W.'Sfc, ^

III mil y

.o

^ -fj

%

<*■

\ \ c/

isslr / X

Cp A

\N b,

A'^ .=*===3^ ^>

A Jpsfttk

.5^

V

^ bk
> ^

A>

,1)1

^ "% H

X <*■

a a

m y \

act ko0
s<^ /^Plfex A

%

>

^ \ *■
Xy _. <fe -

yg&s> x

vw^ vwy \~ *>

ptsHfll

ya

Yc Jp

v

cP %

**0, vv .sn

% «

Q, - _,y.,:

\ A w

A' A,

£> Kf*

xP \

%, sC

A1 ^

-A5

A /

.O

A A^

£■ b.

III Hr %

1 V

I'

,A

* bv @gjfa \ /

xo^ % W ^ ^o

A> b, 'AgSSSA A5" A,

A

% /

A A,

W

&L ',/■ \ AA x iAA

xV ^

A

A %

y \

<> A

,,\' v..-::';;- :;-3., _ (y^ A

* <„ Sfejfl y „. it

.cP V 1

% xb-

1, A> <A

^Tix^ %

% ^

A %

# A

xP A

A

V* ^Sail^ a A ^

A.

A-v A,

O, -V

/V V

.6"

**

v ^

A, >x
^A. A

,e> a i%'-r-b#^| a kf

c^v bp ^ b

Jp-M -„.. <A ; -

«^.gfcafa~ <i\, ^

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 47 • 1962 ® NUMBERS 1-16

PUBLISHED BY THE SOCIETY
The ZOOLOGICAL PARK, New York

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

630 Fifth Avenue, New York 20, N. Y.

PUBLICATION OFFICE

The Zoological Park, Bronx 60, N. Y.

OFFICERS

PRESIDENT VICE-PRESIDENT SECRETARY TREASURER

Fairfield Osbom Laurance S. Rockefeller George W. Merck David H. McAlpin

SCIENTIFIC STAFF:

William G. Conway . . Director,

Zoological Park

Christopher W. Coates. .Director, Aquarium
John Tee- Van ...... General Director

Emeritus

ZOOLOGICAL PARK

Joseph A. Davis, Jr. . . Curator, Mammals

Grace Davall Assistant Curator,

Mammals and Birds

William G. Conway. . Curator, Birds

Peter Ames Assistant Curator,

Birds

Herndon G. Dowling . Curator, Reptiles
Charles P. Gandal. . . Veterinarian

Lee S. Crandall General Curator

Emeritus

Roland Lindemann .... Consultant in Mam-
mal Management

AQUARIUM

James W. Atz Curator

Carleton Ray Associate Curator

Ross F. Nigrelli Pathologist & Chair-

man of Department
of Marine Biochem-
istry & Ecology

Klaus D. Kallman .... Geneticist

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . . Research Associate

in Histology

Sophie Jakowska Research Associate

in Experimental
Biology

Louis Mowbray Research Associate

in Field Biology

GENERAL

William Bridges . . Editor & Curator,

Publications

Dorothy Reville . . Editorial Assistant

Sam Dunton Photographer

Henry M. Lester. .Photographic Consultant

DEPARTMENT OF TROPICAL
RESEARCH

Jocelyn Crane Directing Curator

John Tee-Van Associate

William K. Gregory .... Associate

AFFILIATE

L. Floyd Clarke Director,

Jackson Hole Biological Research Station

EDITORIAL COMMITTEE
Fairfield Osbom, Chairman
James W. Atz Lee S. Crandall

William Bridges Herndon G. Dowling

Christopher W. Coates John Tee-Van
William G. Conway

Contents

Z-<9 />

d? 71/ • D / Q

v: ^7

1.

Part 1. May 9, 1962

PAGE

Serological Correspondence Among Horseshoe “Crabs” (Limulidae). By
Carl N. Shuster, Jr. Plate 1; Text-figures 1 & 2 1

2. Relative Abundance, Microhabitat and Behavior of Some Southern Appa-

lachian Salamanders. By Robert E. Gordon, James A. MacMahon &
David B. Wake. Text-figure 1 9

3. Sexual Discrimination and Sound Production hr Uca pugilator Bose. By

Michael Salmon & John F. Stout. Plate I; Text-figures 1&2 15

4. Low Temperature Effect on the Testicular Cell-components of the Common

Indian Toad, Bufo melanostictus Schneider. By S. L. Basu. Plate I; Text-
figure 1 21

5. Biology and Behavior of Damon variegatus Perty of South Africa and
Admetus barbadensis Pocock of Trinidad, W. I. (Arachnida, Pedipalpi).

By Anne J. Alexander. Text-figures 1-5 25

6. Breeding Activities, Especially Nest Building, of the Yellowtail ( Ostinops

decumanas) in Trinidad, West Indies. By William H. Drury, Jr. Text-
figures 1-4 39

7. Further Observations on the Pilot Whale in Captivity. By David H. Brown.

Plates I & II 59

Part 2. September 15, 1962

8. A Field Study of the Black and White Manakin, Manacus manacus, in

Trinidad. By D. W. Snow. Text-figures 1-21 65

9. Longevity of Fishes in Captivity, as of September, 1956. By Sam Hinton. 105

10. Hybridization Experiments in Acheilognathine Fishes (Cyprinidae.
Teleostei). A Comparison of the Intergeneric Hybrids between Tanakia
tanago and Rhodeus spinalis and Rhodeus ocellatus from Korea and Japan.

By J. J. Duyvene de Wit. Plates I & II; Text-figure 1 117

Part 3. November 29, 1962

PAGE

11. Observations of the Sound Production Capabilities of the Bottlenose
Porpoise: A Study of Whistles and Clicks. By William E. Evans & John

H. Prescott. Plates I-IV; Text-figures 1-6 121

12. Notes on the Biology of Some Trinidad Swifts. By D. W. Snow. Text-

figure 1 129

13. The Genetics of Some Polymorphic Forms of the Butterflies Heliconius

melpomene Linnaeus and H. erato Linnaeus. I. Major Genes. By J. R. G.
Turner & Jocelyn Crane. Plate I; Text-figure 1 141

Part 4.

14. Effects of Hybridization on Pigmentation in Fishes of the Genus Xipho-

phorus. By James W. Atz. Plates I-VIII 153

15. A Field Study of the Golden-headed Manakin, Pipra erythrocephala, in

Trinidad. By D. W. Snow. Text-figures 1-6 183

16. The Natural History of the Oilbird, Steatomis caripensis, in Trinidad, W.I.

Part 2. Population, Breeding Ecology and Food. By D. W. Snow. Plates
I-IV; Text-figures 1-4 199

Index to Volume 47 223

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 47 • PART 1 • MAY 9, 1962 • NUMBERS 1-7

PUBLISHED BY THE SOCIETY
The ZOOLOGICAL PARK, New York

Contents

PAGE

1. Serological Correspondence Among Horseshoe “Crabs” (Limulidae). By

Carl N. Shuster, Jr. Plate I; Text-figures 1 & 2 1

2. Relative Abundance, Microhabitat and Behavior of Some Southern Appa-

lachian Salamanders. By Robert E. Gordon, James A. MacMahon &
David B. Wake. Text-figure 1 9

3. Sexual Discrimination and Sound Production in Uca pugilator Bose. By

Michael Salmon & John F. Stout. Plate I; Text-figures 1&2 15

4. Low Temperature Effect on the Testicular Cell-components of the Common

Indian Toad, Bufo melanostictus Schneider. By S. L. Basu. Plate I; Text-
figure 1 21

5. Biology and Behavior of Damon variegatus Perty of South Africa and
Admetus barbadensis Pocock of Trinidad, W. I. (Arachnida, Pedipalpi).

By Anne J. Alexander. Text-figures 1-5 25

6. Breeding Activities, Especially Nest Building, of the Yellowtail ( Ostinops

decumanus ) in Trinidad, West Indies. By William H. Drury, Jr. Text-
figures 1-4 39

7. Further Observations on the Pilot Whale in Captivity. By David H. Brown.

Plates I & II........... 59

1

Serological Correspondence Among Horseshoe “Crabs” (Limulidae)1

Carl N. Shuster, Jr.

Department of Biology, New York University2

(Plate I: Text-figures 1 & 2)

Introduction

The few researchers, Graham-Smith
(1904), Boyd (1937), Wilhelmi (1942,
1944) and Leone & Webb (1952), that
have reported on systematic aspects of the blood
serum of horseshoe “crabs,” have all compared
Limulus serum with that of animals outside of
the family. Since the present report is the first
on the comparative serology of the Limulidae,
it constitutes an introduction to the serological
relationships among horseshoe “crabs.” Three of
the four extant species, representing the two sub-
families, were serologically compared: the
North American species Limulus polyphemus
(Linnaeus) Muller, Limulinae, and two Indo-
Pacific species, Carcinoscorpius rotundicauda
(Latreille) Pocock, and Tachypleus gigas
(Muller) Leach, Tachypleinae. The fourth, the
remaining Asiatic species, Tachypleus tridenta-
tus Leach, was not studied.

The data, obtained by turbidimetric measure-
ments and agar-diffusion records of the precipi-
tin reaction between antigens and antibodies,
were limited by the available supply of serum
from the Indo-Pacific species.

Our serological findings provide a basis for a
more comprehensive study on the system atics of
the Limulidae. There is questionable support for

1 Based upon Part II of a dissertation in the Depart-
ment of Biology accepted by the Faculty of the Grad-
uate School of Arts and Science, New York University,
in partial fulfillment of the requirements for the degree
of Doctor of Philosophy.

This study was aided by a grant from the Research
Council, during the summers of 1953 and 1954, and is
listed as a contribution from The Serological Museum,
Bureau of Biological Research, Rutgers— The State Uni-
versity of New Jersey.

Contribution No. 21 : University of Delaware Marine
Laboratories.

2 Present address: Department of Biological Sciences,
University of Delaware, Newark, Delaware.

the taxonomy proposed by Pocock (1902), who
divided into three genera the four extant species
previously confounded under Limulus.

Materials and Methods

Hemolymph samples were collected from
adult limuli in eleven populations of Limulus,
from Florida to Maine, during the summer of
1953. Sera for this study were selected from
animals in four widely separated populations:
Pleasant Bay, Massachusetts; Providence River,
Rhode Island; Miles River, Maryland; and Sar-
asota Bay, Florida. Sera from two of the three
Indo-Pacific species, C. rotundicauda (samples
from eleven specimens) and T. gigas (one sam-
ple), were obtained by Dr. D. S. Johnson and
his associates in the Department of Zoology,
University of Malaya. Without this generous as-
sist from Dr. Johnson the present study would
not have been possible; the author owes him a
personal note of gratitude. Similar contacts with
other researchers, to obtain T. tridentatus serum,
proved nonproductive.

The hemolymph was preserved for study in
the following manner (see Boyden, 1953, for
general procedure). (1) A horseshoe “crab”
was held in ventriflexion over a large glass
beaker, or a funnel, and the arthrodial (dorsi-
mesal) membrane in the hinge region cut deeply
to puncture the underlying heart. Hemolymph
usually flowed rapidly, directly from the heart;
the flow could be increased by slowly moving the
opisthosoma up and down like a bellows. (2)
The hemolymph was stored overnight in an ice
chest or a refrigerator to allow the amoebocytes
to coagulate and form a “clot.” (3) Then, after
the clot formed, the serum was decanted into
a graduate. (4) A stock solution of merthiolate
( 1 gram merthiolate powder in 100 milliliters of
distilled water) was then added, 1 milliliter of
the stock solution to each 49 milliliters of serum
and the solution shaken well. (5) The merthio-

1

2

Zoologica: New York Zoological Society

[47: 1

lated serum was bottled and stored in a cold
room until used (temperature regulated at
1°C.).

Antisera to the serum of female Limulus from
six different populations and the Tachypleinae
sera were produced, using rabbits as producers.
A reciprocal testing program was followed, limi-
ted only by the small amount of Tachypleinae
serum available. Full reaction curves for each
of the antigens were run on the Libby photron-
reflectometer (photron’er). In graphing the re-
sults, galvanometer readings on the photron’er
were expressed as turbidity units along the or-
dinate, antigen concentration as antigen dilution
along the abscissa. Each successive cell of anti-
gen dilution, 1 through 15, contained one-half
as much antigen as the preceding cell. The first
cell, 1, was assumed to contain 5% antigen, or
a titer of 1/20. Thus, cell 15 has an assumed an-
tigen titer of 1/327,680. The exact protein con-
centration is not critical when whole titration
curves are available for comparison. Total tur-
bidity values for each curve were obtained by
adding all the individual turbidities of each an-
tigen dilution tube in the test. A value of 100
percent, was assigned to the area under the hom-
ologous curves and the heterologous values then
expressed as percentages of the homologous
total.

These serological tests were conducted at The
Serological Museum, Bureau of Biological Re-
search, Rutgers— The State University of New
Jersey, under the direction of Drs. Ralph J.
DeFalco and Douglas G. Gemeroy in 1954 and
1955. The techniques and procedures elaborated
by Dr Alan A. Boyden, director of The Serologi-
cal Museum, and co-workers (Boyden & De-
Falco, 1943; Bolton, Leone & Boyden, 1948;
and Leone, 1949, 1950) were used. The author
is grateful to Dr. Boyden for reviewing the
manuscript of this report.

Another measure of the precipitin reactions
among Limulidae sera and antisera were ob-
tained from a series of tests using the double-
diffusion, Jennings-modified (Jennings &
Malone, 1954; Jennings, 1954), three-depot
method of Ouchterlony (1949). The desired
test combinations of sera and antisera, each
reactant placed separately in the corner reser-
voirs of a triangular plastic plate, separately
diffuse through the agar. The diffusing reactants
immediately combine in the region of equiva-
lent concentrations (Munoz, 1954). At first in-
visible, this combination appears in time as a
band or zone of precipitin which can be photo-
graphed. Dr. Sheldon A. London conducted
these tests at the University of Delaware in
1957, in conjunction with a problem on micro-

bial serological correspondence. The author is
indebted to Dr. London for the photographs
and the data he obtained.

Results

Reciprocal precipitin reactions between the
sera and antisera of three species of Limulidae,
L. polyphemus, T. gigas and C. rotundicauda,
are summarized in Tables 1 and 2, and are
shown on graphs (Text-fig. 1), a diagram
(Text-fig. 2) and photographs (Plate I).

The curves (Text-fig. 1) obtained from the
precipitin tests are typical unimodal curves
which show no departure in spatial arrange-
ment from a typical reaction curve (see for ex-
ample: Leone, 1949; Boyden & DeFalco, 1956).
The relative positions of the peaks of the curves
are also typical, with the peaks of the heterolo-
gous curves usually to the right of and exceeded
by the peak of the homologous reaction.

Turbidimetric results are summarized by a
diagram (Text-fig. 2) indicating the serological
distance between the species using a modifica-
tion of Boyden’s (1932) method. When the
reaction values between two species are averaged
and subtracted from 100, a serological “yard-
stick” is obtained which indicates the relative
distances between the species. Thus the triangle
in Text-figure 2B gives the relative distance be-
tween the three species according to a turbidi-
metric measurement of the precipitin reaction
of their serum proteins in the reciprocal tests.

Reactions among sera and antisera for three
Limulus from widely separated populations pro-
duce the typical homologous pattern of coalesced
zones and form four precipitin bands in the agar
plates (Plate I, Figures A, B). The crossed
zones produced by tests between the three spe-
cies show the unrelated character of at least
two major components of the reactants (Plate
I, Figures C, D) or one component (Plate I,
Figures E, F). The Tachypleus-Carcinoscorpius
reactants appear to produce the coalesced zones
typical of strong serological correspondence
(Plate I, Figures G, H). The asymmetry of the
zones is ascribed to unequal concentrations of
the reactants, with zone displacement toward
the reservoir containing the reactant of lesser
concentration.

Discussion

An examination of the turbidimetric data
(Table 1 and Text-fig. 1) reveals the marked
difference in the magnitude of the reaction of
each of the antigens to the antisera and the
shift to the right of the heterologous reaction
curves in the region of antigen excess.

1962]

Shuster: Systematic Serology of Limulidae

3

Text-fig. 1. Precipitin titration curves showing the order of serological relationships among
three species of horseshoe “crabs”: Limulus polyphemus, Carcinoscorpius rotundicauda and
T achy pie us gigas. The antisera and test sera (antigens) are given in the legend for each series
of curves. The area under the homologous reaction curve is highlighted by the shaded background.

4

Zoologica: New York Zoological Society

[47: 1

Table 1. Relative Amounts of Precipitin Reaction, Recorded as Turbidity
in Photronreflectometer Tests

These four sets of data (A-D) are represented graphically in

Text-fig. 1.

Antiserum

Antigen

Peak

Tube

Number

1

Peak

urbidity Units
Total

Percent, of
Homologous
Reactions

A. L. polyphemus

L. polyphemus*

(8)

501

2641

100

L. polyphemus *

(8)

481

2502

95

C. rotundicauda

(8)

361

1556

59

T. gigas

(9)

230

1212

46

B. L. polyphemus

L. polyphemus

(5)

305

1924

100

C. rotundicauda

(7)

247

1284

67

T. gigas

(7)

183

1025

53

C. C. rotundicauda

C. rotundicauda

(7)

336

1549

100

T. gigas

(7)

189

871

56

L. polyphemus

(8)

112

477

31

D. T. gigas

T. gigas

(7)

320

1769

100

C. rotundicauda

(7)

171

1208

68

L. polyphemus

(7)

143

616

35

* Serum from specimens in two geographically different populations.

A comparison of all the reaction curves sug-
gests that the several different rabbits employed
in the production of the antisera reacted to the
sera and produced comparable specificities of
antibodies. The shift to the right of the hetero-
logous curves, with respect to the homologous
reaction, generally denotes a lesser amount of
serological reaction between the heterologous
sera and the test antisera. We are concerned
here, however, with the relative placement of
the species, from the homologous to the lesser
reaction of the heterologous antigens, in each
of the apparently linear serological series. These
series are the basis for the two-dimensional dia-
gram (Text-fig. 2) of the relationships among
the three species. For a general discussion of
the principles of systematic serology and the im-

Table 2. Summary of the Turbidimetric Data
on the Reciprocal Precipitin Reactions, Re-
corded as Percent, of the Homologous Reaction.
These data are graphically portrayed in Text-fig. 2.

Antiserum

Antigen

Limulus

Carcinoscorpius

Tachypleus

L. polyphemus

100

59

46

100

67

53

C. rotundicauda

31

100

56

T. gigas

35

68

100

plications of the results of precipitin testing see
Boy den (1942).

The serological placement of L. polyphemus
slightly closer to C. rotundicauda than to T. gigas
(Table 2 and Text-fig. 2B) requires further test-
ing. Evidence (Shuster, 1958) indicates that L.
polyphemus has a greater correspondence in
morphometric and morphological characters
with the genus Tachypleus than with Carcino-
scorpius. Data for a definitive systematic evalu-
ation, however, are not available. For instance,
it is possible that T. tridentatus, which was not
tested, is closer serologically to Limulus than T.
gigas is. Further, since the Indo-Pacific serum
supply was limited and only two rabbits were
employed, one each, for the production of the
antiserum to T. gigas and C. rotundicauda, a
sharply defined serological relationship is im-
possible at this time.

The turbidimetric data are consistent, as the
reciprocal tests show (Table 2). The greater
reaction between Limulus serum and the anti-
serum of the two other species (46 and 59%
or 53 and 67%), than that from the reciprocal
reactions (35 and 31%), may indicate that the
Limulus serum has one or more additional chem-
ical components. The number of these compon-
ents is suggested by several kinds of studies.

According to Allison & Cole (1940), hemo-
cyanin is the only protein present in the serum
of Limulus polyphemus after complete removal
of the clot. They also noted that the clotting pro-
cess, which involves the amoebocytes, does not

1962]

Shuster: Systematic Serology of Limulidae

5

Text-fig. 2. Two stages in the construction of a “serological yardstick” diagram show the relative distances
between the three species: Limulus polyphemus, Carcinoscorpius rotundicauda and Tachypleus gigas.

remove any appreciable amount of hemocyanin
from the serum. Limulus hemocyanin exists,
within the pH range of 5.2-10.5, in four different
and stable molecular species, but dissociation
and a fifth component appear when the pH is
less than 5.2 (Eriksson-Quensel & Svedberg,
1936). The electrophoretic pattern of Limulus
hemocyanin reveals at least five components
(Deutsch & McShan, 1949) and at least four
prominent entire bands are revealed by the hom-
ologous test for Limulus in the agar-diffusion
precipitin technique. The later technique reveals
one and perhaps two major reaction bands that
do not coalesce in any of the reciprocal reactions
between Limulus and the two Indo-Pacific spe-
cies. Reactions between only Tachypleus and
Carcinoscorpius give entire but fewer bands.

Boy den (1943) found that the serological
correspondence among marine crustaceans of
the same genus averaged 46%. Later, in similar
studies, Leone (1949, 1950) distinguished be-
tween species of the same genus within certain
families of Crustacea. In his studies the magni-
tude of serological correspondence between
crustaceans of the same genus was as low as
31%, but most values were above 70%, with
some as high as 89% for closely related species.
In our Limulidae study the heterologous reac-
tions ranged from 30 to 67%. Although not
directly comparable, the serological distance
between species of the same genus of marine
Crustacea, when used as a “ruler” for compari-
son with the values among the Limulidae, sug-
gests that the three species studied are congen-
eric.

Interpretation of the “distances” provided by
the “serological ruler” is arbitrary, since “dis-
tance” depends upon a sliding scale which in
turn is an expression of the specificity, i.e., the
discriminating capacity, of the antiserum which
varies from antiserum to antiserum. Also, there

are too few species and the present data are too
scanty to establish more definite serological
measurements of either species or genera within
the Limulidae. Further, the best available refer-
ence for comparison is a distant taxonomic
group comprised of many families, the Crus-
tacea. The Limulidae is a very small taxon by
comparison, but it is one that has roots in an
old history, as measured by the geologic age of
the Xiphosura, and one that reflects a conserva-
tive evolution.

If we consider that the present serological
findings coincide with the current sub-Limulidae
taxonomic categories given in the beginning of
this report, then the serological distance between
species, genera and families is indeed narrow.

The author wishes to express his gratitude and
sincere thanks to the men most closely asso-
ciated with the inception and progress of his
studies on the Limulidae. Dr. Thurlow C. Nel-
son, Rutgers— The State University of New Jer-
sey, counseled the writer in the formative years
of the research. Dr. Harry A. Charipper, New
York University, was largely responsible for the
culminating stages of the dissertation prepara-
tion. Dr. Alfred C. Redfield, Woods Hole Ocean-
ographic Institution, and Dr. William C. Cole,
Rutgers— The State University of New Jersey,
were instrumental in aiding the study through
counsel and research grants from their respec-
tive institutions.

Summary

From the turbidimetric measurements of pre-
cipitin reactions among the Limulidae sera and
the rabbit-produced antisera it can be concluded
that:

1. The precipitin reactions produced typical
titration curves.

2. No demonstrable differences were found
in the homologous reactions based on Limulus
sera from different populations.

6

Zoological New York Zoological Society

[47: 1

3. The sera of the two Indo-Pacific Limulidae
tested, Carcinoscorpius rotundicauda and Tach-
ypleus gigas, are readily distinguished from the
sera obtained from different populations of Lim-
ulus polyphemus.

4. Serological correspondence is greatest be-
tween the two species of Tachypleinae, but C.
rotundicauda may be serologically closer than
T. gigas to L. polyphemus.

5. The three species of Limulidae are con-
generic when compared to the same magnitude
of serological correspondence among marine
Crustacea.

6. The taxon including the Limulidae is limi-
ted in number of species, yet it exhibits a long
geologic history and a conservative evolution.
If the current classification is correct, wherein
three of the four species are not congeneric, then
the taxonomic categories of the extant species,
genera and families are indeed narrow.

7. The results and other information re-
viewed in the discussion suggests that Limulus
serum may have one or more components not
present in the sera of the two Indo-Pacific spe-
cies.

Data from agar-diffusion records of the pre-
cipitin reaction revealed:

1. Four major bands of reaction or compon-
ents in Limulus homologous reactions.

2. A lack of correspondence for one and per-
haps two major components between the Indo-
Pacific species and Limulus.

3. An apparent complete correspondence be-
tween the two species of Tachypleinae.

The present study was limited by the lack of
an adequate supply of sera from the Indo-Pacific
species. More extensive serological studies would
be of great value in interpreting the systematics
and evolution of the Limulidae. These studies
should include serum of the species Tachypleus
tridentatus and employ, for example, the im-
munoelectrophoretic method (first used by
Grabar & Williams, 1953, and emphasized as
a requirement in systematic serology by
Williams, 1956), among other techniques.

Literature Cited
Allison, J. B., & W. H. Cole

1940. The nitrogen, copper and hemocyanin
content of the sera of several arthropods.
J. Biol. Chem., 135:259-265.

Bolton, E. T., C. A. Leone & A. A. Boyden

1948. A critical analysis of the performance of
the photronreflectometer in the measure-
ment of serological and other turbid sys-
tems. J. Immunol., 58: 169-181.

Boyd, W. C.

1937. Cross-reactivity of various hemocyanins
with special reference to the blood protein
of the black widow spider. Biol. Bull.,
73:181-183.

Boyden, A. A.

1932. Precipitin tests as a basis for a quantitative
phylogeny. Proc. Soc. Exp. Biol. & Med.,
29:955-957.

1942. Systematic serology: A critical apprecia-
tion. Physiol. Zool., 15:109-145.

1943. Serology and animal systematics. Amer.
Nat., 77:234-255.

1953. Zoological collecting expeditions and the
salvage of animal bloods for comparative
serology. Science, 118:57-58.

Boyden, A. A., & R. DeFalco

1943. Report on the use of the photronreflect-
ometer in serological comparisons. Phys-
iol. Zool., 16:229-241.

1956. On measuring serological correspondence
among antigens. Serol. Mus., Bull., No.
16:3-8.

Deutsch, H. F., & W. H. McShan

1949. Biophysical studies of blood plasma pro-
teins. XII. Electrophoretic studies of the
blood serum proteins of some lower ani-
mals. J. Biol. Chem., 180:219-234.

Eriksson-Quensel, Inga-Britta, & T. Svedberg

1936. The molecular weights and pH-stability
regions of the hemocyanins. Biol. Bull.,
71:498-547.

Grabar, P., & C. A. Williams

1953. Methode permettant l’etude conjugee des
proprieties electrophoretiques et immuno-
chimique d’un melange de proteines. Ap-
plication au serum sanguin. Biochemica et
Biophysica Acta., 10:193-194.

Graham-Smith, G. S.

1904. Blood-relationship amongst the lower Ver-
tebraia and Arthropoda. In “Blood Im-
munity and Blood Relationship” (G. H. F.
Nuttall). Cambridge Univ. Press.

Jennings, R. K.

1954. Diffusion-precipitin studies of the com-
plexity of antigen mixtures. II. The num-
ber of zones formed by one antigen. J.
Bacteriol., 67:565-570.

Jennings, R. K., & Frances Malone

1954. Rapid double diffusion precipitin analysis.
J. Immunol., 72:411-418.

Leone, C. A.

1949. Comparative serology of some brachyuran
Crustacea and studies in hemocyanin cor-
respondence. Biol. Bull., 97:273-286.

1962]

Shuster: Systematic Serology of Limulidae

7

1950. Serological systematics of some palinuran
astacuran Crustacea. Biol. Bull., 98:122-
127.

Leone, C. A., & H. E. Webb

1952. Serological relationships among some dis-
tantly related arthropods and mollusks.
System. Zool., 1:184 (Abstract).

Munoz, J.

1954. The use and limitations of serum-agar
techniques in studies of proteins. In “Sero-
logical Approaches to Studies of Protein
Structure and Metabolism.” Rutgers Univ.
Press, New Brunswick.

OuCHTERLONY, 6.

1949. An in-vitro test of the toxin-producing
capacity of Corynebacterium diphtheriae.
Lancet, 256:346-348.

Pocock, R. I.

1902. The taxonomy of recent species of Limu-
lus. Ann. Mag. Nat. Hist., 9:256-262.

Shuster, C. N., Jr.

1958. On morphometric and serological relation-
ships within the Limulidae, with particular
reference to Limulus polyphemus (L.).
Diss. Abs., 18:371-372.

WlLHELMI, R. W.

1942. The application of the precipitin technique
to theories concerning the origin of verte-
brates. Biol. Bull., 82:179-189.

1944. Serological relationships between the Mol-
lusca and other invertebrates. Biol. Bull.,
87:96-105.

Williams, C. A., Jr.

1956. Antigen-antibody reactions in gels and
systematic serology. Proc. Fourteenth In-
ternal. Congr. Zool.: 336-337.

8

Zoologica: New York Zoological Society

[47: 1: 1962]

EXPLANATION OF THE PLATE
Plate I

These retouched photographs of triangular
serum-agar plates show the visible bands of pre-
cipitin reaction among the three species: Limulus
polyphemus, Carcinoscorpius rotundicauda and
Tachypleus gigas. The set of three letters around
each triangle indicate the side from which the serum
of each species (designated by CAPITAL LET-
TERS: L, C, and T) or antiserum to each species
(lower case letters: 1, c, and t) diffused. The eight
sets of reactions, designated by the letters A through
H, are described in the text.

SHUSTER

PLATE I

SEROLOGICAL CORRESPONDENCE AMONG HORSESHOE •'CRABS” (LIMULIDAE)

2

Relative Abundance, Microhabitat and Behavior
of Some Southern Appalachian
Salamanders1

Robert E. Gordon, James A. MacMahon &

David B. Wake

University of Notre Dame and University
of Southern California

(Text-figure 1)

Introduction

SOME 44 years have elapsed since Emmett
Reid Dunn described Plethodon yonah-
lossee and discussed the relative abun-
dance and behavior of the salamanders associ-
ated with it in the Linville, Avery Co., North
Carolina, area (Dunn, 1917). Many of Dunn’s
observations concerning the plethodontid sala-
manders were supplemented in a later publica-
tion (Dunn, 1920) and by other workers
(Breder & Breder, 1923; Bailey, 1937; Gray,
1939; Wood, 1947; Hairston, 1949; and Pope,
1950).

The type locality of Plethodon yonahlossee is
given as “near the Yonahlossee Road about 1 Vi
mile from Linville, North Carolina” (Dunn,
1917) . The general area in which Dunn worked
“most of the time in sight of the road” embraces
a hill which “rises to about 4,400 feet,” with the
road having an average altitude of 4,100 feet.
In August, 1960, the authors made observations
and collections along the Yonahlossee Road
(frequently referred to as “Old Yonahlossee
Road”) in the vicinity of View Rock, a vantage
point overlooking N. C. Hwy. 220. The collec-
tion was prompted by a need for an ontogenetic
series for investigations of osteology and my-
ology (DBW), comparative physiology (JAM)
and behavior and periodicity (REG).

This paper presents ( 1 ) a report of the collec-
tion, (2) a comparative description of the type

1The field work was supported in part by a grant, Na-
tional Science Foundation G-13327, to the first author.

locality after 44 years, and (3) a discussion of
the habitat, relative abundance and behavior of
collected salamanders.

Description of the Collecting Stations

Dunn (1917) gives no detailed description of
the flora of the type locality for Plethodon yon-
ahlossee nor of the surrounding area, but we in-
fer from his remarks (see p. 595) that the mixed
mesophytic forest typical of the undisturbed
southern Appalachian Mountains prevailed.

