Full text of "The Journal of experimental zoology"

See other formats

5 ge ai ees

savas oghet

ig ace Beh S_ 9 P98
segs arate aes
ies
Neal ad ~

fs!

oe eee ee
well eh i er er ae

aid
ms
ot)

~ ee ee
eRe Yas Yate fa Tata had ae
ella ._s ¥

al taek a ha A

Se AA te nee 4 2.33.8 bb ee eb eb ee
}_a_o._6. 5 _o_b_b_ See
oe at

hh te
BE at ee tl ee
Pe ae ee ee
bb eee ee a od
te te AP tee ae nt te a

St ee

we
ee el tt ede

had eal 2. ) AD A
Rut ghee!
yee

pe ee £
VS

(<7-s a
el et at tae

erst
Fh ec de ee Pe t_6s 98
. Fae ob 6 ite
a PaaS J
aaeS,

PF PP PD r
Pe BP ae Fo PD

2A VEAP 59

eee nae

seieiea tee.

ST eerie Heater eree

THE JOURNAL

EXPERIMENTAL ZOOLOGY

HDi D By

WiuiiaM Ef). Caste Jacques Lors

Harvard Univei “7 The Rockefeller Institute
Epwin G. CoNKLIN Epmunp B. WILSON

Princeton University Columbia University
Cuarues B. DAVENPORT Tuomas H. Morgan

Carnegie Institution Columbia University
HERBERT S. JENNINGS GEORGE H. PARKER

Johns Hopkins University Harvard University
Frank R. LILLIE RAYMOND PEARL

University of Chivago Johns Hopkins University

and

Ross G. HaArrRIson, Yale University

Managing Editor

| VOLUME 32
JANUARY—APRIL, 1921

|
|

THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY
PHILADELPHIA, PA.

; i a \
. a
< j }
N if - « | A, ay \
5 ae Aa inhvaSe
} {, , _ ‘? ier
‘, , eg | < Aa j
WY A ys cn % mice ; é
f Sat PA ae
rb df ee

CONTENTS

No. 1. JANUARY

Ross G. Harrison. On relations of symmetry in transplanted limbs. One
huncredvand: thintvestxaneuUees cys... a ie sete atlas ake es eens 1
JosePpH Hatut Bopine. Factors influencing the water content and the rate
Of metabolism oficertain orthoptera. Six figures .,........0742.-.54.0. 137
CHARLES W. Merz anp Jost F. Nonrprz. Spermatogenesis in the fly, Asilus
sericeus Say. Two plates (twenty-two figures) ..................0005. 165

No. 2. FEBRUARY

Haroup H. Proucu. Further studies on the effect of temperature on cross-

INCOVeLwMbhree-fromesre waa os ates eerie: Seas ee eee 187
W. E. BurGcr anp E. L. Burak. An explanation for the variations in the
intensity of oxidation in the life-cycle. One figure.................... 203

Henry LAvRENS AND S. R. DetwiterR. Studies on the retina. The struc-
ture of the retina of Alligator mississippiensis and its photochemical
changes) “lhirteen timbres ere, 5s, cs. myers aise sod oer ae nee en eae 207

Wivsur WILuis SwincLe. The germ cells of anurans. I. The male sexual
cycle of Rana catesbeiana larvae. Two text figures and fifteen plates

(one -hundredvand thighy-one figures): 02.2.0). «5.8 de nosedevec dees nwes 235
J. A. DETLEFSEN AND E. Roperts. Studies on crossing over. I. The effect
of selection on crossover values. Two text figures.................... 309

No. 3... APRIL

RutxH B. Howiann. Experiments on the effect of removal of the pronephros

of Amblystoma punctatum. Mwenty-three figures)... .. o26. 2202 4.. (000
Wm. A. Kepner AnD W. Cart WuttLock. Food reactions of Ameba proteus.

Si plates. (twenty-one memes ie jee sce ccstencs koe ene sue sen cee cae 397
ALFRED O. Gross. The feeding habits and chemical sense of Nereis virens

USP R NE SCs 22 « 2c eS et REE ok RE, cai Ne ee ame a 427
Lesure B. Arny anp W. J. Crozrer. On the natural history of Onchidium. 443

iil

(kee | 15

ee i e fee He ee Pile 7
4 a4 - ; x ve als? .
r, be Bee a ‘i. es

| hee ge pow alg

‘9 awe :

“a - i ti = s
< js ya

_ aE, % y Ki as j : / =i oe TA = is 1
wy) Pra se tr) Aleit’. Ris mays rhe ia
i i ao ’ a oY
? ‘ 2 \ ie b
Aba Gel, tae Se phe Bye
‘ ; ear mi i ;
Aitia® \ +0 ay & HSS f
‘
i 7) ’
- 2 ‘ve , 5 -
/ : = Pee “st }
+ Wy x
7
‘ a ‘"
f 4 aha
, 4 pe 3 sia" 3. 2
j \ ie
{ io 7 ie Ie
£ ’
> ‘ $e
é yoy ea: ~ aes : / j -
het
tat jy pps, iA one }- " yey oat |
A LISP Sal 4 Z ! | ne
‘ , oe rac’ | es aha
. } :
: , ries Shain ee
: :
o- id)
= ihe? < ey, iY? Sb sd
pid te
8
\ 2 j
4 ; ee
jhe ro j
ix nny brah
, '
ir “ : Pr ive j
a ’ ;
Ny : =v
Jy F j
PAln =
‘
of
me i
} + a8
: :
| } ry
ia é

id *
h sPeet
my " :
" Pray
‘ P
* re 4
’ ‘i
or ‘ ol
Milse
- = 4
q
‘4
as r
ih
vai
3 4
ye } ? = ;
* 1
-
i a = i
*, '
rs
&

JANUARY, 1921

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 1

oe? Geel,

= iy
=

i
> oi
=a =

Resumen por el autor, Ross G. Harrison,
Yale University.

Sohre las relaciones de simetria en los miembros transplantados.

El autor ha llevado a cabo los siguientes experimentos con
los esbozos de los miembros anteriores de Amblystoma: Trans-
plante a la superficie lateral del cuerpo de otro embri6n; trans-
plante al sitio normal después de extirpar el esbozo del miembro
que le ocupaba; superposicién de un esbozo sobre otro después
de separar su ectodermo; y transplante de medio esbozo en el
sitio en que existia otro medio esbozo, extirpado previamente.
Los injertos fueron obtenidos en el mismo lado del cuerpo o-
en el lado opuesto e implantados con el eje dorso-ventral situado
normalmente o invertido. Los dos primeros grupos de experi-
mentos indican que cuando el eje dorso-ventral del miembro
no esta invertido persiste la simetria prospectiva originaria, y
que cuando se invierte, la simetria se invierte también (simetria
enantiomorfica). La asimetria esta determinada: 1) Por la
polarizacion del eje antero-posterior del esbozo y 2) Por la
orientacién del esbozo respecto a la polarizacién dorso-ventral
del ambiente orgdinico. Dos combinaciones (armé6nicas) pro-
ducen miembros con la asimetria correspondiente al lado en que
se colocaron; otras dos (desarménicas) producen miembros con
asimetria opuesta. Se encuentran con frecuencia reduplicaciones
el miembro primario sigue en este caso las reglas mencionadas,
mientras que los miembros secundarios simulan sus imagenes
producidas por un espejo, y a veces duplicados a su vez. Existe
conformidad con las reglas de Bateson, las cuales sin embargo
pueden enunciarse mas simplemente para incluir los supernumer-
arios sencillos y dobles. En los experimentos con mitades de
esbozos superpuestas, salvo ciertas excepciones, las combinaciones
armonicas producen miembros sencillos y las desarménicas
reduplicaciones de acuerdo con las reglas. El mesodermo del
esbozo del miembro es un “‘sistema arm6nico equipotencial”’
y excepto para la determinacién de ciertas relaciones axiales,
es autodiferenciable. Su forma, incluso sus relaciones simétricas,
debe estar representada en su estructura intima.

Translation by José F, Nonidez
Cornell Medical College, New York

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, NOVEMBER 1

ON RELATIONS OF SYMMETRY IN TRANSPLANTED

LIMBS

ROSS G. HARRISON
Osborn Zoological Laboratory, Yale University

ONE HUNDRED AND THIRTY-SIX FIGURES

CONTENTS

OIG HOM... ::: < Peer sera eriee oOo ce ec R ie Seite ince fe 5 See oantaa tat
Method svanduserminglocyeri ate Sie tyne ome rare oe ie eee ees eee ee
General features of the development of the transplanted buds..............
XPS GIN ental ets <<, ., AeMPIeNreey wer Ane Oy see err ee cads a Slat a tall hiro ae

A. Limb

buds implanted in abnormal location—heterotopic trans-

PLTGAtIONS eee ne eas ecco Pen een cle as be crea eae

Ot PB OO

B. Limb
th

10.

11.

12.

. Homopleural transplantations, normal or dorsodorsal orien-

FCO 1 RR ce, IL Te Ss ee is ep toe Reng LN eh 2

. Heteropleural transplantations, dorsodorsal orientation........
. Heteropleural transplantations, dorsoventral orientation........
. The shoulder-girdle in heterotopic transplantations.............

Summary of the results of heterotopic transplantations.........
buds implanted in natural location—orthotopic transplantation.
Homopleural transplantations, dorsodorsal orientation..........
Homopleural transplantations, dorsoventral orientation.........
Heteropleural transplantations, dorsodorsal orientation.........
Heteropleural transplantations, dorsoventral orientation........
The shoulder-girdle in orthotopic transplantations..............
Summary of the results of orthotopic transplantations..........

C=cuperposed: limb. bildsuwmewer tte. 02 cares cc akin ods atl bioua. nemotes os

13.
14.
15.
16.
17.

Homopleural transplantations, dorsodorsal orientation...........
Homopleural transplantations, dorsoventral orientation.........
Heteropleural transplantations, dorsodorsal orientation.........
Heteropleural transplantations, dorsoventral orientation........
Discussion of experiments with superposed limb buds...........

Dwiransplantation of halfipmusererne sso. she ce ate Seek sone ee ek

18.
19.
20.
21.

11
14

Ale

1g)
21
25
2d
29
29
30
31
43
55
57

2 ROSS G. HARRISON

Generaldiscussions ic cckac:sescmie ce oek lo Shea RIO oles 5) oe ee 83
E. The rules of symmetry.............. He oA PE OF Seo ese 83

F. The mode of representation of symmetric relations in the limb
TUGTNENG fo 28h. ee oo eco nd Fo a ee 85
G. Reduplication and the problem of polarity and heteromorphosis. 92
H. Form regulation and function in transplanted limbs..,.......... 102
Generalesummary. co cae soe ee ee en in eo 109
Iseirob references fe .2 2s oS Fe 4 ee occ es 0 ee er 115
Appendix—Histories of selected individual cases..............-.2...++00.00. 119

INTRODUCTION

The circumstance that originally suggested the present study
was the apparent difference in the results obtained by Streeter
(07) and by Spemann (710) in their respective experiments with
the amphibian ear vesicle. According to the original account
of Streeter, the otocyst, when taken out of an embryo just after
closure and replaced after having been rotated 180° on any of
its axes, develops in normal posture, though a right vesicle placed
on the left side remains true to its side of origin. According to
Spemann, the inverted vesicle develops in inverted position, the
rudiments of the constituent parts being localized, at the time
of operation. Although subsequent work by Streeter (14) seems
to have shown that the normal development of the inverted
vesicles, found in his cases, was due to their regaining normal
posture by rotation as a whole, the original divergence of results
nevertheless had raised theoretical questions of great interest,
which Spemann has ably discussed. The main question was
whether we might have in the otic vesicle an ‘harmonic equipo-
tential system’ with its future asymmetry in some way stamped
upon its intimate structure. Though Spemann’s analysis an-
swers the question in the negative, as far as the closed ear vesicle
is concerned, it is nevertheless important to determine how far,
if at all, systems of this kind are present in the embryo, for their
study would throw light upon the question of the mode of rep-
resentation of adult form characters in the germ, giving evidence
from a new quarter with regard to the great problems of devel-
opment which have usually been approached by way of experi-
ments upon the unsegmented egg and the early stages of cleav-
age. The method of embryonic transplantation obviously

SYMMETRY IN TRANSPLANTED LIMBS 3

affords a means of studying this question in any organ or part
that in the adult lacks a plane of symmetry. It was with this
purpose in view that the present experiments with the limbs of
Amblystoma were begun. Limb buds were implanted in both
normal and abnormal location, oriented in various ways with
respect to the main axes of the embryo-host, and the form and
posture of the resulting limbs were studied.

It became evident, after the first experiments were made, that
the rudiment of the fore limb behaved differently from the audi-
tory vesicle, no matter whether Streeter’s original interpretation
or Spemann’s was accepted as correct. While it was found that
a certain tendency did exist for inverted limb buds to rotate back
to normal posture during development, this was not the usual
result, nor did the rotation take place in the sense meant by
Streeter in his later publication (14). Furthermore, many
irregularities of development were produced by the operations,
due largely to the power of the limb rudiment to duplicate itself
by budding. On the other hand, it often occurred that buds
transplanted from one side of the body to the other developed in
harmony with their new surroundings, a right limb bud, for ex-
ample, placed on the left side, giving rise to a normal left limb.

The earlier experiments which were made in 1911 and 1912 led
to no satisfactory general conclusion, so that publication of the
work was deferred pending further investigation. Subsequently
numerous additional experiments were made, in which more ef-
fective precautions against regeneration of the limbs from the
host were taken.2 The situation began to clear when in some of
the cases in which the asymmetry of the limb was reversed by

1 Cf. Barfurth (94), who showed that supernumerary limbs could be produced
in amphibians by regeneration after irregular amputation; Tornier (’05), who
obtained multiple appendages by cutting into the limb bud of tadpoles; Braus

(04, ’05, and ’09) and Harrison (’07), who found that transplanted limb buds
frequently give rise to double limbs.

2 The first experiments were reported to the National Academy of Sciences
in November, 1912, at the New Haven meeting. Later reports were made before
the American Association of Anatomists in December, 1915 (’16), and before the
American Society of Zoologists in the following year (’17 a). The main results
have been stated somewhat more fully in the Proceedings of the National Acad-
emy (715 and 717 b).

4 ROSS G. HARRISON

transplantation it was observed that the reversal came about
by a process of reduplication or twinning. By following closely
the history of individual cases, it became evident that the double
formation was not infrequently obscured by the preponderance
of the reduplicating limb bud over the original, so that the former -
grew into a member of opposite asymmetry, while the original
bud was reduced to a mere spur or nodule, which might readily _
be overlooked. The tendency to produce duplicities thus proved
to be even greater than the actual number of fully developed
cases indicated. In other cases the reversal appeared to be more
direct; at least, a limb of the side of origin often failed to appear
as such on the surface, though slight irregularities in the early
stages of development, coupled with an appreciable delay in the
process, showed that some internal readjustment of the grafted
tissue was taking place.

It. seemed that the functional activity of the limb, when trans-
planted to its normal environment, might accentuate the appar-
ently anomalous results just described. In order to eliminate
this factor, a series of experiments in which the limb bud was
grafted on some other part of the body was undertaken. Here
the proper nervous connections did not become established, func-
tional activity was usually lacking or was at best but slightly
developed, and the undisturbed effect, upon development, of the
relative orientation of the tissues of graft and host could be ob-
served. The latter experiments led to the formulation of the
three following simple rules,? which hold for implantations either
in normal or in abnormal location:

Rule 1. A bud that is not inverted (dorsodorsal) gives rise to
a limb of the side of origin of the bud, whether implanted on the
same or on the opposite side of the body.

Rule 2. An inverted bud (dorsoventral) gives rise to a limb
of reversed asymmetry, whether implanted on the same or on
the opposite side of the body.

Rule 3. When double limbs arise, the original one (the one
first to begin its development) has its asymmetry fixed in accord-

3 These rules were phrased somewhat differently in two preliminary communi-
cations (717 a and 717 b).

SYMMETRY IN TRANSPLANTED LIMBS o

ance with rule 1 or 2, while the other is the mirror image of the
first.

Experiments previously reported‘ have shown that the limb bud
is an ‘harmonic equipotential system,’> and additional experi-
ments with inverted buds (p. 87) and with half buds (p. 83)
confirm this result. We must assume, then, that the potencies
of the cells of the limb bud to form the fore limb are in the last
instance represented in their intimate structure and not merely
in their arrangement. ‘The above rules show, however, that not
all essential features are stamped upon the constituent elements
of the rudiment at the time of transplantation. For example,
the difference between the right bud and the left is not an abso-
lute one, since a right limb bud upside down behaves like a left
one right side up and vice versa. From this the conclusion has
been drawn that the elements making up the limb bud are dif-
ferentiated in an anteroposterior direction, 1.e., along the antero-
posterior axis, but are not yet differentiated, at least not irre-
versibly, along the dorsoventral axis at the period of development
at which the transplantations are made. In this one respect the
differentiation of the limb is dependent upon its orientation with
reference to the dorsoventral axis of the embryo; otherwise as
regards its specific form, the limb bud constitutes a self-differenti-
ating system.

These questions will be considered more fully in the concluding
section (p. 85). :

METHODS AND TERMINOLOGY

All experiments were made upon embryos of Amblystoma
punctatum in stages that have been previously defined.®

In performing the operations the embryo which is to receive
the implanted limb bud is first made ready. If the bud is to be
placed in normal location, the wound is prepared as in the extir-
pation experiments referred to above. A circular incision, hay-

4 Harrison, 718.

> “Jedes kann Jedes und alles Einzelne steht in Harmonie zu einander.’’

(Driesch, ’02, p. 229.) See also: 799, p. 72; ’05, p. 679, and ’08 b, p. 120.
6 Harrison, 715 and ’18.

6 ROSS G. HARRISON

ing the diameter of three and a half somites (ca. 0.9 mm.) and
ventral to the third, fourth, fifth, and half of the sixth myotome,
is made, and the dise of tissue, including both mesoderm and
ectoderm, is lifted, after which the remaining mesoderm cells are
carefully cleaned off. ‘This may all be done without injury to
the pronephros, which lies immediately dorsal to the limb rudi-
ment, though if this organ is injured or even extirpated, there is
no noticeable effect on the subsequent development of the limb.
The embryo thus prepared is held in readiness for the grafting,
being secured in position by pieces of silver wire or glass rod bent
into proper shape. The limb bud which is to be transplanted is
removed from another embryo, as described above, care being
taken in lifting it from its bed to take all of the mesoderm pos-
sible. It is then transferred on the point of the scissors to the
first embryo and fitted into the wound. The orientation of the
bud is important and may be earried out as desired by noting
the position of pigment markings in the ectoderm of the graft.
After it has been properly placed, it is held in position for an
hour or more by a single piece of glass rod, bent into such shape
that it straddles the embryo, exerting a light pressure upon the
grafted tissue. The healing of the wound takes place readily,
though frequently a small area of underlying yolk may be left
exposed on the border of the wound. ‘This usually heals in a day
or two and does not seem to influence the result of the experi-
ment, unless the yolk begins to disintegrate, in which case death
of the embryo usually follows.

At the time when the first experiments were made, the condi-
tions necessary to prevent regeneration had not been determined,
so that in a number of cases the extirpated area was too small
or the wound bed was insufficiently cleaned of scattered meso-
derm cells for the result to be conclusive. These experiments
have not been included in the tabulations, but will be referred to
separately in so far as theyare of special interest. In all of the
later experiments the size and character of the wound in the host
was such as to preclude regeneration of the limb from that
source; the resulting limbs must, therefore, have arisen from
the engrafted tissue. Even in the cases where the wound was

SYMMETRY IN TRANSPLANTED LIMBS 7.

not especially cleaned, there is no evidence, aside from certain
exceptional cases, that the tissue of the wound bed displaces the
transplanted bud, though the possibility of its participation in
the make-up of the limb cannot be excluded, and it probably
actually does take place to some extent.

In transplanting the limb bud to a location other than the
normal, the recipient embryo is first prepared as in the other
experiments. A wound of proper size is made, usually in the
flank just below the ventral border of the myotomes, and the bud
grafted in the same manner as described above. In doing this
it is well to avoid injury to the pronephric duct (p. 15).

Three different factors regarding the placement of the limbs
have been considered in these experiments, viz.: 1) location of
the graft in the embryo; 2) the side of the body on which it is
placed (whether the same from which it was taken or the oppo-
site); and 3) orientation with respect to the cardinal points of
the embryo. The experiments thus fall into eight categories,
as follows (fig. 1):

A. Limb buds placed in abnormal location—heterotopic
transplantation.

1. On the same side of body as origin—homopleural (hom.).

a. Dorsal border of limb bud dorsal with respect to embryo—
dorsodorsal (dd.).

b. Dorsal border of limb bud ventral with respect to embryo—
dorsoventral (dv.).

2. On side of body opposite to origin—heteropleural (het.).

a. Dorsal border of limb bud dorsal with respect to embryo—
dorsodorsal (dd.), in which case the anteroposterior axis isreversed.

b. Dorsal border of limb bud ventral with respect to embryo
dorsoventral (dv.), in which case the anterior and posterior points
of the graft correspond, respectively, to those of the embryo.

B. Limb bud placed in natural location—orthotopic trans-
plantation.

The several categories under this head as under A.

According to the rules stated on page 4, two of the combina-
tions (homopleural dorsodorsal and heteropeural dorsoventral)
yield limbs which are of the same side of the body as that on

8 ROSS G. HARRISON

which they are placed (fig. 2), for the non-inverted limb bud
from the same side (hom.dd.) does not have its prospective asym-
metry changed while the inverted limb bud from the opposite
side (het.dv.) does. The limbs which develop in these combina-

D |
|

P RA).
Vv \

ORTHOTOPIC

D D D D
D v D Vv

PP RAJA P(A R pia PIA L PJA PIP L AJA
Vv D Vv Dy
Vv v Vv Vv
HOM. HOM. HET. HET.
D-D o-Vv D-D ov

Fig. 1 Diagram showing the eight different operations. The outline of an
Amblystoma embryo in the operating stage is shown above. The circles within
it represent the limb bud, in the normal (orthotopic) and the abnormal (hetero-
topic) location. The four circles below represent the four different ways in which
limb buds may be oriented with reference to the cardinal points of the embryo;
the letters (D, dorsal; V, ventral; A, anterior, and P, posterior) within the
circles designate the original cardinal points of the transplanted limb, those
outside the corresponding points of the embryo. The operations are represented
to be on the right side. R, right limb bud; ZL, left limb bud; hom., homopleural;
het., heteropleural.

tions thus fit in with their surroundings; they have therefore been
ealled harmonic. The other two combinations (homopleural
dorsoventral and heteropleural dorsodorsal) give rise to limbs
of the side opposite to that of their seat of implantation, for the
inverted bud from the same side (hom.dv.) has its asymmetry
reversed, while the non-inverted bud from the opposite side

SYMMETRY IN TRANSPLANTED LIMBS 9

(het.dd.) remains as it was. The limbs which develop here are
not primarily in harmony with their surroundings, so that these
combinations have been termed disharmonic.

“AIA--- MIRROR PLANE IT

HOM.
o-V

/
|
| |
3
/
\
; .
'
'
' ;
, /
i i

MIRROR

PLARE II
__ MIRROR

PLARE I

Fig. 2. Diagram showing the results of the four operations, heterotopic or
orthotopic, represented as on the right side of the embryo. The circles indicate
the transplanted limb buds, the letters having the same significance as in figure 1.
Thus the two upper figures in the diagram represent homopleural, and the two
lower ones heteropleural transplantations. The two on the left show the trans-
planted bud in upright (dorsodorsal) orientation while the two on the left are
inverted(dorsoventral). The limbs which develop are shown in profile, the ulnar
border being uppermost (dorsal) in all which actually develop. A heavy outline
indicates the primary member, a light outline the reduplicating one. It is to
be noted, however, that the latter develop in by no means all cases, while the
former may be resorbed in the heteropleural dorsoventral combination, leaving
only the reduplicating member present. The broken outlines show the posture
that the limb would have assumed,had it developed as a self-differentiating mem-
ber totally independent of the influence of its surroundings.

The transplanted limb bud is a flattened dise of tissue, and it
is theoretically possible to make eight further combinations by
placing the medial surface of the graft corresponding to the lateral
surface of the embryo. This is impracticable, however, because

10 ROSS G. HARRISON

the mesoderm would thus be brought to the surface and the ecto-
derm buried beneath it. The same effect might be obtained,
however, by transplanting the mesoderm alone. While the dif-
ficulties in this procedure are great, they have now been in a large
measure overcome. The positive experiments are too few in
number to warrant any very definite statement, but they do indi-
cate that it is immaterial which surface of the mesodermal dise
faces laterally.

A much greater variety of experiments could be had by experi-
menting with positions intermediate between the upright and
inverted positions, i.e., with limb buds turned, say, 90° instead
of 180°. Such experiments may yield very interesting results,
but as yet there has not been sufficient time to carry them out,
nor has the effect of implanting the limb exactly in the midline
been studied.’

The experiments with superposed buds were made in the same
way as the above, except that the mesoderm of the host was not
excised. In the case of half buds, more combinations are pos-
sible, as described in the section dealing with this group. Both
here and in the superposition experiments all possible positions
with regard to the placement of the graft within the limitations
stated above were experimented with. Relations of harmony and
disharmony proved to be the same here as in the case of whole
buds.

The total number of cases of which records have been kept is
462. The analysis is based, however, upon the 271 individuals
which yielded positive results. The identity of the individual
cases has been maintained by rearing each in a container by
itself and keeping a separate history of each. These histories
consist in notes and in sketches made from time to time directly
from the living specimens, mostly with the aid of the camera
lucida.

In dealing with so large a mass of material it has of course been
necessary to select typical cases for presentation, and in order
not to interrupt the continuity of the general account, the indi-
vidual histories, as far as given, have been gathered together in
an appendix. The main body of the paper has been divided in

SYMMETRY IN TRANSPLANTED LIMBS ila

accordance with the outline presented above. The larger groups
of experiments have been considered apart from each other, and
each subgroup is treated in a special section. The peculiar fea-
tures of each of the larger groups have been considered at the
beginning, and the results of the experiments summarized sep-
arately at the end of each main section. The more general
questions are treated in the final chapter.

It has been thought best to provide numerous illustrations in
order to avoid lengthy descriptions. Since it was not possible
to keep a complete pictorial history of each case, those were se-
lected for drawing that promised typical or otherwise interesting
results. Unfortunately, however, it was not always possible to
predict what the outcome of an experiment would be, so that
some important cases were not drawn in early stages, while
others of less interest were.’

GENERAL FEATURES OF THE DEVELOPMENT OF THE
TRANSPLANTED BUDS

The development of the transplanted limb buds must now be
considered in comparison with normal development. When the
normal limb bud appears it is a round prominence just below the
pronephros. It soon becomes more sharply marked off from the
background and begins to ‘point’ dorsoposteriorly.’ The radial
border of the fore arm and hand is at first ventrolateral, then ven-
tral, and the first digits to arise are the first and second. The
third and fourth digits appear later on the dorsal border of the
hand, so that there is never any difficulty in distinguishing the
ulnar from the radial border unless the third and fourth digits
are entirely suppressed. The palmar surface of the hand faces
at first ventromedially and later medially.

The transplanted limbs, both heterotopic and orthotopic,
give evidence of their orientation early in development, inasmuch

7 Almost all of the preliminary sketches and many of the finished drawings
were made by Miss Lisbeth Krause. The former, which were pencil sketches,
had to be redrawn for reproduction. For this part of the work and also for a
number of the original drawings I am indebted to Mr. A. Hemberger and Mr. H.

D. Rhynedance.
8 Harrison, 718, p. 419.

12 ROSS G. HARRISON

as their direction of ‘pointing’ is determined principally by the
bud itself. In two of the combinations (homopleural dorsoven-
tral and heteropleural dorsodorsal), they point anteriorly or
dorsoanteriorly; in the other two (homopleural dorsodorsal and
heteropleural dorsoventral) posteriorly or dorsoposteriorly like
the normal. The subsequent development in the latter case is
normal, but in the former there is a tendency for the limb to stick
out more sharply to the side or to rotate more or less from the
position in which it would be found were the position determined
entirely by the orientation of the bud itself. Nevertheless, the
palm tends to face ventromedially, or else the limb is so rotated
that it faces more ventrally or anteriorly. In order to determine
whether the limb is right or left, it is necessary to be able to dis-
tinguish between the palm and the back of the hand, which is
not always so simple as it might seem. It can usually be done,
however, by noting the digits, which are frequently slightly
flexed. When there is uncertainty, it 1s necessary to resort to
sections, in which case there is no difficulty in distinguishing be-
tween the two faces, because of the much greater thickness of
the soft parts on the flexor surface of the skeleton.

The duplicities that arise are of all grades and kinds, and occur
in very different proportions in the several experiments. Some-
times they make their appearance very early, sometimes late in
development. In the orthotopic grafts reduplication is far more
common when the developing limb and the substratum are of
opposite sides. In such cases the doubling member nearly always
appears as a bud posterior to the main limb, growing there into
a limb of proper asymmetry. The extent of reduplication may
include the whole limb from the shoulder down, or only certain
of the digits. The duplicate limb is as if it were mirrored from
the original in a plane which is perpendicular to the plane of the
proximodistal axes of the two limbs® and which cuts the axes
of the two limbs at their junction, at an angle which varies from
almost 0° to 90°. In the former case the two members are
almost parallel, in the latter they diverge in the opposite direction
at almost 180°, the mirror plane bisecting the angle between them

9 Bateson, Materials for the Study of Variation, p. 479.

SYMMETRY IN TRANSPLANTED LIMBS . 13

(fig. 3). In the present paper the relation of the mirror plane
to the long axis of the limb has not been taken into account for
purposes of description, the relation only to the dorsopalmar and
the radioulnar axes being stated; i.e., the degree of divergence
of the two members is not taken into account. Thus, when the
mirror plane is parallel to the radioulnar axis, the limb is said to

A.DU eet) PR MP, (R)

Fig. 3 Diagram showing mode of reduplication. PR, primary limb; P.DU,
posterior reduplicating member; A.DU, anterior reduplicating member; WP;(R),
primary (radial) mirror plane; 1/P2,(U), secondary (ulnar) mirror plane; / to 4,
first to fourth digits, respectively. S, location of section shown in figure 4B.
Dotted lines show the outlines of limbs as they would have been had there been
no coalescence.

be mirrored in a palmar or a dorsal plane, according as the palms
or the backs of the hand face one another; when the plane is
parallel to the dorsopalmar axis, the mirroring is in a radial or
an ulnar plane, according as the radial or ulnar borders of the
limb face one another (fig. 4, 4). Intermediate planes are de-
scribed as radiodorsal, ulnopalmar, etc. (fig. 4, B). No attempt
has been made for the present to measure accurately the angles
of mirroring. It has been found, in agreement with Bateson, that

fee. . ROSS G. HARRISON

when there is a double reduplication, then the two mirror planes
intersect at the bifurcation in a line perpendicular to the proximo-
distal axes; 1.e., so that with reference to the radioulnar and dorso-
palmar axes the planes of reflection face one another (fig. 4).
Considerable deviation from this rule has, however, been noted
in certain cases, and the amphibians do not seem to follow
it with the same regularity as the arthropods, according to
Bateson.!°

MP2 (vu)

PAL PAL MP1 (R) PAL

Fig. 4 Diagram of reduplication, sectional view. In A the mirror planes are
radial (MP) and ulnar (MP2), and a certain amount of coalescence between the
primary and the anterior reduplicating members is shown, as in figure 3. In B
the mirror planes are radiodorsal (MP,) and ulno pulnar (WP2). D, dorsal;
PAL, palmar; R, radial; U, ulnar.

EXPERIMENTAL

A. Limb buds implanted in abnormal location—heterotopic
transplantations

In nearly all of the experiments in this group the limb bud was
implanted on the flank of the embryo at the ventral border of the
myotomes between the region of the fore and hind limbs. Ina
few cases it was placed on the side of the head between the eye
and the ear, but the grafts were absorbed in all of these except

10 Op, cit:, p. 552.

SYMMETRY IN TRANSPLANTED LIMBS 15

one, which yielded an imperfect appendage. They need not be
considered separately here, though a more extensive series of
experiments of the latter type would probably yield different
and more interesting results.

The limb buds transplanted to the flank of the embryo are
placed in an environment similar to that of the normal fore limb,
as far as relations to the body wall and muscle plates are con-
cerned, though they lack the specific blood supply and innerva-
tion of the limb region. Consequently, a very high percentage

TABLE 1
Heterotopic transplantations. Swmmary of experiments

| ee pecans

1 Excluding all cases where death occurred prematurely or where the grafted
limb was resorbed or remained rudimentary. Percentages in all tables have been
calculated on the basis of positive experiments.

2 There is evidence that in this case thére was an error in the orientation of
the bud and that it should therefore be classed in the group het. dd.

of cases yielded only abortive limbs, and those that did develop
rarely showed any functional activity. There is also greater
difficulty in securing good healing of wounds in the intermediate
region, so that a larger proportion of the cases died early. In
many of these cases there is obviously some interference with
the development of the pronephric duct, which becomes blocked.
The secretion which accumulates causes the formation of a cyst
of considerable size, which may interfere with the development
of the limb bud.
The results of the experiments are summarized in table 1.

11 Cf. Detwiler, 719 and ’20.

YW fff f /
/

t G ify iff Li Uf ify
MeO Lf Gi

Wh ff
veel if (MI

AF
YS =
VILE JU fii

PUAN”:
Mi /[\\\\\WNS
( \ S

eee

Se
Pf

Figs.5 to8 Heterotopic transplantation of fore limb; right limb to right side
(hom.dd.). TR, transplanted limb. X 10.

Fig. 5 Exp. Tr. E. 148, eight daysafter operation.

Fig. 6 Same, twenty days after operation.

Fig. 7 Same, twenty-eight days after operation, drawn from preserved
specimen.

Fig.8 Experiment Tr. E. 154, drawn from preserved specimen, killed twenty-
two days after operation. ;

SYMMETRY IN TRANSPLANTED LIMBS 17

1. Homopleural transplantations, normal or dorsodorsal orien-
tation. Nineteen cases were operated upon in this way (table 1).
In all of the cases where observations are recorded (thirteen in
number), the limbs, in the course of their development, gave evi-
dence of their original orientation, in that they pointed posteri-
orly or dorsoposteriorly when they first began to grow out (fig.
5). In the three cases that gave rise to single limbs they contin-

Fig. 9 Heterotopic transplantation (hom.dd.), showing twin limbs from one
implanted bud; PR, primary member; DU, reduplicating member. Exp. Tr. E.
182. X 10.

ued their growth in this direction, developing almost exactly like
the normal (figs. 6, 7, and 8). Likewise in the four cases that
gave rise to double appendages, the transplanted buds first began
to grow in a dorsoposterior direction, and only later did the re-
duplicating buds appear on the anterior border of the original
limb. The original bud developed in each case into a limb of
the same side, and the reduplicating buds became limbs of oppo-
site asymmetry (fig. 9). Histories of typical cases are given in
the appendix (p. 119).

18 ROSS G. HARRISON

(iy

My } i S SSX Vou >) SS gi i W) ee, z,
Hi We 7 nn an willl Co ae
Ni yy

SoS SSO AY

SSS

[a ES

SYMMETRY IN TRANSPLANTED LIMBS 19

2. Homopleural transplantations, inverted or dorsoventral orienta-
tion. Thirty-one experiments of this kind were made, with re-
sults as shown in table 1. Of the twelve cases yielding positive
results, one” gave rise to a pair of limbs and the others to single
limbs in which the asymmetry was reversed; i.e., the right limb
bud when placed upside down on the right side of the body gave
rise directly to a left limb. Even in the case which showed redu-
plication the primary limb of the pair became reversed. In all
of the cases where limbs resulted, the initial direction of point-
ing was anterior or dorsoanterior (figs. 10 and 11); 1.e., nearly
the opposite of normal. In four other cases this was also true.
In only four cases is the direction of pointing recorded as poster-
ior, and from these nothing definite was developed. All limbs
which developed continued their growth in the same general
direction, sometimes being directed more dorsally and sometimes
more sharply anteriorly (figs. 12 to 17). They also showed the
tendency to project more directly to the side than the normal
limbs. The final posture assumed by these appendages varies
considerably and does not seem to be dependent upon the degree
of development attained by the appendage. Two cases, each
having perfectly developed hands, exhibit the following extreme .
conditions: One!’ is practically a perfect mirror image of the nor-
mal right limb both as regards form and posture (fig. 18). The

Figs. 10 to 17 Heterotopic transplantation of fore limb; right limb bud to
right side inverted (hom.dv.), Exp. Tr. E. 219. N, normal limb, right side; 7'R,
transplanted limb. X 10.

Fig. 10 Dorsal view, five days after operation.

Fig. 11 Lateral view, same age.

Fig. 12 Dorsal view, eight days after operation.

Fig. 18 Lateral view, same age.

Fig. 14 Dorsal view, twelve days after operation.

Fig. 15 Lateral view of limbs only, same age.

Fig. 16 Dorsal view, sixteen days after operation.

Fig. 17 Lateral view, same age.

Fig. 18 Heterotopic transplantation; right limb to right side inverted (hom.
dv.), Exp. Tr. E. 139; drawn from specimen preserved twenty-eight days after
operation. X 10.

122 Tr. EH. 220.
13 Tr, BE. 139.

20 ROSS G. HARRISON

upper arm runs dorso-anteriorly and laterally. The elbow bend
is somewhat less than 90° and the fore arm and hand extend antero-
ventrally and laterally. The extensor surface of the elbow-joint
faces dorsally and slightly anteriorly and medially. The palm
of the hand faces medially, anteriorly, and slightly ventrally.

Fig. 19 Heterotopic transplantation (hom.dv.), Exp. Tr. E. 140; drawn from
specimen "preserved twenty-eight days after operation. J'R, transplanted limb.
x10]

The other case has its upper arm transverse and horizontal, and
its fore arm extends ventroposteriorly at an angle of less than
45° to the horizontal axis (fig. 19). The palm looks ventrally
and anteriorly. In order to bring this limb into the position of
the former, it would have to be rotated about the axis of the
humerus 45° or more and then adducted dorsoanteriorly at the
shoulder-joint through about 30°. The difference in position
assumed by the limbs in the various cases is thus due to differ-

14 Tr. E. 140.

SYMMETRY IN TRANSPLANTED LIMBS 21
ences in the amount of rotation, etc., undergone during the
later stages of development.

Histories of these cases are given in the appendix (p. 120).

3. Heteropleural transplantations, dorsodorsal orientation.
Twenty-eight experiments in this class have been made (table 1).
Five of these died prematurely, and in twelve the tissue was
either resorbed or failed to develop beyond the nodule stage.
In one case the bud developed into a stump about as long as the
upper arm, but without digits. Two cases gave double limbs
and eight developed into limbs which preserved their original
prospective asymmetry. ‘Two other cases may belong in this cate-
gory, one in which the original orientation of the bud is recorded
as uncertain!® and another!’ in which it is recorded as dorsoven-
tral probably by mistake.

In the development of the limb buds in this group twenty-one,
in addition to the two doubtful cases just mentioned, are recorded
at first as pointing in an anterodorsal direction, thus preserving
their original tendency in this respect. In the eight cases in
which the pointing was slight and in the five in which no definite
pointing was observed the limbs were abortive or resorbed.

In the eight cases where single limbs of the side of origin de-
veloped they retained their posture, developing as nearly exact
mirror images of the normal fore limb of the side to which they
were transplanted (figs. 20 to 23). The elbow-joint points dorso-
anteriorly, though varying somewhat, and the palm of the hand
faces ventrally, medially, and anteriorly (figs. 24 and 25). Indi-
vidual cases show variations similar to those observed in the
previous group. It is a striking fact that the general type of
development is the same here in the heteropleural non-inverted
buds as in the homopleural inverted bud, which shows that both
the posture and the asymmetry of the limb depend upon some
reaction between the bud and its new environment. (For case
histories see p. 121.)

The cases which showed reduplications, but two in number,
differ considerably from one another. In the first!’ growth was
slow and the resulting limb short with irregular reduplications

str ee Ls. 26.7. ee Tes ees: 18 Tr. Ha 1L9.

ROSS G. HARRISON

N
N

Ss Shee

Z\ 4:
ti)

4
SYMMETRY IN TRANSPLANTED LIMBS 23
in the hand, so that right- or left-sidedness could not be deter-

mined. In the other!® the limb developed promptly and formed
a duplicate member (fig. 26), which was first seen at ten days and

Fig. 26 Heterotopic transplantation (het.dd.), Exp. Tr. E. 127. Drawn from
specimen preserved twenty days after operation. PR, primary; DU, redupli-
cating member; / to 4, digits.

Figs. 20 to 23 Heterotopic transplantation of fore limb; right limb bud to
left side (het.dd.), Exp. Tr. E. 227. N, normal left limb; 7'R, transplanted limb.
x 10.

Figs. 20 and 21 Dorsal and lateral views, respectively, eight days after oper-
ation.

Fig. 22 Ventral view, thirteen days after operation.

Fig. 23 Lateral view, seventeen days after operation.

Figs. 24 and 25 Heterotopic transplantation (het.dd.), Exp. Tr. E. 107. Lat-
eral and ventral views, respectively, of preserved specimen, twenty-six days
after operation. X 10.

edad Be ea) De Bf

24 ROSS G. HARRISON

Wy HL |
’ 4 Uy WY
UY YWYey/ Wp YU

FE IDES
Oe Lah

SYMMETRY IN TRANSPLANTED LIMBS 25

which developed ultimately into a left limb, the mirror image of
the primary member, having been reversed from the original
prospective asymmetry of the transplanted bud.

4. Heteropleural transplantations, dorsoventral orientation. Sixty
operations were done in this series. For some unknown reason a
very large proportion (twenty cases) died prematurely and six-
teen of the survivors yielded only abortive limb buds, leaving
only twenty-four available for consideration. Eight of these are
recorded as imperfect, six produced reduplications to some degree,
and nine, single limbs of reversed asymmetry. Several of the
latter were somewhat defective and others showed slight redu-
plications. Several cases which are exceptional will be consid-
ered below.

In the cases where single limbs arose, development took place
in a manner fundamentally like that of the limb buds normally
oriented (hom. dd.). As the buds grew out, they began to point
in a posterior direction, and so continuing, developed into limbs
in normal posture (fig. 27). There was, however, less regularity
than in the homopleural dorsodorsal group. ‘The direction of
pointing was not always dorsoposterior, as in the normal limb,
but was in many cases inclined more ventrally. There are rec-
ords of pointing in all of the positive experiments and in many
of the negative. In only three cases is the direction recorded

Fig. 27 Heterotopic transplantation of fore limb; right limb bud to left side
inverted (het.dv.), Exp. Tr. E. 193. Preserved specimen killed twenty-four days
after operation. 10.

Figs. 28 to 32 Heterotopic transplantation of fore limb; right limb bud
to left side inverted (het.dv.), Exp. Tr. E. 217. N, normal left limb; TR, trans-
planted limb; PR, primary member; DU, reduplicating member; / to 4, numbers
of digits.

Fig. 28 Dorsal view, five days after operation.

Fig. 29 Lateral view, five days after operation.

Fig. 30 Dorsal view, fifteen days after operation.

Fig. 31 Lateral view of limbs, fifteen days after operation.

Fig. 32. Limb showing beginning of reduplicating digits (DU) on ventro-
anterior border (from a free-hand sketch nineteen days after operation).

Figs. 33 and 34 Heterotopic transplantation (hef.dv.); right limb bud to left
side. Exp. Tr. E. 163. Anomalous result. Primary member (PR) defective;
reduplicating member (DU) reversed. Lateral and ventral aspects, respectively,
drawn from specimen preserved thirty-nine days after operation. X 10.

26 ROSS G. HARRISON

as dorsoanterior; one of these died early and the other two gave
rise to imperfect limbs with indeterminate asymmetry.?°

The individual cases in which limbs of opposite asymmetry
developed were rather more irregular than in the preceding
groups, though the best cases gave perfect reversed appendages. .
In addition to the ones included in the tabulation, there is one
other case that probably belongs in this category. It is one in
which the orientation of the bud at the time of transplantation is
recorded as uncertain.2! The limb that developed is a perfect
one of reversed asymmetry in nearly the same posture as the nor-
mal limb of the side to which it was transplanted. It showed an
unusual amount of motility. In one case, included in the records
of this group,” the transplanted bud developed into a normal
limb of the side from which it was taken. It is believed, however,
that a mistake was made in recording the operation in this case,
and that probably in reality the orientation of the limb was not
inverted. The direction of pointing, as observed on the third
and fifth days after the operation when the limb bud is recorded
as pointing anteriorly, is evidence, though not absolutely conclu-
sive, that an error has been made. If this interpretation is cor-
rect, the case would not be exceptional, but would accord with
the eight cases described in the previous section.

In the eight cases in which reduplications occurrred, the early
stages of development were like the normal (figs. 28 and 29),
the reduplicating buds not being noted until at least twelve
days after the operation. Three individuals showed distinctly
that the primary limb was of reversed asymmetry. In one case
it was so imperfect that it could not be determined to which side
it belonged, but the reduplicating limb was sufficiently devel-
oped to show that it was of the same side as the bud was origin-
ally, indicating that the original member was in all probability
reversed. Another case? gave a limb with nearly symmetrical
reduplication in the hand without anything to indicate which
member was primary (figs. 31 and 32). Two long radial digits
are present in the middle and two short ulnar digits on each side.
Still another case gave a very peculiar result. The primary

20 Tr, EH. 108 and 203. 21Tr, EK. 109. 22. Beles 23°, Boe
24Tr, E. 163.

SYMMETRY IN TRANSPLANTED LIMBS 27

limb bud developed into a long almost filiform structure, with
out digits, that grew posteriorly on the ventral side of the body
not far from the midline. Twenty days after the operation a sec-
ond bud was noticed dorsal to the original, and this developed
into a somewhat peculiarly placed limb. The upper arm runs
transversely and the palm of the hand faces dorsomedially (figs.
33 and 34). This limb is clearly a left; 1.e., its original prospective
asymmetry has been reversed. It therefore constitutes an ex-
ception to the rules, not only because of the position of the
hand, but also because of its particular asymmetry ; for the original
(filiform) member should have been reversed (a left), and the
second one reversed back again to the original asymmetry. How-
ever, the fact that the latter developed at such a considerable dis-
tance from the original member, might be regarded as indicating
that it was beyond its sphere of influence, perhaps having been
split apart from it at an early stage, and that it remained there-
fore as of the same side. Several cases of regeneration after
extirpation of half buds and of transplantation of half buds gave
analogous results (fig. 132).

5. The shoulder-girdle in heterotopic transplantations. The
limb-girdle in the hetrotopic transplantations is developed in
more or less reduced condition, as was first shown by Braus ('09)
in the anurans. Detwiler (718) has studied this question in
Amblystoma, and has found that the degree of development of
the girdle is dependent upon the size of the graft and the region
from which it is taken, the scapula and suprascapula being local-
ized in the tissue dorsal to the normal limb bud and the coracoid
in that ventral to it.22 The form of the reduced girdle derived

25 Cf. Harrison, 718, p. 441 (Exp. Rem. E. 17 and H. R. E. 10), and page 135
of the present paper (Exp. H. R. E. 20).

26 It is a curious fact that in the embryo the limb-girdle has undoubtedly
the character of a mosaic, without totipotence of its parts, while in the adult
Triton, according to Tornier (’06), Fritsch (711), and Kurz (712), a small portion
of the shoulder-girdle can regenerate the whole, including the fore limb. Ac-
cording to the two last-named investigators, even if the whole girdle is removed,
it will be regenerated together with the free appendage. Kurz has found that
this holds for both shoulder and pelvic girdles but that removal of the sacral
portion of the vertebral column prevents regeneration. In the anurans, accord-

ing to Braus (’06), there is considerable variation in the regenerative powers of
the limbs in early stages.

28 ROSS G. HARRISON

from the usual round dise (limb bud) is roughly triangular, as
shown in the figure of Detwiler’s model (his figure 28), with a
ventral process projecting anteriorly, to be identified as a rudi-
mentary coracoid, and a dorsal process, which includes the rudi-
ment of the scapula. In the normally oriented grafts (homo-
pleural dorsodorsal) these processes point anteriorly, with a
single process projecting posteriorly slightly behind the glenoid
cavity. This shows clearly in two cases.27 The question now
arises whether the girdle follows the rules governing the asym-
metry of the free limbs. The results, in the main indicate that
such is the case, though the girdle developed is often so small
and rudimentary, that it is not possible to determine to which
' side it belongs. In the inverted homopleural grafts, which give
rise to reversed limbs, the girdle also seems to be reversed. ‘This
is true in four cases out of the five examined in serial sections.?8

Among the heteropleural dorsodorsal transplantations, five
cases have been examined in sections. In two of them?’ with
well-developed glenoid cavity, the girdle cartilage is mostly ven-
tral and posterior to the joint. This probably represents a cora-
coid with asymmetry corresponding to that of the free limb. One
case,*° with the cartilage projecting both anteriorly and pos-
teriorly from the cavity, gives no evidence as to the side to which
it belongs. One is too rudimentary,*' and one seems to have had
its asymmetry reversed,” though the limb is not reversed. In
the two dorsoventral heteropleural transplants which have been
studied in sections, the side to which the girdle belongs cannot be
determined. Other cases from among the earlier experiments,
where in most instances the size of the transplanted bud was
small, are inconclusive. On the whole, the cases where the asym-
metry can be determined with any degree of certainty seem to
follow the rules. Only a single case thus far examined is clearly
exceptional.

27 Tr. EK. 148 and 154. $0'Tr Hs i20:
28 Tr, EK. 185, 186, 189, and 140. 3 Tr. Hele
29 Tr, HE. 124 and 169. 22D. HalOie

SYMMETRY IN TRANSPLANTED LIMBS 29

6. Summary of the results of heterotopic transplantations. A
survey of all the experiments in this group brings out the following
facts:

Implanted in dorsodorsal orientation, a limb bud gives rise to
an appendage of its original prospective asymmetry, whether
placed on the same or opposite side of the body. Such appen-
dages have a normal posture when placed on the same side of
the body from which they were taken, but when placed on the
opposite side they mirror approximately the limb of that side,
though they often become rotated to quite different postures.
Implanted in inverted (dorsoventral) position, a limb bud gives
rise to an appendage of reversed asymmetry whether placed on
the same or opposite side of the body. When placed on the same
side, such appendages mirror the normal limb of that side, but
when grafted on the opposite side, they assume a posture approxi-
mately identical with that of the limb of that side.

Limbs implanted in any of the four positions here studied may
produce reduplications. As far as it has been possible to deter-
mine, the primary limb of the pair is then of the same side as a
single limb would be according to the foregoing rules. The redup-
licating limb has been found to be, with a single exception, the
mirror image of the first.

Limbs that are grafted in abnormal location have at best very
incomplete function and are often apparently entirely immobile.
They usually do not become so large as those that are implanted
in normal location, and they show defects and evidences of atro-
phy much more frequently.

B. Limb buds implanted in natural location—orthotopic
transplantation

In these experiments the limb bud of the host was first removed
and then either put back in place, or else a bud from another
embryo was grafted into the wound. In all of the earlier cases
the wound bed was not cleaned after removal of the bud, so
that some cells from the host were left to mingle with the tissues
of the transplanted limb rudiment. The later experiments, with

30 ROSS G. HARRISON

but few exceptions, were carried out under precautions necessary
and sufficient to preclude contamination of this kind: the extir-
pated area was three and a half somites in diameter, and the bed
of the wound was carefully scraped after removal of the bud.*8
The results were somewhat different (proportionately) in the

. TABLE 2
Orthotopic transplantations. Summary of experiments

NUMBER OF SINGLE LIMBS SINGLE LIMBS

EXPERIMENTS | NOT REVERSED REVERSED HADDALUELEIKELAE 2

OPERATION —————
Total Posi- | Num-| Per Num- Per Num- | Per
‘ tive ber cent ber | cent ber cent

A. Wound bed cleaned and wound ' not less than 34 somites

Moni. Sar pees ache 9 9 9 {100.0 C | 00.0 0 | 00.0
ROUT. iy; he satus, oe eco. 2 61 38 102) 26.3 iL 2:6:\| 2A geese
et. da made ot eas 49 31 ] 3.2 52 16.7 | 25. sOme
Het veer Ss or caiaoe vie, ase 26 16 O- | 00.0 | La ar9s88 1 6.3
gi Notre) Re 1 AS Sie Bee ora 145 94 20° | 21.3 fe 21. 022.3 | 5a) eae
Average of percentages. 31.6 28.8 39.6

B. Wound bed not cleaned

EL On Gi eee tad. ce 0 0 0 0 0
Hiomiidy2 2a ks.) tee B71 DH 194 | 95.0] Oo | 00.0 1 5.0
Fleteadale ci cee... eet eae 13 2) |) 154s osha 23a 8 | 61.5
Hetenas = eee. ee 21 15 O || OO#0) seri753.3 7 Aga
Faye AL Se ea 75 48+}. 2t | 43.8)" das “\c29-9Nl 6 Saaee

‘ Including three cases in which the primary bud righted itself by rotation and
the duplicate is disharmonic.

* Limbs which became normal by rotation, including one case (I. E. 101) of
hyperdactyly.

’ Normal by resorption of original member of pair.

4 One case included in which the posture of the limb was abnormal.

two classes of experiments and have been summarized separately
in table 2 (A and B). The differences will be taken up in connec-
tion with the consideration of each of the subgroups.

?. Homopleural transplantations, dorsodorsal orientation. This
is in reality merely a control experiment and is a test of the effect

38 Harrison, 715 and ’18, p. 422.

SYMMETRY IN TRANSPLANTED LIMBS 31

of the operation as such on the development of the limb. A fore
limb bud is carefully excised and either replaced in the same
wound or else engrafted in normal position in another embryo
from which the limb bud had been previously removed.

Only nine individuals were operated upon, in all of which the
wounds were carefully cleaned. Normal limbs developed in all
cases, though they were slightly retarded in the earlier stages of
development in comparison with the unoperated limb of the
_ opposite side. In six of the cases the pronephros was removed
and in the other three it was left in. No difference was noted
between the two sets. It may be safely concluded that the
effect of the operation itself upon normal development is prac-
tically negligible.

8. Homopleural transplantations, dorsoventral orientation. In
some of the cases of this series, as in the last, the limb bud was
simply lifted and replaced after rotation through 180°. In the
others the wound bed was first prepared in one embryo and the
bud taken from another. The latter method is preferable and
it was employed in all the later experiments.

The total number of experiments is one hundred and four, of
which sixty-one were with cleaned wounds of proper size. The
latter will be considered first, since the conditions of experimenta-
tion are more definitely known and there can be no doubt that
the limbs were derived exclusively from the transplanted tissue.

Leaving out of consideration the twenty-three cases which died
prematurely or gave rise merely to abortive or rudimentary
limbs, there are thirty-eight cases which yielded positive results,
as recorded in table 2A, The single limbs are in the minority
and are of two kinds, reversed and non-reversed. The most re-
markable case** (history on p. 124), which really gives the clue
to the interpretation of the experiments of this group, is the one
in which a limb of reversed asymmetry developed, a right limb
on the left side, perfectly normal in form, function, and posture,
as far as the last is possible on the wrong side of the body (figs.
39 to 41). The shoulder-girdle of this limb is also reversed and

i) IB De

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, No. 1

By ROSS G. HARRISON

eis Hip)
(

ae Yi YY)
35

e ——- o

di Yi

i lt

Ah
Wes
;

i
1} ily hey
yy y\ 4

iw ye

SYMMETRY IN TRANSPLANTED LIMBS 33

is quite separate from the rudimentary girdle developed from the
tissues of the host (p. 59).

The ten cases in which normal non-reversed limbs developed
are clearly contrary to the rule (p. 4). The records of these
cases show that the end result is reached by a process of rotation
at the shoulder-joint during development (figs. 56 to 58). They
will be considered below (p. 40). The reduplicated limbs, of
which there were twenty-seven, fall, like the single, into two
groups. In the first the original bud developed into a limb of
reversed asymmetry, while in the second it is not reversed.

Observations upon the earlier stages of the operated limbs show
that in those cases which give reduplications, as well as in the
case of the simple limb with reversed asymmetry (figs. 35, 42,
43, 49, 51, and 52), the original direction of pointing is either
dorsal, anterior, or dorsoanterior, and more sharply lateral than
normal. Likewise in the case of those that develop into single
non-reversed limbs, the first pointing is more sharply lateral than
normal, and also more dorsal, though only two are recorded as
pointing slightly anteriorly from the dorsal direction. This
shows that the original tendencies of growth, immanent in the
bud at the time of transplantation, are by no means inactive
when it is in its new position. One or the other of two conse-
quences of this growth tendency now ensues, indicating a sort of
antagonistic reaction between the organization of the transplanted
rudiment and that of the surrounding parts. The limb either
continues to grow in an anterior or anterodorsal direction, in

Figs. 35 to 41 Orthotopie transplantation; left limb to left side inverted
(hom.dv.), resulting in a normal right limb onthe operated side. Exp. I. E. 64.
TR, transplanted limb. X 10.

Fig. 35 Lateral view, five days after operation.

Fig. 36 Ventral view, ten days after operation.

Fig. 37 Dorsal view, sixteen days after operation; transplanted limb covered
by gills.

Fig. 38 Lateral view of limb, sixteen days after operation.

Fig. 39. Dorsal view, twenty-three days after operation.

Fig. 40 Ventral view, twenty-three days after operation.

Fig. 41 Lateral view, twenty-three days after operation.

35 T. E. 49 and 94.

<<
SS

SSSaxq_

ROSS G. HARRISON

Figs. 42 to48 Orthotopic transplantation; left limb bud to left side inverted
(hom.dv.), resulting in duplicate limbs. Exp. I. E.60. N, normal right limb bud;
TR, transplanted bud; PR, primary member; DU, reduplicating member. X 10.

Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.

Dorsal view, five days after Uperation.

Lateral view, five days after operation.

Dorsal view, seven days after operation.

Dorsal view, twelve days after operation.

Lateral view, twelve days after operation.

Lateral view of specimen preserved eighteen days after operation.
Ventral view of limb, preserved specimen.

SYMMETRY IN TRANSPLANTED LIMBS 35

which case its asymmetry is reversed (figs. 36 to 39), just as in
the corresponding class of heterotopic transplantations, or it
gradually rotates towards its normal position while retaining its
original prospective asymmetry (figs. 56 to 59). In the former
alternative duplicate limbs are nearly always formed. Only in
the one case, referred to above, did a perfect single limb arise.
In the other alternative single limbs usually arise, though some
of the cases of reduplication certainly belong to this group.

In the duplicities belonging to the first group the original limb
bud continues to grow in an anterior direction and ultimately
becomes a reversed limb. After a time a reduplicating bud
appears on the posterior border of the original bud (fig. 44) and
in the clearer cases grows into a homopleural limb in approxi-
mately normal posture (figs. 45 to 48). The original bud becomes
a reversed limb which, together with the reduplicating member,
may form an almost symmetrical complex.

Twenty-four of the thirty-one** cases of reduplicated limbs are
probably of this type. Fifteen are certainly so,*7 and in three
others*® that are very similar all that is lacking to place them
unequivocally in this group is a definite observation as to which
bud was the primary one; six more cases*? may also be inter-
preted in the same manner, though they are not sufficiently clear
to insure that this is the only possible interpretation.

The degree of reduplication varies here, as in the other groups
of experiments, from the condition where almost the whole arm
is involved to that in which the hand is only partly double. In
three cases*® the anterior bud was much reduced (p. 49), the
posterior bud becoming a somewhat irregular homopleural limb.
In eleven cases there is only one reduplicating appendage, which
is always posterior to the primary (figs. 43 to 48), while in the
remaining twelve" there are evidences of further doubling, usu-

36 Four cases are considered here which are not included in the tabulation on
account of the fact that the wound was only 3 somites in diameter (I. E. 39, 41,
44 and 45).
37 J. EK. 48, 60, 62, 63, 66, 72, 74, 75, 81, 85, 87, 89, 91, 92, and 96.
387. BK. 44, 45, and 52.
39 T. EK. 39, 68, 70, 93, 100, 102.
40 T, EB. 92, 93, 100.
417, KE. 39, 45, 62, 63, 66, 72, 75, 81, 85, 87, 91, and 93.

36 ROSS G. HARRISON

MULL pp fp
RYY é

WWM UY LY)
YS Ug:
' 7

ayy

X
Ss

SSX

SYMMETRY IN TRANSPLANTED LIMBS 37

ally on the anterior side of the original (figs. 49 to 55 and 61).
When the latter condition arises and the anterior reduplicating
member is sufficiently developed, it is seen that it, too, is mirrored
from the original member and is homopleural (fig. 55). In one
case there are three complete hands, one of which has two of its
digits doubled.” The plane from which the posterior redupli-
eating member is mirrored in the final form of the limb varies
from radial to dorsal (figs. 3 and 4) and is usually intermediate
between these two extremes (p. 13). Nineteen cases follow
this rule, three are indeterminate and there is only one positive
exceptional case, in which the mirror plane is ulnodorsal.**? When
there is also an anterior reduplicating member, it is generally
mirrored from a plane 180° around the limb axis from the first;
i.e., ulnar, ulnopalmar, or palmar.

The reduplications belonging to the second group are more
restricted and less certain of diagnosis. The limb bud retains its
original prospective asymmetry, reaching an approximately nor-
mal position by rotation, and reduplication is much less extensive,
involving in most cases the digits only (fig. 62). Three cases“
almost certainly belong to this group, and there may be two
others.” |

Of the two remaining cases of reduplication, one died too
young; in the second** the supernumerary limb was of the same
side as the primary and was quite distinct from it. This is a
very unusual condition, but the transplanted bud in this case was

Figs. 49 to 55 Orthotopic transplantation; left limb bud to left side inverted
(hom.dv.), resulting in limb with two reduplicating members. Exp. I. E. 63.
N, normal right limb bud; TR, transplanted left bud; PR, primary limb; A.DU,
anterior, and P. DU, posterior reduplicating members. X 10.

Fig. 49 Dorsal view, four days after operation.

Fig. 50 Lateral view, four days after operation.

Fig. 51 Dorsal view, seven days after operation.

Fig. 52 Lateral view, seven days after operation.

Fig. 53 Dorsal view, ten days after operation.

Fig. 54 Lateral view, ten days after operation.

Fig. 55 Dorsal view of specimen preserved seventeen days after operation.
Gills (BR) removed to show limb. 1 to 3, numbers of digits.

27.E.87. “1. E.72. “1. EB. 86,88,90. “I.E. 41,59. “I.E. 38.

ROSS G. HARRISON

YH,
Wij
y

WV
\ aah

\\
i)

LEY DEEL IVI 7/.,
\\ SALSA ES = SEL 7
Why \ a) ‘ : YD) NN NAN
Z| )

SYMMETRY IN TRANSPLANTED LIMBS 39

larger than usual (four somites in diameter), and it is possible
that the reduplicating bud, growing from near its anterior border,
was uninfluenced by the primary limb and hence was not mirrored.

There remain for consideration those cases in which a single
non-reversed limb developed. As in the other cases, the limb
bud in these showed at first the consequences of abnormal orien-
tation. When first observed it pointed more sharply laterally
and more dorsally (less posteriorly) than normal. Two even
pointed dorsally and slightly anteriorly. In the course of devel-
opment the limb gradually changed its posture and ultimately
came to a perfectly normal posture by a process of rotation at the
shoulder-joint (figs. 56 to 59). Ten such cases were obtained,*7
though in several of them?’ it is possible that reversal may have
been brought about by early reduplication and suppression of the
original bud, as deseribed in the next section (p. 49). In one of
these*? a supernumerary radial digit was present, but this is to
be regarded as a case of hyperdactyly rather than one of mirrored
reduplication. In one case*® the limb which originally developed
showed irregularities in the digits. The arm was then ampu-
tated above the elbow, and the appendage which regenerated was
in every respect normal. ‘This case is of considerable interest in
showing that the abnormal condition which produces reduplica-
tion is not necessarily stamped upon the whole structure, but
may be due to some local mechanical disturbance.

In reviewing this group of experiments, it is clear that the first
two results, 1.e., single reversed limbs and most of the reduplica-
tions, come under the same scheme. ‘There is a primary reversal
of asymmetry, without reduplication in the first case and accom-

Figs. 56 to 59 Orthotopic transplantation; left limb bud inverted (hom.dv.).
Limb reaches normal posture by rotation. Exp. I. E. 49. N, normal (right)
limb; 7, transplanted (left) limb.

Fig. 56 Eleven days after operation.

Fig. 57 Twenty-one days after operation.

Fig. 58 Thirty-eight days after operation.

Fig. 59 Preserved specimen, killed at thirty-nine days.

47 J. E. 49, 55, 69, 71, 73, 77, 84, 94, 99, and 101. gig Gul Be bs
48 For instance, I. E. 77. SOOT, Swill:

40 ROSS G. HARRISON

panied by reduplication in the second. In the latter the second-
ary bud, being the mirror image of the other, is again reversed
back to the original prospective asymmetry of the transplanted
bud. This then occupies an approximately normal position and
may function to a considerable extent as a normal limb, though
impeded by the connection with its mate. These two results are.
directly comparable to those of the heterotopic transplantations
of the corresponding class (fig. 2).

The third result, in which normal homopleural limbs develop,
reaching their normal position gradually by rotation, is funda-
mentally different. Here no reversal occurs; the limb bud begins
its development as a self-differentiating system, but later, under
the stress of the changed relation to its environment, it comes
again into normal posture.

What determines whether the limb bud shall reverse its asym-
metry or rotate back to its normal posture? The earlier experi-
ments of this series afforded no satisfactory answer to this ques-
tion. It was certainly not due to the size of the wound, the
mode of preparation of the wound, the presence or absence of
the pronephros, or the age of the embryo. What seemed most
likely was that there were minor accidental differences in the
amount of rotation to which the limb bud was subjected at the
time of operation. It was conceivable, for instance, that if the
dise were rotated anteriorly around the dorsal semicircumference
of the wound a little less than 180° (fig. 60), the reversing effect
of its organic environment might be lessened and the rotation
back to normal position facilitated; in this case a normal non-
reversed limb would result. If, on the other hand, the grafted
bud were rotated 180° or shghtly more, the reversing effect might
be at a maximum and rotation most impeded, in which case a
heteropleural limb or twin limbs would arise.

Experiments made in the spring of 1917 had for their main
purpose the testing of this hypothesis. Operations were done
in pairs; in one case the limb disc was rotated about the dorsal
circumference in a posteroanterior direction slightly more than
180° and in the other slightly less, extremes being probably not
more than 190° and 170°, respectively (see histories on pp. 125-6).

SYMMETRY IN TRANSPLANTED LIMBS 41

The results are not altogether conclusive, though they point
to the correctness of the hypothesis. Twenty-three operations
were done, of which seventeen yielded positive results. Ten
cases harmonize with the hypothesis; four are doubtful, and three
are contrary to it (figs. 61 and 62). When the difficulty of ex-
actly estimating the degree of rotation is considered, many ap-
parent exceptions must be expected, and a far greater number of
experiments would be necessary to eliminate statistically the
effect of this uncertainty. As a matter of fact, the records of the
cases classed as surely exceptional give evidence that the amount

A 180° °° P

Fig. 60 Diagram showing difference in amount of rotation in two sets of
experiments. The circle represents the left limb bud and the arrow the direction
of rotation. A,D,P,V, the direction of the cardinal points of the embryonic
body, anterior, dorsal, posterior, and ventral, respectively.

of rotation at the time of operation was probably not correctly
estimated, for the first direction of pointing (p. 11) in all of
these cases is not according to expectation.

The thirty-seven cases in which the wound bed was not entirely
cleaned of mesoderm may now be considered. These show a
marked contrast to those with cleaned wounds, inasmuch as there
are very few reduplications and a very large proportion of normal
non-reversed limbs. Thus out of twenty cases which yielded
positive results eighteen or 90 per cent are normal, as compared
with 23.8 per cent (ten cases) in the clean-wound class. Only one
case (5 per cent) had a reduplicated limb, as compared with
thirty-one duplicities (73.8 per cent) in the clean-wound class.

ROSS G. HARRISON

: Ey =) )\ ea

Gg Lt —— WWE: LEE

Y, Sree Ss = oe Br Lish—_> =
Pent te SSS
! == — = eee ZB
\\ Ls es Ml) pp Ut |p py phils ps) tft PUL! 244 try fy

— eee Yi WEE ; RES Ssascwe

SSsSSS>—]

=
— Sas

—S== —————— — aD
{! as = :
mn ae Reeeee aw SRS WARS
WH MAY UNO
: = SBE ee

Fig. 61 Orthotopic transplantation; left limb bud to left side (hom.dv.),
rotated slightly more than 180° from its normal position, 7.e. P to A’ in fig. 60.

Exp. I. E. 85. Primary member (PR) is a right; posterior reduplicating mem-
ber (P.DU) is a nearly normal left; anterior reduplicating member (A.DU)
partly coalesced with primary. ;

Fig. 62 Orthotopic transplantation; left limb bud to left side (hom.dv.),
rotated slightly less than 180° (i.e. P to A in fig. 60). Exp. I. E. 86. The trans-
planted limb is primarily a left, having reached its normal posture by rotation;

second (2’) and third (3’) digits, reduplicated. X 10.

SYMMETRY IN TRANSPLANTED LIMBS 43

Since these differences can scarcely be accounted for on the ground
of different degree of rotation of the limb buds at operation (p.
40), it would seem that the few mesoderm cells remaining in the
wound bed must have exerted some influence upon the develop-
ing limb. This does not mean that the limbs which do develop
in such cases arise solely by a process of regeneration from the
host. In fact, the rate of development, which is only slightly
retarded below the normal, precludes such an interpretation.
What probably does take place is an intermingling of cells from
the host and the graft, with the result that the former, acting in
the same sense as the environment with which they are in har-
monic relation, counteract the tendency of the inverted elements
to reverse their asymmetry. This was, however not shown to
the same degree in the corresponding experiments with super-
posed limbs (p. 65).

9. Heteropleural transplantations, dorsodorsal orientation. For-
ty-nine cases were operated upon in this way and thirty-one
lived long enough to yield definite results (table 2). By far the
largest number of these (twenty-five) developed reduplications of
one kind or other. Five cases gave rise to limbs with reversed
asymmetry, 1.e., to limbs which developed to fit their new sur-
roundings, though one of these was considerably underdeveloped.
One yielded a somewhat imperfect non-reversed limb and four
were rudimentary. These results seem altogether divergent from
the corresponding heterotopic transplantations. An examina-
tion of them shows, however, that fundamentally they accord
with the latter, complete agreement being modified, by a second
factor, which may suppress the original bud in favor of the
reduplicating member. The normal environment of the trans-
planted bud and the concomitant normal functioning seem to
facilitate this transformation. Moreover, there is no hard and
fast line between the different results just enumerated, and the
individual cases may be taken as forming a series, beginning with
the single non-reversed and ending with the single reversed limb.
The reduplications are intermediate. They will be considered in
this order.

44 ROSS G. HARRISON

SSS SS
SSS

SSS

65 66

A : |
=<, Sa SSS S \
“ LLL = Wy = .
- LE SS i= Ny
Zc coe ——— =e: \ »\
= == ; SS = Sot = WY
= ~ SS een a Zz
— <S S) Agee

SYMMETRY IN TRANSPLANTED LIMBS 45

The history of the case in which a limb of original prospective
asymmetry developed* is given on page 126. In this individual
the limb bud, as it began to grow, pointed anteriorly (fig. 63),
and continued to grow in that direction. Though it remained
small and imperfect (fig. 64), it is clearly a right lLmb on the
left side (not reversed).

The cases which formed reduplications began their develop-
ment in the same manner. The first direction of poimting is
recorded as anterior in nine cases, anterodorsal in eight, and
anterolateral in five. Three are described as pointing dorsally
and one laterally. Thus these limbs all show in greater or less
degree the initial effect of their original growth tendency.
Growth of the bud continues then for some days in a general an-
terior direction, but sooner or later a reduplicating bud appears,
usually at the posterior border of the original bud, and this
grows in most cases into an appendage equal to or exceeding the
original in size. If the reduplicating bud does not appear
until late, then the original one may attain considerable size and
remain, for some time at least, the principal member (figs. 65
and 66). If it appears earlier, but not until the original. bud has
a good start, then the two members may remain of almost equal
size (figs. 67 to 71). In other cases, where the reduplicating bud
begins to grow early, it soon gains the upper hand, and the orig-
inal may be reduced to an atrophic or rudimentary limb (figs.
72 to 74). This condition leads over to the single reversed ap-
pendage in which the original bud is reduced to a spur or nodule

Figs. 63 and 64 Orthotopic transplantation; right limb bud to left side
(het.dd.). Exp. R. E. 87. Resulting limb, though defective, is reversed. JN,
normal right limb; 7'R, transplanted limb. X 10.

Fig. 63 Dorsal view, seven days after operation.

Fig. 64 Ventral view of specimen preserved sixteen days after operation.

Figs. 65 and 66 Orthotopic transplantation; right limb to left side (het.dd.).
Exp. R. E. 96. Resulting limb reduplicated.

Fig. 65 Ventral view, ten days after operation; primary limb (PR) points
into gills; reduplicating bud (DU), just appearing. X 10.

Fig. 66 Ventral view, nineteen days after operation; primary member (PR)
shows evidence of reduplication of hand; reduplicating member (DU) is in
approximately normal position. X 10.

eT, Wh 87.

ROSS G. HARRISON

—_

: SSS SD
SSSR Rn

SYMMETRY IN TRANSPLANTED LIMBS 47

(p. 49). The reduplicating limb is, of course, mirrored from the
original and hence corresponds to the side of the body on which
it is grafted.

Being placed in the position of the normal limb, the reduplica-
tion is favorably situated with regard ‘to blood and nerve supply,
and it is probably on this account that so many of them develop
into functional appendages. In this respect the experiments of
this group differ considerably from the preceding (homopleural
inverted), where there is a greater tendency for the original mem-
ber to retain its predominant condition. Otherwise the course
of development in the two groups is strikingly alike.

For the study of the details of reduplication, nineteen cases
are available, including one case with small wound not consid-
ered in the table. Seven others were preserved at relatively
early stages in order to investigate the internal processes involved.
Histories of several typical cases are given in the appendix.

Seventeen of the cases conform to the main type and do. not
differ materially from those considered in the last section. As in
the homopleural inverted limbs (p. 35), the degree and character
of the reduplication vary much from case to case. In some the
digits alone are doubled, and at the other extreme we find two
almost entirely separate limbs. In thirteen individuals a second
reduplicating limb formed on the anterior side of the original.
These usually did not develop so completely as the limbs arising
from the posterior buds, and the reduplication often involved
only the distal part of the manus, with the digits more or less sym-
metrically placed. The anterior reduplications are mirrored
from the ulnar or ulnopalmar surface and occasionally from the

Figs. 67 to 70 Orthotopic transplantation; right limb to left side (het.dd.).
Exp. R. E. 70. Resulting limb reduplicated. N, normal right limb bud; TR,
transplanted limb bud; PR, primary member; DU, reduplicating member. X 10.

Fig.67 Dorsal view, five days after operation. The transplanted bud already
gives evidence of reduplication.

Fig. 68 Dorsal view, ten days after operation. Reduplicating bud is large
and in normal position.

Fig. 69 Dorsal view, sixteen days after operation.

Fig. 70 Dorsal view of specimen preserved thirty days after operation.

Fig. 71 Similar case (Exp. R. E. 71); ventral view, twenty-two days after
operation. X 10.

HARRISON

ROSS

S
Ss

SYMMETRY IN TRANSPLANTED LIMBS 49

palm. One case” had three almost complete separate appendages
(figs. 75 and 76).

Two cases of the nineteen gave rise to an anterior reduplicat-
ing bud only, which in both individuals was mirrored from the
ulnopalmar surface. Owing to the position of the reduplicating
bud in front of the original heteropleural limb, it could not be
brought into normal posture (fig. 77).

There remain to be considered the five cases in which the
asymmetry of the transplanted bud was reversed. These are
of the utmost interest in showing how a secondary factor (redu-
plication) may so modify the result that the rules of symmetry
seem not to hold. They show more than any others the neces-
sity of having complete histories in each case, for the manner in
which the end result is reached is of cardinal importance for
the correct interpretation of the process. As stated above, these
cases gradate into those in which duplicate limbs arise, so that
the classification is somewhat arbitrary, the simgle-limb condition
being a masked reduplication. Like the others, they begin their
development with growth of the bud in an anterior direction (figs.
83 and 86). Then a posterior reduplicating bud makes its ap-
pearance, and the original bud is rapidly reduced (figs. 84 and
85 and 87 to 89) in relative importance, becoming a spur or nodule
attached to the latter. The history of a typical case is given on
page 128.

Figs. 72to74 Orthotopic transplantation; right limb bud to left side (het.dd.).
Exp. R. E. 74. Reduplication with atrophic primary member. JN, normal right
limb; PR, primary transplanted limb; DU, reduplicating member. X 10.

Fig. 72 Dorsal view, twelve days after operation; the primary limb already
appears as an appendage of the reduplicating member.

Fig. 73 Ventral view, twenty-one days after operation.

Fig. 74 Lateral view of transplanted limb.

Fig. 74A Dorsal view of same.

Figs. 75 and 76 Orthotopic transplantation; right limb bud to left side (het.
dd.). Exp. R. E. 133. Two almost perfect reduplicating members, one anterior
(A.DU) and one posterior (P.DU) to the primary (PR). The relations of these
limbs are just as in the diagram, fig. 4B. X 10.

Fig. 75 Ventral view, eleven days after operation.

Fig. 76 Ventral view, nineteen days after operation. B, lateral view of pos-
terior reduplicating member.

2 R. E. 133, p. 127. 53 R, BE. 120 and 134.

50 . ROSS G. HARRISON

In one individual* the growth in an anterior direction was well
marked before the posterior reduplicating bud appeared (figs.
78 to 80). The anterior one was finally reduced to a spur, which,
however, was considerably longer than in the next two cases of
the series.*° This is in reality the border-line case and might be

Fig. 77 Orthotopic transplantation; right limb bud to left side (het.dd.).
Exp. R. E. 134. The reduplicating member (DU) is anterior to the primary
(PR). Preserved specimen, eighteen days after operation. X 10.

classed equally well as a reduplication. In one individual’ the
anterior bud was reduced to a slight scar, while in another®’ (figs.
90 to 92) it had only a very slight development and was soon
entirely resorbed.

The cases in which the wounds were not cleaned (for the most
part the earlier experiments made in 1911 and 1912) were seven-

SSRs 108. 55 R. E. 77 and 69. bal ra: Sed o7 RR. a Go.

“Uf

Ziv
4,

MW Z

a af

NY Xs

2d i
AF |W
PRI A iN
\

| \

y \
‘a \
NS

Figs. 78 to82 Orthotopic transplantation; right limb bud to left side (het.dd.).
Exp. R. E. 108. Reduplicating member (DU) in normal position, developed at
expense of original (PR), which is reduced to a long spur. WN, normal right
limb; 1 to 4, numbers of digits. > 10.

Fig. 78 Dorsal view, six days after operation. Reduplicating bud already
more massive, though less prominent, than the primary.

Fig. 79 Lateral view, six days after operation.

Fig. 80 Dorsal view, eight days after operation.

Fig. 81 Ventral view, fifteen days after operation.

Fig. 81A_ Lateral view of limb.

Fig. 81B_ Dorsal view of same.

Fig. 82. Dorsal view, thirty-three days after operation.

Fig. 82A Ventral view of transplanted limb.

Fig. 82B Ventral view of normal right limb.

ROSS G. HARRISON

My
ify

y
a.
iN Wi,

A .

Gees

Nee’

S Z
ee !
SS. tay ,,
es) d
=e FZ Y

Se,
aos }

= A

\— ay}

Sn oe

ee tw

fs yi

\\' '

Figs. 83t085
Exp. R. E. 77.
ing member (DU) which has become a normal left limb. X 10.

Orthotopic transplantation; right limb bud to left side (het.dd.).
Primary member (PR) reduced to a small spur on the reduplicat-

Fig. 83 Dorsal view, eight days after operation.
planted bud already visible. X 10.

Fig. 84 Dorsal view, eleven days after operation..,
than primary.

Fig. 85 Ventral view, twenty-six days after operation.

Figs. 86 to 89 Oriiotopic transplantation; right limb bud to left side (het.dd.).

Reduplication of trans-

Reduplicating bud larger

Exp. R. E. 69. Primary member (PR) reduced to a nodule on the reduplicating
one (DU). X 10.

Fig. 86 Dorsal view, nine days after operation.

Fig. 87 Dorsal view, twelve days after operation.

Fig. 88 Dorsal view, sixteen days after operation.

Fig. 89 Ventral view, twenty-six days after operation.

SYMMETRY IN TRANSPLANTED LIMBS

54 ROSS G. HARRISON

teen in number, in thirteen of which limbs developed. The
distribution of these in the various groups does not show any
significant differences from the cases with cleaned wounds. Eight
(61.5 per cent) gave reduplications and three (23.1 per cent)

Figs. 9) to 92. Orthotopic transplantation; right limb to left side (het.dd.).
Exp. R. E.95. Primary bud (PR) entirely obliterated; the reduplicating member
(DU) a normal left limb. N, normal right limb. X 10.

Fig. 90 Dorsal view, ten days after operation; the primary bud (PR) shows
as a slight nodule.

Fig. 91. Dorsal view, thirteen days after operation.

Fig. 92 Dorsal view, thirty-three days after operation. Owing to weakness
of wrist and hand extensors, the larva has difficulty in bringing its hand to nor-
mal posture.

SYMMETRY IN TRANSPLANTED LIMBS 55

developed into reversed limbs. In at least two of the individuals
reversal was brought about by reduplication. The third is un-
certain. Of the two cases (15.4 per cent) recorded as develop-
ing without reversal, only one is clear. The other died at fifteen
days and was lost, so that the notes made from the living speci-
men could not be verified.

10. Heteropleural transplantations, dorsoventral orientation.
Twenty-six experiments were made in this group. In five out
of the twenty-three individuals that lived the transplanted tissue
was resorbed, and in two others the resulting appendage was im-
perfect or rudimentary, so that sixteen positive cases are available.
Single limbs with reversed asymmetry developed in fifteen, and
only one gave rise to a duplicate structure (table 2).

This group of cases shows that from the first the transplanted
limb buds behave differently from those implanted in dorsodorsal
orientation. When they begin to become prominent, they point
dorsoposteriorly in most cases, though sometimes more sharply
dorsally and frequently more laterally than the normal bud (fig.
93). As the bud grows, it thus occupies a nearly normal posi-
tion, though it may continue for some time to project more
sharply to the side or more dorsally than the normal limb (figs.
94 and 96). When the third and fourth digits develop, they are,
however, not formed on the ventral border of the appendage, as
they would be if the original asymmetry were preserved, but they
come in on the dorsal border, just as in the normal hand of the
side to which they were transplanted (figs. 2, 95, 98, 99, and 102).
The palm of the hand, as in the normal individual, faces the body
of the larva. Besides the one case in which reduplication actu-
ally occurred, there were three others in which slight indications
of doubling appeared, only to disappear later, the more ventrally
lying bud soon being resorbed. Histories of typical cases are
given on page 128. |

In all of these cases there was some retardation of development,
and in some’ it was very marked. A somewhat greater amount
of tissue is lost by disintegration when the limb is placed dorso-
ventrally than when placed otherwise, since the bud does not

BSW Geka gO.

56 ROSS G. HARRISON

fit into the wound so exactly. Besides this there seems to be a
time factor involved in the reversal, which would indicate that
the dorsoventral axis of the limb elements is slightly differenti-
ated, though not irreversibly so.

Figs. 93 to 95 Orthotopic transplantation; right limb to left side (het.dv.).
Exp. R. E. 80. Transplanted limb (7'R) becomes a normal left. N, normal
right limb; 7 to 4, numbers of digits.

Fig. 93 Dorsal view, six days after operation. Transplanted bud much
smaller than normal.

Fig. 94 Dorsal view, fourteen days after operation.

Fig. 95 Ventral view, forty-two days after operation.

—

SYMMETRY IN TRANSPLANTED LIMBS ol

The sole ease in which a double appendage resulted®® is inter-
esting, inasmuch as it shows that the primary bud grows into a
reversed limb, while the reduplicating bud has the original asym-
metry (figs. 103 to 105). This is the opposite of the result ob-
tained when the bud is implanted in dorsodorsal orientation.
(History on p. 129.)

In the experiments with wounds not cleaned the proportion of
reduplications is considerably larger—seven out of fifteen, or
46.7 per cent, as against one case in sixteen in the clean-wound
group. There were eight cases (53.3 per cent) in which normal
limbs with reversal of asymmetry developed, as against fifteen
(94 per cent) in the case of the clean-wound experiments.

11. The shoulder-girdle in orthotopic transplantations. ‘The
above account has dealt only with the external features a the
limb. The shoulder-girdle is likewise of interest.

As the heterotopic transplantations show, a small portion of the
girdle surrounding the glenoid cavity always develops in connec-
tion with the grafted limb. After extirpation of the limb bud,
however, the outlying regions of the girdle, including the supra-
scapula and portions of the procoracoid and coracoid, develop
from cells that are left in the host.°° It was to have been expected,
therefore, that relations of harmony or disharmony would mani-
fest themselves in the shoulder-girdle in orthotopic grafts. Study
of serial sections of some of the cases shows that this is usually
the case. Twelve individuals, belonging to three different groups
have been examined in this way.

The three harmonic grafts (het.dv) all show girdles that are
normal, except that they are somewhat underdeveloped. ‘There
has obviously been a union of host and graft tissues to form a
normal whole, in spite of the fact that the transplanted bud was
from the opposite side of the body.

The nine disharmonie grafts all show some form of irregularity,
and in nearly all cases there is some sort of double girdle with
reversal of the part that is derived from the graft. The condi-
tion of the girdle is complicated by the reduplication of the free

59 R. E. 98.
60 Cf. Detwiler, 718, p. 503, and Harrison, ’18, p. 429.

Wi # ii

} fe pl

100

SSS SERS

——

ROSS G. HARRISON

“

102°. 5

Z LY if 4

SYMMETRY IN TRANSPLANTED LIMBS 59

limb which takes place in most cases (table 2). It is more
readily understood in the two cases in which a single limb of
opposite asymmetry is present.

In the first of these," in which a limb bud from the same side
of the body was implanted in inverted position (p. 32, figs. 39
to 41), there are two entirely distinct shoulder-girdles. The
anterior one has no connection with the other and is undoubtedly
derived from the host, having the characteristics of girdles which
develop after extirpation of the limb bud. The scapula and su-
prascapula are already joined in cartilaginous union with the
procoracoid, but the coracoid is connected with the latter by
ligament only. ‘The girdle belonging to the transplanted limb is
mainly posterior to the other, though there is some overlapping.
It is large to have developed from a transplanted bud, but it
has the characteristics of such. There is a distinct procoracoid
process as well as a large coracoid, both of which project pos-
teriorly from the glenoid cavity. This girdle is clearly reversed,
as is the transplanted limb which is connected with it.

The other single disharmonic limb is the one developed from a
bud taken from the opposite side of the body.” The limb itself
is atrophic (fig. 64). The girdle is double, but the ventral parts
of the two members are fused. The suprascapula, which is single
and belongs to the host, is not connected with the rest. The

Figs. 96 to 99 Orthotopic transplantation; right limb to left side (het.dv.).
Exp. R. E. 107.

Fig. 96 Dorsal view, eight days after operation. Transplanted bud (TR)
smaller than normal (V) and more pointed.

Fig. 97 Dorsal view, thirteen days after operation.

Fig. 98 Lateral view, thirteen days after operation.

Fig. 99 Dorsal view, nineteen days after operation. °

Fig. 99A Lateral view of transplanted limb. / to 3, numbers of digits.

Figs. 100 to 102 Orthotopic transplantation; right limb bud to left side
(het.dv.). Exp. R. E.116. Transplanted limb (7R) becomes anormal left. X 10.

Fig. 100 Dorsal view, seven days after operation. Transplanted bud only
slightly smaller than normal.

Fig. 101. Dorsal view, thirteen days after operation.

Fig. 102 Lateral view, same age.

317. E. 64. oR Sie

60 ROSS G. HARRISON

SS EEE /
GU Yf
Yi if / I,

Figs. 103 to 105 Orthotopic transplantation; right limb bud to left side
(het.dv.). Exp. R. E. 98. Reduplication, the primary (PR) member being a
left (reversed). DU, reduplicating member. X 10.

Fig. 103 Dorsal view, nine days after operation.

Fig. 104 Lateral view, same age.
Fig. 105 Dorsal view, preserved specimen, age fifteen days.

SYMMETRY IN TRANSPLANTED LIMBS 61

ventral part, which forms the glenoid cavity, is in fore and aft
symmetry, with a coracoid and procoracoid process pointing in
each direction. ‘The posterior half of this cartilage has almost
certainly developed in connection with the grafted limb and is
reversed, while the anterior half is derived from the host.

In the disharmonic cases which have reduplicated limbs, the
shoulder-girdles are on the whole less regular, owing to the com-
plex articulations of the double appendages. One of them”
_(hom.dv) is, however, similar to the one first described in having
two entirely separate girdles, one derived from the host and one
from the graft. The suprascapula, procoracoid, and coracoid
of the former are separate chondrifications, situated directly
opposite the corresponding parts of the normal limb. The
girdle of the transplanted limb has a broad flat glenoid cavity for
articulation with the massive humerus. There is a large coracoid
running ventrally from the joint, though without any very well-
marked procoracoid. This girdle is placed some distance pos-
terior to that of the host. Another of these cases (hom.dv) is
more like the second case described above, inasmuch as the dorsal
element (suprascapula) is separate, while the two coracoids
(from host and graft, respectively) are fused. The procoracoid
of the host is a separate cartilage in this case. ‘Two other cases”
are of the same general type with fused coracoids, though they
are rather too young to show all characteristics. Again, two
others** have two scapulae with coracoids fused. There is only
one case*’ that shows in sections practically no sign of doubling
of the girdle, though even in this the coracoid region is thicker
than normal and the glenoid cavity is large in correspondence
with the more massive humerus.

To sum up: The shoulder-girdle in orthotopically grafted
limbs is derived in part from the host and in part from the
transplanted tissue. The former portion retains its normal asym-

& 7, E. 81.

647. E.. 93.

8 T. EK. 68 (hom. dv.) and R. E. 129 (het dd.).
66 R. E. 77 (het. dv.) and R. E. 96 (het. dv.).
877. EK. 60 (hom. dv.).

62 ROSS G. HARRISON

metry, while the latter behaves in accordance with the rules
governing the asymmetry of transplanted limbs. In the dis-
harmonic combinations the portions of the girdles derived, re-
spectively, from the two sources may fuse together or may remain
entirely separate. In the harmonic combinations they unite to
form a single normal girdle.

12. Summary of the results of orthotopic transplantations. The
orthotopic transplantations develop according to the same rules
as the heterotopic. In the homopleural dorsodorsal and the
heteropleural dorsoventral groups rules 1 and 2 (p. 4) are very
closely followed. In the former the limb buds, being right side
up, retain their normal asymmetry; and in the latter, being upside
down, they reverse it. In both groups this results in limbs which
correspond to the side on which they are implanted (harmonic
combinations).

In the other two groups he primary single limbs which
develop do not correspond to the organic environment, since
the homopleural graft, when placed upside down, becomes re-
versed, and the heteropleural graft right side up retains its origi-
nal prospective asymmetry. In these combinations, which
have been called disharmonic, single limbs are, however, the
exception. It is here that rule 3 comes into play. Reduplica-
tions occurred in 71.1 per cent of the cases in the homopleural dor-
soventral group and in 80.6 in the heteropleural dorsodorsal.
The former includes only one case of single limb reversed. In
this class are also five cases of reversed single limbs, which are
fundamentally the same as reduplications, the original limb
having been suppressed or resorbed. The disharmonic relation
thus augments immensely the tendency to reduplicate. In the
case of the heterotopic grafts, on the contrary, the greater pro-
portion of reduplications occurs in the harmonic combinations.
This curious fact will be discussed below (p. 107). The ten cases
of non-reversed single limbs which resulted from homopleural
inverted buds are, as already pointed out, exceptional in that the
limb regained its normal posture gradually during development
by rotation at the base.

SYMMETRY IN TRANSPLANTED LIMBS 63
C. Superposed limb buds

In the preceding study of transplanted limbs certain experi-
ments were described, which showed that the mesoderm from
two limb buds, when fused together; would develop into a single
normal limb. At first larger than normal, the size of such a
limb is soon regulated. In the former communication only those
experiments were considered in which the orientation of the
superposed bud was normal (hom.dd). The effect of the orienta-
tion of the graft will now be taken up.

TABLE 3

Superposed limbs. Summary of results

NORMAL SINGLE
| WITH REDUCED | REDUPLICATED
REDUPLICATION

NUMBER OF NORMAL
EXPERIMENTS SINGLE LIMBS
OPERATION

Posi- | Num- P Num- Pe Num- P
Total ave ee seat ber cent ; bee Baw
lorena ame ios. aks sae 5 5 5 100 0 00 0 00
Ta liniie Cohan eae ne S a 1 20 0 00 4 80
LEIS (Ol oe ee em ts 6 5 0) CO 1 20 4 80
JBI. (0 N7ty Sane eee 9 5 5 160 0 00 0 00
ioelicr ees... ceil Be Oe fel 55 | 5 8") "40

The experiments are summarized in table 3. There were
twenty-five operations, of which twenty are available for the
analysis. Two of the combinations, the ones which the ordi-
nary transplantations have shown to be harmonic (homopleural
dorsodorsal and heteropleural dorsoventral) yielded only normal
appendages (ten cases). The two disharmonic’ combinations
(homopleural dorsoventral and heteropleural dorsodorsal) yielded
reduplications in nine cases out of ten. One case, in which one
member of the duplicate limb was reduced to a spur, is included
among the reduplications. aah,”

13. Homopleural transplantations, dorsodorsal orientation. In
this group development went forward with a minimum of dis-
turbance. The only abnormal feature to note is the large size
of the double bud in certain individuals. In several of the cases

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, No. 1

64 ROSS G. HARRISON

20,

Dp) Ew :

AL,

< Lo

109

1098 \ |

SYMMETRY IN TRANSPLANTED LIMBS 65

the difference from normal size persisted for twelve or more days, ®8
gradually diminishing during that period (fig. 106). In others
the difference was less marked, though in all some difference in
favor of the limb on the operated side was noted.

14. Homopleural transplantations, dorsoventral orientation.
Four out of the five cases in this group gave rise to reduplications
in the grafted limb.

The reduplications vary. One®? is a typical case of radial mir-
roring of the lower part of the forearm and hand (fig. 107).
Another” is similar, except that the anterior member is itself
reduplicated, the hand being a nearly symmetrical complex with
four digits. These two individuals are in every respect like
those cases of reduplication resulting from simple inverted limb
buds, in which the primary member is reversed and is accom-
panied by a non-reversed twin which takes up the normal posi-
tion. In two other cases” the reduplication is less and is of a
character not necessarily attributable to disharmonic combina-
tion, though there is nothing to indicate that it is not due to such
a cause. Principally the digits are involved (fig. 108).

In the remaining case” a normal limb developed. This one
and possibly also the two foregoing are analogous to those cases
of simple transplantation in which the inverted limb bud develops
into a normal limb without reversal by means of rotation (p. 40).

Fig. 106 Superposed limb bud; right limb bud to right side, normal position
(hom.dd.). Exp. 8. E. 3. Normal limb (7R) on operated side. N, normal left
limb. Preserved specimen, ventral view, eighteen days after operation. X 10.

Fig. 107 Superposed limb bud; right limb to right side inverted (hom.dv.).
Exp. 8. E. 18. Reduplicated appendage (TR) on operated side. Preserved
specimen, ventral view, eighteen days after operation. X 10.

Fig. 108 Superposed limb bud; right limb bud to right side inverted (hom.dv.)
Exp. 8S. E. 9. Operated limb (TR) normal except for reduplication of second
digit. X 10.

Fig. 109 Superposed limb bud; right limb bud to left side (het.dd.). Exp.
S. E.6. Reduplication with heteropleural member reduced to spur (S). X 10.

Fig. 109A Outline of limb bud from above. Five days after operation (free-
hand sketch).

Fig. 109B Same, eleven days after (free-hand sketch).

cS Ht. 3. BoD. 1, 18: WSs Bie2. 1§. E, 9 and 14, 725. E. 10.

66 ROSS G. HARRISON

Sita
es UCT M0,
ions

SYMMETRY IN TRANSPLANTED LIMBS 67

15. Heteropleural transplantations, dorsodorsal orientation. In
this group the results of five experiments all fall clearly within the
same category. A normal appendage in approximately normal
position developed, and this had a reduplicating member attached
to its radial border. The differences between the cases consist
in the extent of reduplication and the degree of development
attained by the disharmonic (anterior) member. The most extreme
cases are one” in which the heteropleural member is reduced to a
spur (fig. 109), and one” in which there is almost complete doub-
ling from just below the shoulder down, the heteropleural member
being, however, atrophic (figs. 110 to 113). Two of the three
remaining cases” are very similar to one another, the anterior
member with two digits arising from near the elbow. In the
third individual the hand is reduplicated’® externally, and the
whole arm is somewhat shorter and thicker, indicating some
degree of internal reduplication.

16. Heteropleural transplantations, dorsoventral orientation.
The five cases in this class all gave normal limbs (figs. 117 and
118). Three of them showed slight indication of doubling (figs.
114 and 115) in the early stages of development (four or five days
after operation), but the more ventral-lying prominence had dis-
appeared at the time of the next observation in each individual

Figs. 110 to 113 Superposed limb bud; left to right side (het.dd.). Exp.
S. E.12. Reduplication. WN, normal left limb; 7R, operated limb; HET, hetero-
pleural member; HOM, homopleural member.

Fig. 110 Dorsal view, five days after operation.

Fig. 111 Lateral view, same age.

Fig. 112 Ventral view, ten days after operation.

Fig. 118 Ventral view, seventeen days after operation.

Figs. 114 to 119 Superposed limb bud; left to right side (het.dv.). Exp.
S. E. 11. Operated limb (7'R) of large size; N, normal unoperated limb. X 10.

Fig. 114 Dorsal view, five days after operation.

Fig. 115 Lateral view, same age. A distinct nodule or bud (DU) on ventral
border of limb was soon afterward resorbed.

Fig. 116 Ventral view, ten days after operation.

Fig. 117 Lateral view, same age.

Fig. 118 Dorsal view, seventeen days after operation.

Fig. 119 Lateral view of limb, same age.

eS, Hs 6. Soi. Ae 7% S. E. 8 and 21. #8, E. 15.

68 ROSS G. HARRISON /

(figs. 116 and 117). A sixth case’? is inconclusive owing to gen-
eral weakness of embryo, but it is not inconsistent with the
results in the other cases. A typical history is given in the appen-
dix, (pz 31).

Though more than the normal amount of material is present
in the composite limb bud, in two of the cases of this group the
developing limb is recorded at first as slightly smaller than the
normal. In one case no difference in size was noted, while in two
the limb on the operated side is noted as larger. In the other
harmonic combination (homopleural dorsodorsal) all cases
showed the limb on the operated side to be somewhat larger in
size.

17. Discussion of experiments with superposed limb buds. The
principal differences between these experiments and those of
simple transplantation are that in the former the tissue available
for the formation of the lmb is approximately double in
amount, and there is a mixture of tissues having two different or-
ientations except in the one group, homopleural dorsodorsal. In
the harmonic combinations the amount of tissue is so regulated
that after a time size-differences disappear. The amount of tis-
sue is, moreover, never quite double that of the normal limb be-
cause of the material lost by the operation and of the general
retardation of growth due to the same cause. In the case of
the heteropleural dorsoventral grafts, which are classed as har-
monic, some readjustment must be necessary, as shown by the
amount of retardation.

In all of the disharmonic combinations there is a mixture of tis-
sues differently oriented and with different prospective meaning
as regards the particular asymmetry of the future limb. The
twin limbs that arise are, therefore, not necessarily due to redu-
plication by budding, as they must be in the simple transplanta-
tions, but probably in part at least to the circumstance that one
of the pair develops out of the original limb bud, while the other
is from the transplanted tissue.

LES 135m bop

SYMMETRY IN TRANSPLANTED LIMBS 69
D. Transplantation of half buds

Partly as a further test of the question of equipotentiality and
partly to study more thoroughly the effect of harmonic and dis-
harmonic combinations, a series of experiments with half limb
buds was instituted. Instead of removing the whole circular
disc comprising the limb rudiment, a semicircular piece was cut
out, the wound bed carefully cleaned, and the removed portion
replaced by a piece of similar size and shape from another limb
bud. Considering only vertical and horizontal halves and re-
placing vertical only with vertical and horizontal only with hori-
zontal, there are sixteen different experiments possible, which
have been numbered in the diagram (fig. 120) from 1 to 16.

There are five different pairs of attributes, which appear as
alternatives in the operations. Thus the transplanted half bud
is either—1) homopleural (hom.) or heteropleural (het.); 2) up-
right (dd.) or inverted (dv.) ; 3) homogeneous (homogen.) or hetero-
geneous (heterogen.); 4) vertical (vert.) or horizontal (horiz.); 5)
anterior (ant.), dorsal (dors.) or posterior (post.), ventral (vent.).

This aggregation would consist of 2° or thirty-two classes, were
it possible to combine the attributes of operation independently
without restriction, as would be the case were the pieces rectangu-
lar. Since, however, they are semicircular they fit in only half
the cases, and the total is therefore reduced to sixteen. All of
the possible experiments have been performed.

If both halves of the dise are considered movable, further pos-
sibilities open up. There would then be thirty-two different com-
binations, which, however, in eight cases would be practically
identical with the experiments where the whole disc is transplanted.
None of these experiments have been performed, since the tech-
nical difficulties would be at least doubled, and, as far as the
study of either of the questions at issue is concerned, they would
offer no advantage over those in which only one half of the bud is
transplanted. Again, were the limb disc homogeneous, either or
both halves could be turned inside out and then one hundred
and twenty-eight different combinations would be possible.
These are precluded, as in the case of the whole discs, by the

70 ROSS G. HARRISON

impossibility of grafting successfully pieces with the mesoderm
turned toward the outside and by the difficulty of handling
pieces of mesoderm free from ectoderm without disturbing their
arrangement. Perhaps we may consider ourselves fortunate in
being subject to such restrictions.

HOM.DV

4
HET.DD (A LIRYA

HET.DV

Fig. 120 Diagram showing the sixteen possible combinations (1 to 16) obtain-
able by transplanting half limb buds. The shaded area signifies the stationary
half, the clear area the transplanted half. R, right; L, left; D, Dorsal; V, ven-
tral; A, anterior; P, posterior. The operations are represented as on the right
side of the embryo.

Returning to the experiments actually carried out (fig. 120), we
find that four of them consist merely in replacing the excised
piece with another of exactly the same kind in normal orientation.
These serve, therefore, as controls for testing the effect of the
operation as such on the further course of development. It is
also seen that half of the combinations are harmonic and half dis-
harmonic (p. 8). Half are of course homogeneous or com-

SYMMETRY IN TRANSPLANTED LIMBS 71

posed of two similar halves, while the other half are heterogene-
ous. Of the former, six belong to the disharmonic group and
only two to the harmonic, while of the latter the reverse is the
case, a circumstance that affects the proportionate results of
the experiments.

The effect of removal of the various halves of the limb rudiment
has already been described (Harrison, 718). As shown by such
experiments, any half of the limb bud can give rise to a whole
limb, though quantitatively the material is eecentrically distrib-
uted, there being more limb-forming tissue in the dorsal and
anterior halves than in the ventral and posterior halves, respec-
tively. Accordingly, four of the homogeneous combinations
would have somewhat less than the normal amount of tissue,
while four would have a little more. In the later experiments
an attempt was made to compensate for this by not cutting the
area exactly in half.

Owing to the large number of combinations in the experiments,
it has not been possible to perform a sufficient number of each,
for accurate statistical treatment. The number is sufficient, how-
ever, to compare the more comprehensive groups; for instance the
homogeneous with the heterogeneous and the harmonic with the
disharmonic.

Seventy-nine operations were done, sixty-eight healing success-
fully. Badly defective limbs developed in but four cases, so that
sixty-four remain for the purpose of the analysis. These experi-
ments are summarized in table 4.

From the results of transplanted whole limbs we should expect
the following to take place; the harmonic combinations should
give rise to simple normal limbs, the disharmonic to reduplica-
tions. The homogeneity or heterogeneity of the combination
should not be expected to make any difference in view of the other
tests of the equipotentiality of the system. These expectations
were in the main realized, probably in fifty-five out of the sixty-
four cases (85.9 per cent) (table 7). There are, however, sources
of confusion, which in certain cases make several interpretations
possible, and which for this and other reasons must not be over-
looked. For example, it is known from experiments with whole

12, ROSS G. HARRISON

limb buds, that a normal limb may arise from a disharmonic com-
bination by the suppression of the original bud or by its reduction
to a mere excrescence on the reduplicating member, which latter
may develop into a normal limb. Eight of the ten normal cases
which would otherwise appear anomalous may certainly be thus
explained, and possibly the remaining two. It has been found

TABLE 4
Transplantation of half limb buds. Summary of results of actual experiments
OPERATION RESULTING LIMB
= | Designa- | Nor-
2 Side of : Direction tion of = mal | Re-
N ee oe vane Composition ee (eae No byre dupli- pee Dead
aus half tion
1 | hom. | dd | heterogen. | vertical | ant. 2 0 0 0 1
2 | hom. } dd | heterogen. | vertical | post. 2 0 0 @ 1
3 | hom. | dd | heterogen. | horiz. dors. | 2 0 0 0 0
4 | hom. | dd | heterogen. | horiz. vent. | 2 0 0 0 0
5 | hom. | dv. | homogen.. | vertical | ant. 0 2 2 0 1
6 | hom. | dv | homogen. | vertical | post. | 0 0 oe n0 2
7 | hom. | dv | homogen. horiz. dors. 0 2 2 0 0
8 | hom. | dv | homogen. horiz. vent. | 0 0 4 1 0
9 | het. dd | homogen. vertical | ant. 2 2 0 3 1
10 | het. dd | homogen. vertical | post. | 0 1 SEO 0
11 | het. dd | heterogen. | horiz. dors. 0 0 5} 0 1
12 | het. dd | heterogen. | horiz. vent. | 0 0 4 0 1
13 | het. dy | heterogen. | vertical | ant. 4 0 1 0 1
14 | het. dv | heterogen. | vertical | post. 4 0 0) 0 2
15 | het. dv | homogen. }_ horiz. dors. 5 0 2 0 0
16 | het. dv | homogen. | horiz. vent. | 6 0 0 0 1
Total number of cases, 79; positive cases, 64... .| 29 7 i 28 atl

1 Includes one case of anomalous reduplication.
2 Includes two cases of anomalous reduplication.

also that almost any transplantation or even simple defect ex-
periment may sometimes bring about reduplication. The three
anomalous reduplications, being slight, are probably of this class.
A further source of error might arise from the circumstance, that
either the grafted or the stationary half may in certain cases be
solely responsible for the limb that develops; for it is known, on
the one hand, that any graft may be resorbed and, on the other,

SYMMETRY IN TRANSPLANTED LIMBS 73

that when half the disc is excised, complete suppression of
development may sometimes result, probably through accidental
injury to the remaining part. The result of the former contin-
gency would be confusing, owing to the development of a normal
limb in place of a reduplication. .

In connection with these questions it must also be borne in
mind that the cases of union of two disharmonic halves differ,
with respect to the disharmony of the combination, from those
in which the limb-bud has been transplanted as a whole. In
the former only one half of the rudiment is involved, the other
being in all cases harmonic with the surrounding tissues. We
are, therefore, dealing with a rudiment that is disharmonic in
itself, while in the case of the whole limb the transplanted bud is
harmonic in itself, though disharmonic with respect to the organ-
ismasawhole. This might possibly give rise to some differences
in the results in the two classes of experiments.

