. ^ i • II • • " I «« I ,1.11 i ■■! • T. !,'• . . ■„ ■■.r. .1 n .1 ii>if'i, i, 4 .< 4 <■ 1H V“ ■ ‘ *' ■ r '• ■'•'1 ; - ■ ■■ - ■•. « nni ir,. .■ < ><

Cv;::

i 'V*'^:"i V f' ’'i\'n^''.|^i*:* ," *'^ ‘i'.,^;''.''''’i, i\i’‘i.‘('i ,''.S^‘ I’V'u' i’’ V '"i'V'u .'' A*" ■ '•r:l..i*.''l <' «V'*< ti t t^ij'n'n * L '•'•i i/j»' '<,«.;',i, ^j‘:'',‘< ■’;V;i'*.,'\'*ii’wv/.\i'’ii'fr/^^ n’ViV ''N'«'-m n -rV A>'ii!i* ’ ;4'"V‘ " ’"■ •.■ iV " '• '

■- 1 * . iK--> 'M if'*--* •> 'r • .?■•’■•'

.Vv'iivy., 1<'^ I II ■'-.\ .■”-,V'i. < ^-r ‘l' H,.\- '.\y. y'l t

y,^'y.,V;y.^v^v^'-':yy.vy!;y^^y^s 'yrVV'iA.’iViri/'^^v^oy^cvyir^ v'lViy

; <yVVvV-:i,y.y;i 'r/j'- yy-' ‘y;v v.'^-v\V..V'’' vvvvy vv^i^-ViV^ y

■ 7 '*' -j'.i\\‘i\ <!' | 4 ‘"4 ] ''/'i* ,><'•»,, /<!»'> '• tj •»;•( 1 .^1 ','1 '< ,* i ^'iV'i)^ ’ '* '■ '* •'*'^i'^ i''i(^^ '" i-'''f*''t’ I r '' ,i li l* « '"i - " ” 7 ' ' ■ « ( 'i.<*’< « ••' * N *j, *i'4 .4 'i iMi • ‘ | </ • '1 vv ' ,'• ’’j ' ^V ruV'ii'' V*^-\'iV'y ’*’7^r'lirt ^ i '':‘'i .i • 'J V \

■i

I

' ®ii V ® ,v. v; > v> W'V i , V i :n4:X ■ ■( i <i j •

vv*'.-v«v.?* v.'.'i’iV. ■A'A'i .v?^?4;’

liiilliililiSiiliillill^

xy '';.■V■y•x!y||'*,)y,’'Xy4yvv''■y'v.iyx1^y!l<’^''■’*'•'*‘^[Xxi'*''^‘-y

J ,y*V «1 . ., yy , .^ V>y iyy<VV>,.«A‘^^

y'y‘:':!'y^^vy7vyy^y:K''i^S’y®''^'-yv';y;y;iy;;-i;>y::'v*'KX'C

r, 4. ti l’s ''|’:,^y

,a. 4 «Vl 1 -f %V, '!;*•'*

, v*4 v.i*. •i’4'«i

, ■ i^v-t y i r i'"‘i'‘ vXi.yyyi

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 40 • 1955 • NUMBERS 1 TO 17

PUBLISHED BY THE SOCIETY
The ZOOLOGICAL PARK, ^ew York

/

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

30 East Fortieth Street, New York 16, N. Y.

PUBLICATION OFFICE

The Zoological Park, New York 60, N. Y.

OFFICERS
PRESIDENT
Fairfield Osborn

SECRETARY
Harold J. O’Connell

VICE-PRESIDENTS TREASURER

Alfred Ely David H. McAlpin

Laurance S. Rockefeller
Donald T. Carlisle

ASSISTANT TREASURER
Percy Chubb, 2nd

SCIENTIFIC STAFF: Zoological Park and Aquarium

John Tee- Van Director

Leonard J. Goss Assistant Director

ZOOLOGICAL PARK

Richard H. Manville . Curator of Mammals

Grace Davall Assistant Curator,

Mammals and Birds

James A. Oliver Curator of Reptiles

Leonard!. Goss. .... Veterinarian
Charles P. Gandal. . .Assistant Veterinarian

Lee S. Crandall General Curator

Emeritus

William Beebe Honorary Curator,

Birds

AQUARIUM

Christopher W. Coates . Curator & Aquarist

James W. Atz. Assistant Curator

Ross F. Nigrelli Pathologist

Myron Gordon Geneticist

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . .Research Associate

in Histology

Homer W. Smith Research Associate

in Physiology

DEPARTMENT OF TROPICAL
RESEARCH

William Beebe Director Emeritus

Jocelyn Crane Assistant Director

Henry Fleming ....... Entomologist

Rosemary Kenedy Field Assistant

John Tee- Van Associate

William K. Gregory Associate

AFFILIATES

C. R. Carpenter Co-ordinator, Animal

Behavior Research Programs

L. Floyd Clarke Curator,

Jackson Hole Research Station

SCIENTIFIC ADVISORY COUNCIL
A. Raymond Dochez Caryl P. Haskins
Alfred E. Emerson K. S. Lashley

W. A. Hagan John S. Nicholas

EDITORIAL COMMITTEE

GENERAL Fairfield Osborn, Chairman

William Bridges Editor & Curator , James W. Atz Lee S. Crandall

Publications William Beebe Leonard J. Goss

Sam Dunton Photographer William Bridges James A. Oliver

Henry M. Lester. . .Photographic Consultant Christopher W. Coates John Tee-Van

Contents

Part 1. May 14, 1955

PAGE

1. Revision of the Genus Cosmotoma Blanchard. (Coleoptera, Ceramby-

cidae, Lamiinae). By E. F. Gilmour. Text-figures 1-15 1

2. Two Little-known Selective Insect Attractants. By William Beebe.

Plates I-IV 27

3. A Factor Analysis of the Performance of Dogs on Certain Learning Tests.

By Anne Anastasia, J. L. Fuller, J. P. Scott & J. R. Schmitt. Text-
figures 1 & 2 33

4. The Chemical Nature of Holothurin, a Toxic Principle from the Sea-

cucumber (Echinodermata: Holothurioidea). By Ross F. Nigrelli, J. D.
Chanley, Stella K. Kohn & Harry Sobotka 47

5. The Effects of Holothurin, a Steroid Saponin of Animal Origin, on Krebs-2
Ascites Tumors in Swiss Mice. By T. D. Sullivan, K. T. Ladue & Ross

F. Nigrelli 49

Part 2. August 19, 1955

6. Regeneration of Melanomas in Fishes. By Recai Ermin & Myron

Gordon. Plates I-IV; Text-figures 1-45 53

7. Further Notes on the Pigmentary Behavior of Chaetodiptenis in Reference

to Background and Water Transparency. By C. M. Breder, Jr., & Pris-
cilla Rasquin. Plates I & II 85

8. Special Features of Visibility Reduction in Flatfishes. By C. M. Breder, Jr.

Plates I-III 91

9. Further Chemical Analysis of Holothurin, the Saponin-like Steroid from
the Sea-cucumber. By J. D. Chanley, Stella K. Kohn, Ross F. Nigrelli

& Harry Sobotka 99

Part 3. November 14, 1955

PAGE

10. A Comparative Study of the Morphology and Histochemistry of the

Reptilian Adrenal Gland. By William B. Hebard & Harry A. Charipper.
Plates I-V; Text-figure 1 101

1 1 . Hormonal Control of the Sexually Dimorphic Pigmentation of Thalassoma

bijasciatum. By Louise M. Stoll. Plates I-III 125

12. The Use of Copper Sulfate as a Cure for Fish Diseases Caused by Parasitic
Dinoflagellates of the Genus Oodinium. By Robert P. Dempster. Plate

I; Text-figure 1 133

13. Polymorphism in Reared Broods of Heliconius Butterflies from Surinam

and Trinidad. By William Beebe. Plates I-VI 139

14. Formation of a Mucous Envelope at Night by Parrot Fishes. By Howard

Elliott Winn. Plate 1 145

Part 4. December 31, 1955

15. A Revision of the Surgeon Fish Genus Ctenochaetus, Family Acanthuri-
dae, with Descriptions of Five New Species. By John E. Randall. Plates

I & II; Text-figures 1-3 149

16. Imaginal Behavior of a Trinidad Butterfly, Heliconius erato hydara Hewit-

son, with Special Reference to the Social Use of Color. By Jocelyn Crane.
Plates I-III; Text-figures 1 & 2 167

17. Three More Gymnotid Eels Found to Be Electrogenic. By Christopher

W. Coates. Plates I-III 197

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 40 • PART 1 • MAY 14, 1955 • NUMBERS 1 TO 5

PUBLISHED BY THE SOCIETY
rhe ZOOLOGICAL PARK, New York

Contents

PAGE

1. Revision of , the Genus Cosmotoma Blanchard. (Coleoptera, Ceramby-

cidae, Lamiinae). By E. F. Gilmour. Text-figures 1-15 1

2. Two Little-known Selective Insect Attractants. By William Beebe.

Plates I-IV 27

3. A Factor Analysis of the Performance of Dogs on Certain Learning Tests.

By Anne Anastasia, J. L. Fuller, J. Pj Scott & J. R. Schmitt. Text-
figures 1 & 2 33

4. The Chemical Nature of Holothurin, a Toxic Principle from the Sea-

cucumber (Echinodermata: Holothurioidea). By Ross F. Nigrelli, J. D.
Chanley, Stella K. Kohn & Harry Sobotka 47

5. The Effects of Holothurin, a Steroid Saponin of Animal Origin, on Krebs-2
Ascites Tumors in Swiss Mice. By T. D. Sullivan, K. T. Ladue & Ross

F. Nigrelli 49

1

Revision of the Genus Cosmotoma Blanchard.
(Coleoptera, Cerambycidae, Lamiinae)

E. F. Gilmour

Director, Museum arid Art Gallery, Doncaster, Yorks, England
(Text-figures 1-15)

This paper presents the results of ex-
amination of the genus Cosmotoma
Blanchard, incurred during broad pre-
liminary investigations of the Neotropical
Acanthocinini as a whole.

The species of this genus have proved to be
comparatively rare in collections, at least in
series, although several new species are herein
described, some of which have up to now stood
under other names. The synonymy of the genus
is herein rectified, having up to the present been
to a large extent given wrongly in various
catalogues and papers.

Grateful acknowledgment is made to the fol-
lowing institutions and individuals for their
kindness in sending specimens for examina-
tion:—

The United States National Museum, Wash-
ington.

The American Museum of Natural History,
N. Y.

The Musee Royal d’Histoire Naturelle de
Belgique, Brussels.

Zoologische Staatssammlungen, Munich.

The Hincks-Dibb collection (per Dr. W. D.
Hincks), Manchester.

The Dr. P. Lepesme collection, Paris.

The Dr. J. M. Bosq collection, Buenos Aires.
The specimens used for the basic investiga-
tion are in the author’s collection.

I would also like to take this opportunity of
expressing my very grateful thanks to the New
York Zoological Society for its grant to me
during 1953 from the Society’s Program for
Aid of Biological Research in Europe.

A distribution map of the genus is given
(Text-figure 1) in this paper.

Cosmotoma Blanchard

Blanchard, 1845, Hist. Nat. Ins., 2, 155.—
Bates, 1864, Ann. Mag. Nat. Hist., (3) 13, 147.
— Lacordaire, 1872, Gen Col., 9 (2), 767, 780.
—Bates, 1881, Biol. Centr. Amer. Col., 5, l60.

Beltista Thomson, 1860, Classif. Ceramb.,
16; 1864, Syst. Ceramb., 355.

Description— very robust, a little elon-
gate-oval in shape; pubescent; with erect hairs
scattered throughout. Head moderately con-
cave between the antennal tubercles, which are
not strongly raised; genae elongate; lower lobes
of the eyes small, acuminate interiorly. Anten-
nae from about one and a third times up to
about twice as long as the body in males, slightly
longer than the body to about one and a half
times as long in females; covered with long fine
hairs, most numerous below; the fourth seg-
ment bearing apically a large distinct tuft of
hairs, which often completely encircles the
apex; segments two, three, and sometimes five
and six with a pencil of hairs beneath the
apex; the scape not very swollen, moderately
elongate, only reaching to about the middle of
the pronotum but equal to or longer than the
third segment; the fourth segment longer than
the third, (Lacordaire states that the third is
longer than the fourth, but this is probably an
error in transcription), the fifth to eleventh
gradually decreasing in length. The pronotum
transverse, convex, bearing two discal obtuse
tubercles, narrowed basally and apically; the
lateral tubercle on each side rather strong, con-
ical, placed a little behind the middle. Scutellum
small, sub-triangular. Elytra not very elongate,
slightly less than twice as long as broad; only
moderately convex, sub-parallel or moderately
rounded laterally, declivous and narrowing pos-
teriorly; projecting beyond the pronotal base
a little anteriorly; obliquely truncate apically;

819S6

1

2

Zoologica: New York Zoological Society

[40: 1

Text-fig. 1. Distributional map of the species of Cosmotoma Blanchard.

with a centro-basal tumescence on each elytron,
bearing a fascicule of hairs. Legs of moderate
length; femora pedunculate then strongly ovu-
larly swollen apically; tarsi short, the first seg-
ment of the posterior tarsi scarcely as long or
slightly shorter than the following two united.
Prosternal protuberance rather narrow, curved
posteriorly; mesosternal protuberance broad,
triangular, recurved posteriorly. Apical ventrite
tranverse, sub-triangularly broadly rounded in
male; a little more elongate in the female, and
with a median longitudinal groove on the basal
half.

Genotype: Cosmotoma adjuncta Thomson

The genus Beltista Thomson was only named
by that author because he considered that the
name Cosmotoma was too similar to the name
Cosmisoma (Cerambycidae, Cerambycinae) . It
therefore is a synonym of Cosmotoma Blan-
chard.

The name Cosmotoma was first given by
Dejean in 1836, Col. Cat., ed. 3, 364, but as
neither of the two species, venustulum Dejean
and plumicorne Dupont, listed there are valid,
the name does not become valid until Blanchard
described it in 1845. Blanchard here gives the
species venustulum Dejean in the genus. This
name is however still invalid as no description
has been made of it. In 1860, Thomson des-
cribed the species adjuncta as genotype of his
genus Beltista. Chevrolat then stated (1861,
Journ. Ent., 1, 188) that adjuncta Thomson

was synonymous with venustulum Dejean. The
latter being, however, an invalid name, and
the name adjuncta Thomson the first available
valid one, the latter is consequently the type of
the genus Cosmotoma Blanchard. It appears
that sertifer Serville cannot be considered under
the International Code, Article 30, Section 11,
e (1).

It is interesting to note that Thomson, after
his description of adjuncta (1860, Classif.
Ceramb., 16) states: “Ma collection renferme 6
especes devant rentrer dans ce genre.” Strangely
enough, however, he never described any of
these and it is possible that some of the new
species which I describe herein may be similar
to some of those among the six mentioned.

Key to Species

1. Elytral apex simply obliquely truncate, the
marginal angle not spinously produced .2
Elytral apex with the marginal angle

spinous 6

2. Base of elytra black 3

Base of elytra not black, nor the suture
black on the anterior half 4

3. Pronotum black from the anterior trans-
verse groove; basal blaek elytral area with
a ferruginous prolongation to the humerus
giving the black area a bilobed appearance,
the antero-sutural black portion not join-
ing the posterior black area

suturalis sp. nov.

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

3

Pronotum only black on its posterior half;
basal black elytral area complete, without
a ferruginous prolongation to the humerus,
the suture completely black between an-
terior and posterior black areas

fasciata Fisher

4. The pronotum completely more or less

blackish in color

adjuncta nigricollis Bates
The pronotum not completely black in
color, at most basally and part laterally . 5

5. Pronotum completely light ferruginous in

color adjuncta rubella Bates

Pronotum light ferruginous basally, ex-
tending posteriorly in a triangle to a little
behind the middle, the rest blackish

adjuncta Thomson s. str.

6. Completely more or less unicolorus, with-
out markings 7

Not unicolorus, with light and dark bands
and markings 8

7. Completely blackish in color

nigra sp. nov.

Completely more or less yellowish-fer-
ruginous in color pallida sp. nov.

8. The post-median elytral transverse black

band almost straight anteriorly

melzeri sp. nov.
The post-median dark colored band not

straight, distinctly curved and may be

broken in part or lacking 9

9. Pronotum greenish-black or blackish in

color 10

Pronotum ferruginous in color 11

10. Post-median elytral transverse band dis-

tinct, somewhat irregularly, fairly strongly,
anteriorly curved, distinctly blackish pub-
escent and strongly contrasting from the
rest of the elytra; the two discal prontal
tubercles strongly separately raised; the
pronotum with distinct large pimctures in
the anterior transverse groove

viridana Lacordaire

Post-median elytral band not present as
such, not blackish, slightly posteriorly
curved, but only present as the basal der-
mal olive-green color separated from the
rest by grayish pubescence; the two discal
pronotal tubercles not at all separated
medially, but planely connected; no large
punctures present in the anterior pronotal
groove olivacea sp. nov.

11. Post-median el)^ral band black, narrow,
distinctly strongly and regularly curved,
broadly bordered posteriorly with ferrugi-
nous pubescence; the third antennal seg-

ment slightly shorter than the scape (3

unkown) triangularis sp. nov.

Post-median elytral band almost not pre-
sent, faint, ferruginous pubescent in part
at the suture; the third antennal segment
slightly longer than the scape (?$ un-
known) sertifer Serville

Cosmotoma adjuncta Thomson
Text-fig. 2

Thomson, 1860, Classif. Ceramb., 16.— Che-
vrolat, 1861, Joum. Ent., 1, 188. (nota synon.).
—Bates, 1864, Ann. Mag. Nat. Hist., (3) 13,
148. (nota).

Description. Ma/e.— Ferruginous in color ex-
cept for the posterior half of the pronotum,
pitchy, extending forward laterally from the
middle and narrowing to about the basal fifth at
the border, thus leaving a sub-triangular ferru-
ginous area medio-anteriorly; the elytra pitchy
behind the middle, except usually the extreme
apex, the anterior border of this dark area strong-
ly curved and well defined; the basal declivity
somewhat lighter ferruginous than the rest.
Marked with grayish-white pubescence on the
elytra as follows: the post-median pitchy area
bordered anteriorly with grayish-white, ramify-
ing forward suturally in a fairly regularly placed
lattice-work in which one white band extends to
the margin at about the basal third, one obliquely
forward to the humeri, and one obliquely for-
ward to the suture at about the basal quarter; a
distinct transverse white fascia at about the api-
cal sixth; the centro-basal elytral tumescence
crested with strong black setae. The antennal
segments from the third with grayish pubes-
cence basally, the setae, fasciculae and brushes
of hairs black.

The underside darker ferruginous than above,
sometimes pitchy-brown and the abdomen oc-
casionally almost black. Covered with grayish-
white pubescence, thin in parts and most dense
on the sides of sternum and laterally on the
posterior border of the first abdominal segment.

Antennae one and a half to one and three-
quarter times as long as the body; slender;
fringed below, after the fourth segment chiefly
at the apices; segments two and three with dis-
tinct pencils of black hairs beneath, the latter
at the apex; the fourth segment bearing on a
little more than its apical third a strong dense
brush of long hairs, which encircles the segment,
which is somewhat swollen on the area which
bears the setae; the scape moderately elongate,
reaching to about the middle of the pronotum,
a little swollen, but not strongly; the third seg-
ment about one and a third times as long as
the scape, the fourth segment almost one and a

4

Zoologica: New York Zoological Society

[40: 1

Text-fig. 2. Cosmotoma
adjuncta Thomson s. str.
$ (X 9.6).

half times as long as the third segment, about
twice as long as the scape, and about four times
as long as the fifth segment, the rest gradually
decreasing to the apex; the segments completely
finely and closely punctured, except where the
fasciculae and brushes of setae arise where the
punctures are very much larger. The antennal
tubercles only slightly raised, the head broadly
and slightly concave between; the frons large,
about equilateral, moderately convex, with a
very fine median longitudinal groove; the lower
lobes of the eyes small, narrowing inferiorly,
about one and a half times as long as broad,
about equal in length to the genae; the head
completely finely and closely punctured, bear-
ing a few long setae anteriorly and at the inner
frontal margin of the eyes; finely tawny pubes-
cent.

The pronotum transverse, convex, with two,
well raised, obtuse tumescences on the disc;
with a strong broadly conical, spinous swelling
laterally on each side slightly post-medially;
completely very finely and closely punctured,
with a single row of sparse, very large punctures
apically and one basally in the transverse
grooves, which are broad and rather shallow;
finely tawny pubescent anteriorly, grayish pub-

escent posteriorly on the dark area, most dense
on the posterior border of the lateral spines. The
scutellum sub-triangular to sub-rotundate, fairly
narrowly rounded apically; very finely and
closely punctured; finely grayish pubescent.

The elytra not very elongate, only moderately
convex, rounded laterally; the apices obliquely
truncate, the marginal angle not spinous; the
centro-basal crests strongly raised, black fas-
ciculate; completely finely and closely punc-
tured, the scattered black setae rising from
larger punctures.

The prostemal protuberance narrow, partic-
ularly between the coxae, somewhat broadly
concave medially; strongly curved. The mes-
osternal protuberance very broad anteriorly,
sub-triangular, truncate apically, the trimcature
being somewhat wider than the breadth of the
prostemal protuberance medially. The ventrites
of normal size; the apical segment transverse,
somewhat sub-triangular and broadly rounded
apically. The underside completely very finely
and variably closely punctured; finely grayish-
white pubescent, mostly rather sparse, most
dense on the sides of the metasternum and lat-
ero-posterior border of the first abdominal seg-
ment.

1955]

Gilniour: Revision of the Genus Cosmotoma Blanchard

5

The legs of moderate length; femora strongly
pedunculate; tarsi moderately slender, the &st
segment of the posterior about equal in length
to the following two segments united; aU the
legs very finely and fairly closely punctured;
sparsely and finely grayish-tawny pubescent.

Tema/e.— Similar in color to the male. The
antennae a little less elongate. The apical ven-
trite a little more elongate.

Length, 6.5-8 mm.; breadth, 2.5-3 mm.

Locality.— French Guiana— (Chevrolat)
(Mus. Hist. Nat. Belg.). Brazil— Para (Gilmour
coll.) (15); “Amazon” (Mus. Hist. Nat. Belg.) ;
Santarem (United States National Museum)
(1 5, 4 $); Manaos (Hincks-Dibb coll.) (15).
Peru— Achinamiza (1.XII.26, H. Bassler)
(Amer. Mus. Nat. Hist.) (1 2); Callanga
(Hincks-Dibb coll.) (1 5) . Colombia— Carta-
gena (1.1.21) (Amer. Mus. Nat. Hist.) (12).

Material Examined —Mms. Hist. Nat. Belg.,
2; Amer. Mus. Nat. Hist., 2; Gilmour collection,
1; United States National Mus. 5; Hincks-Dibb
collection, 2; Total: 12.

Cosmotoma adjuncta Thomson var. (?subsp.)

RUBELLA Bates

Text-fig. 3

Bates, 1864, Ann. Mag. Nat. Hist., (3) 13,
14; 1872, Trans. Ent. Soc. Lond., 237; 1881,
Biol. Centr. Amer. Col., 5, 160, pi. 13, fig. 8.

Cosmotoma rubella Bates has stood up to
the present as a good species but so far as I
can ascertain it is structurally quite similar to
Cosmotoma adjuncta Thomson and the distri-
bution areas of adjuncta Thomson and rubella
Bates overlap. Bates, however, states (1872,
Trans. Ent. Soc. Lond., 237) that the specimen
from Chontales, Nicaragua, is “rather darker in
colour of the elytra than specimens from the
Amazons” and in my opinion, in view of the
fact that there is such a wide gap between the
known distributional areas, this specimen might
be of a different species. The coloration of the
figure given by Bates is not particularly good.

Description— The, variety rubella Bates differs
most conspicuously from the typical form of C.
adjuncta Thomson in the pronotum being com-
pletely ferruginous-red in color, without any
trace of dark color anywhere. This would appear
to be quite constant. Also the posterior pitchy
area of the elytra seems to have its anterior
curved border continued suturally a little fur-
ther behind before meeting the suture.

Text-fig. 3. Cosmotoma ad-
juncta rubella Bates. 5 (^
11.0).

6

Zoologica: New York Zoological Society

[40: 1

Length, 5-7.5 mm.; breadth, ?-2.8 mm.

LocaZ/ry.— Brazil— Para (Bates); R. Tapajos
(Bates). French Guiana— (Bates); (GUmour
collection) (19). Nicaragua— Chontales (Belt)
(Bates).

Material Examined— GWmova collection, 1.

CosMOTOMA ADJUNCTA Thomson, var.
(?subsp.) NiGRicoLLis Bates
Text-fig. 4

Bates, 1864, Ann. Mag. Nat. Hist., (3) 13,
148.

Cosmotoma nigricollis Bates, which has pre-
viously been sunk as a synonym of Cosmotoma
adjuncta Thomson, has in my opinion as good
a claim to validity as C. rubella Bates. In the
same way that rubella Bates is a very distinct
form of Cosmotoma adjuncta Thomson, so I
consider nigricollis Bates to be equally distinct,
but being the darker colored form and not
lighter form as rubella Bates. Structurally again
I can find no distinguishing feature for nigricol-
lis Bates from C. adjuncta Thomson or rubella
Bates.

Description— Tht variety nigricollis Bates dif-
fers most conspicuously from the typical form

of C. adjuncta Thomson in the pronotum being
completely pitchy-black in color without any
ferruginous color at all. The posterior half of
the head is also dark colored. This coloration is
constant in all specimens examined. Also the pre-
apical transverse white pubescent band on the
elytra is often more extensive, though the pubes-
cence is thiimer, thus giving the appearance of a
rather narrower posterior elytral dark colored
area. This form is shown in the figure given
herein.

Length, 5.5-6.5 mm. (Bates’ specimens ap-
pear to have been larger, measuring 7.7-8.5
mm.); breadth: 2.2-2.5 mm.

Loca/jty.— Brazil— Upper Amazons, Ega
(Bates); Amazon (Gilmour coll.) (19), (R.
Mus. Nat. Hist. Belg.); (Am.m.!) (Zool.
Staatssamml., Mun.) {!$); Teffe (X1.24, H.
Bassler) (Amer. Mus. Nat. Hist.) (13).

Material Examined.— Amer. Mus. Nat Hist,
1; Mus. Hist Nat Belg., 1; Zool. Staatssamml.,
Mun., 1; Gilmour collection, 1; Total: 4.

Cosmotoma fasciata Fisher
Text-fig. 5

Fisher, 1931, Journ. Wash. Acad. Sci., 21
(2), 23.

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

1

Text-fig. 5. Cosmotoma
fasciata Fisher. $ (X
9.6). Paratype.

Description. Ma/e.— Ferruginous and black
in parts as follows; antennae ferruginous, the
setae black; the head ferruginous. Pronotum
with about the anterior half and posterior bor-
der (narrowly) ferruginous, the rest black;
scutellum black. The elytra with almost the basal
quarter black, the posterior border of this area
almost rectilinear, and about the apical two-fifths
black (the anterior border only slightly curved) ,
these two areas united by a common sutural
black band; the area between these ferruginous.
The legs ferruginous, a little darker basally; the
anterior femora pitchy postero-dorsally. The
ventral surface pitchy to black, except the head
and anterior border of the pronotum.

The head and anterior half of the pronotum
with sparse tawny pubescence, this becoming
denser towards the sides of the pronotum and
forming a vague macula dorso-laterally on each
side. Marked with thin white or grayish-white
pubescence as follows: the pronotum medially
longitudinally and behind the lateral tubercles
on the posterior half; the lateral borders of the
scutellum; the inner side of the humeri; three
somewhat curved irregular transverse fasciae on

the ferruginous elytral area which unite towards
the suture; one or two vague transverse areas
on the apical black area, and a very distinct
densely white pubescent oblique latero-discal
macula at about the apical fifth, and an almost
as distinct, though less clearly defined, common
white sutural macula at about the apical quarter.
The underside more or less completely very
sparsely grayish-white pubescent, which be-
comes densely white at the latero-posterior
angles and posterior border of the sternum and
at the latero-posterior borders of the first and
apical abdominal segments.

Antennae about one and three-quarter times
as long as the body; slender; fringed beneath on
the basal segments; the fourth segment bearing
round its apical half a large brush of black hairs;
the second, third and sixth segments with distinct
apical pencils of black hairs, (Fisher states the
fifth segment, not sixth, but this I think is an
error). The scape elongate, not very swollen,
extending to about the apical third of the pro-
notum; the third segment about one and a third
times as long as the scape, the fourth segment
not quite one and a sixth times as long as the

8

Zoologica: New York Zoological Society

[40: 1

third, about one and two-third times as long as
the scape, and about two and a third times as
long as the fifth segment, the rest gradually de-
creasing to the apex; the fourth segment dis-
tinctly swollen on its apical half; the antennal
segments completely and finely punctured, with
larger punctures from which the setae arise, par-
ticularly on the fourth segment.

The antennal tubercles slightly, but not
strongly, raised, slightly broadly concave be-
tween; the frons about as long as broad, slightly
convex, with a fine median longitudinal groove;
the lower lobes of the eyes small, narrowing in-
teriorly, slightly longer than broad, about equal
in length to the genae; the head completely very
finely and closely punctured, with a few long
erect setae scattered marginally on the frons,
particularly at the anterior border.

The pronotum slightly broader than long, con-
vex, with two well raised obtuse tumescences
on the disc; with a slight, though broadly con-
ical, spinous swelling laterally on each side,
slightly behind the middle; completely very finely
and closely punctured, with a few large coarse
punctures in the anterior and posterior trans-
verse grooves, which latter are broad and only
moderately deep. The scutellum a little longer
than broad, sub-triangular, very broadly rounded,
almost truncate, apically; very finely and closely
punctured.

The elytra moderately elongate, moderately
convex; almost parallel-sided laterally to about
the apical third, then broadly rounded to the
apices, which are obliquely internally truncate,
the marginal angle not spinous; the centro-basal
tubercle on each elytron strongly raised, black
fasciculate above; completely finely and closely
punctured, with a number of larger, coarse,
punctures scattered here and there, from which
erect black setae arise; a distinct sub-sutural line
on each elytron, which is much less distinct pre-
medially.

The prosternal protuberance moderately nar-
row, more so medially, longitudinally concave
medially, strongly curved. The mesosternal pro-
tuberance subtriangular, very slightly tumescent
medially, strongly curved, truncate apically, the
truncature being only very slightly broader than
the breadth of the prosternal protuberance medi-
ally. The apical ventrite very broadly some-
what subtriangular in shape, the apex broadly
rounded. The whole underside completely finely
and closely punctured; the apex of the apical
ventrite with a few large coarse punctures.

The legs of moderate length; the femora
strongly pedunculate; the tarsi fairly slender,
the first segment of the posterior about equal in
length to the following two united; all very finely
and fairly closely punctured.

Fe/naZe.— Apparently unknown. (I have only
seen one male paratype and from Fisher’s orig-
inal description it would appear that all the spec-
imens are males) .

Length, 5-7.5 mm.; breath: 2-3.2 mm.

Locality.— Costa. Rica— Reventazon River,
Hamburg Farm, (4.11.25, F. Nevermann)
(Type Locality). Panama— (15.IV.37), In
Banana debris (W. J. Fisher in Hit.).

Type, two Paratypes and one other specimen
(Panama) in the United States National Mu-
seum. (The F. Nevermann collection was pur-
chased by this institution on Nevermann’s
death). One Paratype (5) in the Gilmour col-
lection. (Exchanged with the United States
National Museum for a paratype of Cosmo-
toma pallida sp. nov.).

This species is most closely allied to Cosmo-
toma suturalis sp. nov. described in this paper,
the differences from it being given with that
species. From Cosmotoma adjuncta Thomson
and the other species it differs at first glance
in both base and apex of the elytra being black.

Cosmotoma suturalis sp. nov.

Text-fig. 6

Description.— Male. Ferruginous and black in
parts as follows: antennae ferruginous, the setae
black; the head ferruginous, a little darker pos-
teriorly; the genae black. Pronotum with the
anterior border ferruginous, in front of the an-
terior transverse groove, and the basal margin
ferruginous, not extending to the posterior trans-
verse groove medially, the rest black; scutellum
pitchy-ferruginous, not quite black. The elytra
black basally, almost up to the basal quarter, and
extending in a first broadening, then narrowing
band along the suture to about the apical two-
fifths; the elytra then ferruginous up to about
the apical two-fifths, with a ferruginous curved
projection anteriorly to the base running on the
inner side of the humeri; about the apical two-
fifths black, the anterior border distinctly curved.
The legs ferruginous, blackish basally and ven-
trally on the anterior femora, and ventro-basally
on the intermediate and posterior femora.

Marked with white or grayish-white pube-
scence as follows: the lateral margins of the
frons and the posterior border of the head; the
anterior border of the pronotum thinly yellow-
ish-white; the postero-superior border of the lat-
eral tubercles; a triangular area on the pronotal
disc, between the tumescences, apex towards
the base; the elytra with whitish pubescent bands,
on the ferruginous area, running from anterior
border, middle and posterior border marginally
to unite on the disc before the sutural black
area and continued forward to the base along
the inner side of the humeri; a distinct trans-

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

9

verse white fascia at about the apical fifth; a
vague oblique patch of tawny pubescence on the
black area in front of this.

Underside completely black, the apical ven-
trite slightly lighter in color— pitchy-ferruginous
apically; covered thinly with grayish pubescence,
which becomes dense at the latero-posterior
angle of the sternum and at the side of the first
and apical abdominal segments.

Antennae slightly more than one and a half
times as long as the body; slender, fringed be-
neath on the basal segments, most densely on
the second and apex of the third segments, be-
coming sparse apically from the fifth segment;
the fourth segment bearing around its apical
half a large brush of black hairs. The scape mod-
erately elongate, a little swollen, reaching to
about the basal third of the pronotum; the third
segment about one and a quarter times as long
as the scape, the fourth segment about one and
a third times as long as the third, about one and
three-quarter times as long as the scape, and
nearly three times as long as the fifth segment,
the rest gradually decreasing to the apex; the
fourth segment distinctly swollen towards the
apex where the brush of setae rises; the fifth
segment rather distinctly curved; the antennal
segments completely and finely pimctured, with

larger punctures from which the setae arise,
particularly those on the fourth segment.

The antennal tubercles slightly raised, but not
strongly, moderately strongly broadly concave
between; the frons very slightly longer than
broad, almost equilateral, slightly convex, with
a fine median longitudinal groove; the lower
lobes of the eyes small, a little narrowing in-
feriorly, slightly longer than broad, about three-
quarters as long as the genae; the head com-
pletely very finely and closely punctured, with
a few long setae scattered marginally on the
frons; sparsely grayish-white pubescent in the
main.

The pronotum transverse, convex, with two
well raised, obtuse tumescences on the disc; with
a strong, broadly conical, spinous swelling later-
ally on each side slightly post-medially; com-
pletely finely and closely punctured, with pos-
terior transverse grooves, which latter are broad
and not very deep. The scutellum slightly longer
than broad, sub-triangular, broadly rounded
apically; very finely and closely punctured;
blackish pubescent, with the lateral borders
finely grayish pubescent.

The elytra not very elongate, only moderately
convex; almost straight-sided laterally to about
the apical third, then broadly rounded to the

10

Zoologica: New York Zoological Society

[40: 1

apices, which are obliquely truncate, the mar-
ginal angle not spinous; the centro-basal tumes-
cences strongly raised, very sparsely black se-
tose; (I believe that in the specimen examined
the fasciculae have been knocked off to some
extent. The setae, composing the normal fas-
ciculae in the species of this genus, appear to
be loosely attached and are easily dislodged);
completely finely and closely punctured, the
scattered black setae arising from slightly larger
punctures; a fairly distinct longitudinal sutural
Carina on the apical half.

The prosternal protuberance narrow, par-
ticularly medially, strongly concave, strongly
curved. The mesosternal protuberance very
broad, particularly anteriorly, sub-triangular,
slightly tumescent medially, truncate apically,
the truncature being almost twice as broad as
the breadth of the prosternal protuberance me-
dially. The apical ventrite transverse, broadly
slightly sub-triangular in shape, the apex broadly
rounded. The whole underside completely finely
and closely punctured; the apex of the apical
ventrite with a few moderately large punctures.

The legs of moderate length; the femora
strongly pedunculate; the tarsi moderately
slender, the first segment of the posterior about
equal in length to the following two imited; very
finely and closely punctured.

Female. Similarly colored to the male. More
robust. The antennae about one and a half times
as long as the body.

The apical ventrite transverse, but slightly
sub-conically triangular in shape, with a fine
median anterior groove, the apex broadly bi-
sinuately truncate; the apex with a number of
very large, coarse, close punctures.

Length, 6.5-8 mm.; breadth, 2.5-S.2 mm.

Loca/ify-— Brazil— Manaos (3) (Holotype);
“Amazonas” ($) (Allotype). Peru— Gancar-
tambo (3) (Paratype).

Holotype, 3, in the Hincks-Dibb collection.
Allotype, $, in the Musee Royale d’Histoire
Naturelle de Belgique. (Coll. Achard) (Coll.
Le Moult Box M.669) (R. Mus. Hist. Nat.
Belg. I.G. 12, 595). Paratype, 3, in the Dr.
P. Lepesme collection, Paris.

Diagnosis.— This beautiful new species is most
closely allied to Cosmotoma fasciata Fisher
from Costa Rica, from which it differs in the
sutural black band on the anterior half not
reaching the apical black area; with a prolonga-
tion of ferruginous color to the humerus; the
pronotum being almost completely black, except
anterior to the transverse groove, and other
differences. From C. adjuncta Thomson and all
the other known species in the genus it differs
conspicuously in markings, through the sutural
black band, pronotal dark and ferruginous areas,
etc.

The paratype is in much worse condition and
of less bright coloration than the holotype and
allotype, probably due to age and dust.

Cosmotoma sertifer Serville
Text-fig. 7

Serville, 1835, Ann. Soc. Ent. France, 4, 59.
— Lacordaire, 1872, Gen. Col., 9 (2), 654, nota
2.— Aurivillius, 1923, Col. Cat. Ed. Junk-
Schenkling, 74, 419.— Melzer, 1927, Rev. Mus.
Paulista, 15, 575 (nota synon.).— Linsley, 1933,
Pan-Pacific Ent., 9 (3), 132. (Synonymy).

Plavilstshikov, 1927, Encycl. Ent.
B.l.-CoL, 2 (2), 59.

setifer Aurivillius, 1923, Col. Cat. Ed. Junk-
Schenkling, 74, 334. {Pogonocherus) .

I do not agree with any of the other authors,
viz., Aurivilhus (1923), Melzer (1927), Lins-
ley (1933), or Blackwelder (1946, BuU. U. S.
Nat. Mus., 185, 617), that Cosmotoma viridana
Lacord. is synonymous with Cosmotoma sertifer
Serville. I regard them as two distinct species.

C. sertifer Serville has not been described or
even mentioned as a collected species from the
time it was described and although ServiUe’s
original description lacks many important points
in view of modern knowledge, there are several
points of note which could not have been missed
out of any description, however early, particu-
larly as regards coloration. Most important and
immediately noticeable is the lack of mention
of any post-median black transverse band. I
have seen the types of viridana Lacord., and this
band is distinctly present. I have further seen a
single specimen which agrees with the descrip-
tion of sertifer Serville and it is quite distinct
from viridana Lacordaire.

I give a translation of Serville’s original de-
scription below for comparison, as well as a
full description of the specimen examined.

Serville placed his species sertifer in his 1st
Division of the genus Pogonocherus, separated
by the following characters:

“Elytra truncate at the apex; the external
angle of the truncature unispinous (p. 57).

“(Length, 2 to 3 lines [i.e., ca. 4.2-6 mm.]).

“Body a little shining, blackish and covered
with long brown hairs, sparse, above; ferrugi-
nous below. Pronotum swollen at the middle,
bordered posteriorly. Elytra slightly bordered
exteriorly and at the suture, each having at the
base a feeble tubercle bearing some long stiff
and brown hairs; they are tinted with greenish
towards their apex. The antennae brown, having
a tuft of black hairs on their fifth segment, the
first greenish. The legs of ferruginous brown,
with some hairs like the body; femora greenish.

“From Brazil. My collection.” (Translation
of original description) .

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

11

There is a very obvious discrepancy in this
description where it states that the fifth segments
of the antennae bear a tuft of black hairs. This
I believe is almost certainly a mistake. If it is
not an error on the part of author or printer,
then the species of Serville does not belong to
this genus, and the specimen which I describe
below will need a new name in being a new
species.

Neither does Serville state the extent of the
brownish pubescence or greenish color of the
elytra. I take it that the head and pronotum are
brown pubescent and the elytra in part, gradu-
ally becoming green towards the apex. In my
specimen the elytra are almost wholly greenish,
but I presume that this is perhaps a rather
variable character.

Af a/e.— Ferruginous and greenish as follows:
head and pronotum closely ferruginous pubes-
cent, the extreme apex of the scape and its
second segment completely ferruginous, the tarsi
in the main ferruginous, the underside com-
pletely ferruginous, the sternum however pitchy-
red; the antennae in the main, and the femora
and tibiae greenish, the elytra greenish in the
main except baso-suturally and narrowly along
the suture, and a vague, ferruginous, irregular,
transverse band at about the apical third, the
elytra also with vague areas of silvery-gray
pubescence, chiefly on the inner side of the

humeri extending posteriorly to about the basal
third, thence branching to margin and suture,
running narrowly along the latter and extending
to the margin again just before the apical third,
and apically. The underside with thin grayish
pubescent, a little more dense on the sternum
and sides of the anterior abdominal segments.

Antennae about one and two-thirds times as
long as the body; slender, fringed beneath, be-
coming sparse after the fomth segment, where
they are chiefly confined to the apices; segments
two, three and five with thin pencils of hairs
beneath the apex, segment four bearing a strong
brush of dense black setae on the inner apical
two-fifths, not extending completely round the
segment; which is somewhat swollen on the
area from which the setae arise. The scape
moderately elongate, reaching to about the
apical quarter of the pronotum, a little swollen;
the third segment about one and a fifth times
as long as the scape, the fourth segment about
one and a half times as long as the third, about
one and three-quarter times as long as the scape,
and twice as long as the fifth segment, the rest
gradually decreasing to the apex; the segments
completely finely and fairly closely punctured,
with somewhat larger punctures from which the
long setae arise. The antennal tubercles only
slightly raised, the head broadly and only very
slightly concave between; the frons large, more

12

Zoologica: New York Zoological Society

[40: 1

or less equilateral, moderately convex, with a
very fine median longitudinal groove; the lower
lobes of the eyes very small, almost sub-
quadrate, only slightly narrowing interiorly,
about three-quarters as long as the genae; the
head completely finely and closely punctured,
bearing a number of distinct tawny long setae
at the inner border of the lower lobes of the
eyes and lower border of the frons; completely
finely and fairly closely ferruginous pubescent.

The pronotum transverse, moderately convex,
with two strongly raised conical discal tume-
scences; with a moderately strong, broadly coni-
cal spinous tubercle on each side slightly behind
the middle; competely finely and closely punc-
tured with a number of fairly close, extremely
large, scattered punctures in the anterior and
posterior transverse grooves, which are broad
and very shallow, except a little deeper medially;
completely uniformly ferruginous pubescent,
with a slight silky sheen in certain lights. The
scutellum sub-triangular, extremely broadly
rounded apically, almost truncate; very finely
and closely punctured; dark ferruginous pubes-
cent, lighter marginally.

The elytra not very elongate, only moderately
convex, slightly rounded laterally, the apices
obliquely truncate, the marginal angle distinctly
spinously produced; the centro-basal crests
strongly raised, black fasciculate; the elytra com-
pletely finely and closed punctured, with an ir-
regular longitudinal band of extremely large
punctures running from the inner side of the
humerus along the outer side of the centro-basal
tubercle almost to the middle of each elytron,
a few slightly smaller and more sparse punc-
tures on the sutural side of the centro-basal tu-
bercle and a number of still slightly smaller punc-
tures, more scattered, on the lateral and apical
half of the elytra, from some of which the long,
erect setae arise.

The prosternal protuberance narrow, particu-
larly between the coxae, where it is sub-parallel
in the main, strongly longitudinally concave,
strongly curved. The mesosternal protuberance
extremely broad anteriorly, sub-triangular, trun-
cate and very slightly emarginate apically, the
truncature only a little wider than the breadth
of the prosternal protuberance medially. The
apical ventrite strongly transverse, a little sub-
triangular, the apex extremely broadly rounded
and very slightly emarginate apically. The un-
derside completely very finely and, in the main,
closely punctured; finely grayish pubescent,
sparse in the main, but a little more dense on
the sternum and sides of the anterior abdominal
segments.

The legs of moderate length; femora very
strongly pedunculate; tarsi moderately slender.

the first segment of the posterior tarsi about
equal in length to the following two united; all
the legs very finely and moderately closely
punctured; sparsely grayish pubescent.

Fenza/e.— Apparently unknown. (As C. viri-
dana has up to the present been synonymous
with this species, it is possible that specimens
of the two species are mixed together in collec-
tions. Serville does not state the sex of his speci-
mens) .

Length, 6 mm.; breadth, 2.2 mm.

Loca/ztj.— Brazil— (Serville) ; Rio de Janiero
(Gilmour coll.) (1 $).

Material Examined— GWmouv collection, 1 .

This is one of the smaller species of the genus.

The specimens listed by Bosq (1944, Rev. Soc.
Ent. Argent., 12, 201) from the Argentine are
C. viridana Lacordaire, and not C. sertifer Ser-
ville. Dr. Bosq has sent me for examination a
typical specimen which is certainly viridana
Lacordaire.

COSMOTOMA TRIANGULARIS Sp. nOV.

Text-fig. 8

Description— Female. Dark-brown to ferru-
ginous and olive-green as follows : head and pro-
notum closely ferruginous pubescent; the elytra
in the main olivaceous in color, except for some
variegated sUky gray pubescence around the
humeri, extending to behind the centro-basal
tubercles, then brokenly running to the margin
at about the middle, a little suturally, and a lit-
tle bordering the post-median dark band an-
teriorly and posteriorly, with the scutellum dark
brown and the area from the base of the scutel-
lum to the apex of the centro-basal tubercles
thence extending to the suture at about the basal
quarter, dark brown, with a little ferruginous
pubescence, thus giving a triangular mark on
each elytron, its base the suture, its apex at the
centro-basal tubercle; also on each elytron, at
about the apical two-fifths, a very distinct trans-
verse, narrow, curved, dark-brown band, this
shading off posteriorly to lighter brown, with
tawny pubescence, to about the apical fifth to
sixth; the antennae greenish in color, with the
extreme apices of each segment dark ferrugi-
nous; the legs dark olivaceous, the tarsi dark
brown; the underside pitchy-black, with the
apex of the last ventrite ferruginous, covered
with thin grayish pubescence, which is most
dense on the sternum and the sides and posterior
borders of first and second abdominal segments.

Comparatively (in this genus) robust, rather
broadly elongate-ovate. Antennae only a little
longer than the body (about one-ninth) , (unfor-
tunately lacking after the third segment in the
Paratype), moderately slender, fringed beneath
up to the sixth segment, with thin pencils of

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

13

setae at the apices of the third and fifth to sev-
enth segments, the fourth segment bearing a
large distinct brush of black setae on almost its
apical half, which does not completely encircle
the segment, but only about the longitudinal
minor half, the area from which the setae of the
brush arise being somewhat swollen. The scape
moderately elongate, extending to about the
basal fifth of the pronotum, not very swollen;
the third segment about a fifth shorter than the
scape, the fourth segment almost one and a half
times as long as the third, only about one and
a sixth times as long as the scape, and twice as
long as the fifth segment, the rest gradually de-
creasing to the apex; the segments completely
finely and fairly closely punctured, with distinctly
larger punctures from which the setae arise. The
antennal tubercles only slightly raised, the head
very broadly and slightly concave between; the
frons large, about equilateral, only moderately
convex, with a very fine median longitudinal
groove; the lower lobes of the eyes very small,
slightly narrowing interiorly, about as long as
broad, about two-thirds as long as the genae; the
head completely finely and closely punctured,
bearing a few long setae at its anterior border
and at the inner margin of the lower lobes of
the eyes.

The pronotum transverse, moderately convex,
bearing two strongly raised conical obtuse
tumescences on the disc; with an only moder-

ately strong, conical, narrowly obtuse, tubercle
laterally on each side slightly behind the middle;
completely finely and closely punctured, with a
number of very large punctures in a somewhat
irregular more or less single row in both anterior
and posterior transverse grooves, and extending
more sparsely between the discal tubercles, and
a few on the inner side of the base of the lateral
tubercles; the anterior and posterior transverse
grooves broad, and very shallow. The scutellum
sub-triangular, very broadly rounded apically;
very finely and closely punctured.

The elytra not very elongate, only moderately
convex, more or less parallel-sided laterally to
behind the middle thence broadly rounded to
the apices, which are obliquely truncate, with
the marginal angle rather stoutly and strongly
spinously produced; the centro-basal tumes-
cences moderately strongly raised and black fas-
ciculate; the elytra finely and closely punctured,
with a number of variably close, scattered, slight-
ly larger punctures here and there and some
much larger sparse punctures from which the
long, erect setae arise.

The prosternal protuberance narrow, particu-
larly between the coxae, where it is sub-parallel,
rather strongly longitudinally concave, strongly
curved. The mesosternal protuberance extremely
broad and rather swollen anteriorly, sub-triangu-
lar, the apex truncate and very slightly emar-
ginate apically, the truncature about equal in

14

Zoologica: New York Zoological Society

[40: 1

width to the breadth of the prosternal protuber-
ance medially. The ventrites of normal size; the
apical segment transverse, broadly rounded lat-
erally, the apex broadly emarginate, the lateral
angles rounded, bearing a median longitudinal
groove on the anterior half. (In the Holotype
this is rather obtuse and ill-defined; in the Para-
type it is much more distinctly marked) . The
underside completely finely, and in general,
closely punctured, with a number of large, rather
close punctures at the apex of the apical ventrite
from which distinct setae arise.

The legs of moderate length; the femora
strongly pedunculate; the tarsi moderately slen-
der, the &st segment of the posterior tarsi about
equal in length to the following two united; all
the legs very finely and fairly closely punctured;
sparsely grayish pubescent.

Length, 10-10.25 mm.; breadth, 3.75-4 mm.

LocaZ/ty.— Brazil— Rio de Janiero (Holotype) ;
(Paratype).

Holotype, $, and Paratype, $, in my collection.

Diagnosis— This new species is easily the larg-
est and most robust so far known in the genus.
It appears to be most closely allied to Cosmo-
toma sertifer Serville (of which I feel quite cer-
tain it is not the female), but differs conspicu-
ously by its large size, possessing a distinct post-
median dark colored transverse elytral band, in
the underside being black with the extreme apex
ferruginous and the third antennal segment
shorter than the scape, etc.

CosMOTOMA vnuDANA Lacordaire
Text-fig. 9

Lacordaire, 1872, Gen. Col., 9 (2), 781, nota
1, p. 108, fig. 4 (non 3).

sertifer Bosq, 1944, Rev. Soc. Ent. Argent.,

12, 201.

This species is certainly, in my opinion, quite
distinct from Cosmotoma sertifer Serville with
which it has been consistently synonymised since
its description, with the exception of Gemmiger
& Harold (1873, Cat. Col., 10, 3153) who,
however, did not have C. sertifer Serville in the
genus Cosmotoma, but retained it in Pogono-
chaerus (sic!) (l.c 3117). Apart from that it has
been always synonymised by other authors who
considered the two to be identical. I further be-
lieve that Argentinian specimens listed as serti-
fer Serville are almost certainly viridana Lacor-
daire.

I am very grateful to the authorities of the
Musee Royale d’Histoire Naturelle de Belgique
for so kindly sending me the Type specimen of
Lacordaire, which is in excellent condition.

In view of the fact that this species and Cos-
motoma sertifer Serville have been for so many
years confounded, and that I have given a trans-
lation of Serville’s original description, I give
also, below, a translation of Lacordaire’s origi-
nal description.

“Gray-green, dark beneath, above silky and
dark mixed, tarsi yellowish; 3rd antennal seg-

Text-fig. 9. Cosmotoma viridana
Lacordaire. $ (X 8.0). Type.

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

15

ment not penicillate; disc of prothorax strongly
binodose; elytral apices spinose, with a single
obtuse costa running down from the middle to
the apex and provided with a black fasciculate
basal crest, a common fascia behind the middle
dark colored. Length, 8 mm. Habit. Brazil
(Santa-Catherina Island). A very close species,
perhaps the same, exist in some collections under
the name of pulchellum Chevrol.” (Translation
of original description) .

The Type of Lacordaire’s description is a
female and the figure given is of this specimen.

Description. Ma/e.— Dark pitchy-brown and
greenish intermixed; the head and pronotum
pitchy; the elytra dark brown baso-suturally,
covering the centro-basal crests and with a very
distinct somewhat undulating transverse black-
ish-brown fascia at about the apical third; the
pronotum with grayish pubescence postero-later-
ally from the apex of the lateral spines to the
base; the elytra with variegated silky grayish
pubescence with the green dermal color inter-
mixed, chiefly grayish from the humeri almost to
the middle, except on the baso-sutural dark area,
and then apically behind the transverse fascia,
medially between more broken and much less
regular and extensive. The underside pitchy-
black to dark ferruginous; covered with thin
grayish pubescence which is most dense on the
sternum and sides of the first abdominal segment.

Antennae about one and a third times as
long as the body; slender; fringed below, most
densely at the apices, and chiefly at the apices
on the fifth to ninth segments; segment two
with a pencil of setae beneath, segments three
and five to seven with apical pencils of setae
beneath, becoming much more sparse towards
the latter; the fourth segment bearing, on about
the apical two-fifths, a strong dense brush of
long setae which is on the inner longitudinal
half of the segment, not completely encircling
the apical portion, which is swollen on the area
from which the setae arise; the scape moderately
elongate, extending to about the middle of the
pronotum, a little swollen; the third segment
about one and a half times as long as the scape,
the fourth segment about one and a half times
as long as the scape, not quite one and a third
times as long as the third segment, and about
two and a quarter times as long as the fifth seg-
ment, the rest gradually decreasing to the apex;
the segments completely finely and closely punc-
tured, except where the setae arise from rather
larger punctures, particularly on the fourth seg-
ment where they are very large. The antennal
tubercles slightly raised, moderately distinctly;
the head moderately and broadly concave be-
tween. The frons large, about equilateral, mod-
erately convex, with an extremely fine median

longitudinal groove; the lower lobes of the eyes
small, narrowing inferiorly, slightly longer than
broad, only about two-thirds as long as the
genae; the head completely fairly finely and
closely punctured, bearing a few long setae an-
teriorly and at the inner border of the lower
lobes of the eyes; finely grayish-tawny pubescent.

The pronotum transverse, convex, with two
strongly raised obtuse discal tumescences; with
a broadly conical, strong, lateral, almost spinous
protuberance on each side slightly post-medially;
completely very finely and closely punctured,
with a number of irregularly scattered very large
punctures in the anterior and posterior trans-
verse grooves, and a few extending, from the
anterior groove, posteriorly between the discal
tumescences; the transverse grooves very broad
and very shallow; dark tawny-brown pubescent
in the main, with slight areas of grayish pubes-
cence latero-anteriorly on the posterior border
and uniting there, and round the base of the
lateral spines, extending to the posterior border.
The scutellum sub-triangular, the apex very
broadly rounded; very finely and closely punc-
tured; blackish pubescent, narrowly margined
with grayish-brown.

The elytra only moderately elongate, not very
convex, narrowing a little, but almost straight-
sided laterally to about the apical quarter, thence
broadly rounded to the apices, which are
obliquely truncate, with the marginal angle pro-
duced into a strong spine; the centro-basal
tumescences moderately strongly raised, strongly
black fasciculate; completely finely and closely
punctured, with fairly numerous very large
punctures scattered here and there, from some
of which arise the long setae.

The prosternal protuberance narrow, particu-
larly between the coxae, broadly longitudinally
concave medially; strongly curved. The meso-
sternal protuberance very broad, particularly
anteriorly, sub-triangular, truncate apically, the
apex slightly emarginate, the truncature about
one and a halftimes as wide as the breadth of the
prosternal protuberance medially. The abdom-
inal segments normal; the apical ventrite trans-
verse, more or less broadly rounded, the apex
slightly emarginate medially. The underside
completely very finely and fairly closely punc-
tured, with a few slightly larger punctures roimd
the apex of the apical segment.

The legs of moderate length; the femora
strongly pedunculate; the tarsi slender, the first
segment of the posterior about equal in length
to the following two united. All the legs very
finely and closely punctured; rather sparsely
grayish pubescent.

Fema/e.— Similar in color to the male. The
antennae slightly shorter, about one and two-

16

Zoologica: New York Zoological Society

[40: 1

fifths times as long as the body. The apical ven-
trite more elongate, sub-triangular, the apex
truncate and fringed with strong, not very long,
hairs.

Length, 6-8.5 mm.; breadth, 2. 1-3.2 mm.
Loca//0’.— Brazil— Santa Catherina (ex. Lacor-
daire coll.) (Mus. Hist. Nat. Belg.) (Type, 9);
Santa Catherina, Corupa (Hansa Humboldt)
(X1.44, A. Mailer) (Amer. Mus. Nat. Hist.)
(29), (19 in Gilmour coll, by exch.) id. loc.
(X.45, A. Mailer) (Amer. Mus. Nat. Hist.)
(13); (ex. Candeze coll.) (Mus. Hist. Nat.
Belg.) (13); Rio Grande do sul (Hincks-Dibb
coll.) (19); Nova Teutonia, (27°11'S, 52°23'W)
(Fritz Plaumann) (J. M. Bosq coll.) (19);
(Amer. Mus. Nat. Hist.) (X-XII,41) (73,89);
Nova Teutonia, Corupa (Hansa Humboldt) (J.
M. Bosq. in litt.); Est. Sao Paulo (J. M. Bosq
in litt.); (Amer. Mus. Nat. Hist.) (X, XII, 44)
(13, 19). Argentina— Misiones (Alta Parana)
(J. M. Bosq in litt).

Material Examined.— Mus. Roy. Hist. Nat.
Belg., 2; Amer. Mus. Nat. Hist., 20; Hincks-Dibb
collection, 1; J. M. Bosq collection, 1; Gilmour
collection, 1. Total 25.

COSMOTOMA OLIVACEA Sp. nOV.
Text-fig. 10

Description. Female.— Dark brown to ferru-
ginous, with olive-green intermixed, particu-
larly on the elytra. The head and pronotum dark

brown, the latter becoming ferruginous on the
anterior and posterior borders. The scutellum
ferruginous. The elytra ferruginous basally be-
tween the humeri and suture and extending a
little suturally, the rest of the elytra laterally and
behind the centro-basal tumescences becoming
dark ferruginous. The head covered with thin
grayish pubescence; the pronotum covered in
the main with dark olivaceous pubescence, va-
riegated with grayish pubescence medially, ex-
tending narrowly posteriorly to the base, a small
amount antero-laterally and postero-laterally on
each side of the disc, and also grayish on the lat-
eral tubercles. The scutellum brownish pubes-
cent, narrowly margined with gray. The elytra
thinly ferruginous pubescent basally; for the rest
in general dark olivaceous pubescent, broken in-
to fine transverse bands by grayish pubescence
as follows: a band from beneath the humeri to
the suture just behind the centro-basal tumes-
cences, a broad one from the basal third and
one from immediately behind the middle which
unite at about the middle of the disc, but do not
reach the suture, but end at the suturo-discal
obtuse carinae, one at about the apical third,
which runs slightly forward to the suture, and
finally a very narrow, less distinct, dark oliva-
ceous band at about the apical eighth which
reaches the carina, turns anteriorly on the inner
side of it and ceases just short of the preceding
transverse band; the rest silky variegated gray

Text-fig. 10. Cosmotoma olivacea sp. nov.
9 (X 9.8). Holotype.

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

17

pubescent. The underside light ferruginous, cov-
ered thinly with grayish pubescence, which is
most dense on the sternum and latero-poste-
riorly on the first and second abdominal seg-
ments. The antennal scape greenish; the rest of
the segments light ferruginous; very thinly gray-
ish pubescent. The femora light ferruginous
basally, the swelling green; the tibiae ferrugin-
ous-green; the tarsi ferruginous; all the legs
covered sparsely with grayish pubescence.

Ovate-elongate; not very robust. The anten-
nae about one and a half times as long as the
body; sparsely fringed below up to the fifth seg-
ment; segment two with a thin pencil of setae
below, and segments three and five to eight with
thin pencils of black setae below at their apices;
the fourth segment bearing a large distinct brush
of long black setae on about its apical two-fifths,
which does not quite surround that apical portion,
but leaves about the outer longitudinal quarter
naked, the area covered by the setae somewhat
swollen; the scape moderately elongate, extend-
ing to about the basal third of the pronotum,
moderately swollen; the third segment about one
and a seventh times as long as the scape, the
fourth segment not quite one and two-thirds
times as long as the scape, about one and two-
thirds times as long as the third segment, and
almost two and a half times as long as the fifth
segment, the rest gradually decreasing to the
apex; the segments completely finely and closely
punctured, except where setae arise from rather
larger punctures, particularly on the fourth seg-
ment where they are much larger. The antennal
tubercles slightly raised, the head slightly and
broadly concave between. The frons large, about
equilateral, moderately strongly convex, with a
very fine median longitudinal groove; the lower
lobes of the eyes small, narrowing a little infe-
riorly, about as long as broad, about two-thirds
as long as the genae; the head completely finely
and closely punctured, bearing a few long setae
anteriorly and at the inner border of the lower
lobes of the eyes.

The pronotum transverse, very strongly con-
vex discally, the two discal tumescences not
separated from one another medially, but form-
ing a large transverse tumescence; with a mod-
erately broadly conical spinous swelling laterally
on each side, slightly behind the middle; com-
pletely finely and fairly closely punctured, with
a somewhat irregular, more or less single, row
of very large punctures only on the posterior
transverse groove, which is broad and only very
slightly broadly concave, the anterior trans-
verse groove not very broad and scarcely at all
concave. The scutellum sub-triangular, ex-
tremely broadly rounded, almost truncate api-

cally; a little longitudinally concave towards the
apex; very finely and closely punctured.

The elytra only moderately elongate, not very
strongly convex, almost parallel-sided to about
the apical third, thence broadly rounded to the
apices, which are obliquely truncate, with the
marginal angle produced into a strong spine;
the centro-basal tumescences moderately
strongly raised, black fasciculate; each elytron
with a distinct (because glabrous — probably
rubbed), longitudinal, fairly broad, very obtuse,
Carina running from immediately behind the
centro-basal tubercles to the elytral marginal
apex, and becoming a little more raised towards
the apex; completely finely and closely punc-
tured, with a number of large punctures scattered
here and there from some of which arise long
setae.

The prosternal protuberance fairly narrow,
particularly between the coxae; almost plane and
scarcely at all concave; strongly curved. The
mesosternal protuberance extremely broad an-
teriorly, sub-triangular, rather swollen, the sides
distinctly bisinuate, towards the apex not imme-
diately truncate, but for a very short but distinct
distance becoming parallel-sided, the apex
broadly truncate, and very broadly and shallowly
emarginate, the truncature about one and a
quarter times as wide as the breadth of the
prosternal protuberance medially. The abdomi-
nal segments of normal size; the apical ventrite
transverse, somewhat broadly emarginate; and
bearing a distinct median longitudinal groove on
its anterior half. The underside completely finely
and fairly closely punctured, with a number of
distinct larger punctures at the apex of the
apical ventrite, which bears short setae.

The legs of moderate length; the femora
strongly pedunculate; the tarsi slender, the first
segment of the posterior about equal in length
to the following two united; all the legs finely
and fairly closely punctured.

Ma/e.— Unknown.

Length, 7 mm.; breadth, 2.5 mm.

Locfl/i/y.— Brazil— Rio de Janiero.

Holotype, $, in my collection. Unique.

This new species at first glance
is very similar to Cosmotoma viridana Lacord-
aire, but on examination is seen to differ dis-
tinctly in not possessing a dark brown pubescent
transverse post-median band, the two pronotal
discal tumescences not separated medially, the
pronotum without large punctures anteriorly,
etc.

Cosmotoma melzeri sp. nov.

Text-fig. 11

Description. Female.— Davis, brown to fer-
ruginous in color, the elytra somewhat dark

18

Zoologica: New York Zoological Society

[40: 1

olivaceous-ferruginous in part. The head dark
ferruginous, the pronotum pitchy-brown in col-
or. The scutellum ferruginous. The elytra some-
what olivaceous-ferruginous in general dermal
color, with the apex of the centro-basal tubercles
black and black fasciculate, also bearing, at
about between the middle and apical two-fifths,
a narrow blackish pubescent transverse band,
which has its anterior border almost straight,
and its posterior border slightly anteriorly
curved; also variegated with silvery-gray rather
thin pubescence which extends rather narrowly
from the humeri round the centro-basal tubercles
posteriorly to the suture about midway between
the centro-basal tubercles, thence extending
rather broadly along the suture to the transverse
dark band which it borders anteriorly rather
narrowly to the margin; the transverse dark band
rather broadly bordered posteriorly with silvery-
gray pubescence, and the elytral apex also rather
irregularly, chiefly apico-laterally, sparsely gray-
ish. The underside ferruginous covered thinly
with grayish-white pubescence, which becomes
most dense on the sternum and latero-posteriorly
on the first abdominal segment. The antennal
scape greenish, the rest of the segments rather
dark ferruginous; with thin grayish-brown pu-
bescence. The femora and tibiae greenish, the
tarsi ferruginous; the legs covered sparsely with
grayish pubescence.

Not very robust, elongate-ovate. Antennae
distinctly longer than the body after about the

Text-fig. 11. Cosmotoma melzeri sp. nov. §
(X 7.2). Holotype.

sixth segment, (unfortunately the two apical seg-
ments are lacking, but the antennae will be prob-
ably a little more than one and a half times as
long as the body when complete) ; rather slender,
sparsely fringed beneath up to the fifth segment,
on which there are only a few setae and only
one or two at the apex of the sixth, the apex of
the third segment with a very sparse pencil of
hairs beneath, the fourth segment bearing a com-
paratively small (in this genus) brush of black
setae on little more than its apical quarter, which
does not completely encircle the segment, only
being present beneath and internally in part, the
segment a little swollen on the area from which
the setae arise. The scape rather elongate and
rather slender, not very swollen, extending a
little past the middle of the pronotum; the third
segment about one and a fifth times as long as
the scape, the fourth segment only about one
and an eighth times as long as the third, about
one and a third times as long as the scape, and
almost twice as long as the fifth segment, the
rest gradually decreasing to the ninth segment,
(thereafter broken, but presumably shorter to
the apex) ; the segments completely very finely
and closely punctured, with larger punctures
from which the setae arise, particularly on the
fourth segment. The antennal tubercles only
slightly raised, the head broadly and shallowly
concave between; the frons large, about equi-
lateral, moderately strongly convex, with an
extremely fine, indistinct, median longitudinal

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

19

groove; the lower lobes of the eyes small, dis-
tinctly narrowing interiorly, in fact almost sub-
triangular, slightly longer than broad, about
equal in length to the genae; the head completely
very finely and closely punctured, bearing a few
long setae at its anterior border and one or two
at the inner margin of the lower lobes of the eyes.

The pronotum not strongly transverse, only
about one and a quarter times as broad as long,
rather strongly convex, bearing two strongly
raised, conical, very obtuse tumescences on the
disc; the lateral border rather strongly rounded
and swollen anteriorly, running almost regularly
to the small lateral spinous conical tubercle,
thence rather strongly narrowed posteriorly to
the base; completely finely and closely punc-
tured, except for the apices of the discal tumes-
cences, which are impunctate and nitid; with a
few sparse very large punctures in the anterior
and posterior transverse grooves, these latter
being extremely shallow, very broad and almost
obsolete. The scutellum sub-triangular, moder-
ately broadly rounded apically; very finely and
closely punctured.

The elytra only moderately elongate, only
moderately convex, almost parallel-sided to
about the apical third, thence broadly rounded
to the apices, which are slightly obliquely trun-
cate, with the marginal angle produced into a
rather broad, strong, pointed spine; the centro-
basal tumescences strongly raised and black
fasciculate; completely finely and closely punc-
tured, with large punctures scattered here and
there from which long erect setae arise.

The prosternal protuberance moderately nar-
row, broadly rounded, broadly and slightly lon-
gitudinally concave medially. The mesosternal
protuberance very broad, somewhat swollen an-
teriorly, sub-triangular, a little sinuate laterally,
the apex broadly truncate, extremely slightly,
scarcely at all, broadly emarginate, the trunca-
ture about one and a half times as broad as the
breadth of the prosternal protuberance medially.
The abdominal segments of normal size; the
apical segment transverse, the apex very broad-
ly truncate, very slightly roundly so; the lat-
eral angles rounded, with a number of large
distinct hair-bearing punctures towards apex,
(the abdomen is somewhat ventrally deflexed so
that the anterior half of the apical ventrite Is
not visible and the groove on the anterior half,
which I think in this species must be very feeble,
is not visible; the strong setae-bearing apical
punctation is, however, normally a female char-
acter in this genus). The underside otherwise
completely very finely and rather variably closely
punctured.

The legs of moderate length; the femora
pedunculate; the tarsi moderately slender, the

first segment of the posterior tarsi about equal
in length to the following two united; all the legs
finely and closely punctured, with a few slightly
larger punctures scattered here and there.

Ma/e.— Unknown.

Length, 8.5 mm.; breadth, 2.9 mm.

Loca//7y.— Brazil— Bahia.

Holotype, $, in my collection. Unique.

Diagnosis.— This distinct new species differs
from all the other known species of the genus in
the much smaller size of the brush of setae on
the fourth antennal segment and in the rather
different pronotal shape, which is more rounded
laterally, etc. From Cosmotoma viridana La-
cordaire, it differs further in the almost straight,
narrow, dark-colored, post-median elytral band,
in being distinctly brownish in appearance and
not greenish, as well as other distinct characters.

Cosmotoma nigra sp. nov.

Text-fig. 12

Description. Ma/e.— Completely pitchy-black
above and below, without any markings of any
kind; the base of antennal segments five to eleven
pale ferruginous-yellow annulate on about their
basal quarter to third; completely covered,
sparsely and extremely finely, with short grayish
pubescence, which is nowhere dense enough to
give a grayish appearance, except perhaps where
a little denser on the upper surface of the inter-
mediate and posterior femora.

Not very robust, somewhat ovate-elongate in
shape, but the elytra somewhat attenuate api-
cally. The antennae about one and a half times
as long as the body, sparsely fringed below up
to about the sixth segment, segments three and
five to seven with thin pencils of setae beneath
their apices; the fourth segment bearing a large
distinct brush of black setae on about its apical
half, which does not completely encircle the
segment, but only about the inner longitudinal
half, the area bearing the setae distinctly swollen;
the scape moderately elongate, extending to
about the basal quarter of the pronotum, moder-
ately swollen; the third segment very slightly
shorter than the scape, the fourth segment about
one and a sixth times as long as the scape, almost
one and a third times as long as the third seg-
ment, and about one and three-quarter times as
long as the fifth, the rest gradually decreasing to
the apex; the segments completely finely and
closely punctured, except where the setae arise
from larger punctures. The antennal tubercles
only very slightly raised, the head slightly and
broadly concave between. The frons large, very
slightly transverse, moderately convex, with a
fine median longitudinal groove; the lower lobes
of the eyes small, only very slightly narrowing
inferiorly, almost sub-quadrate about three-

20

Zoologica: New York Zoological Society

[40: 1

Text-fig. 12. Cosmotoma nigra sp.
nov. $ (X 10.7). Holotype.

quarters as long as the genae; the head com-
pletely finely and closely punctured, bearing
only very few long setae anteriorly and at the
inner margin of the lower lobes of the eyes.

The pronotum transverse, moderately convex,
bearing two strong discal tumescences; with a
strong, broad, conical, obtuse swelling laterally
on each side, slightly post-medially; completely
very finely and closely punctured, with a few
sparse very large punctures scattered in the an-
terior and posterior transverse grooves, and few
anteriorly between the discal tubercles. The
transverse grooves broad, the posterior very
shallow, the anterior more distinct. The scutel-
lum sub-triangular, moderately broadly rounded
apically; finely and closely punctured.

The elytra only moderately elongate, not
strongly convex, narrowing towards the apices,
which are a little obliquely truncate, with the
marginal angle strongly spinously produced; the
centro-basal tumescences moderately strongly
raised and densely black fasciculate; completely
very finely and fairly closely punctured, with
larger punctures scattered here and there from
which erect setae arise.

The prosternal protuberance narrow, particu-

larly between the coxae, slightly longitudinally
concave, moderately strongly curved. The meso-
sternal protuberance very broad and rather
swollen anteriorly, sub-triangular, the apex trun-
cate and very slightly emarginate, the trunca-
ture a little wider than the breadth of the pros-
ternal protuberance medially. The abdominal
segments of normal size; the apical ventrite
transverse, somewhat broadly rounded, its apex
rather broadly emarginate. The underside com-
pletely finely and fairly closely punctured, with
a few slightly larger punctures on the sides of
the abdominal segments.

The legs of moderate length; the femora pedun-
culate; the tarsi slender, the first segment of the
posterior tarsi very slightly longer than the fol-
lowing two united; all the legs fairly finely and
fairly closely punctured, with a number of
slightly larger punctures scattered here and there.

Length, 7 mm.; breadth, 2.5 mm.

Loca///y.— Brazil— Santa Catherina.

Holotype, 5, in the Musee Royale d’Histoire
Naturelle de Belgique. Unique.

Diagnosis— This new species is conspicuously
different from Cosmotoma adjuncta Thomson
and all other known species in the genus in being
uniformly black in color.

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

21

Cosmotoma pallida sp. nov.

Text-fig. 13

Description. Female.— In general appearance
uniform yellowish-brown. The dermal color of
head, pronotum and underside dark pitchy-
brown, of the elytra fulvous-ferruginous; the an-
tennae, femora and tibiae greenish to greenish-
ferruginous, the tarsi light ferruginous, com-
pletely covered with very thin, short, fulvous
pubescence, without any trace of other color ex-
cept extremely slightly darker suturally at about
the apical third, and the black fasciculae.

Ovate-elongate; not very robust. The an-
tennae distinctly longer than the body (unfor-
tunately broken after the eighth segment in the
best specimen, but probably about one and
a third times as long as the body when com-
plete); very sparsely fringed below up to the
fifth segment, the third and fifth and perhaps
the sixth segment bearing thin pencils of setae
beneath; the fourth segment bearing a large
distinct brush of black setae on about its apical
two-fifths, which does not completely encircle
the apical area, but leaves about the outer lon-
gitudinal half naked, the area bearing the setae
distinctly swollen; the scape moderately elon-
gate, extending to about the basal third of the
pronotum, not very strongly swollen; the third
segment about equal in len^h to the scape, the
fourth segment about one and a third times as
long as the scape, about two and a third times
as long as the fifth segment, the rest gradually
decreasing to the apex; the segments completely

finely and closely punctured with larger punc-
tures from which the setae arise. The antennal
tubercles only a little raised, the head broadly
concave between. The frons large, about equi-
lateral, moderately strongly convex, with a very
fine median longitudinal groove; the lower lobes
of the eyes small, narrowing interiorly, about
one and a quarter times as long as broad, about
two-thirds as long as the genae; the head com-
pletely finely and closely punctured, with only
a very few extremely sparse setae at the lower
border and inner borders of the lower lobes of
the eyes.

The pronotum transverse, moderately convex,
bearing two strong discal tumescences; with a
moderately strong spinous swelling laterally on
each side, slightly behind the middle; completely
very finely and closely punctured, with a more
or less single irregular row of very large punc-
tures in the anterior and posterior transverse
grooves, which are broad and not very strongly
concave, also a few of these large punctures
antero-medially almost between the two discal
tubercles. The scutellum sub-triangular, a little
elongate, rather narrowly rounded apically; very
finely and closely punctured.

The elytra only moderately elongate, not very
strongly convex, more or less parallel-sided to
about the apical third, thence broadly rounded
to the apices, which are obliquely truncate, with
the marginal angle produced into a strong sharp
pointed spine; the centro-basal tumescences
moderately strongly raised and black fasciculate;

22

Zoologica: New York Zoological Society

[40: 1

completely very finely and closely punctured,
with larger punctures scattered here and there,
from some of which long erect setae arise.

The prosternal protuberance narrow, particu-
larly between the coxae, where it is rather
parallel-sided for a short distance; rather strongly
curved. The mesosternal protuberance very
broad and rather swollen anteriorly, sub-tri-
angular, the apex truncate and slightly, but
distinctly, emarginate, the truncature only
slightly wider than the breadth of the prosternal
protuberance medially. The abdominal segments
of normal size; the apical transverse, but a little
conical, the apex very broadly truncate, bearing
a distinct median longitudinal groove on its
anterior half. The underside completely finely
and fairly closely punctured, with numerous
large, setae-bearing punctures at the apex of the
apical ventrite.

The legs of moderate length; the femora
pedunculate; the tarsi slender, the first segment
of the posterior tarsi about equal in length to
the following two united; all the legs finely and
moderately closely punctured, with a few slightly
larger scattered punctures.

Male.— Quite similar in color to the female. A
little less robust, and the elytra less parallel-
sided, more attenuate to the apices.

The apical ventrite transverse, less elongate
than the female, more or less broadly rounded,
the apex slightly, but distinctly emarginate, the
lateral angles rounded; lacking, or with very
many fewer, large seta-bearing punctures.

(The comparative antennal length unknown
as the antennae are completely lacking in the
only male specimen examined) .

Length, 7. 5-8. 5 mm.; breadth, 2.8-3. 2 mm.

Loca/ity— Brazil— Santa Catherina (1^, 2$).

Holotype, 2, Allotype, $, in my collection.
Paratype, 2, (from my collection No. 6284) in
the United States National Museum. (Ex-
changed for the paratype of Cosmotoma fasciata
Fisher) .

I have made the female the Holotype because
two almost complete females were seen, the male
lacking its antenna.

Diagnosis.— This new species is conspicuously
different from Cosmotoma adjuncta Thomson
and all the other known species in the genus in
being uniformly fulvous-yellow in color.

COSMOTOMELLA gen. nov.

Description.— NlodevdAely elongate, sub-par-
allel, finely pubescent, silky or with silky reflec-
tions, with erect setae throughout.

Head large, concave between the antennal
tubercles; frons transverse; eyes small, finely
granulated; genae elongate. Antennae of male
about one and a half times as long as the body,

in the female about as long as or slightly longer
than the body; with erect hairs beneath; scape
slender, slightly longer than the third segment,
fourth segment longer than the third, the fifth
to eleventh gradually decreasing; fourth seg-
ment bearing beneath a fascicule of long hairs.
Pronotum about as long as broad, sub-globose,
strongly convex dorsally, the disc bearing two
feeble tubercles; broadly rounded, but only
feebly tuberculate, not spinous laterally. Scutel-
lum sub-triangular in male, broadly rounded in
female. Elytra straight in front, scarcely broader
than the pronotum at widest; moderately elon-
gate, rather straight-sided, sub-parallel laterally,
narrowing gradually to the apices which are
sinuately truncate, the marginal angles spinous;
a strong centro-basal tumescence. Legs of mod-
erate length; setose; femora pedunculate, very
swollen distally; tarsi rather short, the first seg-
ment of the posterior about as long as segments
two and three united. Prosternal protuberance
not very broad, curved posteriorly; meso-
sternal protuberance broad, triangular. Apical
ventrite broadly rounded and emarginate
apically in male, a little more elongate and
broadly rounded apically in female, the last
segment in female with a distinct median longi-
tudinal groove on the basal half.

Genotype: Cosmotomella zikani Melzer,
1927. Brazil.

This new genus, created for the reception of
the species described as Cosmotoma zikani
Melzer, differs conspicuously from Cosmotoma
Blanchard in lacking lateral pronotal spines, in
the centro-basal elytral crest not bearing a fas-
cicule of hairs, in the pronotal shape varying
between the two sexes, and the more elongate
and parallel form.

Table 1 gives the relative proportions of
length to breadth of Cosmotomella zikani
Melzer and the species of Cosmotoma Blanch-
ard. From this it wiU be seen that Cosmotomella
zikani Melzer is always slightly more than three
times as long (in total len^) as broad, whereas
all the species of Cosmotoma Blanchard are less
than three times. Also the elytral length of
Cosmotomella zikani Melzer is always slightly
more than twice the breadth, whereas in the
species of Cosmotoma Blanchard it is less than
twice.

Cosmotomella zikani Melzer
Text-figs. 14 2, 15 5

Melzer, 1927, Rev. Mus. Paulista, 15, 51 A.

A pair of specimens in my collection agree
with Melzer’s description almost completely,
except in one or two, perhaps minor, points.
Examination of these during the course of this
revision showed further sexual dimorphism than

1955]

Gilmour: Revision of the Genus Cosrnotoma Blanchard

23

Table 1. Comparison of Body Proportions be-
tween THE Species of the Genera Cosrnotoma
Serville and Cosmotomella Gilmour.

Body:

Elytra:

Species

Breadth ( 1 )
(Averages)

Breadth (1)

COSMOTOMA

olivacea Gilmour

2.58

1.79

sertifer Serville

2.77

1.95

nigra Gilmour

2.65

1.73

suturalis Gilmour

l.lf>

1.96

adjuncta Thomson

2.69

1.75

rubella Bates

2.69

1.79

nigricollis Bates

2.64

1.72

melzeri Gilmour

2.90

1.85

pallida Gilmour

2.73

1.91

triangularis Gilmour

2.59

1.74

viridana Lacordaire

2.82

1.81

Cosmotomella

zikani Melzer

3.14-3.3

2.14-2.15

noted by Melzer, and further differentiating
features from Cosrnotoma Blanchard, in which
Melzer originally placed the species, while not-
ing that it did not completely agree with the
generic characters of this genus. I have, there-

fore, as described above, created a new genus,
Cosmotomella, for the species zikani Melzer.

Comparison is not possible between the male
antennae, for in my specimen these are lacking
after the scape. Further, Melzer states that the
pronotum is unarmed laterally. I cannot believe
that such an apparently competent observer as
Melzer would overlook any trace of tubercle
and must therefore point out that while there is
only an extremely faint trace in my female
specimen, in the male it is visible, though feeble,
and somewhat more superior than in the genus
Cosrnotoma Blanchard. Perhaps it is possible
that in zikani Melzer there is an individual vari-
ation in pronotal shape. Melzer does not com-
ment on any undue sexual difference in
pronotal shape, but in my two specimens this is
quite conspicuous, as will be seen by the figures,
and yet I am quite sure that my specimens are
the same species, showing also a similar sexual
color difference as noted by Melzer.

In view of these differences I give, firstly a
translation of Melzer’s original description, so
that other workers may draw their own con-
clusions.

“Related to C. viridana Lacord., olivaceous-
piceous, clothed with silky silvery pubescence,
interspersed with long erect setae, the elytra
marked with a black fascia behind the middle;
the head large, concave between the antennae,
the frons transverse, sub-planate, the eyes small.

Text-fig. 14 (Left).
Cosmotomella zikani
Melzer. $ (X 7.0).

Text-fig. 15 (Right).
Cosmotomella zikani
Melzer. g (X 7.9).

24

Zoologica: New York Zoological Society

[40: 1

minutely granulated, the genae elongate; the
antennae of the $ a half longer than the body,
in the $ over-reaching the apex of the elytra by
the last three segments, the scape slender,
slightly shorter than the 3rd segment, the 4th
longer than the preceding, 5-11 sub-equal,
elongately fimbriate beneath, the 4th segment
ornated beneath the apex with a crest of hairs;
the thorax not broader than long, sub-globose,
the base strongly constricted, strongly convex
dorsally and armed with two obsolete tubercles
placed transversely in the middle, unarmed
laterally, the scutellum densely silky-silvery
pubescent; the elytra almost equal to the thorax
at the greatest width, straight truncate basally,
gradually attenuate posteriorly, the apices
themselves singly sinuately truncate, the ex-
ternal angles spinose, the centro-basal carinae
strong, not fasciculate, not costate posteriorly;
the legs subequal, the femora pedunculate,
strongly clavate; the prostemal process mod-
erately broad, the mesosternal process broad;
the metasternum rather densely sUky pubescent,
the mesosternal and metasternal epimera and
also the posterior margin of the 1st abdominal
segment with white pubescence.

“Length, 8.25-10 mm. 13, 2$.

“Locality.— Fazenda. Jerusalem, Estado do
Espirito Santo, Rio Muriahe, Estado do Rio de
Janiero, Rio Jose Pedro, Estado de Minas
Geraes, J. F. Zikan leg.

“Through its principal characters, this longi-
corn appertains to the genus Cosmotoma, but
because of the complete lack of the lateral
spines of the prothorax, etc., fails in similarity.

“Through the more parallel form, through
the lack of the spine on each side of the pro-
thorax, through ttie lack of the fascicules hairs
and of costae on the elytra and further through
the fascicules much reduced on the fourth an-
tennal segment, this species easily distinguishes
itself from C. viridana Lacord.

“The transverse band on the elytra in the
$ is narrow and opaque, in the 3, however, it
is much broader, lustrous and accompanied
through to the suture as if forming a bridge.”
(Translation of the original description.)

Male.— Head and pronotum black; elytra
pitchy to dark ferruginous; antennae, legs and
underside ferruginous. Head and pronotum
with variably dense olivaceous pubescence; on
the head most dense on the genae, between the
antennal tubercles and round the eyes; on the
pronotum most dense on the anterior half and
medially, on the minute lateral tubercle whitish.
The scutellum densely white pubescent. The
elytra with, in general, dark brown pubescence,
marked with olivaceous and white pubescence;
the dark brown pubescence becoming very

dense and forming a somewhat curved post-
median complete fascia which distinctly broad-
ens to the margin; the olivaceous pubescence
as follows: along the anterior border of the
fascia, a vitta from the humeri to the fascia
near the suture, and suturally on the anterior
half; the white pubescence as follows : a narrow
short pre-median vitta at the side of the disc,
and a broad oblique irregular area on the
posterior half, which gradually becomes oliva-
ceous round its borders. The antennae and legs
finely olive-gray pubescent, on the latter be-
coming white on the tibiae and tarsi. The
underside finely olivaceous pubescence; the
mesosternal and metasternal epimera, and the
postero-lateral border of the first abdominal
segment with dense white pubescence.

The antennal scape slender, rather elongate,
extending to about the basal third of the prono-
tum; finely and closely punctured; sparsely but
distinctly fringed beneath; (all the other seg-
ments unfortunately lacking in the only male
examined). The antennal tubercles moderately
raised, widely separated. The frons very slightly
transverse, moderately but distinctly convex;
the lower lobes of the eyes small, narrowing
interiorly, about as long as broad, about two-
thirds as long as the genae; the head completely
finely and closely punctured, with numerous
long, erect setae.

The pronotum about as broad as long, sub-
globose, only slightly rounded laterally to about
the middle, thence distinctly constricted to the
base; bearing laterally, slightly behind the
middle on each side a small obtuse tubercle; the
disc strongly convex, hearing medially on each
side a distinct broad, obtuse tubercle; the whole
very finely and fairly closely punctured, much
more sparsely on the discal tubercles, and with
two or three large punctures anteriorly and a
few posteriorly near the borders. The scutellum
sub-triangular, rounded apically; about as long
as broad; very finely and closely pvmctured.

The elytra elongate, attenuate to the apices;
a little broader basally than the pronotum me-
dially; the humeri distinct; rather convex, but
with a distinct elongate depression from the
inner side of the humeri; somewhat abruptly
declivous at about the apical fifth; the apices
each distinctly, slightly sinuate emarginate, the
marginal angle strongly spinous, the sutural
angle rather narrowly rounded; each elytron
with a strongly raised centro-basal tumescence;
the whole very finely and closely punctured,
with a few very large punctures on the disc just
behind the basal tumescences; with numerous
long erect setae.

The prostemal protuberance moderately
broad, shallowly, longitudinally depressed me-

1955]

Gilmour: Revision of the Genus Cosmotoma Blanchard

25

dially; moderately strongly curved. The meso-
sternal protuberance very broad basally, broadly
sub-triangular, the apex truncate, the truncature
about as broad as the prosternal protuberance
medially, somewhat plane above, gradually
rounded anteriorly. The abdominal segments of
normal size; the apical ventrite slightly trans-
verse, rounded apically, with a very shallow,
obtuse, median emargination. The underside
more or less completely finely and closely punc-
tured, the punctures more sparse on the ab-
dominal segments, the apical ventrite a little
coarsely punctured, particularly towards the
apex.

The legs of moderate length, (comparatively
a little more elongate than in the genus Cosmo-
toma) ; the femora pedunculate, strongly
clavate; the tarsi rather elongate and rather
slender, the anterior the broadest; the first seg-
ment of the posterior tarsi about equal in length
to the following two segments united; the tarsal
claws divaricate. All the legs very finely and
closely punctured; with numerous long erect
setae, particularly on the tibiae, which are also
most densely pubescent.

Female.— More robust than the male, and of
moderately similar aspect, but differing con-
spicuously in many ways on examination.