Our collections were made at two stations.
The first was a gentle slope 0.5 miles below View
Rock (toward Linville); the second was 0.6
miles above View Rock and approximately 1.5
miles from the Linville entrance to the Yon-
ahlossee Road. In our judgment, the second sta-
tion embraces the type locality as designated by
Dunn. The area has been subjected to at least
one lumbering since Dunn’s visit, as indicated
by the presence of well-decayed stumps up to
four feet in diameter and numerous fallen logs.

The flora of both stations consists of a second
growth mixed mesophytic forest, aptly described
as a “rich woods.” At the first locality the col-
lecting was concentrated along and on the down-
hill side of the Yonahlossee Road. We estimate
the area covered as one acre. No detailed notes
of the flora were made, but the area had less
understory, fewer stumps and logs than the
second locality.

Distinctive landmarks encountered during the
collecting permitted us to measure the total area
covered at the second locality. We collected an

9

10

Zoologica: New York Zoological Society

[47: 2

area on the southeast slope (15% gradient) 100
X 200 yards above the road, and 100 X 100
yards below the road. The elevation of the pres-
ent road is 4,300 ft. at this point. The area was
bounded on the east by a fast-flowing stream and
on the west by a dense understory of nettle
which prohibited effective collecting. The pres-
ent road bed was laid down to the north, or
uphill, from that travelled by Dunn, and rem-
nants of the old roadbed formed a distinctive
landmark below the present road, a distance of
100 yards. The total area in which we collected
is approximately 2.1 acres.

The flora of the second station was examined
both at the time of our collection and on a
subsequent trip over the surrounding area. The
canopy is dominated by a mixture of red and
white oaks, maples, buckeyes, ironwood, gum
and tulip trees. A few hemlocks are scattered
through the area. The understory is sparse al-
though seedlings of the above, plus chinquapin,
catbriar and an occasional rhododendron thicket
are present. The most prevalent plants compos-
ing the herbal layer are jewelweed, nettle, shield
fern, pipsissewa, goldenrod, twayblade, bell
flower and clumps of pinesap.

The sandy surface is covered by a 3- to 4-inch
loam on top of which leaf litter, minimally 3
inches in depth, occurs. Stumps and logs in var-
ious stages of decay are numerous and seem to
constitute important physical features of the
habitat of Plethodon yonahlossee. Occasional
outcrops of the underlying granite occur. The
leaf litter, logs, stumps, rock outcrops and bases
of plants conceal the openings to the refugia of
the salamanders.

Methods

Daytime collecting was restricted to the hours
of 4:00 to 6:15 P.M. Both stations were ex-
amined. Six man-hours were spent hand-collec-
ting salamanders exposed by log rolling and rock
turning.

Night collecting with light from two head-
lamps and a Coleman gas lantern was carried
out between 8:30—12:00 P.M. The salamanders
were collected by hand, but few objects were
turned. The entire night sample was taken from
the station 0.5 miles above the View Rock. In
all, 10.5 man-hours were spent in collecting at
night.

The two samples were maintained separately,
so as to permit a comparison between day and
night sampling.

The Sample

The total number of salamanders taken, in
16.5 man-hours of collecting at the two stations,

was 661. The sample includes the following
forms: Diemictylus v. viridescens, Desmogna-
thus ochrophaeus carolinensis, D. m. monticola;
Plethodon c. cinereus, P. g. glutinosus, P. yon-
ahlossee, P. jordani metcalfi, and Eurycea bis-
lineata wilderae.

A difference between day and night collecting
was first noted by Bailey (1937), who obtained
much better results for P. yonahlossee at night.
Nocturnal collecting is effective for almost all
species of salamanders; however, no quantita-
tive reports of collecting results have been made.
Diurnal collecting yielded 28.3 salamanders per
man-hour, whereas nocturnal collecting pro-
duced 46.8 per man-hour. In addition, habitat
disturbance is minimal during night collecting.
The destruction of habitat during diurnal collec-
ting activities has led to reduction of population
size (as measured by availability) in some areas
that have been visited from time to time by one
of us; other areas that have been collected with
the same intensity at night with little or no hab-
itat destruction continue to produce large sam-
ples of animals.

Another important point not mentioned by
previous workers is the composition of a sample
as affected by difference in time of collecting
(Table 1). While it can be argued that animals
taken during the day are not available for
sampling at night, only a part of the daytime
sample (in our best judgment, 20 per cent.)
was taken from the same locality as was sampled
at night. The difference in sample composition
is striking. If the species were ranked from most
abundant to least abundant, not one of the six
most abundant taken in the daytime would re-
tain its position when based on nocturnal abun-
dance. This difference is a result of the behavior
and consequent availability of the species en-
countered. Desmognathus seemed to occupy the
most superficial cover of all species and conse-
quently was more available in the daytime than
when it was active at night. Plethodon yon-
ahlossee and P. jordani retreat into deep bur-
rows and hence are not discovered by rock
rolling, log turning, etc., as is P. glutinosus,
which occupies a more superficial diurnal
refuge. Note that the three most abundant ple-
thodons represent only 36 per cent, of the diur-
nal sample, but 72 per cent, of the nocturnal
sample.

Relative Abundance

Although Dunn collected for three days and
we for only one, we believe that the results are
comparable because the mean number of in-
dividuals for one of Dunn’s days is equivalent to
our single day. Dunn’s sample of August, 1916,

1962] Gordon, MacMahon & Wake: Southern Appalachian Salamanders 1 1

Table 1. Percentage Composition of Sample by Day, by Night,
and Total Compared with that of Dunn (1917)

Percent of Total

Species

D + N*

Dunn, 1917

Day

Night

Diemictylus viridescens (eft)

1.8

0.0

0.5

9.5

Desmognathus ochrophaeus

57.6

24.2

32.8

29.3

Desmognathus monticola

2.9

1.0

1.5

0.0

Plethodon cinereus

1.2

2.3

2.0

10.4

Plethodon glutinosus

10.6

6.7

7.7

10.2

Plethodon yonahlossee

6.5

16.7

14.1

5.0

Plethodon jordani

18.8

48.7

40.9

32.6

Pseudotriton ruber

0.0

0.0

0.0

0.2

Eurycea bislineata

0.6

0.4

0.5

2.8

Total number individuals

170

491

661

462

*Day and night combined.

probably was made during diurnal forays. There
is no indication of time of collecting in either
his 1917 or 1920 paper, and night collecting
for reptiles and amphibians has been prevalent
only in the last twenty years. In view of these
points, we believe that the only valid comparison
that may be drawn is between our daytime data
and those of Dunn. Several differences between
the two sets of data (Table 1) may be noted,
including (1) an increase in the percentage of
Desmognathus ochrophaeus in the sample (29.3
to 57.6%); (2) a decrease in the percentages
of Plethodon cinereus (10.4 to 1.2%) and P.
jordani (32.6 to 18.8%); and (3) the relative
constancy of P. yonahlossee and P. glutinosus.
Eurycea was noticeably rare, but this form is
lacking in samples from other areas of the south-
ern Appalachians which in recent years had been
very productive.

The data may indicate either a gradual shift
in the composition of the population, or a short
term fluctuation that may be cyclic. The latter,
in view of previous collecting (1948-50, by
REG) , seems to be the case for Eurycea.

Whatever drastic effect lumbering might have
had upon the area has been negated by subse-
quent succession, although one is tempted to
suggest that the 20 per cent, decrease in the
combined plethodon group (primarily P. cin-
ereus and P. jordani ) from that of 44 years ago
may be attributed to lumbering. At the same
time, the addition of stumps and logs in various
stages of decay would seem to enhance the posi-
tion of P. yonahlossee, since these appear to
represent conspicuous elements in its habitat
(Pope, 1950).

Pope (1950) discusses the relative abundance
of Plethodon yonahlossee , P. glutinosus and P.
/. metcalfi. His daytime data (obtained in July

and August, 1949) for P. yonahlossee and P.
glutinosus are converted to percentages and
compared to that of Dunn (1917 and 1920,
combined) and our data for 1960 (Text-fig. 1).
The percentage distribution for all daytime
samples (except that of Pope from Comers
Rock) falls into the same general pattern. How-
ever, we believe that the best measure of rela-
tive abundance lies in a combined collecting,
one which samples the individuals not only in
their refugia, but also as they are active on the
surface. A reversal of the pattern occurs when
our data are pooled to illustrate this point. The
reversal would be even greater if the individuals
taken at night were considered separately. The
latter procedure would be misleading, but per-
haps not as much as consideration of a daytime
sample alone.

Minimal Available Density

Because of the paucity of data to indicate the
number of salamanders per unit area in the
southern Appalachian Mountains, we have cal-
culated a density figure for the four most com-
mon species on the Yonahlossee Road (Table
2). We recognize that these data have inherent
weaknesses. The figures do not represent the
total number of salamanders per unit area, and
do not represent crude density. Test & Bingham
(1948), working with P. cinereus in Michigan,
showed that the number of animals present on
the surface (and available for capture) at any
one time represents only a portion of the total
population present in the area. The term mini-
mal density is appropriate because ( 1 ) not every
animal observed was captured; (2) a portion of
the daytime sample was taken from the area
collected at night; and (3) an estimate of the
surface area covered in the daytime collecting
was made. Two sets of figures are presented.

12

Zoological New York Zoological Society

[47: 2

Text-fig. 1. The relative abundance of P. yonahlossee and P. glutinosus at different localities and under
different collecting conditions (A— Iron Mt., B— Buck Mt., C— Comers Rock, data from Pope, 1950; D— Lin-
ville 1917-20, day, data from Dunn; E— Linville 1960 day; and F— Linville 1960 night).

The data for the nocturnal sample were collected
from a measured area (2.1 acres), but the
sample is biased by the daytime collecting in
the area. Those for the total sample were calcu-
lated from an estimated total area of 3.1 acres.
Thus the figures represent minimal available
densities and should be treated with caution.

We suggest that future collectors may find
that sampling (even of the “one-stop” type)
from a measured area will make their data on
relative abundance more meaningful if the min-
imal available density is calculated for each
species.

Microhabitats, Behavior and Competition

Pope (1950) points out that the existence of
P. glutinosus and P. jordani at Linville “is of spe-
cial interest and calls for further investigation in
view of their ecological segregation elsewhere.”
His subsequent remarks (p. 87) imply that the
sympatry of the two may be due to disturbed
conditions brought about by lumbering. That
this cannot be the case is evident from (1)
Dunn’s (1917) statement regarding “. . . the
primitive condition of flora and fauna, and being
rendered accessible by the splendid Yonahlossee

Road, is a paradise . . .” and (2) the existence
of P. glutinosus and P. jordani together in the
second-growth woods in approximately the same
relative proportion today (1:5, respectively) as
existed in Dunn’s sample (1:3).

In view of the inability of competent taxono-
mists to distinguish between P. glutinosus and
the southern representatives of P. jordani, we
prefer to believe that a genetic difference in the
two taxa which permits one to readily distin-
guish between them in the northern portion of
the P. jordani range, also reinforces their ecolog-
ical isolation where the two are sympatric.

Table 2. Minimal Available Density per Acre
for the Four Most Abundant Salamanders on
the Yonahlossee Road, Linville,

North Carolina

Nocturnal

Sample

Total

Sample

Desmognathus ochrophaeus

56.7

70.0

P. glutinosus

15.7

16.5

P. yonahlossee

39.1

30.0

P. jordani

113.8

87.4

1962]

Gordon, MacMahon & Wake: Southern Appalachian Salamanders

13

Pope (op. cit.) postulates competition be-
tween P. yonahlossee and P. glutinosus on the
basis of similarity of “habitat niches” and food
items. He was unable to demonstrate a convinc-
ing difference in diet. The fact that P. glutinosus
is more readily available during the day than
P. yonahlossee (as indicated by our collections
and those of Dunn and Pope) leads us to think
that there is a difference in microhabitat or be-
havior between the two taxa. Vernberg (1955)
found that P. glutinosus was less photosensitive
than P. cinereus. Our collections reveal an earlier
peak of activity for P. glutinosus than for P.
cinereus with P. yonahlossee intermediate. This
seems to indicate that P. glutinosus is less light-
sensitive than P. yonahlossee. P. glutinosus either
(1) does not penetrate the subsurface to the
depths inhabited by P. yonahlossee, or (2) takes
advantage of its relatively less sensitivity to light
and comes to the surface (beneath cover) dur-
ing daylight hours more frequently than P. yon-
ahlossee. In either case, P. glutinosus seems to
be effectively isolated by microhabitat from P.
yonahlossee, at least enough to reduce spatial
competition. As long as an abundant food sup-
ply exists these two species can be considered
only as potential and not actual competitors.

Concerning the habitat of P. yonahlossee, we
are in essential agreement with Pope (1950) that
this species is not “restricted to a zone within 100
feet” of streams as reported by Hairston ( 1949) .
The eastern margin of our plot was bounded by
a stream, but there was no evidence that P. yon-
ahlossee was any more abundant near the stream
than it was toward the nettle patch on the west-
ern boundary.

Our observations of animals retained in the
laboratory are of interest here. The animals were
retained on wet leaves in large finger bowls with
an excessive amount of moisture in the bottom.
A series of pustules appeared on the skin of P.
yonahlossee but not on the other plethodons re-
tained in the same bowls. P. yonahlossee was ob-
served more often on top of the leaves, than were
the other taxa and this may represent moisture
avoidance behavior.

Activity

Time was noted at intervals throughout the
collecting period and determination of activity
peaks of the various species was attempted. If
peaks of abundance may be considered an index
to peaks of activity (see Hairston, 1949), there
is a suggestion that adults of the plethodon taxa
are isolated during the active portion of the diel
cycle on a temporal basis.

The young of all these species were especially
prevalent before 9:00 P.M. Subsequent observa-

tions indicate that the young appear shortly after
dusk and attain a peak of abundance by 8:00
P.M., 1 hour after sundown at this locality.
Adults of P. yonahlossee attained a peak in
abundance between 9:00 and 10:00 P.M., with
a minor peak between 11:30-12:00 P.M. Be-
tween 10:00 and 11:00 P.M. both P. yonahlos-
see and P. glutinosus were often observed in re-
fugia with only their heads exposed. P. jordani
(juveniles and small adults) were present
throughout the evening, but large adults (and
P. cinereus ) were definitely more abundant after
10:00 P.M. Although our sample of adult P.
glutinosus was not large, we have the impression
(at Linville and elsewhere) that the peak of ac-
tivity outside the refugia is slightly earlier than
that of P. yonahlossee (8-9:00 P.M.) These ob-
servations represent our concensus recorded im-
mediately after the end of collecting. The obser-
vations were confirmed by one of us (JAM)
who visited the area during August, 1961.

Our observations on behavior of the pletho-
dons at Linville substantiate and supplement
those of Dunn (1917) and Pope (1950). Pleth-
odon jordani was observed on the leaf litter, near
the bases of trees, logs and in open areas. Be-
tween 8:30 and 10:00 P.M., individuals seemed
sluggish and were captured with the same ease
as an individual found beneath a log during the
day. However, from 10:00-12:00 P.M., this spe-
cies became more active and agile and less easily
caught. This latter behavior never approached
that of P. yonahlossee. P. jordani climbs more
often than the other three species of Plethodon.
Many were collected on trunks and low branches
of shrubs up to 3.5 feet from the ground. Groups
were observed feeding on fungal gnats around
decaying fungi, or on fruit flies and other insects
at the base of trees from which sap flowed.
Plethodon glutinosus was moderately abundant
and easily caught in the early evening. No climb-
ing was observed; practically all individuals were
at the bases of plants, near logs or lying with
their heads exposed in the openings of refugia.
After 11:00 P.M., P. glutinosus was conspicu-
ously absent from the surface area. This obser-
vation is not in accord with that of Hairston
(1949), who reported an abundance peak at
11:00 P.M. in the Black Mountains during late
July.

As described by Dunn and Pope, P. yonah-
lossee is the most agile of all eastern plethodons.
Usually only one opportunity is available to cap-
ture an individual. If the collector misses, the
animal retreats into a refugium for the evening.
Numerous individuals were first seen with then-
heads sticking out of refugia. Others were mov-
ing at the bases of trees, or beside fallen rotted

14

Zoologica: New York Zoological Society

[47:2: 1962]

logs and stumps. In most instances, this species
was associated with a log over 10 inches in diam-
eter, with not more than 1 to 3 inches of the log
below the surface. A thick layer of leaf litter
accumulation at the log— ground interface was a
prerequisite. P. yonahlossee climbs more than
P. glutinosas or P. cinereus, but less than P.
jordani.

Summary

The type locality of Plethodon yonahlossee
near Linville, Avery Co., North Carolina, was
visited in August, 1960. The flora and physical
aspects of two sampling areas, 2.1 acres and an
estimated 1 acre, are described.

Diurnal collecting yielded 28.3 salamanders
per man hour, as opposed to 46.8 salamanders
per man hour at night. Ranking of the species
according to percentage composition of the sam-
ple is shown to vary with the time in which the
sample is taken.

The relative abundance of each species en-
countered is compared with Dunn’s figures of
44 years ago. The most significant differences lie
in a two-fold increase of Desmognathus ochro-
phaeus and a marked decrease in Plethodon
cinereus and P. jordani; Plethodon yonahlossee
and P. glutinosus have remained relatively con-
stant.

The relation of P. yonahlossee to P. glutinosus
is examined in terms of relative abundance in
diurnal collections made by Pope (1950), Dunn
(1917, 1920) and ourselves. Five localities are
involved, yet the percentage distribution of the
two species remains approximately the same at
all but one locality. Competition between these
two taxa, postulated by Pope ( 1950) , is believed
to be reduced by differences in microhabitat and
behavior.

The minimal available density is calculated for
the four most common species at the type local-
ity. This density term is explained and its use as
a quantitative basis for determination of relative
abundance is suggested.

Temporal isolation in activity between age
groups and the different species of Plethodon
was observed and is discussed. General observa-
tions on behavior and microhabitat are offered.

References

Bailey, Joseph R.

1937. Notes on plethodont salamanders of the
southeastern United States. Occ. Pap. Mus.
Zool., Univ. Mich., 364: 1-10.

Breder, C. M., & Ruth B. Breder

1923. A list of fishes, amphibians and reptiles
collected in Ashe Co., North Carolina.
Zoologica, 4(1): 1-23.

Dunn, E. R.

1917. Reptile and amphibian collections from
the North Carolina mountains, with espe-
cial reference to salamanders. Bull. Am.
Mus. Nat. Hist., 37(23): 593-634.

1920. Some reptiles and amphibians from Vir-
ginia, North Carolina, Tennessee and Ala-
bama. Proc. Biol. Soc. Wash., 33: 129-138.

Gray, Irving E.

1939. An extension of the range of Plethodon
yonahlossee. Copeia 1939, (2): 106.

Hairston, Nelson G.

1949. The local distribution and ecology of
plethodontid salamanders of the Southern
Appalachians. Ecol. Monogr., 19(1): 47-
73.

Pope, Clifford H.

1950. A statistical and ecological study of the
salamander Plethodon yonahlossee. Bull.
Chi. Acad. Sci., 9(5): 79-106.

Test, F. H„ & B. A. Bingham

1948. Census of a population of the red-backed
salamander ( Plethodon cinereus). Am.
Midi. Nat., 39: 362-372.

Vernberg, F. John

1955. Correlation of physiological and behavior-
al indices of activity of Plethodon cinereus
and Plethodon glutinosus. Am. Midi. Nat.,
54(2): 383-393.

Wood, John T.

1947. Notes on North Carolina salamanders.
Copeia 1947, (4): 273-274.

3

Sexual Discrimination and Sound Production in Uca pugilator Bose

Michael Salmon & John F. Stout
Department of Zoology, University of Maryland,

College Park, Maryland

(Plate I; Text-figures 1 & 2)

Introduction

THE purpose of this study was to investi-
gate sexual discrimination and to describe
certain display behavior, including sound
production, by males of Uca pugilator. Speci-
mens and models with various combinations of
male and female appendages were introduced
to resident males. Sound production was ob-
served and the sounds produced were physically
analyzed.

The reproductive behavior of fiddler crabs
( Uca sp.) has been investigated by many work-
ers during the last fifty years. In all species,
copulation is usually preceded by display be-
havior, consisting of the “waving” of the male’s
large cheliped. In some species, sounds are pro-
duced which play a role in courtship. “Waving”
behavior is believed to define the territory of the
male and attract females (Crane, 1941).

Sound production in Uca involves movements
of the large cheliped. Dembowski (1925) de-
scribed a peculiar “shivering” of the large
cheliped in U. pugilator, lasting from one to
three seconds, which was used by one crab to
“call” another out of its burrow. Crane (1941)
reported sound production by “rapping” of the
cheliped against the ground in three Pacific
American species of Uca, and in U. pugilator
(Crane, 1943). Burkenroad (1947) made ob-
servations on U. pugilator. He did not believe
that rapping of the cheliped was involved in
sound production in this species, as he failed to
detect any disturbance of sand grains below the
cheliped of a male that had just produced
sounds. No rattling of the dactyl or vibration
of the body could be detected. He concluded that
some other mechanism of sound production was
utilized. Rathbun (1914) described ridges,
thought to be stridulatory, on the large cheliped

and ambulatory legs of male U. musica. Auri-
villus (in Burkenroad, 1947) predicted, on
morphological grounds, that stridulation would
be found in the genus.

Burkenroad (1947) conducted experiments to
investigate sexual discrimination in Uca. He re-
leased normal males, males without the large
cheliped, and females of U. pugilator into an
area occupied by resident males. In all cases,
normal males evoked aggressive responses, but
males without the large cheliped and females
were courted by resident males. Altevogt
(1957), working with U. marionis and U. annu-
lipes, concluded that sexual discrimination by
males was based on the presence of the large
cheliped.

Materials and Methods

All observations and experiments were carried
out on the west side of Pivers Island, Beaufort,
North Carolina, during August, 1959 and 1961.
The population of U. pugilator resided in clear
areas of sand, either just above the high water
mark or about 10 metres inland.

Sexual discrimination by males was investi-
gated. In 1959, experiments were performed
which repeated those of Burkenroad (1947).
Males without the large cheliped, normal males,
and females were released, in that sequence,
and allowed to move through a 3 X 3 metre
study area where over 40 males were exhibiting
waving behavior. This sequence of releases was
repeated ten times and the reactions of the resi-
dent males to all released specimens were re-
corded. In 1961, other introduction experiments
were performed. A male that had previously
exhibited waving behavior was frightened into
its burrow. Two probes were placed in the sand
about 0.6 metre to either side of the burrow

15

16

Zoologica: New York Zoological Society

[47: 3

(Text-fig. 1). A dead specimen, killed by im-
mersion in dilute alcohol, was placed between
the probes. A long piece of thread was tied to
a leg on either side of the body and then placed
around one of the probes on the corresponding
side and back about 1.5 metre toward the ob-
server. The probes were placed in the sand so
that by pulling the thread attached to either
side of the dead crab, the specimen could be
moved toward one or the other probe and within
7.5 cm. of the burrow. Dead specimens were
introduced once to each of five resident males
in this order: normal females; females with an
attached large cheliped; normal males; and
males with the large cheliped removed. Also
introduced were two blue clay models of crab
bodies. One model had two small chelipeds at-
tached and the other, one large and one small
cheliped attached to the anterior face. Each in-
troduction consisted of slowly moving the dead
crab or model from one probe to the other. After
an experiment was completed, the male was
frightened into his burrow. Another dead crab
or model was prepared for the next test which
was performed about five minutes after the male
emerged from its burrow. Other introduction
experiments were performed using live male

Text-fig. 1. A clay model (A) with two small
chelipeds attached is used to demonstrate the tech-
nique employed in introduction experiments. By
pulling the thread (B) from either side, the model
could be moved between the probes (C) and
through the resident male’s territory. The male’s
burrow (D) was about 7.5 cm. from the model
when it passed through the male’s territory..

c

Text-fig. 2. The appearance of the large cheliped
of live male U. pugilator used in introduction ex-
periments to conspecific resident males. In the first
trial, the dactylopod was removed (A); in the
second trial, the anterior half of the protopod (B);
and in the third trial, the protopod was completely
removed (C). The parts of the claw outlined with
dashed lines represent the removed portions of the
appendage.

specimens. These were also tied with thread and
moved between the probes as described above.
Before each introduction, portions of the large
cheliped were removed (Text-fig. 2, A-C) . For
the first introduction, the dactylopod was re-
moved; for the second, the anterior half of the
protopod; and for the third, the entire protopod.
All three introductions were performed once in
the order described until each of five additional
resident males had been tested.

Sounds produced by males of U. pugilator
were recorded with a Magnecorder tape recorder
(model PT63-A) and a Dukane microphone
(model 7A150). The microphone was suspend-
ed above or placed on the sand 2.5 cm. from
the burrow. The sounds were analyzed with a
Kay Electric Company Sonograph Model Re-
corder (B).

Results of Field Observations

Resident males and females defended an area
of 7.5 to 15 cm. in diameter around their bur-

1962]

Salmon & Stout: Sexual Discrimination and Sound Production in Uca pugilator

17

rows from intruders of the same sex. Four resi-
dent females each approached by another
female advanced with chelipeds extended and
open. The intruder female then retreated from
the area. More than five resident males ap-
proached by intruding males exhibited aggres-
sive responses as follows. The resident male
oriented so that the large cheliped was held
slightly away from the body and with its broad
surface facing the intruder. If the intruder came
within 7.5 to 10 cm. of the resident male, the
latter advanced, opening its cheliped, until the
broad surfaces of both individuals’ large cheli-
peds came into contact. The resident then
pushed the intruder from its burrow, using its
cheliped as a shield. Occasionally, the resident
male pointed its opened claw toward the in-
truder. If the intruder advanced, the claws of
both individuals would be locked together and
the intruder was either pushed or flipped away
by the resident. If the intruder was more aggres-
sive, the resident male retreated into his burrow
but with the large cheliped opened and facing
upward out of the entrance. Intruding females
either evoked aggressive or courtship behavior
from resident males, the latter occurring more
frequently. Observations by the authors confirm
those made by Crane (1943) on the features
of waving exhibited by males. The chelipeds
were raised upward and, at the same time, the
animal rose on its “toes.” The claws were then
extended laterally and lowered, as was the body.
Finally, the chelipeds were returned to the an-
terior margin of the body. The frequency of
the waving motion was about once every two
seconds when no female was present. When a
female approached, the frequency increased to
more than twice that rate. The claw was not
brought back to the body, but raised and lowered
while laterally extended. The male oriented so
that the broad face of the claw was turned to
the female. If the female moved toward him,
the male would go toward his burrow in short
spurts of 2.5 to 5 cm., starting as the claw and
body were raised and stopping as they were
lowered. In five observations, the female fol-
lowed about 5 cm. behind the male as he moved
toward his burrow. The male then entered the
burrow, but the large cheliped remained extend-
ed out of the entrance. Waving of the large claw
continued, but when lowered it was vibrated
against the sand at the lip of the hole, producing
a rapid series of thumps. Similar sounds were
heard after the female entered the male’s bur-
row. On four occasions, lone males were ob-
served making these sounds in their burrows at
night by hitting the base of the large claw against
the side of the burrow.

Over fifty sounds produced by one male in his
burrow were recorded during the day. A sono-
graphic analysis of three sounds is presented
(Plate I). Each burst of sound had a duration
of 0.2 to 0.3 seconds. Of the 17 bursts analyzed,
7 consisted of 5 pulses; 5 of 4 pulses; 3 of 3
pulses; and 2 of 6 pulses. Each pulse was pro-
duced when the claw hit the surface of the
ground as it was vibrated. The sound energy
was concentrated between 85 and 2,000 cycles.
With sounds of higher intensities, a suggestion
of harmonics was present with frequencies up
to 10,000 cycles.

Results of Introduction Experiments

The results of introduction experiments per-
formed in 1959 were in agreement with those
of Burkenroad ( 1947) . Females and males with-
out the large cheliped were courted, but normal
males elicited aggressive responses from all resi-
dent males. Results of introduction experiments
performed in 1961 are shown in Table 1. In all
instances where dead crabs of either sex bore
an intact large cheliped, the resident males re-
sponded aggressively. These responses were the
same as those observed under natural condi-
tions. When the dead introduced specimens bore
no large cheliped, the resident males exhibited
waving behavior and sound production typical
of courtship. Neither waving nor aggressive
responses were exhibited toward clay models
with two small chelipeds attached. In two of
five introductions, aggressive responses were ex-
hibited to clay bodies bearing a large cheliped.
Live male specimens with the dactylopod re-
moved, and with the dactylopod and anterior
half of the protopod removed, all evoked aggres-
sive responses from resident males. Live males
with the protopod removed were courted by four
of five resident males.

Discussion

Even though the specimens and models were
not introduced to resident males in a random
order, the following tentative conclusions can
be made. Sexual discrimination in U. pugilator
is based on the presence or absence of the large
cheliped, regardless of the sex of the specimen
bearing the appendage. This appendage releases
aggressive behavior in male-male encounters,
and may be important in inducing sexual recep-
tivity in females. The posterior portion of the
protopod, the broadest surface of the claw, is
capable of eliciting aggressive responses from
other males, so that an intact claw is not neces-
sary for sexual discrimination.

The large cheliped also functions as a sound-
producing organ and the sounds have come to

18

Zoologica: New York Zoological Society

[47: 3

play an important role in courtship. Burkenroad
(1947) suggested that sounds substituted for the
attractive qualities of the large cheliped when
that appendage could no longer be seen by the
female. He found that sounds were produced
more frequently at night than during the day.
The results of our field observations seem to
support this view. When a male produces sounds
during the day, he is partially hidden from the
female by the lip of the burrow so that waving
motions of the claw could no longer be effective
visual stimuli. The waving motion is then modi-
fied to include a rapid vibration of the claw
against the ground. The sounds produced either
substitute for the visual stimuli of waving or
may have other functions in courtship.

The results of sonographic analysis as well
as field observations support the conclusion that

Table 1. Responses of Male Uca pugilator to
Conspecific Males and Females and to
Clay Models1

Number of Responses

Stimulus

Waving

Aggressive

Response

No

Apparent

Response

Dead Female

5

0

Dead Female:
large cheliped
attached

0

5

Dead Male

0

5

Dead Male:

5

0

large cheliped
removed

Clay Model: 5

Two small

chelipeds

Clay Model:
one large, one
small cheliped

Live Male: 0

dactylopod

removed

2 3

5

Live Male: 0 5

dactylopod and
anterior half
protopod re-
moved

Live Male: 4 1

entire protopod

removed

1 All tests with models and dead specimens were per-
formed on the same five resident males. Live individuals
were introduced to five other resident males.

stridulation is not used to produce sounds
occurring during courtship in this species. A
stridulatory sound is usually characterized by
the concentration of the maximum energy in
the higher frequencies. But rapid vibrations of
the cheliped against the ground would be ex-
pected to produce thumping sounds which would
show predominantly low frequency compo-
nents such as those analyzed in this study.

Summary

Courtship behavior of Uca pugilator was ob-
served in a natural population. Waving behavior
by males was exhibited when no female was
present, but when a female approached a male,
the waving motion was modified to produce
sounds by vibrating the claw against the ground.
A sonographic analysis of these sounds from
one specimen showed that the energy of the
sound was concentrated between 85 and 2,000
cycles and had a duration of 0.2 to 0.3 seconds.
The role of sound production in the reproduc-
tive behavior of the species is discussed.