In view of these considerations, we should not expect the trans-
plantation of half buds to give such clear-cut results as the
experiments with whole ones. On the other hand, it must not
be overlooked that the sources of confusion above enumerated,
while accounting for nearly all of the anomalies, also render less
cogent the cases which conform to the rules. Nevertheless,
after taking all circumstances into consideration, it can scarcely
be doubted, that the experiments with half discs do afford a
valuable confirmation of the results obtained from the other
experiments.

18. Homopleural transplantations, dorsodorsal orientation. The
eight cases of homopleural grafts in upright orientation (hom.dd),
two involving each half of the bud, all resulted in normal limbs,
as was to have been expected, for this operation is nothing more
than replacing an excised portion with one exactly similar. Only
slight retardation of development is recorded in some of the cases.

19. Homopleural transplantations, dorsoventral orientation. The
nineteen experiments with homopleural grafts in inverted posi-
tion (hom.dv) resulted, in accordance with expectation, in a
large number (fifteen) of duplicities’® (figs. 121 and 122). The

78H. BE. 29 and 31.

74 ROSS G. HARRISON

remaining four cases were normal. In the early stages, however,
all of the latter gave evidence of reduplication (fig. 123). The
limb bud, when it first appeared, showed two distinct nodules or
prominences, one of which developed into a normal limb. In

Figs. 121 and 122* Transplantation of half limb bud (comb. 6, fig. 120); pos-
terior right to anterior right (hom.dv.). Exp. H. E. 31. Partial reduplication
of hand, mirror plane being radiodorsal. Arm and medial hand homopleural.
1 to 4, numbers of digits of main hand; 2’, 3’, digits of reduplicating member.

Fig. 121 Outline of normal (N) and operated (7R) buds from above, nine
days after operation (free-hand sketch).

Fig. 122. Ventral view, preserved specimen twenty-one days old. X 10.

Figs. 123 and 124 Transplantation of half limb bud (comb. 7, fig. 120); dorsal
right to ventral right (hom.dv.). Exp. H. E.2. Operated limb slightly thicker,
reduplicating member reduced to a nodule (S).

Fig. 123 Outline of operated limb, dorsal view eight days after operation
(free-hand sketch).

Fig. 124 Ventral view preserved specimen, twenty days old. X 10.

SYMMETRY IN TRANSPLANTED LIMBS 75

three cases the other prominence persisted also, at the elbow in
one?’ in the form of a spur (fig. 124), and as a nodule at the
shoulder in two others.8° In the remaining case*! all external
traces of reduplication disappeared.

On the other hand, two of the cases of reduplication are of an
anomalous nature and cannot be regarded as conforming to the
rule. Both of these were experiments in which the anterior half
of the limb bud was replaced by a posterior half. We should
expect in such a case to find posteriorly a homopleural member
developed out of the stationary portion of the bud, while anteri-
orly there should be a reversed limb which might itself be redu-
plicated. . The opposite is, however, true. In both cases the an-
terior member is not reversed. The posterior member is reversed
(a left) in one case® (fig. 126) and in the other* it is itself double,
the anterior portion being reversed and the posterior homopleural
(fies i25)., ;

The operated limb in all of these cases was composed of two
homogeneous halves.

Histories of typical cases are given in the appendix (p. 132).

20. Heteropleural transplantations, dorsodorsal orientation.
This combination, being disharmonic, yielded out of seventeen
cases twelve duplicities (figs. 127 and 129) and three limbs that
became normal by reduction of the reduplicating member (fig.
130). In two individuals* normal limbs resulted without exter-
nal evidence of incipient doubling, and two of the reduplications,
in one of which both members are of the same side in linear series,
are of an anomalous nature. This makes four cases out of sev-
enteen that do not follow the rule. Two of the combinations
are heterogeneous; all of these conform to the rule except one of
the anomalous reduplications. The other three non-conforming
cases belong in the homogeneous class, and it is interesting that
all of the normal cases in this group resulted from the combina-
tion of two like halves.

MAS (, 1, 2 a Jal 18, 16}
80 H. BK. 18 and 21. CM & [al ae
su H. BH. 4. 84H. R. E. 43 and 44.

76 ROSS G. HARRISON

Fig. 125 Transplantation of half limb bud (comb. 6, fig. 120); posterior half
right to anterior right (hom.dv.). Exp. H. E. 5. Two limbs, the posterior of
which has a double hand. Anterior member an almost normal right; of the two
parts of the posterior member, the anterior one is a left hand and the posterior,
aright. X 10.

Fig. 126 Same operation (comb. 6, fig. 120) (hom.dv.). Exp. H. E.13. Ulnar
reduplication! HOM, homopleural hand; / to 3, digits of same; HET, hetero-
pleural hand; 1’ to 3’, digits of same; N, normal (unoperated) left limb xX 10.

Fig. 127. Transplantation of half limb bud (comb. 10, fig. 120) ; posterior half
right to anterior left (het.dd.). Exp. H.R. E.1. Double hand. HOM, homo-
pleural hand with digits (7 to 4); HET, heteropleural hand with digits. (2’ to
Be o< OL

SYMMETRY IN TRANSPLANTED LIMBS Ut!

The double limbs are of various degree and kind. The least -
involved is one in which only the first digit is doubled. In this
individual the ventral half of the limb was replaced by a ventral
half of the opposite side, and in all probability very little limb
material was actually transplanted, since, in the embryo from
which the graft was taken, the operated limb developed almost
asrapidly asthe normal. In five other cases* (fig. 127) the whole
hand is involved, with indications that in three of these at least’’
the internal reduplication extends farther proximally. In two
cases the fore arm and hand®s (figs. 128 and 129) are externally
double, and in one,*? which was not fully developed when pre-
served, doubling would probably have shown from near the
shoulder down. In two of the individuals®? there are secondary
reduplications. In the two anomalous cases"! (figs. 1381 and 132)
there are two entirely separate limbs.

Histories are given on page 134. Unfortunately, external
observation does not always reveal the relations of each of the
two halves of the bud to the developing members.

21. Heteropleural transplantations, dorsoventral orientation.
Out of twenty-two successful experiments in this group nineteen
resulted in the development of normal limbs (fig. 133), which is
according to rule, and only three gave rise to reduplications (fig.
134). Two of the latter’? involved only the radial digits, in
which paimar reduplication was present, the limbs being other-
wise normal. In the remaining one a bifurcated appendage
arose, but the dorsal branch remained merely as a spur attached
to the main limb, which was normal though undersized and with
slight syndactyly.

In one case,*? which has been classed as normal, a filamentous
appendage probably not a limb, developed a short distance ven-
tral to the main limb, which was normal though slightly shorter.

SARS Bi 30: 90H. R. EH, 21 and 47.
86 HW. R. E. 1, 15, 46, 47, and 48. 1H. R. E. 9 and 20.

87H. R. EB. 15, 46, and 47. 2A. 2. LOvand 1G:
88H. R. E. 5 and 21. iy Re Bez:

aOR. B28

78 ROSS G. HARRISON

This group of experiments is interesting because two of the
combinations (ventral half in place of dorsal and dorsal in place
of ventral) are homogeneous. Out of thirteen such cases, normal
limbs developed in eleven.

For histories of representative cases see appendix (p. 136).

13O

Figs. 128 and 129 Transplantation of half limb bud (comb. 11, fig. 120) ; dorsal
half right to dorsal left (het.dd.). Exp. H. R. E.5. HOM, homopleural member;
HET, heteropleural member.

Fig. 128 Outline of limb from above, ten days after operation (free-hand
sketch).

Fig. 129 Preserved specimen lateral view, nineteen days old. X 10.

Fig. 130 Transplantation of half limb bud (comb. 9, fig. 120); anterior half
left limb bud to posterior right (het.dd.). Exp. H. R. E.11. Normal limb with
small spur (S). X 10.

22. Discussion of experiments with half buds. In order to es-
tablish the conclusion stated in the introduction to this section,
that it is the harmony or disharmony of the half-and-half combi-

SYMMETRY IN TRANSPLANTED LIMBS 79

nation and not one of the particular qualities of the operation that
determines whether normal or reduplicated limbs arise, it will be
necessary to examine the numerical results of the experiments
more carefully.

If we take the actual figures of the experiments and examine
the qualities in pairs, we find the actual number of each class and

132

Fig. 131 Transplantation of half limb bud (comb. 10, fig. 120); posterior half
left side to anterior right (het.dd.). Exp. H. R. E.9. Anterior member a right
(homopleural), the posterior one a left (heteropleural), but imperfect. Ventral
view of specimen preserved forty days after operation. X 10.

Fig. 131A Lateral view of limbs of same.

Fig. 132 Transplantation of half limb bud (comb. 11, fig. 120); dorsal half
left side to dorsal right (het.dd.). Exp. H. R. E. 20. Two right limbs, the ante-
rior one imperfect. X 10.

80 ROSS G. HARRISON

the proportion of normal results to be as given in table 5, column
6. ‘Normal by resorption,’ being fundamentally the same as
reduplication, is classed as such.

Fig. 133 Transplantations of half limb bud (comb. 16, fig. 120); ventral half
left side to dorsal right (het.dv.). Exp. H. R. E. 36. Normal limb.

Fig. 134 Transplantation of half limb bud (comb. 18, fig. 120); anterior half
left side to anterior right (het.dv.). Exp. H. R. E. 10. N, normal left arm; TR,
grafted arm with palmar reduplication of .two digits (DU). X 10.

The one thing that stands out is the great difference between
the results of the harmonic and those of the disharmonic combi-
nations. In the case of none of the other attributes of opera-
tion is there anything like the same difference between those of
a pair, though in the case of homogeneity vs. heterogeneity the
difference is considerable (37.1 vs. 64 per cent).

SYMMETRY IN TRANSPLANTED LIMBS 81

However, the comparisons cannot be accurately made without
the same number of experiments in each class, unless made by
means of percentages. This is quite obvious, for instance, in
the case of the homopleural transplantations with dorsodorsal
orientation, which all result in normal limbs. The relatively

TABLE 5
Transplantation of half buds. Comparison of pairs of qualities of operation

° NUMBER OF CASES ASAE
| pall eee
PAIRS OF QUALITIES COMPARED |Normal| = ee FOR INEQUAL-
Norma! by nie Tctal ae SLE SE ea
“tion | cated? “en
omopleuralo. ... 2. cee ae 8 4 11 23 | 34.8 50.0
Heteropleurils.. 22. epeneee sear Al 3 13 37 | «56.8 50.2
Worsodorsale 542 ch. SAE el | nO 3 10 23 | 43.5 56.25
DonsoventGall.): 2)... ea ae eee 19 4 14 ay |) Ged! 43.9
Homogeneous... .))) 92). See eee 13 7 15 30) ode Heth
Hetenozeneous). 4. . = = aocnete ae, 16 0 9 25 | 64.0 72.5
WETTIG Mle ae sicc-cns Mente nokta 14 5 8 2 ole9 53). 09
EMOUENZ ONG ete ce) kes sc Nore hele NLD 2 16 3a | 45.5 46.8
Anterior or dorsal half trans-
JST STG ea? 15 6 11 382 | 45.45 50.2
Posterior or ventral half trans- | ,
plantede a eeaees |e hee 14 1 13 28 | 50.0 50.0
Harmonic (hom. dd and het. dv) | 27 0 3 30 | 90.0 93.9
Disharmonic (hom. dv 2nd het. | | |
Casey ee As) eet el eee DA NTE. 21 30 (Bey 6.25

1 The four cases of anomalous reduplication have been omitted from this
tabulation.

small number of cases in this group affects the record for homo-
pleural transplantations by reducing considerably the number
of cases that would have developed normally, thereby giving
undue weight to the larger number of dorsoventral cases which
result in reduplications. Likewise the dorsodorsal vs. dorsoven-
tral record is influenced by the relatively small number of homo-
pleural cases.

$2 ROSS G. HARRISON

On this account the operations in each class have been reduced
to a common basis. While the probable error of these figures is
in most cases large, the comparisons resulting therefrom are
no doubt much more reliable than those resulting from the
figures of the actual experiments. They are given in the last
column of table 5.

In examining this table we find that there is little or no asso-
ciation between the experimental results and the following quali-
ties of operation: homopleural vs. heteropleural, dorsodorsal vs.
dorsoventral, vertical vs. horizontal, anterior and dorsal vs.
posterior and ventral, the deviation from total lack of associa-
tion (50 per cent) being in the most extreme case but 6.1 per cent.
When we examine the figures with reference to the pair, homo-
geneity vs. heterogeneity, we find that there is a much wider dif-
ference (27.7 per cent as compared with 72.5). This would have
to be regarded as a significant difference but, as will be seen below
it is only secondarily so. The marked association between the
harmonic combinations and normal development (93.9 per cent)
and the very small proportion of normal development (6.25 per
cent) in the disharmonic group, show that it is largely this pair
of attributes that determines whether development will be normal
or not. This quality of harmony or disharmony, however, is
not like the simple qualities of side of origin, orientation, or direc-
tion of the incision, but is itself a combination of two of them.
Those that are harmonic are the homopleural dorsodorsal and
the heteropleural dorsoventral combinations, the other two being
disharmonic, as in the experiments with whole limb buds.

_ When we consider the homogeneous and heterogeneous com-
binations, we find them unevenly distributed with respect to
the harmonic and disharmonic. This is on account of the restric-
tion of operation due to the semicircular shape of the trans-
planted pieces, which makes half of the combinations impossible
of execution. Were these all possible, there would be complete
symmetry in the aggregation as a whole. In reality, it will be
recalled, six of the homogeneous combinations are disharmonic,
while only two are harmonic. On the hypothesis that it is the
harmony of the combination that determines normal develop-

SYMMETRY IN TRANSPLANTED LIMBS 83

ment, and with an equal number of experiments in all of the six-
teen possible classes, the expectation would be that only 25
per cent of the homogeneous and 75 per cent of the heterogeneous
would be normal. This corresponds closely to the figures 27.7
and 72.9, respectively, found in table 5.

As regards the question of equipotentiality, the results of these
experiments are equally striking. The two homogeneous com-
binations which, according to expectation, should yield normal
limbs did so. Thus two ventral halves yielded normal limbs in
all six experiments, as did two dorsal halves in five out of
seven. In three cases of disharmonic homogeneous combination
normal limbs developed by resorption of the reduplicating bud;
two of these were from two anterior halves and one from two
posterior. ‘Two further cases of normal limbs from two anterior
halves developed without external evidence of resorption. While
the last five, if interpreted according to the rules, can only be
accepted as evidence of equipotentiality in so far as they show
that a whole limb can develop out of a single half bud, the
others show that two half buds which are alike except that they
are from opposite sides of the body may give rise, when harmonic,
to a single normal limb.

GENERAL DISCUSSION

In this section the following questions will be considered:
1) the foundation of the rules of symmetry; 2) the mode of repre-
sentation of symmetric relations in the limb rudiment; 3) the
formation of reduplications; and 4) form regulation and function
in transplanted limbs.

E. The rules of symmetry.

The validity of the rules of symmetry which have already
been stated in the introduction (p. 4) will best be realized by
considering the results of the several experiments in tabular
form (table 6). Conformity is shown most strikingly in the
heterotopic group, where there is only a single apparent excep-
tion in forty-five cases; and this exception, as already pointed
out, is probably due to an error in recording the operation.

84 ROSS G. HARRISON

In the orthotopic group the lowest percentage of conformity
(65.8) is found in the inverted homopleural buds (hom.dv),
where the exceptions are due entirely to adjustment by rota-
tion of the limb as a whole. In the superposed buds the sole
exception is due probably to the same cause. The exceptional

TABLE 6
Showing conformity to rules in the several experiments
: OPERATION RESULTING LIMB
Conforming *

Side of | O,jen-| Harmonic to rules ISUTE GES

Type of experiment origin of an or dis-_ ; Total
graft harmonic Aun Be oo PeLeent

Whole bud heterotopic..| hom. | dd | harm. 7 |100.0) 0 7
Whole bud heterotopic..| hom. | dv | disharm. | 12 |100.0} 0 12
Whole bud heterotopic. .| het. dd | disharm.| 10 |100.6) 0 10
Whole bud heterotopic..| het. dv | harm. 15. | 9358) <2.(?)) 6.3.02)) 36
Whole bud orthotopic...| hom. | dd | harm. 9 |100.0} 0 9
Whole bud orthotopic...| hom. | dv | disharm.| 25 | 65.8/13 [34.2 38
Whole bud orthotopie...| het. dd | disharm. | 31 /100.0} 0 31
Whole bud orthotopic...| het. dv | harm. 16 {100.0} 0 16
Superposed buds. ......| hom. | dd | harm. 5 |100.0) 0 5
Superposed buds.......| hom. | dv | disharm. 4 | 80.0) 1 {20.0 5
Superposed buds....... het dd | disharm. 5 |100.0| 0 ~5
Superposed buds....... het. dv | harm. 5 |1C0.0} 0 5
Half buds). a s.cccaes,. ....2-|, Hom dd | harm. 8 |100.0) 0 8
Halfbuds. 5.5. 544... +. >| som dv | disharm. | 15 {100.0} 0! 15
Hate buds. .o. eee. 0... | sete dd-|disharm. | 13 | 86.7] 2! |13.3 15
Halt buds teak. o2.: | hett dv | harm. 19 | 86.4; 3 /13.6 22
Wotala SRE eee eck COs oak accion 199 | 90.9/20 Ort 219

Average Ofspercentaress ali serie, . nents < o2ais 94.5 5.5

1 Four cases of anomalous reduplication have been omitted from this tabula-
tion, inasmuch as they cannot be classified either as conforming or as exceptional.

cases arising after transplantation of half buds have already
been discussed, and, as pointed out above, there are in this group
of experiments obvious disturbing factors which might readily
account for the exceptions. Taking the experiments as a whole,
90.9 per cent conform and 9.1 per cent are exceptional, but if
allowance is made for the difference in the number of experiments
in each class, assuming that each is a fair sample of what would

SYMMETRY IN TRANSPLANTED LIMBS 85

occur in a large number of cases, then but 5.5 per cent are
exceptional.

The behavior of transplanted limb buds in accordance with
the above rules indicates that the posture and the asymmetry
of the limb is determined neither by the limb itself nor by its
surroundings exclusively, but by an interaction between the two.
This is best described by the assumption, that in the stages exper-
imented upon the anteroposterior axial differentiation is already
determined within the limb bud, while the ventrodorsal axis (prob-
ably radio-ulnar of the grown limb) is determined by its orien-
tation with reference to the surrounding tissues of the host (fig.
135). In a given place a right limb bud upside down thus be-
haves like a left limb bud right side up and vice versa (fig. 2).
It is scarcely necessary to point out that this is not a gravity
effect, for the embryo lies on its side during the period when the
dorsoventral axis of the limb is determined, ‘upside down’ being
used here merely with reference to the cardinal points of the
embryo itself.

What the nature of the influence exerted by the organic envi-
ronment may be, has not been determined. Whether it acts upon
the intimate structure of the limb bud or directly upon the differ-
entiating systems contained therein, without affecting the inti-
mate structure as a whole, cannot be answered from the present
data (p. 101). The influence is not sharply localized, for it is
the same both in the limb region itself and elsewhere along the
flank of the embryo, so that it is probably an effect of the axial
differentiation of the tissue elements themselves. It is possible
that light may be thrown upon this question by transplanting
the limb bud to the dorsal or to the ventral midline of the embryo.

F, The mode of representation of symmetric relations in the limb
rudiment

The question whether the adult parts are localized in the
germ, forming a mosaic, must be answered in the negative for
the limb bud, as used in the experiments, i.e., if we consider as
such a dise of tissue, three and a half somites in diameter, cen-
tering ventral to the fourth myotome, and leave out of account
the outlying regions from which certain portions of the shoulder-

86 | ROSS G. HARRISON

girdle develop. This conclusion is based upon the following
evidence derived from the experiments: 1) After extirpation of
any half of the limb bud, a complete normal limb may develop
from the remaining half; 2) fusion of two limb buds by super-
position is followed, if the combination is harmonic, by the devel-
opment of a single normal limb, which at first is usually larger
than normal, but in which there is rapid regulation of size; 3)

Vv
HOM.DV HETDV

Fig. 135 Diagram to show determination of asymmetry of limb. The circles
represent the limb bud, the squares the surrounding part of the embryo. A,D,
P,V, the cardinal points of the embryo—anterior, dorsal, posterior, and ventral,
respectively. The heavy arrows represent the determining axes, i.e., the antero-
posterior axis of the bud and the dorsoventral axis of the surrounding parts;
UL, future ulnar border; D/, approximate direction of outgrowth. The smaller
arrows show the other axes of bud and surroundings, respectively, which are not
effective in determining the axes of the definitive limb.

SYMMETRY IN TRANSPLANTED LIMBS 87

a normal limb usually develops out of two like halves, i.e., two
dorsal, or two ventral halves, if properly oriented, when the
opposite half is entirely missing; 4) after inversion of the limb bud
the material that normally would have formed the radial half of
the limb gives rise to the ulnar half and vice versa, so that prac-
tically no part of the bud has the same fate that it would have
had if it had been left in place; 5) the inoculation of mesoderm
from the limb bud under the skin of the flank of another embryo
may result in the formation of a normal limb, although the inoc-
ulated tissue is badly disarranged by the operation. According
to all tests that have been applied, the embryonic limb rudiment
constitutes, therefore, an harmonic equipotential system, though,
as a whole, it is self-differentiating except for the determination
of its dorsoventral axis. The term ‘harmonic equipotential
system’ is employed here, as defined by Driesch, in the sense that
the potencies of all parts of the system are the same, the constitu-
ent cells being totipotent.*%* Its use does not imply that the
writer attaches to the existence of such systems the same signifi-
cance as Driesch, who considers them as constituting a proof of
the ‘autonomy of life.’ Even without this, however, and even
though the actual system may not reach the abstract perfection
demanded by its definition, 1t remains as a useful conception in
experimental morphogenesis. The existence of the equipotential
system necessitates, in fact, the assumption of some sort of mo-
lecular hypothesis for the representation of adult form in the
germ, and herein lies its importance in connection with the pres-
ent study.” In particular, we must look to the constitution of

% The concept ‘harmonic equipotential system’ is defined by Driesch (’05, p.
679) as follows: ‘‘Bekanntlich nenne ich harmonisch-iquipotentielle Systeme
solehe Formganze, bei denen eine Differenzierungs- oder Wachstumsgesamt-
leistung in ihren Einzelheiten jeweils einzelnen Elementen des Ausgangsganzen
zufaillt, derart, dass jedes Einzelne dieses Ganzen jedes Einzelne jener Leistung
vermag, alles Einzelne aber derart in Harmonie steht, dass die Leistung selbst
ein Ganzes ist.’? The bearing on the question of vitalism is discussed in various
papers, especially: ’99, p. 99; ’01, p. 170; ’08 b, p. 138.

% Child has expressed skepticism as to the very existence of equipotential
systems; for instance: ‘‘I think we may say that there is at present no valid
evidence for the belief that any living system which is undergoing regulation or
development in nature is at any given time an equipotential system”’ (’11, p.
306). Cf. also Child, ’08.

88 ROSS G. HARRISON

the elementary units of the limb bud, rather than to their ar-
rangement, for the representation of those relations of symmetry
that the experiments here described have revealed.** Inother
words, it is the intimate protoplasmic structure that underlies
symmetry.

In an equipotential system without axial differentiation, it is
most natural to assume that the elements themselves are iso-
tropic.°7 Axial differentiation would then result from the grad-
ual modification of these units by reaction with other elements
of the system or through external influences. These differentia-
tions with reference to directions in space may be referred arbi-
trarily to three axes crossing one another at right angles. They
are geometrically of four grades, according to the number of
axes along which polarization has taken place.

Taking the models used in stereochemistry to show the spatial
relations of the atom groups in certain carbon compounds, we
may represent the above four conditions of the elements of the
organism or system by four figures (fig. 136) in which the groups
that determine the axial relations are situated at the four angles
of a tetrahedron. At the center of each tetrahedron we might
by analogy assume a carbon atom linked to the four groups oecu-
pying the angles of the figure, though this is not necessary for the
present purpose. By hypothesis the groups at the angles are
supposed to be at first all alike (fig. 136, 1). If one of them
should be changed by some reaction, the structure of the molecule
would become polarized (fig. 136, 2), and if all the molecules
should assume approximately the same orientation, the system
which they constitute would show a similar polarity. If two of

98 The question whether relations of symmetry of the organism are to be based
upon symmetrical relations of the intimate protoplasmic structure is answered
in the affirmative by Driesch (’08 a, p. 144): ‘‘Wir miissen also alle Symmetrie
und auch alle Wirkungen, die von fusseren #aktoren ausgehen und sich auf
Symmetrie beziehen, auf priformierte, gerichtete Elemente des ‘Protoplasmas’
beziehen und kénnen in jenen Wirkungen nur richtende und umordnende
Geschehnisse sehen.

97 To avoid misunderstanding, it should be stated that when we speak of equi-
potentiality and isotropy, we do not lose sight of the fact that the system in its
entirety is heterogeneous.

SYMMETRY IN TRANSPLANTED LIMBS 89

the groups become differently modified, then the structure be-
comes bilaterally symmetrical (fig. 186, 3). And, finally, if three
become modified, so that all four are different, then the arrange-
ment becomes asymmetrical (fig. 136, 4 and 5) as in the ease of
optically active substances with an asymmetric carbon atom. In
the last phase there are two kindsof individuals, which are exactly
alike in every respect, except that they are the mirror images of

RIGHT

Fig. 136 Diagram to show hypothetical progressive differentiation of the
structural units. 1) condition of isotropy; 2) polarization with reference to one
axis; 3) bilateral symmetry (two axes differentiated); 4 and 5) condition of com-
plete asymmetry (three axes differentiated) giving right and left enantiomorphs.

one another—in other words, rights and lefts. This is expressed
in aggregate form in the right- and left-handed crystals corre-
sponding, respectively, to the dextro- and laevo-rotatory forms
of otherwise identical substances.

The experiments with limbs show that the bud at the time of
transplantation is in either the second or the third phase, probably
the former. There must be a differentiation along the antero-
posterior axis, because if this is reversed the limb shows it by
growing in a direction nearly opposite the normal. The medio-

90 ROSS G. HARRISON

lateral axis is probably not differentiated, though in the absence
of sufficient experiments in reversal of this axis, it is better to
make no definite assumption regarding the intimate structure
in relation to it. The dorsoventral axis is at most but slightly
differentiated, and if it is at all, then the differentiation is revers-
ible.°8 As already pointed out (p. 55), there is some ground
for the latter assumption, for it has been observed that after
transplantations in the heteropleural dorsoventral position (the
harmonic combination in which the dorsoventral but not the
anteroposterior axis is reversed) the adjustment of the tissues of
the limb bud is apparently not immediate, but involves a time
factor, probably not entirely accounted for by the effect of the
operation as such. Whatever the character of this dorsoventral
differentiation may be, itisnevertheless very slight in comparison
with the anteroposterior differentiation, which has become irre-
versible by the time the stage in question is reached.

If we could experiment over a wide enough range of stages,
it should be possible to determine the time limits of the above
phases of axial differentiation of the limb rudiment. At present,
however, there are no data bearing upon the question, for in the
earliest stages in which the transplantation of limbs has been
carried out (embryo with wide open medullary folds), as shown
by Detwiler (18), the limb bud follows in its development the
same rules as here formulated.

Lest the foregoing scheme seem too formal, it may be pointed
out that the model has been chosen to explain solely the relatively
simple characters of polarity and symmetry. Upon this as a
basis, further experimentation may yield facts from which the
mode of representation of more specific form features may be
determined. There is nothing in such a scheme inconsistent
with the fact that the cell itself is not a homogeneous system, for
the model is supposed to represent only that constituent of the
system which determines the adult character in question.

°8 This is perhaps odd in view of the facts brought out by Przibram (’10b),
showing that dorsoventral differentiation is very marked in the animal organiza-
tion, more so, for instance, than the anteroposterior differentiation.

SYMMETRY IN TRANSPLANTED LIMBS 91

The point which it is desired to emphasize is that in an organic
‘equipotential system’ there must be some intimate structural
basis for adult characters in the units that make up the embryonic
rudiment.*? It cannot be in the arrangement of these units; for
in that case marked disturbances of development would be pro-
duced by such operations as removing half of the rudiment, fusing
two buds together, combining two like halves, inverting the
dorsoventral axis, or inoculating masses of mesoderm cells from
the limb rudiment under the skin of the flank; and yet normal
development may follow any of these procedures.

These experiments yield, of course, no information concerning
the localization in the cell of the representatives of the adult form
characters in question. The system here dealt with is a pluri-
cellular one, but it is interesting to find that in the most thorough
and careful studies of polarity and symmetry in the egg, the
basis of these properties is found to be in the cytoplasm of the egg
cell. Lillie (06, ’09) shows that the polarity of the Chaetopterus
egg must be located in the ground-substance, because any amount
of shifting of visible granulations in the egg, such as yolk, oil
droplets, pigment, etc., has no effect on the polarity of the result-
ing embryo. With this conclusion the work of Morgan and his
collaborators (08 b, ’09, 710) on the centrifuged eggs of various
animals, more particularly Arbacia and Cumingia, is in substan-
tial accord. Conklin (716, ’17), in his study of Crepidula, con-
cludes that it is the spongioplasmic framework of the egg-cell that
determines its polarity, though he does not consider how this
quality is determined in relation to the intimate structure of the
spongioplasm. To the extent that Conklin places the seat of
polarity in the more viscid rather than in the more fluid constit-
uent of the cytoplasm, he takes issue with Lillie, but in the main,
there is agreement between these two investigators. Lillie, how-
ever, goes a step further when he says (’09, p. 77): ‘‘The existence
of polarity and bilaterality in an optically homogeneous medium,
and the persistence of both as to orientation under experimental
conditions that seriously modify the quantitative relations of
the oriented medium in different regions (as, for instance, when

93'Cf. Drieseh, 1. ¢c.

Q? ROSS G. HARRISON

the yolk granules are packed closely into the small cell of the
two-celled stage of Chaetopterus) seem to me to argue for a
molecular basis of the fundamental principle of vital organi-
zation.”’ Morgan, likewise, takes this view when he says (’09,
p. 114), “These considerations incline us to the view that there
exist in the molecular constitution of the egg the potential
factors of symmetry.”’ The scheme outlined above is in harmony
with this concept.

On the other hand, Child (18, ’15), as also Della Valle (713)
rejects all such hypotheses, basing the phenomena of axial differ-
entiation upon the occurrence of gradients, which, according to
Child, are primarily of a functional (metabolic) nature. It seems
to the present writer that such gradients may well be an expres-
sion of the polarity rather than its cause.

G. Reduplication and the problem of polarity and heteromorphosis

The reduplications which have been observed in the various
experiments have already been described sufficiently for the pres-
ent discussion (pp. 35, 45, 65, 73). The salient facts are: 1) that
the duplicate is the mirror image of the original limb; 2) that
more than one secondary member may arise by budding from the
same primary bud, in which case both of the former stand in some
relation of symmetry to the original; 3) that the secondary
appendages themselves may be doubled, forming a more or less
symmetrical pair. There are a few exceptional cases, where two
members of the same side stand in linear series, but probably
these have arisen only where the two rudiments are far enough
apart not to influence one another.!°°

100 Several cases are to be considered here. One (H. R. E. 10-) is a case in
which the anterior half of the limb bud was removed. Two limbs developed,
one clearly from the remaining posterior half and the other probably from the
anterior border of the wound (cf. Harrison ’18, p. 441). The operation was done
on the left side and both limbs were lefts, the posterior one being somewhat
defective. In another case, which had an early history similar to the above, the
posterior member was very defective and it was impossible to determine whether
it was aright oraleft. A third case (H. R. E. 20) is the one figured on page
79. This is not a case of regeneration, but one in which the anterior member
probably developed from the grafted half, while the posterior member may have

developed from the stationary ventral half. Both are attached to the same
shoulder-girdle, but there are two separate glenoid cavities.

SYMMETRY IN TRANSPLANTED LIMBS 93

Bateson, in his ‘‘ Materials for the Study of Variation” (’94),
has given an exhaustive review of the literature relating to super-
numerary parts, in which the limbs are fully considered. In this
treatise he has made a masterly analysis of the available material,
particularly with reference to the appendages of arthropods.
The phase of the problem which is especially relevant to the pres-
ent discussion is that concerning what Bateson calls minor sym-
metries, in which the supernumeraries are in some way symmetri-
cal with respect to themselves or to the normal appendages with
which they are associated. The other class of supernumeraries,
in which two identical appendages stand in simple succession to
one another are, according to Bateson, practically unknown,
and even those that have been described are considered by him to
be of somewhat doubtful nature, though many cases of simple
hyperdactyly would seem to belong in this category.

The symmetrical extra appendages fall into two groups: 1)
those in which there is a pair of extra members symmetrical with
themselves, arising from the normal appendage with which one
of the supernumeraries appears to have a definite relation of
symmetry, and, 2) those in which the single supernumerary is
symmetrical with respect to the normal. The former condition
Bateson considers to be the more usual, and, in fact, he accepts
the existence of the latter with a certain skepticism which seems
unnecessary.'" It is true that many cases that apparently fall
within the latter group may upon closer examination be found to
belong in the former, but the converse isalso true, as will be shown
below (p. 97). Bateson has devoted special attention to the
first group, and, on the basis of about one hundred and twenty
cases in insects and a considerable number in the Crustacea, he
has formulated the following rules,! showing the relation of the
supernumerary appendages to each other and to the original
member:

I. The long axes of the normal appendage and of the two extra
appendages are in one plane: of the two extra appendages one is there-

fore nearer to the axis of the normal appendage and the other is remoter
from it.

101 Op, cit. p. 5389 and 553. Consider, however, in this connection, the clear
case described by Bender (’06).
102 Op. cit. p. 479.

94 ROSS G. HARRISON

II. The nearer of the two extra appendages is in structure and posi-
tion formed as the image of the normal appendage in a plane mirror
placed between the normal appendage and the nearer one, at right
angles to the plane of the three axes; and the remoter appendage is
the image of the nearer in a plane mirror similarly placed between the
two extra appendages.

Transverse sections of the three appendages taken at homologous
points are thus images of each other in parallel mirrors.

In the vertebrates Bateson marshals a large amount of mate-
rial, of which about fifty cases are in amphibians.’ At the time
Bateson’s book was written, however, little or nothing was known
regarding the origin of supernumera y appendages in either the
arthropods or the vertebrates. Since then a large amount of
experimental evidence has accumulated to show that they may
be formed by superregeneration, especially by regeneration from
complex or irregular wound surfaces.!°% The evidence all cor-
roborates Bateson’s main gene. alization regarding the relation of
symmetry of supernumerary limbs, and there are practically no
exceptions.!”

The importance of double supernumeraries (Bruchdreifachbil-
dung, la doppia rigenerazione inversa, see p. 95) is emphasized
in the papers by Emmel (’07) and Della Valle (’13), and this con-
ception is given prominence in Przibram’s more general discus-
sion of the question (’09, p. 234).

103 See Bateson (pp. 554-5) for a discussion of the older literature.

104 In the amphibians the investigations of Barfurth (’94), Giard (’95), Tornier
(97, ’00, ’05, 706), Lissitzky (’10), Fritsch (11), Kurz (’12), Della Valle (’13),
and others have added much to our knowledge of the subject. In the crustaceans
Przibram (’02), Reed (’04), Zeleny (’05), and Emmel (’07) have reported experi-
ments which, though not so numerous, are none the less important. The more
recent literature is fully discussed in many of these papers, especially in those
of Lissitzky, Fritsch, and Della Valle, to which the reader is referred for details.

10 A remarkable exceptional case has recently been described by Dawson
(20) in a lobster, in which there is an extra pair of chelipeds attached to the
normal. The two extra chelae are mirror images of one another, but the one
nearer the primary claw is not mirrored from the latter, but is of the same side.
Furthermore, the primary claw is a ‘nipper,’ while the supernumeraries are both
of the ‘crusher’ type, so that the case proves to be likewise an exception to Prai-
bram’s rule (’11), according to which, in heterochelous forms, the extra appen-
dages are of the same type as the primary. The case described by Cole (’10),
also in a lobster, is an almost diagrammatic example of Bateson’s rule, if allow-
ance is made for the effects of torsion.

SYMMETRY IN TRANSPLANTED LIMBS 95

Likewise, the well-known experiments of Tornier (05) upon
‘tadpoles of Pelobates, in which the hind-limb bud was divided
in an early stage, some of the cases of Lissitzky (10), and Della
Valle’s case of reversed regeneration conform closely to Bate-
son’srules. Although the end results of the experiments of Tor-
nier and of Della Valle are analogous, there is, however a sharp
difference of opinion regarding the exact mode of origin of Tor-
nier’s double supernumerary hind legs, Tornier maintaining that
they both arise from the dorsal part of the pelvis, which was
split off by the operation, while Della Valle holds them to be
analogous to his own case.

Della Valle has laid particular stress upon the supposed identity
of change of asymmetry and reversal of polarity, and has sought
to make the various cases of superregeneration which have been
reported fit into his scheme of ‘doppia rigenerazione inversa.’
The case on which Della Valle bases his discussion of these ques-
tions is that of a newt (Triton) in which the left anterior limb was
fractured in the region of the brachium and cicatrization pre-
vented by tying a silk thread around the limb at that level.
Twenty days later the same limb was amputated a short distance
below the point of fracture. There régenerated three perfect
limbs, one from the distal end of the stump and two from the
region of the fracture. Of the latter, one was from the proximal
end of the small portion beyond the ligature and the other was
apparently a continuation of the stump proximal to the fracture.
The first and last of the three were left limbs, i.e., of the same side
as the original, while the one which regenerated in a distoproxi-
mal direction was reversed. The end result was a triple append-
age in which the three members were placed in accordance with
Bateson’s rule.

Della Valle seeks to make the eases of Tornier (’05) and Lis-
sitzky (10) conform to this scheme, and falls into line with
Przibram (’09) who had previously given a schematic represen-
tation of the same phenomena, which he termed ‘Bruchdreifach-
bilding.’ He also interprets in the same way the reduplications
obtained by Braus (’04, ’09) and myself (’07) in the transplanta-
tion of embryonic limb buds. He suggests that when the limb

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 1

96 ROSS G. HARRISON

bud is implanted in normal location, triplicate appendages could
be accounted for in the following way: one member derived —
directly from the grafted bud; one member, of mirror symmetry,
by an inverse regeneration from the base of the bud, and’ the
third member, of original asymmetry, from the wound surface
of the host. This view is, however, not borne out by the present
experiments. !°

When the whole of the evidence bearing on the question is
taken into consideration, one cannot but think that too much
weight has been placed by Bateson and his followers on the double
supernumerary. The other class of cases, where the single
supernumerary is symmetrical with the normal appendage
with which it is associated, while neither so numerous nor so
spectacular, is nevertheless of wide occurrence. Cases reported
by Tornier (’97), Przibram (’02), Reed (’04), Zeleny (’05), and
Megusar (07) show that truly double appendages in mirror
symmetry with respect to each other may be formed by constric-

106 Se dunque noi considerassimo uno di questi innesti praticati invece che in
una regione lontana (come p. es. nell’orbita), nell’immediata vicinanza della
regione donde fu tolto l’innesto, noi osserveremmo I|’uno presso dell’altro lo svil-
uppo oltre che dell’arto normale, anche dell’arto rigenerato dalla superficie di
sezione della regione prossimale del corpo, nonché dell’arto sviluppatosi dalla
superficie di sezione della regione periferica, identico all’arto che lo ha prodotto,
ma con simmetria speculare. La identita anche di questo fenomeno con la dop-
pla rigenerazione inversa dalle due superficie di una ferita risulta in questo modo
evidente. Della Valle: op. cit. p. 125.

There is opportunity to test this hypothesis by comparing the experiments in
which the wound-bed was cleaned with those where it was not. In the former,
regeneration from the host is precluded (p. 6), and triplicate limbs could only
arise by a second reduplication from the base of the graft; whereas in the latter,
regeneration from the host should occur in a large number of cases, if at all, and
thus yield a large proportion of triplicate appendages. An examination of the
results shows that this is not the case. In the first place, as shown in table 2,
the total number of reduplications in the series with cleaned wounds is fifty-three,
which is 56 per cent of the total number of positive experiments, while there are
but sixteen cases (33.3 per cent) in the group with non-cleaned wounds. The
disproportion is much greater when the number of triplicate appendages in each
group is compared. Out of a total of eighty-seven cases old enough to be deter-
mined, there are twenty-five triplicate limbs (28.7 per cent) in the clean-wound
experiments and only three in forty-eight cases (6.25 per cent) in the others.
It is quite clear, then, that leaving in the wound-bed cells that are capable of
giving rise to a new limb reduces greatly, instead of increasing, the chance of
formation of supernumerary limbs, so that Della Valle’s suggestion is untenable.

SYMMETRY IN TRANSPLANTED LIMBS 97

tion of a simple regenerating bud. This harmonizes with
Driesch’s (06) observations on double Echinus embryos.

In the present work, the reduplicated extremities are nearly
all found to be in minor symmetry, and many of those in which
three members are present, if seen only in the fully developed
condition, would appear to be cases of paired supernumeraries,
conforming, though with some aberration, to Bateson’s rules,
The individual histories show, however, that they are mostly
simple duplicities in which the supernumerary mirrors the orig-
inal, and this seems to be the case in Braus’s experiments, too.
Two reduplicating limbs often do develop, but usually each grows
as a bud from the original instead of the two arising as a pair in
themselves. Each of them mirrors the original limb, so that the
two supernumeraries are both of the same side. In other cases
the supernumeraries are themselves double, in which event there
is strict conformity to Bateson’s rule, but the former constitute
a large majority, and conformity there is only superficial, for the
original limb is the middle member and not one of the extremes.

In view of these facts, there is probably no very fundamental
difference between the two classes of reduplications, i.e., between
the double supernumeraries symmetrical with each other and the
single supernumerary symmetrical with the original; had Bate-
son had the developmental stages at his disposal, he himself
might not have drawn so sharp a distinction.

In accordance with the above, Bateson’s rules might be stated
in more general form, so as to include both simple duplicities and
symmetrical pairs, as follows:

1. The long axes of duplex or multiplex appendages lie in one
plane.

2. Two adjacent members form in structure and position the
image of each other, as reflected from a plane mirror bisecting
the angle between the respective axes and perpendicular to the
common plane of the two axes (figs. 3 and 4).

The present experiments show (tables 2, 3, 5, and 8) that,
excepting heterotopic grafts, it is in the disharmonic combina-
tions that reduplications are most frequent. What, now, is the
nature of the disturbance that causes the doubling of transplanted

98 ROSS G. HARRISON

limb buds and of regenerating limbs, which, when it occurs, is_
always combined with reversal of one member? ‘The first visible
sign of reduplication both in the embryonic limbs and in the re-
generating blastema is the presence of two growth centers for the
limb in place of one; each becomes an apex of growth, with a
resulting bifurcation of the appendage as a whole. The question
arises whether the doubling of the growth center is antecedent
to or resultant from the reversal of the asymmetry. From the
fact that mere mechanical division of a simple regenerating cen-
ter??? may bring about doubling, it would seem to be more prob-
able, if not certain, that. the existence of two growth centers
within spheres of mutual influence is the factor that produces
the reversal in one—the one that is less advantageously placed,
or in which differentiation is less advanced.

The problem before us thus resolves itself into two phases:
that of division or repetition of parts and that of symmetry.
This was clearly seen by Bateson, who has emphasized the funda-
mental nature of the power to divide.!°? No attempt will be
made here to analyze this phase of the question. The symmetric
relations of the repeated parts are, however, so definite and of
such general recurrence that they, too, are beyond question of a
fundamental nature.

The phenomenon of reversal of asymmetry has been treated
by many investigators as one with that of axial heteromorphosis,
and yet this is not strictly correct, for the reversal of asymmetry
may be brought about by the interchange of the poles of any one
of the three axes to which the object is referred, and not neces-
sarily the one along which regeneration and differentiation is
taking place. ‘This is true not only when regeneration occurs in
a proximodistal direction, as in the cases of Tornier, Zeleny, and
others, cited above, but also when it takes place distoproximally,
as shown in the two experiments reported by Kurz (’12).!°

107 Cf. Tornier (’97), Przibram (’02), Reed (’04), Zeleny (05), MeguSar (’07).

108 “This power to divide is a fundamental attribute of life and of that power
cell division is a special example.’’ (Problems of Genetics, p. 38.)

109 In somewhat similar experiments by Morgan (’08 a) only the bone, not the

soft parts, was reversed. Nothing is said regarding the exact character of the
limbs regenerated.

SYMMETRY IN TRANSPLANTED LIMBS . 99

The fundamental phenomenon, therefore, is not that a particular
axis is reversed, but that reversal occurs at all, and how it is
brought about.

Organic polarity, in general, has been based either on the sup-
posed polarization of the organic units themselves or upon a
supposed gradient of a functional (Child) or material (Morgan,
05) nature, running from one end of the organism to the other.
There is evidence for the occurrence of both factors, and what
seems most likely is that both are at play. Under certain cir-
cumstances they act in the same direction; under different condi-
tions one may antagonize and retard, or even overcome, the other,
as seems likely in heteromorphic regeneration where polarity is
reversed (earthworm, planarians, amphibian tail, ete.). Prai-
bram (713), who advocates a theory combining the two factors,
which he calls ‘Richtungspolaritit’ and ‘Schichtungspolaritit,’
respectively, nevertheless regards the reversal of polarity as due
to actual rotation of the cells. He (’06, 710 a) cites unpublished
work by Hadzi in support of this view, and adopts it in his dia-
grams (’09) illustrating the five fundamental varieties of opera-
tion leading to regeneration (Biotechnik).

The figures are not convincing, however, for just as much rota-
tion of the cells is shown at the end where polarity is not reversed
as at the other end where it is reversed (Przibram, ’09, pl. XV,
1h-3h), and in fact, as expressed in these diagrams, what turning
is shown is nothing more than a wound-healing process. Until it
is demonstrated that rotation of the cells as a whole takes place
solely in heteromorphic regeneration, it cannot be used to explain
reversal of polarity.

So long as the elementary units of the limb bud have one plane
of symmetry left, and the final asymmetry of the limb remains to
be determined by its relation to certain axes of the embryo,'!° it

110 Tn the case of asymmetric organisms, the elementary units, if representing
the form of the organism at all, must be postulated as asymmetric themselves.
In the case of paired organs, each asymmetric in itself, but symmetrical with
respect to its opposite, polarization on the transverse axis may be assumed as
due to the position of the parts with respect to the other two axes (ef. Przibram,

13, p. 38) and not as necessarily due to actual differentiation of the elements in
the transverse direction.

100 ROSS G. HARRISON

will of course be possible to account for its reversal by rotation of
the elements about the proper axis. As an alternative to the
rotation hypothesis, we might, however, consider reversal as due
to an interchange in position of two of the determining groups in
the elementary units (p. 89, fig. 186). In case of differentiation
on all three of the axes, 1.e., if the units themselves are asymmet-
ric, then reversal could take place only in the latter way, unless
it occurs altogether independently of the intimate structure.

There is an analogy for reversal of this kind in the change of
the asymmetry of organic molecules of known composition, as,
for instance, in the Walden inversion by means of successive sub-
stitutions, or in the conversion of dextrotartaric acid into racemic
acid, by which transformation half of the dextrorotating groups
are changed into the laevo form. Of course, these examples are
mere analogies.

Such questions have been touched upon by many of those who
have studied twins and double monsters, but, unfortunately,
the evidence both as to the cause and as to the occurrence of
reversal of asymmetry is conflicting. In the case of human dupli-
cate twins, it is certain that there is no situs inversus viscerum,
except very rarely, and apparently even in double monsters the
degree of fusion of the two individuals must be considerable for
the asymmetry of the internal organs—heart and alimentary
canal—to be reversed. On the other hand, it has been shown
by Wilder (’04) that in duplicate twins the friction-skin patterns
of the two mates may show mirror imaging, particularly those
on the index fingers. <A similar condition has been found by
Newman (16) in his study of variation of the scutes in armadillo
quadruplets, except that here the matter is further complicated
by the relation, in pairs, of the four individuals of a litter.™

11 “Now in the armadillo there are many definite evidences of a system of
symmetry common to all of the quadruplets, upon which has been superimposed
a secondary symmetry system between twins. This in turn is more or less com-
pletely obliterated later by a tertiary symmetry between the antimeric halves of
the single individuals. In some sets evident traces of the primary system of
symmetry persist as mirror-image relations between individuals of opposite
pairs, but it is more usual to find no trace of the primary system. The secondary
mirror-imaging between pairs is far more commonly in evidence, but is frequently

SYMMETRY IN TRANSPLANTED LIMBS 101

The evidence which Morrill (19) has collected from the study
of double monsters in fish embryos shows that situs inversus does
occur, but that it is the exception, not the rule, and that there is
no ‘‘very precise relation between the amount of separation of
the two components and the occurrence of mirror imaging.’’!!”
This would seem to oppose the view expressed above, that it is
the proximity of two growth centers that causes the reversal of
one. Still, in the absence of statistical data regarding the corre-
lation of the two events, it is unsafeto draw a definite conclusion.

In this connection the recent work of Spemann and Falkenburg
(19) is of the greatest importance. By extension and modifica-
tion of the earlier methods of the former, these investigators ob-
tained a large number of twins in Triton by constricting the eggs
in segmentation stages or in the early blastula. They found that
in a large number of the cases one individual of a pair (the right-
hand member in all cases but one) has complete situs inversus
viscerum. Spemann, after an admirable critical analysis of the
question, reaches the conclusion, that while some asymmetric
intimate structure must be postulated to account for the normal
asymmetry of the vertebrate body, there is no proof from these

obliterated by the tertiary mirror-imaging between antimeric halves of the same
individual, which latter is the prevailing symmetry system. . . . . In gen-
eral, mirror-imaging between individuals of opposite pairs is interpreted as an
evidence of the early system of symmetry present in the embryonic vesicle
before polyembryonic. budding began. When the primary buds are formed they
are the product of the antimeric halves of the undivided embryo and therefore
should have mirror-image relations, but a partial physiological isolation of the two
buds permits a certain degree of reorganization or regulation in the symmetry
relations, that tends partially to obliterate the original symmetry relations of
the undivided embryo. Similarly, when each primary bud subdivides to form
the secondary buds that are the primordia of the definitive individuals, a certain
residuum of the primary bud symmetry system is carried over, manifesting itself
in mirror-imaging between the twins derived from the same primary bud. But
here again a certain amount of regulation occurs so that a third system of sym-
metry, the bilateral symmetry of each individual, tends to obliterate former
systems of symmetry.’? Newmann: op. cit., pp. 200-201. j

12 Tannreuther (’19, p. 359) has recently figured a double chick embryo in
which the two individuals are united only by the posterior tip of the primitive
streak. Although trunk and head are entirely separate, the heart of the embryo
on the left shows situs inversus. No mention is made of this fact in the text.

102 ROSS G. HARRISON

experiments that it has been completely reversed by the opera-
tion. The evidence points, on the contrary, to the introduction
of another factor which is manifested also in the tendency of the
individuals to show defects on the inner side (i.e., the side turned
toward the partner). In the case of the left-hand member,
this acts in the same sense as the innate tendency to asymmetry
of the viscera, while in the case of the right-hand member, it
antagonizes, and in many cases, overcomes the latter.43 Spe-
mann suggests many different experiments to throw light upon
this question. The interesting point in Spemann’s discussion,
in connection with the present work, is that the necessity for as-
suming some sort of intimate structure to account for external
symmetry relations is recognized. In the case of the redupli-
cated limbs it is not clear whether the reversal of the secondary
bud is a result of direct action upon the individual processes of
development going on within it, or whether the influence of the
primary bud actually reverses the intimate structure. If it
should be found that the reaction takes place before the cells of
the limb blastema lose their totipotence, then the latter is
undoubtedly true. Otherwise it may be that the differentiation
in the limb blastema takes place directly under the influence of
the tissues of the environment.

H. Form regulation and function in transplanted limbs

That the limb bud after transplantation becomes adapted in a
measure to the new conditions is obvious from a casual con-
sideration of the experimental results. There are different types
of adaptation, however, for although regulation of form and fune-
tion go largely hand in hand, in some cases there may be very
complete functional regulation without form regulation,!™ and,
particularly in. the heterotopic grafts, form regulation without
function. By form regulation is meant, in the present connec-
tion, the process by which a limb bud that is implanted abnor-

113 Cf, also Pressler (711).
‘14 Wor instance, in Experiment I. E. 64, the single disharmonic limb func-
tioned very actively and effectively. :

SYMMETRY IN TRANSPLANTED LIMBS 103

mally in relation to the embryo as a whole becomes adjusted so
as to give rise to a limb in harmony with its new surroundings.
Perfect adaptation is attained, however, only in cases where there
is functional adaptation as well, 1.e., only when both form and
function are adjusted to the organism as a whole, and this occurs
only in the orthotopic position or in positions very close to it.
Functional regulation is here largely a question of innervation,
and, inasmuch as this has been made a subject of special investi-
gation by Detwiler (19, ’20), it will not be considered at present
further than is necessary in its relation to form regulation.
Considering only the orthotopic grafts, there is no a priori
ground for expecting any of the combinations to yield normally
functional limbs except the homopleural dorsodorsal group, in
which the axial relations of the transplanted bud remain normal
with respect to the cardinal points of the embryo. Nevertheless,
it has been found that inverted limb buds from the opposite side
of the body yield almost as high a proportion of normally formed
and normally functional limbs as the first-named group. This
phenomenon has been interpreted as due not so much to a sec-
ondary regulation as to the structure of the elementary units of
the limb bud, which are supposed to be symmetrical or reversibly
differentiated (p. 89) along their dorsoventral axis. With an
equal number of experiments in each of the four groups and with-
out any disturbing factors, the primary processes of regulation
should therefore lead, in accordance with the fundamental rules
of symmetry, just as often to adaptive (harmonic) as to non-
adaptive (disharmonic) end results, but not more often. This
proportion is, however, exceeded by nearly 20 per cent of the
expectancy, so that, taking the orthotopic experiments as a
whole, 59.6 per cent yielded normally functioning limbs in normal
posture, while 40.4, for the most part reduplications, were non-
adaptive or only imperfectly adaptive (table 7). The proportion
is almost the same in the three classes of experiments, whether
with whole buds after extirpation of the normal bud, with super-
posed buds, or with half-buds, though, as more fully discussed
below (p. 107), it is quite different in the heterotopic grafts.

104. ; ROSS G. HARRISON

TABLE 7
Showing the adaptive or non-adaptive character of the resulting limbs after
orthotopic transplantation

NON-ADAPTIVE (REDUPLI-
CATION OR SINGLE
DISHARMONIC LIMBS)

ADAPTIVE (SINGLE
HARMONIC LIMBS)

TYPE x NT : Es Bes
F EX IME sd =)
x OF EXPER 2 fs 38 be 38
Z it 68 = a oS =
=| oS ee =A oO e 38 SP es 3)
Sees | 2 el See eee
Sacer eo) gt Be lee ey es
Whole buds
ROMO, Hecke: cee ae: 9 9 0 CMC, Oil Co OU GkG
homsay eee Nee 38 On Pel ON 108 |e 2683 he2bmimes DSi
hetidal bose 157 Mee 31 0 5 a] LO eG 26 | 83.9
MEG Civ ic mokctow hth bio UKs) {US 0 15) OBR! ©. |) I WE 03)
LENG Gi wee ht Rear 1] HARA en ye OAT 245 a a SONAL 5) 5 55 | 58.5
Average of percentages. ... 59.0 41.0
Superposed buds
homme xe etc u eaeeent 5 5 0 | LOORO Os 270 0; 0.0
Inve) Cohygammee RS nee 5 0 1 1 2000/40 4 | 80.0
Hed o Aa eine see 5 0 1 ENO). 4b lh) 4 | 80.0
MG EC IV: zsh Sane nee knee 5| 5 0 5 100-0} 0 | 0 0} 0.0
Ly cri] ba ST noi abate 20 | 10 PN SUN ONO . Ne 8 | 40.0
Average of percentages. ... 60.0 46.0
Half-buds
home iGdmrtertaouee eee 8 8 0 S 10020) OF 0 0} 0.0
homo ivan apes oh ls eee 17 0 4 4 | 23.5} 13 0 13 | 76.5
ets dee eccs heehee: il7/ 2, 3 D294 aaa 123) 4056
| OVS ERIC boo ees GOREN AE at 22 19 0 LOR S654: see es By | Bo
"Pobale! aeee py ese She teh 64 |} 29 oN Bor 56.3: 3 28 | 43.8
Average of percentages.... 59.8 | 40.2
——Se Sy a SSS |
otal harmonica seeks ole Gone Ol 0} 61 | 93.9) O| 4 4! 6.1
Total disharmonic....... 113 PAN PPE RG || PASO rey 0 YANN! Hall)
Grand totals re eee LiSh “63. 2a e STARA seo Rasa) 4 Oi Zales
Average of percentages!..... 59.6 40.4

1 As given in the three main groups in the upper part of the table.

SYMMETRY IN TRANSPLANTED LIMBS 105

The secondary regulation of form may be brought about in
one of two ways: either by rotation of the limb as a whole dur-
ing development, whereby it is gradually brought into normal
posture, or by a process of reduplication, which is more compli-
cated. In the latter the original grafted limb bud gives rise
to a secondary or reduplicating bud of mirror symmetry, which
outstrips the former in development, reducing it to a spur or
even practically suppressing it. The process of reduplication
more often yields, however, actual double appendages, and there
is no hard and fast: line between the latter and the single limbs
with the original bud reduced to a spur. The figures given in
the table are therefore somewhat arbitrary in this respect.

Regulation by rotation has been noted only in the inverted
limb buds from the same side of the body (hom.dv), and it has
been further shown that this mode of adjustment probably occurs
when the limb bud at the time of operation is rotated anteriorly
along the dorsal semicircumference somewhat less than 180°
(p. 41). The mechanics of this process is not yet understood.
In the other disharmonic group single limbs are very rare (only
one case), unless reversal by reduplication and reduction occurs,
and no eases of rotation have been observed.

Regulation by reduplication and reduction, which is attended
by reversal of asymmetry, has been observed in five cases out of
thirty-one in the heteropleural dorsodorsal group, but no perfect
cases of single limbs so produced have been found in the other
disharmonic combination (hom.dv.). In the latter group, how-
ever, reduplication is frequent, and the reduplicating member is
often normally attached and assumes the same posture as the
normal limb, but it has never been found to begin its development
early enough to bring about the suppression of the original bud.

In a measure offsetting the regulative cases, there are four
others in which the same process, reduplication, has resulted not
in regulation, but rather in its prevention. In three cases of
inverted buds! (hom.dv.), where there was tendency to regulate
by rotation, this regulative process was rendered futile by second-
ary budding, and also in one case of a heteropleural dorso-

15 7, E. 86, 88, 90 and possibly two other cases. See p. 37.

106 ROSS G. HARRISON

ventral graft, the usual primary regulative result was vitiated by
reduplication. The same was true in three cases of half-buds
inverted on the opposite side of the body (het.dv.). Reduplica-
tion, therefore, is not in itself a regulatory process, but leads
merely to a condition where regulation may take place through
the reduction of the disharmonic member.

The question now arises, whether the experiments give any
ground for assuming that there is anything teleological in these
regulatory processes. While there is no proof that regulation by
rotation is not of this nature, it would be a mistake to draw any
conclusion regarding it until the details of the process are better
understood. With regard to regulation through reduplication
and reduction, it is clear that mechanical factors are sufficient
to account for the process. It is in the disharmonic combinations
that secondary regulation is necessary to produce normal results.
When this is not accomplished by rotation, reduplications almost
always arise, and there are only two cases of single disharmonic
limbs among the orthotopic grafts. The disharmonic relation,
which is brought about by the inversion of the anteroposterior
axis of the limb bud, is thus seen to be a factor of prime import-
ance in the production of reduplications. It is not the only
factor, however, for some reduplications have occurred in har-
monic combinations, due probably to other disturbances at the
time of operation. But quite aside from the latter, it would be
wholly unjustifiable to assume that there is any causal relation
between the possible utility of the process of reduplication for
regulatory purposes and its frequent, or even almost universal,
occurrence where it might lead to this result. What seems to be
the condition that brings about regulation after the reduplicating
bud has arisen is the chance placement of the latter in a position
corresponding to that of the normal single limb. This relation
is attended by a more advantageous situation with respect to
blood and nerve supply than that occupied by the original bud,
and in extreme cases this leads to the resorption or suppression
of the latter. It is only in such cases that complete regulation
takes place and these constitute but about 15 per cent of the
total number of reduplications.

SYMMETRY IN TRANSPLANTED LIMBS 107

In this connection it is important to compare the orthotopic
with the heterotopic operations in respect to the regulatory pro-
cesses Just considered. When the limb bud is implanted in ab-
normal location, functional adaptation does not occur at all, the
limbs rarely showing any motor function whatever, unless placed
close to the normal position of the limb.1"*

Now the records show (table 8) that the two harmonic combina-
tions in the heterotopic series produced ten single harmonic limbs
and twelve reduplications (44.4 and 55.6 per cent, respectively).

TABLE 8

Comparison of heterotopic and orthotopic transplantations with reference to the
relative number of single limbs and duplicities in the harmonic and
disharmonic combinations -

HETEROTOPIC ORTHOTOPIC
CHARACTER OF COMBINATION Single Reduplicated Single Reduplicated
Num- Per Num- Per Num- Per | Num- Per
ber cent ber | cent ber cent ber cent
HArMOWien yA... 0... 2 10! | 44.4 12 55.6 24 96.0 1 4.0
PDISWATIMORIC 96%... =: 2 +s 19 | 86.4 3 13.6 22 3.4 | 573 | 96.6

1 Excluding one anomalous case in which an error of record is probable.

? Excluding five cases which became normal single limbs by development of
the duplicate and resorption of the original bud, and ten cases which became
normal by rotation.

’ Including five cases in which the original member was resorbed and the single
normal limb arose from the reduplicating bud.

In many of the latter, however, the doubling was but slight, in-
volving only the digits. The two disharmonic combinations, on
the other hand, produced nineteen single limbs (86.4 per cent)
and three (13.6 per cent) reduplications. The results obtained
by Detwiler (718), using much younger limb buds from embryos
with open medullary folds, sustain the above results, as far as the
harmonic group is concerned. There are eight cases of single
limbs and ten reduplications. In the disharmonic group, how-

16 This has been subjected to a careful analysis by Detwiler (’19, ’20), who
has shown that the failure to function is not due to lack of peripheral innervation
so much as to the insufficient connections within the central nervous system.

108 ROSS G. HARRISON
ever, there are four normal limbs and three reduplications. If
combined with the present results, this would reduce the dispro-
portion somewhat, though it would still leave it very large.
There are no cases of complete regulation over to the harmonic
position either by rotation or reduplication. Moreover, redu-
plication in over half the harmonic grafts disturbs the possible
harmonic end result.

As pointed out above, the case is very different in the orthotopic
operations. Here the harmonic group produced twenty-four
out of twenty-five (96 per cent) single limbs, while the dishar-
monic group produced but twelve single limbs (17.4 per cent),
ten of which became harmonic by rotation and are therefore
omitted from the tabulation, and fifty-seven reduplications (82.6
per cent), in three of which, however, the primary bud became
harmonic by rotation and a possible adaptive result was pre-
vented by the process of reduplication. Leaving out of consid-
eration the cases in which: single normal limbs were produced by
rotation, there are but two cases left (3.4 per cent) in which the
disharmonie relation failed to be followed by reduplication.

In a word, with orthotopic grafts the almost completely
dominant factor in producing single limbs or reduplications is
the harmony or disharmony of the combination, whereas the
case is quite different in the heterotopic grafts, where the har-
monic group produced a slight preponderance of reduplications
and the disharmonic group a great preponderance of single limbs.

At first sight the above figures might be advanced in support
of some hypothesis of an end purpose in development, inasmuch
as reduplications are produced in overwhelmingly great number
only where they may be taken advantage of to produce an adap-
tive end result. This, however, would be a hasty conclusion to
draw. The orthotopic experiments may be explained as above
(p. 106). There the tendency to reduplicate is due first to the
disturbance of operation, which, being very slight, is almost al-
ways suppressed in the harmonic combinations by the advantage
the primary bud has in connecting normally with the surround-
ings; while in the disharmonic combinations the tendency to re-
duplicate is not only greatly increased by the reversal of the axis

SYMMETRY IN TRANSPLANTED LIMBS 109

of the bud, but the primary bud also gains a headway in develop-
ment that usually cannot be overcome even by the more advan-
tageous position of the reduplicating bud. In the heterotopic
operations the frequency of reduplication in the harmonic group
may be ascribed to the disturbance due to the operation, together
with lack of any special anatomical relations at the seat of implan-
tation that would overcome the tendency to bud by giving the
harmonic member a special advantage. All that remains to be
accounted for is, therefore, the small proportion of reduplica-
tions in the disharmonic group. This should be subjected to
further investigation. Standing alone, it can hardly be advanced
as an argument for a teleological theory of development.

GENERAL SUMMARY

1. The results given below are based upon the following experi-
ments with the fore-limb bud of the embryo of Amblystoma
punctatum :

a. Transplantation to the flank of another embryo posterior
to the normal position of the fore limb (heterotopic transplanta-
tion), the grafted buds being taken either from the same side of
the body (homopleural) or from the opposite (heteropleural), and
implanted with the dorsoventral axis upright (dorsodorsal) or
inverted (dorsoventral).

bh. Transplantation to the normal location of the fore limb
after extirpation of the original fore-limb rudiment (orthotopic
transplantation), with the same variations as in the heterotopic
group.

c. Superposition of one limb bud upon another after removal of
the ectodermal covering of the latter, also with the same varia-
tions as in the previous groups.

d. Transplantation of half of the circular dise constituting the
limb bud, after extirpation of one-half the rudiment. Sixteen
different combinations of this ‘experiment (all possible within the
limitations imposed) were tried (p. 70).

2. In the early stages of development in any position the trans-
planted buds give evidence of their constitution by growing out

110 ROSS G. HARRISON

(‘pointing’) in the direction of what was originally the posterior
pole of the anteroposterior axis. Thus, in two of the combina-
tions (homopleural dorsoventral and heteropleural dordosorsal)
they point anteriorly or dorsoanteriorly, and in the two others
(homopleural dorsodorsal and heteropleural dorsoventral) they
point posteriorly or dorsoposteriorly like the normal. In the
latter case the subsequent development is usually normal, bar-
ring reduplication; in the latter there is a tendency for the limb to
stick out to the side and to rotate more or less from the position
in which it would be found, were the position determined entirely
by the orientation of the bud itself.

3. The palmar surface of the limb tends to develop on the side
turned toward the body of the animal, and the ulnar border is
dorsal, although the rotation mentioned in the previous para-
graph tends to change these positions.

4, The above circumstances determine the asymmetry of the
limb as follows: when the dorsoventral axis is not inverted, the
original prospective asymmetry persists; when the axis is in-
verted, the asymmetry is reversed (rules 1 and 2, p.4). In more
general terms: the asymmetry of the limb is determined by two
factors, the polarization of the anteroposterior axis of the limb
bud and the orientation of the limb bud with respect to the dorso-
ventral polarization of its organic environment (figs. 2 and 135).

5. In two of the combinations (homopleural dorsodorsal and
heteropleural dorsoventral) the asymmetry of the limb which
develops corresponds to that of the side of the body on which it is
placed (harmonic); in the other two (homopleural dorsoventral
and heteropleural dorsodorsal) it corresponds to that of the oppo-
site side (disharmonic).

6. Duplex and multiplex limbs arise frequently from the trans-
planted buds. They are of all grades and kinds and occur in
different proportions in the several experiments. In the hetero-
topic grafts they are more frequent in the harmonic combinations,
while in the orthotopic position they are much more frequent in
the disharmonic combinations.

7. In nearly all cases one member of a pair or group can be
distinguished as the original (primary) and the other one or ones

SYMMETRY IN TRANSPLANTED LIMBS fn i

as buds. The reduplicating bud is in each case the mirror image
of the original, and, when the reduplicating bud is itself doubled,
then the one next to the original is the mirror image of the latter, *
while the one further away is mirrored, with respect to its mate,
approximately in accordance with Bateson’s rule.

8. Limbs placed in abnormal location, where the specific blood
and nerve supply is lacking, are frequently resorbed, and when
they do develop, are usually stiff and functionless, or at best
show imperfect function. The shoulder-girdle in such limbs is
reduced in size and the more outlying elements are lacking.

9. Limb buds placed in normal location (orthotopic) are rarely
resorbed and nearly always become functional.

10. Limb buds from the same side of the body normally ori- .
ented in orthotopic position develop normally with but slight
retardation.

11. When the limb bud from the same side is rotated 180° in
its normal location, the results vary considerably, and in the
majority of cases reduplications occur. The single limbs are of
two kinds, reversed and normal. The former develop in accord-
ance with rule 2 (p. 4), but only one case of this kind has been
observed, the others that conform to the rules being redupli-
cated (rule 3). In the other cases the normal position was
reached by the rotation of the limb as a whole about the shoulder-
joint. These cases are exceptions to the rules.

_12. Of the twenty-seven cases of reduplications in the above
group, the original bud grew anteriorly and was reversed. In the
three remaining cases the primary limb righted itself by rotation
and the reduplicating member was reversed.

13. Regulation by rotation usually takes place when at the
operation the limb bud has been rotated anteriorly over the °
dorsal semicircumference not quite 180°.

14. Limb buds from the opposite side of the body, with the
dorsoventral axis normally oriented, produced but one unre-
versed single limb, in accordance with rule 1 (p. 4), the rest
being reduplications (rule 3). In one-sixth of the latter the original
bud, which was disharmonic, was resorbed and remained as a
small nodule or spur on the reduplicating appendage. The dupli-

12 ROSS G. HARRISON

cate bud in these cases, having its asymmetry reversed and oc-
cupying the right position, became a normally functioning fore
limb, perfectly adjusted both functionally and structurally to
its organic environment.

15. Limb buds, taken from the opposite side of the body and
implanted with the dorsoventral axis inverted, so as to leave the
anteroposterior axis in normal relation, formed, with the excep-
tion of one reduplication, single limbs, all of which were reversed
These limbs were often considerably retarded in development,
but, as regards both function and form, they became perfectly
adjusted to their new surroundings (rule 2).

16. In the superposed grafts two limb buds are fused into one.
In the two harmonic combinations normal single limbs arise.
Though at first usually above normal in size, they soon become
regulated in this respect. In the disharmonic combinations du-
plex appendages were formed in a large majority of cases. One
case of .adjustment by rotation and one case of regulation by
reduction of one member of a pair were found.

17. In experiments with half buds there are sixteen combina-
tions possible with the restrictions imposed by the character of
the experiment. In addition to the two pairs of attributes of
operation common to all of the experiments (hom. or het., dd or
dv) there are three others: the bud may be halved vertically or
horizontally; the anterior or the dorsal half, or the posterior or
the ventral half may be transplanted, the other remaining intact;
two like halves or two unlike halves may be united. An analysis
of the results shows that no one of these qualities in itself deter-
mines the result, but that it is the harmonic or disharmonic
character of the combination that determines whether normal or
reduplicated appendages arise. Thus, allowing for differences
in the number of experiments in each class, 93.4 per cent of the
harmonic combinations produced normal limbs, while in the
disharmonie groups about that same proportion produced redu-
plications, of which, however, a considerable number were regu-
lated secondarily through resorption of the disharmonic member.

18. That the limb bud is an equipotential system is shown by
the fact that a normal limb may develop after the following oper-

SYMMETRY IN TRANSPLANTED LIMBS 113

ations, provided the combination is harmonic: 1) extirpation of
any half of the bud; 2) fusion of two whole buds; 3) combination
of two like halves, the other half being entirely missing; 4) inver-
sion of the limb bud; 5) inoculation of mesoderm cells from the
limb under the skin in some other region of the embryo.

19. Except for the circumstance that the dorsoventral differen-
tiation of the limb bud is a function of the orientation of the bud
with respect to its organic environment, the limb bud is a highly
specific self-differentiating system. Its definitive form must,
therefore, be represented in the organic elements (intimate struc-
ture) of the limb rudiment.

20. One quality of these elements is their polarization, as
shown by the definite relation to the direction of out-growth,
assumed by the anteroposterior axis of the limb bud. It is sug-
gested that the asymmetry of the limb rudiment and of other
similar systems may be gradually brought about by the change
in constitution of the structural elements in a manner similar to
the building up of asymmetric molecules in organic compounds.

21. Reduplications are produced as a result of that funda-
mental attribute of living matter, the power to divide (Bateson).
They are induced, in the case of the limb bud especially, by a
disharmonic relation between graft and host.

22. There is no fundamental distinction between double super-
numerary limbs constituting a symmetrical pair and the single
supernumerary symmetrical with the normal one with which it
is associated. Bateson’s rules may be stated in simplified form
in accordance with this conclusion (p. 97).

23. Exceptions to Bateson’s rule regarding symmetry rela-’
tions of supernumerary parts are very rare. Those found in the
present study, where two limbs of the same side occurred in linear
series, are probably due to the appendages having been far
enough apart not to influence one another in development, and
at the same time having been under the influence of the same
organic environment.

24. Review of the data on regeneration of supernumerary
appendages shows that the reversal of asymmetry in one of the
members of an enantiomorphic pair is not dependent upon the

114 ROSS G. HARRISON

reversal of direction of growth, regeneration, or differentiation.
The reversed member may grow and differentiate in the same
direction as the original, another axis than that on which growth
is taking place being the one that is reversed. Reversal may
thus occur without axial heteromorphosis and vice versa.

25. In any system, like that of the limb bud at the time of
transplantation, in which at least one axis is left undifferentiated,
rotation of the elements of which the system is made up might
account for reversal. The rotation of cells observed by Hadzi
and Przibram is, however, concerned primarily with wound
healing, and heret is no evidence that it is correlated with the
occurrence of axial heteromorphosis or reversal of asymmetry.

26. As an alternative to the hypothesis of rotation, we might
consider reversal as due to reversal of molecular asymmetry
according to analogy with the behavior of optically active com-
pounds.

27. There is an analogy between the production of enantio-
morphic limbs and the production of situs inversus viscerum, as
effected by Spemann. Either the reversal may be due to reversal
of the intimate structure, or it may take place in spite of the
intimate structure through the direct action of mechanical factors
on the individual parts of the differentiating system.

28. The transplanted limbs show both regulation of form and
functional adaptation. The two often go hand in hand, but not
necessarily, for some cases show regulation of form without
function, and others functional regulation without form regulation.

29. Functional regulation is largely a matter of innervation,
and it occurs only in orthotopic grafts or in those approximately
in that position (Detwiler).

30. Form regulation is either primary, as in the case of har-
monic combinations, or secondary, as in the disharmonic. In
the latter it takes place in one of two ways, either by rotation of
the developing limb or by means of reduplication and reduction
of the disharmonic member.

31. Form regulation by rotation has been observed to occur
only in orthotopic grafts; reduplications in disharmonic combi-
nations are more frequent in orthotopic than in heterotopic

SYMMETRY IN TRANSPLANTED LIMBS 115

transplantations. While these circumstances lead to an har-
monic end-result more frequently where there is functional
adaptation as well, this cannot be used as a cogent argument for
a teleological theory of development.

LIST OF REFERENCES

BarrurtH, Dierricu 1894 Die experimentelle Regeneration iiberschiissiger
Gliedmassenteile (Polydaktylie) bei den Amphibien. Arch. f. Entw.-
Mech., Bd. 1.

Bateson, W1LuIAM 1894 Materials for the study of variation. London.
1913 Problems of genetics. New Haven.

Benner, O. 1906 Zur Kenntnis der Hypermelie beim Frosch. Morpholog.
Jahrbuch, B. 35.

Bravs, H. 1904 Einige Ergebnisse der Transplantation von Organanlagen bei
Bombinatorlarven. Verh. d. Anatom. Ges. 18. Versammlung i. Jena.
1905 Experimentelle Beitrige zur Frage nach der Entwicklung peri-
pherer Nerven. Anatom. Anz., Bd., 26.
1906 Vordere Extremitét und Operculum bei Bombinatorlarven.
Morpholog. Jahrbuch, Bd. 35.
1909 Gliedmassenpfropfung und Grundfragen der Skeletbildung. I.
Die Skeletanlage vor Auftreten des Vorknorpels und ihre Beziehung
zu den spiteren Differenzierungen. Morpholog. Jahrbuch, Bd. 39.

Cuitp, C. M. 1908 Driesch’s harmonic equipotential systems in form-regula-
tion. Biol. Zentralbl., Bd. 28.
1911 Studies on the dynamics of morphogenesis and inheritance in
experimental reproduction. I. The axial gradient in Planaria doroto-
cephala as a limiting factor in regulation. Jour. Exp. Zodl., vol. 10.
1913 Studies, ete. VI. The nature of the axial gradients in Planaria
and their relation to anteroposterior dominance, polarity and sym-
metry. Arch. f. Entw. Mech., Bd. 37.
1915 Individuality in organisms. Chicago.

Coxe, L. J. 1910 Description of an abnormal lobster cheliped. Biol. Bull.,
vol. 18.

Conxuin, E.G. 1916 Effects of centrifugal force on the polarity of the egg of
Crepidula. Proc. Nat. Acad. Sci., vol. 2.
1917 Effects of centrifugal force on the structure and development
of the eggs of Crepidula. Jour. Exp. Zoél., vol. 22.

Dawson, A. B. 1920 An exception to Bateson’s rule of secondary symmetry.
Biol. Bull., vol. 38.

Devia VALLE, P. 1913 La doppia rigenerazione inversa nelle fratture delle
zampe di Triton. Bollettino della Societa di Naturalisti in Napoli,
vol. 25 (1911-12).

DetwiLer, S. R. 1918 Experiments on the development of the shoulder girdle
and the anterior limb of Amblystoma punctatum. Jour. Exp. Zodl.,
vol. 25.

116 ROSS G. HARRISON

DetwiterR, 8. R. 1919 The effect of transplanting limbs upon the formation
of nerve plexuses and the development of peripheral neurones. Proce.
Nat. Acad. Sci., vol. 5.

1920 Experiments on the transplantation of limbs in Amblystoma.
The formation of nerve plexuses and the function of the limbs. Jour.
Exp. Zodl., vol. 31.

Drirescu, Hans 1899 Die Lokalisation morphogenetischer Vorgiinge. Ein Be-

weis vitalistischen Geschehens. Arch. f. Entw.-Mech., Bd. 8.

1901 Die organischen Regulationen. Leipzig.

1902 Uber ein neues harmonisch-iquipotentielles System und iiber
solche Systeme iiberhaupt. Arch. f. Entw.-Mech., Bd. 14.

1905 Die Entwickelungsphysiologie von 1902-1905. Ergeb. d. Anat.
u. Entw., Bd. 14.

1906 Studien zur Entwicklungsphysiologie der Bilateralitit. Arch.
f. Entw.-Mech., Bd. 21.

1908 a Zur Theorie der organischen Symmetrie. Arch. f. Entw.-
Mech., Bd. 26.

1908 b The science and philosophy of the organism. London.

Emmet, V. E. 1907 Regenerated and abnormal appendages in the lobster.
Special paper no. 31. Reprinted from the Thirty-seventh Annual
Report of the Commissioner of Inland Fisheries of Rhode Island.

Fritscu, C. 1911 Experimentelle Studien iiber Regenerationsvorgiinge des
Gliedmassenskelets der Amphibien. Zool. Jahrb., Bd. 30.

Giarp, M. A. 1895 Polydactylie provoquée chez Pleurodeles Walthii Micha-
helles. Comptes Rendus de la Société de Biol., T. 10, Sér. 10.

Harrison, R. G. 1907 Experiments in transplanting limbs and their bearing
upon the problems of the development of nerves. Jour. Exp. Zodl.,
vol. 4.

1915 Experiments on the development of the limbs in Amphibia.
Proc. Nat. Acad. Sci., vol. 1.

1916 Onthe reversal of laterality in the limbs of Amblystoma embryos.
Anat. Rec., vol. 10.

1917 a Further experiments on the laterality of transplanted limbs.
Anat. Rec., vol. 11.

1917 b Transplantation of limbs. Proc. Nat. Acad. Sci., vol. 3.
1918 Experiments on the development of the fore limb of Amblystoma,
a self-differentiating equipotential system. Jour. Exp. Zodl., vol. 25.

Kurz, Oskar 1912 Die beinbildenden Potenzen entwickelter Tritonen. Arch.
f. Entw.-Mech., Bd. 34.

Litiiz, FRANK R. 1906 Observations and experiments concerning the elemen-
tary phenomena of embryonic development in Chaetopterus. Jour.
Exp. Zoél., vol. 3.

1909 Polarity and bilaterality of the annelid egg. Experiments with
centrifugal force. Biol. Bull., vol. 16.

Lissitzky, Eugen 1910 Durch experimentelle Eingriffe hervorgerufene iiber-
zihlige Extremititen bei Amphibien. Arch. f. mikr. Anat. u. Entw.,
eye lary

MerauSsar, Franz 1907 Die Regeneration der Coleopteren. Arch. f. Entw.-
Mech., Bd. 25.

SYMMETRY IN TRANSPLANTED LIMBS Ly.

Morean, T. H. 1905 ‘Polarity’ considered as a phenomenon of gradation of
materials. Jour. Exp. Zodl., vol. 2.

1908 a Experiments in grafting. The American Naturalist, vol. 42.

1908 b The location of embryo-forming regions in the egg. Science,
N.S., vol. 28.

1910 Cytological studies of centrifuged eggs. Jour. Exp. Zodl., vol. 9.

Morcan, T. H., anp Spoonsr, G. B. 1909 The polarity of the centrifuged egg.
Arch. f. Entw.-Mech., Bd. 28.

Morritu, C. V. 1919 Symmetry reversal and mirror imaging in monstrous
trout and a comparison with similar conditions in human double mon-
sters. Anat. Rec., vol. 16.

Newman, H. H. 1916 Heredity and organic symmetry in armadillo quadrup-
lets. II. Mode of inheritance of double scutes and a discussion of
organic symmetry. Biol. Bull., vol. 30.

PressLteR, Kurr 1911 Beobachtungen und Versuche iiber den normalen und
inversen Situs viscerum et cordis bei Anurenlarven. Arch. f. Entw.-
Mech., Bd. 32.

PrzipraM, Hans 1902 Experimentelle Studien tiber Regeneration. Arch. f.
Entw.-Mech., Bd. 13.

1906 Regeneration als allgemeine Erscheinung in den drei Reichen der
Natur. Vortrag. Versamm. Deutscher Naturf. u. Aerzte. Stuttgart.
Naturwiss. Rundschau, Bd. 21.

1909 Experimental-Zoologie, 2. Renegeration.

1910 a Die Verteilung organbildender Fiahigkeiten auf Ix6rperre
gionen. Verh. der k. k. zoologisch-botanischen Ges. in Wien.

1910 b Die Verteilung formbildender Fahigkeiten am Tierk6rper in
dorso-ventraler Richtung. Arch. f. Entw.-Mech., Bd. 30.

1911 Experiments on asymmetrical forms as affording a clue to the
problem of bilaterality. Jour. Exp. Zodél., vol. 10.

1913° Experimental-Zoologie, 4. Vitalitat, Kap. 3, Polaritat.

Reep, Marcaret A. 1904 The regeneration of the first leg of the crayfish.
Arch. f. Entw.-Mech., Bd. 18.

SpeMAnN, H. 1910 Die Entwicklung des invertierten Hérgriibchens zum Laby-
rinth. Arch. f. Entw.-Mech., Bd. 30.

SpmeMANN, Hans, UND FALKENBURG, HermanN 1919. Uber asymmetrische Ent-
wicklung und Situs inversus viscerum bei Zwillingen und Doppel-
bildungen. Arch. f. Entw.-Mech., Bd., 45.

Srreeter, G. L. 1907 Some factors in the development of the amphibian ear
vesicle and further experiments on equilibration. Jour. Exp. Zodl.,
vol. 4.

1914 Experimental evidence concerning the determination of posture
of the membranous labyrinth in amphibian embryos. Jour. Exp.
Zool., vol. 16.

TANNREUTHER, GreorGE W. 1919 Partial and complete duplicity in chick
embryos. Anat. Rec., vol. 16.

Torntpr, Gustav 1897 Uber experimentell erzeugte dreischwinzige Hidech-
sen und Doppelgliedmassen von Molechen. Zool. Anz., Bd. 20.

1900 Das Entstehen von Kifermissbildungen besonders Hyperanten-
nie und Hypermelie. Arch. f. Entw.-Mech., Bd. 9.

118 ROSS G. HARRISON

TornieR, Gustav 1905 An Knoblauchskréten experimentell entstandene
iiberzihlige Hintergliedmassen. Arch. f. Entw.-Mech., Bd. 20.
1906 Experimentelles und kritisches iiber tierische Regeneration.
Sitzungsberichte der Gesellschaft Naturforschender Freunde zu
Berlin.

Witprr, Harris HaAwrHorRNE 1904 Duplicate twins and double monsters.
Am. Jour. Anat., vol. 3.

ZELENY, CHARLES 1905 The regeneration of a double chela in the fiddler crab
(Gelasimus pugilator) in place of a normal single one. Biol. Bull.,
vol. 9. ;

SYMMETRY IN TRANSPLANTED LIMBS 119

APPENDIX
HISTORIES OF SELECTED INDIVIDUAL CASES
A. Heterotopic transplantations

1. Homopleural, dorsodorsal

Experiment Tr. HE. 148. May 4, 1915. Right limb bud to right
flank; orientation normal.

May 5. Healed except for two small oval areas.

May 8. Transplanted bud points caudally like normal limb.

May 12. Sketch (fig. 5).

May 15. Transplanted limb slightly bidigitate—not so long or as
far differentiated as normal.

May 21. Transplanted limb has grown considerably, normal pos-
ture; two digits well marked, third appearing.

May 26. Third and fourth digits show more plainly than in sketch
made on May 24 (fig. 6).

June 1. Limb a normal right (homopleural) limb in normal posture
(fig. C):

Experiment Tr. H. 182. April 13, 1916. Right limb ‘bud to right
flank; orientation normal.

April 14. Perfectly healed.

April 16. Tranplanted bud as large as normal, points caudally.

April 25. Not so large as normal, points dorsocaudally.

May 1. Much shorter than normal. Points caudally, arising from
a ridge-like prominence on side of body.

May 7. Two distinct digits and trace of third. A new bud is
erowing from base anterolaterally.

May 1 5. Original limb (the posterior one in fig. 9) has three distinct
digits and beginning of fourth, shows elbow bend. It is a normal
right (homopleural) in nearly normal posture, though it sticks out
stiffly to side. The two limbs are entirely separate from their origin.
The anterior member, which is the reduplicating one, shows only a
faint trace of the third digit on the anterior border.

Experiment Tr. EK. 210. May 10, 1916. Right limb bud to right
flank; orientation normal. :

May 11. Healed except for small area of uncovered yolk at caudal
border of wound.

May 15. Transplanted bud somewhat smaller than normal, points
in same direction. ;

May 18. Posture as of normal right limb. At base a rounded
prominence which points headward.

May 22. The prominence has become a reduplicating limb, showing
first trace of digits.

May 28. Both lmbs well developed. Original is longer and is a
right (homopleural) with three digits and faint trace of fourth. Redup-
licating member branches from other at elbow and is a mirror image of
latter, except that it is less advanced in development.

120 ROSS G. HARRISON

2. Homopleural, dorsoventral

Experiment Tr. E. 219. May 17, 1916. Right limb bud to right
flank, inverted.

May 18. Healing fair; a considerable area anterior to grafted bud
not covered with ectoderm.

May 21. Bud looks nearly normal, but points dorso-anteriorly.
See sketches made on May 22 (figs. 10 and 11).

May 25. Transplanted bud growing as rapidly as normal. The
two point toward each other (figs. 12 and 13). ge

May 29. Transplanted limb growing well; two distinct digits;
points dorso-anteriorly and is apparently a mirror image of normal
limb on same side (figs. 14 and 15).

June 5. Rapid growth has continued; limb reaches gills; third
digit well marked, fourth beginning to show, elbow bend is shght;
hand in an almost vertical plane, palm being anteromedial. Limb is an
undoubted left, i.e., reversed (figs. 16 and 17).

Experiment Tr. E. 139. April 28, 1915. Right limb bud to night
flank, inverted.

April 29. Healing perfect, transplanted tissue rather prominent.

May 5. Limb bud points slightly dorso-anteriorly.

May 8. Points distinctly dorso-anteriorly.

May 17. Limb points laterally at angle of a little less than 60° to
axis of body. Third digit developing.

May 26. Four digits present (fig. 18). Limb does not look atro-
phic, but gives no evidence of motility or sensitivity. Position as when
last observed. Specimen preserved.

Examination of serial sections shows that the anteromedial surface
of the hand is the palm. There can be no doubt, consequently, that
the limb is a normal left, having been reversed.

Experiment Tr. E. 140. April 28, 1915. Right limb bud to right
flank, inverted.

April 29. Healing of wound good; slight areas still uncovered by
ectoderm at ventral border of wound.

May 5. Transplanted tissue growing; but points anterodorsally.

May 11. Digitations plain; limb points as before.

May 17. Limb now points almost transversely, with ulnar digits
on anterior border.

May 26. Perfect limb with extensor surface of elbow and ulnar
digits dorso-anterior. No motility (fig. 19). Sections of preserved
specimen show that ventral surface is palm, and that there can be no
question about its being a left limb, 1.e., reversed.

Experiment Tr. E. 220. May 17, 1916. Right limb bud to right
flank inverted.

May 18. Wound perfectly healed.

May 21. Transplanted bud very prominent. Sticks out to side
and slightly anteriorly.

May 25. Transplanted limb points dorso-anteriorly and not much
to side. A small reduplicating bud points ventroanteriorly from near
base.

SYMMETRY IN TRANSPLANTED LIMBS 1A

May 29. Limb points as before. Two well-marked digits which
are slightly irregular. Probably no reduplication.

June 5. Limb sticks out to side almost horizontally; first two
digits short, third distinct, fourth barely indicated. Reduplicating
hand attached to radial border of the same limb and pointing ventrally
has two digits.

N. B. It is uncertain whether this member developed out of the
bud noted on May 25.

3. Heteropleural, dorsodorsal

Experiment Tr. E. 227. May 19,1916. Right limb bud to left side,
dorsodorsal.

May 20. Well healed.

May 23. Transplanted bud points dorso-anteriorly and is more
massive than normal.

May 26. Grafted limb about same length as normal; points ante-
riorly and laterally, scarcely dorsally (figs. 20 and 21 drawn one day
later).

June 2. Grafted limb now reaches to the axilla of the normal limb;
trace of third digit (fig. 22, made day before).

June 5. Limb a normal right (is not reversed), pointing antero-
laterally, crossing the normal limb on the medial surface (fig. 23).

Experiment Tr. E. 107. April 9, 1913. Right fore limb to left
side, dorsodorsal orientation. Diameter of transplanted bud 3 somites
(3 to 5).

April 10. Crescentic uncovered area of yolk ventral to bud, which
otherwise is well healed in.

April 14. Grafted bud shows indications of pointing headward.

April 21. Limb growing, pointing anteriorly exactly in reversed
position. No digitations yet.

April 28. Limb is clearly a right; third digit now plain, with slight
nodule to indicate fourth.

May 5. Possibly slight motility of grafted limb. It is now seen to
be a perfect right, sticking out sharply to side, upper arm is inclined
dorso-anteriorly, fore arm slightly posteriorly; palmar surface of hand
ventro-anterior (figs. 24 and 25).

Experiment Tr. E. 127. May 12, 1914. Right limb bud to left
side, dorsodorsal orientation; diameter of bud 3 somites.

May 13. Healing perfect.

May 16. Transplanted bud growing well with slight indication of
pointing anteriorly.

May 18. Limb points distinctly anteriorly.

May 22. Two digits show; a reduplicating bud is growing out from
base of graft.

June 1. The primary limb now shows fourth digit and is a perfect
right (fig.26,PR). The reduplicating member (DU) isa left, though not
quite so far along in its development. The primary limb now points
laterally and posteriorly with its ulnar border headward.

122 ROSS G. HARRISON

4. Heteropleural, dorsoventral

Experimental Tr. HE. 198. April 14, 1916. Right limb to left flank,
dorsoventral orientation; diameter of transplanted bud 3% somites.

April 15. Wound perfectly healed.

April 17. Transplanted bud, rather larger than normal, points
posteriorly.

April 21. Transplanted bud has large base, but free portion is
slender; points distinctly posteriorly.

April 28. Transplanted bud growing, though much smaller than
normal. Points dorsoposteriorly and has two digits.

May 1. Digits more distinct. Slight bending at elbow. Has posi-
tion of normal left limb.

May 7. Three digits and trace of fourth. A normal left limb in
almost normal posture (fig. 27).

Experiment Tr. 167. February 24, 1916. Right limb bud to left
flank, dorsoventral orientation.

February 25. Wound completely healed. Transplanted bud prom-
inent.

Feb. 28. Transplanted bud a large blunt prominence which points
posteriorly.

March 8. Transplanted bud is very small; may be resorbed.

March 7. Grafted limb much smaller than normal, but growing.

March 11. Considerable growth, but still smaller than normal.
Slight indication of elbow bend and digits.

March 15. Digits marked. Limb looks like a normal left reversed,
though it sticks out to side more than the normal limb.

March 24, Third digit now plain. Arm points to side. Hand
transverse and vertical with respect to body.

March 29. Preserved. Transplanted limb a normal left in nearly
normal posture. Fourth digit indicated. Upper arm still sticks out
more to side than normal, but otherwise little difference except in size.

Experiment Tr. E. 129. May 18, 1914. Right limb to left side,
dorsoventral orientation. Transplanted dise 3 somites in diameter.

May 14. Wound perfectly healed.

May 18. Transplanted bud has grown considerably, points pos-
terlorly.

May 25. Grafted limb points laterally and about 30° caudally, is
bidigitate. Hand and perhaps forearm are being reduplicated on
anterolateral border.

June 6. Specimen preserved. Imperfectly symmetried double
hand, mirrored from radial plane. Arm stretches out to side with
elbow bending ventrally. The posterior (original) hand is a left
(reversed), as is the arm as a whole; it has four digits normally placed;
the anterior (reduplicating) hand is less perfect with only three digits,
one long one in the middle and a short one on each side, the first digit
being imperfect.

Experiment Tr. E. 168. February 24, 1916. Right limb bud to
left flank, dorsoventral orientation.

SYMMETRY IN TRANSPLANTED LIMBS 123

February 25. Well healed. Very small uncovered area ventral to
bud.

February 28. Transplanted bud larger than normal and a little
further ventral.

March 3. Bud is growing ventroposteriorly.

March 7. Bud long and slender, growing posteriorly almost on
ventral surface of body; probably defective.

March 11. Same.

March 15. Original limb has not changed essentially. A second
outgrowth of considerable length arises some distance dorsal to former
and projects laterally. . Judging by its size, 1t must have been present
and overiooked at last observation.

Mareli 20. Original limb a Jong bent appendage tapering gradually
to a point. New bud has made considerable growth and is bidigitate,
the digits spreading more than normally. Indication of elbow bend.

March 28. Secondary appendage has developed a third digit, which,
with elbow bend, shows that it is-probably a left (reversed).

April 4.. Specimen preserved (figs. 83 and 34). Transverse sections
show that limb is undoubtedly a left, since it is the dorsoposterior
surface of the hand that is the palm—a very unusual posture. The
two limbs have separate girdles which articulate with one another, but
are not fused.

Experiment Tr. E. 217. May 17, 1916. Right limb bud to left
flank, dorsoventral.

May 18. Perfectly healed.

May 21. Grafted bud almost exactly like normal, points dorso-
posteriorly.

May 29. Limb short, sticks out more sharply to side. Digitations
will probably be abnormal.

May 31. Possibly symmetrical reduplication of digits (fig. 31).

June 5. Uncertainty about reduplication. One distinct digit
(third) on posterior (ulnar) border and also a slight nodule (fourth
digit). Prominence on opposite border may not be digits (fig. 32).

June 8. One distinct digit and trace of another formed out of hump
on radial side so that hand is now practically symmetrical.

_ Limb amputated; afterward regenerated; form much like original.

B. Orthotopic transplantations
7. Homopleural, dorsodorsal

Experiment N. E. 3. May 19, 1918. Right limb bud to right side
—normal orientation. Pronephros intact.

May 21. Perfectly healed; limb bud normal.

May 25. Both limbs same except that transplanted one is slightly
shorter.

May 28. No difference between the two limbs.

June 7. Same. Specimen preserved.

124 ROSS G. HARRISON

8. Homopleural, dorsoventral

Experiment I. E. 64. May 13, 1916. Left limb bud to left side,
inverted. Pronephros removed.

May 15. Small uncovered area ventral to bud, which is larger than
normal and slightly posterior to it.

May 18. Limb bud points dorso-anteriorly (fig. 35).

May 24. Transplanted lmb points anteriorly into gills. Slght
trace of digits; no reduplicating bud (fig. 36, drawn one day earlier).
May 28. Limb points anterolaterally below gills.

May 29. Limb is undoubtedly a right (reversed) and is perfectly
normal except as to posture. Ulnar digits appearing on the antero-
dorsal border of hand (figs. 37 and 38).

June 1. Transplanted limb sticks out more to side than bofone and
ean be brought back further toward normal position. Motility was
first observed two days ago.

June 4. Larva has grown rapidly. Transplanted limb functions
well; elbow bends toward tail instead of head. Limb a perfect right
(figs. 39 to 41, drawn one day later).

Experiment I. E.60. May 11,1916. Left limb to left side, inverted.
Pronephros removed.

May 12. Well healed, normal-looking bud.

May 15. Grafted bud points dorso-anteriorly.

May 16. Sketches (figs. 42 and 43).

May 18. Beginning of bud like outgrowth (reduplication) pos-
teriorly (ig. 44, DU).

May 2 Main limb (PR) points more sharply to side and rather
more ee, than normal. Posterior (reduplicating) bud (DU) has
grown considerably (figs. 45 and 46, drawn May 238).

May 29. Specimen preserved. Limb double from lower part of
fore arm. The primary member, which is anterolateral, is a right
(reversed), while the other, produced from the posterior bud, is a left
(reversed back). Former has short first digit and well-marked third.
On the reduplicating hand a trace of third digit present (figs. 47 and
48). Mirror plane is radiodorsal.

Frontal sections show the radii of the two arms are fused at the
proximal end, the two ulnae being entirely separate. The humerus is
single, i.e., fused throughout and is much more massive than normal.
Glenoid cavity is also much larger, but normally oriented. Barring
slight abnormalities, the shoulder-girdle seems to be a normal left
(not reversed).

Experiment I. E.63. May 12,1916. Left limb to left side, inverted.
Pronephros removed.

May 13. Fairly well healed.

May 15. Transplanted bud smaller than normal; points dorsally,
sharply laterally, and slightly anteriorly (figs. 49 and 50, drawn May 16).

May 19. Points dorso-anteriorly. Slight indication of posterior
reduplicating bud at base (figs. 51 and 52).

' SYMMETRY IN TRANSPLANTED LIMBS 1235

May 24. Main limb points much more dorsally and laterally than
normal (figs. 53 and 54, drawn May 22); no digits. Two reduplicating
buds; posterior one (P.DU) points as in I. E. 60 (figs. 44 and 45);
anterior one (A.DU) smaller and spur-like.

May 29. Specimen preserved (fig. 55). Main limb is a right
(reversed) with two long digits and third digit on anterodorsal (ulnar)
border. The posterior reduplicating member with three digits is free
and is in approximately normal position for left limb (reversed back).
It is mirrored in a radiodorsal plane. The anterior hand is bidigitate,
and is mirrored in an ulnopalmar plane. Arm and fore arm of com-
plex is short and thick.

Experiment I. E. 49. May 14, 1915. Left limb bud to left side of
same individual turned 180°. Pronephros removed.

May 15. Fairly well healed; irregular band of uncovered yolk at

posterior border of wound.

May 21. Limb bud developing well, points dorsally with very slight

anterior inclination.

May 25. Limb points dorsally (fig. 56); two digitations—hand flat-

tened in vertical plane, making angle of 45° with body axis.

May 29. Limb points dorsoposteriorly, as if approaching normal

position.

May 30. Limb has rotated still further toward normal position.
June 3. Limb in nearly normal position (fig. 57, drawn June 4).
June 11. Limb tends to le partially adducted with dorsal surface

of manus on bottom.

June 17. Limb now perfectly normal in every respect—form, size,
posture, motility (fig. 58, drawn June 21, and fig. 59, drawn from
specimen preserved thirty-nine days after operation).

Experiment I. E. 55. May 23, 1915. Left limb bud removed and
replaced inverted. Pronephros removed.

May 24. Very well healed. Only a very narrow uncovered strip at
posterior border.

May 27. Limb bud smaller and more pointed than normal.

May 30. Jamb bud increasing, but not so large as normal. Sticks
out more to side.

June 2. Limb bud points dorsoposteriorly, ventral part of bud
prominent.

June 7. Limb has grown considerably, points more posteriorly than
before. Two digits; manus flattened in transverse plane.

June 11. Posture of limb still abnormal. It points dorsopostero-
laterally at about 45° to median and horizontal planes. Trace of third
digit.

June 17. Limb normal in every respect—form, size, posture, motility.

Experiment I. E. 85. April 10, 1917. Left limb bud to left side,
rotated 180°+ (fig. 60). Pronephros removed; pronephros grafted
along with limb bud.

April 11. Well healed.

April 14. Grafted bud about as distinct as normal, points dorsally
and slightly anteriorly.

126 ROSS G. HARRISON

April 16. Bud points dorso-anteriorly ca. 60° to horizontal; attach-
ment slightly posterior to normal. Distinct pronephric swelling ven-
tral to limb.

April 19. Original bud points as before. Posterior reduplicating
bud in approximately normal position. Anteriorly, at base of original
bud, is a third bud pointing anterolaterally.

April 24. Middle (primary) member points dorsally and has two
long digits; anterior member is attached along anterior border of
middle; posterior member, in normal posture shows beginning of digits.

May 7. Specimen preserved (fig. 61). Posterior member, an
essentially norma! left limb from its origin above elbow down, has four
digits: it is mirrored from the middle member in a radiodorsal plane.
Anterior and middle limbs fused except for distal portion of manus;
mirror plane, palmar. The middle member has two long digits and
nodules on ulnar border, and is probably a right (reversed). Anterior
member has three long digits and ulnar nodule, the third being partly
fused with the second of the middle hand. Anterior hand possibly
hyperdactylous.

Experiment I. E. 86. April 10, 1917. Left limb bud to left side,
rotated 180°— (fig. 60). Pronephros removed; pronephros transplanted
with limb tissue. .

April 11. Well healed.

April 14. Limb bud prominent, placed a little further posteriorly
than normal; points slightly dorsally.

April 16. Transplanted bud points dorsoposteriorly, but more
dorsally and slightly more laterally than normal. Pronephrie swelling
ventroanterior to bud.

April 19. Limb points more dorsally than normal. Small nodule
at base is part of pronephros.

April 24. Limb in normal posture not quite so long as’ normal.
Third digit beginning. Reduplicating digit comes off palmar side
between first and second digits.

April 29. Second and third digits reduplicated (palmar) ; limb other-
wise very nearly normal. Motility apparently not so good as normal.

May 5. Limb clearly a left (not reversed); arm a little shorter and
thicker. Motility now better.

May 7. Specimen preserved (fig. 62).

9. Heteropleural, dorsodorsal

Experiment R. FE. 87. May 19, 1915. Right limb to left side, dorso-
dorsal orientation. Pronephros left intact.

May 21. Small round uncovered area posterior to bud.

May 25. Transplanted bud points anteriorly and laterally (ea. 45°).
and is nearly as large as normal (fig. 63, drawn May 26).

May 28. Limb points as before and is growing into gills, one of
which is caught in notch between limb and neck and is bent ventrally.
Limb itself bent slightly at tip. Digitations faintly indicated.

SYMMETRY IN TRANSPLANTED LIMBS 127

June 2. Limb thinner than normal and somewhat limp. Bends
backward at tip. Two digits only.

June 4. Specimen preserved (fig. 64). The arm sticks out hori-
zontally to side; elbow bends posteriorly; two digits in same hori-
zontal plane with very faint indication of ulnar digits. Limb clearly
a right (not reversed), though it is distinctly smaller than normal and
is otherwise defective.

Experiment R. E. 70. May 12, 1915. Operation: right limb bud to
left side, dorsodorsal orientation. Pronephros left in.

May 138. Still a crescentic area of uncovered yolk at ventral border
of bud.

May 15. Transplanted bud more prominent than normal; indica-
tions of pointing anteriorly (fig. 67, drawn May 17, at which time the
posterior bud could just be made out).

May 21. Points anterolaterally. Attached to base is an almost
equally large reduplicating bud, pointing posteriorly and normally
located (fig. 68, drawn May 22).

May 27. Both buds growing. The original (anterior) one is still
a little further advanced than the other. Figure 69, drawn May 28,
shows beginning of digits.

June 7. The original limb is still the larger one, the reduplicated
one has two digits and trace of third. The complex has some motility.

June 11. Specimen preserved (fig. 70). The arm is double from the
elbow down; the posterior or reduplicating member is not so well
developed as the other. The anterior member is a right (not reversed)
while the other is mirrored in a radiodorsal plane and occupies a posi-
tion approximately normal for the left limb.

Experiment R. E. 133. May 25, 1916. Right limb bud to left side
dorsodorsal orientation. Pronephros removed.

May 26. Fairly well healed. Small uncovered area ventral to bud.

May 29. Transplanted bud well developed, points dorsoanteriorly
about 45°. It is in proper position and about same size as normal.

June 2. Limb points dorso-anteriorly, but more sharply dorsally,
extending into notch behind last gill. Two digits show distinctly as
in normal limb. Reduplicating bud from posterior border of base has
position of normal left.

June 5. Original limb (fig. 75, PR) Battie dorsolaterally toward
gills and looks nearly normal. Posterior (reduplicating) bud (P.DU)
is rather short and arises further tailward than normal limb; digitations
beginning. A third small imb (A.DU) anterior to the original, pro-
jects dorsolaterally, parallel to it.

June 7. Original limb a normal right (not reversed) having two
long digits with trace of third and fourth. The posterior bud has
grown considerably and is in normal position. Anterior bud shows
traces of digits.

June 12. All three limbs have grown and differentiated. Posterior
has nodule for third digit and is a normal left (reversed). It is mir-
rored in a radiodorsal plane while the anterior, limb, which is more

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 1

128 ROSS G. HARRISON

nearly parallel to the original, is mirrored in an ulnopalmar plane
(fig. 76, drawn June 13).

Experiment R. FE. 134. May 25, 1916. Right limb bud to left side,
dorsedorsal. Pronephros removed.

May 26. Perfectly healed.

May 29. Transplanted bud, normal size and position, points antero-
dorsally ca. 45°.

June 2. Limb points dorsally and somewhat laterally. Reduplicat-
ing bud on anterior border near base also points dorsally and slightly
anteriorly. ;

June 5. Limb points anteriorly into gills, is short and massive.
Three distinct digits, irregular.

June 12. Specimen preserved (fig. 77). Arm short and_ thick,
internal reduplication, certainly from elbow down, mirrored in ulno-
palmar plane. The two hands nearly separate; each has two long
digits and nodular third. The main limb is a right (not reversed) and
the other a left. No posterior reduplication in this case.

Experiment R. E.69. May 12,1915. Right limb to left side, dorso-
dorsal orientation. Pronephros left in.

May 13. Wound well healed. Only a small area at ventroposterior
border of wound uncovered.

May 15. Transplanted and normal budsa eqully prominent; former
points anteriorly.