The general color lighter. The head and pro-
notum more or less uniformly and densely oli-
vaceous pubescent; this is a little yellowish on
the middle of the head and middle of the pro-
notum. The scutellum similarly densely white
pubescent. The elytral general brown pubes-
cence somewhat lighter; the dark brown trans-
verse fascia just behind the middle narrower,
almost straight, and not widening laterally; the
olivaceous pubescence as found in the male,
more grayish-olivaceous and a Uttle more ex-
tensive, the short white lateral premedian vitta
more or less lacking, the apical third rather
more extensively silky grayish to olivaceous-
gray pubescent. The legs more densely grayish
pubescent. The antennae grayish pubescent, be-
coming light brown from the apical third of the
fourth segment. The underside more olivaceous-
yellow pubescent; the metasternum almost com-
pletely grayish-yellow pubescent; the white
epimera of the mesosternum, metasternum and

first abdominal segment somewhat less distinct
owing to the more dense general pubescence.

The antennae about one and a sixth times as
long as the body; slender; fringed below, al-
though only sparsely and apically on segments
after the fourth; the fourth segment bearing a
moderate brush of setae below on its apical
third; the scape slender, moderately elongate;
the third segment about equal in length to the
scape, the fourth segment about one and a
third times as long as the scape; the following
segments gradually decreasing to the apex, the
fifth segment slightly curved; the segments com-
pletely finely and closely punctured. The frons
slightly transverse, much less convex than in
the male, and the median longitudinal groove
more distinct.

The pronotum much more distinctly globose
than in the male; the lateral tubercle almost
obsolete, scarcely discernible; the two discal
tumescences much smaller and less distinct, the
general convexity being stronger. The scutellum
larger, about as long as broad, not sub-trian-
gular, more or less regularly rounded.

The elytra slightly less attenuate than in the
male, comparatively not so much broader than
the pronotum; otWwise structurally similar
to the male; the centro-basal tumescences with a
few long erect setae, (these are only present on
the left elytron, so it is possible that they were
originally present in the male; they do not form
a fascicule); the lateral apical elytral spine a
little broader and not quite as elongate.

The underside similar to that of the male,
except, the apical truncature of the mesosternal
protuberance broader, about one and a half
times as broad as the prosternal protuberance
medially; the apical ventrite comparatively a
little more elongately conical, bearing a very
distinct median longitudinal groove on its an-
terior half, the apex broadly rounded and with
a number of very large coarse punctures and
short setae.

The legs a Uttle less elongate compared to
the male. Tarsal proportions similar to those of
the male.

Length, 8.5-9.S mm.; breadth, 2.9-3. 1 mm.

Locality.— Brazii: Espirito Santo (13); Rio
de Janiero (12).

Material Examined.— GiXvaoxa coUection, 2.

I

2

Two Little-known Selective Insect Attractants^

William Beebe
Department of Tropical Research
New York Zoological Society, New York 60, N. Y.

(Plates MV)

[This paper is one of a series emanating from the
tropical Field Station of the New York Zoological
Society, at Simla, Arima Valley, Trinidad, British
West Indies. This station was founded in 1950 by
the Zoological Society’s Department of Tropical
Research, under the direction of Dr. William Beebe.
It comprises 200 acres in the middle of the Northern
Range, which includes large stretches of undis-
turbed government forest reserves. The laboratory
of the station is intended for research in tropical
ecology and in animal behavior. The altitude of the
research area is 500 to 1,800 feet, with an annual
rainfall of more than 100 inches.

[For further ecological details of meteorology
and biotic zones see “Introduction to the Ecology
of the Arima Valley, Trinidad, B.W.I.,” William
Beebe. (Zoologica, 1952, Vol. 37, No. 13, pp. 157-
184).]

Contents

Page

I. Introduction 27

II. Fedegoso or Wild Heliotrope Association. . 28

III. Tangerine Association 30

A. Lepidoptera 30

B. Orders other than Lepidoptera 31

1. Coleoptera 31

2. Diptera 31

3. Hymenoptera 31

rv. Summary 32

V. References 32

I. Introduction

INSECT repellents have earned a generous
representation in entomological study and
literature, owing perhaps to their economic
importance.

Attractants appertain rather to pure science
and are more directly related to our studies of

^ Contribution No. 952, Department of Tropical Re-
search, New York Zoological Society.

animal behavior. They are abundant in variety
and varied in efficiency and, at the Zoological
Society’s station at Simla, Trinidad, B.W.I., are
brought to our attention the moment we leave
the laboratory. This will be evident from a few
random examples.

Under odor we may mention tree-bait, with
the old familiar mixture of such ingredients as
molasses and beer, attractive more especially
to moths at night; sexual odors, as the well-
known drawing of males to a caged female
from as far away as two to six miles; plant odors,
operating sometimes indirectly, as when an adult
butterfly oviposits on a larval food-plant.
Finally, we have the olfactory focus of over-ripe
fruit or decaying mouse which influences a
whole complex of eager receptors.

Light is one of the best-known attractors, as
indicated by the swarms around an electric
light globe or reflected light from a sheet. The
power of color is shown when an iridescent mor-
pho butterfly swoops down upon a bit of blue
paper. The auditory attractant of a cicada is
clearly evident to our senses, and the exact
tone of the wings of a female mosquito is an
attractant to her mate.

More mysterious than any of the foregoing
is the impelling force of mass migration, a yield-
ing to an instinct which, at least basically, is
inexplicable. At this moment, thousands upon
thousands of insects of almost all orders, are
pouring southward through a sLxty-foot wide
pass near Rancho Grande, Venezuela, repelled
from their place of birth or attracted to an un-
known area by some all-inclusive impulse.

The chief object of this casual and hetero-
geneous list of attractants is to point out that
certain ones are selective, carrying their mes-
sages to individual species, while the remainder
are more or less general, appealing to a wider
range of recipients.

27

28

Zoologica: New York Zoological Society

[40: 2

II. Fedegoso or Wild
Heliotrope Association

Early in the occupancy of the station at
Simla, we selected the family of moths, Euchro-
midae, for particular study. Among the con-
siderations which prompted this were intricate
instincts of the larvae, day-flying habits of
many of the adults and the frequency of ap-
parent mimicry.

During the first season, from December,
1952 to May, 1953, our collecting was restricted
to three methods: pursuit, in the field, of free-
flying moths; the capture of those which alighted
on the laboratory screens in the daytime; and
capture of the nocturnal forms that came to an
illuminated sheet.

All this was revolutionized the following
season by the use of the common weed, Helio-
tropium indicum Linnaeus, which proved to be
a remarkably efficient and selective attractant.
Our attention was directed to this phenomenon
by the notes of G. Hagmann (1938) and A.
Miles Moss (1947).

Our cultivated garden heliotrope is derived
from the South American wild species, Helio-
tropium peruvianum Linnaeus, and is charac-
terized by the clustered appearance of the blos-
soms, and their strong, sweet, vanilla-like scent.

The Indian heliotrope, Heliotropium indi-
cum, was named by Linnaeus 202 years ago.
It has spread from its native home in Asia, and
has been acclimated in the New World to such
a degree that its present distribution extends
from Virginia and Illinois south through Cen-
tral and South America to Buenos Aires. It has
received many common names, the one we have
chosen being “fedegoso,” although this intro-
duces a semantic misunderstanding. This is a
Portuguese word meaning ill-smelling, whereas
Hagmann characterizes it as “exquisite.” To our
senses the dry foliage of the heliotrope gives
forth a not-unpleasant, somewhat pungent,
musty smell, such as might distinguish a long-
used herbarium. Other names, such as eye-
bright, refer to alleged curative properties. Still
other terms, cocks-comb and scorpion plant,
relate to the shape of the flower sp^e.

Shortly after our return to Simla, on Decem-
ber 24, 1953, Research Assistant Rosemary
Kenedy located a plant of the fedegoso with the
assistance of the Botany Department of the
Imperial College of Tropical Agriculture. From
then on, assiduous search revealed many scat-
tered clumps of this plant. It grows in waste
places, such as old, neglected gardens and fields,
usually singly or in small clumps. It is a typical
weed, wholly undistinguished, without intensive
odor or color. This wild heliotrope is a small

plant, from one to four feet in height, with
single, curved spikes of small, pale lilac blos-
soms. These spikes or racemes are often divided
longitudinally into thirds; the terminal third
with unopened buds, the middle of full-blown
flowers, and the basal third of developing seeds.
The stems are hairy, the branches coarse, the
leaves large, and oval or ovate. The roots are
short and thick and have only a comparatively
slight hold on the soil.

In our use of the weed, uprooting is the first
step. A half dozen plants are pulled up and
shaken free of soil. At Simla the roots are tied
together with twine and the cluster is suspended
upside-down from some low branch or from a
stake driven into the ground. A favorite place
is along an open trail through the jungle or in
an area free of vegetation close to the forest’s
edge. The leaves shrivel soon after the plant
is collected and lose whatever of apparent sym-
metry or character they may have possessed.
During subsequent days of sun and rain the
foliage becomes in succession dry and brittle,
saturated and sodden.

For the first two or three days little activity
is observed around the withered plants. Then
one, two, a dozen butterflies and moths appear,
coming upwind, and all alight. The desiccated
racemes of flowers and seeds seem to exert espe-
cial attraction, but the stems, leaves and roots
are far from neglected.

The dominating point of interest in fedegoso,
as an attractant, is its selective quality. In the
course of five months of observation in and
around Simla, we detected members of only
four families of Lepidoptera coming to the
bunches of dead plants. Two of these were but-
terflies, Danaidae and Ithomiidae, and two were
moths, Euchromidae and Arctiidae. Other
groups, such as heliconids, nymphalids and
pierids, sometimes flew past the dry vegetation,
but no individual ever alighted or even hesi-
tated. Added to this is the fact that in most
classifications the four selected families are
placed at the top of their respective groups, pre-
sumably indicating extreme specialization in the
scale of lepidopteran phylogeny.

This remarkable, selective, attractant phe-
nomenon was reaffirmed during a few weeks of
our experience with fedegoso in Surinam, and
in reports from Para, of Hagmann and Moss.

The selective quality of fedegoso is apparent
m other than lepidopteran families. Details of
specific selection as shown in Euchromidae will
be discussed in future papers. One example will
suffice here. Until we began to use this method
of collecting we had never come across a speci-
men of Sphecosoma trinitatis Rothschild, a close
mimic of a small Polybia wasp. The resemblance

1955]

Beebe: Two Little-known Selective Insect Attractants

29

is so exact that only close examination reveals
the difference between moth and wasp. In the
course of five months we took or observed one
hundred and forty-seven of these wasp-mimick-
ing moths on fedegoso, and these resolved into
three distinct species, two of which had not
heretofore been recorded from Trinidad.

Arctiidae was the only other family of moths
attracted by fedegoso, and this sparsely, as only
fourteen specimens of seven species were taken,
three of which were uniques.

As euchromids were the dominant fedegoso
group among moths, so ithomiids were, far and
away, the more numerous of the butterflies. Fif-
teen species of Ithomiidae have been recorded
from Trinidad. Of these, on March 21, 1954, at
8 A.M., on two adjacent bunches of fedegoso,
I observed or collected eight species, totalling a
minimum of 257 individuals. The shade-loving,
skeleton-winged species Ithomia drymo pellu-
cida Weymer and Hymenitis andromica trifen-
estra (Fox) excelled in numbers, comprising
three-fourths of the total ithomiids.

We recorded small numbers of two species of
Danaidae, the first of which, Danaus plexippus
megalippe (Hiibner), is the tropical represen-
tative of our northern Monarch. The striking
Lycorea ceres ceres (Cramer) was present in
numbers and the same individuals often re-
turned day after day, conspicuous in their
ithomiid type of pattern and coloring.

There seems little doubt that the first appeal
of the attraction of fedegoso is through the
olfactory senses of the Lepidoptera, as evinced
by the approach upwind, together with the total
lack of advertising or directive coloring in the
plant.

Soon after the insect alights, or occasionally
just before, the tongue unrolls, the tip probing
and prodding about as the organ comes into
function. Careful examination of the surface
tissues of fedegoso fails to reveal any evidence
of liquid drops or other source of nourishment.
Nevertheless, so potent is some such aliment
that it affects the whole behavior of the insects.
After a short period of feeding the insect loses
its timidity and will often permit itself to be
picked up by the wings, examined and replaced,
or it may be captured by gently slipping a glass
vial over it. When crowds of ithomiids are clus-
tered close together, they will often buffet one
another without taking flight.

During this first season nothing was done
about chemical or other study of the nutritive
substance of fedegoso. While it exerts a notice-
able effect on the reduction of the escape reac-
tion, this is in no sense through what might be
called intoxication, as in the case of butterflies
feeding on fermenting fruit. The insects show

an obvious reluctance to leave their repast, but
once on the wing, their flight is swift, accurate
and typical. They quickly return to their feed-
ing on leaf or raceme.

A quotation from the notes on fedegoso in
Para, Brazil, by A. Miles Moss will present the
similarities and the differences of our observa-
tions on the same plant. The suspended, dried
plants, he says, “constitute a most remarkable
attraction for the great majority, though not
all, of the species of Syntomidae [Euchromidae],
as well as for the closely related Arctiidae; also
for many Danaiid butterflies, particularly the
Ithomiinae, for wasps of many kinds, for a few
beetles, for grasshoppers, for bugs, mosquitoes
and flies of all sorts.” Aside from Lepidoptera,
there is considerable discrepancy between re-
corded visitors as observed at Para and at Simla.
At the former place in Brazil numerous non-
lepidopteran visitors were observed. In our expe-
rience, such insects are rare or absent. In the
case of mosquitoes, Hagmann indicates that
they divide their interest between fedegoso and
the observer, who is attacked “unmercifully.”

Of the non-lepidopterans we noted only three
species of wasps, which occurred rarely, and on
plants desiccated for more than two weeks, so
old that they had lost their attractiveness for
moths and butterflies. Besides this, there were
two species of small longicoms which, on four
occasions, were found, usually in pairs, wander-
ing about the dry foliage. Now and then we had
hints of minor adaptations and inter-relation-
ships beginning to take shape on our fedegoso.
Twice we found flower spiders and once an orb-
weaver trapping and devouring unwary euchro-
mids. On another occasion a mantid had found
the dry vegetation good hunting and was holding
a euchromid in the grip of its forelegs, and a
ponerine ant was discovered carrying off an
ithomiid. The several times that we discovered
butterflies sleeping on the same twig from which
they had been feeding during the day suggested
opportunities for new arthropod relationships.

Heliotropium is a genus of plants of the bor-
age family, Boraginaceae, and the thought oc-
curred to us that botanical consanguinity might
carry with it some of this mysterious attraction
for insects. Miss Kenedy made a few prelimi-
nary experiments with shrivelled twigs and
leaves of two species of the genus Cordia,
namely, C. alliodora or Cypre, and C. cylindro-
stachya or Black Sage. These gave no positive
results.

Of a third genus of the borage family, Tour-
nefortia, we had no available material. Upon
our return to New York we received a letter
from Dr. P. A. Buxton referring us to notes
made by himself, 1926, and by Mr. G. H. E.

30

Zoologica: New York Zoological Society

[40: 2

Hopkins (1927) on trees of this genus growing
on Samoa and other Pacific islands. Dr. Buxton
writes that he found that “males of the genus
Euploea of several species, in several different
islands in the South Pacific, occur in numbers
on withered twigs of the tree Tournefortia. The
insect does not feed on that tree and the butter-
flies do not visit the flowers of that tree. The
observations were made repeatedly but one can-
not explain them.”

This gives hope that more extensive experi-
mentation with members of the borage family
may reveal other related plants that share this
attractant quality.

III. Tangerine Association

On March 6 and on two successive days I
observed a Cacao Caligo, Caligo eurilochus
minor Kaye, resting on a particular spot on the
branch of a tangerine tree, a citrus. Citrus nobilis
Andre, whose fruit is otherwise known as por-
tugal or mandarin orange. The tree was one of
a row in the citrus grove on Water Trail, a few
yards to the east of Simla Laboratory. The insect
was unusually fearless and permitted close ob-
servation. It was busily probing with its tongue
in a small area of what looked like the exudate
of a spittle insect.

Subsequent observation and consultation with
Dr. Egbert Tai of the Government Agricultural
Experiment Station tentatively identified the
puddle of white foam as probable evidence of
the Crotch Disease of tangerines. Of this malady
it is said that the cause is not known but a virus
is suspected. One of the most conspicuous symp-
toms is a white froth, which may ferment, and
oozes from openings in the bark. From time to
time there may be a rehealing. Examined under
a hand lens, bubbles are seen to surge up from
below, burst, and their place taken by others.
Except for these little scattered lesions the dis-
ease seemed no detriment to either blossoms,
fruit or foliage of the trees under observation.

On the branches of two adjacent tangerine
trees there were several small areas of the pale
exudation. These seemed to bubble and to over-
flow at times, and again in temporary drought
they dried up. At all times they were sources of
intensive attraction to certain insects. As in
fedegoso, the attractive agent was strongly and
definitely selective.

A. Lepidoptera

In the course of several weeks of intermittent
watching I observed fifteen species of butterflies
alighting at one or the other of three small areas
of bubbling froth, small puddles not more than
one-half by one inch. I bored holes through the
bark and wood, and daubed sugar and honey

on the trunks, but failed to distract the attention
of the insects from their chosen ambrosia. Sel-
dom did a butterfly alight on the trees except
within tongue’s reach of the froth. On the occa-
sions when this did happen, there was instant
approach to the attractive substance.

The fifteen butterflies were distributed among
three related families, Nymphalidae, Morphidae
and Brassolidae, in the proportion, of known
Simla species, of 33%, 100%, and 60%. Al-
though few in number of species, this group
nevertheless included the largest and, in color
and pattern, the most striking of Trinidad’s
tropical butterflies.

The following is a list of the species of Lepi-
doptera observed at the tangerine feeding le-
sions, together with a few casual notes.

The large. Green-checkered Nymphalid,
Victorina steneles steneles (Linnaeus), joined
the tangerine association three times. One
marked individual returned repeatedly for a
week. When feeding it kept its wings closed, not
flattened in the normal position of sunning. Also
it confined its visits to the sunniest, warmest
parts of the day.

Peridromia arethusa (Cramer), the Blue-
spotted Black, visited the trees day after day,
singly, or a pair at a time. They alighted in an
inverted position, with wings flat against the
bark. If at a distance from the froth, they ap-
proached by runs or short spurts, with a single
flap of the wings at beginning or end. Occa-
sionally a histerid beetle would push beneath
the wings, causing a momentary flapping but
not distracting the insect from its feeding. Once
a Colobura alighted on the flattened wings, send-
ing the insect into flight, but the Blue-spot re-
turned and in turn drove away its annoyer.

The Red-banded Hindwing, Biblis hyperia
(Cramer), is common about Simla but a rare
visitor to the tangerine trees. Twice it was seen
feeding on the froth, like Peridromia alighting
upside down, with flattened wings.

On April 3, an individual of the Six Orange-
spot, Catonephila numilia Cramer, came to the
bait. This was a new record for Simla and Ariraa
Valley. Two others were seen later.

The Zebra Clicker, Colobura dirce dirce
(Linnaeus) , was the only member of the tan-
gerine group to be found commonly in the gen-
eral vicinity at all times. During the present
period of watching, from two to eight indi-
viduals were present at the bubbling areas. They
had a regular flight routine, a quick, whirling
dash, followed by a swoop to the bark of the
nearest tree, alighting inverted, with closed
wings. They often progressed by a series of
short, quick runs over the branches. Zebra No. 1
had lost a full fourth of wing area, yet flew well.

1955]

Beebe: Two Little-known Selective Insect Attractants

31

It was present on the first day, March 6, and on
the last, May 18, the same individual was seen
feeding while perched on the wing cases of a
large, brown, eyed-elater. This butterfly was
among the most persistent in pushing aside the
beetles and flies which interfered with its feed-
ing. It not only pushed but slapped with side-
ways flicks of its tongue, to obtain for itself
free access to the froth.

A medium-sized leaf -brown butterfly haunted
the tangerines for a week before I could capture
and identify it. It proved to be Anaea morvus
morvus (Fabricius) deserving the vernacular
name of the Tailed Pygmy Morpho. On alight-
ing its wings snapped together and it became as
much of a stemmed, dead leaf as any Kallima.
Above, it was conspicuous, the proximal half
of the wings iridescent morpho blue, the re-
mainder black with two anterior, small, blue
spots. This was the second specimen to be taken
at Simla.

At first glance I thought that the family Itho-
miidae must be included in the tangerine asso-
ciation, but a butterfly observed on March 14
proved to be another nymphalid— Pro/ogon/Mj
ochraceus Butler, tailed and of typical ithomiid
pattern. Several were observed, one of which
crept close alongside a preponid and fed while
brushing wings with the larger species.

Historis odium orion (Fabricius) was re-
corded six times, comprising at least four differ-
ent individuals. In brief glimpses of the orange
and black of upper wings in flight it resembled
a brassolid; when alighted it closely approxi-
mated Prepona.

No day passed without one, two or all three
species of preponids being present, feeding upon
the frothy matter. These were Prepona demo-
phon (Clerk), Prepona antimache (Hiibner),
and Prepona meander (Cramer) . Owing to their
swift flight among the lights and shadows it was
impossible to differentiate the species on the
wing, but after alighting, the under wing pattern
was diagnostic. Especially was this true of
meander, with the sharply demarcated halves of
light and dark brown. All were seen to defend
their position against encroachment by other
butterflies or by beetles and flies. The tongues
of preponids are pale red and seem stout enough
to push and buffet aside any interfering insect.
These usually wary butterflies were all exceed-
ingly tolerant of approach, almost permitting
one to touch them before taking flight.

Two individuals of the Trinidad Morpho,
Morpho peleides insularis Fruhstorfer, were
members of the tangerine association, coming
to drink at odd times, and by sheer size and
weight taking possession of some of the food
areas.

Five Coconut Brassolids, Brassolis sophorae
sophorae (Linnaeus), were visitors to the tan-
gerines. Individual count was made possibly by
the various degrees and positions of wing dam-
age. Two of these insects made return visits
throughout five weeks.

Two out of the three species of Trinidad Owl
Butterflies or Caligos came to drink at the tan-
gerine supply. These were Caligo eurilochus
minor Kaye and Caligo illioneus saltus Kaye.
These were easily identifiable on their alighting,
because of their relative size, minor being ap-
preciably the smaller. The ocellus in saltus is
almost twice the size of the other. These great
butterflies seldom came into contact with the
others because of their crepuscular habits, ar-
riving in early morning and in late evening.

B. Orders Other than Lepidoptera

Less attention was paid to orders other than
Lepidoptera. They were few in number, both
of species and individuals. All were drinking at
the lesions.

1. Coleoptera

Large Green Elater, Chalcolopidius virens
(Fabricius). Three were seen at feeding troughs
on separate occasions. Did not seem to mind
being picked up and replaced.

Large Brown Eyed-elater, Pyrophorus pel-
lucens Eschscholtz. Two seen and captured.

Large Black Cetonia. One was captured on
May 9 and another seen the following day.

Black Histerid Beetle. Feeding almost every
day, usually in pairs.

Small Red-headed Histerid. Several every day.

2. Diptera

Stilt Fly, family Micropezidae, genus Odon-
toloxozus, species probably new. These were
present in small numbers every day, two to six
at each drinking station. They were very active,
moving forward, backward and sideways so
smoothly that they seemed to flow over the bark.
In spite of being constantly flicked aside by the
tongues of the butterflies, they persisted and
occasionally took their stand beneath the bodies
of the larger insects.

Hairy, Blue-bodied Fly, Tachinidae. Several
seen.

Drosophila. A group of a dozen congregated
at one source of nutriment during a week’s time.
Not seen at the other lesions.

3. Hymenoptera

Trigonid. Two seen, one taken. A large nest
of these bees was established a few yards away,
yet none, other than these two, was seen at the
tangerines.

32

Zoologica: New York Zoological Society

[40; 2: 1955]

IV. Summary

Two unrelated plants, under very different
conditions, have been found to be characterized
by their pronounced selective power of attrac-
tion for a few definite groups of Lepidoptera.

The first is fedegoso or wild heliotrope, Helio-
tropium indicum Linnaeus, which exercises a
powerful attraction for two families of butter-
flies, Danaidae and Ithomiidae, and two families
of moths, Euchromidae and Arctiidae. These
four families happen to be the most specialized
in their respective groups. The attraction or lure
becomes effective only on the death of the plant
and after the consequent shrivelling and desic-
cation of the foliage. It persists for ten days or
two weeks.

The second plant is the tangerine orange tree.
Citrus nobilis Andre, afflicted with what is prob-
ably a virus disease. The lure is the fermented
matter exuded from small, bark lesions. In this
case the attraction extends to three families of
Lepidoptera, Nymphalidae, Morphidae and
Brassolidae.

V. References

Beebe, W.

1952. Introduction to the ecology of the Arima

Valley, Trinidad, B. W. I. Zoologica, 37:
157-184.

Buxton, P. A.

1926. Euploea spp. frequenting dead twigs of
Tournefortia argentea in Samoa and
Tonga. Proc. Ent. Soc. London, 1: 35-37.

Dethier, V. G.

1947. Chemical insect attractants and repellents.
The Blakiston Co., Phila., 289 pp.

Hagmann, G.

1938. Syntomideos do Estado do Pard. Livro
jubil. Travassos, Rio de Janeiro, pp. 185-
194.

Hopkins, G. H. E.

1927. Butterflies of Samoa and some neighbor-
ing island-groups. Insects of Samoa and
other Samoan terrestrial Arthropoda. HI.
Lepidoptera. British Museum (Natural
History), London. Ease. 1, p. 15.

Moss, A. M.

1947. Notes on the Syntomidae of Pard, with
special reference to wasp mimicry and
Fedegoso, Heliotropium indicum (Bora-
ginaceae), as an attractant. Entomologist,
London, 80: 30-35.

Plate I

Fig. 1. Fedegoso {Heliotropium indicum Lin-
naeus) growing wild.

Fig. 2. Terminal inflorescence of fedegoso.

Plate II

Fig. 3. Fedegoso suspended from a branch, with
more than 30 butterflies feeding on the
seed panicles.

Fig. 4. Ithomiid butterflies attracted to fedegoso.

Plate III

Fig. 5. Ithomiid butterfly, Ithomia drymo pellu-

cida Weymer, male, with proboscis start-
ing to uncoil, alighting on fedegoso.
Day-flying euchromid moths feeding on
seed panicle of fedegoso. Above: A wasp
mimic, Pseudosphex melanogen Dyar,
male. Below: Dinia menu Hiibner, male.

Plate IV

Fig. 7. Prepona antimache (Hiibner) feeding at
tangerine lesion with histerid beetles.

Fig. 8. Prepona meander (Cramer) feeding at
tangerine lesion with Stilt Flies, Micro-
pezidae.

EXPLANATION OF THE PLATES
Fig. 6.

BEEBE

PLATE I

TWO LITTLE-KNOWN SELECTIVE INSECT ATTRACTANTS

BEEBE

PLATE II

TWO LITTLE-KNOWN SELECTIVE INSECT ATTRACTANTS

BEEBE

PLATE III

TWO LITTLE-KNOWN SELECTIVE INSECT ATTRACTANTS

BEEBE

PLATE IV

TWO LITTLE-KNOWN SELECTIVE INSECT ATTRACTANTS

3

A Factor Analysis of the Performance of Dogs on
Certain Learning Testsi

Anne Anastasi, J. L. Fuller, J. P. Scott & J. R. Schmitt
Fordham University and Roscoe B. Jackson Memorial Laboratory

(Text-figures 1 & 2)

Factor analysis is a statistical tech-
nique for simplifying and clarifying the
description of individual differences by
reducing the number of necessary variables or
dimensions. Such an analysis begins with the
intercorrelations among a set of variables, such
as the scores obtained by a group of individuals
on a series of tests. The object of the analysis
is to find the smallest number of factors or
dimensions which can account for the obtained
correlations among the test scores. Individual
differences may then be described in terms of
this relatively small number of factors, rather
than in terms of all the original tests. For sub-
sequent testing purposes, an effort is generally
made to choose or develop single tests which
provide the best measure of each of the factors
identified in the analysis. Detailed discussions
of the techniques of factor analysis and of its
mathematical foundations have been given by
Thurstone (1947), Holzinger & Harman
(1941), Thomson (1951) and Cattell (1952).
A very lucid elementary introduction to fac-
torial techniques can be found in the recently
published book by Fruchter (1954).

Although originally developed in connection
with the study of human abilities, factor analy-
sis has wide applicability. It has been employed
in such diverse areas as the investigation of
bodily physique and constitutional types, the
classification of psychoses and neuroses, the
identification of emotional and motivational

^ The raw data for this study were obtained at the
Hamilton Station of the Roscoe B. Jackson Memorial
Laboratory, Bar Harbor, Maine, while the statistical
analysis was conducted at Fordham University, New
York City. The cost of IBM computations was covered
partly by the Roscoe B. Jackson Memorial Laboratory
and partly by the New York Zoological Society. As-
sistance in making many of the arrangements necessary
to conduct this research was rendered by Dr. John V.
Quaranta, formerly Research Associate in Animal Be-
havior, New York Zoological Society.

traits, the study of interrelationships among
allergy reactions, the exploration of aesthetic
and humor preferences, the delineation of the
cultural patterns of different nations and the
analysis of the voting records of legislators and
Supreme Court judges. For general surveys of
the applications of factorial methods and for
critical evaluations of results, reference may be
made to Thurstone (1948), Fruchter (1954,
Ch. 10), Anastasi (1948) and Anatasi &
Foley (1949, Ch. 15). The special implications
of factor analysis for test development are con-
sidered in Anastasi (1954, Ch. 14).

Factorial analyses of infrahuman behavior
have been relatively few. A major reason for
the infrequent use of this technique in animal
studies stems from the difficulty of meeting cer-
tain important methodological requirements.
Among such requirements, special mention
should be made of the need for high test relia-
bility, a sufficient number of variables to permit
adequate determination and definition of each
factor, and a large enough group of subjects so
that chance errors of sampling will not loom too
large in the correlation coefficients. Owing to
their failure to meet one or more of these con-
ditions, even the best available factorial investi-
gations of infrahuman behavior must be re-
garded as preliminary and exploratory. And it
should be added that such a characterization
must also be applied to the study which will be
reported in the present paper.

The pertinent animal studies published prior
to 1950 have been summarized by Royce
(1950a). About a dozen investigations con-
ducted before 1935 reported correlations be-
tween two or more measures of learning. All
were concerned with rats, with the exception of
one study in which chicks were employed
(Dunlap, 1933). The correlations were uni-
formly very low, except those between closely
similar tasks, such as different mazes. There
was no evidence of a general learning factor

33

34

Zoologica: New York Zoological Society

[40; 3

(Dunlap, 1933; McCulloch, 1935). To be sure,
no common learning factor has been found in
the case of human subjects either, the abilities
or group factors identified in the human studies
being organized along different lines. Thus a
person who excels in spatial learning, for ex-
ample, may be quite deficient in verbal or
numerical learning. The early animal studies,
however, showed little evidence of any sort of
group factors beyond a few of very narrow
scope. A high degree of specificity seemed to
characterize the behavior measured in these
studies.

The first systematic investigation of animal
behavior by means of current procedures of
factor analysis is to be found in a study of rat
behavior by R. L. Thorndike ( 1935) . A total of
32 scores was obtained from seven experimental
set-ups, including mazes, problem boxes, con-
ditioned response apparatus, activity wheel and
an obstruction box for measuring the relative
strength of different drives. The subjects were
64 albino rats. Factorial analyses indicated the
presence of three factors, which were described
as docility, transfer and a conditioned response
factor.

Van Steenberg (1939) subsequently re-ana-
lyzed Thorndike’s data and rotated the centroid
axes for simple structure in accordance with the
procedures developed by Thurstone. Such a ro-
tation is now common practice in factorial
studies, its object being to obtain a more clear-
cut and easily interpretable configuration of
factors. Van Steenberg’s analysis yielded ten
factors, five of which could, according to the
author, be interpreted with some confidence.
These factors were identified as follows: ability
to profit from visual cues (common to elevated
mazes), adaptability to new situations, speed
of movement, ability to learn a right-left alter-
nation and visual insight or perception of the
total stimulus pattern. Of the remaining five
factors, three were very narrow factors specific
to one kind of apparatus; one admittedly defied
psychological interpretation; and one was re-
garded as a residual factor. It should be added
that the descriptions of the first five factors
themselves fall somewhat short of desirable
clarity. Nor does the extraction of ten factors
from intercorrelations obtained on only 64 rats
appear quite warranted.

In a later study, Vaughn (1937) applied cen-
troid analysis and rotation of axes to the inter-
correlations among a set of 34 measures ob-
tained from 75 rats. An even wider variety of
behavior was covered than had been the case
in Thorndike’s study, although most of the tests
were again concerned primarily with learning.
The apparatus included a wildness tunnel, an

activity cage, a straightaway, a perseverance
box, several types of mazes, a problem box and
a test designed to measure reasoning. Eight fac-
tors were isolated, four of which were tenta-
tively identified as follows: speed, wildness-
timidity, associative or insight learning, and
transfer.

Other ways in which factor analysis may be
applied to the investigation of animal behavior
are illustrated by the work of Wherry (1939,
1940, 1941) and Searle (1949). In Wherry’s
analyses, intercorrelations were found, not
among the scores obtained by each animal, but
among the numbers of errors made by the en-
tire group in different segments of the learning
situation. Thus each blind alley in a given maze
was considered as an “individual,” and the total
number of entrances made during a given trial
or stage of learning was taken as the “score”
for that learning period. Intercorrelations of
“scores” obtained in different periods were
found and submitted to a centroid analysis, with
subsequent rotation of axes. By this procedure,
Wherry sought to investigate changes in the
factorial composition of behavior at different
stages of learning. When applied to published
data from mazes and other types of learning
situations, this procedure yielded remarkably
consistent results. Factors described as forward-
going, food-pointing and goal-gradient pre-
dominated in the initial, middle and final stages
of learning, respectively.

Searle (1947, 1949) applied obverse- factor
analysis to rat learning data. In this method, cor-
relations are found between individuals rather
than between tests or other variables. The pro-
cedure can be visualized if we think of the col-
umns and rows of a table of scores as having
been interchanged prior to the computation of
intercorrelations. Each correlation thus ob-
tained indicates the degree of similarity of the
score patterns or profiles of two individuals.
When such correlations are submitted to a fac-
tor analysis, the resulting factors represent
clusters or “types” of individuals characterized
by similar score profiles.

Factorial techniques have likewise been ap-
plied to the analysis of emotional and motiva-
tional data obtained in animal studies. But the
results in this area are even more tentative than
those in the field of learning. Geier, Levin &
Tolman (1941) factor-analyzed 29 measures of
the behavior of 57 rats in two experimental
set-ups. Both learning and emotionality indices

2 Also known as “Q-technique” and sometimes in-
correctly described as “inverted factor analysis.” In
the terminology of matrix algebra, such an analysis in-
volves, not the inverse, but the transpose of the original
score matrix.

1955]

Anastasi, et al.: Performance of Dogs on Learning Tests

35

were represented in this study. An investigation
concerned only with the factorial composition
of emotionality indices was conducted on 40
rats by Billingslea (1942). Using a procedure
similar to that of Wherry, described above,
Rethlingshafer (1941) compared the factorial
composition of different stages of learning under
conditions of varying motivational strength.
Previously published data on rats were utilized
for this purpose.

More recently, Royce (1950b, 1951) applied
the centroid method of factor analysis to the
intercorrelations among 32 physiological, psy-
chological and social measures of emotionality
obtained from 53 dogs. Rotation of the cen-
troid axes yielded an oblique simple structure.
Of the ten factors thus identified, six were ten-
tatively interpreted as follows: psychophysio-
logical timidity, behavioral timidity, heart reac-
tivity to social stimulation, aggressiveness,
activity level and audiogenic reactivity. Many
of the animals utilized in the Royce investiga-
tion have been included in the sample employed
in the present study.

Procedure

The present study represents an exploratory
factorial analysis of the performance of dogs in
a variety of learning situations. The data were
gathered by members of the research staff of
the Division of Behavior Studies, Roscoe B.
Jackson Memorial Laboratory, Bar Harbor,
Maine, as part of a long-range project on genet-
ics and social behavior in dogs. A brief account
of the over-all research plan can be found in
a report by Scott & Fuller (1951). For descrip-
tions of the physical environment and of the
procedures followed in the care and rearing of
the dogs, the reader is referred to the Manual
of Dog Testing Techniques, edited by Scott &
Fuller (1950, pp. 4-9).

Subjects. — Seventy-three dogs of pedigreed
stock were included in the present sample. All
had been reared under uniform laboratory con-
ditions and had been put through a standardized
system of handling, training and testing. De-
tailed genetic records on each animal are avail-
able at the Jackson Laboratory.

Table 1 shows the breed and sex distribution
of the subjects. It will be noted that the group
comprised 16 Basenjis, 4 Beagles, 18 Cocker
Spaniels, 5 Shetland Sheep Dogs, 7 Wire-haired
Fox Terriers and 23 Basenji-Cocker Spaniel
crosses. There was a total of 34 males and 39
females. The animals employed represent all
those for whom complete data were available
on the variables under consideration.

Tests.— The present analysis is based on the
scores obtained in the 17 variables described

Table 1. Breed and Sex Distribution of Sub jects

Breed

Male

Female Total

Basenji

8

8

16

Beagle

3

1

4

Cocker Spaniel

8

10

18

Shetland Sheep Dog

3

2

5

Wire-haired Fox Terrier

2

5

7

Basenji-Cocker Spaniel Cross

10

13

23

Total

34

39

73

below.3 More detailed descriptions of the tests
from which these scores were derived are pro-
vided in the previously cited Manual of Dog
Testing Techniques (Scott & Fuller, 1950). In
order to facilitate cross-references, the test
names given in the manual have been employed
in the present report, even when an objective
examination of the test might suggest the de-
sirability of a somewhat different name.

1. Habit Formation: Time.— The dog is placed
in a small release cage, while a food box is
placed a few feet away. The release box is re-
motely operated by the unseen observer. Two
trials a day are given over a five-day period,
the food box being placed in one position for
the first two days and in another position dur-
ing the last three. The score is the total time
required to reach the food box in ten trials.

2. Manipulation: Time.— The same apparatus is
used as in variable 1, except that the dog can-
not reach the food without first biting, nosing,
or pawing the food dish out of the food box.
Two trials a day are given for two days. The
score is the total time required to obtain the
food in four trials.

7). Manipulation (String-pulling): Time. — The
same apparatus is again used as in variables
1 and 2, except that the food dish is placed
well back in the food box and must be pulled
out by a string which is attached to the rim of
the dish. The score is the total time required to
obtain the food in two trials.

4. Maze (Second Barrier Test): Errors.— The ap-
paratus for this test consists of a six-unit
T-maze with a food dish at the exit. Since the
barriers are made of poultry netting, the solu-
tion to each part of the maze is visible, but not
that of the whole. Following a two-day orien-
tation period, each dog is put through one trial

3 Other variables were considered and discarded be-
cause of lack of experimental independence of scores.
The measures omitted for this reason include a coyer-
lifting test and three scores obtained in a discrimina-
tion and delayed-response apparatus. In all these tests,
subjects who had failed an earlier related test were not
subsequently tested, but were automatically recorded as
failures in the new test.

36

Zoologica: New York Zoological Society

[40: 3

CC— Closed Corridor. D— Door banged as cue. E— Partition which is swung either
right or left in variable 5 (Motivation) so as to completely obstruct one corridor.

a day for ten days. An error is recorded when-
ever an animal stops or reverses direction. The
score is the total number of errors in the last
nine of the ten trials.

5. Motivation {Discrimination Apparatus): Time.
—This test was designed to measure the motiva-
tional strength of each animal, as indicated by
his running time in escaping from an enclosed
area and in reaching food. The discrimination
apparatus (cf. Text-fig. 1) is utilized as the
enclosed area. This apparatus has a starting
box and two escape corridors. In the motiva-
tion test, one escape corridor is completely
closed off so that the dog has only one possible
route. To provide a visual and auditory cue,
the experimenter bangs the inner door of the
correct escape corridor four or five times, just
as he releases the dog from the starting box.
After escaping from the apparatus, the dog is
allowed to run freely in the room and is fed by
the experimenter near the starting box. The

order of presentation of escape corridors
(RLRRLRLLRL) is designed to avoid the
formation of position habits or of a simple
alternation habit. The score is the median time
required to escape in ten trials.

6. Cue Response {Discrimination Apparatus) :
Trials.— Cue response training is started on the
day following completion of the motivation
test. The apparatus is arranged so that the
center partition extends directly forward, as
shown in Text-fig. 1, rather than being swung
right or left as in the motivation test. The dog
must therefore choose one side or the other
before he can see which outer door is open.
Entrance into the wrong corridor counts as an
error. The cue is given as in the motivation test
by swinging the inner corridor door four or
five times and releasing the dog from the start-
ing box immediately thereafter. The outer door
of the uncued corridor is closed so that the dog
can never escape from the wrong corridor. A

1955]

Anastasi, et al.: Performance of Dogs on Learning Tests

37

uniform random sequence of right and left
escape corridors is again used. The score is the
number of trials required to reach either of two
pre-established criteria of learning in which
the proportion or sequence of correct choices
exceeds chance at the .01 level of significance.

/. Cue Response {Discrimination Apparatus) :
Time.— This variable is the same as variable 6,
except in the scoring. In the present variable,
the score is the median time required for the
last 19 correct trials of cue response training.

8. Leash Control (In) : Trials.— The dog is re-
moved from the outside pen before the leash
is put on him. He is then led on leash over a
short course of outdoor pathways and brought
into the building, where he is fed a small
amount of fish. The training is continued for
ten days. The dog’s performance is rated by
assigning differential weights to errors of vary-
ing degrees of seriousness, including balking,
pulling, dragging, fighting the leash, “sun-
fishing,” crossing in front of or jumping at the
trainer and various kinds of vocalizations such
as whining, yelping or howling. The score is
the number of trials required to reach a per-
formance rating of 2 or less.

9. Leash Control {Stairs): Trials.— During the
last eight days of leash control training, the
dog is also taken through the building and up
the stairs to one of the lofts, where he is fed.
The course includes a flight of stairs interrupted
by a landing. A rating scale similar to that of
variable 8 is used to score the subject’s per-
formance while climbing up and down the
stairs. The score is the number of trials re-
quired to reach a performance rating of 2 or
less on this portion of the course.

10. Leash Control {In): Initial Errors.— This vari-
able is the same as variable 8, except in the
scoring. The score is the error rating obtained
on the first day of training in leash control.

1 1. Leash Control {Stairs): Initial Errors.— This
variable is the same as variable 9, except in the
scoring. The score is the error rating obtained
on the first day of training in stair climbing.

12. Motor Skills: Time.— This test was designed
to measure the dog’s general physical skill,
especially in relation to climbing, jumping
and balancing. Two boxes, each one foot high,
are stacked and a two-by-five-foot ramp leads
to the top box where a food dish is placed.
The dog is released about six feet from the
apparatus and is timed from his release to the
moment when he reaches the food. The score
is the total time on three trials.

13. Fi>j/ Barrier Test {First Problem): Errors.—
This test is of the “Umweg” type and was de-
signed to test performance in a situation which
is totally new to the dog. It may also test
generalization or transfer of training from a
simple situation to a more complex one of
the same type, as represented by variables 14
and 15. The test is conducted within a large

rectangular area, two days being initially em-
ployed to accustom the animal to this area.
The barriers consist of five wood-and-wire
fences, six by three feet, with supports on one
side. Each is covered with opaque brown paper,
except for a one-foot-wide window in the
center of the barrier behind which the dog is
placed. The barriers are always set up so that
the supports are on the side on which the dog
is placed. In the present variable, only one
barrier is used. The dog is first allowed to smell
the food and is then placed on the opposite
side of the barrier, where he can see the food
through the window (cf. Text-fig. 2— First
Problem). The dog can solve the problem by
taking either of the two possible paths to the
food. An error is recorded whenever the ani-
mal stops or reverses direction. The score is
the total number of errors on three trials.

14. First Barrier Test {Second Problem) : Errors.—
This variable is the same as variable 13, except
that three barriers are placed end to end so as
to form a longer straight-line obstruction (cf.
Text-fig. 2— Second Problem). The score is
the total number of errors on three trials.

15. First Barrier Test {Third Problem): Errors.—
The procedure and apparatus are again the
same as those described under variable 13,
except that five barriers are set up in a U shape
(cf. Text-fig. 2— Third Problem). The score is
the total number of errors on three trials.

16. Obedience {Adjusted Stay Score): Time.— A
choke collar with a short lead is placed on the
dog and he is led to a box which is 20 inches
high and 20 X 16 inches on the top. The dog
is lifted to the top of the box and given the
command, “Stay.” The lead is held so that the
dog is choked if he leaps from the box. When
the animal learns to remain on the box for
30 seconds, training for responding to “Down”
is begun. When he stays up for 30 seconds
and jumps promptly at “Down,” training
without a collar is started. The experi-
menter stands within 6 inches of the box but
does not touch the dog or restrain him from
jumping. If the dog remains on the box for
30 seconds and jumps promptly at “Down,”
the distance of the experimenter is increased
on the next trial. The control distances are:
6", 18", 36", 72", 144", and out of sight behind
a screen placed 14 feet from the box (BHS).
A total of three days is devoted to the above
training. On the fourth or test day, the dogs
are tested in the following sequence of control
distances: 6", 18", 72", 144", BHS, 6", 18",
72", 144", BHS. The score employed in this
variable is the “adjusted stay score,” i.e., the
total time during which the dog remains on
the box in the ten 30-second test trials (max.
= 300 sec.), minus a 10-second penalty for
each failure to jump within 10 seconds of the
command “Down.”

17. Obedience: Jumps during Training.— This vari-
able is similar to variable 16, the score being

38

Zoologica: New York Zoological Society

[40: 3

FIRST PROBLEM

D

9

0-F

SECOND PROBLEM

THIRD PROBLEM

S S ^ ^^3

Da irO-F

S 8 3

Text-fig. 2. First barrier test. S— Supports for barriers. F— Food dish. D— Dog. W— Window
in barrier.

the number of spontaneous jumps with collar
on, or at the minimum distance of 6 inches
without the collar, during the training period.
Each such jump constitutes an error.

Results

Conversion of Scores.— Prior to the computa-
tion of intercorrelations, the scores on all vari-
ables except two (variables 3 and 6) were con-
verted to single-digit, normalized standard
scores.^ The converted scale ranges from 0 to 9,

* A list of raw scores, as well as details of the score
conversion and other computational procedures, can
be found in Schmitt (1954).

with a mean of 4.5 and a standard deviation of
2. It will be recalled that the raw scores on all
variables except 16, the adjusted stay scores on
the obedience tests, were expressed so that the
higher the score the poorer the performance. In
the converted scores, however, 9 represents the
best performance and 0 the poorest in all vari-
ables.

Some of the converted distributions retained
a certain amount of skewness or other irregu-
larities. Such variations result from the occur-
rence of an excessive number of identical scores
either at the upper or lower end, or at some

1955]

Anastasi, et ai: Performance of Dogs on Learning Tests

39

Other part of the range. Since all such identical
scores were assigned the same converted score,
the frequency of a given converted score some-
times exceeded that required by the normal
curve transformation. Nevertheless, the con-
verted distributions of 15 variables were deemed
to be sufficiently close to a normal curve for
use in the computation of Pearson correlation
coefficients. In the case of the remaining two
variables, however, the marked skewness re-
sulting from the large number of failures led to
the decision to dichotomize the variables. This
was done for variable 3, string pulling, and vari-
able 6, trials to learn cue response.

Intercorrelations.— The intercorrelations
among the 17 variables were computed by IBM
procedures at the Test Division of The Psycho-
logical Corporation, New York City. All are
Pearson correlations, except that between vari-
ables 3 and 6, which is tetrachoric, and those
between variables 3 or 6 and the remaining vari-
ables, which are biserial. The complete set of
136 correlations is reproduced in Table 2. The
correlations range from 4-.71 to —.43, including
82 positive and 54 negative coefficients. For a
sample of 73 cases, the minimum correlations
significant at the .05 and .01 levels are ±.232
and ±.302, respectively. Reference to Table 2
shows that 44 coefficients reach or exceed the
.05 level of significance; and of these, 27 reach
or exceed the .01 level. By chance, between 6
and 7 of the 136 correlations would be ex-
pected to reach the .05 level, and only 1 or 2
of these should reach the .01 level.

Factor Analysis.— The intercorrelations were
analyzed by Thurstone’s complete centroid
method (Thurstone, 1947, Ch. 8). The criterion
employed for determining how many factors to
extract was that developed by McNemar
(1942). According to this criterion, the sth
factor is significant if the estimated SD of the
partial correlations remaining after the extrac-
tion of s factors exceeds the standard error of
a zero correlation. The SD of the partial cor-
relations is estimated by the following

formula: = in which is the SD

of the sth factor residuals and ^^,2 is the mean
communality of s factors. With 73 cases, the
standard error of a zero correlation is .1179.
This value is slightly less than that of ^ (.1257 ) .
but exceeds that of (.1168). Factorization
was therefore discontinued after the extraction
of the 5th factor.

In Table 3 will be found the centroid factor
matrbc, showing the weight of each of the five
factors in each of the 17 variables, as well as

the communality, or proportion of common
factor variance, in each variable. The centroid
axes were next rotated graphically in such a way
as to maximize the number of zero factor load-
ings (simple structure), while retaining the
orthogonal relationship among the axes. The
rotated factor matrix is reproduced in Table 4.

It will be noted that the mean communality
is .46. Factor II contributes the largest propor-
tion of common variance, .12. Factors I and
IV each contribute .10; and Factor V accounts
for .09. The smallest contribution, .05, is made
by Factor III. The uniqueness of the variables,
including unknown proportions of specificity
and error variance, accounts for as much as 64
per cent of the total variance of the battery.

Interpretation of Factors.— In order to arrive
at a provisional psychological interpretation of
each of the five rotated factors, all variables
having loadings of ±.40 or higher on that fac-
tor were examined. Such a factor loading
accounts for 16% or more of the variance of
the particular variable.

Reference to Table 4 shows that the variables
which meet the above criterion with regard to

Factor I are the following:

13. First Barrier Test (First Problem):

Errors .47

11. Leash Control (Stairs):

Initial Errors .47

9. Leash Control (Stairs): Trials .46

16. Obedience (Adjusted Stay Score):

Time —.44

17. Obedience: Jumps during Training —.53

The type of behavior involved in all these
tests suggests that Factor I may be related to
activity and impulsiveness. In the first barrier
problem, the more active or impulsive animal
is less likely to hesitate or reverse direction in
going to the food dish. It might be added that
the second and third barrier problems (vari-
ables 14 and 15) also show appreciable positive
loadings on Factor I, but of decreasing magni-
tude (.39 and .31). These two problems would
also favor the more impulsive animal, since
hesitations and reversals again constitute the
only errors. Lower loadings would be expected,
however, than on the first problem, which rep-
resents the animal’s initial contact with a rela-
tively strange situation. Moreover, because of
their greater complexity, the second and third
problems may depend more heavily upon cog-
nitive factors than upon mere impulsiveness or
general activity level.

In the two measures of stair climbing (vari-
ables 9 and 11), the more active animal is less
likely to manifest such behavior as balking and
dragging, both of which are scored as errors.

Table 2. Intercorrelations Among the Seventeen Variables

(N = 73)

40

Zoologica: New York Zoological Society

[40: 3

o

O Tf
VO
m

^ VO
On ^
(S

Tf

ON

O

*-< 00
^ o
00 ^
cs ^

<s

VO

<s o

•0 Tf

Ov 00 O

On O

<s

O

r

Os

o

o

o ^

Tf

<N

o

Tf

<s

<s

0\ 00
Tf

VC VO
O <N

<S

Tf

cs

o

00

o

m On
O

m <s

T-j vr>

m 00 0\

ON On

cn TM <s|

Tf

<s

Tf ^

Tf 00

rn VO cT*i

^ n rj

00 On
o

v-i 00 On

VO «n
O VO 00

m »— f ^

I I

O

On

fS

o

VO 00

^ r- Tf

VO ON
-T rs

^ Tf
VO ON
Tf m
^ O

m o

-H On

O

o

Tf

»r>

O

o

H s

O' S PS

00 o 'O

o 't <N

O

r

O OO >ri

o « r-

n VO O'

rj T— •— I

I I

00

Tf

Tf

o

o

<s

r-

<s

Tf

Tf

VO

00

m

ON

00

On

<s

00

cn

o

On

r-

cn

(N

VO

00

o

<s

Tf

O

Tf

1

Tf

1

o

1

o

1

o

1

1

o

i

fS

1

o

1

o

1

<s

»

»

♦

»

*

*

4

*

*

*

«

*

00

ON

o

<s

<s

<s

<s

<s

•o

On

m

o

<S

On

o

ON

On

o

r-

00

VO

m

00

VO

r-

VO

\o

00

r-

Tf

Tf

NO

o

o

<N

o

o

o

cn

o

o

1 I

NO <S
On On <S

VO VO O

m

>

0\ o ^

All correlations are Pearson Product-Moment Coefficients unless otherwise indicated.
* Biserial correlation,
t Tetrachoric correlation.

1955]

Anastasi, et al.: Performance of Dogs on Learning Tests

41

Table 3. Centroid Factor Matrix

Variable

Factor

I

II

III

IV

V

1. Habit Formation: Time

.5116

.2032

-.3358

.3516

.1556

.5636

2. Manipulation: Time

.3383

.3634

-.3985

-.1795

-.1319

.4549

3. Manipulation (String pulling): Time

.4722

.2973

-.4206

-.2253

.0706

.5440

4. Maze (Second Barrier Test): Errors

-.2371

-.3162

-.1768

.1848

.3375

.3355

5. Motivation (Discrim. Appar.): Time

-.3519

.7085

.2551

.1168

-.2046

.7464

6. Cue Response (Discrim. Appar.) : Trials

.1709

.6063

.3021

.1391

-.4237

.6869

7. Cue Response (Discrim. Appar.) : Time

-.2978

.6478

.1535

.1362

-.1217

.5653

8. Leash Control (In) : Trials

-.5104

.3040

.2917

-.2519

.1920

.5383

9. Leash Control (Stairs) : Trials

.2853

.1631

.3100

-.1146

.1405

.2370

10. Leash Control (In) : Initial Errors

-.2687

.4404

.1671

.0547

.2501

.3596

11. Leash Control (Stairs): Initial Errors

.0940

.4113

.3916

-.0576

.3299

.4435

12. Motor Skills: Time

.2439

.1531

-.1104

.1340

.0764

.1189

13. First Barrier Test (First Problem) : Errors

.5578

.2211

.1079

.1337

.2011

.4300

14. First Barrier Test (Second Problem) : Errors

.5917

-.1248

.1292

.1608

.2805

.4869

15. First Barrier Test (Third Problem) : Errors

.6032

-.1463

.1804

.2085

-.0697

.4661

16. Obedience (Adjusted Stay Score): Time

-.4147

.2231

-.2711

.1944

-.0679

.3376

17. Obedience: Jumps during Training

-.6356

.1675

-.3209

.0961

.1735

.5744

On the other hand, the negative weights of the
two obedience measures are understandable,
since an active, impulsive dog is more likely to
jump down and finds it more difficult to remain
motionless for the required period. It may also
be suggested, as a further elaboration or descrip-
tion of Factor I, that this factor indicates con-
fidence in a strange situation and lack of tim-
idity. It is noteworthy in this connection that
all three variables which have high positive
loadings on this factor are based on tests admin-
istered outside the animal’s normal living en-
vironment.

Turning our attention to Factor II, we find
loadings of .40 or more on the following

variables :

8. Leash Control (In): Trials .57

10. Leash Control (In): Initial Errors .56

17. Obedience: Jumps during Training .54

5. Motivation (Discrimination
Apparatus) : Time .54

7. Cue Response (Discrimination
Apparatus) : Time .53

11. Leash Control (Stairs):

Initial Errors .40

15. First Barrier Test (Third Problem):
Errors —.45

This factor appears to involve docility or re-
sponsiveness to a human trainer. Its two highest
positive weights occur in those measures of
leash control in which the animal is led over an

outdoor course. A loading of .40 is likewise
found in initial errors made when being led up
the stairs on leash. Subsequent stair-climbing
performance, however, shows no significant
loading with this factor, probably because such
performance soon becomes primarily a matter
of motor skill or activity rather than responsive-
ness to the trainer. Similarly, the measure based
upon performance during obedience training
(variable 17) has a loading of .54 on this fac-
tor. A lower loading of .32, which may also be
significant, is found in the performance meas-
ure obtained after completion of obedience
training (variable 16).

The motivation and cue response (time)
measures both involve speed of escaping from
the discrimination apparatus. It will be noted
that the pattern of weights on all five factors is
closely similar for these two variables. With
reference to the present factor, it should be re-
called that in both tests the animal receives
food from the experimenter, whom he must
approach specially for this purpose, since the
experimenter does not stand near the exit of the
apparatus. Relation to the human trainer thus
appears to play a more important role in these
tests than in those in which the animal obtains
food impersonally from a dish.

In this connection it is also interesting to
observe the negative weight on the third barrier
problem. The first two barrier problems likewise
have negative weights on this factor, although
the weight is negligible for the first problem. It

42

Zoologica: New York Zoological Society

[40: 3

Table 4. Rotated Orthogonal Factor Matrix

Variable

1

n

Factor

III

IV

V

h^

1. Habit Formation: Time

.10

-.04

.00

.73

.00

.56

2. Manipulation: Time

.20

.00

.41

.37

-.32

.45

3. Manipulation (String pulling): Time

.31

.00

.46

.46

-.13

.54

4. Maze (Second Barrier Test) : Errors

-.34

.09

.00

.03

.47

.35

5. Motivation (Discrim. Appar.): Time

.01

.54

-.27

-.13

-.60

.74

6. Cue Response (Discrim. Appar.) : Trials

.26

.09

-.31

.08

-.71

.68

7. Cue Response (Discrim. Appar.) : Time

-.02

.53

-.20

-.04

-.50

.57

8. Leash Control (In): Trials

.05

.57

-.02

-.46

-.04

.54

9. Leash Control (Stairs) : Trials

.46

.03

-.12

.00

.01

.23

10. Leash Control (In) : Initial Errors

.07

.56

-.15

-.06

-.08

.35

11. Leash Control (Stairs) : Initial Errors

.47

.40

-.23

-.02

.00

.43

12. Motor Skills: Time

.11

.02

-.01

.32

-.03

.12

13. First Barrier Test (First Problem): Errors

.47

-.03

-.17

.42

.02

.43

14. First Barrier Test (Second Problem) : Errors

.39

-.24

-.20

.37

.29

.47

15. First Barrier Test (Third Problem) : Errors

.31

-.45

-.28

.30

.04

.47

16. Obedience (Adjusted Stay Score): Time

-.44

.32

.06

.04

-.19

.34

17. Obedience: Jumps during Training

-.53

.54

.16

-.06

.03

.60

n

.10

.12

.05

.10

.09

.46

will be recalled that in all three barrier prob-
lems, the animal must walk away from the food
in order to circumvent the barrier. In the second
problem he must walk farther than in the first;
and in the third, which presents a U-shaped
barrier, he must turn completely around and
walk in the direction opposite to that of the
food. Moreover, in all these problems, the ex-
perimenter sits by the food dish and is visible
through the window in the screen. An animal
which is unduly dependent upon the human
trainer might thus be handicapped in these
problems— and particularly on the third— since
the correct solution requires that he begin by
walking away from the visible experimenter.
The less docile and more “socially independent”
animal, on the other hand, tends to respond to
the physical elements of the situation, with little
or no regard for the position of the experi-
menter.

The only variables which meet our criterion
for the interpretation of Factor III are :

3. Manipulation (String pulling) : Time .46

2. Manipulation: Time .41

The factor may thus be named manipulation,
in the sense of pawing, nosing, biting or pulling
with the teeth. The measures listed above are
the only variables which require such activities.
To be sure, this factor is underdetermined.

insofar as it has weights of .40 or more in only
two variables. At the same time, the proposed
interpretation of this factor is supported by
the consistent pattern of low negative weights
in variables involving the discrimination ap-
paratus, leash control and the barrier problems.
In all these tasks, any biting, nosing or pawing
behavior would delay the animal, distract him
from the correct solution, or might in some
cases be counted directly as an error, as when
the animal bites or fights the leash.

On Factor IV, the following variables have

loadings of .40 or more:

1. Habit Formation: Time .73

3. Manipulation (String pulling) : Time .46
13. First Barrier Test (First Problem):

Errors .42

8. Leasb Control (In): Trials —.46

Since all tests with high positive loadings on
this factor require the use of vision in locating
objects or in perceiving the relationships among
objects, the factor may be identified as visual
observation. In the habit formation test, the
location of the food dish is changed from trial
to trial, so that the animal must be guided
by visual cues in order to reach the incentive.
In the string-pulling test, the discovery of the
string and its proper utilization to secure the

1955]

Anastasi, et ai: Performance of Dogs on Learning Tests

43

food depend upon visual observation. It will
be noted that the other manipulation test
(variable 2) also has an appreciable positive
weight of .37 on this factor. Similarly, all three
problems of the first barrier test require the
correct visual perception of the spatial relations
between barrier and goal. The first of these
problems has a weight of .43 on this factor.
The second and third have weights of .37
and .30, respectively. Although the three suc-
cessive problems are of increasing difficulty,
it is possible that the benefit to be derived
from visual observation is greatest in the initial
problem, when the animal must first discover
the Umweg type of solution to be followed.

It should also be noted that the motor skills
test has a loading of .32 on this factor. Al-
though this is not a high weight, it is the highest
loading of this test with any factor, all other
loadings being virtually negligible. In this test,
too, the dog must correctly observe the rela-
tion of ramp to food dish. And he must inhibit
any tendency to try to reach the food by jump-
ing directly from the ground to the stacked
boxes, rather than by climbing the ramp. In
this respect the motor skills test might be said
to require that the animal visually recognize
an Umweg-type solution.

The negative weight of Factor IV in the
single measure of leash control (variable 8)
suggests the possibility that the more visually
observant animal is more likely to be dis-
tracted and hence drag, pull, or make similar
errors. Visual distractions of interest to the dog
would probably occur more often in the out-
door course followed in variable 8 than in
stair-climbing (variables 9 and 11). Similarly,
such distractions would not be likely to oper-
ate on the first day of leash training (variable
10) , since the animal’s attention would then
be more completely absorbed by the novelty of
the leash itself.

The variables to be considered in the inter-
pretation of Factor V include:

4. Maze (Second Barrier Test) : Errors .47

7. Cue Response (Discrimination
Apparatus): Time —.50

5. Motivation (Discrimination

Apparatus): Time —.60

6. Cue Response (Discrimination

Apparatus): Trials —.71

The animal which performs well on the
maze is probably one who has good positional
memory for the correct turns. Conversely, the
tendency to take the same path on successive
trials is a handicap on all three variables based
on the discrimination apparatus, since the cor-
rect escape route is varied in random order

from trial to trial. It would thus seem that
Factor V represents persistence of positional
habits.

Breed Differences.— li should be borne in
mind that some of the factors which have been
identified may correspond to characteristic dif-
ferences among the breeds included in the
present sample. Previously published studies on
many of the same dogs employed in the cur-
rent investigation provide evidence of signifi-
cant physiological differences among these
breeds (Fuller, 1951). Observations of the gen-
eral behavior of the dogs have likewise sug-
gested breed differences in such traits as timid-
ity, attraction to human handlers and activity
level (Scott & Charles, 1953). Analyses of
breed differences have also been carried out
on four of the tests included in the present
study, viz.. Maze (Scott & Charles, 1953),
Motivation (Fuller, 1953), Cue Response (Ful-
ler & Scott, 1954) and Leash Control (Fuller
& Scott, 1954). In all of these variables, one
or more significant differences between breeds
were found.

The number of cases available for these
analyses was small, especially in certain breeds.
In the current study, some of these numbers
were further reduced by the necessity of re-
taining only animals with complete records on
all 17 variables. Nevertheless, it may be of
interest to examine the results on breed differ-
ences in the present group. The relevant data
are summarized in Tables 5 and 6, covering
continuous and dichotomized variables, respec-
tively. In Table 5, the results are reported in
the form of median scaled scores for each
breed. It will be recalled that the unit employed
in these scaled scores is .5SD. Table 6 gives
the median raw score, as well as the number
of cases passing and the number failing each
test.

Reference to Tables 5 and 6 suggests that
the Basenjis tend to excel in tasks requiring
independent action and visual observation of
relations, such as habit formation, manipula-
tion, string pulling, the three barrier problems
and the maze. They are especially deficient in
tasks which depend upon responsiveness to
the human handler, such as leash control and
obedience training. And they also do poorly
on the discrimination apparatus tests. It is in-
teresting to note that the inferiority of the
Basenjis on some of these tests is so pro-
nounced that there is no overlapping with the
distributions of high-ranking breeds. This is
true of Basenji-versus-Beagle in the motiva-
tion test (variable 5), and of Basenji-versus-
Wire-haired Fox Terrier in leash control (vari-
able 8). Thus in these two variables the best

44

Zoologica: New York Zoological Society

[40: 3

Table 5. Median Scaled Scores for Each Breed: Continuous Variables*

Variable

Bas.

Bea.

CS

SS

WHT

Bas. X
CS

1.

Habit Formation: Time

5

4

4

2

4

5

2.

Manipulation: Time

6

4.5

4

5

5

4

4.

Maze (Second Barrier Test): Errors

5

6

5

3

2

4

5.

Motivation (Discrim. Appar.): Time

3

7

5

5

6

4

7.

Cue Response (Discrim. Appar.): Time

3

7.5

5

4

5

4

8.

Leash Control (In): Trials

2

6

5

6

5

4

9.

Leash Control (Stairs): Trials

5

7

2.5

5

5

4

10.

Leash Control (In): Initial Errors

4

5.5

6

2

6

4

11.

Leash Control (Stairs) : Initial Errors

4

6.5

4

4

6

6

12.

Motor Skills: Time

4

4

5

3

3

6

13.

First Barrier Test (First Problem): Errors

5

6.5

3

4

6

5

14.

First Barrier Test (Second Problem) : Errors

7

3.5

4

2

5

5

15.

First Barrier Test (Third Problem): Errors

5.5

4.5

3

5

4

5

16.

Obedience (Adjusted Stay Score): Time

3

4.5

5

6

5

4

17.

Obedience: Jumps during Training

4

5

6.5

2

4

4

Number of Cases

16

4

18

5

7

23

* High scores signify better performance. All scores are scaled to an over-all M of 4.5 and cr of 2.

Basenji score falls at least .5SD below the
poorest Beagle or Terrier score, respectively.

Since there were only 4 Beagles in the group,
it is especially hazardous to make any state-
ments about their performance. The exception-
ally high achievement level of these dogs on
several of the tests, however, is very striking.
This is particularly evident in the three discrim-
ination apparatus tests and in leash control. The
Cocker Spaniels excel in obedience training and

leash control, but not in stair climbing, which
yields their poorest score. They are also poor in
cue response (trials), which requires the estab-
lishment of an association between visuo-
auditory cue and open exit; but they do rela-
tively well on the other two discrimination
apparatus tests. In string pulling, they exhibit
the poorest performance in the group; and they
are also below average in the manipulation test.
Any conclusions about the Shetland Sheep

Table 6. Analysis of Breed Differences in Dichotomized Variables

Variable

Bas.

Bea.

CS

SS

WHT

Bas. X
CS

3. Manipulation (String Pulling)

No. failing

2

1

9

2

3

4

No. passing

14

3

9

3

4

19

Median* time in seconds

124

276

477t

199

430

262

6. Cue Response (Discrim. Appar.): Trials

No. failing

5

0

7

1

0

5

No. passing

11

4

11

4

7

18

Median* No. of Trials

113

27

116

30

48

84

Number of Cases

16

4

18

5

7

23

♦ The failures were included in the computations of these medians. The higher the raw scores, the poorer
the performance.

t This median falls midway between a bona fide score and a failure, which was automatically recorded as 480.

1955]

Anastasi, et al.: Performance of Dogs on Learning Tests

45

Dogs must be very tentative because of the small
number of cases. The outstanding finding re-
garding this breed seems to be its relatively
poor performance on many variables.

The Wire-haired Fox Terriers achieve their
best scores on tests which seem to call for con-
fidence in strange situations. These are illus-
trated by the motivation test (which represents
the animal’s first contact with the discrimination
apparatus), initial errors in both leash control
and stair climbing, and the first barrier problem.
They also do well in other tests involving the
discrimination apparatus. On the other hand,
they do particularly poorly on the maze, where
it is reported that they become over-excited and
make many errors (Scott & Charles, 1953).
Little can be concluded regarding the Basenji-
Cocker Spaniel crosses beyond the fact that
they are close to the total group mean on most
measures.

A sharper delineation of breed differences in
behavior characteristics might be obtained
through the application of obverse factor
analysis or Q-technique. For this purpose, it
would be desirable to have scores on a more
extensive set of variables, or at least more part-
scores resulting from further breakdowns of
the present variables. Eventually it would be
advisable to carry out factor analyses similar to
that reported in the present study on each breed
separately. This would, of course, require a
much larger sampling of each breed than is now
available.

Summary and Conclusions

The scores of 73 pedigreed dogs on 17 vari-
ables, most of which were designed as measures
of learning, were submitted to a multiple factor
analysis. The dogs included males and females
of the following breeds: Basenji, Beagle, Cocker
Spaniel, Shetland Sheep Dog, Wire-haired Fox
Terrier and Basenji-Cocker Spaniel crosses. A
centroid analysis of the intercorrelations among
the 17 variables yielded five factors. Following
orthogonal rotation of reference axes, the fac-
tors were interpreted as: activity and impulsive-
ness, docility or responsiveness to a human
trainer, manipulation, visual observation, and
persistence of positional habits. It is pointed out
that one or more of these factors may reflect
breed differences within the population inves-
tigated. An obverse factor analysis would fur-
ther clarify breed differences. If data should
eventually become available on sufficiently large
numbers within each breed, separate factorial
analyses for each breed would be desirable.

Some of the present findings indicate that
tasks which may appear quite similar to the
human experimenter often involve dissimilar

factors for the animal. Moreover, the factorial
composition of the same task may vary consid-
erably at different stages of training, a fact
which was suggested by the earlier results of
Wherry (1939, 1940, 1941) on rats.

Another outstanding finding pertains to the
predominance of bipolar factors. This, too, cor-
roborates earlier factor analyses of animal be-
havior, and sharply contrasts with typical re-
sults on human abilities. In the present study,
negative factor loadings are common and ap-
pear to be psychologically meaningful in terms
of the proposed interpretation of the factors.
Such a finding is probably related to the obvious
intertwining of cognitive with emotional and
motivational factors in animat behavior. On
most of the learning tasks employed in the pres-
ent study, the dogs’ performance reflected
emotional and motivational factors as much as,
or more than, it reflected ability factors. As in
the case of other factor analyses of animal be-
havior, the present findings thus suggest that
the distinction between cognitive and non-
cognitive aspects of behavior is not so sharply
drawn in animals as in humans (cf. Anastasi,
1948). To what extent such a trait differentia-
tion is the product of cultural influences in the
human has not been determined. It is hoped
that it will eventually prove feasible to conduct
longitudinal studies on animals, whose object
will be to alter the subjects’ trait organization by
controlled experiences. Such an approach
should provide the answers to many questions
regarding the nature and organization of psy-
chological traits.

Bibliography

Anastasi, Anne

1948. The Nature of Psychological “Traits.”
Psychol. Rev., 55: 127-138.

1954. Psychological Testing. N. Y.: Macmillan.

Anastasi, Anne, & J. P. Foley, Jr.

1949. Differential Psychology (2nd ed.). N. Y.:
Macmillan.

Billingslea, F. Y.

1942. Intercorrelational Analysis of Certain
Behavior Salients in the Rat. J. comp.
Psychol., 34: 203-211.

Cattell, R. B.

1952. Factor Analysis: An Introduction and
Manual for the Psychologist and Social
Scientist. N. Y.: Harper.

Dunlap, J. W.

1933. The Organization of Learning and Other
Traits in Chickens. Comp. Psychol.
Monogr., 9, no. 4.

46

Zoologica: New York Zoological Society

[40: 3: 1955]

Fruchter, B.

1954. Introduction to Factor Analysis. N. Y.:
Van Nostrand.

Fuller, J. L.

1951. Genetic Variability in Some Physiological
Constants of Dogs. Amer. J. Physiol.,
166; 20-24.

1953. Cross-Sectional and Longitudinal Studies
of Adjustive Behavior in Dogs. Ann. N. Y.
Acad. Sci., 56: 214-224.

Fuller, I. L., & J. P. Scott

1954. Heredity and Learning Ability in Infra-
human Mammals. Eugen. Quart., 1 : 29-43.

Geier, F. M., M. Levin & E. C. Tolman

1941. Individual Differences in Emotionality,
Hypothesis Formation, Vicarious Trial
and Error, and Visual Discrimination
Learning in Rats. Comp. Psychol.
Monogr., 17, no. 3.

Holzinger, K. j., & H. H. Harman

1941. Factor Analysis: A Synthesis of Factorial
Methods. Chicago: Univer. Chicago Press.

McCulloch, T. L.

1935. A Study of the Cognitive Abilities of the
White Rat with Special Reference to
Spearman’s Theory of Two Factors. Duke
Univer. Contrib. Psychol. Theory, 1, no. 2.

McNemar, Q.

1942. On the Number of Factors. Psycho-
metrika, 7: 9-18.

Rethlingshafer, Dorothy

1941. The Learning of a Visual Discrimination
Problem Under Varying Motivating Con-
ditions. J. comp. Psychol., 32: 583-591.

Royce, j. R.

1950a. The Factorial Analysis of Animal Be-
havior. Psychol. Bull., 47: 235-259.

1950b. A Factorial Study of Emotionality in the
Dog. Amer. Psychologist, 5: 263. (Ab-
stract)

1951. A Factorial Study of Emotionality in the
Dog. Unpubl. doctor’s dissert., Univer.
Chicago.

Schmitt, J. R.

1954. A Factor Analysis of Learning in the
Dog. Unpubl. master’s dissert., Fordham
Univer.

Scott, J. P., & Margaret S. Charles

1953. Some Problems of Heredity and Social
Behavior. J. gen. Psychol., 48: 209-230.

Scott, J. P. & 1. L. Fuller

1951. Research on Genetics and Social Behavior
at the Roscoe B. Jackson Memorial Labo-
ratory, 1946-1951— A Progress Report.
J. Hered., 42: 191-197.

Scott, J. P. & J. L. Fuller, (Eds.)

1950. Manual of Dog Testing Teehniques. Bar
Harbor, Maine: Roscoe B. Jackson Me-
morial Laboratory. (Mimeographed)

Searle, L. V.

1947. Application of the “Inverted” Factor
Analysis Technique to the Study of Here-
ditary Behavior Types in Rats. Amer.
Psychologist, 2: 320.

1949. The Organization of Hereditary Maze-
Brightness and Maze-Dullness. Genet.
Psychol. Monogr., 39: 279-325.

Thomson, G. H.

1951. The Factorial Analysis of Human Ability
(5th ed.). Boston: Houghton MiflBin.

Thorndike, R. L.

1935. Organization of Behavior in the Albino
Rat. Genet. Psychol. Monogr., 17: 1-70.

Thurstone, L. L.

1947. Multiple Factor Analysis. Chicago:
Univer. Chicago Press.

1948. Psychological Implications of Factor
Analysis. Amer. Psychologist, 3: 402-408.

Van Steenberg, N. J.

1939. Factors in the Learning Behavior of the
Albino Rat. Psychometrika, 4: 179-200.

Vaughn, C. L.

1937. Factors in Rat Learning. An Analysis of
the Intercorrelations between 34 Vari-
ables. Comp. Psychol. Monogr., 14, no. 3.

Wherry, R. J.

1939. Factorial Analysis of Learning Dynamics
in Animals. J. comp. Psychol., 28: 263-
272.

1940. A Test by Factorial Analysis of Honzik’s
Exteroceptive Data. J. comp. Psychol.,
29: 75-95.

1941. Determination of the Specific Components
of Maze Ability for Tryon’s Bright and
Dull Rats by Means of Factorial Analysis.
J. comp. Psychol., 32: 237-252.

4

The Chemical Nature of Holothurin, a Toxic Principle
from the Sea-cucumber (Echinodermata: Holothurioidea )

Ross F. Nigrelli, J. D. Chanley, Stella K. Kohn & Harry Sobotka
New York Aquarium, New York Zoological Society, and
Department of Chemistry, The Mount Sinai Hospital, New York

Introduction

Atoxic principle in some Holothurioidea
and its effects on fish and other small
animals have been described in the lit-
erature (Saville-Kent, 1893; Frey, 1951). More
recently, Nigrelli (1952) and Nigrelli & Zahl
(1952) have reported on certain chemical and
biological characteristics of a water soluble, heat
stable substance isolated from the Bahamian^
sea-cucumber, Actinopyga agassizi Selenka.
This material, called Holothurin (Nigrelli,
1952), has been found to be toxic to a wide
range of plants and animals, and also appears to
have some anti-tumorous and other pharma-
cological properties^.

Holothurin is a new variation on the theme
of steroid compounds. It consists of several
closely related aglycones or genins of steroid
structure and a chain of several monosacchar-
ides. It is soluble in water and produces a soapy
foam in high dilution. Insofar as can be ascer-
tained, this is the first known saponin of animal
origin.

Chemical Studies

The starting material of these experiments
consists of sun-dried Cuvierian glands of Acti-
nopyga agassizi. It forms pinkish flakes and is
almost completely soluble in water. The mate-
rial is first extracted in 95% ethanol and the
minor fraction is discarded; it is then dissolved

^We wish to thank Dr. C. M. Breder, Jr., Director of
the Lemer Marine Laboratory of the American Museum
of Natural History, Bimini, B.W.I., for the use of the
laboratory facilities in collecting the Holothurin mate-
rial used in these studies.

^We wish to thank Dr. Sophie Jakowska, College of
Mount Saint Vincent, New York City, and Mr. Herman
Baker, Department of Chemistry, The Mount Sinai
Hospital, for their assistance in testing the pharma-
cological action of some of the products derived in
these studies.

in 50% ethanol, leaving behind cellular mate-
rial and other debris. The resulting product,
comprising 60% of the starting material, is neu-
tral, practically free of nitrogen and exerts no
reducing power for sugar reagents; it shows no
absorption in the ultraviolet region of the spec-
trum and its specific rotation is about [«]d ~
—19°. Upon hydrolysis with normal hydro-
chloric acid for three hours, it yields about 60%
of its weight as a mixture of water soluble re-
ducing sugars and 40% as a mixture of water
insoluble aglycones.

The sugar solution, when chromatographed
on paper with butanol-ethanol-water at
pH=3.7, appears to consist of three different
sugars, the fastest moving of which is probably
rhamnose with Rf=0.31. The next one with
R(=ca. 0.19 gives pentose reactions and is pre-
sumably identical with xylose, while the slowest
one with R[=0.14 is glucose and may be re-
moved by fermentation with yeast. Their quan-
tities are in the approximate ratio of 2: 1 : 1. This
mixture of sugars is reminiscent of the carbo-
hydrate portion of cardiac glycosides and other
saponins of plant origin.

The water insoluble moiety resulting from
acid hydrolysis melts, after softening at 200°,
over a wide range from 230° to 250°. This mix-
ture of aglycones (1.0 g.) may be fractionated
by repeated crystallization from acetone, ether
and other organic solvents, and yields four frac-
tions in the following quantities: 130 mg. genin
A, 70 mg. genin B, 90 mg. genin C, and 420
mg. fraction F. These fractions melt respectively
at 256-260°, 235-240°, and 195-220°; fraction
F is an oily substance, soluble in petrolic ether
and partially soluble in aqueous alkali. The
optical rotations are as follows: substance A
[a]n=-27° (in ethanol), substance B
[a]D=-86°, and substance C [a]D=-124°
(both in ethyl acetate). Fraction F presumably

47

48

Zoologica: New York Zoological Society

[40: 4: 1955]

consists of fatty acids and esters, and its weak
optical rotation of [a]c=-8° is probably due
to contamination with some aglycone. Whether
all of fraction F forms an intrinsic part of the
molecule or is merely carried along by virtue
of the detergency of the saponins is undecided.
Substances A, B, C, as well as their acetyl deriv-
atives, show ultraviolet absorption with Xmax =
244 m^ and an inflection at 238 mju,. Extinction
is highest, E/g^= 278, for acetyl-A, and E
194 for compound A itself. The acetyl deriv-
ative of fraction C shows a secondary maxi-
mum at = 300 m/i. with Ej*^ = 19.

On passing an aqueous solution of Holothurin
prior to hydrolysis through a mixed bed ion-
exchange resin column, all sodium and chloride
ions may be removed. During this operation
about 20% of the original solids, comprising
electrolyte as well as much of the fatty mate-
rial, remain on the column. The eluate is acidic;
its optical rotation is [«] d ~ —23°. The tox-
icity of the material seems to be diminished
by this operation. This product has no reducing
power and yields on acid hydrolysis the water
soluble sugars and about 33% of its weight
of the mixed neutral aglycones. Optical rota-
tion [a] D — —77°; elementary analysis: C,
72.27; H, 9.72; calculated for C27H44O5 (aver-
age composition) : C, 72.29; H, 9.88. M.p.
230-250°.

Fractionation of the aglycones yields a crys-
talline substance of m.p. 258-262° and [a]*D =
-19°, which is identical with substance A above.
Elementary analysis: C, 69.87; H, 9.06; cal-
culated for C27H42O6; C, 70.10; H, 9.15. A
second fraction, melting at 228-230°, parallels
fraction B above in its properties. Elementary
analysis: C, 70.90; H, 9.09 and inorganic resi-
due 4.24%. These observations indicate that the
derived aglycones are closely related one to an-
other and consist of a tetracyclic steroid skeleton
with two conjugated double bonds and five to
six oxygen functions mostly of hydroxyl nature.

Discussion

The carbohydrate components, presumably
linked to a hydroxyl group on C,, are easily
hydrolyzed by dilute acid. At the same time a

double bond (C:C or C:0) arises in conjugation
with a previously present double bond. The
wavelength of the resulting absorption maxi-
mum indicates that the two double bonds are
situated in adjoining rings, e.g. in position
C^:C5 (or Cg:C^^) and Cg:Cj. A highly substi-
tuted aijS-unsaturated ketonic structure cannot
be excluded as an alternative. This unsaturated
system may be independent of the glycosidic
linkage, as it once arose on passage through
the mixed bed column, where the glycoside por-
tion of the molecule remains intact. The new
double bond may be formed by saponification
of an ester grouping in angular position and loss
of the resulting free tertiary hydroxyl group by
dehydration, or it may have arisen from the
cleavage of an enol ether and concomitant mi-
gration of a double bond.

Summary

Holothurin, the toxic principle in the
Cuvierian organ of Actinopyga agassizi, a
Bahamian sea-cucumber, is a steroid saponin.
Acid hydrolysis yields a mixture of levorotatory
genins with a conjugated system of two double
bonds and three sugars— rhamnose, xylose and
glucose. Some of the chemical features of the
aglycones are discussed.

This is the first known steroid saponin of
animal origin.

References

Frey, David G.

1951. The Use of Sea Cucumbers in Poisoning
Fishes. Copeia, 2: 175-176.

Nigrelli, R. F.

1952. The Effects of Holothurin on Fish, and
Mice with Sarcoma 180. Zoologica, 37:
89-90.

Nigrelli, R. F. and Zahl, P. A.

1952. Some Biological Characteristics of Holo-
thurin. Proc. Soc. Exp. Biol, and Med.,
81: 379-380.

Saville-Kent, W.

1893. The Great Barrier Reef of Australia, Lon-
don, p. 293.

5

The Effects of Holothurin, a Steroid Saponin of Animal Origin, on
Krebs-2 Ascites Tumors in Swiss Mice

T. D. Sullivan, S.S.E.i, K. T. Ladue & Ross F. Nigrelli
Department of Biology, St. Michael’s College, Winooski, Vermont, and
New York Aquarium, New York Zoological Society

Introduction

The preparation and general character-
istics of the water soluble, thermostable
factor, Holothurin, from the Cuvierian
organ of the sea-cucumber, Actinopyga agas-
sizi Selenka, have been described by Nigrelli
(1952)2 and by Nigrelli, Chanley, Kohn &
Sobokta (1955). Nigrelli & Zahl (1952) re-
ported that Holothurin inhibited growth of
certain Protozoa and that in vitro treatment of
Sarcoma 180 cells with this substance markedly
reduced the subsequent growth of these tumor
cells after inoculation into Swiss mice. The
present preliminary report is primarily con-
cerned with the in vitro and in vivo effects of
Holothurin on the Krebs-2 ascites tumor in
Swiss mice.

Materials and Methods
The Krebs-2 ascites tumor, carried in Swiss
mice and in a breeding colony of Highline, ICR
Swiss mice, was obtained in 1952 from Dr. T. S.
Hauschka. The tumor has been maintained by
weekly intraperitoneal inoculations of 0. 1-0.2
ml. of ascitic fluid into 10-12 mice. All animals
in these experiments were obtained from the
original breeding mice by cousin matings, and
weighed 18-25 grams at the beginning of the
experiments.

The method recommended by Goldberg, Klein
& Klein ( 1950) for counting the tumor cells and

^This investigation was supported in part by grants
C-1745 and C-1745-C from the National Cancer In-
stitute of the National Institutes of Health, Public
Health Service. We wish to thank Mr. Richard
DiLorenzo and Mr. Paul Lachance for technical assist-
ance.

^We wish to thank Dr. C. M. Breder, Jr., Director of
the Lemer Marine Laboratory of the American Museum
of Natural History, Bimini, B.W.I., for use of the labora-
tory facilities in the collection and preparation of
Holothurin.

for the inoculation procedure was used. The
donors of ascitic fluid had carried the tumor for
at least 7 days but not more than 13 days. A
uniform inoculum of 2 X 10® tumor cells
suspended in sterile saline, as suggested by
Sugiura (1953), was injected in all experi-
mental animals. No remission has occurred in
more than 2,000 mice inoculated with 1 to 44
X 10® tumor cells.

Holothurin was dissolved in physiological
saline and autoclaved for 30 minutes at 20 lbs.
pressure. The method of Nigrelli & Zahl (1952)
was used for treating the tumor cells in vitro;
varying amounts of Holothurin were added to
tubes containing 20 X 10® tumor cells diluted
with sterile saline. The tubes were incubated
at 99 °F and shaken for 1 hour. Control tubes
were similarly treated.

In the in vivo experiments 0.01, 0.05, 0.1 and
0.2 mg. of Holothurin were injected daily or
spaced over 7-21 days. Repeated injections were
made intraperitoneally on alternate sides of the
abdomen. Controls with and without tumor cells
were maintained simultaneously. Fifty-six mice
were also inoculated with tumor cells in the sub-
cutaneous areas of the axillary and inguinal re-
gions. These did not survive for more than two
and one-half months.

Results

Toxicity of Intraperitoneal in-

jection of 0.2 mg. of Holothurin in sterile saline
was lethal in 48 hrs. for 6 female mice; injection
of 0. 1 mg. into 1 2 female mice caused no deaths.
The safe upper limit for injection of Holothurin
was taken as 0.1 mg.

In Vitro Experiments.— No deaths were ob-
served in 60 days in female mice inoculated
with 2 X 10® tumor cells treated with 0.1 mg.
of Holothurin. These mice produced normal
litters when mated; the offspring, however,

49

Table 1. Standardization of Mean Survival Times and Body Weight Gains of Swiss Mice Carrying the Krebs-2 Ascites Tumor*

50

Zoologica: New York Zoological Society

G

CS V5

o> ^

S G

- H

«4_i a

O ,>

G >

^ r2

^6
a .

2

O

a

W)

C3

o

M-t

o

d

Z

u

o

Z

X

c/5

o

Z

Q

Q

^ B

s « g

o >,.2

m o
*3 O

qo a

d
o> O

s

ja

• 2 d

P °
^ IS 5

^ O ''3

S 'u’S

s.a 8

ss «

^ Ui

cn O U
MH M

0 1>

01 rn

! ^

:Q

a

V C
00 °

.9 'U "a
&0.2 3
•« S 2

1) ^ ri

o

03

c2 o

^ >'9

<« ce O

° Q -

. SJ

^ §3
S .2 ”

ss

(4-1 U

O ,o
<5 VO

Q 00

1) £
oot:

n j3
CC ^

oi Q

E

00

9

00

o

.9

u

U

e(

O

S

D

H

n

u

Bi

X

n

B4

o

S

D

i-l

D

O

o

z

►H

z

<

o

z

s

o

u

u

c/2

z

o

o

u

o

33

b

O

cn

H

CJ

w

b

b

W

ri

u

>-)

H

i> .2

rt 3

2^3

o

CO

S

C= 2-S

O c 03

d *d
^ H tj

« <4-1 O

o a

a >»;:;

CO cO S
1) O 1/

cS *

>

•r* CO

I ^

c/) Q

a 'IT

Sb

.9

Im

I CO

"ij

K 8

■S'?

