eggs
of litter 175’ are here given as made in salt solution:

1 This is the only instance in all of my records in which the eggs, all removed
at the same time from the animal, consisted of two distinct groups or stages,
separated by a considerable period of development, in this case about twenty-
four hours. There is, of course, a possibility of error on my part due to mixing
of labels in this case.

10 CARL G. HARTMAN

Through shell........ 0.

7 BU 0.70 0.70 0.70 0.68 0.64
Through blastocyst.. 0.54 55

0
0.55x0.4 0.50 0.44 x0.4 0.4 0.47 0.40

No. 189. Received Jan. 22; operation at 10: 45 a.m., Jan. 23; 10 eggs:
3 unfertilized; 7 bilaminar blastocysts, mostly about 1.2 mm. in
diameter; one 0.9 mm. with smaller vesicle probable in dying state
@Gien18; pli is ehes4 pli:

No. 189’. Killed at 10:30 p.m. same day (interval 123 hours); 12
eggs, five of which measured 1.7, 1.7, 1.8, 1.8, 1.9 mm.; stages just
preceding the beginning of mesoderm formation; no record of unfer-
tilized eggs (fig. 20, pl. 13; figs. 4, 4A, 7, 8, 8A, 11, 12A, 12B, pl.22).

No. 191. Jan. 23, 3: 30 p.m., took out left ovary and uterus; 10 eggs:
one a.defective 16-celled stage, others just completed blastocysts of
about 35 cells; recorded measurements in salt solution average 0.56
mm. through shell membrane and 0.14 to 0.15 mm. through ovum
(fig 210); plaalserhio ssl stale):

No. 191’. Jan. 26, 9 p.m., removed right uterus, leaving ovary
(interval 3 days, 53 hours); 11 eggs: 4 bilaminar blastocysts, 1.4 mm.
in diam. in alcohol, almost no albumen; 7 unfertilized eggs. Animal
died Feb. 5; no wound infection.

No. 192. Operated Jan. 23, 4:30 p.m.; 12 eggs: 4 unfertilized; 8
bilaminar blastocysts measuring mostly about 1 mm. in alcohol, but
three measure 0.85, 0.90, and 1.20 mm., respectively. In xylol four
measurements were 1, 1, 1.01, and 1.06, with formative areas 0.62,
0.67, 0.76, and 0.65 mm., respectively.

No. 192’. Killed Jan. 24, 11:30 a.m. (interval 19 hours); 15 eggs: 3
unfertilized; 12 vesicles, of which two measure 1.6 and 2 mm., the
others about 2.4 mm.; pear-shaped embryonic area with primitive
streak.

No. 193. Left uterus and ovary removed Jan. 23, 6 p.m. Number
of eggs not recorded; collection contains 10 poorly fixed preparations,
mostly of small blastocysts of 25 to 36 cells, one egg, however, in the
14-celled stage with two cells in telophase (fig. 8, pl. 3; fig. 9, pl. 13).

No. 193’. Removed remaining uterus Jan. 26, 8:45 p.m. (interval 3
days, 2¢ hours); number of eggs not recorded; five measured in salt
solution 1.15, 1.60, 1.70, 1.70, 1.85 mm.; the first two are bilaminar
blastocysts; the last three have primitive streaks in pear-shaped
areas; one l-mm. blastocyst was dead and one of 1.40 mm. has
imperfect embryonic area; several unfertilized eggs (figs. 1 and 2,
pl. 10; figs. 1,.9, 9A, 9B, 9C; pl. 22).> Ammaldied’ Jan.) 29 hot an
intestinal disease common in cage animals.

No. 194. Jan. 24, 8:30 p.m., found 7 young degenerating eggs like
those shown in figure 6, plate 11, in left uterus which was removed
with the left ovary.

No. 194’. Feb. 9, signs of approaching oestrus returned; Feb. 13,
10 A.M., copulation observed; killed Feb. 17, 25 days after first opera-
tion; 18 eggs: 9 unfertilized; 9 vesicles with entoderm only at em-
byronic area, stage intermediate between Nos. 356 and 352 (fig. 5,

DEVELOPMENT OF THE OPOSSUM ct

pl. 12, figs. 13, 14, 15, pl. 17 and accompanying text fig. 2). Measure-
ments in salt solution and in fixing fluid are as follows:
In Ringer’s solution (average 0.66 mm. and 0.34 mm.).

‘Rhroueheshelle seen On7s) (0268 0265, 0065) 0.65 0.60 0.65 0.65
Through blastocyst........ 0.35 0:35 0:40 0.37 0.35 0.30) 0.32 0.30
In fixing fluid (average 0.57 mm. and 0.33 mm.).
Hill’s fluid Aceto-osm.—biochr
ithroueheshelle ener On COR SonOsoon Room: OFG0R OL50ORGOM O55 0s54
Through blastocyst...... O34 10:32) 0533 0.3) 0733 10.34 10.32 0.35 0.34

No. 203. Received Jan. 26. Removed only left uterus, leaving
ovary, Jan. 28, 8:40 a.m.; about 12 eggs: one with pronuclei (fig. 20,
pl. 14; some 2-celled (text Wed hy tOrd wud. spl. bss hou plik) s
one 3- celled (text fig. 4, 7); others 4-celled (fig. 8, pl. 13: ; fig. 7, ple 15);
one recorded measurement of whole egg is 0.44 mm. ‘through shell
membrane, 0.15 through ovum.

OGOO oO ©

Fig. 2. Five blastocysts with embryonic areas and one unfertilized egg of
litter No. 194’, sketched alive in Ringer’s solution with the aid of the camera
lucida. X 8.

No. 203’. Second operation at 11:45; date not recorded in protocol,
but cage record indicates that the time was 11:45 p.m., Jan. 29; hence
the interval was probably 39 hours; removed only right uterus, leaving
both ovaries. Several young vesicles of about 50 cells; one measure-
ment in salt solution is 0.5 mm. through shell membrane, 0.16 mm.
through ovum. Killed Feb. 16; corpora lutea had almost entirely
disappeared, follicles still small, but mammae very thick as in
pregnancy.

No. 205. Captured by dogs Jan. 28, the skin being ripped at
shoulder; operated Jan. 29, 9:15 p..; 10 eggs: 9 young bilaminar
blastocysts with much albumen at one ‘pole: entoderm quite or nearly
reaching non-formative pole; three measured in salt solution 1.05 mm.
three 1 mm., two 0.90 mm., and one 0.75 mm.; one unfertilized egg
measured 0.72 mm.; in alcohol after two years the eggs measured about
O70rmmra hes, 29° Ti and-12,-pl. 19):

No. 205’. Killed Jan. 30, 10:10 a.m. (interval 13 hours); 13 eggs:
3 unfertilized, the remainder vesicles with faint primitive streak in
rounded areas or with more advanced primitive streaks in pear-shaped
areas. Two of the former measured in alcohol 1.45 and 1.83 mm.

with areas 1.1 and 1.2, Tespectiv ely; one of the latter 2 mm. with area
1.32 x 1 mm.

10; CARL G. HARTMAN

No. 208. Caught Jan. 29; Jan. 30, 11:30 a.m., removed left uterus
containing 4 eggs: 3 unfertilized, measuring in alcohol 1.1 mm., and
one young bilaminar blastocyst measuring in salt solution 0.85 mm.,

“in alcohol 0.8 mm.; size of vesicle in salt 0.65 x 0.6 x 0.5 mm. (figs. 10,
10A, and 10B, pl. 19).

No: 208’. Killed Jan. 31, 1:45 p.m. (interval 26% hours); right
uterus yielded 8 eggs: one unfertilized, one defective vesicle, 1.25 mm.
in diameter, and others like No. 205’, measuring in salt solution 1.30,
1.45, 1.59, 1.59, 1.94, 2.85 mm.

No. 214 (D. marsupialis). Received from south Texas, Feb. 1;
operated Feb. 2; a dozen or more undivided, unfertilized eggs, a slight
degeneration apparent only after sectioning.

No. 214’. Feb. 6, right uterus removed, leaving right ovary; 14 large
eggs with opaque shell membrane, dense albumen and fragmenting ova
(interval 4 days). Killed Feb. 28; after 22 days the completely
hysterectomized and semi-spayed animal had again come into heat.

No. 256. Removed three pouch young Feb. 9; 10 days later, numer-
ous small eggs in early stage of degeneration were found in uteri.

No. 285. Caught Jan. 12; injured. Jan. 13, 10:25 p.m., 10 eggs: 2
unfertilized, the remainder small blastocysts partially lined with ento-

2QODOQVOP

Fig. 3. <A, B, and C, three eggs of litter No. 285; D, E, and F, three eggs of
litter No. 285’, sketched alive in Ringer’s solution with the aid of the camera
lucida: <8!

derm; eggs measured 0.85 to 0.9 mm. in salt solution; no preparations
made of this litter (text fig. 3, A to C).

No. 285’. Killed Jan. 14, 12:30 p.m. (interval 14 hours); 14 eggs: 3
unfertilized; others as illustrated by D, E, and F, text figure 3; ento-
dermal lining complete (least advanced, figs. 3, 3A, and 3B, pl. 20;
most advanced in figs. 7 and 7A, pl. 21).

No. 287. Jan. 15, 8:15 a.m.; 7 or more eggs (the collection contains
7 preparations) undivided, unfertilized, little or no signs of disin-
tegration (fig. 6, pl. 13; fig. 19, pl. 14).

No. 287’. Jan. 18, 6 p.m. (interval nearly 33 days); 13 clear, hyaline
eggs, disintegration evident in ovum.

No. 290. Copulation during night of Jan. 11 to 12; motile two-
headed spermatozoa recovered from vagina A.M., Jan. 12; Jan. 17,
8:45 p.m., 5 eggs: one unfertilized; 2 with small thick-walled vesicles
at one pole, abnormal (compare No. 290 (8), fig. 2, pl. 6); 2 eggs with
normal bilaminar vesicles occupying about one-half of the egg (compare
290 (4), fig. 2, pl. 6; fig. 6, pl. 12).

No. 290’. Killed Jan. 18, 4:30 p.m. (interval 192 hours); 8 eggs:
2 unfertilized; one retarded blastocyst; 5 apparently normal bilaminar

DEVELOPMENT OF THE OPOSSUM 13

blastocysts a little over 1 mm. in diameter; no preparations made of
this litter (figs. 2 and 3, pl. 11*).

No. 292. Caught with male in hollow log Jan. 13; isolated till
operation, Jan. 17, 11 p.m; 10 eggs: of the 4 that were sectioned one
is a defective 7-celled egg, the others normal blastocysts of 40 to 50
cells with none or only one entodermal mother cell (fig. 5, pl. 1*; figs. 1
Deore Tand.G, ple i6):

No. 292’. Killed Jan..21, 10:40 p.m. (interval 4 days); 7 vesicles
about 3 mm. in diameter, late primitive-streak stage; one unfertilized
ege@; one degenerating young bilaminar blastocyst (fig. 6, pl. 1*).

No. 293. Caught Jan. 17; Jan. 18, 8:00 p.m., 13 eggs: of the eight
preparations made all are 4-celled except one 2-celled egg; in one case
one blastomere, in two cases 2 blastomeres are in mitosis (fig. 1, pl. 2*;
figs. 5 and 6, pl. 15).

No. 293’. Killed Jan. 22, 7:30 a.m. (interval 33 days); 17 eggs: 8
unfertilized, one defective; 8 bilaminar blastocysts like those sketched
insiext meure. 3, , KH, (igs 2. pl: 2° )2 9 =

No. 294. Caught Jan. 17; large skin wound on belly; Jan. 18, 8:30
p.M., 15 eggs: 5 unfertilized, 8 with small rounded or irregular blasto-
cysts at one pole, all rather abnormal; 2 apparently normal bilaminar
blastocysts like F, text fig. 3 (fig. 1, pl. 11*).

No. 294’. Killed Jan. 20, 7 a.m. (interval 343 hours); 14 eggs: 7
unfertilized; one 1.4 mm. in diameter and 4 small degenerating blasto-
cysts; 2 bilaminar blastocysts about 1.3 mm. in diameter, of which
only one is perfectly normal. Thus both litters, 294 and 294’, were
mostly abnormal (fig. 4, pl. 11*). No preparations were made of this
litter.

No. 298. First copulation Jan. 14, spermatozoa recovered from
vagina; at 1 p.m., Jan. 15, the double spermatozoa had mostly divided.
Jan. 20, 10: 15 a.m., 14 eggs, of which perhaps 10 are normal blastocysts;
five preparations contain 60 to 120 cells each, showing earliest ento-
dermal proliferation (fig. 7, pl. 2*; figs. 7 and 8, pl. 6; figs. 6 and 20,
ple 16):

No. 298’. Killed Jan. 23, 12 m. (interval about 33 days); 6 eggs: 4
vesicles 4.25 and 4.9 mm. in diameter with medullary groove as long
as primitive streak; 2 smaller vesicles and 2 unfertilized eggs (eggs in
utero, fig. 8, pl. 2*).

No. 299. Caught Jan. 19; Jan. 20, 8:15 p.m., 12 eggs, of which all
of the 7 sectioned are normal 4-celled eggs with small blastomeres
(ie 3) pli hess Grandetepl 3, ties. 13 and 14, pl. 15).

No. 299’. Killed 11:30 p.m., Jan. 24 (interval 4 days, 3% hours); 14
eggs: 6 apparently normal, nearly or quite completed bilaminar blas-
tocysts; 3 abnormal blastocysts; 5 unfertilized eggs (fig. 4, pl. 1*;
figs. 1 and 2, pl. 6; fig. 5, pl. 10; figs. 7 and 8, pl. 12; fig. 16, pl. 13).

No. 303. Caught Jan. 19; Jan. 20, pseudopregnant, 7 degenerating
eggs a week old (fig. 7, pl. 11*); killed Feb. 1, the mammary glands
still very thick, almost as in pregnancy.

14 CARL G. HARTMAN

No. 306. Jan. 21, 12 m.; 11 eggs recorded in notes, but only 3 found
in collection; two of these are 2-celled with both blastomeres in mitosis
(text fig. 4, A to D; fig. 4, pl. 3; fig. 2, pl. 15); one egg is 3-celled (text
fig. 4, K; fig. 4, pl. 15).

No. 306’. Killed Jan. 26, 8:30 a.m. (interval 5 days, 20% hours);
10 eggs: 2 unfertilized; 8 bilaminar blastocysts 0.7 to 0.75 mm. in
diameter, of which one has no embryonic area (fig. 7*, pl. 10; fig. 17,
pl. 13; figs. 2 and 2A, pl. 20; figs. 1 and 1A, pl. 21).

No. 307. Jan. 21, 3:30 p.m.; 11 eggs removed from left Fallopian
tube (fig. 7, pl. 1*; figs. 2 to 6, 9, and 10, pl. 14).

No. 307’. Killed Jan. 27, 9:15 a.m. (interval 52 days); 10 eggs,
unfertilized and fragmenting (fig. 8, pl. 1*).

No. 3138. Caught Jan. 19; Jan. 22, 10: 30 p.m., 9 tubal ova, with
considerable albumen (figs. 1* and 3, pl. 3; figs. 12 and 18, pl. 14).

No. 313’. Animal died during the night; the 11 eggs taken from
remaining oviduct had more albumen than the ‘313’ litter; eggs
poorly fixed (fig. 5, pl. 13).

No. 314. Copulation a.m., Jan. 20; spermatozoa recovered; Jan. 23,
7:30 p.m., 9 eggs (fig. 1*, pl. 5; fig. 1, pl. 6); of the 5 eggs sectioned
one is a blastocyst of 26 cells (fig. 2, pl. 5), 2 contain about 30 cells
each (fig. 4, pl. 6), and 2 are abnormal (fig. 10, pl. 21).

No. 314’. Killed Jan. 29, 10 a.m. (interval 5 days, 14% hours); 6
normal embryos with first rudiment of allantois.

No. 318. Jan. 23, 26 eggs in early stage of fragmentation, 13 from
each uterus, in which involution had already set in (fig. 6, pl. 11*).

No. 320. Received about Jan. 20; Jan. 24, 9 p.m., 13 4-celled eggs
studied in salt solution; subsequent fixation poor (fig. 1, pl. 1*).

No. 320’. Jan. 30, 9:25 a.m. (interval 53 days); 17 eggs: 4 unfer-
tilized; 11 vesicles 2.3 to 2.6 mm. in diameter with well-developed
primitive streak; one egg contains two vesicles and two embryos; two
vesicles have no embryonic area (fig. 2, pl. 1*).

No. 321’. Jan. 25; litter of foetuses near term accompanied by the
«cee shown in fig. 9, pl. 11*; these eggs are, therefore, nearly 10 days

No. 332. Jan. 26, 21 eggs (9 plus 12), degenerating, unfertilized, in
middle stage of pseudopregnancy (fig. 10, pl. 11*).

_ No. 336. Jan. 27, 9 p.m., 14 eggs (one was lost before photograph-

ing); the 6 preparations made are young blastocysts of 17, 26, 29, 30,

mi a 32 cells, respectively (all figures on plate 4*; figs. 18 and 19,
15

No. 336’. Killed Feb. 1, 5:45 p.m. (interval nearly 5 days); s
10-mm. vesicles with small embryos; 2 smaller vesicles.

No. 337. Jan. 28, 10:30 a.m., 8 eggs: a study of them in salt solution
seemed to show that one egg was unfertilized, one 8-celled and six
16-celled; two eggs sectioned are 15- and 16-celled, respectively (fig.
9; pl. 1*; figs. 3* and/4*% ple peal Gp. 5),

No. 337’. Feb. 1, 10:30 p.m. (interval 44 days); 14 eggs: 12 blas-
tocysts, about 2.5 to 4.5 mm., first appearance of medullary groove;
2 defective blastocysts (eggs in utero, fig. 10, pl. 1*).

DEVELOPMENT OF THE OPOSSUM lex

No. 339. Jan. 28, 3:30 p.m., 8 eggs (fig. 6*, pl. 9): 2 unfertilized;
5 eggs with more or less abnormal, round, thick-walled blastocysts at
one pole (fig. 2, pl. 6; fig. 15, pl. 18; figs. 5, 5A, and 6, pl. 19); one quite
normal thin-walled blastocyst with entoderm spread to equator (fig. 2,
pl. 6; figs. 6 and 6A, pl. 18).

No. 339’. Killed 12:30 p.m., Jan. 29 (interval 21 hours); 9 eggs: 6
bilaminar blastocysts measuring about 0.85 mm. in alcohol; one dead
blastocyst 0.75 mm.; 2 unfertilized eggs (fig. 3, pl. 21).

No. 342. Received Jan. 27; Jan. 28, 9:30 p.mM., 19 eggs; of the 4
sectioned specimens two are defective and the other two are blastocysts
of 26 and 28 cells, respectively (figs. 6* and 7*, pl. 5; fig. 20, pl. 15).

No. 342’. Feb. 4, 8:15 p.m. (interval 7 days); 2 dead and 9 normal
embryos, the latter about 7.5 mm., head-rump length.

No. 343. Observed copulation, 4 a.M., Jan. 22; Jan. 29, 2:45 P.M.,
left uterus yielded 15 eggs: 4 unfertilized; one small defective blasto-
cyst; one blastocyst with defective embryonic area; 9 normal bilaminar
blastocysts about 1 mm. in diameter, embryonic areas 0.64 to 0.76 mm.
Gi ple Zoaties o. pl. 21):

No. 343’. Killed 7 hours, 20 minutes later (73? days after copula-
tion); 8 eggs, of which 3 are unfertilized, 5 normal 1.8-mm.blastocysts
‘just preceding proliferation of mesoderm; embryonic areas 1 to 1.1mm.
in diameter (fig. 6, pl. 2*; fig. 2, pl. 22).

No. 344. Received and operated Jan. 29, 4:45 p.m.; 16 eggs: 15
sectioned; of these 6 are unfertilized and fragmenting; 2 are abnormal
blastocysts (fig. 4, pl. 16); 7 are normal blastocysts showing early
differentiation of embryonic and non-embryonic areas; the most ad-
vanced contains 124 ‘ectodermal’ and 45 entodermal cells (figs. 5*, 6%,
and27, pl. Si fies. 14 tor l7, pl. 16).

No. 344’. Killed Feb. 1, 8:30 p.m. (interval 3 days, 32 hours); 7
eggs, all normal vesicles 4 mm. or more in diameter, with short
medullary groove.

No. 346. Received and operated Jan. 29, 8:45 p.M.; 21 eggs: 8
unfertilized; one dead blastocyst; 8 normal 1.5 mm. blastocysts,
embryonic areas about 1 mm.; the other eggs retarded and defective
(igs; ply 25):

No. 346’. Killed next morning at 6:35 o’clock (interval 9$ hours);
22 eggs: 11 unfertilized; 11 blastocysts ranging up to 2.2 mm. in
te all in early primitive-streak stages (fig. 4, pl. 2*; fig. 22,
pl. 13).

No. 347. Jan. 29, 9:45 p.m.; 15 eggs: 4 unfertilized; 11 normal
blastocysts partly or entirely bilaminar (fig. 5*, pl. 9; figs. 5 and 5A, 7
and 7A, pl. 18; figs. 3, 8, and 8A, pl. 19).

No. 347’. Jan. 30, 10:15 p.m. (interval 123 hours); 17 eggs: 4 un-
fertilized; 13 bilaminar blastocysts measuring 1.1 to 1.24 mm. in
alcohol (fig. 1, pl. 22).

No. 349. Front foot wounded in trap; Jan. 30, 3:45 p.m., 5 eggs: 2
unfertilized; one unilaminar blastocyst (fig. 4, pl. 8); 2 blastocysts
with spreading entoderm (fig. 3*, pl. 8; fig. 12, pl. 17).

JOURNAL OF MORPHOLOGY, VOL. 32, No. 1

16 : CARL G. HARTMAN

No. 349’. Killed Feb. 2, 11 p.m. (interval 3 days, 7 hours); 10 eggs:
9 vesicles 8 to 10 mm. in diameter with embryos of about 10 somites;
one small vesicle.

No. 351. Jan. 30, 5 p.M., animal was opened: freshly burst follicles
on left ovary, but no eggs in left oviduct.

No. 351’. Killed at 7:30 p.m., 23 hours later; 14 eggs with a little
albumen found in right oviduct (fig. 2*, pl. 3; fig. 4, pl. 18).

No. 352. Jan. 30, 5 p.m.; 16 eggs: 9 unfertilized; of the 7 remaining,
3 are like eggs No. 40, the others less advanced and perhaps not quite
normal (figs. 1* and 2, pl. 9; fig. 14, pl. 13; fig. 2, pl. 18; fig. 8, pl. 21).

No. 352’. Killed Jan. 31, 8 a.m. (interval 15 hours); 20 eggs: 8 un-
fertilized, one egg with dead vesicle; one egg with two vesicles; the
remainder bilaminar vesicles fill one-half to three-quarters of the egg,
which measured fresh about 0.75 mm. (fig. 4*, pl. 9; fig. 4, pl. 10;
fig. 4, pl. 19).

No. 358. Jan. 30, 7:45 p.m., 16 eggs: 5 unfertilized; 11 bilaminar blas-
tocysts measuring in alcohol 1.1 to 1.8 mm. (figs, 3, 3A, 3B, 3C, pl. 22).

No. 353’. Killed Jan. 31, 1 a.m. (interval 5+ hours); 12 eggs: one
unfertilized; one dead; 10 blastocysts 2 mm. and less in diameter,
showing the proliferation of first few mesodermal cells (fig. 21, pl. 18).

No. 356. Had been in cages some time before first operation, Jan. 30,°
8:45 p.m.; 15 eggs: 10 normal blastocysts averaging about 0.18 mm.
through ovum, with numerous entodermal mother cells at formative
pole and considerable attenuation of non-formative pole; one ab-
normal blastocyst with large blastomere (fig. 14, pl. 22); 4 unfertilized
eggs (fig. 1, pl. 6; figs. 1* and 2, pl. 8; fig. 3, pl. 9; fig. 4, a 12; fig. 12,
pl. 13; figs. Ato ipl. 17).

No. 356’. Killed Wah, 3, 12:30 a.m. (interval 3 days, 33 hours); 6
vesicles about 3 mm. in diameter, with short medullary groove.

No. 360. Jan. 30, 9:30 p.m., 11 eggs: 10 bilaminar blastocysts
about 1.5 mm. in diameter; one unfertilized egg (fig. 6 and stereogram,
fig. 8, ply 10: figsG6, pl. 22).

No. 360’. Killed Feb. 2, 7:30 p.m. (interval nearly 3 days); 2 ab-
normal and 18 normal embryos about’ 5.75 mm. in length with first
rudiment of allantois.

No. 415. Feb. 10, 11 fragmenting eggs in early stage of pseudo-
pregnancy, presented for the false ‘2-celled’ and ‘4- celled’ eggs seen in
figure 5, plate 11*.

3. Material arranged according to stage of, development. The
following tabulation is arranged by stages for ready reference.
Within a given stage the litters are also placed in ascending
order of development.

1. Ripe ovarian eggs: 21, 28.

2: Pubal-ova: S6.76ne07, 301") slo roto

DEVELOPMENT OF THE OPOSSUM Mf

3. Undivided, unfertilized uterine eggs showing little or no
degeneration: 58, 173, 214, 287.

4, Cleavage stages:

a. From one to about four cells: 46, 52, 54, 203, 306, 293,
81, 83, 299, 320.

b. From about 8 to 16 cells: 85, 117’, 337, 342.

5. Young unilaminar blastocysts:

a. Containing from 25 to 35 cells, mostly without entodermal
mother cells: 336, 173’, 191, 198, 203’, 314.

b. Older stages up to 100 cells, mostly with entodermal
mother cells: 50, 83, 298, 292, 88.

6. Young blastocysts with distinct polar differentiation: 344,
144’, 356, 349.

7. Young blastocysts with spreading entoderm: 194’, 339, 352,
43, 294, 175’, 347.

8. The eerie stage: 347, 285, 299’, 205, 208, 290, 293,
43, 306’, 352’, 82, 285’, 290’, 294! 189, 191’, 192, 343, 339), 94,
55, 347, 353; 346, 360, 193’, 3437, 1897, 353! (few mesodenn
cells).

9. Primitive-streak stages: 353’, 346’, 320’, 192’, 193’, 205’,
208’, 337’, 344’, 356’, 292’, 298’.

10. Embryos: 349’, 336’, 314’, 342’, 360’, 321’.

11. Unfertilized and degenerating eggs: 112, 415, 175, 194,
anh) 282300 214", 318,303), 297,832, 7321".

1. Securing the eggs. During the dolleenne season 1916 and
1917 two stages were secured from each female after the method
first employed by Bischoff on the rabbit. As the method has
proved of great value to the writer in securing a complete series
of stages, it is here discussed in some detail.

The female is placed under anesthesia and one uterus is re-
moved under aseptic conditions; the animal recovers and the
eggs are allowed to ‘incubate’ in the remaining uterus for a calcu-
lated period of time. In this way, by utilizing gradually ac-
cumulating data on the rate of development, it became possible
to secure almost any desired stage and thus fill in the gaps still
appearing in the series. Thus, for example, I succeeded in
securing from animal No. 353 eggs in which the mesoderm was

18 CARL G. HARTMAN

just beginning to proliferate. Animal No. 343 had previously
furnished bilaminar blastocysts (fig. 5, pl. 2) from the left
uterus; she was killed seven hours and twenty minutes later and
a litter of large blastocysts, still in the bilaminar stage, was
removed (fig. 6, pl. 2): the interval allowed had been too short.
Animal No. 346 (figs. 3 and 4, pl. 2) had yielded bilaminar blas-
tocysts a little larger than No. 343 and an interval of nine hours
and forty-five minutes had proved to be too long, for, when
the animal was killed, the primitive streak was already well
advanced in the second litter of ‘eggs. Profiting by these two
experiments, when animal No. 353 appeared with large bilaminar
blastocysts about the size of those in litter No. 346, a five and
one-fourth hour interval proved to be the correct one, for the
eggs in litter No. 353’ contain the first anlage of the primitive
streak, one egg having as few as twenty-five mesodermal cells.

In the operations I have found it most convenient to enter
the abdomen through a short slit on one side of the pouch.
For the sake of uniformity I select the left side as a matter of
routine. The animal is shaved over this area and the incision
is made as near the pouch as possible, care being taken not to
cut through the pouch, especially in multiparae, which possess
dilated pouches. The operated animal is bandaged; but it is
impossible to keep the bandage on an animal unless the entire
trunk is covered. I use over the bandage a jacket with holes
cut for head and legs and tied over the back. As the animal
usually sweats with the bandages on, the wound will heal better
if they are removed at the end of three or four days.

If, on opening the animal, the uteri are purplish and flaccid,
the case is one of pseudopregancy and the organs may be left
intact and the animal kept for another oestrus period, which
takes place in about thirty days. If ovulation is recent, however,
one uterus must be removed to ascertain the state of the eggs.
If the appearance of the organs indicates that young stages are
to be expected, the uterus is placed in warm Ringer’s solution
and a slit is made through the musculature and peritoneum
from one end to the other, and this must be done by a rapid
manipulation of the scissors to precent eversion of the mucosa.

DEVELOPMENT OF THE OPOSSUM 19

The pressure now being removed, the hypertrophied mucosa is
pulled apart, preferably under the binocular microscope, with
two pairs of finely pointed forceps, and the lumen exposed.
The eggs may be picked out from among the delicate folds of
the mucosa by means of a pipette. But this method is un-
necessarily tedious; the uterus may instead be simply turned inside
out in the Ringer’s solution and the eggs picked out from the
bottom of the dish. To insure finding all of the eggs, a little
Bouin’s fluid added to the salt solution, after removal of all the
eggs that can be seen, makes any specimens overlooked promi-
nently visible. The uterus should also be shaken out in another |
dish of Ringer’s solution for any eggs that may have been hidden
in the uterine folds. To keep the solution clear of blood, it is
well, before opening the organ, to slit all the superficial blood-
vessels and drain them of blood. I may add that the neck of
the uterus should be ligated with a ‘lifting’ ligature before it is
cut from the body, in order to prevent the loss of eggs through
the os uteri.

Young eggs in cleavage and small blastocysts are mostly
found near the caudal end of the uterus, often closely bunched
together. Hence one cannot speak of ‘implantation’ of the
opossum egg at any early stage. The ‘uterine cups’ described by
Spurgeon and Brooks (’16) do not mark implantation surfaces,
but merely accidental pits produced by pressure into the delicate
oedematous mucosa.

If, on opening the animal, pregnancy seems to be advanced,
in order to remove entire vesicles intact, it is best to slit the
uterus superficially in many places and to trim off the entire
musculature before attempting to remove the vesicles, which
are closely applied to the mucosa, but never fused with it.
This procedure renders the use of a killing fluid to paralyze the
musculature entirely superfluous. With a pair of forceps and a
fine brush an entire litter of delicate vesicles may be removed
intact. They may be transferred to the fixing fluid in a deep
mustard spoon or in a shallow, neckless vial. A collapsed
vesicle may again be dilated in the fixing fluid by injection with
a fine pipette; in fact it is well to irrigate with the fixing fluid the
lumen of every vesicle containing a large embryo.

20 CARL G. HARTMAN

Eggs are easily washed out of the Fallopian tube by means
of a stream of Ringer’s solution, as has been done in other
mammals. .

5. Fixing and staining. I have used the following solutions:
Bouin’s, Bouin’s half strength, increased graduaily to full
strength; Hill’s; Flemming’s; Carnoy’s; Zenker’s; formol-
Zenker; picro-sulphuric; trichloracetic; Bensley’s aceto-osmic-
bichromate. MHill’s fluid is made as follows: Mayer’s picro-
nitric, 96 cc.; 1 per cent osmic, 2 cc.; glacial acetic, 2 cc. I
stated in 1916: ‘‘I have found Hill’s mixture to be the perfect
fixing liquid for the opossum egg.”’ Further experience with it
has led me to give decided preference to Bouin’s for all older
blastocysts; for younger eggs up to the bilaminar stage I get
equally good fixation with both; and I also have made some
poor preparations with either. For all stages Bouin’s is perhaps
the safest solution to use; with it the specimens have the ad-
vantage of toughness and they can be safely transported,
whereas solutions containing osmic acid render the specimens
unduly brittle. The half-strength Bouin is not as good as full-
strength. I have some excellent preparations of material fixed
in Flemming’s fluid, although collapse of blastocysts is more
likely to occur in this fluid than in Bouin’s. My poorest fixation
was with aceto-osmic-bichromate, although superficially the eggs
thus fixed seem well preserved. ‘This fluid has the advantage of
bringing out cell membranes clearly. I have no perfect speci-
mens fixed in Zenker or formol-Zenker, both of which shrink the
material more than any other and render it very brittle. Several
fairly good preparations were made with picro-sulphuric. ‘Tri-
chloracetic has the peculiar property of fusing extruded yolk and
cytoplasm of the blastomeres into an almost undifferentiated
mass (lig. 13; plesiey):

Hematoxylin stains, especially Heidenhain’s iron-alum hema-
toxylin, have proved entirely satisfactory, both for sections and
for surface mounts. Several eggs fixed in solutions containing
osmic acid were stained in acid fuchsin, saffranin, or cochineal
to differentiate the nuclei clearly from the black yolk granules,
but this refinement of technique is not at all necessary.

DEVELOPMENT OF THE OPOSSUM PA

To prevent collapse of the blastocysts, which my photographs
on plates 1 and 2 show to be perfect spheres, it is important to
pass from one medium to another (Bouin’s to alcohol, water to
alcohol, aleohol to xylol, and especially xylol to paraffin) by slow
gradations. I use 5 per cent differences, accurately made up in
stock solutions. The eggs are placed in small vials and each
higher percentage is added gradually. Vesicles from about 1.5
mm. on up may easily be cut in half equatorially with fine
scissors; but such hemispheres, if of approximately equal size,
should not be placed in the same vial, because they are likely
to telescope in such a way that they are hard to separate without
injury to them.

The material was imbeded in paraffin and most of the newer
material was sectioned by Huber’s water-on-the knife method,
which gives incomparably better results than ribboning the
series with the rotary microtome.

6. Unfertilized eggs. Since, unfortunately, the vast majority
of the eggs removed from animals in captivity are unfertilized,
it will not be amiss to mention them here. Such eggs remain
in the uterus for many days, undergoing progressive degeneration
before being discharged or absorbed. For the first two or three
days they are not easily distinguished from normal uterine eggs.
The first sign of degeneration is the breaking up of the ovum
into masses of various shapes and sizes (fig. 5, pl. 11) and the
ovum may flatten out into the shape of a crescent (fig. 6, pl. 11).
Gradually, too, the eggs increase in opacity (fig. 7, pl. 11) and
become covered with white concretions (fig. 8, pl. 1; figs. 8
to 10, pl. 11), so that they are only too prominent when one
opens the uterus hoping to find embryos. The eggs shown in
figure 9, plate 11, accompanied foetuses near term; hence these
eggs are at least nine days old.

7. The illustrations. The drawings (plates 12 to 22) were
made on Ross stipple board No. 2 and reduced to one-sixth the
original size. Korn’s lithocrayon No. 1 (a_paraffin-carbon
pencil) gave the best results, since to be reproduced by the
line process the dots must be absolutely black, a result not
easily attained with a graphite pencil.

Ze, CARL G. HARTMAN

The drawings on plate 12 were made free hand by Mr. T. H.
Bleakney from stained specimens in oil of wintergreen, the
size being calculated from photographs of the living eggs. The
eggs shown in text figure 4 were mostly drawn from wax models
made after the Born method, x 600. This figure was also
drawn on Ross board No. 2, but was reproduced one-third of
the original size. All other drawings, except a few where
especially mentioned, were made from sections and were drawn
as nearly like the specimens as possible, imperfections and all.
The attempt was made to reproduce not only the form, but
also the texture of the specimens, and for this purpose the Ross
board has proved to be a delicate and responsive medium.

The smaller drawings were outlined with the camera lucida at
a magnification of x 300, x 1200, x 3000 (reduced to x 50,
x 200, x 500); the larger sections of blastocysts, since they had
to be drawn at a magnification of 1200 which resulted in draw-
ings more than a meter long, were first sketched x 400 with
a Leitz-Edinger drawing apparatus, then photographed and the
negative finally projected to the desired magnification by the
Edinger apparatus. To facilitate measurements, the scale of a
stage micrometer was drawn beside the first sketch made and
appears upon the negative made from it.

The photomicrographs were made with Spencer lenses, which
afford a flatter field for photographing than the Zeiss microscope.
lenses. Attention is especially directed to the photographs of
living eggs. More than 500 different eggs are represented in
the heliotypes. All of the photographs on plates 1, 2, and 11
and many other figures are magnified eight times and some are
at a higher magnification.

The photographs are unretouched and are reproduced as
exactly like the original as was possible with the process
employed.

To secure an absolutely black background for the photographs
taken by reflected light, we found it best to use a black watch-
glass. The eggs are removed from the uterus and placed in a
deep watch crystal in clear Ringer’s solution free from dirt and
blood-cells. The watch erystal is now set into the watch-glass

DEVELOPMENT OF THE OPOSSUM 23

which must also be filled with the solution; for it is absolutely
essential that there should be no air space for the reflection of
light between the transparent glass holding the eggs and the
black glass serving as a background.

In the photographs of preparations, as well as in the drawings,

for ease of comparison, magnifications of 50, 200, and 500 have
been adhered to with few exceptions. In this connection plates
12 and 13 are especially adapted to serve as a résumé of the
stages covered in this paper. Comparison of these two plates
shows that the young eggs shrink greatly on account of their
delicate albumen layer.

Altogether the twenty-two plates accompanying this paper
contain over 750 representations of more than 600 different
opossum eggs, mostly, of course, in groups on the photographs.
The drawings and the photomicrographs of preparations number
some 240 of over 180 different eggs. .

d. External changes at ovulation in the female opossum

In common with many other wild animals, the opossum does
not breed well in captivity. I have worked with hundreds of
animals kept in cages or in large rooms, isolated or in groups of
dozens or of a hundred or more; yet the number of observed
copulations that I have to record is disappointingly small.
Many births have taken place in the cages, but the cases of
,pseudopregnancy outnumber the cases of true pregnancy many
times over.

In spite of careful personal attention to the habits of the
captive animals, I was unable during the first two years’ col-
lecting to determine from outward signs the sexual state of the
female. In this regard I was at first forced to agree with
Selenka who says: ‘‘Ohne operative Eingriffe ist ttber die Trach-
tigkeit eines Weibchens keine Gewissheit zu erlangen, da man
weder durch Tasten mit dem Finger die weichen Uterushérner
auffiden kann, noch auch an den Milchdriisen eine Verainderung
wahrnimmt, bevor nicht die Embryonen nahezu ausgewachsen
sind.”’ I have since learned, however, that Selenka was wrong

24 CARL G. HARTMAN

in his statement concerning the mammary glands. For, during
the 1916 season, I found that by simple palpation of the mam-
mary glands within the pouch I was able to diagnose with a
high degree of accuracy the state of the imternal reproductive
organs, so responsive are the glands to the physiological changes
going on in the animal just before and after oestrus. By this
method one is enabled to select from the animals on hand those
that are likely to furnish eggs or embryos. Thus out of the
hundred animals Nos. 300 to 400, used at the height of the
breeding in 1917, only a half-dozen failures are recorded. A
typical case of misjudgment is that of No. 326, in which ‘5-mm.
vesicles’ were predicted and ripe follicles found in the ovary;
or No. 354 in which ‘bilaminar blastocysts’ were expected and
the animal was found in pro-oestrus. Sometimes a later stage
than the one predicted will be found, as is, of course, to be
expected from individual variations that are general in all
. physiological processes. The method has resulted in the saving
of a great deal of time, effort, and material, especially during
the last breeding season.

Unfortunately, however, the physiological changes which the
mammary and other reproductive organs undergo are identical
immediately after oestrus whether pregnancy ensues or not.
This holds true for the mammary glands more than for the other
organs, and it is impossible during the first five or six days to
distinguish externally between pseudopregnancy and pregnancy.
As ovulation is always spontaneous, the internal organs behave
the same in both conditions. The vaginal loops begin to retro-
gress even before ovulation. The uteri are almost maximum in
size when the minute eggs first reach them; in pregnancy they
remain bright red and turgid and possess a peculiar luster like
polished red agate; but, if the eggs are unfertilized, the uteri,
after four or five days, become dull and dark red and then
flaccid and collapsed. The corpora lutea are somewhat more
persistent in true pregnancy. But the mammary glands con-
tinue development even after the degeneration of the corpora.
and the involution of the uteri are well under way.

bo
Or

DEVELOPMENT OF THE OPOSSUM

e. Are the eggs of operated animals normal?

The question may well be asked whether we are dealing with
normal material in the case of eggs removed some time after an
abdominal operation under anesthesia or whether such treat-
ment of the mother affects the development of the eggs un-
favorably. It should be emphasized at the outset, however,
that whatever answer we give to the question does not affect the
conclusions reached in this study, which is supported by an
abundance of material from freshly killed animals and by a
large assortment of specimens removed from animals at the first
operation. It should also be noted that the time interval
between the two operations was in many cases only a few hours
or a half-day, so that in this material, too, the chances of modi-
fying the normal course of development of the eggs were reduced
to aminimum. From a careful examination of my notes and a
scrutiny of both classes of material I have concluded that there
is no evidence pointing to deleterious effects of the operation,
and I here present some of the facts that have led me to this
conclusion.

In the first place, the condition of the operated animals was
as good or better than that of non-operated cage animals; for
the former were the choice specimens, vigorous in health and
sexually active. As stated above, I am now able to determine,
with a high degree of accuracy, the near approach of oestrus in
the female opossum. A surprisingly large number of females
are captured (and by the terms of our contract with the hunters
must be purchased) which are either too old or too sick to breed.
Specimens with deep, infected wounds, intestinal diseases,
xerophthalmia (McCollum), or other nutritional disturbances do
not come into heat. Only once or twice have I seen females
with badly infected wounds continue in the oestrus cycle like a
normal animal; but I have records of dozens of cases in which
the normal sexual processes were interrupted by wounds or dis-
-ease during pro-oestrus or dioestrus. Pregnant females, how-
ever, pass successfully through the period of gestation even in

26 CARL G. HARTMAN

the same condition that prevents the onset of oestrus. On the
other hand, operated animals recover quickly, often eating
heartily several hours after the operation. Their wounds heal
readily and the animal comes into heat again even after two
operations and after double hysterectomy, in the same manner
as if the abdomen had not been opened. Certainly, if Bischoff
a century ago was able to secure as many as six different stages
of normal embryos from one rabbit, without anesthesia or
asepsis, successively opening the abdomen and ligating off
segments of the uterine horns ‘‘until inflammation set in,” then.
a very simple operation on the opossum under modern surgical
precautions should have no deleterious influence on the embryos.

If, now, to test the matter further, we compare the proportion
of normal’ eggs secured at the first operation (table 1 above)
with the proportion from the second operation, we find 63.1
per cent normal (item 3, table 1) for the former and for the
latter 67.4 per cent (items 4 and 5, table 1), an unexpectedly
but quite accidentally large percentage of normality for the
operated animals.

These figures, however, include under ‘abnormal’ all unfer-
tilized eggs, which should be left out of consideration, since we
are here testing the effect on the development of the eggs and
embryos. We must, therefore, count only the dead and de-
fective fertilized eggs in given litters, selecting comparable
. stages. Table 2 gives this data for bilaminar vesicles of litters
in which every egg was studied; and cases from the second
operation are selected in which at least one day had intervened
after the first operation. Table 3 gives similar data for primi-
tive-streak stages up to 5 mm. in diameter. The litters are
arranged more or less in order of relative stage of advancement.

From a study of tables 2 and 3 it is apparent that there is a
high rate of mortality in the eggs of the opossum, both of
operated and unoperated animals, but that wholly normal
litters occur in both classes. Some cases are of special interest.
No. 334, for example, yielded a perfect litter of eggs from the .
left uterus, but only a single abnormal vesicle 20 hours later
from the right uterus. On the other hand, No. 344 yielded 7

DEVELOPMENT OF THE OPOSSUM DY}

normal and 2 abnormal eggs of an early stage and after an
interval of three days a perfect litter of 7 normal vesicles.
Litter No. 339 from the left uterus is largely abnormal, whereas
No. 339’ from the second operation was largely normal. Both
batches of eggs from animal No. 294 were mostly abnormal,
possibly on account of a temporary interference with the circu-
lation of the uteri when the animal was twisted out of its lair
TABLE 2
Number of abnormal eggs at the bilaminar stage

TOTAL NUMBER
OF FERTILIZED
EGGS

NUMBER OF
NORMAL EGGS

NUMBER OF

LITTER NUMBER ABNORMAL EGGS

INTERVAL

a. At first operation

days
339 0 1 5 6
205 0 8 1 9
343 0 9 2 11
346 0 8 5 13
353 0 11 0 11
360 0 10 1 11
Mortal east A 47 14 61
23%
b. At second operation
299’ 4 6 3 9
293’ 33 8 1 9
339/ 1 6 1 a
306’ 43 a 1 8
191’ 3t 4 0 4
Motallee ance 33 6 37
10.8%

after a method employed by hunters and applied in this case.
Litters No. 175’ and No. 194’, twenty-five and twenty-eight days,
respectively, after the removal of the left uterus, contained no
abnormal eggs.

Later embryos and foetuses present facts similar to those just
indicated for the younger stages. To refer to special cases
shown in table 4, No. 349, which had furnished only 3 normal

28 CARL G. HARTMAN

eggs out of 5 from the left uterus, had 9 large, normal embryos
and one dead embryo in the right uterus three and one-half
days later, and this in spite of the fact that the animal was
unusually small and had one of its legs wounded in a trap. But
of the six large embryos yielded by the left uterus of No. 379,
an apparently normal female, only one was normal; but 18

TABLE 3

Number of abnormal eggs at primitive-streak stages

TOTAL NUMBER
OF FERTILIZED
EGGS

NUMBER OF
‘NORMAL EGGS

NUMBER OF

LITTER NUMBER ABNORMAL EGGS

INTERVAL

a. At first operation

180 0 5 0 5
338 0 14 4 18
380 0 8 2 10
334 0 11 1 11
211 0 10 1 11
Dil 123 hours 8 0 8
MOG alles arg eto 56 63
11%
b. At second operation
days
208/ 1 7 1 8
292’ 4 6 1 a
320’ 5A tat 2 13
Sole 4t 10 , 12
344! 3 a. 0 Th
356/ 3 6 0 6
298/ 34 3 3 6
180’ 1 11 1 12
334’ 1 0 1 1
hoGaleeeeeeaec 61 11 71
15.3%

hours later all of the 6 embryos in the other uterus proved to

be normal. ;
As such cases could be multiplied, the facts are that there is a

mortality in the opossum ovum at all stages and that the death

DEVELOPMENT OF THE OPOSSUM 29

rate is not affected by the abdominal operations such as em-
ployed in our experiments.

TABLE 4

Number of abnormal embryos of later stages

= NUMBER OF TOTAL NUMBER
LITTER NUMBER INTERVAL lea ae peau cee eet ies

207 0 11 1 i
207’ 2? days 5) 7 10
226 (left) 0 3 4 7
226 (right) 0 0 all

291 0 10 0 10
291’ + day 7 0 if
379 0 2 4 6
379’ 18 hours 6 0 6
314’ 53 days 6 0 6
336/ 5 days 6 2 8
342’ 7 days 9 2 11
349’ 33 days 9 1 10
360’ 3 days 18 2 20

MATURATION AND CLEAVAGE TO THE FORMATION OF THE
BLASTOCYST

a. The ripe ovarian egg

Since the publication of my former paper I have not seen the
first maturation spindle, for the ova of all large follicles thus
far studied in numerous series of ovaries have either germinal
vesicles in the resting stage just preceding maturation or have
already given off the first polar body. Recently collected
ovaries containing large follicles are now being prepared and
will be discussed in connection with a paper on the corpus
luteum.

The ripe ovarian egg is broadly elliptical in form, but may
be nearly spherical, as some dissected specimens indicate.
Measurements were previously stated to average 0.165 x 0.135
mm. or larger than any tubal or uterine ova. This is large in
comparison with Futherian ova, but small in comparison with

30 CARL G. HARTMAN

the egg of Dasyurus, which measures 0.21 x 0.126 to 0.27 x 0.26
mm. The ova shown in figure 1, plate 13, and in figure 1,
plate 14, are unusually large, even for ovarian eggs, measuring
0.183 x 0.156 mm. (average 0.175 mm.) and 0.185 x 0.15 (average
0.167) mm., respectively. That the ovarian ova are on the
average larger than the tubal or the uterine ova would seem to
be the case from the few measurements of ovarian eggs that
have been made. There is, of course, considerable variation in
the sizes of different eggs, both of the same litter and of different
litters (fig. 2, pl. 3, figs. 2 to 6, pl. 13).

The ova are surrounded by a well-defined zona pellucida,
within which the polar body is found. This is given off usually
at one of the ends of the somewhat elongated egg (fig. 1, pl. 13),
but it may be found near the equator. The polar body is small
and flattened, and contains chromatin matter and a minimum of
cytoplasm. The chromosomes of the egg nucleus lie in the
cytoplasm near the polar body, mostly more or less discrete
and arranged in an equatorial plate. In this condition the egg
reaches and traverses the Fallopian tube.

The ovarian egg is, therefore, essentially like the tubal ovum
presently to be described in greater detail. There is no polar
differentiation recognizable except for the location of the polar
body. It is important to note also that the yolk has no tendency
to accumulate at one pole of the egg, as is so strikingly the case
in the mature ovum of Dasyurus and to a slight degree in
certain Eutherian eggs (bat, armadillo). Herein lies the first
striking difference between the eggs of Didelphys and the
Australian Dasyurus.

b. The tubal ovum

1. Material, ovulation, secretion of albumen and shell membrane.
Eggs were removed from the Fallopian tubes of five animals,
but in no case was insemination observed. Unfortunately,
none of the thirty or more eggs sectioned contains any trace of
a spermatozoon. This stage has, therefore, not yet been
observed in the case of any marsupial.

DEVELOPMENT OF THE OPOSSUM 541 |

Eggs were found in both tubes of female No. 56; and, since
both batches of eggs had practically no albumen deposited on
their surface, they must have been discharged simultaneously
from both ovaries a short time before their removal. None of
these eggs possesses any granulosa cells nor was any semblance
of a ‘corona radiata’ ever observed on any tubal ova. Their
naked condition when discharged is positive evidence that
Selenka (87) was in error when he considered the shell mem-
brane of uterine eggs as the modified granulosa cells—an
error made despite Caldwell’s correct interpretation of this
structure.

Litter No. 76 was taken from one Fallopian tube only, the oppo-
site one not yet having received the ova, which had, however,
been discharged from the ovary on that side, as indicated by
the presence upon it of fresh corpora lutea. I infer that the
eggs must have been lost in the body cavity. Since the eggs
secured from the right uterus had already been provided with a
small quantity of albumen, one may assume that they had an- °
ticipated the eggs from the other ovary by a short space of
time.

No. 351 had ovulated at 5 p.m., when the animal was first
opened, but no eggs were found in the left Fallopian tube, they
having also been lost in the handling of the organs. Two and
a half hours later the right tube contained eggs which had a
distinct albumen layer on all sides (fig. 2, pl. 3). If this repre-
sent the amount of albumen deposited in two and a half hours,
it would require twenty-four hours for the eggs to traverse the
Fallopian tube and receive their entire quantum of albumen.
Professor Hill thinks that the eggs of Dasyurus pass through the
tube very quickly, since he has never found a whole litter of
eggs in the tube. But the uterine eggs of Dasyurus are very
scantily provided with albumen; in fact, never relatively more
than the incomplete deposit around the eggs of my litter No. 351
(hoa pl: 3).

Eggs No. 307 (fig. 7, pl. 1) show a greater deposit of albumen
on one side of the egg. Since this is true to some extent in all
the eggs found in the upper part of the tube, and since, later,

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 1

ae CARL G. HARTMAN

the albumen is of about the same thickness on all sides (fig. 1,
pl. 3), the eggs are probably rolled about slowly as they pass -
through the Fallopian tube.

The shell membrane is doubtless secreted and added to the
surface of the albumen in the lower part of the oviduct. In-
semination must of necessity take place soon after the eggs
enter the tube before albumen is deposited; for spermatozoa are
found in some eggs throughout the albumen and most often
nearest the ovum. The eggs of litters Nos. 336 and 356 have
enormous numbers of spermatozoa entangled among the lamellae
of the albumen; in figure 12, plate 13, for example, the sper-
matozoa are seen to occur in thick clusters as well as scattered
singly throughout the albumen.

Usually an ovum is necessary to afford’ the stimulus for the
secretion of the albumen; but in one case a rounded mass of
epithelial cells proved adequate, and there was produced a
structure without an ovum, the cell mass replacing the latter
- in the center of the egg. Epithelial cells from the wall of the
Fallopian tube, enmeshed within the albumen, are of common
occurrence. In another case an ovum and a cell mass and in
a third case two ova were included within the same egg en-
velopes. Both of these latter ova would be likely to develop, if
one may judge from the cases of double ova in the blastocyst
stage shown in figure 2, plate 1, and in figure 4, plate 9. An
egg of a parasitic roundworm once found among tubal ova did
not seem to afford the adequate stimulus for the secretion of
albumen.

2. Size and shape. The ova vary greatly in size and shape,
not only among the different litters, but also among the eggs of
a single litter. They are elliptical, rarely spherical in shape,
as may be seen from the figures in plates 3 and 14. The average
size of thirty-one preparations on the slide is 0.122 x 0.104 (av.
0.113) mm. Twelve eggs of litter No. 56 average 0.128 x 0.109
(av. 0.118) mm.; this list includes two whole mounts which are
nearly round and measure 0.131 mm. (fig. 7, pl. 14) and 0.135
mm., respectively, the latter being the largest tubal ovum in
the collection. The maximum length of elliptical eggs of this

DEVELOPMENT OF THE OPOSSUM 33

batch is 0.142 mm., the maximum width 0.120 mm. The
average of eight eggs of batch No. 76 is 0.134 x 0.113 (av. 0.123)
mm.; the maximum length is 0.148 and maximum width 0.125
mm. But this batch includes also two very small eggs measur-
ing 0.100 x 0.085 (av. 0.093) mm. and 0.090 x 0.083 (av. 0.087)
mm., the latter being the smallest egg in the collection. Three
eggs of batch No. 307 average 0.126 x 0.116 (av. 0.121) mm. in
the preparations; the size of the fresh eggs of this batch, shown
in figure 7, plate 1, cannot be given because the magnification
of the photograph is not known. The two living eggs of batch
No. 313 shown in figure 1, plate 3 measure 0.119 x 0.105 (av.
0.112) mm. and 0.106, respectively; the average of eleven eggs
of batch No. 313’, as photographed in the living state, is 0.108 x
0.099 (av. 0.103). In the preparations, three eggs of batch
No. 313 measure 0.105 x 0.091 (av. 0.098) mm., indicating some
shrinkage of the eggs in the histological processes. Batch
No. 351’ (fig. 2, pl. 3) average 0.110 x 0.096 (av. 0.104) mm. in
the living state; three preparations from this batch measure 15
per cent less, namely, 0.102 x 0.076 (av. 0.089).

3. The distribution of yolk. The opossum egg, in common
with the eggs of other marsupials, is rich in yolk or other lipoid
deposit, which partly accounts for their larger size (figures on
pl. 14). The fat occurs in the form of granules or spherules,
many or perhaps all of which stain black with osmic acid. Eggs
fixed in Bouin’s fluid show numerous vacuoles from which the
fat is dissolved in clearing. The fat content of the eggs renders
them much less transparent; but in the living state the globules
may be seen in the egg and they also appear distinct near their
outer limits of distribution in the photographic negatives. taken
by transmitted light in salt solution. Thus the negative from
which figure 1, plate 3, was made shows in detail oil globules -
quite similar in distribution to those shown in figures 12 and 18,
plate 14. A study of the fresh eggs and photographs of them
convinces me that the fixed and sectioned specimens accurately
show the true details of these eggs, little altered by the histo-
logical processes.

34 CARL G. HARTMAN

Three more or less distinct regions are typically recognizable
in many of the ova (fig. 12, pl. 14). There is a marginal zone,
sometimes very narrow, consisting of granular cytoplasm,
nearly devoid of fat granules. Beneath this is a more or less
diffuse zone of oil globules, which may be very small or very
large or medium in size, as seen from the figures on plate 14.
Some litters show considerable uniformity in this respect (No.
76) and in others there is variation within the litter (No. 313,
figs. 12 and 18, pl. 14). The outer surface of this zone of fat
globules is often marked by a delicate region which may break
down in fixation (light zone in figs. 12 and 19), reminding one
of the delicate deutoplasmic pole of the egg of Dasyurus. In
the living opossum egg this region is a light band interrupted
here and there with oil globules coming near the surface. The
third region is the large central portion of the egg which is
rather uniformly granular and contains few oil globules or
vacuoles.

The tubal ovum, like the ovarian ovum, exhibits no polar
concentration of yolk, which is in striking contrast to the un-
segmented ovum of Dasyurus, in which the deutoplasm is
gathered in a mass at the vegetative pole of the egg and is
bodily extruded just prior to the first cleavage; whereas in the
opossum the yolk is thrown out from both ends or from all sides
in greater or less amounts, as the sequel will show.

4. The polar body. The first polar body, which is present in
all ripe ovarian eggs and in all tubal ova, lies in a spindle-shaped
depression under the vitelline membrane. It is never large,
containing a modicum of cytoplasm, in contrast to the prodigality
with which yolk and cytoplasm are eliminated from the egg in
cleavage. The polar body is usually of such a peculiar color
and homogeneous texture that it is easily recognizable in eggs
fixed in Bouin’s fluid; but if the fixing fluid contain osmic acid
the polar body is seldom recognizable. The chromatin is
usually a deeply staining irregular mass.

Both polar bodies are soon absorbed, disappearing as dis-
tinctive structures in early cleavage. Except for a slight differ-
ence in size, the two polar bodies are practically identical. I

DEVELOPMENT OF THE OPOSSUM Y9)

have seen them in numerous eggs in cleavage, especially in
4-celled eggs. The oldest stage which contains two objects
that I take to be polar bodies is a blastocyst of 34 cells. The
polar bodies are caught between two blastomeres of the vesicle
(fig. 1, pl. 16). The larger of the two is shaped like a bent
spindle and resembles in outline the space which it occupied
while still crowded in the usual periovarial space before cleavage.
I have never seen polar bodies so large that they appear in
cross-section like those figured by Spurgeon and Brooks (’16).

5. The chromosomes. The spindle for the second maturation
division is formed soon after the giving off of the first polar body,
and in this condition the egg reaches the Fallopian tube. The
vesicular or resting stage does not seem to intervene between
the two maturation processes. Insemination was not observed.
In the absence of spermatozoa, the ovum reaches the uterus
unchanged, except for the accession of the egg envelopes. Thus
the young uterine eggs Nos. 58, 287, and others have chromo-
somes practically indistinguishable from those about to be
described for tubal ova.

Three preparations from batch No. 307 (fig. 7, pl. 1; figures
on pl. 14) are especially favorable for a study of the chromosomes
and for determining their number. These are still scattered
along the clearly defined spindles, the equatorial plate being
delayed in its formation. Some of the spindle fibers are thick
and beaded as though they were derived from the fibers of the
preceding division. One spindle is contained in a single section
(fig. 6, pl. 14). There are clearly twelve chromosomes in each
of these eggs. In all other tubal ova the chromosomes are
closely arranged in a more or less definite equatorial plate and
are difficult to count; but I am sure that the number is twelve
in five or six cases, and I can count at least ten or eleven chromo-
somes with distinctness in all cases. Hence I am prepared to
state that twelve is the reduced number of chromosomes in the
egg of the opossum.

Figures 11 and 14, plate 14, represent the usual appearance
of the chromosomes in these specimens, and in these two cases
twelve chromosomes can be clearly made out. Figure 13 shows

36 CARL G. HARTMAN

a side view of a spindle in which eight chromosomes are seen and
short fibers are clearly outlined. The three sections shown in
figures 15 to 17, plate 14, were cut tangentially through the egg,
hence the polar body is cut longitudinally and the chromosomes
are seen in polar view.

The chromosomes in every case are short and thick, never
characteristically rod-shaped. Some are hollow squares with
rounded corners, others more perfectly ring-shaped. In side
view several appear as short, thick rods, slightly constricted in
the middle; others bent or cupped so as to appear narrowly
kidney-shaped. The spindle is usually situated in a granular
area free of vacuoles or fat globules; or, in other words, in the
region of the spindle, the central and the marginal granular
regions are bridged across. The polar body is usually placed
near the chromosomes, as seen in the figures. In one case one
chromosome was extruded with the polar body (fig. 8, pl. 14).

c. The young uterine egg

1. Size and shape. The appearance of young uterine eggs is
well illustrated by the photographs in plates 1, 2, 5, 11, and
others, which represent them with fidelity just as they were
removed from the uterus. In size the eggs, as measured through
the shell membrane, are subject to considerable variation among
the different litters as well as to some extent within a given
litter. Thus litter No. 342, consisting of about the 26-celled
stage, average 0.7 mm., whereas the 4-celled eggs of litter
No. 293 average 0.57 mm. and the eggs of litter No. 292, which
are young vesicles of some 100 cells, measure 0.55 mm. Again,
litters No. 336 and 337 average 0.73 and 0.50 mm., respectively,
although they are in nearly the same stage of advancement and
the ovum proper is of about the same size in the two litters.
The differences in diameter among the eggs is therefore a differ-
ence in the quantity of albumen deposited about the ovum.
Figure 1, plate 12, represents an average unsegmented uterine
ovum.

DEVELOPMENT OF THE OPOSSUM Ot

2. The albumen. The albumen is laid down in delicate, con-
centric lamellae around the ovum (pl. 13 and others). In the
living state it is usually of nearly the same density throughout
the layer (fig. 2, pl. 4) or it may be more concentrated about the
ovum in young eggs (figs. 3 and 4, pl. 5).

At first the albumen is extremely poor in protein content,
for on fixation it usually gathers immediately about the ovum
and the thin and delicate shell membrane collapses and follows
close upon the albumen (fig. 5, pl. 5). This phenomenon is
apparent on comparing plate 12, which represents the living
condition, with plate 138, on which corresponding stages are
shown from sections on the slide. The shrinkage of the ovum
proper or the whole egg in later stages is comparatively slight;
only the albumen of the younger egg suffers great collapse. It
follows, therefore, that the albumen layer gradually increases its
protein content (figs. 14 and 17, pl. 13), and the shell membrane
likewise grows in thickness and resistance. The uterine ‘milk’
doubtless supplies the material thus absorbed. This holds true
for unfertilized eggs also, which continue to grow in diameter
and in density of albumen and shell membrane for a week or
more. Figures on the thickness of the shell membrane have
previously been given (Hartman, ’16) and are not repeated here.
It is subject to great variation, as may be seen from the various
drawings in the plates, where the shell membrane is represented
in correct proportions. |

3. The unsegmented ovum. Unless insemination has taken
place, the uterine differs from the tubal ovum only in the pos-
session of completed albumen and shel envelopes (fig. 19,
pl. 14). My collection contains a number of litters of such eggs.
The first polar body and the second maturation spindle are as
in the tubal ova, although in some case the chromosomes begin
to show a clumping and are surrounded by a light area. The
chromosomes fragment sooner or later, however, and the chro-
matin breaks up and rearranges itself into round lumps simu-
lating nuclei in the resting stage. The cytoplasm breaks up
also, some fragments taking one or several ‘nuclei,’ others none.
Sometimes the fragments are equal or nearly equal in size, so

Bey CARL G. HARTMAN

that such eggs may easily be taken for cleavage stages (fig. 5,
pl. 11). I have never seen a mitotic figure in such degenerating
eggs. The eggs shown in the photograph in figure 8, plate 1,
remained in the uterus until near the end of the sixth day after
ovulation.

4. The pronuclear stage. Several eggs were studied at this
stage, although I did not secure an entire litter of eggs con-
taining pronuclei. The pronuclei at first lie at the periphery
of the ovum in a homogeneous granular area devoid of fat
globules. Eventually they come to lie near the center of the
egg where the first cleavage spindle will form. The chromatin
of the pronucleus is in most cases very diffuse and stains weakly.
Two figures of the stage are given in the writer’s former paper
and figure 20, plate 14, is an ovum with two nuclei which differ
from the nuclei of the other members of this litter (2-, 3-, and
4-celled eggs); hence I regard this egg as in the pronuclear
stage.

d. The first cleavage

1. The first cleavage spindle. No new material containing the
frst cleavage spindle has been obtained recently. In figure 21,
plate 14, I have redrawn specimen 52 (3) as a composite of four
sections, which were taken obliquely through the spindle, which
lies in the central yolk-free zone of the egg. The fat vacuoles
are evenly distributed at the poles and some of the yolk has
already been extruded, chiefly on one side of the egg.

2. The 2-celled stage. The early cleavage material in the
writer’s collection was, however, considerably increased by
recent accessions, namely, from the following litters: No. 173,
which furnished 3- and 4-celled eggs; No. 203, which furnished
four 2-celled, one 3-celled, and the rest 4-celled eggs; and No. 306,
which furnished, besides a number of eggs that were unfor-
tunately lost or misplaced, two 2-celled and one 3-celled egg.
All of these litters are the product of the left uterus and in each
case a later stage was removed from the right uterus, which is
good evidence that we are here dealing with normal fertilized
material.

DEVELOPMENT OF THE OPOSSUM 39

Of the 2- and 3-celled eggs models were prepared and draw-
ings made from the models, which are shown in text figure 4.1,
J and P of this figure were drawn from eggs mounted in toto in
balsam.

The blastomeres of the 2-celled eggs are usually flattened as if
by mutual pressure upon their contact surfaces. They may be
of equal size and shape and practically identical, or they may
be unequal, as the drawings in the figure amply show. If they
differ in size I rather believe this difference to be secondary and
not to unequal cleavage, that is, to the greater amount of yolk
extruded from the smaller blastomere. Thus, in egg No. 203
(13), shown in I, text figure 4, one blastomere has given off a
large mass of yolk at each end, but the aggregate of the masses
in the two halves of the egg is as nearly equal as in the adjoining
fgure of a sister egg.

A study of the eggs in serial section fails to reveal either a
qualitative difference between the two blastomeres or the
slightest indication of polarity within the blastomeres them-
selves (fig. 4, pl. 3; figs. 1 and 2, pl. 15). The yolk granules
occur in equal numbers and sizes at the two poles and the nuclei
are centrally placed. The distribution of the yolk granules is
indeed quite similar to that of the undivided egg, namely, in a
zone toward the margin of the cell (figs. 2 to 4, pl. 15), and this
holds true also for the blastomeres of the 16-celled stage and
even later (fig. 17, pl. 15). In no ease is it possible to distinguish
a more deutoplastic ‘vegetative’ pole and a relatively yolk-free
‘animal’ pole in any stage of segmentation.

There would seem, however, to be a qualitative difference in
the blastomeres, as evidenced by the more precocious division of
one of them in the formation of the 3-celled stage, which in
later divisions leads to the 6- and the 12-celled conditions.

e. The second cleavage

1. The 3-celled stage. The 3-celled egg differs from the
2-celled stage only in the more rapid division of one of the
blastomeres (figs. 3 and 4, pl. 15). In all the 3-celled eggs the

40 CARL G. HARTMAN

large blastomere has two nuclei (text fig. 4, K, L, M) and in
one egg the cytoplasmic division is initiated, as indicated by a
constriction around the cell. The interesting point in these
eggs lies in the position of the blastomeres to one another,
especially in eggs Nos. 306 (3) and 173 (8); for the lines joining
sister nuclei are almost absolutely at right angles to each other.
The usual position of the blastomeres of the 4-celled stage
(text fig. 4 N to P; figures on pl. 15) is, therefore, already
anticipated in the 3-celled egg. The shifting of the blastomeres
in egg No. 203 (3); shown in text figure 4, M, is rather along
the original plane of the 2-celled stage, and such an egg might
develop a 4-celled egg like that shown at O, whereas an egg like
No. 173 (8), shown at L, would be sure to develop into the
typical ovum with cross-shaped blastomeres, as in figures 5
and 6, plate 15.

2. The 4-celled stage. If the number of specimens which the
collector happens to secure of a given stage be any criterion of
the relative length of time which the egg remains in that stage,
then according to my collection the 4-celled condition of the
opossum egg is not passed very quickly. For I have more than
five dozen, mostly excellent preparations of this stage, and have
other eggs still unsectioned. Three whole litters (Nos. 293, 299,
320) furnished only 4-celled eggs so far as these have been
studied. However, this preponderance of 4-celled eggs is prob-
ably quite accidental. Inasmuch as cleavage proceeds irregu-
larly after the 4-celled stage, it would not be fair to compare the
number of 4-celled with the number of 8-celled eggs, for ex-
ample, for a litter preponderatingly 8-celled would be sure to
contain 6-, 7-, and perhaps 10- and 12-celled eggs also (compare
Nos. 85, 117, 342).

The 4-celled egg of the opossum is typically Eutherian in the
cross-shaped arrangement of the blastomeres. This is quite
evident from the figures presented, some of which are drawn
from models, others from in toto preparations and from sections
(text fig. 4 and pl. 15). The arrangement: of the blastomeres is
such that no section can possibly pass through the centers of all
the four blastomeres. If the imbedded egg be so oriented that

DEVELOPMENT OF THE OPOSSUM 4]

a section cuts two blastomeres the other two have a chance to
be similarly cut (figs. 6 and 7, pl. 3; figs. 11 and 12, pl. 15).
Sometimes three blastomeres are found in one section and a
single one in another section (figs. 9 and 10, pl. 15). In figure 5,
plate 3, and figures 7 and 8, plate 15, the knife passed through
the centers of two blastomeres, the top of a third, and the
bottom of the fourth. I have also studied 4-celled eggs in the
living state under strong illumination and have clearly seen that
the crossed arrangement of the blastomeres, sometimes with slight
deviations from 180°, is normal for the opossum egg.

The four blastomeres of any one egg are usually of the same
size; hence one can seldom differentiate a pair of large and a
pair of small cells, and I have searched in vain for any other
‘trace of polarity in these eggs aside from that afforded by the
occasional presence of the polar bodies which, with the shifting
of the cells, has little meaning (figs. 11 and 12, pl. 15). More-
over, the blastomeres are always spherical, except when very
large, in which case they are flattened on contact surfaces by
mutual pressure (figs. 7 and 9, pl. 15). The entire ovum
measures through the zona the same as the undivided tubal or
uterine egg. Among the various litters of eggs there is, however,
a remarkable variation in the relative size of the blastomeres,
which depends upon the amount of yolk extruded. The egg
represented in figure 7, plate 15, has a minimum of eliminated
yolk and the largest blastomeres; figure 14 represents the other
extreme; figure 11 the intermediate condition. The extent of
yolk elimination would seem to be hereditary, for in each batch
of eggs the blastomeres of the individual eggs are approximately
of the same size; thus, in No. 203 they are all very large, in No.
299 all extremely small. Both types are, however, normal, for
sister ova in the right uterus in each case were allowed to develop
and produced normal blastocysts.

The eliminated yolk in the 4-celled eggs seems to be char-
‘acteristic of this stage. It occurs in small rounded lumps of
about equal size, uniformly distributed (figs. 5, 6, and 7, pl. 3;
figs. (eel 12 and 4 pl. 45).

42 CARL G. HARTMAN

f. The origin of the crossed arrangement of the first four
blastomeres

Since in the Eutherian ova there is very little yolk to be
eliminated, even in cases, such as the bat, where the phenome-
non has been described by Van der Stricht, the blastomeres of
the 4-celled stage fill the space within the vitelline membrane
rather snugly. It has therefore been suggested that mutual
pressure is responsible for the shifting of the blastomeres and
that in the crossed arrangement they occupy the minimum space
in the egg. A glance at the specimen photographed in figures 6
and 7, plate 3, will convince one, however, that this mechanical
explanation is inadequte, for certainly here one cannot speak
of mutual pressure of the blastomeres, for they are not even in
contact, and yet in such eggs the shifting also takes place.
Hence we must look for other causes of the shifting movement.

It is, of course, quite possible that there is no shifting at all,
but that the cleavage planes cut the two blastomeres of the
2-celled egg at right angles, as has been suggested by Professor
Hill (10, p. 31). According to this assumption, one of the
blastomeres would be divided meridionally, the other equa-
torially, and the crossed arrangement would obtain from the
beginning. Indeed, a study of the 3-celled eggs described above
would seem corroborative of this view, for here the definitive
arrangement has already manifested itself. But two facts make
this theory untenable. First, a number of 4-celled eggs and
one 3-celled egg I find to deviate less than 180° from the parallel
arrangement; hence for these one would under the theory have
to postulate a backward shifting toward the parallel postion.

But conclusive evidence on the point is furnished by eggs
Nos. 306 (1) and 306 (2), in each of which both blastomeres are
in mitosis. In the latter the spindles are exactly parallel. as
shown by lines connecting their ends in D, text figure 4. In
the former egg (A and B) the spindles in the blastomeres are
36° removed from the parallel. These observations seem to
indicate that division begins in both blastomeres of the 2-celled

DEVELOPMENT OF THE OPOSSUM 43

Fig. 4. A and B, two views of egg No. 306 (1); C and D, two views of egg
No. 306 (2); in both cases both cells are in mitosis and the lines run through the
ends of the spindles (compare fig. 2, pl. 15, and fig. 4, pl. 3). Ev and F, two views
of egg No. 203 (8) (compare fig. 1, pl. 15). G and H, two views of egg No. 203
(4) (compare fig. 7, pl. 13). I, No. 203 (13); one blastomere only has given off
a large amount of yolk. J, egg No. 203 (11). K, egg No. 306 (3) (compare fig.
A, pl. 15). L, egg No. 173 (8) (compare fig. 3, pl. 15). M, egg No. 203 (3). N,
No. 81 (6). O, No. 299 (7) (compare fig. 13, pl. 15). P, No. 46 (7). I, J, and P,
drawn from total preparations; all others from wax models.

44 CARL G. HARTMAN

egg in a single cleavage plane and that secondarily a shifting
sets in early in the process of division.

There is a third possibility, as described by Sobotta (’95)
for the mouse. According to this author, if I follow him
correctly, one of the two blastomeres of the 2-celled egg divides
meridionally, but the other blastomere has the division spindle
at right angles to the first cleavage plane. In other words, the
cleavage plane of the first blastomere to divide stands at right
angles to the first cleavage plane, whereas in the second blasto-
mere it is parallel to it. In such a case some shifting is also
necessary to bring about the typical crossed arrangement of the
4-celled egg. This method does not obtain in the opossum, as is
seen from my description above.

This point would appear to be further complicated by Spur-
geon and Brooks (’16), who describe and figure cleavage stages,
derived apparently from two female opossums. According to
these authors, the second cleavage plane passes through both
blastomeres equatorially and not meridionally, and thus a
fourth method is suggested. I would cheerfully accept the
authors’ conclusions, but for the fact that the eggs described
by them do not appear to me te represent normal fertilized
eggs. I believe their specimens to be fragmenting and unfer-
tilized eggs that have been in the uterus three or four days.
My reasons are as follows: 1) Cases of fragmenting eggs are
extremely common in cage animals and such eggs may fragment
into regular pieces resembling blastomeres of eggs in cleavage,
as I have seen repeatedly in hundreds of such eggs (compare my
photograph in fig. 5, pl. 11). 2) In their illustrations some of
the blastomeres have an additional peculiar nucleus and many of
the nuclei are very eccentric in position. Multinucleated ‘cells’
and those with nuclei placed at a distance from their centers are
quite characteristic of fragmenting eggs. 3) The ‘polar bodies’
represented are peculiar for their large area in cross-section and
for their position at a distance from the periphery of the egg.
4) The authors do not figure any of their 4-celled eggs, of which
they secured four along with other stages, although they present
drawings of six 2-celled and other eggs. 5) The photographs

DEVELOPMENT OF THE OPOSSUM 45

given by the authors in their figures 12, 13, and 14 I recognize
from my experience with hundreds like them as typical pictures
of degenerating eggs; for example, in the thickness of the shell,
which suffers little collapse in fixation; in the peculiar stringy,
not uniformly concentric character of the albumen, and in the
character of the ovum itself, where fragmentation is quite ap-
parent. 6) Finally, the size of the eggs as stated by the authors,
0.75 to 1.5 mm., is far above that of normal eggs in cleavage
and entirely in agreement with my own specimens of fragmenting
eggs. I must, therefore, conclude that the eggs described by
Spurgeon and Brooks do not represent normal cleavage in the
opossum.

I would conclude, therefore, that both blastomeres of the
2-celled opossum egg divide meridionally, but that they shift
their position during division so that the resulting 4-celled ovum
possesses the typical crossed arrangement. ;

g. Comparison of the 4-celled egg of the opossum and of Dasyurus

It is seen from the foregoing that the 4-celled stage of the
opossum is typically Eutherian, at least in the arrangement of
the blastomeres, and quite different in every recognizable way
from the egg of Dasyurus, in which, as described in Hill’s
beautiful monograph, the second cleavage is shown to be
meridional, dividing the egg into four equal cells which exhibit
the same polar differentiation as the 2-celled egg, for each
blastomere possesses a larger, vegetative pole and a smaller,
relatively yolk-free animal pole.

Precisely such an egg is described by Selenka (’87) for the
4-celled stage of the opossum. I can reafflrm my former state-
ment that this is a case of an unfertilized egg undergoing pseudo-
segmentation or amitotic fragmentation, in which the four
pieces or ‘blastomeres’ (pseudoblastomeres) happen to be of
equal size. I have seen such eggs dozens of times. Figure 5,
plate 11, is a photograph of a litter of eggs, palpably frag-
menting, but showing one ‘2-celled’ and one ‘4-celled’ stage
which might easily be mistaken for normal cleavage. ‘This

46 CARL G. HARTMAN

matter is again mentioned and the photograph presented as
further evidence that there is a decided difference between the
normal 4-celled egg in the opossum and that of Dasyurus. What
has been described by various authors as ‘parthenogenetic
cleavage’ in ovarian eggs of mammals may often be merely a
fragmentation process similar to that here described for the
opossum. I have also found just such fragmenting eggs in
atretic follicles of opossum ovaries.

h. Deutoplasmolysis or the elimination of yolk

In the eggs of both Dasyurus and the opossum the extrusion
of yolk proceeds in the manner that one might predict from the
distribution of the yolk in either case. In the ripe egg of
Dasyurus the deutoplasm is collected in a mass at one pole
where it is bodily extruded when the first two blastomeres round
up during the first cleavage. In the opossum the yolk, being
peripherally distributed, is given off from any or all sides. This
happens, in small amounts, as early as the pronuclear stage and
in larger amounts at the first cleavage. At each cleavage stage
some yolk is left within the blastomeres, and it is probable that
with each succeeding division of the blastomeres some additional
yolk masses are eliminated. There seems to be no regularity of
time in the elimination of the yolk, just as there is no regularity
in the relative amounts eliminated; but the greatest quantity
seems to be given off between the 2- and the 4-celled stage.
The blastomeres may be very large, and full of yolk or very
small and proportionally yolk-free; and, since considerable cyto-
plasm is thrown off with the yolk, this would seem to indicate
that a relatively unimportant role is played by the peripheral
cytoplasm in the normal processes of the cells. But the fate of
the yolk is in all cases the same: it is eventually digested and
resorbed, so that in the bilaminar stage only a few granules
occur among the cells of the embryonic area, as will be pointed
out later.

As the eliminated material contains both cytoplasm and yolk
granules, it would seem that whole portions of the cells are

DEVELOPMENT OF THE OPOSSUM 47

dropped bodily. The appearance of these cast-off masses in the
various stages may be seen from the drawings. With tri-
chloracetic fixation, blastomeres and yolk blend into an almost
uniform mass, so that the limits of the cells are recognizable
with difficulty (compare fig. 13, pl. 15, with fig. 14, eggs from
the same litter).

The yolk elimination in marsupials is, of course, striking in
that the mass involved is very large, and this is as one would
expect from the phylogenetic position of the group, as has been
so ably discussed by Professor Hill. This phenomenon has,
however, not entirely disappeared among the EKutheria, as Van
der Stricht’s fine study of the bat ovum amply proves. This
author has shown that there is a polar distribution of the yolk
in the bat egg and this undergoes elimination, a process called
deutoplasmolysis by the author. A similar condition is found
in the ovum:of the armadillo by Newman (’12), but this author’s
statement that the similarity in the distribution of deutoplasm
in the eggs of the armadillo and of Dasyurus argues for the low
phylogenetic position of the Edentata loses some of its force
from the fact that the egg of Didelphys, a marsupial, does not
exhibit a polar concentration of fat.

1. Later cleavage to the formaiion of the blastocyst

An extended description of the later cleavage of the opossum
egg was presented in my former article (Hartman, 716), to
which the reader is referred for details here omitted. The new
material collected in 1916 and 1917 contains eggs from 8 to 26,
28, and more cells; all corroborative of the former account.
These eggs were also carefully studied in the living state and
were photographed in salt solution at high and low magnifica-
tions, and the assurance may be given that the fixed and
sectioned material accurately represents the true morphological
relations. This is well borne out by the photographic repro-
ductions of living eggs and of sections made from them as shown
on plates 4 and 5.

JOURNAL OF MORPHOLOGY, VOL. 22, NO. l

48 CARL G. HARTMAN

The later cleavage is represented in the collection by eggs
with every number: of: blastomeres from the 4-celled stage in
which two blastomeres only are in mitosis (figs. 5 and 6, pl. 15)
to the fully formed blastocysts of about 32 to 36 cells. Cleavage
proceeds very irregularly after the 4-celled stage, which explains
the fact that the 8-celled and the 16-celled eggs are only slightly
in the plurality (compare litter No. 85). There is a retardation
in division of cells at one pole of the egg, presumably among the
lineal descendants of one of the first two blastomeres. In
models of 10- and 12-celled eggs the larger cells are grouped at
one pole, but, aside from this fact, there is nothing that would
point to a polar differentiation, and in the 16-celled stage even
this criterion is lost.

After the 4-celled stage is passed, the ovum of the opossum
behaves no longer as a typical Eutherian, but as a marsupial
ovum. In the former the blastomeres of the successive divisions
cling together to form a solid mass or ‘morula,’ which is soon
overgrown by a layer of cells, Rauber’s layer or the trophoblast.
The mass within is the ‘inner cell mass’ which gives rise to the
embryo and its envelopes. The blastocyst is formed by the
appearance of a cavity between the trophoblast and the inner
cell mass at the lower pole of the egg. |

In the marsupials the morula stage is absent. Already in
the 2- and 4-celled opossum eggs the space between the blasto-
meres represents, potentially, the blastocyst cavity, for at the
16-celled stage, or even earlier, the blastocyst cavity is clearly
indicated. As early as the 6-celled stage the blastomeres mani-
fest a tendency to migrate to the zona pellucida and to apply
themselves to the wall of the ovum (fig. 15, pl. 15). In 12-
and 15-celled eggs the blastomeres are usually well flattened out
at the periphery, as seen in figure 8, plate 3; figure 9, plate 13,
and figure 16, plate 15. At the 16-celled stage it is exceptional
to find rounded cells, and models of such eggs show the outer
surface of the blastomeres molded against the curvature of the
surrounding albumen (compare figs. 17 and 18, pl. 15).

It thus happens that the eliminated yolk comes to lie within
the cavity of the blastocyst, for the blastomeres migrate to their

DEVELOPMENT OF THE OPOSSUM 49

places against the wall of the ovum and here undergo further
division and further flattening until they come into mutual
contact and thus complete the blastocyst wall, leaving the yolk
within the cavity. Figure 19, plate 15, is a section through an
ovum of 26 cells; figure 20 through one of 28 cells. In both
cases there are gaps in the wall of the blastocyst, indicating
that this is not yet complete. ‘The same is true of two eggs of
30 and 32 cells, respectively, in which the gaps are fewer in
number (fig. 3, pl. 4). In figure 10, plate 13, and figure 1,
plate 14, are shown sister ova of 32 and 34 cells, respectively;
their walls are practically continuous and the blastocyst may be
considered complete. Occasionally more advanced blastocysts
still have gaps in their walls, as, for example, the one shown in
figure 6, plate 6, which has 46 cells. We may say, however,
that, on the average, the blastocyst wall is completed when the
32-celled stage is reached or soon thereafter. No polarity is
evident in the egg, the cells being of uniform size and structure
throughout. Not long after this the entoderm formation is
initiated.

Hence, in the opossum the blastocyst is completed at a much
earlier stage than in Dasyurus, where the blastomeres of the
16-celled egg are arranged in two superimposed rings at the
equator of the egg. To form the blastocyst wall they must
proliferate and migrate toward either pole, and the blastocyst
is not completed until the gaps at the two poles are closed. In
the opossum, on the contrary, to complete the blastocyst all
that is necessary is the closing of the gaps between the cells
which are early distributed more or less evenly at the periphery.
The just completed blastocyst of Dasyurus contains more than
three times the number of cells (90 to 130) than does the corre-
sponding stage of the opossum, and it is three times as large.

At this stage in the opossum, neither the ovum nor its en-
velopes have increased perceptibly in size (pl. 12). The albu-
men layer lies over the ovum as thickly as before, again in
striking contrast with the condition in Dasyurus, in which the
albumen layer is completely resorbed when the. egg has reached
the 16-celled stage. The opossum blastocyst is completed about

50 CARL G. HARTMAN

thirty hours after the beginning of cleavage; in one case (No. 314)
such eggs were found three and one-half days after copulation.

j. On the fate of the first two blastomeres

In the Eutherian ovum it seems probable that one of the
first two blastomeres is destined to form the inner cell mass, the
other the trophoblast, as was first pointed out by van Beneden
(75). If, then, Hill be correct in his interpretation of the
embryonic area of marsupials as being homologous with the
inner cell mass of Eutheria (a view in which I join), one might
suppose that the 2-celled stages in the two groups of mammals
are also homologous. But that this does not hold in the case

Fig. 5. To illustrate the probable fate of the two blastomeres of the 2-celled
egg. Polarity is indicated in B, D, and E by the more rapid cell division at
the upper pole. In F, a 16-celled egg, and G, one of 40 to 50 cells polar differ-

of Dasyurus seems clear from the scholarly work of Professor
Hill. In Dasyurus the most reasonable interpretation of the
facts is that the upper poles of the two blastomeres form the
embryonic area and the lower poles the non-embryonic area.
If this view be correct, then both blastomeres contribute equally
to the embryo and to the trophoblast, or, in other words, the
upper halves of the two first blastomeres of Dasyurus are
together homodynamous with an entire blastomere of the
Eutherian ovum, the lower halves homodynamous with the
other blastomere. There would seem, then, to be a fundamental
difference between the 2-celled Metatherian and the 2-celled
Eutherian ovum.

The question arises: Does the opossum ovum follow, in its
behavior, the egg of Dasyurus, with which the opossum is

DEVELOPMENT OF THE OPOSSUM 51

phylogenetically more closely related, or does it follow that of
the Eutherian ovum, to whose indeterminate type of cleavage
it is strikingly and unexpectedly similar?

I have previously taken the latter position, namely, that the
formative area very likely arises from one of the blastomeres,
as in the Eutheria. If one follow a series of models of opossum
eggs in successive stages, such as shown in text figure 5, A to H,
one may visualize the formation of the blastocyst. We may
safely assume that there are an upper and a lower pole in the
eggs A to E, as evidenced by the difference in rate of division,
aside from various other differences which may occur between
the first two blastomeres (in size, amount of yolk extruded, rate
of division). In the 12-celled egg there are eight smaller cells

entiation is lost, soon to be reestablished by the appearance of entoderm. It
seems not unreasonable that the upper pole of H is the product of one of the two
cells in A.

(2 x 4) and four larger cells (1 x 4), and it is evident that such
an egg arose by one division from the 6-celled stage. Polarity
is, therefore, indicated at least to this extent. The four undi-
vided cells may next divide, establishing the 16-celled stage, in
which polar differentiation is lost (F), not to be resumed again
until the entoderm begins to proliferate at about the 50- to
60-celled stage (between G and H in the figure).

It is, therefore, impossible to bridge over the brief gap between
the 16-celled stage, where polarity is lost, and the 50-celled
stage, where it is resumed, and all that may be said is that it
seems more reasonable to assume that all of the slowly dividing
cells are of one kind and have one destiny and that all of the
rapidly dividing cells are of another kind and have a different
destiny. One has the choice between this view and the alterna-

yy CARL G. HARTMAN

tive, that a part of each type of cell goes into the formative and
a part into the non-formative region.

Because of the short period in the opossum egg in which polar
differences are lost, it is, therefore, impossible to demonstrate
cell lineage in the cleavage of the opossum egg. But the same
statement may be made with reference to the egg of Dasyurus,
as pointed out by Hill himself, for, in the Dasyurus blastocyst,
polar differentiation is lost during the long period of growth
from 0.6 to 3.5mm. Hill says (’11, p. 46): “It might therefore
be supposed that the polarity, which is recognized in early blas-
tocysts, and which is dependent on the pronounced differences
existent between the cells of the upper and lower rings of the
16-celled stage, is of no fundamental importance, since it ap-
parently becomes lost at an early period during the growth of
the blastocyst. Such an assumption, however, would be very
wide of the mark. . . . . and, indeed, in view of the facts
set forth, is an altogether improbable one.” There is not the
slightest doubt that Professor Hill’s view is the reasonable and
the probable one. Upon the same grounds, the 2-celled opossum
egg is not homologous with the 2-celled egg of Dasyurus, but
rather with the 2-celled Eutherian egg.

Several abnormal opossum eggs are instructive in this con-
nection, for they are indicative of polar differentiation in young
blastocysts normally devoid of evidences of polarity. They are
abortive attempts on the part of one-half of the egg in each
case to form a blastocyst wall and they were found in litters
of eggs made up for the most part of normal young blastocysts.
Figure 3, plate 16, represents a section through an egg in which
one-half of the blastocyst, consists of twenty cells, which have
flattened normally, whereas the other half consists of eleven
cells which are still rounded as in an earlier cleavage. Similarly,
one-half of another egg (fig. 4, pl. 16) seemed to develop normally,
the other half containing cells with fragmenting nuclei. In still
another the normal half is beehive-shaped and surrounds two
very large and three small cells (fig. 10, pl. 21). Egg No. 356
(2) has two large blastomeres at one pole of the blastocyst
. (fig. 14, pl. 22) and egg No. 88 (6) has a retarded blastomere
enclosed within the blastocoele (fig. 18, pl. 22).

DEVELOPMENT OF THE OPOSSUM a5

These cases perhaps indicate that the cells at one pole of the
blastocyst are all of a distinct type, and it is not a far ery from
eggs of 32 cells to the 2-celled stage, nor is it an unreasonable
assumption, in view of the facts presented, to derive cells of each
type from one of the two blastomeres.

THE FORMATION OF THE ENTODERM
a. General

In his classical work on Dasyurus, Hill has described an
apparently new method of entoderm formation in mammals.
His account is specific and definite, for the entoderm may be
traced from certain unique cells which appear in the blastocyst
wall when the egg has attained a diameter of about 4 mm.
Within the embryonic area of such eggs a number of small
ectodermal cells become modified, leave the blastocyst wall, and
migrate to the inner surface to become the definitive entoderm.
A similar process was independently discovered in the armadillo
blastocyst and described in detail by Patterson (’13), who
showed conclusively that also in this Eutherian mammal the
entoderm forms not by delamination of cells on the surface of
the inner cell mass, but by migration of the cells from the
embryonic ectoderm of the monodermic vesicle.

Selenka, in his work on the opossum, naturally also speculated
upon the method of entoderm formation in this species. His
ideas were based upon one defective 8-celled egg and on two
blastocysts of 42 and 68 cells, respectively. He believed that
the lower half of the 8-celled egg consists of entodermal, the
upper half of ectodermal cells. Each of his two youngest blas-
tocysts has a large cell included within the blastocyst cavity,
and one of his figures is almost identical with my specimen
No. 314 (2), shown in figure 2, plate 5. This included cell, which
he calls ‘Urentodermzelle,’ Selenka believed to be a migrant
from the lips of the ‘blastopore’ at the ‘entodermal pole’ of the
egg.

In my previous publication I reported upon 44 normal young
unilaminar blastocysts, in 39 of which there occurred one or

54 CARL G. HARTMAN

more cells within the blastocyst cavity, as well as other enlarged
and modified cells still within the wall (pls. 7 and 16; compare
Hartman, ’16, p. 36). I conjectured that the free cells might
have arisen by accidental inclusion of a blastomere in about the
16-celled stage (compare fig. 17, pl. 15) or by proliferation from
the large cells within the blastocyst wall, since frequently a
number of cells would be united into a column projecting into
the cavity (figs. 3 and 8, pl. 6). These cells appeared to come
from various points in the blastocyst wall.

My next stage consisted of considerably advanced unilaminar
blastocysts (compare figs. 1 to 4, pl. 18), in which I found ento-
derm in various stages of differentiation, including certain few
cells that appeared to come out of the formative area of the
blastocyst in precisely the same manner described for Dasyurus
by Hill; and I figured cases in point.

With these two considerably separated stages before me, I
concluded that the entoderm arose, as in Dasyurus, after the
formative area had become well differentiated, and hence I
considered the included cells of the young stages as of ‘no
morphological importance.’

Since publishing my report on these young blastocysts, I have
been fortunate enough to collect an unbroken series of tran-
sitional stages between the just completed unilaminar blas-
tocyst and the just completed bilaminar stage, ‘and of especial
interest are litters Nos. 344, 356, 194’, and 349, of which I
possess numerous preparations (pls. 8, 9, 16, 17). I also have
more than five dozen additional young unilaminar eggs of the
stage previously described, so that I now have before me 100
such preparations, besides a considerable number which I did
not consider necessary to section. In these blastocysts I again
find the persistent occurrence of the peculiar included cells such
as previously described, which my new material now teaches are
the true entodermal mother cells of the opossum. What I had pre-
viously described as entoderm formation marks the end and
not the beginning of this process. The true entoderm formation
begins in blastocysts containing 50 to 60 cells within the blas-
tocyst wall; that is, these large modified cells in the blastocyst

DEVELOPMENT OF THE OPOSSUM 55

wall, which proliferate after becoming free, or even in situ,
constitute the first entoderm mother cells. This I am now
able to show from the study of a closely graded series of stages,
as abundantly illustrated by my drawings (pls. 16 to 18) as
well as by photographs of preparations and of living eEes
(pls. 6 to 9).

b. The youngest unilaminar blastocysts

It has been shown above that the blastocyst arises by the
early migration of the blastomeres to the periphery of the ovum,
where they flatten out against the zona pellucida orthe albumen
layer. By further division and spreading, the cells come into
mutual contact, obliterating the spaces between them. The
blastocyst is completed at the 32-celled stage or immediately
thereafter. At first there is no evidence of polarity in the
blastocyst, all of the cells being of the same structure and
thickness throughout.

c. The first entoderm mother cells

At about the 50- or 60-cell stage, on the average, certain cells
within the blastocyst wall undergo modification in situ. They
become larger jutting out more or less into the blastocyst cavity.
On their inner surface they may be rounded (HNT"%, figs. 6
and 7, pl. 16), or they may display an extended tip as if under-
going amoeboid movement (fig. 5, pl. 16). Some eggs show
this tendency only to a slight degree in one or several cells;
in others one or two cells will show more decided enlargement,
projecting as much as two-thirds of ‘the radius of the blastocyst
into the cavity (fig. 4, pl. 7). These cells are the first entoderm
mother cells in the opossum and can be traced in every gradua-
tion from earliest differentiation until they become detached
from their place in the wall. Most of these cells are to be
recognized only by their size and shape, since they have the
same staining reactions as other unmodified cells and they con-
tain apparently the same number of yolk granules. But if
they remain some time in the wall, they elongate greatly and’

=

56 CARL G. HARTMAN

take a much darker stain, as in figure 4, plate 7. This elongated
type of cell is common in the collection. If the attachment of
such a cell in the wall continues, it may give rise by cell division
to columns of three, four, or more cells, as numerous examples
serve to indicate (figs. 3 and 8, pl. 6, and figs. 20 and 21, pl. 16).

d. The detachment of entoderm mother cells

It more commonly happens, however, that the entoderm
mother cells leave their place in the wall soon after attaining
their maximum size, and their behavior at this time constitutes
perhaps the most remarkable phenomenon in the entire de-
velopment of the opossum egg. Their performance at this
stage is little short of spectacular. Such partly or wholly
detached cells are present in nearly every egg of litter No. 88,
which covers this critical period in the formation of the ento-
derm by a series of more than two dozen preparations, and
there are identical cells in numerous other excellent prepara-
tions from various litters. The cells, moreover, have such a char-.
acteristic appearance that I should term them the more typical
entoderm mother cell of the opossum.

After a period of growth the entodermal cell rounds up on all
sides. In this way its contour no longer conforms to the
curvature of the ovum, and as a result, the contact with the
adjoining cells is broken—the cell seems to roll out of its place,
as it were, into the blastocyst cavity. But the gap thus formed
does not long remain, for the vacant spaces are filled at once
by a flowing in of the surrounding cells. This is clearly seen at
A, figures 7 to 11, plate 16, which specimens were not selected
originally with this point primarily in view, but they.illustrate
the phenomenon without exception. The entoderm mother cells,
when they leave their place in the wall, do not, therefore;
leave gaps that may be called ‘blastopores’ (Selenka), and such
gaps as occur in earlier stages, with or without included free
cells, are due to a different cause, as was shown above. Some-
what more advanced stages, moreover, still proliferate cells of the
‘same type, as will appear below (pl. 17).

DEVELOPMENT OF THE OPOSSUM ai

The newly formed entoderm mother cells are sometimes found
in mitosis (fig. 10, pl. 16, and fig. 8, pl. 17); indeed, in egg No. 88
(11) six of the nine entoderm mother cells are in process of cell
division, although most of them have not yet left the blastocyst
wall (fig. 22, pl. 16).

It is thus apparent that the entoderm mother cells, found in
variable numbers within the blastocyst cavity, arise from cells
leaving the blastocyst wall and also as a result of their mul-
tiplication before, during, and after their migration into the
cavity. The specimens figured here as well as numerous others
afford ample evidence of these developmental processes.

The process in the opossum is essentially the same as obtains
in’ Dasyurus, for, at a given stage in both forms, certain cells
in the superficial unilaminar wall become modified and migrate
into the interior of the vesicle. In the opossum we have an
approach to the Eutheria in the early differentiation of the
entoderm; hence we may consider the Dasyurus as exemplifying
the more primitive, the opossum the more specialized condition.

e. Proliferation of entoderm confined to one pole

The small blastocysts of about 0.15 mm. in diameter referred
to above cannot be oriented for sectioning, and hence the plane
of the sections is entirely a matter of chance. It thus happens
that the sections taken tangentially or oblfquely through the
ovum present, in some cases, very confusing pictures; for in
such specimens the entodermal proliferation appears to take
place promiscuously from various parts of the egg, and the
polarity, which is very apparent in favorably cut series, is thus
obscured. In the former the entodermal proliferation is pal-
pably confined to one pole; to ascertain the arrangement in the
‘latter it is necessary to make idealized reconstructions in the
proper plane. ‘This I did from series of camera-lucida drawings.
Five such reconstructions are shown in figures 18 to 22, plate 16.
In every case, without exception, the entoderm proliferation is
confined to one pole, in some cases to exactly one-half of the
blastocyst. We may, therefore, now speak of embryonic and

58 CARL G. HARTMAN

non-embryonic areas, for there is no longer any doubt as to
their identity: the embryonic area is marked by the position of
the entoderm mother cells and the polarity of the ovum is
definitely reestablished.

A study of the young blastocysts just considered, as well as
immediately succeeding stages, seems to show, moreover, that
the first proliferation of entoderm takes place more actively on
the margin of the future embryonic area, for one often finds
them most numerous on opposite sides, as shown, for example,
in figures 11, 18, and 22, plate 16, and figures 8, 10, and 11,
plate 17.

The blastocyst now contains two types of cells: 1) those:
clearly entodermal in destiny, as just described, and 2) the
peripheral or enveloping layer, which, of course, gives rise to
all of the ectoderm, embryonic and trophoblastic. All of the
cells at the lower pole are ectodermal, being trophoblastic. But
the epithelial cells at the embryonic pole, since they will for
some time still continue to: proliferate entoderm mother cells,
are potentially both ectodermal and entodermal and should
better be called entectoderm until the entoderm is fully formed.
They are, however, also potentially mesoderm, as my next paper
will clearly show.

The ovum of the opossum has not grown much in volume since
its discharge from the ovary, being still less than 0.15 mm. in
diameter (pls. 12 and 13), which is in striking contrast with the
blastocyst of Dasyurus, where, at the appearance of the first
entoderm mother cells, the vesicle is nearly 4 mm. in diameter.
In the opossum the period of growth follows the formation of
the entoderm, but in Dasyurus this is preceded by a long period
of growth. The process of entoderm formation in Dasyurus
may, therefore, be studied from surface mounts of pieces easily
cut from the blastocyst wall, as well as from serial sections; but»
the opossum egg at this stage may be studied in section only,
for it is small, covered with a thick layer of albumen, and is
densely packed with more or less opaque yolk. The vesicular
structure can well be made out from in toto preparations, but
for detailed study such preparations are w6rthless.

DEVELOPMENT OF THE OPOSSUM 59

f. Time of appearance

From the foregoing it is apparent that size is as yet no criterion
to the differentiation among the blastomeres. The number of
cells seems, therefore, to be the best means of .establishing the
stage in question. With this in view, a careful count was made
of the number of cells in thirty-three flawless series. By making
camera-lucida drawings of each series sketching in the nuclei,
superimposing the successive drawings, and eliminating dupli-
cates, it is believed that the counts are quite accurate. This
data is presented in table 5.

This table shows that there is a rough correlation between
the number of cells and the extent of entoderm proliferation.
Extreme variations occur, however, as, for example, in the
sister eggs Nos. 298 (1) and 298 (3), which have four entodermal
mother cells each, but the latter totals twice as many cells as
the former (64 and 124, respectively). But it may be stated
in general terms that entoderm proliferation usually begins
when the blastocyst is made up of 50 or 60 cells.

It is thus apparent that in the early differentiation of ento-
derm the opossum again approaches more closely than does
Dasyurus to the condition in the Eutheria. Thus in the
absence of polar differentiation in the undivided egg (with due
consideration to certain exceptions as the bat and armadillo),
in the crossed arrangement of the blastomeres in the 4-celled
egg; in the more or less indeterminate type of cleavage; in the
early proliferation of entoderm—in all of these characters the
opossum egg resembles that of the Eutheria. But in the absence
of the morula stage and in the method of entoderm formation
from definite entoderm mother cells arising from the unilaminar
entectoderm the opossum closely resembles its relative Dasyurus
as described by Hill. It is, of course, possible that Dasyurus
represents the more typical development among the mar-
supials, as would appear also from Hill’s description of some
vesicles of Macropus and Parameles, in which the entoderm is
laid down in vesicles less than 1 mm. in diameter, just as in the
opossum.

60 CARL G. HARTMAN

TABLE 5
Cell counts in young opossum blastocysts

NUMBER OF

NUMBER OF TOTAL

IDENTIFICATION CELLS IN WALL | _ FIGURES AND PLATES WHERE
NUMBER Robins ec a ae Soe ae Beers OE ILLUSTRATED
314 (2) 2 (?) 28 30 2 aN;
314 (5) 0 32 32 4, VI
191 (2) 0 32 32 10, XIII
191 (5) 0 34 34 LXV
292 (4) 0 46 46 6, VI
88 (5) 0 50 50
50 (3) 0 52 52
88 (138) 0 62 62
50 (7) 0 70 70 Devil
50 (5) 1 64 65 1, VII
83 (5) 1 52 53 122 XVI
50 (8) 2 61 63 Davill
88 (10) 2 50 52
88 (23) 2 55 57 10, XVI
298 (1) 4 61 64 7 and 8, VI
298 (3) 4 20 124 20, XVI
356 (3) —— — 100 13s SOWA
88 (20) 5 60 65
88 (7) 5 82 87 ar Wl Mil SWI
83 (1) 5 106 111 19, XVI
88 (1) 6 UP 7 ‘
88 (17) 6 97 103 Pe, WANES 16°, VIL
298 (5) 8 118 126 6, XVI
50 (4) 9 59 68 3, VII
88 (11) 9 94 103 22, XVI
88 (3) 10 59 69 §, VII
SS) ee 10 60 70 8 and 9, XVI
88 (16) 10 72 82 AN NYAS TL AVA
88 (9) 11 s+ 95 106 5 VL
344 (14) 19 174 193 16, XVI and 17
344 (11) 23 141 164 15 OXOVEL
344 (7) 45 124 169 7, VIII
356 (4) 42 241 283 6 and 7, XVII

356 (5) 48 201 249 10 and 11, XVII

DEVELOPMENT OF THE OPOSSUM 61

g. Included cells which may not be entodermal

It sometimes happens that a blastomere in early cleavage
becomes displaced, fails to attain its proper pos tion at the
periphery and thus comes to be surrounded by its fellows. I
have several 16-celled eggs with one such misplaced blastomere
(fig. 17, pl. 15). Another egg, No. 314 (2), shown in figure 2,
plate 5, has two included cells. One of these in the section
figured is near a gap in the blastocyst wall, and it might be
supposed that it had migrated from the unoccupied space. But
this egg is made up of only 28 cells, a stage at which the blas-
tocyst would hardly be expected to be completed. The cells
were probably accidentally included at a somewhat earlier
stage and are probably not typical entoderm mother cells. The
included cell in ovum No. 838 (5) shown in figure 12, plate 16,
has every characteristic of an entoderm mother cell except that
it is unusually large. Some other included cells, as, for example,
those in figures 10, plate 21, and 13, plate 22, are clearly undi-
vided blastomeres of an early cleavage stage, as previously
pointed out. The true entoderm mother cells are quite dis-
tinctive and are not readily mistaken for abnormal cells.

h. Further polar differentiation

After the proliferation of entoderm mother cells is well under
way, the differences between the embryonic and the non-
embryonic areas of the blastocyst become more and more pro-
nounced. ‘The former becomes marked by the large size of its
cells as well as by the presence at that pole of entoderm mother
cells; the non-embryonic portion becomes progressively more and
more attenuated. These changes are readily understood from
plates 16 and 17. The increase in size of the blastocyst is
largely due to the spreading of the non-embryonic or tropho-
blastic ectoderm, and as a result the embryonic area comes to
occupy a more and more restricted proportion of the surface
of the vesicle, and this process continues until the formation
of the bilaminar stage has been completed. The change in

62 CARL G. HARTMAN

proportionate number of embryonic and trophoblastic cells is
apparent from the few examples given in table 6.

If a comparison be made between the facts shown in table 6
and the illustrations referred to therein, it is apparent that the
increase in the number of cells and the differentiation proceed
pari passu. In figure 15, plate 16—a longitudinal section
through ovum No. 344 (11)—the formative area is roughly
marked out by the position of the entodermal cells and the yolk
and coagulum surrounding them; there is some thinning out of
the trophoblastic cells. Ovum No. 344 (14) is slightly more
advanced. It was cut tangentially to the formative area, to
which nine of the twenty-two sections belong, the limits of this
area being determined by the presence in the ninth section of

TABLE 6

Number of embryonic and trophoblastic cells number of cells

IDENTIFICA- = 4! EMBRYONIC |TROPHOBLAS- = .
Savon | xaos |-ar cata | EXTAGEN | omic | “ax Pree
344 (11) 164 23 71 70 15, XVI
344 (14) 193 19 76 98 16, and 17, XVI
356 (4) 283 42 101 140 6 and 7, XVII
356 (5) 249 48 126 75 10 and 11, XVII

the last entodermal cells. The further differentiation between
the two areas, as seen in litter No. 356, is quite apparent from a
glance at plate 17.

In the eggs of litter No. 344, of which I have seven excellent
preparations, the vesicular structure was quite apparent in the
living state as well as after fixation, but it was not possible to
show this in the photographs taken after staining them, since the
albumen also absorbed considerable stain (fig. 2, pl. 6). Two
eges were, however, photographed alive in Ringer’s solution
and are shown in figures 5 and 6, plate 8. The former was taken
in side view and shows the yolk and coagulum hanging in the
vesicle like a bunch of grapes; in the sections of the egg taken
longitudinally the relations were found to be as in life (fig. 7,
pl. 8). The other eggs shown in figure 6 was photographed

DEVELOPMENT OF THE OPOSSUM 63

with the embryonic area uppermost; the dark spot in the center
is the rather opaque yolk mass.:

In the eggs of litter No. 356 the polarity was always apparent
in whatever medium they were placed, whether in Ringer’s
solution immediately on removal from the uterus or in alcohol
after fixation; hence these eggs were readily oriented for section-
ing. One of these eggs was photographed by transmitted light
in salt solution and is shown in figure 1, plate 8. It is a perfect
sphere, situated in the center of the egg as in younger. stages
(fig. 4, pl. 12; fig. 1, pl. 6). The embryonic area is an opaque
mass at one pole and the trophoblastic area is a thin layer
making up the rest of the vesicle. This egg is typical of all
of this litter (fig. 1, pl. 6), all of which measure 0.17 to 0.20 mm.
in diameter through the vesicle. Fixation and staining have not
changed the relation of structures essentially and even the
distortion due to imbedding is very slight (compare figs. 1 and 2,
Plo igs >. plo and fig, 1 plo):

A somewhat transitional stage between Nos. 344 and 356 is
furnished by litter No. 144 (figs. 1 to 3, pl. 17). These eggs
were overfixed in Carnoy’s fluid, but are instructive and cor-
roborative of the trend of development described above.

In all of these litters (Nos. 144, 344, and 356) the entoderm
mother cells are still being formed, as the figures in plates 16
and 17 amply show (ENT”, ENT"). The cells are of the same
type as those previously encountered, namely, rounded and in
process of leaving the periphery. Because of the greater
density of the cytoplasm, these cells often take a deeper
cytoplasmic stain than do the neighboring cells, from which
they also become separated by a more definite cell membrane.
Occasionally the entodermal cells are united into a column, as
at ENT?, figure 5, plate 17, reminding one of such rows of cells
in the younger stages (fig. 3, pl. 6).

In these eggs, too, the margin of the embryonic area seems
to be the region of greatest proliferation. Thus figure 10,
plate 17, represents a section near the margin of the area and
shows a line of primitive entodermal cells; while figure 11, a

JOURNAL OF MORPHOLOGY, VOL. 32, No. 1

64 CARL G. HARTMAN

section nearer the middle of the series, shows entodermal cells
only at the margin.

At the stage just described there is still a considerable quantity
of yolk and coagulum in the egg. This is usually collected near
the inner surface of the embryonic area as well as among or
within the cells of the area, more rarely also in the trophoblastic
cells. Occasionally the yolk is collected in a large spherule,
as in egg No. 356 (11); this spherule measures 0.04 mm. in
diameter and a portion of it, cut tangentially, is shown at Y,
figure 4, plate 17.

1. The embryonic area superficial in position

The question may arise whether there appears at any time
over the embryonic area a transitory layer that may at all be
compared with Rauber’s layer in the Eutherian egg. Professor
Hill has homologized the embryonic area of the Eutherian egg
with the inner cell mass and the non-embryonic area with
Rauber’s layer, and hence he uses the term ‘trophoblastic’ to
designate the latter. According to this view, the embryonic
cells lie upon the surface of the ovum from the beginning and are
potentially ectoderm, entoderm, and mesoderm; in other words,
the embryonic cells are never covered with trophoblastic cells.
I believe this to be the true interpretation of the facts. Since
my collection includes an unbroken series of critical stages on
this point, if there were such a layer, it could not escape de-
tection. Nowhere is there the slightest suggestion of a transitory
layer of cells. Mitoses are always present in the superficial
layer (pl. 17), disintegrating cells never. The very method of
blastocyst formation, as described in these pages, precludes the
probability of a trophoblastic cover over the embryonic area,
which is differentiated very soon after the establishment of the
blastocyst. For, if the upper half of the unilaminar opossum
egg is not embryonic, it is trophoblastic and there can be no
embryonic area; in which case we should be forced to derive the
embryo from the trophoblast, a manifest absurdity. The pre-
ceding and the succeeding stages all show that in the blasto-

DEVELOPMENT OF THE OPOSSUM 65

cysts last described, the superficial cells at the upper pole are
ectodermal except a few which are destined to form entoderm
mother cells.

j. The primitive entoderm

In the stages thus far described the entodermal cells are still
round to polygonal and only occasionally does a cell flatten
out upon the surface of the mass as at LNT", figure 4, plate 17.
We may call these cells primitive entodermal cells (HNT?) as
distinguished, on the one hand, from the large entoderm mother
cells from which they arose (HN7") and on the other, from the
typical, flattened definitive entoderm into which they are about
to develop. The primitive entoderm, through rapid cell division,
becomes more or less crowded and shows a tendency to become
two or three cells deep, as early as the stage represented by
litter No. 356 (pl. 17).

k. Further growth of the blastocyst

When the blastocyst contains less than 200 cells, of which
about 20 would be entodermal (litter No. 344, pl. 16), the
embryonic and the trophoblastic areas each make up about
one-half of the blastocyst wall. When the number approaches
300, including 40 or 50 entodermal cells (litter No. 356, pl. 17),
the latter area has greatly extended so that the embryonic
portion occupies a third or less of the blastocyst wall. The
increase in size of the blastocyst is, therefore, to be attributed
largely to the spreading and attenuation as well as a more
rapid multiplication of trophoblastic cells (table 6).

The eggs of litter No. 194’ are illustrative of the further
development in the direction just indicated and follow close
upon the eggs of litter No. 356. The vesicle has grown from
about 0.23 mm. in diameter as the maximum for litter No. 356
to 0.34 mm. in litter No. 194’, or about double the diameter of
the ovum at cleavage. The blastocyst is still situated in the
center of the egg, which has, however, not yet increased in

66 CARL G. HARTMAN

volume of shell membrane (text fig. 2; fig. 18, pl. 18; fig. 5,
ple): .

The development of the trophoblastic ectoderm is seen to
have continued in the direction indicated above, so that in this
litter of eggs it now occupies about three-fourths of the circum-
ference of the blastocyst. Little more need be said of this
layer. It becomes progressively more attenuated until it may
have the appearance of endothelium, and even at high mag- ~*
nifications appear as a sharp narrow line with here and there a
swelling which marks the location of a nucleus (fig. 1, pl. 18).
The region may come to occupy from four-fifths to five-sixths
of the entire circumference of the blastocyst; and this again
constitutes a point of contrast with the egg of Dasyurus, in
which the formative area occupies, in section, from one-third
to one-half of the blastocyst wall.

In general, the marsupial trophoblast does not differ markedly
in structure from that of Eutherian vesicles, but more interesting
and important changes take place in the embryonic area of the
opossum blastocyst. For a short period, which includes the
stage represented by litter No. 194’ (figs. 13 to 15, pl. 17), these
changes now appear to be chiefly of two kinds: 1) further
proliferation of entoderm mother cells from the peripheral
layer, and 2) multiplication of all types of cells.

The former process gives every evidence of having slowed
down considerably since the preceding stage, the cells which
can be identified as migrating inward from the superficial layer
being of comparatively rare occurrence. Such cells are shown
at HNT”, figs. 14 and 15; they stand with their long axes at
right angles to the surface of the area and project inward among
the primitive entodermal cells now everywhere closely applied
to the ectoderm. It is clear that entodermal proliferation from
the superficial entectoderm is approaching the end.

As a result of the cell multiplication, the embryonic area has
become crowded, so that in places it is three and occasionally
four cells deep; and it may be stated parenthetically that this is
the only stage before the formation of the mesoderm that the
blastocyst wall is anywhere more than two cells deep, as the

DEVELOPMENT OF THE OPOSSUM 67

sequel will show. At this stage there is no regular arrangement
of entodermal cells into an epithelium, and, even in the super-
ficial layer, regularity is only approximated (figs. 13 to 15,
pl. 17). The cells which are not in contact with the albumen
are as irregular in shape and size, at least in my specimens, as
they aré in arrangement; only the nuclei preserve a uniformity
of size and structure.

The superficial cells for the most part are clearly embryonic
ectoderm, and all of the nuclei seen below this layer are primitive
entoderm. Most of them possess rounded nuclei, and only
here and there in the sections is there any indication of cells
which tend to flatten out into definitive entodermal cells (HN T?,
fig. 15). In this respect there has been little progress since the
preceding stage.

Litter No. 194’, was found four days after copulation or about
two days after the beginning of cleavage.

A somewhat more advanced stage is represented by litter
No. 349, one of which is shown photographed in the living state
in figure 3, plate 8. It measures 0.352 mm. through the vesicle.
Figure 4 is a section through the youngest egg of the litter and
belongs to an earlier stage corresponding to litter No. 344.
One of the two eggs like the one in figure 3 was sectioned, the other
was accidently broken and was used for study in toto. The
two are in essential agreement. The formative area, shown as
a distinct opacity in the living egg (fig. 3, pl. 8), is larger in
area than in the litter just described; the trophoblastic area is
thick-walled and less extended than would be expected at this
stage. The embryonic area is crowded, the cells being three
and four cells deep in some places. A large number of cells of
all types—embryonic ectoderm, primitive and definitive ento-
derm—are in mitosis, chiefly in the spireme stage, as though a
wave of cell division had spread over the entire area. Here
and there an entoderm mother cell is still in process of forma-
tion. The definitive entoderm has begun to differentiate and to
spread beyond the area (YNT). In surface view the embryonic
area is approximately round and is sharply marked off from
the surrounding trophoblastic ectoderm. ‘This description shows

68 CARL G. HARTMAN

that the entoderm is present in a watch-crystal-shaped mass at
one pole of the egg in vesicles of 0.30 to 0.85 mm. The mass is
thicker in the middle, being even three to four cells deep. Only
the outer superficial layer is ectodermal, the massed cells
beneath being all entodermal.

The opossum blastocyst differs, then, both from the corre-
sponding stage of Dasyurus, on the one hand, and of the higher
mammals, on the other. In Dasyurus the entodermal cells
flatten out and spread singly as they are formed and never pile
up in a mass as in the opossum. ‘There would seem to be in the
opossum a hearer approach to the Eutherian ovum in its pos-
session of a. kind of ‘inner cell mass’ (fig. 15, pl. 17; fig. 1, pl. 18).

But there are fundamental differences. For in the Eutheria
the entoderm seems to arise only from the cells on the inner
surface of the inner cell mass, presumably from a single layer.
The outer or superficial layer of the blastocyst is Rauber’s
layer; between the two is the embryonic ectoderm, a layer of
cells variable in thickness and at first irregularly dispersed.
Thus, if figure 1, plate 18, represented an Eutherian egg, the
superficial layer would constitute Rauber’s layer; beneath would
be the irregularly disposed ectoderm (cells marked ‘EHNT*”’),
and only the innermost layer would be the entoderm. In the
opossum there are only two layers: 1) embryonic ectoderm, a
superficial layer, one cell deep, and 2) all the remainder which
is entodermal. If there seem to be more than two layers, as in
the figure just referred to, the outer layer is the ectoderm, the
inner the differentiated entoderm, and between the two a mass
of cells which. are still undifferentiated or primitive entoderm
which are yet to spread and form entoderm. The opossum is,
therefore, fundamentally like Dasyurus; but the resemb!ance is
obscured by the temporary massing of entodermal cells, the re-
sulting picture superficially resembling an Eutherian vesicle with
spreading inner cell mass.

l. The end of entoderm formation and the spreading of the entoderm

In my previous publication I described some vesicles from 0.3
to 0.5 mm. in diameter, of the type shown in figures 1 to 4,

DEVELOPMENT OF THE OPOSSUM 69

plate 18, in which certain cells seemed to migrate from the
embryonic ectoderm to take their place among the entodermal
cells. I interpreted these cells as entoderm mother cells and
presented a number of cases which closely parallel the process of
entoderm formation in Dasyurus as described by Hill. Indeed,
Hill states that in Macropus the primitive entodermal cells are
already recognizable as cells situated internally in the blastocyst
of 0.35 mm. and in Parameles in vesicles of about 1 mm., which
would seem to correspond very well with the condition in the
opossum. Certain cells drawn in figures 3 and 4, plate 18,
might conceivably be proliferating entoderm. If this be true,
then certainly these sporadic cases are the last stragglers in the
stream of entodermal cells which arise from the entectoderm.
The climax in the formation of entoderm in the opossum, how-
ever, occurs long before this, namely, in blastocysts between
0.15 and 0.30 mm. in diameter.

The differentiation of primitive entodermal cells into definitive
entoderm takes place with rapidity soon after a diameter of
0.34 mm. is attained (litter No. 194’), so that in vesicles of
about 0.50 mm. the entoderm has largely assumed its squamous
structures and lies closely appressed against the simple uni-
laminar ectoderm. As soon as this differentiation is well under
way, the entoderm at once migrates beyond its region of origin
toward the opposite pole of the vesicle. These changes are
readily observed in some typical examples furnished from litters
Nos. 194’, 349, 40, 48, 175’, 339, 299’, and 347.

In litter No. 194’ as described above, the spreading of the
entoderm has scarcely begun. In litter No. 349 (fig. 12, pl. 17)
a few cells have flattened decidedly, while the majority are still
in the condition of indifferent primitive entoderm. The tendency
to spread is exhibited on the entire margin of the area. As soon
as the entodermal cells have differentiated, they at once stain
much darker, a characteristic which they maintain in sharp con-
trast to the ectoderm throughout the bilaminar stage; this is
true without exception.

A somewhat later stage is represented by egg No. 43 (7),
which is large, and has a greatly attenuated trophoblastic area,

70 CARL G. HARTMAN

and represents the normal condition at this stage. The section
in figure 1, plate 18, is the tenth of thirty-five sections through
the embryonic area; that is, lies to one side of the midline. The
definitive entoderm would seem from this section to clothe the
entire inner surface of the area, but a study of the series dis-
closes the fact that the entoderm has differentiated only in spots,
chiefly at the periphery of the area. That these changes take
place chiefly in the periphery first would appear also from
ovum No. 339 (3) shown in figures 6A and 6 and in figure 2,
plate 6. At HNT? is a group of cells which run through ten
sections; they are primitive entodermal cells not yet differ-
entiated. So also in figure 8 (egg No. 175’ (2) ) the entoderm
has already advanced considerably toward the equator, although
there are still some undifferentiated cells near the middle of the
area. Similar undifferentiated cells are also seen in other
vesicles, as at HNT?, figures 3 and 4.

Litter No. 352 consists of small blastocysts which fall into
the stage under discussion. The group of eggs was photo-
graphed fresh in Ringer’s solution by transmitted light (fig. 1,
pl. 9). The vesicles with their more or less opaque embryonic
areas are very evident. In the largest specimens the ento-
derm has entirely differentiated, except in one (fig. 14, pl.
13; fig. 8, pl. 21) in which a large blastomere has retarded the
spreading of the cells (compare fig. 2, pl. 9, and fig. 2, pl. 18).
In these eggs the entoderm has also advanced some distance
toward the equator of the blastocyst.

m. Maximum attenuation of the blastocyst wall

The trophoblastic ectoderm, it was seen above, begins its
thinning and spreading process soon after the proliferation of
entoderm begins (in 0.15-mm. ova) and reaches its maximum
when the formation of new entodermal cells from entectoderm
ceases, and the entoderm begins to line the lower hemisphere
(0.50-mm. blastocysts). The spreading and attenuation, how-
ever, affect the embryonic as well as the trophoblastic area and
takes place rapidily. while the entoderm is migrating to the

DEVELOPMENT OF THE OPOSSUM (All

opposite pole. In all of the eggs of litters 175’ and 347 (pl. 18)
these facts are clearly shown. Blastocyst No. 175’ (9) is an
extreme case in point (figs. 7 and 7A, pl. 19). In succeeding
stages the embryonic area thickens progressively but slowly,
until it reaches its maximum in blastocysts 1 to 1.5 mm. in
diameter, after which it remains more or less constant until the
embryo begins to differentiate.

n. Cause of spreading of the entoderm

In the spreading of the entoderm the chief factor is the active
migration of the entodermal cells. The passive spreading, due
to the enlargement of the vesicle, as had been suggested in the
case of other mammalian vesicles, is, in the opossum, a negligible
factor. An inspection of plate 18 will make this clear, for the
vesicles are about as large when the entoderm begins to spread
(fig. 1) as when it has reached the opposite pole (fig. 5). In
fact, comparison of eggs of the same litter (figs. 5 and 7) show
that the size of the vesicle bears no relation to the extent of the
entoderm. My observations on numerous eggs at this stage
(Nos. 347 and 299’) go to show that, when once begun, the
spreading of the entoderm proceeds rapidly. I have not been
able to demonstrate amoeboid movements in the cells, but
processes sometimes occur on entodermal cells at about this
stage (fig. 2, pl. 19). |

0. Changing position of the vesicle in the egg

In most of the eggs which mark the early stages in the spread-
ing of the entoderm, the vesicle still occupies practically the
center of the eggs as in previous stages (text figs. 1 and 2). In
litter No. 352 (fig. 1, pl. 9) the tendency of the vesicle to ap-
proach the shell membrane is already manifest to some extent;
also in litter No. 175’ (fig. 7, pl. 19). In all later stages the
vesicle occupies an eccentric position and in most cases it is in
immediate contact with the shell membrane. This contact is
established, therefore, for the first time about the beginning of
the bilaminar stage, or about four days after the beginning of

C2 CARL G. HARTMAN

cleavage. This delay is in contrast, again, with the egg of
Dasyurus, in which the blastomeres in the 16-celled stage have
already establ’shed contact with the shell membrane on_ all
sides, the albumen of the egg, always limited in thickness, having
entirely disappeared. The formative area almost always reaches
the shell membrane first, the exception being very rare. The
albumen, therefore, becomes concentrated at one pole, gradually
decreasing in amount with the growth of the blastocyst. Hence-
forth the stage of advancement of the blastocyst may be gauged
by the amount of albumen which appears as a crescent in the
egg when viewed from the side or when seen in a longitudinal
section (eggs No. 299’, in figs. 1 and 2, pl. 6; fig. 5, pl. 10, etc.).
That this eccentric position is not an artifact due to fixation or
other causes, is shown by the fact that in the living egg the
blastocysts are situated in exactly the same position as after
fixation, as the photographs (fig. 4, pl. 1; figs. 4 to 6, pl. 9)
show.

p. Some abnormal eggs

Since future workers on the opossum are likely to encounter
abnormal material, it is not amiss to describe several abnormal
eggs of about the stage just described.

In the group of eggs in figure 2, plate 6, a number of such
abnormal specimens are shown: Nos. 294 (1), 294 (2) and 294
(3) and 339 (4). The last mentioned is the least abnormal of
all. In the living state the vesicle was spherical and remained
so throughout the process of imbedding (figs. 5 and 5A, pl. 19).
A similar egg is shown in figure 15, plate 18, and the embryonic
area of a third in figure 6, plate 19. These eggs, all from one
litter, are in close agreement and the relation of ectoderm and
entoderm is as in normal eggs of this stage (compare pl. 18).
But the wall of the vesicle consists of unduly inflated cells with
very diffuse cytoplasm and often large nuclei. Egg No. 339 (3)
is exceptional in this litter, for its normal appearance; it doubt-
less represents the normal stage to which the others should have
attained (fig. 6, pl. 9; fig. 2, pl. 6; figs. 6 and 6A, pl. 18).

DEVELOPMENT OF THE OPOSSUM Ta

The other more abnormal eggs referred to above show even
at low magnification evidences of abnormality (fig. 2, pl. 6).
The vesicles are not plump and rounded, but more or less
shriveled and are surrounded by a large ‘perivitelline space.’
That the abnormalities are not due to the method of fixation is
shown by the appearance of the living eggs, of litter No. 294
reproduced in figure 1, plate 11. In sections made of these
eggs the walls are composed of cells swollen to enormous volume
and are extreme cases of the condition shown in fig. 5A, pl.
19. Many cells are in mitosis, with the chromosomes strewn
about pell-mell throughout the cell. Similar eggs are also met
in normal litters (fig. 4, pl. 9). The ‘pear-shaped vesicle’ de-
seribed by Selenka and figured in his Tafel XVIII, Fig. 1 u. 2,
was doubtless an egg of the type just described.

PART IV. THE BILAMINAR BLASTOCYST

GENERAL DESCRIPTION
a. Material

The various stages in the bilaminar blastocyst of the opossum
are represented in my collection by an unbroken series separated
from one another by minutes rather than hours of development.
Two hundred and thirty-five normal eggs were secured from
thirty-two litters of twenty-five different animals; hence it
may be assumed that, the following description gives in detail
the normal opossum egg during these stages. One hundred and
fifty eggs were sectioned or were dissected for study of surface
views; and the former include many that were carried through
the imbedding and sectioning process without collapse and
with the minimum of shrinkage.

b. The living eggs

The general trend of development during this period may
be followed by reference to the photographs of living eggs
presented in plates 1, 2, 9, and 11 of this paper.

74 CARL G. HARTMAN

All of the eggs still lie free in the lumen of the uterus (fig. 10,
pl. 1; fig. 8, pl. 2) and are distributed as in the preceding stages,
often grouped near the os uteri; hence one should not speak of
the ‘implantation’ of the eggs even at the 2-mm. stage. The
shell membrane has attained considerable thickness (Hartman,
16) and throughout the stage in question maintains the shape
of a perfect sphere.

Before the entodermal spreading is well under way the blas-
tocyst occupies approximately the center of the egg (fig. 1, pl. 9).
Before the entoderm has reached the opposite pole of the blas-
tocyst the embryonic area has almost or quite come into con-
tact with the shell membrane (compare figs. 3 and 4, pl. 19),
giving the blastocyst a decidedly eccentric position, It now
fills one-half or less of the egg and has the shape of a bi-convex
lens (fig. 5, pl. 10). This migration of the blastocyst may be
due to the increased metabolism of the more voluminous cells
of the embryonic area, as a result of which the albumen is here
more rapidly digested and absorbed. This position is maintained
in the subsequent stages (compare eggs No. 299’, fig. 1, pl. 6).

The size of the entire egg containing the youngest bilaminar
blastocysts with just closed entodermal sae is very little greater
than the youngest uterine eggs, although the albumen has
become denser and the vesicle wall has become considerably
differentiated. Thus, for example, the diameter of the eggs in
figures 4 to 6, plate 9 (all bilaminar blastocysts), is only a little
greater than that of eggs in cleavage stages (figs. 1, 3, and 5,
pl. 1). Again, the two litters shown in figures 3 and 4, plate 1,
exhibit an evident, but not striking growth in volume, although
they have developed from the 4-celled stage in the former to
young bilaminar blastocysts in the latter in a period of four
days, or 40 per cent of the entire period of gestation!

In all of the young bilaminar blastocysts the embryonic area
is plainly outlined and distinctly marked off at the junctional
line from the trophoblastic region. This differentiation increases
with the growth and development of the egg.

From the beginning, the bilaminar stage is acentalie the
period of growth; this period thus follows the formation of the

DEVELOPMENT OF THE OPOSSUM Co

entoderm, whereas in the Dasyurus it precedes as well as follows
the process. At first the blastocyst grows faster than the shell
membrane, for gradually the albumen disappears before the
advancing trophoblastic area. This growth would seem, there-
fore, to affect the trophoblastic more than the embryonic area, —
which latter lies in contact with the shell membrane from the
earliest bilaminar stage. The embryonic area, however, easily
keeps pace with or even exceeds the rate of growth of the entire
vesicle; for in the 1-mm. eggs it is proportionally larger than
in certain younger stages. It would seem, then, that, despite
the absence of albumen, the area receives sufficient nutriment
for vigorous growth or is indeed better supplied by virtue of its
superficial position in the egg, with the secretion of the uterine
glands.

Eggs about 0.8 mm. in diameter, as illustrated by litter
No. 306’, removed four and one-half days after the beginning of
cleavage, still contain considerable albumen which is readily
visible at all positions of the eggs and may be seen on the pho-
tographs of the eggs (fig. 7, pl. 10; fig. 17, pl. 13; fig. 1, pl. 21).
When the diameter of 1 mm. is reached, the quantity of albumen
has been -considerably reduced and is. mostly confined to the
trophoblastic region below the equator of the egg (fig. 2, pl. 21).
In living specimens of such eggs the albumen is visible as a
narrow crescent, only when viewed from the side, and is not
visible in photographs of living eggs (fig. 5, pl. 2). The re-
duction of albumen continues, and when the diameter of the
egg approaches 2 mm. and the mesodermal proliferation is
about to begin there is only a thin film of albumen left (fig. 6,
pl. 2; fig. 20, pl. 13). The amount of albumen at any stage is,
of course, variable. It may still occur in small amounts in early
primitive-streak stages (fig. 4, pl. 2; fig. 22, pl. 13).

The opossum blastocyst, therefore, begins as a perfect spheer
at about the 32-celled stage. It maintains this shape until the
_ definitive entoderm begins to spread, when the blastocyst as-

sumes a biconvex form, flattened in the direction of the egg axis,
and lies with the formative area against the shell membrane.
The spherical form is again attained when the trophoblastic area

lod

76 CARL G. HARTMAN

has reached the shell membrane, which occurs almost com-
pletely when the egg is 1 mm. in diameter, more perfectly
at a diameter of 1.5 to 2 mm. It then maintains the spherical
‘rom until, through crowding of large litters in the pregnant
uterus, the vesicles are somewhat misshapen through mutual
pressure.

The embryonic area of the larger blastocysts, due to its pro-
toplasmic differentiation, now stands out clearer, so that it is
plainly visible in living eggs. It is recognizable in photographs
of living eggs, but much more clearly in photographs o° eggs
immersed for a few minutes in the fixing fluid (fig. 2, pl. 11).

The size of the embryonic area varies greatly, even in pro-
portion to the total surface area of vesicles in the same litter.
In general its diameter occupies between one-fifth and one-
fourth of the circumference of the egg, occasionally a little less
than one-fifth, sometimes nearly one-third. But it never reaches
the equator as in Dasyurus. An average 1 mm. blastocyst is
shown in figure 9, plate 21. Various measurements are given
under the legends of the eggs illustrated.

Aside from these details, little may be learned from a study
of the living egg, and we must turn to preparations for more
intimate details of structure. The progress of development will
be followed by describing first the youngest stage, then the 1
mm. blastocyst, and lastly the blastocyst just preceding the
proliferation of mesoderm.

THE JUST COMPLETED BILAMINAR BLASTOCYST

The bilaminar stage may be said to begin when the entoderm
has migrated to the trophoblastic pole of the egg opposite its
point of origin, thus forming a closed sac within the ectoderm.
This stage was attained by most of the eggs in litter No. 299’
about four days after the beginning of cleavage. The vesicle
occupies about one-half of the egg contents.

a. The embryonic ectoderm

In the youngest bilaminar blastocysts the embryonic area is
approximately circular in shape and clearly visible, but not as

DEVELOPMENT OF THE OPOSSUM Hae

clear-cut nor as definitely and neatly circular as in later stages.
In surface view of preparations the junctional line between the
two areas can always be made out (figs. 11 and 12, p'. 19); but
in sections some difficulty is experienced in trying to determine
with exactness the marginal cells of either area, for the cells of
the embryonic area have not yet assumed that density of pro-
toplasm characteristic of later stages. The embryonic ectoderm
is at first comparatively thin, consisting of a single layer of
somewhat flattened cells as in figures 5 to 7, plate 18, and 3
and 8 to 10, plate 19. The area gradually thickens and the
cells become cubical in section (fig. 4, pl. 10, fig. 4, pl. 19). In
the younger eggs the nuclei are, therefore, further apart, but,
as they multiply, they become more and more crowded until
they are almost or quite in contact (compare fig. 3B, pl. 20, and
fig. 12, pl. 19). Im surface views the area is studded with
mitotic figures (fig. 3B, pl. 20).

The surface views presented in figures 3B, plate 20, and 12,
plate. 19, of which the latter is the more advanced, show that
the junctional line between the embryonic and the trophoblastic
areas is quite definite, sometimes being marked by a con-
tinuous sharp line around the entire area. Figure 3B, plate 20,
is especially instructive in this connection, since it represents
in surface view (x 500) a portion of the same area shown in
figure 3A in section (X 200); it is the portion removed before
imbedding from point A, figure 3. There is, neither in the
section nor in the surface view, any great dfference of tone
between the two areas, but the junctional line (XX) or margin
of the embryonic area may easily be located where the ecto-
dermal nuclei become less crowded. The junctional line is
always more clearly defined in specimens fixed with a fluid that
will bring out the cell membranes, as in figure 12, plate 19. It
is seen that the line is formed of the cell membranes of con-
tiguous cells bordering the areas. It is a perfectly definite
structure: the marginal cells of the two areas do not intermingle
nor do transitional cells occur between the two types.

78 CARL G. HARTMAN

b. The trophoblastic area

The cells of this area are very attenuated in the late unilami-
nar stage, but as they multiply they also thicken and increase
their volume. The thickening is often most pronounced at the
vegetative pole where numerous mitoses may occur (figs. 5
and 7, pl. 18). This is a matter of some interest, for in larger
blastocysts this region may continue to be more thickened than
the remainder of the trophoblastic area; or certain ‘blisters’ may
occur there, such as will be described under the 1 mm. blastocyst
(O; figs: 2.and:3, pl. 21).

In all stages the cytoplasm of the trophoblastic cells is very
diffuse and loosely reticular, but in the early bilaminar stage
especially it tends to break down into strands and ret’ culum,
leaving the cell membrane collapsed and wrinkled. Only
around the nuclei is there a denser mass of well-fixed cytoplasm
and the nuclei themselves maintain their form and structure as
perfectly as in the embryonic area (fig. 2, pl. 20). Normally,
the trophoblastic layer follows the curvature of the albumen
layer to which it remains closely applied, as is apparent from a
study of the living eggs and photographs of them and as the
best preparations show (fig. 4, pl. 1; fig. 17, pl. 13). Frequently,
however, the vesicle collapses more or less in this area, leaving
as an artifact a space between the vesicle and the albumen
(fig. 4, pl. 10; ART, fig. 1, pl. 19). At this stage the albumen
seems to be most dense nearest the shell membrane and often
very loosely layered near the vesicle, which would account for
the more frequent breaking down of the albumen at this point
in the specimens.

c. The entoderm

Soon after completely lining the blastocyst wall the entoderm
is everywhere the same, passing over the junctional line without
change. This condition remains so throughout the bilaminar
stage, except for certain modified cells to be mentioned in con-
nection with older blastocysts. At first the entoderm is ex-
tremely delicate so that it appears in section to be discon-

DEVELOPMENT OF THE OPOSSUM 79

tinuous, because in the fixing fluid portions of the cells break
down. That this is the correct explanation is seen from surface
views of good preparations, as in figure 11, plate 19, in which only
the entodermal cells are shaded. They are seen to be connected
by fibrous strands, the coagulated portions of the delicate cells.
The entoderm may, therefore, be considered practically con-
tinuous. In most surface mounts the entoderm appears like a
mottled surface, for the cells are thick in the middle, hence
darker, and shade off almost into nothingness toward their
edges (fig. 2, pl. 19).

The entoderm invariably stains darker than the ectoderm.
It is interesting to note that an entodermal cell in mitosis is
darker, often very decidedly so, than its fellows in the resting
stage (fig. 2, pl. 19); while on the other hand an ectodermal cell
is usually much lighter when in mitosis (fig. 12, pl. 19), so that
in some surface views dividing cells look like holes in the wall.
Yolk granules abound among the cells of the embryonic area,
both ectodermal and entodermal, and occur occasionally also in
the trophoblastic region.

A somewhat older egg (figs. 4 and 4A, pl. 20) is presented be-
cause of the degenerating cells included within the cavity.
Such cellular remnants have been noted in apparently normal
- vesicles of Eutheria (Hartman, ’16, pp. 46 and 47). In this egg,
too, the albumen is definitely arranged in three layers of varying
density, a condition noted also in a few other specimens.

THE 1 MM. BLASTOCYST
a. General description

The typical 1-mm. blastocysts contained in litter No. 343.
(fig. 6, pl. 2) were removed seven and a half days after copu-
lation. It has already. been pointed out that only a small
amount of albumen still remains in these eggs and the vesicle
has become very nearly a perfect sphere. The embryonic area
is now more sharply marked off from the surrounding tropho-
blast and lies like a cap at one pole of the egg (fig. 9, pl. 21).
It is, in fact, occasionally in alcoholic specimens raised in relief

JOURNAL OF MORPHOLOGY, VOL. 32, NO. I

80 CARL G. HARTMAN

above the surface of the ovum like a blister, a condition probably
due to its greater density and resistance to shrinkage as com-
pared with the trophoblastic area.

The growing contrast between the two regions of the egg,
which is now as clear-cut in sections as in whole mounts, is due
to the increasing difference in the structure, as well as to the
number of the formative cells. These are taller, much more
crowded, and contain a denser and more granular cytoplasm,
and this contrast in the types of cells is a constant character,
no matter what the fixation, and the differences that exist
among the specimens are those of degree only. ‘These points
are evident from an inspection of figures 1A, 2A, and 4 to 7,
plate 21, which were drawn as nearly as possible in imitation of
the tone of the specimens.

While there is great variability in the thickness of the em-
bryonic areas in 1-mm. blastocysts, it is true that in most cases
the area has become considerably thickened as the vesicle has
grown in volume and as the area- has increased in diameter
(pl. 21). The cells have become mostly tall cubical to columnar
and in the embryonic area are now nearly or quite as much
crowded together as in older stages (compare fig. 12, pl. 19, and
fig ply 22)k

The embryonic ectoderm is arranged strictly in a single
layer, never stratified or pseudostratified. The nuclei are
practically on a level throughout, and this is one of the points
of contrast with the blastoderm of other mammals. Mitotic
spindles usually stand with their axes parallel to the surface of
the egg (fig. 7, pl. 21). Frequently the cell that is in mitosis
juts out above the level of the ectodermal layer (fig. 5, pl. 21).
as in the blastocysts of the rabbit and other mammals, and such
cells almost always stain less deeply, a fact that applies both to
sections and to surface views.

The trophoblastic area has also developed more mass and
thickness, proportionally quite as much as the embryonic area
(figs. 6 and 7A, pl. 21). Typically it is about 8 to 10 » in thick-
ness. It is usually uniform in structure at all points and fits
closely to the albumen (fig. 6), except when artificially separated

DEVELOPMENT OF THE OPOSSUM Sl

from it in fixation (fig. 7A). In the 1l-mm. blastocyst it will
endure fixation better than in younger stages. In all cases, in
contrast with the uniform granulation of the embryonic area,
the trophoblastic cells are reticulated and often possess coarse
meshes or appear highly vacuolated. In extreme cases, es-
pecially when fixed in aceto-osmic-bichromate, the trophoblastic
area may be greatly swollen; but this may also sometimes happen
even in so reliable a fluid as Bouin’s, as in figure 4, plate 22.
The trophoblastic area is thus much more affected by fixation
than the embryonic area.

While, as a rule, the trophoblastic area is rather uniform in
' thickness throughout its extent, there are frequent exceptions
which deserve special mention. The area may gradually thicken
toward the lower pole (fig. 4, pl. 22), or there may be a thick
mass of cells jutting out into the albumen, and even touching
the shell membrane at that point. In such cases the entoderm
is continuous over the mass. In still other cases the ectoderm
at the extreme lower pole is depressed outward into a pocket
which may also come into contact with the shell membrane
(O, figs. 2 and 3, pl. 21). The entoderm bridges over this cavity
in a continuous layer and does not follow the ectoderm into the
pocket. In the whole egg the pocket is quite evident and
looks like a blister on the vesicle. These structures can scarcely
have any special significance, since they are not of constant
occurrence, nor are they situated at a point of special future
importance.

As in both younger and older stages, the entoderm is a con-
tinuous layer lining the entire cavity of the blastocyst. It
consists of a very attenuated layer of large squamous cells quite
typical of the corresponding stage of all mammals. The cyto-
plasm is mostly gathered near the center of the cell, where the
nucleus lies. In surface views the entodermal nuclei usually
appear larger than the ectodermal and the chromatin granules
in them are more evenly distributed. They can be recognized
by this difference as well as by the depth of focus required to
see them. The entoderm always has a stronger staining
reaction than the ectoderm; I find no exception to this rule.

82 CARL G. HARTMAN

b. The bilaminar blastocyst according to Selenka

In his ‘Studien’ (’87) Selenka briefly describes two opossum
blastocysts of about 1.1 mm. in diameter. The lithographs pre-
sented by him are idealized drawings, reconstructed from his
sections, which I judge to have been considerably shrunken by ~
the treatment to which they were subjected. The illustrations
give the correct relation of the structures except for the diffuse
junctional line, although Selenka is in error as to the homology
of the ‘Granulosa membran,’ as he terms the shell membrane.

c. The 1-mm. blastocyst according to Minot

In 1911 the late Professor Minot published a description of
six 1-mm. blastocysts of the opossum, of which two were fixed
in Flemming’s fluid and four in Zenker’s and of the latter, two
were fixed in situ with the uterus.

He gives an adequate description of the embryonic area of
this stage, as did Selenka in 1887. Of especial interest, however,
is Minot’s description and interpretation of certain cells in the
trophoblastic area. In a vesicle which he dissected and mounted
flat on a slide he found numerous large light areas apparent as
‘minute round holes’ when viewed with a hand lens. He inter-
preted these areas as gaps in the ectoderm filled with entodermal
cells which thus reach the surface at these points. Corrobora-
tion was found in the study of the serial sections. The author
furthermore draws a comparison between these large lightly
staining entodermal cells, which rise to the surface in the tropho-
blastic region of the opossum egg, and the small darkly staining
entoderm mother cells which appear in the embryonic ectoderm
of the Dasyurus blastocyst.

While the preparation of the present paper was in progress I
had the privilege of studying Professor Minot’s specimens at
the Harvard Medical School. As I expected, the serial sections
of eggs fixed and sectioned in toto with the uterus are badly
shrunken and filled with coagulum in a manner which never
occurs in eggs removed from the uterus and treated separately.

DEVELOPMENT OF THE OPOSSUM 83

The specimens are unique, too, in that the entoderm is somewhat
lighter in stain than the ectoderm, as described by Minot.

The surface mount is nicely fixed and is, histologically, an
excellent preparation. The light areas are as described by
Minot, and they are even more striking in the specimen than in
his figure 2B. It is my Judgment, however, that the vesicle in
question is not entirely normal, for the reason that among all of
my numerous specimens, I have never encountered any possess-
ing such large light spaces. It is true that normally small
lightly staining areas occur in almost all opossum vesicles of about
this stage; and they usually mark the presence of cells in mitosis
(figs. 12 and 12A, pl. 22), or cells that have just divided or are
preparing to divide, and they are especially prominent in speci-
mens fixed in bichromate mixtures. But they never attain such
size as in Doctor Minot’s unusual specimen. I have no explana-
tion to offer of the phenomenon; I saw no evidence of degenera-
tion of cells at those points.

Again, among all of my specimens I have looked in vain for
entodermal cells coming to the surface in bilaminar blastocysts, -
either in surface views or sections; and I am certain that
normally this does not occur. I have convinced myself, how-
ever, that also in the Harvard specimen the entoderm is nowhere
at the surface and that Doctor Minot was in error in his inter-
pretation. In the first place, by careful focusing with the oil-
immersion lens the (ectodermal) nucleus within the light area is
in several instance seen to be superimposed over an entodermal
nucleus. Furthermore, if one plot the entodermal nuclei of
the embryonic area, it is seen that they are uniformly and
continuously distributed, entirely without reference to the above-
mentioned light areas. These areas are without doubt ecto-
dermal and not entodermal. Hence Minot’s comparison be-
tween the supposedly superficial entodermal cells in the tropho-
blastic area of: the opossum with the entoderm mother cells of
the unilaminar blastocyst of Dasyurus is a futile one.