May 21. Transplanted limb double. Smaller (originally main)
part (fig. 86, PR) points anteriorly and is borne upon the posterior
portion, DU, which is an almost normal looking limb, though smaller
than the normal on the opposite side.

May 24. The posterior bud has grown materially and is normal
looking with two digits. The anterior bud is relatively much less
developed (fig. 87).

May 27.- The original bud is reduced to a small spur (PR) onthe
lateral surface of the reduplicating bud, which is almost as large as the
normal limb on the opposite side (fig. 88, drawn one day later).

June 3. Transplanted limb perfectly normal. Spur much reduced.

June 7. Spur reduced to a nodule (fig. 89, Pf).

10. Heteropleural, dorsoventral

Experiment R. E. 80. May 18, 1915. Right limb bud to left side,
dorsoventral orientation. Pronephros removed.

May 19. Crescentic band ventroposterior to bud still uncovered.

May 21. Limb bud almost normal in size points posteriorly.

May 24. Limb bud points posteriorly, but more sharply laterally
than normal (fig. 98, TR).

May 30. Limb points dorsoposteriorly angle of 45°.

June 2. Limb stiff and projects rather more to side than normal;
has two digits and slight indication of third on dorsal border (fig. 94,
drawn June 1).

SYMMETRY IN TRANSPLANTED LIMBS 129

June 11. Limb mobile, in normal posture. Brachium shorter than
normal. Third digit distinct, fourth indicated.

June 21. Four digits all plain. Upper arm still short (fig. 95 TR).

Experiment R. E. 107. May 10, 1916. Right limb bud to left side,
dorsoventral orientation. Pronephros out.

May 11. Rather large area of yolk, posterior to graft, still uncovered.

May 15. Transplanted bud points a little more dorsally and more
sharply to side than normal.

May 19. Grafted limb (TR) about size of normal, though more
pointed and projecting more dorsally and laterally (fig. 96, drawn
May 18); divided at end (digits or reduplication?).

May 24. Limb on side of operation still projects a little to side.
Normal limb seems to move and transplanted one still stiff (figs. 97 and
98, drawn May 23).

May 28. Motility of operated limb still seems deficient. Third
digit distinct (fig. 99, drawn May 29).

June 2. Motility better. Specimen preserved.

Experiment R. FE. 116. May 16, 1916. Right limb bud to left side,
dorsoventral orientation; pronephros removed.

May 17. Perfectly healed.

May 20. Transplanted bud rather small, points slightly dorso-
posteriorly.

May 24. Transplanted bud slightly smaller than normal and
points a little more to side. Figure 100, drawn on May 23, shows
searcely any difference between the two buds.

May 28. Transplanted limb (7) practically exact counterpart of
normal (figs. 101 and 102, drawn May 29).

May 31. Transplanted limb has third and trace of fourth digit.
Function normal.

Experiment R. E. 93. May 19, 1915. Right limb bud to left side,
dorsoventral orientation; pronephros mostly removed.

May 21. Well healed; bud smooth and mound-lke.

May 25. Transplanted bud points dorsoposteriorly with a second-
ary bud at anterior border of base.

May 28. Posterior bud (PR) is normal in position and orientation,
and is even a little longer than the normal limb on the opposite side.
The other (anterior) nodule (DU) is growing into a limb and also
points posteriorly (figs. 103 and 104).

May 30. Posterior bud growing more rapidly than other; digita-
tions plain. Anterior bud sticks out more to side, shows trace of
digitations.

June 4. Two hands, each with two digits; some excrescences. Arm
a little shorter than normal. The posterior bud is (probably) a left,
and its position is approximately normal. The other is mirrored from
its dorsal surface. Both hands have faint trace of third digit. Excres-
cences which seem to involve mainly the skin make an exact interpre-
tation of this case difficult (fig. 105).

130 ROSS G. HARRISON

C. Superposed transplantations
13. Homopleural, dorsodorsal

Experiment S. E. 3. April 20, 1916. Right limb bud to right side,
normal orientation; mesoderm of host torn in posterior part of wound,
otherwise intact.

April 21. Well healed; limb bud considerably larger than normal.

April 24. On operated side imb bud larger than normal.

April 26. Operated limb slightly larger, otherwise no difference.

May 2. Operated limb still slightly larger.

May 8. The two limbs exactly alike (fig. 106).

14. Homopleural, dorsoventral

Experiment S. HE. 18. April 13, 1917. Right limb bud to right side,
inverted; mesoderm all left in.

April 14. Perfectly healed.

April 17. Operated limb bud sticks out more to the side, and is
more massive than normal.

April 21. Operated limb double; posterior component more mas-
sive, occupies normal position; other is attached to radial border and
points dorsoposteriorly; neither has digits.

April 25. Reduplication of hand and part of fore arm; one com-
ponent is in approximately normal position, the other sticks out to
side.

April 30. Limb has grown; relations essentially the same as before.

May 1. Specimen preserved. The main limb projects horizontally
to side and is inclined about 30° to the transverse plane; palm is ventral,
ulnar border posterior. The reduplicating member with two digits is
attached to the radial border of the wrist, pointing anteriorly snd
laterally (fig. -107).

Experiment S. FE. 10. May 26, 1916. Right limb bud to right side
inverted; mesoderm of host torn along posterior border of wound and a
little tissue lost.

May 27. Perfectly healed.

May 30. Limb on operated side points a little more dorsally than
normal. Ventro-anterior to main bud another fairly prominent mass
of tissue (like S. E. 9, in which an extra digit developed).

June 4. Limb on ‘operated side considerably smaller than normal,
and still points more dorsally; digits just beginning.

June 12. Operated limb still Tess developed than the unoperated;
third and fourth digits beginning to show.

June 15. Practically no difference between the two limbs; the
operated one is possibly a little thicker.

SYMMETRY IN TRANSPLANTED LIMBS 131

15. Heteropleural, dorsodorsal

Experiment S. HE. 12. May 26, 1916. Left limb bud to right side,
dorsodorsal orientation; mesoderm of host not injured at all.

May 27. Perfectly healed.

May 30. Limb on operated side (7) more massive than normal
and points dorsally; on posterior border at base there is a small rounded
prominence or bud (figs. 110 and 111, HOM, drawn May 31).

June 4. Main limb (HET) is spindling and points laterally, though
inclined slightly dorsally and posteriorly; there are three imperfectly
marked digits at the tip. From the bud at the posterior border there
has developed a second limb (HOM) which has normal posture; it is
of considerable size and shows beginning of digitations (fig. 112, drawn
June 5).

June 7. The posterior member is much more massive than the
other, which still sticks out to side and shows a very imperfect hand.

June 19. The posterior member (HOM) is practically normal with
four digits; it is mobile, though probably there is some extensor weak-
ness of hand. The anterior member (HET), a left (not reversed), is
thin and atrophic, the imperfect hand having three digits. It arises
from near the shoulder of the other, the reduplicating plane being
approximately radial (fig. 113, drawn June 12).

Experiment S. E. 6. April 21, 1916. Right limb bud to left side,
dorsodorsal orientation.

April 22. Perfectly healed.

April 26. On operated side there are two projections in limb region
(fig. 109 A).

May 2. Operated limb not quite so advanced as the normal; dis-
tinct spur (fig. 109 6, S) on radial border, probably from the anterior
of the two prominences.

May 8. Specimen preserved (fig. 109). Operated limb (7'R) not so
advanced as normal, digits not so well developed. The spur (S) is
attached to the antero-lateral border of arm above elbow and is as
large as one of the primary digits. It is a radial reduplication which
has remained abortive.

16. Heteropleural, dorsoventral

Experiment S. E. 11. May 26, 1916.. Left limb bud to right side,
dorsoventral orientation; mesoderm torn along posterior border of
wound; no tissue lost.

May 27. Perfectly healed.

May 30. On side of operation a large limb bud points dorsopos-
teriorly like the normal. Ventrally another distinct but smaller bud
(DU) points ventroposteriorly (figs. 114 and 115 drawn May 31).

June 4. Operated limb (TR) longer and further adyanced (digits)
than normal (N). Otherwise no abnormality (fig. 116, drawn June 5).
No trace of ventral bud noted last time.

June 7. Limb on operated side still a little larger than the other.

June 12. Still some difference in size (fig. 118); limb normal in form;
motility not so good as normal, extensors of hand weak.

132 ROSS G. HARRISON

June 19. Function of operated limb very much better, almost, if
not quite, normal. Specimen preserved. No difference in size of
limbs.

D. Transplantation of half-buds
18. Homopleural, dorsodorsal

Experiment H. HE. 6. April 12, 1917. Anterior half of limb bud to
anterior half, normal orientation (fig. 120, 7). Operation on right side.
Pronephros removed and transplanted.

April 18. Perfectly healed.

April 17. Limb bud on operated side normal.

April 20. Same.

April 30. Same.

19. Homopleural, dorsoventral

Experimental H. FE. 31. April 9, 1918. Posterior half of right limb
(inverted) in place of anterior right (fig. 120, 6). ;

April 10. Wound perfectly healed.

April 15. Operated limb bud smaller than normal, and stands out
more sharply from body. It is more distinctly marked off (points)
anteriorly.

April 18. Operated bud distinctly more massive (fig. 121, TR).

April 21. Hand double, but coalesced to end. Digitations indis-
tinct.

April 30. Practically a normal right limb with an accessory hand
growing from back of hand. The preserved specimen shows this to
be a case of reduplication mirrored from a dorsal plane slightly inclined
to the radial. The reduplicating member consists of second and third
digits, (2’ and 3’) the former slightly bifurcated. The first (radial)
digit is not doubled (fig. 122).

Experiment H. E. 29. April 9, 1918. Ventral half (inverted) of
right limb bud in place of dorsal right (fig. 120, 8). Embryo from
which graft was taken, stained in Nile-blue sulphate.

April 10. Healing good. Small round area still uncovered by
ectoderm at ventroposterior border of wound.

April 15. Operated bud has grown considerably, pointing dorsally,
and slightly anteriorly. Whole free portion of bud covered by stained
(grafted) ectoderm.

April 18. Limb on operated side points dorsoposteriorly with redu-
plicating nodule on posterior border.

April 21. Limb short; double hand.

April 26. Well-developed double hand. Anterior member further
developed, sticks out to side; is a left; palm anterior; other a right;
dorsoradial reduplication.

May 4. Almost perfectly symmetrical double hand.

May 6. Specimen preserved. Arm as a whole a right, as isthe
posterior hand. This has three digits and a nodule for fourth, the first

SYMMETRY IN TRANSPLANTED LIMBS 138

two digits being partly syndactylous. NReduplication is radial. The
anterior hand (which is a left) has two long digits and a well-developed
third.

Experiment H. FE. 2. April 18, 1916. Dorsal half of right limb bud
(inverted) in place of ventral right (fig. 120, 7).

April 19. Healing good, but a groove dividing the limb region
horizontally still indicates line of suture. :

April 21. Two rather distinct humps on operated side; larger one
ventro-anterior.

April 26. Operated limb bud nearly normal, but there is a bud-like
projection on the anterolateral border (fig. 1238, S).

May 1. Transplanted limb larger than normal. The anterior proc-
ess is now a spur at elbow of the main limb, which is nearly normal.

May 8. Preserved. The operated limb, especially above the elbow
is a little thicker, but is otherwise normal. The spur is a nodule just
above elbow on radial border (fig. 124).

Experiment H. EF. 5. April 12, 1910. Posterior half of right limb
bud Gnverted) in place of anterior right (fig. 120, 6).

April 13. Perfectly healed except for minute uncovered area at
dorsal border.

April 17. Operated limb bud double; large posterior bud points
normally, but is not so large as normal; anteroventral bud is much
smaller; it is clear that former is developing out of the remaining half
of the limb bud of the host, while the latter is from the graft.

April 20. Both buds have grown; posterior one points more dorso-
laterally than normal and shows first beginnings of digitations; anterior
bud prominent, rounded, no digitations; the two are separate to base.

April 25. Posterior member is fan-shaped, with three long digits,
and beginning of ulnar digits on each margin; anterior member much
shorter, with faint indication of digitations.

April 30. Posterior member has a symmetrical fan-shaped _five-
digitate hand; as a whole it is a right limb, as indicated by elbow bend.
Anterior member much shorter and thicker.

May 21. Specimen preserved. The posterior member has a double
hand, of which the posterior is a right and the anterior a left. The
anterior member is a nearly normal right, with the hand a little twisted
and syndactylous second and third digits (fig. 125). This is an anoma-
lous case.

Experiment H. FE. 13. May 8, 1917. Posterior half of right limb
bud (inverted) in place of anterior half (fig. 120, 6).

May 4. Perfectly healed.

May 7. Two distinct though small limb buds, one ventro-anterior,
the other dorsoposterior.

May 10. Two buds united at base; both project sharply laterally
and slightly posteriorly.

May 14. Two limbs; bifurcation in Y-form about at elbow. Ante-
rior member shows two digits; posterior, none.

134 ROSS G. HARRISON

May 21. Specimen preserved (fig. 126). Hand double; limb (TR)
as a whole looks like a right, including the anterior hand (47UM),
which has two long digits and a distinct third. The posterior hand
(iT) is a left, also with three digits, the first of which is truncated.
Reduplication ulnar.

20. Heteropleural, dorsodorsal

Experiment H. R. E. 1. April 18, 1916. Posterior half of ne
limb bud in place of anterior left (fig. 120, 10).

April 19. Well healed.

April 21. Two distinct humps on operated side; dorsal one larger,
ventro-anterior one quite small.

April 26. Limb on operated side about as large as normal, but pro-
jects laterally.

May 1. Reduplication of hand; three digits, fan-lke.

May 8. Specimen preserved (fig. 127). Arm somewhat thicker
than normal, projects posteriolaterally with hand flexed ventrally.
Hand symmetrical, with five digits, of which middle one (1) is defective.
The posterior half of the hand (HOM) is a left, the anterior (HET) a
right.

Experiment H. R. E. 5. April 19, 1916. Dorsal half of right limb
bud in place of dorsal left (fig. 120, 77).

April 20. Perfectly healed.

April 24. Operated limb bud not quite so large as normal, points
laterally and slightly anteriorly.

April 29. Operated limb double (fig. 128).

May 8. Specimen preserved (fig. 129). The dorsal (posterior)
member (HOM) is a nearly normal left arm; first digit short; third is
small, and fourth not visible. The ventral (anterior) member (HET)
arises Just below elbow, and likewise shows three digits, of which first
is short; plane of mirroring is radial.

Experiment H.R. E.11. April 11, 1917. Anterior half of left limb
bud in place of posterior right (fig. 120, 9). Wound perfectly healed
same evening.

April 14. Operated limb bud very large and prominent, position
normal,

April 16. Limb bud massive, points dorsolaterally; almost perfect
anteroposterior symmetry.

April 23. Limb has now assumed normal position; first two digita-
tions show; a distinct but small spur at elbow on radial border points
laterally.

May 4. Operated limb normal; spur much reduced.

May 7. Specimen preserved. Arm normal; spur a mere nodule
about middle of upper arm, radiodorsal border (fig. 130).

Experiment H. R. E. 43. April 5, 1918. Anterior half of left limb
bud in place of posterior right (fig. 120, 9).

Apri 6. Perfectly healed.

SYMMETRY IN TRANSPLANTED LIMBS 135

April 9. Operated limb a little more massive than normal.

April 11. Limb more massive, rounded, not more distinctly marked
off posteriorly than anteriorly when viewed from above. Points
dorsally slightly posteriorly in lateral view.

April 14. More massive, but otherwise normal.

April 22. Not quite so long; syndactyly first two digits, otherwise
normal.

May 4. Normal except for syndactyly.

Experiment H.R. E.9. April11,1917. Posterior half left limb bud
in place of anterior right (fig. 120, 70). Wound perfectly healed same
evening.

April 14. Operated limb bud looks nearly normal. No definite
pointing.

April 16. Two rather distinct nodules, the posterior one consider-
ably more prominent.

April 18. Posterior bud smaller than normal, points dorsolaterally.
Anterior one not so definite.

April 23. Two entirely separate limbs; anterior one is shorter,
points ventroposteriorly; posterior one points straight to side; no
digitations.

April 28. Two limbs point posterolaterally, parallel to one another;
anterior one thicker, has two digits and beginning of third; apparently
a normal, though much smaller, right limb in approximately normal
position. Posterior limb thinner, rod-like, with no digits.

May 4. Anterior limb a normal right; posterior, very imperfect,
has one long digit and third digit nodule on upper border; elbow bend
shows; limb probably also a right (?)

May 12. Anterior limb has good function.

May 21. Specimen preserved. The anterior member is a normal
right, somewhat smaller than the unoperated limb. Posterior member
has one long and one short digit, and has same general position as the
anterior. Total view leaves in uncertainty which is palmar surface
(fig. 131). Sections show clearly that the ventrolateral surface is the
palm and that the limb is therefore a left.

Experiment H. R. EL. 20. May 2, 1917. Dorsal half left limb bud
in place of dorsal right (fig. 120, 77).

May 3. Wound perfectly healed.

May 7. Operated limb about full size, points anteriorly and slightly
dorsally.

May 10. Projects straight to side. Slght indication of posterior
reduplicating bud from near attachment of limb.

May 14. Main limb points dorsolaterally and slightly posteriorly.
The reduplicating bud is attached near base and in part directly to
body wall; it has grown considerably, but is still without digits.

May 18. Main (anterior) member sticks out to side, but now points
distinctly posteriorly as well, is slender and with two digits. Other
limb shorter, but considerably stouter, with faint indication of two
digits.

136 ROSS G. HARRISON

May 27. Anterior member has remained slender, and still has but
two digits; distinct elbow bend; general shape indicates it to be a right.
The posterior member, practically normal except slightly underdevel-
oped, is an undoubted right; first two digits partly syndactylous (first
short); third is distinct and fourth indicated. Some function in this:
member.

May 28. Specimen preserved; form of limbs shown in figure 132.
Sections show that there are two separate sockets in theshoulder-girdle
for the two limbs. The medial surface of both is the palm. Hence
both are rights.

21. Heteropleural, dorsoventral

Experiment H. R. E. 36. April 4, 1918. Ventral half left limb bud
(inverted) in place of dorsal right (fig. 120, 16). Pronephros removed.

April 5. Perfectly healed.

April 8. Operated bud a little sharper; points normally (dorso-
posteriorly).

April 10. Nearly normal; points slightly more dorsally.

April 13. Normal.

April 16. Normal limb; digitations show.

April 30. Specimen preserved (fig. 133).

Experiment H. R. EH. 10. April 11, 1917. Anterior half left limb
bud (inverted) in place of anterior right (fig. 120, 73). Wound com-
pletely healed same evening.

April 14. Limb bud operated side nearly normal looking.

April 16. Almost perfectly normal.

April 18. Operated limb normal except slightly larger.

April 23. Reduplication of digits radial side.

April 28. Operated limb a normal right with two digits arising from
palmar surface near radial border; arm somewhat thicker than normal.

May 7. Specimen preserved (fig. 134). The first digit is somewhat
small; the reduplicating digits (DU) are united at base.

et gloat: Mden fe Yen: ranierall

ee sid ei Wy ite Age
aah BS Ponti) TT iE ao

a OG butt 15) 9 ne gue i bide witiadt

4 r
“ i Bae oe) 7 4 é ifleg to, [

“val diphee: “a Lie Feith ica! eet
Tibet ee) oat Ae imal j aystal?
re Nei Axle. i 1% alae 4 Ra wis
7 ee uli pi ade

igatogie inl “sane cies Tile
4 }
i git wie aj Pash hte 11% Ve ane

ath

hast Mat mcltt nett 4 itbsihe!
pm Gly Dai lire! il is bvbhain
9 “aA =e | Vi ‘age
re. : ia, wea) ite: api i bs ‘ipeaptiihs bar sata , Re: } ad etl
oir tik aor adele di, Pel marty) tga). mba,
a ee ie ees wi maths ol. ohne
@ul aay fiw HY. F viball vey is ONG) lise ia
eal? CARTE: 20’ ee ‘ii “ ul Silie era -
"he ee a ee ee
| a A AP Mit al a . iar hatiiha:
he ota niet - > ele ae ieydin ris fy mati: Ain
pe? Tint gee ee hike Tad castaviiteil ahuit VU if.
‘tein hacia aroha Daisy Oimig ai .ad ime ot

= (rs enna ri eh BM, rive vi Plo i } ;
: J svgealyreyy

re Laie
= _ . : 3
a ows oa Fy) rer vee 3
a cy
I
iat mise
Pere.
3 wie
hee ee Al temtey he,

138 ; JOSEPH HALL BODINE

special interest are: changes occurring with age, differences due
to species and sex, the effects of starvation and hibernation, as
well as problems dealing with the various phases of metabolism.
Results of such nature obtained for higher forms have contri-
buted much toward the advancement of various theories of
growth, senescence, etc. (Child, 1; Mathews, 2; Minot, 3), as
well as to our knowledge of eranee: in organisms one are
closely related to age, sex, etc. (Hatai, 4).

The present investigation deals with such problems as the
percentage of water during growth, starvation, and ‘hibernation’
in different species of grasshoppers, and also the rate of meta-
bolism as indicated by determinations of the amounts of carbon
dioxide given off by the animals.

MATERIAL AND METHODS
A. Material

Grasshoppers were used in all the following experiments.
These animals have been found to be excellent material because
of the ease with which they can be obtained, kept under usual
laboratory conditions, and handled in experiments. They are
sufficiently large to be used individually, and this is of great
importance because most of the physiological work heretofore
done on insects has been concerned with masses 3 rather than with
individual animals.

The following species were used: Melanoplus femur rubrum,
Melanoplus differentialis, Dichromorpha viridis, and Chorto-
phaga viridifasciata, and they will be Beate in the order
given.

Melanoplus femur rubrum DeGeer, the red-legged locust, is
perhaps the most common grasshopper found throughout the
entire United States. Its general life-history is practically
typical; eggs are laid in the late summer and early fall and remain
over winter; nymphs hatch in early summer, and by the last
of July and early part of August, in the vicinity of Philadelphia,
adults are found. It occurs in rather large numbers throughout
the entire summer. The average length of the body of adult

WATER CONTENT AND RATE OF METABOLISM 139

males is 23.5 mm., and of females, 24.5 mm.; average weights
are: adult males, 0.20 to 0.40 gram; females, 0.25 to 0.65 gram.
Nymphs range in weight up to a maximum of 0.35 gram.!

Melanoplus differentialis Uhler, the largest grasshopper found
in this vicinity, closely parallels Melanoplus f. rubrum in life-
history. In length adult males measure 39 mm., and females,
41 mm. Adult males weigh 0.7 to 1.3 grams, and females, 1.3
to 2.8 grams. |

Dichromorpha viridis Scudder has a general life-cycle similar
to the above-described species. However, it is not as active
an animal and occurs in open wet places. Differences in size
between adult males and females are marked. Adult males
measure 18.75 mm., and females,’ 27.0 mm. In weight adult
males range from 0.15 to 0.20 gram, and females, from 0.15 to
0.55 gram.

Chortophaga viridifasciata DeGeer is quite different in life-
cycle from the above-mentioned species. Eggs are laid in late
spring and early summer; these hatch in later summer and early
fall; the nymphs live throughout the winter, and in spring grow
rapidly, and become adults by early summer. ‘Two-thirds of
their active life, in contrast with other species, is spent as nymphs
and approximately one-third as adults. Two well-marked
varieties oc¢éur, a green form (virginaria Fab.) and a brown form
(infuseata Harris). Most females are green and males brown,
but some are found of each sex in either color, and as a matter
of fact, when green animals are put at a constant temperature
of 38°C. they turn brown in a very short time. Adult. males
measure 25 mm. and weigh 0.10 to 0.20 gram; females measure
30 mm. and weigh 0.15 to 0.45 gram.

For further descriptions of the above species reference is made
to standard text-books on entomology and to the works of
Morse (6) and Lugger (4).

1 Average dimensions of animals used are taken from Lugger (5), while body
weights have been determined by the author.

140 JOSEPH HALL BODINE
B. Methods

The following general description of methods applies to all
experiments and any further details will be given in describing
individual cases.

All animals were caught in the vicinity of Philadelphia during
the summer and fall of 1919-1920, brought into the laboratory,
where they were kept in large screened insect cages, designated
as stock cages. They remained in these cages for at least a day
under the usual laboratory conditions, and were fed during this
time on grass. Inasmuch as grasshoppers normally consume a
great deal, those kept in the laboratory for a day ate large
amounts of the grass, and upon examination the alimentary
canal was found to be filled, thus insuring uniformity as to
initial amounts of food. General laboratory conditions remained
constant throughout the experiments, and any slight temperature
changes, usually occurring at this particular season of the year,
are noted in data following.

Animals were separately weighed in a small covered beaker
on a rather delicate balance, determinations being made to four
places of decimals. After weighing they were marked by gluing
a small numbered tag on the pronotum, which mark could easily
be removed and again attached, thus avoiding confusion in
keeping accurate records of individuals. After initial weighing
they were kept, in groups of five to ten, in small wire insect
cages.

In determining water content, individuals were killed with
chloroform, opened by a longitudinal slit through the abdomen,
and then put in an oven at 95° to 97°C. and left there for a
period of one week. This was found to be more than ample
time for complete desiccation.

Carbon-dioxide determinations were made by the barium
hydrate titration method of Lund (7), using single animals, and
each determination extending over a period of thirty minutes
to one hour. In suspending individuals in the respiration bottle
they were carefully tied around the prothorax by means of a
fine silk thread. This was found to cause them little incon-

WATER CONTENT AND RATE OF METABOLISM 141

venience if properly adjusted, and if it was kept sufficiently short,
they did not move about to any appreciable extent but rested
upon the sides of the bottle. Experiments in which the animal
was confined in a small wire cage, just large enough to accom-
modate it and not allowing body movements, gave results quite
similar to those in which the animal was suspended by the
thread. In any such experiments, however, it is quite impossible
to entirely eliminate movements by the individual, but by careful
handling and manipulation they can be greatly lessened, and
comparable, if not accurate, results obtained. In the following
experiments emphasis is laid upon relative rather than upon
absolute amounts of carbon dioxide given off.

Acknowledgment is made to Prof. C. E. McClung for the
original suggestion of work on the physiology of the grasshopper,
and to Prof. M. H. Jacobs I am deeply indebted for constant
advice and criticism during the course of the work. To the
members of the Zoological Department of the University of
Pennsylvania I am also greatly indebted for generous help and
criticisms.

OBSERVATIONS
A. Water content

The physiological importance of water to organisms is too
well known to require special discussion. In recent years,
many determinations of water content have been made on differ-
ent organisms, and on the same organisms under different con-
ditions, and the results have thrown much light on such questions
of biological interest as the cause of certain tropic responses
(Breitenbecher, 8), the nature of the process of senescence
(Hatai, 4), the question of the influence of different foods (Bab-
cock, 9), ete. Determinations of this sort have been made
mostly on higher forms, and no such observations appear to be
available for grasshoppers, which I have, therefore, studied
rather extensively.

a. Relation between body weight and water content. Table 1
gives the body weights and water contents of 981 individuals,

142 JOSEPH HALL BODINE

comprising nymphs and adults of Melanoplus f. rubrum and
adults of Dichromorpha viridis and Melanoplus differentialis.
It is evident from an examination of this table that considerable
variation in the body weights for the different species exists.
_ These can be explained by a consideration of the differences due
to species, age, and random variations. It will be noted, for
instance, that of the species studied, Melanoplus differentialis
is by far the heaviest, reaching a maximum of 2.9 grams. Range
in weight for Melanoplus f. rubrum is of interest, since weights
for both nymphs and adults are given, and these show that as
the animals become adults there is a progressive increase in
body weight up to a maximum for the species. Males never
reach the same maximum weights as females, and this is not due
primarily to development of masses of eggs by the older females,
since in nymphs this relation also holds. Consequently any
comparison between males and females of the same weights will
not necessarily be between those of the same age. Different
conditions, such as food supply, development of reproductive
elements, etc., modify the maximum weights of animals, causing
some variation among individuals of the same age as shown for
Melanoplus differentialis, where the animals are of approximately
the same age and show rather large variations in body weights.

Closely related to these differences in body weights of the
animals are the changes occurring in the percentages of water.
With increasing body weight and age, a progressive diminution
in the relative water content takes place, as shown especially
by Melanoplus f. rubrum, where nymphs have an average maxi-
mum of 77.6 per cent and adults an average minimum of 72
per cent. That this diminution in water content is related to
age, and not to body weight, is shown by a comparison of these
results with those for Melanoplus differentialis, where the ani-
mals are of the same age, but differ in body weights, and have
approximately the same percentage. of water. <A similar result
has been found by Hatai (4), for the albino rat, in which the
percentage of water during the life-span decreases from 87.2
per cent to 65.3 per cent and bears no relation to body weight,
but depends primarily upon the age of the animal. It will also

WATER CONTENT AND RATE OF METABOLISM

143

be noted that the average percentage of water for the adult
males of Melanoplus f. rubrum tends to be slightly higher than
that for the females, and that for Dichromorpha viridis the
reverse relation is found.

TABLE 1

Showing the percentage of water in-normal grasshoppers of different species.
Figures in () giving the number of animals used

BODY

PERCENTAGE OF WATER

Melanophus f.

Melanophus f. rubrum

Dichremorpha viridis

Melanophus differ-

WEIGHT rubrum (nymphs) (adults) (adults) entialis (adults)
fou 2 (of ie) fon ie) rol Q
grams
0.10-0.15/76.7( 8)/78.5(11) 69.8 (84)|75.1 (25)
0.15-0.20/75.0 (10) |77.9 (14)|74.9 (7) 64.9 (15)|74.3 (15)
0.20-0.25)/74.8 (18 )|74.1(15)|74.2 (19) |76.0(25) 74.3 (40)
0.25-0.30 73.9 (10)|73.6 (82) |74.1 (40) 73.3 (42)
0.30-0.35 74.4 (12)|73.4(47) |74.5 (45) 71.8 (50)
0.35-0.40 73.9(19) |72.5(23) 68 .4 (57)
0.40-0.45 70.8 (12) 67.9 (40)
0.45-0.50 73.1(14) 65.4 (20)
0.50-0.55 TAL PACU) 66.7 (18)
0.55-0.60 ALA U(Gls}))
0.60-0.65 70.3 (18)
0.90-1.10 69.3 (15)
1.10-1.30 69.6 (11)
1.30-1.50 69.7 (10)
1.50-1.70 66.8 (12)
1.70-1.90. 71.0(25)
1.90-2.10 69.2 (18)
2.10-2.30 69.0(10)
2.30-2.50 t 69.4 (84)
2.50-2.70 69.9 (27)
2.70-2.90 69.3 (19)
Average.|75.5 (41) 75.7 (62)|74.0(124)|72.6 (217 )|67.3 (49) 70.8 (307) 69.4 (26) 68.0 (155)

b. Specific differences.

It will be noted from table 1 that the

average percentage of water differs for the various species studied.
For example, Melanoplus f. rubrum has an average water
content of 74.4 per cent, Dichromorpha viridis 69 per cent, and

Melanoplus differentialis 68.9 per cent.

These averages (except

144 JOSEPH HALL BODINE

those for Melanoplus differentialis) are based upon individuals
taken during the entire summer and hence represent values for
both young and old animals and can be considered as fair averages
for the species. That such an average must be based upon the
water contents of both young and old individuals is quite evident,
since the preceding section shows that the percentage of water
of an animal decreases with increasing age. It is of some interest
to mention here that in another species of grasshopper, Chorto-
phaga viridifasciata, the percentage of water during the winter
months drops to a minimum of 65 per cent, while in the same
animal when growing the water content is raised to 75 per cent.
Since the percentage of water of an animal is thus dependent
upon and so easily modified by external factors, such as tempera-
ture, moisture, ete., it is difficult to get data to show that definite
species differences in water content exist. The present results
for Melanoplus f. rubrum and Dichromorpha viridis, where the
animals were under the same conditions, seem to show, however,
that such a condition might occur. Babcock (9), in discussing
‘metabolic water,’ concludes that the water content of insects
depends largely upon the nature of the food eaten, but does not
show that those living upon the same foods may have different |
percentages of water.

c. Changes during the normal life-cycle. Some changes in the
percentage of water during the life-cycle of grasshoppers have
been shown in a preceding section. Chortophaga viridifasciata,
however, because of its peculiar habit of living as a nymph during
the winter and ‘hibernating,’ affords opportunity for further
studying the various changes in water content which normally
occur during such a period.

Table 2 gives results obtained for 565 specimens and shows
that during this period the percentage of water falls from 72
per cent to 65 per cent. This minimum of 65 per cent is perhaps
as far as desiccation can be carried without injurious results, and
is a factor of some importance in the economy of the organism.
Somewhat similar rhythmic variations in the water content of
animals are shown for the frog by Donaldson (10), and for the
potato-beetle by Breitenbecher (8). These ‘hibernating’ animals

WATER CONTENT AND RATE OF METABOLISM 145

with low percentages of water, left at 38°C. for a period of three
weeks and fed grass, grow and become adults, and the various
changes in the water content during this period are easily followed.

Table 3 selected from several such experiments shows that a
growing animal has the maximum percentage of water (75 per
cent), and that a progressive decrease in water content with
age and increasing body weight, to a minimum of 65 per cent,

TABLE 2

Showing the percentage of water and solids of Chortophaga viridifasciata taken at
different periods of the year

PERCENTAGE OF a:
DATE TEMPERATURE 5 ada ad
Water Solids
x
October. 10, 1919: .... 2. lace 22-25 72.0 28.0 50
Octobersile 1919... eee 22-25 72.0 28.0 35
November 4, 1919......... 13.0 69.4 30.6 25
November 5,.1919)... ... 5 1s. 8.0 69.4 30.6 15
November 138, 1919........ 10.0 66.5 33.5 17
November 14, 1919........ 4.5 66.1 33.9 43
November 18, 1919........ a0 69.5 30.5 14
November 19, 1919........ 7.5 67.3 32.7 60
November 23, 1919........ 11.0 67.3 oot 47
November 26, 1919.......-. 9.0 65.7 34.3 62
December 3, 1919.......... 0 65.2 34.8 50
December 8, 1919.......... 5.0 65.2 34.8 34
December 16, 1919......... 0 67.4 32.6 19
December 17, TO19s Suc. 0 66.5 33.5 38
December 22; 1919......:.. 0 62.8 37.2 15
Vantanye+ WLOZ0E Se eee ser 0 65.0 35.0 41

takes place. Figure 1 gives graphically and in more detail,
similar results taken from many experiments which show the
water content and weight relations of animals during the period
from October, 1919, to January, 1920. The results of two
experiments at different intervals, giving the effects of a constant
temperature of 38°C., are also indicated.

From this figure it is to be noted that during this four-month
period, rather marked changes in the water content of animals
take place, while body weights undergo only slight and gradual

146 JOSEPH HALL BODINE

increases. With the approach of cold weather the animal
begins gradually to lose water, and with sudden decrease in
temperatures rather marked drops, to a minimum of 65 per cent,

TABLE 3

Showing the changes in water content of Chortophaga viridifasciata during different
stages of its life-cycle

—_+—

WEIGHT OF ANIMAL STAGE OF GROWTH CONDITION PER CENT OF WATER
grams
0.0598 Nymph Growing 74.9
0.0648 Nymph Growing 76.8
0.0749 Nymph Growing 73.2
0.0800 Nymph Growing 75.2
0.0895 Nymph Growing (AE
0.1100 Nymph Growing Whe)
0.1700 Nymph Growing 73.8
0.2498 Nymph Growing 74.1
0.0700 Nymph Hlbernating 64.8
0.0750 Nymph Hibernating 60.6
0:0750 Nymph Hibernating 65.4
0.0725 Nymph Hibernating 65.6
0.0848 Nymph Hibernating 64.7
0.0840 Nymph Hibernating 66.7
0.0885 Nymph Hibernating 66.0
0.1020 Nymph Hibernating 65.7
0.2780 Adult Growing 76.5
0.3850 Adult Growing 73.4
0.3715 Adult Growing 75.3
0.3150 Adult Growing 74.7
0.4130 Adult Growing 73.4
0.4230 Adult Old 67.0
0.1350 Adult Old 61.5
0.2640 Adult Old 68.6
0.3020 Adult Old 69.3
0.2480 Adult Old 63.2
0.3970 Adult Old 61.6
0.3630 Adult Old 61.5
0.4570 Adult Old 64.9

occur. The water content then remains at this minimum during
the remainder of cold weather. Despite this preparation for
the winter by a falling off in the percentage of water, the process
seems one quite easily changed at any period by exposure to a

WATER CONTENT AND RATE OF METABOLISM 147

higher temperature of 38°C. The two experiments represented
show results typical for many others obtained at different inter-
vals during this four-month period. It will be noted that a
marked and steady increase in body weight takes place, until
at the end of approximately three weeks, the maximum for the
species is attained. The percentage of water, on the other hand,

Percentage of
water |

ue eis All ata a a i 2 ia |
ean be

Body weight Gms

i] \ oO water |
| \ @ body wt
norma) animals

23 et] {--— — animals at 38° 05

\
\
\
\
| \ :
: | {
| ala ‘ 4 + 7 = aT (ee
YX 7 ] 7 ‘i |

A Va
Ye 7
VA wa
- Boa
OCT.8 : NOV.1 DEC.1 : JAN.1

Fig. 1 Curves show the water content and weight relations of Chortophaga
viridifasciata, during four months, from October 8, 1919, to January 1, 1920, out
of doors, and also the effects of 38°C. on these relations in ‘hibernating’ nymphs.
Abscissas, time in weeks indicated. Ordinates, at the left, the percentage of
water. Ordinates, at the right, body weight in grams. For further explanation
see description in text.

undergoes rather striking and regular changes. A slight decrease,
followed by a rapid increase to a maximum of 75 per cent and
accompanied by active growth, and a correspondingly rapid
decrease, closely connected with the later stage of the animals’
life, occur. Such a result strikingly confirms conclusions arrived
at in an earlier section in which such differences in water content
were shown to be correlated with the age and not the body weight
of an animal.

148 JOSEPH HALL BODINE

d. Effects of temperature. As pointed out in the preceding
section, raising the temperature to 38°C. causes marked changes
in the water content of ‘hibernating’ individuals. That such
results can be obtained at other temperatures is shown from a
series of experiments in which ‘hibernating’ nymphs were kept
at 9°, 23°, and 38°C., respectively, for periods of:three weeks
and fed grass. Progressive increases in the percentage gains of
water, body weight, and solids take place in all three series,
being highest, however, at 38°; while with nothing to eat at
these temperatures the animals die, thus showing that during the
winter a real hibernation can hardly be supposed to take place,
since it is generally understood that hibernating animals require
no food other than that already stored in the body.

e. Effects of relative humidity of the air. ‘Hibernating’ nymphs
were put into small cages covered with wire gauze and the cages
were then put into, 1) sealed jars containing wet sand and filter-
paper, 2) sealed jars with dry sand, and, 3) a desiccator contain-
ing calcium chloride, and kept at temperatures from 4° to 388°C.
The animals were not in contact with the sand or wet filter-
paper, and hence any increase in weight or water content cannot
be attributed to imbibed water. Nothing was given the animals
to eat. .Table 4 shows the percentage loss in weight, water,
and solids for ninety-five individuals treated in such a manner.

In general it is found that in the jars with dry sand and in
the desiceators the animals lose weight and water, the losses
being highest at 38°. Losses in water are relatively higher than
those in body weight. In the jars with the wet sand, on the
other hand, marked increases in body weight and water result.
At 4°C., however, a slight loss in weight (1.4 per cent) is noted,
but an increase in water of 3.5 per cent takes place. Such a
slight absorption of water at this lower temperature further
shows how the organisms are protected during winter, preventing
freezing and possible destruction. It is evident from these
results that ‘hibernating’ nymphs are able to take up water
directly from the surrounding medium. Breitenbecher (8) finds
a similar condition in the potato-beetle. It is of interest to note,
too, that old individuals with low percentages of water are unable

WATER CONTENT AND RATE OF METABOLISM 149

to readjust their water relations when exposed to decreased
temperatures. The animal’s ability to regulate its moisture
content seems to be connected with its ability to grow and
withstand adverse conditions.

TABLE 4

Showing the percentage changes in weight, water, and solids in ‘hibernating’ Chorto-
phaga viridifasciata when exposed to differences in relative humidity

PERCENTAGE OF Sane na f
aad Change|Change|Change} 7™® aes aeet ADEA
Water in in in MALS ue
weight | water | solids
grams hours SCe
0.4130 | 66.4 | —2.8) —3.7) —0.7| 48 5 4 In cage with sand
0.4640 | 66.2 | —1.7) —3.3) +1.6) 48 5 4 | In jar with sand
0.2950 | 60.5 |—10.4;—23.9/4+16.1| 48 5 15 | In jar with sand
0.3250 | 61.3 | —8.6)/—15.5) +5.0] 24 5 23 In jar with sand
0.3560 | 59.5 |—14.6)—23.6} +2.1]} 48 5 23 In jar with sand
0.4150 | 60.5 |—24.9/—41.6] +7.1) 24 5 38 | In jar with sand
0.4180 | 65.4 | —2.8)| —5.7| +1.7| 48 5 4 | In desiceator
0.3810 | 60.5 | —9.9|—17.7| +7.0} 48 5 15 In desiccator
0.3695 | 60.0 | —7.3}/—16.2/4+10.5} 24 5 23 | In desiccator
0.4315 | 60.1 |—14.7/—23.5) +2.4| 48 5 23 | In desiccator
0.3730 | 59.9 |—24.1)—34.0) —4.4| 24 5 38 | In desiccator
0.5170 | 68.1 | —1.4) +38.5/—10.7| 48 5 4 | In jar with wet sand
0.42380 | 64.8 | +38.5) +0.5| +9.5) 48 5 15 | In jar with wet sand
0.38380 | 62.5 | -+-0.5| --0.3) -+-0.7| 24 5 23 | In jar with wet sand
0.5810 | 71.6 |4+-14.3/+22.6] —2.3) 48 5 23 In jar with wet sand
0.3950 | 73.0 |+15.5/+28.0) —9.0} 24 5 38 | In jar with wet sand
0.4845*) 64.6 |—14.8/—18.0) —8.1| 24 5 38 | In jar with wet sand
0.4660*| 64.4 |—14.8)—18.7| —6.6| 24 5 38 | In desiccator
0.4840*) 61.9 | —3.7| +1.9/—15.6} 24 5 38 In jar with wet sand

+ = gain, — = loss. * = animals previously kept at 38°C.

B. Effects of starvation

Before dealing with the direct effects of starvation on water
content, solids, and body weight of these animals, some points
of general interest deserve consideration, one of which is length
of the starvation period. It is found that adults of Melanoplus
f. rubrum endure complete starvation approximately 73 hours,
while with water, but no food, they live as long as 144 hours

150 JOSEPH HALL BODINE

with a loss of 30 to 35 per cent in body weight. On the other
hand, Melanoplus differentialis, a larger species, survives com-
plete starvation approximately 96 hours, and with water alone,
about 172 hours with a loss in body weight of 20 to 25 per cent.
Hibernating nymphs of Chortophaga with nothing to eat, at
temperatures from 0° to 9°C., can survive only a little more
than two weeks; at 23° they live about one week, and at 38°
only three to four days. The maximum loss in weight up to
death ranges from 20 to 25 per cent. Such a short survival
period for the grasshopper is in marked contrast with that found
for certain insects and related forms. Dufour (17) for example,
kept bedbugs for a year without food, while Riley and Johannsen
(12) cite examples where certain ticks were kept for over three
years with nothing to eat.

The changes brought about by starvation in the body weight,
water, and solids of the grasshopper are rather striking as the
following results show. In all experiments adult animals were
used and were weighed at twenty-four-hour intervals at the same
time each day. The number for Melanoplus f. rubrum is 250,
and for Melanoplus differentialis, 75. Room temperature during
the experiments remained at 22° to 25°C. Table 5 gives the
percentage losses in body weight, water, and solids for different-
sized individuals and for the two sexes of Melanoplus f. rubrum,
during seventy-two hours of starvation, with and without water.
Figure 2, taken from this table, shows the average losses in body
weight and water, and figure 3 gives the average percentage
loss in weight and also the average percentage loss in weight
per day or the rate of loss, with water alone, for Melanoplus
differentialis.

From an examination of these data we find that losses in body
weight during starvation are marked, and that they increase.
progressively as starvation proceeds up to a maximum for the
species. The rate of loss, indicated in figure 3, is greatest,
however, during the first forty-eight hours and diminishes subse-
quently up to the end of the experiment. Losses in water, as
shown in figure 2, are always relatively greater than those in
weight, and maintain this same general relation throughout

WATER CONTENT AND RATE OF METABOLISM 151

] FIG.2
| PERCENT

® loss inwt. 3

a,b,c,d,wi'th H,O
—--—--O " "4,0 e,f,a.n, “ nothing

i TIME-HRS.

Og 48 "72

Fig. 2. Curves show the average percentage losses in body weight and water
during starvation, with and without water, for both sexes of Melanoplus f. rub-
rum. Abscissas represent period of starvation in hours indicated by numbers.
Ordinates represent percentage of losses occurring during starvation. See
table 5.

PERCENT

a,b, loss inwt. per day
c,d, " " original wt.

Time-hrs .
20
ale T T

24 48

Fig. 3 Curves show the average percentage loss in body weight per day and
the average percentage loss of original weight during starvation, with water, for
both sexes of Melanoplus differentialis. Abscissas represent period of starva-
tion in hours indicated by numbers. Ordinates represent percentage losses
occurring during starvation.

15? JOSEPH HALL BODINE

TABLE 5
Showing the water content and the percentage losses in body weight, water, and
solids in Melanoplus f. rubrum during starvation. Top figures denoting
losses without water and lower ones those. with water

PERCENTAGE OF

INITIAL BODY WEIGHT TIME Water Loss in wt. Loss in HeO Loss in solids
fog Q J e) fou 2 fof 2
grams hours
0.20-0.25 24 73eOU7808 | 722 | 22) 8.01 bo 1 46 leeenn
pp liacies 10.9 14.0 2.5
= 6005 76.7] 13.1 12.8) | AS BAe Gras leo ota
A 63.5 22.6 33.4 aS
95-1
Het est 48 11 76.3 8.7 5.4 Wie
5 65.7 28.3 36.3 6.3
[ 74.1 9.2 8.5 ible
ah 70:0 -)) 7328 el AS24393) Pea) 1415) MO eae
“\| 69.1 10.9 14,7 45.3
P 65.9) (69-2, 201620224) 987 26.012 38
0.5 : 8
Be SURE 48)! 79.5 8.7 9.7 5.3
D 68.8 22.5 Pell 9.6
% 70.9 17 14.7 3o2
o4 {| 67.7 | 72.6 | 13.4 | 15.8 | 20.9 | 15.5 |+7.6 | 18.1
3 70.1 10.5 14.7 +1.9
64.45] 70/8 | 22:8 | B57 S27 | 27.5 |b ye) eos
35-0.¢ 8
Uist es FAD THAN BiG tO 250 | ess |A-IO |e Ore a eae
=: 63.4 | 66.1 | 25.2 | 24.6 | 35.8 | 31.0 |+4.8 7.9
x 70.5 | 74.2 | 11.6 | 15.6 | 15.5 | 13.4 | 0.0] 20.9
Bh 71.9 Isr 12.5 16.7
68.2 11.9 15.3 3.8
. 67.5 23.0 26.6 14.5
Oe as ALS 9.4 8.7 10.6
a 66.2 24.8 29.7 12.9
68.4 18.6 27.8 19.0
24 67.2 17.0 Bes +14
66.1 1s7 Spall agg
He his 65.6 24.6, 32.5 2.9
abate last ee 69.5 12.4 16.6 0.7
= 64.7 26.4 34.9 2.9
66.5 15.3 10.3 370)

WATER CONTENT AND RATE OF METABOLISM 153

TABLE 5—Continued

PERCENTAGE OF

INITIAL BODY WEIGHT TIME Water Loss in wt. Loss in H2O Loss in solids
fou Q on Q fou Q rot Q
grams hours
5) 66.8 133.33 18.6 0.0
- 67.4 |. 18.8 23.2 as
64.6 20.7 28.0 2.5
eecyecre ee 70.4 15.6 16.5 13.2
Ea 6S |) | 38-9 42.8 11.3
a fale 18.0 18.1 17.7
a 66.0 13.9 20.1 Ba ey
67.6 17.9 22.0 73
64.7 18.2 25.5 0.0
Ledeauel He 68.2 11.8 15.4 28
ae 63.5 26.9 34.6 7.5
77 65.8 17.0 93.1 ey
si 69.4 15.6 16.6 12.9
2 66.7 12 at 11.9
65.9 Dp i 27.5 11e3
Pee ze 68.1 9.6 12.4 2.5
ris 64.5 25.6 Bled 10.8
s 66.4 19.7 24'0) 9.3

starvation. Losses in solids, however, are invariably lower than
those in body weight and water. This shows that starvation
in the grasshopper results in a rapid loss in water which has a
decidedly quick and fatal effect. In striking contrast to such
a condition, Hatai (73), with medusae, and Morgulis (14), with
salamanders, find that during starvation the water content is
increased rather than decreased, but it must be remembered that
in these cases we are dealing with aquatic forms. ‘Table 5,
arranged according to body weight, shows that considerable
variation in the losses for different-sized individuals exists, but
that after the first twenty-four hours of starvation, larger animals
tend to suffer the greater relative losses. This is perhaps due
to the fact that the lighter individuals are still growing, and as
pointed out by Donaldson (/5) in experiments with rats, the
loss of water in the nervous system of underfed individuals is

154 JOSEPH HALL BODINE

decidedly less for growing animals, and in growth such losses
are markedly more fatal than when growth has ceased. It is
also of interest to note that the average losses in body weight
and water for males are lower than those for females.

The general effects of starvation, with and without water, are
more graphically shown in figure 2, where some of the data from
table 5 are represented in the form of curves. It is quite evident
from these that grasshoppers must normally require water, and
that any condition which deprives them of it results in marked
losses to the animal, which rapidly become fatal. No results
on ‘metabolic water’ (Babcock, 9) are available, but it appears
that in the grasshopper there is present little of the power shown
by clothes moths, ete., of maintaining the proper degree of
moisture in the body tissues from water resulting from the
oxidation of the organic matter comprising the food and tissues
of the animal.

C. Carbon-dioxide output

The respiratory exchange of animals is of physiological sig-
nificance, since it gives quantitative evidence of the metabolic
processes taking place in the organism. Measurements are made
either of the oxygen consumption or the carbon-dioxide output,
and at present methods for the detection of the latter quantity
have been greatly improved and are especially favorable for
work on lower forms; such as insects. The factor of greatest
importance in such determinations, however, is the functional
activity of the animal. In the organism as a whole, functional
activity can be reduced only to a minimum, and in those animals,
like insects, where narcotization is impossible, only approxi-
mations to this can be obtained.

From the results of various investigators, it is of interest to
note that the respiratory rates for insects are considerably higher
than those for other animals. For example, Vernon (16), finds
that a cockroach, weighing 0.0007 kilogram, gives off 0.470
gram of carbon dioxide per kilogram per hour, while a frog,
weighing 0.004 kilogram, gives off only 0.140 gram. Smaller
and younger individuals of different species tend to have the

WATER CONTENT AND RATE OF METABOLISM 155

higher respiratory rates. It is of importance, however, to
mention here that most of the work heretofore done on insects
has been concerned with masses rather than with individuals,
and that little consideration has been given to results obtained
for different species of the same general group, for different sexes,
and for animals of different ages. The present discussion deals
with the carbon-dioxide output of individual animals of different
t | :

AeM. f. rubrum |

B-D. viridis }
C-M.differentialis

CO. p.gm.p.hr
ro]
a
i)
a
nN
°o
a
ao
le}
ao
a
p
oO

Fig. 4 Curves show the rate of CO» output per gram per hour for nymphs of
Melanoplus f. rubrum and Dichromorpha viridis and for adults of Melanoplus
differentialis. Abscissas represent body weights in grams. Ordinates repre-
sent rates of CO» output in grams CO, per gram total body weight per hour, for
the three species. For further explanation see text.

species, different ages, and of different sexes, under normal as
well as experimental conditions. The following species of grass-
hoppers were studied: nymphs of Melanoplus f. rubrum and
Dichromorpha viridis and adults of Melanoplus differentialis.

a. Carbon-dioxide output of normal animals. The number of
individuals studied is, for nymphs of Melanoplus f. rubrum, 350;
for nymphs of Dichromorpha viridis 300, and for adults-of Melano-
plus differentialis, 85. Figure 4 shows the average rates of
carbon-dioxide output per gram per hour for these animals.

156 JOSEPH HALL BODINE

An examination of this figure shows several interesting facts,
the most striking of which is that a difference in rate of CO,
output is noted between the three species. That such a differ-
ence is not due to body weight is shown by a comparison of the
respiratory rates of nymphs of Melanoplus f. rubrum and Dichro-
morpha viridis, which are of approximately the same weights.
The most plausible explanation of this fact seems to be that this
difference corresponds closely with the mode of life of the two
species, Melanoplus f. rubrum being a very active animal, while
Dichromorpha viridis is a relatively sluggish one. <A point of
further interest is that the rate of CO, output is higher for
lighter animals and decreases progressively as the animals
increase in body weight. As it has already been pointed out
that differences in body weight, especially in nymphs, are closely
correlated with differences in age, we are led to assume that
younger individuals have the higher rate of respiratory output.
Like results are also found for other species. Since differences
in body weight between males and females exist, the question
naturally follows as to whether similar differences in the rate.
of CO, output are found. Figure 6 shows that males tend to
have the higher rate. The animals of the two sexés, in this
case, are of approximately the same age, and differences in weight,
as shown in a previous section, are due mostly to eggs in the
female. Whether any fundamental difference in rate of respi-
ratory exchange exists between the two sexes is somewhat doubt-
ful, but such differences are reported for other animals, including
man (Benedict and Emmes, 17).

b. Rate of output. Much evidence has been accumulated
concerning higher forms and man to show that smaller individuals
have a greater respiratory exchange per unit of weight than
larger ones, and that respiratory exchange is proportional to the
area of the surface of the body (Rubner, 78). For lower forms
few such observations exist, Child (1/9), and Allen (20), for
example, with Planaria, find that respiratory exchange decreases
as the size of the worm increases, but give no calculations show-
ing any possible surface relations. Krogh (21), in summarizing
work done on lower forms, finds that results are conflicting, and

WATER CONTENT AND RATE OF METABOLISM Lov

concludes that no reason exists for assuming a surface relation
to hold.

As already pointed out for the grasshopper, smaller individuals
have per unit of weight a greater CO, output than larger ones.
And since the area of the surface of an animal is usually estimated
from the body weight by means of the formula of Meeh (22),
based on the law that surfaces of similar solids are proportional
to the two-thirds power of volume, it is of some interest to see
in how far the rate of CO, output of the grasshopper can be
thus expressed. In the following table are given a few examples
of the ratio of CO, output to body weight and to the two-thirds
power of the weight, respectively. .

CO:2 ACCORDING TO THE
WEIGHT OF ANIMAL COz2z ACCORDING TO BODY WEIGHT TWO-THIRDS POWER OF THE
BODY WEIGHT

grams

2.16 0.001000 0.001290
1.35 0.001037 0.001204
Teall 0.001174 0.001224
1.08 0.001186 0.001226
1.05 0.001215 0.001236
1.01 0.001219 0.001232
1.01 0.001263 0.001276
0.94. 0.001310 0.001282

It is evident that the more constant values are obtained by

using the two-thirds power of the weight, and so far as the
results here reported are concerned, the conclusion might reason-
ably be drawn, that the surface law holds for grasshoppers as
well as for mammals. But in view of the complex nature of the
problem, more extensive data will be necessary before this
relation can be considered as definitely established.

c. Effects of temperature. The influence of temperature on the
respiratory exchange is a somewhat disputed question because
comparatively few observations are made under standard con-
ditions. Krogh (21), however, in summarizing the work of
various investigators, points out that different animals respond
in different ways, but in general, with cold-blooded forms, in-
creased temperatures cause increased respiratory rates, while

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, No. 1

158 JOSEPH HALL BODINE

with warm-blooded animals, the reverse is the case. The effects
of temperature on the respiratory exchange of insects are espe-
cially marked and in general agree with results for other cold-
blooded forms. But since most of these results are based upon
masses rather than individuals, it has seemed desirableto show
the effects of temperature upon the respiratory exchange of
individual animals. Three hundred and fifty nymphs of Melano-
plus f. rubrum and three hundred of Dichromorpha viridis were

SVCED COz per gm. per hr. | = | =

Fig. 5 Showing effects of various temperatures on the CO: output of nymphs
of Melanoplus f. rubrum and Dichromorpha viridis. Abscissas represent tem-
peratures in degrees centigrade as indicated by the numbers. Ordinates repre-
sent rate of CO, output in grams COs, per gram total body weight per hour. For
further explanation see text.

used. Figure 5, plotted from results, shows the average rate of
CO, output at various temperatures, ranging from 0° to 38°C.

Examination of this figure shows that in general grasshoppers
respond to temperature changes as do other cold-blooded forms;
that is, Increased temperatures cause increased respiratory rates.
At 38° the rate of CO, output is highest and in the interval
from 0° to 15° it is nearly constant. This deviation from a
regular increase from 0° to 15° is difficult to explain and perhaps
is due to the imperfect control over conditions, such as body
movements, etc. However, careful observation during the course

WATER CONTENT AND RATE OF METABOLISM 159

of the experiments seemed to show no appreciable differences
in body movements at the various temperatures. It is interest-
ing to note to what extent these variations in respiratory rates
are directly influenced by the different temperatures, and if a
constant temperature coefficient, similar to that for other bio-
logical processes and chemical reactions, exists. As is well
known, temperature influences on the velocity of certain chemical
reactions can be satisfactorily expressed by the rule of van’t
Hoff, that for an increase in temperature of 10°C., the rate is
approximately doubled or trebled, i.e., there is a constant ratio,
Qo, of 3-2 for the rates at temperatures separated by an interval
of 10°C. Applied to the present results, Qio, varies considerably,
increasing with increasing temperatures, and is highest at 15°
to 25° (1.5), and lowest at 0° to 10° (1.1). These figures, it will
be noted, are somewhat lower than values obtained for chemical
reactions. Values obtained for other biological processes are
varied as the following examples show. Krogh (2/), in experi-
ments on the effects of temperature on the respiratory exchange
of the chrysalids of the meal-worm, finds that Qi for temperatures
from 10° to 30°C., varies from 5.7 to 2.0, being highest for lower
temperatures. Respiration in seedlings from 0° to 40°C. has
a value for, Qi) of 3 — 2, (Clausen, 23) and in a leaf 2.4 — 1.8,
(G. L. Matthaei, 24). Here, too, the values for Qi) are highest
at lower temperatures and decrease as the temperature increases.

There is no fundamental reason why the respiratory exchange
of an animal should follow the rule of van’t Hoff, since we are
dealing, not with a single chemical reaction, but rather with a
group of reactions, most complex in nature. Why the tempera-
ture coefficient for the respiratory exchange of the grasshopper
should be so much lower than that found for other forms is
difficult to explain. ‘Two plausible explanations suggest them-
selves, however. First, grasshoppers may possess some nervous
regulatory mechanism by which their respiratory exchange is
controlled and, secondly, the imperfect control over the animals
during the experiments might account for such results. No
such nervous mechanism is known to exist in insects, and if these
results were due entirely to imperfect control over the animals,

160 JOSEPH HALL BODINE

we should at least expect to find a much greater temperature
coefficient for the higher temperatures, since the animals would
then be most active. Much further investigation is necessary,
however, before any satisfactory conclusions can be drawn.

d. Effects of starvation. Adult Melanoplus differentialis were
starved, with water, for 96 to 120 hours and the rates of CO,
output measured during this period. Some eighty-five speci-
mens were individually studied and results given are taken from
selected cases showing typical conditions. Table 6 gives the
actual amounts of CO, given off and also the rate per gram per
hour for males and females of different weights.

An examination of this table shows that the actual amount
of CO: given off by an animal decreases during successive periods
of starvation. For example, the male weighing 1.0798 grams,
at the start gave off 0.000704 gram of CO., and at the end of
120 hours of starvation, only 0.00033 gram, a decrease of over
one-half of the original amount. Since we already know that
a loss in weight takes place during starvation, it is of interest
to find that a decrease in the rate of CO, output also occurs.
During the early period of starvation this decrease tends to be
rather slight and gradual, but at approximately forty-eight to
seventy-two hours marked drops are noted. This decided
decrease is doubtless due to the fact that at this time all residual
food in the intestine has been utilized and body reserves alone
are being used. Figure 6 shows that males have the higher rate
of CO, output and that these decreases are more marked for them.
This is perhaps closely related to the difference in size between
the animals of the two sexes. [It is evident, then, from these
results, that in grasshoppers as in other cold-blooded animals,
frog (Hill, 25), Planaria (Hyman, 26; Child, 19; Allen, 20),
CO, output decreases during starvation—at first rather rapidly
and later reaching a practically constant level up to the time of
death.

e. Effects of feeding starved animals. It is a well-known fact
that in higher forms, including man, ingestion of food after
starvation results in an increased rate of metabolism. Recently
Lund (27) has found similar results for Paramecium. Various

WATER CONTENT AND RATE OF METABOLISM 161

experiments with starved grasshoppers also show striking results.
Figure 6, taken from typical cases, shows the effect of feeding
animals sprouted oats after periods of starvation, varying from
48 to 120 hours.

It is evident, from figure 6, that feeding increases the CO,
output of starved animals. Some variations in extent of response

TABLE 6
Showing the actual CO, per one-half hour and the rate of COz output per gram per
hour, during starvation, for Melanoplus differentialis. Figures in italics repre-
senting COz in grams per gram total body weight per hour

WEIGHT STARVATION PERIOD TEM-
OF SEX | NORMAL PERA-
SSE 24 hours 48 hours 72 hours 96 hours 120 hours | TURE
grams
2.0500 | 9 Q.0006270)0.0004840/0.0005940)0 .. O006600/0. 0004510/0.0004400 1
: 0.0006116\0.00057 92|0.0006172\0.0006760'0 .0004832\0.0004730)| ~
1.0100 | ¢ 0. 0006380)0.0005390/0. 00072600. 0004620/0.0005830/0. 0005940 1
: 0.0012360)\0.0011170\0.0015120|0.0009150\0.0011770\0.0012310
1.8995 | 9 0. 0008030/0.0007590/0 .0008140/0. 0006160)/0.0005610)0. 0007040 21
; 0.0008454|0.0008116\0 .0009170\0.0007468)0.0007216\0.0007645|; ~
0.8995 | ¢ 0. 0007810)0.0006160/0.0005170|0. 0004840/0. 0003366 99
‘ '10.0017360)\0.0015020\0.0012840\0.0012100\0. 0008634 I *
0.7795 | o 0. 0006490)0.0004840/0 .0004840/0.0004180/0. 0002486 \ 29
; 0.0016650)\0.0014560\0.0013450\0.0012030)|0. 0007 106 i 2,
2.1200 | ° 0. 0010340|0 . 00100700. 0069790\0. 0007040/0.. 0005940|0. 0006270) } 99
: \ 0.0009754|0.0010410\0.0010300\0.0007650|0. 0006562)0 .0007 130 ji it
0.9645 | 2 0. 0006600)0.. 0006600/0. 00064460. 0004620)0 .. 0003740 ) 23
; | |0.0013680)\0 .0016290|0.0015720\0.0011840'0.0009906 f v3
1.9100 | ° 0.0013970\0.0010010)0. 0008860|0 .0006600/0. 0003300/0. 00033900) | 23
; 0.0014620\0.0013080)\0.0011220\0. 0009040)0 .0006970\0. 00067 40 ; %
1.0798 | ¢ 0.0007040)0.0007920)0 .0007100]0..0006600/0. 0005280 \ 1
, 0.0013030)\0 .0014940\0.0013930)\0.0013400|0. 0006004 if
1.2148 0. 0006160)0.0006380)0..0008200/0 . 0006820)0 ..0005500/0. 0005280 1
‘ cs 0.0010140\0.0011650|0.0016750\0.0013640|0.0010420\0.0010770|| ~

162 JOSEPH HALL BODINE

occur, but generally it has been found that the rate of output
is approximately doubled three hours after feeding. The effects
of a single feeding, however, last but a short time, depending
upon the amount of food eaten. No detailed study of the effects
of different amounts of different foods has been made, but an
animal starved for forty-eight or more hours and then fed always
shows an increased output. This increase gradually rises as
the weight increases until the animal gains its normal weight
relations. Such results, showing that starved grasshoppers
respond to ingestion of food by increased production of COu.,
agree with those for other forms, and especially with those of
Allen (20) for Planaria.

«0020

| CO2 p.gm. p. hr.

y t Time-HRS.
5
Se ——t
'N 24 48 72 | 96 120 Maa 168

Fig. 6 Curves show decrease in CO: output by Melanoplus differentialis dur-
ing starvation and increase after feeding. Abscissas represent time in hours
indicated by numbers. Ordinates represent rate of CO. output in grams COs:
per gram total body weight per hour. Time at which feeding was begun indicated
by arrow. See text for further description.

CONCLUSION

The results of the present study, as presented above, seem to
indicate the extent to which comparisons between some of the
physiological phenomena of insects and mammals can be made.
It is found that the percentage of water an animal contains is
characteristic for the particular species, and that it decreases
with age and increasing body weight. When exposed to low
temperatures, the animals respond by a decrease in water content
and are thus prevented from freezing and possible destruction.

WATER CONTENT AND RATE OF METABOLISM 163

Starvation results in marked and rapid losses in body weight,
water, and solids, but the greatest and quickest loss seems to be
of water. Closely correlated with these losses is a decrease in
the rate of CO, output. Various species of animals seem to have
different rates of respiratory exchange, but all show a higher
rate for the younger individuals. Increased temperatures cause
increased rates of CO. output, while lower temperatures seem
to have the reverse effect. Ingestion of food by starved animals
greatly increases the rate of CO. production. By a comparison
of these data with those found for mammals, striking similarities
are found to exist, and these would seem to indicate that the
problem of insect physiology, although at first seemingly unre-
lated to that of mammals, has, in fact, many points in common
with it.

SUMMARY

1. The percentage of water an animal (grasshopper) contains
decreases with age and increasing body weight, up to a minimum
for the species.

2. Different species of the same general group, living upon
similar foods, may have different percentages of water.

3. During the active life-cycle of Chortophaga viridifasciata,
the water content falls to a minimum during ‘hibernation,’
rises again to a maximum when ‘hibernation’ is broken up, and
then again falls to a minimum as the animal grows old. These
changes seem to be due to the effects of temperature and advanc-
ing age.

4. Water and temperature are the controlling factors in
Chortophaga’s emergence from ‘hibernation.’

5. Different species of grasshoppers studied, under similar
conditions, survive starvation for different periods of time, e.g.,
Melanoplus differentialis, 172 hours; Melanoplus f. rubrum, 144
hours, and Chortophaga viridifasciata, 170 hours.

6. Starvation results in losses of body weight, water, and
solids, the greatest relative loss being of water. With water
alone, losses are lower than with nothing.

' 7. Larger individuals tend to lose relatively greater amounts
during starvation.

164 JOSEPH HALL BODINE

8. Rates of CO, output differ for the different species of animals
studied.

9, Lighter and younger animals have the higher rates of CO,
output, and the possible relation of a surface law holding for
grasshoppers is indicated.

10. Temperature influences on the CO, production are rather
marked, higher temperatures cause increased rates of CO: output
and lower temperatures tend to have the reverse effect. How-
ever, the temperature coefficients for these different temperatures
are variable and are also considerably lower nea those found for
other biological processes.

11. Starvation causes a decrease in the rate of CO, output.

12. Feeding starved individuals results in an increase in the
rate of CO, output.

LITERATURE CITED

1 Curip, C. 1915 Senescence and rejuvenescence. Chicago.
2 Marsews 1916 Physiological chemistry, 2nd ed. New York.
3 Minor 1908 Age, growth and death. New York.
4 Harar 1917 Am. Jour. Anat., vol.21, p. 23.

5 Luacer 1897 Orthoptera of Minnesota. .

6 Morse 1896-97 Psyche, vols. 7 and 8.

7 Lunp 1918 Biological Bulletin, vol. 34, p. 105.

8 BREITENBECHER 1918 Carn. Inst. Pub., no. 263

9 Bascock 1912 Research Bull., Agr. Exp. Stat., Univ. Wisconsin, 19 to 24.
10 Donatpson 1911 Jour. Morph., vol. 22, p. 663.

11 Dvurour 183838 A l’Acad. d. Sc., T. 4, p. 129.

12 Rivey AND JoHANNSEN 1915 Handbook of Med. Ent. Ithaca, New York.
138 Harar 1917 Carn. Inst. Publ., no. 251.

14 Morevuus 1911 “Arch. f. Entw. d. Org., Bd. 32, 8. 169.

15 Donaupson 1911 Jour. Comp. Neur., vol. 21, p. 139.

16 Vernon 1897 J. Physiol., vol. 21, p. 443.

17 BrNEpDIcT AND EmMmes 1915 J. Bio. Chem., vol. 20, p. 253.

18 Rusner 1883 Zeit. f. Biol., Bd. 19, S. 536.

19 Curitp 1919 Am. J. Physiol., vol. 48, p. 231.
20 Auten 1919 Am. J: Physiol., vol. 49, p. 420.
21 Krocu 1916 The respiratory exchange of animals and man. Longmans,

Green & Co., London and New York.

Meen 1879 Zeit. f. Biol., Bd. 15, 8. 425.

CriauseN 1890 Landwirtschaftl. Jahrb., Bd. 19, s. 893.

Marrnuarr 1904 Trans. Philos. Soc. London, series B, vol. 197, p. 47.
Hitt 1911 J. Physiol., vol. 43, p. 379.

Hyman 1919 Am. J. Physiol., vol. 49, p. 377.

Lunp 1918 Am. J. Physiol., vol. 47, p. 318.

bo hy hw bh} bb bv
IQ oe & bP

ad ilk milous eit um ia ww} ieeaelt 1"

ae pon? ga sl ae
Chie c 2S Ib
}. me ry Ho) oe Diaehh) Se y

oO
he

“ot

a Siete wie a > hig) 1 es Ge sey

:, é :
“ee tae a ie ( ti) re arent i on

, pit as, oe doth is a ad oti
wi Ging = Sheed byte ae
ts La afditiaahe Fs |
- b * es. AS Js
: rat Ot Te he Rode be

ban a ta)
i 7 '
ie: Sts:

ee ration Ee:

ais pays ig 55,
Ny ‘ wicca eee y
a 4 sf \
in Pinal ie Pa Wig ieee
NG es sea seat Te i pes reat az,
viata FA Mght iD: het pi 0) ean,

RAS ria a chit Rae
i" 3 ; spent ee le i wee Pr APs ie {

Mee

eeriy

* airy

Resumen por el autor, Charles W. Metz,
Estacion de Evolucién Experimental,
Carnegie Institution of Washington.

Espermatogénesis de la mosca Asilus sericeus Say.

Asilus sericeus Say es una especie de Diptero sumamente
favorable para los estudios sobre la espermatogénesis a causa
del considerable numero de células y de la exacta serie de estados
de crecimiento en el testiculo, asi como la ausencia relativa de
estados confusos durante el periodo de crecimiento. Existen
cinco pares de cromosomas. La asociacién de los cromosomas
en parejas se presenta en la espermatogonia y se retiene durante
la ultima divisién espermatogonial. En la telofase de esta
division, la disposicién de los cromosomas en parejas es tan
intima que realmente equivale a la sinapsis, y la unién que tiene
lugar en este momento persiste durante el periodo de crecimiento
que la sigue.

Los autores no han podido encontrar estados leptoténicos o
zigoténicos (sindpticos) propiamente dichos. Los cinco ele-
mentos dobles (bivalentes) que aparecen en la telofase permane-
cen relativamente condensados y son facilmente discernibles
durante todo el periodo de crecimiento hasta la primera divisién
de los espermatocitos. La sinapsis precoz y la omisién del
estado leptoténico ordinario pueden interpretarse como una
manifestacién de la fuerza que origina la asociacién por parejas
caracteristica de los cromosomas de la espermatogonia, oogonia
y células somaticas de los dipteros.

Translation by José F. Nonidez,
Cornell Medical College, New Yosk

AUTHORS’ ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, NOVEMBER 1

SPERMATOGENESIS IN THE FLY, ASILUS SERICEUS
SAY!
CHARLES W. METZ anv JOSE F. NONIDEZ
Station for Experimental Evolution, Carnegie Institution of Washington

TWO PLATES (TWENTY-TWO FIGURES)

CONTENTS
rte OC UT Ora! St Ss |S TSUN OE OU ae Se ceOmCene Mans od LEB OY, Soe he 165
PR SCHIITORUG Pr, 308 ans 212 2s Cepke SAe aa, Ca yd ee ees oe (ERD ble a. al tee 167
POMORLA DLN CAL TEATUTES Eis ao 0 5% sterajureral EO, Seana To aks cclei fs ol dend Reawavers 167
SS OCEELEINEWI{0 00 mr eee gor Re Ck Te ea ae 168
Bimaltspermatozonial telophasess 2.272 0.) tere tetas aioe ew kee ce 169
ME lye Ow bh period :...\..01. 4 Yhd ess Wa ne cae REA Geis ok cena edu eee 170
atemerow tl perrod,. («2d iesuncse as aoe tothoeiee BART ee Ee Peed are oS RTE histoate 174
Wher speEMmabocy ber CiVISIONS Ani.) ee crsee, ale er Ieee cheese eae. Gas 175
ETG: THEOL SOT geen ali a Rea PE He eo ae dn et ice 176
IDWGUSES 0; Ae Sea ee eam ee SRR a ee ee, ch hte eS See 5 eee 177
SLUIACUN Tg cee ao ee ee eee Fn =e ee 179
ther aiaire rented «Avis toe cocks 1 Ban Die Wen ne yet ONE Re eR, AbATy HME ce ch 180
INTRODUCTION

The rapid development of genetical studies on Drosophila in
the last few years has drawn considerable interest to the subject
of gametogenesis in the Diptera. The phenomena of mutation,
crossing over, nondisjunction, ete., occur partly or wholly during
gametogenesis, making it desirable that our knowledge of the
latter be extended as far as possible. Another feature that has
added. interest to the subject is the characteristically ‘paired
association’ of the somatic chromosomes of flies as distinguished
from those of other animals.2. It seemed possible that this pre-
existent association of chromosomes in the soma and early germ
cells might affect the processes of maturation in such a manner
as to throw additional light on their significance.

1 We are indebted to Mr. C. W. Johnson for identifying the species used.

* For discussion of this phenomenon, see Metz, ’16.

165

166 CHARLES W. METZ AND JOSE F. NONIDEZ

These considerations led the senior author some time ago to
begin gathering material for a study of gametogenesis, in con-
nection with other studies on Diptera chromosomes, and the
first results of this study are presented here. From the genetical
standpoint the present paper represents only a short step in the
desired direction, but it may serve as a foundation upon which
to base further work—particularly studies on odgenesis, which
are under way at the present time.

Two reviews of the literature on Diptera chromosomes have
been given recently (Metz, 716, pp. 213, 214; Whiting, 717).
From these it will be seen that most of the work on the subject
has dealt primarily with the sex chromosomes or other special
features, and that our knowledge of gametogenesis, epee
the growth stages, is meager.

In her studies on the sex chromosomes, Stevens (’07, ’08, 710,
11) records several observations on other aspects of spermato-
genesis, but unfortunately these cannot be combined to make a
connected account. The observations of Taylor (’14) and of Lomen
(14) on Culex, in addition to being meager as regards details of
spermatogenesis, are, we believe (see also Whiting, °17), faulty
on account of the poorly fixed material used. Whiting (17), in
a more recent paper on Culex, has given a comprehensive account
of the maturation divisions, beginning with the first spermato-
cyte prophase. His observations on the earlier stages, however,
particularly the earlier part of the growth period, are limited,
probably owing to the fact that Culex does not afford favorable
material for this purpose.

As regards odgenesis in the Diptera, practically nothing has
been published, so far as we are aware.

Unfortunately, the Diptera have long, and justly, been looked
upon as unfavorable objects for cytological study—a fact that
has undoubtedly been responsible for keeping our knowledge of
gametogenesis in this group far behind that of such insects as
the Orthoptera, Hemiptera, and Coleoptera. We have found,
however, that some groups of Diptera are much more amenable
to study than others, and by making selections from these and
by careful attention to technique we have been able to obtain

SPERMATOGENESIS IN ASILUS SERICEUS 167

favorable material. It is hoped that the results obtained from
this will eventually permit of a satisfactory analysis of the pro-
cesses in the difficult Drosophila material.

Our survey of the Diptera has not yet revealed any single
species that is favorable for a study of all stages of spermato-
genesis, but the species considered here combines more favorable
features than any other. In this form the seriation of stages
through the development of the first spermatocyte up to the
maturation divisions is clear-cut and practically complete.
Details of certain stages are not depicted with as great clearness
as they are in some other species (to be considered subsequently),
but in most respects the material is unusually favorable.

TECHNIQUE

Flemming’s strong solution has been found most satisfactory
for fixation and has been used almost exclusively. Heidenhain’s
iron haematoxylin and safranin have been used for chromatic
stains, with or without counter-stains. A more detailed con-
sideration of technique has been given in an earlier paper (Metz,
16, p. 219). All of the accompanying figures are from material
fixed in Flemming and stained in iron haematoxylin.

TOPOGRAPHICAL FEATURES

The testes in Asilus, like those in most other asilids, are a
pair of long coiled tubes, each containing thousands of cells.
The distal end of the testis contains a clearly marked spermato-
gonial region with a central core of giant nutritive cells. Then
comes a narrow transition zone in which spermatogonial ana-
phases and telophases are intermingled with the earliest stages
of the first spermatocytes. The nuclei here are very small.
Following this is a broad zone, containing cells, nearly all of
which are in one stage. From this point the development
involves a gradual transformation, which may be followed with
comparative ease through the long growth period extending far
down the testis. In favorable material all of these stages appear
in One preparation, or even in one section, and since there is such

168 CHARLES W. METZ AND JOSE F. NONIDEZ

a large number of cells available for study, even the intermediate
transition stages, often so difficult to obtain, are nearly all in
serial order. The only exceptions to this rule are found in the
stages immediately following the last spermatogonial anaphases.
Here development proceeds with great rapidity, and the telo-
phase and associated stages are more or less intermingled. But
in spite of this it is possible to obtain a clear conception of what
takes place, for the immediately succeeding stage (6) is perfectly
clear, and with its aid the other figures may be pieced together.
The details of these processes will be given below.

SPERMATOGONIA

The spermatogonia are abundantly represented in our material
and are of ample size for study. Those of the last generation or
two are noticeably smaller than the earlier ones and have the
chromosomes more closely aggregated in metaphase, but other-
wise are not appreciably different from the rest. In a previous
paper (Metz, ’16) the general peculiarities of chromosome behay-
ior in spermatogonia and other diploid cells of flies have been
described in detail. For this reason the spermatogonial stages
will be passed over briefly here. It is important, however, to
keep in mind the fact that the close association of homologous
chromosomes, especially in the resting stages and prophases,
makes the pairs of chromosomes simulate single chromosomes
of other animals.

In spermatogonial metaphases of A. sericeus there are ten
chromosomes, arranged in five pairs (figs. 3 and 4), the smallest
of which is probably the sex chromosome pair, although there
is no evident dimorphism to distinguish it. Occasionally the
arrangement of one or two pairs is disturbed, but normally they
all show the paired association just as in other Diptera (Metz,
716). In anaphase the chromosomes pass to the poles in this
paired condition (fig. 5). Their behavior during the telophase
will be discussed below.

During most of the resting stage the chromatin is diffuse and
stains so faintly that its behavior cannot be observed satis-
factorily. In prophase it becomes aggregated, and each aggre-

SPERMATOGENESIS IN ASILUS SERICEUS 169

gate gives rise to a long double thread by a process of attenu-
ation or uncoiling (fig. 1). This thread shortens up immediately
into a pair of prophase chromosomes (fig. 2).8

FINAL SPERMATOGONIAL TELOPHASES

Since the telophase of the last spermatogonial division is a
stage of particular interest, it may be considered separately.
In the last spermatogonial anaphase, as in preceding anaphases,
the chromosomes go to the poles associated in pairs. In late
anaphase the pairing becomes more intimate, due partly to the
crowding of the chromosomes at the poles. Then in telophase
the crowding is relaxed, the cluster loosens up, and the individual
chromosomes may be observed. They are now intimately
associated in pairs—so intimately, indeed, that the duality is
often obscured. In other words, as the cluster loosens, the
chromosomes separate out as bivalents instead of single elements.
In figures 6 and 7 different degrees of association are represented.
Some of the chromosomes show the dual structure clearly, while
others show it very little or not at all. These nuclei are entire,
or nearly so, and all of the chromatin is represented. Figure
8 is from a slightly later stage in which the paired association
is so intimate that all trace of duality is gone, and only five
chromosomes can be detected.

The staining capacity of the chromatin is greatly decreased
at this time and the chromosomes appear less bulky than before.
Needless to say, such figures are difficult to analyze, but a careful
study has convinced us that the process is as described—that
homologous members join in early telophase and effect an inti-
mate union side by side. This conclusion is based not only on
the duality of the telophase chromosomes, but on the fact that
they are haploid in number (5) instead of diploid (10). The
cells are small, affording plenty of examples of uncut nuclei,
and in no case have we been able to find one in which the chro-
matic bodies approached the diploid number. Indeed, we found
no clear case in which more than five were present.

® These features, together with other details omitted in the present paper, will
be considered more fully in a subsequent publication.

170 CHARLES W. METZ AND JOSE F. NONIDEZ

Probably the density of the stain or degree of extraction has
a good deal to do with the appearance or non-appearance of the
duality in these telophase nuclei, but there can be no doubt that
the union is very intimate. In this relation the chromosomes
pass (fig. 9) into the succeeding stage in which they lose their
staining capacity to a much greater extent, as they enter the
growth period. A careful scrutiny of the late telophase nuclei
reveals very little indication of a spinning-out process or a net-
work formation, except that due to the linin. The chromosomes
appear simply to fade out through loss of color, while retaining,
approximately, their form and position (fig. 9).

It is probable that the above account should not be restricted
to the final spermatogonia, but should apply to’all of the sper-
matogonial telophases. The evidence points consistently in
that direction (Metz, 716), but we have not been able to make
sure of the point in the species under consideration.

THE ,.EARLY GROWTH PERIOD; STAGES A AND B

Following the final spermatogonial telophase there is a very
brief period during which relatively little chromatin is visible
in the nucleus, as indicated by figure 10. This stage, which may
be called stage a, is also characterized by the appearance of a
small nucleolus, as shown by the figures. The nuclei of this
period, together with those of the telophase just preceding, are
the smallest to be found in the testis and cannot be confused
with those of any other stage.

Apparently our stage a corresponds to Montgomery’s stage
a in the Orthoptera and Wilson’s stage a in the Hemiptera (see
below). So far as we can determine, it is structurally similar
to the early resting stage of the spermatogonia. The stage is
so brief that it is only represented by a few scattered groups of
cells at the border-between the final spermatogonia and the
clearly marked region in which the next stage (stage 6) is
represented.

Adjacent to the nuclei of stage a (fig. 10) are others only slightly
larger in which the chromatin becomes progressively more deeply
stained and condensed, revealing the outline of the chromosomes.

SPERMATOGENESIS IN ASILUS SERICEUS CE

These are double and haploid in number, corresponding to the
telophase pairs. The size of the cells and nuclei indicate that
the actual growth period has barely begun when these bodies
become visible (fig. 11), and there seems to be little doubt that
the preceding transition from the telophase stage has not only
been very brief, but has involved little change in the chromo-
somes other than that involved in the loss of staining capacity.
The intimate association in pairs appears to have remained
unaltered.

The chromatin now becomes further condensed, revealing the
size and shape of the bivalents in a more clear-cut manner (figs.
12 and 13). Attached to one of these (apparently the smallest
pair) is the nucleolus, the history of which will be considered
later. This stage may be designated stage 6. Structurally
stage b resembles the late resting stage or early prophase of the
spermatogonia in which the chromatin becomes condensed into
five bivalent aggregates that give rise to the prophase chromo-
somes.*

Since stage b forms a definite point of orientation between
the brief early stages and the more extended later ones, it may
be well to consider events up to this time before describing the
subsequent processes. It is apparent that the synaptic condition
has been fulfilled at the very beginning of the growth period by
the intimate association in telophase of chromosomes that were,
for the most part,® already arranged in pairs. Technically, this
association should be called synapsis, for, as will be seen, the
union effected in telophase persists throughout the growth period
and is responsible for the formation of the bivalent chromosomes
of the first maturation division.

Compared with the corresponding stages in other animals,
this behavior seems to be unique, and it seems legitimate to
infer that it is associated in some causal manner with the other

4 See footnote 3, page 169.

5 It is not justifiable to assume that the paired association in anaphase is
absolute and invariable, for occasionally the two members of a pair may be sep-
arated in metaphase (e.g., fig. 4), and consequently in anaphase. It must be

assumed, however, that in these exceptional cases the paired arrangement is
restored in telophase or soon thereafter.

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 1

ily CHARLES W. METZ AND JOSE F. NONIDEZ

peculiarity of the dipteran chromosomes—their characteristically
paired association in somatic cells.

It is of interest to consider the events up to this point in relation
to those found in other insects, such, for instance, as the Hem-
iptera, Odonata, and Orthoptera. A marked similarity is at
once noticeable in many features, but always with the difference
that the chromatic elements of Asilus are double instead of single.

Thus in Oncopeltus Wilson (’12) describes the stages immedi-
ately following the final spermatogonial divisions as involving
a diffusion of the chromatin in telophase followed by a stage in
which definite prochromosome-like aggregates arise (compare
Wilson’s figures 48 to 51 with our figures 9 to 13.)° These aggre-
gates or masses would correspond to those of stage 6 in Asilus,
but instead of being haploid in number and bivalent in compo-
sition, they are, in Oncopeltus, diploid in number and apparently
univalent in composition.

In Lygaeus among the Hemiptera (Wilson, 712), Anax among
the Odonata (Wilson, 712), Phrynotettix, Dissosteira, and
Chortophaga among the Orthoptera (Davis, ’08; McClung, ’02;
Wilson, 712), and probably in numerous other forms, phenomena
not essentially different from those in Oncopeltus are found, so
that the comparison of Asilus with Oncopeltus may be extended
to include several species representing a widespread type of
spermatogenesis as regards the earliest stages of the growth
period.

Apparently the Coleoptera may also be put in this class,
although there are so many conflicting accounts of coleopteran
spermatogenesis that many cases are open to question. The
essential features, however, namely, the resting stage followed
by the appearance of more or less condensed masses or aggregates
in diploid number, seem to be well established in certain instances
(Stevens, 05, ’06, 08 a, ’09; Nonidez, 714, 715; Goldsmith, 719,
figs. 17, 18, 19).

6 This resemblance is even more strikingly shown by another species of Asilus

(A. notatus) in which the prochromosome-like bivalents are more condensed and
shorter than in A. sericeus (discussion, page 178).

SPERMATOGENESIS IN ASILUS SERICEUS 173

One author (Arnold, ’08) has described in Hydrophilus piceus
(Coleoptera) a precocious reduction of the chromosomes at the
beginning of the growth period not unlike that found in Asilus.
But the brevity of his description together with the fact that no
other observer (Stevens, Vom Rath, 92, Goldsmith, ete.) has
noted such a phenomenon in this or other Coleoptera makes it
seem probable that Arnold is mistaken in his interpretation.

It appears, then, that although a superficial similarity exists
between the early growth stages of Asilus and those of various
other animals, the divergence between the double (bivalent)
chromatic bodies on the one hand and the single ones on the
other separates the representative of the Diptera from all the
other forms.”

If we turn to the plants, however, we find a different situation.
Here, although the evidence is not as clear as might be desired,
some species appear to exhibit a paired association of “‘prochromo-
somes,’ in the early growth period immediately after the last
diploid telophase, somewhat like that found in Asilus. Overton
(05, 09), for instance, records such an association in Thalictrum,
Calyeanthus, Campanula, and Helleborus. In these the last
diploid division is followed by a resting-stage network in which
definite chromatic bodies (prochromosomes) are scattered about.
These are diploid in number, but often, or usually, lie in paired
association. Their shape and degree of condensation differ in
different cases, but their paired association seems to be fairly
constant. In these plants the association may persist from the
last ‘premiotic’ anaphase through the growth period and up to
the metaphase of the reduction division, although Overton does
not commit himself as to the behavior in the telophase and
earliest stage of the growth period, as indicated by the following:

7 The earlier literature of spermatogenesis contains numerous references to
possible or supposed precocious pairing in the last spermatogonial telophases.
For instance, Montgomery, ’00, page 297, on Peripatus, notes a few such ‘excep-
tional’ cases; Blackman, ’03, 710, page 141, on Scolopendra, describes a pre-
cocious telosynapsis; Stevens, ’03, on Sagitta, suspects an early pairing, and
Downing, ’05, on Hydra, makes a similar suggestion. Other and more recent
examples could be cited also, but in no case have we been able to find clear-cut
evidence of such an association as is exhibited by Asilus.

174 CHARLES W. METZ AND JOSE F. NONIDEZ

I have attempted to trace the processes of reconstruction of the
nucleus of the pollen mother-cells from the last pre-meiotic division,
and to compare the structure of these nuclei with that of ordinary
somatic ones, but have experienced considerable difficulty in identify-
ing with certainty the last pre-meiotic divisions. After the formation
of the nuclear membrane and during the period of nuclear enlargement,
the chromatic material becomes rather regularly distributed in the
nuclear cavity, the greater portion of the staimable substances lying in
the prochromosomes, each suggesting by its form and size that it is
derived from a chromosome of the preceding telophases. I am not
prepared to discuss the problem as to how the chromosomes of the
telophases are modified in passing over into the resting nucleus.
(Overton, ’09, pp., 21, 22.)

In Oncopeltus and the other insects mentioned above, the
prochromosome-like bodies of stage b, whether massive (Wilson’s
stage b in Hemiptera and Anax) or more thread-like (Davis’s
stage b in Orthoptera), give rise, by a process of unraveling, to
long, delicate leptotene threads that then undergo synapsis to
form the pachytene or diplotene threads. Since, in Asilus, the
chromosomes are already double (i.e., bivalent) it is of especial
interest to examine their subsequent behavior.

LATER GROWTH PERIOD

The transition from stage b to later stages involves merely a
gradual lengthening out of the five diplotene threads (figs. 13
to 16) and their polarization with reference to the nucleolus.
One member (apparently the smallest) is already attached to the
nucleolus. The others, or at least two or three of them, soon
become attached and extend out like fingers (figs. 16 to 20).
Apparently each thread becomes attached at one end only. No
eases have been found in which a complete loop was formed.
Fortunately, the threads lie close to the nuclear wall and remain
well separated from one another throughout almost the entire
growth period, so they may be examined readily. They show
no indication of dissociation into single (leptotene) threads at
any stage, although their duality is evident throughout. As
may be noted from the figures, the nucleus decreases somewhat
in size instead of enlarging as polarization progresses.

SPERMATOGENESIS IN ASILUS SERICEUS 175

The polarized stage persists almost up to the first spermatocyte
prophase, and is modified, toward the end, by a definite contrac-
tion period (fig. 17) in which the threads draw away from the
nuclear wall and lie close together. Apparently no significance
attaches to this contraction for the threads undergo no visible
changes and soon spread out again into their previous positions
near the periphery (figs. 20 and 21) and condense into the five
prophase chromosomes, ready to go on the spindle.

Although these processes cover about four-fifths or more of
the growth period and are represented by many thousands of
cells, they are so simple and involve such slight changes in the
chromatin, that in essentials the condition found in stage b
(fig. 13) may be said to typify all the succeeding stages up to
the prophase, and the whole series may be represented by a few
figures. The diplotene threads that appeared at the beginning
of stage b have persisted unchanged so far as their diplotene
structure is concerned. The contraction stage, occurring in the
late growth period, if it has any counterpart, outside of the
Diptera, would represent the so-called second contraction, taking
place long after synapsis.

These events seem to resemble those in Thalictrum (Overton,
09) to the extent that the chromatin remains in the form of
relatively condensed, bivalent threads. Compared with animals,
however (other than Diptera), there is no such resemblance, for,
as just mentioned, the leptotene and synaptic stages usually
following stage b are not found in Asilus.

THE SPERMATOCYTE DIVISIONS

Since this paper is concerned primarily with the growth stages,
the maturation divisions will be passed over briefly. So far ~
as known, they present no unusual features. Metaphases of
both spermatocyte divisions are clear, and each shows five
chromosomes. Only the first is represented here (fig. 22). It
is the reduction division, apparently, for no tetrad structure is
evident.

176 CHARLES W. METZ AND JOSE F. NONIDEZ

THE NUCLEOLUS

The history of the nucleolar structures has not been studied
in detail. but the nucleolus is so prominent during the growth
period that a study of the chromosomes must necessarily reveal
the main features of the nucleolar behavior. It is probable
that the chromatic part of the nucleolar complex persists from
the final spermatogonial anaphase in the form of a pair of chromo-
somes, but whether the achromatic portion arises from this or
originates independently we are unable to state. The two are
united from stage b (fig. 11) throughout the remainder of the
growth period. The chromatic portion may be followed directly
to the first spermatocyte metaphase where it becomes one of the
five bivalent chromosomes. During much of the growth period
the nucleolus is plainly compound (fig. 14), being composed of a
large, oval achromatic portion and a smaller dense chromatic
portion to which is attached a chromatic thread or finger-like
projection. The latter is very characteristic and persistent
throughout the growth period. The achromatic portion seems
to diminish gradually during the later stages, and cannot be
detected with certainty in late prophase. However, the degree
of extraction of the haematoxylin has much to do with the appear-
ance of the structure, and it is difficult to say just what becomes
of the achromatic portion.

The chromatic portion is presumably the sex chromosome
pair. At first sight the finger-like process suggests the presence
of an unequal XY pair, but this asymmetry seems to disappear
in metaphase and we are unable to verify the point. Likewise,
the spermatogonial divisions do not reveal any such inequality
in any chromosome pair. It seems more probable that the
finger-like process is due to a difference in the degree of conden-
sation of the two chromosomes, such, e.g., as that shown by the
XY chromosomes of Enchenopa binotata (Kornhauser, 714).

In another species of Asilus it is practically certain that the
chromosomes involved in the nucleolar complex are the sex
chromosomes, and it may be inferred that the same is true in
A. sericeus. This is in agreement with the observations of
Stevens (08 b), who found the sex chromosomes condensed
during the growth period in several species of flies.

SPERMATOGENESIS IN ASILUS SERICEUS 177
DISCUSSION

Asilus sericeus presents the most simple and clear-cut type of
spermatogenesis thus far found in the Diptera, owing to the
fact that during the growth period the chromatic threads do
not spin out and become entangled to such a degree as they do
in other forms.

When compared with animals other than Diptera, the most
outstanding characteristic of the maturation processes in Asilus
is the apparently continuous association of corresponding chromo-
somes in pairs. Superficially some of the stages bear a marked
resemblance to those in various other forms, but on close exami-
nation it appears that only the later growth period and succeeding
stages are actually similar in essential features. Previous to
this there js an underlying difference due to the fact that in
Asilus the chromosomes, whether condensed or thread-like,
maintain an intimately paired association from the telophase
throughout the entire growth period, with the result that the
usual leptotene stage and the subsequent synaptic process seem
to be omitted entirely. This is discussed more specifically above.

Another feature that should be recalled here is the probable
parallelism between the peculiarities in chromosome behavior
observed in Asilus and those found in certain plants as recorded
by Strasburger, Rosenberg, Miller, Overton, and others (see
especially Overton, ’09). Apparently the peculiarities during
the maturation stages are in each known case correlated with a
noticeable paired association of chromosomes in the somatic
cells, which would again lead one to conclude that the two
phenomena are causally connected and are both manifestations
of the same inherent ‘tendency toward pairing.’ As has been
remarked previously (Metz, 716, p. 225), the latter seems to be
an accentuation of the tendency or force that unites correspond-
ing chromosomes during synapsis in most organisms. It seems
to differ mainly in that its effects in the cases mentioned are not
limited to the final germ cells, but are visible in somatic and early
germ cells as well. What this force is, physicochemically, remains
as obscure as ever, although there is very strong reason for believ-

178 CHARLES W. METZ AND JOSE F. NONIDEZ

ing that it is due to a likeness in constitution of corresponding
chromosomes.

Regarding the genetical question of ‘crossing over,’ our obser-
vations afford only negative evidence. Since no leptotene
threads have been observed and nothing like a typical synaptic
stage has been identified, there is little indication of any process
that might bring about crossing over during the early stages.
It is difficult to determine just what takes place in stage a, but
it should be recalled that this stage appears to be like the sperma-
togonial resting stage, and that as far as cytological evidence
goes there is no more reason for expecting it in the former than
in the latter.

In subsequent stages there is some evidence of chromosome
twisting, but more often the threads lie side by side without
twisting, and when they do overlap there is no evidence of break-
ing. On the whole, then, what evidence there is would argue
against the probability of crossing over in the males of Asilus,
which agrees with the genetical results in Drosophila, where
crossing over is found only in the female. But this question,
from the cytological standpoint, is only in the speculative stage,
and will probably remain there at least until further studies are
completed, particularly studies on o6genesis.

In this connection a word should be said regarding the degree
to which the above description may be considered typical of the
Diptera. One other species of Asilus (A. notatus) has been
studied fully and shows certain noteworthy deviations from the
above account. ‘These may be summarized briefly as follows:
In stage a following the final spermatogonial telophase, the
chromatin stains more deeply than in A. sericeus and gives even
clearer evidence of remaining relatively condensed, i.e., not
spinning out into threads. Stage a is very brief and is succeeded
immediately by stage b, in which the chromatin is likewise more
dense than that in the corresponding stage of A. sericeus. It
is in the form of short, thick, bivalent prochromosome-like
bodies, the dual nature of which is very plain. These show a
more marked superficial resemblance to the bodies of stage 6
in the Hemiptera than do those of A. sericeus.

SPERMATOGENESIS IN ASILUS SERICEUS 179

The most noticeable difference between sericeus and notatus,
however, appears in the stage immediately following stage ).
At this time the bivalents in A. notatus, instead of lengthening
only slightly and remaining well separated from one another, as
they do in sericeus, become greatly attenuated and entangled
for a time, making analysis very difficult. Here again the super-
ficial resemblance to phenomena in the Hemiptera is more marked
than in sericeus, although the actual structural characteristics
(persistence of the diplotene condition) appear to agree with
those of sericeus.

When an attempt is made to compare spermatogenesis in
Asilus with that in other genera of flies, confusion enters at
once. Other members of the Asilidae show definite resemblances,
but outside of the family superficial differences are so great that
comparisons cannot be made safely without very careful study.
As a case in point we may mention the genus Drosophila. Super-
ficially spermatogenesis in this group is exceedingly different in
appearance from that in Asilus. Further study and possibly
detailed examination of intermediate forms will be necessary
before the relationships can be determined. Perhaps much of
the apparent divergence between Drosophila and Asilus is due
to difference in the cytoplasm, rate of growth of the spermato-
cytes, degree of staining of the different nuclear elements, and
other secondary features, but there is as yet no certainty that
it may not also include fundamental differences in the chromo-
somal behavior.

SUMMARY

1. The spermatogonial chromosomes of A. sericeus are ten
in number, arranged in five pairs. The sex chromosomes have
not been identified.

2. In the last spermatogonial anaphase, as in preceding ana-
phases, the chromosomes go to the poles associated in pairs.

3. The paired association becomes more intimate in telophase,
giving rise to bivalent chromosomes in haploid number.

4, A brief diffuse stage (stage a) ensues, in which the chromatin
stains only slightly.

180 CHARLES W. METZ AND JOSE F. NONIDEZ

5. Then the double chromosomes reappear, apparently in the
same form and relative position as before, and condense into
bivalent prochromosome-like bodies (stage 6).

6. The ordinary leptotene condition seems to be omitted
entirely.

7. The bivalent bodies of stage b elongate into diplotene
threads that remain relatively condensed and clearly separate
throughout the entire growth period, giving rise to the bivalent
prophase chromosomes.

8. In another species, A. notatus, the process appears to be
essentially the same, but is somewhat confused by a spinning
out and intertwining of the threads in the stage following stage b.

9. The usual synaptic process is entirely wanting. Synapsis
is effected in telophase at the beginning of the growth period by
an intimate association of chromosomes that were already paired
in anaphase.

10. Superficially the early growth stages are not unlike those
in the Hemiptera and other forms, but the chromatic structures
are bivalent instead of univalent.

11. Tetrad structures are not visible.

12. The first division appears to be reductional for all of the
chromosomes.

LITERATURE CITED

ArnNotp, G. A. 1908 The nucleolus and microchromosomes in the spermato-
genesis of Hydrophilus piceus. Arch. f. Zellfor., Bd. 2, No. 1.

BriackMaNn, M.W. 1903 The spermatogenesis of the myriapods II. Biol. Bull.,
vol. 4; pp. 187-218.
1910 The spermatogenesis of the myriapods VI. Biol. Bull., vol. 19;
pp. 138-160.

Davis, H. 8. 1908 Spermatogenesis in Acrididae and Locustidae.. Bull. Mus.
Comp. Zool. Harvard, vol. 52, no. 2.

Downinea, E. R. 1905 The spermatogenesis of Hydra. Zool. Jahrb., Bd. 21;
s. 379.

GoutpsmitH, W. M. 1919 A comparative study of the chromosomes of the tiger
beetles. Jour. Morph, vol. 32, pp. 487-488.

Kornuauser, 8S. I. 1914 A comparative study of the chromosomes in the
spermatogenesis of Enchenopa binotata, ete. Arch. f. Zellf., Bd. 12
s. 241-298.

LomeN, Franz 1914 Der Hoden von Culex pipiens L. Zeits. Naturwiss., B. 52.

SPERMATOGENESIS IN ASILUS SERICEUS 181

McCuune, C. E. 1902 The spermatocyte divisions of the Locustidae. Kans.
Univ. Sci. Bull., vol. 1, pp. 185-240.

Merz, C. W. 1916 Chromosome studies on the Diptera II]. Jour. Exp. Zodl.,
vol. 21, pp. 213-279.

Montreomery, T. H. 1900 The spermatogenesis of Peripatus.
Zool. Jahrb., Bd. 14, s. 277-368.

Nonipez, José 1914 Los cromosomas en la espermatogenesis del ‘Blaps lusi-
tanica’ Herbst. Trab. Mus. N. Cienc. Nat., Madrid, Ser. Zool., num. 18.
1915 Estudios sobre las celulas sexuales. I. Los cromosomas goni-
ales y las mitosis de maduracion en Blaps lusitanica y B. Waltli. Mem.
Soc. Esp. Hist. Nat., T. 10.

Overton, J. B. 1905 Ueber Reduktionsteilung in den Pollenmutterzellen ein-
iger Dikotylen. Jahrb. f. wiss. Bot., Bd. 42, s. 121-153.
1909 On the organization of the nuclei in the pollen mother cells of
certain plants. Ann. Bot., vol. 23, pp. 19-62.

Srevens, N. M. 1903 On the odgenesis and spermatogenesis of Sagitta bipunc-
tata. Zool. Jahrb., Bd., 18, s. 227-240.
1905, 1906 Studies in spermatogenesis. Parts I and II. Carnegie
Institution of Washington, Pub. 36.
1907 The chromosomes of Drosophila ampelophila. Proc. VII Inter-
nat. Zool. Cong. ;
1908 a The chromosomes in Diabrotica. . . . . Jour. Exp. Zodl.,
vol. 5.
1908 b <A study of the germ cells of certain Diptera, with reference
to the heterochromosomes and the phenomena of synapsis. Jour.
Exp. Zool., vol. 5, p. 45% 3379
1909 Further studies on the chromosomes of the Coleoptera. Jour.
Exp. Zodl., vol. 6.
1910 The chromosomes in the germ cells of Culex. Ibid., vol. 8.
‘1911 Further studies on the heterochromosomes in mosquitoes. Biol.
Bull., vol. 20, pp. 109-120.

Taytor, Montca 1914 The chromosome complex of Culex pipiens. Quart.
Jour. Mie. Sei., vol. 60.

Vom Rarn 1892 Zur Kenntnis der Spermatogenese von Gryllotalpa vulgaris
Latr. Arch. Mikr. Anat., Bd. 40; s. 102-132.

Wuirttnc, P. W. 1917 The chromosomes of the common house-mosquito, Culex
pipiens L. Jour. Morph., vol. 28, pp. 523-577.

Witson, E. B. 1912 Studies on chromosomes VIII. Jour. Exp. Zodl., vol. 13,
pp. 345-449.

EXPLANATION OF PLATES

All of the figures were drawn from material fixed in Flemming and stained in
iron haematoxylin; sections in most cases were 5u in thickness. All were drawn
with the aid of a camera lucida, using 1.5-mm. Zeiss apochromatic objective
and no. 12 ocular, with 160-mm. tube length. Drawings were made at table
level and are reproduced without reduction in size.

PLATE 1

EXPLANATION OF FIGURES

Figures 1 to 13, Asilus sericeus. Figures 1 to 5, spermatogonia. Figures 6
to 13, telophases of final spermatogonia and early growth stages of first sper-
matocyte.

1 Late resting stage or early prophase, the chromatin aggregated into five
bodies that represent the five pairs of chromosomes.

2 Two prophases, nuclei entire, each showing the five pairs of chromosomes,
resembling bivalents. Note the differences in degree of condensation of the
different elements.

3 and 4 Early and late metaphases.

5 Final spermatogonial anaphase, showing the paired association of the
chromosomes as they go to the poles; nucleus is cut so that one pair is missing
from the lower pole.

6 Late anaphase or early telophase of the final spermatogonial division.
The nuclei are practically entire and the chromosomes are all represented. The
figure at the right is slightly earlier than the one at the left, but the union of
the chromosomes in pairs is so intimate that only one shows the dual structure.
In the figure at the left the duality is scarcely visible, and not more than the
haploid number of chromosomes can be detected; the two lying side by side at
the lower pole are separate pairs, not members of one pair.

7 Approximately the same stage as 6. The union is progressively more inti-
mate from the lower to the higher of the three nuclei. The nuclei are practically
entire and each shows four or five bivalent chromosomes.

8 An entire nucleus of the same stage, showing the five bivalent chromosomes ;
scarcely any trace of duality is revealed.

9 Four cells in late telophase after the nuclear membrane has appeared and
the chromosomes have moved apart and lost much of their staining capacity.
Only part of the chromatin is represented.

10 Stage a, shghtly later than the preceding; the chromosome remains are
barely visible; nuclei not entire.

11 to 13 Stage b, the chromosomes again taking the stain and reappearing
in the form of five long bivalents, one of which forms part of the large nucleolus;
nuclei entire.

182

SPERMATOGENESIS IN ASILUS SERICEUS

CHARLES W. METZ AND JOSE F. NONIDEZ

Ne
“~
s
or,
& . .
s A }
o
wa

}
~“
™~
/ Ps)
ee.
j Fn
ne 6. j
, eS
4
hie
4a a
(oe
ae HF
/
ys ~
f >

PLATE 1

PLATE 2

EXPLANATION OF FIGURES

Figures 14-22, Asilus sericeus, middle and late growth stages.

14 Late stage b, four of the bivalents drawing out into diplotene threads,
the other attached to the plasmosome, forming the nucleolar complex. Three
of the threads are, respectively, at the top, the extreme left, and the extreme
right of the figure; the fourth is at a low focus, indicated by its light color, and
passes underneath the nucleolus. The nucleus is entire.

15 and 16 Stages progressively later than the preceding, showing the conden-
sation of the threads and their orientation with respect to the nucleolus; nuclei
entire.