Sexual discrimination by males was studied
by introducing models and specimens of either
sex, bearing or not bearing the large cheliped
(typical of male individuals), to resident males.
In all cases, resident males exhibited aggressive
responses to the specimens bearing the large
cheliped and courted specimens without the
large cheliped, regardless of the sex of the intro-
duced specimen. It was concluded that the domi-
nant visual cue in sexual discrimination was the
presence or absence of the large cheliped.

Acknowledgements

The authors would like to thank Dr. C. G.
Bookhout, Director of the Duke University
Marine Laboratory, for his encouragement and
aid in carrying out this study. Drs. Howard E.
Winn and Anthony R. Picciolo read the manu-
script and offered many helpful criticisms. This
study was supported by grants to Dr. Howard
E. Winn from the Office of Naval Research
(N. R. 104-489) and the U. S. Public Health
Service (B 3241).

References

Altevogt, R.

1957. Untersuchungen zur biologie, okologie,
und physiologie indischer winkerkrabben.
Zeitschr. fur Morph, und Okol. der Tiere,
46: 1-110.

Burkenroad, M. D.

1947. Production of sound by the fiddler crab,
Uca pugilator Bose, with remarks on its
nocturnal and mating behavior. Ecology,
28: 458-461.

1962]

Salmon & Stout: Sexual Discrimination and Sound Production in Uca pugilator

19

Crane, J.

1941. Eastern Pacific Expeditions of the New
York Zoological Society. XXVI. Crabs of
the genus Uca from the west coast of Cen-
tral America. Zoologica, 26: 145-207.

1943. Display, breeding and relationships of
fiddler crabs (Brachyura, genus Uca ) in
the northeastern United States. Zoologica,
28: 217-223.

Dembowski, J.

1925. On the “speech” of the fiddler crab, Uca
pugilator. Trav. inst. Nencki, Vol. Ill,
No. 48.

Rathbun, M. J.

1914. New genera and species of American
Brachyrhychous crabs. Proc. U. S. Nat.
Mus., 47: 117-229.

20

Zoologica: New York Zoological Society

[47: 3: 1962]

EXPLANATION OF THE PLATE
Plate I

Fig. 1. Sonographic analysis of three sounds pro-
duced by one male Uca pugilator in its burrow
during the day.

SALMON & STOUT

PLATE I

FIG. I

SEXUAL DISCRIMINATION AND SOUND PRODUCTION
IN UCA PUGILATOR BOSC

4

Low Temperature Effect on the Testicular Cell-components of the
Common Indian Toad, Bufo melanostictus Schneider

S. L. Basu

G. C. Bose Biological Research Unit, Bangabasi College, 19 Scott Lane, Calcutta 9, India

(Plate I; Text-figure 1)

THE degree of warmth, humidity, light
and other climatic features in combina-
tion with the internal rhythm are respon-
sible for controlling the sexual pattern of ani-
mals. The environmental temperature plays a
great role in the reproductive physiology of
poikilothermous animals such as fishes, amphib-
ians and reptiles, although this appears to be
lacking in warm-blooded birds and mammals
(Bullough, 1951, p. 29). The potentially contin-
uous type of spermatogenesis of Rana esculenta
could be suppressed by decreasing the tempera-
ture during summer time, and the reverse during
the winter induced activity (Galgano, 1934,
1935). Similar action on spermatogenesis was
also proved to be true when Triturus crist atus
cranifer (Galgano & Falchatti, 1940; Mazzi &
Galgano, 1949), Triturus viridescens (lift,
1942), Geotriton fuscus, Triturus alpestris,
Rana gracea and Rana latestei (Cei, 1942 a, b,
and 1944) were treated with high temperature
during winter and summer months. In the South
American wood frog, Leptodactylus ocellatus
typica, Cei (1948) and Rengel (1950) observed
a temperature tolerance limit both in summer
and winter and spermatogenesis is reported to
be impaired if the frogs are treated beyond that
range in both the seasons. However, Rengel
(1950) was unable to observe any such effect
when Leptodactylus ocellatus reticulatus was
treated at high tempearture ( ± 30°C) . Similarly,
Cei (1944) could not induce spermatogenesis
in Rana arvalis when it was treated with high
temperatures during winter. Witschi (1924) re-
ported that spermatogenesis is independent of
environmental temperature in Rana temporaria.
This was experimentally supported by Cei ( 1942,
1944) by keeping the frogs at high temperatures
during the winter. But van Oordt (1956a, b)

treated Rana temporaria at 5°C for two months
and observed the absence of spermatogenetic
activities. In Telmatobius schreiteri (Cei, 1949)
and Hyla raddiana andina (Caruso, 1949) sper-
matogenesis is continuous and is not affected by
the considerable low temperature of the high al-
titudes of the Andes mountains.

It is, therefore, evident that spermatogensis is
dependent on the temperature in most of the
Salientia, but with some specific variations. The
above reports are from the temperate zone with
the exception of some South American tropical
examples. Consequently, it appears useful to
study the influence of low temperature espe-
cially on tropical toads like Bufo melanostictus
where spermatogenesis is continuous (Mondal
& Basu, 1960; G. Church, 1960) and is not
affected by an average temperature fluctuation
of 38°-150C.

Materials and Methods

Mature male toads, Bufo melanostictus, were
collected from the vicinity of Calcutta and were
brought to the laboratory the next day. The body
weight and snout-to-vent length of all experi-
mental animals varied from 28-32 gms. and 70-
75 mm. respectively. Secondary sexual charac-
ters were carefully observed before treatment.
A group of ten toads was allowed to stay in a
controlled temperature room (±10°- 15°) and
another batch of five was kept at outside normal
room temperature for one month. This experi-
ment was performed during the months of Aug-
ust-September 1960 (Group A) and April-May
1961 (Group B). At the time of autopsy, body
weight, snout-to-vent length, testicular weight
and secondary sex characters were noted. The
right testis of all the individuals was fixed and
sectioned at 6p and stained with hematoxylin

21

22

Zoologica: New York Zoological Society

[47: 4

Table 1. Observations Taken from the Control and Low Temperature Treated Toads for Thirty Days
DURING THE MONTHS OF AUGUST-SEPTEMBER (1960) AND APRIL-MAY (1961).

Relative
Testis Wt.
(in mg.)

Testis Tubule
Diameter
(in p)

Spermatogenetic Stages!

Percentage of Tissue Components

Interstitium

Testis

Tubule

Misc.f

0

I

ii m

IV

V

August-September (1960) Group A

Control

70

52.5

47.0

26.0

40.5 75.5

4.0

5.2

10.8

87.5

1.7

(5)*

Treated

67

45.5

48.2

25.0

42.5 70.5

2.7

4.0

20.35

78.6

1.15

GO)*

±10°-15oC

30 days.

April-May (1961) Group B

Control

134

57.0

35.0

23.0

40.5 80.2

2.0

3.2

17.5

81.2

1.3

(5)*

(Treated)

135

56.5

37.0

22.5

27.5 75.7

2.2

4.0

18.2

80.5

2.3

(10)*
±10°-15°C
30 days.

^Figures indicate the number of toads used, flncludes blood vessels, vasa efferentia and some empty spaces. JStage O = Primary spermatogonia
at resting phase. Stage I — Secondary spermatogonia less than ten cells in a cell nest. Stage II = Secondary spermatogonia more than ten cells
in a cell nest. Stage III = Primary spermatocyte. Stage IV = Secondary spermatocytes. Stage V = Spermatids.

eosin. Different spermatogenetic stages were
counted and the percentage of tissue compon-
ents was calculated by a planimetric method.
Tubule diameter and relative testicular weight
were also recorded.

Results and Discussion
The relative testicular weight of both control
and treated toads of Group A (vide Table 1)
is well within the normal range, as was pre-
viously reported by Mondal & Basu (I960) in
dealing with the annual spermatogenetic cycle
of the present species. The overall picture of the
different spermatogenetic stages and the average
number show no distinct difference caused by the
treatment. A decrease, although not very re-
markable, in the number of cell nests from stage
III onward is observed among the treated toads
of Group A (August-September, 1960). Sperms
are embedded in the Sertoli cells and the va«a
efferentia remain closed in all the treated and
control toads during the months of August-
September, 1960. The most prominent differ-
ence is noted in the testis tubule diameter, which
is definitely narrow in the treated group (45.5/x)
when compared with the controls (52.5 p). Sim-
ilarly the percentage of interstitium has in-
creased in the treated batch of animals (20.3% )
as compared with that of the controls ( 1 0.8 % ) .
About 10% increase of the interstitial cellular
components due to low temperature and an
average shrinkage of Ip of the tubule diameter
are definitely significant (Text-fig. 1; Plate I,
A & B).

In toads of Group B (April-May, 1961) both
control and treated individuals show differences
of insignificant nature in their relative testis
weight, tubule diameter and also in their per-
centage of tissue components (Table I; Text-fig.
1 ) . So far as the different spermatogenetic stages
are concerned, an abrupt and sudden fall in the
cell nest number of the secondary spermatogonia
of stage II is noted in the treated group. The
number of spermatocytes has also been de-
creased but not as significantly as have the
spermatogonia. The sperms are found both in
scattered and bundled condition in the tubular
lumen, suggesting spermiating activity in the
toads (Plate I, C& D).

The experiments of Galgano (1934, 1935) on
Rana esculenta proved beyond doubt that tem-
perature can induce or suppress spermatogenesis.
On the other hand, in Rana temporaria Witschi
(1924) and Cei (1942, 1944) believed that
temperature caused no significant influence on
spermatogenesis. Later, van Oordt (1956 a, b)
and Galgano & Lanza (1951) proved that the
influence of temperature exists in Rana tempor-
aria if treated for a long period. All the above-
mentioned observations are limited to the frogs
of the temperate zone where normal environ-
mental temperature is very low. On the contrary
the frogs and toads inhabiting the tropical and
subtropical zones generally show a continuous
type of spermatogenesis and their temperature
tolerance limit is also variable. This appears to
be inherent and varies from species to species.

1962]

Basu: Temperature Effect on Testicular Cell-components of Bufo melanostictus

23

100-

90-

60-

70-

60-

50-

40-

30-

20-

10-

6R0UP A B CD E F

Text-fig. 1. Illustrating the change in the tubule
diameter (in p), interstitium percentage and testis
tubule percentage of control and treated toads in
Groups A, C & E (August-September) and Groups
B, D & F (April-May).

The normal cycle of Rana tigrina (Basu & Mon-
dal, 1961) and Bufo melanostictus (Mondal &
Basu, 1960; Church, 1960), inhabiting more or
less the same climatic zone, show a good deal
of difference particularly during the winter
months. A similar condition is reported by
Rengel (1950) with regard to a South American
species of Leptodactylus, which shows contin-
uous spermatogenesis. Rengel (1950) observed
that if Leptodactylus ocellatus reticulatus is ex-
posed to temperature higher than ±30°C there
is no harmful effect on spermatogenesis, but
this temperature appeared to be too high for the
testicular cell divisions of Leptodactylus ocella-
tus typica. The present experiment on Bufo mel-
anostictus reveals that very low temperature
treatment in the warmest days (±36°-40°C) of
the year (April-May, 1961) causes no signifi-
cant effect except the decrease of the secondary
spermatogonial number. But the toads of Group
A (August-September, 1960) under similar
treatment did not show any such remarkable
fall in the cell nest number of stage II. On the
contrary the tubule diameter and percentage of
the interstitium was remarkably affected by low
temperature in Group A. This proves that the
role of low temperatures on male gonads of

I [CONTROL
ITREATED

Testis

tubule

Testis

tubule

diameter

{ in)J)

Inierstitium

%

Bufo melanostictus is different during the sum-
mer and autumn months. In summer probably
the spermatogenetic stages are more susceptible
than is the interstitium of autumn. However, it
is expected that prolonged treatment for several
months may affect all the target organs during
the summer and autumn months. In Rana tem-
poraria, Witschi (1924) concluded that sperma-
togenesisis is to a large extent independent of
the environmental factors and consequently de-
pends on the internal rhythm determined by
genetic factors. But experiments of van Oordt
(1956) suggest that prolonged treatment may
change the condition. Cei (1948) and Rengel
(1950), from their experimental observations,
are of the opinion that there remains a certain
temperature tolerance limit in the spermato-
genetic field of the Salientia. Moreover, from
different reports so far available it appears that
the temperature tolerance range of the Salientia
varies from species to species even in the same
climatic zone. The present experiment also sug-
gests that in Bufo melanostictus seasonal varia-
tion causes some difference in the site of sensi-
tivity and also that it has a wide range of
temperature tolerance which is suitable for its
adaptation to this tropical climate.

Acknowledgements

1 wish to record my sincere thanks to Dr.
S. K. Brahma for photomicrography and to the
authorities of G. C. Bose Biological Research
Unit for extending facilities for this research.

Summary

1. The common Indian toad, Bufo melano-
stictus Schneider, was treated with low tempera-
ture (±10°-15°C) during summer (April-May)
and autumn (August-September) months of the
year. The treatment was continued for one
month.

2. No significant effect on spermatogenesis
was observed except a sudden fall in the number
of secondary spermatogonial cell nests of treated
toads during the summer season.

3. The autumn toads after treatment showed
a decrease in the diameter of the tubules and
an increase in the percentage of interstitium in
the testis tissue components.

4. Possible significance of the changes due
to this treatment and some probable explana-
tions have also been discussed.

References

Basu, S. L. & A. Mondal

1961. The normal spermatogenetic cycle of the
common Indian frog, Rana tigrina Daud.
Folia Biologica, 9 (2): 135.

24

Zoologica: New York Zoological Society

[47: 4: 1962]

Bullough, W. S.

1951. Vertebrate sexual cycle. Methuen & Co.,
Ltd., London.

Caruso, M. A.

1949. Sobre al ciclo sexual anual de algunos
Hylidae del Norte Argentino (Phyllo-
medusa sauvagii e Hyla raddiana). Acta
Zool. Lilloana, 8: 83.

Cei, G.

1942a. L Influenza dei fattori ambientali sulla
spermatogenesi del Geotriton fuscus
Bonap. (Spelerpes fuscus). Arch. Zool.
Ital., 30: 311.

1942b. Richerche biologiche e sperimentali sul
ciclo sessuale annuo dei Triton alpestris
( Triturus alpestris” Laur.) del Trentino e
dell’alto Adigo. Studi Trentini Sci. Nat.,
23: 1.

1944. Analisi biogeografica e richerche biolo-
giche e sperimentali sul ciclo sessuale
annuo della Rana rosse d’ Europa. Monit.
Zool. Ital., 54, suppl. 1.

Cei, J. M.

1948. El ritmo estacional en los fenomenos cic-
licos endocrino-sexuales de la Rana cri-
olla (Leptodactylus ocellatus L.) del
Norte Argentino. Acta Zool. Lilloana, 6:
283.

1949. Sobre la biologia sexual de un batracio de
grande altura de la region andina (“Tel-
matobius schreiteri Vellard). Acta Zool.
Lilloana, 7: 467.

Church, G.

1960. Annual and lunar periodicity in the sexual
cycle of the Javanese toad, Bufo melan-
ost ictus Schneider. Zoologica, 45 (13)-
181.

Galgano, M.

1934. L’influenza della temperatura sulla sper-
matogenesi della Rana esculenta L. Monit.
Zool. Ital., 45 suppl., 82.

1935. Intorno all’influenza del clima sullea
spermatogenesi di “Rana esculenta L.”
Arch. Ital. Anat. Embryol., 35: 511.

EXPLANATION

Plate I

(Photomicrographs of 6p, sections through the testis
of Bufo melanostictus )

A. Section showing the hyperplasia of the inter-
stitium ( vide arrow) and narrow tubule diam-
eter of treated toads during August-September
(Group A). X 100.

B. Section showing the normal condition of the
testis in the control toads of August-September

Galgano, M. & L. Falchelli

1940. L’influenza della temperatura sulla sper-
matogenesi e sopra i caratteri sessuali dei
Triton cristatus Laur. Monit. Zool. Ital.,
51: 166.

Galgano, M., & B. Lanza

1951. Contributi intorno all’azione della tem-
peratura e dell ’ormone folliculostimolante
sulla stasi spermatogenetica in Rana tem-
poraria L. Rend. Acad. Naz. Lincei., (8) :
11, 105.

Ifft, J. D.

1942. The effect of environmental factors on the
sperm cycle of Triturus viridescens. Biol.
Bull. Woods Hole, 83: 111.

Mazzi, V. & M. Galgano

1949. Ultariori observazioni intorno all’effetto
della serra calda sul ciclo sessuale del
Triton cristato. Rend, Acad. Naz. Lincei,
(8): 6, 518.

Mondal, A., & S. L. Basu

1960. Spermatogenetic cycle in Bufo melano-
stictus Schneid. Ind. Jour. Physiol, and
Allied Sci., 14: 2, 43.

Rengel, D.

1950. Accion de la temperatura elevada sobre
la espermatogenesis de des formas de
“Leptodactylus ocellatus L.” Acta Zool.
Lilloana, 9: 425.

van Oordt, P. G. W. J.

1956a. Regulation of the spermatogenetic cycle in
the common frog (Rana temporaria).
Thesis, Utrecht University.

1956b. The role of temperature in regulating the
spermatogenetic cycle in the common frog
(Rana temporaria). Acta Endocrinol., 23:
251.

WlTSCHI, E.

1924. Die entwicklung der keimzellen der Rana
temporaria L. 1. Urkeimzellen und sper-
matogenese. Z. Zell. U. Gewemelchre,
1, 523.

OF THE PLATE

(Group A). Note the undifferentiated inter-
stitium and wide tubule diameter. X 100.

C. Testis sections of treated toads during April-
May (Group B), showing the sudden fall of
secondary spermatogonia (Stage II) in the folli-
cles. X 100.

D. Section through the testis of control toads of
Group B showing very little difference except
the normal number of secondary spermatogonia
(arrow indicates the stage II cell-nests). X 100.

BASU

PLATE I

LOW TEMPERATURE EFFECT ON THE TESTICULAR CELL COMPONENTS
OF THE COMMON INDIAN TOAD. BUFO MELANOSTICTUS SCHNEIDER

5

Biology and Behavior of Damon variegatus Perty of South Africa
and Admetus barbadensis Pocock of Trinidad, W.I.

( Arachnida, Pedipalpi) 1

Anne J. Alexander2

Department of Zoology, Rhodes University, Grahamstown, South Africa
(Text-figures 1-5)

Contents

Introduction 25

Habitat 25

Feeding 26

Washing 27

Intraspecific Behavior 27

Courtship 29

Development 32

Phylogenetic Discussion 34

Summary 36

References 36

Introduction

TAILLESS whip-scorpions or scorpion-
spiders are dorsally flattened, cryptically
colored arachnids, found in tropical and
sub-tropical regions. Their systematic position
is controversial so that a study of their biology
and behavior is not only interesting in its own
right, but might also be relevant to this question.
The scorpion-spiders have traditionally been
placed in the arachnid order Pedipalpi, but in
1949 Millot replaced this by two new orders, the
Uropygi and the Amblypygi, the latter contain-
ing the scorpion-spiders. Subsequently Petrunke-
vitch ( 1955) has divided the Pedipalpi into three
separate orders, the Thelyphonida (true whip-
scorpions), the Schizomida and the Phrynichida

Contribution No. 1,015, Department of Tropical Re-
search, New York Zoological Society.

2 Some of the observations recorded here were made
in Trinidad on a visit to the field station of the Depart-
ment of Tropical Research and thanks are due both the
staff of the station and the Society for making this visit
possible. Financial assistance is also gratefully acknowl-
edged from the National Science Foundation, the South
African Council for Scientific and Industrial Research
and the Royal Commission for the Exhibition of 1851.

(scorpion-spiders). He agrees with Millot, how-
ever, that the first two are more closely related
to each other than either are to the scorpion-
spiders. Both of these proposals imply that the
previous order Pedipalpi reflected similarities
which had in fact arisen in two separate lines
of evolution. All these conclusions have been
reached from a consideration of the morphology
of the animals, the more recent suggestions tak-
ing into account the internal as well as external
features. It seemed possible that behavioral char-
acteristics might supply evidence for or against
the suggested convergence of the Schizomida and
Thelyphonida with the Phrynichida. Few records
have been made of the behavior of the Pedipalpi
as a whole and this dearth of information is espe-
cially noticeable in the case of the Phrynichida.
The present observations on Admetus barbaden-
sis and Damon variegatus, both members of the
family Tarantulidae, are offered in view of this
lack. It is hoped that if they serve no other pur-
pose, they may provide some incentive to record
the equivalent information for the other two
groups; for until comparative studies have been
made on the animals as a whole, much of what
is said here must remain a speculative contribu-
tion to the subject of pedipalp relationships.

Habitat

In the field these scorpion-spiders can be found
beneath loose pieces of wood or leaves, more
especially under forest cover. There is no indi-
cation that they ever construct burrows for them-
selves either in the wild or in the laboratory, as
has also been reported by Gravely (1915) for
the East Indian scorpion-spiders. If the cover is
suddenly removed from a specimen, it may stay
absolutely still, presumably employing behavior

25

26

Zoological New York Zoological Society

[47: 5

which would sometimes lead to its being over-
looked because of its general flatness and cryptic
coloration. If it moves, it does so quickly, run-
ning suddenly sideways onto another surface of
the covering object or away to another hiding
place. It never threatens unless it has actually
been picked up and then only occasionally.

D. variegatus is more markedly synanthropic
than A. barbadensis and in many parts of Natal
can be collected very readily from cellars, out-
houses or man-holes where it lives either in crev-
ices or freely on the walls if the place is fairly
dark. This acceptance of human habitation is a
biological feature which contrasts with the be-
havior of the schizomids and thelyphonids, nei-
ther of which has been reported to occur asso-
ciated with man.

In laboratory conditions both A. barbadensis
and D. variegatus show a diurnal rhythm of ac-
tivity in which the active phase is nocturnal and
since the former species has been found wander-
ing about at night in the forests of Trinidad, it
is assumed that the activity pattern shown in the
laboratory is a natural one and not induced by
disturbances during the day. Similar habits occur
in the other two groups but it is common for
most arachnids to be nocturnal.

Feeding

There is no information about what the scor-
pion-spiders eat in the field. Prey is, however,
caught at night and is almost certainly living.
Animals such as moths, crickets, spiders, cock-
roaches and beetles are accepted in the labora-
tory and probably form part of the normal diet.
In stalking prey, the scorpion-spider approaches
directly, i.e., not sideways as in escape. The tips
of the first pair of legs3 tap the prey so gently
that they seldom disturb it. When the scorpion-
spider is an inch or two from its prey, it suddenly
throws itself upon the insect, clutching at it with
the exposed spines of both pedipalps. Sometimes,
when the prey is especially large, the scorpion-
spider attacks several times in this manner, re-
treating between each attack. More usually the
insect is impaled on the pedipalpal spines at the
first onslaught. Once the prey is caught, the pedi-
palps fold, pulling it towards the mouth and re-
taining their hold of it while the needle-like cheli-
cerae alternately dig down into it.

After the meal the corpse is often left as a
mangled mass of exoskeleton but this is not in-
variably so; sometimes it remains almost entire,

8 The first legs are very long and antenniform and will
be referred to throughout this study as the “feelers;”
see Patten (1917).

only showing external damage at the points
where the chelicerae had punctured it. This is
strongly reminiscent of what occurs among the
spiders and suggests that scorpion-spiders may
also rely on extra-oral digestion to a considera-
ble extent.

Such a consideration leads directly to the prob-
lem of how the food is conveyed into the gut,
once it has been liquified by the digestive juices.
Just behind the mouth opening there is a typi-
cally arachnid sucking pharynx and ingestion it-
self consists of drawing up the fluid contents
of the corpse into this sac.

Both scorpions and scorpion-spiders possess
“pseudotracheal” areas on certain limbs. In the
scorpions such filter-like areas are located on the
coxal endites of the second legs. During feeding
either digestive juices pass through these and
onto the food or the liquified food passes through
them into the gut: there is no decisive evidence
as yet to distinguish between these possibilities.
Limb movements, which occur during feeding,
could help move the fluid in either direction. In
the scorpion-spiders there is no limb movement,
yet on each coxa of the pedipalps they have a
pseudotracheal area structurally resembling the
pseudotracheae of a scorpion. The main channel,
like that in the scorpions, opens into the base
of the esophagus.

In captivity scorpion-spiders frequently drink
water. In this behavior, as in the eating pattern,
there are alternate movements of the chelicerae.
These cease for short intervals as fluid is drawn
up into the gut. The significance of the cheliceral
movements is obscure; perhaps they are merely
a reflection of the fact that the animal uses the
motor pattern of normal feeding when drinking
free water. Conversely, the similarity of the eat-
ing and drinking patterns is in keeping with the
suggestion that it is a fluid which is taken up
into the mouth during feeding and that extra-
oral digestion must occur.

As with the detection of prey, the detection of
water appears to be done by sense organs on
the feelers. This may be demonstrated as fol-
lows : A desiccated animal is put on the bottom
of a dish which has a number of small holes
bored in its lid and many drops of water placed
between these holes. During its investigation of
the dish, the scorpion-spider will eventually put
one of its feelers up through a hole and into a
drop of water. Immediately, the behavior pat-
tern of drinking can be seen in the animal below
and it will make efforts to get at the water. The
sense organs involved have not yet been identi-
fied, nor is it clear whether the feelers only can
be used for detecting water.

1962]

Alexander: Biology and Behavior of Damon variegatus and Admetus barbadensis

27

Washing

This behavior may be seen most frequently
after the animals have been drinking or eating,
though they will interrupt other behavior, such
as courting or fighting, to wash themselves. The
feelers and walking legs are pulled to the mouth
by the pedipalps and then drawn between the
chelicerae, being cleaned by the medial brushes
which occur on these appendages (Text-fig. la).
During washing, the chelicerae move up and
down as they do in eating or drinking.

Text-fig. 1. a— Lateral view of the left chelicera of
A. barbadensis, showing the mesial brushes on both
first and second segments, b— Lateral view of the
last segment of the pedipalp of A. barbadensis,
showing the structure used by the animal for wash-
ing its limbs.

The catching of a leg or feeler by the pedipalp
may occur as a preliminary to its being washed
between the chelicerae, as described above, but
the action may also be repeated over and over
again without the chelicerae being involved at
all; the limb is pulled a short distance towards
the mouth by the pedipalp and then released.
Here the pedipalp itself is doing the cleaning:
the apparatus concerned is a little close-haired
brush on the last segment (Text-fig. lb)4. The
limb is drawn through this brush as it returns to
its position after being pulled towards the mouth.
The important function of the cleaning brush,
however, is the washing of the pedipalps them-
selves. Though the brushes of the chelicerae are
able to clean some of the spines on the inner
surface of the pedipalp, this cleaning action is
incomplete and many of the pedipalpal spines

4 It is this structure that Barnard, Ann. Mag. nat. hist.
(6), 11, 28-30 (1893), suggested was a sense organ, and
the homologue of the adhesive organ on the last segment
of the pedipalp of the solifugid.

are not reached at all. If the pedipalpal spines
are to act as weapons, their sharpness is all-im-
portant. This must depend on their being kept
free of congealed remnants of prey and dirt and
it is desirable that there should be a mechanism
for cleaning them properly. It is the pedipalpal
brushes which do the major part of this cleaning,
wiping each of the large spines in turn, then
cleaning the outer surface of the pedipalp itself.
The brushes are themselves cleaned on the cheli-
ceral hairs, these being moistened at the mouth.
It seems possible that the brush on the pedipalp
of D. variegatus and A. barbadensis has been
evolved primarily for the cleaning of these pedi-
palpal spines. Millot (1949) shows the presence
of the brush in his illustration of Charinus milloti
Fage, a member of the Charontidae, the second
family of Phrynichida. So pedipalpal brushes
may occur throughout the whole order; possibly
they were one of the prerequisites for the capital-
ization of the pedipalpal spines as weapons.

Intraspecific Behavior

Aggressive and threatening behavior occurs in
encounters between two males or two females,
between an adult and a young animal or even
between two young animals, and the same be-
havior comprises the first part of courtship in
any pair of scorpion-spiders. Hence a descrip-
tion of actual courtship behavior will be left un-
til after consideration of intraspecific behavior
which does not lead up to mating.

In an encounter between individuals that are
clearly unevenly matched, (i.e., a small and a
large individual or one injured and one intact),
regardless of which touches the other first, the
result is almost always the same. The “inferior”
animal runs off sideways for a short distance and
then extends the nearer feeler towards the supe-
rior animal, quivering it violently in the air
over the body of the other. Then, without neces-
sarily any further move from its antagonist, the
weaker will suddenly fold in its feelers and run
off as far as possible from the other. Sometimes
the superior animal may unfold its pedipalps
and threaten or even rush and fling itself at the
inferior. Invariably, however, the latter animal
escapes uninjured from the first encounter. In
laboratory observations it may subsequently be
killed, but such an event should perhaps be re-
garded as something that would not normally
occur in the field; it seems to happen only when
the weaker animal is allowed no space for escape.

When the contestants are more evenly
matched, the first encounter may, as in the pre-
vious case, include violent pedipalpal attack.
Here the two animals will strike at each other

28

Zoologica: New York Zoological Society

[47: 5

with one or both pedipalps. The spines on these
appendages may tear the carapace or abdominal
tergites or rip open the swollen intersegmental
membranes of the pedipalps where they are ex-
posed in threatening. Such wounds are often
fatal and, if the stricken animal is not eaten im-
mediately by its opponent, it will usually be dead
by the following day.

In addition to these direct attacks, however,
there are numerous conflicts that end without any
blood being shed. Such encounters may be de-
scribed as threat fights and they precede almost
all true courtship. There are several variations,
but in general the victor must beat its opponent
into submission using only its delicate feelers.
Although the blows may be given with such
force that the body of the opponent sways be-
neath them, it is obvious that little discomfort,
let alone hurt, could come from them. This
would appear to be a case of highly ritualized
threatening which does not involve the weapons
of offence themselves.

The commonest form of threatening is what
is here called “side-tapping.” The animal con-

cerned faces about 45° away from its opponent,
(Text-fig. 2a), bends its abdomen so that it is
more nearly in line with the opponent, extends
and opens the pedipalp furthest from the enemy
and taps or beats the latter’s body with the nearer
feeler, which is stretched out in front. The body
of the animal doing the tapping is nearly always
held very close to the ground, as if it needs
support.

The opponent normally does one of two
things: either it returns the tapping, using pre-
cisely the same pose as does the first animal so
that a very symmetrical effect is achieved (Text
fig. 2b) or it stands up, holds its body well away
from the ground and allows the first to beat at
it (Text-fig. 2c). If the latter response is made,
the animal which is being beaten usually extends
its pedipalps to some extent, drawing the tips
fairly close together and keeping the claws
(formed by the spines) closed, i.e., not using the
normal threat pose in which the pedipalps are
both extended and opened.