The entoderm, therefore, never comes to the surface in the
bilaminar stage of the opossum egg. Entodermal cells in
Dasyurus and in the opossum, and doubtless in all marsupials,

84 CARL G. HARTMAN

occupy the superficial position only as undifferentiated entoderm
mother cells from which all of the entoderm is destined to be
formed. ‘The formative area is, from the beginning, potentially
ectoderm, entoderm, and mesoderm, giving rise first to the
entoderm and later in quite a similar manner to the mesoderm,
the residue becoming definitive ectoderm. ‘The trophoblastic
area consists of a single layer, the ectoderm, until lined with the
entoderm arising from the embryonic area.

THE LATE BILAMINAR BLASTOCYST
a. General description

Passing now to the later stages, we note that superficially the
blastocyst appears to have changed but little, except in. size
(compare figs. 3, 5, and 6, pl. 2). The embryonic area remains
prominent at the upper pole and less and less albumen remains
at the lower pole. Important changes in the blastocyst wall
_ are, however, to be discovered from a study of the sections.

In blastocysts of 1.2 to 1.5 mm. the ectoderm, both embryonic
and trophoblastic, has attained its maximum thickness for the
bilaminar stages. Two eggs shown in plate 22 illustrate this
point. Egg No. 353 (4), shown in figures 3, 3A, 3B, 3C measured
1.22 mm. in alcohol. The embryonic area consists of tall cells,
for the most part of the columnar type; the trophoblastic area
is the same as in smaller blastocysts above described. Practically
the same holds true for egg No. 360 (4) (fig. 6), which measured
about 1.3 mm. in diameter (compare stereogram fig. 8, pl. 10).
The wall of a somewhat smaller egg, No. 347’ (1), 1.1 mm. in
diameter in alcohol, has a somewhat thinner embryonic area
(fig. 5, pl. 22). There is, however, considerable variation in this
respect, even within the same litter of eggs of equal size.

In all of the larger blastocysts the entoderm can be followed
as a continuous layer completely lining the vesicle. Except
where pulled away in the preparations, the entoderm fits
closely against the ectoderm and is always distinctly recognizable
at all points (fig. 6, pl. 10).

DEVELOPMENT OF THE OPOSSUM . 85

b: The central light field in the embryonic area

When the diameter of the egg approaches 1.8 mm., certain
changes of importance have taken place, for in such eggs the first
proliferation of mesoderm is usually observed. The eggs of
litters Nos. 189’ and 353’ are of this size; in the last litter meso-
dermal cells occur and the primitive streak is faintly indicated
(M, fig. 21, pl. 13); but the other litter lacks a few minutes of
development to have reached this stage.

The premesodermal changes in the blastocyst are best illus-
trated by a typical and favorably sectioned example, namely,
egg No. 193’ (2), measuring about 1.4 mm. in alcohol. This is
one of the two eggs shown in figures 1 and 2, plate 10. In surface
view there is within the embryonic area, a large light field,
more plainly visible by transmitted light. Such areas have been
described for other mammalian vesicles of a corresponding stage.
They are usually somewhat eccentric, sometimes very con-
siderably nearer one side than the other. I am convinced that
the point where the light field comes nearest the margin of the
area markes the posterior portion of the embryonic area, and I |
have therefore, in figure 9, plate 22, oriented the surface view of
such an egg with the posterior end down. At a slightly later
stage, perhaps an hour later, the primitive streak would have
appeared as a faint tongue-shaped clouding projecting upward
from the lower margin of the formative area into the light field
in question, as in the eggs of litter No. 353’, to be described in
_ the next number of these studies.

The light field is due entirely to a thinning of the embryonic
ectoderm (7, fig. 9A, pl. 22), a condition already seen in small
eggs in figure 6, where at 7’ the area is palpably thinner than at
either end of the section. At 7, figures 4A and 7, the sections
also pass favorably to show this central thinner field. In this
region, too, the nuclei are usually farther separated, whereas at
the margins they are so crowded as to form a continuous chain
like a string of beads, although not quite so uniformly arranged.

86 CARL G. HARTMAN

c. Modified entodermal cells

In many of these larger eggs the entoderm undergoes slight
differentiation at one point. Over the junctional line which
marks the border of the embryonic area there is often a group
of entodermal cells which attract attention by virtue of their
number, the roundness of their nuclei and the volume of the
cytoplasm (HNT, fig. 9A). They are sometimes found in eggs
of 1 mm. (ENT, fig. 5, pl. 21), more often in larger blastocysts
(ENT, figs. 4A, 6 and 7, pl. 22). Similar entodermal cells have
been described by Van Beneden for the bilaminar blastocyst of
the rabbit. He states that they mark the anterior end of the
area and that the future primitive streak appears at the opposite
side. In the preparations made from litter No. 353’, in which
there occurs the first anlage of the primitive streak, these cells
appear in the region where the mesodermal cells are found,
hence not in the anterior, but in the posterior portion of the
embryonic area. I shall treat this subject further at a later
date.

d. The ectoderm of late bilaminar blastocyst

As was stated in the preceding section, the entoderm attains
its maximum thickness in vesicles of about 1.3 mm. diameter
(figs. 3A and 3B, pl. 22). Larger vesicles may have thinner
formative areas or they may remain about the same, although
they are apparently more slender because of their length in
sections. In some of the eggs, especially in litters Nos. 193’
and 343’, the trophoblastic areas are as greatly attenuated as in
the 0.8-mm. stage (figs. 1 and 9A, pl. 22). Sometimes the
trophoblastic area becomes gradually thicker towards the lower
pole (fig. 4, pl. 22), or there may be ectodermal pockets or
‘blisters’ at this point, as described above in connection with
the 1-mm. stage.

As seen in surface view, the distribution of cells is practically
the same as in the 1-mm. blastocysts (figs. 12A and 12B, pl. 22).
All of the cells of the embryonic ectoderm are crowded closely
together and are darker than the trophoblastic cells because

DEVELOPMENT OF THE OPOSSUM 87

they are thicker and uniformly granular, but the cell boundaries
are not as apparent as those of the large flat trophoblastic cells.
The nuclei of the latter are flatter, but of uniform roundness,
unchanged by mutual pressure, and possess fewer chromatin
granules and larger light spaces than the embryonic nuclei,
otherwise the nuclei of the two areas are very much alike. The
entodermal nuclei, as a rule, appear larger in surface view than
those of the ectoderm and they possess a more uniform
granulation.

It should be noted that the embryonic ectoderm is still a
single layer of cells with the nuclei mostly at nearly the same level.
In the corresponding stage of other mammals the embryonic
area is considerably thickened as in a pseudostratified epithelium
(rat, bat). This simple arrangement has the decided advantage
for the observer in that the very first mesodermal nuclei which
drop down out of the ectoderm may be located instantly and
with certainty.

e. Yolk spherules in ectoderm and entoderm

In many surface views of bilaminar blastocysts round dark
objects, usually as large as a nucleus or smaller, frequently meet
the eye. Sometimes these bodies stain like the cytoplasm, or
they may be much darker in preparations fixed in osmic acid.
They are found in both ectoderm and entoderm (figs. 1A and 2A,
pl. 21; fig. 10, pl. 22), sometimes free, sometimes within the
cytoplasm and partly enveloped by the nucleus. The inclusions
are often surrounded by a light zone as though they were partly
digested and absorbed (fig. 3B, pl. 20); indeed, vacuoles, instead
of solid masses, in similar situations are not uncommon (V,
figs. 11 and 12B;pl. 22).

Blastocyst No. 189’ (12) is worthy of special notice. It
appeared to be normal in every respect, and the embryonic
area, which measures 0.75 mm., was dissected off and stained
and mounted intact in balsam. The interesting feature of this
vesicle is the large number of these dark bodies that are mostly
observed in connection with the entodermal cells. The majority
of these cells underlying the embryonic area are each provided

88 CARL G. HARTMAN

with large or small masses, about which the nucleus lies as if
about to engulf it. In figure 11, plate 22, are several typical
cases drawn with the aid of the camera lucida. At A two bodies
are found in connection with a single nucleus; and a similar case
is seen in section of a somewhat younger blastocyst at A,
figure 10. At B, figure 11, the entodermal cell behaves toward
a vacuole as toward a solid mass, a phenomenon by no means rare.
The large cell at C seems to have completely ingested a mass,
the body in the center being not a nucleolus, but a typical
inclusion like those marked Y in the other figures.

I have looked through the whole series of stages from the first
appearance of entoderm to the largest bilaminar blastocyst and
find that the foreign bodies just described are present in nearly
all cases, whether in total preparations or in serial sections. I
am convinced that they are only remnants of undigested yolk.
If one recall the young blastocyst containing 40 or 50 entodermal
cells (litter No. 356) at a stage when the trophoblast has become
considerably attenuated (pl. 17) one notes that the included yolk
is almost entirely confined to the embryonic area. So in suc-
ceeding stages, numerous granules of yolk are found among the
embryonic, seldom within the trophoblastic, cells. More such
granules are, with some exceptions, found in the younger than
in the older blastocysts. Thus, while the albumen melts away
before the embryonic area of the young bilaminar blastocysts,
the yolk granules maintain their identity in small rounded
masses for’a longer time.

f. Mesoderm formation initiated

With the appearance of the first mesodermal cells about six
days have elapsed since ovulation, about five days since the
beginning of cleavage, or about one-half of the period of gesta-
tion, which I am tentatively stating to be ten days in the
opossum. The formation of the mesoderm will be treated in the
next number of these studies.

DEVELOPMENT OF THE OPOSSUM 89

SUMMARY

1. Several thousand eggs were removed from several hundred
pregnant and pseudopregnant animals during the collecting
seasons 1914 to 1917, an average of 11.5 per litter, or 23 from
each animal.

2. At least one-third of the average litter of eggs are unfer-
tilized or abnormal (table 1).

3. Six hundred and forty-one normal eggs, from tubal ova
to the bilaminar blastocyst, form the basis of the present study.

4. Collection of embryological material from the opossum has
become greatly facilitated because of the discovery that the
mammary glands of this animal hypertrophy at the approach
of ovulation, so that the sexual condition of the female may
be predicted with a high degree of certainty, without sacrificing
the animal or without loss of time and effort, by simple though
trained and practiced palpation of the glands. But the behavior
of the mammary glands as well as the other reproductive organs
is the same in the early stages of pseudopregnancy and in
- pregnancy. Ovulation is always spontaneous.

5. A series of photomicrographs of eggs in the living state is
presented in the plates 1 to 11.

The development of ten litters for given periods of time is
shown in plates 1 and'2 and in figures 1 and 4, plate 9.

Plates 12 and 13 are intended to serve as a résumé of the
stages covered by the present study.

The development of the opossum egg is illustrated in one
series by the photographic plates 3 to 10 and in another series
by the drawings plates 14 to 22.

Six hundred different opossum eggs are shown, including 240
illustrations of some 180 different preparations.

6. The rate of development was determined in a number of
cases in which the eggs of the right uterus were allowed to
develop a given period of time after the removal of the left
uterus and its contents. This method contributed in no small
part to the success in securing an unbroken series of stages.

7. The stages covered in this paper comprise about the first

90 CARL G. HARTMAN

half of the ten-day period of gestation. More exact figures
have not as yet been worked out.

8. The first polar body is given off in the ovary, the second
in the Fallopian tube as in other mammals (pl. 14).

9. In the Fallopian tube much albumen and the shell mem-
brane are added to the ovum. It probably requires about
twenty-four hours for the passage of the egg to the uterus.

10. The haploid number of chromosomes in the opossum is
twelve (pl. 14).

11. At no stage in the unsegmented egg is there any evidence
of polarity in the distribution of the yolk, as in Dasyurus, the
bat and some other mammalian eggs, although in the opossum
the yolk is abundantly present (pls. 13 and 14).

12. The egg varies considerably in size, but on the average it
is about 0.12 mm. in diameter through the ovum and 0.6 mm.
through the shell membrane. Some normal eggs attain the
diameter of 0.73 mm., owing to the larger amount of albumen
deposited (compare fig. 2, pl. 4, and fig. 3, pl. 5).

13. As in probably all marsupials, the egg reaches the uterus
unsegmented, hence at an earlier stage than in any of the
Eutheria.

14. The pronuclei at first occupy a yolk-free area at the
periphery of the egg; then migrate to, the yolk-free central
portion, where the first cleavage spindle is later to be seen
(figs. 20 and 21, pl. 14).

15. Deutoplasmolysis or elimination of yolk begins at the
pronuclear stage, continues at the 2-celled stage, and reaches
its maximum during the second cleavage (pls. 3 and 15).

16. The quantity of yolk and surrounding cytoplasm extruded
varies greatly, hence the size of the blastomeres varies in inverse
ratio to the extent of deutoplasmolysis (pl. 15).

17. Deutoplasmolysis occurs by elimination of masses of
various sizes on all sides of the egg, not at any particular spot
or pole, as in Dasyurus and the bat (fig. 4, pl. 3).

18. The two blastomeres of the 2-celled stage are usually of
the same size, shape and‘structure, or they may differ in size.
This difference is probably due chiefly to the difference in the
amount of yolk extruded (text fig. 4).

DEVELOPMENT OF THE OPOSSUM 91

19. One blastomere sometimes anticipates the other in divi-
sion, and as a result 3-celled eggs are found, but not nearly in
as large numbers as eggs in the 4-celled stage.

20. The second cleavage plane is at right angles to the first
and the spindles in the two cells lie parallel; but the shifting of
the blastomeres soon begins, so that in the 3-celled stage the
crossed arrangement may already be attained (K and JL, text
fig. 4).

21. The crossed arrangement of the blastomeres in the 4-celled
egg is, therefore, in the opossum not due to the direction of the
cleavage planes in the second cleavage, but is secondarily caused
by the shifting of the blastomeres (compare B and D, text
fig. 4).

22. The shifting of the blastomeres is not due to mutual
pressure, for in many 4-celled eggs the blastomeres are very
small and not even in contact (figs. 6 and 7, pl. 3).

23. There is no morula stage in the marsupials. The blasto-
cyst cavity is virtually present in the 4-celled egg as the space
between the blastomeres. In the 16-celled stage, or earlier, in
the opossum, the blastomeres have migrated to the periphery
and have applied themselves to the zona pellucida. The
structure of the blastocysts is clearly indicated. The extruded
yolk now lies within the cavity (pls. 4 and 15).

24. The blastocyst wall is usually fully formed at about the
32-celled stage, when all the gaps between the cells are closed by
the flattening and multiplication of the cells of the late cleavage
stage. This marks the end of cleavage as such, which requires
nearly thirty hours of development (fig. 1, pl. 16).

25. During cleavage the only evidence of polarity lies in the
difference in the rate of division among the cells at the two poles.
The more rapidly dividing cells are probably embryonic and
arise from one of the first two blastomeres (text fig. 5).

26.- Definite polarity is established at about the 60- to 70-
celled stage with the first appearance of the entoderm. One
litter at this stage was found six days after copulation, doubtless
a case of retarded ovulation, as a later stage was to have been
expected (pl. 16).

92 CARL G. HARTMAN

27. The entoderm arises from entoderm mother cells of very
characteristic appearance. They are cells in the blastocyst wall
which round up and usually roll out of their place, as it were,
into the blastocyst cavity, as in certain invertebrates, or they
may remain attached to the wall for some time, in either case
multiplying by mitotic division (pls. 7 and 16).

28. The entoderm mother cells all arise from one-half of the
egg, the future embryonic area (figs. 15 to 22, pl. 16).

29. The area that remains free of entoderm mother cells is
the trophoblastic area; it soon begins to thin and spread out so
that the growth of the ovum now begins. Growth is, therefore,
at first due to the spreading of the trophoblastic area (pls. 16
and 17).

30. Since the entodermal cells spring from the superficial
epithelial layer in the embryonic area, this would better be
termed embryonic entectoderm.

31. When the blastocyst has attained a diameter of 0.3 to
0.35 mm., the entoderm is several cells deep, being crowded into
a mass which superficially somewhat simulates an Eutherian
inner cell mass in the process of spreading. In the opossum
only the superficial cells are embryonic ectoderm, all the rest
are entodermal (figs. 13 to 15, pl. 17).

32. Such a stage was removed from an animal four days after
copulation, or about a day and a half after the beginning of
cleavage.

33. The superficial layer of cells is never transitory; it is
embryonic ectoderm and not Rauber’s layer; it is in active
state of mitosis throughout. Rauber’s layer is homologous with
the non-embryonie or, as Hill has expressed it, the ‘tropho-
blastic’ area.

34. The proliferation of entoderm is at an end when the
blastocyst has attained a diameter of 0.45 to 0.5 mm., when the
trophoblastic area has attained its greatest degree of attenuation
(pl. 18).

35. The entoderm now spreads by an active migration of the
flattened, definitive entodermal cells toward the opposite pole of
the egg (pl. 18).

DEVELOPMENT OF THE OPOSSUM 93

36. When the spreading is well under way, the blastocyst,
previously spherical and centrally placed, is usually flattened
like a thick biconvex, lens at one pole of the egg, with the em-
bryonic area in contact with the shell membrane (figs. 1 and 2,
pl. 6; fig. 4, pl. 10; pls:.12 and 19).

37. Eggs in which the entoderm has just become closed at
the lower pole, and which are thus in the beginning of the
bilaminar stage, are still about the same size as in the cleavage
stages, but the albumen has become more dense and the shell
membrane thicker and more resistant (figs. 14 and 17, pl. 13).
The albumen disappears with the growth of the blastocyst and
egg (pl. 13).

38. The bilaminar blastocyst is simply a double-walled sac
consisting of ectoderm without and entoderm within. The two
layers are closely applied to each other and to the shell membrane
and albumen, and any variation from this condition is due to
shrinkage or to abnormality of the egg. There is no ‘perivitel-
line’ space in the normal opossum egg, but frequently occurs in
abnormal material (pls. 19 and 20).

39. The bilaminar stage is the period of growth, little dif-
ferentiation occurring until near the first appearance of mesoderm
in vesicles 1.5 to 1.8 mm. in diameter (pls. 19 to 22).

40. The 1-mm. stage was once found seven and one-half days
after copulation (litter. No. 343, fig. 5, pl. 2); the mesoderm first
appears about eight hours later (compare litters 343’, 346, 346’,
353, 353’); the 0.8-mm. stage was once removed about five days
after the beginning of cleavage (litter No. 306’, fig. 17, pl. 13;
figs. 1 and 1A, pl. 21), and 1.4-mm. blastocysts were found at
about four and a half days after the beginning of cleavage
(compare litters Nos. 191 and 198, figs. 1 and 9, pl. 22).

41. The embryonic area grows in extent with the growth of
the egg, so that in the later bilaminar stage its diameter is about
one-fifth to one-fourth of the circumference of the egg (compare
pl. 18 and figs. 1 to 3, pl. 21, with figs. 1 to 4, pl. 22).

42. As the egg develops, the embryonic area becomes increas-
ingly more sharply set off from the trophoblastic -area (fig. 9,
pl. 21). The embryonic ectoderm becomes thicker, the cells

94 CARL G. HARTMAN

cubical to columnar and more densely granular, whereas the
trophoblastic ectoderm remains flat and its cytoplasm reticular
(pl. 21).

43. The entoderm in the early stages is everywhere the same,
consisting of the typical squamous cells with swellings at the
nuclei. In surface view the flatter entodermal nuclei, as a rule,
appear larger than the ectodermal. The entoderm nowhere
comes to the surface of the blastocyst (pls. 10 and 19 to 22).

44. In blastocysts over 1 mm. in diameter the entoderm is
often modified at one side of the embryonic area. These cells
increase in number, thickness of nuclei, and density of the
cytoplasm, and I believe them to mark the future posterior, not
the future anterior margin of the embryonic area. The primitive
streak will be laid down here (ENT, fig. 9A, pl. 22).

45. These eggs also exhibit a clear field a little to one side of
the center in the embryonic area (figs. 1 and 2, pl. 10; fig. 9,
pl. 22). This is due to a thinning out of the embryonic ectoderm.
Where the light field comes nearest to the margin is the posterior
margin of the area, for here the modified entoderm is also found
(pl. 22).

46. Yolk spherules occur in the bilaminar stage even in the
largest specimens. They are remnants of the extrusions of
cleavage stages. They are found often within the cells, usually
of the embryonic area only, both ectodermal and entodermal,
and the nuclei frequently surround the masses as if to engulf
them (figs. 9B, 10 to 12, pl. 22).

47. The egg of the opossum is like that of Dasyurus in its
possession of a large amount of yolk, in the absence of the morula
stage and in the formation of entoderm from entoderm mother
cells coming from the wall of the unilaminar blastocyst. But it
differs in many regards: in the absence of polarity, in the
uniform distribution of the yolk, and in the consequent manner
of deutoplasmolysis; in the indeterminate type of cleavage; in
the crossed arrangement of the 4-celled egg; in the early period
in which the blastocyst is formed; in the very early formation
of the entoderm; in the simulation of an inner cell mass due to
the crowding of the primitive entoderm cells,—in these respects

DEVELOPMENT OF THE OPOSSUM 95

the opossum egg is much more Eutherian than Dasyurine. Of
these perhaps the most striking feature is very early differentia-
tion of the entoderm. Since in other marsupials, according to
Hill, the entoderm is formed early (Macropus, Perameles) it
seems probable that when the other marsupials have been more
thoroughly studied it will be found that the opossum is more
typical of the marsupials in general and that Dasyurus represents
a more primitive, albeit, therefore, an even more interesting
form.

ADDENDUM

Since the completion of my manuscript there has come to
hand Prof. J. P. Hill’s paper on ‘The Early Development of
Didelphys aurita,’’ published in the April, 1918, number of the
Quarterly Journal of Microscopical Science. This work is based
on eggs secured from six females: one animal furnished unseg-
mented, unfertilized eggs; another numerous 2-, 3-, and 4-celled
eggs; from two others, cleavage stages of 4 to 16 cells were
taken, and from two animals bilaminar blastocysts about 1 mm.
in diameter.

It appears from this contribution that the developmental
stages of the South American opossum, so far as Professor Hill’s
material goes to show, are closely duplicated by my own speci-
mens of the local species. The same is true of a number of
somewhat later stages (primitive streak) which I have secured
from the small black D. marsupialis occurring in south Texas
and Mexico.

In most respects my own work finds full corroboration as well
as interesting extensions in the careful study made by Professor
Hill. I wish briefly to refer to several points discussed by our
able British colleague.

In his analysis of the 2-, 4-, 8-, and 16-celled stages, he presents
additional evidence of polarity in the opossum egg; for he finds
that in a large proportion of such eggs the cells are plainly made
up of two groups, differing somewhat in size, the smaller cells
being considered by Hill as constituting the upper or formative
pole of the egg. He also finds that the majority of eggs in

JOURNAL OF MORPHOLOGY, VOL. 32, No. 1

96 CARL G. HARTMAN

later cleavage show an accelerated rate of division at one pole.
The evidence of polar differentiation in the opossum egg through-
out cleavage seems, therefore, to be complete. Professor Hill
recognizes fully the difference between the cleavage of the
opossum and of Dasyurus and joins me in deriving the formative
and the non-formative areas each from one of the two blastomeres
of the 2-celled egg.

As to the method of deutoplasmolysis, Hill considers that‘‘ the
yolk spheres are budded off from a narrow, clear zone which
has made its appearance at the exposed surfaces of the blasto-
meres” and his ‘‘figures shown undoubted yolk spheres in direct
continuity with the lighter peripheral zone.’”’ I have noted the
same phenomenon, although never as pronounced as in the cases
illustrated by Hill in his figures 11 to 138, plate 8. In most of
my hundred specimens, the smooth, unwrinkled cell membrane
can be followed clearly around the blastomeres. Professor Hill
is correct in assuming that the egg No. 50 (6) upon which I
based my former conclusion on the method of yolk elimination
(Hartman, 716, page 23, and fig. 9, pl. 5) is probably not quite
normal; in fact, the specimen was considerably retarded in
development as compared with its fellows in the 50 to 70-celled
stage. Such retarded eggs are always to be regarded with
suspicion. I therefore no longer regard yolk elimination as due
to the “formation of a new cell membrane, . . . . ata
distance from the original surface of the blastomeres,” but
believe with Professor Hill that masses of variable size are
extruded from different places on the exposed surfaces of the
blastomeres.

On the origin of the crossed arrangement of the blastomeres
in the 4-celled egg, Hill presents evidence, which taken by itself,
would be conclusive of the fact that the blastomeres do not
attain this position by shifting, but assume it from the beginning
by virtue of a meridional division of one blastomere and an
~ equatorial division of the other. For out of his ten 2-celled eggs,
five have both blastomeres in the process of division and in these
the axes of the blastomeres are already nearly or quite at right
angles to each other. Three of these eggs have completed their

DEVELOPMENT OF THE OPOSSUM 97

nuclear division, the nuclei being in the resting stage; the other

two are in late anaphase.

However, two of my own specimens, both of which contain
short spindles in each blastomere, cast doubt upon Hill’s view
as stated above, with which I had agreed before I came into
possession of the eggs (from No. 306). For in one of these eggs
the spindles are exactly parallel; in the other they deviate 36°
from the parallel.

It is, therefore, apparent that, unless we assume a rapid
shifting of the blastomeres during the early phases of the second
cleavage, the matter must for the present remain in doubt.

It is interesting to note that D. aurita has two breeding seasons
a year, whereas D. virginiana has but one.

Professor Hill states it as his belief that twelve is the reduced
number of chromosomes in the opossum, and with this I fully

agree.

LITERATURE CITED

CaLpweLL, W. H. 1887 The embryology of Monotremata and Marsupalia,
Part I. Phil. Trans. Soy. Soc., vol. 178 B.

HarTMAN, Cart G. 1916 Studies in the development of the opossum, Didel-
phys virginiana L., Parts I and II. Journ. of Morph., vol. 27.

Hiuz, J. P. 1910 The early development of the Marsupalia, with special refer-
ence to the native cat (Dasyurus viverrinus). Quart. Jour. Mier.
Sci., vol. 56.

Mrnot, CHarLes R, 1911 Note on the blastodermie vesicle of the opossum.
Anat. Rec., vol. 5.

PaTTERSON, J. P. 1913 . Polyembryonic development in Tatusia novemcincta.
Journ. of Morph., vol. 24.

Sevenka, E. 1887 Studien in der Entwicklungs geschichte der Thiere, Band
4, Das Opossum (Didelphys virginiana). Wiesbaden.

SPURGEON AND Brooks 1916 The implantation andearly segmentation of the
ovum of Didelphys virginiana. Anat. Rec., vol. 10.

Van DER Stricut 1909 Lastructure de l’ceuf des Mammiféres (Chauve-souris,
Vesperugo noctula): Troisiéme Partie., Mem. de |’Acad. roy. de Bel-
gique, IIe ser., t. 2.

Plates 1 to 11 contain 82 figures nearly two-thirds of which are from photo-
graphs of living eggs. Plates 1 and 2 show the development of nine litters for
given periods of time. Plates 3 to 10 are arranged by stage of development, and
the same stages are shown in drawings on plates 14 to 22. Plate 12 was drawn
from 8 specimens as they appeared in the living state. Plate 14 is a résumé of
the stages covered in this paper as drawn from sections.

PLATE 1
EXPLANATION OF FIGURES

Photomicrographs of living eggs in Ringer’s solution; the ten figures show two

litters from each of five different animals; figs. 1 to 9 X 8; fig. 10, natural size.
1 Litter No. 320; 4-celled eggs.

Litter No. 320’; interval 5} days; primitive-streak stage.
Litter No. 299; 4-celled eggs.
Litter No. 299’; interval 4 days, 3{ hours; blastocysts partly bilaminar.
Litter No. 292; young unilaminar blastocysts containing 40 to 50 cells.
Litter No. 292’; interval ‘4 days; primitive-streak stages.
Litter No. 307; tubal ova with a little albumen on one side.
Litter No. 307’; interval 5? days; unfertilized, fragmenting eggs; albu-
men rather opaque. ‘

9 Litter No. 337; 8- to 16-celled eggs.

10 Litter No. 337’; vesicles with primitive streak and very short medullary
groove; photographed in open uterus, natural size.

CON OD Ot e W LO

GENERAL ABBREVIATIONS

A, designates particular portions of HNT?, undifferentiated primitive en

various drawings to which reference todermal cells (pls. 17 and 18)
is made in the text ENT, flattened definitive entodermal
ALB, albumen cells (pl. 16)

ART, artifact
C, coagulum
CH, chromosomes

GR, granulosa cells of discus proligerus
O, ‘blister’ in trophoblastic ectoderm

EMB.A, embryonic area PB, polar body
EMB, ECT, embryonic ectoderm or SM, shell membrane
entectoderm TR.A, trophoblastic or non-embryonic

ENT, definitive eritoderm; in plate 18 aren
it designates the limit of spread of pp ATC trophoblastic ectoderm
entoderm; in plates 21 and 22 it re- ;

F sae : V, vacuole
fers to certain specialized entodermal

X, placed at limits of embryonic area

cells A :
ENTA, entoderm mother cells in wall (junctional line)
of blastocyst (pls. 16 and 17) XX, embryonic area of sections, junc-
ENT',entoderm mother cells that have tional line of surface views
migrated into cavity of blastocyst Y, yolk masses
(pls. 16 and 17) ZP, zona pellucida

98

DEVELOPMENT OF THE OPOSSUM PLATE 1

CARL HARTMAN

PLATE 2
EXPLANATION OF FIGURES

Photomicrographs of living eggs; in Ringer’s solution by reflected light; in
the eight figures two litters are shown from each of four animals; all figures
X 8, except fig. 8, which is X 2.

1 Litter No. 293; 4-celled eggs.

2 Litter No. 293’; interval 3} days; young bilaminar blastocysts.

3 Litter No. 346; bilaminar blastocysts.

4 Litter No. 346’; interval 93 hours; early primitive-streak stage.

5 Litter No. 343; 1 mm., bilaminar blastocysts.

6 Litter No. 343’; interval 7 hours and 20 minutes; bilaminar blastocysts

just preceding first proliferation of mesoderm.

7 Litter No. 298; unilaminar blastocysts of 60 to 120 cells with entoderm
mother cells.

8 Litter No. 298’; interval 34 days; vesicles in opened uterus; primitive
streak and short medullary groove. X 2.

100

DEVELOPMENT OF THE OPOSSUM
CARL HARTMAN

PLATE 3
EXPLANATION OF FIGURES

Photomicrographs of tubal ova and early cleavage stages in uterine eggs.

1 Two eggs of litter No 313, photographed in the living state in Ringer’s
solution by reflected light; considerable albumen has been deposited. 130.

2 Litter No. 351’, photographed in Ringer’s solution by transmitted light;
a small ring of albumen is seen. XX 56.5.

3 Section throught ovum No. 313 (4); 10th section (total in series 19); polar
body shown above; the albumen is darkly stained with Delafield’s haematoxylin ;
yolk stained with osmic acid; Hill’s fluid; 54. X 200.

4 2-celled ovum No. 306 (2); 10th section (total 21); compare C and D, text
fig. 4; Hill’s fluid; 54. X 200.

5 4-celled ovum No. 173 (5); 11th section (total 19); aceto-osmic-bichro-
mate; du. X* 200.

6 and 7 Sections 9 and 14 through ovum No. 299 (5) showing four small
blastomeres and much extruded yolk; Hill’s fluid; 54. X 200.

8 14-celled ovum No. 193 (6); 8th section (total 17); two large cells (of which
one is cut longitudinally in section) have undergone nuclear division; aceto-
osmic-bichromate. X 200.

102

DEVELOPMENT OF THE OPOSSUM PLATE 3
CARL HARTMAN

PLATE 4
EXPLANATION OF FIGURES

Late cleavage stage as illustrated by litter No. 336.

1 The living eggs as photographed in Ringer’s solution by transmitted
light. xX 36.

2 Two eggs of the same litter; the peripheral arrangement of the blasto-
meres is apparent. X 82.

3 32-celled ovum No. 336 (5); blastocyst still incomplete; 9th section (total
20); Flemming; 5u. X 200.

4 30-celled ovum No. 336 (4); incomplete blastocyst; section taken through
middle of ovum (about 19 sections); Flemming; 54. X 200.

5 The eggs of litter No. 336 photographed by reflected light in Ringer’s
solution. X 8.

104

4

PLATE

DEVELOPMENT OF THE OPOSSUM

CARL HARTMAN

PLATE 5
EXPLANATION OF FIGURES

Photomicrographs of late cleavage stages; all figures except 2 and 5 were
photographed in the living state in Ringer’s solution; figs. 1 and 7 by reflected
light; figs. 3 and 4, by transmitted light.

1 Litter No. 314. x8.

2 Ovum No. 314 (2); 13th section (total 19); 28 cells in the incomplete
blastocyst wall and 2 cells in cavity, of which only one cell is shown; the cells
are highly vacuolated; Bouin; 5u. X 200.

3 Several eggs from litter No. 337; about 16-celled stage; the peripheral
arrangement of the blastomeres is well seen; compare fig. 4, below, and fig. 9.
pli. X82:

4 Litter No. 337; one egg with the albumen has been removed from its shell
membrane. X 36.

5 Several eggs in cleavage stages and young blastocysts, stained in Dela-
field’s haematoxylin and photographed in oil of wintergreen; the albumen is
darkly stained in some cases; see accompanying text fig. 6for key. X 16.

f= \-344

VON 3398
\
a

2930) @- © @
es
1)
3396) “e NF ip
ie
Ze
7 @\
(Ge mS] aye 3440)
tS
w= Maa7 0) oe
(@ \ )-2930

Fig. 6. Key to figure 5, plate 5.

6 One egg of litter No. 342 shown in fig. 7; late cleavage stage, about 26 cells.
x 36.
7 Litter No. 842. X 8.

106

DEVELOPMENT OF THE OPOSSUM PLATE 5
CARL HARTMAN

is
Be
4

a
p
fe

PLATE 6
EXPLANATION OF FIGURES

Mostly unilaminar blastocysts.

1 Group of young unilaminar to early bilaminar blastocysts photographed
in oil of wintergreen; fixation mostly by solutions containing osmic acid; for
identification of individual eggs see accompanying illustration, text fig.7. X 16.

2 Group of blastocysts stained in Delafield’s haematoxylin and cleared in
oil of wintergreen; Bouin; see accompanying illustration, text fig. 8 for key.
x 16.

3 Section 13 through ovum No. 88 (7), reconstructed in fig. 21, pl. 16; 18
sections In series; 87 cells, including 5 entoderm mother cells; Hill’s fiuid. > 200

REIS |
fy \ 2 f

356(@)

Z Ya
43903} 356(D
© JLsse
292(3) ASN 7%
= eS Ca:
KN 8 « @ 292
‘t e as \ iS j
SALON “314 05)
(—292¢2)
ars
\ 3144)
)@)
Ye 320% ;
Fig. 7. Key to figure 1, plate 6. Fig. 8. Key to figure 2, plate 6.

4 Just completed 32-celled blastocyst No. 314 (5), 9th section (total 19 sec-
tions); Hill’s fluid;-54; compare fig. 1, pl. 5, and fig. 2 above. X 200.

5 Litter No. 292, photographed alive in Ringer’s solution. X 8.

6 Section 10 through the 46-celled, incomplete blastocyst No. 292 (4), also
seen in fig. 2 above; 19 sections in series; no entodermal cells; Hill’s fluid; 5y.
x 200.

7 and 8 Ovum No. 298 (1); fig. 7, whole egg, X 30, in aleohol (compare fig.
7, pl. 2); fig. 8, 6th section showing column of entoderm mother cells; 21 sections
in series; 64 cells, including 4 entoderm mother cells; Hill’s fluid; 54. X 200.

108

PLATE 6

DEVELOPMENT OF THE OPOSSUM

CARL HARTMAN

PLATE 7
EXPLANATION OF FIGURES

Mostly blastocysts with entoderm mother cells.

1 Portion of 13th section through ovum No. 50 (5), showing its only entoderm
mother cell; total 21 sections; 65 cells; Bouin; 5u. XX 500.

2 Portion of 13th section through ovum No. 88 (17), taken as indicated by
parallel lines on fig. 18, pl. 16; characteristic entoderm mother cell is shown;
total 18 sections; 103 cells, of which 6 are entoderm mother cells; Hill’s fluid;
Su. X 500.

3 Section through 6 of the 9 entoderm mother cells of ovum No. 50 (4); 16th
section (total 20 sections); 67 cells; Hill’s fluid; 5u. X 200.

4 The 15th section through ovum No. 88 (16), of which the 11th section is
shown in fig. 11, pl. 16; total 20 sections; large entoderm mother cell is shown in
blastocysts wall; 82 cells, including 10 entoderm mother cells; Bouin; 5u. X 500.

5 23-celled incomplete blastocyst No. 173’ (7); 13th of a total of 23 sections;
aceto-osmic-bichromate; 5u. * 200.

6 The 12th section through ovum No. 88 (3), showing several entoderm
mother cells and much yolk and coagulum; 23 sections in series; 69 cells, of which
10 are entoderm mother cells; Bouin; 5u. X 500.

110

DEVELOPMENT OF THE OPOSSUM PLATE 7
CARL HARTMAN

PLATE 8
EXPLANATION OF FIGURES

Photomicrographs showing progress in entoderm formation and polar differ-
entiation; figs. 1, 3, 5, and 6, photographed in the living state in Ringer’s solu-
tion by transmitted light.

1 An egg of litter No. 356, showing opaque embryonic area and thin tropho-
blastic area. XX 82.

2 Blastocyst No. 356 (7); 16th section (total 32); Flemming 54; X 200; com-
pare with fig. 1.

3 One of two identical eggs from litter No. 349; opaque embryonic area to
the left; compare fig. 12, pl. 17. X 36.

4 Egg No. 349 (5), the least developed egg from litter No. 349; polar differ-
entiation is well under way; retarded in development as compared with fig. 3;
14th section (total 25); Bouin, 54. X 200.

5 Egg No. 344 (7), lateral aspect, as seen alive; the longitudinal section of
this egg is shown in fig. 7. x 82.

6 Egg No. 344 (8), as viewed with opaque embryonic area uppermost. X 82.

7 Egg No. 344 (7), in section, seen alive in fig. 5; 10th section (total 23) ;
Flemming; 5 pl. X 500.

DEVELOPMENT OF THE OPOSSUM PLATE 8
CARL HARTMAN

PLATE 9
EXPLANATION OF FIGURES

Photomicrographs of late unilaminar and young bilaminar blastocysts; figs.
1, 4, 5, and 6, photographed alive in Ringer’s solution; the first by transmitted
light, X 36, the last three by reflected light, X 8.

1 Litter No. 352, blastocysts with bilaminar embryonic area (indicated by
dark region at one pole of the vesicle); trophoblastic area very attenuated;
compare fig. 5, pl. 12.

2 Section through embryonic area of egg No. 352 (11); 44th section of egg
(total 96); 26th section of blastocyst (total 66); 16th section of embryonic
area (total 34); entoderm has begun to migrate beyond area; ,half-strength
Bouin; 54. X 100.

3 Section 12 of ovum No. 356 (5), of which sections 13 and 18 are shown in
figs. 10 and 11, pl. 17; total 21 sections; Bouin; 64. X 500.

4 Litter No. 352’; interval 15 hours; compare fig. 1; young bilaminar blasto-
cysts; one egg has two blastocysts.

5 Litter No. 347; partially and entirely completed bilaminar blastocysts. °

6 Litter No. 339, a little younger than litter No. 347, shown in fig. 5.

114

DEVELOPMENT OF THE OPOSSUM PLATE 9
CARL HARTMAN

PLATE 10
EXPLANATION OF FIGURES

Photomicrographs of bilaminar blastocysts.

1 and 2 Two eggs of litter No. 193’, photographed unstained in alcohol by
transmitted light; note light field near center of embryonic area; compare fig.
GU 22 ew a3

3 1.15-mm. blastocyst No. 360 (5), stained in Delafie!ld’s haematoxylin, cut
in two horizontally and photographed in alcohol by transmitted light. = 16.

4 Egg No. 352’ (10), one of the litter shown in fig. 4, pl. 9; 46th section of
vesicle (total 92); 29th section of embryonic area (total 54); section is oriented
with embryonic area to left; Hill’s fluid; 5u. X 100.

5 Egg No. 299’ (6); one of litter shown in fig. 4, pl. 1; shown in toto in fig. 1,
pl. 6; 61st section of vesicle (total 120); 41st section of embryonic area (total 81);
the entoderm has not quite reached the lower pole; Hill’s fluid; 54. X 100.

6 Section of egg No. 360 (5) shown in surface view in fig. 3; section is ori-
ented with embryonic area uppermost; a little albumen is left at lower pole;
55th section of embryonic area (total 121). > 100.

7 Litter No. 306’; interval 5 days, 20} hours, after beginning of cleavage;
photographed alive in Ringer’s solution. X 8.

8 Stereogram of ova No. 360 (7), (8), and (9); photographed in alcohol;
embryonic area seen in two of the eggs; Zenker. X 6.3.

116

PLATE 10

DEVELOPMENT OF THE OPOSSUM
CARL HARTMAN

PLATE 11
EXPLANATION OF FIGURES

Photomicrographs taken alive by reflected light in Ringer’s solution; except
fig. 2. Figs. 1 to 4, bilaminar blastocysts; figs. 5 to 10, unfertilized eggs. Mag-
nification, X 8, except fig. 5.

1 Litter No. 294, mostly abnormal blastocysts with still largely unilaminar
walls (compare fig. 2, pl. 6).

2 Part of litter No. 290’, photographed a few minutes after immersion in
Hill’s fixing fluid; embryonic area is well seen.

3 Litter No. 290’; bilaminar blastocysts (compare fig. 2, pl. 6).

4 Litter No. 294’ (interval 341 hours; compare fig. 1); mostly abnormal
bilaminar blastocysts.

5 Litter No. 415; unfertilized eggs in early stage of fragmentation; shows
false ‘1-celled,’ ‘2-celled,’ and ‘4-celled’ eggs.

6 Litter No. 318; early eggs in fragmentation; note that the ovum proper is
no longer spherical.

7 Litter No. 303, with opaque albumen and fragmenting ova.

8 Litter No. 297; old fragmenting eggs with white concretions on shell mem-
brane.

9 Four degeneration eggs that accompanied foetuses one day before birth;
Litter No. 321’.

10 Litter No. 332; degenerating eggs nine or ten days old.

118

PLATE 11

CARL HARTMAN

PLATE 12
EXPLANATION OF FIGURES

Résumé of stages in the development of the opossum eggs from cleavage to
the completed bilaminar blastocysts; drawn from actual specimens cleared in
oil of wintergreen, the measurements being made from photographs of the living
egg. X 50.

1 The unsegmented uterine egg.

2 The 4-celled ovum.

3 The just completed unilaminar blastocyst of about 32 cells.

4 Blastocyst with polar differentiation well under way; primitive entoderm
present; drawn after No. 356 (6) (compare fig. 2, pl. 6).

5 Blastocyst with attenuated unilaminar trophoblastic area, bilaminar only
in the embryonic region (after litters Nos. 194’, 175’, and 352).

6 More advanced blastocyst with spreading entoderm; after No. 290 (4),
photographed in fig. 2, pl. 6; the flattened shape of the vesicle is the usual one at
this stage.

7 Similar stage with more unusual spherical blastocyst, drawn after No. 299’
(2), shown photographically in fig. 2, pl. 6.

8 Completed bilaminar blastocyst, drawn after No. 299’ (1), shown photo-
graphically in fig. 2, pl. 6.

120

DEVELOPMENT OF THE_OPOSSUM
CARL G. HARTMAN

121

JOURNAL OF MORPHOLOGY, VOL. 32, No. 1

PLATE 13
EXPLANATION OF FIGURES

Résumé of stages in the development of the opossum egg as drawn from sections
of representative specimens from the ovarian egg to the primitive streak stage.
x 50.

1 Ovarian egg from specimen No. 28; Hermann’s fluid.

2to 5 Respectively the following tubal ova: No. 76 (8), Hill’s fluid; No.
56 (4), Bouin’s; No. 351 (1); Hills; No. 313’ (1), Bouin’s.

6 Unsegmented uterine egg No. 287 (1); Hill’s fluid; 5 wu.

7 2-celled ovum No. 203 (4); thirteenth section (total 20); 5 yu.

8 4-celled ovum No. 203 (5); tenth section (total 21); Hill’s fluid; 5 p.

9 14-celled ovum No. 193 (6); eighth section (total 17); aceto-osmic-
bichromate.

10 Just completed blastocyst No. 191 (2); eleventh section (total 18); 32
cells; Bouin; 6 yu.

11 63-celled blastocyst No. 50 (8), of which one of the two entoderm mother
cells is shown in fig. 5, pl. 16; ninth of 17 sections; Hill’s fluid.

12 Egg No. 356 (4), the seventeenth section (total 25); compare figs. 6 and 7,
oll, I/3 (0 ja.

13 Blastocyst No. 194’ (3); ninth section of vesicle (total 34); fifth section
of embryonic area (total 13); aceto-osmic-bichromate (?); 5 uw.

14 No. 352 (7); detail invfig. 8, pl. 21, q. v.

15 Egg No. 339 (5); thirtieth section of blastocyst (total 65); not perfectly
normal; one-half strength Bouin; 5 pw.

16 No. 299’ (5), shown in toto in fig 1, pl. 6; sixty-second section of egg
(total 122) and thirty-fifth section of blastocyst (total 83); Hill’s fluid; 5 uz.

17. No. 306’ (2), also shown in fig. 2, pl. 20, q. v.

18 Bilaminar blastocyst No. 189 (6), the embryomie area of which is shown in
fig. 4 pla21> qv.

19 Bilaminar blastocyst No. 55 (19), showing (at right) a mass of cells at
lower pole sixty-sixth section of egg (total 139); twenth-fourth section of em-
bryonic area (total 77); Bouin.

20 Bilaminar blastocyst No. 189’ (10) approaching time of mesoderm forma-
tion; 130th section of egg (total 282); ninety-third section of embryonic area
(total 165); vesicle wall very thin; egg slightly damaged.

21 Eeg No. 353’ (6), blastocyst with about 140 mesodermal cells; 1.6 mm.
in aleohol with embryonic area 1.1 mm.; ninety-second section (total 205); M.
mesodermal cells, Bouin; 6 pw.

22 Eee No. 346’ (6); section taken through primitive steak; 1.5 mm. in
diameter in aleohol; embryonic area 1.1 mm.; Bouin; 5 «; compare fig. 4, pl. 2.

122

DEVELOPMENT OF THE OPOSSUM PLATE 13
CARL G, HARTMAN

ee EMB.ECT

4
<a
:

PLATE 14
EXPLANATION OF FIGURES

Maturation and fertilization.

1 Large ovarian egg with discus proligerus, from No. 21;5 4. X 200.

2 Ninth of 23 sections through tubal ovum No. 307 (1); polar body is in
nineteenth section; egg is surrounded with thin albumen layer; Bouin; 5 pu. —
x 200.

3 Portion of fig. 2; 7 chromosomes are seen; 5 yu. X 500.

4 Tenth section of same egg; 5 chromosomes; zona pellucida is a beaded
line with darkly staining granules; 5 y.

5 and 6 Eleventh section (total 20 sections) through ovum No. 307 (3); the
second maturation spindle has 12 chromosomes; Bouin; 5 u. X 200 and X 500,
respectively.

7 Sketch of ovum No. 56 (11), from total preparation drawn with focus on
middle of egg; chromosomes and polar body; Bouin. X 200.

8 Sixth section through ovum No. 76 (1); total 23 sections; 6 chromosomes of
this section are in ovum and one in polar body; Bouin; 5 yu. X 200.

9and10 Second maturation spindle of egg No. 307 (2); tenth and eleventh
sections (total 23 sections); 12 chromosomes; Bouin; 5 4. X 500.

11 Portion of ninth section of ovum No. 76 (8); total 15 sections; 12 chromo-
somes in homogeneous granular area; little albumen at left; Hill’s fluid. X 500.

12 Ovum No. 313 (2); eleventh section (total 20); marginal granular zone
limited within by reticulated region; oil globules of medium size; a little albumen
at left; Hill’s fluid; 5 yu. X 200.

13 Ovum 76 (6); composite of fifth and sixth sections (total 21); polar body
and short spindle with 7 chromosomes 5 »; Bouin. X 2500.

14 Ovum No. 56 (6); portion of fourth section (total 17); there are 12 chro-
mosomes; Bouin. XX 500.

15, 16, and 17 Ovum No. 76 (4); 2nd, 3rd and 4th sections tangentially (total
25 sections); polar body and equatorial plate of maturation spindle; marginal
granular zone, vacuoles and oil globules; Hill’s fluid; 54. X 500.

18 Ovum No. 313 (5) from same litter as fig. 12; 5th section (total 20); large
fat globules; albumen layer thick (compare fig. 1, pl. 3).

19 Young unfertilized uterine ovum, No. 287 (5); 9th section (total 16);
concentric lamellae of albumen; compare fig. 12; Hill’s fluid; 54. X 200.

20 Ovum No. 203 (1) with pronuclei; 10th section (total 20); 54. X 200.

21 Ovum No. 52 (3); composite of sections 12 to 15 (total 20) taken obliquely
through first cleavage spindle; a little yolk has been extruded (Y); Bouin; 5 pz.
x 200.

DEVELOPMENT OF THE OPOSSUM PLATE 14
CARL G, HARTMAN

SEO Gg S89 eS GS
oa” bs 25%

oo Bed Oo 82

j Op ©

©

PLATE 15
EXPLANATION OF FIGURES

Cleavage stages; all magnifications X 200.

1 2-celled ovum No. 203 (8), 11th section (total 20); compare figs. E and F,
text fig. 4; Hill’s fluid; 5 yu.

2 2-celled ovum No. 306 (1); 9th section (total 23); compare figs. A and B,
text fig. 4; Hill’s fluid; 5 p.

3 3-celled ovum No. 173 (8); 10th section (total 21); compare fig. L, text 4;
aceto-osmic-bichromate; 5 yu.

4 3-celled ovum No. 306 (3); 11th of 22 sections; compare K, text fig. 4; Hill’s
fluid; 5 wu.

5 4-celled ovum No. 293 (2), with two blastomeres in mitosis; drawn from
clay model; total 18 sections, Bouin; 5 1; compare fig. 1, pl. 2.

6 4-celled ovum No. 293 (4), with two cells in mitosis; drawn from clay
model; Bouin; 18 sections; 5 4; compare fig. 1, pl. 2.

7 4-celled ovum No. 203 (7); 10th section (total 20); see text; Hill’s fluid; 5 yu.

8 4-celled ovum No. 83 (7); 11th section (total 18); blastomeres cut as in
* preceding; one polar body; Bouin; 5 wp.

9 and 10 5th and 9th sections through ovum No. 17’ (7) one of 438 very
small eggs from one ovary; total 11 sections.

11 and 12 7th and 14th sections (total 19) through ovum No. 83 (8), with
portions of shell membrane and albumen; two polar bodies; Bouin; 5 u.

13 4-celled ovum No. 299 (7), also shown reconstructed in 0, text fig. 4; 9th
section (total 20); trichloracetic; 5 pu.

14 4-celled ovum No. 299 (5), of which sections 9 and 14 are shown in figs. 6
and 7, pl. 3; 13th section (total 22); Hill’s fluid; 5 u.

15 6-celled ovum No. 85 (5); 7th section (total 16); blastocyst formation
already anticipated; Bouin.

16 15-celled ovum No. 337 (1); 8th section (total 19); one-half strength
Bouin; 5 zu.

17. 16-celled ovum No. 85 (12); 8th section (total 21) ; blastomeres still rounded;
one misplaced cell; Hill’s fluid; 5 uw.

18 17-celled ovum No. 336 (1); 9th section (total 18); Bouin; 5 z.

19 26-celled ovum No. 336 (2); 11th section (total 17); Bouin; 5 z.

20 28-celled ovum No. 342 (1); 8th section (total 18); half-strength Bouin;
45) [te

PLATE 15

DEVELOPMENT OF THE OPOSSUM

CARL G, HARTMAN

127

PLATE 16
EXPLANATION OF FIGURES

The formation of entoderm initiated. All magnifications are X 200, except
figs. 5 to 10 which are * 500.

1 Completed blastocyst No. 191 (5); 34 cells; 9th section (total 24); Bouin;
5 1

2 Large blastocyst No. 50 (7); 70 cells, but no entoderm; 11th section
(total 23); Hill’s fluid.

3 Half-normal blastocyst No. 88 (18); 8th section (total 19); Hill’s fluid; 5 p.

4 Half-normal ovum No. 344 (12); 6th section (total 16); half-strength
Bouin; 5 wu.

5 Portion of 6th section (total 17) through ovum No. 50 (8) showing entoderm
mother cell ENTA; 63 cells including 2 entoderm mother cells in wall; Hill’s
fluid; compare fig. 11, pl. 13.

6 Portion of 8th section (total 22) of ovum No. 298 (5), showing entoderm
mother cell leaving its place in blast. wall; 126 cells, of which 8 are free entoderm
mother cells and several are in process of formation; Bouin, 5; ef. fig. 7 pl. 2.

7 Portion of 6th section (total 22) of ovum No. 88 (9); 106 cells of which 11
are free entoderm mother cells; Hill’s fluid; 5 u.

8and9 Portions of the 10th and 12th sections (total 18) through ovum No. 88
21); 70 cells, including 10 more or less detached entoderm mother cells; Hill’s
fluid.

10 Greater part of 4th section (total 15) through ovum No. 88 (23); 57 cells
including the two detached entoderm mother cells here shown; Bouin.

11 Blastocyst No. 88 (16), having 82 cells; 6 of the 10 entoderm mother cells
are here shown; 11th section (total 20); section 15, fig. 4, pl. 7; Bouin, 5 yp.

12 Ovum No. 838 (5), containing 53 cells, including the one large binu-
cleated entoderm mother cell (?) here shown; 9th section (total 20).

13° Ovum No. 356 (3), most retarded member of litter No. 356;about 100
cells; 12th section (total 20); Bouin; 5 un.

14 Ovum No. 344 (4); small blastocyst with numerous entoderm mother cells;
8th section (total 18); Hill’s fluid; 5 wu.

15 Longitudinal section of ovum No. 344 (11), showing definite polar differ-
entiation; typical entoderm mother cells; 7th section (total 19); half-strength
Bouin; 164 cells:

Embyonic ent-ectoderm.............. 71-cells of which 7 are in mitosis
Trophoblgstic ectoderm.............. 70 cells of which 8 are in mitosis
mC Od erie 2.035 ine. een net arian eA Ce 23 cells of which 3 are in mitosis

16and17 The 7th and the 16th sections (total 22) taken horizontally through
ovum No. 344 (14), slightly more advanced than preceding; fig. 16, through
embryonic area; fig. 17, through trophoblastic area; half-strength Bouin; 5 «; 193
cells:

Hmbryonic ent-ectoderm: 2-0. 5. 2000s ane se oe 76 cells, 8 in mitosis
rophoblasticgechod ernie er rane 98 cells, 13 in mitosis
HO COR IIE Franson RS eR ne ne A 19 cells, 1 in mitosis

18 to 22. Reconstructions from blastocysts to show the polar distribution of
entoderm mother cells. Fig. 18, ovum No. 88 (17), 103 cells, of which 6 are
entoderm'mother cells; section indicated by parallel lines is shown in fig. 2, lee
Fig. 19, ovum No. 83 (1), 111 cells, of which 4 are free entoderm mother cells.
Fig. 20, ovum No. 298 (3), 124 cells, of which 4 are entoderm mother cells. Fig.
21, ovum No. 88 (7), 87 cells, including 5 entoderm mother cells; the section in-
dicated by lines is shown in fig. 3, pl. 6. Fig. 22, ovum No. 88 (11), 103 cells, of
which 9 are more or less free entoderm mother cells and 7 of these are in mitosis.

128

DEVELOPMENT OF THE OPOSSUM PLATE 16
CARL G, HARTMAN

PLATE 17
EXPLANATION OF FIGURES

The formation of entoderm (concluded).

1 The 7th section longitudinally through ovum No. 144’ (1); ENT!, dividing
entoderm mother cell; 16 sections in series; overfixed in Carnoy. X 200.

2 The 14th section longitudinally through ovum No. 144’ (8); 19 sections in
series; Carnoy. X 200.

3 Section taken tangentially through embryonic area of ovum No. 144’ (10);
14th section (total 19); Carnoy. X 200.

4 and 5 Sections 9 and 11 (total 25) cut longitudinally through blastocyst
No. 356 (11); Hill’s fluid; 5 4; ENT’, primitive entoderm cell tending to flat-
ten out; HNT,? row of entoderm mother cells similar to those in fig. 3, pl. 6.
x 200. ’

6 and 7 Details of ovum No. 356 (4) shown in fig. 12, pl. 18. Fig. 6, 10th
section (total 25), X 200; fig. 7, 16th section, X 500, with spermatozoa in albumen
layer; mitosis In embryonic entectoderm; Bouin; 6 ; 283 cells:

Embryonic ent-ectoderm...5.......5..4:.605.:- 101 cells, in mitosis 3
simoploplasiChechod erie: sneer nee rene 140 cells, in mitosis 8
LBA R(oxo Kone age Aerts ck ret ae ere I Re te Sree Sons hata. 42 cells, in mitosis 2

S8and9 Portions of sections 10 and 17 (total 29) longitudinally through ovum
No. 356 (9), shown whole in fig. 1, pl. 6; Flemming; 5 yp. X 500.

10 and 11 Sections 18 and 13, respectively (total 21), longitudinally through
ovum No. 356 (5), section 12 of which is shown in fig. 3, pl. 9; mitoses in ent-
ectoderm; Bouin; 6 w; 249 cells:

Limibynonickent=ectodermesss:.) seen ee eee 75 cells, in mitosis 14
Trophoblastic’ectoderm. ......¢. ....... osc. see aso. 126 cells; an! mitasiaen
) Baal oY Kenta eee ren reh in ota ears Cm eae, Pemaans Soar: 48 cells, in mitosis 4

12 Longitudinal section of egg No. 349 (2) like the one shown in living
stage is fig. 3, pl. 8; 31st section through vesicle (total 48); 16th section through
embryonic area (total 28); Bouin;5u. X 200.

13. Longitudinal section through ovum No. 194’ (4); 7th section through
embryonic area, (total 20); Hill’s fluid; 74. X 200.

14. The 9th of a total of 13 sections through the embryonic area of ovum
No. 194’ (8). X 500.

15 The 12th of a total of 20 sections through the embryonic area of blastocyst
No. 194’ (6); 38 sections through vesicle; 74. X 500.

130

DEVELOPMENT OF THE OPOSSUM PLATE 17
CARL G. HARTMAN

Ss

SEE E TONED

z

TRECT

EMBECT

PLATE 18
EXPLANATION OF FIGURES

Stages from the spreading of the entoderm to the just completed bilaminar
blastocyst. Whole sections (figs. 5A, 6A, 7A) X 50; vesicles oniy & 200; ENT,
limits of distribution attained by the entoderm; ENT°, undifferentiated primi-
tive entoderm not yet spread.

1 Blastocyst No. 48 (7); 10th section through embryonic area; Bouin.

2 Blastocyst No. 352 (12); 42nd section through egg (total 98), 33d, section
through vesicle (total 80), and 18th section through embryonic area (total 50);
half-strength Bouin; 5 uw; compare fig. 1, pl. 9.

3 Blastocyst No. 40 (1); 15th section through embryonic area (total 30);
Hill’s fluid.

4 Blastocyst No. 40 (2); the 19th section through embryonic area (total 36);
Carnoy.

5A and 5 Blastocyst No. 347 (2); earliest stage of the completed bilaminar
blastocyst; 22nd section through vesicle (total 57); Bouin; 7 u; compare fig. 5,
pl. 9.

6A and6 Blastocyst No. 339 (8). Fig. 6A, 66th section of vesicle (total 94)
and 37th section of embryonic area (total 51); fig. 6, 68th section of vesicle; Bouin;
5 w; compare fig. 6, pl. 9, and fig. 2, pl. 6.

7A Egg No. 347 (1); 52nd section through egg (total 127); 46th section through
vesicle (total 103); entoderm spread to equator.

7 Blastocyst No. 347 (4); nearly the same stage as fig. 7A; 32nd section
through vesicle (total 90); Bouin; 7 x.

8 Blastoeyst No. 175’ (2); 9th section through embryonic area (total 25)
and 30th section through vesicle (total 56); aceto-osmic-bichromate; 6 u.

132

DEVELOPMENT OF THE OPOSSUM PLATE 18
CARL G. HARTMAN

JOURNAL OF MORPHOLOGY, VOL. 32, No. 1

PLATE 19
EXPLANATION OF FIGURES

Partially completed and just completed bilaminar blastocysts.

1 Longitudinal section through ovum No. 205 (4); 41st section (total 93);
Bowing 0)

2 Surface view of trophoblastic area of an egg from the litter No. 205; ento-
derm shaded; ectodermal nuclei unshaded. > 500.

3 Detail of embryonic area of blastocyst No. 347 (4), shown in fig. 7, pl. 18.
x 500.

4 Detail of section through embryonic area of ovum No. 352’ (10), shown
in’ fie. A pl. 10-5500:

5and 5A Entire section, 50, and vesicle only, 200, through the middle
of ovum No. 339 (4), photographed in toto in fig. 2, pl. 6; note swollen cells;
Bouin; 5 wp.

6 Embryonic area only of similar egg No. 339 (2); Bouin; 5 4. X 200.

7 Blastocyst No. 175’ (9), with very attenuated, mostly unilaminar wall;
52nd section through vesicle (total 89); aceto-osmic-bichromate (?);5 4. X 50.

7a Embyronic area only of same egg. X 200.

Sand 8A_ Entire section, X 50, and embryonic area (XX), X 200, of ovum
No. 347 (5); entoderm has not yet reached equator; 30th section of vesicle (total
119) and 18th section through area (total 29); Flemming; 5 yu.

9 The embryonic area of ovum No. 205 (6); 49th section of blastocyst (total
79); Bouin; 5 uw; compare fig. 1.

10 Just completed bilaminar blastocyst No. 208 (1); 56th section through
vesicle (total 117); Bouin. X 50.

10A and 10B_ Details of embryonic area and trophoblastic area of same egg.
xX 200.

11 Portion of surface view of ovum No. 205 (7); XX, junctional line; only the
entorderm is shaded; ectodermal nuclei unshaded circles. 500.

12 Surface view at junctional line (XX) of ovum No. 205 (9); entire ecto-
derm shaded; embryonic nuclei very dark, trophoblastic nuclei very light;
entodermal nuclei intermediate in tone.

134

DEVELOPMENT OF THE OPOSSUM PLATE 19
CARL G. HARTMAN

135

PLATE 20
EXPLANATION OF FIGURES

Completed bilaminar blastocyst.

1 Section of embryonic area (XX) of blastocyst No. 82 (13), nearly like
fig. 1, pl. 21; 54th section of egg (total 100) and 38th section of embryonic area
(total 58); Bouin. X 200.

1A Detail of same. X 500.

2 Blastocyst No. 306’ (2) shown in fig. 17, pl. 13; 0.77 mm. in diameter in
alcohol; 18th section of embryonic area (total 49) and 63d through vesicle (total
118); compare fig. 7, pl. 10; Hill’s fluid; 54. X 200.

2A Detail of same near junctional line (X). X 500.

3 Ovum No. 285’ (1); 72nd section through blastocyst (total 107); Bouin;
5 uw; the portion of vesicle marked by dotted line was dissected off before in-
bedding and was stained and mounted in toto (fig. 3B). X 50.

3A Same blastocyst. X 200.

3B Surface view from point A, fig. 3; junctional line XX; ectoderm lightly
shaded; entodermal nuclei darkly shaded. 500.

4. The 92nd section (total 140) through blastocyst No. 48 (10); two de-
generating cells at A; Hill’s fluid. X 50.

44 The embryonic area (XX) of same. X 200.

136

DEVELOPMENT OF THE OPOSSUM
CARL G, HARTMAN

~ex Of

[OF

sor)

Tacs

137

PLATE 20

PLATE 21
EXPLANATION OF FIGURES

The 1-mm. bilaminar blastocyst.

1 Egg No. 306’ (3); 0.85 mm. in diameter in alcohol; 44th section of embryonic
area (total 70) and the 84th section of vesicle (total 135); Bouin; 5u. X 50.

1A Embryonic area (XX) of same section. X 200.

2 Egg No. 55 (20); 0.87 mm. in diameter in alcohol; 57th section of vesicle
(total 121) and 33d section of embryonic area (total 77); O, pocket in tropho-
blastic ectoderm; Flemming; 64. X 50.

2A Embryonic area (XX) of same section; at A, unusual crowding of ecto-
derm. XX 200.

3 Egg No. 339’ (3); 0.85 mm. in alcohol; 64th section of vesicle (total 118)
and 45th section of embryonic area (total 71); 0, pocket in ectoderm; half-
strength Bouin; 64. X 50.

4 Embryonic area (XX) of egg No. 189 (6), shown in fig. 18, pl. 13; 74th
section of embryonic area (total 94); 1.02 mm. in alcohol; Hill’s fluid; 5 u. X 200.

5 Embryonic area (XX) of egg No. 343 (4), about 1.0 in alcohol (compare
fig. 5, pl. 2, and fig. 9 below); 99th section of vesicle (total 189) and 67th of
embryonic area (total 126); Bouin; 5 yu. XX 200.

6 <A portion of trophoblastic area of a 1.0 blastocyst No. 55 (6), showing
remnant of albumen; compare fig. 10, pl. 22; Hill’s fluid. > 200.

7 and 7A Sections through embryonic and trophoblastic areas of ovum
No. 285’ (6); 65th section of vesicle (total 119) and 26th through embryonic area
(total 65); Hill’s fluid; 6. X 200.

8 Embryonic area (XX) of ovum No. 352 (7); 35th section of egg (total 80);
24th section of vesicle (total 49) and 13th section of embryonic area (total 21);
in aleohol egg measured 0.585 mm. through shell membrane and 0.325 * 0.370
through vesicle; 64; Bouin. X 200.

9 Surface view of a typical 1 mm. blastocyst, showing embryonic area;
compare fig. 2, pl. 11, and fig. 3, pl. 10. X 16.

10. Half-normal blastocyst No. 314 (3); Bouin; 5 4. X 200.

138

PLATE 21

DEVELOPMENT OF THE OPOSSUM

CARL G. HARTMAN

aw Big)

~~ <— s,

KSA so =

139

PLATE 22
EXPLANATION OF FIGURES

Advanced bilaminar blastocysts, to the beginning of mesoderm proliferation.
Figs. 2, 4, and 9 represent eggs only a few minutes removed from the first appear-
ance of mesoderm.

1 Egg No. 193’ (4), similar to fig. 9 below; a 92nd section of egg (total 184)
and 55th section of embryonic area (total 87); Hill’s fluid; 7. X 50.

2. Egg No. 343’ (2), one of the five shown in fig. 6, pl. 2; 1.5 mm. in alcohol;
embryonic area, 0.87 mm.; 130th section of vesicle (total 259) and 89th section
through embryonic area (total 161); Bouin; 5 yu. X 50.

3 Egg. No. 353 (4); several hours preceding first appearance of mesoderm;
diameter 1.22 mm. in alcohol; 91st section of vesicle (total 135) and 53rd section
of embryonic area (total 85); Flemming; 6. 50.

3A and 3B Details of trophoblastic and embryonic areas, respectively, of
same section. X 200.

38C A detail of fig. 3B. X 500.

4and4A Egg No. 189’ (1); 118th section of vesicle (total 200) and 62nd section
of embryonie area (total 116); aceto-osmic-bichromate. X 50 and X 200.

5 Embryonic area (XX) of egg No. 347’ (1); 1.1 mm. in alcohol; 74th section
of vesicle (total 126) and 55th section of embryonic area (total 82); Flemming;
Napa 200:

6 Thick embryonic area (XX) of egg No. 360 (4); 71st section of vesicle (total
176) and 41st of embryonic area (total 130); 7, thinning near middle; Hill’s fluid;
5 pu. X 200.

7 The 79th section through embryonic area (total 172) of egg No. 189’ (9).
X 200.

8 Large embryonic area of egg No. 189’ (4); 157th section of egg (total 311)
and 59th section of area (total 169); Hill’s fluid; 54. XX 200.

9 Drawing made from photograph of an egg in litter No. 193’; shows central
light field in embryonic area; compare figs. 1 and 2, pl. 10. X 16.

9A Section of ovum No. 193’ (2); 61st section of embryonic area (total 160).
xX 200.

9B and 9C_ Details of portions of fig. 9A. X 560.

10 Portion of embryonic area of 1-mm. blastocyst No. 55 (6), showing yolk
granules (Y) in ectoderm, and entoderm; compare group A with A, fig. 11.

11 Surface view of some entoderm cells from below the embryonic area of
egg No. 189’ (12), showing reaction of cells to yolk remnants (Y).

12A and 12B Surface views, from within, of embryonic and trophoblastic
areas of egg No. 189’ (11); entodermal nuclei (mostly the larger) are seen above
the entoderm; embryonic area measures 0.96 mm.; aceto-osmic-bichromate.
* 500.

13 Defective blastocyst No. 88 (6) with large included blastomere; Hill’s
fluid; 5 yu. > 200.

14 Half-normal blastocyst No. 356 (2); Bouin; 54. X 200.

140

DEVELOPMENT OF THE OPOSSUM
CARL G. HARTMAN

141

142

“>

Resumido por el autor, Edward Phelps Allis, jr.

Los labios y orificios nasales en los peces gnatostomos.

En los vertebrados existen tres clases de labios funcionales,
primarios, secundarios y terciarios. Los labios primarios estan
colocados en posicién inmediatamente aboral al arco cuadrado-
mandibular; son funcionales en los Cicléstomos, en la porcién
media de la hendidura bucal de los Plagiostomos y probablemente
también en los Condrésteos; estan situados siempre en posicion
oral respecto a los orificios nasales. Los labios secundarios estan
formados por un pliegue del dermis externo, el cual primitiva-
mente cruza el angulo lateral de la boca, pero mas tarde se ex-
tiende hasta que encuentra el del lado opuesto con el que se une
en forma de sinfisis. La posici6n de estos labios es aboral con
relacién a los primarios y estan representados en los embriones
por los procesos maxilar, mandibular y fronto-nasal; la banda de
dermis externa situada entre ellos y los labios primarios se in-
cluye secundariamente en la cavidad bucal. Estos labios son fun-
cionales en las porciones laterales de la hendidura bucal de los
Plagiostomos y en toda la longitud de dicha hendidura en los
Teleostomos, Anfibios y Amniotos; estan colocados en posicién
inmediatamenta aboral al arco maxilo-dentario, y el labio supe-
rior en posicion oral con relacion a los dos orificios nasales (Teleos-
tomos, la mayor parte de los Plagiostomos) o entre dichos ori-
ficios (Heterodontus, Anfibios, Amniotos). Los labios terciarios
estiin situados en posicién aboral con relacién a los secundarios y
orificios nasales y se encuentran solamente en la mandibula de los
Dipnoos. Cuando la cresta del pliegue del labio superior secun-
dario encuentra el orificio oro-nasal se interrumpe y de este modo
se origina un sureo naso-bucal. En los Holocéfalos un conducto
nasal secundario, colocado entre ambos orificios nasales, se desar-
rolla exteriormente al puente nasal de los peces, entre éste y un
pliegue naso-labial que le recubre, y este conducto se transforma,
aparentemente, en el conducto nasal definitivo de los Anfibios.

Translation by Dr. José F. Nonidez
Columbia University

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MARCH 17

THE LIPS AND THE NASAL APERTURES IN THE
GNATHOSTOME FISHES ~

EDWARD PHELPS ALLIS, JR.

Palais de Carnolés, Menton, France
SIXTEEN FIGURES

His (92 b), in a work largely embryological and embracing all
classes of the Craniata, came to the conclusion that there were
four kinds of lips in these vertebrates:

1. Die Lippe der héheren Wirbelthiere und der Amphibien, welche
durch Verschmelzung des mittleren Stirnfortsatzes mit den Ober-
kieferfortsitzen entsteht und die vor den primiren Choanen liegt.

2. Die Lippe der Knochenfische, an deren Bildung der mittlere
Stirnfortsatz zwar Theil nimmt, aber deren Ort unterhalb der primaren
Choanen fallt.

3. Die Oberlippe der Selachier, welche ohne Betheiligung des mitt-
leren Stirnfortsatzes unterhalb der Riechgrube entsteht. Wenn wir
die erste Form als ‘Gesichtslippe’ bezeichnen, so kénnen wir die Formen
2 und 3 vielleicht ‘Gaumenlippen’ nennen.

Kine vierte Form ist die ‘Rauchenlippe’, welche wir weiter unten
bei Besprechung der Petromyzontenschnauze werden. kennen lernen;
sic hat ihren Ausgangspunkt hinter dem Eingang in die Rathke’sche
Tasche.

Keibel (93), in a work relating to this same subject, quotes
these four paragraphs from His’s work and then says that he
considers ‘‘die Hauptfrage durch His gelost,’’ but that he differs
from him in regard to certain points, one of which is that the
upper lips of the Teleostei and Selachii develop in a strictly
similar manner and are accordingly homologous instead of non-
homologous. In a later work Keibel (’06) reconsiders this sub-
ject, at somewhat greater length, and reasserts his earlier con-
clusions regarding it.

My work, limited almost exclusively to the adults of the
genathostome fishes, leads me to quite different conclusions. The
conditions found in these fishes will be first described, and then
brief comparison made with the conditions in higher vertebrates.

145

JOURNAL OF MORPHOLOGY, VOL. 32, NO. |

146 . EDWARD PHELPS ALLIS, JR.

PLAGIOSTOMI

In Chlamydoselachus the lips are much thicker at the angle of
the gape of the mouth than in their more anterior portions, the
angle of the gape thus being a relatively long line. The inner
end of this line forms the functional angle of the gape when the
mouth is widely opened, and the outer end of the line the func-
tional angle when the mouth is closed, and from this outer angle
the outer edge of each lip converges toward the inner edge until
the lips attain their normal thickness. This is readily seen in
the accompanying figures 1 and 2, as also in two similar figures
given by Garman in 1885, and it is there also seen that what is
actually a portion of the external surface of the head when the
mouth is widely opened, becomes enclosed between the lips when
the mouth is closed.

The cause of this thickening of the lips at the angle of the gape
was, in the first place, the inevitable formation of a fold in the
loose dermis at that angle, such a fold being well shown in
Miiller and Henle’s (1841) figure of Pristis antiquorum, repro-
duced in the accompanying figure 3, but this slight fold was
later enlarged, apparently because of the pressure of the thick
concave anterior edge of the musculus adductor mandibulae,
where it passed from the upper to the lower jaw, against the
internal surface of the fold. Because of this pressure, the fold
was forced outward and forward, and, when the mouth was
closed, bulged across the primary angle of the gape, its
anterior surface being presented symphysially and internally
and added to the lips at the angle of the gape. The crest of this
fold then formed, when the mouth was closed, a secondary angle
of the gape, which lay antero-lateral to the primary angle, and
short secondary upper and lower lips ran forward from it and
joined the primary lips. That portion of the external surface
of the head which lay between these secondary lips and the
primary ones was then first added to the lips at the angle of
the gape but later incorporated in the buccal cavity by the
formation of a cheek. The secondary lips, because of the
manner of their formation, are not at first represented by the

LIPS AND NASAL APERTURES IN FISHES 147

crest of the fold which gives origin to them, that crest diverging
more or less from the primary lips and the fold gradually spread-
ing out upon the external surface of the head and vanishing.
There is accordingly quite frequently a break in the definitive lip,
particularly the upper lip, between the primary lips and the
crest of the fold of the secondary lip; as seen in the accompanying
figures of Mustelus and Scyllium (figs. 5, 8).

In all of the few other Plagiostomi that I have been able to
examine, secondary lips are found which are strictly comparable
to those above described in Chlamydoselachus, but the short
secondary lips of the latter fish may be extended much fa. ther
forward, and there are marked variations in the upper lip due
mainly to the varying relations of the nasal apertures to the
upper edge of the mouth. These variations, in the few fishes I
have been able to examine, will be considered in connection with
the descriptions of the nasal apertures, but it may here be stated
that, as a general rule, where the nasal apertures lie at a con-
siderable distance from the upper edge of the mouth, the
secondary upper lip passes between those apertures and that
edge of the mouth, but when the oral nasal aperture lies near the
upper edge of the mouth, the fold of the secondary upper lip is
interrupted or displaced by its encounter with that aperture.

Related to the secondary angle of the gape there are, as is
well known, in most, but not all of the Plagiostomi, dermal
furrows, more or less developed. One of these furrows lies in
the upper jaw, dorsal and internal to the one or two upper labial
cartilages, and it usually turns downward, posterior and internal
to the articulating hind ends of those labials with the man-
dibular labial, and then forward (symphysially) a short distance
aboral and internal to the latter liabal. A dermal flap, or fold, en-
closing the articulating hind ends of the upper and lower labials,
is thus formed, and it may be called the labial fold. The related
furrow is indistinctly separated into two parts in all the Selachu I
have examined, these two parts being confluent in their superficial
portions but slightly separated from each other in their deeper
portions. They are, however, apparently simply parts of a single
furrow and can be called, together, the labial furrow; the two

148 EDWARD PHELPS ALLIS, JR.

parts, where necessary to designate them separately, being
called the supralabial and the postlabial furrows. <A third furrow
lies in the lower jaw immediately and hence superficial to
the mandibular labial. It extends posteriorly as far as the
ventro-anterior end of the postlabial furrow, but ends dorsal
(oral) to that furrow, not running directly into it, the two fur-
rows being confluent superficially, but distinctly separate in
their deeper portions. This third furrow may be called the
supramandibular furrow and the related fold the supraman-
dibular fold, the term mandibular being avoided because the fold
and furrow both lie external and hence anterior to the man-
dibular labial, and the term premandibular not being used
because it implies something belonging to a premandibular arch
or region. The labial and supramandibular folds together, form,
in the Plagiostomi, a single large fold which is usually but not
always separated into dorsal (maxillary) and ventral (man-
dibular) portions by a posterior continuation of the line of the
angle of the gape. Two other furrows are usually found, one in
each jaw, running forward (symphysially) from the line-of the
angle of the gape, not far from its inner end. They are both
short, and were apparently primarily simply creases in the
dermis between the folds of the primary and secondary lips. The
crease in the upper lip runs symphysially, diverging slightly from
the line of the primary upper lip. The crease in the lower lip,
in the few specimens I have examined, curves aborally and
approaches the outer end of the line of the angle of the gape, thus
circumscribing a small islet of dermis which lies immediately
symphysial to the line of the angle of the gape and external to
the primary lower lip; the islet accordingly belonging to the
tissues of the secondary lip. These little furrows will hereafter
be referred to as the maxillary and mandibular preangular
labial creases.

In Chlamydoselachus these several furrows are not well de-
veloped, the supralabial and supramandibular furrows being
simply creases in the dermis which do not run together posterior
and internal to the hind ends of the labials. There is accord-
ingly no postlabial furrow, and hence no labial fold, properly

LIPS AND NASAL APERTURES IN FISHES _ 149

so-called, in this fish. In Mustelus (probably vulgaris) and
Triakis fasciatum I find all the furrows well developed, and they
are shown in the accompanying figures of Mustelus (fig. 8). In
Seyllium canicula (fig. 5) I find the supramandibular furrow and
the two preangular creases well developed, but there is no post-
labial or supralabial furrow. In Raia clavata none of these
furrows are found as such, but the naso-buccal groove has .
probably absorbed the maxillary preangular crease, as will be
later explained.

In the adults of all of the Plagiostomi the primitive single ex-
ternal opening of the nasal pit is more or less completely, but
never completely, separated into two parts either by the well
known nasal flap, which projects from one side of the primitive
nasal opening and rests upon a flap seat on the other side, or by
the nasal flap and seat together with two deeper-lying flaps, one on
either side, which together form what I shall eall the nasal valve
and its valve seat. The nasal flap and nasal valve of one side
of the nasal pit, and the flap-seat and valve-seat of the other,
correspond to the two halves of the well known nasal bridge of
the Teleostei, but these two halves of the bridge never fuse
with each other in the Plagiostomi, the two nasal apertures
never, In consequence, being completely separated from each
other. ;

The two nasal apertures le, in the adults of all of the Plagi-
ostomi that I have examined or can find described, one lateral
or antero-lateral to the other, and it is always the lateral one
of the two which serves for the ingress of the current of water
passing through the nasal pit and the other for its egress. The
nasal groove of embryos of these fishes, as shown in figures,
always runs from in front orally and mesially, as does the line
of the external nasal apertures of the adult, but the line of the
groove is always inclined to the axis of the body at a smaller
angle than the line of the apertures. The line of the median
raphe of the Schneiderian membrane, lies, in the adult, ap-
proximately in the plane of the long axis of the fenestra nasalis,
_and crosses the line of the external apertures at a variable
angle, running, where the mouth is ventral and in the few speci-

150 EDWARD PHELPS ALLIS, JR.

mens I have examined, from behind mesially and more or less
~ aborally, apparently tending to become approximately parallel
to the upper edge of the mouth.

Neither the line of the external apertures of the adult nor the
line of the median raphe of the Schneiderian membrane thus lies
in the direction of the nasal groove of embryos, this apparently
having been caused by, or being related to, a change in direction
of the long axis of the external opening of the nasal pit. The
descriptions and figures of embryos do not permit the several
stages in this change in direction of this axis to be followed,
but that there is such a change, and that it has the character of
a partial rotation of the axis of the opening in the plane of that
opening is evident from a comparison of the conditions shown by
Berliner (02) in embryos of Acanthias with those found in the
adult of that fish; the central line of the opening of the nasal
pit of embryos, and the line of the median raphe of the
Schneiderian membrane, both lying approximately in the line of
the nasal groove and hence directed from in front orally and
mesially, while in the adult the long axis of the fenestra nasalis
and the line of the median raphe are directed aborally and
mesially, the line of the external apertures crossing this line at a
considerable angle and being directed mesially. The appearance
is accordingly that of the long axis of the fenestra nasalis having
rotated from left to right through a considerable angle, carrying
the median raphe with it and dragging, at either end, the related
external aperture a certain distance from its embryonic position.
The rotation of the nasal apertures is accordingly less extensive
than that of the axis of the fenestra, and the passage from each
aperture into the nasal capsule is, in consequence, pulled out and
lengthened to a variable extent, the two passages being directed
in opposite directions.

In the adult Mustelus (probably vulgaris) the long axis of the
fenestra nasalis extends from in front orally and laterally, lying
approximately parallel to the upper edge of the mouth and
coinciding in direction with the line of the median raphe of the
Schneiderian membrane (figs. 8 and 9). The line of the centers of _
the external nasal apertures crosses this line at a considerable an-

LIPS AND NASAL APERTURES IN FISHES il

gle, extending from in front mesially and slightly orally, the antero-
lateral aperture leading orally into the postero-lateral end of
the fenestra nasalis and the postero-mesial aperture leading
aborally into its antero-mesial end. The long axis of the fenestra
nasalis and the median raphe have accordingly here both*swung
partly round a circle, dragging the external apertures after them,
as in Acanthias.

The postero-mesial two-fifths of the edge of the fenestra nasalis
of this fish is of membrane, the remaining three-fifths of cartilage.
The ala nasalis (Nasenfliigelknorpel) encircles about four fifths
of the fenestra and fits against the inner edge of its cartilaginous
portion, the mesial end of its oral limb projecting mesially beyond
the cartilaginous portion of the fenestra and there lying largely
external to the fenestra and hence outside the nasal capsule.
That part of the ala nasalis which lies against the inner edge of
the cartilaginous portion of the fenestra is there strongly at-
tached to the inner surface of the nasal capsule by connective
tissues, but it is nowhere fused with the capsule. It lies against
the internal surface of the inner lining membrane of the capsule ’
and is strongly attached to it, and this membrane, in my pre-
served specimens, lies closely against the membrane forming the
membranous postero-mesial portion of the capsule. The two
membranes can, however, be easily separated from each other,
and it is to the inner lining membrane and not to the outer that
the ala nasalis is here attached. It is furthermore this inner lining
membrane of the capsule which alone connects the free mesial
ends of the oral and aboral limbs of the ala nasalis, no membrane
representing an unchondrified portion of the nasal capsule, such
as Gegenbaur (’72) describes in this and others of the Plagiostomi,
existing here. There is however a stout thick membrane,
which doubtless includes the perichondrial membrane, which lies
closely upon the external surfaces of the nasal capsule and the
ala nasalis, thus binding them together, but this membrane is
not an unchondrified portion of either of these cartilages and can
be easily stripped from them.

The lateral portion of the ala nasalis, together with the pro-
cesses a, a’ and 6 of Gegenbaur’s descriptions, encircles the in-

ha2 EDWARD PHELPS ALLIS, JR.

current nasal aperture, the process a not however entering the
nasal flap or being capable of being turned back as shown in
Gegenbaur’s figure 6, plate 17. The process a’ of this figure of
Gegenbaur’s projects internally and orally into the nasal capsule,
and, together with the adjoining lateral portion of the ala nasalis,
forms a broad plate which supports the internal and aboral surface
of the incurrent aperture, the aperture thus being a short funnel-
shaped passage which inclines from without orally and is consid-
erably contracted internally. The excurrent aperture is a similar
short and funnel-shapetl passage, but the funnel is here inverted,
the smaller end lying at the external end of the passage, and the
passage is directed from without aborally and hence in the oppo-
site direction to the incurient passage. The external and aboral
wall of the excurrent passage is supported by a broad and stout
plate which forms that part of the aboral limb of the ala nasalis
that lies mesial to the process a, the oral edge of the passage
being bounded by that short portion of the oral limb of the ala
nasalis which hes mesia! to the process 6. A short prong-like
process rises from the outer edge of the ala nasalis immediately
lateral to the process a, and, projecting ventrally (externally)
and aborally, abuts against the internal surface of the nasal
latero-sensory canal, apparently being developed in supporting
relation to that canal.

The nasal flap arises from the free orally directed edge of the
process a of Gegenbaur’s figure, and from the corresponding edge
of that part of the ala nasalis which lies mesial to that process,
and, projecting laterally and orally, rests upon the tissues cover-
ine the base of the process 8. The processes a’ and 6 project
uiwerd toward each other and are each clothed with mucous
membrane which is prolonged mesially beyond the process, the
mucous folds thus formed being apposed, valve-like, so as to
form, in part, a roof to the Schneiderian membrane and, in part,
to separate the two nasal passages from each other.. The base
of the process a is connected with the process a’ by a thin fold of
mucous tissue, this still farther separating the two nasal passages
from each other.

LIPS AND NASAL APERTURES IN FISHES 153

The current of water which enters the incurrent nasal aperture
is accordingly at first directed against the oral wall of the nasal
capsule, then turned aborally and mesially internal to the nasal
_ valve, in the direction of the long axis of the fenestra nasalis and
hence in the direction also of the median raphe of the Schneiderian
membrane, and then again turned orally in order to issue through
the excurrent aperture. The current thus has a zig-zag course
through the nasal capsule, entering it at the oral and lateral end
of the fenestra nasalis and leaving it at its mesial and aboral end.

The secondary upper lips of this fish lie, as in Chlamydo-
selachus, oral to the nasal apertures, and, as also in that fish, they
do not extend to the median line.

In Heptanchus the two nasal apertures lie near the lateral
edge of the ventral surface of the snout, the incurrent aperture
approximately lateral to the excurrent one. In Chlamydoselachus
the two apertures lie still farther laterally, the incurrent aperture
lying dorsal to the lateral edge of the snout and the excurrent
aperture ventral to that edge. The positions of these apertures
in these two fishes might accordingly be considered to represent
two stages in a migration of the apertures from the ventral to
the dorsal surface of the snout, such as is found in the ontoge-
netic development of the Teleostei (His, 92 b), but this is quite
certainly not the case, for this change in position of the apertures
in the Teleostel is apparently due wholly to an unrolling of the
cranial flexure, and not to a migration of the apertures, while
the change in position in the Selachii, such as it is, is due wholly
to their actual migration.

In both Heptanchus and Chlamydoselachus, the median raphe
of the Schneiderian membrane crosses the line of the external
nasal apertures almost at a right angle, running forward and
slightly mesially, approximately parallel to the upper edge of
the mouth. In both fishes also the processes a and a’ of the
ala nasalis are fused, the process a forming the external edge of
the combined: processes; and this combined process and the
process 6 project into the nasal capsule and form, together with
that part of the ala nasalis which bounds the lateral (dorsal in
Chlamydoselachus) half of the incurrent aperture, a broad cylin-

154 EDWARD PHELPS ALLIS, JR.

drical band, slit along the surface presented toward the excurrent
aperture. The two edges of this slit are apposed and form the
nasal valve and its valve-seat. Short horn-shaped processes
arise from the mesial (ventral in Chlamydoselachus) surface of
this cylinder, one on either side of the valvular slit, and, pro-
jecting mesially (or ventrally) and curving toward each other,
partly surround the excurrent aperture. The mesial (or ventral)
ends of these two latter processes are not connected with each
other in my specimens of either of these fishes, thus completing
the alar rg as shown in Gegenbaur’s figure of Heptanchus, and
the cartilage is nowhere fused with the outer edge of the nasal
capsule. The cartilage is strongly attached to the outer edge of
the capsule by ordinary connective tissues, and it is also attached
to the inner lining membrane of the capsule. In both Hep-
tanchus and Chlamydoselachus the lateral edge of the nasal flap
is attached to the external edge of the process aa’, the full length
of the process a, and, because of this attachment, there is no
passage connecting the two nasal apertures, between the flap
and the nasal valve. No part of the ala nasalis actually enters
the nasal flap, but the process a lies along the internal surface of
the lateral edge of the flap.

In Chlamydoselachus the secondary upper lips are, as- slvaade
stated, short, and they lie oral to the nasal apertures. In Hep-
tanchus these lips also lie oral to the nasa' apertures, but they
are much longer than in Chlamydoselachus, and, so far as I can
tell from my one much dissected specimen, they extend forward
to the symphysis and there fuse with each other, a continuous
band of the external surface of the head, concentric with the
upper edge of the primary cavity of the mouth, thus here being
added to that cavity.

In Sceyllium canicula (fig. 5) I find the ala nasalis practically as
described by Gegenbaur (’72), the process a lying in the lateral
edge of the nasal flap and the process 6 lying in the lateral edge
of a groove which forms the seat for the flap. The nasal valve
is formed by a small fold of mucous tissue which projects mesially
from the internal surface of the nasal flap and crosses the aboral
end of the process 6, and there is no cartilaginous process a’ de-

LIPS AND NASAL APERTURES IN FISHES 155

veloped in relation to it. The oral limb of the ala nasalis is not
prolonged mesially beyond the process 8, this apparently being
related to the presence of a naso-buccal groove, but the aboral
limb of the ala nasalis is thus prolonged beyond the process a
and there sends a second long process into the nasal flap, this
process and the process a both being thin and flexible. The line
joining the centers of the external nasal apertures is more nearly
parallel with the median raphe of the Schneiderian membrane
than in Mustelus, and the current of water passing through
the nasal pit does not have the markedly zig-zag course which
it has in that fish.

A secondary upper lip is found in normal position in this
fish, and extends from the secondary angle of the gape to the
lateral edge of the naso-buccal groove, where it ends abruptly
against the lateral wall of the groove. The crest of the fold of
this lip runs directly toward the process 8, which lies in the line
of the fold, and a well marked crease, extending from the line of
the angle of the gape about half way to the process £8, cuts
across the definitive lip and separates the crest of the fold of the
secondary upper lip from the primary lip. The nasal flap
extends to the upper edge of the mouth, completely covering
both the postero-mesial nasal aperture and the naso-buccal
groove, and its oral edge has the appearance of forming a part of
the secondary upper lip. It, however, forms no part of the fold
of that lip, as comparison with Miiller and Henle’s (’41) figures
of Scyllium edwardsii, Scyllium catulus, and Seyllium africanum
will show, for in these several fishes, notwithstanding that the
oral edge of the nasal flap lies not far from the upper edge of the
- mouth, the fold of the secondary upper lip runs forward, without
interruption, oral both to the nasal flap and the nasal apertures,
exactly as it does in Chlamydoselachus and Mustelus. There
is no naso-buccal groove in either of these three species of
Scyllium, the presence of this groove in Scyllium canicula thus
being related to a nasal flap which extends to the upper edge of
the mouth, or, more properly, to the presence of a nasal-flap
furrow which has that extent, that furrow lying beneath the
nasal flap (Allis, 716).

156 EDWARD PHELPS ALLIS, JR.

The naso-buceal groove of Scyllium canicula is short, the
postero-mesial edge of the nasal. capsule lying not far from the
upper edge of the mouth. The groove is bounded laterally by the
abruptly ending anterior end of the fold of the secondary upper
lip, and bounded mesially by the base of the nasal flap. The
postero-mesial nasal aperture has the full width of the naso-
buccal groove, and, as the passage leading from the nasal pit to
this aperture is always directed orally, the oral edge of the
aperture must have lain primarily near the upper edge of the
mouth. It is therefore this edge of this aperture that primarily
interrupted the fold of the secondary upper lip as it pushed for-
ward toward the symphysis, and the encounter of the fold with
the aperture raised the lateral edge of the aperture to such an
extent that the oral edge of the aperture became a groove, the
mesial edge of the groove being formed by the mesial edge
of the nasal-flap furrow. The fold of the secondary upper lip
could not cross the postero-mesial aperture and reappear mesial
to it, because of the barrier formed by the nasal flap. The naso-
buccal groove of this fish is thus not an independently developed
structure especially designed to connect the nasal pit and the
cavity of the mouth, as is generally assumed to be the case, but
is simply the oral edge of the postero-mesial nasal aperture and the
corresponding edge of the nasal-flap furrow transformed into a
eroove by the encounter of the fold of the secondary upper lip
with the lateral edge of the nasal aperture.

In Raia (species not given) Gegenbaur (’72) shows the ala
‘nasalis completely fused with the outer edge of the nasal capsule.
In two specimens of Raia clavata I find it wholly separate from
the capsule, but strongly attached to the inner lining membrane
of the capsule. The process a is long and flexible and lies in the
lateral edge of the nasal flap, as shown in the figures in my
work on the labial cartilages of this fish (Allis, 716). On the
internal surface of this part of the nasal flap is a large pad of
tissue, the thicker, aboral portion of which lies directly above
the nasal pit while the less tall, oral portion rests in a depression
on the opposite side of the nasal opening, immediately mesial
to the base of the process 6. The process 6 projects into the

LIPS AND NASAL APERTURES IN FISHES Bliss

nasal capsule and supports a fold of mucous tissue which repre-
sents one-half of the nasal valve. On the opposite side of the nasal
opening, and beneath the nasal flap, is another mucous fold, which
forms the other half of the nasal valve, but is not supported by
cartilage. These two halves of the nasal valve do not, in my
specimen, meet in the median line of the nasal pit, but they and
the thicker part of the pad on the internal surface of the lateral
edge of the nasal flap together form a partition across the pit.
In contact with the lateral edge of the ala nasalis, but not fused
with it, a narrow band of cartilage arises, and running inward,
is at first closely attached to the lining membrane of the capsule,
but soon separates from that membrane and lies in relation to
the median raphe of the Schneiderian membrane. ‘This raphe is
directed mesially and slightly orally, and coincides, in direction,
with the line of the external nasal apertures. The postero-
mesial edge of the nasal capsule is membranous, and the naso-
buccal groove passes over this membranous portion of the
capsule.

The naso-buccal groove is large, and extends orally and slightly
laterally from the postero-mesial nasal aperture to the anterior
edge of the mouth. In my work on the labial cartilages of this
fish (Allis, 16) I called this groove the nasal-flap furrow, the naso-
buccal groove being considered to be a secondary differentiation
of this furrow and to be represented in a deeper, lateral portion
of the entire furrow. My present work confirms the opinion
there expressed that the entire groove is primarily derived from
the nasal-flap furrow, and that the lateral and deeper portion of
the groove is a secondary differentiation, but comparison with
the conditions in Seyllium canicula, as now interpreted, shows
that it is the mesial portion only of the entire groove which is
derived from the nasal-flap furrow, the deeper, lateral portion
of the groove being formed by the crease which, in Seyllium,
separates the crest of the fold of the secondary upper lip from
the primary lip, together with the maxillary preangular labial
crease. . If these two creases of Scyllium were to coalesce and
then be extended forward until they fell into the naso-buccal
groove of that fish, the naso-buccal groove of Raia would be

158 EDWARD PHELPS ALLIS, JR.

formed. The conditions in the two fishes are so strictly similar
that it seems to me there can be no possible doubt of this, and
the formation of the groove in Raia is related, as it is in Scyllium,
to a nasal-flap furrow which extends to the upper edge of the
mouth. The fold of the secondary upper lip does not however,
in Raia, abut against the naso-buccal groove and end there, as it
does in Scyllium, for it has been deflected from its forward
course by the coalescence, with the naso-buccal groove, of the
crease between the secondary and primary upper lips. The
crest of the fold of this lip accordingly retains its normal relations
to this crease and runs aborally along the lateral edge of the
groove. In those of the Raiidae in which the nasal-flap furrow
does not extend to the upper edge of the mouth there is no
naso-buceal groove, and the secondary upper lip runs forward,
oral to both nasal apertures, exactly as it does in those Selachii
in which this groove is not found.

In Heterodontus (probably francisci), a single specimen of
which I have had at my disposal (figs. 6 and 7), the outer end of
the line of the angle of the gape lies at a relatively considerable
distance from the lateral edge of the palatoquadrate, and it has
been carried forward into the transverse plane of the hind end
of the nasal capsule, or even slightly anterior to that plane.
Because of this shortening of the length of the gape without a
corresponding shortening of the line of the angle of the gape, the
fold of the secondary upper lip is tall, and in running forward,
it immediately reaches the process 6 and is there directed
between the two nasal apertures. On the mesial side of the
nasal apertures the line of this secondary fold is continued by
a well developed fronto-nasal flap, or so-called process. A well
developed primary upper lip runs forward along the external
edge of the palatoquadrate dental arcade until it has passed
the postero-mesial nasal aperture, where the fold spreads out
on the internal surface of the fronto-nasal flap and vanishes as a
distinct fold. The edge of the fronto-nasal flap is certainly not
formed, in any part, by this lip, and it must accordingly either
represent an anterior continuation of the crest of the fold of the
secondary upper lip, or be a special and independent formation.

-

LIPS AND NASAL APERTURES IN FISHES 159

Its position, definitely in the line of the fold of the secondary
upper lip, is strongly in favor of its being an anterior continua-
tion of that fold, and such I consider it to be, since it is not
found in any fish I know of in which the fold of the secondary
upper lip has not been interrupted by meeting some part of the
nasal groove. The fold of the secondary lower lip is continued
forward approximately to the level of the anterior end of the
distinctly evident portion of the primary upper lip, and from
there onward the lower lip of the fish is a primary one.

The fenestra nasalis of this fish is long and narrow, and’ its
long axis is directed from in front postero-laterally, approximately
parallel to the secondary upper lip and hence diverging laterally
both from the primary upper lip and the lateral edge of the
palatoquadrate. The incurrent nasal aperture lies aboral to the
line of the secondary upper lip, the excurrent aperture oral to
that line; and the planes of the two apertures are inclined to
each other to such an extent that the internal angle between the
two planes is less than a right angle. The two apertures are in-
completely separated from each other, as in the other Selachii,
by an incompletely formed nasal bridge, which lies in the line
of the crest of the fold of the secondary upper lip. The excurrent
aperture is surrounded by a frill of dermal tissues, apparently a
modification of the nasal flap, the frill being continued, as a
fold, partly around the incurrent aperture. The nasal section
of the buccalis latero-sensory canal passes oral to this frill,
between it and the lateral edge of the palatoquadrate, thus
encircling the oral edge of the excurrent aperture, as it does in all
other Plagiostomi.

The excurrent nasal aperture is thus enclosed within the
buccal cavity when the mouth is closed, but it lies definitely
between the secondary and primary upper lips, aboral to the
latter, and hence in the same relation to it, to the palatoquadrate
dental arcade, to the buccalis latero-sensory canal, and to the
incurrent aperture which the corresponding aperture has in the
other Plagiostomi considered above. There is accordingly no
possible doubt that the excurrent aperture of Heterodontus is
the strict homologue of the corresponding aperture in other

1c0 EDWARD PHELPS ALLIS, JR.

Plagiostomi, that it hes on what is, in them, a part of the ex-
ternal surface of the snout, which has here been secondarily in-
cluded in the buccal cavity. There is no naso-buceal groove
connecting this aperture with the upper edge of the primary
cavity of the mouth, The fold of the secondary upper lip,
passing as it does between the two nasal apertures, might be con-
sidered to correspond to the lateral edge of the snout of Chlamy-
doselachus, which also passes between the two apertures, but
these two edges are not homologous, for they both exist con-
temporaneously and independently in Chlamydoselachus.

In Ginglymostoma concolor and Stegostoma fasciatum the
fold of the secondary upper lip has, as shown in Miller and ©
Henle’s (41) figures, approximately the course which it has in
Heterodontus, but it apparently crosses the nasal pit slightly
oral to the process 8, and the fronto-nasal flap is not so well
defined as in Heterodontus.

It is commonly said of Heterodontus, that the nasal and
buceal cavities are confluent, that the two nasal apertures are
connected by a naso-buccal groove, or that the excurrent aper-
ture has shifted orally until it has cut through the upper lip
and so come to lie on the internal surface of the lip; the upper
lip and the buccal cavity of this fish being considered to be the
strict homologues of the lip and cavity of other Plagiostomi.
These assumptions are, however, all incorrect.

Huxley (76) considered the excurrent aperture of this fish as
formed by the incomplete bridging of a naso-buccal groove, and he
compared it with the posterior nasal aperture of Ceratodus. This
implies two assumptions; first, that this aperture of Heterodon-
tus represents the oral end of a naso-buccal groove which has been
incompletely bridged by the arching over of its opposite edges;
and, second, that it.is the homologue of the posterior aperture
of Ceratodus, the latter aperture then being the oral end of a
canal formed by the completed bridging of a naso-buccal groove
similar to the one assumed to be found in Heterodontus. The
first of these two assumptions is incorrect, as explained above.
The second assumption is probably correct in so far as the
homology of the posterior nasal apertures of Heterodontus and

LIPS AND NASAL APERTURES IN FISHES 161

Ceratodus are concerned, but incorrect as to the formation of
this aperture, in the latter fish, by the bridging of a naso-buccal
groove, as Greil (713) has shown and as will be fully discussed
later. Furthermore, it may here be stated that the assumption,
frequently made, that the bridging of a naso-buccal groove, as
that groove is currently described in certain of the adult Plagi-
ostomi and in embryos of these and other vertebrates, could
produce two nasal apertures the homologues of those actually
found in the adult gnathostome fish is an error. The naso-
buccal groove, as described both in the adult and in embryos,
is said to extend either from the oral edge of the oral (posterior)
nasal aperture, or from that edge of the nasal pit, to the upper
edge of the mouth, the primitive oral (posterior) nasal aperture
accordingly lying aboral to the aboral end of the groove, and
between that end of the groove and the incompletely formed
nasal bridge. If then this nasal bridge were to be completely
formed, and the naso-buccal groove were to be bridged by the
fusion of its opposite edges, the fusion of this bridge with the
nasal bridge would give rise to a secondary posterior nasal
aperture which would not be the homologue of the aperture
actually found in fishes, while the formation of a naso-buccal
bridge alone, without the formation of a proper nasal bridge,
would give rise to an external nasal aperture which would corre-
spond to the undivided primitive single opening of the nasal
pit, and to an internal aperture which would have no homologue
in fishes.

The ala nasalis of Heterodontus (Cestracion) philippi has
been carefully described and figured by Gegenbaur (’72), and it
is said by him to be a complete ring, surrounding both nasal
apertures, and quite extensively fused, at two points, with the
cartilage of the nasal capsule. Huxley (’76) also described and
figured this cartilage in this fish, but as a partial and not a
complete ring, and he makes no mention of its being anywhere
fused with the edge of the nasal capsule; both of which details
are in accord with his conclusion that this cartilage is an upper
labial cartilage. Daniel (15) has also described and figured this
cartilage in Heterodontus francisci, and he also does not find it

JOURNAL OF MORPHOLOGY, VOL. 32, No. 1

162 EDWARD PHELPS ALLIS, JR.

either a complete ring or anywhere fused with the edge of the
nasal capsule.

In my specimen of Heterodontus francisci the fenestra nasalis
is, as already stated, a long and relatively narrow opening, and
the nasal capsule is relatively deep. The oral and mesial walls
of the capsule are largely membranous, much as shown in Gegen-
baur’s figure 2, plate 16, but unfortunately this membrane had
been dissected away along the edge of the ala nasalis before my
attention was called to the importance of preserving it, and I
can not tell whether it extended to that cartilage, as Gegenbaur
states, or not. The conditions in the other Plagiostomi ex-
amined would however indicate that it did not. The median
raphe of the Schneiderian membrane lies in the line of the long
axis of the fenestra nasalis, and the folds of that membrane are
so long that they extend outward almost to the inner edge of
the ala nasalis. The membrane is, as in the other Plagiostomi,
attached to the ala nasalis, and it is so closely applied to the
inner surface of the cartilaginous portion of the fenestra nasalis
that where it projects beyond the fenestra it appears as an
outward membranous extension of the walls of the nasal cap-
sule, and this may be what led Gegenbaur to conclude that the
membranous portions of the capsule are directly attached to the
ala nasalis.

The ala nasalis is, in my specimen, a complete ring, the
postero-mesial portion of which is thin and flexible, and it
is completely fused at one point with the outer edge of the
nasal capsule. The processes a and 6 are as described and
figured by Gegenbaur and Daniel, and there are projecting
mucous folds which form the nasal valve and its seat, but they
are without cartilaginous support. The incurrent passage is
directed orally, passes through that part of the alar ring which
lies lateral to the processes a and £, and leads to the postero-
lateral end of the fenestra nasalis. The excurrent passage is
directed from without aborally, passes through that part of the
alar ring which lies mesial to the processes a and 6, and leads to
the antero-mesial end of the fenestra nasalis. The little process
of cartilage, shown in my figure projecting laterally from the

LIPS AND NASAL APERTURES IN FISHES 163

mesial border of the alar ring, supports the external (ventral)
wall of this latter passage and corresponds to the flat recurved
end of the alar cartilage shown in Daniel’s figure of this fish.
The point where the alar cartilage is fused with the outer edge
of the nasal capsule lies between this little process and the
process 6.

The ala nasalis of all of the Plagiostomi was considered by
Gegenbaur to be a part of the chondrocranium, and not, as J.
Miiller (34) had previously concluded, an originally independent
skeletal element which had secondarily fused with the outer
edge of the nasal capsule. Huxley (76) and Parker (’76) must
both have accepted Miiller’s view, for they both (Huxley in
Heterodontus and Parker in Scyllitum and Raia) describe this
cartilage as a labial. Gaupp (06) however considers Gegen-
baur’s conclusion to be confirmed by conditions in an 8 cm.
embryo of Mustelus, the Nasenfliigelknorpel being said to there
be ‘‘in kontinuierlicher Verbindung mit der Nasenknorpel.”’ In
a 122 mm. embryo of Mustelus vulgaris I find the cartilage every-
where definitely and distinctly separate from the nasal capsule,
but the two cartilages are connected by a line of tissue which
is a continuation of two thin layers of tissue which are closely
applied, one to the external and the other to the internal surface
of both these cartilages, and which apparently represent thick
perichondrial membranes similar to the one that I have described
on the external surface of these cartilages in the adult Mustelus.
In a 55 mm. embryo of this same fish the ala nasalis is also
everywhere definitely separate from the capsule, but at one
point the two cartilages closely approach each other and are
connected by dense tissue continuous with that which lines both
surfaces of these cartilages. This tissue does not however here
undergo chondrification, for it persists as fibrous or connective tis-
sue in the olderembryo. The alar cartilage is thus quite certainly
not cut off from the edge of the nasal capsule in the ontogenetic
development of this fish, but it is developed in a layer of em-
bryonic tissue which is continuous with that in which the cap-
sule and adjacent portions of the chondrocranium are developed.
There seems however no more reason, simply because of this,

164 EDWARD PHELPS ALLIS, JR.

to consider the ala nasalis to be cut off from the outer edge of
the nasal capsule than there is to consider the vertebrae to be
segmented from the hind end of the cranium.

The posterior upper labial of Heterodontus francisci projects
dorsally and somewhat posteriorly from the angle of the gape,
lies against the postero-lateral surface of the nasal capsule, and
hence has the position shown in Huxley’s figure of Heterodontus
Philippi and not that in Gegenbaur’s figure of the same fish
(72, fig. 3, pl. 12). The anterior upper labial lies antero-mesial
to and parallel with the posterior one, as shown in Huxley’s
figure. Thus both these two labials project dorso-posteriorly
from the angle of the gape, instead of anteriorly, as they do in
the other Selachii above considered, this apparently being due to
the fact that the dorso-posterior ends of these labials were attached
posterior to the nasal capsule, and that accordingly, when their
ventro-anterior ends were carried forward by the shortening of
the gape, the labials swung forward around their dorsal points
of attachment as centers and so became directed dorso-pos-
teriorly. Because of this position of the labials, there is no
supralabial furrow. A deep postlabial furrow lies internal to
the mandibular labial and extends to its anterior end. A small
supramandibular furrow also occurs, external to the mandibular
labial and appearing as a groove on the external wall of the
large postlabial furrow. It is accordingly not seen unless the
latter furrow be forced open. There is a maxillary preangular
crease in the upper jaw and a corresponding crease in the lower
jaw.

HOLOCEPHALI

In the Holocephali the lips and the nasal apertures differ
markedly from those in the Plagiostomi. In Chimaera colliei
(figs. 18 to 16) the long axis of the fenestra nasalis lies in a plane
approximately parallel to the lateral edge of the palatoquadrate,
as it does in Heterodontus, but the line joiming the external
nasal apertures lies approximately in the same plane, instead
of, as in Heterodontus, crossing it at a considerable angle. The
two nasal apertures are accordingly, one postero-lateral, and the

LIPS AND NASAL APERTURES IN FISHES 165

other antero-mesial, instead of, as in all the Plagiostomi, one
antero-lateral and the other postero-mesial, and it is the antero-
mesial aperture of Chimaera which is currently considered to be,
and probably is, the incurrent aperture. The postero-lateral
aperture lies not far from the lateral edge of the palatoquadrate,
nearer to it than the antero-mesial aperture, and a short naso-
buccal groove leads from it to the lateral edge of the palatoquadrate
immediately anterior to the inner end of the line of the angle
of the gape. The line joining the two nasal apertures lies in a
nearly horizontal position, extending from the antero-mesial
aperture, laterally and posteriorly and inclining slightly toward
the lateral edge of the palatoquadrate. The Schneiderian mem-
brane lies at the bottom of a relatively deep nasal capsule, as an
elliptical rosette, the long axis of which lies in a nearly horizontal
position, transverse to the axis of the body, and at a marked
angle with the line of the nasal apertures, thus having practically
the position that it has in most of the Plagiostomi.

The middle portion of the upper lip is formed by a thick pad
of tough dermal and subdermal tissues which has the width of
the vomerine teeth, this portion forming the premaxillary upper
lip of Huxley’s (76) descriptions, and bounding antero-mesially
the naso-buccal groove. Along the oral edge of this lip there is
usually, but not always, a shallow but well defined suleus which
separates the lip into thin oral and thick aboral portions. The
oral portion is certainly a primary upper lip. The aboral por-
tion has the position of a secondary upper lip, but as it is quite
certain that it is formed, not by the fold of the secondary lip,
but by the oral edge of the nasal flap which has been turned
back upon the dorsal surface of the snout and has there coalesced
with the external dermis, it will be best to call it the aboral pre-
maxillary lip. The primary lips of opposite sides are directly
continuous with each other in the median line. The aboral
premaxillary lips are there separated from each other by a long -
and deep median incisure. At the lateral edge of the pre-
maxillary lip these two parts of the lip turn toward each other and
coalesce, thus forming a rounded fold, the oral end of which forms
a slightly projecting angle. Beyond this point the deeper portion

166 EDWARD PHELPS ALLIS, JR.

of the primary lip continues dorso-posteriorly along the lateral
edge of the palatoquadrate, there forming the floor of the naso-
buecal groove. This latter groove certainly includes, as in the
Plagiostomi, a nasal-flap furrow, the mesial edge of which forms a
crease beneath the rounded fold arising where the superficial
portion of the primary lip turns outward to coalesce with the
aboral premaxillary lip. Immediately posterior to this point the
lateral edge of the palatoquadrate is overlapped by,a little pro-
jecting point on the lateral edge of the mandibular dental plate,
shown in Dean’s figure 102, (06) of this fish, this point cutting
slightly into the primary lip.

The rounded fold formed by the coalescence of the primary
and aboral premaxillary upper lips runs at first aborally, and
hence in the direction of the mesial edge of the nasal-flap furrow
of the Plagiostomi, and then turns posteriorly (absymphysially),
approximately parallel to the lateral edge of the palatoquadrate,
and forms the oral edge of the postero-lateral nasal aperture.
At its dorso-posterior end this fold is continuous with the dorsal
end of a transverse ridge on the internal surface of the large
naso-labial fold, to be described later, which lies, when the naso-
labial fold is closed, in a nearly vertical position and hence
diverging posteriorly at an angle to the oral edge of the postero-
lateral nasal aperture. The aboral edge of the latter aperture
begins along the anterior (symphysial) edge of the transverse
ridge on the internal surface of the naso-labial fold, and, running
antero-ventrally, forms the oral edge of the large valvular proc-
ess, to be described below. When the naso-labial fold is closed,
the summit of the transverse ridge rests against that part of the
primary lip which forms the floor of the naso-buccal groove, and
in that position it forms a wall which closes the passage from
the postero-lateral nasal aperture into the buccal cavity while
still leaving a free passage through the aperture into the nasal
pit, and also from the aperture into a canal between the naso-
labial fold and the external surface of the valvular process.

Immediately posterior to the point where the transverse ridge
on the internal surface of the naso-labial fold reaches the oral
edge of that fold, a normal secondary upper lip begins, and from

LIPS AND NASAL APERTURES IN FISHES 167

there to the secondary angle of the gape is formed by the oral edge
of the labial portion of the naso-labial fold, which fold is formed
by the fusion of a normal labial fold with another, the probable
origin of which will be considered later. This large naso-labial
fold completely covers the postero-lateral nasal aperture, the
naso-buecal groove, and the nasal valvular process, and its
antero-mesial edge runs antero-dorso-mesially around the lateral
edge of the antero-mesial nasal aperture and then about half
way around its dorsal edge. The fold is separated into nasal
and labial portions by a slight transverse and vertical furrow on
its external surface, which extends orally from the aboral edge
of the fold about half way across it. The transverse ridge on
the internal surface of the fold lies at the posterior end of its
nasal portion, and hence immediately anterior to the vertical
furrow referred to above. Anterior to this ridge a smaller trans-
verse ridge abuts against the external surface of the nasal val-
vular process.

A primary lower lip extends the full length of the primary gape
of the mouth, and a secondary lower lip the full length of the
secondary gape, each of these lips extending forward to the
median. line and there being continuous with its fellow of the
opposite side.

Aboral to the large naso-labial fold, and aboral also to the
antero-mesial nasal aperture, is another dermal fold, continuous
in the median line with its fellow of the opposite side, the united
ventral edges of the folds of the two sides forming a somewhat
semicircular line which arches over and frames the nasal apertures.
and the mouth. This fold can be called, for reasons to be given
later, the supramaxillary fold, the furrow separating it from the
underlying tissues being called the supramaxilary furrow. The
fold, as here developed, is a characteristic feature of all the
Holocephali that I find figured, and in Chimaera it lodges the outer °.
ends of a series of ampullary tubules, which open on the external
surface in a line of ampullary pores immediately dorsal (aboral)
to the edge of the fold. The latero-sensory canals all lie aboral
(dorso-posterior) to the fold and hence aboral also to both the
nasal apertures, the antorbital section of the buccalis latero-

168 EDWARD PHELPS ALLIS, JR.

sensory canal not passing, as it does in all of the Plagiostomi and
Teleostomi, between the upper edge of the mouth and the nasal
apertures.

The supramaxillary furrow is always deepest in its posterior
portion, diminishing anteriorly to a shallow suleus, and it varies
in depth in different specimens, this apparently being wholly
due to a greater or less oral extension of the outer edge of the
fold, the fold in some specimens only overlapping the aboral edge
of the labial fold while in others it extends beyond the middle
line of that fold. The furrow extends dorsally into the tissues of
the head, the side walls of the furrow lying parallel to the ex-
ternal surface of the head, and the fold being, in consequence, a
thin sheet of integumental tissues.

A postlabial furrow occurs in normal position, internal to the
hind end of the labial fold, the end of the fold enclosing the
ventral end of'a labial cartilage which I consider, with Vetter
(78), to be the mandibular labial, instead of enclosing, as in the
Selachii, the articulating hind ends of that labial and one or both
of the upper labials. Starting from below and running upward,
the bottom of the postlabial furrow crosses the internal surface of
the mandibular labial, and, on reaching its dorso-posterior edge,
closely approaches the bottom of the supramaxillary furrow, but
it never, in my specimens, falls directly into that furrow. The
outer portions of the two furrows are however here confluent.
The postlabial furrow then turns anteriorly and, crossing the
external surface of the mandibular labial, reaches a point imme-
diately ventral (symphysial) to the line of articulation of the
mandibular and posterior upper labials, the latter labial being
‘the maxillary labial of Vetter’s descriptions. There the furrow
falls, at nearly a right angle, into the vertical furrow already
referred to as separating the large naso-labial fold into nasal and
labial portions. This vertical furrow lies in the direction pro-
longed of the line of articulation of the mandibular and posterior
upper labials, and directly external to the line between the sup-
plementary secondary upper and lower lips, to be described later.
It extends about half way across the naso-labial fold, and has no
homologue in the Selachii.

LIPS AND NASAL APERTURES IN FISHES 169

A supramandibular furrow occurs in normal position, external
to an anterior process of the mandibular labial, shown in Vetter’s
figures of this fish. The furrow is deep, its deeper portion ending
posteriorly dorsal to the ventro-anterior end of the postlabial fur-
row, while its larger, superficial portion is confluent with the corre-
sponding portion of the latter furrow. The supramandibular
fold differs from that in the Selachii in that its hind end has been
pushed downward on to the external surface of the mandible,
this being associated with a change in direction of the line of the
angle of the gape, which, in the Plagiostomi, is directed from
within antero-laterally. In Chimaera it is directed ventrally
and but slightly laterally, the inner end of the line having been
carried upward and forward, while its outer end has dropped
dewnward upon the external surface of the mandible. The
mandibular preangular labial crease starts from near the inner
end of this line and, after running at first symphysially along the
lateral edge of the primary lower lip, turns outward across the
outer edge of the secondary lower lip to a point dorsal (oral) to
the anterior end of the supramandibular furrow. There the
anterior end of the crease and furrow are connected by a slight
depression in the dermis. This crease thus cuts out of the
secondary lower lip an important preangular portion which corre-
sponds to the little islet cut out of this lip by the crease in
Mustelus. The labial fold overlaps the larger part of this islet,
the part so overlapped being thinner than the part beyond it
and being separated from it by a low ledge. This ledge corre-
sponds to the outer edge of the secondary lower lip of the Selachii,
and forms the edge of the functional lip of Chimaera when the
mouth is widely opened; but it is not the functional lower lip
when the mouth is closed, the functional lip then being formed
by the mesial edge of the preangular islet. In other words, the
posterior portion of the secondary lower lip has been turned
downward upon the outer surface of the mandible, and a broad
and V-shaped portion of its inner surface is presented externally.
The angle of the V is directed symphysially, and its mesial arm
forms the functional lip when the mouth is closed and the lateral
arm the functional lip when the mouth is opened. The former

170 EDWARD PHELPS ALLIS, JR.

lip may be ealled the supplementary, and the latter the actual
secondary lower lip.

When the mouth is closed, the supplementary lower lip, as
above defined, abuts against the posterior surface of the larger
of the two transverse ridges on the internal surface of the naso-
labial fold, this posterior surface, abutting as it does against
the supplementary secondary lower lip, thus being a supple-.
mentary secondary upper lip. The corresponding portion of
the actual secondary upper lip is formed, as already stated,
by the ventral edge of the labial portion of the. naso-labial fold,
and it extends forward from the secondary angle of the gape to
the vertical furrow which separates this portion of the fold from
the nasal portion. Anterior to this vertical furrow the line of
the secondary upper lip is continued onward along the rounded
ventral edge of the nasal portion of the naso-labial fold, but this
edge, although, like the oral edge of the nasal flap in Seyllium
and Raia, it forms part of the upper edge of the mouth, is no
morphological part of the fold of the secondary upper lip. The
vertical furrow cn the external surface of the naso-labial fold
lies directly external to the line between the short supplementary
secondary upper and lower lips, and has evidently been retained,
though, as will be later explained, probably not caused, by the
tissues of the thin labial fold there being creased by falling
slightly in between the two lips.

The supplementary secondary upper and lower lips form the
bounding side wall of the buccal cavity when the mouth is closed,
a supplementary gape of the mouth, which lies between the
primary and secondary gapes, thus being formed. The line of
this supplementary gape runs dorso-posteriorly at a considerable
angle to the line of the gape of the jaws, its inner end turning
dorsally and but slightly posteriorly, and the mandibular and
posterior upper labials articulate with. each other immediately
dorsal (morphologically posterior) to the inner end of the line.
These labials thus lie not far from the inner end of the line of the
angle of the gape, instead of near the outer end, as in the
Plagiostomi, and they lie, when the mouth is closed, external to
the palatoquadrate. The mandibular labial has, in consequence,

LIPS AND NASAL APERTURES IN FISHES 171

been pulled upward through the hind end of the labial fold, and
its ventro-anterior end, instead of its dorso-posterior end, lies in
the hind end of that fold.

The nasal apertures of Chimaera are separated from each other
by a stout broad valvular process, already several times referred
to, which projects ventro-antero-mesially from the aboral margin
of the nasal pit. The external surface of this valvular process is
concave, and its outer end is also concave or V-shaped, this end
fitting against a corresponding surface on the dorso-lateral
corner of the thick premaxillary lip of the fish, which forms the
valve-seat process. On the internal surface of the valvular proc-
ess there is a longitudinal ridge which fits into a corresponding
depression on the premaxillary lip, a second valvular surface
thus being formed which Jies ventro-lateral and internal to the
V-shaped valvular surface. Because of the markedly concave
external surface of the valvular process, a relatively large passage
is left between it and the overlapping naso-labial fold, which
communicates at one end with the postero-lateral nasal aperture,
and at the other end with both the antero-mesial aperture and
the exterior. This passage is so large that it would seem as if
it must give regular passage to water, either incurrent or excurrent
but exper!ments on the living fish can alone decide this. The
posterior edge of the valvular process is formed by a delicate
fold of mucous tissue which encloses a delicate piece of cartilage,
the cartlage ‘1’ of Hubrecht’s (’77) descriptions. There is, in all
my specimens, a small teat-like eminence on the mesial wall of the
antero-mesial nasal aperture, the possible significance of which
will be explained later.

The antero-mesial nasal aperture is encircled by the cartilage
‘kn’ of Hubrecht’s descriptions of Chimaera monstrosa, called by
him the ‘Nasenmuschel’ and certainly corresponding to some part
of the ala nasalis of the Plagiostomi. In Chimaera colliei, this
alar cartilage has the form of an oblique section of a cylinder, the
axis of the cylinder lying approximately in the line of the
trabeculae and hence at a marked angle to the plane of the
fenestra nasalis, this giving to the cartilage the appearance of
having been pulled forward and upward, almost completely out

ieee EDWARD PHELPS ALLIS, JR.

of the capsule. It is slightly concave on its dorso-lateral sur-
face, the concavity being bounded dorso-posteriorly by a slight
eminence on the cartilage and ventro-anteriorly by two sharply
pointed processes of the cartilage. The rounded eminence is
bound by ligamentous tissues to a corresponding eminence on
the internal surface of the cartilage ‘fg’ of Hubrecht’s descrip-
tions, and the two pointed processes support the tissues of the
V-shaped valvular surface of the valvular process. Ventral to
those pointed processes the cylinder is slit its full length, the cut
edges of the cartilage supporting the longitudinal valvular sur-
face and its valve-seat. The alar cartilage projects slightly
into the nasal capsule, the inner lining membrane of which is
firmly attached to it.

The conditions in Chimaera are thus, as already stated,
markedly different from those in the Plagiostomi, but it would
nevertheless seem as if they could have been derived from those
in certain of the latter fishes. In Miiller and Henle’s (41)
figure of Chiloscylliium punctatum, reproduced in the accom-
panying figure 4, a dermal fold is shown which encircles the
lateral edge of the antero-lateral nasal aperture; it may be
referred to as the nasal fold in order to distinguish it from the
nasal flap. The antero-mesial end of this fold lies anterior
(aboral) to the nasal flap and the related process a of the ala
nasalis. Posterior to this fold, and lying slightly deeper than it,
is the flap-seat and the related process 6 of the ala nasalis, and
posterior to the flap-seat there is a labial fold, the supralabial
furrow and the furrow of the nasal fold apparently being con-
tinuous at their adjoining ends. In this fish the long axis of
the fenestra nasalis has certainly rotated a certain distance in
the same direction that it rotates in other Plagiostomi, and if
this rotation were to be continued until the axis of the fenestra
had acquired the position that it has in Chimaera, and approxi-
mately has in Heterodontus, and if each of the two nasal apertures
were to follow that end of the axis of the fenestra to which it is
related until it came to lie directly external to it, as the two
apertures approximately do in Chimaera, the antero-lateral
aperture would pass internal to the nasal fold, and the nasal

LIPS AND NASAL APERTURES IN FISHES E7433

flap and its related process a would acquire the position of the
large valvular process of Chimaera. The position of the flap-
seat and the process 6 would not be affected by this change in
position of the nasal flap, but the antero-lateral nasal aperture
would be distorted and would lie somewhat perpendicular to the
external surface, a passage leading aborally and mesially from it
to the exterior beneath the overlapping nasal fold. An increased
development of the nasal fold would then produce the naso-
labial fold of Chimaera, and the flap-seat (process 8) of Chilo-
scyllium would become the transverse ridge on the internal
surface of the fold of Chimaera. The postero-mesial nasal
aperture would, in the meantime, have been crowded mesially
beneath the nasal flap and would push it backward, much as it
is shown artificially pushed back on one side of Miiller and
Henle’s figure of Chiloscyllium, and the process a, as shown in
that figure, would swing orally and then. mesially until it came
in contact with the base of the turned back nasal flap, and a new
flap-seat would be formed there.

If the nasal flap, turned back as above assumed, were to fuse
with that part of the external surface of the snout which lies
beneath it, the edge of the flap would in part form what I have
described as the aboral premaxillary lip, and the median incisure
of that lip would represent the line of incomplete fusion of the flaps
of opposite sides of the head. The remainder of the edge of the
flap would encircle a part of the original postero-mesial nasal
aperture, separating from it a new and smaller aperture, the
anterior-mesial aperture of the adult Chimaera. The postero-
lateral nasal aperture of the adult Chimaera would then be
formed by the naso-buceal groove of Chiloscyllium; that*is, by
the oral edge of the original postero-mesial aperture together
with the nasal-flap furrow. This postero-lateral aperture would
lie internal to the persisting antero-lateral aperture, these two
apertures having one edge in common, formed by the postero-
lateral edge of the large valvular process (process a), while the
other edges of the two apertures would be formed, in the one case
by the little fold which crosses the floor of the naso-buccal groove
of Chimaera and in the other by the transverse ridge on the

174 EDWARD PHELPS ALLIS, JR.

internal surface of the naso-labial fold of that fish. The fold of
the secondary upper lip of Chimaera would then have been inter-
rupted, as it is in the Plagiostomi, by its encounter with the
original postero-mesial nasal aperture and not by its encounter
with the antero-lateral aperture, as it would have been if that
aperture of the Plagiostomi were, as Hubrecht considered it to
be, the homologue of the postero-lateral aperture of Chimaera.
The folding back of the nasal flap would give rise, as already
stated, to the rounded fold at the lateral end of the thick pre-
maxillary upper lip of Chimaera, and the latero-oral corner of
the fold might have been perpetuated in the little teat-like
eminence on the mesial surface of the newly formed antero-
mesial nasal aperture.

The ala nasalis would naturally undergo modifications during
these changes in the nasal apertures, that part of it which en-
circled the original antero-lateral aperture undergoing reduction,
and the part which encircles the newly formed antero-mesial
aperture undergoing special development. This latter aperture
would be external and aboral to the position occupied by the
original postero-mesial aperture, and that part of the ala nasalis
which encircles it (cartilage ‘kn’) would have the appearance of -
having been pulled or stretched upward and outward from the
nasal capsule, as it actually has in Chimaera. <A remnant of the
process 6 would remain in the ridge on the internal surface of
the naso-labial fold, and such a remnant is actually there found.
The process a’ would become the longitudinal valvular surface,
found, in Chimaera, on the internal surface of the large valvular
process (process a), and it would acquire a new valve-seat near
the lateral edge of the newly formed premaxillary lip. The
cartilage ‘l’ of Hubrecht’s descriptions, above referred to, and ~
possibly also the little eminence on the dorsal surface of the
ala nasalis (cartilage ‘kn’), would represent persisting remnants of
that part of the ala nasalis which originally encircled the antero-
lateral nasal aperture.

During these changes in position of the nasal apertures, the
antero-lateral and incurrent-aperture would remain connected
with the exterior through the passage between the external sur-

LIPS AND NASAL APERTURES IN FISHES ges)

face of the process a and the overlapping naso-labial fold, and
would remain an incurrent aperture until such time as the
originally excurrent aperture had acquired its definitive position
and so become better situated to receive the inflowing current
of water; provided, of course, that this aperture of Chimaera is
actually incurrent and not still excurrent. The furrow related
to the nasal fold would probably become the supramaxillary
furrow, and if the bottom of this furrow were directed aborally,
_ it would give rise to a supramaxillary fold which would partly
overlap the labial fold, as it actually does in Chimaera. The
space in chiloscyllium, between the nasal and labial folds would
mark the place of, or actually represent, the vertical furrow
on the external surface of the nasal-labial fold of Chimaera.

How these changes could affect the buccalis latero-sensory
canal to such an extent as to deflect it from its normal course
and turn it aboral to the nasal apertures is not apparent, but
the cause, whatever it may have been, must have also been
operative in the Dipneusti, for this sensory line there also passes
aboral to both the nasal apertures. In the Amphibia this
sensory line always lies aboral to the posterior (internal) nasal
aperture and apparently usually aboral to the anterior (ex-
ternal) aperture also, but the descriptions that I find are not
definite as to this.

TELEOSTOMI

In the Teleostomi the nasal pit is said (His, ’92b) to lie, in
embryos, on the ventral surface of the snout, as it does in the
Plagiostomi and Holocephali. Peter (’06) says that the pit
develops late in these fishes, and that when the two nasal aper-
tures later shift from the ventral to the dorsal surface of the
snout they always retain their relative positions in relation to
the upper edge of the mouth, the anterior aperture of embryos
thus being the posterior aperture of the adult. There is ap-
parently a slight rotation of the line of these apertures in
the opposite direction to that in which they rotate in the
Plagiostomi, and there may be some rotation of the line of the
median raphe of the Schneiderian membrane, for Burne (’09)

176 EDWARD PHELPS ALLIS, JR.

says that this raphe is sometimes transverse to the line of the
apertures. The aboral, or posterior nasal aperture of the adult
Teleostei thus corresponds to the aboral, or antero-lateral aper-
ture of the adult Plagiostomi, and the current of water through
the apertures of the former fishes is the reverse of that in the
latter. The nasal pit is said by Burne to be completely bridged
in nearly all, but not all, of the Teleostei, and this bridge is
currently considered to be the homologue of the two half bridges
of the Plagiostomi fused with each other above the nasal groove.
No alar cartilage is however ever found, so far as I know, related
to this bridge in these fishes. Burne describes mucous folds
which project inward from the internal surface of the nasal
bridge and that would seem to correspond to the nasal valves of
the Plagiostomi.

In the Holostei and Crossopterygii the nasal apertures and
nasal bridge are apparently strictly similar to those in the
Teleostei.

Labial and supramandibular folds and furrows are well de-
veloped in many if not in all of the Teleostei, Holostei and
Crossopterygu, and I have, in an earlier work (Allis, ’00),
described them in certain of these fishes. The secondary upper
lip is represented, in all of these fishes, in the ventral edge ofthe
labial fold, and it is always continuous in the median line with
its fellow of the opposite side. It passes between the nasal
apertures and the upper edge of the primary cavity of the mouth,
and the space included between it and the primary lip forms a
secondary addition to the buccal cavity. The maxillary and
premaxillary bones lie in this secondary upper lip, as do the
labial cartilages of the Selachii in the secondary upper lips of
those fishes, and, where teeth are developed in relation to these
bones, they form a dental arcade which lies external to and
concentric with the primary, palatoquadrate arcade.

A supramaxillary fold is found, more or less developed in
many of these same fishes, the related furrow there extending
upward internal to the lacrimal bone, or to it and the anterior
suborbital bone. This furrow I have already described in
Scomber (Allis, 703, p. 64) as an important fold which “extends

LIPS AND NASAL APERTURES IN FISHES iW E

upward between the outer surfaces of the maxillary and pre-
maxillary and the inner surface of the lachrymal,”’ but I did not
then recognize that the projecting fold which encloses the ventral
edge of the lacrimal bone had any morphological significance. It
is, however, quite unquestionably the homologue of the supra-
maxillary fold of the Holocephali, the fold in the one being
related to a latero-sensory line and in the others to an ampullary
line, for the line of ampullary pores related to this fold in
Chimaera marks the primitive position of the related ampullary
sacs (Allis 01). In Amia this fold and furrow are much less
developed than in Scomber, but they are both still related to the
lacrimal bone, and the base of the shank of the maxillary bone
passes upward in the furrow, internal to the fold. In Gadus the
fold has been extended forward until it meets in the median line
and is there continuous with its fellow of the opposite side, the
fold lying internal to the ventral edges of the lacrimal and first
suborbital bones. The lacrimal bone here extends far forward
between the upper edge of the mouth and the nasal apertures,
and is traversed by the buccalis latero-sensory canal. This
canal does not extend, in this fish, anterior to the lacrimal bone,
the antorbital and dermal ethmoid bones of Amia, and the
sections of latero-sensory canal related to them, not being
found here. The supramaxillary fold is thus related, in both
Gadus and Amia, to the suborbital portion of the buccalis
latero-sensory canal and not to its antorbital section, and, as
this section of the canal in Amia turns upward in the lacrimal
bone posterior to the posterior nasal aperture, the supramaxillary
fold also turns upward there, while in Gadus, because of the
different position of the lacrimal bone, the fold runs forward,
oral to both nasal apertures. In Ophidium, Merlangus, Ammo-
dytes and Pleuronectes I find the supramaxillary fold in the same
position as in Gadus, and also continuous with its fellow of the op-
posite side, and as I do not find it so in any of the other Teleostei
at my disposal, this is apparently a characteristic of the Anacan-
thini. The position of the fold would here seem to be determined
by the position of the canal, rather than the position of the canal
by that of the fold. No teeth are ever found, so far as I know,

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 1

178 EDWARD PHELPS ALLIS, JR.

developed in relation to the supramaxillary fold, but marked
tooth-like spines may be developed in relation to it, as on the
lacrimal bone of Scorpaena (Allis, ’09).

In the Chondrostei the conditions are markedly different from
those in the other Teleostomi, and I am unable to give a definite
opinion regarding them. In my earlier work (Allis, ’00), I came
to the conclusion that the labial fold, there called the maxillary
fold, was represented in a fold of dermal tissues shown by
Parker (’82) extending across the snout of larvae of Acipenser,
immediately anterior to the barbels of the fish. This fold is not
shown either in my figures (Allis, ’04) of the adult Acipenser or
in those of Scaphirhynchus, but in both of these fishes the
anterior portion of the buccalis latero-sensory canal has ap-
proximately the position of the fold in larvae of Acipenser. It
is therefore probable that ‘this fold in the larvae of Acipenser is
the supramaxillary fold of the present descriptions, and hence
not a maxillary fold as I formerly concluded; and it passes, as
the fold does in Gadus, between the upper edge of the mouth
and the nasal apertures. The primary lips of both Acipenser
and Scaphirhynchus are certainly represented in some important
part of the lips of the suctorial mouth of these fishes. Whether
or not secondary lips are also represented in some part of the
lips I can not determine. If present there, they are quite cer-
tainly rudimentary and found only at the angle of the gape, and
it may be that they are wholly wanting. In Polyodon, also,
the secondary lips, if present, are found only at the angle of the
gape, this rudimentary condition of the secondary upper lip
doubtless accounting for the absence of a premaxillary bone in
all these fishes, and for the peculiar position and character of
the maxillary splint in Polyodon. The conditions in these fishes
however need further investigation.

DIPNEUSTI

In embryos of Ceratodus the nasal groove, as shown by
Semon (’93) and Greil (’13), is, when first formed, directed
postero-mesially, as it is in embryos of the Selachii, but it later

LIPS AND NASAL APERTURES IN FISHES 179

becomes directed postero-laterally. The anterior end of the
groove in the younger embryos shown by these authors apparently
corresponds to the same end of the groove in the older ones, the
aboral (anterior) nasal aperture of the adult fish (the mceurrent
aperture), accordingly corresponding to the antero-lateral and
also incurrent aperture of the Selachii. The line of the nasal
groove shown in the older embryos, which coincides in direction
with the line of the nasal apertures of the adult, accordingly
rotates in this fish in the opposite direction to that in which it
rotates in the Selachii and Holocephali. In the Teleostei the
line of the nasal apertures also rotates slightly in this same
direction, as already stated, and this is also the case in the
Amniota. There must then be some reason for this difference
in the direction of rotation of this line in these different verte-
brates, and it would seem to be related to the position of the
septum nasi, for this septum, when present in the Plagiostomi
and Holocephali, lies ventral to the trabeculae, while in Cerato-
dus and in the Teleostomi and Amniota it lies dorsal to the
trabeculae. .

Ginther (71) says that both the nasal apertures of the adult
Ceratodus lie within the cavity of the mouth. Huxley (’76)
later concluded that ‘‘the anterior nares can in no sense be said
to open into the cavity of the mouth, inasmuch as they lie outside
the premaxillary portion of the upper lip, and are not enclosed
by the maxillary portion of that lip. They are not even placed
between the upper and the lower lips, inasmuch as the vaulted
flap, on the under side of which they lie, is not the upper lip, but
the anterior part of the head.’’ Semon (’93) says that, in em-
bryos of this fish, the anterior aperture lies anterior to the upper
edge of the mouth and the posterior aperture posterior to that
edge, and that the two apertures arise through the coalescence
of the opposite lips of a naso-buccal groove. Gegenbaur (’98)
says that the posterior aperture, alone, is a choana, the anterior
aperture lying on the edge of the lip, and hence not in the cavity
of the mouth, and representing the primitive opening of the
nasal pit; but he was uncertain as to whether or not the posterior
aperture was derived, as he says it is in certain of the Selachii

180 EDWARD PHELPS ALLIS, JR.

and in the Holocephali, from the coalescence of the lips of oppo-
site sides of a naso-buceal groove. Greil (13) says that, in
embryos, both apertures lie in the roof of the mouth, that they
arise, as in the Teleostei, by the coalescence of the opposite
edges of a nasal groove formed by the elongation of the primitive
nasal pit, and that no naso-buccal groove is ever developed in
this fish.

Huxley (’76) described two cartilages in this fish which he
considered ‘to be the homologues of the anterior and posterior
upper labials of the Selachii, one of them lying between the
anterior and posterior nasal apertures and the other posterior to
the latter aperture. Bridge (98) later concluded that the
anterior of these two cartilages was not a labial cartilage, but a
persisting remnant of the ventral wall of the nasal capsule, and
he called it the subnasal cartilage. The posterior cartilage he
was inclined to consider, with Rose (’92), to be the homologue
of the so-called antorbital process of Lepidosiren and Protop-
terus, but he suggested that it might be the homologue of a
wholly separate and independent cartilage, found, in the latter —
fishes, which he considered to be unquestionably the homologue
of one of the upper labial cartilages of the Selachii. Firbringer
(04) accepted Bridge’s conclusion regarding the anterior of
these two cartilages of Ceratodus. Regarding the posterior
cartilage, he says that it is the homologue of the upper labial
cartilage of Bridge’s descriptions of Lepidosiren and Protop-
- terus, but that this cartilage is, in all these fishes, a detached
portion of the chondrocranium and not the homologue of either
of the upper labials of the Selachii. Because of this derivation,
of the cartilage he calls it the postnasal cartilage, He describes
and figures a cross-bar of cartilage lying lateral to the posterior
nasal aperture and connecting the subnasal and postnasal carti-
lages, and he says that it is a secondary, protective arrangement,
developed in relation to these cartilages and the nasal apertures.
Huxley neither describes nor shows this cross-bar of cartilage
connecting his two upper labials.

In a large but not well preserved head of this fish I find the
cross-bar of cartilage described by Fiirbringer, but in my speci-

LIPS AND NASAL APERTURES IN FISHES 181

men it is simply a somewhat posteriorly directed process, or
prolongation, of the so-called subnasal cartilage, the outer end
of this process reaching and being strongly attached by liga-
mentous tissues to the so-called postnasal cartilage of Fir-
bringer’s descriptions, but not being continuous with it. The
lateral prolongation of the subnasal cartilage beyond this cross-
bar, shown in Fiirbringer’s figure, is, in my specimen, simply a
band of tough ligamentous tissue which runs forward along the
lateral edge of the anterior nasal aperture and is lost in the
tough tissues of the upper lip. The latero-sensory canals of this
fish all lie aboral (dorso-posterior) to both the nasal apertures,
and aboral also to the upper lip, this being the relations that
they have, in the Holocephali, to the nasal apertures and the
supramaxillary fold of those fishes.

Comparing the conditions in this fish, as thus described, with
those in the Plagiostomi, it seems certain that the anterior and
posterior nasal apertures of these fishes are respectively homolo-
gous. The so-called subnasal cartilage of Ceratodus must
then be a remnant of the alar cartilage of the Plagiostomi, for
that this cartilage of Ceratodus, spanning as it does the primitive
single nasal opening, can be a persisting remnant of any part of
the walls of the nasal capsule is evidently impossible, those walls
encircling the primitive nasal opening and not spanning it.
This cartilage is said to be found, in Lepidosiren, fused at its
lateral end with the edge of the nasal capsule and there appear-
ing as a process of that capsule, and this would be in accord with
Gegenbaur’s derivation of the alar cartilage from the outer edge
of the nasal capsule, but in Ceratodus it is an independent
cartilage, and this would seem to be its primitive condition.
The posterior process of this cartilage, as I find it, the cross-bar
of Furbrimger’s descriptions, must then also be a part of this .
alar cartilage, and the ligament which lies along the lateral edge
of the anterior nasal aperture, a further but unchondrified por-
tion of it. The postnasal cartilage is, for reasons given im-
mediately below, quite certainly not a part of this alar cartilage,
and would seem to be the homologue of the cartilage ‘f’ of
Hubrecht’s (77) descriptions of Chimaera and Callorhynchus,

182 EDWARD PHELPS ALLIS, JR.

this cartilage being said to be fused with a cartilage ‘g’ in
Chimaera, but a separate and independent cartilage in Callor-
hynchus. Hubrecht considered both these cartilages of the
Holocephali to be alar cartilages (‘Nasenfliigelknorpel’), and
Schauinsland (’03) calls them nasal cartilages, but I consider
the cartilage ‘f’ to be an upper labial cartilage.

It is evident that the upper lip of Ceratodus, as defined Ne
Ginther and confirmed by Greil, corresponds to the supra-
maxillary fold of Chimaera. The upper lip as defined by Huxley
was said by him to consist of two parts, one represented in a
transverse integumental fold lying immediately anterior to the
so-called vomerine teeth, and -the other by a fold extending
forward from the angle of the gape of the mouth along the
lateral edges of the nasal apertures. The former fold, as I find
it, I consider to be a part of the primary upper lip of the fish,
the other, which is hardly recognizable in my specimen and found
only at the angle of the gape, being a secondary lip. The
functional upper lip is then, as Huxley concluded, simply a fold
of the dermis on the anterior part of the head, and is accordingly
a tertiary upper lip lying anterior to the secondary one and
circumscribing a second band of the external surface of the
head which is here added to and included in the cavity of the
mouth.

These lips and the labial folds and furrows of Ceratodus are
shown in the accompanying figure 12, and it is there seen that
there are two angles to the gape of the mouth, one the actual angle
and the other the apparent angle when the mouth is closed.
The actual angle corresponds to the angle of the secondary lips
of the Selachu, and immediately posterior to it there is a short _
flat flap of dermal tissues which is the hind end of the labial fold.
This latter fold lies parallel to the dorsal surface of the mouth
cavity, and its lateral edge lies slightly internal to the edge of
the functional, or tertiary upper lip. Dorsal and posterior to
this fold is the external opening of the so-called labial cavity
of Ginther’s descriptions, that cavity lying dorsal to the cavity
of the mouth and being the supramaxillary furrow of the present
descriptions. In an earlier work (Allis, ’00) this furrow was

LIPS AND NASAL APERTURES IN FISHES 183

also called by me the supramaxillary furrow and was said to be
the homologue of the similarly named furrow in Amia, but I
now recognize that there are, in Amia, two partly confluent
furrows here, one being the labial furrow and the other the
supramaxillary furrow of the present descriptions. From the
external opening of this furrow the postlabial portion of the
labial furrow runs forward in the lower jaw to the hind end of a
supramandibular fold, and is there continuous with the supra-
mandibular furrow. When the mouth is closed the functional,
or tertiary upper lip extends posteriorly across the actual angle
of the gape and ends at the hind end of the supramaxillary
furrow, where that furrow and the postlabial furrow are fused
in their superficial portions, this point forming a tertiary angle
of the gape and lying posterior to the actual, or secondary
angle. When the mouth is open, the line of the tertiary upper
lip is seen to be interrupted for a short distance immediately
posterior to the secondary angle of the gape. The lateral end
of the so-called postnasal cartilage of Fiirbringer’s descriptions
passes between the labial cavity (supramaxillary furrow) and
the cavity of the mouth, and ends in the mesial edge of the
short labial fold. This cartilage is thus quite unquestionably
an upper labial which has been pulled postero-mesially out of ~
the labial fold, but there is nothing in this fish definitely to show
whether it is an anterior or a posterior upper labial. It, how-
ever, evidently corresponds to the cartilage ‘f? of Chimaera,
above referred to, and as that cartilage is certainly not a pos-
terior labial, both it and the cartilage of ceratodus must be
anterior upper labials.

SUMMARY AND COMPARISONS

It is thus seen that there are three distinctly different types
of lips in the gnathostome fishes, a primary lip, a secondary lip,
and a tertiary lip, and a band of the external surface of the head
is added to the functional cavity of the mouth between the
primary and secondary lips and a second such band between
the secondary and tertiary lips. The primary lips, as functional

184 EDWARD PHELPS ALLIS, JR.

lips, are found only in the antero-mesial portions of the lips of
the Plagiostomi, and probably also in those of the Chondrostei.
The secondary lips are probably found in the postero-lateral
portions of the lips of the Chondrostei; are found either in the
corresponding portions only of the lips of the Plagiostomi, or
extending the full length of the gape of those fishes; extending
the full length of the gape in all of the Teleostei, Holostei and
Crossopterygii that I have been able to examine; and extending
the full length of the gape in the lower jaw of the Holocephali,
but not the full length of the gape in the upper jaw.