17 Contraction stage, nucleus entire.

18 to 20 Successive stages following the contraction. The diplotene threads
extend out like fingers from the nucleolus; nuclei entire.

21 Prophase of the first spermatocyte division; nucleus entire. The four
long threads have broken loose from the nucleolus; the latter has become smaller
and is mostly chromatic.

22 First spermatocyte metaphase showing the five bivalents.

184

SPERMATOGENESIS IN ASILUS SERICEUS
CHARLES W. METZ AND JOSE F. NONIDEZ

PLATE 2

alee j
it ee \
ee. ult ;
ry j f
ih . st ¥ ;
A ’ ‘
em i‘ ”
i D iy J] me
‘Sp
: he tea oy
ee i
L; URNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2, iV
FEBRUARY, 1921 cpt v
ne oe .
whi
7
i A |
e -~ +

Resumen por el autor, Harold H. Plough,
Amherst College.

Nuevos estudios sobre el efecto de la temperatura en el crossing
over.'

Los datos presentados en este trabajo constituyen un suple-
mento a un trabajo precedente del autor sobre el efecto de la
temperatura sobre el crossing over del segundo cromosoma de
Drosophila melanogaster. Empleando el mismo método del
trabajo mencionado, el autor ha estudiado los efectos de dicho
agente sobre prdcticamente el total de la longitud conocida de
los cromosomas segundo y tercero, sometidos a una temperatura
de 31.5°C. Los resultados obtenidos indican que ni-la temper-
atura ni la edad de la hembra madre de una generacion, producen
una variaciOn significativa en el crossing over de parte alguna
del cromosoma sexual. Solamente la regién media del tercer
eromosoma presenta un aumento marcado en el crossing over
como resultado de la exposicién a una temperatura elevada y
una variacion con la edad. Las regiones de los cromosomas
que son ‘‘sensitivas”’ a los cambios del medio ambiente presentan
también una proporcién elevada de crossing over sencillo y
doble. Es probable que donde el crossing over es menos libre,
se pueden observar los efectos del medio ambiente.

1Con este nombre se designa el entrecruzamiento de los cromosomas en las
células sexuales de la hembra. Nos parece mas conveniente conservar la palabra
inglesa, puesto que ha adquirido un significado preciso que se pierde al tradu-
cirla. (N. del T.)

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, NOVEMBER 15

FURTHER STUDIES ON THE EFFECT OF
TEMPERATURE ON CROSSING OVER

HAROLD H. PLOUGH
Department of Biology, Amherst College

THREE FIGURES

CONTENTS
LiNGTG is (Gia pole ee Nh ee hay Se epee Rael Eee ew eae SIM 187
LOSSES Ten Str Nes AR AC ee ee PR CO a PL 73 AG CM pec eee 188
Interpretation of the curves of crossing over... .......6.-ceees-eeeeessan 189
Reaction to temperature and high coincidence......................--000- 194
WIS CUSSION Amines. > os oe AT iese MRAP Sica revs = OE One ae SPE Pooh s SANG. coat ntetye Pare 198
Age wnd: temperature effects compared . ..))./..\.'8oh. hak She apardea ce tioete. ¢ 199
STEUER Toy iter ea hr Sid eae ys Meee. | ASO Wei Colley chy gla 201
INTRODUCTORY

In an earlier study of the effect of temperature on crossing
over in Drosophila melanogaster I showed that temperature
both above and below the optimum (22°C.) caused a significant
increase in the amount of crossing over between certain genes
located in the second chromosome (Plough, ’17). Preliminary
work on the first and+third chromosomes indicated, however,
that crossing over in these groups was not visibly affected by
temperature. No reason for this unexpected result could be
assigned, and it seemed worth while to test the first and third
chromosomes by the same accurate methods which had been
used with chromosome IT. Such data would also give an accurate
basis for checking the fact reported by Bridges (’15) that in
chromosome J, unlike the second chromosome, there is no signifi-
cant variation in crossing over due to the age of the female
parent.

The large amount of breeding work with Drosophila has
resulted—especially through the work of Bridges—in making
available a large number of easily workable mutant characters

187

188 HAROLD H. PLOUGH

with excellent viability. The most valuable mutants of each
chromosome have been assembled into multiple stocks, the use
of which has made it possible to determine the effect of environ-
mental changes on linkage relations over approximately the
whole known lengths of each of these chromosomes in a single
experiment. At the same time in these multiples the distances
between the different genes are generally not sufficiently great
to cause complications due to unobserved double crossing over.
My present more accurate data establish the truth of the earlier
observation that crossing over in chromosome I is not influ-
enced by temperature, but show that there is a section of chromo-
some III in which crossing over is increased in the same way as
in chromosome II.

EXPERIMENTAL

The mutant stocks used for the tests were the sex (or first)
chromosome multiple stock, scute-echinus-cut-vermilion-garnet-
forked, and the third chromosome multiple stock, sepia-spineless-
sooty-rough. Scute shows an absence of scutellar bristles, and
echinus, a roughened condition of the facets of the eyes. The
other mutations have been described. Dichete, a dominant
character, was introduced in a small number of preliminary
third chromosome tests based on ten-day brood counts. A
glance at the chromosome maps in figure 3 will show that the
genes used cover both chromosomes at fairly even intervals
throughout a large portion of their known lengths.

The method of making the tests was essentially the same as
that used in my earlier work with the second chromosome.
Virgin sister females of the normal wild stock were mated to males
of the multiple mutant stock to be tested, and allowed to lay
in one set of bottles for about three days. This first set of bot-
tles was kept at the control temperature. The P, pairs were
next transferred to another set of bottles which was kept con-
tinuously at the high temperature. Virgin female offspring .
from each set were then isolated and back crossed to males of
the original mutant stock used. These back crossed pairs were
placed in quarter-pint milk bottles containing banana agar, and

EFFECT OF TEMPERATURE ON CROSSING OVER 189 .

kept at the control temperature. With the exception of the
preliminary ten-day brood test, the pairs were changed from
one set of such bottles to another at the end of successive three-
day periods throughout the life of the females. (The first
change was made in all cases at the end of the fourth day.) The
counts of the successive sets of offspring of these pairs furnished
the data for determining the effect of the high temperature on
crossing over in the developing eggs of the heterozygous females.

The control temperature was approximately 24°C. maintained
in a wooden stock cabinet controlled by an electric heater with
a thermostat. It varied between 22°C. and 25°C. throughout
the experiments. The high temperature was 31.5°C. maintained
in a Freas electric incubator. This varied as much as 1° above
and below, though the normal variation was about 0.5° either
way. |

The results given by the tests of the first chromosome regions
are given in table 1. The bottle counts for each three-day period
are added and the percentages of crossing over for each of the
five regions calculated and listed in the columns at the right.
The successive percentages of crossing over for each of the five
regions are plotted as curves in figure 1, the dotted line in each
case being the experimental value and the full line the control.
The results of the tests of the multiple third chromosome stock
are given in tables 2 and 3, with the percentages of crossing over
in the columns headed per cent 1, etc. Table 2 summarizes the
results of a ten-day brood count made with the sepia-Dichete-
spineless-sooty-rough stock, and table 3 shows the three-day-
interval results from the same stock without Dichete. The
crossover values of table 3 are plotted as curves in figure 2.

INTERPRETATION OF THE CURVES OF CROSSING OVER

An examination of the curves in figure 1 demonstrates the
following facts. First, the full and dotted lines for each of the
five regions show no significant differences. The work on the
second chromosome showed that for a region that was sensitive
to temperature, the dotted line was significantly higher than the

8°S19 TIP E19 STP SO;CTLT

ZL F180
O'SI8°%

Beem |

€°S16 E1/0 60/ZaT
S'TLIP 91S 90/887

6° E1/G ST/9 €1\G 02/9 L0/Z0E
9°OT/0 STE SIZ FIL ITSP
SCL Il6 P16 STZ LO)2h2

LOLF 606 TTF FILE SO/6TE
VILL St0ST/9° 21/8 90/90€

TOT OZ ELF 91/9 60/S8E

O'Slh ELE FIG FIIs LO|69P

8° S119 TT/9°ST9 STII 20)90¢

€ cle OL ELF 9/0 OL LELZ

aad
udd
udd
dad

ddd
TVLOL

¢ INGO
f INGO
¢ INGO
Z INGO
[. INGO

A]Snonul}Uo0d *),FZ—S}UN0d poosg Avp-Us} 9WOSOMMOAYD Say JO SotreuTUING

I
(¢-F-8)

OT O/LP O/ZS 'O/S2"0/TZ0/9Z 0/98 0|8E" T/82°Oj}0F'O

(G--8) | ($-4-Z)

(G-¥-8) | (S-€-T)

(S=Z=P) (G=c- 1)

Sa Td UL

relbol| Ahea Ev

TZ | OL

&@ | ST

"+" g9uaplouloy

SOT/OPTISST| 10/18} 962

CO <0) alfa esata at cm LO Ye item TRC IA OV I S(0)e= eye cP oP) teal i Roa hica threo) 18-46
I GON GAO Ae ieee I | UL G | 0 |6€ |8F /6¢ |6S |ST} 68 | F2-ZT
(¢-§-Z)
Si & -@ cies | GO ace ROM ILGE SI Pee Gira GulescienieAl—el
|! 0 I (6 T | 2. |.00 | 8 | l2esi98 \tF |\SP 21) 906 | €L-01
I Geli et TE 96.3) -F= TTS | aE ee 2G 129 SE TG \Sai SSeS) | OLeL
l! I G I Seal On ay Gu | Te RO WZ GGn | SGe Ose ONS Z es eZ 7,
I OG Ws mG A eS {l OP OMS SCy| 8G 26s Si RGSIes |S <1
S
‘DoF 1B pope “HG TE 78 poyoyey—poyeo1y yoy] .
qT 0 I Were ND PO ots I tO) Kaizer Sie G2 ray Age |) ZOE
($-4-T)
T qT Hye Wakes | EE WO ee Wie lk Ne IT (|&P |€ \6P \ES |ZT) Tea | OT-2
(¢-4-€)
I 0) ts LEE ts) I § |24 |% | 0 |9S \PP ISS |v9 \ec| cS | 2H
(S-4-Z)
Do fs yf 7 COLE H tee Ae Ce Mitel 74 toy AS OS NBS Wie ay) sas | ae IE
A[snonurjuod “9, ¢Z—]o1}U0y
UGAO | ONILYW
SP | SE || #8 | Se |.7-e | S-s | ST | FT | er | ST | S| 7] 8 | CS IT | -ssouo | uanay
NON BAVaG
J 3 A 49 09 oS

$91LaS BWOSOWOLYD BS.LYf fO saludDUUNY

T WIaViL

EFFECT OF TEMPERATURE ON CROSSING OVER 191

full line for the first eight or ten days. It later coincided with
the control after all the eggs were laid which had passed the
critical period—when crossing over probably occurs—at the
high temperature. No section of the first chromosome tested,
therefore, shows any effect of high temperature on the amount
of crossing over. This is a complete confirmation of my earlier
conclusion, made on less exact data, and for a much shorter

scute-echinus

echinus-cut

Fig. 1 Curves of crossing over for different regions of chromosome I. The
control values are the solid lines; the values from the heat-treated lines are
dotted. The abscissas are days after mating, the ordinates percentages of
crossing over.

section of the chromosome. Second, a comparison of the differ-
ent control lines with each other indicates no significant variation
as the female grows older. The control line for second chromo-
some regions dropped steadily up to about the tenth day, and
gradually rose up to about the twenty-second. The sex chromo-
some control values show no significant nor uniform changes,
and confirm the conclusion of Bridges (’15) that in this chromo-
some the age of the female has no influence on the amount of
crossing over.

192 HAROLD H. PLOUGH

TABLE 2
Preliminary test of chromosome III, ten-day broods
se ss es ro
D

CENT 4

266 | 30| 48| 49| 7o| 10 | 6 | 12] 7} 17] 6 | 0/2) 0]2 525/11.4|16.4|13.3)20.8

2-3-4
TOTAL
PER

1-4

Control—22°C. continuously

Coincidence... ....|1.22 0.78 0.55 |

Hatched at 31.5°C., mated at 22°C.

103 | 15| 63| 19| 31] 9 | 8 | 9] 13) 15] 6 | 3 | 8] 1 | 2 (305/17.3/33.814.9123.6

Coincidence....../1.02 0.99 0.78

TABLE 3
Summaries of third chromosome series
se ss e& ro :

DAYS
AFTER
MAT-
ING

NON
CROSS-
OVER

1-3

Control—24°C. continuously

1-4 337 | 129| 47 | 96] 26 | 40
Aa 493 | 180| 97 | 154] 32 | 49
7-10| 507 | 183] 82 | 136] 25 | 45
10-13 | 455 | 157| 72 | 126] 30 | 27
13-16 | 506 | 198] 95 | 123] 24 | 53
16-19 | 380 | 146] 71 | 93] 11 | 28
19-26 | 185 74| 25 | 66] 11 | 18

0 | 677|28.7\11.0/18.9|1.20/1.09/0. 14
2 |1008)26.1)/13.1)/20.4/0.93)0.90/0.04
0 | 979/25.7|10.9)18.5/0.89|0.95|0.05
2 | 872/24.7/12.1|18.0/1.14)0.69/0.14
0 |1004/27.4|12.6)18.3/0.69/1.08)0.47
3 | 732/25.6)/11.6|16.9/0.55|0.87|0.30
0 | 379/27.1/09.5|22.1)1.13/0.79

oo OW Fe tb

Heat treated—hatched at 31.5°C., mated at 24°C.

1-4 195 97| 34 | 54) 20] 31 | 3
AG 370 | 228} 67 | 140| 37 | 86 | 11
7-10 2570 |alsO)rady) 76|29)) 37) 6
10-13 257 96} 45 | 78) 15 | 23 | 4
13-16 212 92) 36 | 55) 16 | 22] 2

435/34 .2/13 .3/20.4|1.01/1.02/0.25
948/39 .0/13.1/26.6|0.76|0.94/0.31
587/33 .5|14.8/20.4|0.99)0.92/0.33
518/25 .8/12 .4|20.2/0.90/0.84/0.32
436|30.0)12.6/18 .3]0.98/0.91)0.21
446/28 .8/11.8/19.9}0.65/0.78}0.09
887 |26 .2|11.1|16.9/0.52/0.79/0.27

BREE OrFOrF

16-19 208 98) 41 | 67} 10 | 20
19-26 200 81| 34 | 49) 6 | 14

EFFECT OF TEMPERATURE ON CROSSING OVER 193

An examination of tables 2 and 3 and of figure 2 discloses an
interesting situation in the third chromosome. It seems clear
from table 2 that high temperature causes a definite increase in
crossing over between sepia and Dichete and a very marked one
between Dichete and spineless, but little if any change in the
remainder of the chromosome. ‘Table 3 and a comparison of the
full and dotted lines in figure 2 bring out this fact even more
definitely, but without separating sepia and spineless by the

sepia-spineless

spineless-sooty

sooty-rough

Fig. 2 Curves of crossing over for different regions of chromosome III. The
control values are the solid lines; the values for the heat-treated lines are dotted.
The abscissas are days after mating, the ordinates percentages of crossing over.

Dichete factor. The dotted line for the sepia spineless region
begins at a point about 6 units higher than the control, rises
to a difference of 18 units, and drops sharply to about the
same point at about the tenth day. This indicates, as in the
second chromosome, that the eggs which go through the critical
period at the high temperature show a much increased crossover
ratio, but that those which pass through that period subsequently
(i.e., after the females are replaced at the control temperature)
are not affected. The dotted line for the spineless sooty region,
on the other hand, shows no significant difference. That for

194 HAROLD H. PLOUGH

sooty rough shows a rise of nearly six units in the four-to-seven-
day period, but since no difference appears either before or after
this time, it is probable that this has no significance. The
data indicate, therefore, that the percentage of crossing over is
increased by exposure to high temperature at one end of chromo-
some III, but not throughout the remainder of its length.

It is interesting to note that the control line for the sepia
spineless region shows the age variation observed in the second
chromosome. The value gradually drops to the tenth day and
then rises. The rise apparently takes place somewhat earlier
than in chromosome II. The other two regions show no signifi-
cant difference as the female grows older.

The results of the tests may be summarized as follows: a) the
sex chromosome shows no significant increase in the percentage
of crossing over as a result of the exposure of the developing
eggs to high temperature; b) the third chromosome shows an
increase in crossing over in the sepia spineless region, but nowhere
else; c) a variation in crossing over with the age of the female
occurs in those regions which show a reaction to temperature only.

REACTION TO TEMPERATURE AND HIGH COINCIDENCE

In figure 3 I have drawn to the same scale comparative maps
of the principal chromosome regions whose reactions to high
temperature have been tested. The percentages of crossing
over in chromosome I have been calculated from the ten-day
brood counts summarized at the end of table 1, and those for
chromosome III from table 2. The map of chromosome II is
taken (for the points indicated) directly from the very accurate
one given by Bridges and Morgan (p. 302). The regions for
which a rise in the percentage of crossing over as a result of
exposure to a temperature of 31.5°C. has been recorded—either
in this or my former paper—are indicated by diagonal lines,
while those which are not changed are solid black. As noted
previously, it may develop that one or both of the long regions
at either end of chromosome II will show a reaction to temper-
ature if they can be broken into short blocks. A rise in total
crossing over may be obscured by a compensating rise in double

EFFECT OF TEMPERATURE ON CROSSING OVER 195

crossing over, so that no result appears in the count. With the
exception of these two regions, in which Bridges (715), Plough
(17), and Bridges and Morgan (’19) have recorded an age differ-
ence, the diagonal lines also indicate the regions which show a
variation in crossing over with age.

0.0 Star
0.Ogrscute 00° sepia
8. echinus 11.4 Dichete
40
102
122
26
24 cut 8
783 spineless
46.5
22 black
-4 4 vermilion 2.7 4 purple
6 55
41.1 sooty
96
ap uck: pa waic 65.04 vestigial
3.5 curved
1. forked , 1.9 rough
48

Chromosome I Chromosome III

185.5 speck

Chromosome II

Fig. 3 Chromosome maps showing all the important regions whose reaction
to high temperature has been tested. The regions which show a significant in-
crease in crossing over after exposure to high temperature are ruled with diag-
onal lines those not affected are solid black. The coincidence values are given
outside of brackets enclosing the different pairs of adjacent regions.

I have recently laid some emphasis on the fact that in the
sensitive section of chromosome II the percentage of increase
in crossing over due to high temperature was roughly in inverse
proportion to the length of the region involved (Plough, 719).
The results in chromosome III demonstrate very plainly, however,

196 HAROLD H. PLOUGH

that the masking effect of increased double crossing over is not
the only reason why certain regions remain unchanged after
exposure to high temperature. Table 2 shows that in chromo-
some III the Dichete spineless region (16.4 units) shows an
increase of more than 100 per cent, while sepia Dichete (11.4
units) is increased only about 50 per cent, and spineless sooty
(13.3 units) is practically unchanged. It is apparent that there
are other factors than the mere length of the region responsible
_ for the difference in reaction to high temperature. Not only are
different chromosomes unlike, but within each chromosome cer-
tain regions show distinct differences in behavior from certain
other regions. This fact has been apparent to several Drosophila
workers in connection with the investigation of the coincidence
values, and has been the subject of a special study by Bridges,
the results of which are not yet published. It is of some interest
to compare the differences in reaction to temperature with the
coincidence results.

Figure 3 shows also the coincidence of double crossing over
for each pair of adjacent regions as a percentage value outside
a bracket enclosing the two regions for which it has been calcu-
lated. The significance of the value for coincidence has been
discussed in detail by Bridges (15), Muller (’16), Weinstein
(18), and Bridges and Morgan (19). It represents the per-
centage of the expected number of double crossovers actually
observed. The size of the coincidence value has been shown
by a number of observers to be in proportion, up to a certain
point at least, to the distance apart of the opposite boundaries
of the regions tested (i.e., to the lack of interference). For
instance, a glance at the ten-day counts in table 1 shows that
the coincidence value is high when the scute-echinus and vermil-
ion-garnet regions (138 per cent) or the scute-echinus and garnet-
forked regions (86 per cent) are figured, but low when the regions
are closer together.

A comparison of the coincidence values given for approximately
equal lengths of chromosome shows that they do not correspond.
Double crossing over per unit of distance is apparently much

EFFECT OF TEMPERATURE ON CROSSING OVER 197

more frequent in certain regions than in others, and this is true
not only when the regions compared are in different chromo-
somes, but when they are in different portions of the same
chromosome. For instance, for the black-purple-vestigial region
of chromosome II we get a coincidence of 66 per cent,! while
for distances slightly longer in chromosome I-scute to cut or
cut to garnet—the values are 40 per cent and 22 per cent, respec-
tively. The black-purple-curved region in chromosome II
gives a coincidence value of 96 per cent (Plough, 717, table 8),
while a similar length in chromosome I—-echinus-cut-vermilion—
gives 26 per cent. Even more striking, however, is the fact that
the region at the upper end of the third chromosome on the
map—sepia-Dichete-spineless, a distance of 28 units—gives a
coincidence value of 122 per cent, yet the longer lower region—
spineless-sooty-rough (33 units) gives only 55 per cent coinci-
dence. This indicates that double crossing over is at least as
frequent as though there were no interference at the sepia end
of chromosome II, but interference is almost as high at the rough
end as in chromosome I. A difference of the same order is
apparent in chromosomes I and II.

It will now be obvious that those sections of the chromosomes
mapped in figure 3 which show a relatively high coincidence per
unit of distance are the same ones which show a change in the
amount of crossing over as a result of high temperature and of
the age of the female. In no case where the coincidence value
for a continuous region of 30 units or less is below about 60 per
cent do we find an increase in crossing over due to high temper-
ature, or, with the possible exceptions noted in the second chromo-
some, a change due to age. The chromosomal regions which are
‘sensitive’ to environmental effects all show a minimum of influ-
ence of one crossover on another simultaneous crossover in the
same region.

1 Bridges and Morgan, ’19, table 42—not 61 per cent, as they give it.

198 HAROLD H. PLOUGH

DISCUSSION

The fact that high coincidence and sensitiveness to environ-
mental effects are found in the same chromosomal regions sug-
gests that certain structural features of the crossing over process
determine each. Bridges and Morgan (p. 188) suggest that
the difference in the amount of interference for short regions
between the first and the second chromosomes may be inter-
preted in two ways. We may assume that the average length
of loop between simultaneous crossovers is the same in each,
which means that a region having a given coincidence value in
chromosome II is actually the same length as one having the
same coincidence in chromosome I. On the other hand, we may
hold that the length of loop between simultaneous crossovers
is relatively shorter’in chromosome II than in chromosome I, -
which means that equal amounts of crossing over then indicate
equal lengths of chromosome. Either of these alternatives holds
also for the different sections of chromosome III. According
to the former interpretation, crossing over takes place relatively
less freely in the regions ruled with the diagonal lines and they
are actually much longer than the map length indicates. Ac-
cording to the latter view, crossing over takes place relatively
more freely, and the map lengths are accurate. Bridges (’19,
and from subsequent unpublished data) distinctly favors the
former interpretation. The effect of high temperature in causing
an increase in these regions does not give any clear evidence
for either view, though it would seem to support Bridges’ inter-
pretation. It is hardly possible that temperature does not act
on the whole chromosome equally. Any observed differences
between different regions would seem to be due to the fact that
slight effects are registered in certain regions and not in others.
It is reasonable to suppose that the regions in which a change
is observable should be those in which crossing over is less free.

It is of some interest to consider what structural conditions
in the chromosomes could result in regions of decreased freedom
of crossing over. Bridges and Morgan (p. 198) and Bridges
(719) suggest that the reason for the difference in behavior of the
black curved region in chromosome II may lie in the fact that this

EFFECT OF TEMPERATURE ON CROSSING OVER 199

region is near the middle of the chromosome, “with the spindle
fiber attachment, and that this middle region is the last part to
undergo synapsis.”” Bridges has subsequently applied this same
idea to chromosome III and decided its middle point is close to
the locus for Dichete. In the latter case the conclusion as to
the midpoint of the chromosome has been definitely confirmed
with the finding by Strong (’20) of the locus for roughoid at
24.9 units beyond sepia. If, as Bridges suggests, crossing over
is less in this middle region because synapsis fails or is slight,
the decreased freedom of crossing over might be consistently
explained. On the other hand, it should be definitely borne in
mind that such behavior is an observable phenomenon, which
is susceptible of cytological demonstration. The demonstration
that the process of crossing over is accomplished by a simple
twisting separation, and reunion of chromosome strands is still
incomplete, and we have no cytological data which indicate
that in Drosophila the middle region is the last part to undergo
synapsis. At the early stage in the growth period of the egg
at which crossing over apparently takes place it seems altogether
unlikely that the spindle fiber suggested by Bridges is present
at all. Until we know more of the actual cytological features
of the crossing-over process and of the spindle fiber attachments
in Drosophila, such suggestions must be regarded as highly
speculative.

AGE AND TEMPERATURE EFFECTS COMPARED

It has been demonstrated above that in general both age and
temperature affect the amount- of crossing over in the same
chromosome regions—those probably in which there is a mini-
mum of crossing over. It is to be expected, therefore, that the
freedom of crossing over is modified by both agents. It is of
some interest to note that Bridges and Morgan (p. 199) and also
Bridges (19) in identical language conclude that the age variation
is probably due ‘‘to a lengthening of the average length of the
section of chromosome between simultaneous crossovers,”’ while
temperature causes an increase in the freedom of crossing over
with no difference in the length of loop. The clearest evidence

200 HAROLD H. PLOUGH

for this conclusion is stated to be found in a calculation of the
coincidence values for my two-day-interval experiment for the
black-purple-curved region reported in my former paper. They
state (p. 199):

In this experiment the curve of variation in coincidence was the
mirror image of the curve of variation (in crossing over) for age. The
curve of coincidence corresponding to the curve of temperature varia-
tion found by Plough seems to be a straight line cutting the rises and
falls of the temperature curve and independent of them.

TABLE 4
Coincidence values for control and heat-treated lines in two-day-interval experiment.
(For actual counts cf. Plough, ’17, table 14)

b pr c
PARENTS HATCHED AT 22°C., EXPOSED
CONTROL—22°C. CONTINUOUSLY To 31.5°C. FROM 3RD TO 11TH Day
NUMBER DAYS AFTER MATING
AFTER MATING
Per cent of crossing Coincidence Coincidence Per cent of crossing
over—b, pr region b, pr-pr, e b, pr-pr, c over—hb, pr region
3 8.3 0.91 Ot Teil
15) 4.9 0.94 1.31 4.8
7 6.8 Ids 0.53 3.8
9 5.8 1.03 1.37 3.8
11 4.2 1.06 0.63 8.8
13 5.1 0.80 0.94 13.9
15 5.3 1.40 0.99 19.2
17 4.2 0.92 0.95 20.0
19 7.3 0.93 0.73 17.5
21 8.2 1.04 IL 5i7/ 6.8
23 7.9 0.63 0.27 4.9
25 7.0 0.98 1.4

I have calculated the coincidence values for the experiment
cited and the results are given in table 4. A comparison of my
coincidence values with the crossover percentages for the black-
purple region shows that the coincidence value varies within
rather wide limits in both the control and experimental lines.
A smoothed curve gives some suggestion of the relation claimed
by the writers quoted, but its significance is doubtful. The same
comparison may be made between the similar lines in chromo-
some III (table 3). The coincidence values for the different
pairs of regions are given in the last three columns. Here there

EFFECT OF TEMPERATURE ON CROSSING OVER 201

is surely no mirror-image relation in the control series. In
addition, the coincidence is subject to so high a probable error
that it would take very marked and constant differences to
establish such a conclusion as the one stated. It seems clear
that much weightier evidence than that quoted must be given
before it can be established that age and temperature act in
different ways on the crossing-over process. It is more con-
sistent with the results here given that each causes a variation
in the actual freedom of crossing over and that the changes ‘in
coincidence recorded are without significance.

\
SUMMARY

1. It has been shown that a temperature of 31.5°C. causes
little or no effect on crossing over in any part of the sex chromo-
some, nor is there any significant variation with the age of the
female.

2. Crossing over in the sepia-spineless region of chromosome
III is increased by a temperature of 31.5°C., the effect being
most marked between Dichete and spineless.

3. The same region shows a variation in crossing over with
the age of the female parent. —

4. Crossing over in the remainder of chromosome III is influ-
enced neither by temperature nor age.

5. The chromosomal regions which are ‘sensitive’ to temper-
ature and to age all give a very high ratio of double to single
crossing over.

6. This is interpreted as indicating that the effects of environy
ment cause observable differences in crossing over only where
crossing over occurs least freely.

7. It is shown that the view that temperature and age act
op crossing over in different ways is not established.

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

202 HAROLD H. PLOUGH

LITERATURE CITED

Bringces, C. B. 1915 A linkage variation in Drosophila. Jour. Exp. Zodl.,
vol. 19, no. 1, July.
1919 The genetics of purple eye color in Drosophila. Jour. Exp. Zoél.,
vol. 28, no. 2, May.

Bripaes, C. B., anp Morcan, T. H. 1919 The second chromosome group of
mutant factors. Publ. no. 278, Carnegie Inst. Wash.

Mouuuer, H. J. 1916 The mechanism of crossing over. Amer. Nat., vol. 50.

Pioucu, H. H. 1917 The effect of temperature on crossing over. Jour. Exp.
Zool., vol. 24, no. 2, November.
1919 Linear arrangement of genes and double crossing over. Proc.
Nat. Acad. Sci. U. 8., vol. 5, May. ;

Srronec, L. C. 1920 Roughoid, a mutant located to the left of sepia. Biol.
Bull., vol. 38, no. 1, January.

WEINSTEIN, A. 1918 Coincidence of crossing over in Drosophila. Genetics, 3,
March. .

Resumen por los autores, William EK. Burge
y Emma Longfellow Burge,
Universidad de Illinois..

Una explicacién de la variacién en la intensidad de la oxidacién
durante el ciclo vital.

Es un hecho conocido que la oxidacién o metabolismo es muy
baja en el 6vulo no fecundado, mientras que aumenta notable-
mente a raiz del proceso de la fecundacion; que el metabolismo
del nifio recién nacido es también muy bajo, pero que aumenta
ripidamente llegando a ser muy elevado durante la ninez, dis-
minuyendo después gradualmente desde la edad adulta hasta la
vejez. Los autores han observado que 0.5 gramos de los huevos
no fecundados de Leptinotarsa, macerados, desprenden 18 cc.
de oxigeno en diez minutes cuando se tratan con perdxido de
hidrégeno, y que 0.5 gramos de huevos fecundados desprenden
35 ec. durante el mismo tiempo. 0.5 gramos de larvas recién
salidas del huevo, durante la cuarta parte, mitad, tres cuartas
partes del desarrollo y larvas completamente desarrolladas
desprenden 280, 800, 1250, 1725 y 1750 cc., respectivamente,
y que las ninfas, adultos e individuos viejos desprenden 1800,
1750 y 900 ec. de oxigeno, respectivamente. Comparando estas
figuras puede comprobarse que el huevo no fecundado contiene
mucha menos catalasa que el fecundado; que el contenido de
catalasa en las larvas recién salidas del huevo es menor que el
de las larvas mds avanzadas y que el contenido de catalasa en
el individuo viejo es menor que el del adulto mas j6ven.

La reducida cantidad de oxidacién en el huevo no fecundado
se debe probablemente a su escaso contenido de catalasa. La
oxidacién aumentada del huevo fecundado y su desarrollo consi-
guiente se atribuyen a un aumento de catalasa introducida por
la estimulacién del huevo para la produccién mayor de esta
enzima por parte.del espermatozoide. Del mismo modo el
aumento del metabolismo respiratorio u oxidacién en el jéven
y su disminucién con la edad avanzada, se atribuyen a un au-

mento de catalasa en el j6ven y a su disminuci6n en el de mas
edad.

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHORS’ ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 6

AN EXPLANATION FOR THE VARIATIONS IN THE
INTENSITY OF OXIDATION IN THE LIFE-CYCLE

W. E. BURGE anv E. L. BURGE
Physiology Laboratory, University of Illinois

ONE FIGURE

As a result of the work of a great number of observers, particu-
larly of Hasselbalch (1), Magnus-Levy and Falk (2), and of
Warburg (3), it is now known that oxidation or metabolism is
very low in the unfertilized ovum, while it increases greatly
following the process of fertilization; that the metabolism of
the newly born infant also is very low, but increases rapidly,
becoming very high during childhood and then gradually decreas-
ing from maturity to old age. The present investigation is an
attempt to find an explanation for the variation in the intensity
of oxidation under the conditions named.

Since we (4) had found that whatever increased oxidation in
the body, the ingestion of food, for example, produced an increase
in catalase, an enzyme possessing the property of liberating
oxygen from hydrogen peroxide, by stimulating the alimentary
glands, particularly the liver, to an increased output of this
enzyme, and that whatever decreased oxidation, narcotics, for
example, diminished catalase by decreasing its output from the
liver and by direct destruction, we naturally turned to catalase
in seeking an explanation for the variations in the intensity of
oxidation at different periods in the life-cycle.

The Colorado potato beetle (Leptinotarsa decemlineata) was
used in this investigation. Catalase determinations were made
of the following materials gound up in a mortar: unfertilized and
fertilized eggs, quarter, Analf, three-quarter, and full-grown
larvae, as well as pupae, adult, and very old beetles. Five-
tenths gram of the ground material were added to neutral hydro-

203

204 W. E. BURGE AND E. L. BURGE

gen peroxide in a bottle and the amount of oxygen liberated in
ten minutes was taken as a measure of the catalase content of
the material. .

AMOUNTS OF CATAL ASE
MERSUREL (NV CC. OF
OXYGEN.
UNFERTILIZEDEGG ———- =~ —— /6
FERTILIZEDLGG ——- ~ —— JS
Newty Harcueo LARVA —— = —— 280
QUARTER Grown LARVA —— ete — 800

tt FS he ake

Fur. GRownlakVa— ame

Pupa —— &
Aputt BEETLE —

OL0 BEETLE —

Fig. 1 The figures in the chart indicate amounts of oxygen liberated from
hydrogen peroxide in ten minutes by 0.5 gram of the material ground in a mortar.

The results of the determinations as well as photographs’ of
the beetles, pupae, larvae, and eggs are shown in figure 1. It
may be seen that 0.5 gram of the unfertilized eggs liberated 18
cc. of oxygen in ten minutes from hydrogen peroxide and 0.5 -
gram of the fertilized eggs, 35 cc.; that 0.5 gram of the newly’
hatched, quarter, half, three-quarter, and full-grown larvae

INTENSITY OF OXIDATION IN LIFE-CYCLE 205

liberated 280, 800, 1250, 1725, and 1750 ec., and that the pupae,
adult, and old bugs liberated 1800, 1750, and 900 cc. of oxygen,
respectively.

By comparing these figures it may be seen that the unfertilized
egg contains much less catalase than the fertilized. This is in
keeping with the fact that the oxidative processes are much less
intense in the unfertilized egg than in the fertilized one, as
observed by Warburg. It may be seen further that the catalase
content of a newly hatched larva is less than that of the older
larvae in keeping with the fact that in the newly born, and
presumably in the newly hatched, oxidation is very low and that
it increases very rapidly shortly after birth. It may also be
seen that the catalase content of the old bug is less than that of
the younger adult in accordance with the fact. that oxidation or
metabolism is less in a person of advanced age than in one in
middle life.

Tt should be mentioned in this connection that our observation
of the low catalase content of the unfertilized potato-beetle egg
and the high catalase content of the fertilized egg is in keeping
with the observation of Winternitz (5), who found that the
unfertilized hen’s egg showed no catalytic activity even after
prolonged incubation, whereas the incubated fertilized egg
rapidly acquired the power of decomposing hydrogen peroxide.
They agree also with the observations of Battelli and Stearn
(6), who found that the catalase content of most of the tissues,
and particularly of the liver, of newly born pigs is lower than the
corresponding tissues of the mother, but that the catalase activity
rapidly increased, until at the end of the seventh or eighth day
it was as high as that of the adult.

J. Loeb (7) attributes the development of the fertilized sea-
urchin egg to the increase in oxidation, and the increase in
oxidation to a change in the cortex of the egg which makes the
entrance of oxygen, and hence oxidation, possible, while R.
Lillie (S) holds that the cortical layer of the unfertilized egg
prevents the diffusion of CO, from the egg and that this CO,
prevents oxidation, and hence development. A more plausible
explanation for the increased oxidation or metabolism in the

206 W. E. BURGE AND E. L. BURGE

fertilized egg, and hence for the development of the egg, would
seem to be that the spermatozoon furnishes a substance which
stimulates the egg to an increased formation of catalase. Fur-
ther evidence that might be presented in support of this view is
afforded by the fact that the very same chemicals : (amines,
alkalies, acetates, butyric acid, ete.) which Loeb found would
bring about increased oxidation and artificial parthenogenetic
development of the egg, we found, when introduced into the
alimentary tract of animals, stimulated the alimentary glands,
particularly the liver, to an increased output of catalase with
resulting increase in oxidation.

SUMMARY

The low rate of oxidation in the unfertilized ovum is attributed
to its low catalase content. The increased oxidation in the
fertilized ovum, with resulting development, is attributed to
an increase in catalase brought about by the stimulation of
the egg to an augumented production of this enzyme by the
spermatozoon.

Similarly, the increase in the respiratory metabolism or oxi-
dation in youth and decrease in old age is attributed to the
increase in catalase in the young and its decrease in the aged.

BIBLIOGRAPHY

HassetBpatcH 1904 Bibliotek for laeger, Copenhagen, vol. 8, p. 219.

Maanus-Levy unp Fark 1899 Arch. f. Anat. u. Physiol., Suppl. 315.

WarsurGa 1908 Zeitschr. f. physiol. Chem., Bd. 57, No. 6.

Buree 1918 Am. Jour. Physiol., vol. 45, no. 4; vol. 47, nos. 1 and 3.

WINTERNITZ AND Rocers 1910 Jour. Exper. Med., vol. 12, no. 12.

BATTELLI E STEARN 1995 Arch. di Fisiol., T. 2, p. 471.

Lors 1913 Artificial parthenogenesis and fertilization. University of Chi-
cago Press.

8 Lititme 1910 Am. Jour. Physiol., vol. 27, p. 289.

IOooa fh WN

‘ hy os fe, bt M
i

mbaie oe "7

rs Mt 2 las Hk tm

Sue a .
a if

a) nn yi meer Tver im: bi
anys ah wigilose My): Hi hiantr

i a i ef oe Lasers ts,

vith aie bn

uf bt

ean
“foe ‘ ye

| miei nats Pye}
crit itn, DN aL hie Mt Pr is

aoe nod
MEO. 1M

y ' ; " 7 i

sett ae | Pues yay e th stp, iia. Ce a

| a pee hs) d a eHKnf i Ny PP MD
; Bi ; ie beat Ai , a at ne c rey wee el | rT] Datura ar en ‘s nt vo
‘ Se : ? ; ha ‘ mM, ie

| shy ase loltwamne: | |

* f hi c Nic
‘ tah ences ‘ oc heu tia
as (a4 F ity ‘2 i iy } bi Lavy x nt pic
woe i cag ay aay tru t
: f r i
1 , ay ry
eh nae’ ae 4 7h
rid Th ete ean one
‘6 py “ u A Ps Ls
pays a ie j ;
.

' cli es 7 40 ig i
rs ae ae basen vo
f " ated) od mh ay th. 4
yon 3 ieee. © me tee } >
ers Y ES fl mids i Wy mee, Cnet He yd 4 pits ah ‘a
/ i

a i Ts sei

j yg me i) NG ree Fi 7 ;
Ta ain 0 2
7 ni A d ; " vey 64 a ' iu? We ; tay

alien me i t 7 m

A r as al ith uit Pi

ae “a
Cet (Oe A wi ' rave
to | wk wie ins Phech Matty, Phat .

4
oe.
LA wod HAP yh eee 1
7 4
t : | ' é
3
geet
’ i
i i
i 4
} ’
\ j (oho
a” im ¥
} i 1 "
é con os
; A ty Wh
~* y ~
i a
i, 3
‘
i -
1
i . l
Ae - ods '

Resumen por los autores, Henry Laurens y 8. R. Detwiler,
Universidad Yale, New Haven.

Estudios sobre la retina. |

La estructura de la retina de Alligator mississipiensis y sus
cambios fotomecdnicos.

El ojo de Alligator posee un tapetum retinal bien desarrollado,
formado por la inclusién de guanina en las células epiteliales de
las poreciones dorsal y posterior, a una distancia de 1.5 mm.
de la entrada del nervio 6ptico. El pecten consiste en una
especie de copa pigmentada ligeramente elevada, que cubre la
entrada del nervio 6ptico. En toda la retina se encuentran
conos y bastones, pero la proporcidn de ambos es diferente en
distintas regiones. Los bastones son todos del mismo tipo, los
conos de dos tipos: grueso y delgado. Los primeros son mas
numerosos, presentdndose especialmente en las regiones posterior
y ventral de la retina. Los conos del segundo tipo se encuentran
solamente en la porcién ventral y no son numerosos. ‘También
existen conos dobles. Ninguno de los conos y bastones contiene
gotas de grasa.

Los nticleos de los bastones son de forma oval y la mayor
parte de ellos se proyectan a través de la membrana lmitante
externa en una extensién variable. Los niicleos de los conos son
piriformes y ocupan un nivel mds profundo que el de los nticleos
de los bastones, formando una segunda fila. Los bastones
presentan un cambio de longitud media de unas 4 micras, y son
mas largos a la luz. Los conos sencillos presentan un cambio
medio de 2.1 micras y son mds cortos.a la luz. Los miembros
mas pequefios de los conos dobles presentan un cambio medio
de longitud de 3.5 micras, y los mayores de 2.7 micras. La
emigracién del pigmento es ligera, y su media es 1.6 micras,
pero cuando se combina con el cambio de longitud de las células
visuales produce una emigracién efecti va igual a su suma. El
trabajo termina con consideraciones tedricas sobre los cambios
fotomecdnicos y la teorfa de la duplicidad bajo un punto de
vista comparado.

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHORS’ ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 6

STUDIES ON THE RETINA

THE STRUCTURE OF THE RETINA OF ALLIGATOR MISSISSIPPIENSIS
AND ITS PHOTOMECHANICAL CHANGES

HENRY LAURENS AND S. R. DETWILER

Osborn Zoological Laboratory, and the Anatomical Laboratory, School of Medicine,’
Yale University

THIRTEEN FIGURES

INTRODUCTION

Owing to the variation in kind and distribution of the visual
cells, the reptilian retina offers an interesting field of investi-
gation in structure and function, with particular reference to
the probable functions of the rods and cones under diurnal and
nocturnal conditions (duplicity theory). Detwiler (16) studied
the structure and photomechanical changes in the retina of a
number of turtles and of lizards, and with the idea of making
another contribution to our knowledge of the conditions holding
in the reptiles the present investigation was undertaken. The
work was started in the spring of 1917, but was necessarily
abandoned, and only recently has it been possible to resume it.

The literature on the crocodilian eye is small and rather
unsatisfactory. Heinemann (’77), in a study of the eyes of
Mexican reptiles, described the visual cells of Crocodilus rhom-
bifer cuv. as consisting of characteristic rods alternating with
considerably fewer and shorter cones. In addition to the typical
rods he found a much less numerous kind, with very long outer
segments of platelet structure. The cones are described as
being also of two kinds, thick and slender, neither of which
contain colored oil drops, and which occur singly and together
to form double cones. There is an ellipsoid with a small central
body in the peripheral portion of the inner segment of both and

207

208 HENRY LAURENS AND S. R. DETWILER

a paraboloid in the slender cones. He could distinguish the
cone nuclei from the rod nuclei by their more spherical form
and larger size, but describes both as occupying a single layer.

Tafani (’83) studied the retina of the crocodile, Champsia
lucius, and found rods to be the predominant type of visual cell.
In the anterior part of the retina there are practically no cones,
but as the fovea is approached they become gradually more
numerous than the rods. He found, unlike Heinemann, only
one kind of rod. He describes the cones as being short, with a
barrel-shaped inner segment, but with an outer segment similar
to that of the rods, although his figures do not bear out his
description. The differences between the nuclei of the cones and
of the rods, which occur in a single layer, he considers too slight
and inconstant to be considered as of any significance.

Chievitz (89) described in some detail the pigmented epi-
thelium and the tapetum. In the eye of Alligator mississip-
piensis the tapetum extends through the entire upper half of the
retina in the form of a bright band, reaching nearly to the ora,
while its lower margin lies about 2 mm. above the entrance of
the optic nerve in a 32-cm. specimen. In this bright band he
found a fovea in the form of a very superficial, narrow furrow
with thickened edges, and running horizontally across the entire
tapetum about 1 mm. from its lower edge. He could not decide
whether rods as well as cones occurred. In a vertical section
of the eye of Crocodilus intermedius the tapetum is seen as a
longitudinal bright stripe in the middle of the pigmented epi-
thelium. In this region the middle part of the epithelial cells
contain a number of fine, whitish, opaque granules of guanin,
which when removed leave the middle portion of the cells color-
less, while the choroidal and the vitreal portions contain melanin.
The nucleus lies in the guanin-containing portion, directly against
the basal pigment. At the margin of the tapetum black pigment
is present in almost the entire cell; toward the middle, the vitreal
pigment is gradually reduced and eventually exists only in the
form of isolated, iregularly distributed small masses, between
which the guanin comes to the edge of the cells. In the alligator
the pigment in the choroidal portion of the cells is sparse and

RETINA OF ALLIGATOR MISSISSIPPIENSIS 209

the nuclei are very close to the basal cell boundary. The pig-
ment processes reach as far as the inner segments of the visual
cells, the outer segments being deeply imbedded in the guanin-
containing portion of the epithelium.

Krause (’93) described the visual cells in the retina of Alligator
mississippiensis and quotes from Hofmann’s description of the
retina of Crocodilus vulgaris. In the alligator, Krause considers
that the rods could be taken for small cones, because the slender
inner segments are slightly tapering, while the outer segments
are almost cylindrical. The inner segments of the cones, on the
other hand, are thick and the outer segments short and pointed.
Hofmann describes the rods of the crocodile as numerous except
in the fovea and the surrounding regions of the retina. They
are very similar to the red rods of the frog, and Krause reproduces
a figure from Hofmann of a rod and cone. ‘The cones are single
and double. Krause reproduces (again from Hofmann) cones
with very long, pointed outer segments from the fovea. There
are no rods in the area and only single cones, the inner segments
of which become narrower as the fovea is approached. Accord-
ing to Krause, the nuclei of the visual cells in Alligator mississip-
piensis are all in one row, the cone nuclei being rounder than those
of the rods. In the crocodile, Hofmann says that they occupy
two rows, with the rod nuclei next to the external limiting
membrane.

Abelsdorff (98) considers that very strong support is given
by the conditions in the reptilian retina to the view first put
‘forth by Max Schultze, that the rods serve for the reception of
weak and colorless light stimuli. He ealls attention to the fact
that most reptiles have practically only cones, the exceptions
being the geckos (in some of which the cones seem to be entirely
lacking), the crocodiles, and the boa. These are all nocturnal
animals. The crocodile, he says, on account of its rod-rich
retina, is not only capable of seeing in a very weak light, but can
find its way about in pitch darkness, this property being enhanced
by the light reflecting tapetum in the upper portion of the eye,
the rods being thereby doubly stimulated. Abelsdorff points
out that it is particularly in water that the upper part of the eye

210 HENRY LAURENS AND 8. R. DETWILER

needs an increase in intensity of the light impression more than
does the lower part of the eye, because the upper portion receives
only what little light may be reflected from the bottom.

He figures the tapetum in a sagittal section where it can be
seen in the upper portion of the eye between the choroid and
retina proper, going over in the lower portion of the eye, with
a gradually increasingly thick black border, into the guanin-free,
melanin-containing portion of the pigmented epithelium. At-
tempts to demonstrate a change in position of this pigment in
light and darkness were unsuccessful. It is interesting to note
in this connection that Garten (’07, p. 89) considered it worth
while to have this experiment repeated, which he did, with
results (p. 108) substantiating those of Abelsdorff. Abelsdorft
describes the rods of the alligator as being similar to those of
frogs, but of narrower diameter. He found that the visual
purple, investigated by direct observation of the opened eye
as well as ophthalmologically (assisted therein by the presence
of the tapetum), was not confined to the upper portion of the
eye, but, by turning the retina over and looking at it from the
visual cell side, could be seen as well in the lower portion. He
investigated the bleaching of the purple in daylight as well as
the relative amount of bleaching and the time relations upon
exposure to light of various wave length. The fact that the
purple was seen throughout the eye would seem to indicate that
the rods occurred throughout the retina, although, if one chose
to follow Edridge-Green, it might be assumed that the purple
diffused into the regions where the rods were not present.

Finally, Garten (’07, p. 109) describes the visual cells in
Alligator lucius. In the upper part of the retina (guanin portion)
there are large cylindrical rods only, which are surrounded in
the light as well as in the dark eye by a mantle of guanin. This
part of the retina is absolutely cone free. In the lower portion
of the retina the visual cells are relatively small, possessing a
very thin tapering outer segment, which in light as well as in
dark eyes is buried in pigment. Garten considers these to be
all cones. He refers to Abelsdorff as having described visual
purple in the lower portion of the retina, and thinks, since he

RETINA OF ALLIGATOR MISSISSIPPIENSIS Oat

(Garten) has shown that there are only cones in this region, that
this matter should be reinvestigated.

The conditions in the alligator eye Garten uses in substanti-
ation of his conception of the functional value of the migration
of pigment in connection with the movement of the visual cells.
Since he finds the two parts of the retina containing exclusively
cones or rods, he argues that pigment migration therefore should
not take place, because the perceiving power of the eye would
not thereby be in anyway enhanced. Garten calls attention
to the importance of the fact that the conical visual cells go over
very gradually into those of rod form, and points out that it is
exactly at this transition place that Chievitz localized the fovea.

In two general reviews by Piitter (12) and Franz (13) the
conditions found in the crocodilian retina are summarily given,
but no new matter contributed. It should be recalled that
Piitter is of the opinion that, although some of the elements in
the reptilian retina may be cylindrical in form (that is rod-like),
they are all nevertheless to be regarded functionally as cones,
on account of their dendritic mode of connection with the bipolar
. cells.

METHODS

The alligators, which were between 45 and 55 cm. long, were
treated as follows. Two animals were placed in a dark room
for twenty-four hours, at the expiration of which time one of
them was removed to bright diffuse daylight for seven hours.
‘At the end of this time both were killed. The upper jaw, with
the eyes, was removed with a pair of large bone forceps, bisected,
and dropped into a large dish containing an abundance of Per-
enyi’s fluid. The time required for this operation did not exceed
thirty seconds, and in the dark was carried out in the weak
light from a photographic lamp. The pupil of the allgator is
vertical, and when the animals are placed in light. the aperture
remaining after a few seconds’ exposure is but a mere slit. The
pupillomotor reaction is so decisive, characteristic, and easily
measurable, that experiments have been begun on the relative
efficiency of spectral lights upon it. These will be reported
later.

| es HENRY LAURENS AND S. R. DETWILER

The halves of the upper jaw containing the eyes were allowed
to remain in the fixing fluid for an hour, in light or darkness,
respectively, without being disturbed. After the expiration of
this time, they were carefully dissected out and dropped into

Fig. 1 Diagrammatic drawing of sagittal section of the epithelial pigment
layer. The guanin is indicated by the lightly stippled, the melanin by the heavily
stippled area; the choroid is shown by horizontal lines. X 11.

fresh fixative, where they remained from four to five hours
longer. Sagittal and horizontal sections 104 thick were made,
stained in eosin and toluidin blue, iron haematoxylin and eosin,
or in Ehrlich’s haematoxylin and eosin. All methods yielded
good results.

RETINA OF ALLIGATOR MISSISSIPPIENSIS PALL

ANATOMICAL

Epithelial pigment layer. A retinal tapetum occurs in the
dorsal and posterior portions of the retina to within 1.5 mm.
of the entrance of the optic nerve. It is formed by the inclusion
of guanin in the cells of the epithelial layer. The relative amount
and distribution of the guanin and the ordinary melanin is shown
in figure 1. In the region designated by B the epithelium is
relatively devoid of melanin, which forms a narrow border of a

Fig. 2 A portion of the tapetum designated by the letter B (fig. 1), showing
broad zone of guanin, a narrow vitreal border of melanin, and a few scattered
needles of pigment near the choroidal margin. 665.

Fig. 3 A portion of the epithelial layer designated by letter C (fig. 1), show-
ing choroidal guanin-containing portion and vitreal melanin-containing portion.
X 665.

Fig. 4 <A portion of the epithelial pigment layer corresponding to the region
designated by letter D (fig. 1). > 665.

few needles along the vitreal margin and occurs also as scattered
needles here and there in the choroidal portion of the cells (fig. 2).
As the optic nerve is approached, the amount of melanin grad-
ually increases as the guanin decreases, until within about 1.5
mm. above the entrance of the optic nerve guanin is no longer
found. The gradual increase in the amount of melanin as the
optic nerve is approached is seen in figures 2, 3, and 4, which
show in detail the condition as found at the levels B, C, and D
in figure 1. Above the level B in figure 1 the guanin again
gradually decreases in amount and the melanin shows a corre-
sponding increase (level A, fig. 1).

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

214 HENRY LAURENS AND S. R. DETWILER

The guanin is light grayish-brown in color (stained sections),
finely granular and fairly uniformly distributed through the
cell (fig. 2). In the tapetum (level B, fig. 1), the guanin-con-
taining protoplasm, although covering over the outer segments
of the visual cells, does not show the finger-like projections which
so typify the melanin-containing portions of the epithelium
(figs. 2, 3, and 4).

In the lower portion of the retina the epithelial layer is entirely
devoid of guanin. Here the melanin, in the form of delicate
brownish-black needles, occupies the entire cell body and finger-
like processes of the cells which project over and embrace the
outer segments of the visual cells. The nuclei of the epithelial
layer are spherical and occupy the choroidal portion of the cell
body. Light and darkness have no effect on their shape and
position.

Visual cells. The retina of Alligator mississippiensis contains
both rods and cones, differing in this respect from the retinae
of turtles and lizards (Detwiler, ’16). The two kinds of visual
cells in the alligator are not uniformly distributed, the cone-rod
ratio changing in different parts of the retina. Histological
examination of the retina has yielded the significant fact that
no portion is rod or cone free and that there is no gradual transi-
tion from the conical elements into rod-like forms, as Garten
(07) claims. There are, however, areas which predominate in
rods as well as areas which contain only a few rods. Viewing
the retina as a whole, it can be said with justice that it is char-
acteristically a red-retina.

Rods. The structure of the rod is uniform throughout the
retina. It consists of an inner segment composed of a cylindrical
myoid and an ellipsoid and a cylindrical outer segment (fig. 5).
No rods with conical outer segments could be found. The
rod nuclei are typically oval in shape and le just beneath the
external limiting membrane projecting above it for variable
distances in both dark and light eyes, and thus form the outer
part of the external nuclear layer.

Cones. There are two kinds of cones, of which the predomi-
nating type is a large thick visual element very similar to that

RETINA OF ALLIGATOR MISSISSIPPIENSIS 215

found in the turtle retina. The inner segment consists of a
short broad myoid, a broader refractive paraboloid, and an
ellipsoid, while the outer segment is relatively short and conical
(fig. 5). This type of cone is found particularly in the posterior
and ventral portion of the retina. The second type of single
cone (not very numerous), which is found only in the ventral
portion of the retina, has a considerably longer myoid and a

Fig. 5 A portion of the visual layer showing rods and large single cones.
< 935. Drawing made at about 2.5 mm. above the entrance of the optic nerve
(level C, fig. 1).

Fig. 6 Elongated single cone and a neighboring rod taken from the ventral
portion of the retina. X 935. Animal in darkness for twenty-four hours.

Fig. 7 A double cone taken from the lower portion of the retina. Animal
in darkness for twenty-four hours. X 935.

Fig. 8 A double cone taken from the tapetal portion of the retina. Animal
in diffuse light for seven hours. X 935.

narrower paraboloid. The conical outer segment is long, slender,
and pointed (fig. 6).

The double cones consist of a larger and a smaller member.
The former are very similar in shape to the large single cones.
The latter is characterized by a long slender myoid and the
absence of a paraboloid and has not been observed to occur
singly (figs. 7 and 8). All cones lack oil drops, differing in this
respect from those of the turtle and lizard.

216 HENRY LAURENS AND S. R. DETWILER

Relative distribution of rods and cones. In the upper peripheral
portion of the retina there are very few cones. In the region
designated by A (fig. 1), the ratio of rods to cones is about
95:5. The cones in this region, mostly single and of the large
type, are irregularly distributed. In the typical tapetal portion
(fig. 1, B) they are shghtly more numerous (about 15 per cent).
The majority of the cones, however, in this region are double,
unlike those from the more peripheral portion of the retina. As
the optic nerve is approached the number of cones shows a
corresponding increase. In the region designated by C (fig. 1)

TABLE 1!)

DISTANCE FROM
NUCLEUS TO
OUTER SEGMENT
IN

DISTANCE FRO}
Rents He DISTANCE FROM NUCLEUS |___
x TO PARABOLOID IN |DISTANCE FROM

ELLIPSOID IN ROD NUCLEUS
TO NEAREST

PIGMENT
Sieein Double cones NEEDLE IN
Cones | Rods | Cones | Rods Paes =e

Large Small

IDEN. cols oanmee ISH We || Webs | Woes || OLS | Go) USO TLRS
Tightens ecetes WD | eee) || WAIL | aly |) abe | eh | O55 9.9
IDISREAG. Ses aesl| Ze 20 | Beth | 2G | 2) oth |) Bee 1G
1 Total number of rod measurements, 280.
Total number of cone measurements, 190.

Total number of pigment measurements, 60.
Measurements made at about 2 to 3 mm. distance from the entrance of the
optic nerve, in the region designated by C in figure 1, with the exception of the
measurements of double cones which were taken from region B.

approximately 40 per cent of the visual cells are cones. Here
the double cones are still more numerous than the single cones.
In the guanin-free portion above the optic nerve (fig. 1, D)
cones and rods are about equal in number, the single cones again
exceeding in number the double. The cones outnumber the
rods in the lower portion of the retina, the number increasing
from the region of the optic nerve toward the ora serrata, and
including all types. In this region only a few scattered rods are
present. About 75 per cent of the cones in this region are single,
the majority of which are like that shown in figure 5. The
small type (fig. 6) is relatively scarce.

RETINA OF ALLIGATOR MISSISSIPPIENSIS PALet

The cone nuclei are easily distinguished from the rod nuclei;
the former being somewhat pear shaped, the latter more or less
oval or elliptical. Furthermore, the cone nuclei occupy a deeper
level than the rod nuclei and constitute a second row.

EXPERIMENTAL

Effects of light: Rods and cones. When sections of eyes of
animals which have been exposed to diffuse light are compared
with sections of eyes of animals kept in darkness, it is seen that
there is an average difference in the lengths of the rod myoid of
4u (table 1). The relative lengths of the rod myoids in the dark
and light conditions are shown in figures 9 and 10, as well as

TABLE 2
DISTANCE FROM ROD NUCLEUS TO ELLIPSOID
(MYOID) IN
Region
3mm from optic Tapetum (region B,
nerve fio)
IPT TSS st depeche cists tenacers eievores aston Gos oketien st S1oKs 11.5 8.9
1D aire ea A Go ree ee ae ees el de oe C25 6.9
DIET ETE COA Ee chs c cleus oiacusie apekelearteneteers 4.0 2.0

diagrammatically in figure 11. The change in the length of the
rods is found to be less extensive in the tapetal area (fig. 1, B),
where rods predominate, than in the region close to the optic
nerve, where the cones are considerably more numerous (table 2).

The effect of light on the cones is not so easily demonstrable.
The results, however, of several series of measurements (table 1)
show that the cone myoids of light eyes are slightly shorter
(2.1u) than those of dark eyes (figs. 11, 12, 13). The contraction
is found to occur in the double as well as in the single cones.
The measurements of the double cones show that the myoid of
the smaller member has shortened more than that of the larger
(figs. 7 and 8). Further evidence of changes in the length of the
visual cells in light and darkness is afforded by a study of the
relative positions of the cone and rod ellipsoids. In the dark

218 HENRY LAURENS AND S. R. DETWILER

condition (fig. 9) the ellipsoids of the single cones are usually
found to be on the same level as the rod ellipsoids. On the other
hand, in the light condition (fig. 10) the cone myoid is seen to
occupy a level closer to the external limiting membrane than that
of the rod ellipsoid. This change in relative position is the result
of the combined effect of rod elongation and cone contraction,

ad i
> ae
=o

eae:

Mack,
5S

ere

eer)

—

HE 4

roe

Fig. 9 A portion of the retina 3 mm. from the entrance of the optic nerve
(region C). Animal in darkness for twenty-four hours. XX 9365.

Fig. 10 <A portion of the retina from a region corresponding to that designated
in figure 9. Animal in diffuse light for seven hours. X 935.

and is clearly shown by an examination of the rod and cone
shown on the right-hand side of the diagrammatic figure 11.
Pigment migration. The differences in position of the pigment
in light and dark eyes is slight. A series of measurements of the
distance between the rod nuclei and the nearest pigment needle
(table 1), as well as measurements of the distance between the

RETINA OF ALLIGATOR MISSISSIPPIENSIS 21

external limiting membrane and the nearest pigment needle,
show that the pigment in the light eye is about 1.6 nearer the
external limiting membrane than in the dark eye. This almost
insignificant amount, however, when combined with the distance
the rod myoid has elongated in the light (4u) gives an effective
migration which is clearly illustrated in figures 9 and 10, as well
as in figure 11. An examination of these figures shows that in
the dark condition the choroidal portion only of the rod ellipsoid
is covered with pigment, while in the light condition the entire

Fig. 11 Diagrammatic drawing compiled from 530 measurements showing.
the effects of light and darkness on the visual cells and on the position of the
pigment. The left-hand side of the figure represents the dark condition; the
right side, the light. The drawings were accurately laid out on codrdinate
paper from measurements presented in table A, each millimeter being given a

value of 0.5.
Fig. 12 Single cones. Animal in diffuse light for seven hours. > 935.
Fig. 13 Single cones. Animal in darkness for twenty-four hours. XX 935.

ellipsoid and a portion of the myoid is ensheathed. The amount.
of cone contraction in the light, being only slightly more than
the extent of the pigment migration (table 1), the relation
between the position of the pigment and the cone ellipsoid is
practically the same in both dark and light eyes, the choroidal
portion of the cone ellipsoid being in both covered by pigment
(figs. 9, 10, and 11). The description of the position of the
pigment in relation to the ellipsoids of the visual cells in both
dark and light eyes pertains only to conditions found in the
posterior part of the retina (about 2 to 3 mm. from the entrance
of the optic nerve). Changes in the position of the pigment in.

220 HENRY LAURENS AND 8S. R. DETWILER

the more peripheral region of the retina could not be demon-
strated. Near the margin of the retina the visual cells are
greatly shortened and the pigment is found to extend down
almost to the external limiting membrane.

DISCUSSION

Photomechanical changes of the retina. The question of the
functional significance of pigment migration and the changes in
position of the visual rods and cones in light and dark adaptation
is one about which much has been written (Garten, ’07, and
Helmholtz, ’11). It may therefore appear redundant to add
anything in the way of a theoretical consideration of this function.
But there are a few points which still lack clarity.

In the eyes of those animals in which these changes take place
they represent a mechanism for the adaptation of the eye to day
and to twilight vision (Herzog, 705; Exner and Januschke, ’06).
In dim light (twilight vision), when the rods alone are capable
of being stimulated to any degree, or in complete darkness, the
pigment moves back and leaves free the spaces between the
rods, resulting in a less complete insulation of these elements.
Under these conditions, with the entrance of a small amount of
light the part played by the individual rods in the reception of
the light is greater, owing to refraction and diffusion, than if the
rods were covered over by a thick mantle of pigment, in which
case only the light which passes through the retina in the direction
of the long axis of the rods could enter them. ‘The presence of
a reflecting tapetum further enhances the favorable conditions.
The cones in twilight vision are not functional, on account of
their high threshold, and they elongate and thus move out of the
way. The rods contract and thereby optimum conditions are
presented for their stimulation.

In bright light (day vision) the pigment by migrating forward
protects the rods, which have a low threshold and which have
been made particularly sensitive by the accumulation of visual
purple in the dark, from too strong stimulation by absorbing
the direct and scattered light. The rods elongate, while the
less sensitive cones are drawn out of the pigment by the con-

RETINA OF ALLIGATOR MISSISSIPPIENSIS Pega

traction of their myoids and are thereby made freely accessible
to the stronger light stimulus, thus presenting optimum con-
ditions for their stimulation.

The fact that. these photomechanical changes have not been
demonstrated in the eyes of man and mammals does not con-
stitute a denial of their taking place in the eyes of other animals,
and there is no ground against explaining as above the phenomena
in such animals. Adaptation-of the mammalian eye is brought
about by different means, viz., the formation and bleaching of
visual purple.

The process of light and dark adaptation is not dependent upon
the phototropic movements of the elements of the retina, but
these movements may take part in the process of adaptation,
that is in the formation and bleaching of visual purple. That
the pigment, as such, has anything to do with the formation of
visual purple is not probable, because visual purple occurs in
the eyes of albinos and in the pigment-free portions of the retina
of many animals, for example, the cat. The significance of the
pigment is probably a purely optical one concerned with the
absorption of scattered light.

In connection with the function of the retinal epithelium in
the formation of visual purple, the paper by Kolmer (’09) is to
be noted. Kolmer finds numerous droplets and granules on and
between the rods in the retina of various vertebrates. These he
regards as secretion products of the pigment epithelium. In
the retina of frogs kept in darkness the droplets and granules are
larger and more numerous than in the illuminated retina, and
after illumination of the eye with direct sunlight are not to be
seen at all. Since they are lacking in the eye of lizards and
snakes, Kolmer assumes that they have something to do with
the visual functions of the rods, the organs of twilight vision, and
perhaps with the appearance of visual purple.

It is interesting to note that pigment migration is still assumed
by some to take place in the human eye. Ramon y Cajal (’11,
p. 363) believes that the function of the pigment is to prevent
the impression of halo, and that the dazzling sensation which one
experiences on going from a dimly lighted place into a bright one

Dao HENRY LAURENS AND S. R. DETWILER

is due to the fact that the pigment, which has moved back into
the body of the epithelial cells as an effect of darkness, requires
some time to be brought out again into the prolongations of the
cells to ensheath the individual visual elements.

Bard (719), in a highly theoretical paper, in many respects
offering views widely divergent from those usually held concern-
ing vision of form and of color, explains many things by the
assumption that both pigment migration and cone contraction
take place in the human eye.

Cobb (19, p. 444) also states that pigment migration takes
place in the human retina. He says:
aside from the changes in the size of the pupil there are two anatomical
factors undoubtedly concerned in dark and bright adaptation: the
exhaustion and regeneration of visual purple (or possibly other photo-
chemical substance); and the migration of the pigment of the hexagonal
cells. This last may be a protecting device that acts fairly promptly,
and has the effect of enclosing the retinal rods, and by its own light-
absorbing qualities reducing the amount of light absorbed by the indi-
vidual rods. It is conceivable that asudden flash of light might antici-
pate this action and produce a strong destruction of the photochemical
material in a short time, before the pigment cells have had time to

react, while with gradual onset of hight the time is adequate for the
pigment cells effectively to assume this protective function.

And on page 445:

Some of the curves strongly suggest two factors playing a part in
dark adaptation . . . . allowing the interpretation that the
results are arising from two more or less independent mechanisms one
of which overtakes the other in effect, at the end of about four minutes.

We believe that in the migration of pigment, the contraction
of the cones, and the elongation of the rods there is exemplified
the response of irritable protoplasm to a definite, adequate
stimulus. In some cases the response is very marked, though
of varying degree (fish, frog, bird); in others it is not demon-
strable (man and mammals). In some it may serve an easily
comprehended purpose; in others, in terms of the theory explain-
ing it, it may seem to be useless. Nevertheless, if it can be
demonstrated (as in the turtle, lizard, and alligator) it cannot
be explained away or ignored because it seems to serve no useful
purpose.

RETINA OF ALLIGATOR MISSISSIPPIENSIS 228

In the eye of the alligator the migration of the pigment and the
change in position of the visual cells seem to be correlated with
the relative distribution of rods and cones. The rods show the
greatest difference in position in light and dark eyes, in the regions
designated by C in figure 1 (rod-cone ratio 60 to 40), and by D
(where the rods and cones are about equal in number), much
less in the region B (rod-cone ratio 85 to 15), and not demon-
strably at all in the region designated by A, or in the region
(E and F) below the optic nerve, where the rods represent only
about 5 per cent of the total number of visual cells. The pig-
ment can be demonstrated to move forward only in the posterior
portion of the retina (regions C and D), thus corresponding to
the regions where the maximum change in position of the rods
takes place. The cones throughout the retina show the same
(slight) degree of shortening in the light, except that the double
cones, which are most numerous in the regions B and C, show a
slightly greater amount of change in length.

Garten (’07, p. 38) weakened the general application of the
suggestion put forth by Herzog (’05) and Exner and Januschke
(06) by observations which seemed to show an extremely high
sensitivity to stimulation by weak light, so that the light con-
dition of the visual cells was considered as assumed in dim light.
Arey (’19) has recently brought forward evidence indicating that
the sensitivity of the retinal pigment and of the rods and cones
is nowhere nearly so high as is generally believed, and the con-
ception that the changes observed in those eyes where marked
effects of light are obtained are adaptive has been thereby placed
on much surer ground.

We should not, however, a priori, deny that light effects similar,
if less marked, changes in retinae so constituted that there can
be no, or little, question of any advantage to be gained by a
correlative shifting of the position of the visual elements. Garten
cites the facts that in the selachians, which presumably have
pure-rod retinae, although the literature on the subject is not
without disagreement,! there is little, if any, pigment, and that

1 Schultze, ’66; Krause, ’76 and ’95; Neumayer, 97; Schaper, 99; Hesse, ’04;
Franz, 705; Retzius, ’05; Garten, ’07; Cajal, 711, and Piitter, 712.

224 HENRY LAURENS AND S. R. DETWILER

in pure-cone retinae, where the pigment, as assumed, is necessary
for the absorption of the ight scattered by the highly refractive
cones, there is practically no pigment migration. But Detwiler
(16) found a demonstrable pigment migration and cone con-
traction in the eyes of both turtles and lizards. Garten further
argues (p. 109) that photomechanical changes should not take
place in the crocodilian eye, because of the exclusive presence
of rods or cones in the various portions of the retina. But, as
we have demonstrated above, the structural conditions, at least
in Alligator mississippiensis, differ from his description, and
changes in the position of pigment and of visual cells do take place
in light and darkness.

The duplicity theory. The duplicity theory of von Kries, or
the theory of the double retina of Parinaud, based on the findings
of Max Schultze (’66), is of the greatest importance in compara-
tive work on vision. The hypothesis is generally regarded as
well substantiated, particularly by the facts of twilight and
day vision. For brief accounts and references to the literature
of the theory and its development the reader is referred to Nagel
(05), Helmholtz (11), and Parsons (’15). Briefly stated, the
theory holds that the rods are sensitive only to ight and darkness,
and by virtue of their power of adaptation in the dark through
the regeneration of visual purple they form the apparatus for
vision in dim light. The cones, on the other hand, are the
apparatus subserving bright vision as well as the perception of
color. But in another way, the rods are the apparatus for
achromatic scotopic vision (twilight vision), the cones the
apparatus for photopic vision (day vision). The cones are not
necessarily assumed to be utterly useless at night, but only rela-
tively so, being quickly fatigued, on account of their high
threshold.

The presence and relative distribution of rods and cones is
therefore a matter of the first importance. But without prej-
udice it can be said that this is a very unsatisfactory matter as
far as the comparative literature is concerned. Early contri-
butions to the histology of the visual neuro-epithelium either
have not been reinvestigated, but assumed to be correct, or

RETINA OF ALLIGATOR MISSISSIPPIENSIS 225

when investigated a divergence in the opinions of later investi-
gators is the rule rather than the exception.

The question as to what constitutes a rod and what a cone
would seem to be a simple matter, but the great variety in the
forms of the cells in different animals makes difficult a gener-
alized classification. The contentions of Pitter (12) for a
functional classification based on the type of connection of the
visual cell with the bipolar, rather than a structural one, are
good if kept within limits, but it seems to us that they are carried
too far. As everyone knows, the foveal cones of man and mam-
mals are cylindrical in shape, and are therefore much more like
rods in general appearance than cones. But their known func-
tion fits in with the general conception of the physiology of the
apparatus for color and bright-light vision. ‘roland (’17) has
pointed out that the shape of the foveal cones suggests that the
function of the cone figure is structural rigidity rather than
differentiation of response.

Whether a visual cell is a rod or a cone is determined by the
presence of one or more of three structural factors, viz., 1) the
shape of the outer segment; 2) the shape of the inner segment,
and, 3) the mode of connection between the visual cell and the
bipolar cell. When we find visual cells which, from their general
form (outer and inner segments), would be called rods, possessing
terminal connections typical of cones (e.g., frogs, diurnal birds,
see Ramon y Cajal, 94, pp. 31, 164, and ’11, pp. 340, 327),
there is no, or very little, reason why they should be called cones
simply because they terminate in dendrites, and considerable
reason for continuing to designate them as rods on functional
as well as structrual grounds. The conditions in the geckos
may be cited as another example. From description and illustra-
tion it would seem as if no more typical rods could be found.
Coupled with their structure there are functional characteristics,
which will be referred to later, and which, from all that we know
about rod vision, indicate that the visual cells are as functionally
typical rods as can be found. Our work on the retina of the
alligator shows that rods as well as cones occur there, structurally
as well as functionally, as exemplified in the inverse changes in

226 HENRY LAURENS AND S. R. DETWILER

light and darkness. But Piitter would call them all cones
because their centripetal termination is similar to that found in
the cones of man.

The designating a visual cell as a rod or as a cone on morpho-
logical grounds is not therefore useless, as Pitter claims, but,
as he also points out the structural basis (form of inner and outer
segments), is brought into line with the functional by what we
know of the respective functions of the visual cells in man, viz.,
threshold, visual acuity, ability to see movement, and vision of
color. The rods are visual elements with a low threshold, but
with possibilities of summated conduction, due to the connection
of more than one of them with a single bipolar cell; the cones
are visual elements with a high threshold and isolated conduc-
tion, based on the histologically found type of connection.

Pitter, in speaking of the conditions in the nocturnal birds,
admits that the visual elements have knob-like endings, and that
the visual cells are morphologically typical cones, although they
have assumed what he regards as the most distinctive character-
istic of rods. Pitter reverses himself here and is, as well, incom-
plete, because, as Ramon y Cajal (’94, p. 104) points out, in
these retinae there are rods ending like those of mammals, while
the cones which have almost entirely similar endings, reach
deeper and come into connection with a different set of bipolars,
so that there is a further morphological differentiation here
between rods and cones.

It does not seem at all certain to us that Hess ('10 and ’13),
by his work on the turtle and on birds, has disproved or weakened
the general truth of the duplicity theory. He claims to have
demonstrated an adaptation in the turtle retina, where there
are cones only. The claim that he and Katz and Révész (13)
make, that adaptation in diurnal birds is a function of the cones,
does not seem warranted, owing to the fact that rods are present
in considerable numbers, as Hess himself describes, particularly
in connection with the presence of visual purple. ‘The phenom-
enon, similar to the Purkinje phenomenon, which they state
to have observed, may therefore, and most naturally, be a
function of the rods and not of the cones. Katz and Révész

RETINA OF ALLIGATOR MISSISSIPPIENSIS apa

(13) also state that the rods of nocturnal birds (owls) in bright
light approach, or are similar, in function to cones. But this
is without anatomical foundation for the simple reason that
the retinae of such birds contain numerous cones (Garten, 07;
Hess, 713, p. 581) which show, with the pigment, photomechanical
changes. In this connection the view of Parsons (15, p. 204)
may be quoted:

If we regard the rods as the more primitive type of visual neuro-
epithelium, as we are probably justified in doing, the persistence of
recognizable rod attributes in the cones, even if modified, differentiated,
and rendered more complex, might well be expected. Apart therefore
from the difficulties of isolating the physiological results of excitation of
the rods from those of excitation of the cones it may be anticipated
that the latter cells will retain some measure of the functions which are
in the highest degree characteristic of their prototypes. Hence, if it
should ever be conclusively proved that the rod-like foveal cones of the
human eye possess some trace of visual purple and are endowed with
some slight degree of light-adaptation it would not be surprising;
neither, on the other hand, would it militate seriously against the view
that the rods and cones have become essentially diverse in function.

Troland (’16) has demonstrated by careful experiments, cor-
roborating the earlier work of v. Kries and Nagel, that the phe-
nomenon of Purkinje does not take place in the rod-free portion
of the human retina. And Watson (715) and Lashley (16)
show that Hess’ contention that the spectrum is shortened for
the bird’s eye as compared with the range of wave lengths seen
by the human eye, is not correct.

It is not out of place to add that in the condition known as
night-blindness, in which the rods are insensitive, or practically
so, dark adaptation is almost abolished or is much slower than
normal, and that Purkinje’s phenomenon is much less marked
than in the normal eye or absent altogether.

One further remark concerning Hess’ work. He (10) claims
that many turtles are nocturnal and cites authorities supporting
his contention. Ramon y Cajal (11, p. 361) says that reptiles
in general (in which of course he is incorrect, witness the alli-
gator and the gecko) do not see in darkness. Rochon-Duvig-
neaud (717) does not believe that turtles can be called nocturnal
because they are incapable of flight from an enemy or of pursuing

228 HENRY LAURENS AND S. R. DETWILER

prey, and he thinks that they detect the plants and insects upon
which they feed by the sense of smell. Rochon-Duvigneaud
asks the very apt question in reference to Hess’ claim of adapta-
tion in the fowl, why it is, if they possess the power of adaptation
almost as well-developed as that of man, that they go to roost
long before man ceases to enjoy good vision.

With reference to the duplicity theory and the distribution
of rods and cones, the work of Abney (16 and 717) is most
interesting and important. By determining the minimum
intensity of ight of various wave length which can be perceived
at the fovea and up to ten degrees from its center Abney and
Watson (716) obtained results indicating that in some cases the
fovea of man is free from rods, which increase rapidly as the
fovea is left, while in others there is a plentiful supply of rods
at the fovea, their distribution, at any rate up to ten degrees,
being very nearly uniform, and, if anything, in excess at the
fovea. In the first group the light appears colored as long as
it is visible at all, particularly in the green. In the second group
the light loses color a considerable time before it is extinguished,
except in the red. In other words, there is an achromatic inter-
val. Abney (717) later examined persons suffering from night-
blindness, in which the rods are generally believed to be non-
functional, for extinction of color from the red to the blue.
The light was found to vanish when its color was extinguished,
so that the same reduction in intensity of the light was the
threshold for both light and color, similarly to the cases men-
tioned above where there are no rods at the fovea, indicating
that there is an absence of sensitive rods in the whole retina of
the night-blind.

If dark adaptation is directly associated with visual purple
the vision of an animal possessing rods only as compared with
that of an animal with cones only, both in respect to ability to
see in light and darkness, after longer or shorter adaptation to
the one or the other condition, and as well the relative stimulating
value of spectral lights of equal energy, should be expected to be
markedly different, quantitatively as well as qualitatively. Now
we have in the reptiles admirable subjects for investigating this

RETINA OF ALLIGATOR MISSISSIPPIENSIS 229

very question, especially in the lizards, for example, the geckos
as compared with other lizards and particularly the horned toad
which has a retina, from all structural indications, peculiarly
adapted for day vision only. By investigating in the gecko
and the horned toad the relative visibility of wave lengths of
equal energy and the relative powers of adaptation, we will
obtain information concerning the question of the selective
response of rods to different wave lengths as compared with
that of cones. Visual purple absorbs all wave lengths except
a little red and violet. The rods therefore are presumably
sensitive to all wave lengths except the extreme red and violet.
Since rods are in general ‘color-blind,’ there is opportunity here
of differentiating between wave length and stimulating value.

In connection with the question as to which kind of visual
cell represents the more primitive condition, the histogenesis
of the retina is worthy of investigation. But the histogenesis
of the neuro-epithelium is a subject about which our knowledge
is most imperfect. In general, the rods are regarded as the more
primitive type of visual cell (Graham Kerr, 719), while the cones
are considered as specialized rods.

According to Leboucq (’09), the two kinds of visual elements
develop simultaneously and are distinguishable only by the fact
that the axis of the diplosome is perpendicular to the surface in
the case of the rod and parallel to it in the case of the cone.

Cajal (11, p. 356) says that the cones and rods evolve in the
same way and that it is difficult to distinguish between them at
the beginning (also Fiirst, ’04).

Cameron (711) reiterates a view earlier championed by him-
self (705) as well as by Bernard (’03) to the effect that the cones
represent early stages in the formation of rods. We mention
this here because of the fact that in looking over some slides of
early amphibian embryos, the eyes do seem to contain nothing
but cones or conically shaped elements. Graham Kerr (p. 137)
finds that the visual elements (rods) in the retina of Lepidosiren
shorten in the light and elongate in the dark, which is similar to
to the usual behavior of cones and contrary to that of rods.

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

230 HENRY LAURENS AND S. R. DETWILER

The shape of the pupil is a subject of interest in a paper dealing
primarily with the eye of the alligator because of the vertical
slit form of the pupil in this animal. In two recent articles
reference is made to this subject. Rochon-Duvigneaud (17),
in listing the characteristics of the eye of the geckos which make
them adapted to nocturnal vision, includes the form of the
pupil—a vertical slit—which shows a rapidity of movement
surpassing that of the human pupil and approaching that of
birds. In dim light the pupil is a large oval, or even round, as
in the cat. In bright light it is closed completely. According
to Rochon-Duvigneaud, a round pupil can dilate as well as an
oval one, but it cannot be entirely closed, and he believes that
it is in the way of a protection against an excess of light in an
animal adapted to twilight vision that an oval pupil finds its
chief function. It is possible to imagine that a pupil in the form
of a vertical slit can be opened wider than if it were round (for
example, the cat).

Hartridge (’19) views the function of a vertical slit pupil (as
seen in the cat) from another angle, viz., the function of the lens
and the aberrations caused thereby, and the habits of life of
the cat family in the nature of their being tree-climbing and tree-
dwelling animals which hunt their prey chiefly at night. An
oval pupil in which the long axis is vertical will cause the lens
system to form images in which the aberrations of horizontal
contours are greater than those belonging to vertical contours.
The contours of trees and their branches are principally vertical,
therefore if the illumination of the image formed on the retina
could be increased by sacrificing the definition of horizontal
contours it would be an advantage. ‘This is effected by the use
of the oval pupil since the aberration of vertical contours is little
greater than that of a circular pupil of the same horizontal
diameter, while the intensity of the image formed on the retina
is as much the greater as the vertical diameter of the oval is
greater than that of the circular pupil.

It seems more likely to us that the function of a vertical slit-
shaped pupil is for protection; that is, to permit of its being
almost, if not entirely, closed. In thinking of animals that have

RETINA OF ALLIGATOR MISSISSIPPIENSIS 231

this type of pupil, we find that many of them are essentially
nocturnal in habits—the gecko, the alligator, the cat—animals
in which the rod functions are predominant over those of the
cones. At night or in dim light the pupil of the cat is wide open
and round. Furthermore, in the daytime, when the pupil is a
vertical slit or oval, cats hunt and catch a great deal on the ground,
for example, birds and squirrels, as well as chase the rapidly
swirling leaves. The cat is furthermore said to have very defec-
tive daylight vision and to be colorblind (DeVoss and Ganson,
15). The alligator hunts along horizontal contours, and yet
one finds that the shape of the pupil is a vertical slit.

SUMMARY

1. The eye of the alligator possesses a well-developed retinal
tapetum in the dorsal and posterior portions of the retina to
within 1.5 mm. of the entrance of the optic nerve. It is formed
by the inclusion of guanin in the cells of the epithelial layer
(figs. 1, 2, 3, 4). The pecten consists of a slightly raised pig-
mented cap covering the entrance of the optic nerve.

2. Typical rods and cones occur throughout the retina, the
cone-rod ratio being different for different regions, but character-
istic for particular regions (p. 216).

3. The rods are all of one type (fig. 5), the cones of two, thick
and thin, of which the first is by far the more numerous, occurring
particularly in the posterior and ventral portions of the retina.
Those of the second type are found only in the ventral portion
and are not numerous (fig. 6). Double cones also occur (figs.
7 and 8). None of the cones contain oil drops. The rod nuclei
are oval or elliptical in shape, and the majority of them project
through the external limiting membrane for a variable extent,
the rest of them being just under it. The cone nuclei are pear
shaped and, occupying a deeper level than the rod nuclei, con-
stitute a second row (figs. 5, 6, 9, 10).

4, The rods show a change in length of their myoids averaging
4u, being longer in the light and shorter in the dark (figs. 9, 10,
11 and table 1). The single cones (thick and thin) show an
average change of 2.lu (figs. 11, 12, 13 and table 1). The

Dae HENRY LAURENS AND ‘S. R. DETWILER

smaller member of the double cones shows an average change
in length of 3.5u, the larger member of 2.7u (figs. 7, 8 and table 1).

5. The actual change in position of the pigment between light
and dark eyes is slight, averaging but 1.64; but when combined
with the change in length of the visual cells, gives an. effective
migration equal to the sum of the two (figs. 9, 10, 11).

6. A theoretical consideration of photomechanical changes
and of the duplicity theory from a comparative point of view
is appended.

BIBLIOGRAPHY

ABELSDORFF, G. 1898 Physiologische Beobachtungen am Auge der Krokodile.
Arch. f. Anat. (u. Physiol»), 8. 155-167.

ABNEY, Sir W. DE W., anpD Watson, W. 1916 The threshold of vision for dif-
ferent colored lights. Philosoph. Trans. Roy. Soc., vol. 216A, pp. 91-
142.

ABNEY, Sir W. pe W. 1917 Two cases of congenital nightblindness. Proc.
Roy. Soc., vol. 90B, pp. 69-74.

Argy, L. B. 1919 A-retinal mechanism of efficient vision. Jour. Comp. Neur.,
vol. 30, pp. 348-354.

Barp, L. 1919 Du réle des bétonnets et des pigments rétiniens dans la per-
ception des formes et des couleurs. Jrl. de Physiol. et de Path. gén.,
T. 18, p. 276-294.

Baytiss, W. M. 1918 Principles of general physiology. Second edition.
Longmans, Green & Co.

BrernarD, H. M. 1903 Studies in the retina. Q. J. Micro. Sci., N. S. vol. 46,
pp. 25-75.

CaMERON, J. 1905 The development of the retina in amphibia. Jrl. Anat.
and Physiol., vol. 39, pp. 185-153, 332-348, 471-488.
1911 Further researches on the rods and cones of vertebrate retinae.
Jr]. Anat. and Physiol., vol. 46, pp. 45-53.

Cuinvitz, J. H. 1889 Untersuchungen iiber die Area centralis Retinae. Arch.
f. Anat. (u. Physiol.), Suppl. Bd., S. 139-196.

Coss, W. C. 1919 Dark-adaptation with especial reference to problems of
night-flying. Psychol. Rev., vol. 26, pp. 428-453.

DetwiterR, 8. R. 1916 The effect of light on the retina of the tortoise and the
Herik Jour. Exp. Zoél., vol. 20, pp. 165-191.

DeEVoss, J. C., anp Ganson, R. 1915 Color blindness in cats. Jour. Animal
Behav., vol. 5, pp. 115-139.

Exner, S., UND JANUSCHKE, H. 1906 Die Stibchenwanderung im Auge von
Abramis brama bei Lichtveriinderungen. Ber. d. k. k. Akad. d. Wiss.
zu Wien, Math.-Nat. K1., Bd. 115, S. 269-280.

Franz, V. 1913 Sehorgan. Oppel’s Lehrb. d. vergl. mikro. Anat. d. Wirbel-
tiere. G. Fischer. ~

RETINA OF ALLIGATOR MISSISSIPPIENSIS Daa

First 1904 Zur Kenntnis der Histogenese und des Wachstums der Retina.
Lund’s Universitets Arsskrift, Bd. 40, adfeln 1. No. 1 (cited from
Cajal, 711).

GarTEN, S. 1907 Die Verainderungen der Netzhaut durch Licht. Graefe-
Saemisch Handb. d. ges. Augenheilk. I. Teil, III. Bd., XII. Kap.
Anhang., S. 1-130.

Harrripvce, H. 1919 The shape of the pupil in various animals. Jour. Phys-
iol., 53, vi.

HEINEMANN, C. 1877 Beitrige zur Anatomie der Retina. Ar. f. mikroskop.
Anat., Bd. 14, S. 409441.

Hetmuoutz, H. von 1911 Handbuch der physiologischen Optik. Dritte Aufl.,
Bd. 2. L. Voss.

Herzoc, H. 1905 Experimentelle Untersuchungen zur Physiologie der Bewe-
gungsvorgiinge in der Netzhaut. Ar. f. (Anat. u.) Physiol., S. 413-
464.

Hess, C. 1910 Untersuchungen iiber den Lichtsinn bei Reptilien und Amphib-
ien. Pfliiger’s Archiv, Bd. 148, S. 255-295.
1913 Gesichtssinn. Winterstein’s Handb. d. vergl. Physiol., Bd. 3,
S. 555-840. G. Fischer.

Hesse, R. 1904 Ueber den feineren Bau der Stibchen und Zapfen einiger Wir-
beltiere. Zool. Jahrb., Suppl. Bd. 7, 8. 471-511.

Katz, D., UND Rfévész, G. 1913 Ein Beitrag zur Kenntnis des Lichtsinns der
Nachtvégel. Zeit. f. Sinnesphysiol., Bd. 48, S. 165-170.

Kerr, J. GrauHAM 1919 Text-book of embryology, vol. 2, Macmillan & Co.

Koutmer, W. 1909 Ueber einen sekretartigen Bestandteil der Stabchen-zapfen-
schicht. » Pfliiger’s Archiv, Bd. 129, S. 35-45.

Kravusp,W. 1876 Die Nervenendigung in der Retina. Ar. f. mikroskop. Anat.,
Bd. 12, S. 742-790.
1893 Die Retina der Reptilien. Intern. Monatschr. f. Anat. und
Physiol., Bd. 10, S. 13-31, S. 33-62, S. 68-84.
1895 Die Retina. Nachtrage. Ibid., Bd. 12, S. 176-186.

Lasauey, K.S. 1916 The color vision of birds. I. The spectrum of the domes-
tic fowl. Jour. Ani. Behav., vol. 6, pp. 1-26.

Lesouce, G. 1909 Contribution a l’étude de l’/histogénése de la rétine chez les
Mammiféres. Arch. d’Anat. Microscop., T. 10, p. 555-605.

Nace, W. 1905 Handbuch der Physiologie des Menschen, Bd. 3. Fr. Vieweg
& Sohn. ,

Neumayer, L. 1897 Der feinere Bau der Salachier-Retina. Ar. f. mikroskop.
Anat., Bd. 48, S. 83-110.

Parsons, J. H. 1915 An introduction to the study of colour vision. Putnam’s
Sons.

Pirrer, A. 1912 Organologie des Auges. Wm. Engelmann.

Ramon y Casat, 8. 1894 Die Retina der Wirbelthiere. Bergmann.
1911 Histologie du Systéme nerveux de homme et des Vertébrés.
A. Maloine.

Rerzius, G. 1905 Zur Kenntnis vom Bau der Selachier Retina. Biol. Unters.
(N. F.), Bd. 12, S. 55-60.

234 HENRY LAURENS AND S. R. DETWILER

Rocuon-DuvicNeaup, A. 1917 Les fonctions des cones et des batonnets. Indi-
cations fournies par la physiologie comparée. Ann. d’oculistique,
Nov., p. 16.

Scuarger, A. 1899 Die nervésen Elemente der Selachier-Retina in Methylen-
blaupriparaten. Festschr. z. 70. Geburtstag von C. von Kuppfer,
S. 1-10.

Scuuttzr, M. 1866 Zur Anatomie und Physiologie der Retina. Ar. f. mikro-
skop. Anat., Bd. 2, S. 175-286.

Tarant, A. 1883 Parcours et terminaison du nerf optique dans la rétine des
crocodiles (Champsia lucius). Ar. ital de Biol., T. 4, p. 210-233.

TroLanp. L. T. 1916 Apparent brightness; its conditions and properties.
Trans. Illum. Eng. Soc., Vol. 11, pp. 947-966.
1917. The nature of the visual receptor process. Jour. Opt. Soc. of
America, vol. 1, pp. 3-15.

Vincent, 8. B. 1912 The mammalian eye. Jour. Ani. Behav., vol. 2, pp. 249-
255:

Watson, J. B. 1915 Studies on the spectral sensitivity of birds. Carnegie
Institute of Washington, Dept. of Marine Biol., no. 211, pp. 85-104.

fon sna’

f WMA av ag

slot Prt a
‘i fe

iia a al ee, rr ee we bee
eae winroLigtedy t

mo Isbagtos wale)
spanorwista sag Oh: 5, ae oalied dba sane avis : pbaattie
Sas Cate tiie sabat rst pega Bibs. ays ae oe nein y,

ihe uy “id ‘sr datlaineed eh is unlit. ‘eal 1 At ORO iid a hy) | ft
} Miylées ast cab Pipe bes Licaon pte ol my" a): q Stat
Pr Mi ty SO a shivihy ta sex Fy’ sdeevts a 7 a

: prin. ati pis avin sabia toy use AN Us A Hut m3 i ao we ie V

By b8e 4 Hie) Some ait, EvU IS va ad wigs in iE by arial SL. ie oe .
Sy eaten maith ainsi Bam a iyece “ait hiss Boog esg: as (nba 4 se
Sa ceajliia.a 3 Vaile pa: oiled. > willingna a ston. ri ae bay yuk maursieel hs i
Bim rg HOw, rwbienoy asl < f= seh Tee BiG os We : i ai {Be ny ‘pot nie. ue :
ite Bis) yotegabg: andy ite me: fae ry re; Hs nae ae
pees) enene cial soiae head iia be giaers{, kee dihuzhe ‘late. ee tae
' Bah aun. Beni; Fe afin sf nt aan wat he wiley ; ac lad yA UF

i r we f 4 Hi
he ALTA) a ace we his iourtis sail ry by it wag! 1, rar .3

> mY deh) mis Seite Bhi Latae: r a, Lauer v9! vin os ? vib : oe
ne ay TUF SSS |) vedas: ue ee?! CG ey eet wor hahha Fuh: eg sia it : & me
“seal seas sant Baas eos eee is Riso IL 2) ais algal | hari

; lone Oye Ha) ry) has Ue A ti Mey abe soe | Ai daar sbi Py 4 Ry Vethuges a ; a
a ph tin nis Pee. | wis ah eeate ihe sani Disa Spiny Lyte Fey “wi eag
plane etiviny ois eolat anh ices 7 “tebiberih ey cone ne aM

LG ee ae wh Ne: eat —- aly + he oben! i Ob id Hoe di : dogs, ye a y:

Bp aA is babisauemany i thiths anna agheb 9-0 ee

nae armeione lye by. eared ab rT ati] any igs ay Le
ie d ia Shang. aa dtp 3 crt HA’ Leisy aah oe . ie:
Arn pig a host “A se ide evi oe Dipys ae
ai “bp ante adh semis tha fay: oth sii By .ar Fs |
Me yi inn ey ret eed 2 pine he oh sta 0 Pe
ve Bey POA a th ang HEO) a9 | ab 2 a
yg Nad a Net ; oT 4 onus as aieanciiis, t a Ns she natit itil ay
sane HY, Gly ip hes wt ang tanual’ Tomi hye irs lie: 78
is 3 ith gy: elatio” bah sl daly Se nse ae oh

0 ‘ . y, Be
<i a ry we ade cbt add my f | eli eal or ale a ae
h. “ j e : , a, Rg |
ee ‘ “a me a) wnat 4 Stok uh rd vd Py is : : 3 ; Y
AFL, ve ES aS 6c athe eM hn h bya Medes ie f
ye Ae eo eee ay © :
4 5 " + Pit ve. \
- \ be | 7 ; i:
Ne : H
£ ‘ oe mh ;
4 * anh us

Resumen por el autor, W. W. Swingle,
Princeton University.

Las células germinales de los Anuros.
I. El ciclo sexual del macho de Rana catesbiana.

Las células germinales aparecen primeramente bajo la forma
de una cresta mediana de células semejantes a las del endodermo,
situado encima del techo del arquenterio, en los embriones de
7mm. La cresta estd separada del endodermo subyacente por:
(1) La oclusién de las placas laterales y la formacién del mesen-
terio, y (2) Por la emigracién activa de las células germinales.
La cresta germinal se divide longitudinalmente, y las mitades
se separan para formar las crestas gonadiales pares de la larva.
En la larva de la primera estacién (de cuatro a seis meses de
edad) las gonadas son simplemente sacos huecos cuyas paredes
estdn formadas de una capa sencilla o doble de células sexuales.
A pesar del caradcter no diferenciado de las gonadas y de la falta
de madurez de las larvas, las células sexuales pasan por un
ciclo sexual muy precoz y abortivo, el cual termina con la degen-
eracién y reabsorcién de las células.

Ki] fenémeno de la maduracién es normal hasta la primera
division de maduracion, cuando la fragmentacién del centrosoma,
con formacién consiguiente de poliasters, tiene lugar acom-
pafiada de la destruccién de los cromosomas. Se forman unas
cudntas espermitidas gigantes mediante crecimiento de una
fibra axial que crece del centrosoma de los espermatocitos pri-
marios no divididos. Las células y los cromosomas se parecen
mucho mas a las de los Urodelos que a las células y cromosomas
de los anuros adultos. Unas cudntas espermatogonias, descen-
dientes lineares de las células derminales primordiales, persisten
sin cambiar durante el ciclo sexual abortivo y producen una
segunda generacién de células germinales en las larvas de dos
anos de edad. Muchas de estas células pasan por un segundo
ciclo de desarrollo y dan lugar a espermatozoides en el rena-
cuajo. Por consiguiente, en la larva de la rana toro existen dos
ciclos sexuales: El primero es muy precoz y abortivo, el segundo
esnormal. El] autor interpreta este fendmeno como una recapitu-
lacién en el ciclo de las células germinales de condiciones filo-
géneticas que han pasado en los anuros.

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHOR'S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 6

THE GERM CELLS OF ANURANS
I, THE MALE SEXUAL CYCLE OF RANA CATESBEIANA LARVAE

WILBUR WILLIS SWINGLE

Department of Biology, Princeton University

TWO TEXT FIGURES AND FIFTEEN PLATES (ONE HUNDRED AND THIRTY-ONE FIGURES)

CONTENTS
TNC LOCUCtTOnM EP al adel eS ae ae Sis Ue POE ORG See te tea te OO
Dini sionoteimexproblem<)\ yes: ae ee Se sec Nt aa eameat tee ape ae 237
Nate andamethOds. | 25 sii. Dace eet Hee © oes. ays agora mete ams ous aS erating «Sage «Nanay als 238
Resume of a few of the more important points in the developmental history
of the germ glands and germ cells of Rana catesbeiana larvae........... 241
Observations. Sexual cycle first-year larvae. Primary and secondary
EGGERS CMUNCDS aye MEA ect Phe RG enor woh singe eRe «Sede Chncaechart« oad as 246
Last spermatogonial division. First-year tadpoles................... 249
DiakinesisssHormationlofthemteurads sect eerie aes eae eit eee 256
The heterotypic mitosis of first-year larval spermatocytes............. 260
Fragmentation of the centrosome of first-year larval spermatocytes.... 261
Cytoplasmic and nuclear changes in the degenerating spermatocytes of
first=year larvaer tcc Ave scien tale Oe Meee ee a ere AHR ie cial entetas 263
Maturation cycle of second-year larvae and formation of functional
SWELIMACOZ OB ya psc. o sae hee eee se sliee Ore lei oievapelig ea: 6 Sa Bee Ce Susi spe! eyelets 265
Diseussion- Ol DSPrv ablOnsne. te hi 2 oe Me hE Ee sits Sas sae Beer 268
RANTING OSIS scouts ree RN ee TOE ae EAS Serer ech ed Haley sees 268
2. ihe polymorphic nucler of amphibians: 25.122 /.:0) oa 2 15 hes ae 269
Bo SWATH OSI He ra Men ee cain Sc oor: b p-ticta OMEN BoM OEE Ete DERE et or RePeRE Bene ae 272
cede chromoplasts 5 <'(? MMA yOSOMES 21s, 4 a. oesd dacleis Sa dtes le + Shee oleae 275
5. Significance of the maturation cycle in the larvae.................. 276
6. Is there a precocious sexual cycle in other anurans?................ 277
7. Significance of degeneration of sexual elements derived from primor-
Aaialeeenmycell sists pcs. 5 CAR ee eerie ie yo eara ale aps oa ahaa Fab cts 279
Somamany at @oneli sions is 4.5. ss Ree Ren 6 aise egsts coslanctuceioale aus ee ote can 291
A. The origin and fate of the primordial germ cells.................. 291
B. The chromosomes and larval sexual cycles...... shea: bagasse 234 293
TEL AT TUES OMS HEC eros. so. Sys vs Yo Z ctk ee a oc oe pare rakes eke eToys ie ais eae lee etal oa 296

235

236 WILBUR WILLIS SWINGLE ~

INTRODUCTION

Several years ago, while engaged in experimental work involv-
ing the germ glands and germ cells of anurans (Swingle, 17, also
17-18), the writer was somewhat hampered by lack of definite
criteria for differentiating the sexes in young larvae. In so far
as the cytological conditions presented by the germ cells were
concerned, it was impossible at that time, to distinguish clearly
male from female tadpoles. The literature concerning sex in
larval Anura was found to be voluminous and contained a great
variety of opinions, many of which were mutually exclusive,
others evidently based upon scanty evidence of somewhat dubi-
ous value, and none in any sense adequate to account for the
conditions presented by my material. In the summer of 1917,
therefore, an attempt was made to clear up the puzzling ques-
tion of sex differentiation, but the effort proved abortive owing
to lack of sufficient material. Certain cell stages occurred in my
larval material which had been a source of mystification to the
writer and to many others as well who had examined the mate-
rial; these stages had apparently never been observed or at any
rate reported by previous workers on anurans. Fortunately,
an opportunity soon presented itself of working with Prof. E. G.
Conklin, of Princeton University, who made a suggestion that
further investigation has since shown to be correct, i.e., that I
was dealing with a precocious maturation cycle in anuran lar-
vae. Professor Conklin’s suggestion throws an entirely new
light upon the question of sex differentiation and development
in the Anura, and brings the sexual conditions of these forms
more nearly into line with those described for other vertebrates.

It is a pleasure to acknowledge my indebtedness to Professor
Conklin for this illuminating suggestion, and for many others as
well, which have made this work possible, for the time he has
spent looking over material, and for the keen interest displayed
in the progress of the work.

To Prof. N. P. Sherwood and Dr. Cora Downs, of the Depart-
ment of Bacteriology of the University of Kansas, I am greatly
indebted for aid in collecting 2000 tadpole specimens from the
outlying districts of Douglas County, Kansas, during the summer

GERM CELLS OF ANURANS aat

of 1918. To Professors Allen and W. R. B. Robertson, of the
Department of Zoology, the University of Kansas, Iam also
indebted for aid in collecting tadpoles and newly metamorphosed
bullfrogs at various times.

DIVISION OF THE PROBLEM

The subject of sex in larval anurans: is such a complex one
and the literature on the question so vast, that no attempt will
be made to deal with all the aspects of the problem in this paper.
Instead, the material has been so arranged that different phases
will be taken up and discussed separately in a series of papers.
This paper is concerned chiefly with the more usual phases of the
sexual cycle of the male Rana catesbeiana, both in the larvae
and newly metamophosed animals, with especial reference to
chromosomal conditions. The broader questions of hermaphro-
ditism, alleged to exist normally as a developmental phase of
anurans, reversal of sexuality, anomalous sex ratios and their
experimental modification, Bidder’s organ, and other interesting
problems will not be touched upon here, save perhaps inciden-
tally, and then only in the briefest fashion. It will be recalled
that Pfliger reported years ago, that there occur normally in
newly metamorphosed frogs three kinds of individuals, males,
females, and hermaphrodites, the two latter forms much more
numerous in early stages than the males. In the course of fur-
ther development the hermaphrodites become either definitely
male or female, as the sex ratio for adult frogs is approximately
50-50. The investigations of R. Hertwig, Kuschakewitsch, and
Witschi not only confirmed Pfliiger’s work, but extended it by
showing that anurans apparently first develop solely as females
and sexual intermediates, the males only later differentiating
from the females and hermaphroditic forms. Moreover, these
investigators described in great detail modification of the sex
ratios by environmental changes, such as extremes of tempera-
ture and late fertilization. All of these alleged facts have given
rise to the belief that anurans in their sexual development differ
greatly from other vertebrates. These questions are reserved

238 WILBUR WILLIS SWINGLE

for a later paper, which will be a consideration of the develop-
mental history of the male and female sex glands, neoteny, Bid-
der’s organ, and an attempt at a reinterpretation of the problems
stated in the light of certain phenomena described below. The
writer regards the second part of this work as perhaps the most
interesting from a theoretical standpoint and as comprising the
main portion; however, for sake of clarity in presentation, divi-
sion of the subject has been found essential. It is necessary to
give in detail the normal germ-cell cycle before discussing its
aberrations or more unusual modifications.

MATERIAL AND METHODS

During the course of the work only one species of anuran has
been employed to any extent, i.e., Rana catesbeiana. Other
forms have been examined for comparison with the bullfrog, but
not for the phase of the problem treated in this paper, so they
need not concern us here. Rana catesbeiana in its larval stage
has no equal among other frogs in respect to the peculiar fitness
of its germ cells for this sort of study. The sex cells of the Uro-
dela have long been noted for their size and fitness for cytologi-
cal study, whereas the cells of adult frogs and toads have received
scant attention. Yet it is a fact that the germ cells of larval
bullfrogs, in regard to the size of cells and chromosomes, are
little surpassed by even the best urodele material, and in this
respect they more nearly resemble the caudate forms than the
conditions presented by adults of their own species. In the
adult frog or in newly metamorphosed animals the size of cells,
nuclei, and chromosomes is distinctly less than in the larvae.
The germ cells of sexually mature bullfrogs are in this respect
like those of a different animal group when compared with larval
stages. The explanation of this peculiarity will be discussed in
its proper place.

Another interesting feature about the bullfrog that makes it an
especially favorable object for study is its remarkable long larval
life. This species usually spends several seasons as a larva, and
is a tadpole for approximately two years. Sometimes these ani-
mals pass through almost three years as tadpoles, though this is

GERM CELLS OF ANURANS 239

a rare condition and probably a result of defective thyroid devel-
opment. The animals are abundant, are easily caught, and
readily adapt themselves to laboratory conditions. ‘Tadpoles
caught in the autumn need not be fed more than once a month
throughout the winter to keep them in good condition. First-
season tadpoles rarely attain a greater length than 35 to 40 mm.;
‘second-season specimens average 65 to 85 mm.; mature tadpoles,
100 to 154 mm. It is rare to find larvae with a greater length
than 145 mm., though the writer recently caught two male speci-
mens measuring 159 and 165 mm., respectively, from snout to
tip of the tail; both had ripe spermatozoa in the gonads.

It will be shown later in this paper that the long larval life of
Rana catesbeiana is correlated with a very interesting and sug-
gestive phase of the germ-cell cycle—a phase which, while nor-
mally occurring in other anurans and probably in many other
vertebrate forms, is brief, and apparently obscured by other
developmental phenomena, hence not so easy of interpretation
as the same condition in the bullfrog larva.