Whether there is mutual side - tapping or
whether only one animal taps, the behavior lasts

Text-fig. 2. Diagrammatic representation of an encounter between two specimens of D. variegatus. Animal
I is above, with its feelers represented in black, Animal II is below and is the smaller. In all cases the full
length of the feeler of Animal I is not shown in proportion to its body, a— Animal I is side-tapping, leading
with the left feeler while the right is folded back, b— Animal I is side-tapping slightly more actively, as can
be seen from the fact that its body is more fully oriented towards II and it has opened and extended its right
pedipalp further. Animal II is now returning the side- tapping so that there is a mutual exchange of blows,
c— Animal I has gone into what is virtually the extreme side-tapping pose while II has taken up the “pas-
sive” stance, with legs raising the body high above the substratum while the pedipalps are half extended
and the claws almost completely closed. The feelers curve forward gently and touch at II.

1962]

Alexander: Biology and Behavior of Damon variegatus and Admetus barbadensis

29

only a short time and then there is a reorganiza-
tion. In the case of mutual side-tapping, the pair
pull in their feelers and face each other. Then
they gradually revert to side-tapping once more
but this time leading with the other feeler. Be-
fore the new orientation and as the animals face
each other, the feelers may exchange a few beats.
Occasionally, after coming forward to change,
the pair revert to leading with the same feeler
as they were using before. Less frequently, one
animal will change its leading feeler while the
other will not, thus producing a thoroughly dis-
organized side-tapping.

If only one animal of the pair has been tap-
ping, and the other is passive, the roles are re-
versed periodically; thus the passive scorpion-
spider begins to tap while its partner rises up on
its legs, brings both pedipalps to the half-ex-
tended position and holds its feelers partly back,
i.e., it assumes the passive role.

Another behavior pattern can be distinguish-
ed; this seems to occur in animals which are
both more active than the couples described
above. Each stands with its pedipalps open and
partly, or even widely extended and then the
animals lunge in turn towards each other. Some-
times they spike the body or one of the limbs
but quite frequently they miss altogether. Some-
times both animals strike together and the spines
of the two pairs of pedipalps become entangled.
A slight variation of this pattern is one in which
both animals move sideways, facing each other
and occasionally lunging across the space that
separates them. In some cases the same dance-
like steps occur but the animals merely beat at
each other with their feelers instead of using
their more offensive pedipalps.

These various patterns may merge into one an-
other or change abruptly from one to the other;
there is as yet no clear explanation of the sig-
nificance of each, or of why or how the changes
are initiated. The sequence does not seem to
have any precise bearing on the final result of
the encounter. This latter takes one of three
forms. Firstly, there may be a serious pedipalpal
exchange between the pair and in this one or both
may be injured or killed. Secondly, one of the
pair may suddenly turn and run off rapidly (be-
havior that does not normally follow a series of
active tapping at the opponent). Lastly, one or
both of the animals may just wander off slowly
and pay no further attention to the other.

Courtship

True courtship is far more difficult to observe
than the encounters recorded so far. The chances
of seeing this behavior, however, are increased
if the observations are made at night and under

a very dim or red light. In neither D. variegatus
nor A. barbadensis can the sexes be distinguished
from a glance at the dorsal surface and so, for
experimental work, it is convenient to deter-
mine the sex of each animal and mark its back
with appropriate paint. Sexing can be done using
the differences in the details of the furrows on
the genital operculum, as was suggested for
D. variegatus by Lawrence (1949). With live
animals, however, it is better to hold the animal
on its back and lift the edge of its operculum
gently; if it is a male, the genital organs will
be extruded (Text-fig. 3a). These male organs
will be referred to here as “genital cones.” (The
term “penes” used by Lawrence (1949) is un-
acceptable for, as will be seen later, they do not
function as intromittent organs). They are rela-
tively larger in A. barbadensis than in D. var-
iegatus and consist of a pair of conical structures
(Text-fig. 3b) almost joined together at their
bases, each opening towards the midline by way
of many peculiar foldings. The male ducts lead
into the bases of these genital cones.

The under surface of the female operculum
bears no such organs, but in the female A. bar-
badensis there is a pair of small dark-colored
sclerites lying obliquely just under the opercu-
lum (Text-fig. 3c) . These point towards the mid-
line and are attached anteriorly. Their function
will be discussed later in relation to the egg-case.
The female D. variegatus lacks these sclerites
and has instead a number of complicated fold-
ings of the cuticle, some of which are sclerotized
(Text-fig. 3d). In either case the single opening
of the female duct lies at the base of the oper-
culum.

Returning to the actual courtship, it has al-
ready been said that this is normally preceded
by certain behavior patterns which also occur in
encounters between animals of the same sex or
immature stages. In the first part of an en-
counter the degree of violence may vary widely,
but eventually, if genuine courtship is to follow,
the animals will reach a stage when the female
response to the side-tapping of the male becomes
one of submission. Instead of holding her body
stiffly away from the ground, she allows it to
come to rest, spreading her legs out sideways.
Her pedipalps are folded from the semi-extended
position into one of rest and her feelers become
motionless. The male then straightens his body
and comes forward to face her; he beats her
intermittently with both feelers. Presently he
extends his pedipalps fairly slowly and lunges
at her, pushing her with his body as if testing
her passivity. During these advances, the female
may show signs of recurring aggression and
immediately the male will move back and again

30

Zoologica: New York Zoological Society

[47: 5

Text-fig. 3. a— Genital area of A. barbadensis showing the extruded genital cones, b— View of the genital
cones from the undersurface of the genital operculum The male ducts and accessory glands would empty
into the cones through that part which is shown by cross-hatching in the diagram. Part of the slit-like
opening of the cones can be seen, si. c— Undersurface of the genital operculum of a female A. barbaden-
sis to show the sclerites, sc, which hold the anterior end of the egg-case, d— Undersurface of the genital
operculum of a female D. variegatus showing the foldings and sclerotization which occur there in the
formation of a socket, so, into which the anterior end of the egg-case is moulded. The position of the first
pair of book-lungs can be seen, bl.

begin his intermittent beating. Such upsets oc-
cur more and more infrequently until, eventual-
ly, after some hours, she lies passively under all
his advances.

D. variegatus completed courtship in the lab-
oratory on one occasion only and then a part of
the behavior was not seen. Courtship climaxed
by mating has, however, been watched in A. bar-
badensis on five occasions. The patterns of
courtship and mating were virtually identical—
except for points of difference introduced by
deliberate interference. It seems likely that the
behavior in the two species is the same.

There is no mating grasp at all in these
scorpion-spiders and the animals remain un-

connected throughout as in some of the pseudo-
scorpions and mites.

Once the female is truly passive, mating can
begin. The male lowers his body and touches it to
the ground several times. He gets up, moving
closer to and almost touching the body of the
female who remains motionless. He turns round
so that he faces directly away from her and,
lowering his body, crouches down on the sub-
stratum as before. A slender, transparent sperm-
atophore is then extruded from his genital open-
ing. Once the distal end of the spermatophore
is cemented onto the ground, the male scorpion-
spider raises himself and turns round so that
he again faces the female. He then moves for-

1962]

Alexander: Biology and Behavior of Damon variegatus and Admetus barbadensis

31

ar

Text-fig. 4. Spermatophore of A. barbadensis. a— Lateral view; the right-hand side is that from which
the female approaches to take up the sperm, b— Dorsal, slightly lateral, view showing the face approached
by the female. The position of the two sperm masses is indicated by the cross-hatched areas just above
the capsule portion of the spermatophore, e, where they are normally held by the arms, or, of this region.
Stem of spermatophore, s; flat basal portion which is cemented down onto the substratum, w.

ward and crouches on top of the spermatophore,
moving his body slightly as if orienting him-
self correctly in respect to it. Once settled, the
male normally remains motionless for up to five
minutes. Apparently it is during this time that
he places two masses of sperm in position on
the proximal end of the spermatophore. Cer-
tainly up to this point the spermatophore has
been empty of sperm.

The male slowly rises from the spermatophore
and steps back a few centimeters. He quivers
and the female which has so far remained im-
mobile comes forward, her feelers guiding her
towards the spermatophore (Text-fig. 4). Fin-
ally she crouches over it, applying her genital
region to its proximal end. She jerks forward
against it, usually several times in quick suc-
cession though sometimes there is a thirty-second
interval between the first and second attempt.

She then stands up with the two masses of
sperm caught partly beneath her genital oper-

culum. The male becomes active once more and
taps the female rapidly for about fifteen seconds.
When she moves away from the empty sperma-
tophore, the male eats it. The two animals then
separate and there is no evidence of aggression.

Finally, in this section on the reproduction of
the Phrynichida, it may be mentioned that elec-
trical stimulation of the genital area of a living
male D. variegatus can cause the production and
extrusion of a spermatophore. When low in-
tensity shocks are first applied, the operculum
will be lifted slightly and the genital cones be ex-
truded. Presently, from between and posterior
to these, a larger organ begins to appear. This
is a many-lobed structure which will be referred
to here as the “genital body” (Text-fig. 5). Fur-
ther gentle stimulation causes the secretion of an
amount of transparent, very sticky material,
presumably that which normally cements the
spermatophore to the substratum. Then, also
from between the lobes of the genital body, the

32

Zoologica: New York Zoological Society

[47: 5

Text-fig. 5. The genital region of a male D. variegatus showing the “genital
body”, gfo, which has been extruded from the genital opening after electrical
stimulation. The lobes of the genital body are somewhat asymmetrical in this
example, so that the flap, f, of the right side is still tucked in on the left, uf.
The genital cones have been displaced laterally and anteriorly and are only
just recognizable, gc, in the drawing. The spermatophore is extruded at the
point marked, gd, in the mid-line and between the various lobes. The stria-
tions, bl, represent part of the first pair of book-lungs which are visible under
the genital operculum.

actual spermatophore itself is extruded. In such
artificial conditions the production is never suc-
cessful and the structure could clearly not be
used for transference of sperm by the animal
concerned. The use of such stimulatory tech-
niques was reported by Piza ( 1950) working on
a scorpion, Tityus bahiensis. Here a structure,
which subsequent work has identified as the
spermatophore, was partly extruded from the
animal. In the case of a South African scorpion
subjected to similar treatment, not only was the
skeleton of the spermatophore ejected, but also
a large part of the glandular apparatus which
secretes it. Thus it should be borne in mind that
the eversion of the “genital body” in D. var-
iegatus may well be an artifact due to the ab-
normal stimulation and that in natural produc-
tion of a spermatophore this does not occur.

Development

It is known that in the Phrynichida, Schizo-

mida and Thelyphonida, the eggs are attached
beneath the abdomen of the female after they
are laid, a phenomenon that occurs also in the
pseudoscorpions. The eggs of A. barbadensis
and D. variegatus are held together in a fairly
tough egg-case consisting partly at least of
chitin. According to Millot (1949) and Law-
rence (1953) the cavity within the egg-case is
continuous with that of the female genital organs
until the young are freed. In the case of A. barba-
densis this is not so; the anterior end of the egg-
case is a tough dried stalk which is held in
position just beneath the genital operculum by
the pair of sclerites mentioned earlier (p. 29
and Text-fig. 3c). This coupling of the anterior
end may occasionally come loose and the animal
is apparently incapable of hooking it up again.
When this happens, the case is merely left at-
tached by sticky threads to the abdomen. Occa-
sionally, however, the whole attachment fails
and the egg-case is dropped to the ground. In

1962]

Alexander: Biology and Behavior of Damon variegatus and Admetus barbadensis

33

one such instance, the female then ate the young
and their case.

As Lawrence and Millot have observed, the
abdomen of D. variegatus may be considerably
hollowed out so that the egg-case does not pro-
ject much below the normal level of the ab-
domen. The lateral margins tend, in fact, to
overlap the edges of the egg-case and thus insure
better protection. In A. barbadensis this hollow-
ing of the abdomen is far less marked and the
egg-case tends to hang down below the normal
level of the abdomen, especially just before
hatching.

The egg-case is secreted when the eggs are
laid. It appears as a semi-transparent, colorless
material which toughens and darkens during the
subsequent twenty-four hours. It consists of two
layers, an outer one which encloses the whole
and a thinner inner layer which is continuous
with packing material which lies between the
eggs and the stalk attaching the case to the
female.

It is not known how long the eggs remain
within the case but they can certainly be carried
for as long as a month before hatching. Emerg-
ence from the egg-case takes place during the
night or early morning as a rule. The young
come out through a ragged slit at the posterior
end of the case. About twenty-four hours before
the hatching this part of the egg-case can in fact
be seen to have become softened and partly freed
from the abdomen. Emergence may take as long
as the whole night, the young easing slowly out
of their own first exuvia as well as the partially
liquified case. As they become free, they climb
back onto the female. The remnants of the egg-
case, the egg-shells and exuvia remain attached
to the female until the young desert her several
days later.

The young scorpion-spiders, A. barbadensis
numbering from 15-30, D. variegatus up to 50,
cling onto the abdomen of the female, covering
both the ventral and dorsal surfaces, a phenom-
enon which occurs also among thelyphonids,
as shown by the illustrations given by Strubell
(1926) andYoshikura (1958). They have never
been seen on the prosoma and if one is dropped
there experimentally, it will immediately climb
back to the abdomen. It is not clear why this is
so, either in terms of what controls the behavior
of the young or what selective advantage is
given by such a distribution of the young on the
parent. Among scorpions, the female does not
appear to be incommoded by the young which
cling onto her prosoma any more than by those
on the abdomen.

Newly hatched A. barbadensis are very soft

and dusky pink; in D. variegatus the abdomen
is light green. The abdomen is relatively much
longer at this stage than it is in the adult and
eleven segments are far more easily recognized.
The feelers are folded up and, like the pedi-
palps, they are not used during this instar.
The legs, however, attach the young very
effectively to the female and one of the young
animals can be disengaged only with great diffi-
culty. If indeed a few are detached and strewn
around the female, they make but feeble attempts
to climb up again and only occasionally succeed.
The female does not help in this at all, and in fact
she merely shakes them off her legs if she can.
She may touch one tentatively with her feeler
for several seconds, then suddenly lunge for-
ward, catch it on the spines of the pedipalps
and immediately begin to eat it. Observations
in which one or more of the young were eaten
were made on four females and there is no
evidence that they would not have eaten the
entire broods if given the chance. In two cases
the young had been hatched outside the labor-
atory so that there is little possibility that the
behavior was abnormal. This leads to the con-
clusion that there is no maternal behavior
towards young scorpion-spiders that leave their
perches on the back of their mother.

This is in marked contrast to the complex
maternal patterns shown by many other ara-
chnids, such as some of the spiders and scor-
pions. In terms of selective advantage, the ex-
planation for this does not seem to lie in the
young scorpion-spiders being more firmly fixed
to the parent than are young scorpions; both
appear equally well attached. It seems possible,
however, that the two different types of response
can be correlated with the reactions of the
mothers to a threatening danger or a disturbance.
In such circumstances, a scorpion will stand and
threaten with claws or sting, otherwise it re-
mains immobile. In only a few cases will it run
away, and if flight occurs, it is usually brief.6

A scorpion-spider, on the other hand, stands
and tries to defend itself only if it has already
been partly damaged. Very occasionally it will
remain still and perhaps be overlooked; almost
always, however, it starts to run immediately.
The flight is broken up into short dashes side-
ways but as a rule the animal finally comes to
rest several yards from the site of the original

5 Possible exceptions to this generalization are such
flattened, rock-dwelling scorpions as the South African
genus Hadogenes, which may display an escape res-
ponse. This, however, is not shown by females carrying
their young, so that at such a time their behavior is
similar to that of the other scorpions mentioned above
and not to be compared with that of the phrynichids.

34

Zoologica: New York Zoological Society

[47: 5

disturbance. The point is that, if young scorpion-
spiders are knocked off their mother, they usu-
ally have no chance to climb back because their
mother will normally be some distance away by
the time they are ready to do so. It is of no
advantage for the female to have developed
behavior which would allow her to distinguish
her own young from any other helpless, wiggling
arthropod which might be food on the ground.
In the case of the scorpion, mother and young
are still together after an “attack” has passed
and it would be advantageous for a female to
possess a behavior pattern that actually helps
the young to remount, and even more so for
her to recognize that they should not be eaten.
A further consideration which may prove im-
portant in such an explanation of why the
phrynichid female has evolved no recognition
of her young is that these animals, unlike
scorpions, possess no “homes” in which they live
and to which they return. It would thus be
of interest to know of the Thelyphonida and the
Schizomida— more especially as females from
both these orders are reported to live in burrows,
at least when they have eggs (see Gravely, 1915;
Millot, 1949; and Yoshikura, 1958).

On their mother’s back, the young phrynichids
make almost no movement during the second in-
termolt period. This lasts four to six days in
A. barbadensis, up to 12 days in D. variegatus.
The second molt appears always to begin during
the morning. The length of time taken for any
one individual molt varies considerably, from
six minutes to almost three hours. There seems
to be a preferred site for this event— the posterior
part of the abdomen of the female— and most of
the young do not begin to shed their skins until
they can move towards this position. The molt
begins with a number of cheliceral movements.
Then a blister-like swelling arises on the cara-
pace. This pulsates slightly and a split gradually
appears round its anterior margin. The animal
bulges out of this slit. Of the limbs the first to be
freed are the chelicerae, then the pedipalps and
lastly the legs. As the old cuticle is sloughed
back, fluid within the body of the young can be
seen to be pulsating rhythmically. When the new
cuticle is uncovered, it is almost colorless but
rapidly becomes greenish-blue and darkens over
the next couple of days to a brilliant metallic
green. The old skins are not eaten by mother or
young and they fall from the mother’s back as
pieces of pink fluff. Like the first exuviae of
scorpions and unlike their own late stages, the
skins that are left after this molt are very thin,
soft and flexible. This is presumably a reflection
of the fact that the animals are markedly un-
sclerotized during their early instars.

Unlike many scorpions which remain for
some time on the female after they molt, the
young scorpion-spiders climb off within a few
minutes of freeing themselves of their cast skins.
This climbing down is no accident and is quite
“deliberately” done, for a young third instar
nymph will immediately run off its mother if
replaced.

There is a marked change in the behavior of
the phrynichids after their second molt. Once
they reach the ground, they run around quickly,
feelers are unfurled and they tap tentatively at
objects near them. There is a noticeable avoid-
ance of light. None of this behavior was of
course present when the young were on their
mother. Within two days they are capable of
catching and eating termites provided for them,
and are clearly independent of the female and,
indeed, can be found at this stage running about
by themselves in the field.

Phylogenetic Discussion

It has already been said that the study of the
behavior of the Pedipalpi might throw light
upon the phylogenetic relations of the different
groups which have been lumped in this order.
Many of the observations here recorded merely
serve to emphasise the need for comparable
studies on the other two members of the Ped-
ipalpi, the Thelyphonida and the Schizomida.
Of the biology that has been described, how-
ever, three aspects appear as though they might
provide material of significance in relation to the
phylogeny of these groups, namely, feeding,
sexual behavior and the general reactions of
defence and offence. The importance of feeding
has already been implied in the stress which
Petrunkevitch lays on the feeding organs in his
classificatory system. It is, however, abundantly
clear that more work is still needed on the
mechanism of feeding in the Phrynichida, while
even less is known of the details in thelyphonids
and schizomids (Snodgrass, 1948). It is per-
haps significant that although the filter-plate
apparatus in the Phrynichida is superficially like
that of scorpions, there are no associated limb
movements in the former.

Courtship likewise can be but sketchily com-
pared in the three groupings because actual mat-
ing has not yet been seen in the thelyphonids.
From descriptions of the preliminary behavior
(Fischer, 1911; Gravely, 1915) it seems that
the male in this order grasps the female first
with his pedipalps and subsequently with his
chelicerae, during which the two animals face
each other. Schizomids also promenade before
mating takes place (Sturm, 1958) and a cheli-

1962]

Alexander: Biology and Behavior of Damon variegatus and Admetus barbadensis

35

ceral grasp occurs, but here, however, the grasp
is quite different from that reported for thely-
phonids. It is the female which holds the male
and the part which is grasped is the specialized
portion of his tail; consequently both animals
face in the same direction. Hence it is possibly
of little significance that the Phrynichida also
differ widely, in having no mating grasp at all.
Indeed such variations in detail of courtship
may occur even among arachnids which are un-
doubtedly closely related; for instance among
the pseudo scorpions a mating grasp may or
may not be present.

Insemination is indirect by way of a sperma-
tophore stuck down onto the substratum. Con-
sidered alone, however, the fact is of little sig-
nificance as this would appear to be the primitive
method among the Arachnida (Alexander &
Ewer, 1957) if not among terrestrial arthropods
generally (Angermann & Schaller, 1956; Gila-
rov, 1958). If the arachnids were originally
aquatic forms in which fertilization was effected
by “casual” meeting of sperm and eggs after
both had been liberated into the sea, then the
first step in the evolution of the spermatophore
is perhaps exemplified by the behavior of water
mites and some pseudoscorpions in which the
male deposits spermatophores apparently hap-
hazardly around him, with or without the pres-
ence of a female. With a drier and wider habitat
there would be increased danger of desiccation
of the sperm and perhaps decreasing chances
of a female coming across such casually depos-
ited spermatophores. Thus there would be selec-
tion for an association between male and female
to be established before deposition of a sperma-
tophore occurs and it is to be expected that
such associations may have been independently
evolved several times. Nevertheless the phryn-
ichids and schizomids might be regarded as re-
lated insofar as the male has his posterior end
towards the anterior end of the female while
he deposits a spermatophore. Such behavior is
in contrast to that of the pseudoscorpions, the
mites and especially the scorpions where the
spermatophore would not function if the female
were forced to approach it from the opposite
direction. However, the very specialized mating
grasp of the schizomids makes it improbable
that the position taken up during spermatophore
extrusion bears any direct relationship to that
of the phrynichids.

The marked dissociation of the acts of sperm-
atophore deposition and sperm extrusion is
unique and it is hard to see any advantage to the
species from such behavior. It is, however, clear
that the two events are completely independent.
Thus on two occasions a male Admetus re-

turned, in one case three times, to a spermato-
phore when he apparently been unable to load it
successfully at the first attempt. Another time a
female Admetus walked away while the male
was depositing sperm; the male chased after her,
courted her once more to passivity and then
returned to his depositing of sperm. This dis-
sociation possibly offers the advantage that
should anything cause a faulty deposition of a
spermatophore or disrupt the mating, sperm will
not have been wasted. A somewhat similar situ-
ation occurs in Tityus trivittatus in which
Biicherl (1956) has described a “reloading” of
the spermatophore after the female had removed
the first mass of sperm.

Petrunkevitch has postulated a close associa-
tion between the Phrynichida and the Araneae,
and the present observations can be read in
support of such an opinion. It has been suggested
(Alexander, 1962) that the sexual behavior of
spiders may be derived from that of forms in
which there was a dissociation of the act of
loading a spermatophore from its deposition.
This is the condition in the phrynichids. Further,
it is easier to envisage the course of the evolu-
tion of spider mating were it derived from a
pattern in which the members of a pair did not
grasp and this again is a phrynichid character.

Finally we will consider what Manton (1958)
would term “habits of life.” It is this last type
of behavior that is hardest to categorize dis-
tinctly, the facts are less definite, actions less
stereotyped and more difficult to describe. As
Manton says: “Habits of life which appear to
have been of evolutionary significance vary
greatly in their ease of recognition because they
may not be exercised all the time.” Nevertheless
it is from this level of activity that an under-
standing of the general biology and the direction
of morphological and behavioral evolution will
emerge.

In considering the life habits of the groups
of the Pedipalpi it is desirable to point out that
the natural micro-ecological distribution of the
animals is not markedly different. What is signifi-
cant, however, is that the Schizomida and Thely-
phonida remain within their burrows or in
spaces in the soil while the Phrynichida run
freely over it or hide in crevices— the animals
may be referred to as being “of fixed abode” and
“vagrant,” respectively. Defensive behavior can
be correlated with this difference. The “fixed
abode” groups stay to defend themselves with
cheliceral pedipalps or repugnatorial glands
when concealment is no longer possible. The
Thelyphonida are aggressive in captivity and
large specimens may even damage human hands
with their pedipalps. This is in complete contrast

36

Zoologica: New York Zoological Society

[47: 5

to the emphasis among the phrynichids on cam-
ouflage or flight; the scorpion-spiders either sit
still and are overlooked or get away quickly and
unexpectedly with sudden and disconcerting
sideways darts— behavior for which their flat-
tened form, short abdomen and the sideways
extension of their legs are well adapted.

The absence of attack as a defensive response
in phrynichids is also partly reflected in an
“inefficient” use of the pedipalps during prey-
catching, for the scorpion-spiders frequently
fail to hold prey at which they grasp. In natural
conditions there must often be unsuccessful
attempts or the prey must be weaker, smaller
and consequently hunted more frequently. In
either case accurate locating and careful stalk-
ing is important and for these two phrynichid
characteristics are essential, the habit of roam-
ing freely about and the possession of elongated
feelers. Together with the latter must be con-
sidered the toilet behavior patterns involved in
keeping them clean as well as the specialized
pedipalpal brushes.

Though it is freely admitted that this is but
the beginning of an understanding of phrynichid
life habit, the number of anatomical and beha-
vioral characteristics which can be correlated
with this habit suggests that it has been of prime
importance in controlling the direction of phry-
nichid evolution. When similar and more com-
plete analyses are available for the Thelypho-
nida and Schizomida, it may prove possible to
comprehend the true inter-relationships of these
groups and their relationships with the other
Arachnida.

Summary

1. The general biology and behavior of two
species of phrynichid, Admetus barbadensis Po-
cock and Damon variegatus Perty, have been
studied. A description is given of their habitat,
food and manner of feeding, their drinking and
their cleaning patterns.

2. Behavior which does not lead up to court-
ship but occurs between two members of the
same species is described; much of this behavior
consists of threatening contests in which the deli-
cate feelers are used as “weapons.”

3. The main course of courtship is described.
Insemination is achieved by means of a sperma-
tophore which the male deposits on the sub-
stratum and from which the female picks up the
sperm mass. This method of mating is compared
and contrasted with those occurring in the
schizomids, scorpions, pseudoscorpions and
some of the mites.

4. Events associated with the electrical stimu-
lation of the genital area of a male D. variegatus

are described and discussed in relation to the
production of a spermatophore.

5. The manner in which the eggs are cared
for after they have been laid and the hatching
of the young is described, as well as general ob-
servations on their behavior in contrast to that
of the adult animals. An hypothesis is put for-
ward attempting to explain, in terms of selective
advantages, the absence of any maternal reaction
from the female phrynichid to one of her young
struggling on the ground in front of her.

6. The behavior recorded here is discussed
in relation to its bearing on conclusions about
the inter-relationships of the Phrynichida, Schi-
zomida and Thelyphonida.

References

Alexander, A. J.

1962. Courtship and mating in amblypygids
(Pedipalpi, Arachnida). Proc. Zool. Soc.
Lond. (in press).

Alexander, A. J., & D. W. Ewer

1957. On the origin of mating behavior in spi-
ders. Amer. Nat., 91: 311-317.

ANGERMAN, H. & F. SCHALLER

1956. Spermatophorenbau und bilding bei Ar-
thropoden mit indirekter Spermatophoren-
iibertragung. Ber. 100-jaarf. Deutsh. en-
tomol. gesell. Berlin, 228-237.

Bucherl, W.

1956. Escorpioes e Escorpionismo no Brazil.
Mem. Inst. Butantan, 27: 121-155.

Fischer, G. E. C.

1911. The courtship of whip-scorpions. J. Bom-
bay Nat. Hist. Soc., 20: 888-889.

Gilarov, M. S.

1958. Evolution of the insemination character
in terrestrial arthropods. Zool. Zh., 27:
707-735. (In Russian, with English sum-
mary).

Gravely, F. M.

1915. Notes on the habits of Indian insects,
Myriapods and Arachnids. Rec. Ind. Mus.,
11: 483-539.

Lawrence, R. F.

1949. Notes on the whip-scorpions (Pedipalpi)
of South Africa. Trans. Roy. Soc. S. Afr.,
32: 1-11.

1953. The biology of the cryptic fauna of for-
ests. Cape Town: Balkema.

Manton, S. M.

1958. Habits of life and evolution of body de-
sign in arthropoda. Joum. Linn. Soc. Lond.
Zool., 44: 58-72.

1962]

Alexander: Biology and Behavior of Damon variegatus and Admetus barbadensis

37

Millot, J.

1949. Ordre des Uropyges et Ordre des Ambly-
pyges. In Traite de Zoologie. 6. Ed. P. P.
Grasse. Paris: Masson et Cie.

Patten, B. M.

1917. Reactions of the whiptail scorpion to light.
J. Exp. Zool., 23: 251-275.

Petrunkevitch, A.

1955. In Treatise on invertebrate Paleontology.
Part P. Lawrence: Univ. Kansas Press.

Piza, S. de T.

1950. Reproductive organs and reproduction in
Tityus bahiensis (Scorpiones, Buthidae).
Proc. 8th Int. Congr. Ent. Stockholm,
1948: 1026-1027.

Snodgrass, R. E.

1948. The feeding organs of Arachida, including
mites and ticks. Smithsonian Misc. Coll.,
110, No. 10, 1-93.

Strubell, A.

1926. Thelyphonus caudatus L- eine biologische
Skizze. Verh. Nat. Ver. Bonn, 82: 301-314.

Sturm, H.

1958. Indirekte Spermatophoreniibertragung bei
dem Geisselskorpion Trithyreus sturmi
Kraus (Schizomidae, Pedipalpi). Natur-
wissenschaften, 6: 142-143.

Yoshikura, M.

1958. Observations on the breeding habits of a
whip scorpion Typopeltis stimpsonii Wood.
Acta Arachnol. 16: 1-7.

6

Breeding Activities, Especially Nest Building, of the Yellowtail
( Ostinops decumanus ) in Trinidad, West Indies1, 2

William H. Drury, Jr.

Massachusetts Audubon Society

(Text-figures 1-4)

Most authors agree that two families
excel at building complex nests— the
Icteridae of the New World and the
Ploceidae (True Weavers) of the Old World.
This paper reports nest building and some asso-
ciated breeding activities of the Yellowtail,
Ostinops decumanus, of the Icteridae, as ob-
served during ten days in late January, 1960,
at the New York Zoological Society’s Depart-
ment of Tropical Research field station at Simla,
Arima Valley, Trinidad, W. I., and discusses
relevant literature.

Since F. M. Chapman (1928) made a three-
year study of the courtship and breeding activi-
ties of Wagler’s Oropendola, Zarliynchus wag-
leri, on Barro Colorado Island, Panama Canal
Zone, the most comprehensive reports on the
flamboyant displays and the complex hanging
nests of this group of the Icteridae are by Skutch
(1954) on the Montezuma Oropendola, Gym-
nostinops montezuma, and the Yellow-rumped
Cacique, Cacicus cela; by Tashian (1957) who
studied the Yellowtail at Simla; and by Schafer
(1957) who studied Ostinops decumanus and
Psarocolius angustifrons in detail. Although
these papers do not include critical analyses of
nest-building motions, they allow me to make
comparisons of the nest-building techniques.
(These papers are referred to below without
citation).

The four species of oropendola— angustifrons,
decumanus, montezuma and wagleri— appear to
be very closely related, and as a systematist

Contribution No. 1,016, Department of Tropical Re-
search, New York Zoological Society.

Contribution No. 36 from the Hatheway School of
Conservation Education, Massachusetts Audubon So-
ciety, South Lincoln, Massachusetts.

trained in a field other than ornithology, I would
not hesitate to include them in the same genus.
However, the currently accepted generic names
are used here. Table 1 lists outstanding features
of the behavior during courtship and nest build-
ing shared by members of this group. Nest
building suggests that they are close to the
caciques— just how close may be revealed in fur-
ther studies.