6

Z

9

■S ’'s

O "

1-1

h-H

ob

s

X

4>

CO

u

%

o

z

v-i

9

wn

»o

<N

<N

<S m

VO

VO

o CO

fO 0\

o vd

© 00

m ts

rs ^

<s ^

CN ^

Sp3

0/Dio

fO

<s

VS

VO

O

(N

0\

1-4 00

^ <y\

O 00

ri os

^ Os

(S ^

<s ^

fS ^

<s ^

cs ^

Sc^

Sps

Sc2

Sc2

On

^7

o

00

Vi

Vi Vi

VO ^

fn O

Vi ^

On

1

^ 1-H

00

1-M 1—1

<S Vi

cn 00

Sc2

Sfti

Sc2

Sc2

Xfi

C5 to
•2 ^
o -o
.«00
c

CO

d 5«
.2 ^
o

•5i,oo

c

CO

CO

O ^

5 -

03

^ *o
o

0> VO

PU

CO

CO

O ^

c ^

c«

— -O
o

(U 1-4

•d<^

CO

ob

ob

ob

ob

’o

a

a

a

a

d

«-H

Vi

o

o

o

o

O

1-H

U

d

d

d

d

[40: 5

*M is the mean cJay of death or mean weight of mice; R is the range of variation

Table 3. Effects of Holothurin on Swiss Mice Carrying an Inoculum of 2 X 10^ Krebs-2 Ascites Tumor Cells

1955]

Sullivan, et al.: Effects of Holothurin on Krebs-2 Ascites Tumor

51

H u.i
^ ^ 3

r- °-2

O

^ H g

c

. I-H

c ST'—
S Q

3 Q
CO W

§ i

D .3

5 .

o §

o s
o

HM (1)

o

-a

o ^
"o o

ffi’a

X

o

CO

>>

CQ

c

o

u

c

o

U

m O
r4 <N

p lAi
<N

<— I

fN <N

X-l
tr, <N

O
(M n

^ V-> ^

^ »ri
^ r- (S’-"

O cQ
w *o
u O

•S

CUD

E

U<

o

ftb

o

Tf

s

.a

a>

I

a

>

t

s

o

Z

showed no resistance to the tumor. In a control
group of 6 mice, all were dead by the 18th day,
with a mean survival time of 12.5 days.

Two deaths occurred in a group of 12 female
mice that were given 2X10® tumor cells treated
with 0.05 mg. of Holothurin. The amount of ma-
terial needed to irreversibly inactivate 20 X 10®
tumor cells at 99°F. for 1 hr. is between 0.05
and 0.1 mg.

In Vivo Experiments.— The efficacy of Holo-
thurin as a tumor growth inhibitor was judged
on the basis of a comparison of the body weight
gain on the 1 0th day after tumor cell inoculation
and of the mean survival time in treated and
untreated mice.

A group of 186 female mice (31 cages of
6 mice/cage) was inoculated with the standard
amount of tumor cells. All mice died between
the 6th and 26th day after inoculation. The re-
sults, shown in Table 1, were used as the base-
line for comparison with the individual controls
and experimental mice. Another group of 120
animals gave similar results. Two mice in this
group, however, survived past the 24th day,
dying on the 33rd and 36th day after inoculation.

A series of 7 injections of 0.01 mg. of Holo-
thurin started on the day after inoculation and
continued for 7 days gave an apparent reduction
of the mean survival time (Table 2). However,
this mean survival time (8.5 days) falls very
close to the lower limit of variation found among
the 186 control mice (see Table 1). It is impos-
sible, therefore, to conclude that there has been
a true reduction in mean survival time with the
amount of Holothurin used in this series.

A similar series of injections of 0.05 mg. of
Holothurin caused a slight increase in mean
survival time and a marked reduction in gains
in body weight on the 10th day after inoculation
(Table 2).

The daily injection of 0. 1 mg. Holothurin was
found to be lethal. However, a series of 7 injec-
tions given on alternate days resulted in some
increase in mean survival time and a marked
reduction of mean body weight on the 10th day.
A series of 10 injections of 0.1 mg. spaced over
21 days produced a considerable increase in
mean survival time and a very marked reduction
of mean body weight gain on the 10th day after
inoculation (Table 2).

The results from another group of 24 female
mice inoculated with the standard amount of
tumor cells and given a series of 10 injections
of 0.1 mg. of Holothurin over a period of 20
days are shown in Table 3. The increase in
survival time and the marked reduction in gain
of mean body weight on the 10th day is clearly
evident when compared to the experimental and

52

Zoologica: New York Zoological Society

[40: 5: 1955]

baseline controls. Four of the 24 treated mice
showed no observable growth during a period
lasting over 6 months.

Discussion

The unequivocal validity of a test based on
1 86 animals is, of course, open to question. The
repetition of the test on 120 animals gave nearly
identical values, but there were two control
animals that survived past the 24th day after
inoculation, dying on the 33rd and 36th day.
There have been no remissions, however, in
such inoculated animals, nor in more than 2,000
mice inoculated with Krebs-2 ascites tumor cells
in amounts from 1 to 44 X 10®.

The four remissions observed cannot be en-
tirely explained by accidental subcutaneous
inoculations. Of a group of 56 mice inoculated
subcutaneously in the axillary and inguinal re-
gions, none were living after two and one-half
months.

Summary

Injection of 0.2 mg. of Holothurin in sterile
saline into female Swiss mice (Highline, ICR)
is lethal within 48 hrs.

The treatment of 20 X 10® Krebs-2 ascites
tumor cells in vitro with 0.1 mg. of Holothurin
by shaking such suspensions for 1 hr. at 99°F.
results in inactivation of the tumor cells. The
inoculation of such treated cells into normal
mice produced no observable growth during a
period of 60 days.

The daily injection of 0.01 mg. of Holothurin
for 7 days, beginning the day after inoculation
of the tumor cells, had no effect on the progres-
sive growth of the tumor.

The daily injection of 0.05 mg. of Holothurin

for 7 days, beginning the day after inoculation
of tumor cells, brings about an increase in sur-
vival time and a marked reduction of mean body
weight gain on the 10th day.

The spacing of one day between injections of
0.1 mg. of Holothurin in a series of 7-10 injec-
tions results in a marked increase in mean sur-
vival time and reduction in mean body weight
gain on the 10th day after inoculation.

Four remissions of the tumor of more than
6 months duration occurred in a group of 24
mice treated with 0.1 mg. of Holothurin in a
series of 10 injections over a period of 20 days.

References

Goldberg, L., Eva Klein & G. Klein

1950. The Nucleic Acid Content of Mouse
Ascites Tumor Cells. Exp. Cell Research,
1: 543-570.

Nigrelli, Ross F.

1952. The Effects of Holothurin on Fish, and
Mice With Sarcoma 180. Zoologica, 37:
89-90.

Nigrelli, R. F., J. D. Chanley, Stella Kohn &

Harry Sobotka

1955. The Chemical Nature of Holothurin, a
Toxic Principle from the Sea-cucumber
(Echinodermata: Holothurioidea). Zool-
ogica, 40: 47-48.

Nigrelli, Ross & Paul Zahl

1952. Some Biological Characteristics of Holo-
thurin. Proc. Soc. Exp. Biol. & Med., 81:
379-380.

SUGIURA, K.

1953. Effect of Various Compounds on the Ehr-
lich Ascites Carcinoma. Cancer Research,
13: 431-441.

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

30 East Fortieth Street, New York 16, N. Y.

PUBLICATION OFFICE

The Zoological Park, New York 60, N. Y.

OFFICERS
PRESIDENT
Fairfield Osborn

SECRETARY
Harold J. O’Connell

VICE-PRESIDENTS TREASURER

Alfred Ely David H. McAlpin

Laurance S. Rockefeller
Donald T. Carlisle

ASSISTANT TREASURER
Percy Chubb, 2nd

SCIENTIFIC STAFF: Zoological Park and Aquarium

John Tee-Van Director

Leonard J. Goss Assistant Director

ZOOLOGICAL PARK

Grace Davall Assistant Curator,

Mammals and Birds

James A. Oliver Curator of Reptiles

Leonard J. Goss Veterinarian

Charles P. Gandal. . .Assistant Veterinarian
Lee S. Crandall General Curator

Emeritus

William Beebe Honorary Curator,

Birds

AQUARIUM

Christopher W. Coates . Curator & Aquarist

James W. Atz. Assistant Curator

Ross F. Nigrelli Pathologist

Myron Gordon Geneticist

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . .Research Associate

in Histology

Homer W, Smith Research Associate

in Physiology

DEPARTMENT OF TROPICAL
RESEARCH

William Beebe Director Emeritus

Jocelyn Crane Assistant Director

Henry Fleming Entomologist

Rosemary Kenedy Field Assistant

John Tee-Van Associate

William K. Gregory Associate

AFFILIATES

C. R. Carpenter Co-ordinator, Animal

Behavior Research Programs

L. Floyd Clarke Curator,

Jackson Hole Research Station

SCIENTIFIC ADVISORY COUNCIL
A. Raymond Dochez Caryl P. Haskins
Alfred E. Emerson K. S. Lashley
W. A. Hagan John S. Nicholas

EDITORIAL COMMITTEE

GENERAL Fairfield Osborn, Chairman

William Bridges Editor & Curator, James W. Atz Lee S. Crandall

Publications William Beebe Leonard J. Goss

Sam Dunton Photographer William Bridges James A. Oliver

Henry M. Lester. . .Photographic Consultant Christopher W. Coates John Tee-Van

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 40 • PART 2 • AUGUST 19, 1955 • NUMBERS 6 TO 9

ro -

r ,• i

PUBLISHED BY THE SOCIETY
Tie ZOOLOGICAL PARK, New York

Contents

PAGE

6. Regeneration of Melanomas in Fishes. By Recai Ermin & Myron

Gordon. Plates I-IV; Text-figures 1-45.. 53

7. Further Notes on the Pigmentary Behavior of Chaetodipterm in Reference

to Background and Water Transparency. By C. M, Breder, Jr., & Pris-
cilla Rasquin. Plates I & II 85

8 . Special Features of Visibility Reduction in Flatfishes. By C. M. Breder, Jr.

Plates I-III 91

9. Further Chemical Analysis of Holothurin, the Saponin-like Steroid from
the Sea-cucumber. By J. D. Chanley, Stella K. Kohn, Ross F. Nigrelli

& Harry Sobotka 99

6

Regeneration of Melanomas in Fishes^

Recai Ermin & Myron Gordon

Zoological Institute of the University of Istanbul, Turkey, and the Genetics Laboratory
of the Aquarium, New York Zoological Society

(Plates I-IV; Text-figures 1-45)

COMPARATIVE pathologists who have
studied the black pigment cell tumors of
sharks, fishes, amphibians, reptiles, birds
and mammals, have discovered that the mel-
anocyte is the common cell type of the mela-
nomas of vertebrate animals, including those of
man (Gordon, 1951b). The living melanocytes
from the melanomas of man, mouse and fish,
when studied by tissue culture methods, have
been shown to be similar morphologically
(Grand, Gordon & Cameron, 1941, and Attardi
& Moro, 1953) . A comparative study of the
enzyme reactions of melanomatous tissues from
mammals and fishes also revealed that both are
quite similar with respect to tyrosinase activity
(Fitzpatrick, Lerner, Calkins & Summerson,
1950).

An outstanding peculiarity of the fish mel-
anoma, in contrast with human melanoma, is
the presence of macromelanophores. The large
size of these melanophores and their capacity
for drastic changes in the shape of their pseudo-
podial processes precludes their being confused
with the much smaller melanocytes or the mel-
anin-laden macrophages. From previous studies
it was not possible to state whether macromel-
anophores and melanocytes were related onto-
genetically, but it was Imown that melanomas
arose in %brid fishes only when macromelano-
phores were present.

Regeneration studies make possible the retrac-
ing of the progressive growth stages in the de-
velopment of a melanoma. It will be shown that,
in part, they confirm the observations of Gordon
& Smith (1938) who demonstrated that in the

^Supported by grants from the National Cancer
Institute of the National Institutes of Health, U. S.
Public Health Service, and the Anna Fuller Fund. The
work was aided by the laboratory facilities of the
American Museum of Natural History, New York 24,
New York.

earliest stages of this neoplasm, macromelano-
phores in their abnormal growth replace the
normal tissues of the corium almost completely
and create a state of melanosis. In later stages,
Gordon & Smith showed that the large black
pigment cells invade the deeper body areas, mov-
ing along the fascial tissues of the muscles. The
macromelanophores eventually surround the
muscle fibres and destroy them. Nodular lesions
leading to the formation of melanomas usually
arise in these primary zones of diffuse melanosis.
During the early stages in the growth of the
melanoma, the macromelanophores are the prin-
cipal cells involved. As the growth of the mel-
anoma progresses, the macromelanophores ap-
parently redifferentiate into cells which have the
morphological properties of melanocytes.

Study of these suspected cellular transforma-
tions of normal macromelanophores to malig-
nant melanocytes, made possible by histological
examination of the progressive growth of mel-
anomas, obviously required additional study by
more direct methods. Observations of regenerat-
ing tissues following the amputation of mel-
anomas in the dorsal fin have made it possible to
retrace the ontogenetic relationships between
macromelanophores and melanocytes. By ex-
periments that will be described, the close as-
sociation of these pigment-forming cells has
been confirmed. In an independent study, Mar-
cus & Gordon (1954) followed the fate of mel-
anoma fragments after they were transplanted
into clear, transparent normal areas of the fish’s
skin. They, too, found that some melanoc)des
are capable of transforming into macromelano-
phores and vice versa.

Material and Methods

Fifty-three xiphophorin fishes were studied;
they consisted of young and adult platyfish,
Xiphophorus maculatus, and platyfish-swordtail

53

is 1

54

Zoologica: New York Zoological Society

[40: 6

hybrids, X. maculatus—X. helleri. Their histories
are listed in Table 1.

The fishes chosen for the study of the effects
of amputation of normally and atypically pig-
mented dorsal fins were obtained from genetical-
ly known stocks maintained at the Genetics
Laboratory of the Aquarium, New York Zoolog-
ical Society, by methods described by Gordon
(1950a). The most important pigmentary pat-
terns for these studies were those that are pro-
duced in the dorsal fins by two types of spe-
cialized pigment-forming cells, the micromel-
anophores and macromelanophores. Various
combinations of these two kinds of pigment ef-
fector cells in the dorsal fins of fishes were avail-
able, so that it was possible to compare the re-
sults of amputation and regeneration of normal-
ly pigmented dorsal fins with those that were
either in a state of melanosis or had melanomas.
In those that had melanomas, it was also possible
to compare those that had the typical black
coloring with those that had amelanotic mel-
anomas.

The fishes were first aresthetized with a 1 : 2000
solution of MS 222 (Triacine Methanesulfonate
produced by Sandoz) and placed on a wet cotton
pad. The dorsal fins were removed close to the
body, using iridectomy scissors. The regenerat-
ing fins were observed at intervals under the
binocular microscope and drawings were made
of the progressive regrowths.

For histological study, tissues amputated and
regenerated were fixed in Bouin’s solution, de-
calcified in nitric acid-phloroglucin, dehydrated
in dioxane, embedded in paraffin and sectioned
at 8 to 10 micra. Sections were studied either
with Mayer’s hematoxylin (hemalum) and
eosin, or with a modification of the Masson’s

trichrome stain. In order to obtain the cellular
details of extremely black tumors, the sections
were first treated with celloidin and then
bleached in a solution of potassium chlorate
crystals in hydrochloric acid and 70% alcohol.

Regeneration of Dorsal Fins in Platyfish

The spotted dorsal fins of two female and
three male platyfish were amputated. Table 1,
No. 1. Each fin had several groups of normal
macromelanophores as well as many micromel-
anophores. Text-fig. 1. Within 24 hours wound
healing began and within five to seven days the
blastema, about 0.5 mm. in height, was formed.
The blastema contained some pigment granules
and the stumps of the fin rays, each of which
was composed of two lepidotrichial elements.
Within two weeks the blastema grew to about 1
to 1.5 mm. in height. Micromelanophores ap-
peared along the regenerating fin rays, in the fin
membranes and on the fin’s distal edge. Text-fig.
lb. Within one to one and a half months the
fin grew to 2 to 3 mm.. Text-figs. Ic, Id. In two
animals some macromelanophores migrated in-
to the base of the regenerating fin from the
region ventral to the dorsal fin. Text-fig. Id. The
fin rays began to bifurcate at their tips. At three
to four months in four out of five fish, the new
dorsal fin reached its original size, 5 to 6 mm.
in height. Only eight fin rays regenerated in one
fish to replace the original 10, Text-fig. If. In
one animal originally having several spots on
its dorsal, within eight months, the macromel-
anophores formed one big spot at the posterior
base of the regenerated fin. Text-figs. Ig, Ih. In
another fish the large spots did not form at all.
Usually macromelanophores appeared in the

Table 1.

Genetic

Histories of Fishes Used in Regeneration Studies of Normal and Atypical
Pigment Cell Growths in the Dorsal Fins^

Group

No.

Pedigree

Genetic Constitution

Melanophores

1.

5

163

-t- -b Spotted-dorsal

Micro. -F Macro.

2.

2

306

+ Co + Comet

Micro.

3.

3

306

+ Co E Wagtail

Micro.

4.

5

306

Sd Co -f- Spotted-dorsal, Comet

Micro, -b Macro.

5.

3

306

Sd Co E Spotted-dorsal, Wagtail

Micro, -b Macro.

6.

2

311,316

+ -f Unpatterned

Micro.

7.

18

311,316

Sd Melanosis, variable (1, 2,

3)2 Micro. -F Macro.

8.

10

311,316

Sd -F Melanomas

Micro. -F Macro.

9.

5

311,316

Sd i Amelanotic melanoma

Not evident

^Summary of 53 histories. Groups 1 to 5 represent Xiphophorus maculatus, 6 to 9 represent X. maculatiis—
X. helleri hybrids. The individual records are presented in Tables 2 to 6.

^The degree of melanosis is indicated arbitrarily: 1, severe; 2, intermediate; 3, slight; see text for criteria.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

55

Text-fig. 1. Regeneration of the spotted dorsal fin of a platyfish, genetically Sd co e. Macromelano-
phores produce the spotted pattern in the dorsal fin. Micromelanophores are also present in the fins and
on the body. 10 X. a— Before amputation, b— Within 13 days following amputation, c— Within 40 days,
d— Within 63 days, e— Within 90 days, f— Within 120 days, g— Dorsal fin of another fish of the same
stock before amputation, h— The regenerate of the fin shown in figure g within 8 months following
amputation.

56

Zoologica: New York Zoological Society

[40: 6

Table 2. Regeneration, after Amputation, of Dorsal Fins and the Spotted Patterns in Platyfish^

Regeneration

No.

Pattern

In days

Of Fin

Of Pigmentation

1.

Spotted-dorsal

240

Complete

Similar

2.

Spotted-dorsal

240

Complete

Lacking

3.

Spotted-dorsal

229

Complete

Lacking

4.

Spotted-dorsal

188

Complete

Lacking

5.

Spotted-dorsal

62

Complete

Lacking

^Xiphophoriis maculatus of pedigree number 163, see Table 1, item 1, for genetic constitution.

regenerated tissues, if at all, much later than the
micromelanophores.

The dorsal fins of two comet platyfish with
micromelanophores only, Table 1, No. 2, were
amputated for purpose of comparison. The
comet pattern is made up of closely grouped
micromelanophores and is confined to the
caudal fin only. Text-fig. 2. Within a week a
blastema developed, 0.3 to 0.5 mm. in height,
which contained fin rays and some micromel-
anophores at the base. Text-fig. 2b. The number
of micromelanophores increased along the fin
rays and they covered the base of the fin, Text-
fig. 2c; they reached the distal edge of the new
fin in two weeks. Within a month, the height of
the regenerate was 3.5 to 4 mm., and some of
the regenerating fin rays had bifurcated. Text-
fig. 2d. Within two to three months the new fin
reached its original height, 5 to 6 mm., and
pigmentary pattern. Text-fig. 2e. Within four
to seven months there were no further ap-

preciable changes except for the secondary bifur-
cation of some fin rays.

The dorsal fins of three platyfish with the
wagtail pattern were amputated next. Table 1,
No. 3. The dorsal fins of the wagtail platyfish
had micromelanophores only, but they were
much more numerous in all the fins than in the
fins of the comet platyfish. The fins of the wagtail
regenerated normally; however, the rate of the
micromelanophore restoration and the number of
regenerating micromelanophores were greater
than in the comet platyfish.

The heavily spotted dorsal fins of five platyfish
were amputated; they had both micro- and mac-
romelanophores in the dorsal fin and a comet
pattern in the tail. Table 1, No. 4. One specimen
had macromelanophores that had spread down
from the dorsal fin to both sides of the body.
There they had infiltrated the underlying tissues,
creating a condition of melanosis. Within a week
following amputation of the dorsal fins, blas-

Table 3. Regeneration, after Amputation, of Dorsal Fins and Pigmentary Patterns in Platyfish'^

Regeneration

No.

Pattern

In days

Of Fin

Of Pigmentation

6.

Comet*

233

Complete

Complete

7.

Without Comet

23

Incomplete

Fixed®

8.

WagtaiF

190

Complete

Complete

9.

Wagtail^

7

Incomplete

Fixed®

10.

Wagtail^

2

Incomplete

Fixed®

11.

Spotted-dorsal, Comet

180

Complete

Incomplete

12.

Spotted-dorsal, Comet

90

Incomplete

Incomplete

13.

Spotted-dorsal, Comet

1

Incomplete

Fixed®

14.

Spotted-dorsal, Comet

21

Incomplete

Incomplete

15.

Spotted-dorsal, Comet

15

Incomplete

Fixed®

16.

Spotted-dorsal, Wagtail

180

Complete

Incomplete

17.

Spotted-dorsal, Wagtail

104

Complete

Incomplete

18.

Spotted-dorsal, Wagtail

55

Complete

Incomplete

^Xiphophorus maculatus of pedigree number 306, see Table 1, items 2, 3, 4, 5.
^Cornet and Wagtail patterns are made up of micromelanophores only.
^Sacrificed for histological study.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

57

2

b

c

Text-fig. 2. Regeneration of the unspotted dorsal fin in a comet platyfish, genetically sd Co e. No
macromelanophores are present in this specimen. 10 X. a— Before amputation, b— Within 1 week follow-
ing amputation, c— Within 2 weeks, d— Within 30 days, e— Within 60 days.

temas developed that were 0.3 to 0.5 mm. and
contained some pigment particles and fin ray
stumps. Within two weeks the blastemas were
0.5 to 1 mm. in height and contained many
micromelanophores and four to five macromel-
anophores that had migrated from the base of
the fin. Within two to three months the new fins
of four of the platyfish were as high as the
originals; one of them regenerated in slightly
more than three months. Within two months

the macromelanophores in the fins reformed the
typical spotted dorsal pattern. Almost no micro-
melanophores were found in the spotted region,
whereas many of them were present in the
distal region of the fin. As time went on, more
macromelanophores continued to migrate into
the new fins, increasing the size of the spots, but
even at six months the size of the macromelano-
phore spottings in the dorsal fin was not as large
as in the originals.

58

Zoologica: New York Zoological Society

[40: 6

Text-fig. 3. Regeneration of spotted dorsal fin of a wagtail platyfish, genetically Sd Co E. This type
of platyfish has many more micromelanophores in its fins than the Sd co e fish shown in Text-figure 1.
10 X. a-Before amputation, b— Within 10 days following amputation, c— Within 33 days, d— Within
62 days, e— Within 120 days, f— Within 150 days.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

59

The heavily spotted dorsal fins of three wag-
tail platyfish, Table 1, No. 5, were amputated. A
strong melanosis produced by both micro- and
macromelanophores existed in the dorsal fins
Text-fig. 3a. Within a week a blastema formed,

0. 5 to 1 mm. in height, which contained pig-
ment particles and stumps of the fin rays. After
10 days the fin rays had begun to regenerate in
the normal way and the micromelanophores
reached the distal edge of the fin, producing a
“Pigmentsaum” effect, a term used by Bosen-
berg (1938); during the same period macro-
melanophores also migrated into the fin from
its base. Text-fig. 3b. After two weeks the re-
generated fin was 1.5 mm. in height. Within a
month it was 3 to 4 mm. and the micro- and
macromelanophores had increased in number
and the latter had begun to form black spots
at the base of the fin. Text-fig. 3c. The macro-
melanophores had migrated over and along the
fin rays; as a consequence, the basal part of the
rays was completely covered. Within two to
three months the new fin attained its previous
size, 4.5 to 5 mm. in height. The macromelano-
phores continued to invade the regenerated fin.
Text-figs. 3d to f. But even six months after the
amputation the melanosis was not as strong as it
was in the original fin. Text-fig. 3g. It appears,
then, that after the fins regenerate normally they
are invaded by macromelanophores which pro-
duce a somewhat lesser state of melanosis.

Regeneration of the Dorsal Fins in Species
Hybrids between the Platyfish
AND Swordtail

Some spotted-dorsal hybrids exhibited mel-
anosis, others typical melanomas, while still
others showed amelanotic melanomas in their
dorsal fins. For purposes of comparison, how-
ever, two platyfish-swordtail hybrids without
macromelanophores but with micromelano-
phores in the dorsal fins were first studied. Table

1, No. 6. Within one week after amputation,
blastemas 0.2 to 0.5 mm. were formed. Within
two weeks the blastemas were 1 to 1.5 mm. and
contained fin rays. Micromelanophores soon
reached the distal edge of the regenerating fins,
forming pigmented borders. The pigment cells
migrated along both sides of the fin rays, leaving
clear areas between the rays. Within two months
the fin rays began to bifurcate. Within three
months the fins reached their previous size of 5
to 6 mm.; their former pigmentary patterns were
at that time restored.

The regeneration process was observed in
eighteen platyfish-swordtail hybrids with dorsal
fin melanosis, Table 1, No. 7. In some hybrid
fish the melanosis was strong; in others just a
few discrete macromelanophores were present in

the dorsal fins. Most hybrids had, in addition to
a melanosis of the dorsal fin, a melanosis of the
tail and body, especially in the region under the
dorsal fin. Text-fig. 4.

After amputation of the dorsal fins of fish
with incomplete melanosis, blastemas contain-
ing some pigment particles were formed within
five to seven days. Within 10 to 15 days the
regenerated fins were 1 to 1.5 mm. and con-
tained micromelanophores and fin rays. Macro-
melanophores began to appear at the base of the
regenerating fins within 14 days in those that
had been highly melanotic, and within 25 to 30
days in others. Within about three months the
regenerated fins reached their original size, 8 to
10 mm. The new pigment patterns produced by
the macromelanophores in the dorsal fins were
not always the same, nor was the degree of
melanosis as great as in the originals.

In those fish showing strong melanosis in their
dorsal fins, the regeneration of the macromel-
anophores was more rapid; one. Text-figs. 4 to
4g, will be described in detail. During the healing
of the wound, the epidermis contained dense
pigment particles derived from the degenerat-
ing melanophores; this cleared up within one
week. Dendritic processes of the macromelano-
phores, located just below the removed fin, pene-
trated the blastema and later whole cells entered
it. Text-figs. 4b, 4c. Within two weeks the re-
generated fin was 3 mm. and had many micro-
melanophores in its distal part; discrete macro-
melanophores appeared between the rays and
formed a melanosis at the base of the fin. Text-
fig. 4d. The fin reached its full height within
two months. Text-figs. 4e to 4g, during which
time melanosis was intensified. The 814 -months-
old regenerated fin was almost as melanotic as
the original fin. Text-fig. 4g.

The migration paths of the macromelano-
phores from their prior positions on the body
just ventral to the fin into the regenerated fin
between rays 7 and 8 may be seen in Text-fig.
4h. Distally some discrete macromelanophores
contain fine grayish melanin granules that repre-
sent newly formed macromelanophores; these
have been transformed from melanocytes. With-
in two months the previously produced darker
macromelanophores that had originated from
the base of the fin, and the newly formed gray
macromelanophores, came together and pro-
duced a state of melanosis. Text-figs. 4i, 4j.
Thus, there are two sources of macromelano-
phores in the regenerate: from those that were
already present below the fin and from macro-
melanophores that are formed in situ from mel-
anocytes in the regenerate itself.

The development of melanosis was faster in
young hybrids than in mature ones. For ex-

60

Zoologica: New York Zoological Society

[40: 6

ample, in young animals of 16 to 20 mm. in
standard length with complete melanosis in the
dorsal fin, as well as in the body below the
dorsal fin, the macro melanophores migrated into
the blastema within four days after amputa-

tion. Within three weeks the large black pig-
ment cells reproduced an almost complete mel-
anosis. The regeneration of the dorsal fin of one
young fish of 21 mm. in length was peculiar.
The anterior part of its original fin and the

Text-fig. 4. Regeneration of the spotted dorsal fin of a platyfish-swordtail hybrid with severe melanosis
in the dorsal fin. a to g, 5X. a— Before amputation, b— Within 2 days after amputation, c— Within 7 days,
d— Within 14 days, e— Within 21 days, f— Within 30 days, g— Within 250 days; note that melanosis is
almost the same as it was in the uncut fin, figure a.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

61

Table 4. Regeneration, after Amputation, of Dorsal Fins and Pigmentation
IN Platyfish-Swordtail Hybrids^

Regeneration

No.

Pattern

In days

Of Fin

Of Pigmentation

19.

Micromelanophores^

165

Complete

Complete

20.

M icromelanophores^

120

Complete

Complete

21.

Melanosis (3)®

257

Complete

Melanosis (2)

22.

Melanosis (2)

196

Complete

Melanosis (1)

23.

Melanosis (2)

180

Complete

Melanosis (2)

24.

Melanosis (2)

25

Incomplete

Fixed

25y.*

Melanosis (3)

30

Incomplete

Melanosis (3)

26.

Melanosis (2)

1

Incomplete

Fixed

27.

Melanosis (3)

1

Incomplete

Fixed

28y.

Melanosis (3)

21

Complete

Melanosis (3)

29y.

Melanosis (3)

6

Incomplete

Fixed

30.

Melanosis (3)

27

Incomplete

Fixed

31.

Melanosis (3)

30

Incomplete

Melanosis (2)

32.

Melanosis (3)

4

Incomplete

Fixed

33.

Melanosis (3)

210

Complete

Melanosis (3)

34y.

Melanosis (2)

30

Bilobed

Melanosis (1)

35y.

Melanosis (2)

30

Complete

Melanosis (1)

36.

Melanosis (2)

30

Incomplete

Melanosis (1)

37.

Melanosis (2)

9

Incomplete

Fixed

38.

Melanosis (2)

280

Complete

Melanosis (1)

^Xiphophorus maculatus—Xiphophorus helleri hybrids, pedigree numbers 311 and 316, see Table 1, items 6, 7.
2pish numbered 19 and 20 had no macromelanophore patterns in dorsal fins; used for controls.

®The severity of melanosis is indicated by numbers in parentheses.

*y represents an immature specimen.

Text-fig. 4 (Continued), h— The proximal area of the regenerate between 7th and 8th fin rays as seen
under higher magnification, after 30 days. The macromelanophores are the dark cells which had moved up
from the base of the fin. The melanocytes are the gray cells. These contain fine, dispersed melanin
granules, i— Same after 38 days. Newly formed macromelanophores and melanocytes are located between
the rays, j— After 60 days. The growth of macromelanophores has created a state of melanosis.

62

Zoologica: New York Zoological Society

[40: 6

Table 5. Regeneration, after Amputation, of Melanomas in the Dorsal Fins
OF Platyfish-Swordtail Hybrids^

Regeneration

No.

Pattern

In days

Of Fin

Of Melanoma

39.

Melanoma on
and below fin

75

Bilobed

Melanoma

40.

Melanoma

300

Bilobed

Melanoma

41.

Melanoma

180

Bilobed

Melanosis (3)

42.

Melanoma

346

Complete

Melanoma

43.

Melanoma

69

Bilobed

Melanosis (3)

44.

Melanoma on
and below fin

46

Complete

Melanosis (3)

45.

Melanoma on
and below fin

1

Incomplete

Fixed

46.

Melanoma on
and below fin

20

Incomplete

Fixed

47.

Melanoma

198

Complete

Melanosis (3)

48.

Melanoma

13

Incomplete

Fixed

^Xiphophorus maculatus-helleri hybrids of pedigree number 311, see Table 1, item 8.

body just below it were entirely melanotic,
whereas the posterior part of the fin and the
body below it were not. After amputation the
middle part of the fin did not regenerate. Within
a month the anterior part of the new fin grew to
3 mm. and showed an almost complete mel-
anosis, whereas the posterior part, which was
2 mm., showed a much lesser degree of mel-
anosis. These observations suggest that when the
tissues underlying the removed fin are complete-
ly melanotic at the time of operation, the mel-
anosis in the regenerated fin develops more
rapidly and becomes complete.

The regenerations of fins in 10 platyfish-
swordtail hybrids with melanotic melanomas in
their dorsal fins were observed. Table 1, No. 8.
In some of these hybrids the dorsal fins were
destroyed to various degrees, and in others
secondary melanomas had developed on various
parts of their bodies. Text-figs. 5 and 6.

In hybrids that had melanomas in the dorsal
fin only, the regenerated fins were usually ab-
normal. Within a week after a melanomatous
fin was amputated, a blastema formed which
contained much pigment cell debris. Text-figs.
5a, 5b. Within one month the anterior and
posterior parts of the fin regenerated but not
the middle. Text-figs. 5c, 5d. After two months
the regenerated fin became melanotic. Text-fig.
5e. Within 10 months a melanoma formed at
the base of the anterior part of the fin, and a
second nodular melanoma developed on the
posterior part. Text-fig. 5f. In another animal,
within six months following amputation, the re-

generated fin was in a state of melanosis; it was
then fixed for the histological study. After am-
putation of a destructive, bilobed, melanotic
melanoma in the dorsal fin of a third hybrid, a
non-tumorous, bilobal, deeply pigmented fin re-
generated within 11 months. It measured 7.5
mm., whereas the removed fin had been only 1.5
mm. posteriorly and 5.5 mm. anteriorly.

In hybrids that had melanoma both in the
dorsal fin and on the body ventral to it, the
redevelopment of tumors in the regenerated fins
was more rapid. For instance, one hybrid had a
melanoma of the dorsal fin as large as the fin it-
self, and had infiltrated the tissue just below the
fin. Text-fig. 6. After the fin together with the
growth had been amputated, a melanotic bilobed
fin developed at first. Text-figs. 6b to 6e, and
then after 75 days a melanoma developed at the
base of the anterior lobe and another around
the posterior part, Text-fig. 6f. These observa-
tions suggest that when the melanoma is re-
stricted only to the dorsal fin, the period of re-
development of the tumor in the regenerated fin
is longer than in those fish that exhibit mel-
anoma both on the dorsal fin and on the body
below that fin.

The fins of five platyfish-swordtail albino
hybrids that had amelanotic melanomas in their
dorsal fins were amputated, Table 1, No. 9. The
melanomas were heavy and pink-colored; they
had partly invaded the bodies below the dorsal
fins. Each fish that had its melanomatous dorsal
fin amputated reacted differently, but the re-
sponses were only slightly different from those

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

63

Text-fig. 5. Regeneration of melanomas in the dorsal fin of a platyfish-swordtail hybrid. 5X. a— Before
amputation, b— Within 7 days following amputation, c— Within 14 days, d— Within 30 days, e— Within
60 days, f— Within 10 months.

of the fish having typical melanomas that were
previously described.

In one albino hybrid the original tumor was
nodular and restricted to the anterior part of the
fin. This part of the fin was either surrounded or
destroyed by the tumor; the posterior part was
also tumorous but the fin rays were still visible.
The entire tumor of the dorsal fin contained
large branching blood vessels clearly visible near

its surface. Within a week following amputation,
a 0.6 mm. blastema formed which contained no
visible blood vessels or fin rays. At two weeks
the regenerated fin measured 1 mm.; within one
month it was 2 mm. and had five visible fin rays
on its distal edge. Apparently the blastema and
the tumor developed simultaneously, because
within two months the basal one-third of the
regenerated fin had a heavy melanoma; the re-

64

Zootogica: New York Zoological Society

[40: 6

Text-fig. 6. Regeneration of a melanoma in the dorsal fin of a platyfish-swordtail hybrid in which the
melanoma had extended into the body ventral to the fin. 10 X. a— Before amputation, b— Within 7 days
following amputation, c— Within 14 days, d— Within 21 days, e— Within 30 days, f— Within 45 days.

Table 6. Regeneration, after Amputation, of Amelanotic Melanomas in the Dorsal Fins

OF Platyfish-Swordtail Hybrids^^

Regeneration

No.

Pattern

In days

Of Fin

Of Melanoma

49.

Amelanotic melanoma

90

Complete

Amelanotic melanoma
at base of fin

50.

Amelanotic melanoma
on, below fin

90

Complete

Amelanotic melanoma

51.

Amelanotic melanoma
on, below fin

60

Complete

Amelanotic melanoma

52.

Amelanotic melanoma
on, below fin

20

Incomplete

Fixed

53.

Amelanotic melanoma
on, below fin

30

Incomplete

Amelanotic melanoma

^Xiphophorus maculatus-helleri hybrids of pedigree 311, see Table 1, item 9.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

65

Text-fig. 7. Regeneration of an amelanotic melanoma in the dorsal fin of a platyfish-swordtail
hybrid. 6 X. a— Before amputation, b— Within 1 week following amputation, c— Within 17 days, d— Within
1 month, e— Within 2 months, f— Within 3 months.

mainder was normal and contained 12 rays. All
parts of the new growth were highly vascularized,
but its general appearance was entirely different
from that of the original fin. The fish died two
months after its fin was amputated.

In another hybrid, an amelanotic melanoma
in the anterior region of a 10-ray ed dorsal fin
was amputated. Text-fig. 7. A blastema contain-
ing four fin ray stumps developed within a week.
Text-fig. 7b. Within two weeks the regenerated
fin grew to 1 mm. and contained six fin rays.
Text-fig. 7c. Within a month it was 3 mm. and
had 13 rays, three more than the original one.
Text-fig. 7d. A melanoma had developed at its
former site, at the base of the anterior region
of the fin. Text-fig. 7e. Within three months the
regenerated fin reached its previous height of
5 mm., at which time the fish was fixed for
histological study.

Another albino had a heavy, highly vascu-
larized melanoma in its dorsal fin in which some
fin rays were visible, particularly their tips. The

tumor had infiltrated the body below the fin and
a secondary melanoma was present on the dorsal
edge of the tail. Text-fig. 8. Within a week after
its amputation a blastema developed into an
amorphous mass about 1 mm.. Text-fig. 8b;
within 17 days it was 2 mm., more vascularized,
and contained four fin rays. Text-fig. 8c. Within
one month the regenerating fin had six rays and
a larger melanoma. Text-fig. 8d. At two months
the new growth measured 4 mm.; it had de-
stroyed all but three of the regenerated rays.
Text-fig. 8e, and these were reduced to two
within three months. During these three months
the highly vascularized amelanotic melanoma
reached its original size of 5.5 mm.. Text-fig. 8f.
The fish was then fixed for histological study.

The dorsal fin of another albino hybrid with
an amelanotic melanoma measuring 1.5 mm.
grew back after amputation to 4 mm., a size con-
siderably larger than that of the original fin.
The regrowth of the fin is thus not necessarily
impeded by the simultaneous growth of a mel-
anoma.

66

Zoologica: New York Zoological Society

[40: 6

8

Text-fig. 8. Regeneration of a large amelanotic melanoma in the dorsal fin of a platyfish-swordtail
hybrid. 6 X. a— Before amputation, b— Within 1 week following amputation, c— Within 17 days, d— Within
1 month, e— Within 2 months, f— Within 3 months.

Histological Observations of
Regenerated Dorsal Fins
Histological observations of the regeneration
process of the normal dorsal fins of platyfish and
platyfish-swordtail hybrids having raicromelano-
phores only (represented in Table 1, Nos. 2, 3
and 6) may serve as a basis for comparison with
the regeneration process in fish having abnorm-
ally pigmented fins (represented in Table 1, Nos.
7, 8 and 9).

Within 12 hours after amputation of the
normally pigmented fins, macrophages phago-
cytize melanin particles and other cell debris.
The squamous cells of the epithelium moving
from both sides of the wound grow over its sur-

face to reform the epidermis. In the epidermis,
the epithelial cells are fibrillar and loosely ar-
ranged; the macrophages that contain melanin
particles are round or oval and have eccentric
nuclei.

Within two or three days the basal cells of
the adjacent epidermis move under the squa-
mous epithelial cell layer. Collagenous fibres be-
come more abundant in the connective tissues
below the regenerating epidermis. Between these
fibres are scattered fibroblasts, lymphocytes, gran-
ulocytes, macrophages and unengulfed melanin
particles. The basal cells, some of which show
mitotic figures, are 8 to 9 micra and their round
or oval nuclei contain one or two nucleoli. The

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

67

Text-fig. 9. Details from blastema of regenerating dorsal fin of a comet platyfish, after one week,
refer to Text-fig. 2b. Me— Melanocyte between the blastema cells. Bl.c— Blastema cells. Co.f— Collagenous
fibers at the base of blastema. Er— Erythrocyte.

Text-fig. 10. Regenerating epidermis in the blastema, see Text-fig. 2d. Note one blastema cell in mitotic
division. Me— Melanocyte. Ba.c— Basal cells. Bl.c— Blastema cell. Pm— Pigment mass.

squamous cells are 5 to 7 micra and have
elongated nuclei. The fibroblasts are round or
oval, 5.5 to 7.5 micra, and have round or oval
nuclei containing one or two nucleoli. The fibro-
blasts show some mitotic figures and are most
active in the formation of the blastema proper,
Plate I, Figs. 1, 2. The initial blastema con-
tains no melanophores although its epidermis
contains some free melanin and some macro-
phages with ingested pigment. Eventually the
free melanin picked up by macrophages passes
through the epidermis in the manner described
by Bosenberg (1938) and by Gordon & Lansing
(1943). Following this, melanocytes and young
micromelanophores appear in the base of the
blastema. The melanocytes are 14 to 15 micra;
they are oval or spindle-shaped, but later they
become dendritic. In Text-fig. 9 and Plate I, Fig.
3, a melanocyte is shown between blastema cells
and collagenous fibres at the base of the regen-
erating fin, and in Text-fig. 10 another is shown
just under the basal cells of the epidermis; here
one blastema cell is in a stage of mitotic division.
During the first week, blood capillaries appear in
the blastema. Micromelanophores continue to
increase as they move in along the connective
tissue of the regenerating fin.

With regard to the regeneration of the fin rays,
within one to two weeks after amputation, mac-
rophages accumulate about the stumps of the fin
rays and phagocytize the lepidotrichial debris.
Later, the distal blastema cells begin to form
actinotrichia and then lepidotrichia in a manner
described by Blanc (1949). At first the regen-
erating lepidotrichia are not in contact with the
ray stumps. Within three weeks those which
form below the epidermis approach each other
along their convex surfaces to reconstitute a fin
ray. The proximal tips of the regenerated fin

rays approach the distal tips of the old ray
stumps and the distance between them is filled
with additional skeletal elements from the
blastema cells.

In Text-fig. 11, a diagonal cross-section of a
dorsal fin 23 days after amputation, the lepido-
trichia of two successive fin rays are shown. The
epidermis of the fin is hyperplastic; the blastema

Text-fig. 11. A diagonal section through a regen-
erated dorsal fin of a comet platyfish 23 days after
amputation, see Text-fig. 2d. Act— Actinotrichia.
Bl — Blastema. E — Epidermis. Lep — Lepidotrichia.
Mi— Micromelanophore.

68

Zoologica: New York Zoological Society

[40: 6

Text-fig. 12. A section through a regenerating
dorsal fin of a platyfish-swordtail hybrid four days
following amputation of the fin, which had been in
a state of melanosis, see Text-fig. 4c. Bo.e— Basal
cell. Bl.e— Blastema cell. E— Epidermis. Mph— Macro-
phage. P— Processes of macromelanophore. Pm—
Pigment mass in the epidermis.

cells show mitotic divisions. Blood capillaries
have developed. New micromelanophores are
located in the blastema and macromelanophores
are present in the base of the fin. Most of the
macrophages and free pigment masses that were
in the epidermis have been eliminated.

Melanocytes, the precursors of micro- and
macromelanophores, are present between the
blastema cells and in the epidermis, Text-figs. 14
and 15. They measure 11 to 22 micra and are
round, oval cells, or spindle-shaped; later they
may develop dendritic processes. It can not be
said whether melanocytes come from the tissue
under the epidermis or from the epidermis itself.
Sometimes macromelanophores are located in
the epidermis of the regenerated fin. Text-fig. 16.
All of the macromelanophores apparently do not
originate from melanocytes because macromel-
anophores that had been in the underlying tis-
sues of the removed fin migrate into the regen-
erate, Text-figs. 4 and 17. As regeneration pro-
ceeds, the macromelanophores accumulate in

cells are present in the distal part. The paired
lepidotrichia of the old fin rays are located at
the base of the fin. In this figure some micro-
melanophores are shown in the connective tis-
sues around the fin rays. Within two to three
months the regenerating fin reaches its previous
dimension and then ceases to grow; the epi-
dermal cells are no longer hyperplastic, the
mucous and sensory cells reappear. Within four
to five months the former pigmentation is re-
stored in the regenerated fin.

From histological observations of the regen-
eration process, it appears that the most active
growth takes place along the distal margin of
the blastema, for it is there that mitotic figures
are most frequently found.

Hybrids with melanosis of the dorsal fin re-
acted in a manner similar to the previous group
following amputation, except for differences in
pigment cell development. Within 12 to 48 hours
following amputation of a melanotic dorsal fin,
melanophore debris, free melanin, pigment-con-
laining macrophages and collagenous fibres are
present in the wound, Plate I, Fig. 4. Within
three to four days the wound is covered by an
epidermis and basal cells are restored. Macro-
phages and intercellular pigment masses are
still present but they are eventually eliminated,
Plate I, Fig. 5. The blastema cells accumulate
under the basal cells of the epidermis. Text-fig.
12. This figure also shows some processes of
macrophages that have entered the area of re-
generation from below. Within six to seven days
the blastema, which is covered by a hyperplastic
epidermis, is formed. Text-fig. 13. Mucous cells
in the epidermis are now present. Some blastema

Text-fig. 13. Cross-section of a blastema formed
within six days in a similar fish, see Text-fig. 4.
Be— Blood capillary. Bl— Blastema. Bl.c— Blastema
cell. E— Hyperplastic epidermis. Ma— Macromelano-
phore. Mi— Micromelanophore. Mit— Mitotic divi-
sion in blastema cell. Pm— Pigment mass.

the connective tissue, where their pseudopodial
processes anastomose and they recreate a state
of melanosis in the regenerated fin. Text-fig. 18.

Two melanoblasts, 14 micra, each with fine
melanin in granules, were found around a blood
capillary of a one-month-old regenerated fin,

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

69

Text-fig. 14. Details of melanocytes in the hyper-
plastic epidermis in the blastema of similar fin,
see Text-fig. 4c. Me— Melanocyte. E— Epidermal
cell. Bl— Blastema cell.

Text-fig. 19. This suggests that new pigment cells
in the blastema may develop in situ.

Fin rays may regenerate in the presence of
melanosis, but in a few of these regenerating
fins all the rays did not redevelop. Lepidotrichia
first reform in the distal part of the fin, while

pigment and epidermal cells generally reappear
in the proximal region. Scales that have been
destroyed at the base of the fin may also regen-
erate. Within two to three months the regener-
ated fin attains its former height, but the degree
of its pigmentation is different. The develop-
mental rate of melanosis in the regenerating fin
depends upon the age of the animal and upon the
degree of melanosis that exists on the body ven-
tral to the amputated fin.

In hybrids that have melanotic melanomas in
their dorsal fins, the regeneration process is
similar, in the earlier stages, to that in hybrids
whose dorsal fins show strong melanosis. Within
24 hours after amputation of the fin, the re-
generating epidermis covers the wound surface.
Text-fig. 20 and Plate I, Fig. 6. From these
figures it may be seen that beneath the ampu-
tated fin, in the body proper, connective and
muscle tissues had been replaced by tumor cells.
The regenerating epidermis is hyperplastic and
the cells are vacuolated. There are many macro-
phages and pigment masses. A bleached section
of the fin is shown in Plate II, Fig. 1 ; the paired
lepidotrichia appear darker here than in un-

Text-fig. 15. Details of melanocytes in the hyperplastic epidermis in the blastema of similar fin, see
Text-fig. 4c. Me— Melanocyte. Ba.c— Basal cells of the epidermis.

Text-fig. 16. Macromelanophore in the regenerating epidermis of a blastema of a similar fish a month
later, see Text-fig. 4f. Ma— Macromelanophore.

Text-fig. 17. Cross-section of similar fish showing part of the regenerated dorsal fin within one month
following amputation and part of dorsal region of body. Macromelanophores have migrated into the
regenerated fin from the region of melanosis below dorsal fin, see Text-fig. 4f. Be- Blood capillaries.
E— Epidermis. Ma— Macromelanophore. Se— Scale.

Text-fig. 18. Cross-section through the regenerated dorsal fin after 6 months, see Text-figs. 4f, 4g.
Sc— Scale.

Text-fig. 19. Two melanoblasts around a blood vessel of a one-month-old regenerated dorsal fin, see
Text-fig. 4f.

70

Zoologica: New York Zoological Society

[40: 6

Text-fig. 20. Cross-section through the regenerated epidermis of an amputated dorsal fin after 24 hours.
The dorsal fin had a melanoma which extended into the region below the fin, see Text-figs. 6, 6a, 6b.
E— Epidermis. Lep— Lepidotrichia. Pm— Pigment mass. Sc— Scale.

Text-fig. 21. Distal region of a 13-day-old blastema of a melanomatous fish, see Text-figs. 5b, 5c.
Act- Actinotrichia. lep— Lepidotrichia. Ba.c— Basal cells. Bl.c— Blastema cells. Er— Erythrocyte.

Text-fig. 22. Hyperplastic epidermis in a similar fish after 45 days, see Text-fig. 5e. Me— Melanocyte.
Ba.c- Basal cells. Muc— Mucous cell.

Text-fig. 23. Macromelanophore surrounding a blood capillary in the connective tissue of a 13 -day-old
regenerated dorsal fin, see Text-fig. 5c.

Text-fig. 24. Basal region of a regenerated dorsal fin, 13 days after amputation of a melanoma. The
tumor cells’ path of invasion of the regenerated fin follows the connective tissue, see also Text-fig. 6c.
H.c— Hypertrophic fibroblasts. Ct.c— Fibroblasts, some are hypertrophic. Mel— Melanomatous tissue.

bleached sections, Plate I, Fig. 6. Within six to
seven days the basal cells of the new epidermis
appear over the wound’s surface and the
blastema is formed. The blastema cells, as is
usual, originate from fibroblasts. In cases where
the connective tissue beneath the amputated fin
has been destroyed by tumor tissue, the blastema
cells migrate from the nearest undestroyed con-
nective tissues. Within 10 to 15 days actino-
trichia appear, followed by the formation of
paired lepidotricha. Text-fig. 21. During their
regeneration some abnormalities occur, for ex-
ample, sometimes three fin rays may regenerate
to replace two.

Melanocytes and micromelanophores first ap-
pear in the blastema. Later, within two to three
weeks, macromelanophores either migrate into

the blastema from the region immediately below
or they are reformed by melanocytes in situ. A
melanocyte in the hyperplastic epidermis of a
2.5 months’ old regenerated fin is shown in Text-
fig. 22; Text-fig. 23 represents a macromelano-
phore around a capillary in the connective tissue.

After the condition of melanosis has returned
in the regenerated fin, the hyperplastic growth
of macromelanophores continues, the cells di-
viding amitotically. They become the principal
tumor cells, replacing the connective tissue in
the regenerated fin. The early fibroblasts become
hypertrophic, Text-fig. 24; they measure 22
micra, are spindle-shaped, and their oval or
round nuclei measure 8 to 10 micra. The ex-
tended pseudopodial processes of the macro-
melanophores first surround the fibroblasts, Text-

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

71

Text-fig. 25. Pigmented hypertrophic fibroblasts
at the base of a 13-day-old regenerated dorsal fin,
see Text-figs. 6c and 24. A— Hypertrophic fibro-
blast surrounded by the pigmented processes of
macromelanophores. B— Pigmented hypertrophic
fibroblast, its melanin acquired from adjacent pig-
ment-producing cells, macromelanophores. C— Pig-
mented hypertrophic fibroblast with nucleus in a
state of karyolysis. D— Pigmented hypertrophic
fibroblast with central mass of pigment granules.

Text-fig. 26. Bleached section showing pigmented
hypertrophic fibroblasts in amitotic division. The
nucleus on the right is that of a macromelanophore
which surrounds the secondary pigmented fibro-
blast, see also Text-fig. 6d.

Text-fig. 27. Multinucleated spindle-shaped cell in
the connective tissue of the regenerated dorsal fin,
shown in Text-figs. 24 and 6c.

Text-fig. 28. Giant cell with 7 nuclei from regen-
erated melanoma, see Text-fig. 6f.

fig. 25a. The fibroblasts obtain their melanin
pigment by these contacts, Text-fig. 25b. The
fibroblasts, which may be round, oval or poly-
morphic, increase to 25 to 35 micra and their
nuclei measure 10 to 15 micra. Their nuclei
divide amitotically with the result that in some
cells two or more nuclei may be formed; this is
illustrated in Text-fig. 26 which shows a
bleached preparation. The nucleus at the periph-
ery of the fibroblast shown in this figure also
represents part of the macromelanophore which
surrounds it. The fibroblasts which appear dur-
ing the early stages of melanoma formation in
the regenerate may be called “pigmented hyper-
trophic fibroblasts.” The nuclei of some of these
cells may disappear by the process of karyolysis.
Text-fig. 25c. Melanin granules may reach the
center of the fibroblasts and become concen-
trated there. Between this central mass of mel-
anin in the pigmented hypertrophic fibroblasts
and the surrounding macromelanophores, a non-
pigmented ring may appear. Text-fig. 25d. The
pigmented hypertrophic fibroblasts do not pro-
duce their contained melanin, but acquire it
from adjacent pigment-producing cells. Some
pigmented hypertrophic fibroblasts that devel-
oped in a melanoma are shown from a bleached
section in Plate II, Fig. 2.

In the melanoma where pigmented hyper-
trophic fibroblasts develop, certain relatively
small and polynucleate cells, measuring 18 to 19
micra, are occasionally found. Text-fig. 27. The
cytoplasm of these rare cells stains more in-
tensely with hemalum-eosin than the other cells
in the regenerating tissues. These multinucleate
cells possibly originate from the hypertrophic
fibroblasts. In addition, some large polymor-
phous and multinucleate cells, which are about
500 micra in size and are called “giant cells,” ap-
pear in the tumor tissue. One giant cell with
seven nuclei is shown in Text-fig. 28. Probably
these large cells are also derived from fibro-
blasts.

The connective tissue in the regenerated fin
forms the fibrillar stroma of the developing mel-
anoma. The connective tissue cells are relatively
small, being 10 to 15 micra, and have nuclei 3
to 8 micra. Like fibroblasts, they may become
pigmented during the hyperplastic growth of the
macromelanophores. It is the hyperplasia of the
tumor cells described above that causes a thick-
ening of the fin which may even become nodular.

Macromelanophores are the principal tumor
cells of the melanomas that redevelop in the re-
generating dorsal fins. They measure 300 to 470
micra, their nuclei are 15 to 20 micra; they
possess fine or lobulated dendritic processes. In
the tumor tissue, they are found as individual
cells or in complex aggregations forming syn-

72

Zoologica: New York Zoological Society

[40: 6

V'X '

200^

Text-fig. 29. Cross-section of regenerated melanoma that developed after 2.5 months in the dorsal fin
in hybrid shown in Text-figs. 6, 6f. Lep— Lepidotrichia. Sc— Scale.

Text-fig. 30. Pigmented hypertrophic fibroblast with one nucleus from melanoma shown in Text-fig. 29.
Text-fig. 31. Pigmented hypertrophic fibroblast with two nuclei from melanoma shown in Text-fig. 29.

Text-fig. 32. Bleached section of the dorsal fin melanoma shown in Text-figs. 6f and 29. Several macro-
melanophores form a syncytium within a fibrillar connective tissue stroma. A pigmented hypertrophic
fibroblast is shown in the central area.

Text-fig. 33. Bleached section showing an isolated macromelanophore from a melanomatous dorsal fin
shown in Text-fig. 5f.

Text-fig. 34. Bleached section showing an isolated macromelanophore with three nuclei from the
melanoma shown in Text-fig. 5f.

Text-fig. 35. Epidermis of a regenerated melanomatous dorsal fin, see Text-figs. 5e, 5f. E— Epidermis.
Mph— Macrophage.

cytial masses in which cell boundaries can not be
determined, Plate II, Figs. 3 and 4. In a bleached
section of the melanoma, hypertrophic pig-
mented fibroblasts, connective tissue stroma cells
and syncytial macromelanophores may be seen,
Plate II, Fig. 5, The fibrillar structure of the
cytoplasm of the macromelanophores may be
seen clearly in this photomicrograph.

Two types of recurrent melanomas may be
distinguished with regard to their origin and
rapidity of growth. The fast-growing type is
mainly composed of macromelanophores that
have migrated into the reformed tumor from
melanomatous tissues in the stump of the orig-
inal dorsal fin and in the body just below, as
shown in Text-figs. 6 and 29 and Plate II, Fig. 6.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

73

Text-fig. 36. Principal cells of the recurrent mel-
anomas in the dorsal fins of platyfish-swordtail
hybrids following amputation: a and b— Macro-
melanophores. c— Four melanocytes, d— Pigmented
hyperplastic fibroblast, e— Giant cell, f— Stroma cell.

The tumor tissue in the regenerated fin, as well
as in the basal region of the fin, have the same
cellular elements and melanotic appearance. The
processes of the macromelanophores lie parallel
to each other or form swirls, Plate II, Fig. 4.
Their cell boundaries are indeterminate, as in-
dicated in the study of bleached sections. Most
of these cells have migrated from their previous
position in the basal parts of the removed fin and
body, Plate III, Fig. I. Other macromelano-
phores, arising from melanocytes, also partici-
pate in recreating the new tumor in the regen-
erated fin, but since they form a syncytial mass
with the macromelanophores that migrate into
the regenerating fin, it is difficult to distinguish
between them. In this type of melanoma, pig-
mented hypertrophic fibroblasts, giant cells and
connective tissue stroma cells are found. Two
pigmented hypertrophic fibroblasts, one of
which contains two nuclei, are shown in Text-
figs. 30 and 31; Text-fig. 32, drawn from a
bleached preparation, shows a pigmented hyper-
trophic fibroblast with several macromelano-
phores in a fibrillar stroma of connective tissue
cells.

The more slowly-growing melanomas are
those that require approximately ten months to

redevelop in situ. These develop in hybrid fish
exhibiting melanomas that were originally re-
stricted to the dorsal fin. Text-fig. 5. They con-
sist of isolated polymorphic macromelanophores
which originate in situ from melanocytes in the
regenerated fin, Plate III, Fig. 2. One hybrid
developed a nodular melanoma in the posterior
part of the regenerate. Text-fig. 5. Cross-sections
of this tumor, Plate III, Figs. 2 and 3, show that
it is not completely pigmented. The almost
round or oval macromelanophores with fine proc-
esses are imbedded in a fibrillar and less pig-
mented connective tissue stroma. This may be
seen in a partially bleached section of this tumor,
Plate III, Fig. 4. Macromelanophores containing
one or more nuclei are observed in other
bleached sections. Text-figs. 33 and 34. The posi-
tion of the lepidotrichia determines a radial
arrangement of the fine processes of the macro-
melanophores in parts of the melanoma, as
shown in Plate III, Figs. 4 and 5.

In both types of recurrent melanoma in the
regenerated dorsal fins, certain tumor cells re-
veal a degenerative process in action, as in-
dicated by pyknosis, karyorrhexis or karyolysis
of their nuclei. Blood capillaries are well dis-
tributed throughout the melanoma and granu-
locytes and free erythrocytes are also present.
The recurrent melanoma is usually covered by
a thin epidermis, but in some instances the epi-
dermal cells may be hyperplastic and may con-
tain some macrophages and melanin masses, as
shown in Text-fig. 35. The skeletal elements
usually resist infiltration but they, too, may be
destroyed by the tumor cells.

Thus, the first step in the recreation of mel-
anoma in a regenerated dorsal fin is the de-
velopment of a melanosis which is characterized
by the proliferation and accumulation of macro-
melanophores. The macromelanophores then
undergo progressive hyperplasia inducing, in
turn, the hypertrophic growth of fibroblasts and
giant cells. The macromelanophores invest these
and other non-pigment-producing cells with
melanin particles.

The most important pigmented cellular ele-
ments observed in the regenerating melanoma
are as follows and are represented in Text-
fig. 36:

a. Melanocytes: These are pigment-produc-
ing cells of 1 1 to 22 micra, with round or oval
nuclei of 3 to 5 micra. They are usually oval
or spindle-shaped, but may have wide lobulated
processes. Melanocytes are capable of trans-
forming into micro- and macromelanophores.

b. Melanophores: These are derived from
melanocytes and are of two kinds:

Macromelanophores are large polymorphic
cells measuring 300 to 470 micra, with round.

74

Zoologica: New York Zoological Society

[40: 6

lOOJU.

1 i v C- V

; M'

100^

■ykry'^ii'

'■ _ . \
5 0 0>

Text-fig. 37. Cross-section of an amelanotic melanoma in the dorsal fin of a hybrid that reformed
within 20 days after amputation, see Text-fig. 7c. E— Hyperplastic epidermis. Lep— Lepidotrichial element.
Co.f— Collagenous fibers. Sc— Scale.

Text-fig. 38. Section of the amelanotic melanoma shown in Text-fig. 37, under higher magnification.

Text-fig. 39. Section of unamputated amelanotic melanoma tissue within a scale pocket at the base of
the regenerated fin, see Text-fig. 7c. E— Epidermis. Be— Blood capillary. Gc— Giant cells. Sc— Scale.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

75

oval or polymorphic nuclei that are 15 to 20
micra; they contain one to two nucleoli and a
loose network of chromatin. They have fine
or lobulated dendritic processes. These are the
cells that initiate the development of melano-
mas in platyfish-swordtail hybrids.

Micromelanophores are smaller, about 100
micra.

c. Pigmented hypertrophic fibroblasts: These
cells have no dendritic processes. They are
round, oval or polymorphic and measure 25
to 35 micra; their round or oval nuclei are 10
to 15 micra. They acquire their melanin from
adjacent macromelanophores.

d. Giant cells: These are large polymorphic
and multinucleated cells of about 500 micra.
Their nuclei, which are round, oval or occa-
sionally polymorphic, and are 5 to 10 micra in
size, have a loose network of chromatin with
one or two nucleoli. The giant cells may develop
dendritic processes rarely; they also acquire
pigment granules from true pigment cells.

e. Pigmented connective tissue stroma cells:
These are spindle-shaped cells of 10 to 15 micra.
Their oval or round nuclei measure 3 to 8
micra and contain dense granular chromatin.
They, too, may acquire their small number of
melanin particles from adjacent pigment-pro-
ducing cells.

The regeneration process and recurrence of
amelanotic melanoma in the dorsal fins of albino
hybrids are essentially similar to these processes
in dark hybrids with typical black melanomas.
However, the cellular details may be more easily
seen in sections of amelanotic melanomas and
compared directly with bleached sections of
ordinarv melanoma.

j

The epidermis of a 20-day-old regenerated
fin that had had an amelanotic melanoma is
hyperplastic and many collagenous fibres are

present. Text-figs. 37 and 38. The connective
tissue and capillaries may be seen, as well as
tumor cells, which remain below the amputated
fin, in scale pockets and under the epidermis
between the collagenous fibres. Text-figs. 39
and 40. The manner of infiltration by tumor
cells into the connective tissue of a regenerated
fin of the hybrid shown in Text-fig. 7 may be
seen in Text-fig. 41. Some connective tissues
and collagenous fibres which had formed in the
regenerated fin are redestroyed and the fibro-
blasts are hypertrophic. Text-fig. 42. The nodu-
lar melanoma that developed in three months
in the albino hybrid, shown in Text-fig. 8, when
sectioned revealed the following details. The
epidermis covering the overgrowth is almost
normal but the vascularity of the amelanotic
tumors is apparently greater than in the typical
melanoma. Text-fig. 43 and Plate III, Fig. 6.
Only three irregularly aligned dorsal fin rays
are present in the tumor, the others being de-
stroyed, Text-fig. 43. The connective tissue be-
tween the rays is replaced by the tumor cells,
Plate III, Fig. 7. The trunk musculature at the
base of the regenerated fin is separated from
the melanoma by a layer of collagenous fibres.
Text-fig. 43. In some areas this layer is de-
stroyed by amelanotic melanoma cells which
then infiltrate the musculature, Plate IV, Fig. 1 .

The amelanotic melanomas in regenerated
fins have a sarcomatous appearance histolog-
ically. The principal tumor cells are the amel-
anotic melanocytes and amelanotic macromel-
anophores. Some of these cells migrate into the
regenerated fins from tumor cells below the
fins, others are formed in situ. The colorless
melanocytes and macromelanophores form syn-
cytial masses in which cell boundaries are in-
determinate. Their processes are sometimes
parallel to each other; some form swirls, as

Text-fig. 40. Tumor cells surrounded by collagenous fibers at the base of a regenerated amelanotic
melanoma, refer to Text-fig. 37.

Text-fig. 41. Cross-section of a regenerated amelanotic melanoma in the dorsal fin after 3 months, shown
in Text-fig. 7f. Amela— Amelanotic melanoma cells of the body invading the regenerating fin. Co.f—
Collagenous fibers. E— Epidermis. Lep— Lepidotrichia. Sc— Scale.

Text-fig. 42. Same as Text-fig. 41 but under higher power of magnification. Amela— Amelanotic mela-
noma cell from below the dorsal fin now in fin. Co.f— Collagenous fibers partly destroyed by tumor
cells. Be— Blood capillary.

Text-fig. 43. Cross-section of amelanotic melanoma that developed in the dorsal fin of an albino hybrid
within three months following amputation, see Text-fig. 8f. Amela-Amelanotic melanoma. Be-Blood
capillary. E— Epidermis. Lep— Lepidotrichial element.

Text-fig. 44. Two types of tumor cells in amelanotic melanoma, refer to Text-fig. 8f. a— Cell with homo-
geneous cytoplasm, homologous to the pigmented hypertrophic fibroblast found in melanotic melanomas,
b— Cell with radial fibrillar cytoplasm.

Text-fig. 45. Giant cell with three nuclei in the amelanotic melanoma shown in Text-fig. 43.

76

Zoologica: New York Zoological Society

[40; 6

shown in Plate III, Fig. 7, and Plate IV, Figs.
1 to 3. The cytoplasm of the macro melano-
phores has a fibrillar appearance. The melano-
cytes show mitotic figures, whereas the macro-
melanophores show amitotic ones. The cyto-
plasm of the non-pigmented hypertrophic fibro-
blasts in amelanotic melanomas is homogeneous
and stains more lightly with hemalum-eosin
than the other tumor cells. Text-fig. 44a. The
non-pigmented giant cells have lobulated nu-
clei; some may be multinucleate, their nuclei
dividing amitotically. Text-fig. 45, Plate IV,
Fig. 4. Certain polymorphic cells (30 to 50
micra in size) in the amelanotic tumor have
their counterparts in the pigmented hypertrophic
fibroblasts of the typical melanoma. These poly-
morphic cells have fibrillar structures that radi-
ate from a center (which can not be seen in
the black melanoma). Text-fig. 44b and Plate
IV, Fig. 5. Ermin (1946) described these pig-
mented hypertrophic fibroblasts as “sekundar”
pigment cells which have one or more nuclei
that measure 9 to 10 micra. Only the distal
surface of the amelanotic melanoma was ne-
crotic. Various types of cell degeneration, such
as pyknosis and karyorrhexis, were observed.
Sections of amelanotic melanoma, when com-
pared with bleached sections of melanotic ones,
reveal the fact that both types of tumor have
almost the same fibrillar structure. Compare,
for example, Plate II, Fig. 4, with Plate IV,
Figs. 4 to 5. The lacunae described by Breider
(1938) in gray melanomas on the bodies of
“black albinos” and by Levine (1948) in amel-
anotic melanomas on the body proper— both
of which are essentially the same— were not seen
in the dorsal fin melanomas.

Finally, the amelanotic melanomas have some
coarse granulocytes “in a discharging state”
which are like those Catton (1951) described
in various normal fishes and Aronowitz, Ni-
grelli & Gordon (1951) found in a spontaneous
epithelioma in the platyfish Xiphophorus vari-
atus.

Discussion

For studies of regeneration of the basic tis-
sues of the dorsal fins and their pigmentation
following amputation, certain fishes were chosen
from many genetic stocks with a view of tracing
the histories of two kinds of pigment cells, the
micro- and macromelanophores. At first, fishes
with micromelanophore patterns were studied;
of these there were three kinds : ( 1 ) Wild type,
St -b +; (2) Wild type with a comet pattern,
St Co +; (3) Wagtail, St Co E, see Table 1.

It should be noted that the micromelanophore
pigmentation of the dorsal fin of the first two
types is similar because the comet pattern is

restricted to the caudal fin. In both dorsal and
caudal fins of the wagtail, however, the number
of these small pigment cells is much greater.
Following amputation of the dorsal fins in all
three types, the reformation of the fins and
pigmentation requires about one month.

The sources of micromelanophores in the
growing blastema and in the regenerated dorsal
fin are: (1) from pre-existing melanophores
on the base of the amputated fin and on the
dorsal ridge of the body, just below the dorsal
fin; and (2) from melanoblasts (or melano-
cytes) which come in with other cells to re-
establish the dorsal fin, these pigment cells
transforming in situ into melanophores.

This interpretation of the sources of melano-
phores in regenerated fins of fishes is essentially
the same as those suggested by other observers,
specially Bosenberg (1938), Goodrich & Nich-
ols (1931), Wunder & Schimke (1935), Grimm
(1949), Wunder (1951), Goodrich, Hine &
Reynolds (1950), Goodrich & Bresinger (1953)
and Goodrich, Mazullo & Bronson (1954),
based on work on various species of freshwater
and marine fishes. Bosenberg suggested that the
melanophores were not migratory hut were
carried along with other cells into the regenerat-
ing fins. Goodrich, et al (1954), however, de-
clared that some melanophores may enter as
propigment cells and differentiate later; they
are usually first observed as lightly pigmented
cells having the migratory or ameboid form.
Marcus & Gordon (1954) traced the movements
of melanocytes in melanoma transplants and
found that some melanocytes, after they trans-
formed into melanophores, ceased moving.

The neural crest origin of pigment cells in
lampreys and fishes was first suggested hy Bor-
cea (1909) and has been supported by Weiden-
reich (1912), Lopashov (1944),Newth (1951)
and Orton (1953). But there may be other,
as yet indefinite, sources of these cells, according
to Oppenheimer (1950) and Goodrich (1950).

In the second series of experiments, the his-
tory of the restoration of the macromelano-
phore pattern in dorsal fins was studied in two
genetic strains of platyfish: (4) Comet with
spotted-dorsal, St Co -f Sd and (5) Wagtail
with spotted-dorsal, St Co E Sd. These were
produced by intermating platyfish from two
different geographical populations. The result
in the next generation of this intermating, as
shown by Gordon (1951a), is an increase in
the intensity of macromelanophore pigmenta-
tion which reaches a point of atypical growth,
that is, a low degree of melanosis. The spotted-
dorsal fish with the comet and wagtail patterns
were essentially similar in this respect. Some-
what similar also were the platyfish-swordtail

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

77

hybrids with various degrees of melanosis, item
7 of Table 1, Sd.

The restoration of pigmentation after ampu-
tation of the dorsal fin in Sd fishes depended
primarily upon the state of original melanosis.
For example, if the original melanosis was rela-
tively light, then the regenerated fin was less
pigmented. This confirms preliminary studies
of this problem made by Goldsmith, Gordon &
Nigrelli (1947). They found that the melanotic
dorsal fins of five-month-old unoperated control
platyfish-swordtail hybrids had more macro-
melanophores than the regenerated fins of 11-
month-old sibling hybrids. They suggested that
the difference might lie in the fact that the tis-
sues comprising the regenerated fins are chrono-
logically younger than those in the controls. In
the present observations, if the dorsal fin orig-
inally had a strong melanosis, regeneration of
pigmentation was more rapid and reached al-
most equal intensity after an 8V2 month period.
Younger hybrids had the capacity to develop
the original state of melanosis more rapidly
than older fish. This is interesting because Scott
(1907) discovered that in Fundulus hetero-
clitus the regeneration rate of normal fins is
greater in younger than in older fishes which,
he believed, is in line with the theory that re-
generation is a growth phenomenon.

In the reformation of the state of melanosis
in the regenerated fins, one source of the pig-
mented cells, including the specific melanosis-
producing macromelanophores, was through
the migration of large melanophores from then-
position below the dorsal fin. Silber (1951)
also found that macromelanophores entered
the dorsal fin blastema of platyfish-swordtail
hybrids from “pigment cell depots” located at
the base of the fin. We have found a second
source of macromelanophores in the regenerated
fin, namely, that they are formed in situ from
melanocytes. Recently Marcus & Gordon
(1954) have also found that some melanocytes
present in a transplanted melanoma transform
into macromelanophores in host tissues.

After amputation, the reformation of a mela-
noma or an amelanotic melanoma in the dorsal
fin of Sd hybrids (list in item 8, Table 1) de-
pends upon the degree of tumor involvement
not only of the original dorsal fin but of the
body just below that fin. If the tumor develop-
ment in the body below the fin is pronounced,
the regenerated melanoma or amelanotic mel-
anoma often exceeds the size of the amputated
tumor, sometimes almost by three times. If
there is little tumor involvement of the body
below the fin, the reformed dorsal fin melanoma
usually does not exceed that of the original
tumor.

It is an interesting fact that the progressive
growth of a melanoma may be halted tem-
porarily by amputation of most of its tissues.
Immediately after the operation, apparently
there is sufficient normal tissue available in the
stump of the dorsal fin to recreate the whole
fin. Subsequently, the tumorous tissue that has
not been removed grows, invades and destroys
the newly regenerated tissues.

In the process of the reformation of the
melanoma following amputation, the same se-
quences of tumor development were found as
in the formation of the original melanoma. In
hybrid fishes, the development of spontaneous
and of regenerated melanomas is preceded by
a premelanomatous state of melanosis in spe-
cific areas. In fishes carrying the sex-linked,
dominant gene, Sd, the dorsal fin is the site of
the melanosis, a condition brought about by
the rapidly proliferating macromelanophores.

Gordon (1951a) pointed out that in pure
platyfish, macromelanophores appear in the
dorsal fin in response to the presence of the Sd
gene only after three to five months, whereas
in inter-racial hybrid platyfish with the Sd gene,
these large pigment cells may appear in two
weeks. In platyfish-swordtail, inter-specific hy-
brids, the macromelanophores in the Sd fish
appear still earlier, some on the day of birth.
Gordon (1948, 1950b) suggested that the rate
of macromelanophore proliferation is accel-
erated in proportion to the strength and fre-
quencies of other genes that modify pigment
cell growth.

The atypical growth of these large pigment
cells leads to a state of melanosis which may
be destructive to adjacent normal tissues. The
latter may be destroyed and replaced by them
(Gordon & Smith, 1938). In the melanosis
produced in platyfish-swordtail hybrids, the con-
centration of macromelanophores is so great,
that practically no other types of pigment cells
can be distinguished. If the melanosis appears in
the dorsal fin, as it does in Sd fishes, the fin may
be destroyed at various levels, but there is no
swelling of tissues. Gradually this phase changes
and there appears a noticeable swelling. When
sections are cut through the swollen areas and
studied histologically, they reveal a significant
change in cellular components. The outstanding
feature of the new growth, as Reed & Gordon
(1931), Gordon & Smith (1938) and Grand,
Gordon & Cameron (1941) (by tissue cul-
tures) have pointed out, is the preponderance of
melanocytes and the relatively small number of
macromelanophores .

The importance of understanding the transi-
tional steps, from the appearance of macro-
melanophores in a genetically susceptible an-

78

Zoologica: New York Zoological Society

[40: 6

iraal, through their hyperplastic growth to the
formation of a melanosis and, finally, to the
development of a definitive melanoma, has been
appreciated in former studies (Gordon, 1951b).
But until the present work on regeneration and
on transplantation (Marcus & Gordon, 1954),
the cellular elements involved and their rela-
tionships to each other could not be properly
evaluated.

There are two categories of pigment-carrying
cells in the melanoma. One group contains those
cells which not only carry melanin granules
but are capable of synthesizing melanin pig-
ment. This group includes melanocytes and
melanophores, both large (macro.) and small
(micro.). Pigment cells of the second group
do not synthesize the few or many melanic
granules that they carry. This group includes
pigmented hypertrophic fibroblasts, pigmented
giant cells, pigmented connective tissue stroma
cells and pigmented macrophages (melano-
phages). These secondary pigment cells acquire
their pigment by contact with cells of the first
category through various processes.

In the transition from the state of melanosis
to that of melanoma, the macromelanophores
in their atypical growth form a dense syncytial
mass in which their dendritic processes anasto-
mose. Sometimes the melanophore processes
are parallel, sometimes they form swirls. The
cell membranes are indeterminate. The fibro-
blasts in contact with and in response to the
progressive atypical growth of the macromelan-
ophores, become hypertrophic and pigmented,
possibly by cytocrine activity on the part of
the dendritic, pigment-forming melanophores.
Although we utilize Masson’s (1948) concept
of cytocrine activities, our use of the term mel-
anophore is not the same as his; Gordon (1953)
pointed out that many human pathologists have
used the term melanophore to denominate what
biologists call macrophage. The primary pig-
ment cells, melanocytes and melanophores, have
the property of liberating some of their melanin
by clasmatosis (Grand, Gordon & Cameron,
1948). Melanin particles so released may be
picked up by adjacent cells in the tumor, such
as the fibroblasts, giant cells and other connec-
tive tissue cells. The pigmented hypertrophic
fibroblasts may divide amitoticaUy. These cells
may lose their nuclei by the process of
karyolysis. In some areas of the melanoma the
fibroblasts form large oval bodies in which the
acquired melanin is both peripheral and central.
Similar pigmented bodies have been seen by
Breider (1938, 1939, a, b), Ermin (1946) and
Levine ( 1948) . In another variation of the dorsal
fin melanoma which externally appeared nodu-
lar, we have found macromelanophores of vari-

ous configurations and with fine dendritic
processes within a dense fibrillar connective tis-
sue stroma.

The amelanotic melanomas, listed as 9 in
Table 1, do not differ fundamentally from the
typical melanomas except, of course, in the
amount of melanin contained in the various
cells. The details presented by Levine (1942)
for amelanotic melanomas on the bodies of
platyfish-swordtail hybrids have also been found
to hold for those of the dorsal fin, except that
the latter have a more prominant fibrillar net-
work.

A comparison of the progressive stages in
the regeneration of teleost fins with their nor-
mal development— based on the observations of
many authors from Ryder (1885) and Harrison
(1893, 1895) to Okado (1943) and Blanc
(1949), and on the present studies— reveals
that the regeneration process is essentially a
repetition of the normal ontogenetic process.
This similarity also holds for the restoration of
the normal pigmentation patterns of the fins.
Moreover, results obtained through the ampu-
tation of abnormal fins in a state of melanosis
or with melanoma, show that here, too, the
fundamental repetitive ontogenetic processes
are evident.

It is well known that in the normal develop-
ment of the teleost fin, temperature and other
exogenous factors influence the nature of the
growth process. Higher temperatures during
certain critical development stages, for exam-
ple, result generally in a smaller number of fin
rays in the adult, and lower temperatures pro-
duce a higher count of rays, according to Hubbs
(1922), Gabriel (1944), Taning (1952) and
others. Recently Buser-Lahaye (1953) sug-
gested that external influences (such as tem-
perature and light) are not applied directly in
the regeneration processes but are mediated
through the endocrine glands among which the
thyroid has a special role. In addition, although
the part that nerve cells have been shown to
play in regeneration in the amphibia has not
been evaluated in fishes, it is probable that these
cells are equally influential in teleosts. Indeed,
the failure of some of the fin rays to regenerate
in certain fishes may possibly be attributed to
this factor.

Melanomas and other pigmented tumors in
man and the normal pigmentation of the human
body have been variously interpreted by pathol-
ogists with regard to cellular components and
their embryological origin. This is evident by
reading the more recent statements of Willis
(1948), Masson (1951), Ito (1951), Becker
(1948, 1953), Raven (1953) and Allen & Spitz
(1953, 1954).

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

79

The subject is too involved for review here,
but it seems to us that no discussion of pigment
cells is complete without reference to Dawson’s
(1925) remarkable studies on human mela-
nomas. We have found his well illustrated analy-
sis of the progressive growth of human cutane-
ous melanomas most instructive, because he not
only described the disease in its final, definitely
pathological phase, but also he traced its onto-
logical development from its earliest, appar-
ently innocuous state. Studied by this dynamic
method, the progressive history of human cu-
taneous melanoma shows striking parallels to
those of fish. In comparing these histories it
must be remembered that Dawson’s term mel-
anophore is equivalent to the biologists’ macro-
phage. No true melanophores are found in mam-
mals; melanophores are specialized effector cells
characteristic of fishes, amphibians and reptiles.
One of the important findings from studies of
regeneration and transplantation of fish melan-
omas is that the true melanophore is related di-
rectly to the melanocyte and melanoblast. Thus
it may be re-emphasized that the fundamental
cell type in the melanomas of all vertebrate
animals is the melanocyte.

Summary

1. The variously pigmented dorsal fins of
fifty-three young and adult platyfish (Xipho-
phorus maculatus) and platyfish-swordtail hy-
brids (X. maculatus-X. helleri) were amputated.
Their regenerated fins and pigmentary patterns
were studied histologically.

2. It was possible to compare the results of
amputation and regeneration of normally pig-
mented dorsal fins with those that were either
in a state of melanosis or exhibited melanomas.
Among the latter, it was also possible to compare
the regeneration process in those that had typical
black melanomas with those that had amelanotic
melanomas.

3. All of the amputated dorsal fins regenerated
in about two to three months, with essentially the
same pigmentary pattern they showed originally.
In some of the fishes with melanomas, regenera-
tion was abnormal but it was not necessarily im-
peded by the simultaneous growth of a mel-
anoma.

4. The restoration of melanosis in the regen-
erated dorsal fins was faster in young hybrids
than in mature ones. The state of melanosis was
incomplete in the regenerated fins of those fish
in which the melanosis was confined to the fins
alone. Melanosis was more complete in the re-
generated fins of those fish that had melanosis in
the fin and on the body below the dorsal fin as
well.

5. The reformation of a melanoma in regen-
erated dorsal fins was more rapid in hybrids that
originally showed a melanoma both in the dorsal
fin and on the body ventral to the fin than in
hybrids that had a melanoma in the fin alone.

6. The regeneration process and the recur-
rence of amelanotic melanomas in the ampu-
tated dorsal fins of albino hybrid fish were es-
sentially similar to those in hybrids with typical-
ly black melanomas. The progressive growth of
the melanomas was halted but only temporarily
by amputation.

7. The same sequences were found in the re-
formation of the remissive melanomas that were
observed in the development of the original
melanomas.

8. After the dorsal fin was amputated, squa-
mous cells of the epithelium moving from both
sides of the wound grew over its surface and re-
formed the epidermis. Basal cells of the adjacent
tissues moved under the squamous epithelial cell
layer.

9. Melanocytes and micromelanophores ap-
peared in the blastema before the macromelano-
phores. These are all true pigment cells that
produce the melanin they carry.

10. Melanophores were derived from two
sources: from normal pigmented areas immedi-
ately below the regenerating fin and from mel-
anocytes that develop in situ in the blastema.

1 1 . Some of the pigmented hypertrophic fibro-
blasts in the melanoma acquired their pigment
from contact with macromelanophores either
through the process of clasmatosis, cytocrine ac-
tivity or both. This is also true of the pigmented
connective tissue cells of the stroma and of the
giant cells.

12. The regenerated amelanotic melanoma
has a sarcomatous appearance histologically, as
does the original tumor. The principal cells of
the amelanotic melanoma are the amelanotic
melanocytes and amelanotic macromelano-
phores.

Acknowledgements

We wish to thank James W. Atz for reading
the manuscript, and Cafer Turkmen for aid in
making the photomicrographs.

References Cited
Allen, A. C., & Sophie Spitz

1953. Histogenesis and clinicopathologic corre-
lation of nevi and malignant melanomas.
Amer. Med. Assn. Archives Derm, and
Syphil. 69: 150-171.

1954. Malignant melanoma. Cancer, 6 (1):
1-45.

80

Zoologica: New York Zoological Society

[40: 6

Aronowitz, O., R. F. Nigrelli & Myron Gordon

1951. A spontaneous epithelioma in the platy-
fish, Xiphophorus (Platypoecilus) varia-
tus. Zoologica, 36: 239-242.

Attardi, G. & F. Moro

1953. La coltivazone “in vitro” di melanomi
maligni della coroide. Arch. “De Vecchi;”
per L’Anatomia Patologia e La Medicina
Clinica, 20 (1): 125-180.

Becker, Wm.

1948. Dermatological investigations of melanin
pigmentation. In Biology of Melanomas,
Spec. Publ. N. Y. Acad. Sci., 4: 82-125.

1953. Microscopic analysis of normal melano-
blasts, nevus cells and melanoma cells. In
Pigment Cell Growth: 109-120. New
York, Academic Press, Inc.

Blanc, M.

1949. :fitude histologique de la regeneration des
nageoires chez quelques poissons teleo-
steens. Arch. Anat. Microsc., 38: 52-64.

Borcea, M. I.

1909. Sur I’origine du coeur, des cellules vascu-
laires migratrices et des cellules pigmen-
taires chez Teleosteens. Comp. Rend.
Acad. Sci. Paris, 149: 688-689.

Bosenberg, K.

1938. Die Auspigmentierung von Schwanzflos-
senregeneration bei Macropodus opercu-
laris L. (Teleost). Zeits. f. Zellforsch. u.
mikr. Anat., 27: 14-45.

Breider, H.

1938. Die genetischen, histologischen und zy-
tologischen Grundlagen der Geschwulst-
bildung nach Kreuzung verschiedener
Rassen und Arten lebendgebarender Zahn-
karpfen. Zeits. f. Zellforsch. u. mikr.
Anat., 28: 784-828.

1939a. tiber die Vorgange der Kernvermehrung
und -degeneration in sarkomatosen Mak-
romelanophoren. Zeits. f. wiss. Zool., 152:
89-106.

1939b. fiber die Pigmentbildung in den Zellen
von Sarkomen albinotischer und nichtal-
binotischer Gattungsbastarde lebendge-
barender Zahnkarpfen. Zeits. f. wiss. Zool.,
152: 107-128.

Buser-Lahaye, J.

1953. fitude experimentale du determinisme de
la regeneration des nageoires chez les
poissons Teleosteens. Anales de I’lnst.
Oceanographique, Nouvelle Serie, 28:
1-61.

Catton, W. T.

1951. Blood cell formation in certain teleost
fishes. Blood, 6: 39-60.

Dawson, J. W.

1925. The melanomata, their morphology and
histogenesis. Edinburgh Medic. Jour., New
Sen, 32: 501-732.

Ermin, R.

1946. fiber erblich begingte Tumoren bei
Eischen (Histologie und Cytologie von
Melanomen und Erythrophorengeschwuls-
ten). Rev. Fac. Sci., Univ. Istanbul,
Sen B, 11: 147-179.

Ermin, R., & Myron Gordon

1952. Study of the regeneration processes in
fishes after amputations of dorsal fins with
and without melanotic tumors. Cancer
Res., 12: 260-261 (Abstract).

Eitzpatrick, T. B., a. B. Lerner, E. Calkins &

W. H. SUMMERSON

1950. Occurrence of tyrosinase in horse and fish
melanomas. Proc. Soc. Exp. Biol, and
Med., 75: 394-398.

Gabriel, M. L.

1944. Factors affecting the number and form
of vertebrae in Fundulus heteroclitus.
Jour. Exp. Zool., 95 (1): 105-147.

Goldsmith, E. D., Myron Gordon &

R. F. Nigrelli

1947. The effect of amputation of the dorsal fin
upon the development of Sd melanomas
in hybrid fishes. Anat. Rec., 97: 388-389
(Abstract).

Goodrich, H. B.

1950. Problems of origin and migration of pig-
ment cells in fish. Zoologica, 37: 171
(Abstract).

1951. Histology and regeneration of color pat-
terns in certain tropical marine fish. Anat.
Rec., Ill: 53 (Abstract).

Goodrich, H. B., Ruth L. Hine & Jack Reynolds

1950. Studies on the terminal phases of color
pattern formation in certain coral reef fish.
Jour. Exp. Zool., 114 (3): 603-626.

Goodrich, H. B., & D. I. Bresinger

1953. The histological basis of color patterns
in three tropical marine fish with observa-
tions on regeneration and the production
of the blue color. Jour. Morph., 93 (3):
465-488.

Goodrich, H. B., Carmela M. Marzullo &

W. R. Bronson

1954. An analysis of the formation of color
patterns in two freshwater fish. Jour. Exp.
Zool., 125 (3): 487-506.

Goodrich, H. B., & R. Nichols

1931. The development and the regeneration of
the color pattern in Brachydanio rerio.
Jour. Morph., 52: 513-518.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

81

Gordon, Myron

1948. Effects of five primary genes on the site
of melanomas in fishes and the influence
of two color genes on their pigmentation.
In Biology of Melanomas, Spec. Publ. N.
Y. Acad. Sci., 4: 216-268.

1950a. Fishes as laboratory animals. In Care and
Breeding of Laboratory Animals, John
Wiley and Sons, New York, 14: 345-449.

1950b. The origin of modifying genes that in-
fluence the normal and atypical growth
of pigment cells in fishes. Zoologica, 35
(1): 19-20.

1951a. The variable expressivity of a pigment cell
gene from zero effect to melanotic tumor
induction. Cancer Res., 11: 676-686.

1951b. Genetic and correlated studies of normal
and atypical pigment cell growth. Growth,
10 (Supplement): 153-219.