It should be stated here that there is apparently no seriation
of germ-cell stages anteroposteriorly in the testis of larval or
newly metamorphosed Rana catesbeiana such as has _ been
described for various urodeles. The testis of a 40 to 50-mm.
larvae is a narrow, flat, ribbon-like structure, gray-white in
color, somewhat convoluted, attached by a mesentery to the
inner edge of the ventral surface of the mesonephros. It bears
little resemblance to the testis of the adult and is longer than
the gonads of newly metamorphosed frogs. The relation of
the glands of first-year animals to those of second-year larvae
and newly metamorphosed frogs is indicated in text figure 1.
The internal structure of these gonads is indicated in photo-
graphs (33 to 35, explanation of figures), where it will be readily
seen that the center of the gonad consists of a large hollow (sec-
ondary ‘genital cavity) surrounded by a germinal epithelium con-
sisting of a single or double layer of germ cells in 40 to 50 mm.
tadpoles and of many layers of cells in 80 to 90 mm. animals. In
mature larvae and newly metamorphosed frogs the central cavity
of the testis is obliterated at definite intervals by migration of

240 WILBUR WILLIS SWINGLE

mesodermal cells from the mesentery and mesonephros (the so-
called sex cords, a misnomer for they are in reality the anlagen
of the rete or efferent apparatus).

Various fixatives have been employed, such as Flemming, Bou-
in’s, Ezra Allen’s (16) modification of Bouin’s fluid, and others.
The best results were obtained with the last two fluids. The
mesonephros was usually left attached to the testis. Sections
were cut at a thickness of 8 tol0u. Even at this thickness it

Fig. 1 a) Gonads of animal of first year. Average total length of larvae, 40
to 50 mm.; b) gonads of tadpoles 70 to 95 mm. total length; c) gonads of tadpoles
nearing metamorphosis; total length, 120 to 150 mm.

is necessary to reconstruct the nucleus in most cases because of
the large size of the cells. All spermatogonial counts, however,
were made from complete cells.

The larvae used for the present work were taken from various
localities and during different seasons of the year. Four hundred
larvae measuring from 70 to 110 mm. were taken from the ponds
of the State Fish Hatchery, Pratt, Kansas, during the month of
September, 1917; 300 larvae averaging 100 mm. were taken from
pools in the vicinity of Lawrence, Kansas, in the fall of 1916; a
group of 1500 larvae averaging 70 mm. was caught in August,

GERM’ CELLS OF ANURANS 241

1918, from a pool in Douglas County, Kansas; 1700 larvae meas-
uring from 60 to 165 mm. total length were taken from a pond on
the University Campus at Princeton during the months of July,
August, and October, 1919. Only a comparatively small num-
ber of animals from these various groups were examined micro-
scopically, the remainder were preserved for a study of the sex
ratios and so-called hermaphroditism at various developmental
stages—phases of the subject not dealt with here, but which make
up the subject-matter of a later communication.

The size or length of tadpoles is not a good criterion of their
‘age because of the size variability shown by anuran larvae of
similar age,reared under identical environmental conditions. The
writer was, until last year (1919), unable to get the eggs of the
bullfrog in sufficient quantity to rear the tadpoles artificially.
Hence the age of the older larvae given in this account is only
approximate, for they are classified according to size and stage of
development, as first- and second-year tadpoles.

RESUME OF A FEW OF THE MORE IMPORTANT POINTS IN THE
DEVELOPMENTAL HISTORY OF THE GERM GLANDS AND
GERM CELLS OF RANA CATESBEIANA LARVAE

A brief summary of the developmental history of the gonads
and sex cells may prove useful in elucidating some of the peculi-
arities of the sexual cycle described later in the paper. Only
a few of the more important stages will be considered here, and
then only in a very brief and sketchy way.

1. The primordial germ cells of the embryo are first distin-
guishable from other yolk-laden entoderm ceils as a ridge just
dorsal to the cavity of the archenteron and ventral to the
aorta, separating the two lateral plates of mesoderm (text fig. 2,
A). The medial growth of the two lateral plates and formation
of the mesentery together with probably an active migration
dorsally of the germ cells themselves, cuts off this germ-cell ridge
from the underlying entoderm (text fig. 2, Band C). As devel-
opment proceeds this median ridge of germ cells splits longitudi-
nally and the cells of the two halves then migrate laterally on
either side to form two independent ridges invested with a cov-

242 WILBUR WILLIS SWINGLE

ering of peritoneum. In cross-section each ridge is seen to be
made up of several large yolk-laden germ cells and a few small
deeply staining peritoneal cells.

2. The two germ ridges enlarge considerably by proliferation
of the cells and also by migration of mesoderm cells into the

Fig. 2 Origin of the germ-cells. A. Cross section through germ-cell region
of 7-mm. larvae. D, the aorta; HE, cardinal veins; F, wolffian duct; G, lateral
plate; H, entoderm cells. B. Transverse section through germ-cell region of
8-mm. larvae. J, germ-cells containing yolk; M, mesentery. C. Low magnifica-"
tion of stage shown in B. J, germ cells. .

ridges from the mesonephros and peritoneum. ‘The ridges pro-
ject into the body cavity and take on the character of germ glands.
The germ cells lose their yolk at about this time and divide
actively.

3. As development progresses, the glands grow rapidly, the
number of germ cells greatly increasing. Large cavities are

GERM CELLS OF ANURANS 243

formed in the gonads, the so-called secondary genital spaces, lined
by small non-sexual cells which have migrated into the gland from
the mesonephros by way of the mesentery. At this stage the
gonads of both sexes are hollow sacs (surrounded by peritoneum,
the so-called germinal epithelium), the walls of which consist of
one, two, or three layers of sex cells, depending upon the stage
studied (figs. 33 and 34).

In female larvae the cavity is later obliterated by growth of
the oocytes. In male animals, these secondary genital spaces
persist until shortly before metamorphosis, when they also are
obliterated, chiefly by increased division of the germ cells, and the
ingrowth of cells from the mesonephros, which form anastomos-
ing cords throughout the testis, the future rete or efferent ducts
(fig. 35 shows the obliteration of the testis cavity).

4. In young first-season tadpoles, the sexes are indistinguish-
able, though later males and females are easily separated by
microscopic examination. The female glands grow very fast and
greatly enlarge, owing to oocyte formation, becoming irregular. in
outline. On the other hand, the male gonads remain small, are
fairly regular in outline, but do not generally assume the shape
characteristic of the adult testis until some months previous to
metamorphosis, i.e., until the larvae are about two years of age
(fig. 35). Also font figure 1, C.

5. The germ cells of piriae: taken in summer of the stead
season, both male and female are found to be undergoing simul-
taneous maturation changes. ‘This is a most unusual phenome-
non, and so far as the writer is aware, unique among the verte-
brates, though common enough perhaps among the invertebrates.
In no other group of the Chordata has anything analogous tothe
simultaneous maturation changes of male and female germ cells
of larval anurans, such as here described, been reported, although
on certain theoretical grounds based on a study of the sexual
cycle of the larval bullfrog, the writer ventures to suggest that
analogous phenomena are likely to be found in the myxinoids,
larval petromyzonts, and eels.

The early maturation stages preceding the growth period of the
oocyte in female animals, such as leptotene, amphitene, pachy-

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

244 WILBUR WILLIS SWINGLE

tene, and diplotene, are said to occur normally extremely early in
most vertebrates, in some mammals before birth. In male indi-
viduals, according to the usual accounts in the literature, the
same stages of the maturation cycle do not usually take place
until shortly before the attainment of sexual maturity.

It has long been known that the germ cells of female frogs
undergo the earlier maturation changes while the animal is still
a tadpole. The growth period of the oocyte in this group of
amphibia is assumed to last a very long time, though further
work may show this not to be altogether true. According to
several investigators of European frogs, the eggs are not ready
for fertilization until the fourth or fifth season after metamor-
phosis, when the first polar body is extruded shortly before fer-
tilization. In this connection Gatenby (14) says of Rana
temporaria:

Though one cannot be certain, I believe that an odcyte takes two
years at least, and more probably three to become mature. It is
evident, therefore, that the young odcytes formed in April or May in
the adult will not be used for spawning next March, but certainly for a
spawning several years ahead. The first ova derived from primordial
germ cells would not be spawned till three years after the hatching of
the tadpole, since the frogs around Oxford seem to become mature in
three years.

However this may be, the germ cells of male Rana catesbeiana
larvae enter maturation simultaneously with those of female tad-
poles—long before metamorphosis, or before the gonads have
even differentiated sufficiently to resemble a testis. Until the
onset of maturation the gonads of the two sexes are morpho-
logically identical. Following the precocious maturation cycle,
it becomes easy to differentiate the sexes, as the female germ
cells soon enter the growth period and become oocytes. The
gross appearance of the glands of the two sexes changes at this
time. The male germ cells pass through all stages of maturation,
leptotene, amphitene (synapsis), pachytene, diplotene, tetrad
formation, up to the first maturation division in a perfectly
normal manner. During the anaphase of the heterotypic mitosis,
or at the earlier period of spindle formation, the spermatocytes

GERM CELLS OF ANURANS 245

undergo degeneration owing to fragmentation of the centrosome
and consequent formation of polyasters which lead to aberrant
divisions. Practically all of the first generation of male germ
cells, i.e., those derived from the primordial germ cells of the
entoderm ridge, pass through this abortive larval sexual cycle
and degenerate in the act of division. A few of these cells give
rise by direct transformation, without the intercalation of the
first or second maturation division, to gigantic spermatids with
axial filaments. Such spermatids possess fourteen tetrads. A
very few of the primordial germ cells fail to pass through the
precocious maturation cycle, and probably persist unchanged,
apparently giving rise later by repeated division to a second gen-
eration of germ cells in the male. It may be remarked here that
many cells of this second generation look as if they take origin
from germinal epithelium elements, i.e., appear to be transformed
mesothelial cells. This point, however, is still under investiga-
tion as morphological methods are not sufficient to determine
whether or not such transformations actually occur. The mode
of origin of the definitive germ cells of the adult is not strictly
germane to this particular paper, and the question must be left
undecided, pending results of experimental investigation.

This second generation of male sex cells, and this is the impor-
tant point here, no matter whether they be lineal descendants of
the primordial germ cells or transformed mesothelial elements,
undergo a second maturation (sexual) cycle in larvae just
ready for metamorphosis, i.e., in second-year tadpoles, and this
generation of cells, oddly enough, gives rise to normal sex prod-
ucts, spermatids, and spermatozoa. The maturation cycle is
normal in every respect. From the time of metamorphosis on to
sexual maturity the young male frog apparently ripens his sex
products continuously—this despite the fact that for a year or
so, owing to his small size contrasted with that of mature females,
he is probably unable to copulate.

It will be recalled that in the female sexual cycle the stage
corresponding to the first spermatocyte division of the male is the
stage of polar body formation which occurs normally at the time
of copulation, presumably several years after metamorphosis.

246 WILBUR WILLIS SWINGLE

Just why there should be this difference in time of maturation
between the male and female sexual cycles of the tadpoles the
writer is unable to say, though from certain data to be consid-
ered hereafter, obtained from studies on birds and mammals, it
would not be surprising if the young female frog some months
after metamorphosis likewise showed an abortive maturation
cycle culminating in degeneration of the oocytes.1_ This point is
now under investigation. Bearing in mind, then, this outline
sketch of the developmental history of the gonads and germ cells
of both sexes in the bullfrog tadpole, the following detailed
account of the cellular changes involved in the larval maturation
cycle of the male becomes more intelligible.'

OBSERVATIONS. SEXUAL CYCLE FIRST-YEAR LARVAE. PRIMARY
AND SECONDARY SPERMATOGONIA

The primary spermatogonia found in such gonads as shown in
figures 33 and 34 and text figures 1 A and B are much larger
‘than the later generation of cells to which they give rise. In
general these primary cells are more lightly staining. than other
elements of the gonad, and are peculiar, moreover, in that they
are usually surrounded by a follicle made up of small, flattened,
deeply staining stroma or peritoneal elements separating them
one from another. ‘This is true of this generation of cells in both
larval and adult frogs.

The primary spermatogonial nuclei are large and very irregu-
lar in outline, presenting marked lobulations and indentations—
the so-called polymorphism of the nucleus. Study of these poly-
morphic nuclei during early prophase stages of division has led to
the conclusion that the lobulations and consequent polymorphism
are due merely to large chromosomal vesicles or to the partial
fusion of such vesicles, for from each of these lobulations a
chromosome or pair of chromosomes appear in division pro-
phases. The resting nucleus contains considerable karyolymph,

1 Recently the writer has observed typical tetrad formation and a few first-
maturation spindles and chromosomes in oocytes_of female larvae. Such cells
degenerate in the act of division just as do the laval spermatocytes of the male
tadpole of the first year.

GERM CELLS OF ANURANS 247

and an irregular linin network upon which is scattered chromatin
granules of various size and shape, together with one or more
nucleoli. The nuclear size is in many instances enormous, com-
pletely filling the cytoplasm except for a narrow peripheral border
(iressolees bil rand slg):

The character of the attraction sphere and centrosome pre-
sents nothing unusual and conforms to the type described for
amphibians by earlier workers, hence it need not detain us here.

Division of the primary spermatogonia is always mitotic, and
_amitosis, though described for this type of cell in amphibians by
La Valette St. George (’85), Meves (’91), Benda (93), and
McGregor (’99), has not been observed in Rana catesbeiana.

The somatic or diploid number of chromosomes in the male
bullfrog larva is twenty-eight, and presumably this number is
characteristic of the adult also, though no counts have been made
on metamorphosed animals. <A few years ago the writer found
that twenty-six is the male diploid number for Rana pipiens,
the leopard frog (Swingle, ’17). Parmenter (’20) has recently
confirmed this count for parthenogenetic frogs of the same species.
According to King (’07), the somatic number in Bufo is twenty-
four. This last number has also generally been regarded as charac-
teristic for urodeles, such as Triton and Salamandra. Recently
Snook and Long (’14) described twenty-eight chromosomes in
the urodele Aneides lugubris, and Parmenter (’20) finds the same
number in the larva of Ambystoma tigrinum. Levy (’14~15)
states that the diploid number in male Rana temporaria is twenty-
five. The writer does not regard Levy’s evidence as above crit-
icism, and is much inclined to consider this statement as possibly
amistake. It would be odd if the males of all other amphibians,
both urodeles and anurans so far studied, possessed an even num-
ber of chromosomes, and one species, Rana temporaria, possessed
an odd number. Levy regards this species as having an accessory
chromosome.

The writer described an odd chromosomal body in the germ
cells of Rana pipiens as an accessory chromosome (Swingle, 717),
but has since been in doubt in regard to this matter. The body
described by me is probably a precociously dividing chromosome,

248 WILBUR WILLIS SWINGLE

one-half of which sometimes migrates toward the pole of the
spindle more quickly than does the other half to the oppo-
site pole. The figures of Levy indicate that the body described
by him as the sex chromosome is in all probability of the same
nature as the precociously dividing chromosome described by
myself. Further work on Rana temporaria will in all likelihood
bring it into line with other species in regard to chromosomal
constitution.

The twenty-eight somatic chromosomes of Rana catesbeiana
may be divided into four groups: 1) Large V- or J-shaped ele-
ments; 2) intermediate sized J’s; 3) small J’s and, 4) slightly
curved rods. ‘These chromosomes appear to be definitely paired
according to size and shape, and in this respect resemble those of
other amphibians. It should be stated, however, that the chro-
mosomes, though occurring in pairs in regard to size and shape
relations, are not always found side by side within the nucleus.
Many times the members of a pair are widely separated and may
be on opposite sides of the nucleus. In general, though, the two
homologues are usually near one another. Certainly, the inti-
mate pairing of somatic chromosomes, such as described by Metz
(14) for Drosophilia and by Whiting (717) for the mosquito, does
not occur in the Anura (figs. 4 to 6).

The size and shape of the spermatogonial chromosomes vary
somewhat with the fixative used, particularly if the fixation is not
of the very best. The size variation is due to the preserving
fluid and not to any real variation of chromosomal size or shape
in the living tissue. In extreme cases the chromosomes may
appear as short blocks, and their characteristic shape is entirely
lacking (fig. 5). It is interesting in this connection to compare
King’s (’07) figure 10, plate 1, with my figure 5. King regards
the chromosomes figured by her as those by young spermato-
cytes before the stage of reduction (p. 368). They look very
much like the short dumpy chromatin blocks of my figure 5.
This cell is an ordinary spermatogonium in prophase, in which
the spireme segments have either been greatly condensed by
imperfect fixation or else the cell was abnormal, probably the
latter is the case, as such cells appear in otherwise excellently

GERM CELLS OF ANURANS 249

fixed material. The chromatin masses are readily counted and
are of the diploid number. This type of cell is unusual in the
larvae, and has never been observed in metamorphosed frogs.

In metaphase the apices of the J-chromosomes are oriented
toward the center of the spindle, and spindle fiber attachment is
non-terminal.

The spermatogonial chromosomes are occasionally split into
two elements twisted about each other as apparently is the usual
condition in Ambystoma (Parmenter, ’20).

Variations in the chromosome number have been observed in
but two cases: once in a spermatocyte which contained eighteen
tetrads, possibly the result of fusion of two adjacent cells, and once
in a spermatogonium containing thirty-six or more chromosomes
(fig. 19).

It is doubtful if variation of chromosomal number occurs in
normal cells within one and the same individual, save perhaps in
those cases where a single chromosome may occasionally undergo
fragmentation. Even in such cases there is apparently no real
variation in quantity of chromatin mass. Such chromosomal
fragmentation as is described by recent writers, notably Hance
(718), has not been observed in the bullfrog except in degenerating
first spermatocytes where the multipolar spindles literally tear
the chromosomes to pieces (fig. 115). -

The multiplication of secondary spermatogonia in the larval
gonad, and this is especially true of first-year tadpoles, does not
continue long enough to obliterate the lumen of the gland or to
crowd the cells together owing to greatly increased numbers.
During the second season the spermatogonial divisions come to
a close and maturation begins in the type of gonad shown in
figures 33 and 34, also text figure 1, A.

Last spermatogonial division. First-year tadpoles

The telophases of the last larval spermatogonial divisions
differ in no respect from other similar stages in the mitosis of the
primary and secondary spermatogonia. ‘The period of nuclear
reconstruction, however, presents marked structural changes dif-
ferentiating it from all previous stages, in that the nucleus enters

250 WILBUR WILLIS SWINGLE

the so-called ‘resting’ period, preparatory to undergoing the com-
plex phenomena of maturation. As stated before, the nuclei of
the primary and secondary spermatogonia soon become poly-
morphic in character, following division, and the chromatin mate- -
rial is found scattered throughout the nucleus in bead-like masses
or granules, attached to an irregular linin network. In sharp
contrast to this type of nuclear reconstruction, nuclei of the last
spermatogonial telophase are round or oval in shape and of small
size. The chromatin is in the form of small lumps or blocks
(Janssens, ’03) somewhat irregular in outline. In especially fav-
orable cells the number of these blocks can be made out with a
fair degree of accuracy. Their number is certainly diploid. The
writer is inclined to regard these chromatin blocks as representing
individual chromosomes at this stage. In early stages they are
independent of one another, but very soon anastomosing linin
fibrils appear between them.

The preleptotene period (Grégoire, ’07) marks the first indica-
tions of resolution of the blocks. They become woolly or mossy
in appearance, delicate, much coiled, and tangled thread-like
processes appear, as if spinning out from the chromatin material
in the mass. During the course of these changes the nucleus
increases in size (figs. 8 and 47). These tangled threads so
characteristic of the preleptotene elongate, lose their spiral-like
character, and extend across the nucleus in loops. At this period
there is no definite orientation of the leptotene filaments. Appar-
ently, for the writer cannot speak with certainty on this point,
each of the chromatin blocks of the telophase nuclei gives origin
to a single thread. It is difficult, if not impossible, to unravel
the snarl of elongating threads crowding the nucleus at the time
of their first appearance. Wenrich (’16) has been able to trace
the origin of the leptotene threads with considerable clearness in
Phrynotettix magnus, and he is of the opinion that a single
chromatin block gives rise to a single filament. This view seems
very probable, when consideration is taken of the fact that the
number of leptotene filaments is diploid and corresponds closely,
if not exactly, to the number of blocks. Judging from Wenrich’s
figures, conditions.in Phrynotettix at this period are much more
favorable for study than in Rana catesbeiana tadpoles.

GERM CELLS OF ANURANS 251

Shortly after their formation the leptotene threads tend to show
a definite orientation of their free ends toward the centrosome
and sphere in many cells, giving the appearance of a series of
delicate loops. Janssens (’05) has characterized this orientation
as the bouquet gréle or leptotene bouquet. The chromatin por-
tion of the looped threads is in the form of very minute particles
distributed at more or less regular intervals along a central linin
core or fibril. Usually at this stage one or more nucleoli are pres-
ent, though they differ from ordinary nucleoli in being in intimate
connection with the chromosomes. ‘These bodies have been
termed chromoplasts by Eisen (’00), who first studied them in
Batracoseps. We shall have more to say about these bodies
‘ later (figs. 10 and 55).

Following the period of the leptotene bouquet, there occurs in
amphibians an extremely important and interesting stage, first
observed by Janssens (’05) and named by him amphitene. This
stage marks the first formation of the thickened pachytene thread
and corresponds to the zygotene of Grégoire’s (’07) terminology.
In Rana catesbeiana larvae the amphitene constitutes a very defi-
nite and well-marked period in the maturation process—one that is
easily differentiated from the leptotene preceding or the pachy-
tene following. Judging by descriptions of various investigators
of the zygotene in different animal groups, the amphitene of the
amphibian germ-cell cycle is a prolonged transition stage between
leptotene and pachytene. In typical amphitene nuclei one finds
the nucleus marked off into two more or less distinct portions by
the type of chromatin thread present. At the proximal pole of
the nucleus, i.e., that side nearest the centrosome and sphere,
the delicate leptotene filaments have disappeared, and one finds
only the thickened pachytene threads; conversely, at the distal
pole of the nucleus, i.e., the pole opposite the sphere, the leptotene
condition persists. By focusing through a single cell, it is possible
to bring into view now a leptotene, now a pachytene condition
(figs. 11 to 13, 36 to 38). The explanation of this apparently
anomalous condition is simply that the thick pachytene threads
of the proximal pole of the amphitene nucleus represent the longi-
tudinal fusion (parasynapsis) of two originally distinct leptotene

252, WILBUR WILLIS SWINGLE

filaments. The side-by-side fusion or synapsis begins at the
ends of the threads nearest the centrosome, and extends distally
until fusion is complete throughout the length of the conjugants.
Thus the amphitene is essentially a transition period in which
the pairing of chromosomes in the stage of leptotene filaments is
progressing. In the distal portion of the nucleus, where typical
leptotene conditions persist, parasynapsis has not yet occurred.

The evidence for this point of view is quite conclusive in Rana
catesbeiana larvae: a) The leptotene threads are certainly nearer
the somatic number than the haploid number; b) the thickened
pachytene loops represent the haploid or reduced number; c) the
thickness of the pachytene elements is just twice that of a single
leptotene filament, d) and, perhaps most conclusive, it is not dif-
ficult in studying amphitene nuclei, to trace the two unpaired ends
of the leptotene threads from the distal pole into a single thick-
ened pachytene thread at the proximal pole (figs. 11 and 12).
Janssens (’05) has figured this stage clearly in his figures 20, 21,
22, and 23, plate IV. Wilson (’12) observed the same thing in
Batracoseps material obtained from Janssens, and states that
the conditions described are even clearer than Janssens figured
them. Apparently analogous conditions are figured by Wenrich
(16) (fig. 77) and designated by him as zygotene stages showing
incomplete conjugation of chromosomes.

The leptotene threads appear to coil or twist about each other
corkserew fashion so tightly that all trace of their double nature
is lost and the resulting thickened thread appears single. (Fig. 12.)

Many investigators of amphibian spermatogenesis have de-
scribed other methods of synapsis for this group of vertebrates.
However, the period assigned, which has usually been considered
as identical with synapsis in urodeles and anurans, is in all prob-
ability an artifact due to imperfect fixation of material, poor stain-
ing, or both. This so-called synaptic period corresponds to what
McClung has termed synizesis or the ‘‘unilateral or central con-
traction of the chromatin in the nucleus during the prophase of
the first spermatocyte.’ Nuclear conditions are at this time
extremely difficult to make out, to say nothing of interpreting
correctly. The pachytene spireme is considered as evolving out
of this contracted nuclear condition (King, ’07, figs. 24 and 25).

GERM CELLS OF ANURANS 253

Janssens (’01) first called attention to this condition in urodeles
and considered it a definite stage of the germ-cell cycle. Later
(05) he reversed his earlier opinion and stated the condition
described earlier was due to poor fixation.

In the bullfrog larva there can be no question that synizesis is
an artifact due to poor penetration of fixatives. For instance, in
well-preserved material it is impossible to find contraction stages;
where large pieces of the gland are used, generally the peripheral
portion of the tissue will show no synizesis, whereas the central
portion will show numerous contraction figures. A comparative
study of reagents, such as Bouin’s or, better, Ezra Allen’s modi-
fication of Bouin’s fluid without urea which is a good fixative for
frog material, with Flemming’s osmic fixative, a rather poor pene-
trant, on similar sized pieces of gonad, gives illuminating results
in regard to contraction stages. In Rana catesbeiana and Rana
pipiens slow penetration of fixatives clumps the delicate loops
of the leptotene bouquet into a typical synizesis figure.

The condition described here for anurans possibly is not com-
parable to a somewhat similar clumping of nuclear contents in
other forms described by various investigators. The writer has
had the opportunity, through the courtesy of Dr. E. L. Shaffer,
to examine synizesis stages in Cicada material. The conditions
presented by this form are hardly comparable to those described
here for anurans, and it may well be that in certain groups syni-
zesis is a definite stage in the maturation cycle.

The partial synapsis of leptotene threads in the amphitene is
completed in the pachytene stage which immediately follows.
The threads of this period are thickened throughout uniformly
and usually show no trace of their dual nature, save perhaps in
respect to size. It is odd that in a fully formed pachytene spi-
reme there is usually no indication of the leptotene threads which
entered into its formation (fig. 39, also 11). Most animals show
distinct traces of a primary longitudinal split or line of fusion
between the conjugants. For example, Wenrich (’16), describing
the pachytene stage in Phrynotettix magnus, states: ‘‘The line
of separation between the threads which have conjugated (i.e
the primary longitudinal split) remains visible throughout the
pacytene stage.”

254 WILBUR WILLIS SWINGLE

It is generally only in the amphitene and diplotene that the
line of fusion of the conjugants is visible in the bullfrog larvae.
In this connection it is interesting to note that Janssens (’05),
in his study of Batracoseps, was unable to detect any indication
of a paired condition. Wilson (’12), in his examination of the
same animal, agrees with Janssens that the pachytene threads
appear as if single.

The pachytene period in anurans larvae is in all respects like
that described for urodeles. In many cases the free ends of the
thickened threads are applied close to the nuclear membrane at the
proximal pole, the broad loops extending distally, thus giving rise
to the pachytene bouquet. Janssens (’05) divided the pachytene
in Batracoseps into two distinct periods: the ‘bouquet orienté,’
corresponding to the condition just described, and the ‘bouquet
transverse,’ in which the nuclear contents have apparently rotated
in relation to the sphere, so that the bouquet instead of being
oriented toward the centrosome and sphere is turned at right
angles to it. The writer is unable to say definitely whether the
period of the transverse bouquet does or does not represent a
well-marked stage in the maturation cycle of Rana catesbeiana.
Very probably this stage is more marked in urodeles than in
anura.

In Bufo, King (’07) derives the pachytene spireme from the
irregular, deeply staining, confused chromatin mass of the synize-
sis period. Her figure 25, plate 1, is a clear expression of her idea
regarding the derivation of the pachytene threads. According to
her account, it is a continuous spireme, does not show any evi-
dence of longitudinal splitting, and later in the course of develop-
ment segments transversely into the reduced (haploid) number of
chromosomes. If this account of conditions in Bufo is correct,
then this anuran differs from other amphibians, both caudate and
tailess, in respect to formation of the pachytene spireme and the
tetrads. The writer is under the impression that the difference
between Bufo and other forms rests upon a misinterpretation of
synizesis and synapsis, and if reexamined Bufo will very likely
be found to conform to the amphibian type of maturation cycle.

GERM CELLS OF ANURANS 255

Following the pachytene is the period of exconjugation or dis-
junction of the homologous chromosomes, i.e., the longitudinal
splitting of the thick double threads into two thin threads which
diverge in the center, but remain united at both ends (figs. 14,
15, and 40 to 51). This stage corresponds to the diplotene of
Winiwarter (’00) or the prostrepsinema of Janssen (’05). The
pachytene threads split longitudinally, the split first appearing
apparently at the distal pole of the nucleus and extending prox-
imally. The line of cleavage might possibly be looked upon as
marking the earlier line of fusion of the two originally unpaired
leptotene threads, and hence be regarded as the line of disjunc-
tion (figs. 14 and 15, also 40 to 51). This is only guesswork,
however, because in general the fused leptotene threads show no
sign of separation in the pachytene as they do in other forms;
i.e., the primary longitudinal split is usually invisible at this
stage.

The diplotene stage in Rana catesbeiana larvae is marked by
extreme growth of the cell, especially the nucleus which reaches
gigantic proportions in many instances. In general the cells of
the pachytene stage, though larger than those of the leptotene,
do not present such marked size differences over leptotene stages
as do the diplotene nuclei over both pachytene and leptotene.

In early diplotene, when the primary longitudinal split is just
making its appearance, there is somewhat superficial resemblance
to the amphitene stage. The similarity is, however, slight, and
one could hardly confuse the two periods. The longitudinal split
of the diplotene appears first at what corresponds to the distal
pole of the double thread. In the amphitene just the reverse
condition is presented, the initial pairing of the leptotene filaments
begins first at the proximal pole. The very obvious difference
in the size of the nuclei of the two periods is an excellent criterion
for distinguishing the two stages. Also separation is never com-
plete in the diplotene, as the homologues remain united at their
ends; conversely, in the amphitene the unpaired leptotene threads
at the distal pole of the nucleus diverge widely from one another.

Shortly after the process of disjunction, a secondary longitudi-
nal splitting of each member of the pair appears, forming the

256 WILBUR WILLIS SWINGLE

tetrad-complex, made up of four chromatids (McClung, ’00)
united at the ends. This secondary split marks the line of
separation of the chromatids in the subsequent equational or
homotypic division, i.e., the second maturation division. In
each chromosome this second split is apparently at right angles
to the primary split. Coincident with the appearance of the
secondary split is a process of shortening and thickening of the
diffuse, thread-like tetrads. This shortening and condensation of
the chromosomes marks the end of the diplotene (figs. 16 and 17).

Following the stage just described, there occurs a series of
transition stages leading up to the complete formation of the
heterotypic tetrads on the mitotic figure. These stages are known
by various names, but for present purposes Hicker’s (95) term
‘diakinesis’ will be employed as including that period in the germ-
cell cycle, from the first formation of the tetrads to their definitive
arrangement upon the first maturation spindle.

Diakinesis—formation of the tetrads

The ring tetrads, so characteristic of the Amphibia, are formed
by the disjunction of the homologous chromosomes that paired
during the amphitene and pachytene and separated during the
diplotene, except at their ends which remained in contact (figs.
14, 16, 41 to 51). Thus, in the writer’s opinion, the annular
space represents, in Rana catesbeiana, the space between homolo-
gous chromosomes. In other words the space between the rings
represents the ‘primary longitudinal split’ and probably the orig-
inal line of fusion in parasynapsis of the autosome pairs. The
first maturation division in the bullfrog larva is heterotypic or
reductional for most of the tetrads in the sense that entire chromo-
somes are separated. This conception has been held by various
workers on urodele spermatogenesis. Thus Janssens holds this
view for the urodele Batracoseps and the anuran Alytes and
Montgomery for Plethodon cinereus and Desmognathus -fuscus,
though the latter writer arrived at this conclusion by assuming
telosynapsis occurs first. He interpreted the pachytene loops
correctly as bivalent chromosomes, but he misinterpreted the
nature of the double spireme, in considering each loop as two

GERM CELLS OF ANURANS 257

univalent chromosomes united telosynaptically at the angle of the
loop. According to this point of view, the space between the two
arms of the loop is the space between two univalent chromo-
somes, but does not represent the line of fusion of originally
separated leptotene threads. This view, while erroneous (admit-
ted to be so by Montgomery himself, ’12), leads to essentially the
same end results as those stated by the writer.

There are fourteen typical rings and crosses plainly of tetrad
nature in the spermatocytes of the bullfrog larvae and in some
cells a large rod-shaped body may appear (fig. 26 to 28). The
rings are of large size as compared with similar chromosome
stages of adult frog material and are practically identical with
those of urodeles in regard to size and shape. In the larvae
these rings can be grouped according to their size relations—one
very large ring (fig. 18), five intermediate in size (fig. 20), and
eight smaller ones (fig. 21). The size relations of these rings in
various cells is apparently constant for the species in cells of the
same size, and this is an important point, for there is a variation
of chromosomal size in cells of different size. The amount of
volume of cytoplasm has much to do with the size of the chromo-
somes. Figures 29, 30,31, and 32 bring out clearly this difference
in chromosome size when two cells in identical stages but of dif-
ferent size are compared. The thinness of the chromosome
group in figure 78 is not entirely due to stretching on the spindle.
The chromatin mass varies in proportion to cellular size, i.e., the
larger the cell the larger the chromosomes, and vice versa. Conk-
lin (12) has clearly shown and discussed this point also in numer-
ous other papers. The extremely large cells shown in plates 7
to 14 have relatively large tetrads, conversely the smaller cells
figured in plate 15 have much smaller tetrads. The photographs
of plates 7 to 14 are of larval cells, those of plate 15 of the germ
cells of animals at the time of metamorphosis.

Ring tetrads in amphibia have been described by several previ-
ous writers, so only a brief discussion is needed here. Following
the separation save at the synaptic ends in the diplotene of the
paired elements, and the appearance of the secondary split, the
rings open in two planes at right angles to each other: 1) In the

258 WILBUR WILLIS SWINGLE

center probably along the original line of fusion of the homo-
logues; 2) along the line of the secondary split after the fashion
described by Robertson (’14) and Wenrich (’16) for the Orthop-
tera. Condensation of the chromatin begins at this stage. In
early stages of ring formation the tetrads stain rather lightly
and are somewhat (fig. 17) diffuse, but as condensation proceeds
they readily take up the basic dyes. ‘The tetrad character is
obvious from a study of the synaptic ends of the homologues.
The larger rings in middle prophase stages are generally in the
form of figure 8’s, and this character may be maintained up to the
metaphase. The smaller tetrads early assume the character of
the rings; other shapes, such as crosses and y’s, and in some cells
a rod, appear. There is usually a single Y and a single cross-
shaped tetrad in every spermatocyte, though these may appear
much like small rings. Other shapes that appear are transitory
stages in ring formation, or else portions of rings viewed from
various angles. Large crosses, for instance, sometimes appear
in early prophase, and are generally true rings viewed ‘en face,’
the arms of the cross being the long synaptic ends of the paired
chromosomes. Such crosses are not comparable to cr to be con-
fused with true cross-shaped tetrads (figs. 19).

In sections slightly overstained, the smaller rings appear solid,
the synaptic ends being represented merely by rounded knobs.
At times such rings may even appear like dumbbells, and this is
notably true of the second-year spermatocytes, i.e., those that
give rise to true spermatozoa at the second ripening of the germ
cells of the larvae (figs. 120 to 128). Indeed, it is not improbable
that those investigators of anuran spermatogenesis who have
described prophase tetrads as solid and of dumbbell shape were
perhaps dealing with either overstained or imperfectly fixed
material.

Spindle-fiber attachment is non-terminal usually, but may occur
anywhere in spermatocytes of the first-year larvae showing cen-
trosomal fragmentation. There are no normal spermatocytes in
the first maturation cycle, so any discussion of spindle-fiber
attachment is useless (figs. 29, 31, and 32).

GERM CELLS OF ANURANS 259

During late stages of diakinesis the cells in many cases become
greatly enlarged and in many instances are of giant proportions
(figs. 64, 67, 112). The increase in volume may affect either nu-
cleus or cytoplasm or both. It is rather common in my material
to find over one-half or two-thirds of a nucleus in a single section
because of the size, and it may be added that my material was sec-
tioned at a thickness of 8 to 10 u. These large spermatocytes
of the larvae resemble those of urodeles more than adult anurans.
It is an interesting and suggestive fact that near the period of
metamorphosis the elongated, more or less ribbon-like testis be-
comes transformed into a very small typically shaped frog tes-
tis. The shortening process may require a considerable time,
though the writer is inclined to doubt this on account of the
absence of transition stages. The shortening progresses from
posterior to anterior and may amount to as much as 1 mm.
Figure 35 is a section through gland from a newly metamorphosed
animal; figure 34 a gland of the second season (before meta-
morphosis) ; figure 33 a section of a gland of a first season larva.

Examination of the small, fully formed testes (full formed
except for the efferent or rete apparatus) of recently metamor-
phosed animals reveals some interesting size differences of the cells
compared with those of the gonads of young larvae of the first
season. The cellular elements of the small gonads are more nearly
like those of the adult, and it is rare to find the giant spermatocytes
of the type figured in plates 7 to 13. The primary spermatogonia
are of about equal size with those of the younger tadpoles, the
chief differences are in the spermatocytes and diplotene and
pachytene nuclei. These small testes first appear in tadpoles
measuring 120 mm. or more from snout to tip of tail. Such ani-
mals are about a year and a half old or perhaps somewhat younger
and are due to metamorphose the following summer. The hind
legs are on the average about 25 to 30 mm.; the fore limbs are not
visible. Now, oddly enough, some of the male animals are
mature, in so far as the possession of ripe spermatozoa is con-
cerned. And, as we shall shortly see, this character marks off
this type of gland from those of the first-season larvae. As was
stated before, the bullfrog tadpole passes approximately two

260 WILBUR WILLIS SWINGLE

years as a larva, and each year is marked by a seasonal ripening
of sexual products. Attempts of first-year tadpoles to ripen
their sex products is abortive; the second year’s attempt is suc-
cessful (plate 15) at the time of metamorphosis.

The heterotypic mitosis of first-year larval spermatocytes

First maturation division metaphases are very abundant in
first-year larvae, especially in young larvae 45 to 60 mm. total
length which have passed through one winter as larvae. In such
animals entire cysts are found with completely formed tetrads
and spindles all ready for division, and many in the act of divid-
ing; yet, oddly enough, careful examination of many hundreds of
sections of such larval gonads fails to show stages of the first
maturation mitosis beyond very early anaphases (plates 9, 10,
11). This is an interesting fact, and it is strange to see entire
cysts of apparently normal spermatocytes in metaphase or early
anaphase, and yet never find telophases of such divisions, inter-
kinesis stages, or any indications of secondary spermatocytes.
Cells which have developed thus far in a perfectly normal man-
ner, save for precocity of the maturation cycle and are apparently
in possession of the requisite mechanism for cell division, are
unable to complete the process. During late prophase and meta-
phase the achromatic elements, that is, the machinery of cell divi-
sion and chromosomal separation, break down and the tetrads go
to pieces before the telophase. It is rare to find complete
separation of the homologous components of the tetrads. Some-
times the smaller ring elements do separate (figs. 29 and 30) and
in rare cases an early anaphase is reached. In general the degen-
erative processes set in shortly after the time of spindle formation
when the chromosomes are arranged in a typical metaphase plate,
their long axis parallel to the long axis of the spindle, or else
when they are scattered irregularly through the cytoplasm, fol-
lowing the disappearance of the nuclear wall (figs. 100 to 110).
The spermatocytes may even go to pieces in late diakinesis,
though such cases are not common.

GERM CELLS OF ANURANS 261

The obvious cause of degeneration of the spermatocytes of first-
season larvae is to be found in the abnormal behavior of the
centrosomes. Very early in the work it was observed that multi-
polar mitotic figures were exceedingly frequent, and, indeed,
these came to be regarded as the rule rather than the exception
in the maturation cycle of first-year tadpoles. Triasters and
tetrasters with striking and bizarre chromosomal arrangement
proved so common that attention was focused upon the centro-
some as the primary seat of degenerative processes. It may be
added here that these aberrant polyasters are very favorable
objects of study in Rana catesbeiana larvae because of their size
and number and should prove of interest to anyone concerned
with cellular mechanics.

Fragmentation of the centrosome of first-year larval spermatocytes

A study of the centrosome of the first-season spermatocytes
proved very fruitful in several respects: 1) it gave the clue to
correct interpretation of the anomalous behavior, i.e., the failure
to divide of the spermatocytes; 2) the results of such study
explained on sound mechanical grounds the presence of the poly-
asters; 3) it led to the discovery of certain giant spermatid-like
structures.
~ In most, if not all, of the first-season larval spermatocytes,
the centrosome behaves abnormally, rarely does it pass through
the normal cycle and give rise to a typical bipolar spindle. The
usual thing is fragmentation of one or both halves of the divided
centrosome; figures 18, 21, 79 show such fragmentation. There
may be a central granule, surrounded by four or five others, all
connected to one another by very delicate filaments. Each of
these granules may or may not form a tiny aster in the cytoplasm.
In other cells a typical spindle may be formed at one pole with
numerous smaller spindles at the other pole. There are several
variations of this type (figs. 18, 21, 79).

Perhaps the most peculiar condition noted in the centrosomal
behavior of the spermatocytes was the tendency to form axial
filaments or tails. It required much searching to find anything

262 WILBUR WILLIS SWINGLE

like ‘tailed cells,’ and in the writer’s experience they occur rather
infrequently, although on theoretical grounds one would expect
to find them numerous. In ‘tailed cells’ the centrosome, instead
of fragmenting or forming a multipolar spindle, sends out a long,
somewhat spiral filament that grows outside the cell like the axial
fiber of aspermatid. These filaments are extremely delicate struc-
tures and difficult to make out. The stage at which these axial
fibers grow out from the centrosome may vary somewhat. In figure
23 the filament had evidently formed during late diakinesis, as
there is a nuclear wall present. Figure 22 shows the fiber extend-
ing out from the periphery of the nucleus—an unusual condition.
Some of these figures correspond to, in fact are practically identi-
cal with, Broman’s (’00) drawings of giant spermatids in adult
Bombinator ingenus material. Compare my figures 22 and 23
with his figures.

The type of cell represented in these figures is very abundant
in larval material, especially following the period of greatest abun-
dance of aberrant spermatocyte divisions. Not all such cells show
axial filaments, indeed, they are rare. Such cells with filaments
growing from the centrosome may be regarded as giant spermatids
resulting directly from transformed first spermatocytes which have
not undergone either first or second maturation division. Compar-
ison of the stages figured in plate 13 brings out this point clearly.
The same type of cell but without axial filaments is quite abun-
dant; these originate in the same manner as described above, but
cannot be spoken of as spermatids in the absence of the axial
fibers. Broman has observed several filaments growing out from
a single cell in his adult toad material; so far such cases have not
appeared in my material, and it has been a source of some wonder
on my part why such cells are not of greater frequency. Cellu-
lar conditions in the first-season larvae are ideal for the devel-
opment of such structures in abundance. The relative infre-
quency of the tailed cell may perhaps be correlated with the
fragmentation of the centrosome.

The giant spermatids are non-functional and usually undergo .
no further metamorphosis, but degenerate and are resorbed.
Stages in the process are shown in figures 103 to 110. In very

GERM CELLS OF ANURANS 263

rare cases these abnormal spermatid-like bodies apparently give
rise by condensation of the nuclear material and elongation of
the cytoplasm to structures bearing a faint resemblance to the
apyrene spermatozoa of certain prosobranchs.

The degeneration of the first-season spermatocytes at the meta-
phase is somewhat analogous to the degeneration of that type of
ova that requires the stimulation of a spermatozoon to enable it
to complete the developmental cycle. Mead (’98) and Conklin
(05) both observed ova with perfectly formed spindle and chro-
mosomes go to pieces at this stage without further development
unless fertilized. In the case of the larval bullfrog spermato- ~
cytes, fragmentation of the centrosome is the immediate cause of
the failure of the cells to divide and of the resulting degeneration.
Professor Conklin has suggested that perhaps the non-fertilized
ova observed by him also go to pieces because of centrosomal
fragmentation. It is known that the entering spermatozoan
brings in a centrosome which takes part in the cleavage process.
In regard to the abortive maturation cycle of the first-year bull-
frog larvae it is possible that it is a vestige of an early repro-
ductive cycle inherited from remote ancestors. Centrosomal
fragmentation is merely the more obvious morphological cause
of the degeneration of the spermatocytes and itself a symptom
of a deeper-seated derangement of cellular life.

Cytoplasmic and nuclear changes in the degenerating spermatocytes
of first-year larvae

The nuclear changes are more or les‘characteristic of degenerat-
ing cells in general, including those just described as ultimately
destined to form the spermatid-like bodies. The initial stage in
degeneration apparently first affects the centrosome, the chromo-
somes and cytoplasm are later attacked. To take a typical
example of degeneration in a spermatocyte (omitting those poly-
asters where the chromosomes are pulled to pieces), there is first
fragmentation of the centrosome and formation of polyasters,
followed by shortening and thickening of the tetrads accom-
panied by increased staining capacity. The chromosomes soon
lose the ring-tetrad structure, and the annular space entirely dis-

264 WILBUR WILLIS SWINGLE

appears as though there had occurred a running together of the
chromosomes. ‘The lugs or knobs marking the synaptic ends of
the chromosomes round off and disappear, leaving an oval-shaped
shiny mass, resembling a heavily stained oil drop. Such masses
tend to run together, forming larger units, until in final stages
of the process the original fourteen tetrads are represented by
three or four large deeply staining spherical masses (fig. 22 and
23). During this time the nuclear wall may or may not have
disappeared, depending upon the age of the spermatocyte when
degeneration began. Where this process sets in after spindle
formation there is no nuclear wall. In cases where the nuclear
membrane is present, the entire nucleus is excentric in position
(fig. 106). In those cases where degeneration begins after the
complete formation of the mitotic figures, and after the tetrads
are arranged on it, the history of the degeneration is slightly dif-
ferent from that just described. ‘The essential difference is that
there is no nuclear wall present, the spindle apparatus is resorbed
into the surrounding cytoplasm, and the tetrads go to pieces in
situ. The latter take on the appearance of oily masses which may
or may not fuse together. In final stages, all that remains of
the fourteen tetrads and mitotic figure is a group of oily vacuoles
grouped together much in the same fashion as the chromosomes
were grouped on the spindle. In many instances the number,
and even the size relations of the vacuoles corresponds to the first-
maturation chromosomes (figs. 109 and 110).

The cytoplasmic changes accompanying these regressive
nuclear phenomena are interesting. The normal, clear, lightly
staining protoplasm becomes yellow in color, much vacuolated,
and numerous spherical droplets of yolk-like substance appear.
In this connection it is interesting to note that an essentially
similar yolk-like material has been described and figured by prac-
tically all workers on apyrene spermatozoa. For instance,
Gould’s (18) description and figures for Crepidula plana and
Reinke’s (712) figures of Strombus. However, as the substance
in question occurs in degenerate and functionless cells in all
cases, the presence of similar degeneration by-products is to be
expected and has no special significance. The writer doubts if
this substance is yolk, though it does resemble it.

GERM CELLS OF ANURANS 265

Maturation cycle of second-year larvae and formation of functional
spermatozoa

Much that should more properly have been discussed in this
section has been referred to here and there earlier in this paper
in order to clear up certain sources of confusion which might arise.

The second-year larval sexual cycle differs from that of the
first year in two ways: 1) The germ cells of the second maturation
cycle are considerably smaller, the tetrads are consequently much
smaller than those of the first maturation cycle, and have less the
appearance of rings than of dumbbell-shaped bodies when at-
tached to the first maturation spindle; 2) mature spermatozoa are
produced, there is little cell degeneration, but few polyasters
occur, and hence few cases of fragmentation of the centrosome.
All maturation divisions are normal. It is obvious that there is
a vast difference between the first and second maturation cycles
of the male larvae; the first is aberrant, the second normal; one
culminates in degeneration, the other in the production of func-
tional male sex cells.

The smaller size of the germ cells of second-year larvae is not
difficult to explain. During the period of the first sexual ripening
practically all of the germ cells in the gonads are affected, and
consequently destined to degenerate and disappear. ‘There are,
however, a few primary spermatogonia with polymorphic nuclei,
lineal descendants of the primordial germ cells, scattered here
and there through the gonad, which fail to undergo the preco-
cious maturation cycle. These cells are generally, though not
always, found near the sex cord region.- In the interval be-
tween the first and second larval sexual cycles these cells appar-
ently divide rapidly and spread through the gonads. It is
probably the repeated division of these cells, and their prog-
eny that brings about the marked reduction in cell size, so
noticeable at the second maturation cycle. The proliferation of
germ cells is so extraordinarily rapid in the gonads of tadpoles
just about to metamorphose that the cellular size becomes re-
duced to a size scarcely larger than that characteristic of the
larger stroma or peritoneal cells. . Indeed, conditions are such in

266 WILBUR WILLIS SWINGLE

the gonads at this time, and especially in certain individuals,
that a great many of the definitive sex cells appear to arise by an
actual transformation of mesothelial elements into germ cells.
This question is still under investigation, for it is exceedingly
difficult to determine definitely whether this is or is not the case
from morphological data alone. Certainly in my material there
is very suggestive morphological evidence that such transforma-
tions may possibly occur, but whether such transformations actu-
ally do occur is an entirely different thing.

Another factor to be considered in regard to the reduction in
cell size of the first-year germ cells is the fact that the entire gonad
undergoes a striking diminution in size during the second year of
growth, taking on the character of the adult testis. There is a
great loss of water from the tissues at the time of metamorphosis
and consequent shrinkage of the cells of the animal in volume.

The size of the chromosomes depends upon the volume of the
surrounding cytoplasm and of the nucleus, hence the smaller size
of the second maturation cycle tetrads. These tetrads are of the
short dumpy type found normally in adult frogs and toads (figs.
120 to 128).

The interesting fact that the sex products of the first larval sex-
ual ripening are all abortive, while those of the second larval cycle
are normal is something of a puzzle, and the only explanation
occurring to the writer is based upon the phylogenetic history of
Anura and will be discussed later in this paper along with some
data of a somewhat similar nature regarding mammals and birds.

There is one type of cell in the gonads of second-year larvae
and metamorphosed frogs that may remain about equal in size to
the germ cells of the younger larvae—the primary spermatogonia.
This type of cell is large in frogs and larvae of any age.

During the month of August, 1919, several very large tadpoles
were captured with a total length of 140 to 160 mm. Examina-
tion of the testes of male individuals showed many normal sper-
matocyte divisions, spermatids in all stages of development, and
a few mature spermatozoa. At this time the efferent ducts of
the testis were not yet fully developed. The gonads were ex-
tremely small and immature looking. Female gonads of larvae
of similar size showed only oécytes undergoing growth.

GERM CELLS OF ANURANS 267

The age of these animals could only be estimated by their size
and developmental stage, and were probably about two years old.
Some of the tadpoles would have undergone metamorphosis within
a short time but for the lateness of the season. The writer has
known for several years that young male bullfrogs shortly after
metamorphosis are sexually mature, though in regard to size they
are pygmies compared to the adults and it is difficult to see how
they could possibly copulate with mature females. The female
Rana catesbeiana apparently does not become sexually mature
and ready for copulation until several years after metamorphosis,
and in this respect resembles the European frogs, such as Rana
esculenta. According to the observations of R. Hertwig, Witschi,
and others, the female of Rana temporaria and Rana esculenta
does not become fully mature and ready for copulation until the
fifth season after metamorphosis. Despite the fact that the
females are sexually immature, the young male bullfrog apparently
ripens his sexual products continuously, beginning with the first
year of larval life, though the first-year sexual cycle is abortive.

The writer has observed somewhat similar phenomena in the
leopard frog, Rana pipiens. It is not uncommon to find very
small immature looking individuals of this species with spérm in
their testes. It has been known for several years that prolonga-
tion of the larval life of tadpoles of this species by thyroid extirpa-
tion does not prevent the normal seasonal ripening of their sex
products. Such ripening corresponds to the second larval sexual
cycle in Rana catesbeiana, for it is probable that the tadpoles
of this species (i.e., Rana pipiens) undergo a very precocious and
abbreviated maturation cycle very early in larval life. This
early cycle in Rana pipiens would correspond to the first sexual
cycle of Rana catesbeiana; the prolonging of the larval life of
the leopard-frog tadpoles leads to the second seasonal ripening
and production, of normal spermatozoa just as in the bullfrog,
though in the latter species the period of larval existence covers
the second seasonal ripening of the germ cells..

The germ-cell cycle of the second year in Rana catesbeiana
larvae, as was previously stated, is normal in every way, hence no
description will be given of the process here. The secondary

268 WILBUR WILLIS SWINGLE

spermatocytes give rise to normal spermatids and some sperma-
tozoa, and the writer has nothing to add to this phase of the sub-
ject that has not been described many times before in papers
concerned with the spermatogenesis of other amphibians.
Unquestionaly, these larval sperm cells are functional, because
morphologically they are indistinguishable from spermatozoa of
adult frogs (figs. 118 to 131).

DISCUSSION OF OBSERVATIONS
1. Amitosis in anurans

It has frequently been stated by earlier investigators working
with amphibian material that amitosis occurs quite commonly in
the testis cells of urodelesand anurans. Several writers have even
asserted that at certain seasons of the year amitosis is the sole
method of division (La Valete St. George, Meves, Benda and
McGregor). It has even been seriously stated that the primary
spermatogonia not only divide amitotically, but the results of
such direct division become functional spermatozoa. Meves and
Benda state that amitosis occurs by means of the constrictive
force of a ring-shaped centrosome in Salamandra. McGregor
states that the nucleus is divided by a cleft into two approxi-
mately equal parts. Oddly enough, in view of these positive
statements, the writer has never observed anything in primary
or secondary spermatogonia or in follicle or stroma cells that is
comparable in any way to amitosis. The polymorphic nuclei of
the spermatogonia do somewhat superficially seem to be con-
structed into two halves at times, and the constriction may be
deep enough to give (fig. 1) the appearance of separate nuclei
in the same cell. Careful study reveals connecting portions lying
at deeper levels. It is evident from descriptions of amitosis in
amphibians that polymorphism of the nucleus has been mistaken
for direct division. ‘The writer takes the position, perhaps ex-
treme, that in the spermatogenesis of anura, amitosis does not
normally occur, and if it ever occurs in these forms, it is an
extremely rare and aberrant condition, save in senescent cells
such as those of Bidder’s organ, and even in this degenerate struc-
ture direct division is uncommon.

GERM CELLS OF ANURANS 269

2. The polymorphic nuclei of amphibians

Practically all investigators of the germ cells of urodeles and
anurans have described and figured the bizarre and striking
lobulation of the spermatogonial nuclei of these forms, but few
have attempted any explanation of the peculiarity. Most of
the earlier investigators regarded the nuclear polymorphism as
stages in amitotic division. The writer has observed, however,
that the striking nuclear lobulations of the spermatogonial nuclei
in Rana catesbeiana larvae, are nothing more or less than chromo-
somal vesicles and fusions of such vesicles. In reality, the poly-
morphism of nuclear structure is due to these vesicles, remaining
more or less distinct and independent from one another. That
this view is essentially correct is readily seen by examination of
early prophase and late telophase spermatogonial divisions. In
early prophases the individual chromosomes arise by condensa-
tion of the chromatin material of the vesicles. Later telophases
clearly show the formation of the vesicles which increase greatly
in size as the cell grows.

Conklin (’02) has called attention to the occurrence of such
chromosomal vesicles in Crepidula. He states in this regard:

1. The chromosomes, consisting of chromatin enclosed in a linin
sheath, divide and move to the poles of the spindle, where they par-
tially surround the spheres. 2. Here they become vesicular, the
interior of the vesicle becoming achromatic, though frequently contain-
ing a nucleolus-like body, while the wall remains chromatic. 3. These
vesicles continue to enlarge and then unite into the ‘resting’ nucleus.
The nuclear membrane is composed of the outermost walls of the
vesicles, while the inner walls stretch through the nucleus as achromatic
partitions.

Again on page 47 of his (’02) communication, Conklin writes:

It sometimes happens, especially in eggs in which more than the
normal number of centrosomes and asters are present, that some or all
of the chromosomal vesicles do not fuse, but remain distinct through
the whole of the resting period. In such cases each of the vesicles
behaves like a miniature nucleus, absorbing achromatic material and
forming a net-work of chromatin either within the vesicle or on its
walls. In this growth and differentiation the vesicles keep pace, step
by step with the normal nucleus, so that one must regard the resting
nucleus as virtually composed of vesicles, though their union may be
so intimate as to hide this structure.

270 WILBUR WILLIS SWINGLE

The resting nucleus is not, therefore, a single structure any more
than is the equatorial plate. It is composed of units, each of which,
so far as known, has the properties of the entire nucleus, and the union
of these vesicles into a single one may be considered as a secondary
character. It is altogether probable that the chromosomes, and hence
the chromosomal vesicles preserve their identity throughout the resting
period, and I venture the suggestion that the daughter chromosomes
will be found to arise within the chromosomal vesicles.

The description just quoted, of the formation and behavior of
chromosomal vesicles in gasteropod molluscs, applies equally as
well to the conditions in the bullfrog larva, and certainly cannot
be better stated than Professor Conklin’s description (figs. 1, 2,
oy.

Hicker (’95) reported that the chromosomes of the early cleay-
ages of Cyclops brevicornis formed two groups of vesicles, one
group from the paternal, the other from the maternal pronuclei.

More recently, Wenrich (’16) has reported that each chroma-
some in Phrynotettix becomes surrounded, as early as the ana-
phase, by a hyaline region; that this region expands in the
telophase; that the chromatin of each chromosome becomes
diffused within its own region; that a membrane becomes formed
at the boundary between the hyaline region and the cytoplasm,
producing the chromosomal vesicle. ‘The nuclear membrane
consists of the outer walls of the vesicles at the periphery of the
nuclear group. Wenrich concludes that the hyaline region is
formed at the expense of the cytoplasm and that the material
of each chromosome tends to remain within the space of its own
vesicle, a core of chromatin being particularly noticeable in the
center of this region, and that the prophase chromosome subse-
quently formed, is developed out of the substance of one, and
only one, of the previously existing telophase chromosomes.

Conditions in Rana catesbeiana larvae, while not so clearly
marked, in regard to individual chromosome vesicles, as those
described for Orthopteran material, nevertheless strongly indi-
cate that the ‘polymorphic’ nucleus of amphibians is nothing
other than a group of large chromosomal vesicles, more or less
independent, the outer walls of the outermost vesicles forming the
nuclear membrane.

GERM CELLS OF ANURANS Det

Some of these vesicles are of exceedingly large size in the prim-
ary spermatogonia, and represent possibly, two or more individual
vesicles pressed so tightly together as to appear as a single vesicle.
At times the individuality of the larger number of vesicles is
obscured or may disappear altogether, not to appear again until
the early prophase of the succeeding division when the chromo-
somes reform within the vesicles. Figures 1 and 2 give a goodidea
of the enormous size attained by these vesicles in certain cells.
The combined size of the vesicles is so great that the nucleus com-
pletely fills the cytoplasm except for a narrow peripheral border.

These chromosomal vesicles are the means by which the indi-
viduality of the chromosomes is maintained from cell generation
to cell generation. During the so-called ‘resting’ stages of the
cell, when the chromosomes appear to have lost their identity,
and merged with the other elements of the nucleoplasm, they are
in reality diffused within little sacs or vesicles, and probably
thus remain entirely separated from one another throughout this
diffuse period. No one nowadays seriously maintains that
chromosomes maintain a strict morphological identity, i.e., ap-
pearance, throughout all stages of cell life, but that they do main-
tain a genetic continuity or individuality throughout ‘resting’
stages of cellular life by means of chromosomal vesicles cannot be
seriously questioned.

The chromosomes that arise from chromosomal vesicles during
prophase stages of mitosis are the same chromosomes that went
into them during the telophase of the preceding division. The
writer can speak only for the conditions presented by the frog,
but the accounts in the literature indicate that this statement
probably holds true for a great many forms, possibly all. The fact
that such chromosomal vesicles have not been found in certain
groups, as for instance, coelenterates, according to G. T. Hargitt
(20), is no sure indication they do not exist in that group.

272 if WILBUR WILLIS SWINGLE

3. Synapsis

This question needs little discussion here, considering the beau-
tiful work of Janssens (’05), Janssens and Willems (’08), the
Schreiners (’06), and Snook and Long (14) on various amphib-
ian types. The conditions in Rana catesbeiana larvae are essen-
tially similar to those described by these works for the urodeles.

Some of the earliest observations on the problems of synapsis
were made on amphibian material, chiefly urodeles, and an end to
end conjugation of chromosomes or telosnapsis was generally
conceded to occur. Janssens (’05), working with Batracoseps,
demonstrated parasynapsis, and later, working in collaboration
with Willems, showed this to be true of the anuran Alytes. Wil-
son (712) after a study of Janssen’s material, agrees with this
author regarding parasynapsis in Batracoseps. Montgomery
(11) who previously (’03) had described telosynapsis in urodeles,
reversed his earlier opinion and states (p. 753): ‘‘ During the past
year I have also convinced myself of the occurrence of parasyn-
desis in Plethodon, such as Janssens had described for this object
and the Schreiners for Salamandra.” Snook and Long (14)
describe the same kind of evidence for parasynapsis in Aneides
lugubris as that presented by Janssens and the Schreiners.

It is interesting to note that King (07) denies the existence of
parasynapsis for Bufo lentiginosus, and states that telosynapsis
is the method of chromosome pairing in this form. King regards
the period of synapsis in Bufo as occurring coincidently with
synizesis—a condition now generally regarded as an artifact, at
any rate in Amphibia. The tetrad chromosomes of the first
maturation mitosis, King thinks, arise by transverse segmentation
of the thick spireme. If this description of conditions in the toad
is correct, then this form differs markedly from other amphibians.
Recently, however, the writer has had an opportunity of examin-
ing preparations of Bufo, and is convinced that parasynapsis
occurs in this animal as the normal method of chromosome conju-
gation. Amphitene stages are abundant in the material exam-
ined by me, and this is the true period of synapsis in anurans
(figs: 11 12513536 to 38).

GERM CELLS OF ANURANS 273

The twisting together of the leptotene threads to form the
double pachytene spireme which occurs during the amphitene in
anura (fig. 12) seems to the writer to be the period when the
mechanical conditions for the ‘chiasma-type’ theory of Janssens
are present, and not during the later stage figured by this writer.
This theory of ‘Chiasma-type’ has been extensively employed by
Professor Morgan and his co-workers to explain ‘crossing-over’ in
Drosophila. It has repeatedly been observed that genetic fac-
tors belonging to a certain group, and presumably carried by a
single chromosome, go into a mating together, but do not always
reappear together, as they should, if carried by a single chromo-
some that has maintained its individuality throughout. Janssens
endeavored to explain the anomalous genetic behavior of such fac-
tors on mechanical grounds, i.e., by showing that in the behavior
of the chromosomes, at certain stages in the maturation cycle,
it is possible for actual ‘crossing-over’ of parts of homologous
chromosomes to occur, and'this exchange of parts of chromosomes
he termed ‘Chiasmatype.’ This theory is based upon a study of
certain postspireme (strepsinema) stages in the spermatogenesis
of the urodele Batracoseps. In this form, after the secondary
longitudinal split (equational split) has taken place, the tetrads
are composed of four separate strands or chromatids. These
strands may cross each other at certain places, and, owing to
strains or weakness at the point of contact, break, subsequently
recombining in such a manner as to form threads composed of
parts of both original strands. ‘That is to say, parts of the two
strands ‘crossed over’ and became incorporated as a portion of the
opposite chromatid. Janssens has carefully figured many such
apparent ‘cross-overs’ in the postspireme stages of Batracoseps.

There can be no reasonable doubt of the accuracy of the genetic
evidence for ‘crossing-over,’ nor of the general truth involved in
the chiasma-type theory. The point to be considered here is
whether or not the cytological evidence for this view is not more
convincing if a stage in the maturation cycle of the chromosomes
is used as the basis of cross-over, earlier than the early tetrad
state employed by Janssens. The mechanical conditions fur-
nished by the amphitene period in the bullfrog, for crossing over of

274 WILBUR WILLIS SWINGLE

parts of homologous chromosomes is well-nigh perfect, certainly
more so than the later stages. In the first place, the side-by-side
pairing of the two leptotene threads is accomplished by a process
of twisting together, and not merely by a side-by-side union; sec-
ondly, the twisting of the homologous threads is so tight that all
trace of their double nature is lost and the two elements appear
as one. It is only later at the time of separation during the
diplotene that the dual character of the pachytene thread again
becomes apparent. It would be odd if during the process of
separation of the tightly twisted threads ‘crossing-over’ did not
sometimes occur (figs. 11, 12, 13, 36 to 38).

Aside, however, from the ideal conditions presented by the
amphitene stage for exchange of parts of chromosomes, the
period figured by Janssens as furnishing actual cytological evi- .
dence of such exchange is not entirely satisfactory. The chief
objection here is that study of the postspireme chromosomes of
the maturation cycle in the larval bullfrog fails to support the
contention that breaking and recombination of the chromatids
occur at this particular period. ‘The strepsinema stages of the
bullfrog tadpole are very similar to those figured by Janssens,
but the writer is of the opinion that crossing-over does not occur
here; at any rate, no good evidence for it has been observed.
The tetrads are of the non-cross-over type like those figured by
Wenrich (716) and Robertson (714) for grasshoppers.

Strepsinema stages of tetrad formation are perfectly definite
and characteristic periods in the maturation cycle, and so far as
my;material is concerned, the chromatids appear to preserve their
identity through this period (figs. 16, 40 to 63).

There is certain genetical evidence indicative of chromosomal
‘crossing-over’ during the early synaptic stages of the odcytes of
Drosophila, such, for instance, as Plough’s (’17) experiments.
He found that environmental changes such as low or high tem-
perature markedly increased the percentage of ‘cross-overs’ in
the second chromosome of Drosophila melanogaster (ampelo-
phila). The temperature apparently increased the amount of
‘crossing-over’ at a definite stage of oogenesis, and Plough’s evi-
dence suggested strongly that the chromosomal exchange takes

GERM CELLS OF ANURANS 275

place at the stage when the chromosomes of Drosophila are
known to be finely drawn-out threads. In other words, he local-
izes the period of ‘crossing-over’ in the stage of oogenesis when
twisting together of the homologous threads is possible.

It matters little, in so far as the validity of the genetical evi-
dence is concerned, at exactly what stage in the germ-cell cycle
‘crossing-over’ may take place, for that such a process does occur
can scarcely be denied in view of the mass of positive evidence.
It is not impossible that the phenomenon may take place at sev-
eral different stages.

It is an odd fact that ‘crossing-over’ of genetic factors appar-
ently does not occur in the male Drosophila, but is confined
solely to the female. Oddly enough, the chiasmatype theory
invoked to explain it is based upon conditions observed in male
Amphibia. So far no one has advanced a satisfactory explana-
tion to account for the apparent absence of this phenomenon in
the male Drosophila. Nabours has reported evidence for ‘cross-
ing-over’ in the males of grouse locusts, Castle for the male rat,
hence it is evident that it is not confined solely to females.

4. The chromoplasts: (2?) karyosomes

Regarding the true nature of these bodies and their relation
to the chromosomes, the writer is in doubt, Janssens (’05)
who has made a-careful and detailed study of the origin and fate
of these structures in Batracoseps states:

Que le chromoplaste prend naissance aux derniers télophases sperma-
togoniales et qu’il résulte d’un empatement df an dépdt d’une sub-
stance sidérophile entre les pointes des V chromosomes aux poles de la
figure.

Qu’ & mesure que le chromosome se remplit de substance sidérophile
le chromoplaste diminue de volume. II] est done naturel de la con-
sidérer comme une substance destinée 4 étre absorbée par les chromo-
somes & la fin du stade auxocytaire comme il semble qu’elle 4 étre
excrétée par eux au commencement de ce stade.

This view is an interesting one; however, the writer has not
paid sufficient attention to the chromoplasts and nucleoli in
Rana catesbeiana to make any statement regarding the origin
and fate of these structures.

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

276 WILBUR WILLIS SWINGLE

The view of Eisen (’00) that the definitive chromosomes of
the spermatocyte are derived from the chromoplasts and that
‘‘chromoplasts guide the formation of the chromosome just as the
archosomes guide the formation of the spindles,” does not seem to
be entirely substantiated by conditions in the bullfrog tadpoles.
In this form the chromoplasts appear to have little to do with the
origin of the definite chromosomes in so far as the chromatin
material is concerned, for this originates from the preexisting
chromatin blocks of the last spermatogonial telophases by a spin-
ning-out process of the leptotene threads. But it is very likely,
and my own observations bear this out, that the chromoplasts
do give up substance to the chromosomes, though just what the
nature of this substance is the writer is unable tosay. Itis doubt-
ful if the chromoplasts are composed of true basi-chromatin.

In early stages of chromosome formation, such as the prelepto-
tene and leptotene, the chromoplasts are usually large heavily
staining bodies, to which are attached several chromatin threads.
As development of the threads proceed, the chromoplasts become
smaller and take the stain with less avidity. In still later stages
they become vacuolated as if being drained of their contents by
the growing threads. In final stages these bodies disappear.

5. Significance of the maturation cycle in the larvae

It is possible that the precocious, seasonal ripening of the
male germ cells of larval bullfrog represents a recapitulation in
ontogeny of a primitive, phylogenetic sexual cycle of ancestral
forms, when the Anura were sexually mature and reproduced
as larvae, much in the same fashion as does the axolotl to-day.
Few biologists would hesitate nowadays to deny that the latter
is not merely a neotenous, gigantic, sexually mature larva of the
urodele Amblystoma tigrinum, in view of the work of Chauvin
(75) and Duméril (65). The question why this animal some-
times fails to undergo metamorphosis in certain districts does not
concern us here (Swingle, 719). Besides the axolotl there are
numerous other instances on record of neotenous, sexually ma-
ture amphibians that have failed to metamorphose at the proper
time. So far as the writer is aware, such individuals are confined

GERM CELLS OF ANURANS ied

to the urodeles, with the exception of the bullfrog larvae described
here. In other anurans permanent retention of larval characters
may be experimentally produced by prolonging the larval life
by thyroid extirpation; the retention of the larval somatic charac-
ters has no effect upon the germ cells. Similar results were
obtained by the writer (Swingle, ’17~18), where it was shown
that acceleration of metamorphosis by thyroid feeding does not
accelerate the normal course of events in the germ-cell cycle.

It appears possible that the precocious seasonal ripening of the
male tadpoles germ cells is a recapitulation, just as the tadpole soma
is possibly a recapitulation of an earlier phylogenetic stage when
the present-day Anura were more like the Urodela than they are at
present, both in regard to body form and sexual conditions. It
would be interesting to know whether or not larval urodeles show
any such precocious sexuality as described here for anurans. It
is not improbable that other vertebrates with larval periods of
development, such as some of the eels and petromyzonts, will be
found to present analogous conditions to those described for
the bullfrog larva. In fact, judging by certain facts to be pre-
sented hereafter, it seems likely that all the vertebrates present
some such precocity of the germ-cell cycle as described here. If
the phenomena described here are in any way rooted in past
phylogenetic conditions, it is a much more remote past than
anything represented by any living Urodele type.

6. Is there a precocious sexual cycle in other anurans?

This question must be answered at once in the affirmative.

In Rana catesbeiana the larval period is the longest of any
other anuran known, and, as a consequence, the precocious sexual
cycle of the tadpole is carried farther than in other frogs. There
is no question but that if other anurans presented conditions in
their germ-cell cycle as marked and unmistakable as those in the
bullfrog, such conditions would have been reported years ago.
One could not easily overlook cysts of spermatocytes in which the
tetrads are of sufficient size to permit counting with a one-sixth
Leitz objective and a no. 5 ocular. Although conditions in other
frog species are not so plain and easy of interpretation as those
presented by the bullfrog larvae, yet nevertheless such species

278 WILBUR WILLIS SWINGLE

apparently show essentially identical phenomena as described
by myself. The trouble heretofore has been one of interpre-
tation. In Rana temporaria and Rana esculenta tadpoles the
same precocity of the sexual cycle, as is presented by the bull- |
frog has been described many times by various investigators
of these European frogs, but has been interpreted in a manner
entirely different from the explanation here given. The fig-
ures and descriptions of Bouin, Kuschakewitsch, Witschi, Schmidt-
Marcel, and others on the history and-development of the germ
cells of these frogs plainly indicate the precocious maturation
process in the larvae. However, the ripening of the tadpole
germ cells of the species studied goes only up to and including
the pachytene, according to their figures. These writers proba-
bly misinterpreted the sexual conditions, and this has led to some
bizarre theories of sex differentiation in frog larvae. According
to the interpretation of this school, all larvae whose germ cells
presented auxospireme, i.e., leptotene and pachvtene maturation
stages, are to be regarded as females, because it is only female
animals that show such maturation changes in larval or embry-
onic life. Similar maturation stages of male sex cells do not
occur until near the period of sexual maturity, according to them.
In the anurans studied by these writers the precocity of the sex-
ual cycle is very marked, and the germ cells do not go beyond
the pachytene and form tetrads and first-maturation spindles as
normally occurs in Rana catesbeiana. Consequently they were
not aware that the male larvae exhibits a precocious maturation
cycle coincident with that of the females, when the cells of the
latter go through the early stages of odcyte formation. As a_
consequence of this developmental peculiarity, i.e., curtailment of
the maturation cycle to the early stages of the process, without
exception these writers, being unable to differentiate male from
female, concluded that all frog tadpoles first develop as females,
then later half of the female tadpoles must transform into males,
because the sex ratio of adult frogs is approximately 50-50.
Their conclusions were logical enough, even though probably
erroneous, considering their premise that early maturation stages,
leptotene and pachytene, are solely characteristic of female
animals, when found in immature or embryonic forms. From a

GERM CELLS OF ANURANS 279

study of the serm cells of Rana pipiens the writer arrived at
essentially the same conclusions, and only after a study of the
germ-cell cycle of the bullfrog was it possible to unravel the puzzle.
The writer does not believe that females transform into males or
vice versa, nor that tadpoles develop solely as females during early
stages. A correct interpretation is possible only by comparing
the bullfrog tadpole with other forms. In the germ-cell cycle of
the larvae of most species of anurans, for example, forms like Rana
pipiens with short periods of larval life, the precocity of the mat-
uration cycle is apparently very marked and evanescent, hence it
is more obscure and difficult to interpret than that of the bullfrog.
But perhaps the most remarkable example of precocity of the
larval germ-cell cycle is presented by Bufo. In the toad the
early maturation phenomena of the germ cells, 1.e., leptotene;
amphitene, and pachytene stages, appear in extremely young
larvae, about two weeks after hatching and are confined chiefly
to the anterior end of the male and female gonads, i.e., that por-
tion which develops into the organ of Bidder. The ripening pro-
cess does not go beyond the pachytene stage apparently before the
cells become senescent. ‘The question whether or not Bidder’s
organ is or is not a rudimentary hermaphrodite gland is reserved
for discussion in another paper. A very brief and curtailed
sketch of the writer’s view of this structure in the toad and the
so-called hermaphroditism of the Anura will be found in the
American Naturalist for July-August, 1920.

There appears to be some sort of correlation in anurans
between length of larval life and stage in development to which
the ripening of the germ cells is carried before degeneration.
The whole problem of sex development and sex differentiation of
anuran larvae needs reinvestigation in the light of these obser-
vations on the bullfrog tadpole.

7. Significance of degeneration of sexual elements derived from
primordial germ cells

It is with considerable reluctance that the writer touches upon
this particular phase of the problem in this paper, which is to be
regarded as a cytological introduction to a more extensive report

280 ; WILBUR WILLIS SWINGLE

later upon the entire developmental history of the male and
female sex glands and cells in anurans. The problem stated in
the heading of this section is of such importance as to deserve a
much more detailed discussion than is possible here. However,
a brief statement of the more important theoretical considera-
tions suggested by the results obtained in the study of the larval
sexual cycle of the bullfrog cannot well be avoided. There are
certain obscure and little-known phenomena occurring in several
other classes of vertebrates of a similar, if not identical nature
with those reported here for the Anura. When the germ-cell
cycle of some of the higher vertebrates is correlated with the
maturation cycle of the larval frog, there is much that suggests
+o the writer that possibly we are here dealing with a fundamental
principle in germ-cell development, of widespread, perhaps of
universal occurrence among the vertebrates. Let us examine
some of this evidence.

There are two important theories concerning the origin of the
germ cells of vertebrates, each backed by considerable amounts
of evidence not lightly to be disregarded: 1) The first view is
that the definitive sex cells of the gonads are derived from pri-
mordial germ cells which have originated elsewhere in the organ-
ism, probably from entoderm, and have migrated into the genital
ridges and there differentiated into odcytes or spermatocytes
as the case may be. These primordial germ cells of the embryo
are distinct from other surrounding mesothelial cells and have a
- separate origin from cells of this type. The advocates of this
view are many, and a great deal of valuable data in support of
this theory has been collected of late years. 2) The second view
regards the germ cells as differentiated products of the germinal
epithelium, i.e., that they arise by direct transformation of sexu-
ally indifferent cells of mesodermal origin.

The advocates of the first point of view, we may, for the sake
of convenience, term the ‘entodermists,’ though not all believers
in the Keimbahn theory are agreed that the germ cells take origin
from entoderm; the adherents of the second theory we shall call
‘mesodermists.’ Between these two groups there is little or no
common ground, but instead a great deal of controversy.

GERM CELLS OF ANURANS 281

The first view can be traced back to Nussbaum (’80) who
claimed that germ cells are not derived from the soma, but are
early differentiated segmentation products which take no part in
body formation and retain their primitive embryonic character.
Many facts supporting this view have come to light through
study of the origin of the germ cells in almost all classes of verte-
brates. There can be no question that in most vertebrates the
sex cells are early set aside during development, and later
migrate into the germ ridges. In the embryo frog this process is
readily traced through every stage; the same is true of other forms,
also, as, for instance, certain ganoids, reptiles, and birds.

The second or mesodermal view, viz., that the germ cells arise
from an original sexually indifferent germinal epithelium origi-
nated with Waldeyer (’70) and has been held by a great many
observers to be the true method of germ-cell origin. Oddly
enough, practically every animal claimed by the ‘entodermists’
as illustrating their view has also been claimed by the ‘mesoder-
mists’ as illustrating their theory. For instance, take the Am-
phibia: Allen, King, and Witschi hold the view that the primordial
germ cells arise from the entoderm and give rise to the definitive
sex cells; on the other hand, Semon, Bouin, Dustin, Kuschake-
witsch, Champy, and Gatenby are all equally certain that the
definitive sex cells of the amphibia arise from the germinal epi-
thelium. These views are diametrically opposed, consequently
both cannot be either entirely true or entirely false, and the
problem is to point out if possible the source of confusion. To
take the case of the frog again, in this form the primordial sex
cells unquestionably arise from the entoderm. ‘There is abso-
lutely no indications of germ cells arising from the germinal epi-
thelium in early larval stages. The two types of cell in the gonad,
mesothelial and sexual, are entirely distinct and it would be
difficult to confuse them. All increase in germ cell number is
by mitotic division of preéxisting, differentiated primordial germ
cells. This state of affairs persists until the tadpole is about 45
to 60 mm. total length. Thus far it is obvious that the evidence
derived from the frog is decidedly in favor of the ‘entodermists.’
However, practically all of the germ cells derived from the ento-

282 WILBUR WILLIS SWINGLE

derm in the male bullfrog larva undergo a precocious and abor-
tive maturation cycle ending in degeneration and absorption. <A
very few descendants of the primordial line of cells fail to maturate
or degenerate, and apparently give rise to a new generation of
sex cells in the tadpole. Whether or not this second generation of
germ cells is derived entirely from the few left-over cell descendants
of the primordial line is difficult to say, for in my material at this
time there is a marked increase of germ cells which looks suspic-
iously like an active transformation of epithelial elements into sex
cells. The new germ-cell generation undergoes another matura-
tion cycle at the time of the metamorphosis of the tadpole, or
very shortly afterward, and gives rise to normal sex products.
Thus it is to be seen that the advocates of the germinal epithelium
theory may not be entirely mistaken, for in the light of events
in the germ cycle of the Anura, it is not sufficient to trace the
primordial sex cells into the genital ridges and there leave them.
Without following their later history, it is unjustifiable to assume
that they do give rise to the definitive sex products. This is
what most workers on the origin of germ cells in vertebrates have
done. On theoretical grounds the writer doubts if mesothelial
cells transform into sex cells but the morphological evidence from
my material does not rule out the possibility of such transfor-
mations.

The evidence presented by the precocious sexual cycle of the
bullfrog leads the writer to believe that further investigation
may perhaps show the existence of analogous phenomena through-
out the various groups of Chordata. It may not be too much
to state as a sort of working hypothesis for further investigation
of germ-cell development in the vertebrates, that most of the
primordial germ cells, i.e., those arising in early stages of embry-
onic development, undergo a precocious and abortive develop-
mental cycle, culminating in degeneration, or else degenerate
before the precocious maturation cycle has had sufficient time to
manifest itself. Furthermore, that the abortive developmental
phenomena are perhaps evidences of a phylogenetic regression
in the germ-cell cycle to remote ancestral conditions.

The question immediately arises if such an hypothesis, no mat-
ter how tentatively stated, is in any sense justified by evidence

GERM CELLS OF ANURANS 283

presented by merely a single group of vertebrates. The answer
is, that scattered here and there through the literature is consider-
able evidence of an extremely suggestive nature, derived from
study of the germ cells in widely separated groups of chordates,
which, when taken in conjunction with conditions existing in
anurans, suggest at once the working hypothesis stated above.
Some of this evidence will be briefly reviewed here.

In 1908 Von Winiwarter and Sainmont described two prolifera-
tions of cells from the germinal epithelium of the cat ovary; the
first formed the medullary, the second the cortical cords. <Ac-
cording to their account, the germ cells and follicle cells of the
first proliferation and the sex cells of the second all degenerate
and disappear when the young kitten is only a few months of
age. About three and a half to four months after birth, the
germinal epithelium of the ovary shows marked activity, and
proliferates a third generation of germ cells, from which the
definitive sex cells of the adult are derived. This third genera-
tion forms the cortex of the ovary in grown animals. In a short
paper published at nearly the same time as their monograph they
state that the definitive ova of the sexually mature cat are derived
either entirely from this third proliferation of the germinal epi-
thelium, or else partly from it and from undifferentiated cells left:
over from the second proliferation, i.e., those that failed to degen-
erate. For example, these authors state (p. 616):

Ks tauchen nun jetzt in den Epithelhaufen und Straingen der Corti-
calis kleine Gruppen von Zellen aut, deren Kerne im staubférmigen
oder deutobrochen Stadium sind. Diese Formen waren schon seit °
langer Zeit nicht mehr vorhanden, und da sie den ersten Stufen des
Wachstums des Oocyten entsprechen, ist es augenscheinlich, dass sie
mit einer Neubildung von Hiern zusammenhingen . . . . Wir
glauben bewiesen zu haben, dass in Séiugetierovarium nicht nur simt-
liche Markstrange, sondern auch alle Kier und Follikel der primitiven
Corticalis dem Untergang anheimfallen. Die definitiven Hier ent-
stammen entweder von undifferenzierten Zellen der zweiten Prolifera-
tion (Pfliigersche Schliuche) oder von Zellen der dritten Wucherung
oder Invaginationsepitheles. Es ist uns nicht méglich, wenigstens mor-
phologisch, die Elemente der einen und anderen zu unterscheiden.

Now, oddly enough, all of the embryonic (primordial) germ
cells described by Winiwarter undergo the characteristic nuclear

284 WILBUR WILLIS SWINGLE

changes of odcytes, such as leptotene, pachytene, and diplotene,
before undergoing degeneration.

Rubaschkin (’12) confirmed the conclusions of Von Winiwarter
and Sainmont that the cells of the first and second proliferations
in the cat degenerate. In the ovary of the guinea-pig this same
investigator observed a third proliferation of germinal cells from
the germinal epithelium which occurs before birth and which he
considers the source of the definitive sex cells.

Firket (’14) using female chick material, showed that the pri-
mordial germ cells pass through the first stages of maturation
previous to odcyte formation, leptotene, pachytene, etc., enter
the growth period and then degenerate. They all disappear in
the chick fourteen days after hatching. The odcytes of the cor-
tical zone (second embryonic proliferation) practically all degen-
erate, although he states that he cannot be sure that they all do.
There is a new formation of germ cells in the cortical region,
from cells derived from the germinal epithelium, and from these
the definitive odcytes develop; but it is not improbable, at least,
that a small number of the primordial germ cells are differenti-
ated into definitive ova. One of the conclusions Firket (14)
draws from his work is of considerable interest from the stand-
point of the results on the frog recorded here:

I] faut, done, morphologiquement parlant, considérer les gonocytes
primaires des Vertébrés comme étant un rappel phyolgénique des
gonocytes définitifs des classes inférieurs, notament des Cyclostomes et
des Acraniens. L’epuisement gradual, dans la série phylogénique
des éléments de cette lignée a nécessité l’apparition, au cours de l’onto-
@énése, d’une seconde lignée de gonocytes, moins précoces (pp. 330, 331).

Recently there came to my attention an abstract of a paper as
yet unpublished by this same author. I shall quote the abstract
entire because of the striking similarity of the conclusions of this
author, to some recorded by myself in this paper, both independ-
ently conceived:

In the testis and the ovary of the chick there are two generations of
germ cells: primary germ cells, which appear in very early stages,
before the genital ridge is formed, and secondary germ cells, which are

derived from the so-called ‘germinal epithelium.’ The former are
able to become odcytes, or spermatocytes, but while most of them

GERM CELLS OF ANURANS 285

degenerate, it is not possible to determine if any of them give rise to
definitive germ cells, because at a certain stage it is impossible to dis-
tinguish the former and the latter from each other. In the white rat
(male) the same two generations occur, but primary germ cells degen-
erate before they reach the period of growth and only secondary cells
become the definitive germ cells. That primary germ cells disappear,
in the ontogenesis, earlier in mammals than in birds, seems to show
that they must be considered as being cells in ‘phylogenetic regression.’

Two interesting papers by Kingery on the female white mouse
show that the phenomenon of primordial germ-cell degeneration
is found in the mouse, and perhaps even more important in this
connection is the fact that certain degenerating cells of this prim-
itive germinal line may undergo abortive maturation stages even
to the formation of first polar bodies.

This author (’14) in a study of the so-called parthenogenesis
in the mouse found that the degenerating primordial germ cells,
i.e., those of embryonic origin, undergo a degenerative fragmen-
tation and may even form a first polar body and second polar
spindle, and may even break up into fragments with or without
nuclei in much the same fashion as described by me for the larval
spermatocytes of the frog. It is interesting to compare the fig-
ures in Kingery’s paper with those of my own in degenerating
spermatocytes.

In a later paper by this same author (’17—’18), evidence of
the kind described for the cat by Winiwarter and Sainmont (’08)
and Rubaschkin (712), for birds and the white rat by Firket (14,
also ’20), and the male mouse by Kirkham (’16) is presented.
The first or embryonic set of germ cells in the female mouse pass
through early maturation stages, leptotene, pachytene, and dip-
lotene, enter the growth period of the odcyte, then degenerate.
The second generation of germ cells arise from the germinal epi-
thelium after birth and give rise to the definitive sex cells of the
adult female. He says:

The evidence shows that all these germ cells formed before birth
degenerate and are resorbed, none of them developing into definitive
ova. This degeneration takes the form of atrophy and resorption in
some cases, but in others there may occur atresia folliculi; accompanied
by the formation of a first polar body, and a degenerative fragmenta-
tion of the egg-cells, simulating more or less closely a parthenogenetic
cleavage.

286 WILBUR WILLIS SWINGLE

Kirkham (716) observed that the primordial germ cells of the
mouse first appear on the eleventh day after fertilization. In
male embryos these primitive germ cells all degenerate, and none
persist by the eighth day after birth. The definitive sex cells
of the male arise from undifferentiated epithelial elements accord-
ing to this account, whereas the definitive o6dgonia are direct de-
scendants of the primordial germ cells. It will be seen that
Kingery’s account of the definitive germ cells of the female mouse
agrees with Kirkham’s account for the corresponding conditions
in the male.

Felix states that in the human embryo the primordial germ
cells degenerate (no details given) and a new generation of sex
cells arises from the germinal epithelium which give origin to the
definitive sex cells (Keibel and Mall, Embryology).

Now it is obvious that evidence of this sort obtained by dif-
ferent investigators, working on vertebrate forms as widely sep-
arated as amphibia and mammals, must be of some significance.
In all vertebrates a definite Keimbahn probably exists; this is
certainly true of the frog, but the important question is, do the
primitive products of the keimbahn and their lineal descendants
in these vertebrate forms early undergo an abortive maturation
or developmental cycle which ends in degeneration such as occurs
in the bullfrog tadpole? Certainly, the evidence looks sugges-
tive. Apparently in the male bullfrog larvae this precocious
sexual cycle is carried further than in any other form so far
reported. The figures of Kingery for the female mouse and of
Winiwarter and Sainmont for the female cat indicate plainly
that the primordial germ cells are undergoing a precocious matura-
tion cycle. These figures show every phase in the maturation
cycle of normal eggs, such as leptotene, pachytene, diplotene,
and growth of the odcyte, yet, just as happens in the male tad-
pole, these early maturating cells degenerate. The same con-
dition is reported in birds. It is difficult to avoid the suspicion
that we are here concerned with a fundamental principle of germ-
cell development. The question arises, why should practically
all of the primordial germ cells of vertebrate undergo an abor-
tive sexual cycle long before the animal is mature and ready for

GERM CELLS OF ANURANS 287

reproduction, and then degenerate? The evidence from the
maturation cycle of the tadpole is again suggestive on this point.

In the tadpole we may assume, in so far as it is safe to assume
anything in biology, that the abortive and precocious sexual cycle
is possibly a case of phylogenetic regression to ancestral condi-
tions when the Anura were permanently of the caudate type and
lived and reproduced normally as Urodele-like creatures. ‘The
carrying over into the ontogeny of the anuran larva’s sexual cycle
of this phylogenetic vestige is not surprising, considering the
heavy impress of phylogeny upon the tadpole soma. Though
this explanation may be involved with much plausibility to
explain the larval sexual conditions of the bullfrog, is it in
any sense adequate to account for the apparently analogous
germ cycle of the Sauropsida and mammals, forms which do not
have a larval period? I believe the same explanation applies
to these forms also, and that the precocious developmental cycle
and degeneration of the primordial germ cells described in the
Amniota differ not in kind, but merely in the degree to which
the maturation cycle is carried from the larval sexual cycle of
the Anura.

In the Amniota we cannot speak of a precocious and abortive
larval germ-cell cycle, but we can speak of an abortive embryonic
sexual cycle, which, like that of thestadpole, possibly bears the im-
press of past phylogenetic conditions. And why not? If the em-
bryo of the higher vertebrates can develop gill clefts and a thousand
and one other evanescent phylogenetic vestiges in the course of
somatic development, it should not be regarded as extraordinary
if the germ-cell cycle likewise presented similar ‘ancestral remi-
niscences,’ and did a little recapitulating on its own account.

However, it must be confessed that ‘phylogeny,’ ‘recapitula-
tion,’ ‘ancestral reminiscences,’ and other vague and more or less
mystical terms of a kindred nature are after all merely conveni-
ent pegs upon which to hang our ignorance. There are immediate
physicochemical reasons for the degeneration of the primordial
germ cells or their abortive sexual cycle, but what these reasons
are is unknown, and in view of a better or more plausible, hypo-
thesis to account for this phenomenon, the one presented above
is advanced tentatively.

288 WILBUR WILLIS SWINGLE

Another point is worthy of consideration here, and that is the
possibility of bringing about some measure of reconciliation be-
tween the ‘entodermists,’ or advocates of the Keimbahn, and the
‘mesodermists.’ In view of the evidence presented by study of
germ-cell origin in all classes of vertebrates, there can be no
reasonable doubt that the primordial sex cells are products of
the entoderm, and probably migrate into the germ ridges at an
early period of development. However, according to the hypoth-
esis advanced here these primordial cells, after a period of mul-
tiplication, undergo an abortive developmental cycle and for the
greater part degenerate—perhaps, in mammals, entirely degen-
erate. The new cell generation destined to give origin to the
definitive sex cells may possibly arise in part from the germinal
epithelium by direct transformation of mesothelial elements.
The evidence for this point of view is suggestive, at any rate,
judging by reports on conditions in the birds and mammals,
and there is little evidence to the contrary, but many pure
assumptions.

In the bullfrog the writer prefers to believe that some cells of —
the primordial germ-cell line persist unchanged through the phase
of maturation and degeneration, and ultimately, by repeated
mitosis are the chief, and probably only contributors to the cells
of the definitive sexual line. There is considerable evidence for
this view, because a few primordial spermatogonia or at any rate
lineal descendants of these cells can be traced through the sexual
cycle easily enough, but it is by no means certain that they are
the sole contributors to the definitive line of germ cells.

Thus it appears possible that there is some basis here for rec-
onciliation between the entodermists and mesodermists regarding
germ-cell origin and development. The former have been at fault
by contenting themselves with tracing the primordial sex cells
into the genital glands and there leaving them, with the assump-
tion that they persist and form the sexual elements of the adult
organism. The mesodermists, working chiefly on mammals,
have for the most part ignored the contributions of the ‘ento-
dermists’ because they have been unable to trace the germ cells
back to the very earliest stages such as described for the lower
vertebrates.

GERM CELLS OF ANURANS 289

No investigation of the germ-cell cycle in the Chordata should
be regarded as complete or as being more than a half-truth
which does not take into consideration the entire history of the
germ-cell cycle, from the origin of the primordial germ cells to
the formation of the definitive sexual elements of the adult. The
investigators of the keimbahn have not gone far enough, for
between the origin of the primordial germ cells and the formation
of the ripe sexual products there is a critical stage in the germ-cell
cycle, characterized by. a precocious and abortive maturation,
degeneration, and reformation of a new line of germ cells, per-
haps by transformation of mesothelial elements, but more proba-
bly by active mitosis of a few left-over cells of the primordial line.

According to Hegner (Germ-cell cycle of animals, p. 99), ger-
minal epithelium theories of germ-cell origin have little if any evi-
dence in their favor, since no one has actually observed a trans-
formation of peritoneal or mesoblast cells into germ cells. ‘On
the other hand there is an abundance of proof that these cells
(germ cells) migrate from some distance into the position of the
sex glands.”

The writer is quite in agreement with Hegner regarding the
existence of a keimbahn in vertebrates, but is not so sure that no
one has actually observed a transformation of mesoblast cells
into germ cells in late stages of development or that germinal-
epithelium theories have little if any evidence in their favor. My
observations on the larval bullfrog have taught me caution in
regard to dogmatizing on this problem. Odd as it may seem, it
is not impossible, in the light of conditions described above for
the bullfrog larvae, that the primordial germ cells of vertebrates,
i.e., the keimbahn elements discussed by Hegner, may possibly be
found upon further investigation to contribute little if any to
the definitive sex products of the adult organism. Further inves-
tigation of the germ-cell cycle of the chordates may possibly
enable the ‘mesodermists’ to turn the tables on the ‘entoder-
mists’ with a vengeance by showing that no one has actually
observed a transformation of keimbahn cells into definitive sex
products. Though regarding himself as an entodermist, and tak-
ing the point of view that the keimbahn is probably continuous in

290 WILBUR WILLIS SWINGLE

vertebrates, and that there is no actual transformation of meso-
thelial elements into sex cells, the writer admits that conditions are
such in the bullfrog that it is impossible to state positively that the
primordial germ cells of the bullfrog tadpole do give rise to the
definitive sex cells of the adult frog. Certainly, this is the more
probable view, though the burden of proof rests with those of us
who hold that the keimbahn is continuous.

It would seem from this that the crux of the whole problem is
to determine whether or not germ cells can develop in an organ-
ism after the primordial germ cells have been destroyed. If
they do develop, then the doubtful question of transformation
of mesothelial cells into germ cells is settled in favor of the meso-
dermists, but if they do not develop, and the gonad is sterileand
remains so up to the period of sexual maturity, then the decision
is in favor of the entodermists. It is not sufficient to extirpate
the primordial germ cells or otherwise destroy them, as was done
by Reagan in the embryo chick, and then report the resulting
sterile gonad as conclusive evidence against the idea of a trans-
formation of epithelial elements, because proof positive can only
be had by rearing the animals to sexual maturity.

The only adequate method of attack upon this problem is by
experimental methods. Morphological methods are not sufficient
to determine whether or not a germ cell in the germinal epithelium
or sex cord tissue is a transformed epithelial element or a small
germ-cell descendant of the primordial line. Transition stages,
nuclear configuration of the cell, size, position, and such like may
be illusory. In my material there is apparently every transition
stage between peritoneal and true germ cells, use whatever
morphological criterion you please, at certain developmental
stages of the tadpoles, and such transition stages almost fill the
gonads, but always the question arises upon examining these
‘apparent transition stages—they look exactly like mesothelial
cells transforming into germ cells, but are they? If one must
judge from morphological data alone, the answer is that they
could very readily be taken for mesothelial elements transforming
into germ cells, but, as stated before, the morphological criterion
alone does not furnish sufficient evidence to nett one to make
a definite answer.

GERM CELLS OF ANURANS 291

SUMMARY OF CONCLUSIONS
A. The origin and fate of the primordial germ cells

1. The primordial germ cells of the embryo bullfrog are first
distinguishable from other entodermal elements in embryos of 7
mm. total length. They arise from the entoderm as a median
ridge of yolk-laden cells just dorsal to the roof of the archenteron,
ventral to the aorta, and separating the two lateral mesodermal
plates from each other.

2. In embryos of 8 mm. total length, the germ-cell ridge be-
comes separated from the underlying entoderm forming the roof
of the archenteron, partly by the median growth of the two
lateral plates which pinch off the ridge and also by active
migration of the germinal elements themselves. In cross-section
at this stage, the germ cells are found at the root of the forming
mesentery as an unpaired ridge, consisting of two or three large
yolk-laden cells.

3. As development progresses, this median ridge of germ cells
splits longitudinally and the cells of the two halves migrate lat-
erally on either side to form two independent ridges, invested
with peritoneum. ‘This stage is represented in embryos of 9.5
mm. total length.

4. The two germinal ridges project into the coelomic cavity
and enlarge considerably by increase in number of their cellular
elements. The primitive sex cells actively divide and there is
also a migration of mesenchymal cells into the ridges from the
mesonephros and peritoneum. ‘These conditions are found in
14 to 15 mm. tadpoles. The germ cells have lost their yolk in
the meantime.

5. The gonads greatly increase in size. Large cavities are
formed, the secondary genital spaces, lined by small non-sexual
cells which have migrated into the gonads from the mesonephros
by way of the mesentery suspending the gland. When the tad-
pole has attained a length of 30 mm., the gonads are hollow sacs
surrounded by a single layer of peritoneum and one or two layers
of germ cells.

292 WILBUR WILLIS SWINGLE

6. All increase in the number of germ cells in male larvae up to-
the 40-mm. stage is beyond question, by mitotic division of the
preexisting sexual elements derived from the primordial germ
cells of the entoderm ridge. A 40-mm. larva is about one vear
of age.

7. At the 40-mm. stage, despite the fact the tadpole is an imma-
ture larva and the gonads mere hollow sacs and in no way re-
semble testes, the germ cells enter maturation and pass through
every stage of the maturation cycle in a normal manner, up to
the first maturation division. In the act of division, the sperma-
tocytes go to pieces and are resorbed.

8. Practically all of the germ cells derived from the primordial
sex cells pass through this abortive maturation cycle and degen-
erate. A very few germ cells lineal descendants of the primordial
embryonic sex-cell line persist unchanged, i.e., remain as sperma-
togonia through the maturation cycle and do not degenerate.
Later these few cells give rise to a second generation of smaller
germ cells.

9. This second generation of germ cells shortly before meta-
morphosis of the larvae undergoes a second sexual cycle, charac-
terized by the production of normal spermatozoa. Thus there
are two larval sexual cycles: one occurring in immature larva of
40 to 60 mm. and ending in degeneration, the other appearing
shortly before metamorphosis, i.e., in larvae 140 mm. total length
and ending in the production of normal sex products.

10. In the interval between the first and second larval sexual
cycles following the degeneration of large numbers of maturation
cells the gonads become filled with small cells which, because of
their size, nuclear structure, and staining capacity, appear as
transition stages between mesothelial cells (germinal epithelium
and sex cord elements) and true germ cells. The later history
of these cells shows them to be germ cells, but their origin is open
to two interpretations and is not as clear as could be desired.
The writer considers these cells as small germ-cell descendants
of the primordial sexual elements, and not as transformed germi-
nal epithelium elements, but admits that the evidence from his
material is equally strong in support of the germinal epithelium
view-point.

GERM CELLS OF ANURANS 293
XN

11. The sexual] cycles of the larval bullfrog are tentatively
interpreted, in lieu of a more satisfactory hypothesis, as recapitu-
lations of the germ-cell cycle to past phylogenetic sexual condi-
tions when the vertebrates ripened their sex products at an earlier
developmental stage than at present.

12. An analogous precocity of the maturation cycle probably
exists in all of the vertebrates, Amniota as well as Anamnia.
Evidence for this hypothesis is presented in detail.

B. The chromosomes and larval sexual cycles

1. The diploid number of chromosomes in the male larva is
twenty-eight. The elements are J-:and V-shaped and curved
rods. Portions of certain chromosomes do not take the stain
under any circumstances and may give the appearance of frag-
menting into two or more parts. Such appearances of fragmenta-
tion are illusory.

2. Spindle-fiber attachment is non-terminal.

3. The chromosomes exist in definite pairs according to size and
shape, i.e., there are fourteen pairs of homologues.

4. The homologues of any pair are not invariably found side
by side within the nucleus, though in general they are near
together.

5. The size and shape relations of the chromosomes are per-
fectly definite throughout all cell generations, and this is probably
true not only for the individual, but for the larvae of the species
as a whole.

6. As an illustration of the statement just made (number 5),
see chromosome pair marked A in figure 6. These chromosomes
are peculiar in that the knob-like end-piece is separated from the
main body of the chromosome by a clear, non-stainable area.
This peculiarity is probably constant in the cells of the larvae,
and has been observed in the spermatogonia of twenty-nine indi-
viduals of various ages and stages of development.

7. The resting nucleus of the Anuran germ cell is a polymorphice,
much-lobulated structure, made up entirely of chromosome vesicles
which are incompletely fused, and in many cases the vesicles are

294 WILBUR WILLIS SWINGLE

entirely independent. By means of these vesicles the chromo-
somes preserve their identity through the so-called resting stages.

8. The cells and chromosomes of the larvae are considerably
larger than those of the adult frog, and more nearly resemble the
cells and chromosomes of urodeles than those of the adult of their
own species.

9. This size difference between the larval and adult cells and
chromosomes is explained in detail in the text and is considered
to be due in part to the number of intervening cell divisions,
with reduction of cell and chromosome size the greater the num-
ber of divisions, and to reduction in cell size at the time of
metamorphosis due to loss of water from the tissues.

10. The haploid number of chromosomes in the larvae is four-
teen. The tetrads of the first larval sexual cycle are extremely
large and of the open-ring type characteristic of urodeles. They
differ markedly from the type of tetrad appearing in the second
larval sexual cycle.

11. Conjugation of the chromosomes is by parasynapsis, and
occurs in the amphitene stage, when the leptotene threads twist
together to form the pachytene.

12. Evidence of ‘crossing-over’ during diakinesis, such as fig-
ured by Janssens for Batracoseps, has not been observed, or
rather, has been observed but not interpreted as such. The
chiasma-type which appears during diakinesis stages of the bull-
frog larvae has been interpreted by the writer as tetrads opening
out in two planes at right angles to one another thus giving the
appearance of ‘crossing-over’ of the chromatids.

13. It is suggested that ‘crossing-over’ occurs during the dragltee
tene stage, when the conjugating leptotene threads coil tightly
about each other corkscrew fashion.

14. The first larval maturation cycle is normal in every respect
save for the size of the cells and chromosomes, up to the formation
of the first maturation spindle. 'Thespermatocytes degenerate in
the act of division.

15. The cause of the degeneration of the larval spermatocytes
is the abnormal behavior of the centrosome which fragments,
forming accessory asters and spindles. It is recognized that the

GERM CELLS OF ANURANS 295

centrosomal behavior is but a symptom of a deep-seated proto-
plasmic disorganization of the larval sex cells.

16. Giant spermatid-like structures are formed by the suppres-
sion of the first and second maturation divisions and the growth
of an axial fiber from the centrosome. These bizarre structures
degenerate.

17. The cells of the first larval sexual cycle degenerate and dis-
appear gradually. A few cells, lineal descendants of the primor-
dial germ cells, persist unchanged through the cycle of matura-
tion and degeneration, and give rise by repeated mitosis to a
second germ-cell generation in larvae just about ready for meta-
morphosis. This second cell generation is small in size.

18. Shortly before metamorphosis, this second generation of
germ cells undergoes a second sexual cycle, characterized by the
formation of normal spermatozoa. The cells and chromosomes
are comparable in every way with those of the adult frog and
are smaller than the larval cells and chromosomes.

19. The second larval sexual cycle is normal in every respect.
There is no degeneration of the sexual elements. The matura-
tion cycle is normal, as are also the spermatozoa, despite the fact
the animal is a larva with the efferent ducts of the testis incom-
pletely formed.

20. In so far as the possession of ripe spermatozoa is concerned,
the larval bullfrog at metamorphosis may be said to be mature,
and in this respect resembles the axolotl.

21. The germ cells of female larvae at the time of metamor-
phosis are not mature, but are young o¢ccytes undergoing growth.
The writer has some evidence that, like the male, the female
larvae may also show a precocious and abortive maturation cycle.
This point is now under investigation.

22. The question of hermaphroditism and the sex ratios of the
Anura is not dealt with in this paper, but forms the subject-matter
of a later communication.

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

296 WILBUR WILLIS SWINGLE

LITERATURE CITED

Apter, Leo 1916 Untersuchungen tiber die Entstehung der Amphibienneo-
tenie. Pfliiger’s Archiv fiir die gesamte Physiologie, Bd. 164.

Acar, W. E. 1911 The spermatogenesis of Lepidosiren paradoxa. Quart.
Jour. Mier. Sc., vol. 57.

ALLEN, B. M. 1906 The origin of the sex-cells of Chrysemys. Anat. Anz., Bd. 29.
1997 An important period in the history of the sex-cells of Rana
pipiens. Anat. Anz., Bd. 31.

1909 The origin of the sex-cells of Amia and Lepidosteus. Anat.
Rec., vol. 3. :

Amma, K. 1911 Uber die Differenzierung der Keimbahnzellen bei den Cope-
poden. Arch. f. Zellforsch., Bd. 6.

BauiowiTz, E. 1905 Die Spermien des Batrachiers Pelodytes punctatus.
Anat. Anz., Bd. 27.

1906 Uber das rege]miissige Vorkommen heteromorpher Spermien
im reifen Sperma des Grasfrosches (Rana muta). Zool. Anz., Bd. 30.

BAUMGARTNER, W. G. 1911 Spermatogenesis in the male cricket. Science,
INS Soy Ol, BS.

Brenna, C. 1893 Zellstructuren und Zellteilungen des Salamanderhodens.
Verhandl. der Anat. Gesellschaft.

Bonnevib, K. 1908 Chromosomenstudien. I. Chromosomen von _ Ascaris,

Allium und Amphiuma. Arch. f. Zellforsch., Bd. 1.
1908 Chromosomenstudien. II. Heterotypische Mitose als Rei-
fungs-Character. Nach Untersuchungen an Nereis limbata Ehlers,
Thalasema mellita Conn, und Cerebratulus lacteus. Arch. f. Zell-
forsch., Bd. 2.