The secondary upper lip passes between the upper edge of the
mouth and the nasal apertures in all of the Teleostei, Holostei
and Crossopterygii that I have been able to examine, as it also
does in the Chondrostei, if there present. In all of the Plagios-
tomi that I have been able to examine in which there is no
naso-buccal groove, excepting only Heterodontus, the secondary
upper lip also passes between the upper edge of the mouth and
the nasal apertures; but in those of these fishes in which there is
a naso-buccal groove, the lip either abuts against the groove
and ends there (Scyllium), or runs along the lateral edge of the
groove (Raia); the naso-buccal groove resulting from the en-
‘ counter of the fold of the secondary upper lip with the postero-
mesial (oral) nasal aperture. In Heterodontus the secondary
upper lip passes between the two nasal apertures and is con-
tinued mesial to those apertures as the so-called fronto-nasal
flap, or process. In the Holocephali this lip is represented in
the oral edge of the labial portion of the naso-labial fold, and it
was primarily interrupted by its encounter with the postero-
mesial nasal aperture, as it is in those of the Selachii in which a
naso-bucecal groove is found. :

The tertiary lip is found only in the Dipneusti, and even there
only in the upper jaw, the lip passing aboral to both the nasal.
apertures. In adults of the Holocephali and Teleostomi (ex-
cepting the Chondrostei), and probably also in embryos of
Acipenser, this tertiary upper lip of the Dipneusti is represented
in a fold of the dermis on the external surface of the head which
is the supramaxillary fold of the present descriptions.

LIPS AND NASAL APERTURES IN FISHES 185

The primary upper lip lies immediately external to the teeth
developed in relation to the palatoquadrate. The secondary
upper lip lies immediately external to the maxillary and pre-
maxillary teeth. The tertiary upper lip has no teeth developed in
relation to it, but in the Teleostei and Holostei the corresponding
supramaxillary fold encloses the oral edges of the lacrimal and
anterior suborbital bones.

In the Amniota the functional lips are the secondary ones,
and the secondary upper lip passes, in all these vertebrates,
between the two nasal apertures. Maxillary and premaxillary
teeth or bones may be developed, as in the Teleostei, Holostei,
and Crossopterygil, in relation to this secondary upper lip. In
most of the Sauropsida both of these latter bones are actually
developed in relation to this lip, and there are, accordingly,.in
the upper jaw of these vertebrates, as in the Teleostei, Holostei, .
and Crossopterygii, two arcades, with or without teeth, an
inner and primary arcade formed by the bones developed in
relation to the palatoquadrate and an outer and secondary
arcade formed by the maxillary and premaxillary bones; and
the posterior nasal apertures lie between these two arcades. In
certain of the Sauropsida a secondary palate is formed by ventral
plates of the vomer and palatine, and the definitive choana lies
posterior to the plate so formed, but the primary choana never-

theless still les anterior to the dorsal and primary portions of
those bones. In the Mammalia this same relation of the
posterior nasal apertures to the two arcades must also persist,
but I am not familiar enough with these vertebrates to discuss
the conditions there. Comparison with fishes would however
suggest that the presence of a cheek in the Mammalia ditremata
is due to a marked reduction of the maxillary bone, as in
Polypterus (Allis, ’00), and its fusion with the pterygoid, this
then accounting for the absence in these animals of the latter
bone, as claimed by Gaupp (710). And it may be further
mentioned, as a curious coincidence, that a dimple is found in
the cheek of man in approximately the position of the post-
labial furrow of fishes. In the Mammalia monotremata, where
the pterygoid persists in normal reptilian position (Gaupp,

186 EDWARD PHELPS ALLIS, JR.

10), there are said to be no ‘lips’ (Géeppert, ’06, p. 79), and
hence, of course, no cheek.

In embryos of all of the gnathostome fishes above considered,
the primary lips of either side of the head are at first represented
in the corresponding half of the edge of the primary stomodaeum.
When the so-called maxillary and mandibular processes of either
side later begin to develop, they overlie and include the absym-
physial portions of this edge of the stomodaeum, leaving the
symphysial portions of that edge exposed between their anterior
(symphysial) ends and the corresponding ends of the processes
of the opposite side, as shown in figures of embryos of all of these
fishes. The mandibular processes of opposite sides always, in
these figures, ultimately meet and coalesce at the symphysis,
but the conditions in the adult show that this can not take place
in all of the Plagiostomi. The maxillary processes of opposite
sides, on the contrary, do not always meet and coalesce at the
symphysis in these embryos, as is well shown in Géeppert’s
(06) figures of embryos of Torpedo and Mustelus and Keibel’s
(06) figures of embryos of Acanthias, a portion of the edge of
the primary stomodaeum, which represents a corresponding part
of the primary upper lip, always remaining exposed between
them and forming the median portion of the definitive upper lip.
Keibel (06, p. 157) calls attention to the fact that, in embryos
of Acanthias, the middle portion of the upper lip is not formed -
by the maxillary processes, and, although he could not determine
from what it was formed, he questions its origin from the
fronto-nasal process. It is, in fact, not formed by that process,
properly so-called, for the oral edge of the process is formed by
the crest of the fold of the secondary upper lip, and the fronto-
nasal process, properly so-called, only occurs where the secondary
upper lip crosses some part of the nasal groove and has been cut
in two by its encounter with it. This takes place in Hetero-
dontus, and probably also in certain others of the Plagiostomi
in which there is a naso-buccal groove, but it does not take place
in any of the Plagiostomi in which there is no naso-bucéal groove,
nor in any of the Teleostomi, the secondary upper lip there
always passing between the upper edge of the mouth and the

LIPS AND NASAL APERTURES IN FISHES 187

nasal apertures. When, in the adults of any of these fishes, the
secondary upper lips meet and coalesce at the symphysis, the
maxillary processes also meet there and coalesce, and that band
of the external surface of the head which, in earlier embryos,
lies between their anterior ends is enclosed in the buccal cavity
as a part of the secondary addition to that cavity. If the
anterior ends of the maxillary process were to approach each
other closely, but not fully to meet and coalesce, a median
incisure would be left in the definitive upper lip.

In embryos of all of the Amniota the fold of the secondary
upper lip has been cut into maxillary and fronto-nasal portions
by its passage across the nasal groove, as it is in certain fishes,
and as each of these two portions of the lip, or so-called proc-
esses, lies at first oral to the corresponding nasal process, as
those processes are defined by Peter (’06), the fold of the
secondary upper lip must here have passed across the oral nasal
aperture and not between the two apertures. In Mammalian
embryos the fronto-nasal and mesial nasal processes later fuse
completely with each other to form the processus globularis of
His’s (92 b) descriptions. This globular process then arches
over the nasal groove and fuses with the maxillary and lateral
nasal processes, either singly and separately or after those two
processes have fused with each other, and a nasal bridge is
formed which is certainly the strict homologue of the bridge in
the Teleostomi, for the fold of the secondary upper lip can, at
the most, simply have caused a widening of the bridge. The
fact that, in the Mammalia, the posterior nasal aperture is
temporarily closed by the contact of the cut ends of the fold of
the secondary upper lip, and that a membrana bucco-nasalis is
formed there and later broken through, is certainly simply a
modification of the normal process of development, as found in
the Teleostomi, and is due to the fold of the secondary upper lip
not passing across the center line of the definitive nasal bridge.
Keibel also considers this manner of formation of this aperture
of no morphological significance, for he says (’93, p. 477):

so erscheint es mir von Wichtigkeit, dass festgestellt wurde, dass der
laterale Stirnfortsatz an der Bildung des primitiven Gaumens beteiligt

188 EDWARD PHELPS ALLIS, JR.

ist, ja beim Séuger diese Bildung einleitet. Mindere theoretische
Bedeutung kann ich der Thatsache beimessen, dass die primitive
Choane erst secondér durchbricht, und dass die Nasenhdéhle in ersten
Stadium ihrer Entwickelung nicht durch eine Spalte, sondern durch
eine solide Epithelleiste mit der Mundhdéhle in Verbindung steht. Es
handelt sich hier nicht um principielle Verschiedenheiten. Ob die
verbindung zwischen zwei Hohlriumen durch eine Epithelleiste oder
durch eine Spalte hergestellt wird, kommt in vielen Fallen im Grunde
auf dasselbe hinaus.

The primary lips of embryos of the Amniota are represented,
as they are in fishes, in the deeper portions of the mandibular,
maxillary and fronto-nasal processes.

In embryos of Amphibia the conditions, as described by
authors, are totally different from those above considered. In
the Gymnophiona the conditions are simpler than in the Urodela
and Anura. In the former (Hypogeophis) a large lateral process
(laterale Stirnfortsatz) is said by Hinsberg (02) to project across
the oral edge of the nasal pit and to fuse completely, in its
deeper portion, with the fronto-nasal process. The superficial
portions of the two processes do not however fuse, this leaving,
between the processes a shallow groove which extends from the
nasal pit to a depression in the roof of the mouth (Gaumendach)
which lies between the bases of the two processes; and it is im-
portant to note that this shallow groove connects with the outer
edge of the nasal pit, and not with its deeper portion. The fold
of the secondary upper lip passes along the middle line of the lat-
eral process and continues beyond it across the fronto-nasal process,
the fold thus lying between the nasal opening and the depression
in the roof of the mouth. In slightly older embryos the lateral
and fronto-nasal processes project above the shallow groove
described above and there again fuse with each other, thus
enclosing the groove beneath the epidermis, the nasal end of
the groove becoming a closed canal surrounded by epithelial
tissues, and its oral end becoming a solid cord of epithelial tissue.
The processes do not so project and fuse with each other above
the depression in the roof of the mouth, that depression still
persisting between the bases of the processes. In still older
embryos the lumen in the epithelial cord is prolonged orally,

LIPS AND NASAL APERTURES IN FISHES 189

and finally reaches and opens into the depression in the roof of
the mouth, there forming the primitive choana.

The lateral process of these embryos of Hypogeophis thus
fuses in two places with the fronto-nasal process, one internal
to the canal leading from the nasal pit to the primitive choana
and the other external to it, the canal thus lying between the
two points of fusion and leading primarily from the definitive
external nasal opening to the choana without traversing the
nasal pit. This canalthus strongly recalls the passage which, in
the adult Chimaera, lies between the valvular nasal process and
the nasal portion of the naso-labial fold and leads from the
antero-mesial nasal aperture to the postero-lateral one, and I
consider it to be its homologue.

If this be so, the conditions in Hypogeophis would be derived
directly from those in Chimaera by the fusion, first, of the
valvular process of Chimaera with its valve-seat, this forming a
nasal bridge and being represented in Hypogeophis by the first
of the two fusions of the lateral and fronto-nasal processes, and
second, by the fusion of the nasal portion of the naso-labial fold
of Chimaera with the premaxillary lip, this being represented in
Hypogeophis by the second of the two fusions of the two proc-
esses. By an extension of the first of these two fusions, the
postero-lateral nasal aperture of Chimaera is occluded, this
leaving the naso-buccal groove external to the nasal bridge and
an oral remnant of it persisting as the depression in the roof of
the mouth of Hypogeophis. The large lateral process of embryos
of Hypogeophis thus represents, in its deeper portion, the lateral
nasal process of the Plagiostomi, and in its superficial portion
the nasal portion, of the nasolabial ‘fold of Chimaera. The
primitive choana of Hypogeophis would then correspond to the
oral end of the naso-buccal groove of Chimaera, and would
accordingly be the homologue of the choana of the Amniota,
but its connection with the nasal pit would be by a canal which
passes external to the nasal bridge instead of internal to it.

In the Urodela and Anura the conditions differ from those in
the Gymnophiona simply in that the shallow groove which
becomes the choanal canal of the latter has been obliterated

199 EDWARD PHELPS ALLIS, JR.

instead of being invaginated, this necessitating a secondary
reopening of the occluded postero-lateral nasal aperture of
Chimaera. The fact that the opening so formed is said (Hins-
berg, ’01) to lie primarily in the dorsal wall of the anterior end
of the alimentary canal, and hence posterior to the oral plate
(bucco-pharyngeal membrane), can not affect this homology, for
the choana of the adult still lies anterior to the bones developed
in relation to the palatoquadrate, and hence quite certainly
anterior and not posterior to the primary upper lip, between that
lip and the secondary one.

If this be the manner in which the nasal apertures of the
Amphibia have been developed, and it seems to me that it must
be, the embryological processes simply being obscured by con-
densations and abbreviations, then the Amphibia must either
be descended from some selachian similar to the one from which
Chimaera is descended, or directly from some early Chimaeroid.
The apparent similarity, in Chimaera and the Amphibia, in the
relations of the buccalis latero-sensory line to the nasal apertures,
is in favor of the latter assumption, and, furthermore, it seems
improbable that these complicated and peculiar nasal apertures
would have been twice developed in the vertebrate series. This
origin of the Amphibia would also probably explain the palato-
quadrate, the upper and lower labials, and the horny jaws of
larvae of the Anura.

From the preceding descriptions of embryos and adults, it
is evident that the primary lips of all of the gnathostome verte-
brates must lie primarily at or but slightly. anterior to the oral
plate of embryos, for as the mandibular arches lie morphologically
posterior to the plate, the cartilaginous bars of those arches must
also have primarily had that position. It would then be natural
to conclude that the primary upper lips, which lie immediately
anterior to the teeth developed in relation to the cartilaginous
mandibular bars, are developed from tissues which lie oral to
the hypophysial invaginations, but that this is so in all verte-
brates can not be definitely determined from the descriptions
given of embryos.

LIPS AND NASAL APERTURES IN FISHES 191

Dohrn (’04, fig. 9, pl. 16), in a median sagittal section of an
embryo of Torpedo, shows what would seem to be the primary
upper lip lying posterior to the hypophysis, and it is shown as
formed of ectoderm on its external, and of entoderm on its
internal surface, as if it were a remnant of the oral plate. His
however shows (92 b, fig. 26), in a median sagittal section of an
embryo of Pristiurus, a dermal fold immediately anterior (aboral)
to the hypophysis which he calls the upper lip of the fish, and in
a similar figure (His, 792 a, fig. 14) of an embryo of Scyllium,
what is apparently this same fold, but not named, is shown in a
similar position. There is, however, in both these figures of
His’s, a larger but lower fold of the ectoderm, oral to the hypoph-
ysis, in the place occupied by the fold in Dohrn’s figure of Tor-
pedo. It may then be that the fold called by His the upper lip in
his figure of Pristiurus is a secondary and not a primary lip.
Comparison with Lundborg’s (’94) figures of Salmo salar would
seem to show that'this is the case. In this latter fish Lundborg
shows, in median sagittal sections, a large rounded eminence,
rather than a fold, immediately anterior (aboral) to the hypo-
physial invagination, and anterior to it there is a small ecto- |
dermal fold projecting postero-ventrally from the posterior
surface of the rounded anterior end of the snout. The low and
rounded eminence is said to later become a part of the dorsal
surface of the buccal cavity, and it seems quite unquestionable
that the small ectodermal fold immediately anterior to this
eminence becomes the maxillary breathing-valve. The func-
tional upper lip of the fish, which is here unquestionably a
secondary one, must then lie anterior to this breathing-valve,
and hence must be developed from tissues on the ventral edge of
the rounded anterior end of the snout, and it would seem to be
shown, in process of development, in His’s (’92 b, fig. 31) figure
of a sagittal section of a 20 mm. trout. If then the small fold
which I take to be the maxillary breathing-valve be that valve
and not the primary upper lip, the latter lip must either be repre-
sented in the low and rounded eminence which becomes part of
the dorsal surface of the buccal cavity, or it must lie posterior to
the hypophysis; and as, in the younger embryos figured by

192 EDWARD PHELPS ALLIS, JR.

Lundborg, there is a considerable distance between the hypo-
physial invagination and the oral plate, and as the mandibular
branchial bars certainly lie morphologically posterior to the oral
plate, it would seem as if the primary upper lip must lie between
the plate and the hypophysis.

The conditions in the embryos above referred to thus give
conflicting evidence as to the relations, in the Plagiostomi and
Teleostei, of the primary and secondary upper lips to the hypoph-
ysis, but in embryos of Amia and Acipenser positive evidence
is given of these relations, in these fishes, by Reighard and Mast,
and von Kupffer, respectively. In Amia, Reighard and Mast
(08) say that the adhesive organ is developed from entoblastic
tissues which probably represent the anterior head cavities of
the fish, and that this organ only secondarily acquires connection .
with the ectoblast. This ectoblastic connection, when acquired,
lies between the fundament of the hypophysis and the stomo-
daeum, and the adhesive disk of larvae is developed there; and
as my figures of these larvae of Amia (Allis, ’89) show that the
lips of this fish, both primary and secondary, lie oral to the
adhesive disk, they must both necessarily develop from tissues
oral to the hypophysis. In Acipenser, also, the adhesive organ
lies between the hypophysis and the stomodaeum (von Kupffer,
93), and as the adhesive organ of embryos is said to become the
barbels of the adult (Reighard and Phelps, ’08), and as the
upper lip of the adult lies oral to these barbels, this lip of this
fish, whether it be simply a primary one or a primary and
secondary combined, must also lie oral to the hypophysis.

It thus seems certain that the secondary upper lip, at least,
varies, in different fishes, in its relations to the fundament of the
hypophysis, lying oral to it in the Ganoidei and aboral to it in
Salmo; and this difference in position is associated with the
presence or absence, in embryos of these fishes, of an adhesive
organ. The fold of the secondary upper lip must accordingly
be developed later than either the hypophysis or the adhesive
organ, and, as it pushes forward from the angle of the gape, it
passes oral or aboral to the hypophysis accordingly as that organ
is more or less remote from the oral plate. Keibel’s statement

LIPS AND NASAL APERTURES IN FISHES 193

(06, p. 157) that, in all vertebrates, from the Selachii upward,
the upper edge of the mouth les anterior to the hypophysis, is
accordingly not wholly correct.

The primitive invagination of the mouth, or primary stomo-
daeum, certainly lies posterior to the primary lips, whatever
the relations of those lips to the hypophysis may have been, and
it is represented only in the deeper portion of the depression
enclosed between the so-called mandibular, maxillary and fronto-
nasal processes of embryos, the remainder of that depression
representing the space between the primary and secondary lips.
The stomodaeum of current descriptions of vertebrate embryos
is accordingly something more than the primary stomodaeum,
being, in all the gnathostome vertebrates excepting the Dipneusti,
largely a portion of the external surface of the head which is in
process of being secondarily enclosed between the primary and
secondary lips, and, in the Dipneusti, being largely a space in
process of being enclosed between the primary and tertiary
lips.

The Schnauzenfalte of His’s (’92 b) descriptions of vertebrate
embryos is quite certainly, in certain instances, either the
primary or secondary upper lip. In embryos of the Mammalia
it has strikingly the position of the median portion of the supra-
maxillary fold of the Holocephali and Dipneusti, but it seems
probable that it is not that fold but a special formation peculiar
to the Mammalia, the supramaxillary fold being limited to the
extent that it has in the Teleostomi, and being represented in
that portion of the maxillary process of embryos of the Mammalia
which bounds the lacrimal groove laterally and which is well
shown in Keibel’s (’93) figures of embryos of the pig. The fact
that the lacrimal canal of the adult mammal lies between the
lacrimal bone and the nasal spine of the maxillary bone, and the
probability that the lacrimal and antorbital bones of Amia repre-
sent, respectively, the lacrimal bone and the nasal spine of the
Mammalia, is in favor of this supposition, for it is at just this
place that the anterior end of the supramaxillary furrow turns
upward and ends in Amia. If this be so, it seems worthy of
note that a furrow beneath a fold definitely related, in the

JOURNAL OF MORPHOLOGY, VOL. 32, No. J

194 EDWARD PHELPS ALLIS, JR.

Holocephali, to the tubules of ampullary sacs which lie on the
dorsal surface of the snout, is represented, in the Mammalia, by
a groove which later becomes connected with a glandular
structure also lying in this region.

The position of the secondary upper lip—in certain instances
oral to the two nasal apertures and in others passing between
them—depends upon the position, at the time the fold of this
lip pushes forward, of the nasal apertures relative to the upper
edge of the mouth, and also upon the height of the fold and the
length of the gape of the mouth. Where the gape is short and
the fold of the lip is high, as in Heterodontus, the fold naturally
passes between the two apertures. The extent of the cranial
flexure at the time of the formation of the fold may also have
some influence on its relations to the nasal apertures.

The importance and wide distribution of the labial and supra-
maxillary folds would seem to indicate that the furrows related
to those folds can not be simple adventitious creases in the
external dermis, and the evident inference is that they may
represent persisting remnants of a premandibular cleft or clefts.
This, if so, would not affect any of the conclusions I have arrived
at, for the related arch or arches would still necessarily le
morphologically posterior to the oral plate of embryos. The
mouth could not, however, in that case, be developed from the
mandibular branchial clefts fused with each other in the mid-
ventral line, for the edge of mouth lies anterior to all the labial
folds and furrows. The mouth would then, of necessity, be a
terminal opening formed by the breaking through of the
anterior wall of the gut, that wall being represented in the oral
plate of embryos.

In the Cyclostomata the upper lip lies between the hypoph-
ysis and the oral plate, and it is highly probable that it repre-
sents the primary lip of all vertebrates. If this be so, the lips
in these fishes are, as compared with those in other vertebrates,
primitive and not degenerate structures. His calls this lip the
‘Schnauzenfalte,’ but it is certainly not the ‘Schnauzenfalte’ of
his descriptions of the Mammalia. The supramaxillary fold of
the Holocephali and Dipneusti is perhaps represented in the

LIPS AND NASAL APERTURES IN FISHES 195

slight fold shown projecting ventro-anteriorly dorsal to the nasal
epithelium in His’s figure of a median sagittal section of
Ammocoetes.

Palais de Carnolés,
Menton, France.
Dec. 1, 1916.

LITERATURE CITED

Anus, E. P., Jr. 1889 The anatomy and development of the lateral line system
in Amia calva. Jour. Morph., vol. 2.
1900 The premaxillary and maxillary bones, and the maxillary and
mandibular breathing valves of Polypterus bichir. Anat. Anz., Bd. 18.
1901 The lateral sensory canals, the eye-muscles, and the peripheral
distribution of certain of the cranial nerves of Mustelus laevis.
Quart. Jour. Microsc. Sci., vol. 45.
1903 The skull, and the cranial and first spinal muscles and nerves in
Secomber scomber. Jour. Morph., vol. 18.
1904 The latero-sensory canals and related bones in fishes. Intern.
Monatsschr. f. Anat. u. Physiol., Bd. 21.
1909 The cranial anatomy of the mail-cheeked fishes. Zoologica,
lal, Gy
1913 The homologies of the ethmoidal region of the Selachian skull.
Anat. Anz., Bd. 44.
1916 The so-called labial cartilages of Raia clavata. Quart. Jour.
Microse. Sei. (in press).

BERLINER, K. 1902 Die Entwickelung des Geruchsorganes der Selachier.
Arch. mikros. Anat. Bd. 60.

Born, G. 1879 Die Nasenhéhlen und der Thrinennasengang der amnioten
Wirbelthiere. Morphol. Jahrb., Bd. 5.

Braver, A. 1899 Beitrige zur Kenntniss der Entwicklung und Anatomie der

Gymnophionen. JI, Die Entwicklung der dusseren Form. Zool.

Jahrb., Bd. 12.

BripGeg, 'T. W. 1898 On the morphology of the skull in the Paraguayan Lepi-
dosiren and in other Dipnoids. Trans. Zool. Soe., London, vol. 14.

Burne, R. H. 1909 The anatomy of the olfactory organ of Teleostean fishes.
Proc. Zool. Soc., London.

Daniet, J. F. 1915 The anatomy of MHeterodontus francisci: II, The
endoskeleton. Journ. Morph., vol. 26, no. 3.

Dean, B. 1906 Chimaeroid fishes and their development. Carnegie Inst.,
Wash.

Dourn, A. 1904 Studien zur Urgeschichte des Wirbelthierkérpes. 23. Die
Mandibularhohle der Selachier,—24. Die Primindibularhéhle. Mitth.
aus. der Zool. Stat. zu Neapel, Bd. 17.

FURBRINGER, K. 1904 Beitrige zur Morphologie des Skeletes der Dipnoer nebst
Bemerkungen iiber Pleuracanthiden, Holocephalen und Squaliden.
Jena. Denksch., Bd. 4.

196 EDWARD PHELPS ALLIS, JR.

GaRMAN, 8. 1885 Chlamydoselachus anguineus Garm. A living species of
Cladodont shark. Bull. Mus. Comp. Zool. Harvard Coll., vol. 12.

Gaupp, E. 1906 Die Entwickelung des Kopfskelets. Handbuch d. vergl. u.
exp. Entwickelungslehre d. Wirbeltiere von Oskar Hertwig. Bd. 3,
Meil).2:

1910 Siugerpterygoid und Echidnapterygoid nebst Bemerkungen iiber
das Siuger-Palatinum und den Processus basipterygoideus. Anat.
Hefte, Bd. 42.

GEGENBAUR, C. 1872 Das Kopfskelet der Selachier. Untersuchungen zur
Vergleichenden Anatomie der Wirbeltiere. Heft 3, Leipzig.

1898 Vergleichende Anatomie der Wirbelthiere mit Beriicksichtigung
der Wirbellosen. Bd. 1, Leipzig.

GOpreRT, E. 1906 Die Entwickelung des Mundes und der Mundhohle mit
Drusen und Zunge: die Entwickelung der Schwimmblase, der Lunge
und des Kehlkopfes bei den Wirbeltieren. Handbuch d. vergl. u.
exper. Entwickelungslehre d. Wirbeltiere von Oskar Hertwig. Bd. 2,
Teil 1. Jena.

GrReEIL, A. 1913 Entwickelungsgeschichte des Kopfes und des Blutgefisssys-
temes von Ceratodus forsteri. Zweiter Teil: Die epigenetischen
Erwerbungen wiihrend der Stadien 39-48. Jenaische Denkschriften,
Bd. 4.

Gintuer, A.C. L.G. 1871 Description of Ceratodus, a genus of ganoid fishes
recently discovered in Queensland, Australia. Phil. Trans. Royal
Soe., London.

Hinspura, V. 1901 Die Entwicklung der Nasenhéhle bei Amphibien. Theil I
und II: Anuren und Urodelen. Archiv f. Mikr. Anat. u. Entwickl.,
Bd. 58.

1902 Die Entwicklung der Nasenhodhle bei Amphibien. Teil III:
Gymnophionen. Arch. Mikr. Anat., Bd. 60.

His, W. 1892a Zur allegemeinen Morphologie des Gehirns. Archiv. Anat.
Phys., Anat. Abtheil.
1892b Die Entwickelung der menschlichen und thierischer Physiog-
nomien. Archiv Anat. Phys., Anat. Abtheil.

Husprecut, A. A. W. 1877 Beitrag zur Kenntniss des Kopfskeletes der Holo-
cephalen. Niederl. Archiv f. Zoologie, Bd. 3.

Huxtey, T. H..1876 On Ceratodus forsteri, with observations on the classi-
fication of fishes. Proce. Zool. Soe. London.

Kerset, F. 1893 Zur Entwickelungsgeschichte und vergleichenden Anatomie

der Nase und des oberen Mundrandes (Oberlippe) bei Vertebraten.
Anat. Anz., Bd. 8.
1906 Die Entwickelung der fusseren K6érperform der Wirbeltierem-
bryonen, insbesondere der menschlichen Embryonen aus den ersten
2 Monaten. Handbuch der vergl. u. experim. Entwickelungslehre d.
Wirbeltiere von Oskar Hartwig, Bd. 1, Teil. 1, Zweite Hilfte.

Kurprrer, C. von 1893 Studien zur vergleichenden Entwicklungsgeschichte des
Kop’es der Kranioten: Die Entwickelung des Kopfes von Acipenser
sturio an Medianschnitten untersucht, Heft 1. Munchen.

LIPS AND NASAL APERTURES IN FISHES 197

Lunpsore, H. 1894 Die Entwicklung der Hypophysis und des Saccus vascu-
losus bei Knochenfischen und Amphibien. Zoolog. Jahrb., Abth. f.
Aniaiten dade

Mituer, J. 1834 Vergleichende Anatomie der Myxinoiden, der Cyclostomen
mit durchbohrtem Gaumen. Berlin.

Mi.tuer, J. anp Hentz, J. 1841 Systematische Beschreibung der Plagiosto-
men. Berlin.

Parker, W. Kk. 1876 On the structure and development of the skull in sharks
and skates. Trans. Zool. Soc., vol. 10.

1878 On the structure and development of the skull in the common
snake (Tropidonotus natrix). Phil. Trans. Royal Soe., London.

1882. On the structure and development of the skull in sturgeons
(Acipenser ruthenus and A. sturio). Phil. Trans. Royal Soe., London.

Peter, K. 1906 Die Entwickelung des Geruchsorgans und Jakobson’schen
Organs in der Reihe der Wirbeltiere. Bildung der dusseren Nase und
des Gaumens. Handbuch der Vergl. u. Experim. Entwickelungslehre
d. Wirbeltiere von Oskar Hertwig, Bd. 2, Teil 2.

REIGHARD, J. AND Mast, 8. O. 1908 Studies on Ganoid fishes. II. The de-
velopment of the hypophysis of Amia. Journal of Morphology,
vol. 19, Philadelphia.

REIGHARD, J. AND PHELPS, J. 1908 The development of the adhesive organ
and head mesoblast of Amia. Journ. Morphol., vol. 19.

Rose, C. 1892 Uber Zahnbau und Zahnwechsel der Dipnoer. Anat. Anz.,
Bd. 7, J.

SCHAUINSLAND, H. 1903 Beitrage zur Entwicklungsgeschichte und Anatomie
der Wirbeltiere, 1, 2,3. Zoologica, Bd. 16 (Heft 39).

Semon, R. 1893 Die dussere Entwickelung des Cenatodus Forsteri, Jenaische
Denkschriften, Bd. 4.

THANe, G. D. 1893 Quain’s Anatomy. Osetology, vol. 2, Pt. 1, London.

Verrer, B. 1878 Untersuchungen zur vergleichenden Anatomie der Kiemen-
und Kiefer muskulatur der Fische. II. Tiel, Jena. Zeitschr. f. Natur-
wiss., Bd. 12.

PLATE 1

EXPLANATION OF FIGURES

1 Lateral view of head of Chlamydoselachus anguineus. 4.
2 Front view of same, with mouth forced widely open. X 3.

ABBREVIATIONS

ala, antero-lateral nasal aperture

ama, antero-mesial nasal aperture

a process a of ala nasalis

a’ process a’ of ala nasalis

8 process of 6 of ala nasalis

lag, line of angle of gape

lfd, labial fold

mple, mandibular preangular labial
crease

nbg, naso-buccal groove

nfd, nasal fold

nfl, nasal flap

nfr, nasal frill

nlfd, naso-labial fold

pag, primary angle of gape

pla, postero-lateral nasal aperture

pll, primary lower lip

pma, postero-mesial nasal aperture

pml, premaxillary lip

pul, primary upper lip

sag, secondary angle of the gape

sll, secondary lower lip

smf, supramandibular furrow

smfd, supramandibular fold

smafd, supramaxillary fold

ssll, supplementary secondary lower lip

ssul, supplementary secondary upper
lip

sul, secondary upper lip

tag, tertiary angle of the gape

tul, tertiary upper lip

vp, valvular process

198

LIPS AND NASAL APERTURES IN FISHES PLATE 1
EDWARD PHELES ALLIS, JR.

6
7

PLATE 2

EXPLANATION OF FIGURES

Ventral view of head of Pristis antoquorim. Copied from Miller and
Henle.

Ventral view of head of Scyllium ecanicula. X 2.

Ventral view of head of Chiloseyllium punctatum. Copied from Miiller
and Henle.

Ventral view of head of Heterodontus francisci. X 2.

The same, the lower jaw removed and the nasal frill pulled back on left

hand side of figure. X 2.

200

LIPS AND NASAL APERTURES IN FISHES
EDWARD PHELPS ALLIS, JR.

f am Dan. Sa REN cn 7, dehoaaeeeeeml

LES RRAERR Ses 2.
en eng Sas

el

PLATE 2

PLATE 3

EXPLANATION OF FIGURES

8 Ventral view of head of Mustelus (probably vulgaris). > 1.

9 The same, dissected so as to show the ala nasalis on left hand side of figure,
and the nasal pit on right hand side. X 1.

10 Ventral (external) view of left ala nasalis of Mustelus. X 2.

11 Dorsal (internal) view of the same. X 2.

12 Lateral view of the head of Ceratodus forsteri. X 1.

i)
S
bo

LIPS AND NASAL APERTURES IN FISHES 4 PLATE 3
q EDWARD PHELPS ALLIS, JR.

PLATE 4

EXPLANATION OF FIGURES

13 Lateral view of the head of Chimaera colliei. X 1.

14 Chimaera colliei; view perpendicular to the ventral surface of the snout,
naso-labial fold turned back on Jeft hand side of figure.

15 Chimaera colliei; lateral view of the nasal region, the naso-labial fold
turned back. X 2.

16 The same; the valvular process also turned back. X 2.

204

LIPS AND NASAL APERTURES IN FISHES PLATE 4
EDWARD PHELPS ALLIS, JR.

205

Ms y
.
‘i
z
ein
4
- i
an
ie
i= A
*
Fi
‘
‘
ro
:

ens

ot ray

:

7

+“
ent

ca

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MAY 1

THE MYODOME AND TRIGEMINO-FACIALIS CHAM-
BER OF FISHES AND THE CORRESPONDING
CAVITIES IN HIGHER VERTEBRATES

EDWARD PHELPS ALLIS, JR.

Menton : Pra nee

A functional myodome, or so-called eye-muscle canal, is a
structure peculiar to fishes, and even among them it is limited,
in the fishes I have examined to Amia and the non-siluroid
Teleosts. It is such an important organ in those fishes in
which it is found that it has necessarily received considerable
attention and various suggestions have been made regarding
its origin and development. In my work on the Mail-cheeked
_ Fishes (Allis, ’09) it was discussed at considerable length, and
I came to the conclusion that it was primarily a subpituitary
and intramural space which had been secondarily invaded by
certain of the rectus muscles, entrance to it having been ac-
quired, on either side, through a foramen that transmitted a
cross-comimissural vein which drained the pituitary region and
more particularly the hypophysis. That any part of the de-
finitive myodome formed part of the cavum cerebrale cranil,
that any part of it had been excavated by certain of the rectus
muscles in previously solid portions of the basis cranii, simply
in order to aquire more favorable points of origin, or that any
part of it had been enclosed by the growth of bone or cartilage
developed for that special purpose, I did not believe. I accord-
ingly did not, at the time my manuscript was sent to press,
accept Swinnerton’s (02) contention that, in Gasterosteus, the
anterior portion of the myodome was an actual derivative of the
cavum cerebrale cranii, while its posterior portion was an extra-
mural space secondarily enclosed between the basioccipital and
the underlying parsphenoid. I, however, later received Gaupp’s
CG5 b) work on Saimo, and when I found that he had arrived

207

208 EDWARD PHELPS ALLIS, JR.

at practically the same conclusion as Swinnerton, I added the
following foot-note to my own work (Allis, ’09, p. 195):

Gaupp, in Bd. 3 of Hertwig’s Handbuch der vergleichenden und
experimentellen Entwickelungslehre der Wirbeltiere, a work that I
have only seen since this manuscript was sent to press, describes prac-
tically similar conditions in Salmo [to those described by Swinnerton
in Gasterosteus], and arrives at practically similar conclusions regard-
ing the homologies of the parts. This would seem to establish the fact
that the basioccipital portion of the myodome is extracranial in origin.
Regarding the prootic portion of the myodome, Gaupp’s descriptions
would seem to confirm my contention that it is an intramural space
and not an intracranial one.

According to the views set forth in the works above referred
to, both Swinnerton and Gaupp maintain that the myodome
owes its origin to the fact that certain of the muscles of the eye-
ball, which primarily had their points of origin on the external
surface of the chondrocranium, forced their way into the cavum
cerebrale cranii, foreed the brain upward considerably above
the basis cranii, and then, after having thus displaced and
certainly disturbed the delicate central nervous organ, forced
their way out of the chondrocranium to again acquire origin
on its external surface, and then became secondarily enclosed
there in a canal developed for that special purpose. This has
always seemed to me improbable, notwithstanding my provi-
sional and somewhat qualified acceptance of it, and I have
long intended investigating the development of this canal when-
ever I could obtain suitable material, my series of sections of
somewhat advanced teleostean embryos not being considered
suitable for the purpose.

I, however, recently’ had occasion, in connection with other
work, to examine a series of sections of a 5l-mm. specimen
of Hyodon tergisus, and noticed that the myodome was di-
rectly continuous, posteriorly, with a groove on the ventral
surface of the basioecipital that lodged the anterior portion of
the median dorsal aorta. This at once suggested that the basi-
occipital portion of the telostean myodome, and hence possibly
the entire myodome, might be a canal of vertebral origin com-
parable to the haemal canal of the tail, for that canal is not

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 209

always limited to the caudal region, as the conditions in Aci-
penser show (Bridge, ’(04, fig. 115, p. 200).

Accordingly, with this idea definitely in view, I have care-
fully examined the myodome, not only in this series of sections
of Hyodon and in a single prepared skull that I have of the adult
fish, but also in all other embryos and adults of the Holostei
and Teleostei that I at present have at my disposal that seem
to be of interest in this connection, and the result has been not
only to strongly favor this interpretation of the conditions,
but also to give a conception of the myodome itself and the bones
related to it quite different from that I formerly held. The
functional myodome, as found in the fishes examined, will first
be quite fully described, and then comparison made with the
descriptions of the corresponding parts in certain other fishes
and in certain of the higher vertebrates.

The preliminary examination of the serial sections used in
connection with the work was wholly done by my assistant,
Mr. John Henry, camera drawings being made of many sec-
tions of each series. The drawings used for the figures are by
my assistant, Mr. Jujira Nomura.

HYODON TERGISUS

The myodome of Hyodon has never been described, so far as
I can find, except by Ridewood, (’04), who only says that, in
Hyodon alosoides: ‘‘The parasphenoid underlies but a small
portion of the basioccipital, and the eye-muscle canal opens
at its posterior end by an oval foramen.”’

In my skull of the adult Hyodon tergisus, the anterior open-
ing of the myodome is triangular and unusually large and tall
for the size of the skull. Posterior to this opening the myodome
diminishes rapidly in size, and finally becomes continuous with
an open groove on the ventral surface of the basioccipital. This
groove extends to the hind end of the basioccipital and there cuts
through the ventral edge of the vertebra-like hind end of the bone
to open upon its posterior surface. This open groove forms no
part of the myodome as described by Ridewood, his myodome
ending at the oval foramen described by him, which leads from

210 EDWARD PHELPS ALLIS, JR.

the myodome directly into the anterior end of the groove. The
groove quite unquestionably lodged, in the fresh specimen, the
anterior portion of the dorsal aorta, as it does in my 51-mm. speci-
men, and it may accordingly be called the aortal groove, the term
myodome being limited to that canal as decribed by Ridewood.

The anterior portion of the floor of the myodome, as thus
defined and limited, is formed by the parasphenoid, the ascending
processes of which form the ventral portions of its side walls.
The middle portion of the floor is probably formed by the syn-
chrondosis, in the mid-ventral line, dorsal to the parasphenoid,
of the ventral ends of the ventral processes of the prootics, for
that is the condition in my 5l-mm. specimen, but, as I do not
wish to destroy my one skull of the adult of this fish, I cannot
definitely say that this is so. A short posterior portion of the
floor is probably formed, like its anterior portion, by the para-
sphenoid, for it is so formed in the 51l-mm. specimen. Slightly
posterior to the sutural line between the hind edges of the pro-
otics and the anterior end of the basioccipital, the parasphenoid
separates into two diverging hind ends which extend poste-
riorly a certain distance, there resting upon the ventral edges
of the bounding walls of the aortal groove.

The dorsal portions of the side walls of the anterior portion
of the myodome and the entire side walls of its posterior portion
are formed by the ventral processes of the prootics, which are
overlapped externally by the lateral edges of the parasphenoid.
The roof of the myodome is formed by the horizontal proc-
esses of the prootics, the so-called prootic bridge or shelf, but
whether these processes suturate with each other in the median
line or are separated by a median line of cartilage, I cannot tell
from my specimen for the reason above given. The prootic
bridge is perforated near its anterior edge by a small median
foramen, the so-called pituitary opening of the brain case of
my descriptions of other fishes, the bridge thus having post-
pituitary and prepituitary portions.

The basisphenoid, which, as Ridewood says, has no verti-
cally descending process, suturates posteriorly with the ante-
rior edge of the prepituitary portion of the prootic bridge; later-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 211

ally, on either side, with the related alisphenoid, and anteriorly
with the orbitosphenoid. It is perforated by a single median
foramen which transmits the two optic nerves, this apparently
agreeing with Ridewood’s description of this bone, for his
statement that it ‘‘forms the superior edge of the optic foramen’”’
must mean that the two optic nerves traverse it through a single
opening.

As already stated, the groove on the ventral surface of the
basioccipital of my 5l-mm. specimen lodges the anterior por-
tion of the median dorsal aorta. When, proceeding anteriorly,
the aorta begins to widen, preparatory to separating into a
lateral dorsal aorta on either side, it recedes from the groove
and is replaced by the hind ends of the musculi recti externi;
these muscles soon occupying the entire groove, the aorta lying
ventral to them and outside the groove. The lateral edges of
the groove give insertion to the tunica externa of the air-
bladder, the tissues of the tunica forming, in the posterior, but
not the anterior portion of the groove, an arched bridge beneath
the aorta and so enclosing it in a canal; this being as described
by Bridge (99) in. Notopterus. The notochord, enclosed in
the basioccipital, lies directly above the bottom of the groove,
separated from it by only a thin layer of bone of perichordal
origin.

In sections through the extreme hind end of the basioccipital
(fig. 12) the aortal groove is shallow and les directly beneath
the notochord, between blocks of cartilage which are unques-
tionably the homologues of the lower arches, or basiventrals,
of current descriptions of the vertebrae, but which I shall refer to
as the ventrolateral vertebral processes. On each of these proc-
esses there are two slight ridges: a ventromesial one, clothed
with perichondrial bone that represents a part of the hind end
of the basioccipital, and a dorsolateral one, not clothed with bone,
which gives attachment to a ligament running outward in an
intermuscular septum and doubtless representing either a dorsal
or a ventral rib. In sections slightly farther forward (fig. 11)
the ventrolateral cartilaginous processes have entirely disap-
peared, but are represented by parts of the basioccipital which

212 EDWARD PHELPS ALLIS, JR.

are of membrane origin. The dorsolateral ridge of the process of
the preceding figure has disappeared, but the ventromesial one is
represented by a tall ridge of bone which bounds laterally on
either side a deep aortal groove. The dorsolateral vertebral
processes (upper arches, basidorsals) are here represented by
two large blocks of cartilage in relation to which the exoccipi-
tals are developed.

Each exoccipital is perforated by two occipital nerves, the
anterior one represented by a ventral root alone and the poste-
rior one by both dorsal and ventral roots. Anterior to these
two nerves a delicate ventral root arises from the medulla, but
it: does not reach the internal surface of the cranial wall.

On the ventral surface of the first free vertebra there is a
slight depression, but no aortal groove, the space between the
ventrolateral vertebral processes of opposite sides being com-
pletely filled by bony deposits.

Comparing these conditions with those in the adult Amia, it is
evident that the ventromesial ridges of the ventrolateral cartilag-
inous processes of Hyodon, bounding laterally the aortal groove,
are the homologues of the little cartilaginous-processes on the ven-
tral surface of the hind end of the basioccipital of Amia. There
are in Amia, as is well known, two pairs of these little carti-
laginous processes—called by Hay (95) aortal supports, by
Schauinsland (05) haemal processes—and Schauinsland says
that they are related to certain vertebrae that were said by
Sagemehl (’83) to have fused with the hind end of the primor-
dial cranium of this fish. Schauinsland further says that Sage-
mehl found one pair of these processes, but that he himself finds
two; but these two pairs had already been described by both
me (’97) and Schreiner (02). The space between each pair of
these processes is almost completely filled by bony deposit, the
aortal groove thus being obliterated here, as it is also shown
by Hay (95, fig. 1) to be in a transverse section of one of the
anterior dorsal (trunk) vertebrae of a 12.5-em. specimen of
this fish. In the last dorsal vertebra of this same specimen,
Hay shows (l.c., fig. 6) this space not so completely filled by bony
deposit, and in the first caudal vertebra it is even still less so

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 213

(l. c., fig. 7). If, in those vertebrae that are known to have
fused with the occipital region of the cranium of this fish, the
deposit of bone between the aortal supports had been as re-
stricted as it is in the first caudal vertebra, conditions similar
to those actually found in Hyodon would have arisen. The pos-
terior portion of the aortal groove of Hyodon is thus certainly
enclosed between processes of vertebral origin, but whether these
processes are the exact homologues of the haemal processes of
the tail is open to some question, for there is marked want of
accord in the descriptions of the formation of the latter.

In Amia the aortal supports (haemal processes, Schauins-
land) are said by both Hay and Schauinsland to be primarily
cartilaginous, and to be simply differentiated parts of the bases
of ventrolateral vertebral processes (lower arches). Posterior °
to the twenty-fourth vertebra, these supports are said by Hay
to be developed apparently independently of the main mass
of the ventrolateral vertebral processes (lower arches), and in
the posterior region of the trunk they are said to be forced away
from the notochord by bony deposits, and to each there become
attached to the ventral surface of the remaining portion of the
related ventrolateral process, which is then called a parapoph-
ysis. In the tail region the aortal supports are said by Hay
to entirely disappear, and this one statement, together with
the several figures given, would lead one to suppose that it is
the remaining portions only of the ventrolateral processes, the
so-called parapophyses, that form the haemal arches of the tail.
The descriptions are, however, not clear as to this. What Hay
actually says is (95, p. 16):

In the vertebrae of the tail the cartilages [aortal supports] are miss-
ing. There,is, however, in my younger specimen, what seems to be
vestiges of them in the first caudal vertebra. Nothing, however, can
be more certain than that the lower arches of the trunk are bent down

to form the arches of the tail, and that the aortal supports have
nothing to do with the formation of the caudal haemal arches.

On a later page he, however, says: ‘‘In the tail the halves of
each lower arch have united at their distal ends, so as to enclose
the blood vessels.” It may accordingly be that Hay con-

214 EDWARD PHELPS ALLIS, JR.

sidered the haemal arches to be formed by the entire ventro-
lateral processes, and this is what Schauinsland (’05) says of
these arches in all fishes. In Laemargus, Schauinsland even
shows (l. c., p. 411) the aortal supports (his haemal arches) pro-
jecting mesially from the mesial surfaces of the entire ventro-
lateral processes and partly separating the haemal canal into
dorsal and ventral compartments which lodge, respectively,
the aorta and the caudal vein.

In Polypterus the haemal arches have, as described by Bud-
gett (02), a totally different origin from that above set forth.
In a 30-mm. specimen of this fish Budgett finds three distinctly
separate series of cartilaginous vertebral processes, one dorsal,
one lateral, and the other ventral. The lateral processes bear
the upper ribs, which have the positions of the ribs in the Se-
lachi. The ventral processes bear the lower ribs, which have
the position of the ribs in the Teleostei. It is said that, in the
caudal region, ‘‘the lateral series of cartilages are not found, while
the ventral cartilages, though retaining their position, become
the greatly enlarged haemal arches.” These latter arches are
thus here formed by processes that are certainly not the homo-
logues of the so-called lower arches of the Selachii. In the
trunk region of specimens of Polypterus older than the 30-mm.
one, the ventral processes are said to be forced away from the
notochord by bony deposits formed in relation to the lateral
processes and these bony deposits are shown, in Budgett’s
figures, forming so-called aortal supports on either side of the
aorta. The ventral processes thus forced away from the
notochord are then found as blocks of cartilage in the bases of
the ventral ribs at some distance from the notochord and loosely
attached to the under sides of the lateral processes; these ven-
tral processes of this fish thus strikingly resembling the aortal
supports of Hay’s descriptions of Amia.

The aortal supports and haemal arches, as those terms are
employed by English authors, may thus be of different origin
in different fishes, but, whatever their origins and homologies
may be, the lateral walls of the aortal groove of Hyodon, in the
posterior basioccipital region here under consideration, are quite

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 215

certainly formed by aortal supports, the remainder of each
primitive ventrolateral vertebral process being represented in
the little cartilaginous ridge that gives attachment to the
ligament which runs outward, rib-like, in the related intermus-
cular septum.

Returning now to the descriptions of Hyodon and proceeding
forward in the sections, figure 10 shows that a ridge of bone grad-
ually appears lateral to the tall ridge that bounds on either side
the aortal groove, the appearance in these sections somewhat
suggesting, excepting in the absence of cartilage, the conditions
shown in Hay’s figures 5 and 6 of trunk vertebrae of Amia,
where the aortal supports are attached to the ventromesial sur-
faces of the lower arches (parapophyses, Hay). Still farther
forward, in sections through the hind ends of the parasphenoid
(fig. 9), the mesial one of these two processes has disappeared
while the lateral one persists as a stout low process, this giving
a broad ventral edge to the aortal groove. The lateral walls of
the aortal groove are now formed by the entire ventrolateral
processes, and not simply by the aortal supports, and the bony
deposits on either side that fill the space between these proc-
esses and the dorsolateral vertebral processes (upper arches)
has been excavated to form the recessus sacculi. Hence the
basioccipital is here W-shaped in transverse section, the two
grooves on the dorsal surface of this W each forming the ven-
tral portion of the related recessus sacculi, and the grooves of
opposite sides being separated from each other by a tall median
plate formed by part of the basioccipital. The notochord lies in
the ventral end of this median plate, dorsal to the bottom of the
aortal groove, but is here represented simply by a notochordal
space. The exoccipital of either side has vertical and horizon-
tal plates, the former forming the dorsal portion of the side wall
of the related recessus sacculi and the mesial portion of the other
the roof of the recessus, the mesial edge of the latter plate rest-
ing upon the dorsal end of the tall median plate of the basioc-
cipital and forming, with its fellow of the opposite side, the
floor of the cavum cerebrale cranii. The median groove on
the ventral surface of the W here lodges the hind ends of the

216 EDWARD PHELPS ALLIS, JR.

musculi recti externi, with the dorsal aorta lying ventral to
them.

Proceeding anteriorly from this point to sections through the
bases of the diverging hind ends of the parasphenoid (fig. 8),
the bony bounding walls of the aortal groove are gradually re-
placed by cartilage lined with thin plates of perichondrial bone
which form parts of the basicccipital. Angles in this cartilage
and bone now replace the two bony ridges, just described, in more
posterior sections. The perichondrial bone then disappears, in
the region of the hind end of the myodome, leaving the bounding
walls of the groove entirely of cartilage, and slightly anterior
to that point the remaining portions of the basioccipital also
vanish. The notochord extends forward nearly to the hind end
of the myodome, its anterior end lying dorsal to the bottom of
the aortal groove and hence in the level of the roof of the myo-
dome and not in that of its floor. In this region the aorta has
separated into a lateral dorsal aorta on either side.

Anterior to the bases of the diverging hind ends of the para-
sphenoid, the aortal groove is closed ventrally by the latter
bone, and, still lodging the musculi recti externi, becomes the
hind end of the myodome. Except that the groove is here
closed ventrally by the parasphenoid and that it lies in the pro-
otic region, there is no line of demarcation between it and the
open canal in the basioccipital region, and each broad ventral
edge of the open groove, lying between the two angles above .
described, is continued forward as the ventral edge of the lateral
bounding wall of the myodomie canal.

Proceeding forward in the sections, there is gradual ventral
growth of the cartilaginous side walls of the myodomiec canal,
this growth taking place between the two little angles above
described. This gives rise to a flange of cartilage on either edge
of the primitive groove, the flange projecting ventrally and
slightly mesially beneath the level of the dorsolaterally pro-
jecting basal portion of the lateral wall of the cranium, the base
of that portion of the wall lying in the level of the ventral edge
of the primitive aortal groove (fig. 7). Proceeding anteriorly,
these flanges increase gradually in actual height, and appear to

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 217

gain additional height because of the gradual widening of the
dorsal portion of the myodomic cavity, cross-sections through
which change gradually from oval to pear-shaped and then to
triangular. The roof of, the myodome thus becomes flat (fig. 6,)
instead of being arched (fig. 7). The cartilage forming the top
of the arched roof is continued forward as the median portion of
the flat roof, and is enclosed between plates of perichondrial
bone which do not meet in the median line and which form the
distal (mesial) portions of the horizontal processes of the pro-
otics. The lateral portions of these processes and the dorsal
portions of the side walls of the myodome are now each formed
by two plates of bone, doubtless of perichondrial origin but
without enclosed cartilage, this bone replacing the cartilage
of the preceding sections and forming the dorsolateral corners
of the myodomie cavity. The horizontal portion of each of
these angles of bone forms the basal (lateral) portion of the
horizontal process of the prootic of its side, and arises from
the base of the lateral wall of the cavum cerebrale cranii. The
vertical portion of the angle of bone forms a wall between the
dorsal portion of the myodome and the ventral portion of what
is, in the prepared skull of the adult, a large bay on the exter-
nal surface of the cranium. This bay forms that part of the
large auditory fenestra of Ridewood’s (’04) descriptions that
lies anterior to the so-called vertical lamina of the prootic, and
its floor is formed, in my embryo as in the adult, by a laterally
projecting, horizontal shelf of the prootic. This bay of this
fish corresponds to the facialis part of the trigemino-facialis
chamber of my description of Scomber and the mail-cheeked
fishes, and occupies a position, relative to the cranial: walls,
similar to that of the recessus sacculi, the floor of the bay being
an anterior continuation of that of the recessus. The truncus
hyomandibularis facialis enters this bay through a foramen
in its mesial, cranial wall, and runs outward above its floor.
The vena jugularis, traced from behind forward, enters the bay
over the posterior edge of its floor, accompanied by a sym-
pathetic nerve, a communicating branch from the nervus glosso-
pharyngeus to the nervus facialis, and the arteria carotis ex-
terna, this artery lying ventromesial to the other structures.

218 EDWARD PHELPS ALLIS, JR.

The foramen faciale perforates the prootic posterior to the
postorbital process of the neurocranium, the foramen trigem-
inum perforating the same bone anterior to that process, the
two foramina both leading directly into the cavum cerebrale
eranii. Between these two foramina the postorbital process
of the cranium is perforated by a short canal, the floor of which
lies at the level of the roof of the myodome and hence dorsal
to the posterior portion of the floor of the facialis bay. The
later bay leads directly into this canal, the canal itself leading
into the orbit and transmitting the vena jugularis, the arteria
carotis externa, a communicating branch from the nervus fa-
cialis to the nervus trigeminus, and a sympathetic nerve. This
canal is thus a jugular canal through the prootic bone, and it
represents all there is, in this fish, of the trigemino-facialis
chamber of my description of others of the Teleostei. There
are in this region of the cranium of fishes three distinctly dif-
ferent chambers. One is the trigemino-facialis recess of my
descriptions of the Teleostei and Selachii; another is the jugular
canal through the prootic, just referred to and which I have here-
tofore called the teleostean trigemino-facialis chamber; and the
third is a chamber formed by the fusion of the other two, and is
the trigemino-facialis chamber of my descriptions of Amia. It is
accordingly necessary to distinguish between these several cham-
bers, and the term trigemino-facialis chamber will hereafter be
limited to the chamber as found in Amia, the two parts of the
chamber being called its pars ganglionaris and pars jugularis.

The ventral edges of the side walls of the myodome of Hyo-
don are nowhere enclosed in perichondrial bone, cartilage always
projecting ventrally beyond the related bone and abutting
against dense connective tissue that separates it from the par-
asphenoid. This tissue is apparently all skeletogenous, for
there is no definite perichondrial membrane separating it from
the cartilage. The parasphenoid develops in the outer layers
of this tissue, and the bases of the diverging hind ends of that
bone are connected by it across the median line (figs. 8 and 9),
the tissue there forming a dense and well-defined band-like layer.
Farther posteriorly this transverse band becomes less dense,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 219

and then practically disappears, but there is always a line con-
necting the two ends of the parasphenoid, and hence also con-
necting the ventral edges of the aortal groove.

Approximately in the transverse plane of the foramen fa-
ciale, a cartilaginous process projects mesially from the ventral
end of each cartilaginous side wall of the myodome and meets
its fellow of the opposite side in the median line, but it does not
completely fuse with it, a slight line of separation always re-
maining evident. These two processes thus together form a
cartilaginous floor to the myodome, the parasphenoid lying
against the ventral surface of this floor, but separated from it
by the dense skeletogenous tissue above referred to. Imme-
diately posterior to this point, the pharyngobranchial of the first
branchial arch articulates with the dorsal portion of the side
wall of the myodome, there lying between the vena jugularis
and the external and internal carotid arteries; and immediately
posterior to that, the pharyngobranchial of the second branch-
ial arch articulates with the ventral surface of the parasphe-
noid (fig. 7). The external and internal carotid arteries sepa-
rate from each other slightly anterior to the latter point, both
lying along the lateral surface of the lateral wall of the myo-
dome. The external carotid runs forward and upward, ventral
to the nervus facialis, and, joining the vena jugularis, traverses,
with that vein, the short canal which represents the pars jugu-
laris of the trigemino-facialis chamber. The internal carotid
continues forward and downward along the side wall of the myo-
dome until it reaches the hind edge of the ascending process
of the parasphenoid, where it traverses a foramen which is, as
in the adult, entirely enclosed in that bone.

Beginning immediately posterior to this foramen (fig. 5) for
the internal carotid artery and proceeding forward in the sec-
tions, the cartilage forming the floor of the myodome, and also
the ventral ends of its. side walls, gradually disappears and is
replaced by the dense skeletogenous tissue already referred to
several times, and in it a cavity appears, bounded on all
sides by the tissue and lying between the parasphenoid and the
myodomic cavity. The floor and side walls of this cavity

220 EDWARD PHELPS ALLIS, JR.

form a matrix, in relation to which the body and ascending
processes of the parasphenoid are developed, and from” here
forward for a certain distance teeth are found developed in re-
lation to this bone. The roof of the cavity forms a membrane
which extends transversely from the ventral end of one per-
sisting cartilaginous side wall of the myodome to the other, this
membrane being horizontal in position in its posterior portion,
but arching upward anteriorly to such an extent that, in the
subpituitary region, its summit reaches nearly to the middle
of the height of the entire myodomic cavity. The parasphenoid
has in this region been inclining quite rapidly ventrally, this,
and the arching upward of the membrane, leaving a space be-
tween the two and separating the myodomic cavity into dor-
sal and ventral compartments. The ventral compartment, lim-
ited to the region of the ascending processes of the parasphe-
noid, is bounded both laterally and ventrally by that bone.
The dorsal portion of the dorsal compartment is bounded later-
ally by the ventral processes of the prootic bone, its ventral por-
tion being bounded in part by the ventral portions of those proc-
esses, overlapped externally by the ascending processes of the
parasphenoid, and in part by the latter processes only. The
dorsal compartment still lodges the recti externi, the ventral
compartment lodging the hind ends of the recti interni and the
internal carotid arteries, the two being separated by a delicate
line of tissue (fig. 3).

Slightly posterior to the internal carotid foramina in the para-
sphenoid, the roof of the myodome, formed by the horizontal
processes of the prootics, is traversed at each lateral edge by
both the nervus abducens and the ramus palatinus facialis, appa-
rently through a single foramen (fig. 4). The abducens goes
directly to the muculus rectus externus. The palatinus facialis
runs ventrally along the internal surface of the side wall of the
dorsal compartment of the myodome, passes through a notch
in the anterior edge of the ventral end of the prootic portion of
that wall, which is wholly of cartilage, and then, continuing
ventrally between that cartilaginous wall and the ascending
process of the parasphenoid, enters that portion of the ventral

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 221

compartment of the myodome that is occupied by the inter-
nal carotids (fig. 3). In its passage along the mesial surface of
the lateral wall of the myodome it lies between the wall and a
delicate layer of connective tissue which everywhere lines the
myodomie cavity, thus apparently not definitely entering the
central cavity of the myodome.

Slightly anterior to the point where the prootic bridge is per-
forated by the nervi abducens and palatinus facialis, and slightly
anterior also to the transverse plane of the internal carotid
foramina, the median cartilaginous portion of the prootic bridge
ceases, and the roof of the myodome is then perforated by what
is, in the prepared cranium of the adult fish, the pituitary open-
ing of the brain case (fig. 1 to 3). This opening is closed, in
fresh specimens by a portion of the dura mater that projects
ventrally into the myodome and so forms a pit-like depression
in the floor of the cavum cerebrale cranii, in which the pituitary
body lies; thus forming the actual pituitary fossa. It will, how-
ever, be best to call it the pituitary sac, for the term pituitary
fossa, and its equivalent sella turcica, has been given to the de-
pression that, in the floor of the cartilaginous or osseous cranial
cavity, lodges this pituitary sac, and the two are not always
coincident. The sac forms the roof of this part of the myodome
of Hyodgn, and a median vertical membrane descends from its
ventral and anteroventral surfaces. Anteriorly this membrane
is directly continuous with the membranous interorbital septum;
posteroventrally it is continuous with the anterior edge of the
median portion of the horizontal myodomic membrane, the
lateral portions of the latter membrane here being so broken up
and interrupted by the muscles and vessels entering or leav-
ing the myodome that they cannot be followed in the sections.
The vertical membrane does not at this point extend ventrally
to the floor of the myodome, but in the transverse plane of the
hind edge of the basisphenoid (fig. 1) it becomes the interorbital
septum, and there its flaring ventral edges are each attached to
a ridge on the related lateral edge of the dorsal surface of the
parasphenoid. In the triangular space enclosed between the
latter bone and the V-shaped ventral end of the septum lies,

JOURNAL OF MORPHOLOGY, VOL. 32, No. 2

222 EDWARD PHELPS ALLIS, JR.

on either side, the related nervus palatinus facialis, the nerves
of opposite sides being separated from each other by a median
ridge on the dorsal surface of the parasphenoid, and each ac-
companied by a small branch of the internal carotid, given off
just before that artery enters its foramen in the parasphenoid.
This branch does not enter the myodome, but runs forward in
a canal in the parasphenoid between what seem to be portions
of the bone that are the one of membrane and the other of dental
origin.

The rectus inferior muscle of either side has its origin on the
dorsal portion of the interorbital septum, near the level of the
posterior edge of the basisphenoid. The rectus superior has its
origin on the anterior edge of the horizontal myodomic mem-
brane. The pituitary vein of either side enters the dorsal com-
partment of the myodome in the subpituitary region, and, run-
ning posteriorly in it, joins its fellow of the opposite side pos-
terior to the membranous pituitary sac, there forming a large
sinus. From this sinus branches are sent to the rectus externus
muscles, and from its anterior end a small median branch is sent |
upward, through the membranous roof of the myodome, into
the cavum cerebrale cranii, where it immediately breaks up and
cannot be followed in the sections.

The pituitary vein of either side is joined by veins from the
eyeball and the eye-muscles, these together forming what Allen
(05) has called the internal jugular vein. I have also employed
his term in certain of my works, in others calling it simply the
jugular vein. This latter term is certainly the only one that can
be appropriately employed, for the vein is the definitive vena
jugularis, and as it is formed in part by the vena capitis media
and in part by the vena capitis lateralis, neither of these terms
can be employed excepting to designate certain sections of it.

The internal carotid artery of either side traverses its foramen
at the hind edge of the ascending process of the parasphenoid
and enters the ventral portion of the ventral compartment of the
myodome. There it gives off the orbitonasal artery and then,
running forward into the subpituitary portion of the myodome,
turns upward in that part of the median vertical myodomic

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 223

membrane which forms the anterior wall of the membranous
pituitary sac. In this part of its course, and while still enclosed
in the median vertical membrane, it anastomoses with its fel-
low of the opposite side, and then, separating from its fellow,
enters the cavum cerebrale cranii (fig. 2). There it immedi-
ately divides into anterior and posterior divisions. The pos-
terior division sends branches to the hypophysis, and then itself
separates into anterior and posterior branches. The anterior
branch runs forward along the floor of the cavum cerebrale
eranii and sends a branch outward in the body of the optic
nerve. Other branches are sent to the brain, one of them join-
ing, anterior to the nervus opticus, the terminal portion of the
anterior division of the entire artery. In one of two specimens
examined, the latter division of the artery immediately issued
from the cranial cavity by passing ventrally across the posterior
edge of the basisphenoid, while in the other specimen it perfor-
ated that bone near its hind edge. In each case the artery then
ran forward ventral to the horizontal plate of the basisphenoid,
enclosed in the dense fibrous tissues that there form the dorsal
edge of the interorbital septum. While in this tissue a branch
is sent outward to the eyeball, the artery then issuing from the
fibrous tissue, passing across the posterodorsal surface of the
nervus opticus, and entering the cranial cavity through the fora-
men for that nerve. There it joins and fuses with the anterior
branch of the posterior division of the entire artery, just de-
cribed, the artery so formed then running forward along the
floor of the cavum cerebrale cranii. The branch sent to the
eyeball from the anterior division of the artery, enters it close
to the point of entrance of the nervus opticus, and there im- :
mediately forms a slight enlargement which somewhat resembles
a glomus. From this glomus a branch arises and unites with
the small artery accompanying the nervus opticus, the two to-
gether forming the arteria centralis retinae.

I cannot recognize the anterior division of the internal ca-
rotid artery, above described, in any descriptions that I have of
the adults of fishes, and yet it is found in all of the non-siluroid
Teleostei that I have examined in serial sections in connection

224 EDWARD PHELPS ALLIS, JR.

with the present work. The fact that, in one of my two speci-
mens of Hyodon, it perforates the basisphenoid, is peculiar, and
it is to be noted that in that specimen this bone has a greater
anteroposterior extent than in the other, extending posteriorly
beyond the sutural line between the alisphenoid and prootic,
instead of ending anterior to that line, as in the other specimen.
This, when compared with the conditions in the other fishes ex-
amined, to be described later, would seem to indicate that the
basisphenoid of Hyodon is not strictly comparable to that bone
in those other fishes.

On one side of my 51l-mm. specimen of Hyodon the efferent
pseudobranchial artery entered the ventral compartment of the
myodome with the internal carotid, through the foramen for
that artery. On the other side it perforated the ascending proc-
ess of the parasphenoid anterior to the internal carotid, sepa-
rated from it by a narrow column of bone. Having entered the
ventral compartment of the myodome, in its subpituitary por-
tion, it passes ventral to the orbitonasal artery and is connected
with its fellow of the opposite side by a cross-commissural ves-
sel which passes anteroventral to the internal carotids. The
efferent pseudobranchial artery then itself runs outward into the
orbit, as the arteria opthalmica magna, to enter the eyeball and
there supply the chorioid gland.

In this 5l-mm. embryo, as in the adult, the alisphenoid bone
has no pedicel (parasphenoid leg), this pedicel being represented
by membrane only, as it is, wholly or in part, in many other Tele-
ostei (Allis, 09). The pedicel or so-called vertical descending
process (Ridewood) of the basisphenoid is also wanting, as al-
- ready stated, that bone being represented by its horizontal plate
alone. In those Teleosts in which this bone has a pedicel, its
hind edge forms the median vertical anterior boundary of the
myodome, and the anterior edge of the median vertical myo-
domic membrane is attached to it. When the pedicel is want-
ing, as in Hyodon, the vertical myodomic membrane runs in-
sensibly into the membranous interorbital septum, and there
is nothing to mark definitely its anterior limit.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 225

The myodome of my 51-mm. specimen of Hyodon thus lies
in part beneath the pituitary opening of the brain case and in
part posterior to that opening, a part of it thus being prechor-
dal and the remainder chordal in position, or, in the terminology
employed by Froriep (’02 a), the one prespinal and the other
spinal. The spinal portion is formed, throughout part of its
length, by two distinctly different parts, one dorsal and the other
ventral, the two being completely separated from each other
in part by cartilage and in part by membrane which forms a
direct anterior prolongation of the cartilage. In the prespinal
portion these two compartments of the spinal portion are
confluent because of the breaking down of the separating wall
(the horizontal myodomic membrane) by the structures that
here enter or leave the dorsal compartment. The canal tra-
versed by the internal carotid arteries as they run upward in the
median vertical myodomic membrane lies in the level, anteriorly
prolonged, of the dorsal myodomic compartment, but it forms
no part of either compartment of the myodome.

The dorsal compartment of the myodome of Hyodon is lim-
ited to the subpituitary and postpituitary regions, and although
both of these parts le in the prootic region, the postpituitary por-
tion, which lies beneath the prootic bridge, may be alone re-
ferred to as the prootic portion of the compartment. The ventral
compartment has prootic, subpituitary, and prepituitary por-
tions. The dorsal compartment is directly continuous posteri-
orly with the anterior end of the aortal groove, which extends
the full length of the basioccipital region. The ventral com-
partment is not continuous with the groove and it does not
extend posteriorly as far as the dorsal compartment. It lies
between the floor of that compartment and the parasphenoid,
lodges the hind ends of the recti interni, and is traversed by the
internal carotid arteries, the palatine branches of the faciales,
and the efferent pseudobranchial arteries. The dorsal com-
partment lodges the pituitary veins and the musculi recti ex-
terni, these muscles entering it from the orbits and leaving it by
its posterior opening. The nervus abducens of either side perfo-
rates the roof of this compartment to reach and supply the rectus

226 EDWARD PHELPS ALLIS, JR.

externus. The ramus palatinus facialis also perforates the roof
and traverses the dorsal compartment in order to reach the
ventral one, but it is separated from the central cavity of the
dorsal compartment by a membrane, apparently of skeletog-
enous character, the nerve thus probably lying morphologically
in the wall of this compartment of the myodome and not actu-
ally traversing it.

The recti externi, after issuing through the posterior opening
of the dorsal compartment of the myodome, extend posteriorly a
certain distance, there lying in a part of the aortal groove which
differs slightly in character from the part posterior to it. This
anterior part of the groove is, however, so evidently an anterior
prolongation of its posterior portion that the two parts must
be of similar origin, and as the posterior portion of the groove
has certainly not been developed in any relation whatever to
any of the muscles of the eyeball, it is certain that the anterior
portion also has not been so developed. This is, furthermore,
confirmed by the conditions in Polypterus, in which there is no
functional myodome, but there is both a cavity corresponding
to the dorsal compartment of the myodome of Hyodon and a
closed and wholly separate canal lodging the cranial portion
of the aorta and corresponding to the aortal groove of Hyodon.
This myodomic cavity and aortal canal have both been referred
to and discussed in certain of my earlier works (Allis,’08 a,’09),
and I now find, on reexamining my sections of a small specimen
of this fish, that the enclosing walls of the aortal canal give even
more positive evidence of having been formed by vertebral proc-
esses than do the walls of the groove of Hyodon.

There thus seems little doubt that the bounding walls of the
aortal groove of Hyodon are formed by processes similar to
those which enclose the haemal canal of the tail, and that those
bounding walls are formed either by the entire ventrolateral
processes of vertebrae which here have been incorporated in the
neurocranium, or by aortal supports developed in relation to
those processes; and if the walls of this groove are so formed, it
would seem as if the side walls of the prootic portion of the dor-
sal compartment of the myodome, evidently an anterior con-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 227

tinuation of the walls of the aortal groove, must be of similar
origin. Why the aorta has been excluded from this prootic
portion of the myodome is not apparent, but it would seem as
if it might be related to the development of the hypochorda.
According to Stohr (’95), the hypochorda of Rana, when first
formed, is attached to the dorsal wall of the alimentary canal
by a series of tubular bridges, which persist longer in the ante-
rior than in the posterior region of the trunk, and, for a time,
there prevent the lateral dorsal aortae from fusing with each
other in the median line excepting between the bridges. In the
head region the hypochorda is said to develop later than in the
trunk, and the related bridges would hence there also, while they
persisted, prevent the lateral dorsal aortae from fusing with each
other excepting between the bridges. It may then be that,
such a bridge persisting in the prootic region, the lateral dorsal
aortae could not there fuse with each other, and before this
bridge had disappeared they had become fixed in position by
the early development of the anterior aortic arches. Anterior
to the spinal region of the cranium they, however, fused with
each other, in certain fishes, that point either representing an
interval between two hypochordal bridges, or lying anterior to
the anterior bridge, as the case may be. This would then not
only explain the formation of the circulus cephalicus, but also
account for its position external to the ventral processes of the
prootics.

In further support of the assumption that the ventral proc-
esses of the prootics are formed by ventrolateral vertebral proc-
esses 1s the fact, possibly significant, that these processes, like
the neural processes in the spinal region, enclose a large
canal between their proximal portions and a smaller one be-
tween their distal ends, the two cavities being separated from
each other by a horizontal partition. In my 5l-mm. specimen
of Hyodon this partition is partly of cartilage and partly of
membrane. In all the other fishes examined it is wholly of
membrane, excepting as that membrane may have’ undergone
ossification as part of the parasphenoid, a median longitudinal
opening thus being left, when the parasphenoid is removed, be-

228 EDWARD PHELPS ALLIS, JR.

tween the ventral ends of the ventral processes of the prootics.
This opening is the hypophysial fenestra of Sagemehl’s descrip-
tions of Amia and the Teleostei, and I have always employed
that term for it in all my works. This fenestra has, however,
in a considerable part of its length, no relation whatever to the
hypophysis, and it will be later shown that, in all probability,
it does not even contain the so-called fenestra hypophyseos of
early embryos of these fishes. The term hypophysial fenes-
tra is thus inappropriate, and I shall hereafter refer to it as
the fenestra ventralis myodomus. To facilitate the descriptions
and comparisons, the ventral processes of the prootics will be
considered to be ventrolateral vertebral processes, notwithstand-
ing that this is not definitely established by my present work.

SCOMBER SCOMBER

In the adult Scomber I found (Allis, 03) the myodome ex-
tending nearly to the hind end of the basioccipital but not open-
ing posteriorly; and, doubtless in direct correlation with this,
the parasphenoid of this fish does not have diverging hind ends.
That part of the myodome that is related to the basioccipital is
enclosed between ventral flanges of that bone which closely re-
semble the ventral processes of the prootics and form a direct
posterior continuation of them. Two membranes, one vertical
and the other horizontal, were said to extend the full length of the
myodome. The horizontal membrane was said to separate the
myodome into dorsal and ventral parts which lodged, respec-
tively, the recti externi and interni. The vertical membrane
was said to arise from the hind edge of the pedicel of the basi-
sphenoid and, lying between the recti interni, to bisect the ven-
tral part of the myodome. ‘The recti inferiores were said to arise
partly from the interorbital septum, between the foramen op-
ticum and the anterior edge of the basisphenoid, and partly
from a ligament or tendon which arises from the dorsal end of
the pedicel. of the basisphenoid. The recti superiores were said
to have their origins from the anterior edge of the horizontal
membrane.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 229

The above statements all referred only to the adult of this fish,
for I at that time had no specimens small enough to be sectioned.
I have, however, since prepared a series of transverse sections
of a 65-mm. specimen, in which [ now find the vertical mem-
brane above referred to descending from the ventral surface of
the membranous pituitary sac, as in Hyodon. That part of it
which, in the adult, lies beneath the horizontal membrane and
extends to the hind end of the myodome is, in this 65-mm. spec-
imen, simply a delicate line of connective tissue, but it would
nevertheless seem to represent a remnant of a wall which pri-
'marily separated this part of the myodome into two parts, one
on either side, as will be later explained. The horizontal mem-
brane is practically as I described it in the adult. In its ante-
rior portion it is not strongly developed, and there arises, on
either side, from a layer of tissue which lines the internal surface
of the side wall of the myodome cavity and is continued outward
around the ventral end of the wall and then upward a certain
distance along its external surface. The parasphenoid rests, on
either side, upon the ventral surface of this tissue, and a longi-
tudinal ridge on either side of the dorsal surface of the bone
projects upward into that part of the tissue which lines the in-
ternal surface of the side wall of the cavity; this part of the par-
asphenoid certainly being an ossification developed in relation
to the tissue. Along the line of origin of the horizontal mem-
brane, the cartilage of the side wall of the myodome is slightly
constricted and imperfect, suggesting a segmentation line similar
to that shown by Schauinsland (’05) where a rib is in process
of being segmented from a lower arch in the vertebral region
of certain other fishes. Near the hind end of the myodome,
beginning slightly anterior to the point where the recti interni
terminate, that part of the cartilage of each lateral wall of the
myodome that lies ventral to this segmentation line gradually
passes, without any line of demarcation, into dense fibrous tis-
sue which forms the ventral end of each lateral wall of the my-
odome. The parasphenoid here lies against the ventral surface
of this tissue, and the longitudinal ridge on either side of the

dorsal surface of the bone extends upward along the mesial sur-
¥

230 EDWARD PHELPS ALLIS, JR.

face of the tissue. The horizontal membrane is strongly devel-
oped here, and extends across the median line between the
ventral ends of the persisting portions of the cartilaginous side
walls, thus lying at a certain distance dorsal to the parasphenoid.
Proceeding posteriorly from here, in the sections, the recti interni
disappear, leaving a space between the horizontal membrane
and the parasphenoid. The latter bone then shortly disappears,
and the myodomiec cavity is then closed ventrally by the hori-
zontal membrane only, this condition possibly persisting to the
hind end of the myodome, but my sections are here imperfect
and I cannot definitely determine this.

The conditions in Scomber would thus arise from those in
Hyodon if the ventral compartment of the myodome of the lat-
ter fish were extended posteriorly nearly to the hind end of the
aortal groove, and the continuous myodomic-aortal cavity so
formed where closed ventrally, to that point, by the para-
sphenoid.

The internal carotid artery of either side enters, as in Hyodon,
the ventral compartment of the myodome, runs forward in it
into the prespinal portion of the myodome, and there turns up-
ward in the median vertical myodomic membrane, anastomos ng,
while in the membrane, with its fellow of the opposite side.
Leaving its fellow, it separates, as in Hyodon, into anterior and
posterior divisions both of which run upward, posterior to the
basisphenoid, and enter the cavum cerebrale cranii. The pos-
terior division sends branches to the hypophysis and then sepa-
rates into anterior and posterior branches, the anterior branch
running forward along the floor of the cavum cerebrale cranii,
sending a branch outward with the nervus opticus, and then
joining the anterior division of the artery, this anterior pro-
longation of this branch of the artery not being found in Hyodon.
The anterior divison of the artery runs forward, dorsal to the
basisphenoid, and, anterior to that bone, enters the thick, dense
tissues forming the floor of the cavum cerebrale cranii and the
dorsal end of the interorbital septum, its course and distribu-
tion from there onward being as in Hyodon. The fact that
this anterior division of the artery runs forward dorsal to the

MYODOME AND TRIGEMINO-FACIALIS CHAMBER Pay A

basisphenoid would seem to show that, as already stated, the
anterior portion of the basisphenoid of Hyodon did not pri-
marily form part of this bone.

‘The cross-commissure connecting the efferent pseudobran-
chial arteries of opposite sides traverses the ventral part of the
‘prespinal portion of the myodome, there passing ventral to the
orbito-nasal artery and anteroventral to the internal carotids.
The pituitary veins enter the dorsal compartment of the my-
odome and there anastomose with each other, a small branch
being sent upward into the cavum cerebrale cranii and appar-
ently going to the hypophysis.

The nervus palatinus facialis of the adult traverses a canal
in the prootic which begins in the floor of the pars jugularis of
the trigemino-facialis chamber and opens on the mesial surface
of the ventral process of that bone in the plane of the hind edge
of the pituitary opening of the brain case, the nerve thus ap-
parently not traversing the dorsal compartment of the myo-
dome. In the 65-mm. specimen the nerve does not enter the
pars jugularis of the trigemino-facialis chamber, perforating the
roof of the dorsal compartment of the myodome and traversing
it, as in Hyodon, but, as also in Hyodon, there lying between
the ventral process of the prootic and the lining membrane of
the myodomic cavity.

MAIL-CHEEKED FISHES (LORICATI)

In the adults of Scorpaena scrofa, Trigla hirundo, and Cot-
tus octodecimospinosus, I found (Allis, 709) the myodome to
extend nearly to the hind end of the basioccipital, and there
open ventrally. In Scorpaena the origins of all the rectus
muscles were given, the external and internal ones extending
posteriorly in the myodome, the external somewhat farther than
the internal. Nothing was said of a horizontal membrane sep-
arating the myodome into dorsal and ventral compartments,
such as I had previously described in Scomber and now find
in Hyedon. The orbital opening of the myodome was said
to be closed by a strong membrane which the recti externi and
interni perforated to reach their points of origin.

232 EDWARD PHELPS ALLIS, JR.

In a 40-mm. specimen of Scorpaena scrofa I now find the recti
externi and interni as described in my earlier work, but they are
separated by a membrane, delicate in places but well developed
in others, which corresponds to the horizontal membrane of
Scomber and Hyodon and separates the myodome into dorsal
and ventral compartments. The rectus superior of either side’
arises in part from the anterior edge of this membrane and in
part from the dorsal surface of the parasphenoid at or near the
line where the lateral edge of the membrane is attached to it,
this line being marked by a slight longitudinal ridge on the
dorsal surface of the bone. The rectus inferior of either side
arises from a median vertical membrane similar to that de-
scribed in Hyodon and Scomber.

The internal carotid artery, after traversing its foramen at
the hind edge of the ascending process of the parasphenoid,
passes across an internal carotid incisure at the antero-
ventral corner of the prootic cartilage, as in the adult (Allis,
09, p. 411), and enters the ventral compartment of the myo-
dome, its farther course and distribution being as in Secomber.
The cross-commissure of the efferent pseudobranchial arteries
traverses the subpituitary portion of the myodome, as in Hyodon
and Scomber. The pituitary veins anastomose with each other
in the dorsal compartment of the myodome, but they do not
there form an important sinus.

In the adult the ramus palatinus facialis traverses a canal
in the prootic which begins in the trigemino-facialis recess (pars
ganglionaris of the trigemino-facialis chamber) and opens on
the internal surface of the ventral process of the prootic. In
my 40-mm. embryo this nerve perforates a membranous por-
tion of the prootic bridge, and, running ventrally between the
side wall of the dorsal compartment of the:-myodome and the
lining membrane of that cavity, as in Hyodon and Scomber,
enters the ventral compartment in the subpituitary region and
then escapes into the orbit.

In a 40-mm. specimen of Trigla hirundo the conditions are
practically as in Secorpaena, excepting that the anterior branch
of the posterior division of the internal carotid artery is inter-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 233

rupted, either anterior or posterior to the branch sent out with
the nervus opticus, as it is in Hyodon, while a branch, not found
in either Hyodon or Scorpaena, is sent outward, anterior to the
nervus opticus, to join a branch of the orbitonasal artery and
then go to the eyeball.

In a 63-mm. specimen of Trigla hirundo the conditions dif-
fer in that the depression in the dura mater which lodges the
hypophysis has a relatively wide and flat floor from which three
membranes arise, one median and one at each lateral edge of
the floor. These membranes are each inserted on a correspond-
ing ridge on the dorsal surface of the parasphenoid, and the
space enclosed, on either side, between them and the para-
sphenoid, lodges the rectus internus. Thus the ventral com-
partment of the myodome here rises to the ventral surface of
the pituitary depression, and hence lies between right and left
halves of the dorsal compartment. This condition continues
posterior to the hypophysis for a certain distance, the roof of
the ventral compartment of the mycdome there forming the
median portion of the floor of the cavum cerebrale crani; but
at the membranous anterior edge of the prootic bridge, the roof
of the compartment begins to recede from the floor of the cavum
cerebrale cranii, and, the lateral halves of the dorsal compart-
ment uniting with each other above it, the conditions become
as in Scorpaena. Apparently because of this intercalation of
the ventral compartment between the anterior ends of the
dorsal compartment, the pituitary veins are greatly reduced,
the hypophysis being drained in part by the encephalic veins.

In this embryo of Trigla the nervus palatinus facialis per-
forates the floor of the pars jugularis of the trigemino-facialis
chamber and enters the dorsal compartment of the myodome,
this apparently being as I found this nerve in the adult Secomber.

In the adult Cottus octodecimospinosus I found (Allis, ’09)
the myodome continued posteriorly a short distance in the
basioccipital, and not opening posteriorly on the ventral sur-
face of the cranium. The prootics have perfectly normal hori-
zontal processes, and they are shown, in my figures, preformed
in cartilage and forming the roof of the myodome. ‘The para-
sphenoid has diverging hind ends.

234 EDWARD PHELPS ALLIS, JR.

In a 20-mm. specimen of Cottus scorpeus I now find the pro-
otic portion of the myodome separated from the cavum cere-
brale cranii by membrane only, no cartilaginous or osseous pro-
otic bridge being as yet developed. In its basioccipital portion
the myodome lies in a groove on the ventral surface of the basi-
occipital (fig. 18), which opens posteriorly between the diverg-
ing hind ends of the parasphenoid, but is there closed ventrally
by membrane which extends horizontally between those ends.
This part of the myodome lodges the hind ends of the recti
externi, the two muscles being separated from each other by a
delicate vertical membrane. Proceeding anteriorly in the sec-
tions, the thin cartilage forming the roof of the myodomic
groove runs gradually into membrane (fig. 17), the entire basis
cranii thus here being perforated by a longitudinal opening
that might be considered to be a fenestra ventralis myodomus.
This is, however, not the case, for the recti externi lie definitely
in this opening and not above it. The bounding walls of the
opening accordingly represent the side walls of the myodomic
groove, and the space between the ventral edges of the side walls
alone represents the fenestra ventralis myodomus. The space
between the dorsal edges of the opening is a perforation of the
floor of the primordial cranium, and the membrane extending
horizontally between the edges forms part of the floor of the
cavum cerebrale cranii and also the roof of the dorsal compart-
ment of this basioccipital portion of the myodome. The recti
externi lie between this membrane and the parasphenoid, and
they are still separated from each other by a median vertical
membrane. The saccus vasculosus is large, lies in the cavum
cerebrale cranii, and projects posteriorly slightly beyond this
point.

Proceeding anteriorly in the sections to the region between
the saccus vasculosus and the hypophysis (fig. 16), the mem-
branous roof of the myodome becomes somewhat arched, and
it now has its attachment, on either side, on the dorso-internal ,
surface of the cartilage of the basis cranii, at some distance
dorsolateral to the midventral perforation of the cartilage, that
perforation now being definitely a fenestra ventralis myodomus.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 235

A second membrane, evident also in the preceding figure, here
closes the fenestra ventralis myodomus, the parasphenoid lying
directly against its ventral surface. The median vertical mem-
brane still extends between these two membranes, separating
the myodome into lateral halves.

Still farther forward in the sections, in the region of the hind
end of the hypophysis (fig. 15), the median portion of the mem-
brane that forms the roof of the myodome, and hence also the
median portion of the floor of the cavum cerebrale cranii, grad-
ually descends, between the recti externi, on to the membrane
forming the floor of the myodome, the two membranes fusing
with each other there, and so forming a thick membrane which
is both the median portion of the floor of the cavum cerebrale
cranii and the median portion of the membrane closing the
fenestra ventralis myodomus, the parasphenoid lying directly
upon its ventral surface. The myodomic cavity, which here still
belongs only to the dorsal compartment of the myodome, is
thus separated into lateral halves, the hypophysis projecting
ventrally between the two halves of the compartment, and
each half lodging the related musculus rectus externus.

Still farther forward in the sections (figs. 13 and 14), the hind
ends of the recti interni appear between the parasphenoid and
the membrane closing the fenestra ventralis myodomus, that
membrane thus being the horizontal myodomic membrane,
and the space beneath it the ventral compartment of the my-
odome. The membrane forming, on either side, the roof of
the related lateral half of the dorsal compartment of the myo-
dome, now has its mesial attachment on the dorsal surface of
the horizontal myodomic membrane, the latter membrane thus,
in its lateral portions, separating the two compartments of
the myodome, while its median portion forms part of the roof
of the ventral compartment of the myodome and part of the
floor of the cavum cerebrale cranii. The conditions here are
accordingly similar to those in the 63-mm. specimen of Trigla
hirundo.

Still farther forward, that lateral part of the horizontal mem-
brane that, on either side, separates the two compartments

236 EDWARD PHELPS ALLIS, JR.

of the myodome, breaks down, but its median portion still
persists as part of the floor of the cavum cerebrale cranii, and,
anterior to the hypophysis, it is perforated by the internal
earotids in their passage from the myodome into the cavum
cerebrale cranu.

Thus the prootic bridge of this small specimen of Cottus
scorplus is nowhere formed by cartilage, and if it be of carti-
lage in the adult, it must be a later chondrification of the mem-
brane that, in this specimen, forms the floor of the cavum cere-
brale cranil. That this does take place is probable, for a car-
tilaginous prootic bridge is developed relatively late in other
fishes also, as will be explained later.

In a 37-mm. specimen of Clinocottus analis the conditions
resemble those in Cottus scorpeus, differing only in that the
horizontal myodomic membrane takes no direct part in the
formation of the floor of the cavum cerebrale cranii, simply
arching upward to such an extent that it is in contact with, and
partly fused with, the membranous prootic bridge, thus sep-
arating the myodome into median and lateral, instead of dorsal
and ventral compartments (fig. 19).

The internal carotid arteries of Cottus and Clinocottus are
strictly similar in their course and branches to those of Trigla.
The cross-commissure of the efferent pseudobranchial arteries
has a position strictly similar to that of the latter fish, and the
nervus palatinus facialis of Cottus is as in Scorpaena, while
that of Clinocottus is as in Trigla. An anterior portion of the
ascending process of the parasphenoid has the position of, and
replaces, the alisphenoid of Amia.

In none of these small specimens of the Loricati is either the
dorsal or the ventral compartment of the myodome definitely
closed toward the orbit by membrane, as, in my earlier work on
these fishes, I said was the case in the adults. Both in embryos
and the adult the spinal portion of each compartment opens
into the prespinal portion, and in embryos this latter portion
is largely open toward the orbit. In the adult the myodome
is doubtless closed toward the orbits by connective tissues which
develop around the rectus muscles as they enter it. In the

MYODOME AND TRIGEMINO-FACIALIS CHAMBER Dok

adults of these fishes I described a well-developed trigemino-
facialis recess. In embryos this recess is not evident, but it
must necessarily exist, potentially.

SYNGNATHUS ACUS

In a 115-mm. specimen of this fish the myodome and the fen-
estra ventralis myodomus are both limited to the prootic re-
gion. Posterior to the hind end of this fenestra is a shallow
median groove on the ventral surface of the cartilaginous basis
eranii, which extends into the basioccipital region and there
lodges the hind end of the parasphenoid. This bone is tri-
angular there, in transverse section, the apex of the triangle
directed dorsally. In sections passing through the posterior
portion of the fenestra ventralis myodomus, the parasphenoid
is still triangular, and the cartilage bounding the fenestra on
either side becomes entirely enclosed in perichondrial bone
which forms part of the prootic. The parasphenoid lies directly
between the ventromesial edges of these prootic bones, dense
connective tissue filling the space between the parasphenoid
and either prootic and also extending dorsally across the para-
sphenoid, there filling and closing the fenestra ventralis myo-
domus. Proceeding anteriorly in the sections, the parasphe-
noid becomes flatter and wider, and the cartilage in the ventral
ends of the prootics vanishes. Further forward in the sections,
a little space appears in the dense connective tissue that covers
the dorsal surface of the parasphenoid, and in this space the
hind ends of the recti externi soon appear (fig. 23), lying directly
above the parasphenoid and separated from the cavum cere-
brale cranii by membrane which continues the full length of
the myodome and represents the prootic bridge.

Proceeding forward from this point, the parasphenoid begins
to widen and at the same time to thicken dorsoventrally, and
it soon has, in sections, a median circular portion with laterally
projecting flanges, each flange being formed of external and
internal plates which receive the ventral end of the ventral
process of the prootic between them (fig. 22). In the rounded
median portion of the bone a median cavity forms, and lodges

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

238 EDWARD PHELPS ALLIS, JR.

the hind ends of the recti interni, the recti superiores arising
from the lateral walls of this cavity, near its anterior end. The
recti externi lie dorsal to the parasphenoid, between it and the
membrane which everywhere forms the roof of the myodome.

Still further forward, the bony roof of the median cavity in
the parasphenoid is gradually replaced by a horizontal myo-
domic membrane which separates the myodome into two com-
partments, a dorsal one lodging the recti externi and a ventral
one lodging the recti interni and superiores, the ventral compart-
ment forming a semicircular depression in the floor of the entire
myodomie cavity (fig. 21). The hind end of the ventral com-
partment is thus completely enclosed in the parasphenoid, and
it seems absolutely certain that that part of the bone forming
the roof of this compartment is simply an ossification of the
horizontal myodomic membrane.

Proceeding anteriorly, the region of the ascending processes of
the parasphenoid is soon reached, these processes rising to the
level of the membranous roof of the myodome and suturating
with the ventromesial edges of the prootics (fig. 20). The
myodome is here semicircular in transverse section, its side wall
and floor being formed wholly by the parasphenoid and its roof
by membrane that separates it from the cavum cerebrale cranii.
A median vertical membrane here descends from the membranous
roof of the myodome, and in connection with it the recti infe-
riores have their origins.

The course and the main branches of the internal carotid
artery are as in Scomber, except that the artery separates into
its anterior and posterior divisions while still within the canal
in the median vertical myodomic membrane, and that the
anterior division then immediately enters the tissues forming
the floor of the cavum cerebrale cranii, thus not actually enter-
ing the latter cavity.

HIPPOCAMPUS GUTTULATUS

In a 20-mm. specimen of this fish the myodome begins pos-
teriorly beneath a part of the basis cranii that is of cartilage
lined, on either side, with perichondrial bone. The parasphe-
noid lies at a certain distance ventral to this part of the basis

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 239

eranil, the space between the two being filled with dense con-
nective tissue which is bounded laterally at its dorsal edges
by little projecting flanges of perichondrial bone developed
in relation to the cartilage of the basis cranil. The hind end
of the myodome lies in this tissue, and lodges the hind ends
of the recti externi. Proceeding anteriorly from this point,
the cartilaginous roof of the myodome soon vanishes and is re-
placed by a thick layer of fibrous tissue. Cartilage is, however,
now found in the ventral portion of each lateral wall of the myo-
domic cavity, this cartilage being enclosed between projecting
flanges of the parasphenoid, one of these flanges lying along
the external surface of the cartilage and the other along its
internal surface. The internal flange lies in the fibrous tissue
that lines the myodome, and is certainly developed in relation
to it.

‘The internal carotid artery traverses its foramen at the hind
edge of the ascending process of the parasphenoid, and then
immediately enters and runs upward in the median vertical my-
domic membrane, its course and distribution there being as
in Sygnathus. The recti interni, superiores and inferiores
have their origins anterior to this ascending column of the artery,
close together, from the dorsal surface of the parasphenoid.
The ventral compartment of the myodome is thus here wholly
prespinal in position, for the foramen for the internal carotid
artery lies in the plane of the pituitary opening of the brain
case. The relations of these two openings to each other varies
considerably in different fishes, the foramen for the artery lying
markedly anterior to the pituitary opening in Scorpaena, but
posterior to that opening in Scomber.

CATOSTOMUS

In Catostomus teres, Sagemehl (’91) describes a myodome
that is everywhere closed ventrally by the parasphenoid, is
bounded dorsally by the horizontal processes of the prootics,
and apparently extends posteriorly slightly into the basioccipital.
The basioccipital has a large pharyngeal process, perforated
by a short canal which encloses the dorsal aorta, and Sagemehl

240 EDWARD PHELPS ALLIS, JR.

says (91, p. 516) that this relation to the aorta at once suggests
a lower vertebral arch. He, however, says that he finds weighty
reasons against the assumption that it is such an arch. One
of these reasons is that, excepting in this region of the Cypri-
nidae and in the tail region of all fishes, the lower arches always
enclose the body cavity, and not simply the aorta. A second
reason is that he himself finds, in embryos of Chondrostomus
nasus, the pharyngeal process not preformed in cartilage, as
the lower arches always are. Sagemehl accordingly concludes
that the pharyngeal process of the Cyprinidae is not a lower
vertebral arch, and he considers it to be a bone formed by
the fusion of pharyngeal bones of dermal origin with another
bone formed by the ossification of a ligament which, in the
Characinidae, extends from the hind end of the basis cranii to
the swim-bladder, embracing the aorta in its course.

In a 57-mm. specimen of Catostomus occidentalis I find the
phyaryngeal process formed by two ventrally projecting lon-
gitudinal flanges of bone which arise from a layer of bone sur-
rounding the notochord, and, diverging slightly and straddling
the aorta, abut against and fuse with the dorsal surface of a
curved and porous plate which les parallel to the dorsal sur-
face of the pharynx (fig. 29). The aorta is thus enclosed in
a canal that corresponds strictly to the aortal groove of Hy-
odon, except in that it is closed ventrally by the formation of a
horizontal floor across its outer edges, and if the one is of verte-
bral origin, as I consider it to be, the other certainly also is.
Whether the floor of the canal has been developed in primary
continuity with its lateral walls, or as an independent dermal
formation, as Sagemehl suggests, cannot be told from my sec-
tions. The lateral walls of the canal are prolonged anteriorly
beyond its floor, and the aorta there lies (fig. 28) in an open
groove similar to that of Hyodon, the lateral walls of the
groove gradually diminishing in height and vanishing approx-
imately in the level of the anterior end of the persisting noto-
chord. Anterior to the point where the vacuolated contents of
the notochord can last be recognized in the sections, the noto-
chorda_ space still continues a certain distance, and in sections

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 241

passing through this region the hind ends of the parasphenoid
are cut, one lying on either side of a median ridge on the ventral
surface of the basioccipital, the median dorsal aorta lying ven-
tral to the ridge (fig. 27). Here, unfortunately, one or more
sections are missing in my series. In the next anterior existing
section (fig. 26) there is a circular space in the basioccipital,
in exactly the position of the notochordal space in the next pos-
terior section of the series, but somewhat larger than it, and this
space lodges the hind ends of the recti externi. ‘The aorta here
begins to separate into a lateral dorsal aorta on either side.

Proceeding anteriorly in the sections, the myodomic cavity
increases in size, and that part of the basioccipital in which it
lies forms a large median, dorsally projecting, and rounded
ridge. The parasphenoid now extends across the median line
fig. 25). Still farther forward the fenestra ventralis myodomus
begins, the parasphenoid closing it ventrally and having a
broad but low median ridge which projects upward into the
fenestra. This median ridge then sends upward a longitudinal
ridge on either side, and in the space between these two ridges
the recti interni make their appearance, separated from the recti
externi by loose connective tissue which does not form a definite
membrane (fig. 24).

Farther forward, the two lateral ridges on the dorsal surface
of the parasphenoid vanish, leaving a flat median ridge, and
the hind end of the hypophysis is there cut in the sections. This
latter organ is large, lies wholly in the myodome, and projects
posteriorly ventral to the roof of the myodome, here formed by
the horizontal processes of the prootics. Anterior to the an-
terior edges of these latter processes the hypophysis is connected
with the brain by a small stalk of nervous material, which per-
forates the membrane which there forms the roof of the myo-
dome and the floor of the cavum cerebrale cranii.

The myodome of this fish is thus, up to this point, strictly
normal, except that the hind end of its dorsal compartment
is enclosed in the basioccipital, and that the recti externi appa-
rently have their origins on the anterior end of the notochord
instead of ventral to it.

242 EDWARD PHELPS ALLIS, JR.

In sections immediately anterior to those that cut through
the nervous stalk of the hypophysis, the median ridge on the
dorsal surface of the parasphenoid extends upward to the mem-_
branous floor of the cavum cerebrale cranii, and so occupies
the position of a basisphenoid, which bone is said by Sagemehl
to be absent in all of the Cyprindidae. A slight line separates
this projecting process from the remainder of the parasphe-
noid, vaguely suggesting a fusion of two bones.

BLENNIUS GATTORUGINI

In Blennius gattorugini I described (’09) a myodome, the roof
of which was said to be formed by membrane. This is correct,
but it was also said that this membrane was attached, on either
side, to the dorsal edge of a groove on the ventral edge of the
prootic, and that that edge which represented the horizontal proc-
ess of the bone. This is incorrect, for, on reexamining my mate-
rial, I find that this groove simply lodges the related portion of the
lateral edge of the parasphenoid and that the membrane repre-
senting the horizontal processes of the prootics is attached to
a slight ridge on the side wall of the cranial cavity at a some-
what higher level, in direct posterior continuation of the line
of the horizontal portion of the basisphenoid. The cavity thus
formed lodges the recti externi only, and hence represents the
dorsal compartment of the myodome. Having no sections
of this fish, the arteries, veins and nerves, and the myodomic
membranes could not be properly traced, but the conditions
are apparently similar to those in Hippocampus, above de-
scribed. Starks (’05) says that in six genera of the Bleniidae
examined by him there was no myodome.

ARGYROPELACUS

In Argyropelacus a myodome is frequently referred to by
Handrick (’01), but not particularly described, and I made
brief reference to it in my work on the mail-cheeked fishes.
The neurocranium of this fish is said by Handrick to be wholly
of cartilage, no bone being found in any part of it, and the my-
odome lies external to this chondrocranium. Its hind wall is

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 243

said to be formed by the anterior, external wall of the bulla
acustica, and it lodges the extracranial semilunar and ciliaris

ganglia. The posterior portion of its roof is evidently formed
_ bya horizontal plate of cartilage shown, in the figures given
by him (i. c., figs. 7 to 9, Pl. 1), lying ventral to the hypophysis,
and which must accordingly represent the prootic bridge. In
a figure of a transverse section in the postfacialis region a ven-
trally projecting process is shown at each lateral edge of this
prootic bridge, and its ventral portion has apparently been cut
off in the figure. These two processes certainly represent
transverse sections of ventral processes of the prootics, sim-
ilar to those found in other Teleostei, and they must form the
lateral walls of the so-called extracranial myodome. Each
process lies mesial to the foramen faciale of its side, and lateral
to this foramen and also lateral to the foramen trigeminum
there is a slight ridge of cartilage which must represent a dorsal
portion of the lateral wall of the pars jugularis of a trigemino-
facialis chamber. The ganglion trigeminum thus probably les
in the orbital opening of this chamber and not in the dorso-
lateral corner of the myodome, as Handrick concluded. The
basis cranil is perforated by a so-called ‘Pituitargrube,’ which
is said to extend from the foramen trochleare nearly to the fo-
ramen trigeminum, is shown closed by a membrane which is per-
forated by the nervi optici, and extends posteriorly to the an-
terior edge of the prootic bridge. This so-called pituitary fossa_
is thus simply a perforation of the primitive cranial wall which
has been formed by the fusion of the pituitary opening of the
brain case with the foramina optici.

Supino (’01), in a work I did not have at my disposal when
my paper on the mail-cheeked fishes was sent to press, finds
several bones developed in relation to the neurocranium of this
fish, two of them being the prootics and one the parasphenoid.
This latter bone must evidently lie ventral to the myodome,
and in a figure giving a ventral view of the entire neurocran-
ium, extensive ventral processes of the prootics are shown which
must form the lateral walls of the myodome. The foramina
for the nervi trigeminus and facialis are said to perforate the

244 EDWARD PHELPS ALLIS, JR.

prootic, but they are not shown in the figure. It is however
probable that the conditions resemble those in Hyodon, the
myodome evidently being large and having a large orbital open-
ing on either side. Whether or not there are both dorsal and
ventral compartments to the myodome cannot be told, but
they are probably both present. Supino says that a basi-
sphenoid is found, which must accordingly separate the so-called
pituitary fossa of Handrick’s descriptions into a pituitary open-
ing of the brain case and a foramen formed by the fusion of the
foramina optici, and he adds that: ‘‘ Posteriormente l’estremita
delle porzione impari del basisfenoide si congiunge, nel Chau-
loides e Argyropelacus, con la cartilagine che si trova nella
cavitaé dei muscoli oculari.”’ This, while not quite clear, would
seem to mean that cartilage formed some part of the floor of
the myodome.

ESOX

In the adult Esox the myodome is large and extends pos- -
teriorly into a conical excavation in the anterior end of the basi-
occipital, as Huxley (71) has stated. A horizontal membrane
separates it into dorsal and ventral compartments, the dorsal
one being large and lodging the recti externi, while the ventral
one is short, extending, posteriorly only to the hind edges of the
foramina for the internal carotid arteries. From there the
membrane separating the two compartments rises rapidly to
the ventral surface of the relatively deep membranous pitui-
tary sae, fuses with that surface, and then appears as two sep-
arate membranes, each having its mesial attachment on the
ventrolateral surface of the sac. The recti interni have their
origins beneath this membrane, on the floor of the ventral com-
partment of the myodome, near its hind end. The recti sup-
periores have their origins from the anterior edge of each half
of the horizontal membrane, near its dorsomesial end, and the
recti inferiores from the lateral walls of the spreading dorsal
end of the median vertical membrane, immediately an-
terior to the membranous pituitary sac. This median vertical
membrane is party fused with the median portion of the hori-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 245

zontal membrane, and it is traversed by the internal carotid
arteries. ‘The membranous roof of the subpituitary portion
of the myodome extends forward from the horizontal proc-
esses of the prootics to the hind edge of the horizontal plate
of the basisphenoid. The conditions in this fish are thus wholly
normal.

Starks (’05) says that the dorsal end of the basisphenoid
(dichost, Starks) of this fish is ‘free,’ and he suggests that this
should be examined in connection with a ‘myodome septum’,
formed of connective tissues, said by him to be found in this
region and to be continued forward as the interorbital septum.
Just what this myodome septum is is not clear, but it would seem
to be the membranous roof of the myodome. Starks further
says that “the dichost (=basisphenoid of Huxley) is always
absent when the myodome is.’’ No particular cases are cited,
but it is evidently assumed that there is no myodome whenever
a prootic bridge (shelf) is not found in the prepared skeleton
of the cranium. ‘This is incorrect, and the statement should
probably be that there is no basisphenoid whenever the roof
of the myodome is wholly of membrane. Whether or not this
statement is true, even in this form, I do not know, my material
being too limited to permit me to form an opinion.

GASTEROSTEUS ACULEATUS

The early stages of the development of the myodome in Gas-
terosteus aculeatus are quite fully described, and the myodome
of the adult briefly described by Swinnerton (’02). The trabec-
ula and parachordal of either side are said by him to be, when
first formed, wholly independent cartilages, and their adjoin-
ing ends are shown in the figures lying approximatively in the
tranverse plane of the tip of the notochord. The posterior
halves of the trabeculae are said to enclose the infundibulum
and the pituitary body, and the infundibulum is shown lying
posterior to the pituitary body and reaching to the tip of the
notochord. ‘The trabeculae and parachordals soon fuse with
each other, and there is then a marked anterior growth of the
parachordals which carries the trabeculae forward considerably

246 EDWARD PHELPS ALLIS, JR.

anterior to the tip of the notochord. The pituitary body still
lies between the hind ends of the trabeculae, in the so-called
pituitary fossa, but the infundibulum now lies dorsal to the
anterior end of the notochord. The space between the anterior
ends of the parachordals is now called the interparachordal
fossa, and this and the pituitary fossa are not only continuous
with each other in these early stages of development, but are
considered to continue so to be even in the adult. As the term
fossa is probably here employed strictly in the sense of fenestra,
these two so-called fossae will hereafter be referred to as the
fenestrae interparachordalis and hypophyseos.

In the third and fourth stages considered by Swinnerton
(embryos 6.6 to 25-mm. in length) it is said that the intra-
cranial notochord has undergone no further change beyond
a slight increase in absolute length, and further, that it under-
goes no actual suppression or reduction even in later stages of
development. It is also said that: ‘‘The interparachordal
fossa has been carried some distance in front of the notochord;
and the parachordals themselves have united across the inter-
vening space and across the end of the notochord in such a way
that this projects below, but close against the basis cranii.”’
A transverse plate of parachordal cartilage is thus formed, and
a median sagittal section through it in a 14-mm. specimen is
shown in one of the figures given (I. c., fig. 38, pl. 30.). The
parasphenoid is shown lying slightly below the parachordal
plate, and the hind end of the musculus rectus externus is in-
serted on the dorsal surface of the parasphenoid beneath the
anterior edge of the plate. Somewhat anterior to this point,
a process of bone is shown projecting dorso-anteriorly from
the dorsal surface of the parasphenoid, and it is called the me-
dian process of that bone. In the space between this process
and the anterior edge of the plate of parachordal cartilage, a
section of the basal portion of the brain is shown, and, although
not index lettered, it must represent the pituitary body and in-
fundibulum of the earlier stages, the infundibulum here slightly
differentiated as the saccus vasculosus. In a sagittal section
through this same region of the adult (J. c., fig. 37), the rectus

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 247

externus is shown lying between the parasphenoid and the plate
of parachordal cartilage, and extending posteriorly beyond the
prootic region into the anterior end of the basioccipital region.
The plate of parachordal cartilage now forms a prootic bridge,
but how it has been developed is not explained. In the base
of the median process of the parasphenoid is a block of cartilage
said to represent the anterior end of the parachordal, the median
process of the paraphenoid thus lying posterior to the fenestra
hypophyseos. The pituitary body (hypophysis) and infundib-
ulum (saccus vasculosus, both shown lying posterior to the
median process of the parasphenoid, must then also lie poste-
rior to the fenestra hypophyseos, and this is apparently also
their position in advanced embryos and the adults of certain
other, if not all fishes, as will appear later.

Swinnerton does not describe the internal carotid arteries,
but it seems certain, both from his figures and from the condi-
tions in a 40-mm. specimen of this fish, described immediately
below, that these arteries pass upward between the hind ends
of the trabeculae and that they are never there enclosed in
cartilage.

In the earlier stages considered by Swinnerton the rectus
externus muscles are said to be inserted into each other and
into the tissues filling the hind part of the fenestra interpara-
chordalis. In the third and fourth stages they extend poste-
riorly so that their hind ends lie beneath the posterior border
of that fenestra, and hence along the ventral surface of the basis
cranii. The eyeballs have in the mean time descended to a
level relatively lower than in the earlier stages, and this is said
to cause the eye-muscles to press upon the anterior prolonga-
tions of the parachordals and, depressing them, institute the
beginning of the formation of the myodome. It would nat-
urally be supposed that this depression would affect the hind
ends of the trabeculae, with which the parachordals are fused,
and this is what actually takes place in Salmo, as described by
Gaupp and to be later considered. In Gasterosteus, on the con-
trary, the hind ends of the trabeculae have not been in the least
depressed in the oldest stages shown by Swinnerton in which

248 EDWARD PHELPS ALLIS, JR.

they still persist, an embryo belonging to his third stage and
said to be 6.6 mm. in length (l. ¢., fig. 8). The lateral edges
of the anterior prolongations of the parachordals are also not
affected, as shown in that figure, their mesial edges alone being
depressed. This depression of the edges must then represent
a ventral growth of the cartilage, for it is difficult to compre-
hend how it could have been the result of any pressure of the
rectus muscles. |

In Swinnerton’s fourth stage, those parts of the trabeculae
which border on the fenestra hypophyseos are said to have been
suppressed, and it is said (/. ¢., p. 518) that:

In the hinder or parachordal portion, the interparachordal fossa
has been carried so far away in front of the notochord that the plate
formed by the median union of the parachordals now furnishes a con-
siderable portion of the basis cranii. Those parts lying immediately
on either-side of the fossa have now begun to undergo a movement of
depression, by which they have already come to he slightly below the
level of the basis cranu.

Swinnerton says that this movement of depression is perhaps
associated with a similar movement on the part of the rectus»
muscles, but, as just above stated, this ventral growth of the
parachordal cartilage begins in earlier stages, and it seems im-
probable that it can there be due to any action of these muscles.

In later stages of development, it is said (l. c., p. 527) that:

The process of depression of those parts bounding the interpara-
chordal fossa laterally has continued, so that this region now appears
to be a mere downward process of the prootic, with its cartilaginous
extremity mortised into the sides of the parasphenoid. This appear-
ance is enhanced by the fact that posteriorly each process is continued
into a ridge running along the under surface of the hinder portion of
the prootic. These two ridges are continuous with those already de-
scribed under the basioccipital, and there is a channel thus formed which
runs a considerable length of the basis cranii, is closed ventrally by the
parasphenoid, and opens anteriorly into the cavum cranii by means
of the interparachordal fossa.

It is then further said (I. ¢., p. 528) that:

In the larval Amia this canal is not present, but there is a well-
marked interparachordal fossa to which the eye muscles bear the same

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 249

relations as in the stickleback. It is probable, therefore, that in this
fish also a process of depression and secondary growth goes on on
either side of the fossa and below the prootic; but that, whereas in the
other types the fossa persists and transmits the eye muscles back again
out of the cranial cavity beneath the basis cranii, in Amia it disappears,
owing to continuous cartilaginous growth. As far back as the so-called
prootic bridge these muscles may be said to run in an actual derivative
of the cranial cavity; behind that they run in an extracranial space
secondarily enclosed.

The myodome of Gasterosteus is thus conceived by Swinner-
ton to be a space, the anterior portion of which is bounded
laterally by the bent-down anterior prolongations of the para-
chordals, and the posterior portion by secondary ventral down-
growths of the parachordals posterior to those anterior pro-
longations. These two portions of the myodome are thus of
totally different origin, and the anterior portion is considered,
because of its relations to the parachordals, to be an actual
derivative of the cavum cerebrale cranii. The figures given
show that it lodges the pituitary body, and that it is prechordal
in position, but, as the arteries and veins of the region are not
shown or particularly described, it is impossible to compare
the conditions here with those in the fishes that I have con-
sidered above. I have accordingly examined this region in a
series of transverse sections of a 40-mm. specimen of this fish,
and as the conditions there present certain new features, they
will be quite fully described.

In this 40-mm. specimen of Gasterosteus I find the recti
superiores, inferiores, and interni all arising from a thick median
vertical membrane which descends from the ventral surface of
the anterior portion of an unusually large membranous pitu-
itary sac, the recti interni having their origins posteroventral
to the recti superiores and inferiores. Posterior to the points
of origin of these muscles, the large hypophysis projects ven-
trally into the membranous pituitary sac, which lies dorsal to,
and in large part posterior to, the dorsoanterior edge of that
transverse ridge on the dorsal surface of the parasphenoid which
Swinnerton calls its median process. This process begins near
the hind edges of the ascending processes of the parasphenoid,

250 EDWARD PHELPS ALLIS, JR.

and, projecting dorso-anteriorly, extends approximately to the
transverse plane of the anterior edge of the hypophysis, where
it reaches to about the middle of the height of the myodome.
The large membranous pituitary sac rests upon its dorsal sur-
face, that surface being presented dorsoposteriorly. The space
beneath this ridge opens anteriorly into the subpituitary por-
tion of the myodome.

The ascending processes of the parasphenoid have their
greatest dorsal extent anterior to the transverse ridge on its
dorsal surface, and Swinnerton says that these processes of
Gasterosteus are not the homologues of the processes of the
bone of Amia. Swinnerton based this conclusion wholly upon
the fact that each process of the bone of Gasterosteus lies ante-
rior to the foramen for the nervus trigeminus, while in Amia
it lies posterior to it; but he overlooked the fact that a part of
the process of Gasterosteus, as shown in his figure 19, plate 29,
projects dorsally posterior to the foramen trigeminum, this
part of the process thus corresponding to the process of Amia.
The anterior portion of the process of Gasterosteus lies lateral
to the oculomotorius, trochlearis, and profundus nerves, and
also lateral to the vena jugularis and the rectus muscles, be-
tween them and the nervus trigeminus, thus having exactly
the relations to these several structures as does the pedicel of
the alisphenoid bone of Amia. This part of the process of Gaster-
osteus thus replaces functionally a pedicel of the alisphenoid,
and it has certainly been developed in relation to tissues that
represent, in this fish, that bone of Amia. The orbital open- .
ing of the myodome of Gasterosteus thus differs from the open-
ing in all the other fishes so far considered, except Cottus and
Clinocottus, in which latter fishes the pedicel of the alisphenoid
of Amia is also represented by a process of the parasphenoid. The
orbital opening of the myodome of Gasterosteus, and also that
of Cottus and Clinoccottus, does not, however, correspond
strictly to that of the myodome of Amia, for, as will be shown
later, there has been added to its ventral portion the canals
traversed, in Amia, by the internal carotid arteries and the
palatine branches of the faciales.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 251

In my 40-mm. specimen of Gasterosteus the internal caro-
tid and efferent pseudobranchial arteries of either side perfor-
ate the base of the ascending process of the parasphenoid
through a single foramen, and enter the space beneath the trans-
verse ridge on the parasphenoid. There the pseudobranchial
artery is connected by a cross-commissure with its fellow of the
opposite side, and then runs forward into the orbit as the ar-
teria ophthalmica magna. The internal carotid gives off, before
entering its foramen, its orbitonasal branch, which traverses
the foramen with the internal carotid and efferent pseudo-
branchial arteries, and then runs forward along the floor of the
myodome to enter the orbit. The internal carotid, after giving
off this branch and having entered the space beneath the trans-
verse ridge on the parasphenoid, turns upward in the median
vertical myodomic membrane, and, while in that membrane,
anastomoses with its fellow of the opposite side. It then sepa-
rates from its fellow and, while still in the membrane, divides
into two parts, one of which at once enters the cavum cerebrale
cranli, and is the posterior cerebral artery. The other part
runs forward in the thick median portion of the membranous
floor of the cavum cerebrale cranii, and, issuing beneath it, sends
two branches to the eyeball, one of them accompaying the nervus
opticus. The remainder of the artery then enters the cavum
cerebrale crani through the foramen opticum, and is the ante-
rior cerebral artery. No positive and definite connection between
the anterior and posterior cerebral arteries was seen, the ante-
rior branch of the latter artery, found in the other fishes, not
occurring here.

The ramus palatinus facialis arises from the trigemino-facialis
ganglionic complex, and passing lateral and then ventral to the
vena jugularis, runs ventromesially along the internal surface
of the prootic bone and perforates the dorso-anterior portion
of the transverse ridge on the parasphenoid to enter the space
beneath it and then to escape into the orbit. This nerve, in this
fish, thus lies lateral to the vena jugularis, while in all others
in which it was traced (Hyodon, Scomber, Scorpaena, Cottus,
Catostomus, and Amia) it lies mesial to that vein. This is, how-

252 EDWARD PHELPS ALLIS, JR.

ever, unquestionably related to the fact that the vena jugularis
lies ventral (mesial) to the nervus facialis in Gasterosteus, while
in the other fishes mentioned above, except Catostomus, it lies
dorsal (lateral) to that nerve. In Catostomus the vein lies. ven-
tral (mesial) to the nervus facialis but lateral to the nervus pala-
tinus, this thus being a variation in the transformation of the
primitive vena cardinalis anterior into a vena capitis lateralis.

A delicate median vertical membrane separates the space
beneath the transverse ridge on the dorsal surface of the para-
sphenoid into lateral halves, this membrane being continuous
anteriorly with the membrane that gives insertion to the rectus
muscles. At its hind end this membrane ossifies as a short
median ridge on the dorsal surface of the parasphenoid.

The space beneath the dorso-anteriorly projecting ridge on
the dorsal surface of the parasphenoid thus corresponds strictly
to the ventral compartment of the myodome of the other Tele-
ostei so far considered, but the roof of that compartment, which
is of membrane in those other fishes, has here been ossified as
part of the parasphenoid. The recti externi and the pituitary
veins run posteriorly dorsal to this ridge, and hence lie in the dor-
sal compartment of the myodome, the pituitary veins lying along
the lateral surfaces of the pituitary sac, and forming, posterior
to it, a large median sinus.

Posterior to the hind end of the ventral compartment of the
myodome, and hence posterior also to the ascending processes
of the parasphenoid, a tall median ridge of the latter bone, flat
on its dorsal surface, projects upward between the ventral ends
of the ventral processes of the prootics, its dorsal surface there
forming the floor of the dorsal compartment of the myodome.
Up to this point the ventral processes of the prootics are wholly
of bone, but cartilage now appears in their ventral halves, as
shown in Swinnerton’s figure 35, plate 30. The hind end of
the membranous pituitary sac is here cut in the sections. Pro-
ceeding posteriorly, the median ridge on the dorsal surface of
the parasphenoid gradually becomes wider, and, arching up-
ward, projects into the myodome between the ventral ends of
the ventral processes of the prootics. ‘The membrane which,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 253

anterior to this point, formed the roof of the myodome, is now
replaced by membrane bone which forms the anterior portion
of the prootic bridge, and the nervus abducens perforates it,
on either side, to enter the dorsal compartment of the myo-
dome. Farther posterior in the sections, a median plate of
cartilage appears in the prootic bridge, enclosed between dor-
sal and ventral plates of perichondrial bone, and in this trans-
verse plane the cartilage in the ventral ends of the ventral proc-
esses of the prootics disappears and is replaced by membrane
only. The ventral processes of the prootics are accordingly
now formed by short processes of bone, partly of membrane
and partly of perichondrial origin, that are prolonged ventrally
by membranes, continuous ventrally with the lateral edges
of the parasphenoid. Proceeding posteriorly, the median plate
of cartilage expands laterally, on either side, and becomes the
cartilaginous basis cranii, here still enclosed, on either side,
between plates of perichondrial bone which form parts of the
prootics. The myodome still continues onward, in a median
groove on the ventral surface of this cartilage, there lodging
the recti externi and being bounded laterally in part by mem-
brane only and ventrally by the parasphenoid.

At the extreme hind end of the myodome a circle of bone appears
in the sections, this bone forming part of the basioccipital and
lying in the groove on the ventral surface of the cartilaginous basis
crani. From this shell of bone a median plate is sent down-
ward between the diverging hind ends of the parasphenoid, and
the shell of bone then fuses with perichondrial bone developed
in relation to the overlying cartilage and forming part of the
basioccipital. The conditions are thus here practically as de-
seribed and figured by Swinnerton in his 16-mm. specimen of
this fish (1. c., fig. 36, pl. 30).

The myodome of Gasterosteus is thus strictly comparable to
that in the other Teleostei described, except that it has a greater
anterior extension than it has in any of them, Cottus and Clin-
ocottus excepted, this being due to the ossification, as part of
each ascending process of the parasphenoid, of tissues representing
the pedicel of the alisphenoid. The parasphenoid has under-

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

254 EDWARD PHELPS ALLIS, JR.

gone special development in this fish, and the conditions here
show, even more positively than in the others considered, that
part of this bone may be developed in definite relations to the
membrane separating the myodome into dorsal and ventral
compartments. It is accordingly certain that this bone is here
developed, in part, in relation to axial skeletogenous material,
and is not a simple dermal bone primarily developed in relation
to the mucous lining membrane of the pharynx, and which .
sank gradually inward to its actual position.

DACTYLOPTERUS VOLITANS

The conditions in Gasterosteus, as above explained, seeming
to offer an explanation of the somewhat exceptional conditions
that I described in Dactylopterus in my work on the mail-cheeked
fishes, I have reexamined my material of that fish. In that
earlier work I described a transverse ridge on the dorsal sur-
face of the parasphenoid that was said to project dorsopos-
teriorly and to form the posterior wall of the myodome. Be-
cause of the position of this wall, I concluded that the post-
pituitary portions of the horizontal processes of the prootics had
been depressed and appressed upon the underlying ventral
flanges of those bones, and that the latter flanges had under-
gone marked reduction. I now find that the ventral flanges of
the prootics have not undergone any particular reduction, and
that there has been no depression and appression of the hori-
zontal processes of the prootics, which are represented by a
well-defined membrane forming the floor of the cavum cere-
brale cranii.. The anterior end of this membrane passes over
the dorsal edge of the transverse ridge on the parasphenoid,
closely adherent to it, and is then continuous with the mem-
brane that I described as closing the pituitary opening of the
brain case. Beneath the part of this membrane that repre-
sents the horizontal processes of the prootics, and between it
and the parasphenoid, is a space which must represent some
part of the dorsal compartment of the myodome, this space
being shut off from the subpituitary portion of the myodome
by the transverse ridge on the parasphenoid. The hypophy-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER Zao

sis lies anterior to this ridge, the cross-commissure of the pitui-
tary veins lying ventral to the hypophysis and separated from
it by the dura mater. The recti externi have their insertions
on a median vertical membrane, immediately posterior to the
pituitary veins and immediately anterior to the summit of the
transverse ridge. They are surrounded by connective tissue
that resembles the fatty tissue found abundantly in this fish,
but are not otherwise separated from the other rectus muscles,
the dorsal and ventral myodomic compartments thus appa-
rently here being confluent.

SALMONIDAE

In Salmo, Parker (’73) and Stohr (’82) found the trabeculae
and parachordals primarily independent of each other. Stohr
also found the anterior portions of the parachordals—the parts
corresponding to the anterior prolongations of the parachordals
of Swinnerton’s descriptions of Gasterosteus—primarily inde-
pendent of the posterior portions, and he considered them to
represent the ‘Balkenplatten’ of the Amphibia. They are
said by him to fuse, first, with the posterior portions of the para-
chordals and then with the trabeculae. Of the adult Salmo,
Parker’ says (i"c., ‘p. 102):

One remarkable change in the investing mass, as a whole, is the
growth downward of a lamella on each side, thus forming a covered
archway; for in front of the retiring notochord the moieties of cartilage
meet, and this viaduct is floored by the submucous bone which has
been removed, the parasphenoid. All the true axial parts of the skull
cease at the front edge of the investing mass behind the pituitary space;
all the rest has a facial foundation, is built on the trabeculae, or has a
secondary character as a development of the cranial wall.

This so-called covered archway is the myodome, which is thus
considered by Parker to be bounded laterally by downgrowths
of the parachordal cartilage and not by those cartilages, them-
selves, bent down.

Schleip says (’04, pp. 355 to 359) shied in trout embryos,
12 to 14-mm. in length, the parachordals ae trabeculae form
the floor of the primordial cranium, and that the cartilages of

256 EDWARD PHELPS ALLIS, JR.

opposite sides are separated from each other by a fissure (Fis-
sur) the posterior, interparachordal portion of which is en-
tirely filled by the projecting anterior end of the notochord.
There accordingly is, as described by Schleip, no interparachordal
fenestra in these embryos. The intertrabecular portion of the
fissure is called the pituitary fossa and it is said to be closed
ventrally by the parasphenoid, which, at these stages, extends
from the orbit only to the tip of the notochord (Il. c., p. 354).
The rectus muscles, in running from their insertions on the
bulbus to their points of origin, are said to lie, in their ante-
rior portions, either above the trabeculae or above the fissure
(so-called pituitary fossa), and farther posteriorly to lie in the
fissure itself; the recti externi extending still farther posteri-
orly so that their hind ends lie under the notochord and hence
beneath the basis cranii. That part of the space above the
trabeculae, or above the fissure, thus occupied by the rectus mus-
cles is said to form a part of the cranial cavity, but to be closed
toward the brain by a membrane which, in these embryos, ex-
tends posteriorly to the tip of the notochord. Anteriorly, the
edges of this transverse membrane are said to be attached to
the side walls of the cranium, above its floor, the membrane
thus there separating the cranial cavity into dorsal and ventral
parts. Posteriorly, the membrane is said to stretch from one
trabecula to the other, there closing the intertrabecular fissure,
(the so-called pituitary fossa) and taking part in the formation
of the basis crani. It is, however, further said that, in later
stages (embryos 18-mm. long), this same posterior portion of
the membrane chondrifies, and that the cartilage so formed
extends posteriorly to the tip of the notochord and there forms
both the roof of the eye-muscle canal (myodome) and the floor
of the cranial cavity. This cartilage is thus evidently the pro-
otic bridge, and as the bridge cannot possibly have been formed
by the chondrification of a membrane extending from one tra-
becula to the other, there is some error in the descriptions.

The myodome, as above described, is said by Schleip to pre-
sent three sections: an anterior one, intracranial in position,
but separated from the brain by the transverse membrane above

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 257

referred to; a middle section, ‘‘der in der Fissur der Schiidel-
basis, bezw. in einem nach unten offenen Sulcus liegt,’ and a
posterior section which les wholly beneath the basis eranii.
Reference is here made by Schleip to a series of half schematic
figures (l. c., pp. 355 to 357), and consideration of them shows
that the so-called anterior section of the myodome is what I
have called, in the fishes described by me, its prespinal section.
The middle section is apparently my prootic portion of the spinal
section, and the posterior section what I have called the basioccip-
ital portion of that section. A membrane is shown in these
figures extending transversely between the ventral ends of the
ventral processes of the prootic cartilages and separating the
recti externi from the recti interni. This membrane is the hori-
zontal myodomic membrane of my descriptions, but I cannot
find that Schleip refers to it in his text, for the transverse mem-
brane of his descriptions is said to form the roof of the myodome.

In a 25-mm. embryo of Salmo salar Gaupp finds conditions
strikingly similar to those described by Swinnerton in Gaster-
osteus, and he arrives at practically similar conclusions regard-
ing the development of the myodome, without, however, here
making special reference either to that author’s or to Schleip’s
conclusions regarding it. Like Schleip, Gaupp (05 b, pp. 665
to 669) separates the myodome into anterior, middle, and pos-
terior sections. ‘The anterior section is said to lie in the poste-
rior portion of the orbitotemporal region, and its floor to be
formed by the two trabeculae and a membrane which extends
transversely between them. ‘The space between the two tra-
beculae is called by Gaupp the fenestra basicranialis anterior,
or fenestra hypophyseos, and it corresponds to the intertra-
becular, or pituitary fossa of Swinnerton’s descriptions of Gas-
terosteus. It, however, apparently corresponds to the anterior
portion only of the intertrabecular, or pituitary fossa of Schleip’s
description of the trout, the hind ends of the trabeculae of
Schleip’s account corresponding to the anterior prolongations of
the parachordals of Swinnerton and Gaupp. The membrane
said by Gaupp to close the fenestra hypophyseos of Salmo is
shown by him, in a figure of a cross-section through this region

258 EDWARD PHELPS ALLIS, JR.

(1. c., fig. 8345, p. 669), lying ventral to the musculi recti interni,
and it is furthermore said that it gives attachment on its dorsal
surface to the metachiasmatic (posterior) portion of the inter-
orbital septum. The posterior portion of the interorbital sep-
tum here referred to is evidently the median vertical myodomic
membrane of my account, the membrane that closes the fenes-
tra hypophyseos then being represented in the layer of skeletog-
enous tissue forming the floor of what I have called the sub-
pituitary portion of the ventral compartment of the myodome.

The roof of the anterior section of the myodome, as thus de-
cribed by Gaupp, is said by him to be formed by the mem-
branous floor of the cranial cavity, and its side walls by the
ventral portions of the cartilaginous side walls of the cranium,
which are said to here extend between the otic capsules and the
trabeculae. The eye muscles are said to have forced the brain
upward from the basis cranii, the hypophysis being carried
with it and so lifted out of the fenestra hypophyseos.

The middle section of the myodome is said by Gaupp to lie,
in part, in the labyrinth region and, in part, in the extreme pos-
terior (hintersten) portion of the orbitotemporal region. Its
floor is said to be formed by the anterior prolongations of the
parachordals (vordere Parachordalia) which have been forced
ventrally by the pressure of the musculi recti externi, exactly
as Swinnerton had previously said was the case in Gasterosteus.
Gaupp, however, shows the hind ends of the trabeculae—the
parts bounding the fenestra hypophyseos—forced ventrally
to the same extent as the parachordals. Because of its rela-
tions to the anterior parachordals, this middle section of the
myodome is said to lie between the primordial basis cranii and
the brain, and hence to form a part of the primordial cranial
cavity. Its side walls are described as formed, on either side,
by two basicapsular commissures, which extend from the otic
capsule of their side to the anterior prolongation of the related
parachordal, and lie, one between the nervi trigeminus and
facialis, and the other between the latter nerve and the otic
capsule. Its roof is formed by the membranous floor of the
cavum cerebrale cranii, this membrane arising, on either side,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 259

from the side wall of the cranium, extending posteriorly dorsal
to the notochord, and anteriorly passing into the supraseptal
membranous floor of the cavum cerebrale cranii in the orbito-
temporal region. In this membranous roof of this middle sec-
tion of the myodome a transverse bridge of cartilage, the prootic
bridge, is later developed, and it is said to lie above and anterior
to the tip of the notochord. In the adult this bridge extends
forward to the hind edge of the hypophysis, as shown in Parker’s
figure of a bisected skull (’73, fig. 4, pl. 7), and the pituitary
opening of the brain case lies considerably posterior to the an-
terior edges of the ventral processes of the prootics. It must
then be that, as in the adult Gasterosteus, the hypophysis of
the adult Salmo lies dorsal to the interparachordal fenestra and
not dorsal to the fenestra hypophyseos.

The middle section of Gaupp’s descriptions of the myodome
thus apparently corresponds to the subpituitary portion of the
prespinal section, and to all of the prootic portion of the spinal
section, of my descriptions. Gaupp says that, primarily, the
nervus palatinus facialis issues from the cranial cavity along
the lateral edge of the anterior parachordal, but that, as the
myodome gains in height and breadth, the nerve becomes in-
cluded in it, then entering it by perforating its membranous
roof and leaving it through a foramen in its floor. The course
of the internal carotid arteries is not given, but as there are no
special perforations of the basis cranii for them, they must pass
upward through the fenestra hypophyseos. Gaupp_ shows,
in his figure of the entire chodrocranium, a foramen lying be-
tween the foramen for the nervus facialis and the incisura pro-
otica, and it is said to give passage to the vena jugularis, coming
from the anterior portion of the cranial cavity. This vein is,
however, certainly not the jugularis of current descriptions of
fishes, and is probably the encephalic vein of Allen’s (’05) de-
scription of the Loricati. It cannot be the pituitary vein,
for that vein does not extend into the anterior portion of the
cranial cavity.

The posterior section of the myodome is said to lie beneath
the basis cranii, between it and the parasphenoid, and to com-

260 EDWARD PHELPS ALLIS, JR.

municate with the middle section through the fenestra basi-
cranialis posterior, which lies between the anterior ends of the
parachordals and apparently corresponds to the posterior por-
tion of the intertrabecular fissure of Scheip’s descriptions. The
recti externi pass through this fenestra, and, beyond it, lie be-
neath the basis cranii. There, as they increase in size, they
are said to push both the middle portion of the basis cranii
upward and the parasphenoid downward. The basal plate,
formed by the parachordals and the enclosed notochord, then
thickens along each lateral surface of these muscles, and so forms
the lateral walls of this section of the myodome.

Comparing the conditions in this fish with those in Gaster-
osteus, it is seen that, in both fishes, the anterior portions of
the parachordals lie, when first formed, at a certain distance
lateral to the anterior end of the notochord, which projects anteri-
orly between them. In later stages of both fishes these project-
ing portions of the parachordals are said, by both Gaupp and
Swinnerton, to be depressed, but the figures given by both show
that this depression effects only the mesial edges of the para-
chordals, their lateral portions retaining their primitive posi-
tions in the level of the notochord. Between these higher lying
portions of the cartilages, the prootic bridge is later developed.
How this bridge is developed in Gasterosteus is not stated by
Swinnerton. In Salmo Gaupp says it is formed by the chon-
drification of a part of the membrane forming the roof of the
middle section of the myodome and also the floor of the cavum
cerebrale cranii. Gaupp says it lies, when first formed, an-
terior to and above the tip of the notochord, and it is shown,
in one of his figures, separated from that tip by the anterior
portion of an open space that is prolonged posteriorly on either
side of the anterior end of the notochord. The posterior portion
of this space may possibly form part of the fenestra basicranialis
posterior of Gaupp, but its anterior portion certainly does not, for
Gaupp says that this fenestra les between the edges of the bent-
down parachordals and gives passage to the recti externi from
the middle to the posterior section of the myodome. There
are, then, four distinctly different fenestrae in this myodomic

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 261

region. One of them is the fenestra hypophyseos of Gaupp,
which lies between the hind ends of the trabeculae and hence
in the floor of the anterior section of the myodome; and this
must be, in early embryos, traversed by the internal carotid
arteries, for Parker (’73) shows those arteries, in this fish, run-
ning upward anterior to the hypophysis, and I so find them in
all the Teleostei I have examined. A second fenestra is the
fenestra basicranialis posterior of Gaupp, which les partly in
the floor of the middle section of the myodome and partly be-
tween that section and the posterior section. <A third fenestra,
not described by Gaupp, lies in the floor of the posterior section
of the myodome, and this, together with that part of the second
fenestra that lies .n the floor of the middle section of the myo-
dome, forms the fenestra ventralis myodomus, the so-called hypo-
physial fenestra of Sagemehl. The remainder of the second fenes-
tra—the part leading from the middle section of the myodome
into the posterior one—is simply a transverse section of the con-
tinuous cavity of the myodome and does not open on to the
ventral surface of the cranium. The fourth fenestra lies in
the roof of the middle section of the myodome, and this alone
is the homologue of the fenestra basicranialis posterior of the
Sauropsida. This is evident from Sonies’s (’07) description
of this fenestra in the chick and duck, to be discussed later, and
from Gaupp’s (00) account of it in Lacerta. In Lacerta the
fenestra is said by Gaupp to be bounded anteriorly by the crista
sellaris, and to be closed by a membrane (Gewebe) everywhere
continuous with the perichondrium of the bounding cartilages,
and that represents an unchondrified portion of the primordial
cranium. The anterior end of the notochord is enclosed in this
membrane, and lies, in part of its course, so close to its ventral
surface that it forms a longitudinal ridge along it The fenes-
tra accordingly les in what corresponds to the roof of the myo-
dome of fishes, and not to its floor, and hence cannot be the
homologue of the similarly named fenestra of Gaupp’s descrip-
tions of Salmo. In the Urodela, also, the fenestra basicranialis
posterior is said by Gaupp (’05b, p. 692) to be a perforation
of the basal plate, traversed by the notochord, and lies pos-
terior to its tip, as it does in Lacerta.

262 EDWARD PHELFS ALLIS, JR.

AMIURUS

In the adult Amiurus the myodome was briefly considered
by me in my work on the mail-cheeked fishes, and I there said
(Allis, 709, p. 200) that;

In the anterior three-fifths, approximately, of its length, the ven-
tral edge of the prootic does not meet its fellow of the opposite side,
a wide hypophysial fenestra, closed ventrally by the parasphenoid,
being left between the two bones. Posterior to this fenestra, the ventral
edges of the prootics meet in the middle line, and the two bones there
form, on the floor of the cranial cavity, a prominent transverse bol-
ster which has closely the position of the cross-canal of Lepidosteus;
and it is certainly in this bolster that MecMurrich found the small
cavity that he considered to be a rudimentary myodome.

In the specimens that I examined at that time I found but slight
indication of this cavity, but I nevertheless considered it to have
existed previously in the transverse bolster and to have been sup-
pressed by invading growth of the surrounding cartilage.

In my work on the pseudobranchial and carotid arteries of
this fish, I said (Allis, 08 b, p. 259) that the external carotid
artery

does not apparently traverse a trigemino-facialis chamber, for al-
though it would seem as if that chamber must be present in some form,
there is no proper indication of but one cranial wall in this region, and
that one wall would seem to be the inner wall of the chamber; for both
the external carotid and the jugular vein lie external to it.

It was further said (p. 259) that:

The parasphenoid of Ameiurus is peculiar in that the base of the
ascending process of the bone, which begins immediately posterior to
the so-called orbitosphenoid, is formed of two plates which enclose
within them the hind end of the Ssubopticus (trabecular?) bar of car-
tilage. The bone is here apparently not of perichondrial origin, but
the inner plate nevertheless lies internal to the cartilage of the skull
and there forms part of the immediate bounding wall of the cranial
cavity. Posterior to the hind end of the trabecular (?) cartilage there
is, for a few sections, a vacant space between the two plates of the proc-
ess of the parasphenoid, and then those plates, the inner one of which
gradually diminishes in height, enclose the anterior portion of the
prootic (parachordal?) cartilage. It is perhaps this portion of the
bone of the adult that led Me Murrich to conclude that the basisphenoid
was here anchylosed with the parasphenoid.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 263

I have now reexamined my sections of young specimens of
this fish, but the material was evidently not in a good state of
preservation when sectioned, for the membranes in the myo-
domic region are all more or less disintegrated. The cartilages
I tentatively identified in my earlier work as trabecular and para-
chordal are certainly those cartilages, as currently conceived,
for the one lies wholly anterior to the hypophysis and the other
along the lateral wall of, and posterior to, that organ. The
anterior end of the parachordal cartilage is, as I stated, enclosed
between external and internal plates of the parasphenoid, but
neither plate is adherent to it, and, in the adult, the cartilage les
in a little pocket on the dorsal surface of the parasphenoid and
ean be easily withdrawn from it without breakage. Posterior
to this pocket, the cartilage, in embryos, gradually becomes
enclosed between plates of perichondrial bone which form part
of the prootic, the internal plate of the parasphenoid gradually
diminishing in height and finally vanishing. The ventral edges
of the prootics form the lateral boundaries of the fenestra ven-
tralis myodomus. The hypophysis is large, lies directly above
this fenestra, upon the dorsal surface of the parasphenoid, and
extends posteriorly to the anterior surface of the transverse
bolster described in my earlier work. This bolster is but slightly
developed in my young specimens, but it is evidently formed
either by the fusion of the horizontal and ventral processes of
the prootics or by the horizontal processes alone, the ventral
processes of the prootics, in the latter case, here vanishing. The
cavity described by MeMurrich (’84) in this bolster would then
seem to represent the prootic portion of the dorsal compart-
ment of a myodomic cavity. A ventral myodomic compart-
ment is wholly wanting, for that part of the parasphenoid lying
in the prootic region has certainly been developed in the skeletog-
enous tissue which, in the other Teleostei described, forms the
horizontal myodomic membrane, this part of the parasphenoid
of Amiurus thus corresponding to the transverse ridge on the
dorsal surface of the bone of Gasterosteus. This, then, accounts
for the fact that both the internal carotid artery and the ramus
palatinus facialis of Amiurus lie everywhere external to the para-

264 EDWARD PHELPS ALLIS, JR.

sphenoid instead of passing internal to the ascending process
of that bone. Whether or not there is a subpituitary portion
of the dorsal myodomic cavity I cannot determine, the mem-
branes being in a more or less disintegrated condition. It is,
however, apparently wanting, for there are no veins compar-
ble to the pitutary veins of the other Teleostei considered, the
pituitary region being drained by veins definitely in the cavum
cerebrale cranii. Furthermore, the membranous pituitary sac
apparently forms the perichondrial lining of the pituitary fossa,
as it also does of the larger part of that fossa in the Selachii.
There is no pars jugularis of a trigemino-facialis chamber.

The internal carotid artery gives off, as described earlier
(Allis, ’08 b), an orbitonasal artery, sends two branches to
the eyeball, and then enters the cranial cavity through the fo-
ramen opticum, behind the nervus opticus. This latter nerve
certainly lies dorsal to the trabecula. The internal carotid
artery of this fish must then also have that relation to that car-
tilage, the artery accordingly entering the cranial cavity by
passing, first lateral and then dorsal to the trabecula.

AMIA CALVA

In Amia (Allis, 97, etc.) the myodome lodges the external
rectus muscles and the pituitary veins, and corresponds to
the dorsal compartment only of the myodome of Teleosts.
It has no basioccipital extension, being limited to the subpitui-
tary and prootic regions. The: hypophysis and saccus vascu-
losus, both enclosed in the membranous pituitary sac, project
ventrally into this myodomie cavity, the hypophysis lying im-
mediately posterior to the presphenoid bolster and the saccus
projecting posteriorly beneath the prootic bridge. The other
rectus muscles have their origins on the lateral surface of the
presphenoid bolster. The myodome has a large orbital open-
ing, bounded laterally by the pedicel of the alisphenoid, and
the nervus profundus and vena jugularis traverse this opening
to enter the trigemino-facialis chamber, the recti externi tra-
versing it to enter the myodome, and the oculomotor and troch-
learis nerves traversing it to reach their primary foramina,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 265

which lie in the membranous wall of the cavum cerebrale cranii.
All of these structures thus pass mesial to the pedicel of the
alisphenoid, as do also, morphologically, the pituitary vein
and the abducens nerve. The arteria carotis externa and the
nervi maxillaris and mandibularis trigemini, on the contrary,
pass lateral to this pedicel to enter the trigemino-facialis chamber.

The trigemino-facialis chamber is not separated by a wall
of bone into ganglionaris and jugularis parts, as in most of
the Teleostei, and, because of the absence of a bony floor, the
chamber is in direct communication with the myodomic cavity.

A ventral compartment of the myodome, as a functional myo-
domie cavity, is wholly wanting in this fish, but is represented
in certain canals traversed by the internal carotid and efferent
pseudobranchial arteries, the internal carotid artery of either
side being accompanied, in part of its course through its canal,
by the palatine branch of the facialis and the pharyngeal branch
of the glossopharyngeus. These several canals were fully
described in an earlier work (Allis, 97, p. 496) and were there
called the palatine, internal carotid, and efferent pseudo-
branchial canals. The palatine canal of either side, as there
described, les between the parasphenoid and the ventrolateral
surface of the chondrocranium, and the nervus palatinus facialis
enters it at a certain distance anterior to its‘ hind end, the pos-
terior portion of the canal lodging only the internal carotid
artery and the pharyngeal branch of the nervus glossopharyn-
geus The internal carotid canal arises from this palatine canal
and, running upward, traverses the cartilaginous presphenoid
bolster to enter the cavum cerebrale cranii. The efferent pseu-
dobranchial canal is in two sections, one of which traverses the
lateral bounding wall of the. orbital opening of the myodome,
while the other penetrates the presphenoid bolster to fall into
the internal carotid canal. My palatine canal is the canalis
parabasalis of Gaupp’s (’05 a) account of Lacerta, and conditions
in other vertebrates, to be later considered, show that it should
be considered as formed by the fusion of two canals, one tra-
versed by the nervus palatinus facialis and the other by the
nternal carotid artery.

266 EDWARD PHELPS ALLIS, JR

In the adult Amia the efferent pseudobranchial artery gives
off its ophthalmica magna branch as it traverses the orbital
opening of the myodome, the artery and this branch thus both
appearing to here lie dorsal to the trabecula. In 8-mm. and
10-mm. embryos I, however, find the artery passing ventral to
the trabecula and there falling into the internal carotid as it
turns upward to pass between the trabeculae. From the artery
so formed the arteria ophthalmica magna arises, and runs out-
ward, dorsal to the trabecula, thus lying, in Amia, on the op-
posite side of the trabecula to that in which it is shown by Dohrn
in a 10-mm. embryo of the trout (Dohrn,’86, fig. 2). The
development of these arteries and their relations to the trabee-
ulae need further investigation.

The prootic bridge, which forms the roof of the prootice por-
tion of the myodome of the adult Amia, is of relatively late
formation, for it is not found in a 40-mm. specimen. In a
43-mm specimen it has been formed, and, as in Salmo, lies
at a certain distance dorsal to the fenestra ventralis myo-
domus and separated from the tip of the notochord by
an open space, closed by membrane, which is the homologue
of the fenestra basicranialis posterior of the Sauropsida. The
saccus vasculosus lies, in this specimen, wholly anterior to the
anterior edge of the prootic bridge, directly in line with it and
embedded in the dorsal surface of loose stringy connective tis-
sue which fills this posterior portion of the myodome. The
recti externi, which, in the adult, extend to the hind end of the
myodome, do not, at this age, extend posteriorly even as far as
the hind end of the saccus vasculosus, having their origins ap-
proximately in the transverse plane of the posterior opening of
the trigemino-facialis chamber.

From these conditions in Amia, it is evident that the pre-
spinal and prootice portions of the normal teleostean myodome
would arise if the cartilage which, in Amia, separates the myo-
dome from the canals for the internal carotid and efferent pseu-
dobranchial arteries were to be resorbed, leaving more or less
developed membranes in its place. This cartilage is known to
undergo resorption during the ontogenetic development of cer-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 267

tain of the Teleostei (Salmo, Gasterosteus), and skeletogenous
tissues capable of taking a membranous form would certainly
be left in its place. The membranous tissues that would then
represent the presphenoid bolster would not offer a firm point
of attachment for the rectus muscles, and it would be wholly
natural for certain of them to seek more solid points of origin,
and one of them actually has, in most of the Teleostei, acquired
such an origin by first creeping downward on to the dorsal sur-
face of the parasphenoid and then pushing posteriorly in the
open end of the persisting remnant of the palatine canal of my
descriptions. This muscle actually is the rectus internus, but
it is possible that it was primarily the rectus inferior, that muscle
and the rectus internus undergoing an exchange of function
and so giving rise to that manner of innervation of these muscles
that I have described in several of these fishes (Allis, ’03, ’09),
and which I now find to be apparently definitely related to
the presence of a functional ventral myodomic compartment.
Where that compartment is wanting, as in Amiurus, or present
but non-functional, as in Lepidosteus, Polypterus, Polyodon,
Acipenser, and higher vertebrates, these muscles are innervated
approximately as they are in Amia (Allis, 708 b).