1953. Preface. In Pigment Cell Growth, Aca-
demic Press, Inc., New York.

Gordon, Myron, & W. Lansing

1943. Cutaneous melanophore eruptions in
young fishes during stages preceding mela-
notic tumor formation. Jour. Morph., 73:
231-245.

Gordon, Myron, & G. M. Smith

1938. Progressive growth stages of a heritable
melanotic neoplastic disease in fishes from
the day of birth. Amer. Jour. Cancer, 34:
255-272.

Grand, C. G., Myron Gordon & G. Cameron

1941. Neoplasm studies. VIII. Cell types in
tissue culture of fish melanotic tumors
compared with mammalian melanomas.
Cancer Res., 1: 660-666.

Grimm, Hans

1949. Die “Berliner Zuchtform” des Schwert-
tragers (Xiphophorus helleri Heckel) als
giinstiges Objekt fiir Regenerationsstudien.
Deuts. Aquar.-Terrar.-Zeits. 2 (11): 186-
187.

Harrison, R. G.

1893. tJber die Entwicklung der nicht knorpelig
vorgebildeten Skeletteile in den Flossen
der Teleostier. Arch, mikrosk. Anat., 42:
248-278.

1895. Die Entwicklung der unpaaren und
paarigen Flossen der Teleostier. Arch,
mikrosk. Anat., 46: 500-578.

Hubbs, C. L.

1922. Variation in the number of vertebrae and
other meristic characters of fishes corre-
lated with temperature of water during
development. Amer. Nat., 56: 360-372.

Ito, Minor

1952. Studies on melanin. Tohoku Jour. Exp.
Med., 55 (Suppl.): 1-104.

Levine, M.

1948. The cytology of the typical and the
amelanotic melanoma. In Biology of
Melanomas, Spec. Publ. N. Y. Acad. Sci.,
4: 177-215.

Lopashov, G. V.

1944. Origin of pigment cells and visceral car-
tilage in the teleosts. Compt. Rend. (Dok-
lady) Acad. Sci. U. S. S. R., 44: 169-172.

Marcus, T. R., & Myron Gordon

1954. Transplantation of the Sc melanoma in
fishes. Zoologica, 39 (10): 123-131.

Masson, P.

1951. My conception of cellular nevi. Cancer,
4: 9-38.

Newth, D. R.

1951. Experiments on the neural crest of the
lamprey embryo. Jour. Exp. Biol., 28:
247-260.

Okada, Yo, K.

1943. Regeneration of the tail in fish. Annota-
tiones Zool. Japonenses, 22: 59-68.

Oppenheimer, J. W.

1950. Atypical pigment cell differentiation in
embryonic teleostean grafts and isolates.
Zoologica, 35: 22-23.

Orton, G. L.

1953. Development and migration of pigment
cells in some teleost fishes. Jour. Morph.,
93: 69-96.

Raven, R. W.

1953. Problems concerning melanoma in man.
In Pigment Cell Growth; 121-138. Aca-
demic Press, Inc., New York.

Reed, H. D., & Myron Gordon

1931. The morphology of melanotic over-
growths in hybrids of Mexican killifishes.
Amer. Jour. Cancer, 15: 1524-1546.

Ryder, J. A.

1885. The development of the rays of osseous
fishes. Amer. Nat., 19: 200-204.

Scott, G. G.

1907. Further notes on the regeneration of the
fins of Fundulus heteroclitus. Biol. Bull.,
12:385-400.

SiLBER, D.

1951. Regeneration of melanotic and amelano-
matous dorsal fins of platyfish-swordtail
hybrids. Manuscript deposited in Genetics
Laboratory of the New York Zoological
Society.

Taning, a. V.

1952. Experimental study of meristic characters
in fishes. Biol. Rev., 27 (2): 169-193.

82

Zoologica: New York Zoological Society

[40: 6

Weidenreich, F.

1912. Die Lokalisation des Pigmentes und ihre
Bedeutung in Ontogenie und Phylogenie
der Wirbeltiere. Zeits. f. Morph, u. Anth.,
Sonderheft., 2: 59-140.

Willis, R. A.

1948. Pathology of tumors. 1st Ed., Butterworth
and Co., London.

WUNDER, W.

1951. tJber Degeneration und Regeneration der
schwarzen Pigmentzellen in der Haut des
Karpfens {Cyprinus carpio L.). Verb.
Deut. Zool. Ges., Zool. Anz., 15. Suppl.
Bd.: 100-105.

WUNDER, W., & H. SCHIMKE

1935. Wundheilung und Regeneration beim
Karpfen. Arch. f. Entw.-mech., 133: 245-
268.

EXPLANATION OF THE PLATES

Photomicrographs of sections stained either with
hemalum-eosin or by Masson’s method.

Plate I

Fig. 1. The regenerated epidermis five days fol-
lowing amputation of dorsal fin of a
comet platyfish. The blastema cells are
below the epidermis. (Corresponds to
Text-fig. 2b, an earlier blastema stage).
240 X.

Fig. 2. Bud formed of blastema cells pushes the
regenerated epidermis upward. (Corre-
sponds to Text -fig. 2b). 600 X.

Fig. 3. A melanocyte (Me) in a one-week-old
regenerating dorsal fin. (Corresponds to
Text-fig. 2b). 1000 X.

Fig. 4. Epidermis four days following amputation
of the dorsal fin that exhibited melanosis.
The collagenous fibers below the epider-
mis and the epidermis cells surround free
melanin masses. (Corresponds to Text-figs.
4b, c). 300 X.

Fig. 5. The elimination of pigment particles
through the epidermis of a nine-days-old
regenerating dorsal fin of a platyfish-
swordtail hybrid that had a melanosis.
(Corresponds to Text-fig. 4d, at an earlier
stage). 600 X.

Fig. 6. The regenerating epidermis 24 hours fol-
lowing amputation of a dorsal fin of a
hybrid that had a melanoma in the fin
and in tissues below the fin. The remains
of one fin ray may be seen within the
melanoma tissue. (Corresponds to Text-
fig. 6b at an earlier stage). 100 X.

Plate II

Fig. 1. A bleached section through the regener-
ating dorsal fin shown in PI. I, Fig. 6.
100 X.

Fig. 2. Part of the melanoma in the ventral re-
gion of a regenerated dorsal fin of an Sd
hybrid with melanoma. Numerous pig-
mented hypertrophic fibroblasts are pres-
ent in the tumor. (Corresponds to Text-
fig. 6e). 440 X.

Fig. 3. Section showing parallel arrangement of
dendritic processes of macromelanophores
in the dorsal fin of an Sd hybrid with mel-
anoma. (Corresponds to Text-fig. 6f).
440 X.

Fig. 4. A bleached section of melanoma of the
same specimen showing swirl-like arrange-
ment of processes of macromelanophores.
440 X.

Fig. 5. Pigmented hypertrophic fibroblasts and a
syncytium of macromelanophore pro-
cesses in a bleached section of a regener-
ated melanoma of an Sd hybrid that had a
melanoma in its dorsal fin. (Corresponds
to Text-fig. 6f). Ma— Macromelanophore.
P.h.f. — Pigmented hypertrophic fibro-
blast. S.C.— Stroma cell. 1000 X.

Fig. 6. Part of regenerated tumor after 2.5 months
in an Sd hybrid that exhibited melanoma.
(Corresponds to Text-fig. 6f). 600 X.

Plate III

Fig. 1. Bleached section of the three-week regen-
erated dorsal fin of an Sd hybrid showing
a macromelanophore that has migrated
from the base of the fin. (Corresponds
to Text-fig. 6d). 600 X.

Fig. 2. Section of the nodular melanoma in the
10-month regenerated dorsal fin of an Sd
hybrid showing macromelanophores de-
veloped in the tumor. (Corresponds to
Text-fig. 5f). 600 X.

Fig. 3. Nodular melanoma reformed in the pos-
terior part of the same hybrid as shown in
Fig. 2. (Corresponds to Text-fig. 5f).
50 X.

Fig. 4. Same nodular melanoma shown in Fig. 3
under higher magnification showing radial
arrangement of processes of macromela-
nophores. 100 X.

Fig. 5. Part of a bleached section of the same
nodular melanoma shown in Fig. 3.
600 X.

1955]

Ermin & Gordon: Regeneration of Melanomas in Fishes

83

Fig.

Fig.

Fig. 1

Section of amelanotic melanoma that had
regenerated in the dorsal fin of an albino
(Sd i) hybrid. Note the hyperplastic blood
vessels. (Corresponds to Text-fig. 8f).
100 X.

Section of amelanotic melanoma of an
albino hybrid showing the replacement
of the connective tissue between the lepi-
dotrichia by tumor cells. (Corresponds to
Text-fig. 8f). 600 X.

Plate IV

Section of amelanotic melanoma that re-
developed in the regenerating dorsal fin
of an albino {Sd i) hybrid, showing the

penetration and destruction of collage-
nous fibers at the base of the fin. (Cor-
responds to Text-fig. 8f). 600 X.

Fig. 2. Section of an amelanotic melanoma show-
ing amelanotic macromelanophores in the
tumor tissue. (Corresponds to Text-fig.
8f). 950 X.

Fig. 3. Another part of the amelanotic melanoma
showing fibrillar structures in the tumor
tissue. 600 X.

Fig. 4. Two giant cells in an amelanotic melano-
ma. (Corresponds to Text-fig. 8f). 950 X.

Fig. 5. Amelanotic melanoma showing cell (A)
with radiating fibrillar structures. (Cor-
responds to Text-fig. 8f). 440 X.

ERMIN a GORDON

PLATE I

FIG. 1 fig. 2

REGENERATION OF MELANOMAS IN FISHES

ERMIN a GORDON

PLATE II

FIG. 3 FIG. 4

FIG. 5

FIG. 6

REGENERATION OF MELANOMAS IN FISHES

ERMIN a GORDON

PLATE III

FIG. 1 fig. 2 p,g 3

FIG. 4

FIG. 5

FIG. 6

FIG. 7

REGENERATION OF MELANOMAS IN FISHES

ERMIN a GORDON

PLATE IV

FIG. 1

FIG. 2

FIG. 5

'A ’ .s'. -:’ ' V- • >■':' #

FIG. 3

■:4

REGENERATION OF MELANOMAS IN FISHES

7

Further Notes on the Pigmentary Behavior of Chaetodipterus in
Reference to Background and Water Transparency

C. M. Breder, Jr., & Priscilla Rasquin
The American Museum of Natural History, New York 24, N. Y.

(Plates I & II)

IT has been shown by Breder (1946) that
very small individuals of Chaetodipterus
faber (Broussonet) appear in a coal black
stage under certain conditions. When viewed
against a light sand background on which
similar-sized black gastropods and black infertile
pods of the red mangrove were scattered, the
small fish effectively vanished from sight. This
black color phase was seen in fishes which were
of about 10 mm. in length and which always
inclined to one side. At larger sizes they were
found to show the characteristic black and white
vertical bars and upright position. The above
observations were made on the west coast of
Florida.

Subsequently, observations differing from
these were made in the Bahamas, at the Lerner
Marine Laboratory (Breder, 1948 and 1949).
Larger-sized fishes, up to and including those
a foot or more in length, were found there to
appear not infrequently in a similar black phase.
At such times they always lay on one side on
the bright, light colored sand, appearing much
like a piece of drifting trash, and could be easily
overlooked. In Breder’s 1949 paper it was sug-
gested that the difference between the behavior
of these fishes at the two places might be asso-
ciated with the difference in the transparency of
the water, that on the Florida west coast being
notably turbid while the Bahamas water is re-
markably transparent.

Various experiments were undertaken with
some of these Bahaman fish in an effort to deter-
mine more clearly the basis of the differential
behavior. In aquaria, under the most diverse
conditions, they were found to swim upright
and show at least some traces of the vertical
bars, although the bars were generally less dis-
tinct when a minimum of dark objects was seen

by the fish. This is, of course, in keeping with
what had already been found.

Some of the experiments included presenting
the fishes with variously painted backgrounds,
such as broad dark vertical bars. In one series
of experiments, quite accurate drawings of
Chaetodipterus in groups in the solid black
phase and in the barred phase were exhibited to
a similar-sized test fish. None of these tests
was able to make the fishes, solitary or in
groups, obliterate the vertical bars, or recline as
do the dark ones in the sea against a light sand
background, or show especially vivid bars. It
is to be noted that this is one of the species
which does not show the classical concentration
of melanophore granules in response to adren-
alin (Breder & Rasquin, 1950).

A favorable situation led to the following
clarification of some of the details of the char-
acteristics which an environment must have to
elicit the full expression of this dark coloration
and reclining posture. A single individual about
3 inches long, in the dark phase, found reclining
on its side near the laboratory dock, was trans-
ferred to a shallow circular concrete pool 12
feet in diameter. Here, over a bottom of clean
light sand, the fish continued to perform as it
had in the sea. Plate I, Figure 1, shows its ap-
pearance at this time. Both in this pool and
earlier in the sea it permitted a very close ap-
proach and sometimes netting by the observer,
simply lying very quiescent, but in about half
the approaches it would dart away a short dis-
tance when very closely and persistently pur-
sued. In this pool it normally took a position
somewhere near the center and huddled toward
the light blue walls of the construction.

After being transferred to a small aquarium,
2X1X1 feet, with a similar white sand bottom.

85

86

Zoologica: New York Zoological Society

[40: 7

the fish retained both its position and coloration
as is shown in Plate I, Figure 2. These it held
even when presented with various targets of
black stripes and the figures of banded fishes
which had been used earlier in the tests on other
fishes. It did, however, show light bands but not
very strongly, seemingly as a “fright” reaction
in response to such things as taps on the glass
walls of the tank.

On the coming of night and consequent low-
ering of light intensity the fish always swam
erect and showed its bands. Such a condition
is shown in the photograph in Plate I, Figure 3,
taken by flashlight. Thus this behavior is evi-
dently associated with light intensity, “fright”
and a variety of visual stimuli, the latter perhaps
being the least potent of the three influences.
The puzzling responses studied earlier, wherein
the fishes showed bands under all manner of
aquarium fittings, were thus evidently actually
in response to both fright and lowered light
intensity. In these former experiments the
aquaria were wrapped with paper on at least
three sides and sometimes four, to insure the
fish seeing a minimum of distracting objects,
thus rather drastically reducing the light
intensity.

Inasmuch as light intensity is evidently in-
volved, and the turbidity of the water on the
west coast of Florida reduces underwater illu-
mination as compared with the Bahaman situ-
ation, the following experiment was undertaken.
A suspension was made of pulverized charcoal
and this was poured into the aquarium in broad
daylight when the fish was in a dark, side-resting
condition as shown in Plate I, Figure 2. The
suspension made the water turbid and the
larger particles settled out, transforming the
aquarium bottom from one of nearly white
sand to a mostly dead black condition. Imme-
diately the fish erected itself and displayed the
light, nearly white vertical bars, as is shown
in Plate I, Figure 4, where the fish can be only
faintly seen because of the turbid water. The
following day the water had cleared but the
bottom was largely black and the fish retained
its stripes in bright daylight but over this dark
bottom. This is shown in Plate I, Figure 5.
Light sand was then introduced into the aqua-
rium gently, through a small pipe, in order to
bury the layer of charcoal. As this was late in
the day, an artificial light was arranged so as
to eliminate the effect of the decreasing light
intensity. The fish returned to its black colora-
tion and retained it long after daylight had
faded. This is shown in Plate I, Figure 6. When
this special light was extinguished, the fish then
showed its bands as it had on other nights.

The following day the fish showed its day-

time black color appropriate to a light sand
background. Netting caused bars to reappear.
It was then returned to the circular pool where
it immediately resumed its dark phase and acted
as it had before the aquarium experiments were
undertaken. Late this day a small dark piece
of Sargassum weed was dropped into the pool
and by the following morning it had drifted
to the outlet pipe. The fish was found under it
and beside the black outlet pipe, vertical and
in the strong black and white banded phase.
When the fish was chased away from this
shelter, it immediately obliterated the bands and
reclined as a black object near the center of
the pool on the light sand. It is to be especially
noted that the outlet and inlet pipes in this pool
were vertical and black (hard rubber) but that
no attention had been paid to them by the fish
until a sheltering, and shadow-casting, object
was also associated with them. Similarly, Breder
(1948) observed a large black fish, off the labo-
ratory dock, which drifted slowly along the
bottom on its side until it approached the pilings.
It then became erect and showed its strongest
bars as it swam among the rather thin piles
supporting the dock. At that time it was thought
that it was the sight of the dark vertical lines of
the piles which elicited the response. In the
light of the present series of experiments it
would seem that this behavior is more probably
referable to the shadow of the dock and the
consequent lower light intensity, than to any
definite retinal image.

The bold pattern referred to in this discussion
is not to be confused with minor lightenings of
the light barred areas which flash faintly in an
evanescent manner following all manner of
stimuli, including minor “frights” or the sight
of food. These lesser responses are evident if
watched for closely, and probably have a signifi-
cance analagous to the twitching of a fin which
these, and in fact most fishes, show under similar
situations. They seem to be nothing more than
nervous “starts.” This individual, it must be em-
phasized, was from the first a most tractable
aquarium inmate. An hour after it had been
placed in the aquarium the fish acted as though
it had always been there. It would investigate a
finger outside the glass and it fed freely from this
time on. When the light sand was introduced in
the course of the experiments, the fish butted
and bit at the pipe. It was clearly not nearly so
timid as the fishes examined previously, which
are referred to earlier in these notes.

The above experiments were performed on
this single individual in November, and the fish
was maintained in the laboratory until May
when it was reintroduced to the circular pool
with a light sand background. It responded in

1955]

Breder & Rasquin: Pigmentary Behavior of Chaetodipterus

87

a manner strictly comparable to its former be-
havior and differed only in that the blackening
was not quite so complete and the reclined posi-
tion not so nearly horizontal. This may be asso-
ciated with the greater age of the fish, as there
is considerable evidence that this behavior is
most definite in the smaller sizes. On the other
hand, it may be more truly associated with the
long sojourn in aquaria. It is a commonplace
among aquarists that many fishes after long
residence in aquaria tend to show less vigor in
their various responses than do wild fishes. Such
a slackening of behavior vigor may be associ-
ated with waning health but certainly, in many
cases, it is not so modified. In these instances it
would seem to be more a matter of dropping old
habits and developing new in accordance with
the radical change in environment, brought
about by moving from the open sea to a small
aquarium, with consequent absence of predators
and complete change in the manner in which
food is presented or found.

Whether any or all of the above noted matters
incident to captivity had anything to do with
this slight change in pigmentary behavior is un-
certain but they are mentioned here to indicate
that the authors have not been unmindful of the
possibility of such influences affecting the re-
sults. It is believed, moreover, that in the present
instance the reduced light to which the fish was
exposed in the laboratory aquarium for the
period of some months was sufficient to induce
a considerable reduction in the numbers of
dermal melanophores. It may be noted that
Chaetodipterus lives well, and for years, in
public aquaria but little by little becomes much
lighter. This lightening evidently results in part
at least from the elimination of melanophores
in the comparatively low light intensities of such
places.

Because of the “opposite” pigmentary be-
havior as compared with that of usual back-
ground-matching species, the influence of mel-
anophore-affecting hormones and other sub-
stances was investigated, as has been noted in
passing by Breder & Rasquin (1950). The pre-
cise nature of these experiments is given below.

One specimen which weighed 276 grams and
measured 192 mm. in standard length was in-
jected with 2.8 cc. adrenalin 1:1,000. Through-
out the observation period of four hours the
injected fish remained noticeably darker than
the uninjected control. Two minutes after in-
jection, the injected fish was darker over the
dorsal surface than the control; after six minutes
it was still darker than the control but showed
small white patches in the light bars. Ten
minutes after injection the fish was darker
than the control from the mid-dorsum down to

approximately the lateral mid-line. It remained
definitely gray where the control showed bold
white bars. The black bars of the injected fish
seemed less definitive than those of the control,
although this may have been owing to less con-
trast of color offered by the injected fish. Two
hours after injection the iris of the injected fish
was white and the whole animal was quite dark,
although not sufficient to eliminate the barred
pattern entirely. After three hours the colora-
tion of the iris had returned to normal while
a mottled appearance was still evident on the
body. After four hours the coloration of the in-
jected fish was nearly back to normal, that is,
it was nearly like that of the control. The fol-
lowing morning the injected fish and the con-
trol were indistinguishable.

Some observations were made on the pig-
mentary reaction to different backgrounds of
three small individuals in a 15-gallon tank. No
adrenalin injections were made. In a tank de-
void of any plants or shells, with a bottom of
white sand and with clear glass sides, the three
fish assumed an all-over dark coloration with
the lighter bands showing faintly. With a white
sand bottom and with the sides of the tank cov-
ered with white paper, the coloration remained
very dark with faint lighter bands. The fish ap-
peared somewhat disturbed by this environ-
ment; they huddled together and were inclined
to lean over to one side (Plate II, Figure 1).
They were seen in the same dark color phase
at night when the lights were suddenly flashed
on in the laboratory.

In a tank with the slate bottom uncovered
and the glass sides covered with black paper,
the fish became lighter and the bands were more
clearly marked (Plate II, Figure 2) . They kept
this banded condition when surrounded by black
but quickly darkened when one paper side was
removed for observation. All the fish swam up-
right in the dark tank and were not seen to lean
over to one side as they did in the white tank.

With the sides of the tank covered with two-
inch vertical black and white stripes, the fish
assumed their bold black and white pattern. This
reaction was not a quick one, but took about a
half hour to occur. The fish again darkened
quickly when one paper side was removed.

Paper images of the fish of approximately the
same size were introduced into the experiment.
They comprised white images on a black back-
ground, black images on a white background
and white-barred images on a black background.
The pigmentary reactions of the fish were the
same when they were surrounded by any of
these as backgrounds. They responded with the
bold banded phase, even when black images on

88

Zoologica: New York Zoological Society

[40: 7

a white background were used (Plate II, Figure
3).

One of these three fishes never gave as posi-
tive reactions as did the other two; it always
remained somewhat darker and was seen more
often in an inclined position than either of the
others. It is possible that this was also an “emo-
tional” reaction of some sort, especially as a
dark phase is generally typical of most teleosts
in the lowest position in a hierarchy, for this
particular fish was annoyed and pecked at by
the other two.

This type of pigmentary behavior, of which
the present authors are evidently the only ones
to take cognizance, at first glance might seem
at wide variance with many of the pigmentary
studies of recent years which have been broadly
summarized by Sumner (1939), Walls (1942),
Parker (1948) and Fox (1953). It is believed
that the described observations on behavior can
be entirely explained on the basis of experi-
ments already performed on other fishes by
various investigators. A considerable portion of
it can be ascribed to the now well-established
fact that tested fishes of various species show
pigmentary responses both to light intensity and
to the ratio of incident light to that reflected
from the background. Walls (1942) treated
the matter as follows : “If the fish were respond-
ing merely to the amount of light entering the
eye, it should give the same responses to a bright-
ly illuminated dark background as to a dimly
illuminated white one— which would not adapt
the fish at all! Instead, however, the shade as-
sumed by the skin of the fish is always (unless
the intensity of the incident light is very low or
extremely high) in accordance with the albedo
of the substrate— the percentage of incident light
which the substrate reflects.” Brown (1936)
nicely demonstrated that the minnow Erycimba
clearly responded to both variation in light in-
tensity over a single background and to varia-
tions in the ratio of incident to reflected light,
by using uniform light intensity and backgrounds
of various degrees of reflectivity. Both his series
of experiments were carried to the limit of the
ability of the melanophores to respond; that is,
to disperse or concentrate their melanin gran-
ules. Carried beyond, there was no further
change. That is to say, below 0.000053 foot
candles they showed no further concentration,
as such was evidently impossible, while on the
same black background, above 1.75 foot candles
there was no further dispersion. The dispersion
of the melanin was found to be propor-
tional to the log of the foot candles at inter-
mediate light values. Using the minimum light
value necessary to produce complete dispersion
on a black background, Brown varied the latter

through various shades of gray to white; a light
gray background which reflected 0.1411 foot
candle of the incident 1.75 foot candles (an
albedo of 12.4) was sufficient to produce full
concentration of the melanin granules. The be-
havior at the other end of the series could not
be tested because of the nature of the experi-
mental arrangement whereby the minimum light
intensity necessary to produce complete disper-
sion on black was employed.

The above-described experiments are suffi-
cient to explain why dark or completely black
fishes may show a pattern when the light has
been reduced to a point where the presence of
such a pattern cannot be seen. This would seem
to have nothing to do with whether the fish
tends to match the background or to contrast
with it, but suffices to explain Figure 3 of Plate
I. The phenomenon of a fully black fish on a
light ground is apparently confined to places
subject to very great light intensities. These often
run up to and in excess of 6,000 foot candles in
summer and usually well over 2,000 in winter
at Bimini. Also the albedo is much smaller in
such places than in most other natural environ-
ments, reaching values at least as small as 3.00.
There is thus much less contrast between back-
ground and incident light than in places with
darker background and less transparent water.
That is, the sensible differential is much reduced,
and the retinal polarizing effect is minimized. As
noted, at these times the fishes recline so that
one eye looks skyward and the other down,
rather effectively neutralizing the polarization
of the retina as compared with the condition
when the fish is in the usual vertical position.
Each eye, while seeing a different field, one
differing from the other by the difference be-
tween the incident light and the reflected light,
is nevertheless seeing a comparatively uniform
field of very considerable brightness. This in it-
self may make it impossible for the fish to re-
spond clearly to the albedo, responding instead
overwhelmingly to the great light intensity, re-
sulting in its darkening irrespective of the fact
that the background is very light. This is the
equivalent of saying that the polarizing effect of
the visual field is essential in order for a fish
not to respond only to the light intensity. Evi-
dently only in regions of extremely clear water
with an exceedingly light bottom is the described
phenomenon possible.

The fact that these fish show their pattern
in turbid water would follow from the above
as turbidity would increase the value of the
albedo because the incident light, whatever its
value, only passes from the surface to the eye
of the fish, while that of the reflected light has
the much greater path; from the surface to the

1955]

Breder & Rasqnin: Pigmentary Behavior of Chaetodipterus

89

bottom and back to the eye of the fish. Only
in perfectly transparent water is the consequent
reduction in light negligible. This is because
the longer passage of the reflected light in even
very slightly turbid water results in much greater
filtering. Another type of behavior of Chaeto-
dipterus which is encountered at Bimini and
other places with similarly clear water, which
has not yet been noted, is associated with deep
water. In such places schools of five hundred
or more individuals of medium to very large
size may sometimes be seen, usually resting
quietly and all headed into the current. Such
schools have not been seen in depths of less than
about six fathoms and the fish were always in
a strongly barred phase. At such depths the
growth-incrusted bottom does not reflect nearly
as much light as the sandy shoal waters and
consequently the albedo is of appreciable size.

That such fishes as Chaetodipterus do not
show a general concentration of their melanin
granules in response to an injection of adrenalin
may be interpreted as follows. After injection
of adrenalin the actual dispersion of the melanin
granules in the lighter bars, the only place in
the body where such activity could be detected,
can only mean that the locus of the differential
behavior between the two types of fishes is not
to be found so much in a differently performing
endocrine system but rather in a differently re-
acting set of target organs. That is to say, the
same hormone, adrenalin, induces concentra-
tion of melanin granules in the dermal melano-
phores of Gambusia, but its dispersion in equiv-
alent melanophores in Chaetodipterus. It is also
to be recalled that melanophores of the iris and
the meninges in both cases respond by concen-
trating their granules, as shown by Breder &
Rasquin (1950).

In addition to this difference in cellular re-
sponse it is probable that there is also a more
subtle difference in the responsiveness of these
pigmentary effects to nervous control. It is ob-
vious that there is here an indication of a lessen-
ing endocrine control and an increasing nervous
control with the teleosts of the more advanced
grades. All the forms known to the authors
which show this type of reversal of reaction are
acanthopterygians, while the most typical ex-
amples of the simple background-matching
types are non-acanthopterygians. The fact that
these fishes showed faint evidences of moment-
ary lightish bands as a reaction to “fright” when
in the black phase, and a darkening of the bands
under similar stimulae when in the banded
phase, is suggestive of strong nervous control.
Such clearly nerve-controlled changes are not
evident in the more slowly responding melano-
phores of most non-acanthopterygians. Actu-

ally fishes from more turbid regions, Chaeto-
dipterus included, generally show all-over lighter
phases. In this extremely clear water, how-
ever, these fishes which are so much exposed
to strong light have no doubt built up their
complement of melanophores to maximum. The
concentration of melanin in the areas which
remain dark even when the fishes show then-
light bars is so great that it is doubtful if any
hormonal application could cause a visible
change in a short time. Evidently a sufficient
reduction in the number of melanophores, to
permit the noting of hormonally induced
changes, could be effected by keeping the fishes
under low light intensities for the required time.
Whether there is an inverse behavior shown by
the guanine present, as Hitchings & Falco
(1944) showed for other fishes, cannot be es-
tablished by these considerations.

A significant side to the differences in chro-
matic behavior which these fishes show under
different conditions of turbidity is the fact that
on the west coast of Florida Chaetodipterus is
known to the natives as “white angel.” This
name is not used in Bimini, but instead the
species is called “chirivita.”i This appelation,
elsewhere used for the very dark colored Pom-
acanthus, has evidently become transferred to
Chaetodipterus at this place, while Pomacan-
thus is called “black angel.”

The intermedin used (Choay) was found
to be inactive on the melanophores of many
different species of acanthopterygians. It was
found to modify the melanophores of the fresh-
water characin Astyanax and the marine cy-
prinodont Gambusia but not to the extent of
full granule dispersion. The biological assay for
this hormone is performed on the erythrophore
system of Phoxinus laevis and all fishes on
which it was used responded by dispersion of
granules in the erythrophores and xanthro-
phores. Because Chaetodipterus did not respond
to this preparation of intermedin, it is not to
be inferred that the melanophores of this spe-
cies are not under control of the intermediate
lobe of the pituitary, although, as has been
noted, the nervous system plays a more domi-
nant role.

The melanophore - dispersing hormone of
Armour was not available at the time the above
experiments and observations were carried out.

We wish to express our thanks to Miss Carol
Mosher for technical assistance in connection
with the November experiments and to Mr.
William Clarke for observations in May which
we were unahle to make personally.

1 Pronounced by the natives as “cherry-wheat-ah,”
which is in close agreement with their general habits
of pronunciation.

90

Zoologica: New York Zoological Society

[40: 7: 1955]

Summary

1. The melanophores of the meninges and iris
of Chaetodipterus faber concentrate their mel-
anin granules on the injection of adrenalin, but
those in the dermis simultaneously disperse their
granules.

2. This species responds both to changes in
light intensity and to the ratio of incident to
reflected light from the background or albedo.

3. These conditions are evidently responsible
for the fish showing a black and white banded
phase in moderate light intensity with a large
albedo, as in turbid water against a dark or
mottled background, and in very low light in-
tensities.

4. Alternatively they also cause the fishes to
show a solid black phase in intense light against
a very light background in clear water where
the value of the albedo is very small.

5. These conditions lead to the interesting
situation under natural conditions of a banded
fish becoming inconspicuous against a variety
of mottled backgrounds and again inconspic-
uous against a very light background by becom-
ing uniformly black and appearing as a bit of
sea bottom litter.

6. Accompanying these chromatic changes
are appropriate attitudes, in the banded phase
the fish swimming upright in ordinary fish
fashion but in the black phase reclining quietly
on the bottom or drifting slowly close to it pro-
pelled only by the byline dorsal and pectoral
fins.

Literature Cited
Breder, C. M., Jr.

1946. An analysis of the deceptive resemblances
to plant parts, with critical remarks on
protective coloration, mimicry and adap-

tation. Bull. Bingham Oceanogr. Coll.,
10 (2): 1-49.

1948. Observations on coloration in reference
to behavior in tide-pool and other marine
shore fishes. Bull. Amer. Mus. Nat. Hist.,
92 (5): 281-312.

1949. On the relationship of social behavior to
pigmentation in tropical shore fishes. Bull.
Amer. Mus. Nat. Hist., 94 (2): 83-106.

Breder, C. M., Jr. & P. Rasquin

1950. A preliminary report on the role of the
pineal organ in the control of pigment
cells and light reactions in recent teleost
fishes. Science, 111: 10-12.

Brown, F. A., Jr.

1936. Light intensity and melanophore response
in the minnow, Ericymba buccata Cope.
Biol. Bull., 70: 8-15.

Fox, D. L.

1953. Animal biochromes and structural col-
ours. Cambridge Univ. Press,: i-xiv— 1-
379.

Hitchings, G. H. & E. A. Falco

1944. The identification of guanine in extracts
of Girella nigricans. The specifieity of
guanase. Proc. Nat. Acad. Sci., 30: 294-
297.

Parker, G. H.

1948. Animal colour changes and their neuro-
humors. Cambridge Univ. Press: i-x— 1-
377.

Sumner, F. B.

1939. Quantitative effects of visual stimuli upon
pigmentation. Amer. Nat., 73 (746):
219-234.

Walls, G.

1942. The vertebrate eye. Bull. Cran brook Inst.
Sci., (19): i-xiv— 1-785.

EXPLANATION OF THE PLATES

Plate I. Pattern reactions of a single individual
C haetodipterus.

Fig. 1.

Fig. 2.

Fig. 3.
Fig. 4.

Fig. 5.
Fig. 6.

Behavior in a circular pool with light
sand background. Fish black and reclin-
ing, as it did in the ocean.

Identical behavior in a small aquarium
floored with light sand.

Nocturnal pattern, taken by photo-flash.
Pattern in daytime after water had been
turbidified.

Effect of darkened bottom and clear water.
Effect of renewed light bottom at night
with artificial light. Aquarium photo-
graphs by C. Mosher.

Plate II. Pattern reactions of Chaetodipterus.
Fig. 1. Darkening of three individuals with the

aquarium surrounded by white paper and
light sand bottom, with no dark objects
in the visual field. In these three photo-
graphs the paper on one side of the
aquarium was necessarily removed just
before the photograph was taken. This
did not visibly affect the first two but did
lighten the last, which effect increased
after photography.

Fig. 2. Pattern of three individuals in an aqua-
rium surrounded by black paper and a
black slate bottom. The light appearance
of the bottom is a refractive effect.

Fig. 3. Similar pattern shown by a single indi-
vidual to which are displayed two black
paper cut out “fish” with the aquarium
surrounded by white paper and with a
light sand bottom.

BREDER

PLATE 1

FIG. 1

FIG. 2

teigsifti

k*7--s^. «!*-*' ..■■iif ■■*•..• \>f.

■:'I^%L. ......

>^W"'

FIG. 4

SPECIAL FEATURES OF VISIBILITY REDUCTION IN FLATFISHES

BREDER

PLATE II

FIG. 3

FIG. 4

FIG. 1

FIG. 2

SPECIAL FEATURES OF VISIBILITY REDUCTION IN FLATFISHES

FIG. 3

FURTHER NOTES ON THE PIGMENTARY BEHAVIOR OF CHAETODIPTERUS
IN REFERENCE TO BACKGROUND AND WATER TRANSPARENCY

BREDER a RASQUIN

PLATE I

FIG. 2

FIG. I

FIG. 3

FIG. 4

FIG. 5 FIG. 6

FURTHER NOTES ON THE PIGMENTARY BEHAVIOR OF CHAETODIPTERUS
IN REFERENCE TO BACKGROUND AND WATER TRANSPARENCY

SEP 9

fdS5

8

Special Features of Visibility Reduction in Flatfishes

C. M. Breder, Jr.

The American Museum of Natural History, New York 24, N. Y.

(Plates I-III)

Introduction

The fishes of the order Heterosomata,
the flatfishes, have many members which
are remarkable for their ability to match
closely the color of the background on which
they rest and are evidently unique in their abil-
ity to match also the texture of the pattern of
the background. Parker (1948), in noting this
ability, wrote, “Responses to differences in pat-
tern do not seem to have been observed in
animals other than flatfishes; hence the excep-
tional nature of this group of teleosts.” Mast
(1916) published a series of plates in both
black and white and color which clearly estab-
lished the reality of this striking activity of the
pigmentary system of flatfishes. Others, before
and after him, who have contributed to our
present understanding of background matching
in this group of animals include Sumner (1910,
1911), Kuntz (1917), Schafer (1921), Cun-
ningham (192lb Hewer (1927, 1931), Meyer
(1930, 1931) and Osborn (1939a, 1939b,
1939c, 1940, 1941a, 1941b).

These responses to background are exceed-
ingly rapid and should not be confused with
relatively permanent patterns which render a
variety of other animals inconspicuous on some
definite type of background. In the cases of the
animals with a fixed pattern the disappearance
is accomplished by the animal fitting itself against
a suitable background and not by immediate and
complex neural adjustments of the pigmentary
system. A long list of forms which possess such
comparatively stable patterns may be found in
Cott (1940). These are widely distributed
throughout the animal kingdom. Reports on
various special cases of this type of behavior in
fishes have been summarized by Atz (1951)
and still others added by Uchida (1951). Ob-
viously there is no sharp dividing line between

animals which simply “melt” into the back-
ground and those which closely “mimic” some
object, a division between such categories being
often merely subjective.

There are, in addition to the problems of
background matching, which have been studied
by the workers mentioned above, other pecu-
liarities in the pigmentary behavior of the
Heterosomata. These, which include pigmentary
behavior of the young and larval stages, have
evidently never been studied nor related to the
pigmentary behavior of the adults. Attention
is herewith called to certain of these peculiarities
and to some of the problems they pose in the
Bothidae and the Achiridae.

Observations on Pigmentation
AND Behavior

The following data are based on original ob-
servations on pigmentary and other behavior
of three species of the family Achiridae and
one of the family Bothidae.

Patterns and Pigment in Achirids

Small individuals of Trinectes lineatus (Lin-
naeus)^ of about 20 mm. in standard length
are occasionally to be found near the laboratory
and, so far, always in a solid black condition.
One such individual, taken on January 15,
1952, was kept under continuous observation
for more than five months in an aquarium pro-
vided with running sea water. This aquarium
was floored with the typical light sand of the
region, against which the fish was notably con-
spicuous. The fish at no time hid in corners of

1 There has been considerable confusion in the no-
menclature of these fishes. According to Chabanaud
(1941), this form should be known as Trinectes lineatus
browni (Gunther), which see for a discussion of the
matter.

91

92

Zoologica: New York Zoological Society

[40: 8

the aquarium but regularly stationed itself on
some centrally located portion of the aquarium
bottom. Here it would lie, exposed on the very
light sand, very often with no sand cover at
all and seldom with more than a few grains. It
was never seen to cover itself completely, a
performance otherwise common in flatfishes.
Some act disturbing the fish, such as pushing
it about with a light rod, would not cause an
alteration of this general behavior. It would
usually move off slowly when the rod was
thrust at it, with a peculiar form of locomotion
resembling that of a large broad flatworm. It
would so flute its marginal fins as to appear
ruffled. Evidently the mechanics of this loco-
motion, which has been discussed in detail by
Chabanaud (1941), is actually close to that of
similarly-shaped flatworms. Once when so dis-
turbed it shook off what few grains of sand
were resting on it and rushed to nearly the exact
center of the aquarium bottom, where it lay
fully exposed and very conspicuous.

Lest it be thought that this was merely a
melanistic individual which was unable to
change its coloration at all, it may be pointed
out that after dark the fish, discovered by flash-
light, would always be found to be exhibiting
the more usual cross bands (see Plate I, Figures
1 and 2).

Since most blinded fish in strong light develop
their darkest phase, on which subject Parker
(1948) gives a brief historic review, the pres-
ence of vision in this individual was carefully
checked. This could be readily established by
the accuracy of its striking at such things as
Anemia as well as by the associated eye and
body movements.

The above notes and photographs were made
during January. By May the fish had altered its
pigmentary behavior to the following. During
the daytime it alternated between burying itself
in the sand, so that it could only be seen par-
tially or not at all, and exposing itself. At these
latter times it came to show its former night
time pattern with increasing frequency during
the daylight periods. The coal-black phase ap-
peared less and less often and usually the fish
was a slate gray with black bars or dark brown
bars. Sometimes it would swim about holding
little piles of sand on various places on its flat
body. The aquarium in which all this change
transpired was near, but not directly, under a
skylight. The significance of the resultant re-
duction of light intensity, as compared with the
open harbor, is considered later.

Another sole, Trinectes inscriptus (Gosse),
with a fairly inconspicuous pattern and color,
always buried itself completely and with ex-
treme rapidity, in the same sand. This action

was so fast and with such immediate vigor that
it was impossible to photograph the fish against
a sand background. The photograph of this fish,
Plate I, Figure 3, which appears to be that of
the fish resting on the sand, was actually taken
when an expedient was devised to circumvent
the fish’s rapid response. The fish was placed
in a glass bowl which contained water only
and then the bowl was placed on top of a pile
of sand. The violent action of the fish in its
fruitless attempts to bury itself was “stopped”
photographically by using an electric photo-
flash. Although this fish, as may be noted from
the photograph, is not exactly conspicuous on
the sand of the area it inhabits, it still does
not nearly approach the better background
matches so usual in these fishes. As far as could
be determined from aquarium observations,
this fish was completely nocturnal.

On the west coast of Florida, at the Palmetto
Key Laboratory of the old New York Aquarium,
in 1940, eggs of Trinectes lineatus collected by
tow net were carried through metamorphosis
m finger bowls. At the time when the one eye
began its migration to the other side of the
head, general pigmentation was well advanced
so that the fish was fully pigmented before the
transformation was completed. This fish, as the
eye migrated and the pigmentation increased,
behaved in a manner suggesting that this period
of change in the relationships of the visual fields
was one of considerable behavioral diflSculty. At
first it swam at an angle as though attempting
to retain a horizontal axis between the eyes.
After this there followed a period, as the eye
came over the dorsal profile of the head, in
which the fish would lie down briefly, then get
up and swim about erratically and then lie down
again. Finally these periods of reclining became
longer and longer, until the fish began to behave
in a typically flounder fashion, the wandering
eye, meanwhile, attaining the other side of the
head. The water at this place is very turbid
because of the presence of an extremely rich
plankton, a condition which, as will be devel-
oped, has a distinct bearing on these pigmen-
tary matters. Pigmentation appears in this
form before the egg is hatched and Plate I,
Figure 4, shows its extent five days after hatch-
ing, but long before any evidence of transfor-
mation appeared. The developmental stages of
this species have been figured by Hildebrand &
Cable (1938)2 and the present material is in
complete agreement with their illustrations.

Transparent and Pigmented Bothids

Small planktonic and fully glass-like individ-

2 The same species in the usage of Chabanaud (1941).

1955]

Breder: Visibility Reduction in Flatfishes

93

uals of Platophrys ocellatus (Agassiz) may be
taken about submerged lights in Bimini harbor
with some degree of regularity. They are in an
advanced stage, with the eyes on one side, and
generally are found floating passively just under
the surface of the sea. When placed in a bowl
or aquarium they promptly settle to the bottom.
Four views of such an individual are given in
Plate II, two by transmitted light and two
against a dark background. Oblique light was
used in all but the first in order to take maxi-
mum advantage of the structural interference
with light. It will be noted that the eyes are fuUy
pigmented, the photographs with the dark
background showing that the color of the in-
vesting tissues produces a light eyeball. This
is a near cream color and very close to that of
the light sand of this region.

In this stage these fish are extremely interest-
ing to watch through a low power binocular.
Mostly they rest quietly and show much oculo-
motor activity. Their extremely mobile and, of
course, independently moving eyes, are capa-
ble of looking straight up. It is almost startling
to have one such eye appear suddenly to be
looking back at one through the other end of
the microscope. That this great ocular activity
is utilitarian is indicated by the fact that nearly
always some particle is clearly in the line of
fixation. That this really is the case may be
demonstrated by the introduction of a quantity
of newly hatched Artemia. The near ones be-
come the objects of fixation just a moment be-
fore a small darting movement and engulfment
of these food objects takes place. Because of the
fish’s great transparency, the swallowing of the
Artemia may be easily followed to where it
quickly lodges in the stomach. As the stomach
becomes packed with Artemia and they die and
are acted on by digestive juices they become
a pale “boiled shrimp” pink which beautifully
outlines the stomach. As the stomach does its
work, small boluses of this material are seen
to tear away from the main mass as the pyloric
sphincter opens to pass the food along into the
intestine. The whole process of digestion may
be followed in this manner as though viewed
on the screen of a fluoroscope.

If these fish are permitted to live under such
conditions but with the normal sand back-
ground they nearly completely disappear, only
the pigmented eyes being recognizable but, even
so, extremely inconspicuous. Plate III, Figure
1, shows such a situation. The eyes may be
seen as indicated but the mottling about them
consists of the sand grains seen through the
body of the fish. When this photograph was
taken a few pigment spots had appeared, but
earlier even these were not to be found. The

fish may continue to live in this condition for
a protracted time, not changing in pigmentation
as rapidly as do many other planktonic larval
fishes when brought in contact with a shore-like
environment in an aquarium. These two fishes
were taken about a submerged light on Novem-
ber 22, 1953. The photographs of one of them
in Plate II, as noted in the legend, were taken
on Nov. 23, 24, 26 and 27 respectively. Figure
1 of Plate III, which includes the two fish, was
taken on Dec. 12, twenty days after they had
been living in a sand-floored finger bowl. Two
days later the next photograph, Plate III,
Figure 2, was taken with the condition prac-
tically unchanged. Finally on December 19,
after 27 days in the bowl (Plate III, Figure 3)
the larger individual was seen to be fully pig-
mented and the smaller nearly so. In both the
transparent and pigmented states and in the in-
termediate conditions these fish seem to be just
about equally invisible. At this time it was
necessary to abandon this series of observations,
but these examples were carried to a point where
the essential pattern of the adult had estab-
lished itself. It may be noted that in all the fishes
shown in Plate III, there is a small dark median
spot about two-thirds of the distance from the
eyes to the caudal peduncle. This made its ap-
pearance between the time the first of these
pictures was made and the last of Plate II.

Discussion

From the foregoing descriptions it is evident
that all flatfish pigmentary behavior is not
uniformly a matter of matching the color and
texture of the background. The following
tabulation indicates the situations involving
pigmentary and other behavior with regard to
various environmental conditions which have
been herein described.

Pigmentary

Location

Condition

OF Fish

Light

Trinectes lineatus

Uniformly black Exposed

Intense

Barred

Exposed

None

Barred

(Exposed

(Buried

Moderate

Trinectes inscript us

Reticulated

Buried

Intense or

Moderate

Reticulated

Exposed

None

Platophrys ocellatus

Transparent

Exposed

Moderate

Background matching Exposed

Moderate

The normal habitat of all was one of light
sand, extremely clear water and extremely
bright sunlight, excepting that of the Floridian

94

Zoologica: New York Zoological Society

[40: 8

Trinectes. These reactions, which at first might
seem to be merely fortuitous and without any
regularity, can in fact be readily interpreted and
understood in the light of the known behavior
of teleost pigmentary systems. The explanations
covering the situation in each of the four spe-
cies examined follow.

The case of the Bahaman Trinectes, black on a
light sand background, is completely explained
on the same basis as the similar behavior of
Chaetodipterus discussed at length by Breder &
Rasquin (1955). As they indicate, in a fish re-
sponding to both light intensity and albedo under
conditions of intense light and very light back-
ground, the melanophores will show maximal
dispersal and the fish become substantially black.
In very low light or none, as at night, what basic
pattern is present but hidden reappears. There-
fore there is no necessity to attempt to ascribe
some biological utility to the seemingly pointless
matter of “displaying” a pattern when it cannot
be seen. After this fish had been kept in a lab-
oratory aquarium for a long time it evidently
reduced its amount of melanin in response to
the lesser light intensity and showed its pattern
in the daytime.

Trinectes inscriptus, which showed no evident
pigmentary responses, simply remained noc-
turnal, with a pattern which was neither strongly
contrasting with the background nor closely
matching it.

Trinectes lineatus, living in a place of rela-
tively low light and high albedo, showed both
good matching and burying habits and had well
pigmented larvae.

The glass-like planktonic young Platophrys
ocellatus may settle to the bottom in that con-
dition and so remain for some time and then
gradually develop the extremely accurate bot-
tom matching behavior so characteristic of the
group.

One of the most interesting matters that
developed in connection with the contemplation
of these fishes was that their behavior, in each
case, was appropriate to their pigmentation.
The transparent, the closely bottom matching
and the strongly bottom contrasting fishes all
exposed themselves and each, in its way, there-
by became less visible. Only those which were
neither extremely contrasting nor extremely
well matched to the bottom, hid or buried
themselves. Note especially that the Bahaman
Trinectes, at first so extremely black and “bold,”
became barred after a sojourn of some months
in the less intense light of the laboratory, and
changed its behavior appropriately. These clear
changes in behavior in accordance with pig-
mentary differences add emphasis to the ob-
servations that fishes may employ either an

extremely well matching pigmentation or a
violently contrasting one to disappear effectively.
That there are other fishes which darken on a
light background in a strong light and thereby
become inconspicuous was recognized by Breder
(1948, 1949) and by Breder & Rasquin (1950,
1955).

The Trinectes taken from turbid water with
a comparatively dark bottom was kept in a light
colored bowl in the laboratory. It regularly
lightened at night and darkened in the daytime.
Here it seemed to respond chiefly to light in-
tensity in a manner comparable to that of clear
water Trinectes but less vigorously under the
lesser stimulus. This followed from the extreme-
ly shallow water in the light bowl being insuf-
ficiently turbid to produce an albedo of large
value, such as was normal to the environment.

In the case of the transparent Platophrys, as
a purely practical matter it would seem to make
little difference whether these fish remained
transparent so that the bottom could be seen
through them or whether they developed a very
complicated pigmentary system which resembles
the bottom under them as much as possible. The
reasons for this change from the transparent to
the pigmented, both of which evidently have
equivalent value to the possessor, is probably
to be found elsewhere. There may be two chief
reasons for this change. It is conceivable that
it would be difficult to prevent the development
of pigment in the bright environment in which
they live. It may be impossible for any large fish
to remain transparent in such places, which cir-
cumstance would in effect relegate this device to
application by only small organisms and for
periods of relatively short duration. It also may
not be very practical behavior to display the
physiology of digestion through transparent
sides on a sufficiently large scale. This demon-
strated digestive activity might very well be at-
tractive to the numerous marauding crabs of
the region. Aside from these possibilities, there
may be purely physical obstacles to the main-
tenance of transparency for long in such an
environment. That is, some of the radiation may
be physically injurious to certain tissues if long
continued. It is thought, at least, that the com-
paratively subdued light of the laboratory made
it possible for these fishes to retain their trans-
parency longer than they could have in the wild
state. The fact that Platophrys showed no change
in general behavior, in passing from its trans-
parent to its closely matching pigmentary pat-
tern, would seem to indicate that the fishes were
equally “secure” at all times during the transi-
tion.

In a more general consideration of teleost
pigmentary reactions, reference may be made

1955]

Breder: Visibility Reduction in Flatfishes

95

to some earlier studies of such matters. Breder
(1949) found it convenient to prepare a table
which listed the conceivably useful reac-
tions a fish might find it possible to make in
response to various stimuli. Pigmentary response
was only one of the items listed and only that
part of it which refers to reactions to back-
ground pertains to the present considerations.
This fragment of the tabulation may be re-
phrased for present purposes as follows.

Possible pigmentary reactions in reference to
background:

A. Matching background

a. In general tone

b. In pattern detail

B. Opposing background

C. Indifferent to background

The data presented here show the fishes to
subscribe to the items in this tabulation as be-
low.

Trinectes lineatus B., A., a.

Trinectes inscriptus C.

Platophrys oce//amj— Transparent C.

—Pigmented A., b.

It is thus evident that each of the categories
listed is represented in various ways by these
fishes. The transparent Platophrys is considered
“indifferent to background” in that it does not
respond, not having any chromatophores with
which to respond. Trinectes inscriptus, in the
same category, is pigmented but there are no
prompt pigmentary reactions. Each of the cate-
gories, as represented by these fishes, is primarily
appropriate for the “freezing” type of organism,
except that shown by Trinectes inscriptus, as an
example of “C.” This is the only fish which
showed violent burying attempts. It is to be
noted in this connection that Trinectes lineatus
showed pronounced burying attempts only after
it passed from its extreme background con-
trasting condition.

As a matter of general observation involving
such species as Pseudopleuronectes americanus
(Walbaum) and Paralichthys dentatus (Lin-
naeus), both excellent bottom matchers, the fol-
lowing conditions seem to obtain. They may
be found either exposed and bottom matching
or partially or fully buried with only their eyes
above the sand. If pursued in shallow water over
a uniform bottom they are not likely to bury
themselves unless the pursuit is very vigorous.
However, if they are driven onto a differently
colored or textured bottom they are almost cer-
tain to bury themselves. This is, of course, in
keeping with the data herein presented. Evi-
dently the burying behavior supplements the
pigmentary behavior in accordance with the
requirements of the moment. This inter-relation

of locomotor and pigmentary response to back-
ground has been shown, in a different guise, to
oe present in such simple background-matching
fishes as Gambusia, by Breder (1947), and in
Cyprinodon, by Breder & Rasquin (1951). In
both cases the activity of the fishes was found
to greatly alter when they passed over a non-
matching background. Also related to this is the
reluctance of many fishes to move over a back-
ground toned differently from the one on which
they have been swimmmg for some time, a mat-
ter discussed in the references above.

Some of the features associated with the
development of glass-like transparency, com-
mon to the planktonic young of many fishes,
have evidently not been previously discussed.
Since various details of the present data sug-
gest some of the necessary concomitants of a
practically invisible body, they are discussed
below.

It is noteworthy that in all the innumerable
fishes that have such transparent early stages
the eye is well pigmented. This is, of course,
essential to image formation on the retina, and
hence to vision. Evidently not one known fish
has developed transparent but sightless eyes.
Such a development would clearly extend the
transparency to practical perfection. From the
fact that this has not been done, it is deduced
that vision plays such an important role in the
lives of these fishes as to forbid its abrogation
in the interest of perfection of transparency as
a protective mechanism. There would thus seem
to be a balance between the two conditions, in
which the value of the retention of vision over-
rides the significance of cues which a predator
might find from the presence of two “disem-
bodied” black spots moving along a fixed
distance apart. As an item of interest bearing
on the above, very often, on examining a bucket
or bowl of plankton, the first knowledge that
the viewer has of the presence of the more ex-
tremely transparent fishes is just this indica-
tion of paired black dots moving along together
without any apparent reason. This is, of course,
a sophisticated recognition which the uninitiate
do not usually make. If predators on these fishes
are able to use similar means of detection it is
evidently not sufficiently important to detract
seriously from the survival value of combined
transparency and vision.

Considerations of the relationship between
the transparent fishes and the opaque ones fol-
low. It would appear that certain limited tissues,
such as muscle, cartilage and connective tissue,
may be rendered completely transparent while
others, by their very nature, cannot. The latter
type includes at least erythrocytes with then-
contained haemoglobin. In the transparent fishes

96

Zoologica: New York Zoological Society

[40; 8

such unavoidably colored tissues are evidently
kept at a minimum. Rasquin (in press), for ex-
ample, in discussing the leptocephalus of Albula,
shows that the red bloou cells are remarkably
few. With such structures kept at the minimum
and dispersed in the animal, its visual density at
macroscopic levels may be no greater than that
of the surrounding water. To attain this con-
dition the opaque tissues would have to be dis-
persed in the body of the animal in a manner
approximately equivalent to the dispersion of
particles suspended in the water and of their
order of magnitude.

In a similar sense, pigmented fishes, both
background matching and background con-
trasting, may disappear by mingling with par-
ticles of detritus which are of their own order
of corporal magnitude. As found in a state of
nature in water, most bodies either float or sink.
Because of this, such association with particles
about the size of the fishes in question is prac-
tically absent from mid-water but such hiding at
the surface or on the bottom is common and
widespread.

The next step in such a series by a single fish
is the matching of a large number of smaller
particles or objects. This evidently is done by the
animals which closely resemble an area of
background with which they are associated. This
frees them from the need of having something
sufiiciently like themselves in size with which
to mingle. There is no better illustration than a
flounder matching a pebble background.

These observations lead to a general con-
sideration of fishes living in extremely clear and
brightly lighted water where optical cues are
probably dominant or at least have much greater
range and importance than in turbid water. That
the interplay of the three most prominent
methods of reducing conspicuousness, i.e., trans-
parency, background matching and background
contrasting, is most evident in brightly lighted
clear water, should be expected. Likewise it
should be expected that these features would be
relatively reduced in turbid waters. The turbidity
itself sharply restricts the range of vision while
the increase in albedo which it induces, elimi-
nates or restricts to very shallow water the back-
ground contrasting behavior. The data pre-
sented herewith and the pertinent references snp-
port such a correlation of light intensity and
albedo with the locomotor activity and pig-
mentary behavior of fishes in reference to back-
ground. In a broader sense it would seem that
these features of the behavior of the pigmentary
system, together with the interacting nervous
and endocrine systems, must have a very in-
fluential bearing on the evolution of pattern
in fishes.

Summary

1. An entirely black Trinectes lineatus (Lin-
naeus) positioned itself conspicuously on a
light sand background and showed a pattern
of light cross bars only when the light intensity
fell to a value too low for the human eye to see.

2. A Trinectes inscriptus (Gosse), with a
network of fine lines, which was not notably
conspicuously nocturnal, remained buried com-
pletely out of sight in the daytime.

3. An as yet untransformed Trinectes linea-
tus, taken in a tow net, developed pigment as
its eye migrated and by the time this was com-
plete the fish was fully pigmented.

4. Completely transparent and glass-like
planktonic post-larvae of Platophrys ocellatus
(Agassiz), with the eye transformation com-
pleted, on taking up residence on the same light
sand did not develop pigment for a long time
but nevertheless were practically invisible be-
cause of their glass-like transparency, but later
when they did develop typical flounder back-
ground-matching color and pattern were just
about as hard to find.

5. These apparently unorthodox pigmentary
behavior patterns are discussed in reference to
more usual modes and are found to be simply
special or limiting cases, fully referable to
known reactions and influences in the general
behavior of teleost pigmentary systems.

Acknowledgements

The studies on which this paper is based
have, for the most part, been carried out at
the Lerner Marine Laboratory on Bimini in the
Bahamas. I am indebted to Mr. Marshall Bishop
for collecting some of the material and to Miss
Carol Mosher for technical aid.

Bibliography

Atz, J. W.

1951. Fishes that look like plants. Animal King-
dom, 54 (5): 130-136.

Breder, C. M., Jr.

1947. A note on protective behavior in Gam-
busia. Copeia, 4: 223-227.

1948. Observations on coloration in reference
to behavior in tide-pool and other marine
shore fishes. Bull. Amer. Mus. Nat. Hist.,
92: 281-312.

1949. On the relationship of social behavior to
pigmentation in tropical shore fishes. Bull.
Amer. Mus. Nat. Hist., 94: 83-106.

Breder, C. M. Jr. & P. Rasquin

1950. A preliminary report on the role of the
pineal organ in the control of pigment
cells and light reactions in recent teleost
fishes. Science, 111: 10-12.

1955]

97

Breder: Visibility Reduction in Flatfishes

1951. A further note on protective behavior in
fishes in reference to background. Copeia,
1: 95-96.

1955. Further notes on the pigmentary behavior
of Chaetodipterus in reference to back-
ground and water transparency. Zoologica,
40 (7): 85-90.

Chabanaud, P.

1941. Sur le comportement en aquarium d’un
teleosteen de la famille des Achirides.
Bull. Soc. Centr. Agricult, et Peche, 1-12:
11-27.

COTT, H. B.

1940. Adaptive coloration in animals. London,
Methuen and Co., Ltd., xxxii-508 pp.

Cunningham, J. T.

1921. Hormones and heredity. London, Con-
stable and Co., Ltd., xx— 246 pp.

Hewer, H. R.

1927. Studies in colour changes of fish. II. An
analysis of the colour patterns of the dab.
III. The action of nicotin and adrenalin
in the dab. IV. The action of caffeine in
the dab, and a theory of the control of
the colour changes in fish. Philos. Trans.
Series B, 215: 177-200.

1931. Studies in colour-changes in fish. V. The
colour-patterns in certain flatfish and their
relation to the environment. Jour. Linn.
Soc. (ZooL), 37: 493-513.

Hildebrand, S. F. & L. Cable

1938. Further notes on the development and
life history of some teleosts at Beaufort,
N. C. Bull. U. S. Bur. Fish., XLVIII (24) :
503-642.

Kuntz, a.

1917. The histological basis of adaptive shades
and colors in the flounder Paralichthys
albiguttus. Bull. U. S. Bur. Fish., 35: 1-29.

Mast, S. O.

1916. Changes in shade, color and pattern in
fishes and their bearing on the problems
of adaptation and behavior, with special
reference to the flounders, Paralichthys
and Ancyclopsetta. Bull. U. S. Bur. Fish.,
34:

Meyer, E.

1930. Ueber die Mitwirkung von Hormonen
beim Farbwechsel der Fische. Forsch.
und Fortschr., 6: 379-380.

1931. Versuche fiber den Farbwechsel von
Gobius und Pleuronectes. Zool. Jb. (Abt.
allg. Zool. Physiol.), 49: 231-270.

Osborn, C. M.

1939a. The experimental production of melanin
pigment on the ventral surface of the
summer flounder (Paralichthys dentatus).
Anat. Rec., 75: 60.

1939b. The physiology of color changes in flat-
fishes. Joum. Exp. Zool., 81: 479-515.

1939c. The effects of partial and total blinding
on the color changes of the summer
flounder (Paralichthys dentatus). Anat.
Rec., 75: 136.

1940. The experimental production of melanin
pigment on the lower surface of summer
flounders (Paralichthys dentatus). Proc.
Nat. Acad. Sci. Wash., 26: 155-161.

1941a. Factors influencing the pigmentation of
regenerating scales on the ventral surface
of the summer flounder. Biol. Bull. Woods
Hole, 81: 301-302.

1941b. Studies on the growth of integumentary
pigment in the lower vertebrates. I. The
origin of artificially developed melano-
phores on the normally unpigmented
ventral surface of the summer flounder
(Paralichthys dentatus). Biol. Bull.
Woods Hole, 81: 341-351.

Parker, G. H.

1948. Animal colour changes and their neuro-
humors. A survey of investigations 1910
to 1943. Cambridge University Press, i-x:
1-377.

Rasquin, P.

In press. Observations on the metamorphosis of
the bonefish Albula vulpes (Linnaeus).
Joum. Morph.

Schafer, J. G.

1921. Beitrage zur Physiologie des Farben-
wechsels der Fische. I. Untersuchungen
an Pleuronectiden. II. Weitere Untersuch-
ungen. Pflfig. Arch. ges. Physiol., 188:
25-48.

Sumner, F. B.

1910. Adaptive color changes among fishes.
Bull. N. Y. Zool. Soc., (42): 699-701.

1911. The adjustment of flatfishes to various
backgrounds. A study of adaptive color
change. Jour. Exp. Zool., 10 (4): 409-
505.

UCHIDA, K.

1951. Notes on a few cases of mimicry in fishes
(In Japanese with English summary).
Sci. Bull. Fac. Agri. Kyushu Univ., 13
(1/4): 294-296.

98

Zoologica: New York Zoological Society

[40: 8: 1955]

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

EXPLANATION OF THE PLATES

Plate I. Trinectes

1. T. lineatus (Linnaeus) in a typical pose
and coloration on light sand in daylight.
X 11/2.

2. The same fish in darkness. Picture made
by means of prearranged photo-flash.

3. T. inscrip tus (Gosse) as seen against a
background of the native sand in a pose
never permitted by the fish under normal
circumstances. X

4. T. lineatus before metamorphosis, show-
ing the early development of extensive
pigmentation, five days after hatching at
a length of 2.5 mm.

Plate II. Platophrys ocellatus (Agassiz)

1. An individual in the extremely transparent
stage seen against a dark background with
reflected light, showing the light pig-
mentation of the eyes and some very pale
chromatophores. November 23, one day
after capture.

2. The same fish against a dark background
with extremely oblique reflected light.
November 24.

3. The same fish by transmitted oblique
light, partially against a dark background,
showing the glass-like transparency. No-
vember 26. X 3V1.

Fig. 4. The same fish by transmitted light so
arranged as to bring out the pale chroma-
tophores shown in Figure 1. Here they
show as mere small smudges. November
27. Photographs by Carol Mosher.

Plate III. Platophrys ocellatus (Agassiz)

Fig. 1. The fish of Plate II and another larger
individual against a background of light
sand. This is the background normal to
these fishes and on which they virtually
disappear. They are here nearly trans-
parent and the sand under them is plainly
visible. The eyes of the smaller one can
be found at the right of the picture, the
fish facing up. Those of the larger are at
the left, the fish facing down. December
12. X

Fig. 2. The same fish. They are in the central
portion of the picture. The smaller, to
the right, is facing left, and the larger,
to the left, is facing the lower left-hand
corner. Slightly more pigment is present.
December 14.

Fig. 3. The same fish, the larger showing full
pigmentation and the smaller partial pig-
mentation. The smaller fish is to the left
and facing a little to the right. The larger
is to the right and facing a little left of
upward. December 19. Photographs
Carol Mosher.

BREDER

PLATE III

FIG I

FIG. 2

FIG. 3

SPECIAL FEATURES OF VISIBILITY REDUCTION IN FLATFISHES

9

Further Chemical Analysis of Holothurin, the Saponin-like
Steroid from the Sea-cucumber

J. D. Chanley, Stella K. Kohn, Ross F. Nigrelli & Harry Sobotka

Department of Chemistry, The Mount Sinai Hospital, New York; and
New York Aquarium, New York Zoological Society

IN continuation of our work (Nigrelli, et al.,
1955) on the toxic principle of the sea-cu-
cumber Actinopyga agassizi, we have found
that the molecule contains a sulfuric acid group
in ester linkage. Holothurin, after passing
through a mixed bed ion exchange column,
shows the following elementary composition;
Found: C, 51.75; H, 8.00; S, 2.54; 2.86; acid
equivalent, 1150. Calculated for C50H82O26S:
C, 52.2; H, 7.8; S, 2.78; acid equivalent, 1148.

That the sulfur is in the form of a sulfate is
shown by the isolation of barium sulfate from
the hydrochloric acid hydrolysate of the above
material. Furthermore, hydrolysis without addi-
tion of acid by heating of the steambath with
reflux for 54 hours, and subsequent removal of
the precipitated aglycone, produces an increase
of titrable acidity by 88 percent., showing the
liberation of H2SO4. The original neutral com-
pound thus consists of the sodium salt of a sul-
furic acid ester.

The aglycone, obtained by hydrolysis with
methanolic hydrochloric acid, seems to be es-
sentially uniform and corresponds to substance
A (Nigrelli, et al., 1955), for which we pro-

pose the name holothurinogen. Its rotation is
[a] = — 17° (in CCI3H) . Upon treatment with

ether its melting point rises from 258° to ca.
274-276°, indicating the loss of some crystal
methanol.

The infrared spectrum of holothurinogen
shows bands at 3500 (hyroxyl) , 1740 and 1080
(carbonyl), and 1635, 995 and 945 cm"’^ (con-
jugated double bonds). The position of these
functions in the molecule is under investigation.

Summary

Holothurin is found to contain a sulfuric acid
group in ester linkage. The infrared spectrum of
the aglycone indicates the presence of a car-
bonyl group and a conjugated system of two
double bonds. The name holothurinogen is pro-
posed for the principle aglycone.

Reference

Nigrelli, Ross F., J. D. Chanley, Stella K.

Kohn & Harry Sobotka

1955. The Chemical Nature of Holothurin, a
Toxic Principle from the Sea-Cucumber
(Echinodermata: Holothurioidea). Zoo-
logica, 40: 47-48.

99

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

30 East Fortieth Street, New York 16, N. Y.

PUBLICATION OFFICE

The Zoological Park, New York 60, N. Y.

OFFICERS
PRESIDENT
Fairfield Osborn

SECRETARY
Harold J. O’Connell

VICE-PRESIDENTS TREASURER

Alfred Ely David H. McAlpin

Laurance S. Rockefeller
Donald T. Carlisle

ASSISTANT TREASURER
Percy Chubb, 2nd

SCIENTIFIC STAFF: Zoological Park and Aquarium

John Tee-Van Director

Leonard J. Goss Assistant Director

ZOOLOGICAL PARK

Grace Davall Assistant Curator,

Mammals and Birds

James A. Oliver Curator of Reptiles

Leonard J. Goss Veterinarian

Charles P. Gandal. . .Assistant Veterinarian
Lee S. Crandall. .... General Curator

Emeritus

William Beebe Honorary Curator,

Birds

AQUARIUM

Christopher W. Coates . Curator & Aquarist

James W. Atz Assistant Curator

Ross F. Nigrelli Pathologist

Myron Gordon Geneticist

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . .Research Associate

in Histology

Homer W. Smith Research Associate

in Physiology

DEPARTMENT OF TROPICAL
RESEARCH

William Beebe Director Emeritus

Jocelyn Crane Assistant Director

Henry Fleming Entomologist

Rosemary Kenedy Field Assistant

J ohn Tee- V an A ssociate

William K. Gregory Associate

AFFILIATES

C. R. Carpenter Co-ordinator, Animal

Behavior Research Programs

L. Floyd Clarke Curator,

Jackson Hole Research Station

SCIENTIFIC ADVISORY COUNCIL
A. Raymond Dochez Caryl P. Haskins
Alfred E. Emerson K, S. Lashley
W. A. Hagan John S. Nicholas

EDITORIAL COMMITTEE

GENERAL Fairfield Osborn, Chairman

William Bridges Editor & Curator, James W. Atz Lee S. CrandaU

Publications William Beebe Leonard J. Goss

Sam Dunton Photographer William Bridges James A. Oliver

Henry M. Lester. . .Photographic Consultant Christopher W. Coates John Tee-Van

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 40 • PART 3 • NOVEMBER 14, 1955 • NUMBERS 10 TO 14

PUBLISHED BY THE SOCIETY
The ZOOLOGICAL PARK, New York

Contents

PAGE

10. A Comparative Study of the Morphology and Histochemistry of the

Reptilian Adrenal Gland. By William B. Hebard & Harry A. Charipper.
Plates I-V; Text-figure 1 101

1 1. Hormonal Control of the Sexually Dimorphic Pigmentation of Thalassoma

bifasciatum. By Louise M. Stoll. Plates I-III. , 125

12. The Use of Copper Sulfate as a Cure for Fish Diseases Caused by Parasitic
Dinofiagellates of the Genus Oodinium. By Robert P. Dempster. Plate

I; Text-figure 1 133

13. Polymorphism in Reared Broods of Heliconius Butterfiies from Surinam

and Trinidad. By William Beebe. Plates I- VI 139

14. Formation of a Mucous Envelope at Night by Parrot Fishes. By Howard

Elliott Winn. Plate 1 145

10

A Comparative Study of the Morphology and
Histochemistry of the Reptilian Adrenal Gland

William B. Hebard^ & Harry A. Charipper
Department of Biology, Graduate School of Arts and Science,
New York University, New York, N. y.

(Plates I-V; Text-figure 1)

Introduction

There is considerable evidence indicat-
ing the importance of the mammalian
adrenal cortex in regulating certain as-
pects of protein, carbohydrate and electrolyte
metabolism (Sayers, 1950). It would seem that
any general concepts of adrenal histophysiology
would be enhanced by comparative studies in
all vertebrate classes. The function of the equiv-
alent cortical tissue in the poikilothermous rep-
tiles presents a challenging problem that has
received little attention. Moreover, the mor-
phology of the reptilian adrenal gland has not
been adequately described for many common
North American species. A basic knowledge of
normal morphology should be a prerequisite
for accurate interpretation of experimental mod-
ifications.

The purpose of this study is to describe the
normal anatomy, histology, cytology and his-
tochemistry of the adrenal glands of certain
common North American representatives of the
various orders of reptiles in order to determine
the nature of the similarities and differences
manifested by the adrenals. By this approach
it was hoped that types especially suitable for
future experimental studies would be discovered.

Knowledge of adrenal morphology in the
Crocodilia is based primarily upon a brief ana-
tomical and histological description of the
adrenals of the alligator. Alligator mississip-
piensis, by Reese (1931). Lawton (1937), also,
has described the gross anatomy and histology
of the the alligator adrenal with special refer-
ence to the adrenal-autonomic complex and vas-
cular supply. The histogenesis of the alligator
adrenal has been described by Forbes (1940).

^Accepted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy at New York
University, June, 1954.

There is a paucity of information regarding
the chelonian adrenal. Present knowledge is
based principally on the early description of the
adrenal in the turtle, Emys orbicularis, by Vin-
cent (1898) and the description of the develop-
ment of the adrenals in the loggerhead turtle,
Caretta caretta, by Kuntz (1912).

The adrenals of the Serpentes have received
little attention. Present knowledge is based
mainly on the early descriptions of colubrid
snakes by Vincent (1898), Minervini (1904)
and Radu (1934); recent observations by Valle
& Souza (1942) on the histology of the adrenals
in the snake, Dryophylax sp.; and brief notes
on the morphology of certain South American
snakes by Uchoa Junqueira (1944).

The adrenal morphology of the Sauria has
received some attention and histological descrip-
tions are available for a few species of lizards.
The first histological description of the reptilian
adrenal is that of the European lizard, Lacerta
agilis, by Braun (1882). Kraus (1921) gave a
rather complete histological description of this
species. Bimmer (1950) has described the sea-
sonal variations in the adrenals of the closely
related lizard, Lacerta vivipara. Vincent (1896,
1898) described the histology of the adrenals in
Chameleon vulgaris, Anguis fragillis, Lacerta
agilis and Uromastix hardwickii. More recently,
Retzlaff (1949) described the histology of the
adrenals in the alligator lizard, Gerrhonotus
multicarinatus. Hartman & Brownell (1949)
gave a brief description of the adrenals in the
Gila monster, Heloderma suspectum, and the
iguanid, Anolis carolinensis. Perhaps the most
complete discussion of reptilian adrenal mor-
phology is that by Miller (1952), who described
the normal histology of the viviparous lizard,
Xantusia vigilis, and its experimental modifica-
tion by hypophysectomy, starvation and the ad-

101

N0V2?1955

102

Zoologica: New York Zoological Society

[40: 10

ministration of ACTH, Cortisone and DOCA;
however, no histochemical studies were made.

The authors wish to express thanks to Drs.
Ross F. Nigrelli and James A. Oliver of the New
York Zoological Society, Mr. Charles M. Bogert
of the American Museum of Natural History,
and Dr. Royal B. Brunson of Montana State
University, for specimens generously contrib-
uted for this study. Acknowledgement is also
made to Mr. Albert H. Stenger for the photo-
micrography.

Materials and Methods

The reptiles used in this study were obtained
from various sources. Young alligators. Alli-
gator mississippiensis, from Schriever, Louisi-
ana, were obtained through local animal deal-
ers. Horned lizards, Phrynosoma cornutum,
were obtained from the vicinity of Abilene,
Texas. Skinks, Eumeces fasciatus, were obtained
from the vicinity of Mena, Arkansas. Painted
turtles, Chrysemys picta, were collected in the
vicinity of Lakehurst, New Jersey, and were
also obtained from collectors near Oshkosh, Wis-
consin. Garter snakes, Thamnophis sirtalis, were
collected in the vicinity of Glen Harbor, Long
Island, New York, and supplemented by col-
lections obtained from the vicinity of Oshkosh,
Wisconsin.

A total of 304 animals was used in the study.
These may be classified under the following
orders, families and species, conforming with
the recent checklist of North American reptiles
(Schmidt, 1953):

Order Crocodilia
Family Crocodilidae

16 Alligator mississippiensis Daudin,
Alligator

10 Caiman crocodilus Linnaeus, South
American Caiman

Order Chelonia
Family Emydidae

36 Chrysemys picta Schneider, Eastern
Painted Turtle

12 Pseudemys scripta Schoepff, Terra-
pin

4 Terrapene Carolina Linnaeus, Box
Turtle

4 Clemmys guttata Schneider, Spotted
Turtle

Family Chelydridae

6 Chelydra serpentina Linnaeus, Snap-
ping Turtle

Family Kinosternidae

4 Kinosternon flavescens Agassiz, Mud
Turtle

2 Sternotherus odoratus Latreille,
Musk Turtle
Family Trionychidae

4 Trionyx ferox Schneider, Soft-shelled
Turtle

Order Sauria

Family Iguanidae

36 Phrynosoma cornutum Harlan,
Horned Lizard

12 Anolis carolinensis Voigt, Green
Anolis

12 Sceloporus undulatus Latreille,
Fence Lizard

Family Scincidae

24 Eumeces fasciatus Linnaeus, Five-
lined Skink

6 Eumeces obsoletus Baird & Girard,
Great Plains Skink

12 Lygosoma laterale Say, Little Brown

sidnk

Family Gekkonidae

1 Tarantola mauretanica Say, Euro-
pean Gecko

1 Gonatodes fuscus Hallowell, Yellow-
headed Gecko

1 Sphaerodactylus cinereus Wagler,
Ashy Gecko

4 Hemidactylus turcicus Linnaeus,
Turkish Gecko

Family Anguidae

12 Gerrhonotus coeruleus Wiegmann,
Alligator Lizard

6 Ophisaurus ventralis Linnaeus, East-
ern Glass Snake
Family Teidae

8 Cnemidophorus tessellatus Say,
Checkered Race Runner

Order Serpentes
Family Colubridae

48 Thamnophis sirtalis Linnaeus, East-
ern Garter Snake

6 Matrix sipedon Linnaeus, Common
Water Snake

6 Lampropeltis doliata Lacepede, East-
ern Milk Snake

6 Opheodrys vernalis Harlan, Smooth
Green Snake
Family Crotalidae

2 Crotalus viridis Rafinesque, Prairie
Rattlesnake

Family Boidae

2 Boa constrictor Linnaeus, Boa Con-
strictor

4 Charina bottae Blainville, Rubber
Boa

1955]

Hebard & Charipper: Reptilian Adrenal Gland

103

The animals, upon arrival in the laboratory,
were kept in large vivaria, given ample water and
supplied with food. Every attempt was made to
utilize them as soon as possible after arrival,
and only those specimens appearing to be in a
healthy condition were used for histological
study. The animals were anesthetized by intra-
peritoneal injections of sodium Nembutal in a
1:100 solution in 0.9% saline. The adrenals
and immediately-surrounding tissues were rap-
idly removed and placed in fixative. Routine
fixation fluids included 10% neutral formalin,
Bouin’s and Zenker-formal.

Tissues for general histological study were
dehydrated and cleared by the dioxane method
(Lillie, 1948). The tissues were then imbedded
in 56-58° C. Tissuemat, serial sectioned at 5, 7
and 10 microns and the sections attached to
slides with Mayer’s albumen glycerol (Lillie,
1948) . The following stains were used routinely:
Mayer’s acid hemalum and eosin, Masson’s tri-
chrome stain and Heidenhain’s modification of
Mallory’s azocarmine-aniline blue technique
(Lillie, 1948).

The distribution of connective tissue fibers
was studied by a number of methods. Collagen
was demonstrated by the Masson trichrome
stain and the Van Gieson picro-acid fuchsin
collagen method (Lillie, 1948). Weigert’s iron
hematoxylin was used as a nuclear stain with
both methods. Elastic fibers were stained by
Verhoeff’s elastic tissue stain (McClung, 1952) .
To demonstrate reticulum, the Foot modifica-
tion of Hortega’s silver carbonate method was
employed (McClung, 1952). Nerve fibers
within the adrenal gland were stained by the
Pearson-O’Neill (1946) silver-gelatine method
following Bouin fixation.

For the study of mitochondria, tissues were
fixed either in Regaud’s fluid for 4 days and post-
chromated for 7 days in 3% potassium di-
chromate (Lillie, 1948), or by Zenker-formal
fixation followed by post-chromation for 48
hours at 37° C. in 3% potassium dichromate.
The tissues were then dehydrated in graded al-
cohols, cleared in cedarwood oil and transferred
through xylol to paraffin. Sections were cut at
3 microns. Mitochondrial stains used included
Regaud’s hematoxylin, Heidenhain’s iron hema-
toxylin and aniline-acid fuchsin-methyl blue
(Lillie, 1948).

Histochemical studies of cortical lipids were
carried out on tissues fixed for 3 days in 10%
formal-calcium (Baker, 1944). The tissues were
then embedded in gelatin (Zwemer, 1933) and
sectioned on the freezing microtome at 10 and
15 microns. The sections were attached to gel-
atinized slides as recommended by Baker (1946)
and stored in Baker’s formal-calcium-cadmium

preservative. Histochemical tests were applied
to the sections within a two-week period after
the time of fixation. Prolonged storage in for-
malin resulted in a diffuse background staining.

The sudanophilic lipids of the cortical cells
were demonstrated by Sudan black B in propy-
lene glycol without counterstaining (Pearse,
1952). Cholesterol and its esters were demon-
strated by the Shultz test (Lillie, 1948). The
distribution of carbonyl-containing lipids was
investigated by means of the Schiff leucofuchsin
reagent (Lillie, 1948).

Observations

A. Crocodilia

(a) Alligator mississippiensis
Anatomy

The adrenals of the young alligator, Alligator
mississippiensis, are elongate bodies located in
a retroperitoneal position on the midventral sur-
face of the kidneys and extending slightly ante-
rior to the latter with the right adrenal extending
slightly more cephalad than the left. They are
in close association with the ventrally situated
gonads and the laterally placed gonaducts. The
adrenal gland is a discrete structure, separated
from the kidney and reproductive organs by a
distinct connective tissue capsule. The medial
surface of the adrenals borders on the post caval
vein (Fig. 1).

In shape, the adrenals appear ellipsoid in
longitudinal view and ovoid-to-round in cross-
section. The surface of the gland is smooth. In
animals ranging in body length from 60 to 75
cm., the glands vary in size from 10 to 12 mm.
in length and from 2 to 3 mm. in width. They
can be distinguished from the gonads and kidney
by their pinkish-white to yellowish-orange color.

The vascular supply consists of both arterial
and venous blood. Several adrenal arteries
branch off each side of the dorsal aorta and
ramify over the dorsal surface of the gland in
the connective tissue capsule. Small arterioles
branch off and penetrate the parenchyma of the
gland, finally connecting with venous sinuses.
Arising from the dorsal body wall lateral to the
midline, a medium-sized afferent vein enters the
anterior dorsal surface of the adrenal and finally
branches into a capsular network. The right
adrenal is in intimate contact with the vena cava
with a large central venous sinus of the gland
opening directly into the vena cava through a
series of openings or ostia in the wall of the vena
cava. The left adrenal gathers blood into a large
ventro-mesial capsular vein which empties into
the vena cava near the center of the adrenal
gland (Fig. 1).

The adrenal is innervated from the paired
lateral sympathetic ganglia in the region of the

104

Zoologica: New York Zoological Society

[40: 10 !

kidney. Sympathetic nerve fibers from at least
four sympathetic ganglia supply each adrenal,
thus producing a more or less segmented inner-
vation. Large nerves consisting principally of
unmyelinated fibers ramify through the peri-
capsular region (Fig. 2). In frozen sections
stained with Sudan black B, thin rings of su-
danophilic material are often noticed in the
nerve fiber bundles. The nerves appear to be
encapsulated in a sheath of collagenous fibers.
Sympathetic ganglion cell clusters are located on
the dorsal side of the gland in the connective
tissue capsule. Following silver impregnation, a
network of argyrophilic fibers can be distin-
guished. Coarse nerve fibers tend to follow small
blood vessels toward the center of the gland and
then branch out to form networks over the sur-
face of cortical cords and medullary cell masses.
From these networks fine thread-like fibers
branch off and penetrate between the cortical
and medullary cells, terminating as very fine
threads.

Histology

The stroma of the adrenal consists of a dis-
tinct connective tissue capsule composed of col-
lagenous and reticular fibers. The collagenous
fibers are continuous with the surrounding con-
nective tissue while the reticular fibers are more
or less limited to the inner zone of the capsule
and are continuous with the parenchymal reticu-
lum. An inner parenchymal meshwork of
reticular fibers supports the cortical cords and
medullary cells (Fig. 3). While collagenous
fibers are the predominant type found in the
capsule, occasional short elastic fibers are
noticed in the walls of capsular veins. No muscle
tissue, other than that occurring in the media
of small arteries, is present. Infrequent bundles
of collagenous fibers form trabeculae which
penetrate into the center of the gland carrying
arterioles and nerve fibers. Fine collagenous
fibers course between the parenchymal cell cords
as well as in the walls of the vascular sinuses.
Heavy reticular fibers in the inner layers of
the capsule extend inward and provide an ex-
tensive network to support the cortical cells and
medullary masses. Coarse fibers, stretching
lengthwise between the cortical cords, give off
a network of fine reticular fibers at right angles
which enclose the cortical cords and medullary
cell masses in a reticular framework (Fig. 3).

The alligator adrenal is a composite endo-
crine gland composed of cortical and medullary
cells. The medullary tissue consists of a peri-
pheral layer that encircles the cortical tissue
together with small clumps and strands of me-
dullary cells scattered throughout the interspaces
between the cords of cortical cells and the
vascular sinuses (Fig. 2). Intertwining and an-

astomosing solid cords measuring from 25 to
35 microns in diameter constitute the cortical
tissue. Occasional cortical cords extend through
the peripheral medullary tissue and touch upon
the capsular connective tissue. Small nodules
of undifferentiated cortical cells are often found
scattered in the capsule. An intricate network of
capillaries and venous sinuses pervades the cor-
tical and medullary tissue.

The zonation typical of the mammalian ad-
renal is lacking in the alligator adrenal. The
peripheral cords tend to be made up of elong-
ated cells in a more or less radial arrangement
while the central region of the gland consists
of cords made up of irregularly dispersed poly-
hedral cells having centrally placed nuclei. These
central cords often appear to be oriented toward
the side of the gland bordering on the vena
cava (Fig. 1) . The venous sinuses are lined
with an endothelium having flattened granular
nuclei.

Cytology

The medullary tissue is composed of two
cell types with different tinctorial and morpho-
logical characteristics. These differences are best
observed after Zenker-formal fixation and
staining with Masson’s or with Heidenhain’s
azan stain. A peripheral zone of yellow-stain-
ing cells can be readily distinguished from the
more centrally situated reddish-brown cells
(Fig. 2). The yellow-stained cells will be re-
ferred to as the “light” cells and the reddish-
brown cells as “dark” cells. The “light” cells are
15 to 18 microns in diameter with a centrally
located nucleus measuring 5 to 7 microns and
are spherical to polyhedral in shape. These
“light” cells have a finely granular cytoplasm
and occur singly, in small clumps and strands,
and in rather large masses in the peripheral re-
gions of the adrenal (Fig. 2). The majority of
them are restricted to the capsular region; how-
ever, occasional strands extend inward among
the cortical cords, especially in the connective
tissue trabeculae penetrating the gland. The
“light” medullary cells are often found in as-
sociation with sympathetic ganglion cells in the
capsule or in the adjacent connective tissue.

The “dark” cells occur singly or in small
clumps or strands in the spaces and sinuses
which pervade the cortical tissue (Figs. 2 & 5).
This close association of “dark” medullary cells
with the cortical tissue is strikingly evident in
all tissues examined. These cells are, as a rule,
smaller in size, measuring from 10 to 12 mic-
rons, and are generally polyhedral in shape. The
granules in the cytoplasm are coarser and ap-
pear reddish-brown after staining with Mas-
son’s. The nucleus measures from 4 to 5 mic-
rons in diameter and is spherical, with 1 or 2
nucleoli.

1955]

105

Hebard & Charipper: Reptilian Adrenal Gland

In sections, the cords of cortical cells appear
as spherical-to-elongate anastomosing masses of
cells usually consisting of 1 to 4 tiers of cells
(Fig. 2). The shape of the cells in the cord
varies from irregular to greatly elongated
spindle-shaped or wedge-shaped. The latter con-
figurations are probably due to compression
forces, since they are found more often in the
tips of peripheral cords abutting against the cap-
sule. The nuclei are centrally located, measur-
ing 5 to 6 microns in diameter, and usually
possess only one large fuchsinophilic nucleolus.
Following Zenker-formal fixation and Masson’s
or Heidenhain’s azan stain, the cytoplasm has
a slightly fuchsinophilic and finely reticulated
appearance.

Mitochondria are best demonstrated after
post-chromation of Zenker-formal fixed mate-
rial and staining with aniline-acid fuchsin and
methyl blue. The small, granular, red-staining
mitochondria are scattered throughout the cyto-
plasm (Fig. 5). The amount and distribution of
mitochondria varies with the lipid concentration
in the cortical cells. Lipid-poor cells exhibit a
strong fuchsinophilia with the mitochondria
concentrated in the perinuclear zone. Lipid-rich
cells have granular mitochondria dispersed in
the cytoplasmic matrix between the lipid vacu-
oles. One to several relatively large fuchsino-
philic spheres are often seen in the cytoplasm in
the vicinity of the nucleus.

In Baker’s formal-fixed frozen sections stained
with Sudan black B, the medullary cells and
surrounding tissue remain colorless in contrast
to the intense blue-black staining of the cortical
cells (Fig. 4). The sudanophilic lipids are evenly
distributed throughout the cortical tissue. Under
high magnification the sudanophilic droplets
appear to be more or less evenly distributed
throughout the cytoplasm. There are differences
in the amount of sudanophilic substances pres-
ent: small undifferentiated cells possess little
sudanophilic material, whereas other cells (par-
ticularly in peripheral cords) are filled with
sudanophilic droplets.