Habitat

The Yellowtail nests in large numbers in the
erythrina trees, Erythrina micropteryx, also
called “bois immortel,” introduced into the
Arima Valley to provide shade in the cocoa and
coffee plantations (Text-fig. 1). Although the
cocoa, Theobroma cacao, coffee, Coffea arabica,
and banana, Musa paradisica, trees planted
under the erythrina are also introduced, all are
readily accepted by the native birds. The under-
growth of most of Trinidad is cleared several
times a year with long knives, locally called cut-
lasses. The cutting suppresses the heavy second-
ary growth that would compete with the local
crops. Rising above the understory to an average
height of 50 to 150 feet are the erythrina, whose
orange-red blossoms cover their crowns early in
the year at the close of the rainy season. Schafer
describes identical habitat for Yellowtails in
Venezuela.

The ends of the erythrina branches are ideally
suited for the attachment of the Yellowtail’s
nest, because (Text-figs. 3, 4) of the whorls of
stiff, dead leaf- or flower-bases that extend two
to five inches along the branch at the bases of
the smaller branches. The bases of the nests we
saw were woven into these burr-like structures.
Furthermore, the trees are tall, with smooth
bark, and at the top they spread umbrella-like

39

40

Zoologica: New York Zoological Society

[47:6

Text-fig. 1. Part of Yellowtail nest colony in Erythrina tree. Nests 1 and 18 were studied during site estab-
lishment; nests 9, 10, 18 and 19 were studied during building of the sleeve, closing the entrance, and weav-
ing the bag; nests 13, 15 and 19 were studied during weaving the bag and closing the bottom.

crowns which do not interfinger with each other.
As Skutch points out, these characteristics pro-
vide sites that protect the nests from predators
(chiefly lizards and snakes). He describes the
destruction of a colony of Yellow-rumped
Caciques to which a snake found access by vines
growing up the trunk. Schafer comments on the
form of branch-tip preferred for nest establish-

ment and the umbrella shape of trees chosen by
decumanus, comparing them with the require-
ments of the forest-opening species, angusti-
frons, which places its nest on the wall of vege-
tation along roadsides.

Territorial Behavior of the Males
Erythrina trees are scattered 40 to 50 yards

1962]

Drury: Breeding Activities of Yellowtail (Ostinops decumanus)

41

ing intruder. C. Low level song posture given when alone, primarily territorial. D. High level song posture
given when near a female, primarily sexual. E. Anxious and aggressive female in squabble near her nest.
F. Soliciting female.

apart on the steep hillsides of the Arima Valley.
Each dominant male Yellowtail appeared to oc-
cupy a territory covering two or three adjacent
trees. The males in the colonies which Chapman
and Skutch studied did not seem to have a terri-
tory or a set relation to any particular group of
females. I agree with them that as soon as one
male made a supplanting attack on another,
the attacked bird left, suggesting dominance and
territory defense, but I saw no territorial fights,
nor did Schafer. Schafer’s detailed discussion

of territory in decumanus and angustifrons
shows site tenacity and hierarchical arrangement
among dominant and subordinate males around
a colony tree. There are differences of degree,
but in both species subordinate males intrude
almost undisturbed, especially at the period of
copulation.

When an “intruding” male came into a tree
(even the nesting tree itself), he often spent as
much as ten minutes there without being at-
tacked. If the resident male did pay attention to

42

Zoologica: New York Zoological Society

[47:6

the newcomer, however, he moved into the tree
or branch, flying directly at the intruder or
perching 20 yards away, then walking or hop-
ping toward him with neck swollen and head
point up slightly (Text-fig. 2B). Soon the in-
truder (Text-fig. 2A) left and shortly afterward
both males sang. If, however, an intruding male
bowed and sang, he was often rushed by the
territory owner. In one instance the intruder
left soon after the owner bowed and called twice.

The male has half a dozen song or patrol
perches at which he spends most of his time.
When active, he spends four to ten minutes at
each perch. When inactive, he may be difficult
to find— either he is away from his tree or sitting
quietly on one perch. Schafer has notes on the
schedule of activities of territorial males and fe-
males.

I heard some sing all day, and the most active
singing between 0600 and 0900, and between
1600 and 1930, local time, as did Schafer.
Tashian recorded the frequency of singing and
reported no peaks, but that may have been be-
cause his study was made earlier in the repro-
ductive season.

Song and Display.— The song is described in
detail by Tashian, who used moving picture
film to analyze and time the postures which
accompany it. Schafer separates territorial from
nuptial song. As a bird starts his nuptial song,
he drops his head steeply between his feet and
raises his tail over his back, stiffly fanning his
under tail coverts which, like his tail, are yellow.
The bird then usually gives two hollow, gurgling
calls and, standing in a deep bow with the
feathers of his neck stiffly raised, gives a rattling
trill ( eeeeoooo-eeeeoooo , or eedy-eedy-ooo) ,
which grades into a continued rustling made by
flapping his relaxed but raised wings (Text-fig.
2D) ; then he stands up. He may repeat this call
as often as every twenty seconds for two hours,
but usually the call is relatively infrequent, given
about once every three or four minutes. Schafer
reports every three minutes, or 10-20 times an
hour, averaging 100 times a day. I found this
call most frequently directed toward a single
female or group of females. Schafer found the
noisy flight between perches so regular a prelude
to this song as to be a part of it. He found that
this call was less frequently given than the terri-
torial song, and I agree.

The territorial call starts with (1) a rattling
gurgle, or (2) a hoarse tsteeee or tsreee-kleee,
ending in a series of “plop’Mike calls -ka-wow-
wow-wow. I did not see the posture which ac-
companied it and heard no wing-rustling with it.
Usually it is given by an isolated male and, I
believe, not necessarily in the presence of fe-

males. Schafer says that the cry is given with
raised and weakly beating wings, and with plu-
mage not fully displayed (Text-fig. 2C).

The singing bird, especially away from a nest-
ing tree, may alternate his song with feeding in
the blossoms of the erythrina tree, but usually
when he is singing he spends the time between
songs moving among his trees, standing, peering
around in the tree, preening wings, flanks and
neck, ruffling feathers, scratching, pecking at his
feet( between the toes) or wiping his bill. When
in a nesting tree, the male occasionally flies vig-
orously (his wings make a deep resonant sound)
to perch on the side of a nest that is completed
or is being built. He lands with head already
down and neck swollen, and gives his full song,
perches stiffly, with crest raised, neck swollen
and blue eye glaring, for ten or fifteen seconds,
and then flies, usually to perch on the branch
at the base of this same nest. I saw a male fly
at and replace a female who had just flown in
and landed on the nest, and several times a male
responded to the arrival of a female or a group
of females with a bow and a song. The male was
especially likely to respond with song to the
arrival of females if he had recently driven
away another male.

Most of the time, even when the male flies to
her nest, the female pays no overt heed to his
activity. Occasionally she is evidently interested
and watches him, usually with her head up and
feathers sleeked (head-up “threat” of icterids,
with some “fleeing tendency,” Text-fig. 2E). I
found no direct relation between song display
and copulation. Schafer does not suggest any
close relation and points out that during the
copulation period territoriality seems to be min-
imal.

During the day, males are usually isolated on
territories, although many (young and inferior
males, Schafer suggests) visit other males’ terri-
tories or nesting trees. In the evening males stay
isolated in the crowns of the erythrina trees until
after the evening roosting flights of females,
which take place between about 1800 and 1830.
Then males leave their territories and fly singly
to the communal roosts which, in the Arima Val-
ley, are in a large clump of bamboo at the bot-
tom of the valley below Simla. On their roost-
ing flights, females in groups of three to
thirty come from several miles, usually in short
flights between crowns of erythrinas. They stop
in the territories of actively singing males and
often engage in precopulatory actions. Males are
the last to settle in the bamboo roosts. They start
to sing on their trees again before the sun is up.

The females of any one nesting tree readily
visit territories of other males, and I saw copula-

1962]

Drury: Breeding Activities of Yellowtail (Ostinops decumanus)

43

tion less often in nesting trees than in other trees
where a male was on station but where there
were no nests.

Territorial Behavior of the Females

1 counted 1, 3, 5, 22 and 43 nests in five
different trees, each in a singing male’s territory
(Text-fig. 1). Females usually in groups of a
dozen, but at times several dozen, fed in the
cocoa and citrus trees and moved along the
ridges down into the groves to gather nesting
material and food. Actively building females
spent much of the day working at their nests,
while groups of females not especially attached
to the nesting tree visited it for several periods
of twenty minutes to half an hour.

As Skutch and Chapman report for their oro-
pendolas, a number of Yellowtail females oper-
ated as a group and placed their nests on the
same branch or on a series of branches close
together. These nests were consistently placed on
the leeward side of the tree and on the outer
and upper part of the crown, at least 40 feet
above the ground. Schafer agrees, and has many
details on choice of colony trees, location of
nests, calendar and schedule of events which led
up to colony establishment. He stresses the im-
portance of weather changes which start the
cycle, but does not mention subgroups of fe-
males within a colony.

When close together in what appears to be a
squabble over the nest site, the females stand
with heads raised to a 45° angle and with tail
partly raised (Text-fig. 2F). I saw no fighting
such as recorded by Skutch and Chapman, but
frequently heard a nasal hiss, garreeoo or aaah,
from birds in this circumstance. It resembled
part of the song of a Starling, Sturnus vulgaris.
The females crowded their nests together even
within the limits of the group and four of the
nests I watched being built were woven into a
neighbor’s or a previous year’s nest. Schafer
mentions hostility of females to males and their
fierce attacks on inept immature males that tres-
pass at their nests.

Progress of the Breeding Cycle

Chapman, Skutch, Beebe (in Tashian) and
Tashian agree that nest building starts at the
very end of December and in early January. This
coincides with the end of the rainy season. Chap-
man believed that for Wagler’s Oropendola the
start of the nesting season is more accurately
associated with date than with the last of the
rains, but Schafer shows in detail how the down-
pours that come with cold, northerly winds start
the cycle in decumanus and how humidity con-
trols the start in angustifrons. Chapman reported

the beginning of the breeding cycle as being sig-
nalled first by the arrival of individual males in
the colony tree. Both Chapman and Skutch say
that the real start is the arrival of groups of
females to inspect the branches of the colony
tree. Schafer shows clearly that in angustifrons
the males’ territorial activity is a critical stimulus
for the start of the nesting activity by females.
If he stops his display, they stop.

Copulation— The male Yellowtail sings in his
tree, occasionally flies noisily and perches with a
group of females who are preening or moving
among the branches. The females may ignore
him or may move away nervously, with their
heads slightly raised and feathers sleeked (Text-
fig. 2E). This activity continues through the
early stages of nest building. Schafer shows that
copulation receptivity appears in females of both
decumanus and angustifrons as they finish the
nest bag and line it.

When receptive, the female flies to a branch
near a male. Twice I saw a female fly to a male
from inside her nest, and on other occasions
females came “out of the blue” into an isolated
song tree. The female perches above or near the
male with feathers sleeked, head raised and tail
raised or horizontal, and may almost imper-
ceptibly flutter her wings. The male, with neck
swollen, hops and walks along the branches to
perch below and beside her. The female squats
and flutters her wings, with head and tail slightly
raised; the male mounts for about ten seconds
while both birds flutter their wings. The male
dismounts and, in my observations, flies rapidly
away at once. On landing, he wipes his bill and
continues his patrolling and singing. After
ruffling and shaking her feathers, the female
usually either started to preen or flew off.
Schafer’s notes are very brief but agree with
these. Tashian reports that the male sang and
displayed and pecked at the female’s cloaca be-
fore copulation, and that he displayed again
after copulation. I have notes on several se-
quences in which a male approached a female
as in copulation and then pecked violently at
her cloaca, but these preliminaries did not lead
to copulation and suggest rejection by the fe-
male.

Asynchronous Activities of Groups of Females.
—All authors agree that many females start nests
much later in the season than the main group.
Chapman felt that some may have been second
nests. Schafer emphasizes the place of imma-
turity. In one of Chapman’s groups of late-nest-
ing females, building was interrupted and then
the beginning of the rainy season caused them to
abandon the colony site. Skutch found a colony
in northern Honduras feeding nestlings in Sep-

44

Zoologica: New York Zoological Society

[47:6

tember. As with so many birds, the breeding
cycle of the oropendolas must be started by
a regular annual time-giver, and breeding con-
tinue until interrupted by some internal or en-
vironmental time-giver— in this case, the next
rainy season.

There is also variation in the timing of an
individual female’s breeding during the species
peak in late January. At the tree I watched,
there were about a dozen females prospecting,
four nests being started, six nests being built,
about thirty idle nests or with females incuba-
ting, and at least five nests with females feeding
young. Females regularly came and perched
near the idle nests and clucked, or they flew in
and perched head down in the entrance for
twenty seconds, turned and perched looking out
for thirty seconds to a minute, then flew off.

Nest Building

Building Material— (See Table 1 for details
of nest-building materials used by the five spe-
cies. Schafer describes them in detail for angus-
tifrons and decumanus ) .

Schafer says that decumanus collects mate-
rials among the trees and off the ground, at a
distance from the colony, while angustifrons
collects material low in undergrowth or on the
ground under the colony. Skutch describes the
collecting of fibers underneath a banana frond
by a female montezuma, and the actions are
similar to those used by ( 1 ) the Baya, Ploceus
philippinus, in India when nipping and tearing
strips from the base toward the tip of a rice leaf,
Oriza sativa, (Ali, 1931); and (2) by the Village
Weaver, Textor cucullatus, tearing leaves of
Elephant Grass, Pennisetum purpureum, (Col-
lias, 1959).

My observations showed that decumanus uses
chiefly long, fibrous strips and grass. The ma-
terials are green when brought to the nest, but
turn brown in a day or two. In many cases
the fruiting heads of the grass (resembling
Panicum ) were still visible. On other occasions
small vines with tiny green leaves (resembling
the Solanum family) were used. Schafer says
that the supports and weave are chiefly (80%)
liana tips, and that the packing between is en-
tirely Spanish moss, Tillandsia usneoides if avail-
able. He describes the material used by decum-
anus as slender and. fine, in contrast to the char-
acteristically coarse and turgid material available
in the forest interior and used by angustifrons.

Skutch describes montezuma folding over the
strips from the banana frond in order to carry
long pieces. Yellowtails brought their material
in unfolded or gathered into loops in the bill,

and often the material streamed far out behind
the bird as she flew (Text-fig. 1).

When the female Yellowtail starts to build,
she uses pieces that average 30 to 45 cm. long;
then, as she finishes the base and builds the sup-
ports at the entrance, she brings fewer pieces
per trip and they are 60 to 100 cm. long. She
brings many shorter pieces again as she weaves
the sides. When weaving the bottom, she brings
a few long pieces (100 to 150 cm.). These ap-
pear to be strips of banana leaf.

Skutch speaks of violent squabbles over nest-
building materials in montezuma and no squab-
bles in the Yellow-rumped Cacique. Schafer de-
scribes frequent violent fights in angustifrons
and few in decumanus. Chapman describes the
stealing of loose ends from slovenly nests or
from those of absent birds by wagleri, as does
Skutch in montezuma. These two species may
demolish a messy nest. Schafer comments that
stealing is usual in angustifrons, infrequent in
decumanus. I saw very little grabbing of material
from other nests. On one occasion a female re-
peatedly flew past a nest with a moplike mass
hanging below the finished entrance, and each
time she grabbed a hanging, loose end and tried
to fly off with it, usually unsuccessfully. On
another occasion, a female perched several min-
utes on the outside of a nearly completed nest,
pecked and seemed to try to pull out the tiny,
still-green fibers sticking out or looped through
the brown weave. Another time a female
perched on the outside of the nest and repeatedly
pushed her bill through the weave, opening it
and making a series of fairly large holes, but
not pulling anything out. In each case the visitor
left when the builder came back, but in no case
did the builder chase her. Schafer says that
young males make holes in the sides of nests
during the period when females are receptive.

Schafer discusses nest building in decumanus
and angustifrons, emphasizing differences in lo-
cation, placing of the nest, and the materials
used. His data are very largely on angustifrons,
and although he describes the phases of building
(anchoring the nest and weaving; building the
apron; building the ring and the future entrance;
building the bag; and bringing the nest lining),
he does not treat in detail the movements in-
volved.

Phase 1: Site Establishment.— When the fe-
male is establishing her nest site, she spends
from five to fifteen minutes perched at a fork
in a branch at the tips of the long branches of
the erythrina (Text-fig. 3A and B). She walks
and hops, alternating her feet, along the main
part— peering, looking under and over the forks

1962]

Drury: Breeding Activities of Yellow tail (Ostinops decumanus)

45

near the tip. She spends much time peering
out across the valley. Skutch and Chapman de-
scribe frequent squabbles in groups of females
inspecting branches. Schafer describes threat
postures, but no real fights. I found the females
inspecting alone, and saw no squabbles among
the few birds prospecting. There were several
squabbles among the females on lower branches
where nests were already established.

When the female first brought material (long
grass or vines), she perched and looked around,
moved, looked around, flew to another branch,
stepped on the grass and gathered another loop,
then flew to another part of the tree. At two
sites, where females had been prospecting on
previous days, I watched a female bring long
grass in her bill and, perching on the branch,
look, move, look, then push it into the dead leaf
petiole bases at the joint (Text-fig. 4D). These
females brought several loads before the grass
remained in place. The female may or may not
step on the grass with one foot (Text-fig. 3C) ;
she may grasp an end in her bill, gather another
loop, pick up all the material, and fly off again;
she may do no more, and just fly off; or she may
take bits of the grass and push them down into
the clump of dead leaf bases, then reach around
the branch and pull a strand up around the
branch (Text-figs. 3E & 4A); or she may pull
a loose end around a leaf base and poke it into
the mass of material with a short shake (four
or five times) of her bill— the tremble-shove
(Lorenz, 1955). When building the base of the
nest, the female pulls ends around and uses the
bill-shake more conspicuously than when weav-
ing the sides of the nest.

I found the females hesitant and irregular in
their activity when establishing the nest site, but
I think this may have been caused by my pres-
ence. Schafer suggests that activity during this
period is readily interrupted.

Phase 2: Establishment of the Nest Base.— The
female brings grass and may put her foot on
the newly-brought material (Text-fig. 3C), or
may merely push it into the tangle already in
the leaf base. Soon after some material hangs
down, she pushes the new material into the mop
(apron, tablier) just under the branch. When
the mop is short, she usually pulls only a little
and just “fiddles” with loose ends, then flies off.
As the mop lengthens, she tucks the new material
under the branch and, reaching over to the other
side, grasps a piece of grass in the very tip of
her bill (Text-fig. 3D), often with a little shake
as she takes it. She pulls this gently or firmly
toward her and tucks it into the grass on her side
of the branch, with two or three pokes and four
or five shakes of her bill. Or she takes a piece

of grass from the side nearest her, pulls it over
and tucks it into the grass around and behind
the branch, with the same bill-shakes. Us-
ually she worked two, three or four times from
her side of the branch, then shifted to reach
behind and work from there. She often worked
for several minutes without touching the green
grass which she had just brought. I did not
establish that she worked with the same piece
of grass, but the fact that she did no alternate
pecking on one side and tucking into the other
with the reverse, suggests that she was not work-
ing with a particular blade.

The only action I saw during the building of
the base was “peck-pull around-tuck”— with the
grass held in the leaf base or under the female’s
feet. She put her foot on the nesting material
only during the first day or two (as Schafer also
reports), but worked on the base for twice that
length of time. As more grass was added, the
female pulled harder at the loose and long ends
of the grass.

Comment.— My notes show consistent differ-
ences in the construction of the nest base by the
Yellowtail, when compared to the other oropen-
dolas and cacique. According to Chapman and
Skutch, wagleri and montezuma weave the pliant
material around and around the branch, and
they suggest that the bird wraps an individual
piece and ties it to the branch before working
with a new piece. Schafer’s notes agree more
closely with mine, but in angustifrons he suggests
that when the female brings in the first long
fibers, she does give each one individual atten-
tion in wrapping them around the branch, and
further, that she may perch and swing on the
hanging end as if testing its security.

Phase 3: Transition to the Sides of the Nest.—
Soon there is a loose mop of fibers about 10 cm.
long hanging from the fork of the branch and
extending 7 to 10 cm. along both sides of the
fork, and then the female shifts to the next
phase of building. Now, instead of perching
with both feet on the branch (Text-figs. 3C, D,
E & 4A) , she perches with one foot on the branch
and the other on the mop which is hanging from
the branch (Text-fig. 4B). Her weight on the
hanging fibers causes the apron to elongate. The
female still pokes the grass into the woven ma-
terial near the branch, but now pokes it only
into the mop and not on top of the branch. She
still consistently pulls the grass around some-
thing (either the branch or the edge of the ma-
terial hanging from the base) and uses the peck-
pull around-tuck movement. Thus she weaves
over the branch and also binds or “overcasts”
the edge of the material on the two “open” sides of
the nest’s horseshoe-shaped base (Text-fig. 3F).

46 Zoologica: New York Zoological Society [47:6

Table 1. Comparisons of Behavior in Courtship and Nest Building

Gymnostinops

Zarhynchus

Ostinops

Psarocolius

Character

montezuma

wagleri

decumanus

angustifrons

Cacicus cela

Roosts

Bamboo

Bamboo

Palms

Male’s call

chuck

chac

quic

Alarm

cack

chack-chack

kak-kak-kak

chak-chak

Male’s alarm panics colony

P

P

P

a

Male’s song

tsu ta ta ooo

Deep, liquid.

Melodious

Bell-like

Brilliant and

tsreee kleee a

“hope you

tschuudu

melody 3-6,

varied

wow wow

choke,”

du du du

soft thick

WOW

sputtering

tshuuii wup

crescendo to

cackle, crash

wup wup

explosive 5-6

Territory

Several males, Several males. Several males.

Several males, Several males.

1 tree

1 tree

several trees.

several trees.

1 tree

1 male domi-

1 male domi-

nant and sub-

nant and 2-3

ordinate

subordinates.

males present

Territorial song

eedy eedy ooo

Melodious

cherie du

and shorter

du du

than song.

wup wup wup

emphasis on

2 & 3

Song bow

P

P

p

pa

Raises on toes

P

P

P

pa

P

Neck swollen

P

p

pa

P

Eyes glare

Blue

Blue

Green-brown

Tail up during song

p Flicked

P

Vi

P

Wing-waving during song

P

a

P

pa

P

Male flight noisy

P

p

P

pa

Enlarged bill

P

p

P

P

P

Presence of crest

p Small

p

P

p small

Courtship away from colony

P

p

P

a

P

Dark central tail feathers

P

p

P

P

Females outnumber males

Several times

6 to 1

P

P

P

At least polygamous

P

P

P

P

P

Sex size-difference

P

P

P

P

P

Territorial fights

a

a

a

a

a

Female preens male’s neck

p

a

P

Male pecks female’s cloaca

(a) P

a

P

Females associate in flocks

P

p

P

pa

P

Female threat

raah whine

raaah whine

tcherie

Hiss

garreeoo or raah

Female alone builds

P

P

P

P

P

One tree

P

P

P

a

Nest at end of rain

P

P

P

a

P

Short-distance migrant

a

P

P

a

Female territory

P

P

P

Height of nests

15-35 m.

> 35 m.

10-20-35 m.

5-8 m.

> 15 m.

Roof over entry

a

a

a

a

P

Female chooses site

P

P

Terminal hanging branch

P

P

Isolated tree

P

P

P

a

P

1962]

Drury: Breeding Activities of Yellow tail (Ostinops decumanus)

47

Table 1. Comparsions of Behavior in Courtship and Nest Building ( continued )

Gymnostinops

Zarhynchus

Ostinops

Psarocolius

Cacicus cela

Character

montezuma

wagleri

decumanus

angustifrons

Nests close together

P

P

P

5-10 over
100-200 m.

P

Leeward side of tree

P

P

P

a

Base in whorl of leaf bases

P

P

Previous year’s base

P

P

P

Materials

Slender & fine

Coarse &

turgid

Palm strips and fibers

P

P

P

P

Tendrils— vines

P

20-25 cm. long

80% at start

P

Bark strips

P

P

Air roots
Weed stalks

P

P

P

Grass and sedge

P

P

Bromeliads

Nearly all of

interstices

Lining

Leaves

Dead & dying

Soft leaves,

Dry

d /Bromeliad

Kapok

P

fibers

P

£ < Sedge
O IHeliconia

cotton

Bark

P

Collect

At distance

Near nest

Steal loose ends

p Often

p Often

Rarely

p Often

Nest Building

Base of wrapped tendrils

p (knots)

P

P

P

P

Head over limb, reaches under

P

P

P

P

to grasp
Apron

P

P

P

P

Convert to loop

P

P

P

P

P

By uniting

By uniting

edges

edges

Standing in ring

p

P

a

a

P

Work head down

P

P

(p) a

a

P

Enter by door

P

P

P

May enter

May enter

below

below

Enters on wing— no pause

P

P

a

a

Only if
hurried

Perches and pauses as leaves

P

P

P

P

Female works inside

P

P

P

P

P

Building time

14-16

23-25

9-25

19-33 ad.
17-51 yg.

Nest length

60-120 cm.

55-100 cm.

125-137 cm.

76-140 cm.

30-45 cm.

Large diameter

17-23 cm.

20 cm.

20-22 cm.

20-22 cm.

Incubation days

@ H

17

17-19

19-20

Nestling days

@ 30

36

28-34 or

25-30

31-36

p = action or article is present,
a — action or article is absent,
pa = action or article may be present or absent.
Blank means no observation is available.

48

Zoologica: New York Zoological Society

[47:6

Text-fig. 3. Nest-prospecting and nest-building postures. A. & B. Investigating possible nest site, Phase 1,
at Nest 1. (Note cluster of previous year’s leaf bases). C. Looking over valley before leaving, Phase 2,
when nest base is still a ball-like snarl at the fork. D. & E. Reaching over and down for a loose end to pull
up over and tuck, Phase 2. F. Perched on the hanging apron, reaching across to pull around and tuck— the first
stages of closing the entrance, Phases 4 and 5. G. Pattern of grasses in Nest 5, with female perched in
entrance feeding young. H. Detail of pattern of weave in the side of a nest. 1. Detail of the weave at the
base of the entrance— Nest 9.

At first ( at least one day) she perches as often with
both feet on the branch as with one foot down,
but as the material gradually lengthens she
spends more time perched on the hanging ma-
terial. Later she may again perch with one foot
on the branch, even after the two halves are
partially joined together (see below).

Phase 4: Buldiing the Sides of the Entrance —
Now, as the female returns with fibrous material,
she perches with both feet on the hanging,
tangled mop. She works between her feet with
her head up, pokes her load of grass into the
woven material, reaches out and around, pecks,
pulls out and up, brings head in, and pokes with
three or four shakes of her bill; or she pecks,
pulls out and around behind, and pokes in be-
hind with a shake.

At this stage she adds a new motion : she first
stuffs the large beakful of rather short material
into her work; then she pushes her bill through
the material, grasps a bit in the tip of her bill,
pulls it back toward her, then moves it horizon-
tally and pokes it into the weave again— with or
without a short (four or five) shake. I refer to
this as “horizontal peck-pull-poke.” Schafer con-
siders this action specific to the wadding or
filling {hour rage intercalaire) .

As long as she weaves with both the peck-pull
around-tuck and the horizontal peck-pull-poke,
she weaves the supporting structure— the horse-
shoe-shaped cross-section— with overcast edges.
This part of the nest often has holes in it because
the bird spends most of her time overcasting
the edges. She spends less time working with

1962]

Drury: Breeding Activities of Yellowtail (Ostinops decumanus)

49

Text-fig. 4. Nest-building postures. A. Working on the nest base. Phase 2 at Nest 1. B. Working on the
apron, Phase 3 at Nest 9, when still perching on the branch. C. Perched on the two halves of the apron,
Phase 5 at Nest 10, starting to work the edges together. D. Branch tip of Erythrina, showing this year’s
cluster of leaves and last year’s shriveled leaf bases. E. & F. Perched inside nest, working on the nest bag
below the entrance hole. Phase 6 at Nests 13 and 19. G. & H. Reaching below feet to pull hanging pieces
in closing the bottom of the nest, Phase 7 at Nests 13, 15, 19.

the horizontal peck-pull-poke which weaves the
fabric of the bag. The supports of the nest are
thus straplike, tied together tightly by connec-
tors, but they do not form an evenly woven
bag (Text-fig. 3H) ; crudely, they resemble the
braided handles of a string shopping bag. Schafer
says that decumanus usually works head-down
(in contrast to my observations) , but that angus-
tifrons works on the apron head-up.

Phase 5: Closing the Base of the Entrance.—
(Schafer calls this the ring and future entrance) .
To start joining the two sides below the entrance,
the female uses the horizontal peck-pull-poke on
the material which hangs loose and frayed across
the open bottom of the horseshoe-shaped sleeve,
and, as Schafer says, she uses long pieces at
this stage. At first she perches on one side and
takes a bit of grass from the same side, pulling

it across to poke it into the other (Text-fig. 3F) .
Finally, she perches with one foot on each side
of the horseshoe, and works strips across from
one side into the woven part on the other (Text-
fig. 4C) . She may work either on the outside or
the inside of the sleeve, and may spend all of
several visits below the entrance, working hor-
izontally on the lower part of the structure. This
transition may be seen as a gradual increase in
use of the horizontal peck-pull-poke, until it is
applied all around the bag and replaces the peck-
pull around-tuck. She closes the gap, still using
a combination of the horizontal weaving and
overcasting actions, but her attempt to join the
sides is not immediately successful; usually she
starts to bring the two sides together about a
foot above the place where they are ultimately
joined. The entrance may be 30 to 45 cm. long

50

Zoologica: New York Zoological Society

[47:6

if she starts to close it soon after the apron is
large enough for her to perch on, or it may be
60 to 90 cm. long if she continues to weave with
the overcasting action. This accounts for differ-
ences in lengths of the hanging nests. Nests vary
from just over two to more than four feet long
(60 to 120 cm.). Schafer emphasized that the
difference in experience between young and old
females may explain the observed differences in
length.

When the base of the entrance is closed and
the apron has become a sleeve, the female spends
a long time working with the overcasting motion
on the margin at the base of the entrance. At
this point the pull-around is vertical. Thus she
weaves a buttonhole-like stitch along the bottom
of the entrance (Text-fig. 31). Her weight, com-
ing and going, pulls the material down, and by
pulling it thus makes the dense weaving stronger
where she perches to enter. Before the entrance
is well sewed in, she usually still hitches up
woodpecker-like to the branch to perch and look
around before she flies off. Entrances are on the
lee side of the nest.

Phase 6: Building the Sides of the Nest.—
While she is working on the sides and supports
of the nest, the female characteristically works
on the inside, but often works on the outside.
As soon as the bottom of the entrance is firmly
closed, she does almost all her work on the in-
side. As far as I could see, she uses chiefly the
horizontal peck-pull-poke action and brings
larger beakfuls of material (on an average 30
to 60 cm. long).