The definitive rectus internus of the Teleostei, in thus shift-
ing its point of origin, passed dorsal to the efferent pseudobran-
chial artery and dorsolateral to the internal carotid. The mem-
branous tissues representing the presphenoid bolster were then
pressed together in the median line by these muscles, and be-
came the median vertical myodomic membrane, the internal
carotid arteries still being enclosed in it, in a membranous canal,
the homologue of the cartilaginous canals of Amia fused to
form a single canal. The floor of the myodome of Amia be-
came the horizontal myodomic membrane, which becomes
adherent to the ventral! surface of the membranous pituitary
sac and seems to end there. It, however, certainly continued,
primarily, beyond that point and was continuous with the ven-
tral end of the interorbital septum. The efferent pseudobran-
chial arteries were necessarily pressed ventrally by the recti in-
terni, and, losing their connections with the internal carotids,

268 EDWARD PHELPS ALLIS, JR. .

acquired a cross-commissural connection with each other. A
basioccipital portion of the dorsal compartment of the tele-
ostean myodome would then be added to the prootic portion,
developed as above set forth, whenever an aortal groove similar
to that in Hyodon had been developed and retained; and a pos-
terior extension of the ventral compartment would be acquired
by the recti interni pushing posteriorly between the floor of
that compartment and the underlying parasphenoid. The
many variations that I have described above in the myo-
dome of the non-siluroid Teleostei would then all arise by dif-
ferent degrees of ossification of the several membranes in this
region.

As already stated, the internal carotid arteries of Amia tra-
verse the presphenoid bolster in order to enter the cavum cere-
brale cranil, and, although the development of this bolster has
not yet been worked out, there seems no question that it is
formed by the hind ends of the trabeculae The median ver-
tical myodomic membrane of the Teledstei, which in those
fishes represents the presphenoid bolster of Amia, would then
also represent the hind ends of the trabeculae. The basisphenoid
of the Teleostei cannot then be the exact homologue of the pre-
sphenoid bolster of Amia. The fenestra ventralis myodomus
of the adult Amia les posterior to the presphenoid bolster;
it must then be bounded laterally by the so-called anterior
prolongations of the parachordals, and hence correspond to
Gaupp’s fenestra basicranialis posterior in embryos of Salmo, the
fenestra hypophyseos of these embryos apparently being repre-
sented in the internal carotid canals of Amia. The fenestra
hypophyseos is said by Gaupp (’05 b, p. 585) to be a persisting
portion of the large fenestra basicranialis anterior of early em-
bryos, and it is said by him to be always traversed by the ecto-
dermal stalk of the hypophysis. The hypophysis must then
lie, in these early embryos, dorsal to this fenestra, and as the
internal carotid arteries, in the Holostei and Teleostei, run
upward anterior to the hypophysis, they must traverse the
fenestra. The hypophysis must then have later shifted pos-
teriorly to a position dorsal to the fenestra interparachordalis

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 269

(fenestra basicranialis posterior, Gaupp), leaving the carotid
arteries behind it, in persisting remnants of the fenestra hypoph-
yseos which I have called, in Amia, the internal carotid canals.

The carotid arteries do not, in either Amia or the Teleostei,
enter any part of the dorsal myodomic cavity. In certain other
fishes and in higher vertebrates they become included in that
cavity. The arteries must accordingly there have either shifted
posteriorly, with the hypophysis, out of the fenestra intertra-
becularis into the fenestra interparachordalis, or the inner walls
of the canals traversed by them in Amia, both the carotid canals
through the presphenoid bolster and those parts of the para-
basal canals which lodge those arteries, must have been re-
sorbed, the canals thus being added to the dorsal myodomic
cavity The arteries would then lie dorsal to the cartilaginous
floor of the myodomic cavity, instead of, as in Amia, ventral
to it; their foramina would lie near the hind end of the subpitui-
tary portion of the pituitary fossa, instead of anterior to it;
and a part of the ventral compartment of the teleostean myo-
dome would be added to the definitive myodomie cavity.

The septum interorbitale may now be considered, for it forms
a direct anterior prolongation of the median vertical myodomiec
membrane and hence must be of similar origin. This septum
is said by Gaupp (05 b, p. 585) to characterize the tropibasie
cranium, and to be found in many of the Selachii (Plagiostomi ?),
in the Ganoidei, the Teleostei, excepting the Siluridae and
Homaloptera, and the Amniota. The platybasic cranium, in
which this septum is wanting, is said to be found in many
of the Selachii and in all of the Amphibia. In the Teleostei
the septum is said to lie above the trabeculae (Gaupp, ’05 b,
pp. 667 and 762), between them and the cavum cerebrale cranii.
The septum must then be formed by the ventral portions of the
side walls of the primordial cranium pressed together in the
median line, and this is in accord with Gaupp’s conclusion in
his work on Lacerta, where he says (’00, p. 553) that this septum
must either be a wholly new formation of the tropibasic (tro-
pidobasic) cranium or be formed from material derived from
the side walls and floor of the platybasic (homalobasic) cranium,

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

270 EDWARD PHELPS ALLIS, JR.

and he definitely decides in favor of the latter supposition.
Fuchs, however, decides just as definitely in favor of the first-
mentioned supposition, for he says (12, p. 104) that, in Che-
lone, the trabeculae take no part in the formation of the septum;
that the septum is a new formation, peculiar to the tropibasic
cranium; that it first appears as a keel-shaped outgrowth (Vor-
wolbung) on the ventral surface of the primordial basis cranii,
and that it increases in height by growing upward. How a
ridge on the ventral surface of the basis cranii could increase
in height by growing upward is not at first quite clear, but in
certain of the figures given by Fuchs the fundament of the
septum is shown lying between the trabeculae, and hence ca-
pable of growing upward between them. This would of course
leave the trabeculae near the ventral end of the septum, and
this is the position in which they are shown in one of the figures
given by Fuchs (I. c., fig. 16 b). It is further said that, in later
stages of development than that shown in the above-mentioned
figure, the trabeculae are no longer recognizable in the optic
region, but persist in the region of the hypophysis and from
there run forward and fuse with the lower, thickened portion
of the septum interorbitale. It is, however, particularly said
by Fuchs that the ventral portions of the side walls of the cra-
nium are here formed by the trabeculae, and that, in the em-
bryo shown in his figures 16a and 16b, the trabeculae, in the
region anterior to the nervus opticus, are reduced to connective
tissue cords which lie near the upper end of the septum.

There is thus a difference of opinion as to the manner in
which this septum arises, and there would also seem to be some
confusion in Fuchs’s statements regarding it. My own work
leads me to suggest that the epichordal and hypochordal bands
of skeletogenous material, known to be developed in the spinal
region of embryos, are continued forward into the prespinal
region, and that the trabeculae are there developed from them.
These two morphologically distinct portions of the trabeculae
are fused to form the basis cranii in the orbital region of the
platybasic cranium, just as they are always fused, in embryos,
to form the parachordal plate in the prootic region, and usually

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 21

so fused in the basioccipital region also. In the tropibasic cran-
ium they have been forced apart, doubtless by pressure of the
eyeballs, and the interorbital septum is formed from the ma-
terial of the hypochordal bands and the tissues between them
and the epichordal bands, the latter bands forming the floor
and side walls of this part of the cranial cavity. The trabec-
ulae might then be said by certain authors to lie at the ventral
end of the interorbital septum, and by others to lie at its dorsal
end. This would also explain how, in fishes where the inter-
orbital and internasal septa are directly continuous with each
other, the trabeculae are said by certain authors to form the
ventral edge of the internasal septum, and by* certain others
to form its dorsal edge (Allis, 713).

LEPIDOSTEUS OSSEUS

In Lepidosteus there is no functional myodome, but the pre-
existing spaces which correspond to both its dorsal and ven-
tral compartments occur and were fully described by me in my
work on the mail-cheeked fishes. The space representing the
dorsal compartment lies, as does the functional myodome of
Amia, dorsal to the cartilage which actually forms the basis
cranil, the space that represents the ventral compartment lying
ventral to that cartilage, between it and the underlying para-
sphenoid, and lodging, as in Amia, the internal carotid arteries
and the palatine branches of the facialis nerves. Veit (’07), in
a work that did not appear until after my own was sent to press,
had previously described, in a 150-mm. specimen o’ this fish,
the space representing the prootic portion of the dorsal com-
partment, calling it the cavum saccivasculosi, and he later (711),
described, in younger specimens, the development of the car-
tilages that bound that space.

In 8 to 16-mm. embryos of this fish, Veit (’11) says that the
notochord is the only recognizable skeletal element; and it ends
with a blunt point against the hind wall of the infundibulum, its
extreme tip turning slightly ventrally. In embryos 10 to 11
mm. in length the notochord is in similar position, but three
cartilaginous elements have now developed on either side of

272 EDWARD PHELPS ALLIS, JR.

the brain: a parachordal cartilage which extends from the
transverse plane of the root of the nervus glossopharyngeus to
that of the root of the nervus trigeminus; a polar (Pol) cartilage,
which lies lateral to the anterior end of the notochord and ex-
tends, in a direct anterior prolongation of the line of the para-
chordal, from the root of the trigeminus to about the middle
of the length of the hypophysis; and a trabecular cartilage,
which, lying in the line prolonged of the other two cartilages,
extends from about the middle of the length of the hypophysis
to a point in front of the nervus opticus. The fundament of
the musculus rectus externus of either side lies directly against
the related polar cartilage.

In embryos of this fish 11 to 12-mm. in length the adjoining
ends of the parachordal polar, and trabecular cartilages of
either side have fused with each other to form a continuous
cartilage, the part formed by the polar and trabecular cartilages
lying, as shown in the figures, slightly dorsal to the level of the |
anterior end of the notochord. The trabeculae of opposite
sides have fused with each other anterior to the recessus pre-
opticus, thus enclosing a large fenestra basicranialis, into the hind
end of which the anterior end of the notochord projects. The
polar cartilages now occupy the positions of the so-called an-
terior prolongations of the parachordals of Swinnerton’s and
Gaupp’s descriptions of Gasterosteus and Salmo, and hence
of the ‘Balkenplatten’ of Stohr’s descriptions of Salmo. The
recti externi have now become inserted on the polar cartilages,
and, because of this or for some other reason, the fenestra basi-
cranialis is there slightly constricted. The fenestra encloses the
ventral portions of the infundibulum and recessus preopticus,
and in later stages the hypophysis and saccus vasculosus come
to lie, respectively, in the interpolar and interparachordal por-
tions of it.

In embryos 14 to 20-mm. long the region under consideration
has not changed in any important respect. The planum orbi-
tonasale, formed by the fusion of the anterior ends of the tra-
beculae, begins immediately anterior to the recessus preopticus
and extends forward beyond that part of the lamina terminalis

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 2738

which forms part of the actual ventral surface of the brain,
this lamina being bent at nearly a right angle and presenting
surfaces that are the one actually ventral and the other an-
terior. A prootic bridge has not yet begun to be formed,
but it is shown by Parker (’82) in somewhat older embryos,
and is there at first separated from the otic portion of the para-
chordal basal plate by a large fenestra basicranialis posterior
similar to the one in Amia and Salmo.

Veit does not give the relations of the internal carotid and
efferent pseudobranchial arteries to the cartilages bounding
the fenestra basicranialis of his descriptions, but in small em-
bryos of this fish (size not given) I found (’09) the internal ca-
rotid running forward beneath the basis cranii, being there
joined by the efferent pseudobranchial artery, and the artery
so formed then turning upward through the fenestra basicrani-
alis. Whether the part of the fenestra so traversed lies between
the polar or trabecular cartilages cannot be definitely told by
comparison with Veit’s figures, but it would seem as if it must
be between the hind ends of the trabeculae, the membranous
pituitary sac lying dorsal to the polar cartilages. The con-
ditions in this fish thus differ from those in the adult Amia only
in that the efferent pseudobranchial artery does not traverse
a foramen and canal in the cartilage of the basis cranii before
falling into the internal carotid, and in that the recti externi have
not invaded the myodomic space.

POLYPTERUS

In the neurocranium of the adult Polypterus there is a large
pituitary fossa, the posterior portion of which is roofed by a
horizontal bridge of the so-called sphenoid bone. The hind end
of the pituitary body projects posteriorly beneath this bridge,
and Waldschmidt (’87) shows it there surrounded by what he calls
‘maschiges, fettartiges Gewebe.’ This tissue apparently fills
the space between the membranous pituitary sac and the walls
of the cartilaginous pituitary fossa, the space thus correspond-
ing to the prootic portion of the functional myodome of Amia.
In his text figure 8, Waldschmidt shows the side wall of the

274 EDWARD PHELPS ALLIS, JR.

pituitary fossa perforated by a cord of tissue. It is not said
what this cord of tissue is, but it is undoubtedly the pituitary
vein described by me (Allis, 08a) in a 75-mm. specimen of Polyp-
terus senegalus. This pituitary vein falls into a vein that I
called the internal jugular, but which is more appropriately called
the vena orbitalis inferior. This vein comes from the orbit, ac-
companied by the internal carotid artery and the nervus palati-
nus facialis, and after receiving the pituitary vein, is joined
by a vein that I called the external jugular, but which is a vena
orbitalis superior and is accompanied by the external carotid
artery. The yein formed by the fusion of these two is the vena
jugularis of the present descriptions. Running posteriorly,
it traverses a short canal in the cartilaginous portion of the
lateral wall of the chondrocranium, between the foramina by
which the nervi trigeminus and facialis traverse that wall, and
issues from the cranium, with the nervus facialis, at the hind
edge of the ascending process of the parasphenoid. The ex-
ternal carotid artery unites with the internal carotid, and the
artery so formed continues posteriorly in a canal through the
ascending process of the parasphenoid, accompanied by a sym-
pathetic nerve. At the hind end of this canal it receives the
efferent artery of the hyoid arch, and then, becoming the lateral
dorsal aorta, enters the aortal canal in the basioccipital, already
referred to when describing the conditions in Hyodon.

The conditions in this fish are thus markedly different from,
but nevertheless strictly homologous to, those in the other fishes
so far considered. ‘There is a dorsal myodomic cavity strictly
similar to that in the Holostei, and a ventral compartment rep-
resented by the canals through the ascending processes of the
parasphenoid. The median portion of the parasphenoid and
the lateral walls of the canals through the ascending processes
of that bone must then, together, correspond to the parasphenoid
of Amia, the mesial walls of the latter canals corresponding to
the ascending processes of the parasphenoid of Amiurus. The
canal traversed by the vena jugularis, which lies partly in the
lateral wall of the chondrocranium and partly between that
wall and the lateral wall of the ascending process of the para-
sphenoid, is the pars jugularis of a trigemino-facialis chamber.

bo

MYODOME AND TRIGEMINO-FACIALIS CHAMBER res,

CHONDROSTEI

The descriptions that I find of the pituitary region of the
cranium of the Chondrostei are incomplete, and but little can be
said about it. A slight pituitary fossa is:sshown by Bridge (’79) in
the chondrocranium of Polyodon, and both I (711) and Dan-
forth (12) have described the arteries in this fish. In embryos
of from 150-mm. to 170-mm. in length the internal carotid runs
forward along the ventral surface of the neurocranium, at first
ventral to a short lateral process of the parasphenoid, and then,
anterior to that process, in a groove on the ventral surface of
the lateral edge of the basis cranii, lateral to the lateral edge
of the parasphenoid. The artery there becomes enclosed in
dense fibrous tissues which are attached to the cranial wall,
and while in the canal thus formed, it is joined by the nervus
palatinus facialis, which issues from the cranial cavity through
a special perforation of the cranial wall. The internal carotid
then enters a canal in the cranial wall, receiving while in it,
the efferent pseudobranchial artery, and then immediately
gives off the arteria ophthalmica magna. <A small pituary vein
is sent outward from the pituitary fossa, through a special
foramen in the cranial wall, and falls into the vena jugularis.
The nervus abducens traverses a short canal in the cartilage
of the basis cranil and, issuing from it, apparently again lies
in the cavum cerebrale crani, from which it definitely issues
with the main root of the nervus trigeminus.

There is thus evidently, in Polyodon, a subpituitary space
corresponding to the myodomic cavity of Amia, but the con-
ditions need further investigation. The ventral compartment
of the teleostean myodome is represented in the canal of fibrous
tissue traversed by the internal carotid artery and the nervus
palatinus facialis, this apparently corresponding to the canal
through the ascending process of the parasphenoid of Polypterus.

276 EDWARD PHELPS ALLIS, JR.

PLAGIOSTOMI

Gegenbaur (’72) describes, in the Selachii, a large pituitary
fossa (Sattelgrube), which extends from the postclinoid wall
(Satellehne) to a traverse presphenoid bolster (Praesphenoid-
vorsprung) which lies slightly anterior to the foramina optica,
and is said to lodge the lobi inferiores anteriorly and the pitui-
tary body posteriorly. The presphenoid bolster is said to vary
greatly in importance in different species of the Selachii and
to be wholly wanting in some of them, the Scylliidae being in-
cluded among the latter. The pituitary fossa is, in certain of
these fishes, everywhere lined with the dura mater, this mem-
brane forming both the perichondrial lining of the fossa and
the sae which encloses the pituitary body. In others of these
fishes there is a deeper, posterior portion of the fossa, shut off
from the cavum cerebrale cranii by a portion of the dura mater
which extends dorsoposteriorly from its anterior edge to the
summit of the postclinoid wall. This subdural portion of the
fossa is said to be traversed by the arteria carotis interna (vor-
dere Carotis), by a vein, and by a lymph canal which Gegen-
baur calls the canalis transversus. When this subdural space
is wanting, the canalis transversus and the internal carotid ar-
teries are separately enclosed in the cartilage of the basis cranil.

Parker (’76) later described the conditions in Scyllium canic-
ula, and in his figures of embryos of that fish he shows condi-
tions in the pituitary region strictly similar to those described
and figured by Gegenbaur in the adult of Scyllium catulus. In
two figures of the adult, Parker, however, shows a small pitui-
tary fossa which lodges the pituitary body and is separated
from the so-called infundibulum by a tall preclinoid wall. I
have heretofore always considered this condition in this fish
to be either an abnormality in the particular specimen ex-
amined by Parker, or a condition due to great age, for Parker
shows both the preclinoid and postclinoid walls strongly calci-
fied. I have, however, now examined two small adults of this
fish, and I find the canalis transversus of Gegenbaur’s descrip-
tions occupying exactly the position of Parker’s pituitary fossa,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER DET

and it is unusually large in both these specimens. It there-
fore seems certain that, if Parker’s figures and descriptions be
not wholly wrong in this particular respect, the specimen ex-
mined by him must have been exceptional and abnormal.

In my work on Mustelus (’01), I found the canalis trans-
versus of Gegenbaur’s descriptions traversed by the pituitary
veins, and not by a lymph vessel, and this was later confirmed
by work on other Selachii (Allis, ’14 a). In this latter work I
found the deeper, posterior portion of Gegenbaur’s descriptions
of the pituitary fossa particularly well developed in Chlamy-
doselachus, and I said of it that it had ‘‘the appearance of
being a somewhat separate and independent fossa.’ It is sub-
pituitary, as well as subdural in position, and is filled with tissues
that seem to be in part tough connective tissues and in part
of a different character.

In all the Selachii I have examined or can find described,
the internal carotid arteries always le anteroventral to the pitui-
tary veins, as they do in the Teleostei and Holostei, and they are
always separated from those veins by either membrane or carti-
lage. They always either fuse with each other in the median
line, or are there connected by cross-commissure, and this fusion
of the arteries is certainly not due, as it apparently is in the
Teleostei, to any pressure of the muscles of the eyeball. In
Heptanchus, Mustelus, and Acanthias I found these arteries
joined by the efferent pseudobranchial arteries, either while
still in the cartilage of the basis cranii or while lying between
that cartilage and the lining membrane of the cavum cerebrale
cranil. The internal carotids of these fishes thus do not enter the
cavum cerebrale cranii until after they have received the efferent
pseudobranchial arteries, which perforate the side walls of
the pituitary fossa slightly anterior to the internal carotid
canals, approximately in the region between the hind ends of
the lobi inferiores and the pituitary body. In Chlamydose-
lachus I found the internal carotids entering the cavum cerebrale
cranii before they received the efferent pseudobranchial arteries,
but I now think this may be an error. The nervus palatinus
facialis does not, in any of these fishes, come into any relation

278 EDWARD PHELPS ALLIS, JR.

to any part of the pituitary fossa, running forward, after issu-
ing from the cranial cavity, ventral to the chondrocranium.

The early development of the cartilages in this region of
these fishes differs somewhat from that in the Teleostei and
Holostei. According to Sewertzoff (99), the trabeculae, when
first formed, are independent cartilages, which lhe oral, and
hence morphologically ventral, to the hypophysis, and because
of the marked cranial flexure at this stage of development,
these cartilages lie ventral to the parachordal plate and per-
pendicular to it, slightly posterior to its anterior edge. In
later stages of development the anterior portions of the tra-
beculae are said by him to fuse with each other, their hind ends
still remaining separate, but having now fused with the ventral
surface of the parachordal plate. An opening is thus enclosed
between the trabeculae and the parachordal plate, and the
hypophysis is said to traverse it. It is called by Sewertzoff the
intertrabecular basal fontanelle, and, as shown by Parker in
Seyllium (’76, fig. 6, pl. 35), has approximately the extent of
the pituitary fossa of the adult fish. In later stages this large
fontanelle is greatly reduced by progressive fusion of the tra-
beculae, both anterior (ventral) and posterior (dorsal) to the
hypophysis, but Sewertzoff says that the hypophysis still pro-
jects through it, and he so shows it in transverse sections of
embryos of Acanthias (I. c., figs. 14 and 15, pl. 30). The stalk
of the hypophysis is said to run forward from this point and to
end blindly, and it apparently does not traverse the persisting
portion of the basal fontanelle in Acanthias, but it does in Pris-
tiurus (I. c., figs: 23 to 25, pl. 31).

The course of the internal carotid and efferent pseudo-
branchial arteries is not given by Sewertzoff, but the internal
carotids must certainly have traversed the posterior portion
of the large primitive fontanelle, and hence that part of that
fontanelle which persists in the oldest embryos of Acanthias
described by him. It would, however, seem as if they could
not have traversed that part of the fontanelle that persists in
Pristiurus, for that part lies considerably anterior to the hy-
pophysis, between the hind edges of the fenestrae opticae (I. c.,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 279

fig. 27, pl. 31), and hence at the anterior end of the pituitary
fossa of the adult.

The pituitary veins, also, are not described by Sewertzoft,
and neither they nor their foramina are indicated in his figures.
They are, however, apparently shown by Baumgartner (’15)
in sagittal sections through this region in embryos of Acanthias.
In that author’s figures 2 to 9, he shows a vessel ventral to the
anterior end of the notochord, and morphologically posterior
to the hypophysis. This vessel is not lettered in the figures,
but it must certainly be a cross-section of the venous commissure
formed by the pituitary veins. In Baumgartner’s figure 9, it
is shown lying between the parachordal plate above and a ventro-
anteriorly directed process of cartilage that is apparently con-
sidered by Baumgartner to be of parachordal origin, but which
must represent a section through that part of the trabecular
cartilage of Sewertzoff’s descriptions which is formed by the
fusion of the trabeculae of opposite sides dorsal (posterior) to
the hypophysis. This commissural vein would then pass dorsal
to the trabeculae, as it normally should. The process shown
by Baumgartner forms the posterior boundary of an opening
between it and the hind end of the trabecular cartilage, and is
hence the intertrabecular basal fontanelle of Sewertzoff’s descip-
tions, and a vessel, possibly the internal carotid artery, is shown
lying directly in it.

The so-called intertrabecular basal fontanelle of these embryos
of the Selachii would then seem to correspond to the fused anterior
and posterior basicranial fenestrae of Salmo and Gasterosteus,
the definitive fenestra of Pristiurus corresponding to the fenestra
hypophyseos of Salmo and Gasterosteus, and the definitive
fenestra of Acanthias corresponding to the fenestra basicranialis
posterior of those fishes. This latter fenestra is, as will be later
shown, the fenestra hypophyseos of the Dipnoi, Amphibia, and
Sauropsida, in which the internal carotid arteries traverse the
fenestra along its posterior border, sometimes separated by a
median cartilage called the intertrabecula.

Neither Sewertzoff nor Baumgartner describe polar cartilages
in these fishes, but van Wijhe (’05) describes them in early

280 EDWARD PHELPS ALLIS, JR.

embryos of Acanthias, between the trabeculae and parachordals
and primarily independent of those cartilages, thus correspond-
ing to the hind ends of the trabeculae of Sewertzoff’s descriptions.
and apparently also to the median, ventro-anteriorly directed
process of the parachordal of Baumgartner’s description.

The conditions in these fishes thus show that chrondification
has taken place to such an extent in the prootic and subpituitary
regions that the dorsal compartment of the teleostean myodome
has been reduced, either to canals traversed by the pituitary
veins or to some part of a deeper, posterior portion of the pituitary
fossa of the chondrocranium. The remainder of the deeper
portion of the fossa represents an anterior extension of the dorsal
myodomie cavity which has been developed in some relation
to the enclosure of the internal carotid arteries in it. The sub-
dural canals traversed by those arteries after they leave this
subpituitary space evidently form anterior prolongations of it,
and were they to be added to it, and the pituitary fossa reduced
to the proportions in Ceratodus and higher vertebrates, the
arteries would traverse a peripituitary space separated from the
cavum cerebrale cranii by the dura mater. The conditions
here thus seem to indicate that the fenestra hypophyseos of
these fishes is the homologue of the fenestra interparachordalis
of the Holostei and Teleostei, and not of the fenestra hypoph-
yseos of those fishes. The foramina carotica of Amia and
the Selachii are then not homologous.

No ventral myodomiec cavity is found in these fishes, except
as it may be represented in a part of the canals traversed by the
internal carotid arteries. The cross-commissure between these
arteries has a position which suggests that it may have been
utilized, in the Teleostei, to form the cross-commissure between
the efferent pseudobranchial arteries.

In certain of these fishes, a canal in the lateral wall of the chon-
drocranium, traversed by the vena jugularis, represents, as in
Polypterus, a pars jugularis of a trigemino-facialis chamber
(Allis, ’14b).

In the Batoidei, the pituitary fossa, as shown in Gegenbaur’s
figures, is but slightly developed, but as he says that a canalis

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 281

transversus is found in these fishes, as in the Selachii, the con-
ditions are probably strictly similar.

DIPNOI

In Ceratodus, the so-called pars ascendens of the anterior
process of the palatoquadrate of Greil’s (13) descriptions forms
the lateral wall of a space which, in an earlier work (Allis, ’14 ¢),
I showed to be the homologue of the trigemino-facialis chamber
of the Holostei. In early embryos of Ceratodus this chamber
has anterior and posterior openings which Greil calls, respec-
tively, the foramen sphenoticum commune and the foramen
praeoticum basicraniale. In older embryos the foramen sphen-
oticum commune becomes separated into four parts by bars of
cartilage developed in the connective tissues surrounding the
nerves and vessels which traverse the foramen. One of these
parts, called by Greil the foramen sphenoticum majus, transmits
all the branches of the nervi maxillo-mandibularis and lateralis
trigemini and the vena and arteria temporalis, the latter artery
being the carotis externa of my descriptions of other fishes. <A
second foramen, called the foramen sphenoticum minus, trans-
mits the nervus profundus and the vena capitis media, this
latter vein being also called the vena pterygoidea. A third
foramen, called the foramen hypoticum, transmits the nervus
oticus trigemini; the fourth foramen transmitting the nervus
abducens. The posterior opening of the chamber, the fora-
men praeoticum basicraniale, does not undergo subdivision in
the oldest embryos considered by Greil, and it is traversed by the
nervus facialis, the ramus palatinus facialis, the arteria tem-
poralis (carotis externa), and the vena capitis lateralis; the latter
vein being a posterior continuation of the vena capitis media
(pterygoidea), and the two together forming the vena jugularis of
my descriptions of other fishes. The floor of the trigemino-
facialis chamber is formed by the processus basalis of the pala-
toquadrate, and the palatinus facialis, after issuing through
the posterior opening of the chamber, runs forward ventral to
this floor, between it and the underlying parasphenoid.

282 EDWARD PHELPS ALLIS, JR.

In these embryos the hypophysis lies at the hind end of a large
fenestra basicranialis, and even projects posteriorly slightly
beyond and beneath the tip of the notochord. The fenestra
basicranialis is bounded laterally by cartilages which Greil
considers of trabecular origin, the parachordal cartilage not
extending anteriorly beyond the tip of the notochord. A vena
hypophyseos is said to arise in the neighborhood of the hypoph-
ysis and to issue from the cranial cavity through a foramen
sphenolaterale, which lies dorsal to the trabecula and anterior to
the foramen sphenoticum minus. ‘This vein falls into the vena
pterygoidea (jugularis), and although it is not said to be con-
nected with its fellow of the opposite side by a cross-commissural
vessel, it is certainly the pituitary vein of my descriptions. There
is no indication, in the figures given, of amembrane separating
this vein from the cavum cerebrale cranii, but this membrane
must certainly exist, for it occurs in all other fishes so far
considered.

In early embryos the arteria carotis interna is connected
with its fellow of the opposite side by a cross-commissural ves-
sel, immediately posterior to the hypophysis and immediately
ventral to the tip of the notochord, but Greil says this cross-
commissure has aborted in the oldest embryos examined by him.
Anterior to this cross-commissure, the artery gives off an arteria
palatina, which runs forward ventral and mesial to the trabe-
cula. The artery itself then runs upward mesial to the trabe-
cula of its side and is distributed mainly to the brain, one branch,
however, the arteria orbitalis, being sent outward through the
foramen sphenolaterale with the pituitary vein, and a second
branch, the arteria ophthalmica sent outward with the ner-
vus opticus through the foramen opticum. Before passing up-
ward through the fenestra basicranialis, the internal carotids
are said to lie between the ventral surface of the chondrocranium
and the underlying parasphenoid.

In the adult, the large fenestra basicranialis of the embryo
is shown entirely closed by cartilage in the median vertical sec-
tions given by Giinther (’71), Huxley (’76), and Bing (’05),
and each of these authors shows a deep pituitary fossa with

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 283

pronounced posteclinoid and preclinoid walls. Bing says the
hypophysis lies in the posterior portion of this fossa, the an-
terior portion being filled with arachnoidal tissue (Arachnoideal-
maschen). The postclinoid wall is evidently formed by growth
of the epichordal and hypochordal bands of parachordal carti-
lage said by Greil to enclose the tip of the notochord in embryos.
The preclinoid wall had not begun to be developed in the oldest
embryos described by Greil. No foramina leading into the
pituitary fossa are shown or described by any of these three
authors.

I find, in an old and somewhat dissected skull of this fish,
a perforation of the cartilage of the basis cranii at the bottom
of the posterior portion of the pituitary fossa, and it is closed
by tough membrane. <A small canal in the cartilage leads from
either orbit to the edge of this membrane and must certainly
have transmitted a vein which either traversed the membrane
or passed dorsal to it, in order to reach and drain the hy-
pophysis. The space traversed by this vein, wherever it may
be, is a dorsal myodomic cavity. The internal carotid artery
of either side passes internai to the parasphenoid, is there
joined by the efferent pseudobranchial artery (mandibular aortic
arch of Greil’s descriptions), and then becomes embedded in the
cartilage of the basis cranii and covered externally by membrane.
The arteries of opposite sides are connected by a cross-commis-
sural vessel which lies posterior to the median perforation in
the floor of the pituitary fossa, the canal traversed by this cross-
commissure representing part of a ventral myodomic cavity.
Anterior to this cross-commissure each artery runs forward
ventral to the pituitary vein, sends forward the arteria palatina,
and then certainly enters the pituitary fossa through a foramen
that I find lying anterolateral to the median perforation in the
floor of the fossa, but, as my skull had been cleaned and the
arteries removed, I cannot definitely establish this. If it tra-
verse this foramen, as seems certain, it must enter and traverse
that anterior portion of the pituitary fossa which Bing says is
filled with arachnoidal tissue, this part of the fossa then repre-
senting the internal carotid canals of Amia, fused with each

284 EDWARD PHELPS ALLIS, JR.

other and become part of the pituitary fossa of the chondro-
cranium. Whether this part of the fossa is separated from the
cavum cerebrale cranii by the dura mater or not cannot be told
from my specimen, but comparison with other fishes and with
higher vertebrates show that it must be.

There are thus, certainly, in this fish, both dorsal and ven-
tral myodomic cavities, and the dorsal cavity has apparently
fused with the prepituitary portions of the canals traversed by
the internal carotid arteries to form a single peripituitary space
similar to that found in higher vertebrates and represented in
the cavernous and intercavernous sinuses of man, as will be
explained later. The foramina carotica lie at the hind edge of
the pituitary fossa, as they do in higher vertebrates. The cross-
commissure connecting the internal carotids is evidently the
homologue of the cross-commissure in the Selachii, and probably
not the homologue of the anastomosis of the arteries of opposite
sides in the Teleostei.

The bar of cartilage separating the foramina sphenoidea
majus and minus is the homologue of the pedicel of the alisphenoid
of Amia, and if the anterior edge of this bar of cartilage were
to grow forward so as to pass beyond the foramina for the pit-
uitary vein and the oculomotorius and trochlearis nerves, it
would give rise to the orbital opening of the myodome of Amia.

AMPHIBIA

In the Amphibia there apparently is no vein comparable to
the pituitary vein of fishes, for I find no such vein described,
and the pituitary region is said to be drained, in certain of these
vertebrates, by branches of intracranial veins. It might be
assumed that the myodomic conditions here were as in Amiu-
rus, where the pituitary veins are also wanting, but it seems
much more probable that the ventral processes of the prootics
have here been wholly suppressed, and that the basis cranii
corresponds to the roof of the dorsal compartment of the myo-
dome of fishes, and hence represents the primary basis cranii.
The course of the internal carotid artery in Rana, and that of
the nervus abducens both in Rana and Salamandra, favor this
interpretation.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 285

The internal carotid artery of Rana is said by Gaupp (’93 b),
p. 403) to pass upward, in early embryos, mesial to the trabe-
cula of its side, but, because of enveloping growth of the trabe-
cular cartilage, soon to become enclosed in a primary foramen
caroticum. Having traversed this foramen and entered the
cranial cavity, the artery gives off the arteria carotis cerebralis-
and then itself issues from the cranial cavity through the foramen
oculomotorium as the arteria ophthalmica. In later stages,
that part of the trabecula between the foramina caroticum and
oculomotorium is resorbed, and the internal carotid is said then
to lie in the orbit and to send its cerebral branch inward through
the foramen oculomotorium. Comparing these conditions in
Rana with those I have described in the Teleostei, it is evident
that the primary foramen caroticum of Rana must he in what
corresponds to the floor of the cavum cerebrale cranii of Amia
and the Teleostei, for that floor, alone, is continuous with that
part of the cranial wall which is perforated by the foramen
oculomotorium.

The nervus abducens of Rana is said by Gaupp to issue from
the cranial cavity in the sheath of the ramus orbitonasalis tri-
gemini, and to pass, with that nerve, under, and hence morpho-
logically anterior to, the processus ascendens quadrati. In fishes
the corresponding branch of the trigeminus (nervus profundus)
passes mesial and anterior to the pedicel of the alisphenoid,
and always lies dorsolateral to the myodome, never traversing
it. Comparison of these conditions would accordingly indi-
cate that the dorsal myodomic cavity is wanting. in Rana. In
Salamandra, Fuchs (10) shows the nervus abducens perforat-
ing the basis cranii and then lying mesial to the arteria carotis
interna in a canal between the basis cranii and the parasphenoid,
the nervus palatinus facialis lying lateral to the carotis interna.
The hypophysis lies in a perforation of the basis eranii, and
even projects ventrally slightly beyond it, lying in a slight con-
cavity on the dorsal surface of the parasphenoid. The dorsal
myodomic cavity must accordingly be wholly suppressed here
by failure of the ventral processes of the prootics to develop,
the canal which lodges the internal carotid artery and the nervus

JOURNAL OF MORPHOLOGY, VOL. 32, No. 2

286 EDWARD PHELPS ALLIS, JR.

palatinus lying directly beneath the floor of the cavum cere-
brale cranii. Thus this canal is, as Fuchs says, not the homo-
logue of the canalis parabasalis of reptiles, and also not the
homologue of that same canal in fishes.

The fenestra hypophyseos of the Amphibia is then the homo-
logue of the pituitary opening of the brain case of fishes, and
not of either the fenestra hypophyseos or the fenestra ventralis
myodomus.

The antrum petrosum laterale of Driiner’s (’01) descriptions
of the Urodela represents some part of a trigemino-facialis cham-
ber, and quite certainly its pars jugularis only (Allis, ’14 d), the
pars ganglionaris of the chamber then being enclosed within
the cranial wall. The lateral wall of the pars jugularis of the
chamber is formed by that part of the palatoquadrate terminat-
ing in the processus oticus, the processus ascendens quadrati,
which is the homologue of the pedicel of the alisphenoid of fishes,
forming the lateral wall of a space which corresponds to the
orbital opening of the myodome of Amia. There, however,
apparently is, in these vertebrates, no cartilage corresponding
to the floor of that opening of Amia.

In the Anura the conditions are apparently similar to those
in the Urodela, for, in embryos of Rana, Gaupp (’93 b) shows
the trigemino-facialis ganglion lying within the chondrocranium.

REPTILIA

In my work on the mail-cheeked fishes, I came to the conclu-
sion that there was, in the pituitary region of the chondrocranium
of Lacerta, a ‘space of uncertain dimensions’ which corresponded
to a part, if not the whole, of the myodome of fishes. This
space was between the cartilaginous floor of the cranial cavity
and a membrane which was assumed to overlie it and to form the
actual floor of the cavum cerebrale cranii, but I could not then
find this membrane described. It is, however, shown by Gaupp
(02, fig. 6, p. 172), well developed, in a figure of a cross-
section through the prootic region of a 32-mm. embryo of
Lacerta, and in the space between it and the cartilaginous
basis cranii the hypophysis and the nervi abducentes are shown.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 287

The foramina for the internal carotid arteries are cut in the sec-
tion, lying in the floor of this space and separated from each other
by a median piece of cartilage which lies at a slightly lower level
than the cartilage on either side of it. The internal carotid
arteries are shown lying ventral to the basis cranii, each artery
accompanied by, and lying mesial to, the nervus palatinus faci-
alis of its side. No parasphenoid bone is shown, but compari-
son with a figure of a 47-mm. embryo (Gaupp, 705 b, p. 763)
shows that that bone lies ventral to the nerve and artery and
forms the floor of the canalis parabasalis of Gaupp’s later de-
scriptions (’05 a, p. 292), this canal of Lacerta thus being the ho-
mologue of the palatine canal of my descriptions of Amia. The
piece of cartilage between the foramina carotica is the intertra-
becula of Fuchs’s (712) descriptions of Chelone, and, as it forms
part of the floor of the little space here under consideration, it
cannot be part of the crista sellaris, as the lettering in Gaupp’s
figure of the entire chondrocranium of Lacerta (’00, fig. 1)
would lead one to suppose. The nervus abducens enters the
space here under consideration by traversing a foramen which
perforates the cartilage of the chondrocranium, lateral to the
lateral end of the crista sellaris, and issues from it into the orbit.
No pituitary veins are shown in Gaupp’s figure of a cross-sec-
tion through this region in Lacerta, but in an earlier work (’93,
p. 571) he fully describes them. A vein lies along each lateral
surface of the middle lobe of the hypophysis and is connected
with its fellow of the opposite side by several cross-commissures,
the largest of which lies posterior to the hypophysis. From
either end of this posterior cross-commissure an important vein
leads into a large vein which drains the blood from the orbital
venous sinus, and the vessel so formed falls posteriorly into the
vena jugularis interna. These veins thus must traverse the
space of uncertain dimensions mentioned in my earlier work,
which is a dorsal myodomic cavity. The internal carotid ar-
teries run upward through the fenestra hypophyseos, and then
along the lateral surfaces of the lateral lobes of the hypophysis,
lying, in their course, anteroventral to the pituitary veins.

288 EDWARD PHELPS ALLIS, JR.

A ventral myodomic cavity is represented in those parts of
the canales parabasales posterior to the foramina carotica.
The antipterygoid is said by Gaupp (’00, pp. 541 and 542) to
be the homologue of the ascending process of the quadrate of
the Amphibia and to be wholly wanting in the cranium of mam-
mals. The ala temporalis of the mammalian cranium is con-
sidered by him to be represented, in reptiles, by the processus
basipterygoideus. Fuchs (712, pp. 91 to 95), on the contrary,
maintains that the antipterygoid (epipterygoid, Fuchs) is the
homologue of the mammalian ala temporalis, and that the pro-
cessus basipterygoideus is the homologue of the processus alaris
of the ala temporalis. To explain the difference in the relations
of the nervus maxillaris trigemini to the antipterygoid and ala
temporalis, he assumes that the nerve has, in mammals, slipped
over the top of the antipterygoid in early stages of development.
I formerly concluded (14d) that the antipterygoid of La-
certa was the homologue of the pedicel of the alisphenoid of
Amia, and the processus basipterygoideus the homologue of the
floor of the orbital opening of the myodome of Amia. The pars
ascendens of the quadrate formed the lateral wall of the post-
trigeminus portion of a trigemino-facialis chamber, as in the
Amphibia. My present work leads me to consider these con-
clusions correct, but to consider the trigemino-facialis chamber
of these vertebrates to be the homologue of that chamber of
Ceratodus and the Holostei, and not of the chamber of the Am-
phibia and Teleostei; for the lateral wall of the chondrocranium,
both of Lacerta and Crocodilus (Shiino, 714), is certainly the
primitive cranial wall and not the outer wall of a trigemino-
facialis recess. The processus basitrabecularis of Crocodilus
would then represent a part of the floor of that chamber, and
the processus pterygoideus quadrati a part of its lateral wall.
The vena cardinalis anterior of Lacerta is said by Gaupp (00,
pp. 547 and 548) to run posteriorly dorsal to the processus
basipterygoideus and then along the external surface of the chon-
drocranium, thus lying wholly external to that cranium. ‘This
is exactly as it should be under my interpretation of the condi-
tions, for this vein is the vena jugularis of my descriptions of

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 289

fishes, and in Amia it enters the orbital opening of the myodome
and then traverses the trigemino-facialis chamber, lying always
external to the wall of the cavum cerebrale craniil. Gaupp con-
siders this vein to be the homologue of the sinus cavernosus of
mammals, and as that sinus is intracranial in position, he con-
cludes that the space traversed by the vein in Lacerta, which
is actually extracranial, has been added to the cranial cavity in
mammals. The sinus cavernosus is, however, a branch of the
vena cardinalis anterior (capitis media), and not that vein
itself, as will be later explained.

MAMMALIA

Properly to explain the conditions in mammals it is necessary
first to consider the ala temporalis. This element of the cra-
nial wall has been considered by many authors to have its homo-
logue in the antipterygoid of reptiles, but Gaupp considers it, as
stated above, the homologue of the processus basipterygoideus
of those vertebrates. A well-recognized objection to its being
the homologue of the antipterygoid of the Reptilia is that
the nervus maxillaris trigemini (second branch of the trigemi-
nus) les posterior to that element of the reptilian cranium, but
anterior to the ala temporalis of mammals. Gaupp accounts
for this by saying that, because of the absence of an antiptery-
goid in mammals, there was no intervening skeletal element,
and the nerve has simply joined the first branch of the trigem1-
nus instead of remaining with the third. Other authors have
suggested that the nerve has either cut through or slipped over
the top of the antipterygoid, or simply, for some unknown rea-
son, chosen a presumably more direct or advantageous course
on the other side of it. My work leads me to quite a different
conclusion, and I look for the homologue of the ala temporalis
in a part of the lateral wall of the trigemino-facialis recess of
fishes.

In all of the lower vertebrates there is apparently always either
a trigemino-facialis chamber, a pars ganglionaris of that chamber
(trigemino-facialis recess), or both partes ganglionaris and jugu-
laris separated from each other by a wall of bone. The outer

290 EDWARD PHELPS ALLIS, JR.

wall of the pars jugularis of this chamber of fishes, and the
pedicel of the alisphenoid are represented, respectively, in the
Amphibia by the otic and ascending processes of the quad-
rate, the latter process being the homologue of the antiptery-
goid of the Reptilia (Allis, 714 c). These two portions of the
neurocranium of fishes are thus secondarily acquired additions
to it, and one or the other, or even both of them, is frequently
wanting. In the Selachii, the pars ganglionaris of the tri-
gemino-facialis chamber may be separated, by a partition of mem-
brane or cartilage, into trigeminus and facialis portions, the
latter portion then fusing with an acusticus recess to form an
acustico-facialis recess.

Assume that, in a piscine skull, the pedicel of the alisphenoid
and the lateral wall of the pars jugularis of the trigemino-faci-
alis chamber are both wanting, as is actually the case in certain
of the Teleostei; that independent trigeminus and acustico-
facialis recesses have been formed, as in certain of the Selachii;
that the muscles of the eyeballs have not acquired entrance into
the preexisting myodomie cavities, as in many fishes; that these
cavities have been reduced to the conditions found in Cerato-
dus; and that the trigeminus recess has been enlarged to such
an extent that its floor projects ventrally below the level of the
pituitary fossa (sella turcica), as it actually does in many of the
Mammalia. If the wall separating the trigeminus and acusti-
co-facialis recesses were then to be perforated, the facialis por-
tion of the latter would be in communication with the trigemi-
nus recess, and conditions would arise similar to those described
by Voit (09) in rabbit embryos, where the cavum epiptericum
(trigeminus recess) and the cavum supracochleare (facialis
recess) form a continuous cavity which communicates with the
meatus acusticus internus (acusticus recess) through a foramen
faciale primitivum. The facialis nerve would then issue from
the facialis portion of this continuous cavity through a foramen
faciale secundarium, the profundus nerve (first branch of the
trigeminus) and trigeminus issuing from the trigeminus portion
of the cavity, and their foramina of exit lying at the hind end
of the orbit and not far from the foramina of the pituitary vein

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 291

and the oculomotorius, trochlearis, and abducens nerves. In
the Teleostei and Selachii these several last mentioned fora-
mina may lie relatively close together, and the chondrification
or ossification of the tissues of the cranial wall may actually give
rise to marked variations in the number and arrangement of
the definitive foramina. Assume that the tissues surround-
ing the nervi maxillaris and mandibularis trigemini, as they
issue from the trigeminus recess, chondrify to form a vertical
bar of cartilage; that this bar grows forward so as to shut in
the other foramina mentioned above, as the pedicel of the ali-
sphenoid actually does in Amia; and that the tissues separating
these other foramina from each other and from the nervus
maxillaris persist as membrane. This would give rise, in this
hypothetical cranium, to three fenestrations of the cranial wall
which would be strictly similar, so far as the nerves travers-
ing them are concerned, to the fissura orbitalis superior and the
foramina ovale and faciale secundarium of Voit’s description
of embryos of the rabbit. If, then, the venous and arterial
vessels of the region also have the same relations to these fora-
mina that they do to the foramina in the rabbit, there would
seem to be no reasonable doubt that the foramina, and hence
their bounding walls, are strictly homologous.

In fishes the vena jugularis always runs posteriorly mesial to
the pedicel of the alisphenoid, when the pedicel exists, and then
always traverses the pars jugularis of the trigemino-facialis
chamber, when it is present and independent of the pars gan-
ghonaris. When the pars jugularis of the chamber is wanting, the
vein passes along the lateral wall of the neurocranium, whether
that wall be formed by the primary wall of the cranial cavity or
by the lateral wall of a trigemino-facialis recess, never enter-
ing either the recess or the cavum cerebrale cranii. The pit-
uitary vein arises from this vena jugularis and perforates the
cranial wall, anterior to the trigemino-facialis chamber, to enter
the dorsal myodomic cavity, never itself entering either the ca-
vum cerebrale cranii or any part of the trigemino-facialis cham-
ber. A branch is, however, sent into the cavum cerebrale cranii.
to drain the hypophysis, and, in certain Teleostei, this branch

292 EDWARD PHELPS ALLIS, JR.

is connected with an intracranial vein, the encephalic vein of
Allen (05), which enters the trigemino-facialis recess, perforates
its lateral wall posterior to the nervus trigeminus, and falls
into the vena jugularis. In other Teleostei the pituitary vein is
connected with intracranial veins which issue through the fora-
men vagum there to fall into the vena jugularis. If either of
these two connections were to become important, the flow of
blood in the pituitary vein would be reversed, and a vein would
be formed whieh would drain the hypophysial region and would
issue, in the one case, through a foramen jugulare spurium, and,
in the other, through a foramen jugulare.

In the Amphibia and Reptilia the vena jugularis always passes
mesial to the ascending process of the palatoquadrate, or its
homologue, the antipterygoid, and then, in each case, traverses
the pars jugularis of the trigemino-facialis chamber, never there
traversing any portion of the lateral wall of the neurocranium.

The arteria carotis externa of fishes, like the vena jugularis,
always traverses the pars jugularis of the trigemino-facialis
chamber, when that part of the chamber has been separated
from the pars ganglionaris, never traversing the pars ganglio-
naris. On issuing from the chamber into the orbit, it always runs
outward, posterior and lateral to the pedicel of the alisphenoid.
In the Amphibia and Reptilia it traverses the pars jugularis of
the trigemino-facialis chamber, always lying lateral to the
lateral wall of the neurocranium, and issues from the chamber,
posterior to the ascending process of the palatoquadrate in the
Amphibia, or to the antipterygoid in the Reptilia, thus lying
lateral to that element of the cranial wall.

In embryos of the porpoise the vena jugularis of fishes is rep-
resented in the vena capitis media plus the vena capitis lateralis,
and, as described by Salzer (’95), all the cerebral veins empty
into it, some anterior, some posterior to the nervus trigeminus,
between it and the nervus facialis, and some in the region of the
nervus vagus. The anterior of these three connections with
the primitive vena jugularis loses its importance in later stages
of development, the other two increasing, but varying in rela-
tive importance at different stages of development, and appar-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 293

ently also in different species of the Mammalia, this giving rise
to a vena jJugularis interna which issues, either through a fora-
men jugulare spurium or a foramen jugulare, or even through
both those foramina; these two connections with the primitive
vein thus evidently corresponding to those referred to above
in the Teleostei. The sinus cavernosus is said by Salzer (I. ¢,
p. 252) to be formed from the veins which primarily collected
the blood from the eyeball and the orbit, and which acquire
a secondary connection with the sinus petrosus. This secondary
connection must certainly be formed by a vein, the homologue
of the pituitary vein of fishes, which has become important be-
cause of the abortion of the short vertical venous commissure
which primarily connected the venae capites media and later-
alis between the trigeminus and facialis ganglia. I do not find
that Salzer mentions the abortion of this connection, but his
figures show that it is absent in older embryos. Thus the sinus
cavernosus of mammals is the pituitary vein of fishes, and it
is said by Salzer (1. c., p. 242) primarily to have delivered the
blood from the orbital veins into the sinus petrosus. Later,
the flow of blood is reversed, in the porpoise, and the sinus
cavernosus and the orbital veins are drained by the facial vein,
the flow of the blood in the sinus cavernosus thus now being
in the same direction as in the pituitary vein of fishes.

The sinus cavernosus of mammals thus certainly contains
no part of the primitive vena jugularis, but a persisting portion
of that vein forms the connection between it and the orbital
veins. In the Sauria the sinus cavernosus is said by Grosser
and Brezina (’95, p. 323) to be perhaps a remnant of the vena
cardinalis anterior, and there to be extracranial in position (I.
¢., p. 321); neither of which statements is correct, for the con-
ditions are here certainly as in the Mammalia. Gaupp (’00,
p. 548) quotes Grosser and Brezina as here saying that the
sinus cavernosus is actually (wohl) a part of the vena cardinalis,
and adds that he has himself confirmed this, as well as its
extracranial position, in embryos of Chelone.

These statements regarding this sinus led me formerly to
conclude (Allis, ’09, p. 193) that the venous vessel which tra-

294 EDWARD PHELPS ALLIS, JR.

verses the sinus cavernosus of man was the homologue of the
vena jugularis of fishes; that the intercavernous sinuses repre-
sented the pituitary veins of fishes; and that the cavernous and
intercavernous sinuses and the cava Meckelii_ together
represented the myodome of Amia together with its so-called
upper lateral, or trigemino-facialis chamber. This is, how-
ever, an error, for the so-called cavernous and intercavernous
sinuses together represent a dorsal myodomic cavity plus
the internal carotid canals, and the venous vessels tra-
versing this cavity are, together, the homologues simply of
the pituitary veins of fishes. The cavum Meckelii is then sim-
ply a trigeminus recess and not a trigemino-facialis chamber.

In Thane’s figure (94, fig. 405, p. 523) of a transverse sec-
tion through the sinus cavernosus of the adult man, the outer
wall of the sinus, formed by the dura mater, is thickened and
is traversed by the oculomotorius, trochlearis, profundus (first
branch of the trigeminus), abducens and maxillaris trigemini
nerves. The inner wall of the sinus is continued across the dor-
sal surface of the sella turcica, and is there separated by a nar-
row space from the membranous pituitary sac, this space being
traversed, on either side of that sac, by the intercavernous
sinuses. The internal carotid artery enters this sinus through
the inner part of the foramen lacerum, runs forward in the carotid
groove on the lateral surface of the body of the sphenoid, and
turns upward in a semicircular notch on the posterior surface of
the preclinoid wall, this notch representing a remnant of the
internal carotid canal of Amia. The artery lies lateral to the
pituitary vein, but if the myodomic cavity were convex on its
ventral surface, as it is in fishes, instead of concave, as in man,
the artery would lie ventral and internal to the loop formed by
the veins of opposite sides, as it does in fishes. The external
carotid artery lies everywhere external to the cranial wall, as
does also the vena jugularis externa, terminal branches only
being sent into the cranial cavity.

The relations of the veins and arteries of man to the crania
wall are thus, like those of the nerves, strictly similar to those
in the hypothetical piscine cranium here under consideration,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 295

and no suppositions have been made in regard to the latter that
are not warranted by conditions actually found in fishes, ex-
cepting only the formation of a bar of cartilage between the
nervi maxillaris and mandibularis trigemini and the fusion of
the foramen for the nervus maxillaris with certain other foramina
to form a single large fenestra; and, as already stated, marked
variations in the fusions and groupings of the foramina in this
region are of constant occurrence in fishes. It thus seems cer-
tain that the foramina in this region in the two crania are homol-
ogous, and it follows that the lamina ascendens of the ala tem-
poralis of mammals is a bar of cartilage formed between the nervi
maxillaris and mandibularis trigemini as they issue from a
trigemino-facialis recess, and this element of the cranium is
apparently characteristic of these vertebrates. The processus
alaris of the ala temporalis must then be represented in some
ventral portion of the basicapsular commissures of fishes, and
apparently in that part which, in Amia, lies between the
palatine foramen and the floor of the orbital opening of the
myodome. If it includes the latter floor, it must include the pro-
cessus basipterygoideus of reptiles, which seems improbable.

Certain other features of the region, which favor this inter-
pretation of the conditions, may now be considered.

The myodomic cavity of the mammalian cranium, corre-
sponding to the so-called cavernous and intercavernous sinuses
of man, must necessarily extend, on either side, beyond the
lateral edge of the foramen caroticum, and its roof is thus formed
by what Terry (17) has recently described as the spreading
basal portion of his membrana limitans. The carotid foramen
accordingly hes in the floor of this myodomic cavity and not,
as Voit (09) concluded was the case in rabbit embryos, in the
floor of the cavum epiptericum. The nervus petrosus super-
ficialis major (nervus palatinus facialis) of the rabbit does, how-
ever, perforate the floor of the cavum epiptericum, as Voit con-
cluded, this being in accord with its course in the Teleostel,
where it usually perforates the floor of the pars ganglionaris of
the trigemino-facialis chamber, but may occasionally perforate
the floor of the pars jugularis.

296 EDWARD PHELPS ALLIS, JR.

The arteria carotis interna of the rabbit is said by Voit to run
upward through the foramen caroticum into the cavum epip-
tericum, which, as explained above, is certainly incorrect. The
artery is said then to run forward dorsal to the processus alaris
of the ala temporalis, which is in accord with my interpretation
of the conditions, for that process forms part of the floor of the
dorsal myodomic cavity. The artery is said by Voit to lie lat-
eral to a cartilage ‘c,’ which Voit considers to form part of the
lateral wall of the cavum cerebrale cranii. This cartilage is,
however, certainly a chondrification of a membrane shown, in
one of Arai’s figures of this animal (’07, fig. 6, p. 482), running
upward between the hypophysis mesially and the arteria caro-
tis interna and the pituitary vein (so-called sinus cavernosus)
laterally. This membrane is continued mesially between the
hypophysis and the dorsal surface of the sella turcica, and is
shown as a single membrane, but it must necessarily be formed
by the fusion of two membranes, one forming the floor and the
other the roof of the subpituitary myodomic cavity. The carti-
lage ‘c’ is evidently a chondrification of some part of this mem-
brane, and may therefore represent a chondrification of either
one of its two components; and its position and its coalescence
with the floor of the sella turcica seem to indicate that its basal
portion belongs to both membranes while its dorsal portion be-
longs to the dorsal membrane only and forms part of the roof
of the myodomic cavity and hence of the wall of the cavum cere-
brale cranii. The internal carotid accordingly here lies in a
lateral portion of the myodomic cavity which has been sepa-
rated from the median portion of the cavity by this wall of car-
tilage. The cartilages ‘a’ and ‘b’ of Voit are, as he concluded,
remnants of the mesial wall of the cavum epiptericum (tri-
geminus recess).

Because of the passage of the internal carotid through what
Voit considered to be a part of the cavum epiptericum, he con-
cludes (’09, p. 551) that this artery of the rabbit, and hence also
that of others of the Mammalia ditremata, must run upward
lateral to the trabecula, the internal carotid of these animals
thus not being the homologue of the similarly named artery of

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 297

the Mammalia monotremata and of lower vertebrates. Gaupp
had previously suggested that the trabecula had here simply
‘cut through’ the artery, but Voit is not inclined to accept this
suggestion. The bounding walls of this foramen are, however,
under my interpretation of the conditions, of parachordal (polar)
and not of trabecular origin, and there is, accordingly, no
question here of its lying on one side or the other of the trabec-
ula. It does, however, apparently lie lateral to the polar ecar-
tilage, and hence morphologically lateral, instead of mesial, to
the trabecula, and a possible explanation of this will be given
when the polar cartilages are considered later.

The processus pterygoideus arises from the ala temporalis at
the base of its lamina ascendens, and hence, under my inter-
pretation of the conditions, from the ventral edge of the lateral
wall of the trigemino-facialis recess. Its position, alone, thus
indicates that it is a remnant of the lateral wall of the pars jugu-
laris of a trigemino-facialis chamber, and its relations to the
nerves, arteries, and veins are in accord with this conclusion.
The several branches of the nervus trigeminus all lie dorsal to
it, as they should; the nervus petrosus superficialis major runs
forward ventral to it; and the vena capitis media of embryos
must necessarily have passed dorsal to the place where the proc-
ess later develops, for that vein lies directly ventral to the ner-
vus trigeminus. The relations of the arteria maxillaris interna
(carotis externa of fishes) to the process vary. In embryos of
the rabbit the artery perforates the process (Voit). In embryos
of the dog it is said by Olmstead (’11) to traverse a canalis alaris
s. alisphenoideum, which begins on the external surface of the
lamina ascendens of the ala temporalis and issues on its anterior
edge. The foramen rotundum opens into this canal, and the
second branch of the trigeminus, passing through this foramen,
enters the canalis alaris and, accompanying the arteria maxil-
laris interna, issues through its anterior opening into the orbit.
In Vespertilio the artery is said by Grosser (’01) to enter the
cranial cavity through the foramen ovale, then to run forward
ventral to the second branch of the trigeminus, and to issue from
the cranial cavity through an opening which corresponds to the

298 EDWARD PHELPS ALLIS, JR.

fissura orbitalis superior of man plus the foramina rotundum
and opticum. In the Macrochiroptera the artery traverses a
canalis pterygoideus in the basis cranii of this region (Grosser),
while in Rhinolopas it lies, asin man, wholly free along the lat-
eral wall of the cranium. ‘This is, then, wholly in accord with
the varying relations of this artery to the lateral wall of the cra-
nium in fishes, the artery traversing the pars jugularis of the
trigemino-facialis chamber in all of the Teleostei in which that
part of the chamber occurs, traversing a foramen in its lateral
wall in Amia, entering it with the nervus palatinus facialis in
Lepidosteus, and lying wholly external to the lateral wall of the
cranium in those fishes (Cottus, Amiurus) in which the pars jugu-
laris of the trigemino-facialis chamber is not enclosed. It is
thus evident that, both in the dog and in the Macrochiroptera,
the processus pterygoideus has fused with the lamina ascendens
of the ala temporalis and so has enclosed the external carotid in
a canal which corresponds to a part of the pars jugularis of a
trigemino-facialis chamber, and that, in Vespertilio, the mesial
wall of this canal has been resorbed, the artery then lying in a
part of a trigemino-facialis chamber.

The fovea epitympanica of rabbit embryos is a depression on
the lateral surface of the chondrocranium, said by Voit (’09, p.
450) to lie between the crista facialis and the tegmen tympani.
The tegmen tympani is said to arch over the upper edge of the
fovea, and it is so shown in his figures, the tegmen apparently
forming the dorsal portion of the lateral wall to the fovea. It is,
however, said (I. c., p. 449) that the tegmen is perforated by the
foramen faciale externum s. secundarium, but as that foramen
~ lies in the plane of the mesial wall of the fovea epitympanica,
it would seem as if there must be some error in the descriptions.
But however this may be, the fovea lodges the upper ends of the
malleus and incus, and these two cartilages lie external to the
nervus facialis, to the posttrigeminus portion of the vena capi-
tis lateralis, and to the arteria stapedialis (maxillaris interna,
carotis externa). The fovea and the space traversed by this
nerve, vein, and artery thus together form a cavity which has
the relations to the cranial wall of the pars jugularis of a tri-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 299

gemino-facialis chamber, the tympanic cavity of mammals thus
being a derivative of this chamber of fishes. The tegmen tym-
pani and the malleus, incus, and stapes are then quite cer-
tainly parts of the outer wall of this cavity, and hence derived
from the quadrate, this being as Driiner (04) has maintained
for the malleus, incus, and stapes. It would also seem as if the
annulus tympanicus must have the same origin, thus complet-
ing the outer wall of the cavity and encircling the part that was
broken up to form the auditory ossicles. The fact that the
stapes may be traversed by the arteria stapedialis is in ac-
cord with the perforation, in Amia, of the lateral wall of the
trigemino-facialis chamber by the external carotid.'

The tympanic cavity is traversed, in mammals, by the chorda
tympani, and Jacobson’s nerve and sympathetic fibers enter
it. In fishes the pars jugularis of the trigemino-facialis chamber
is traversed by a sympathetic nerve and frequently (always ?)
also by a communicating branch from the nervus facialis to the
nervus trigeminus, and Jacobson’s nerve enters it as a part of
the truncus facialis. The communicating branch from the ner-
vus facialis to the nervus trigeminus must then be the chorda
tympani, and that nerve must be a prespiracular one, for in
fishes it certainly is prespiracular. The chorda tympani must
then be represented, in fishes, in the ramus mandibularis inter-
nus trigemini of my descriptions of Amia (Allis, ’01, p. 188).

In fishes the spiracular canal or a diverticulum of it may lie
along the lateral wall of the trigemino-facialis chamber. If
a diverticulum of either of those canals were to expand into the
pars jugularis of the chamber, it would evidently give rise to a
tympanic cavity connected with the pharynx by an eustachian
tube, or the same result would be obtained by the expansion in-

‘Later work has somewhat modified this opinion and convinced me that the
incus, alone, corresponds to the lateral wall of the trigemine-facialis chamber
of fishes, both structures being derived from the posterior branchial-ray bar of
the mandibular arch. The malleus and the teleostean quadrate both represent
the epal element of the mandibular arch. The styloid and mastoid processes
are, respectively, the anterior and posterior branchial-ray bars of the hyal arch,
and the stapes probably the pharyngohyal. The chorda tympani is a posttre-
matic nerve.

300 EDWARD PHELPS ALLIS, JR.

to the chamber of a diverticulum of a plica hyomandibularis
(Driiner, ’03).

The conditions in Echidna remain to be considered. In an
earlier work (Allis, 14 b) I came to the conclusion that the cavum
epiptericum of Gaupp’s descriptions of embryos of this animal
was the strict equivalent of the trigemino-facialis chamber of
Amia less its pars facialis, this conclusion being based on my
interpretation of Gaupp’s descriptions of the venous vessels of
the region. According to him (’08, p. 598), there is, in the
cavum cerebrale cranii of this animal, a large cross-commis-
sural venous vessel, anterior to the hypophysis and issuing on
either side through the fenestra pseudo-optica into the cavum
epiptericum. There, one part of this vessel turns forward and
passes into the orbit, the other turning posteriorly in the cavum
epiptericum and becoming the sinus cavernosus. This so-
called sinus cavernosus is said to pass ventral to the ganglion
trigeminum, and it is shown, in a figure of a transverse section
through this region, lying ventrolateral to the base of the taenia
clino-orbitalis, the hypophysis lying mesial to the taenia. Pos-
terior to this point, and hence apparently posterior to the sella
turcica, the sinus cavernosus turns laterally and falls into the
sinus transversus, the latter sinus descending almost vertically
in front of the otic capsule. The fusion of these two veins is
said to form the vena capitis lateralis, which issues from the
cranial cavity through the hindermost corner of the fenestra
sphenoparietalis and immediately enters the sulcus facialis on
the external surface of the chondrocranium. In a slightly older
embryo the sinus cavernosus is said (l.c., p. 629) still to be
connected with the sinus transversus, but to be now also pro-
longed posteriorly as the sinus petrobasilaris, which runs pos-
teriorly in the cavum cerebrale cranii, sends a branch outward
through the foramen jugulare, and then itself issues through the
foramen occipitale magnum.

From these descriptions I concluded (’14 b) that the so-called
sinus cavernosus, plus the vena capitis lateralis, must form a vein
the homologue of the vena jugularis of fishes. That vein could
not then enter the cavum cerebrale cranii, as Gaupp says it does,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 301

and I concluded that there must be some error in the descrip-
tions, for I did not question the identification of the veins. Be-
cause of the position of this vein, I concluded that the cavum
epiptericum was a trigemino-facialis chamber. ‘This is, how-
ever, wrong, for the so-called sinus cavernosus is, in reality, the
homologue of the pituitary vein of fishes, and not of the vena
jugularis. This vein of Echidna must then traverse a myo-
domic cavity, as it does in the Mammalia ditremata, and there
must be a membrane separating it from the cavum epiptericum,
that membrane being a part of the membrana limitans of Terry’s
(17) descriptions of the cat and forming the roof of a myodomic
cavity which is the sinus cavernosus properly so-called. The
pituitary vein then traverses this cavity, as it does in man, and
that part of the so-called sinus which Gaupp says turns later-
ally and falls into the sinus transversus, is the vena encephalica
of fishes, this latter vein falling into the vena capitis lateralis
(vena jugularis of fishes) after and not before, it issues from the
cavum epiptericum. The vena capitis lateralis has here, as
in the Mammalia ditremata, lost its primitive continuity with
the vena capitis media; the persisting portions of these veins
both lie external to the cranial wall; and the cavum epiptericum
is a trigemino-facialis recess. The conditions in this animal
are then stricty similar to those in the Mammalia ditremata
except that a taenia clino-orbitalis has been formed, compar-
able to, but somewhat different from, the cartilage ‘c’ of Voit’s
descriptions of the rabbit.

In the adult Echidna it would seem, from Gaupp’s descrip-
tions, as if certain of the bones forming the lateral wall of the
cranium were developed in the lateral wall of the caxum epip-
tericum (trigeminus recess), and certain of them in the lateral
wall of the pars jugularis of a trigemino-facialis chamber, for
certain of the bones are said (I. c., p. 650) to be ossifications of
the membrana spheno-obturatoria, which is said to lie external
to the ala temporalis. The taenia clino-orbitalis is said by
Gaupp (l. c., p. 647) to have fused, in the adult, with the lateral
edge of the sella turcica along the full length of the sella, the
fissura pseudo-optica thus being greatly reduced in size; the

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

302 EDWARD PHELPS ALLIS, JR.

development of this bar of cartilage doubtless accounting for
the suppression of a posthypophysial commissure between the
pituitary veins of opposite sides of the head. |

The so-called parasphenoid of Echidna is considered by Gaupp
(05 a) to be the homologue of the ascending process of the
parasphenoid of the Sauria, and also of the mammalian ptery-
goid, the latter bone not being the homologue of the pterygoid
of reptiles. The bone of Echidna is said to lie, in embryos,
directly upon the cartilage of the basis cranii, without inter-
vening connective tissue, and later to fuse with the sphenoid
(Keilbein) as part of its processus pterygoideus. No cartilage
has been found in this bone in Echidna, but it is said to be found
in the pterygoid of mammals. The bone lies anterior to the
foramen caroticum, the internal carotid arteries accordingly
not coming into any relations to it. The nervus parabasalis
(palatinus facialis) is said to run forward external to the pos-
terior portion of the bone, but, anterior to the point of exit of
the nervus opticus from the cranial cavity, it perforates the
bone through a foramen parabasale, and so enters the anterior
portion of the cavum epiptericum. There is thus no canalis
parabasalis in this animal, and the relations of the parasphenoid
to the chondrocranium, to the internal carotid arteries, and to
the ramus palatinus facialis all show that it corresponds to the
ascending process of the parasphenoid of Amiurus, and to the
mesial plate of that process of the parasphenoid of Polyp-
terus, and that it is accordingly an ossification in the roof of a
ventral myodomie cavity and not in its floor.

CARTILAGINES POLARIS AND ACROCHORDALIS

Polar cartilages were, as already stated when describing the
Selachii, first described by van Wijhe (’05) in embryos of Acan-
thias, where the cartilage of either side is said by him to lie be-
tween the trabecular and parachordal cartilages, but it soon
fuses with both those cartilages and then forms, with the tra-
becula, the ventral border of the orbital fenestra. The pos-
terior border of the orbital fenestra is said to be formed by the
lamina antotica, which is an outgrowth of the anterior end of

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 303

the parachordal of its side. Nothing is said of the relations of
the eye-muscles, arteries, and veins to these cartilages.
Sewertzoff (’97, ’99), in his work on embryos of this same fish
(Acanthias), did not find these cartilages, and he says that the
alisphenoid, which is van Wijhe’s antotica, is primarily a wholly ,
independent cartilage and hence not an outgrowth of the par-
achordal. He considers it to be a prechordal structure, and
says that it is apparently developed in close relations to the eye-
muscles, the four rectus muscles and the obliquus superior all
having their insertions on it. The only other fish in which this
cartilage has been described is, so far as I know, Lepidosteus,
where it has been described by Veit and has been already re-
ferred to when considering that fish. The cartilage is there said
to give insertion to the rectus externus, this cartilage of this fish
thus apparently corresponding, functionally, to the base of the
alisphenoid cartilage of Sewertzoff’s descriptions of Acanthias,
as it does also to the eye stalk of the adult selachian. A polar
cartilage, although only described in these two fishes, has been
recognized and described in certain of the Sauropsida and
Mammalia.

In 5-mm. chick embryos and 8 to 9-mm. embryos of the duck, |
Sonies (’07) finds no cartilage as yet developed in the cranial
region. The notochord is said to extend far up in the plica
encephali ventralis, and its tip is there bent slightly ventrally
and is lost in connective tissues behind the hypophysis. In
slightly older stages, an unpaired cartilage, the cartilago acro-
chordalis, develops around the anterior end of the notochord,
the cartilage inclining dorso-anteriorly and the notochord per-
forating it from its dorsal surface. The parachordals are said
to then develop, posterior to this acrochordalis cartilage, as a
simple unpaired median plate, for, although always thickest
along their lateral edges, they are always continuous with each
other dorsal to the notochord and, in most instances, also con-
tinuous ventral to it. These two primarily independent and
unpaired cartilaginous plates, the acrochordalis and parachor-
dalis, then become connected with each other, on either side,
by a short cartilage which is called the cartilago basiotica, these

~

304 EDWARD PHELPS ALLIS, JR.

two paired cartilages developing as independent pieces in the
duck, but in the chick in continuity with the anterior edge of
the parachordal plate. The basal plate is thus completed, and
it is perforated by a median space, traversed longitudinally
by the notochord, which is said to be the fenestra basicranialis
posterior and which has the position of that fenestra in Lacerta
and the Amphibia.

The trabeculae appear as independent paired cartilages at
about the same time as the cartilagines basioticae, lying ros-
tral to the nervi optici and nearly at right angles to the basal
plate. An independent cartilago polaris then develops, in the
duck, on either side of the hypophysis, between the trabeculae
and the ventral surface of the basal plate, and later fuses with
both of those cartilages, usually first with the trabeculae, but
occasionally first with the basal plate. In the chick the carti-
lago polaris is, from the very first, continuous with the hinder
end of the trabecula of its side. The fusion of the polar carti-
lages with the basal plate takes place in the line of the fusion of
the cartilagines acrochordalis and basioticae, and a fenestra
hypophyseos is thus enclosed, which lies nearly at a right angle
to the fenestra basicranialis posterior and is separated from it
by the cartilago acrochordalis (I. c., p. 426). The side walls of
this fenestra hypophyseos are at first formed both by the polar
cartilages and the hinder ends of the trabeculae, but, as the tra-
beculae gradually fuse with each other in the median line, that
part of the fenestra which was primarily enclosed between them
is gradually suppressed, the trabeculae then only forming its
anterior wall, the tuberculum sellae. The cartilago acrochor-
dalis, projecting dorso-anteriorly, is said to form the dorsum
sellae. A processus infrapolaris develops later on either side,
from the posteroventral surface of the polar cartilage, and in
Sterna projects posteriorly beneath and parallel to the basal
plate, its hind end fusing with it on either side of the fenestra
basicranialis posterior. A somewhat vertical, subparachor-
dal plate is thus formed which is perforated by a large opening,
traversed by the arteria carotis interna. That artery, after
traversing this opening, passes through the fenestra hypo-

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 305

physeos, posterior to the hypophysis, sends a cross-commissural
branch to its fellow of the opposite side, another branch, the
arteria opthalmica interna, outward dorso-anterior to the polar
cartilage, and is then itself distributed to the brain. It passes
mesial to the polar cartilage of its side, but lateral and dorsal
to the processus infrapolaris. If this latter process were to be-
come the only connection between the polar cartilage and the
basal plate, the artery would pass lateral to the polar cartilage,
and this is apparently what actually takes place in the Mam-
malia, as will be explained later.

Comparing these conditions in the chick and duck with those
in embryos of the Teleostei and Holostei, it is at once evident
that the cartilago acrochordalis of the former must be the homo-
logue of that cartilaginous prootic bridge of the latter which
forms the beginning of the definitive prootic bridge. The re-
lations of these two cartilages to the other skeletal elements,
and to the brain, are too strictly similar to leave any reasonable
doubt as to this, the differing relations of the cartilages to the
notochord evidently being related to the early development of
the cartilage in the chick and duck and its late development in
the Teleostei and Holostei. The space which, in the Teleostei
and Holostei, lies between this bridge and the otic portion of the
basal plate must then be the homologue of the fenestra basi-
cranialis posterior of the chick and duck, as has already been
stated, and the side walls of this fenestra the homologues of
the basiotic cartilages; these latter cartilages being prolonged
ventrally, in fishes, by the ventral processes of the prootics,
and, in the chick and duck, by the infrapolar processes. These
latter processes, together with the polar cartilages, are then the
so-called anterior prolongations of the parachordals of Swinner-
ton’s and Gaupp’s descriptions of Gasterosteus and Salmo, there
apparently developed in continuity with the basiotic cartilages,
as they are said to be in certain of the Aves. The so-called
fenestra basicranialis posterior, or fenetra interparachordalis,
of Gaupp’s and Swinnerton’s descriptions of fishes is then the
homologue of the fenestra hypophyseos of the chick and duck
and not of the fenestra basicranialis posterior, and the fenes-

306 EDWARD PHELPS ALLIS, JR.

tra hypophyseos of embryos of fishes has been suppressed in
advanced embryos of the chick and duck.

In Talpa, Noordenbos (’05) finds the parachordals of oppo-
site sides, when first developed, united with each other, ventral
to the notochord, and not extending to its tip. The tip of the
notochord reaches, at this stage, to the hypophysis, and is said
to represent, in a certain sense, the morphological anterior end,
or anterior pole, of the embryo, the hypophysis being an organ
at that pole. In slightly older embryos the notochord is some-
what withdrawn from the hypophysis, and its tip then doubt-
less lies posterior to the infundibulum. The parachordal plate
has at the same time grown rostralward, and, turning upward
at its anterior end, now surrounds the notochord, which tray-
erses it from its dorsal to its ventral surface and extends
anteriorly beyond it.

The trabeculae first appear as a single median plate between
the nasal sacs and extending posteriorly to the recessus preop-
ticus. In the space between the trabecular and parachordal
plates, ventral to the hypophysis and at a slightly lower level
than the parachordal plate, two pairs of little cartilages, the
insulae polares, later appear, and soon fuse to form a polar plate
which is at first perforated by a median fenestra hypophyseos,
which soon becomes closed by growth of the bounding cartilage.
This polar plate fuses, soon after its formation, with the tra-
becular plate, and in the line of fusion a slight transverse fur-
row is formed which lodges the chiasma opticum. The hind
edge of this furrow is slightly raised, and forms the tuberculum
sellae, which thus lies on the anterior end of the polar plate and
not, as in the chick and duck, on the hind ends of the trabec-
ulae. No cartilago acrochordalis has yet been formed, and the
polar plate accordingly cannot fuse with the basal plate along
the line of fusion of that cartilage with the cartilagines basi-
oticae, as it does in the chick and duck. Accordingly, a direct
fusion of the polar plate with the basal plate does not take
place, and connection with the latter plate is acquired through
the intermediation of a delicate Y-shaped mass of cartilage, the
arms of which fuse with the projecting anterior ends of the

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 307

parachordal plate. This Y-shaped cartilage, probably together
with the posterior pair of insulae polares, thus corresponds to
the infrapolar processes of the chick and duck, and the internal
carotid arteries run upward lateral to them, as they do in the
chick and duck and as explained just above.

The Y-shaped cartilage of Talpa, by its fusion with the
anterior end of the parachordal plate, encloses a circular open-
ing which Noordenbos calls the fenestra basicranialis posterior.
The anterior end of the notochord les directly above this fenes-
tra, which it would not do were the fenestra the homologue of
the similarly named fenestra in the chick and duck. Further-
more, Noordenbos says (’05, p. 385) that the hypophysis lies
in a slight fossa, bounded anteriorly by the tuberculum sellae
(a ridge formed, as above stated, on the anterior end of the polar
plate) and posteriorly by the anterior end of the parachordal
plate, thus necessarily lying directly above the so-called fenes-.
tra basicranialis posterior, instead of, as in the chick and duck,
definitely anterior to it. This fenestra of Talpa must then be
ah opening corresponding to some part of Gaupp’s fenestra
basicranialis posterior of Salmo, and apparently to that part
of it which he says leads from the middle into the posterior
sections of his descriptions of the myodome. The fenestra of
Talpa is, in any event, not a perforation of the floor of the cavum
cerebrale cranii, as it is in the chick and duck, and that per-
foration, and a cartilago acrochordalis are both wanting in these
embryos. The dorsum sellae of these early embryos is then
not the homologue of the posteclinoid wall of Amia and the Tel-
eostei, nor of the dorsum sellae of Sonies’s descriptions of the
chick and duck. It is, however, possible that a cartilago acro-
chordalis may be developed in later stages than those described
by Noordenbos, for Voit shows this cartilage in his figures of
embryos of the rabbit, there perforated by an opening, the evi-
dent homologue of the fenestra basicranialis posterior of Sonies’s
descriptions of the chick and duck; and Faweett (’10), in a work
I have not been able to consult, is said by Kernan (16, p. 621)
to have found the dorsum sellae separated from the crista trans-
versa in 19-mm. and 21-mm. human embryos, the dorsum sellae

308 EDWARD PHELPS ALLIS, JR.

then there representing the cartilago acrochordalis, and the crista
transversa representing the anterior end of the parachordal
plate.

Sonies calls attention (’07, p. 406) to the unusual position
of the fenestra basicranialis posterior in Talpa, and suggests
that the posterior pair of insulae polares correspond to the car-
tilago acrochordalis of his own descriptions of the chick and duck.
Terry (’17) says that this fenestra lies, in embryos of the cat,
between the anterior end of the parachordal plate and the
cartilago polaris (hypophyseal cartilage), thus agreeing with
Noordenbos in his identification of it, and he says that it lies
‘not within the basal (parachordal) plate, but anterior to it,
as Noordenbos insists.”’ In his figure of a median vertical
section of a 12-mm. embryo (l. c., fig. 17) he, however, shows
it lying definitely beneath the turned up anterior end of the
parachordal plate, in exactly the position I have assigned to it.

Polar cartilages, lying between the trabeculae and parachor-
dals, have thus been identified in Acanthias and Lepisdosteus
among fishes, and in several of the Sauropsida and Mammalia,
and it is probable that they form an integral element of the
cranium in all of the Gnathostomata, though probably not
always developed as wholly independent cartilages. The two
cartilages embrace the ectodermal stalk of the hypophysis, the
openang between them thus being the fenestra hypophyseos proper-
ly so-called, but this fenestra is continued both anteriorly and pos-
teriorly, at certain stages of development, in most of the Gna-
thostomata. The anterior prolongation of it les between the
hind ends of the trabeculae, and although it is generally con-
sidered to persist, in the Teleostei and Holostei, as part of the
fenestra ventralis myodomus of the adult, it is probable that
it becomes largely, if not entirely, suppressed by fusion of the
trabeculae. The posterior prolongation of it lies, in fishes,
between the ventral edges of the ventral processes of the prootics,
and, in the Aves and Mammalia, between the corresponding
edges of the infrapolar processes. These processes must then
be homologous structures, and if the ventral processes of the
prootics of fishes are ventrolateral processes, as I conclude,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 309

the intrapolar processes, and hence also the polar cartilages,
must be hypochordal and not parachordal structures. This,
then, is in accord with, but an extension of, Terry’s conclusion
(17, pp. 344 and 396) that there seems no doubt of the presence
of a cartilaginous hypochordal layer in the occipital region of
mammals generally, and that the basal plate of the occipital
region falls into the category of arch structures, not centra.

The notochord and trabeculae may now be considered. Swin-
nerton says that the notochord in Gasterosteus undergoes no
actual reduction from the earliest to the latest stages examined
by him, its relatively less extensive anterior prolongation in older
embryos being wholly due to an anterior prolongation of the
parachordal cartilages. Froriep, however, says (’02 a, ’02 b)
that in Torpedo there is an actual disintegration of the anterior
portion of the notochord. According to him, in early embryos
of that fish, the notochord is separated into two definite regions,
one of which he considers to be spinal and the other prespinal.
The spinal region is said to begin at the dorsorostral corner of
the first visceral pouch, this point coinciding with that in which
the dorsal wall of the foregut, in early embryos, bends abruptly
ventrally in an obtuse angle, the notochord there also bending
ventrally at the same angle. Posterior to this point the noto-
chord develops a cuticular -sheath immediately after its con-
striction from the dorsal wall of the foregut, and is persistent
throughout the life of the individual. Anterior to this point,
and hence in the prespinal region, the notochord presents two
different conditions, one related to the region in which the man-
dibular head cavities develop and the other to that in which
the premandibular cavities develop. In the mandibular region
a chorda entoblast is said to be constricted from the dorsal
wall of the foregut exactly as in the spinal region, but it does
not undergo further differentiation and later entirely disinte-
grates. In the premandibular region, according to Froriep
(792\b, p. 55):

Kommt es nicht einmal zur Bildung einer primitiven Chordaan-
lage, sondern deren Bildungsmaterial sowohl wie dasjenige des Meso-
blasts bleibt ungesondert in der Wand des Vorderdarms enthalten.

310 EDWARD PHELPS ALLIS, JR.

Diese indifferente Urdarmmasse schniirt sich zu Ende des Stadium F
von den Gebilden der Mandibularregion vollstaéndig ab und stellt nun
die Anlage der Priimandibularen Kopfhohle Balfour’s oder dasI. Somit
van Wijhe’s dar.

The protovertebrae are said by Froriep to extend the full
length of the persisting notochord, and not to extend beyond
that point; the whole animal being, at this stage, vertebral
column. The prespinal, or head region is said to contain the
matrix in which all the visceral arches and the mandibular and
premandibular head cavities are developed.

Katherine M. Parker, in the latest work I know of relating
to this subject, also finds, in the Marsupiala, the notochordal
tissue extending anteriorly beyond the end of the persisting
notochord, for she says (’17, p. 24):

The primitive relation of the tip of the notochord is one of continuity
with the protochordal plate, and in Perameles continuity is retained
between the chorda and the derivatives of the protochordal plate (pre-
chordal plate and Seesel’s pocket). As a secondary condition, con-
tinuity may be established between the chorda and the hypophysis.

His, in a much earlier work, also came to a similar conclusion,
for he says (’92, p. 348) that the notochord, in all early vertebrate
embryos, ends anteriorly in a tapering point which les imme-
diately posterior to a transverse basal ridge (Basilarleiste) of
the brain which lies at the extreme anterior end of the ventral
surface of the neural tube. This basal ridge is in contact, either
with the dorsal end of Seesel’s pocket or with a strip of entoderm
(Entodermstreife) which replaces that pocket, and His shows
the tip of the notochord wedged in between his basal ridge and
Seesel’s pocket in two different figures, one said to be a general
vertebrate schema and the other to show an actual median
sagittal section of the head of an embryo of Pristiurus 26-mm.
in length. Seesel’s pocket lies at the dorsal edge of the oral
plate, and is said to be not only topographically, but also genet-
ically, an anterior continuation of the notochord (J. ¢., p. 350),
the notochordal tissue thus extending to the level of the anterior
end of the ventral surface of the neural tube. This primitive
topographical relation of these four structures, the basal ridge,

MYODOME AND TRIGEMINO-FACIALIS CHAMBER oul

the tip of the notochord, Seesel’s pocket, and the dorsal edge
of the oral plate, is said to be subject to marked changes in
later stages and in different vertebrates, Seesel’s pocket shifting
either anteriorly or ventrally (rachenwirts) relatively to the
basal ridge. In the latter case an ectodermal fold is formed
between it and the basal ridge, and becomes the hypophysial
invagination (Rathke’s pocket), which extends posteriorly be-
yond the basal ridge, forcing the tip of the notochord away
from the ventral surface of the brain, and even forcing it upward
into the plica encephali ventralis. The relations of the brain
to the notochord, in the adult vertebrate, are accordingly said
by His not always to be the primitive ones, and he (l.c., p. 358)
considers only those parts of the brain of the adult to be pre-
chordal which lie anterior to the basal ridge, and which therefore
formed primarily a part of the anterior surface of the neural tube.
Those parts are said by him to be the regions of the recessus infun-
dibuli, the chiasma opticum, the recessus opticus, the lamina ter-
minalis, and the olfactory lobes. The saccus vasculosus lies
posterior to the basal ridge and belongs morphologically, as well as
actually, to the ventral surface of the brain, the line between the
morphologically ventral and anterior surfaces of the brain thus
lying between the saccus and the recessus infundibuli.

There is thus reason to believe that the notochord extended
primarily to the level of the anterior end of the primitive gut,
and that, accordingly, the epichordal and hypochordal bands
of skeletogenous material, developed in relation to it, had a
similar extent. The polar and trabecular cartilages must then
be developed from some part of these anterior extensions of
these bands, and the polar cartilages quite certainly, as already
stated, from the hypochordal bands alone. The trabeculae,
in crania of the platybasic type, would seem to be developed
from both these bands of tissue. In crania of the tropibasic
type the two bands seem to have been forced apart, by pressure
of the eyeballs, the epichordal bands lying at the top of the
interorbital septum and the hypochordal bands at the bottom
of that septum.

It is furthermore to be noted that the trabeculae do not lie

312 EDWARD PHELPS ALLIS, JR.

along the ventral surface of the brain, as that surface is defined by
His; since as, when first formed, their hind ends apparently always
lie anterior to the recessus infundibuli, they must themselves lie
either definitely on the anterior surface of the brain or along
the lateral surface of its extreme anterior end. In the latter
case they would actually have, to the neural tube, the relations
of dorsal vertebral arches. It does not, however, necessarily
follow that they are such arches, for their relations to the brain
may be wholly due to a cranial flexure so sharp and pronounced
that it has turned the anterior surface of the neural tube down-
ward upon cartilages which primarily lay either in the line of
the axis of the body, or projected ventrally beneath it.

SUMMARY

A functional myodome is found only in fishes, and even among
them it is limited, in-those I have examined, to Amia and the
non-siluroid Teleostei.

The myodome is always separated from the cavum cerebrale
cranii by membrane (dura mater), cartilage, or bone, and the
separating wall is in part spinal and in part prespinal in position.
A depression in the prespinal portion lodges the hypophysis
or both the hypophysis and saccus vasculosus, and this part of
the wall never undergoes either chondrification or ossification,
a more or less developed pituitary sac always projecting into
the myodome.

The myodome is found in its most complete form in the Tel-
eostei, and there consists of dorsal and ventral compartments
which are usually separated from each other only by membrane,
but that membrane, the horizontal myodomic membrane, is
capable of either chondrification or ossification. The dorsal com-
partment lodges the hind ends of the musculi recti externi and
is always traversed by a cross-c ommissural venous vessel formed
by the pituitary veins. The ventral compartment lodges the
hind ends of the musculi recti interni and is traversed by the
internal carotid and efferent pseudobranchial arteries and the
palatine branches of the facialis nerves.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 313

The parasphenoid forms the floor of the ventral compartment ,
of the myodome, and whenever the horizontal myodomic mem-
brane undergoes ossification, the bone so formed forms part of
the parasphenoid. This bone is thus certainly, in some fishes,
in part of axial origin, and not simply a dermal bone which has
gradually sunk inward to its actual position.

In the Siluridae (Amiurus) there is apparently a much reduced,
but non-functional, dorsal myodomic compartment, but no
ventral compartment, that portion of the parasphenoid which
lies in the prootic region being developed in what corresponds to
the horizontal myodomic membrane of others of the Teleostei.

In Amia the myodome corresponds to the dorsal compart-
ment only of the teleostean myodome, and a strictly similar,
but non-functional myodomic cavity is found in Lepidosteus
and Polypterus. The ventral compartment of the teleostean
myodome is represented, in each of these three fishes, by a canal,
on either side of the head, which is traversed by the internal
carotid artery, and which corresponds to the canalis parabasalis
of Gaupp’s descriptions of higher vertebrates.

The myodomiec cavity is limited, in the Holostei and Cros-
sopterygii, to the prootic region, and is there in part subspinal
and in part prespinal and subpituitary in position. In the non-
siluroid Teleostei examined, the dorsal compartment of the
myodome is always more or less prolonged posteriorly into the
basioccipital region and the ventral compartment frequently
so prolonged.

The posterior part of the basioccipital portion of the myodome
lies between ventrolateral vertebral processes which are quite
certainly the homologues of the haemal arches of the tail. In
Hyodon this part of the myodome is an open groove and lodges
the anterior portion of the median dorsal aorta. In the Cy-
prinidae part of this groove has become enclosed to form a short
canal which is traversed by the median dorsal aorta, the enclos-
ing bone forming the pharyngeal process.

The conditions in these fishes thus lead inevitably to the
assumption that the entire dorsal myodomic cavity is a sub-
vertebral canal similar to the haemal canal in the tail, and that

314 EDWARD PHELPS ALLIS, JR.

the dorsal aorta has been excluded from it because of the for-
mation of a circulus cephalicus. What the primary relations
of the hypophysis and pituitary veins to this preexisting canal
were is problematical, but they became lodged in its anterior
portion and so gave rise to the conditions actually found in
Lepidosteus and Polypterus. The musculi recti externi then
secondarily invaded this space by traversing the foramina for
the pituitary veins, the other rectus muscles retaining their in-
sertions on the external surface of the preclinoid wall, and so
gave rise to the conditions found in Amia. The conditions in
the non-siluroid Teleostei then arose as a result of the resorption
of the cartilage which, in Amia, forms the preclinoid wall, the
pedicel of the alisphenoid, and those ventral portions of the
basicapsular commissures which form the lateral walls of the
subpituitary portion of the myodome. Because of the resorption
of the preclinoid wall, and its replacement by membrane, the
musculi recti interni, which in Amia have their points of in-
sertion on either lateral edge of that wall, have first sought firmer
attachment on the dorsal surface of the parasphenoid, and have
later pushed posteriorly in the open ends of the persisting por-
tions of the canales parabasales. The fusion of these two canals
with each other has formed a ventral myodomic compartment
which, in early embryos, is separated from the dorsal and pri-
mary compartment by membrane only; but this membrane may
undergo either partial chondrification (Hyodon) or ossification
(Gasterosteus), the bone, in the latter case, forming a transverse
and inclined ridge on the dorsal surface of the parasphenoid.
The membranes resulting from the resorption of the preclinoid
wall were then pressed together in the median line by the recti
interni, and form a median vertical myodomic membrane which
encloses the internal carotid arteries in a membranous canal,
the homologue of the cartilaginous canals of Amia. The efferent
pseudobranchial arteries, pressed downward by the recti interni,
lost their connections with the internal carotids and acquired a
cross-commissural connection with each other. The membrane
resulting from the resorption of the anterior portions of the
basicapsular commissures of either side ossified as part of the

MYODOME AND TRIGEMINO-FACIALIS CHAMBER ey

ascending process of the parasphenoid, and the tissues resulting
from the resorption of the pedicel of the alisphenoid ossified, in
certain fishes (Cottus, Gasterosteus), to form an anterior portion
of that process. i

The myodomie cavities of the Holostei and Teleostei are rep-
resented in the Selachii either by canals in the basis cranii which
are traversed by the pituitary veins and the internal carotid and
efferent pseudobranchial arteries or by a posterior and deeper
portion of the large pituitary fossa of the chondrocranium which
is shut off from the cavum cerebrale cranii by the dura mater,
and is traversed by the pituitary veins and the internal carotid
arteries.

In embryos of Ceratodus there is a subpituitary space, trav-
ersed by the pituitary veins, which corresponds to the dorsal com-
partment of the teleostean myodome, and the internal carotid
canals of Amia have been added tosit. This fusion of these
canals with the dorsal myodomic cavity is due, either to the
resorption of the cartilage that separates them in Amia or to a
shifting posteriorly of both the hypophysis and the internal
earotids from a position between the hind ends of the trabeculae
to one between the so-called anterior prolongations of the para-
chordals.

In the Amphibia the basis cranii apparently corresponds to
the roof, and not to the floor, of the dorsal myodomie cavity of
Amia and the Teleostei. The fenestra hypophyseos of these
animals is then the homologue of the pituitary opening of the
brain case of fishes.

The Reptilia and Mammalia have a dorsal myodomic cavity
similar to that in Ceratodus. In man it is represented in the
cavernous and intercavernous sinuses, and the venous vessels
that traverse the sinuses are the homologues of the pituitary
veins of fishes.

The cartilago acrochordalis of Sonies’ and Noordenbos’ de-
scriptions of birds and mammals, respectively, is the homologue
of the cartilaginous prootie bridge of embryos of fishes. The
open space between this cartilage, or bridge, and the anterior
end of the parachordal plate is the fenestra basicranialis poste-

316 EDWARD PHELPS ALLIS, JR.

rior proper. This fenestra is a perforation of the roof of the
myodomic cavity, and hence is not the homologue of the so-
called fenestra basicranialis posterior of embryos of fishes, which
is a perforation of the floor of that cavity. This latter fenestra
of embryos of fishes is the homologue of the fenestra hypophyseos
of birds and mammals, the so-called anterior prolongations of
the parachordals of fishes being the homologues of the polar
cartilages of birds and mammals.

In certain of the Selachii there is an acustico-trigemino-faci-
alis recess, and there may be certain canals in the cranial wall
traversed by the vena jugularis and the external carotid artery.

In Amia the trigemino-facialis portion of this recess has fused
with the canals for the vena jugularis and the external carotid
artery to form a trigemino-facialis chamber; this chamber has
become continuous with the myodome, and the large chamber
so formed has been prolonged anteriorly by a space between the
pedicel of the alisphenoid and the primitive side wall of the
neurocranium. The foramina for the pituitary vein and the
oculomotor and trochlear nerves open into this anterior pro-
longation of the chamber, and through its orbital opening into
the orbit. The vena jugularis traverses this opening to enter
and traverse the trigemino-facialis chamber; the musculus rectus
externus traverses it to enter the myodome, and the nervus pro-
fundus traverses it to join the ganglion, or root of the nervus
trigeminus. The nervus trigeminus and the external carotid
artery issue from the trigemino-facialis chamber posterior to
the pedicel of the alisphenoid and run forward lateral to it.

In the non-siluroid Teleostei the trigemino-facialis chamber
is not continuous with the myodome, and it has been separated
by a wall of bone into ganglionaris and jugularis parts which
correspond, respectively, to the trigemino-facialis recess and the
jugular and external carotid canals of the Selachii. The pedicel
of the alisphenoid is incomplete or wholly wanting, but it may
be replaced by an anterior prolongation of the ascending process
of the parasphenoid. In the latter case the nerves, arteries,
veins, and muscles all have the same relations to this process
that they have to the pedicel of the alisphenoid of Amia. ‘The

MYODOME AND TRIGEMINO-FACIALIS CHAMBER Siz

lateral wall of the pars jugularis of the trigemino-facialis chamber
is always less extensive than in Amia and may be wholly wanting.

In Ceratodus there is a trigemino-facialis chamber similar to
that in Amia, and there is a bar of cartilage which corresponds
to the pedicel of the alisphenoid of that fish.

In the Amphibia there is a trigemino-facialis recess, and the
pars ascendens of the quadrate forms the lateral wall of a space
corresponding to the pars jugularis of the chamber of the Teleos-
tei. The ascending process of the palatoquadrate is the homo-
logue of the pedicel of the alisphenoid of fishes.

In the Reptilia there apparently is no trigemino-facialis recess,
the lateral wall of the neurocranium being the primitive cranial
wall. The pars ascendens of the quadrate forms the lateral
wall of a trigemino-facialis chamber. The antipterygoid (col-
umella) is the homologue of the pedicel of the alisphenoid of fishes,
and the processus basipterygoideus the homologue of the floor
of the orbital opening of the myodome of Amia.

In the Mammalia there is a trigemino-facialis recess formed
by the cava epiptericum and supracochleare. The ala tem-
poralis is peculiar to mammals; it is a bar of cartilage formed be-
tween the nervi maxillaris and mandibularis trigemini as they
issue from the trigemino-facialis recess, the processus alaris cor-
responding to some part of the side wall of the prespinal portion
of the myodome of Amia. The ala temporalis has been pro-
longed anteriorly so as to enclose a space anterior to the tri-
gemino-facialis recess, and the foramina for the pituitary vein
(sinus cavernosus) and the nervi oculomotorius, trochlearis and
profundus (first branch of trigeminus) open into this space and.
from it into the orbit. The cavum tympanicum is the pars
jugularis of the trigemino-facialis chamber, and the processus
pterygoideus, the malleus, incus, and stapes, and probably also
the annulus tympanicus, are quite certainly portions of the lat-
eral wall of that part of the chamber. A diverticulum of the
spiracular canal, or an independent evagination of the pharynx,
has expanded into this part of the chamber and so formed the
middle ear. The chorda tympani must then correspond to that
communicating branch from the nervus facialis to the nervus

THE JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

318 EDWARD PHELPS ALLIS, JR.

trigeminus which, in fishes, traverses the trigemino-facialis cham-
ber, and hence must be a prespiracular nerve.

The internal carotid artery enters the cranial cavity, in most
vertebrates, by passing upward mesial to the related trabecula,
or mesial to the posterior prolongation of the trabecula formed
by the polar cartilage, but in Amiurus it enters the cranial cavity
through the foramen opticum, and hence would there seem to
pass lateral and then dorsal to the trabecula. In embryos of
the Mammalia ditremata this artery is said to also pass upward
lateral to the trabecula, but it is probable that it here simply
passes lateral to the infrapolar process of the polar cartilage, the
latter cartilage not itself fusing directly with the parachordal
plate, and its direct relations to the artery thus being obscured.

Palais de Carnolés, Menton, France,
May 1, 1918

LITERATURE CITED

ALLEN, W.F. 1905 The blood-vascular system of the Loricati, the mail-cheeked
fishes. Proc. Washington Acad. Sci., vol. 7.

Auuis, E. P., Jr. 1897 ‘The cranial muscles and cranial and first spinal nerves
in Amia calva. Jour. Morph., vol. 12.
1901 The lateral sensory canals, the eye-muscles, and the periph-
eral distribution of certain of the cranial nerves of Mustelus laevis.
Quart. Jour. Microsc. Sci., vol. 45.
1903 The skull, and the cranial and first spinal muscles and nerves
in Scomber scomber. Jour. Morph., vol. 18.
1908 a The pseudobranchial and carotid arteries in Polypterus.
Anat wAnZ Bassas:
1908 b The pseudobranchial and carotid arteries in Ameiurus. Anat.

Anz., Bd. 33.
1909 The cranial anatomy of the mail-cheeked fishes. Zoologica,
Hft. 57.

1911 The pseudobranchial and carotid arteries in Polyodon spathula.
Anat. Anz., Bd. 39.

1913 The homologies of the ethmoidal region of the selachian skull.
Anat. Anz., Bd. 44.

1914a Certain homologies of the palato-quadrate of Selachians.
Anat. Anz., Bd. 45.

19146 The pituitary fossa and trigemino-facialis chamber in Se-
lachians. Anat. Anz., Bd. 46.

1914 ¢ The pituitary fossa and trigemino-facialis chamber in Cera-
todus forstri. Anat. Anz., Bd. 46.

1914d The trigemino-facialis chamber in A nphibians and Reptiles.
Anat. Anz., Bd. 47.

€

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 319

ArAl, H. 1907 Der Inhalt des Canalis cranio-pharyngeus. Anat. Hefte, Bd.
3a.

BAUMGARTNER, HE. A. 1915 The development of the hypophysis in Squalus
acanthias. Jour. Morph., vol. 26.

Bine, R., anp Burckuarpt, R. 1905 Das Centralnervensystem von Cerato-
dus forsteri. Jenaische Denkschriften, Bd. 4.

Bripce, T. W. 1879 The osteology of Polyodon folium. Phil. Trans. Royal
Soe. London, vol. 169.
1899 The air-bladder and its connection with the auditory organ in
Notopterus borneensis. Jour. Linn. Soe. Zool., vol. 27.
1904 Fishes, Cambridge Nat. Hist., vol. 7.

Bupcett, J. 8. 1902 On the structure of the larval Polypterus. Trans. Zool.
Soc., vol. 16.

Danrortu, C. H. 1912 The heart and arteries of Polyodon. Jour. Morph.,
vol. 23.

Dourn, A. 1886 Studien zur Urgeschichte des Wirbeltierkérpers. XI. Mit-
theil. Stat. Neapel, Bd. 7.

Driner, L. 1901 Studien zur Anatomie der Zungenbein-, Kiemenbogen-
und Kehlkopfmuskeln der Urodelen, I. Teil: Zoolog. Jahrb. Abt. f.
Anat., Bd. 15.
1903. Uber die Muskulatur des Visceralskelettes der Urodelen. Anat.
Anz., Bd. 28.
1904 Uber die Anatomie und Entwickelungsgeschichte des Mittel-
ohres beim Menschen und bei der Maus. Anat. Anz., Bd. 24.

Fawcett, E. 1910 Notes on the development of the human sphenoid. Jour.
Anat. and Physiology, vol. 44.

Frortep, A. 1902a Zur Entwickelungsgeschichte des Wirbeltierkopfes. Anat.
Anz. Erginz. zum 21. Band.
1902 b Einige Bemerkungen zur Kopffrage. Anat. Anz., Bd. 21.

Fucus, H. 1910 Uber das Pterygoid, Palatinum und Parasphenoid der Quad-
rupeden, insbesondere der Reptilien und Siugetiere, nebst einigen
Betrachtungen iiber die Beziehungen zwischen Nerven und Skelett-
eilen. Anat. Anz., Bd. 36.
1912 Uber einige Ergebnisse meiner Untersuchungen iiber die Ent-
wickelung des Kopfskelettes von Chelone imbricata. Anat. Anz.
Ergiin. zum 41. Band.

Gaupp, E. 1893a Uber die Anlage der Hypophyse bei Sauriern. Archiv
mikros. Anatom., Bd. 42.
1893 b Beitrige zur Morphologie des Schadels. 1. | Primordial-Cran-
ium und Kieferbogen von Rana fusca. Morphol. Arbeiten, Bd. 2.
1900 Das Chondrocranium von Lacerta agilis. Ein Beitrag zum
Verstindnis des Amniotenschiidels. Anat. Hefte, Bd. 14.
1902 Uber die Ala temporalis des Siugerschidels und die Regio or-
bitalis einiger anderer Wirbeltierschidel. Anat. Hefte, Bd. 19.
1905 a Neue Deutungen auf dem Gebiete der Lehre vom Siugetier-
schidel. Anat. Anz., Bd. 27.
1905 b Die Entwickelung des Kopfskelets. Handbuch d. vergl. u.
experim. Entwickel. d. Wirbeltiere von O. Hertwig. Bd. 3.

320 EDWARD PHELPS ALLIS, JR.

Gavupp, EH. 1908 Zur Entwickelungsgeschichte und vergeleichenden Morphologie
des Schadels von Echidna aculeata var. typica. Jenaische Denks-
chriften, Bd. 6, Teil 2. ;

GEGENBAUR, C. 1872 Das Kopfskelet der Selachier. Untersuchungen zur
Vergleichenden anatomie der Wirbeltiere. H. 3., Leipzig.

GreiLt, A. 1913 Entwickelungsgeschichte des Kopfes und des Blutgefiiss-
systemes von Ceratodus forsteri. Zweiter Teil: Die epigenetis-
chen Erwerbungen wihrend der Stadien 39-48. Jenaische Denks-
chriften, Bd. 4.

Grosser, O. 1901 Zur Anatomie und Entwickelungsgeschichte des Gefiiss-
systemes der Chiropteren. Anat. Hefte, Bd. 17.

Grosser, O., AND Brezina, E. 1895 Uber die Entwicklung der Venen des
Kopfes und Halses bei Reptilien. Morphol. Jahrb., Bd. 23.
GintHer, A. C. L. G. 1871 Description of Ceratodus, a genus of ganoid fishes
recently discovered in Queensland, Australia. Phil. Trans. Royal

Soc. Lond.

Hanprick, K. 1901 Zur Kenntnis des Nervensystems und der Leuchtorgane
von Argyropelecus hemigymnus. Zoologica, Heft 32.

Hay, O. P. 1895 On the structure and development of the vertebral column
of Amia. Field Columbian Museum, Chicago vol. 1, no. 1, Zool. Ser.

His, W. 1892 Die Entwickelung der menschlichen und thierischer Physiog-
nomien. Archiv Anatom. u. Entwickel.

Huxtey, T.B. 1871 A manual of the anatomy of vertebrated animals. London,
1876 On Ceratodus forsteri, with observations on the classification
of fishes. Proce. Zool. Soe. London.

KERNAN, J. D., Jr. 1916 The chondrocranium of a 20-mm. human embryo.
Jour. Morph., vol. 27.

McMurricu, J. P. 1884 The osteology of Amiurus catus (L.) Gill. Proce.
Canadian Instit. vol. 2, Toronto.

Noorpensos, W. 1905 Uber die Entwickelung des Chondrocraniums der
Siugetiere. Petrus Camper. Deel. 3.

OumstEap, M. P. 1911 Das Primordialcranium eines Hundeembryo. Hin
Beitrag zur Morphologie des Siugetierschidels. Anat. Hefte ,Bd. 43.

Parker, K. M. 1917 The development of the hypophysis cerebri, pre-oral gut
and related structures in the Marsupalia. Jour. Anatomy, vol. 51,
pt. 3. London.

Parker, W. K. 1873 On the structure and development of the skull in the
salmon (Salmo salar). Phil. Trans. London.

1876 On the structure and development of the skull in sharks and
skates. Trans. Zool. Soc. London, vol. 10.

1882. On the development of the skull in Lepidosteus osseus. Philos.
Trans. Royal Soc. London, vol. 173.

Ripewoop, W. G. 1904 On the cranial osteology of the fishes of the families
Mormyridae, Notopteridae, and Hyodontidae. Journ. Linn. Soc.
London, Zoology, vol. 29.

SacemMEnHL, M. 1883 Beitriige zur vergleichenden Anatomie der Fische. 1.
Das Cranium von Amia calva (L.) Morph. Jahrb., Bd. 9.

1891 Beitrige zur vergleichenden Anatomie der Fische: IV. Das
Cranium der Cyprinoiden. Morph. Jahrb., Bd. 17.

MYODOME AND TRIGEMINO-FACIALIS CHAMBER 321

Sauzer, H. 1895 Uber die Entwicklung der Kopfvenen des Meerschweinchens.
Morph. Jahrb., Bd. 23.

Scuarer, HE. A. 1894 Organs of the senses. Quain’s Anatomy, vol. 3, pt. 3,
10th edition, London.

SCHAUINSLAND, H. 1905 Die Entwickelung der Wirbelsiiule nebst Rippen und
Brustbein. Handb. Entw. Wirbelth. O. Hertwig 1906, Bd. 3, Abth.
2, Jena.

Scuierp, W. 1904 Die Entwickelung der Kopfknochen bei dem Lachs und der
Forelle. Anat. Hefte, Bd. 23.

ScHREINER, Ix. . 1902 Einige Ergebnisse tiber den Bau und die Entwick-
lung der Occipitalregion von Amia und Lepidosteus. Zeit. wiss.
Zool., Bd. 72.

Sewerrzorr, A. N. 1897 Beitrag zur Entwickelungsgeschichte des Wirbel-
tierschadels. Anat. Anz., Bd. 13.

1899 Die Entwicklungsgeschichte des Selachierschidels. Festschrift
C. von Kupffer, Jena.

Suino, Kk. 1914 Studien zur Kenntnis des Wirbeltierkopfes. 1. Das Chon-
drocranium von Crocodilus mit Beriicksichtigung der Gehirnnerven
und der Kopfgefiisse. Anat. Hefte, Bd. 50.

Sonies, F. 1907 Uber die Entwickelung des Chondrocraniums und der knor-
peligen Wirbelsiule bei den V6geln. Petrus Camper, Deel. 4.

Srarks, EH. C. 1905 Some notes on the myodome of the fish cranium. Science,
WO, Wil, joy, ‘Wk, 765).

Stour, P. 1882 Zur Entwicklungsgeschichte des Kopfskeletes der Teleostier.
Festschr. der Julius-Maximilians-Universitait zu Wiirzburg. Leipzig.
1895 Uber die Entwicklung der Hypochorda und des dorsalen Pank-
reas bei Rana temporaria. Morph. Jahrb., Bd. 23.

Suprno, F. 1901 Ricerche sul cranio dei Teleostei. 1. Scopelus, Chauliodus,
Argyropelecus. Ricerche Lab. Anat. Roma, T. 8.

SwinnerTON, H. H. 1902 A contribution to the morphology of the teleostean
head skeleton, based upon a study of the developing skull of the three-
spined stickleback (Gasterosteus aculeatus). Quart. Jour. Mic.
Sci., vol. 45.

Terry, R. J. 1917 The primordial cranium of the cat. Jour. Morph., vol. 29.

THane, G. D. 1894 Arthrology—Myology—Angeiology. Quain’s Elements of
Anatomy, vol. 2, pt. 2, London.

Veir, O. 1907 Uber einige Besonderheiten am Primordialeranium von Lep-
idosteus osseus. Anat. Hefte, Bd. 33.

1911 Beitrige zur Kenntnis des Kopfes der Wirbeltiere. 1. Die
Entwickelung des Primordialeranium von Lepidosteus osseus. Anat.
Hefte, Bd. 44.

Vorr, M. 1909 Das Primordialeranium des Kaninchens unter Beriicksichtig-
ung der Deckknochen. Ein Beitrag zur Morphologie des Siéuge-
tierschidels. Anat. Hefte, Bd. 38.

Watpscumipt, J. 1887 Beitrag zur Anatomie des Centralnervensystems und
des Geruchsorgans von Polypterus bichir. Anat. Anz,, Bd. 2.

vAN Wine, J. W. 1905 Uber die Entwickelung des Kopfskelets bei Selachiern.
Compte Rendu, 6ieme Congrés Internat Berne, 1904, published
Geneva.

O22 EDWARD PHELPS ALLIS, JR.

Figs. 1 to 12. Semidiagrammatie figures of transverse sections through a 51-
mm. specimen of Hyodon tergisus, showing the myodomic cavities and adjacent
regions, and selected at intervals between the hind end of the interorbital sep-
tum and the hind end of the basioccipital.

Figs. 13 to 18 Similar sections of a 20-mm. specimen of Cottus scorpius.

Fig. 19 Similar sections of a 3l-mm. specimen of Clinocottus analis.

Figs. 20 to 23 Similar sections of a 115-mm. specimen of Syngnathus acus.

Figs. 24 to 29 Similar sections of a 57-mm. specimen of Catostomus occiden-
talis.

ABBREVIATIONS

ab, air-bladder ng, nervus glossopharyngeus
as, aortal support nocm, nervus oculomotorius
ba, basioccipital nt, nervus trigeminus
br, brain ona, orbitonasal artery
dlp, dorsolateral vertebral process pb. I, pharyngobranchial of first bran-
ear, parts of membranuous ear chial arch
ec, external carotid artery pb. II, pharynogbranchial of second
eff, I, efferent artery of first branchial branchial arch
arch pf, ramus palatinus facialis
epsb, efferent pseudobranchial artery php, pharyngeal process
exo, exoccipital pro, prootic
hmd, hyomandibula prob, prootic bridge
hmy, horizontal myodomice membrane ps, parasphenoid
hy, hypophysis psb, pseudobranch
ic, internal carotid artery pv, pituitary vein
ic. a, anterior division of internal car- re, musculus rectus externus
otid artery res, recessus sacculus
ic. p, posterior division of internal car- inf, musculus rectus inferior

otid artery
ios, interorbital septum
ja, Jacobson’s anastomosis
l, ligament
lda, lateral dorsal aorta
mda, median dorsal aorta

rint, musculus rectus internus

rs, musculus rectus superior

sv, saccus vasculosus

syc and sy-c, sympathetic nerve and
communicating branch from N. tri-

nab, nervus abducens geminus to N. facialis
nc, notochord or notochordal space vj, vena jugularis

nf, nervus facialis vlp, ventro-lateral vertebral process

MYODOME AND TRIGEMINO-FACIALIS CHAMBER
EDWARD PHELPS ALLIS, JR.

PLATE 1

PLATE 2 MYODOME AND TRIGEMINO-FACIALIS CHAMBER
EDWARD PHELPS ALLIS, JR.

324

MYODOME AND TRIGEMINO-FACIALIS CHAMBER PLATE 3
EDWARD PHELPS ALLIS, JR.

Ld
JO /

>» il tint Pe

ac 19 PS ry rint Ts

PLATE 4 MYODOME AND TRIGEMINO.FACIALIS CHAMBER
EDWARD PHELPS ALLIS, JR.

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, APRIL 7

ON THE NATURE, OCCURRENCE, AND IDENTITY OF
THE PLASMA CELLS OF HOFBAUER

ARTHUR WILLIAM MEYER

Department of Embryology, Carnegie Institution of Washington, and Department of
Anatomy, Stanford Medical School

The history of these cells illustrates very well how a re-discovery,
when accompanied by a fuller description, succeeds in domiciling
itself in anatomical literature as an original discovery. As we
shall presently see, Hofbauer (’05) was impressed especially
by a conspicuous phase in the life history of a particular
cell. He noted its reaction in the fresh state, to certain stains,
described it more fully, and speculated with some freedom on its
functional réle; but he did not discover this cell, as he supposed,
in 1903. Although Hofbauer refers to his address given in 1903
in his book published in 1905, he does not refer to or list the
paper based on this address, published in 1903, in the title of which
he refers to these cells as ‘hitherto unknown’ and as ‘constantly
occurring.’ His failure, in 1905, to recognize earlier workers
was, I presume, an oversight, which apparently led Essick (715)
and others to assume that ‘‘Hofbauer first called attention to
specific round cells appearing in the human placenta toward
the end of the fourth week of pregnancy.”’

The type of cells which in recent years has been designated
with Hofbauer’s name was known previously especially as Wan-
derzelle and had been represented by various investigators.
Minot (’12), in a footnote, refers to the latter fact and rightly
adds: ‘‘It has long been known that strikingly large free cells
appear in the mesenchyme of the chorion. They are pictured
in my Human Embryology.’’ Reference to the illustration in
this work shows a large, rather granular cell, with a somewhat
eccentrically placed, vesicular nucleus, but without vacuoles.
Moreover, previous to the publication of the Embryology,

327

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

328 ARTHUR WILLIAM MEYER

Minot (’89) not only spoke of large, granular, wandering cells
in the stroma of the chorion, but also represented them. From
Minot’s familiarity with the work of Langhans (’77) and of
Kastschenko (’85), it does not seem unlikely that, among others,
he had these investigators particularly in mind when he referred
to the earlier descriptions.

In the absence of a more discriminating term for these er-
ratic and ephemeral elements, the original designation of wan-
dering cell would seem far preferable to the designation lipoid
interstitial cells, used by certain Italian writers. The former
is a non-committal term, and, although too inclusive, is for this
reason no more objectionable than the expression giant-cell.
Although these cells may not—indeed, probably do not—wan-
der in the sense of the amoeba or the leucocyte, they neverthe-
less may change their location decidedly. The qualification,
intersititial, is objectionable for the very reason for which it
was chosen—the alleged analogy to the interstitial cells of testis
and ovary, and since they may contain lipoid substances merely
because they are degenerate, the adjective lipoid is equally
objectionable. For reasons to appear later, the designation
plasma cell, used by certain Italian writers after Hofbauer,
would not seem to be justified.

Virchow (’67) stated that isolated cells with clear vesicular
spaces in their protoplasm are found in the stroma of the villi
in cases of hydatiform degeneration, and identified them with
certain other cells, physaliphores—previously described by him.
He found these bubble-like cells, as he called them, also in the
thymus of the new-born, in cancer, etc., and, according to Vir-
chow, they were not merely vacuolated cells. He seems to have
regarded these cells as identical also with the vacuolated syncy-
tial masses, for he stated that Miiller described them as occur-
ring in the chorionic epithelium. Since syncytial elements not
rarely are found in the stroma, instances of confusion of these
two cell types can be found in contemporary literature also.

Langhans (’77), in describing the stroma of the villi, said
that it contained ‘“‘sharply delimited large cells with many
granules in the protoplasm. Their form is variable—circular,

PLASMA CELLS OF HOFBAUER 329

spindle and star-shaped.’’ These cells were said to lie mainly
near the periphery. However, Langhans, who was interested
mainly in other problems, did not represent them nor discuss
their probable significance. But Kastschenko (’85) represented
them and described them as being about 9u large, and as cor-
responding exactly in form and size to the white blood cells
of the same embryo. According to Kastschenko, the cytoplasm -
is reduced in quantity after the first month, so that the nucleus
no longer is surrounded by it. The nuclei, also, are said to
undergo a change and to appear later as solid structures. The
latter observation cannot fail to remind one of pycnosis and
of one of its well-known significances. Kastschenko found
these cells mainly near the epithelium of the villi and stated
that they vary greatly in size, number, and occurrence in the
same placenta. The fact that Kastschenko identified the cells
found in the mesenchyme of the embryonic villi as leucocytes
might seem to indicate that what he saw and described were
other than Hofbauer cells. However, his illustrations, espe-
cially when considered in connection with those of earlier inves-
tigators and those of Minot, leave little doubt that all these
investigators saw the same type of cell. Moreover, it is not
improbable that Kastschenko was influenced in his interpre-
tation of these cells by the origin and current use of the term
Wanderzelle. It will be recalled that von Recklinghausen (’63)
showed that the leucocyte preeminently belonged in this class
of cells, but even at the time that Kastschenko was writing and
far later, all cells which were regarded as foreign to the tissue
in which they lay still were included in the designation Wan-
derzelle. Reference to the literature of that period will make
this fully evident.

Tne presence of these cells in pathologic ‘ova’ was noticed re-
peatedly by Mall (’08), who also designated them as wander-
ing or migrating cells in his earlier protocols. Chaletzky (91)
also saw and described these cells, but perhaps the best de-
scription from an earlier date is that given by Kossman (’92),
who also refered to the Hofbauer cells as Wanderzellen, and
gave excellent representations of them. Indeed, from an inspec-

330 ARTHUR WILLIAM MEYER

tion of the latter alone there can be no question as to the iden-
tity of these Wanderzellen and the Hofbauer cell. In speaking
of them, Kossman said:

: Auffallend sind zahlreiche grosse Zellen, die eine sehr
wechselnde, oft amdboide, niemals sternformige Gestalt haben. Die
Filarmasse ihres Protoplasma’ s ist durchaus fein netzartig angeordnet
und firbt sich stark in Hamatoxylin. Die Zellen enthalten einen
oder mehrere grosse blasenartige Hohlriume, von denen ich !eider
nicht sicher sagen kann, ob sie Fett fiihrten, da sie mir erst nach Be-
handlung des Priparats mit Xylol auffielen. Der Kern dieser Zel-
len enthielt stets Nucleoli. Die Zellen sind also jedenfalls nicht in
lebhafter Vermehrung; wahrscheinlich sind es Wanderzellen und da sie
auf einem wenig dlteren Stadium wieder fehlen, mag ihr Vorkommen
in einigem Zusammenhange mit der um diese Zeit beginnenden
Vascularisation des Stroma’s stehen.

Merttens (94) found the same cells in abortuses, and, in de-
scribing the stroma of the villi of his first case said:

An den Ernahrungszotten ist es kernreich, vielfach aufgelockert,
mit stern- und spindelformigen Zellen, in den Maschen jene oben fiir
die normalen ersten Stadien beschriebenen grossen, runden oder poly-
edrischen Zellen mit kérnigem oder auch vacuoléirem Protoplasma mit
grossem, blaschenférmigem, rundem Kern.

Merttens seems also to have suggested that these cells are
swollen stroma cells, but since he made this observation some-
what disconnectedly I am not quite certain of his meaning;
yet the mere suggestion is particularly interesting in view of
Minot’s special emphasis upon the degenerate character of the
Hofbauer cells. Marchand (’98) also wrote: ‘‘ Die durchsich-
tigen hellen Zellen im Stroma normaler oder pathologischer
Zotten sind mir wohlbekannt, sie kjnnen denen der Zellschicht
sehr ahnlich sein; ich halte sie jedoch fiir gequollene, rundlich
gewordene Bindegewebszellen, da mann Ubergiinge zu solehen
findet, ebenso wie in andern Schleimgeweben.”’

Ulesco-Stranganowa (’96), who also saw these cells, says that
if one compare the Langhans cells with round nuclei with these
cells scattered about the stroma of the villi, and which have
been named Wanderzellen by Kastschenko, one becomes con-
vineed of the identity of these two types of cells. According

PLASMA CELLS OF HOFBAUER bel

to Ulesco-Stranganowa, then, the Hofbauer and Langhans cells
are identical. Mall (’15) also called attention to this possibil-
ity, for, when speaking of the invasion of the mesoderm of the
villi by trophoblast, he called attention to the presence of numer-
ous Hofbauer cells, and added: ‘‘It would seem possible that
these Hofbauer cells are free trophoblast cells within the meso-
derm of the villus, an opinion already expressed in my paper
on monsters.’’ Neumann (’97) also noticed these cells and
referred to Virchow’s opinion regarding them, and von Lenhos-
sek (’02) is credited in 1904 by the reviewer of his paper with
having examined a large series of young human embryos, and
having suggested that what Kastschenko regarded as Wander-
zellen were mesenchyme cells. It should be noted, however,
that von Lenhossek apparently came to this conclusion largely
because of the absence of blood-forming organs or lymphatic
centers in embryos the villi of the chorionic vesicles of which
contained these cells. Strangely enough, Kworostansky (’03)
also recorded the presence of these cells, and after describing
the stroma of the villi wrote:

Zwischen den genannten Bindegwebszellen giebt es in der wolki-
gen Grundsubstanz Liicken, und am Rande oder im Winkel derselben
sitzen frele andere Bindegewebszellen, die sehr gross sind, lappige,
runde Form, wabenartiges Protoplasma und gleiche Kerne wie andere
Bindegewebszellen haben; ihre Kerne werden auch, hie und da stern-
formig getheilt. Da sie stets nur in Gewebsliicken gefunden werden,
so glaube ich, sie als Lymphgefassendothelien, oder vielleicht als Lymph-

ocyten bezeichnen zu diirfen. Man findet sie in spaiteren Stadien
der Placenta nur sind dann natiirlich die zellen nicht mehr gross.

The illustration which accompanies Kworostansky’s article,
as well as his description, leaves no doubt that the cells seen by
him are the same as those which we are considering, although
his surmise that they are lymphocytes and that they arise from
the endothelium of the lymphatics may, upon first thought,
seem rather irreconcilable with such an interpretation.

From these references alone it is evident that Minot’s state-
ment, that the so-called Hofbauer cells were repeatedly men-
tioned in the earlier literature, is well founded. Muggia (’15)
states that these cells were described also by Guicciardi (99),

332 ARTHUR WILLIAM MEYER

Clivio (03), Stoffel (05),! Vecchi (’06), and Pazzi (’04). In-
deed, many other names could be added, for surely any one of
the many who studied even a small series of chorionic vesicles
must have seen some of them in some villi, especially in unrec-
ognized cases of hydatiform degeneration, but since they have
been referred to as Hofbauer cells, it is his description that es-
pecially interests us. In describing the chorionic villi, Hof-
bauer (’05) spoke of certain gaps or spaces between the meshes
of the mesenchyme of the villi which he thought might belong
to the lymphatics or contain tissue fluid. In these spaces he
found certain granular, round cells arranged longitudinally. He
thought they often were spherical with a diamenter of 10.5 u to
12.5. but more commonly star-shaped or branched. By means of
these branches they come into direct relation with other similar
cells or with connective-tissue cells. However, Happe (’06)
stated that he could not with certainty find cells united by their
processes, as described by Hofbauer, in preparations stained
after Hansen. According to Hofbauer, the cell processes are
delicate, and the cells contain one or two nuclei from 4.7 u to
5.7 « in diameter, oval or circular in form, eccentric in position,
with a definite membrane and a dense chromatin network. Mi-
toses were common, and fragmentation of nuclei and indications
of pluripolar mitoses also were seen. Hofbauer emphasized that
the most characteristic thing in these cells which he regarded
as being specific was the presence of vacuolation in the ‘plasma’
and the existence of a perinuclear clear zone, which was said
to be the result of fusion of ‘small light spots.’ As the cyto-
plasm becomes vacuolated the nucleus is said to become pyc-
notic, which stage is followed by failure to stain and finally by
complete disappearence. Hofbauer also noticed the presence
of granules and fat droplets, and regarded the life history of
the cell as a circumscribed one. He did not find them present
in real young villi. They were said to appear at the end of the
fourth week, and were more common in young than in old placen-
tae. They reacted to vital stains like plasma cells, and Hof-

1A rereading of Stoffel’s article shows quite conclusively that he did not de-
scribe the plasma cell of Hofbauer.

PLASMA CELLS OF HOFBAUER 333

bauer regarded the vacuoles as having an assimilative and di-
gestive function. A reference to the plates accompanying
Hofbauer’s monograph, however, shows that vacuolation was
not always present, and that the largest of the cells were almost
twice the size of the smallest. ‘

In his earlier paper Hofbauer (’03) also said that his prepa-
rations taken from material from the fourth to the ninth weex
of pregnancy, and obtained at operation, showed these cells
in all stages of mitotic division. Hofbauer further wondered
whether the spaces surrounding these cells are lumina of capil-
laries, added that the cells discovered by him undoubtedly are
found in capillaries, and made some rather unguarded surmises
concerning them. —

Berlin (’07), in writing on the changes in retained placentae,
also spoke of large swollen, hydropic cells which lie in spaces.
These cells she regarded as undoubted mesenchyme cells. How-
ever, Berlin did not believe that they are degeneration products,
although her description certainly would lead one to suppose
that they were such. Even when she states that they bear no
sign of degeneration, emphasizing that the chromatin network
is fine, she speaks of swollen nuclei which have gathered a larger
amount of protoplasm about them, phenomena which she re-
garded as signs of luxurious nutrition. Moreover, Berlin never
observed mitoses and never found the nuclei increased in villi
containing many of these cells, an observation wholly in har-
mony with that of others and directly opposed to proliferation.

Grosser (710), who was plainly aware of the fact that Hofbauer
was not the discoverer of these cells, also represented a cell
which, however, is non-vacuolated and binucleated, and added
that their significance is still unknown.

I have given Hofbauer’s description, partly to emphasize
the vacuolation, for it was this which also impressed Minto,
(04), who rightfully stated:

We frequently find in the literature mention of wandering cells
with vacuolated protoplasm, but they seem not to have been recog-
nized as degenerating cells. . . . . The disintegration by vacu-
olation has, so far as known to me, not been described heretofore, and

334 ARTHUR WILLIAM MEYER

consequently may be treated somewhat more fully. Renewed investi-
gation has led me to the conclusion that we have to do with erythro-
cytes which have gotten into the mesenchyma and, remaining there,
have swollen by imbibition and are undergoing degeneration by vacu-

olization of their protoplasm. . . . . We can explain the appear-
ance of these cells by the assumption of imbibition, in which the
nucleus has participated. . . . . Since I have found similar cells

in a considerable number of placentas, I draw the conclusion that they
are constant and normal. I regard the interpretation of the pictures
unattackable as proof of progressive degeneration. —

In association with these remarks, Minot represented a series
of cells showing progressive degeneration, beginning with the
nucleated red cells and ending with a highly degenerated, but
nevertheless nucleated, Hofbauer cell which apparently is in
process of disintegration. These cells were seen by Minot
especially in a human embryo of 15-mm. length, from the Mall
collection.

As shown in the references to the literature above, it is not
quite correct to say that the degenerate character of vacuolation
has not before been recognized, for the surmises that Hofbauer
cells contain fat granules may, and that they are swollen mesen-
chyme cells must, carry this implication. Moreover, those
familar with the effects of inanition know that investigators
of this subject long ago called attention to vacuolation as one
of the evidences of degeneration although, certainly, no one
contends that it always is such.

Instead of regarding these cells as degeneration products,
certain Italian writers (notably Acconci ’14) regard cells which
they found, especially in the first half of pregnancy, as morpho-
logically and functionally comparable to the intersitital cells of
the ovary and testis. Acconci believed that certain cells which
he and other Italian writers after him designated lpoid-inter-
stitial cells, probably produce a special internal secretion. He,
like Hofbauer, found these cells to contain lipoid granules, and
regarded them also as equivalent to certain cells “‘described by
Ciaccio in various parts of the organism, or by Brugnatelli
in the interstitial tissue of the mammary gland.” Acconci
further emphasized certain similarities between the syncytium

PLASMA CELLS OF HOFBAUER 335

and the interstitial cells, both of which he conceived as exercis-
ing a protective réle. Muggia (’15), too, instead of regarding
the lipoid interstitial cells of Acconci as degenerate, emphasized
his belief that they are particularly resistant to degeneration,
being found perfectly preserved in the midst of detritus. Since
the young connective-tissue cell loses, or rather retracts, its
processes as it becomes converted into a Hofbauer cell, it need
not surprise us that the latter survives the former. Retraction
of the processes contributes to the apparent increase of cyto-
plasm of the rounded swollen cell and also is involved in the
formation of the spaces in which these cells usually lie. Mug-
gia, who considered the cells found by him in great numbers in
a case of partial hydatiform degeneration, as identical with
those described by Acconci, gave a fine detailed description
absolutely typical of the cells described in greatest detail by
Hofbauer. Moreover, the excellent illustrations which accom-
pany Muggia’s article leave no doubt as to the identity of the
cells or of their degenerate character. Muggia stated that these
cells in normal villi increase until the end of the fifth month,
when, according to Savare, they are most numerous. Muggia
further found numerous cells very similar to the interstitial
cells of Acconci, or “the plasma-like cells of Hofbauer,’ which
he says are regarded by some as early stages of interstitial cells
and by others as mast cells, although he regarded them as par-
tially differentiated interstitial cells.

Until I had seen sections of the chorion of embryo no. 1531,
I was largely at a loss to know why Hofbauer cells so frequently
were described as lying in gaps or spaces in the mesenchyme.
However, in this specimen cross-sections of a number of villi
showed splendid examples of this condition, which alone made
the cells very conspicuous. The cells often were very numer-
ous, in fact more numerous than the mesenchyme cells which
remained, although some well-preserved villi contained no Hof-
bauer cells whatever. Some of the younger specimens also
contained none. This was true of a chorionic vesicle with an
embryo 1 mm. in length. They were found most commonly
in the villi, but not infrequently some of them lay in areas of

336 ARTHUR WILLIAM MEYER

the chorionic membrane which had undergone degeneration.
They were not so common here, but sometimes were exceed-
ingly numerous in small areas. They were found in the am-
nion also, in the umbilical cord, and in the tentorium cerebelli,
and as isolated specimens in embryonic mesenchyme elsewhere.
As emphasized by other investigators there seemed to be nothing
particularly characteristic about their distribution except that
they were more common in places where the mesenchyme was
degenerating. Sometimes a considerable number were con-
tained in one villus and none in an adjacent one. As many as
twelve might lie in one field and none in the next. Very rarely
was there a solid mass of them, but usually they were scattered
about at random, although groups also were seen. The better-
preserved cells were small, the poorer-preserved larger, the size
varying from 8.5 4 to 30 u. The smaller cells usually were
quite circular in outline, stained evenly and possessed a non-
granular cytoplasm with a nucleus quite centrally located. Bi-
nucleate cells, as described by Grosser, were not uncommon,
and multinucleated cells—fusion products—also were found.
The nuclei of the latter frequently were more unequal in size,
and usually also more oval in outline, than the single nucleus
of the typical Hofbauer cell. Measurements of the larger cells
made with a micrometer caliper, gave the following results which
are considerably above those given by Hofbauer, whose estima-
tion of 10.5 uw to 12.5 uw applies to the average-sized cell.

Size of the larger Hofbauer cells in micra

25.5 X 20.4
30.4 X 27.5
18.0 X 12.0
21.5 X 25.5
18.0 < 14.0

However, the size of the cells varied from specimen to specimen
of chorionic vesicle, but not nearly so much as their state of
preservation. ‘This no doubt, partly is due to the varying state
of preservation of the villi themselves.

PLASMA CELLS OF HOFBAUER So

In outline they varied from irregular to circular, as stated by
Hofbauer, and as represented by Minot (’11) in his series showing
progressive degeneration. - Although it was easy to distinguish
the vacuolated Hofbauer cell from the well-preserved mesen-
chyme cell with cylindrical nucleus and many processes, speci-
mens which represent transition forms as stated by Marchand,
were quite common. ‘The latter generally were oval or slightly
irregularly formed cells with a number of short processes, which
latter, as well as the character of the nuclei and the form of the
cell itself, certainly suggested a mesenchymal origin. They
were also most numerous in villi the stroma of which had be-
come vacuolated or fenestrated. Here the reciprocal numer-
ical relationship between the Hofbauer and the mesenchyme
cells often was especially evident. In certain areas in which
almost no mesenchyme cells remained intact, numerous Hof-
bauer cells occurred in all stages of degeneration. In other por-
tions of the chorionic membrane or of the villi, mesenchyme
cells with processes in all stages of retraction also were clearly
outlined in the homogeneous ground substance. Such evidences
naturally remind one of Hofbauer’s statement that “Marchand
called his attention to the fact that these cells were mesenchyme
cells, a conclusion which Hofbauer accepted. My implication,
however, is not that degeneration of the mesenchyme or of
individual mesenchyme cells can proceed only through a Hof-
bauer stage, but that, especially in the chorionic villi, a form of
degeneration of the mesenchyme seems to occur which gives
rise to this peculiar cell form, the degenerate character of which
rightly impressed Minot. This relationship also attracted the
attention of Mall (’15), who represented degenerating villi and
stated

The core of the villus gradully breaks down and disintegrates. While
this process is taking place we often see scattered through the stroma
of the villus large protoplasmic cells. . . . . These cells. which
I have repeatedly seen in the villi of pathological ova, may be a type
of wandering cells; at any rate, when the villus is being invaded by
the leucocytes and trophoblast it might be thought that they arise
from the latter, but this is improbable.

338 ARTHUR WILLIAM MEYER

It is of particular interest in this connection that Virchow
(63) stated that Schroeder van der Kolk (’51) had concluded
that large clear cells in the stroma of the villi, later classed among
the physaliphores by Virchow, occurred too frequently to be
correlated with hydatiform degeneration. This suggests that
the so-called Hofbauer cells were known since the early days
of cytology, and that some one must have noticed, even at that
early date, that they were very common in some hydatiform
moles. Whether or not this was van der Kolk I am unable to
say, but that Hofbauer cells are especially numerous in some
cases of hydatiform degeneration is undoubted. But it does
not therefore follow that they constantly are present in this
condition. Large numbers of Hofbauer cells occurred in seven-
teen out of the sixty-one cases of normal and pathologic chor-
ionic vesicles in which they were especially studied. Of these
seventeen cases fourteen later were independently identified
as showing hydatiform degeneration, and the other three were
considered as possibly such. In other words, every case of this
ser.es of so-called normal and pathological chorions in which
the Hofbauer cells were numerous, was one showing hydati-
form degeneration of the villi. It also is true, however, that
thirty-four cases containing a few or some Hofbauer cells were
not identified as being hydatiform moles, although three cases
containing smaller numbers of these cells were so recognized.
Moreover, not a single case of this series of sixty-one specimens
which contained no Hofbauer cells whatever was later identi-
fied as showing hydatiform degeneration.

Somewhat similar evidence was afforded by the study of the
twenty-two cases in the protocols of which Mall had noted that
Hofbauer cells were present. Of these twenty-two cases, thir-
teen later were identified as showing this degeneration. How-
ever, since a total of 112 cases of hydatiform degeneration were
identified among the 313 classed as pathologic among the first
thousand accessions in the Mall collection, it is evident that the
presence of Hofbauer cells was especially noted in but a rela-
tively small percentage of the series of embryos classed as path-
ologic. If we include certain other cases in which they came

PLASMA CELLS OF HOFBAUER 339

to attention later, the percentage becomes 26.7; that is, 30 out
of 112 cases of hydatiform degeneration. Of these thirty cases
containing Hofbauer cells in sufficient numbers to attract atten-
tion in the course of a routine examination made for other purposes,
seventeen or 56.6 per cent, were later identified as instances of
hydatiform degeneration. Since the sixty-one cases in the first
series were examined especially for the purpose of study of Hof-
bauer cells, the higher percentage of correlation observed in
this series may be due partly to this fact. At any rate, that
such a correlation exists seems to be quite clear, although I do
not conclude that the two conditions necessarily or invariably
are associated.

It is interesting that Pazzi (’04) considered a distrophy of
the connective tissue with the development of cellular elements
“not very well differentiated, but like the plasma cell of Hof-
bauer,” as the initial and pathognomonic change in hydati-
form degeneration. Pazzi further stated that the plasma cell
of Hofbauer may be in a state of hyperactivity or of de-
generation, and questioned the statements that Hofbauer cells
appear only at the end of the fourth week and that they have a
short life. Pazzi regarded the Hofbauer cell as fundamentally a
constituent of the villi, as the decidual cell is of the decidua.
He, like Essick, attributed their origin to the endothelium of
the vessels, and Pazzi suggested that the Hofbauer cell may have
a special internal secretion intended to preserve the stroma of
the young villus against degeneration. Pazzi further considered
the question whether a Hofbauer cell can transform itself into an
epithelial cell and finally into a syncytial cell, adding that the
invasion of the stroma of the villus by epithelial growth, is only
a special development of Hofbauer cells!

As already stated, Muggia also found these cells very abun-
dant in a case of partial hydatiform degeneration, and held that
their appearance and condition was correlated with the pro-
liferation and vacuolation of the syncytium, maintaining that,
as the latter becomes vaculolated the lipoid interstitial cells
of Acconci appear, the changes in the two being wholly parallel.

340 ARTHUR WILLIAM MEYER

Since thirty-two of the fifty-one specimens in this series of
sixty-one containing a few, some, or many Hofbauer cells had
been classed among the pathologic, it follows that these cells
were noticed more frequently in the pathological than in speci-
mens classed as normal. This becomes especially evident if
we exclude from this series of fifty-one cases all those contain-
ing some or many Hofbauer cells, for of twenty-seven of these,
nineteen, or 70.4 per cent, had been classed among the patho-
logic. Moreover, since the great majority of the conceptuses
classed as normal belong among abortuses, one would be en-
tirely justified in questioning the strictly histologically normal
nature of the chorionic vesicles which accompany some embryos
classed as normal. At any rate, it is evident that the plasma
cell of Hofbauer is associated with degenerative changes in the
mesenchyme of the villi. Since such changes are more common
in pathologic abortuses it is not surprising that Hofbauer cells
are more common in the latter than in normal specimens, and,
since degenerative changes in the stroma are especially pro-
nounced in advanced cases of hydatiform degeneration, it is
still less surprising that Hofbauer cells are particularly common
in this condition. But they are not necessarily pathognomonic
of hydatiform degeneration, although it is true that when at
all numerous they are associated with hydatiform degeneration

in about 75 per cent of the cases.

A ter a careful survey of a considerable number of speci-
mens, both normal and pathologic, ectopic and uterine, of
human conceptuses of widely different ages, Iam led to concur
entirely in the opinion of Minot that the typical vacuolated
cell, as described by Hofbauer, is a degeneration product, though
usually not a degenerate erythroblast, as Minot concluded. Rarely
have I seen a chorionic vesicle in which the rather small, clear,
isolated Hofbauer cells scattered throughout the stroma of a
villus undoubtedly were erythroblastic in origin. In these villi
capillaries in various stages of disintegration were present, and
the erythroblasts could be traced directly to these degenerate
capillaries. In the earlier stages of this degeneration these
degenerating erythroblasts are not surrounded by spaces, how-

PLASMA CELLS OF HOFBAUER 341

ever, and this is true also of early stages in the degeneration of
the fixed or already detached mesenchyme cell, which later
forms the typical, degenerating, wandering cell. However, it
represents but one stage in this degeneration.

It is significant that, although Hofbauer suggested that these
cells might have a digestive or assimilative function, he, too,
frequently found fragmentation of the nuclei and complete dis-
appearance of the cytoplasm and even of the cell itself. All
stages of degeneration, as manifested by crenation of both cyto-
plasm and nucleus, even to complete disappearance of the cell,
can easily be found. Signet-ring forms are common, and the
nuclei are found in all stages of extrusion and degeneration.
The cell boundaries are often ragged, the nuclei crenated and
pyenotic, the cytoplasm granular, vacuolated, webbed or fenes-
trated, until finally nothing but a faint ring or shadow form
without a trace of a nucleus remains. However, in these trans-
parent or shadow forms the nuclei, if not previously extruded
or dissolved, are frequently represented by a mere outline or
by a faint trace of one. Since all stages between the latter and
the well-preserved cells, without vacuoles and well-preserved
nucleus and cytoplasm, and also with processes, occur in good
material, one can scarcely doubt their origin.

Undoubted instances of mitoses were never seen in any Hof-
bauer cells, no matter how well preserved. This no doubt can
be accounted for by the fact that from the time the mesenchyme
cells retract their processes and become isolated in the ground
substance of the villus, they are in a stage of degeneration.
Under such circumstances one would hardly expect to see in-
stances of cell division, although these possibly may be simu-
lated by necrobiotic phenomena.

Hofbauer (’05), and also in his first publication, stated that
the cells described by him increase by mitoses which are fre-
quent. He also found examples of what seemed to be instances
of pluripolar mitoses, but also noted fragmentation of the nuclei.
Acconci (714) also found mitotic figures in cells designated lipoid
interstitial cells by him, but most investigators say nothing
about this. On the contrary, a number of them specifically

342 ARTHUR WILLIAM MEYER

state that they could not find an actual increase in the number
of nuclei present in the stroma of villi containing large numbers
of these cells. Furthermore, every one except Muggia (and
also he in his description and illustrations, as also Acconci) has
noted characteristics, and described the cells in such a way as
to suggest the presence of degenerative changes. When at all
distinct they are of various shapes and sizes, and are surrounded
by a relatively large clear zone. Their occurrence is erratic
and they contain lipoid granules or vacuoles, and have nuclei
varying considerably in size, position, and staining reaction,
as does also the cytoplasm. They are most frequent in degen-
erate villi and not infrequently lie in detritus. The better pre-
served the stroma the fewer one finds, and in these observa-
tions on this rather large series of chorionic vesicles, some of which
were obtained fresh—one living—in hysterectomly specimens,
I have only found a few instances of what possibly could be
regarded as mitotic figures. Since almost all are agreed that
these cells are of mesenchymal or connective-tissue origin, it
is easy to see that considerable difficulty must be encountered
in deciding just when to regard a mesenchyme cell, which is
the precursor, as a Hofbauer cell. However, since I have not
made this aspect of the question a particular subject of investi-
gation, I have no other evidence to offer.

Since some of these cells, during the early period of degenera-
tion, after they have become quite circular in outline and the
nucleus has taken an eccentric position, have.a decidedly granu-
lar or even a lumped cytoplasm, the confusion with plasma cells,
or their earlier de:ignation as granular wandering cells, need
not surprise us. Nevertheless, the term plasma cells is hardly
applicable, as many of them are not granular. Moreover, no
one has shown that in fixed preparations these cells take the
stains specific for plasma cells. Indeed, although he stained
material with borax methylene-blue after Jadassohn, Happe
(06) did not find any of the Hofbauer cells impregnated. It
must be remembered, however, that failure to stain may be
dependent very largely upon the degree o’, degeneration which
the particular cells have undergone, for, as already stated, Hof-

PLASMA CELLS OF HOFBAUER 343

bauer found that in fresh material these cells reacted as plasma
cells to vital stains.