Frozen sections subjected to the Shultz chol-
esterol test develop an intense blue-green color
in the cortical tissue only. Medullary tissue, con-
nective tissue and kidney tissue give a negative
reaction to this procedure. The Shultz-positive
droplets have a uniform distribution throughout
the cytoplasm of the cortical cells, similar to the
sudanophilic droplets. An intense and selective
staining of frozen sections is observed upon
their treatment with Schiff leucofuchsin. A
purplish-red color develops throughout the
cortical tissue. Except for the interesting obser-
vation that the elastic fibers of blood vessels
stain a purple color, all other tissues in the sec-

tions are Schiff-negative. The distribution par-
allels that of the sudanophilic lipids.

(b) Caiman crocodilus

The animals used in this study were very
young specimens, still having an internal rem-
nant of the yolk-sac and a visible umbilical scar
on the abdomen. The adrenal of the immature
caiman is similar in anatomical and histological
features to the adrenal of the alligator. The only
significant difference noted was in connection
with the cortical tissue. The cortical cells are not
as well differentiated in the center of the gland
as are the peripheral cells. The centrally located
cells are small and polyhedral and irregularly
arranged in cords, while those at the periphery
are of a tall columnar type.

B. Chelonia

1. Family Emydidae
(a) Chrysemys picta
Anatomy

The adrenal glands of the painted turtle,
Chrysemys picta, are paired bodies, yellowish
in color, extending along the ventro-mesial sur-
face of the kidneys in a retroperitoneal position
(Fig. 6). They have an irregular outline, often
appearing to consist of partially-fused spherical
or irregularly-shaped masses imbedded around
the numerous efferent renal veins. The gland is
not completely encapsulated, but there is a close
association and often an intermingling of renal
tissue and adrenal tissue on the dorsal side of
the adrenal. The ventral surface of the gland is
covered by the pleuroperitoneum which, at the
ventro-lateral edge, gives rise to the mesentery
supporting the gonad.

There is an interesting difference in the an-
atomical relationships of the kidney, adrenal and
gonad in the turtle, when compared with those
of the alligator. The gonad of the young alli-
gator is attached to the ventral surface of the
adrenal, whereas the gonad in the turtle is sepa-
rate and suspended in a mesenteric fold of the
pleuroperitoneum (Figs. 1 & 6).

The arterial blood supply consists of several
small adrenal arteries branching off from the
dorsal aorta, as well as small branches arising
from the genital and renal arteries that border or
penetrate through the adrenal. Because of the
unique location of the turtle adrenal with respect
to the blood vessels of the kidney, it is difficult
in histological sections to determine the identity
of the veins and venous sinuses penetrating
through the parenchyma of the gland. Doubt-
less, most of the sinuses represent efferent renal
sinuses; however, some may be either sinuses
of the renal portal veins or sinuses of afferent
adrenal veins, all of which constitute a complex
venous plexus in the region of the adrenal.

106

Zoologica: New York Zoological Society

[40: 10

Numerous sympathetic nerve ganglia are
found in the turtle adrenal, occuring most often
in the connective tissue between the adrenal and
the kidney or in the trabeculae which penetrate
into the adrenal (Fig. 8). Following the silver-
gelatine technique, a dense network of nerve
fibers can be distinguished surrounding the
medullary and cortical masses. A heavy network
of fibers encircles and penetrates among the
medullary cells, while in contrast only occa-
sional fine, faintly-stained fibers penetrate the
cortical tissue.

Histology

The stroma of the turtle adrenal is similar to
that described for the alligator, except that in
the turtle the gland is not completely encapsu-
lated. A distinct capsule composed of collagen-
ous and reticular fibers covers the ventral sur-
face. Numerous trabeculae of collagenous and
reticular fibers extend into the parenchyma,
often appearing to separate the gland into sepa-
rate masses of tissue. The columns and clumps
of cortical and medullary tissue are outlined by
a dense network of reticular fibers which serve
as a framework to maintain the structural unity
of the gland. Elastic fibers are limited to the
adventitia of arteries and veins occurring in the
capsule and trabeculae.

The adrenal of the painted turtle consists of
intermixed cortical and medullary tissue (Fig.
7 ) . The medullary tissue is not concentrated in
any particular region of the gland, but is irregu-
larly distributed among the cortical tissue as
clusters of medullary cells. These clusters are of
irregular size and shape, varying from small
groups containing a few cells to large clumps
and strands. It often appears that these clusters
and strands lie in the vascular sinuses (Fig. 7).

The cortical tissue is composed of anastomos-
ing and intertwining columns varying in width
from approximately 40 to 70 microns. The col-
umns are made up of several tiers or rows of
cells which usually have an irregular arrange-
ment (Figs. 7 & 8). Sometimes the cells in peri-
pheral cords in longitudinal section have the
appearance of being compressed into a single
row of greatly elongated and flattened cells
(Figs. 7 & 10).

Cytology

The medullary tissue consists of two cell-
types which can be distinguished principally by
their staining properties (Fig. 7) . The peripheral
medullary cells, after Zenker-formal fixation
and Masson’s stain, have a light brownish-yel-
low color in contrast to the darker yellowish-
brown cells found in the more central region of
the gland. The “light” cells occur singly or in
small clumps or strands in the peripheral cap-

sular region, and sometimes extend for a short
distance in toward the center of the gland in
the intercortical spaces, especially in trabeculae
that penetrate the gland. The “dark” cells occur
singly or in small clumps or strands either im-
bedded in the side of cortical cords or in the
intercortical spaces and sinuses.

A striking histological feature of the “dark”
cells is the fact that they are usually in close
contact with cortical tissue. On the other hand,
the “light” cells are more often separate from
the cortical tissue, either in the peripheral cap-
sule or in the surrounding connective tissue.
They are also found in association with sym-
pathetic ganglia (Fig. 8) , in the adventitia of the
dorsal aorta and in the parenchymal connective
tissue of the kidney. Both cell types are poly-
hedral-to-spherical in shape, varying from 9 to
12 microns in diameter. The nucleus varies in
size from 5 to 7 microns.

The cortical cells are variable in size and
shape. The most common morphological type is
an irregularly-shaped polyhedral cell (Figs. 7
& 10). Sometimes, due to compression forces
on the column, the cells assume a tall, columnar-
to-spindle-shaped form (Fig. 10). The centrally
located nuclei are 5 to 6 microns in diameter,
are slightly reticulated and usually have a single
large hyaline nucleolus. Fuchsinophilic spheres
are commonly found in a juxtanuclear position.

The mitochondria appear as small granules
scattered throughout the cytoplasm, with the
number of mitochondria depending upon the
amount of lipid in the cell (Fig. 10). Highly
vacuolated cells have few mitochondria, where-
as cells having little lipid are mitochondria-rich.
Only an occasional mitotic figure was noted in
the cortical cells. The lack of any concentration
of undifferentiated cells in the columns and a
random dispersion of mitotic figures suggests
that cell replacement must take place within the
cords in a random fashion.

The cortical cell lipids are sudanophilic,
Shultz-positive and Schiff-positive, with a uni-
form and parallel distribution throughout the
cortical tissue. Under high magnification, the
sudanophilic droplets appear to be evenly dis-
tributed throughout the cytoplasm and no zona-
tion of the gland is seen (Fig. 9).

(b) Pseudemys scripta, Clemmys guttata,
Terrapene Carolina

The adrenals of the other members of the
family Emydidae included in this study, namely,
Pseudemys scripta, Clemmys guttata and Ter-
rapene Carolina, resemble (both anatomically
and histologically) the descriptions for Chry-
semys picta. The only significant histological
difference is found in the adrenals of the spotted

1955]

Hebard & Charipper: Reptilian Adrenal Gland

107

turtle, Clemmys guttata. In this species there
is considerably more medullary tissue. In trans-
verse sections, the intermingling clumps and
strands of medullary and cortical cells appear
to be approximately equal in size and number.

2. Family Chelydridae

The adrenals of the snapping turtle, Chely-
dra serpentina, show no significant morpholog-
ical differences from the description of Chry-
semys picta.

3. Family Kinosternidae

The morphology of the adrenals of the mud
turtle, Kinosternon flavescens, and the musk
turtle, Sternotherus odoratus, resembles that of
Chrysemys picta in general features. The glands
in transverse sections appear more elongated and
closely associated with the kidneys. The medul-
lary tissue is often located in the adjacent renal
tissue as yellow clumps among the renal tubules.
The cortical cells show a tendency toward a
regular arrangement in cords measuring approx-
imately 50 microns in diameter. The cells are
tall, columnar in shape and usually radially ar-
ranged in the cords so that in longitudinal sec-
tions they appear as a double row of cells. The
nuclei are usually located near the end of the
cell distal to the vascular sinus.

4. Family Trionychidae

The adrenals of the immature soft-shelled
turtle. Trionyx ferox, resemble the adrenals of
immature Chrysemys picta both in anatomical
and histological features. In young animals such
as these, the cortical cells are grouped in narrow
cords and are not well differentiated because
they have little cytoplasm or lipid material.

C. Sauria

1. Family Iguanidae
(a) Phrynosoma cornutum
Anatomy

The adrenal glands of the horned lizard,
Phrynosoma cornutum, are paired bodies lo-
cated near the anterior ends of the gonads and
enclosed in the mesorchia or mesovaria sup-
porting the gonads. In the male, the adrenals
are intimately associated with the sperm ducts
(Fig. 11) . The right adrenal is slightly cephalad
to the left, extending along the vena cava to
the right lobe of the liver.

The glands are elongate, oval-shaped bodies
measuring approximately 2 to 3 millimeters in
length and 1 millimeter in width. The combined
weight of the freshly-dissected glands in adult
animals is approximately 1 milligram.

The blood supply consists of both arterial and
venous blood. Each adrenal receives a branch

from the dorsal aorta. These branches arise
from the aorta at the origin of intervertebral
arteries. In some specimens, the left adrenal re-
ceives two arteries from the aorta. A small af-
ferent vein arises from the dorsal body wall
cephalad to the adrenal and enters the anterior
pole of the gland. The left adrenal is in con-
tact with the cephalad end of the left efferent
renal vein into which blood from the gland
drains directly. The right adrenal is situated at
the junction of the efferent renal veins and an-
teriorly along the vena cava. Blood from this
gland drains directly into the vena cava.

Histology

The adrenal gland is enclosed in a thin con-
nective tissue capsule composed of an outer
layer of collagenous fibers which are continuous
with the surrounding connective tissue, and an
inner layer of reticular fibers which are contin-
uous with the reticular framework supporting
the parenchyma of the gland. Numerous trabe-
culae containing collagenous fibers and retic-
ular fibers penetrate the parenchyma (Fig. 12).

Sympathetic ganglion cells are aggregated
principally in the dorsal region of the capsule
near the adrenal artery (Fig. 12). Occasional
sympathetic cells are found in association with
groups of medullary cells in other regions of the
capsule. The adrenal gland of the horned lizard
consists of cortical and medullary tissue ar-
ranged in a typical saurian pattern (Fig. 12).
The medullary component comprises an incom-
plete dorsal capsule of primarily “light” type
medullary cells with a few strands often extend-
ing inward from the periphery. In addition,
small groups and individual cells of the “dark”
type are interspersed in the parenchyma.

The cortical tissue consists of interwining and
anastomosing cords measuring 30 to 50 microns
in diameter (Fig. 12). In transverse section, a
typical cord appears as a radially arranged
group of 15 to 20 tall columnar cells with very
indistinct cell outlines. The nuclei are regularly
arranged in a zone near the periphery of the
cord. In longitudinal section, a cord often pre-
sents the characteristic appearance of a double
row of tall columnar cells with nuclei arranged
in a double row near the peripheral edge of the
cord. The central area of the cord, following
Masson’s staining, has a pale fuchsinophilic
vacuolated appearance.

The cephalic pole of the right adrenal is often
fused with the caudal tip of the dorsal lobe of
the liver, both tissues being in contact with and
partially encapsulating the vena cava in this
region. The liver parenchyma is not distinctly
separated from the cortical tissue by a connec-
tive tissue capsule and as a result small groups

108

Zoologica: New York Zoological Society

[40; 10

of cortical cells in many sections appear to be
isolated in the liver parenchyma.

Cytology

The medullary tissue consists of two cell-types
(Fig. 12). The strands and masses of medullary
cells in the peripheral capsular region are of the
“light” type, having few granules and appearing
light-yellow after Zenker-formal fixation. The
medullary cells in the central part of the paren-
chyma are of the “dark” type, staining a red-
dish-brown after the Masson or azan technique.
Both cell types are similar in their morphology,
consisting of polyhedral cells which are 9 to 12
microns in diameter. The centrally located nu-
cleus varies from spherical to oval in shape and
measures 4 to 5 microns in diameter and has a
granular appearance with one or more nucleoli.
Granular mitochondria are dispersed throughout
the cytoplasm.

The anastomosing cords of cortical tissue are
composed of polyhedral cells of varying height
(Fig. 12). The most frequently observed type
is a tall, wedge-shaped columnar cell approxi-
mately 20 microns in height and 7 microns in
width at the base. The spherical-to-oval-shaped
nucleus is 4 to 5 microns in diameter and con-
tains one or more nucleoli. Generally, only a
single large fuchsinophilic nucleolus is present.

The mitochondria in the cortical cells appear
as fine granules, either concentrated in the per-
inuclear zone or dispersed in the delicate cyto-
plasmic network surrounding the lipid vacuoles.
The lipid droplets are evenly distributed through-
out the cytoplasm and give a strong, positive
color reaction to the Sudan procedure (Fig. 13) .
There is a general and parallel distribution of
sudanophilic lipids, carbonyl lipids and choles-
terol esters.

(b) Anolis carolinensis

The adrenals of the green anolis, Anolis carol-
inensis, are similar in anatomical features to
Phry nosoma (Fig. 14) . The histological picture
differs somewhat in that the cortical cords ap-
pear to be narrower (25 to 35 microns) and are
made up of radially arranged low columnar
cells. The nucleus is centrally located or may, in
some cords, be located near the edge of the cell
bordering on the blood supply. Fewer medullary
cells are seen in the parenchyma of the gland
with most of them being concentrated on the
dorsal surface of the gland.

(c) Sceloporus undulatus

The adrenals of the fence lizard, Sceloporus
undulatus, resemble, in general, those of Anolis.
In this species the cortical cells are of a low col-
umnar type with a tendency toward exhibiting
a polarity by the display of nuclei at the peri-
pheral edges of the cortical cells.

2. Family Scincidae

Anatomy

The anatomical relationships of the adrenals
in the three species of skinks that were studied,
namely, Eumeces fasciatus, Eumeces obsoletus
and Lygosoma laterale, are in general similar to
the saurian pattern described previously for the
iguanid lizard, Phyrnosoma. The glands are
elongate bodies closely associated with the
gonads and gonaducts. The right adrenal in this
group of lizards does not make contact or fuse
with the dorsal lobe of the liver; however, the
mesentery of the right gonad is continuous with
the mesentery supporting the dorsal lobe of the
liver. A single afferent adrenal vein enters the
caudal pole of each adrenal. A large sympathetic
ganglion is located near the cephalic pole of each
adrenal.

Histology

The gland is enclosed in a thin capsule con-
tinuous with the surrounding mesentery and
connective tissue. The parenchyma of the gland
is supported by a network of thick reticular
fibers.

The disposition of the cortical and medullary
tissue corresponds to the saurian pattern de-
scribed for Phrynosoma. The medullary cells
form a partial capsule around the dorsal surface
of the cortical tissue. Some medullary tissue in
the form of individual cells and small clumps is
diffused throughout the parenchyma. The med-
ullary cells are of two tinctorial types: the “light”
cells which are confined chiefly to the periphery;
and the deeper-lying cells which are usually of
the “dark” type.

The arrangement of the cortical cells within
the cords is a unique and significant feature in
the Scincidae (Figs. 18 & 23). In transverse sec-
tion, a cord consists of 10 to 12 tall, wedge-
shaped columnar cells arranged in a radial pat-
tern with the nuclei abutting against the peri-
pheral edge of the cell. In longitudinal section,
the cord appears to consist of two rows of tall
columnar cells with parallel rows of nuclei at
the edges of the cord. This marked polarity of
the cortical cells is present in all three species of
skinks examined. Occasionally, a lumen is ob-
served in the center of a cord, giving the cord
an appearance that resembles an exocrine gland.
These central cavities are not a consistent fea-
ture and probably are artifacts resulting from
shrinkage.

Cytology

The cortical cells display a marked polarity
with the nucleus abutting against the edge of the
cell facing the vascular supply. The region of
the cell distal to the nucleus contains a faintly-

1955]

Hebard & Charipper: Reptilian Adrenal Gland

109

staining cytoplasmic network surrounding lipid
vacuoles of varying size. Following aniline-acid
fuchsin and methyl blue, the red-staining mito-
chondria are found widely scattered in the cyto-
plasmic network (Fig. 19). The typical cortical
cell is a lipid-rich, mitochondria-poor type. Un-
differentiated cortical cells are not commonly
seen in the capsule, which would seem to suggest
that cortical cell replacement occurs within the
individual cords and not by migration.

The unique polarity of the cortical cells is
clearly demonstrated by histochemical tests of
the lipid droplets. In frozen sections where a
cord happens to be cut lengthwise, the blue-
black sudanophilic lipids are confined to a cen-
tral zone of the cord and the nuclear zone at the
periphery fails to stain (Fig. 20). A similar pat-
tern of distribution of Shultz-positive and Schiff-
positive droplets is found. No significant dif-
ference in the uniformity of staining is apparent.
The medullary and surrounding connective tis-
sue do not give a positive reaction to any of the
tests applied.

An intense blue-green color reaction is ob-
tained by the Shultz test for cholesterol and its
esters. The distribution of the blue-green drop-
lets is localized in the cytoplasm distal to the
nucleus. Following application of the Schiff
leucofuchsin technique, an intense red color de-
velops exclusively in the cortical tissue. The
intensity and specific localization of the color
reaction in the cytoplasmic region of the cortical
cells indicates that the cortical tissue, following
Baker’s fixation, contains a high level of car-
bonyl lipids.

3. Family Gekkonidae

A limited number of specimens from this
family was available for histological study. The
series consisted of a single adult European
gecko, Tarantola mauretanica; one adult yellow-
headed gecko, Gonatodes fuscus; one adult ashy
gecko, Sphaerodactylus cinereus; and four adult
Turkish geckos, Hemidactylus turcicus. The
anatomical relationships in these specimens re-
semble the general saurian pattern already de-
scribed. The disposition of the cortical and
medullary tissue is similar to that noted in other
groups of lizards.

A special and significant feature of the gecko
adrenal is the extreme polarity of the cortical
cells (Figs. 21 & 22). TTie arrangement of the
cells in the cords is similar to that described for
the Scincidae. The cortical tissue, in transverse
or longitudinal sections, appears as spherical-
to-oblong anastomosing cords measuring ap-
proximately 50 microns in diameter. The cells
in longitudinal sections of a cord appear as tall,
wedge-shaped columnar cells with their nuclei

abutting on the peripheral edges of the cells
facing a capillary or venous sinus. Because of
the extremely small size of these animals and
the consequent small size of the gland, a typical
transverse section appears to consist of 3 or 4
anastomosing cords.

A single specimen of the Dibamidae, Dibamus
novae-guineae, and two representatives of the
Pygopodidae, Lialis burtoni and Pygopus lepi-
dopus, were examined histologically and found
to have a gekkonid arrangement of the cortical
cells.

4. Family Teidae

The anatomical features of the adrenals of
the checkered race runner, Cnemidophorus tes-
sellatus, conform to the saurian pattern already
described for Phrynosoma. The microscopic
anatomy, however, shows certain characteristics
which are distinctive of this species (Fig. 17).
The medullary tissue is more widely dispersed
within the cortical tissue. In transverse sections,
a dorsal aggregation is present, as well as numer-
ous clumps scattered throughout the paren-
chyma of the gland. Both “light” and “dark”
cells are present with the “light” cell type more
or less restricted to the dorsal aggregation or to
extra-adrenal locations.

The cortical cells do not exhibit a consistent
regular arrangement, but generally appear to
have a more or less random dispersion (Fig. 17) .
Occasional cords, 20 to 30 microns in diameter,
present a radial and polarized arrangement of
tall columnar cells with nuclei aligned on the
periphery. Most of the cells are low columnar to
polyhedral in shape, measuring approximately
10 microns in height and 7 microns in width.
The nuclei are usually centrally located or are
found toward the edge of the cell facing a vas-
cular sinus.

5. Family Anguidae
(a) Gerrhonotus coeruleus

The morphology of the adrenals of the alli-
gator lizard, Gerrhonotus multicarinatus, has
been described by Retzlaff (1949). They exhibit
the general saurian anatomical relationships ob-
served for other lizards. The significant and spe-
cific characteristic feature of the anguimorphid
type of adrenal is illustrated by Gerrhonotus
coeruleus. The cortical cells are irregularly
shaped and do not arrange themselves in definite
cords or columns (Fig. 15). Only cortical cells
bordering on a vascular sinus tend to arrange
themselves in a regular fashion around the sinus
with the nuclei near the side of the cell which
faces the sinus. Most of the cells have centrally
located nuclei. An unusual feature of the cortical
tissue of Gerrhonotus is the presence in large
individuals of a narrow peripheral zone of cor-

110

Zoologica: New York Zoological Society

[40: 10

tical cells that stain less intensely than the rest
of the cortical cells (Fig. 15). No histochemical
tests were made on the few animals available to
determine the nature of the lipids in this peri-
pheral zone.

(b) Ophisaurus ventralis

The adrenals of the glass snake, Ophisaurus
ventralis, exhibit the general saurian pattern. In
this species, the body is greatly elongated like
that of a snake, yet the adrenals and associated
reproductive organs display only a slight asym-
metry—the right gonad being slightly more
cephalad than the left, but no more so than the
usual disposition in the Sauria. The microscopic
anatomy of this species is similar to the anguid
lizard, Gerrhonotus, just described.

D. Serpentes

1. Family Colubridae
( a ) T hamnophis sirtalis
Anatomy

The adrenals of the Eastern garter snake,
T hamnophis sirtalis, are discrete bodies closely
associated with the asymmetrically-located
gonads (Fig. 24). They can be distinguished
from the latter by their golden-yellow color and
difference in texture. The left adrenal lies an-
terior to the left kidney near the cephalic end
of the left efferent renal vein and extends ante-
riorly for a short distance alongside the gonad.
With reference to the ventral scutes or gastro-
steges, the left adrenal lies approximately be-
tween the 31st to 34th gastrostege in females
and between the 35th to 37th gastrostege in
males, counting anteriorly from the anal plate.
The right adrenal lies alongside the vena cava
just anterior to the junction of the efferent renal
veins and is closely associated with the caudal
end of the right gonad. The right adrenal lies
approximately between the 45th and 48th gas-
trostege in both sexes. The glands are contained
in the mesovaria or mesorchia supporting the
gonads (Fig. 24). In the male, there is a close
association with the vas deferens.

The adrenals are elongate bodies varying in
length from 5 to 15 mm., depending on the body
length of the snake. Usually the right adrenal
is slightly longer than the left. Occasionally a
gland appears to be constricted into a bilobed
shape. In transverse section, the gland is ellip-
tical-to-round with a diameter of 1 to 3 milli-
meters. Adrenal weights for a series of 13 garter
snakes of mixed sex range from 0.03 to 0.08 per
cent, of the body weight, with an average of
0.045 per cent.

The garter snake adrenal is an extremely vas-
cular organ having both an arterial and a venous
blood supply. The dorsal aorta gives off to each

adrenal a branch (the spermatic or ovarian
artery) which enters the mid-dorsal region of the
gland, bifurcates, sending a branch anteriorly
through the dorsal pericapsular connective tis-
sue, and continues into the testes or ovary (Fig.
24) . The posterior branch continues caudally
through the dorsal pericapsular layer of the
adrenal and finally joins the gonaduct arteries.
Numerous arterioles branch off from these
longitudinal arteries to enter the parenchyma of
the gland.

A separate venous supply can readily be seen
on gross examination. Two (sometimes only
one) afferent veins emerge from the dorsal body
wall slightly lateral to the midline and enter the
mid-dorsal region of the gland. Usually the
second vessel emerges from the body wall just
posterior to the adrenal and joins a longitudinal
vein in the mesotuberium which connects with
the anterior branch from the body wall.
Branches from this venous supply connect with
the venous sinuses and capillaries in the paren-
chyma of the gland.

The blood from the left adrenal gland is car-
ried by a number of small, extremely short, ef-
ferent adrenal veins opening into the left efferent
renal vein. The blood from the right adrenal
drains into the postcaval vein either by means
of numerous short efferent veins or is so inti-
mately associated with the vein that blood from
the parenchymal sinuses drains directly into the
vena cava through small openings (Fig. 24).

A large aggregation of sympathetic ganglion
cells is located in the dorsal capsular region,
usually near the artery. Occasional sympathetic
nerve cells are seen in the capsule of the gland;
however, they were not found in the paren-
chyma of the gland. The medullary cells are
innervated by an extensive network of nerve
fibers, although nerve endings to the cortical
cells cannot be definitely distinguished in the
silver carbonate preparations.

Histology

The garter snake adrenal, fully encapsulated
by the connective tissue, is almost completely
enveloped in the surrounding mesentery (Figs.
24 & 25). The capsule consists chiefly of colla-
genous fibers with a few fine reticular fibers in
the innermost layers. The parenchyma of the
gland is supported by a thick network composed
primarily of collagenous fibers. Fine reticular
fibers are found in the walls of the vascular
sinuses, where they form a fine network envel-
oping the cords of cortical tissue.

The medullary and cortical tissues are dis-
posed in a pattern somewhat similar to that of
the lizards (Fig. 25). The greater part of the
medullary tissue is concentrated in the dorsal

1955]

He bard & Charipper: Reptilian Adrenal Gland

111

pericapsular region in the vicinity of the sper-
matic or ovarian artery. In transverse section,
a large mass of “light” yellow cells is seen
aggregated around the dorsally located artery.
Medullary cells of the “dark” type are found
scattered in and among the cords of cortical
cells singly or in small groups of three or four
cells (Figs. 25 & 28).

The cortical tissue exhibits a unique and char-
acteristic arrangement (Figs. 25 & 28). The
cortical cells are arranged in anastomosing and
intertwining cylindrical cords measuring 40 to
50 microns in diameter. In transverse section, a
cord appears spherical and to consist of 12 to 15
radially-disposed tall, wedge-shaped columnar
cells. The nuclei are regularly arranged in a
small circle in the center of the cord. In longi-
tudinal section, a cord appears to consist of two
rows of tall columnar cells with a double row of
nuclei running lengthwise in the center of the
cord. A common occurrence in the garter snake
adrenal is the presence of strands and large
masses of undifferentiated cortical cells in the
capsule (Fig. 25). These masses of cells, if fol-
lowed in serial sections, are often found to be
continuous with the parenchymal cortical tissue.
The imdififerentiated cells have a very limited
amount of cytoplasm which, after Masson’s
staining, appears strongly fuchsinophilic.

The garter snake adrenal is an extremely vas-
cular gland and this is made apparent in histo-
logical preparations by the large and extensive
network of vascular sinuses that pervades the
gland (Fig. 25). These sinuses are lined with a
thin endothelium having extremely flattened
nuclei.

Cytology

Two tinctorial types of medullary cells are
found in the garter snake adrenal. The cells ag-
gregated near the dorsally-located artery have a
yellow, finely-granular appearance after Zenker-
formal fixation and Masson’s stain. The small
clumps and individual cells associated with the
cortical tissue possess coarse granules and stain
reddish-brown. Both types are generally poly-
hedral in shape and measure 12 to 14 microns in
diameter.

The cortical cells are usually of a tall col-
umnar type measuring 20 to 25 microns in
height. There is neither notable sexual variation
nor discernable seasonal difference in cell height
in the series of garter snakes which was collected
in April, July and October. The cells exhibit a
marked polarity with the nucleus usually lo-
cated near the end of the cell distal to the blood
supply (Figs. 25 & 28). The nuclei are 4 to 5
microns in diameter, spherical-to-oval in shape,
and exhibit from one to three prominent
nucleoli.

The mitochondria are granular bodies occur-
ring throughout the cytoplasm (Fig. 28). In
those cells with a high lipid concentration, the
mitochondria are few in number and widely
scattered in the cytoplasm surrounding the lipid
vacuoles. In the lipid-poor undifferentiated cells,
a heavy concentration of mitochondria pervades
the cytoplasm. There is an inverse ratio between
the amount of mitochondria and lipid in the
cortical cells. Fuchsinophilic spheres larger
than mitochondria occur frequently in the
cytoplasm.

An outstanding feature of the garter snake
adrenal is the distribution of lipid inclusions in
the cortical cells. The lipid droplets are osmio-
philic, sudanophilic, Schultz-positive, and ex-
hibit an intense positive reaction to the Schiff
leucofuchsin reagent. The lipid droplets appear
to be evenly distributed throughout the cortical
tissue and no tendency toward zonation is ob-
served (Fig. 26).

Within the cells, the lipid droplets are con-
centrated at the peripheral end of the cell and
at the end opposite the nucleus. This polarity is
strikingly demonstrated in frozen sections
stained with Sudan black B (Fig. 26), or by the
Shultz technique for demonstrating cholesterol
and its esters (Fig. 27). A similar distribution
of Schiff-positive carbonyl lipids is seen, there
being a parallel pattern of distribution of the
sudanophilic, Schultz-positive and Schiff-posi-
tive lipids. By chance, one of the frozen sections
tested for cholesterol was observed to have a
mass of undifferentiated cells in the capsule
(Fig. 27) . Only a slightly positive reaction was
seen in this nodule.

(b) Matrix sipedon

The adrenals of the common water snake,
Matrix sipedon, are in general similar in mor-
phology to those of Thamnophis. Medullary tis-
sue is aggregated in the dorsal region of the cap-
sule and small groups of cells are scattered
throughout the cortical tissue. The cortical tissue
exhibits a regular arrangement of polarized col-
umnar cells into anastomosing cords. The polari-
zation is similar to that of the garter snake.

(c) Lampropeltis doliata

The morphology of the adrenals of the East-
ern milk snake, Lampropeltis doliata, is similar
to that of Thamnophis and Matrix.

(d) Opheodrys vernalis

The adrenals of the smooth green snake,
Opheodrys vernalis, are similar in anatomical
features to those of the other colubrid snakes
mentioned. Histologically, however, the paren-
chyma of the gland is more compact. The cor-
tical tissue is arranged in radial cords of low

112

Zoologica: New York Zoological Society

[40: 10

columnar cells. The nuclei are located distal to
the capillaries.

2. Family Crotalidae

Only two specimens of the prairie rattlesnake,
Crotalus viridis, were studied. Both of these
animals had been preserved in 10% formalin
for an unknown period of time. The adrenals
were poorly fixed, yet on staining with Mayer’s
acid hemalum and eosin, sufficient detail was
present to warrant comment. Anatomically, the
glands are similar to those described for the
garter snake. Microscopically, a typical ophidian
disposition of cell types is seen. The medullary
tissue is confined primarily to the dorsal capsular
region while the cortical cells are regularly ar-
ranged in anastomosing cords. The nuclei show
a similar disposition and polarity to that noted
for the garter snake.

3. Family Boidae

(a) Boa constrictor

The adrenals of the boa constrictor are
similar, for the most part, to those of the colu-
brid snakes. There is a tendency toward an
elongation of the right gonad and a consequent
elongation of the right adrenal. The microscopic
anatomy is the typical ophidian type with a
dorsal cap of medullary cells plus small clumps
and single cells scattered in the cortical tissue.
The cortical tissue consists of anastomosing
cords made up of radially arranged low col-
umnar cells exhibiting a slight polarity. The
nuclei are arranged near the center of the cords
or distal to the blood sinuses.

(b) Charina bottae

The adrenals of the rubber boa, Charina
bottae, are elongated thread-like glands located
in the mesenteries supporting the gonads. The
left adrenal measures approximately 10 milli-
meters in length and 1 millimeter in width,
whereas the right adrenal is approximately 25
millimeters in length and 1 millimeter in width.

The vascular supply of these glands consists of
a single artery to the left adrenal and three
arteries to the right adrenal. Blood leaves the
glands by a series of short efferent adrenal veins,
6 for the left adrenal and 12 for the right
adrenal. The glands are not attached directly to
the vena cava, as noted in other snakes, but both
glands empty into the vena cava, the left adrenal
lying just anterior to the junction of the efferent
renal veins.

In order to determine the arrangement of the
cortical and medullary tissue, histological prep-
arations were made of glands removed from
museum specimens. Formalin-fixed tissues such
as these will, on staining with Weigert’s iron

hematoxylin, give fairly good differentiation of
cortical and medullary cells. The medullary cells
exhibit a strong basophilia and stain dark blue.
The adrenal has a regular ophidian distribution
of glandular tissue with a dorsal aggregation of
medullary cells and the cortical cells arranged
in cords. In sections the cords appear to be made
up of several layers or tiers of cells.

Discussion

Anatomy

The adrenal glands of reptiles exhibit a con-
sistent association with the reproductive organs
and the venous system. These anatomical rela-
tionships are found in all four orders of the
reptiles studied.

In the Crocodilia, the adrenals are discrete,
elongate bodies located in a retroperitoneal posi-
tion on the midventral surface of the kidney in
close association with the ventrally situated
gonads and gonaducts. This relationship has
been described for the alligator (Pettit, 1896;
Spanner, 1929; Reese, 1931; Lawton, 1937),
and the caiman (Pettit, 1896), and is confirmed
in the present study.

A consistent anatomical relationship is found
in all the species of North American turtles
examined in this study. The adrenals are located
on the ventral surface of the kidneys in close
association with the gonaducts and gonads. The
latter, however, are separate from the adrenal,
being suspended ventrally from the adrenal-
kidney complex by the mesorchia or mesovaria.
The adrenal is not as well organized into a dis-
crete gland as is that of the Crocodilia, but,
rather, usually consists of partially fused spher-
ical-to— irregular-shaped masses intimately asso-
ciated with the vascular plexus on the ventral
surface of the kidney. This close relationship
with the kidney resembles the condition found
in the amphibians (Radu, 1931; Villee, 1943).

There are few anatomical descriptions of the
adrenals of the Chelonia in the literature.
Vincent (1896), Minervini (1906) and Thomp-
son (1932) give brief descriptions of the genus
Testudo. Ogushi (1911) describes their rela-
tionship to the kidneys in the soft-shelled turtle,
Trionyx japonicus. Kuntz (1912) describes the
embryonic development and gross anatomy of
the adrenals in young specimens of the logger-
head turtle, Caretta caretta. The gross anatom-
ical relationships in Emys europaea have been
reported by Spanner (1929). These early de-
scriptions have been reviewed by van der
Sprenkel (1934). In the marine tortoise, Caretta
caretta, the adrenals are usually united into a
single structure, according to Holmberg & Soler
(1942).

The adrenal glands in the Sauria are found

1955]

Hebard & Charipper: Reptilian Adrenal Gland

113

enclosed in the mesorchia or mesovaria sup-
porting the gonads. This consistent relationship
has been reported in the agamid lizard, Uromas-
tix hardwickii (Vincent, 1896), the Chamae-
leon sp. (Minervini, 1904), various iguanids
(Brooks, 1906; Reynolds, 1938), in the Lacer-
tidae (Braun, 1882; Pettit, 1896; Vincent, 1896;
Krause, 1922; Radu, 1934), the teid lizard,
Cnemidophorus gularis (Brooks, 1906), the
anguimorph lizards, Gerrhonotus (Retzlaff,
1949) , Heloderma (Hartman & Brownell, 1949)
and members of the Varanidae (Pettit, 1896).

The adrenals of the Serpentes bear the same
relationship to the reproductive organs as was
noted in the lizards; however, in this group the
elongation of the body has resulted in an asym-
metrical placement of the paired internal organs.
The right adrenal is always situated cephalad to
the left. This arrangement has been observed in
many colubrid snakes: Matrix natrix (Pettit,
1896; Spanner, 1929; Radu, 1934), Coluber sp.
(Vincent, 1896; Minervini, 1904; Uchoa
Junqueira, 1944), Dryophylox sp. (Valle &
Souza, 1942), the garter snake, Thammophis
sirtalis (Atwood, 1918) and the water snake,
Natrix sipedon (Bragdon, 1953).

The relative position of the adrenals in Tham-
nophis sirtalis can best be correlated with the
ventral scales, in which case a consistent rela-
tionship is noted. The left adrenal lies between
the 35th to 37th gastrostege in males and be-
tween the 31st to 34th in females. In both sexes,
the right adrenal lies between the 45th and 48th
scute, counting anteriorly from the anal plate.
This separation of approximately ten scutes has
been reported also for the common water snake,
Natrix sipedon (Bragdon, 1953).

While extensive data on adrenal weights
were not obtained, the average determined—
0.045 per cent, of the body weight— agrees with
the range of 0.04 to 0.05 per cent, reported for
other colubrid snakes by Naccarati (1922).

Bourne (1936, 1949) states that the adrenals
of the boa constrictor exhibit a primitive ana-
tomical condition in that they are located along
the lobes of the attenuated kidneys. This asso-
ciation with the kidneys, however, was not found
in the specimens of the boa constrictor and the
rubber boa, Charina bottae, that were examined
in this study. Both species have a typical serpen-
tine arrangement with the adrenals located
along the greatly elongated and asymmetrically-
placed gonads. This disposition of the adrenals
has previously been reported for the boa con-
strictor by Poll (1905) and the python (Pettit,
1896; Beddard, 1906a, 1906b; Spanner, 1929).
In the very young anaconda, Eunectes notaeus,
Beddard (1906a) noted that the adrenals are
associated with the persistent mesonephros but

that this condition is not found in adult animals.
This would suggest that Bourne probably based
his descriptions on very young animals that still
possessed mesonephros kidney tissue.

The vascular supply of the adrenals of reptiles
consists of both an arterial and venous supply.
In the Crocodilia, the arterial supply consists of
several pairs of arteries which arise from the
dorsal aorta and enter the dorso-mesial region of
the capsule. This general picture has been de-
scribed by Pettit (1896), Reese (1931) and
Lawton (1937). The arterial supply of the
adrenal in the Chelonia consists of several small
paired adrenal arteries from the aorta, as well
as small branches from the renal and genital
arteries. Very few observations on the arterial
supply have been reported in the literature.
Thompson (1932) described the arterial supply
to the adrenal in the tortoise, Testudo graeca.

The arterial supply to the adrenals in the
Sauria exhibits some variation among the vari-
ous species studied. In the horned lizard, Phry-
nosoma cornutum, each adrenal receives one
(sometimes two) arteries from the dorsal aorta.
These arteries also supply the gonads and per-
haps should be referred to as the spermatic or
ovarian arteries. The adrenals receive blood
from branches arising from these arteries, as the
former are in close contact with the adrenals. In
Anolis carolinensis, separate arteries supplying
the adrenal are present. Pettit (1896) reported
finding three arteries supplying the adrenal in
Varanus niloticus and only a single artery to
each gland in Varanus salvator. Bhatia (1929)
observed that branches of the spermatic arteries
supply the adrenals of the agamid lizard, Uro-
mastix hardwickii. According to Retzlaff ( 1949) ,
the arterial adrenal blood in the alligator lizard,
Gerrhonotus multicarinatus, is derived from
branches from the aorta, renal and mesenteric
arteries.

In the Serpentes, the adrenal arterial supply
comes from branches of the spermatic or
ovarian artery as it courses in an anterior-
posterior direction along the dorsal region of
the connective tissue capsule. This general pat-
tern has been reported for Natrix natrix
(O’Donoghue, 1912), for the rat snake, Ptyas
mucosus (Ray, 1934), and previously reported
for the garter snake, Thamnophis sirtalis, by
Atwood (1918).

A unique feature of the vascular supply of
the reptilian adrenal is the presence of an
adrenal portal system. This system of vessels has
been carefully studied by a number of morphol-
ogists and their observations have been reviewed
by van der Sprenkel ( 1934) . In the alligator and
caiman, a single adrenal portal vein arises from
the dorsal body wall anterior to the adrenal.

114

Zoologica: New York Zoological Society

[40; 10

enters the anterior pole of the adrenal, and
branches into a plexus connecting with the vas-
cular sinuses in the parenchyma of the gland.
This arrangement has been described by Pettit
(1896), Lawton (1937) and Spanner (1929)
and is confirmed in the histological preparations
of this study.

In the Chelonia, Spanner (1929) pictures
three adrenal portal veins branching off from the
renal portal veins. The presence of a portal
system in the turtles which were studied could
not be definitely determined by the histological
approach employed.

In the Sauria, a variable number of portal
veins branch off from the dorsal body wall and
enter the adrenals. Beddard (1904b, 1904c,
1905) and Spanner (1929) examined a number
of species of lizards and noted considerable
species variation in the number and relationship
of the adrenal portal veins. In the present study,
a single adrenal portal vein was observed en-
tering the anterior pole of the adrenal of Phry-
nosoma cornutum. Retzlaff (1949) does not
mention a portal supply for Gerrhonotus nor
does Miller (1952) for Xantusia.

In the Serpentes, the adrenal portal system is
well developed. For example, in the garter snake,
Thamnophis sirtalis, two relatively large veins
enter each adrenal from the dorsal body wall.
From one to three veins have been reported
entering the snake adrenal (O’Donaghue, 1912;
Ray, 1936; Beddard, 1904a, 1904d, 1906a).

According to Beddard (1906a), the adrenal
portal veins in reptiles are persistent parietal
branches of the postcardinal veins associated
with the embryonic mesonephros kidneys. Their
adrenal portal relationship is a secondary func-
tion taken over in the mature animal.

It is interesting to note that the adrenals of
reptiles are always located in the region where
the efferent renal veins join to form the inferior
vena cava. In the Chelonia, both adrenals are in
intimate contact with the efferent renal veins. In
the Crocodilia, the adrenals are associated with
the vena cava, the right adrenal being more in-
timately joined to the vena cava than the left.
In the Sauria and Serpentes, the left adrenal
usually drains either via short efferent veins or
directly into the anterior portion of the left ef-
ferent vein while the right adrenal usually drains
directly into the vena cava.

The nerve supply to the adrenal is derived
from lateral sympathetic ganglia, the extent of
innervation differing somewhat among the
various orders of reptiles. In the Crocodilia,
Lawton (1937) described the innervation of
the alligator adrenal as arising from four to five
ganglionic clusters. Fibers from these ganglia
enter the gland and associate themselves with

medullary tissue. This segmental pattern of in-
nervation was found in both the alligators and
caiman examined during this study.

The innervation of the chelonian adrenal is
more complex than that of the other reptiles.
Because of the close association of the gland
with the kidney, it is difficult to determine the
exact relationship of the numerous ganglia which
are located in or near the adrenal.

In the lizards and snakes, sympathetic ganglia
are found chiefly on the dorsal surface of the
gland in association with the arterial blood ves-
sels.

The exact nature of the numerous bundles of
nerve fibers seen coursing in the capsule and
between the parenchymal units is extremely
difficult to ascertain from histological prepara-
tions. Sudan-stained frozen sections reveal thin
sudanophilic sheaths in some of the nerve
bundles, which is indicative of the presence of
both myelinated preganglionic and unmye-
linated postganglionic fibers in the adrenals of
reptiles.

The medullary cells receive an abundant sup-
ply of unmyelinated fibers which end as dense
networks around the cells. Kolossow (1930) ob-
served that the nerve fibers terminate as oval
patches on the medullary cells of the turtle,
Emys europaea. In the alligator lizard, Gerrho-
notus multicarinatus, Retzlaff (1949) noted that
fine nerve fibers terminate on the medullary cells.
In the alligator, painted turtle and garter snake,
nerve fibers are found running along between
the cortical strands and often fine side-branches
are given off at right angles, which appear to
penetrate into the cortical tissue. The innerva-
tion of the cortical cells is a moot question. Some
investigators claim that nerve fibers terminate on
the cortical cells (Alpert, 1931, in the human;
and Retzlaff, 1949, in the alligator lizard). On
the other hand, Hollinshead (1936) and Swin-
yard (1937) found no evidence of cortical tis-
sue innervation in the cat adrenal. Also, Kolos-
sow (1930) was not able to discover nerve end-
ings to cortical cells of the turtle, Emys europaea.

The identification of argyrophilic nerve fibers
with specific cortical cells is a difficult problem
inasmuch as the fibers may not be destined for
the cortical cells, but rather for some other near-
by tissue such as blood vessels. Swinyard (1937)
cautions that the silver techniques employed
may possibly stain reticular fibers which may be
incorrectly identified as nerve fibers.

Histology

The reptilian adrenal gland is enclosed in a
thin connective tissue capsule continuous with
the enveloping mesentery or surrounding con-
nective tissue. The capsule is, in general, com-

1955]

Hebard & Charipper: Reptilian Adrenal Gland

115

posed of an outer layer of collagenous fibers con-
tinuous with the adjacent connective tissue and
an inner layer of reticular fibers continuous
with the reticular framework supporting the
glandular parenchyma. Elastic fibers are con-
fined to the media of blood vessels that are found
in the capsule and trabeculae. The relative
amount of collagenous and reticular fibers in the
gland varies with both the species and the age of
the animal. In young animals, the parenchyma is
supported by a thin network of reticular fibers
while in old adult animals the reticular frame-
work is supplemented by thick collagenous
fibers, the amount of which varies from species
to species. In adult turtles, there is considerable
infiltration of thick collagenous trabeculae from
the capsule, which tend to partition the cortical
tissue.

Some species variation is found in the lizards.
For example, Retzlaff (1949) observed in Ger-
rhonotus heavy bundles of collagenous fibers
surrounding the cortical cells, with fine fibrils
originating from these bundles passing among
the cells. This pattern is also found in the an-
guimorphid lizard, Ophisaurus, examined in this
study, but it is not present in the Scincidae. In
the skink, Eumeces fasciatus, the cortical cords
are enclosed in a network of collagenous fibers;
however, fibers do not appear to penetrate
among the individual cells comprising a cord.

An interesting and significant departure from
the hereinbefore described parenchymal stroma
is found in the garter snake. The cortical cords
are usually enclosed in a loose connective tissue
network composed chiefly of collagenous fibers.
This collagenous network is infiltrated with nu-
merous vascular sinuses supported by a fine re-
ticular network. This same arrangement has been
reported for the European snake. Coluber sp.,
by Minervini (1904).

The reptilian adrenal gland consists of three
kinds of cells— cortical, medullary, and sym-
pathetic nerve cells— arranged in definite patterns
which appear to be characteristic for the major
phylogenetic groups. In the Crocodilia, the
medullary tissue consists of an encircling periph-
eral zone in addition to small clumps, strands
and individual cells scattered in the interspaces
and sinuses between the columns of cortical
cells. This general crocodilian pattern is found
in the alligator and caiman. Reese (1931) and
Lawton (1937) have also described this arrange-
ment in the alligator; however, no descriptions
of Old World crocodiles were found in the
available literature.

A random intermixing of the medullary and
cortical components is typical of the Chelonia.
The medullary tissue is not concentrated in any
particular region of the gland but rather is ir-

regularly distributed among the cortical cell
groups as clusters and strands of medullary cells.
These clusters are of irregular size and shape,
tending to exhibit some species differences. For
example, the spotted turtle, Clemmys guttata, is
unique in having larger and more numerous
clumps of medullary cells than those observed
in other species of turtles studied. This general
chelonian arrangement has been described for
several species of turtles: Testudo tabulata and
Malaclemys terrapin, (Vincent, 1896), and
Emys europaea (Poll, 1906; Radu, 1934). More-
over, a similar random disposition of medullary
and cortical components is also present in the
avian adrenal (Hartman, et al., 1947).

In the Squamata (lizards and snakes), the
medullary tissue is aggravated principally on the
dorsal side of the cortical tissue with the degree
of intermingling with the cortical tissue varying
from species to species. In the horned lizard,
Phrynosoma cornutum, there is considerable in-
filtration, while in Anolis carolinensis little med-
ullary tissue is found in the cortical paren-
chyma. This scarcity is also true of the skinks,
Eumeces obsoletus, E. fasciatus and Lygosoma
laterale. Among the anguimorphid lizards, Retz-
laff (1949) reported that the medullary cells of
the alligator lizard, Gerrhonotus multicarinatus,
are intermingled with the cortical portion. This
was also noted in this study for Gerrhonotus
coeruleus as well as for the related species,
Ophisaurus ventralis. In addition, Hartman &
Brownell (1949) noted some intermingling in
the Gila monster, Heloderma suspectum.

In theSerpentes.the medullary cells are chiefly
aggregated in a compact unit associated closely
with arterial blood vessels located on the dorsal
side of the capsule. A more or less uniform scat-
tering of small medullary clusters of one to four
cells is found in the cortical parenchyma. This
arrangement is observed in the garter snake,
water snake and other colubrid snakes examined,
as well as in the boa constrictor.

The cortical tissue exhibits considerable var-
iation in shape and arrangement of the cells,
with an appearance of a more or less character-
istic pattern for each species. In the Crocodilia,
namely, the alligator and caiman, the cortical
tissue consists of anastomosing cords composed
of polyhedral-to-tall columnar cells having a
more or less regular arrangement. Probably be-
cause of compression, the outer cords in the
adrenal often appear to have a more regular
cellular arrangement. In this instance, the periph-
eral cords may consist of a single row of
greatly elongated and flattened cells, as noted
for the alligator, or as a double row of cells, in
the case of the caiman.

In the Chelonia, the cords usually consist of

116

Zoologica: New York Zoological Society

[40: 10

irregularly-shaped cells which, in longitudinal
section, consist of several rows of cells having
very indistinct cell outlines and an abundant
and irregular distribution of nuclei. The more
centrally located columns usually consist of more
than two tiers of cells; on the other hand, periph-
eral cords (as a result of compression) often
consist of a single row of greatly elongated and
flattened cells. In the family Kinosternidae, the
cells tend to have a more regular arrangement,
with the cells appearing to be radially disposed
in the cords.

The disposition of the cortical cells in the
Sauria is of particular interest for the reason that
certain phylogenetic trends are apparent. On
arranging the various species of lizards (de-
scribed in this study and elsewhere in the litera-
ture) into groups on the basis of the disposition
and morphology of the cortical cells, they fall
into three general categories having similar his-
tological features: the gekkonid, anguimorphid
and iguanid types.

These histological trends appear to substan-
tiate the suggested phylogenetic relationships of
the various lizard groups. According to Camp
(1923) , the Sauria constitute a natural and di- j
versified group arising from some unknown
Permian stem reptile which soon divided into I
two main groups, the Ascalobota and the Autar- |
choglossa. The former includes the Gekkonidae,
Agamidae, Chamaeleontidae and Iguanidae, '

with the Gekkonidae considered to be the most
primitive family. The Autarchoglossa are sub- j
divided into the Scincomorpha, which includes
the Xantusidae, Scincidae, Lacertidae and
Teidae, and the Anguimorpha, which includes
the Anguidae and the platynotid lizards, of
which the Varanidae are a part. The Xantusidae
are believed to be an intermediate group be-
tween the Ascalobota and Autarchoglossa, These
suggested relationships are shown in Text-fig. 1,
modified after Camp (1923). Only those fam-
ilies are included for which histological data are
available for comparison.

IGUANID GEKKONID ANGUIMORPHID

ADRENAL TYPE ADRENAL TYPE ADRENAL TYPE

Text-fig. 1. Comparison
of adrenal histological
types with suggested phy-
logenetic relationships of
the lizards. (Phylogenetic
chart modified from
Camp, 1923).

1 r

UJ

<

o

Q

O

CL

o

o

>-

UJ

<

o

z

o

ic:

UJ

UJ

<

o

CL O

UJ

UJ

<

Q

Ui

UJ

<

s

UJ

<

<

<

<

01

Q

o

to

z>

t-

Q

O

o

3

UJ

o

o

z

<

<

z

z

o

_l

01

OQ

<

o

z

UI

<

O

X

to

<

X

>

UJ

t-

z

s

(O

GEKKOTA SCINCOMORPHA ANGUIMORPHA

PERMIAN ANCESTOR

1955]

Hebard & Charipper: Reptilian Adrenal Gland

117

The characteristic histological feature of the
gekkonid type of adrenal is the radial arrange-
ment of tall wedge-shaped columnar cells into
cords having a consistent disposition of the
nuclei at the periphery of the cords. This ar-
rangement is found in the four species of the
Gekkonidae described in this study as well as in
the three species belonging to the family Scin-
cidae. The gekkonid pattern is also observed in
the Dibamidae and Pygopodidae, and has been
described in the Xantusidae by Miller (1952).

The anguimorphid pattern consists of an ir-
regular arrangement of polyhedral low columnar
cells with no general tendency to form cords
consisting of regularly arranged tiers of cells.
This arrangement is found in the glass snake,
Ophisaurus ventralis, and the closely related
species, Gerrhonotus coeruleus. Retzlaff (1949)
has also described this irregular disposition of
cortical cells in the latter genus. Hartman &
Brownell (1949) found a similar pattern to
exist in another anguimorphid lizard, the Gila
monster, Heloderma suspectum.

The disposition and morphology of the
adrenal cortical cells in the Iguania is somewhat
variable, even within family groups; however,
all the species that have been described appear
to have a more or less regular arrangement of
the cells into cords. In the Agamidae, Vincent
(1896) describes the cortical tissue of
Uromastix hardwickii as comprised of cords of
cells consisting of a double row of tall columnar
cells each having a centrally placed nucleus.
Minervini (1904) pictures the cortical cords of
the Chameleon sp. as composed of small poly-
hedral cells with centrally located nuclei. In the
Iguanidae, the cortical cells of Anolis caro-
linensis and Sceloporus undulatus would fit more
or less the above description. On the other hand,
cortical cells of the horned lizard, Phrynosoma
cornutum, are more inclined to be of a tall
columnar type with the nuclei tending toward a
peripheral position.

While certain anatomical similarities suggest
that the Lacertidae and Teidae are closely re-
lated to the Scincidae (Camp, 1923), the his-
tological features of the adrenal gland of these
two families resemble more closely the iguanid
type rather than the gekkonid pattern.

According to Camp (1923), the Gekkonidae
represent the most primitive group with the
Xantusidae being intermediate between the Gek-
konidae and the Scincidae. The Lacertidae are
considered an offshoot from the Scincidae and
the Teidae as a branch of the Lacertidae. The
Anguimorpha are believed to have separated
from the Scincomorpha very early in the
Mesozoic era.

If, as Camp points out, the Gekkonidae re-

present the most primitive group, then it would
seem justifiable to consider that the gekkonid
type of adrenal found in the Gekkonidae rep-
resents a primitive morphological condition.
This type is also present in the following groups:
Xantusidae, Scincidae, Dibamidae and Pygo-
podidae. The Iguania, on the other hand, have
diverged from the gekkonid pattern, showing
only an occasional similarity as, for example,
in Phrynosoma cornutum. Furthermore, the
Anguimorpha (Anguidae and Helodermatidae)
show little similarity to the gekkonid or iguanid
types and have in their long geologic separation
either undergone an evolutionary modification
from the primitive gekkonid pattern or else may
have been derived from primitive Ascalobota
differing entirely from the primitive Gekkonidae.

The cortical cells of the Serpentes are ar-
ranged into cords having a characteristic pattern.
In the garter snake, Thamnophis sirtalis, the cells
are radially arranged into cylindrical cords which
exhibit a striking regularity and polarity. In longi-
tudinal sections of a cord, the cells are disposed
in two rows of tall columnar cells having a
double row of nuclei aligned lengthwise near
the center of the cord. In transverse section, a
cord consists of a radially arranged group of
tall, wedge-shaped cells having a centrally lo-
cated ring of nuclei. This consistent pattern is
also found in the water snake, Matrix sipedon,
and the rattlesnake, Crotalus viridis. In the
rubber boa, Charina bottae, the cords appear to
be made up of several tiers of cells having a
more irregular arrangement.

Cytology

(a) Medullary cells

Two types of medullary cells can be distin-
guished in all of the species of reptiles studied
on the basis of their tinctorial characteristics:
a “light” type of cell usually light yellow in color
and having fine granules; and a yellowish-brown
“dark” type having coarse granules. The degree
of staining intensity and the relative number of
each type varies from order to order and from
species to species. The two cell types are best
distinguished after fixation in solutions contain-
ing chromium salts, such as Zenker’s, followed
by stains containing acid fuchsin (for example,
Masson’s). Classical literature often refers to the
“chromaffin reaction” as a selective staining of
the cytoplasmic granules by the chrome salts;
however. Coupland (1953) points out that the
yellow-to-brown color is the result of the oxida-
tion of a certain amount of catechol amines into
corresponding quinones and other complex oxi-
dation products, such as pigments.

The occurence of different medullary cell
types has been reported in mammals (Bander,

118

Zoologica: New York Zoological Society

[40: 10

1951; Bourne, 1949); birds (Muller, 1929);
reptiles (Radu, 1934; Reese, 1931; Retzlaff,
1949; Miller, 1952) ; and amphibians (Fustinoni
& Porto, 1938).

(b) Cortical cells

The most characteristic feature of the cortical
cells is the presence of lipid droplets and gran-
ular mitochondria, there being an inverse re-
lationship between the amount of each of these
cytoplasmic inclusions present. Histochemical
studies indicate that the droplets are a mixture
of neutral fats, cholesterol and cholesterol esters,
and carbonyl-containing lipids. This description
portrays, in general, the fundamental functional
unit of the cortical tissue found throughout the
various vertebrate classes: shark (Dittus, 1941;
Hayes, 1941), lungfish (Holmes, 1950; Gerard,
1951), frog (Radu, 1931), bird (Muller, 1929;
Miller & Riddle, 1942; Knouff & Hartman,
1951) and mammal (Creep & Deane, 1949;
Nicander, 1952).

The number of lipid droplets may vary
slightly from cell to cell; however, there is no
tendency toward a zonation similar to that found
in the mammals. On the other hand, the dis-
tribution of the lipid droplets within the cells
varies significantly among the different species
of reptiles that were studied, and it is possible
to distinguish at least three basically different
cytological cell types.

The cortical cells of the garter snake are of
special interest for they exhibit a polarization
in which the lipid droplets are localized at the
free end of the cell in closest contact with the
blood stream, while the nucleus is placed near
the base of the cell. The granular mitochondria
are scattered in the interstices between the lipid
droplets. This unique disposition of the lipid
droplets, mitochondria and nucleus suggests that
an exchange of materials occurs at the periph-
eral surface adjacent to the blood supply while
presecretory synthetic processes occur in the re-
gion between the nucleus and the free end of
the cell. A similar type of functional polarity
has been described for the brown pelican
(Knouff & Hartman, 1951).

A functional polarity the reverse of that just
mentioned is found in certain species of lizards.
In the skink, Eumeces ohsoletus, the lipid drop-
lets and mitochondria are concentrated near the
base of the cells with the nucleus located at the
peripheral edge next to the blood stream. This
interesting cytological arrangement is also found
in other species of skinks, geckos, xantusids,
dibamids and pygopodids. A satisfactory ex-
planation for this unique functional adaptation
cannot be derived from histological studies alone.
Miller (1952) found that the lipid content of the

cortical cells of Xantusia can be experimentally
altered; however, the mechanism as to how
secretory products reach the blood stream still
remains unknown.

The third functional type has the nucleus
centrally located and exhibits an even distribu-
tion of lipid droplets and mitochondria through-
out the cytoplasm. This disposition is found in
the cortical cells of the alligator, painted turtle
and green anolis.

Summary

1. The adrenal glands of reptiles exhibit a
consistent association with the reproductive
organs.

2. The vascular supply consists of both arte-
rial and venous blood with the origin and num-
ber of vessels supplying the adrenals varying
from species to species. A close association with
the efferent renal veins and vena cava is common
to all species.

3. The nerve supply to the adrenal is derived
from lateral sympathetic ganglia with the extent
of innervation differing somewhat among the
various orders of reptiles. The medullary com-
ponent of the gland is innervated by networks
of unmyelinated fibers; however, only a few
fibers are found in the cortical tissue.

4. The reptilian adrenal gland is enclosed in
a thin connective tissue capsule composed of an
outer layer of collagenous fibers continuous
with the adjacent connective tissue and mesen-
teries, and an inner layer of reticular fibers con-
tinuous with the reticular framework support-
ing the parenchyma of the gland. The relative
amount of collagenous and reticular fibers in
the gland varies with both the species and the
age of the animal.

5. The reptilian adrenal gland consists of
three kinds of cells— cortical, medullary and sym-
pathetic nerve cells— arranged in definite pat-
terns which appear to be characteristic of the
major phylogenetic groups.

6. In the Crocodilia, the medullary tissue con-
sists of an encircling peripheral zone in addition
to small clumps, strands and individual cells
scattered in the cortical parenchyma. A random
intermixing of the medullary and cortical com-
ponents is typical of the Chelonia. In the Sauria
and Serpentes, the medullary tissue is aggre-
gated principally on the dorsal side of the cortical
tissue, with only a slight intermixing.

7. The cortical tissue exhibits considerable
variation in shape and arrangement of the cells
with the appearance of a more or less character-
istic pattern for each species.

8. The disposition of the cortical cells in the
Sauria is of particular interest for the reason
that certain phylogenetic trends are apparent.

1955]

Hebard & Charipper: Reptilian Adrenal Gland

119

The Sauria can be divided into three general
groups: the gekkonid, anguimorphid and

iguanid types, on the basis of similar histo-
logical features.

9. Two types of medullary cells can be dis-
tinguished in all of the species of reptiles studied
on the basis of their tinctorial characteristics.

10. The cortical cells are characterized by the
presence of lipid droplets and granular mito-
chondria, there being an inverse ratio between
the amount of each of these present. Histo-
chemical studies indicate that the droplets are
a mixture of neutral fats, cholesterol esters and
carbonyl-containing lipids. The number of lipid
droplets varies from cell to cell; however, there
is no tendency toward a zonation.

11. Polarity differences in the distribution
of the lipid droplets, mitochondria and nuclei
in the cortical cells of the various species studied
illustrate fundamental variations in the cytology
of cortical cells.

Literature Cited

Alpert, L.

1931. The innervation of the suprarenal gland.
Anat. Rec., 50: 221-234.

Atwood, W. H.

1918. The visceral anatomy of the garter snake.
Trans. Wise. Acad. Sci., 19: 531-552.

Baker, J. R.

1944. The structure and chemical composition of
the Golgi element. Quart. J. Micro. Sci.,
85:1-71.

1946. The histochemical recognition of lipine.
Quart. J. Micro. Sci., 87: 441-470.

Bander, A.

1951. fiber zwei verschiedene chromaffine Zell-
typen im Nebennierenmark und ihre Be-
ziehung zum Adrenalin und Arterenolge-
halt. Anat. Anz., 97 (Suppl.): 172-176.

Beddard, F. E.

1904a. Contributions to our knowledge of the cir-
culatory system in the Ophidia. Proc. Zool.
Soc. London, 1: 331-370.

1904b. Contributions to the anatomy of the La-
certilia. ( 1 ) Qn the venous system in cer-
tain lizards. Proc. Zool. Soc. London, 1:
434-450.

1904c. Contributions to the anatomy of the La-
certilia. (3) On some points in the vascular
system of Chamaeleon and other lizards.
Proc. Zool. Soc. London, 2: 6-22.

1904d. Notes upon the anatomy of certain snakes
of the family Boidae. Proc. Zool. Soc.
London, 2: 107-121.

1905. Some additions to the knowledge of the
anatomy, principally of the vascular system
of Hatteria, Crocodilus and certain Lacer-
tilia. Proc. Zool. Soc. London, 2: 461-489.

1906a. Contributions to the anatomy of the Ophid-
ia. I. Notes on the vascular system of the
Anaconda, Eunectes notaeus. Proc. Zool.
Soc. London, 1: 12-44.

1906b. Contributions to the knowledge of the vas-
cular and respiratory systems in the Ophid-
ia and to the anatomy of the genera Boa
and Corallus. Proc. Zool. Soc. London, 2:
499-532.

Bhatia, M. L.

1929. On the arterial system of the lizard, Uro-
mastix hardwickii. J. Morph, and Physiol.,
48: 283-316.

Bimmer, Edith

1950. Metrische Untersuchungen iiber die Ent-
wicklung der Nebenniere und der ihr be-
nachbarten Organe bei Eidechsen. Anat.
Anz., 97: 276-311.

Bourne, G.

1936. The phylogeny of the adrenal gland. The
Amer. Nat., 70: 159-178.

1949. The Mammalian Adrenal Gland. Oxford
University Press, London.

Bragdon, D. E.

1953. A contribution to the surgical anatomy
of the water snake, Natrix s. sipedon;
the location of the visceral endocrine
organs with reference to ventral scutel-
lation. Anat. Rec., 117: 145-161.

Braun, H.

1882. Bau und Entwicklung der Nebennieren
bei Reptilien. Arb. a. d. Zool. Zoot. Inst.
Wurzburg, 5: 1-30.

Brooks, B.

1906. The anatomy and internal urogenital or-
gans of certain American lizards. Trans.
Tex. Acad. Sci., 8: 23-38.

Camp, C. L.

1923. Classification of lizards. Bull. Amer. Mus.
Nat. Hist., 48: 289-481.

Coupland, R. E.

1953. On the morphology and adrenaline-norad-
renaline content of chromaffin tissue.
J. Endocrinology, 9: 194-203.

Dittus, P.

1941. Histologic und Cytologic des Interrenal-
organs der Selachier unter normalen und
experimentellen Bedingungen. Zeit. f. wiss.
Zool. Abt. A., 154: 40-124.

Forbes.T. R.

1940. Studies on the reproductive system of the
alligator. IV. Observations on the devel-
opment of the gonad, the adrenal cortex,
and the Mullerian duct. Contrib. to
Embryol. No. 174. Carnegie Inst, of
Washington, XXVIII: 131-157.

120

Zoologica: New York Zoological Society

[40: 10

Fustinoni, O., & J. Porto.

1938. Morfologia de las glandulas adrenales del
sapo, Bufo arenarum (Hensel). Rev. Soc.
Argent, de Biol., 14: 315-320.

Gerard, P.

1951. Sur la cortico-surrenale du Protopt^re
(Protopterus dolloi Blgr.). Arch. Biol.
(Paris), 62: Zll-311.

Creep, H. O., & Helen W. Deane.

1949. The cytology and cytochemistry of the
adrenal cortex. Ann. N. Y. Acad. Sci.,
50: 596-615.

Hartman, F. A., & K. A. Brownell.

1949. The Adrenal Gland. Lea and Febiger,
Philadelphia.

Hartman, F. A., R. A. Knouff, A. W. McNutt

& J. E. Carver.

1947. Chromaffin patterns in bird adrenals.
Anat. Rec., 97: 211-222.

Hayes, E. R.

1941. The elasmobranch interrenal; a preliminary
note. The interrenal body of Alopias vul-
pinus (Bonnaterre). Biol. Bull., 81: 299.

Hollinshead, W. H.

1936. The innervation of the adrenal gland.
Jour. Comp. Neur., 64: 449-467.

Holmberg, a. D., & F. L. Soler.

1942. Some notes on the adrenals. Presence of a
united adrenal in the marine tortoise.
Contributions from the laboratory of
anatomy, comparative physiology and
pharmacodynamics. Univ. Buenos Aires,
20: 457-469, 667-675.

Holmes, W.

1950. The adrenal homologues in the lungfish,
Protopterus. Proc. R. Soc. London, Ser.
B., 137: 549-562.

Knouff, R. A. and F. A. Hartman.

1951. A microscopic study of the adrenal of the
brown pelican. Anat. Rec., 109: 161-188.

Kolossow, N. G.

1930. Zur Frage der Innervation der Neben-
nieren. Zeit. f. Micros. Anat. Forsch.,
20: 107-121.

Krause, R.

1922. Microskopische Anatomic der Wirbeltiere.
II. Vogel und Reptilien. Berlin und
Leipzig.

Kuntz, a.

1912. The development of the adrenals in the
Loggerhead Turtle, Thalassochelys ca-
retta. Am. J. Anat., 13: 71-89.

Lawton, F. E.

1937. The adrenal-autonomic complex in Alli-
gator mississippiensis. J. Morph., 60: 361-
370,

Lillie, R. D.

1948. Histopathologic Technic. The Blakiston
Company, Philadelphia.

McClung, R.

1952. McClung’s Handbook of Microscopical
Technique, 3rd Edition. Paul B. Hoeber,
Inc., New York.

Miller, M. R.

1952. The normal histology and experimental
alteration of the adrenal of the viviparous
lizard, Xantusia vigilis. Anat. Rec., 113:
309-323.

Mn-LER, R. A., & O. Riddle.

1942. The cytology of the adrenal cortex of
normal pigeons and experimentally in-
duced atrophy and hypertrophy. Am. J.
Anat., 71: 311-341.

Mi-nervini, R.

1904. Des capsules surrenales. Development-
structure-fonctions. Jour, de I’anat. et de
la physiol., 40: 449-492, 634-667.

Muller, J.

1929. Die Nebennieren von Gallus domesticus
und Columba livia domestica. Zeit. f.
Micros. Anat. Forsch., 17: 313-352.

Naccarati, S.

1922. On the relation between the weight of the
internal secretory glands and the body
weight and brain weight. Anat. Rec., 24:
255-260.

Nicander, L.

1952. Histological and histochemical studies on
the adrenal cortex of domestic and labo-
ratory animals. Acta Anat., 14 (Suppl.
16): 1-88.

Ogushi, K.

1911. Ueber die Nebennieren und Nierenpfor-
tader des Trionyx japonicus. Anat. Anz.,
39: 183-190.

O’Donoghue, C. H.

1912. The circulatory system of the common
grass-snake, Tropidonotus natrix. Proc.
Zool. Soc. London, 2: 612-647.

Pearse, a. G. E.

1952. Histochemistry, Theoretical and Applied.
Little, Brown and Company, Boston.

Pearson, A. A., & Sarah L. O’Neill.

1946. A silver-gelatin method for staining nerve
fibers. Anat. Rec., 95: 297-302.

Pettit, A.

1896. Recherches sur les capsules surrenales.
Jour, de I’anat. et de la physiol., Paris, 32:
301-362, 369-419.

1955]

Hebard & Cliaripper: Reptilian Adrenal Gland

121

Poll, H.

1905. Die Anlage der Zwischenniere bei der
europaischen Sumpfschildkrote (Ernys
eiiropaea).lnsi. Monatschr. Anat. Physiol.,
21: 195-291.

1 906. Die vergleichende Entwickelungsgeschich-
te der Nebennieren-systeme der Wirbel-
tiere. In Hertwig’s Handbuch d. vergl. u.
exper. Entwg. d. Wirbeltiere, 3: 443-616.

Radu, V.

1931. fitude cytologique de la glande surrenale
des Amphibiens, anoures. Bull. d’Hist.
appliq., 8: 249-264.

1934. Les glandes surrenales des reptiles (note
preliminaire). Annales Scientifiques de
rUniversit6 de Jassy, 19: 378-381.

Ray, H. C.

1934. On the arterial system of the common
Indian Rat-Snake, Ptyas mucosus (Linn.).
J. Morph., 56: 533-575.

1936. On the venous system of the common
Indian Rat-Snake, Ptyas mucosus (Linn.).
J. Morph., 59: 517-547.

Reese, A. M.

1931. The ductless glands of Alligator mississip-
piensis. Smithsonian Misc. Coll., 82: 1-14.

Retzlaff, E. W.

1949. The histology of the adrenal gland of the
alligator lizard, Gerrhonotus multicari-
natus. Anat. Rec., 105: 19-34.

Reynolds, A. E.

1938. Some gross anatomical relations of the
male urogenital system and other internal
organs in Eumeces fasciatus. Proc. Indiana
Acad. Sci., 49: 233-242.

Sayers, G.

1950, The adrenal cortex and homeostasis.
Physiol. Rev., 30; 241-320.

Schmidt, K. P.

1953. A Check List of North American Am-
phibians and Reptiles. The Univ. of Chi-
cago Press, Chicago.

Spanner, R.

1929. liber die Wurzelgebiete der Nieren, Ne-
bennieren und Leberpfortader bei Rep-
tilien. Morph. Jarb., 63: 314-358.

SWINYARD, C. A.

1937. The innervation of the suprarenal glands.
Anat. Rec., 68: 417-427.

Thompson, J.

1932. The anatomy of the tortoise. Sci. Proc.
Royal Dublin Soc., 20: 359-461.

UCHOA JUNQUEIRA, L. C.

1944. Nota sobre i morfologia das adrenals dos
ofidios. Rev. Brasil. Biol., 4: 63-67.

Valle, J. R., & P. R. Souza.

1942. Observations upon the endocrine system
of the Ophidia. Rev. Brasil. Biol., 2: 81-88.

VAN DER SpRENKEL, H. S.

1934. II. Nebenniere und Paraganglien. In Bolk’s
Handbuch der vergleichende Anatomie
der Wirbeltiere. Band 2: 791-794. Urban
u. Schwarzenberg, Berlin.

ViLLEE, C. A., Jr.

1943. The effect of adrenocorticotrophic hor-
mone on the interrenals (cortical tissue)
of Triturus torosiis. J. Elisha Mitchell
Sci. Soc., 59: 23-26.

Vincent, S.

1896. The suprarenal capsules in the lower ver-
tebrates. Proc. Birm. Nat. Hist, and Phil.
Soc., 10: 1-26.

1898. The comparative histology of the supra-
renal capsules. Intern. Monat. f. Anat. u.
Physiol., 15: 282-324.

Zwemer, R. L.

1933. A method for studying adrenal and other
lipoids by a modified gelatin embedding
and mounting technique. Anat. Rec., 57:
41-44.

EXPLANATION OF THE PLATES
Plate I

Fig. 1. Transverse section through alligator ad-
renal (a) showing relationship to kidneys
(k), gonads (g), dorsal aorta (d), vena
cava (v) and sympathetic ganglia (s).
Bouin’s fixation and Masson’s stain. X24.

Fig. 2. Transverse section through adrenal of
alligator showing sympathetic nerve fibers
in capsule, peripheral zone of “light” med-
ullary cells, strands of “dark” medullary
cells extending into cortical tissue and
irregular shape and arrangement of light-
staining cortical cells. Zenker-formal fixa-
tion and Masson’s stain. X130.

Fig. 3. Transverse section through adrenal of alli-
gator showing distribution of reticular
network supporting parenchyma of gland.
Note heavy reticular fibers running length-
wise with cords and fine network of reticu-
lar fibers that encircles the cortical cords.
Bouin’s fixation and Foot’s modification of
Hortega’s silver carbonate technique.
X240.

Fig. 4. Frozen section of alligator adrenal show-
ing distribution of sudanophilic lipids in
the cortical cells. Note absence of stain
in medullary cells. Baker’s formalin fixa-
tion and Sudan black B stain. X240.

Fig. 5. Transverse section of alligator adrenal
showing centrally located nuclei and the
uniform distribution of granular mito-
chondria in tJie lipid-rich cortical cells.

122

Zoologica: New York Zoological Society

[40: 10

Note cluster of “dark” medullary cells at
upper left and strand of “light” medullary
cells at right-hand edge. Zenker-formal
fixation, post-chromated for 48 hours and
stained with aniline-acid fuchsin and
methyl blue. X690.

Plate II

Fig. 6. Transverse section through adrenal (a) of
painted turtle, Chrysemys picta, showing
relationship to kidneys (k), gonads (g),
dorsal aorta (d) and vena cava (v).
Zenker-formal fixation and Masson’s stain.
X24.

Fig. 7. Transverse section through adrenal of
painted turtle showing disposition of med-
ullary and cortical tissue. Note irregu-
lar arrangement of cortical cells into anas-
tomosing cords. Zenker-formal fixation
and Mayer’s acid hemalum and eosin.
X230.

Fig. 8. Transverse section through adrenal of
painted turtle showing close association of
medullary cells and sympathetic ganglion
cells. Zenker-formal fixation and Mayer’s
acid hemalum and eosin. X230.

Fig. 9. Frozen section of adrenal of painted turtle
showing distribution of sudanophilic lip-
ids in cortical tissue. Note single sudano-
philic nerve fiber in sympathetic nerve at
upper edge of photograph. Baker’s forma-
lin fixation and Sudan black B stain. X240.

Fig. 10. Transverse section of adrenal of painted
turtle showing distribution of granular
mitochrondria in lipid-rich cortical cells.
Note centrally located nuclei in cortical
cells and clump of “dark” medullary cells
at upper edge of photograph. Zenker-
formal fixation, post-chromated for 48
hours and stained with aniline-acid fuch-
sin and methyl blue. X690.

Plate III

Fig. 11. Oblique section through adrenal (a) of
horned lizard, Phrynosoma cornutum,
showing association with vas deferens
(v.d.). Note peripheral zone of darker-
staining medullary tissue as well as scat-
tered masses among the cortical cells.
Zenker-formal fixation and Masson’s stain.
X24.

Fig. 12. Transverse section of adrenal of homed
lizard showing sympathetic ganglion cells
near artery. Note peripheral zone of
“light” medullary cells and clumps of
“dark” medullary cells lying between the
cortical tissue cords. Note regular arrange-
ment of tall columnar cortical cells. Zen-
ker-formal fixation and Masson’s stain.
X230.

Fig. 13. Frozen section of adrenal of homed lizard
showing intense sudanophilia of cortical
cells. Note negative-staining sympathetic
ganglion. Baker’s formalin fixation and
Sudan black B stain. X240.

Fig. 14. Transverse section of adrenal of lizard,
Anolis carolinensis, showing dorsal sym-
pathetic ganglion and medullary tissue.
Note elose association of cortical tissue
with vena cava and cord-like arrangement
of cortical cells. Zenker-formal fixation
and Mayer’s acid hemalum and eosin.
X230.

Fig. 15. Transverse section through adrenal of al-
ligator lizard, Gerrhonotus coeruleus,
showing pigmented capsule, peripheral
zone of medullary cells, peripheral zone
of “light-staining” cortical cells and irreg-
ular arrangement of cortical cells in the
parenchyma of the gland. Bouin’s fixation
and Masson’s stain. X230.

Fig. 16. Transverse section through adrenal of the
glass snake, Ophisaurus ventralis, showing
aggregation of medullary cells, large sinus
opening into ventral vein and the irregular
arrangement of the cortical cells. Bouin’s
fixation and Mayer’s acid hemalum and
eosin. X230.

Fig. 17. Transverse section through adrenal of the
lizard, Cnemidophorus tessellatus, show-
ing dorsal aggregation of medullary cells
and irregularly arranged cortical cells.
Bouin’s fixation and Mayer’s acid hema-
lum and eosin. X230.

Plate IV

Fig. 18. Transverse section through adrenal of the
skink, Eumeces obsoletus, showing typi-
cal radial arrangement of the tall col-
umnar cortical cells into anastomosing
cords. Note disposition of nuclei at the
peripheral edge of the cords. Zenker-for-
mal fixation and Masson’s stain. X230.

Fig. 19. Transverse section of adrenal of skink,
Eumeces fasciatus, showing peripheral
position of nuclei and scarcity of granular
mitochondria in lipid-rich cytoplasm of
cortical cells. Zenker-formal fixation, post-
chromated for 48 hours and stained with
aniline-acid fuchsin and methyl blue.
X690.

Fig. 20. Frozen section of adrenal of skink, Eu-
meces obsoletus, showing characteristic
localization of sudanophilic lipids in cen-
tral region of cortical cords and clear
peripheral nuclear zone. Baker’s formalin
fixation and Sudan black B stain. X240.

Fig. 21. Transverse section through the adrenal of
the gecko, Hemidactlyus turcicus, show-
ing dorsal medullary tissue and cords of
radially arranged tall columnar cortical
cells with peripherally located nuclei.
Zenker-formal fixation and Masson’s stain.
X230.

Fig. 22. Transverse section through adrenal of the
gecko, Tarentola mauretanica, showing
similar disposition of cortical cells as noted
for 'Hemidactylus and Eumeces. Zenker-
formal fixation and Masson’s stain. X230.

1955J

He hard & Charipper: Reptilian Adrenal Gland

123

Fig. 23. Transverse section through adrenal of
skink, Lygosoma laterale, showing dorsal
medullary tissue and cords of radially ar-
ranged tall columnar cortical cells with
peripherally located nuclei. Zenker-formal
fixation and Masson’s stain. X230.

Plate V

Fig. 24. Transverse section through the adrenal
(a) of the garter snake, Thamnophis sir-
talis, showing relationship to ovary (o)
and mesovarium (mes.). Note ovarian
artery (o.a.) in dorsal capsular region
and intimate association of adrenal with
the ventrally located vena cava (v).
Bouin’s fixation and Masson’s stain. X28.

Fig. 25. Transverse section through the adrenal of
the garter snake showing large capsular
nodule of undifferentiated cortical cells,
dorsal aggregation of medullary cells at
right of nodule and cords of radially
arranged tall columnar cortical cells with
nuclei arranged in regular rows at center
of cord. Note small groups of dark-staln-
ing medullary cells dispersed between the

cortical cords and the extreme vascularity
of the gland. Bouin’s fixation and Mas-
son’s stain. X 120.

Fig. 26. Frozen section of garter snake adrenal
showing characteristic localization of sud-
anophilic lipids at periphery of cortical
cords and clear nuclear zone in center of
cords. Baker’s formalin fixation and Sudan
black B stain. X240.

Fig. 27. Frozen section of garter snake adrenal
showing distribution of Shultz-positive
substances. Note slight staining of undif-
ferentiated cortical cells in capsular con-
nective tissue. Baker’s formalin fixation
and Shultz cholesterol technique. X240.

Fig. 28. Transverse section of garter snake adrenal
showing characteristic arrangement of tall
columnar cortical cells. Note centrally lo-
cated nuclear zone and uniform distribu-
tion of granular mitochondria in lipid-
rich cytoplasm. Zenker-formal fixation,
post chromated for 48 hours and stained
with aniline-acid fuchsin and methyl blue.
X690.

/

(

HEBARD a CHARIPPER

PLATE 1

FIG. 2

FIG. 5

FIG. 4

A COMPARATIVE STUDY OF THE MORPHOLOGY AND HISTOCHEMISTRY
OF THE REPTILIAN ADRENAL GLAND

HEBARD a CHARIPPER

PLATE II

FIG. 9

FIG. 10

A COMPARATIVE STUDY OF THE MORPHOLOGY AND HISTOCHEMISTRY
OF THE REPTILIAN ADRENAL GLAND

HEBARD & CHARIPPER

PLATE III

FIG. 11

FIG. 12

FIG. 13

FIG. 15

FIG. 14

A COMPARATIVE STUDY OF THE MORPHOLOGY AND HISTOCHEMISTRY
OF THE REPTILIAN ADRENAL GLAND

HEBARD a CHARIPPER

PLATE IV

FIG. 20

FIG. 21

FIG. 22

A COMPARATIVE STUDY OF THE MORPHOLOGY AND HISTOCHEMISTRY
OF THE REPTILIAN ADRENAL GLAND

HEBARD a CHARIPPER

PLATE V

FIG. 25

FIG. 24

FIG. 26

A COMPARATIVE STUDY OF THE MORPHOLOGY AND HISTOCHEMISTRY
OF THE REPTILIAN ADRENAL GLAND

11

Hormonal Control of the Sexually Dimorphic
Pigmentation of Thalassoma bijasciatutn

Louise M. Stoll

The American Museum of Natural History, New York 24, N. Y.

(Plates I-ni)

The bluehead, Thalassoma bifasciatum
(Bloch), shows the blue-headed condi-
tion only in the adult males while imma-
ture fish and females show quite other colors
and patterns. The fact that the males are sexu-
ally mature some time before they develop their
strikingly different secondary sex characters has
led to a considerable taxonomic confusion. The
females and young have gone under the name of
T. mWa (Gunther) and T. nitidissima (Goode) ,
depending on the color phase they showed. The
males have been referred to as T. subfurcatus
Nichols. This situation was recognized by
Longley (1914, 1915) but not generally ac-
cepted by ichthyologists. As recently as 1930
Nichols still considered the possibility of them
being separate species. Breder (1927) and
Beebe & Tee-Van (1928) were disposed to the
view that they were all one species. Tee-Van
(1932) referred T. nitidum and T. subfurcatus
Nichols to the synonomy of T. bifasciatum.
T. nitidissima was added to the synonomy by
Eeebe & Tee-Van (1933).

The coloration and pattern of the juvenile,
female and young male Thalassoma is princi-
pally that of a yellow-backed fish with or with-
out a dark lateral stripe. This stripe may be
broken up into blocks. Changes in color pat-
terns are observable in the living individuals.
Variations of this type of pattern are not the
concern of the present study, however.

The presence of males with two different
colorations in a single species suggested that the
factor instituting the change from small yellow
male to larger blue male was hormonal rather
than chromosomal. If this were the mechan-
ism, injections of sufficient androgenic hormone
given to yellow-phase fish should cause the

^This work has been supported in part by a grant
from the National Science Foundation.

blue phase to replace it and would clearly estab-
lish these two forms as individuals of the same
species, accounting for the presence of two types
of males and only one type of female.

Darby (1935-1936) reported Testosteron
(Ciba) effective in causing "... the immature
bluehead (Thalassophryne sp.)” [«c] to take
on mature coloration. The designation of the
fish used as Thalassophryne sp. is probably in-
correct as this form is a genus of toadfishes
found in the Pacific. There is one report (Giin-
ther, 1861) of a single Atlantic species, Thalas-
sophryne maculosa Gunther, from Panama, but
this report is in question. There have been no
reports of Thalassophryne sp. in the vicinity of
the Dry Tortugas. Thalassoma was probably
the “bluehead” used. No subsequent publication
on the problem has been found.

I should like to express my appreciation to
Dr. C. M. Breder, Jr., and Miss Priscilla Ras-
quin for their suggestions and advice during the
course of this study.

Materials and Methods

The blueheads used in this investigation were
collected from tidal flat Thallasia beds and
around small coral growths in Bimini, Baha-
mas, B. W. I. Experimental work was done
in the Lerner Marine Laboratory at Bimini.
While under observation, the fish were kept in
aquaria with running sea water. Conch shells
were provided for shelter.

Four groups of fish were used. Group 1 com-
prised “nitidum” and “bifasciatum” fish col-
lected in February and March. “Nitidum” and
“bifasciatum” designate color phases and are
described in a separate section. These fish were
used for study of normal tissue. Group 2 con-
sisted of “nitidum” fish used in a preliminary
experiment to determine dosage and medium for
suspension of the hormone. Metliyl testosterone

125

126

Zoologica: New York Zoological Society

[40: 11

was dissolved in 95 % alcohol and suspended in
physiological saline. Injections were given intra-
peritoneally without anesthesia. Dosage was not
equal for all fish, as the saline caused a partial
precipitation of the hormone and additional
95% alcohol was added to restore the suspen-
sion. The alcohol was found to be toxic and
fatalities were numerous. Another set of fish was
designated as Group 3. In this case methyl test-
osterone was dissolved in 95% alcohol, sus-
pended in sesame oil and incubated at 37° C.
to evaporate the alcohol. A 1 % solution of ethyl
urethane dissolved in sea water was used as an
anesthetic. A group of fish was collected in June
for comparison with Group 1. These fish were
designated as Group 4. The ratio of males to
females in June differed from the ratio in the
February-March period. In March, 28 fish were
collected of which nine were males and 19
females. Two to three months later, 18 fish were
collected and ten were males, seven females and
one a true hermaphrodite having one lobe of
testicular tissue and one of ovarian tissue (PI.
Ill, Figs. 15 & 16).

The numbers of fish, dosages and descriptions
of fish in all groups are given in Table 1 with
dates of collection, treatment and sacrifice.

All fish were killed by a lethal dose of ethyl
urethane. This caused the dispersion of melanin
granules, making the fish assume their darkest
coloration. Photographs in PI. I were taken after
the fish had been killed, in order to show them
uniformly in this darkest and most marked
pattern.

Bouin’s solution was used for fixation. The
gonads were dissected from all fish, imbedded
in paraffin and sectioned at Ig. Sections were
stained with Harris’s hematoxylin and eosin and
with Masson’s trichrome stain for connective
tissue.

Observations

Color and Pattern.— The predominant color
of the “nitidum” stage is yellow. A dark median
lateral stripe marks the fish from the snout to
the caudal fin, running through the eye. The
stripe is sometimes modified into six rectangular
areas separated by gray or yellow bars. These
intermediate markings are the basis for the dis-
tinction between “nitidum” and “nitidissimum,”
the former notably displaying the gray and the
latter the yellow. In “nitidum,” the dorsal fin is
bordered by light blue and the caudal fin has a
dark margin. In shape, the caudal fin is slightly
concave as the median rays are shorter than the
peripheral rays. Ventral and anal fins are trans-
parent but the pectoral fins have slightly dusky
tips. A fish in the “nitidum” stage is shown in
Pl..I„Fig. 1.

“Nitidissimum” would appear to be a modi-

fication of “nitidum” in which bands of yellow
replace the gray ones. This stage is more apt to
be a product of activity and environment than
of age or sex. Longley & Hildebrand noted that
this phase was shown most frequently by resting
fish. For the purposes of this study, “nitidissi-
mum” fish are included in the “nitidum” classi-
fication.

Blue is the principal color of the “bifasciatum”
stage. Longley & Hildebrand described a blue or
violet coloration on the head and the throat ex-
tending to the base of the dorsal fin. Fish in the
“bifasciatum” phase kept in the aquaria seldom
showed the violet coloration but the blue was
vivid. The body of the fish is banded by two
black bars situated behind the pectoral fins. The
black bars are separated by a blue bar. The re-
mainder of the body of the fish from the second
black bar is greenish. As in the case of the violet
eolor, green is seldom observed to be displayed
by fish in aquaria; blue or yellow is the color
shown. The pectoral fins have distinct black tips.
The caudal fin is markedly forked. The black
bars of the body extend through the dorsal fin
which is otherwise transparent. Fish in this
phase are larger than “nitidum” phase fish. PI. I,
Fig. 4, pictures an untreated “bifasciatum” fish.