As the female works inside the bag, she sits
woodpecker-like (her tail may be bent up be-
hind her in the the sleeve and occasionally her
wings may be partly opened (Text-fig. 4E & F).
She works four or five minutes, reaching slowly
and carefully through the weave to grab a bit
of fibrous material in the tip of her bill, pull
it toward her, then poke it through again, often
with a little shake. There is less pulling, tugging
and wrapping of long pieces, and more delicate
attention to detailed weaving. Also, she spends
much less time, proportionately, gathering build-
ing materials at this stage.

Now she starts to add a new movement. After
about every ten to fifteen horizontal weaving
actions, she reaches down between her feet to
peck, pull up and poke (Text-fig. 4G & H). The
longer she continues to use only horizontal
weaving, the longer the bag of the nest.

Where two nests are built immediately next
to each other, the female may carry her weaving
action across into the material of the adjacent
nest, as Skutch pointed out. I observed this in

the case of nests #7 and #13 (Text-fig. 1).
Skutch described how one bird, building thus
into a neighbor’s nest, seemed to make its own
nest an additional length below the bottom of
the neighbor’s nest so that there was a “proper”
length of bag hanging down. The two females I
watched did the same. This suggests that a
definite length of bag stimulates the bird to the
next activity, which leads to closing the bottom.
Crook’s (1960) discussion of the change of
stimuli which leads to changes in the building
techniques used by weaver finches agrees with
this.

Phase 7: Closing the Bottom— During the
building of the sides, the female perches with
both feet at the same level, above or just below
her bill, and works reaching ahead, beside or
between her feet. In closing the bottom, she
reaches way down between her feet, almost
doing a somersault (Text-fig. 4G and H), grasps
the long, hanging strands, pulls them up to eye-
level, and weaves them into the side— pulling
and poking with little or no evident shaking of
her bill. As she moves around she gradually
takes more and more bits from one side and
weaves them into the opposite side. This closes
the bottom. But because of her weight and her
inefficient actions, she is not immediately suc-
cessful. The stage at which she shifts to closing
the bottom and her success with the cross-
weaving govern the length of the hanging bag.
She continues to weave horizontally, alternating
work on the sides and on the bottom. Schafer
reports that young female angustifrons may have
to try several bottoms before they succeed.

Chapman, Skutch and Schafer commented
that their birds entered the nest only through
the entrance— once the entrance is formed.
While building the sides, the Yellowtail usually
enters and leaves by the entrance, but may often
fly in and out through the bottom of the nest,
even after she has started to close the bottom.

When the first strands are hooked across the
bottom, the female spends much time weaving
between her feet, pulling especially hard on the
grass with which she is working. During this
action her wings are often partly raised and
her tail is pushed hard against the side of the
bag (Text-fig. 4G). Her pushing movements
inside and her weight as she climbs up the sides
combine to enlarge the gourd-like bottom, and to
pull the strands taut. Schafer comments on the
joggling and pushing inside the nest during this
phase, which forms the 25 cm. diameter of the
bag.

Phase 8: Lining the Nest.— My observations at
one nest suggest, as do Tashian’s, that the birds

1962]

Drury: Breeding Activities of Yellowtail (Ostinops decumanus)

51

use bits of leaves, chiefly erythrina, to line the
nests. I could not see the motions used in form-
ing this lining because the nest screens them.
Movements visible on the outside of the bag,
however, showed that the bird frequently
changed her position, joggling the nest, and that
her wings and tail were raised. Other birds,
whose nest-molding activity has been recorded
in detail, thrust the breast forward by pushing
the feet back and up, and Schafer’s observa-
tions show that decumanus tramples the nest-
lining materials (torn fragments of dried, brown
leaves) while turning around and around.

This is the period of the female’s receptivity
to copulation.

Attentiveness to Building— The birds build
fitfully as they weave the base and may be absent
for several days. Schafer says the period lasts
one to nine days, according to the level of
stimulation and to the female’s age. As the sides
of the nest are started, they still build sporad-
ically and may be absent for a day or two, but
then work with energetic spurts of concentrated
activity. Schafer says the closing of the ring
takes one to six days and that an inexperienced
angustifrons worked 16 days and built down
1 m. before successfully closing the entrance.
After the birds start to weave the hanging sleeve,
they work constantly until the bottom of the nest
is partly closed, which takes four to five days
(Schafer). Then they may be absent for a day
or two before the bottom is brought together.
Once finally started on the bottom, they work
constantly until it is closed, but they may pause
again before weaving the thick bottom and lin-
ing the bag which takes two to six days. Even
when the heavily woven bottom is finished, the
female may spend hours slowly and carefully
weaving horizontally part way up the bag.

Discussion

Function of the Territory. — Skutch and
Chapman commented on the peculiar territory
structure in the oropendolas. They were unable
to establish whether a male took up a territory
and defended it, and whether pairing was prom-
iscuous or polygamous. Schafer compares terri-
torial behavior in decumanus and angustifrons
in detail but does not comment.

Males seem to give their attention temporarily
to one group of females and readily shift atten-
tion to a new group. In the Yellowtail, my ob-
servations point to isolation of males on their
own defended territories, but although I watched
for 35 hours at one tree with 43 nests (long
enough to expect to observe frequent copula-
tions if they were restricted to the male whose

territory included the nesting tree), I saw copu-
lation there only twice. During the same period
I saw eight copulations in trees with no nests.
This suggests either (a) that the females are
nesting in one particular male’s territory but
each is paired to a specific male whose territory
may be elsewhere, or (b) that the females nest
together in one male’s tree by flock formation
among the females without regard to sexual
relations with any particular male. There may
be polygamy, no pair formation or “standard”
pair formation. The restriction of the term “pair
formation” to those cases in which copulation
takes place only with the pair partner, may not
necessarily apply to these birds. Such limitations
gratuitously suggest some form of propriety—
which may be anthropomorphic. My notes, and
Schafer’s, show a hierarchy of several males
associated with a colony tree. Thus there is
territory but its exclusiveness is modified.
Schafer points out the lowering of territorial
“jealously” during the copulation period in
decumanus and what appears to be complete
promiscuity in both species where inferior males
intrude to copulate with receptive females while
the dominant male is occupied.

Chapman and Skutch suggest that groups of
females establish a new nesting site and that the
presence or absence of the male is unimportant;
in fact, the males seem to follow groups of fe-
males. Schafer’s observations, especially of an-
gustifrons, show the necessity of the territorial
males’ constant stimulation to arouse and carry
through the females’ interest in nesting.

These observations and mine lend support to
Tinbergen’s (1957) explanation of the function
of territory; namely, the combination of a need
for the male to act in a specific way recognizable
by females and other males during pair forma-
tion, coupled with a need for the male to have
a fixed location. The site, in the case of the
oropendolas studied by Skutch and Chapman,
seems to be a group of females rather than a
map area. With this change in structure, it may
become selectively advantageous to the species
to de-emphasize the aggressive aspects of the
display of the male or lower his tendency to
exclude other males. Thus, the male goes
through his song display to stimulate the female,
but he is not necessarily bound to drive away
other males, and the female does not necessarily
restrict her attentions to one male or the male
in whose territory she builds. I can see no way
in which food enters directly into the selective
advantage of territory in this case.

Closing the Ring.— The building of the en-
trance by the Yellowtail contrasts with that
described by Skutch and Chapman for their

52

Zoologica: New York Zoological Society

[47:6

species. Their birds weave an apron across the
fork and then, standing on the material, push
an entrance through the partly woven material
which then becomes a circle or loop. This loop
is used as a perch, and the sleeve is woven
downward from it— the bird hanging head down
into the sleeve. Although they do not dwell on
details, the descriptions of these authors suggest
that their species use techniques of building the
base and loop which are similar to those used
by the True Weavers (Ploceidae) as reported
by many authors (Friedmann, 1922, Ali, 1931;
Grzimek, 1952; Collias, 1959; and Crook,
1960).

The techniques which Schafer and I describe
are similar, and differ from those of the weavers.
While the weavers make a ring first and build
the nest bag out from the ring, which becomes
the mouth of the bag, decumanus and angusti-
frons build a hanging apron (horseshoe-shaped
in cross-section) directly down from the foun-
dation on the branches and join the free ends
to form the ring and the entrance. The tech-
niques of weavers and oropendolas resemble
each other in the weaving of free-hanging ends
from one side into the other, but differ in the
location of the ring in forming the nest foun-
dation, and (at least in decumanus) differ also
in that the female oropendola works with her
head at the level of her feet or above, while the
male weaver works with his head down below
his feet.

Skutch watched one montezuma after her
properly formed circle was broken when her
neighbor stole some loose ends. The robbed fe-
male perched with one foot on each side of the
ring and in this “uncomfortable” position closed
the two sides in the same way as I observed
the Yellowtail to do. This suggests that the two
systems may not be fundamentally different in
the oropendolas.

Entering and Leaving— Both Skutch and
Chapman observed that the returning bird darts
swiftly into the nest entrance as soon as it has
been formed. They suggest that the fast dis-
appearance into the bag and the long look
around before leaving serve to avoid predation.
In building the foundation and sides of the nest,
however, the female Yellowtail perches on the
outside and is conspicuously exposed. Even so,
she peers around over the valley just before fly-
ing away. Also, later in the cycle, when she
visits the completed nest or comes to feed the
young, she perches for ten to twenty seconds
with her head down inside the bag and her
yellow tail hanging out conspicuously. This
action must deny the significance of the fast dart

into the entrance as only to avoid predation.
There may be advantage to the peering around
before leaving, but it would seem that the tech-
nique of entering may be dictated as well by
the bird’s heavy wing-loading which exposes
her to the danger of stalling as she flies up
sharply to the entrance.

Techniques of Nest Building.— Herrick (1911)
pointed out that the stereotyped movements
used by birds to build their nests are con-
venient tools for the comparative study of be-
havior and its evolutionary aspects, yet little
work has been done since. Later, Laven (1940a)
repeated this suggestion. Several authors have
described the nest-building activities of birds
that nest on open ground, especially the non-
passerine species, e.g., Selous (1902) and Brock
(1911) for the Lapwing, Vanellus vanellus;
Portielje (1925) for the Cormorant, Phalacro-
corax car bo, and (1928) for the Herring Gull,
Larus argentatus; and Tinbergen (1931) for the
Common Tern, Sterna hirundo, and (1936) for
the Herring Gull. Several studies have shown
how universal certain nest-building actions are.
Although the loons, Gavia stellata, (Huxley,
1923) and grebes (Huxley, 1914; Selous, 1901)
merely drop their nesting weeds, moss or mud
on the nest edge, most birds add them with some
form of tremble-shove: Cormorant (Portielje,
1925, and Kortlandt, 1940); herons (Lorenz,
1955); storks (Schiiz, 1943); and perching
birds, Raven, Corvus corax (Lorenz, 1940).
Friedmann’s (1922) study of the building actions
of the Ploceidae was one of the very few on
perching birds until the Second World War. In
1943, Nice mentioned the appearance of gener-
alized nest-building actions in the developing
behavior of young Song Sparrows, Melospiza
melodia, and since then several authors have re-
ported on the early appearance of these funda-
mental actions (Dilger, 1956; Goodwin, 1954;
Kramer, 1950; Nicholai, 1956; and Schiiz, 1943).

Nest-building actions, more or less modified,
are used by many species as part of courtship
actions, e.g., Great Crested Grebe, Cormorant,
herons, woodpeckers, Lapwing and other shore-
birds, Alcidae and estrildid and ploceid finches.
In addition, certain of the actions associated with
nest building, and thus presumably primarily
sexual, have been transferred to aggressive ac-
tion, e.g., scraping by Ringed Plover, Charadrius
hiaticula, (Laven, 1940), and Killdeer, Chara-
drius vociferus; nestling and scraping by Col-
lared Flycatcher, Ficedula albicollis, (Lohrl,
1951; Curio, 1960); and grass pulling by Her-
ring Gulls (Tinbergen, 1951 and 1952). Moyni-
han (1955) argues, however, that grass pulling
by Herring Gulls is not transferred from sexual

1962]

Drury: Breeding Activities of Yellow tail (Ostinops decumanus)

53

motivation, but is directly aggressive as redi-
rected attack.

Kluijver (1949/1955) seems to have started
the revival of detailed studies of nest-building
techniques. He describes the building actions of
the Great Reed Warbler, Acrocephalus arundi-
naceus; van Dobben ( 1949) describes the build-
ing actions of the Icterine Warbler, Hippolias
icterina and Chaffinch, Fringilla coelebs; and
Kramer (1950) describes the nest-building ac-
tions of the Red-backed Shrike, Lanius collurio.

Kluijver, van Dobben and Kramer define the
fundamental actions of nest building in passerine
birds as three: (1) nestling— the bird presses its
breast down into the nest-cup, usually with bill
and tail pointed upward; (2) trampling— the bird
presses its breast to the bottom of the nest-cup
and kicks vigorously and repeatedly with each
leg, backward and upward (3) pecking, tugging
and tucking— the bird reaches forward and
grasps nest material, pulls it one way or the
other, and tucks it into the nest again. These
actions are the same as Herrick (1911) de-
scribes for the nest building of the American
Robin, Turdus migratorius, Red-eyed Vireo,
Vireo olivaceus, and Baltimore Oriole, Icterus
galbula. Additional detailed studies of a number
of passerine species show how widely distributed
these actions are : Sylviidae— Lesser Whitethroat,
Sylvia curruca, and Blackcap, Sylvia atricapilla,
Dechert (1955), Icterine Warbler, van Dobben
(1949); and Great Reed Warbler, Kluijver
(1949/1955); Paridae— Long-tailed Tit, Aegi-
thalos caudatus, Maxse (1951), Tinbergen
(1953b), and Bearded Tit, Panurus biarmicus,
Koenig (1952); Sittidae— European Nuthatch,
Sitta europaea, Lohrl (1958); Ploceidae— Baya,
Ali (1931), and Village Weaver, Grzimek
(1952), and other weavers, Crook (1960);
Estrildidae of several species, Kunkel (1959);
Turdidae— European Blackbird, Turdus merula,
E. and I. Messmer (1956); Muscicapidae,
Lohrl (1951), Curio (1960); Fringillidae—
Chaffinch, van Dobben ( 1949) , Marler ( 1956) ;
and Emberizidae— Song Sparrow, Nice (1943).

All these studies show that the movements
used by passerine birds in placing nest material
and forming the nest-cup are uniform and wide-
spread, but as Dechert (1955) illustrated, the
important thing is the sequence of these actions,
the materials used and how they are used. Her
study of the Lesser Whitethroat and Blackcap
showed that these closely related species used
nearly identical actions but different materials,
and that they used the actions in different pro-
portions, thus creating quite different nests. A
further illustration of the importance of the ma-
terial chosen is given by Lorenz (personal com-

munication) who, when first keeping some Red-
billed Weaverbirds, Quelea quelea, discovered
that they were unable to build their huge nests
because the grass native to Germany did not
“stick.” He obtained some of the grass the birds
use in their native Africa and found that the
leaves of this grass are “retrorsely scabrous” on
their margins (have small tooth-like spines on
their edges) , causing the leaves to cling together.
Ali (1931) realized the importance of the
scabrous margins of rice leaves in the nest build-
ing of the Baya. Schafer’s study also shows the
importance of different nesting materials in simi-
lar species, but his study also shows use of dif-
ferent actions (in anchoring the base). Nickell,
(1958) examined nesting materials and nest
types of 169 species of eastern North American
species.

In a number of families scattered among the
perching birds, some of the generalized actions
have atrophied and some actions have been
added to the basic repertory. The estrildid
finches (Kunkel, 1959) built their messy,
domed nests with the following actions: (1) the
bird, standing in the middle of its nest, grabs
material with its bill and pushes it away, or it
may simply push at the wall and roof of the nest
with its head; (2) the bird grasps material and
pulls sideways, either to the left or to the right;
(3) the bird pulls material directly toward itself
into the cup. In these actions, the birds have lost
certain of the basic behavior sequences and their
nests seem to reflect this.

It is interesting, in terms of the former classi-
fication of the estrildid finches with the ploceid
finches (Steiner, 1955), that they share two
unusual nest-building actions: (1) termination
(Crook, I960)— a grass stalk held in the beak
is moved by a rotating motion of the bill until it
is held at one end; (2) to stretch the building
material and form the pocket-like nest, the
builder pushes against the walls with its head or
bill.

The highest development in nest building,
nearly all authors agree, occurs in two families
—the Old World True Weaver Finches (Plocei-
dae) and the New World blackbirds, troupials
and orioles (Icteridae). Ploceids such as the
Red-billed Weaverbird studied by Friedmann
(1922), the Baya described by Ali (1931), the
Village Weaver described by Grzimek (1952)
and studied and photographed by Collias
(1959), and the species discussed by Crook
(1960), use additional actions to weave and tie
knots in their material in order to fasten it to
the foundation branches. According to Fried-
mann (1922) and Crook (1960), the True
Weavers take a fiber in the bill, hold it at the end,

54

Zoologica: New York Zoological Society

[47:6

place it on a branch (sometimes holding it with
the foot), then take one end of the fiber and
push it to the far side of the branch. Reaching
around the other side, they take the strand and
tuck it under the part they are standing on, then
pull taut the knot that has been made. The fiber
is repeatedly drawn round and round, each time
being inserted within the previous loop. The end
may also be looped in and out through the al-
ready-woven fabric.

Collias (1959) describes “four basic mech-
anisms in working a grass strip into its nest. He
tends (1) to bend the strip about some object,
either a twig, another grass strip, or his own leg;
(2) to double a strip back on itself; (3) to alter-
nate the direction in which he winds the strip
about objects such as twigs or other grass strips;
and (4) to poke and pull a strip through holes,
normally the interstices of the nest materials. It
is in the possession of mechanisms (3) and (4)
that a true weaver (Subfamily Ploceinae) differs
from other weaverbirds. . . The end of a strip
is often looped back on itself in such a way that
pulling on the strip tightens its attachments. This
is essentially a hitch type of knot. Knots of other
types than the hitch are rare.” In this specific
treatment of an individual fiber, the action of
these birds differs from that of many Icteridae.
They tuck a bill full of fibers into the nest mate-
rial and then may ignore them until dealt with
again “by accident.”

In the ploceid finches, the foundation of the
nest is made by forming a loop on which the
bird stands and which it uses as an entrance
from which to build the sleeve that will become
the nest bag. The caciques and oropendolas
studied by Skutch and Chapman resemble the
True Weaver Finches in building a similar loop.
These authors also suggest that the birds wind
the strands around the branches to form the
foundation of the nest. However, my study of
the Yellowtail, and Schafer’s, show that such
behavior is not characteristic of decumanus and
angustifrons.

The New World orioles have specific actions
peculiar to themselves which are slight modifica-
tions of the generalized ones of most perching
birds: pecking, pulling around and tucking
(Herrick, 1911, Baltimore Oriole). Herrick
showed that after the nest bag was constructed,
the Baltimore Oriole also used the standard
trampling, nest-molding technique. The fact that
the enormous and complex nest of the Yellow-
tail is built by simple actions which are used by
other species to make much simpler nests, points
to the generalization that variation is built out
of specializations of a few fundamental “inven-
tions.”

To emphasize this, Selous (1902) and later
Kramer (1950) pointed out that the scraping
actions of plovers and sandpipers, gulls and
terns, are homologous with the trampling actions
of perching birds making their cup-shaped nests.
Similar trampling actions are found among a
number of the Laro-limicolae: Lapwing, Selous
(1902), Brock (1911), Rinkel (1940) and
Laven (1941); Turnstone, Arenaria interpres,
Bergman (1946); Northern (Red-necked)
Phalarope, Phalciropus lobatus, Tinbergen
(1935); Common Tern, Tinbergen (1931);
Caspian Tern, Hydroprogne caspia or tsche-
grava, Bergman (1953); and Herring Gull,
Tinbergen (1936). Furthermore, these studies
show that there are other nest-building actions
in gulls and terns (and my own studies show
that the same is true in sandpipers, Calidris
bairdii, and plovers, Charadrius vociferus and
Ch. melodus, similar to those of perching birds.
When sitting on its scrape, the bird picks up
material and either drops it over the shoulder
or pulls it immediately in front of itself and
drops it. These actions are identical to the side-
ways pulling or the peck-pull around-tuck action
of perching birds. The big difference, of course,
is in the choice of nest-building material, the
uniformity of treatment and concentration given
to the material. Shorebirds and gulls either use
no material, cast the material aside, or do not
pursue the treatment to a final resting place.

Many small perching birds may spend two or
three days building; the Red-eyed Vireo spends
approximately five. The Yellowtail may spend
three to five weeks. A Killdeer may spend two
to three weeks scraping, and then continue for
another three weeks throwing nest material over
its shoulder, or pecking, pulling and dropping
material on the edge of the nest after the eggs
have been laid until they hatch. The bird usually
kicks with its feet five to ten times in the tramp-
ling movements each time it settles on the nest.
In the Piping Plover the persistence of this
action is functional. It is used to uncover eggs
that have been buried by blowing sand. Similar
actions are used to uncover buried eggs by the
Kentish Plover, Charadrius alexandrinus, and
the Little Ringed Plover, Charadrius dubius,
(Walters, 1956), and by Kittlitz’s Sandplover,
Charadrius pecuarius, (Hall, 1958), which regu-
larly covers its eggs when frightened from them.

Clearly, the type of action and the amount
of time spent on the action does not control the
end product. But there seems, in fact, to be a
sequence in the intensity of attention paid to
the nesting material from (a) the very careless
handling by shorebirds and gulls through (b)

1962]

Drury: Breeding Activities of Yellowtail (Ostinops decumanus)

55

the sloppy work of the estrildid finches, to (c)
the situation found in most perching birds, and
finally (d) to the True Weavers and icterids.
In the evolution of nest building (which is at
least in part independent of the evolution of
those factors used in classification), similar
motions have been applied to different types of
material, and a change has been made accord-
ing to differences in the “plan in mind” and
perseverance to that plan: specialized actions
added, and new reactions appearing to specific
building situations. The difference is not in the
motion (the tool) but in the central nervous
system, which in turn is modified, within limits,
by changes in environment which influence the
living bird (see especially the differences in
the nest building of young and old female
angustifrons studied by Schafer).

Stereotyped Behavior and Brain Structure.— In
considering the modification of behavior patterns
in evolution and according to experience of the
individual organism, the basic difference in brain
structure between birds and mammals must be
taken into account, as described by Cobb
(1960). In birds, the basal area of the fore-
brain, the corpus striatum, has been greatly
enlarged. This is the area associated with quick,
complex physical co-ordination. In contrast, the
roof of the forebrain in mammals has been
developed— an area whose specialization has
tended toward associations developed during the
experience of the individual.

The impression of bird behavior is of stereo-
typed responses (consistent and efficient) that
are uniform as they occur, but whose occurrence
is modifiable between species and according to
experience of the individual bird. Certain stimuli
or events change the bird’s instructions and its
behavior shifts. When watching the Yellowtail
construct its nest— operating steadily and mech-
anically with a smooth, weaving action for a
period of time, and then shifting over to a new
system of weaving— I was strongly reminded of
the control on an automatic loom by a card.
The machine’s motions— and the bird’s— do not
have the ability to change the card. The ability
to change the card must, in a very crude way,
be the organism’s ability to learn and associate.
The bird’s actions are efficient in its present cir-
cumstances and, through evolution, suitable to
its environment. During the bird’s life, it seems
that inherited material is as if carded and in
units, and during maturation the bird “learns”
what conditions, stimuli or parts of the environ-
ment are suitable for the expression of that
inborn activity. The nest building of the Yellow-
tail is an illustration of the complex train of
activity patterns that is changed to another

complex train by a stimulus, rather than an
illustration of a series of simple actions built
up by association into complex trains, as seems
to be the product of much mammalian learning.

But this illustration does not clarify the crux
of the questions presented by the shifts in nest-
building behavior by the Yellowtail. What com-
bination of “things” leads to the decision to
change the “card?” How does the bird treat
inconsistent information upon which it must
base its decision? For example, the behavior
of a bird whose nest has been partly destroyed
by the robbing of nest-building material by its
neighbor. There must be a “look” (Thorpe,
1956; Crook, 1960; ? Gestalt) about the mate-
rial and the state of construction which influ-
ences the sequence of activities that the bird fol-
lows. This hypothesis is supported by several
natural tests resulting when several males work
at the same nest (Crook, 1960). The bird’s
actions are neither guided exclusively as if by the
unravelling of a string, as seem to be the activ-
ities of an insect, nor yet with the gradual brick-
by-brick building of the association of learned
actions into a whole, as may be the case with
much mammalian development.

Summary

1. Yellowtails nest in colonies, 2-43 nests in
the Arima Valley, chiefly in Erythrina micro p-
teryx, an introduced shade tree towering over
the cocoa, coffee and banana tree plantations.

2. A dominant male defends twoor three neigh-
boring trees but allows intrusion of subordinate
males if the intruder does not display. Males
have a territorial and a sexual song.

3. Males hold territories in nestless trees, and
females copulate with males other than the dom-
inant male of the colony tree.

4. Groups of females within a colony, nest ac-
cording to a schedule independent of other
groups.

5. Nest-building material is collected in valley
bottoms away from the colony tree; it is chiefly
grass, sedge and thin vines. The building mate-
rial averages 30-45 cm. when the female starts
the base, 60-100 cm. when building the supports
and sides of the entrance, 30-45 cm. in weaving
the bag, and 100-150 cm. when closing the
bottom. The nest appears to be lined with dried
leaves.

6. Detailed observations of nest-building tech-
nique show that a few simple movements are
used and that the fibrous strips are dealt with
by chance— not woven in individually. Actions
are:

56

Zoologica: New York Zoological Society

[47:6

(a) Push billful of material into the work,
reach over and grasp any end, pull around and
tug, tuck in with tremble-shove. The female may
perch on the branch or on the apron that she
in making when using this technique;

(b) Push billful into the work, push bill
through and grasp end, pull out and move to
the side, tuck with a shake. The female perches
woodpecker-like on the side on the nest;

(c) When closing the entrance, the female
weaves across the two free edges, then overcasts
the edges to bind the entrance with a buttonhole-
like stitch;

(d) She uses chiefly the horizontal move-
ments in making the bag, whose length seems
to be measured by distance below the entrance.
In closing the bottom, she reaches down between
her feet to peck, pull up and poke.

7. The bird’s attentiveness to nest building
varies, being least at the start, strongest when
building the sides of the nest, and interrupted
before the bottom is closed and the nest lined.

8. Discussion:

(a) Territory in Yellowtails suggests purely
courtship function. Copulation appears to be
promiscuous;

(b) Closing the entrance and weaving the
sides of the nest, as observed by Schafer and
myself, use different techniques than have been
described so far;

(c) It is not clear why females hesitate on
the nest entrance when entering and leaving;

(d) A review of nest-building techniques
shows that three fundamental movements are
widespread even among non-nest builders such
as gulls and plovers. To these movements some
groups have added actions by which they con-
struct sloppy or elegant nests. The fine nests of
the ploceid weaver finches are based on per-
sistent individual attentions which tie knots in
the fibers. Among the Icteridae, the fiber gets
attention by accident, once it has been pushed
into the work. Closeness of weave and security
of attachment result from persistent repetition
of actions.

Literature Cited

Ali, S. A.

1931. The nesting habits of the Baya (Ploceus
philippinus). Joum. Bombay Nat. Hist.
Soc., 34 (4): 947-964.

Bergman, G.

1946. Der Steinwalzer, Arenaria i. interpres
(L.), in Seiner Beziehung zur umwelt.
Acta. Zool. Fenn., 47: 3-144.

1953. Verhalten und Biologie der Raubsee-
schwalbe ( Hydroprogne tschegrava).
Acta. Zool. Fenn., 77: 3-45.

Brock, S. F.

1911. Lapwings ( Vanellus vulgaris) in the pair-
ing season. The Zoologist, 15: 296-304.

Chapman, F. M.

1928. The nesting habits of Wagler’s Oropendola
(Zarhynchus wagleri) on Barro Colorado
Island. Bull. Am. Mus. Nat. Hist., 58:
123-166.

Cobb, S.

1960. Observations on the comparative anatomy
of the avian brain. Perspectives in Biology
and Medicine, III (3): 383-408.

COLLIAS, N. E. & E. C. COLLIAS.

1959. Breeding behaviour of the Black-headed
Weaverbird, Textor cucullatus graueri
(Hartert), in the Belgian Congo. Proc.
First Pan-African Cong., Ostrich, Supp.
No. 3: 233-241.

Crook, J. H.

1960. Nest form and construction in certain
West African Weaver Birds. Ibis, 102 ( 1 ) :
1-25.

Curio, E.

1960. Ontogenese und Phylogenese einiger T rieb-
ausserungen von Fliegenschnappern.
Journ. fur Ornith., 101: 291-309.

Dechert, G.

1955. Beitrage zur nestbautechnik deutscher
Sylviiden. Journ. fur Ornith., 76: 186-206.

Dulger, W. C.

1956. Nest-building movements performed by a
juvenile Olive-backed Thrush. Wils. Bull.,
68: 157-158.

Friedmann, H.

1922. The weaving of the Red-billed Weaver-
bird in captivity. Zoologica, 2: 355-372.

Goethe, F.

1937. Beobachtungen und Untersuchungen zur
Biologie der Silbermowe ( Larus a. argen-
tatus Pontopp.) auf des vogelinsel Mem-
merstand. Joum. fUr Ornith., 85 (1): 1-
119.

Goodwin, D.

1954. Juvenile Mistle Thrushes showing repro-
ductive behaviour-patterns. Brit. Birds,
47: 81-83.

Grzimek, B.

1952. Zum Balz verhalten des Westafrikanischen
Textor-Webers ( Hyphanthornis ( Ploceus )
cucullatus.) Zeits. Tierpsychol., 9: 289-
294.

Hall, K. R. L.

1958. Observations on the nesting sites and nest-
ing behaviour of the Kittlitz’s Sandplover
{Charadrius pecuarius) . Ostrich, 29: 113-
125.

196,2]

Drury: Breeding Activities of Yellowtail (Ostinops decumanus)

57

Herrick, F. H.

1911. Nests and nest building in birds. Journ.
Animal Behavior, 1: 159-192, 244-277,
336-373.

Huxley, J. S.

1914. The courtship habits of the Great Crested
Grebe ( Podiceps cristatus ); with an addi-
tion to the theory of sexual selection. Proc.
Zool. Soc., London, 1914, H: 491-562.

1923. Courtship activities in the Red-throated
Diver ( Colymbus stellatus Pontopp.); to-
gether with a discussion of the evolution
of courtship in birds. Journ. Linn. Soc.,
35: 253-291.

Kluijver, H. N.

1949. In het Voetspoor van Thijsse. H. Veen-
man, Wageningen.

1955. Das Verhalten des Drosselrohrsangers,
Acrocephalus arundinaceus (L.), am Brut-
platz mit besonderer Beriicksichtigung
des Nestbautechnik und Rivierbehaupt-
ung. Ardea, 43: 1-50.

Koenig, I. O.

1952. Okologie und Verhalten der Vogel des
Neusiedlersee - Schilfgurtels. Journ. fur
Ornith., 93: 207-289.

Kortlandt, A.

1940. Fine Ubersicht der angeborenen Verhalt-
ensweisen des Mittel-europaischen Kor-
morans ( Phalacrocorax carbo sinensis,
Shaw & Nodd.), ihre funktion, onto-
genetische Entwicklung und phylogene-
tische Herkunft. Arch. Neerl. Zool., IV:
401-442.