Boutin, M. 1900 Histogénése de la glande genitale femelle chez Rana tempo-
raria. Arch. de Biol., T. 17.

Broman, J. 1900 Uber Riesenspermatiden bei Bombinator igneus. Anat.
ANIWAn, IBGlG Iie
1900 Uber Bau und Entwickelung der Spermien von Bombinator
igneus. Anat. Anz., Bd. 17.

Branca, A. 1907 La spermatogénése chez l’axolotle. Arch. Zool. Exp. et
Générale. Notes et Revues.

CaroTHERS, EH. ELEANOR 1916-17 The segregation and recombination of homol-
ogous chromosomes as found in two genera of Acrididae (Orthoptera).
Jour. Morph., vol. 28.

1913 The mendelian ratio in relation to certain orthopteran chromo-
somes. Jour. Morph., vol. 24.

Carnoy, J. B., rt Lesrun, H. 1900 La vesicule germinative et les globules
polaires chez les Batrachiens. 2 partie, les Anoures. La Cellule,
gt bf

Cuavyin, Marie V. 1875 Uber die Verwandlung des mexikanischen Axolotl in
Amblystoma. Zeitschr. f. wissensch. Zool., Bd. 25, Suppl., 1875, und
Bd. 27, .1876.

1885 Uber die Verwandlungsfihigkeit des mexikanischen Axolotl.
Zeitschr. f. Wissensch. Zool., Bd. 41.

GERM CELLS OF ANURANS 297

Cuampy, C. 1913 Recherches sur la spermatogénése des Batrachians et les
éléments accessoires du testicle. Arch. de Zool., Exp. et Générale,
T. 52.
1908 Sur la dégénérescence des spermatogonies chez la grenouille
verte. Compt. Rend. Assoc. Anat. Mar.

Conxuin, E. G. 1897 The embryology of Crepidula. Jour. Morph., vol. 13.
1896 Cleavage and differentiation. Woods Hole biol. lecture.
1901 The individuality of the germ nuclei during the cleavage of the
egg of Crepidula. Biol. Bull., vol. 2.
1901 Centrosomes and spheres in the maturation, fertilization and
cleavage of Crepidula. Anat. Anz., Bd. 19.
1902 Karyokinesis and cytokinesis in the maturation, fertilization
and cleavage of Crepidula and other Gasteropoda. Jour. Acad. Nat.
Sci. Philadelphia, vol. 12.
1915 Why polar bodies do not develop. Proceedings Nat. Acad.
Sciences, vol. 1.
1917 Effects of centrifugal force on the structure and development
of the eggs of Crepidula. Jour. Exp. Zodl., vol. 9.

Denorne, A. 1911 Recherches sur la division de la cellule. I. Le duplicisme
constant du chromosome somatique chez Salamandra maculosa et
chez Allium cepa. Arch. f. Zellforsch., Bd. 6.
1910 Le nombre des chromosomes chez les batrachiens et chez les
larves parthénogenetiques de grenouille.

Dumérit, A. 1865 Nouvelles observations sur les Axolotls nés a la ménagerie.
Comp. Rend., T. 61.

Eisen, G. 1900 The spermatogenesis of Batrachoseps. Jour. Morph., vol. 17.

Fick, R. 1893 Uber die Reifung und Befruchtung des Axolotl. Zeit. f. wiss.
Zool., Bd. 56.

Firxet, J. 1914 Recherches sur l’organogénése des glandes sexuelles chez les
oiseaux. Arch. de Biol., T. 29.

Fiemmine, W. 1887 Neue Beitrige zur Kenntniss der Zelle. Arch. f. mikr.
Anat., Bd. 29.

Gricotre, V. 1905 Les résultats acquis sur les cinésis de maturation dans les
deux régnes, Mém. I. Les cinésis de maturation dans les deux régnes.
L’unité essentielle du processus meitotique. Mem. II. La Cellule,
26:
1907 La formation des gemini hétérotypiques dans les végétaux. La
Cellule, T. 24.

HAcxker, V. 1895 Die Vorstadien der Hireifung. Arch. f. mikr. Anat., Bd. 45.

Hance, R. T. 1917 The fixation of mammalian chromosomes. Anat. Rec.,
vol. 12.
1917 The diploid chromosomes of the pig (Sus scrofa). Jour. Morph.,
vol. 30.
1917 The somatic mitosis of the mosquito, Culex pipiens. Jour.
Morph., vol. 28.
1918 Variations in somatic chromosomes. Biol. Bull., vol. 35.

Hartmann, F. A. 1913 Variations in size of chromosomes. Biol. Bull., vol.
24.

298 WILBUR WILLIS SWINGLE

HinpvererR, H. 1891 Ein Fall von Festhalten der Larvenform bei Frosch
lurchen. Blatter f. Aquarien- u. Terrarienkunde, Bd. 2 (quoted from
Adler).

JANSSENS, F. A. 1901 La spermatogénése chez les Tritons. La Cellule, T. 19.
1905 Evolution des auxocytes males du Batracoseps attenuatus.. La
Cellule, T. 22.

1909 La Théorie de la chiasmatypie. La Cellule, T. 25.

JANSSENS, F. A., pt Dumez, R. 1903 L’élément nucléinien pendant les cinésis
de maturation des spermatocytes chez Batracoseps attenuatus et
Plethodon cinereus. La Cellule, T. 20: ;

JANSSENS, F..A., pT WituEMs, J. 1908 La spermatogénése L’Alytes Obstetri-
cans. La Cellule, T. 25.

JULLIEN 1868 Observations de tétards de Lissotriton punctatus reproduisant
l’espéce. Paris. Compt. rend., T. 68.

Kingery, H. M. 1914 So-called parthenogenesis in the white mouse. Biol.
Bull., vol. 27.

1917-18 Odgenesis in the white mouse. Jour. Morph., vol. 30.

Kinessury, B. F.. 1899 Reducing division in the spermatogenesis of Desmo-
gnathus fusca. Zool. Bull., vol. 2.

1901 The spermatogenesis of Desmognathus fusca. Am. Jour. Anat.,

WOle wis
KrrkHamM, W. B. 1916 The germ cell cycle in the mouse. Abstract in Anat.
Rec., vol. 10.

Kuscnakewitscu, 8. 1910 Die Entwicklungsgeschichte der Keimdriisen von
Rana esculenta. Festschr. f. R. Hertwig.

Koxtimann, J. 1884 Das Uberwintern von europiischen Frosch- und Triton-
larven und die Umwandlung des mexickanishen Axolotl. Verhandl.
d. Naturf. Gesellsch., Basil, 1884.

KKAMMERER, P. 1906 Experimentelle Verinderung der Fortpflanzungstitigkeit
bei Geburtshelferkréte (Alytes obstetricans) und Laubfrosch (Hyla
arborea). Arch. f. Entwicklungsmech., Bd. 22.

Kine, H.D. 1907 The spermatogenesis of Bufo lentiginosus. Am. Jour. Anat.,
Vol”, No. 3.

1912 Dimorphism in the spermatozoa of Necturus maculosus. Anat.
Rec., vol. 16.

KowatskI, F. 1904 Reconstitution du noyau et formation des chromosomes
dans les cinéses somatiques de la larvae de Salamandre.

Kostanecxi, V. K. 1892 Uber Kernteilungen bei Riesenzellen. Anat. Hefte,
Bd. 1. :

Laucun, A. 1913 Uber pluripolare mitosen in Hodenregeneraten von Rana
fusca. Arch. f. mikr. Anat., Bd: 82.

Mack, J.B. 1914 A study of the dimensions of the chromosomes of the somatic
cells of Amblystoma. Kans. Univ. Science Bull., vol. 9.

Meap, A. D. 1898 The origin and Behavior of the Centrosomes in the Annelid
Egg. Jour. Morph., Vol. XIV, No. 2.

Meves, F. 1891 Uber amitotische Kerntheilung in den Spermatogonien des
‘Salamanders, und das Verhalten des attractionssphiren bei derselben.
Anat. Anz., bd. 5.

GERM CELLS OF, ANURANS 299

Meves, F. 1896 Ueber die Entwicklung der minnlichen Geschlechtszellen von
Salamandra maculosa. Arch. f. mikr. Anat., Bd. 48.
1911 Chromosomenliingen bei Salamandra, nebst Bemerkungen zur
Individualititstheorie der chromosomen. Arch. mikr. Anat., Bd. 77.

Metz, C. W. 1914 A preliminary study of five different types of chromosome
groups in the genus Drosophila. Jour. Exp. Zoél., vol. 17.

McCuune, C. E. 1899 A peculiar nuclear element in male reproductive cells
of insects. Zool. Bull., no. 2.
1900 The spermatocyte divisions of the Acrididae. Kansas Uni-
versity Quarterly, vol. 9.
1914 A comparative study of the chromosomes in orthopteran sper-
matogenesis. Jour. Morph., vol. 25.
1917 The multiple chromosomes of Hesperotettix and Mermiria
(Orthoptera). Jour. Morph., vol. 29.

Montaomery, T. H. 1903 The heterotypical maturation mitosis in Amphibia
and its general significance. Biol. Bull., vol. 4, no. 5.
1911 The spermatogenesis of an Hemipteran, Euschistus. Jour.
Morph., vol. 22.

Moraan, T. H. 1917 Heredity and sex. Columbia University Press. .

Nowuin, Napine 1908 The chromosome complex of Melanoplus bivitattus Say.
Kansas Univ. Science Bull., no. 4.

Nusspaum, M. 1880 Zur Differenzierung des Geschlechts im Tierreich. Arch.
f. Mikr. Anat., Bd. 18.

Payne, F. 1914 Chromosomal variations and the formation of the first sperm-
atoecyte chromosomes in the EKuropean earwig, Forficula sp. Jour.
Morph., vol. 25. ;

PARMENTER, C. L. 1919 Chronftosome number and pairs in the somatic mitosis
of Amblystoma tigrinum. Jour. Morph., vol. 33, no. 1.
1920 The chromosemes of parthenogenetic frogs. Jour. Gen.
Physiol., vol. 11, no. 3.

Proucu, H.H. 1917 The effect of temperature on linkage in the second chromo- .
some of Drosophila. Proc. Nat. Acad. Science, vol. 3.

Rawirz, B. 1895 Centrosoma und Attraktionssphire in der ruhenden Zelle
des Salamanderhodens. Arch. f. mikr. Anat., Bd. 44.

Rogpertson, W. R. B. 1915 Chromosome studies. III. Inequalities and defi-
ciencies in homologous chromosomes. Jour. Morph., vol. 26.
1916 Chromosome studies. I. Taxonomic relations shown in the
chromosomes of Tettigidae and Acrididae: V-shaped chromosomes and
their significance in the Acrididae, Locustidae and Gryllida: Chromo-
somes and variation. Jour. Morph., vol. 27.

Rupascukin, W. 1912 Zur Lehre von der Keimbahn bei Siiugetieren. Anat.
Hefte, Bd. 46.

ScHREINER, A. AND K. E. 1906 a Neue Studien uber die Chromatinreifung der
Geschlechtszellen. I. Die Reifung der minnlichen Geschlechtszellen
von Tomopterus onisciformis. Arch. de Biol., T. 22.
1906 b Ibid. II. Reifung der minnlichen Geschlechtszellen von Sal-
amandra maculosa, Spinax niger and Myxine glutinosa. Ibid.

Snook, H. J., anp Lona, J. A. 1914 Parasynaptic stages in the testis of Aneides
lugubris (Hallowell). University California, pub., vol. 15.

300 WILBUR WILLIS SWINGLE

Stevens, N. M. 1908 A study of the germ cells of certain Diptera, with refer-
ence to the heterochromosomes and the phenomena of synapsis. Jour.
Exp. Zodl., vol. 5.

Swineite, W. W. 1917 The accessory chromosome in a frog possessing marked
hermaphroditic tendencies. Biol. Bull., vol. 33.
1918 The effects of inanition upon the development of the germ
glands and germ cells of frog larvae. Jour. Exp. Zodl., vol. 24, no. 3.
1918 The acceleration of metamorphosis in frog larvae by thyroid
feeding and the effects upon the alimentary tract and sex glands.
Jour. Exp. Zoél., Vol. 24, no. 3.
1919 Studies on the relation of iodine to the thyroid. I and II.
Jour. Exp. Zoél., vol. 27, no. 3.
1919 Todineandthethyroid. I]I]and IV. Jour. Gen. Physiol., vols.
2 and 3.
1920 Homoplastic thyroidal transplants and the metamorphosis of
Rana catesbeiana. Anat. Assoc. Abstracts.
1920 Neoteny and the sexual problem. Amer. Nat., July-August.

FIscHER-SIGWART, H. 1896 Die Fortpflanzung und die Entwicklung der Larv-
en vom Molge vulgaris. Das Uberwintern der Larven. Der Zool.
Garten, No. 37.

SHuFretpT, R.W. 1885 Mexican axolotl and its susceptibility to transformation.
Science, vol. 6.

Von Ratu, O. 1893 Ueber die Bedeutung der amitotischen Kerntheilung im
Hoden. Zool. Anz., Bd. 14.
1893 Beitrige zur Spermatogenese von Salamandra. Zeitschr. f.
Wiss. Zool., Bd. 57.

Von WINIWARTER ET SatnMontT 1908-1909 Nouvelles recherches sur l’ovo-
génése et l’organogénése de l’ovaire dans mammiféres (Chat). Arch.
de Biol., T. 24.

Vespévsky, F. 1911-12 Zum problem der Vererbungstriger. Bohm. Gesell.
Wiss., Prag.

Wenricu, D. H. 1917 Synapsis and chromosome organization in Chorthippus
(Stenobothrus) curtipinnis and Trimerotropos suffusa (Orthoptera).
Jour. Morph., vol. 29. ,
1916 The spermatogenesis of Phrynotettix magnus with special ref-
erence to synapsis and the individuality of the chromosomes. Bull.
Mus. Comp. Zo6l. Harvard, vol. 60.

Witson, E. B. 1906 Studies on chromosomes. III. The sexual differences of
the chromosomes in Hemiptera, etc. Jour. Exp. Zodl., vol. 3.
1909 Studies on chromosomes. V. The chromosomes of Metapodius,
etc. Ibid., vol. 6.
1910 Studies on chromosomes. VI. A new type of chromosome com-
bination in Metapodius. Ibid., vol. 9.
1911 Studies on chromosomes, VII. A review of the chromosomes of
Nezara, etc. Jour. Morph., vol. 22.
1912 Studies on chromosomes. VIII. Observations on the matura-
tion phenomena in certain Hemiptera, ete. Jour. Exp. Zodl., vol. 13.

Wintwarter, H. von 1900 Recherches sur l’ovogénése et l’organogénése de
Vovaire des mammiféres (Lapin et Homme). Arch. de Biol., T. 17.

GERM CELLS OF ANURANS 301

Witscu1, Emin 1914 Experimentelle Untersuchungen uber die Entwicklungs-
geschichte der Keimdriisen von Rana temporaria. Arch. f. mikr.
Anat., Bd. 85.
Wricut, A. H. 1914 North American Anura. Life-histories of the Anura of
Ithaca, New York. Publ. Carnegie Institute, Washington.
Wa.terstorrr, W. 1896 Uber die Neotenie der Batrachier. Der Zool. Gar-
ten, no. 37.

Wicuanp, B. 1906 Uber Neotenie bei Tritonen. Blitter fiir Aquarien- und
Terrarienkunde (quoted from Adler, Leo).

Wootsey, C.I. 1915 Linkage of chromosomes correlated with relation in num-
bers among the species of a genus, also within a species of Locustidae.
Biol. Bull., vol. 28.

EXPLANATION OF PLATES

All drawings were outlined with camera lucida; 2-mm. oil-immersion objective
used, ocular 12; hence are of the same magnification. Plates 1, 2, and 3 have been
reduced one-third.

PLATE 1
EXPLANATION OF FIGURES

1 Primary spermatogonium surrounded by follicle. Note polymorphic
nucleus, i.e., chromosomal vesicles. From first-year larvae.

2 Diagram of early division prophase of primary spermatogonium. The
chromosomes appear in vesicles which make up the entire nucleus and give it the
lobulated appearance.

3 Diagram of later prophase. Chromosome vesicles have disappeared.

4and6 Equatorial plates showing twenty-eight chromosomes. Note Chro-
mosome pair marked A. The peculiarity of an end-piece attached by a non-
staining area is constant. ,

5 Odd type of spermatogonial prophase chromosomes may appear in best
fixed material, Chromosomes appear as solid balls. Abnormal cell evidently
degenerating.

7 Odd cell division. Spindle oriented in short axis of cell.

8 Resting nucleus after last spermatogonial telephase. Note the chromatin
blocks and linin fibrils. Larvae 45 mm. total length.

9 Preleptotene stage showing resolution of the chromatin blocks into fine
threads. Larvae 40 mm. total length.

10 Large cell with leptotene threads. Note the chromoplasts with attached
fibrils.

11 Amphitene nucleus. Pachytene loops at proximal pole, unpaired lepto-
tene filaments at distal pole. This cell iilustrates the formation of the pachytene
bouquet. See plate 5, fig. 39. Larva 45 mm. total length.

12 Isolated pachytene threads showing unpaired leptotene filaments at dis-
tal ends. Compare with figures 36 to 39, plate 5.

302

GERM CELLS OF ANURANS PLATE 1
WILBUR WILLIS SWINGLE

303

PLATE 2
EXPLANATION OF FIGURES

13 Amphitene nucleus. Larva 45 mm.

14 Diplotene stage showing splitting of the thick pachytene threads and their
separation in the middle. Note attachment at synaptic ends.

15 Diplotene nucleus. Compare with figures in plates 5 and 6. Larvae 45
to 60 mm.

16 Diplotene nucleus showing early formation of tetrads. Larvae 45 to 60
mm. No evidence of crossing-over.

17 Condensation of the type of tetrad shown in figure 16. Note the woolly
appearance. Larva 74 mm.

18 First spermatocyte tetrads and tripolar spindle. Fourteen tetrads pres-
ent. First-year tadpole. See figure 79.

19 Spermatocyte with eighteen tetrads. Very unusual condition. Larvae
80 mm. total length.

20 Tetrads of larvae of first year. See plates9toll. X in figure 20 indicates
persisting karysomal structure of unknown origin and fate.

21 Spermatocyte with multiple asters. First-year larvae 40 to 50 mm. total
length.

22 Giant spermatid-like body resulting from degeneration of larval sperm-
atocytes like those shown in figures 106 and 108. See also plate 13.

23 Different type of spermatid-like body. The black masses represent the
ring tetrads which have run together. See plate 13 also. Larva 80 mm.

304

305

PLATE 3
EXPLANATION OF FIGURES

Fig. 24 and 25 Large larval spermatocyte. Cell cut. Both figures are of
one cell. Note the body resembling a yolk nucleus in figure 25. Larva 100 mm.

26 and 27 Both figures of same cell. Note rod-shaped tetrad in figure 26.
Unusual condition. Larva 80 mm.

28 Giant spermatocyte of first-year larva. Note rod tetrad and extreme
nuclear size. This cell shown in photographs 64 and 67. Larva 80 mm.

29 and 30 Sections of the same cell. Note the Y-shaped and cross-shaped
tetrads. Cell of unusual size. Note similarity of tetrads to those of urodeles.
Same cell shown in photographs 87 and 88. Larva 100 mm.

31 and 32 Sections of same cell. Note the rod-shaped and cross tetrads.
Cell of unusual size. Figure 114 is a photograph of a portion of this cell. Larva
95 mm.

306

WILBUR WILLIS SWINGLE

PLATE 4
EXPLANATION OF FIGURES

Microphotographs on this plate made at magnification of 50 diameters. No
reduction.

33 Transverse section through male gonad shown in text figure 1,a. Note
the very large secondary genital cavity surrounded by double layer of germ cells.
Animal 40 to 50 mm. total length. Maturation of the germ cells begins in many
instances in undifferentiated glands of this type.

34 Cross-section through male gonad of first-year tadpole (text fig. 1, b).
The large cavity is disappearing, owing to rapid proliferation of germ cells. The
cavity is lined by non-sexual mesodermal cells which have migrated in from
the mesonephros. Practically all the germ cells in this type of larvae gonad are
maturating. Animal 70 to 95 mm. total length.

35 Section through male gonad of second-year tadpole approaching meta-
morphosis (text fig. 1, C). The cavity has disappeared. Note the testicular
ampullae. The second sexual cycle occurs in this type of testis. Animals 120
to 150 mm. total length.

308

GERM CELLS OF ANURANS PLATE 4

WILBUR WILLIS SWINGLE

309

PLATE 5
EXPLANATION OF FIGURES

Microphotographs made at a magnification of 1500 diameters. No reduction.

36 Amphitene nuclei. Note the thick pachytene thread at one pole and the
fine leptotene filaments at the opposite pole. Synapsis occurs at this stage.

37 and 38. Amphitene nuclei from first-year larvae. No effort has been made
to show the cytoplasm.

39 Pachytene nuclei. Synapsis complete. The threads appear single.

40 Transition stage from pachytene to diplotene nuclei. The thick pachy-
tene threads of figure 39 are in figure 40 sphttion into two thin threads.

41 Diplotene stages showing the disjunction of the leptotene threads paired
in figure 39. Note at X the split in the thick thread. This is the primary longi-
tudinal split.

42 Diplotene stage showing splitting of the thick pachytene threads.

43 and 44 Diplotene stages showing the figure-8 configuration of the splitting
threads.

310

PLATE 5

GERM CELLS OF ANURANS
WILBUR WILLIS SWINGLE

‘=
S

311

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

PLATE 6
EXPLANATION OF FIGURES

45 The large cell near the upper central edge of the photograph is a stage
in middle diplotene, showing the disjunction of the homologous chromosomes
in the shape of fine threads. The threads, it will be observed, are united at
their ends. Several pachytene nuclei are visible at the lower left-hand corner.
In the lower right-hand corner are shown portions of three ring tetrads.

46 Gigantic diplotene cell. This type of cell is very common; the size of the
nuclei is remarkable in some instances. Note the split threads extending across
the nucleus, but united at the ends by two dark staining knobs.

47 Portion of a diplotene nucleus together with two early preleptotene nuclei.
Note the size difference.

48 Diplotene nucleus.

49 Part of a diplotene nucleus, showing splitting of the threads.

50 and 51 Large diplotene nuclei. Larva of 80 mm.

312

GERM CELLS OF ANURANS PLATE 6
WILBUR WILLIS SWINGLE

PLATE 7
EXPLANATION OF FIGURES

52 to 54 Portions of large diplotene nuclei, showing condensation and great
thickening of the threads to form the tetrads. Note the size of the chromosomes.
All are cut in these photographs; the size of the nuclei is so great that each one of
these cells is cut into three parts when sectioned at a thickness of 8to10u. In
figure 54 note that the condensing chromosomes are split again—equational
split.

55 and 56 Portions of large cells in diakinesis. The ring tetrads are shown.

57 Large diplotene nuclei. Nucleus almost completely fills cytoplasm.

314

GERM CELLS OF ANURANS PLATE 7
WILBUR WILLIS SWINGLE

315

PLATE 8
EXPLANATION OF FIGURES

58 Portion of large nucleus in diakinesis stage. Practically no cytoplasm
surrounding nucleus.

59 to 63. Smaller cells in diakinesis. Note figure-8 chromosomes.

62 Portion of large nucleus in diakinesis. Note the extraordinary large
figure 8 in this cell.

64 Portion of large spermatocyte, showing ring tetrads.

65 and 66 Portions of spermatocytes showing chromosomes.

67 Same cell as figure 64 photographed at a different focus. Note follicle
cells and nuclear size. Compare these cells with those of plate 15.

316

GERM CELLS OF ANURANS PLATE 8
WILBUR WILLIS SWINGLE

317

PLATE 9
EXPLANATION OF FIGURES

68 Portion of large spermatocyte nucleus.

67 to 76 Ring tetrads of first-year larval spermatocytes. Note the large
open rings and especially the size of certain of the tratrads. In adult frogs the
tetrads on the spindle are not in the form of open rings of this type, but resemble
dumbbells and are very much smaller in size. Compare with plate 15.

318

GERM CELLS OF ANURANS PLATE 9
WILBUR WILLIS SWINGLE

PLATE 10
EXPLANATION OF FIGURES

77 Large first spermatocyte in act of division. Note the size of the tetrads
and spindle. These cells degenerate in the act of dividing at this stage.

78 to 80 Large spermatocytes of first-year larvae in act of division. Figure
79 shows a tripolar spindle and irregular arrangement of the tetrads.

8l and 82 Larval spermatocytes. Figure 82 is of a cell in process of degener-
ation. The tetrads have lost their annular appearance.

320

GERM CELLS OF ANURANS PLATE 10
WILBUR WILLIS SWINGLE

321

PLATE 11
EXPLANATION OF FIGURES

83 Giant spermatocyte in division. Note size of the cell and its follicle.

84 <A portion of large spermatocyte preparing for division. All such cells
degenerate before completing the process.

85 to 88 Large larval spermatocytes preparing for division. All degenerate
before completing the process.

89 and 90 Are isolated tetrads of larval germ cells in similar stages of devel-
opment. Note variation in size of the tetrads in these cells. Figure 90 is from
animal in second year. It is rare to find this type of tetrad in second-year larvae.
Compare with plate 15.

GERM CELLS OF ANURANS PLATE 11

WILBUR WILLIS SWINGLE

PLATE 12
EXPLANATION OF FIGURES

91 to 94 Larval spermatocytes dividing.

95 Single large ring tetrad. Note the size and the lugs marking the synaptic
ends of the homologous chromosomes.

96 to 99 Larval spermatocytes dividing abnormally. Note in figure 97 the
great size and thickness of the central chromosome. This tetrad is cut in half
longitudinally.

100 and 101 Abnormal spermatocyte. Many tiny asters scattered through
the cytoplasm.

102. Large ring tetrad from second-year larvae. Just beneath is a small
germ cell of the second year; note the size difference. The spermatocyte in
which this tetrad was photographed is not shown here and was in early stages of
degeneration. This type of tetrad rare in second-year tadpoles and found only
in degenerating cells, which persist from first sexual cycle.

324

GERM CELLS OF ANURANS PLATE 12
WILBUR WILLIS SWINGLE

PLATE 13
EXPLANATION OF FIGURES

103. Larval spermatocyte in early stages of degeneration. Tetrads condens-
ing or running together.

104 and 105. Further stages in spermatocyte degeneration.

106 to 108 End stages of degeneration of larval spermatocytes. The type
of cell depicted in figures 106 and 108 sometimes show long axial filaments as out-
growths of the centrosome. The black bails are the remains of the tetrads.

109 and 110 Stages in the degeneration of the larval spermatocytes of the first
sexual cycle. Note in figure 109 the condensed group of vacuoles in the large
clear area. The clear area represents the original size of the cell; the vacuoles
are the remains of the tetrads that went to pieces in situ on the first maturation
spindle.

326

GERM CELLS OF ANURANS PLATE 13
WILBUR WILLIS SWINGLE

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

PLATE 14°
EXPLANATION OF FIGURES

111 So-called ‘oocyte’ in larval testis. The character and history of these
eelis will be discussed in another communication.

112 Giant larval spermatocyte, showing size of the spindle.

113 Giant spermatogonium from testis of first-vear larvae.

114 Large spermatocyte dividing. Only half of the cell is shown. Note
the extreme length of spindle.

115 Spermatogonium showing polymorphic nucleus. At left of picture is
an abnormal spermatocyte division. The chromosomes are torn to pieces by
polyasters in the cytoplasm.

116 Gigantic spermatocyte dividing, showing tripolar spindle. Note distri-
bution of chromosomes on the two spindles.

328

GERM CELLS OF ANURANS PLATE 14
WILBUR WILLIS SWINGLE

329

PLATE 15
EXPLANATION OF FIGURES

117. Primordial spermatogonium with large vesicular nucleus. This type of
cell, in very small numbers, persists unchanged through the first-year maturation
cycle of the larvae, and by repeated divisions probably contributes the cells of
the second sexual cycle of the larvae.

118 and 119 Small germ cells of second-year tadpoles preparing for the second
sexual cycle. Compare these cells with figure 117. Both types of cell are found
in the same gonad. The smaller elements appear to originate in part from
transformed mesothelial cells (?).

120 to 123. First spermatocyte prophases. Note the small size of the cells
and tetrads. Compare with figures 112 or 114 or any cell on plates 7 to 13.

124t0128 First spermatocyte divisions of second larval sexual cycle, or newly
metamorphosed frogs. Note extremely small size of cells and tetrads. Compare
with plates 7 to 13. Clearness of detail of the chromosomes has been sacrificed
by overdevelopment to show the spindles and cell outlines.

129 and 130 Spermatids of second sexual cycle.

131 Spermatozoa of second larval sexual cycle. These are sometimes formed
in large numbers in the larvae, but are more numerous shortly after metamor-
phosis.

330

GERM CELLS OF ANURANS PLATE 15
7 S SWINGLE

Resumen por los autores, John A. Detlefsen y Elmer Roberts,
Illinois Agricultural Experiment Station.

Estudios sobre el crossing over.
I. El efecto de la seleecién sobre los valores del crossing over.

El tanto por ciento de cross overs de la combinacion de ojos
blancos y ala en miniatura de Drosophila melanogaster es pr6xi-
mamente 33, y la “distancia de mapa’”’ hallada por Morgan y
Bridges es préximamente 36. La seleccién de hembras que
presentaban valores bajos en el crossing over redujo el tanto
por ciento de estos Ultimos casi a O en la serie A y su derivada
serie A’. Estas dos series reprodujeron este valor reducido del
crossing over durante tres y nueve generaciones, respectiva-
mente. En otra serie independiente B, el valor del crossing over
se redujo a 5 o 6 por ciento en 28 generaciones y el tronco selec-
cionado ha continuado produciendo dicho crossing over, sin
mas seleccién, durante 22 generaciones. La serie C, con crossing
over alto, no produjo por seleccién un aumento en los valores de
crossing over, pero produjo en la F, nueve pares que sumaban
26 cross overs; 1055 individuos—2.46 por ciento de crossing over.

El crossing over es por consiguiente muy variable y manifiesta
los efectos de la seleccién. La seleccién de valores bajos ha
eliminado practicamente el crossing over en las series A y A’,
y le ha reducido considerablemente en al serie B. La seleccién
de un valor elevado no ha aumentado los: valores de crossing
over en la serie C, pero probablemente ha producido mas crossing
over doble en algunas hembras, que resulta en una disminucién
del valor de dicho proceso con relacién a los dos genes escojidos.
Los autores proponen varias hipdtesis, pero parece sumamente
probable que los factores multiples regulan o por lo menos
influyen considerablemente sobre el crossing over. Si esta
explicacién es correcta, debemos modificar nuestra opinién sobre
el crossing over en relacién a las distancias que separan a los
genes.

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHORS’ ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 6

STUDIES ON CROSSING OVER

I. THE EFFECT OF SELECTION ON CROSSOVER VALUES!
J. A DETLEFSEN AND E. ROBERTS?

TWO TEXT FIGURES

CONTENTS
LEN TTEYC HU ELENOTT AE ee sai ie Seen eich co etre etnias 4 nial alah hl NBR ke Bich gS 330
MiaterralssancemethodSeiistatates. hice crin Seoeans aoe aes secur ek nels eee meas 335
su eR artamepe et te ts He She to eet ter chine ki Eh Ce EEy Ry Oe atee wince Rh ree teas 338
SenlespAc slows SelechiOmye sie sue Ney cs, ites tee et ene Osuna Scheer te tasaroee 338
Series A’; low selection; derived from series A................-..-06-- 341
Senles* lows selechlom seas cm occ toe oe oe cee termes wi ot aebeteteucls where ara ete 343
Senies? CS hight selections sy4ae cco ge tena eche ads eee nee Saas cabs tote 348
DISCHSSTOMEAT CAS UNIT ATE ier crus tae en ate cohort ORIN aie receded cya cheolep ele awe 349
PT RESUME CROLL Clic oe. ON gy ate a cera stoic s SeSteve Ie Me eRe ed gave ae cal Pcs w whe cavede eyes 304
INTRODUCTION

‘The experiments described in this paper were undertaken in
an effort to answer the question: can the percentage of crossing
over be modified by selection. The significance of an answer
(either affirmative or negative) and its relation to our present
concepts on crossing over, locus, chromosome mapping, etc.,
were apparent and seemed to justify the time and attention
necessary to carry these investigations over so long a period.
They were planned and begun in February, 1916, and have
involved the classification of over 300,000 individuals. Had
the results been negative, the experiment would have been

1 Paper no. 12, from the Laboratory of Genetics, Illinois Agricultural Experi-
ment Station.

2 We wish to give credit to the following graduate students for aid in these
investigations: Mr. A. T. Fishman carried series A for three generations; Mr. L.
E. Thorne earried series C for seven generations. The war called both men
from their work. The late Prof. B. O. Severson carried series B from the begin-
ning to the Fy, generation. In the death of Professor Severson both genetics
and scientific agriculture have lost a capable and enthusiastic student and
investigator.

333

Spiel J. A DETLEFSEN AND E. ROBERTS

dropped much earlier, but the effects of selection were conspicu-
ous and prompted us to carry the experiments through to their
logical end.

Whenever one crosses an individual with the linked factors,
AB, to the double recessive, ab, then the heterozygote, ABab,
will form four sorts of gametes: AB + ab, the parental types,
and aB + Ab, the recombinations or crossovers. The relative
frequencies of these gametes will depend upon the distance
between the loci for A and B, at least according to the commonly
accepted hypothesis. If a distance on the chromosome, which
gives 1 per cent of crossovers is adopted as an arbitrary unit,
then the distance between genes on a chromosome may be deter-
mined in terms of this arbitrary unit, and the map of a chromo-
some may be plotted, as has been done by investigators working
with Drosophila. Repeated trials using large numbers, with
comparable stocks and controlled environmental conditions,
have shown that the ratio of crossovers to total gametes is uni-
form enough to suggest that the distance between two genes is
fairly constant. However, the phenomenon of crossing over is
not as simple as was first supposed, for a number of genetic and
environmental influences have been found to affect crossing over
markedly, at least in Drosophila melanogaster. Bridges (715)
stated that crossing over varied with age, for second broods
showed a rather consistent decrease. Plough (’17) found that
low and high temperatures (below 17.5°C.,. and above 28°C.)
increase the amount of crossing over. Sturtevant (719)
found in the second chromosome of Drosophila one gene to the
left of purple and one to the right, both of which lower the per-
centage of crossing over in that portion of the chromosome in
which they lie. He also found a similar factor in the third chro-
mosome. Furthermore, an incompletely investigated case dis-
closed a dominant third chromosome gene which increased the
amount of crossing over between purple and curved-in the second
chromosome. Gowen (’19) measured the amount of variability
shown in a population of 240 Drosophila females with respect to
crossing over between fixed points in the third chromosomes
and found a very high degree of variability. His data show that

EFFECT OF SELECTION ON CROSSOVER VALUES 330

a change in genes between two or more fixed points may be
accompanied by a slight disturbance of the crossing-over ratios
between these fixed points. Sex, to be sure, has a striking effect
on crossing over, for the male Drosophila does not show this
phenomenon even in the autosomes.

Whenever one observes a large number of Drosophila females

=, it is common to
find much variability with respect to the amount of crossing
over, even though the cultures are kept at the usual normal
temperature and no striking genetic modifiers of crossing over
are known to exist. Just what this variability is due to is not
known. Some of it may represent fluctuations of sampling and
some of it may be due to age, but very frequently the devia-
tions are so wide as to arouse a suspicion that hitherto unknown
causes may be effective. If this variability is due, at least in
part, to genetic causes, then selection should have an effect,
particularly if environmental fluctuations do not mask or obliter-
ate the effect of genetic modifiers. It was with this thought in
mind that the senior writer began to select for high and low
crossover values.

: A
of the generalized zygotic formula aie

MATERIALS AND METHODS

The selection experiments consisted of four series:
Series A, low selection;
Series A’, derived from series A in F;;
Series B, low selection, a second experiment duplicating series A;
Series C, high selection.

Each series began with a single white-eyed miniature-winged
female mated to a wild red long male. These strains were
chosen because the characters are easily recognized, show little
or no variability, and have at least fair viability. To classify
any female with respect to her ‘crossover capacity’ requires the
classification of all the progeny which we can obtain from her.
Thus in F,, series A (table 1), we classified 8660 offspring to
obtain the necessary data on fifty-six F; females for the purpose
of selection. In the usual selection experiment, individuals are

336 J. A DETLEFSEN AND E. ROBERTS

chosen on the basis of external characters which can be deter-
mined by direct observation. To select for high and low cross-
over values is rather more tedious and time exacting because
the individual cannot be classified directly with respect to its
crossover potentiality. Its character is disclosed only after
obtaining a reasonably large progeny. Characters which could
be recognized easily and classified rapidly and accurately were
indispensable. The two allelomorphic pairs, white eye vs.
red eye, and miniature wing vs. long wing seemed to fulfill these
conditions, and they have the added advantage of giving a large,
initial, normal percentage of crossovers (about 33 per cent),
which means that variations are thus more readily detected.

The procedure followed in the low-selection series A is typical
of all the series and can be taken as a sample. A single white
miniature female mated to a red long male gave F; white minia-
ture males, w m, and long red females ae = the latter
being double heterozygotes. The F, sibs were mated in pairs
in 8-drachm homeopathic vials, and the pairs were removed to
new vials about every three days. The culture methods were
those commonly used with Drosophila. The F, offspring from
each vial were classified daily until a fair sample of each F,
female’s ‘crossover capacity’ was obtained. As expected, the
offspring were of four kinds: the parental types, red long, and
white minature and the crossovers red miniature and white
long. It was impossible to anticipate which F; female was going
to be selected because of her low crossover ratio determined by a
reasonably large progeny, and it was likewise virtually impos-
sible to continue mating in pairs the sibs from each F; female
until we could find out which line was going to be used to con-
tinue the selection. Thus in F, series A, low selection (table 1),
there were twenty-eight pairs of F; individuals, several of which
appeared to be promising material, but we eventually chose
pair 15, which gave 21:98 = 21.43 per cent. By the time we

’In giving crossover values we shall put the data in the following order
throughout this paper: crossovers: total-per cent of crossing over. The classes
are always the same and repetition can thus be avoided.

EFFECT OF SELECTION ON CROSSOVER VALUES rent

were in position to know that pair 15 would be selected, prac-
tically all of its F, offspring had emerged. ‘Therefore it seemed
expedient to mate en masse the F, offspring (i.e., red long females
W m
Ww WM’
promising pairs, to perpetuate the promising lines. Hence, in
table 1, the F, sibs came from the selected F; pair, no. 15, and
were mated en masse, giving 25.46 per cent crossovers. The F;
offspring were then mated in pairs, and selection was again
exercised. This means that an odd-numbered generation (Fj,
F,, etc.) in table 1, for example, represents the mating of pairs,
while an even-numbered generation represents the mating en
masse of sibs from the selected pair. It will be clear that inbreed-
ing was very intense throughout all series, for the pair gave sibs,
and the sibs from the selected pair mated en masse gave a popu-
lation in which the most remote relationship could be double
cousins, but it might be as close as sibs again. Thus we had
alternate generations of double cousins (or nearer relatives)
mated in pairs of which we selected the offspring (a sibship) of
the most promising pair to mate en masse. Selection therefore
really took place in alternate generations. While we recognized
that this procedure was not ideal theoretically, at least from the
point of view of a strict selection experiment, the advantages out-
weighed the disadvantages, inasmuch as it made the whole selec-
tion experiment possible in a practical sense and yet maintained
inbreeding. The chief disadvantage lies in the fact that this
method precludes calculation of the parent-offspring correlation
and regression coefficient for any two successive generations.
In all these selection experiments, after the P; generation, all

miniature white males w m) from each of several

C WwW m
of the matings were of the type W M
long females heterozygous in white miniature mated to white
miniature males, except where special tests were made for the
sake of genetic analysis. This type of mating gave crossovers
among the offspring of both sexes and thus a more effective
criterion for selection, since numbers were doubled. It also
gave the doubly heterozygous females and the ultimate recessive

SW. m; 1.e., red

338 J. A DETLEFSEN AND E. ROBERTS

males as two of the four most frequent classes, which was very
convenient, since these were used again for mating in the next
generation. Writing the form of all matings for every genera-
tion in tables | to 6 in the usual Mendelian terms we have:

w m wm
WM
red long 9 white miniature @

w m wm Wee a Warn
wm Ww mM W m w M

ae ate

white miniature 2 + red long 9 + red miniature @ + white long ?
ote ae
AP +

Ww M

red long &

wm W m + w M

white miniature red miniature o + white long

—

= <a —V

Non-crossovers Crossovers

THE DATA

Series A; low selection

Table 1 and text figure 1 give the main facts of this selection
experiment. The F, generation consisted of twenty-eight pairs
whose total progeny showed 27.11 per cent crossovers. This is
a little lower than might be expected in a general population, but
the difference between this ratio and Sturtevant’s (given in
Morgan and Bridges, ’16) ratio of 32.8, based on 41,034 progeny,
is no greater than that recorded by Bridges (Morgan and Bridges,
16), who gives data showing 38.3 per cent crossover. This
same stock has repeatedly given crossover ratios close to 33
per cent.

The crossover values in this and other similar tables are
treated as variables and classified in frequency distributions in
which the class interval is 3 per cent. The average of each class
is placed at the head of the columns, e.g., 1.5, 4.5, 7.5, ete., which
means that the class ranges were 0-3, 3-6, 6-9, and so forth.
The crossover values in the F; ranged from 10 per cent to 36.8
per cent. ‘There is no doubt but that some of these ratios have
little meaning, for they are based upon small totals. We have

EFFECT OF SELECTION ON CROSSOVER VALUES 339

TABLE 1
Series A: low selection

E
» | THE DISTRIBUTION OF CROSSOVER VALUES a aa
= IN EACH GENERATION = a 8
< < ah
Zz 3 an > Po | THE SELECTED
Cc B & rs PAIR GAVE
e |g : . | ee
al a g 4 Smulrae
F, | 28 TU) TY) Ay a Gh e3] hl a ay 4271 ,575,27 .11| 56.3,21: 98=21.43
F. x 97| 381/25 .46
1a || Ae 18) a 5| 8] 4] 5] 1] 2} |1 ,087|8 ,760)27 .58/139 .3 34:189=17 .99
F, x 206) 787|26.18
F; | 56) 1 1 3| 7| 9} 5| 9/11} 4] 2} 2} 2/2 ,188/8 660/25 .27|154 .6)44:210=20.95
F x 197} 920,21.41
F, | 45/10 if LG Aa 7) Ay 2) 2 905/4 ,234/21 .37| 94.1) 9:104= 8.65
Fs x 125} 612)/20.42
Fy | 438/74) 1) 2 3] 5} 4) 7| 3) 4 478/2 ,899 16.49) 67.4) 0:131= 0.00
Fio x 0; 53) 0.00
Fi 5| 6 0} 87} 0.00} 17.4) 0: 12= 0.00
Fie x 2| 148) 1235

Per cent of crossovers
e)
28
26
24
22
20
18
16
14
12

Series B

1 3 5) 7 g ata 13 15 Ati 19 21 23 P43) 27 2

Generations

Fig. 1 Series A, A’, and B, low selection

340 J. A DETLEFSEN AND E. ROBERTS

recourse to at least two methods of dealing with such unpro-
ductive pairs. We can either include in our frequency distri-
bution only those females on which we have ample data to give
a somewhat reliable crossover value and ignore all pairs giving
less offspring than a fixed minimum (fifty individuals for,
example), or we can simply include all females and thus withhold
no data. The latter course seemed preferable and we have
followed it. There were five pairs showing lower crossover
values than the one we selected, as follows: 10.0; 12.5; 16.0;
16.6; 20.7. We did not always select the lowest absolute value,
for in many cases this was based upon an insufficient number of
offspring. It was also necessary to keep fertility in mind, in
order to insure the perpetuation of our selected line. This
explanation will make clear why we could not always choose
the lowest absolute crossover value in the frequency distribution
of any given generation. In table 1 the italicized frequency in
each distribution shows the relative position and value of the
selected pair. No dispersion can be given for the Fs, Fu, Fe,
etc., since these represent en-masse matings. An x represents
the point to which the progeny of the pair selected in the preced-
ing generation regressed. The average number of offspring per
pair shows how reliable the crossover values usually were in this
experiment. The crossovers, total, and the crossover value for
each selected pair are also given in the last column. Those
generations which have any number of pairs entered under that
heading are generations in which all matings consisted of pairs,
while the other alternating generations were en-masse matings.
Since the crossover value of a female may be based upon a small
number of offspring in some cases, and thus give an apparently
wide deviation which has little significance, we have not calcu-
lated the variability of each generation in this paper. For
example, a female showing a crossover value of 10 per cent based
upon twenty offspring might well show 30 per cent if one hundied
and fifty offspring had been secured, since age and changing
temperature affect crossing over; or she might even show 10 per
cent based upon twenty offspring as a sheer fluctuation of
sampling.

EFFECT OF SELECTION ON CROSSOVER VALUES 341

The first two selections seemed to show little or no effect.
Although the values of the selected pairs were low, their progeny
regressed practically to the parental average. Possibly this
means that all wide deviations were not necessarily due to
genetic causes and that we had difficulty in distinguishing
between wide environmental variates and wide variations due to
genetic causes. Selection was thus effective only when by chance
we chose a wide variate due to the latter set of causes. For
example, in the F;, we chose a female showing 17.99 per cent cross-
overs, but her progeny gave an average of 26.18 per cent. After
the F;, progress was very rapid. The F, gave 16.49 per cent, and
the Fi, to Fi; gave about 0 per cent. These last generations
in this series were based upon small totals, because the excessive
heat (90° to 100°F., day and night) for long-continued periods
reduced fertility to a minimum and eventually annihilated our
stock in this one. However, series A’, which was derived
directly from series A, gave just as low crossover values with
larger numbers and under better conditions. We may be quite
sure that temperature was not the cause of low crossing over;
for, if we may anticipate, series B showed effects of selection
under normal temperature conditions.

Series A’, low selection; derived from series A

In the F; generation of series A, two selections were made.
One female (2 14) gave 9:104 = 8.65 per cent, and a second
female (2 10) gave 1:91 = 1.10 per cent. The former was
used to continue series A, while the latter was used to begin a
new series, A’. Table 2 and text-figure 1 give the main facts
pertaining to series, A’. We began this series to insure keeping
alive some of the low crossover material of series A during con-
tinuously hot weather. Our facilities did not permit controlling
temperature, and the whole experiment was in a precarious
situation during the early summer months of 1916. We found
that mating a number of females en-masse assured more progeny
than the same number of females mated in individual bottles—
evidently because the larger number of larvae carried the yeast
through the culture and kept molds down. Hence, during the

342 J. A DETLEFSEN AND E. ROBERTS

summer months of 1916, we made numerous en-masse matings
in this series to insure keeping the stock alive. Beginning with
a single F; pair of series A showing 1:91 = 1.10 per cent, the
new series A’ was run for nine generations. All generations
were en-masse matings except F, and Fy, in which paired mat-
ings were made to ascertain what the crossover values of the
individual females might be in this line. In the F, the average
crossover value for the total population was 8:397 = 2.02 per
cent. The wide dispersion in this generation does not carry

TABLE 2
Series A’: derived from series A

THE DISTRIBUTION OF CROSSOVER
GENERA- NUMBEE) VALUES IN EACH GENERATION cet ere || Are CROSSOVER
Paes 4.5 @.5 | 16-5 | 13:5) 16:5 | 19:5

I, 1 1 91 1.10
Fs x 1 86 1.16
Fo 18 iA | 1 1 8 397 2.02
Fio x 0) 61 0.00
Fi x 0 133 0 00
Fis eel 4 373 1.07
Fi3 x 9 1,473 0.61
Fis 25) 25 | 10 2p25e 0.44
Fi; xX 0 289 0.00

EIR te Fee Rs Sots a ae od Sere IE AR Sin 33 5 ,156 0).64
1See text.

much weight because cultural conditions were poor and fer-
tility was low. Pair no. 4, for example, gave 3:15 = 20 per cent,
but such a pair might well give a much lower crossover value
with a larger number of offspring. The Fi; gave 10:2253 = 0.44
per cent, and the numbers are large enough to be significant.
This generation included twenty-five pairs which gave a total of
2:977 = 0.20 per cent, and an en-masse mating which gave 8:1276
= 0.63 per cent. There can be no doubt but that an original
crossover value of 33 per cent has been changed by selection, at
least, that a marked change has followed selection. For nine
generations the stock bred true to about 0 per cent crossover.
The totals for series A’ were 33:5156 = 0.64 per cent.

EFFECT OF SELECTION ON CROSSOVER VALUES 343

Series A’, like series A, was eventually lost in the latter part
of the summer of 1916 because of an unavoidable succession of
events. We regretted the loss of this stock because we had
hoped to make a genetic analysis of the last generations in an
attempt to learn what was taking place during selection. How-
ever, the data as they stand indicate that crossing over is not a
very stable phenomenon and that it can be rather easily modi-
fied. We surely cannot concur in Morgan’s (’19) view that
crossing over ‘‘gives numerical results of extraordinary con-
stancy.”

We immediately began a new selection experiment, hoping
that we could duplicate the results of series A and A’.

Series B; low selection

Series B, like the preceding series A and A’, began with the
mating of a single white miniature female and a wild red long
male. In fact, as a prelude to series B, we made eighty such
paired matings, for we had found some non-disjunction in our
original stocks and in series A and A’. Since non-disjunction
theoretically lowers the percentage of crossing over (Bridges, ’16),
we wished to assure ourselves, if possible, that this cause might not
be operative in producing low crossover values in our selection
experiment. Of the eighty white miniature females tested we
found eleven giving either matriclinous daughters or patri-
clinous sons or both. This must mean secondary non-disjunc-
tion in the white miniature stock, for the exceptions were -too
numerous to be considered primary. We chose white miniature
2 53 mated to a wild male as the foundation pair for our experi-
ment, because this pair gave fifty-two wild-type daughters and
seventy-eight miniature white sons. While they showed no
exceptions, it does not prove that 2 53 may not have been non-
disjunctional (XXY), for a ratio of 0: (52 + 78) might well
occur as a chance ratio where an average of 4.3 per cent of excep-
tions is expected from XXY females (Bridges, ’16). However,
in the present paper we are concerned only with the question
whether selection based on variable crossover ratios can be effec-

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 2

344 J. A DETLEFSEN AND E. ROBERTS

tive. Whether non-disjunction has any necessary relationship
to our result will be discussed in a subsequent paper.

TABLE 3

Series B: low selection

nm | THE DISTRIBUTION OF CROSSOVER VALUES
5 IN EACH GENERATION
z 3 2D |UD | ad | 2D | Ad | 1D] 29 | 29 | ag | 29 | 19 | 20 | ap | 19 °
| at ES) Sah Sh Se Pa ee ess Sai) iS)
F, 34 3] 5} 3/10) 6} 3} 2| 2'2 ,056
F, 1 129
F3 | 47 HUE 1) 9] 5} 8} 7] 6} 6} 2) 1/2 ,204
F, 1 330
1B Hokey al) al at Al19/17| 5/19}11) 2 5,798
F, itil al 379
Bie | ola: 1} 7) 4) 4/18/18) 9) 8)-1) 1) |5,490
Fs iM Ulett) I 650
Fy | 88 1|12/12|13|17|13} 9| 3] 2 1/3 ,230
Fio i) SAP PAN AL 595
Fu | 84 8) 6)13/24/14/10)10) 3) 1 | {1,896
Fie 8) 3) 1) 1 815
Fi3 | 63} 3} 9/13|16)10} 6) 5 759
Fiy Lies 1 1 149
Fis 1 68
Fig | 79} 5) 7/22)29)12) 2) 1) 1 1,416
Fy7 6| 1 642
Fig | 74) 6} 5/28/22) 9) 1) 2) 1 1,195
Fi5 3] 9] 4| 2 975
Feo | 52] 4} 5)12|16) 9} 5} 1 641
Fo | * 1) 5) 6 1,108
Foo | 67) 3) 6)15/22/12) 7) 2 1232
Fo3 il Ap 4 1 392
Fes | 47/10)13)16) 5) 3 431
Fo; TNT e/a aL 223
Foe | 38} 5)13)12) 6} 1 1 280
Foz Le 65
Fes | 46/10) 8/12] 7] 6) 3 215
Fag Q| 10

ATION

Totals

7,189
» 530
8 ,089
1,141
23 ,618
1,411
21,974
2,858
17 ,550
3,502
11 937
4558
7,439
790
609

14 ,765
6 027
12 166
8 ,786
6 533
11,407
1 717
3,649
6,801
3,119
3,903
1,152
2 650
158

THE TOTAL POPULA-
TION IN EACH GENER-

Crossover
values

28 .60
24.34
25 .02
28 .92
24.55
26.86
24.98
22 74
18.40
16.99
15.88
17.88
10.20
18.86
7
9.59
10.65
9.82
11.10
9.81
9.71
10.51
10.74
6.34
7.15
rds
5.64
8.11

AVERAGE NUMBER OF
OFFSPRING PER PAIR

L749)
Os
144.7) 0:
39:
OPAs7/\) “We

43:

6.33

211.4) 61:

THE SELECTED PAIR
OR SELECTED EN
MASSE GAVE

219 2a
129:
N74. Na br Aaye
330:1141 =28 .92
288.0} 48:
296 :1130=26.19
360.0) 36:
217:1185=18 .31
Pea! ities
191:1179=16.20
LADEN eo:

a i tala
PLSE A abe
Do:
68:
186.9} 12:
122:1009=12 .09
164.4] 4:
47:
125. 6|) 20:
199:2027= 9.82

580= 24 .34
645 = 27.13

230= 20.87

251= 14.34

276= 6.16

96= 5.20
512=12.89
197= 7.61
207 = 12.08
609=11.17
169= 7.10

77= 5.19
464=10.13
179=11.17

63= 1.59
768= 6.64
107= 0.00
454= 8.59
141= 4.26
873= 4.93
39= 0.00

: 129= 6.20

The F,; sibs from @ 53 were mated in thirty-four pairs and
gave as a whole a crossover value of 28.60 per cent (table 3).

EFFECT OF SELECTION ON CROSSOVER VALUES 345

In order to further test our foundation stock, the F, offspring of
2 25 (one of the eighty Pi ? 9, and similar to 2 53) were
tested en-masse and gave 1142: 3553 = 32.14 per cent. The F;
offspring of several other F; 2 2 were mated en-masse and
gave 830: 2923 = 28.40 per cent. All of these facts indicate
that our foundation stock was quite normal with respect to
crossing over and gave crossover values of the same general
magnitude as those ordinarily used in plotting maps of the sex
chromosome. .

The main facts pertaining to series B are given in text figures
1 and 2 and in tables 3 and 4. Table 3 was constructed in the

Per cent of crossovers

§ Series B

SU aliSdow So RST DGD ere Al SaqreeSinh “40s 249
Generations

Fig. 2. Series B, continued, low selection

same way as table 1, with the following exception: in series B,
records of the en-masse matings of the offspring from several
promising pairs were kept and the crossover values of all these
are put in the form of a frequency distribution, but the italicized
frequency shows the position of the en-masse mating which was
derived from the pair eventually selected to continue the experi-
ment. The italicized frequency in the distribution of the pairs
likewise shows what the value of the chosen pair was. The
first three columns at the right of table 3 give the data for the
total population in each generation.’ The average number of
offspring per pair shows that the fertility was high and selection
was based upon what seemed to be adequate numbers. © The last
column gives the number of crossovers total, and crossover value

346 J. A DETLEFSEN AND E. ROBERTS

of the selected pair in each generation and the same data for
en-masse matings from these selected pairs. Text figures 1 and
2 give a graphic representation of the progress made in series B.
The graphs are based upon the crossover values in the selected
line; i.e., ali en-masse matings except the selected one have been
neglected in plotting the graph. In other words, the graph
relates only to the actual line of selection, and all side lines have
no weight in determining the coordinates. It will be clear that
those generations in table 3 which have any number of pairs
entered under that heading were generations made up entirely
of paired matings, while all other generations were en-masse
matings. ;

The first three selections had little or no effect, but it cannot
be said that selection was very rigid during these generations.
In F, we selected a pair giving 36:251 = 14.34 per cent and
made some progress, for the next seven generations (F;—Fy,)
fluctuated between 10 per cent and 23 per éent. The subse-
quent nine generations (I'j;—F.3) fluctuated around 10 per cent.
Selection was carried on up to Fs, and the last six generations
(F..-F2,) varied around 6 per cent. After that we simply
carried the stock without selection, and have found it to breed
quite true to low crossover for twenty-two generations. The
F.5—-F 59 have given values around 6 per cent. These last twenty-
one generations are shown in table 4.

There are some features of tables 3 and 4 which require com-
ment for the sake of clearness. Temperature conditions made
it necessary to breed the offspring of the selected pair in the Fi;
for two generations by the use of en-masse matings. Hence,
the matings in the F,, and F,; show no pairs and selection was
interrupted. This was the only case in which the usual sequence
of selecting in alternate generations was not followed. The Fs;
showed a rather abrupt rise in crossover value (12.50 per cent),

4 An independent mutation of gray to yellow which occurred in the F2; should
perhaps be put on record. One female (9 no. 30) proved to be heterozygous for
yellow, and this gene was linked to white and miniature. Hence the mutating
gene came through the spermatozoon from the gray white miniature father of

2 30. This new gene for yellow proved to be identical with the original yellow
mutation found by Wallace in 1911 (Morgan and Bridges, 716).

EFFECT OF SELECTION ON CROSSOVER VALUES 347

which was without doubt due to high temperature, as our records
indicate. The fertility was low and we obtained with much
effort from en-masse matings in the F3, and F3; only forty-eight
and eighty individuals, respectively, while under ordinary con-
ditions several thousand would have been possible. As soon as
normal conditions were restored, the usual low crossover values
were again found. The Fu showed a rather unexpected rise

TABLE 4
Series B—Continued

GENERATION CROSSOVERS TOTALS CROSSOVER VALUES

Big ; 6 144 4.17
Fa Vf 171 4.09
Fis 4 48 8 33
ie 10 80 12.50
Lora 52 ‘ 643 8.09
Ris 48 1,147 4.18
Fe 55 1,032 5.33
By; 46 814 5.65
Bye 39 697 5 60
F 39 55 954 5 77
By, 72 1,074 6.70
Bi 94 1,015 9 26
Be 463 8 ,564 5.41
we AT 901 5 22
Ba 103." 1,312 7 85
Fis 43 661 6 51
By 59 992 5 95
Faz 69 1,021 6 76
ie 45 734 6.13
Ba 81 1,081 7 49

Fo 96 1,375 6.98

(9.26 per cent), but since there were no unusual temperature
conditions, we must regard this somewhat higher value as with-
out peculiar significance. The subsequent generations dropped
to about 6 per cent again.

348 J. A DETLEFSEN AND E. ROBERTS

Series C; high selection

In the F, generation of series A, pair number 21, was chosen
to begin a high-selection series, series C. While this series was
carried for only eight generations, and then discarded in order
to devote time to the other series, nevertheless brief mention
should be made because the results may aid us in interpreting
series A, A’, and B. We were not able to make progress in
selecting upward, as the averages of table 5 show. (Table 5
was constructed in the same way as tables 1, 2, and 3.) On
the contrary, we were much surprised to find that in the F,
generation a number of pairs suddenly dropped to very low

TABLE 5
Series C: high selection

GENERATION NUMBER OF PAIRS CROSSOVERS TOTALS CROSSOVER VALUES

Fi 28 427 1,575 27.11
F, 162 512 31.64
F; 35 1,407 4,842 29 .06
F, : 436° 1,355 Some
ia 90 6 465 21,071 30.68
Fe 684 2 267 30.17
F, 72 3,893 13 ,705 28.41
Fs 296 1,298 22 .80

crossover values; in fact, much lower ratios than one would
find in any ordinary population such as our original stocks or our
F, of table 1. The distribution of the F, in series C is given in
table 6. It will be noted that nine pairs gave values lower than
6 per cent. Their values were as follows: .

4:72 = 5.56
5:279 = 1.79
9:164 = 5.49
1:142 = 0.70
1:82 = 1.22
0:123 = 0.00
4:104 = 3:85
2:40 = 5.00
0:49 = 0.00
Total, 26:1055 = 2.46

EFFECT OF SELECTION ON CROSSOVER VALUES 349

There can be no doubt that these crossover values are signifi-
cantly different from any ratio in the F; in table 1, or from the
usual ratios shown by random stock females. Furtherrnore,
there is an interval of about 10 per cent between the lower and
higher groups of table 6, in which we found no crossover values.
The natural inference is that any attempt to increase the amount
of crossing over leads to double crossing over, and thus to very
low crossover values (practically zero). That is, these nine
females showed a marked decrease in crossover values, despite
high selection, because they gave almost nothing but double
crossovers. In other words, their low crossover values are,
after all, the result of:effective high selection. -Mr. L. E. Thorne,
who had this series under observation, was called into military
service and we did not make any further tests on this material.

TABLE 6
The distribution of crossover values in the F; generation of series C

THE DISTRIBUTION OF CROSSOVER VALUES

SSHOEAEGSTE: » | ec DM ie a SS ER ae ae
OF PAIRS | 19/19] 15 CROSSOVERS TOTAL
ai[ale

CROSSOVER
VALUE

72 5| 4 1 1| 6| 8/15/16} 8) 6} 1) 1) 3,893 13 ,705 28.41

We hope, however, to repeat the high-selection experiment and
test out the region between white and miniature in such females
which apparently give uniform double crossing over in a region
in which single crossing over is the rule.

DISCUSSION AND SUMMARY

As far as we are aware, there is only one record of a similar
selection experiment. Gowen (719) selected for high and low’
crossover values, but his results and conclusions are diametri-
cally opposed to ours, since: he found selection ineffective, and
concluded there were no differences in modifying factors for
crossing over in his experiment. He continued selection for
only five generations in the low series and six in the high, using
the region of the third chromosome between sepia and rough.

350 J. A DETLEFSEN AND E. ROBERTS

While it is possible that this chromosomal region may fail to
show the same phenomenon which we found in the sex chromo-
some, we are rather inclined to believe that the difference between
our results and Gowen’s is more likely due to differences in the
method of procedure, for Gowen states that his ‘“‘chief difficulty
lies in the few individuals that it was possible to include in a
given generation.”’ Gowen gives only the mean total crossing
over in each generation, and we do not know how rigid his selec-
tion may have been, for he does not state how many pairs were
included in each generation nor does he give the frequency dis-
tribution for crossover values. We suspect that he found the
same impediments in using strict brothet-and-sister matings
which we found and which prompted us to use en-masse matings
in alternate generations to increase our numbers. We are carry-
ing on selection experiments in other regions of the sex chromo-
some and in the autosomes, which should decide whether other
regions and chromosomes are similarly affected by selection.
We have no reason to suppose that they will not be.

The effects of selection upon crossover values may be due to
one or a number of causes, some of which suggest themselves
almost immediately. It would hardly be profitable to expatiate
on these, since we are making tests, which we hope may indicate
what has really happened in the course of selection. Briefly
stated, we think of the following possible causes which may have
been operative in modifying our crossover values:

1. We may perhaps have dropped out a large part of the
chromosome between white and miniature, thus bringing these
two genes closer together. We can probably disregard this as a
cause, for although ‘deficiency’ reduces crossing over (Bridges,
17) nevertheless the lethal action of deficiency would be seen in
a disturbed sex ratio. We found no such disturbance.

2. Is it possible that we may have moved the locus of the
genes on the chromosome? This would mean that the locus of a
gene is not permanently fixed, but that a given gene is found
in a characteristic position in the majority of cases. If we have
done this, and at the same time have not moved other genes,
then linkage tests should disclose this fact, for the order of the
genes would be changed.

EFFECT OF SELECTION ON CROSSOVER VALUES 351

3. In series A and A’ we found much evidence of non-disjunc-
tion. Bridges (’16) stated that XXY females should logically
show a decrease in crossing over, because heterosynapsis takes
place in about 16.5 per cent of the cases and precludes crossing
over in these cases. However, Bridges also showed that the
experimental results disagree with such an expectation, for cross-
ing over was not decreased among the regular sons of XXY
females, but as far as the evidence goes it was slightly increased.
For some time we labored under the impression that much, if
not all, of our decreased crossing over was associated with the
presence of non-disjunction (Detlefsen and Roberts, ’20). We
are now rather inclined to believe we were in error. It should
not be a difficult matter to free our low crossover stock in series
B from non-disjunction and thus dissociate this possible cause
from the others. We could in this way demonstrate that non-
disjunction was only accidentally present in our experimental
material and that our results are quite independent of non-dis-
junction.

4. Have we reduced the frequency of the usual single ‘chromo-
some twist’ between white and miniature to a minimum? Wein-
stein’s (718) results indicate that crossing over takes place in the
sex chromosome with about forty-six units as the modal distance
between successive crossovers. Similarly, Gowen (719) found
twenty-five to thirty units in the case of the third chromosome.
We began with two genes which were about thirty-three units
apart, and which should therefore show a-single crossover as the
characteristic or modal number. This would mean that in
series A, A’, and B we have practically eliminated the usual
single crossover in this region, while in series C we were on the
way to increasing it to two crossovers (i.e., a double crossover),
which would give us no crossing over as far as these two genes
were concerned. Does this mean that we can decrease or
increase the amount of ‘twisting’ which members of an homo-
logous pair of chromosomes may show, and which is supposed
to be responsible for crossing over according to the chiasmatiype
theory? If selection can accomplish this, then we may reason-
ably suppose that numerous hereditary modifying factors are

352 J. A DETLEFSEN AND E. ROBERTS

present in a general population and are the basis and cause of
this variable chiasmatype relationship. If this explanation is
correct (and we are inclined to believe it the most plausible one
of those we have suggested here), then we cannot escape a
marked change in our view-point on crossing over and related
phenomena. If, for example, all of the difference between prac-
tically no crossing over in our series A and A’ and normal cross-
ing over (33 per cent) is due to numerous modifying factors, then
we naturally begin to wonder just what part ‘distance between
two genes’ on a chromosome may play in determining linkage
values. Our current view is that ‘‘the percentage of cases in
which two linked genes separate (amount of crossing over between
them) is necessarily proportional, other things being equal, to
the distance between the genes,” (quoted directly from Weinstein
(’18)). » But evidently the percentage of crossing over is a vari-
able which is the expression of different possible combinations
of multiple modifying factors; hence the percentage of crossing
over cannot be proportional to the distance if the distance
remains uniform. For example, in series B we find 6 per cent
crossing over, and so we should conclude that the distance in
this stock is 2/11 or 18 per cent of what it was when we began
selection! Thus, to maintain our original position, we must
conclude that the percentage of crossing over and distance are
correlated variables, if the proportion between the two is to
remain reasonably constant. We then naturally begin to wonder
what has happened to all of the distance (and the genes) between
0 and 33 in series A and A’ where crossing over has been prac-
tically eliminated. In view of these considerations, it would
perhaps be simpler to conclude that the percentage of crossing
-over is not necessarily proportional to the distance. ‘The dis-
tance may remain fairly constant, but the amount of crossing
over (i.e., twisting of the chromosomes) will depend upon numer-
ous hereditary factors.

One recalls in this connection Goldschmidt’s (17) suggestive
paper in which he postulated variable forces that hold genes to
their customary loci on the chromosome and which allow an
exchange between allelomorphs in a certain average percentage

EFFECT OF SELECTION ON CROSSOVER VALUES 353

of cases. While we cannot subscribe fully to this theory for
cogent reasons advanced by Sturtevant (717), Bridges (’17), and
Jennings (718), nevertheless Goldschmidt’s proposed theory
would not appear entirely. supererogatory, for a crossover value
is apparently a variable and the variation is due to or controlled
by multiple hereditary factors. A cross between low crossover
stocks and the original population, and testing out a large num-
ber of F, segregates should throw the desired light on this ques-
tion. Unpublished data indicate that segregation in crossover
values does take place as one would expect on the basis of the
multiple-factor explanation.

5. May we suppose that we have been taking advantage of
small mutations in the nature of modifying factors arising during
the course of selection? While this is possible we are inclined to
doubt it, for favorable mutations evidently do not take place
in the direction of selection as readily as this view would imply
(ef. Muller and Altenburg, 719).

The following conclusions may be drawn from the data of this
paper:

1. Crossover values are very variable and part of this vari-
ability is due to genetic causes.

2. Low selection has been effective in two entirely independ-
ent series, A and B.

3. The low crossover stock bred true to about 0.6 per cent
(almost zero) for nine generations in series A’ (derived from
series A).

4. The low crossover stock bred true to about 6 per cent for
twenty-two generations in series B.

5. High selection probably induces double crossing over, as
shown by series C.

6. Crossing over in the various regions of the sex chromosome
(and other chromosomes?) is probably controlled by multiple
incompletely dominant factors.

354 J. A DETLEFSEN AND E. ROBERTS

LITERATURE CITED

Brivces, C. B. 1915 A linkage variation in Drosophila. Jour. Exp. Zodl.,
vol. 19, pp. 1-21.
1916 Non-disjunction as proof of the chromosome theory of heredity.
Genetics, vol. 1, pp. 1-52, 107-163.
1917 An intrinsic difficulty for the variable force hypothesis of cross-
ing over. Am. Nat., vol. 51, pp. 370-373.
1917 Deficiency. Genetics, vol. 2, pp. 445465.

DetLersEN, J. A. AND Roperts, E. 1920 Variation in the percentage of cross-
overs and selection in Drosophila melanogaster. Anat. Rec., vol. 17,
p. 336. :

GoupscuMipT, R. 1917 Crossing over ohne Chiasmatypie? Genetics, vol. 2,
pp. 82-95.

Gowen, J. W. 1919 A biometrical study of crossing over. On the mechanism
of crossing over in the third chromosome of Drosophila melanogaster.
Genetics, vol. 4, pp. 205-250.

JENNINGS, H. S. 1918 Disproof of a certain type of theories of crossing over
between chromosomes. Am. Nat., vol. 52, pp. 247-261.

Moraean, T. H., AND Bripaes, C. B. 1916 Sex-linked inheritance in Droso-
phila. Publ. Carnegie Inst. Wash. D. C., no. 237, pp. 1-87, 8 fig., 2 pl.

Moraean, T. H. 1919 The physical basis of heredity. J. B. Lippincott Co.,
Philadelphia and London. 305 pp., 117 fig.

Mutuer; H. J., anp ALtTENBURG, EX. 1919 The rate of change of hereditary
factors in Drosophila. Proc. Soe. Exp. Biol. and Med., vol. 17, pp.
10-14.

PuioueH, H. H. 1917 The effect of temperature on crossing over in Droso-
phila. Journ. Exp. Zool., vol. 24, pp. 147-210.

Sturtevant, A. H. 1917 Crossing over without chiasmatype. Genetics, vol.
3, pp. 301-304.
1919 Inherited linkage variations in the second chromosome. Pub.
Carnegie Inst. Wash. D. C., no. 278, pp. 305-341.

WernsTEIN, A. 1918 Coincidence of crossing over in Drosophila melanogaster
(ampelophila). Genetics, vol. 3, pp. 135-172.

Me a’ : -)

' ‘ > ~
hs 4 « ate M f : = as
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 3,
APRIL, 1921 ; d
i \

Resumen por la autora, Ruth B. Howland,
Osborn Zoological Laboratory, Yale University.

Experimentos sobre el efecto de la extirpacién del pronefros de
Amblystoma punctatum.

Las condiciones que siguen a la extirpacion bilateral del pro-
nefros de Amblystoma punctatum demuestran claramente que
estos O6rganos son necesarios para la vida del embrién. Todos
los embriones desprovistos de rifones cefdlicos presentan debili-
dad cardiaca y edema, muriendo al cabo de doce dias. La
doble extirpacién de los tibulos pronéfricos no afecta al desa-
rrollo normal de los glomérulos. La presencia de un solo pro-
nefros es suficiente para mantener la vida del animal. Después
de las operaciones unilaterales el resto del pronefros presenta
una marcada hipertrofia, aumentando un 100 por ciento el
Area de la superficie secretora; el contenido ctbico de la masa
de células 63 por ciento, y la longitud de los tubulos 21 por ciento
sobre lo normal.

El conducto segmentario procedente del é6rgano hipertrofiado
posee un dimetro medio mucho mayor que el de cualquiera de
los dos conductos del animal normal. El] riién hipertrofiado
también presenta indicios de una pequefa cantidad de hiper-
plasia, puesto que el nimero de nucleos presente en él es 16 por
ciento mayor que el normal. En el lado operado los glomérulos
se desarrollan normalmente, y los embudos anterior y posterior
pueden regenerar a expensas del epitelio celbmico. La condicién
del conducto segmentario varia considerablemente. En _ los
casos extremos esta representado solamente por una pequena
masa de células degeneradas. En los embriones en que se ha
extirpado el rudimento del corazén en un estado muy temprano
del desarrollo, el desarrollo inicial de los glomérulos es normal,
pero pronto se altera a causa de la enorme distensién de los
vasos sanguineos.

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MARCH 14

EXPERIMENTS ON THE EFFECT OF REMOVAL OF
THE PRONEPHROS OF AMBLYSTOMA
PUNCTATUM!

RUTH B. HOWLAND

Osborn Zoological Laboratory, Yale University

TWENTY-THREE FIGURES

CONTENTS
PANEL GBC DION Ear re cee 5 2G aco clases Veet oote ers oie Maven, AP erGee neptta eam ra soe A, nl sesoee cia aye 355
Materialemethods, and, normal/deyelopment. 4.2. +25ss66- 0642006650046 oe 361
Bilateralexcision of pronephric rudiments. 2. 2226-65. Ss foc de lees 365
Mod Ckoio PebatlOlen: ah Yosrgs eau see ot a oS elas hit ceca 365
Mieetrorbilateraleremovale. 4-4 elke: kee COE Ee ees Oe oe em eta 366
Unuilatendleexcisionsol pronephricudiments. +. 4 eee cess oe oe eee 371
Postoperative effect on the embryo as a whole....................... 342
Effect of unilateral excision on the remaining Prancohnas: Peony ole
Effect of unilateral excision on the glomerulus........ ae Od
Effect of unilateral excision on the other components of nics Sy ae hsitky Oud
Effect of removal of the heart on the development of the glomerulus. ..... 381
SUM MAryaaind TeONClUISION Sey revises rs seis eat lei os le ee eo ochre a action 382
Wer a tUTeV ChE yori s nara ascrs wn. tro 5 hycuil sicce SoS ERI Aordeco aes eee ss 384
INTRODUCTION

The common occurrence, among the lower orders of verte-
brates, of a more or less persistent head kidney or pronephros
has led to the accumulation of a very comprehensive literature
on this subject from the standpoint of pure morphology. Little
evidence has been furnished in any group, however, of the role
which these organs play in the life of the embryo.

Price (710), working on the head kidney of one of the myxi-
noids, Bdellostoma stouti, a form in which it persists throughout
adult life, follows up his earlier descriptions of the development

1 A preliminary report of the results obtained was published in 1916. (How-
land, R. B.,’16. On the Effect of Removal of the Pronephros of the Amphibian
Embryo. Proc. Nat. Acad. Sc., vol. 2.)

309

356 RUTH B. HOWLAND

of the excretory system with a study of both the structure and
the function of this organ in its fully formed state. The head
kidney of Bdellostoma is shown to be a composite structure,
possessing at its earliest appearance all the characteristics of a
pronephros, with the single exception of the typical glomerulus,
but later fusing with the anterior end of the mesonephros and
losing all connection with the exterior. Structurally, then, the
head kidney in at least one of the myxinoids is rendered incap-
able of playing the réle of an excretory organ, but since it is
connected with both coelomic cavity and the circulatory sys-
tem, and since, also, it has been proved possible to transfer
certain substances from the coelomic fluid directly into the cir-
culation, Price concludes that its probable function is the trans-
ference of lymph from the body cavity into the blood-vessels.

Since the discovery of the organ by Johannes Miiller (29),
the origin, development and morphology of the amphibian
pronephros have been described by many investigators, chief
among whom are von Wittich (’52), Fiirbringer (’78 a), and Field
(91). An excellent review of the early controversies concerning
its structure is also given by Field. ‘The presence of a well-
developed pronephros in the amphibian embryo, its early appear-
ance, and its relatively large size have led to the general assump-
tion that it is a functioning organ. Its characteristic structure,
consisting in a glomerulus which extends freely into the coelomic *
cavity, a coiled tubule furnished with open ciliated funnels for
the intake of coelomic fluid, and a simple duct establishing
direct connection with the exterior, further points to its function
as excretory in nature. Still, from the physiological viewpoint,
no experimental evidence as to the extent to which the embryo —
is dependent upon it for the elimination of excretory products
had been offered at the time of publication of my first note (’16);
Since then Swingle (19), working independently upon the embryo
of Rana sylvatica, has obtained results which in the main agree
with my own.

In the series of experiments described in the present paper,
the necessity of these organs for the life of the embryo has been
proved by the fact that death follows, in time, after the removal

REMOVAL OF PRONEPHROS OF AMBLYSTOMA aon

of both pronephroi. The uniform occurrence of a pronounced
edema after bilateral extirpation, similar to the condition which
follows certain pathologic conditions in the permanent kidneys
of the higher animals, suggests a further parallel between the
larval kidneys on the one hand and the permanent kidneys on
the other with respect to their function. Extirpation of the
coiled portion of one or both pronephroi has also afforded the
opportunity of investigating the question of correlation in develop-
ment through a study of the effect of its removal on the other
components of the excretory system. ‘The response of one kidney
in cases where it has been left functioning alone has further led
to the consideration of the factors involved in the restoration
of the normal secreting area through the process of compensatory
hypertrophy.

Although no invariable rule can be formulated as to the type
of regulation which may be anticipated as a consequence of the
abnormal conditions imposed by extirpation of an embryonic
region or organ, a survey of the results obtained in the many
instances already investigated shows that, in a large proportion
of cases, there occurs a more or less complete regeneration of the
excised part. Byrnes (98 b) and Harrison (18) have shown,
for instance, that the limbs of the amphibian embryo, if removed
at an early age, will soon be replaced through the regenerative
capacity of the surrounding tissue. ‘This is also true of the audi-
tory and nasal placodes, the lens, and the gills. The possession
by the amphibia, of the regenerative power to such a high degree
would naturally lead to the presumption that removal of the
pronephric rudiment might result in a similar replacement of
this organ. This, however, as will be shown later, is not the
case in the Amblystoma larva, for the adjustment consequent
on removal of the pronephros is not in the nature of a restitution,
but is a compensating hypertrophy? of the remaining head kid-

2 A peculiar instance of compensatory hypertrophy in another organ is cited
by Kochs (’97) in his work on Triton, where the amputation of the fore leg often
resulted in a marked hypertrophy of the tail. Retardation or acceleration in
the growth of a fore leg in the larvae of Rana esculenta and Bufo viridis may be

induced by removal of a hind leg, according to Kammerer (’05), the hastening or
arrest of growth depending on the rapidity of wound healing.

358 RUTH B. HOWLAND

ney and of its duct—such compensation as is common in the
adult kidney of the higher vertebrates. The degree of com-
pensation which has been attained by the single kidney has been
estimated in terms of increase in the secreting surface of the
pronephric tubules, as well as by measurement of the volume of
the cells making up the walls.

Although the literature on the subject of hypertrophy in the
kidney of the higher forms is too extensive to permit of a dis-
cussion in any detail in the present paper, it may be well to
mention several of the more important standpoints from which
the subject has been treated.

The question of direct causal connection between the demand
for increased functional activity and the changes in the com-
pensating organ has been definitely settled by such experimental
studies as those of Sacerdotti (96), in which the kidneys of
unoperated dogs were stimulated to compensatory overgrowth
by injection of the blood from completely nephrectomized ani-
mals. In this instance, as in all the early pathological literature
dealing with this subject, the exact nature of the histological
changes evoked in the stimulated organ was given only second-
ary consideration. Recently, however, with the general accep-
tance of the distinction between the terms hypertrophy and
hyperplasia, more accurate observations of the condition of the
kidney constituents consequent on increased activity have been
made. Both forms of regulation may occur in the same organ,
each being limited to a definite area. Wolff (00), in his con-
tributions on the macroscopic and microscopic conditions of the
hypertrophied kidney after resection, draws a sharp distinction
between those changes which occur in the region of the lesion,
and those occurring in the uninjured portions of a resectioned
kidney. In the former location he observes that mitoses appear
at the end of two days, resulting in the formation of new epithe-
lium. In the uninjured portion of the remaining kidney tissue,
however, no new formation of either urinary tubules or of glo-
meruli takes place as compensation for those excised, but here a
sufficient restitution occurs through increase in size, and the
normal balance is restored. From the histological viewpoint,

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 359

this process is sure to be almost entirely one of hypertrophy, not
a hyperplasia, of the kidney elements. In the glomerulus and
urinary tubules the former process occurs exclusively, in the
epithelial cells themselves, hypertrophy with a negligible amount
of hyperplasia.

The clearest and by far the most accurate statement of the
exact histological conditions found in the hypertrophied kidney
is given, however, by Galeotti and Villa Santa (’02). These
authors approached the problem from a widely different view-
point, their main object being to determine whether the hyper-
trophied kidney of an adult animal would show the same histo-
genetic potency as that of an animal which had not attained its
full growth. From their study of the kidneys of young and
adult dogs and rabbits, careful estimates were made of the num-
ber of glomeruli found in sections of normal and hypertrophied
kidneys and of the relationship between the average surface
area of the glomeruli and the number present. Furthermore,
the diameter of the lumen in the tubuli recti was accurately
measured and the secreting surfaces in the two kidneys obtained
for comparison. The volumes of the cell walls in the two cases
were also computed and contrasted. The thoroughness of the
methods used in these computations gives added weight to their
conclusion that, whereas in the young kidney hyperplasia may
occur, the adult kidney has lost its potency for addition of new
parts, and can only respond by the enlargement of those elements
already present.

Kittleson (’20), in his recent report on the effects of inanition
and refeeding on the growth of the kidneys in young rats, con-
firms the opinion of former workers that starvation inhibits the
formation of renal corpuscles, and further concludes that ‘‘refeed-
ing after stunting results not only in a hypertrophy of the renal
corpuscles but also in.an increase in number, which may even
exceed the normal.’”’ This would indicate the possibility of both
hypertrophy and hyperplasia of the same kidney element,
induced by these abnormal conditions.

No instances are on record of experiments dealing with the
production of hypertrophy or hyperplasia in the adult am-

360 RUTH B. HOWLAND

phibian kidney, although Levi (’05) claims that the anuran
mesonephros has, in one species, the power of regeneration after
injury. He destroyed both the urogenital anlage and that
of the wolffian duct and tubules in Bufo larvae by means of a
red-hot needle, and obtained after a certain time complete
regeneration of the excretory organs. His method is, however,
open to the criticism that no accurate estimate of the actual
extent of the injury done to the organs by this type of operation
can be made. The introduction of a hot needle may cause only
minor displacement or destruction of a few cells of the duct or
coils. . Certainty as to the degree of injury can be assured
only by complete excision.

Removal of a portion of a given organ or system has been
found to have a varied effect on the formation of its other constit-
uents. ‘The extent to which the growth of one part is influenced
during its development by other developing portions of the
embryo varies widely, the scale of difference ranging from com-
plete interdependence to those extremes in which each con-
stituent possesses the potency for self-differentiation without
the influence of any formative stimuli.

In view of this variation in the degree of correlation exhibited
by closely related parts during their early growth, it was of interest
to consider, particularly in cases of bilateral extirpation of the
pronephros, the effect of the absence of the tubules on the forma-
tion of the glomeruli. The operated animals showed without
exception a differentiation of glomerular tissue perfectly normal
in position and size. The glomerulus therefore possesses the
power of self-differentiation, and is entirely independent of the
presence of the tubular elements.

It gives me great pleasure to acknowledge my indebtedness to
Prof. R. G. Harrison, at whose suggestion this investigation was
begun, for his helpful and constructive criticism during the
course of my work.

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 361

MATERIAL, METHODS AND NORMAL DEVELOPMENT

Embryos of Amblystoma punctatum were used for all the
experiments. The stages chosen for operation varied from the
condition in which the first loop of the pronephric tubules
appears as a slight, ventrally directed curve of the duct (fig. 1,
stage 30)% to that in which the two funnels, together with the
first loop, appear as a broadened Y (fig. 2, stage 32). In all
cases embryos were used before contraction of the body muscles
began, as movement not only hindered the operation, but often
tore open the wound after successful removal of a kidney.

Figs. 1 and 2 Embryos in the stages used for operating. PR, pronephros
located below the third and fourth myotomes. Figure 1, earliest stage (stage
30); figure 2, latest stage (stage 32).

Anaesthetics were unnecessary, and the slight motion due to the
ciliated epithelium was controlled by holding the animal in the
field with an operating needle. The body tissues in these early
stages are easily distinguished from each other through slight
differences in pigmentation, and, in addition, are so loosely
bound together as to allow removal of the pronephric mesoderm
without dislocating the cells of contiguous regions. In a few
‘Instances portions of the somatopleural layer ventral to the
pronephric rudiment were included in the tissue removed, result-
ing in retarded development, in abnormalities, or even in total
absence of the limb on this side (Harrison, 718).

3 See Harrison, R. G., 718, Jour. Exp. Zodél., vol. 25, no. 2, p. 417, footnote 9.

362 RUTH B. HOWLAND

The general methods employed in operating are so well known
that no detailed description of them is necessary here. The
special technique required in removal of the pronephros and in
the construction. and measurement of the models will be described
in later sections.

The pronephric swelling is one of the earliest and most clearly
defined of the developing organs. Its position may be accu-
rately located at a stage not long after the closing over of the
neural folds, when the first eight pairs of muscle plates may be
seen, and, like these, it differentiates in an anteroposterior
direction. It is found in the region immediately underlying the

Figs. 3 and 4 Embryos showing the first looping of the tubule, due to rapid
growth of the cells just posterior to the nephrostomes. Figure 3, a—a, original
axis of the pronephric rudiment. Figure 4, b—b, direction of first bend of the
tubule.

third and fourth myotomes as a bulbous thickening tapering
posteriorly into a short thickened ridge. Operations at this
period, although having the advantage of not interfering with
developing nerves or blood-vessels, are inadvisable, since the
mesoderm is still so compact that excision results almost invari-
ably in the removal of more than the pronephric rudiment. In
succeeding stages, the delimitation of the segmental duct pro-
gresses, and the original bulbous enlargement becomes pitted in
in two places on its coelomic border, establishing the nephros--
tomal openings into the anterior and posterior funnels. These
two openings lie opposite the midline of somites three and four.
At the same time the tubule, which originally lies along the
longitudinal axis of the body just below the muscle plates (fig.

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 363

3, a—a), lengthens rapidly in the region of the fourth myotome,
and bending outward and downward in an acute angle over the
upper surface of the yolk (fig. 4, b—b), forms a U-shaped loop.
A little later this is bent over anteriorly, and may even come to
lie shghtly farther forward than the anterior nephrostome.
The pronephros of Amblystoma differs from that of certain
of the Anura in the possession of two instead of three nephros-
tomal canals,‘ and in the absence of the common chamber or
‘pronephric pouch,’ the funnels, instead, narrowing directly

Figs. 5 and 6 Diagrams showing the region where greatest growth occurs in
the early and late stages of development of the pronephros. Figure 5, condi-
tion before the coiling of the longitudinal tubules connecting the funnels. Fig-
ure 6, growth of these tubules in the older kidney. a.f., anterior funnel; p.f.,
posterior funnel; c.t, longitudinal tubule; s.d. segmental duct; x-y, region of great-
est growth during early development resulting in the formation of the ventro-
lateral portion of the pronephrie coil, p.c. Shaded areas drawn from wax models,
pronephric coil indicated by curved lines.

into the U-shaped tube just described. With further multi-
plication of the cells just below the funnels, two longitudinal
tubules are established (figs. 5 and 6, c.t., and fig. 17, L.T.),
separating the anterior and posterior nephrostomes from each
other and from their original point of junction with the looped
tubule (x), as they grow. This growth is at first a very slow
process as compared with that of the U-shaped portion. In the
latter region the active proliferation of cells results in a rapid

* One instance is on record of the presence of a third funnel on both sides in
an Amblystoma larva. See Field (’91).

364 RUTH B. HOWLAND

coiling of the tubule, the early increase in size of the kidney
being limited mainly to this region, between the connecting
tubule and the proximal end of the segmental duct (figs. 5 and
6, x to y). This eventually forms the ventrolateral region of
the fully formed organ, which will again be referred to in con-
nection with the discussion of edema.

In still later stages® the nephrostomal canals and their connect-
ing tubule also elongate, and are thrown into loops and folds
(fig. 6, l.¢.), retaining their dorsal position and extending slightly
laterally over the coils already formed. This portion may
then be termed the dorsolateral region, as contrasted with the
ventrolateral portion already mentioned. That part of the
tubule which is a direct continuation of the segmental duct
never becomes strongly convoluted, but retains its original posi-
tion along the ventrolateral boundary, slanting obliquely toward
the dorsal surface over the kidney from the anterior margin.
Minor folds may occasionally occur along its course. Increase
in growth is outwardly evidenced by a more and more pro-
nounced swelling in the pronephric region. Operated speci-
mens may be easily distinguished from normal animals, even
after healing is complete, through the absence of this thickening
on the operated side. Posterior to the pronephric coils, the
segmental duct extends backward along the body just below
the ventral surfaces of the muscle plates. The junction between
pronephric coils and the proximal end of the segmental duct is
always in the immediate region of the posterior funnel. With
the subsequent downgrowth of the myotomes the formation
of the shoulder-girdle and anterior limb buds, the pronephros
becomes partly covered, and comes to lie deeper in the body,
and nearer the midline of the embryo. The edges of the myo-
tome also extend downward over the segmental duct, making its
removal in this stage extremely difficult.

® Excised kidneys in the older stages may be slightly stained, and the capsular
nuclei, thus made visible, removed. The tubules may then be easily uncoiled
for observation in water or weak alcohol. Oil is an unsatisfactory medium, both
for examination and preservation, as it not only increases the brittleness of the
tubules, but renders them too transparent for clear definition.

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 365

Parallel with this development, although not appearing at so
early a stage, the rudiment of the glomerulus is formed. This
first appears as a thickening in the opposite or splanchnic wall
of the coelom, extending over an area as long as the distance
between the anterior and posterior funnels. Vascular cavities
soon appear, and at an early stage these become continuous
with branches of the dorsal aorta. The arterial supply is derived
directly from this source; the venous supply from the postear-
dinal vein, which enlarges in the region of the pronephros, form-
ing a large venous sinus, into which the fully developed kidney
projects. The tubules are thus continually bathed in the blood
returning from the posterior part of the body (Field, 791).

BILATERAL EXCISION OF PRONEPHRIC RUDIMENTS
Mode of operation

The first experiments consisted in the removal of the proneph-
ric rudiments on both sides, to test the functional necessity of
these primary organs in the life of the embryo.

In the largest proportion of cases, a period of several hours, or
even a day, was allowed to elapse between the removal of the
right and left pronephros. However, the two excisions may
follow each other immediately without incurring serious results.

Sharp-pointed needles, inserted in glass rods, were substituted
for the more generally used iridectomy scissors. Controls were
kept under identical conditions of light, temperature, water,
etc. ‘lwo methods were employed in removing the pronephros.
In the first, three straight cuts were made, one beneath and one
along each side of the pronephric swelling. The flap of ecto-
derm thus defined was loosened from the underlying mesoderm,
and the organ removed from below. In the second and more
satisfactory method, a single incision was made, dorsal to or
immediately over the thickening, the tubule raised upward from
below, pulled outward, and excised. In loosening the nephros-
tomal surfaces of the funnels, as much of the tissue in a dorso-
median direction was removed as seemed possible without
disturbing the splanchnopleuric wall, since in this region the
glomerulus normally arises.

366 RUTH B. HOWLAND

Larvae in which the triple-incision method was used healed
much more slowly than those to which the second method was
applied, since contraction of the cut edges left a gaping wound
much greater in extent and permitted more oozing of the yolk.
On the other hand, when only one cut was made, a critical
inspection of the excised tubules was necessary to make sure of
total removal. In the majority of cases the incision is entirely
healed at the end of an hour and a half. :

Effect of bilateral removal

Most conspicuous of any of the postoperative conditions
resulting from bilateral excisions was a pronounced edema, par-
ticularly in the anteroventral region. It is well known that
edema of the amphibian embryo commonly occurs as the result
of a variety of causes. Narcotized embryos reared in a solution
of acetone-chloroform (chloretone), though structurally normal
in other particulars, show a slight edema and pericardial effusion
as a result of weakened heart action (Harrison, 04). More
pronounced abnormalities result when early developmental
stages are exposed to the heat of direct sunlight, or may be
induced experimentally by exposing the embryos to radium
rays (O. Hertwig, ’11). McClure (719), in his recent work on
edema in anuran larvae, draws attention to the ‘‘less extensive
tubular complex which normally occupies a dorsolateral position
in the pronephros, and into which the nephrostomal canals
directly open,” and the ‘‘tubules which normally constitute the
greater portion of the kidney and which occupy a medial and
ventral position.”” From a study of the histological conditions
existing in edematous frog larvae, he concludes that there is a
functional as well as a morphological difference between these
two regions, for in all of the embryos in which edema had become
apparent, the ventrolateral tubules were either entirely absent
or but poorly developed. From this he argues that deficiency
in the ventrolateral tubules alone may be the cause of edematous
conditions.

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 367

In bilaterally operated Amblystoma larvae, the swelling which
first appeared in the region of the wounds progressed gradually
forward, the pericardial cavity soon becoming enlarged (figs. 7

Figs. 7 and 8 Section and entire sketch of embryo, from which both kidneys
had been removed, to show pericardial effusion. Figure 7, section through peri-
cardial (p.c.) region. Figure 8, camera drawing of embryo, showing swelling in
the heart region.

Fig. 9 Section through an embryo from which both kidneys had been removed.
Both glomeruli (gl) are present, extending out into the enlarged coelomic cavities
(c); A, aorta.

and 8). Later the fluid caused a pronounced distention of the
abdominal cavities (fig. 9), and in extreme cases the gills also

became swollen and distorted. Slowing or entire absence of
circulation accompanied this condition and sloughing of the

368 RUTH B. HOWLAND

ectoderm was not infrequent. Microscopic examination of
sections through edematous embryos showed the tissues of the
body to be in various stages of degeneration. Pressure of the
accumulated fluid often forced the intestine ventrally or to one
side and the fibers of the muscle plates were separated by large
vacuoles. The muscle fibers themselves also became vacuolated,
and in extreme cases the whole region was reduced to a spongy
mass of irregular fibers with scattered nuclei.

Although the splanchnopleural mesoderm which gives rise to
the glomeruli had been left intact during the operations, the
question nevertheless arose as to whether normal conditions of
development would obtain for these organs in the absence of
other parts so closely allied with them. Sections made through
the operated region in embryos killed four days and six days
after double excision showed capillary tufts extending out into
the much-dilated coelomic cavity (fig. 9, gl). The tubular
region cannot, therefore, be considered to exert any influence in
the nature of a formative stimulus on the development of the
glomeruli, since these parts of the system arise quite inde-
pendently. Furthermore, the presence of the glomeruli in these ~
operated cases would tend to strengthen the view supported by
McClure (19) that the glomerular filtrate, given off directly
into the coelomic cavity, collects here in excess, producing the
typical edematous condition already described. ‘These embryos
also showed well-developed anterior and posterior funnels,
extending laterally into the regions from which the tubules had
been removed, and ending blindly there. Segmental ducts were
present, in some cases the lumina being flattened dorsoven-
trally, in other instances the cells of these tubules showing
marked signs of atrophy. The condition of the funnels and
segmental ducts after extirpation of the kidney will be dealt
with more fully in connection with the question of unilateral
removal.

Efforts were made to bridge over the interval between the
operations and the beginning of functional activity of the meso-
nephros. The first means applied was that of pricking the body
wall as soon as abnormal distention was evidenced. ‘The larvae

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 369

were immersed in 0.4 per cent NaCl in an attempt to balance the
loss of essential salts through the escape of the glomerular
filtrate. Although slightly stimulated heart action resulted,
probably due both to the stimulus of operation and to relieved
pressure in the pericardial cavity, only temporary benefit was
derived in this way, for, with the accumulation of new fluid, the
former pathologic condition was restored, and death followed
after a short interval. It is quite possible that if a sufficient
number of experiments were made, varying the constituents of
the solution in which these larvae were kept, a satisfactory
medium might be found for prolonging the life of these animals.
A determination of the optimum salt percentages of such a solu-
tion has not yet been undertaken.

The second means employed was the transplantation of the
pronephric rudiment to the region of the mesonephros. It was
by this means possible to test the capacity for reestablishment
of function, through union with the segmental duct. In a series
of thirty embryos, the right head kidney, without the ectodermal
covering, was placed under the skin farther back on the same
side. Twenty-four hours later, the left pronephros was removed
from those embryos which had responded well to the first opera-
tion and were apparently recovered. The transference and
proper orientation of the kidney in these operations was easily
accomplished and the wounds healed entirely in the usual short
time, but the general edematous condition common to embryos
on which only the bilateral excision had been performed sub-
sequently developed. With the exception of a few which were
preserved and sectioned at the end of a week, all of this series
died within twelve days, showing no indication of resumption of
function by the transplanted tubule. The transplanted tubule
still retained its identity, although as a general rule the cells
were pressed together in a solid mass, and only in a few instances
a distinct lumen was visible. Removal of the pronephros
resulted here in the partial or total atrophy of the segmental duct,
to which attention will be called in greater detail below. How-
ever, in cases where the transplanted tubule had been placed in
the immediate region of the segmental duct, no connection was

370 RUTH B. HOWLAND

restored between these two components, nor did the duct,
posterior to the transplant, give any evidence of functioning.
A slightly different method was applied in a third series of experi-
ments. The pronephros was not alone transplanted, but was
taken together with the overlying ectoderm, the surrounding
mesoderm, and even small portions of the ventral myotomal
walls. This transplant was procured from another animal, and
was transferred into a previously prepared incision and held in
place until healed. On the next two successive days the left
and right pronephric rudiments were removed. No appreciable

a, a
s'p| YY? sD
a

Fig. 10, Aand B- Diagrams to show the location and extent of operations made
in removal of different segments of the embryonic pronephros. In A the seg-
ment removed, x, was a part or the whole of the U-shaped tubule anterior to the
segmental duct. In B the segment removed, x, consisted of a part or the whole
of the rudiment of the funnels.

difference was noted in the ensuing condition of this series, and itis
safe to conclude that under these circumstances the excised tubule
is unable to readjust itself and function in its new location.
Interruptions to the development of the pronephros by a less
radical operation also go to strengthen the belief that the regen-
erative capacity of the kidney tubule is either very limited or
very slow in taking place. In a number of embryos from which
one pronephros had been extirpated, a small portion of the
opposite rudiment was also excised. The segments removed (2)
were at two levels, as designated in figure 10, A and B. The
funnels were undisturbed in one group (A), a short piece of the

® This does not apply to the coelomic epithelium which lines the nephrostomal
opening, as will be shown in a later section.

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 371

first loop being removed, while in the second group (B), the
funnel rudiment was cut off just anterior to the first bend. Of
the twenty specimens used, all exhibited symptoms identical
with those induced by bilateral excision.

The extent of regeneration which would occur in a defective
tubule if the opposite kidney were allowed to remain intact is
still a matter for investigation, but it is not improbable that
a given portion of a tubule may possess a prospective potency
which would insure the restoration of an excised section. How-
ever, in dealing with an organ where the demand for functional
activity follows so closely upon this disturbance of normal con-
dition, the requisite time for readjustment may be the factor
lacking. Accumulation of excretory fluid may inhibit the regen-
eration which might be the normal consequence, if excretory
activity were maintained by an undisturbed kidney. A funda-
mental difference thus places excision of the kidney in a category
apart from the majority of regeneration or transplantation
experiments upon the amphibian embryo which have been
reported up to the present, for the effects consequent on extirpa-
tion and transfer of limb rudiments, optic vesicles, or nasal pits,
though abnormal, are not of a nature to interfere with any of the
vital functions of the embryo.

UNILATERAL EXCISION OF PRONEPHRIC RUDIMENTS

Conclusive evidence having been obtained as to the essential
nature of the pronephros in the life of the embryo, a further
study of the correlation of the development of this organ with
that of the other components of the excretory system was then
undertaken.?7 Unilateral excision of the pronephric rudiment
served.as a practical means to this end.

The technique of operating has already been discussed in the
previous section, but a word of explanation is necessary regard-
ing the controls used in this series. Since a more or less pro-.
nounced retardation in growth was the unavoidable consequence

7 As has been previously stated, the glomerulus was found to develop normally
even in the absence of the pronephrie coil.

ote RUTH B. HOWLAND

of such operations, the controls were always more advanced
than the operated embryos, so that, for the comparison of the
excretory organs, others had to be selected as described below
(p. 373).

Postoperative effect on the embryo as a whole

Every outward evidence of successful readjustment to the
new conditions imposed was shown by the operated larvae.
Adverse symptoms, such as edema and general sluggishness, were
absent, and, barring the slight retardation already mentioned,
normal progress continued, except in those cases in which the
limb bud was disturbed or entirely removed. The ectodermal
surface which showed a slight concentration of pigment in the -
initial stages of wound healing gradually became indistinguish-
able from the surrounding regions, and differed from the opposite
side only in the absence of the distention caused by the under-
lying pronephric coils.

Effect of unilateral excision on the remaining pronephros

The pronephros remaining after removal of one head kidney
obviously takes over the function of excretion usually per-
formed by the two organs. Beginning with the fourth
day after excision, operated embryos were killed for obser-
vation each day for a period of two weeks. Sections showed
distinct changes in the several remaining components of the
excretory system, particularly in the head kidney functioning
alone, the size of which was indicative of a marked compensa-
tory hypertrophy. Since the controls taken frcm the same
egg mass and carried along under the same conditions as those
to which the operated forms were subjected invariably showed on
sectioning a more advanced stage of development, the first step
in determining the nature and extent of the change in the oper-
ated kidney was the establishment of a criterion for comparison
of an operated with a normal embryo. An operated individual
(PN 7) was chosen as a typical case and a large number of nor-
mal larvae of apparently the same age was examined to obtain

REMOVAL OF PRONEPHROS OF AMBLYSTOMA ate

one in which the stage of development was identical. In many
embryos where superficial features, such as length, breadth, and
condition of limb and gill rudiments, were the same as those of
PN 7, it was found on sectioning that the internal organs varied
widely in degree of development. The normal larva finally
selected (PN 7 d) tallied* not only in external measurements,
but showed the several internal organs (retina, lens of eye,
digestive tract, etc.) to be in a stage corresponding to those of
BINT

As a further check against the possibility of error in the choice
of a normal duplicate, a second duplicate was chosen, and the
respective volumes of the kidneys of the two roughly compared
by the following method:? On drawing-paper of uniform thick-
ness the serial sections of the entire kidney of PN 7 d and of the
second duplicate (PN 7 d, no. 2) were projected and the lumen of
the tubule outlined. The drawings of each kidney were then
carefully cut out. No attempt was made to assemble them in
the form of a model, but the weight of the paper used for each
was taken as a standard for comparison. The weight of PN 7 d
was 2.35 grams and that of PN 7 d, no. 2, 2.26 grams, giving a
difference of only 0.09 gram, or about 4 per cent—a variation so
small as to be considered negligible. The larger normal kidney
(PN 7 d) was used for comparison with the hypertrophied one
in order to lessen the possibility of exaggerating the difference
between the two.

After the normal duplicate (PN 7 d) had been selected, several
methods were open for the determination of the nature and
degree of the hypertrophy of the remaining pronephros in the
embryo from which the organ on one side had been removed.

8 The slight variation would tend rather to minimize the contrast than to
accentuate it, since, if either, PN 7 d is the more advanced.

* In connection with the review of Kittleson’s paper (see previous reference),
I find that somewhat the same method was employed by him in his estimation
of the relative surface areas and weights of the kidneys of rats. The weight in
grams was reduced to square centimeters by estimating the average area in
square centimeters of one gram of paper, and from this the total volume of the
kidney was estimated.

374 RUTH B. HOWLAND

Wax models of the unoperated right pronephros of PN 7 and
the corresponding organ in PN 7 d were constructed by means
of the Born method, at a magnification of 200 (figs. 18 to 21).
Unlike the normal model, to which reference has already been
made in an earlier section, these two are reconstructions of the
lumen of the tubules without enclosing walls. The model of
the kidney of the operated embryo not only showed a consider-
able increase in the thickness of the tubule as contrasted with
that of the normal, but its length is also appreciably greater.
This was determined by taking the average of five measure-
ments. - A flexible but inelastic cord was pinned along the sur-
face of the wax for its entire course, and its length thus recorded.
On each measurement the cord was pinned along a different sur-
face, so that the data would be of a representative nature. For
the model of the normal kidney the average length was found
to be 155 em., at a magnification of 200, with a probable error
of +0.181: for PN 7, 188 cm., at a magnification of 200, with a
probable error of +0.155, showing an increase of 21 per cent
over the normal condition.

Microscopic examination of the tubules showed the walls in
the normal organ (PN 7 d) to be relatively thick and made up of
cuboidal cells. The hypertrophied tubules (PN 7) were thinner
walled proportionately, the cells often flattened and elongated,
and the lumen strikingly larger than that of the unoperated
specimen!® (figs. 22 and 23). Outline drawings were made of
the outer and inner boundary of the walls of the hypertrophied
and of the normal tubules at a magnification of 600 (figs. 11 and
12). A non-elastic cord was then pinned at frequent intervals
along the inner lines, removed and measured, and the circum-
ference obtained. Five different sections were used at different
levels in each case, and in each section from three to five tubules
were measured, making a total of twenty-one measurements for
each kidney. The average measurement obtained for PN 7 d

10 Tn a series of experiments reported by Detwiler (’18), the pronephros was
often carried along in the transplantation of the limb rudiments. The enlarge-
ment of the undisturbed pronephros and its contrast with the normal condition

may also be seen on examination of his plates (Jour. Exp. Zo6él., vol. 25, 1918,
pl. 3, figs. 18 and 19).

REMOVAL OF PRONEPHROS OF AMBLYSTOMA By a

was 78 mm. (0.13 mm. in actual size), as contrasted with 130
mm. (0.216 mm. in actual size) for PN 7—a fact suggesting the
large percentage of increase in functional capacity of the two
kidneys determined and described below. From these projec-
tions also the thickness of the walls was estimated by taking the
average of sixty measurements in each kidney. The average
thickness of the wall of the normal kidney is 10.9 mm. at a mag-

oo.
oe Se

Fig. 11, PN 7d Normal tubules with thick walls, and cells bulging into the
lumen. X 240.
Fig. 12, PN 7 Hypertrophied tubules with large lumens and thinner walls.

nification of 600, or 0.0181 mm. in actual measurement. For
the hypertrophied kidney, the average thickness is 7.8 mm. at
a magnification of 600, or 0.013 mm. in actual measurement.
Having the length of the two kidneys, determined from the
models, and considering each kidney as a simple cylinder, with
the average measurement just obtained as its circumference
and the length as its height, the areas of the inner surfaces in.
each were computed. In the normal kidney the area of the
secreting surface was found to be 1.007 sq. mm. as contrasted

376 RUTH B. HOWLAND
with 2.037 sq. mm. in the compensating organ—an increase of
more than 100 per cent (table 1).

A comparison of the volumes of the cells making up the walls
of the two kidneys is likewise of great importance in determining
the nature of the response brought about by a unilateral opera-
tion. Although in the calculation of the surface area of the
tubules it seemed sufficiently accurate to regard them as cylin-
ders with the average circumference as their boundary, it did not
seem possible to apply such a geometrical method in estimating
the volume of the walls, for the cells (figs. 11 and 12), especially
in the normal kidney, often bulge out into the lumen, making
this, as seen in projected outline, very irregular. ‘The method
already described in the selection of a normal model was again

TABLE 1

Showing the area of the inner or secreting surfaces of the normal kidney, PN 7 d,
and the hypertrophied kidney, PN 7

INNER OUTER AREA OF ACTUAL

SERIES NUMBER eter eee (5¢ 600) eNet AREA OF

(X 600) (X 600) S360 1000) ql oe eee

cm. cm. cm. sq.cm. sq. mm.