The opinion of Minot that Hofbauer cells are degenerating
erythroblasts probably can be accounted for by the fact that in the
chorionic vesicle from which Minot’s series, showing a progres-
sive degeneration of the latter into the former, was obtained,
it was impossible to distinguish between the two. This diffi-
culty was due partly to the poor state of preservation of the
particular specimen, a larger survey, especially of better mate-
rial, would have revealed the fact that Hofbauer cells are found
in villi, the blood-vessels of which contain no erythroblasts.
Moreover, as will appear later, the distribution of these cells in
the villi is not such as one rightfully would expect if they have
their source in the vessels. However, since the final form of
the typical Hofbauer cell is a mere shadow cell, it necessarily
may be impossible to determine the kind of cell from which this
shadow form arose, for, as is well known, the end forms in
processess of degeneration of many different types of cells are
indistinguishable. Consequently, a group of swollen, highly
vacuolated cells also may contain among them degenerated,
nucleated red blood cells, as Minot held. Indeed, degenerat-
ing erythroblasts which are indistinguishable from some Hof-
bauer cells, can be seen occasionally not only in the vessels,
but in the heart and also within the cavity of the chorionic ves-
icle; but such occurrences do not prove that the Hofbauer
cells of the villi arise from erythroblasts. That this usually
is not the case follows also from the fact that well-preserved, non-
vacuolated Hofbauer cells occur in villi which have not become
vascularized or which, as stated above, no longer contain ves-
sels. It is true that it often is impossible to distinguish be-
tween degenerate erythroblasts within the vessels and Hof-
bauer cells lying outside of, even if near to them, in the stroma
of the villus. This difficulty is entirely avoided by examin-
ing the older specimens without nucleated reds, for, since Hof-
bauer cells always are nucleated except in their very last stages,
confusion with nucleated cells thus is avoided.

THE JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

344 ARTHUR WILLIAM MEYER

Although the elimination of the erythroblast as the source
of the Hofbauer cell was thus very easy, some difficulty was
encountered, strangely enough, with regard to polymorphonu-
clear leucocytes. This largely is due to the fact that the nu-
cleus of the latter often ceases to be polymorphous as the cell
degenerates. Instances of this kind are quite common, espe-
cially in the membranes of hemorrhagic or infected abortuses.
They are, however, also met with in the decidua. Since the
polymorphous character of the nuclei of these leucocytes usually
can be recognized without difficulty in degenerated accumula-
tions of pus, I was at first predisposed against regarding a circu-
lar nucleus as possibly polymorphous in origin, but careful
scrutiny of numerous specimens in which these misleading de-
generation forms occurred soon left no doubt as to the facts.

As stated above, Hofbauer cells were found in the cavity of
the chorionic vesicle in abortuses which contained blood or had
become infected. In these specimens the degenerated poly-
morphonuclear leucocytes usually lie in groups, or more com-
monly in rows along the inner borders of the chorionic mem-
brane, or in long narrow clefts or folds of the same. Some also
were scattered about among the degenerating erythrocytes,
but an examination of the contained blood usually surprises
one by the entire absence not only of well-preserved poly-
morphonuclear leucocytes, but of all leucocytes whatsoever.
This, to be sure, is in marked contrast to what is found in the
case of ordinary hemorrhages and is a fact full of significance
for the question under discussion. Most of the degenerated
polymorphonuclear leucocytes, many of which contain un-
doubted evidence of phagocytosis, possess a relatively small,
circular, vesicular nucleus which often is eccentric in position.
Others are filled with a granular cytoplasm, or even with very dis-
crete golden granules, while still others are filled with dark,
black pigment granules corresponding in size to the golden ones.
Here and there the field of degenerating erythrocytes may also
be studded with masses of pigment which clearly declare their
origin by the presence of all manner of transition forms, between
the well-preserved, easily recognizable polymorphonuclear leu-

PLASMA CELLS OF HOFBAUER 345

cocytes and the disintegrated pigmented detritus. The phago-
cytic nature of these cells is especially noticeable in the specimens
of young chorionic vesicles with nucleated reds, stained with
iron hematoxylin, for in these the leucocytes are often seen
filled with a mass of nuclei only.

Similar appearances can also be seen occasionally in the decidua
from cases of endometritis, as well as in portions of the decidua
in which the glands have undergone considerable maceration and
degeneration. In the former the polymorphonuclear leucocyte is
the misleading form, while in the latter the degenerating, cast-off
glandular epithelial cells simulate Hofbauer cells in almost every
morphological detail. I have also seen similar specimens of degen-
erated polymorphonuclear leucocytes in ill-preserved hemorrhagic
lymph nodes, especially from cases of septicemia, and, until the true
nature of such degenerate leucocytes became evident, it was very
puzzling to see why the Hofbauer cell, which never was found to
contain evidences of phagocytosis when lying in the stroma of
the villus, should become phagocytic when contained in a de-
generated amniotic or chorionic membrane or when lying in
a hemorrhagic area. Undoubted instances of phagocytic Hof-
bauer cells were never seen, although certain misleading forms
other than those already mentioned were encountered also in
pregnant tubes and in an ovarian pregnancy. Among these
misleading forms were specimens of binucleate cells in which
one nucleus had undergone almost complete chromatolysis,
leaving only a nuclear membrane. These nuclear remnants
or so-called nuclear shadows, can easily simulate a phagocy-
tosed erythrocyte. The same is true of small areas of cyto-
plasm which stain but faintly, and hence look more translu-
cent, and particularly of vacuoles themselves.

Essick (15) found what he regarded as morphologically sim-
ilar cells in transitory cavities in the corpus striatum, and
believed them be macrophages. Consequently, he concluded
that Hofbauer cells also are phagocytic and regarded them as
having an endothelial origin. I have not been able to find any
evidence for the latter origin, however, for in specimens in which
the capillaries are plugged with degenerate endothelial cells

346 ARTHUR WILLIAM MEYER

or in which they are composed of a layer of greatly enlarged
oedematous endothelial cells, so as to make the cross-section of the
vessels look not unlike that of a duct, Hofbauer cells never
were found in close proximity to capillaries or other vessels or
in unusual numbers elsewhere in the stroma of such villi. Nor
did I see any evidence for such an origin in villi from hemor-
rhagic or inflammatory cases, and although Hofbauer cells often
lay near to, or even in extravasations in the villi, they never
were found engorged with erythrocytes or pigmented. Never-
theless, if Hofbauer cells arise from mesenchyme cells, it stands
to reason that they at least may be potentially phagocytic, and
failure to find them so may be accounted for by the fact that
they possess a lowered vitality in consequence of degenerative
changes.

I am prompted to suggest, in connection with the question of
phagocytosis, that, unless we regard the process as other than
an actively vital movement on the part of the cell for the pur-
pose of engulfing things, we have undoubtedly misused the term.
That the mere possession of parts of cells, or even of whole cells
within the cytoplasm, is not sufficient evidence for the posses-
sion of phagocytic activity on the part of a particular cell, seems
to me to be beyond question. In some instances, for example,
degenerating phagocytic leucocytes fuse with each other in
groups of twos, threes or even in greater number, thus forming
multinucleated and not infrequently vacuolated complexes.
Similar phenomena can be seen also among degenerated ery-
throblasts and trophoblast cells. Although it would be incor-
rect to regard these degenerate fusion products as true, living
giant-cells, they nevertheless simulate such very closely indeed.
Moreover, when these larger fusion products fuse with an in-
dividual cell of the kind that gave rise to them, it would be quite
natural to regard them as being phagocytic, while, as a matter
of fact, the process is merely one of degeneration. Another
example of what we may call pseudophagocytosis is that repre-
sented by the isolated erythroblasts rarely seen in the stroma of
a villus. In some instances two or three cells, whose bounda-
ries for the most part still are clearly outlined, can be seen to

PLASMA CELLS OF HOFBAUER 347

have partly fused, forming a so-called giant-cell. All transi-
tion forms and stages can be found, and were it not for this fact,
the resultant large multinucleated fusion product, if seen to join
with an isolated trophoblast cell, might be regarded as being
phagocytic. Other instances of a similar nature were discussed
briefly elsewhere (Meyer,’18), and I am inclined to believe that
the non-vital character of this kind of cell formation, which occurs
under conditions of cell degeneration, needs further emphasis.
It certainly would seem to be a non-vital, rather than a vital
phenomenon. It is indicative of degeneration and death, rather
than of regeneration and life.

Cells which are morphologically identical with certain stages
in the degeneration of the Hofbauer cell can also be found in
entirely different locations than those mentioned. Such in-
stances occur in the Graafian follicle. In some of these, germinal
epithelial cells which have become detached and displaced in
the liquor folliculi become swollen and transparent and the
nucleus takes an eccentric position. In all details of structure
and ordinary staining reaction, as shown by hematoxylin and
eosin, by iron hematoxylin, by van Gieson, and by Mallory,
these cells are identical with phases in the typical Hofbauer
cells. This, however, does not justify us in designating them
as such, unless we wish to extend the use of this name to degen-
erating and disintegrating forms of cells of very many different
types and origins.

348 ARTHUR WILLIAM MEYER

LITERATURE CITED

Acconci, G. 1914 Sulla fine struttura della placenta. Internat. Monatschr.
f. Anat. u. Physiol., Bd. 30, and Folia Gyn., t. 9.

Bertin, Marta 1907 Uber Verinderungen in retinierten Placenten. Inaug.
Diss., Leipzig.

CuaLeTzky, Eva 1891 Hydatidenmole. Inaug. Dissert., Bern.

Essick, CuarLes R. 1915 Transitory cavities in the corpus striatum of the
human embryo. Contributions to Embryology, no. 6, Carnegie In-
stitution, vol. 2, p. 97.

Grosser, O. 1909 Vergleichende Anatomie und Entwicklungsgeschichte der
Eihaiute und Plazenta. 8vo. Leipzig. u. Wien, W. Braumiiller.

Harpe, H. 1906 Beobachtungen an Eihiuten junger menschlicher Lier.
Anat. Hefte, 1. 32, S. 171-209.

Horsavrer, J. 1903 Uber das konstante Vorkommen bisher unbekannter
zelliger Formelemente in der Chorionzotte der menschlichen Placenta.
Wiener klin. Wochenschr., Bd. 16.

1905 Grundziige einer Biologie der menschlichen Plazenta. Leipzig
u. Wien, W. Braumiiller.

KastscHenko, N. 1885 Das menschliche Chorionepithel und dessen Rolle
bei der Histogenese der Placenta. Archiv f. Anat. u. Physiol., Anat.
Abth., 8. 451-480.

KossMaNn, Rossy 1892 Zur Histologie der Chorionzotten des Menschen. Inaug.
Diss. Universitit zu Géttingen. Leipzig, W. Engelmann.

Kworostansky, P. 1903 Uber Anatomie und Pathologie der Placenta.
Syncytium in dem schwangeren Uterus. Wirkung der Herz- und
Nierenkrankheiten auf die Musculatur und Placenta. Atonie des
Uterus, Placentar-Adhirenz, Uterus-Ruptur. Archiv. f. Gynaek.,
Bd. 70, S. 113-192.

LaNGHANS, THEopoR 1877 Untersuchungen iiber die menschliche Placenta.
Archiv f. Anat. u. Physiol., Anat. Abth., S. 188-267.

v. Lennossek, M. 1904 Uber das Chorion-epithel. Rev. in Zentralbl. f. Gyn.,
Bd. 28}.

Matt, FRANKLIN P. 1908 A study of the causes underlying the origin of human
monsters. Jour. Morph., vol. 19.1916 On the fate on the human em-
bryo in tubal pregnancy. Contributions to Embryology, no. 13,
Carnegie Inst. of Washington.

Marcuanp, Franz 1898 Uber das maligne Chorio-epitheliom. Zeitschr. f.
Geb. u. Gyn., Bd. 39.

Mertens, J. Beitrige zur normalen und pathologischen Anatomie der men-
schlichen Placenta. Zeit.f. Geburtsh. u. Gyn. Bd. 30 und 31.

Meyer, A. W. 1919 Hydatiform degeneration in tubal pregnancy. Surg.,
Gyn. and Obst., vol. 28.

1918 A report on over one hundred new cases of hydatiform de-
generation. Amer. Jour. Obst., vol. 78.

Minot, Cuarues 8S. 1889 Uterus and embryo. I. Rabbit. IJ. Man. Jour.
Morph., vol. 2.

1911 Die Entwicklung des Blutes. Keibel-Mall Handbuch, Leipzig
and Philadelphia, 8. 483-517.

PLASMA CELLS OF HOFBAUER 349

Moeara, Virainio 1915 Contributo allo studio della parziale degenerazione
vesiculare della placenta. Folia Gynaecologica, T. 11, Fasc. 3.

Pazzt, M. 1908 Vescicole molari iniziali e nuova orientazione della teoria
patogenetica della mola vescicolare e del corionepitelioma. Folia
Gynaecologica, T. 1, Fasc. 3.

von RECKLINGHAUSEN, F. 1863 Uber Eiter- und Bindegewebskérperchen.
Archiv f. path. Anat. u. Physiol., Bd. 28.

ScHRODER VAN DER KoutK 1851 Waarnemigen over het maaksel van de men-
schlijke placenta. Amsterd., p.49, Taf. V, fig.26. Cited by Virchow.

UxeEsco-Stranconowa, K. 1896 Beitrige zur Lehre vom mikroskopischen
Bau der Placenta. Monatschr. f. Geburtsh. u. Gynaek., Bd. 3.

Vircuow, Rupotr. 1863 Die krankhaften Geschwiilste. Bd. 1, Berlin.
1871 Die Cellularpathologie und ihre Begriindung auf physiolog-
ische und pathologische Gewebslehre dargestellt. 4th ed., Berlin.

Resumido por el autor, A. R. Ringoen.
FE] desarrollo de las glandulas gastricas de Squalus acanthias.

Los primeros vestigios de las glandulas gastricas pueden dis-
cernirse en el epitelio estomacal de los embriones de Acanthias
de 133 mm. de longitud. Dicho epitelio se caracteriza en este
estado por una gran actividad que se traduce por cambios locales.
Grupos de células epiteliales presentan cambios en su reaccién
hacia el colorante empleado (hematoxilina férrica-naranja G).
Tan pronto como una célula epitelial entra a formar parte de
uno de estos grupos, tanto el citoplasma como el nticleo dis-
minuyen su afinidad con el colorante citado. En los embriones
de 137 mm. de longitud el epitelio gdstrico esta’ sembrado de
grupos celulares bien definidos, separados por intervalos regu-
lares. Tales grupos no estan ya formados por células epiteliales
sino que representan los rudimentos de glandulas. Su diferen-
ciacién ulterior comprime los extremos libres proximales de las
células del epitelio. A consecuencia de esta compresién, las
células situadas entre dos rudimentos glandulares adquieren
una disposicién en abanico cuando se observan en corte trans-
versal. Los rudimentos glandulares crecen y penetran en el
tejido mesodérmico subyacente produciendo las glindulas gas-
tricas. A medida que estas invaden el tejido mesodérmico las
células que las constituyen sufren una rotacion, al final de la
cual, las del fondo de la glandula se colocan con sus ejes mayores
formando un angulo recto con la luz glandular. Una glandula
gdstrica de Acanthias completamente desarrollada es una estruc-
tura no ramificada. La diferenciacién de dos tipos celulares no
se lleva a cabo nunca.

Translation by José F. Nonidez
Columbia University

AUTHOR’S ABSTRACT OF THIS PAPHR ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MAY 1

‘THE DEVELOPMENT OF THE GASTRIC GLANDS IN
SQUALUS ACANTHIAS

ADOLPH R. RINGOEN

Department of Animal Biology, and the Institute of Anatomy, University of
Minnesota

SEVEN FIGURES (THREE PLATES)

CONTENTS
[rave oXe LOWOLYoy Neh VOM oe ire eerie es 3 i co en RA ARE es lo P-ReRe oP e dA Are UME Sn RM Eee 351
Eerie BT CeNN Pe A eee ERCP rene ES cha rch etend ietbelba ay rece ECHR SEO ae RG Oe OT fs aya 352
PORE CY INO Tis: Sh Bite 2 CES nd oc oss Qa eels mM, el ASE 2k a, 353
A. Harly ehanges in the gastric epithelium... 2.0... 6..5...5.0200n0008 353
Beehormsation ot dehnitersland snidiments:s..+48) neta. se eee ee 355
C. Influence of the gland rudiment on the epithelium.................. 356
D. Subsequent history of the gland rudiments......................005 307
HP VOcat ONTO fCe lls wMberrs crite tear recs Aris aeciteusscegs's a atte reken ret omens ake edo 359
F. Histology of the fully fumoriaeau SASULC CAN eee set ans fee 362
Discussioniorresults anagiliserature lie 2602s ead AS aoe Ae ee 363
SUIT aTaves Hs ear ¥o sean tenet scenes eyonyan stares, Macioiaiae’s oud wma S82 aap eee 368
BT SISRe yea To RR ec ica Re eRe eve eae Sete CTP Ee en or ERI NE A a a ERNE NS 370
INTRODUCTION

The digestive tract of fishes, although the subject of a con-
siderable amount of research, still affords opportunity for further
study. The very early investigators of the subject were in-
terested in its glands, but no serious attempts: were made to
determine their origin and further differentiation. A study
of the literature shows that there is still comparatively little
written on the histogenesis of these structures. In view of these
facts, the present paper attempts to elucidate the sequence of
events that takes place in the development of the gastric glands
as found in Squalus acanthias.

The material used in this study was placed at my disposal
by the Department of Anatomy, University of Minnesota. For
the study of the glands, the stomach was removed from the

351

352 ADOLPH R. RINGOEN

specimen, embedded in paraffin, and cut in serial sections 5 y»
in thickness. The sections were stained with iron-haematoxy-
lin. Both erythrosin and orange G were employed as counter-
stains. For a large part of the prepared Acanthias material
I am personally indebted to Dr. Richard E. Scammon, who
has so kindly permitted me to use numerous series from his
private collection (S. C.).

It is with great pleasure that I express my appreciation to
Dr. Scammon for suggesting this investigation, for the loan
of material, and for the interest shown during the progress of
this study. I also wish to express my indebtedness to Mrs.
Helen Sanborn Chapman for the accurate drawings.

LITERATURE

Our conceptions of the histogenesis of the glandular elements
of the digestive tract are based largely on observations made on
the study of mammalian material. A comprehensive review
of the literature bearing on the subject would be foreign to the
purpose of this study. Later an attempt will be made to in-
dicate the present status of our knowledge in this field, in so
far only as it may be essential to a clearer understanding of
gland formation in Squalus acanthias, by brief reference to a
few papers.

Sprott Boyd (’36) made the first observations on the presence
of gastric glands in mammals and fishes. Following Boyd’s ob-
servation, Bischoff (’38) studied the mucous lining of a great
many species of fishes. In some species he was unable to find
glands, while in others they were abundant.

In 1852 Leydig discovered gastric glands in Squatina an-
gelus and Torpedo galvani. Later he referred to these glands
as ‘Labdriisen,’ thinking that they were comparable to the
gastric glands of mammals.

Edinger (’77), in studying the mucous membrane of the
stomach of fishes, was unable to distinguish the chief and cen-
tral cells as discovered by Heidenhain and Rollet in the mamma-
lian stomach. According to Edinger, among the Teleosts there
are a number of forms that possess no glands in the stomach.

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS 353

He believes that the gastric glands appear phylogenetically
first in the selachians.

The literature on the glands of the Selachii has been to a
large extent reviewed by Oppel (’96), but one finds very little
information relative to the origin of the gastric glands themselves.

Sullivan (’07), in the course of a study devoted to the diges-
tive tract of Elasmobranchs, considers primarily the physio-
logical features of the glands. He makes no comment upon
the development of the glands.

Peterson (’09) is, as far as I am able to ascertain, one of the
few investigators to consider in some detail the development
of the gastric glands in selachians. In his studies on the
histogenesis of the glands in a number of these forms he de-
scribes and figures epithelial outgrowths as the rudiments of
glands.

OBSERVATIONS

In Squalus acanthias, gland development proceeds in a very
different manner from that commonly described for a number
of mammals, including man. The conditions in the gastric
epithelium of embryos 133 mm. in length indicate that in the
selachians gland formation is not associated with the formation
of gastric pits, as has been claimed for man. Since Acanthias
specimens of this length show the first traces of gland dif-
ferentiation, I begin with this stage and follow out in later stages
the complete evolution of a gland.

A. Early changes in the gastric epithelium

The early stages in the development of gastric glands are
clearly followed in Squalus acanthias embryos 133 mm. long.
In fact, the very beginnings of the glands are discernible here
as differentiating in the gastric epithelium itself. At this par-
ticular stage the epithelium of the stomach is characterized
by its great activity in the way of undergoing definite local
changes throughout its entire extent. Prior to the 133-mm.
stage there were no apparent variations or irregularities in it;

354 ADOLPH R. RINGOEN

all of its constituent cells- presented similar morphological fea-
tures and identical staiing reactions. The nuclei also pre-
sented their own characteristic configuration and staining re-
actions. With the establishment, however, of local changes
in the gastric epithelium (133-mm. embryos) there appear cer-
tain definite modifications in its cellular make up. Figure 2,
from a 133-mm. embryo, shows a number of these modifications.
The two cells on the extreme right and left, respectively, of this
figure represent the typical columnar epithelial cells so char-
acteristic of the selachian stomach. They are long and narrow,
and are without a basement membrane. The cytoplasm stains
a faint grayish tint with iron haematoxylin-orange G. The
nuclei are long and slender; they present a deeper gray tint
than does the cellular cytoplasm. The four cells just described
are destined to remain as epithelial cells. Interpolated between,
however, are a number of other cells which are characteris-
tically different from the former in both their cytoplasmic and
nuclear staining reactions. A number of these cells may still
show a close relationship to the neighboring epithelial cells
as exemplified by their behavior toward the iron haematoxylin-
orange G stain. In most cases, however, the cytoplasm and
the neuclus do not stain in the same manner as in the
adjacent epithelial cells. The whole tendency of staining vari-
ations in this direction is for an epithelial cell to decrease
in its avidity for the stain as soon as it is called upon to
assume a different rdle from that of an ordinary epithelial cell.
At times such a change in the staining reactions of the cell body
may precede somewhat that of corresponding changes in the
nucleus or vice versa, and again the changes may proceed rather
synchronously in both the cytoplasm and the nucleus.

In addition to the above-described staining reactions, there
is also a further change in the form of the nucleus, as shown
in figure 2. On the right-hand side of the figure, in the third
cell from the margin, is a nucleus which has not changed its
staining reactions, although the cell body has progressed to
some extent in that direction. It is apparent from the figure
that the nucleus is undergoing a change in its shape. No longer

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS 355

does it correspond in its configuration to the long narrow epi-
thelial nucleus contained in the cells at the extreme right and
left of the figure. A decided change in shape is clearly seen
in those cells which have changed both their cytoplasmic and
nuclear staining reactions.

If the conditions in the gastric epithelium of 133-mm. <Ac-
anthias embryos, as I have described them in figure 2, are inter-
preted properly, they would indicate that in this group gland
development is of a rather primitive nature. No specialized
cells are set apart in early embryonic life for the origin of glands,
but the general epithelium is endowed with the capacity of
transforming certain patches of its constituent cells into definite
gland rudiments at the proper time. When the proper time
is at hand, small groups of these apparently similar epithelial
cells change their staining reactions and the shape of their nuclei,
and differentiate in another direction. They continue to evolve
in a very definite direction, because they are incapable of giv-
ing rise to any other structure than the particular gland rudi-
ment towards which their potentialities are directed.

B. Formation of definite gland rudiments

The characteristic changes which appear in the gastric epi-
thelium at the 133-mm. stage are even more pronounced in
slightly longer embryos. In 137-mm. specimens the epithe-
lium has been modified only at those points where glands are
to be formed. It is surprising with what regularity the ap-
portioning of the general epithelium into glandular and non-
glandular areas has taken place at this stage. The number of
epithelial cells which intervene between two potential gland
areas is practically the same in all cases. Just what factors
determine the selection of certain groups of epithelial cells as
the precursors of glands to the absolute exclusion of others
obviously can no more be answered than why certain entodermal
cells will differentiate into liver cells, while other similar cells,
at least so in their early stages of differentiation, will give origin
to pancreatic tissue.

356 ADOLPH R. RINGOEN

On comparing figures 2 and 3 (the latter an embryo of 137
mm.), it is apparent that gland rudiments are now sharply
marked off from the neighboring epithelial cells. In embryos
of 137mm. in length, such areas are very numerous. The
entire epithelium of the stomach is literally studded at regular
intervals with them. At this time none of the cells making
up a gland rudiment present staining characters bordering on
those of the epithelial cells. The nuclei present about the same
staining reactions as portrayed in figure 2. There has been,
however, a considerable progressive change in their shape.

Figure 3 shows two gland rudiments embedded in the epi-
thelium. They present such a striking appearance in sections
of the stomach (137-mm. specimens) that one cannot fail .to
notice them. The regularity in their distribution is indeed
striking. Never have I seen similar rudiments lying at the
base of the epithelium. Obviously, there would be no reason
for such a location, since the early gland rudiments are simply
transformed epithelial cells. In the same figure the two gland
cells at the extreme right represent only a portion of a gland
rudiment, due to the plane of sectioning.

That the potentialities of the epithelial cells are by no means
the same is particularly evident in the gastric epithelium of
Acanthias specimens 137 mm. in length. Every gland of the
adult specimen is represented at this stage by its own epithelial
modification or gland rudiment.

C. Influence of the gland rudiments on the epithelium

The differentiation and presence of the gland rudiments in
the epithelium has had a profound influence on the final con-
figuration of the epithelium itself. During the early stages of
differentiation it increases in thickness; at no time is it strat-
ified, although in the early stages of embryonic development
the disposition of the nuclei in several planes simulates strat-
ification. In Acanthias embryos 133 mm. in length, the epi-
thelium consists of columnar cells with their lateral surfaces
closely approximated (fig. 2). The differentiation and pres-

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS 357

ence of gland rudiments have changed this simple uniform
condition. No longer are the epithelial cells arranged in the
form of a single continuous row in embryos 137 mm. long. On
comparing figures 2 and 3, it is evident that as soon as a gland
rudiment is well marked out in the epithelium (fig. 3) it forms
a bottle-like plug. The lower expanded portion of the plug
tends to press upon the neighboring epithelial cells, thus crowd-
ing them closer and closer together. This lateral displacement
of the epithelial cells proceeds unhampered, as they are not
anchored fast by a basement membrane, but may extend freely
down into the underlying mesodermic tissue.!. As a result of
compression exerted by the expanded part of the bottle-like
plug of the gland rudiment, or for a lack of space, the epithelial
cells are closely approximated at their bases. Since the distal
parts of the epithelial cells have not been affected by the mechan-
ical forces involved in the compression phenomena, it naturally
follows that the portion of epithelium intervening between two
gland rudiments takes on a fan-shaped form in cross-section
(fig. 3). This peculiar arrangement of the epithelial cells, as
compared with the simple uniform condition in the earlier
stages, is maintained in the later stages.

D. Subsequent history of the gland rudiments

The subsequent history of the gland rudiments will now be
considered in detail. Acanthias embryos 137 mm. long furnish
very favorable material for such considerations. At this stage
many of the gland rudiments are undergoing a great change in
their length; their constituent cells are growing out into the
mesodermic tissue. During the course of a very short period
of time every gland rudiment will have elongated and burrowed
its way into the underlying tissue.

Figure 4 represents one of the so-called epithelial outgrowths
in the gastric epithelium of an Acanthias embryo 137 mm.

1 Hopkins (95) believes that a basement membrane does not exist in the

ganoids. According to Edinger (’77), fishes possess no basement membrane,
but the epithelium borders directly upon the underlying tissue.

358 ADOLPH R. RINGOEN

long. At this stage the constituent cells of a gland rudiment,
and as illustrated in figure 3, have simply elongated and are
pushing their way out into the underlying mesodermic tissue.
During the pushing and burrowing of an outgrowth down into
the mesodermic tissue there is an actual migration of its cells.
Now the nuclei have increased greatly in size, as compared with
those shown in the gland rudiment of figure 3, and at this stage
they tend to occupy the most distal parts of the outgrowth. -
No doubt the increase in the size of the nuclei at this time—
preparatory to division—is associated with the fact that the
cells which contain them are to grow out still further at a sub-
sequent time, and obviously this requires additional cells (see
fig. 5, also from a specimen 137 mm. long, for documentary
evidence on the further growth of a gland rudiment; increase
in the number of nuclei with a distinct change in their size and
shape, as compared with fig. 4). Cell outlines are just as dis-
tinct here as they were before the gland rudiment cells began
to grow out into the mesodermic tissue (compare ftg. 2 with fig. 3).

In slightly more advanced stages of differentiation than is
represented in figure 4, an outgrowth becomes more or less
tubular inform. This condition is represented in figure 5, also
from a 137-mm. specimen. With still further differentia-
tion, as I shall attempt to show at a subsequent time, this simple
condition is radically changed.

Another point of considerable interest in connection with
figure 5 is the position of the nuclei in the tubular outgrowth.
Those located in the region where the gland rudiment first
burrowed do not seem to present any definite arrangement
in their distribution within the outgrowth. On the other hand,
most of the nuclei which are found in the lower half of the out-
growth show that their long axes are placed parallel to the long
axes of the tube. In slightly more differentiated outgrowths
this parallel arrangement is decidely modified. Even as shown
in figure 5, there is a slight tendency in this direction, for
several of the nuclei are on the verge of shifting their axes.’

2 The nuclei, as will be pointed out later, are not actually changing their
long axes, but the cells containing them are shifting.

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS 359

During the extension of a glandular outgrowth down into
the mesodermiec tissue it meets with a small amount of resistance
for two reasons. In the first place, the constituent cells of an
outgrowth are not anchored fast by any structure comparable
to a basement membrane, and in the second place, the meso-
dermic tissue is of such a loose character that it is simply pushed
back and more or less condensed by the burrowing cells (figs. 4,
556, amd 7).

E. Rotation of cells

Figure 6, from an embryo 137 mm. long, shows a glandular
outgrowth at its maximum length. On comparison with figure
5, it is apparent that the staining reactions of both the nuclei
and cytoplasm are identical. The morphological features of
the nuclei are also the same. The glandular outgrowth is no
longer a tubular structure; the shape of its proximal end has
been characteristically modified. This change of form may be
ascribed to a rotation of cells in the lower two-thirds of the out-
growth. Although cell boundaries are not distinguishable in
this figure, it is apparent on comparison with figure 5, in which
instance cell boundaries are also absent, that in following the
course of the nuclei in figure 6, as compared with those shown
in figure 5, one is simply tracing out the movements of the cells
that contain them. A number of cells, as represented in figure
6, have rotated through an angle of about 90°. The completion
of the rotation process is seen in figure 7 (from a specimen 146
mm. long). In this stage cell boundaries are evident; the long
axes of the nuclei, instead of being placed parallel to the long
axes of the future gland, as they were in the tubular outgrowth
represented in figure 5, are now placed at right angles to the
lumen of the gland. Secammon (15), in the course of a study
devoted particularly to the histogenesis of the selachian liver,
describes a similar rotation of cells in tubule anastomosis. He
does not think that the nuclei shift their axes within the cells,
for he frequently finds, in cases where faint cell boundaries can
be made out, that the cells show the same changes in position
as do the nuclei.

THE JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

360 ADOLPH R. RINGOEN

In order to elucidate more clearly the sequence of events that
takes place in the evolution of a selachian gland from the simple
tubular condition on through to the typical flask-shaped form,
it has seemed feasible to submit a number of diagrams illustra-
tive of the rotation processes. A casual reference to these
diagrams will enable one to comprehend at a glance, without
the aid of detailed figures and descriptions, that an actual rota-
tion of cells play a fundamental rdle in molding the fundiec por-
tion of the Selachian gastric gland.

In figure 1, diagram A, the glandular outgrowth has extended
only a short distance, and in most instances the long axes of
the nuclei are parallel to the long axes of the outgrowth. This
is the condition one would expect to find, since in a slightly
younger stage (fig. 3) the nuclei are also parallel to the long axes
of the gland rudiment. Diagram 1B shows that the glandular
outgrowth has not only increased in length, but that a number
of the nuclei are beginning to change their long axes with ref-
erence to the long axes of the outgrowth. With reference
to these shifting movements Scammon (715) states “that it
is hardly to be considered that the nuclei shift their axes within
the cells.’’ Since cell boundaries were not seen in the speci-
men, illustrated in figure 1B, it may be reasonably assumed
that the shifting movements of the nuclei are but the shifting
movements of the cells which contain them. The next diagram
(fig. 1C) distinctly shows that the cells in the region of what is
to form the future fundice portion of the gland have rotated
through an angle of about 90°. It is very probable that the
cells located at. what may be called the center of the base of
the fundus do not change their positions, since their nuclei
show no shifting movements at any time. Diagram 1D shows
that all of the nuclei have passed through a change in their
original axes of about 90°. The long axes of the nuclei now
lie at right angles to the lumen of the gland. In figure 7 the
outlines of the cells were drawn just as they appeared in the
actual specimen. This shows that the rotation of the nuclei
is to be interpreted only as the shifting movements of the cells
that contain them.

‘uoumdeds *WIW-9F] B UWLOIJ QT pure ‘YYsuo] ul ‘WUT YET SoAIqUID WOT; OT pus qT
‘yjSus] Ul “UIU EET OAIGUIO UB WOIJ UMVIP ST WT ‘OOS X ‘UOT}BOYTUSeUT OURS OY} YIM UMVIP puB
BPION] VIOWIVI OY} YIM PoUT[JNO 919M ToJONU oY} JO []Y ‘Suoutroeds [enjoe WOA, UMVAP oe SUTBISVIP

asoy} JO [[V ‘“PUBLS O11]SvS UBIYOVIOS B JO UOTJN[OAD OY} UL SasBIS MOF SUIMOYS SUIVISBICE [SIT

d V

361

362 ADOLPH R. RINGOEN

F. Histology of the fully differentiated gastric gland

Figure 7 (embryo 146 mm. long) represents the fully differen-
tiated gastric gland. In specimens of this length the vast
majority of glands present such an appearance, although there
may be slight variations in the length of the neck and in the
size of the fundic portion. These slight variations are, however,
of no fundamental importance, for they do not signify that the
glands differ in character.

All of the gastric glands are unbranched, flask-shape struc-
tures. Never have I been able to find the bifurcated forms
described by Peterson(’09). He believes that the neck cells
of the gland are of a rather primitive character during their
early history, and that they may give origin to a side bud
which with further differentiation, gives to the gland a forked
appearance.

At no time have I been able to distinguish more than one
type of cell in a gland (fig. 7). All authors who have investigated
the gland cells of lower vertebrates are quite agreed in the
occurrence of a single cell-type.* According to Edinger (’77),
in the fishes, this single type is homologous neither to the chief
nor to the parietal cell of the mammalian stomach. His con-
clusion has been generally accepted. It appears from the in-
vestigations of numerous other observers that the differentiation
of two the cell-types in the gastric glands of fishes does not
obtain. This specialization probably appears phylogenetcally
much later.

As far as Squalus acanthias is concerned, and as Peterson
(09) maintains for a number of selachians, the neck cells of
a gastric gland are not concerned in the formation of any specific
secretion. This activity appears to be confined to the cells mak-
ing up the fundic portion of the gland. In all the cases that
have come under my observation, I have found that the secre-
tion granules are elaborated exclusively in these cells. As to
the precise character of these granules, I have no definite
knowledge.

3 Oppel (96) agrees with Edinger in the finding of only one type of cell in

the glands of fishes, but states that this single type appears to possess relation-
ships with the parietal cell of mammals.

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS 363

DISCUSSION OF RESULTS AND LITERATURE

Miss Ross (’02), in her investigations on the development
of the gastric glands in Desmognathus, Amblystoma, and
pig, finds that the round granular cells which give origin to the
glands appear as distinct cells with the differentiation of the
entoderm. These gland cells occur at the base of the epi-
thelium in close contact with the basement membrane.

In addition to Miss Ross (’02), Toldt (’81) also maintains that
the beginnings of the gastric glands in the cat are represented
by granular cells, which are interpolated at the basal parts
of the epithelial cells. There are, however, numerous other
investigators who ascribe the origin of glands to epithelial pro-
jections, or outgrowths,‘ among whom may be mentioned K@6l-
liker (52), Brand (78), Griffini and Vassale (’88), Oppel (’96),
Minot (’02), and Peterson (09). Johnson (’10) finds in a human
embryo of 120 mm. that the beginnings of the gastric glands
appear as knob-like outgrowths at the bottoms of the gastric
pits. It is generally admitted by those who believe that the
glands arise from downgrowths of the surface epithelium, that
the gastric pits—later elongated to form grooves—give origin
to the glands in this manner.

In specimens of Squalus acanthias there are no gastric pits;
the longitudinal folds, which are presumably the forerunners
of the villi in higher forms, are variable in their number, location,
and size. In all probability, their presence is due to the con-
traction of the muscular coats. Consequently, gland develop-
ment must go on in a very different manner from those forms
where all these features can be readily identified. Peterson
(09) states that in a number of selachians he has found that
they arise from epithelial outgrowths directly. With regard
to the origin of the outgrowths themselves, and their further
history as based more particularly on a study of Acanthias
vulgaris, he states:

4 Bensley (’00) finds in a Urodele larva 11 mm. in length that the oesopha-
geal glands also appear as tubular downgrowths of the foregut entoderm. He

believes that these glands represent gastric glands whose development has been
arrested.

364 ADOLPH R. RINGOEN

es erreichen nicht alle Zellen die freie Oberfliche, son-
dern viele bilden eine untere Schicht und entbehren des Pfropfes.
Diese lassen die Magendriisenzellen aus sich hervorgehen. Sie wandern
aus, ein kurzer, halbkugeliger Fortsatz ragt aus dem Epithel heraus,
vergrossert sich, der Kern riickt nach. Andere dicht daneben lie-
gende tun dasselbe, so dass eine Knospe an der Epithelbasis zum
Vorschein kommt. Die Zellen bleiben mit einem lang ausgezogenen
Ende zunichst noch mit dem iibrigen Epithel in Verbindung. Immer
mehr Zellen riicken nach und drangen die erst ausgewanderten
weiter :

Since I have ‘already discussed the origin of gland rudiments
in Squalus acanthias specimens, no further attempt will be made
here to consider them in detail. Suffice it to say that in speci-
mens 133mm. long, gland rudiments differentiate in the gas-
tric epithelium itself—not from special embryonic cells set apart
with the early differentiation of the entoderm (as Miss Ross
(02) maintains for Desmognathus, Amblystoma, and pig),
—but from typical epithelial cells (fig. 2).

It may be remarked that Peterson’s observation that in
Acanthias vulgaris embryos of 55 to 70mm. in length not all
of the epithelial cells reach to the free surface—lumen side of
the stomach—‘‘sondern viele bilden eine untere Schicht und
entbehren des Pfropfes. Diese lassen die Magendriisenzellen

aus sich hervorgehen . . . ,” is not in accordance with

what I find in the epithelium of Squalus acanthias. For it
is clearly seen in embryos of less than 133 mm. in length that
all of the epithelial cells at this stage present the closest morpho-
logical features and staining reactions. They are all of equal
length, and, therefore, they all reach to the surface.2 Even

° In support of my results I wish to quote from Kirk’s (’10) paper with par-
ticular reference to Toldt’s (’81) work on the rudiments of glands in the fundic
portion of the cat’s stomach as large, eosinophilic cells interpolated at the base
of the surface epithelial cells. ‘“Toldt is sure these cells are of epithelial origin,
but believes they at no time reach the surface, being always shut off from the
latter by the overhanging distal ends of the tall pyramidal surface epithelium;
he suspects that they arise from young Ersatzzellen. His Ersatzzellen have
almost certainly been shown by the work of Stéhr (1882) and Bizzozero (1888)
to be ‘Wanderzellen.’ Griffini and Vassale maintain that Toldt’s figures and
text harmonize remarkably with their findings, except that Toldt, through use
of oblique sections, erroneously concluded that these primary gland cells do not
reach the surface, and that their lumen is thus not at first continuous with the

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS 365

at the 133-mm. stage, and as shown in figure 2, all of the epi-
thelial cells still reach to the surface. There are, however,
considerable changes in the staining reactions of small groups
of them. ‘These changes are very significant, for they show
that, in spite of their morphological similarities, not all of the
epithelial cells are endowed with the same specificities for fur-
ther differentiation. Small groups of them will remain as epi-
thelial cells, while other groups change their staining reactions,
and the shape of their nuclei, and finally form a bottle-like plug
which is embedded in the epithelium (fig. 3). All of these
changes were considered in detail in connection with figures
2 and 3.

Since Peterson maintains that the gastric glands grow out
from portions of the epithelium whose cells do not reach the
surface, it would be interesting to know how he would account
for certain nuclear variations. In figure 5 the nuclei are numer-
ous in the glandular outgrowth. On comparing one of them
with the epithelial nucleus on the left-hand side of the same
figure it is apparent that there are fundamental differences
with reference to their form, and also in the amount and distribu-
tion of their chromatin material. These differences were also
mentioned in connection with figures 2 and 3. How would
Peterson explain such differences? As far as the nuclear dif-
ferences in figures 2 and 3 are concerned, he woud be unable
to give any logical explanation, because he failed to find similar
stages in his material. Now, if the gastric glands in Selachians
really arise by means of epithelial outgrowths, as Peterson
holds, one would naturally expect to find similar nuclei in both
the outgrowths and the general epithelium. Figures 4 and 5,
however, show that this is not the case. They are very dif-
ferent in their form, size, and in their arrangement of chro-
matin material. This is precisely what one would anticipate
after studying stages similar to the ones depicted in figures 2
and 3 of this paper.
stomach lumen. Griffini and Vassale found many such groups with lumina

apparently shut in on all sides, but reconstructions always demonstrated con-
tinuity with the stomach lumen from the first.’’

366 ADOLPH R. RINGOEN

With reference to the formation of the gland lumen, I find
that it arises in the manner described by Peterson. The early
formation of it is frequently seen as a slight indentation on the
free surface of a well-defined gland rudiment, shortly after it
has pushed out into the underlying mesodermic tissue. The
further extension of the gland rudiment is followed by a similar
extension of the gland lumen. As to whether the extension
of the lumen into the fundic portion is responsible for the shift-
ing movements of the cells, I am unable to say. At all events
the extension of the lumen into the fundic portion of the glands
does not cause the cells there to take on a ‘schriage’ position
with reference to the lumen, as Peterson maintains I have
already pointed out that in the fully differentiated gastric gland
the cells at the fundic portion are placed at right angles to the
lumen of the gland with reference to their long axes.

In spite of his observations relative to the origin of gastric
glands in selachians, Peterson is at a loss to account for the
great numbers of these structures that he finds in old specimens.
To quote from his paper:

Wie erfolgt nun die weitere Vermehrung der Driisen? Bei einem
jungen Tiere hat die Schleimhaut dasselbe Aussehen wie bei einem
alten. Es kommen genau so viel Driisen auf den Quadratmillimeter;
auf einer Schnittstrecke von 1mm. (Schnittdicke 10 uw) liegen quer-
wie lingsgeschnitten 20 Schnitte 1m Durchschnitt, also 400 auf den
Quadratmillimeter. Ein Magen eines jungen Tieres (Magenlinge
5em.) habe n qmm Fliche, der eines alten (Magenlinge 10 cm.)
sagen wir dann 4 n qmn Flaiche, so kommen auf den einen Magen
400 n, auf den anderen 1600 n Driisen, wo kommen diese 1200 n
weiteren Driisen her?

In order to account for these 1200 new glands Peterson thinks
that there may be a longitudinal splitting of the young glands.
As a result of this cleavage process there are now two glands
where before there was only one. He also thinks that ‘ Ver-
zweigungen der Driise in den mittleren Partien kénnten als
Anhaltspunkt dienen”’ for increasing the numbers of the glands.
Furthermore, Peterson thinks that the neck cells of a young
gland may contribute to the formation of additional glands:
‘Sie haben wir oben als indifferente Reste der alten Anlage

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS Gt

kennen gelernt, und sie kénnten also auch spiterhin Driisen-
knospen entwickeln. Die anderen Zellen der basalen Reihe
sind vollkommen verschwunden—aufgebraucht.’? He _ states,
however, that he has no proof for such activity on the part of
the neck cells.

_Peterson’s statement, ‘‘bei einem jungen Tiere hat die Schleim-
haut dasselbe Aussehen wei bei einem alten,’’ is absolutely
not in accordance with my observations. Figures 2 and 3
show that there are decided morphological and staining dif-
ferences, and in these specimens, from which the drawings were
made, there is only a difference of 4mm. in length. In figure
2 (embryo 133-mm. long) the epithelium is undergoing definite
changes, while in figure 3 (137-mm. long) it shows more decided
changes—the rudiments of glands.

It is unnecessary to build any elaborate theories with ref-
erence to the manner in which the number of glands is increased,
as Peterson has attempted to do. Any attempt with such an
aim in view is diametrically opposed to what one finds in the
gastric epithelium of Acanthias specimens 137 mm. in length.
At this stage the epithelium has been modified at only those
points where glands are to be established, and, indeed, it is
surprising with what regularity and precision these changes
occur. The entire epithelium is literally studded at regular
intervals with them. Every gland of the adult animal is rep-
resented at this stage by its own epithelial modification, or
gland rudiment. In specimens 146mm. long and above this
length, all the rudiments have given origin to glands, and the
neck portion of every gland occupies the same relative position
in the epithelium as did the early gland rudiment. The entire
process of glandular development takes place in just as orderly
a manner as does the differentiation and development of any
other vertebrate organ or structure.

Peterson’s entire difficulty in failing to be able to account
for the vast numbers of glands in old specimens is easily
explained. Although he saw numerous epithelial buds (‘Knos-
pen’) in young specimens, they were by no means sufficiently
numerous to account for all the glands in the adult specimen.

305 ADOLPH R. RINGOEN

Undoubtedly the specimens that he examined showed com-
paratively few buds. I have also frequently found that in em-
bryos 137mm. in length they are not especially numerous.
Therefore, since Peterson did not discover the precursors of
the buds themselves, or the gland rudiments as I have called
them throughout this paper, he is tempted to theorize with
reference to the manner in which the number of glands is
_increased.

SUMMARY

1. In the selachians, as represented by Squalus acanthias,
no specialized cells are set apart in early embryonic life for
the formation of glands (as described by Miss Ross for Des-
mognathus, Amblystoma, and pig), but the general epithelium
is endowed with the capacity of transforming certain groups
of cells into definite gland rudiments.

2. The apportioning of the gastric epithelium into glandular
and non-glandular areas is evident in Acanthias embryos 133 mm.
in length. At this stage the epithelium is undergoing local
modification in that small groups of its constituent cells change
their staining reactions (iron haematoxylin-orange G) as com-
pared with the adjacent epithelial cells (fig. 2). As soon as an
epithelial cell is called upon to become a contributory member
toward the formation of such a group of cells, both its cyto-
plasm and nucleus decrease in their avidity for the above-men-
tioned stain.

3. In addition to this change of staining reactions, there is
a further change in the morphology of an epithelial nucleus.
Many a nucleus is seen in the process of changing its shape
from the typical narrow, elongated type, so characteristic of
the young epithelial cell, to the plump nucleus of a gland rudi-
ment cell.

4. That the potentialities of the epithelial cells are by no
means the same is particularly evident in the gastric epithelium
of Acanthias specimens 137 mm. in length. At this stage the
epithelium is studded at regular intervals with well-defined
groups of cells. These are no longer to be considered as epi-
thelial cells, but the rudiments of glands (fig. 3).

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS 369

5. The presence and further differentiation of the gland rudi-
ments in the epithelium play an important part in causing the
proximal free ends of the epithelial cells to be laterally com-
pressed. As a result of compression, the epithelial cells inter-
vening between two gland rudiments are forced to take on a
fan-shaped arrangement in cross-section (fig. 3).

6. Acanthias embryos 137 mm. in length show that many of
the gland rudiments are growing out into the underlying meso-
dermic tissue to form the future gastric glands (figs. 4, 5, and 6).
In the course of a very short time all the gland rudiments have
given origin to glandular outgrowths. Peterson recognized
similar outgrowths, but he thinks that they arise from certain
epithelial cells which do not reach to the free surface—lumen
side of the stomach. His interpretation is obviously incorrect
because, on the basis of it, he is unable to account for all the glands
in older specimens. Peterson failed to discover stages similar
to those shown in figures 2 and 3, and these are the critical stages
in establishing the number of glands for any given specimen.

7. As the glandular outgrowths invade the mesodermic tissue
there is an actual rotation of their cells (fig. 6 and 7). In many
outgrowths it is not possible to distinguish cell boundaries, but
in instances where they are evident, it is apparent that the
shifting movements of the nuclei are but a rotation of the cells
that contain them (Scammon, 715). The rotation processes
are confined chiefly to the cells at the lower two-thirds of the
glandular outgrowth. The fundic portion of the gland takes
on a flask-like form as a direct result of these rotation proc-
esses (figs. 6 and 7).

8. With the completion of rotation, the cells at the fundic
portion of the gland are placed, with reference to their long
axes, at right angles to the lumen of the gland (fig. 7).

9. The fully differentiated gastric gland of Squalus acanthias
is an unbranched, flask-shaped structure. The differentiation
of the two cell-types does not occur.

370 ADOLPH R. RINGOEN

BIBLIOGRAPHY

Benstey, R. R. 1900 The oesophageal glands of Urodela. Biol. Bull., vol.
2.
1902 The cardiac glands of mammals. Am. Jour. Anat., vol. 2.

Berry, J. M. 1900 On the development of villi in the human intestine. Anat.
Anz., Bd. 17.

Biscuorr, Tu. W. L. 1838 Uber den Bau der Magenschleimhaut. Millers
Arch. f. Anat. und Phys.

Boyp,S. 1886 On the structure of the mucous membraneof thestomach. Edin.
Med. and Sur. Jour., vol. 46.

BRAND, E. 1877 Beitraige zur Entwicklung der Magen- und Darmwand. Diss.
Wiirzburg, 1877.

Epincer, L. 1877 Uber die Schleimhaut des Fischdarmes, nebst Bemerk-
ungen zur Phylogenese der Driisen des Darmrohres. Arch. f. mikr.
Anat., Bd. 13

Grirrint, L., unpD VassaLn, G. 1888 Uber die Reproduktion der Magenschleim-
haut. Beitrage zur pathol. Anat. u. zur allg. Pathol., Bd. 3.

Hopkins, G. 8. 1895 On the enteron of American ganoids. Jour. Morph.,
vol. 11

Jounson, F. P. 1910 The development of the mucous membrane of the
oesophagus, stomach and small intestine in the human embryo.
Amer. Jour. Anat., vol. 10.

Kirk, E. G. 1910 On the histogenesis of gastric glands. Am. Jour. Anat.,
vol. 10.

Leypic, F. 1852 Beitrige zur mikroskopischen Anatomie und Entwicklung
der Rochen und Haien. Leipzig.

Minot, C. 8. 1892 Text-book of human embryology. New York.

OppEL, A. 1896 Vergleichende mikrosk. Anatomie der Wirbeltiere, Erster
Teil, Jena, 1896.

Peterson, H. 1909 Beitrige zur Kenntnis des Baues und der Entwickelung
des Selacherdarmes. Jena. Zeitschr. f. Naturwiss., Bd. 44.

Ross, M. J. 1902 The origin and development of the gastric glands of Des-
mognathus, Amblystoma and pig. Biol. Bull., vol. 4.

Scammon, R.E. 1911 Normal-plates of the development of Squalus acanthias.
Heft 12, Normentaf. d. Ent. d. Wirbeltiere, Jena.
1915 The histogenesis of the Selachian liver. Am. Jour. Anat., vol. 17.

Sutuivan, M. X. 1907 The physiology of the digestive tract of elasmobranchs.
Bull. U. S. Bureau Fish., vol. 27.

Toutpt, C. 1881 Die Entwicklung und Ausbildung der Driisen des Magens.
Sitzungsber. d. k. k. Akad. d. Wissensch., math. naturw. KI. Abt. 3,
Jahrg. 1880. Wien.

PLATES

PLATE 1
EXPLANATION OF FIGURES

The figures are all taken from the gastric epithelium of embryos fixed in a
mixture of 10 per cent formalin and 2 per cent chromic acid.

2 The earliest recognizable stage in gland development of an embryo 133
mm. long. Iron haematoxylin-orange G stain. > 1000.

3 The formation of definite gland rudiments in the epithelium of a specimen
137 mm. in length. Iron haematoxylin-orange G stain. »X 1000.

4 QOutgrowth of a gland rudiment and extension into the underlying meso-

dermie tissue in an embryo 137 mm. long. Iron haematoxylin-orange G stain.
x 1000.

372

PLATE 1

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS

ADOLPH R. RINGOEN

373

PLATE 2

EXPLANATION OF FIGURES

5 The glandular outgrowth has assumed a tubular form in a specimen 137

mm. in length. Iron haematoxylin-orange G. X 900.
6 Shows the change in form of a glandular outgrowth as based on a rotation
of cells in an embryo 137 mm. in length. Iron haematoxylin-orange G. > 900.

374

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS

PLATE 2
ADOLPH R. RINGOEN

ra)

375

PLATE 3
EXPLANATION OF FIGURE

7 The fully differentiated gastric gland of a specimen 146 mm. long. Iron
haematoxylin-erythrosin. X 1170.

BYES)

ORIGIN OF GASTRIC GLANDS OF ACANTHIAS PLATE

ADOLPH R. RINGOEN

Resumido por el autor, Gilman A. Drew.
Actividades sexuales del calamar, Loligo pealii (Les.)
II. El espermatéforo: su estructura, eyaculacién y formacion.

Un espermat6foro esta’ compuesto de una masa de espermato-
zoides, una masa de material de cemento, un grupo de mem-
branas que forman en conjunto un aparato eyaculador y vainas
para cubrir la masa de esperma, y tunicas envolventes que sus-
ministran fuerza para la eyaculacién. Durante esta ultima la
masa de esperma y el cemento son expulsadas por un tubo que
se evagina en el extremo oral del espermatéforo. La masa de
esperma se rodea de ciertas membranas que forman el reservorio
espermatico y se fija sobre la hembra por una masa de cemento
que se acumula en el extremo del reservorio espermatico. El
esperma sale por un orificio situado en el extremo libre del
reservorio. Los espermatéforos se forman en el 6rgano esper-
matoférico, porcién especializada del espermoducto. Este 6r-
gano presenta un cierto numero de divisiones, cada una de las
cuales forma una porcién del espermatdéforo. La cantidad de
esperma suficiente para cada uno de estos entra en el 6rgano de
una vez. Por accion ciliar este filamento de esperma se arrolla
en una masa espiral cilindrica y apretada que prosigue girando
sobre su eje longitudinal durante su paso por el 6érgano. A esta
estructura se suman materiales producidos en ciertas partes del
organo y de este modo se arrollan membranas y tunicas alrededor
del espermatoforo en vias de formacién. Después que este se
ha formado por completo, la tunica externa se contrae produci-
endo la turgescencia de dicha estructura. Los espermatdéforos
una vez formados se acumulan en el saco espermatoférico, en
donde permanecen hasta que son utilizados.

Translation by José F. Nonidez
Columbia University

AUTHOR'S ABSTRACT OF THIS PAPER ISSUBD
BY THE BIBLIOGRAPHIC SERVICE, APRIL 7

SEXUAL ACTIVITIES OF THE SQUID LOLIGO PEALII
(LES.)

Il. THE SPERMATOPHORE, ITS STRUCTURE, EJACULATION AND
FORMATION

GILMAN A. DREW

Marine Biological Laboratory, Woods Hole, Massachusetts

FORTY-ONE FIGURES (SIX PLATES)

The spermatophores of Cephalopods have been mentioned
in zoological literature by many writers, but most of the ac-
counts are short, incomplete, and inaccurate, so there seems to
be no real need for reviewing the literature as a whole.

By far the best account of the structure and ejaculation of
spermatophores of Cephalopods I have seen is given by Emile-
G. Racovitza, for Rossia macrosoma (’94) in a paper dealing
with the habits and reproduction of the species. His descrip-
tions and figures of the spermatophores are much more complete
and accurate than those of earlier writers and there seems to be
nothing added by later writers of very great importance.

To follow the method by which so complicated a structure is
formed by added secretions on the inside of a duct, it is neces-
sary to have rather elaborate figures representing the structure
and ejaculation of the spermatophores, so it has seemed best to
consider the whole question of structure, ejaculation, and forma-
’ tion together. A previous paper (Drew’11) on the sexual ac-
tivities of the squid deals with the copulation, egg-laying, and
fertilization, and might well follow the observations given
in this paper as it deals with the use made of the completed
spermatophores.

Part of the laboratory work, which forms the basis of this
paper, was done at the University of Maine while I was connected

379

380 GILMAN A. DREW

with the Biological Department of that university, and I am
grateful for working space furnished me by the University of
Arizona during the winter of 1917. By far the greater part of
the work has been done at the Marine Biological Laboratory
at Woods Hole, Massachusetts, and all of the material has been
obtained at that station.

This species of squid is very abundant in the vicinity of Woods
Hole, and any mature male taken from early in the spring until
as late as September, and frequently later than that, is sure to
have an abundance of spermatophores in the spermatophoric
sac.

The spermatophores vary in size according to the size of the
animal from which they are taken. Those from very small
animals may not be over 8 mm. in length and those from large
animals may be as much as 16 mm. in length. They are of
course all similar in structure, but the small ones are softer and
are not so easily handled in ejaculation observations as the larger
ones. As might be expected, there are some slight individual
variations in shape and size of parts in the spermatophores of
different individuals. The spermatophores of each individual
are practically identical in shape and appearance, but may
vary slightly in size.

The number of spermatophores carried by each individual
varies with the size of the animals (the smaller having fewer than
the larger), with the season of the year, and with the frequency
of copulation. May and June are probably the months when
sexual activity is at its greatest. Usually at this time the sper-
matophoric sacs are gorged with spermatophores. A large
individual may have as many as four hundred fully formed
spermatophores stored at one time.

The formation of the spermatophores is evidently rather
rapid. Several, perhaps several dozen, may be formed in a day.
It is difficult to determine with anything like accuracy what
the rate may be, but after a male has used large numbers of sper-
matophores in repeated copulations, a day or two is sufficient to
bring the supply almost if not quite to normal.

SEXUAL ACTIVITIES OF THE SQUID 381

STRUCTURE

Each spermatophore consists of a white opaque mass of sper-
matozoa surrounded by almost transparent liquids and mem-
branes, a mass of material, rather opaque but not so opaque as
the mass of spermatozoa, which lies against one end of it, and
a brownish spiral filament together with a number of membranes
of varying degrees of transparency on the other side of the body
just mentioned. For convenience we may speak of these main
divisions as the sperm mass (fig. 1, SM), the cement body (CB)
which lies at one end of the sperm mass and was called by Raco-
vitza the ‘faux boyaux,’ and the ejaculatory apparatus (HA),
composed of the complicated group of membranes and the spiral
filament which joins the cement body and occupies the smaller
end of the spermatophore. It is also convenient to speak of
the portion of the spermatophore occupied by the ejaculatory
apparatus as the oral end, since from this end ejaculation takes
place.

Figure 1, a complete spermatophore, shows that the aboral
is considerably larger than the oral end. The sperm mass is
little more than half the diameter of the spermatophore and does
not extend to the extreme aboral end of the spermatophore.
There is some difference in the amount of space posterior to the
sperm mass, and likewise in the diameter of the portion of the
spermatophore in which the sperm mass lies in different sper-
matophores. This is due largely to the fact that the spermato-
phores, when liberated take up water rapidly, and consequently
swell and change in shape. This condition is easily controlled
by passing the spermatophores into solutions of formaldehyde.
Full strength formaldehyde will do no harm, but 10 per cent
formalin is sufficient for ordinary purposes.

In studying the structure I have found that staining the forma-
lin-treated spermatophores in dilute solutions of Ehrlich’s triple
stain or in Ehrlich’s triacid stain and mounting in glycerin
jelly has been very helpful. The stains are selective and stain
different parts different colors or shades. The glycerin jelly
will clear and preserve without shrinking the specimens badly.

382 GILMAN A. DREW

The stains are not permanant, but when kept in the dark will
last for some months. Formalin will harden gelatin, so mounts
of formalin specimens become quite permanent, unless they are
kept in a very damp place without being sealed.

The spermatophore is very turgid and elastic. This is due
to the outer covering, the outer tunic (fig. 2, OT), which is a
tough and elastic membrane. It is transparent, rather thin,
and of about even thickness except at the extreme anterior
end, where it becomes thickened and sculptured for the attach-
ment of membranes of the ejaculatory apparatus, and where
it is modified to form the cap (C) covering the oral end. The
cap ultimately loosens and allows the spermatophore to ejacu-
late. The name cap which is applied to the portion that covers
the oral end is somewhat confusing. The covering is formed
by winding and cementing down around this end a thin leaf of
outer tunic material which is continued as a long thin thread,
the cap thread (CT), from the oral end of the spermatophore.
This thread serves to loosen the winding of the cap when it is
pulled, and ejaculation processes are immediatley started.

The outer tunic gives the strength and elasticity to the sper-
matophore. The very great turgidity is due to the strain to
which this tunic is subjected. When this tunic is punctured
or cut, the contents escape rapidly and the tunic shrinks be-
cause of its elasticity.

Inside the outer tunic is the middle tunic (figs. 1 and 2, MT).
This is closely applied to the outer tunic, thickened over the
whole aboral end and gradually thins out orally, after the sperm
mass is passed. The aboral portion is rather thick and granular,
but transparent. After the region of the cement body is reached.
the granular character is lost, and it is very difficult to deter-
mine whether a membrane is present or whether the space is
filled with liquid. However, it can sometimes be easily traced.
Under certain conditions of distension or ejaculation of the sper-
matophore, a line appears that can be accounted for only as
the inner border of such a membrane. This portion is quite
transparent, differing greatly in appearance from the granular
membrane in the aboral end, but under favorable conditions

SEXUAL ACTIVITIES OF THE SQUID 383

it may be traced to the oral end of the spermatophore, where
it seems to end against a ridge on the inside of the outer tunic.
The middle membrane. of the ejaculatory apparatus, to be de-
cribed later, is firmly attached to the other side of this same
ridge. The middle tunic is soft, evidently elastic, forming an
elastic cushion, and is evidently capable of taking up water
rapidly. When a spermatophore is removed from a spermat-
ophoric sac and placed in sea-water, the middle tunic imme-
diately begins to increase in thickness and soon the spermatophore
begins to ejaculate, or, if the cap holds, the outer tunic is rup-
tured by the increased internal pressure.

If the outer tunic be cut in such a spermatophore, the sperm
mass is driven through the opening and the middle tunic thickens
to occupy the space formerly occupied by the sperm mass. The
combined swelling of the middle tunic and the elastic shrinking
of the outer tunic nearly obliterate the space formerly occupied by
the sperm mass. Evidently these two tunics are concerned in sup-
plying the force that causes the ejaculation of the spermatophore.

Closely applied to the sperm mass is a very thin and not very
definite membrane, the inner tunic (fig. 2, 77’). This is fre-°
quently hard to identify, as it has nearly the same appearance
as the mucilaginous material with which the sperm are mixed,
and it is very closely applied to the mass. In the figures the
part of the inner tunic covering the sperm mass is represented
by a single line.

During the formation of the spermatophore, when the ma-
terial of this tunic is wrapped around the sperm mass, it is easily
distinguished in sections. In spermatophores which have not
been completely formed, before final shrinking takes place,
the coiling of the sperm thread of the sperm mass is quite dis-
stinct, and here the inner tunic is seen in the spaces between
the coils of the sperm thread, separated slightly from the sperm
mass.

Connecting the sperm mass and the cement body is a thin
cylinder of transparent material, the connecting cylinder (fig.
6, CC) which seems to be a continuation of the core of the muci-
laginous material with which the spermatozoa are mixed. The

384 GILMAN A. DREW

inner tunic is seen here as a thin sheet surrounding this cylinder
and extending from the sperm mass to the cement body.

On the cement body the inner tunic becomes applied to the
outer membrane (fig. 2, OM) which covers this body. From
this point on the inner tunic (/7’) is thicker and very easily
seen. It, together with the outer membrane, to which it is
closely applied, leaves the cement body where it abruptly narrows
and continues toward the oral end of the spermatophore nearly
to its extremity, as part of the ejaculatory apparatus. The
two structures are very similar in appearance, but the dividing
line between them is distinct.

Just before reaching the point where the ejaculatory apparatus
is thrown into loops, both the inner tunic and the outer mem-
brane thicken to form a distinct ring. The dividing line between
the two structures is easily followed for some distance into the
thickening and is then hard to trace. I find, however, in the
forming spermatophore, and occasionally in completely formed
specimens which have been mounted some months in glycerin
jelly, that the line of separation can be indistinctly traced nearly
through this thickening, and I am inclined to think that the
oral end of this thickening may be taken to be the extremity of
the inner tunic. Beyond the thickening there is no indication of
a double character and this part is probably the continuation
of the outer membrane only. In the thickening the two struc-
tures are evidently more closely applied than elsewhere if, indeed,
they are not fused.

The free end of the outer membrane is further on among the
loops of the ejaculatory apparatus. It is easily seen in a speci-
men which is cut so that the loops of the ejaculatory apparatus
straighten out, (fig. 3). This free ending of the inner tunic and.
outer membrane on the ejaculatory apparatus is of importance
in studying the method of ejaculation.

The middle and inner tunics are entirely separate from one
another. Where they touch they do not adhere. Between
them is an actual or potential space (figs. 1 and 2, SL) filled
with clear liquid. This space is always visible behind the sperm
mass, in the region of the cement body, and along the ejacu-

SEXUAL ACTIVITIES OF THE SQUID 385

latory apparatus. It is most easily seen behind the sperm mass,
but there is evidently considerable liquid along the sides of
the ejaculatory apparatus. This is especially shown during
ejaculation.

The sperm mass (figs. 2 and 23 A, SM) is really composed
of a thread or sheet of spermatozoa spirally coiled around a
core of mucilaginous secretion, with the secretion extending
between the loops of the sperm thread to the surface. The
sperm forms a sort of elongated thin plate, coiled edgewise so
one edge rests against the core and the other comes to the sur-
face. The plate does not meet the surface and core at right
angles, but is tipped slightly so the surface edge lags behind
the core edge. By means of the secretion the plate composed of
individual spermatozoa is consolidated into a very flexible
cylindrical rod that has the character of a unit body. The
secretion is tenacious and penetrates in between the spermatozoa,
but the core and secretion between the loops of the sperm thread
have very few spermatozoa embedded. Because of the appear-
ance the sperm mass thus formed is frequently called the sperm
rope. In the squid the loops of the spiral in the completed
spermatophore are not very easily seen and, as a rope does not
consist of a single strand coiled in this way, the term has no
significance and may be misleading. The secretion mixes
freely with sea-water, so the spermatozoa obtain their individ-
ual liberty promptly when the time comes for them to perform
their function.

Just oral to the sperm mass is the cement body. This is
more or less definitely attached to the sperm mass by the core
of mucilaginous material, which extends orally from the sperm
mass as a definite cylindrical filament, and by the inner tunic
which continues from one to the other. The connection be-
tween the two structures is slight, but sufficient to connect them
definitely together.

The cement body (fig. 2, CB) the function of which was not
determined by Racovitza and was called by him the ‘faux
boyaux’—is a somewhat elongated pear-shaped body with
the base somewhat less in diameter than that of the sperm mass

386 GILMAN A. DREW

and the narrower elongated oral portion. in contact with the
end of the conspicuous spiral filament of the ejaculatory ap-
paratus (figs. 2 and 23 A CB).

In the forming spermatophore the narrow end of the cement
body, the hyaline core, (HC) is continued to the oral end of
the spermatophore as a narrow cylindrical rod, which stains
about the same as the material of the cement body. This fills
the space inside the loops of the spiral filament and, oral to the
spiral filament, the continuation of the lumen of the ejacula-
tory apparatus. I take it from Racovitza’s figures and de-
scriptions that this, which he calls the hyaline core, is persistent
in the fully formed spermatophore of Rossia. In Loligo it is
transient, disappearing, evidently by liquification, almost
simultaneously with the completion of the formation of the
spermatophore. The space it occupied remains as the lumen
of the ejaculatory apparatus, which is probably filled with the
liquid in the completed spermatophore.

In spermatophores in process of formation and frequently
in freshly stained fully formed specimens, the cement material
can be seen to be spirally wound. There is a central core, evi-
dently a continuation of the cylinder extending between the
sperm mass and the cement body, which is continued as a nar-
row cord the whole length of the hyaline core which extends
orally from the cement body (figs. 23 and 23 A). In fully formed
specimens the cement material usually does not show the spiral
character plainly and the inner core may not be visible.

The cement body is probably covered entirely by the inner
membrane, but over the large aboral end, where it is likewise
covered by the outer membrane and the inner tunic, and where
these membranes are fused tightly together, it cannot be traced
as a separate membrane.

Over the narrow oral end of the cement body it is visible as a thin
membrane which continues orally after leaving the cement body
over the outer surface of the spiral filament. Still further orally,
where there is no spiral filament, this membrane lies next to
the lumen of the ejaculatory apparatus (fig. 2, 7M). Formerly
it was in contact at this point with the hyaline core.

SEXUAL ACTIVITIES OF THE SQUID 387

The spiral filament (fig. 2, S/), while formed as a separate
structure, evidently sticks to the inner membrane which covers
it. The filament is brown, more or less granular, and is not of
equal size and shape throughout. It is heaviest, with the coils
most open, midway in its course, with both ends rather crowded.
The loops of the spiral on the end next the cement body are
very closely crowded, flattened, and the extreme end sometimes
has the continuity of the thread broken so it is made up of con-
secutive fragments. The loops of the oral end of the filament
become closely crowded, then more open and finally fade away
so gradually that it is hard to determine where the filament ends.
The function of the filament is not easily determined. It is
evidently not a coiled spring and it seems to have very little
elastic value. It seems probable that its chief function is to
hold the lumen of the ejaculatory apparatus freely open so that
evagination, to be described later, can be accomplished with-
out tearing the membranes concerned. The rapidity of the
ejaculation must be slowed somewhat to allow time for the on-
coming cement body and sperm mass. The resistance caused
by breaking the spiral filament into small fragments probably
accomplishes this purpose.

Between the inner and the outer membranes is the middle
membrane (fig. 2, MM). This is very transparent and fre-
quently shows longitudinal striations, indicating the position
of the successive windings of the sheet of which it is composed.
It is much thicker than the other membranes and, while capable
of much stretching, is evidently tough. It extends from the
point where the outer membrane, together with the inner tunic,
leaves the cement body, to the oral end of the spermatophore.
At the aboral end, the tube formed by this membrane is closed
by the oral end of the cement body. At the oral end this tube,
which was open in formation (fig. 23, MM), is closed and closely
apphed to the inside of the cap where it spreads out laterally
and is fastened by its lateral margin to the ridge of the outer
tunic (fig. 2, MM). In this spreading and flattening process
the lumen of the tube is also pressed out laterally so that in form
the end is something like a pressed-in hollow rubber ball with

€

388 GILMAN A. DREW

the tube formed by the middle membrane extending back from
the concave side of the ball.

The term ‘ejaculatory apparatus’ has been applied to the inner
membrane, together with the spiral filament, the middle mem-
brane, and the outer membrane and inner tunic to their junc-
tion with the cement body. This is not a very satisfactory term
as the spermatophore acts as a unit in ejaculation. That is, there
is no one part that is active while the remainder are passive.
The ejaculatory apparatus could not possibly deliver the sperm
mass in position were it not for the elastic outer and middle
tunies and their relations to liquids and structures. The term
may, however, stand for want of a better one, since this portion
is mostly concerned in ejaculation.

It may aid somewhat in understanding the arrangement of
the parts of a spermatophore if we consider what is present in
optical cross-sections through: 1, the region of the sperm mass;
2, the region of the aboral end of the cement body; 3, the region
of the oral end of the cement body; 4, the region of the spiral
filament, 5, the region just posterior to the cap. The parts cut
will be mentioned in turn from the outside to the median axis
(figs. 21 to 18).

1. The region of the sperm mass: 1, Outer tunic; 2, middle
tunic; 3, space (actual or potential) filled with liquid; 4, inner
tunic; 5, sperm mass.

2. The region of the aboral end of the cement body: 1,
Outer tunic; 2, middle tunic; 3, space (usually actual) filled
with liquid; 4, inner tunic; 5, outer membrane (if actually
present fusee with the inner tunic), 6, inner membrane (prob-
ably); 7, cement body.

3. The region of the oral end of the cement body. 1,
Outer tunic; 2, middle tunic; 3, space (usually actual) filled
with liquid; 4, inner tunic; 5, outer membrane; 6, middle mem-
brane; 7, inner membrane; 8, cement body.

4. The region of the spiral filament: 1, Outer tunic; 2, mid-
dle tunic; 3, space (actual) filled with liquid; 4, inner tunic;
5, outer membrane; 6, middle membrane; 7, inner membrane;
8, spiral filament; 9, lumen, probably filled with liquid, formerly
filled with hyaline core.

SEXUAL ACTIVITIES OF THE SQUID 389

5. The region just aboral to the cap: 1, Outer tunic; 2,
middle tunic; 3, space (actual) filled with liquid; 4, middle mem-
brane (the inner tunic and outer membranes do not extend
this far); 5, inner membrane; 6, lumen (the spiral filament does
not extend this far).

If each structure is considered in turn in their longitudinal
relations we find:

1. The outer tunic is continuous over the whole spermato-
phore except at the oral end where there is a modification, the
cap, which is spirally wound around the otherwise open tunic
to form a closing mechanism. ‘The cap has attached to it a
long thread which, when pulled, serves to loosen the cap and thus
liberate the enclosed mechanisms.

2. The middle tunic is continuous throughout the length of the
spermatophore up to the thickened ridge on the outer tunic
near the oral end, which it joins. .

3. The inner tunic is continuous over the region of the sperm
mass as a closely investing, thin membrane. It becomes thicker
over the posterior end of the cement body. After leaving the
cement body as an investing membrane it becomes a little thicker
and closely covers the outer membrane. Near the region of
the anterior extremity of the spiral filament this tunie ends with
open mouth closely associated with the outer membrane.

4. The outer membrane probably begins at the aboral end
of the cement body, but cannot be definitely distinguished from
the inner tunic until near the place where both of these struc-
tures leave the cement body and, together, give the appearance
of a double membrane. After leaving the cement body the
outer membrane is applied to the middle membrane. A short dis-
tance orally from the end of the inner tunic the outer membrane
also ends with an open mouth. It is important to understand
that the oral portions of the inner tunic and the outer membrane
together forma tube, closed at the aboral end where they are
united to the cement body, and open at the oral end. The
opening is of course closed by the other structures. There is,
however, no organic union between these structures and the
other membranes from the point where the outer membrane and

390 GILMAN A. DREW

the inner tunic leave the cement body to invest the middle
membrane.

5. The space, potential or actual, between the middle tunic
on the outside, and the inner tunic, outer membrane and middle
membrane on the inside, is continuous throughout the sper-
matophore. The liquid enclosed in this space serves the mechan-
ical purpose of a lubricant and at the same time an easily flow-
ing substance to which pressure is applied. The elastic force
of the outer and middle tunics is transmitted through this liquid
to the sperm mass and other structures during the act of
ejaculation.

6. The sperm mass extends through the aboral two-thirds
of the spermatophore, inside the inner tunic, which is closely
applied and united to it.

7. The cement body is just oral to the sperm mass and at-
tached to it by a connecting cylinder. The aboral end of the
cement body is covered by the inner tunic, part of it at least
by the outer membrane and possibly also by the inner mem-
brane. If all are present, they are closely fused so they are
hard to distinguish. The oral end of the cement body is covered
by the inner membrane, outside of which come, in order, the
middle membrane, outer membrane, and inner tunic. These
are all easily distinguished from one another at this point.

8. The middle membrane forms a tube extending from the
position where the inner tunic and the outer membrane leave
the cement body to the oral end of the spermatophore. Just
beneath the cap the oral end, which, although formed as an
open tube, is now closed, becomes closely applied to the inside
of the cap and is spread out laterally to the ridge on the outer
tunic to which it is firmly attached. The open aboral end
of the tube formed by the middle membrane is plugged by
the small oral end of the cement body which is covered by the
inner membrane. It has no organic connection with the outer
membrane or with any part of the cement body, except for about
one-third of the length of the surface that is applied to the inner
membrane where it covers the cement body (fig. 23 A, PA). The
portion next to the open mouth of the middle membrane ad-

SEXUAL ACTIVITIES OF THE SQUID 391

heres to the inner membrane covering the cement body firmly,
and in ejaculation is liberated only by the rupture of this mem-
brane. It is important to understand that the open mouth of
the middle membrane is directed aborally and the open mouths
of the inner tunic and the outer membrane are directed orally.
The one fits inside the other. The oral end of the middle mem-
brane is a closed structure, like an indented hollow rubber ball,
attached by its margin to the outer tunic and with the thin con-
vex side (fig. 2, MM‘) applied to the under surface of the cap
which closes the outer tunic. This part ruptures when the cap
is loosened, so the lumen of the ejaculatory apparatus is opened
to the outside (fig. 5, MM").

9. The inner membrane and spiral filament are united to the
middle membrane. Their positions and relations are better
shown by figures than by description. In-ejaculation the inner
membrane ruptures at the point where it joins the cement body
(fig9, PH):

With these points in structure in mind we can now proceed
with the method of ejaculation.

EJACULATION

In delivering the spermatophores the male grasps a bundle
of them with the tip of the left ventral arm and quickly passes
them into position (Drew, ’11). The spermatophores leave
the sexual duct of the male aboral ends first and the long threads
connected with their caps, embedded in the secretions of the
spermatophoric sac, drag behind. The sharp pull occasioned
by the movement of the arm pulls on these threads and causes
the caps to loosen. This starts the process of ejaculation.
Under normal conditions the process is very rapid, occupying
about ten seconds. This very rapid action of course makes it
impossible to follow the ejaculation under a microscope even
though the spermatophore is held in position and the thread
pulled when all is ready. It was accordingly necessary to devise
some method to slow down the movements.

Evidently the great tension of the elastic outer tunic has much to
do with the process of ejaculation. Inasmuch as spermatophores

THE JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

392 GILMAN A. DREW

carefully removed from the spermatophoric sac without pulling
the cap thread and placed in sea-water are very likely to ejacu-
late soon, and placed in fresh water will either ejaculate or burst
very promptly, it is evident that osmotic action, in which the
middle tunic is probably involved, plays an important part.

Elasticity and osmotic action accordingly have to be con-
sidered in searching for some method to slow down the action.
It was found that formaldehyde affects probably both the elas-
ticity and the osmotic properties, but that it evidently hardens
the cement holding the cap, so it is very difficult to open it when
needed, and that the various membranes of the spermatophore
are soon so changed, possibly by hardening certain colloids,
that ejaculation is not likely to be completed. These very prop-
erties were, however of the umost value in the studies, for
full-strength formaldehyde thrown on an ejaculating spermat-
ophore will cause it to stop all action rather promptly. After
some experience it became possible to allow the required amount
of time to stop a spermatophore at the required stage of ejacula-
tion by squirting full-strength formaldehyde on it just before
it reaches the stage wanted. Such a spermatophore may then
be stained in aqueous stains and mounted in glycerin jelly for
study at leisure. It, of course, requires a great deal of time and
patience to get some of the stages, but, as the figures showing
the stages of ejaculation accompanying this paper are camera-
lucida drawings of such specimens, it will be seen that it is pos-
sible to get the stages by this method.

Very many chemicals were tried to get the required slowing
effects. Sugar. solutions were good, but the membranes were
soon weakened so the elasticity was destroyed. Magnesium
chloride solutions have given the best results. The strength
of the solution that works best seems to differ with spermato-
phores from different individuals, but a one-fourth saturated
solution in sea-water has been very good.

The spermatophores are received from the spermatophoric
sac directly into this solution, and in two or three minutes they
will be ready for use. The turgidity is evidently effected, and
where the spermatophores are left several hours, there may be

SEXUAL ACTIVITIES OF THE SQUID 393

changes in elasticity and the freedom with which membranes
will move on each other may be disturbed. After the sper-
matophores have been in the solution some minutes, if they are
to be used for work for a long period, more sea-water may be
used to dilute the solution. Generally it is best to use material
that has recently been put into the solution.

The method used in studying ejaculation was to remove the
spermatophore from the magnesium chloride solution to a watch-
glass with a little sea-water, placed on a black background.
The cap thread was then grasped with forceps and the whole
spermatophore shaken. With a reasonably powerful engraver’s
glass held on the head with a spring the process of ejaculation
may be watched, and with a large-mouthed pipette filled with
formaldehyde the process can be stopped when desired. The
time for ejaculation may be slowed down to take from a minute
to two minutes, so it is possible to supplement observations
made on the fixed material by observations on the ejaculating
spermatophores.

It is necessary to concentrate attention on one portion at a
time, but there is no difficulty in following movements of parts
under the lens of a compound microscope. The chief trouble
is in focusing attention on particular parts, for everything is
moving at the same time and the mechanism is too complicated
to be taken in at a glance and too large for all to be under a lens
of sufficient power at one time.

As the process of ejaculation is somewhat complicated, a
series of diagrams are given on plate 6, from which all portions
not essential to understanding the process have been eliminated.
By referring to these diagrams at this time it will be easier to
follow the processes of ejaculation as they are given in other
figures and descriptions.

The cap end of the cap thread is flattened and is apparently
applied and cemented to the outer tunic in a somewhat spiral
manner so the otherwise open end of the tunic is held shut.
When the thread is pulled it loosens where cemented (fig. 4),
and the end of the outer tunic is allowed to open. There is
evidently some tearing, but not much.

394. GILMAN A. DREW

With the opening of the tunic the portion of the middle mem-
brane applied to the inside of the cap ruptures so that the lumen
of the ejaculatory apparatus is opened to the outside of the sper-
matophore and the ejaculatory apparatus immediately begins
to evaginate because of the pressure on the inside of the sper-
matophore (fig. 5) The evagination of that portion of the
ejaculatory apparatus which is oral to the spiral filament is so
rapid in the untreated spermatophore that the eye cannot follow
it, but in the specimens treated with magnesium chloride it may
be slowed down so the gradual evagination can be followed easily.
As the ejaculatory apparatus evaginates, the diameter of the
tube is greatly increased and the walls are correspondingly
thinned.

There is a distinct pause in the evagination when the region
of the spiral filament is reached (fig. 6). This is probably
largely due to the stiffness of the filament itself, but may be
influenced by the fact that other membranes are involved at
about the same point.

Evidently the evagination of the first part of the ejaculatory
apparatus is due to the pressure of the liquid between the mid-
dle and inner tunics that is in the oral end of the spermatophore.
This is shown by the fact that the action is so rapid and by the
further fact that the cement body and sperm mass are drawn
away from each other (fig. 6). The sperm mass lags behind,
so the connection between it and the cement body is stretched
to its full extent.

After an instant’s delay when the region of the spiral filament
is reached, the tube continues to evaginate. The evagination
here is continuous, but not nearly so rapid as the first part.
the evaginated portion of the tube increases greatly in diameter,
the walls become correspondingly thinner and the spiral fila-
ment is broken into minute fragments which continue to adhere
to the outside of the evaginated tube, (fig. 7). As this process
goes on, the free edge of the outer membrane adheres to what
is now the inside of the evaginated middle membrane and is .
reflected so that this membrane, together with the inner tunic
with which it is associated, is turned inside out (figs. 7 and 8,
OM and IT).

SEXUAL ACTIVITIES OF THE SQUID 395

Evidently the force that causes the evagination is still the
elastic and osmotic force in the outer and middle tunics of the
spermatophore acting through the liquid which fills the space
between the walls of the evaginating tube.

The part played by the spiral filament seems to be largely,
if not wholly, that of keeping the tube from collapsing with the
pressure, but there may be some elastic force that aids in the
evagination. The torsion that would be caused by the turning
of a spiral spring might aid in the evagination, when once started,
but there is no evidence that the filament is particularly elastic
or partakes of the nature of a spring. The very fact that it is
broken into minute fragments during the process of the evagina-
tion of the tube indicates that it can have no very great elastic
properties, and probably indicates that it retards rather than
accelerates the evagination of this part. It is necessary that
evagination of this portion shall not be too rapid as the sperm
mass must gain momentum and move along at a corresponding
rate.

That such a filament may serve a very useful purpose in keep-
ing the tube from collapsing or folding is evident. The freedom
of the movements of the membranes concerned would be seri-
ously interfered with if the tube were allowed to collapse or
kink. The oral end of the tube does not need such a mechan-
ism, as it is short and simple in construction and would natu-
rally evaginate quickly with the pressure of the liquid between
it and the outer wall. The same condition would not hold true
for the much longer and more complicated tube that has the
spiral filament.

It is of passing interest to note that there is a very general
impression among zoologists who have no personal acquaint-
ance with Cephalopod spermatophores that this spiral filament
is really a spring, that it is used in discharging the sperm mass
in the same mechanical way that a spring gun discharges its
projectile, and that the discharge is through the end of the sper-
matophore farthest from the spring. There is, of course, no
foundation of fact whatever for such an impression. It is simply
arriving at conclusions from superficial appearances rather
than by study and experimentation.

396 GILMAN A. DREW

When evagination has proceeded as far as the oral end of the
cement body, so that the end of this body begins to project
through to the outside, the inner membrane is ruptured, so that
this membrane, with the remnants of the spiral filament, is
separated from the cement body (fig. 9, PR).

As evagination now proceeds, the oral end of the cement
body projects into the sea-water. At this point ejaculation
is retarded until the very great pressure behind the sperm mass
forces this mass against the aboral end of the cement body,
to which the inner surface of the middle membrane is attached,
and so causes the middle portion of the cement body to be drawn
out and around the sides of the aboral portion of the cement
body in the form of a cap (fig. 11). The extreme oral end of the
cement body, from which the inner membrane and spiral fila-
ment have been torn, appears as a knob or button on the other-
wise smooth surface of the cement body.

It may be well, before proceeding with the other changes
-that are taking place, to call attention to the position of the
oral portions of the inner tunic and outer membrane. In the
evagination that has taken place the free ends of the inner tunic
and outer membrane which originally enclosed the middle mem-
brane have been turned back by the evaginating tube, so that
the opening is directed toward the aboral end of the spermato-
phore, and, together, they form the inner lining of the evaginated
tube as it now appears. In this process the sperm mass is
being carried through the opening of the inner tunic and outer
membrane and is being forced into the sac formed by them (figs.
9, 12, and 39).

As has already been pointed out, the inner tunic and outer
membrane are firmly attached to the sides of the aboral end
of the cement body so, when ejaculation has proceeded to this
point, the membrane cannot be stripped further aborally. The
part which has been turned back with the evaginating middle
membrane thus forms a sac, with the cement body firmly at-
tached to the closed end, and the pressure from behind forces
the sperm mass into this sac.

SEXUAL ACTIVITIES OF THE SQUID 397

The adhesion of the end of the middle membrane to the mid-
dle portion of the cement body not only serves to draw this
cement body around the end of the sperm mass, but holds the
sac in position to have the spermatozoa thoroughly and completely
forced into it by pressure behind. It will be noticed that dur-
ing this process the diameter of the aboral end of the spermato-
phore is greatly reduced, due to the elasticity of the outer tunic,
and that the middle tunic swells, loses its granular appear-
ance, and comes to occupy the space vacated by the sperm mass.
At the same time the outer end of the evaginated ejaculatory
apparatus becomes considerably expanded as the sperm mass
is crowded into it (figs. 12, 13, and 14).

Continued pressure causes the walls of the cement body to
burst (fig. 15). The end of the middle membrane is thus re-
leased and the sperm mass enclosed in the reflected inner tunic
and outer membrane, smeared with cement from the ruptured
cement body on its larger closed end, glides rapidly through
the middle membrane and is free from all other mechanisms.

The covered sperm mass, which may be called a sperm res-
ervoir (fig. 17), is usually somewhat coiled. The closed end
is large and covered with cement, and the open end is small,
with a thickened portion just behind the opening. The thick-
ened portion seems to correspond to the thickened portion of
the inner tunic and outer membrane, described in connection
with the structure of the spermatophore, that lies a little aboral
to the free end of this membrane. The thickened walls prob-
ably tend to prevent too rapid escape of the spermatozoa.

From the open mouth of this sperm reservoir of untreated
specimens the spermatozoa escape in a constant cloud which
reminds one of the smoke from an evenly discharging factory
chimney. The discharge may go on for hours. When care
is taken to provide an abundance of sea-water, such reservoirs
will still be discharging twenty-four hours and more after they
were liberated from the spermatophores.

Referring to the methods of copulation of the squid, given in
a former paper (Drew, ’11), it will be seen that when the sper-
matophores are carried to the mantle chamber of the female

398 GILMAN A. DREW

they are held in position by the male long enough for them to
discharge and to have the sperm reservoirs fixed by the cement
on them to the tissue near the oviduct of the female. In this
position each gives out its small cloud of sperm for some hours.
If the female lays her eggs within the time they are active, in-
semination is assured. On the other hand, if the spermato-
phores are transferred to the region of the buccal membrane
of the female, they are held in position by the male until they
discharge and the sperm reservoirs are attached to the walls
arranged for them. Here, as they discharge, the spermatozoa
are directed, evidently by ciliary action, into the sperm recep-
tacle where they are stored for future use.

The discharged empty case (fig. 16) is much smaller, especially
in diameter, than it was before ejaculation. The outer tunic
appears about as it did. The middle tunic is clear, not granu-
lar, and occupies most of the space inside the outer tunic. The
evaginated tube that adheres to the oral end of the outer tunic
is likewise less in diameter than it was at the time of evagination
when there was pressure inside. The end of the tube attached
to the outer tunic is clear and corresponds to the oral unorna-
mented portion of the tube in the spermatophore. The region
of the spiral filament is shown by the broken fragments adher-
ing to the tube, and the outer end of this marked portion
represents the end of the spiral filament and inner membrane
that was attached to the oral end of the cement body. This
is the point of rupture (PR). The remaining unornamented
flaring tube is the part of the middle membrane which was in
contact with the cement body. The outer third of this portion
was firmly attached to the inner membrane that covered the
cement body. By the breaking of this attachment the sperm
reservoir, with the cement at the closed end, became free to
be forced out of the case by the pressure behind it.

SEXUAL ACTIVITIES OF THE SQUID 399

FORMATION

At first thought it is very difficult to understand how so com-
plicated a structure as a spermatophore, with its numerous
coats and structures, can be formed as a secretion inside the
lumen of a glandular duct. To make the process clear it is
necessary to know the structure of the duct in some detail.

The parts of the duct have received different names by dif-
ferent writers, and, inasmuch as the functions of the parts were
not well Minced at the time, the names that have been
applied to them are generally not significant and should, I think,
be abandoned as misleading.

A recent writer (Marchand, ’07), who has covered this sub-
ject much more fully than has previously been done and who
has made careful comparisons of the male ducts of a large number
of Cephalopods, had analyzed the names previously given and
made selections that suit his purpose, but as these names are
applied without definite knowledge of the functions of the parts
receiving the names and as more than one function is performed
by a part to which he gives a single name, following the names
he gives would seem to lead to even more confusion than to
again change them.

The male genital organs of the squid are asymmetrical, only
the testis and duct on the left side being present. The testis
lies far posteriorly and dorsally (the terms posterior, anterior,
dorsal, and ventral are used in the apparent rather than the
true morphological sense). Just beyond the testis capsule the —
vas deferens shows a slight swelling, the ampulla of the vas
deferens. From this point the vas deferens, at first a wavy
and then a closely plaited tube, extends around the left side
of the visceral mass to a point just posterior to the left branchial
heart. Here the sexual duct enlarges to form a complicated,
folded gland in which the spermatophores are formed.

The whole mass is frequently referred to as the spermatophoric
gland. This is proper in the sense that the spermatophores
are formed here, and it is not proper, inasmuch as it is not a
single gland, but a series of glands and mechanical contrivances,

400 GILMAN A. DREW

each portion of which has a definite individual function in the
formation of spermatophores. For convenience we will call
it the spermatophoric organ. This name will be applied to
the whole structure, consisting of various glands and mechan-
isms, which extend from the vas deferens to the duct that car-
ries the completed spermatophores to the spermatophoric (Need-
ham’s) sac. The spermatophoric organ is rather transparent,
like most tissues of the squid. The forming spermatophores
may be easily seen in the different parts of the organ. It is
possible to cut the organ away from the visceral mass, place
it in a watch-glass of sea-water and, under a compound micro-
scope, see somewhat clearly the structures and positions of a
forming spermatophore. By keeping the water changed on
such an organ, its movements, which are very vigorous, will
be kept up for nearly an hour and the forming spermatophores
during the interval will move some distance. Within this organ
the spermatophores are formed and completed.

The duct leading from the spermatophoric organ to the sper-
matophoric sac, which will be called the spermatophoric duct
(frequently called the vas efferens, and by Marchand the distal
vas deferens) carries the completed spermatophores for storage
in the spermatophoric sac.

The vas deferens (figs. 29 and 30, VD) passes dorsal to the sper-
matophoric organ (between it and the general visceral mass) |
for about three-quarters of the length of the organ, where
it joins the first of a series of structures that together form this
organ (fig. 32, VDO).

The portion of the spermatophoric organ joined by the vas
deferens I will call the mucilaginous gland (figs. 29 to 36). It
secretes a sticky substance which is mixed with the spermatozoa
and forms the material in which the sperm thread is imbedded,
the cement body, and the hyaline core around which the spiral
filament is wound. This gland is composed of two parts. One
part (MG?) extends from the vas deferens to the next portion
of the spermatophoric organ. This is referred to by Marchand
as the second division of the spermatophoric gland (vesicula
seminalis). Marchand uses the term spermatophoric gland

SEXUAL ACTIVITIES OF THE SQUID 401

for three parts of what is here called the spermatophoric organ.
The term does not serve my purpose, for the portion is more
than a mere gland or indeed a series of glands. There are more
parts to be decribed than the three divisions given by Marchand,
and the term spermatophoric gland would indicate that the sper-
matophores are formed here, while they are only partly formed
here.

The other portion of the mucilaginous gland (MG') forms a
large outgrowth from the side of the portion just described. The
opening of this portion is near the opening of the vas deferens
and a considerable portion of the gland extends back between
the viscera and the portion just described. This is called the
first part of the spermatophoric gland by Marchand, and will
be called the first part of the mucilaginous gland here.

In structure the two parts are much alike. Both have thick
walls, thrown into folds on the inside. These folds are fre-
quently joined by bridges, and in many places the deeper por-
tions of the depressions between the ridges form pouches or
sometimes tubules (Williams ’08). The whole is, however, too
open to form a true racemose or tubular gland. The cavity of
each portion of the gland is extensive, forming a pelvis or basin in
which the secretion is poured. The whole interior of the gland is
ciliated, but the pelvis is particularly well supplied with cilia.
The spermatozoa, entering from the vas deferens, pass into the
pelvis of the second part of the mucilaginous gland (fig. 32),
where they are mixed with secretion and the moving thread
of sperm is covered with it. The spermatozoa do not enter part
one of the mucilaginous gland, but are passed along a groove
in part two past, but a little to one side of, the opening of part
one. In the region of the groove, and for some distance along
the side, especially along the side nearest part one, the cilia
are large and numerous and serve to move the mixed sperm
and secretion continuously toward and along this groove through
this portion of the mucilaginous gland.

Possibly one-third of the distance from the vas deferens to
the distal end of part two the walls of the groove are thrown
into a few spiral ridges (fig. 34, /’), between which the spermatozoa

402 / GILMAN A. DREW

are passed. The sperm thread is here flattened between the
ridges and wound edgewise so one edge becomes the center and
the other edge of the surface of the spirally wound sperm mass.
As the spermatozoa are covered by the secretion from the gland,
the secretion along that edge which forms the center becomes
a continuous core in which there are few spermatozoa, and the
secretion on the flat applied sides of the sperm thread stick the
successive loops of the coil together. The sperm mass, as coiled,
does not lie with flat applied surfaces of the loops at right angles
to the central core, but the edge applied to the surface lags a
little behind the edge at the central core. The surface edge
is accordingly nearer the cement body than the core edge.

In longitudinal section the sperm mass thus appears like a
series of small open funnels with the small ends directed toward
the aboral end of the spermatophore. ‘The spaces between the
funnels, together with the core that would occupy the small
open ends of the funnels is filled with sticky material furnished
by the mucilaginous gland. From the location of the ridges
which serve to coil the thread, through the remainder of the
mucilaginous gland and through succeeding glands, the sperm
mass is rotated on its longitudinal axis by the action of the cilia
in the groove in which it les. It is through this longitudinal
rotation that the sperm mass, molded by the ridges between
which it passes, is coiled into the form described. The sperm
mass, coiled in this way, is usually called the sperm rope. It
should be borne in mind that the coil consists of a single flat-
tened strand and not a number of strands as is the case with a
rope (fig. 23 A; SM).

Spermatozoa continue to issue from the vas deferens and the
sperm mass continues to form until the end first formed reaches
some distance past the limit of the mucilaginous gland to a
point about opposite the notch (fig. 32, C1). The sphincter
muscle around the opening of the vas deferens then contracts
and no more spermatozoa are allowed to enter. When the free
spermatozoa are wound into the coil the charge of sperm for
one spermatophore is complete and the sperm mass is in final
form. ,

SEXUAL ACTIVITIES OF THE SQUID 403

As completed the sperm mass is cylindrical, with slightly
tapering ends. The surface is smooth, the coiled thread being
visible, but the coils are not prominent. The free surface is
covered by a small amount of the mucilaginous material. In
staining, the spermatozoa take haematoxylin, or other nuclear
stains, and the mucilaginous material eosin. Scattered sper-
matozoa are found in the mucilaginous material, but there are
not many of them. In this condition the sperm mass appears
much as it does in the completed spermatophore, except that
the coils of the sperm thread are a little more open and more
easily seen. ‘The change is to be accounted for by the pressure
applied in the completed spermatophore by the elastic outer
tunic.

As the sperm mass passes back through the mucilaginous
gland, the groove in which it lies is formed by an overhanging
ridge, an arrangement that becomes very prominent in the suc-
ceeding part of the spermatophoric organ (fig. 36, GR).

As the sperm mass passes out of the mucilaginous gland the
cement body is attached to the end which leaves the gland last.
This body is evidently formed by the mucilaginous gland, but
I have not observed the actual process of formation. It has
been seen immediately after it has left this gland, and, as it
must be formed before the coiled filament is laid down, and the
coiled filament is formed just beyond the mucilaginous gland,
there can be no alternative as to its place of formation.

So far, I have not been able to determine whether parts one
and two of the mucilaginous gland have the same function.
Possibly one of these portions is concerned in the formation of
the cement body alone, but I have not been able to find evidence
on the point. With various stains these glands appear alike
and the secretion in the sperm mass and the material of the
cement body have similar affinities for stains. There seems to
be a difference in composition however, for the cement hardens
so as to stick permanently to bodies in sea-water, while the
muci aginous material mingled with the sperm mixes freely
with sea-water and liberates the spermatozoa. This differ-
ence In composition has led me to search diligently for the exact

404 GILMAN A. DREW

place and method of formation of the cement body, but thus far
I have not been successful.

As the sperm mass and cement body leave the mucilaginous
gland and are passed along the spermatophoric organ, a thin
thread of mucilaginous material is formed which is continuous
with the cement body. This continues to be formed as the forming
spermatophore passes on, and becomes the hyaline core (fig. 23
HC) around which the spiral filament is wound. Racovitza,
(94) calls this the hyaline core in his description of the sper-
matophore of Rossia, where it evidently persists in the fully
formed spermatophore. In the squid it is present only during
the formation of the spermatophore and disappears before the
spermatophore becomes functional.

The part into which the sperm mass and cement body is passed
from the mucilaginous gland is thick-walled and granular, but
the inside is smooth, not thrown into ridges and grooves as in
the mucilaginous gland, nor are there sacules or tubules in its
structure. The inner surfaces are smooth and strongly ciliated.
The upper surface of the wall (the surface toward the visceral
mass) is thrown into a very prominent ridge (figs. 32 and 34,
GR) very similar in appearance to the typhlosole in the intestine
of an earthworm, except that it is not bilaterally symmetrical.
One margin of the ridge is drawn to the side and overhangs to
form a very definite: ciliated groove (fig. 34, @), along which the
forming spermatophore is passed, moved by the cilia and by
movements of the organ, and kept constantly rotating on its
longitudinal axis.

The general structure of this portion of the spermatophoric
organ is essentially the same from the mucilaginous gland to
the narrow duct near the anterior end of the organ, but at least
two divisions may be recognized in it. Externally the bound-
aries of these divisions are roughly marked by constrictions,
the first of which (figs. 29 and 32, C1) may be taken as the bound-
ary of the mucilaginous gland and the second (C?) the boundary
between two functional parts which show very similar structure.
Marchand refers to these two divisions jointly as the third part
of the spermatophoric gland. As the two parts are functionally

SEXUAL ACTIVITIES OF THE SQUID 405

quite different it will be convenient to refer to them by different
names.

The first part (figs. 29 and 32, HG) is slightly swollen and in
it are formed the membranes of the ejaculatory apparatus. I
therefore call it the ejaculatory apparatus gland. It is true
that one of the membranes, the inner tunic, continues over the
sperm mass so this portion actually forms more than the ejacula-
tory apparatus, but this term answers very well. The remaining
portion (figs. 29 and 32, MTG) forms the middle tunic and will
be called the middle tunic gland.

The sperm mass and cement body enter the ejaculatory appa-
ratus gland, with the hyaline core still being formed in the mucil-
aginous gland, and moves slowly through it, receiving the inner
tunic at the distal end of this gland, near the notch which sepa-
rates this gland from the middle tunic gland. The material
secreted by the glandular walls of this portion of the organ is
moved by the cilia over the edge of the ridge. The slowly
rotating sperm mass thus has this material wound around it as
a thin sheet. Parts of the gland between this point and the
mucilaginous gland are at the same time secreting materials that
are being wound into other parts. Bear in mind that after the
cement body leaves the mucilaginous gland, the hyaline core
continues to be secreted by it.

The ridge, under the edge of which the sperm mass and cement
body have passed, has a groove across its convex surface at a
point about opposite the notch marking the boundary between
the mucilaginous and the ejaculatory apparatus glands. This
groove (fig. 32, SFG) is not very deep, but is easily seen in dis-
sections of spermatophoric organs which have been preserved
in formalin. It extends diagonally from one side of the ridge
to the other and, on the side where the forming spermatophore
passes, is deep enough to join the groove in which it lies. Just
after the cement body passes its end the material that forms the
spiral filament passes along this diagonal groove and, because
of the rotation of the forming spermatophore, is wound around
the hyaline core.

Immediately beyond this groove the material for the inner
membrane is secreted and wound on as a sheet. The inner

406 GILMAN A. DREW

membrane thus covers the cement body and the outside of the
spiral filament, to both of which it adheres firmly. Orally
to the spiral filament, the inner membrane covers the hyaline
core.

A little further on the middle membrane is formed. The sheet
of which it is formed is thin, but is wound around many times
in building up this comparatively thick membrane. What
causes the aboral end to be so definitely limited has not been
determined.

The portion of the gland immediately following that which
forms the middle membrane forms the outer membrane and
that which follows, as already stated, forms the inner tunic.

All of these structures (figs 23 and 23 A), the inner membrane
(IM), middle membrane (MM), outer membrane (OM), and
inner tunic (/7’), are formed in the same manner and the gland
in which they are formed shows no definite change in structure
from one part to the other. Apparently all are being formed
at the same time and the formation of each part stops when it
is completed. There seems to be nothing visible that limits the
extent of the formation of each structure.

Some membranes adhere to others with which they come in
contact and some do not. Thus the inner membrane forms
a covering for the cement and adheres to the spiral filament
and middle membrane. The outer membrane adheres to the
inner membrane over the cement body, where they come in
contact, and to the inner tunic, but not to the middle membrane.

It is perhaps as well to call attenton to certain peculiarities
in forming structures here as anywhere. The core of mucilagi-
nous material in the sperm mass seems to be continued for-
ward into and through the cement body (fig. 23 A). The con-
necting cylinder between the two parts is very prominent. The
cement is seen to be spirally wound around this core in the par-
tially formed spermatophore and the core is continued on through-
out the length of the hyaline core as a much smaller core. The
hyaline core is evidently continuous with the cement material.

The inner core does not stain heavily with any of the stains
and seems to be distinct in composition from the cement material.

SEXUAL ACTIVITIES OF THE SQUID 407

It is much more like the mucilaginous material in the sperm
mass, but it has not just the same staining properties.

How these parts are formed is not known. Possibly the
mucilaginous substance binding the sperm mass is continu-
ous as a core and the cement substance and hyaline core are
similar substances wound around the central core. If this be
the case the mucilaginous gland must consist of two functional
parts.

A second point has to do with the spiral filament. This seems
to lie directly against the hyaline core, with the inner membrane
covering it. The space between the loops of the spiral filament
which extends from the inner membrane to the hyaline core
is evidently filled with some substance that never stains and is
apparently liquid. The hyaline core never bulges much between
the loops of the spiral filament, and the inner membrane is never
pressed in much between these loops. With the pressure that
is put upon the contents of the spermatophore when it is com-
pleted—by the elastic outer tunic, even before the hyaline core
disappears—there would be distortions were there not a support-
ing liquid in this space.

It is not difficult to understand how each of the layers described
are formed when we bear in mind that each is wound around
the slowly rotating mass as it proceeds through the duct. The
invisible part, the part connected with the nervous mechanism
that sees to it that each secretion is started and stopped at the
proper time to make the whole a complete, well-formed com-
plicated machine, is not more remarkable than many other
nervously controlled mechanisms.

The forming spermatophore has now passed well back into
the middle tunic gland, and by the time the structures described
have been completed the first formed end of the sperm mass
lies near the distal end of this gland.

As previously stated, the structure of the middle tunic gland
(figs. 29 to 36, MTG) is essentially the same as that of the ejac-
ulatory apparatus gland. The middle tunic is formed by wind-
ing a sheet of secretion around the rotating mass as in the mem-
brane described. There is a little liquid between the middle

THE JOURNAL OF MORPHOLOGY, VOL. 32, NO. 2

408 GILMAN A. DREW

and inner tunics so the two do not adhere at any place. The
middle tunic is of about uniform thickness over the part occupied
by the sperm mass. It is thinner and less granular from this
point to the oral end (figs. 23 and 23A, MT).

The forming spermatophore apparently remains in this part
of the organ for some time; the sperm mass, cement body and
part of the ejaculatory apparatus lying in the middle tunic
gland, and the forward part of the ejaculatory apparatus lying
in the ejaculatory apparatus gland. At this time the forming
spermatophore is very much larger than the completed structure.
It is sticky and soft so that when it is removed from the organ
it remains bent in any shape in which it is placed, provided the
bends be not abrupt. Before it is completed and functional,
the forward end becomes much folded so the length is greatly
decreased and all is shrunken so it is much less in diameter. The
shrinking must effect length as well as diameter. All these
changes are associated with putting on the outer tunic.

Before leaving the middle tunic gland, mention should be
made of a narrow tube that joins its distal end (figs. 31, 32, and
34, X). The lumen of this duct is lined with epithelium lying
directly on connective tissue. The walls are not glandular and
the epithelium, which is ciliated, is evidently not composed
of actively secreting cells. I am unable to assign any function
to this tube. It has been suggested, by Marchand, that it may
represent a degenerated part of the originally paired sexual ducts,
only the left of which is functional. I have no information that
throws light on this subject, but the point of junction in the
course of a highly modified section of the duct is not what might
be expected if this were the case.

I have not been able to observe the actual formation of the
outer tunic. In the specimens I have examined the outer tunic
is never present while the forming spermatophore is in the mid-
dle tunic gland. The outer tunic is always present in a sper-
matophore that has reached the next large division, which,
though I am not entirely sure of its function, I call the harden-
ing gland (figs. 29 to 36, HG). Marchand speaks of this gland
as the accessory gland (prostata), a term with no functional

SEXUAL ACTIVITIES OF THE SQUID 409

meaning. It can in no way be compared with the prostate
gland of vertebrates. The inner walls of the gland are marked
by various connecting ridges which project into the large cavity
of the gland. The gland forms a blind sae with only one duct.

Only the end of the spermatophore that contains the sperm
mass is pushed back into the hardening gland, and, as has been
said, when the spermatophore is pushed into this gland the
outer tunic is always present.

The connecting duct between the middle tunic gland and
the hardening gland is relatively small, but the walls are highly
glandular. The duct forms a bent cylindrical tube with a lumen
that corresponds pretty well with the diameter of the sperma-
tophore. I have not succeeded in removing spermatophores
passing through this portion without injuring them and in sec-
tions the injury of the spermatophores is usually considerable.
I find, however, that the material of the outer tunic is present
on parts of the spermatophore that have not reached the harden-
ing gland, so this narrow gland must be reponsible for its forma-
tion. It may therefore be called the outer tunic gland (figs.
29 to 33, OTG). Inasmuch as spermatophores are seldom found
in this gland, they probably pass through it rather rapidly.

The end containing the sperm mass is passed back into the
hardening gland to about the level of the cement body. The
region of the cement body and ejaculatory apparatus never
enter this gland. The aboral end of a spermatophore, when
present in this gland, projects into its lumen from the narrow
outer tunic gland without touching its walls. The sperma-
tophore has definite outlines, the outer tunic is fully formed
and not sticky, and the liquid in the lumen of thegland is trans-
parent and not noticeably viscid. When the gland is opened
in sea-water, the secretion that mixes with the water is visible
only because it has a different refractive index. It mixes readily
with the water and disappears. In sections of the organ the
contents of this gland frequently show some coagulated and
stained material which probably comes from the secretion.

While the aboral end of the spermatophore is in the harden-
ing gland the oral end is passed along the outer tunic gland. to

410 GILMAN A. DREW

the position of the opening from the side of this gland. This
opening communicates with a complicated portion in which
the spermatophore is completed. Marchand calls this (appen-
dix) the blind sae of the distal vas deferens. It is not a true
blind sac, as it has two openings, and the term appendix, which
has been applied by other writers, has no meaning. I will call
this (figs. 29 to 32, FG) the finishing gland. In passing the oral
end of the spermatophore from the outer tunic gland into the
opening of the duct leading to the finishing gland (fig. 32, FD)
this end of the spermatophore is considerably folded, and, as
the further movement is now with this end directed forward,
the folds are held and compressed while the outer tunic hardens.
around them.

Just how much of the gland is responsible for the formation
of the outer tunic is uncertain, but judging from the structure
of the gland, the appearance of the tunic as seen in sections, and
the appearance of spermatophores removed from the gland,
I am inclined to think that the whole structure, from the end
of the middle tunic gland to the end which extends into the
hardening gland, is very active and that the duct leading from
this portion to the finishing gland, the finishing gland duct,
and the finishing gland itself, adds to the outer tunic over the
oral end of the spermatophore.

Spermatophores taken from this position in the organ exhibit
great differences in the appearance of the oral ends, and, as
spermatophores. are common in this position, the meaning
probably is that the spermatophore is held here until the oral
end is shaped and covered. It is then passed, oral end first,
down the duct to the pointed end of the finishing gland.

The duct to the finishing gland is much larger than the lumen
of the outer tunic gland and has a very definite groove along
the side away from the visceral mass, which ends on the side
of the finishing gland in a pouch (figs. 29 to 32, PF). The
spermatophore usually lies in the part of the duct away from this
groove and pouch, but in a few cases I have found the oral end
of the spermatophore in this pouch. This position may not
have been normal, for, in opened animals, the mechanism con- .

SEXUAL ACTIVITIES OF THE SQUID 411

trolling the movements of the parts of the spermatophorie organ,
and accordingly the forming spermatophores, must be badly
interfered with.

In a spermatophore removed from this position the outer
tunic over the aboral end is well formed and normal in appear-
ance. That over the oral end is thin, somewhat opaque, and
adheres to a needle. It is most difficult to get specimens at
this stage of formation free without injury. Figure 24 shows
the oral end of the only really perfect specimen I have been able
to remove.

In passing into the finishing gland the oral end of the sper-
matophore is pressed forward into the pointed end of the gland
and evidently receives further additions to the outer tunic. The
aboral end of the spermatophore is now free from the hardening
gland and the secretions from this gland are free to find their
way into the finishing gland. Whether this actually takes place
I do not know, but the oral end, which just before was covered
by a thin, somewhat opaque and sticky outer tunic, changes
in form and appearance to that of the completed spermatophore.

The last processes in the change have to do with the formation
of the cap. As shrinkage takes place, the oral end of the ejac-
ulatory apparatus becomes further coiled and the cap region
is bulged outward by the end of the tube formed by the inner
and middle membrane (fig. 25). The cap thickens and the tube
in question is forced over and finally pressed out sidewise so
the lumen of the tube is spread to correspond to the shape of
the cap (figs. 26 and 27). The margins of the tube that come
in contact with the margins of the swollen cap are fastened
to the ridge of the outer tunic along the borders of the cap, and
further shrinkage brings the spermatophore into functional
form.

It is while in the finishing gland that the spermatophore
shrinks into the finished size and the outer tunic becomes nor-
mally turgid and elastic. Here the hyaline core disappears
probably becoming liquid. When the spermatophore starts
down the spermatophoric duct it is completed in form and ca-
pable of normal ejaculation. A slight shrinkage, especially in

412 GILMAN A. DREW

the region of the cap (fig. 28) will take place, but otherwise all
is completed.

I have not been able to determine just how the thread that
extends free from the cap is formed. It was first seen shortly
before the oral end of the spermatophore enters the spermato-
phoric duct. A small glandular tube (figs. 31 to 33, Y) lies along
the spermatophoric duct, and opens into the finishing gland
near where this gland opens into the spermatophoric duct. The
lumen of this duct is flattened in cross-section and the position
of its opening is so near the point where the thread is first seen
that I have been inclined to the belief that secretions from
this gland form the thread. J have, however, no real evidence.

The spermatophore passes down the spermatophoric duct
and enters the spermatophoric sac oral end first, with the
cap thread lying by its side. Here it reverses ends again as
the spermatophoric sac extends posteriorly beyond the sper-
matophoric duct a distance equivalent to the length of a
spermatophore.

Each successive spermatophore crowds its predecessor side-
wise and by forcing its oral end into the posterior pointed end
of the spermatophoric sac causes the preceding spermatophore
to move, aboral end forward, further into the spermatophore
sac. Successive spermatophores are thus arranged in a spiral
manner inside the sac, and the cap threads trail back from
their oral ends. The last spermatophore to enter the sac has
its oral end slightly posterior to the oral end of the spermato-
phore that preceded it into the sac.

The walls of the spermatophoric sac are muscular, and spiral
lamellae, extending into its interior (fig. 33 to 36, SS), keep
the spermatophores in position, practically parallel to each
other, but spirally arranged, with the aboral ends moving for-
ward. The muscular action of the spermatophoric sac is evi-
dently responsible for the most of the movements of the sper-
matophores it contains.

Where the spermatophoric sac joins the outer muscular duct,
commonly called the penis (a term somewhat misleading as to
function), the spermatophores largely lose their spiral arrange-

SEXUAL ACTIVITIES OF THE SQUID 413

ment and become arranged in groups of from twenty to forty
or more, parallel to each other and filling the lumen of the duct.
Thus, when they are ejected from the penis and grasped by the
hectocotylized arm, an even group, with their aboral ends for-
ward and the threads still embedded in the secretion of the penis,
is presented to the grasping arm.

SUMMARY

Spermatophore structure. The contents of the spermato-
phore are referred to as the sperm mass, cement body, and ejac-
ulatory apparatus.

The sperm mass consists of the spermatozoa surrounded by
and mixed with a mucilaginous material which mixes readily
with water. It is the proper delivery of the sperm mass that
is the essential action of the spermatophore.

The cement body contains the sticky material that finally
sticks the reservoir, into which the sperm mass is forced, in
position on the female.

The ejaculatory apparatus consists of membranous’ tubes
and structures that together form by their evagination, the
conducting tube through which the cement body and sperm
mass are forced, and the sperm reservoir into which the sperm
mass is forced.

The contents of the spermatophore, as described, are en-
closed inside a very elastic outer tunic and a middle tunic that
is elastic and capable of taking up water rapidly. Together
these tunics supply the power necessary for ejaculation of the
spermatophore.

The outer tunic is closed by a cap which is cemented in posi-
tion and may be loosened by pulling the thread connected with
the cap.

Spermatophore ejaculation. When the cap loosens the force
supplied by the elastic outer tunic and osmotic middle tunic
causes the ejaculatory apparatus to evaginate. In doing so the
two outer coats of this apparatus, the inner tunic and the outer
membrane, are reflected to form the sperm reservoir into which
the sperm mass is forced. The continued action of the outer

414 GILMAN A. DREW

and inner tunics forces the reservoir containing the sperm mass
out, ruptures the cement body and smears the cement over the
closed end of the sperm reservoir.

In this condition this body is freed from the remainder of the
spermatophore and is normally stuck in position on the female
by the cement.

Reference to the diagrams on plate 6 will aid in understand-
ing the essential processes of ejaculation.

Spermatophore formation. The spermatophore is completely
formed inside of that portion of the sexual duct called the sper-
matophoric organ. This is a complicated series of continuous
glands, in the lumens of which the forming mass is kept rotat-
ing on its longitudinal axis. By this rotation the sperm mass,
cement body, spiral filament, and the various enclosing mem-
branes are spirally twisted and wrapped into position as the
mass moves along the lumen of the organ.

When fully formed, the whole spermatophore undergoes a
shrinking process by which the elastic outer membrane is left
in a state of high tension which makes the whole spermatophore
turgid and ready to ejaculate.

LITERATURE CITED

Papers that mention Cephalopod spermatophores and spermatophorie or-

gans are very numerous, but only a few have been cited in this paper.

Drew, GILMAN A. 1911 Sexual activities of the squid, Loligo pealii (Les.)
I. Copulation, egg-laying and fertilization. Jour. Morph., vol. 22.

MarcHanp, WERNER 1907 Studien tiber Cephalopoden. I. Der minnliche
Leitungsapparat der Dibranchiaten. Zeit. f. wiss. Zool., Bd. 86.

Racovirza, Emine-G. 1894 Notes de Biologie. III. Moeurs et Reproduc-
tion de la Rossia macrosoma. Arch. d. Zool. Exper. et Gen., (3), T. 2.

Wituiams, Leonarp W. 1903(?) The anatomy of the common squid, Loligo
pealii (Les.). Amer. Mus. Nat. Hist.

SEXUAL ACTIVITIES OF THE SQUID 415

EXPLANATION OF FIGURES

With the exception of figures 32 and 37 to 41, all of the figures were outlined
with the aid of a camera lucida. Where cut surfaces are shown on spermato-
phores and spermatophoric organs the relations were worked out by study and
shown for convenience of those interested in the paper. It is hardly necessary
to say these were added after the camera-lucida sketches were made.

The sizes of spermatophores and spermatophorie organs differ with the sizes
of the individuals from which they were obtained. This explains the differences
in the size of the figures. The spermatophores were all drawn with the same
magnification except figures 1, 12, 16, and 17, which are not so highly magnified.
The sections of the spermatophores shown in figures 18 and 22 were consider-
ably broken in preparation. While their outlines were obtained with the aid
of a camera lucida, the damage was repaired by study. The figures are not from
sections of the same spermatophore.

Large spermatophores measure 16 mm. in length, but 13 mm.is more common.
During formation they are larger, but they shrink to their final size after they
are otherwise fully formed.

ABBREVIATIONS

C, cap covering the oral end of the spermatophore.

C1, constriction separating the mucilaginous gland from the ejaculatory appara-
tus gland. ‘i

C2, constriction separating the ejaculatory apparatus gland from the middle
tunie gland,

CB, cement body.

CB', cement liberated by rupture of cement body.

CC, connecting cylinder between the sperm mass and the cement body. In
forming specimens the connecting cylinder is continuous with material
which extends the length of the sperm mass, cement body, and hyaline core.
In fully formed specimens this material may sometimes be distinguished
in places.

CT, cap thread. When this is pulled the cap is normally loosened and the sperm-
atophore ejaculates.

EA, ejaculatory apparatus. This term is slightly misleading, as the process of
ejaculation is not confined to this part.

EG, ejaculatory apparatus gland. This term is not quite accurate, as other
portions than the so-called ejaculatory apparatus are formed by this gland.

F, folds that serve to wind the sperm thread into a spiral. In the position shown
in figure 34 they are much smaller than they are a little further along in the
groove.

FD, finishing gland duct, connecting the finishing gland with the outer tunic
gland.

FG, finishing gland. Where the cap and cap thread are formed and where the
shrinking of the spermatophores is completed.

(7, groove along which forming spermatophores pass. In some figures the forming
spermatophores are présent.

416 GILMAN A. DREW

GR, gland ridge; typhlosole-like in appearance. Under one edge of this ridge
is the groove along which the forming spermatophores are passed.

HC, hyaline core. Present in forming spermatophores, but later disappears,
probably by liquefication, possibly by withdrawal to the cement body.
HG, hardening gland. This may not be properly named. Only the aboral end
of the spermatophore is thrust into this gland. In this position the aboral
end of the spermatophore is always covered by the outer tunic, which is
smooth elastic, and not sticky. The hardening of the oral end of the sper-
matophore takes place in the finishing gland, possibly by secretions deliv-
ered with the spermatophore from the hardening gland, possibly by secretions

furnished by the finishing gland itself.

IM, inner membrane. A membrane of the ejaculatory apparatus and a cover-
ing for at least a portion of the cement body. On its inner surface it bears
the spiral filament. It is so thin it has been represented by a line.

IT, inner tunic. Inconspicuous and represented by a line over the sperm mass
and connecting cylinder, becoming thicker and more conspicuous over the
ejaculatory apparatus, where, with the outer membrane, a double membrane
is formed. This becomes part of the covering of the sperm reservoir when
this is discharged from the spermatophore.

MM, middle membrane. A conspicuous membrane of the ejaculatory appara-
tus. The tube formed by it is firmly attached to the outer tunic at the
oral end and has its open mouth applied to the shoulder of the cement body
beneath the outer membrane.

MM‘, middle membrane, cap end. This nation ruptures when ejaculation of
the spermatophore begins.

MT, middle tunic. Probably of a highly osmotic material that furnishes part
of the power which causes ejaculation of the spermatophore.

MTG, middle tunic gland.

MG‘, mucilaginous gland, part one.

MG?, mucilaginous gland, part two. The separate functions of these two parts
have not been determined, but together they form the secretions with which
the spermatozoa are mixed, and which form the,connecting cylinder, the
hyaline core, and the cement body.

OM, outer membrane. A portion of the ejaculatory apparatus. For most of
its length it is intimately associated with the inner tunic so the two appear
as a double membrane. The tube which it forms is applied to the middle
membrane and ends with a free opening near the oral end of the spermato-
phore. With the inner tunic it forms the sperm reservoir.

OT, outer tunic. A highly elastic tough outer covering. This, together with
the middle tunic, furnishes the power that causes ejaculation. When the
spermatophore nears completion this tunic shrinks until it is under great
tension and the spermatophore becomes very turgid as the result.

OTG, outer tunic gland. It is possible this may not be responsible for the for-
mation of the outer tunic, but it probably is.

PA, point where adhesion betwee: the middle and inner membranes covering
the cement body becomes strong. Form this point to the end of the middle
membrane they adhere firmly. As the spermatophore nears completion,
the point of adhesion is not so easily seen, but during ejaculation the adhesion
is seen to be strong.

SEXUAL ACTIVITIES OF THE SQUID 417

PF, pouch on the finishing gland, of unknown function. This pouch is connected
with the lumen of the finishing gland duct and with the finishing gland it-
self as a sort of diverticulum.

PR, point of rupture of the inner membrane. During ejaculation the inner mem-
brane and the spiral filament separate from the oral end of the cement body,
As ejaculation proceeds they form the outer covering of the tube through
which the sperm mass is forced. The extreme outer end of this tube is free
from them as the middle membrane, which forms the extreme outer end,
extends along the cement body past the point of rupture.

S, 18, 19, 20, 21, 22, lines on figure 1 and figure 16 indicating the planes of sections
of spermatophores represented by figures bearing the same numbers.

S, 33, 34, 35, 36, lines on figure 29 that indicate the planes of sections of the sper-
matophoric organ represented by figures bearing the same numbers.

SD, spermatophorie duct, connecting the finishing gland with the spermato-
phorie sac. When a spermatophore starts into this duct from the finishing
gland it is completely formed, except that a slight shrinking, especially in
the region of the oral end, will still take place. By the time the spermato -
phoric sac is reached the shrinking is complete.

SF’, spiral filament. This is fastened to the inner membrane and seems to serve
to keeping lumen of the ejaculatory apparatus open. The material of
which it is composed is brittle and the filament is broken into small frag-
ments as the tubes composing the ejaculatory apparatus are everted. The
same letters have been used for the filament while the coils are distinct and
for the broken fragments that remain sticking to the outside of the evagi-
nated inner membrane. See figure 7.

SFG, spiral filament groove. The material from which the filament is formed is
passed along this groove to the forming spermatophore which is passing along
the groove underneath the ridge across which the spiral filament groove
cuts.

“SL, space filled with liquid. This liquid originates in the middle tunic gland.
It does not stain and evidently has only the double purpose of lubrication
and transmission of pressure.

SM, sperm mass. In lettering the same letters have.been used for the mass of
spermatozoa, whether in position in the spermatophore, during ejaculation,
or in the sperm reservoir after ejaculation is complete. It should be borne
in mind that the arrangement is changed so the original sperm mass is dis-
organized entirely by the time it reaches the sperm reservoir. As the dis-
organization is a continuous process in ejaculation, it seems more confusing
to attempt to designate it by different letters and names than to use the
same letters with this explanation.

SS, spermatophoric sac. The receptacle in which the completed spermatophores
are stored. Because of size it is shown only in the figures of cross-sections
of the spermatophoric organ. It is really not a part of the spermatophoric
organ, but a storage receptacle. It receives the spermatophores from the
spermatophoric duct which comes from the finishing gland, and delivers them
through the penis.

VD, vas deferens. This plaited tube joins the testis with the spermatophoric
organ and delivers the completely formed spermatozoa to it.

418 GILMAN A. DREW

VDO, vas deferens opening into the mucilaginous gland. The opening is pro-
vided with a sphincter muscle and the spermatozoa are allowed to enter only
at definite intervals.

X, a duct of unknown function. A ciliated, not glandular, duct which opens
into the distal extremity of the middle tunic gland. It has been suggested
that this represents the vestige of the right vas deferens, but this seems
rather doubtful.

Y, a glandular duct of unknown function that joins the finishing gland near the
opening to the spermatophorie duct.

PLATE 1
EXPLANATION OF FIGURES

1 Spermatophore completely formed as taken from the spermatophorie sac.
The cap thread is shown broken at a little less than one-half the normal length.
At this magnification the outer tunic is represented by a single line and ejacula-
tory apparatus details are not shown. X 20 diameters.

2 Oral end of a spermatophore. X 70 diameters.

3 Oral end of a spermatophore. Represented as cut when fresh so the ejac-
ulatory apparatus has expanded, uncoiled and thrust back through the cut ends
of the outer and middle tunics. In this condition the free oral ends of the outer
membrane, OM, and the inner tunic /7’, are more easily seen. X 70 diameters.

4 Oral end of a spermatophore. Shown with the cap thread loosenipg. X
70 diameters.

5 Oral end of a spermatophore, after the cap has opened and the ejacula-
tory apparatus had begun to evaginate. X 70 diameters.

6 Oral end of a spermatophore, after the oral unornamented portion of the
ejaculatory apparatus has evaginated and before the portion bearing the spiral
filament has begun to evaginate. There is a slight pause at this stage of ejac-
ulation. X 70 diameters.

7 Oral end of a spermatophore , when the portion of the ejaculatory appara-
tus bearing the spiral filament is evaginating. The spiral filament is broken
into small fragments in the act of evagination. The fragments, which remain
sticking to the inner membrane (now on the outside), are responsible for the
broad, indefinite, spiral ornamentation on the outside of the evaginated tube.
X 70 diameters.

'8 A portion of the evaginating ejaculatory apparatus at a slightly later in-
terval than shown in figure 7. This shows the relation of the membranes in the
evaginating position. > 70 diameters.

SEXUAL ACTIVITIES OF THE SQUID
GILMAN A, DREW \ PLATE 1

‘MCN IT SU MT OT ..C- iem
CB MMOMIT SL MT wi

419 420

PLATE 2

EXPLANATION OF FIGURES

9 Extremity of the evaginating portion of a spermatophore at the instant
the cement body has reached the end of the nearly evaginated ejaculatory ap-
paratus. The inner membrane and spiral filament have broken from the tip of
the cement body and are seen at the limit of ornamentation on the outside. The
middle membrane adheres closely to the cement body, but the pressure from
behind has not yet caused the cement body to change shape. > 70 diameters.

10 A portion of the evaginated ejaculatoy apparatus of the same specimen
shown in figure 9, taken some distance from the oral extremity, at the point
where the thickened portions of the outer membrane and inner tunic now lie.
Evagination has turned these membranes back on the inside of the ejaculatory
apparatus where they now form a reservoir wall into which the disorganized
sperm mass is being forced. X 70 diameters.

11 Extremity of the evaginating portion of the spermatophore an instant
later than shown by figure 9. The pressure from behind has caused the adher-
ing middle membrane to draw out the oral portion of the cement body at the
sides. X 70 diameters.

12 A whole spermatophore shown at a stage of ejaculation just a little more
advanced than shown by figure 11. For convenience in placing on the plate the
spermatophore is drawn as if cut in two parts. The position of the sperm mass
which is being forced through the evaginated ejaculatory apparatus is shown.
x 20 diameters.

13 Extremity of the evaginating portion of the spermatophore shown in
figure 12. This is an instant later than the stage shown by figure 11. The ad-
hering middle membrane has drawn the cement body out to form a cap over
the end of the sperm mass which is being forced against it. XX 70 diameters.

14 A slightly later stage than that shown by figure 13. X 70 diameters.

15 Extremity of the evaginating portion of aspermatophore immediately after
the pressure has caused the cement body, to which the middle membrane has
adhered, to burst the inner membrane which has confined its viscid cement ma-
terial. This act at once liberates the cement, which is spread over the end of
the reservoir wall that encloses the sperm mass, and frees the reservoir wall,
which consists of the outer membrane and inner tunic, now stretched and forced
together so their individuality can no longer be distinguished, so it may slip out
of the evaginated middle membrane, against which the outer membrane lies.
< 70 diameters.

16 The empty case, consisting of the outer and middle tunics, and the evag-
inated middle and inner membranes with the broken fragments of the spiral
filament, after the sperm mass with the enclosing membranes and cement have
been discharged. XX 20 diameters.

17 Thesperm mass with the enclosing membranes and with the cement spread
over the closed end, after being ejected from the case. This mass may be called
the sperm reservoir. The walls consist of the stretched outer membrane and
inner tunic, which are open at the pointed end. Here spermatozoa leaves as
they are mixed with sea-water and become active. The cement hardens in sea-
water and sticks the reservoir in place. The thickened portion near the open-
ing, with the constricted portion immediately beyond it, is characteristic of
the reservoirs. It may have something to do with the thickened portions of

(Continued on page 426)
422

SEXUAL ACTIVITIES OF THE SQUID
GILMAN A. DREW

HN Suns
» Oem

THE JOURNAL OF MORPHOLOGY, VOL, 32, No. 2

PRM pou

424

(Continued from page 422)

the inner tunic and the outer membrane, but I am not certain this is the ex-
planation. X 20 diameters.

18 Transverse section of a spermatophore through the region of the spiral
filament. For position see figure 1. X 70 diameters.

19 Transverse section of a spermatophore through the region of the oral end
of the cement body. For position see figure 1. X 70 diameters.

20 Transverse section of a spermatophore through the region of the aboral
end of the cement body. For position see figure 1. X 70 diameters.

21 Transverse section of a spermatophore through the region of the sperm
mass. For position see figure 1. % 70 diameters.

22 Transverse section through the case of an ejaculated spermatophore. For
position see figure 16. % diameters.

PLATE 3
EXPLANATION OF FIGURES

23 and 23A Two continuous portions of the oral end of the same spermat-
ophore dissected from a spermatophoric organ. The oral end of 23 was near the
point (’, figure 32. The cement body was just beyond the point C?. The aboral
end of the spermatophore (not represented in the figure) was near the distal end
of the middle tunic gland. As the middle tunic is evidently almost completely
formed, it would probably soon have been passed on to the outer tunic gland.
x 70 diameters.

24 Oral end of a forming spermatophore dissected from a spermatophoric
organ. The oral end had passed through that portion of the outer tunic gland
that connects with the middle tunic gland, and had passed into the finishing gland
duct. Theextreme end was near the point where this duct widens into the finishing
gland. The aboral end of the specimen (not shown in the figure) was well back
in the hardening gland and the cement body region was in that part of the outer
tunic gland that extends up into the hardening gland. The outer tunic was elastic
and would not adhere to a needle on that portion in the hardening gland, and was
soft and sticky over the whole oral extremity. This portion had ‘not been in
the hardening gland and would not have entered it. X 70 diameters.

25 Oral end of a forming spermatophore dissected from a spermatophoric
organ. The oral extremity had reached the pointed end of the finishing gland and
the aboral end (not shown in the figure) was in the finishing gland duct. The
whole of the outer tunic, oral as well as aboral end, was elastic and would not
adhere to a needle, but the spermatophore was not yet as turgid nor as small as
those more fully formed. The shrinking of the oral end, the formation of the
cap and the accompanying changes in the ejaculatory apparatus are the chief
features to be understood. X 70 diameters.

26 Oral end of a forming spermatophore dissected from a spermatophoric
organ. It was contained entirely within the finishing gland and had its ex-
treme oral end very near the opening of the spermatophoricduct. X70diameters.

27 Oral end of a forming spermatophore dissected from a spermatophoric
organ. The oral end had entered the spermatophoric duct. 70 diameters.

28 Oral end of a spermatophore dissected from a spermatophoric duct. A
slight shrinkage, especially at the oral end, is the only change to take place in
completing the spermatophore. X 70 diameters.

426

SEXUAL ACTIVITIES OF THE SQUID
PLATE 3

GILMAN A. DREW

OT NN IN

OT

PLATE 4

EXPLANATION OF FIGURES

29 Spermatophoric organ seen from the surface that is free from the vis-
ceral mass. X 4 diameters.

30 Spermatophorie organ seen from the surface that is applied to the vis-
ceral mass. XX 4 diameters.

31 Finishing gland end of a spermatophoric organ seen from the same posi-
tion as that is figure 29, but showing cut ends of the parts exposed by cross-
section. X 12 diameters.

32 Semidiagrammatie view of a spermatophoric organ with the parts sepa-
rated and the walls cut away to show the internal arrangements of the parts.
The second division of the mucilaginous gland has had the wall cut away so as
to expose the vas deferens as it approaches its entrance to this gland. The
figure is made from the study of many dissections, and reconstructions from the
study of sections of the organ.

430

SEXUAL ACTIVITIES OF THE SQUID PLATE 4
GILMAN A. DREW

PLATE 5

EXPLANATION OF FIGURES

33, 34, 35, 36 Transverse sections of a spermatophoric organ. For posi-
tions of sections see figure 29. The spermatophoric sac, which is not a portion
of this organ, is shown in the figures. % 12 diameters.

432

SEXUAL ACTIVITIES OF THE SQUID PLATE 3
GILMAN A, DREW

PLATE 6

EXPLANATION OF FIGURES

Diagrams showing the successive stages in the ejaculation of a spermatophore.
For simplicity everything not necessary for understanding the process has been
omitted from these figures. Thus the cap, the inner membrane and spiral fila-
ment, and the inner tunic are not shown.

37 Diagram of the spermatophore ready for ejaculation.

38 Beginning of the process of ejaculation. The evagination of the ejac-
ulatory apparatus, represented here by the middle and outer membranes.

39 Near the end of the evagination of the ejaculatory apparatus, showing
the position of cement, sperm reservoir membrane (represented here by the outer
membrane), and spermatozoa.

40 Sperm reservoir with filled spermatozoa, and the cement body ruptured.

41 The end of the process. The separation of the discarded case and the
filled reservoir.

454

SEXUAL ACTIVITIES OF THE SQUID PLATE 6
GILMAN A. DREW

saat.
Sox,

Pe a

Fs bf

-SM

-OM

~MM

OT
“} MT

OL

a

a , = eh.

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JULY 21

A COMPARATIVE STUDY OF THE CHROMOSOMES OF

THE TIGER BEETLES (CICINDELIDAE)

WILLIAM M. GOLDSMITH
Colorado State Normal School

From the Zoological Laboratory, Indiana University!

ONE HUNDRED TWENTY-SEVEN FIGURES (TEN PLATES)

CONTENTS

MAI CTOAUC TION 2254.8 ela PR ROE POS BM A ES re be Senet 438
2. Synopsis of previous papers on Coleopteran cytology................... 438
“7 Oeil Per te Rene ae ME een BENS OEA PO ciG hed bine ati A Same Metin) yok Ihe 439
SIESTA UN AP en a sca ka tos 2 cuss sik co chp Pe aan Sea geen heres! <i el RID OPAE Sai sieiloh bv 44)
Pipe DINE ee SOREN oS Le I AS Le MEME Det cote aaa, Ses et hoenmepeer se 441
Beat onidl aaclemethouser. Rell)... ENT Sole OTA ORE ND SE DREN ha 443
AO) HSeLVicitlONs Oni the sperm ahO OTA) ses <9 eee ei ies Ces eiieices oeke ier 445
PAP SV ARC Miu Teer aend Se akct © EN nak epee BEE «Tas nleod Bd ida, CASI oak aid ag hae ape he ae 445

Bee arhyes permeate Onisestas «ace eo Scr taney monn eam any ceepeliey ere 447

Cat be SPELMI UO ROTI sd late allies tee RD creck tse We ea ear aye cn goes aoe 449

5. Growth period of the primary spermatocyte..................-0022 ee eee 450
Gee Divistonrofnet Sperm atOCy te wissen cece ters oe er crepe cae ores ee eee ores 451
rar AAI LOROMIEHM | ek) pe ee seo ay ath Dane re tare Ue and Ren ected eas eae ce ort E 451
Byebhedonublerodd=chromosomeyyiccece ae ay) oe Beale sae le ke isee 452

7, Division. of the secondary spermatocyte. ..\. o. 0.0.26 1d ko oid yee eet hee 3 ale os 454
52 Lhe metamorpnosis Of the spermatids... § Slcsioa. «ce hon,- shee pace gee ote 455
PMOUSerVALIONS Gn the OOZONIA 2/155 icv hawt seek fat csine + sce einem secede eee 457
A. Oogoni4l growth period and prophase.....:..... 0.02 ...0052-e008- 457

DPB COON Ale HT OMOSOINERE 504). So aang] lenge loens chains © Sartore end arta aa oe 457

40. Growth period of the primary Oo0C¢yte!..30 25 2 ac. beats sa eee Seed eaaes 458
Atetorma tion ol leptovene threads ernie sn tsin ee cls caer hike eset « 458

BMC X-CNTOMORGINES Hr IN. bt Saat. cats ae Sede ee as eaeleu cee aw aa 5 459
(CBouguet, synizesis, and, later stagesy iting nas sk ders. semen ciara. Oe! 460
HeeOpservations on the somatie.cellsyi6 2. 5. Jina dees sieusseig side pct o ok afewine 460
Mies a MEME PRI REN PARIERHG OSES 7 IE ee yay oo ca tegibe AG sue cae he Re cote eR EUS BUchele Bale Be sgese & 462
UB is ist iiaaina nis) Sar peas gah a ee ae ee ely SERRE e ds Ui ae aie a aR eS 464
[eh aa nisign oie nets Seng AeA | ta eet ee aaa 7 SR a 2 he ea 465

1 Contribution No. 162. These investigations were carried on under the

direction of Dr. Fernandus Payne to whom the writer is indebted for many
valuable suggestions.

437

438 WILLIAM M. GOLDSMITH

1. INTRODUCTION

The present paper follows the development of the germ cells
of the male from the early spermatogonial stage to the mature
spermatozoon, presents, so far as is possible, a study of the germ
cells of the female, and considers the somatic divisions. The
points of chief interest and importance are as follows:

1. The formation of syncytia in the early spermatogonia and
their relation to the later stages of cellular development.

2. The marked difference between the early and last sperma-
togonial mitoses.

3. The numerical relations between the spermatogonial, oogo-
nial, somatic, and first and second spermatocyte chromosomes.

4. The striking contrast between the behavior of the chroma-
tin material in the growth period of the primary spermatocyte
and that in the primary oocyte.

5. The presence and behavior of a double odd-chromosome.?

6. The great variety of abnormal mitotic figures and their
relation to the centrosomes and spindle fibers.

2. REVIEW OF COLEOPTERAN CYTOLOGY

Although about seventy species of beetles are referred to in
cytological literature as having been studied, the detailed his-
tory of the germ cells has been followed in only three or four
species. Practically every worker seemed to consider his task
completed when he had demonstrated the presence of an unequal
pair of heterochromosomes.

Of the sixty-eight species listed by Harvey (16) forty-five are
attributed to Miss Stevens. Since she was primarily concerned
with the sex-chromosomes, her observations on the other parts
of the cell were only incidental and usually very meager. She
gives the spermatogonial, oogonial, and first and second sperma-
tocyte numbers of chromosomes in only seven species.

4 Since this body behaves as the odd-chromosome in many forms, but in the
present material is bipartite and may at certain times appear as two distinct
chromosomes, the term ‘double odd-chromosome’ is used. Throughout this
paper large ‘X’ is used to indicate the larger element of this bivalent body and
small ‘x’ the smaller element.

CHROMOSOMES OF TIGER BEETLES 439

The chromosomes of the beetles may be divided into three
types, depending upon the behavior of the sex elements.

Type I

Stevens finds thirty species, ranging through thirteen families,
which possess an unequal pair of sex-chromosomes (‘hetero-
chromosomes’). In each case the two members appear united
on the first spermatocyte spindle, but separate and pass un-
divided to opposite poles. The union and separation of these
members vary in different species. In Haltica chalybea (Chry-
somelidae—Hydrophilidae according to Harvey, 716) the X and
Y elements are widely separated in prophase and metaphase,
but come together.in early anaphase and then again separate
and pass to opposite poles after the autosomes have divided.
In Doryphora decemlineata and D. clivicolis, X and Y appear
united in prophase and separate at various points on the spindle.
Sometimes they seem not to separate at all, but the count in the
second spermatocyte division shows that they do. In late
anaphase the larger heterochromosome is often outside the polar
mass, as is the odd-chromosome in the Orthoptera.

Wieman (’10) considers that Stevens has erroneously inter-
preted the behavior of the chromatin material in this form. He
says:

The great similarity between the telophase of the first division, as
represented by Stevens in figures 175 and 176 of her paper, and the
corresponding stage in L. signaticolis, led me to examine the ovaries
and testes of D. decemlineata. I.found the nucleolus of the primary
spermatocytes to accord with Stevens’ description as far as the resting
stage is concerned, but that its wnequal components separate in the first

division, does not seem to be the case, and in this regard I cannot agree
with her observations.

In Trirhabda virgata the X is larger than the autosomes,
while the Y is very small. In this case they are closely united
to a plasmosome. At metaphase they separate and pass to
opposite poles. In T. canadense, however, the two elements are
more loosely attached to a plasmosome. In this condition they
pass about half-way to one pole, then take their position on the
spindle, and separate before the autosomes divide.

440 WILLIAM M. GOLDSMITH

Nowlin’s work (’06) on Coptocycla aurichalcea and C. guttata,
though deficient in details, conforms with that of Stevens on
thirty species, of which those considered above are types. In
these forms Nowlin finds the unequal pair of heterochromosomes
united, as usual, at metaphase; then they separate, pass to
opposite poles, and divide normally in the second spermatocyte
division.

Miss Nichols (’10) finds that in Euchroma gigantea, ‘‘the
small heterochromosome is separated from the larger in the first
division.”’ Since she presents no further evidence and her
figure 21 shows the two elements still united, her conclusion can
have but little weight. The further meagerness of her observa-
tions is evinced by the fact that neither the spermatogonial nor
the second maturation divisions are figured. Even the number
of chromosomes is not given.

In Cicindela primeriana (’06) and C. vulgaris (09), according
to Stevens, the heterochromosome group is trilobed in the meta-
phase of the first maturation division. From this mass a small
spherical chromosome separates and leaves a larger V-shaped
one. Since each of these elements passes to opposite poles and
divides normally in the second maturation division, the Cicin-
delidae are placed with this type. The observations set forth in
the present paper, however, do not substantiate these conclusions.

It is observed that in each of the families Buprestidae, Ceram-
bycidae, Chrysomelidae, Cicindelidae (according to Stevens,
06, 09), Coccinellidae, Melandryidae, Meloidae, Scarabaeidae,
and Staphylinidae an unequal pair of heterochromosomes occurs.
The larger member of the pair is the maternal homologue of the
odd-chromosomes in the Dytiscidae, Elateridae, and Lampyridae.

The following is a typical fertilization formula for these
families:

Chelymorpha argus (Stevens, ’06)

Sperm Egg
(10+ Y)+ 00+X) =20+X4+Y=22¢
(10 +X) + (10+ X) = 200+ 2X £=22 9

CHROMOSOMES OF TIGER BEETLES 441

Type II

As was noted above, Doryphora decemlineata (Wieman, ‘10)
contains, in the growth period, a conspicuous basic-staining
bipartite body. This element takes its position on the spindle
as in the thirty or more species reported by Stevens. However,
instead of the two parts separating, the entire body passes to
one pole and divides normally in the last maturation division,
giving dimorphic spermatozoa with sixteen and eighteen chromo-
somes. D. decemlineata is the only species of Coleoptera that
has been recorded, prior to the appearance of the present paper,
in which a bivalent sex-chromosome passes undivided to one
pole. In this instance, however, neither the spermatogonial nor
oogonial number is given.

As will be shown in this paper, the double odd-chromosome
in the Cicindelidae takes the characteristic position on the first
maturation spindle and then, in contradiction to the records of
Stevens, passes to one pole undivided.

In the Carabidae, Hydrophilidae, Silphidae, and Tenebrion-
idae some species have an odd-chromosome; others, an unequal
pair of heterochromosomes. The Cicindelidae and at least one
Hydrophilid (Wieman, ’10) possess a double odd-chromosome
which passes to one pole of the first maturation spindle.

The typical fertilization formula of the last-mentioned forms
-in which the bivalent body passes undivided to one pole is as
follows:

Cicindelidae (present paper)
Sperm Egg .
104+ 104+X+x =204+ X+ x=22¢
10+X+x+(00+X +x) = 204 2X 4+ 2x = 24 9

Type III

The third type is characterized by the appearance of a single
odd-chromosome. This type is represented by twelve species
from seven families. Ten of these species were reported by
Stevens. In all cases except two, the odd-chromosome passes
undivided to one pole in the first maturation division and divides

442 WILLIAM M. GOLDSMITH

normally in the second. The bahavior, however, varies some-
what in these forms. In Limoneus griseus, for example, the
odd-chromosome, which is larger than the autosomes, lies later-
ally and in advance of the other chromosomes in metaphase, but
lags behind in anaphase and telophase. In the second division
it divides later than the other chromosomes. In Necrophorus
sayl the odd-chromosome passes to the pole in the first division
simultaneously with the autosomes, but at the periphery of the
plate; while in Chrysomela similis it passes laterally and in
advance of the other chromosomes.

Photinus consanguineus and P. pennsylvanicus are marked
exceptions to all other forms, according to Stevens (’09). In the
former the odd-chromosome divides late in the first division and
passes to the pole in the second in advance of and lateral to the
autosomes. The behavior of this body in P. pennsylvanicus is
the same as in P. consanguineus, except that there is greater
delay in passing to the pole in the second division. It will be
noted that the behavior in these two forms would be identical
with that in the type considered above were the maturation
divisions reversed.

Schafer (07) finds another variation from this third type in
Dytiscus marginalis. He describes thirty-eight chromosomes
in the spermatogonial and nineteen in each of the spermatocyte
divisions. The odd-chromosome divides in both maturation
divisions. !

Fernandez-Nonidez (’14) finds the odd-chromosome in Blaps
lusitanica, attached to another pair on the first maturation
spindle.

The families Dytiscidae, Elateridae, Lampyridae, possess the
single odd-chromosome.