Kramer, G.

1950. Der Nestbau beim Neuntoter ( Lanius col-
lurio L.). Orn. Berichte., Ill (1): 1-14.

Kunkel, P.

1959. Zum Verhalten einiger Prachtfinken (Es-
trildinae). Zeits. Tierpsychol., 16: 302-
350.

Laven, B.

1941. Beobachtungen uber Balz und Brut beim
Kiebitz ( Vanellus vanellus L.). Journ. fiir
Ornith. Erganzungsband III (1941): 1-64.

Laven, H.

1940a. Studies on nest building (Nestbaustudien).
Ornith. Monatsberichte, 48: 128-131. (A
Review).

1940b. Beitrage zur Biologie des Sandregenpfei-
fers ( Charadrius hiaticula L.). Journ. fiir
Ornith., 88: 183-287.

Lohrl, H.

1951. Balz und Paarbildung beim Halsbandflieg-
enschnapper. Journ. fiir Ornith., 93 : 41-60.

1958. Das Verhalten des Kleibers (Sitta europaea
caesia Wolf). Zeits. Tierpsychol., 15: 191-
252.

Lorenz, K.

1940. Die Paarbildung beim Kolkraben. Zeits.
Tierpsychol., 3: 278-292.

1955. Morphology and behavior patterns in
closely allied species. In: “Group Proc-
esses,” Trans. 1st Conf., 1954: 168-220.

Marler, P.

1956. Behaviour of the Chaffinch, Fringilla coe-
lebs. Behaviour, Suppl. V, 184 pp.

Maxse, V.

1951. The British Long-tailed Tit ( Aegithalos
caudatus rosens ) : How its nest is built.
Proc. 10th Int. Orn. Cong., Uppsala, 1950:
564-566.

Messmer, E., & I. Messmer

1956. Die Entwicklung der Lautausserungen und
einiger Verhaltensweisen der Amsel (T Ur-
dus merula merula L.) unter naturlichen
Bedingungen und nach einzelauf zucht in
schalldichten Raumen. Zeits. Tierpsychol.,
13: 341-441.

Moynihan, M.

1955. Remarks on the original sources of dis-
plays. Auk, 72: 240-246.

Nice, M. M.

1943. Studies in the life history of the Song Spar-
row (Part II). Trans. Linn. Soc., N. Y.,
Vol. VI. viii + 329 pp.

Nicholai, J.

1956. Zur Biologie und Ethologie des Gimpels
( Pyrrhula pyrrhula L.). Zeits. Tierpsy-
chol., 13: 93-132.

Nickell, W. P.

1958. Variations in engineering features of the
nests of several species of birds in relation
to nest sites and nesting material. Butler
Univ. Bot. Studies, 13 (2): 121-140.

Portielje, A. F. J.

1925. Zur Ethologie bezw. Psychologie van Pha-
lacrocorax carbo subcormoransis (Brehm).
Ardea, 16: 107-123.

1928. Zur Ethologie bezw. Psychologie der Sil-
bermowe Larus a. argentatus Pontopp.
Ardea, 17: 112-149.

Rinkel, G. L.

1940. Waarnemingen over het gadrag van de
Kievit ( Vanellus vanellus L.) gedurende
de broedtijd. Ardea, 29: 108-147.

Schafer, E.

1957. Les Conotos — Bonner Zoologische Beit-
rage. Sonderheft 5, 148 pp.

58

Zoologica : New York Zoological Society

[47: 6: 1962]

Schuz, E.

1943. Bewegungsnormen des Weissen Storches.
Zeits. Tierpsychol., 5: 1-37.

Selous, E.

1901. An observational diary of the habits—
mostly domestic— of the Great Crested
Grebe ( Podiceps cristatus ) and of the
Peewit ( Vanellus vulgaris ) with some
general remarks. Zoologist, 1901: 161-
183; 339-350; 454-462 (for Great Crest-
ed Grebe).

1902. An observational diary of the habits—
mostly domestic— of the Great Crested
Grebe ( Podiceps cristatus ) and of the
Peewit ( Vanellus vulgaris) with some
general remarks. Zoologist, 6: 133-144
(for Peewit).

Skutch, A.

1954. Life Histories of Central American Birds.
Part I. Families Fringillidae, Thraupidae,
Icteridae, Parulidae, and Coerebidae.
Pacific Coast Avifauna, No. 31. Cooper
Ornithological Society, Berkeley, Calif.
448 pp.

Steiner, H.

1955. Das Brutverhalten der Prachtfinken, Sper-
mestidae, als Ausdruck ihres selbstand-
ingen Familiencharakters. Acta XI Congr.
Int. Ornith., 1954: 350-355.

Tashian, R. E.

1957. Nesting behavior of the Crested Oropen-
dola ( Psarocolius decumanus) in north-
ern Trinidad, B.W.I. Zoologica, 42 (7):
87-97, 2 pis.

Thorpe, W. H.

1956. Learning and instinct in animals. Har-
vard Univ. Press, Cambridge, Mass., 493
pp., illus.

Tinbergen, N.

1931. Zur Paarungsbiologie der Fluseeschwalbe
(Sterna h. hirundo L.). Ardea, 20, 1-18.

1935. Field observations of East Greenland
birds. 1. The behaviour of the Red-necked
Phalarope (Phalaropus lobatus L.) in
Spring. Ardea, 24: 1-42.

1936. Zur Sociologie der Silbermowe (Larus a.
argentatus Pontopp.). Beitr. Fortpfl. Biol.
Vogel, 12: 89-96.

1951. The study of instinct. New York: Oxford
Univ. Press, 228 pp., illus.

1952. Derived activities; their causation, bio-
logical significance, origin, and emancipa-
tion during evolution. Quart. Rev. Biol.,
27: 1-32.

1953a. The Herring Gull’s world. London: Col-
lins New Naturalist Monograph No. 9:
xvi & 255 pp.

1953 b. Specialists in nest building. Country Life,
30: 270-271. Jan. 1953.

1957. The functions of territory. Bird Study, 4
(1): 14-27.

van Dobben, W. H.

1949. Nest-building technique of Icterine War-
bler and Chaffinch. Ardea, 37: 89-97.

Walters, J.

1956. Eiruckgewinnung und Nistplatzorientier-
ung bei See- und Flussregenpfeifer. Lim-
osa, 29 (4): 103-129.

7

Further Observations on the Pilot Whale in Captivity1

David H. Brown

Curator of Mammals, Marineland of the Pacific,
Marineland, California

(Plates I & II)

Introduction

RECORDS on the behavior of a captive
Atlantic pilot whale, Globicephala ma-
croryncha Gray (Kritzler, 1952), and of
captive Pacific pilot whales, Globicephala scam-
rrtonii Cope (Brown, 1960), have shown that
these delphinids generally adapt readily to con-
finement in large aquaria.

At Marineland of the Pacific, Marineland,
California, Pacific pilot whales are exhibited
with striped dolphins, Lagenorhynchus obli-
quidens Gill, in a circular tank 80 feet in diam-
eter and 19 feet deep. Glass windows placed in
three levels of corridors permit observations of
what transpires beneath the surface. Observation
of surface activity can be made from a “top
deck” area, where seating facilities are provided.

Prior to January, 1959, the exhibit consisted
of two female pilot whales and four striped dol-
phins (Brown, 1960). This paper describes the
capture of a male pilot whale and the activity
observed upon its introduction into the tank.
An account of the sickness and death of a fe-
male pilot whale, including findings at necropsy,
is given. The behavior of the male pilot whale
at the time of this death is also described. Also
presented are remarks on laryngeal withdraw-
al in certain odontocetes.

Behavior During Capture

On January 21, 1959, the Marineland col-
lecting vessel, Geronimo, while operating in
the Catalina Channel, California, encountered
a very diffuse school of approximately 40 pilot
whales. Several distinct pods were involved,
some being almost a mile apart. Family groups

Contribution No. 18, Marineland of the Pacific Bio-
logical Laboratory.

were seen in which a male could be observed
swimming with mother and young.

The men maneuvered their vessel close to
what appeared to be a family group, consisting
of a female, a calf and a larger male. The collec-
tor had the opportunity of rapidly snaring the
latter animal, which for the first 35 minutes af-
ter becoming ensnared, towed the Geronimo
around in large circles. The animal was quickly
brought alongside the boat by winching in the
nylon lead rope. No sounds were heard coming
from him during any of these procedures, and
while still attached to the line his dives consumed
about five minutes. The animal had a tendency
to dive and then rest at the end of the dive, leav-
ing the line quite slack. The large size of the
pilot whale made it impossible to lift him aboard,
so, to effect his transport to Marineland, a de-
flated 20-foot rubber life raft was pulled be-
neath the animal, quickly inflated, and thus he
was safely secured during the return trip to port.

The animal was measured and found to be
17 feet 3 inches long; his weight was estimated
to be about 3,000 pounds.

Introduction into the Exhibit

The pilot whale made no movements in the
raft during the return journey or while he was
hoisted into the exhibit tank. Upon being re-
leased he quickly dived and began to swim slow-
ly in a head down position, only inches from
the floor of the tank. After being submerged
for five minutes he surfaced to blow, and then
again resumed the head down position, which he
maintained for several hours, interrupted only
by excursions to the surface to breathe.

At this time the female pilot whales appeared
to show fear and swam rapidly together around
the periphery of the tank, accompanied by the

59

60

Zoologica: New York Zoological Society

[47:7

striped dolphins which also appeared to show
apprehension at the presence of the new arrival.
Several hours later, toward the end of the day,
the female whales commenced to rub their flip-
per tips against the male. The females were also
seen to butt the melons of their heads against
the larger melon of the male.

By the following morning the animals were
quite familiar with one another and in the en-
suing weeks, between feedings, spent their time
swimming (Plate I, Fig. 1) or lying together in
close formation on the surface in what appeared
to be a resting position.

Feeding the New Arrival

Inducing newly arrived cetaceans to accept
nourishment in capitivity is often difficult, ow-
ing to the stresses of capture, unusual confines
of captive environment and the type of food the
animal is expected to consume.

Inanition results rapidly in emaciation with
most cetaceans. Pilot whales, probably as a re-
sult of their less active behavior, seem to with-
stand prolonged fasting with far less weight loss
than the small delphinids (Brown, et al., I960;
Brown & Norris, 1956). The first pilot whale
captured by Marineland in 1957 responded to
forced feeding on its eighth day in captivity
(Brown, 1960) while the other female com-
menced to accept dead squid voluntarily after 14
days of fasting. No major loss of body weight
occurred in either case.

The first attempt to feed the new pilot whale
was made six days after his introduction into the
tank; this was effected each evening by throw-
ing large quantities of squid into the water. The
active feeding behavior demonstrated by the
other animals would, we hoped, initiate a favor-
able response in the new arrival. This technique
proved successful, and on February 1 he began
to feed and thereafter regularly accepted 100
pounds of squid each day.

The rapidity with which Globicephala usually
adapts to captivity has already been discussed
(see Kritzler, 1952, and Brown, 1960). The
male pilot whale also quickly responded to the
attendant’s summons and within four weeks of
his capture learned to lunge half of his body
clear of the water to accept food from his train-
er’s hand.

It is remarkable that what must have been
the animal’s well established behavior pattern
did not offer more resistance to such major en-
vironmental change. It is conceivable that the
presence of the female pilot whales already es-
tablished in the exhibit contributed largely to the
adult animal’s remarkably rapid integration into
the group.

Sexual Behavior

Sexual behavior among the pilot whales in
Marineland of the Pacific has been seen sporad-
ically throughout the spring and winter months.
The first of the following observations was made
only 25 days after the male’s admission into
the tank.

February 15, 8:00 a.m. Both female pilot
whales were observed swimming around and
rubbing against the ventral surface of male’s
body. After several minutes of this activity the
three animals swam almost to the floor of the
tank and commenced to butt their heads to-
gether with great force. The larger female moved
away and began to rub her genital opening on
the floor of the tank. The smaller female and
the male commenced to slowly swim together,
the smaller animal swimming upside down, in a
position immediately under and considerably
forward of the male. In this position she was
seen to arch her body and vigorously rub her
genital opening between the flippers of the larger
animal. At this time the male extruded its penis,
and when the female moved back to a position
where her ventral surface approximated his,
several attempts at intromission were made. De-
spite the female’s apparent sexual receptiveness,
however, penetration was not achieved.

April 1, 9:45 a.m. This morning a loud creak-
ing sound was heard coming from the circular
tank; investigation revealed the large pilot whale
very active, rubbing himself against the females.
The vocalization emanated from the large whale
and was not accompanied by any emission of
air. The sounds produced were similar to the
creaking of a rusty hinge. This activity appeared
to disturb the striped dolphins which were rap-
idly swimming around the tank in close forma-
tion. The glans penis was protruding from the
male’s genital opening at this time.

April 17, 10:00 a.m. The large female was
swimming against and rubbing her genital open-
ing on the ventral surface of the male. The
striped dolphins again appeared very active and
were also seen to rub against the male pilot
whale. The two larger animals then sank to
floor of tank and upon confronting each other
some 30 feet apart swam at each other, butting
the melons of their heads together with great
force. On several occasions the smaller female
through force of impact was thrust backward
through the water several feet; the whip-like
crack produced by these collisions could be
clearly heard outside the tank. After indulging
in this behavior several minutes, the male had
an erection, and although the female drifted pas-
sively in the water and appeared receptive, many
attempts at intromission again proved unsuc-
cessful.

1962]

Brown: Observations on the Pilot Whale in Captivity

61

October 28, 7:30 a.m. This morning the
smaller female was observed rubbing the ven-
tral surface of her body against male’s and trail-
ing her flippers and dorsal fin over the larger
animal’s genital opening. After two or three
minutes of this play, the male extruded about
18 inches of his penis and tried to effect intro-
mission. Penetration occurred briefly, where-
upon the male returned to a resting position on
the surface, and was eventually joined by the
female.

November 7, 11:30 a.m. The male pilot whale
was swimming around tank with his penis ex-
truded about twelve inches. The smaller female
forcibly butted him on his side and then assumed
the position noted on February 15, 1959. Intro-
mission was again attempted.

November 14, 9:30 a.m. There was consid-
erably activity in the whale tank this morning.
Both female pilot whales were rubbing against
male, who maintained a head down position.
The animals were extremely vocal and produced
a bleating or crying sound and a loud buzzing
noise, which had not been heard before. These
vocalizations were not accompanied by any
emission of air from the animal’s blowholes.

At 10:30 a.m. diver G. McLaughlin reported
the male lying on his back at the surface with
his penis extruded about 24 inches.

Other behavior of probable sexual signifi-
cance has been observed many times in both
male and female pilot whales. McBride & Hebb
(1948) recorded similar behavior in male Tur-
siops while Tavolga & Essapian, (1957) de-
scribed unpaired female T ursiops engaged in un-
usual activities of a sexual nature.

The male pilot whale is often observed rub-
bing his genital opening on the floor of the tank
and sometimes on these occasions the animal
utilizes a piece of stone; by means of body move-
ments he attempts to rub this against the wall
of the genital orifice or the glans penis, which
is usually extruded at this time. The females
also employ the tank floor in a similar manner.
The diving ladder has been rubbed by the smaller
whales so violently it has become loosened at
its mountings.

Brown & Norris (1956) record sexual stimu-
lation of male T ursiops by underwater petting.
On one occasion a diver while swimming with
the male pilot whale and patting his sides elicited
an identical response; the male extruded his
penis and moved towards the man.

Tavolga & Essapian (1957) have described
the sexual behavior of the bottlenosed dolphin,
Tursiops truncatus Montagu, and they state
that the dominant role during courtship is as-

sumed by the male. The observations at Marine-
land on interspecific mating activity between
female Lagenorhynchus and male Tursiops
(Brown & Norris, 1956) and more recently be-
tween Globicephala are contrary to those de-
scribed in Tursiops by the above authors. Ob-
servations made at Marineland generally show
the female to be responsible for the initiation
of precopulatory behavior.

It would appear that the wide range of sexual
behavior recorded in male T ursiops by McBride
& Hebb (1948) and Brown & Norris (1956) is
not so apparent in Globicephala, but it must
also be noted that this statement is made after
observing only one captive adult male. It is pos-
sible that younger males demonstrate a broader
field of sexual proclivity.

Observations on the mating activity of Tur-
siops truncatus show that little difficulty is ex-
perienced by these smaller active animals in
effecting penetration. An individual upon pro-
ducing an erection is able to swing the penis
around in a half circle and thus can achieve
intromission with an approach from either
side, and by swinging on its back under the
female is able to contrive intromission by a
powerful vertical flexure of its body.

The male pilot whale appears to lack this
underwater dexterity, and the females seem
to play the major role in precopulatory body
positioning. Before erection, the male’s penis
is enclosed within the genital sheath. When
fully extended it is about 24 inches long and
approximately six inches in diameter at its base.
The shaft tapers to an apex surmounted by a ver-
miform glans some four inches long. Unlike
that of Tursiops the penis is not rigid but flex-
uous, upon extension its contortions giving the
false impression of independent movement.
During the several attempted and one success-
ful intromission, strong pelvic thrusts were seen.

Aggressive Behavior

In a previous paper (Brown, 1960), aggres-
sive behavior in a female pilot whale after
months of solitary confinement was recorded.
After being transferred to a tank with another
whale, these aggressive actions ceased. Except
for one incident, to be discussed later, the male
has not directed aggressive behavior towards
any man, and has allowed divers to grasp his
flippers or dorsal fin as he swam around the
tank. This was stopped after the small female
made attempts to ram and bite the man involved.

Interspecific Relationship and Play

Schools of Globicephala and Lagenorhynchus
are commonly seen swimming and feeding to-

62

Zoologica: New York Zoological Society

[47:7

gether in the Catalina Channel. In February,
1957, during the capture of a pilot whale, sev-
eral striped dolphins accompanied the snared
whale to within a few feet of the hull of the
Geronimo.

Interspecific behavior has also been seen in
the circular tank. Of particular interest is the
association between a female whale and a fe-
male dolphin. This first became apparent during
the spring of 1960 when they spent many days
playing together, their favorite sport being to
catch the air emitted from each other’s blowhole.
A variation of this play took place below the
water-inlet jets where each animal in turn re-
leased air into the current.

Behavior of a possible homosexual nature
has occurred between this pair; the dolphin
nuzzled the genital opening of the whale, which
then did the same to her companion.

On occasion the dolphins “tease” the pilot
whales by biting their flukes. At this writing, the
male is subject to this treatment, and he is seen
to pursue the agile dolphins, who easily evade
his rushes.

The diver’s hose provides a constant source of
interest to the whales, who loop the hose around
their pectoral flippers and rub it with their
bodies. On one occasion a man wearing a Desco
mask was pulled around the tank for several
minutes by the male, who seized the airline in
his mouth.

Sickness and Death of a Pilot Whale

The large female pilot whale had lived at
Marineland since January, 1957. During the
initial months of capitivity she swallowed a rub-
ber innertube, but after oral administration of
mineral oil, the tube was regurgitated (Brown,
1960) . Following this incident, the animal main-
tained good health until inception of the fatal
sickness now to be described.

On March 2, 1960, symptoms of a gastroen-
teric disorder became apparent and vomiting oc-
curred regularly after feeding. This made oral
medication difficult. However, some mineral oil
given in squid was retained. The animal showed
no other abnormality and continued swimming
activity around the tank.

March 3. The whale’s condition not im-
proved, emesis occurs after eating
even small quantities of food. At-
tempts at medication unsuccessful.
Mineral oil and kaopectate regurgi-
tated almost immediately upon ad-
ministration.

March 4. Still avidly accepts food, all of
which is lost shortly after ingestion.
Today quite inactive between at-
tempted feedings and lies on surface
with eyes closed.

4:30 p.m. The antispasmodic,
methyl - atropine - nitrate (Metro-
pine), given in food and apparently
retained. An additional feeding at
5:00 p.m. induced vomition. Metro-
pine again given in food at 12:00
a.m.; air and cloudy fluid expelled
at 12:15 a.m.

March 5. 9:00 a.m. This morning general
condition seems improved, animal
active and swimming with other
whales. At 9:30 a.m., Metropine
given in food; some loss of fluid
and discharge of air seen shortly
after feeding. Animal retained all
of food given during remainder of
day. Apart from oral air and fluid
loss, behavior quite normal.

March 6. Animal vomited during the night;

considerable quantities of predi-
gested food found floating in center
of tank. Refused to swim to the
feeding platform and spent the day
lying on the surface with eyes
closed, occasionally sounding to the
floor of the tank.

March 7. The whale’s condition remains un-
changed. At 9:45 a.m. vomits
cloudy fluid, and at 12:15 a.m. seen
slowly swimming with male, eyes
closed; At this time vocalized sev-
eral times, making high pitched
squeals which were accompanied
by air emissions from blowhole.

At 2:00 p.m. returned to surface
and stayed in same position for re-
mainder of the day.

March 8. 4:00 a.m. Female lying with other
whales at the surface of the tank,
respiration rapid and shallow. This
was the last observation made prior
to death.

Behavior of the Male Pilot Whale

At 5:30 a.m. on March 8, the male pilot
whale was first seen transporting the dead fe-
male to and from the surface by grasping her
flipper in his mouth (Plate 1, Fig. 2). He carried
her in this fashion for the next five hours. An
erection occurred at 8:15 a.m., and on several

1962]

Brown: Observations on the Pilot Whale in Captivity

63

occasions he effected intromission with the dead
animal. A diver entered the tank at 8:45 a.m.
and endeavored to pass a rope around the fe-
male’s tail. On seeing this, the male momen-
tarily released his burden and tried to strike the
man with his head. To entice the male to leave
the dead animal, squid was thrown into the
water, but he ignored this, seized the female
and continued to carry her around the tank.
Another diver entered the water, and the whale
became more active. Swimming at great speed
around the tank, he emitted bleating sounds and
recommenced copulation with the dead female.
The other female at this time began to vocalize
and tried to push a diver with her head.

While endeavoring to evade the divers, the
big male dropped the dead animal, which lodged
between the rocks cemented to the floor of the
tank. He then seized the caudal peduncle in
his mouth and again succeeded in lifting the
body to the surface, where he changed his hold
firstly to the dorsal fin and then the flipper. In
recovering the body he displaced several rocks;
one large boulder weighing at least 700 pounds
was propelled completely across the bottom of
the tank. At 9:45 a.m. the diver finally suc-
ceeded in passing a line around the dead female’s
tail. The male then offered little resistance to
her removal from the tank, and shortly there-
after accepted food in a normal way.

Our collectors have seen behavior of a similar
nature in wild pilot whales, and report two cases
of female whales supporting dead young. Sev-
eral authors (see McBride &Hebb, 1948, Moore,
1953, and Hubbs 1953) have also recorded such
actions in Tursiops.

While these mother-young relationships are
not beyond interpretation, the behavior of the
male pilot whale is difficult to translate. The
complex social behavior of this species and the
environmental restrictions of captivity without
doubt furnished behavioral stimuli at this time.

Necropsy

The dead female pilot whale had increased
23 inches in total length during the three years
and two months she had been in captivity and
weighed 1,360 pounds. Postmortem examina-
tion was performed on the day of death by vet-
erinary pathologists of the Los Angeles County
Livestock Department. Necropsy showed chron-
ic enteritis. Submucosal hemorrhages with pete-
chiae up to 3 mm. in size were found throughout
the entire intestinal tract (Plate II, Fig. 3). Fur-
ther investigation however, revealed laryngeal
occlusion as the primary cause of death.

The following is an excerpt from the path-

ologist’s report: “When the whale’s head was
totally severed from the body, a piece of rock
fell to the floor, its origin unknown at that time.
I approached the severed head from its posterior
aspect and proceeded to dissect the remaining
part of the trachea, larynx and esophagus. I was
able to enter my hand into the pharyngeal area.
By grasping the elongated epiglottis and aryte-
noid structures, then depressing and retracting,
I was able to force this anterior part of the lar-
ynx out of the pharyngeal chamber. Further
traction on the larynx with one hand and in-
cision of the ventral attaching structure with
the other hand resulted in my freeing the afore-
mentioned structures.

“Dissection and examination of these struc-
tures revealed a well defined localized lesion of
the mucous membrane on the floor and partly
on the side walls of the larynx in the region of
the thyroid cartilage.

“This lesion was approximately 5 cm. long,
extended up each side wall about 3 cm. at its
longest points. It was irregular in form but well
demarcated. The pathological change of the mu-
cous membrane was that of passive congestion,
and edema not to the point of macroscopic
necrosis. It was evident this lesion was due to
pressure of the rock. Thorough examination of
the esophagus, pharynx and trachea revealed
no gross lesions.”

Discussion

During the months preceding the death of the
large female, the whales were often seen mouth-
ing fragments of stone separated from the reefs
on the floor of the tank. The female pilot whale
seemed particularly prone to this activity and
was frequently seen to ingest, and then expel,
the stone involved.

The fatal consequences of this behavior were
revealed by the pathology just described, and
it is evident that laryngeal occlusion induced by
inspiration of a stone was the primary cause of
death. It is assumed that the enteric lesions were
secondary in this case.

The etiologic implications are apparent. It is
quite plain that the stone gained access through
the glottis into the aryteno-epiglottid tube, after
entry into the palatopharyngeal region via the
mouth. Oral entry was supported by subsequent
investigation, which showed the stone too large
to pass the bony nares of the skull (Plate II,
Fig. 4). It is not illogical to assume that vomi-
tion occurred after the stone had been swal-
lowed; during emesis, and involuntary retraction
of the aryteno-epiglottid tube, the stone passed
into the dilated sphincter of the naso-pharynx.
Reinsertion of the tube and respiratory action

64

Zoologica: New York Zoological Society

[47:7: 1962]

caused its migration into the laryngeal struc-
ture.

Since the death of the pilot whale, laryngeal
retraction has been recorded in a male bottle-
nose dolphin. This was observed while treating
a lesion in the commisure of the animal’s mouth.
While his jaws were being forcibly held apart,
he convulsed and vomited. When the animal’s
jaws were released, he violently exhaled and
expelled vomitus through his blowhole.

It is evident that the opening of the mouth
withdrew the lamyx from the nasal cavity; the
enforced withdrawal prevented respiration, in-
ducing convulsion and regurgitation. Vomitus
then entered the vacated nasopharynx. Upon
allowing the mouth to close, the aryteno-epi-
glottid tube reentered the narial cavity. The
powerful exhalation cleared the nasal passage.
It would then appear that on full expansion of
the jaws the tube is retracted by action of the
throat muscles connected to the hyoid bone.

The ability to disengage the larynx from the
narial cavity by full expansion of the jaws is
perhaps possessed by all odontocete whales, and
possibly facilitates the passage of large items of
food through the esophagus.

Acknowledgements

The author wishes to express his sincere ap-
preciation to Dr. Rankin W. McIntyre and
Dr. Maurice Barenfus for their skillful post-
mortem examination of the dead pilot whale,
and also to Mrs. Muriel Johnson for her help
in preparing the manuscript. He also wishes to
thank Captain Frank Brocato and Frank Calan-

EXPLANATION

Plate I

Fig. 1. Male pilot whale accompanied by females
in the circular tank, Marineland of the
Pacific. (Photograph by Peter Stackpole).

Fig. 2. Male pilot whale carrying dead female in
circular tank, Marineland of the Pacific.
(Photograph by Robert Vanderhoof).

drino of the Geronimo for observations made

during the capture of the male pilot whale.

References

Brown, D. H.

1960. Behavior of a captive Pacific pilot whale.
Jour. Mamm., 41: 342-349.

Brown, D. H., R. W. McIntyre, C. A. Delliquadri

& R. J. Schroeder

1960. Health problems of captive dolphins and
seals. Jour. Amer. Vet. Med. Assoc., 137
(9): 534-538.

Brown, D. H. & K. S. Norris

1956. Observations of captive and wild ceta-
ceans. Jour. Mamm., 37: 311-326.

Hubbs, C. L.

1953. Dolphin protecting dead young. Jour.
Mamm., 34: 498.

Kritzler, H.

1952. Observations on the pilot whale in cap-
tivity. Jour. Mamm., 33: 321-334.

McBride, A. F. & D. O. Hebb

1948. Behavior of the captive bottle-nosed dol-
phin Tursiops truncatus. Jour. Comp, and
Physiol. Psych., 41: 111-123.

Moore, J. C.

1953. Distribution of marine mammals to Flor-
ida waters. Amer. Mid. Nat., 49: 117-158.

Tavolga, M. C., & F. S. Essapian

1957. The behavior of the bottle-nosed dolphin
( Tursiops truncatus ): mating, pregnancy,
parturition and mother-infant behavior.
Zoologica, 42 (2): 11-31.

OF THE PLATES

Plate II

Fig. 3. Opened lower intestine of pilot whale,
showing enteritis and submucosal hemor-
rhage. (Photograph by J. Courtland
Beazie).

Fig. 4. Skull of pilot whale demonstrating the
size of stone in relation to the bony nares.
(Photograph by J. Courtland Beazie).

BROWN

PLATE i

FIG. 1

FIG. 2

FURTHER OBSERVATIONS ON THE PILOT WHALE IN CAPTIVITY

BROWN

PLATE II

FURTHER OBSERVATIONS ON THE PILOT WHALE IN CAPTIVITY

,V.V

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

630 Fifth Avenue, New York 20, N. Y.

PUBLICATION OFFICE

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

OFFICERS

PRESIDENT VICE-PRESIDENT
Fairfield Osborn Laurance S. Rockefeller

SCIENTIFIC STAFF:

John Tee-Van General Director

William G. Conway. . Director,

Zoological Park

Christopher W. Coates. .Director, Aquarium

ZOOLOGICAL PARK
Joseph A. Davis, Jr. . . Associate Curator,
Mammals

Grace Davall ...... Assistant Curator,

Mammals and Birds
William G. Conway . . Curator, Birds
Herndon G. Dowling . Curator, Reptiles
Charles P. Gandal. . . Veterinarian
Lee S. Crandall. .... General Curator,

Emeritus

William Beebe Honorary Curator,

Birds

AQUARIUM

James W. Atz Curator

Carleton Ray Associate Curator

Ross F. Nigrelli Pathologist & Chair-

man of Department
of Marine Biochem-
istry & Ecology

C. M. Breder, Jr. Research Associate

in Ichthyology

Harry A. Charipper . . . Research Associate

in Histology

Sophie Jakowska Research Associate

in Experimental
Biology

SECRETARY TREASURER

George W. Merck David H. McAlpin

Klaus D. Kallman. . . . Research Associate

in Genetics

Louis Mowbray ..... Research Associate

in Field Biology

GENERAL

William Bridges . , Editor & Curator,

Publications

Dorothy Reville . . Editorial Assistant

Sam Dunton ..... Photographer

Henry M. Lester. .Photographic Consultant

DEPARTMENT OF TROPICAL
RESEARCH

William Beebe Director Emeritus

Jocelyn Crane Assistant Director

John Tee-Van Associate

William K. Gregory Associate

AFFILIATE

L. Floyd Clarke. Director,

Jackson Hole Biological Research Station

EDITORIAL COMMITTEE

Fairfield Osborn, Chairman
James W. Atz William G. Conway

William Beebe Lee S. Crandall

William Bridges Herndon G. Dowling

Christopher W. Coates John Tee-Van

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 47 • PART 2 • SEPTEMBER 15, 1962 • NUMBERS 8-10

PUBLISHED BY THE SOCIETY
The ZOOLOGICAL PARK, New York

Contents

PAGE

8. A Field Study of the Black and White Manakin, Manacus manacus, in

Trinidad. By D. W. Snow. Text-figures 1-21 65

9. Longevity of Fishes in Captivity, as of September, 1956. By Sam Hinton. 105

10. Hybridization Experiments in Acheilognathine Fishes (Cyprinidae.
Teleostei). A Comparison of the Intergeneric Hybrids between Tanakia
tanago and Rhodeus spinalis and Rhodeus ocellatus from Korea and Japan.