PN 7 d (normal). 7.8 13.6 465 3627 1.007
PN 7 (hyper meohied).. 13 18.6 564 7332 2.037

used (p. 373). Paper of uniform weight was obtained, and to
insure further accuracy, section number. one of the normal
kidney was projected on one half of a sheet, section one of the
hypertrophied organ on the same sheet of paper, and so on
through the series. The walls of the two kidneys were then cut
out, and the aggregate of each weighed separately. The paper
representing the walls of PN 7 d weighed 20.39 grams that of
PN 7, 33.4 grams, showing an increase of 63 per cent in the
weight of the latter. Since the weight in this case is in direct
proportion to the volume, the hypertrophied kidney may then
be considered as showing an increase of almost two-thirds beyond
the normal. A count of the nuclei in each kidney, all the sec-
tions of which were cut at 10u, shows a small percentage (16 per
cent) of increase in the number of cells found in the larger kidney,

REMOVAL OF PRONEPHROS OF AMBLYSTOMA oid

there being 1950 nuclei in PN 7 d and 2252 in PN 7. Although
this may indicate a certain amount of hyperplasia, that is,
increase in actual number of cells present, it is not large enough
to account for the great percentage of increase in both secreting
surface and in volume. This increase must be mainly attrib-
uted to actual hypertrophy or enlargement of the cells already
present.

Effect of unilateral excision on the glomerulus

A study of the condition of the glomeruli subsequent to double
excision was necessarily limited by the early death of the embryos.
The occurrence of the glomerulus in these cases has already been
noted. On examining larvae from which only one pronephros
had been removed, both glomeruli were found to be invariably
present. The one on the operated side, however, exhibited less
uniformity in size and shape than the normally functioning one.
In some eases the outer layer of the glomerulus and the epithelial
lining of the body wall had coalesced, this being the case not
only on the operated, but on the normal side as well. On the
operated size, however, the distance to be bridged is very greatly
increased, since the absence of the pronephric swelling increases
the width of the coelomic cavity there (fig. 13).

Since the development of this part of the kidney unit is quite
independent of the presence of the tubular portion, its functional
activity can be counted on to continue undisturbed. No hyper-
trophy in glomerular structure, then, would be anticipated, nor
was such found to be the case, for the structural conditions
existing allow free passage of the filtrate ventrally from one side
to the other through the coelomic cavity. Filtration continues
on both sides, the demand for increased physiological activity
falling only on the tubules of the unoperated side.

Effect of unilateral excision on other components of this system

Removal of one head kidney has a widely varied effect on the
formation of the segmental duct on the operated side. The
process of development of the non-functioning ducts took place

3/8 RUTH B. HOWLAND

irregularly, and, in general, only to a limited extent. As shown
in table 2, every gradation occurred, ranging from a condition
in which the lumen, though small and flattened dorsoventrally,
appeared throughout the entire length (A 51, 5 days) to a con-
dition where only the occasional presence of a few degenerating
cells indicated the location of the atrophied rudiment (A1, 10
days).

Fig. 18, PN 2 Showing the increased width of the glomerulus (gl) on the oper=
ated side. Description in text. X 40.

Increased activity of a single kidney also has a definite effect
on the segmental duct of that side. Cross-sections of the duct
of an individual with unilateral operation, when compared with
either of the ducts of a normal larva of the same age, show a
marked increase in diameter.

Although in the excision of the pronephric rudiment great care
was taken to remove as much of the somatopleural mesoderm
as seemed feasible without disturbing the underlying tissues, an
examination of the operated embryos showed that a large num-

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 379

ber possessed well-developed anterior and posterior nephrostomes
ending blindly (figs. 14 and 15 and table 2). This would indi-
cate that the adjacent coelomic endothelium possessed the

\ }

Un Neat / ty, ae? r) 6 /{ } ;

NY Lee OZ
: S \ ai ah f oi, Tf lif /

ss c/

Fig. 14, PN 2. Showing double anterior funnel (d.f.). X 30.
Fig. 15, PN 7 A, regenerated anterior funnel (a.f), on the operated side. B,
regenerated posterior funnel (p.f) on the operated side. X 30.

capacity for regeneration of this portion of the organ, even
though, as we have already seen, this property is not shown by
the tubules themselves. Of the fifteen embryos tabulated

TABLE 2

Showing condition of ducts, nephrostomes, and glomeruli in embryos from which
one pronephros has been removed

SERIES
NUMBER

A 54

A 51

PN 3

PN 4

PN 5

PN 6

PNG

DUCT ON
OPERATED SIDE

Canalized at in-
tervals, flat-
tened

Small, canalized,
flattened

Small, flattened,
atrophied pos-
teriorly

Small, canalized

Present anteri-
orly atrophied
posteriorly

Present anteri-
orly, atrophied
posteriorly

Small, shghtly
canalized

A few degenerat-
ing cells

Small, irregu-
larly canalized

Very small but
canalized pos-
teriorly

Small, discontin-
uous anteriorly

Very small,
(whole mount)

Small (whole
mount)

Small anteriorly,
central portion
discontinuous

Small anteriorly,
none posteri-
orly

DUCT ON
UNOPERATED
SIDE

Round,
canal-
ized,
definite

Large,
round,
canalized

Large,
round

Very large

Large,
round

Large, but
flattened
- anteri-
orly,
Large,
(whole
mount)
Large,
(whole
mount)
Flattened
anteri-
orly
Large,
round

1 For PN 7, 9 days, see figs. 18 and 20.

380

NEPHROSTOMES

GLOMERULI

Anterior present,
posterior ab-
sent

Anterior present,
posterior pres-
ent

No anterior, pos-
terior present

Anterior present,
posterior ab-
sent

Anterior present,
posterior pres-
ent

Anterior absent,
posterior pres-
ent

Small anterior,
no posterior

Anterior present,
posterior ab-
sent

Anterior present,
posterior pres-
ent

2 anterior, 1 pos-
terior

Anterior present,
posterior pres-
ent

No anterior, pos-
terior present

Anterior present,
posterior pres-
ent

Both present

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 381

(table 2), twelve had well-formed anterior, and twelve had pos-
terior nephrostomes. In one instance (PN 2, 17 days) the
anteror nephrostome had doubled, suggesting the three pro-
nephric openings normally found in anuran larvae (fig. 14 and
table 2).

A study of the development of the mesonephros in operated
animals will be the subject of further work.

EFFECT OF REMOVAL OF THE HEART ON THE DEVELOPMENT OF
THE GLOMERULUS

In a series of experiments dealing with the effect of removal
of the heart on certain other organs of the embryo, Doctor Harri-
son removed the rudiment of the heart in larvae of stages 29 to
30. Through his kindness, these embryos were made available
to me for a study of the effect produced on the glomeruli. With
the incoming of new material (March, 1920), these cases were
further augmented by additional experiments.

The glomerulus in Amblystoma punctatum, as has been stated
in an earlier section, normally begins to differentiate from cells
of the splanchnopleural wall below and at each side of the aorta,
in stage 36. Within these clusters of cells vacuolated areas
soon appear, and in a short time connect with the aorta."

In embryos from which the heart has been removed before any
contraction of the cardiac muscles occurred, the initial develop-
ment of the cell groups is normal. However, as the connection
with the aorta is established, the more or less compact nature of
the tufts can no longer be maintained, but from pressure of the
blood plasma which has collected in and is distending the blood-
vessels, the vacuolated centers of the glomeruli are torn apart.
As this accumulation of fluid increases, the outer walls of the
tufts become more and more flattened, and consequently less
easily distinguishable from the wall of the aorta, with which
they are still continuous, finally losing their identity as separate
organs. It is of interest, however, to note their early formation
under these circumstances as additional proof of their independ-
ent power of development.

11 A detailed description of this process together with plates is given by Field,
91 (pl. 1, figs. 8, 9, 10; pl. 6, figs. 48, 49, 50, 52, etc.).

382 RUTH B. HOWLAND

SUMMARY AND CONCLUSIONS

From a study of the results obtained after bilateral and uni-
lateral extirpation of the head kidney and of the heart rudiment
of Amblystoma larvae, the following conclusions may be drawn:

1. Conditions ensuing on bilateral removal of the pronephros
show clearly that this organ is necessary to the life of the embryo,
although the presence of one pronephros suffices to keep the
organism alive and in a healthy condition. All embryos from
which both head kidneys had been extirpated died within from
eight to twelve days, evidencing during that interval weakened
heart action, edema, and effusion into the pericardial and abdom-
inal cavities. Pricking the body wall to relieve the edematous
condition proved ineffective.

2. Double extirpation does not affect the normal development
of the glomeruli. These appear in embryos killed four days
after the operation.

3. The pronephros remaining after the removal of one head
kidney takes over the function of excretion usually performed by
the two organs, and, concomitant with the increased physiolog-
ical activity, presents marked morphological changes.

4. The adjustment consequent on unilateral removal consists
not in the regeneration of the lost part, but in compensatory
hypertrophy of the remaining organ, a response which has long
been known to occur in the adult kidney and in other glandular
organs, both paired and unpaired.

5. The area of the secreting surface in the hypertrophied
kidney shows an increase of over 100 per cent when contrasted
with the normal (2.037 sq. mm.; 1.007 sq. mm.).

6. The cubie content of the mass of cells constituting the
hypertrophied kidney as shown by their relative weight is
increased 63 per cent above the normal.

7. The length of the tubules shows an increase of 21 per cent.

8. The number of nuclei in the hypertrophied kidney exceeds
that of the normal by 16 per cent, due to the occurrence of a
small amount of hyperplasia. |

9. In single, as well as in double, extirpations, the glomerulus
develops normally in the absence of the pronephric tubules.

REMOVAL OF PRONEPHROS OF AMBLYSTOMA 383

10. Anterior and posterior nephrostomal funnels are regen-
erated from the coelomic endothelium in a large proportion of
operated embryos.

11. The segmenta’ duct on the operated side shows great
variation in development, ranging from a condition in which the
lumen, though small and flattened dorsoventrally, appears
throughout the entire length, to a condition where only the
occasional presence of a few degenerating cells indicates the
location of the atrophied duct.

12. Increased activity of a single kidney also has a definite
effect on the segmental duct of the same side. Cross-sections
of the duct of an individual with unilateral operation, when com-
pared with either of the ducts of a normal larva of the same
age, show a marked increase in diameter.

13. In embryos from which the heart rudiment has been
removed in a very early stage, the initial development of the
glomeruli is normal. Subsequent distention of the aorta tears
the cells apart and they soon lose their identity as lateral capil-
lary tufts. —

384 RUTH B. HOWLAND

LITERATURE CITED

Byrnes, EstHer F. 1898 On the regeneration of limbs in frogs after the extir-
pation of limb rudiments. Anat. Anz., Bd. 15.

Derwiter, S. R. 1918 Experiments on the development of shoulder-girdle and
the anterior limb of Amblystoma punctatum. Jour. Exp. Zod6l., vol.
25; N10. 2:

Fietp, H. H. 1891 The development of the pronephros and segmental duct in
Amphibia. Bull. Mus. Comp. Zoél. Harvard. Coll. 21.

Firprincer, M. 1878 Zur vergleichenden Anatomie und Entwicklungs-
geschichte der Exkretionsorgane der Vertebraten. Morph. Jahrb.,
Bd. 4.

GaLrorti, G., UND Vinia-Santa, G. 1902 Uber die kompensatorische Hyper-
trophie der Niere. Beitr. z. path. Anat. u. allg. Pathol., S. 121-141.

Harrison, R. G. 1904 An experimental study of the relation of the nervous
system to the developing musculature in the embryo of the frog. Am.
Jour. Anat., vol. 3.
1918 Experiments on the development of the fore limb of Amblystoma,
a self-differentiating equipotential system. Jour. Exp. Zodl., vol. 25.

Hertwie, O. 1911 Die Radiumkrankheit tierischer Keimzellen. Arch. mikr.
Anat., 2te Abt., Bd. 77.

How.tanp, Rurx B. 1916 On the effect of removal of the pronephros of the
amphibian embryo. Proc. Nat. Acad. Se., vol. 2.

KAMMERER, P. 1905 Uber die Abhiingigkeit des Regenerationsvermégens der
Amphibienlarven von Alter, Entwicklungsstadium und spezifischer
Grosse. Roux’s Archiv, Bd. 19, 2. Heft.

KirrLeson, Jonn A. 1920 Effects of inanition and refeeding upon the growth
of the kidney of the albino rat. Anatomical Record, vol. 17.

Kocus, W. 1897 Versuche iiber Regeneration von Organen bei Amphibien.
Arch. mikr. Anat., Bd. 49.

Levi, Gruseprpe 1905 Lesioni sperimentali sull’ abbozzo urogenitale di larve
di Anfibi e loro effetti sull’ origine delle cellule sessuali. Roux’s Archiv,
Bd. 19.,

McCuovre, C. F. W. 1919 Experimental production of edema in larval and
adult anura. Jour. Gen. Physiol., vol. 1, no. 3.

Miiier, JOHANNES 1829 Ueber die Wolff’schen Kérper bei den Embryonen
der Frésche und Kréten. Meckel’s Arch. f. Anat. u. Physiol., Jahrg.
1829, S. 65-70. Taf. III.

Pricz, G. C. 1910 The structure and function of the adult head kidney of
Bdellostoma stouti. Jour. Exp. Zodl., vol. 9.

SacerpoTTi, C. 1896 Ueber die compensatorische Hypertrophie der Nieren.
Virchow’s Archiv, Bd. 146, S. 267.

Swinete, W. W. 1919 On the experimental production of edema by nephrec-
tomy. Jour. Gen. Phys., vol. 1, no. 5.

Wotrr, M. 1900 Die Nierenresektion und ihre Folgen. Berlin. Hirschwald.

WirticH, von 1852 Beitrige zur morphologischen und histologischen Entwick-
elung der Harn- und Geschlechtswerkzeuge der nackten Amphibien.
Zeitschr. f. wiss. Zool., Bd. 4, S. 125-167, Taf. X, XI.

PLATES

ABBREVIATIONS
A, anterior PC, pronephrie coil
AF, anterior funnel PF, posterior funnel
LT, longitudinal tubule SD, segmental duct
P, posterior
hr
385

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 3

PLATE 1
EXPLANATION OF FIGURES

Model of young normal pronephros, ventrolateral view.
Model of young normal pronephros, dorsolateral view.

386

125.
x 125.

REMOVAL OF PRONEPHROS OF AMBLYSTOMA PLATE
RUTH B. HOWLAND

PLATE 2
EXPLANATION OF FIGURES

18 Model of hypertrophied pronephros, ventrolateral view (PN 7).
19 Model of normal pronephros, of same age as figure 18, ventrolateral view
(PINEZ,"d)3

REMOVAL OF PRONEPHROS OF AMBLYSTOMA PLATE 2
RUTH B. HOWLAND

5389

PLATE 3
EXPLANATION OF FIGURES

20 Model of hypertrophied pronephros, dorsolateral view (PN 7).
21 Model of normal pronephros, of same age as figure 20, dorsolateral view
(PN 7, dd;

390

REMOVAL OF PRONEPHROS OF AMBLYSTOMA PLATE 3
RUTH B. HOWLAND

era
ke

Pil

eed

—s

ON OF FIGU

€

ATI
pronephros

; EXPLA
Ss

ypertroph

ion

22 Sect

cathe mona aatie
a x =
are

Fee Eke qe ry hb att ae ot ge ee oar |

REMOVAL OF PRONEPHROS OF AMBLYSTOMA PLATE 4
RUTH B. HOWLAND

nee

23 “Section of normal pron

wy
”
.
' »
: t
* ua ‘ 9 are
. 5 he
~ 4
é a cee
Tee ee .
#
tn -
’ b. =
Seay
: ried
i <e oe
-
Cae <P
—_ - ee
. 4 bs
‘ a“ ’
- - prt ay
Ms - ~ -
~ 8 enced
»
'
. 4
:
:
x 3 A
4 nS
> oe i. 4
>. .
‘
+
= y . -

PLATE St.) oo

EXPLANATION OF FIGURE. ae el

ephros (PN 7, d). SNe ie

oa ‘
5 s €
~
o =
:
= eae
.
‘
‘
~
'
5 ‘ .
Man
t ’
' i
* ra =
*
.
* a a
‘ ‘ ‘ :
394 3 - = ee
; .
4

REMOVAL OF PRONEPHROS OF AMBLYSTOMA PLATE 5

RUTH B. HOWLAND

395

Resumen por los autores, W. A. Kepner y W. Carl Whitlock,
Universidad de Virginia.

Reacciones alimenticias de Ameba proteus.

En este organismo existen dos tipos generales de reaccién en
presencia de los alimentos: (a) Cuando la presa no puede escapar
la amiba la rodea estrechamente; (b) Cuando puede escapar la
amiba corta su retirada envolviéndola con sus pseud6podos, y
entonces la presa queda capturada. Estos dos tipos de reaccion
alimenticia no son fijos, sino que varian notablemente. Al
reacclonar en presencia de un objeto que se mueve general-
mente en un plano horizontal, la amiba rodea la presa primero
en este plano, y después corta su retirada en un plano vertical.
Generalmente una reaccidén tiene lugar mediante cooperacién
del ectoplasma y el endoplasma, aunque el primero por si solo
puede llevar a cabo una reaccién del segundo tipo. Tanto el
ectoplasma como el endoplasma son muy contractiles cuando
las condiciones lo exigen.

La fragmentaciOn de un animal como Paramoecium en dos
pedazos es primariamente un proceso fisico y no quimico, y la
digestion comienza después que la presa ha sido fragmentada.
El proceso de la ingesti6n del alimento es reversible. Alimento
medio ingerido, casi ingerido o completamente ingerido puede
ser expulsado. Las reacciones de Ameba difieren de los fené-
menos fisicos y quimicos en que son cualitativas mas bien que
cuantitativas y se llevan a cabo en interés del organismo.

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHOR’S ABSTRACT OF THIS PAPER {SSUED
BY THE BIBLIOGRAPHIC SERVICE, FEBRUARY 7

FOOD REACTIONS OF AMEBA PROTEUS

WM. A. KEPNER AND W. CARL WHITLOCK
University of Virginia

SIX PLATES (TWENTY-ONE FIGURES)

The observations presented in this paper are selected from
many that have been recorded by members of the staff of this
laboratory. Some of the observations were taken from speci-
mens in Petri dishes, some in uncovered drops, still others under
cover-glasses, while many were secured from amebas that had
been kept in hanging drops until time was available for making
observations.

We have found Ameba proteus reacting to two types of food.
The first type embraces the following forms: desmids, Mouge-
otia, quiet Oscillatoria, encysted Chlamydomonas, and bacterial
gleas; while the second group of food bodies comprises flagellates
like Chilomonas, Peridinium, and Euglena, ciliates like Parame-
cium caudatum, Colpidium, Cyclidium, and rotifers. The first
of these groups of food objects is characterized by being non-
motile, the second group by being motile. Some of the non-
motile objects give off oxygen, while others give off carbon diox-
ide. The same may be said of the motile group; it, too, may be
subdivided into the forms that give off oxygen and those that
yield carbon dioxide to the surrounding medium. Therefore the
most conspicuous difference between these two groups of food
is the non-motility of the first group and the motility of the sec-
ond group.

Correlated with this conspicuous difference between the two
types of food of Ameba there is a two-fold food reaction on the
part of these rhizopods. Ameba’s conduct toward non-motile
food is much less complex than its conduct toward motile food.
The less complex type of reaction is concerned with ingesting
forms that do not set up currents in the surrounding water and

397

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, No. 3

398 WM. A. KEPNER AND W. CARL WHITLOCK

that do not present the contingency of escape. ‘The more com-
plex type of food reaction of Ameba is concerned with the cap-
ture of forms that set up currents in the surrounding water and
that do present the contingency of escape. It was interesting to
us to find that Leidy (’79) had shown in his figure 5, plate 1, two
Urocentra captured according to our second type of food reac-
tion, while a green plant cell was ingested apparently by the first
type of food reaction.

EXAMPLES OF THE FIRST TYPE OF FOOD REACTION

a. Objects that yield oxygen to the water

On March 15, 1919, we found that in a Petri dish there were
many filaments of Oscillatoria that were quite quiet. None of
these filaments were to be seen moving as Oscillatoria filaments
frequently do. An ameba had ingested one end of one of these
quiet filaments (fig. 1). In endeavoring to take this specimen
from the Petri dish on to a slide, the capillary canula of the
pipette dragged over the free end of the algal filament in such
manner as to tear the ameba from the substratum and turn it
through about 180 degrees. The ameba was now given time to
fix itself again to the bottom of the dish. The free end of the
filament was then pushed against with the canula of the pipette.
This time instead of the ameba’s being torn from the bottom of
the dish, the part of the ameba’s body that surrounded the fila-
ment was bent from position a to position 6 (fig. 2.) The ameba
was now drawn up into the capillary pipette and transferred to
a hanging drop. The compound microscope showed that, de-
spite this relatively rough handling, the ameba yet held on to the
filament of Oscillatoria. Within two minutes after the cover-
glass was placed over the glass ring, the Oscillatoria was egested.
Soon after this the ameba again ingested one-third of the length
of the filament and then threw it out a little later. The ameba
a third time set to ingesting the plant. When about one-fourth
of the filament was within the body of the ameba, a paramecium
collided sharply with the projecting end of the alga at right

FOOD-REACTIONS OF AMEBA PROTEUS 399

angles, and the ameba then gave up its efforts to lay hold of this
food. Throughout all of the time that the ameba was working
on this rather long filament of Oscillatoria it had within its body
a filament of Oscillatoria that was 30u long when first seen.
Within the course of our observation, this ingested filament was
broken up into three pieces, one 10y, one 5u, and one 1duz.

After we had secured several observations showing that the
ameba laid hold of quiet Oscillatoria filaments tightly, we called
in some of our colleagues in this laboratory to make observations.
Six others verified our results by making similar observations.
The most conspicuous of these corroborative records was made
by Dr. I. F. Lewis. He was given an ameba that had ingested
an end of a very long filament, indicated as broken off in figure 3.
He took a fine glass rod and bent the plant to contour b (fig. 3),
at which point the tension of the alga caused it to spring back as
a straight rod. ‘‘Twenty big bends, some like this, others differ-
ent, were made as the ameba gradually lost its hold.”

No such large filaments have been ingested wholly. The
ameba sometimes travels from end to end along such long objects,
sometimes making several trips, and then leaves the food behind.
Frequently small fragments have been seen in different stages of
digestion within amebas (fig. 5, O). It would seem that the
ameba seeks the planes of fission of the Oscillatoria filaments to
break off fragments for food. Such may not be the case, how-
ever, for we have seen an ameba travel along a Mougeotia fila-
ment in a similar manner, and there are no fission planes in
Mougeotia. No Mougeotia filaments or fragments were ever
seen completely ingested. Large desmids were also ingested in
part and then rejected. On one occasion, January 28, 1919, two
amebas began to ingest opposite ends of a large Penium syn-
chronously. The lower half of the desmid was ingested within
twenty minutes by one of the amebas. During this period the
upper ameba ingested about one-third of the desmid. Both
amebas were closely embracing the plant, but they eventually
rejected the object by withdrawing from it. Small desmids have
been observed by us being ingested by Ameba proteus. A Chla-
mydomonas within its gelatinous sheath was also ingested. Nei-

400 WM. A. KEPNER AND W. CARL WHITLOCK

ther the small desmids nor the encysted Chlamydomonas rolled
before the advance of the ameba, and they too were ingested
_ within a closely fitting food vacuole. The observation, based
upon the ingestion of an encysted Chlamydomonas, makes an
interesting contrast with Jennings’s (’04) observation of an
ameba ingesting an encysted Euglena. In the latter case the
encysted alga rolled ahead of the advance of the ameba, and
here Jennings saw a cup form behind the algal cysts. This con-
trast between our observation and that of Jennings suggests that
even the type of reaction involved in ingesting non-motile objects
may be modified to meet an unusual turn of events. There are
some non-motile food objects which give off carbon dioxide. Of
these, bacterial gleas form common examples. February 16,
1918, Dr. R. D. Mackay observed an ameba glide over a glea.
As it was about to leave the glea, two embracing pseudopods
were sent out about the bacterial mass. These pseudopods lay
close up to the sides of the rounded mass and eventually con-
stricted a small portion from the glea as the enclosing pseudopods
began to converge (fig. 4, a and b).

In the above reactions we have the ameba responding to non-
motile objects that gave off either oxygen or carbon dioxide. In
reacting to this class of food, the amebas seized the objects in an
intimate embrace.

The following constitute a list of motile objects to which Ameba
proteus has been seen reacting: a) Euglena viridis, Peridinium,
and diatoms; b) rotifers, Paramecium caudatum, Urocentrum,
Glaucoma scintillans, Colpidium, and two species of unidentified
ciliates, Chilomonas, Codosiga, Euglena acus, and two species
of unidentified flagellates. Of this group of motile food objects,
a) forms a subdivision of forms that give off oxygen, while b)
forms a subdivision of forms that yield carbon dioxide to the
surrounding medium.

The reaction of ameba to diatoms has been rather indefinite.
The ameba seems to react to these motile plants as if they were
non-motile. We have, however, obtained but two observations
based upon diatoms, and in both of these cases the diatoms,
while they were being intimately embraced, escaped.

FOOD-REACTIONS OF AMEBA PROTEUS 401

Except for the diatoms, we have seen that there is a wide range
of motile food bodies to which Ameba proteus displays a general
type of response. ‘The following observations have been chosen
as examples of the ameba’s second type of food reaction and also
to display the range of adaptive modification this type of reac-
tion may present.

The ameba seems to have a marked preference for Chilomonas
paramecium. It will readily accept one of these little ciliates,
though it has been feeding on a non-motile object or other motile
objects. On March 19, 1919, we observed a specimen that had
been feeding upon Oscillatoria. A Chilomonas swam into a bay
between two stout, short pseudopods and lay in the position
shown in figure 5. The ameba immediately sent two secondary
pseudopods, A and B, out toward each other and behind the
Chilomomas. These pseudopods met and fused; the ciliate was
thus surrounded on all sides. It was next overarched by a thin
sheet of ectoplasm. When all lines of retreat were thus cut off
from the Chilomonas, the ameba reduced the size of the large
vacuole, within which the prey had been captured, to that of the
usual food vacuole. Both ectoplasm and endoplasm entered the
formation of the pseudopods A and 8B in this reaction. ‘This is
the manner in which the enclosing pseudopods are usually con-
structed. But even the structure of the pseudopods may be
modified to meet the needs of a peculiar situation.

In one instance we observed an ameba approach two Chilo-
monases in the shallow margin of a hanging drop. In this case
ectoplasmic pseudopods a and a’ were sent out about the Chilo-
monases (fig. 6). As a grew down to contour b, an overarching
layer of ectoplasm, c, was formed above the prey. The internal
margins thus formed eventually fused as b grew down to divide
the enclosed space into two food vacuoles. The animal then
moved out into deeper water. The unusual feature of this reac-
tion is not that the overarching protoplasm is ectoplasmic, for
that and the underhanging wall of the forming food vacuole are
usually ectoplasmic. The unusual feature is the fact that the
ectoplasm formed all sides of the forming food vacuoles. These
vacuoles were thrown into the endoplasm when the animal moved

402 WM. A. KEPNER AND W. CARL WHITLOCK

out into the deeper regions of the drop after capturing the two
flagellates.

During the course of this observation it was noticed that an
ameba does not of necessity react to an object that is setting up
currents in the surrounding water or that is colliding with the
ameba repeatedly; for before, during, and after the reaction of
the ameba to the above two Chilomonases, a very active, dense
swarm of bacteria plied to and fro against the side of the ameba
making frequent contacts with it. At none of these contacts
did the ameba react to this highly motile mass. It mattered not
whether the contact were made at an angle between pseudopodia,
as at A, figure 6, or at the tips or sides of the pseudopods.

A newly formed pseudopod that is taking part in the forma-
tion of a food vacuole may further react to cooperate with a
part of the body proper to construct a second food vacuole.
That such is the case is shown by the following example. Two
Chilomonases were being surrounded by pseudopods a and a’
(fig. 7). When a had grown to contour 6, a third Chilomonas
came up by the side of a’. In reacting to this third Chilomonas,
the body proper threw out pseudopod c’, while pseudopod b sent
out c to meet c’. In this manner all three flagellates were
captured.

On March 19, 1918, we saw an advancing pair of pseudopods,
a and 6b, encounter a relatively large piece of foreign matter as
they advanced about a Chilomonas which lay in position indi-
cated in figure 8. At this synchronous contact of the two
pseudopods the one, b, was arrested while a advanced to contours
c and d, d finally fusing with the body proper. The Chilomonas
was next overarched and captured.

Perhaps a more striking example of a reaction involving for-
eign matter is presented in our observation of an ameba ingesting
a paramecium that lay in a shallow bay by the side of a large
brown mass of detritus (fig. 9). The ameba was advancing in a
general way toward the paramecium along pseudopods 1, 2, and
3. As it approached the ciliate, pseudopods 1 and 2 widened
and partly fused to form a large bi-lobed extremity, m-—m’.
When this extremity had nearly touched the paramecium, it sent

FOOD-REACTIONS OF AMEBA PROTEUS 403

out a small secondary pseudopod, a, beneath the prey, and }
anterior to it (fig. 10). When the pseudopods a and b came in
contact with the detritus, they moved apart and became much
stouter (fig. 11). In the meantime a third pseudopod, c, appeared
projecting from between a and b over the dorsal side of the para-
mecium, while a pocket was formed within the body proper of
the ameba at the bases of these three pseudopods. The para-
mecium first jumped to position 2, figure 11. The excited para-
mecium next backed into the pocket of the body proper, 3, and
a, b, and c closed in and surrounded it completely.

Usually ameba reacts to a free-swimming Euglena viridis by
sending out pseudopods that widely embrace it. Sometimes,
however, the embracing pseudopods close in upon the Huglena
to hold it in a tight grip behind the position of the gullet, and
this though the flagellum be quite active. On March 17, 1919,
we saw a Euglena caught in this manner at its anterior end.
The projecting part of the flagellate’s body was passive, but the
fiagellum was very actively lashing within the enclosed bay.
All movement for the time being had ceased in the gripping:
pseudopods. This observation had lasted for but a minute more:
or less when a large Paramecium, coming up at right angles
to the Euglena, collided with it at the point indicated by the
arrow in figure 12, and dragged the Euglena free from the ameba’s
grip. This was apparently the first step in the process of chang-
ing the second type of reaction into the first type. Mr. C. O.
Dean, a student in this laboratory, observed an ameba that had
thus gripped a Euglena viridis and thereby cut off its chance of
escape. After the ameba had thus laid hold of the Euglena, its
“eetoplasm flowed out around the Euglena” on all sides and so
close to the wall of the Euglena that there was no water present
between ‘‘the surfaces of the two organisms.”’ This is not com-
parable to the food-taking by means of invagination as Prenard
(05) and Grosse-Allermann (’10) have described for Ameba
terricola.

In 1900 the senior author observed a relatively small ameba
ingest a relatively large Paramecium caudatum. In this case
the ciliate was surrounded by pseudopods that were sent out.

404 WM. A. KEPNER AND W. CARL WHITLOCK

about it, but not touching it, about as Blochmann (’94) and
Mast and Root (’16) indicate to be the usual method of ingesting
paramecium. ‘The latter authors saw some very interesting
exceptions to this method of swallowing paramecium. We, too,
have observed departures from this type of reaction. On May
2, 1919, we had a hanging drop in which there had been many
Colpidia, but which were now dying off. ° The dead ones, though
frequently encountered by the ameba, were not in any case
ingested. The living Colpidia were frequently accepted in wide
embraces. The paramecia in this hanging drop were peculiar
in that they were wider than normal ones and rather sluggish.
Then, too, their bodies were so pliable that an ameba’s pseudo-
pod, advancing against the dorsal side of one of them, would
indent it. Moreover, when the paramecia were crowded between
two amebas, they became greatly flattened and even in some
instances bent upon themselves at right angles. The cilia and
contractile vacuoles of these peculiar paramecia were active.
The amebas attacked these relatively inactive paramecia over
and over; but in each instance their attack was peculiar in that
they attempted to surround these ciliates closely or intimately.
Because of this unusual method of attempting to capture the
paramecia they caught none, for after two-thirds or less of the
length of the paramecium’s body had become involved in the
embrace of the ameba, the paramecium would slowly glide out
and remain by the side of the ameba until it would again be
partially enclosed in a second embrace, when it would move out
of the enclosing arms of its would-be captor. The conduct of
the amebas toward these unusual paramecia is itself peculiar
and exceptional. Here for some reason the ciliary disturbance
of the water by the paramecia has not resulted in stimulating
the amebas in such manner that they sent out about the prey
remote encircling pseudopods.

A further departure from the usual method of ameba in cap-
turing paramecium was observed March 19, 1919. This ameba
was first seen at 10:10 a.m. It was then perfectly quiet, spending
all of its available energy upon the partly constricted para-
mecium. The ameba showed no cytoplasmic movement (fig.

FOOD-REACTIONS OF AMEBA PROTEUS 405

13). At 10:15 a.m. the paramecium was further constricted and
the ameba quiet (fig. 14). At 10:25 a.m. the cytoplasmic isth-
mus of the paramecium’s body was stretched and the ameba dis-
played a little movement along pseudopod c and threw out pseu-
dopods a and 6 about the projecting portion of the paramecium,
the cilia of which were quite active (fig. 15). Pseudopods a and
b were soon withdrawn. At 10:35 a.m. the constricting and
stretching of the isthmus of cytoplasm were increased and the
isthmus was flexed (fig. 16). By 10:48 a.m. the flexing of the
enclosed cytoplasm had become very conspicuous (fig. 17). Two
minutes later a pseudopod, d, was sent out along one side of the
projecting lobe of the paramecium’s body, the cilia of which were
yet quite active. This secondary pseudopod was at once with-
drawn, while the paramecium was further stretched and bent.
At this phase of the reaction a Cyclidium darted into the field
and lay near the free end of a large ‘anterior’ pseudopod. The
ameba reacted to this animal at once by sending out pseudopods
e and f and capturing the smaller ciliate (fig. 18). The para-
mecium was now released by the ameba as it ingested the Cy-
clidium. The constricted, elongated portion of the mutilated
paramecium shortened greatly and the large ciliate swam off
under its ‘own steam,’ having a contour about like the outline
given in figure 19. No trace of cilia could be seen on the part
of the paramecium’s body that had been ingested by the ameba.

An ameba may ingest food at different parts of its body syn-
chronously. We have observed one ingesting five Chilomonases
at one time and at five different regions of its body. Moreover,
the two types of food reactions may be carried on simultaneously.
On January 28, 1919, while an ameba was ingesting a quiet fila-
ment of Oscillatoria, a Chilomonas came to lie beneath the fila-
ment at a position indicated in figure 20. The Chilomonas was
lying beneath the plane in which the filament of Oscillatoria lay.
The ameba advanced about the plant until pseudopod b was
formed. This pseudopod then sent out an encircling wall of
cytoplasm about the Chilomonas and then overarched it with
an ectoplasmic film. The space within which the Chilomonas
was thus taken was next divided into a larger and a smaller

406 WM. A. KEPNER AND W. CARL WHITLOCK

vacuole, the prey being in the smaller vacuole. The Chilo-
monas was not disturbed until it was thus enclosed within the
smaller vacuole. The filament of Oscillatoria was further in-
gested, but it was finally rejected. ‘Thus, while a reaction to a
non-motile object was being carried on, the ameba completed a
food reaction of the second type, in capturing a passive motile
object.

Chilomonases have been seen to swim in beneath unattached
regions of amebas’ bodies. In such cases, when the amebas react
positively to the flagellates, a curtain of cytoplasm is dropped
down around the prey, the lips of which turn in beneath the food
body and fuse without disturbing the Chilomonas.

Perhaps the most interesting reaction we have seen was that
of an ameba reacting to a Chilomonas that had come to lie against
the tip of a pseudopod (fig. 21, 7). The ameba sent out two
pseudopods in response to the stimulus. The smaller pseudopod
arose from the side of the parent pseudopod and a little behind
its end, while the larger secondary pseudopod came out quite a
distance behind the tip of the parent one. ‘The interesting fea-
ture of this reaction is the fact that the parts reacting to the
source of stimulation are parts least stimulated; indeed, the
greater reaction was displayed by the least stimulated part. The
quiet Chilomonas could stimulate the parent pseudopod in two
ways: either chemically by means of its metabolic by-products,
or physically by means of slight vortices that the play of its
flagella may set up. In either case the end and not the sides of
the parent pseudopod would be most affected by these stimuli.
Moreover, we have studied the types of vortices set up in the
water by quiet Chilomonases. This study showed that in all
cases the strength of the currents thus set up was greatest at the
anterior end of the Chilomonas. Finally, as the two secondary
pseudopods were coming out by the sides of the Chilomonas, a
second Chilomonas came to lie at position 2, figure 21, and thus
double, or at least increase, the sources of stimulation; but this
did not modify the conduct of the two secondary pseudopods.
These facts indicate that the ameba’s reaction is a qualitative
and not a quantitative one.

FOOD-REACTIONS OF AMEBA PROTEUS 407

DISCUSSION

It has been the tendency of recent work on the ameba to
reduce the conduct of the ameba to simple terms. For example,
Loeb (05) says—‘‘As a criterion for ‘living matter’ we might
use the irritability or spontaneity. But as the ‘spontaneity’ of
living matter is in its simplest form (in Amoebae) apparently not
different from the physical phenomenon of spreading, for this
criterion the limits of divisibility of living matter coincide with
the limits of purely physical phenomena”’ (p. 321). McClendon
(09) tries to explain food-getting of ameba in the following man-
ner: ‘‘Chemical and physical influences of the medium cause a
hardening and shrinkage (by loss of water) of the ectosare
(Rhumbler’s ‘Geletinisrungsdruck’). Chemical processes within
prevent this hardening from extending to the endosare, and dis-
solve portions of the ectosare that are displaced inward. The
medium affects different portions of the surface to different
degrees, causing regional differences in degree of hardening and
shrinking, thus producing amoeboid movements. <A food body
being protoplasmic and therefore similar to the substance of
the Amoeba might, in lying near an Amoeba, protect it from
outside influences. The protected region would become more
fluid, and shrinkage of other regions of the surface would press
it out toward the food until it touched it. The food would be
pushed along and sometimes rolled over and would rub on the
surface of the pseudopod producing mechanical stimuli of suffi-
cient frequency to cause a local shrinkage of the ectosare. This
stimulus would spread through the protoplasm but being very
weak and rapidly growing weaker would cause the contraction
of only asmall area. Beyond the contracted area the protoplasm
would continue moving toward the food and surround it from
the sides”? (pp. 268-269). But our observations indicate that
the movement of an ameba about a food particle has little in
common with ‘‘the physical phenomenon of spreading’ and
demands more than surface phenomena for their explanation of
the food reactions of Ameba proteus. Of very recent date the
students of the ameba have quite given up the idea that the

408 WM. A. KEPNER AND W. CARL WHITLOCK

phenomena of spreading and surface tension are adequate for the
explanation of the movement and food getting of this rhizopod.
Hyman (’17) says “I am in favor of Clowes’s” (16’ a, 716 b)
“interpretation, already referred to, that in passing from the
surface of the protoplasm to the interior, a reversal of phase
occurs, the colloidal material forming the outer phase, or dis-
perse medium, in the surface layers, while in the interior it forms
the disperse phase and water containing a variety of materials
in solution and suspension, is the disperse medium”’ (p. 87).

It appears to us.that even colloidal phenomena cannot be
called upon to explain the phenomena involved in the food reac-
tions of this ameba, because of their qualitative character. The
qualitative nature of these phenomena becomes apparent when
we compare the reactions of ameba to various quiet paramecia.
In these reactions there appears a marked disparity between
their variability and the degree of variability of the stimuli
arising from the quiet animals that are about to be captured.
An ameba may react to a quiet paramecium in three ways:
1) by forming a pocket within its own body within which the
ciliate will be driven (figs. 9, 10, 11); 2) by sending encircling
pseudopods about the prey and then roofing over and flooring
the enclosed space with ectoplasm before disturbing the prey,
and, 3) by closing in upon the paramecium with the advancing
tips of two pseudopods until the prey is held fast in a grip of the
pseudopods’ ends. After the paramecium is thus caught, it is
very tightly closed in upon and constricted (figs. 13 to 18).
Paramecia concerned in 2 and 3 should not in a highly variable
manner stimulate the ameba. It might be held that the feeding
vortex of a paramecium lying with its ventral side directed toward
a mass of detritus, as in 1, is of a different character from that of
a paramecium lying free in the open. ‘Then, too, the disturbance
of the water by paramecium that had been caught at its girdle
would be changed as soon as the vortex became divided by the
advancing tips of the pseudopods and thus set up a new type of
stimulation. There is weight to these possible objections that
may be raised against the idea that the ameba’s reactions toward
paramecia are qualitative.

FOOD-REACTIONS OF AMEBA PROTEUS 409

But in the case of the variability of the ameba’s conduct toward
Euglena there is less weight to these criticisms. Frequently
Euglenas are taken into wide vacuoles while their flagella are
lashing vigorously. When a Euglena is caught by the advancing
ends of an ameba’s pseudopods, as shown in figure 12, there is
no reduction in the intensity of the stimulation, for the flagellum
is quite as active as it was before the body of the Euglena was
laid hold of. In this case the water in the forming vacuole is,
if anything, more greatly disturbed than when a vacuole is form-
ing about a Euglena that has not thus been grasped. Despite
this lashing of the flagellum within the vacuole that had been
formed, in the beginning, with reference to a Euglena free to
move out of its embrace, the type of reaction was changed as
soon as the body of the flagellate was held in the ameba’s grip.
This sort of modification of the ameba’s conduct gives it a quali-
tative character rather than a quantitative one.

Finally the conduct of ameba toward Chilomonas paramecium
indicates the qualitative character of its food reactions.

Chilomonas is captured only as it lies apparently quiet. This
little saprophytic flagellate is a very active creature. Kent (’80)
says this ‘“‘animalcule rushes to and fro, though with the anterior
end foremost, at a speed too rapid almost for the eye to follow,
while at the next moment it comes as it were abruptly to anchor,
with its body perfectly quiescent and one flagellum adherent to
the glass slide or covering glass, while the other maintains a
vibratory motion” (p. 425). By placing these animals in a
mixture of India ink and water or by studying them in water
containing many non-motile bacteria, we were able to determine
the extent to which these apparently quiet Chilomonases dis-
turbed the surrounding water. By this method, together with
checks or controls of specimens in aquarium water, we were able
to determine that but a very few anchor themselves as Kent
describes. The smaller or ventral flagellum in this case lashes
the water slowly in such manner that a vortex of water arises
beyond the anterior tip of the Chilomonas, for about the length
of its body, and this is drawn down over the gullet; from here it
passes, at right angles to the axis of the body, to the dorsal side.

410 WM. A. KEPNER AND W. CARL WHITLOCK

This current gains in velocity as it approaches the gullet; its
velocity is greatly checked at the gullet, and by the time that
it has passed the width of the body beyond the dorsal side of
the Chilomonas, it has become quiet. The apparently quiet
Chilomonas in this condition, therefore, disturbs the water within
a very restricted area. The great majority of quiet Chilomonases
are anchored by both flagella. These display a vibratory move-
ment of short amplitude. But even in these cases we find the
water is not remotely disturbed by the vibrations of the anchored
Chilomonas. Not even the smallest suspended particles in the
water were disturbed if they lay a body’s length away from the
vibrating animal. There is no other variation in the manner in
which these animals disturb the surrounding water. The effect
of. Chilomonas upon the surrounding water seems, therefore, to
be rather restricted (at no time passing beyond a body’s length
from the margin of the animal) and constant.

The great variability of the reactions of ameba to the stimuli
arising from this rather restricted and constant disturbance of
the water stands in sharp contrast to the constancy of the source
of its stimulation. Kepner and Taliaferro (’13) showed that
the reactions of Ameba proteus toward Chilomonas are quali-
tative ones. Our own observations on ameba feeding upon Chilo-
monas indicate that these reactions are qualitative. For exam-
ple, in figure 5 we have represented a situation in which the
ameba’s pseudopods are traveling about the Chilomonas on all
sides at a distance of about the length of the latter’s body. It
would appear, therefore, that in this case the ameba is following
the limits of the waves that radiate from the vibrating animal-
cule. Many other examples might be given of amebas’ making
a much wider embrace of the Chilomonas, but there are some
observations to be made that are more conspicuous in their
contrast to the reaction recorded in figure 5. In the reaction
shown in figure 21 there is no relation between the extent of the
water’s disturbance by the Chilomonas and the ameba’s mode of
capturing its prey. For here the regions of the ameba’s body
that reacted by sending out secondary pseudopods lay down
beneath the tip of the broad parent pseudopod, so that these

FOOD-REACTIONS OF AMEBA PROTEUS 411

regions, if not completely protected from the waves that radiated
from the Chilomonas, were much less stimulated than was the
broad end of the parent pseudopod; in addition to this, the
smaller of these secondary pseudopods lay nearer the Chilomonas.
This should have been the larger had the reaction been a quan-
titative one. Moreover, this reaction presents another inter-
esting phase, for after it had well set in, a second Chilomonas
entered the bay between the advancing secondary pseudopods.
The amplitude of the water’s agitation must now have been
relatively greatly increased and yet the conduct of the two small
pseudopods was not altered and they finally converged, though
the stimulus had been increased. Here the qualitative character
of the reaction is displayed in a manner exactly the opposite of
that in which Kepner and Edwards (’17) saw a Pelomyxa act
in a qualitative manner toward Paramecium, where, “though
the stimulus was weakened the pseudopods continued to diverge
as they grew” (p. 394) about the remaining paramecium of the
three that were present at the inception of the reaction. Because
of the qualitative character of the ameba’s food reactions, it
appears to us that these reactions toward an animal that presents
the possibility of escape are modified with reference to meeting
that contingency. The ameba’s reactions differ, therefore, from
chemical reactions in that they are made in the interests of the
ameba and may be suspended or even reversed when its own
interests demand.

SUMMARY

1. There are two general types of reaction to food: a) when no
contingency of escape is presented by the prey, the ameba tightly
surrounds the food; 6) when such contingency is presented, a
wide embrace is made and the prey is disturbed only when
retreat is cut off.

2. These two types of food reaction are not fixed, but vary
greatly. .

3. In reacting to an object that usually moves in a horizontal
plane, the ameba surrounds the prey in this plane first and next
cuts off its vertical paths of escape.

412 WM. A. KEPNER AND W. CARL WHITLOCK

4, A reaction is usually brought about through the coopera-
tion of both ectoplasm and endoplasm, though the ectoplasm
alone may carry out a reaction of the second type.

5. Both the ectoplasm and the endoplasm are highly contrac-
tile when conditions demand it.

6. The cutting of an animal like paramecium into two is pri-
marily a physical and not a chemical process—digestion setting
in after the prey has been defragmented.

7. The process of ingesting food is a reversible one. Food
half, almost, or wholly ingested may be egested.

8. An ameba’s reactions differ from physical and chemical
phenomena in that they are qualitative rather than quantitative,
and are made in the interests of the acting organism.

LITERATURE CITED

BLocHMANN, F. 1894 Kleine Mitteilungen iiber Protozoa. Biol. Centralb.,
Bd. 14.

Crowes, G.H.A. 1916a Protoplasmic equilibrium. J. of Phys. Chem., vol. 20.
1916 b Antagonistic electrolyte effects in physical and biological sys-
tems. Science, N. S., vol. 43.

GROSSE-ALLERMANN, WiLHELM 1910 Arch. Protistenkunde, Bd. 17.

Hyman, Lissis H. 1917 Metabolic gradients in ameba and their relation to
the mechanism of ameboid movement. Jour. Exp. Zodél., vol. 24.

Jennines, H. 8. 1904 Behavior of the lower organisms. New York.

Kent, W. Savitte 1880 A Manual of the Infusoria. London.

Kepner, Wo. A., AND Epwarps, J. GRAHAM 1917 Food reactions of Pelomyxa
carolinensis Wilson. Jour. Exp. Zodél., vol. 24.

Kepner, Wm. A., AnD TaLiAFERRO, W. H. 1913 Reactions of Ameba proteus
to food. Biol. Bull., vol. 24.

Leipy, JosepH 1879 Fresh-water Rhizopoda of United States. Washington,
Dp: C:

Lors, Jaques 1905 Studies in general physiology. Univ. of Chicago Press.

McCuenpon, J. F. 1909 Protozoan studies. Jour. Exp. Zodl., vol. 6.

Mast, 8. O., AnD Root, F. M. 1916 Observations on ameba feeding on rotifers,
nematodes and ciliates, and their bearing on the surface-tension the-
ory. Jour. Exp. Zodl., vol. 21.

PRENARD, E. 1905 Fauna Rhizop. du Bassin du Leman, Généve, Kundig 1902
Arch. Protistenkunde, Bd. 6.

“413

_ THE JOURNAL OF EXPERIMENTAL ZOOLOGY, Vou. 32, No. 3

PLATE 1
EXPLANATION OF FIGURES

1 First position of specimen with two small fragments of Oscillatoria within
its body and in the act of ingesting a second filament of Oscillatoria. X 100.

2 Second position after the ameba had been turned through 180 degrees by
pushing against the partly ingested algal filament, a. The ameba was given time
to fix itself to the substratum in this position, when the Oscillatoria filament
was a second time pushed upon. This time the body of the ameba was not torn
from the substratum and turned, but the tip of the ameba bent as the filament
was pushed to position b. X 100.

3 A very long, quiet Oscillatoria filament was being ingested when the pro-
jecting part of the filament was bent about the ameba to about the position
shown in contour b. When bent to this extent, the filament would slip from the
glass rod and spring back as a straight rod. The filament was thus bent and
released twenty times before the ameba released its hold on the plant. X 100.

414

FOOD-REACTIONS OF AMEBA PROTEUS PLATE 1
WM. A. KEPNER AND W. CARL WHITLOCK

PLATE 2
EXPLANATION OF FIGURES

4 Specimen reacting to a non-motile bacterial glea by constricting it with
the pseudopods a and b. The larger lobe of the glea was ingested and broken
up and the fragments delivered to small food vacuoles within the endoplasm of
the ameba. X 200.

5 Specimen with fragments of Oscillatoria (O) in various stages of digestion
(as shown by color) within the endoplasm. The ameba is cutting off the retreat
of a quiet Chilomonas by means of the advancing pseudopods A and B. XX 100.

6 An ameba against the side of which a motile mass of bacteria (A) was play-
ing. The animal did not react to this stimulation. In reacting to some Chilo-
monases, that lay at the margin of a hanging drop in water that was shallow,
the ameba sent out ectoplasmiec pseudopods, a and a’. When a had grown to
contour b an ectoplasmic wall was thrown over the top of the flagellates. These
animals were thus caught in an ectoplasmic enclosure instead of one that was
formed of both ectoplasm and endoplasm. X 200.

7 As the ameba was sending pseudopods a—b, and a’ about two Chilomonases,
a third Chilomonas came to he along the outer margin of a’. In reacting to this
third flagellate pseudopods c¢ and c’ were thrown out. X 200.

8 While pseudopods a and b were advancing along each side of a Chilomonas,
they collided at the same time with a solid body. The growth of pseudopod b was
now inhibited, while a advanced to contours ¢ and d and finally surrounded com-
pletely the prey. X 200.

416

FOOD-REACTIONS OF AMEBA PROTEUS PLATE 2
WM. A. KEPNER AND W. CARL WHITLOCK

417

PLATE 3
EXPLANATION OF FIGURES

9,10, 11 y, margin of a mass of detritus by which a paramecium was lying.
9, shows ameba advancing toward paramecium; 10, ameba sending pseudopods
a and b forward until they made contact with detritus, after which they were
moved apart and enlarged. In the meantime a pocket was formed within the
body of the ameba into which the paramecium moved as it moved to positions
2 and 3. X 200.

12 A ciliate being captured by pseudopods a, 6b, c, d, while a Euglena has
been grasped as it was retreating from a forming food vacuole. X 200.

418

FOOD-REACTIONS OF AMEBA PROTEUS PLATE 3

WM. A. KEPNER AND W. CARL WHITLOCK

419

; PLATE 4
; or Me,
EXPLANATION OF FIGURES

13, 14,15 Figures show phases in the process of an ameba tearing a parame-_
cium into pieces. X 200. ;

420

FOOD-REACTIONS OF AMEBA PROTEUS PLATE 4
WM. A. KEPNER AND W. CARL WHITLOCK

421

PLATE 5
EXPLANATION OF FIGURES

16, 17, 18 Show process of constricting and stretching the paramecium con-_
tinued. e and f, pseudopods advancing about a ciliate. As this small ciliate
was being captured, the paramecium was released by the ameba. X 200.

19 Shows the shape the living paramecium assumed after it had been released
by the ameba and swam away. X 200.

FOOD-REACTIONS OF AMEBA PROTEUS PLATE 5
WM. A. KEPNER AND W. CARL WHITLOCK

423

PLATE 6
EXPLANATION OF FIGURES

20 As the ameba was advancing about an Oscillatoria filament, a Chilomonas
came to lie to the side of and beneath the alga. When the ameba had sent out
the large pseudopod, b, it sent down beneath the alga a part of its body which
enclosed the Chilomonas within a large vacuole before the prey was disturbed.
Both types of food reactions were here carried on by the ameba synchronously.
x 200.

21 An ameba capturing two Chilomonases which lay off the end of the parent
pseudopod. XX 200.

424

FOOD-REACTIONS OF AMEBA PROTEUS PLATE 6

WM. A. KEPNER AND W. CARL WHITLOCK

Resumen por el autor, Alfred O. Gross,
Searles Biological Laboratory, Bowdoin College.

La alimentacion y el sentido quimico en Nereis virens Sars.

Nereis virens no es un gusano carnivoro, como han creido
varios autores. Se alimenta principalmente de plantas que
crecen en las proximidades de los agujeros que habita, o de frag-
mentos de plantas transportados cerca del gusano por la marea.
El sentido del gusto o sentido quimico no juega en Nereis una
parte muy activa en el hallazgo o la seleccién del alimento.
Nereis es fuertemente quimotropico negativamente hacia los
Acidos, hidréxidos y sales. Es estimulado con mayor intensidad
por el cloruro potisico que por el cloruro sédico, al contrario de
lo que sucede en el caso de la lombriz de tierra. Esta diferencia
esta relacionada con el tanto por ciento elevado del cloruro
sddico en el agua de mar, a la cual esta adaptado Nereis.

El tegumento general de Nereis es sensitivo a la accién de la
estimulacién quimica, pero existe una localizacién o concentra-
cidn del sentido quimico en los palpos y tentaculos, cuya con-
dicién esta relacionada con la rica inervacién de estos apéndices
y la relacién de sus nervios con el cerebro. Aunque existe una
tendencia hacia la localizacion del sentido quimico, este animal
no posee receptores especializados para recibir los estimulos
quimicos. La localizacién del sentido del gusto en los palpos y
tentaculos debe explicarse mediante la existencia de alguna
cualidad especifica diferenciada del protoplasma de_ estos
apéndices.

Translation by José F. Nonidez
Cornell Medical College, New York

AUTHOR’S ABSTRACT OF THIS PAPER ISSUED
BY THE BIBLIOGRAPHIC SERVICE, MARCH 14

THE FEEDING HABITS AND CHEMICAL SENSE OF
NEREIS VIRENS, SARS

ALFRED O. GROSS

Searles Biological Laboratory, Bowdoin College, Brunswick, Maine

Nereis virens is a very common marine worm distributed
along the Atlantic coast from Virginia northward to the Arctic
regions. On the Pacific coast of America it is less common, but
there are records of its occurrence from California northward
to Puget Sound, Washington. In all favorable places of its
range it occurs under stones or in burrows in the sand and mud
of the intertidal areas.

Nereis is a very favorable animal for use in experimental work
because of its abundance and the ease with which it may be
kept alive in the laboratory for long periods of time. Since it is
commonly used in the zoological laboratories as a type for dissec-
tion, a study of its habits seems desirable.

The experimental work on the chemical sense was conducted
at the Marine Biological Laboratory, Woods Hole, Massa-
chusetts. J wish to express my gratitude to Prof. G. H. Parker
who suggested the problem and who has given me helpful
criticism.

The fishermen and clam diggers along the New England coast
believe that Nereis is dependent on the clam for its existence,
hence the common name, ‘clam worm.’ Situations favorable
for the clam are also attractive to Nereis, and as many of the
worms find their way into the interior of dead snail shells or
into the mud and sand between the two valves of the dead
clams, the layman concludes that living molluscs are preyed
upon and killed by Nereis. Zoologists, if they have any con-
ception at all of the feeding habits of Nereis, believe it to be a
carnivorous worm, whose powerful jaws are for the purpose of
capturing and tearing other marine animals. In all probability,

428 ALFRED O. GROSS

this idea has arisen from certain published statements such as
the following made by Prof. A. E. Verrill on page 318 of his
report upon the invertebrate animals of Vineyard Sound. ‘‘It
is a very active and voracious worm, and has a large, retractile
proboscis, armed with two strong, black, hook-like jaws at the
end, and many smaller teeth on the sides. It feeds on other
worms and various kinds of marine animals. It captures its
prey by suddenly thrusting out its proboscis and seizing hold
with the two terminal jaws; then withdrawing the proboscis,
the food is torn and masticated at leisure, the proboscis, when
withdrawn, acting somewhat like a gizzard.” This statement
apparently was taken at its face value, and we find it copied into
the various text-books and natural histories, of which the follow-
ing taken from the Standard Natural History (vol. 1, p. 229) is
one of many examples: ‘‘It is a very active and voracious worm
terrible to smaller animals upon which it preys capturing them
by its large proboscis which it suddenly thrusts out seizing its
victim with the two large jaws which arm the tip of its efficient
weapon of attack,” etc. Verrill’s statement has also misled
investigators who have taken it for granted that the food of
Nereis is animal.

Prof. 8. 8. Maxwell, in his paper on the physiology of the brain
of annelids, quotes Verrill, and later, on page 283, he describes
the normal feeding reactions of Nereis virens as follows: ‘‘ Wenn
man ein Stick Futter, z. B., ein kleines Stick von einem Wurm,
auf eine Nadel spiesst und vorsichtig einem normalen Wurm
reicht, kann man den Fressvorgang leicht sehen. Wenn man
das Futter den Spitzen der vorgestreckten Fihler nihert, kommt
der Wurm gewohnlich ruhig niher. Dann zieht er den Kopf
ebenso ruhig ein wenig zuriick, legt die Fiihler an den Korper
und 6ffnet den Rachen, um die Nahrung zu fassen.”

It was with the above conception that Nereis virens was a
carnivorous worm, that the author began experiments on the
sense of taste. A voracious worm whose food is other animals
would be expected to have well-developed organs of taste.
Various experiments were subsequently devised in an attempt
to study the normal feeding reactions of the worms. Entire, as

FEEDING HABITS OF NEREIS VIRENS, SARS 429

well as extracts and ground-up messes of marine worms, crus-
taceans, molluscs, fish, etc., would not tempt even a semistarved
individual to eat. Though Nereis never utilized the animal
substances provided as food, it ate freely of the sea lettuce which
had been introduced into the dishes to aerate the water. Think-
ing that possibly laboratory conditions had so altered the physio-
logical conditions of the worms, that its feeding habits were
abnormal, observations were made in the field. In no case
was a Nereis seen to prey upon living animals, but many were
observed to eat vegetable matter. To substantiate these obser-
vations examinations were made of the intestinal contents of a
number of worms collected from various situations in several
localities as shown in the following table:

TABLE 1

- NUMBER
LOCALITY OF WORMS
EXAMINED

CONTENTS OF THE ALIMENTARY TRACT GIVEN IN
APPROXIMATE PERCENTAGES

Naushon Island, Woods 35 Eel-grass, 75 per cent
Hole, Mass. Various algae, 15 per cent
Sand, mud and miscellaneous material,
10 per cent
Juniper Point, Woods Hole, 10 Nemaleon (an alga), 75 per cent
Mass. Sphacelaria (an alga), 15 per cent
Sand, egg masses and miscellaneous
material, 10 per cent

Eel Pond and Buzzards 38 Roots and blades of eel-grass, 95 per cent
Bay, Mass. Sand, mud, sponges, Bryozoans and mis-
cellaneous material, 5 per cent

Lynn Beach, Lynn, Mass. 12 Eel-grass, 55 per cent

Algae, 25 per cent

Sand and mud, 15 per cent
Miscellaneous material, 5 per cent

New Meadows River (salt), 20 Rock weed, 95 per cent
Brunswick, Maine Mud and miscellaneous material, 5 per
cent

The results of this examination show conclusively that Nereis
virens is not an animal feeder, but is primarily a vegetarian.

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32. NO. 3

430 . ALFRED O. GROSS

Furthermore, this worm is able to adapt itself to a large range of
plants for food and utilizes that which is abundant and most
convenient to its burrows. The jaws and proboscis are used
extensively in excavating burrows, but, as compared with the
earthworm, a relatively small amount of sand and mud is ingested
by Nereis. The animal materials, such as bryozoans, sponges,
and egg masses, found in the intestine were originally attached
to the plants eaten by Nereis and were probably an accidental
element of the food. The worms seemed to exhibit no preference
for eel-grass covered with bryozoans and. egg masses, nor did
they shun such material when they chanced to come upon it.

When a number of Nereis are crowded into a small dish they
may, especially if mechanically or chemically stimulated, vio-
lently thrust out the proboscis, extend the jaws, and bite the
body of a fellow worm so severely as to sever it in two parts. I
have seen a worm bite its own body in two when placed under
pressure or treated with a strong acid or alkali. In such cases
it may incidently take into its proboscis some of the flesh which
is grasped. Very often, when ejecting extracts of animal juices
from a pipette toward the head of the worm, it would thrust out
its proboscis, Just as it did when treated with an acid or alkali.
These thrusts I soon learned were not attempts at securing food,
but were acts of self-defense and, it is very probable they often
serve the worms as an effective protection against enemies as
large or much larger than itself. The feeding response is a much
more deliberate act. Is it not possible that an observation,
such as noted above, and the fact that other species of Nereis
have been reported as animal feeders, may be primarily respon-
sible for Professor Verrill’s erroneous statement, a record which
has been copied so many times without any attempt at veri-
fication?

The jaws, though not used in capturing animal prey, are
employed in tearing out bits of the plants used as food. In the
intestines of some of the larger individuals it was not uncommon
to find pieces of eel-grass or other vegetation 1 to 2 cm. in length.

A number of experiments were made with the natural food in
an effort to localize the sense of taste, but the worms showed no

FEEDING HABITS OF NEREIS VIRENS, SARS 431

consistent responses to food of any kind. They fed freely upon
sea lettuce and other plants placed in the aquaria, but the find-
ing of it was more or less accidental. Food hidden in sand,
placed in cheese-cloth bags, or otherwise concealed was, as far
as could be determined, never detected by the worms. Ani-
mals from which one or all of the pairs of cephalic appendages,
such as the tentacular cirri, palps, and tentacles, were removed,
fed and thrived as well as normal animals.

It is evident that the sense of taste, or chemical sense, of
Nereis virens does not play an important role in locating and
selecting food. It is conceivable, however, that a chemical
sense may be developed which enables the worm to detect certain
unfavorable environmental conditions of the water and mud in
which it lives.

To test the chemotropism of Nereis, simple reagents, such as
HCl, KOH, NaOH, KCl, NaCl, and NH,Cl, were used. The
worms were tested by the various methods used by Parker,
Hurwitz, Shohl, Crozier, and Irwin. on the earthworm. The
fence method used by Shohl proved to be the most satisfactory
for the experiments on Nereis. For this purpose a rectangular,
shallow glass tray was divided into two compartments by a paraffin
partition, a quarter of an inch wide. A notch about three-quar-
ters of an inch long and reaching within a half inch of the bot-
tom, was cut in the middle of the partition. The entire tray
was covered over with a thin layer of paraffin, to prevent the
liquids from wetting the walls. Sea-water was placed on one
side of the fence and sea-water containing the stimulating sub-
stance on the other side. The worms were transferred from the
individual dishes in which they were kept to the notch in the
fence by means of two paraffin-covered wooden spatulas. The
worms thus placed were free to crawl into the liquid toward
which their anterior end was directed or to withdraw into that
on the opposite side of the fence. Nereis was found to be very
strongly negatively chemotrophic to all the reagents used. The
reaction times of the worms, which increased inversely as the
strength of the stimulating substances, were recorded by means
of a stop-watch controlled by foot pressure.

432 ALFRED O. GROSS

When the worms were placed on the fence with sea-water on
one side and ordinary tap-water or distilled water on the other,
the worms quickly withdrew into the sea-water, indicating that
the latter has a marked disturbing effect on Nereis. Because of
this condition, the special substances used as stimuli were always
added. to the sea-water. It may not be safe, by this method, to
make a comparison of the relative stimulating efficiency of one
acid with another or with an alkali, hydroxide, or salt, because of
the many substances in solution in sea-water which might affect
the reagent. One can, however, make comparisons of the rela-
tive sensitiveness of the worms under different conditions. As
long as there is a constant stimulating liquid in the mixture of a
measured quantity of sea-water and a definite amount of the
chemical, it makes no difference for this purpose what the result-
ing chemical combinations and mixtures may be.

For convenience of comparison the various reagents were
made up in molecular solutions and these solutions were added
in definite quantities, by means of a burette, to the sea-water
in the following proportions.

TABLE 2
1 ee. mol. HCl to 300 ec. sea-water
lec. mol. KOH to 10 ec. sea-water
1 cc. mol. NaOH to 10 cc. sea-water
1 ec. mol. KCl to 10 ec. sea-water
1 ec. mol. NaH,Cl to 3. ec. sea-water
1 ec. mol. NaCl to + ec. sea-water

The worms exhibited a marked reaction when tested with
mixtures of sea-water and molecular solutions in the proportions
shown in the above table.

The reaction times of the worms when tested with these solu-
tions were short, but not too short to be accurately measured by
means of a stop-watch. Though these concentrations of salts
produced only approximately similar reaction times, it is inter-
esting to note that it required about fifty times as great a concen-
tration of NaCl as KCl to produce an approximately similar
reaction time on Nereis; whereas Parker and Metcalf found
NaCl to be more stimulating than KCl to the dung earthworm,
Allolobophora foetida. This striking difference is probably

FEEDING HABITS OF NEREIS VIRENS, SARS 433

correlated with the high percentage of NaCl and the low per-
centage of KCl in the sea-water to which Nereis is adapted.
The author hopes to perform experiments on this interesting
and important aspect of the problem which involves the rela-
tions of osmotic pressure and sense of taste to the stimulating
substances used as well as the relative stimulating efficiency of
the various reagents.

This paper involves only those experiments made on Nereis
in an effort to determine whether the chemical sense is localized
in certain cephalic appendages or in other parts of the worm.
For a preliminary test twenty-four worms of a uniform size
(10 to 12 cm. long) were numbered and: placed in separate finger-
bowls, each containing 20 to 25 ec. of sea-water and a small
piece of sea lettuce. The latter aerated the water and provided
food for the worm. Each individual of the whole series of Nereis
was given one test in its turn with the HCl, then one test with
the KOH, and so on until the entire set of readings for each of
the two reagents was obtained. The order of the worms in the
test was reversed each time a new series of readings was taken.
All the experiments were made under conditions controlled for
temperature and light, and, as far as possible, free from mechan-
ical stimulation. After each individual test, the worm was
rinsed in fresh sea-water before being returned to its bowl. The
sea-water containing the stimulating substance, as well as the
plain sea-water, was renewed after each set of readings, since a
small amount of the liquid was carried from one side to the other
by the worm.

After the reaction time of the worms had been determined,
the tentacular cirri, palps, and tentacles of the first twelve of
the twenty-four Nereis were removed, while the other twelve of
the series were retained in a normal condition to be used as a
control. The entire series was left undisturbed for a period of
six days, a length of time more than sufficient for the wounds
made by the operations to heal. Readings were then made as
before to determine what effect, if any, the removal of the
cephalic appendages had on the sensitiveness of the worms to
chemical stimulation.

434 ALFRED O. GROSS

The average reaction time was very much lengthened in the
ease of the twelve individuals from which the appendages were
removed, but it remained practically unchanged in the unoper-
ated animals used as a control. After these determinations
were made, the worms were put aside until the appendages
were completely regenerated. With the regrowth of the appen-
dages the sensitiveness of the worms to chemicals was restored,
as evidenced by the reaction time which became about equal to
that of the control animals and to that of the same worms before
the operations were made. The results of this preliminary
experiment indicate that certain of the appendages of the head
region are more sensitive to chemical stimulation than the gen-
eral integument of the worm. In order to determine whether
the chemical sense is shared equally by all the appendages or
more strongly developed in some than in others, the experiments
were repeated, but. with this difference, only one of the three
pairs of cephalic appendages was removed from the worms of
any one series. The anal cirri were likewise tested for their
sensitiveness to chemical stimulation. The following tables
contain the results of the tests made upon Nereis virens under
the various conditions indicated. In each case the number of
animals used, the number of readings made, and the mean of
the reaction times with the probable error is given. It is deemed
impracticable to publish the individual readings which, with the
preliminary tests, involve more than 2000 determinations. For
convenience in comparing the reaction times of the worms under
two conditions, the difference of the means of the reaction times
and the ‘significance factor’ are also given in the tables.

The significance factor involves a comparison of the means of
the reaction times including their probable errors. The value of
this factor is obtained by dividing the difference of the means of
the reaction times of the normal and operated animals by the
square root of the sums of the squares of the probable errors of
these two reaction times. As an example, take the case of the
palps in table 3 in which the worms were tested with KOH.
The reaction time of the normal animals is 9.44 seconds with a
probable error of 0.479, and the reaction time of the same set of

435

a ee eee eee

a) G GT |T8?7°0 \28°2 8h [ttt ts" ((BurL0u) Tor}WOD [she | Ze°9 | 8PF Pee he 2 oe BULIOU STOmgLOS)
S| —T |c0'0 |T2r'0 |62'8 OO | POAOUISE AROS SOY SOMO ieesn| 09 G po (eurrou) Tur yeuy
wn :
wD G |IS'T |ccr'0 |§'°6 8h fo **** ([wulLoU) [orZMOD |66Z'0 | Z8°L | SPF Bop" ((eurrow) [oryw0—D
a GL 619 |$9F'0 |9Z°cI | 09 |°°""°**** * peAourer sopoequay, |T8e°O | 249°9 | 09 ¢ |' ‘[BUIIOU suvsIo [[e ‘sep>ey UAT,
po
> T |96°0° |79F'0 |FS OT | OF |" **** *([eUALOM) JoryU0H |209°0 | 89°6 | OF Q ft tt ss ([eudtou) [oryu0D
2D 8 (SlCr | SPT 4913 | Go joc *pasoures sdjeg |62F'0 | 6 | Sz G | "°° ‘[BuUILoU suBsi0 [TV ‘sdjeg
ey
A T |96°0° |79F°0 |FS OL | OF | *"********* (TBUIXOM) Jor;MOD |20¢°0 | 89°6 | Os 9 roeerssssess (TBuLTOW) [OI}WOD
a G |€8T [€9°0 |60°OL | $% |°*** poAoutod T1419 AB[NOVZUaT, |9ZG'O | 93'S | GZ Sigh ie: ata: Saga, aes a PERCE Om
e SUBSIO ][B ‘“TAIIO IBpNOeZUIT,
m 2 3 Ke ie Z fe 3 2 z
i= a 8 cs Hq 4 cas Hq
e fe |Sbe|a8e| Be | 2: BHo| Ee | 3a | &
< B° |ete|Seh| Bo 5 Set | Bo B | &
ES RMR tae epee te tale e
da | Be] | gee 2 3 22 ey 3 =

Se a a ee ee Se i a Se eh Se ee eee
eS 1ajoM-vas [0°99 OT 02 HOM oynsajow

"99 [ “uoyvyuaou0) ‘aprcouphiy unissnjod 0) sua.ia svai0 N fo Suoyova. ay) UO spuausadaa ay} fo sy)nsO.A Bururvnzuos 9790],

§ ATAVL

GROSS

ALFRED O.

436

STP CAO M=SOIeOWOV IG | Shas om oe ([euIoU) [OUCH |7F0'0 | SBS) BF | F | ({BUILOU) [O1}U0/
—{ | 70'0 | 280'0| ez | 09 | °° °°" peAourer T4119 TeUW |G60'0 | 862 | 09 g | [euLoU suvSIO ]][B “T10 [VU
B 19e 0 86G0:0\"61sG | ESP | ((BUTOW) s[OLIMOD) TSO LO) COST | Se Wy Pe ee ge: ([BUuIOU) [oLpWOH
OP eT meee Ol SGec OOM jet rmee te poAoulel sopovjuay, |990'0 | SF IT 09 G ‘*]BULLOU SUBSLO [[B ‘SO[Iv}UI TY,
Se | ORO} BRO 0)S2e0 | nOge | ({ewIoU) JorRUOD |OTT'O] S8T| O€§ | 9 PO ([eur10U) Jor} MOH
OL MG PSO veasal GG. I> oie e ‘poaowor sdjeq | #10 | €F T | ia al, meee [vultou suRsio [Te ‘sd
Salen ROL OMeGSO RO) saz OG, Wes se. | ceca ((BULIOm) [TorUOD OTT 01) G85) Of jy Oe ee ({BUALOU) [O1}WO
¢ | #6'0 | 260°0| 90°% | go |''*' PeAOUOI JaII0 AvfMOLZUEL |T80'O | GT) $s | S fo [eto
suvsIO ]]e ‘dtd AvpNoOBVyJUIT,
a i yg S A o SI 4 4
o%. |age}/>Hs | 3 hs 33] 35 sg g
a? eae lees | = Set} ®o e 5
8 g z fo) 4 te Ss 3 = } 4 2 fl ry Fe S
oe lie ae eee 3 3 225 5 a >
5 a 3 a oe 5 =) 5 o2 Ns Ee] Z
be. (eae egeg |e te & aes ae es | &
a] ae | eee ee Geile gas lesbian

wayon-nas {0 *99 00 02 IDH “wynaajow
‘29 T ‘UOYD.UaDU0D) “prop aisojyoouphy 07 Suasia sialaN fo SwoyoDas ay} WO spuamriadxa ay) fo sqjnsas Buyurzuod ajqn],

’ ATAVL

437

SaWIL

NOIDVau Jo

NVGW

O& Oe cee tebe "s* (7BULLOU) [OI}ZUOD
GG ¢ “7 euToU suBsIO TV ‘sdjeg
O& Oe a ae 2 em “({euloU) ToLyWUOD
ez Pe Goo jeunou
SsuBSIO [[v ‘dtd a@vpnoRzUaT,
es 4
Hd q
25 z
2 ie<}
F F
°
2} "
2 >
Fs 4
a g
a a
‘i w

99 [ “UOUDAPUIIUN) =" APLLOJYI UNUOUWUD 0} SUdLIA SiadaN fO SUOIDAL AY) WO SpuaUisadxa ay} fo S]]NnSaL BuruiwjUod 2)qQD,],

om I | $41 | 99°0 |26°9T | O€ {°° °°°° "7" ([ewMA0U) JoyUOD | gO'T

a Gid= | 78ST) 2850 |1Z-6G) So. 3)" "= a peAomorisdjeg | 2650

ea}

= [En Neereyae TOMS 2) (OVA G) AEM (0 Ce alle See cg ([vuLLoW) [OA}UOD | EO'T

D —T | 190 | 980 |29° FL | Go | peAOUOL T1110 IBTNOVJUOT, | LTT

A of BEE ae 28 mr: »ak

s° «¢ |a8e|S8e|] ge | Zé S58

Bee eee ee | eel oe ee

~ Ba | 28s | age Bs r) Peps

fea] eae a 5 5 S| oS)

Be ae) Coa ipeee | me: [cee | 2

sel 7 os cm) y y ag

ob wajpn-nas fo *99 & 02 1Q* HN ppnaajow

5 G¢ WIAVL

438 ALFRED O. GROSS

animals with their palps removed is increased to 21.57 seconds
with a probable error of 1.45. Substituting the values in the
formula as above stated, we have

Zip — 9.44 = 12437) *

A879)? 4- (145/212

A significance factor greater than about 3 signifies the results
are to be considered of scientific value. No importance is to
be attached to the difference in reaction times if the signifi-
cance factor falls below 3. Furthermore, if this factor becomes
greater than 3 in the two sets of readings of the control,
then the results of the experiments become questionable, either
because of lack of care in performing them or because certain
factors, such as light, temperature, etc., were not properly kept
under control. An examination of the tables at once reveals
the fact that the palps and tentacles are so highly sensitive to
chemical stimulation that their removal causes a marked change
in the reaction times of the animals. The tentacular cirri, which
together have a much greater exposed surface than the ten-
tacles and palps combined, are sensitive to a much less degree;
indeed, in only the HCl test was there a noticeable change in the
reaction time when the eight tentacular cirri were removed. —
The significance factor in this case is only 4, so even here the
difference in reaction time becomes of doubtful value. The
anal cirri, though they are sensitive to chemical stimulation,
are not sensitive to the degree that their removal causes a meas-
urable change in the responsiveness of the worms.

DISCUSSION

The fact that the palps and tentacles are much more sensitive
to chemical stimulation than the tentacular cirri becomes of
more interest when the innervation of these appendages is con-
sidered. The palps and tentacles are supplied with well-devel-
oped nerves, which arise directly from the supra-oesophageal
ganglion or brain, whereas the two pairs of tentacular cirri are
innervated by nerves which have a very different origin. The

FEEDING HABITS OF NEREIS VIRENS, SARS 439

nerves of two pairs of tentacular cirri arise from the sub-oesoph-
ageal ganglion, and those of the others take their origin from the
circumoesophageal connectives. In the higher animals, the nerves
of special sense, such as sight, taste, ete., are directly connected
with the brain. It is reasonable to infer that in a highly organ-
ized worm like Nereis, we have the beginnings of a concentra-
tion of sense receptors into more or less limited regions which
have become secondarily but directly related to a centralized
brain. This localization of the chemical sense has not pro-
gressed to any great degree, for the whole general integument of
Nereis, though less sensitive than the palps and tentacles, is
open to chemical stimulation. The same is true with the light
sense. The general integument of Nereis is sensitive to light,
yet there is a tendency toward a localization of the light sense
in the presence of two pairs of relatively well-developed eyes.
These eyes are innervated by large nerves which connect directly
with the brain. The conditions of these sense organs in Nereis
are intermediate between those forms in which there is only the
general integumentary sense and the higher forms in which the
chemical sense is vested solely in special sense organs innervated
by cranial nerves.

Maxwell has attempted to show that the feeding responses,
that 1s responses due to chemicals or substances given off by
food, cease when the supra-oesophageal ganglion is removed.
Maxwell’s statement is as follows: ‘‘Operirte Wiirmer beachten
dagegen angebotenes Futter gar nicht, es sei denn, dass man es
unsanft auf sie wirft und sie dadurch erschreckt. Sie kriechen
uber Stiicke Futters, die in ihrem Wege liegen, als ob es Steine
oder anderes lebloses Material wire. Obschon ich diese Wiirmer
viele Wochen hindurch gehabt habe, ist es mir nicht in einem
einzigen Falle gelungen, sie zu fiittern. Mit dem Verlust des
supradsophageschen Ganglions scheint das Thier die Fahigkeit
verloren zu haben die spezifischen Reaktionen auf die chemischen
Reize, die vom Futter ausgehen, zu zeigen.”

Maxwell’s experiments show that the removal of the supra-
oesophageal ganglion changes the responsiveness of the worms to
chemical stimulation—a result which is in direct line with what

440 ALFRED O. GROSS

I have found. When the above ganglion is removed, the animal
is less sensitive to chemicals, because that part of the chemical
sense which resides in the palps and tentacles is lost. The
nerves of the tentacular cirri were left intact, as evidenced by
the fact that the cirri were still responsive to mechanical stimula-
tion. Unfortunately, Maxwell’s experiments are of less value
from the standpoint of their bearing on the sense of taste because
his observations, as before noted, are not of the feeding reactions
of Nereis. The responses he secured by holding a piece of worm
flesh in front of Nereis was merely the characteristic defensive
thrust, due to chemical stimulation or irritation. These
responses are, at best, very irregular and erratic and cannot be
used in careful comparative work. In the tests on the operated
worms Maxwell placed the stimulating substances, 1.e., pieces
of worm flesh in the sea-water containing the Nereis. Under
such conditions it is difficult to detect and impossible to meas-
ure quantitatively the effect of chemical stimulation on the
worm. ‘To such a liquid the worms soon became adapted and
not stimulated at all. That the operated worms exhibited no
feeding reactions under these conditions is perfectly obvious,
because even a normal Nereis does not feed upon flesh, with
which Maxwell tested the worms. I have found that worms
still respond to chemical stimulation after the brain is removed
if tested by the method previously described. This chemical
sense of the general integument evidently works through a
ganglionic reflex, that is, through the ganglia of the ventral
nerve cord.

In addition to the rich innervation of the palps and tentacles,
as shown by Retzius, there is an abundance of diffuse integu-
mentary sense organs to be found on these appendages. Lang-
don has shown these organs to be especially numerous on the
tentacles and on the tips of the palps. But since these organs
are also abundant on the tentacular, parapodial, and anal cirri,
their significance in connection with the sense of taste is at
least a doubtful one. From the standpoint of distribution,
the evidence is to the contrary, and I am inclined to believe
these integumentory sense organs, which are also abundant

FEEDING HABITS OF NEREIS VIRENS, SARS 441

in the earthworm, are purely tactile. Any evidence that the
so-called ‘spiral organs’ are chemical receptors is also lacking.
The localization of the sense of taste in the palps and tentacles
must be explained by some differentiated specific quality of the
protoplasm of these appendages. In Nereis there is a beginning
of the localization of the sense of taste or chemical sense, but
there have not as yet developed specialized receptors (taste
buds) for the reception of chemical stimuli.

CONCLUSIONS

1. Nereis virens is not a carnivorous worm as stated by Verrill
and others.

2. Nereis feeds chiefly upon plant life.

3. The sense of taste or chemical sense of Nereis plays a small
part, if any, in locating or selecting food.

4. Nereis is strongly negatively chemotrophic to acids, hydrox-
ides, and salts.

5. Nereis is stimulated much more by potassium chloride
than by sodium chloride—a reverse of the conditions found in
the earthworm. This difference is correlated with the high
percentage of sodium chloride in the sea-water to which Nereis
is adapted.

6. The general integument. of Nereis is sensitive to chemical
stimulation, but there is a localization or concentration of the
chemical sense in the palps and tentacles—a condition correlated
with the rich innervation of these appendages and the relation
of their nerves to the brain.

7. Though there is a tendency for a localization of the chem-
ical sense, there are no specialized receptors, taste buds, for
receiving chemical stimuli in Nereis virens.

442 ALFRED O. GROSS

BIBLIOGRAPHY

Crozier, W. J. 1916 Cell penetration by acids. Jour. Biol. Chem., vol. 24,
pp. 255-279.

Hurwitz, 8. H. 1910 The reactions of earthworms to acids. Proc. Am. Acad.
Arts and Sciences, vol. 46, pp. 67-81.

Irwin, M. 1918 The nature of sensory stimulation by salts. The Amer. Jour.
of Physiol., vol. 17, pp. 265-277.

Lanapon, F. E. 1900 The sense-organs of Nereis virens Sars. Jour. Comp.
Neur., vol. 10, pp. 1-78.

Maxwe tt, 8.8. 1897 Beitrige zur Gehirnphysiologie der Anneliden. Archiv
fiir die gesammte Physiologie, Bd. 67, S. 263-297.

Parker, G. H., anp Metcatr, C. R. 1906 The reactions of earthworms to
salts. Amer. Jour. Physiol., vol. 17, pp. 55-74.

Retzivus, G. 1895 Zur Kenntniss des gehirnganglions und des sensiblen Ner-
vensystems der Polychiten. Biol. Untersuchungen, Bd. 7, S. 6-11.

Suout, A. T. 1914 Reactions of earthworms to hydroxyl ions. Am. Jour.
Physiol., vol. 34, pp. 384-404.

Sranparp Naturat History 1885 Lower invertebrates, vol. 1, pp. 1-889.
Boston, S. E. Cassino & Co.

VerRILt, A. E. 1873 Report on the invertebrate animals of Vineyard Sound
and the adjacent waters. U.S. Comm. of Fish and Fisheries, part 1,
pp. 295-778 (Nereis, p. 318).

et pyc iia A seit: ao su arasttersyy
“het - van he oe +t : i my!
- da Neel ue Ms * : aa r 4 mn Hy omy ri he } :
eel is i
eta): oy a i ih, Aes ay aul * 7, ita
or an Oy oe Fis

TEE smbyhe bide, vai» TERE ee 5 age aged
al ve ae We tes Hee Pe ie ae | ;
eb" ALE Ni sab binif airs ial UAT IAT) ethan ia
_ i al | Tye vel it nice tin olan NTH UBTH the eee a
BL, ih plang ba 9 Mayssetig ioehivi eget a MIMO TAS abt bron s
ile

"1,

tee . a rif" Welaelie i
ee 9 nee uy i dtobdy THC NTL ee Bete Sa
2 per bis wien rit 1a eth al Wr oye

| "i $ |
; = ur oT,
oh anid: a wih BUTS rie we a a ao RTA SODA) Voy
Aa 2 oS Ah i
pi j i.

‘ ‘
J a J ‘ 5 a f ‘ 1%) ree fj ey 4, *
at \ eae APR
| i it i f wey pe)
by v t c ¥ : a) i 7 , ( .
rT) | . i i Thi j Aa
i) bin ¥ : Vale ;
, ; i Dy Z iv :
: : ’ ire ‘
- valet a P ’
4 S au 1 @ * ‘i ee f
j , 1
bate he aa Ley ekg,
j ie _— ‘ ms ;
te as pdr ral
. * = * ' bey '
mr a ia az ( y ;
ae | i _? : on wei? na
-uarin ety ad y poids ou }
ra ¢ io
; diate i } : ; ' : on
Aare ay i vy ; up § , . a i
q a f nt { {) oy 4 ‘ i. ie
" ‘ 4 ; 4
iy K teat a
er ve .
4

i a nepdluth + os hye suey Dy ii inet. ith a, ais } | ne ¥
pr lp'aiys nly: LAI! 0 LAMA AEN t=. Ps ee pul
rare Py pate bis ‘iss yn oteleisdeetiat -

Resumen por el autor, W. J. Crozier,
Rutgers College.

Sobre la historia natural de Onchidium.

El] presente trabajo contiene una descripcién de las activi-
dades de Onchidium floridanum, miembro de un grupo de pul-
monados semi-marinos notables por sus rasgos estructurales de
naturaleza enigmatica, y, como demuestra este trabajo, también
por sus costumbres no menos notables. A una discusién de
las reacciones sensoriales sigue un andlisis comparativo del
heliotropismo y especialmente de la conducta de este animal al
buscar una habitacién. El autor insiste especialmente sobre el
cardcter no adaptativo del heliotropismo en Onchidium y sobre
la interpretacién mecanistica de los fendmenos que exhibe al
buscar un albergue.

Translation by José F. Nonidez
Cornell Medical College, New York

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

ON THE NATURAL HISTORY OF ONCHIDIUM!

LESLIE B. AREY anp W. J. CROZIER

Northwestern University Medical School and Hull Zoélogical Laboratory, The
University of Chicago

CONTENTS
[AGRE CHICIT GLO 5 APNE GT RRR. che, Semper 4 Nc Ane Oe Se OR 0 Co a 443
Occurrence; ‘homing’ activities; coloration; repugnatorial glands.......... 445
SEMSOLVALCSDONSCSI ae hs eee epee ioe -scie ay chkosrebererne cae arches cto. saetetstereis efelees 459
feae\ViechsmicaitexcheahlOMs. Oe e112 too ae. - «Skee ANRENAT 22 2 = Sel aa « dees 459
PMO GCL RCIA 1 ON spt sient Sete oak MELA Se 2h oS 2) ase ceene heks Sen oy hey Sevrey A 464
SMM eMM alex CUDA LION skis she Asia ects, << w a's repens aie NS «os vuole Rich see tole Gack deste 474.
Ham OLeMiC Alege xCloAULOM 1 ae tematic sorte cists cco cue errors ie nker oe te Sere NE saps al os 47
Ontthe- analysis of the hontmg behavior.).........2 000. fh. Se 479
MICO RTO fe Mot ot. oh artis « I MER: CARRE s oo SARS Dae b oie Set See e. MeR« eee 489
ie Origimuor One brain: AM eek certs k aen,0 Sethe, vf? Qaereree hia eos se Se ee 489
2. Heliotropism and the analysis of conduct. FR ee ryt Aah Sa Sts, AEE OD
3. On the nature of heliotropic inhibition nl SE UeEEl biccel 2 omket bese AON
4, The question of persisting rhythms of behavior.................... 493
5. The comparative physiology of homing movements................ 493
SLID hla b Gtr Gere Poe BEG DOE DOH GOS Rei anircoecing Soe Horror Eyer 497
Literature ade dictate ae REE, Cit BU ch OE aI or RIA eS toh Cee aA rte 499

INTRODUCTION

Onchidium (Onchidella) floridanum Dall belongs to a group
of naked pulmonates (Pelseneer, ’01, p. 21) which, after the long
discussion concerning the obscure morphology of their respira-
tory apparatus (cf. Joyeux-Laffuie, ’82; Bergh, ’95; v. Wissel,
98), have been chiefly remarkable for their littoral marine
habitat, and, more conspicuously, for the eyes—of a structural
type unique among gastropods (Semper, ’77; Stantschinsky,
07; Hirasaka,? ’12)—developed by some of them upon their

‘ Contributions from the Bermuda Biological Station for Research, no. 126.

2 We are indebted to Mr. T. Minoura, of the zodlogical department, Univer-
sity of Chicago, for a knowledge of Hirasaka’s paper, and for his kind translation
of essential portions of its contents, as well as for his translation of Fujita’s (’97)
account of the respiration of the Japanese Onchidium.

443

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 3

444 LESLIE B. AREY AND W. J. CROZIER

dorsal surface. In connection with these strange eyes and their
supposed functional significance (Semper, ’81; Bretnall, ’19) a
few observations and some suppositions have from time to time
been recorded with reference to the behavior of Onchidium;
but knowledge of its life and habits, which we find to exhibit
some perplexing features of curious interest, has been strangely
meager; nor has proof been offered that the ‘eyes’ are indeed
photo-sensitive (cf. Crozier and Arey, 719¢e).

At Bermuda, O. floridanum is quite generally distributed along
the shores of the islands, commonly inhabiting protected loca-
tions where the intertidal shore zone is covered by a layer of
sun-bleached algae, although it is also found in intertidal habitats
of other kinds (Crozier and Arey, ’19a). O. floridanum is an
Antillean species closely resembling others of this genus found
in the Pacific (Dall and Simpson, ’01). Like most of its rela-
tives, this species is strictly intertidal; some forms have been
recorded as dwelling beneath low water.

Onchidium is a good laboratory animal. It can be maintained
in small dishes with a little sea-water for a month or longer,
even if starved. In one case several individuals were kept in
small (50-ce.) bottles, tightly stoppered, and half full of sea-
water, in which diatoms had been planted. They were alive,
active, and of normal appearance at the end of six weeks. The
mollusks were observed to creep above the water in the bottles
at irregular intervals, later returning under water and feeding
upon the plant grow “lhe there.

For experimental observations it is nevertheless desirable
that freshly collected animals be employed, and this has been
the rule in the work forming the basis of the following discussion.
Our material came, in 1914-17, from Little Agar’s Island, and
in 1918 from the shore of Dyer Island, in each instance very
close to the locations of the Biological Laboratory in these
years. In addition, much of what we have learned concerning
the behavior of Onchidium is the result of sun-baked vigils on
the shore rocks in other parts of Bermuda. The work was
initiated in 1914 by L. B. A., and has since then been extended
by W. J. C. The observations here collected pretend to no

NATURAL HISTORY OF ONCHIDIUM 445

exhaustiveness; they do, however, throw valuable light upon
some vexed questions, particularly in connection with the theory
of adaptation. It is therefore important to remark that upon
some of the peculiar features in this account of the behavior of
Onchidium our respective observations, secured in an entirely
independent manner, have been found in essential agreement at
every point of overlapping. Elsewhere we have already com-
mented on certain points in the behavior of Onchidium (Arey
and Crozier, 718; Crozier and Arey, 19a, 1919c). It is our
purpose here to bring together in a unifying way the results of
our previous communications, presenting fully the evidence
concerned and indicating its place in a more systematic account
of the natural history of this animal.

OCCURRENCE;. HOMING ACTIVITIES; COLORATION;
REPUGNATORIAL GLANDS

If, during the period of low water, on a sunny day, one inspect
the shore rocks at Bermuda laid bare by the tide, at a place
protected from the dashing of the surf, one frequently has little
difficulty in discovering numbers of small blue-black slug-like
Onchidia, about 15 mm. long or smaller. They feed quietly
upon the thin coating of felted algae or creep about in a manner
which at first sight seems aimless. Favorable localities are found
on the lee shores of smaller islands within the semi-enclosed
sounds (Great Sound, Castle Harbor, but not in Harrington
Sound, where there is practically no tidal rise and fall and where
no Onchidia occur). The animals also live, however, in bays on
the north and south shores of Bermuda, but not where they
would be directly exposed to the ocean surf.

The body of the adult Onchidium when resting undisturbed
is dome-like, oval in outline, at most 17 mm. long xX 12 mm.
broad X 6mm. high. At the anterior end, two slender tentacles,
with knobbed ends, project from under the mantle-fold. Two
large ‘oral lappets,’ the ‘cephalic lobes’ of Pelseneer (’01, p. 20),
overlie the mouth region; in creeping they are constantly in slight
contact with the substratum. The margin of the encircling
mantle-fold is serrated; certain of the marginal projections rep-

446 LESLIE B. AREY AND W. J. CROZIER

resent the location of large mantle-glands, which give rise to a
repugnatorial secretion; there are fourteen of these glands in O.
floridanum, seven on either side.

Although at about the time of low water Onchidium is often
abundant upon shore rocks (ef. Semper, ’81; Eliot, 99), and in
many places even conspicuous, during high tide it is invisible.
It never wanders above high-water mark nor beneath the water
level. In these respects it differs sharply from a much smaller
species (Onchidiella) of intertidal habitat also found at Bermuda
(Arey and Crozier, 719 a, p. 163), the latter creeping about when
covered by the sea, but being sheltered within dead Serpula
tubes, barnacle shells, and the like during low tide. Eliot (’99)
has briefly referred to the occurrence of O. tonganum on the
tidally exposed reefs at Apia, Samoa. O. floridanum also lives
upon more or less isolated rocks and islets, within Great Sound,
but does not occur on the Bermuda ‘coral’ reefs, which are not
exposed at low water.

We were soon struck by the fact that sometimes, even when
the tide was not high, no Onchidia were procurable in places
known from preceding and from subsequent observations to be
thickly populated by them. There is in fact a very exact rela-
tion between the appearance of the Onchidia upon the rocks and
the state of the tide. This relation involves some curious and
precise ‘homing’ habits, of a kind hitherto unsuspected among
mollusea (Arey and Crozier, ’18).

Onchidium lives in communities numbering a dozen or more
individuals, with pocket-like ‘nests’ in the eroded shore rock.
The openings to these nests are as a rule very inconspicuous.
The mollusks creep out of their nest only when the tide has a
little more than half ebbed, that is, about 2 to 2.5 hours before
low tide. For any particular nest the time of emergence depends
upon the nearness of the nest to the high-water mark. Those
Onchidia living in nests located nearest to the high-water level
appear in the open sooner than do those situated farther down.
It is the actual position of the cavity of the nest itself, in rela-
tion to the tidal level, rather than the location of its external
opening, which regulates the moment of emergence.

NATURAL HISTORY OF ONCHIDIUM 447

A common habitat of Onchidium is on isolated rocks which are
more or less completely submerged at high water and are covered
by a yellowish-brown felt-work of algae. On such islets, and on
more extensive protected shores of similar appearance, numbers
of tiny crevices are almost invariably lined with a layer of
Modiolus,? and these frequently contain a passageway to an
Onchidium nest. The eroded cavity in the limestone forming
the irregularly shaped nest is sometimes of the bigness of a man’s
head, though usually much smaller. The external openings of
the passageways are quite inconspicuous, for they are not only
small, but they are further masked and partially choked by the
growth of Modiolus. It is astonishing through how small an
opening an Onchidium can slowly make its way, as it insinuates
itself into the tiny spaces between the mussels. An individual
that is 5 mm. high when normally creeping can squeeze through
the space between the edges of two glass slides held 1 mm.
apart. In nature the process seems even more startling. When
a group of Onchidia emerges from the nest the individuals
appear one at a time in continuous series.

Colonies were also found established at sheltered spots where
loose stones were held together by red clay, the nest being here
a deep crevice between two stones. This type of habitat is
less common than that afforded by eroded limestone.

Following the emergence from its nest, a colony of Onchidia
wanders in various directions over the rock. Some individuals
may creep a meter from the nest. They remain exposed for a
certain length of time, and then, s¢multaneously, return directly
to the nest from which they came.

The individuals emerging from one nest sometimes become
more or less scattered, separated, and even somewhat mingled
with others derived from other nests. Among the components
of any one community, however, the coincidence of the return
to the home nest is in most cases, if not indeed in all, remarkably

3 This Modiolus is sometimes found underneath small stones and on the under
sides of rock slabs, where numbers of stones occur piled together. In such dark-
ened situations the color of the mussel is not black, but, on the contrary, retains
the reddish-brown cast of the juvenile shell.

448 LESLIE B. AREY AND W. J. CROZIER

close; one is tempted to compare this rather startling exhibition
to the effect of a thunder-storm in causing human families to
retire to their respective homes. On the other hand, the colonies
which, in any given place, are the first to appear are likewise
the first to retire, so in this way a separation of the different
communities is effected, favorable to their exact observation.
Independent study has convinced us that there is probably no
‘mixing up’ of the individuals from different communities; in
fact, we have never seen an instance in which an individual
coming from one nest and carefully watched during the whole
of its unmolested perambulations, failed to return to its original
home.

The process of return to the nest has a highly determinate
aspect. The return course is direct and as straight as is per-
mitted by the physical imperfections of the substratum. In
creeping back to the nest an Onchidium may move toward the
sun, although if removed from the rock and immediately placed
on a glass plate it will be found negatively heliotropic. Helio-
tropism has nothing to do with the direction of normal creeping
(Crozier and Arey, 719 ¢).

In the immediate vicinity of the entrance to the home cavity,
the Onchidia frequently make use of a natural groove in the rock.
In the case of small colonies, all the individuals may utilize this
trail. The members of larger colonies, however, may simul-
taneously approach the nest from several widely separated
points. Arriving at the entrance to the nest, they crowd about
it in an absurd orderly fashion, and without restlessness ‘await
their turn’ to enter. Owing to the fact that the opening of the
nest is usually quite small, several minutes may be required for
an individual to insinuate its body into the opening. Hence
the disappearance of the whole colony within the nest usually
occupies some time. Once, however, they have collected about,
or upon, the mass of Modiolus which commonly surrounds or
even partially occludes the entrance to the nest, the Onchidia,
because of the similarity of their coloration to that of Modiolus,
may readily be overlooked.

NATURAL HISTORY OF ONCHIDIUM 449

The following notes of one emergence of the Onchidia inhabit-
ing a section of Little Agar’s Island will give a sufficient general
idea of the phenomenon, which we have repeatedly observed:

July 6, 1914. A cloudy, but fairly bright day. Low tide at 11:30

A.M.

9:57 a.m. A few (4 or 5) Onchidia ‘out,’ others*in process of emer-
ence.

10:02 About a dozen out. Many at once begin to climb straight
upward. Others wander in devious paths; if on a flat rock-
shelf, may start toward the water, but do not actually go
downward.

10:12 In one cove on S. E. side of island 24 individuals seen. Over
100 seen on flat rocks at opposite side of the cove, to the west-
ward; at 10:02, only 1 or 2 were to be seen here.

10:17 A few stragglers are still appearing.

10:32 <A few (those which were the first to appear) are beginning to
retire.

10:37 Some Onchidia from nests lower down than those previously
concerned are beginning to come out. For these, the tide is
as much lower than the nest as it was in the case of those
first appearing.

11:02 Excepting a few, the Onchidia seen out at 10:12 have all
returned. At 11:09, only 3 of this first group remained out.

11:42 A community seen coming out at 10:30 has now completely

‘retired.

12:20prp.mM. About 10 Onchidia, as nearly as could be determined,

are now to be seen on the entire island.

The foregoing record illustrates some of the general features of
the appearance of the Onchidia upon the rocks and of their
return to their nests. Several further records may be cited which
exhibit the remarkable synchronous character of the return to
the nest on the part of the different individuals of a colony, even
when these individuals may be separated from one another by
a relatively considerable distance. These records are typical
of the many observations made upon this point, and no facts
discordant with them have ever been encountered.

July 2, 1914:

9:30 a.m. A bright sunny day. Only a very few animals to be seen

on Little Agar’s Island. Low tide occurred at 8:00. [Doubt-

- less the few animals visible were late comers from nests near
low-tide level, but no note was kept of this.]

9:40 Ona rock 3 Onchidia, in a group, were seen creeping back and
forth; one of these wandered two feet away from the others.

450 LESLIE B. AREY AND W. J. CROZIER

9:55 The two animals turned and went directly back toward their
nest. The solitary one, two feet away, turned back at about
the same time.

10:06 All three reached entrance of the nest, by a straight route in
each case.

April 16, 1918. Dyer Island.

Low tide 5:30 p.m. Several rocks some feet from the main shore-
line, on the southern side of Dyer Island, afforded convenient sta-
tions for study of Onchidium colonies occurring more closely grouped,
on the whole, than was the case on Little Agar’s. At the top of one of
these rocks there was a single colony, well removed from the closely
clustered nests around the lower edge of the rock. At 4:46 the mem-
bers of this topmost group began to emerge; by 4:53 the whole colony,
numbering 17 individuals, 3 to 13mm. long, had emerged. They scat-
tered in various directions over the rock; some going 70 cm. away from
the nest entrance. No animals from other nests became mingled with
them. By 5:37 one Onchidium of this group had turned and begun
to creep toward home; at 5:48 the last one of the 17 had done like-
wise, the whole number finally arriving at the entrance to their nest
by 5:55.

Hirasaka ('12) has noted that the Japanese O. verruculatum
Cuv., though easily obtainable at low tide, seemed at high water
to have ‘disappeared.’ He also records that this species is not
seen, even at low tide, during stormy weather, and suggests that
the animals retire to clefts in the rocks. This corresponds to
the behavior of O. floridanum, with the difference that in stormy
weather this species does not emerge from its nests on a given
bit of shore if the latter be exposed to wind or surf. Rain or an
overcast sky has a negligible effect, if indeed any at all, upon
the emergence of the Onchidia. It is stated that O. verrucu-
latum is most abundant from April 15th to October 15th, and
Hirasaka suggests that this species, like other pulmonates, passes
through a period of hibernation (also Fujita, ’97).

O. floridanum, in the warmer Bermuda region, does not hiber-
nate although some of our field notes suggest that on some of
the colder winter days the animals may fail to emerge. Further
observations would be necessary to settle this point.

The intertidal rock crannies inhabited by Onchidium celticum
were carefully described by Joyeux-Laffuie (82), who detected
also the fact that these animals ‘‘abandonnent leur retraite en
moyenne une heure ou une heure et demie apres que la mer a

NATURAL HISTORY OF ONCHIDIUM 451

commencé & baisser,” and that they again seek shelter in these
cavities with the return of high water. But the peculiarly
significant fact that each Onchidium returns to a particular
crevice, Joyeux-Laffuie did not discover (if indeed it occurs in
his species). He pointed out that in O. celticum, as in our
form, the attachment of the foot to the substratum is feeble, so
that the snail if exposed under water would find it impossible to
retain its footing. He states also that ‘‘Le moment ot elle sort
des fentes des rochers et celui oti elle s’y refugie sont trés variables,
suivant la température,” this species rarely emerging in winter;
moreover, ‘“‘par un temps couvert et humide, elle se promenent
beaucoup plus longtemps que par un temps clair et sec.” For
O. floridanum we can agree that temperature is probably a factor
in this matter, but the degree of atmospheric humidity has
seemed quite insignificant. According to Bretnall (719), two
species of Onchidium observed by him in Australia did not show
the possession of ‘homing habits.’ No details are given.

Further data regarding the curious homing behavior of Onchid-
ium, together with such analysis of the situation as may be
attempted on the basis of our inquiries, are best deferred until
something has been said concerning the sensory reactions of
Onchidium. Some additional features of the creature’s natural
history must first be presented.

Locomotion

The mechanism of the pedal creeping of Onchidium has been
noted by several observers (cf. Parker, 711; Olmsted, ’17a). As
in the case of many other pulmonates, progression 1s accom-
plished by means of pedal waves originating at the posterior end
of the foot and coursing anteriorly (direct waves). Ordinarily
but one of these waves, extending across the whole width of the
foot (monotaxic), is present on the foot at a time. Another
wave is initiated just as its predecessor reaches the anterior end
of the foot. Sometimes two waves are visible at one time, the
second having been commenced just before the disappearance
of the first. A wave causes the posterior extremity of the foot

452 LESLIE B. AREY AND W. J. CROZIER

to be contracted anteriorly for a distance of 2.5 to 3 mm.; the
posterior end, moving forward more quickly just before the
termination of its ‘step’ than at the beginning, is then attached;
as the wave begins to move anteriorly, the animal appears to
brace itself against its point of attachment, the posterior end of
the foot spreading out against the substratum and seeming as a
whole even to move backward slightly. The pedal wave—as
may readily be seen either when Onchidium is creeping on a glass
plate in air or when placed upon its dorsum so that righting
movements are begun—represents an area of the foot lifted out
of contact with the substratum (ef. Olmsted, 17a, p. 235). When
an Onchidium is placed on its back, wave movements appear
on the foot, and are somewhat magnified as compared with their
normal size; inspection of the waves produced in this way shows,
under the binocular microscope, that the posterior zone of the
wave concavity is undergoing longitudinal contraction in the
anterior direction, as it should according to Parker’s (11) view
of the mechanics of progression in such cases. The pedal wave
traverses a foot surface of 13 to 14 mm. length in about 5 seconds
(i.e., at a rate of 16+ cm. perminute). The rate of progression
of adult Onchidia, on a smooth surface, is found to be approxi-
mately 5 em. per minute, agreeing with that calculated from the
preceding data regarding the frequency of the waves (20 per
minute) and the distance observed to be transversed (2 to 5
mm.) as the result of a single wave (cf. Peyréga et Vlés, 713).
Like most gastropods, Onchidium creeps only in the anterior
direction; its creeping appears incapable of reversal.