The injection of methyl testosterone produced
“bifasciatum” coloration and pattern in “niti-
dum” phase fish regardless of sex. PI. I, Fig. 2,
shows a treated male; Fig. 3 shows a treated
female. The first sign of the “bifasciatum” col-
oration became apparent four days after injec-
tion. During these four days the yellow color
and “nitidum” pattern gradually faded. The blue
coloration started in the head region and three
days later the black bars in the region of the
pectoral fins began to appear. The initial black
of the bars originated at the point where the bar
would have bisected the stripe of the “nitidum”
phase. The tips of the pectoral fins which cover
this region became darker because of an increase
in the number of melanophores. PI. II, Figs. 5
and 6 show the difference in the amount of
pectoral fin pigmentation between a “nitidum”
stage and an experimentally-induced “bifasci-
atum” stage. In the ten days following the ap-
pearance of the black bar, the blue coloration
spread over the entire body of the fish, gradually
increasing in intensity.

In the aquaria, the fish were light blue and
the “bifasciatum” pattern was evident but not
intense. When sacrificed in urethane with the
resulting full expansion of the pigment cells, the
fish were vivid blue and the pattern well defined
in all but two of the 1 5 fish. These two fish were
blue but showed both faint black bars and faint
black stripes. The experimentally-induced blue
color, however, was never as intense as the blue

Table 1. Summary of Four Groups of Thalassoma bifasciatum.

1955]

Stoll: Hormonal Control of Pigmentation of Thalassoma bifasciatum

127

c

o

a

a» a
^6

o

>

O

o> a

^6

CO CO
<U 0>

H o
>

H

or-- -H

to CO • o
(U ^(U X3 .ts
to

o cd 55 o
H > H 2

O ^

cd

Q

-D

O

cd

s

c ^

d> X

S B

a

t-Q

c c

o o

Z Z

<u
eb o
S

OO

• C
o

6

<u

C

S "

M S5

'T* I—. '

Q> O

6

•sl'

I

C\3 ^

»n

Os

— J3

^ fli

u

Cd .£3

E

C

o

Sift'S 8

2 S E S

E o

•rd cd

Tt*

r-

o

r--

GO
)U C
GO
Cd
Im

^ 53

Cd

E (u
E E

cC

X>

E

Z

Cd

•C

CL*

E

D

.2 E

O 3

S 3

^ • ti

S ’c

3

’O

3

T3

3

Cd rti

o

CL,

P

o

u

O

CJ

o

t-l

a

lo

CJ

Oh

D

O

0

^ c

Ti 3
cd

PP

CL.

3

O

a

128

Zoologica: New York Zoological Society

displayed by the wild fish in the “bifasciatum”
phase. Comparison of the wild and the experi-
mental “bifasciatum” coloration showed that
the green element was lacking in the experi-
mentals.

Goodrich & Biesinger (1953), working on
the histology of the coloration of Thallasoma
bifasciatum, noted that the underlying tissue of
the green scales had a layer of guanophores as
well as the xanthophores and melanophores
present in the black and the blue areas. The
number of melanophores present in the green
areas was less than the number present for the
other two colors: 200 per sq. mm. in green
areas and 450-500 per sq. mm. in black areas.
However, the number of xanthophores present
in the green areas was greater than that present
in black or blue areas; 350 per sq. mm. for
green, 15 scattered per sq. mm. for blue and 45
per sq. mm. for black. All the numbers given
were approximations.

There are two possible explanations for the
nonappearance of the green color: (1), that the
two weeks’ time the experimental fish were
maintained after injection was not long enough
to build up the concentration of xanthophores
and guanophores necessary to produce green
coloration; and (2), that an improper diet
caused loss of color in the xanthophores (Fox,
1953). The one fish from Group 2 which was
maintained for more than six weeks did show
some green coloration when anesthetized in
ethyl urethane.

Histology— In histological section, distinction
could be made between normal “nitidum” testes
and normal “bifasciatum” testes. The testes of
males in the “nitidum” phase contained large
reservoirs of mature sperm aggregated in the
lumen of the sperm duct and held in the tubules.
The bulk of the testes was comprised of large
cysts with thin membranes containing mature
sperm. Very few early or intermediate stages of
spermatogenesis could be noted (PI. II, Fig. 7) .

Small reservoirs of mature sperm were seen
in the tubules of “bifasciatum” testes. The
greater part of the testes, however, was made up
of cysts in early or intermediate stages of
spermatogenesis. A few cysts were present
which contained mature sperm and these ap-
peared to be in the stage immediately preceding
rupture. The sperm cells were clustered on or
near the walls of the cysts while the centers of
the cysts were empty (PI. II, Fig. 8). The testes
of fish in this stage were half as large as testes
from fish in the “nitidum” phase. No interstitial
cells like those described by Courrier (1921)
were identified. The specific source of the andro-
genic hormone responsible for the blue color
and pattern is as yet unknown.

[40: 11 I

'

The presence of large quantities of sperm in
the testes of “nitidum” males indicates that sex-
ual maturity for the males is attained in this
phase. The absence of large sperm reservoirs in \
“bifasciatum” fish suggests that some spawning |
has occurred during the “nitidum” phase to de- }
plete the sperm reservoirs. i

Normal ovaries showed early and interme-
diate maturation stages as well as mature eggs. |
The follicles were large and the ova were sur- |
rounded by heavy chorionic membranes and en-
circled by follicular cells (PI. II, Fig. 9). ■

Sesame oil injections used for control pur- !
poses produced variations from the normal in j
both “nitidum” males and females. The testes j
of the two males of the group showed all ma-
ture sperm, indicating that spermatogenesis had
been accelerated. The sperm were held within
the cysts and the membranes were well-defined
and thick. Only a few small sperm reservoirs
could be seen. Spaces formerly occupied by the
sesame oil which had been dissolved by the his-
tological procedure were scattered throughout
the gonads (PI. II, Fig. 10). The size of the
testis was not affected by the sesame oil.

Ovaries from sesame oil-injected females
were about one-quarter the size of the normal
ovaries. Each ovary was sac-like in structure and
enclosed a large empty lumen. The walls of the
sac were as thin as three to four cell layers in
some fish. No mature eggs were seen; however,
loose follicular cells were present. Very little
connective tissue was seen. A few chorionic
membranes were observed, which suggested that
some eggs had been resorbed (PI. Ill, Figs. 11
& 12).

The testes of males injected with methyl
testosterone showed only mature sperm. The
membranes of the cysts were thin and there were
large reservoirs of mature sperm (PI. Ill, Fig.
13). The androgenic hormone accelerated the
production of mature sperm. Early and inter-
mediate stages of maturation were absent.

The gonads of the 13 other fish which had
received injections of methyl testosterone
showed ovarian tissue interspersed with matura-
tion stages of spermatogenesis (PI. Ill, Fig. 14).
The predominance of ovarian tissue indicates
that these fish were females before injection.
The ovaries were drastically reduced in size to
less than a quarter the size of ovaries of normal
females collected at the same time and location.
The ovary appeared as a sac-like structure,
hollow and collapsed. In section, maturation
stages, mature eggs and degenerating eggs could
be seen. Collapsed chorionic membranes were
plentiful and there was an abundance of con-
nective tissue. All stages of spermatogenesis
could be identified in the ovaries and small res-

1955]

Stoll: Hormonal Control of Pigmentation of Thalassoma bifasciatum

129

ervoirs of mature sperm were present. The
testicular tissue was present throughout the
organ and did not appear to be especially abun-
dant in any specific part. This tissue is thought
to have arisen from primordial germ cells.

Discussion

Previous experiments with sex reversal in
fishes have been carried out primarily on repre-
sentatives of the poeciliid fishes. Particular em-
phasis has been given to Xiphophorus helleri
(Heckel). The development in the female of
the secondary sex characters of the male of this
species has served as an external index of pos-
sible sex reversal.

Essenberg (1926) reported spontaneous sex
reversal from female to male in X. helleri. The
reversal occurred even after the female had
given birth to one or more broods. After reversal
the secondary sex characters of the male were
exhibited by the fish but the female shape was
retained. On the basis of these observations,
Essenberg concluded that sex in this species was
determined and controlled hormonally and not
genetically and stated that any agent or condi-
tion which tends to decrease the capacity of the
female sex hormone secretion beyond a certain
limit becomes an immediate factor in sex re-
versal in the female.

In 1937, Witschi & Crown found that non-
pregnant female X. helleri subjected to testos-
terone propionate (Ciba) dissolved in the
aquarium water, absorbed their eggs and the
ovaries resembled testes although no spermato-
genesis was observed. The secondary sex char-
acters of the male were displayed. Pregnant fe-
males under similar treatment aborted or ab-
sorbed their eggs within one to two days.

Baldwin & Goldin (1939) reported histo-
logical changes in ovaries of 50% of a group of
virgin female X. helleri injected with testoster-
one propionate dissolved in sesame oil. The
changes included absorption of the gonad and
the presence of some phases of spermatogenesis.
Baldwin & Li (1942) demonstrated the possi-
bility of complete sex reversal in adult female
X. helleri treated with gonadotrophic (human
chorionic) hormone, and later (1945) cited
two cases of ovotestes in males that had been
injected with alpha-estradiol benzoate.

Burger (1942) treated male Fundulus with
testosterone propionate and found that it had
only a slight stimulating effect on the male germ
cells. An increased coloration was noted and an
increase in the extent of the testicular duct
system.

Mature female Gambusia holbrooki Girard
developed masculinized anal fans when treated
with testosterone propionate. Immature females

showed inhibition of growth (Hamon, 1946).
Female Gambusia affinis maintained in a solu-
tion of ethynyl testosterone for two to four days
developed masculine secondary sex characters
and retained these characters up to 60 days after
they had been replaced in fresh water. On the
basis of this experiment, Turner (1946) stated
that this response was a clear indication that the
genetic factors for the male characters were
present in the female but normally did not de-
velop because of insufficient androgenic hor-
mone.

Zei (1950) described typical sex reversal in
three Mediterranean fishes, Maena smaris
(Linnaeus), M. chryselis (Cuvier & Valen-
ciennes) and Pagellus erythrinus (Cuvier &
Valenciennes) . The fish are female for the first
half of their lives and male the second. Trans-
formation takes place at the 13-15 cm. size. The
males may be identified by brighter color, larger
size and better developed anal and dorsal fins.

The histology of hermaphroditism in Ser-
ranus and Sargus was described by Van Oordt
(1929). Lavenda (1949) reported the presence
of developing testicular tissue in functional fe-
male sea bass, Centropristes striatus. The tes-
ticular tissue arose from the epithelial lining of
the oviduct.

According to D’ Ancona (1950) hermaphro-
ditism in teleosts is found only in some species of
Sparida and Serranida. In the case of the
Sparidae, he assumes two different germinal
areas for gonad origination reaching maturity
successively and each producing distinct sex
differentiators known as gynogenine and andro-
genine.

Courrier (1921) described interstitial cells
in the testes of representatives of the Gobiidae,
Callionymidae, Cottidae, Cichlidae and Gaste-
rosteidae. The function of control of the sec-
ondary sex characters was ascribed to the inter-
stitial cells. Van Oordt (1925) found no
interstitial cells in the testis of X. helleri I
(Heckel). Craig-Bennett (1931), studying
Gasterosteiis aculeatus Linnaeus, found inter-
stitial cells most abundant during the breeding
season and inconspicuous in quiescence.

The existence of teleost androgens has been
experimentally demonstrated. In 1937, Hazle-
ton & Goodrich accelerated comb growth of two
capons with an extract of the testes of salmon,
Oncorhynchus kisutch. Potter & Hoar (1954)
reported androgens in the testes of Oncorhyn-
chus keta Walbaum. Histological examination
of the testes showed that interstitial cells and
extracts produced comb growth in baby chicks.
The number of interstitial cells was dependent
on the season, as with Gasterosteus aculeatus
Linnaeus.

130

Zoolog ica: New York Zoological Society

[40: 11

Results of the present report show that T.
bifasciatum is not a hermaphroditic fish but is
possibly a progynous species in which all indi-
viduals start fife as female and later become
male. Development of the “bifasciatum” phase
is under the control of the androgenic hormone
as injection of male hormone brought about the
male coloration regardless of the sex of the
“nitidum” fish used. However, the appearance of
the blue, “bifasciatum” coloration is not concom-
itant with sexual maturity, because mature testes
are found in the yellow phase, but it may be a
function of age and length of time the andro-
genic hormone has been in action.

The sac-like structure and reduced size of the
ovary of the hormone-treated fish may be the re-
sult of the sesame oil in which the hormone was
suspended. The ovaries of females treated with
sesame oil alone showed a similar modification
in size and structure. It is of interest to note that
the sesame oil was inert in regard to testicular
tissue and caused no change in size or structure
of testes.

Summary

1. All Thalassoma bifasciatum with blue
heads are males but fish with largely yellow
coloration are male, female or juvenile.

2. Large reservoirs of mature sperm were
found in testes of yellow males, indicating that
sexual maturity is reached in this coloration
phase. The testes of blue males contained num-
erous early and intermediate stages of sperma-
togenesis but very few mature sperm. The testes
of yellow males were twice as large as the testes
of blue males.

3. Intraperitoneal injection of methyl testos-
terone in yeUow-phase fish produced blue-phase
color and pattern in both sexes.

4. Androgenic hormone is responsible for the
change from yellow to blue phase. The first in-
dication of color change was noted on the head
region four days after injection, the first indi-
cation of pattern change was observed three days
later, on and under the tips of the pectoral fins
with an increase in the number of melanophores
present.

5. Methyl testosterone produced the develop-
ment of testicular tissue and the regression and
absorption of ovarian tissue in yeUow females
and acceleration in the production of sperm in
yellow males.

6. Testicular tissue that developed in yellow
females evidently arose from the primordial
germ cells.

7. The experimentally produced ovotestes con-
tained large amounts of connective tissue absent
in the gonads of both normal females and males.

8. No interstitial cells for androgenic hormone

elaboration could be found in testes from blue-

phase fish.

References

Baldwin, Francis M., & H. S. Goldin

1939. Effects of testosterone propionate on the
female viviparous teleost, Xiphophorus
helleri Heckel. Proc. Soc. Exp. Biol, and
Med., 42: 813-819.

Baldwin, Francis M., & M. H. Li

1942. Effects of gonadotropic hormone in the
fish, Xiphophorus helleri Heckel. Proc.
Soc. Exp. Biol, and Med., 49: 601-604.

1945. Induction of ovotestis in estrogen treated
adult teleost, Xiphophorus helleri. Amer.
Nat., 79: 281-286.

Beebe, William, & John Tee-Van

1928. The fishes of Port-au-Prince Bay, Haiti.
Zoologica, 10: 1-279.

1933. Nomenclatural notes on the shore fishes
of Bermuda. Zoologica, 13: 133-158.

Breder, C. M., Jr.

1927. Scientific results of the first oceanographic
expedition of the “Pawnee” 1925. Fishes.
Bull. Bing. Oceanogr. Coll., 1 (1): 1-90.

Burger, J. W.

1942. Some effects of androgens on the adult
male Fundulus. Biol. Bull., 82: 233-242.

COURRIER, R.

1921. Sur I’existence d’une glande interstitielle
dans le testicule des poissons. C. R. Soc.
Biol. Paris, 85: 939-941.

Craig-Bennett, a.

1931. The reproductive cycle of the three-spined
stickleback, Gasterosteus aculeatus Linn.
Phil. Trans. Roy. Soc. London, B 219:
197-279.

D’Ancona, U.

1949. Ermafroditismo ed intersessualita nei Te-
leostei. Experientia, 5: 381-389.

Darby, H. H.

1936. Sex hormones in fishes. Annual Report of
the Tortugas Laboratory. Carnegie Inst.
Wash. Division of Animal Biology, Year
Book no. 35 for 1935-1936: 84.

Essenberg, J. W.

1926. Complete sex reversal in viviparous
teleosts. Biol. Bull., 51: 98-109.

Fox, Denis L.

1953. Animal biochromes and structural colours.
Cambridge University Press, Cambridge:
xiv. 378.

1955]

Stoll: Hormonal Control of Pigmentation of Thalassoma bifasciatum

131

Goodrich, H. B., & Dorothy L. Biesinger
1953. The histological basis of color patterns in
three tropical marine fish with observa-
tions on regeneration and the production
of the blue color. Jour. Morph., 93 (3):
465-487.

Hamon, M.

1945. Effets morphologiques du propionate de
testosterone sur la femelle de Gambusia
holbrooki Gir. C. R. Soc. Biol. Paris,
139: 110-111.

Hazleton, L. W., & F. J. Goodrich

1937. A note on the presence of male sex hor-
mones in fish testes. Jour. Amer. Pharm.
Assoc., 26: 420-421.

Lavenda, N.

1949. Sexual differences and normal protogyn-
ous hermaphroditism in the Atlantic sea
bass, Centropristes striatus. Copeia, 3:
185-194.

Longley, W. H., & S. F. Hildebrand

1941. Systematic catalogue of the fish of Tor-
tugas, Florida. Papers from the Tortugas
Laboratory, volume XXXIV. Carnegie
Inst. Wash. Publ. 535: 196-198.

Nichols, J. T.

1930. The fishes of Porto Rico and the Virgin
Islands. Scientific Survey of Porto Rico
and the Virgin Islands, New York Acad.
Sci., X, pt. 3: 316-317.

Potter, G .D., & W. S. Hoar

1954. The presence of androgens in chum sal-
mon (Oncorhynclms keta Walbaum).
Jour. Fish Res. Bd. Can., 11: 63-68

Tee- Van, John

1932. Color changes in the blue-head wrasse
Thalassoma bifasciatum (Bloch). Bull.
New York Zool. Soc., 35 (2): 43-47.

Turner, C. L.

1946. Retention of response to ethynyl testoster-
one in females of Gambusia affinis. Jour.
Exp. Zool., 102: 357-369.

Van Oordt, G. J.

1925. The relation between the development of
the secondary sex characters and the
structure of the testis in the teleost X. hel-
ZcriHeckel. Brit. Jour. Exp. Biol., 3:43-59.

1929. Zur mikroskopischen anatomie der ova-
riotestes von Serranus und Sargus. Zeit-
schr. Mikros. Anat. Forsch., 19: 1-17.

WiTSCHi, E., & E. N. Crown

1937. Hormones and sex determination in fishes
and frogs. Anat. Rec., 70 ( 1 ), Suppl. ( 1 ) :
Abstr. 205

Zei, M.

1950. Typical sex-reversal in teleosts. Proc.
Zool. Soc. London, 119: 917-920.

132

Zoologica: New York Zoological Society

[40: 11; 1955]

Plate I

Fig. 1. Female bluehead in the “nitidum” phase.

X .71.

Fig. 2. “Nitidum”-phase male three weeks after
injection of 2 mg. methyl testosterone. The
black lateral stripe has been obliterated
and the color and pattern of the “bifas-
ciatum” stage has begun to appear. X .74.

Fig. 3. Female “nitidum”-phase fish after injec-
tion of 4 mg. methyl testosterone, show-
ing pronounced “bifasciatum” color and
pattern. The blue color is indistinguish-
able in black and white reproduction.
X .71.

Fig. 4. Untreated male in “bifasciatum” phase.

X .84.

Plate n

Fig. 5. Tip of pectoral fin of a “nitidum”-phase
fish. X 100.

Fig. 6. Tip of pectoral fin from a fish in an
experimentally - produced “bifasciatum”
phase. X 100.

Fig. 7. Testis of untreated “nitidum”-phase fish
showing relatively large numbers of ma-
ture sperm and a few maturation stages.
X 600.

Fig. 8. Testis of untreated “bifasciatum”-phase
fish showing predominance of early stages

of spermatogenesis and relatively few
mature sperm. X 600.

Ovary of untreated “nitidum” fish. X 90.
Testis of “nitidum”-phase fish three
weeks after injection of 0.5 cc. sesame oil.
X 100.

Plate III

Fig. 11. Ovary three weeks after injection of 0.5
cc. sesame oil, showing decrease in size
of the organ and decrease in size and num-
ber of eggs. X 100.

Fig. 12. Detail of Figure 11, showing collapsed
chorionic membrane. X 600.

Fig. 13. Testis after injection of 2 mg. methyl
testosterone, showing only mature sperm
and no stages of early spermatogenesis.
X 600.

Fig. 14. Gonad, designated as an ovotestis, three
weeks after injection of 2 mg. methyl
testosterone. Note early stages of sperma-
togenesis and degenerating ova. X 600.

Fig. 15. One lobe of gonad of hermaphroditic
“nitidum” fish, showing only testicular
tissue. X 110.

Fig. 16. The other lobe of the gonad pictured
in Figure 15, showing typical ovarian
structure. X 110.

EXPLANATION OF THE PLATES

Fig. 9.
Fig. 10.

STOLL

PLATE I

FIG. 1

FIG. 2

FIG. 3

FIG. 4

HORMONAL CONTROL OF THE SEXUALLY DIMORPHIC
PIGMENTATION OF THALASSOMA BIFASCIATUM

STOLL

PLATE II

FIG. 9

FIG. 10

HORMONAL CONTROL OF THE SEXUALLY DIMORPHIC
PIGMENTATION OF THALASSOMA BIFASCIATUM

STOLL

PLATE III

FIG. 12

FIG. 13

FIG. 14

HORMONAL CONTROL OF THE SEXUALLY DIMORPHIC
PIGMENTATION OF THALASSOMA BIFASCIATUM

FIG. 16

12

The Use of Copper Sulfate as a Cure for Fish Diseases
Caused by Parasitic Dinoflagellates of the Genus Oodinium

Robert P. Dempster

Steinhart Aquarium, California Academy of Sciences, San Francisco
(Plate I; Text-figure 1)

Introduction

CORAL-REEF fishes from the Hawaiian
Islands have been received by the Stein-
hart Aquarium for many years and
until 1951 they were relatively free from any
disease that could be considered as epidemic.
In that year, however, about ten days after an
exceptionally large shipment arrived from Ho-
nolulu, a gill disease broke out in the tanks. At
first the fishes were not suspected of having any
specific disease. Many of them congregated near
the surface and were observed to be respiring
very rapidly, obviously in great distress. It was
apparent that either the water was deficient in
oxygen or the fishes for some reason were not
able to utilize the oxygen that was present. On
the assumption that the tanks had become con-
taminated, they were drained and refilled with
fresh sea water and the circulation of water was
substantially increased, but the fish were not
relieved of their respiratory trouble. As time
went by, more individuals congregated near
the surface, gasping for air, and it was not long
before some of them died.

Microscopic examination revealed innumer-
able minute, oval, parasitic organisms clinging
to the gill filaments of the dead fish (Plate I) , so
thickly planted that they were obviously inter-
fering with the respiratory function of the gills.
They were determined to be a species of Oodin-
ium, possibly ocellatum. Left overnight in a
dish of water, they were re-examined the next
morning and it was found that some had di-
vided once and many twice, so that all had
passed into the 2- or 4-cell stage of development
within the period of about 24 hours. Nigrelli
(1936) found that each Oodinium organism,
after becoming detached from the gills of a fish,
settled to the substratum, where it gave rise to

palmella stages of 2, 4, 8, 16, 32, 64 and 128
cells. One more palmella division took place to
form 256 flagellated, free-swimming dinospores.
Later the dinospores settled to the bottom, where
they developed into typical peridinian dinoflagel-
lates, the infective form. These dinoflagellates,
which are also free swimming, apparently in-
vade the branchial chamber of the fish, become
attached to the gill filaments and metamorphose
into the parasitic form.

According to Jacobs (1946), Oodinium ocel-
latum is the first dinoflagellate known to para-
sitize marine vertebrates.

Treatment of Marine
Oodinium Infestations

The immediate problem was to determine how
to relieve the fish of this very prolific parasite
and how to eradicate it from the Aquarium’s
water system. Several methods of treatment
were tried; the most effective entailed the use
of copper sulfate. That copper is highly toxic
to fish and that it must be used with extreme
caution is well known. After making a consid-
erable number of tests with tropical marine
fishes, I have found that 0.5 p.p.m. is a safe con-
centration and is lethal to Oodinium. When
treating large volumes of water, especially in an
aquarium where the amount of untreated in-
coming water may fluctuate greatly, it may be
difficult to maintain the concentration exactly
at that level. Excellent results can be achieved
even though the copper concentration is allowed
to fluctuate from 0.4 p.p.m. to 0.8 p.p.m.,
although 0.8 p.p.m. should be considered the
upper limit and should not be maintained for
long periods.

Within the allowable range of concentration,
the copper induces the fish to secrete a copious

133

134

Zoologica: New York Zoological Society

amount of mucus which causes the parasites to
become detached. After they are sloughed from
the body and settle to the bottom, cell division
takes place and development proceeds normally
to the free-swimming dinoflagellate stage, and
at this point the copper sulfate apparently be-
comes lethal to them. Since it takes about seven
days for the Oodinium organisms to develop into
free-swimming dinoflagellates, it is necessary to
maintain the copper concentration in the tank
for at least that long. A ten-day treatment with
copper sulfate is recommended.

Nigrelli (1936) reported that numerous
marine fishes in the New York Aquarium were at
one time heavily infected with Oodinium ocel-
latum and that the infection was not confined to
the gills but was found on almost any part of
the body. On fishes collected in the vicinity of
the Hawaiian Islands I have found Oodinium
only on the gill filaments. However, in May,
1954, some very sick Clown Fish (Amphyprion
percula) that had been collected near Singapore
were brought to the Aquarium for examination,
and both gills and body were found to be cov-
ered with Oodinium. Treatment with copper
sulfate solution was begun immediately. Two
days later the parasites had disappeared from the
body and gills. These fish were covered with tiny
pit marks caused by the parasitic organisms, and
one fish, more heavily pitted than the others,
died two days later from severe fungus infec-
tion apparently resulting from the minute skin
punctures. The others lived for several months
after treatment, without recurrence of Oodin-
ium.

Treatment of Freshwater
Oodinium Infestations

Oodinium limneticum, a species described
by Jacobs (1946), attacks many species of ex-
otic freshwater fishes and causes a malady
known to fish fanciers as velvet disease. Jacobs
states that this is the first parasitic dinoflagel-
late known to attack freshwater fishes. It may
occur on all external portions of the body, in-
cluding the fins, trunk, eyes, mouth and gills.
A fish parasitized with O. limneticum somewhat
resembles one with an Ichthyophthirius infec-
tion, and because of this the disease may be
mistakenly diagnosed; microscopic examina-
tion is necessary in order to make a positive
diagnosis. Treatments most commonly used to
cure Ichthyophthirius disease usually do not cure
fishes with velvet disease.

While this paper was being prepared, velvet
disease occurred only once in the Steinhart
Aquarium. In this instance a tank of Glassfish
{Chanda lala) became infected. Copper sulfate
was added to the tank at a 0.5 p.p.m. level and

[40: 12

two days after treatment was begun the disease
disappeared completely. The fish suffered no ill
effects from the copper. However, since our
experience is so limited, we recommend cau-
tion when treating tropical freshwater fish with
copper sulfate.

Determination of Concentration
OF Copper

When treating fish with copper sulfate, it is
extremely important to make a daily chemical
analysis of the water in the treatment tanks so
that a constant level of copper may be main-
tained. In calculating the amount of copper sul-
fate needed to make up a solution containing a
given amount of the metal, one must take into
consideration the fact that the copper represents
approximately only 25% of the total weight of
CuSOr • 5H2O. To determine the amount of
CuSOi • 5H2O that must be added to a given
volume of water in order to produce a desired
concentration of copper in p.p.m., the following
equation may be used:

x =: V X p X 3.93
1000

X - weight of CUSO4 ♦ 5H2O in grams.

V = volume in liters,
p = parts per million of copper desired.

3.93 = number of grams of CuSO4*5H20 con-
taining 1 gram of copper.

Copper does not long remain in solution in the
presence of excess amounts of carbon dioxide
and carbonates; therefore, if practicable, the
fish to be treated should be transferred to a clean,
nonmetallic tank devoid of all coral, shells, etc.
If these substances are present, the copper may
react with them to produce insoluble carbonates.
Should it not be practicable to remove the fish
from such an environment, copper sulfate may
have to be added daily to compensate for the
loss through precipitation.^ If freshwater fishes
infected with Oodinium limneticum are being
treated, it is important to know whether the
water is hard or soft. Copper is readily precip-
itated from hard water.

The concentration of copper in water may be
determined by a colorimetric method using
sodium diethyldithiocarbamate as the indicator.
The reagents necessary for this are prepared in
the following ways:

Sodium diethyldithiocarbamate solution: Dis-
solve 1 g of N(C2H6)2 CS2Na in 100 ml of
copper-free distilled water and keep in a bot-
tle of dark glass protected from sunlight. Add

1 Precipitation of copper may be substantially de-
creased by the addition of citric acid to the copper
sulfate solution before it is added to sea water. One
part by weight of citric acid to 100 parts of copper sul-
fate usually suffices.

1955]

Dempster: Copper Sulfate as a Cure for Certain Fish Diseases

135

NH4OH until the pH reaches 9.6-10. This
retards decomposition of the carbamate. This
reagent will remain stable for approximately
40 days when stored at this pH in a dark
place.

Copper-free distilled water-. Redistill distilled
water, using an all-glass still.

Copper sulfate standard solution: Dissolve
0.393 g of CuSOi • 5H2O in 1 liter of distilled
water (copper-free) . Dilute 25 ml to 250 ml.
One ml of the diluted solution contains 0.01
mg cu.

Ammonium citrate buffer 20%: Dissolve
200 g of citric acid in about 500 ml of cop-
per-free distilled water. Add C.P. NH4OH
until the pH reaches 9.0-9.2. Add copper-free
distilled water to 1,000 ml.

Carbon tetrachloride solution— reagent grade.
The detailed procedure for colorimetric analy-
sis may be described as follows : In a separatory
funnel, place 50 ml of the water to be analyzed,
5 ml of ammonium citrate buffer, and mix. Then
add 1 ml of sodium diethyldithiocarbamate re-
agent and mix again. (The carbamate reagent,
when introduced into the water sample contain-
ing copper, produces an amber color, the in-
tensity of which is in direct proportion to the
amount of copper present). To this mixture add
exactly 10 ml of carbon tetrachloride and shake
for at least 2 minutes to extract the color com-
pletely. The solution should be allowed to stand

for 10 minutes to achieve complete phase separa-
tion. Carefully drain off the colored layer into a
test tube and read in the colorimeter.^ The color-
imeter must have been previously adjusted so
as to read 100 percent light transmission when
a test tube containing the reagent grade carbon
tetrachloride has been inserted. To interpret the
readings in the colorimeter, it is necessary to
refer to a calibration graph based on water sam-
ples containing known amounts of copper (Text-
fig. 1). These water samples are prepared by
adding carefully calculated quantities of the
standard copper sulfate solution to 50 ml quan-
tities of copper-free distilled water. The samples
are then analyzed in the colorimeter and the
readings plotted. Several colored carbon tetra-
chloride samples were measured at different
wave lengths of light, and maximum light ab-
sorption was observed at 415 to 450 mu. Table
1 indicates that a wave length of 435 would be
ideal. This agrees closely with the findings of
Chow & Thompson (1952).

Discussion

It is probable that there is wide variation
among different species of fishes in their toler-
ance to copper. Some tropical marine species
begin to show distress at 1 p.p.m. and salmonoid
fishes are adversely affected by concentrations

- A Bausch & Lomb spectronic twenty colorimeter
was used for this purpose.

Text-fig. 1. Calibration graph in-
dicating the percentage of light
transmission through samples of
known concentration of copper-
diethyldithiocarbamate. Samples ex-
tracted with carbon tetrachloride
from distilled water.

c

g

’(/>

in

£

in

c.

o

SI

O’

<D

cn

o

c

CL>

O

p.p.m. of copper

136

Zoologica: New York Zoological Society

[40: 12

Table 1. Percentage of Light Transmission of Different Wave Lengths of Light through
Various Concentrations of Copper-diethyldithiocarbamate Solution. Samples Extracted with
Carbon Tetrachloride and Measured with B. & L. Spectronic Twenty Colorimeter.

Wave Length

p.p.m.
of copper

375 mu

400 mu

415 mu

425 mu

435 mu

450 mu

475 mu

500 mu

.05

97%

92%

88%

87%

86%

89%

94%

98%

.1

95

82

72

69

67

73

85

90

.2

91

75

60

55

54

61

75

88

.3

89

65

46

39

38

46

66

82

.4

87

60

38

32

32

39

61

76

.5

84

53

31

26

25

32

55

73

.6

82

48

25

20

19

24

47

69

.7

81

43

22

15

14

19

42

65

.8

74

36

16

11

10

15

37

58

.9

71

33

13

9

8

12

31

57

1.

70

29

11

7

6

11

30

54

of less than 0.5 p.p.m. Brook Trout fingerlings
will die in concentrations above 0.1 p.p.m.

Oodinium disease appears to be very widely
distributed throughout the tropic seas and has
been reported from aquariums in different parts
of the world. The Director of the Taraporevala
Aquarium at Bombay reported in 1954 that this
gill disease had affected some of the fishes on ex-
hibition there, and the Honolulu Aquarium is
periodically invaded by Oodinium. Reports of
Oodinium disease have also come from the
Zoological Society of London’s Aquarium, the
oceanarium (Marine Studios) at Marineland,
Florida, and the New York Aquarium. Although
this dinofiagellate seems to occur primarily in
warm water, it invaded the temperate water
system in the Steinhart Aquarium in at least one
instance. Shortly after the severe attack on reef
fishes in 1951, it was found on some of our local
coastal fishes. Several Striped Bass (Roccus sax-
atilis) , Rubberlip Seaperch {Rhacochilus toxo-
tes) and Lingcod {Ophiodon elongatus) died
from the disease. The water temperature in then-
tanks was approximately 65° Fahrenheit at the
time of infection. It should also be noted that
Nigrelli (1936) found Oodinium on fishes col-
lected in Sandy Hook Bay, New Jersey, during
the summertime. From this evidence it does not
seem unreasonable to believe that this gill dis-
ease could invade the cooler waters along the
California coast and cause serious damage to an
important food fishery.

Summary

At the Steinhart Aquarium, copper sulfate,
in concentrations ranging from 0.4 p.p.m. to
0.8 p.p.m., has been found relatively non-toxic
for tropical marine fishes, yet effective in eradi-

cating the gill and skin infections of the dino-
flagellate, Oodinium. Because of the toxicity of
higher concentrations of copper, it is essential
to control the amount present in the aquarium
water. To make this possible a colorimetric
method of determining the quantity of copper
present in water is described. Velvet disease,
which is caused by a freshwater dinofiagellate,
O. limneticum, also appears to be cured by
treatment with copper sulfate.

Acknowledgements

I wish to extend my thanks to Dr. Earl S.
Herald, Curator of Aquatic Biology, Steinhart
Aquarium, under whose direction the present
work was done, for his interest, advice and
sound judgment. I should also like to express
my appreciation to Dr. Albert E. Bagot, Chemist,
North Point Sewage Treatment Plant, San Fran-
cisco, for his assistance in working out a simpli-
fied method for copper determination in sea
water, and to Dr. Jerald A. Ballou, Associate
Professor of Physical Sciences, San Francisco
State College, for his helpful advice regarding
all chemical problems.

Literature Cited

Chow, TsAmwA J., & Thomas G. Thompson

1952. The determination and distribution of
copper in sea water. Part 1. The spectro-
photometric determination of copper in
sea water. Jour. Marine Res., 11(2): 124-
137.

Jacobs, Don L.

1946. A new parasitic dinofiagellate from fresh-
water fish. Trans. Amer. Microscopical
Soc., 65(1): 1-17.

1955]

Dempster: Copper Sulfate as a Cure for Certain Fish Diseases

137

Nigrelli, Ross F.

1936. The morphology, cytology and life-his-
tory of Oodinium ocellatum Brown, a
dinoflagellate parasite on marine fishes.
Zoologica, 21(3) : 129-164.

138

Zoologica: New York Zoological Society

[40: 12: 1955]

EXPLANATION OF THE PLATE
Plate I

Fig. 1. Gill arch of a fish, showing Oodinium
parasites clinging to the gill filaments.

DEMPSTER

PLATE I

FIG. 1

THE USE OF COPPER SULFATE AS A CURE FOR FISH DISEASES CAUSED BY
PARASITIC DINOFLAGELLATES OF THE GENUS OODINIUM

I

-.■i'

4 V

13

Polymorphism in Reared Broods of
Heliconius Butterflies from Surinam and Trinidad'

William Beebe

Department of Tropical Research,

New York Zoological Society, New York 60, N. Y.

Plates I- VI

[This paper is one of a series emanating from
the tropical Field Station of the New York Zoolog-
ical Society at Simla, Arima Valley, Trinidad, Brit-
ish West Indies. This station was founded in 1950
by the Zoological Society’s Department of Tropical
Research, under the direction of Dr. William Beebe.
It comprises 200 acres in the middle of the Northern
Range, which includes large stretches of undis-
turbed government forest reserves. The laboratory
of the station is intended for research in tropical
ecology and in animal behavior. The altitude of the
research area is 500 to 1,800 feet, with an annual
rainfall of more than 100 inches.

[The present paper is chiefly concerned with the
results of an eighteen-day trip to Surinam, which
was undertaken in mid-April, 1954. Headquarters,
as on a previous visit in 1953, were at the Moengo
mine of the Surinaamsche Bauxite Maatschappij,
where we were the guests of Mr. and Mrs. H.
Meijer. The trip was made possible by the generous
gifts of Mr. C. R. Vose, and by the cooperation of
the Alcoa Steamship Company.

[The major part of the actual collecting of living
Heliconius material in 1954 was undertaken by
Henry Fleming, and the painstaking care of the
eggs, larvae, pupae and ultimate emergence of the
imagos was borne by Jocelyn Crane and Rosemary
Kenedy. Future papers will deal with immature
stages and rearing methods. For details of keeping
and maintaining adult heliconids, see “Construc-
tion and Operation of Butterfly Insectaries in the
Tropics,” Crane & Fleming (Zoologica, 1953, Vol.
38, No. 14, pp. 161-172.)]

Forty years ago two able entomologists, J.
Joicey and W. J. Kaye, made a thorough
study of a collection of butterflies from
French Guiana. The three paragraphs below are
from their paper (1916):

^Contribution No. 953, Department of Tropical Re-
search, New York Zoological Society.

“The following account is concerning a collec-
tion made during the months July, August and
September, 1915, between the places St. Jean and
St. Laurent on the Maroni River in French Guiana.
The distance between the two places is about twelve
miles or rather less, and the distance of St. Laurent
(the nearer place) from the coast is about twenty
miles. The collection, which contained numbers of
specimens of other families, was, however, chiefly
remarkable for the vast numbers and variety of
forms of Heliconius melpomene and Heliconius
erato. A few other species of Heliconius were ob-
tained, but only a very few specimens of each.

“As it is, there are 731 specimens, which show
a most wonderful range of variation. Many forms
are new, and others graduate completely into these
as well as to all the other known forms that have
ever come from French Guiana.

“In comparison with the very large numbers of
melpomene specimens the number of erato forms
is small, being only 155 against 731 melpomene."

The object of these quotations is to empha-
size several facts. First, the closeness of the
locality to that in which we collected. The
Moengo Mine Road is 30 miles in length, run-
ning east and west, beginning at Moengo and
ending on the bank of theMarowyne (=Maroni)
River at Albina, directly opposite St. Laurent
on the French side. Our collecting was all done
along the central 10 miles of the Moengo-
Albina Road. The two collecting areas are thus
only a few miles apart, separated by merely a
change in direction and the width of the
Marowyne River.

Secondly, the identification of the insects by
Joicey & Kaye, insects with such a bewildering
maze of colors and patterns, was necessarily
based on individuals of unknown parentage or
inter-relationship. The result of this was a pleth-
ora of bi- and trinomial names divided into

139

140

Zoologica: New York Zoological Society

[40: 13

various categories, such as species, subspecies,
variety, form, type, stage and aberration.

From our studies of Heliconius behavior in
insectaries, we came to this aspect of extreme
variation convinced that the final solution de-
pended on one method of approach. This was
actually to breed the butterflies, and to deter-
mine the composition and possible variation in
breeds from individuaUy-known parents, both
if possible, but at least the females.

In 1953 our first attempt at breeding Surinam
heliconids failed, and only a single butterfly
lived to reach the laboratory at Simla, where
she survived only an hour. The elaborate notes
we made on the behavior and other phases of
ecology of these easily tamed butterflies more
than compensated for the one set-back.

In 1954, eleven days were spent in field work
in Surinam. In a few days we had 53 butterflies
in captivity, feeding and in health. Twelve of
these were females. Day by day, as they flour-
ished, they began to lay eggs, and we were ulti-
mately rewarded with 54. Every one of these
was correlated with an individual female, and
we knew even the sequence of deposition.

In the course of our experience we learned
many niceties of butterfly preferences and re-
quirements. They did not favor low-growing
egg-plants, but more readily accepted those at
higher levels. We found that by following them
around with small passion vines growing in pots,
they would often desist from their lantana
flower feeding and give their whole attention
to alighting on a tendril and depositing an egg.
They had a preferred time schedule, and 12
to 12:30 PM was the period most likely to pro-
duce results (cf. Seitz, 1913, p. 377). An ex-
cess number of individuals or a superabundance
of flower food were distinct deterrents to ovi-
position.

Obtaining a supply of new-laid fertilized but-
terfly eggs was one thing. Transporting them to
Simla, Trinidad, was quite another. One batch
was taken on board Alcoa’s river steamer, and
cared for throughout the three days’ trip to Port-
of-Spain. The second lot was flown direct from
Paramaribo to Trinidad, via the Dutch airline.
The latter had the best of it, for the flight lasted
only four hours. On the vessel which carried
some of us and our heavy luggage the eggs
began to hatch in our cabins on the first day
out.

Forty-five larvae were reared from six broods.
They are discussed in detail below.

Brood A
(Plate I)

Female Parent'. H. erato (Field No. 7);
Surinam. This individual bears a close resem-

blance to the 13 offspring of Surinam male No.
20 crossed with Trinidad female No. 45 (see
below) . It shares characters of androdaixa,
andremona and amazona (Seitz, plate 78d).
Wing spread 72 mm.

Offspring'. One of the eggs of this female
hatched into a larva which was reared. This
proved also to be a female, but details of the
caterpillar made it certain that both parent and
offspring are erato and not melpomene.

The single offspring approximates leda (Seitz,
plate 78d). The general color is not Scarlet-Red
but Flame-Scarlet (Ridgway). The pale fore-
wing markings are pale lemon yellow. They dif-
fer not only from Seitz’s illustration but, when
examined minutely, vary on the right and left
wings, and still more so on the underside of the
wings. Wing spread 65 mm.

Eight leda-Vike individuals taken in the same
locality on the expeditions were all different
from one another in minor details, as well as in
number and size of the various spots.

Brood B
(Plate II)

Parents: Male: H. erato (Field No. 20);
Surinam. Female: H. erato (Field No. 45);
Trinidad. The Surinam male parent (No. 20),
approximates H. amazona (Seitz, plate 78d)
both in general pattern and color. In the single,
pale, mid-forewing spot, and in the apparent
perforation of two of the hindwing markings,
it resembles andremona (plate 78d). It posses-
ses five instead of six radii. The color of the light
forewing spots is Pale Green-Yellow (Ridgway)
instead of Light Lemon Yellow. Wing spread
75 mm.

The character most consistently different from
Seitz’s figures, and present in both male parent
and all the offspring, is the large size and spindle
shape of the two outmost forewing spots. These
point toward and are closest to the mid outer
margin of the wing, and show little variation in
shape and position. They are decidedly unlike
the corresponding spots in Seitz’s androdaixa,
andremona and amazona.

The Trinidad female parent (No. 45) is a
typical Heliconius erato hydara Hewitson. In
Seitz it is closest to viculata (plate 78b), except
that the red band is solid, with the veins almost
obliterated. The shape of the various compo-
nents of the scarlet mosaic is similar to that of
colombina. Wing spread 61 mm.

Offspring: The caterpillars of all 13 offspring
were typical erato. The imagos have all wing
markings entirely scarlet, lacking yellow. In
pattern they closely approximate their male
parent, and although each show minor distin-
guishing characters, in pattern they come closest

1955]

Beebe: Polymorphism in Heliconius Butterflies

141

to andremona (Seitz, plate 78d). For example,
8 out of the 13 have the mid-forewing spot with
a trace of a double, and in 8 there are 6 instead
of five radii.

The wing spread varies from 55 to 69 mm.,
with an average of 63 mm. The forewing spots
which are Pale Green-Yellow in the male parent
are, on the underside of the forewing of the
offspring, varying shades of pale pink.

Brood C
(Plate III)

Female Parent: H. erato (Field No. 30);
Surinam. This is a black-hindwing, subsequently
proved erato, as close to H. viculata as anything
in Seitz (plate 78b) , sharing minor characters of
the red band with colombina and erythrea (plate
78b and c). Wing spread 72 mm.

Offspring:Fo\ix caterpillar offspring were typ-
ically erato, and the imagos reared from these
are divided sharply into two and two. The first
two are identical, black-hindwinged erato, and
the remaining two are rayed-hindwinged, but
differing strongly from each other. Offspring
One, Two and Four are females, while Three is a
definite erato male, confirming the identification
of the female parent.

One and Two are identical with their female
parent, closer in fact, even in minute differ-
ences of the red wing-band, than to any Seitz’s
figures. The wing spreads are, respectively, 63
and 68 mm.

Imago Three is closest to H. udalrica (Seitz,
plate 78c), but with the forewing pattern less
pronounced and more scattered. The color is
Scarlet instead of Apricot Orange. The basal,
under aspect of the hindwings shows the four
proximal scarlet dots as distinct, and forming
heads for four of the six hindwing radii. The
wing spread is 73 mm.

Offspring Four is smaller than Three, 60 as
compared with 73 mm. in wing extent, but the
red forewing pattern is much the same. The
most striking difference is in the presence of
three white spots which occupy the center of the
three proximal spots in the wing pattern. The
one along the costal margin of the wing is linear;
the next below, in the discal area, is crescentic,
and the third is a long oblong, occuyping the
central two-thirds of the spot on the midwing.

On the underside of the forewing the red is
subordinated and the white is dominant. As in
Offspring Three the basal scarlet spots are nearly
or quite separate.

Brood D
(Plate IV)

Female Parent: H. erato (Field No. 41) ;
Surinam. A large, black-hindwinged erato.

closest, in red band outline, to erythrea (Seitz,
plate 78c). Identical with erato female parent
No. 30. Identified as erato both by larvae and by
male offspring. Wing spread 80 mm.

Offspring: Thirteen out of 15 larvae reared,
with an average wing spread of 65 mm. These
offspring divide sharply into two groups, 8 black-
hindwings and 5 rayed-hindwings. In the ratio
of males to females, the blackwings show 5 to
3, the rayed wings 3 to 2. The order of egg de-
position reveals the following succession of
black and rayed individuals: R,R,B,R,R,B,B,R,
B,B,B,B,B,.

Considering the red wing band in the eight
black-hindwing offspring, no two are exactly
alike. As a whole they roughly divide into two
divisions. The first group consists of three indi-
viduals with forewing spots so tightly joined that
the veins are obliterated, exactly as in their fe-
male parent. In the remaining five the band is
very broad, irregular as to outline, and loose
and open in the joints, revealing the veins to a
less or greater extent.

In the case of the five rayed-hindwing off-
spring, the same general distinction holds, two
butterflies, like their maternal parent, showing
close-knit units of the red band, and three pre-
senting varying stages of disorganization and
dissolution of the vein-separated spots. None of
these can be correlated with any of Seitz’s plates.

Brood E
(Plate V)

Female Parent: H. erato (Field No. 51);
Surinam. A black-hindwinged erato, 82 mm.
wing spread, with red spot identical with that of
female parent No. 30.

Offspring: Four offspring, all black-hind-
winged, consisting of two males and two
females, are very close to their female parent
and to one another. The average wing spread
is 67 mm.

Brood F
(Plate VI)

Female Parent: H. melpomene (Field No.
49); Surinam. On April 17 a black-hindwinged,
female heliconid was taken at Moengo. It was
large, with a wing spread of 82 mm. The scarlet
wing band was unusually broad, and irregular
along the distal margin.

Offspring: Ten were successfully reared to
maturity. The larvae showed the color distinc-
tion of melpomene, not erato, and all, in black-
hindwing and general characters, resemble their
maternal parent. Eight of the offspring are
males, and two females. The male imagos con-
firm the larval identification of melpomene.

142

Zoologica: New York Zoological Society

[40: 13

Offspring Numbers One to Five were reared
on Surinam passiflora, and showed an average
wing spread of 77.4 mm. The larvae of Num-
bers Six to Ten were fed on passiflora from
Simla, Trinidad, and are uniformly smaller,
showing an average wing spread of 63.4 mm.
This dissimilarity in size bears no relation to
variation in relative shape or size of the scarlet
band. Were the 10 individuals not members of
the same brood, 4 of their band variations would
be considered worthy of some degree of recog-
nition.

Major characteristics of the six broods may
be summarized as follows, omitting lesser de-
tails of color and pattern:

Heliconius erato group

Brood A: Female parent: red radiations on
fore- and hindwings; broken fore-
wing red band. Male unknown.

Offspring: 1 is rayed like parent; fore-
wing band represented by small,
scattered, whitish spots.

Brood B: Male parent (Surinam): red radia-
tions on fore- and hindwings; fore-
wing band represented by a broken
area of large, scattered, creamy
spots.

Female parent: (Trinidad) : red fore-
wing band only.

Offspring: 13 with red radiations on
fore- and hindwings; broken, red,
forewing band.

Brood C: Female parent: red forewing band
only. Male unknown.

Offspring: 2 like parent.

I with broken, red, forewing band;
red radiations on fore- and hind-
wing.

1 with broken, red, forewing band
with whitish spots; red radiations
on fore- and hindwing.

Brood D : Female parent: red forewing band
only. Male unknown.

Offspring: 8 like female parent; red
forewing band varying from solid
to broken.

5 with red variations on fore- and
hindwings; red fore wing band vary-
ing from solid to broken.

Brood E: Female parent: red fore wing band
only. Male unknown.

Offspring: 4 like female parent.

Heliconius melpomene group

Brood F : Female parent: red forewing band
only. Male unknown.

Offspring: 10 like female parent, with
much variation in red band.

Summary

Six broods of heliconid butterflies were reared
from parents taken at Moengo, Surinam. The
one exception was a brood of 13, with the fol-
lowing parents: a male Surinam Heliconius
erato amazona, mated to a female Trinidad
Heliconius erato hydara.

Four broods of 1, 4, 4 and 13, had typically
black-hindwinged, Heliconius erato, Surinam,
female parents. The sixth brood of 10 had a
Heliconius melpomene, Surinam, female parent.

The distinction between Heliconius erato and
melpomene was established by means of differ-
ences in the larvae, as well as by the scent scales
of the male offspring.

The offspring frequently differed radically
both from the parent and from one another.
These differences were tentatively correlated
with illustrations in A. Seitz, “Macrolepidop-
tera of the World,” Vol. 5, plates: “The Amer-
ican Rhopalocera,” plate 78. New sibling rela-
tionships were therefore established for types of
individual patterns and colors, to which have
heretofore been applied terms such as species,
subspecies, variations, forms, types, stages and
aberrations.

Further data and interpretations are antici-
pated in cross-breeding future broods, resulting
in Fa generations.

References

JoiCEY, J. J. & W. J. Kaye

1916. On a collection of Heliconine forms from
French Guiana. Trans. Ent. Soc. London,
412-531.

Ridgway, R.

1912. Color standards and color nomenclature.
Publ. by the author. Washington, D. C.

Seitz, A.

1913. Heliconiinae. Macrolepidoptera of the
World, Stuttgart, 5: 375-395; pi. 78.

1955]

Beebe: Polymorphism in Heliconius Butterflies

143

EXPLANATION OF THE PLATES

The light wing markings on the figures of aU the
plates are red or scarlet, with the following excep-
tions:

Plate I. Lower Figure: Pale spots on forewing
are Light Lemon Yellow (Ridgway).

Plate II. Upper Left Figure: Forewing spots are
Pale Green-Yellow (Ridgway).

Plate III. Lower Right Figure: The three pairs
of light spots on forewing are white.

Plate I

Brood A. Heliconius erato group

Upper Figure. Female parent.

Lower Figure. Single offspring.

Plate II

Brood B. Heliconius erato group

Upper Left Figure. Male parent.

Upper Right Figure. Female parent.

Lower Figures. Thirteen offspring.

Plate III.

Brood C. Heliconius erato group
Upper Figure. Female parent.

Lower Figures. Four offspring.

Plate IV

Brood D. Heliconius erato group
Upper Figure. Female parent.

Lower Figures. Thirteen offspring.

Plate V

Brood E. Heliconius erato group
Upper Figure. Female parent.

Lower Figures. Four offspring.

Plate VI

Brood F. Heliconius melpomene group
Upper Figure. Female parent.

Lower Figures. Ten offspring.

BEEBE

PLATE I

POLYMORPHISM IN REARED BROODS OF HELICONIUS BUTTERFLIES
FROM SURINAM AND TRINIDAD

BEEBE

PLATE II

POLYMORPHISM IN REARED BROODS OF HELICONIUS BUTTERFLIES
FROM SURINAM AND TRINIDAD

BEEBE

PLATE III

POLYMORPHISM IN REARED BROODS OF HELICONIUS BUTTERFLIES
FROM SURINAM AND TRINIDAD

BEEBE

PLATE IV

POLYMORPHISM IN REARED BROODS OF HELICONIUS BUTTERFLIES
FROM SURINAM AND TRINIDAD

BEEBE

PLATE V

POLYMORPHISM IN REARED BROODS OF HELICONIUS BUTTERFLIES
FROM SURINAM AND TRINIDAD

BEEBE

PLATE VI

POLYMORPHISM IN REARED BROODS OF HELICONIUS BUTTERFLIES
FROM SURINAM AND TRINIDAD

14

Formation of a Mucous Envelope at Night by
Parrot Fishes

Howard Elliott Winn

The American Museum of Natural History, New York^

(Plate I)

Introduction

IT is well known that many parrot fishes
rest on the bottom at night while leaning
against various objects such as rocks, coral
or shells. During the summer of 1954 at the
Lerner Marine Laboratory, Bimini, B. W. I.,
several species of these fishes were observed in
such positions at night with a large and con-
spicuous mucous fold around their bodies (Plate
I) . The formation of this envelope apparently
represents a specialized function of the mucus-
secreting system. When the parrot fishes, sur-
rounded by mucus, were first observed in an
aquarium at night, it was thought that this condi-
tion was pathological, but their respiration and
reactions appeared normal. After repeated ob-
servations of this envelope formation, it became
clear that this was normal behavior. Observa-
tions were made with a flashlight, and the trans-
parent folds were difficult to see until the light
rays were directed at an angle which made them
more easily visible.

C. M. Breder, Jr., Priscilla Rasquin, Louise
Stoll and Carolyn Winn kindly read and made
suggestions on the manuscript, which are grate-
fully acknowledged.

Observations

Five species of parrot fishes were placed in
laboratory aquaria. Four of these. Scams
croicensis Bloch, S. punctulatus Cuvier & Valen-
ciennes, Pseudoscarus guacamaia (Cuvier), and
Sparisoma pachycephalum Longley,^ formed
conspicuous mucous folds only in the dark, ex-
cept under certain special conditions. The no-
menclature is that used by Longley & Hildebrand

^Present address: Department of Zoology, University
of Maryland, College Park, Md.

“The identification of the individual used, which was
6 cms. in standard length, is only tentative.

(1941). Five individuals of S. croicensis were
also observed at night in dead finger coral
{Porites sp.) branches beside the laboratory
dock by means of goggles and an under-
water flashlight. All were surrounded by a
mucous sheath. One species, Sparisoma chrysop-
terum (Bloch & Schneider), did not form
such a structure in aquaria or in the water off
the dock, where two individuals were observed.
The standard length of the various fish studied
varied from about 6 to 21 cms. The water tem-
perature in the aquaria was 30 to 32° C.

The structure of the mucous envelope was
similar in the four species that produced it. It
started as a fold at the mouth and went back-
ward to surround the body of the fish completely.
A little flap with a hole in its center covered
the open mouth and moved in and out as the
fish breathed. Posterior to the caudal fin, an
opening of one to several centimeters in di-
ameter was present through which the expired
water left. Thus, a flow of water over the gills
was insured. As Plate I shows, the folds were
complex and extended up to several centimeters
away from the body, depending on the size of
the individual. The envelope seemingly con-
sisted of a series of layers of mucus and varied
in shape to some extent. In a specimen of S.
punctulatus, about 18 cms. in standard length,
the maximum length, width and depth of the
envelope was 25 X 13 X 9 cms. In some in-
stances the mucus near the bottom had sand
grains attached to it and silt particles often set-
tled on the upper part, which made the envelope
more readily observable.

The fish usually leaned against the aquarium
wall or drainpipe. However, they swam into
conch shells or in among the branches of finger
coral which they used for support if available.
The mucus was then secreted in these positions.

145

146

Zoologica: New York Zoological Society

[40: 14

In one instance, two S. croicensis leaned against
each other, one being against the aquarium wall,
and formed what appeared to be a single fold
around both of them.

The mucus was transparent and gelatinous.
Large amounts could easily be picked up in the
hand. Sometimes when a fish moved out of the
fold, the structure partially collapsed into a
large ball (Plate I, Fig. 2), but at other times
it remained temporarily in the expanded con-
dition.

The exact formation of the envelope was not
observed because a light could only be turned
on for a few seconds every five minutes or so.
If the light was left on, the fish stopped the se-
cretion and soon became active enough to break
out of the fold. It appeared to form &st around
the head region and then pass back over the
body, but it was not determined what groups of
mucus cells were involved.

The consistency of the formation of the struc-
ture was established by nightly observation for
the different fish as follows: four individuals of
5. croicensis and one of S. punctulatus from July
19 to 28; five S. croicensis from August 17 to
27; one S. punctulatus from August 18 to 23;
and one P. guacamaia from August 21 to 27.
Other individuals were watched only occasion-
ally.

Certain differences were noted in the time
that some species required to form the com-
pleted envelope in the dark and the time re-
quired to break out of the folds after the lights
were turned on. Although aU the species were
not studied comparatively, it was demonstrated
that S. croicensis took longer to form the folds
and broke out more quickly than either S. punc-
tulatus or P. guacamaia. The individuals of S.
croicensis normally took more than one hour
to complete the fold. Under 0.3 footcandles of
light, one individual formed it in 80 minutes and
another between 75 and 145 minutes. Under
these conditions the other two species completed
the envelope usually within 30 minutes (P.
guacamaia in 20 mins., S. punctulatus in 25 and
30 mins.). When the lights were turned on, S.
croicensis usually broke out in less than 30
minutes (5, 10, 25 and 28 mins.), whereas the
other two species usually required a longer
period (P. guacamaia more than 30 mins.). It
should be noted, however, that two specimens of
S. punctulatus and P. guacamaia were consider-
ably larger than those of S. croicensis.

At night the respiratory rate of the parrot
fishes was considerably reduced. One S. punctu-
latus inspired 76 and 73 times per minute during
the day and 47 times at night. One S. croicensis
inspired 100 and 144 times per minute during
the day and 56 times at night.

In several instances, under apparently anoxic
conditions, some of the parrot fish formed the
mucous envelope around their bodies in day-
light. One specimen of S. pachycephalum and
three of S. croicensis were observed to do this
in aquaria where the running water supply had
shut off. The water had been turned off about
an hour and the temperature of the water had
increased in two tanks each containing two fish.
One specimen of P. guacamaia (about 21 cms.
st. 1.), presumably in a state of anoxia in a
bucket, formed the envelope. In at least one
instance, three individuals of S. croicensis, not
under anoxic conditions but wedged in finger
coral, formed the structure in daylight. It was
also usually produced in tanks darkened during
the day, although the formation time was about
doubled.

When a light is turned on after the slime
covering is formed, the fish breaks out in such
a manner as to indicate that the envelope is suf-
ficiently resistant to modify the first swimming
movements. In 4 out of 7 instances, parrot fishes
moved backwards out of the mucous fold. Twice
they wiggled out sideways and once one swam
forward out of it.

In one instance, the mucus at the front of a
parrot fish was separated from the mouth. The
fish then moved forward slightly so that the
mucus again came in contact with the mouth
and a new hole was made for breathing.

Discussion and Summary

The formation by parrot fishes of a large
mucous envelope after dark is a remarkable
behavioral trait that evidently has not been pre-
viously recorded. It has been observed both in
aquaria and in the field. Four species appar-
ently do this normally every night, whereas
Scams brachiale does not form the envelope.
Most fishes are covered with a thin coat of
mucus and its secretion is a slow, possibly con-
tinuous process. The behavior recorded here is
a divergence from this pattern and seems to
represent a rare specialization of the mucus-
secreting system. More refined chemical differ-
ences may be present, but these have not been
adequately investigated.

The structure has been described in the pre-
vious section. The fish utilize any object to swim
into or lean against at night before the mucus is
secreted. Precisely how the envelope is formed
has not been determined. At the resumption
of daytime activities the fish’s behavior is modi-
fied in that the fish has to back out or wiggle
vigorously forward to rid itself of the mucous
coat. Preliminary data show that S. croicensis
takes longer to form the structure and takes
less time to break out than either S. punctulatus

1955]

Winn: Formation of Mucous Envelope by Parrot Fishes

147

or Pseudoscarus guacamaia. This seems to in-
dicate that the latter two species have a higher
threshold of reactivity to light stimuli under
these conditions. The ecological significance of
these differences is unknown. Two questions that
arise are the nature of the physiological mechan-
isms involved and the function of this conspicu-
ous mucous envelope.

Reid (1894) and Uhlich (1937) indicated
that there is a direct reaction to touch or chemi-
cal stimulation of the mucus cells and possibly
a nervous control. However, as they indicate,
their experiments of direct stimulation of intact
and excized skin do not seem to present critical
evidence bearing on the latter conclusion. The
reactions of parrot fishes to light and anoxia
strongly suggest that there is a nervous control,
at least in these fishes, although the possibility
of hormonal influence cannot be eliminated.
The lack of light, which is usually necessary for
the formation of the envelope, leads one to be-
lieve that the stimulus is mediated through the
eyes to the nervous system and finally to the
mucus cells. However, it appears that anoxic
conditions produce the same effect and pre-
sumably stimulate the respiratory system which
relays to the nervous system. When in the dark,
the fish’s respiration is considerably lowered so
that there may not be transmission through the
optic system to the mucus cells, but instead the

stimulus may be mediated by the respiratory
system because of the reduced oxygen supply as
under anoxic conditions.

The function of the mucous envelope is not
clear and can only be speculated upon at this
time. The envelope may protect the fish from
nocturnal predators, since it would be particu-
larly vulnerable when lying on the bottom. Pro-
tection against the settling of silt is one of several
other possibilities. The nature of the mucous
secretion in the parrot fishes may make it pos-
sible to test some of these supposed functions in
future experiments.

Literature Cited

Longley, William H. & Samuel F. Hildebrand

1941. Systematic catalogue of the fishes of Tor-
tugas, Florida, with observations on color,
habits and local distribution. Carnegie
Inst. Wash., Publ. 535 (Papers from Tor-
tugas Lab. 34): 1-331.

Reid, E. W.

1894. The process of secretion in the skin of the
common eel. Phil. Trans. Roy. Soc. Lon-
don (B), 185 (1): 319-354.

Uhlich, H.

1937. Uber die Schleimsekretion bei Fischen.
Zool. Jahrbuch, 57 (4): 417-456.

148

Zoologica: New York Zoological Society

[40: 14: 1955]

EXPLANATION OF THE PLATE

Fig. 1. Large mucous envelope around a rainbow
parrot fish (Pseudoscarus guacamaia), stand-
ard length, 21 cms.

Fig. 2. On the left, mucous fold around a parrot fish,
Scarus croicensis. On the right, another in-
dividual has freed itself from the mucous
fold, which remains as a nearby oval mass
of slime.

WINN

PLATE I

FIG. 1

FIG. 2

FORMATION OF A MUCOUS ENVELOPE AT NIGHT BY PARROT FISHES

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

30 East Fortieth Street, New York 16, N. Y.

PUBLICATION OFFICE

The Zoological Park, New York 60, N. Y.

OFFICERS

PRESIDENT VICE-PRESIDENTS TREASURER

Fairfield Osbom Alfred Ely David H. McAlpin

Laurance S. Rockefeller
Donald T. Carlisle

SECRETARY ASSISTANT TREASURER

Harold J. O’Connell Percy Chubb, 2nd

SCIENTIFIC STAFF: Zoological Park and Aquarium

John Tee- Van Director

Leonard J. Goss Assistant Director

ZOOLOGICAL PARK

Richard H. Manville . Curator of Mammals

Grace Davall Assistant Curator,

Mammals and Birds

James A, Oliver Curator of Reptiles

Leonard J. Goss Veterinarian

Charles P. Gandal. . .Assistant Veterinarian

Lee S. Crandall General Curator

Emeritus

William Beebe Honorary Curator,

Birds

AQUARIUM

Christopher W. Coates . Curator & Aquarist

James W. Atz Assistant Curator

Ross F. Nigrelli Pathologist

Myron Gordon Geneticist

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . .Research Associate

in Histology

Homer W. Smith Research Associate

in Physiology

GENERAL

William Bridges Editor & Curator,

Publications

Sam Dunton Photographer

Henry M. Lester. . .Photographic Consultant

DEPARTMENT OF TROPICAL
RESEARCH

William Beebe Director Emeritus

Jocelyn Crane Assistant Director

Henry Fleming ....... Entomologist

Rosemary Kenedy Field Assistant

J ohn Tee- Van A ssociate

William K. Gregory Associate

AFFILIATES

C. R. Carpenter Co-ordinator, Animal

Behavior Research Programs

L. Floyd Clarke Curator,

Jackson Hole Research Station

SCIENTIFIC ADVISORY COUNCIL
A. Raymond Dochez Caryl P. Haskins
Alfred E. Emerson K. S. Lashley
W. A. Hagan John S. Nicholas

EDITORIAL COMMITTEE

Fairfield Osbom, Chairman
James W. Atz Lee S. Crandall

William Beebe Leonard J. Goss

William Bridges James A. Oliver

Christopher W. Coates John Tee- Van

200L0GICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 40 • PART 4 • DECEMBER 31, 1955 • NUMBERS 15 TO 17

PUBLISHED BY THE SOCIETY
Hhe ZOOLOGICAL PARK, New York

Contents I

k

PAGE V

15. A Revision of the Surgeon Fish Genus Ctenochaetus, Family Acanthiiri- |

dae, with Descriptions of Five New Species. By John E. Randall. Plates |
I & II; Text-figures 1-3 149 |

16. Imaginal Behavior of a Trinidad Butterfly, Heliconius erato hydara Hewit- >

son, with Special Reference to the Social Use of Color. By Jocelyn Crane. |

Plates I-III; Text-figures 1 & 2 167

17. Three More Gymnotid Eels Found to Be Electrogenic. By Christopher

W. Coates. Plates I-III 197 |

15

A Revision of the Surgeon Fish Genus Ctenochaetus,
Family Acanthuridae, with Descriptions of
Five New Species^

John E. Randall

University of Hawaii, Honolulu 14, T.H.

(Plates I & II; Text-figures 1-3)

The genus Ctenochaetus is distinguished
from other surgeon fish genera chiefly
by its unique teeth, which are numerous,
loosely attached in the jaws and very elongate,
with expanded incurved tips. Little difficulty is
experienced in assigning specimens to this
genus; however, insufficient attention has been
given to identification at the species level. Two
specific names, striatus Quoy & Gaimard and
strigosus Bennett, have been applied indiscrim-
inately. Generally only one or the other name is
used and some authors have asserted their be-
lief that the genus is monotypic. In this paper
the validity of C. striatus and C. strigosus is
established and five other species are described
as new.

Several species of Ctenochaetus are very
abundant in the tropical inshore waters of the
Pacific and Indian Oceans and the genus is
extensively recorded in the literature. Many
records are listings of names only and others
contain too little descriptive information to pro-
vide identification. Such records have been
omitted, generally, unless the specimens re-
ported on were examined by me.

Drawings of adults of only two of the seven
species are given. The others have either been
pictured previously or present insufficient dif-
ferences in external appearance from remaining
species to warrant the preparation of figures.
The shape of the teeth (principally the number
of denticulations on the lateral edge of the ex-
panded tips) has been found to be the most
important basis for the separation of species.

1 Contribution No. 58, Hawaii Marine Laboratory,
in co-operation with the Department of Zoology and
Entomology, University of Hawaii.

Text-fig. 1 consists of drawings of the teeth of all
seven species. The number of teeth (Text-fig. 2)
and fin ray and gill raker counts (Tables 1-3)
are also of diagnostic value.

It is with pleasure that I acknowledge the
guidance and assistance given by Leonard P.
Schultz, Ernest A. Lachner and Robert H.
Kanazawa of the United States National
Museum, where much of the work on this paper
was done. William A. Gosline of the University
of Hawaii provided helpful counsel, and thanks
are due Boyd W. Walker of the University of
California at Los Angeles and George S. Myers
of the Stanford Natural History Museum for the
loan of specimens.

The types of the new species are deposited in
the United States National Museum.

Methods of Counting and Measuring

Each ray of the dorsal and anal fins with a
distinct base was counted regardless of how
close adjacent rays might be. In cases where two
rays branch from a common base, they were
counted as one. At times dissection was nec-
essary to determine whether the last two rays
originate from a single basal element. Pectoral
fin ray counts include the two uppermost un-
branched rays, the first being a very short bony
spicule. Few gill raker counts were made, be-
cause it is difficult to remove a complete gill arch
from an acanthurid without damaging the speci-
men. The gill rakers are small and occur in two
distinct series, anterior and posterior (actually
more medial on the arch than posterior). All
rakers (including rudiments) in each series were
counted separately. In many instances it was
not possible to determine which raker occupied
the position at the angle of the arch, therefore

JWII0I956

149

150

Zoologica: New York Zoological Society

[40: 15

Text-fig. 1. Camera lucida drawings of the teeth of Ctenochaetus. A. striatus, 141 mm. speci-
men, Marshall Islands. B. strigosus, 1 19 mm. specimen, Hawaiian Islands. C. cyanoguttatus, type.
D. hawaiiensis, type. E. magnus, type. F. lominiensis, type. G. binotatus, type. 1. lateral view
of upper tooth. 2. inner view of end of upper tooth. 3. lateral view of lower tooth. 4. inner view
of end of lower tooth. All of the teeth were taken from the left side of the jaw near the center
of the mouth.

only the total count is recorded. Occasional
specimens showed a fusion of rakers on the
dorsal end of the medial series where the rakers
normally become broad, low ridges; no counts
from such specimens are included in the table.
At least for C. striatus, no obvious increase in
gill raker counts occurs with increasing size
after the fish have completely transformed from
the late postlarval or acronurus stage. The
state of being completely transformed is easily
distinguished by the presence of fully formed
scales on the body.

The standard length was measured from the
tip of the snout to the posterior end of the
hypural plate. Head length was taken from the
tip of the snout to the most posterior part of the
opercular membrane. Body depth is the distance
from the natural groove at the base of the sec-
ond anal spine to a similar groove at the base of
the dorsal fin. Caudal concavity is the distance
between vertical lines passing through the tips
of the shortest middle caudal ray and the longest
ray of the dorsal lobe of the caudal fin. Propor-
tional measurements are based on specimens
above 70 mm. in standard length.

Counts of the number of teeth were not made

in specimens with tightly closed or damaged
jaws. Even the best specimens generally had
missing teeth; however, these were detected by
empty sockets in the gums and by gaps in the
series and were included in the total counts.

The number of teeth was found to increase
with increasing size in striatus and strigosus.
Text-fig. 2 represents a graph of the relation-
ship between the number of teeth and the stand-
ard length of these two species. Such data from
the available specimens of the other five species
are also included in the graph, but there are not
enough of these of different lengths to clearly
show this relationship.

Scale counts are difficult to make, for the
scales are small and not in even rows. Approxi-
mate scale counts are recorded as the number
of vertical scale rows from the upper end of
the gill opening to the posterior end of the
caudal peduncle spine.

In the description of new species, data in
parentheses are referable to paratypes. Propor-
tional measurements and counts of scales were
made on five paratypes, if available, including
the largest and smallest specimens. Fin ray
counts are based on all paratypes. Tooth counts

1955]

Randall: Revision of CtenochaeUis; Five New Species

151

o

<\)

'■j

o

oj

§

o

e',

o

CD

o

vs

§

o

§

o

e

o

VT)

O

CD

O

V9

standard LENC5TH Cmrn)
Text-fig. 2. Tooth counts of Ctenochaetiis.

152

Zoologica: New York Zoological Society

[40: 15

and gill raker counts of paratypes may be found
in Text-fig. 2 and Table 3, respectively.

Genus Ctenochaetus Gill
Ctenodon Klunzinger, Synopsis der Fische des
Rothen Meers, 1870; 509 (proposed as a
subgenus; preoccupied by Ctenodon Wagler
1830).

Ctenochaetus Gill, Proc. U. S. Nat. Mus., 7;
279. 1885 (Type species, Acanthurus strigo-
sus Bennett).

Quoy & Gaimard (1824) described and
figured the first species of this genus as Acan-
thurus striatus. They remarked on its unusual
dentition. E. T. Bennett (1828) named another,
Acanthurus strigosus, from Hawaii, and he ac-
curately described the teeth. Although Gunther
(1861) separated two of the species, strigosus
and ctenodon Cuvier & Valenciennes, in a key
on the basis of dentition, he left them in
Acanthurus. Gill (1884) established the genus
Ctenochaetus, designating A. strigosus Bennett

Table 1. Counts of the Soft Rays of the Dorsal and Anal Fins for Species of Ctenochaetus

Species and locality

Dorsal soft rays

Anal soft rays

24 25 26 27 28 29 30 31

21 22 23 24 25 26 27 28

striatus

Wake I.

3 6 4

2 6 4 1

Marshall Is.

1 14 39 8

1 3 34 23 1

Gilbert Is.

1 9 13 4 1

1 5 21 1

Line Is.

2 1 1

1 2 1

Phoenix Is.

2 2 1

1 3 1

Samoa Is.

4 9 11

7 7 1

Tuamotu Arch.

2 7 2

6 5

Society Is.

3 4

1 1 5

Austral Is.

1

1

Solomon Is.

2 1

3

New Hebrides

1 1

1 1

Palau Is.

1 1

1 1

Mariana Is.

5 8 11 4

1 5 16 6

Philippine Is.

12 24 14

3 23 22 2

Formosa

2

2

East Indies

2 2 3

2 4 1

Egypt, Red Sea

3 2

4 1

Mauritius

2 2 2 1

3 2 2

Madagascar

1

1

strigosus

Hawaiian Is.

4 19 23 1

9 32 6

Johnston I.

8 1

2 6 1

Tuamotu Arch.

4

1111

Philippine Is.

3 6 4

7 6

& East Indies

Mauritius

1

1

cyanoguttatus, sp. nov.

Gilbert Is.

2

2

Phoenix Is.

1

1

Cocos I.

1

1

hawaiiensis, sp. nov.

Hawaii

1 3

3 1

magnus, sp. nov.

Malden & Jarvis I.

3

2 1

Cocos I.

1

1

tominiensis, sp. nov.

Celebes

1 5

4 2

binotatus, sp. nov.

Philippine Is.

2 4 16 8

1 9 17 3

& Molucca Is.

1955]

Randall: Revision of Ctenochaetus; Five New Species

153

Table 2. Counts of Pectoral Fin Rays for the
Species of Ctenochaetus

Species and locality

Pectoral
fin rays
15 16 17

striatus

Marshall, Mariana, and Samoa Is.

6 29

Philippine Is. and East Indies

15 24

Egypt, Red Sea

3 4

Mauritius

3 1

strigosus

Hawaiian Is.

1 22

Philippine Is.

13

Mauritius

1

cyanoguttatus, sp. nov.

Gilbert, Phoenix, and Cocos Is.

1 3

hawaiiensis, sp. nov.

Hawaii

4

magnus, sp. nov.

Malden, Jarvis, and Cocos 1.

1 3

tominiensis, sp. nov.

Celebes

2 4

binotatus, sp. nov.

Philippine and Molucca Is.

3 27

as the type species. He was apparently unaware
of A. striatus Quoy & Gaimard. Prior to Gill,
Klunzinger (1870) set up Ctenodon as a sub-
genus for strigosus and ctenodon. This was not
the Ctenodon of Swainson (1839), the type
species of which was Acanthurus sohal (For-
skal), and possibly not of Bonaparte (1833)
(for which there was no description nor type
specified). However, Wagler’s use in 1830 of
Ctenodon for a reptile invalidates subsequent
generic use of this name.

Ctenochaetus is characterized as follows:
body compressed, elliptical, the depth contained

1.7 to 2.2 in standard length; head length 3 to

3.8 in standard length; caudal peduncle armed
on each side by a single, sharp, folding spine;
length of caudal peduncle spine 2.3 to 4.3 in
head length; least depth of caudal peduncle 1.9
to 2.7 in head length; mouth small, terminal, and
only slightly protractile; jaws equal; teeth nu-
merous, in a single series in each jaw, movable,
elongate, with tips expanded, incurved, and
denticulate on the lateral margin; eye diam-
eter contained about 3 to 5 times in head
length; interorbital space 2.6 to 3.3 in length of
head; gill openings restricted to the sides; gill
membranes confluent and attached to the isth-
mus; scales ctenoid, very small, and not in
regular rows; head scaled, though not conspic-
uously; lateral line complete; dorsal fin single,
continuous, unnotched, with 8 spines, the first

small and easily overlooked; dorsal soft rays 24
to 3 1 ; anal fin with 3 spines, the first very short,
and 21 to 28 soft rays; the first 2 to 6 dorsal and
anal soft rays unbranched; pectoral fins long,
their length contained 2.7 to 3.2 in standard
length; pectoral rays 15 to 17; pelvic fins I, 5,
relatively long and pointed, and only slightly
posterior in their insertion to the base of the
pectoral fins; caudal fin with 16 principal rays
and varying in shape from lunate to nearly
truncate; 22 vertebrae; stomach subspherical to
oval with very thick walls.

The single erectile spine on each side of the
caudal peduncle is a prominent feature of four
of the acanthurid genera, Acanthurus, Cteno-
chaetus, Paracanthurus and Zebrasoma. The
latter genus differs from the other three in hav-
ing 4 or 5 dorsal spines instead of 7 to 9. The
monotypic Paracanthurus is readily separable
with its pelvic count of I, 3 and its remarkable
coloration. Ctenochaetus, as noted, is distin-
guished primarily by its unusual dentition.
Usually the genus has 8 dorsal spines (in counts
of over 300 specimens two had 9 and three had
7 dorsal spines). In the 28 species of Indo-
Pacific Acanthurus the usual count of dorsal
spines is 9. Two species, A. pyroferus Kittlitz
and A. sohal (Forskal), have 8 and another, A.
nubilus (Fowler & Bean), has 6 or 7.

Ctenochaetus striatus and certain other spe-
cies of this genus are similar in general body
shape and color pattern to Acanthurus nigroris
Cuvier & Valenciennes (.—Hepatus atramen-
tatiis Jordan & Evermann, 1905, and Acan-
thurus lineolatus Gunther, 1873). The simi-
larity may be seen even in the transforming
acronurus stages of these fishes by comparing
Fig. A (C. striatus) and Fig. C (A. nigroris) in
Plate 62 of Schultz & Woods (1953). Other
Acanthurus species such as lineatus and sohal
may be striped in the postacronurus form, but
the stripes do not angle downwards as they
pass posteriorly on the body. Inconsistent with
the apparent close relationship of Ctenochaetus
to A. nigroris is the finding that the roundish,
heavy-walled stomach of the former does not
resemble the elongate, thin-walled stomach of
A. nigroris but rather that of such acanthurids
as A. olivaceus and A. gahhm.

Although no existing Acanthurus species may
be considered the direct progenitor of the genus
Ctenochaetus, it seems reasonable to conclude
(as has Aoyagi, 1943) that Ctenochaetus was
derived from Acanthurus-\\\iQ stock. This view
is supported by the fact that postlarval Cteno-
chaetus have teeth much like those of postlarval
Acanthurus, and these transform into the typical
adult Ctenochaetus dentition (see Text-fig. 3
and discussion under striatus) ,

154

Zoologica: New York Zoological Society

[40: 15

A B C

Text-fig. 3. A. Upper tooth of 27 mm. acronurus larva of Ctenochaetus
strigosus. B to D. Upper teeth from 31 mm. postacronurus stage of
Ctenochaetus striatus. Dotted lines indicate margins of pulp cavity.

The food habits and method of feeding of
Ctenochaetus strigosus were investigated. The
stomach contents of seven adult specimens from
different localities in the Hawaiian Islands were
analyzed. All but one fish contained a very
large amount (up to 90%) of fine inorganic
sediment; the remaining detrital material was
mostly algal, consisting of diatoms and small
fragments of many kinds of red, green and blue-
green algae. About 1 to 2 per cent was soft,
unidentifiable organic matter. There were oc-
casional tiny molluscs and crustaceans, sponge
spicules, holothurian plates, pedicellaria frag-
ments, etc. The stomach of one specimen was
filled primarily with a fine red alga (Ceramium
sp.), though there was still a large amount of
sediment. When a thallus of fine filamentous
red algae {Polysiphonia sp.) was placed in an
aquarium containing two adult specimens of
strigosus, the fish attempted to feed upon it.
Their slender movable teeth, not able to ef-
fectively bite off pieces, soon became tangled in
the algae, resulting in very little being ingested.
When fine particles of the alga were put in the
tank and allowed to settle, the fish fed in the
following manner: the body was elevated to a
near-vertical position about 15 mm. above the
bottom, there was a pause, then the fish pecked
at a small area, the teeth and lips scraping over
the surface. Such an area was not only cleaned
of particulate algae but also of very fine sedi-
ment that had collected there, suggesting that
a suction mechanism is involved as well as a
scraping one. There was definitely no lateral
plowing or sieving action by the teeth as their
comb-like structure might suggest. Sand on the
bottom was generally avoided, but if picked
up, most was usually forcefully ejected.

Key to the Species of Ctenochaetus
la. No prominent blackish spot at base of last
3 to 7 rays of both the dorsal and anal fins.
(Juvenile and young adult striatus have a
small black spot basally at rear of dorsal
fin only).

2a. Teeth of the upper jaw with 4 to 7 den-
ticulations (including tip) on lateral
edge of their distal expanded ends.

3a. Teeth of the upper jaw with 5 to 7
denticulations on lateral edge of
their distal expanded ends; body
with numerous pale longitudinal
stripes (often faint or not visible in
preserved specimens) ; spots, if pres-
ent, occur only on head or anterior
part of body; inter-radial mem-
branes of pectoral fin hyaline;
length of longest dorsal ray con-
tained 3.6 to 4.4 times in standard
length.

4a. Teeth of upper jaw with 6 den-
ticulations (rarely 5 or 7) ; teeth
of lower jaw with 4 denticula-
tions (including tip) ; caudal fin
lunate, caudal concavity (see
section on methods of measuring
and counting) contained 3.7 to
6 times in standard length; body
depth contained 1.9 to 2.3 in
standard length; dorsal fin rays
VIII, 27 to 31 (usually 28 to 30) ;
anal fin rays III, 24 to 28 (usu-
ally 25 to 27)

Ctenochaetus striatus (Quoy &
Gaimard) .

4b. Teeth of upper jaw with 5 dentic-
ulations; teeth of lower jaw with

1955]

Randall: Revision of Ctenochaetus; Five New Species

155

3 denticulations (including tip);
caudal fin moderately concave,
caudal concavity contained 5.7
to 10 times in standard length;
body depth contained 1.7 to 2.0
in standard length; dorsal fin
rays VIII, 25 to 28 (usually 26
or 27); anal fin rays III, 21 to

25 (usually 23 or 24)

Ctenochaetus strigosus (Ben-
nett) .

3b. Teeth of upper jaw with 4 denticula-
tions on lateral edge of their distal
expanded ends; body without stripes,
when alive speckled with numerous
bright blue spots which may or may
not persist as pale spots in preserved
specimens; inter-radial membranes
of pectoral fin dark brown; length of
longest dorsal ray contained about

5.2 times in standard length

Ctenochaetus cyanoguttatus, sp. nov.

2b. Teeth of upper jaw with 3 denticulations
(including tip) on lateral edge of then-
distal expanded ends.

5a. Ratio of number of teeth in
lower jaw to number in upper
jaw about 2:1; caudal fin
slightly emarginate, caudal
concavity contained 1 8 to 40
times in standard length;
longest dorsal ray contained
4 to 5 times in standard
length; inner surfaces of lips
plicate, margins crenulate;
distance from base of upper
lip to distal end of upper
teeth contained 3.1 to 3.7
times in head length; length
of snout contained 3.6 to 3.9

in standard length

Ctenochaetus hawaiiensis, sp.
nov. (Plate II, Fig. 2)

5b. Ratio of number of teeth in
lower jaw to number in upper
jaw about 1.2:1; caudal fin
lunate, caudal concavity con-
tained 6 to 7 times in stand-
ard length; longest dorsal
ray contained 5.5 to 6 times
in standard length; inner sur-
faces and margins of lips
smooth; distance from base
of upper lip to distal end of
upper teeth contained 4.5 to
5.3 times in head length;
length of snout contained 4. 1

to 4.8 in standard length ....
Ctenochaetus magnus, sp.
nov.

lb. A prominent blackish spot at base of last 3
to 7 rays of both the dorsal and anal fins,
these spots extending narrowly on adjacent
regions of caudal peduncle.

6a. Membranes of caudal fin
and posterior parts of dor-
sal and anal fins pale;
margins of lips papillate;
enlarged distal curved
portions of each tooth of
upper jaw with lower half
smooth and blade-like
and upper half with 3
(rarely 2) lateral denticu-
lations . . . Ctenochaetus
tominiensis, sp. nov.

6b. Membranes of caudal,
dorsal, and anal fins dark
brown; margins of lips
smooth; enlarged distal
curved portion of each
tooth of upper jaw divid-
ed into 6 approximately
equal lateral denticula-
tions .... Ctenochaetus
binotatus, sp. nov.

Ctenochaetus striatus (Quoy & Gaimard)

Text-fig. 1 A; Plate I, D-F; Text-fig. 3 B-D

Acanthunis argenteus Quoy & Gaimard, Voy.
Uranie, Zool.: 372, pi. 63, fig. 2. 1824
(Guam); Cuvier & Valenciennes, Hist. Nat.
Poiss., 10: 239. 1835.

Acanthunis striatus Quoy & Gaimard, Voy.
Uranie, Zool.: 373, pi. 63, fig. 3. 1824 (type
locality, Guam) ; Cuvier & Valenciennes,
Hist. Nat. Poiss., 10: 229. 1835 (not Ha-
waiian Is.); ?Giinther, Cat. Fish. Brit. Mus.,
3: 334. 1861 (Borneo).

Acanthurus ctenodon Cuvier, Regne Animal.,
2: 224. 1829; Cuvier & Valenciennes, Hist.
Nat. Poiss., 10: 241, pi. 289, 1835 (Caroline
Is. and New Guinea); Gunther, Cat. Fish.
Brit. Mus., 3: 342. 1861 (Ceylon and East
Indies) ; Bleeker, Ned. Tijdschr. Dierk., 1 :
156. 1863 (Halmahera); Playfair, Fishes of
Zanzibar: 57. 1866 (Zanzibar) (as variety
d) ; Day, Proc. Zool. Soc. London: 688. 1870
(Andaman Is.).

? Acanthurus fiavoguttatus Kittlitz (not of Herre,
1936, or Fowler, 1944), Mus. Senckenb., 1 :
193, pi. 13, fig. 5. 1834 (Caroline Is.).

?Acanthurus Ketlitzii Cuvier & Valenciennes,
Hist. Nat. Poiss., 10: 222. 1835.

156

Zoologica: New York Zoological Society

[40: 15

Ctenodon Cuvierii Swainson, Nat. Hist. Class.

Monocard. Animals 2: 256. 1839.
Acanthurus strigosus Cuvier & Valenciennes
(not of Bennett), Hist. Nat. Poiss., 10: 243.
1835 (New Guinea); Bleeker, Nat. Tijdschr.
Ned. Indie, 4: 264. 1853 (New Guinea, not
Hawaiian Is.); ibid., 6: 102. 1854 (East
Indies); ?Gunther, Cat. Fish. Brit. Mus., 3;
342. 1861 (New Guinea, not Hawaiian Is.);
Kner, Novara Exped. Fische, 1: 211. 1865-
1867 (Tahiti); Gunther, Journ. Mus. Godef-
froy, 2 (3) : 116, pi. 79. figs. B and C. 1873
(Indo-Pacific) ; Day, Fishes of India, 1 : 207,
pl. 47, fig. 2. 1876 (Andaman Is.);— Fauna
Brit. India, 2: 143. 1889 (India); Harden-
burg, De Tropische Natuur, 22; 155, fig. 1.
1933.

Acronwus argenteus GUnther (in part). Cat.

Fish. Brit. Mus., 3: 346. 1861 (Marianas).
Acanthurus (Ctenodon) ctenodon Klunzinger,
Synops. Fische Rothen Meeres: 509. 1871
(Red Sea) .

Acanthurus (Ctenodon) strigosus Klunzinger,
Fische Rothen Meeres, 1: 85. 1884 (Red
Sea); Weber, Siboga Exped. Fische, 57: 319.
1913 (Indo-Australian Arch.).