The following is a typical fertilization formula for these families:

Photinus (Stevens, ’09)
Sperm Egg
9+(9+xX)=18+ X=19¢
.9+X)+9+X) = 18 + 2X = 20 9

CHROMOSOMES OF TIGER BEETLES 443

The review of Harvey (’16) shows that no sex-chromosome is
figured for a large number of species of beetles. In practically
every instance, however, the observations were either meager or
made by early workers. It might be concluded, therefore, that
up to the present time the absence of the sex-chromosome has
never been conclusively demonstrated in any species of beetles.

3. MATERIAL AND METHODS

The specimens used as a basis of this study were collected at
various points in Indiana, and a record of their distribution and
life habits is published in a separate paper (Goldsmith, 716 b).
The most important points, however, in connection with the
breeding habits of the tiger beetles should be mentioned, as a
knowledge of life histories is of fundamental importance in mak-
ing collections for cytological investigation. Shelford (08) sepa-
rates the life-histories of the members of this family into three
types as follows:

(a) Eggs laid in the late spring or early summer; larvae hibernate
usually in the third stage, pupate in the second summer; imagos emerge
about a month after pupation, hibernate, and become sexually mature
late in the third spring,—larval life lasts twelve to thirteen months,
adult life ten months,—two years between generations.

(b) Eggs laid in mid-summer; larvae hibernate usually in the third
stage, pupate in the following June; imagos emerge in early July and
become sexually mature very soon,—larval life ten months, adult life
two months, one year between generations.

(c) Eggs laid in mid-summer; larvae hibernate in the second stage,
reach the third stage early in the second summer, hibernate again and
pupate in the following May; imagos emerge in the early part of the
third summer and become sexually maturesoon,—larval life twenty-one
months,—adult life two months,—two years between generations.

It will be noted that, from the standpoint of cytological work,
it is practically impossible to discriminate between the form of
type ‘b’ and type ‘c,’ as the adult life is the same, even though
the larval life differs by eleven months. For convenience of
discussion, therefore, I have elsewhere (Goldsmith, ’16 b) used
the term ‘double-brooded’ to apply to all forms under type ‘b’
and ‘c,’ having an adult life of two months. Since the imagos of

444 WILLIAM M. GOLDSMITH

Cicindela repanda, C. purpurea, C. ancocisconensis, and C.
vulgaris emerge in late summer or early fall, hibernate, and reap-
pear in the spring, they are unquestionably classed under type
‘a’ and thus double-brooded. Specimens dissected from fall
collections of these species proved very immature, but were very
favorable for a study of the early spermatogonia and oogonia.
On the other hand, the spring collections from these double-
brooded forms were found to be of great value for a study of the
later stages of the germ cells.

Since Cicindela punctulata and C. sexguttata appear in Indi-
ana in late spring or early summer and die about eight weeks
later, they are spoken of as single-brooded. Shelford’s observa-
tions on the larvae and pupae of the latter form indicate that
this species is in reality double-brooded. However, the young
adults of the northern range do not dig their way out in the fall,
but remain in the pupal burrows until spring. Since the single-
brooded forms mature at once after appearance, there is a range
of only about three or four weeks in each year in which favor-
able cytological material can be collected.

The technical side of the study of Coleopteran cytology is
very difficult and disappointing in its results. All fixations in
common use were tried under various conditions, but none
proved entirely satisfactory. Several hundred specimens of
Cicindelidae were dissected, resulting in but a few good fixations
from each species. All of the best preparations were fixed in
Flemming’s fluid (strong). This method gave best results when
the warm solution was dropped into the body cavity of the live
specimen before dissecting out the gonads. After the fluid had
had time to penetrate slightly all parts of the body, the gonads
were removed and placed in cold Flemming for two hours. Even
under these conditions good fixations were exceptional, and no
explanation could be given for a good preparation when obtained.
One of the most perfect fixations (C. sexguttata) was one of
seventeen specimens collected, dissected, and treated under the

same conditions. The other sixteen were absolutely worthless.
Iron-haematoxylin, with orange G as a counterstain, when
needed, was used most extensively.

CHROMOSOMES OF TIGER BEETLES 445
4, THE SPERMATOGONIA

Of the five species presented in this study, only a few minor
differences in the cellular behavior were found. ‘The differences
were not sufficiently great to warrant a separate discussion for
each species. Unless otherwise specified, the descriptions and
drawing are based upon C. sexguttata.

A. Syncytia

The anterior end of each bipartite testis is a loosely arranged
tubular coil containing early spermatogonial cells. The young-
est of these are crowded with cells having no perceptible cell
wall (figs. 5, 6, and 7); the wide internuclear protoplasmic spaces
are homogeneous save for a few scattered unknown chromatin
staining bodies (figs. 5, 6, and 7). As the cells further mature,
light streaks may occur here and there in the cytoplasm, appear-
ing as cytoplasmic fibrillar bridges. With the increase in age
and size of the cells, these bridges become more dense and as-
sume a definite arrangement about a number of cells. This
continues until the entire tubule is subdivided into a large num-
ber of syncytia—cysts containing cells without perceptible cell
walls.

Wieman (710) presents a study of the cyst formation in one of
the Chrysomelid beetles, Leptinotarsa signaticollis,.in which he
concludes that the process is carried on by amitotic cells multi-
plication. He says: ‘‘At any rate, in the earliest stage at which
the cysts can be recognized, they are filled with cells undergoing
amitosis.”” Disregarding the controversy over amitotic divi-
sions in the primordial germ cells, the fact that such divisions
were found in the cyst when it is first recognizable, does not seem
to justify the conclusion that these cells are fundamentally con-
cerned in cyst formation. No such cell divisions were found
in the formation of the testicular syncytia of the tiger beetles.
In the stage represented in figures 5, 6, and 7 the syncytia are
somewhat elongated, and contain from five to eight giant nuclei
in cross-section or a total of from twenty to thirty nuclei.

446 WILLIAM M. GOLDSMITH

As to the relation between the early spermatogonia and the
syncytial membrane, we can only surmise. If, however, such a
membrane does not exist prior to its apparent formation—and
such seems to be the case—it does not seem justifiable to con-
clude that the containing cells of each cyst were all derived from
a single primordial germ cell. Here again, Wieman seems to
conclude prematurely that, “‘the contents of each cyst are the
descendants of a single mother-cell.’”’ Since the cysts were
filled with cells when first recognized, it does not seem possible
to determine whether or not all cells of a cyst were direct de-
scendants of a single cell. In fact, if the membrane of the syncy-
tium is indiscriminately formed among the early spermatogonial
cells, as seems to be the case in the Cicindelidae, this conclusion
cannot be justified.

These syncytial membranes become more and more defined
ahd persist throughout maturation. At the close of the sperma-
togonial divisions, these membranes usually are separated at
places by wide protoplasmic non-cellular spaces which increase
in size throughout spermatogenesis.

The testis of the imagos contain spermatogonial cells which are
differentiated by neither a perceptible cell wall nor a syncytial
membrane. Any region of such a testis may show various
stages of cell division. Sometimes one or more neighboring cells
in a prophase field may be in metaphase. This suggests that
there is little relation between the adjacent cells with reference
to their sequence of development. On the other hand, after
the syncytial membranes are formed, it is true, as shown in
figures 5, 6, and 7, that the stages of cellular development in
each early syncytium are the same. This unity of cellular
development persists until late in the maturation period. Fur-
thermore, a cross-section of a testicular tubule usually shows as
many distinct stages as there are syncytia represented. These
results would suggest that the contents of each syncytium, rather
than the cell itself, constitute a unit of cellular activity.

CHROMOSOMES OF TIGER BEETLES 447

B. Early spermatogonia

In the early spermatogonial nuclei there is a fine chromatin
network, not evenly distributed but more or less in clumps. In
practically every instance it was either attached to, or most
abundant near the nuclear wall and had a fibrillar connection
throughout the nucleus. As the nucleus grows, this network
becomes. broken here and there, leaving the central area almost
free from chromatin (figs. 1 and 5). Simultaneously and asso-
ciated with the breaking of the central network, the chromatin
aggregations become more conspicuous and come to lie nearer
to, or in contact with, the nuclear wall. In some instances,
however, one or more clumps of chromatin remain some distance
from the nuclear wall, but in all observed cases, they were con-
nected with this membrane by anastomosing fibrillar bridges
(fig. 1). Since the small particles of chromatin are also de-
posited on the nuclear wall, this membrane soon becomes much
more conspicuous than in earlier stages. The staining capacity
continues to increase until the nucleus reaches its maximum size
(figs. 1 and 2). This condition, as well as the formation of the
spermatogonial chromosomes, is highly suggestive of the con-
cluding stages of the growth period in the Hemiptera and other
forms. It is interesting to note that the prophase spermato-
gonial nuclei are much larger than those in the corresponding
stage of the first spermatocyte division (figs. 29 and 30). The
mean diameter in case of the former is lly, while the latter
measured only about 6u.

The above chromatin aggregations are further differentiated
by the smaller particles uniting with them, leaving the inter-
vening spaces clear. This method of intensifying the nuclear
wall by the addition of chromatin granules and the disappear-
ance of it with the withdrawal of the chromatin material again
suggest that ‘‘this membrane may be, at least in part, chromatic”
(Goldsmith, ’16 a). .

The irregular chromatin masses now begin to suggest the
shape of spermatogonial chromosomes (figs. 2, 3, and 4). These
more or less definite but granular bodies assemble about the

448 WILLIAM M. GOLDSMITH

central part of the former nuclear area and the granular cyto-
plasm crowds in from all sides. The spindle fibers now appear
and the twenty-two compact chromosomes are drawn into meta-
phase (figs. 8, 9, and 10).

The shape and size of these chromosomes vary from the large
asymmetrically armed V’s to very small spheres. Intermediate
between these two extremes are the hooked or J-shaped, the
uniformly rod-shaped, the pointed rods or club-shaped, and the
circular V’s or U’s of various shapes and sizes.

Although definite pairs of chromosomes can readily be recog-
nized in every clear spermatogonial metaphase plate, the arrang-
ing of all the chromosomes into a paired series is very unsatis-
factory. This is due principally to the fact that they vary
somewhat in shape in different plates. Since the pairing is an
arbitrary matter, the discussion on this point is confined to the
larger pairs which are recognized with greater certainty. Pair
‘A’ (figs. 8 and 9) is composed of large V-shaped chromosomes
constricted at the base and increasing in size from that point
outward. The arms seem to be about equal in size, but each
possesses its characteristic shape (fig. 10); one is somewhat
crooked, having the concave side inward, while the other is
club-shaped. The arms normally stand about 10° or 20° apart,
but the angle of divergence may vary from zero to 170°. The
V’s which are opened widely are usually found with the apex at
the periphery of the plate and with the arms extending left and
right (fig. 13). They are ofttimes constricted at the apexes to
such an extent that they appear, under low power, as two pointed
rod-shaped chromosomes with the sharp ends touching. This
condition is more often found when the arms of the V’s are
pressed almost together. This accounts for the large number of
plates that seemingly present twenty-three and twenty-four
chromosomes.

Although the hook (fig. 9, B) and pointed rod-shaped (fig. 9,
C’) chromosomes vary, they are readily recognized as pairs. The
former vary from straight rods to hooks, while the latter vary
from pointed clubs to blunt rods. The characteristic shapes of
the remaining chromosomes are not sufficiently prominent to
warrant a comparative study.

CHROMOSOMES OF TIGER BEETLES 449

A study of the anaphase stage clearly shows that the diploid,
and not haploid, number passes to each pole. It will be noted in
figure 15 b that, even though the chromosomes have coalesced,
the diploid number still persists. These observations corroborate
those of Metz on the corresponding stages in the Diptera, but
oppose the view of Lomen (’14) and Taylor (714).

The position of spindle fiber attachment varies with the shape
of the chromosomes. In case of the unequal armed V’s (pair A)
and the hooks or J’s (pair B), the fibers are attached at the apex
of the angle formed by the arms. In the open U’s the fibers
also seem to arise from the median region. In the straight rod
chromosomes the spindle attachment is terminal.

C. Late spermatogonia

The last spermatogonial cells and mitotic figures, as well as
the entire cysts, are characteristically different from those of
earlier stages. The cysts are much more pronounced, having
large intervening non-cellular spaces; they are very large and
contain many more cells than in the earlier stages. ‘The number
of cells was counted in fifty typical cross-sections, and the aver-
age for each was thirty-seven. Since each cyst continues through
about eight 5u sections, the total number of cells in each cyst
would approximate 250, allowing for those which might appear
in two sections. However, the cells here are only 7 in diameter,
as opposed to 11, in the former stages.

The secondary spermatogonial mitotic figures are much
smaller than the primary. This renders an analysis of the
spindle content much more difficult, as the same number of
chromosomes is crowded into a much smaller space. It was
impossible to study the chromosomes except at metaphase, and
even here the entire plate could not be analyzed. Figures 12
and 13 represent the most favorable plates. Neither of these,
even though they are clearer than revealed by the microscope,
shows twenty-two chromosomes. There is no doubt, however,
that the full number could be counted, were it possible to obtain
sufficiently differentiated material. Further, the chromosomes
of the late divisions differ in shape from those of the earlier.

450 WILLIAM M. GOLDSMITH

The rods are shorter and thicker, the U’s are more bean- or kid-
ney-shaped, and the conspicuous V’s are more widely open.
These changes are due, no doubt, to the increased pressure and
to the crowded condition.

5. GROWTH PERIOD OF THE PRIMARY SPERMATOCYTE

The spermatogonial telophase chromosomes, though drawn
out and confused, show in cross-section a certain degree of in-
dividuality (fig.15b). At this pot they still appear somewhat
compact, but readily change to a woolly appearance (fig. 16 b).

In the earliest growth period the chromatic material presents
itself as faint, delicately coiled threads, having no perceptible
limiting membrane (fig. 17). Only under the most favorable
conditions are the sex-chromosomes discernible. During the
formation of the leptotene stage these fibers increase in size and
staining reaction, and thus present a very crowded nucleus
(figs. 18 and 19). The entire mass of chromatic threads now
gradually contracts and culminates in a typical synaptic knot
(figs. 20 and 21). This is usually spherical and lies against the
nuclear membrane which made its appearance in the late pre-
synaptic stage. In some cases the synaptic knot is very irregu-
lar or flattened (fig. 20) and extends across the central part of
the nucleus. Here, as in case of the majority of presynaptic
leptotene nuclei, the chromatin nucleolus cannot be identified
with certainty.

The postsynaptic pachytene is inaugurated by the gradual
loosening of:the chromatic fibers of the synaptic knot. The
outer loops, which seem to persist throughout synizesis, first
recede from the chromatic bundle, giving room for the loosening
of the central fibers. This proceeds until the nucleus is again
crowded with chromatic threads (figs. 25 and 26), much coarser
and less numerous, however, than in the former leptotene nucleus.
A number of instances were noted in which the chromatin fibers
did not loosen from the synaptic knot. In these cases practically
the entire nuclear content formed a very dense sphere, and then
the entire cell degenerated (figs. 23 and 24). As the nucleus

CHROMOSOMES OF TIGER BEETLES 451

reaches the prophase condition without any perceivable split in
the chromatin threads, the diplotene stage is imperceptible and
the diffuse condition entirely lacking. ‘The heavy, densely stain-
ing chromatic rods of the prophase stages seem to be derived
directly from the pachytene nucleus. The woolly chromatic
strands of the early prophase are gradually transformed into
more definite late prophase chromatic bars (figs. 26 to 31). The
most marked differences, however, between the early and late
prophase cells are the gradual increase in nuclear size and the
development of the nucleolus from an almost non-perceptible
body to its conspicuous and characteristic late prophase form
(figs. 29 and 30).

6. FIRST SPERMATOCYTE DIVISION
A. The autosomes

Near the close of the growth period, the cytoplasm seems to
pass to one end of the cell, leaving the nuclear wall and the cell
membrane almost or quite in contact, giving the cell an elongated,
triangular appearance (figs. 29 and 30). <A single centrosome is
sometimes seen in the central part of this cytoplasmic mass some
distance from the nuclear wall. The large, woolly, chromatin
rods often give indications of polarization in the vicinity of the
appearing centrosome (fig. 29). The chromatic nucleolus now
assumes its characteristic bivalent (but unequal) appearance of
the first spermatocyte mitosis. Upon the appearance of the
spindle fibers and the breaking down of the nuclear membrane,
the cell reassumes its somewhat symmetrical form, and the
irregular chromosomes make their appearance (fig. 31). As the
eleven first spermatocyte chromosomes take their position on the
metaphase spindle, they represent almost as many types as there
are individuals, but the shape and size of each is fairly constant
at each corresponding stage of division. Figure 37 shows in
detail the characteristic shape of the average chromosomes at
metaphase and the approximate point of spindle attachment.
Spindles showing all of these elements in the same phase are
exceedingly rare, as the shape is constantly changing with the

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

452 WILLIAM M. GOLDSMITH

progress of the division of the chromosomes. For example,
many cases were found in which certain chromosomes would
bear a marked resemblance to others, but as the division pro-
gressed, the characteristic shape would be assumed. This con-
dition caused the various spindles to present, seemingly, a
great variety of chromosomes. A study of the metaphase
plate yields little results, as the number and shape of the visible
‘chromosomes depend entirely upon the point of cross-section and
the stage of development. The drawn-out chromosomes of late
anaphase soon break and form the irregular chromatic masses of
the early telophase (figs. 44 and 45). In late telophase (fig. 46)
the spindle usually condenses and gives rise to a faint midbody.

B. The double odd-chromosome

The double odd-chromosome cannot be recognized in prophase
on account of the confused condition of the autosomes; in later
phases, however, it is very conspicuous (figs. 36, 39; 41, 42, and 43)
and is surrounded by a clear area, leaving it seemingly free from
spindle attachments (fig. 42). Its disposition is, therefore, left
to the law of chance, and thus the body may appear at any
point on the spindle. In metaphase it usually appears eccentric
and in advance of the other chromosomes. On account of its
position, it may be separated from the other chromosomes either
by cutting the spindle crosswise or sagittally. This, no doubt,
accounts for the fact that many of these bodies are found seem-
ingly free in the cytoplasm, while innumerable spindles are
found which seem to lack them (figs. 32 and 38).

Under ordinary staining conditions, the double odd-chromo-
some appears as a spherical chromosome attempting an unequal
division, but when more stain is extracted, the bivalent nature
becomes more apparent. This is especially true in the anaphase
stages, when the body appears as a large flattened and a small
spherical chromosome stuck together by achromatic material.
The larger part of the element frequently shows an invagination
in the central region opposite the point of attachment of the
small member (figs. 39 and 41), which gives it the appearance of
a small single and a large double chromosome. It was, no

CHROMOSOMES OF TIGER BEETLES 453

doubt, the extreme of this condition which attracted the atten-
tion of Stevens (’09) when she suggested that, ‘‘In the first
spermatocyte spindle [of C. vulgaris] the conspicuous elements
are the trilobed heterochromosome group and a four-lobed or
cross-shaped macrochromosome’”’ (Stevens, fig. 88, reproduced
fig. 40). The “four-lobed or cross-shaped macrochromosome”’
(fig. 40, h) is evidently an early stage in the formation of the
ring chromosome (fig. 37, h, 7, and J).

There is a remarkable similarity between the first spermato-
cyte chromosomes of C. sexguttata and those of Coptocycla
aurichaleea (Nowlin, ’06) and a number of forms worked by
Stevens (06 and ’09). In the latter, however, the small and
large elements separate in anaphase and go to opposite poles,
thus giving two types of spermatozoa. One contains the large,
and the other, the small element. In the case under considera-
tion no evidence of separation has been found; while on the other
hand it is very difficult, from direct observations, to establish
the fact that such does not occur. However, the following
facts seem to be sufficient to prove conclusively that this double
element passes undivided to one pole. First, the number of
second spermatocyte chromosomes is clearly ten and twelve
(figs. 48 to 51). If the two parts of the double odd-chromosome
should pass to opposite poles, all second spermatocyte divisions
would be eleven, since the spermatogonial number is twenty-
two. Second, two chromatin nucleolei are observed in the
maturation stages of the female (figs. 95 to 99); while only one
is found in the male. Third, there is no uniformity in the orien-
tation of the double odd-chromosome (figs. 36, 39, 41, and 42).
This body is often surrounded by a clear space, and this is seem-
ingly free from fiber attachments (fig. 41). Fourth, the double
odd-chromosome has been observed at or near the pole while
the other chromosomes were in anaphase (fig. 48).

Brief reference should be made to some of the forms, exclu-
sive of the beetles, whose sex-chromosomes behave somewhat
similar to those of the Cicindelidae.

Wallace (05) claims that in Agalena naevia two large ele-
ments pass undivided to the same pole in both the first and sec-

454 WILLIAM M. GOLDSMITH

ond divisions, thus entering only one-fourth of the spermatozoa.
The number of chromosomes are given as follows: spermato-
gonial 40, first spermatocyte 19 and 19 + 2X, second spermato-
cyte 19 and 19 + 2X. Boring (’07) doubts this observation.

Davis (’08) finds in Arphia tenebrosa two bodies which may
pass to the same or opposite poles in the first division. In this
case the number of chromosomes is given as follows: spermato-
gonial 24; first spermatocyte 13, and second spermatocyte 11,
12, and 13. If the observations of Davis be correct, this one
animal is the. only exception of this nature found among the
Orthoptera.

In the pig (Wodsedalek, ’13) the double X-element of the
first spermatocyte spindle is smaller than the autosomes. Though
the parts are not united, they pass to the same pole, eccentric
and in advance of the other chromosomes.

In Syromastes marginatus (Wilson, ’09 a) the ‘double acces-
sory’ passes as a single body to one pole in the second division,
while in Phylloxera caryaecaulis and P. fallax (Morgan, ’15) and
also in Dolomedes fontanus (Painter, ’14) these clement pass
undivided to the pole in the first division.

7. SECOND SPERMATOCYTE CHROMOSOMES

The second maturation division follows immediately after the
telophase of the first, with no reconstruction of the nucleus.
The representative number of chromosomes of this division is
ten and twelve (figs. 48 to 51), while a number of plates were
found which showed eleven and thirteen chromosomes. The
cause of these aberrant numbers can be explained on the basis of
faulty technique or precocious splitting and of overlapping and
fusion of chromosomes. Observations show that the two ele-
ments of the double odd-chromosome which pass to the pole in
advance of the autosomes in the first division, separate and act
as single chromosomes in the second division. In the cells con-
taining the twelve, the X elements cannot be definitely dis-
tinguished from the other chromosomes.

The second spermatocyte chromosomes present no such
irregularities as are found in the first division, but more nearly

CHROMOSOMES OF TIGER BEETLES 455

resemble the small chromosomes of the spermatogonia. When
viewed from the side, the metaphase chromosomes appear
exceedingly uniform, but a polar view usually shows one V-shaped
and a number of irregular chromosomes. Although the present
material is not especially favorable for a comparison of the
diploid and haploid chromosomes, a close study of figures 48
and 50 might suggest that the V’s and U’s of the latter are the
diploid pairs A and B, respectively.

The metaphase chromosomes appear on the spindle as biva-
lents (fig. 52), the elements of which pass irregularly to opposite
poles (figs. 52 to 55). It is thus difficult to find an anaphase
plate showing all the chromosomes in one plane. Late ana-
phases and telophases (figs. 55 to 58) are quite uniform, however,
in comparison with the drawn-out and massed condition found
in the first maturation division (figs. 44 and 45).

8. THE METAMORPHOSIS OF THE SPERMATIDS

At the close of the second maturation division the telophase
chromatin mass is transformed into a dark spermatid nucleus,
containing large granular strands of chromatin (fig. 59). This
_ heavy network soon shows light areas, indicating a loosening of
the nuclear content (fig. 60). The chromatin strands then
become more conspicuous near the nuclear membrane, leaving
the central part almost clear (figs. 61 and 62). They continue
to condense until chromatin aggregations are formed, which
resemble prophase chromosomes of very ‘small cells (fig. 62).
Whether or not the number of aggregations of chromatin at this
point represent the haploid number of chromosomes we can
only surmise. A large chromatin nucleolus is usually very
prominent throughout these stages.

- In the early spermatid the nucleus takes a position at one
side of the irregular cell (figs. 59 to 61). The cytoplasmic part
of the cell elongates, leaving only a very thin film on one side
of the nucleus (figs. 60 to 63). In this large mass of cytoplasm
and near the nuclear wall of the early spermatid can be observed,
under very favorable conditions, a small area which seems to be
‘less granular or fibrillar than the remaining cytoplasm. This

456 WILLIAM M. GOLDSMITH

area gradually enlarges with the increase in length of the cyto-
plasmic tail, until a very conspicuous sphere is formed (figs. 60
to 63). Although this body could not be followed in other stages,
it is assumed, by comparison with other forms, to be mitochon-
drial. The increase in the radius of this mitochondrial sphere
places it in closer proximity to the nuclear wall. <A faint filament
—the first rudiment of the future axial filament—is now present
(figs. 62 and 63). At the point of attachment of the filament to
the nuclear wall, one or more small irregular bodies which are
later concealed in the middle piece can usually be observed.
Figure 63 represents the culmination of this entire process, both
in the formation of the chromosome-like bodies of the spermatid
nucleus and also in the development of the extranuclear sphere.
The chromatin masses again become less compact and are dis-
tributed quite evenly throughout the nucleus. The nucleus now
becomes more pointed at the end opposite the place of attach-
ment of the axial filament (figs. 63 to 66). The cytoplasmic-
like sphere and the cytoplasm elongate and condense to form the
tail.

The junction between the extranuclear body and the basal
part of the nucleus now becomes very dense (fig. 66). The
following processes are so exceptional that it seems impossible
for this plate to remain as the middle piece of the mature sperma-
tozoon. Especially well differentiated material revealed the
fact that this body was not a uniform plate as it usually
appeared, but that it was made up of chromatin-staining bodies
of various sizes. These chromatin-staining bodies, to which the
axial filament seems to be attached, later appear more compact,
move to one side, and then toward the anterior end of the nu-
cleus (fig. 67). This change in position causes the filament to
shift to one side and finally to come in contact with the wall of
the elongated mitochondrial body (figs. 66 to 69). As the
chromatin-staining mass proceeds forward (figs. 67 to 70), it
gives indications of a bivalent nature and soon presents two
conspicuous bodies which take their place near the forward end
of the spermatozoon (figs. 71 and 72). Asa result of this migra-
tion, the middle piece seems to become drawn out into a long

CHROMOSOMES OF TIGER BEETLES 457

granular thread which becomes continuous with the axial fila-
ment. It can only be hoped that an interpretation of these
two bodies and the significance of the migratory movement will
be revealed by future researches on this and similar forms.

A third body, which may be the acrosome, makes its appear-
ance at this time (fig. 71) and later fuses with the other two.

9. OBSERVATIONS ON THE OOGONIA
A. Oogonial growth period and prophase

The resting oogonia present little similarity to the correspond-
ing cells of the male. There is no trace of a syncytium or even
a cyst wall, but every cell is surrounded by its own conspicuous
membrane (figs. 80 to 82). In view of the suggestion that each
spermatogonial syncytium acts as a unit in the process of cell
growth, we should expect a greater irregularity in the develop-
ment of adjacent oogonial cells. This, indeed, is the case, for
even though certain regions of an ovarian follicle show in gen-
eral the same stage of development, it seems to be a matter of
rare chance for adjacent cells to proceed with their development
with the precise unity found in the corresponding spermatogonial
cells.

The oogonial nucleus is much smaller than the spermatogonial
and usually contains two well-defined chromatin nucleolei.
The remaining chromatic material is scattered more or less in a
fibrillar form throughout the nucleus. Upon the approach of
the prophase condition, this chromatic material collects in
masses, usually at the periphery of the nucleus. These chromatic
aggregations gradually condense into the prophase chromosomes.
The nuclear wall is now practically invisible as in the male
(fig. 82). :

B. Oogonial chromosomes

The oogonial number of chromosomes was practically estab-
lished when the behavior of the double odd-chromosome in the
male was determined. In order to further substantiate the
earlier observations, special effort was made not only to obtain

458 WILLIAM M. GOLDSMITH

the female count, but to extend the study of the cells of this sex
as far as possible. Over 200 slides were made before a satis-
factory count of the oogonial chromosomes was obtained. Even
after the difficulty of poor fixation was partially overcome, the
overlapping and irregular arrangement of the crowded meta-
phase chromosomes rendered the count practically impossible.
The chromosomes of a single plate were never found in the same
plane. Figure 85 shows two large chromosomes lying across
the central part of the plate, while other figures (83 to 86) show
a number of chromosomes lying at various angles to the meta-
phase plane. This suggests that they have passed the typical
metaphase condition and are approaching early anaphase. If
this be true, and there is a stage in which the chromosomes are
arranged in a single plane, a sufficiently large number of divi-
sions have been studied to justify the statement that the meta-
phase condition is practically instantaneous. The smaller chro-
mosomes often lie in such close contact with the end of the
larger that an especially good differentiation is required to dis-
tinguish the bivalent nature. A number of instances were also
noted in which a small chromosome is above, below, or in con-
tact with a larger one (fig. 84). A number of plates were also
found which showed more than the normal number of chromo-
somes. The explanation of such cases is obvious from figure
83, in which the V’s are almost perpendicular to the plane of
the plate, thus a number of the arms are cut, causing each V to
appear as two spherical chromosomes. <A very careful study,
however, of a large number of plates fully establishes the female
number as twenty-four.

10. GROWTH PERIOD OF THE PRIMARY OOCYTE
A. Formation of leptotene threads

The reconstructing nuclei at the close of the last oogonial
division differ from those of the earlier divisions. The telophase
chromosomes remain for some time as compact, irregular chro-
matin masses, with woolly or fibrillar connections (fig. 95). As
the cell begins to grow and the nuclear membrane becomes

CHROMOSOMES OF TIGER BEETLES 459

more conspicuous, definite granular threads radiate from these
ehromatin masses (figs. 95 and 96). As the number and prom-
inence of these chromatin threads increase, there is a corre-
sponding decrease in the chromatic masses (figs. 96 and 97).
This observation indicates that each leptotene thread is derived
directly from an oogonial chromatic mass. This seems espe-
cially conclusive since the estimated number of leptotene threads
approximate the number of oogonial chromosomes. In the typ-
ical leptotene nuclei (figs. 97 and 98) the long threads are twisted
and irregularly arranged throughout the nucleus. They are
usually attached to the nuclear membrane at one or both ends
.by an accumulated mass of chromatin-staining material. A
cross-section not only shows a number of cut ends, but threads
at various foci, depending upon the loop of the threads and angle
of the section. It will be noted that these threads are rendered
much more conspicuous by the granular enlargements—the
“chromomeres.

B. Sex chromosomes

The telophase of the last oogonial division shows all the chro-
matin masses of about the same density. However, in very
early growth period, only one or two condensed bodies remain,
and these show little relation to the chromatin fibers. This
becomes especially apparent a little later when all other chro-
matin bodies have been transformed into the leptotene threads.
In the earlier stages these bodies are usually compact (figs. 95,
96, and 97), but in later growth period they appear irregular
and woolly (fig. 105).

Although the evidence is not conclusive that these bodies are
the sex-chromosomes, it seems reasonable to assume that such
is the case. According to the observational evidence illustrated
in the fertilization formula, the cells of the female should con-
tain twice the amount of X-chromatin as those of the male.
That is, the male possesses X + x, while in the female cell there
should be found 2X + 2x. In accordance with these conclu-
sions, observations further indicate that the chromain content
of the chromatin nucleolus of the female will approximate twice

460 WILLIAM M. GOLDSMITH

that found in the male. In view of the fact that only two of
the four female elements are visible, we might assume that the
two large X’s have fused to form the large nucleolus, and the
two small x’s to form the small one. The great variation, how-
ever, in the size of these two bodies tends to weaken these
assumptions.

C. Bouquet, synizesis, and later stages

It has been noted that in the typical condition the leptotene
threads are scattered loosely throughout the nuclei and may be
attached to the nuclear wall at any point. Immediately follow-
ing this typical leptotene condition, the nuclear wall on the one
side becomes free from leptotene threads (fig. 99). At this
time more threads than usual are attached by only one end, the
other end floating free in the cell sap. This free end soon finds
its way to the ‘polarized’ side of the nucleus where the opposite
end is usually attached. By this method the large loops of the
bouquet stage are formed, and the nucleus is cleared on one side
of chromatin fibers (fig. 100). This method also clearly accounts
' for the fact that loops, rather than the ends of the leptotene
threads, extend outward from the chromatin mass in the bouquet
stage. As the loops are never drawn tightly together, there
never appears a compact bouquet as described in other forms.
Figure 101 is perhaps the most typical case found. No indica-
tions of a pairing of these threads were observed.

From the bouquet stage the threads emerge in broken pieces
of more or less faintly stained chromatin rods (figs. 102 to 104).
These appear very irregular and feathery, until the stage repre-
sented in figure 105 is reached. In this and later stages the
chromatin material is scattered uniformly throughout the large
nucleus in the form of faint anastomosing aggregations.

11. THE SOMATIC CELLS AND MITOSES

The follicular tissue of the ovaries proved very satisfactory for
a study of the somatic cells. Although the majority of the
cells are in a resting condition, mitotic divisions are compara-

CHROMOSOMES OF TIGER BEETLES 461

tively abundant.. The active cells show little indication of
rhythm of cell activity.

The somatic resting cells (fig. 74) are much smaller than the
germ cells. ‘They usually possess two irregular feathery nucleoli.
The chromatin material is scattered somewhat uniformly
throughout the nucleus with here and there slight feathery
aggregations of fibrillar material. The nuclear wall is very
conspicuous, while the cell wall is somewhat less apparent. The
entire cell, in case of the ovarian follicles, is rectangular and
flattened, caused by the normal growth of the ovary. The
prophase stage is inaugurated in the usual way by the gradual
accumulation of chromatin material at various points in the
nucleus and by a further transformation of these aggregations
into irregularly shaped chromosomes. The formation of the
prophase somatic chromosomes differs from that of the germ
cells, especially in the testis, in that the chromatin aggregations
are formed indiscriminately throughout the nucleus rather than
in contact with or near the nuclear membrane. The somatic
cells are so small that a satisfactory study of the prophase chro-
mosomes is impossible.

Although the metaphase chromosomes were very difficult to
study on account of the crowded and flattened condition of the
cells, a number of plates were found in which the theoretical
number, twenty-four, could be definitely established (figs. 76).
The somatic chromosomes possess, in general, the characteristics
of those of the germ cells, but show much greater variations in
size, shape, and general arrangement. This is largely due to
the crowded condition of the growing ovarian tissue. Many
instances were noted in which the lateral pressure had been
sufficiently great to force the metaphase plates into an extremely
elongated form. Regardless of the variation in size and shape,
a number of pairs of somatic chromosomes can be definitely
determined. This observation indicates that the chromosomes
of the somatic and germ cells possess the same general character-
istics and that the somatic number in the female is the same as
the oogonial number (figs. 83 to 86). The anaphase condition
is characterized by a very early fusion of the chromosomes

462 WILLIAM M. GOLDSMITH

(figs. 78 to 79). A few anaphase cells were found in which the
chromosomes stood apart (fig. 77), but the number (diploid)
was too great to permit a detailed study.

The late telophase chromosomes pass directly into the diffuse
condition characteristic of the normal resting cell. There seems
to be no further changes until the prophase chromatin aggrega-
tions are formed. No indications of synizesis (as Taylor, 714,
p. 391, finds in the somatic cells of Culex pipiens) were observed
in well-fixed material.

12. ABNORMAL MITOSES

In a number of specimens a variety of abnormalities was
noted in the first spermatocyte mitoses. Figure 119 shows a
typical multiple chromosome group. The normal number of
chromosomes for this division is eleven, but twenty-two are
clearly shown in this plate. Abnormalities of this type have
been reported a number of times from other material. Metz
(16) finds (in the Diptera, notably in Sarcophaga and Funcellia),
‘certain cases of multiple chromosome numbers (tetraploid, or
higher multiple). In these cases corresponding chromosomes
were associated in prophase in aggregates of four, eight, etc.,
instead of being arranged in pairs.”’ Wilson (’06) reports in
Anasa tristis a number of oogonial cells containing forty-four
chromosomes, when the normal number is twenty-two. He
suggests that the presence of these multiple chromosome groups
is due to the fact that, ‘‘all the chromosomes divided once with-
out the occurrence of cytoplasmic division.’’ Wilson also finds
nine chromosomes in Lygaeus turcicus and in Coenus delius,
when eight is the normal number. He says, “the presence of
this additional chromosome is probably due to a failure of synap-
sis between two of the spermatogonial chromosomes which
normally conjugate to form a bivalent body, and it is evidently
to be regarded as an abnormal condition.”’

Randolph (’08) finds in the earwig, Anisolabis maritima, occa-
sional giant nuclei with double the normal number of chromo-
somes. ‘There also occur in the first spermatocyte divisions

CHROMOSOMES OF TIGER BEETLES 463

of this material, tripolar or multipolar spindles, which probably
explain a certain irregularity in the number of chromosomes.

Figures 120 to 127 show the extreme varieties of multipolar
spindles found in the material under consideration. Figure 120
shows a first spermatocyte telophase more than twice the nor-
mal size. Size relations seem to indicate that figure 124 is an
anaphase of this same condition. A large number of spindles
(figs. 121, 122, and 125) showed indications of being absorbed
in the cytoplasm. This destruction seemed always to take
place about the time the chromosomes were in anaphase, as no
telophase multipolar spindles were found. Even though indi-
vidual chromosomes, or rather chromatin bodies, were found
near the pole, the central region still held other chromatin bodies,
which seemed to be attracted equally by all poles (figs. 122,
125 to. 127). In figure 127 is a collection of irregular bodies
being acted upon by six different centrosomes.

Figure 117 shows an elongated abnormal spindle, the fibers
of which are bent around another abnormal spindle (fig. 118)
shown in cross-section. The spindle in figure 118 stands per-
pendicular to the plane of that in figures 116 and 117.

An attempt to ascribe a cause for these abnormal processes
would be, in general, very unsatisfactory. Randolph (’08) says,
“In one case of abnormal spindle it is known that the material
came from an earwig which had very recently molted; and it is
possible that there is a connection between the two facts.”
Since molting is a normal process, and further since only one
abnormal spindle was observed in one specimen, it seems pre-
mature to suggest even a possible association of the two facts.

It has been suggested that the chromosomes of the primary
spermatocyte divisions in the tiger beetles are crowded and very
irregular. This crowded condition was characteristic of all
species studied. In many cases the chromosomes were so inter-
laced that the whole spindle presented a very abnormal appear-
ance. In fact, it may be possible that the large number of
abnormal mitotic figures found in the first spermatocyte divi-
sion were conditioned partly by the crowding and interlacing of
these irregular chromosomes.

464 WILLIAM M. GOLDSMITH
13. SUMMARY

1. The behavior of the chromosomes of the beetles has been
divided into three types; each is represented by a typical fertil-
ization formula.

2. The early spermatogonia possess neither perceptible cell
walls nor syncytial membranes. However, the latter soon forms
and divides the testicular tubule into definite syncytia. All
the cells in each syncytium are in exactly the same stage of devel-
opment. This synchronism is broken, however, in the late
maturation divisions after the cell walls become apparent.
These observations would suggest that the contents of each
syncytium, rather than the cell itself, constitute a unit of cellular
activity. The early oogonia possess very definite cell walls,
there being no indication of a syncytium or a cyst.

3. The spermatogonial number of chromosomes for each of
the five species studied is twenty-two. The oogonial and the
female somatic number is twenty-four each. Two distinct
types of spermatogonia were found. The late spermatogonial
cells are much smaller and stain more intensely than do those of
the earlier divisions. The chromosomes of the late divisions are
crowded and very difficult to figure.

4. Definite pairs of chromosomes are readily recognized in
every clear spermatogonial, oogonial, and somatic metaphase
plate.

5. The eleven first spermatocyte chromosomes are very
irregular in shape and especially difficult to figure. Autosomes
in the form of complete and incomplete V’s of various sizes,
rings, hooks, and rods were figured from side views of the spindles.
The secondary spermatocyte numbers of chromosomes are ten
and twelve. They are much more uniform than those of the
first division.

6. The ‘sex-chromosome’ appears on the first spermatocyte
spindle as a double body, the two elements (X, x) of which are
very unequal in size and loosely united.

These elements neither divide nor separate in the first divi-
sion, but pass to one pole in advance of the autosomes, giving

CHROMOSOMES OF TIGER BEETLES 465

secondary spermatocytes, with ten (10) and twelve (10 + X + x)
chromosomes, respectively. In the second division the com-
ponents of the bipartite body separate, and both divide in this
division with ten and twelve chromosomes.

7. The germ cells of the female seemed to contain approxi-
mately twice the amount of X chromatin as those of the male.
This is in accordance with the fertilization formula considered
in the text. The behavior of the chromatin of the growth
period was followed through the leptotene, bouquet, and syn-
izesis stages, to the breaking up of the late synaptic threads to
form the faintly staining chromatin masses, characteristic of
the prophase egg nucleus.

8. A number of abnormal mitotic figures were observed of
which the following are types: metaphase plates containing
multiple chromosome numbers, abnormally large telophases,
spindles being absorbed in the cytoplasm, dipolar spindles con-
taining only one large chromatin elements, and spindles with
three, four, and six poles.

LITERATURE CITED

Borpas, L. 1900 Rétherches sur les organes reproducteurs males des

’ Coleoptéres. Ann. Sc. Nat. Zool., T. 11.

Borine, Autic—e M. 1907 A study of the spermatogenesis of twenty-two species
of the Membracidae, Jassidae, Cercopidae, and Fulgoridae, with
special reference to the behavior of the odd-chromosome. Jour.
Exp. Zool., vol. 4.

Davis, H. 8. 1908 Spermatogenesis of Acrididae and Locustidae. Bull.
Museum Comp. Zool., Harvard College, vol. 52.

FERNANDEZ-NONIDEZ 1914 Blaps lusitanica. Trab. Mus. Nae. de C. Nat.
de Madrid, Ser. Zool., No. 18.

Gieiio-Tos, ErmManno -1908 Mitochondri nelle cellule seminali maschili di
Pamphagus marmoratus. Biologica, vol. 2.

GotpsmiTH, WituiAM M. 1916a Relation of the true nucleolus to the linin

- network in the growth period of Psellidoes cinctus. Biol. Bull., vol. 31.
1916 b Field notes on the distribution and life habits of the tiger
beetles of Indiana. Proc. Indiana Acad. of Sci., 1916.

Guyer, M. F. 1910 Accessory chromosomes in man. Biol. Bull., vol. 19.

Harvey, Erase, Browne 1916 A review of the chromosome numbers in the
Metazoa. Jour. Morph., vol. 28.

Hoy, W. E., Jr. 1916 A study of somatic chromosomes. Biol. Bull., vol. 31.

LomEN, Franz 1914 Die Hoden von Culex pipiens L. Jena. Zeits. f. Naturw.,
Bd. 52.

466 WILLIAM M. GOLDSMITH

Metz, CHarLtes W. 1914 Chromosome studies in the Diptera, I. Jour. Exp..

Zool., vol. 17.
1916 Chromosome studies in the Diptera, II. Jour. Exp. Zool.,
vol. 21.

Montcomery, T. H. 1912 Human spermatogenesis. Jour. Acad. Nat. Sci.
Phil., vol. 15, 2nd ser.

Morean, T. H. 1915 The determination of sex in Phylloxerans and Aphids...
Jour. Exp. Zool., vol. 19.

Nicuots, M. Loutsr 1910 The spermatogenesis of Euchroma gigantea. Biol.
Bull., vol. 19.

Nowuin, W.N. 1906 A study of the spermatogenesis of Coptocycla aurichalcea.
and C, guttata with especial reference to the problem of sex deter-
mination. Jour. Exp. Zool. vol. 3.

Payne, FerRNANDUS 1909 Some new types of chromosome distribution and
their relation to sex. Biol. Bull., vol. 16.

1916 A study of the germ cells of Gryllotalpa borealis and Gryllotalpa
vulgaris. Jour. Morph., vol. 28.

Ranpoueu, H. 1908 On the spermatogenesis of the earwig, Anisolabis mari-
tima. Biol. Bull., vol. 15.

Suetrorp, V. E. 1906 Life histories and larval habits of the tiger beetles.
(Cicindelidae). Linnean Soe. Journ., Zool., vol. 30.

Stevens, N.M. 1906 a Studies in spermatogenesis, II. Carnegie Inst., Wash.,
Pub. 36, no. 2.

1908 b The chromosome in Diabrotica vittata, D. soror and D. 12
punctata. Jour. Exp. Zool., vol. 5.

1909 Further studies on the chromosomes of the Coleoptera. Jour.
Exp. Zool., vol. 6, ®

Taytor, M. 1914 The chromosome complex of Culex pipiens. Quart. Jour
Mier Sci., vol. 60.

Watuace, L. B. 1905 The spermatogenesis of the spider. Biol. Bull., vol. 8.
1909 The spermatogenesis of Agalena naevia. Biol. Bull., vol. 18.

Wireman, H. L. 1910 A study of the germ cells of Leptinotarsa signaticollis.
Jour. Morph., vol. 21.

Witson, E. B. 1905 Studies on chromosomes. I. The behavior of the idio-
chromosomes in Hemiptera. Jour. Exp. Zool., vol. 2.

1909 a The ‘accessory’ chromosome in Syromastes and Pyrrhocoris
with a comparative review of the types of sexual differences of the
chromosome groups. Jour. Exp. Zool., vol. 6.

1909 ec The female chromosome groups in Syromastes and Pyrrhocoris.
Biol. Bull., vol. 16.

WopsEDALEK, J. E., 1913 Spermatogenesis of the pig with special reference to
the accessory chromosomes. Biol. Bull., vol. 25.

| JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3 : a age ee

PLATE 1
EXPLANATION OF FIGURES

Spermatogonia

1 Early resting spermatogonial nucleus.

2, 3, and 4 Formation of prophase spermatogonial chromosomes.

5,6, and 7 Syncytia containing resting nuclei (fig. 5), prophase nuclei (6),
and mitotic figures (fig. 7).

468

CHROMOSOMES OF TIGER BEETLES PLATE 1
WILLIAM M. GOLDSMITH

469

PLATE 2

EXPLANATION OF FIGURES

Spermatogonia—Continued

8,9, and 10 Metaphase plates of the early spermatogonia. Figures 8 and 9
show the approximate pairing of the chromosomes.

11 Side view of the same stage.

12 and 13 Metaphase plates of the late spermatogonial mitoses.

14 and 15 Late anaphases showing the chromosomes approaching the poles

15 b Cross-section of an early telophase.

14, 15, and 16 Formation and development of midbodies.

16a Late spermatogonial telophase showing the formation of the cell wall
by the midbodies.

16 b Cross-sections of the nuclear region of the same stage showing the
woolly appearance of the chromatic bodies.

Spermatogonial growth period

17 The diffuse postspermatogonial stage.
18 Early leptotene nucleus.

470

CHROMOSOMES OF TIGER BEETLES PLATE 2
WILLIAM M. GOLDSMITH

PLATE 3

EXPLANATION OF FIGURES

Growth period—Continued

19 Later leptotene showing the process of unwinding and expansion completed.

20 and 21. Two cells in synizesis.

22 and 25 Pachytene nuclei.

23 and 24 Degenerating cells.

25 to 28. Early prophase nuclei. The sex-chromosomes usually appear in
this stage as a more or less spherical body.

29 and 30 Later prophases showing elongation of the cell and position of the
early centrosome. The nucleus has taken a peripheral position. The sex-
chromosomes have become more conspicuous and very irregular in form.

Primary spermatocyte

31 Prophase chromosomes being drawn into metaphase position.
32 Side view of a first spermatocyte metaphase.
33 Polar view of the same.

472

<HROMOSOMES OF TIGER BEETLES

PLATE 3
WILLIAM M. GOLDSMITH
pe 5 | | ae —. th a
a CN i
es | :
\ : | , we
WX: || :
22 > 93 24
x { ad
er
25 26 27

28

3| 32 33

473

PLATE 4

EXPLANATION OF FIGURES

Primary spermatocyte—Continued

34 and 35 Polar views of two first spermatocyte metaphase plates.

36, 38, 39, 41, 42, and 43 Early and late metaphase spindles showing the
irregularity in shape of the autosomes and the position of the ‘double odd-chro-
mosome’ (X).

37 Various forms of first spermatocyte chromosomes drawn from typical
metaphase spindles. 6 is a later stage of a; and d a later stage of c; h, 7, and
j are stages in the development of the ring chromosome shown in figures 38
and 41.

40 First spermatocyte spindle of C. vulgaris (Stevens, ’09, fig. 88), showing
the ‘‘Trilobed heterochromosome group (x) and a four lobed or a cross-shaped
macro-chromosome (h).’’ Compare X with X in figures 39, 41, and 42; also note
the similarity between h in figures 40 and h in figure 37.

44 and 45 Typical late anaphase spindles.

46 Telophase.

Secondary spermatocyte

47 An exceptional secondary spermatocyte metaphase with eleven chromo-
somes.
48 <A typical metaphase containing twelve chromosomes.

474

CHROMOSOMES OF TIGER BEETLES PLATE 4

WILLIAM M. GOLDSMITH

40 al AQ

PLATE 5

EXPLANATION OF FIGURES

Secondary spermatocyte—Continued

49 Secondary spermatocyte metaphase—twelve chromosomes.

50 and 51 Secondary spermatocyte metaphases—ten chromosomes.

52 A typical secondary spermatocyte metaphase spindle, side view.

53 and 54 Early anaphases.

55 Later anaphase.

56 and 57. Telophases.

58 An exceptional telophase.

59 Late telophase showing the reconstruction of the two daughter-cells.

Metamorphosis of the spermatids

60 Early spermatid showing loosening of nuclear content and also first
appearance of the supposed mitochondria in the cytoplasm.

60 to 63 Development of the ‘mitochondrial mass’ and elongation of the
cytoplasm. Condensation of chromatin to form definite chromatin bodies
characteristic of the spermatid nucleus.

63 <A very typical stage at the conclusion of the formation of the mitochon-
drial mass. The nucleus at its maximum size showing indications of elongation.

476

CHROMOSOMES OF TIGER BEETLES PLATE 5:

WILLIAM M. GOLDSMITH

9, a ) o- @
“e °¢ @ e® 9° ry
e@ eae ef
aS 50° 5l
,
% \ GR |
% @ |
) | 2.0 %, 3
| \ ®
ee ofen \ a3 <<
52 “
ee 68. 54.
Ate | a -
‘e . ad
™ >
55 Z 56 57
be 4 | ‘ & 59 60
&
6! 62 63

477

PLATE 6

EXPLANATION OF FIGURES

Metamorphosis of the spermatids—Continued

64 and 65 Elongation of the spherical nucleus and extranuclear body, to form
the typical head and cytoplasmic-like body shown in figure 66.

66 to 72 Further transformation of the spermatid into the spermatozoan,
showing method of elongation, disappearance of the neck plate and cytoplasmic
body, and the diffusion of the chromatin. Migration of the two unknown chro-
matin-staining bodies to the anterior end of the spermatozoan, resulting in the
production of a long, granular fiber (figs. 69 and 70) from the plate-like middle
piece (fig. 66).

73 Mature spermatozoan showing the ‘false’ head. The true head consti-
tutes, perhaps, the entire drawing, the tail being many times longer than the
illustration.

Somatic mitosis

74 <A resting somatic cell. All somatic figures are from ovarian follicles.
75 Side view of a metaphase spindle.

76 Metaphase plate—twenty-four chromosomes.

77 An exceptional anaphase.

78 and 79 Early telophases.

478

CHROMOSOMES OF TIGER BEETLES PLATE 6

WILLIAM M. GOLDSMITH

"hie . ee es: 79

479

PLATE 7

EXPLANATION OF FIGURES

Oogonia

80 Resting oogonial cell showing the two chromatin nucleoli and the arrange-
ment of the chromatin material. A very conspicuous cell membrane is found in
all the early oogonial cells.

81 Early prophase showing the newly formed chromatin aggregation.

82 Formation of prophase chromosomes—nuclear wall practically invisible.

83 to 86 Oogonial metasphase plate showing the overlapping and irregular
arrangements of the chromosomes (figs. 83 and 85, vulgaris; 84 and 86, ancocis-
conensis) .

87 Late metaphase, side view (punctulata).

88 Anaphase showing first visible appearance of the granular enlargements
of the spindle fibers which form the ‘Zwischenkorper.’

88 to 92 Successive stages in the development of the ‘Zwischenkérper.’

90 to 92 Typical telophase oogonia, showing relation between the develop-
ing Zwischenkorper and the cell wall.

93 Daughter-nuclei reconstructed before the complete division of the cell or
the disappearance of the Zwischenkorper.

94 Midbodies surrounding a bundle of spindle fibers lying free in the
cytoplasm.

480

CHROMOSOMES OF TIGER BEETLES PLATE 7
WILLIAM M. GOLDSMITH

>

eet SN ¢ =e x
(am, \ et \) 7 4 -
tee Can tt)
Veo .. . + id
¥ ie — % ¥

eae —_#

PLATE 8

EXPLANATION OF FIGURES

Oogonial growth period

95 Leptotene threads forming from the chromatin masses following the last
spermatogonial division. Chromatin nucleoli plainly visible.

96 Further spinning out of the chromatin masses to form leptotene threads.

97 Typical nuclei of leptotene stage.

98 and 99 Transition of the leptotene threads into loops in the formation of
the bouquet stage.

100 The typical, loose bouquet stage.

101 Synizesis stage.

102 to 104 Stages in the breaking up of the synaptic threads to form the
faintly staining, anastomosing chromatin masses, characteristic of later stages.

105 Typical egg which has passed through the preceding stages and is rapidly
increasing in size. Chromatin nuclei are usually very irregular and woolly.

CHROMOSOMES OF TIGER BEETLES
WILLIAM M. GOLDSMITH

7) _ » P| mae
, | Pain ~
% fe c A Ye |
af ° | ee lay
es * boo Sl 4 ty

95 || #96

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

PLATE 8

PLATE 9

EXPLANATION OF FIGURES

Spermatogonial chromosomes of other species.

106 Splitting of the metaphase chromosomes—punctulata.

107 Typical anaphase—punctulata.

108 and 109 Metaphase showing the usual twenty-two chromosomes—
punctulata.

110 and 111 Metaphase—ancocisconensis.

112 and 113 Metaphase—vulgaris.

114 Metaphase—purpurea.

Abnormal mitoses

115 A spindle containing only one large chromatin element.
116, 117, and 118 An interesting arrangement of three abnormal spindles.
Figure 118 stands perpendicular to the plane of figures 116 and 117.

484

PLATE 9

CHROMOSOMES OF TIGER BEETLES
WILLIAM M. GOLDSMITH

> << oW
prs We a iP
@ ay \ } ‘Se e é °¢
cr =) ae
oe ]
106 107 log

ae e ¢ |
ad
%) a
09g HI

a Ya

nN

Co) =
\ am

: _ |

®
ue ilketie

485

PLATE 10

EXPLANATION OF FIGURES

Abnormal mitoses—Continued

119 A multiple chromosome group.

120 Telophase more than twice the normal size.

121, 122, and 125 Abnormal spindles being absorbed in the cytoplasm.
124 Anaphase with an excessive amount of chromatin.

123, 126, and 127 Typical multipolar spindles.

486

CI:ROMOSOMES OF TIGER BEETLES PLATE 10
WILLIAM M. GOLDSMITH

120 121

129 123

124 }25

126 127

487

Resumen por el autor, Bennet M. Allen.
Universidad de Kansas.

Desarrollo de las glindulas tiroides de Bufo y su relacién normal
con la metamorfosis.

La acumulaci6n de material coloide en las glandulas tiroides de
los renacuajos de sapo coincide con la aparicién de los rudimentos
de los miembros posteriores. Las masas coloides aumentan en
tamafio y numero hasta que los miembros anteriores perforan
la piel. Esta acumulacién de material coloide esta acompanada
de un mareado aumento de tamafio en las glandulas, el cual
parece ser un resultado directo de aquella. El hecho aparente-
mente paraddéjico de la cesacién de crecimiento y actual disminu-
cién de tamafio de las glandulas tiroides y de las masas coloides
en el momento en que el proceso de la metamorfosis es mas activo,
podria explicarse en parte como el resultado de un proceso par-
cial de desecacién debido a la emergencia de los renacuajos fuera
del agua, si la salida de estos fuera de dicho medio no tuviese
lugar en un estado ulterior. La reduccién de tamafio est’ pues,
realizindose antes de que el factor citado pueda ser efectivo.
Es mucho mas probable que tal disminucién se deba a la absor-
cidn por la sangre de una cantidad considerable del coloide alma-
cenado en las glidulas, en el momento en que dicho material
puede producir mas efectos. La cola aumenta continuamente
de tamafio hasta un cierto momento, presenta una ligera, dis-
minucién y desaparece después rdpidamente. La secrecci6n
tiroidla no es causa de la disminucién de tamafio de la cola ni
aleanza unvolimen considerable antes de ser suficiente para pro-
ducir tal resultado. Es indudable que el desarrollo de los miem-
bros y el proceso de desaparicién de la cola siguen a la acumula-
cién de coloide en las glindulas tiroideas de Bufo.

Translation by José F. Nonidez
Carnegie Institution of Washington

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, AUGUST 11

THE DEVELOPMENT OF THE THYREOID GLANDS OF
BUFO AND THEIR NORMAL RELATION
TO METAMORPHOSIS!

BENNET M. ALLEN
Department of Zoology, University of Kansas

ONE PLATE (SIX FIGURES) AND ONE TEXT FIGURE

In recent years much light has been thrown upon the influence
of the thyreoid gland upon growth. Most striking are recent
studies upon the influence that this gland exerts upon meta-
morphosis in the amphibians. Gudernatsch (712) showed by
experimental means that feeding thyreoid preparations of sheep
to tadpoles of Rana greatly accelerated their metamorphosis.
This experiment has been repeated by others (Swingle, ’18) and
completely verified. In 1916 the writer was successful in re-
moving the thyreoid-gland anlagen from young tadpoles, which
were then reared to a period long after that at which meta-
morphosis should normally occur, one tadpole being kept alive
until fourteen months after the operation. In all successful
cases there was total failure to metamorphose, although the
hind limbs underwent a limited amount of growth and the fore
limbs began to develop without, however, breaking through the
skin.

Hoskins and Morris (16) suecessfully accomplished the re-
moval of the thyreoid gland in a similar way at the same time
in Rana and Amblystoma. They had some difficulty in rearing
the operated tadpoles, but were successful the following season.
Their work was reported at the December, 1916, meeting of the
American Association of Anatomists and again at the December,
1917, meeting.

In the spring of 1917 the writer repeated his earlier experi-
ments upon Rana and also performed similar operations upon

1Contribution No. 318 from the Department of Zoology, University of

Kansas.
489

490 BENNET M. ALLEN

Bufo. These experiments clearly demonstrate that the thyreoid
glands exert a marked influence upon development—an excess
of thyreoid (feeding) accelerating metamorphosis and the re-
moval of the thyreoid gland producing decided retardation of
metamorphosis. It is thus seen that these two methods of
investigation give results that are fully corroborative of one
another.

There are certain points in the solution of this problem that
can be best attacked by a study of the normal relation of these
glands to metamorphosis. It was shown in Rana pipiens that
the thyreoidless and normal control tadpoles show no differences
up to the time when the limb buds begin to develop, but as soon
as these have made their appearance in the thyreoidless tad-
poles they show a marked retardation in growth, while the body
continues to grow in size. As far as our observations go, it was
found that all somatic features of tadpole development were
thus retarded, the gonads alone being unaffected (Allen, Rogers
and Terry, 718). In order properly to interpret investigations
of this kind, it is necessary to have definite data regarding the
normal relation of the thyreoid glands to development.

The development of the thyreoid glands of Anurans was
described by Goette in Bombinator. A paper by W. Miller
gave the first complete account of the process. This was worked
out in Rana temporaria. Miller found that the thyreoid devel-
oped from a ventral downgrowth of the floor of the pharynx.
This anlage was at first unpaired, but later became divided into
two parts by the development of the copula of the hyoid car-
tilage. He traced the further development of the gland from
a condition of a solid mass to a looser texture, accompanied by
the disappearance of the pigment cells characteristic of the early
stages, and he found later that it consists of a network of twisted
cords which are surrounded by looser connective tissue. He
traced the development through later stages in which vesicles
had developed. These were composed of a single layer of epi-
thelium and filled with colloid. He found that in young frogs
immediately after metamorphosis, the thyreoid gland is wholly
made up of the vesicles containing colloid.

THYREOID GLANDS OF BUFO 491

De Meuron (’86) also described the early stages in the devel-
opment of the Anuran thyreoid glands.

Maurer (’88) worked out the development of the thyreoid
gland in Rana esculenta. He found the division of the median
anlage to take place in the 13-mm. larva. At this time the
cells were deeply pigmented, loosely arranged, and showed the
first beginnings of vesicle formation. His account of the further
development did not take it ap in great detail, but showed that
by the time of metamorphosis the thyreoid was well developed,
being composed of a large number of follicles containing colloid.
This was very clearly illustrated in one of his beautiful figures.
We thus see that the general development of the thyreoid has
been pretty thoroughly worked out. It remains, however, to
show the relation between thyreoid development and the general
body features which become modified during metamorphosis. <A
study of this kind involves close attention to the length of body,
length of tail, length of limb, ete., and a comparison of these
features with the volume of the thyreoid at various stages of
development. In order to have any force in showing the rela-
tion of the thyreoid gland to metamorphosis, this work must be
done in a roughly quantitative fashion. None of the investiga-
tors up to the present time have attacked the problem from this
angle with one exception. Leo Adler (’14) made a few observa-
tions upon the size of the thyreoid gland in different stages of
Rana temporaria. His series was made up of one each, the
length of the thyreoid gland being given in parenthesis after the
total length dimensions of each stage. His measurements were
as follows: 20 mm. (0.07 mm.); 23 mm. (0.09 mm.); 25 mm.
(0.10 mm-); 28 mm. (0.16 mm.); 30 mm. (0.21 mm.); 33 mm.
(0.82 mm.); 35 mm. (0.28 mm.); 40 mm. (0.24 mm.). He states
that the 28-mm. tadpoles have hind legs which show a division
into joints, but no statement is made as to their length. The
33-mm. tadpoles have completely developed hind legs, while
the fore legs are visible through the skin. Adler states that the
last two tadpoles (85 mm. and 40 mm.) are abnormal in size.
They had been hindered in metamorphosis, at first by tempera-
ture that was too high and, later, by temperature that was too
low.

492 BENNET M. ALLEN

MATERIAL AND METHODS

A very complete collection of tadpoles of Bufo lentiginosus
gathered in Lawrence, Kansas, afforded all of the stages, from
the first appearance of the hind-limb buds to the completion of
metamorphosis. A number of specimens were fixed in Flem-
ming’s fluid and others in bichromate acetic. Large quantities
of Bufo material fixed in 5 per cent formalin were used for a
study of the gross features of the development of the thyreoid
gland. ‘These proved to be most valuable.

I wish to express my grateful acknowledgment of facilities and
assistance accorded me by the Department of Anatomy of the
University of Illinois Medical School. The greater part of the
sections and illustrations used in this work were made by their
technicians and artists. A series of gross dissections of Bufo
were made in our University of Kansas Zoological Laboratory.

The sections were cut at a thickness of 104 and were for the
most part stained with haematoxylin and eosin. In some cases
Heidenhain’s iron alum-haematoxylin was used.

The dissections were made under a binocular microscope in
such a fashion that the exposed thyreoid glands were left adher-
ent to the hyoid cartilage, the whole being stained with alum-
cochineal. These dissections were dehydrated and then cleared
in oil of wintergreen. They were preserved and finally studied
in this fluid.

Measurements were made by means of a micrometer eyepiece,
the maximum length, breadth, and thickness being determined
in each case. In making a measurement of the thickness, the
hyoid cartilage was held vertical between two small pieces of
glass and the extreme thickness was measured in optical section.
The accompanying table 1 gives the measurements obtained from
a study of the dissections just mentioned. Length, breadth, and
thickness of the gland were multiplied together to give a rough
approximation of the volume, in effect the volume of a paral-
lelopiped that would contain the gland. The latter is flattened
oval, somewhat irregular in some instances, but for the most
part of relatively constant shape.

TABLE 1

Table of measurements of Bufo lentiginosus larvae and newiy metamorphosed
individuals

NUMBER

II a

II b

Il ec

II d

TOTAL
LENGTH

BODY
LENGTH

TAIL
LENGTH

mm,

21.6

21.1

20.9

20.1

22.2

20.8

20.1

19.5

20.4

26.2

26.1

25.0

26.0

mm.

Ceti

9.6

O29

SEG

9.9

9.9

10.0

9.7

9.1

9.5

10.5

12.3

11.2

10.8

10.4

10.4

10.9

14.4

14.2

13.9

14.1

HIND LEG
LENGTH

mm,

1.386

1.353

1.386

1.188

1.056

4.521

2.739

4.026

2.574

THYREOID

Length

mm,

R.

0.273

. 0.322

0.217
0.189

. 0.280

0.322

Breadth

VOLUME

AVERAGE
VOLUME

mm.

0.119
0.112

0.112
0.112

0.112
0.091

0.147
0.119

0.119
0.126

0.133
0.112

0.126
0.091

0.105
0.119

0.077
0.084

0.091
0.133

0.231
0.182

0.203
0.224

0.231
0.231

0.217
0.231

cmm.

0.0029
0.0032

0.0017
0.0017

0.0015
0.0014

0.0045
0.0030

0.0017
0.0021

0.0019
0.0018

0.0012
0.0008

0.0017
0.0008

0.0005
0.0005

0.0009
0.0015

0.0138
0.0155

0.0099
0.012

0.0087
0.0094

0.0117
0.0109

cmm.

0.0030

0.0017

0.0014

0.0037

0.0019

0.0018

0.0010

0.0013

0.0005

0.0012

0.0146

0.0109

0.0090

0.0113

TABLE 1—Continued

NUMBER

TOTAL

BODY

TAIL

HIND LEG

THYREOID

LENGTH| LENGTH | LENGTH | LENGTH Teneen Biscath Thick-

mm, mm. mm. mm. mm. mm. mm.

TE en) 25.85) Weal) A384 1.22673") R30:31on| 0 AS25072

L. 0.385 | 0.196 | 0.112

Thee W2359- | 5. MEA eB 2a RN Ors (Ss NOZie WO Se

L. 0.392 | 0.196 | 0.112

II g | 26.2 | 12.8] 13.4 | 4.889 | R. 0.336 | 0.210 | 0.147

L. 0.315 | 0.224 | 0.182

II h | 25.2 | 11.0] 14.2 | 4.818 | R. 0.364 | 0.252 | 0.168

L. 0.651 | 0.315 | 0.182

Ili | 25.1] 11.2| 13.9 | 2.640 | R.0.420 | 0.231 | 0.133

L. 0.378 | 0.161 | 0.122

II j | 25.8] 11.2] 14.6 | 4.884 | R. 0.385 | 0.203 | 0.161

L. 0.497 | 0.266 | 0.175

III a | 27.2 | 12.2 | 14.0 | 7.227 | R. 0.504 | 0.308 | 0.206
L. 0.434 | 0.266 | 0.168

III b | 25.7 | 12.4] 13.3 | 7.326 | R. 0.525 | 0.315 | 0.217
L. 0.511 | 0.301 | 0.161

III c | 26.0] 12.1] 13.9 | 7.194 | R. 0.462 | 0.278 | 0.175
L. 0.455 | 0.287 | 0.189

III d | 27.2 | 12.2) 15.0) 7.194 | R.0.455 | 0.301 | 0.189
L. 0.406 | 0.301 | 0.168

Ill e | 27.1] 12.4] 14.7 | 8.283 | R.0.413 | 0.301 | 0.182
L. 0.455 | 0.308 | 0.189

TIT £-) 26.9 |) 11:25 15.7) 8.283 | R. 0.427 (0.294 | 0.196
L. 0.511 | 0.294 | 0.245

TIT g 27 | ADA Ae V7 AGL) ReOc476 | Ol 3225 0152
L. 0.441 | 0.329 | 0.189

III h } 28.1} 12.2 | 15.9 | 7.524 | R. 0.420 | 0.294 | 0.182
L. 0.399 | 0.252 | 0.175

TIL i) 27.85) 10.9) "1579" | 85679 | 7. 085189) 02350), 0).224
L. 0.238 | 0.154 | 0.119

~| VOLUME

cmm

0.0061
0.0084

0.0108
0.0086

0.0107
0.0123

0.0153
0.0363

0.0102
0.0074

0.0123
0.0222

0.0325
0.0197

0.0355
0.0245

0.0211
0.0248

0.0256
0.0209

0.0211
0.0256

0.0249
0.0355

0.0276
0.0276

0.0219
0.0170

0.0400
0.0043

AVERAGE
VOLUME

cmm.

0.0072

0.0097

0.0115

0.0258

0.0088

0.0172

0.0261

0.300

0.230

0.0232

0.0233

0.0302

0.0276

0.0194

0.0221

494

TABLE 1—Continued

THYREOID

TOTAL BODY TAIL HIND LEG VOLUME AVERAGE
LENGTH| LENGTH | LENGTH | LENGTH Thick- ; VOLUME

Length Breadth ness

NUMBER

mm. mm, mm, mm, mm. mm. mm. cmm, cmm,

DE e27 Ly Wi W546. 745: | R08 42¢ | 02329) | 07259) | 0.0369
L. 0.409 | 0.357 | 0.238 | 0.0354 | 0.0361

DV a) 2828) 10.9 | 1222-) 9.636 0.280 | 0.147

. 0.371 | 0.350 | 0.210 | 0.272 | 0.0263

es 0.308 | 0.206 | 0.053

0.385 | 0.273 | 0.154 | 0.0154
. 0.392 | 0.266 | 0.133 | 0.0187 | 0.0145

IVb | 25.3} 10.5] 14.8 | 9.702

eile eh eee

. 0.413 | 0.259 | 0.168 | 0.0181
0.427 | 0.259 | 0.168 | 0.0190 | 0.0185

LV e-|°23.6)) 10:6-) 13.0") '8.948

ba

. 0.574 | 0.343 | 0.210 | 0.0407
. 0.623 | 0.322 | 0.206 | 0.0418 | 0.0412

[Vd | 25.4 | 11.3} 14.1 | 9.306

Ho

. 0.567 | 0.364 | 0.210 | 0.0433
0.609 | 0.385 | 0.206 | 0.0487 | 0.0460

iWee 2628) £1-9)| 14.9),)\9.966

. 0.497 | 0.322 | 0.210 | 0.0336

TV £) 2625 )-12-0),| \ 14.5.) 8.778
0.525 | 0.343 | 0.224 | 0.0389 | 0.0362

HR ne

0.518 | 0.329 | 0.175 | 0.0292
0.525 | 0.266 | 0.154 | 0.0211 | 0.0251

UVew | 25,00" 13s 1387 182943

ae

. 0.504 | 0.357 | 0.217 | 0.0396
0.490 | 0.315 | 0.175 | 0.0258 | 0.0327

TV jh. 25.6.) 11-5 | 14.1 110.131

IVi | 24.6} 10.4] 14.2 | 8.844 0.532 | 0.322 | 0.208 | 0.0406

. 0.490 | 0.315 | 0.231 | 0.0349 | 0.0377

He oe

.0.511 | 0.315 | 0.206 | 0.03832
0.560 | 0.336 | 0.217 | 0.0419 | 0.0375

Vey, 20-7.) 1069 | 1008) | 7.448

Vea o12298| 1026 1.3 |10.956 . 0.385 | 0.329 | 0.182 | 0.0226

. 0.413 | 0.301 | 0.210 | 0.0258 | 0.0242

0.497 | 0.3864 | 0.203 | 0.0362
0.413 | 0.413 | 0.225 | 0.0403 | 0.0381

Mob.) 4ese lO 3.3 |10.560

sisi [otis sits

. 0.490 | 0.273 | 0.175 | 0.0225
0.364 | 0.287 | 0.168 | 0.0177 | 0.0201

Wee 1) 1459h| iG 3.3 | 9.735

Vid [212297 les 1.1 |10.923 . 0.602 | 0.371 | 0.217 | 0.0488

. 0.483 | 0.371 | 0.175 | 0.0302 | 0.0395

HA

tS
(ie)
Or

TABLE 1—Concluded

THYREOID
NUMBER |, owarn|LENeTE | LuNenH | LaN@TH |]. pice. | YOL™M® | youuu:
Length Breadth Bea
ae. mm. mm, mm, mm, mm, mm, mm. eae cmm.,
Ve |13.8| 11.9] 1.9 [10.890 | R. 0.483 | 0.378 | 0.175 | 0.0310
L. 0.434 | 0.336 | 0.238 | 0.0351 | 0.0480
Vf |14.8] 11.9] 2.9 |10.857 | R. 0.392 | 0.329 |.0.161 | 0.0106
L. 0.409 | 0.315 | 0.182 | 0.0229 | 0.0167
Vg 12.3] 11.7] 0.6 11.550 | R. 0.476 | 0.336 | 0.280 | 0.0457
L. 0.409 | 0.357 | 0.287 | 0.0438 | 0.0442
Veoh | 13.405) 11-7] 44) 9801: 0.49741 0287-10-18? "010261
L. 0.532 | 0.336 | 0.231 | 0.0414 | 0.0337
Vi |12.9] 11.5] 1.4 | 9.207 | R. 0.371 | 0.266 | 0.168 | 0.0240
L. 0.427 | 0.294 | 0.196 | 0.0249 | 0.0244
Vj |13.0] 11.1 | 2.9 10.395 | R. 0.392 | 0.259 | 0.182 | 0.0182
L. 0.364 | 0.287 | 0.182 | 0.0188 | 0.0185
VI a 12.8 10.659 | R. 0.623 | 0.406 | 0.368 | 0.0788
L. 0.567 | 0.413 | 0.287 | 0.0679 | 0.0733
VI b 13.5 13.299 | R.0.616 | 0.385 | 0.287 | 0.0681
L. 0.497 | 0.441 | 0.673 | 0.0594 | 0.0637
Vic 1308 12.419 | R.0.651 | 0.392 | 0.294 | 0.0735
L. 0.567 | 0.385 | 0.301 | 0.0650 | 0.0692
VId| .5 | 12.9] .5 [12.507 | R.0.413 | 0.308 | 0.294 | 0.0368
2 2g L. 0.448 | 0.294 | 0.234 | 0.0287 | 0.0327
mH i)
[o) [o)
Vie | & | 12.3] 8 [11.055 | R. 0.658 | 0.448 | 0.204 | 0.0719
cS ne L. 0.497 | 0.455 | 0.294 | 0.0667 | 0.0792
o o
—) ~
VIf | & | 12.7] 3 {10.923 | R. 0.483 | 0.364 | 0.252 | 0.0432
Be: E L. 0.511 | 0.399 | 0.301 | 0.0612 | 0.0522
o oO
Vig| ‘sd | 12.3] ‘S (12.474 | R.0.518-| 0.315 | 0.301 | 0.0484
= EH L. 0.581 | 0.322 | 0.308 | 0.0520 | 0.0502
VI h 11.4 11.088 | R. 0.462 | 0.273 | 0.224 | 0.0273
L. 0.532 | 0.294 | 0.206 | 0.0323 | 0.0298
VI i 11.9 11.979 | R. 0.477 | 0.322 | 0.294 | 0.0445
L. 0.448 | 0.308 | 0.301 | 0.0418 | 0.0431
VI j 12.2 10.527 | R. 0.504 | 0.357 | 0.294 | 0.0506
L. 0.455 | 0.392 | 0.266 | 0.0474 | 0.0490

ass
eo)
[o>)

THYREOID GLANDS OF BUFO 497

The tadpoles selected from a jar containing many hundred
were closely matched in six representative stages with ten speci-
mens of each stage. The measurements of lengths of hind leg,
body, and total length give a basis for comparing these lots.
Of these criteria the most constant is the length of the hind
legs. These show a continuous growth, while the total length
and body length are modified by the process of metamorphosis,
the body showing distinct reduction for a time. The stages
chosen may be described as follows:

I. Hind-limb buds very small, the longest showing but faint
indications of differentiation into parts. No evidence of fore
legs (fig. 1 a).

II. Total length and body length increased. Hind limbs
showing differentiation into parts. Toes well differentiated.

III. Hind limbs decidedly larger than in preceding stage.
Continued increase in total and body length. Fore limbs formed
beneath the skin, but not yet broken through (fig. 2 a).

IV. Continued increase in length of hind limbs. Fore limbs
through the skin. Slight decrease in total length and body
length.

V. Continued increase in size of limbs. Slght increase in
body length, but marked decrease in tail length.

VI. Completion of metamorphosis (fig. 3 a).

In the main there is little need of comment upon the figures in
the accompanying tables. In stage I (fig. 4) there is little
colloid present in the follicles of the thyreoid glands and many
of the follicles are not yet formed, being represented merely by
small scattered masses of cells. They lie all in one plane at
this time, except in a few cases where they are beginning to
arrange themselves in two layers. This process is completed in
stage II, where the thyreoid glands show a distinct increase in
size and in the number and size of the component follicles (fig. 5).
This is continued through later stages. It will be seen that
there are many cases where the volume of the thyreoid glands is
not proportional to the relative length of the legs or of the body.
While this is true in a comparison between the members of group
I and group II (with the single exception of II h), there are no

498 BENNET M. ALLEN

members of the former group that have a thyreoid gland volume
approaching that of any member of group II. The same is true
in comparing group II with group III (with the single exception
of II h). These statements do not hold true, however, in com-
paring groups IV, V, and VI. In these there are a number of
cases in which members of a younger stage will show a greater
volume of the thyreoid glands than do certain individuals in the
higher groups. In fact, the average volume of the thyreoid
glands in group V is less than that of group IV. This point.
will be discussed later. Even among the metamorphosed toads,
VI h, for instance, shows a thyreoid gland volume less than the

TABLE 2

Dimensions of thyreoid glands and body measurements of Bufo lentiginosus larvae

TAIL BODY LEG > THYREOID | THYREOID:
I BODY G ID
STAGE Eee PROPOR- Paes PROPOR- eo PROPOR- pi girs 3 PROPOR-
TION TION > TION : V VOLUME TION
mm. mm. mm, cmm.

I 11.09 | 0.742 | 9.69 | 0.770 | 1.21 | 0.104 | 0.00139) 0.0518 | 0.0635
Th 13.88 | 0.954 | 11.65 | 0.926 | 3.67 | 0.314 | 0.01270) 0.5026 | 0.616
Ill 14.95 | 1.000 | 12.07 | 0.959 | 7.76 | 0.664 | 0.02612) 0.6390 | 0.771
IV 13.63 | 0.912 | 11.13 | 0.885 | 9.17 | 0.785 | 0.03161} 0.6811 | 0.834
Vv 2.91 | 0.195 | 11.58 | 0.920 | 10.49 | 0.898 | 0.03077| 0.6753 | 0.828
VI 0.00 | 0.000 | 12.58 | 1.000 | 11.69 | 1.000 | 0.05427} 0.8158 | 1.000

averages of groups IV and V, while the thyreoid volume of VI d
is surpassed by the thyreoid volume of several individuals in
each of the two preceding classes. It may be pointed out that
this is partially to be explained by the fact that these three
stages are really passed through in a relatively short period of
time. Figure 6 shows the thyreoid glands of VI f.

Table 2 shows the average dimensions of the body and of the
thyreoid gland in each of these groups. The actual dimensions.
are given, and in the following column is shown, in each case,
the proportional size of the feature as compared with the size at.
the stage when it shows its maximum development.

Text figure A gives a graphic representation of these features
as seen in table 2. In this case the growth of the hind limbs was.
taken as a standard for determining the relative stage of devel-

THYREOID GLANDS OF BUFO 499

opment in each group. With the material at hand it was impos-
sible to judge the age of the specimens. This would be quite an
unsatisfactory method of seriating material, even in laboratory-
reared specimens, because of the large amount of individual
variation in the rate of growth of tadpoles. ‘Temperature con-
ditions play a large part in determining the rate of growth. Any
attempt to regulate this factor would probably entail abnormal

Relation between thyreoid growth and metamorphosis in Bufo lentiginosus

Stage I II Ta IV Ve Wat

Hind leg length Curves mark ratio of dimen-
veoreeeee Tail length sions at various stages to di-
=——— Body length mensions of structure at com-
mae Thyreoid volume pletion of metamorphosis.

Fig. A Curves to show the relative rate of growth of the total length, body
length, and hind-limb length as compared with the growth of the thyreoid gland
during metamorphosis in Bufo lentiginosis.

conditions that would modify growth in other ways. It is
difficult, at best, to bring about normal development of tadpoles
under laboratory conditions. For these reasons it was decided
to use specimens caught under natural conditions and seriated
as indicated above.

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

500 BENNET M. ALLEN

The growth of the hind limbs was arbitrarily represented by
a straight line inclined at an angle of 45°. Upon this were marked
off intervals indicating the average length of the hind limbs at
the given stage as compared with the average length at the
stage of metamorphosis (stage VI). From each of these points
a perpendicular was dropped to the base line. These perpendicu-
lars then served to indicate the six stages chosen. Their distance
from each other serving to indicate the probable time intervals
between the different stages upon the assumption that the growth
of the hind limbs takes place at a uniform rate. Points es-
tablished upon these verticals serve to indicate the average
dimensions of various features at each of the six stages studied,
the height from the base line showing the proportion that the
dimensions of any given feature of that stage bear to its dimen-
sions at the time of metamorphosis—stage VI—maximum de-
velopment in the case of tail length. Curves were constructed by
joining these points, thus giving the proportional rate of growth
of each feature. The cube root of the volume of the thyroid
gland was employed because it would represent one dimension
of a cubical figure whose volume would roughly represent the
volume of the thyreoid gland. This appeared to be the best
criterion of comparison, because each of the other features was
represented by a one-dimension value as body length, tail length,
and hind leg length. In reality all of these features have length,
breadth, and thickness. Any influence that the thyreoid gland
would exert upon their growth would be the influence of one
solid body upon another. The length of the thyreoid gland
could not be taken as a criterion of comparison, because it in-
creases little during the stages, while the volume of the gland
increases greatly, owing to growth in thickness. Thus it seems
that the fairest basis of comparison would be to compare the
cube root of the volume of the thyreoid with the length dimen-
sions of the body, tail, and hind limb.

It is noted that the cube root of the thyreoid volume shows
a marked rise from stage I to stage II, from which the rise con-
tinues strongly to stage III, then more strongly to stage IV.
There is a slight fall in the curve from stage IV to stage V, with

’

THYREOID GLANDS OF BUFO 501

a sharp rise to stage VI. The body length shows a steady rise
to stage III, when it falls off quite distinctly. This is probably
due to the shrinkage of the intestine which brings the cloacal
opening closer to the root of the tail than it had previously been.
It is just at this time that the fore limbs have first appeared.
The partly metamorphosed toads are leaving the water at stage
V, and an appreciable loss of water from the tissues must take
place. However, the body has really begun again to increase
in size at stage V, and by stage VI it has exceeded the length
attained in stage III. The tail reaches its maximum length
in stage III, and then rapidly diminishes to the vanishing point.

The cube root of the thyreoid gland increases more rapidly
than does the length of the hind legs during the interval between
the first and second stages. This is significant in that it corre-
sponds with the results of experimental work which show that
the hind limbs develop very slowly in tadpoles from which the
thyreoid glands have been extirpated. It is thus seen that
growth of the hind legs is to a very large extent dependent upon
the growth of the thyreoid gland.

A study of sections of the thyreoid glands shows that colloid
begins to form at about the time when the hind limbs commence
to develop. Compare figures 1 b, 26, and 36 and table 3. It
increases in amount as growth continues. Measurements of
the diameter of the larger colloid masses at different stages of
development show a steady increase in size, very rapid in the
early stages, as seen in table 3. In each specimen micrometer
measurements were made of ten of the larger colloid masses.
An average was then calculated in each case (table 3).

The colloid masses increase in size until stage III, while in
V there is a diminution in size. This is observed even in some
of the metamorphosed toads, while in others the colloid masses
have reached a size beyond that found in stage III. This table
is clearly based upon too few observations to prove in itself
of much value. It is of significance, however, in that it corre-
sponds in a general way with the results of table 1. We see
that the increase in size of the thyreoid gland corresponds with
an increase in size of the colloid masses. It appears that this

502 BENNET M. ALLEN

growth of the glands is, in fact, brought about by the accumula-
tion of colloid substance. In the first stages of colloid formation
there were only from five to ten colloid masses. These soon be-
came very numerous, as shown in figures 1 b, 2b, and 3b. No
satisfactory conclusions were drawn from a study of the epithe-
lium of the follicles, although it is quite possible that important
points might be gained by an application of special methods
of technique upon the problem of the manner in which they
elaborate the thyreoid secretion.

TABLE 3
Table of measurements of colloid masses in the thyreoid gland of Bufo lentiginosus

AVERAGE COLLOID

TOTAL LENGTH BODY LENGTH HIND-LEG LENGTH DIAMETER
é
mm, mm. mm. mm,
8.4 3.4
8.4 Bere No colloid present
8.9 3.8
9.1 4.3
11.4 5.5 0.245 0.0122
14.6 6.9 0.357 0.0138
15g 8.4 0.0147
20.8 9.5 0.0284
20.8 10.6 2.805 0.0258
23.6 11.8 4.884 0.0559
24.8 11.4 6.270 0.0603
13-4 10.1 7.623 0.0499
12.0 12.0 10.164 0.0478
12.3 12.3 11.055 0.0672

SUMMARY AND CONCLUSIONS

The accumulation of colloid material in the thyreoid glands
of toad tadpoles begins just as the hind limb buds appear. The
colloid masses continue to increase in size and number until
the fore hmbs break through the skin. This accumulation of
colloid material is accompanied by a marked increase in the size
of the thyreoid glands, which appears in the main to be a direct
result of it. .

A series of observations upon the effect of thyreoid extirpation
in Rana and Bufo upon limb development have shown that the

THYREOID GLANDS OF BUFO 503

limb buds appear simultaneously in both the control and oper-
ated tadpoles. Soon after their appearance the limb buds of
the thyreoidless tadpoles lag far behind those of the normal
controls. They finally grow to an appreciable degree in spite
of the absence of the thyreoid gland, but never so fast nor to
any degree approaching the length relative to body length at-
tained in the normal controls. These observations will be ex-
tended during the coming season and published later. It is
clear, however, that the effects of thyreoid removal first become
evident in Bufo at the period when colloid normally begins to
accumulate. A comparative study along this line would give
some valuable hints upon the real significance of colloid secretion
and accumulation.

We have next to consider the apparently paradoxical fact
that there is a cessation in growth and an actual diminution in
the size of the thyreoid glands and of the colloid masses at the
very time when the process of metamorphosis is most active
(stage V). This might in part be explained as the result of a
partial drying process due to the emergence of the tadpoles from
the water, were it not for the fact that they do not emerge upon
the land until stage V. The reduction in size is thus under way
before this factor could prove effective. It is much more prob-
able that this diminution may be due to the absorption of an
unusually large amount of stored colloid at this time when it
would prove most effective. It is quite conceivable that sub-
stances might be elaborated in the blood that would enable it
to more readily dissolve the colloid and that its solvent power
might decrease again after metamorphosis has been completed.
Of course this is pure conjecture, but it is put forth in the hope
that it may: prove suggestive.

The development of the tail presents an interesting problem.
It steadily increases in size until stage III, shows a slight diminu-
tion to stage IV, and then quickly disappears. It might be
assumed that a certain amount of thyreoid secretion must be
elaborated before the absorption of the tail can be accomplished,
or, if our assumption of a more solvent condition of the blood
should prove true, it might serve to explain this point. What-

504 BENNET M. ALLEN

ever the means by which it is accomplished, we should have to
choose between two alternatives: either the thyreoid secretion
does not cause the shrinkage of the tail, or it must reach a con-
siderable volume before it is able to accomplish that result. It
is certain that limb development and the process of disappear-
ance of the tail follow the accumulation of colloid in the thyreoid
glands of Bufo.

BIBLIOGRAPHY

ApLER, Leo 1914 Metamorphosestudien an Batrachierlarven. I. Exstirpa-
tion endokriner Drusen. B. Exstirpation der Thymus. Archiv f.
Entw.-Mech. d. Organismen, Bd. 40.

Auten, B. M. 1916 The results of the extirpation of the anterior lobe of the
hypophysis and of the thyroid of Rana pipiens larvae. Science,
Noy. 24th.

1918 The results of thyroid removal in the larvae of Rana pipiens.
Jour. Exp. Zool., vol. 24.

GorTtge, A. 1875 Entwickelungsgeschichte der Unke. Leipzig, 1875.

GupERNaAtscH, J. F. 1912 Fiitterungsversuche an Amphibienlarven. Cen-
tralbl. f. Physiologie, 1912.

1912 Feeding experiments on tadpoles, etc. Arch. f. Entw.-Mech.,
Bd. 35.

Maurer, F. 1888 Schilddriise) Thymus und Kiemenreste der Amphibien.
Morph. Jahrb., Bd. 18.

Meuron, D. pe 1886 Recherches sur le développement du thymus et de la
glands thyreoide. Receuil Zool. Suisse, 1 Ser. V. 3.

Miuier, W. 1871 Uber die Entwicklung der Schilddriise. Jenaische Zeitschr.,
Bd. 6.

Hosxins, E. R. anp Marcaret Morris 1917 Thyroidectomy in Amphibia.
Anat. Rec., vol. 11.

Rocers, JAMES B. 1918 The effects of the extirpation of the thyroid upon the
thymus and the pituitary glands of Rana pipiens. Jour. Exp. Zool.,
vol. 24.

Swinete, W. W. 1918 The acceleration of metamorphosis in frog larvae by
thyroid feeding and the effects upon the alimentary tract and sex
glands. Jour. Exp. Zool., vol. 24.

TerRY, GreorGE S. 1918 Effects of the extirpation of the shania gland upon
ossification in Rana pipiens. Jour. Exp. Zool., vol. 24.

ROMA ct) Babe i
PE MN Lea ake: ‘2
Sean)
4 : iy 2 i . | ee
ce i ie Nahi? ete | | ‘

NT
gt

ie ee NK Did Wy Sepa

Jose BAe 9 rf ity hits: ete * Ma Ty: ii)
et Mics: : veay ast Rt pn : Ww } i ea 4

‘ Sa? teil Ay Mecho; ’ ‘ ‘ a (ithe
\ { ws F }
bah iC, ai WES he N igabey | ne :
thea mgt A ih i 4 ey At] ah) x9!

bt abe
ray
iT
: RE NTS
4 4 J Li
a ? 4 {
: Ha igs
ais (if ,, } ¥)
- re ;
pe ya ¢
; : ‘i
ry i
a FES 4s
-
‘ i
a’
1m s* 4 4 ; a) e ey
' y , ’
. i i + | a)
% Aye
ns ay
gi na i 4 J :
i ripe ; y - ied y ‘ ¥
TA Wwe
Re PLATE | is
iV eG , sit
F) f x r
ah | ‘ he Mid q 3 by $5
i
a }
4 i
ih 2 ib - ? > ‘ ‘ Mi
Y al A ba r J 7 fhll |
ig inal reat LAr
eg ys IL: 2M. 4 Ele ORTON Oo Care, 1 Oaeiyy Sy Tite os Te) ; i
’ re Valdes ' > b ) a p ain 5 TF
. Tas ‘Ve ‘t iret hap ae fie vadkt hs: f 2 ; ry |

i q f f : }
e 0 4 i arly AG fi 4 5 | 1 %
ell / ea , ; eg

J Y ie a

; me ‘ 7 Me

y; poet pide Ry Srey? ey.) | aN
4 i i

Sarl yt Pay A ae Ha PBN ae Soe ii: i.
:

PLATE 1
EXPLANATION OF FIGURES

la, 2a, and 3a Drawings to show typical stages in the development of Bufo
lentiginosus. Figure la represents stage I. Figure 2a represents stage III.
Figure 3a represents stage VI. All drawn to scale and magnified X 3.

1b, 2b, and 3b Drawings to show transverse sections of the thyreoid glands of
la, 2a, and 3a, respectively. Figure 1b, transverse section of the thyreoid gland
of la. Figure 2b, transverse section of the thyreoid gland of 2a. Figure 3b,
transverse section of the thyreoid gland of 3a. All drawn to scale and magnified
X 75.

4, 5, and 6 Whole mounts of the thyreoid glands of the stages represented
above. Figure 4, whole mount of thyreoids of I b. Figure 5, whole mount of
thyreoids of II h. These thyreoids have a volume almost identical with the
average in group III. Figure 6, whole mount of thyreoids of VI f. All drawn
to scale and magnified.

THYREOID GLANDS OF BUFO PLATE 1
BENNET M. ALLEN ‘

507

Resumen por el autor, Waro Nakahara.
Universidad Cornell, Ithaca.

Estudio de los cromosomas en la espermatogénesis de Perla
immarginata Say, con especial mencién del problema
de la sinapsis.

El presente trabajo es un estudio del elemento cromatico en
la espermatogénesis de Perla, hasta el final de la segunda divisién
de los espermatocitos. En el complejo espermatogonial existen
diez cromosomas (incluyendo los cromosomas Xe Y). El esper-
matocito de primer érden posee cinco cromosomas bivalentes;
los cromosomas X e Y estén fusionados entre si y aparecen como
un solo elemento. En el espermatocito de segundo 6rden existen
cinco cromosomas univalentes; cada espermatocito recibe uno de
los cromosomas X e Y. El autor discute el modo de forma-
cién de los cromosomas bivalentes. En Perla los cromosomas
del complejo espermatogonial forman parejas. Los cromosomas
homdlégos se unen por telosinapsis en el espirema del esperma-
tocito de primer 6rden, y mas tarde se incurvan uno hacia el otro
en el punto sindptico para formar anillos y tetradas antes de la
metafase de la divisién. No hay pruebas sobre la parasinapsis
en un estado temprano de la divisi6n.

Translation by José F. Nonidez
Carnegie Institution of Washington

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JUNE 21

A STUDY ON THE CHROMOSOMES IN THE SPER-
MATOGENESIS OF THE STONEFLY, PERLA IMMAR-
GINATA SAY, WITH SPECIAL REFERENCE TO THE
QUESTION OF SYNAPSIS

WARO NAKAHARA
Department of Histology and Embryology, Cornell University, Ithaca, New York.

THREE PLATES (FIFTY-ONE FIGURES)

CONTENTS
Introduction oss. srits ocd ask ee AGT SRR Mae oer crate oes Me en izist 509
“1RGYA VITO LLORAS OS NEES Bete POOR MIE AIS tt dos CaRL an Cee nt ANS © hens clomiotes 510
NUL EON es trad che RSE CE ¢ TAOS 4 pote manera Wate cie ese ee ale aees 510
rHDral CONSIGEERUIONS 2 oot iG fee eitie ce Se fet ette sce bern ree bbe o ttem ete He 513
Pairing of chromosomes and the probability of synapsis................ 513
BUGEMESLEC: MGUECS. Of SV MAUSIO GY. boon. oic's:- Sapa etyag by gcc Paes Noein she tad aaes 515
MOREE oR MITEIEEE DR Wits eae ed NA oe Sos Sed hcvocaneBORgnp) WIPE eres Se al cha, Shae eG 518
ie ret LOMIALKS Cea ata te Skee Re ket Ce Fee ace ote meee ad eae 519
ageraruncwered aia %, sahsds EER e ee. AEC Recher Reis eh 2 as ea 520
INTRODUCTION

The need of accurate knowledge of the chromosomes, and
especially their behavior in the maturation divisions, can hardly
be overemphasized. The chromosome theory of inheritance, so
beautiful a unification of biological knowledge as it is, cannot be
fully established until the true nature of synapsis and other
phenomena involved in the course of maturation divisions are
satisfactorily understood.

The present contribution to the study of the chromosomes
is based on my observations on the spermatogenesis of the
stonefly, Perla immarginata Say (Plecoptera). The fact that,
notwithstanding recent cytological activity, the order Plecoptera
has not been made a subject for chromosome study led me to
make some preliminary observations on a few species of this
group of insects during the fall of 1916. Perla immarginata,

509

510 WARO NAKAHARA

which is one of the commonest stoneflies in Ithaca (Smith, ’13),
was then determined as best fitted for minute investigations,
because of the favorable condition of its chromosomes. Since
the spring of 1917 observation on this form has been carried on
in the Laboratory of Histology and Embryology, Cornell Uni-
versity, under the supervision of Prof. B. F. Kingsbury, whose
helpful suggestions and criticisms given to me during the course
of the work are most sincerely appreciated. A prolonged obser-
vation was made with the greatest care I am capable of, and
the numerous sketches made were carefully compared, and the
working out of the history of the chromatin element was thus
ventured.
TECHNIQUE

Flemming’s strong fluid has proved to be the best for fixing
the testis dissected out in normal salt solution. Bouin’s picro-
aceto-formol has also been used very frequently. Notwith-
standing the powerful penetration of the fluid, the result was
no better than that obtained from Flemming’s fluid. Bouin’s
fluid tends to make the split line of the spireme more or less
obscure, and the material of this fixation was thus found to be
unfavorable for observations of certain critical stages. As far
as the penetration is concerned, even Flemming’s fluid seems
to be powerful enough, when applied on dissected material,
to say nothing about Bouin’s. Addition of a small amount of
urea to Bouin’s fluid (Hance, 717) did not make any noticeable
change in the fixation from the original formula.

Sections were cut from 7 to 10 uw thick and stained with Heiden-
hain’s iron hematoxylin. The longer method with this stain,
resulting in black coloring of the section, was more favorable
than the shorter ‘blue’ method, especially for the observations
on the nucleus in the early prophase of mitosis.

DESCRIPTIVE

Ten chromosomes appear in the spermatogonial division (figs.
1 and 2). Looking at a metaphase plate from either pole, all
of these ten can be individually recognized in every case. The

CHROMOSOMES OF PERLA 511

chromosome group consists of two pairs of V’s, a pair of rods,
two spherules (m-chromosomes), and two unpaired rods, one
of which is much longer than the other. These last I interpret
as X- and Y-chromosomes, respectively.

Figures 3 to 5 show the changes in the nucleus subsequent
to the last spermatogonial division. The nucleus in the late
anaphase of the division (fig. 3) gradually enters into the resting
stage (figs. 4 and 5), when the reticular appearance of the nucleus
is resumed.

The chromatin reticulum then begins to form the double
spireme. The process of the formation of the double spireme
is illustrated in figures 6 to 13, and may be best described as a
development of dual threads out of the reticulum. There is
no sign of two threads coming to conjugate side by side. It
is also impossible to say that the process involves the .actual
splitting of a simple thread. As may be seen very clearly, there
are no definitely formed and separate simple threads (i.e.,
chromosomes) in the nucleus before the development of the
double spireme commences.

The stage of contraction seems to follow the completion of
double threads. This seems to correspond to the period of
synizesis of some authors. The nuclei in this condition are
shown in figures 14 and 15, and it will be readily seen that the
contraction has nothing to do with the formation of the double
threads. The spireme at this stage is unquestionably already
double in structure. More or less regularly accompanying the
contraction nuclei, there are a number of nuclei in the process
of degeneration. Four degenerative cells are shown in figures
16 to 18.

The duality of the spireme (zygotene thread) seems to be
maintained all through the later stages (figs. 19 to 23).

The actual number of the prochromosomes could be counted
at the pachytene stage (figs. 24 to 26), when the nuclei enlarge
a little, and the spireme threads become more compact. As
can be seen in figures 24 to 34, there are six separate segments
recognizable at this stage. The smallest one, which is sometimes
attached to the largest (fig. 32), will be seen to represent the

512 WARO NAKAHARA

m-chromosome. Two other small segments, which are often
seen joined together (fig. 26), can be interpreted as representing
the X- and Y-chromosomes.

Excepting the X- and Y-segments (and the m-segment, of
which nothing definite was observed), halves of each segment
bend toward each other, until they come to lie closely side by
side. This interpretation of the process of tetrad formation
may receive justification through a comparison of figures 24
to 33.

The tetrads thus formed now arrange themselves on the
equatorial plate of the spindle (fig. 33). Polar view of the plate
shows five chromosomes as distinct bodies (figs. 35 to 37). The
X and Y are joined to each other and appear as a single body.
Of the remaining four, one is decidedly smaller than the others
and undoubtedly identifiable with the m-chromosome. Three
others, although they vary more or less in their appearances,
may therefore be considered as representing the three pairs of
the diploid chromosomes (two pairs of V’s and a pair of rods
in the spermatogonial group). Looking from the side, it will
be seen that the X and Y separate with the division of the biva-
lents (figs. 34, 38 to 40). Figures 41 to 44 show the anaphase
of the first division and the prophase of the second, in which
there is no resting stage.

If the process of ring and tetrad formation be that of an open-
ing out of a split chromosome, as it is frequently interpreted,
the space enclosed by a ring must correspond to the longitudinal
split in the zygotene thread. That this interpretation does not
hold in the case of Perla may easily be seen from a comparison
of figures 21, 27, 33, and 34. The longitudinal split of the
zygotene spireme persists as such in the ring and even in the
chromosome on the metaphase plate, and may be best inter-
preted as a precocious split for the second spermatocytic division,
which follows the first division without the ‘resting stage.’ It
has nothing to do with the space enclosed by a ring, which is
secondarily formed when the two arms of a bivalent segment
become joined.

CHROMOSOMES OF PERLA ole

The chromosome number in the second spermatocytic division
is five. The five consist of two V’s, one rod, ‘m’, and X- or
Y-chromosome (fig. 46). From what has been observed in the
preceding division, it is to be expected that half the number
of the second spermatocytes should contain one accessory chro-
mosome, and the other half of them the other, and this is ap-
parently what takes place here.

All five chromosomes divide equationally in the second sper-
matocytic division, neither of the accessories is heterotropic (figs.
44 and 45).

The anaphase of the division (figs. 48 and 49) is immediately
followed by the formation of the spermatids (figs. 50 and 51).

The further history of the spermatids has not been followed.

GENERAL CONSIDERATIONS
Pairing of chromosomes and probability of synapsis

The idea of the paired association of chromosomes was first
suggested by Sutton (’02), when he noted in Brachystola that
all chromosomes could be associated into pairs according to the
size characters. More data were accumulated later from both
zoological and botanical sides, and there seems to be no doubt
at present that, where chromosomes of different sizes and shapes
are present, there are always two of each kind (excepting acces-
sory chromosomes). In his extensive work on the topic, Metz
(16) has shown with special clearness in about eighty species
‘of Diptera which he examined, that the chromosomes are uni-
formly associated in pairs in diploid cells, in all tissues, somatic
as well as germinal, and in all stages of ontogeny (from egg to
adult), and he stated that pairing chromosomes give an actual
demonstration of a side-by-side approximation of corresponding
chromosomes.

The probability of synapsis becomes stronger, when inquiries
are made as to the nature of the pairing. In a certain Hemip-
teron, Wilson (’09) has described, beside the regular coupling
of idiochromosomes of unequal sizes, that a small supernumerary
chromosome which is indistinguishable from the m-chromosome

514 WARO NAKAHARA

always couples with the much larger idiochromosome, but never
with the m-chromosome, and suggested that the coupling results
from definite affinities among the chromosomes. He said:

The possibility no doubt exists that the couplings are produced by
extrinsic cause (such as the achromatic structure), but the evidence
seems on the whole opposed to such a conclusion. I consider it more
probable that they are due to intrinsic qualities of the chromosomes
and that the differences of behavior shown by different forms may
probably be ceo ded as due to corresponding physico-chemical differ-
ences

Very similar statements were made by Metz (’16), who said:

Pairing (of chromosomes) is not due to purely mechanical causes,
but is dependent in some way upon the qualitative nature of the chro-
mosomes. This conclusion seems evident from the fact that paired
chromosomes are corresponding or similar chromosomes. It is diffi-
cult to conceive how purely mechanical forces can cause anything more
than random pairing, while as a matter of fact the actual pairing is
selective to the highest degree. That this association is not merely
assortment according to size is shown by the pairing of unequal sex-
chromosomes in the male, where X is several times as large as Y.

Metz said further that the paired chromosomes are qualita-
tively similar and ‘‘their association is dependent upon, although
not necessarily caused by, this relation.”’ Convincingly support-
ing this statement, he pointed out that in the tetraploid groups
in Diptera,
two of the four chromosomes are sister halves of the other two, and
hence are respectively similar to them in make up. But all four of
these chromosomes associate in essentially the same manner, iLe.,°

paired chromosomes are indistinguishable from sister chromosomes in
their manner .of association.

Turning our attention to the case of Perla, we see at once that

1. The ten chromosomes in the spermatogonial group may
be grouped in pairs, and that

2. Each of the pairs is represented by a single chromencine
of corresponding appearance in the spermatocytic groups (ex-
cepting the X- and Y-chromosomes).

The X- and Y-chromosomes are seen actually coupling in
the late prophase and in the metaphase of the first spermatocytic

CHROMOSOMES OF PERLA Byles

division. The fact that these two chromosomes are of totally
different sizes and shapes, and that the coupling takes place
most regularly between these two, seems to signify much, be-
cause these afford a complete demonstration of the occurrence
of synapsis, in so far as these two chromosomes are concerned.
Although this does not show that a similar process must take
place in other pairs of chromosomes, it does, nevertheless, add
more to the probability of the general occurrence of synapsis.
However, it is evident that the case of the ordinary chromosomes
must be established through their direct study, for it is not without
reason to suspect that some differences may be found in the
process from that seen in the case of the accessory chromosomes.

Suggested modes of synapsis

In the case of Perla, facts show for certain of the chromosomes,
and hence with probability for all the other chromosomes, that
synapsis does take place in the maturation of the germ cells.
Before entering into the closer examination of the critical stages
where the process of synapsis may possibly be involved, it would
be well to review briefly the interpretations of some of the pre-
vious authors.

Vejdowsky (07) is of the opinion that the chromosomes in
normal number conjugate parasynaptically and fuse completely.
The mixochromosomes (haploid number) thus produced split
longitudinally at both divisions.

Bonnevie (07, ’08a, ’08 b, 711) considers that the diploid
chromosomes conjugate parasynaptically, and although the con-
jugants fuse completely in the maturation period, during which
they do not separate, they ultimately become distinct.

According to Henking (90-92) and Korschelt (95), the
spireme segments into the diploid number of chromosomes, all
of which undergo longitudinal splitting and remain separate until
the metaphase, when they conjugate and appear again in the
haploid number. The conjugants may separate at the first
division. Goldschmidt (08a, ’O8b) is of the same opinion,
but he maintains that the separation of the conjugants takes
place at the second division.

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

516 WARO NAKAHARA

Rickert (’92, ’93), Haecker (95), and vom Rath (’95) believe
that the spireme first splits longitudinally and later segments
into the haploid number of the chromosomes. Each segment
then undergoes transverse segmentation. At the first matura-
tion division the separation takes place along the longitudinal
splitting which first appeared and the transverse division is
effected at the second division.

Paulmier (’98, 99) and Foot and Strobell (05) consider that
the haploid chromosomes are produced by telosynapsis of the
diploids and later they split longitudinally. The first division
takes place along the line of the conjugation and the second one
along the line of the splitting.

In the views of Montgomery (’04), Farmer and Moore (’05),
Mottier (05, ’07, ’09), Schaffner (07), Gates (’08, ’09, 710),
Yamanouchi (’09), Farmer and Schove (714), and Nothnagel
(16), the spireme segments into a haploid number of loops,
each loop consisting of two chromosomes united end to end.
These later bend to a side-by-side position and separate at the
first maturation division. The duality of the spireme thread
before it segments into loops is regarded as a precocious split
for the second division.

According to Winiwarter (’00), Gregoire and Wygart (’03),
Gregoire (’04, 710), Berghs (04), Schreiners (’06 a, ’06 b), Rosen-
berg (’04, 08), Overton, (’05, ’09), Allen (05), Miyake (’05),
Tischler (05), Strasburger (05, ’08, 709), Janssens (05, ’09),
Yamanouchi (08), Montgomery (711), Stevens (712), Wilson
(12), Kornhauser (714, 715), Robertson (716), and Wenrich (’16,
17), diploid chromosomes conjugate parasynaptically early at
the leptozygotene stage. The spireme segments into the haploid
number of pieces, each of these opens out along the line of the
original conjugation, and the conjugants finally separate at the
first maturation division.

The citations given cover only a small part of the entire litera-
ture relating to the topic, but they nevertheless represent the
several modes of synapsis that have been suggested, and in-
clude only those that are based upon comparatively accurate
observations.

CHROMOSOMES OF PERLA a WA

A glance at the above review may suffice to reveal that there
are two fundamentally different views regarding the modes of
synapsis—those of the parasynaptist and of the telosynaptist.

The opinions of the telosynaptists have been strenuously
opposed by von Winiwarter, Gregoire, Janssens, and others,
because earlier authors of this school have more or less entirely
overlooked a certain stage in the early prophase, which para-
synaptists claim as supporting their views. The fact that this
critical stage is observable only with difficulty may well add‘to
the dignity of the teaching of the parasynaptists, although the
very same fact may also let one doubt as to the reality of the
conception.

There is no doubt that the majority of cytologists to-day
feel quite justified in accepting the universal occurrence of para-
synapsis, probably partly due to the development of the chromo-
some theory of inheritance. It is significant, therefore, that in
spite of the overwhelming number of the parasynaptists, there
are a few who still insist upon the truth of telosynapsis.

Arnold (09), for instance, concisely discussing Planarian
spermatogenesis, concludes that “‘the spireme is gradually elabo-
rated out of a reticulum and is in the earliest stage in which it can
be recognized as spireme, composed of several separate segments,”
which are in haploid number, and never do the segments in lepto-
tene nuclei pair up longitudinally.

One of the best botanical works supporting telosynapsis is
that of Nothnagel (16). He asserted that in Allium the double
thread in the premeiotic nucleus is due to the splitting of the
single thread, by means of essentially the same process as in
somatic mitosis, and each segment appearing in the haploid
number represents two diploid chromosomes united end to end.

Some seem to believe that both para- and telosynapsis may
take place for different chromosomes in the same cell. Gates
(11) tries to show that the modes of synapsis may differ according
to the sizes of the chromosomes. Wilson (’12) seems to admit
parasynapsis for autosomes, describing at the same time an
actual telosynapsis for accessory chromosomes. Payne (14),
finally, describes two different methods of ring formation in the

518 WARO NAKAHARA

first spermatocytic division in Forficula, namely, by bending
of a rod and by opening up of a longitudinally split thread.

Critical points

In the entire history of the maturation of the germ cell the
points where the interpretations diverge are, (1) an early stage
when the double spireme develops, and (2) a later stage when
the tetrad becomes formed. ‘These two stages will be designated
for the sake of convenience as the ‘lepto-zygotene’ and ‘pachy-
streptotene’ stages, respectively.

The condition of the chromatin threads at the ‘lepto-zygotene’
stage was considered, not only by Meves (’96, 707, ’11), Kings-
bury (98, 702), Duesburg (08), Fick (07, ’08), and others, but
also by telosynaptists, as representing a longitudinal splitting
and as essentially the same as in the corresponding stage of
prophase in homotypic mitosis. Parasynaptists claim that this
is the stage when a parallel conjugation, two by two, of simple
chromatin threads takes place.

As stated before, there is no evidence of conjugation, nor of
splitting, in the case of Perla. The condition here might best
be described as the development of a double spireme out of the
chromatin reticulum of a resting nucleus, although the duality
of the spireme may be best interpreted as a precocious splitting.
In the first place, there are no definitely formed fine undivided
leptotene threads in the nucleus before the double threads begin
to appear. It is true that there are many ‘“‘thick and often
double threads terminating in two undivided diverging thin
threads like the branches of a Y, which often separate at a wide
angle and may be traced for a long distance,’ but these are
hardly adequate to base the conclusion that parasynapsis is
taking place, because the condition may just as well, or better,
be attributed to the rearrangement of the reticulum into double
spiremes. As a matter of fact, the thick threads may be seen
diverging into more than two thin threads, as observations by
Fick (07) and also by Janssens and Dumex (’03) have shown.
It must be concluded that, at least in the case of Perla, there is
no evidence of synapsis at this stage. Also, since there are no
clearly differentiated simple threads before the development of

CHROMOSOMES OF PERLA 519

the double threads, an actual longitudinal splitting of such
threads cannot account for the production of the double spireme.

According to the view of parasynaptists, the ‘pachystreptotene’
is the stage when each spireme segment becomes open along the
line of synapsis. Modern telosynaptists believe that each seg-
ment with two chromosomes conjugated end to end up to this
stage bends at the synaptic point, and finally the constituent
chromosomes come to lie side by side. The space enclosed by
a ring at this stage should result from a longitudinal opening
up of a spireme segment, if parasynaptists are correct in their
interpretation, or, if we take the telosynaptists’ view, this should
be the consequence of the bending of the spireme segement, two
arms of which coming in contact with each other to form a ring.
The process taking place in Perla is in accordance with the second
view, as it is described in the last section.

It must be noted, however, that the conclusion that chromo-
somes conjugate telosynaptically can be only indirectly supported
in the light of later history of the haploid spireme segments.
The actual process of end to end conjugation of chromosomes
has not been observed, and telosynapsis, therefore, shall still
remain as an hypothesis.

CONCLUDING REMARKS

I have come to agree with the view of the telosynaptists in
the case of Perla immarginata, reaching the conclusions that:

1. Homologous chromosomes are connected to each other
telosynaptically in the spireme.

2. That later they bend toward each other at the synaptic
point and become reunited parasynaptically before the meta-
phase, thus forming rings and tetrads.

If there be no error in my observation, and should my inter-
pretation prove to be correct, the feeling is irresistible that at
least some of the recent parasynaptists are misinterpreting the
relation of the so-called ‘primary’ and ‘secondary’ splits in the
tetrads and the nature of the split in the early spireme. It
seems also possible that some of the very convincing figures of
early prophase stages, those of Wenrich (717, p. 517, figs. 1 to 4),
for instance, may be found to be partially incorrect.

520 WARO NAKAHARA

LITERATURE CITED

ALLEN, C. E. 1905 Nuclear division in pollen mother cells of Lilium cana-
densis. Ann. Bot., vol. 29.

ArNoLp, G. 1909 The prophase in the ovigenesis and spermatogenesis of
Planaria lacteus, O. F. M. Arch. f. Zellforsch., Bd. 3.

Bercus, J. B. 1904 La formation des chromosomes hétérotypiques. II.
Depuis la sporogonie jusqu’au spiréme definitif. La Cellule, T. 21.

Bonnevigz, K. 1907 ‘Heterotypical’ mitosis in Nereis limbata. Biol. Bull.,
vol. 13.
1908 a Chromosomenstudien, 1 Chromosomen von Ascaris, Allium
und Amphiuma. Arch. f. Zellforsch., Bd. 1.
1908 b Chromosomenstudien, II. Heterotypische Mitosen als Reif-
ungscharakter. Arch. f. Zellforsch., Bd. 2.
1911 Chromosomenstudien, III. Chromatinreifung in Allium cepa.
Arch. f. Zellforsch., Bd. 6.

DuerssBerG, J. 1908 Les divisions des spermatocytes chez le rat. Arch. f.
Zellforsch., Bd. 1.

Farmer, J. B., AND Moors, J. 8S. 1905 On the meiotic phase (reducing divi-
sions) in animal and plants. Quart. Jour. Mic. Scei., vol. 48

_ Farmer, J B., aNnp Scuove, D. 1914 On the structure and development of the
somatic and heterotypie chromosomes of Tradescantia virginica.
Quart. Jour. Mic. Sci., vol. 48.

Fick, R. 1907 Vererbungsfragen, Reduktions- und Chromosomen-hypothesen.
Bastard Regeln, Merkel u. Bonnet’s Ergebnisse, Bd. 16.
1908 Zur Konjugation der Chromosomen. Arch. f. Zellforsch., Bd. 1.

Foot, K., AND Stropeuyi, E. C. 1905 Prophases and metaphases of the first
maturation spindle of Allolobophora fetida. Am. Jour. Anat., vol. 4.

Gates, R. R. 1908 <A study of reduction in Oenothera rubrinervis. Bot. Gaz.,
vol. 46.
1911 The mode of chromosome-reduction. Bot. Gaz., vol. 51.

Goxpscumipt, R. 1908 a Uber das Verhalten des Chromatins bei der Hireifung
und Befruchtung des Dicrocoelium lanceatum. Arch. f. Zellforsch.,
Badal:
1908 b_ Ist eine parallele Chromosomenkonjugation bewiesen? Arch.
f. Zellforsch., Bd. 1.

GREGOIRE, V., et Wyaart, A. 1903 La reconstruction du noyau et la formation
des chromosomes dans les cinéses somatique. La Cellule, T. 21.

GrecoirE, V. 1904 Le réduction numérique des chromosomes et les cinéses
de maturation. La Cellule, T. 21.
1907 La formation des gemini hétérotypiques dans les végétaux.
La Cellule, T. 24.
1910 Les cinéses de maturation dans les deux régnes. II. La Cellule,
B-326;

Harcker, V. 1895 Die Vorstadien der Eireifung. Arch. f. mikr. Anat., Bd. 45.

Hance, R. J. 1917 The fixation of mammalian chromosomes. Anat. Rec.,
vole 122

Henkina, H. 1890-92 Untersuchungen iiber die ersten Entwicklungsvorginge
in den Eiern der Insekten. Zeit. f. wiss. Zool., Bd. 49, 51, 54.

CHROMOSOMES OF PERLA HZ

JANSSENS, F. A. 1905 Evolution des Auxocytes males du Batracoseps at-
tenuatus. La Cellule, T. 19.
1909 La théorie de la chiasmatypie. Nouvelle interprétation des
cinéses de maturation. La Cellule, T. 25.

JANSSENS, F. A., et Dumez 1903 L’élément nucléinien pendant les cinéses de
maturation des spermatocytes chez Batracoseps attenuatus et Pletho-
don cinereus. La Cellule, T. 20.

Kinassury, B. F. 1899 The reducing divisions in the spermatogenesis of
Desmognathus. Zool. Bull., vol. 2.
1902 The spermatogenesis of Desmognathus fuscus. Am. Jour.
Aa tena; Olen ile

Kornuauser, 8. I. 1914 A comparative study of the chromosomes in the
spermatogenesis of Enchenopa binotata (Say) and Enchenopa (Cam-
pylenchia Stal) curvata (Fabr.). Arch. f. Zellforsch., Bd. 12.
1915 A cytological study of semiparasitic copepod. Hersilis apodi-
formis (Phil.), with some general considerations of copepod chromo-
somes. Arch. f. Zellforsch., Bd. 13.

Korscuett, E. 1895 Uber Kerntheilung, Eireifung und Befruchtung bei
Ophryotrocha puerilis. Zeit. f. wiss. Zool., Bd. 60.

Metz, Cu. W. 1916 Chromosome studies on the Diptera. II. The paired
association of chromosomes in the Diptera and its significance. Jour.,
Exp. Zool., vol. 21.

Meves, F. 1907 Die Spermatocytenteilungen bei der Honigbiene, nebst Bemer-
kungen tiber Chromatinreduktion. Arch. f. mikr. Anat., Bd. 70.
1908 Es gibt keine parallele Konjugation der Chromosomen! » Arch.
f. Zellforsch., Bd. 1.

Mryaxe, K. 1905 Uber Reduktionsteilung in der Pollenmutterzellen einiger
Monokotylen. Jahrb. wiss. Bot., Bd. 42.

Montaomery, T. H. 1904 Some observations and considerations on the matu-
ration phenomena of the germ cells. Biol. Bull., vol. 6.
1911 The spermatogenesis of an Hemipteron, Euchistus. Jour.
Morph., vol. 22.

Mortinr, D. M. 1907 The development of the heterotype chromosomes in
pollen mother cells. Ann. Bot., vol. 21.
1909 On the prophases of the heterotypic mitosis in the embryo-sac
mother cell of Lilium. Ann. Bot., vol. 23.

NotTHNAGEL, M. 1916 Reduction divisions in the pollen mother cells of Allium
tricoccum. Bot. Gaz., vol. 61.

Overton, J. B., 1905 Uber Reduktionsteilung in der Pollenmutterzellen
einiger Dikotylen. Jahrb. wiss. Bot., Bd. 42.
1909 On the organization of the nuclei in the pollen mother cells of
certain plants, with special reference to the permanence of the chro-
mosomes. Ann. Bot., vol. 23.

Pauutmigr, F. C. 1898 Chromatin-reduction in the Hemiptera. Anat. Anz.,
Bd. 14.
1899 The spermatogenesis of Anasa tristis. Jour. Morph., vol. 15.

Payne, F. 1914 Chromosomal variations and the formation of the first spermat-
ocyte chromosomes in the European earwig, Forficula sp. Jour.
Morph., vol. 25.

522 WARO NAKAHARA

Vom Ratu, O. 1895 Neue Beitrige zur Frage der Chromatinreduktion in
der Samen- und Eireife. Arch. f. mikr. Anat., Bd. 46.

Rosertson, W. R. B. 1916 Chromosome studies. I. Taxonomic relationships
shown in the chromosomes of Tettigidae and Acrididae: V-shaped
chromosomes and their significance in Acrididae, Locustidae, and
Gryllidae: chromosomes and variation. Jour. Morph., vol. 27.

RosenBerG, O. 1904 Uber die Tetradenbildung einer Drosera. Ber. d. deutch.
bot. Ges., Bd. 22.

Rickert, J. 1892 Zur Entwickelungsgeschichte des Ovarialeies bei Selachiern.
Anat. Anz., Bd. 7.

1893 Uber die Verdoppelung der Chromosomen im Keimblischen
des Selachiereies. Anat. Anz., Bd. 8.

Scuuerp, W. 1906 Die Entwicklung der Chromosomen im Ei von Planaria
gonocephala. Zool. Jahrb., Bd. 23.

1907 Die Samenreifung bei den Planarien. Zool. Jahrb. Bd., 24.

ScHREINER, A. UND K. E. 1906a Neue Studien titber Chromatireifung der
Geschlechtszellen. I. Die Reifung der minnlichen Geschlechtszellen
von Tomopteris onisciformis. Arch. Biol., Bd. 22.

1906 b Neue Studien itiber Chromatireifung der Geschlechtszellen.
II. Die Reifung der minnlichen Geschlechtszellen von Salamandra
maculosa, Spinax niger und Myxine glutinosa. Arch. Biol., Bd. 22.

Smitu, L. W. 19138 The biology of Perla immarginata. Ann. Ent. Soc. Amer.,
vol. 6.

Stevens, N. M. 1912 Supernumerary chromosomes, and synapsis in Ceutho-
philus. Biol. Bull., vol. 22.

STRASBURGER, EK. 1905 Typische und allotypische Kernteilung. Jahrb. f.
wiss. Bot., Bd. 41.

1908 Chromosomenzahlen, Plasmastrukturen, Vererbungstriger und
Reduktionsteilung. Jahrb. f. wiss. Bot., Bd. 45.

1909 Zeitpunkt der Bestimmung des Geschlechts, Apogamie, Par-
thenogenese und Reduktionsteilung. Jena.

Surron, W. 8. 1902 On the morphology of the chromosome-group in Brachy-
stola magna. Biol. Bull., vol. 4.

Vespowsky, F. 1907 Neue Untersuchungen iiber die Reifung und Befruchtung.
K6nigl. Bbhmische Ges. d. Wiss., Prag.

Wenricu, D. H. 1916 The spermatogenesis of Phrynotettix magnus, with
special reference to synapsis and the individuality of the chromosomes.
Bull. Mus. Comp. Zool., vol. 60.

1917 Synapsis and chromosome organization in Chorthippus (Steno-
bothrus) curtipennis and Trimerotropis suffusa (Orthoptera). Jour.
Morph., vol. 29.

Witson, E. B. 1912 Studies on chromosomes. VIII. Observations on the
maturation phenomena in certain Hemiptera and other forms, with
considerations on synapsis and reduction. Jour. Exp. Zool., vol. 13.

von Wintwarter, H. 1900 Recherches sur l’ovogenése et l’organogenése des
mammiféres (lapin et homme). Arch. d. Biol., T. 17.

YamMaAnoucui, 8. 1908 Sporogenesis in Nephrodium. Bot. Gaz., vol. 45.

1909 Mitosis in Fucus. Bot. Gaz., vol. 47.

PLATES

All the figures were drawn with the camera lucida on the level of the table,
and with Zeiss apochromatic 1.5-mm. oil-immersion objective and no. 8 com-
pensating ocular.

523

PLATE 1
EXPLANATION OF FIGURES
Spermatogonial division

A metaphase plate, with ten chromosomes.
Two cells in metaphase (right and middle), and one in anaphase (left).

Early telophase.
4 Late telophase; the nucleus is beginning to resume the appearance of

‘resting.’

—_

WwW bh

First spermatocytie division

5 A ‘resting’ cell, with its chromatin substance in the form of a reticulum.

6 to 13 Cells in early prophase (lepto-zygotene stage). Rearrangement of
the reticular chromatin substance into a double spireme is taking place. Figures
12 and 13 show the polarized condition of the forming double spireme.

14 and 15 ‘Contraction stage.’ The duality of the spireme is already com-
pletely established.

16 to 18 Degenerating cells.

524

CHROMOSOMES OF PERLA PLATE 1
WARO NAKAHARA

Su | |
SS My See
Mg

| ; 7 ~

an
3 Ki}
¢. +4,
ee re J A;
3 ¢ S A 4
Tt. # AY
; y4*

hia ss Os A
7 | =e eer
9
8
K! ~ Of i ae
Sy nit wey | 9789)
= : t v4 p
Soe Re aS? aed
2
10 i

PLATE 2
EXPLANATION OF FIGURES
First spermatocytie division—Continued

19 to 23 The formation of a double spireme completed (zygotene stage).
There is a peculiar chromatin body in the nucleus at this stage, as is represented
by the nucleolus-like structure in figures 19, 20, and 23, or by a modified portion
of the spireme in figure 22. The exact nature of these structures has not been
worked out, although their genetic relation with accessory chromosomes seems
very probable.

24 to 26 Breaking up of the spireme into segments (pachytene stage). The
segments are in reduced number.

27, 31, and 32. Halves of each spireme segment bending toward each other,
the original split of the spireme is clearly recognizable (pachystreptotene stage).

28, 29, and 30 Transformation of spireme segments into tetrads; the narrower
lateral split in the tetrads undoubtedly corresponding to the original split of
the spireme.

33 A cell entering the metaphase.

34 Metaphase spindles. The split of the early spireme is still visible on the
chromosomes.

526

CHROMOSOMES OF PERLA PLATE 2
WARO NAKAHARA

PLATE 3
EXPLANATION OF FIGURES

First spermatocytie division—Continued

35 to 37. Metaphase plates.

38 to 40 Metaphase spindles. These may show that the split line of the
spireme has nothing to do with that of the division of chromosomes. Figures
39 and 40 illustrate the separation of the X- and Y-chromosomes.

41 Early anaphase of the division.

42 Later anaphase.

Second spermatocytic division

43 The stage immediately following the anaphase of the first division. Bi-
partite chromosomes are already visible as such.

44 A cell just before the metaphase (left) and another in metaphase.

45 A metaphase spindle, showing equational division of chromosomes.

46 Metaphase plates with five chromosomes, including the Y.

47 Ditto, including the X.

48 Anaphase of the division.

49 Later anaphase.

50 to 51 Spermatids.

CHROMOSOMES OF PERLA
WARO NAKAHARA

38

49

39

PLATE 3

37

Resumen por el autor, Sidney Isaac Kornhauser,
Universidad del Noroeste, Chicago, e Instituto
de Brooklyn, Long Island.

Los caracteres sexuales del membracido Thelia bimaculata (Fabr.)

I. Cambios externos inducidos por Aphelopus theliae (Gahan).

El drinido poliembrionario Aphelopus theliae, deposita un
huevo en la ninfa del membracido Thelia bimaculata. Un solo
huevo produce de cincuenta a setenta y cinco larvas que viven
en el abdomen del animal parasitado. Los individuos parasitados
de Thelia se transforman a menudo en adultos, pero se modifican
por la presencia de las larvas de Aphelopus. Los machos se
parecen a las hembras por el color, tamafio, forma y costumbres.
La transformacién se extiende hasta los pequenos detalles del
exoesqueleto. Los 6rganos genitales externos en ambos sexos
se reducen considerablemente en tamafio y pierden sus caracteres
especificos. Los 6rganos genitales de la ninfa durante la Ultima
muda son inhibidos a menudo en su desarrollo hasta que llegan
a parecerse a los de la muda anterior. Los pardsitos nhibeni
también el crecimiento de als gonadas en ambos sexos. Los tes-
ticulos sufren una degeneraci6n parcial o completa, pero en nin-
gin caso los espermatocitos producen células parecidas a los
ovocitos. Los machos parasitados almacenan grasa en sus ab-
démens hipertrofiados y son menos activos quelosnormales. Las
modificaciones mas mareadas tienen lugar en los caracteres es-
pecificos de los machos; después siguen las de los detalles de los
érganos genitales externos de ambos sexos; los caracteres menos
modificados son los de las células germinales no maduras. Los
primeros procesos ontogénicos no se modifican bajo la accion de
los pardsitos en vias de crecimiento pero los caracteres que apa-
recen mis tarde en la ontogenia se modifican profundamente.
El cambio de metabolismo inducido por los pardsitos es mas anor-
mal en el macho que en la hembra y produce un efecto marcado
sobre aquel ser altamente especializado inhibiendo la accién de
los genes filogenéticos recientes que le comunican sus caracteres
sexuales especificos.

Translation by José F. Nonidez
Carnegie Institution of Washington

AUTEOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JULY 21

THE SEXUAL CHARACTERISTICS OF THE MEM-
BRACID, THELIA BIMACULATA (FABR.)

I. EXTERNAL CHANGES INDUCED BY APHELOPUS THELIAE (GAHAN)

SIDNEY I. KORNHAUSER

The Zoological Laboratory of Northwestern University and the Biological Laboratory
of the Brooklyn Institute of Arts and Sciences at Cold Spring Harbor,
Long Island

FIFTY-FOUR TEXT FIGURES

CONTENTS

ee O CH GELOM, 228s. ot Senin wns tie ain cueiae <2 ames ashes Saas ate ee aghraie «eee ae 531
2. Synopsis of previous papers on the sex of Al a BA oe Como
3. Material and methods. . : oT iteoly Joba ay Pisaa
4. Brief account of the life history, meet habits a holier: : 545
5. Aphelopus theliae (Gahan), a Peat cpaseinde Omer fe shite

history and habits.. : . 547
6. Changes induced by Wanelonast in Mine jeolont ad nunc 6 che meee

and appendages of Thelia.. mrt BS. SELSOG Ye eer SU An one
7. An account of two significant asda} Sieg Sedge eS CORE een
Se, LD EST ELTSIEE iN Bae DE RE Gee Cotes pitas sere A ee einer tte oe ey. ee Oe 2 ae 599
Ub, QUI DIPRE TS peal eee a ARR erin teem ead ox DiI ell ng fs edly ino era 629

1. INTRODUCTION

In the insects and especially in the Homoptera, the supposed
physical mechanism for sex determination is well known. Most
of the forms show a single x-chromosome in the male diploid
group and two x-chromosomes in the female diploid group.
A wide difference of opinion exists as to the part these chromo-
somes play in sex differentiation. One investigator will speak
of them as sex determiners, another will allow that they are
merely associated with sex, a morphological expression of the
underlying sex and comparable only to other structural sexual
differences. The present paper is an attempt to throw some
light on sex determination in insects.

531

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

532 SIDNEY I. KORNHAUSER

Let us first define the characteristics which distinguish male-
ness from femaleness. The male germ gland produces sperm
cells, the female produces eggs, and that certainly is the primary
difference. But in addition there are sexual differences in the
soma and these are striking in the group of insects under con-
sideration. These somatic differences are generally designated
as secondary sexual characteristics. ‘They may be divided into
two categories: those immediately concerned with the transfer
of the spermatozoa or the laying of the egg may be called genital
secondary characteristics, and those differences in color, orna-
mentation, and general form of body may be called extragenital
secondary sexual characteristics or, if one choose, tertiary sexual
differences. One of the problems, stimulated by remarkable
results in the vertebrates, has been to determine in the insects,
whether or not the development of the somatic sexual differ-
ences is dependent upon or independent of the primary sex gland.

Attempts to alter these secondary and tertiary characteristics
of insects by experimental castration or the transplantation of
gonads have not proved successful: the soma seems to be fixed,
either male or female, and not dependent for its development
upon the gonad nor upon some hormone developed by the gonad.
We have, therefore, no means, except through hybridization,
under our direct control of determining whether or not the
secondary sexual characteristics of the opposite sex may be
present in a latent form in either male or female insect. Cer-
tainly, latent characteristics cannot be brought out, as they may
be in birds and mammals, by removal or transplantation of the
gonads. Accordingly, it is of interest to observe the results of
an experiment in nature in which the seemingly fixed and strik-
ing sexual characteristics of the male insect have been lost and
those of the female have appeared in their place; this change
being brought about by the action of internal hymenopteron
parasites.

In the summer of 1911, while collecting Thelia at Cold Spring
Harbor for a cytological study, the writer was struck with the
fact that more gray females were to be seen on the locust trees
than handsome brown and yellow males, and this in midseason,

SEXUAL CHARACTERISTICS OF THELIA 533

a time when the males should be at their maximum. Exami-
nation of the gray individuals showed that some were altered
males, each of which, when dissected, revealed the presence of
thirty-five to fifty larvae in its abdomen. A great deal more
material was observed and collected in the summers of 1914-17,
and it is upon this material that the present paper is based.

We are first concerned with the external changes which the
parasites bring about in their host, and to this I shall attempt
to confine the present study; secondly, with the internal changes
produced, to which I shall at times refer and upon which it is
hoped more data can soon be procured; and, thirdly, to the
nature of the parasites themselves, about which a number of
facts are now known and are briefly given in part 5.

2. SYNOPSIS OF PREVIOUS PAPERS ON THE SEX OF ARTHROPODS

- Experimental observation on the relation between germ gland
and soma in insects goes back to Oudemans (’98), who castrated
both male and female caterpillars of gypsy-moths. Although
the objection may be raised that the majority of the larvae were
castrated only on one side, yet neither these nor the more
important, completely castrated individuals showed in their
adult form any deviations from the normal structure or instincts.
Morphological and psychological sexual characteristics were not
in the least altered by the absence of the gonads. Oudemans
reviewed in his paper many of the striking cases of gynandro-
morphism then known and used these as additional evidence to
prove that in insects the soma is independent of the germ plasm
in its development. Cramptons’ paper (’99), while not dealing
with the castration of Lepidoptera, has an interesting bearing
on the question of sex in the insects. He grafted the pupae of a
number of our common moths in pairs arranged in tandem and
side by side. The members of a pair were often opposite sexes
of the same species, others were opposite sexes of different spe-
cles. In some cases the ovaries of the female component grew
into the male portion of the graft, and yet in none of the cases
in which the imaginal stage was successfully reached were the
colors of the components in the least altered from the normal.

534 SIDNEY I. KORNHAUSER

Likewise, small ingrafted portions of integument retained their
original: color and the transfusion of haemolymph was also
without effect on the color of the adult. In the bisexual grafts,
a free flow of haemolymph from one component to the other
was possible, and the conclusion to be drawn is that the presence
of the soma, haemolymph, and gonads of one sex will not inhibit
the development of the characteristics of the other component
belonging to the opposite sex.

Kellogg (04), by the use of a heated needle, castrated the
larvae of silk worms. The animals which survived the opera-
tion and became adults did not show any alterations from the
normal in their sexual characteristics. Meisenheimer (09)
published the results of extended researches upon the castration
and transplantation of gonads in moths. He used a platinum
needle heated by electricity to burn out the small gonads, and
this he could do successfully on minute caterpillars after their
first molt. Castration was often followed by the implantation
of gonads of the opposite sex. Testes developed normally in
the female soma, likewise ovaries developed in the abdomen of
males, but the eggs were always smaller and fewer in number
than normal. Castration alone or castration followed by in-
grafting gonads of the opposite sex had no effect on the second-
ary sexual characteristics; structure and breeding instincts re-
mained unaltered. <A series of experiments in which the anlage
of a wing was destroyed in each larva is significant. The anlage
regenerated in individuals of three types, those with normal
gonads, those without any gonads, and those provided with
gonads of the opposite sex, and yet in all cases where an adult
wing was formed, it bore the original coloration and pattern of
the sex operated upon. In his general consideration of the soma
and germ plasm of insects, Meisenheimer reviewed important
papers on arthropod gynandromorphs and assembled in his
publication many of the best illustrations of external and in-
ternal conditions found in these anomalous individuals.

Regen (’09 a, 09 b) castrated nymphal crickets before their final
molt or before their penultimate molt. He allowed the individ-
uals which bore identification marks to mature in their natural

SEXUAL CHARACTERISTICS OF THELIA 535

habitats. In his second paper he is very positive that the
absence of the gonads has no effect upon the adult. Castrated
males chirped as loud as ordinary males; their mating instincts
were normal; they produced spermatophores, although no
sperm was present to fill them, and the stridulating apparatus
remained unchanged. Castrated females also had their normal
structure and habits, and even bored in the ground with their
ovipositors, although no eggs were present in their bodies.
Regen’s work is the only one on paurometabolic insects.
Turning again to the Lepidoptera, Kopeé has an interesting
series of papers (’11, 713 a, 713 b), mainly on Lymantria dispar L.
and Gastropacha quercifolia L. He developed a remarkable
technique in the removal of the gonads, using a sickle-shaped
hook and for the smallest larvae a hook of silver wire. For the
transplantation of the gonads, he used sterile pipettes. Castra-
tion and implantation of gonads of the opposite sex were per-
formed on larvae of first, second, or third larval stage. Often
he would repeat the ingrafting of gonads so that the abdomen
and thorax of the operated individual would contain many in-
stead of two gonads. Testes grafted into castrated females
often grew to be larger than normal testes and in their histo-
logical structure were normal. Ovaries which developed in the
bodies of castrated males were always much smaller than normal
ovaries. There were seldom more than one-fourth or one-fifth
the normal number of ova developed and these were small. Sec-
tions showed that their yolk granules were fewer and smaller
than in normal eggs. In Gastropacha quercifolia, the ova
developed in the male soma were yellow instead of green. The
small size of the implanted ovaries, Kopeé maintains, is due
entirely to lack of space for development in the male abdomen.
The haemolymph or extract of triturated gonads, when injected
into castrated individuals of the opposite sex, produced no
effect upon the adult structures. In all his experiments (castra-
tion, castration followed by implantation, and transfusion)
the results are in perfect agreement with those of his predeces-
sors and strengthen the idea that in insects the development of
the secondary sexual characteristics is in no way dependent on

536 SIDNEY I. KORNHAUSER

the gonads nor on hormones from the gonads. In more recent
experiments Kopeé (’13 a) removed the imaginal disc of the left
antenna from larvae which were castrated and into which gonads
of the opposite sex were grafted. The regeneration of the
antenna supported his former conclusions as to the independence
of the soma, in that most of the antennae were normal in form
and also showed their characteristic coloration—light in, the
male and dark in the female. A few females developed light
antennae, and Kammerer (713) attacked the conclusions of
Kopeé, using these individuals for hisargument. Kopeé (713 b),
by a series of check experiments, successfully answered Kam-
merer’s objections by showing that in control females, not cas-
trated, the regenerated antenna at times were light instead of
dark. This condition is, therefore, not due to the absence of the
normal gonad, but doubtlessly is caused by a trophic effect
brought about by the operation.

Another interesting and instructive line of work throwing
light on the physiology of the sexes in the insects is that of Steche
(12) and of Geyer (’13), who experimented with the haemo-
lymph of various insects, mainly Lepidoptera. Steche noted
that the haemolymph of male larvae of Lymantria dispar was
yellow, that of the females green. The yellow pigment was
shown to be xanthophyll, the green pigment a metachlorophyll
formed from the leaves eaten by the caterpillars. Besides this
color difference, there are protein differences between the sexes,
shown by bringing together in a watch-glass the haemolymph of
male and female larvae. At contact there was thrown down a
‘veil-like’ precipitate. The stiffening of larvae into which for-
eign haemolymph is injected may be explained by the formation
of this precipitate. The sexual differences in color and protein
content of the blood were shown to be independent of the gonads,
for they remained unaltered in castrated individuals and cast-
rated individuals with implanted gonads of the opposite sex.
Steche attributes these sexual differences to the somatic cells
which produce the haemolymph; thus, cells of the female diges-
tive tract allow the chlorophyll to pass through quite unchanged,
whereas only the xanthophyll passes through the cells of the

SEXUAL CHARACTERISTICS OF THELIA 537

male digestive tract. Thus the somatic tissues of insects are
clearly sexually differentiated, and this differentiation is inde-
pendent of the gonads. It is, therefore, superfluous, according
to the author, to speak of primary and secondary sexual differ-
ences in the insects, for all differences are primary—those of the
soma as well as those of the germ plasm. Geyer (713) extended
the observations of Steche to other Lepidoptera and also to
other orders of insects, including rapacious forms in which no
color differences existed in the haemolymph. He also carried
out extensive castrations, transplantation of gonads, and trans-
fusion experiments. His results are in perfect agreement with
those of Steche. Even where no color difference exists, a pre-
cipitate is formed in bringing together the haemolymph of oppo-
site sexes. This precipitate is often quite as dense as that
formed by mixing the haemolymph of different species or genera.
He also showed that, where color differences existed in the blood,
it was not due to an enzyme in the male destructive to the color
found in the female haemolymph, for in no case could the meta-
chlorophyll be bleached by the addition of male haemolymph.
These observations of Steche and Geyer would seem to indicate
that a male soma would be unable to furnish the coloring matter
or the complete protein requirements for ova transplanted into
such a soma. An interesting question might also be raised in
regard to the characteristics of the haemolymph in perfect
lateral gynandromorphs of Lepidoptera, such as described by
Toyoma (’05), and in which doubtless half of the cells of the
digestive tube are male and the other half female.

Turning now to Nature’s own experiments on sex, gynandro-
morphs, we have a definite line of evidence in the insects and
Crustacea supporting the independence of somatic development
from gonad influence. A complete analysis of the individuals,
including the description of the gonads as well as external char-
acteristics, is to be sought for in a study of the biology of sex.
Such a description is given by Wenke (’06), whose article ade-
quately describes and illustrates the conditions found in his
Argynnis gynandromorphs. A perfect lateral gynandromorph
contained a single well-developed ovary which did not in the

538 SIDNEY I. KORNHAUSER

least interfere with any of the secondary sexual characteristics
of the male half of the animal. As will be pointed out later,
such examples as these have an important bearing upon the
reasons given for changes seen in parasitized crustaceans and
upon theories of sex based on general metabolic differences.
‘Many cases similar to that of Wenke, and others presenting
different internal conditions, namely, the presence of testes or
the presence of both testes and ovaries, have been described and
many of the most important cases up to 1909 reviewed by Meisen-
heimer (’09). Recently, Duncan (715) has reported on several
interesting gynandromorphs in Drosophila. Male and female
soma together may be associated with the presence of either
testes alone or ovaries alone. In the Lepidoptera, Cockayne
(15) has presented every possible association of somatic sexual
mixture with gonads of one or both sexes. In the lower Crus-
tacea, Bremer ('14) describes two individuals of Diaptomus:
the first had male somatic characteristics associated with the
presence of an ovary; the second a female abdomen which con-
tained a testis. Such anomalous forms of arthropods have
interested not only entomologists and students of sex, but also
geneticists and cytologists. The causes of gynandromorphism
are generally looked upon as having a chromosomal basis.
Boveri (715) believed that, in the lateral gynandromorphic
EKugster bees, the spermatozoon united with one of the two daugh-
ter nuclei formed by the parthenogenetic division of the female
pronucleus, and that this zygote gave rise to the female half of the
individual, whereas the unfertilized daughter nucleus produced
the male half. His contentions are upheld by showing that the
external characteristics of the female half are hybrid; those of
the male half, maternal. Morgan (16), basing his contentions
on Toyama’s (’05) silkworm gynandromorphs produced by
crosses of different races, believes the male half in these cases to
be formed by a supernumerary spermatozoon developing partheno-
genetically in the egg cytoplasm, as the male half is paternal in
its external characteristics. That parthenogenesis in moths can
give rise to males has been shown by Goldschmidt (’17 a). Prob-
ably the 2x condition is obtained in the male half by a doubling

SEXUAL CHARACTERISTICS OF THELIA 539

of the chromosomes introduced by the spermatozoon. Sex inter-
grades, which are generally mosaics of the soma of both sexes, often
containing abnormal gonads of hermaphroditic character, have
been studied by Goldschmidt (’16, °17 b) in the moth Lyman-
tria, and by Banta (716) in the crustacean Simocephalus. Al-
though these forms, as in the case of gynandromorphs, support
the idea that the somatic cells are physiologically independent
of the gonads in development, still the explanation of these
mosaics on a cytological basis must await further investigation
of their chromosomal make-up. Banta believes the environment
plays an important réle, whereas Goldschmidt (17 b), although
believing in a chromosomal explanation, is skeptical of obtain-
ing a visible demonstration of size differences of chromatic
elements in the components of the mosaics of Lepidoptera.
Although the sexual characteristics seem to be fixed, there are
nevertheless three lines of evidence which indicate that in every
normal individual the determinants for the opposite sex are
present in a hidden condition. First, the males may be pro-
duced by parthenogenesis, as is the case in the rotifers, in many
Entomostraca, and in numerous insects, especially homopterans
and hymenopterans, also occasionally in Lepidoptera and arti-
ficially induced in frogs (Loeb, ’16, Gatenby, 17). The mechan-
ism of parthenogenetic male production has been most fully
solved in the aphids (von Baehr, ’09) and in the phylloxerans
(Morgan, 09). That the egg before maturation is equivalent to
the genetic constitution of the cells of the female which produced
it is agreed upon by all biologists. In the production of a male
from such an unfertilized egg, something must be eliminated to
allow the hidden male characters to appear. Thus Morgan and
von Baehr have shown that there is a differential maturation and
that every small (male) egg throws off a whole x-chromosome or
a group of x-chromosomes in the polar body. We can hardly
escape from the belief that the presence or the absence of a par-
ticular x-chromosome determines whether the male or the female
characteristics shall develop in the mature egg. ‘The size of the
egg, however, regulates the maturation, so it seems; since the
small eggs always extrude one x-chromosome or one group of

540 SIDNEY I. KORNHAUSER

x-chromosomes, whereas in the large egg all the chromosomes
divide equationally and a female results. This idea does not
assume that the x-chromosomes actually contains the deter-
minants for the sexual characteristics, primary or secondary, but
that they merely influence the development of these characteris-
tics, which are in all probability borne by the autosomes.

The second line of evidence is gained from breeding experi-
ments. Harrison and Doncaster (14) have shown in Ithysia
zonaria, in which the females alone are wingless, that the male
of this moth (when crossed with the female of Lycia hirtaria
winged in both sexes) transmits to his daughters a characteristic
of the zonaria females, namely, small flightless wings, much
smaller than those of their parents. Foot and Strobell (14,
715, 717 a, 717 b) have crossed several species of Euschistus in
which they have studied the inheritance of two ‘exclusively male
characters.’ They have shown that the length of the intro-
mittent organ and a black spot on the male genital segment may
be transmitted through the female as well as through the male.
What weight the term ‘exclusively male character’ carries is hard
to gather from the papers, since, according to the interpretation
of the authors, it includes characteristics at one time equivalent
to primary sexual characteristics and at another time equivalent
to sex-linked characteristics. The intromittent organ we would
call a genital secondary sexual characteristic, whereas the spot
we would call an extragenital secondary sexual characteristic or
even a tertiary sexual characteristic. Foot and Strobell’s
arguments against the chromosomal basis of heredity will be
considered in the discussion. The facts of their breeding experi-
ments show very nicely that, just as in birds and mammals, the
female may transmit the male characteristics of her species in
crosses. The determiners for these characters must, therefore,
be present in her genetical make-up, although they were not
expressed in her soma.