Onchidium creeps over substrata of varied texture, such as
bare rock, corraline algae, or felted enteromorpha, and while
entering or leaving its ‘nest’ passes over groups of Modiolus
presenting a sharply serrated surface. In agreement with the
possibility of the mollusk’s living upon such surfaces, it is found
that the foot does not serve as a hold-fast through its action as
sucker. To some extent, as in Chiton (Arey and Crozier, ’19 a),
the mantle serves as a hold-fast. If the animal be touched, or
shaded (vide infra), the mantle is quickly applied to the rock
surface, very much as that of Chiton is. This reaction is inter-

NATURAL HISTORY OF ONCHIDIUM 453

rupted if the dorsum of the animal is pinched or prodded, since
the periphery of the mantle with its repugnatorial glands is then
elevated and caused to bend toward the focus of irritation.
The animal nevertheless continues to adhere with some firmness to
the rock. The major share of this adhesion is apparently due to
slime, asin other pulmonates. Onchidiumcan maintain a position
upside down on a glass surface when less than one-third of the
foot is in contact; it can creep over a crack between two glass
plates; it creeps undisturbed over a hole in a glass plate, the sub-
stance of the foot being pressed firmly into the vacant space.
The adhesion of the foot is due, then, in the first place, to the
close contact brought about between the foot and the rock or
other surface, and, secondly, to the slime which the foot secretes;
possibly local suction on the part of minute areas of the foot is
also concerned, as in Chiton, although this cannot easily be
tested. Usually, if not invariably, the attachment of the foot
can begin at its anterior end only, although the animal may for
some minutes maintain its weight by means of adhesion through
the posterior third of the foot surface alone.

When rolled over on its back, an Onchidium usually begins
soon to right itself. Some individuals, in the laboratory, are
quite inactive, however, and will not begin to right themselves
for along time. If turned over while under water, an Onchidium
seemingly active in other ways may remain on its back for sev-
eral hours, much longer than when in air. In righting, pedal
waves of considerable amplitude are formed and continue to pass
over the foot until righting is accomplished. When first dis-
placed, the animal tends to curl up like a Chiton, but soon
stretches out again. The anterior end is twisted and contracted
on one side until the anterior portion of the foot can be attached.
Righting then proceeds as the attachment of the pedal surface
progresses toward the posterior end, so that when half attached
the body of the Onchidium is twisted through an angle of 180°.
Under ordinary conditions an Onchidium will right itself in
15 to 30 seconds (mean, 25 seconds). It should be noted that
the back of an Onchidium is strongly arched, so that when dis-
placed from the rock, and caused to ‘curl up’ with its fringe of

454 LESLIE B. AREY AND W. J. CROZIER

poison glands projecting, the body is admirably shaped for rolling
over to one side or the other, even before twisting is begun, the
form thus tending to facilitate righting. We have never seen an
Onchidium in nature disturbed to an extent sufficient to bring
this behavior into play. It is unlikely that they are ever dis-
lodged by wave action, even though the force of attachment
through the foot is not great, since they do not creep upon the
exposed rocks under water; nor do they come out of their ‘nests’
at low tide when there is a stormy wind; nor do they on the
whole inhabit places where wave action is severe, but rather the
reverse.

In nature, Onchidium must spend a good part of each day
under water, sometimes several days continuously. Certain of
the nests may entrap a small amount of air as the tide rises, but
this cannot be the case with most of the nests. In aquaria, these
animals will live for at least two weeks under sea-water, without
visible impairment. In several instances groups of them were
forced to do so by being placed upon a stone suspended by a
string under water. A group isolated in this way stays during
daylight hours, and also for the greater part of the night, on the
under (shaded) side of the stone.

When beneath the water surface an Onchidium tends to creep
more slowly than when in air, and remains for long periods ‘at
rest,’ with girdle depressed, dorsum arched, tentacles retracted.
While creeping under water the tentacles are never so far pro-
truded as they characteristically are in air. As already noted,
the attachment of the foot to the substratum is not very firm in
Onchidium; it is, however, quite sufficient to enable the creature
to creep on the under surfaces of stones in air as well as under
water.

Onchidium emerges from its nest at low tide even though rain
may be falling heavily, provided no wind is blowing strongly
on the bit of shore concerned. It is therefore interesting to
determine the toxicity of rain-water for Onchidium, especially
since the porous rocks containing the nests of these mollusks
must frequently permit the seepage of rain-water into the Onchid-
ium shelters. Experiment showed that an Onchidium placed

NATURAL HISTORY OF ONCHIDIUM 455

in 200 ec. of rain-water would live in some eases for 4 hours (at
27°C.). When first immersed the animal partly rolls up, lying
on its side or back, and it so remains until dead. If stimulated
by touching, it momentarily writhes about, then returns to
quietness. A fair degree of sensitivity is retained for at least 3
hours in rain-water. The intertidal Chiton tuberculatus is
similarly resistant (4 hours), but is in addition protected by the
completeness with which the girdle can exclude fluids from con-
tact with the animal’s. soft tissues (Arey and Crozier, ’19).
The nudibranch Chromodoris is killed by 45 minutes’ immersion
in rain-water (Crozier and Arey, ’19b). Resistance to rain-water
may clearly be of bionomic importance, but that it has originated
adaptively is in no degree certain; more probably, it depends
upon the organization which Onchidium has inherited from
ancestors among the land pulmonates (Perrier, ’17).

Certain phenomena of coloration in O. floridanum are not
without interest, especially in view of the fact that there is
present in this snail an active system of repugnatorial glands.
We have recently published some discussion of this matter
(Crozier and Arey, 719 a), and need refer here merely to the
chief points involved.

At Bermuda this species exhibits two fairly distinet types of
pigmentation, a pale type verging upon dull olive-yellow, and,
much more abundant, a type of dark blue-black appearance.
These two kinds of coloration are found in other species of
Onchidium (cf. Eliot, 99; Dall and Simpson, ’01) and in the
related genus Onchidiella. The lightly pigmented type is often
concealingly adjusted to its background, but not in every case.
The color of the dark variety might also be considered a conceal-
ing match for the mussels? which cluster about the entrance to
its nest. But no correlation can be established between specific
substrata and the pigmentation of the Onchidia which creep over
them at low tide. In view of this fact and of the further obser-
vation that the mud-encrusted slime pellicle investing the back
of Onchidium is commonly removed in mechanical fashion as
the snail creeps out of its nest, whereas if the pellicle should

456 LESLIE B. AREY AND W. J. CROZIER

remain it would add decidedly to the creature’s homochromicity,‘
it is important to note that the powerful repugnatorial mantle-
glands are found developed to an equal degree in Onchidia of
whatever variety of pigmentation. Bretnall (’19) states that
O. diimelii and O. chameleon exhibit ‘‘the chameleon-like prop-
erty of changing their colors, especially when disturbed’’ or
placed upon a different background. No basis for such color
modification is known, nor is one described by Bretnall; nothing
of this sort occurs in O. floridanum.

These facts are incompatible with the notion that the colora-
tion of Onchidium involves adaptive restraint. Some reasons
have been given (Crozier and Arey, ’19 a) for regarding the color-
ation of an Onchidium as the result of genetic factors primarily.
It is therefore noteworthy that the breeding habits of this snail
may provide a mechanism for the perpetuation of a racial type,
lightly pigmented, which probably would behave as a recessive
in crosses with dark-hued forms. Onchidia are ‘simultaneous’
hermaphrodites, exhibiting reciprocal insemination (cf. also
Joyeux-Laffuie, ’82). Joyeux-Laffuie states that O. celticum
conjugates during its periods of emergence upon the rock. We
have never seen this in O. floridanum, and believe that copula-
tion occurs within the nests. Even if it should occur in the
open, however, each colony is in large measure prevented by its
habits from mingling with the members of other communities,
so that an appreciable degree of inbreeding may safely be pos-
tulated. Eggs are deposited during July within the nests,
attached in pearly masses to the upper surfaces of the rock
cavities, and the creature which emerges from the egg membrane
has already the form of an adult.

The intimate physiology of the peripheral glands, referred to
as ‘repugnatorial glands,’ will be considered in a subsequent
‘paper.. It is pertinent, however, to mention here some observa-
tions upon their use in nature.

Semper (’81) regarded the glands as of service in warding off
the attacks of certain intertidal fishes. According to Semper,

4It may be pointed out that the non-removal of the slime coating of the
snail’s back might interfere with dermal respiration.

NATURAL HISTORY OF ONCHIDIUM 457

the shadow of an approaching fish induced the discharge from
the glands of a liquid spray which drove off the fish. This notion
was extended by Semper so as to provide a mechanism explain-
ing the development of the mantle-eyes of Onchidium. He
considered that the distribution of those species of Onchidium
possessing mantle-eyes coincided with that of the intertidal
blenny Periopthalmus. This remarkable fish skips along the
beach zone laid bare by the falling tide (cf. figure in Hess, ’12),
and was said by Semper to prey upon Onchidium. ‘The testi-
mony of later naturalists does not favor this view. Periopthalmus
lives upon mud flats and frequents the margins of mangrove
swamps (Murray, p. 489; Eliot, 99), whereas Onchidium lives
upon rocks and along the edges of reefs, where the blenny is not
found (Eliot, 99). Periopthalmus, moreover, feeds on arthropods
(Murray, loc. cit.). Australian species of Onchidium possess
mantle-eyes, and in regions where no Periopthalmus-like fishes
are known (Bretnall, 1919). Although Onchidium is quite sen-
sitive to shading (Crozier and Arey, ’19c), O. floridanum does
not discharge its poison glands when stimulated by a sudden
decrease of light intensity; tactile excitation is the form of stim-
ulation pre-eminently successful in eliciting discharge of the
glands. The dorsal surface of O. floridanum is, however, the
part sensitive to shading, while one of the most conspicuous
features of the animal’s response to shading is the retraction of
the tentacles, the latter organs being, nevertheless, themselves
devoid of direct excitability by shading.

Neither Periopthalmus nor any other fishes of analogous habit
occur at Bermuda. It is improbable that fishes are able to enter
the majority of the Onchidium nests, because of the minuteness
of the entrances. It seems likely that it is in air, rather than
when under water, that the glands are most effective. The gland
contents are discharged in a stream which in air breaks up into
a fine, almost invisible, spray, which may be thrown as far as
15 cm., or about ten times the length of the Onchidium. Under
water, long threads of secretion are expelled, which do not form
a spray and fail to travel more than a centimeter or so from
the apertures of the glands. Tactile stimulation of the dorsal

458 LESLIE B. AREY AND W. J. CROZIER

surface of the mantle is the kind of excitation most effective for
gland discharge. Among the animals noted as frequenting the
Onchidium zone were included: isopods (Ligia, especially),
crabs (Sesarma, Panopeus, Porcellana, in some places Pachy-
grapsus, and an occasional Portunus), Chiton tuberculatus,
Coscinasterias (during its breeding season—January to February—
often left above low water, in a depression between rock slabs
and in similar places), and the mud-dauber wasps (Polistes),
which gather moist silt from cool, shaded, intertidal spots near
the entrances to caves. Unequivocal instances have been found
in which crabs, isopods, starfish, and wasps came into contact
with Onchidium; in most, if not indeed in each of these instances
there occurred immediate moderate discharge of the glands,
followed by the retreat of the animal from the Onchidium. It
must not be supposed that these creatures were endeavoring to
devour the Onchidia; rather, it seemed important for the snails
to avoid being accidentally pushed off the rock into the water, for,
as previously noted, Onchidium does not adhere with any great
firmness to the algae-covered rock. When purposely pushed
off, into the water, an Onchidium is not able to return to its
nest. It is entirely probable that most of the creatures acci-
dentally touching one of these snails would have retreated even
in the absence of the repugnatorial secretion, but the importance
of the discharge is nevertheless clear.

The gland secretion was obtained in a ‘pure’ state by holding
a glass slip over the back of an Onchidium stimulated mildly,
in air, by means of faradic shocks. The glands individually
turn their apertures dorsalward and their axes converge in such
fashion that the several discharges impinge very nearly at a
single point immediately above the site of stimulation. This
conspicuous accuracy involved in release of the gland content is
a noteworthy feature of the use of the glands. Small bits of
crab- and mussel-meat were smeared with the secretion and
were found to be vigorously rejected as food by sea-anemones,
star-fish, crabs, and fishes, including forms which could not by
any possibility have previously encountered this material, as, for
example, Fundulus from landlocked brackish ponds at Ber-

NATURAL HISTORY OF ONCHIDIUM 459

muda (Crozier, 719 b). When received upon one’s tongue the
repugnatorial spray is found to sting like wild mustard, and with
considerable persistence (Crozier and Arey, 719 a).

It is not likely that many creatures able to inflict damage
upon Onchidia can gain access to them while they are concealed
within a ‘nest’ and covered by the sea. Among those which
need be mentioned in this connection one of the most interesting
is the curious littoral chilopod Hydroschendyla.’ These rare
forms occur at Bermuda between tides, in crevices and within
the muddy interstices of much-eroded sandstone blocks. Like
their geophilid relatives, they devour annelids, for they have
been uncovered in the act of biting into the sides of Leodocid
worms which occur between the aeolean strata.’ On several
occasions Hydroschendyla has been obtained within Onchid-
ium nests when these were chiseled open. No indication was
had, however, of either symbiotic or predatory connection
between these forms—their association seemed entirely accidental.

SENSORY RESPONSES
1. Mechanical excitation

The responses of Onchidium to tactile excitation are of some
diversity, depending upon the part activated. The surface of
the foot responds by attachment when brought into contact
with a surface sufficiently large; the glands upon the periphery
of the mantle become erected and release their contents when
the mantle is stroked or pinched, but otherwise the reactions to
touch are of the more or less local and negative (withdrawing)
type. General mechanical activation by water currents induce
negative rheotropism. A kind of anemotropism, involving the

5 We are indebted to Dr. R. V. Chamberlin, Museum of Comparative Zodlogy,
for the identification of this chilopod. (Cf. Chamberlin, 1920.)

6 Geophilus has been figured wrapped spirally about the body of an earth-
worm which it had begun to devour. Hydroschendyla, however, seems merely
to bite into the body of Leodocids, whereupon the worms conveniently auto-
tomize at that place, the anterior end creeping away while the centipede sucks
the juices of the abandoned tail. It is interesting to observe that allied Schen-
dylids are known to frequent caves (Ribaut, 15).

460 LESLIE B. AREY AND W. J. CROZIER

stimulation of the tentacles, occupies a well-defined place in the
bionomics of Onchidium. Geotropism is not well defined and
reactions to vibratory stimuli are but poorly represented.

When the back of an Onchidium is momentarily touched, a
slight local depression is formed. The dorsum is not very active
to a single light touch. If the dorsum be ‘scratched,’ how-
ever, or stroked several times in succession with a blunt point,
the particular area affected becomes to some extent contracted;
but the most obvious response is from the margin of the mantle—
the mantle-fold is erected, forming a saucer-like rim about the
body, so that the now erected repugnatorial glands come to
point in a general way toward the spot irritated.

The marginal zone of the mantle is much more sensitive to
touch than is the back of the animal. If the anterior end of a
creeping individual be lightly touched, the tentacles and oral
lappets instantly retract, the mantle-fold is depressed to the
substratum, the back of the animal becomes strongly arched,
and locomotion ceases. Three or four gentle stimulations in
succession are required to induce the completion of this form of
response, but a single touch is sufficient to bring about the expres-
sion of its initial phases. Almost immediately after the stimu-
lation has ceased, the head and tentacles are protruded from
beneath the mantle and locomotion is resumed. If the animal
is, to begin with, not creeping, but quietly attached in its char-
acteristic attitude (the head and tentacles being withdrawn, the
body then appearing oval or circular in outline), a touch causes
the edge of the mantle to be retracted and the back more decid-
edly arched. Stimulation of the posterior end of the mantle of a
resting Onchidium causes the part affected to be drawn forward
and curled under the body. In a creeping individual, if the
posterior end be touched several times in succession, the .whole
posterior third of the body is contracted, so that in outline,
seen from above, it is pear-shaped; but the locomotion does not
cease, and the animal continues to creep with its posterior part
contracted in this way for several minutes. The peripheral
edge of the mantle is quite sensitive to touch, reacting by local
retraction.

NATURAL HISTORY OF ONCHIDIUM 461

The ventral surface of the projecting portion of the mantle
when stimulated undergoes reactions of a sort similar to the
preceding at the anterior or the posterior extremity. At the
sides of the body, the mantle locally bends ventralward toward
the substratum when its lower surface is activated. If the
periphery of the foot is touched, the substance of the foot is
puckered away from the source of stimulation, and the mantle
is depressed at this point. The ends of the foot are more sensi-
tive than its lateral edges. Stimulation of the end of the foot
causes the animal as a whole to contract, arching its dorsum.
The ventral surface of the foot, which may be studied by allow-
ing the Onchidium to creep over a gap between two glass plates,
reacts negatively to the tactile activation of a blunt-pointed
instrument, and the lateral margins of the foot on both sides at
the level of stimulation contract locally toward the median line.

The tentacles and the oral lappets are the parts most sensitive
to touch. A tentacle stimulated at its tip or at a point along
its stalk is quickly rolled inward, glove-finger fashion, like the
tentacle of a snail; it is then re-extended more slowly. The
response is unilateral. Unsymmetrical tactile excitation of the
tentacles may be used to direct the path of locomotion, as was
attested by the fact that an Onchidium moving away from a
source of light could be made to alter its direction by repeatedly
touching one tentacle; the animal turns away from the stimu-
lated side.

Activation of the anterior portion of the mantle-fold at one
side results in the contraction of the tentacle, of the oral lappet,
and of the head as a whole on the homolateral side; the mantle-
fold is itself at the same time locally depressed. More intense
stimulation or a light touch repeated four to six times leads to a
similar response from the opposite side of the head as well.

If a tentacle be very lightly touched, it alone responds; if
somewhat more strongly stimulated, the homolateral oral lappet
is also involved in the reaction. An oral lappet, however, will
respond repeatedly without the homolateral tentacle being impli-
cated. The decided parallelism between these peculiarities—
homolaterality of response and irreciprocal conduction between

462 LESLIE B. AREY AND W. J. CROZIER

tentacle and oral lappet—and the corresponding behavior of the
analogous organs of a nudibranch, such as Chromodoris (Crozier
and Arey, ’19b), should be noted here.

The high tactile irritability of the oral lappets was noted by
Joyeux-Laffuie (’82, p. 311), who described also the richness of
their epithelial innervation and the manner in which, as we
have LT EM described, these organs are constantly ees in
‘feeling over’ the substratum during creeping.

Onchidia placed in a trough through which a gentle current of
sea-water is maintained soon become oriented by the current
and creep with it. It is not altogether certain to what extent
such orientation may be a purely passive one, for to an under-
water surface the foot of Onchidium is not very firmly attached,
so that, as the anterior end of the foot is sometimes lifted, this
part of the body may be mechanically swung around with the
current.

A location affording on a calm, sunny day several hundred
Onchidia may be searched in vain for a single one if a strong
wind be blowing from such a direction as to impinge upon this
particular stretch of shore, while at other places, sheltered from
the wind, the usual complement of feeding snails is seen at low
water. With the idea that perhaps the explanation might here
be found for the non-emergence of Onchidium during stormy
periods, individuals were taken from the laboratory stock and
allowed to creep upon a horizontal slab of stone freely exposed
to the wind. They became promptly oriented so as to head
away from the wind. The tentacles were sharply retracted
when first struck by the breeze, then subsequently slightly
extended. When the tentacles were removed by a quick snip
with scissors, these animals were not longer oriented by the
wind. If but one tentacle was removed, an Onchidium was
found to go through a sort of ‘cireus movement,’ tending strongly
to bend toward the unstimulated (non-tentacled) side. The
tentacles seem therefore to serve as anemotropic receptors. In
agreement with this a number of rocks containing Onchidium
colonies were found to have their upper surfaces and windward
sides free of exposed Onchidia, whereas the protected leeward

NATURAL HISTORY OF ONCHIDIUM 463

faces of these stones bore many feeding individuals. In one
case three of these had their tentacles removed, and when placed
in exposed situations crept into the breeze without hesitation.
It is probable, however, that a still stronger wind would orient
detentacled Onchidia. (No tests were made of the orientation
of detentacled snails in a water current.)

It should be mentioned here that in the tips of the tentacles
there are found considerable numbers of nerve cells; the possi-
bility is suggestive that they may be important of anemotropic
sensitivity.

With O. celticum, Joyeux-Laffuie noted that certain stretches
of shore might harbor many hundred Onchidia, whereas nearby
stations exhibited few or none; he did not believe that the char-
acter of the food supply was instrumental in determining the
erratic character of this distribution, but gave no evidence bear-
ing upon the real explanation. Similar facts are quite evident
at Bermuda, and we regard it as clear from our own studies that
the degree of exposure to wind, and possibly surf, is a prepon-
derating factor in the matter. Protection from the force of
wind and surf is essential, and at stations notably deficient in
this particular no Onchidia are found.

Onchidia placed at the bottom of a battery jar soon climb its
side, and generally do so in a straight line perpendicular to the
ground. If the dish be covered they will accumulate at the
top, some remaining on the vertical surface, others upside down
on the glass cover; particularly during the first hours of such a
test it is rare to see even one individual moving downward again,
and the whole group will usually stay at the top for days. In
the absence of a cover, however, the same animals readily creep
over the edge and on, downward, to the table. Onchidia placed
on a glass plate seemed on the whole to be negatively geotropic,
since in many instances appropriate reversals in the direction of
creeping could be induced by turning the plate; such results
were not always forthcoming, however. The natural situations
of the Onchidium nests involve normally some degree of upward
or downward creeping, since not infrequently the opening of a
nest will be on an almost vertical surface. But aside from the

464 LESLIE B. AREY AND W. J. CROZIER

fact that the snails never creep down to the actual water level,
they do creep upward or downward with seeming indifference.
Only a limited degree of negative geotropism may therefore be
postulated for Onchidium.

Vigorous vibratory mechanical stimulation serves to interrupt
the creeping of an Onchidium, but neither the sharp tapping of
a glass beaker containing some of the animals nor forcible blows
struck upon natural rock bearing them leads to any more pro-
nounced form of response. Although the snail may altogether
cease creeping for some minutes when disturbed in this way,
the tentacles are not retracted, nor is the mantle-fold depressed.

2. Photic excitation

The data presented in this section were all secured under
laboratory conditions. In nature the heliotropism of Onchidium
is inhibited (Crozier and Arey, ’19c), although its reactivity to
shading is quite pronounced.

Onchidia gather in groups on the side of aquaria away from
the light, and, once there, in the majority of cases stay on that
side. Diffuse daylight and brilliant sunlight induce the same
form of response, which obtains whether the animal is in air or
under water. After reaching the side of the container farthest
from the light, they usually continue to creep upward to the
water edge, where a brief halt commonly ensues, and then on
up. To a source of unilateral horizontal light, an individual
orients sharply, precisely, without ‘trial’ movements. Having
oriented, the animal proceeds to move in a straight line away
from the light source.

In a dark room the surface of Onchidium was explored with a
minute and fairly intense beam of light (cf. Patten, 715). The
diameter of this beam was about 0.8 mm. The anterior end of
the mantle was found the most sensitive part of the creature’s
surface. Even with this minute source of stimulation it was
possible to make the animal move in any desired direction by
appropriately placing the spot of light upon one side of the
anterior end of the mantle. The photic excitability of the
posterior part of the mantle was somewhat lower than that of

NATURAL HISTORY OF ONCHIDIUM 465

the anterior end of the mantle. The stimulation of the posterior
region of the mantle led, however, to a type of response not seen
when the anterior portion was activated. When the lght-spot
was applied to a point anterolateral with respect to the mid-
posterior point, the anterior end of the Onchidium was swung
sharply toward the stimulated side; when the body was then,
immediately afterward, straightened, the stimulated spot was
swung out of the region of activation. The body of Onchidium
is not readily twisted sideways, and it seems that when such
twisting does occur it is always initiated by the anterior end, and
is of such a character as to contract the animal on the side of
the stimulated spot. This spot must, however, be located on
the posterior half of the body. In the case of photic irritation,
the resulting maneuver is very efficient in withdrawing from
activation a given stimulated posterior part, and seems pur-
posive; but it appears to be the only form of reaction possible
under the circumstances. At the anterior end of the mantle,
however, the reaction elicited by a spot-light is a more purely
directive one, leading to less pronounced differential body con-
traction, but to more vigorous locomotor movements.

The whole mantle dorsum is photosensitive. When one point
was activated by the spot-light, a puckered depression quickly
appeared and extended as a furrow transversely across the
animal’s back.

Stimulation of the anterior end of the foot or of the head
region (beneath the mantle) was, so far as could be detected,
without directive effect.

The tentacles, also, seemed themselves to be non-reactive to
illumination, nor when stimulated either by the spot-light or by
concentrated sunlight did they induce directive locomotion.

The tentacles are not important for orientation by light.
Onchidia from which the tentacles have been amputated still
orient away from the light as sharply and as promptly as with
the tentacles intact.

These experiments suggest that the dorsal mantle surface
contains the only photoreceptors important for phototropism.
This is confirmed by the following experiment:

466 LESLIE B. AREY AND W. J. CROZIER

Animals from which the tentacles had been removed had the right
or the left half of the dorsal surface smeared with a mixture of lamp-
black and vaseline. When a narrow beam of sunlight was allowed to
fall on the blackened side of the body, no response of any kind was
observed. When the unpainted side was exposed to the light, how-
ever, these animals behaved like normal ones—they oriented precisely
‘and always toward the darker side.

The distribution of photic irritability, which thus seems to be
confined to the dorsal surface of the mantle, is sufficient to show
physiologically that differentiated receptors are concerned in
the reactions of Onchidium to lght. The exclusive photic
irritability of the dorsal surface of the body is important in
connection with the well-known mantle-eyes present in some
members of this genus. Two views have been advanced con-
cerning the phylogeny of these eyes: (1) that under bionomic
stress the eyes have developed from some less specialized integu-
mentary photoreceptors, this being the essence of Semper’s
idea, and (2) that eyes of this type were early developed by the
primitive Onchidium stock, and have subsequently become in
some species lost or rudimentary (cf. Stantschinsky, ’08). In
O. floridanum there are no differentiated mantle-eyes, and,
although tentacular eyes are present, no evidence has heen
forthcoming to show that they actually play a part in the
creature’s activities, or indeed that they are in any degree photo-
sensitive. A physiological analysis of the photic sensitivity of
Onchidia possessing undoubted mantle-eyes should afford some
data valuable in this connection, but has yet to be made.

When the light falling upon an Onchidium quietly creeping
in air is suddenly decreased, the tentacles of the mollusk are
quickly and forcibly withdrawn beneath the mantle, the head is
retracted, locomotion stops, and the mantle is lowered into con-
‘tact with the substratum. In the case of a snail which has for
some time been undisturbed in this way, a very. slight decrease
in light intensity induces response of almost maximum ampli-
tude. In any event, the snail quickly resumes the attitude it
had previous to stimulation, even though the reduced illumina-
tion be maintained. If a shadow be cast on the anterior end
only, the head is sharply withdrawn, locomotion ceases, the

NATURAL HISTORY OF ONCHIDIUM 467

mantle-fold being promptly depressed; at the posterior end,
when shaded locally (from behind), the posterior part of the
body is likewise contracted and depressed, but the contraction
of this portion of the body is not so pronounced as when,
in the case of anterior shading or of the shading of the whole
body, the anterior part of the body seems to be drawn backward.

No reaction follows the shading of the tentacles alone, but if
the anterior edge only of the mantle be shaded, a normal reac-
tion follows. Onchidia from which the tentacles have been

TABLE 1

Showing the course of exhaustion of the response to shading in three individuals
(Onchidium) in bright sunlight; shaded at three-minute intervals

RESPONSE
NUMBER OF -

STIMULATION

Animal 1 Animal 2 Animal 3
1 Complete reaction Complete reaction Complete reaction
2 Complete reaction Only tentacles re- Weak total response
tracted (shghtly)

3 Reaction mostly from | Just perceptible re-| Good complete reac-
the tentacles sponse of tentacles tion

4 Tentacles retract One tentacle partial- | Tentacles only, re-
slightly ly contracted tracted

5 Tentacles retract just | One tentacle slightly | Tentacles only, re-
perceptibly bent away tracted

6 One tentacle caused
to bend to one side

7 One tentacle caused
to bend to one side

amputated are normally responsive to shading, even within five
minutes subsequent to the removal of the tentacles.

Reactivity to shading is quickly exhausted by rapidly repeated
activation. In bright sunlight, one or two ‘complete’ reactions
are all that are usually obtainable when the shadings succeed
one another at intervals of 3 mm. (cf. table 1). Successive shad-
ings evoke responses gradually less complete.

When the receptivity of the anterior end for shading stimula-
tion is completely exhausted, this part is nevertheless fully
responsive to delicate tactile activation. Many tests were

468 LESLIE B. AREY AND W. J. CROZIER

made with the aid of a glass plate upon which there was a small
opaque dot of india-ink, thus enabling a small shadow to be
cast upon the back of an Onchidium. In a resting animal, if
the shadow spot so produced was made to move slightly upon
the back of the mantle, a response was usually provoked, in
the form of a slight retraction of the tentacles with or without
the depression of the mantle.

Under water the tentacles of Onchidium are never extended so
fully as when the animal is in air. The response to shading is
therefore not so conspicuous when the animal is under water,
since the tentacular retraction is not so obvious; otherwise, the
behavior of Onchidium when peeieL is identical under water
and in air.

Onchidium gives no response whatever to increase in light
intensity, as such.

Onchidium is but one of a number of animals in which it has
-been demonstrated that precise negative orientation by lght
may occur simultaneously with the presence of definite and
conspicuous negative responses to decrease of light intensity,
reactions initiated by increase of light intensity as such being
absent (Euglena, Bancroft, 713; the leech Dina, Gee, 713; holo-
thurians, Crozier, ’14, ’15; Chiton, Crozier and Arey, ’18).
Onchidium resembles Chiton and the holothurians especially in
the fact that the regional distributions of the two kinds of irri-
tability are seemingly identical. Phenomena of this kind nullify
the supposition that the orienting stimulus can originate, for
these animals, in the changing intensity of hght; because the
sense of their only known form of response to changes of light
intensity is incompatible with the manner in which photic orien-
tation occurs. The significance of this fact, especially as demon-
strated in bilaterally symmetrical iil? seems to have been
unduly ignored.

Additional evidence was secured from Onchidia repeatedly
shaded at various rates until they ceased to respond to shading
at all. The orientation of these individuals in a field of light
was in no particular different from that of snails with undimmed
reactivity to shading.

NATURAL HISTORY OF ONCHIDIUM 469

If a sharply defined narrow beam of sunlight, about 1.5 times
the width of the animal, is reflected into a dark chamber con-
taining an Onchidium, the snail orients and creeps away from
the light. Slight deviations bring the animal within the zone of
shadow, whereupon it retracts sharply, being thus confined to
the light beam. After a few shading stimulations, however, this
kind of reactivity is exhausted, a slight deviation puts the ani-
mal’s anterior end in the shade and the rest of the body follows.
Similarly, if an animal be shielded from the light on one lateral
half, and then illuminated from behind, it turns and creeps into
the shade.

The photic orientation of Onchidium is therefore a purely
tropistic process, determined through photochemical transfor-
mations localized in specific receptors. In the similar case of
Holothuria it has been suggested (Crozier, 715) that the same
photochemicéal system may afford the receptive basis for both
the continuous action of light and the sudden decrease of light
intensity; final proof for this suggestion cannot as yet be offered.

According to the theory of responses obtainable under these
conditions, the musculature upon either side of an unequally
illuminated, bilaterally symmetrical animal undergoes differ-
ential contraction, resulting in forced movements of orientation
with reference to a single source of light (Garrey, 718). With
such an organism as Onchidium it might then be conceived that
when an individual placed upon its dorsum begins to carry out
righting movements, the direction in which righting occurs
should be strongly influenced by the relative illumination of the
two sides of the body. We have made experiments of this kind
with Onchidium. ‘Tests of this point have been made previously,
principally with echinoderms. ‘There are reasons for regarding
such organisms as the starfish as not well adapted for this pur-
pose. The superiority of Onchidium consists in the fact that it
is (minor morphological points aside) a pronouncedly bilateral
animal. It is necessary to point out that tactile stimulation of
the dorsal surface may complicate such tests. The righting of
an Onchidium involves the lateral twisting ventralward of the
anterior end of body and foot, so that the anterior portion of the

470 LESLIE B. AREY AND W. J. CROZIER

foot is enabled to attach itself. If illuminated from one side
only, Onchidium almost invariably contracts the musculature of
the opposite side when righting is begun.

We may consider at this point the nature of certain evidence
sometimes adduced in adverse criticism of the idea that photo-
chemical transformations are responsible for the activation pro-
ducing heliotropic movements. The theory of heliotropic move-
ments centers upon the fact of orientation in a field of light; if
the light for any reason produces unequal effects in the symmet-
rical receptive areas, the rates of photochemical transformations
will not be the same in these areas ‘‘and the rate at which the
symmetrical muscles of both sides of the body work will no
longer be equal; as a consequence the direction in which the ani-
mal moves will change.’”’ It has been deduced from such state-
ments that the rate of locomotion of an animal, once oriented
(or already oriented), should be proportional to the acting light
intensity (Dolley, ’17).

This idea is not necessarily correct, and it may be pointed out
that there are forms for which it can at most have but a lmited
applicability.

The locomotor progress of Onchidium is determined by the
succession of transverse neuromuscular waves upon the sole of
the snail’s foot. These waves traverse the length of the foot
at a rate of about 16 cm. per minute, succeeding one another at
intervals of approximately five seconds (at 27°C.). The speed
of these waves and their frequency are very largely independent
of the proximal stimulus, once a certain threshold of activity has
been exceeded. But in orientation the bending of the body is
due to the differential contraction of symmetrically located
parietal muscles, quite distinct from those producing the pedal
waves. The rapidity of orientation thus depends upon the
differential action of the symmetrical halves of the neuromus-
cular mechanism which controls the lateral bending of the body
—upon the degree to which one side is contracted, the other
side reciprocally relaxed. It should therefore not prove surpris-
ing to find the speed of photic orientation (within limits imposed
by the snail’s anatomy) proportional to the intensity of the

NATURAL HISTORY OF ONCHIDIUM 471

orienting illumination, but the rate of progression, with orien-
tation established, relatively independent of this intensity.

In agreement with this conception, we have found differences?
in the speed with which a dark-adapted Onchidium is oriented
by light of different effective intensities, but with very slight

TABLE 2

Showing the result of one experiment in which five dark-adapted Onchidia were ori-
ented and allowed to creep in illuminated fields of three intensities of horizontal
light. The different intensities were secured by placing the center of the obser-
vation stage at distances of 12, 24, and 36 inches, respectively, from a water-screened
oil-lamp. At the beginning of each test the light impinged wpon one side of the
animal, so that ‘complete orientation’ required a turning through 90°. When
orientation had been completed, the rate of creeping over a 6-cm. stretch was
measured in the same illumination

I Il Ill

ANIMAL SE a Se
Orientation | Locomotion | Orientation | Locomotion | Orientation | Locomotion
1 orl ee 20 7.0 iL. D2
2 4.8 4.1 Bsa one Weil mili
3 Dao Dae, Bial 6.1 0.9 4.3
4 4.2 6.1 HAG: 4.8 ikail 6.0
5 5.6 4.9 2.9 cll) 0.8 4.5
Means...... 5.0 5615) B30. 5.6 1.0 5.2
100
Orient. SVE a Sa ae PL se a. ee
sec. for complete orientation.
cm.
tocom—
sec.

If the rate of orientation were directly proportional to the light intensity,
rates under I, II, III, should be in the proportion 9:4:1; actually they are as
5.0:3.5:1. More extensive tests might provide a more complete agreement.

The rate of progression is practically independent of the light intensity.

differences in the speeds of progression under different intensities
of light, provided the intensity be sufficient to keep the snail
moving continuously. There is undoubtedly, as in Chiton
(Arey and Crozier, ’19), some correlation between speed of pro-

7In making this statement we must for the present regretfully rely mainly
upon the results of qualitative experiments. When this work was being done
at the Dyer Island laboratory no electric light or other suitable light source was
available. WKerosene lamps did not afford light of adequate intensity.

472 LESLIE B. AREY AND W. J. CROZIER

gression and light intensity, but a limit is quickly reached beyond
which no sensible increase is possible in the rate at which pedal
waves succeed one another upon the foot. Moreover, it is very
probable that the pedal nervous mechanism, as is to some degree
indicated by other responses, is ‘set off’ as a whole, and not bilat-
erally, by impulses originating in the mantle and passing through
the central ganglia. So long as the photoreceptive mechanism
is in bilaterally balanced excitation, the rate of operation of the
pedal musculature might then proceed upon the ‘all-or-none’
principle, uninterfered with by the mechanism concerned in
turning movements. This conception is already well founded
for vertebrates (cf. Brown, 714). The rhythm of the ‘scratch
reflex’ of mammals is independent of the frequency of the excit-
ing stimuli (Sherrington, ’06). More generally stated, the rate
of contraction of symmetrical locomotor muscles is very nearly
constant so long as the rate of stimulation of the two sides of
the animal is the same; the bilateral halves of the central nerv-
ous mechanism of control together behave as an independent
unit so long as they are not stimulated differentially. Con-
siderations of this sort make it clear why careful measurements of
the locomotor rate of insects oriented in a field of light show no
correspondence between rate of creeping and the light intensity
(Dolley, 717; Minnich, ’19). One must insist that before an
animal reaction can properly be made the basis of quantitative
experiments, the nature of the reaction must be carefully reviewed
as to its suitability for the purpose in mind; measurements of a
phenomenon not itself sufficiently understood are likely to
prove a waste.of time. Minnich (19, p. 406) has also pointed
out that light intensity affects the posture of the legs of the
bee, but not their rate of rhythmic activity.

The speed and precision of the orientation vary in a char-
acteristic manner with the wave length of the light used. The
experiments upon this topic were qualitative in nature, but were
repeated a sufficient number of times to make sure of the essen-
tial features of the results, which were as follows: When sunlight
or light from a tungsten filament is made to pass through dif-
ferent ray-filters, blue and green lights are very powerful in induc-

NATURAL HISTORY OF ONCHIDIUM 473

ing orientation, yellow light is less so, and red light is relatively
ineffective. In different series of experiments ray-filters of
several kinds were employed; thus in the first series of trials we
used the colored glasses previously described in the work on
Chiton and on Chromodoris (ef. Arey and Crozier, ’19; Crozier
and Arey, 719 b), with this result:

Blue. When Onchidia, in the dark, are made to crawl directly
toward the future source of light, and the blue light is then suddenly
turned on, the animals stop, pivot sharply through 180°, without
creeping, then move directly away from the light source. If started
in such a way as to creep in a direction perpendicular to the future light
source, the animals sometimes hesitate, lift the anterior end, and swing
it sharply away from the light source, thus making a precise 90° turn.

Green. When creeping is so begun as to lead the animal into the
light, a prompt turning through 180°, with subsequent locomotion
away from the light, at once follows the admission of the green light.
On the whole, the response is not quite so sharply carried out as with the
blue light. To unilateral light the response is as in the case of blue.

Yellow. To light impinging directly on the anterior end of the ani-
mal, Onchidium responds by creeping in a complete circle away from the
light, locomotion continuing along a path parallel to that at first pursued.
With unilateral light, the process of orientation is equally precise, but
less rapidly effected than in the case of green.

Red. In the majority of the trials, although Onchidium invariably
creeps away from the light, the orienting process is sometimes quite
slow, the animal occasionally continuing to move onward for several
millimeters after direct red light is turned on—then going off at a right
angle or turning back on its path. No cases were observed in which the
Onchidia pivoted immediately away from the light, as with the blue;
but on the contrary they maintained a steady creeping progress during
the course of orientation. When orientation is completed, the new
path does not quite coincide with the old, as in the case of blue or green
light, but in addition to being in the reverse direction lies within 10°
to 45° on one side or the other of the original course.

These experiments were made in a dark chamber, the only
source of light being a slit of appropriate size covered by the
appropriate ray-filter. The result of these tests, showing the
higher stimulating power of blue and green light, agrees with
what was found in the case of shading: under the blue or the
green ray-filter used in the phototropism tests, decrease of light
intensity led to a normal reaction on the part of the Onchidium,
while under the red or the green filter only a slight response, or

474. LESLIE B. AREY AND W. J. CROZIER

none whatever, could be secured; this was true even when the
light falling on the blue filter was made very weak, while that
admitted to the red filter was much stronger.

3. Thermal excitation

Under water Onchidium remains ‘normal’ in its behavior, so
far as can be told, at temperatures between 17° and 36°C. If
placed in sea-water cooled to 5°, the animal becomes instantly
immobile, and does not respond to touch. After ten to fifteen
minutes at this temperature, Onchidium quickly recovers if
transferred to sea-water at the normal temperature (26° to 27°);
in the early stages of this recovery no reactions of any sort could
be elicited by tactile agents, but good reactions, involving con-
traction and curling up of the whole body, were elicited by the
local application of small volumes of dilute HNO; (N/200 +). .
Sea-water cooled to 10° has practically the same effect. At 15°,
an Onchidium suddenly placed in water of this temperature
usually exhibits a few sluggish contractions of the foot muscula-
ture; if the animal is expanded, it contracts a little; if con-
tracted, it relaxes somewhat and then remains quiet. Tactile
irritability is present, but the responses are of slight amplitude.

Above 17°, probably, certainly above 20°, and until a tempera-
ture of 35° to 36° is used, no changes in the behavior of Onchid-
ium indicative of sensory activation by heat are obtainable.

At 35° to 38°, Onchidia transferred to sea-water of these tem-
peratures quickly become motionless, in the expanded state,
except that if placed ventral surface uppermost there seems to
be a slight but detectable increase in the peristaltic activity of
the foot. They may make attempts to right themselves, but
the bending movements involved in this process do not continue
after two to three minutes.

With temperatures up to 45°, the result of immersing the
animal suddenly is to call forth rather violent general contrac-
tions, more powerful the higher the temperature, but lasting
less than one minute at 45°. Onchidium endures exposure to
water of the latter temperature for about forty-five minutes;
even after thirty minutes’ exposure, it still reacts to touch and
to stimulation by weak acid (HNO;) solutions.

NATURAL HISTORY OF ONCHIDIUM 475

The lowest of the high temperatures which leads to death
rapidly (i.e., within one minute) is very nearly 48°, although
47° is withstood for ten to fifteen minutes.

Local application of ‘heat’ or ‘cold’—the tests being made by
bringing a warmed or cooled glass rod into close proximity with
a part of the snail, in air, or, since it was possible by careful
manipulation to avoid tactile complications, bringing the glass
rod directly into contact with the skin—provided no evidence of
delicate thermal sensitivity. A rod cooled approximately to
0° C. called forth no detectable response whatever. A rod
heated to 60°, or even 50°, did, however, cause prompt responses
from all parts. The type of reaction was for each part of the
body the same as in the case of touch, but more vigorously car-
ried out. A light touch leading to no response at all calls forth
a powerful reaction when administered with a warmed glass
rod. The mimimal temperature effective in this way is about
45°.

In the case of an Onchidium detached from the substratum
and resting on its back, the lower surface of the mantle may be
touched repeatedly on one spot until the maximal amount of
rolling-up occurs; there is then no further response obtainable to
repeated mechanical stimulation, but a rod warmed to 45° (in
reality, perhaps somewhat cooler at the moment of application)
induces a sharp local contraction and drawing away of the
mantle. Similarly, the back of an Onchidium may be repeatedly
stimulated by pressing on it with a blunt point, by prodding or
stroking it, until a response fails to be elicited ; but a warmed
rod, however lightly applied, calls forth a deep retractive puck-
ering of the mantle over a relatively wide area.

The findings in these somewhat tedious experiments indicate
the presence of a ‘heat’ sense physiologically distinct from that
involved in touch, although (perhaps owing to the method of
experimentation) the relative sensitivity of the various parts of
the body seems the same as that already described for mechan-
ical excitation. Sensitivity to ‘cold,’ on the other hand, is but
poorly defined, if indeed it can be said to be present at all.

THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 32, NO. 3

476 LESLIE B. AREY AND W. J. CROZIER

The resistance of Onchidium to high temperatures is decidedly
greater than that of Chiton (Arey and Crozier, 719), which is
worth pointing out because these animals live almost side by
side, the Chitons probably being exposed to higher temperatures,
and for longer periods, than Onchidium. The blue-black hue of
most of the Onchidia may lead to the absorption of heat energy,
however, during the intervals of creeping in the open. An
adaptive basis might therefore conceivably be found for the
relatively high heat resistance of the dark Onchidia used in these
tests, if it should be found that these dark forms show a heat
resistance superior to that of the pale variety obtained at nearby
stations.

4. Chemical excitation

Onchidium does not react in any definitely detectable way
when a small volume (0.5 ec.) of rain-water is applied to any
part of its surface. Sea-water itself, so applied, may provide a
mechanical stimulus unless the pipet current be very gentle;
this effect, however, may readily be eliminated. Sugars (mal-
tose, lactose, sucrose) in 1 M solution in rain-water were not
found to induce responses. Sea-water concentrated to one-half
its volume by evaporation and reaerated by bubbling air through
it called forth contractile movements when applied to the head
and lips; sea-water whose osmotic pressure had been increased
by the addition of 0.5 M glycerin similarly induced slight reac-
tions when applied to the lip region.

The osmotic disturbances which may induce responses about
the mouth are much less effective as activating agents for Onchid-
ium than are various chemicals, including alkaloids, anaes-
thetics, alcohols, acids, alkalies, salts, and such irritants as
H,O, and various-essential oils. The whole surface of Onchid-
ium is open to sensory activation through the agency of dilute
solutions, such as those found effective for numerous other inver-
tebrates (cf. Arey and Crozier, 719).

Small volumes of sea-water which had been shaken with ether,
chloroform, carbon bisulphide, aniline oil, oils of thyme, ori-
ganum, juniper, bergamot, pennyroyal, cloves, or cassia, were

NATURAL HISTORY OF ONCHIDIUM 477

allowed to flow over a portion of the surface of Onchidium; the
reactions were in every case very vigorous, no matter what part
was immediately concerned in the activation—the foot, for
example, being found extremely sensitive.

Low concentrations of electrolytes, however, are more effec-
tive on the dorsal mantle and in the mouth region than on the
foot. The limiting concentrations beyond which the further
dilution of some representative materials no longer constitutes
sensory stimulants under the conditions of these tests—in a
general way similar to those found for other animals—are:

HCl—The foot ceases to react when the solution is more dilute than
M/300; the remainder of the body surface responds to dilutions as
great as M/500.

KOH—The limiting concentrations are about the same as for HCl,
but somewhat higher.

KCl—For the foot, M/3; for the rest of the body, roughly M/5 or
M/6.

Picric acid—At M/50 a weak response is still obtainable from the
foot; the limiting dilution for the general body surface is about M/3000;
at M/2000 the stimulation induced is still quite decided.

The animals were tested in air. The solutions were made up in
rain-water; with picric acid, solution in sea-water gave results
identical with those already tabulated.

It was attempted, for comparison with conditions in other
forms, to establish the relative stimulating efficiency of some of
the commoner ions. Solutions of the neutral salts of the alka-
lies, 5/8 M in rain-water, induced responses from all parts of
Onchidium. MgCl, even in 1 M solution, gave weak reactions
only, and none whatever upon the foot.

NaCl and LiCl led to the explosive discharge of the repugna-
torial glands, whereas KCl and NH,Cl as a rule did not. When
local applications of the two latter salts were followed by a gland
discharge, it could be seen that the discharge was a secondary
phenomenon, due to the general squeezing of the glands through
the muscular contraction of the mantle, rather than to the
(normal) contraction of the investments of the glands them-
selves, the latter holding with NaCl and LiCl. NaCl in small
quantities led, on the whole, to more violent gland discharge

478 LESLIE B. AREY AND W. J. CROZIER

than did the same amount of LiCl. NaCl was more effective,
also, on animals previously fatigued by repeated stimulation.

The foot is the part of Onchidium least responsive to ionic
excitation; therefore it is in some respects a region advantageous
for comparing the relative effectiveness of different substances.
NaCl and LiCl induced pronounced puckering contractions
when applied to any part of the foot, the animal being also induced
to curl up, pill-bug fashion. KCl and NH,Cl were very much
milder in their effects; with KCl, tiny surface puckerings were
induced, local wrinklings, but no deep local retractions such as
were occasioned by NaCl; the efficiency of NH,Cl was in this
respect invariably less than that of KCl.

We arrive therefore at the following series, which expresses
the relative efficiency of these cations at uniform concentration:

Nae) Sn ke Ni

The halides of potassium were used to obtain a corresponding
series for the anions. KI and KCl were more effective than
KNO; or KBr in stimulating the foot of Onchidium. While
these two groups were sufficiently distinct, the further separa-
tion of the anion effects was a matter of some difficulty. The
series eventually chosen as best expressing the results, in terms
of the relative magnitudes of the reactions induced, was this:

I > Cl NO; >) Br

The cation series is practically the inverse of that usually
found in such eases, and of that which we have found for the
sensory activation of several other mollusks (Arey and Crozier,
19; Crozier and Arey, ’19b), employing comparable criteria.
The anion series is more like that commonly reported. The
relatively great stimulating power of NaCl may be related to the
fact that Onchidium creeps upon rock surfaces wet with sea-
water, but exposed to the evaporating power of the sun. In
this connection it should be noted that the mouth region is
decidedly the most sensitive part, and also that in an actively
creeping animal it is not possible, as a rule, to obtain sensory
responses from activating substances at concentrations so dilute

NATURAL HISTORY OF ONCHIDIUM 479

as in the case of resting individuals. These features may play a
part in directing the movements of Onchidium in nature.

There is some evidence for the view that receptors for general
‘irritating’ activation are distinct from chemoreceptors proper.
Many substances which provide powerful sensory excitants for
Onchidium do not induce discharge of the repugnatorial glands.
Methyl or ethyl alcohol, however, at 5 M concentration in rain-
water, do lead to such discharge; whereas pure amyl alcohol,
directly applied, does not, although the general responses of the
animal are in the case of the amyl aleohol much the more vigor-
ous. Chloretone also, and urea, in relatively concentrated solu-
tions, do not lead to discharge of the poison glands, although
they do induce powerful reactions on the part of the animals as
a whole.

An effort to effect physiological separation of tactile- and
chemo-reception was without decisive result.

It may be noted that, according to Joyeux-Laffuie (’82, p.
238), O. celticum exhibits in its feeding a certain amount of pref-
erential selection of Ulva from among a diversity of algae. He
regarded this fact as evidence of a certain degree of ‘gustatory’
discrimination, associated with the mouth parts, but pointed
out in addition that the oral lappets were employed in feeling
over bits of algae before their being submitted to the radula.
This latter observation we can confirm for O. floridanum, but
we have seen no evidence of:selective activity respecting food,
perhaps because the thin algal carpeting of the rock in Onchid-
ium habitats presents so uniform a field. Some of the older
writers have referred to the ‘eating’ of silt by Onchidium; we
have already pointed out (Crozier and Arey, ’19a) that shore-line
silt is ingested, but in a purely fortuitous manner, through its
adherence to the seaweeds.

ON THE ANALYSIS OF THE HOMING BEHAVIOR

An attempt to interpret or to reconstruct the natural activi-
ties of Onchidium on the basis of analytical study of its modes of
response when removed from its habitual environment is con-
fronted at once by some profound inconsistencies and by the

480 LESLIE B. AREY AND W. J. CROZIER

puzzle of ‘homing’ behavior. Although we have to offer some
suggestions toward such an interpretation, they are presented
with due reserve as to their probable finality. No more striking
instance is known to us of apparent incompatibility between the
results of controlled experimentation, repeatedly verified, and
the most obvious activities of the same animal’s natural life.

First, as to heliotropism. In the laboratory Onchidium
behaves as a typical negatively heliotropic animal. In nature
its behavior is in every essential at variance with this finding
(Arey and Crozier, ’18; Crozier and Arey, 719 ¢). Not only does
this mollusk creep out of its dark nest into the glaring sun, when
the tide is falling, to feed, but it does so only during daylight
hours, and never at night, no matter how bright the moon.
During the period of its emergence an Onchidium’s movements
seem not in any degree influenced by the sun. An individual
quietly creeping in brilliant sunlight may be shaded by a board,
and after re-expansion following its response to the shading, it
creeps on as before. If now sunlight be thrown suddenly on
this animal from a new direction, with a mirror, it may momen-
tarily ‘hesitate’ somewhat, perhaps turn very slightly to one side,
but soon it continues as before.

It is not merely the presence of a normal kind of shore sub-
stratum which determines this suppression of heliotropism,
because animals brought from laboratory stock (secured several
days before) and placed upon the same rock are oriented pre-
cisely by the sun. The presence of the specific rock surface
normal to the Onchidia emanating from a given nest, namely, of
that surface within a certain small radius of their nest entrance,
is the deciding influence. An Onchidium moving toward its
nest entrance, but heading directly into the sunset, may be
picked up and then replaced upon the rock without interfering
with its course. But if a glass plate be slipped under the ani-
mal before replacing it, it is now at the mercy of its heliotropism.
If an Onchidium from one section of the shore be quickly trans-
ferred to a strange zone, again the animal is oriented by the
light. Obviously, there is no question here of the mere rapid
exhaustion of the heliotropic mechanism or other kind of ‘light
adaptation.’

NATURAL HISTORY OF ONCHIDIUM 481

It is not that the Onchidium, ‘seeking its best interests,’
comes out into the light to feed and in so doing ignores the
dictates of its negative heliotropism. Definite evidence, on the
contrary, is available to the effect that under normal circum-
stances the central nervous mechanism of heliotropic movements
is inhibited, perhaps ordinarily by impulses originating in the
oral lappets which compete successfully for the control of the
body musculature. ‘This inhibition can be abolished (‘reversed’)
by means of strychnine (Crozier and Arey, ’719c). Reversal of
inhibition within the ganglia of mollusks is known in Chiton,
in Chromodoris, and in Cephalopods (ef. Crozier, ’20).

A possible explanation for the existence of negative heliotro-
pism under any circumstances may be found in an imperfection
of the photoreceptive system. The response of Onchidium to
sudden shading is clearly of conceivable survival value; it leads
to the retraction of the tentacles, the cessation of locomotion,
and the depression of the mantle-fold to the substratum; the
mantle is thus enabled to assist as a hold-fast, for the suction
power of the foot is but poorly developed, and in any event the
algal-covered surface affords a difficult field for contact attach-
ment by the small foot of an animal so light in weight that it
does not flatten out the algae. Our suggestion as to the nature
of the heliotropic irritability, depending on the continuous:
chemical activity of incident light, involves the assumption that
the stimulation so produced is in this case bound up with sub-
stances forming a necessary part of the receptive mechanism for
the response to shading. The obviously efficient manner in
which heliotropic impulses are normally blocked, in such fashion
that they play no detectable part in controlling the creature’s
movements, may account for the fact that this deficiency in the
photosensitizing mechanism has failed to suffer selective elimina-
tion. A balanced system of photocatalizable reactions was
postulated to account for the phenomena of photic irritability
in Holothuria (Crozier, 714). The general features of such a
system have been (in Mya) admirably treated in a quantitative
manner by Hecht (19). For Holothuria it was suggested that
both phases of activity within the photosensitive system—

482 LESLIE B. AREY AND W. J. CROZIER

photolysis in light of given intensity, and resynthesis in light of
lower intensity—might be involved in stimulation by light and
by sudden decrease of light intensity, respectively; here, both
forms of irritability play rdles of bionomic importance (Crozier,
15), whereas, according to this notion, but one phase of the
matter, namely, sensitivity to shading, is permitted to exert an
influence upon the normal behavior of Onchidium. The mechan-
ism whereby the possibility of heliotropic response is normally
prevented has already been discussed (Crozier and Arey, ’19¢e).
It appears to depend upon specific impulses originating in the
oral lappets, at their points of contact with the substratum, for
when these lappets are removed or are anaesthetized by MgSOs,
the Onchidium becomes photonegative, even if replaced upon the
specific rock surface from which it was taken.

The rhythm of the tides, ordinarily well defined, controls the
emergence of Onchidium upon its feeding ground. Under cer-
tain conditions the orderly succession of periods of low water is
seriously interfered with. High winds and accompanying ocean
currents, during times of storm, not infrequently cause such a
‘piling up’ of water within the semi-enclosed sounds at Bermuda
that the water may fail to fall appreciably for several tides; such
a period is followed, also, by a certain irregularity in the tidal
sequence. Only when the water level has become lowered to
the proper degree, previously indicated, do the snails emerge.
Conversely, extensive periods of low water, notably occurring
at spring tides, may leave the rocks uncovered along the Onchid-
ium zone for an uncommonly lengthy interval. The duration
of the emergence of a given colony is practically fixed, however,
and gives not the least indication of being normally terminated
by the rising of the tide; rather, the duration of the feeding time
is determined in a quite different way, which we shall shortly
consider.

The facts thus far presented do not exhaust the curious intri-
eacies of the behavior of our snails. It has been mentioned that
Onchidium comes out from its nest only during the daytime,
and never at night. Were it not for certain serious obstacles,
all this might be understood in terms of an hypothesis advanced

NATURAL HISTORY OF ONCHIDIUM 483

by Joyeux-Laffuie (82), namely, that the tentacular eyes enabled
the mollusk to distinguish between the darkened interior of its
shelter and the illuminated outside feeding ground, thus guiding
its emergence.’ The difficulties facing such interpretation are:
1) the fact that the tentacles of O. floridanum, although there
are tentacular eyes, seem non-photosensitive; 2) the fact that
the creature is never positively heliotropic and, 3) the fact that
emergence does not always occur when the tide is out during
daylight hours. Good instances of the sort last mentioned
were available in midsummer, when for several days at each
lunar interval both tides were seen to expose the beach zone
while the illumination was still quite good. At these times the
snails were always found to emerge during but one tidal ebb,
never during both. Once in each twenty-four hours is the
maximum frequency of emergence. Even in the absence of
conditions imposed by winds and currents impeding the escape
of tidal water from the sounds, however, this rhythm does not
represent the minimum frequency of emergence. For at neap
tides the Onchidium nests highest above water may fail for some
twenty-four hours to be submerged at flood tide, and in that case
the Onchidia there located do not emerge until after their nest has
been submerged. Moreover, in winter both morning and even-
ing tides may fail for a day or so to occur during daylight hours;
in this event, so far as we have studied the point (Dec., Feb.,
1918), the Onchidia do not emerge at all until one tide occurs
while the sun is up.

The diurnal rhythm thus clearly established ‘in defiance of?
the snail’s heliotropism completely disappears in the laboratory.
Fair-sized slabs of stone, a foot or so in breadth, were placed in
aquaria containing freshly gathered groups of Onchidium. ‘The
animals always collected after a short time on the shaded, under
side of such a slab, whether under water or in air. In the dark
(artificial darkness or at night), they crept actively over the
surface of their stones, feeding; the admission of light quickly
caused typical photonegative retreat to the dark, under surface

8 The fact that emergence occurs only during the daytime was not known to
Joyeux-Laffuie.

484 LESLIE B. AREY AND W. J. CROZIER

of the rock. It was not possible, although several times the
attempt was made, to establish an artificial tidal rhythm by
periodically lowering the water level in such aquaria; this phase
of the work will, we hope, be continued. Experiences of this
sort plainly indicated the existence of some very specific corre-
lation between the natural behavior of an Onchidium and the
features of its ordinary home.

In our opening description of the chief phases in the daily life
of O. floridanum we have already given a brief account of the
most remarkable aspect of this specific correlation, namely, the
snail’s ‘homing’ behavior. It was obviously necessary to inquire
into the nature of this peculiar activity. Although our results
are not in any sense exhaustive, for we were unable to complete
the series of experiments planned, the evidence we do command
nevertheless permits a fairly precise characterization of the
major aspects of the homing process. Involved in this matter
are: 1) the fact of almost simultaneous commencement of the
return to the ‘nest’ on the part of the scattered members of one
colony, 2) the fact that the duration of the feeding interval
seems automatically fixed without reference to the rising of the
tide, and 3) the evidence concerning central inhibition, already
referred to in connection with the normal abeyance of heliotropic
movements. To these points we shall return, after dealing with
the directed creeping toward the nest.

The fact that Onchidia are able to return to their nest after
being picked up and replaced at some other point within a cer-
tain radius of the nest aperture is best appreciated from the
perusal of such records as the following:

July 1, 1914. Little Agar’s Island.

Five individuals were seen returning to a nest. Two of these were
removed to the rock surface above the high-water mark (where, in
our experience, these animals never wander naturally); the distance
of the new point of departure was perhaps 40 cm. from the aperture
of the nest. One of the displaced Onchidia turned directly toward
the nest and crept straight back to it; the second one ‘lost its way,’
and wandered off in a strange direction. The three other members of
the original group of five marched in a sort of triangular formation

toward the nest aperture, two going to one side of the opening, the
remaining one to the other side, then all three crept into the nest.

NATURAL HISTORY OF ONCHIDIUM 485

The nest was now broken into with a chisel. The three individuals
inside were removed, and placed on a flat rock surface above the nest
opening, and at three different points each some 46 to 50 cm. from
the nest. All three Onchidia succeeded in effecting a return to the
mouth of the nest. Near the nest two of them followed a slightly
grooved trail; this trail or channel had also been used by the one return-
ing individual described in the preceding paragraph. The channel led
directly to the mouth of the nest. In getting into this trail, each indi-
vidual had to change its course greatly. When they had reached the
region about the nest aperture where the rock had been broken in
examining the interior of the cavity, the Onchidia became much ‘con-
fused’ and merely wandered about on the outside rock, where they were
left as the tide rose.

July 2, 191 4:

A group was noted returning to a nest, and one individual was
picked up and replaced on the rock on the opposite side of the mouth
of its nest at a distance of 1 meter therefrom. The animal returned
directly to its nest.

In subsequent years many trials of this sort were carried out,
and always with essentially the same result. It is possible, but
not probable, that some interesting results would have been
obtained by comparing carefully the homing capacity of Onchidia
of different ages. In the autumn three. groups of fairly distinet
size are noticeable in the Onchidium population, so these snails
probably live two years at least, if not more. Nevertheless, our
experiments did not disclose any differences in homing ability
among the individuals of different sizes. Factors which more
noticeably affect the ability to ‘home’ after experimental dis-
placement are the natural extent of the normal feeding area and
the degree to which this area is populated with nests. These
two factors are usually correlated quite closely. Boulder-like
rocks more or less isolated from the shore are frequently so
eroded as to present a veritable honeycombed aspect; a rock 3
feet by 2 in cross-section, projecting some two feet above m.l.w.,
was found to harbor about thirty Onchidium nests, if not more;
less eroded rocks, often affording considerable expanses of flat
surface, were seen to shelter an Onchidium population much
less dense. In a habitat of the latter sort the distance limiting
successful homing was about 1 meter, while experiments on rocks
of the type first mentioned (on the south side of Dyer and of

486 LESLIE B. AREY AND W. J. CROZIER

Tucker’s Island) showed that homing from distances greater
than 30 to 40 em. was not obtainable.

The fact that the course of an Onchidium when creeping out to
feed may be quite ‘haphazard,’ zigzagging here and there, while
the homeward course is usually as direct as the substratum
allows, as well as the findings in experiments just cited, shows
that it is not necessary for Onchidium to follow its own slime
track. Limpets do adhere to their own tracks (ef. Davis, 795;
Bethe, 798; Orton, ’14, ete.); Bethe (loc. cit.) thought that
limpets were guided in their return journeys by a sort of chemo-
taxis, which led them to follow their own slime trails. An
Onchidium picked up when on its homeward journey and placed
upon a clear glass plate in diffuse light does not tend to adhere
to its own slime track, nor to the slime tracks of other indi-
viduals. The same result obtains with paper or the surface of a
brick. Nor do these snails ‘favor’ a wet surface over a dry one
(glass or filter-paper). An individual from a strange section of
the shore put down on rock near an Onchidium nest will creep
without hesitation across fresh trails of others.

All the facts which we have been able to gather about the
homing of Onchidium may be brought into relation according to
the hypothesis which we now set forth. Complete demonstra-
tion of the validity of this notion involves further experimenta-
tion, the nature of which we indicate.

The Onchidia in any one colony emerge from their nest after
the tide has fallen so far as to have left it above water level for
about a half hour to an hour. They scatter over the rock sur-
face and feed. In the unfed condition certain sensory impulses
otherwise directing and controlling the creature’s movements in
such fashion as to cause it to return to the nest are inhibited. —
The possibility of such central inhibition is given from the
‘reversal of inhibition’ with respect to phototropism seen under
strychnine action while the snail is on the surface normal to it.
After having fed for a certain time, substances derived from
materials ingested while feeding pass into the juices of the
snail’s body and produce a ‘reversal of inhibition’ so far as the
‘homing’ impulses are concerned. Reversals of behavior follow-

NATURAL HISTORY OF ONCHIDIUM 487

ing feeding are known in such animals as the Porthesia cater-
pillars of Loeb (18, p. 116), and the Planarians studied by Olm-
sted (17b). This hypothesis readily accounts for the fact that
the period of feeding lasts very nearly the same length of time
in all the members of a group.

The sensory impulses thus conceived to be released from
central inhibition through the results of feeding are regarded as
originating in the oral lappets. These well-developed ‘cephalic
tentacles’ are constantly in touch with the algal carpet of the
stone. If they are cut off, the Onchidium is ‘lost,’ unable to
return to its home. The removal of the dorsal tentacles, some-
times regarded as the seat of ‘smell’ in snails, has no such effect.

These impulses must be regarded as possessing the character-
istics of ‘contact odors’ (meaning thereby that perhaps both con-
tact and ‘olfactory’ stimuli of a certain kind must be received
simultaneously). The reason for this assumption is twofold:
in the first place, an Onchidium beneath which there is shpped a
glass plate is left thereby at the mercy of its heliotropism; sec-
ondly, an Onchidium will ‘home’ from points which it has not
previously visited; therefore, the aereal dissemination of some
guiding substance must be presumed. The olfactory com-
ponent of such a complex must be regarded as more important
than the tactile, for the rock surface above high-water mark is
not covered by algae as is the surface natural for Onchidium,
nevertheless the snails will ‘home’ from points on the former
surface, although in the ordinary course of events they never go
above high-water mark. They will not home when put under
sea-water, even if quite near their nest.

The substance providing a tropistic guide for a fed Onchidium
must be granted some highly specific quality. In view of
Bethe’s findings for ant colonies (’98), such a supposition need
not be thought preposterous. Moreover, it is supported by some
striking results in our experiments on homing. In a number of
trials an Onchidium from one community was so placed that it
was forced to creep across the sunken gully leading to the opening
of a strange nest. Sometimes such a snail was found to follow
the new ‘trail,’ after a certain amount of preliminary turning

488 LESLIE B. AREY AND W. J. CROZIER

back and forth or ‘hesitation,’ and even to creep within the new
entrance. But never, in these tests, did such a snail remain in
the strange nest. Not infrequently several journeys were made
about the foreign opening (Arey and Crozier, 18).

A further significant result was that in several of the tests
made upon Onchidia taken from nests subsequently found to
possess two apertures, the successfully homing snails gained
entrance by way of a cleft different from that which they had
followed in their undisturbed homeward trip.

The specificity of the assumed ‘olfactory’ substance is not
‘remembered’ by an Onchidium after twenty-four hours’ con-
finement to the laboratory. This point was repeatedly tested
at Dyer Island. Such confinement obliterates the possibility of
homing to the old nest, even from distances of a few centimeters.

Our conception of the rdle of the oral lappets might be taken
to explain the functional significance of certain curious glands
located on these organs, in certain species. Plate (94), with
Oncis, and later Pelseneer (01, p. 20), with Oncidiella patelloides,
found on the sides of the oral lappets a pair of symmetrical aper-
tures, the orifices of glands compared by Pelseneer to the anterior
tentacular glands of Vaginula, but of unknown function. In
QO. floridanum, however, these glands are not present, otherwise
one might suggest that these peculiar organs furnish a mucous
covering for the oral lappets, perhaps containing a material
serving as a specific solvent for hypothecated odoriferous emana-
tions from the nest. It would be interesting to know how
widespread the ‘homing’ may be among these related species.

This hypothesis not only accounts for all the facts known to
us, as already stated, but obviously avoids reference to such
obscurely defined notions as ‘muscular memory’ and the like.
The hypothesis could best be tested by means of experiments
upon the homing tendencies of Onchidia which had not been
permitted to feed, and by the attempted discovery of the sub-
stance naturally responsible for our ‘reversal of inhibition.’ It
should be noted that we distinctly avoid saying whether or not
such substance may be derived from’ the algae ingested, because
a certain amount of caleareous mud is also swallowed while
feeding (Crozier and Arey, ’19 a).

NATURAL HISTORY OF ONCHIDIUM 489

DISCUSSION

1. It is desirable to deal, as briefly as may be, with certain of
the more general implications of the conclusions provided by our
inquiry into the habits of Onchidium.

The Onchidiide are a group well calculated to cause the zoélo-
gist trouble. For a long time the taxonomic affinities of these
organisms were hazy and in dispute, for it was by some (Bergh,
95; Fujita, 97) supposed that, in addition to dermal respiration
accomplished through mantle papillae (conspicuously developed
in certain species) when under water, air breathing was also
carried out, but by the organ regarded as a kidney—an idea once
used as the foundation of von Ihring’s class ‘Nephropneusten,’
but now known to have resulted from an imperfect acquaintance
with the difficult morphology of the true lung (Plate, 94; von
Wissel, 98; Pelseneer, 01). Thus we are probably dealing with
a member of a typical land group, Pulmonata, which has second-
arily taken up the habit of living on the seashore. It would be
valuable to know whether Onchidella is a more archaic type than
Onchidium proper (Plate, ’94), or the reverse. According to
Bretnall (19), Onchidium dimelii lives either altogether below
low water or between tidal limits.

2. The activities of these animals are not less curious than
their presumptive evolutional history. In the case of numerous
invertebrates of the shore zone it has seemed possible to provide
a clear description of behavior in terms directly stated by the
outcome of analytical experiments. In fact, this general method
of study has been made the basis of much recently published
work in animal ecology. The ethology of Chiton (Arey and
Crozier, ’19) and of Chromodoris (Crozier and Arey, ’19b)
can be followed with gratifying completeness from relatively
simple experimental results. With Onchidium the situation is
more subtly complicated, and for the purpose of ecologic inter-
pretation the isolated results of such a method are here almost
meaningless, as shown conspicuously by the creature’s helio-
tropic responses. Michael (’16) and others have recently been
to some trouble to emphasize the fairly obvious point that no

490 LESLIE B. AREY AND W. J. CROZIER

amount of mere laboratory investigation makes it absolutely
certain that we may predict the movements of an animal in
nature. In reality, however, it is largely a question of the rela-
tive completeness with which experimentation is conducted; nor
does it require much penetration to discover that the necessary
degree of completeness may differ in various cases.

A point of some interest, although perhaps unduly speculative,
concerns the historical source of Onchidium’s heliotropic machin-
ery on the receptor side. That the possibility of heliotropic
orientation does not entrain adaptive consequences, seems ade-
quately demonstrated by a previous discussion (Crozier and
Arey, 719 c). But most land pulmonates are negatively helio-
tropic. Might it then be conceived that the sensory organs
involved in this form of irritability are mere ‘holdovers’ from the
more ancient stock? Aside from the fact that the skin of at
least some snails and slugs is photosensitive (Yung, 710), very
little information useful in this connection has been discovered.
It must be remarked that the mechanisms for sensitivity to
light and to shading are seemingly closely connected, if not
identical, in Onchidium; nothing of this sort is known for other
pulmonates. More important, however, is the conclusion of
Stantschinsky (07) regarding the origin of the dorsal eyes
(mantle-eyes) in the family of Onchidia: he has shown it prob-
able, on general morphological grounds, that the more highly
developed forms of mantle photoreceptors are indeed primary,
rather than a secondary development, and that species, there-
fore, such as O. floridanum, which lack the dorsal eyes have
arrived at this condition through secondary degeneration.

Yung (713) holds that certain gastropods are ‘blind,’ their
tentacular eyes being non-functional, and that this is due to the
fact that the optic nerve fibers fail to penetrate the basement
membrane of the retina. The lack of apparent functional
activity in connection with the tentacle eyes of O. floridanum
might be of interest in relation to this conception, were it not
for the fact that the conditions here may not involve a complete
absence of innervation of the retinae. So far as they have been
made out from carefully studied sections, the relations of the

‘optie’ nerve in Onchidium seem to be as follows: The tentacular
nerve, entering at the base of a tentacle, runs mainly to the
periphery of the tentacle, ramifies there, and ends in intimate
association with numerous large, clustered nerve cells near the
tip of the tentacle; at the level of the eye-cup, a small ramus is
split off from the main course of the nerve and passes to the
eye, but an actual connection with it, such as is easily seen in
many molluscan eyes, is exceeding difficult to demonstrate; our
evidence seems to show, however, that a few fibers perhaps do
actually enter the optic cup. This structural state may be
indicative of degeneration.

If the mantle receptors of O. floridanum must be regarded as
mere remnants of the original photosensitive equipment of this
stock, the possibility of their connection with a primitive helio-
tropic mechanism in ancestral pulmonates acquires an unprofit-
able vagueness. We have thought it necessary to raise this
point because it has sometimes been held that non-adaptive
responses ‘‘have been inherited from ancestors in which they
were adaptive” (meaning that the mechanism for response has
been so inherited). For Onchidium such interpretation is
highly improbable.

3. Neither can the heliotropism of Onchidium be dismissed as
a mere ‘laboratory product.’ Some writers have endeavored to
account for heliotropic orientations as found in various animals
on the basis that determinate movements of this character must
be the result of ‘abnormal’ conditions (cf. Franz, 713). It will
be obvious that notions of this sort cannot affect the analysis of
the mechanism of photic orientation, but can refer only to the
role of heliotropism as an ethologic factor. It is only in a very
limited sense that the heliotropism of Onchidium may be regarded |
as ‘unnatural.’ It is not that laboratory conditions artificially
imposed determine the orientations so produced, but on the
contrary that in surroundings other than the immediate environ-
ment of the ‘home’ nest some specific factor producing central
nervous inhibition of what may, for convenience, be termed the
(sensory) heliotropic impulses, fails to appear. It is sufficient to
remember that an Onchidium need only be transferred to a new

492 LESLIE B. AREY AND W. J. CROZIER

section of the shore in order to witness the complete unmasking
of its heliotropic impulses.

Since strychnine has in some instances been shown to produce
negative phototropism, even in animals naturally indifferent to
light (Moore, 12), it should be clearly understood that the
strychnine effect in Onchidium cannot be regarded as of this
sort.®

Certain animals are known to become photonegative upon
immersion in sea-water. Isopods of the genus Ligia, which in
certain places occupy territory also frequented by Onchidium,
have been said by Abbott (18) to reverse their phototropism,
perhaps under control of humidity, in the sense that at low tide
they come out from hiding places above flood-tide level and
wander over the exposed intertidal zone. That this behavior
really involves phototropism of any kind, and is not rather a
case similar in certain features to that of O. floridanum, remains
to be proved. It would be of interest to test this matter, for
Abbott states that ‘‘in the laboratory they (Ligia) give a nega-
tive reaction to sunlight.”” Moreover, an understanding of
the situation in Onchidium may be important for the elucidation
of other curious cases in which an animal’s heliotropism seems
fundamentally at variance with its mode of life (e.g., Para-
vortex, described by Ball, 716, p. 464).

According to Mitsukuri (’01), the specific habitat of Littorma
is determined by changes in its phototropism, from negative in
air and under water to positive when splashed by waves. Here,
again, the evidence that phototropism is really primarily involved
is somewhat defective. The heliotropism of Onchidium is in no
respect altered by complete immersion in sea-water.

9 Whether the action of strychnine in producing negative hehotropism with
an animal naturally photopositive or even indifferent to light (Moore, 712, 713)
can be always referred to chemical modifications within the primary receptors,
rather than to some more strictly.central nervous (synaptic) effect, remains
unanswered. Even with the human eye, where visual acuity (retinal resolving
power) is notably augmented by strychnine, one cannot at present be sure that
the removal of certain central inhibitions is not at bottom responsible. As an
example of the inhibition of one sensory impulse by another, we might cite the
heightened tactile responsiveness of certain de-eyed fishes (Crozier, 718).

NATURAL HISTORY OF ONCHIDIUM 493

For reasons already amply set forth, we must reject the notion
that the movements of Onchidium involve, or depend upon,
any ‘reversal’ of phototropism. From the standpoint of adap-
tation, the heliotropic mechanism must be regarded as a most
interesting example of a perfectly definite functional character-
istic which proceeds automatically from the given physicochem-
ical composition of the organism (cf. Loeb, ’16), without refer-
ence to adaptive requirements (cf. Arey and Crozier, ’18; Crozier
and Arey, ’19c). Since the young of Onchidium (developing
within the nest) emerge from the egg capsule with the form of
the adult, and not as veligers (Joyeux-Laffuie, ’82), and since
we have found very tiny individuals (2 mm. long) emerging from
nests with adults, it cannot be said that perhaps at an early
stage these animals are normally photonegative, by this means
first becoming established in their definitive nest.

4. A number of instances are on record of the preservation in
the rhythmic activities of animals of some diurnal or tidal
rhythmicity inherent in the environment (Wilson, ’00; Gamble
and Keeble, ’00; Schleip, 710; Keeble, ’10; Esterly, ’17; Cary, in
Dahlgren, ’16, reprint p. 11, etc.). Unfortunately, a number of
such reports, especially those concerning the persistence of
environmental rhythms in actinians, have proved to be the
result of erroneous observation (Parker, 719). We were inter-
ested to discover if, in a form like Onchidium exhibiting such
complex responses, there would be found any persistence of
either tidal or nycthemeral rhythms of activity and repose, in
the absence of the rhythmic excitations normally associated. It
can be said with confidence that no rhythms of this character
are maintained by Onchidium when removed to the laboratory.

5. For some time it has been known that the limpets and their
allies inhabiting the tidal zone may at times wander for some
little distance from, and subsequently return to, the ‘scar’
indicating their definite ‘home.’ The literature of this subject
has been reviewed in an interesting way by Piéron (’09c). A
certain complication enters here, for some limpets creep forth
from their scar when covered by the sea, others only when left
bare by the tide. Piéron (loc. cit.) has given a plausible account

494 LESLIE E. AREY AND W. J. CROZIER

of these differences, though certain of his statements could be
better weighted with evidence.

The general fact of the wandering of limpets from their scars
has been known since the time of Aristotle, and the fact of hom-
ing has more recently attracted the notice of a number of natural-
ists (Bethe, ’98; Davis and Fleure, ’03; Piéron, ’09a, ’09b,
’09c,’19; Bohn, ’09; Orton, ’14; Billiard, ’14; Wells, 717). It is all
the more curious that so little close experimental work has been
devoted to the elucidation of the matter. Homing activities
are shown by a number of more or less distinctly related forms—
Patella, Siphonaria, Helicon, Fissurella, Calyptraea; among
these, various degrees of ‘homing’ ability are manifest, and in
Acmaea testudinalis, according to Willcox (’05b), there is no
evidence of this activity at all. ‘Homing’ is found to be success-
fully executed by these animals even when they are artificially
removed from their scar or from some point along their feeding
path and replaced within a reasonable distance of the scar. For
Patella, Piéron (’09c) records successful returns following a
displacement of 12 cm., while, in certain localities, the extent of
the natural feeding journeys may be as great as 55 to 90 cm.,
according to various observers quoted by Piéron. For other
genera less distances limit the molluse’s successful return to its
home subsequent to experimental shifting—with Siphonaria,
2 to 3 inches or perhaps a little more; for Fissurella, 2 inches
(Willcox, ’05 a).

This kind of ‘homing’ has certain resemblances to that of
Onchidium, yet there are noteworthy differences. Piéron notes
that some individuals (Patella) wander little or not at all in
securing food; these are easily ‘lost,’ and do not succeed in
returning to their scars if even slightly displaced. Piéron
regards the homing as dependent upon a permanent memory of
the topography of the habitual situation, and upon a very
exact memory of the relief of the spot on which the Patella
orients itself according to the irregularities of its shell.1° He
succeeded in demonstrating that a Patella could so orient itself,

10Cf. Piéron, 719.

NATURAL HISTORY OF ONCHIDIUM A95

even when the margin of the shell had been chipped away, and
concluded therefore that the ‘topographic memory’ involved
must be a sensory affair. The deduction is reasonable that the
cephalic tentacles are the essential guiding organs, particularly
in creeping, and that the pallial tentacles serve this function
while the Patella is adjusting its irregularly outlined shell to the
depression of the scar. But it should not be forgotten that the
experimental test of this conclusion, particularly in so far as it
pertains to the somewhat obscure ‘topographic memory,’ has
yet to be instituted.

With reference to the bearing of these findings upon the
analysis of ‘homing’ in Onchidium, we need only point out that
the homeward creeping of the latter has a much more flexible
cast than is true of the behavior of the limpets, especially
in those experiments where an Onchidium removed from one
entrance of its nest and replaced upon the rock was found to
gain the nest again, but by a second, different aperture. The
distances from which a successful return is effected are also
notably greater in the case of Onchidium. Nevertheless, the
probability of any deep-seated ‘memory’ of the location of the
nest is negatived by the fact that confinement to the laboratory
for twenty-four hours obliterates the capacity for return to the
nest. Tests of this point with Fissurella and Patella are of great
interest; according to Piéron (’09b), the ability of Patella to
return to a particular ‘home’ can survive some days’ removal
from the scene. Even bees are said to lose their memory of
specific locations after being anaesthetised (quoted from Min-
nich, ’19).

The behavior of limpets is of greater significance in connec-
tion with the possible evolution of the ‘homing’ capacity. Some-
thing of this sort seems to have been in Wells’s (17) mind. It
can be pointed out that several grades of increasing precision
and complexity of ‘homing’ activity are shown by molluscs
(Arey and Crozier, 718). Beginning with Chiton tuberculatus
(Crozier, ’21), in which there can be found something like ‘hom-
ing,’ but of a rather vague type and pretty certainly the result
of immediate stimulations, a series comprising also Patella,

496 LESLIE B. AREY AND W. J. CROZIER

Onchidium, and Octopus exhibits more and more highly devel-
oped ‘homing’ propensities. The return of a Patella, Fissurella,
Siphonaria, or Calyptraea to its specific site cannot be accom-
plished beyond a relatively slight distance; these creatures also
tend to follow fairly definite paths in their excursions and to
adhere to these paths when returning; and some of them creep
but slightly, if at all, away from their scars. Onchidium’s
behavior is obviously an advance in respect to complexity.
Analogous behavior has been described for snails and slugs (as
in the famous story of the sick snail and its companion, cited
by Darwin, ’71, p. 316, and by others; cf. also Cooke, ’95, and
Scharff, 07). The investigation of this matter in snails and
slugs holds the possibility of considerable interest. Finally, the
behavior of Octopus (cf., e.g., Cowdry, 711); which returns to its
nest after extensive forays and from considerable distances,
under circumstances such that direct vision of the nest entrance
is completely excluded, represents the most complex form of this
activity among molluscs.

There has been a tendency to regard any series of this kind as
exhibiting stages in the evolution of a particular response, or
even of an instinct. To speak of a ‘homing instinct’ is little
short of a perversion of sense. Such a view-point is very likely
quite incorrect. Much more probable is it that this series of
forms displays merely stages in the evolution of the central
nervous machinery making possible more and more complicated
behavior. ‘The phrase ‘evolution of an instinct’ tends to obscure
the real basis of the matter. . Moreover, in the special instance
under discussion, it is not at all obvious that the actual ‘homing’
performances of the several types named are in any sense geneti-
cally connected; any relation with the mechanism of homing in
higher forms, birds, for example, is in the highest degree improb-
able. Even the homing of ants involves certain characteristics,
such as those described by Cornetz (14), which are not in any
sense represented in the behavior of Onchidium.

We early recognized the simulation of associative memory in
the activities of Onchidium (Arey and Crozier, ’18) with refer-
ence to its nest. If the notion of such memory or ‘beginnings of

NATURAL HISTORY OF ONCHIDIUM 497

intelligence’ be valid for cephalopods (v. Uexkill, ’01; Poli-
-manti, 710), it is legitimate to inquire if anything of this nature
can be imputed to Onchidium. According to Miss Thompson
(17), the snail Physagyrina, although in maze experiments it
gives no evidence of learning, does exhibit the establishment of
simple associations when tested by Pavloff’s method of ‘con-
ditioned responses.’ There is no real evidence favoring the
idea of memory as evinced in the ‘homing’ of O. floridanum.
One adequate test of the conceivable action of associations or
even of primitive intelligence has occurred to us. When an
Onchidium is picked up and put down on a strange portion of
the shore, it cannot, of course, find its old nest; but other nests
and various unoccupied crannies are available for shelter. The
fact is, however, that instead of seeking the shelter afforded by
‘strange’ crevices, the snail is on the contrary at the mercy of
two chief modes of response: its negative phototropism and its
withdrawing reaction when shaded; that specific quality of its
own particular nest which probably determines homing makes
it possible for the creature to enter its own nest notwithstanding
its photic sensitivities. Strayed Onchidia do not find shelter in
new cavities of the rock, but on the contrary creep about on the
shore until covered and washed off by the returning tide. Evi-
dence of intelligence or of adaptive use of associative memory is
completely absent, although, as we have elsewhere remarked
(Arey and Crozier, ’18), the close simulation of behavior of that
order is certainly deceptive.

SUMMARY

Onchidium (Onchidella) floridanum is a small naked pulmo-
nate inhabiting the intertidal shore zone at Bermuda. ‘The indi-
viduals of this species are grouped together into colonies num-
bering about a dozen individuals, more or less, in each. A
colony during high water occupies a ‘nest,’ in the form of an
eroded cavity in the shore rock or a cleft between clay-cemented
stones. During the day-time only, and at most but once in the
twenty-four hours, the Onchidia emerge from their nest after the
falling tide has left it above water for about an hour. The
animals feed for a fixed period of about one hour, then those

498 LESLIE B. AREY AND W. J. CROZIER

individuals emanating from a given nest begin simultaneously
to execute a direct return to the nest from which they originated.
These animals will not enter a ‘foreign’ nest.

When tested apart from their specific normal environment the
Onchidia are always negatively phototrophic. In the natural
state their movements are entirely independent of heliotropism.
This independence can be obliterated by injected strychnine,
which produces ‘reversal of inhibition.’ Similarly, the simul-
taneous return to the nest on the part of the various members
of a colony can be understood on the assumption of a ‘reversal
of inhibition’ brought about by substances derived from mate-
rials ingested while feeding.

The impulses which, on this hypothesis, suffer central inhibi-
tion in the outwardly creeping snail may be identified with those
which normally control the determinate character of the home-
ward course. These impulses probably originate in the oral
lappets, and are taken to have the character of a ‘contact odor’
(see text) specific for each particular nest.

This is the only hypothesis which can account for the observed
peculiarities of the movements of Onchidium and for the out-
come of the experiments concerning homing reported in this
paper.

There is no evidence of associative or persisting memory in
connection with homing, nor do other activities of Onchidium
point to the existence in this form of anything approaching
intelligent behavior. Responses to immediate stimulations are
adequate for the analysis of the situation. |

The negative heliotropism of Onchidium, apparently devoid
of adaptive significance, is accounted for in terms of a photo-
sensitive receptor system enabling these snails to respond to
shading by an effective use of the mantle as a hold-fast, supple-
menting the weak suctional efficiency of the foot. The existence
of receptors making negative heliotropism possible cannot be
understood as a condition persisting from ancestral pulmonates
normally responding in this way.

Mantle eyes are absent in this species, and although the ten-
tacular eyes are perhaps of normal structure, no photic sensi-
tivity has been discovered in connection with them.

NATURAL HISTORY OF ONCHIDIUM 499

The snails do not emerge to feed when high winds and surf
are directed against their feeding zones. The tentacles are
anemotropic organs, which prevent the emergence of the Onchidia
in such circumstances.

An analysis has been given of some of the snail’s modes of
response, and of certain general implications of the remarkable
activities of Onchidium.

LITERATURE CITED

Axppott, C. H. 1918 Reactions of land isopods to light. Jour. Exp. Zodl.,
vol. 27, pp. 193-246.

Arey, L. B. 1918 The multiple sensory activities of the so-called rhinophore
of nudibranchs. Amer. Jour. Physiol., vol. 46, pp. 526-532.

Arey, L. B., anp Crozier, W. J. 1918 The ‘homing habits’ of the pulmonate

mollusk Onchidium. Proc. Nat. Acad. Sci., vol. 4, pp. 319-321.

1919 The sensory responses of Chiton. Jour. Exp. Zoél., vol. 29,

pp. 157-260.

Baz, 8. 1916 The development of Paravortex gemellipara (Graffilla gemelli-
para Linton). Jour. Morph., vol. 27, pp. 453-557.

Bancroft, F. W. 1913 Heliotropism, differential sensibility; and galvanotro-
pism in Euglena. Jour. Exp. Zodél., vol. 15, pp. 383-428.

Bereu,R. 1895 Uber die Verwandtschaftsbeziehungen der Onchidien. Morph.
Jahrb., Bd. 10, S. 172-181.

Betue, A. 1898 Diirfen wir den Ameisen und Bienen psychische Qualititen
zuschreiben? Arch. f. ges. Physiol., Bd. 70, p. 15-100. ;

BituiarD, J. 1914 Sur la locomotion chez les patelles. Bull. Soc. Zool. France,
no. 7, pp. 325-326.

Boun, G. 1909 De Vorientation chez les patelles. C. R. Acad. Sci., T. 148,
pp. 868-870.

BRETNALL, R. W. 1919 Onchidiide from Australia and the Southwestern Pa-
cific Islands. Rec. Austral. Mus., vol. 12, pp. 303-328.

Brown, T. G. 1914 On the nature of the fundamental activity of the nervous
centers; together with an analysis of the conditioning of rhythmic
activity in progression, and a theory of the evolution of function in
the nervous system. Jour. Physiol., vol. 48, pp. 18-46.

CHAMBERLAIN, R. V. 1920 The myriopod fauna of the Bermuda islands. Ann.
Ent. Soc. Amer., vol. 13, pp. 271-302.

Cooks, A. H. 1895 Molluscs, in Cambr. Nat. Hist., vol. 3, pp. 1-459.

CorneEtTz, V. 1914 Les explorations et les voyages des Fourmis. Paris.

Cowpry, E. V. 1911 The color changes of Octopus vulgaris Lmk. Contrib.
Bermuda Biol. Sta., no. 22 (vol. 2); from: Univ. Toronto Stud., Biol.
Ser., no. 10, 153 pp.

Crozinr, W. J. 1914 The orientation of a holothurian by light. Amer. Jour.
Physiol., vol. 36, pp. 8-20.

1915 The sensory reactions of Holothuria surinamensis Ludw. Zool.
Jahrb., Abt., Physiol., Bd. 35, S. 233-297.

500 LESLIE B. AREY AND W. J. CROZIER

Crozier, W. J. 1917 The nature of the conical bodies on the mantle of certain
nudibranchs. Nautilus, vol. 30, pp. 103-106.
1918 On tactile responses of the de-eyed hamlet (Epinephelus stri-
atus). Jour. Comp. Neur., vol. 29, pp. 163-173.
1919 a On the use of the foot in some mollusks. Jour. Exp. Zodl.,
vol. 27, pp. 359-366.
1919b On the resistance of Fundulus to concentrated sea-water.
Amer. Nat., vol. 53, pp. 180-185.
1919 ce On the control of the response to shading in the branchiae of
Chromodoris. Jour: Gen. Physiol., vol. 1, pp. 585-591.
1920 On the analysis of neuromuscular mechanisms in Chiton. Jbid.,
vol. 2, pp. 627-634.
1921 The question of ‘homing’ eiaea in Chiton. Amer. Nat. (In
press. )

Crozter, W. J., AnD Arny, L. B. 1918 On the significance of the reaction to
shading in Chiton. Amer. Jour. Physiol., vol. 46, pp. 487-492.
1919 a Onchidium and the question of adaptive coloration. Amer.
Nat., vol. 53, pp. 415-4380.
1919 b Sensory reactions of Chromodoris zebra. Jour. Exp. Zodl.,
vol. 29, pp. 261-310.
1919 c¢ The heliotropism of Onchidium: a problem in the analysis of
animal conduct. Jour. Gen. Physiol., vol. 2, pp. 107-112.

DanucREen, U. 1916 The production of light by animals (Porifera and Coe-
lenterata). Jour. Franklin Inst., Feb., pp. 1-19.

Dat, W. H., anp Srupson, C. T. 1901 The mollusca of Porto Rico. U. S.
Fish. Comm. Bull. for 1900, vol. 1, pp. 351-524.

_ Darwin, C. 1871 The descent of man, vol. 1. New York.

Davis, J. R. A. 1895 The habits of limpets. Nature, vol. 51, pp. 511, 512.

Davis, J. R. A., AND Firvre, H. J. 1903 Patella. Liverpool M. B. C. Mem.,
X. London.

Dotiny, W. J., Jr. 1917 The rate of locomotion in Vanessa antiopa in inter-
mittent light and in continuous light of different illuminations, and
its bearing on the ‘continuous action theory’ of orientation. Jour.
Exp. Zoél., vol. 23, pp. 507-518.

Exiot, GC. 1899 Notes on tectibranchs and naked mollusks from Samoa. Proc.
Acad. Nat. Sci. Phila., 1899, pp. 512-523.

Esterty, C.O. 1917 The occurrence of a rhythm in the geotropism of two spe-
cies of plankton copepods when certain recurring external conditions
are absent. Univ. Calif. Publ., Zoél., vol. 16, pp. 393-400.

Franz, V. 1913 Die phototaktischen Erscheinungen im Tierreiche und ihre
Rolle in Frei-leben der Tiere. Zool. Jahrb., Abt. Physiol., Bd. 33,
S. 259-286.

Fusira, T. 1897 [On the method of respiration in Onchidium.| (Japanese.)
Dobutsu Gaku Zasshi (Zodl. Magaz.), vol. 7.

GamBie, F. W., anp Kresie, F.W. 1900 Hippolyte varians: a study in color-
change. Quart. Jour. Micr. Sci., N. S., vol. 49, pp. 589-699.

Garrey, W. 1918 Light and the muscle tonus of insects. The heliotropic
mechanism. Jour. Gen. Physiol., vol. 1, pp. 101-126.

Grn, W. 1913 The behavior of leeches, with special reference to its modifia-
bility. Univ. Calif. Publ., Zodl., vol. 11, pp. 197-805.

NATURAL HISTORY OF ONCHIDIUM 501

Hecut, S. 1919 Sensory equilibrium and dark adaptation in Mya arenaria.
Jour. Gen. Physiol., vol. 1, pp. 545-558.

Hess, C. 1912 Gesichtssinn, 3: Akkommodation. In: Winterstein, Handb.
vergl. Physiol., Bd. 4, 9, S. 709-840.

Hirasaka, K. 1912 [On the structure of the dorsal eye of Onchidium.] (Jap-

anese.) Dobutsu Gaku Zasshi (Zool. Magaz.), vol. 24, pp. 20-35.

_ Jennines, H. 8. 1907 Behavior of the starfish, Asterias forreri de Loriol.
Univ. Calif. Publ., Zodl., vol. 4, pp. 53-185.

Joyrux-LaFrrulkz, J. 1882 Organisation et développement de l’Oncidie. Arch.
Zool. expér. et gén., T. 10, pp. 225-383.

Karka, G. 1914 Einfiihrung in die Tierpsychologie, Bd. 1: Die Sinne der Wir-
bellosen. Leipzig.

Kress, F. 1910 Plant-animals: a study in symbiosis. Cambridge.

Lors, J. 1916 The organism as a whole. New York.
1918 Forced movements, tropisms, and animal conduct. Philadel-
phia.

Micnaet, E. L. 1916 Dependence of marine biology upon hydrography and
necessity of quantitative biological research. Univ. Calif. Publ.,
Zool., vol. 15, pp. 1-xxill.

Minnicu, D. E. 1919 The photic reactions of the honey-bee, Apis mellifera L.
Jour. Exp. Zo6l., vol. 29, pp. 343-425.

Mirtsuxour!, K. 1901 Negative phototaxis and other properties of Littorina as
factors in determining its habitat. Annot. Zool. Japon., vol. 4, pp.
1-19.

Moors, A. R. 1912 Negative phototropism in Diaptomus by means of strych-
nine. Univ. Calif. Publ., Physiol., vol. 4, pp. 185-186.
1913 The negative phototropism of Diaptomus through the agency of
caffein, strychnine and atropin. Science, N. §., vol. 38, pp. 131-133. —

Murray, J. 1885 Narr. of the Chall. Expedition, vol. I, pt. 1 (Onchidium,
p. 489).

OumsteD, J. M.D. 1917a Notes on the locomotion of certain Bermudian mol-
lusks. Jour. Exp. Zodl., vol. 24, pp. 223-236.
1917 b Geotropism in Planaria maculata. Jour. Anim. Behav., vol.
7, pp. 81-86.

Orton, J. H. 1914 Note on the feeding habits of Patella. Jour. Mar. Biol.
Assn., N. S., vol. 10, pp. 254-257.

Parker, G. H. 1911 The mechanism of locomotion in gastropods. Jour.
Morph., vol. 22, pp. 155-1709.
1919 The elementary nervous system. Philadelphia.

Parren, B. M. 1915 A device for projecting-a small spot of light suitable for
exploring photosensitive areas. Science, N.8., vol. 41, pp. 141-142.

PELSENEER, P. 1901 Etudes sur les gastropodes pulmonés. Mem. Acad. Se.
Roy. Belg., T. 54.

Perrier, E. 1917 Sur les échanges de faune entre la mer et les eaux donces
et les conséquences qu’ils entrainent au point de vue de la sexualite.
C. R. Acad. Sci., T. 165, pp. 748-751.

Preyrica, E., er Vues, F. 1913 Notes sur quelques relations numériques rela-
tives aux ondes pédieuses des Gastéropodes. Bull. Soc. Zool., Fr.,
1913, pp. 151-254.

502 LESLIE B. AREY AND W. J. CROZIER

PrérRon, H. 1909a Sens de l’orientation et mémoire topographique de la Pa-
telle. C. R. Acad. Sci., T. 148, pp. 530-532.

1909 b Contribution 4 la biologie de la patelle et de la Calyptrée.
Le sens de rétour et la mémoire topographique.. Arch. Zool. expér.
et gen., Notes et Rev., ser. 5, T. 1, pp. xviii-xxix.

1909 e¢ Contribution a la bidlosie de la patelle et de la calyptrée,
Bull. Sci. Fr. et Belg., T. 48, pp. 183-202.

1919 De Vimportance ceded des divers facteurs sensoriels dans
le sens du retour de la Patelle. C. R. Soc. Biol., T. 82, pp. 1227-1230.

Puate, L. 1894 Studien iiber opisthopneumone Lungenschnecken, II. Zool.
Jahrb., Abt. Anat., Bd. 7, S. 93-234.

PotimantI,O. 1910 Les céphalopodes ont-ils une mémoire? Arch. de Psychol.,
T. 10, pp. 84-87.

Risaut, H. 1915 Biospeologica: xxxvi. Notostigmorpha, Scolopendromorpha,
Geophilomorpha. Arch. Zool. expér. et gén., T. 55, pp. 323-346.

Romanes, G. J. (1912, repr.) Animal intelligence. New York.

ScuarFr, R. F. 1907 European animals. New York. 14 + 258 pp.

Scuierp, W. 1910 Der Farbenwechsel von Dixippus morosus (Phasmidae).
Zool. Jahrb., Abt. Zool., Bd. 30, pp. 45-132.

SempER, K. 1877 a Reisen im Archipel der Philippinen, Theil 2, Bd. 3, Land-
mollusken, Ergiinzungsheft, S. 146, Taf. A-E.

Semper, C. 1877 b Ueber Schneckenaugen vom Wirbelthiertypus nebst Bemer-
kungen iiber einige andere histologische Eigenthumlichkeiten ver-
schiedener Cephalophoren. Arch. mikr. Anat., Bd. 14, pp. 117-124.
1881 Animal life as affected by the natural conditions of existence.
New York.

SHERRINGTON, C.8. 1906 The integrative action of the nervous system. New
York.

SranrscHinsky,W. 1907 Zur Anatomie und Systematik der Gattung Oncidium.
Zool. Jahrb., Abt. f. Syst., Bd. 25, p. 353-402.

1908 Uber den Bau der Riickenaugen und der Histologie der Riicken-
region der Oncidium. Zeit. wiss. Zool. Bd. 90, S. 137-178.

THompson, ExvizaBperaH lL. 1917 Experiments with Physagyrina. Behav.
Monogr., Cambr., ili, no. 3, p. 1-89.

UEXKULL, J. von. 1909 Umwelt und Innenwelt Tiere. Berlin.

Wetts, M. M. 1917 The behavior of limpets, with particular reference to the
homing instinct. Jour. Anim. Behav., vol. 7, pp. 387-395.
Witicox, M. A. 1905 a Homing of Fissurella and Siphonaria. Science, N.S.,

vol. 22, pp. 90, 91.
1905 b Biology of Acmaea testudinalis Miller. Amer. Nat., vol. 39,
pp. 325-333.

Witson, C. B. 1900 The habits and early development of Cerebratulus lacteus
(Verrill). A contribution to physiological morphology. Quart. Jour.
Mier. Sci., N. 8. vol. 48, pp. 97-198.

WissEL, K. von 1898 Beitriige zur Anatomie der Gattung Oncidiella. Zool.
Jahrb., Suppl. 4 (Fauna Chilensis, Bd. 1), S. 583-640.

Yune, E. 1910 La sensibilité des gastropodes terrestres pour la lumiére. Arch.
Sci. Phys. et Nat., T. 30, pp. 617-618.

1913 La cécité des gastéropodes pulmonés. Arch. Sci. Phys. et Nat.,
To35 pedde

SUBJECT AND AUTHOR INDEX

Earner ey mississippiensis and its pho-
tomechanical changes. Studies on the
retina. The structure of the retina of.. 207

Amblystoma punctatum. Experiments on

the effect of removal of the pronephros of. 355
Ameba proteus. Food reactions of.......... 397
Anurans. I. The male sexual cycle of Rana

catesbeiana larvae. The germ cells of.... 235
Arny, Lestip B., AND Croziur, W.J. On the

natural history of Onchidium............ 443
Asilus sericeus Say. Spermatogenesis in the
IVA coer aroe iis nists ie le sis siotohorelaysjerais ss) 165

ODINE, Josrern Hau. Factors influenc-
ing the water content and the rate of

metabolism of certain Orthoptera...... 137
Bures, E. L., Burce, W.E., anp. Anexpla-
nation for the variations in the intensity

of oxidation in the life-cycle............. 203
Bure, W. E., anp BurcE, E. L. An expla-
nation for the variations in the intensity

of oxidation in the life-cycle............. 203

ELLS of anurans. I. The male sexual
cycle of Rana catesbeiana larvae. The

germ

Changes. Studies on the retina. The struc-
ture of the retina of Alligator mississip-
piensis and its photomechanical.......: 207

Chemical sense of Nereis virens Sars.
Feeding habits sad. ock ecitec eeeeis sere syeuyaie et 427

Content and the rate of metabolism of cer-
tain Orthoptera. Factors influencing the

RU LETAN herrea Me ecene i hiatoe siete ne 137
Crossing over. Further studies on the effect _

ofstem perature Ole tncaes case clos stn 187
Crossing over. I. The effect of selection on

crossover values. Studies on............. 333
Crozier, W. J., AREY, Lestie B., AND. On

the natural history of Onchidium........ 443
Cycle of Rana catesbeiana larvae. The germ

cells of anurans. I. The male sexual.... 235

ETLEFSEN, J. A, aND Roperts, E.
Studies on crossing over. I. The effect

of selection on crossover values........ 333
DerwiteR, 8. R., LaAurRENs, HENRY, AND.
Studies on the retina. The structure of
the retina of Alligator mississippiensis and

its photomechanical changes.............. 207

EEDING habits and chemical sense of
Nereis virens Sars. The............... 427

Food reactions of Ameba proteus............. ¢97

ERM cells of anurans. I. The male sex-
ual cycle of Rana catesbeiana larvae.
AY ce 5 tae coe ene 6 WE esa asta ee 235
Gross, ALFRED O. The feeding habits and
chemical sense of Nereis virens Sars..... 427

| eee and chemical sense of Nereis vi-
rens Sars. The feeding............... 427

Harrison, Ross G. On relations of sym-
metry in transplanted limbs..............

503

History of Onchidium. On the natural...... 443

Howxanpd, Ruta B. Experiments on the
effect of removal of the pronephros of %
Amblystoma punctatum................. 355

| (ee es of oxidation in the life-cycle.
An explanation for the variations in the.. 203

EPNER, Wm. A., AND WuuitTtock, W.
Caru. Food reactions of Ameba pro- ™

| Pig Henry, AND DEeTwILer, S. R.
Studies on the retina. The structure of
the retina of Alligator mississippiensis
_, and its photomechanical changes......... 207
Life-cycle. An explanation of the variations
. in the intensity of oxidation in the....... 203
Limbs. On relations of symmetry in trans-
plantedhepere cee... '.8 sa a eee 1

ETABOLISM of certain Orthoptera.
Factors influencing the water content
angdithe ratelOtine.: ccc sae sce sears eels 137

Mzrz, CHARLES W., AND Nonipez, José F.

Spermatogenesis in the fly, Asilus sericeus

ROT Bs sooenes ou SeOee matoe De saae Soa LOD,

Nee eS history of Onchidium. On the 443

Nereis virens Sars. The feeding habits and

SERS EE ino nwoc: 2 ODP ARMOR en SO ETS aE SCOOT 165
(@yanenes On the natural history of 443

Orthoptera. Factors influencing the water
content and the rate of metabolism of cer-
UCN OITA Ae 5 3 ek CORE RR CERIO ce DOS COE 137

Over. Further studies on the effect of tem-
perature On crossing..........-.-.---+-+-- 187

Over. I. The effect of selection on crossover
values. Studies on crossing.............. 333

Oxidation in the life-cycle. An explanation
for the variations in the intensity of...... 203

De ive a changes. Studies
on the retina. The structure of the ret-
ina of Alligator mississippiensis and its. 207
PLovucu, Harotp H. Further studies on the
effect of temperature on crossing over.... 187
Pronephros of Amblystoma punctatum. Ex-
periments on the effect of the removal of

Proteus. Food reactions of Ameba........... 397
Punctatum. Experiments on the effect of
removal of the pronephros of Amblystoma 355

Re catesbeiana larvae. The germ cells
ofanurans. I. The male sexual cycle of 235

Reactions of Ameba proteus. Food.......... 397
Removal of the pronephros of Amblystoma
punctatum. Experiments on the effect of 355

504 INDEX

Retina. The structure of the retina of Alli-
gator mississippiensis and its photome-
chanical changes. Studies on the........ 207

Rosekrts, E., DETLEFSEN,J.A, AND. Studies
on crossingover. I. Theeffect of selection

GNICIOSSOVEL) VALUESs--ceee eee annex on 333
ARS. The feeding habits and chemical
sense of Nereis virems................0. 427
Selection on crossover values. Studies on
crossing over. I. The effect of........... 333
Sense of Nereis virens Sars. The feeding
habitsiancrchemical-sseeeereren cs ase. ce 427
Sericeus Say. Spermatogenesis in the fly,
ANSI aes, toe ee pole rie Sion 165
Sexual cycle of Rana catesbeiana larvae. The
germ cells of anurans. I. The male...... 235

Spermatogenesis in the fly, Asilus sericeus Say 165

Swincie, WILBUR WILuis. The germ cells of
anurans. I. The male sexual cycle of
Rana catesbeiana larvae................. 235

Symmetry in transplanted limbs. On rela-
HONS Of As5s). Mosel ee cee ee ee

ERE ES ATURE on crossing over. Fur-
ther studies on the effect of............ 187

Transplanted limbs. On relations of sym-
Metry- IN). hehe ee eee

Vee in the intensity of oxidation
in the life-cycle. Anexplanation for the 203

Virens Sars. The feeding habits and chemical
sense:of Nereis:04)..4,32h ono eee 427

ATER content and the rate of metabo-
lism of certain Orthoptera. Factors

® “influenempqthee». 1 taee aun eee eee 137
Wuittock, W. Cart, Kepner, Wn. A., AND.
Food reactions of Ameba proteus........ 397

)

5 WHSE 020

pa)
fe}
=
=
4
ray
=

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "The Journal of experimental zoology"

See other formats