Ctenochaetus strigosus Seale, Occ. Pap. B. P.
Bishop Mus., 1; 109. 1901 (Guam); Fowler
& Ball (in part), B. P. Bishop Mus. Bull.,
26: 19. 1925 (Wake I., not Hawaiian Is. or
Johnston I.); Herre, Philipp. Journ. Sci., 24:
438, pl. 15, figs. 2 and 3. 1927 (Philippine
Is.); Fowler (in part), Mem. B. P. Bishop
Mus., 10: 247. 1928 (Oceania, not Hawaiian
Is.) ; Fowler & Bean (in part) , Bull. U. S. Nat.
Mus., 100, 8: 200. 1929 (Philippine Is. and
East Indies); Herre, Field Mus. Nat. Hist.
Zool. Ser., 21 : 247. 1936 (Polynesia and East
Indies) ; Fowler, Acad. Nat. Sci. Phila. Mon.,
2: 76, 147, 173, 185, 202, 215. 1938 (Mar-
quesas Is., Society Is. and Christmas I. in the
Line Is.) ; Poll. Mus. Roy. Hist. Nat. Belgique,
18: 11. 1942 (Society Is.); Aoyagi, Coral
Fishes: 218, pl. 6, fig. 20 (teeth only). 1943
(Riu Kiu Is.); Schultz (in part), U. S. Nat.
Mus. Bull., 180: 161. 1943 (Phoenix and
Samoa Is.); Hiyama, Poisonous Fishes South
Seas: 83, pl. 19, fig. 52. 1943 (Marshall Is.) ;
Smith, Sea Fishes S. Africa: 240, pl. 33, fig.
614. 1949 (Natal); de Beaufort, Fishes Indo-
Aust. Arch., 9: 128, fig. 24. 1951 (Indo-
Pacific, not Hawaiian Is.); Harry, Atoll Res.
Bull., 18: 150. 1953 (Raroia, Tuamotus).
Ctenodon ctenodon Fowler, Journ. Acad. Nat.

Sci. Phila., 12; 545. 1904 (Sumatra).
Ctenochaetus striatus Jordon & Seale (in part).
Bull. Bur. Fish., 25: 355. 1906 (Oceania, not

Hawaiian Is.); Seale, Occ. Pap. B. P. Bishop
Mus., 4: 67. 1906 (Society Is., Cook Is., New
Hebrides, Austral Is. and Solomon Is.);
Evermann & Seale, Bull. Bur. Fish., 26: 97.
1907 (Zamboanga, P. I.); Kendall & Rad-
cliffe, Mem. Mus. Comp. Zook, 35: 145.
1912 (Manga Reva) ; ?Jordan, Tanaka &
Snyder, Cat. Fishes Japan: 634. 1913 (Mi-
saki and Sagami, Japan) ; Fowler & Bean,
Proc. U. S. Nat. Mus., 62: 57. 1922 (Zam-
boanga, P. I.); Fowler, Proc. Acad. Nat. Sci.
Phila., 75: 42. 1923 (Madagascar); Herre,
Philipp. Journ. Sci., 34: 437, pl. 13, fig. 2.
1927 (Philippine Is.);— Field Mus. Nat. Hist.
Zool. Ser., 21 : 246. 1936 (Tuamotus and
Moorea) ; Schultz & Woods, U. S. Nat. Mus.
Bulk, 202: 620, pl. 61, fig. A, pl. 62, figs.
A and B. 1953 (Marshall and Mariana Is.) .
Teuthis striatus Barnard, Ann. S. Afric. Mus.,
21: 780. 1927 (Natal).

ICtenochaetus ctenodon Whitley, Journ. Pan-
Pacific Res. Inst., 3: 12. 1928 (Santa Cruz
Is.); Whitley & Colefax, Proc. Linn. Soc.
New S. Wales, 63: 294. 1938 (Nauru).
Diagnosis.— No black spot at axil of both
dorsal and anal fins; teeth of upper jaw with 6
denticulations, of lower jaw 4, on lateral edge of
expanded tips; membranes of pectoral fin hya-
line; caudal fin lunate, caudal concavity 3.7
to 6 in standard length; dorsal soft rays usually

28 to 30.

Description.— T>ors2i\ rays VIII, 27 to 31; anal
rays III, 24 to 28; pectoral rays 16 or 17;
anterior gill rakers 27 to 36; posterior gill rakers

29 to 42; scales from gill opening to posterior
end of caudal peduncle spine 104 to 122.

Body depth 1.9 to 2.3, snout 4.2 to 4.8, pelvic
fin 3.3 to 3.7, eighth dorsal spine 4.9 to 5.8,
longest dorsal ray 3.6 to 4.4, third anal spine
5.4 to 6.7, longest anal ray 4.4 to 5.2, caudal
concavity 3.7 to 6, all in standard length.

Margins and inner surfaces of lips smooth.
Upper teeth with lateral edge of distal expanded
end with 6 (rarely 5 or 7) denticulations. Lower
teeth with 4 lateral denticulations at end (a small
fifth one is sometimes evident) .

Color in alcohol dark brown with numerous
pale lengthwise slightly irregular lines on body
(which are often faded and difficult to see);
median fins brown; traces of about 5 lengthwise
bands may at times be seen in the soft dorsal and
anal fins; pectoral fin with rays light brownish
(except uppermost principal ray which is edged
in dark brown) and membranes hyaline; pelvic
fins brown; young specimens may show a small
black spot at the base of the last few dorsal rays.
Color of juveniles dark brown with 8 to 12 pale
longitudinal stripes on the body about % as

1955]

Randall: Revision of Ctenochaetus; Five New Species

157

a

-c

o

o

Cj

u

&

u

ffi

H

oc

o

iu

K

u

X

<

O

H

c/5

&

£

m

X

H

S

0

01

Ph

(fi

01
Ui
U
<

d

O

Um

O

c/3

H

2
D
O
U

u

cc

<

H

d) \,o
cd

•5b ^
^-1
O

c/3

d>

c/3 m

a

<

TD

c

CO

Oh

C/5

lo m ri

r<3 VO m CM ^

»-H lO T-H (S fN)

^ rn rl ^ .

5 5/5

O o

E ^
« c
c/5 *—1

C cO
CO pj

G T3

CO 5

CO

T3

S3 C 3
CO ^ • —

> .i:

2

C (U

. o

Vi CU

o

U

ns

c

p-

CO .

-S Oh /-V ^

g u S3 ^ D S;

ft 5 00 “ O -

Cu

04

^ S

“ XJ

cu

^ d)

50

&>5 .-4

II

I X

05

I

C cO
. ^
Oh _-

M a
^ u

"O

S

o .

C

(u ^ 0«

X| ^ cu

a ^ a
u sS

Tf rri

m r-t

r4

ro T—

^ ri n

C T3

CO C
C/5 *—•

C CO
CO PQ
CO

G ^

CO ®
u ” d>

CO • c/5

2 "D

S .S 3
S Oh -S

•ft &<■!-'•[:

>■ .2

2 c

C lo

o

Onf!
c/5 O*

O

u

OS

o

"" ^2 '2 ^
d> 35 C

C ^ « .S3
., 04^3 t: 52 :
CO Oh 5 d)

cO

E? ^ -ft! 00 :g ^

O cO ^ J33 ?t cO

OS cO

£ la 2 K

C; 2

O CO

c/3 C
^ d)
>}

5 ^

cs

3: 0)

E O
c

2 S

s:

158

Zoologica: New York Zoological Society

[40: 15

broad as mtermediate brown areas, which angle
slightly downward as they pass posteriorly. At
a standard length from 5U to 65 mm. the num-
ber of pale lines is suddenly doubled by the
appearance of thin whitish lines in the center of
the brown intermediate areas; these narrow lines
become as broad as the pale lines of the first set,
and more lines are added above and below,
especially above, until the adult number of about
30 to 40 is attained. The ones added above the
original set tend to angle upwards.

Color in life dark olive brown with blue
lengthwise lines on the body and small orange
spots on the head and nuchal region; soft dorsal
and anal fins with about 5 lengthwise dark bluish
lines; pectoral fin yellowish.

Juvenile specimens from the Gilbert Islands
of about 40 mm. length were observed with the
first set of pale lines red in color, narrow light
bluish lines in the intermediate brown areas, and
red tips to the lobes of the caudal fin. It is this
color pattern that Herre (1927) probably tried
to show in his Figure 2 of Plate 1 3 and which led
him into the mistaken and persistent contention
that this juvenile form was a different species.
The error is an easy one to make in view of the
variability in length at which adult coloration
and configuration are assumed.

One specimen in the acronurus stage (trans-
parent with silvery abdomen) was secured by
night lighting at anchorage at Sydney Island in
the Phoenix Islands by personnel of the Pacific
Oceanic Fishery Investigations. This specimen,
USNM No. 163616 and 32 mm. in standard
length, is shown in Plate I, D. It is interesting
to note that there is already a change-over
taking place from larval dentition to adult-type
dentition even though there is no evidence of
any other bodily transformation to adult form.
This change in dentition may be seen in only
a slightly more advanced state in the trans-
forming specimen of Plate I, E, from which
the three teeth in Text-fig. 3 B to D were drawn.
These teeth were all taken from the upper jaw
of this specimen and show early stages in the
transformation of tooth form. The inward bend-
ing of the expanded tips, already apparent in C,
can only be seen in a side view of the teeth.
The tooth drawn in A of Text-fig. 3 was taken
from the strigosus acronurus of Plate I, D.
All of its teeth are essentially alike; there are
about 18 in each jaw. Probably this specimen is
in an earlier stage of development than the
striatus acronurus of Plate I, D. It is sus-
pected that the teeth of striatus younger than
the latter will bear 3 denticulations like the
strigosus acronurus. The striatus acronurus has
about 18 upper teeth and 16 lower teeth with
more apparent just beneath the gums at the ends

of the jaws. The lateral teeth are more adult-like
in form than more medially-located teeth. The
most striking thing about the dentition was the
finding, upon dissection, of good-sized adult-
type teeth imbedded in the soft tissue within
the bones of the jaws. Five were found from
one side of the upper jaw just above the upper
teeth and six from one side of the lower jaw just
below the lower teeth. Only the most medial of
these imbedded teeth showed a departure from
the Ctenochaetus pattern, for not all of the den-
ticulations of this tooth were restricted to the
lateral side. Subsequent dissection of the pre-
maxillary and dentary bones of juvenile and
adult specimens of C. striatus and other Cteno-
chaetus of all sizes showed these teeth to be
present, in greater number and larger size in
larger specimens, and it is believed that they
function as replacement teeth. Specimens of
species of Acanthurus, including acronurus and
postacronurus forms, were examined, and adult-
type teeth characteristic of this genus were also
found imbedded in the bones of the jaws.

Remarks —C. striatus is an exceedingly abun-
dant and widespread species, probably occuring
in the entire tropical Indo-Pacific region except
the Hawaiian Islands. Jordan, Tanaka & Snyder
(1913) record what is probably this species
from Japan. J. L. B. Smith in a letter states that
it appears to be the most abundant acanthurid
on the reefs of tropical east Africa. At Onotoa
Atoll in the Gilbert Islands, after two months
of collecting, I found it was the dominant fish
on a weight basis among all the reef fishes taken.
It was very common on the benched reef slope
(Cloud, 1952) of the windward reef of the
atoll and about coral heads in the lagoon.

In the Red Sea striatus seems to have a
lower gill raker count (Table 3) and also a
greater percentage of specimens with 17 pectoral
rays than 16 (Table 2) . More material is needed
from the Indian Ocean and Arabian Sea to
clarify these and other differences.

Better established is the demarcation on the
basis of dorsal and anal fin ray counts (Table
1) between populations in the Philippines and
East Indies and the rest of Oceania. Also from
the examination of fin ray counts, southeast
Oceania (i.e.. Society Islands and Tuamotus)
may represent a differentiated population.

The largest specimen seen by me was 195 mm.
in standard length. It was taken in the Philip-
pines.

Taxonomic Discussion. — In all probability
Acanthurus argenteus Quoy & Gaimard is a
transforming Ctenochaetus striatus. These au-
thors suspected that it was the young of some
species they had not collected, but the nature of

1955]

Randall: Revision of Ctenochaetus; Five New Species

159

the striped color pattern and the high fin ray
counts leave little doubt that it is the same as
striatus. I retain the name striatus because it has
been commonly used, whereas argenteus has
been recognized as a name applying to an im-
mediate postacronurus form (the name referring
to the silvery coloration of the abdomen) and
has never been applied to any adult surgeon fish.

The consideration of Acanthurus flavoguttatus
Kittlitz {A. Ketlitzii Cuvier & Valenciennes) as
a synonym of C. striatus is based on the Kittlitz
figure which is greenish with yellow spots on
the head and yellow lines on the body. Lack of
reference to dentition and the peculiar fin ray
counts account for the uncertainty in my deci-
sion. The type of flavoguttatus was not located.

Acanthurus ctenodon Cuvier & Valenciennes
was referred to synonymy by Kner (1865) and
by Gunther (1873).

Much confusion has resulted from the fre-
quent use of the name strigosus for the species
striatus. This seems to stem from Gunther
(1873) who considered striatus the young of
strigosus and in the overlooking by Gill (1884)
of the Quoy & Gaimard species altogether in
his erecting of the genus Ctenochaetus. Jordan
& Evermann (1905) perpetuated the error by
copying Gunther’s plate of striatus and using
it to represent the common Hawaiian species
which, in reality, is strigosus. Fowler (1928) did
not help by referring striatus to the synonymy
of Acanthurus lineatus.

Herre’s (1927) separation of strigosus and
striatus was based on specimens not exceeding
60 mm. in length and is erroneous.

Ctenochaetus strigosus (Bennett)

Plate II, Fig. 1; Text-fig. 1 B; Plate I,

A-C; Text-fig. 3 A

Acanthurus strigosus Bennett, 7joo\. Journ.,
4: 41. 1828 (type locality, Hawaiian Is.);
Cuvier & Valenciennes (in part?). Hist. Nat.
Poiss., 10: 243. 1835 (Hawaiian Is., prob-
ably not New Guinea); Gunther (in part?),
Cat. Fish. Brit. Mus. 3: 342. 1861 (Hawaiian
Is., probably not New Guinea).

Acanthurus (Etenodon) strigosus Steindachner,
Denkschr. Akad. Wiss., 70: 495. 1901 (Ho-
nolulu).

Ctenochaetus strigosus Jenkins, Bull. U. S. Fish.
Comm., 22: 480. 1903 (Honolulu); Snyder,
Bull. U. S. Fish. Comm., 22: 534. 1904
(Honolulu); Fowler & Ball (in part), B. P.
Bishop Mus. Bull., 26: 19. 1925 (Laysan,
French Frigate Shoals and Johnston I., but
not Wake I.) ; Fowler (in part), Mem. B. P.
Bishop Mus., 10: 274. 1928 (only Hawaiian

Is.); Fowler & Bean (in part), U. S. Nat.
Mus. Bull., 100, 8: 200. 1929 (Philippines
and East Indies) ; Fowler, Mem. B. P. Bishop
Mus., 1 1 : 344. 1931 (Honolulu) ;-Acad. Nat.
Sci. Phila. Mon., 2: 233 (only). 1938 (Hono-
lulu);—Proc. Acad. Nat. Sci. Phila., 93: 257.
1941 (Honolulu); Whitley, Proc. Roy. Soc.
New S. Wales: 23. 1954 (Great Barrier Reef,
Australia) .

Ctenochaetus striatus Jordan & Evermann, Bull.
U. S. Fish. Comm., 23: 398, (Fig. 174 after
Gunther, not strigosus) , 1905 (Hawaiian Is.) ;
Jordan & Jordan, Mem. Carnegie Mus., 10:
66. 1922 (Hawaiian Is.).

Ctenochaetus flavicauda Fowler, Acad. Nat. Sci.
Phila. Mon., 2: 104, pi. 10, fig. 24. 1938 (Tak-
aroa, Tuamotus) ;— Mem. B. P. Bishop Mus.,
12: 104. 1949 (Takaroa, Tuamotus).
Ctenochaetus sp. Harry, Atoll Res. Bull., 18:
151. 1953 (Raroia, Tuamotus).

Diagnosis.— fAo black spot at axil of dorsal
or anal fin; teeth of upper jaw with 5 denticula-
tions, of lower jaw 3, on lateral edge of ex-
panded tips; membranes of pectoral fin hyaline;
caudal fin moderately concave, caudal con-
cavity 5.7 to 10 in standard length; dorsal soft
rays usually 26 or 27.

Description.— Dorsal rays VIII, 25 to 28; anal
rays III, 21 to 25; pectoral rays 16 (rarely 15);
anterior gill rakers 27 to 34; posterior giU rakers
28 to 33; scales from gill opening to posterior
end of caudal peduncle spine 85 to 96.

Body depth 1.7 to 2, snout 4.3 to 4.8, pelvic
fin 3.1 to 3.4, eighth dorsal spine 4.8 to 5.5,
longest dorsal ray 3.8 to 4.2, third anal spine
5.3 to 6.5, longest anal ray 4.4 to 4.7, caudal
concavity 6 to 10, all in standard length.

Margins and inner surfaces of lips smooth;
upper teeth with 5 denticulations (occasional
teeth from Philippine specimens will show tiny
6th denticulation) ; lower teeth with 3 denticula-
tions (including tip) .

Color in alcohol brown with about 35 narrow
pale bluish longitudinal lines (about Va as broad
as the alternate brown bands) on the body
which angle upwards on the basal part of the
dorsal fin and downwards on the basal part of
the anal; small pale spots on the head (and
anteriorly on the body in Philippine specimens) ;
a pale ring, broader posteriorly, around the eye
in Hawaiian and Tuamotu specimens, restricted
to the posterior edge of the eye in Johnston
Island material, and absent in Philippine speci-
mens; median fins brown (except caudal fin of
Tuamotu specimens which is abruptly pale,
white in life); pectoral fin pale except edge of
uppermost principal ray which is almost black;
pelvic fins brown.

160

Zoologica: New York Zoological Society

[40: 15

Color in life from a 35 mm. kodachrome
transparency of a Hawaiian specimen brown
with narrow pale blue longitudinal lines on the
body and basally on dorsal and anal fins; pur-
plish region on chin; blue spots on head; yellow
ring around eye; caudal fin brown; pectoral fin
rays brownish, membranes orange-yellow; pelvic
fins brown.

Hawaiian specimens in life show considerable
variation in ground color. Some, especially in
a light-colored environment (as areas of high
coral cover) , become a light tan in color. Others
may be dark brown. The ground color of the
species at Raroia, Tuamotus, was observed by
Harry (1953) to be black.

Remarks.— Along with the color differences
noted above, and meristic data, differences in the
shape of the caudal fin may be seen among the
specimens from the Philippines and East Indies,
Hawaiian Islands, Johnston Island, Tuamotus
and Mauritius which might form the basis for
the recognition of subspecies when more ma-
terial is available and the range of the species
more completely known. The single specimen
from Mauritius has the least concave caudal fin,
the caudal concavity being contained in the
standard length about 20 times. The caudal con-
cavity of specimens from the Philippines and
the Hawaiian Islands is 7 to 10. Johnston Island
specimens have a caudal concavity which ranges
from 5.7 to 9 in the standard length; Tuamotu
specimens have the most lunate caudal fins, the
caudal concavity contained about 5.7 times in
the standard length.

Harry (1953) reported acute pain and swell-
ing in the hand and arm when cut on the hand
by the caudal spine of this species in Raroia;
the pain did not subside until the second day
and persisted for a week. He added that the only
other surgeon fish he encountered producing
such effects was C. striatus (recorded as strigo-
siis), though pain from this species lasted only
3 to 4 hours and was not so intense. I tested the
poisonous qualities of the caudal spine of Ha-
waiian strigosus by gingerly inserting the tip of
the spine into my palm; a stinging sensation was
soon experienced, and the experiment was car-
ried no further.

This species is extremely abundant in the
Hawaiian Islands where it is known by the local
name Kole. Elsewhere, except perhaps Johnston
Island, it is not at all common. Only 13 speci-
mens from the Philippines, Moluccas and Cele-
bes were found at the U. S. National Museum
among the vast collections made by the Alba-
tross Philippine Expedition (1907-1910); only
four are known from the Tuamotus, and one
from Mauritius. The single specimen from the
Great Barrier Reef was identified by Whitley

with the aid of a manuscript key such as appears
in this paper and was recorded from Australia
by him (1954).

Although more collecting in the Indo-Pacific
region may reveal new localities for strigosus, it
is believed that it is not in continuous distribu-
tion throughout its range. The differentiation
which has taken place in five of its known locali-
ties supports this contention. The great abun-
dance of this species in Hawaii stands in sharp
contrast to its apparent absence from much of
the Indo-Pacific and its scarcity in the few areas
where it does occur. This may be associated with
the absence of striatus from the Hawaiian
Islands.

The acronurus specimen (Plate I, A), U.S.-
N.M. 118040 and 27 mm. in standard length,
was taken at night from the steamer Albatross
at the surface at Diamond Head Light, Oahu,
Hawaiian Islands, on May 6, 1902. The trans-
forming specimen of Plate I, B (University
of Hawaii No. 1877) was collected by W. A.
Gosline on the reef at Diamond Head, Oahu,
May 16, 1950. It is 28 mm. in standard length.
The recently transformed juvenile (Plate I, C) ,
U.S.N.M. 167199, was taken by the author at
a depth of 40 feet, Waikiki, Oahu, on June
4, 1952. As no small juveniles have been seen
in tidepools or very shallow water, it is sus-
pected that the acronurus transforms to the
juvenile stage at a moderate depth on the reef.
Postacronurus and juvenile striatus, on the other
hand, occur in immense numbers in tidepools
and protected shallow-water areas.

The largest specimen of strigosus seen by me,
139 mm. in standard length, was collected by the
Albatross Philippine Expedition at Luzon.

The food habits of this species are discussed
in the general section on the genus.

Ctenochaetus cyanoguttatus, sp. nov.

Text-fig. 1 C

?Acanthurus guttatus Kittlitz (not of Bloch &
Schneider), Mus. Senckenb., 1: 193, pi. 13,
fig. 4. 1834 (Luganor I.).
lAcanthurus marginatus Cuvier & Valenciennes,
Hist. Nat. Poiss., 10: 221. 1935 (new name
for Acanthurus guttatus Kittlitz); Gunther,
Cat. Fish. Brit. Mus., 3: 333.1861 (Luganor).
lAcanthurus ctenodon Var. b Playfair, Fishes
of Zanzibar, 57. 1866. (Zanzibar).
Ctenochaetus strigosus Snodgrass & Heller, Proc.
Washington Acad. Sci., 6: 402. 1905 (Cocos
L, Costa Rica) ; Schultz (in part), U. S. Nat.
Mus. Bull., 180: 161. 1943 (Hull!.).
Ctenochaetus sp. Hiyama, Poisonous Fishes
South Seas: 92, pi. 19, fig. 53. 1943 (Mar-
shall Is.) .

1955]

Randall: Revision of Ctenocliaetus; Five New Species

161

//o/otype.— U.S.N.M. No. 167178, Aunteuma
Island, Onotoa Atoll, Gilbert Islands, lee side,
depth of water about 5 feet in poorly-defined
surge channel region, August 1, 1951, spear,
John E. Randall, 170.5 mm. in standard length.

Para/3;pcj.— U.S.N.M. No. 167179, same data
as holotype except date, August 11, 1951, 166
mm.; U.S.N.M. No. 115152, Hull Island, reef,
July 13, 1939, Leonard P. Schultz, 157 mm.;
Stanford Mus. No. 12280, Cocos Island, Costa
Rica, Hopkins-Stanford Galapagos Expedition,
1898-1899. (Precise proportional measurements
and scale and gill raker counts were not made
on the Cocos Island specimen) .

Diagnosis.— No black spot at axil of dorsal
or anal fin; teeth of upper jaw with 4 denticula-
tions, of lower jaw 3, on lateral edge of ex-
panded tips; membranes of pectoral fin dark
brown; caudal concavity 5.8 to 6.2 in standard
length; dorsal soft rays 27 or 28.

Description.— Dorsal rays VIII, 27 (27 to
28); anal rays III, 25; pectoral rays 17 (16 to
17) ; anterior gill rakers 28 (26 to 29) ; posterior
gill rakers 35 (34 to 37) ; scales from gill open-
ing to posterior end of caudal peduncle spine
95 (94-104); upper teeth 44, lower teeth 60.

Depth of body 1.85 (1.94 to 1.95), length
of head 3.48 (3.27 to 3.61), length of snout
4.34 (4.36 to 4.74), length of pectoral fin 3.10
(3.02 to 3.13), length of pelvie fin 2.94
(2.91), length of 8th dorsal spine 5.09 (5.24
to 5.27), length of longest dorsal ray 5.14 (4.62
to 4.89), length of third anal spine 5.89 (5.81
to 6.15), length of longest anal ray 5.01 (5.42
to 5.44), caudal concavity 6.20 (5.81 to 5.93) ,
all in standard length. Greatest diameter of eye
4.41 (4.26 to 4.40), width of inter-orbital space
3.06 (2.87 to 3.1), least depth of caudal pe-
duncle 2.27 (2.19 to 2.34), length of caudal
peduncle spine 2.90 (2.42 to 2.67), distance
from base of upper lip to distal ends of upper
teeth 5.1 (4.9 to 5.1), all in length of head.

Inner surfaces of lips and margin of upper
lip smooth; margin of lower lip, especially
posteriorly, papillate. Upper teeth with 4
(rarely 3) denticulations (the one at the tip
being longest) ; lower teeth with 3 denticulations
(counting tip).

Color in alcohol dark brown with numerous
small pale spots on the body and pectoral fin
(these spots are faint and difficult to see even
in freshly preserved fish) ; all fins brown; dorsal
and anal fins with about 9 or 10 longitudinal
dark brown lines (fewer anteriorly in these fins) .

Color in life dark brown with head, body,
and pectoral fins profusely covered with small
bright blue spots, hence the name, cyanogut-
tatus.

Retnarks.-Jhis species was observed by me in

its natural habitat on two occasions at Onotoa,
Gilbert Islands. It was seen in small rapidly-
moving schools in moderately rough water in
broad shallow surge channels on the lee side of
the atoll. It was a mode of life similar to that
oi Acanthurus guttatus, the latter species school-
ing in the rougher water of sharply-defined
surge channels on the windward side of the
atoll. The spotting of the body of both these
species might be associated with the masses of
small swirling air bubbles in the water that char-
acterize the surf zone where they live.

Taxonomic Discussion.— Acanthurus guttatus
Kittlitz was a brown fish with blue spots and
8 dorsal spines. Cuvier & Valenciennes, realizing
that this could not be the A. guttatus of Bloch
& Schneider, gave the species the name margina-
tus. It is possible that this species was Cteno-
chaetus cyanoguttatus. But no mention was made
of the important item of dentition, although the
teeth in Kittlitz’ figure are drawn fairly long.
The count of dorsal spines cannot be considered
too diagnostic, for species of Acanthurus have
been recorded with 8 instead of 9 dorsal spines
because of the inconspicuous nature of the first
spine. Even the blue spots are not unique, for
adult Acanthurus nigroris Cuvier & Valen-
ciennes {=Hepatus atramentatus Jordan &
Evermann) at Wake Island had the usual blue
lines broken up into spots, which gave this
species an appearance much like C. cyanogut-
tatus. In view of this, and the fact that the
Kittlitz specimen has not been located, describ-
ing the species as new seems in order. The Kitt-
litz specimen is not in the Senckenberg Museum.
Gunther (1861) stated that the type of Acan-
thurus pyroferus described by Kittlitz from the
same publication was in the old St. Petersburg
Museum. The specimen of A. rnarginatus may
be there; I have been unable to complete cor-
respondence on the matter.

The Cocos Island specimen of Snodgrass &
Heller (1904) is referable to this species, al-
though there is no record of the color in life
and the spots which were probably present have
now faded. G. S. Myers, who kindly loaned the
specimen (SU 12280) to me, states that only
one other of the four specimens collected from
Cocos is now located at the Natural History
Museum, Stanford University.

CtENOCHAETUS HAWAIIENSIS, sp. nov.

Plate II, Fig. 2; Text-fig. 1 D
Ctenocliaetus hawaiiensis Brock, Journ. Wildlife
Man., 18: 307. 1954 (nomen nudum; name
used by Brock from a personal communica-
tion before species published) .

Ho/o/ype.— U.S.N.M. No. 167180, entrance

162

Zoologica: New York Zoological Society

[40: 15

to Keauhou Bay, Hawaii, Hawaiian Islands,
depth of water about 12 feet, February 8, 1952,
spear, John E. Randall, 197.5 mm. in standard
length.

Pu/'utjpes.— U.S.N.M. No. 167181, Keahole
Point, Hawaii, Hawaiian Islands, depth of water
40 feet, June 18, 1953, spear, Vernon E. Brock,
2 specimens, 204 and 209 mm.; Standford Mus.
No. 47661, same data as other paratypes, 202.5
mm.

Diagrtosis.—No black spot at axil of dorsal or
anal fin; teeth of upper and lower jaws with 3
denticulations on lateral edge of expanded tips;
membranes of pectoral fin dark brown; twice
as many teeth in lower jaw as upper; caudal fin
slightly emarginate to nearly truncate; dorsal
soft rays 27 or 28.

Description.— Dorsal rays VIII, 28 (27 to
28) ; anal rays III, 25 (25 to 26) ; pectoral rays
16; anterior gill rakers 25 (23 to 25); posterior
gill rakers 28 (25 to 27) ; scales from gill open-
ing to posterior end of caudal peduncle spine
139 (115 to 131); upper teeth 26; lower teeth
56.

Depth of body 1.81 (1.82 to 1.84), length of
head 3.00 (3.27 to 3.32), length of snout 3.66
(3.81 to 3.92), length of pectoral fin 3.05 (3.07
to 3.21 ) , length of pelvic fin 3.87 (3.61 to 3.78) ,
length of eighth dorsal spine 4.60 (4.50 to 4.80) ,
length of longest dorsal ray 4.12 (4.18 to 4.21),
length of third anal spine 6.49 (5.96 to 6.97),
length of longest anal ray 4.49 (4.50 to 4.86),
caudal concavity 18.8 (15.0 to 40.8), all in
standard length. Greatest diameter of eye 4.86
(4.39 to 4.85), width of inter-orbital space 3.30
(2.97 to 3.14), least depth of caudal peduncle
2.30 (2.17 to 2.38), length of caudal peduncle
spine 3.14 (2.87 to 3.75) , distance from base of
upper lip to distal ends of upper teeth 3.66 (3.15
to 3.64), all in length of head. The last-men-
tioned measurement, which might also be
loosely termed the height of upper lip, is greater
in this species than any other in the genus (this
distance in other species is contained in the head
length 4.5 to 6.6 times) .

Margins of lips finely crenulate; distal one-
fourth of inner surfaces of lips plicate, the ridges
running perpendicular to the margin. Upper and
lower teeth with 3 denticulations (including
tips); lower teeth about twice as numerous as
upper teeth.

Color in alcohol very dark brown with many
narrow pale longitudinal lines faintly visible
on head and body; (these lines tend to be less
irregular than those on other species; the lines
on the head are somewhat diagonal); all fins
dark brown.

Color in life dark olive brown (appearing al-

most black underwater) with fine yellowish-gray
lengthwise lines on the head and body.

Remarks.— This species is thus far known
only from the island of Hawaii in the Hawaiian
Islands, where it is common; it is named for
this locality. Fishermen have told me that it is
seen rarely at the island of Maui. It certainly
does not seem to be present in the waters around
the island of Oahu, which are well-collected.

I have observed hawaiiensis underwater on
only four occasions, three times as a solitary
fish and once as a group of three fish. Vernon E.
Brock, Division of Fish and Game, Territory of
Hawaii, has told me that he has observed the
species in schools.

A smaller specimen (estimated 100 mm. in
standard length), yellowish-brown in color, and
possibly hawaiiensis, was sighted by me at a
depth of 70 feet in Kealakekua Bay, Hawaii,
but was not taken.

CtENOCHAETUS MAGNUS, sp. nov.

Text-fig. 1 E

Ctenochaetus strigosus Fowler (in part), B. P.

Bishop Mus. Bull., 38: 20. 1927 (Jarvis L).

Holotype.—D.S.'N.M. No. 163614, Malden
Island (4° 03' S., 154“ 59' W.) , January 27,
1951, spear, Herbert Mann and Joseph E. King,
Pacific Oceanic Fishery Investigations, 225 mm.
in standard length.

Paratypes.—Stanford Mus. No. 47662, same
data as holotype, 219 mm.; B. P. Bishop Mus.
No. 4308, Jarvis Island (0° 23' S., 160° 02' W.),
August 15, 1924, Whippoorwill Expedition,
205.5 mm., U.S.N.M. No. 163659, Cocos Is-
land, Costa Rica, December 28, 1952, Bruce W.
Halstead and Norman C. Bunker, 209 mm.

Diagnois.—l^o black spot at axil of dorsal or
anal fin; teeth of upper and lower jaws with 3
denticulations on lateral edge of expanded tips;
membranes of pectoral fin dark brown; 1.2
times as many teeth in lower jaw as upper jaw;
caudal concavity 6 to 7 in standard length;
dorsal soft rays 26 or 27.

Description.— Dorsal rays VIII, 27 (26 to

28) ; anal rays III, 25 (24 to 25) ; pectoral rays
17 (16 to 17); anterior gill rakers 28 (28 to

29) ; posterior gill rakers 36 (35 to 36); scales
from gill opening to posterior end of caudal
peduncle spine 159 (149 to 164); upper teeth
59; lower teeth 68.

Depth of body 2.14 (2.02 to 2.13) , length of
head 3.14 (3.32 to 3.48), length of snout 4.60
(4.18 to 4.76), length of pectoral fin 3.13 (3.05
to 3.17); length of pelvic fin 2.87 (2.98 to
3.16), length of eighth dorsal spine 5.00 (5.76
to 5.97), length of longest dorsal ray 5.92 (5.54
to 5.63), length of third anal spine 6.52 (6.26 to

1955]

Randall: Revision of Ctenochaetus; Five New Species

163

6.74), length of longest anal ray 6.00 (5.02 to
6.44), caudal concavity 5.63 (6.04 to 7.45),
all in standard length. Greatest diameter of eye
5.07 (4.50 to 4.88), width of inter-orhital space
2.87 (3.00 to 3.07), least depth of caudal pe-
duncle 2.27 (2.22 to 2.47), length of caudal
peduncle spine 2.36 (2.33 to 2.57), distance
from base of upper lip to distal ends of upper
teeth 5.28 (4.5 to 5.3), all in length of head.

Margins and inner surfaces of lips smooth.
Upper and lower teeth with 3 denticulations
(including tips).

In alcohol the color of this species is uniform
dark brown. The only record of color in life
which I have is that given by Fowler (1927) for
the Jarvis Island specimen. He states that the
body including the pectorals was covered all
over with fine blue-gray dots.

Remarks— Thus far Ctenochaetus striatus
(and C. strigosus as well) is not known from
Malden, Jarvis and Cocos Islands, in spite of its
common occurrence elsewhere in the tropical
Pacific. These islands are the sole known locali-
ties for C. magnus.

Assuming that Ctenochaetus is Indo-Pacific
in origin, which seems reasonable in view of the
distribution of the species of this genus and the
Acanthuridae in general, one must explain how
magnus (and cyanoguttatus) crossed the East
Pacific barrier (Ekman, 1953). Malden Island
and especially Jarvis Island are near the coun-
ter-equatorial current which could conceivably
carry fish larvae the great distance to Cocos
Island (possibly via the Galapagos Islands).
Herre (1940) discusses this mode of transport.

This species is named magnus in reference
to its large size.

Ctenochaetus tominiensis, sp. nov.

Text-fig. 1 F

Ctenochaetus strigosus Fowler & Bean (in part) ,

U. S. Nat. Mus. Bull., 100, 8: 200. 1929

(Gulf of Tomini, Celebes, only).

Ho/otype.-U.S.N.M. No. 136112, Sadaa Is-
land, Gulf of Tomini, Celebes, November 17,
1909, dynamite. Albatross Philippine Expedi-
tion, 98 mm. in standard length.

U.S.N.M. No. 136093, Buka Is-
land, Gulf of Tomini, Celebes, November 20,
1909, dynamite. Albatross Philippine Expedi-
tion, 4 specimens, 74 to 100 mm.; Stanford
Mus. No. 47663, same data as other paratypes,
87 mm.

Diagnosis.— A black spot at axil of both the
dorsal and the anal fins; teeth of upper jaw
with 3 denticulations on upper lateral edge of
expanded tips; teeth of lower jaw with 3 dentic-
ulations; membranes of caudal fin, posterior

parts of dorsal and anal fins and pectoral fins
pale; margins of lips papillate; dorsal soft rays
24 or 25.

Description.— DorsdiX rays VIII, 24 (25) ; anal
rays III, 22 (22 to 23); pectoral rays 15, 16
(15 to 16) ; anterior gill rakers 21 (20 to 21);
posterior gill rakers 20 (20); scales from gill
opening to posterior end of caudal peduncle
spine 91 (83 to 86) ; upper teeth 33; lower teeth
32.

Depth of body 2.04 (1.80 to 2.00), length
of head 3.49 (3.22 to 3.37), length of snout
4.56 (4.40 to 4.65), length of pectoral fin 3.06
(2.69 to 3.03), length of pelvic fin 3.50 (3.28
to 3.59), length of eighth dorsal spine 6.12
(5.33 to 6.06) , length of longest dorsal ray 3.50
(3.57 to 4.40) , length of third anal spine 6.95
(5.87 to 6.90), length of longest anal ray 3.92
(4.00 to 4.11), caudal concavity 5.03 (4.35 to
6.52), all in standard length. Greatest diameter
of eye 3.37 (3.17 to 3.53), width of inter-orbital
space 2.92 (2.81 to 3.09), least depth of caudal
peduncle 2.29 (2.23 to 2.44), length of caudal
peduncle spine 2.55 (2.50 to 3.25), distance
from base of upper lip to distal ends of upper
teeth 6.20 (6.00 to 6.10), all in length of head.

Margins of lips papillate or crenulate, inner
surfaces smooth. Upper teeth with distal half of
the expanded ends smooth and blade-like and
basal half divided into 3 (rarely 2) lateral
denticulations; lower teeth with 3 (occasionally
4) denticulations (including tip).

Color in alcohol brown with a jet black spot
at the base of the last few dorsal and anal fin
rays, these spots extending slightly on to the
caudal peduncle; caudal fin pale yellowish,
gradually becoming brown basally, outer por-
tions of the soft dorsal and anal fins pale yel-
lowish, especially posteriorly, basally brown like
body; outer portion of the brown part of these
fins with about 3 to 5 narrow pale horizontal
bands (difficult to see on some specimens)
which become confluent with the pale distal
region of the fins as they pass posteriorly;
pectoral fins with rays brownish, membranes
pale; pelvic fins brown, a little darker terminally.

The following color note from fresh speci-
mens was taken from Fowler & Bean (1929) ; it
was associated with this species by means of
Albatross field numbers. “Brownish, dark in
life, spotting of side of head indistinct, lower
part slightly paler. Dorsal olive, with 5 or 6 bars,
beginning as darker olive and on soft fin become
cadmium or orange, fuse on fin posteriorly and
terminally to form entire color; extreme fin
edged narrowly black above; black blotch at
axil. Anal like dorsal. Caudal fades whitish.
Ventral blackish terminally, membranes hyaline

164

Zoologica: New York Zoological Society

[40: 15

and scattered small, orange, basal spots, rays
probably black in life.”

Named tominiensis for the Gulf
of Tomini, Celebes.

Ctenochaetus binotatus, sp. nov.

Text-fig. 1 G

Ctenochaetus strigosus Fowler & Bean (in part) ,
U. S. Nat. Mus. Bull., 100, 8: 200. 1929 (Phil-
ippine Is. and East Indies) .
Holotype.-U.S.'NM. No. 136125, Pagapas
Bay, Luzon, Philippine Islands, February 20,
1909, dynamite. Albatross Philippine Expedi-
tion, 111 mm. in standard length.

Paratypes.-U.S.N.M. Nos. 136061, 136064,
136073, 136075 to 136080, 136083, 136095,
136099, 136102 to 136105, 136109, 136111,
136114, 136116, 136119, Philippines and

Moluccas, January 20, 1908, to November 24,
1909, all taken with use of dynamite except one
specimen from market at Cebu Island, Alba-
tross Philippine Expedition, 27 specimens, 79
to 141 mm.; Stanford Mus. No. 27174, Jolo,
Philippines, August 15, 1931, A. W. Herre, 2
specimens, 106 and 121 mm.

Diagnosis— A black spot at axil of both the
dorsal and anal fins; teeth of upper jaw with 6
denticulations, of lower jaw 3, on lateral edge
of expanded tips; membranes of caudal fin and
dorsal and anal fins dark brown; membranes of
pectoral fin hyaline; margins of lips smooth;
dorsal soft rays 24 to 27 (mostly 26).

Description.— Dorsal rays VIII, 25 (24 to 21),
anal rays III, 23 (22 to 25), pectoral rays 16
(15 to 16) , anterior gill rakers 29 (23 to 29),
posterior gill rakers 24 (22 to 27), scales from
gill opening to posterior end of caudal peduncle
spine 102 (93 to 100), upper teeth 39, lower
teeth 42.

Depth of body 1.93 (1.88 to 2.03), length
of head 3.47 (3.29 to 3.79), length of snout
4.74 (4.77 to 5.47), length of pectoral fin 2.77
(2.86 to 3.10), length of pelvic fin 3.47 (3.44 to
3.78) , length of eighth dorsal spine 5.84 (5.64
to 6.15) , length of longest dorsal ray 3.77 (3.61
to 4.58), length of third anal spine 5.84 (5.81 to
6.21), length of longest anal ray 4.21 (4.20 to
4.74), caudal concavity 4.27 (4.51 to 6.15), all
in standard length. Greatest diameter of eye
3.41 (3.12 to 3.91), width of inter-orbital space
2.91 (2.60 to 3.16), least depth of caudal pe-
duncle 2.19 (1.97 to 2.32), length of caudal
peduncle spine 2.74 (2.43 to 3.60), distance
from base of upper lip to distal ends of upper
teeth 6.4 (5.5 to 6.6) , all in length of head.

Margins and inner surfaces of lips smooth.
Upper teeth with expanded distal part divided
into 6 approximately equal lateral denticula-

tions; lower teeth with 3 denticulations (count-
ing tip) .

Color in alcohol brown with pale lengthwise
lines on the body, (much as in strigosus ex-
cept these lines are about twice as broad as the
alternate darker brown lines); a prominent
black spot at the base of the last few dorsal and
anal fin rays, these spots extending narrowly on
to the caudal peduncle; median fins brown;
pectoral rays light brownish, the membranes
hyaline; margin of uppermost principal pectoral
ray dark brown; pelvic fins yellowish, lateral
edges and ends of rays brownish.

There is no positive record of the color in
life. Fowler & Bean (1929) give a color note for
three specimens, one of which was identified as
the type of binotatus by virtue of the Albatross
field number; however, one of the remaining
specimens (the other was not located) proved
to be C. striatus. The color description is as
follows: “Fine stripes of light bluish on the
shoulder at and below the pectoral base and
body posteriorly greenish. On dorsal 5 or 6
densely greenish stripes, similarly on anal. First
pectoral ray very dark, center of fin yellow.
Ventrals like body. Some examples with fine
lines on body very deep violet.” Both the type
of binotatus and the one specimen of striatus
faintly show in the preserved state the narrow
lengthwise lines posteriorly on the body as well
as the shoulder and pectoral region. It seems
odd that the posterior markings would not be
apparent in life.

Remarks.— lAameA binotatus tor the two black
marks, one at the axil of the soft dorsal fin and
the other at the axil of the anal fin.

A recently-transformed 35 mm. specimen of
Ctenochaetus (U.S.N.M. No. 167177), col-
lected in the lagoon at Onotoa Atoll, Gilbert
Islands, by the author was noted to have a
brilliant yellow caudal peduncle and caudal fin
and stood out in sharp contrast to the young
of C. striatus picked up at the same time. Exam-
ination of the preserved specimen some months
later revealed the last few rays of the soft dorsal
and anal fins to be colorless and a small spot to
be present in each fin at the base of these rays;
small dark brown spots were observed on the
head and anteriorly on the body. Meristic data
are as follows : D VIII, 27 ; A III , 25 ; pectoral rays
15; anterior gill rakers 27; posterior gill rakers
27, upper teeth 22, lower teeth 22. The upper
teeth bear 6 or 7 denticulations (usually 7) and
the lower teeth 3 or 4 (mostly 4) . Of the known
species of Ctenochaetus, this specimen would
best receive the label of binotatus-, however,
there are some obvious differences such as col-
oration and structure of the teeth. Since no
specimens of binotatus below 79 mm. in stan-

1955]

Randall: Revision of Ctenochaetus; Five New Species

165

dard length are available for comparison and
none of any size from Gilbert Islands (if the
species occurs there), the differences observed
may be due to the juvenile nature of the spec-
imen or the geographical separation of the
Gilbert Islands from the Philippines and East
Indies. It is certain that the specimen is not
Ctenochaetus flavicauda Fowler, for the latter
is a variant of C. strigosus, and probably had a
white tail in life instead of a yellow one.

References^

Aoyagi, H.

1943. Coral Fishes, viii -|- xii -|- 224 pp., 37 pis.,
54 text figs. Tokyo.

- See synonymy for taxonomic references.

Cloud, P. E., Jr.

1952. Preliminary report on geology and marine
environments of Onotoa Atoll, Gilbert
Islands. Atoll Res. Bull. No. 12: 1-73,
8 figs.

Ekman, S.

1953. Zoogeography of the Sea. xii -f- 417 pp.,
121 figs. London.

Herre, a. W.

1940. Distribution of fish in the tropical Pacific.
Proc. Sixth Pac. Sci. Congr., 3: 587-592.

Schultz, L. P. and collaborators

1953. Fishes of the Marshall and Marianas
Islands, U. S. Nat. Mus. Bull. (202), Vol.
1 : xxxii + 685 pp., 74 pis., 90 text-figs.

166

Zoologica: New York Zoological Society

[40: 15: 1955]

EXPLANATION OF THE PLATES
Plate I

Acronurus, postacronurus, and juvenile stages of
Ctenochaetus. A to C. strigosus. D to F. striatus.
Natural size.

Plate II

Fig. I. Ctenoc/iae/MJ (Bennett). 95 mm.

specimen from the Hawaiian Islands
drawn by Miss Marion Adachi.

Fig. 2. Ctenochaetus hawaiiensis, sp. nov. Draw-
ing of holotype by Miss Marion Adachi.

RANDALL

PLATE I

A REVISION OF THE SURGEON FISH GENUS CTENOCHAETUS, FAMILY ACANTHURIDAE.
WITH DESCRIPTIONS OF FIVE NEW SPECIES

RANDALL

PLATE II

- - ■ ■ ’► ■ -7^ '

FIG. I

FIG. 2

A REVISION OF THE SURGEON FISH GENUS CTENOCHAETUS. FAMILY ACANTHURIDAE,
WITH DESCRIPTIONS OF FIVE NEW SPECIES

16

Imaginal Behavior of a Trinidad Butterfly, Heliconius erato hydara
Hewitson, with Special Reference to the Social Use of Color ^

Jocelyn Crane
Department of Tropical Research,

New York Zoological Society, New York 60, N. Y.

(Plates I-III; Text-figures 1 & 2)

[This paper is one of a series emanating from the
tropical Field Station of the New York Zoological
Society, at Simla, Arima Valley, Trinidad, British
West Indies. This station was founded in 1950 by
the Zoological Society’s Department of Tropical
Research, under the direction of Dr. William Beebe.
It comprises 200 acres in the middle of the Northern
Range, which includes large stretches of undisturbed
government forest reserves. The laboratory of the
station is intended for research in tropical ecology
and in animal behavior. The altitude of the research
area is 500 to 1,800 feet, with an annual rainfall of
more than 100 inches.

[For further ecological details of meteorology
and biotic zones see “Introduction to the Ecology
of the Arima Valley, Trinidad, B.W.I.,” William
Beebe. (Zoologica, 1952, Vol. 37, No. 13, pp. 157-
184.)].

Contents

PAGE

I. Introduction 167

II. Historical Review 168

III. Materials and Methods 169

IV. Vision; Spectral Range and Color

Perception 170

V. Behavior of Imagos: General Account. . 174

A. Introduction 174

B. Habitat and Flight Habits 174

C. Relation of Behavior to Meteoro-
logical Conditions 176

D. Feeding 176

E. Social Behavior 177

F. Defense 180

G. Oviposition 182

H. Longevity 183

VI. Experimental Analysis of Social Behavior 183

A. Introduction 183

B. Work with Concealed Females. ... 183

C. Work with Painted Butterflies. ... 183

D. Work with Butterfly Models 185

VII. Conclusions Concerning Social Behavior 192

VIII. Summary 193

IX. References 194

I. Introduction

A MONO the most interesting of neotropical
f\ insects are the predominantly red and
jL black butterflies occurring in the genus
Heliconius. Because of the great variety of pat-
terns found in closely related or identical species,
the group poses a famous problem in systema-
tics. All of these butterflies are apparently
strongly aposematic, and with their basic color-
ing of scarlet on black, they are considered a
classic example of warning coloration. They
also serve as one of the most often quoted il-
lustrations of Mullerian mimicry, since the high-
ly variable members of two distinct sections of
the genus show many closely similar pairs of
forms.

Their general ecology has long been familiar
to collectors in the damper parts of the new
world tropics. The butterflies are usually com-
mon in the various types of rain, montane rain,
seasonal and swamp forests. They rarely occur
in the depths of the forest itself, however, prefer-
ring its edges, glades, clearings, trails and road-
sides. They also are typical inhabitants of well-
grown open second growth, and of lands devoted
to such crops as citrus, cocoa and bananas, pro-
vided only that woodland is in the vicinity and
that the crops are not too cleanly cultivated.

Except for records of roosting aggregations, it
appears that nothing has been published on the
social behavior, including the courtship, of Heli-
conius. The purpose of the present paper is to de-
scribe and analyze this behavior in the common
Trinidad form, H. erato hydara Hewitson, with
special emphasis on the function of color in in-

iContribution No. 963, Department of Tropical Re-
search, New York Zoological Society.

168

Zoologica: New York Zoological Society

[40: 16

traspecific relations. Because of the color vari-
ability in other localities and in related sympatric
species, the possible role of this characteristic as
a social releaser is considered to be of special in-
terest from an evolutionary point of view. This
is particularly so because one function of the
garish colors is held by many, including the
writer, to be aposematic.

In contrast to some continental forms (see
especially Beebe, 1955) the Trinidad subspecies
is little variable in color or pattern, being black
with an unbroken red band running transversely
across the middle of the forewing. Except for
minor differences in the red, yellow and white
dots and bars of head, thorax and basal under-
wings, macroscopic variation is confined to the
irregularities of the margins of the red forewing
bands, and to their exact width. A rare local
variant has a single tiny red spot, less than 1 mm.
across, near the anterior margin of the hind-
wing. The subspecies, as at present defined,
ranges across northern South America from
Panama to Trinidad. The species in the broadest
sense apparently flies throughout the neotropics
from Guatemala south, exclusive of the West
Indies.

Because of the emphasis in the present study
on the function of color, it was essential to con-
duct basic experiments on the existence and ex-
tent of color perception in this species. These
results are included in the present report as
prerequisites to the work on social behavior.
However, a thorough study of spectral limits,
the division of the spectrum into hues, and the
precise boundaries between innate and learned
aspects of feeding behavior must await further
work. A basic remaining requirement is the
making of electroretinograms.

Also only briefly treated in the present paper
are the three or more odors so far detected in
this species, as well as analytical work on roost-
ing behavior.

It is a pleasure to thank the Research Labora-
tories of the Interchemical Corporation, New
York, for contributing spectrophotometric
curves of the numerous color samples used in the
experiments, and Mr. Ernest C. Crocker, of the
Flavor Laboratory of Arthur D. Little, Inc.,
Cambridge, for his report on the odors of Heli-
conius scent organs.

Deep appreciation goes to Dr. William Beebe,
who inaugurated and took part in many of the
observations recorded in the following pages.
Hearty thanks also go to all members of the
Department staff. Dr. Beebe, Mr. Henry Flem-
ing. Miss Rosemary Kenedy and Dr. Richard
Tashian, for their help in assembling and rearing
specimens, for witnessing key experiments and
for many most helpful suggestions.

II. Historical Review

Butterfly behavior has rarely been systematic-
ally observed, much less analyzed by experi-
mental means. A principal exception is a study
of the satyrid, Eumenis semele (Tinbergen et ah,
1941) . In this contribution, the feeding and
courtship behavior were described, and the re-
spective roles of odor, movement, form and
color as feeding and social releasers investigated.

Use’s pioneer work on butterfly color vision
and behavior is also outstanding (1928, 1932.1,
1932.2, 1937). She established beyond question
the fact of butterfly color perception through
their responses to artificial flowers. The models
were made from series of standard gray and
colored papers. Differences among the vanessids.
pierids and papilionids in innate color prefer-
ences were found to occur, and foundations laid
for determination of the number and boundaries
of the hues distinguished.

Eltringham (1919, 1933) and Eltringham and
Ford (Ford, 1945) performed experiments
which produced positive responses in Argynnis
euphrosyne to dyed specimens which had been
previously dried and bleached, and to photo-
graphic models, painted to resemble living but-
terflies in color. Their results also indicated that
these butterflies respond to color irrespective of
motion or odor. However, the spectrophoto-
metric reflectances of the paints and dyes used
were not known, nor was hue completely sepa-
rated from either brightness or pattern. In none
of the above studies was the possible role of the
ultraviolet investigated.

Similar experiments were carried out by Use
(1937), by Eggers (1938.1, 1938.2) and by
Petersen, Tornblom & Bodin (1952), all of
whom worked with pierids. The latter contribu-
tion, based on responses of wild polymorphic
species to dried specimens and to models, makes
a needed beginning on the vital problem of
visual releasers in polymorphic butterflies. The
role of the low ultraviolet reflectance of white
pierid wings remains uncertain (see also Crane,
1954, p. 108 and ref.).

Records of butterfly courtships, mostly de-
scriptions of necessarily incomplete observations
in the field, are scattered through the literature,
particularly in the field accounts of the older
naturalists. Carpenter (1935) gives a general
discussion and a good bibliography. Courtships in
the heliconines, however, have apparently been
hitherto unknown. The only reference to mating
behavior appears to be that of Edwards (1884)
who reported on males of H. charithonius
gathering on the chrysalids of females and mat-
ing with them on emergence. Territoriality is
suggested by Seitz (1913, p. 377) who observed

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

169

heliconine males “promenading” up and down
individual beats in the forest, day after day.

The emission of scent by butterflies, apparent-
ly sometimes as sexual attractants and some-
times for protection against predators, has been
described from time to time. Notable are Long-
staff’s (1912) observations, which include de-
scriptions of the strong, witch-hazel-like odor
which some Heliconius emit when captured (pp.
492, 503-504). Poulton (1925, 1931.2), Long-
field (1926) and Collenette (1929) added more
details and opinions on the scents of heliconiids,
including the subject of this paper. Muller’s
(1878) descriptions and figures of scent glands
and scales in danaids and heliconines are class-
sics; translations have been published by Long-
staff as appendices (loc. cit.). Eltringham (1925,
1926) published further studies of both the
glands and their odors in H. erato hydara, as
well as in the genera Coloenis, Dione and Euei-
des; these three genera included species now re-
ferred to Dry as, Agraulis and Heliconius. Fin-
ally Barth (1953) discussed the form and func-
tion of the scent organs of male Lepidoptera
from a more generalized and comparative point
of view.

Gregarious roosting habits have been reported
for a number of butterflies, including representa-
tives of the danaids, heliconiids, ithomiids and
nymphalids (e.g. Beebe, 1918, pp. 203 ff.;
Myers, 1930; Jones, 1930, 1931; Poulton,
1931.1, 1931.2; Poulton and others, 1933; Gup-
py, 1932 and Carpenter, 1932). The only ex-
perimental work has been that of Jones ( loc.
cit.) who determined that H. charithonius in
Florida was attracted on successive nights by a
general locality recognition of some kind, rather
than to special twigs or branches. Hence the at-
traction was not to an odor left on particular
resting places.

Gatherings of various butterfly species on
damp sand or around mudholes are also well
known (e.g. Collenette & Talbot, 1929) , as are
the winter aggregations of monarch butterflies,
Danaus plexippus, as mentioned in, for example,
Swain (1948, p. 109). Records of butterfly mi-
grations in all part of the world are exceedingly
numerous, the most intensive study having been
made by Williams (1930). No migrations of
heliconines appear to have been reported, except
the occasional flights, small and moderately
large, reported at Portachuelo Pass, Rancho
Grande, in Venezuela by Beebe (1950).

Food plant selection and the general egg-
laying habits of female butterflies have received
the widest notice, chiefly because of the fre-
quent economic importance of the subject. Seitz’s
(1913, p. 377) observations on Heliconius seem
to be the only field notes for that genus, al-

though the caterpillars are fairly well known.
He observed these butterflies near Belem lay-
ing eggs on Passi flora, usually around noon;
they apparently chose high-climbing vines.

A paper on the spectral composition of cer-
tain butterfly colors (Crane, 1954) includes an
analysis of the red band in H. erato which was
a prerequisite to the investigation of the band’s
function in intraspecific relations. It proved to
have a minimum reflectance from the ultra-
violet through the green, low in the yellow and
high in both the orange and red.

III. Materials and Methods

Two circumstances permitted the undertak-
ing of the present study. Of first importance was
our location in a permanent tropical field sta-
tion in an area where Heliconius erato is com-
mon. This contribution is the result of observa-
tions and experiments extending over parts of
five years. The second factor was our mainten-
ance of two large insectaries for the study of
captive butterflies.

The construction and operation of these fly-
ing cages have been described elsewhere (Crane
& Fleming, 1953). Here it need only be said
that the structures measure 12' X 18' and 24' X
36', respectively, and are built of fine wire mesh
on a wooden framework. Being floorless, they
are planted with a profusion of herbs, shrubs
and vines approximating natural conditions.
Ample sun, shade and wind protection are es-
sential as well as the maintenance of high humi-
dity. Butterflies caught in the field are placed in
envelopes, kept cool and released in the insec-
taries as soon as possible.

Heliconius erato proved to be the most amen-
able to life in the insectaries of numerous spe-
cies tested. Once these large cages have been
“seasoned,” presumably through clinging species
odors, newcomers settle in at once, feeding and
successfully avoiding the wire within minutes of
their release. Reared specimens are equally
adaptable.

The favorite local food flower of these butter-
flies is Lantana camera Linnaeus; also popular
are Bidens pilosa Linnaeus and Verbena spp. All
of these are either grown in the insectaries or
provided daily as freshly cut flowers.

Small cubic cages measuring about three feet
on each side served to isolate newly emerged
imagos in preparation for experimental work.
Individuals could be thus kept for about two
days without harm to their subsequent vitality,
providing that the cage stood in the shade, was
partly covered with green branches and was
freely sprinkled with water during the heat of
the day. Lantana was supplied on a shelf six
inches below the roof, since butterflies kept in

170

Zoologica: New York Zoological Society

[40: 16

such small spaces do not readily feed from bou-
quets on the floor. Although mating was once se-
cured in a small cage, normal courtship was
absent, and no female was ever induced to
oviposit under such conditions.

A description of the growth stages of this
species, as well as an account of the rearing
technique, is in preparation. Here it is appropri-
ate to mention the general precautions necessary
in order to produce the vigorous adults which
are essential for behavior study : the larvae must
be reared in individual, closely covered contain-
ers; provided amply at least once a day with
freshly picked, or briefly refrigerated, leaves of
appropriate sizes from the vine Passiflora tube-
rosa Jacquin; furnished with suitably high
amounts of light and humidity; and kept clean.
The chrysalids also must be kept in a humid en-
vironment. In short, this is not a species that
can be healthfully reared en masse in a shoebox.

The experimental techniques employed in the
study of vision and of social behavior will be
described under the appropriate sections (be-
low and pp. 183-190).

IV. Vision: Spectral Range and
Color Perception

In this section will be briefly considered cer-
tain characteristics of the visual sense of Heli-
conius erato, with brief mention of other but-
terflies.

A few experiments were made, as described
below, only to establish with certainty that color
perception and sensitivity from the near ultra-
violet at least to the orange exist in this species
and play a role in both innate and acquired
feeding responses. Detailed studies of color vis-
ion and feeding behavior are reserved for the
future.

A. Limits of the Spectrum. From time to
time various species of butterflies, including H.
erato, were tested for responses to light of vari-
ous spectral characteristics. In each experiment
a wooden box was used, measuring either 3' X 3'
X 2' or 4' X 3' X 3'. Each box was completely
light-tight, all cracks and corners being rein-
forced with black photographic tape, except for
the following apertures: One 5" X 5" opening
in a corner of the top, for covering with a filter
which, in turn, was sealed in place with tape; a
light-tight door, near the bottom, for insertion
of butterflies; a suitably-sized hole for insertion
of a flashlight, which was fastened tightly with
the switch extending outside; and finally an eye-
hole. The latter measured about 1% inches in
diameter and was fitted with a short length of
flexible black tubing; when not in use the ex-
terior end of the tube was covered with a black

paper cornucopia. Before every test the box
was checked for light-seepage by an independent
observer.

After the introduction of one to three but-
terflies the box was left undisturbed in the open
air for a maximum of five minutes. When the
ultraviolet filter was used on dull days, the natur-
al daylight was supplemented by placing a port-
able ultraviolet lamp close above the filter. After
a few minutes, the positions of the insects were
observed by means of the eye-hole and, in the
case of the ultraviolet filter, by use of the in-
serted flashlight. The filter was then covered with
a piece of wood and, when the observer’s hand
was inserted through the butterfly entry door,
the butterflies were disturbed and forced to seek
new positions away from the filter. The same
procedure was then repeated twice more with all
individuals tested.

In the tests involving H. erato, 1 1 out of 1 2
specimens individually introduced under each of
the filters (not necessarily on the same day)
batted against both the Corning Ultraviolet
#5860-7-37 and Corning Red #2408-2-60.
These filters pass, respectively, no light of wave-
lengths longer than 390 mju, or shorter than 610
m/x. Most of the responses to the ultraviolet oc-
curred within 30 seconds. Strong positive re-
sponses were also obtained to filters transmit-
ting freely in intermediate spectral regions. The
following other species were tested with sim-
ilarly unequivocal results: Heliconiidae: Heli-
conius Sara (3 individuals) , H. ricini (2) ,Agrau-
lis vanillae (3). Danaidae: Danaus plexippus
(4). Ithomiidae: Tithorea mopsa (1). Papilio-
nidae: Papilio neophilus (3), P. anchises (2),
P. anchisiades (3). Pieridae: Eurema albula
(2), Phoebis sennae (2), Anteos maerula (1) .

Although it was expected that butterflies
would prove to be sensitive to the ultraviolet, in
common with the numerous insects of other
orders which have to date been tested, no re-
ports on such responses in Rhopalocera appear
hitherto to have been published.

The tests established conclusively that for H.
erato the visible spectrum extends from at least
the near ultraviolet at least up to 610 m/x, which
marks the extreme lower transmission limit of
the red filter employed. The experiments yield
no data at all, of course, concerning the rela-
tive brightness for the insect of the various
spectral regions, nor do they indicate whether
any of these regions are distinguished qualita-
tively, that is, as different colors.

B. Color Perception. Since existence of a
color sense in the heliconiids has never previ-
ously been tested, basic experiments were under-
taken to determine whether the capacity is pres-

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

171

ent in this family. The same general system was
used as that employed by von Frisch and others
with bees (von Frisch, 1948, 1950 and ref.),
and by Use with butterflies (1928). In all of
these, colored samples, in our case paper flowers,
were offered among a large number of gray
samples ranging in brightness from “white”
(positively ultraviolet), to “black.” No training
was undertaken in the present series, since un-
der certain conditions the species came spon-
taneously to the colors, and since it was only
desired to determine whether or not they could
distinguish color from any shade of gray, rather
than one color from another.

grays; apparently for them, as for bees, negative-
ly ultraviolet whites are colored. This never oc-
curred with the India ink series, the only re-
sponses to uncolored (for humans) flowers be-
ing to a model painted with Chinese white (zinc
oxide); as is well known, this substance has
minimal reflectance in the ultraviolet, and so, by
subtraction, apparently appears colored to the
butterfly.

In order to control the possibility that the
odor of the India ink was a deterrent to the
butterflies or that, conversely, some of the paints
or papers had an odor attractive to them, the
same ink, in various dilutions, was applied in

In the present tests, advantage was taken of
the tendency of H. erato to come to small color-
ed objects of certain forms— i.e., of flower-like
shape. As has been shown in bees (von Frisch,
1948, 1950 and ref.) and butterflies (Use,
1932.2), the most successful were small circles
with fairly numerous converging petal-like divi-
sions. Our most successful models were more
complex than those previously described, and
strongly three-dimensional. Flat ones, although
they evoked a few responses, were far less pop-
ular.

Series of these model flowers were fashioned
from Stoelting Psychological Test papers in 17
colors, and from papers painted in a variety of
opaque water colors and with Floquil paints.
All of these were analyzed spectrophotometric-
ally in the visible and near ultraviolet through
the courtesy of the Research Laboratories of
the Interchemical Corporation.

A series of 20 gray steps was also made, by
hand, from highly ultraviolet reflectant blotting
paper dipped in a measured succession of pro-
gressively higher dilutions of India ink with
distilled water. This laborious procedure was
necessary because a commercially manufactured
gray-step series with the necessary characteris-
tics could not be located. The only one found
in the United States (the Munsell series) with
a sufficiently matte surface proved upon spectro-
photometric analysis to have a very low reflec-
tance in the ultraviolet. In fact some results ob-
tained during one season were misleading until
this characteristic was ascertained, since the but-
terflies came rather freely to some of the lighter

Text-fig. 1. Pattern used in making
three dimensional paper flower. The
fringed strip is rolled up and fastened
at the base with scotch tape and a
paper clip. Natural size.

alternate rows of “petals” on single colored
flowers, while bits from the most popular col-
ored flowers were placed invisibly but open to
the air beneath as well as within some of the
gray models. The butterflies’ responses were un-
affected. The preference of the butterflies for
three-dimensional models made impractical von
Frisch’s use of a glass plate over flat models.

The technique of the tests was as follows:
they were run in two parts. First, a group of
experienced butterflies, well settled in an insec-
tary, was deprived of food for a few hours. The
best time for tests was between nine and eleven
o’clock on a sunny morning; therefore food was
usually removed the preceding afternoon. The
artificial flowers were displayed in checkerboard
fashion on an inverted square tray made of wire
netting with a one-quarter- or one-half-inch
mesh. The flower models were attached to the
meshes by paper clips. They were displayed both
with and without basal circles of black paper for
contrast, without noticeable difference in the
results. Sufficient tests of this aspect, however,
have not yet been made which would enable
the statement of reliable conclusions.

Freshly picked Lantana leaves were used to
cover the netting at the base of the models, since
experienced butterflies made extremely few re-
sponses to models without the added stimulus
of odor. Amyl acetate was also used with some
success, being sprayed on a pan of damp sand
or earth in which the models were set. In half
the tests no reward was provided; in the other
half, all the models— both colored and gray—
were fastened to half-dram vials filled with a
10% solution of white sugar.

172

Zoologica: New York Zoological Society

[40: 16

H

g

<

CL,

06

CL

<

Oh

lb

o

ei

u

z

<

h

U

Ui

nJ

lb

lb

06

H

U

lb

Oh

C/5

T3

a

CO

D

o

o

s E
g z

§

S o
cb

S 2
w I

3 O

I u
;| S

o

X)

CO

0> CO
O «4-4
CO b

3

B a

« E

E 43

wi

•- cS
3 i

tC 43
1 00
•S 3
3 2
o

E^

c/3

<Z)

CO Uh

w 0>

Cb

IS.

0) c5
CO O

T3

u>

'S

*S

Im

3

> ^
3:§

o *o

3> co

S

O u

■w (U

O^'S

6 w

S|

Ph E

s:

"S. •
-

c

Q> 4:3
3 OJD

cr -r

CQ

CO

O
O
DO O

o ^

II

S^‘"

3 3 3
o** o

*b

^ V

Cb

o

3
Ph cO

(X, Cb^
3 O ■

(1>

c

55

‘3 ^ n
3*^ 2

CO X) CO X) O CO bC

Ui ^

3

X)

4>

3

X

DO

.s

>

DO

c

X

G

*C0

Cb

o>

*R

o

o

a

X

0

u *0
o o

S

X >.

CO

'o S

Q> X

CO g

o> X

0>

cO

S 'O
CO a
^ 3
o> o
3

CT* c/5
CO 3

8-^
^ V

• -> >
bN

S -

*0

O X

BB

>%T3
tp 3

CO

3

li cr

DO

'o O

If

45 •- iT. .2? 3

PC ^ ffi

■H ^

3

S3®

E S

■s «

&

O -T3

CO X

X «
■&

3 ai

B S

<0 ^

1 3

> a
3

° ^
CO X
X 6J

w Ih
*rt

S 0>
X *0

O *SL
CO >
o CO

CbX
Cb 3
3

- «
3 -a

2 •«

> 0

a.

a>

2 g

«n M
<u >->

li CO
1-*

tS Of

!§13

> h

2 3

X o>

3 ^

X 00

.SP^

X <4-1
43 °

X c/5

:3 2 S

^ ^ .s ^

X ^

0) o
g-3
S V

Im

Ui u

«2 3

o> o>
c« 3

=> o-

CO
Cb
u O
> VT

*0

G

3 ^

«'2
Cb 9

CO ^
Cb

<b

o>

o

(L>

flb >
Cb ^
3^ O

^ X
Cb'C

o o
r^3 Cb
H g
Cb

^ tj

^ 3
C'3
3

3

3 B

^ .2

3 ^
*b u

C/5 X

jLi ^

.-H CO

< E

u

O

X

o

s

D

So

UJ

Z

0

<

2

b

0

€

w

u

Z

<

CJ

s

b

W

01

8

DO

C

G

O

«N

X

8

VO

O

ffl 60

PP

•3 o

so <N
a

o

\o

§

Tl*

o

<N

Tf

8

8

Cb

6

G

C/5

o\ r- 00 cN VO

0\ <S O X 0\ ^

DO o\ X o o

»ocsr^'0^^xr^<N

00 00 r- fN <N 00

00 00 On O X

o^ ^ Tf

«n <N
(S *n m

s

00 0\ o\

0)

Cb

-S

0

Cb

0

Cb

.S

.S

G

cO

G

3

cO

cO

Ch

Cb

Ob

Cb

Cb

Cb

r- o 00

T3 0) « 3 *3

o Cb Cb .5 o

X CO 3 3 7^

2 G G G

H 0 (D <b

2 000

4- j< & & M

^^^S>3333333
0 _0 0333333_33

■>i>>S5mSmmSS

-HfSm'0'‘o'or~600\o^

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

173

'o ts o

'O Os

■o o\ 0\

00 cn o

On m 00 O

^ ^ ^ rn ^

O o

fS

CN r*» r'l

trj Tt

^ VO 00

oo 00 Tj- O On O

NO r- oo m NO

00\f^^Of^*^vow->0\<St^^ON

00 00

^r^^rsoor-«o«r)^ON

oo NO On

Nor^oo*— '’ONO'^‘riTj-OfN

O On

Nor-^r-Nor*-r^ONONO'^NOTj-i-H

ON'00\r**v*>oor^Ov->»nTfr^T}*m

NO 00 00

ooNf^'Ooot^'nor^ONOv^'^ooTfNo

*'^ot^OfsoNoor^'Ooor-«noNoovo‘'NTtoNTi-vo

OOONNONOOONONOONO^t'-VNTfO’^NOm^

VO

ONOOfNNOOOr^VNOt^<NONOmO^NO

00

r^ONONNOoor^<noNt^r'ic^^vor-Tt

Tt NO r- O

'^■^oo’^f^)^«^(^^(3^v^oo^^Not^t^<^oo'Ovo^^^«rN^<^T^Ti■o

•5 -9

C«

CU CL,

ri *S W O O

U|

u

Cl,

-S

u,

0>

CL,

u,

0>

Oh

1m

Dh

2

_e

5

O

u,

<l>

CL,

1m

a.

u ^
CL, .5

.a

o

u,

o

CU

C3

*c5

d

C3

d

c3

'3

a

cd

cd

cd

cd

cd

O.

CIh

Oh

CL,

Oh

CU

CL,

(3li

D

Ph

CL.

Ph

Ph

Pk

U

Ph

8 o

2 ft-

H

•!>

Oh

c3

CL,

G C

g 8
a o

a S
8 5
M "S

i «

:§ 8

a >H O >H >H >- >H

^fO-sJ-l/1VOI^OOO\

R

0/}

C

rt

u>

O

o

"B

0> 0> O 4> 0^

O 4> Q> U O 4>

ClO OX) OX) OjO 0/)

ccccad^^^

csrjrtgrtg-g-g-g

>->-OOOOOOe!C!^0!!

(S <S fS

o c

3 «

” 8

■S "S o a s ""

eJ Pi > Oh pL,

Table 1. Spectral Reflectance of Papers, Pajnts and Textiles Used in Heliconius Experiments.

Explanation: From spectropholometric curves furnished by the research laboratories of the Interchemical Corporation, New York, N. Y. Sources and

general type of the materials used were as follows:

Papers: a. Stoelting Psychological Test series (a smooth-finish matte surface).

b. A series of school construction papers (a rough-finish matte surface).

PainLs: a. Floquil “Flopaque" (a quick-drying opaque lacquer).

b. Ricbart opaque water colors.

c. "Stroblite" fluorescent paint. (Translucent). (Bright pink).

Textiles: a. Light-weight canvas.

b. Heavy, opaque, felt-like cotton.

The papers were used for experimental flowers and markings on butterfly models; opaque lacquer used especially for painting living butterflies, but also
for paper flowers; opaque water-colors principally for paper flowers; canvas and felt served as foundations for butterfly models.

Also used were (1) white paper and paint with high ultraviolet reflectance which approached that of wavelengths in the visible; (2) zinc oxide paint with
minimal ultraviolet reflectance; and (3) a series of 18 neutral grays, with positive ultraviolet reflectance (see text, p. 171).

REFLECTANCE (% OF MAGNESIUM OXIDE)

Ultraviolet

Violet

Blue

Blue-

Green

Yellow

Orange

Red

green

Sample

Color

Material

340

360

380

400

420

440

460

480

Wave-length (m^)
500 520 540

560

580

600

620

640

660

680

700

Violet

Paper

18

20

25

26

27

30

28

20

13

7

8

4

9

11

13

18

25

Violet

Paint

33

31

38

43

46

45

36

23

14

9

8

8

9

12

19

35

57

74

Violet-blue

Paper

9

10

14

19

22

23

22

18

14

11

9

8

10

17

23

27

33

4

Blue

Paper

9

10

16

35

46

52

51

39

24

14

10

7

6

6

6

7

12

18

Blue

Paint

8

12

23

28

29

28

25

21

17

15

12

11

10

9

8

7

Blue

Paint

52

55

59

56

61

66

64

52

37

24

16

12

11

10

11

12

15

20

28

Blue

Cloth

21

24

30

31

35

36

36

33

29

25

21

18

16

15

15

16

17

19

21

Blue

Paper

17

20

24

24

25

38

45

46

45

40

30

20

12

9

7

6

12

23

9

Blue-green

Paper

10

10

14

23

26

29

35

46

54

53

47

40

32

27

24

22

24

28

32

10

Bluc-grccn

Paint

7

14

22

24

25

28

32

33

31

28

25

21

18

16

14

13

12

12

11

Bluc-grccn

Cloth

18

17

15

14

14

16

22

30

34

32

27

22

17

13

12

13

14

16

18

j

12 Green

1 3 Green

14 Green

15 Yellow-green

16 Green-yellow

17 Yellow

18 Yellow

19 Yellow

20 Yellow

21 Yellow

22 Yellow-orange

23 Orange

24 Orange

25 Orange

26 Orange

27 Orange

28 Orange-red

29 Red

30 Red

31 Red

32 Red

33 Red

34 Violet-red

35 Pink

36 Pink, fluo-

rescent

Paint 6 6

Paint 14 14

Cloth 8 8

Paper 14 22

Paper 13 14

Paper 7 7

Cloth 9 9

Paper 9 9

Paint 5 6

Paint 8 8

Paper 7 7

Paper 6 5

Paper 7 9

Paint 7 7

Paint 13 12

Cloth 8 12

Paper 6 6

Paper 6 6

Paper 7 7

Paint 4 4

Paint 15 20

Cloth 3 4

Paper 14 16

Paper 14 17

Translucent 10 10

Paint on
White Paper

6 5 5 6

13 11 10 18

8 7 7 8

24 20 20 21

14 13 12 13

9 8 9 10

8 9 8 9

12 16 17 17

6 6 6 6

8 8 8 8

7 6 7 7

5 6 5 5

10 10 10 10

7 6 6 7

12 11 10 10

10 7 6 6

6 5 5 5

5 4 4 4

10 10 9 8

4 4 4 4

16 16 16 16

4 3 3 3

22 21 21 23

18 17 16 16

15 21 25 31

7 11 17 18

18 27 38 38

11 17 22 23

24 30 45 56

15 19 35 61

14 25 41 59

11 15 18 34

19 26 36 44

6 7 9 25

9 11 22 50

7 7 9 25

5 6 6 9

8 7 7 7

7 7 8 11

10 10 11 12

5 6 6 8

5 6 6 7

4 4 4 5

7 6 5 5

4 4 4 4

13 11 10 9

3 3 2 2

22 16 11 8

14 15 11 10

34 27 22 20

IS U

31 23

20 16

58 45

67 62

67 70

41 46

51 57

52 65

75 83

54 68

10 15

9 19

17 29

14 23

10 22

7 9

5 6

6 7

5 5

9 8

2 3

7 6

9 11

19 20

9 8

17 13
12 10

40 38

55 51

71 71

50 53
62 66

71 73

84 85

72 74

33 58

39 55

45 61

51 75

41 55

16 39

8 19

II 17

8 13

18 38

4 10

10 22

19 40

20 40

7 6

11 9

9 9

37 38

48 46

70 70

55 58

69 72

74 75

85 86

75 76

70 73

62 66

71 76

82 84

60 62

65 75

45 65

26 38

33 57

70 82

25 40

45 56

64 70

63 68

6 6

9 12

9 10

41 44

50 59

72 75

60 62

74 77

76 76

88 90

77 78

75 77

69 72

78 79

85 88

63 65

78 79

72 75

50 59

67 72

86 88

50 57

61 64

73 75

72 75

6

18

11

48

69

77

64
79

77
91

79

78
74

80
90
68
80
78

65
74
90
62
67

76

77

Zoologica: New York Zoological Society

174

Zoologica: New York Zoological Society

[40: 16

In preliminary tests five models were dis-
played, representing general spectral regions:
blue, blue-green, yeUow and orange-red, all ex-
cept some blues having minimum reflectance in
the ultraviolet. As always they were exhibited
with the complete series of gray models, plus
one of Chinese white. In subsequent tests four
models, of roughly similar hues but of varying
brightnesses and texture, were displayed among
the grays, such as four yellows, or four orange-
reds and reds. Generally each test was limited to
15 minutes. Hungry butterfles on bright days
would return and probe a paper flower— whether
or not sugar water was attached— a number of
times; well-fed individuals in cloudy or oppres-
sive weather would give very few or no re-
sponses, even though they were repeatedly flush-
ed from resting positions. Responses were di-
vided into three parts (Table 2), “dips,” in
which the butterfly changed course and fluttered
down within three inches of a particular model,
alightings and probes; in the latter actual and
often vigorous attempts were made to feed,
with the proboscis uncoiled; when no sugar
water was furnished the attempts were some-
times much prolonged.

Second, inexperienced butterflies, which had
never previously fed on or seen any flowers, or
even any other butterflies, having been kept in
isolation cages (p. 187), were introduced singly
into an insectary. Individuals, especially males,
allowed to go about fifty-four hours without
food upon emergence, sometimes came spon-
taneously straight over to the experimental tray
and probed repeatedly at several differently col-
ored models in succession, a variety of hues
being displayed among the grays. These re-
sponses were without the added attraction of
either Lantana leaves or amyl acetate, and no
sugar water was provided. In fact, no chem-
ical stimulus was ever employed with the
inexperienced individuals; therefore their posi-
tive responses were all to visual characteristics
alone. However, younger butterflies had to be
tested by fastening the test flowers in succession
on the end of a stick; the model was then
brought, sometimes several times over, slowly
up to the head of the resting butterfly. The order
of color presentation was varied in different in-
dividuals, and with the same individual in dif-
ferent test sessions, without allowing either real
food or other insects in the insectary until after
the conclusions of the tests.

The results of both these groups of prelimi-
nary tests to determine the presence of color
vision in H. erato were conclusive:

1. Neither experienced nor inexperienced but-
terflies ever dipped toward, settled on or probed
an uncolored paper flower, except for the model

painted with negatively ultraviolet Chinese white
(zinc oxide) .

2. Positive probing responses were obtained
for literally all regions of the spectrum, in-
cluding some with relatively high ultraviolet re-
flectance, except blue-greens and greens, and
by inexperienced as well as experienced but-
terflies, without the addition of odor. Feeding
response to color is therefore innate.

3. The largest numbers of responses were re-
ceived, in this order, by yellow, orange-yellow,
orange and Chinese white. Blue, whether posi-
tively or negatively ultraviolet, was next to blue-
green and green least popular of the colors. No
actual proportions of responses to the different
models can be given; the variations of presenta-
tion were not made with sufficient care to en-
sure a random, or other system of rotation, and
a given butterfly, or group of butterflies, in-
variably ceased to respond before an entire
series could be run. Table 2 gives an adequate
sample of typical series of responses.

4. The preferences held among the inexperi-
enced as well as the experienced and show that
the attraction for “yellow” in food is innate and
not only a learned association with such favorite
food-flowers as Lantana, Bidens, etc. It is, of
course, by the terms of the experiments, not yet
known whether the unpopularity of the blue and
green regions is due merely to a low retinal
sensitivity. It is interesting to recall that in Use’s
experiments, vanessids— which of her groups are
nearest the heliconiids— found the blue and yel-
low regions of the spectrum most attractive. It
will be noted that blue is not a favorite with the
heliconiids. The papilionids and pierids on the
other hand came most freely to red. Casual but
repeated observations in the field, garden and
insectaries in Trinidad support this red prefer-
ence in the tropical papilionids and pierids.

V. Behavior of Imagos: General Account

A. Introduction. The following account of
the adult behavior of Heliconius erato is based
principally on the activities of specimens ob-
served in the Trinidad insectaries. However, all
the general observations, including food flower
preferences, time of day and meteorological
conditions governing roost-leaving, feeding,
courting, egg-laying and roosting, have been
checked a number of times in the field, by other
staff members as well as by the author. These
checks were found to agree with the more de-
tailed observations made possible by the insec-
taries.

B. Habitat and Flight Habits. Heliconius
erato flies typically along the edges of seasonal,
lower montane and swamp forests, and of well-

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

175

Table 2. Feeding Responses of Experienced H. erato hydara to 4 Colored Artificial Flowers
AMONG A Series of 18 Graduated Grays and Zinc Oxide White.

The data below are typical of the results obtained in about twenty similar tests. See text, p. 170 ff. and
Text-fig. 1. Numbers under colors refer to samples analyzed spectrophotometrically in Table 1, p. 172.

Test conditions: Five hungry individuals were used in each 15-minute test. No individual had been
tested within the preceding five days and all were habituated to insectary conditions. Tests were con-
ducted only on calm, sunny mornings between 9:30 and 11:30 A.M. No food was furnished with the
flower models, although Lantana leaves were placed evenly around their bases. The models were arranged
in the usual checkerboard (Roman square).

In no test was there any response whatever either to models of the gray series, including posi-
tively ultraviolet white, or to Lantana leaves.

Test a

Dips

Alightings
Proboscis probes

Total

Responses to Violet, Blue and Zinc Oxide White

Violet Violet Blue Blue

(No. 1) (No. 2) (No. 6) (No. 13)

- 1 - 1

- 1 - 1

Zinc Oxide
White

2

2

4

Test b

Responses to Blue, Blue-green and Zinc Oxide White

Blue

Blue

Blue-green

Blue-green

Zinc Oxide

(No. 5)

(No. 9)

(No. 10)

(No. 12)

White

Dips

1

—

—

—

3

Alightings

—

1

—

—

2

Proboscis probes

—

—

“

4

Total

1

1

-

-

9

Test c

Responses to Yellow and Zinc Oxide White

Yellow

Yellow

Yellow

Yellow-orange

Zinc Oxide

(No. 19)

(No. 20)

(No. 21)

(No. 22)

White

Dips

2

8

1

13

2

Alightings

-

-

1

2

1

Proboscis probes

16

2

21

6

2

Total

18

10

23

21

5

Test d

Responses to Orange, Red and Zinc Oxide White

Yellow-orange

Orange

Orange-red

Red

Zinc Oxide

(No. 22)

(No. 23)

(No. 28)

(No. 29)

White

Dips

7

4

2

1

1

Alightings

1

2

-

-

-

Proboscis probes

9

5

-

-

-

Total

17

11

2

1

1

established second growth. These characteristic
localities occur along roads, open trails, streams
and old, small clearings; the butterfly is also
common in overgrown citrus and cocoa cultiva-

tions. It does not fly in the depths of thick
forests, since partial sunlight is an essential. In
the Arima Valley it has not been seen above
1,200 feet, although it was fairly common at

176

Zoologica: New York Zoological Society

[40: 16

Rancho Grande, Venezuela, at 3,600 feet. It
is a low flier, usually fluttering between about
three and seven feet off the ground, and rarely
is seen as high as 20 feet. Females, when about
to lay, skim along a few inches to two feet above
the ground, since the local food plant is usually
low-growing.

In the insectaries a combination of ample
shade and sunlight is required plus a high hu-
midity and an abundance and variety of growing
herbs and shrubs. Given these conditions H.
erato thrives in captivity better than any of the
remaining fifty-odd species so far tested, al-
though all the heliconines do well.

C. Relation of Behavior to Meteorologi-
cal Conditions. Temperature, light and hu-
midity are all important in the behavior of these
completely diurnal butterflies. Temperature is
the critical factor which governs their first
morning activity, while light intensity controls
roosting in the late afternoon. At Simla they are
always inactive, regardless of the light intensity,
at temperatures lower than 67° F. (19.4° C.);
the usual minimum for activity however is 70° F.
(21.1° C.) . This temperature has never, between
December and lune, which includes the coolest
season, been reached later than 8 A.M. The
usual time of first morning activity changes with
the seasons. In January it usually occurs between
7 and 8, in April between 6:30 and 7:30, and
in May and June the first butterflies frequently
leave the roost at 6:45. In hot weather, when
the early morning temperature does not fall
below 70° F. (21.1° C.), light is the crucial
factor. Ground light must usually be around
3.2 by Weston exposure meter, south light 25-50,
and overhead sky light (through the insectary
roof), 25-100; however ground light may be as
little as 1.6 and sometimes registers more than
6.5 before the first flights are made. A single
butterfly once flew, at 5:45 A.M. in late April
at a temperature of 73° F. (22.3° C.) when the
ground light registered only .4, the south light
3.2 and the sky between 3.2 and 6.5. For com-
parison, noon conditions on a bright sunny day
with a typical rainforest combination of some
clear sky and some clouds show the following
readings inside the insectary: ground 25-50,
south 200, overhead sky 800. Full morning ac-
tivity by all healthy individuals is reached when
the temperature is at least 73° F. (22.3° C.)
and the ground at least 13 Weston in some areas,
south light 50 and overhead sky more than 50.
Activity is always higher on sunny days.

The butterflies become inactive in rainy, dull
or windy weather. They also often become in-
active up to an hour before a heavy squall, even
while the weather remains calm and bright. An
attempt to link this with barometric pressure

failed. On hot, dry afternoons the butterflies
tend to fly high and erratically when the hu-
midity drops below fifty per cent., batting against
the roof and becoming exhausted. Sprinkling the
insectary at once corrects this condition.

In bad weather, roosting— that is, hanging up-
side down from a twig or tendril, with the wings
closed— may begin as early as 2:30 in the after-
noon. On bright or moderately cloudy days,
however, the &st butterflies may start gathering
in the roosting place around 4:45 P.M. m the
average month of April, while others are still
feeding. The majority hang up around 6:00 P.M.
and all are in their final positions at 6:15. The
time varies only about twenty minutes one way
or the other from these hours on the longer and
shorter days around the solstices. The tempera-
ture records during the going-to-roost period
have ranged from 80° F. (26.7° C.) down to
73° F. (22.8° C.), and, as was said above, did
not apparently affect the roosting behavior,
which was primarily controlled by the decreas-
ing light. When the ground-vegetation light re-
corded between 6.5 and 13 Weston, with the
south over 25 and the sky overhead more than
50, nearly all the butterflies continued to feed,
and courting sometimes took place. At intensi-
ties less than the above, the butterflies gradually
gathered to roost. In one instance, an individual
fed accurately and subsequently found the roost
at the extremely low intensities of ground-
vegetation .4, south 3. 2-6.5 and sky 13. Feeding
with a ground reflectance of 1.6, south 6.5-13
and sky 13-25 was not uncommon. It will be seen
that these intensities are sUghtly lower than those
at which the first morning activity usually oc-
curs, when the tempertuare is nearly always
lower than the low of 73° F. (22.8° C.) recorded
for going to roost.