We now come to the third line of evidence: the effect of para-
sites on the sexual characteristics. There are three important
papers dealing with the strepsipteran parasites of Hymenoptera.
Perez (’86) describes and pictures the modifications in Andraena

SEXUAL CHARACTERISTICS OF THELIA 541

brought about by Stylops. The head showed a reduction in
size, the abdomen became more globose with the puncturing less
strongly marked, and the villosity increased. The scopa, or
pollen-carrying apparatus, of the hind tibia of the female was
reduced in parasitized females. These individuals also lost the
pollen-gathering instinct. In the males, on the contrary, the
narrow hind tibia was increased by the presence of Stylops.
The clypeus of the female lost its black color and gained the
yellow color of the male, whereas the clypeus of parasitized
males showed merely a reduction in the extent of the yellow
pigment. The sting of the female was reduced greatly in size.
This was also the case with the external genitalia of the male.
Perez contends, therefore, that not only is there a loss of sec-
ondary sexual characteristics due to the parasites, but also in
certain cases there is the assumption of characteristics of the
opposite sex. A study of the gonads showed that in the male
one testis might continue functional, whereas in the female only
a minute rudimentary ovary remained.

Wheeler (10) studied the effect of Xenos on Polistes. He
also gives a most excellent review of work done on the castra-
tion of insects. The parasitized Polistes failed to give the inter-
esting series of changes we might expect. They merely assumed
a reddish tinge to the abdomen and face. Wheeler’s work does
not, of course, invalidate that of Perez; it merely fails to extend
the known changes on bees to the wasps. Smith (14) studied
three species of Andraena infested with Stylops. A large part of
his study is devoted to the development and habits of the para-
site and the remainder to the internal and external changes
wrought in the host. He reviews in detail the work of Perez and
reproduces several of his figures. Of the parasites themselves,
two facts are of especial interest. Stylops carries on its respira-
tion with the external world through two tubercles on the head,
which extend between two abdominal segments of the host, and
therefore does not use the haemolymph of its host for respiratory
changes. The sex of the parasite is not a factor in considering
the changes brought about in Andraena. Smith does not de-
scribe modifications as extensive as those given by Perez, but in

542 SIDNEY I. KORNHAUSER

two changes both authors agree: the reduction of the scopa of
the female and the assumption of the male color for the clypeus
of the female in Andraena chrysosceles and Andraena, labialis.
Stylopized males show no reduction of the testes and may have
functional sperm; but in the female the ovary, which is normally
a hundred times the size of the testis, is greatly reduced through
lack of nourishment and produces only minute functionless
eggs. Smith seizes upon this fact as the cause for the changes
wrought in the female characteristics. He argues that, as in
birds (Goodale, ’16) the ovary inhibits the development of male
characteristics, so also in Andraena the absence of the ovary
allows the male characteristics to develop. This assumption
will be considered later.

Most closely associated with the study undertaken in the
present paper is the work of Giard (’89). He described the
effect of the internal parasitic dryinid, Aphelopus melaleucus,
and the parasitic dipteran, Atelenevra spuria, on the homopterans
Typhlocyba hippocastani and Typhlocyba douglasi. In females
of both species of Typhlocyba infested with Aphelopus, the
Ovipositor was much reduced. Atelenevra had much less effect
on this organ. In Typhlocyba hippocastani the oedagus is a
complicated forked organ, and this is greatly altered by the
parasites, the forks being reduced from eight branches to six,
four, or even three. In the males of both species there occurs
on the ventral wall of the abdomen a pair of organs of unknown
function, perhaps homologous with the sound-producing appa-
ratus of male cicadas. Ordinarily, these extend from the first
to the posterior extremity of the fourth somite. In parasi-
tized males these enigmatical organs seldom reach beyond the
middle of the first somite, being reduced to two small pockets.
Matausch (’09, ’11) described the effect of insect parasites on
Membracids. In his first paper he believed that he was dealing
with gynadromorphs, but later (11) discovered that the abnor-
malities were caused by parasites.

Changes similar in character, but even more striking than
those described in the insects, have been studied in crabs infected
with rhizocephalans, parasitic barnacles. Giard (’86, ’87 a, ’87 b,

SEXUAL CHARACTERISTICS OF THELIA 543

88), Potts (06), and Smith (10) show conclusively that in-
fected males develop the secondary sexual characteristics of the
female. The abdomen assumes the general form characteristic
of the female, even bearing biramous abdominal appendages.
The large chela of the male is replaced by the slender claw of
the female. Smith (11, ‘13) and Robson (11) have studied the
effect of the parasite on the lipochromes, fats, and glycogen
content of Carcinus and Inachus. In parasitized Carcinus males
the yellow lipochrome characteristic of the female blood ap-
peared. The fat content of the blood and liver increased, whereas
the glycogen content decreased. These changes show that
the metabolism of the parasitized animals had become female.
Smith believed that the roots of the parasites made a demand
upon the soma of the host similar to the demand for fat made by
an ovary. The response to this demand, the assumption of
female metabolism, carried with it the production of female
secondary sexual characters in the morphology of the host.
Altered metabolism brought about changes in the blood of the
male which stimulated the development of the latent female
characteristics. It is not the absence of the testes, but the
presence of the parasite acting like an ovary which brings about
the changes in the males.

Not only rhizocephalans, but also protozoa may alter sexual
characteristics. Thus Smith (’05) described changes in the crab
Inachus due to a gregarine. The abdomen and claw of the
male were altered much as described above for the barnacles.
Only those individuals in which sporozoites were liberated in
the haemolymph showed modifications. No case exactly parallel
to this is known in insects. Grassi and Sandias (’93) maintained
that the presence of Protozoa in the intestinal caecum prevents
the full development of both internal and external genitalia in
termite workers. When the Protozoa are killed or removed by
feeding saliva, the purged individuals become sexually mature
substitute kings and queens. Wheeler (’10) was inclined to
believe that the dimorphism in the males of Forficula, based on
length of the forceps, might be due to the gregarines infesting
their alimentary tracts. Brindley and Potts (’10) do not believe

544. SIDNEY I. KORNHAUSER

in this assumption since they found no correlation between the
length of the forceps and the number of gregarines in adult
insects. It might here be interjected that, since the adult struc-
tures of insects cannot be altered, it would be necessary to
know how many gregarines had been present in the alimentary
tract previous to the final molt, at which time their presence
might influence the imaginal structure.

3. MATERIAL AND METHODS

A brief description of the life history and habits of Thelia and
its parasite Aphelopus will be given in parts 4 and 5. The
principal collecting ground was situated about two and a half
miles from the laboratory, and material was procured practi-
cally every second day and brought in alive. For further obser-
vations on living specimens, the insects were placed in cages on
cut locust branches or put upon branches of locust trees growing
near by, and enclosed in bags of cheese-cloth or mosquito net-
ting. The branches in the cages must have the cut ends in
water and the leaf surface reduced to remain fresh. They must
be renewed every second or third day to keep the animals in
good condition so that they will grow and molt.

For a study of the parasitized adults and normal individuals,
first the pronotum with head and prothorax attached was re-
moved, pinned, numbered, and shielded from the light to pro-
tect the colors, which fortunately keep well in dried specimens.
The body was placed immediately in a dish of physiological salt
solution or in Ringer-Locke solution, and dissected under a
binocular microscope. The abdomens were cut open dorsally
with a fine microscissors so as not to injure the genitalia. A
careful search was then made for any remnants of gonads in
parasitized individuals. The light from a Nernst glower directed
upon the interior of the animal by a condensing flask added
greatly in discovering any minute gonads which might be pres-
ent. The gonads discovered were removed and fixed for section-
ing: testes in Bouin’s fluid, ovaries in Gilson’s fluid. The body
and parasites were separately preserved for further study, usu-
ally being put into 80 per cent alcohol. Body, pronotum, and

SEXUAL CHARACTERISTICS OF THELIA 545

gonads of an individual were all given the same number and a
careful record kept of date, fixation, characteristics of pronotum
and genitalia, and size of parasites, together with any excep-
tional condition worthy of record. This was done for every
specimen.

Nymphs of all stages were fixed whole in Petrunkevitch’s
fluid warmed to 50°C. The duration of fixation varied, with the
size of the nymphs, from one hour for the smallest up to twenty-
four hours for the largest.

In making preparations of the genitalia, the non-chitinous
portions were removed by heating in a solution of caustic soda
and then trimming the preparation with a fine scissors and scalpel
under a binocular microscope. They were then dehydrated and
mounted in balsam.

Sections were made through the bodies of adults and nymphs.
In this work the celloidin-paraffin method (Kornhauser, ’16)
was invaluable. Sections were made 10u in thickness, and it is
possible to cut the hardest chitin without tearing the ribbons
or nicking the knife. The gonads were cut 6, in thickness, and
stained in Heidenhain’s haematoxylin and Congo red.

All the figures (with the exceptions of numbers 7, 8, 18, 32 to
35, 53, 54) are untouched photographs made with Spencer
micro-teleplat objectives 8 mm., 24 mm., and 60 mm. In the
24-mm. and 60-mm. objectives, iris diaphragms were inserted.
The non-photographic figures are drawings made with the aid
of a camera lucida.

4, BRIEF ACCOUNT OF THE LIFE HISTORY AND HABITS OF THELIA

Thelia bimaculata is the largest and one of the commonest
membracids of northeastern United States. It feeds exclusively,
as far as is known, on the sap of the common locust (Robinia
pseudo-acacia L.). It is found on the trunk or larger branches
of this tree. Adults occur from July to October. Association
with other organisms is nicely seen in this docile, domesticated
homopteran which is constantly attended by ants andis imposed
upon by internal and external parasites. Hymenoptera live
within its body, mites attach themselves to the exterior, and

"546 SIDNEY I. KORNHAUSER

predacious dipterans often pounce down and carry the Thelia
away without much trouble.

In my principal collecting fields, Formica truncicola Nyl.
subsp. obscuriventris and Cremastogaster lineolata Say were
the chief ants associated with Thelia. When tapped by the
antennae of the ants, the Thelia nymph or adult exudes from
the anal tube a drop of clear fluid which is taken by the ant with
great alacrity. Toward the middle of June, the ants build
collars about the bases of the locust trees, and inside these collars
in the cracks of the bark are to be found hundreds of Thelia
nymphs of third to fifth instar, quietly feeding and undisturbed
by the numerous ants in attendance. In this moist situation,
protected from many of their enemies, the nymphs thrive.
Formica builds the protecting collar of leaves, twigs, and bits
of wood; Cremastogaster builds of sand grains cemented to-
gether. When one breaks the collar, many ants swarm out and
attack the intruder, Formica biting one’s fingers ferociously,
while others grab the Thelias and drag them into underground
passages. These pugnacious ants seem to have complete mas-
tery of the Thelia nymphs.

After completing its growth in the fifth instar, the Thelia
emerges from the collar of leaves or sand, climbs higher on the
trunk or branches of the tree, and molts into an adult. Mature
males are first to appear, generally being found early in July.
The females and parasitized adults of both sexes mature a week
or two later. In July and early August, the adults sit motion-
less on the bark and are not found in groups. One can often
catch them. between the forefinger and thumb. The males,
however, are more active than the females and hop or fly at less
provocation. Both sexes are more active on hot, sunny days.
Toward the end of August and in September, the individuals
gather into groups on the branches or the trunk of the tree, and
there is evidence of courtship, for very often one sees a single
female surrounded by several males. In September, individuals
mating can occasionally be found. The male and female face
in opposite directions and the tip of the abdomen of the male is
placed beneath the ninth abdominal segment of the female.

SEXUAL CHARACTERISTICS OF THELIA 547

Thelia lays its eggs in late September and in October. At
this time the males are already becoming lessnumerous. The
female lays her eggs in the bark of the small branches of Robinia.
I have never seen eggs being deposited at the bases of trees as
described by Funkhouser (’15), who has written a very good
account of the life history of Thelia and has described the five
instars. With her sword-like ovipositor the female makes a slit
through the bark, longitudinal to the branch and tangential to
the underlying wood part. The total length of the egg is 2.4
mm. The chorion forms a tube 0.4 mm. beyond the contained
ovum, which is, therefore, but 2 mm. in length. This chorionic
tube projects from the slit in the bark and probably aids in the
respiratory changes of the developing embryo. From three
to sIx eggs are deposited in a single slit and one Thelia will lay
between thirty and forty eggs at a time, judging from dissections
of adults previous to laying. The eggs remain in the bark over
winter. In early June they hatch, and the small shiny brown
nymphs begin to feed out on the small branches of the tree.
They occur in cracks in the bark at the bases of thorns, or at the
edges of healed wounds where the bark is thin and succulent.
First, second, and third instars occur on the branches, constantly
attended by ants. Soon, however, the ants begin to build the
collars at the bases of the trees and third to fifth instars are found
in abundance only inside these collars. As described above,
they emerge as full-grown nymphs in July, crawl higher on the
tree and molt into adults.

5. APHELOPUS THELIAE (GAHAN), A POLYEMBRYONIC PARASITIC
DRYINID; ITS LIFE HISTORY AND HABITS

The life history and habits of Aphelopus theliae were gradu-
ally worked out by the author until at the present time we have
a fairly complete story. The difficulty in getting the series of
events in the life cycle complete was due to the fact that until
the past summer (1917) the stay at Cold Spring Harbor had not
begun early enough in June to obtain adults and to see the lay-
ing of the Aphelopus egg in the nymphs of Thelia.

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

548 SIDNEY I. KORNHAUSER

Aphelopus theliae belongs to the Dryinidae, characteristically
parasitic on homopterans, but never before known to be poly-
embryonic. This species was named and described by Mr. A.
B. Gahan (’18) from material reared and sent by the author.

Fig. 1 Megagnathie larvae in abdomen of Thelia. Terga, digestive tube,
and parasites dorsal to digestive tube have been removed. X 9.

Fig. 2 Eruciform larvae as they emerge from their host. X 8.7.

Fig. 3. Male Aphelopus theliae. Taken from pupa case before complete
transformation accomplished. X 10.1.

Fig. 4 Female Aphelopus theliae. Taken from pupa case before its ova had
reached maturity. X 10.1.

Imagoes are shown in figures 3 and 4. They are jet black.
The female is 2.2 mm. in length, the male slightly shorter.
The specimen shown in figure 4 was taken from its pupa case
before complete maturity was reached. This accounts for the

SEXUAL CHARACTERISTICS OF THELIA 549

space between the abdominal terga and sterna; the fat distending
the abdomen shows through laterally at this stage, but not in
sexually mature individuals.

In early June, when the young Thelia nymphs of first to third
instar are abundant on the smaller branches, the female Aphe-
lopus hurries nervously up and down the branches of the locust
tree hunting its prey. Finding a nymph, she taps it with. her
antennae, much as does an ant, and quiets it; she then climbs
upon the dorsum of the nymph, feels its head with her large
powerful mandibles, to test, I believe, whether or not the nymph
is soon to molt. If the parasite decides to deposit an egg, she
grasps the caudal part of the nymph’s abdomen between her
mandibles, and, holding firmly with her legs to the abdomen of
the host, tries to thrust her sword-like ovipositor cephalad
through the intersegmental membranes of any two abdominal
terga. The nymph struggles as the ovipositor pierces, and from
the anal tube exudes a drop of liquid. This, the Aphelopus
grasps In her mandibles and is gone in a second. If the nymph
is a small first, second, or third instar, the whole act of oviposit-
ing may take but a few seconds; but, when a fourth or fifth
instar is attacked, a long struggle often follows and it may take
several minutes before the small energetic parasite accomplishes
her task. Having examined the nymph and decided to oviposit,
the Aphelopus gets hold of a tibia with its mandibles and legs
and hangs on, no matter how hard the Thelia kicks or even if
both go rolling over and over. The parasite then tries to force
its ovipositor through the soft membranes between the tro-
chanter and femur or between the coxa and thoracic sternum.
After ovipositing, the Aphelopus generally mounts the abdomen
of the nymph to secure a drop of excrement. After a struggle
with a large nymph, the parasite remains on the bark or on a
leaf, cleans her wings, legs, antennae, and mandibles, and espe-
cially rubs her ovipositor vigorously with her hind legs. The
whole process of oviposition was watched many times in the
laboratory under a binocular microscope by putting a female
Aphelopus into a test-tube containing a locust twig upon which
several Thelia nymphs were feeding. The parasite would hunt

550 SIDNEY I. KORNHAUSER

over the twig, and finding the nymph would oviposit just as in
nature. This procedure was continued by renewing the nymphs
until the Aphelopus was exhausted. She might oviposit again
the next day, laying in all from twenty-five to fifty eggs, gener-
ally one in each nymph. It was found best to work the para-
sites as hard as possible the day they were captured, for they
were most active then and would live only a few days in the
laboratory. One lived five days, but thatwas unusual. The
only successful way to capture the adult female Aphelopus was
to place the mouth of a small vial over the individual as soonas
she was discovered running over the bark, and thus try to corner
her so that she would run up into the vial. One must not hesi-
tate a second nor await a more favorable opportunity, for,
should an ant come along, the parasite for which one may have
been hunting several hours would hop or fly in terror. Aphe-
lopus has a keen fear of ants, especially of Formica. This was
tested in the laboratory by putting an ant and an Aphelopus into
a tube. The ant immediately took the offensive, showing that
an enemy of Thelia is not to be tolerated.

When one dissects, under a high-power binocular microscope,
a nymph which has just been stung, one will find a thin-shelled
oval egg, 145u in length and 60u in diameter, and also several
spheres covered with a chitinous shell and filled with yolk-like
material. These spheres are developed in the female Aphelopus
from single cells in a sack-like pocket ventral to and leading
into the posterior portion of the oviduct, Just below the opening
of the spermatheca. The function of the spheres, which vary
from 25u to 35u in diameter, has not yet been determined. The
egg! absorbs fluid from its host and the thin shell swells. Within,
total cleavage takes place and the sphere of cells formed soon
develops into a polygerm mass. This mass becomes oval, and
then angular and irregular in outline, as it starts to form branch-
ing chains of embryos. These chains are composed of spheres
connected and covered by a placental envelope several cells in

1 A detailed study of the cleavage of the Aphelopus egg and the formation of

the polygerm is contemplated, suitable material having been secured from lab-
oratory-stung specimens during the past summer.

SEXUAL CHARACTERISTICS OF THELIA aol

thickness. This envelope constricts between the embryos and
‘the chains are broken up into separate individuals, each of which
develops into a larva surrounded by its nutritive envelope. The
details of this growth and differentiation will be left to a subse-
quent paper.

If the Thelia nymph is stung during the first or second instar,
the abdomen of the fourth instar will be filled with forty to
sixty larvae bent in a half-circle, ventrad and measuring 0.75
mm. from the cephalic to the caudal end. Each embryo is
enclosed in a semitransparent envelope of cells which doubt-
lessly serves as an organ of nutrition and respiration. The
mouth parts are not chitinized at this stage. In the fifth instar
or the adult Thelia the larvae reach their maximum development
as internal parasites (fig. 1, p. 548). They grow tremendously
and acquire large, hard, brown, chitinous mouth parts, well-
defined stigmata, and a circumcephalic plate (Keilin and Thomp-
son, 715) which in its dorsal region is brownish and thrown into
many transverse folds. On account of the prominent mouth
parts, we may designate the larvae of this stage as megagnathic
larvae. Surrounded by fat, they are packed close to one another,
about thirty-five being in contact with the abdominal sterna of
the host, fifteen or more occupying the region lateral and dorsal
to the digestive tube, while occasionally a few are present in the
thorax. Reaching their maximum size, they distend the ab-
dominal segments of the host to such an extent that the inter-
segmental membranes are clearly visible between the abdominal
sclerites. This must certainly interfere with the respiratory
movements of the host.

The megagnathic larvae have a well-defined alimentary tract
which ends blindly caudad. The tract is filled with shining
crystals which appear white in reflected light. These crystals
are kept in constant motion by peristaltic waves which may be
viewed nicely in living larvae in Ringer’s solution. A deep
constriction of the digestive tube runs caudad, then reverses and
runs cephalad. The crystals in the tube are very insoluble,
persisting in specimens kept for several years in 80 per cent alco-
hol and resisting in section-making all the ordinary reagents.

OZ SIDNEY I. KORNHAUSER

They are probably some of the katabolic substances formed in
the parasites’ development which are stored in an insoluble con-
dition rather than being thrown into the haemolymph of the
host, as the host might be unable to rid itself of these waste
products.

The final molt of the parasitic larvae and their escape from
the host is the next act. This generally occurs during the fifth
nymphal instar of Thelia, but if oviposition of the parasite
occurs late in the ontogeny of the host, the Aphelopus larvae
reach maturity in the adult Thelia and escape often with some
difficulty from the imagos. The fact that the parasites may be
present in adults and are found there in various stages of devel-
opment makes this paper possible. In Lepidoptera, for instance,
the larvae or pupae are always destroyed by their polyembryonic
parasites, and so we cannot tell how the sexual characteristics of
the imago might have been affected. The Thelia nymph or
adult from which the larvae are soon to emerge leaves its com-
rades, climbs to a solitary twig or leaf, and fastens itself firmly
with its tarsi. Small elevations running in rows across the
abdominal sterna appear. A little later each elevation becomes
a hole from which the caudal end of a green or yellowish eruci-
form larva emerges. Before escaping the larvae devour the
entire contents of the host, leaving only an empty shell clinging
to the leaf or twig. By wriggling motions, the larvae work their
way out of the host and also out of the integument of the mega-
gnathic stage. The integument is broken in the dorsal region
behind the circumcephalic plate and the eruciform larva leaves
its exuvia at the hole through which it emerges from the Thelia.
The chitinous jaws of the exuviae plainly mark the holes made
in the sternites of the host (figs. 48 and 47). The eruciform
larvae (fig. 2) are entirely different in form from the mega-
gnathic larvae, having small jaws and a very bristly integument
marked into distinct segments by rings of spines. They are
yellow or light green, depending upon the amount and color of
the pigment which was present in the body of the host. On the
average, thirty-five larvae free themselves simultaneously, the
abdomen of the host bending until the ventral surface is par-

SEXUAL CHARACTERISTICS OF THELIA 58

allel to the ground and allowing them to drop to the earth.
These are followed by the remaining individuals which did not
have a ventral position in the host next to the abdominal
sterna. .

The eruciform larvae crawl at a good rate, hunting any small
opening in the ground. In the laboratory it was found advis-
able to make small holes in the soft earth in jars into whichthe
larvae dropped. Finding these holes, they burrow half an inch
or an inch through the soil and spin a little straw-colored cocoon,
2.36 mm. in length. The cocoon is almost oval, but a trifle con-
stricted about the middle where a thickened portion forms a
little white transverse band. Often the larvae spin their cases
against stones or the side of the jar furnished the laboratory
cultures. One could follow the color changes of the enclosed
larva or pupa, for the cocoon is quite transparent along its sur-
face of adherence.

By the end of September, the larva has transformed itself into
a white-bodied pupa with red or chocolate-colored eyes. The
abdomen is filled with fat and the intestine is distended with
crystals and dark unorganized waste material. Evidently the
winter is passed in this state, the transformation to the jet-black
adult taking place in the spring. In the laboratory cultures
kept in a warm greenhouse, development contined. Thus, by
December the chitinous portions of the Aphelopus were com-
pletely formed, but the abdominal sclerites were still distended
with the enclosed fat. In the males sexual maturity was al-
ready reached, spermatogenesis having been completed; but the
females contained only minute ova. In April the adults hatched
in the laboratory; the females then contained fifty to seventy
full-grown eggs. From these specimens reared in 1916, before
the adult had been seen in nature, the species was named and
described by Mr. Gahan.

Female Aphelopus hunting Thelia nymphs, taken in June,
1917, were dissected and revealed the fact that in some the
spermatheca was filled with sperm and in others it was empty.
Probably fecundated females lay eggs which develop into fe-
males, whereas virgin females lay eggs which develop partheno-

554 SIDNEY I. KORNHAUSER

genetically into males. The offspring of a single egg, as tested
in laboratory cultures, were all of the same sex. However, as
noted by Patterson in Paracopidosomopsis (’17), both sexes may
emerge from a single host, and this would not be impossible in
the case of Aphelopus, for at times a female might oviposit in a
nymph already containing an Aphelopus ovum. This was
demonstrated in the laboratory and both eggs were recovered by
dissection. Twice in parasitized Thelia secured in nature Aphe-
lopus larvae of two distinct stages of development have been
found in single individuals. In these cases only those coming
from the first ovum would become adults, for the host would
be killed at their emergence; but, should two ova be laid within
a short time in the body of a Thelia, it would be possible that
some offspring of both ova would reach maturity at the same
time and produce a mixed brood.

6. CHANGES INDUCED BY APHELOPUS IN THE COLOR AND FORM
OF THE INTEGUMENT AND APPENDAGES OF THELIA

A. The thorax

In Thelia, as in most membracids, the pronotum forms a
conspicuous part of the organism. It is continued forward and
upward as a horn projecting beyond the head, and it also ex-
tends caudad as the posterior process covering the dorsum of
the entire thorax and abdomen. The pronotum constitutes one
of the most conspicuous differentiating characteristics of the
extragenital type between the sexes. In the male (fig. 5) it has a
uniform, chocolate-brown color with a conspicuous orange-
yellow vitta on each side, extending caudad from the humeral
angle usually about half-way to the tip of the posterior process.
The length and form of the pronotal horn as well as the angle it
makes with the rest of the body are subject to the greatest vari-
ation, as 1s shown in figure 7. Sometimes it is long, curved,
slender, and erect, (7, d, 7, h); again, short, straight, and blunt
(7,e, 7,1). The form of the vitta is also variable, the yellow
color showing many degrees of extension toward the tip of the
posterior process. Figure 7 is a series arranged from a to l to

-———

SEXUAL CHARACTERISTICS OF THELIA ro T9T9)

show this variation. Yet, in spite of this lack of uniformity,
the vitta is always present in normal males and the rest of the
pronotum is a uniform brown. The light areas in figure 5 out-

Fig. 5 Normal male Thelia bimaculata. X 6.
Fig. 6 Normal female Thelia bimaculata. X 6.

side the vitta are merely reflections of light. The whole pro-
notum is covered with coarse punctures which are nicely seen
on the vitta of figure 5. If, now, we examine a vertical section
of a portion of the pronotum, including part of the vitta and

556 SIDNEY I. KORNHAUSER

Fig. 7 Normal males and pronota of normal males, showing variation in
extent of vitta and form of pronotal horn. Extension of yellow pigment in vitta
arranged in series from atol. X 1.5.

Fig. 8 Vertical section of cuticula of pronotum of male, passing through
part of vitta. X 144.

Fig.9 Vertical section of cuticula of pronutum of female, passing through
part of vitta. X 144.

SEXUAL CHARACTERISTICS OF THELIA Dol

the bordering brown area (fig. 8), we shall see that the brown
pigment is contained in the upper layers of the chitin in an
amorphous form. ‘This is present over the whole surface outside
the vitta in the punctures and intervening areas. This brown
pigment is melanin, resisting strong alkalis and acids. In the
hypodermal spaces below the brown chitin red granules are
embedded in a clear matrix which surrounds the tracheae extend-
ing through the pronotum. In the vitta the chitin is trans-
parent, no melanin being present, and the hypodermal spaces
are filled with an orange-yellow granular pigment which shows
through the chitin and gives that area its striking color. This
hypodermal pigment is easily destroyed by acids and alkalis
and gradually loses its color in aleohol. In figure 8 the hypo-
dermal pigment, which may vary from a light yellow to a deep
orange-yellow, is shown in the right half of the section. The
punctures in the vitta are very transparent, there being no
hypodermal pigment present and just a trace of melanin in the
most superficial layers of the chitin, which produces a yellowish
tinge.

In the female (fig. 6) the pronotum has a gray tone, harmon-
izing nicely with the bark upon which the animai rests. The
vitta is not conspicuous, being but a trifle lighter than the rest
of the body. A vertical section (fig. 9) shows us how the gray
coloration is brought about. The melanin of the entire pro-
notum is restricted to the punctures and the edges of these
punctures, whereas the hypodermal spaces are partly filled with
a yellow-green granular pigment. Some red granules may also
be found in the hypodermal matrix immediately surrounding
the puncture, and must in some way be associated with the
presence of melanin in the cuticula above. The combination
of the brown punctures and greenish-yellow areas produces a
gray tone in the pronotum.

One of the most striking effects of Aphelopus is causing the
pronotum of the male to assume the pigmentation of the female
pronotum. Many steps in the transformation have been seen
in parasitized adults and several are shown in figures 10 to 12.
Some individuals are but slightly affected (fig. 10), others have

Or
or
0/6)

SIDNEY I. KORNHAUSER

Figs. 10,11, and 12 Pronota of parasitized males, showing loss of male char-
acteristics and assumption of female pigmentation and size. Figure 10, slight
loss of uniformity of melanin and loss of hypodermal yellow. Figure 11, greater
loss of uniform melanin and further encroachment of melanin on vitta. Figure
12, complete loss of male characteristics and complete assumption of those of
female. X 6.

SEXUAL CHARACTERISTICS OF THELIA 559

perfect female coloration (fig. 12), while many show merely a
medium condition (fig. 11). In slightly modified males the
yellow hypodermal pigment of the vitta becomes fainter and less
abundant, and melanie spots appear in the cuticula of the vitta
(fig. 10). In individuals showing greater assumption of female
coloration the melanin loses its uniform distribution outside the
vitta, becomes restricted more and more to the punctures, and
encroaches still farther upon the vitta. Yellow-green hypo-
dermal pigment forms and shows through the cuticula no longer
impregnated with melanin (fig. 11). Finally, in completely
altered males, the punctures alone are brown (fig. 12) and the
hypodermal pigment is exactly like that of the normal female in
color and distribution. These changes are summarized in figures
13 and 14. Figure 13 represents ten vittae seen in reflected
light, and figure 14, the same in transmitted light. Yellow pig-
ment appears light in figure 13, and dark in figure 14, since it
absorbs the actinic rays of the transmitted light. Melanin is
dark in both figures. The clear punctures of the male vitta are
light in transmitted light (fig. 14, a). Vitta ain both figures is
that of a normal male; vitta 7, that of a normal female. Vittae
b to h, inclusive, are parasitized males and illustrate the gradual
loss of yellow accompanied by the coming in of melanin in the
punctures. In g and h complete assumption of female colora-
tion has taken place. Vitta 7 is that of a parasitized female.
It shows no tendency toward the assumption of male pigmenta-
tion. At most parasitized females show smaller punctures with
restricted melanic pigment and a thinner and weaker cuticula.
This is doubtlessly due to an interference with the normal nutri-
tion of the hypodermal cells which produce the chitin, as a sim-
ilar condition may often be noted in parasitized males. In these
cases there may be an actual scarcity of the necessary materials
for chitin and pigment production, caused by the presence of the
parasites.

It must be clearly borne in mind that no modification in the
integument can be effected by the parasites after the host has
become an adult. The degree of change is, therefore, dependent
upon the activity of the parasites previous to the final molt of

560 SIDNEY I. KORNHAUSER

the Thelia. It is during the fifth nymphal instar that prepara-
tion for the most striking feature of the metamorphosis of the
homopteran occurs (compare fig. 15 with figs. 5 and 6). The

Fig. 138 Vittae mounted in balsam, seen in reflected light. a, normal male;
j, normal female; b to h, inclusive, parasitized males showing various degrees of
loss of yellow hypodermal pigment (light in color in photographs) and increase
of melanin in punctures; g and h, complete change; 7, parasitized female.
16.6.

sexual differences of the pronota as well as many other remark-
able changes in the integument appear first at the final molt.
If, for example, the parasites in a male fifth instar are large

SEXUAL CHARACTERISTICS OF THELIA 61

megagnathic larvae while the preparation for this transformation
is going on, female coloration and many other changes to be
described are brought about. If, on the other hand, the Aphe-

Fig. 14 Same as figure 13, seen in transmitted light. Chitin without melanic
pigment (punctures in normal male), light; yellow pigment acting as absorber of
actinic rays appears dark between punctures when present; melanin also appears
dark. X 16.6.

lopus has oviposited in a nymph during its fourth or fifth instar,
the parasites will be small and have less effect upon the adult
structures.

562 SIDNEY I. KORNHAUSER

The changes in the coloration of the male pronota above de-
scribed cannot be referred to retardation of development for the
integument of the fifth instar (figs. 15 to 18) is entirely different
in structure and coloration from that of the adult. The cutic-
ula covering the entire dorsum of the nymph has long spines
projecting from its surface (figs. 17 and 18). These spines are

Fig. 15. Lateral view of fifth nymphal instar, showing mottled pigmentation
of body and appendages due to large areas devoid of melanin. 5.3.

Fig. 16 Cuticula of pronotum of fifth nymphal instar. Each small spot on
the dark background represents the base of a long jointed hair. The large light
spots which cause the mottling are plainly shown. X 14.7.

Fig. 17 Small area of cuticula seen in figure 16, under greater magnification,
showing details of long jointed hairs in dark and light areas. X 66.6.

Fig. 18. Cuticula from sixth abdominal tergum of nymph in fifth instar,.
showing jointed hairs, their bases, and distribution of meanic pigment. X 600.

~

SEXUAL CHARACTERISTICS OF THELIA 563

jointed near their bases. The melanin is very irregularly dis-
tributed, always being absent about the bases of the spines and
also in larger irregular areas, which produces a mottled appear-
ance over the whole nymph (figs. 15 and 16). Parasitized adults
possess none of these juvenile characteristics, but males affected
by Aphelopus assume female coloration through the loss of male
characteristics and the addition of those of the adult female
integument.

Female Thelia are larger than males. Thus the pronota of
111 normal females, measured with a micrometer caliper from
the tip of the horn to the end of the posterior process, averaged

TABLE 1
VOMBER OF| AVERAGE AVERAGE
INDIVID- LENGTH OF} WIDTH OF
UALS PRONOTUM | PRONOTUM
mm, mm,
Normal males.......... th ge fa 11.55 4.73
Parasitized males shards. craaiiia or F complete
change to female coloration........ 98 12.24 5.03
Parasitized males with male eu loncinon still pre-
dominating. . ay, es a Ror aa cee leucine aon ise 29 11.68 4.85
Total senesced ales shore aS ay See Sees et 127 patil 4.98
Normal females. . Soetins itt 13.39 5.41
Parasitized femelle with pecdediy Rodieae - Ovi-
J OXSISITLHON ES se ence maaeehis Ciao aNd Bec amanactor Rares 100 Se wel 5.33

13.389 mm., and the average width across the humeral angles
was 5.41 mm. (table 1). Corresponding measurements on 114
normal males gave 11.55 mm. as average length and 4.73 mm.
as the width. This shows that the pronota of normal females
are about 15 per cent longer and broader than those of normal
males. If, now, we examine parasitized males with a well-
defined change in color (medium to complete), we see that they
are both longer and wider than normal males, the increase being
about 6 per cent. That this is due to the action of the parasites
is shown by the fact that in males parasitized late in their on-
togeny (those with male coloration still predominating) we find
but a slight increase in size. Only when large parasites are

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

564 SIDNEY I. KORNHAUSER

present during the fifth nymphal instar is there a decided in-
crease in length of the pronotum. The effect cannot be pro-
duced by mechanical means, because the parasites are present
in the abdomen, whereas the pronotum is attached to the pro-
thorax alone and receives its material for growth through the
haemolymph coming into it through the prothorax. The in-
crease in the size of the pronota in parasitized males is of still
greater interest when we see that in the parasitized females
there is no increase, but rather a decrease of about 2 per cent in
both length and breadth. We are, therefore, led to the con-
clusion that the 6 per cent increase in size in the males infected
with parasites is a partial assumption of a female characteristic.

TABLE 2

NUMBER OF| AVERAGE

INDIVID- LENGTH OF

UALS FOREWING
mm.
INiorinallbmalese te eek sit eee eh ae Patan, Sepa he at eat nt ee 25 8.13

Parasitized males showing medium or complete change to

female: colonatlone sais. cote ee eee On oe oe 25 9.09
Parasitized males with male coloration still predominating... 15 8.39
INormialltfemialesemare tthe sCre ieias iret aan iene Mie as arcana ee 2a 9.78
Parasitized females with decidedly reduced ovipositors..... | 25 9.81

Corresponding changes in size and color hold also for many
other parts of parasitized male Thelia. The length of the fore
wing was next examined. Measurements were made by remov-
ing the wings, laying them on a scale divided into tenths of
millimeters and taking the reading under magnification. The
length from the posterior tip to the anterior end of the tegula
was used as the basis of comparison. The averages are given in
table 2 and a typical example shown in figure 19.

The fore wing of normal females is on the average 20 per cent
longer than that of normal males. Parasitized males with
changed color show an increase of 12 per cent in length over
normal males. Parasitized males with but slight change in
color have wings only 3 per cent longer than normal males. In
the case of the females, the parasites cause almost no change,
those measured showing an increase of three-tenths of 1 per cent

ae eee

SEXUAL CHARACTERISTICS OF THELIA 565

Fig. 19 Typical fore wings, showing increased !ength in parasitized male
(p. o&). Compare with normal male and normal female wings. From left to
right, normal female, parasitized male (p. o’), and normal male. 1/, 2, 3, 4, 4,
apical areoles. X 6.

Fig. 20 Typical hind legs, showing size difference between male and female
and increase in parasitized male. From left to right, normal male, parasitized
male (p. co’), and normal female. cz., coxa; tr., trochanter; fm., femur; ¢b.,
tibiastay, tarsuse >< ale

Fig. 21 Typical acrotergites, showing approach toward female size in para-
sitized male. From top to bottom, normal female, parasitized male (p. <=),
normal male. X 7.95.

566 SIDNEY I. KORNHAUSER

over normal females. The wing length in parasitized males
behaves just as did the size of the pronotum and the degree of
change is correlated with the change in the color of the pro-
notum. Pronotum color is the best index one can find for prac-
tically all the changes in the male.

A qualitative change in the wing is also seen in parasitized
males (fig. 19). Whereas in normal males melanic pigment
extends diagonally through the second apical areole to the base
of the fourth apical areole, in the female it forms only a spot in
the second and is restricted to the distal two-thirds of the fourth
areole. Parasitized males show a distribution of pigment like
that of normal females.

Turning to the other thoracic appendages, the legs, we find
that those of normal females are longer and stouter in every
segment than those of the male. The size relation in the third
pair is shown in figure 20. Similar to what was observed in the
pronota and wings, parasitized males exhibit an increase in the
size of the legs. This increase is especially noticeable if we
compare the length of the tibia and tarsus taken together in
each of the classes shown in figure 20.

Not only do the thoracic appendages increase in size in para-
sitized males, but the thorax itself becomes larger. This is best
measured by comparing the acrotergites of normal males, nor-
male females, and parasitized individuals. This plate, which
extends from the metanotum ventrad between the mesothorax
and metathorax and serves for the attachment of locomotor
muscles, gives us a good idea of the cross-section of the thorax
and the relative surface provided for the thoracic musculature.
A comparison was made of thirty acrotergites removed entire,
ten from each of the three classes represented in figure 21. Not
only do parasitized males show a noticeable increase in the size
of the acrotergite, but even the form of the aperture through
which the digestive tube passes and the contours of the thick-
ened ribs of the chitinous plate become female in character.
The increased acrotergite indicates, I believe, that the muscles
which move the enlarged appendages have become larger than
the muscles of normal males. Although dipterous parasites are

SEXUAL CHARACTERISTICS OF THELIA 067

known to cause the degeneration of wing muscles in certain
Acridiidae (Kiinekel d’Herculais, 94), making volitation impos-
sible, parasitized Thelia can both fly and jump quite as well as
ordinary females. To move the enlarged wings and transport
the increased bulk of the body certainly larger muscles are
necessary.

B. The head

If we examine the sexual distinctions in the heads of Thelia
we find differences of color, pattern, and size. The hypodermal
pigment is similar to that of the pronotum in the respective
sexes, being orange-yellow on the vertices and clypeus of the
male and greenish-yellow in the female. ‘There is also a sexual
difference in the distribution of the melanic pigment of the face
(fig. 22.), which is not only darker brown in the male, but also
more abundant and less scattered than in the female. On the
vertex about each ocellus the male has a well-defined spot, and
along the medium suture between the vertices there is a distinct
line of brown pigment. From this vertical line there are two
diverging arms bordering the upper edge of the clypeus and
forming an inverted Y. The male clypeus has two distinct
bands extending ventrad from the arms of this inverted Y.
Along its lower angle and the borders of the genae it is deeply
pigmented. In the female, the melanic pattern is less distinct,
being present in smaller and more irregular patches. Especially
is this noticeable on the lower border of the clypeus which is a
mottled light brown in the female, deep solid brown in the male.
Parasitized males not only lose the orange-yellow hypodermal
pigment which is replaced by greenish-yellow pigment, but in
fully altered specimens also exhibit the melanic pattern char-
acteristic of the female head. These changes cannot be ascribed
to the retention of juvenal characteristics, for the color, pattern,
and structure of the integument of the nymphal head (fig. 22, ny.)
differ greatly from those of the adult and resemble the nymphal
integument of the thorax and abdomen as described on page 563.

A comparison of head widths, measured at the level of the
compound eyes, reveals the fact that parasitized males show an

568 SIDNEY I. KORNHAUSER

Fig. 22 Typical heads, showing male and female characteristics and effect
of parasites on male head. Nymphal head included for comparison. Above in
middle, normal male; below, nymph (ny.); left, normal female; right, parasitized
male (p. o). oc., ocellus; vt., vertex; clp., clypeus; gn., gena; lbr., labrum;
lbm., labium. X 7.5.

Fig. 23 Labia, to show increased length in parasitized males. From left to
right, parasitized male (p. =), normal female, normal male, parasitized male

(Os Sie 3 ALY.

SEXUAL CHARACTERISTICS OF THELIA 569

increase in width approaching that characteristic for the female.
This same relation of size increase is seen to exist when we com-
pare lengths of labia of normal males, normal females, and para-
sitized males (fig. 23). As was the case in the pronota and
wings, parasitized females show no marked changes in either
size or pigmentation of the head.

C. Extragenital abdominal characteristics

The effects of the parasites are also very marked upon the
abdomens of Thelia. The changes occurring in the extragenital
secondary sexual characteristics or tertiary sexual characteristics
will be described first.

The abdomens of the two sexes present very different appear-
ances and, directly or indirectly, many of these differences are
associated with reproduction. The female abdomen must be
capable of holding, on the average, thirty-five ova, 2.4 mm. in
length—a bulk much greater than that formed in the male
abdomen by the testes, seminal vesicles, and accessory glands.
In the male the reproductive apparatus is mature when the
nymph molts to an adult, and is contained in the abdomen
without distention; but in the female the ova are very minute
at the beginning of imaginal existence and continue to grow,
filing the abdomen and extending the chitinized abdominal
sclerites to such an extent that the intersegmental membranes
show as lighter bands between these pigmented plates. Thus
the female abdomen is much larger than that of the male,
and its cuticula is far more pliable than that of the male. This
pliability, a necessity in accommodating the growing ova, is
accompanied by a pigmentation of the abdominal sclerites,
lighter than that of the male abdomen (figs. 24and 28). Strength,
firmness, and rigidity of chitinous parts in insects is always
accompanied by a heavy melanic pigmentation, as in mandibles,
ovipositing apparatus, and muscle attachments; whereas the
absence of melanin leaves the chitin much weaker and more
pliable.

570 SIDNEY I. KORNHAUSER

Fig. 24 Abdomen of normal male, lateral view of left half, compressed under
cover-glass. oe., oedagus; cl., clasper or style; v., ventral valves; s., sternum;
p., pleuron; t., tergum. 7, 8, 9, refer to abdominal somites to which sclerite or
appendage belongs. X 9.

Fig. 25 Abdomen of normal male, ventral half, not compressed under cover-
glass. Abbreviations as in figure 24. X 9.

Fig. 26 Abdomen of parasitized male, lateral view of left half, compressed
under cover-glass. Abbreviations as in figure 24. X 9.

Fig. 27. Abdomen of parasitized male, ventral half, not compressed under
cover-glass. Abbreviations as in figure 24. XX 9.

SEXUAL CHARACTERISTICS OF THELIA 571

Each typical abdominal segment consists ventrad of a sternum
and two pleura (figs. 25 and 29, s, p) and dorsad and laterad of a
tergum, bent into an arch which is somewhat more pointed at
its apex in the female than in the male. The sterna of the
female are much broader than those of the male. With the
pleura which bear the spiracles, the sterna form a flat ventral
surface which, at its union with the terga, forms a sharp ventro-
lateral angle. Thus the female abdomen is almost triangular in
cross-section. In the male the sterna are bent slightly dorsad
and the pleura are also bent upward, making them lateral rather
than entirely ventral (fig. 25). No sharp angle is formed at the
union of the male pleura with the terga. The abdomen is rather
subovoid in cross-section. The sterna and particularly the
terga of the male abdominal somites are more deeply pigmented
than those of the female.

When we examine the abdomens of parasitized males (fig. 27)
we see that the sterna increase greatly in width and that the
abdomen in cross-section becomes similar to that of the female.
The pleura become flattened plates entirely ventral in position
and form with the terga a sharp ventrolateral angle. All the
abdominal sclerites show a loss of pigmentation and a corre-
sponding decrease in strength or rigidity. The terga (fig. 26)
have even less melanin than those of most normal females, and
the pigment remaining is restricted chiefly to the regions of
muscle attachments.

Two of the changes effected by the parasites on males, namely,
increase in size of the abdomen and decrease in firmness, strength,
and pigmentation of the sclerites, are to be found even in indi-
viduals still showing predominatingly male characteristics in the
thorax and head (figs. 36 to 39). Broad translucent sterna
through which greenish fat and some red pigment show are an
infallible clew to the presence of parasites. These changes are
requisite to the development of Aphelopus. The larvae must
have sufficient space in which to develop, and when full grown
present a bulk quite as great as the ova of a mature female
Thelia. The narrow abdomen of the male would be insufficient
for the development of the polyembryonic brood which, before

Fig. 28 Abdomen of normal female, lateral view of left half, compressed
under cover-glass. gn. 8, gonapophysis of eighth segment forming left half of
sheath surrounding ovipositor; gn. 9, ovipositor composed of anterior gonapo-
physes of ninth segment; gn. 9; posterior gonapophysis of ninth segment which
partly covers gn. 8 and gn. 9 when in natural position, as in figure 29; s., sternum;
p., pleuron; ¢t., tergum; 7. 8, 9, refer to abdominal somites to which sclerites or
appendages belong. X 9.

Fig. 29 Abdomen of normal female, ventral half, not compressed under
cover-glass. Abbreviations as in figure 28. X 9.

Fig. 30 Abdomen of parasitized female, lateral view of right half, compressed
under cover-glass. Abbreviations as in figure 28. X 9.

Fig. 31 Abdomen of parasitized female, ventral half not compressed under
cover-glass. Abbreviations as in figure 28. X 9.

072

SEXUAL CHARACTERISTICS OF THELIA 573

emerging, distends even the enlarged abdomen to its full capac-
ity. Moreover, the larvae would be unable to bore through
normally chitinized sterna, so these plates are thin and trans-
lucent in all parasitized individuals.

The abdomens of parasitized females (figs. 30 and 31) remain
in cross-section and size similar to those of normal females. A
loss of pigment is seen in all the sclerites and the sterna remain
thin and delicate as in parasitized males to permit the escape
of the larvae.

Turning now to a consideration of the details of the abdominal
integument, we note two distinct sexual differences. The ab-
dominal sclerites of the second to eighth segments, inclusive,
possess long, scattered, chitinous hairs directed caudad, and in
addition to these there are minute hairs forming very distinct
patterns (figs. 32 to 35). These minute hairs are very differently
arranged in the two sexes. Figures 32 to 35 are camera-lucida
drawings representing a corresponding region of the sixth ab-
dominal tergum of typical individuals. Figure 32 represents the
arrangement found in normal males, fairly straight and com-
pact rows running laterad and ventrad. The pattern on the
female terga is strikingly different (fig. 34). The spines here
form a network, made up roughly of rows of half-circles arranged
alternately in such a way that the ends of the half-circles of one
row touch the apexes of those in the row more anterior.

In parasitized males (fig. 33) there is a complete loss of the
characteristic arrangement of these spines and almost a com-
plete assumption of the female pattern. The only difference
to be noted is that the ends of the ares are not so well formed,
usually lacking a few spines to complete the articulation with
the row in front. This change of pattern is one of the clearest
qualitative sexual changes in the abdomen of parasitized males
and cannot be ascribed to anything except the assumption of
a female characteristic. Parasitized females are similar to nor-
mal females in the above-described characteristic (fig. 35).

Short spines or hairs are also present on the sterna. In the
male these are arranged in rows running in straight lines across
the sterna from pleuron to pleuron. Camera-lucida drawings

KORNHAUSER

SIDNEY I.

574

yO ~ ~ .

yidi, Dos \
Wie ieee Aa re SG Mie May),
is ie ee ay NY > ae tice
Ny “0. NVA) erin: 4 A > = AM, thay, WWihay \
Wu? ale Why Sra Say ise ees Pe
" cyte ee M4 ys NO? Ye mh ie
ee ass l COD oe cS NY
Sy = ey y \s \ ~ s Wily
Prus +" dU) Tula a = *f > Da Sb (Wg ve \\
= Le ys v N : :
: Y ? q az 4 z : e | \
v1 +» “ yer xdy wis Re = aula) s R Vi \\
4, alii yey, re t Seesaw ba ‘4 w ay ae, &
Siaees to evita oe mer ONe , . is
y Ny ye Md. \dd ol egy aa: ys. AY, x
3 Went plld/],. > : es yl My me ME 4 3 Pe
, » = i . te ‘é ‘y**s
a> oe pla & a > 4, NJ ‘dy > \
ne 7 SOS S WOU oe,
e e ld, » r
Pe . e e
Y We Os “ty e:
+
VING »
Weise — aly NVyy) in)
© BAN HaS wads id
“By Muee tees y 8
. anil\\va e ee
eee. iy Mui shh L\ie\.. 1 1 “4 14! goer 4 im
NY ans! \\ ludtvayge Vy\ hi oe % Y N\ M4 Ws =
sobele Jt vi Biri bh Ab = tale, ay Y yhle 5 °
\\\. bau y J] Ql > EN 2
mdb aval n: yriblas blr eaney as > rene Neh ae
vee Me olh ial o> sbavdeldadyd ad “Gq AS Ree Von a
Nat d 4 = Why
Me ood 4¥d ouabh Pry NOW TEINUY/y “de, Nod My Siri, ays iz E
o-*” a AES SOUNTL Ving bbz e i uy. ; dy A
eal TW NN eo Whyy Thy Nay So Ay #
ep SILL I. “ales Woodys Bacal nie sey ” A y 0g,
Weryppars7 17 o@y dled rypel> ON X\ 5 : hee é
e 4 a
> 4

All draw-
Corresponding

Figure 32, normal male; figure 33, parasitized
x 480.

figure 35, parasitized female.

.
b)

Surface view of integument of sixth abdominal tergum, show-

ing typical arrangement of minute hairs in each class represented.

Figs. 32 to 35
The dotted circles with light centers represent the long hairs which are scattered

ings from a corresponding dorsolateral area of the sixth abdominal segment.

area of nymphal integument shown in drawing in figure 18.

irregularly over the integument.
male; figure 34, normal female

SEXUAL CHARACTERISTICS OF THELIA 575

show these rows to be on the average 7.5u apart. In the females
these rows arch cephalad on the sterna and are 9.34 apart on
the average. In parasitized males, instead of forming straight
lines, they are curved as in the female and are 9.5 apart.

These changes in the arrangements of the minute spines of
male abdomens may not be ascribed to the retention of juvenal
characteristics, for the integument of the fifth nymphal instar of
both sexes is quite different from that of the adult. The terga
of the nymph are covered with large jointed hairs (fig. 18, p. 562)
and lack entirely the minute hairs arranged in patterns in adult
terga. Also the nymphal sterna are clothed with peculiar
triangular spines quite unlike those of the adult and not arranged
in definite rows.

The sclerites of the eighth and ninth abdominal segments and
the sterna of the seventh segment are quite dissimilar in the
two sexes, being modified in each sex to accommodate the gona-
pophyses. Thus the sterna of the seventh and eighth abdominal
somites are rectangular in the male (figs. 24 and 25, s. 7, s. 8),
not unlike those more anterior. The sternum of the ninth
segment of the male abdomen is a single small heart-shaped
sclerite (fig. 24, s. 9), which articulates laterad with the claspers
and caudad at its apex with the oedagus. In the female the
immense ovipositor extends so far cephalad that the sternum of
the seventh somite (fig. 28, s. 7) is deeply indented, its caudal
border forming an are cephalad which almost divides the sternum
into two plates at its median plane. The sternum of the eighth
somite in the female (fig. 28, s. 8) is represented by two laterally
situated plates, each articulating with a gonapophysis of the
pair belonging to that somite (fig. 28, gn. 8). In the ninth
somite also the sternum is divided into two parts, each having
the outline of a Roman lamp. Each of these plates (fig. 28,
s.9) is covered by the pleuron of its respective side and has
articulations with two gonapophyses (gn. 9 and gn. 9’), one at
its cephalic end and one at its caudal end.

In parasitized males, although the genitalia are not produced
cephalad, the sternum of the eighth somite is often almost divided
into two plates by a deep notch from its caudal border (fig. 27,

576 SIDNEY I. KORNHAUSER

s. 8). The sternum of the ninth segment, instead of being a
single heart-shaped plate, is represented by two minute sclerites,
each placed laterad against the clasper of its corresponding side
and about equidistant from the ends of this organ. It would be
difficult to explain these changes in infected males as being
due to mechanical necessity caused by alterations of the gona-
pophyses, and they can only be understood, I believe, in looking
at them as a partial assumption of female characteristics.

The sterna of parasitized females show the following changes:
that of the seventh segment bears merely a notch in the median
plane running cephalad from its caudal border; that of the eighth
remains a single sclerite, being not quite bisected by the gona-
pophyses (fig. 31, s.7, s.8) which are greatly shortened by the
action of the parasites. The ninth segment presents the sternum
in two sclerites, retaining their original position and articula-
tions, but being somewhat larger than normal and less well
chitinized.

A distinct sexual difference is found in the tergum and pleura
of the ninth abdominal segment. In the male (fig. 24, t. 9, p. 9)
the tergum and pleura of this segment are separate sclerites
which are united by a thin, pliable, chitinous membrane. The
pleuron is almost semicircular and projects caudad beyond the
tergum. In the female (fig. 28, ¢ 9, p. 9) these sclerites are
much larger and quite different in form from those of the male.
They are fused into one, the suture on each side being visible
as a curved line passing from the anal tube to the lower border
of the eighth tergum. The integument on each side of this
suture is of a distinct character; the dorsal portion being similar
to that of the other terga, whereas the ventral portion is beset
with long hairs. The pleuron, unlike the semicircular sclerite
of the male, is drawn into a long plate, extending almost the
entire length of the external genitalia (fig. 29, p. 9).

In parasitized males (figs. 26 and 27, t. 9, p. 9) the tergum and
pleura fuse as in the female. Moreover, both plates increase in
length, the pleura lose their semicircular form and fail to pro-
ject caudad beyond the tergum. Thus all the male characteris-
tics are lost and an approach to the female condition takes place.

SEXUAL CHARACTERISTICS OF THELIA 577

In parasitized females (figs. 30 and 31, t. 9, p. 9) no qualitative
change occurs; the tergum remains practically normal and the
pleura decrease in length, accommodating themselves to the
shortened gonapophyses.

Summing up the changes in the extragenital abdominal char-
acteristics in parasitized males, we find that the abdomen in-
creases in size; that the pleura become flat and ventral in posi-
tion and in cross-section the abdomen is similar to that of the
female; that the hard dark brown integument becomes thin and
pliable and lightly pigmented; that the patterns formed by the
minute spines on the terga and sterna assume the arrangements
characteristic of the female; that the sternum of the eighth
abdominal segment is almost divided into two plates, while that
of the ninth often separates completely, forming two minute,
lateral sclerites; that the tergum and pleura of the ninth seg-
ment fuse, increase in length, and partially assume the form of
the corresponding sclerites of the female.

In the females the parasites do not cause the abdomen to
increase in size or change its form in cross-section. However,
the integument becomes thinner and there is some loss in pig-
mentation. Since in the female the pigmentation is normally
much less intense than in the male and the cuticula more pli-
able, these changes are not nearly so radical as those in the male.
The arrangement of the minute hairs on the abdominal sclerites
remains unaltered in the female. The sternum of the seventh
abdominal segment is not so deeply notched at its caudal border
and that of the eighth segment only partially divided into two
sclerites by the ovipositor, which is much shortened. The form
of the tergum of the ninth segment remains unchanged, while
the pleura decrease in length.

D. The genital appendages

As has been intimated in the foregoing pages, the external
genitalia suffer a considerable reduction in size in parasitized
individuals of both sexes. This is quite in contrast in the male
to the reaction of the extragenital secondary sexual characteris-

578 SIDNEY I. KORNHAUSER

tics. Since the head, thorax, and abdomen of parasitized males
show such a decided change toward the female condition, it was
of course puzzling to understand why the genital appendages
showed not the slightest tendency toward such a transformation.
A study of the structure of the adult gonapophyses of both sexes
and their mode of origin was undertaken. The results of this
study not only justify the classification of sexual characteristics
used in this paper, but also show why the two categories of
secondary sexual characteristics behave differently under the
effects of parasitism.

The external genitalia consist of three pairs of appendages in
both sexes. In the male they are terminal, located on the last
complete abdominal segment, the ninth. In the female they
are produced cephalad on the ventral surface, reaching into the
indentation of the seventh sternum. They are, however, con-
nected entirely with the sclerites of the eighth and ninth ab-
dominal somites.?

The male genitalia (figs. 24 and 25) consist of a pair of ventral
valves, a pair of styles or claspers, and an unpaired oedagus.
The ventral valves (fig. 24, v.) are flattened sclerites united in
the median plane for more than half their length, and produced
caudad into a pair of free, narrow appendages, composed of two
layers of chitinous cuticula closely approximated and envelop-
ing the hypodermis which produced them. ‘The claspers are
strong chitinous rods, each produced into a hook laterad to the
oedagus (figs. 24 and 25, cl.). Each articulates with the side of
the heart-shaped sternum of the ninth segment and extends

2 The terms used in reference to the genital appendages are merely descrip-
tive and bear no phylogenetic implications. From a study of a more primitive
homopteran, such as the ordinary Cicada linnei Grossb., one may see that the
gonapophyses originally arose from the eighth, ninth, and tenth abdominal
somites. This is very clearly seen in the female, although a trifle obscured in
the male, where the appendages are so modified and specialized that they are of
great use to the systematist in a study of the Cicadidae. It is the writer’s belief
that in the membracids the gonapophyses originated from the primitive eighth,
ninth, and tenth abdominal appendages, even though in the male they all seem
to arise from the ninth segment. A careful study of the embryonic stages might
reveal a migration and persistence of the cells of the limb buds of the segments in
question.

SEXUAL CHARACTERISTICS OF THELIA 579

cephalad within the abdomen half-way through the eighth
somite. To this internal portion muscles are attached which
move these appendages. The oedagus is a heavily chitinized,
tubular organ containing the penis. Its proximal end articulates
with the caudal tip of the ninth sternum. The oedagus bends
sharply dorsad and comes into close approximation with the
anal tube (figs. 24 and 25, oe.).

The female external genitalia are seen in figures 28 and 29.
The appendages of the eighth segment are two sharply pointed
plates, each articulating with the lateral remnant of the eighth
sternum (fig, 28, gn. 8). They form a strong, closely fitting
sheath about the ovipositor, being interlocked along the median
plane by a sort of dove-tail union. The anterior appendages of
the ninth segment form a tube, the ovipositor, open along its
ventral border at both its proximal and distal ends (fig. 28, gn. 9).
Its proximal portion consists of two strong, chitinous rods which
diverge and articulate laterad with the two small sclerites repre-
senting the ninth sternum. The distal third of the ovipositor
bears three pairs of well-defined teeth along its dorsal border.
The third pair of gonapophyses, the posterior pair of the ninth
segment, are flattened sclerites, laterally placed, partly covering
the ovipositor and its sheath (fig. 28, gn. 9’). Each articulates
with the caudal portion of the ninth sternum of its respective
side.

In parasitized males the oedagus and claspers are greatly
reduced in size and are located on the ventral portion of the
enlarged eighth somite (figs. 26 and 27). The oedagus is short-
ened to form a stout tube, and the claspers do not reach even to
the base of the ninth segment, but retain quite well their orig-
inal form. The ventral valves show much less reduction in
size than the oedagus or clasper. They, however, present many
irregularities in that the free caudal projections aze often mis-
shapen and of unequal length. On the average, the ventral
valves are reduced only 12 per cent and the maximum reduction
found was 24 per cent of the normal length. Their reduction is
not as constant as that of the other male genital appendages, for
in males, otherwise greatly altered, they may at times be prac-
tically normal in length (fig. 26, v.).

JOURNAL OF MORPHOLOGY, VOL, 32, NO. 3

580 SIDNEY I. KORNHAUSER

If, now, we examine parasitized males, which have retained
the more striking tertiary sexual characteristics of color, form,
and size (figs. 38 and 39), we find that they, too, generally have
much reduced genitalia. Occasionally, male colored individuals
with parasites present only a partial reduction of the gona-
pophyses and these most probably represent those parasitized
shortly before their mcult to the adult form (figs. 36 and 37).
The above facts show that the genital appendages are very
sensitive to the influence of Aphelopus, being reduced even in
individuals in which the effect of the parasites was not sufficient
to alter the pigmentation of the pronotum or face.

In parasitized females, all three pairs of genital appendages
show a great reduction in size (figs. 30 and 31). They are some-
times weakly chitinized, bent, and misshapen. The ovipositor
(fig. 30, gn. 9) does not always form a tube, but consists of two
separate plates diverging distally. All three pairs of append-
ages are reduced proportionately, there being no one pair as
little affected as the ventral valves of the male genitalia.

The above facts show that, although there is a decided reac-
tion of the external genitalia due to parasitism, vet there is not
the slightest tendency of these appendages to assume the char-
acteristics of the opposite sex. They are, therefore, entirely
different in the male from the extragenital secondary characters
in their behavior. The reason for this is to be sought in a con-
sideration of the origin and development of these two cate-
gories of characteristics. The extragenital secondary or tertiary
sexual characteristics, such as color of pronotum and face, ar-
rangement of spines on the abdominal sclerites, and many others,
arise during the fifth nymphal instar and make their first appear-
ance in the adult Thelia. If the nymph be a male and contain
well-grown Aphelopus larvae, the resulting adult will exhibit
female extragenital sexual characteristics; but the genitalia,
though greatly reduced in size, will unquestionably remain male
in character, because these appendages did not arise in the fifth
nymphal instar, but were laid down early in ontogeny, either
before the parasites were present or while the parasites were
minute and incapable of exerting any marked influence.

SEXUAL CHARACTERISTICS OF THELIA 581

Figs. 36 to 39 Abdomens of parasitized males which still retained almost
normal coloration of pronotum. Compare with figures 24 to 27. Figure 36,
lateral view, abdomen compressed, only partial reduction of gonapophyses.
Figure 37, ventral view, abdomen not compressed, partial reduction of gona-
pophyses. Figure 38, lateral view, abdomen compressed, extreme reduction of
gonapophyses. Figure 39, ventral view, abdomen not compressed, extreme
reduction of gonapophyses. oe., oedagus; cl., clasper or style; v., ventral valve;
s. 8, sternum of eighth abdominal somite. X 9.

582 SIDNEY I. KORNHAUSER

Many descriptions of membracid nymphs have been written
and a detailed account of the five nymphal instars of Thelia
has recently been published (Funkhouser, ’15), but no mention
is made of characteristics which may be used to distinguish the
sexes. Neither in coloration, nor in the form of the integument,
nor in the thoracic appendages is any clew given to the sex of
the nymph. In the fifth instar the female abdomen is noticeably
larger and wider than the male abdomen; but an infallible test
for the sex of the nymph is to be found by examining the ventral
surface of the eighth and ninth abdominal segments. The
genitalia are so well developed in the nymphs and are so dis-
tinctly different in male and female that, as early as the third
instar, one can determine the sex by an examination of the ven-
tral surface of the eighth and ninth somites with a hand lens.
Figures 40, 41, and 42 show the form of the genitalia in male
nymphs of the third, fourth, and fifth instars, respectively,
whereas the corresponding stages in the female are shown in
figures 44, 45, and 46. The sexes are easily recognizable, not
only by the form and position of the appendages, but also by the
pigmentation of the eighth sternum.

Examining the characteristics of the three stages of the male
represented in figures 40, 41, and 42, we find that the genitalia
arise in a triangular-shaped area on the ninth abdominal seg-
ment. Small chitinous pockets pointing cephalad contai then
hypodermal cells which produce the genitalia. In each instar

Fig. 40 Abdomen of male third nymphal instar, ventral view, caudal portion
showing genital area (g. a.) and characteristic pigmentation of sternum of eighth
somite (s. 8). XX 26.2.

Fig. 41 Abdomen of male fourth nymphal instar, caudal segments, ventral
view, showing inner (dorsal) pockets as smaller pair, and outer (ventral) pockets
of genital area. Medial partition in each pair. X 26.2.

Fig. 42 Abdomen of male fifth nymphal instar, ventral view of entire ab-
domen. Inner pockets now three in number, median and two lateral; outer
pockets still right and left of median partition. Pigment confined to ends of
pockets and partitions. X 9.7.

Fig. 43 Abdomen of parasitized male fifth nymphal instar, ventral view of
entire abdomen, showing extreme reduction of inner pickets of genital area due
to action of parasites. The holes in the sterna were made by escaping eruciform
larvae, which left their exuviae (those of megagnathic stage) at puncture holes.
x 9.7.

SEXUAL CHARACTERISTICS OF THELIA 583

584 SIDNEY I. KORNHAUSER

these cells within the pockets produce the appendages of the
next stage. In molting the newly formed appendages are with-
drawn from these pockets cephalad, and the soft, wrinkled in-
tegument unfolds to produce the larger gonapophyses. Thus
there is a gradual growth of the external genitalia. In the male
third instar there are two pairs of pockets, one placed dorsally to
the other on the ninth segment. Each pair is separated into its
right and left components by a median chitinous partition extend-
ing cephalad from the apex of the triangular genital area. The
inner pair of pockets is smaller than the pair more ventrally
placed. There is no evidence for any appendage arising from
the eighth somite in the male. The ninth segment has a band
of melanie pigment about its anterior portion and this pigmented
band is slightly more intense where it crosses the genital area.
The sternum of the eighth segment bears a pigmentation char-
acteristic of this stage and of the male sex. There is a lght
medium stripe and a light patch extending from each side toward
the median plane, and this is quite different from the pattern
on the eighth sternum of the female third instar. In the fourth
instar (fig. 41), the triangular genital area increases considerably
in size and the two pairs of chitinous pockets are more easily
seen at its apex. The pair more dorsally placed does not extend
quite so far caudad as the larger ventral pockets. The pig-
mented ring of the ninth segment remains and broadens out
on the genital area, while the melanin on the eighth sternum
retains its former pattern, but is less intense. In the fifth instar
(fig. 42), the ventral pockets which produce the ventral valves
of the adult retain the form seen in the previous stage, a median
partition separating the right from the left, but the dorsal pair
change greatly. Instead of a median chitinous partition, there
are two lamellae, which divide the pocket into there subdivi-
sions: a larger median compartment and two smaller lateral
compartments. In the median pocket the oedagus develops,
and in the lateral pockets, the claspers. The melanie pigment is
restricted to the caudal end of the genital area.

Turning now to the development of the ovipositing apparatus,
we find in the third instar (fig. 44) a pair of darkly pigmented,

Fig. 44 Abdomen of female third nymphal instar, ventral view of caudal
portion, showing gonapophyses of eighth (gv. 8) and ninth (gn. 9) segments, and
pigmentation characteristic of sternum of eighth somite (s. 8). X 26.2.

Fig. 45 Abdomen of female fourth nymphal instar, ventral view of caudal
portion. All three pairs of gonapophyses represented: one pair of the eighth
(gn. 8), and two pairs on the ninth segment (gn. 9, and gn. 9’). The adult ap-

pendages which these form are shown with similar designations in figures 28 and.

29. X 26.2.

Fig. 46 Abdomen of female fifth nymphal instar, ventral view of entire ab-

domen, showing growth caudad of gn. 8, which now cover and extend beyond
gn. 9. gn. 9’ now partly overlap the two median pairs of gonapophyses. X 9.7.

Fig. 47 Abdomen of parasitized female fifth nymphal instar, from which
larvae have emerged. Retardation of gonapophyses, which remain similar to
those of fourth instar. > 9.7.

585

586 SIDNEY I. KORNHAUSER

chitinous pockets arising from the caudal border of the eighth
sternum. The presence of these tell us immediately that the
nymph is a female, as does also the pigmented area on the eighth
sternum anterior to these pockets. On the ninth segment is a
pair of narrow pockets more deeply pigmented at their posterior
border. In the fourth instar the three pairs of nymphal append-
ages which form the three pairs of adult gonapophyses can be
most plainly seen (fig. 45).

The pair arising from the eighth somite, which in the adult
form the sheath covering the ovipositor, have grown consider-
ably caudad, partly overlapping the median and anterior pair of
the ninth segment which eventually produce the ovipositor
itself. Lateral to the anterior pair of the ninth segment are
two lightly pigmented, slender, curved pockets, and in these
the posterior gonapophyses of the ninth segment are produced.
The pigmented area of the eighth sternum is similar to that
of the third instar. ‘The fifth instar (fig. 46) shows a con-
siderable growth and closer approximation of all three pairs of
appendages. Those of the eighth segment extend caudad be-
yond the narrower pair of pockets which form the ovipositor.
The posterior appendages of the ninth segment now form broad
lateral pockets, partly overlapping the two median pairs. As
the adult appendages develop they are seen through the
nymphal cuticula, the most anterior gonapophyses extending
from the pigmented cephalic border of the eighth segment be-
yond the tip of the ovipositor.

Thus we see that the genitalia develop quite differently in the
two sexes and are determined quite early in ontogeny. The
above description goes back merely to the third instar, at which
stage one can distinguish the sexes with ease from surface views;
but long before this the anlage for the gonapophyses are formed.
Careful examination of whole mounts of second instars also
reveals the differences: the form of the genital area on the ninth
segment is broad and in the form of a semicircle in the male,
whereas in the female it is much narrower and resembles that of
the third instar. In the female one can also see the beginnings
of the gonapophyses of the eighth segment as two minute chit-

SEXUAL CHARACTERISTICS OF THELIA 587

inous elevations on the caudal border of the eighth sternum.
Going back still farther to the first instar, longitudinal sections*
show thickenings of the hypodermis corresponding to the append-
ages seen in surface views of later stages, and, since the gonads
are already differentiated into minute ovaries or testes, one
knows positively the sex with which he is dealing.

In considering the effects of the parasites on the gonapophyses,
it might be inquired whether or not these appendages are modi-
fied in their nymphal form by the larvae of Aphelopus. As was
stated on page 552, nymphs stung during their first or second
instars do not become adults, for the internal larvae reach ma-
turity in the fifth instar of the host and emerge, thereby killing
the Thelia. Two such nymphs from which the larvae have
emerged are shown in figures 43 and 47. In the male (fig. 43)
there are two changes in the nymphal genitalia. The dorsal
pockets suffer a great reduction in size and remain in part sim-
ilar to those of the fourth instar in that the median partition
still persists. The two pockets are each again subdivided by a
partition corresponding to those of the normal fifth instar, but
these are very small and run parallel to the long axis of the
animal rather than at an angle toward the median plane, as in
normal nymphs. The condition is, therefore, intermediate
between the fifth and fourth instars, for the medium partition
is characteristic of the fourth instar and is not present in the
fifth instar, whereas the lateral partitions are characteristic of
the fifth instar and not present in normal fourth instars. The
persistence of a characteristic of the fourth instar indicates
retardation of development. In the parasitized female nymph
(fig. 47) not only are the nymphal genitalia reduced in size, but
their form corresponds exactly to that of the normal fourth
instar, each pair of gonapophyses remaining distinct and easily
differentiated. After having ascertained that during the fourth

3 At the suggestion of the author, Mr. E. D. Churchill made a histological
and gross study of the development of both external and internal genitalia of
the membracid Vanduzea arquata (Say). In this he traced the gonapophyses
from the first instar in histological sections. His work substantiates exactly
what has been found true of Thelia.

588 SIDNEY I. KORNHAUSER

instar the parasites exert an influence upon the host sufficient to
change the appendages at the succeeding molt, an attempt was
made to see if this could be carried back to still earlier stages.
Accordingly, a number of male and female nymphs of fourth
instar each containing parasites of maximum size for that stage
(0.75 mm.) were prepared and mounted similar to the normal
individuals shown in figures 41 and 45. Careful camera-lucida
drawings were then made of twenty individuals, ten parasitized
and ten normal. The width and length of the nymphal append-
ages were measured on the drawings, and in neither sex could
any alteration from the normal in the size or form of the append-
ages or in the pigmentation of the integument be found. From
this we may infer that in the third instar the parasites are still
too minute to exert any influence upon the cells producing the
integument and appendages of the fourth instar. Since the
genital appendages are laid down even before the third instar,
we can see that the parasites could in no way interfere with the
formation of the anlage of the gonapophyses. Only during the
fourth and fifth instars do the parasites affect the genitalia, and
this effect appears in the retardation of development and reduc-
tion in size found in the fifth nymphal instar and the great
reduction in size in the adult. Even minute parasites, still
spherical embryos just becoming separate individuals, cause the
adult genitalia to be reduced. ‘This is probably due to the fact
that the greatest step in the formation of the gonapophyses
comes in the development of the adult from the fifth nymphal
instar. That the genitalia are reduced in size but retain their
general form in parasitized individuals is probably to be ascribed
to their history.

It is generally conceded that organs formed early in ontogeny
are phylogenetically older than those appearing late in develop-
ment. The extragenital secondary sexual characteristics which
arise during the fifth instar certainly belong to the species. The
genital appendages are relatively older and appear early in
ontogeny. In the later stages of development various specific
modifications may slightly alter the gonapophyses. Systema-
tists have shown that the Membracidae possess a relatively

SEXUAL CHARACTERISTICS OF THELIA 589

ancient type of genitalia found in practically all primitive ho-
mopterans. Funkhouser (’17) considers that the family Mem-
bracidae is the lowest of the Homoptera with the exception of
the Ciecadidae. One of his four reasons for this assumption is
that in the Membracidae ‘‘the genital organs are simple. Little
progress has been made in developing these structures from an
ancient type.” Considering, then, that the gonapophyses of
Thelia are still of the same type as that established when the
order Homoptera evolved in the distant past, we see why, under
the effects of parasitism, they remained more constant than did
the extragenital sexual differences. The formation and sexual
differentiation of these appendages are started before the para-
sites are present. The mechanism once under way is not re-
versed by parasitism, but the resultant products are greatly
reduced in size and suffer, as will be shown, the loss of certain
adult characteristics specific for Thelia bimaculata.

Giard (’89) pointed out that parasitized individuals of Typh-
loeyba showed a reduction of the external genitalia in both
sexes. The male of Typhlocyba hippocastani has a very com-
plicated oedagus produced into eight branches at its distal end,
and this organ is specific for hippocastani, distinguishing it
nicely from T. douglasi. Males parasitized by Atelenevra show
a reduction of the branches of the oedagus to six, four, or three.
Thus the specific character of the oedagus is ‘profoundly modi-
fied,’ so that parasitized hippocastani may be confused with
Typhlocyba rosae L. or Typhlocyba lethierryi J. Edw.

The specific characteristics of the gonapophyses first appear
in the final molt, and they are probably recent from a phylo-
genetic standpoint. As other specific characteristics, one would
expect that they would fail to develop or develop but partially
in parasitized individuals. Thus a comparative study of the
gonapophyses of Thelia bimaculata Fabr. and its nearest avail-
able relatives was undertaken, to ascertain if bimaculata pos-
sessed any specializations which might be looked upon as specific
modifications. Through the generosity of Dr. E. D. Ball various
membracids of the tribe Telamonini were made available, and
these fortunately included the rare Thelia uhleri Stal. Normal

590 SIDNEY I. KORNHAUSER

individuals of both sexes belonging to the following species were
prepared and mounted in balsam as was done earlier for Thelia
bimaculata (figs. 24 and 28): Thelia uhleri Stal, Glossonotus
acuminatus Fabr., Glossonotus godingi Van D., Telamona
querci Fitch, Telamona reclivata Fitch, Archasia _ belfragei
Stal, and Carynota mera Say.

The one outstanding characteristic of the male genitalia of
Thelia bimaculata which distinguishes it from the others studied
is the shape and length of the oedagus. All the Telamonini
have the oedagus bent sharply dorsad toward the anal tube.
At the bend, in the forms above named with the exception of
Thelia bimaculata, the oedagus is narrowed and becomes thicker
and bulbous toward the distal end. Viewed in profile, the
inner surface of the portion dorsad to the bend is practically a
straight line, whereas the outer or free surface is greatly curved.
Thelia uhleri Stal, the nearest relative to Thelia bimaculata
Fabr., also possesses a bulbous oedagus narrowed at the bend
dorsad; but T. bimaculata has a slender almost tubular oedagus,
not bulbous at its distal end or narrowed at the bend dorsad.
Its inner surface viewed in profile is not a straight line, but is
curved so as to be practically parallel to the outer or free surface.
In comparison to its diameter, the oedagus of Thelia bimaculata
is longer than in any of the other forms studied, and its distal
third bends cephalad as well as dorsad. In parasitized individ-
uals the oedagus is reduced in length much more than in diam-
eter. The inner edge shows a decrease in its curvature so that
the thick bulbous form found in the other Telamonini becomes
approximated in the reduced organ. Thus the specific form of
the oedagus so characteristic for Thelia bimaculata is lost.

The female gonapophyses of the various Telamonini avail-
able were likewise studied. Of all the forms Thelia bimaculata
had the bluntest, most rounded tip at the distal end of its ovi-
positor. The ends were found to be strongly chitinized and
reinforced by longitudinal ribs. In all the forms, with the
exception of Thelia bimaculata, near the tip of the ovipositor on
its ventral border there is a thin translucent area. This is
directly anterior to a chitinous rib which runs diagonally on

SEXUAL CHARACTERISTICS OF THELIA 59L

each side of the ovipositor from a median rib toward the ventral
border. Thelia uhleri has a pointed ovipositor with a trans-
lucent area behind a diagonal rib near the tip. This seems to
be a general characteristic of the Telamonini studied. Ovi-
positors of parasitized Thelia are not only reduced in size, but
they are more pointed at their distal ends than those of normal
individuals. The chitinous ribs are reduced in size, and one
may easily discern a thinner area near the tip on the ventral
border.

Thus both oedagus and ovipositor lose through parasitism
characteristics specific for Thelia bimaculata. These specific
characteristics of the genitalia must be recent acquisitions in
the genetic make-up of this highly specialized form. They
merely modify and add details to the gonapophyses, organs long
established and rather stable in most homopterans. Whereas
the genitalia of parasitized individuals retained the general form
found in the tribe Telamonini, they lost their specific character-
istics through the action of the parasites.

7. AN ACCOUNT OF TWO SIGNIFICANT INDIVIDUALS

Two extraordinary individuals, one captured in the summer
of 1916 and the other in the summer of 1917, throw consider-
able light on the relation of the gonads to the soma in Thelia,
and are too important to omit even in this paper which deals
mainly with the effect of the parasites upon external features of
the host.

The first individual was an adult parasitized male. The para-
sites were of medium size, measuring 0.75 mm. from the post-
cephalic region to the posterior end of the abdomen. Their
mouth parts were not chitinized and the curved bodies of the
parasites were enclosed in nutritive envelopes. However, the
larvae had already exerted a marked influence on the host.
The pronotum (fig. 48) showed considerable change in colora-
tion toward the female condition. The bright yellow was absent
from the vitta, and the melanic pigment over the rest of the pro-
notum was no longer uniform and encroached upon the clear
chitin of the vitta. The length of the pronotum was 11.90 mm.,

092 SIDNEY I. KORNHAUSER
which is 0.35 mm. greater than the average length in males.

The fore wing showed even a more marked increase in length,
being 9 mm. long, which is nearer the average length in normal

-

Figs. 48 to 50 Views of a parasitized male which contained one full-sized
testis and showed somatic changes toward the female condition. Figure 48,
pronotum. X 6. Figure 49, abdomen, lateral view of left half compressed
under cover-glass; 0e., oedagus; cl., clasper; p. 9, pleuron of ninth segment;
t. 9, tergum of ninth segment. X 9.7. Figure 50, four tubules of testis as seen

in section. X 84.

SEXUAL CHARACTERISTICS OF THELIA 593

females (9.78 mm.) than the average length in normal males
(8.13 mm.). The pattern of the face also showed a decided loss
of male characteristics and an assumption of those of the female.

Turning now to the abdomen (fig. 49), we find there also strik-
ing changes. The terga form a sharp ventrolateral angle where
they join the pleura and the cuticula shows a marked reduction
of melanic pigment and has become more pliable. Examined
microscopically, the terga exhibited the arrangement of the
minute spines in the pattern characteristic of the female (figs.
32 to 35). The tergum and pleura of the ninth abdominal seg-
ments fused together and became longer than those of normal
males. The oedagus and claspers were greatly reduced in size.

The most interesting feature about the male Thelia described
above is that in the right half of the abdomen there was located
an entire testis, normal in size. This was immediately removed
and placed in Bouin’s fluid, drawn with the aid of a camera
lucida to obtain its exact dimensions, and then imbedded and
sectioned. All stages of active spermatogenesis were found.
There were spermatogonia quiescent and in mitosis, spermato-
cytes in growth and maturation, spermatids undergoing trans-
formation, and mature spermatozoa in great numbers. Figure
50 is a photograph of four of the tubules, and, even at the com-
paratively low magnification used, one may distinguish the
cysts of spermatozoa with their deeply staining heads lying side
by side. The mitoses found in this testis were in every way
normal and the large, lagging, unpaired x-chromosome is as
noticeable as in any normal first spermatocyte division (fig. 51,
e and f).

The testis above described was the largest ever found in a
parasitized Thelia which showed marked somatic changes.
Other testes have been taken from altered males, but these in-
variably showed many abnormal mitoses, many stages of fatty
degeneration, and broken-down cysts devoid of spermatozoa.
These have been sectioned and studied and will be reported upon
at a future date. The important fact to be noted about this
male is that, in spite of the presence of a normal testis, the para-
sites exerted a marked influence on the developing soma. This

594 SIDNEY I. KORNHAUSER

is evidence that the effect of the parasites is not accomplished
through the destruction of the testes of the host (castration
parasitaire), but that the parasites have a direct effect upon the
developing tissues.

The second significant individual was a nymph of the fourth
instar. The author was dissecting a large number of nymphs

Fig. 51 Cells from normal Thelia gonads and soma. a, spermatogonium, 21
chromosomes including one x-chromosome, the largest of the group (x); b, so-
matic cell from developing external genitalia, removed from fifth nymphal instar
and stained with acetic-carmine, 21 chromosomes, one large x-chromosome;
c and d, primary spermatocytes showing large x-chromosome; e, primary sper-
matocyte, lateral view of first maturation spindle showing large unpaired x-chro-
mosome; f, late anaphase of first spermatocyte division; g, spermatid which has
received x-chromosome; h, o6gonium, 22 chromosomes including two x-chromo-
somes (the largest pair); 7, somatic cell from developing external genitalia, acetic
carmine, 22 chromosomes. X 1980.

of this stage when he was surprised, upon opening the abdomen
of a normal-appearing female, to find that a pair of testes, in-
stead of ovaries, were present. ‘The testes were quickly re-
moved, one preserved in Bouin’s fluid and the other prepared
for immediate observation in Schneider’s aceto-carmine. The
Bouin material sectioned (fig. 53) proved far more valuable than

SEXUAL CHARACTERISTICS OF THELIA 595

the freshly stained material. The body of the nymph was
preserved in Gilson’s fluid and later sectioned, with the exception
of the caudal end of the abdomen, which was prepared for whole
mount (fig. 52).

The structure of the gonads of this individual was compared
with that of testes and ovaries of normal fourth instars. The
normal testis is composed of tubules which are almost spherical
and are bound together by their efferent ducts which unite
eentrally to form the chief duct. This duct runs in the direction
of the long axis of the testis, which is roughly an ellipsoid in
form. The testis from surface view looks like a bunch of grapes.
Each tubule is composed of a number of well-differentiated
cysts, marked off by definite walls. Each cyst is filled with
spermatogonia or spermatocytes, the cells in any one cyst being
approximately in the same stage of growth or mitosis. The
ovary presents a very different structure. The most conspic-
uous portion consists of terminal chambers, elliptical in longi-
tudinal section and placed side by side in a dorsoventral plane
of the abdomen. From the anterior end of each terminal cham-
ber runs a terminal filament, and from the posterior end, an
ovarial tubule which later contains the large odcytes. The
terminal filaments of each ovary converge and form a single
support for the gonad, and the tubules likewise converge caudad
to form an oviduct. <A section of the terminal chamber shows a
wall of tall epithelial cells with clear cytoplasm. These cells are
unlike the flattened epithelial cells covering the testicular tubules.
In the region of the attachment of the terminal filament, one
encounters oogonia in various phases of mitosis, and the rest of
the chamber is filled with nurse cells and small odcytes pre-
paratory to growth. ‘There are no subdivisions of the chamber
into cysts. Thus, macroscopically and microscopically, the
gonads of the two sexes are so distinctly different that they are
not easily confused. That the gonads of the peculiar nymph
under discussion were testes, there is not the least doubt. Cer-
tain peculiarities of these testes will be noted later.

We are, therefore, considering an individual with female soma
and male germ glands. As already noted on page 586 (fig. 45),

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

596 SIDNEY I. KORNHAUSER

the three pairs of appendages which form the adult external
genitalia may be seen on the ventral side of the eighth and ninth
abdominal segments of female fourth instars. These append-
ages and the pigmented area of the eighth abdominal sternum
comprise the external sexual characteristics of the nymph.
When compared by means of camera-lucida drawings with nor-
mal female individuals similarly mounted, the external genitalia
and pigmented area of the eighth sternum of this anomalous
nymph (fig. 52) were not in any way abnormal. This is a dis-
tinct proof that in Thelia the testes do not pour forth a secretion
which influences the developing soma, or, in other words, the
soma is independent of the gonads in its development. And we
may also conclude from this individual that the development of
the gonad is to a large extent independent of the sex of the body
in which it is found.

An inquiry into the cause of the development of testes in this
peculiar nymph with female soma was next undertaken. A
cytological study of the testis proved most instructive. Sper-
matogonia in mitosis were very abundant and handsomely pre-
served so that the chromosomes stood out with diagrammatic
clearness. Eight of the most favorably situated metaphase
plates are shown in figure 54. These represented about one-
fourth of the number of sepermatogonia in which the chromo-
somes could be accurately counted. The drawings were all
made with a camera lucida, and in no cases are they reconstruc-
tions from two or more sections. Neither are the chromosomes
shifted from their original positions. As is easily seen in figure
54, these cells are true spermatogonia containing twenty-one
chromosomes, the largest being the unpaired x-chromosome.
For comparison, the chromosomes of soma and germ-plasm of

Figs. 52 to 54 Abdomen, testis, and spermatogonia from anomalous fourth
instar with female soma and male gonads. Figure 52, vental view of caudal
portion of abdomen, showing external genitalia and pigmentation of eighth
abdominal sternum; compare with figure 45. X 30.6 Figure 53, longitudinal
section of one of the testes, showing eight tubules, each divided into cysts. X
140. Figure 54, eight spermatogonia from testis, a section of which is shown in
figure 53. Each spermatagonia contains 21 chromosomes, the largest being the
unpaired x-chromosome (#); compare with figure 51, a. X 2200.

SEXUAL CHARACTERISTICS OF THELIA 597

598 SIDNEY I. KORNHAUSER

normal Thelia are shown in figure 51, most of the figures being
reproduced from a former paper (Kornhauser, 714). Figures
51, a, and 51, b, show the diploid male group taken from
testis and developing gonapophyses, respectively. These show
twenty-one chromosomes, and the largest of each group is the
unpaired x-chromosome. Diploid groups from the ovary and
soma of the female are shown in figures 51, h, and 51, 7, and
exhibit twenty-two chromosomes, which include two large
x-chromosomes.

In this anomalous nymph the testes contained typical sper-
matogonia, and, since these cells are direct descendants of the
primary germ cell which gave rise to the gonads, it was most
probably the chromatic make-up of this primordial cell which
caused the development of testes instead of ovaries. It is the
author’s opinion that the zygote from which the nymph arose
was female and that an abnormal mitosis gave rise to a pri-
mordial germ cell lacking one of the x-chromosomes. Having
the male diploid group of chromosomes, this cell proceeded to
form male gonads even though nourished and enclosed within
the body of a female. An attempt was made to analyze the
chromatic composition of the soma of this nymph, but on the
whole the cells were not favorable for accurate counts. No
metaphase plates were clear enough to count. Two good pro-
phases indicated the presence of twenty-two chromosomes in
body cells, but counts of such cells cannot be regarded as entirely
satisfactory. Nevertheless, since the soma was purely female,
we may feel fairly certain that the cells which composed it had
the female chromosome complex, just as the gonads which were
male had the male chromosome complex.

Meisenheimer (’09) and Kopeé (711) maintained that ovaries
transplanted into castrated males produced small ova because
there was not sufficient room for development, and did not
ascribe any influence to the metabolism of the soma in which
they were placed. Their view seems rather extreme, and in the
case of the testes of the nymph with female soma we find quite
an effect produced by the soma on the size of the gonads. In
Thelia, as in most insects, the testes develop much more rapidly

SEXUAL CHARACTERISTICS OF THELIA 599

in ontogeny than do the ovaries, so that when the imaginal form
is reached the testes have already reached their maximum size
and contain mature sperm, whereas the ovaries are still very
immature. This same discrepancy in rate of development may
also be noted in the various nymphal instars. Thus in the.
fourth instar of Thelia the testicular tubules average 0.2 mm. in
length, whereas the terminal chambers of the ovaries average
0.1 mm. in length. In the testis of the anomalous nymph: the
tubules were on the average 0.15 mm. long, making the testes
quite noticeably undersized. This would indicate that the
rapid growth of the testes in normal males is not due entirely to
the properties of the germ cells, but that the soma which pro-
vides the nutriment contributes in no small measure to the
growth of these cells. In this nymph with female soma the
testes did not get all the nourishment they required, but prob-
ably they did get more material for growth than is ordinarily
supplied to an ovary at this stage. The lack of sufficient nutri-
ment evinced itself in the presence of many cysts of degenerating
spermatogonia, such as may often be found in mature testes,
but not in normal testes of nymphs. Previously (Kornhauser,
14, p. 251) it was suggested that the degeneration of cysts of
spermatogonia, found so often just at the close of the period of
multiplication in testicular tubules, was due to the lack of suffi-
cient cytoplasmic materials, and the present case seems to bear
out this presumption.

The conclusions we may draw from the two significant individ-
uals described in the previous pages are: that the development
of the female soma of Thelia is not influenced by the presence of
gonads of the opposite sex; that the changes in parasitized male
Thelia are not due to the absence of testicular tissue, and that
the development of the gonad into ovary or testis is primarily
determined by the chromosomal composition of its cells, but the
growth of the testis is in part dependent upon the soma which
nourishes the germ cells.

600 SIDNEY I. KORNHAUSER

8. DISCUSSION

In the arthropods transformation of various male sexual char-
acteristics to corresponding female characteristics has been
known and adequately described only in crustaceans infested
with parasitic cirripeds. Giard in a series of papers and Smith
more recently have identified themselves enduringly with this
problem. Smith not only described the external changes of the
hosts, but went into an analysis of the effects of the parasites
on the hosts’ metabolism, using both histological and chemical
methods of attacking the problem. He also opened up the field
for a discussion as to the manner in which female characteristics
are made to appear in the male.

Early in the study of the effects of parasites on sexual char-
acteristics, the term ‘castration parasitaire’ became associated
with this phenomenon, chiefly through the work of Giard. This
term leads one immediately to look upon the transformation as
being due to castration effected by parasites. This association
of castration with changes in sexual characteristics may be
attributed to our knowledge of the important influence exerted
by the gonads of vertebrates, and especially mammals, upon the
development of the soma. Whether effected experimentally or
by parasites, castration produces striking results in the higher
animals. It is therefore natural, since such parasites as the
rhizocephalans or in the case of Thelia, Aphelopus, generally
cause the gonads of the host to undergo reduction or complete
obliteration, that we should associate the somatic changes
incurred with the loss of the testes or ovaries. Still at the
present time there is no evidence in the work on arthropodsto
support such an inference. It is therefore urged that the term
‘castration parasitaire’ be set aside in considering the alterations
induced by parasites on the sexual characteristics of arthropods
There are several lines of evidence drawn from our knowledge of
the crustacea and insects to support this contention.

That the gonads of insects do not produce hormones which
shape the development of the secondary sexual characteristics
is shown by the numerous cases of successful experimental
castration of immature individuals, followed by development

SEXUAL CHARACTERISTICS OF THELIA 601

into adults normal in every way—in color, form, and psychie
traits. Castration was often followed by the implantation of
gonads of the opposite sex and still no effect was induced. Like-
wise, the injection of gonad extracts into castrated individuals
proved ineffective. The most noteworthy of these experiments
has been reviewed in some detail in part 2. As intimated, these
results are contrary to what is known in vertebrates.

The literature dealing with castration and transplantation of
gonads in vertebrates is so voluminous that no attempt will be
made here to review the entire subject and only the general
principles founded upon these researches will be cited. Several
lines of work on vertebrates indicate that the interstitial cells of
the gonads secrete important hormones which influence the
development of the soma. Castration itself practiced on im-
mature animals causes the retention of juvenal characteristics,
and a partial or complete lack of the appearance of morpholog-
ical and psychic sexual characteristics which normally appear at
sexual maturity. The long bones of the body continue to grow
and reach a length greater than the average in both sexes; in the
male, an unusual amount of fat may be laid on, the external
genitalia may be greatly reduced in size, and the male tempera-
ment suppressed. However, some characteristics of the sex
operated upon may be retained in part, such as the shape of
the pelvic bones. Evidence that the interstitial cells and not
the growing germ cells themselves play the important part in
hormone production is presented by the following phenomena.
In cryptorchism, or in testes subjected to ultraviolet rays, active
germ cells may be entirely absent while the interstitial cells
remain intact. In such animals the secondary sexual character-
istics develop normally and it is believed that the activity of the
interstitial cells accounts for this. Tandler and Gross (13)
have presented this subject in a most illuminating manner.
That hormones play an important part in the early embryonic
development of mammals has been shown in the case of two
sexed twins of cattle having anastomoses between the circula-
tory vessels of the two individuals (Lillie, ’17; Chapin ’17).
The more rapidly developing male by its hormones affects the

602 SIDNEY I. KORNHAUSER

development of the external genitalia, the oviduct and uterus,
and the structure of the gonad of the female twin which becomes
a free martin. Not only the interstitial cells of the testis pro-
duce hormones, but it is known that the ovary likewise makes
important secretions. Steinach (’12) feminized rats by cas-
trating young males and implanting ovaries. These animals
later displayed many female characteristics of skeleton and hair
pattern; they developed mammary glands and psychic activities
of the female, while the external genitalia remained small and
undeveloped. Im birds Goodale (’16) has conclusively shown
that the absence of the ovary allows the male secondary sexual
characteristics of the species to be fully developed in the female.
Goodale believes that the ovarian hormone acts as an inhibitor,
which normally prevents the appearance of those characteristics
that we associate with maleness.

Since such intimate association of somatic characteristics with
the presence or absence of gonads exists in the vertebrates, it is
not surprising that the results on the castration of insects pre-
senting absolute independence of somatic characteristics should
be doubted and criticised. KKammerer (’12) would lead us to
believe that those who followed in Oudemans’ (’98) footsteps
and extended the work on the castration of insects, investigators
who improved the methods and succeeded in the implantation
of gonads of the opposite sex, had interpreted their results blindly,
merely accepting the belief in the fixity of the somatic sexual
characteristics of insects as propounded by Oudemans, and
extending his observations and ideas without question as to
their correctness. Certainly, Kammerer’s attitude seems rather
extraordinary, for the researches of Meisenheimer (’09) and
Kopeé (11, ’13, a, ’13, 6) seem strictly scientific, well planned,
accurate, and convincing. But aside from these experiments,
gynandromorphs, such as are occasionally met with in the ar-
thropods, are convincing evidence of the independence of somatic
development in this group of animals. One half of the body
may be perfect in its male characteristics and the other half
female, and within the abdomen there may be present either
testes or ovaries or gonads of both sexes. In fact, all possible

SEXUAL CHARACTERISTICS OF THELIA 603

internal combinations may be associated with a bisexual soma.
Likewise, such an anomalous individual as the Thelia nymph
shown in figures 52 to 54 is most convincing proof that the testes
produce no hormone which influences somatic development.
In this individual, a nymph of the fourth instar with soma of a
female and gonads of a male, the female secondary sexual
characteristics were perfectly formed. Had the same conditions
existed in this Thelia nymph as exists in developing vertebrates
(Lillie, 17), the testes even though immature would have greatly
modified the female somatic characteristics. It seems, there-
fore, most probable that, in the insects and probably in the
arthropods in general, the development of the secondary sexual
characteristics is independent of the gonads. If we accept this,
it follows that the effects of parasitism upon the sexual charac-
teristics cannot be traced directly to the destruction of the
gonads.

A second objection to the term ‘castration parasitaire’ as a
descriptive term for the effect of parasites on arthropod hosts is
the fact that often the gonads are not wholly destroyed. Smith
(10) showed that if the parasitic barnacle were removed from
the host, the germinal epithelium might regenerate and produce
germ cells which however in the male might grow into oécytes,
instead of spermatocytes. In bees parasitized by Stylopidae the
gonads of the male may still be functional, but in all cases the
ovaries are described as being very minute. In Thelia generally
the presence of large Aphelopus larvae had as its accompanying
condition the entire absence of gonads. Nevertheless, occasion-
ally in males remnants of testes were found, varying from minute
clumps of germ cells imbedded in a mass of fatty tissue, to a
full-sized testis, such as was shown in figure 50. In parasitized
females minute ovaries were at times discovered. The largest
odcytes found in such ovaries were 0.4 mm. in length. That a
full-sized testis appearing normal in every way and filled with
sperm was found in a parasitized male showing considerable
somatic alteration is, I believe, convincing proof that the somatic
changes found in parasitized male Thelia are not due to the
absence of testicular tissue.

604 SIDNEY I. KORNHAUSER

Since somatic changes may occur in animals not castrated by
the parasites, we must look for these changes as being due to
other factors than the lack of gonads. Smith, in his Studies on
the Experimental Analysis of Sex, always denied the production
of hormones by arthropod gonads, and still, in his discussion of
the effect of Stylops on Andraena (’14) he concluded that the
reduction of the pollen-gathering apparatus of females and the
appearance of male coloration of the face, as described by both
Perez (’86) and himself, were accounted for by the reduction of
the ovaries through parasitism; just as in the birds the absence
of the ovary or the presence of a small non-functional ovary has
as its consequence the appearance of the male secondary sexual
characteristics. This explanation of stylopization is open to
several objections. It omits consideration of the results of
experimental castration of female insects. Likewise, in the
Hymenoptera, to which order of insects Andraena belongs,
gynandromorphs have been frequently found (Wheeler, ’03):
the female portions of the soma appearing perfectly developed,
although dissection of several individuals showed the absence of
ovarial tissue. Likewise, in social bees, if the lack of develop-
ment of ovaries would bring about the appearance of male char-
acteristics or the reduction of female characteristics, the workers
which are sterile females with small undeveloped ovaries might
be expected to have small scopae and present various character-
istics of the male. Smith’s standpoint, if we are to interpret it
from his other papers on sex, might be expressed as follows:
the absence of the ovary brings about these changes not because
an ovarial hormone has been eliminated, but because there is no
ovary present which makes a demand upon the vegetative tissues
of the organism. This lack of demand for food material to be
stored by the ovary brings about a change of metabolism, and
also has as its consequence the somatic alteration already re-
ferred to. Whether such a demand is normally made by the
gonads of arthropods upon the soma will be considered later.

All evidence points toward the conclusion that the soma of
insects is in no way dependent on the gonad in its development.
If this be true, we must dismiss in our analysis of the changes in

SEXUAL CHARACTERISTICS OF THELIA 605

parasitized Thelia all explanations based upon the experimental
evidence gained from operations on vertebrates. As was pre-
viously stated, vertebrates and especially mammals castrated as
immature individuals often retain juvenal characteristics and
continue to grow beyond the normal size, and males may store
up a large amount of adipose tissue. Throughout the descrip-
tion of the changes incurred by adult parasitized Thelia, the
normal nymphal characteristics have been presented for com-
parison, and it may safely be stated that in no case were the
alterations of males referable to the retention of juvenal charac-
teristics. Only in one case was arrest of development encoun-
tered, and that in the genitalia of the fifth instars which contained
large Aphelopus larvae. These individuals had genitalia resem-
bling those of the fourth instar. It is unfortunate that such
individuals are always killed by the emergence of the eruciform
larvae, for it would be extremely intersting to know, could they
become adults, whether they would possess genitalia similar to
those of the fifth instar instead of reduced adult genitalia, such
as are met with in all infected adults.

It was seen-that parasitized male Thelia grew to a greater size
than normal males, but this can hardly be ascribed to continued
growth similar to that occurring in castrated mammals. Such
an explanation would not account for the fact that parasitized
female Thelia are not larger than normal females and parasitized
females lack gonads even more frequently than do infected
males. Likewise, no increase in size of experimentally castrated
insects has ever been reported, and it is quite safe to say that
parasitized male Thelia are larger than normal males, not be-
cause the testes may be degenerate, but because the parasites
exert some positive effect upon the soma which in part develops
certain female extragenital secondary sexual characteristics of
which greater size is one. This increase in size of parasitized
males was noted in the pronotum, wings, head, acrotergites,
hind legs, and abdomen. Not merely, therefore, is the region
containing the parasites, the abdomen, enlarged, but the most
remote portions of the body respond. As far as could be ascer-
tained, this is the first case of this sort found in parasitized

606 SIDNEY I. KORNHAUSER

insects. One might offer a simple explanation, that the ab-
domen increases because the parasites make this a mechanical
necessity and this change induces an adaptive increase in other
parts of the body: the wings must be larger to carry the increased
bulk of the abdomen through the air; the thorax must be larger
to contain wing muscles of increased strength. But there are
serious obstacles to such an explanation. Very often the para-
sites do not distend the abdomen of the host at its final molt
and still the sclerites are so enlarged that later these larvae which
increase tremendously can be accommodated. The influence
is, therefore, not by mechanical stimulus. In many cases of
parasitism in insects it has been observed that, although the
abdomen may be greatly enlarged, the other parts of the body
either remain normal in size or are even reduced. Wheeler (10)
cites several cases in which ants with enlarged abdomens con-
taining mematodes possess heads of normal size and wings greatly
reduced. Perez (’86) states that in stylopized Andraenae the
heads of the bees in both sexes are smaller than normal. It is, .
therefore, rather remarkable that male Thelia increase in all
parts of the body when infested by Aphelopus. We would
naturally expect that the drain upon the nutritive material of an
immature host would result in a starved undersized adult.

In Thelia, although the infected males are larger than normal,
still, in this sex as well as among the females, the parasites induce
certain changes due to their demands upon the host. The
gonads which are not essential to the life of the host are the
first tissues to suffer and the material which would go in to the
formation of countless numbers of spermatozoa or be used in
the growth of ova doubtlessly affords one of the chief sources of
nutriment for the parasites. The chitin, too, is often thinner
than in normal individuals, this being noted on the pronota of
females as well as males. The punctures may be smaller and
shallower and the amount of melanic pigment restricted to the
depressions of the punctures. Lack of materials demanded by
the growing larvae might explain the absence of yellow pigment
in the vittae of parasitized males but it would not explain the
assumption of the female pigmentation nor would it account for

SEXUAL CHARACTERISTICS OF THELIA 607

any of the qualitative changes previously described. Female
secondary sexual characteristics are stimulated to development
in male Thelia containing Aphelopus larvae, and these charac-
teristics are not only those of size, coloration, and pattern, but
metabolic characteristics also. The change of male metabolism,
I believe, is necessitated by the demand for food by the parasites.

The metabolic differences between male and female tissues are
known in many groups of the animal kingdom. These differ-
ences are associated primarily with gamete production, the act
of fertilization, and the rearing of the young. ‘The female must
not only provide material for her somatic needs, but she must
often store up a large quantity of food in the formation of ova or
furnish it to the growing embryo. The male produces micro-
gametes generally in immense numbers and actively seeks the
female, and, with the fertilization of the eggs, often his function
is performed. The sexes of Thelia exhibit many differences
associated with gamete production. The female develops less
rapidly than the male, but the resulting adult is of greater size.
Upon reaching the imaginal stage only minute ova are present
in the gonads, but the tissues have the capacity for the storage
of reserve food, and this storage continues for several weeks,
during which time a large amount of yolk and fat is accumulated
due to steady feeding and low oxidation of the ingested food.
Male Thelia develop more rapidly and appear in July, generally
before any females are to be seen. They are already sexually
mature at molting and are much more active than females,
flying or jumping more easily upon being disturbed. This
activity is doubtlessly associated with seeking the females in
mating. The males are also less long-lived than the females,
disappearing in late summer, whereas the females are still abun-
dant in the fall. The more rapid development of the male, the
rapid division of the spermatogonia, the short growth period of
the spermatocytes, early sexual maturity, great activity, and
earlier death indicate high'’metabolism and high oxidation. As
the season progresses, the gonads of the male become smaller,
many cysts of germ cells undergo degeneration, and the stored
adipose tissue surrounding the testes decreases in amount.

608 SIDNEY I. KORNHAUSER

Thus, as adults, the metabolism of the two sexes stands in con-
trast: the female is continually storing food material in the for-
mation of ova, the male is using up available material faster
than it is being supplied by intake of food. With this difference
in metabolism, it is clear that, while the adult female can supply
the growing Aphelopus larvae which may happen to be present
in her abdomen, the male is not so well equipped by nature.
The male tissues must change their metabolic level if the host
itself is to survive and if the parasites are to obtain sufficient
food for their growth. This changed metabolism expresses
itself in various ways. As was stated on page 546, parasitized
adult males make their appearance in the field along with the
females, a week or two later than normal males, indicating that
their development has been retarded. They are less active
than normal males, sitting immovable on the branches and feed-
ing continuously. They become animals of high storage capac-
ity, as is indicated by the increase in adipose tissue. The portion
of the abdomen normally occupied by the testes is filled with a
mass of greenish or yellowish fat, often containing remnants of
testicular tissue. The parasites themselves are literally im-
bedded in fat, and a comparison of histological cross-sections of
abdomens of parasitized and normal males shows the great
increase of adipose tissue in the former. I am greatly indebted
to Dr. Oscar Riddle for a quantitative chemical analysis of the
alecohol-soluble substances in normal and parasitized Thelia.
Lots of ten individuals minus pronotum and head were preserved
in alcohol for analysis. Two lots of parasitized males and one
lot of normal males were analyzed. The parasitized males
showed on the average an increase of 47 per cent of lipoids over
the normal males. While the samples analyzed were too small
and too few in number to be entirely satisfactory, yet they
indicate the true state of affairs.

Another change brought about by the parasites is seen in the
production of melanin, which is reduced over the entire body of
males with the exception of the punctures on the vitta. These
are pigmented in parasitized, but not in normal males. This
general reduction of melanin may be due to a decrease in the

SEXUAL CHARACTERISTICS OF THELIA 609

amount of either the base of melanin, probably tyrosin, or the
oxidase, tyrosinase, or possibly to a decrease in both. It has
been found that the amount of the ferment present has a great
influence on the melanin produced (Kastle, ’09), increasing
pigment production up to a certain concentration of the tyro-
sinase, and then inhibiting the reaction. It is also known that
even weak acids prevent melanic pigment production, and the
parasites of Thelia probably bring about a condition of acidosis
in the host, which has as its most important result fatty infiltra-
tion of tissues, and which also may bring about inhibition of
melanin formation.

Recognizing that the metabolic level of parasitized male
Thelia has been altered from the normal, are we to refer the
changed morphological somatic characteristics as being due
directly to this change? The answer to this question rests
entirely on our conception of the origin and meaning of sex itself.
If we believe that the underlying difference between the sexes is
one of metabolism and not one of gamete production, then high
metabolism has as its consequence maleness and sperm produc-
tion, while low metabolism has femaleness and egg production
as a consequence. The opposite view is that primarily the
male is a sperm producer, the female an egg producer. The
cells of an individual, somatic as well as germinal, are of one
sex, either male or female. The metabolic level of the organism
is merely one of the many expressions of sex, but a very impor-
tant one in a consideration of gamete production and the nour-
ishment of the offspring. This difference in metabolic pitch
certainly is more important than other secondary sexual differ-
ences, as size, pattern, coloration, or psychic characters, and
probably preceded these in the phylogenetic development of
sex. It has probably become more divergent in the two sexes
as the gametes have been modified in evolution, as active seeking
of the female by the male and internal fecundation has become
established in higher forms and the development of the embryo
has become dependent upon the female. Just as the form of
the sperm or egg must be molded by the genes in the cells of
the individual which produces them, likewise, it is not improb-

610 SIDNEY I. KORNHAUSER

able that the metabolic level also depends on genetic factors
probably present in the chromatin.

The excellent researches of Smith on sacculinized crabs indi-
cated that the metabolic level of infected male crabs was lowered.
Smith would ascribe the appearance of female secondary sexual
characteristics in these crabs as being referable to the underlying
causes of the changed metabolism. Lowered metabolism evinced
itself in various ways. It was first detected by the presence of
lipochromes characteristic of the blood of the female (Smith, ’11,
713, and Robson, 711). Then it was shown that the percentage
of fat in the blood of infected males increased, approaching that
characteristic of females producing ova. Histological sections
of the liver (Robson, ’11) showed a great increase of fat globules
in the cells of parasitized males, bringing about a condition sim-
ilar to that of normal females. Smith (13) made quantitative
analyses of the livers of normal and infected individuals and
showed that fat production was stimulated in parasitized male
crabs and that glycogen production was depressed bringing
about a metabolism characteristic of the female. Smith con-
tended that the Sacculina roots act as does a normal ovary.
They make a demand on the tissues of the host and in the male
alter the substances carried by the blood and body fluid. This
demand has two results, it brings about the increase of fat mole-
cules and it stimulates the production of certain female second-
ary sexual characteristics. In Smith’s theoretical considera-
tions of the problem he explained the demand made by the
Sacculina roots and the response of the tissues upon the Ehrlich
theory of immunity reactions. The Sacculina roots in para-
sitized males acting upon protein molecules in the crab’s blood
stimulate the production of fat links, which travel through the
blood to the host’s liver, where each receives a fat molecule
which it transports back through the blood to the Sacculina fat
chains. Here the fat molecule is given up and the fat link again
travels toward the liver. The constant increase of fat links
stimulates an increase of fat molecules in the liver of the host.
Smith clearly states that not the increased fat itself in the blood
stimulates the appearance of the female secondary sexual char-

to ie ied

SEXUAL CHARACTERISTICS OF THELIA 611

acteristics, but the deep-seated changes or underlying causes
which involve increase of fat stimulate the appearance of the
secondary sexual characteristics. Just how the Sacculina roots
or normal ovary bring about the production of changes in the
blood is not gone into, but certainly the whole differs little from
the hormone theory which Smith himself opposed vigorously.
In Smith’s explanation, the Sacculina roots or ovary act upon
molecules in the blood and the changed blood acts upon the
somatic tissues. In the hormone theory, the gonad produces
substances within its own cells, and these substances trans-
ported by the blood which acts merely as a carrier affect the
somatic cells to which they are brought. The most questionable
feature of Smith’s theoretical considerations is the demand of
the ovary upon the soma, followed by a response on the part of
the somatic cells. In the arthropods we are wholly without
evidence that there is any demand coming from the ovary which
can alter the metabolism of the vegetative cells or stimulate the
production of certain somatic characteristics.

Such an individual as that described by Wenke (’06), whose
work was reviewed in part 2, offers difficulties to a belief in
ovarian influence exerting an effect upon the developing soma.
The individual described was a perfect lateral gynandromorph
with the male somatic characteristics perfectly developed in one
half, although the only gonad present was one well-developed
ovary containing eggs almost mature. Such combinations of
male and female soma are not altogether uncommon in the
insects and higher crustacea and indicate that in the arthropods
the genes in the cells producing the somatic structures are not
influenced either directly or indirectly by the gonads.

Another line of convincing evidence on the question of the
independence of somatic tissues in insects is that brought for-
ward by the works of Steche (12) and Geyer (’13), as previously
reviewed on pages 536 to 537. These investigators proved that
the haemolymph of insects was unlike in the two sexes, that
sometimes color differences were evident in phytophagous species
and that there were always protein differences demonstrable by
precipitation tests. Castration failed to alter the characteristic

JOURNAL OF MORPHOLOGY, VOL. 82, No. 3

612 SIDNEY I. KORNHAUSER

color or composition of the haemolymph. Likewise, the im-
plantation of testes into castrated female caterpillars’ or the
implantation of ovaries into castrated males failed to alter the
characteristics of the haemolymph in either of the sexes operated
upon. The conclusion of both authors was that the somatic
cells which produced the haemolymph, referring principally to
the cells of the digestive tube, were either male or female, and
by their activity the haemolymph was also male or female in its
characteristics. The function of these digestive cells was clearly
shown to be independent of the ovary or testis. Surely, if the
ovary were capable of creating a demand upon the soma, we
would expect that an ovary transplanted into a castrated male
would so affect the cells of the digestive tube that the necessary
materials for provisioning the ova with yolk materials and pig-
ment would be supplied. But as a matter of fact, implanted
ovaries in male somas fail to get their necessary supplies and they
fail also to influence those somatic cells which, instead of being
nutritive in function, have for their mission the production of
the more permanent structures which include many of the sec-
ondary sexual characteristics. In the work of Kopeé (’11) we
see a strengthening of the conclusions of Geyer and Steche
upholding the physiological difference between the somatic
cells of the two sexes. It is well known that ovaries trans-
planted into castrated males never attain normal size, being
generally a third or a fifth as long and containing but a fourth
or fifth the normal number of ova which are much undersized.
A further observation of Kopeé (711) is interesting, namely,
that in Gastropacha quercifolia such ova are yellow instead of
green and have fewer and smaller yolk granules than eggs of
normal females. This latter condition, I believe, may be traced
to the fact that the cells which provided the haemolymph were
male instead of female and thus the failure to supply the green
pigment and necessary yolk materials. Kopeé (711) and Meisen-
heimer (’09) would refer the smallness of the implanted ovaries
to lack of space for development in the smaller male abdomen,
but if the ovary were capable of making a demand as postulated
by Smith we would expect that those male larvae in which’

SEXUAL CHARACTERISTICS OF THELIA 613

ovaries had been implanted would as adults possess abdomens
capable of accommodating the mature ovary. In sacculinized
male crabs and in male Thelia parasitized by Aphelopus the
abdomen is enlarged, but there is little evidence to support the
idea that the enlargement is called forth by an influence similar
to that supposedly exerted by an ovary. The physiological
activity of the somatic cells, although correlated with gamete
production, does not, in insects, seem to be governed by the
gonads, but is probably to be referred to the genetic constitution
of the cells upon which the gametes themselves may be partially
dependent for their normal development. ‘This is illustrated by
the anomalous Thelia nymph described in part 7. This individ-
ual, a nymph of fourth instar, had perfect female soma and con-
tained two testes which were undersized for this stage and con-
tained many degenerating cysts. If we compare the normal
development of the gonads in the two sexes, we see that the testes
develop rapidly and are full sized at the final molt, but the ova-
ries progress very slowly and are smaller than the testes during
the entire nymphal life. In the adult female they grow to rela-
tively enormous size, but in the fourth instar they are consid-
erably smaller than the testes. Thus, in the anomalous nymph,
I believe, we must ascribe the smallness of the testes to the
fact that they were provided with only as much nutriment as
a female soma normally provides to its contained ovaries. The
degenerating cysts indicate that cell division proceeded faster
than materials for growth were supplied. Here surely the testes
failed to influence the soma to change its normal metabolism
or to produce any male characteristics. If, then, the develop-
ing soma of arthropods is entirely independent of any influ-
ence emanating from the gonads, we cannot explain the modi-
fications produced by parasites on the assumption that the
parasites act as do the gonads of one sex or the other.

The sexual characteristics of insects must, I believe, depend
entirely upon the chromatic makeup of the cells composing the
individual. Of all the insects, no groups show a more uni-
versal visible chromatic difference between the cells of the two
sexes than do the Hemiptera and Homoptera. It is not my

614 SIDNEY I. KORNHAUSER

purpose to review here the many papers on hemipteran and
homopteran chromosomes, so it will suffice to say that, in spite
of the many combinations found, the presence of one x-chromo-
some in the cells of the male and two x-chromosomes in the cells
of the female is the essential difference. It is true that each
x-chromosome may be represented by a group of separate chro-
matic elements and that the x-chromosome may be accompanied
in the cells of the male by a y-chromosome, but these are rather
subordinate details. We also know that this sexual difference
extends not only to the cells of the gonads, but to the somatic
cells as well (Morrill, 710, Hoy, ’16). The writer has definitely
demonstrated that this also holds true for Thelia. The growing
gonapophyses of the fifth nymphal instar were slipped out of
their chitinous coverings and mounted in acetic carmine. Many
clear and handsome metaphase plates were studied and drawn
with the aid of a camera lucida. The female somatic cells
showed regularly twenty-two chromosomes including two large
x-chromosomes, whereas the male cells exhibited twenty-one
chromosomes, of which the largest one was the x-chromosome.
Two of these cells are shown in figure 51, 6b and 7. Thus the
sex of the soma and gonads of insects is determined in the zygote,
and there is at no time an indifferent sex gland which may be
molded into testis or ovary by the soma in which it develops.
This idea of the indifferent gonad was put forth by Doncaster
(14), who, after reviewing the work on the arthropods describ-
ing the physiological differences between males and females,
states that perhaps the physiological differences are the primary
sexual differences and that the ‘primitive gonad’ develops into
an ovary on one hand or into a testis on the other in consequence
of this. The peculiar nymph described on pages 594 to 598 of
this paper illustrates nicely that the chromosomes determine the
character of both the soma and the germ plasm. ‘This individ-
ual with the soma of a female had male gonads and a cytological
examination showed that these gonads had the male comple-
ment of chromosomes. In some way the primitive germ cells
which formed the gonads failed to receive two x-chromosomes,
and so testes developed instead of ovaries. That the soma was

SEXUAL CHARACTERISTICS OF THELIA 615

female did not prevent the development of testes, nor did the
presence of testes in any way interfere with the development of
the female somatic characteristics. In insects the chromatin of
the germ cells represents a mechanism for shaping the gametes,
and this mechanism is extremely difficult to upset, and cannot
be interfered with by metabolic changes in the soma.

This is nicely demonstrated in parasitized male Thelia. Many
careful microscopic examinations of some forty testes and rem-
nants of testes from parasitized males have failed to reveal any
cells taking on the characteristics of odcytes. That the metab-
olism of these males was altered by the growing Aphelopus
larvae cannot be in the least questioned. They were no longer
animals of high oxidizing powers, but stored fat in large quan-
tities and in many other ways approached female metabolism.
Still spermatocytes failed to grow beyond their normal size, and
often, when conditions were not too adverse, divided into sper-
matids which differentiated into mature spermatozoa. Frequent-
ly, fatty infiltration caused the spermatocytes to degenerate,
and one might infer that conditions in parasitized individuals
always made continued growth of the germ cells impossible.
This, however, cannot be the case, for in parasitized females
odcytes measuring 390u by 46. have been found, and these
had grown in the adult females from cells no larger than 56u
by 36, which is the maximum size of odcytes at the final
molt. Therefore, storage is possible in the germ cells of para-
sitized adults. But spermatocytes are not induced by changed
conditions to approach in character the germ cells of the female,
although the spermatogonia may have been subjected to the
changed environment for many generations. That the gamete-
forming mechanism is rather stable is shown also by Gold-
schmidt’s intersexual moths (’17b). He states that the last
organ to be changed to that of the opposite sex is the gonad.
Likewise in vertebrates, where the development of the sexual
characteristics is largely dependent on hormones, the gameto-
cytes are not easily influenced. Thus in the gonads of the
freemartins, although the general structure of the ovary is pro-
foundly changed by the male hormones, still the germ cells fail

616 SIDNEY I. KORNHAUSER

to develop into gametes of the opposite sex (Chapin, 717). This
stability exhibited by the germ cells is probably due to some
fundamental difference early established in the evolution of
sex between spermatogonia and oégonia or between spermato-
cytes and odcytes.

Were it permissible to speculate upon the origin of the sexual
differences, the writer would consider the production of two
sorts of gametes as the first and fundamental step. It would
be assumed that originally either in the protozoans or lowest
metazoa the individuals were isogametic and that zygotes were
formed by the union of two similar, rather large, and not. ex-
tremely active gametes, as still exist among many green algae
and certain protozoa. If by a mutation one individual pro-
duced many small, highly motile gametes which sought the
larger less active ones in conjuguation, the number of zygotes
might be increased and the species thereby benefited. This
original mutation which established microgamete production
would require a more rapid cell division and a shorter growth
period in gametogenesis. The gene bringing this about would
be represented in the constitution of all the cells of the individ-
ual, somatic as well as germinal, and might influence in many
ways the form and physiology of the whole individual. Recently
Morgan (717), in speaking of the manifold effects of each gene,
citing an example from Drosophila, says that, ‘‘whatever it is in
the germ plasm that produces white eyes, it also produces these
other modifications as well and modifies not only such ‘super-
ficial’ things as color, but also such ‘fundamental’ things as
productivity and viability.” A gene in one chromosome may
influence or inhibit the genes located in other chromosomes and
change in many ways the characteristics or constitution of the
mutant. But even in the origin of heterogametic forms should
the soma of the individual producing the microgametes (the
male) remain similar in form to the macrogamete producer
(the female), as it does in many marine forms, including even
annelids and echinoderms, still there would remain that genetic
difference in all the cells of the individual, and any future muta-
tion would arise either in the presence or the absence of the
gene affecting the fundamental difference between the sexes.

SEXUAL CHARACTERISTICS OF THELIA 617

It must be assumed that the gametes themselves have under-
gone an evolution and that many genetic changes have occurred
to modify and specialize the form of the mature germ cell. The
multitudinous forms and the intricate apparatus which sper-
matozoa exhibit lend credence to this idea. Likewise, the ovum
must have been changed from its original state. One of the first
processes, one which has now become universal in animal ova,
must have been the production of cells of unequal size in matu-
ration. The polar cells represent abortive ova, whereas the one
functional ovum receives the nutritive supply originally intended
for four cells of equal size. Thus fewer and larger ova would
result from this change. In many animals the period of growth
was prolonged so that larger ova resulted, finally culminating in
the immense egg cells of birds. Undoubtedly mutations affect-
ing the storage power of the somatic cells as well as the germ
cells have aided in this specialization and caused the female to
diverge in its physiological constitution from the male. This
difference is still demonstrable between the male and female
somas of mammals. Although the ovum no longer stores yolk,
still the body tissues must provide food for the growing embryo.
An economy is perhaps hereby effected in that fertilization must
first be insured before prolonged storage of food materials can
take place.

With the specialization of the gametes changes in the acces-
sory tissues of the gonads, in their ducts and glands, must have
becomes established. These tissues minister to the needs of the
gametes and serve to carry them to the exterior. In land ani-
mals and also in some aquatic animals copulation and internal
fertilization made the development of external genital apparatus
essential and the establishment of instinct of sex necessary.
Last of all, we are to look upon the extragenital sexual char-
acteristics, including all forms of ornamentation, as coming into
existence.

According to the views expressed above, sex has evolved
through a series of genetic changes accompanying the evolution
of the various groups of the animal kingdom. It is not assumed
to have sprung up independently whenever a difference in the

618 SIDNEY I. KORNHAUSER

rate of metabolism of the germ cells of individuals in the various
groups appeared (Riddle, ’17). Metabolic differences are de-
monstrable and measurable between the sexes of highly special-
ized forms which store yolk in large ova or provide nutriment to
the zygote. This metabolic level is assumed to be an expression
of sex rather than a causal factor.

Genetic changes are most probably to be sought in changes
in the chromatin. In the evolution of sex, genes located in
various chromosomes have undoubtedly played a part. In
nematodes, insects, spiders, and some mammals visible chromatic
differences have led to the belief that the unpaired element
of the male cells, the x-chromosome, is intimately connected
with sexual differentiation. Let us inquire what is known
about the x-chromosome and the location of other genes affect-
ing principally sexual characteristics. In spermatocytes the
x-chromosome would seem entirely different from ordinary
chromosomes. It does not form a typical leptotene thread, and
even if a y-chromosome is present it fails to form a double syn-
detic thread with this element. Should a fairly long growth
period follow syndesis and the autosomes become very indistinct,
the x-chromosome remains a compact deeply staining mass.
This is true for Thelia and is shown in figure 51, c and d. Still,
the x-chromosome is not an inert mass of chromatic material
differing in character from that of the autosomes. When paired
as in the odcytes, they behave exactly like any of the other chro-
mosomes. They form leptotene threads which conjugate. In
Drosophila (Morgan and Bridges, ’16) linkage and crossing
over in the paired x-chromosomes has been shown to occur just
as in the autosomes. By investigations of linkage many genes
which bring about sex-linked characteristics have been located
in the x-chromosome and these genes seem to have a definite
linear arrangement. Furthermore, Bridges’ (’16) remarkable
observations on ‘non-disjunction’ have definitely demonstrated
that one x-chromosome is absolutely necessary in the formation
of the male and that two and only two are essential in the pro-
duction of a female. From the behavior of the x-chromosome
in spermatogenesis one might infer that this element were in-

SEXUAL CHARACTERISTICS OF THELIA 619

active, an unessential in the male, and that males might be
produced lacking this element, yet no zygote in the ‘non-dis-
junction’ experiments formed by the union of an ovum minus
an x-chromosome and a sperm also without this chromosome
developed. Likewise, ova containing two x-chromosomes never
developed if fertilized by spermatozoa containing an x-element.
Thus females with three x-chromosomes are not formed. These
same experiments of Bridges also demonstrated that the y-chro-
mosome is without influence in sex production for xxy females
and xyy males are in no way different from normal individuals
except in the types of gametes produced. No genes have been
located in the y-chromosome, yet from its behavior in sper-
matogenesis it seems to have some affinity for the x-chromo-
some, and it is not improbable that the y-chromosome represents
an altered x-chromosome rendered inactive through some change
in its makeup. Its affinity for the x-chromosome is shown in
the formation of an unequal xy tetrad in the first spermatocyte
division, as in the Coleoptera, or an unequal xy diad in the second
spermatocyte, as in the Heteroptera. In Enchenopa binotata
(Kornhauser, 714) these two chromosomes unite end to end in
syndesis and form a tetrad composed of two elements similar in
size. ‘Thus, I believe, we are to look upon the y-chromosome as
having originally been the partner and homologue of the x-chro-
mosome, but that a change, perhaps the loss of a single gene,
made it inactive as a carrier of genes. Inactive and therefore
unimportant, it might undergo many chance variations or losses
which might culminate in the final disappearance of the y-element,
a condition not at all uncommon in the insects. Bridges (’17)
has recently shown that a chromosome may become deficient
as a bearer of genes. A race of flies was produced in which the
x-chromosome was abnormal in that a particular ‘measurable
section of genes’ was either inactivated or lost. This experi-
ment further demonstrates that ‘deficient’ x-chromosomes pro-
duce normal sex ratios, and he concludes that the determiner of
sex is not the ‘x-as-a whole,’ but that in some definite part or
parts of the x there are specific sex-differentiators.

620 SIDNEY I. KORNHAUSER

If the production of an animal with but one active x-chromo-
some had as its consequence, perhaps through further muta-
tions in the one-x-individual, the formation of a microgamete-
producing individual, the beginning of sexual differentiation
would be established. The inactivation of the x-chromosome or
the production of the y-chromosome may merely have brought
about a less stable condition in the cells of the individual possess-
ing the inactive x-element, and offered a basis for further muta-
tions leading to the differentiation of two types of individuals,
one producing microgametes and the other macrogametes.
The further divergence of the sexes in the form of the gametes
and soma would be partially dependent upon this primary
difference in stability or constitution, for new genes would
necessarily arise and be expressed either in cells with one or
with two functional x-chromosomes. That there is something
vital in the x-chromosome upon which development depends
was shown nicely by Bridges, who proved that zygotes without
an x-chromosome or with more than two x-chromosomes failed
to develop. It is therefore not unlikely that cells having two
sets of x-genes should be somewhat differently constituted than
those possessing but a single set. ‘The male often shows a much
greater tendency toward variation than does the female, and
this most probably is referable to a greater conservativeness in
the composition of the female cells. In the family of the Mem-
bracidae, with which the author is fairly familiar, the variability
of the males is much greater than that of the females. In Thelia
the form of the pronotal horn and the extent of the vitta, as
illustrated in figure 7, page 556, exhibit far less constancy than
in the female. Fowler (’08) has described and figured many
types of the membracid Umbonia orozimba which in the male
show many types of variations in color, form, and size. Some
greatly resemble the females, but the series extends to forms so
unlike the female that they were originally placed in another
genus, Physoplia, by Amyot and Serville (’43). As stated on
page 589, the membracids are characterized by conservativeness
of the genital appendages. Lately the author has studied the
genitalia of both sexes of forms from the various subfamilies

SEXUAL CHARACTERISTICS OF THELIA 621

of the Membracidae, and in more detail the available genera of
the tribe Telamonini, to which Thelia belongs. The outstand-
ing result is that, whereas the female genitalia show a remark-
able conservativeness to a general type, even in the most widely
separated forms, the male genitalia are not nearly so stable.
All the genital appendages of the male, the oedagus, the claspers,
and ventral valves, show a diversity of form not approached by
the female gonapophyses. Even the abdominal sclerites of the
male become involved and form accessory apparatus to the
genital appendages. The greater inconstancy of the male is, I
believe, due to the fact that new genes may find expression in in-
dividuals possessing a single x-chromosome in all their cells which
could not find expression in the presence of two x-chromosomes.

Factors to be important in heredity must arise in the germ
cells. They must arise therefore either in the spermatogonia
and oogonia or in spermatocytes and oocytes. New genes must
have their origin in cells which are male, or cells which are female.
It is conceivable that a mutation might arise in a germ cell of a
male which could not arise in a female germ cell owing to greater
stability and conservativeness of the latter, due to its chromatic
constitution. Likewise, the expression of any new gene must
often be entirely dependent upon those genes already established
in the heredity of the individual. Those genes already fixed in
the evolution of the family, genus, or species form, as it were,
the internal environment for the new gene. The production of
characteristics in the development of an individual are believed
to come through a series of changes in the protoplasmic mass
governed and controlled by the genes located in the chromatin.
Each step must be dependent upon those steps which preceded,
and thus the most recent characteristics must be dependent on
the changes produced in the cells by the genes older phylo-
genetically.

The egg of an insect even before fertilization becomes highly
differentiated. This subject has recently been presented by
Hegner (’17). The regions of the cytoplasm become special-
ized, polarity and bilaterality are established. We look upon
the force which determines and brings about this organization

622 SIDNEY I. KORNHAUSER

as emanating from the nucleus of the growing odcyte. After
fertilization the division of the cleavage nucleus gives rise to a
number of nuclei, and these possessing complete diploid sets of
chromosomes migrate to various parts of the cytoplasm which
become more and more specialized as development proceeds.
This specialization becomes rigid, as shown by the experiments
on centrifuging insect eggs, destroying portions of developing
eges (Hegner, 717). In the latter class of experiments, the
uninjured portion continues to develop and forms only that
part of the embryo which would be formed by it in a normal
egg. However, we do not believe that even in insects the cells
of the embryo act as independent units, although this condition
becomes almost realized in the formation of the adult. In the
embryo the cells codperate to form various organs, and there
must be some vital intercellular connection. As development
proceeds in an insect egg the internal mechanism of the indi-
vidual cells seems to be of the greatest importance in shaping
the most highly differentiated tissues of the adult. Crampton
(99) showed this nicely in his lepidopteran experiments. Hypo-
dermis grafted onto a transforming pupa of another species or of
opposite sex developed as it would have in the individual from
which it came, as shown by the pigmentation and cuticular
out-growths of the ingrafted portions.

Must we not also look upon the development of sex in an indi-
vidual as a continuous series beginning with the differentiation
of the gonads, proceeding in the formation of the important
accessory sexual organs and culminating in the expression of
various specific secondary sexual characteristics? Those sexual
differences old phylogenetically are developed early in ontogeny
and each of the more recent genes affecting sexual characteristics
must be dependent upon the entire series of changes which
preceded. Every cell of the individual possessing a complete
set of chromosomes probably has a double set of genes repre-
senting the total hereditary basis of the species. There is a
single exception to this, namely, that in the male the genes
located in the x-chromosome are present only once. We know
from the various lines of evidence presented on pages 539 to

SEXUAL CHARACTERISTICS OF THELIA 623

544 that each sex possesses the genes for the secondary sexual
characteristics of the opposite sex, but that normally one set
finds expression in the male and the other set in the female.
Thus they differ from sex-linked characteristics which may be
present in either males or females, although this distinction is
sometimes not recognized. The ‘exclusively male character-
istics’ studied by Foot and Strobell (15, ’17 a, 717 b) belong to
the secondary sexual characteristics which in the cases best
known have their genes located in the autosomes. The above-
named investigators have shown in their Euschistus crosses that
the gene or genes controlling the length of the intromittant
organ may be transmitted from father to son, and that, since
the son receives no x-chromosome from the male pronucleus,
this gene or group of genes must be located in the autosomes.
It does not seem improbable or impossible that the expression
of a particular set of genes for the secondary sexual character-
istics, even though located in the autosomes, should be dependent
upon some controlling gene or genes in the x-chromosome.
With the knowledge of the multiple effects of a single gene, and
the presence of modifying and inhibiting factors, more and more
evidence is being accumulated to show that a particular factor
in one chromosome may influence the expression of genes in
other chromosomes. The writer would contend that the impor-
tant and vital genes of the x-chromosomes, not those forming the
sex-linked characteristics, but those which must be present
either once or twice for development to take place at all, create,
if present once, an environment for a series of changes leading
to the expression one by one of the male characteristics, and, if
present twice, lead to the development of the female character-
istics. If we consider for a moment the ontogenetic series in
Thelia, we find that in the embryo one may see the differentia-
tion of the gonads into ovaries or testes. The arrangement of
the gonia, the form and method of attachment of the tubules,
the structure of the gonaducts, and the position of the genital
apertures early distinguish the males from the females. The
external genital appendages, developing most probably from the
primitive abdominal limb buds of the eighth to tenth somites,

624 SIDNEY I. KORNHAUSER

early become distinguishable as male or female, and this dif-
ference becomes more and more apparent as development pro-
ceeds. In the fourth and fifth nymphal instars the somatic
cells as a whole show physiological differences in that the females
grow more slowly, but grow to a greater size. Last of all, with
the final molt and sexual maturity, comes the expression of a
host of secondary sexual characteristics. Pigmentation, size,
form, and detail of the sclerites show various sexual differences
and the gonapophyses exhibit their specific characteristics. The
difference in metabolic pitch now becomes very evident as the
female stores yolk material, while the more active male uses up
his stored energy. Behavior differences connected primarily
with the act of reproduction also appear in the imagos.

In the insects we know of no hormones produced in the gonads
either early or late in ontogeny, which, circulating in the blood,
constitute a factor in the internal environment of the somatic
cells stimulating certain genes to find expression and suppressing
others. In vertebrates such hormones certainly exist, and are
generally believed to be produced in the interstitial tissue of the
testis or ovary. They are different in the two sexes, as shown by
their effects upon developing somas. ‘These hormones must be
produced through the influence of certain genes which are active
in the interstitial cells. That the hormone may be modified by
new genes has been demonstrated by Morgan’s (’17) experi-
ments on hen-feathered Seabright cocks. In these birds the
testis produces a hormone which prevents the development of
the normal male plumage and causes a large proportion of steril-
ity, according to Goodale (16). That the sex-hormone-forming
genes of the male find their expression in the presence of one
x-chromosome while those of the female depend on two x-chro-
mosomes seems not unlikely. The hormones from the gonads
and probably in some cases from other endocrine glands circulate
in the blood and form a very important step in sexual develop-
ment, creating at times a necessary factor for the expression of
those genes which follow in their activity in the series of develop-
mental changes. But while no definite hormone forms an im-
portant step in the sexual development of insects, still each gene

SEXUAL CHARACTERISTICS OF THELIA 625

as it comes into activity must cause a change in the protoplasmic
mass which it influences and, since the cells are not independent
but in physiological continuity, this constitutes a definite step
also. Since each gene is represented in every cell of the indi-
vidual, its coming into particular activity in one part of the
organism might be accompanied by minor changes due to this
same gene in other parts of the organism. In both arthropods
and vertebrates I would look upon the sexual development of
the individual as a continuous series, each step depending on the
steps preceding. The presence of one x-chromosome would
form the basis for the inauguration of the male series, beginning
in the formation of a sperm-forming gonad and ending with the
production of specific secondary sexual characteristics. Like-
wise, the presence of two x-chromosomes in the zygote would
start the female series of developmental changes.

Returning finally to the effect of Aphelopus upon Thelia
bimaculata, will the foregoing considerations help us any in
understanding the changes suffered especially by the male in its
secondary sexual characteristics? It was shown that the germ
cells of the male were not changed toward those of the female
type, nor were the male gonapophyses altered so as to be sim-
ilar to those of the female, although they were reduced in size.
These reduced external genitalia, although still retaining the
general form common to the membracids, lost their specific
characteristics. In the fifth instar it was noted that parasitized
individuals of both sexes showed a retardation in the develop-
ment of the gonapophyses. The striking changes brought about
were in those characteristics of the male which first appear at
the final molt and belong to the category of the extragenital
characteristics. Especially noteworthy was the loss of the male
coloration and the assumption of the female pigment and pat-
tern. Likewise, the female arrangement of the spines on the
abdominal sclerites and increase in size of all parts of the body
were plainly observed in parasitized males. It must also be
remembered that parasitized males formed a complete series
from those but slightly altered to those with extreme change.
Those but slightly changed, if taken shortly after the final molt,

626 SIDNEY I. KORNHAUSER

showed the presence of only small parasites, and it seemed reas-
onable to believe that the degree of change was largely due to
the size of the parasitic larvae during the fifth instar. Not by
their size merely, but by their greater physiological effect, would
the presence of the large parasites greatly alter the internal
conditions. The degree of change in the host’s form would be
directly referable to the degree of internal alteration effected by
the parasites. It might be expressed as a quantitative reaction
depending on the concentration of the metabolic products of the
Aphelopus larvae. The sexual characteristics of Thelia which
appeared early in ontogeny would be subjected to the products
of the parasites’ metabolism during several molts, although the
concentration of these products must be rather less in the earlier
instars. Yet in spite of this the older characteristics we found
were not reversed. Spermatogonia were surely subjected to
conditions where the growth of odcytes would be possible and
still only spermatocytes developed. The external genitalia
were greatly reduced in size, but still retained the general form
found in the Membracidae. Their reduction in size is probably
due to the fact that the step from the genitalia of the fifth
nymphal instar to the adult form is a big one, the size difference
being rather remarkable in the two stages; and, just while this
growth is taking place in parasitized individuals, the Aphelopus
larvae must exert a greater influence than they did in the pre-
vious instar or instars when smaller in size. It may also be true
that it is difficult to reverse the development of an organ old
ontogenetically and well established in the group after it has
once got under way. There also seems to be good reason to
believe that the more recent the characteristic and the more
specific it is the more it will be altered. The details of sexual
differentiation often represent various specializations of phylo-
genetically older differences. The specific characteristics of the
gonapophyses come under this category, and thus we find that
the oedagus and ovipositor of Thelia bimaculata are altered in
regard to these details when parasites are present. Gliard (’89)
showed that the elaborate oedagus of Typhlocyba hippocastani
lost its specific characteristics when parasitized by Aphelopus,

SEXUAL CHARACTERISTICS OF THELIA 627

so that this organ could no longer be used in distinguishing hip-
pocastani from certain other species.

The color reversal in Thelia bimaculata was very complete in
many parasitized males. If we examine various examples of the
tribe Telamonini, to which Thelia belongs, we find that the
coloration of the male is not often strikingly different from that
of the female. In Thelia bimaculata we have an exceptional
and extreme case of dimorphism in coloration. Thus in Thelia
uhlert the nearest relative both sexes are of a reddish-brown
color. Microscopical examination shows that a faintly out-
lined lateral area on the pronotum corresponding to the vitta of
bimaculata exists and that the melanic pigment of the pronotum
is chiefly restricted to the punctures. No sexual difference
either in the pigment or in the pattern on the pronotum or face
could be found. The coloration certainly resembles that found
in the female of bimaculata much more than the male, and I
believe we are to look upon Thelia uhleri as representing a more
ancestral type. There are several reasons for this. The gona-
pophyses are quite similar in form to those of other genera of
the Telamonini, as indicated by the bulbous oedagus and pointed
ovipositor. Also the arrangement of the spines on the abdom-
inal terga shows a greater similarity in the patterns of the sexes
than was found in Thelia bimaculata. The rows in the male
are not so straight and close together and the characteristic
network arrangement of the female may also be found over
parts of the male terga. Thus I believe that the coloration and
arrangement of the abdominal spines in the male of Thelia
bimaculata represent recent changes and that the female char-
acteristics represent the more ancient and less modified con-
dition, ‘These recent sexual characteristics of the male must
have arisen through genetic changes modifying characteristics
once common to both sexes. The genes for these characteristics
of the male find their expression in the final ontogenetic step in
the presence of one x-chromosome, and they probably depend
upon all the changes in development which preceded them. Is
it not likely that the internal upset caused by the parasites
interferes with the normal expression of these male genes?

JOURNAL OF MORPHOLOGY, VOL, $2, NO. 3

628 SIDNEY I. KORNHAUSER

Let us briefly review the possibilities of internal alteration
which might affect the cells. The growing larvae depend en-
tirely upon the host for their respiratory and nutritive needs.
Oxygen must be extracted from the haemolymph or produced
by the reduction of carbohydrates in the haemolymph. Carbon-
dioxide must be poured back into the haemolymph by the para-
sites. It is not unlikely that full-grown larvae distending the
abdomen of the host interfere with its respiratory movements.
Should an insufficient supply of oxygen be furnished the ¢ells
of the host, a condition of acidosis would result in the tissues.
This would bring about the accumulation of droplets of fat in
the cells and lead to the degeneration of many cells. Micro-
scopic examination shows much fatty infiltration and the accu-
mulation of adipose tissue around the Aphelopus larvae. This
tissue probably forms one of the chief sources of energy for the
development of the parasites. But not only energy is needed,
proteins must be supplied as well by the haemolymph of the
host. In growing the parasites must give rise to nitrogenous
waste materials. The premegagnathic larvae, which can alter
an adult considerably if present before the final molt, probably
excrete these waste materials directly into the haemolymph of
the host. In the megagnathic larvae at least part of the kata-
bolic products are deposited as insoluble crystals in the digestive
tract. There is a possibility that nitrogenous excretions from
the larvae might be toxic. Parasitic worms have long been
known to give rise to toxic excretions (Firth, ’03). These toxins
might act indirectly upon the developing cells. In Thelia, as in
all homopterans probably, there exists a tissue well organized
and of considerable size, arranged segmentally in the abdomen,
and known as the pseudovittelus. This tissue, in which sym-
biotic fungi develop, is thought to be very important in the metab-
olism of the animal. In parasitized Thelia the pseudovittelus
is much reduced, although I have seen no indications of fatty
degeneration in its cells. This reduction of the pseudovittelus
might also alter the composition of the haemolymph. With the
possibility of so many alterations in the haemolymph which
bathes the cells and furnished them with food and oxygen, a

SEXUAL CHARACTERISTICS OF THELIA 629

complete solution of the problem as to the one important change
seems rather remote. We would maintain that the changes
were not effected through the production of some sex hormone
and that the parasites do not act like an ovary to stimulate the
development of the female secondary sexual characteristics in
males. It seems rather that the altered haemolymph failed to
provide the necessary conditions for the expression of the char-
acteristics most recent in the evolution of Thelia bimaculata,
namely, the specific characteristics of the gonapophyses and the
male extragenital sexual characteristics. These new character-
istics may have come into the species by the origin of genes
modifying established characteristics, and the failure of these
genes to come into activity would produce an organism retain-
ing ancestral characteristics. The male and femaleindividuals
would then be more similar in color, size, and in other details of
their integument. A slight change in the environment would
permit a partial expression of the newest genes, and thus inter-
mediates were found in males containing small larvae.

9. SUMMARY

1. Thelia bimaculata is parasitized by the dryinid Aphelopus
theliae. An egg of the parasite may be deposited in an immature
Thelia of any of the five nymphal instars. The Aphelopus ovum
undergoes polyembryonic development and gives rise to about
fifty larvae which reach their maximum development either in
the fifth nymphal instar of the host or in its adult stage. The
Aphelopus larvae escaping from their host, thereby killing it,
drop to the ground, burrow in, pupate, and become mature the
following summer.

2. Thelia parasitized by Aphelopus show many alterations.
Most interesting are the changes wrought in males which reach
the adult form. These assume either partially or completely
many sexual characteristics of the female, much as do sacculi-
nized crabs. The degree of change depends upon the size of the
parasites during the fifth nymphal instar of the host. If the
Aphelopus ovum is deposited early in the nymphal life of the

630 SIDNEY I. KORNHAUSER

host, the parasitic larvae will be large and the assumption of the
female characteristics pronounced; if deposited late, the altera-
tions will be less marked.

3. Of the changes in parasitized males none is more striking
than the assumption of the pigmentation of the female. The
character of the pigment and its distribution on the pronotum
and head may duplicate exactly that of the female.

4, Parasitized males increase in size, approaching but not
reaching the size characteristic for female Thelia. Measure-
ments show this increase in the pronota, wings, heads, legs,
acrotergites, and abdomens. Thus all regions of the body are
influenced and the amount of increase is correlated with the
degree of alteration of the pigmentation, those with complete
female coloration being largest.

5. Parasitized female Thelia show no assumption of male
pigmentation, nor do they change in size.

6. Minute spines on the abdominal sclerites of parasitized
males take on the arrangements characteristic of the female.
The shape, pigmentation, and texture of the abdominal sclerites
of parasitized males become female in character and various
sclerites of the terminal somites associated with the genital
appendages show considerable change toward the opposite sex.

7. Parasitized individuals of both sexes sometimes show a
weakened cuticula and a reduction of the melanic pigment.

8. None of the changes described are due to a retention of
nymphal characteristics.

9. The genital appendages of parasitized males are not changed
to those of the opposite sex. They are reduced in size and lose
their specific characteristics, but retain the general form found
in male Membracidae. Likewise, the gonapophyses of infected
females show a similar decrease in size and a loss of specific
characteristics, but retain the general form found in female
membracids.

10. The above (9) may be partly explained by a history of
the gonapophyses. The genital appendages are laid down early
in ontogeny and become clearly sexually differentiated in young
nymphs. Thus the sexes may be easily ascertained by an

SEXUAL CHARACTERISTICS OF THELIA 631

external examination of the nymphal gonapophyses of the third
instar and by an examination of sections much earlier than this.
These genital appendages are ancient ontogenetically and phylo-
genetically as well, belonging to a type primitive for the group
of the Homoptera. The small parasites of the younger instars
are not capable of changing the growing sexually differentiated
nymphal gonapophyses to those of the opposite sex. They may
retard the development of these appendages in both sexes, so
that fifth instars with full-grown parasites present gonapophyses
quite similar to those found in fourth instars.