The conditions controlling first morning ac-
tivity and roosting are summarized in Table 3.
Although observations on other species are in-
complete, it may be said here that forest-living
ithomiids are active at both lower light inten-
sities and lower temperatures, while Heliconius
Sara and H. ricini, which fly in slightly more
open areas than H. erato, require either more
light or higher temperature or both for activity.
These characteristics lead, in the insectaries, to
the ithomiids flying both earlier in the morning
and later in the afternoon than erato, while sara
and ricini both usually leave the roost later and
hang up on it earlier than does erato.

D. Feeding. H. erato is altogether a flower
feeder, never coming to fruit. From observations
on feeding preferences in the insectaries, garden
and in the field. Table 4 was compiled. Lantana
has been observed to be strongly attractive in

1955]

Crane: Iniaginal Behavior of Heliconius erato liydara

111

Table 3. Relationship of Diurnal Activity to Temperature and Light in Heliconius erato hydara

Based on observations made in Trinidad, between December and June, altitude 800 feet. All

light readings made under wire netting roof of insectary, with Weston photometer pointing directly over-
head. For fuller data, see text.

Morning

Afternoon

(First activity: 5:45-8:00 AM)

(Last activity: 4:30-6:00 PM)

Light always more than 25 Weston

Temperature always more than 73°F.

Activity

Temp.

Temp.

Temp.

Light

Light

Light

< 67° F.

o^

1

>73°F.

> 100 W.

25-100 W.

< 13

Full

—

—

X

X

—

—

Slight

-

X

-

-

X

-

None

X

_

—

—

—

X

the field in both Venezuela and Surinam, as tvell
as in Trinidad, and is far and away the favorite
in the insectaries, being used as the staple diet.

Feeding takes place all during the day, but is
especially active on bright mornings between 9
and 1 1 A.M. When left undisturbed in the insec-
taries, neither sex seeks food until 20 to 24 hours
after emergence; however, when presented with
food some individuals will uncoil the proboscis
and feed about six hours after emergence.

E. Social Behavior. Social behavior may be
divided into three categories, namely, courtship,
social chasing and roosting. The following para-
graphs are descriptive statements; the activities
will be analyzed in subsequent sections (p. 183
ff.).

1. Courtship. In the normal sequence, two
alternative beginnings are possible. First, a male
gives chase to a passing receptive female and
pursues her closely. She usually settles on a
fairly broad green leaf within a few moments,
sometimes apparently urged to alight by the
male’s flying slightly above her and by actually
brushing her wings with his on the downbeats,
thus forcing her downward. Alternatively, a male
approaches a female at rest on a leaf; her wings
may be open or closed. Females which have
already been stimulated by unconsummated
courtships often will remain on the same leaf
for hours, responding promptly to subsequent
males.

At the male’s approach, the female usually
opens and closes the wings slowly several times.
If she is receptive and begins to court, however,
both pairs are held closed over the back. The
male begins (Plate I, Figs. 1-3) by flying in
short, repeated spurts close behind her, hovering
in place, his forewings sometimes almost touch-
ing her hindwings and creating a current of air
against them. Occasionally on a downbeat the
male wings actually cover those of the female
(Plate I, Fig. 4). During this stage, as well as

in any preliminary chasing of the female, his
fore- and hindwings are not separated to expose
the silvery friction surfaces which apparently
distribute the products of the scent scales (see
p. 181). The male harpes are not in evidence,
the tip of the abdomen remaining tightly closed,
and no odor is detectable in this sex at this stage
to the human observer.

The response of a receptive female to this
buffeting is shown in two ways: first the abdo-
men is elevated and a gland, bright chrome yel-
low in hue, is extruded dorsally at the junction
between the penultimate and distal segments.
Two bulbous excrescenses project from it, one
on each side of the dorsal midline, and are
arhythmically inflated and partially deflated with
waxing and waning stimulation. Below and be-
hind them is the pair of tiny “stink clubs,” de-
scribed, along with the gland, in other heli-
conines by Muller (1877, 1912), and in this
species by Eltringham (1925). These clubs are
extremely difficult to see, even with a hand-lens,
in the actively courting butterfly, and, indeed,
are not exserted except during maximum excite-
ment; however they are unmistakably in evi-
dence in some frames of our motion pictures of
courting females. The clubs show up as yellow-
ish brown, in contrast to the chrome yellow of
the gland proper.

Usually no odor is detectable to the human
observer during this response of the female;
rarely a musky odor is evident at very close
range in individuals less than about two days old.
Apparently this is the same odor which is de-
tected when females are allowed to emerge from
the chrysalid in a closed box; young males have,
to human nostrils, a similar odor. Only in fe-
males which have been already mated and are
being persistently courted on the following day,
or, rarely, in which advanced courtship is dis-
turbed so that the insects fly off with the female
gland exserted, is a witch-hazel-like odor dis-

178

Zoologica: New York Zoological Society

[40: 16

cernible during courting. This odor, well-known
in the species, has no apparent effect on the
courting male. (cf. “Defense,” p. 180). Ford
(1945, p. 96) has remarked that, unlike the
products of male scent scales, the chemical at-
tractants of female butterflies have not been
reported as perceptible to human observers. Ex-
cept for the circumstances just noted, this is
true also of H. erato.

As the tip of the abdomen is elevated and the
gland exserted, the anterior margins of the fore-
wings, as they are held closed above the back,
are pressed closely together. The more posterior
portions of the forewings are allowed to bell
outward. Simultaneously the hindwings are part-

ly opened and rapidly vibrated for a second or
two with brief intervals between vibrations.

The next stage of courtship. Stage II, may be
omitted or strongly curtailed by excited butter-
flies when the female is highly receptive. Never-
theless, it is usually both well-marked and
characteristic. The male shifts position from the
rear forward, so that his flapping continues
above her and, especially, immediately in front
of her. He continues to face in the same direc-
tion as the female, or sideways, and he may or
may not at times touch her forewings with the
tips of his hindwings. His fore- and hindwings
are now separated, sometimes widely, exposing
the silvery friction surfaces (Plate I, Fig. 5;

Table 4. Feeding Preferences of Heliconius erato hydara in Trinidad.

(Representative examples of both wild and garden species are included. Colors are roughly indicated of
entire inflorescence, including bracts or rays, to give an idea of their variety. None of the whites and very
few of these colors reflect more than minute amounts of ultraviolet. The important roles played by form
and odor in the relative attractiveness of the different species are not considered here.)

Loganiaceae.

Verbenaceae.

Rubiaceae.

Boraginaceae.

Compositae.

Zingiberaceae.

Orchidaceae.

Asclepiadaceae.

Solenaceae.

Gesneraceae.

Cucurbitaceae.

Musaceae.

Bignoniaceae.

Rubiaceae.

Papilionatae.

Convolvulaceae.

Bignoniaceae.

Malvaceae.

Plumbaginaceae.

Campanulaceae.

Acanthaceae.

Frequently Visited

Buddleia variabilis Faranchet.
Lantana camara L.

Verbena spp.

Cephaelis tomentosa Willd.
Warszewiczia coccinea. (Wahl) Kl.
Hamelia erecta Jacq.

Cordia cyclindrostachya. R. & S.
Bidens pilosa L.

Senecio confusus Britten.

Buddleia. Purple.

Lantana. Yellow & orange.
Vervain. Purple; white.

Wild Ipecancuanha. Yellow & red.
Wild Poinsettia. Yellow & red.
Wildclove. Orange.

Black sage. White.

Spanish Needles. Yellow & white.
Gem of the Rio Grande. Orange.

Occasionally Visited

Costas spiralis Rose.
Epidendron fragrans Swartz.

Oncidium luridum Lindl.
Asclepias curassavica L.
Browallia americana L.
Tussacia pulchella Rchb.

Momordica charantia L.

Scarlet Cane Reed. Yellow & red.
Purple Streak Orchid.

White with purple streaks.

Brown Bee orchid. Yellow & brown.
Milkweed. Yellow & orange.

False violet. Yellow & violet.
Harlequin Flower. Yellow, orange &
purple.

Carilla. Yellow.

Rarely Visited

Heliconia hirsuta L.

Heliconia humilis Jacq.

Tabebuia serratifolia (Vahl).

Ixora coccinea L.

Paradise flower. Yellow & orange.
Swamp ballsier. Yellow & orange.
Poui. Yellow.

Ixora. Red.

Apparently Never Visited

Erythrina micropteryx Poepp.
Convolvulus spp.

Tabebuia pentaphylla Hems).
Hibiscus spp.

Plumago capensis L.
Centropogon surinamensis (L.)
Pachystachys coccinea Ns.

Mountain immortelle. Orange-red.
Morning glory. Blue; yellow; white.
Poui. Pink.

Red; pink; orange; yellow; white.
Plumbago. Blue.

Presl. Crepe Coq. Red.

Black Stick. Red.

1955]

Crane: linaginal Behavior of Heliconius erato hydara

179

Plate 11, Fig. 11) . Both fore- and hindwings are
in rapid motion (Plate 11, Fig. 12), unlike nor-
mal flight, where the hindwings are almost mo-
tionless. A flowery fragrance can rarely be de-
tected by the observer at this stage, probably
emanating from the now-exposed friction sur-
faces. The tip of the male abdomen is still tightly
closed. When both sexes are subsequently ex-
amined, both fore- and hindwings, especially
near the bases and along the anterior margins
of the hindwings, as well as the thorax, are found
to be similarly fragrant. The female fragrance
develops even in uncourted individuals, although
to a lesser extent, and hence must be a product
on her own, rather than merely sprayed on by
the male in Stage II (see p. 182).

During Stage II the female continues to vi-
brate the hindwings while the anterior margins
of her forewings remain closely apposed. When
Stage II is prolonged, however, in its extreme
form where the male fans her from a position in
front of her, rather than directly above, she
withdraws the abdominal scent gland and partly
lowers the abdomen. Her antennae are occa-
sionally lowered during Stage II, again prin-
cipally when the male is courting from in front
of her.

The final sequence. Stage III, immediately
precedes copulation. It begins with the male’s
alighting and moving backward beside the fe-
male, on either side, but more frequently on the
right. He still faces in the same direction as she,
and he usually comes to rest with his eyes near
the level of her thorax. He continues flapping
his wings with the friction surfaces exposed and
she once more elevates the abdomen and ex-
trudes the scent gland. Fragrance is most likely
to be detected at this stage, but its source cannot
be determined because of the insects’ activity.
If Stage II has been omitted, the male simply
alights beside her, his friction surfaces now
being exposed for the first time. In any case he
extrudes the harpes, twists his abdomen side-
ways, and seeks to curve it forward between the
posterior margins of the female’s now folded
hindwings until the harpes can engage the fe-
male’s abdomen. In fully receptive females the
abdomen is by now lowered, the scent gland
being simultaneously withdrawn. The harpes
grasp the abdomen’s lower tip, gripping it by the
latero-ventral portion of the penultimate seg-
ment.

During copulation, which usually continues
more than an hour and rarely overnight, both
pairs of wings in both sexes are normally closed
and held erect over the body in typical daytime
rest position. The male swings promptly around
so as to face in the opposite direction from the

female. When disturbed a short flight is made
to another resting place, the male carrying the
inert female. No odor is apparent during copu-
lation except for an occasional faint fragrance
when, if slightly disturbed, the male slowly
waves his wings.

Persistently courting males which are not ac-
cepted by females, or old males which have
made repeated unsuccessful attempts to engage
the female’s abdomen, or males which have been
courting females of another species (namely,
Heliconius melpomene, H. ricini, H. isabella,
and Dry as julia) sometimes come to rest in front
of the female, facing her, which is a position
atypical for this species. Alternatively, they may
sit quietly beside her, their heads on a level with
the female’s head, or, more often, her thorax.
From either position the male sometimes pal-
pates the female’s head, antennae or thorax with
his antennae, and, rarely, with his uncoiled pro-
boscis (Plate III, Figs. 13, 14). Another type
of atypical behavior often occurs among the old
males, which court facing the female, and flying
both in front of, above and behind her, in ir-
regular alternation; the males’ friction surfaces
remain exposed regardless of their position.
None of the atypical behavior described in this
paragraph occurs in the course of a normal
courtship. It appears exceedingly likely that dis-
placement behavior is involved, particularly
when the proboscis is inappropriately extruded.

Courting may take place at any time during
the day but is especially common during the late
morning and after 2:30 P.M. Males almost
never court females engaged in feeding at
bunches of flowers, although they are not simi-
larly restricted if the female is on an isolated
blossom. There is no courting on the roost (see
below) .

Sometimes a receptive female which has not
been courted will flutter the hindwings at the
close passing of almost any species of butterfly;
if she is more than two days old and still un-
mated, she may freely chase passing H. erato
of any age; younger females, however, fly and
feed very little.

Males and females both can mate at least
twice, the males from the second to at least the
thirty-first day after emergence. Some individ-
uals court persistently but unsuccessfully for
another month. Females have been seen mating
from 45 minutes after emergence through the
seventh day. The hindwing flutter and extrusion
of the scent gland in response to courting takes
place through the eighth day. However, the an-
terior forewing margins are held less and less
tightly together following the third day, and the
abdominal gland is extruded less and less com-
pletely.

180

Zoologica: New York Zoological Society

[40: 16

2. Social Chasing. Chasing-and-circling, which
seems not to be directly of a sexual nature, is a
frequent activity of healthy male butterflies
throughout their lives and of females which have
completed egg-laying. The males and old fe-
males chase both one another and non-receptive
females indiscriminately in short flights, some-
times with mutual circlings. Young males and
newcomers to the insectary are especially subject
to chasing. There is no apparent attempt to force
a chased butterfly to alight, as in preliminary
courtship, and the roles of chaser and chased are
frequently reversed several times during a flight.
The abdominal gland of the female is not exsert-
ed, nor are the male harpes, the friction sur-
faces are not exposed and there is no detectable
odor of any kind. The flights are never the result
of two or more males courting a single female,
although this multiple courting often takes place.
The chasings and circlings invariably end quite
simply with the two or three butterflies involved
going their separate ways.

Sometimes a butterfly will approach a resting
individual, which is not a receptive female, from
the rear, as in the first stage of courtship. The
resting butterfly then opens and closes the wings
slowly a few times, whether they have been held
open or shut, whereupon the approaching indi-
vidual flies away.

Except for the apparently non-courtship rela-
tionships described above, no approach has been
found to inter-male threat display or fighting,
nor is there evidence in the insectaries of terri-
toriality or of a dominance order. Males and
old females are far more active in both chasing
and feeding than females which have not com-
pleted egg-laying.

3. Roosting. During the late afternoon H.
erato follows the family habit of gathering in
groups for the night. In the insectaries as in the
field the butterflies often crowd the same dead
twigs or vine tendrils, night after night, hanging
upsidedown, the wings closed. Species of the
same genus may roost with them. In the insec-
taries males and old females tend more to
occupy crowded roosts together, while younger
females, except when at their most receptive
(the second through the fourth days), show less
gregariousness. However, these receptive fe-
males usually roost on the more crowded
perches, or sometimes next to a somewhat iso-
lated male. No courtship takes place on or near
the roosts, although there is always a great deal
of activity, buffetting and pushing and jockeying
for position, as the butterflies gather.

F. Defense. No organized work on the dis-
tastefulness of H. erato has yet been undertaken
in the course of the present study. However,
many casual tests have been conducted in our

laboratories, both at Rancho Grande, Venezuela,
and in Trinidad. Almost all of these confirm the
heliconiid reputation of distastefulness to their
natural predators. Butterflies of both sexes and
all ages have been refused by various individuals
of the lizard, Polychrus marmoratus. Some were
seized by the wings or body and dropped, while
others were disregarded altogether. In some
butterflies the scarlet band was cut off by the
observer before the insect was offered to the
lizard. Individuals were also refused by the frog,
Hyla maxima. Usually, but not always, they
were refused, or examined and dropped, by cap-
tive capuchin monkeys. The butterflies were
always either dropped or disregarded by the
mantids Stagmatoptera septentrionalis, Stagmo-
mantis Carolina and Oxyopsis rubicunda. All of
these laboratory test animals, from lizards to
mantids, freely ate Lepidoptera of unprotected
families and comparable size following the re-
fusal of Heliconius erato. Ponerine ants, how-
ever, attacked and ate ailing Heliconius in the
insectaries, and various non-ponerines scavenged
them freely. Also, several species of epeirid
spiders were major insectary enemies.

In H. erato three or four different odors may
be distinguished by the human observer. All of
them probably play roles in the insect’s sexual
or social patterns or in both. At least two of
them and possibly all appear to be concerned
also in the butterfly’s defense. These odors and
their sources will now be considered in turn.

1. Abdominal Glands. It has been generally
held in the literature (see ref., p. 169) that the
glands of the penultimate abdominal segment in
the female, and in each harpe of the male are
aposematic in function, discouraging attack by
emitting an unpleasant odor. The scent has been
variously compared with that of carbylamine,
phenyl carbylamine, witch hazel, cashew oil
from the shells, and, when less strong after a
lapse of time, sweet briar. Some have considered
the smell to be exceedingly unpleasant, others
pleasant. However, as Eltringham (1925) point-
ed out, an odor which seems agreeable to one
person may be obnoxious to another. For ex-
ample, he himself did not like the scent of witch
hazel, which is rated either “pleasant” or “not
unpleasant” by all five of the present members
of the Tropical Research staff.

The following account of the abdominal
glands is derived from our Trinidad studies.

In H. erato hydara these glands in both sexes
sometimes give off a strong odor which is prac-
tically never discernible in the course of court-
ship, and never at the roost or in social chasing.
To us it always resembles witch hazel in the
living or freshly dead insect. As noted by pre-
vious observers of this and other subspecies

1955]

Crane: Imaginal Behavior of Heliconhis erato hydara

181

(Longfield, 1926; Collenette, 1929), the odor
is stronger and occurs more frequently in fe-
males than in males. In the Trinidad form it
appears in the female only after mating; it can
be detected as a faint fragrance, similar to that
of the wings (see below), almost immediately
after copulation. It then develops slowly, becom-
ing strongest in the midst of the oviposition
period, then declines and has not been detected
after the thirtieth day of the imaginal period.
It does not deter either subsequent courtings or
at least one additional mating. It never develops
in unmated females, even after death; however,
it may be strong in the glands, excised and
bottled, of males and mated females for many
weeks although, as will be discussed below, it
alters its witch-hazel-like characteristics. The
odor can only rarely be detected in living males
of any age; those which emit the odor in
response to seizures are usually moderately but
not extremely young.

The abdominal glands are everted in apparent
defense only when the insect is seized either near
the base of the wings, or by the head, thorax or
abdomen. The seizure may be by any means,
whether a natural enemy, fingers or forceps.
Grasping the tips of the appendages, or simply
touching or tapping any part of the insect, does
not evoke the response. Neither does a purely
visual stimulus, or the odor of a lizard’s cage.
Also, as previously stated, it is a rather uncom-
mon response, being fairly predictable only in
moderately young, mated females, or when the
abdomens of either sex are squeezed so forcibly
that the extrusion may be considered a purely
mechanical effect. It is only with the stronger
stimuli that the “stink clubs” (p. 177) are ex-
serted.

On the other hand, the ventral curving of the
abdomen, which always accompanies extrusion
of the gland, occurs to a greater or lesser extent,
and without extrusion, whenever the insect is
seized as described above. It will be noted that
this abdominal curving-under is in the opposite
direction from the upward tilting of the female
abdomen when the gland is everted during court-
ship (p. 177) . The restricted use of the gland
in defense is true of field-caught as well as
insectary-reared specimens, although it may be
that male responses are even rarer in these
“tame” butterflies. However, insectary females
of appropriate age respond fully to seizure, and
evert the glands even late in life, when no odor
can be detected.

Dr. Eltringham’s (1925, 1926) dissections
and serial sections of the abdominal glands in the
female of this species and subspecies showed
only a single pair of glands, although there were
two pairs in related genera. It seems likely there-

fore that the protective function, if any, of the
strong odors evolved directly out of the sexual
function, and that their chemical composition is
not necessarily different from the substance(s)
obviously used in courtship but not detectable
to the human sense of smell. The actual courting
use of the glands in the female, apparently re-
ported for the first time in the present study,
confirms Eltringham’s surmise concerning such
a function in addition to that of defense.

In comparison with H. besckei in Brazil
(Muller, 1877, 1912), //. erato has this defense
mechanism much less well developed. It may
be that H. erato represents a preliminary stage in
evolution, in which the everting of the glands
when the insect is prevented from escape, and
the concomitant discharge of abundant odorifer-
ous material, is scarcely more than displacement
behavior in which a sexual response is given. In
other species of the group it has probably de-
veloped into a highly evolved defense. Accord-
ing to Muller it was a habitual defense response
with both sexes.

Mr. Ernest C. Crocker of the Flavor Labora-
tory of Arthur D. Little, Inc., in Cambridge,
Mass., has most kindly given his impressions on
samples sent him of the harpes of a 72-hour-old
male and the abdominal scent gland of a 16-
hour-old, mated female. Mr. Crocker and two
assistants examined the material within five to
six days after the specimens were killed and sent
to him in small vials, via airmail. He writes as
follows (personal communication) : “The . . . .
Heliconius erato specimens are all animal-like,
earthy, musty, and dulcy, and yet somewhat
flowery. Their distinctive features follow . . .

“The harpes, 72 hours old, have distinctly
phenyl carbylamine character (see next descrip-
tion) .

“The scent glands 5, 16 hours old, have par-
ticularly sharp, strong phenylcarbylamine (phe-
nyl isocyanide, CeHs— N=C type) odor. This
odor also suggests styrene (phenyl ethylene)
and phenyl proprionaldehyde. Enclosed is some
of that aldehyde and also some brom-styrene. In
our opinion, this insect odor must be due to a
phenyl compound, with no more than 2 or 3 car-
bons on the side chain and possibly an oxygen
or nitrogen.”

The odors of the samples, when compared
with living specimens in Trinidad, did not bear
any obvious similarity to those of the abdominal
glands of either sex; however, within a few days
the odor of the excised glands decidedly re-
sembled the brom-styrene sample.

2. Scent Scales. As in other species of the
genus (Muller, loc. cit.), specialized scent scales
are present in the male only, and only on the
friction surface of the anterior portion of the

182

Zoologica: New York Zoological Society

[40: 16

upper hindwing. In the present species they are
small, dark, and mostly concealed by the larger
uncolored scales of the friction surface. Al-
though they apparently play a definite role in
courtship (p. 179), their odor is only rarely and
questionably detectable in courting males. In
fresh or long-preserved friction surfaces, how-
ever, the odor is moderately strong, and to us
appears decidedly flowerlike. Mr. Crocker ex-
amined samples of this area, from the 72-hour-
old male mentioned above, and noted that the
general character was that of the abdominal
glands, namely “animal-like, earthy, musty and
dulcy, and yet somewhat flowery.” In distinc-
tive features they were “tobacco-like, somewhat
fecal (scatole) and have some finnanhaddick
amine character. No carbylamine-like odor
noted.”

3. Odor of Thorax and Wings. A fragrance,
apparently distinguished here for the first time,
is clearly perceptible in both sexes on the dorsal
side of the thorax and on the upper surfaces of
all the wings, especially basally. To our percep-
tion it is not distinct from the fragrance of the
scent scales proper, except that it is weaker. In
dried specimens or detached wings, even many
weeks after death, there may be, as in the ab-
dominal glands, a resemblance to the odor of
brom-styrene. In living butterflies this fragrance
does not become apparent until about the sec-
ond day in females and the third in males. The
odor does not depend on mating or even court-
ing to be evident in females; therefore at least
in these females it cannot be merely the result
of fanning the products of the scent scales by
courting males.

4. Odor of Young Imagos. A musky odor is
evident in very young imagos; usually it is not
detectable after about the third day. It has not
been definitely localized except that rarely it
has seemed strongest at the tip of the female
abdomen.

To summarize: Three or four distinct odors
are discernible to human beings in H. erato; at
least two of them are probably involved in de-
fense. First, the abdominal glands emit an odor
which, in the intensity detectable to human be-
ings, is at least indirectly a product of the male
harpes; it reaches its maximum development in
mated females. It is a phenyl compound which
changes, for human sense, from a witch-hazel-
like to a phenyl-carbylamine-type odor within
a few days. The odor which almost certainly is
emitted by the extruded gland of unmated fe-
males during courtship is not detectable by the
human sense of smell. Second, as in other mem-
bers of the genus, fragrant scent scales are pres-
ent on the anterior upper portion of the male
hindwing. Third, a fragrance, apparently sim-

ilar both to that of the male scales and, faintly,
to that of the abdominal glands, develops on the
thorax and upper wing surfaces of both sexes,
mated and unmated. Fourth, a musky odor is
apparent only in recently emerged imagos,
especially females, and is possibly strongest at
the distal end of the abdomen.

Any defensive use of the witch-hazel-like odor
of the abdominal glands seems to be quite obvi-
ously a secondary development from an orig-
inally sexual function. A puzzling point is that
the odor reaches maximum strength in oviger-
ous females. Since these are biologically the
most valuable members of the species, it is pos-
sible that their increased odor has indeed some
selective significance. However individuals of
this age do not seem to be more resolutely re-
fused by predators than are others of the species.

It seems that the basis of the aposematic taste
of these butterflies may very likely be a secre-
tion of the thorax, or of the thorax and the wing
bases, which may be also responsible for the
fragrance of these regions in both sexes. Perhaps
it is contained in the weather-proof oily coating
of the scales. The odor probably also has a
sexual, or at least a social significance within
the species.

The function of the male scent scales is prob-
ably sexual only. The musky odor of young
adults may well be of value both socially, sexual-
ly and aposematically.

An interesting related question is that of the
general function of odor as an aposematic signal
in various Lepidoptera. Some authors have re-
marked both on this protection and, simultane-
ously, on the avian beak marks which often
show on the wings of living butterflies, and so
indicate successful past escapes. Assuredly birds
must be the chief natural enemy of adult diurnal
Lepidoptera. Yet these vertebrates, according to
the best current knowledge, cannot smell. The
chemical source of the nauseous taste must
therefore extend to the insects’ wings’ them-
selves, being perhaps contained, as suggested
above, in the oily scale coating. The odor(s) of
the insect, as opposed to the taste, must be
chiefly of use in deterring reptilian and mam-
malian enemies.

It also seems clear, as suggested by Jones
(1930, 1931), to anyone who has observed
sleeping heliconiids in the field, that the protec-
tion afforded by roosting aggregations is that of
reinforced odor rather than of conspicuous
warning color. Roosting heliconiids, of whatever
species, are exceedingly inconspicuous objects
even in daylight.

G. OviPOsmoN. The foodplant of H. erato
hydara in Trinidad is Passiflora tuberosa (Jac-

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

183

quin). Eggs are laid only on the youngest
shoots and tendrils, generally one to a plant.
The vines seldom grow taller than six feet, each
shoot dying back after flowering and the young
shoots tending to come from the ground or near
it. Hence laying females usually fly only a few
inches or several feet above the ground. Most
eggs are laid between noon and 2:30 PM, which
is otherwise a period of relative quiescence for
the species. One egg, however, was laid at 8:45
in the morning and another as late as 3:30 PM.
The first egg normally appears on the eleventh
or twelfth day after emergence, although a sin-
gle individual, mated with a Surinam male 45
minutes after emergence, laid the first egg on the
fourth day; another female commenced laying
on the tenth. In no other females, among more
than 20 recorded layers of known age, was an
egg laid earlier than the eleventh day. Egg-lay-
ing continues at the rate of 0 to 4 eggs, very
rarely 6, daily for about two weeks or slightly
more, until a total of up to about 24 eggs have
been deposited. A typical female laid 18 eggs in
1 5 days, with a final gap of four days between
the seventeenth and eighteenth egg. The higher
numbers, two to four a day, are laid in the first
half of the period.

H. Longevity. Under ideal conditions, males
usually live in the insectaries slightly more than
a month after emergence although rare indi-
viduals live more than two months. Females
habitually live six weeks or more, again if no
untoward circumstances, such as long rainy or
windy spells, occur. The record for any indi-
vidual was 91 days in the imaginal state, attained
by a female reared in captivity. She had been
mated on her third day and laid about 20 eggs
on schedule. Counting the total 26 to 30 days
of the developmental stages, this gives a record
lifespan of about four months.

VI. Experimental Analysis of Social
Behavior

A. Introduction. Experimental work was
undertaken in order to determine the releasing
mechanisms of courtship and other types of
social behavior. In particular it was desired to
discover whether any role was played by color in
these activities, and if so, whether it was as an
innate or acquired response.

The evidence was accumulated in three ways :
first, through concealing the female so that only
her odor could guide the male, the sense of sight
being altogether or partially eliminated. Second,
the color and pattern of both sexes were changed
in various ways, in order to establish the relative
importance, if any, of hue, brightness and mark-

ings. Third, the responses of both sexes were
tested with models.

B. Work with Concealed Females. 1.
Method. Six virgin females, each between 24
and 72 hours after emergence, were placed,
singly, in a box 2X2 inches in size, which gave
them sufficient space in which to move their
wings freely. In the first tests the cardboard cov-
ering top and bottom was punctured with about
a dozen nailholes, through which the female was
quite invisible; in other tests the top was cov-
ered with khaki-colored mosquito netting, which
presumably made the odor of the female more
discernible, but through which she was dimly
visible to the human observer. Each of the but-
terflies, before being placed in the box, had been
tested with one or more young males and found
to be normally attractive— that is, she elicited
Stage I courting behavior.

2. Results. When the box with an enclosed
female was left on a stool in the insectary, in the
path of freely flying males, none ever dipped
toward it, or paid it any other evidence of at-
tention, whether the box was covered with punc-
tured cardboard or netting. Similarly, virgin
females kept in a cage 18 X 18 X 12 inches,
with one of the long sides covered with wire
netting, did not attract the attention of males.
It is concluded that female odor alone is not
sufficient in this species to inaugurate courtship.

C. Work with Painted Butterflies. 1.
Method. A variety of unsuccessful preliminary
attempts was made to change the color of a
living butterfly. Sometimes the color, such as
water color or India ink, would not adhere to
the oily surface of the scales. Removal of the
film with weak acid damaged the wing, as did
bleaches strong enough to fade out the black
pigment. Painted tissue attached to the wing
with an appropriate cement was so heavy that
the butterfiy could not fly. Certain paints, such
as artists’ oils, did not dry fast enough. Finally,
the butterfly was often so badly shocked by
the procedure— apparently because of the nec-
essary handling— that it never recovered normal
activity levels and died more or less promptly.

The following technique, however, was slowly
perfected and can now be recommended (PI.
Ill, Fig. 15). The main precautions to be ob-
served are to work slowly and to touch the
butterflies with the fingers as little as possible.
Properly done, no manual handling whatever is
necessary, after the butterfly has been placed in
the envelope, and hardly any touching with
forceps. Stroking with the paint brush should
be feather-light.

Step. 1. The butterfly is placed in a glassine
envelope, the wings folded over the back, as

184

Zoologica: New York Zoological Society

[40: 16

soon as captured, either in the field or the in-
sectary; handling and warming are avoided.
With specimens taken from the insectary, cap-
ture and painting is done with the least shock at
night, the insect being taken from the roost.
After painting it can frequently be rehung in
approximately the same place.

Step 2. The envelope is placed in a flat wood-
en box about 3X3X1 inches, with a glass bot-
tom. The box is turned upside down so that the
butterfly can be observed through the glass.
Through one side of the box enters the nozzle
of a short tube attached at the other end to a
small fire extinguisher or bicycle pump filled
with carbon dioxide. The stopcock of the
cylinder has previously been turned on and the
gas regulated to a very low rate of flow. The
butterfly in the envelope is left in the box for
three to five minutes, or until a minute or two
after all movement has stopped. If the butterfly
becomes active during painting, it can be further
anaesthetized merely by placing the glass-bot-
tomed box briefly over the insect on the spread-
ing board (see below). Deep anaesthetization
should be avoided, since butterflies so treated do
not always regain their normal vigor, and may
not court.

Step. 3. A piece of oiled paper is laid across
the slot and breadth of the butterfly spreading
board, to support the body in a sling and protect
the wings from the wood. The butterfly is
shaken onto the paper and, with the wings re-
maining folded, fastened into place with two
strips of paper pinned across, but not through
the wings, in such a way that the area of under-
wing to be painted remains uncoverd.

Step 4. With a fine paint brush ether is
brushed lightly several times across the area to
be painted.

Step 5. The underwings of one side are
painted as desired with a waterproof, fast-dry-
ing lacquer. In the present experiments, “Flo-
paque” was used (manufactured by Floquil
Products, Inc., Cobleskill, New York). The
paint should be applied thinly and restroking
avoided.

Step 6. Papers are removed, the butterfly
flipped over, and the underwings of the other
side similarly treated.

Step 7. By careful manipulation of paper
strips, the butterfly’s wings are now opened
and fastened to the board, without the toueh of
fingers or forceps. Extreme care must be used
not to exert pressure on head or body by pinning
their paper covering too tightly or by letting the
wax paper sling under the body become taut.
A strip loosely covering eyes and body, how-
ever, helps keep the insect quiet when it starts to
emerge from the anaesthetic. Brushing with

ether and painting then proceeds as on the
underwings.

Step 8. After painting is completed the set-
ting board is taken to the insectary and the but-
terfly released by unpinning the paper strips,
thus avoiding further handling. When the pro-
cedure is properly carried out, the insect is
usually capable of flying at once to a resting
place. When the painting has been done at
night, the butterfly can often be simply rehung
with forceps on the roost. In any case, the but-
terfly usually remains quiescent for some hours,
although females give normal courting re-
sponses sooner. Recovery should be altogether
complete by the next day. When possible, but-
terflies should not be painted sooner than 48
hours out of the chrysalid, since the wings of
freshly emerged butterflies are very easily dam-
aged and the insect is more subject to after-
effects of the treatment. However, for special
procedures employed to ensure a butterfly’s
being unconditioned to the color “red,” see p.
187.

If the use of other types of paints is desirable,
such as fluorescent paints, or water colors with
particular spectral characteristics, they may
sometimes be successfully applied as follows:
the coloring matter is painted on a sheet of lens
tissue and allowed to dry, in several coats if
necessary, so that the tissue becomes opaque.
The paper is then cut in pieces of the desired size
and shape. The butterfly is anaesthetized and
placed on the spreading board as for painting.
The painted bits of tissue are fastened in place
with rubber cement diluted with xylene. The
method is only rarely satisfactory, since the
extra weight often prevents the butterfly from
flying well. No successful attempts have been
made to remove the insects’ scales before at-
taching the paper, and so reducing the weight
of paint needed to efface the natural color; the
wing tissue is always too much weakened either
by the use of chemicals or by stripping with
Scotch tape.

Table 4 gives a summary of the various color
and pattern changes effected; Table 1 (p. 172)
shows a spectrophotometric analysis of the
paints used.

No positive responses are included where there
was danger of the female’s showing effect of
summation. The experiments were conducted
both with individuals which had emerged and
been kept in isolation, out of sight of their own
species, and not allowed any sight of the color
red, and those which had not been so isolated
(see p. 187).

2. Results. In brief the following principles
may be stated, concerning the social responses to
painted butterflies.

1955]

Crane: Imaginal Behavior of Heliconiiis erato hydara

185

a. Courtship. These responses are given in
summary in Table 4. Positive responses were
counted for the sex involved if the female, when
approached or actively courted by a painted
male, elevated the abdomen, extruded the yellow
organs, apposed the forewings and vibrated the
hindwings. If only the abdomen was elevated
and the yellow organs extruded, a half-response
was counted (the hindwings were never flut-
tered without the abdominal response). The
very rudimentary response of merely raising the
abdomen slightly, without extruding the yellow
glands, was not counted, since this is some-
times done when the insect is resting alone, or
feeding.

Positive responses were counted for males to
painted females if he approached the female
and made repeated, continuous courting dashes
at her, in at least the typical Stage I rear posi-
tion (p. 177).

It will be seen from Table 5 that all changes
of color and pattern attempted resulted in posi-
tive responses in at least one sex. What cannot
be fairly indicated in the table, because of the
difficulties of the procedure and the many vari-
ables of weather, physiological state of the but-
terflies and so on, is the relative popularity of
the various colors. This can be recorded only
in general terms, resulting from the author’s
prolonged observation of painted butterflies over
a three-season period. It was borne out in every
respect in the amount of general social chasing
of the variously painted individuals by other
members of the group after the active breeding
age was passed (p. 180).

Briefly, the farther the altered color of the
forewing pattern is distant from the normal
scarlet in the spectrum, the less notice is taken
of the butterfly, either by the opposite sex or as a
subject for general social chases. That is, butter-
flies with the forewing band painted orange or
yellow were almost or quite as successful social-
ly as butterflies repainted in natural colors. In
most trials the orange was more successful than
yellow, that is, more promptly responded to by
members of the opposite sex, and slightly more
subject to social chasing than yellow. Yellow-
green bands were less successful, while greens,
blues, violet, positively ultraviolet white and all
black were least so. Negatively ultraviolet white
was, again, moderately successful, but since for
this color Chinese white (zinc oxide) water
color on bits of tissue had to be used, the but-
terflies were both overburdened and scarcely
weather-proof; hence their general social life
when fitted with this color was not subject to
valid comparisons.

All-black butterflies and those with incon-
spicuous short-wave markings, such as all black

with dull green radiations on the hindwings,
were notably unsuccessful. In fact no complete-
ly black female, with even the pinkish under-
hand of the forewing eliminated, ever drew a
positive response from a male; the one case of
mating given in the table was a young female
which had been given only a thin wash of black,
permitting a pinkish tinge to show both above
and below. Out of eight individuals painted
more or less black, this was the only positive re-
sponse. Black females were repeatedly ignored,
even though, at the close passing of males, all the
usual Stage I responses were given: apposed
forewings, gland extruded from the elevated
abdomen and vibrated hindwings. Occasionally
these unnoticed females even chased after such
passing males, yet never attracted more than
the briefest dips in their direction.

b. Social Chasing. Behavior toward males and
non-receptive females which had been painted,
paralleled that of courting individuals. The far-
ther away the spectral reflectance of the paint
used from the orange-red region, the less was
the butterfly chased. This was regardless of
brightness: pale blues, greens or green yellows
were disregarded in comparison with deep or-
anges and reds. Grays were treated like black.
Often an ignored painted female would chase
other butterflies, which even then would only
rarely take any apparent notice of it. There was
no reciprocal circling. Blue-marked butterflies,
which often lived weeks after painting, were
ignored almost or quite as completely as all-
black individuals.

c. Roosting. In this activity the painting of
individuals made no apparent difference. All
were accepted on the roost, and made and held
for themselves positions in close juxtaposition
both to normal butterflies and to those otherwise
colored.

D. Work with Butterfly Models. 1.
Method. The most successful models were made
of heavy black or colored cotton felt or canvas
cut roughly to the shape and size of a Heliconius
erato. To these were fasted bits of colored paper
in various sizes and patterns. Along the longi-
tudinal axis of the “body” of each model was
threaded a black insect pin. The other essential
part of the apparatus was a flexible 2Vi-foot
wand of split bamboo. To the small end of the
wand was attached a small magnet, such as is
used for fastening notes to bulletin boards. A
petroleum-base adhesive attached the magnet to
the bamboo, both magnet and wand being
painted dark green and black. By means of the
magnet and the insect pin through the body of
each model, a series of different models could be
presented to a butterfly in as rapid succession as
desired.

186

Zoologica: New York Zoological Society

[40: 16

Table 5. Alterations Effected in Color of Living Heliconius erato hydara.

Each color change indicated by “X” was effected in at least one butterfly of each sex. The responding
individual, painted or not, had not been allowed previously to see the color red, except that of its own
wings. Positive courting responses, sometimes including copulation, were obtained at least once for each
color change effected in each sex. However, responses were exceedingly rare, and almost always weak to
the blues, greens and blacks, while oranges and yellows were treated almost like normally colored
butterflies. Change in pattern was of relatively little importance. Females were in general less influenced
by color change than were males. Numbers in parentheses after color names refer to paints used (For
spectrophotometric analyses, see Table 1, p. 172).

No Color Added

Red (31)

Orange (25)

Yellow (20)

Green (12)

Blue (5)

Blue-green (10)

White (UV Positive)

White (Zinc Oxide)

Pink (36)

Red forewing band
unaltered

Hindwing
with spot

X

X

Hindwind
with radiations
Under fore- and/or
hindwing with spots,
crosses, dots,
in many combinations
(for identification)

X

X

X

X

X

X

X

X

X

Forewing band
altered in color
(upper & underwings)

Hindwing unaltered

Hindwing with spot
of same color

X

X

X

X

X

X

X

X

X

X

X

Hindwing with radiations
added

X

Upper forewing band
altered in color,
blackened on underwing

Hindwing unaltered

X

X

X

Upper forewing band
blackened

Underforewing band
painted

X

Underforewing band
unaltered (pale pink,
pos. UV)

X

Forewing band
blackened on
upper & underwings

Hindwing unaltered

Forewing with colored
spot, hindwind unaltered

Hindwing with spot

Hindwing with radiations

X

X

X

X

X

X

X

Fore & hindwing with
single spots

X

Upper forewing band
altered in color,
underforewing band
painted white (pos. UV)

X

X

X

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

187

The most successful, and in general the only
successful technique, consisted in jiggling the
wand gently up and down near the butterfly to
be tested by holding the wand in the left hand
and tapping it rapidly near the base with the
right forefinger. This caused the model at the
tip to flutter up and down in a fairly realistic
way. When models made entirely of paper had
to be used (because of desired spectrophoto-
metric characters), an underlayer of flapping
black felt beneath the stiff paper gave the air
current which is an important factor in the
species’ courtship pattern.

As in the work with paints, spectrophoto-
metric reflectance curves were secured for the
colored textiles and papers used (Table 1, p.
172).

The experiments were conducted in three
groups. First, with individuals which had not
been kept isolated, after their emergence, from
normally colored members of their species, and
which hence might have become conditioned to
the color “red.” Second, with individuals which
had been kept in isolation from the time of their
emergence until the hour of the experiments.
These butterflies were not fed at all, or allowed
to see any red object, even being isolated from
other chrysalids, lest the red wing band which
shows through the chrysalid on the last day be-
fore emergence should be seen by and possibly
imprinted on an emerging neighbor. The iso-
lated individual, after 24 to 48 hours, was then
painted or not, depending on the needs of the
subsequent experiment.

The isolation of unpainted examples was
maintained as follows: The cages were kept at
a distance from one another and from the in-
sectaries, and the sides facing occupied cages
were masked with branches. Wild erato almost
never enter the garden where the small cages
are kept, so there was very slight chance of con-
ditioning by the brief sight of a passing member
of the species.

The third group of test butterflies was even
more rigidly guarded against possible condition-
ing to red. In this group the possibility was
eliminated that a just-emerged butterfly might
be conditioned or imprinted with the red color
of its own wing markings, whether the scarlet
forewing bands or even the tiny red dots and
bar near the base of the underwings. It proved
difficult to control this factor, but the following
technique finally resulted in active, uninjured
butterflies, fully protected against a previous
sight of red, in about seventy-five per cent of
the individuals treated.

Caterpillars of the imagos to be tested under
these conditions were reared in complete isola-
tion, as in the second group described above.

and allowed to pupate hanging from a four-by-
four inch piece of glass. On the night before
emergence, the glass with its dangling chrysalid
was placed over a hole in the lid of a light-tight
shoebox. On top of the glass was placed a six-
by-six-inch Corning Glass filter (Col. spec. 5-60,
“H.R. Lantern Blue”); while transmitting the
near-ultraviolet, violet and blue freely, it has a
sharp cut-off in the green at 520 m^u. The filter
was tightly sealed into place with black photo
tape. Under this blue filter the test butterflies,
in turn, emerged on schedule and were found on
release to be less battered than those which
emerged in complete darkness. Each individual,
when examined the same evening under an
identical blue filter fastened over an electric
desk lamp, was found to be fully expanded with
the wings relatively firm. The room was kept in
complete darkness except for the blue light, so
that the butterfly at no point had an opportunity
to catch sight of its own scarlet markings. The
butterfly was then removed from the box, anaes-
thetized with carbon dioxide and fastened on the
spreading board, as in the usual painting tech-
nique (p. 183). The only difference was that
after each wing surface had been arranged in
position for painting, the eyes were covered
closely with criss-crossed strips of black felt.
This enabled actual painting to proceed under
ordinary electric light, without the use of the
blue filter. A blue paint, with relatively low re-
flectance in the long wave regions, was used
(Sample 5, Table 1). After completion of the
painting, the butterfly was hung in the insectary,
from which all red or orange flowers had been
removed, as well as all other butterflies. The
young butterfly was allowed to feed well the
following morning. Usually these butterflies
which had emerged in a small box, been kept
there all day and painted less than twelve hours
after emergence were not in condition for test-
ing until the day following feeding, that is, 48
hours after emergence. One did not become
fully active until his seventh day, when he was
successfully tested. The mortality is in any case
high from this rough treatment, but the results,
from the seven specimens which survived it,
seem conclusively to be based on innate, and not
learned or imprinted, behavior. These results
(Table 6) agree extremely well with those ob-
tained for the more numerous individuals tested
in the other two groups, in which previous ex-
posure to red was either partially or not at all
controlled.

Throughout the experiments in all three
groups, no positive responses were counted
where there was danger of the female’s showing
summation, or heterogeneous summation. A
model, normally unsuccessful, such as plain

188 Zoologica: New York Zoological Society [40: 16

Table 6. Responses of Blue-painted H. erato hydara, Reared in Isolation, to Model Butterflies

Part a: Field Data

Explanation: Each of the 7 specimens was maintained in isolation, without exposure to orange-red
or red, until the time of the test. See text, p. 187 tf. All tests were conducted in calm, sunny weather in
the large insectary. Each model (p. 185, Text-fig. 2, PI. Ill, Fig. 17) was presented for a maximum of
30 seconds. A response, if any, usually occurred within five seconds. In males, a single slight dip
toward the model was counted as a “minimum” response, several dips or short chases as “good,” and
definite Stage 1 courtship, with persistent pursuit, as “strong.” The single female’s responses were similarly
gauged from the strength of her courting behavior, shown by the degree of abdomen elevation, extru-
sion of yellow gland, apposition of forewings and fluttering of hindwings. Color sample numbers refer
to spectrophotometric analyses in Table 1, p. 172.

Individual

Presen-

Color

No.

tation

Order

Sample

No.

Model Type

Model Hue

Response

KS)

1

4

2 colored bands on black

Blue

None

2

8

(1

Blue-green

None

3

15

4i

Yellow-green

None

4

28

44

Orange-red

Good

5

17

M

Yellow

None

6

22

44

Yellow-orange

Minimum

7

9

44

Blue-green

None

8

33

Solid color

Red

Good

9

—

“

Black

None

10

23

2 colored bands on black

Orange

Good

2($)

1

4

2 colored bands on black

Blue

Minimum

2

15

44

Yellow-green

None

3

17

44

Yellow

Minimum

4

8

44

Blue

Minimum

5

28

44

Orange-red

Strong

6

22

44

Yellow

Minimum

7

33

Solid color

Red

Strong

8

4

2 colored bands on black

Blue

Minimum

9

23

“

Orange

Good

10

8

Blue

None

11

29

Red

Strong

3(3)

1

7

Solid color

Blue

None

2

14

«

Green

None

3

18

“

Yellow

None

4

27

“

Orange

Good

5

7

Blue

None

6

33

“

Red

Good

7

36

“

Pink

Minimum

8

22

“

Yellow

None

9

33

“

Red

Strong

10

4

2 colored bands on black

Blue

Minimum

11

16

“

Green-yellow

None

12

17

“ Yellow

None

13

22

“ Yellow

Minimum

14

8

“ Blue

Minimum

15

15

“ Yellow-green

None

16

23

“ Orange

Good

17

16

“ +UV White

Minimum

18

4+28

2 colored bands on color

Orange-red on blue

None

19 28+22

Colored stripes on black
(Text-fig. 2c)

Red; Yellow

Minimum

20

—

2 colored bands on black Zinc oxide white

None

21

28

“ Orange-red

Strong

22

16

“ Green-yellow

None

23

34

“ Violet-red

Strong

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

189

Table 6. (Continued)

4(3)

1

7

Solid color

Blue

None

2

11

Blue-green

Minimum

3

18

Yellow

Minimum

4

27

Orange

Strong

5

7

Blue

None

6

33

4(

Red

Strong

7

4

Blue

None

8

17

Yellow

None

9

22

“

Yellow

None

10

28

Orange-red

Strong

11

9

44

Blue-green

None

12

36

44

Pink

Good

13

17

Black stripes on color

Yellow

None

(Text-fig. 2d)

14

1

Solid color

Violet

None

15

16

2 colored bands on black

Green-yellow

None

16

34

“

Violet-red

Good

17

_

“

Zinc oxide white

Minimum

18

28

Orange-red

Good

5(3)

1

7

Solid color

Blue

None

2

18

“

Yellow

None

3

27

“

Orange

Strong

4

14

“

Green

None

5

33

Red

Strong

6

18

Yellow

None

7

28

2 colored bands on black

Orange-red

Strong

8

22

“

Yellow

None

9

33

Solid color

Red

Strong

10

23

2 colored bands on black

Orange

Strong

11

22

Yellow

Good

12

16

Green-yellow

Minimum

13

8

Blue

None

14

33

Red

Good

6(3)

4

1

7

Solid color

Blue

Minimum

O

14

“

Green

None

3

18

Yellow

None

4

27

Orange

Good

5

7

Blue

None

6

36

“

Pink

Minimum

7

33

Red

Strong

8

4

2 colored bands on black

Blue

None

9

—

4-UV White

None

10

22

Yellow

Good

11

17

Yellow

None

12

29

Red

Good

13

74-28

2 colored bands on color

Orange-red on blue

None

14

28

2 colored bands on black

Orange-red

Strong

15

34

Violet-red

Good

16

16

“

Green-yellow

None

17

23

Orange

Strong

18

33

Solid color

Red

Strong

7(5)

1

33

Solid color

Red

Strong

2

27

Orange

Strong

3

7

Blue

None

4

14

44

Green

None

5

18

44

Yellow

None

6

36

et

Pink

Minimum

7

7

44

Blue

None

8

27

44

Orange

Good

9

33

Red

Strong

190 Zoologica: New York Zoological Society [40: 16

Table 6. (Continued)

10

22

2 colored bands on black

Yellow

Minimum

11

4

Blue

None

12

23

Orange

Minimum

13

17

<6

Yellow

Minimum

14

—

((

Zinc Oxide White

Minimum

15

29

((

Red

Minimum

16

16

((

Yellow-green

Minimum

17

28

((

Orange-red

Good

18

33

Solid color

Red

Good

19

28+4

2 colored bands on color

Orange-red on blue

None

20

34

2 colored bands on black

Violet-red

Good

Part b: Summary

Color

Sample

No.

Hue

Responses*

None

] Minimum |

Good

Strong

—

+UV White

1

1

—

Zinc oxide White

1

2

—

1

Violet

1

__

_

4

Blue

4

3

—

_

7

Blue

8

1

_

_

8

Blue

3

2

—

9

Blue-green

2

—

—

—

11

Blue-green

-

1

—

—

14

Green

4

—

—

15

Yellow-green

3

—

—

—

16

Green-yellow

4

2

—

—

17

Yellow

4

2

_

—

18

Yellow

5

1

—

_

22

Yellow-orange

3

4

2

—

23

Orange

—

1

3

3

27

Orange

-

—

3

3

28

Orange-red**

—

—

3

5

29

Red

—

1

1

1

33

Red**

4

9

34

Violet-red

_

__

2

1

36

Violet-pink

-

3

2

-

*Black models with colored bands and solid colored models only.
**Used frequently, to check condition of specimen.

black felt or a stick, was used as a test in sus-
pected cases. Contrariwise, a butterfly’s thresh-
old was gauged periodically during an experi-
mental session by presenting a highly successful
model. If a positive response was given to this,
the insect was considered still to be in a mood
adequate to continue the session.

In the entire series of tests, 53 young butter-
flies were used, including 30 males and 23 fe-
males. Of the total, 25 had not been isolated
from their owu species, 21 had been isolated

from red except for the sight of their own scarlet
wing band, and 7 were both isolated and painted
blue.

2. Results. Since the factor of odor was re-
moved, and movement controlled, the results
were more clean-cut than in the case of painted
specimens.

It was found through the use of models that
both motion and hue are of importance in court-
ship. In order to elicit the hindwing flutter of
females, the model had to be waggled up and

1955]

Crane: Iniuginal Behavior of Heliconhis erato hyduru

191

down immediately behind, but not touching, her
hindwings; the determining factor in this stage
was the air current thus formed. A model, other-
wise highly successful, which was vibrated to one
side or in front of her, as in later stages of court-
ship, did not cause her either to raise the ab-
domen, extrude the yellow gland, appose the
forewings or flutter the hindwings.

The response of males to models was partly
to the fluttering movement of the model. Mo-
tionless models, tied by black threads to twigs or
leaves, or left on the ground, only very rarely
elicited a dip from a passing male. Models flut-
tered on the wand, however, were often chased
and more rarely drew the first stage of courtship,
that is, repeated short dashes with the friction
surfaces not exposed.

The various models themselves were success-
ful in direct proportion to their similarity to the
size and color of living butterflies, with one ex-
ception; the most successful model of all was
made entirely of negatively ultraviolet red felt
without any pattern whatsoever; in fact it acted
as a supernormal stimulus, eliciting a response
from either sex when not only all other models
failed, but often when the threshold was so high
that a living member of the opposite sex in an
appropriate physiological state did not win a
response. It was accordingly used to test the
continued responsiveness of an individual from
time to time during a series of tests, or to deter-
mine its threshold before an experimental
session.

Models with forewing bands of naturalistic
size and position but varying only in color were
successful in about the same sequence as were
butterflies with the bands painted: the most suc-

cessful were the naturalistic orange-red. How-
ever, the red, orange and yellow-orange were
almost equally successful, lemon yellow and
greenish yellow notably less so, while greens,
blues, violets and positively ultraviolet white
least so. In general, the colors reflecting ultra-
violet the most strongly were less popular in
comparison with purer colors of the same gen-
eral spectral region. Negatively ultraviolet white
drew responses rather unpredictably; on the
whole they were moderately successful with in-
dividuals of low threshold. No response was ob-
tained to an all-black model except weakly, by
several females of very low threshold which
were already highly stimulated by preceding
models or by actual courting. These individuals
responded equally well at that stage to prac-
tically any small object moved close behind
them, including dead twigs and the observer’s
finger. Pattern was of little importance, so long
as the background was black. Text-fig. 2 shows
the type of variations used. However— except
for the solid color red models, approaching the
natural color of the butterfly’s band— solid col-
ors, with or without contrasting or gray bands
or other markings, were less successful than
their counterparts in which the model was black
with a band of unnatural color. For example, an
all-yellow model was less successful than a black
model with a yellow band. An all-orange model
however usually elicited a strong positive
response.

Models with a wingspread of from about two-
thirds to one and a half times normal were of
acceptable size; models notably smaller or larger
than these were unsuccessful, regardless of color
or the individual’s threshold.

Text-fig. 2. Examples of models used in butterfly courtship e.xperlments. A, normal type: black with
colored bands; B, solid color; C, colored stripes on black; D, black stripes on color. Natural size.

192

Zoologica: New York Zoological Society

[40: 16

Table 6 gives the series of responses by the
seven individuals in which all chance of condi-
tioning to red was avoided. Because of the im-
portance of the order of model presentation the
full field data are given in Part a. In the sum-
mary of the table, Part b, it will be seen that the
results follow closely those summarized in the
text above for all the 53 young butterflies tested.

Models elicited a minimum amount of atten-
tion from old males and old females, and were
disregarded by all when manipulated before the
roost.

VU. Conclusions Concerning Socul
Behavior

The principal conclusions to be drawn from
the preceding observations and experiments on
the social behavior of H. erato hydara are as
follows:

1. Motion, odor and hue are the most im-
portant factors in courtship, and play roles both
in the males and females.

2. Minor elements are size, pattern and shape.

3. Motion is the single most nearly essential
element; without it neither odor nor hue can
elicit courtship, and even the combination is of
no value except in cases of low-threshold males
mating with nearly motionless, freshly emerged
females.

4. Odors function importantly as releasers in
both sexes, and are apparently essential for car-
rying courtship to completion: no male ever
tried to copulate with a model and the hindwing-
flutter-elevated-abdomen response of females to
models was always of short duration. Part of
this, however, may have been due to the fact
that the motions of the models were necessarily
only rough approximations of the flutterings of
live butterflies. On the other hand, motion and
odor without color badges practically never
elicit courtship in the male and odor alone never
does. Females, however, quite readily accept
courting males lacking in or different in color,
providing the other elements in the pattern are
represented and the female threshold is very
low. Odor is certainly the primary stimulus in
roosting, as it is in the non-social activity of
oviposition.

5. Color badges act rather as a directive stimu-
lus, or as a preliminary releaser at most; how-
ever in this role they are important to both
sexes. Although under low threshold conditions
either sex will respond with the first stage of
courtship to models or individuals marked with
practically any color, the most successful are
those most nearly approaching the natural black
marked with orange-red. The striking exception
is the high sucess of all-red models, which, com-
bined with the unimportance of pattern, indi-

cates that the hue “red” itself is a decided re-
leaser. (cf. the discovery of Tinbergen et al.
( 1941 ) concerning the supernormal black model
in the grayling butterfly). The least successful
of both models and painted individuals are those
marked or solidly colored with hues reflecting
mostly in the short-wave end of the spectrum,
or colored very dark or black. Negatively ultra-
violet white is rather unpredictable in eliciting
responses. As might be expected, the toleration
for unnaturally painted but living butterflies is
wider than for odorless, clumsily moved models.

The conditions of the experimental work
show that this preference for orange-red and
near-by similar colors is distinctly innate, and
that it is not a question of mere visibility: H.
erato will come to paper flowers of any color
except green and probe accurately. Living blue
flowers of certain species, such as vervain, are
visited freely, especially in the absence of more
favored species. The color preference in feeding
tests is distinctly for yellow flrst and orange
second, in contrast to the species attraction for
orange-red flrst, then orange and red, then yel-
low.

There was no apparent difference in social
responses of butterflies conditioned to red by the
sight of flowers, other individuals, or simply to
their own wing colors, and those which had had
no previous experience of red.

The red was of similar value in “social chas-
ing” as in courtship, but of none in roosting.

It must be kept in mind that the “red” of
erato is a nearly pure orange-red: its spectrum
shows exceedingly low to negative reflectance
in the ultraviolet, violet, blue, blue-green and
green, very little in the yellow, and relatively
strong in the orange and red, from 600 m/z. up
(Crane, 1954, p. 97, text-fig. 9a). Therefore,
even if these butterflies prove to have, like bees
and other insects, very weak visual sensitivity
above 650 m/x, their perception of the orange
region would still be adequate to make the
orange components of their bands easily visible
to them. The low reflectance of these bands in
the yellow presumably would differentiate them
adequately to the butterfly from the color ap-
pearing “orange” to human eyes; this orange,
in the test material used, included moderately
high reflectance in the yellow and sometimes in
the yellow-green, as well as in both orange and
red. It will be remembered that the color orange
is not quite as attractive in erato courtship as a
natural erato orange-red. However, until electro-
retinograms can be made of the butterflies’ eyes,
there seems to be no satisfactory method of
determining the limits of their spectral sensi-
tivity.

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

193

Perhaps the most interesting aspects of the
results obtained are the following;

First, the selection of “red” as a releaser in
courtship and social chasing indicates a sec-
ondarily evolved use of aposematic coloring. It
is assumed here that aposematism in this and
other red-and-black heliconiids, and their cor-
related development of Mullerian mimicry, oc-
curred in evolution before the development of
red as a social releaser. This is considered a
likely assumption because of the widespread de-
velopment of aposematism in the family. The
members of the family show a variety of strik-
ing colors and patterns which very frequently
do not include red, but rather browns and yel-
lows all characterized by the inclusion of yellow
and green, in addition to orange and red, in
their spectral reflectances.

Second, both color pattern and secondary
colors in addition to red, lack importance in
social behavior, although black as a background
has value. This is to be expected in a species
which in other parts of its range, e.g. Surinam,
is subject to conspicuous variations in pattern
(Beebe, 1955).

The vitally interesting questions concerning
the functions and origins of the various behavior
elements, and of the relation of courtship and
social behavior to phylogeny in butterflies, are
left for a subsequent paper. It concerns com-
parative behavior in a number of species of
heliconiids and is now in preparation.

VIII. Summary

1. Methods of rearing and maintaining broods
of the butterfly, Heliconius erato hydara, are
described which yield healthy adults suitable for
observation and experiment. The caterpillars are
reared singly and fed on Passiflora tuberosa. The
adults are maintained in large, open-air insec-
taries.

2. It was established by behavior responses to
filtered light that H. erato and a number of other
butterfly genera are visually sensitive to light
from at least 366 m/x to at least 600 m^x.

3. Experiments involving the use of colored
paper flowers of known spectral reflectances
among a full series of gray models established
the existence of color discrimination in Heli-
conius erato. Models representing all spectral
regions as well as negatively ultraviolet white
(zinc oxide) were distinguished from all grays.
Yellow is the preferred color in feeding re-
sponses, and the preference appears to be innate.

4. The general habits of flight, feeding and
roosting of the butterfly are described. Daily
activity is governed by temperature in the early
morning and by the reduction of daylight in the
afternoon. As usual in butterflies, the highest

activity occurs on sunny days. Feeding is most
prevalent during the morning, courting in the
late morning and after 2:30 P.M. Egg-laying
usually occurs around noon. These Heliconius
are entirely flower-feeders, preferring yellow
and negatively ultraviolet white blossoms; a few
blues are visited; reds very rarely.

5. Aposematism and defense through unpleas-
ant odor and taste are briefly discussed.

6. Social behavior is of three types— courting,
social chasing and roosting.

a. Courting depends in both sexes on both
visual and scent cues. The most important visual
stimuli are motion and color. Motion releasers
include, in the female, air currents made by the
fanning wings of the male. Color approaching
in hue the orange-red of the butterfly’s forewing
band is an important releaser in both sexes.
Minor visual releasers are form, size and pat-
tern. At least several odors are involved, eman-
ating at least from a special yellow gland in the
tip of the female abdomen and from scent scales
on the anterior margins of the male hind wings.
A more general body odor is also important.
Although all the releasers are mutually depend-
ent components of the courtship pattern, motion
is the one most nearly indispensable. Visual cues
are more important in the early stages, odor in
the later ones. Courtship cannot be initiated by
odor stimuli alone.

b. Experiments with painted living butterflies
and with artificial models all underline the fact
that the purer orange-reds are important re-
leasers in both sexes. A solid-color orange-red
model of normal size acts as a supernormal
stimulus. The farther a color lies from this re-
gion in the spectrum, the less strong is usually
the response. Greens, however, are even less
popular than blues. This responsiveness to
orange-reds is unquestionably innate.

c. Ultraviolet appears to be of no importance
in the life of this butterfly, except to the extent
that its absence (e.g. in negatively ultraviolet
white) affects the perception of a color for the
butterfly.

d. Social chasing is common, especially among
aged individuals of both sexes. Color preferences
are equivalent to those characteristic of court-
ing. No evidence of inter male threat display or
of territoriality was found. Territoriality may,
however, prove to occur in wild populations.

e. Odor is of great importance, color appar-
ently of none, in roosting behavior.

f. Evidence of displacement behavior is found
in atypical courting, where the proboscis is
extruded.

g. It is also suggested that H. erato, in which
the apparently aposematic scent of females de-

194

Zoologica: New York Zoological Society

[40: 16

velops only after mating, is evidence that the
protective function of this odor developed out
of a sexual function. The role of red as a sexual
releaser, however, is held to be a secondary de-
velopment from its original role in warning
coloration.

IX. References

Barth, R.

1953. Consideragoes gerais sobre os organs
odoriferos sexuais dos machos dos Lepi-
dopteros. Mem. Inst. Oswaldo Cruz, 51:
187-202.

Beebe, W.

1918. Jungle Peace. New York: Henry Holt &
Co., 297 pp.

1950. Migration of Danaidae, Acraeidae and
Heliconiidae (Butterflies) at Rancho
Grande, North-central Venezuela. Zoo-
logica, 35: 57-68.

1952. Introduction to the ecology of the Arima

Valley, Trinidad, B.W.I. Zoologica, 37:
158-184. .

1955. Polymorphism in reared broods of Heli-
conius butterflies from Surinam and Trini-
dad. Zoologica, 40: 139-143.

Carpenter, G. D. H.

1932. Acraeine butterflies congregating in a
small area for the night’s rest, observed
by Dr. Hale Carpenter in Portuguese East
Africa. Proc. Ent. Soc. London, 6: 71.

1935. Courtship and allied problems in insects.
Trans. Soc. British Ent., 2 (2): 115-135.

Collenette, C. L.

1929. 3. On the odour of two species of Heli-
conius. (In “Observations on the biono-
mics of the Lepidoptera of Matto Grosso,
Brazil.” by C. L. Collenette & G. Talbot).
Trans. Ent. Soc. London. 76 (2) ; 409-410.

COTT, H. B.

1940. Adaptive coloration in animals. New
York. Oxford University Press, ix-f508pp.

Crane, J.

1954. Spectral reflectance characteristics of but-
terflies (Lepidoptera) from Trinidad,
B.W.I. Zoologica, 39: 85-115.

Crane, J., & H. Fleming

1953. Construction and operation of butterfly
insectaries in the tropics. Zoologica, 38:
161-172.

Edwards, W. H.

1875. The butterflies of North America, 2.
Eggers, F.

1938.1 Zur Frage des biologischen sinnes der
fliigelfarbung tagfliegender Lepidopteren.
Zool. Jahrb. Abt. Syst., 71,: 277-290.

1938.2 Zur biologischen bedeutung der fliigel-
f arbung tagfliegender Lepidopteren. (Film-
vorfiihrung.). Verb. VII Intern. Kongr.
Entom. Berlin II, 688-691.

Eltringham, H.

1919. Butterfly vision. Trans. Ent. Soc. London,
1-49.

1925. On the abdominal glands in Heliconius
(Lepidoptera). Trans. Ent. Soc. London,
1925: 269-275.

1926. On the abdominal glands in Colaenis,
Dione, and Eueides (Lepidoptera). Trans.
Ent. Soc. London, 74: 263-266.

1933. The senses of insects. Methuen, London.

126 pp.

Ford, E. B.

1945. Butterflies. Collins, London, xiv -f 368 pp.
Frisch, K. von

1948. Aus dem leben der bienen, 4th edition.
Wien: Springer-Verlag. 196 pp.

1950. Bees: their vision, chemical senses, and
language. Cornell Univ. Press, Ithaca,
N. Y. xiii + 119 pp.

Guppy, P. L.

1932. The gregarious sleeping habits of a heli-
conian and an ithomiine butterfly in Trini-
dad. Proc. Ent. Soc. London, 6: 68-69.
(Quoted in Poulton, 1931.2).

ILSE, D.

1928. Uber den farbensinn der tagfalter.
Zeitschr. Vergl. Physiol., 8: 658-692.

1932.1 Fine neue methode zur bestimmung der
subjektiven helligkeitswerte von pig-
menten. Biologisches Zentralblatt, 52:
660-667.

1932.2 Zur “formwahrnehmung” der tagfalter.
1. Spontane bevorzugung von formmerk-
malen durch vanessen. Zeitschr. Vergl.
Physiol., 17: 537-556.

1937. New observations on responses to colors
in egg-laying butterflies. Nature. 140:
544.

Jones, F. M.

1930. The sleeping heliconias of Florida. Natu-
ral History, New York, 30: 635-644.

1931. The gregarious sleeping habits of Heli-
conius charithonia L. Proc. Ent. Soc.
London, 6: 4-10.

Klots, a.

1951. A field guide to the butterflies. Houghton
Mifflin Co., Boston, xvi -f 349 pp.

LoNGI lELD, C.

1926. The witch-hazel (Hamamelis)-y\^t smell
of Heliconius h. hydarus Hew., and H. h.
columbinus Stand. Proc. Ent. Soc. Lon-
don, 1:20.

1955]

Crane: Imaginal Behavior of Heliconius erato hydara

195

Longstaff, G. B.

1912. Butterfly hunting in many lands. New
York: Longmans, Green & Co., xx + 729

pp.

Muller, F.

1877.1 Die Maracujafalter. Stettin Ent. Zeit. 38:
492-496. (Transl. in Longstaff, 1912, pp.
651-654, “The ‘Maracuja [or Passion
Flower] Butterflies’”).

1877.2 Beobachtungen an Brachlianilchen Sch-

metterlingen. Kosmos, 1: 391-395.

(Transl. in Longstaff, 1912, pp. 655-659,
“The Scent-scales of the Male ‘Maracuja
Butterflies’ ”).

1878. Die Stinkolbchen der Weiblichen Mara-
cujafalter. Zeitschr. Wiss. Zool., 30: 167-
170. (Transl. in Longstaff, 1912, pp. 664-
666, “The Stink-clubs of the Female ‘Ma-
racuji Butterflies’ ”).

Myers, J. G.

1930. Epigamic behavior of a male butterfly
(Terias) and the gregarious habit during
rest of a heliconine butterfly, in Cuba.
Proc. Ent. Soc. London, 5: 46-48.

Petersen, B., O. Tornblom, & N. O. Bodin

1952. Verhaltensstudien am Rapsweissling und
Bergweissling. {Pieris napi L. und Pieris
bryoniae Ochs). Behaviour 4 (2): 67-84.

POULTON, E. B.

1925. The scents emitted by the neotropical but-
terfly Heliconius erato hydarus Hew.

Trans. Ent. Soc. London, 1925: xxxvii-
xlii.

1931.1 The gregarious sleeping habits of Heli-
conius charithonius L. Proc. Ent. Soc.
London, 6: 4-10, also 68 and 71.

193 1.2 The gregarious sleeping habits of a heli-
conian and ithomiine butterfly in Trinidad,
observed by P. Lechmere Guppy. Proc.
Ent. Soc. London, 6: 68-69.

PouLTON, E. B. and others

1933. The gregarious resting habits of danaine
butterflies in Australia; also of heliconine
and ithomiine butterflies in tropical Amer-
ica. Proc. Ent. Soc. London, 7: 64-67.

Seitz, A.

1913. Subfamily: Heliconiinae. In “Macrolepi-
doptera of the World. The American Rho-
palocera”, 5: 375-402.

Swain, R. B.

1948. The insect guide. New York: Doubleday
& Co., 261 pp.

Tinbergen, N., B. J. D. Meeuse,

K. K. Boereme, & W. W. Variosseau

1943. Die balz des sampfalters, Eumenia (Saty-
rus) semele (L). Zeitschr. Tierpsychol., 5:
182-226.

Williams, C. R.

1930. The migration of butterflies. London:
Oliver & Boyd, xi + 473 pp.

196

Zoologica: New York Zoological Society

[40: 16: 1955]

EXPLANATION OF THE PLATES
(Photographs by Rosemary Kenedy)

Plate I

Courtship of Heliconius erato hydara: Sequence
of Stages (Photographed at 1/2000 sec.)

Fig. 1. Stage I. Male approaches resting female
from rear.

Fig. 2. Stage I (cont.). By rapid fluttering of
wings, male sends current of air against
female.

Fig. 10. Female giving practically full response:
Characteristics of Fig. 9, plus apposition
of forewings (1/2000 sec.).

Fig. 11. Male (upper) entering Stage II, showing
exposed end of silver friction surfaces,
bearing invisible scent scales on anterior
margin of upper hindwings. Female is at
left, back to camera, abdomen slightly
elevated (1/2000 sec.).

Fig. 3. Stage I (cont.). Male backs off for another
forward dart. Female shows beginning of
response by starting to flatten hindwings.
Meanwhile her abdomen is erected, the
scent gland extruded from the penultimate
segment and the forewings held apposed
(see Plate II, Fig. 10). Male’s partly un-
coiled proboscis is unusual in normal
courtship (cf. Plate II, Figs. 13, 14).

Fig. 4. Stage I (cont.). Male again flutters wings
against female; sometimes, as in this pic-
ture, actually touching her wings with his
own.

Fig. 5. Stage II. Male moves above and toward
front of female; now the scent scales on
the usually overlapping margins of his
fore- and hindwings are uncovered.

Fig. 6. Stage II (cont.). Note outward expansion
of female hindwings, and her partly de-
pressed antennae.

Fig. 7. Stage III. Male prepares to alight beside
female. Note elevated female abdomen
with extruded scent organ visible as pale
spot at tip.

Fig. 8. Stage III (cont.). This is followed by
copulation (see Plate III, Fig. 16).

Plate II

Heliconius erato hydara, courtship: Details, and

atypical courtship behavior.

Fig. 9. Female (right) giving partial response to
courting male. Stage I: Her abdomen is
elevated, the scent gland extruded and
hindwings lowered and vibrated; the fore-
wings however, are not apposed (cf. Fig.
10) (1/2000 sec.).

Fig. 12. Courtship, Stage II, photographed by slow
flash (1/200 sec.), indicating speed of
male motion. Note that speed almost
stops fluttering of female hindwings.

Fig. 13. Atypical courtship behavior: Male in final
stages of courtship, never completed, pal-
pates female head, thorax and forewings
with his forelegs and uncoiled proboscis
(1/2000 sec.).

Fig. 14. Atypical courtship behavior: Male alights
in front of female, facing her, palpating
her at intervals as above. Female is making
partial courting responses, the scent gland
partially extruded, hindwings slightly
opened, forewings not apposed (1/2000
sec.).

Plate III

Fig. 15. Setup for painting wings of living butter-
fly: Left, carbon dioxide cylinder with
glass-topped box for anaesthetizing insect.
Center, spreading board with butterfly
held in place by strips of paper. Back-
ground and to right, fast-drying lacquers,
solvent, and brushes for painting; ether is
used for partially removing oily coating of
scales before painting.

Fig. 16. A male butterfly with forewing band
painted blue above, black below, copu-
lating with normally colored female.

Fig. 17. All-red felt model butterfly being attached
by insect pin to magnet on end of split
bamboo wand.

Fig. 18. Black felt model with scarlet paper spots
fastened to wand, being fluttered behind
female, to simulate Stage I of normal male
courtship (1/2000 sec.).

CRANE

PLATE I

FIG. 2

FIG. 1

FIG. 3

FIG. 4

FIG. 7

FIG. 6

FIG. 8

IMAGINAL BEHAVIOR OF A TRINIDAD BUTTERFLY. HELICONIUS ERATO HYDARA HEWITSON,
WITH SPECIAL REFERENCE TO THE SOCIAL USE OF COLOR

CRANE

PLATE II

itiviiuaiiiiiiimtiiitiiiuiitHMBiiMMfa**

■SnUtSHminiiiiiiaiiiBSMuiiaiiiiinitiit^

ilniutifmtMavafitMMMif

iMliiVMtfnnaBtaataMBaiBBiaaitfHMBVitaiatBfi

inni II iS n rililS iiMS liU li bl

^ffiSlSiSKSSM

PI|viiiBBfi«B«ifivi|BiatBin
Qn»miunaBBiaBiiHBr~^
i«iiaii«aiiuniitaiiliii
|iBiiaMiiafnM^aatt^BaBi

iiBiet&Ti.biexc'

inBBraBBBVlBSBa^'

laiiiiiiiafii’iiM

, Jflr^lBIBBBBB'Bfa

tiav naiButiitiBBt

la■k«■ft1.BB9k

iiUBiiii.tBait

IBBItIBBBkVBBI

jmtUBBBI'.': Rt

.tBnBIBBBBBIMtSBTI
— mii'fliaiiBBi
.-jatBBkinaiBBr
iBlnBBaaBBBILBBB.
■BlBlBIBIBiaiaBB

•aatBiBiBtnilBi

I *B B« BB n «• M 9

BinminttaaMMi

■ CllfiMMBtBBBBItl

v>7t'tiiiiBnBitinnMnBt
eii BGIIKIIIimr ‘-1BBI
«MBBt?KsnBy>«BIC.'. If Bl

FIG.

FIG, II

FIG. 10

FIG. 12

FIG. 13

FIG. 14

IMAGINAL BEHAVIOR OF A TRINIDAD BUTTERFLY. HELICONIUS ERATO HYDARA HEWITSON.
WITH SPECIAL REFERENCE TO THE SOCIAL USE OF COLOR

' ':r'r '

s

r;

O'

■ t

t

f'

i

V,

"Cf;

•U

'••• I

A

/

r<>B

CRANE

PLATE III

IMAGINAL BEHAVIOR OF A TRINIDAD BUTTERFLY, HELICONIUS ERATO HYDARA HEWITSON,
WITH SPECIAL REFERENCE TO THE SOCIAL USE OF COLOR

3

K

:■

*> ■■■:• .;.- -. “• <r

.. '^AiiV'V' .■;

J‘'«

' .„ ifSPSAii. 'T; . • . • .:;v,©.iV-' ,. 'i g

■f ■!

1 '

' ji/.4' V' „■ ■■ V ,

■ J". .

ii?

} y..

?s;.^--"

!%'o, i

i^ - ,:'fr

-■3"

fflSe. a

(t . ■ ., '.

'S';.,'

'^•:;^i4 'At'

„ ;V u'll -'<■

.■'Vi

>:

17

Three More Gymnotid Eels Found to Be Electrogenic

Christopher W. Coates

New York Aquarium, New York Zoological Society
(Plates I-III)

IT has been reported previously (Lissmann,
1951; Coates et ah, 1954; Coates, 1954)
that three species of knifefishes, Family
Gymnotidae, other than the well-known Electric
Eel, Electrophorus electricus, are electrogenic.
Three more species, never before regarded as
electrogenic, are here reported as producing
electric discharges more or less continually, and
apparently in patterns peculiar to each species.

The three species are Apteronotus albifrons,
of which two individuals were examined, Steato-
genes elegans and Sternarchus oxyrhynchus. All
three were found to emit electrical impulses of
low amplitude and great regularity.

The discharges were measured by electrodes
inserted into the fresh water in which the fishes
were swimming freely, connected to a cathode
ray oscillograph. The temperature of the water
was 25° C. and the water was that to which the
fishes were accustomed. The electrodes were in-
sulated to their tips, which were silver, and were
spaced at varying intervals as indicated in the
figures. They were held as close to the test fish
as possible without disturbing it, both when the
fishes were swimming and when they were lying
quietly at the bottom of the tank. Voltages
measured in our experimental conditions did not
exceed 400 millivolts.

A. albifrons was represented by two indivi-
duals, one of which was 215 mm. long and had
been in captivity for more than two years. This
specimen was presumed to be adult. The other
was 50 mm. long and was obviously immature.
Oscillographic traces of the discharges of these
are shown in Plate I, Figs. 1 & 2. The regularity
of the discharges of the adult is obvious, in con-
trast to the irregularity and lack of pattern of
the other. There was no apparent difference
between recordings made while the fish were
either resting or swimming.

The pattern of discharge exhibited by Steato-

genes elegans, Plate I, Figs. 3 & 4, showed that it
discharged regularly while resting (Fig. 3) and
equally regularly, but about four times as fast
when it was deliberately disturbed, (Fig. 4).
This is an indication that the frequency of the
discharges is centrally controlled.

Plate II shows the varying magnitude of the
discharges of A. albifrons when the electrodes
were placed near different parts of the body. This
suggests that the organ producing the discharge
is located in the caudal extremity, as might be
expected in this family. The extreme regularity
of the discharges is quite apparent in these fig-
ures.

The discharge of Sternarchus oxyhynchus,
Plate III, Figs. 11 & 12, shows the same regu-
larity and almost the same frequency as that of
A. albifrons, that is, about 1,000 discharges per
second. Such high, steady rates of discharge are
rarely encountered in any physiological systems.

References

Coates, C. W.

1954. What we are learning about electric
fishes. Animal Kingdom, 57 (6): 182-186.

Coates, C. W., M. Altamirano & H. Grundfest
1954. Activity in electrogenic organs of knife-
fishes. Science, 120 (3125): 845-846.

Lissmann, H. W.

1951. Continuous electrical signals from the
tail of a fish, Gymnarchus niloticus Cuv.
Nature, 167 (4240): 201-202.

EXPLANATION OF THE PLATES
Plate I

Oscillographic traces of discharge of Apteronotus
albifrons. Time scale (lower line), 1,000 c. p. s.

Fig. 1. Adult fish.

Fig. 2. Immature fish.

197

198

Zoologica: New York Zoological Society

[40: 17: 1955]

Oscillographic traces of discharge of Steatogenes
elegans. Time scale (lower line), 100 c. p. s.

Fig. 3. While resting.

Fig. 4. When disturbed.

Plate II

Localization of discharge of Apteronotus albifrons.
Time scale (lower line), 1,000 c. p. s.

Fig. 5. Drawing showing segments measured on
oscillograph.

Fig. 6. With electrodes on “A” in drawing.

Fig. 7. With electrodes on “B” in drawing.

Fig. 8. With electrodes on “C” in drawing.
Plate III

Oscillographic traces of discharge of three species,
with electrodes 2.3 cm. apart.

Fig. 9. Apteronotus albifrons, slow sweep.

Fig. 10. Apteronotus albifrons, fast sweep.

Fig. 11. Sternarchus oxyrhynchus, slow sweep.
Fig. 12. Sternarchus oxyrhynchus, fast sweep.
Fig. 13. Steatogenes elegans, slow sweep.

Fig. 14. Steatogenes elegans, fast sweep.

COATES

PLATE I

THREE MORE GYMNOTID EELS FOUND TO BE ELECTROGENIC

/

(:

I

,\S..

?

■’’A

COATES

PLATE II

THREE MORE GYMNOTID EELS FOUND TO BE ELECTROGENIC

COATES

PLATE III

THREE MORE GYMNOTID EELS FOUND TO BE ELECTROGENIC

[1955]

Zoologica: Index to Volmiie 40

199

INDEX

Names in bold face indicale new
genera, species or varieties; numbers
in bold face indicate illustrations,-
numbers in parentheses are the series
numbers of papers containing the
plates listed immediately following.

A

Actinopyga agassizi, 47, 49, 99

Agraulis vanillae, 170

Alligator mississippiensis, 103, (10)

PI. I

Anaea morvus morvus, 31
Anolis carolinensis, 108, (10) PI. Ill
Anteos maerula, 170
Apteronotus albifrons, 197, (17) Pis.
MU

Asclepias curassavica, 178

B

Biblis hyperia, 30
Bidens pilosa, 178
Boa constrictor, 112
Brassolis sophorae sophorae, 31
Browallia americana, 178
Buddleia variabilis, 178

C

Caligo eurilochus minor, 30, 31
illioneus saltus, 31
Catonephila numilia, 30
Centropogon surinamensis, 178
Cephaelis tomentosa, 178
Chaetodipterus faber, 85, (7) Pis. I
& II

Chalcolopidus virens, 31
Charina bottae, 112
Chelydra serpentina, 107
Chrysemys picta, 105, (10) PI. II
Citrus nobilis, 30, 32
Clemmys guttata, 106
Cnemidophorus tessellatus, 109, (10)
PI. Ill

Colobura dirce dirce, 30
Convolvulus, 178
Cordia alliodora, 29

cylindrostachya, 29, 178
Cosmotella adjuncta nigricollis, 6
Cosmotoma, 1
adjuncta, 2, 4
rubella, S
lasciata, 6, 7
melzeti, 17, 18
nigra, 19, SO
olivacea, 16
pallida, 21
sertifer, 10, 11
suluralis, 8, 2
triangularis, 12, 13
viridana, 14
CosmotomeUa, 22
zikani, 22, S3
Costus spiralis, 178
Crotalus viridis, 112

Ctenochaetus binotatus, 150, 164
cyanoguUaius, 150, 160
hawaiiensis, 150, 161, (15) PI. II
ntagnus, 150, 162
striatus, 150, 154, 155, (15) PI. I
strigosus, 150, 154, 159, (15) Pis. I &
II

tominiensis, 150, 163
D

Danaus plexippus, 170
plexippus megalippe, 29
Dibamus novae-guineae, 109
Dinia mena, (2) PI. Ill
Dryas julia, 179

E

Epidendron fragrans, 178
Erythrina micropteryx, 178
Eumeces iasciatus, 108, (10) PI. IV
obsoletus, 108, (10) PI. IV
Euploea, 30
Eurema albula, 170

G

Gerrhonotus coeruleus, 109, (10) PI.

Ill

multicarinalus, 109
Gonatodes fuscus, 109

H

Hamelia erecta, 178
Heliconia hirsuta, 178
humilis, 178

Heliconius erato hydara, 140, (13)

Pis. I-V, 167, (16) Pis I & II

Isabella, 179

melpomene, 141, (13) PI. VI
ricini, 170
sara, 170

Heliotropium indicum, 28, (2) Pis.
Mil

Hemidactylus turcicus, 109, (10)

PI. IV

Hibiscus, 178

Historis odium orion, 31

Hyla maxima, 180

Hymenitis andromica trifenestra, 29

I

Ithomia drymo pellucida, 29, (2)

PI. Ill

Ixora coccinea, 178

K

Kinosternon flavescens, 107

L

Lantana camara, 178
Lampropeltis doliata. 111
Lialis burtoni, 109
Lycorea ceres ceres, 29
Lygosoma laterale, 108, (10) PI. IV

M

Momordica charantia, 178
Morpho peleides insularis, 31

N

Nalrix sipedon. 111

O

Odontoloxozus, 31
Oncidium luridum, 178
Oodinium limneticum, 134
ocellatum, 133, (12) PI. I
Opheodrys vernalis. 111
Ophisaurus ventralis, 110, (10)

FI. Ill

Oxyopsis rubicunda, 180

P

Pachystachus coccinea, 178
Papilio anchises, 170
anchisiades, 170
neophilus, 170
Passiflora tuberosa, 193
Peridroma arethusa, 30
Phoebis sennae, 170
Phrynosoma cornutum, 107, (10)

PI. Ill

Plalophrys ocellatus, 93, (8) Pis. II
& III

Plumago capensis, 178
Polyohrus marmoratus, 180
Prepona antimache, 31, (2) PI. IV
demophon, 31
meander, 31, (2) PI. IV
Protogonius ochraceus, 31
Pseudemys scripta, 106
Pseudoscarus guacamaia, 145, (14)
PI. I

Pseudosphex melanogen, (2) PI. Ill
Pygopus lepidopus, 109
Pyrophorus pellucens, 31

S

Scarus croicensis, 145, (14) PI. I
punclulalus, 145
Sceloporus undulatus, 108
Senecio confusus, 178
Sparisoma chrysopterum, 145
pachycephalum, 145
Sphaerodaclylus cinereus, 109
Sphecosoma trinitatis, 28
Stagmatoptera septentrionalis, 180
Stagmomantis Carolina, 180
Steatogenes elegans, 197, (17) Pis. I
& III

Sternarchus oxyrhynchus, 197, (17)

PI. UI

Sternotherus odoratus, 107

T

Tabebuia pentaphylla, 178
serratifolia, 178

Tarantola mauretanica, 109, (10) PI.

rv

Terrapene Carolina, 106

200

Zoologica: Index to Volume 40

[1955]

Thalassoma bifascialum,

Pis. I-UI

Thamnophis sirtalis, 110,
Tithorea mopsa, 170
Tournefortia, 30
Trinecies inscriptus, 92,
lineatus, 91, (8) PI. I

125, (11)

(10) PI. V

(8) PI. I

Trionyx ferox, 107
Tussacia pulchella, 178

V

Verbena, 178

Victorina sleneles sleneles, 30
Victorina steneles, 30

W

Warszewiczia coccinea, 178

X

Xiphophorus maculatus, 53, 55, 57,
58, 60, 63-74, (6) Pis. MV
X helleri, 54

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

30 East Fortieth Street, New York 16, N. Y.

PUBLICATION OFFICE

The Zoological Park, New York 60, N. Y.

OFFICERS
PRESIDENT
Fairfield Osborn

SECRETARY
Harold J. O’Connell

VICE-PRESIDENTS TREASURER

Alfred Ely David H. McAlpin

Laurance S. Rockefeller
Donald T. Carlisle

ASSISTANT TREASURER
Percy Chubb, 2nd

SCIENTIFIC STAFF: Zoological Park and Aquarium

John Tee- Van Director

Leonard J. Goss Assistant Director

ZOOLOGICAL PARK

Richard H. Manville . Curator of Mammals

Grace Davall Assistant Curator,

Mammals and Birds

James A. Oliver Curator of Reptiles

Leonard J. Goss Veterinarian

Charles P. Gandal. . .Assistant Veterinarian

Lee S. Crandall General Curator

Emeritus

William Beebe Honorary Curator,

Birds

AQUARIUM

Christopher W. Coates . Curator & Aquarist

James W. Atz Assistant Curator

Ross F. Nigrelli Pathologist

Myron Gordon Geneticist

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . .Research Associate

in Histology

Homer W. Smith Research Associate

in Physiology

DEPARTMENT OF TROPICAL
RESEARCH

WiUiam Beebe Director Emeritus

Jocelyn Crane Assistant Director

Henry Fleming Entomologist

Rosemary Kenedy Field Assistant

John Tee-Van Associate

William K. Gregory Associate

AFFILIATES

C. R. Carpenter Co-ordinator, Animal

Behavior Research Programs

L. Floyd Clarke Curator,

Jackson Hole Research Station

SCIENTIFIC ADVISORY COUNCIL
A. Raymond Dochez Caryl P. Haskins
Alfred E. Emerson K. S. Lashley
W. A. Hagan John S. Nicholas

EDITORIAL COMMITTEE

GENERAL Fairfield Osborn, Chairman

William Bridges Editor & Curator, James W. Atz Lee S. Crandall

Publications William Beebe Leonard J. Goss

Sam Dunton Photographer William Bridges James A. Oliver

Henry M, Lester. . .Photographic Consultant Christopher W. Coates John Tee-Van

•i

i

»';■■■ ' V,";-

)■