11. The parasites generally cause the degeneration of the
gonads and bring about an accumulation of fat in the abdomen
of the host. Testes and ovaries in various stages of disappear-
ance were studied. Never were cells similar to odcytes found
in any testes, but normal spermatogenesis proceeded as far as
possible under adverse conditions. Likewise, odcytes retained
their characteristic features and even grew for a time in some
parasitized adult females.

12. One parasitized male was found which, although con-
siderably changed toward the female condition, contained a
full-sized normal testis with many spermatozoa. Another in-
dividual, a fourth instar, showed perfect female soma, but con-
tained male gonads composed of cells with the characteristic
male complex of chromosomes. The first individual indicates
that the changes wrought by the parasites are not due directly
to the destruction of the gonads, while the second individual
lends support to the idea that the soma of arthropods is inde-
pendent of the gonads in its development. Likewise, there is
no evidence that the demand for food made by the parasites on
the soma of male hosts stimulates the development of female
secondary sexual characteristics. Such a demand has been
assumed by other investigators to emanate normally from the
ovaries, calling forth the development of the female characteristics.

13. It is not thought probable that the lowering of the meta-
bolic level in parasitized males could account for the®hanges
described.

632 SIDNEY I. KORNHAUSER

14. Male membracids show a greater degree of variation than
do females. Males seem to have a less stable constitution and
have departed from generalized types farther than have females.
This difference in constitution probably rests on the chromatic
makeup of the two sexes—the xx condition being more stable
than the xy or xo condition. The male of Thelia bimaculata
shows extreme departures from the female in many character-
istics of form, size, color, and pattern. The male extragenital
sexual characteristics and the specific characteristics of the
gonapophyses first appear in the final molt. The extragenital
sexual characteristics may be due to genes which in many cases
modify characteristics once common to both sexes, but now
exhibited in a primitive condition only in the more stable female.
The metabolism of the parasites, altering the constitution of the
host’s haemolymph which bathes the developing cells, may offer
an environment unsuitable for these recent genes to find their
expression and leave the male individual in a more primitive
state exhibiting characteristics now found in the female of the
species.

Evanston, Illinois,
May, 1918

BIBLIOGRAPHY

Amyot, C. J. B., anp Servitun, A. 1843 Histoire Naturelle des Insectes
Hémiptéres. Paris. 675 pp., 12 pl.

Barur, W. B. von 1909 Die Oogenese bei einigen viviparen Aphididen und
die Spermatogenese von Aphis saliceti, mit besonderer Beriicksich-
tigung der Chromatinverhidltnisse. Arch. f. Zellf., Bd. 3, S. 269-333,
Taf. 12-15.

Banta, A. M. 1916 Sex intergrades in a species of Crustacea. Proc. Nation.
Acad. Sci., vol. 2, pp. 578-583.

Boveri, Tu. 1915 Uber die Entstehung der Eugsterschen Zwitterbienen.
Arch, Entw.-Mech., Bd. 41, Hft. 2, S. 264-311, Taf. 7-8, 2 Textfig.

Bremer, Hans 1914 Zwei Faille von Pseudohermaphroditismus bei Diaptomus
vulgaris Schmeil. Zool. Anz., Bd. 44, 8. 572-574, 3 Textfig.

Bripees, C. B. 1916 Non-disjunction as proof of the chromosome theory of
heredity. Genetics, vol. 1, no. 1, pp. 1-55, and vol. 1, no. 2, pp. 107-163,
J pl., 9 text figs.
1917 Deficiency. Genetics, vol. 2, no. 5, pp. 445-465.

Brinpuey, H. H., anp Ports, F. A. 1910 The effect of parasitic castration in
insects. Science, N. S., vol. 32, No. 832, pp. 836.

SEXUAL CHARACTERISTICS OF THELIA 633

CHAPIN, CaTHAaRINE L. 1917 A microscopic study of the reproductive system
of foetal free-martins. Jour. Exp. Zodél., vol. 28, no. 2, pp. 453-478,
16 text figs.

Cockayne, E. A. 1915 Gynandromorphism and kindred problems. Journ. of
Genetics, vol. 5, no. 2, pp. 75-128, 4 pl., 10 text figs.

Crampron, H. E. 1899 An experimental study upon Lepidoptera. Arch.
Entw.-Mech., Bd. 9, 8. 293-318, pl. 11-13, 13 text figs.

Doncaster, L. 1914 Chromosomes, heredity and sex: a review of the present
state of the evidence with regard to the material basis of hereditary
transmission and sex-determination. Quart. Journ. Micr. Sci., N. 8S.,
vol. 59, no. 236, pp. 487-521, 4 text figs.

Duncan, F. N. 1915 A note on the gonads of gynandromorphs of Drosophila
ampelophila. Amer. Natural., vol. 49, no. 583, pp. 455-456.

Foot, KATHERINE, AND STROBELL, E. C. 1914 Preliminary report of crossing
two hemipterous species, with reference to the inheritance of a second
exclusively male character. Biol. Bull., vol. 27, no. 4, pp. 217-234,
pl sk;

1915 Results of crossing two hemipterous species, with reference to
the inheritance of two exclusively male characters. Journ. Linn. Soc.
London, Zool., vol. 32, pp. 457-487, pl. 41-47.

1917 a Results of crossing Euschistus variolarius and Euschistus
servus, with reference to the inheritance of an exclusively male char-
acter. Journ. Linn. Soc. London, Zool., vol. 32, pp. 337-370, pl. 28-
34, 2 text figs.

1917 b Results of crossing Euschistus variolarius and Euschistus
ictericus, with reference to the inheritance of two exclusively male
characters. Biol. Bull., vol. 32, no. 5, pp. 322-336, pl. 1+.

Fowier, W. W. 1903 Biologia Centrali-Americana. Insecta. Rhynchota.
Hemiptera-Homoptera, vol. 2, part 1, 339 pp., 21 pl.

FunkHovuser, W. D. 1915 Life history of Thelia bimaculata Fab. (Mem-
bracidae). Ann. entom. Soc., Amer., vol. 8, no. 2, pp. 140-151, pl. 9.
1917 Biology of the Membracidae of the Cayuga Lake basin. Cor-
nell Univ. Agric. Exp. Stat. Memoir 11, pp. 181-445, pl. 23-44, 6 text
figs.

Firtu, Orro von 1903 Vergleichende chemische Physiologie der niederen
Tiere. Fischer, Jena. 670 S.

Ganan, A. B. 1918 An interesting new hymenopterous parasite. Canad.
Entom., vol. 50, no. 5, pp. 151-152.

GaTENBY, J.B. 1917 Note on the sex of a tadpole raised by artificial partheno-
genesis. Quart. Journ. Micr. Sci., N. S., no. 246, vol. 62, pt. 2, pp.
213-216, 5 text figs.

Geyer, Kurr. 1913 Untersuchungen iiber die chemische Zusammensetzung
der Insektenhaemolymphe und ihre Bedeutung fiir die geschlechtliche
Differenzierung. Zeitschr. f. wiss. Zool., Bd. 105 8. 350-499, Taf. 20-
22, 58 textfig.

GrarpD, A. 1886 De ’influence de certains parasites rhizocéphales sur les char-
actéres sexuels extérieurs de leur héte. C. R. Acad. Sci. Paris., T.
103, pp. 84-86.

634 SIDNEY I. KORNHAUSER

Grarp, A. 1887 a Parasitic castration and its influence upon the external char-
acters of the male sex, in the Decapod Crustacea. Ann. Mag. Nat.
Hist., ser. 5, vol. 19, no. 113, pp. 325-345, 7 text figs. Trans. by W. S.
Dallas from Bull. Scientifique du Nord. ser. 2, an. 10. (Original not
available.)

1887 b Sur la castration parasitaire chez l’Eupagurus Bernhardus
Linné et chez la Gebia stellata Montagu. C. R. Acad. Sci. Paris, T.
104, pp. 1113-1115.

1888 Sur la castration parasitaire chez les Eukyphotes des genres
Palaemon et Hippolyte. C. R. Acad. Sci. Paris, T. 106, pp. 502-505.
1889 Sur la castration parasitaire des Typhlocyba par une larve
d’Hyménoptére (Aphelopus melaleucus Dalm.) et par une larve de
Diptére (Atelenevra spuria Meig.). C. R. Acad. Sci. Paris, T. 109,
pp. 708-710.

Goutpscumipt, R. 1916 Experimental intersexuality and the sex-problem.
Amer. Natural., vol. 50, no. 600, pp. 705-718.

1917 a On a case of facultative parthenogenesis in the gypsy-moth
Lymantria dispar L. with a discussion of the relation of partheno-
genesis to sex, Biol. Bull., vol. 32, no. 1, pp. 35-48.

1917 b A further contribution to the theory of sex. Jour. Exp.
Zool., vol. 22, no. 3, pp. 593-611, 53 text figs.

Goopatr, H. D. 1916 Gonadectomy in relation to the secondary sexual char-
acters of some domestic birds. Carnegie Inst. Washington, Publ.
no. 243, pp. 5-52, 7 pl.

Grassi, B., anp Sanpras, A. 1893 The constitution and development of the
society of termites: observations on their habits; with appendices on
the parasitic protozoa of Termitidae, and on the Embiidae. Quart.
Journ. Mier. Sci., N. S., vol. 39, pp. 245-315, pl. 12-16, vol. 40, pp. 1-75.

Harrison, J. W. H., anp Doncaster, L. 1914 On hybrids between moths of
the Geometrid sub-family Bistoninae, with an account of the behaviour
of the chromosomes in gametogenesis in Lycia (Biston) hirtaria, Ithysia
(Nyssia) zonaria and in their hybrids. Journ. of Genetics, vol. 3,
no. 4, pp. 229-247, pl. 17-18.

Heener, R.W. 1917 The genesis of the organization of the insect egg. Amer.
Natural., vol. 41, no. 611, pp. 641-661; no. 612, pp. 705-718; 12 text figs.

Hoy, W. E. 1916 A study of somatic chromosomes. I. The somatic chromo-
somes in comparison with the chromosomes in the germ cells of Anasa
tristis. Biol. Bull., vol. 31, no. 5, pp. 329-363, 10 text figs.

Kammerer, P. 1912 Ursprung der Geschlechtsunterschiede. Fortschritte d.
Naturwiss. Forsch. (Abderhalden), Bd. 5, S. 1-240.

Kastie, J. H. 1909 The oxidases, and other oxygen-catalysts concerned in
biological oxidations. Bull. no. 59, Hygienic Lab., Washington, D. C.,
164 pp.

Keiitn, D., rr THompson, W. R. 1915 Sur le cycle évolutif des Dryinidae,
Hyménoptéres parasites des Hemiptéres homoptéres. C. R. soe. biol.
Paris, T. 78, pp. 83-87, 10 text figs. i

Ketioae, V. L. 1904 Influence of the primary reproductive organs on the sec-
ondary sexual characters. Jour. Exp. Zodél., vol. 1, no. 4, pp. 601-605.

SEXUAL CHARACTERISTICS OF THELIA 635

Kopr¢é, Steran 1911 Untersuchungen iiber Kastration und Transplantation
bei Schmetterlingen. Arch. Entw.-Mech., Bd. 33, S. 1-116, Taf.
1+4, 19 Textfig.

1913 a Untersuchungen iiber die Regeneration von Larvalorganen
und Imaginalscheiben bei Schmetterlingen. Arch. Entw.-Mech.,
Bd. 37, S. 440-472, Taf. 12-14, 6 Textfig.

1913 b Nochmals tiber die Unabhingigkeit der Ausbildung sekundirer
Geschlechtscharaktere von den Gonaden bei Lepidopteren (Fihlerre-
generationsversuche mit Kastration und Keimdriisentransplantation
kombiniert). Zool. Anz., Bd. 48, No. 2, 8. 65-74.

Kornuavser, 8. I. 1914 A comparative study of the chromosomes in the sper-
matogenesis of Enchenopa binotata (Say) and Enchenopa (Campy-
lenchia Stal) curvata (Fabr.). Arch. f. Zellf., Bd. 12, Hft. 2, S. 241-
298, pl. 18-22, 7 text figs.

1916 Celloidin paraffin method. Science, N. S., vol. 44, no. 1124,
pp. 57-58.

KtnckeEt p’Hercunats, M. J. 1894 Les Diptéres parasites des Acridiens les
Muscides vivipares 4 larves sarcophages. C. R. Acad. Sci. Paris, T.
118, pp. 1106-1108.

Lituiz, F. R. 1917 The free-martin; a study of the action of sex hormones in
the foetal life of cattle. Jour. Exp. Zodl., vol. 23, no. 2, pp. 371-452,
29 figs.

Lors, J. 1916 The sex of parthenogenetic frogs. Proc. Nation. Acad. Sci.,
vol. 2, pp. 313-317, 3 text figs.

Martavuscyu, I. 1909 Gynandromorphic Membracidae. Jour. N. Y. Entom.,
Soc., vol. 17, pp. 165, pl. 18.

1911 The effects of parasitic castration in Membracidae. Journ.
N. Y. Entom. Soc., vol. 19, no. 3, pp. 194-196, pl. 6.

MEISENHEIMER, J. 1909 Experimentelle Studien zur Soma- und Geschlechts-
Differenzierung. Erster Beitrag. Gustave Fischer. Jena. S. 1-147,
2 Taf., 55 Textfig.

Morean, T. H. 1909 A biological and cytological study of sex determination
in Phylloxerans and Aphids. Jour. Exp. Zoél., vol. 7, no. 2, pp. 239-
352, 1 pl., 23 text figs.

1916 The Eugster gynandromorph bees. Amer. Natural., vol. 50,
no. 589, pp. 39-45.
1917 The theory of the gene. Amer. Natural., vol. 51,no. 609, pp.
513-544, 9 text figs.

Morean: T. H., anp Bripges, C. B. 1916 Sex-linked inheritance in Droso-
phila. Carnegie Instn. Washington,. pub. 237, 87 pp., 2 pl., 8 text
figs.

Morritt, C. V. 1910 The chromosomes in the odgenesis, fertilization and
cleavage of Coreid Hemiptera. Biol. Bull., vol. 19, no. 2, pp. 79-126,
2 pl., 12 text figs.

OvuprEmMans, J. TH. 1898 Falter aus castrirten Raupen, wie sie aussehen und
wie sie sich benehmen. Zodél. Jahrb., Abt. f. Syst., Bd. 12, S. 71-88,
Taf. 3-5, 2 Textfig. i

Parrerson, J. T. 1917 Studies on the biology of Paracopidosomopsis. I.
Data on the sexes. Biol. Bull., vol. 32, no. 5, pp. 291-305.

636 SIDNEY I. KORNHAUSER

Pérez, J. 1886 Des Effets du Parasitisme des Stylops sur les Apiaires du Genre
Andrena. Actes soc. linn. Bordeaux, T. 40, ser. 4, T. 10, pp. 21-60, pl. 1-2.

Potts, F. A. 1906 The modification of the sexual characters of the hermit
crab caused by the parasite Peltogaster (castration parasitaire of
Giard). Quart. Journ. Micr. Sci., N. S., no. 200 (vol. 50, part 4), pp.
599-620, pl. 33-34.

Recen, J. 1909a Kastration und ihre Folgeerscheinungen bei Gryllus cam- .
pestris L. Zool. Anz., Bd. 34, 8. 477-478.

1909 b Kastration und ihre Folgeerscheinungen bei Gryllus cam-
pestris L. II. Mitteilung. Zool. Anz., Bd. 35, S. 427-482.

Rippie, O. 1917 The control of the sex ratio. Journ. Washington Acad. Sci.,
vol. 7, no. 11, pp. 319-356.

Rogson, G. C. 1911 The effect of Sacculina upon the fat metabolism of its
host. Quart. Journ. Micr. Sci., N. S., no. 226 (vol. 57, part 3), pp.
267-278, 2 text figs.

Smitu, G. 1905 Note on a Gregarine (Aggregata inachi, n.. sp.) which may

cause the parasitic castration of its host (Inachus dorsettensis).
Mitt. Zool. Stat. Neapel, Bd. 17, S. 406-409, Taf. 26.
1910 Studies: in the experimental analysis of sex. Quart. Journ.
Mier. Sci., N. S., no. 216 (vol. 54, part 4), pp. 577-604, pl. 30.
1911 Studies in the experimental analysis of sex. Part 7. Sexual
changes in the blood and liver of Carcinus maenas. Quart. Journ,
Micr. Sci., N. 8., no. 226 (vol. 57, part 3), pp. 251-265.
1913 Studies in the experimental analysis of sex. Part 10. The
effect of Sacculina on the storage of fat and glycogen, and on the for-
mation of pigment by its host. Quart. Journ. Micr. Sci., N. S., no.
234, vol. 59, pp. 267-295, 1 text fig.
1914 Studies in the experimental analysis of sex. Part 11. On
Stylops and Stylopisation. Quart. Journ. Micr. Sci., N. S., no. 239
vol. 60, pp. 485-461, pl. 32-35.

Srecur, O. 1912 Die ‘sekundiren’ Geschlechtscharaktere der Insekten und
das Problem der Vererbung des Geschlechts. Zeitschr. f. indukt.
Abstamm. u. Vererb., Bd. 8, Hft. 3, S. 284-291.

Srminacu, E. 1912 Willkiirliche Umwandlung von Siéugetier-Mannchen in
Tiere mit ausgeprigt weiblichen Geschlechtscharakteren und weib-
licher Psyche. Arch. ges. Physiol., Bd. 144, S. 71-108, Taf. 3-8.

TaNnDLER, J., AND Gross, 8. 1913 Die biologischen Grundlagen der sekundiren
Geschlechtscharaktere. Berlin. Julius Springer. 169 S., 23 Textfig.

Toyama, K. 1905 Studies on the hybridology of insects. I. On some silk-
worm crosses, with special reference to Mendel’s law of heredity.
Bull. Coll. Agric. Tokyo, vol. 7, pp. 259-393, pl. 6-11.

Wenke, K. 1906 Anatomie eines Argynnis paphia-Zwitters, nebst vergleich-
end-anatomischen Betrachtungen iiber den Hermaphroditismus bei
Lepidopteren. Zeitschr. f. wiss. Zool., Bd. 84, S. 95-138, Taf. 7-8,
15 Textfig.

Wueeter, W. M. 1910 The effect of parasitic and other kinds of castration in
insects. Jour. Exp. Zoél., vol. 8, pp. 377-438, 8 text figs.

1914 Gynandromorphous ants described during the decade 1903-1913.
Amer. Natural., vol. 48, no. 565, pp. 49-56.

ry we pose ae Nee
2 ; cia onus fam mt ; A ,
bia We Gamal" eat ia he sa Aa eee i ae
he’ Hits ? 5, “Wii's cake i piped ‘
. A iF aude f Polio i
i | oe ee , . ay ay ae 4 tris ‘ 4
AL ye { . SN ey A a iets 2 Te
iy, seit | Atal ‘ig * spasisiesr ips, shims
ve oe xe i) Py OT ee rey as ea ty Lr Mon ine
sicis ae cro a bi eas om ane ae SMa wes 5) *f wend : wae vr ie r
he Lied ibe eh a a ate 1b if; sink sesso itite Bel ui!
haan a Sie an jens UP sie i yn aay Pt. pk ve tts i - me iv) au oe A he"

ayVID Ra

A eh, oy

ROT EES A Sige Pear On ae Pa suey J
f P .j at ke rime acts ie 73 shit de ‘ tt TIC} OS
FLyamantp Sf sa ab. v2)

the Hira) Te Peet oa erty 4%, “eye | ' rs shat sy fit sity

r Spy : : j 5 ‘
Tea RT Tee ce ay) Be Tae eT rie
Beas jy tery fatron Rie ORT BETS bgt prerehjiae Py 1
be Oh holes haerchastee expan

Gf is ih AMID 4 Fy vidi he dass AON MEDD nis

vial a apiaanehh 220 F ako RO ORD GF eaeky

WG ae sah) oaibee aN opti! 4 oh rity ‘abr
fo o Nien vies, abhi Mi owl % by oy palonett ‘ an
Liye shale yeti hae ta ‘coats Poul peggy e yy.
; um Raa mht) , ; Don a
ra. 7 mane fon rah Hepee wa ae
vey ene lar j panei )

: Ai ba ‘.,

Resumen por el autor, Tokuyasu Kudo,
Instituto Anatémico de Migata, Japon.

La musculatura facial de los japoneses.

El presente trabajo trata de los musculos faciales y se basa en
la diseccién de quince japoneses, tres chinos y cinco europeos,
compardndose los resultados obtenidos con los de otras razas.
La musculatura de los mongoles esta generalmente menos difer-
enciada y es mas primitiva que la de los europeos, aunque en
diversos casos sucede lo contrario. El autor dd detalles sobre
cada uno de los miuisculos de la cara.

Translation by José F. Nonidez
Carnegie Institution of Washington

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, AUGUST 11

THE FACIAL MUSCULATURE OF THE JAPANESE

TOKUYASU KUDO

Formerly Assistant in the Anatomical Institute of the Imperial Japanese University
of Kyoto (Acting Professor in the Anatomical Institute
of Migata)

FIVE TEXT FIGURES AND THREE PLATES

Since Chudzinski and Giacomini made their contribution on
the muscular system of the negro, many works have been pub-
lished, especially on the head musculature of the intra-European
races (Forster, Birkner, Fischer, Eggeling, Fetzer, Loth, H. Vir-
chow). These authors have uniformly proved that racial dif-
ferences are related to facial muscles. The following study is in-
tended to be a small contribution to the solution of the problem.

The material at my disposal consists of five Europeans (male),
three Chinese (male), and fifteen Japanese (ten male and five
female). All are adult. The Japanese cadavers came mostly
from hospitals, others from prisons (among these two that had
been hanged). I arranged them in the following order under the
respective abbreviations: for Europeans, EI to V; for Chinese,
CI to CIII; and for Japanese, JI to JXV (and likewise the males
JI to JX, females JXI to JXV).

The faces have been dissected on both sides or on a single side
(usually the right); the condition of nourishment of Europeans
and Chinese is good; that of the Japanese is as follows:

dils body no. 3778, age 51, emaciated.
JII, body no. 3768, age 18, good.
JIII, body no. 3048, age 44, good.

JIV, body no. 3750, age 18, moderate.
JV, body no. 3780, age 21, moderate.
JVI, body no. 3765, age 51, emaciated.
JVII, body no. 3762, age 30, moderate.

637

638 TOKUYASU KUDO

JVIII, body no. 3145, age 20, good.
JIX, body no. 3756, age 45, good.
JX, body no. 3786, age 21, emaciated.
JXI, body no. 3772, age 20, moderate.
JXII, body no. 3781, age 32, good.
JXIII, body no. 3770, age 61, emaciated.
JXIV, body no. 3659, age 21, moderate.
JXV, body no. 3729, age 28, good.

The Japanese heads were separated from the bodies and in-
jected with a carbol-glycerin-alcohol solution. The European
and Chinese heads were first salted and frozen, were then ex-
posed for a long time to formol fumes, and have been preserved
for many years in alcohol. In order to preserve the color of the
muscles, several heads were immersed in sodium chloride or
sodium nitrate solution. In the preparation for study a prepa-
ration needle was also used; furthermore, I have studied the more
delicate fibers in alcohol or water as Eisler has done.

I have made measurements of the muscles of the European and
Chinese, mostly on both sides of the face; but with the Japanese
I have always made the measurements on one side only. At-
tention may be especially called to the fact that the differences
of muscularization have been confirmed, mainly by direct com-
parison of the preparation with one another by the unaided
eye. The proportions based on measurements are not sufficiently
trustworthy.

Furthermore, it is to be noted that the majority of the European
authors in their studies of the facial muscles of extra-European
races do not show how they have compared these with the Euro-
peans. One cannot discover whether the authors have prepared
the European faces specially for this or not, what kind of prepa-
rations they were, nor how many heads of Europeans were used
for the comparison. In the comparison of the (evidently not
numerous) relationships of facial muscles, it is absolutely neces-
sary that they may be worked out by a juxtaposition of the
dissections.

FACIAL MUSCULATURE OF THE JAPANESE 639

At this place it is fitting that I express my heartiest thanks to
Professor Doctor Adachi for his manifold suggestions and the
liberal gift of the excellent material for unconditional use.

My study divides itself into two portions:

1. The individual facial muscle.

2. The facial musculature as a whole.

1 THE INDIVIDUAL FACIAL MUSCLES
Platysma myoides! (figs. 1 to 3)

In all of my material this muscle shows the greatest variety
in the muscle configuration of the face, hence I have not as yet
been able to discover any apparent difference between the Mon-
golian (Japanese and Chinese) and European types. lLoth has
rightly observed that it is erroneous to suppose that in a race in
which one expects a primitive musculature of the face, the
platysma should always be well developed.

The negro has, at times, well-developed muscles (Chudzinski,
Giacomini, Turner, Eggeling, Loth), at other times they are very
poorly developed (Giacomini, Loth). Among the Papuans, as
far as dissections have shown, the muscle is said to be well de-
veloped (Forster, Fischer, Steffens, and Kérner-Eckstein). On
the other hand, Fetzer has not expressly mentioned the primitive
character of the platysma in the study of seventeen Hottentot
heads.

In general, in the Japanese and Chinese, the platysma fibers
arising at the edge of the jaw are extensive and form a closed
muscle plate, a fact which Birkner has observed in three Chinese
heads. J have made a similar observation also on three Kuropean
preparations (EIJI, EIV, EV).

The fibers diminishing toward the corner of the mouth and
directed toward the lower lip (pars labialis) run more trans-
versely or upward and forward in the Mongolians and are not
plainly separated from the bundles ascending in the orbitotem-

1 For practical purposes the platysma is visible only in the cheek and neck
portions of the separated head.

640 TOKUYASU KUDO

poral direction (pars aberrans ascendans). The latter section
radiates at various elevations near the region of the cheek.

I have found in five Japanese and two Chinese? (figs. 1, 3, 4,
6, and 7) well-developed fibers which cross a line drawn from the
corner of the mouth to the outer auditory opening, but I have not
found these in European heads. The platysma which reaches
this line has been observed in four Japanese and three Euro-
peans.*? J have observed in six Japanese (fig. 5), a Chinese (fig.
2), and two Europeans (fig. 6)‘ a lesser extension of the muscle,
which usually exists as the pars labialis and does not reach a line
drawn from the corner of the mouth to the auditory organ.

I have arranged these three developmental stages after Loth,
as follows:

JAPANESE CHINESE NEGRO EUROPEAN

(KUDO) (KuUDO) (LOTH) (KUDO)
HFONg 58 22.0 GES Le REET. 5 4 18 0
Mediumeee ones ees 4 0 6 4
\GEy eee traps dia are on ey eee Ten 6 2 2 3
Number of half faces...... 15 6 26 7

The radiating form of the cheek division of the platysma is
subject to many variations. These have been arranged by
Bluntschli-Loth in the following series of types:

I. Usually transverse course of the upper platysma-fibers
(fig. 8).

Ia. Weakening of type I.

The above two types belong to a primitive group.

II. Acquisition of an ascending direction of the fibers with loss
of the transverse. Pars aberrans ascendans is well developed
(figs. 1, 4, 6, and 7).

Ila. Pars aberrans is lost or only weakly developed.

III. Strengthening of pars labialis after loss of pars aberrans
(fig. 5).

IIIa. Weak development of the platysma, the fibers of which
hardly extend over the edge of the mandible.

2 JII, JIV, JVI, JXI, JXV, and CII, CII.

8 JVII, JIX, JX, JXV, and EIII, EIV, EV.
‘JI, JIL, JIV, JV, JX, FXII, CI, and EI, EII.

FACIAL MUSCULATURE OF THE JAPANESE 641

However these types grade into one another and there are
other possibilities of platysmal relationship. The divisions are
therefore more or less arbitrary.

JAPANESE| CHINESE ROE AN NEGRO
(KUDO) (KUDO) (LOTH)
(Birkner!) | (Kudo)
As Acvew I. Sapa ane Ree, Ob neem neer ats oO 0 0 0 0 5
yg OY St) 0s caches ee as cea eke Ca 2 0 0 0 5
“ID NWf OS) LIL ooie a aeeaied Nectar eR eo WE 6 + 2 0 12
‘TPS GOTE LL Ws ete Sem AUS Poe ae irae eet aE 5 2 1 4 2
‘TDN 0 (eV UH Een Teen ey EO aa 1 0 0 2 12
1 syq a eye} Ed Wade ar eitsia kus Ae ae an ee 1 0 0 1 0
Number of half faces............. 15 6 3 a 36
9

1 After Birkner’s figures.

Loth contends that the primitive relationships, which recall
the primates, are relatively abundant in the negro. He calls
special attention to the fact that types IIa, and IIIa are present
in large numbers in Europeans. Likewise it is not exceptional
that four of the seven half faces of my Europeans belong to type
Ila.

In the Mongolian types II and Ila are common. It may be
noted here that the pars aberrans ascendens directed upwards is
often bent forward over the cheek region like a bow and in two
Japanese it divides the platysma-risorius (see also M. risorius).

In its distribution the platysma is variously related to the
adjacent muscles. In the negro the interweaving of the muscle
with the M. triangularis, zygomaticus, quadratus labii sup., or-
bicularis oculi, etc., has been observed many times. The muscle,
in the Mongolian, is always covered with the triangularis and
usually with the fibers of the risorius (figs. 1, 2, 3, and 7). In
two Japanese and two Chinese (fig. 3) the platysma reaches the
zygomaticus, where the fiber ends of the former cover those of
the latter. Furthermore, in a Chinese (CII, fig. 7), the platysma
reaches to the splitting up of the bundles of the orbicularis oculi;
in a Japanese (JVI), beyond that point.

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

642 TOKUYASU KUDO

The cervical region of the platysma has been observed in three
Japanese (figs. 1, 2), a Chinese (fig. 7), anda European. In these
Chinese (fig. 7) the cervical bundles are not directly connected
with the cheek region of the platysma. In this (CII), as fig. 7
shows, the neck fibers of the platysma first run ventral from the
front, then, forming an open angle above the ear, arise from the
anterior region dorsad, and finally interweave with the radiating
bundles of the orbicularis oculi.

Under the chin the fibers of either side cross over to the oppo-
site side in three Chinese heads. I found this crossing to be the
rule in the Japanese cadavers.

Fig. 1. Male Japanese VI, fifty-one years old. Platysma well developed;
pars aberrans ascendens rises above the line from the angle of the mouth to the
auditory meatus. The risorius is spread over the platysma in several strands;
intermediate strands occur between this and the triangularis. The zygomaticus
is penetrated at its insertion by the caninus and split into a superficial and a
deeper layer of the quadratus labii superioris; the three heads together form a
nearly continuous whole. The caput zygomaticus, at its origin, is close to the
radiating bundles of the orbicularis oculi anterior. The bundles which turn
downward in the middle of the course of the caput zygomaticus, pass below the
zygomaticus; those which extend from the orbicularis oculi are well developed
at the lateral (especially the superior lateral) and median inferior portion of the
muscle. The M. occipitatis is moderately broad; its fibers are transverse. A
proper M. transversus nuchae lies under the posterior bundles of the occipitalis
at the same level as the lowest belly of the auricularis posterior; other fibers,
which may be the remains of the platysma bundle run close beneath, obliquely
downward and forward. The larger auricularis posterior is divided into three
parts and, in an almost fleshy condition, reaches the ear cartilage.

Fig. 2. Chinese I. The platysma is poorly developed and runs almost trans-
versely in front; the pars aberrans is lacking. The triangularis is covered at
its origin by risorius and platysma fibers. The risorius is distinct from the
triangularis. The zygomaticus is divided into two portions which pass into
each other; the upper part extends close to the caput zygomaticus of the quad-
ratus labii superioris and at its insertion, becomes lost in the caninus; the lower
portion extends downward toward the triangularis. The caput zygomaticus of
the quadratus labii superioris connects with the caput infraorbitale. The
orbicularis oculi forms a well-closed ring, although the lateral bundles enter the
caput zygomaticus. The lateral short strands of the frontalis are extended
toward the temple. The lower bundle of origin of this muscle joins the upper
lateral part of the orbicularis ring. The M. occipitalis spreads out fan-like and
reaches the anterior fibers of the auricularis superior. The auricularis superior
and anterior is fibered vertically; the auricularis posterior is plainly separated
into two long bellies. The transversus nuchae is lacking. The cervical part of
the platysma is directed obliquely over the insertion of the sternocleidomastoid.

FACIAL MUSCULATURE OF THE JAPANESE 643

Eph

Figure 1

Fig. 3. Chinese III. On the whole, the facial musculature has quite firm
fibers. The platysma forms a single mass; the pars aberrans, divided like a V,
extends beyond the line from the angle of the mouth and the auditory meatus;
the posterior portion reaches the lateral marginal bundle of the orbicularis oculi,
and the anterior passes, beneath the orbicularis, beyond the zygomaticus major
and minor. The M. triangularis is broad at the edge of the jaw. The risorius
shows considerable development; the zygomaticus is penetrated at its insertion
by the insertion bundles of the risorius and triangularis. The three heads of the
quadratus labii superior, in part covered by the orbicularis ring, form a con-
tinuous muscle. The coarsely bundled orbicularis oculi is radiately extended
at its upper lateral part. Noteworthy are the large parallel fibers which have
lost their connection with the auricularis and course over the temple in the same
direction as the latter. Finally, oblique fibers are directed toward the frontalis;
father forward a transversely fibered extends toward the front under the in-
sertion of the frontalis. All such thin-fibered sheets may belong to the auriculo-
frontalis.

644 TOKUYASU KUDO

Figure 2

645

FACIAL MUSCULATURE OF THE JAPANESE

646 TOKUYASU KUDO

M. mandibuli-marginalis Blundschli (M. menti accessorius
Kelchi) (fig. 6)

This muscle is found in Europeans only as an infrequent vari-
ation (Kelechi, Wood, Henle, Testut, Ruge, Seydel, Blundschli).
Ruge has observed the muscle as a variety arising from the
platysma, while Seydel considers it the remnant of a sphincter
colli superficialis.

I have encountered the muscle twice in the Japanese (in J VIII
and J XII) in fifteen half faces. Thus it appears to be no rarity
in the Japanese. In the latter (fig. 6) many isolated fibers ex-
tend from the prolongation of the triangularis toward the ear.
They lie superficially on the cheek portion of the platysma and
cross that muscle. In the former two separate muscle strands
radiate from the margin of the jaw bone upward in the form of
a weak bow, concave on the anterior face. In a Chinese head
Birkner has seen similar compact muscle strands directed toward
the ear.

M. triangularis (figs. 1 to 7)

In Mongolians this muscle is mostly fan-shaped behind, below,
and in front. As far as can be judged in my material, it is better
developed in the Chinese than in Europeans. Likewise Birkner
found this muscle well developed in three Chinese heads; the
same is the case in the negro (Flower and Murie, Hamy, Hart-
mann, Chudzinski, Popowsky, Eggeling and Loth) and in the
Hottentots (Fetzer).

In the Japanese and Chinese (figs. 3 and 7) the triangularis is
commonly associated with the risorius; in Europeans, on the
other hand, the muscle is usually isolated (four out of five heads).
I have seen the well-known variations of the insertion ends of
the fibers (Hisler), namely, the transition of bundles into the M.
caninus, in five Japanese, one Chinese and three European heads;
the transition into M. zygomaticus in one Chinese and seven
Japanese; the transition into M. orbicularis oris sup. in nine
Japanese, two Chinese, and two European heads. I have never
met with a tendinous interruption of the muscle at the angle of

PE

FACIAL MUSCULATURE OF THE JAPANESE 647

Figure 4

Fig. 4. Female Japanese XII, thirty-two years old. Platysma well de-
veloped; pars ascendens arcuate forward and extends to the zygomatic. The
zygomaticus is easily distinguished from the caput zygomaticus of the quadratus
labil superioris. The insertion of the zygomatic is separated by the caninus
into a superficial and a deeper layer. The risorius consists of only two arcuate
muscle bands and hes over the insertion of the triangularis. The three heads
of the quadratus labil superioris are comparatively easily distinguished; the
caput zygomaticus at its origin radiates toward the temple. The diverging
marginal bundles of the orbicularis oculi, like the upper lateral part, are es-
pecially large in the lower medial portion; the latter pass over the infra-orbital
head of the quadratus labii super. to the caput zygomaticus. The pars trans-
versus of the M. nasalis is present. The occipitalis is weakly developed and
short-fibered. The M. auricularis posterior has two bellies. The long cervical
portion of the platysma passes forward in a light curve over the insertion of the
sternocleidomastoid under the ear.

648 TOKUYASU KUDO

the mouth. The fibers of the triangularis, already mentioned,
which run into the orbicularis bundle have frequently been found
before in the negro (Chudzinski, Popowsky, Loth). The in-
sertion of the muscle in Mongolians is usually at the upper
median part of the risorius bundle (figs. 1 to4and 6). I have not
observed the division of the triangularis into several large por-
tions (Macalister) in the Mongolians.

In order to show the extent of the muscles, the following meas-
urements (in millimeters) have been inserted, although they are
merely of general interest.

1. The breadth of the muscle at the lip commissure, according
to Chudzinski, is 15 (12 to 19) in the negro on the average; 10
(6 to 15) inthe Japanese (fifteen individuals) ; on theright side 6, 10,
12, on the left side 9, 12 in the Chinese (three individuals) ;
on the right side 7, 11, 14, on the left side 7, 8, 9, 10 in Euro-
peans (five individuals) according to Kudo, according to Chud-
zinski 11 on the average.

2. The breadth of the muscle at the place where the risorius
diverges, is 20 (10 to 33) in the Japanese; on the right side 15,
25, 20, on the left side 10, 27 inthe Chinese; on the right side 14,
20, 20, on the left side 13, 19, 15 in Europeans; 11 to 15 in the
Hereros (Eggeling).

3. The breadth of the muscle in the region of radiation is 38
(29 to 55) according to Chudzinski, 37 (31 to 438) according to
Eggeling and Loth in the negro; 45 (30 to 60) in the Japanese;
on the right side 40, 65, 45, on the left side 35, 70 in the Chinese;
on the right side 31, 35, 44, on the left side 30, 25, 40, 35 in the
European, and according to Chudzinski an average of 38.

Fig. 5. Male Japanese III, forty-four years old. The pars aberrans as-
cendens of the platysma is lacking. The risorius consists of two arcuate bundles.
Zygomatie separated at its insertion into a superficial and a deeper layer by the
caninus. The caput infraorbitale is separated in the region of origin from the
other two heads. The lateral marginal bundles, and especially the upper, are
developed. The M. frontalis extends farther up, its lateral fibers are short and
arcuate toward the ear. The auricularis is spread out like a fan; its posterior
fibers are transverse; the anterior are interrupted for a distance. The occipitalis
is strongly developed; its hinder fibers are more vertical, the anterior are in-
clined forward, interlaced with the auricularis superior, and finally reach the
conch in a fleshy condition. The auricularis posterior is divided into two
portions, the lower of which undergoes a tendinous interruption.

649

FACIAL MUSCULATURE OF THE JAPANESE

ig|

Figure 5

650 TOKUYASU KUDO

4. The divergence of the median muscle edge from the middle
of the chin amounts to 14 (9 to 19) in the negro (Chudzinski) ;
13 (5 to 20) in the Japanese; on the right side 6, 6, 12, on the left
side 12, 14, 15, 9 in the Chinese; on the right side 6, 10, 9, on the
left side 12, 14, 15, 9 in the European, and according to Chud-
zinski an average of 12.

M. risorius (figs. 1 to 8)

Opposite the risorius of Santorini, which arises from the trian-
geularis, Ruge has distinguished a platysma-risorius, in which
Forster and Loth support him. This is said to arise by a con-
tinuous separation from the arcuate platysma bundles extending
over the cheek toward the angle of the mouth. Such a dis-
tinction, however, appears to have no significance, since there
are cases where actual confirmation is difficult, even impossible.
Such a recognition would be especially difficult in the Mongolian,
because here the platysma bundle is directed very generally
toward the corner of the mouth, covering the risorius. Still the
typical platysma-risorius is occasionally present, as I have seen
it in at least three cases in Japanese preparations. These three
cases are grouped according to the arrangement of Blundschli,
one (JIV) of type VI, 2 (JV and JX) of type IV.

In the Mongolians the M. risorius is recognized, as is the tri-
angularis, by its stronger development. My material allows the
following arrangement. Observations on fifteen half faces of
Japanese:

1. Risorius lacking (fig. 8) : twice in Japanese, once in European.

2. Risorius is weakly developed and entirely isolated from
the triangularis (fig. 4): four times in Japanese, two times in
Europeans.

In JIII, JIV, and EI, EIII, platysma bundles are inserted
between M. risorius and triangularis and have a convergent
course, together with the former, toward the corner of the mouth.
They may function possibly as a risorius.

3. Risorius is spread out radially in the region of its origin;
several intermediate bundles are present between risorius and tri-
angularis (figs. 1 and 6): 6 times in Japanese, 2 times in Europeans,
once in Chinese.

FACIAL MUSCULATURE OF THE JAPANESE 651

4. Risorius spreads out into a fan pattern, forming together
with the triangularis a closed muscle plate (figs. 3 and 7); 3 times
in Japanese, 2 times in Chinese, none in Europeans.

Birkner is also struck by the more powerful development of
the risorius. According to the account of Chudzowski, this
muscle in the negro is also more markedly developed; yet, as
Loth remarks, this author has not seen the true risorius. Egge-
ling has found the risorius better developed but once in five
Hereros.

The M. risorius is more rarely lacking in the Mongolian than
in the black race. I missed the muscle only twice in fifteen
Japanese heads, not once in my three and Bickner’s three Chinese
heads. In the negro Loth enumerated fifteen cases out of thirty-
five half faces, in which the muscle was entirely lacking (thus
about 43 per cent). In the Hottentots Fetzer usually found the
muscle lacking; that is, he found the muscle on both sides in only
four heads out of fifteen Hottentots and on one side in only two
more. Unfortunately, we have no account which tells us how
often the muscle is lacking in Europeans.

Blundschli divides the muscle into six types. The compiling
of this classification is rendered difficult, however, by various
transition forms. The approximate frequency of the types is
brought together in the following table:

NEGRO JAPANESE | CHINESE EUROPEAN
(LOTH) (KUDO) (KUDO) (KUDO)
‘TA 81S t al Uaioe arrears As tetak ae Reenter 15 2 0 1
Ey ote Ih Renters acca a aaa 5 2 0 0
‘TENG ofS) 1) OB LE doe ices tet Goa Nea 10 1 1 1
HL ated INOS cata cenceet eee Paar ice teectees 5 3 2 1
1
IM OS Nic oer ree een eer tre 1 a 1 2
ALN SX Wares researc eer 0 3 1 0
Number of half faces...... 36 15 6 5)

652 TOKUYASU KUDO
M. transversus mentt

With regard to the frequency of this muscle, according to
accounts of Thiele, Schmidt, and LeDouble, fifty-six cases (60
per cent) in ninety bodies have been found with the muscle
present (Hisler). But the number is too high according to
Eisler, because it encloses radiating bundles of the platysma and
triangularis. The muscle occurs eight times out of twenty-one
heads in the negro, according to the computation of Loth. Fetzer
failed to find the muscle only once in seventeen Hottentot heads,
twice it was only rudimentary. In the Japanese and Chinese
the muscle was present without exception; but twice it was only
vestigial. Birkner found the muscle readily demonstrable in the
Chinese head. It is evident, then, that the muscle is more
generally present in the Mongolians and Hottentots than in the
Europeans and negro.

The relationship of the transversus to the triangularis and
platysma is variable. In the Japanese the muscle is at times
derived entirely from the more anterior triangularis fibers, which
run in an arcuate manner along the submental region; more
frequently (six Japanese heads) the triangularis fibers pass over
in part into the transversus. Eggeling has observed a similar
condition twice in the negro. The muscle is often entirely iso-
lated, or occasionally inserted at both ends into the edge of the
under jaw through the platysma fibers. This muscle has been
observed twice in the negro (Eggeling and Loth) without a con-
nection with the triangularis.

The development of the muscle is computed by its breadth,
which on the average amounts to: 7 mm. in the negro (Eggeling
and Loth); 5 mm. in the Japanese; 3, 4, 5, 4, 2 mm. in the Euro-
pean, and 3, 5, 3 mm. in the Chinese (Kudo).

M. quadratus labii superior (figs. 1 to 8)

All authors state that this muscle is less frequently divided
into three portions among colored races than in Europeans
(figs.6and8). Fetzer, on the other hand, notes no peculiarities
among Hottentots.

FACIAL MUSCULATURE OF THE JAPANESE 653

In the Mongolians the differentiation of the muscle parts is
very slight (figs. 1, 3, 4, 6, and 7). Birkner has observed an ap-
parently united quadratus in a Chinese. In the negro the three
distinct portions appeared in only three out of twenty-six indi-
viduals. I have found this muscle divided into three parts
three timesin the Japanese. Asarule, among Japanese the muscle
likewise is fused into an entire muscle plate, so that the super-
ficial layer, caput angulare and caput zygomaticus, and especially
near the origin, the caput infraorbitale, covers it; if the former is
more strongly developed, then the latter is entirely covered by
it (in JII, JXIV). In the following lines the individual muscle
portions are considered:

The caput zygomaticum (M. zygomaticus minor) is well de-
veloped and is never lacking in the Mongolians. In the negro
(Chudzinski, Loth) the caput was absent in two out of forty-
eight half faces (7.7 per cent), while, on the other hand, in
Europeans (LeDouble) it was lacking twenty-two times in 100
individuals (22 per cent). The separation of the caput from the
M. orbicularis oculi is usually difficult, since the radiating bundles
of the latter spread out over the surface of the so-called caput;
only in the case of three Japanese was the head of the muscle
perfectly distinct.

The caput infraorbitale (figs. 1, 4, 5, 6, and 8) is strong in the
Mongolians, coarse-fibered, and usually covers the caput zygo-
maticum (figs. 3 and 6). The caput in the superficial layer in
the case of two Chinese is strengthened by means of the orbicu-
laris bundle (fig. 7); but I have never observed a caput divided
into two distinct portions in the Japanese. The breadth at the
point of origin is as follows:

Japanese (Kudo) ca. 16 (10 to 26) mm.

Europeans (Kudo) right 12, 26; left 12, 22, 21, 16 mm.
Europeans (Chudzinski) 22 mm.

Chinese (Kudo) right 23; left 25 mm.

Negro (Chudzinski) 32 (22 to 39) mm.

The caput infraorbitale is difficult to isolate from the other
two capita in the case of the Mongolians (figs. 1, 6, and 7). The
well-separated cases are divided into the following ratio:

654 TOKUYASU KUDO

Japanese (Kudo) two times in fifteen half faces (13.3 per
cent).
Chinese (Birkner and Kudo) none in six half faces.
European (Kudo) three times in five half faces.
Negro (Loth) four times in forty-seven half faces (8.5 per
cent).
The small muscle which is figured in atlases or in the text-
books has been observed by me only once in the Mongolians.
The caput angulare (figs. 1, 4, 5, 6 and 8) may have its origin
above the hgamentum papebrale mediale. Its connection with
the adjacent muscles is apparently intimate; the union with the
frontalis has often been observed. The caput is so intimately
connected with the orbicularis oculi that it not only unites with
it, but also receives superficial bundles from it.

M. orbicularis oris (figs. 1, 4, 5, 6, and 8)

Birkner has observed a well developed muscle in the Chinese.
In the negro the muscle is well developed, in connection with
the thick lips; according to Eggeling, even curved, with its free
ends respectively somewhat outward and upward in the upper
lip, downward in the lower lip.

In the Japanese and Chinese a stronger development as con-
trasted with the Europeans has not been demonstrated. The
muscle appears here as a plate with nearly parallel fibers which,
as in Europeans, has an anterior position. In the negro the
orbicularis is more fully developed in the under lip than in the
upper (Giacomini). I have not been able to establish a clear
distinction of development between the upper and the lower
lip.

M. quadratus labu inf. (figs. 1, 3, 4, 5, 6, and 8).

In the preparation of the superficial layer this muscle shows no
noteworthy differences between the Japanese and the Europeans.

M. zygomaticus

In the Japanese and Chinese this muscle forms a relatively
powerful strand and often grows together with adjacent muscles

FACIAL MUSCULATURE OF THE JAPANESE 655

(figs. 6 to 8). As to the development of the muscle, its breadth®
at different muscle levels is arranged as follows:

Breadth at origin:

Japanese (Kudo) 6.4 (5 to 10) mm.
Chinese (Kudo) right 9, 8, 10, left 8, 9, 5 mm.
European (Kudo) right 7, 7; left 7, 10, 10, 5, 5 mm.
European (Chudzinski) average 8 mm.
Negro (Chudzinski) 10.5 (8 to 18) mm.
Breadth in the middle:
Japanese (Kudo) 7.7 (6 to 14) mm.
Chinese (Kudo) right 8, 12, left 11, 9 mm.
European (Kudo) right 9, 10; left 8, 9, 11, 5, 5 mm.
European (Chudzinski) average 7 mm.
Negro (Chudzinski) 10.4 (6 to 14) mm.
Breadth at lip insertion:
Japanese (Kudo) 10.5 (6 to 20) mm.
Chinese (Kudo) right 10, 10, 12; left 11, 9, 6, mm.
European (Kudo) right 18, 14; left 11, 9, 6 mm.
European (Chudzinski) average 25.35 mm.
Negro (Chudzinski) 26 (13 to 32) mm.

This muscle is never lacking in the Mongolians which have
been investigated, and is very seldom absent from the Europeans
(Otto, Macalister). The division into two (Macalister), three
or four portions (Chudzinski, Popowsky) has not been observed
in the Japanese and Chinese; but the eccentric orbicularis bundles
often merge with this muscle (JV, JVI, JXI, and CIT) (figs. 1, 6,
and 7), just as Birkner has observed in a Chinese head. In a
Japanese head (JVII) the orbicularis bundles entering the muscle
turn upward and mesad into the orbicularis ring. In a Chinese
and in European heads the muscle is entirely separate from or-
bicularis fibers (figs. 2 and 8).

The muscle is often combined (in JIX, CII, EIT) with the
platysma bundle; it is also covered by it. The direction of the
fibers is more crosswise or transverse. Chudzinski has measured
the space between the head of the muscle and the anterior end

® Measurement is the breadth of the deeper head portion of the muscle where
it is covered with the orbicularis bundles.

656 TOKUYASU KUDO

of the porus acusticus externus. The data secured from this are
as follows:
Japanese (Kudo) 50.9 (44 to 66) mm.
Chinese (Kudo) right 48, 40, 50; left 50, 47, 55 mm.
European (Kudo) right 50, 51; left 53, 55, 51, 54 mm.
European (Chudzinsk1) average 47 mm.
Negro (Chudzinski) 40.3 (33 to 46) mm.

The relation of the zygomaticus in the region of insertion
affords great interest. A short distance from its insertion into
the lip it is penetrated by the caninus, so that at its end it is split
into a superficial and a basal layer (figs. 1, 4, 5, and 7). Such
a condition was found three times in the Japanese, only once in
the Chinese and Europeans. An insertion, with the superficial
layer the stronger, has been found in three Japanese and one
Chinese; one with a weaker superficial layer in three Japanese.
The superficial layer may be lacking; the muscle end crowds its
way likewise deeply toward the buccalis (in a Japanese, a Chinese,
and a European, fig. 8). The absence of the deeper portion has
has been established in three Japanese and one European heads.
However, the muscle is always penetrated more or less by the
caninus.

A well-isolated zygomaticus is very rare in the Japanese, just
as in the Chinese. The following table concerning this may be
instructive:

JAPANESE (KUDO) NEGRO (LOTH) EUROPEAN (KUDO) CHINESE (KUDO)
15 half faces 45 individuals 5 half faces 3 half faces
4 (26.7 per cent) | 11 (23 per cent) 1 0

Loth believes that the fusion of the muscle with the contig-
uous structures takes place only exceptionally in Europeans.

M. orbicularis oculi (figs. 1 to 6 and 8)

This muscle consistently shows a moderately strong devel-
opment in the Mongolians. It forms a powerful broad ring
around the eye. In stating the strength of the muscle, up till

FACIAL MUSCULATURE OF THE JAPANESE 657

now, its breadth from the edge of the eyelid has been taken as a
criterion by authors. The following tables give the result of such
measurement:

1. Distance of the middle of the upper lid margin from the
upper edge of the muscle (with the eye closed) :

Negro (Chudzinski, Popowsky, Eggeling, Loth), 28.2 mm.
Japanese (Kudo), 27.2 mm.

Chinese (Kudo), right, 30 mm; left, 24, 25 mm.
European (Kudo), left, 22, 30, 29, 30, 19 mm.

European (Chudzinski), 16 mm.

2. Distance of the middle of the lower lid margin from the

lower edge of the muscle (with the eye closed):
Negro (Chudzinski), 25.5 mm.
Japanese (Kudo), 29.6 mm.
Chinese (Kudo), right, 30, 37, 35; left, 30, 32, 35 mm.
European (Kudo), right, 28; left, 29, 23, 25, 37, 28 mm.
European (Chudzinski), 26 mm.

3. Distance of the corner of the eye from the outer edge of the
muscle (which can be computed quite exactly, because the contour
of the muscle through the eccentric orbicularis bundles or the
growth with contiguous muscles) is very little modified.

Negro (Chudzinski), 26.1 mm.

Japanese (Kudo), 31.1 mm.

Chinese (Kudo), right, 30, 30, 30; left, 31, 32, 35 mm.
European (Kudo), right, 28, 33; left, 29, 14, 27, 23 mm.
European (Chudzinski) 22 mm.

From this table is it apparent that the muscle in the Mongolian
is broader in the lateral portion; the upper orbicularis portion is
smaller than the under; in the negro the relationship is reversed.

According to Chudzinski, the muscle is more powerfully de-
veloped in the negro than in the European; likewise in the Hot-
tentots. This is apparently the case also in the Mongolian. In
the primates the muscle element situated over the edge of the
orbit is generally weakly developed (Ruge). I have directly
computed the distance of the outer edge of the pars orbitalis
from the edge of the orbit; the result obtained is about the same
as the measurement at the edge of the lid as given above.

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

658 TOKUYASU KUDO

The scattering bundles at the margin of the orbicularis radiate
in very different ways in the Mongolian; but strongly developed
bundles are relatively rare at the margin of the eyelid. For com-
parison with the tables of H. Virchow and Loth I separate the
bundles into the following groups:

1. Upper lateral marginal bundle. A =the one directed dorsad
(mesad). B=the one directed ventrad (lateral).

HOTTENTOT

JAPANESE (KUDO) CHINESE (KUDO) | EUROPEAN (KUDO)| NEGRO (LOTH) (FETZER)

14 half faces 4 half faces 5 half faces 38 half faces | 17 individuals

13 (92.9%) 20 (52.5%) | 14 (82.4%)
fA.18 3 aa - ne
\B.0 B.0 B.0

2. Lower lateral marginal bundle. A =the one directed dorsad
(laterad). B=the one directed ventrad (mesad).

HOTTENTOT

JAPANESE (KUDO) | CHINESE (KUDO) | EUROPEAN (KUDO)}| NEGRO (LOTH) (FETZER)

14 half faces 4 half faces 5 half faces 88 half faces | 17 individuals
10 (71.4% 33 (86.9%) 10 (58.8%)

A. 0 (Aun Aya
es 10 : = 3 - i. 1
3. Lower median marginal bundle. A =the one directed dor-
sad (laterad). B=the one directed ventrad (laterad).

HOTTENTOT

JAPANESE (KUDO) | CHINESE (KUDO) | EUROPEAN (KUDO)| NEGRO (LOTH) (FETZER)

14 half faces 4 half faces 5 half faces 88 half faces | 17 individuals
6 (42.9%) 29 (76.38%) 15 (88.2%)

PAC eek JA. 0 A SEL
a 5 = 0 2 a 1
4. Upper median marginal bundle. A =the one directed dor-
sad (laterad). B=the one directed ventrad (mesad).

HOTTENTOT

JAPANESE (KUDO) CHINESE (KUDO) | EUROPEAN (KUDO)} NEGRO (LOTH) (FETZER)

14 half faces 4 half faces 5 half faces 38 half faces | 17 individuals
4 (28.6%) 6 (15.8%)

A. 4 ag Ieee fA. 0
i 0 ; it 0 z ee 1 0{R 0

FACIAL MUSCULATURE OF THE JAPANESE 659

This shows that the outermost bundles at the lower medial
margin, in Japanese and Chinese, are, as a whole, more strongly
developed and often consist of larger fibers than in Europeans.
At their origins, the bundles, together with the depressor super-
cilii, often extend over the ligamentum palpebrale mediale; also,
united with the caput angulare laterally below, they extend
farther than the fasciculi deflexi toward the M. zygomaticus
major (JVII, fig. 7) or minor (JVIII). According to Fetzer, the
lower median bundle is usually present in Hottentots; this appears
to be the case not infrequently in Europeans.

The adjacent orbicularis, well distinct from the contiguous
muscles in Europeans (Loth), is only distinguished with diffi-
culty in Mongolians, as also with the negro (Loth, compare fig.
8 and figs. 6 and 7) especially on the upper and lower margins
the boundary of the muscle is not distinct. ‘The muscle often
unites with the quadratus labii superius and zygomaticus to a
compound muscle (fig. 7).

M. depressor supercilir

This is always found as a triangular muscle above the liga-
mentum palpebrale mediale in Japanese and Chinese; no dif-
ferences have been observed between Mongols and Europeans.
- Loth has observed it considerably developed in a negro.

M. corrugator supercilit

This muscle, which comparative anatomy regards (Ruge) as
an offshoot of the bundles of the orbicularis oculi occurring above
the cleft of the lids, is frequently, both in Mongolians and Euro-
peans (Macalister), not separated from the orbicularis oculi.
Similar relations have been shown by Popowsky and Loth for the
negro. According to Eisler, this muscle is readily distinguished
from the frontalis by its coarse bundle formation. In the Mon-
golian the muscle, together with the depressor, is generally fused
with the frontalis, as in the negro (Chudzinski). It is seldom
lacking in Europeans (Macalister) ; in the negro once out of five
in the individuals (Eggeling). I have always found it in my
material.

660 TOKUYASU KUDO
M. nasalis

In general this muscle appears to be as well developed in Mon-
golians as in Europeans. The development of the pars trans-
versa (figs. 1 to 6 and 8) fluctuates widely. Thus size and shape
of the nose are no criteria for the development of this muscle.
According to Eggeling and Loth, it is well developed in the negro.
Measurements of the greatest width of the pars transversa follow:

Negro: 12.5 mm. (Flower and Murie) ; 15 mm. (Loth).
Japanese: 12 mm. (Kudo).

Chinese: Left, 10, 10, 12 mm. (Kudo).

Europeans: Left, 10, 9, 12 mm. (Kudo) ; 11 mm. (Chudzinski).

Muscles of the two sides which meet in the middle line on the
bridge of the nose have been observed in few instances in the
negro, while I have met this connection frequently in Japanese.
This muscle farther shows an intimate relation to other muscles
in the yellow races. Frequently (eight times in fifteen half faces
of Japanese) it passes over into the superficial fibers of the M.
procerus. Eggeling has found the same in two cases out of five
Herero heads. The muscle is also comnected laterally with the
caput angulare.

I find the pars alaris weakly developed, and frequently it could
not be distinguished. Still it is possible that it had been removed
in the preparation. According to Macalister, this part of the
muscle can be absent in Europeans.

M. frontalis (figs. 2, 3, and & to 8)

Several authors have described a strong development of this
muscle in negroes and Hottentots (Eisler, Loth, Fetzer). It is
also well developed in the Japanese. The following measure-
ments give an idea of the extent of this muscle:

1. The depth of the muscle fibers in the middle line, from the
base of the nose outward

a. Measured as one banded group:

Japanese 81.8 (73 to 122) mm. (Kudo).
Chinese: right 75, 77 mm.; left 75 mm. (Kudo).

FACIAL MUSCULATURE OF THE JAPANESE 661

European: left 50 mm. (Kudo); 82.5 mm. (Chudzinski).
Negro: 78 mm. (Loth); 78 (62-93) mm. (Chudzinski).
b. Measured in a curve:

Japanese 76.1 (50-100) mm. (Kudo).

Chinese: right 71, 71 mm. (Kudo).

European: left 47 mm. (Kudo).

Negro: 60 mm. (Eggeling’s 4 Hereros); 50 mm. ? (Livini);
57 mm. (Loth, 2 negroes).

The muscle is shortest in the median line and becomes in-
creasingly deeper laterally so that the passage of muscle fibers
into the galea forms an obliquely arcuate line (figs. 2, 3, 6, and 7).

2. The greatest breadth of the muscle, about perpendicular
to the course of the fibers, is about 55.4 mm. in the Japanese.

A complete separation of the muscle into two halves has never
been recognized in the Japanese and Chinese, likewise not in
seventeen Hottentot heads, according to Fetzer.

M. procerus nasi (figs. 1, 2, 3, and 6 to 8)

This muscle is always well developed in the Mongolians; in
the Europeans it may be occasionally wanting in one or both
sides (Harrison, Macalister, Le Double). In the Japanese there
is occasionally a considerable inequality of antimeric muscles.
The passage of the muscle into the frontalis is the rule in
Mongolians according to the literature, a complete separation
appears opposite the frontalis; but in the European it seldom
occurs. The facts in the case of the negro are not clear. In
the Mongolians the superficial fibers of the muscle may reach
the pars transversa of the M. nasalis farther up on the nasal
cartilage, as is true of the negro, according to Loth. The direct
junction of the procerus with the caput angularis, which seldom
occurs in Europeans (Hisler), has been observed only twice in
negroes (Loth), while I have come across it four times in the
Japanese. .

The antimeric muscles are hard to separate in the Mongolians.
According to Chudzinski this muscle in negroes is on both right
and left sides.

662 TOKUYASU KUDO

The breadth muscle of the two sides amounts to 10 mm. in
Japanese; the smallest breadth is as follows:
Japanese: 6.6 (4 to 10) mm. (Kudo).
Chinese: right 4, 6 mm.; left 4, 5 mm. (Kudo).
European: left 3, 3 mm. \Gaadove 70.4 mm. (panda
Negro: 5.4 os to 9) mm. (Chudzinski.)

M. occipitalis (figs. 1, 2, and 4 to 8)

This muscle is represented in Mongolians by a strong, coarse-
bundled plate, which has varying outlines—often elongate quad-
rangular, more commonly triangular or lunate or somewhat
notched. Chudzinski has not given accurate details for his
material as to the measurements of this muscle. I have taken
measurements as follows:

Depth (in front*):

Japanese (Kudo): 25.1 (12 to 37) mm.
Chinese (Kudo): right, 19, 18; left 18, 21, 21 mm.
European (Kudo): right, 21, 14; left, 16, 20, 22, 27 mm.
European (Chudzinski) : 21 and 22 (17 to 28) mm.
Depth (behind) :
Japanese (Kudo): 28.5 (16 to 46) mm.
Chinese (Kudo): right, 18, 23, 30; left, 23, 30, 23 mm.
European (Kudo): right, 18, 23, 30; left, 16, 24, 25, 24,
23 mm.
European (Chudzinski): 27.7 mm.
Greatest depth:
Japanese (Kudo): 34.6 (23 to 53) mm.
Chinese (Kudo): right, 33, 37; left, 31, 33, 36 mm.
European (Kudo): right, 29, 38, 33, 46; left, 29, 35, 31
23mm.
European (Chudzinski) : 32.6 mm.
Negro (Chudzinski) : 37 (34 to 47) mm.

6 T have measured the depth of the front and hinder sections of the muscle at
about 5 mm. from the marginal bundles of either side and parallel to the course
of the fibers.

FACIAL MUSCULATURE OF THE JAPANESE 663

Upper breadth’:
Japanese (Kudo): 63.6 (50 to 78) mm.
Chinese (Kudo): right, 76, 67; left, 64, 85, 66 mm.
European (Kudo): right, 68, 64, 77; left, 64, 75, 68, 56,
66 mm.
Negro (Chudzinski): 71.4 (61 to 81) mm.
Lower breadth’:
Japanese (Kudo): 56.3 (53 to 72) mm.
Chinese (Kudo): right, 64, 90; left, 61, 65, 86 mm.
European (Kudo): right 56, 60, 68; left 52, 53, 60, 45,
73 mm.
European (Chudzinski) 65 mm.
Negro (Chudzinski) : 63.8 (43 to 96) mm.

As a vestige of the auriculo-occipitalis, which is well developed
in apes, the anterior section of the occipital fibers toward the
ear (namely, the M. auricularis posterior), shows a changing
condition. In the posterior position of the muscle bundles
are arranged more nearly vertically than in the anterior portion,
and, directed obliquely upward and forward, paralleling each
other, lie close together. The more anterior bundles, which
only in a single Japanese (JX) course almost vertically, incline
strongly forward. They are almost transverse in five cases
(fig. 1), even obliquely ventral (in a Chinese and a European,
fig. 8). In a Japanese head the anterior muscle part is sepa-
rated by an interruption of continuity from the more posterior
one.

The anterior muscle bundles run not only parallel with the
upper margin of the auricularis posterior, but even fuse with it
(in CII, JIX, and JI). In JIII such bundles become entirely
separated from the hinder element and reach the ear (fig. 5).
In Europeans this muscle usually unites with the M. auricularis
posterior (Le Double). Eisler, on the other hand, does not con-
firm this for adults.

In seven out of 100 Europeans Austoni saw the median bundle
of the occipitalis running dorsad to and parallel with the pos-

7 The distances, respectively, between mae upper and lower ends of the marginal
bundles of the two sides.

664 TOKUYASU KUDO

terior margin of the auricularis. In negroes, Eggeling, who never
observed an approximation of occipital fibers and the auricularis
posterior, found a close approximation of this muscle to the
auricularis superior (three cases in five Herreros). The occipi-
talis fibers which reach the auricularis were often found (three)
times) in Japanese (fig. 5) and in two Chinese (fig. 7); in JI they
extend under the latter muscle.

In the negro (Chudzinski, Popowsky) the anterior bundle
reaches the otic conch. According to Eggeling, these bundles
are continued to the conch, sometimes by distinct, sometimes by
weakly developed tendinous strands. In a Japanese (JI) and a
European head (EV) I found the muscle fleshy, even till it reaches
the ear cartilage.

M. transversus nuchae (figs. 1, 4, and 7)

This muscle has many variations; in text-books and atlases it
is variously figured. Two types, however, may be distinguished.

The first type, the transversus nuchae (fig. 13) is a derivative
of the auricularis posterior, that is, a median part (posterior)
of the auricularis which becomes interrupted by the interme-
diate tendon. This type is always connected with the auricu-
laris bundles by means of transverse tendinous fibers.

The second type (M. corrugator posticus Santorini, figs. 4
and 8) has no direct genetic connection with the muscles of the
ear. It is the residue of the platysma fibers which radiate pos-
teriorly into the neck region. The ventral end of the fibers may
connect directly with the principal bundle of the platysma or
may be separated from the latter (Pabis and Ricci).

Thus two types may always be readily distinguished. The
muscle of the second type lies under that of the first, and, sur-
rounded by thick, felted subcutaneous connective tissue, runs
more obliquely caudad to the fascia parotida (M. occipito-
parotoidea Chudzinski) or to the platysma. In the Mongolians
it never arises directly nor by tendinous connections from the
bones, but the fiber ends diverge more or less over the fascia of
the neck. These differences of the two types are especially
evident in the preparations which possess the muscles of both
types, as I have observed in two individuals.

FACIAL MUSCULATURE OF THE JAPANESE 665

I have found the first type seven times in the Japanese (on an
average, the greatest length is 23.7 (11 to 32) mm.; greatest
breadth is 10.6 (4 to 18) mm.) and once in a Chinese (CII,
greatest length is 10 mm., greatest breadth is 4mm.) In Euro-
peans Schulze describes eighteen cases in twenty-five individuals,
Macalister seven in thirty individuals (23.3 per cent); but the
conclusions of the latter have not been recognized by Knott and
Le Double. The second type is rarer than the first; I found it
three times in the Japanese, in one Chinese, and in one European.

In a Japanese (JVI) the transversus fibers, which probably
belong to the first type, course along the posterior margin of the
linea nuchae suprema ventromedial; at times tendinous platysma
fibers are demonstrable, which show no connection with the
platysma.

The relative frequency of the muscle in both types is about as
follows:

EUROPEAN

JAPANESE (KUDO) CHINESE (KUDO) (LE DOUBLE, ETC.) (CHUDZINSKI, ETC.1)
14 3 89 24 individuals
7 (50%) 1 33 (386.7%)? 14 (58%)

1 Chudzinski, Turner, Hartmann, Papowsky, Eggeling, Loth.
2 LeDouble, Macalister, Schulze.

According to the number of half heads:

JAPANESE (KUDO) CHINESE (KUDO) (LE POSE A aie) lonuinamaee. ETC.)
14 half faces 3 half faces 118 half faces 34 half faces
7 (50%) iL 48 (40.7%) 19 (56%)

From this it is evident that, based on per cent, there is a rad-
ical difference between the colored and white peoples.

M. auricularis posterior (figs. 1, 2, and 4 to 8)

This muscle is rather well developed in Mongolians and occurs
in all of my material. Entire absence of this muscle is very rare
in Europeans (Macalister, Le Double). The insertion on the

666 TOKUYASU KUDO

eminentia conchae by a long or short tendon always occurs.
The origin lies behind the ear, along the neck line for a varying
distance. The greatest breadth of the longest belly averages
0.5 mm. in the Japanese. The greatest length of the same belly
is as follows:

Japanese (Kudo): 14 half faces, 30 (5 to 60) mm.

European (Kudo): right, 29; left, 26, 25, 38, 22 mm.

Chinese (Kudo): left, 36, 42 mm.

Negro (Chudzinski) : 41.1 (23 to 46) mm.

Negro (Eggeling, Popowsky, Loth) 11 half faces 40 (23 to

46) mm.

The auricularis posterior may reach to the auhnhenaene
occipitalis externa by evident growth. This apart from a con-
sideration of the separation by an intermediate tendon into
auricularis posterior and transversus nuchae.

With reference to the well-known division of the muscles into
several parts (figs. 1, 2, and 4 to 8), I have grouped them as
follows:

2. - : =
Ag EEN GLE See ae oe
ceoege | ee Cees: | ea oot
Japanese (Kudo) 14 half faces........... 6 (42.9%) | 6 (42.9%) | 2 (14%)
Chinese (Kudo) 13 half faces............. 1 1 1
European (Kudo) 5 half faces............ 1 2 2
Negro (Loth)' 30 half faces.......:....... 12 (40%) 11 (86%) 30 (25%)

The entire muscle generally possesses a broad, thick belly;
divided muscle parts often unite with one another at their origin
(fig. 8). I find-a similar condition in the five Europeans which I
have studied; also in the Japanese.

M. auricularis superior et anterior (figs. 2, 3, and 6 to 8)

All authors agree that these two muscles are incompletely
separated from each other in negroes. In general, in the yellow
race as in Europeans, I find a deep, broad, thin muscle plate which
may join the occipitalis dorsally, the frontalis in front (figs. 2,
and 5 to 7). On the whole, no peculiar differences between the
white and yellow races are demonstrable in my material

FACIAL MUSCULATURE OF THE JAPANESE 667

M. auricularis inferior (Le Double)

I have never been able to find anything like the so-called M.
auricularis inferior in Mongolians. Of course, I have occasionally
observed a portion of the platysma which runs close below the

ear conch.

M. auriculo-frontalis Gegenbaur (figs. 3 and 6 to 7)

I have found this muscle distinct six times in fifteen Japanese
(fifteen half faces), besides twice in Chinese (three half faces).
The extent of the muscle varies with the individual. In com-
plete development (in a Japanese) it appears as a single muscle
plate over the temple and crown, arising in front from the orbic-
ularis oculi, and uniting behind with an auricularis superior
(fig. 6). )

In two cadavers (a Japanese and a Chinese) the auricular
bundles radiate fan-like on the lower frontal section (fig. 3). As
a rule, the thin pale muscle runs, with parallel fibers, over the
temple and loses connection with the auricularis (fig. 5).

The M. auriculo-frontalis has been found in six out of thirteen
negroes (Chudzinski, Eggeling, Popowsky, Loth) and in two out
of seventeen Hottentots (Fetzer). In Europeans only Ruge
states that the muscle occurs ‘nicht ganz selten.’ Sappey con-
siders it constant, since, on temples apparently free from muscles,
the microscope still demonstrates muscle bundles; but he says:
‘‘Mais sa minceur est extréme, et telle, que huit fois sur dix
e’est 4 peine si l’on peut le distinguer 4 l’oeil nu.”

2 THE FACIAL MUSCULATURE AS A WHOLE

Before considering the facial musculature as a whole, I desire
to give the following short résumé of the literature relating to
this subject.

In the negro (better, black race) the superficial muscles of the
head, according to Chudzinski and others, are strong and greatly
developed. Giaccomini (cited by Loth), on the other hand, is
the only one who has not expressly mentioned the primitive
character of the facial muscles of the black races.

668 TOKUYASU KUDO

On the basis of the examination of four Hereros and a Herero
child, Eggeling says that the relatively frequent presence of
certain features, which have been regarded by Ruge as primitive,
and the similar coincidence of several such characters in the same
half faces distinguish the facial musculature of the Hereros from
that of the European. Loth, who has compiled the scattered
literature on the muscle system of various blacks, with refer-
ence to his own observations on the negroes, demonstrated a
tendency toward a fusion of the single muscles or the formation
of an almost united muscular layer in the face, a conception
which has been confirmed by most of the students of the negro
(Hamy, Chudzinski, Popowsky, Eggeling). Eckstein worked
on the muscular system of a negro foetus, Hans Virchow on the
facial musculature of sixteen negroes; the latter author states
that the muscles of the negro are always more strongly developed
than those of the stronger Europeans, are inclined to strati-
fication, and that the fibers are coarse and are not exactly parallel.

As for the Papuans, the facial musculature of two newly born
individuals studied by Forster constitutes “das klassische Bei-
spiel atavistischen Zustinde.’’ He mentions a stronger develop-
ment, a very plump appearance, yet often no plain separation
of single bundles to form special muscular entities, which could
cause the delicate shades of facial expression. Fischer also finds
in two adult Papuans an essential agreement with Forster’s
newly born individuals. Steffens and Korner, studying a newly
born Papuan, contradict much of Forster’s account which he
regarded as pithecoid in character.

In the Hottentots, as the result of Eggeling (a child) and Fetzer
(seventeen individuals) show, the muscles of the face, lying
between and around the mouth and eyes are somewhat bulky
and undifferentiated; the single units are not so isolated and not
so widely separated from each other as we are accustomed to see
in Europeans. He declares, with apparent probability, that
here the type of facial muscles corresponds to a lower grade of
development of the human race.

In the Mongolians Birkner has already shown that his three
Chinese heads are distinguishable from those of Europeans by

FACIAL MUSCULATURE OF THE JAPANESE 669

a more restricted division of the facial musculature. In his in-
vestigations on European, Chinese, and Japanese heads (the first
two on which have been used by me in this study) Adachi has
recognized two types of facial musculature: 1) All facial muscles
strongly developed, coarse-fibered, and criss-crossed; 2) the
facial muscles weakly developed, finely fibered, and little criss-
crossed. He found, however, no special racial difference and
only states that the first type usually appears in the broader-
faced forms and the second in the narrower faces.

In the Europeans the head muscles have been exhaustively
investigated from many angles. Since the results have, for the
most part, not been given on a per cent basis, it will not do, in
the search for racial anatomical differences, to evaluate the con-
tributions of literature directly as a basis of comparison.

The results reached by my own investigations make it im-
possible for me to set up a single conclusive racial difference
between Mongolians and Europeans. However, I will not
deny simply on the ground of facial musculature that racial
differences are present. I find differences between the Mon-
golians and Europeans which cannot well be explained as pure
individualities. If, for example, one compare figure 7 (Chinese)
and figure 8 (European), it is evident at first glance that in the
former the face is strongly muscularized and little differen-
tiated; in the latter, on the other hand, it is delicately built and
well differentiated. The fact that these two chosen extreme
cases belong to two different races may not be entirely casual.
Among the fifteen Japanese neither the case present in figure 7
nor figure 8 is found. I have observed cases similar to those
above cited very often. At least, it is an extreme case when the
two conditions cited have not been found.

Also, in a general consideration of the musculature of the face
as a whole, it is not too venturesome to assert that the Japanese
and Chinese are separated from the Europeans by a somewhat
smaller differentiation; that is, a tendency of single muscles to
fuse superficially into a single plate; also, by greater development
and greater extent of the musculature, just as other students
have proved for the black race and for others.

670 TOKUYASU KUDO

This tendency, it seems to me, is stronger in the Chinese than
in the Japanese. I found a highly developed crossing or extreme
radiation of muscle fibers in two Chinese and in a few Japanese
but never in Europeans. Perhaps the described differences
would all be sharper if the comparisons of the musculature had
been carried out on more abundant material.

In a strongly muscularized head the muscles are usually coarse-
bundled. Nevertheless, on a basis of fineness or coarseness of
bundles alone, I could not base any racial difference. When the
thickness of the muscle bundles and the differentiation of single
muscles, etc., are taken into account, the racial differences are
insignificant in my material.

With reference to single variations among Mongolians, all
types of varieties are found which are weakly expressed in the
facial muscles of the Europeans; on the other hand, no varieties
are shown in the Mongolians which have not been described for
the Europeans. But we find that the presence of certain varie-
ties or characteristics which manifest themselves only occasion-
ally or seldom in Europeans or negroes, are observed regularly
in the Mongolians.

In conclusion, I sum up the observations on the separate
regions and the arrangement of the entire facial musculature of
fifteen Japanese, three Chinese, and five European heads as
follows:

1. The platysma which takes part in the structure of the cheek
region, consists, for the most part, in the Mongolian of a con-
tinuous muscle plate, the same asin Europeans. Well-developed
platysma fibers which extend in a line drawn from the corner of
the mouth to the outer ear opening or course above it have been
found in five Japanese and two Chinese.

Most of the cases of the aberrant platysma strands, which
rise orbitotemporalward and may often reach the zygomaticus
or orbicularis oculi, have been observed in the Japanese (eleven
out of fifteen half faces), and constantly in the Chinese. I
have nothing special to contribute with respect to the frequency
of the neck portion in the Mongolians.

The M. mandibulo-marginalis has been found twice in fifteen
half faces of the Japanese. It is rarer in Europeans.

FACIAL MUSCULATURE OF THE JAPANESE 671

2. The muscles of Mongolians (Japanese and Chinese) which
function as dilators of the mouth appear to be less divided than
in Europeans. In the Mongolians the muscles are generally
difficult to distinguish from one another, are more extensive and
coarser. In the Mongolians the triangularis fibers, for the
most part, are spread out, fan-shaped, along the margin of the
jaw.

The M. risorius is generally present in Mongolians (twice in
fifteen Japanese half faces, never in Chinese, 43 per cent in negroes,
33 per cent in seventeen Hottentots. The M. transversus
menti also occurs frequently (without exception in fifteen Jap-
anese and three Chinese, 60 per cent in Europeans, 30 per
cent in the negro). In Mongolians, as a rule, the three parts of
the quadratus labii superior fuse into a single plate; further,
the caput zygomaticus, constantly present in Mongolians, is
distinguishable with difficulty from the neighboring muscles.

°3. The musculature around the eye is more strongly devel-
oped in Mongolians (especially in Chinese), as I have found by
comparison with five Europeans. The bundles radiating at the
lower median part are especially strongly developed. The
separation of the muscle from its surroundings is usually not def-
inite (connection with the M. zygomaticus and M. quadratus
labii superioris).

4. The epicranius shows no noteworthy difference between
Japanese and Europeans. The junction of the muscle of either
side along the median line in the region of the middle third of
the muscle follows the same plan as that in the European (and
also in the negro).

5. Likewise, I find no special difference in the muscles in the
vicinity of the conch in my material. Nevertheless, it might be
desirable to undertake an investigation of the ear muscles in
more extensive material, in which eventually a racial difference
might be discovered.

It may be noted that the M. transversus nuchae is more
frequent in the Japanese (negro 58 per cent, Japanese in half of
the cases of half faces, Europeans 37 per cent). The (M. auric-
ulo-frontalis was found six times in the Japanese (fifteen half
faces) and once in the Chinese (three half faces).

672 TOKUYASU KUDO

6. In spite of a considerable difference in form of the nose,
nothing noteworthy has been found with respect to the muscles.

7. The facial musculature as a whole in the Mongolians
appears to show only individual minor differences.

Thus, in general, the facial musculature of the Japanese pre-
sents a more primitive type than that of the European. It is
to be noted, however, that in certain parts, the reverse holds.

LITERATURE CITED

Apacui, E. 1905 Preliminary notes on the facial muscles of the Japanese and
the Chinese. Jour. Anthrop. Socy. of Tokyo, vol. 20.

_Agsy, C. 1879 Die Muskulatur der menschlichen Mundspalte. Arch. mikr,
Anat., Bd. 16.

Austoni, A. 1908 Muscoli auricolari estrinseci dell’ uomo. Arch. Ital. Anat.
e Embriol, vol. 7. Cited from Eisler.

BrirKNER, F. 1906 Beitrige zur Rassenanatomie der Chinesen. Arch. f.
Anthropol., N. F., vol. 4.

Buuntscuu, H. 1910 Beitrige zur Kenntniss der Variation beim Menschen.

_ Morph. Jahrb., Bd. 40.

CuupzinskI, T. Cited from Le Double, Loth, ete.

EcxkstEeIn, A. 1912 Bemerkungen iiber das Muskelsystem eines Negerfetus.
Anat. Anz., Bd. 41.

Eacetinec, H. 1909 Anatomische Untersuchungen an den Kopf von vier
Hereros, einem Herero- und einem Hottentotkind. Denkschr. med-
naturwiss. Gesellsch. Jena, Bd. 15.

E1ster, P. Die Muskeln des Stammes. Bardeleben’s Handb. Anat. Menschen,
2. Bd Abt..25 Veil 1:

Frerzer, C., 1914 Rassenanatomische Untersuchungen an 17 Hottentoten-
képfen. Zeitsch. Morphol. u. Anthrop., Bd. 16.

FiscHer, E. 1905 Anatomische Untersuchungen an Kopfweichteilen zweier
Papua. Korr.-Bl. deutsch. Gesellsch. Anthropol. Ethnol. u. Urge-
schichte, Bd. 36.

Fiower, W. H., anp Murin, J. 1867 Account of the dissection of a Bushman.
Jour. Anat. and Physiol., vol. 1.

Forster, A. 1903 Kurzer Bericht ttber das Muskelsystem eines Papua-Neu-
geborenen. Anat. Anz., Bd. 24.

1904 Beitrag zur vergleichenden menschlichen Anatomie. Nova
Acta. Leop. Carol. Akad., Bd. 86. Cited from Eggeling and Fischer.

Froriep, A. 1877 Uber den Hautmuskeln des Halses und seine Beziehungen
zu den unteren Geschichtsmuskeln. Arch. f. Anat.

Furamura, R. 1906 Uber die Entwickelung der Facialmuskulatur des
Menschen. Anat. Hefte, Bd. 30.

GEBENBAUR, C. 1899 Lehrbuch der Anatomie, 7 Aufl.

GiaAcoMInI, C. 1882 and 1884 Annotazioni sopra ]’anatomia del negro.

Hamy, E. T. 1870 Muscles de la face d’un négrillon. Bull. Société d’Anthrop
Ade a)

FACIAL MUSCULATURE OF THE JAPANESE 673

HartTMANN, M. 1883 Die menschlichenihnlichen Affen und ihre Organisation
im Vergleich zur menschlichen. Internat. wiss. Bibliothek, Bd. 60.
Cited from Loth.

Hentz, J. 1879 Muskellehre. Handb. syst. Anat. Mensch., 2 Aufl.

Keucn, G. 1813 Beitrige zur pathologischen Anatomie. Berlin. Cited from
Eisler.

Knott, J. F. Cited from Hisler.

Le Dovusite 1897 Traité des variations du systéme musculaire de homme.
Paris.

Livint, F. 1899 Contribuzioni all’anatomia del negro. Arch. p. l’anthrop. e
Vetnol. Firenze. Cited from Loth.

Lorn, E. 1911 Anthropologische Beobachtungen am Muskelsystem der Neger.
Korr.-Bl. deutsch. Anthropol. Gesellsch., Bd. 42.

Macauister, A. Cited from Eisler and Loth.

Orto, A. W. 1816-24 Seltene Beobachtungen zur Anatomie, Physiologie und
Pathologie Breslau. 1, (1816; 2, 1824) cited from Hisler.

Porowsky, J. 1890 Les muscles de la face chez un negri Achanti. L’Anthro-
pologie, T. 1.

Ruace, G. 1889 Untersuchungen iiber die Gesichtsmuskulatur der Primaten.
Leipzig.
1911 Geschichtsmusculatur und N. facialis der Gattung Hylobates.
Morph. Jahrb., Bd. 44.

Sappy, Pu. C. 1888 Traité d’anatomie descriptive, T.2. Paris.

Scumipt, W. 1894 Uber das Platysma des Menschen, seine Kreuzung und
seine Beziehung zu Transversus menti und Triangularis. Arch. f.
Anat.

ScuuuzeE, F. EK. 1865 Der M. transversus nuchae. Schmidt’s Jahrbuch, Ros-
tock, Bd. 127. Cited from WHisler.

SrypaL, O. 1894 Uber eine Variation der Platysma beim Menschen. Morph.
Jahrb., Bd: 2%.

STEFFENS, F., AND KORNER,O. 1910 Bemerkungen iiber das Muskelsystem eines
Papua-Neugeborenen. Anat. Anz., Bd. 36.

THIELE, Fr. W. 1841 Die Lehre von den Muskeln. In S6mmering’s Vom Baue
des menschlichen Koérpers. Leipzig. Bd. 3, Abth. 1. Cited from

Hisler.
Trestut 1884 Les anomalies musculaires chez lhomme. Paris. Cited by
Bluntsehli.

Turner, W. Cited by Loth
VircHow, H. 1908 Gesichtsmuskeln und Gesichtsausdruck. Arch. f. Anat.
1912. Uber Gesichtsmuskulatur von Negern. Verh. Anat. Gesellsch.,

Bd. 26.
Woop, J. 1867 Variations in human myology. Proc. Roy. Socy. London,
vol. 15.

JOURNAL OF MORPHOLOGY, VOL. 32, No. 3

EXPLANATION OF PLATES 1 TO 3

Comparison of the three plates (1, Japanese; 2, Chinese; 3, European) shows
at the first glance great differences in the facial musculature asa whole. Inthe
Japanese and Chinese the facial muscles are strongly developed, coarsely bundled,
spread widely, with a tendency toward a fusion of the separate muscles into a
single sheet, often (especially in the Chinese) radiated and interlaced. In the
European (pl. 3), on the other hand, the separate muscles are delicate and well
differentiated. I selected extreme cases for my figures of the different races;
naturally, such great differences are not universal. The separate muscles are
described in detail as follows:

PLATE 1
EXPLANATION OF FIGURE

6. Female Japanese XI, twenty years old. Platysma forms an entire plate,
the pars aberrans surpasses the mouth angle-auditory meatus line in a fan-like
arrangement. The triangularis is broad at its origin. Between this muscle and
the risorius (the latter consists of two bundles) are intermediate marginal fibers.
It is noticeable that several isolated fiber strands of the M. mandibulomargi-
nalis, partially covered by the triangularis, run obliquely on the platysma. The
coarsely bundled zygomaticus is scarcely separated from the caput zygomaticus
and the orbicularis oculi. The superficial layer of the caninus, at the insertion
of the zygomatic, consists of weak fibers. The caput angulare of the quadratus
labii superioris is strong. The lateral marginal bundles of the orbicularis oculli
are compact and pass over below to the zygomaticus. Also there are laterally
directed radial bundles on the lower lateral quadrant of the orbicularis. The
medial lower bundles are compact and somewhat swollen. The M. frontalis is
relatively fine-fibered. The lateral part of this muscle, the auriculofrontalis
and the auricularis superior and anterior come in contact with each other and
form a thin connected sheet over the temple. The medial vertical fibers of it,
along the temporal vessels, are bent toward the ear; the fibers of the auriculo-
frontalis which run forward, end on the lower region of insertion of the frontalis,
beneath the orbicularis oculi The three parts of the M. auricularis posterior
are not sharply separated from each other.

674

PLATE 1

FACIAL’ MUSCULATURE OF THE JAPANESE

TOKUYASU KUDO

675

PLATE 2
EXPLANATION OF FIGURE

7. Male Chinese II. The platysma is coarsely bundled and forms an entire
plate. Its upper part extends above the mouth-angle meatus line toward the
zygomatic; the inferior part passes under the triangularis. The triangularis,
together with the M. risorius, forms a fan-formed muscle mass; some inter-
mediate bundles between the two muscles extend to the platysma. At its origin
the zygomaticus is overlaid by and partially interlaced with the orbicularis
oculi. Strong radiating and often crossing bundles on the lateral part of the
orbicularis oculi are striking. Marginal bundles from the auriculofrontalis
radiate upward; the curved lateral fibers interlace with ascending platysma fibers.
Those running ventrad extend farther on the zygomaticus, even to the quad-
ratus labii superioris. The lower medial bundles are also well developed; the
superficial layer is transected in the illustration and the deep portion passes
under the lateral marginal bundle and then interlaces with the cervical part of
the platysma. The laterally directed marginal bundles of the orbicularis oculi
lie deep between it and the ear, and are visible in the plate through an artificial
opening. The M. frontalis is connected laterally with the higher auricularis
bundles. The auriculofrontalis runs between the two muscles, joins the auricu-
laris anterior, and continues forward and upward to the lateral part of the
frontalis. The temple is also well muscularized. The extent of the occipitalis
is noteworthy; the posterior bundles are more vertical and the anterior incline
more forward and reach the auricularis superior. At last they unite trans-
versely with the auricularis posterior. The auricularis is inclined forward and
downward and consists of four parts, not readily separable. The cervical part
of the platysma is broadly extended under the ear, the hinder limb approaches
the median line above the insertion of the sternocleidomastoid; the anterior
diverges at the hinder part of the cheek region of the platysma. Some bundles
of the latter ascend further and cross the diverging bundles of the orbicularis
oculi.

for)
=]
(op)

PLATE 2

FACIAL MUSCULATURE OF THE JAPANESE

TOKUYASU KUDO

<h.. Ug, pete

677

PLATE 3
EXPLANATION OF FIGURE

8. Male European II. The platysma runs forward as an entire plate, but
lacks a pars aberrans. The triangularis is small. The zygomatic is entirely
separate from the orbicularis oculi and the caput zygomatici. The three heads
of the quadratus labii superioris are well differentiated. The orbicularis oculi
forms a closed ring around the eye, diverging bundles are present. The frontalis
extends far toward the vertex. There is no musculature on the temporal region.
The auricularis superior and the occipitalis are moderately developed; the latter
does not reach the ear and the occipitalis. The auricularis posterior consists
of two parts, connected with each other at the origin.

678

-

PLATE 3

FACIAL MUSCULATURE OF THE JAPANESE

TOKUYASU KUDO

4

679

Resumen por el autor, Clarence Lester Turner.
El ciclo estacional en el espermario de la perea.

Il presente trabajo es un estudio de la variacién volumétrica
del testiculo de la perea. El tamanho minimo se encuentra a
principios de verano. El comienzo del aumento de tamafo es
brusco y este aumento es muy rapido, aleanzando el maximo de
tamano en la tiltima parte de Noviembre en la cual el peso de
los testiculos es 60 veces mayor que su peso minimo. El volu-
men del espermario decrece a consecuencia de la puesta (desde
primeros de Marzo hasta tltimos de Mayo). El cord6n de célu-
las germinativas que susministra anualmente elementos al tes-
ticulo est: situado centralmente a este 6rgano; las células germi-
-nativas emigran dentro de este ultimo, siguiendo exteriormente
las paredes de los l6bulos y alojandose en los lobulos de la peri-
feria. A partir de esta, el testiculo est’ ocupado por células ger-
minativas, que se transforman en espermatogonias, las cuales
forman cistos. El comienzo del aumento de tamano del testiculo
coincide con la formacién de las espermatogonias. No hay peri-
odo de crecimiento antes de la maduracioén. La espermatogéne-
sis tiene lugar en los primeros meses del otoho. El comienzo de
la disminuci6én anual de temperatura (a ultimos de Agosto) coin-
cide con el aumento volumétrico estacional del testiculo. El
comienzo de la disminuci6n estacional del testiculo (a primeros
de Marzo) coincide con el aumento de temperatura en el agua.

Translation by José F. Nonidez
Carnegie Institution of Washington

AUTHOR’S AESTRACT OF TH!IS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, JULY 21

THE SEASONAL CYCLE IN THE SPERMARY OF THE
PERCH?

CLARENCE L. TURNER

Zoological Laboratories of the University of Wisconsin

CONTENTS

Tealimtro cliuGhlOnvs,.Sts3 ta eaete ree bates sc unpeek | Pees AN ENA ROR ROE ie Rlear i Pa. neta: eee OSIL
if Misterialbandemethodse PEE) OO ERAGE MPT ees 15S)
Ill. Anatomy SPnCMeehe eet ea MOPAR oi hemo? A ewe OSL
7 Generalltrelationship suena 455) oe Meee a ee ee Se OA
RE AVOUMELEUC Vera Dil OMmngs hace gece aus ee ii one eae omens OGL
3. Microscopical amenity ge AG eee ae ee OOO

4. Discussion of seasonal Micolenical hs ar ges aad comparison with
OuWerzanimMall oro UD Sateen serena ace see len ec ae OOO
Wes Seasonalehistonysot cenncelllsimss erate tie stot ie teeter ts eee OOD
[PBRestimeustacerandapenlodaolennlo Tat One merrier eee ere OO
2. Period of proliferation and Broth st ton ae a eee 695
33, IMSIOMTAAE AKIN Wine) SOCAN HOPAONMUE, AR. Jodeb beep dos gdesouse ooo OOM
ES Spermatogonia........ : Exot SRS Ne TC A Tae mee OOS
5. Synopsis and the mene ion Aineua. BOs Raa eee ta pana ee SOOO
(6 ee Myer es hl Ce ee ea mete RETR GG
Us Sie Mane. oe Bee ane e701)
VY. Discussion of posnible Paeiora) mnienen ing tines germ- eel cules inne Dated Neh 700
AVES ikl aaa aWZe seme Sete coe Me eet oR cele nee nee aE NEE inet AMER A MRE edleait cet) arama eens iTS)
WARS illo ora. jo his 2 he-eeae rem sna ay eed eae ache cise Salis Cacetne AES eae comteera ey GOA

I. INTRODUCTION

The study of the germ cells in vertebrates has been confined
principally to two lines of investigation. The first has under-
taken to recognize the germ cells in the embryo and to follow
their lineage, on the one hand, backward through the younger
embryos to the first stages in which they may be positively
identified as germ cells and, on the other, to trace them to the

1A Thesis submitted to the Graduate School of the University of Wisconsin
in partial fulfillment of the requirements for the degree of Doctor of Philosophy.

2 A part of this work was carried out in the laboratories of the Department of
Anatomy and Biology, Marquette University School of Medicine.

681

682 CLARENCE L. TURNER

stage in which the germ gland is first formed. The interest has
centered principally about the location in which the germ cells
apparently originated, in their active migration, in the paths
traversed in reaching their final place of lodging, and in the
number which entered the germ glands. Eigenman (’92, ’96),
Wheeler (99), Dustin (07), Beard (00, ’02), Woods (’01),
Jarvis (08), Allen (’06, ’09), Dodds (710), and Swift (’14) have
all contributed to the early history of the germ cells in different
vertebrates. Embryologists and histologists have contributed
to the further history of the germ cells up to the time in which
they become definitely placed in the fully developed germ
glands.

The second line of investigation has been cytological. The
problems of spermatogenesis—synapsis, chromosome number,
the reduction process, nuclear and cytoplasmic inclusions, ete.—
have so occupied the attention of the cytologist that he has
confined his study mainly to the testis of the sexually mature
animal or to those stages in the history of the testis which fur-
nished satisfactory cytological material.

Among the vertebrates are some forms which have lent them-
selves admirably to the investigation of the origin of the germ
cells and to cytological studies. Other animals, because of their
small germ cells or the difficulty of obtaining complete embry-
ological series have been abandoned by both the embryologist
and the cytologist. Eigenmann and Dodds have given excellent
accounts of the germ cells and their behavior in the embryos of
teleosts. Many other papers have dealt with the fertilization
and cleavage of the teleostean egg and the embryology in a num-
ber of teleosts has been described in detail. There is a dearth
of literature, however, on the morphology of the adult teleostean
testis and the germ cells in the adult have been investigated but
little.

It is a familiar fact that the ovaries and testes of teleosts as
well as those of amphibians and of some other vertebrates differ
in size at different seasons of the year. Advantage has been
taken of the volumetric variation in the ovary of the fish to
work out some valuable data concerning the growth and dis-

SEASONAL CYCLE IN PERCH SPERMARY 683

tribution of fish at different seasons. So far as the writer can
determine, however, no thorough study has been made upon
the seasonal changes that take place in the teleostean testis.

It is the main purpose of this paper to describe the changes
that have been observed in the testis of the yellow perch (Perca
flavescens) during the different seasons of the year with refer-
ence to the gross and microscopical anatomy and the cytology.

II. MATERIAL AND METHOD

The material for this investigation was collected from Lake
Mendota at Madison, Wisconsin, and from Lake Michigan at
Milwaukee, Wisconsin, between May 1, 1915, and October 1,
1917. From five to twenty specimens were examined each week
during this period. AIl were weighed fresh and the testes were
weighed before fixation. Spermaries of other teleosts (carp,
sunfish, crappie, pickerel, and lake trout), as well as those of
Necturus, of two species of turtles, of several birds, and of the
muskrat, were examined from time to time for comparison.

For histological study material was fixed in 10 per cent formalin
and in a picro-acetic acid mixture. Haematoxylin and haemalum
stains were used with an eosin counterstain. Resorsin-fuchsin
was used to stain the elastic tissue.

In fixing material for cytological detail, Flemming’s, Gilson’s,
and Bouin’s mixtures were used, the best results being obtained
with Bouin’s mixture to which had been added a small propor-
tion of urea (Allen, 716). Flemming’s triple stain, saffranin
counterstained with licht griin and haematoxylin counterstained
with eosin or licht griin were employed. Gentian violet alone
also proved valuable. Sections from 5yu to 15u thick were used
in the histological work. Both smears and sections were used
in the cytological study, smears being prepared between cover-
slips as described by Agar (711).

The writer is indebted to Prof. A. S. Pearse for aid in collect-
ing specimens in Madison, to Prof. M. F. Guyer for suggestions
in technique and for criticisms, and to Prof. E. A. Birge, who
kindly loaned a record which summarized the last fifteen years’
data on the temperature of the water of Lake Mendota. <Ac-

684 CLARENCE L. TURNER

knowledgment is also made to Mr. Leo Massopusst, artist at
the Marquette University School of Medicine, for instruction in
his method of illustrating.

Ill. ANATOMY OF THE TESTIS
1. General relationships

The testes are two elongated white bodies situated in the pos-
terior part of the body cavity just ventral to the swim bladder.
They fuse together at their posterior ends, forming a single
body. They are oval in shape and taper toward their posterior
ends. Usually they are about equal in size, but cases have been
noticed in which one testis was nearly twice the size of the other.
They are attached to the swim bladder by two delicate mesorchia
which converge posteriorly into one at the point where the testes
fuse. Anteriorly the mesorchia extend beyond the limits of the
testes and support a sheath containing the genital arteries and
veins.

Posteriorly the testes communicate with the urogenital open-
ing by a thin-walled but capacious sinus. <A horizontal septum
separates this sinus from a dorsal chamber which receives the
common ureter from the Wolffan bodies. Both chambers ter-
minate in the common urogenital opening.

2. Volumetric variation

The volumetric variation of the testis from one season to
another is a conspicuous feature. Just after the spawning sea-
son, which occurs in late April or early May, the testis is de-
pleted and its volume is slight. During May, June, July, and
early August there is practically no change, except for a slight
individual variation. Late August witnesses the initiation of a
sudden increase in volume, which proceeds so rapidly that by
the latter part of September the weight is between thirty and
thirty-five times as great as it was six weeks earlier. By Janu-
ary the maximum weight has been attained, when it is nearly
forty-five times as great as the weight of the depleted testis.
At this time it represents from 4.58 to 5.9 per cent of the entire
eross weight of the body.

SEASONAL CYCLE IN PERCH SPERMARY 685

The weight and size of the testis is proportionate to the weight
and size of the body. For example, at one season of the year the
weight of the testis of a fish weighing 97 grams was 5.25 grams
while that of a fish weighing 40 grams was 2.25 grams. In the
first case the weight of the testis was one-eighteenth, and in the
second it was one-seventeenth of the gross weight of the body.
While the actual weight of the testis cannot be taken as a reliable
criterion, therefore, for the volumetric variation of the testis, the
proportion between the weights of the body and of the testis
will be fairly constant for a large number of individuals for any
given season. Hence, in constructing a curve to show the varia-
tion in the volume of the testis in different seasons, the propor-
tion which the weight of the testis forms of the total body weight
has been taken as the unit by which points may be determined
for a curve. Figures 5-22 represent a series of camera-lucida
drawings showing the variation in size and in shape during the
different seasons. Figure 7A represents the variation graphi-
cally, and the accompanying table gives the figures upon which
this curve was based.

= == a ai AVERAGE RATIO

Dare werent or | werant or | ,BETWEEN

musth ANOISNAS BODY WEIGHT
grams grams per cent
aI ull yale ews Petet ep eeges 2 He Oey, A Ole Aah 78.4 0.1 0.12
UDI ace Wayne alba an 9 i oa eNO ae pana ee ot 68.2 0.1 0.14
TNT OV LAT ISH A Seer amo eas eens eee se ee eee ics (ay Onl 0.14
AANTDTEATISI SS eee esr eat arse rae vane A Ae en ahh aN chen 72.48 0.177 0.21
Sepuemiben Wh apres tines geo aun 89.7 0.37 0.47
Sepvemben lost ja arene Was ame ree: 60.98 0.726 1.19
Octobertli peters eek ein eee ae 72.00 2.46 3.44
Octobe sl ony nei verre eines Se ME mesh & 78.2 4.15 DEo2
INI Oy Sita over eI GS A se ee ae ere UD sh 4.44 5.88
DE cemib erplieptrc erm erent tes 84.00 4.2 5.00
ROMMUTANay alae Mert ny Athen oti sacahe oo Aun wesiehs cite & Ges 67.5 3 2al 4.76
SHG NE Meer tay oulseeante SAC epee Aes gen ach as 72.6 3.16 4.34
Nair isan, Seba des ante eee oe 91.4 4.15 4.54
ZsNy oC [RR hg ih, cae i RM APU: Sed gt 75.3 2.78 3.70
ANVONENIP TIES) Seve ene IR cee Clclgs RABAT A Sis oe 68.8 2.86 4.16
INTe ihe ares s, Mey pore sete cient Mant fat 1 76.7 1.91 2.50
IN aiy all SARs cre in emer et adnate (ees 0.84 1.62
JUTe Re er pas 7, saa Ce Ree ay ee eee a 81.4 0.26 0.338
Jiu eye ea cori ee Pe eee pty cg eo eee 85.00 0.15 0.17

686 CLARENCE L. TURNER

The spaces between the ordinates represent intervals of ap-
proximately thirty days while the spaces between the abscissae
represent variations of one per cent in the ratio between the
weight of the testis and the weight of the body.

The maximum weight of the testis is reached about November
1. From November 1 to April 1 there is a gradual decline in
weight. The irregularity in the curve between December 1 and
March 1 is occasioned by individual variation. During these
months difficulty was experienced in getting specimens, and it is
probable that there would have been no irregularity had enough
material been at hand. The declining curve from March 1 to
July 1 does not indicate that there is a gradual expulsion of the
spermatozoa by each individual between these dates, but rather
that a few fish discharge their spermatozoa as early as March 1
while others do not discharge until late in May. The curve
declines as the proportion of those which have discharged in-
creases. Although the point in question was not verified by
actual observation, it is probable that each fish has a series of
discharges, for a great many testes showed different stages of
depletion during the spawning season.

No attempt was made to keep separate data for fish of differ-
ent ages when it was found that specimens varying in weight
from 60 grams to 300 grams did not offer any essential differ-
ences from the general course described.

3. Microscopical anatomy

The testis is a sack enclosed in a connective-tissue sheath
which contains a large proportion of elastic elements. Within
each testis, on its ventral side, is a connective-tissue core, deeply
imbedded, from which septa radiate. These septa extend to
and join the testis-wall, dividing the entire organ into lobules.
This connective-tissue core is somewhat comparable to the me-
diastinum testis of the higher vertebrates. The cores of the two
testes converge and fuse posteriorly at the point of fusion of the
testes. The elastic fibers are very abundant throughout the en-
tire testis. Two heavy cords of elastic fibers are located in the
mesorchia which suspend the testes from the swim bladder.

SEASONAL CYCLE IN PERCH SPERMARY 687

They are fused together for a short distance at the point of the
fusion of the testes, but they redivide, one cord passing into the
connective-tissue core of each testis and, subdividing within,
send one branch anteriorly and one posteriorly. Small branches
extend into each of the connective-tissue septa in the form of

JUNE JULY AUG. SEPT. OCT NOV. DEC. JAN. FEB. MAR APR MAY

Fig. 1. A. Graphical illustration of volumetric variation in testis. Each
space on the ordinate represents a period of one month. Each space on the
abscissa represents a variation of 1 per cent in the ration between the body
weight and the weight of the testis. B. Graphic representation of the seasonal
variation in temperature in the water of Lake Mendota at a depth of 6 meters.
Each vertical space represents a variation of 5°. Each horizontal space repre-
sents a period of one month.

Fig. 2. Sketch illustrating casts of lobules. A. Segment showing lobules in
place. B. Surface view of a group of lobules. C. Side view of same group of
lobules.

688

SEASONAL CYCLE IN PERCH SPERMARY 689

cords and bands. The amount and distribution of the elastic
tissue is important, since, in the absence of muscle fibers and of
special ejaculatory organs, it must play some part in the expul-
sion of the spermatozoa. The reduction of the testis to its nor-
mal small size after its distention must also depend upon the
elastic tissue.

The genital arteries and veins passing into the connective-
tissue core at the anterior end of the testis give off minute
branches into the testis as they proceed posteriorly. Elongated
bodies of adipose tissue are also located in the connective-tissue
core of the testis and send out digitate processes into the septa
between the lobules.

The lobules are irregular-shaped spaces with their apices at
the center of the testis and their broader ends directed toward
the periphery. The precise shape can be obtained in the form of
casts by allowing a testis to soften for a time, fixing it in alcohol
and then tearing it apart. The hardened mass of spermatozoa,
representing a cast of the lobule, may then be dissected out
with ease (fig. 2 a, b, and c). The same procedure may be car-
ried out with success in testes which contain only the early,
transforming germ cells. The casts of the lobules appear as
flattened leaves, joined at their apical ends, diverging and branch-
ing toward the periphery of the testis. As many as five branches
may be given off from a single central trunk, and these branches
may again bifurcate before coming into contact with the testis
wall. Some of the branches end blindly. In the pickerel the
branches of the tubules at the periphery of the testis are much
more finely divided and are convoluted, giving the appearance
of seminiferous tubules (fig. 23).

4. Discussion of seasonal histological changes and comparison with
other animal groups

From the foregoing description it follows that the perch testis
differs radically from that of the Sauropsida and the mammals.
The fact, too that the testis undergoes such a complete seasonal
change places it in a class far removed from the amniotes. It

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

690 CLARENCE L. TURNER

resembles the testis of the Amphibia in so far as there is a lack
of seminiferous tubules and in the presence of lobules which be-
come divided up into cysts as soon as the maturation process
has begun. While the amphibian testis undergoes a seasonal
change, there is no such seasonal variation in volume as there is
in the perch. There is also an anteroposterior seriation in the
testis of the urodele, while the testis behaves as a unit in the
perch.

It is evident from only a cursory examination, however, that
there are variations in different teleosts which differ considera-
bly from the course of the changes which occur in the perch.
After some germ cells have migrated to the periphery of the
testis in the perch there is-a proliferation which ‘increases their
number. There is also a constant addition occasioned by the ©
arrival of new migrating cells. This process continues until there
is a solid cord of germ cells which fills each lobule of the testis
at its normal small size (early August). Some connective-
tissue cells. are found among the germ cells. The transformation
of the germ cells into spermatogonia is contemporaneous with
the beginning of the increase in the volume of the testis (late
August). After spermatogonia are formed the process of sper-
matogenesis takes place rapidly and the volume of the testis
increases apace. The spermatogonia arrange themselves into
cysts which are imperfectly divided by connective-tissue cells.
Each cyst behaves as a unit during the maturation process, all
of the cells passing through the same stage at the same time
(figs. 25 and 26). The cysts at the center complete their mat-
uration and form spermatozoa first, but there does not seem to
be any definite seriation from the center to the periphery of the
lobule.

The formation of cysts within the lobules is not as clearly
shown in the perch testis as it is in the testis of the sunfish
(fig. 24). There is a close resemblance between the cyst forma-
tion in the testis of these teleosts and in the testis of many
arthropods.

When the germ cells first become lodged at the periphery of
the testis in the perch they form a lens-shaped mass which

SEASONAL CYCLE IN PERCH SPERMARY 691

conforms to the shape of the cavity at the peripheral end of the
lobule. In the pickerel the peripheral ends of the lobules be-
come subdivided into pockets which are long and tubular.
When the germ cells reach these pockets in the pickerel they
become arranged along the inner walls, leaving a lumen in the
center. Because of the peculiar form of these peripheral sub-
divisions of the lobules and the relations of the germ cells to
them, there is a resemblance between them and the seminiferous
tubules of the higher vertebrates (fig. 23).

There is a close resemblance between the cords of germ cells
formed annually in the lobules of the testis of the adult perch
and the formation of cords of primordial germ cells in the em-
bryonic testis of mammals. In both cases the cord is formed by
germ cells which have migrated into the testis from without.
Here the resemblance ceases, for a part of the germ cells in the
cord of the embryonic mammalian testis are destined to form
nurse cells, while in the perch all the cells give rise to sexual
products.

There is a marked change in the somatic structures of the tes-
tis during the changes in volume, but this does not involve
growth. The interlobular walls become thin and the wall of the
testis decreases in thickness. The blood-vessels dilate, appar-
ently to meet the needs of the rapidly dividing germ cells. ‘There
is no increase in the adipose tissue. The entire testis seems to
be in a state of tension while it is at its maximum size, as evi-
denced by the stretched condition of the interlobular and testis
walls and by the deflection at the peripheral ends of the branches
of the lobules (fig. 2, c). When the spermatozoa are expelled
the region nearest the connective-tissue core on the ventral side
of the testis is depleted first. When all the spermatozoa have
been expelled, the normal thickness of the interlobular walls
and of the testis walls is restored and the small size of the testis
is resumed.

It is interesting to speculate as to the character of the changes
which occur in those teleosts which spawn but once in their
lives. The Pacific salmon and the eel would furnish material
for such an investigation.

692 CLARENCE L. TURNER

IV. HISTORY OF THE GERM CELLS
1. Resting stage and period of migration

A cord of germ cells outside of the testis was found in a single
specimen which was killed on May 5. Unfortunately, this was
the only fish taken at this date and, though the cord has been
sought in specimens taken at other dates, it has not been found.
Consequently, further investigation of this point must be post-
poned until a time which will again furnish favorable material.
In this specimen there were no less than 5400 germ cells by
actual count. The cells, imbedded in a connective-tissue ma-
trix, varied in size as well as in shape (fig. 27). Most of them
were apparently at rest, although the irregular shape of some
seemed to indicate migration. No dividing cells were found
among them.

In these cells the cytoplasm is hyalin or reticular and stains
very lightly. The nuclei vary in shape, some being spherical,
others oval, and a few irregular, having a marked indentation
on one side. The plasmosome is a conspicuous object and lies
near the nuclear wall. Most nuclei contain a single plasmosome,
but it is not unusual to find two. The chromatin is well dis-
tributed, being scattered along the linin threads and massed
together in some places to form chromatin knots. The linin
threads appear to be in contact with the nuclear wall and radiate
from the region occupied by the plasmosome.

The actual migration of the germ cells from this cord is as-
sumed upon the grounds of the following observations:

1. Germ cells are found in various locations at different
periods.

a. A mass of germ cells larger than any other to be found at.
that time occurs in the cord outside of the testis (fig. 27, fig. 31,
a and b). The question as to whether there is a progressive
depletion of the germ cells in the cord could not be settled be-
cause material was not preserved which would show this point.

b. Germ cells of a peculiar shape are found along the septa of
the lobules from the center to the periphery of the testis during
the time in which clusters of germ cells are formed and increased
at the periphery of the testis (fig. 32).

SEASONAL CYCLE IN PERCH SPERMARY 693

c. Clusters of germ cells occur at the periphery of the testis
and their numbers are greatly increased at the time when the
germ cells are found along the septum walls. Few mitotic
figures occur, and it is evident that the original clusters of germ
cells at the periphery do not give rise to all the cells found in
this location a short time afterward.

2. Germ cells along the septa between the lobules have an
elongated and an irregular shape which suggests amoeboid
motion (fig. 28).

3. The germ cells in the region of the junction of the inter-
lobular septa and the periphery of the testis (fig. 29) seem to
show direct transitional stages between their migratory form and
their resting form.

During the migration there seems to be a slight increase in
the volume of the cells. There is no change in the character
of the cytoplasm, but the nucleus is a little more hyalin owing
to the more even dispersion of the chromatin along the linin
threads.

After migration, the germ cells come to lodge at the distal
end of the lobules. Some of the lobules do not extend entirely
to the periphery of the testis, but end blindly a short distance
from the center (fig. 2, c). In consequence, the germ cells, after
their migration, occur in pockets some distance from as well as
at the periphery. It seems that the tendency to migrate ceases
only when the cells have definitely come into contact with an
obstruction at the end of the lobule. This fact would seem to
furnish evidence in favor of the germ cells accomplishing migra-
tion actively rather than behaving in a purely passive manner.
During the migration most of the cells are found along the in-
side of the lobule, but some are actually within the intervening
septa. Some recent work in tissue culture has shown that-in
cells cultivated in vitro, migration is facilitated by the presence
of strands along which the cells may move. The strands. of
fibrous connective-tissue in the walls of the lobules would fur-
nish admirable supports of this character. No mitotic figures
- were observed in the migrating cells and it is assumed that they
do not divide during this period. In this connection it is sig-

694 CLARENCE L. TURNER

nificant that germ cells in embryos generally do not divide while
they are migrating.

Once the germ cells have become lodged they undergo an im-
mediate change. A comparison of a and 6 in figure 3 indicates
that there is a slight increase in volume. There is also the for-
mation of some darkly staining spherules which accumulate
around the nuclear wall and pass out into the cytoplasm (fig.
33). The actual formation and extrusion of these spherules is,
of course, a matter of interpretation, but no spherules are found
in the nucleus of the migrating cells except the plasmosome, and
the cytoplasm is entirely free from them. Such a change might
well be brought about by a change in the metabolism of the cell.
During migration the energy would be consumed in locomotion, —
but when the cell becomes sedentary the energy would be di-
verted into growth and a reorganization of the cell contents
preparatory to division.

There is a striking parallel between the behavior of the germ
cells in the adult perch and the embryonic cells described for
Lophius by Dodds (710). He remarks as follows:

In all vertebrates examined, this period (of growth) corresponds to
the time during which the germ cells are in active migration, and it
has been suggested that possibly the energy of the cell is expended in
locomotion rather than in growth and cell division.

The above discussion of conditions observed in these cells during the
rest period offers no explanation why this period of suspended activity
begins, nor why after a time it comes to an end. At its beginning,
before there are any differences we can detect with the eye, there must
be an unseen physiological difference which determines the future be-
havior of the cell. Whatever the nature of the difference, it is one of
the earliest of which we have evidence in the cleaving egg of Lophius.

The migration occurring in the perch seems to correspond to
the period of rest. Dodds has also called attention to the fact
that the germ cell retains its embryonic character longer than
any of the other tissues of the body. It is possible that the germ
cells migrate seasonally because they have not differentiated,
even in the adult perch, and that they are only fulfilling their
innate tendency to migrate whenever the opportunity offers or .
when there is a proper stimulus provided. The problem en-

-

SEASONAL CYCLE IN PERCH SPERMARY 695

countered by Dodds as to why this period of migration and cessa-
tion from division should be inaugurated and why it should
come to an end is also met in the present study. The point
will be discussed further in another part of this paper. ‘The be-
havior of the migrating germ cells in the perch would suggest,
however, that the capacity for migration had not been exhausted
when the embryonic germ cells had reached the germ gland, but
that under favorable circumstances they might again undertake
locomotion after a.period in which their energy had been used
in a static condition resulting in growth and cell division.

The amoeboid cells have been observed along the lobule walls
shortly before the spermatozoa are discharged in April and con-
tinuously until the period of spermatogenesis, which begins
about the first of September.

2. Period of proliferation and growth

This period extends from the time in which the migrating
germ cells begin to collect at the periphery (early April) till
they are transformed into spermatogonia (early August).

The transformation which the migrating germ cells undergo
when they reach the periphery of the testis has already been de-
scribed above. Active growth starts as soon as the cells are
lodged, and by early May a small proportion have become con-
siderably enlarged (fig. 3, a and b). The clusters at the periph-
ery at this stage contain less than a hundred cells. By early
July the clusters have increased considerably (fig. 30, G.c.) and
theré is a larger proportion of the larger cells (fig. 3, c). Pro-
liferation is going on, but very slowly, and it is evident that the
increase in the clusters is partly due to the arrival of new mi-
grating cells. The only mitotic figures in which a definite
chromosome count could be made are furnished by these larger
cells (figs. 35, 36). Twenty-seven chromosomes could be counted
distinctly. Nucleus and cytoplasm maintain their former volu-
metric proportion during growth, and the plasmosome also in-
creases in size. The darkly staining bodies which appear at the
edge of the nucleus when the cell first becomes lodged disappear
during growth.

696 CLARENCE L. TURNER

Fig. 3. Camera lucida drawings to show size relationships of cells during
different periods. All figures except e and f represent outlines of nuclei and all
are drawn to the same scale with an X8\ocular and 1.9 mm. oil immersion lens.
A. Germ cells in cord outside of testis (May 5). B. Germ cells at ‘periphery of
testis (May 5). C. Germ cells during period of growth and proliferation (July
3). D. Germ cells during latter part of period of growth and proliferation.
Maximum size of cells shown (August 5). E. Outline drawings illustrating
comparative size of largest and smallest cells in a lobule, also volumetric rela-
tions of cytoplasm, nucleus and plasmosome; cyt., cytoplasm; n., nucleus; pl.,
plasmosome. F. Drawing illustrating size relations during : transformation of
germ cells into spermatogonia; vac., clear vacuole of liquid surrounding trans-
forming cell; cyt., cytoplasm; n., nucleus. G. Group of spermatogonia. H.
Group of primary spermatocytes. I. Group of secondary spermatocytes. J.
Group of spermatids. K. Group of spermatozoa.

SEASONAL CYCLE IN PERCH SPERMARY 697

The germ cells come to their maximum size about August 5
(fig. 34, a, b,c). At this time the entire testis in its normal small
size is filled with solid cords of germ cells (fig. 30, G.c.). About
5 per cent of the entire number have reached the maximum size;
approximately 15 per cent are-still very small and apparently
have but recently migrated, while the remainder represent in-
termediate sizes.

The cytoplasm of the largest cells is reticular and is free from
darkly staining inclusions. The nucleus is remarkably hyalin.
The linin threads are attenuated and the chromatin is well dis-
tributed. In most cases the plasmosome is a large spherical
structure which takes an acid stain, but in some cases the larger
structure is absent and its place is taken by two or three smaller
spheres. Figure 34 represents three cells taken from the same
lobule, c representing a young germ cell before growth; 6, an
intermediate stage, and a, a cell at its maximum size.

Immediately upon. reaching their maximum growth the cells
are transformed into spermatogonia. The proportion of the
largest cells never increases to more than 5 per cent and the trans-
forming cells are few in number. This would indicate that the
time occupied in the final stage of growth and in the trans-
formation into spermatogonia is short. The smaller and the
intermediate cells are growing meanwhile, and there is a contin-
ual procession of growth and transformation until late in No-
vember. During late September and October each lobule shows
a profusion of young growing germ cells, of germ cells at their
maximum size, of transforming germ cells, of spermatogonia
and of all the succeeding stages of spermatogenesis, including
the mature spermatozoa.

‘ 3. Period of transformation into spermatogonia

During the transformation into spermatogonia two features of
the process are outstanding: 1) There is a definite reduction in
size. 2) There is a change in the chemical composition of the
cells as shown by its capacity to acquire a deeper stain.

698 CLARENCE L. TURNER

Three stages of the transforming cells are shown in figure 37.
In the first stage there is a contraction of the entire cell and prob-
ably the extrusion of a clear fluid. At any rate the boundaries
which marked the limits of the cell at its maximum size are
maintained, and there is a space between this boundary and the
contracted cell, giving the appearance of a cell suspended in a
chamber of clear liquid. The fact that only the transforming
cells present this appearance, while all the surrounding cells are
normal, would preclude the possibility that the condition is an
artif act. Both nuclei and cytoplasm become more densely
staining as further contraction takes place and darkly staining
spherules appear in the cytoplasm and in the nucleus. The re-
duction in volume affects both the cytoplasm and the nucleus.

If changes in volume of the cell, the appearance of spherules
in both the cytoplasm and the nucleus and the acquisition of the
capacity to take a denser stain may be considered criteria for
metabolic activity, the cells have undergone a marked change
in their metabolism.

4

4. Spermatogonia

It is impossible to determine the exact number of generations
through which the spermatogonia pass before spermatocytes are
produced, but it is evident that there are at least five or six.
As each spermatogonium gives rise to a group of descendants
they form a cyst and all pass through the same stages of division
at the same time (fig. 25). Consequently, it is possible to esti-
mate the approximate number of descendants to which a single
spermatogonium has given rise by counting the number con-
tained in a cyst.

In the dividing spermatogonium the chromosomes are so
massed as to preclude a definite count. The most favorable
cells were those prepared in smears, stained with iron haemo-
toxylin and viewed from the pole. The number found in the
dividing primordial germ cells, twenty-seven, would probably
appear if the chromosomes could be separated so as to permit a
count. Dividing spermatogonia are shown in figure 39.

SEASONAL CYCLE IN PERCH SPERMARY 699

5. Synapsis and maturation divisions

There is no growth period after the spermatogonia are formed.
In figure 3 all cells have been drawn to the same scale and there
is a decrease in size from the largest germ cell to the mature
spermatozoon. In this respect the perch differs from the am-
phibia, from the dipnoans where the primary spermatocytes are
described as nearly three times as large as the spermatogonia,
and from many other vertebrates where there is a considerable
growth. A condition somewhat similar to that in the perch is
found in some of the insects.

Some entire cysts show nuclei in which the contents are dis-
tributed as slender threads. In many of them the threads are
equally distributed (fig. 26), in others there is a contraction |
stage in which the threads have been drawn into a ‘bouquet’
at one pole (fig. 40, a), while in still other stages there is the last
degree of contraction. Here about three-fourths of the nuclei
is clear, while the remaining fourth contains the threads drawn
together into a dense mass (fig. 40, 6).

The small size of the spermatocytes and the tenacity with
which the chromosomes adhere to each other make any detailed
study of the maturation divisions impossible. Figure 41, a and
c, represent a polar and an equatorial view of a primary sperma-
tocyte division in a metaphase. Figyre 41, 6, is an anaphase of
the same division. No data were collected which would point
toward the presence of a sex chromosome. A large number of
dividing spermatocytes were examined, especially in the meta-
phase and the anaphase, but no lagging chromosomes were
observed nor any chromosome proceeding toward the pole more
rapidly than the general mass.

6. Spermatids

The period in which spermatids are present in any one cyst,
like the duration of the spermatocytes in any one cyst, is very
short. The entire period in which they may be found in the
testis, however, lasts from early September to the middle of
December.

700 CLARENCE L. TURNER

After the last division the chromatin collects at one side of the
nucleus. There is a gradual reduction in size accompanied by a
denser accumulation of chromatin at one side of the nucleus
(fig. 42). The nucleus stains more darkly with each stage of
progress toward the mature spermatozoon.

7. Spermatozoa

The mature spermatozoon has a blunt, kidney-shaped head
which stains an intense black with iron-haematoxylin, a small,
lightly staining middle piece and a short tail (fig. 43). The sper-
matids of a single cyst, while transforming into spermatozoa,
collect into masses which resemble parachutes (fig. 44). The
heads all point in the same direction while the tails are drawn
together. As the mature spermatozoa are formed the para-
chutes become more compact, and in fixed material they may
be teased apart without losing their shape. ‘These structures
are probably comparable to the spermatophores described in
some fishes.

The first spermatozoa are formed about September 10th and
are present until their expulsion takes place the following spring.
The expulsion is not complete and a few scattered spermatozoa
are still to be found in the testis during the early summer months,

V. DISCUSSION OF FACTORS INFLUENCING CYCLE

The three critical points in the variation of the testis, volu-
metrically, are indicated in figure 1A. The first occurs in the
latter part of August when the sudden increase in the size of
the testis is started. The second is the beginning of the reduc-
tion in size which occurs about November 1, and the third
occurs about the middle of March when the volume suddenly
begins to drop.

In figure 4 an attempt is made to correlate the internal proc-
esses with the curve in figure 14. Undoubtedly the tremendous
increase in the volume of the testis is contemporaneous with the
formation of the spermatogonia and the rapid subsequent divi-
sions. While the maturation divisions are taking place new

701

SEASONAL CYCLE IN PERCH SPERMARY

‘S]]09 U1IOS JO UOI}BANAVUT UI SosvyS JO UOTJVINP SUIYBIYSN][I BUIMVIG F “SIT

“ BOZULYWaAAdS

“ SdILEWaadS

« SALAOOLUWUAdS

INASSad INOSYLUWAadS
VINODVLYWAAdS oumt NOLLVWYOISNYAL
HL MOY) 2 NOILPATSITOYd +0 dOlddd

GOlaid NOMPYSIN

‘MUW “Gad ‘NUP ‘D4d AON LOO dds ‘Onb AINC INN’ AVW Adv

702 CLARENCE L. TURNER

cells are starting through the procession of changes and divisions
which occur before mature spermatozoa are formed. The mi-
gration of new germ cells into the testis ceases about the first
or the middle of August. The end of this period of migration
may or may not have something to do with the decline in the
volume of the testis which starts at the same time. It is certain
that the sharp decline occurring from March to June is caused
by the expulsion of the spermatozoa.

It is not satisfactory merely to point out internal changes in
the testis and to correlate them with periodical variations in the
volume of the testis unless it be argued that the germ-cell cycle
is an automatic and inherent one which will take place inde-
pendently of the factors of external environment. The very
fact that there are seasonal changes justifies a closer scrutiny of
the seasonal environment of the fish.

The perch is a bottom dweller. A gill net suspended in the
water a few feet above the bottom will trap only a few fish, while
one suspended at the bottom in the same locality will trap great
numbers. It has been found by Prof. A. 8. Pearse that the
perch occur in the greatest numbers in Lake Mendota at the
bottom of the lake at a depth of from 20 to 35 feet. The rec-
ords of the Wisconsin Geological and Natural History Survey
show that the variation in the temperature between the surface
and the bottom of the lake at any given date is a matter of only
two or three degrees. The variation in the temperature at the
bottom of Lake Mendota at a depth of 6 meters is shown graphi-
cally in figure 1B. The curve is based on the mean temperature
for the last fifteen years at a depth of 6 meters. The base line
represents 0 degrees, centigrade, and each space an interval of
5 degrees. The spaces between the ordinates represent inter-
vals of thirty days. During the winter the temperature at the
bottom remains slightly above zero. During the latter part of
March there is a rise in temperature, and the rise continues till
about the third week in August. It will be seen at a glance
that the two critical points are early in the spring and late in
August and-a comparison of these two periods with the critical
points on the curve in figure la shows a coincidence in the criti-

SEASONAL CYCLE IN PERCH SPERMARY 703

cal points of the two curves. It cannot be positively affirmed
from this evidence that changes in temperature are responsible
for the initiation of certain processes in the testis, but it is
significant that the tremendous synthesis of material which takes
place in the testis is started in late August, i.e., in the period in
which the temperature of the water surrounding the perch has
reached its highest point and has begun to decline, and that the
expulsion of spermatozoa takes place at the precise time in
which the temperature of the water is beginning to rise. It has
long been known at the University of Wisconsin that the date of
the spawning of the cisco (Coregonus artedi) may be predicted
rather accurately by following the temperature of the lake.
During the fall of 1916 the temperature of the lake was a little
higher than usual and numbers of ciscos were found by Mr. A.
R. Cahn to be resorbing their eggs. It seems reasonable to
presume that changes in temperature may influence the repro-
ductive processes of the male perch when the reproductive proc-
esses In the female cisco are dependent upon such delicate
changes in temperature.

SUMMARY

1. A great seasonal variation exists in both volume and in the
internal processes in the testis of the perch. The minimum size
occurs from late June to late August. The maximum size is
attained early in November.

2. The testis is divided by connective-tissue partitions into
lobules, but there are no seminiferous tubules.

3. There is a cord of germ cells outside the body of the testis
from which the testis is periodically supplied.

4, An active migration of germ cells occurs from the cord,
outside of the testis, to the ends of the lobules at the periphery
of the testis.

5. The lobules of the testis are gradually filled with an accu-
mulation of germ cells. Transformation of the germ cells into
spermatogonia and the process of spermatogenesis take place
immediately after the accumulation has filled the lobules.

704 CLARENCE L. TURNER

6. During spermatogenesis the lobules are partitioned into
cysts resembling those found in the testis of insects.

7. There is no period of growth after the spermatogonia are
formed.

8. The diploid number of chromosomes is 27.

9. The beginning of the period of spermatogenesis is con-
temporaneous with the beginning of the seasonal reduction of
the temperature of the water in which the perch is found.

10. The expulsion of the spermatozoa occurs at the same time
as the seasonal rise in the temperature of the water. The be-
ginning of the sudden increase in the size of the testis is simul-
taneous with the beginning of the seasonal drop in the tempera-
ture of the water in which the perch is found.

VII. BIBLIOGRAPHY

ALLEN, Ezra 1916 Studies on cell division in the albino rat. Anat. Rec.,
vol. 10.

Agar, W. E. 1911 The spermatogenesis of Lepidosiren. Quart. Journ. Mic.
Sc., vol. 57.

ALLEN, B. M. 1903 The embryonic development of the ovary and testis in
mammalia. Biol. Bull., vol. 5.
1905 The embryonic development of the rete-cords and sex cords in
Chrysemys. Am. Jour. Anat., vol. 5.
1906 The origin of the sex cells of Chrysemys. Anat. Anz., Bd. 29.
1909 The origin of the sex cells of Amia and Lepidosteus. Anat.
Rec., vol. 3.

Barry, D. T. 1910 Morphology of the testis. Journ. Anat. and Physiol.,
vol. 44.

Brarp, J. 1900 The morphological continuity of the germ cells in Raja batis.
Anat. Anz., Bd. 18.
1902a The germ cells of Pristiurus. Anat. Anz., Bd. 21.
1902 b The numerical law of germ cells. Anat. Anz., Bd. 21.
1902 c The germ cells. Part 1. Raja batis. Zool. Jahr., Bd. 16.

BreMER, JOHN Lewis 1911 Morphology of the human testis and epididymis.
Am. Journ. Anat., vol. 2.
1911 Seminiferous tubules of man. Am. Journ. Anat., vol. 2.

Bucuner, Paut 1910 Keimbahn und Ovogenese von Sagitta. Anat. Anz.,
Bd. 35.

Dopps, Gipron S. 1910 Segregation of the germ cells of the teleost, Lophius.
Journ. Morph., vol. 21.

DuesBerG, J. 1917 Chondriosomes in fish embryos. Am. Journ. Anat., vol. 21.

Dustin, A. P. 1907 Recherches sur l’origin des gonocytes ches les amphibiens.
Arch. Biol., T. 23.

SEASONAL CYCLE IN PERCH SPERMARY 705

EIGENMANN, C. H. 1891 On the precocious segregation of the sex cells in
Micrometrus aggregatus. Jour. Morph., vol. 5.

Feperow, V. 1907 Uber die Wanderung der Genital-Zellen bei Salmo fario.
Anat. Anz., Bd. 38.

Feuix, W. 1897 Beitrige zur Entwickelungsmechanik der Salmoniden. Ana-
tomische Hefte, Bd. 8.

Heener, Rost. W. 1914 Studies on germ cells. Journ. Morph., vol. 25.

Huser, G. Cart 1913 Morphology of the seminiferous tubules of mammals.
Anat. Rec., vol. 7.

Huser, G. C., anp Curtis, M. C. 1913 Seminiferous tubules of the adult
rabbit. Anat. Rec., vol. 7.

Huser, G. C. 1916 Notes on the seminiferous tubules in birds. Anat. Rec.,
vol. 11.

Jarvis, May M. 1908 Segregation of the germ cells of Phrynosoma cornutum.
Biol. Bull., vol. 15.

JuNGERSEN, H. F. 1889 Beitriige zur Kenntnis der Entwicklung der Ge-
schlechtsorgane bei den Knochenfischen. Arbeit. zool. zoot. Inst.
Wiirzburg., Bd. 9.

Retzius, Gustar 1905 Die Spermien der Leptokardiei, Teleoster und Ganoi-
den. Biol. Untersuch., N. F., Bd. 12.

1909 Zur Kenntnis der Spermien der Elasmobranchier. Biol. Unter-
such., Bd. 14.

RusascuK1n, W. 1907 Uber das erste Auftreten und Migration der Keimzellen
bei Végelembryonen. Anat. Hefte, Bd. 41.

Swirt, Cuas. H. 1914 Origin of the primordial germ cells in the chick. Am.
Jour. Anat., vol. 15.
WHEELER, W. M. 1899 Development of the urino-genital system in lampreys.

Zool. Jahr., vol. 13.

Woops, F. A. 1902 Origin and migration of the germ cells in Acanthias. Am.
Jour. Anat., vol. 1.

Von BERENBERG-GosSLER 1912 Die Urgeschlechtzellen des Hiihnerembryos
aus 3 und 4 Bebriitungstage. Arch. mik. Anat., Bd. 81.

JOURNAL OF MORPHOLOGY, VOL. 32, NO. 3

PLATE 1
EXPLANATION OF FIGURES

Figures 5 to 22 inclusive, represent a series of camera lucid: drawings show-
ing the volumetric increase and decline during the different seasons. The series
also represents several variations in form.

5and 6 Average size of testis on July 30.
7and 8 Average size of testis on August 30.

9and 10 Average size of testis on September 25.

ll and 12 Average size of testis on October 13.

13 and 14 Average size of testis on October 21.

15 and 16 Average size of testis on November 20.

17 and 18 Average size of testis on January 5.

19 and 20 Average size of testis on March 30.

21 Average size of testis on April 20.

22 Average size of testis on May 5.

706

SEASONAL CYCLE IN PERCH SPERMARY PLATE 1
CLARENCE L, TURNER

PLATE 2
EXPLANATION OF FIGURES

23 Section of pickerel testis near periphery. Con.lis.nuc., connective tissue
nucleus; g.c., germ cells; lob., lobule; sp.m., sperm mass.

24 Section representing part of a lobule in sun fish testis. con.tis.nuc.,
connective tissue nucleus; g.c., germ cells; spg.c., developing cyst of spermato-
gonla.

25 Camera lucida drawing of a single lobule of a perch testis during period
of transformation. XX 680. g.c., germ cells; div.g.c., dividing germ cells:
trans.g.c., transforming germ cells; spg., spermatogonia; div.spg., dividing sper-
matogonia; lob.w., wall of lobule.

26 Camera lucida drawing of peripheral portion of perch testis. X SSO.
Material fixed November 15. prim.g.c., primitive germ cells; spg., spermato-
gonia; div.spg., dividing spermatogonia; sp.td., spermatid; spz., spermatozoa,
syn., nuclei during contraction stage of synapsis; fes.wall, testis wall.

27 Camera lucida drawing of a portion of germ cell cord. X 567. g.c., germ
cells; con.tis., connective tissue.

28 Camera lucida drawing of peripheral portion of perch testis. X 680.
Material fixed July 3. bl.c., blood cells; bl.v., blood vessel; con.tis.nuc., connec-
tive tissue nucleus; con.tis.fib., connective tissue fiber; lob.w., wall of lobule:
m.g.c., Migrating germ cell; s.g.c., stationary germ cell; tes.wall, testis wall;
spr.sp., space formerly occupied by spermatozoa.

* 29 Camera lucida drawing of section of peripheral portion of perch testis
< 680. Material fixed May 5. mz.g.c., migrating germ cell; s.g.c., stationary
germ cell; bl.cell, blood cells; elas. fib., elastic fiber in testis wall.

30 Camera lucida drawing of section of perch testis at periphery. X 680.
Material fixed August 5. ad.t., adipose tissue; bl.v., blood vessel; g.c., germ
cells; lob.wa., wall of lobule; tes.w., wall of testis. (Note: figures 31 to 42 are
drawn to the same scale. )

708

PLATE 2

SEASONAL CYCLE IN PERCH SPERMARY

CLARENCE L. TURNER

Se en

rN
+ yo

#9 984

| Con Tis Rb r

*

bh

PLATE 3
EXPLANATION OF FIGURES

31 aandb. Germ cells in cord outside of testis before period of migration.

32 a,b,candd. Shapes assumed by germ cells during migration to periph-
ery of testis.

33. aandb. Germ cells located at the periphery of testis just after period
of migration.

34 Germ cells during the period of growth and proliferation. a, maximum
size; b, intermediate size; c, minimum size.

35 Polar view of dividing germ cell showing 27 chromosomes.

36 Equatorial view of early anaphase of dividing germ cell.

37 Three stages in the transformation of the germ cells into spermatogonia
a, early stage; b, late stage; c, intermediate stage.

38 aandb. Resting spermatogonia.

39 Dividing spermatogonia. a., polar view of metaphase; b, equatorial view
of metaphase.

40 aandb. Two nuclei in the bouquet stage of synezesis.

41 Dividing spermatocytes. a, polar view of a metaphase; b, equatorial
view of an anaphase; c, equatorial view of a metaphase.

42 Transforming spermatids.

43 Single spermatozoon greatly enlarged.

44 Group of spermatozoa.

710

PLATE 3

SEASONAL CYCLE IN PERCH SPERMARY

CLARENCE L. TURNER

. Sen, GK ESD rewse er ee te OR ROD Og OR thy AF SD
BPN Fea Vey age MaRS 8 Cs REDE NAVAN toe

gp ah ee eS ig Ta Beh A eas Sd aie i A
PEE ERR Re ES ALY ees SES

711

‘ ae) 4
“
.
é .
. =
, fi
? i
7 {
ey sr ea] ‘
fii
x ‘
-
5 4
.
. * ,
‘ ,
fe
' way
” » ~
) ,
; ‘ "
’ ~

SUBJECT AND AUTHOR INDEX

CANTHIAS. The development of the

gastric glands in Squalus............... 351

Activities of the squid, Loligo pealii (Les.)
II. Thespermatophore: its structure, ejac-
ulation, and formation. Sexual

ALLEN, BENNET M. The development of the
thyreoid glands of Bufo and their normal
relation to metamorphosis................

ALLIS, EDWARD PHELPS, JR. Thelipsand the
nasal apertures in the gnathostome fishes .

Ais, EpwarRD PHELPs, Jr. The myodome
and trigemino-facialis chamber of fishes
and the corresponding cavities in higher

379

489
145

VEELC ILA COS 700 eres ie cee ettuee cites cletsreists 207
Apertures in the gnathostome fishes. The
lipsiandthemasalle e-s ae ehieien see cicie 145

Aphelopus theliae (Gahan). The sexual char-
acteristics of the membracid Thelia bi-
maculata (Fabr.). I. External changes
INGUCERN DYE c cere ciate cine eisistn eines eee

ec cae. Studies in the develop-
ment of the opossum (Didelphys virgin-
iana L.). III. Description of new
material on maturation, cleavage, and
entoderm formation. IV. The bilaminar.
Bufo and their normal relation to metamor-
phosis. The development of the thyreoid
BlAN SOL ew me occ aes Se eae ee esle nie eee wlere

HAMBER of fishes and the corresponding
cavities in higher vertebrates. The
myodome and trigemino-facialis........

Changes induced by Aphelopus theliae (Ga-
han). The sexual characteristics of the
membracid,..Thelia bimaculata (Fabr.).

489

SE xternalllt 38.0 5 Soe eran eee ite 531

Characteristics of the membracid, Thelia bi-
maculata (Fabr.). I. External changes
induced by Aphelopus theliae (Gahan).

MVE SOR oN rece Wve ntinealahte oe eI o a aes 531

Chromosomes in the spermatogenesis of the
stonefly, Perla immarginata Say, with
special reference to the question of synap-
sis.

Cbromoemes of the tiger beetles (Cicindeli-

ae).

(Cicindelidae). A comparative study of the
chromosomes of the tiger beetles...........

Cleavage, and entoderm formation. IV. The
bilaminar blastocyst. Studies in the de-
velopment of the opossum (Didelphys vir-
giniana L.). III. Description of new
material on maturation...................

sonal

EVELOPMENT of the gastric glands in

Squalus acanthias. The............... 3

Development of the opossum (Didelphys vir-
gininaL.). III. Description of new mate-
rial on maturation, cleavage, and ento-
derm formation. IV. The bilaminar
blastocyst. Studies in the...............

Development of the thyreoid glands of Bufo
and their normal relation to metamorpho-
Sissey De years cai alselgninacie manasa

AtStudyzonbhere-cenese Seen ences 5

A comparative study of the....... 437

437

681

(Didelphys virginiana L.) III. Description
of new material on maturation, cleavage,
and entoderm formation. IV. The bi-
laminar blastocyst. Studies in the devel-
opment of the opossum..

Drew, Gi~MAN A. Sexual activities of the
squid, Loligo pealii (Les.). II. The sper-
matophore: its structure, ejaculation, and

FOLMALLOME jac aet onoe a Ske nisin DO ates chro
NTODERM formation. IV. The bilam-
inar blastocyst. Studies in the devel-

opment of the opossum (Didelphys
virginiana L.). III. Description of new
material on maturation, cleavage, and....

ACIAL musculature of the Japanese.
Ht I oY paces eile eae neice aarti an Sri eerie

Fishes and the corresponding cavities in higher
vertebrates. The myodome and trigem-
ino-facialis chamber of

Fishes. The lips and the nasal apertures in
the enathostometa.c nse seen eee

Formation. IV. The bilaminar blastocyst.
Studies in the development of the opos-
sum (Didelphys virginiana L.). III. De-
scription of new material on maturation,
cleavage, and entoderm...................

ASTRIC glands in Squalus acanthias.
The development of the...............

Glands of Bufo and their normal relation to
metamorphosis. The development of the
thy Trev eae wie heen ee ee erect eters

Gnathostome fishes.
ADEIECUTER AIM GNEN eer ceva nig ere eas on hele

GoupsmitH, WittiaM M. A comparative
study of the chromosomes of the tiger
beetles (Cicindelidae)...............-..5--

ARTMAN, Cart G. Studies in the de-
velopment of the opossum (Didelphys
virginiana L.). III. Description of

new material on maturation, cleavage,

and entoderm formation. IV. The bi-
laminamiblastocystesaqves cece eerie
Hofbauer. On the nature, occurrence, and
identity of the plasmacells of............

Al oes: The facial musculature of
t

ORNHAUSER, Srpney I. The sexual
characteristics of the membracid, The-
lia bimaculata (Fabr.). I. External
changes induced by Aphelopus theliae
(Gahan) Sone anene bea aetees < eee

Kupo, Toxkuyasu. The facial musculature

Of the Japaneses cc Me cet e emis «id ateleeiors

IPS and the nasal apertures in the gna-
thostome pshesy elbhetae emer cr cee

379

207

=

637

714

ATURATION, cleavage, and entoderm
formation. IV. The bilaminar blas-
tocyst. Studies in the development of

the opossum (Didelphys virginiana L.)

III. Description of new material on... 1

Membracid, Thelia bimaculata (Fabr.). I.
External changes induced by Aphelopus
theliae (Gahan). The sexual characteris-
ticsiof theteeeaw cee te ee ee
Metamorphosis. The development of the
thyreoid glands of Bufo and their normal
relation tO ews tio ae ee ee
Meyer, ARTHUR WILLIAM. On the nature,
occurrence, and identity of the plasma
cellsiof slotbauers..caee ee ee ee
Musculature of the Japanese.
Myodome and trigemino-facialis chamber of
fishes and the corresponding cavities in
higherivertebrates) “Ghe!..).:..-..e:eee

Noe Waro. A study on the
chromosomes in the spermatogenesis of
the stonefly, Perla immarginata Say,
with special reference to the question of
SVMAPSISM Sik. oc eE Se ee ee eae
Nasal apertures in the gnathostome fishes.
Wheplipssandiithey eae. sweet eae

489

327
637

207

509
145

Orn oe (Didelphys virginiana L.).
III. Description of new material on
maturation, cleavage, and entoderm for-
mation. IV. The bilaminar blastocyst.
Studies in the development of the........ 1

pecs The seasonal cycle in the sperm-
APVROMUNC soe csc Nate oe ee eee

Perla immarginata Say, with special reference
to the question of synapsis. A study on
the chromosomes in the spermatogenesis
ofsthemmpnetiv. ses sn ont ane meee ee

Plasma cells of Hofbauer. On the nature, oc-
currence, and identity of the............. 32

eee. AvourH R. The development
of the gastric glands in Squalus acan-

thias 351

INDEX

EXUAL activities of the squid, Loligo

pealii (Les.). II. The spermatophore:
its structure, ejaculation, and formation. 379
Sexual characteristics of the membracid, The-
lia bimaculata (Fabr.). I. External
changes induced by Aphelopus theliae
(Gahan)! “They, 12) 2:8 Os! ee eee
Spermary of the perch. Theseasonal cycle in

531

681

Spermatogenesis of the stonefly, Perla immar-
ginata Say, with special reference to the

A study on the

question of synapsis.
chromosomes in thé
Spermatophore: is structure, ejaculation, and
formation. Sexualactivities of the squid,
Loligo peal (Ges:)) Ik. ‘Lhe. 2.23
Squid, Loligo pealii (Les.). II. The sperma-
tophore: its structure, ejaculation, and
formation. Sexual activities of the......
Stonefly, Perlaimmarginata Say, with special
reference to the question of synapsis. A
study on the chromosomes in the sperma-
togenesis ofthe sizn ane eee one eee
Synapsis. A study on the chromosomes in the
spermatogenesis of the stonefly, Perla im-
marginata Say, with special reference to
theiquestion.ofy.. Mtoe see ee eee ee

379

509

fee bimaculata (Fabr.). I. External
changes induced by Aphelopus theliae
(Gahan), The sexual characteristics of
whe smenrbracid yeas alae oe eee ee
Theliae (Gahan). The sexual characteristics
of the membracid, Thelia bimaculata
(Fabr.) I. External changes induced by
Aphelopus is: i202 see ae eee
Thyreoid glands of Bufo and their normal re-
lation to metamorphosis. The develop-
ment: ofjthess sets. sear eee eee
Tiger beetles (Cicindelidae). A comparative
study of the chromosomes of the......... 437
Trigemino-facialis chamber of fishes and the
corresponding cavities in higher verte-
brates. The myodome and
TURNER, CLARENCE L. The seasonal cycle
in the spermary of the perch............. 681

531

489

fe

o

oa

BL WHOI Library -

IIMA

5 WHS EE 04709