By J. J. Duyvene de Wit. Plates I & II; Text-figure 1 117

8

A Field Study of the Black and White Manakin,
Manacus manacus, in Trinidad1,2

D. W. Snow

Department of Tropical Research,

New York Zoological Society, New York 60, N. Y.

(Text-figures 1-21)

Introduction

The Black and White Manakin ( Man-
acus manacus ) is one of the commonest
forest birds in Trinidad. It is also one of
the easiest to study in the field, and it was this
fact, combined with the outstanding interest of
its communal courtship displays, that led me to
devote much of my time to it during AV2 years’
residence in the Arima Valley, the site of the
New York Zoological Society’s Tropical Field
Station in the center of the Northern Range of
Trinidad (Text-fig. 1).

It is a small, stockily built bird, with short
wings and relatively long, strong legs (Text-fig.
2). The sexes differ markedly, the male being
black and white and the female olive-green; the
legs are orange in both sexes. Though it occurs
in forest of all kinds, it probably reaches its
greatest abundance in secondary forest. Like
other manakins, it subsists largely on fruit which
it takes in flight. It bathes in shallow streams,
and drinks occasionally from streams but usu-
ally from water collected in the leaves and bracts
of plants. Though it will feed on occasion in the
tops of lofty trees, most of its life is spent within
25 feet of the ground. In particular, it nests
only a few feet above the ground, and displays
on and within a few inches of the forest floor.
It is communal in its courtship; each male dis-
plays at a cleared “court” on the forest floor
within a few yards, or even a few feet, of his
neighbors. Groups of courts constitute “display
grounds,” which are used year after year. The

1 Contribution No. 1015, Department of Tropical
Research, New York Zoological Society.

2 This study has been supported by National Science
Foundation Grants G4385 and G21007.

females visit the display grounds to mate. No
pairs are formed, and the male takes no part in
the nesting.

One other member of the family occurs in
Trinidad, the Golden-headed Manakin ( Pipra
erythrocephala) . Both occur in the same habi-
tats and have largely the same feeding and nest-
ing habits. The Golden-headed Manakin, how-
ever, displays 25 feet or more up in the trees,
and also tends to feed and nest higher than the
Black and White Manakin, so that, although it
is even more abundant, it is less suitable for de-
tailed study. A shorter account of its biology
will be presented in a later publication.

There have been some previous accounts of
the biology of Manacus. Chapman’s pioneer-
ing study of the courtship of Gould’s Manakin
(M. vitellinus ) in Panama was the first detailed
study of any of the Pipridae (Chapman, 1935).
Although there had been some earlier accounts,
it was this paper which first drew attention to
the extraordinary development of courtship be-
havior in the family. Working on specimens
sent to him by Chapman, Lowe (1942) de-
scribed the musculature of the wing-feathers
and other specializations responsible for the
loud mechanical noises made by the males dur-
ing their displays. In Trinidad, Chapman
(1894) had earlier made some brief observa-
tions on the display of M. manacus, and sub-
sequent writers have occasionally referred to the
males’ courtship gatherings. A preliminary ac-
count of the present study, dealing only with
display, has already been published (Snow,
1956). Darnton (1958) has also published some
observations made at a Trinidad display ground
of M. manacus. Finally, Sick (1959) has dealt
with Manacus among other genera in his review
of displays in the Pipridae as a whole.

65

66

Zoologica: New York Zoological Society

[47: 8

Text-fig. 1. Trinidad, showing the study area (S)
and the annual isohyets. The Northern Range ex-
tends along the whole of the northern side of the
island, the two highest points being shown (Aripo
and El Tucuche).

It may be noted that, although M. manacus
and M. vitellinus have usually been treated as
separate species, there are no important differ-
ences in behavior between them (Appendix 1)
and morphologically they differ only slightly,
M. vitellinus having the plumage suffused with
orange-yellow which is lacking in M. manacus.
They would thus reasonably be considered as
conspecific, but for evidence that they are sym-
patric in Colombia (de Schauensee, 1950).

The present study is based on observations
continued over years. Display was watched,
and filmed, from hides at display grounds, the
most complete observations being made at one
display ground that was visited regularly over
the whole period and weekly, except for a few
gaps, from June, 1958, to September, 1961.
Trapping was an important part of the work; a
total of 271 individuals were caught in mist-nets
in the study area, many of them repeatedly.
Some 150 more were trapped in other parts of
Trinidad, mainly in connection with the work
of the Trinidad Regional Virus Laboratory.
Each bird trapped in the study area was given a
different color combination, and many of them
were seen subsequently in the field. Many of the
males occupying courts at the display grounds
under observation were eventually color-ringed.
Trapped birds were also examined for moult and
other details of plumage, and were weighed.
Breeding was studied by systematically search-
ing for nests along certain stretches of forest
stream, and recording their fate. Searching was
continued in all months of the year, not only at
seasons when manakins were known to be nest-
ing. Twenty-eight birds were ringed as nestlings.

Some observations on behavior at the nest were
made from a hide. Food was studied by direct
observation of birds feeding, and by collecting
the regurgitated remains of food from display
grounds and from below nests.

Throughout the work, attention was directed
to certain problems of general biological interest
concerning the ecology of birds in tropical for-
est. These include; numbers, reproductive rate
and the control of the population; the food sup-
ply throughout the year; the breeding season
and the factors controlling it; the function of
communal displays.

I am much indebted to my wife for help in
the field work, especially the trapping and the
finding and inspection of nests; to Dr. William
Beebe and Miss Jocelyn Crane for placing at my
disposal all the facilities of the New York Zoo-
logical Society’s Tropical Field Station; to Dr.
W. G. Downs and Dr. T. H. G. Aitken for the
opportunity to examine birds trapped during the
field studies of the Trinidad Regional Virus
Laboratory; and to Mr. N. Y. Sandwith, Mr.
N. W. Simmonds and Dr. J. J. Wurdack for many
plant identifications. My thanks are due also to
Dr. N. P. Ashmole, Mr. R. E. Moreau and Dr.
D. Lack for valuable criticisms of the paper in
draft. The whole work was generously sup-
ported by the National Science Foundation.

The Environment

The lower part of the Arima Valley at 400-
1,800 feet above sea level, the site of the study
area (Text-fig. 3), has a natural vegetation
transitional between lower montane rain forest
and lowland seasonal forest (Beard, 1946). As
most of the area is government forest, the
greater part of this natural vegetation remains
little altered except for a limited amount of lum-

Text-fig. 2. Male Black and White Manakin; rest-
ing posture.

1962]

Snow: Field Study of the Black and White Manakin

67

Text.-fig. 3. The study
area. Manakin display
grounds are shown as cir-
cles and are lettered. Dot-
ted line: boundary be-
tween St. Patrick’s Estate
and Government Forest.
Broken lines: pathways
and roads. Hatched area:
unforested country.

bering, which has reduced the number of large
trees of the economically important species (es-
pecially Cordia alliodora, Cedrela mexicana and
Carapa guianensis) . The forest is rich in tree
species, many of which bear berries of suitable
size for manakins to eat. The country is hilly and
steep, the underlying rocks being rather soft
schists.

The main Arima River flows with a gently
winding course down the valley and is joined by
numerous steep-sided tributaries, which in turn
are joined by numerous deep side-gullies. The
whole area is thus divided and subdivided into
numerous ridges. The main river and the larger
side-streams have water all the year round, while
the gullies fill up only after heavy rain. Along the
side-streams and in the gullies the dampest con-
ditions are found; the vegetation includes tree-
ferns and many smaller ferns, Carludovica,
Cyclanthus bipartitus and species of Heliconia.
At the other extreme, up on the well-drained
ridges there is a thinner forest of trees tolerant
of desiccation, among which the Yellow Poui
(Tabebuia serratifolia) is prominent during the
dry season, when it loses its leaves and produces
its yellow blossoms in two or three bursts of
flowering. In the dry season there is a marked
contrast between the ridges, where the forest
includes an important deciduous element, and
the lusher vegetation along the stream bottoms.

The main road up the valley follows the river.
As a consequence, the forest along the river has
been much opened up and the natural vegetation
modified. In particular there are great clumps
of an introduced bamboo.

St. Patrick’s Estate, where the two main dis-
play grounds that were studied were situated and

all the trapping was carried out, is an area of
secondary forest adjoining the government
forest. Some large trees remain, but mostly the
trees are in various stages of growth, up to a
height of about 60 feet. This secondary forest is
even richer in berry-bearing trees and shrubs
than the primary forest. Melastomaceae are es-
pecially abundant, being represented by twelve
common tree species and several shrubs. Among
the lower-story trees and shrubs, in both primary
and secondary forest, the berry-bearing Rubi-
aceae are important and are represented by
many species.

Along the Northern Range the rainfall in-
creases steadily from west to east. Port-of-Spain,
at the foot of the hills at the western end, has an
average annual precipitation of 55 inches, while
at the eastern end the average is well over 100
inches (Text-fig. 1). The Arima Valley, half
way along, has an intermediate rainfall, but
closer to that of the eastern end than to the
western. The average precipitation at Verdant
Vale Estate, adjoining St. Patrick’s, for the six
years 1935-40 was 108.1 inches. The four years
1957-1960 were considerably drier, the average
being only 89 inches (Text-fig. 4). Probably a
long-term average would be around 100 inches
at an altitude of 500 feet above sea level.

There is one main dry and one main wet sea-
son. The dry season begins in January and ends
variably, usually in May. The wet season lasts
for the rest of the year, but is usually broken by
a spell of dry weather in September or October
(the “petit careme”) . Monthly totals of only one
or two inches, sometimes less, are recorded in
February and March; in the wet season totals of
15 inches or more are common. There is, how-

68

Zoologica: New York Zoological Society

[47: 8

MONTHS

Text-fig. 4. Mean monthly rainfall, 1957-1960.

ever, considerable variation in the rainfall from
year to year, especially in the length and sever-
ity of the dry season. At all times of the year
there is much sunshine.

Shade temperatures rarely rise above 88° F.
and on wet and cloudy days they remain in the
seventies. Nightly minima are usually between
65° and 75°, falling a few degrees lower on the
coldest nights of the months December-March.
Mean temperatures rise steadily from December
to May, remain steady from May to September,
then fall more rapidly from October to Decem-
ber (Text-fig. 5). Relative humidity is very high
at nights, being always close to saturation; the
day minimum varies much with location, season
and weather. In two years’ records an absolute
minimum of 43% was recorded in March at
Simla, half a mile from the study area (Beebe,
1952).

MONTHS

Text-fig. 5. Mean monthly temperatures. St. Pat-
rick’s Estate. Upper line: mean daily maxima.
Lower line: mean daily minima.

The seasons affect the appearance of the
forest strikingly. Most of the trees have well-
defined flowering and fruiting seasons, and all
the trees of any one species are usually well syn-
chronized. Thus when the Yellow Poui flowers,
towards the end of the dry season, the forest
for a few days is dotted with patches of brilliant
yellow. A little later another common tree, the
White Olivier ( Terminalia obovata), loses its
leaves and comes into flower, and pale yellow

patches appear all over the forest. In general,
there is most flowering and most loss of leaf
during the dry season, and most renewal of leaf
during the wet season. In prolonged dry spells,
especially towards the end of the dry season, the
ground becomes parched and hard, the dead
leaves on the forest floor crackle underfoot, and
the leaves of the shrubs and smaller trees be-
gin to wilt. The first heavy rain after such spells
has a greatly stimulating effect on all plant and
animal life, and makes the sharpest seasonal
change of the whole year.

Distribution, Numbers and
General Ecology

No detailed survey was made of the distribu-
tion of the Black and White Manakin in Trini-
dad, but visits to many parts of the island
showed that it occurs generally in all kinds of
forest, primary and secondary, from sea level
to about 2,000 feet, except for the dry forests
at the western end of the Northern Range to the
west of El Tucuche (Text-fig. 1). It is especially
abundant in the forests of the central and eastern
parts of the Northern Range, at moderate alti-
tudes, and in the adjacent low-lying forests.
This is the sector of highest rainfall, up to 100
inches or more. Above 2,000 feet, where the
rainfall is considerably higher and the forest
changes to the montane type, the Black and
White Manakin becomes rather rare. The two
highest display grounds found were on Aripo
at about 2,000 feet and on El Tucuche at about
2,200 feet. Very few birds were seen on walks
through the mountain forests above 2,000 feet.

Where manakins are common, display grounds
are scattered at intervals through the forest. In
the hilly country of the Northern Range they are
mainly situated on ridges, either on the crest or
a little way down one of the sides. Those in the
study area (Text-fig. 3) were situated on minor
side-ridges not more than two or three hundred
feet above a stream bed. In low-lying country,
however, display grounds are common in per-
fectly level forest. The preference for a situation
on a ridge, where it is available, may be due to
the commanding position of such a display
ground over the streams below, along which the
females prefer to nest.

More critical than the position of the display
ground with respect to ridge or stream bed is
the structure of its lower vegetation. The Black
and White Manakin depends for its display on
small upright saplings up to about 20 mm. in
diameter. As will be seen later, two or more of
these must be available within a few feet of each
other to allow a bird to display fully. Further, a
communal display ground cannot be established

1962]

Snow: Field Study of the Black and White Manakin

69

unless there are many such groups of saplings
within a few yards of each other.

The same display grounds seem to be used as
long as the vegetation remains unchanged. All
those that were under observation persisted for
up to AV2 years, and local information indi-
cated that they were already old. During a visit
to Panama in 1958 the display ground of M.
vitellinus studied by Chapman in 1932 was
found to be active exactly in the position which
he described.

In the course of the AVi years’ observations at
the main display ground that was studied, the
same individual saplings continued to be used,
and though they were not measured when ob-
servations began they showed no apparent
growth. In this, as in other respects, the tropical
forest is an extremely constant, unchanging
habitat. Only a few courts were abandoned, due
to the fall of a mass of vines or a small tree, and
some new ones were cleared, but these new ones
were mainly peripheral and they did not all last
long. The central nucleus of courts remained
almost the same.

Text-figure 3 shows the distribution of dis-
play grounds in the area of forest, comprising
some 450 acres, that was most thoroughly cov-
ered. It is unlikely that any large display grounds
were not found, but small, intermittently-used
display grounds (the “practice display grounds,”
p. 80) may easily have been missed. From the
number of courts at these main display grounds
it is possible to assess the number of adults in the
area.

There were approximately 205 courts at the
seven main display grounds known in the area.
At the display ground studied in most detail (A
in Text-fig. 3) it was found that in addition to
the established adult males there were about a
quarter as many unestablished adults. Thus 205
courts probably corresponds to some 250 adult
males. Trapping showed almost exact equality
between the sexes (p. 97). Thus there were
probably about 500 adult birds in the 450 acres
of forest.

There was no evidence for any considerable
change in numbers in the course of the study.
The main display ground under observation usu-
ally had from 24 to 28 active courts over the
four years, and another that was observed less
closely (B in Text-fig. 3) had from 30 to 33
over three years.

Although individuals are in general seden-
tary, there was nevertheless too much local move-
ment for the trapping figures to be used for an
independent assessment of the population. Espe-
cially during the moult there must be some indi-
vidual wandering, as the proportion of birds

ringed then but never retrapped nor seen again
was higher than at other seasons (47% for the
months of moult, July-October, compared with
19% for the other months). For feeding, too,
birds may move several hundred yards. Two
males trapped at the main trapping place (T in
Text-fig. 3), one of them as a juvenile, were later
found established at display ground E, nearly
half a mile away, and three females were found
nesting about half a mile from where they were
trapped. All these were probably feeding at the
time that they were trapped. None was found at
any greater distance from the trapping place.

In the breeding season, which lasts for sev-
eral months, males and females are differently
distributed in the forest for the greater part of
the day. The males keep to their display grounds,
up on the ridges, while the females hold small
territories along the streams and gullies by which
they mostly nest. The sexes meet when feeding
and bathing, and also, for short periods, at the
display grounds.

When feeding and bathing, males from the
two display grounds A and B kept mainly to
separate areas (Text-fig. 6). Thus at the main
trapping place (T) many males from display
ground B were trapped while feeding, most of
them repeatedly, but only one male from dis-
play ground A was ever trapped there, once
only. No males from display ground A were
ever seen at the two bathing places nearest to
T, but males from display ground B were seen
bathing at them 13 times. At the bathing place
nearest display ground A, on the other hand,
there were 13 records of males from A bathing,
and only 8 of males from display ground B.

Though no figures are available for compari-
son, there is little doubt that the density of the
manakin population in the forests of the Arima
Valley, and probably over much of Trinidad, is
unusually high. Manacus was found to be much
more sparsely distributed in various forest habi-
tats in British Guiana, Surinam and Panama,
and the published accounts from other areas are
in agreement. Two factors are probably re-
sponsible for the high Trinidad population.
First, the forest in Trinidad is much broken up,
and there is a high proportion of second growth.
Clearings, paths and roads are numerous. Where
lumbering is carried out, the trees that are re-
moved are mostly species whose fruits are un-
suitable for manakins to eat (especially Cordia
alliodora, Cedrela mexicana, Carapa guianensis,
Sterculia caribaea and Mora excelsa) . Many of
the manakins’ main food trees are characteristic
of secondary forest and of road-edges, being
relatively rare in untouched forest; this is par-
ticularly true of the most important family of

70

Zoologica: New York Zoological Society

[47: 8

Text-fig. 6. Foraging and bathing ranges of 9 males from display ground A (left) and 13 males from dis-
play ground B (right). Figures in squares: numbers of observations at bathing and trapping places. Dots:
sight records of feeding birds.

all, the Melastomaceae, which flourish in second
growth in areas of high rainfall and reach a
great abundance in the wetter parts of Trinidad.
Thus a moderate amount of clearing and lum-
bering benefits the manakin population. Sec-
ondly, with a reduced avifauna, compared with
neighboring parts of Venezuela, Trinidad shows
to some extent the phenomenon characteristic
of islands: few species, but large populations of
those species that are present. This must be
attributed to reduced competition from other
species, but the way in which this reduction of
competition operates is far from clear.

DISPLAY

First Impression of a Display Ground

It is difficult to describe the activities at a dis-
play ground of Black and White Manakins in
terms suitable for a scientific journal and at the
same time give any adequate impression of the
total effect of the bizarre postures and move-
ments, the sharp snaps and cracks, and the ex-
traordinary vivacity of the whole performance.
Yet undoubtedly this combined effect is of para-
mount importance, for communal display would
not have evolved if the effect of a group of males
displaying at close quarters had not, for the fe-
male manakin, greater attractive power than that
of a single male displaying by itself.

The first impression is bewildering. Spaced
only a few feet apart (if the courts are well con-
centrated), a group of small black and white
birds are seen leaping about and performing oth-
er evolutions with extraordinary rapidity, within
a few inches of the ground. Accompanying these
movements, some of which are too rapid to fol-

low in detail by eye, are a variety of sharp cracks,
like percussion caps exploding, rolling snaps,
and whirring and grunting noises, as well as a
chorus of excited high-pitched calls.

Closer inspection shows that each bird is per-
forming on and around a small area of bare soil
and rootlets, which shows up plainly against
the surrounding leaf litter, and that two or three
saplings around the edges of each cleared area
are the chief perches of the displaying birds.
More prolonged observation shows that the
birds’ bewildering evolutions can be resolved
into a number of highly stereotyped movements.
Surprisingly, the whole performance may be tak-
ing place, and may continue to do so, without
the presence of a single female. But if a female
does appear, the intensity of display will at once
increase.

Anatomical Specializations of the Male
and the Mechanical Sounds

Lowe (1942) made a detailed examination of
the plumage and musculature of Manacus vitel-
linus and described the structures responsible
for its specialized displays. In its morphology
M. manacus seems to be identical with M. vitel-
linus, and its displays are almost identical. Here
only a short summary will be given of the ana-
tomical findings. It is still not certain how all the
mechanical sounds are made. As Sick (1959)
remarks, full elucidation will depend on experi-
ments on birds displaying in aviaries.

Both the primaries and secondaries are modi-
fied in the male. The four outer primaries are
very narrow and stiff, the outer webs especially

1962]

Snow: Field Study of the Black and White Manakin

71

being extremely narrow. Due to this, the male
makes a grasshopper-like whirring in level flight
which is automatic and not a part of any display.
Females and young males, which have unspe-
cialized primaries, make only a low whirr when
they fly, and moulting males that have lost some
of their outer primaries also fly comparatively
silently. However, the attenuated outer prima-
ries are probably also responsible for the sound
made during “fanning,” a display to be described
later.

The secondaries are more highly modified.
Their shafts are unusually thick, and the outer
webs very stiff. In addition, the ends of the quills
are not attached to the ulna but pass dorsally
to it and are loosely attached to a tendon that
runs along the dorsal side of the radius. The
muscle slips attached to the bases of the secon-
daries are very highly developed. The secon-
daries thus have great mobility.

These modifications of the secondaries are un-
doubtedly responsible for the two loudest me-
chanical noises made in display, the single
“snap” and the “rolled snap.” A single “snap”
is made each time the displaying bird jumps
from one perch to another; it is also made, much
less often, by perched birds. In either case the
wing movement is much too rapid for the details
to be seen. The “rolled snap” is also made when
perched or just before taking off for a leap or
short flight. As it is made the wings can momen-
tarily be seen raised above the back and vibrat-
ing rapidly. Both in the single and the rolled
snap the actual sound is probably made by the
stiff outer vanes of the secondaries brushing
against one another rapidly as the wing is
opened and closed, partly through the action of
their muscles and partly due to the nature of
their attachment along the radius.

More general muscular modifications are also
important for the display. In particular there
is a great development of the pectoral muscula-
ture (responsible for the very rapid wing-actions
needed for “snapping”), the thigh musculature
(responsible for the powerful leaps from perch
to perch) , and the muscles moving the tail (con-
nected with the rapid turning about in mid-air
during jumps between perches).

The appearance of the head is much altered
during display. Most striking are the elongated
feathers of the throat, which are puffed out to
form a white “beard” protruding beyond the tip
of the beak. Connected with this, Lowe reported
a noticeable development of the unstriated mus-
cles activating these feathers, and also of the
muscles of the hyoid apparatus. The neck feath-
ers, and the feathers of the sides of the head,
are also greatly puffed out, so that the black cap

is transformed from a broad oval to a narrow
slip of black tapering to a point at the posterior
end. Seen from the front the bird appears all
white with a small black lozenge in the middle.
Probably the neck feathers too are provided
with a specialized musculature.

Voice

Black and White Manakins utter a variety of
rather simple, mainly monosyllabic, calls. Un-
doubtedly the very loud mechanical snaps have
functionally replaced the loud and more com-
plex calls of some other manakins and the re-
lated Cotingidae. Nevertheless some of the calls
are quite loud, and they make up an important
part of the total volume of sound heard at an
active display ground.

The only call of the adult that is unconnected
with display and uttered by both sexes is a rather
plaintive monosyllabic “peerr,” slightly trilled
at the end. It is uttered by alarmed birds, such
as females whose nests have been approached or
males disturbed at a display ground, by solitary
birds away from display grounds or nests, and
by males sitting quietly at the display ground
between bouts of display. In aggressive encount-
ers between males, it is uttered by the subordi-
nate bird.

As the excitement of a male mounts, his calls
change from the unexcited “peerr” to a louder,
higher-pitched and untrilled “chwee.” This is the
dominant vocal sound made during the out-
bursts of calling and snapping which greet the
arrival of a female at a display ground.

The “peerr” call has other variants that sound
like “pee-you” or “pee-yuk,” being more disyl-
labic, with the trilled ending modified into a
more distinct separate note. Chapman described
these as “notes of awareness, address and re-
sponse, or of inquiry or protest”; certainly they
have a varied social significance that cannot be
simply defined. They are uttered by males at
display grounds in a variety of circumstances,
but not in the immediate presence of a female.

A quite distinct call, a disyllabic “chee-poo,”
typically marks the beginning of a bout of dis-
play. Frequently it follows immediately after a
rolled snap. It is not uttered in other circum-
stances.

Juvenile males, when displaying, sometimes
utter a rather soft, plaintive “pu.” In M. vitel-
linus this call is associated with a stereotyped
display (Appendix 1), but in M. manacus it
seems almost to be obsolescent, since it was
rarely heard, and then only from juveniles when
engaged in uncoordinated display with other
young birds.

The begging calls of young birds, in the nest

72

Zoologica: New York Zoological Society

[47: 8

and when being fed after leaving the nest, are
described elsewhere (p. 89 and 90).

The Court

The general character of the display ground
has already been described, and the constancy
of the courts from year to year mentioned. The
spacing of the courts at different display grounds
is variable. If there are sufficient suitable sap-
lings, most courts may be within a few feet of
each other and clusters of two or three may be
in contact with each other. If groups of suitable
saplings are fewer, courts may be many yards
apart. At the largest display ground seen in Trini-
dad (H in Text-fig. 3) there were some 70
courts, many of them almost touching each
other, within an area of about 20 by 10 yards.
More commonly courts are on average from one
to five yards apart. The display grounds of M.
vitellinus studied by Chapman in Panama were
far more scattered, and also very small, with not
more than five courts spaced as much as 200 feet
apart. This situation is probably unusual; at an-
other display ground in Panama visited in 1958,
the spacing of the courts was the same as in
Trinidad.

The size of a court is also variable. Most are
roughly oval and measure about 3 by 2 feet.
From this area all dead leaves are carried away
and most of the rootlets are stripped white by
the owner’s continual picking at the bark. All
that remains is bare earth, rootlets and such
sticks as are too large for the bird to carry away.

One of the upright saplings growing round
the court is more important than the others, in
that some of the main displays, and mating it-
self, take place on it. This “main upright,” as
it will be called, must be fairly straight and
smooth; those that were measured were not less
than 6 mm. or more than 20 mm. in diameter
near the ground. Its bark is usually worn smooth
by the continual rubbing of the bird’s feet, and
below it the ground is kept especially well
cleared.

The birds clear their courts at intervals
throughout the day, usually during bouts of dis-
play. Typically, the bird jumps down onto the
ground beside the object, usually a leaf, which
is littering the court, picks it up and flies to a
perch a few feet away, dropping it as it lands.
Or it may lean down from a low perch at the
base of a sapling, pick up an object and fly away
with it. Less often, a bird flies down and without
alighting, or alighting for only a fraction of a
second, picks up a leaf and flies off with it, the
whole movement being so rapid that it is impos-
sible to see in detail.

The Elements of the Display

The different displays were seen very many
times and were recorded on 16 mm. movie film,
usually at 24 frames per second (the light at the
display grounds was not good enough for ex-
posures shorter than 1/50 second) . This account
differs to some extent from previous accounts
of the display of M. manacus. Previous observ-
ers have not seen all the displays, and their ac-
counts of some that they have seen have been
inaccurate. This is not surprising; conditions of
observation are often not easy, and in any case
very fast display movements need to be seen
many times, and over a long period, before their
exact nature becomes clear.

There are also some differences between this
account and Chapman’s account of the display
of M. vitellinus. Some of these seem to represent
real differences between the behavior of the two
species; they are dealt with fully in Appendix 1.

The “ snap-jump ”

This is the most frequently used of the display
movements. Typically, the bird perches in a
horizontal position across one of the upright
saplings round its court, with its head thrust
forward and beard extended, then suddenly with
a loud snap leaps to another perch, reversing
its position as it reaches the other perch so that
it lands facing the way it came from (Text-fig.
7). The complete leap from one perch to an-

3

*

Text-fig. 7. The “snap-jump”: a male turning in the air and landing after the jump. The bird approaches
the perch in an upright position (1), tucks tail under, at the same time turning head and neck (2 and 3),
and lands facing the way it came from (4). (Drawn from movie film.)

1962]

Snow: Field Study of the Black and White Manakin

73

other usually lasts from one-sixth to one-quarter
of a second, depending on the distance between
the two perches. During the jump, the whirring
noise made during normal flight is not heard.

Many snap-jumps may be repeated in quick
succession, so that the bird crosses and recrosses
its court with great rapidity. When more than
two perches are available, there is a tendency
for the sequence of snap-jumps to follow a fixed
pattern from one perch to another.

Because the snap and the jump are simul-
taneous, it would seem that the snap depends
on the wing-movements involved in the leap.
Very occasionally, however, a bird makes a snap
and remains perched, but such snaps are less
loud than those accompanying a jump.

The “rolled snap”

The bird leans forward on its perch, raises its
wings above the back, and with a vibrating
movement that is almost too rapid to see emits

a loud rolling snap, like a succession of single
snaps run together. It frequently takes flight
immediately afterwards, or if it does not fly
utters a call, “chee-poo,” indicative of mounting
excitement. The rolled snap typically introduces
a bout of display after a period of inactivity.

The “grunt-jump”

This follows after a sequence of snap-jumps
between the saplings round the court. The bird
jumps or flies to the “main upright,” landing
transversely on it within a few inches of the
ground, then with beard extended, body tensed
and even quivering slightly, as if bracing itself
for the effort, it leaps down to the ground, turn-
ing in the air as it does so, lands momentarily on
its feet facing the perch, and leaps back to a
higher position on the same perch (Text-fig. 8).
As it leaves the ground it emits a curious sound
between a grunt and a whirr. The whole se-
quence, from leaving the perch to landing on it

\

\

\

\

\

\

\

\

\

\

\

\ »

\v

£

Text-fig. 8. The beginning of the “grunt-jump,” from two viewpoints, showing the bird’s trajectory (right,
upper). Just before jumping, the bird raises its head and neck and half-opens its wings. (Drawn from movie
film.)

74

Zoologica: New York Zoological Society

[47: 8

again higher up, is extremely rapid and difficult
to follow by eye; it lasts just over one-third of
a second.

The same grunting sound is occasionally made
by perched birds, and is accompanied by a sud-
den humping up of the back, the head being held
forward and rather low. At the same time the
wings are moved, but too rapidly for the details
to be seen. The grunting sound thus made is a
little lower-toned, lasts longer and is more vi-
brant than that made on an upward jump. A
sound approaching the grunt is also made when
a bird rapidly flies upward from the ground to
a perch, or from one perch to another. Thus the
sound is certainly mechanical in origin but it is
not clear how it is made.

The “slide down the pole”

Immediately after landing on the main up-
right, after the “grunt-jump,” the bird may move
down the perch for several inches in a head-
downward position, with the wings beating,
moving with such short steps that it appears to
slide. As it approaches the bottom of the sapling
it turns to a more horizontal position (Text-fig.
9). This display is comparatively infrequent.

The “grunt-jump” followed by the “slide
down the pole” represent the culminating phase
of the courtship display in that they immediately
precede copulation, as described later (p. 77).
The male “slides down the pole” on to the back

of the female, who perches below him on the
sapling.

“Fanning”

This is a very distinct display. The male leans
forward on his perch, holding his head low, with
the neck retracted and the beak pointing slightly
upwards, and sways his body from side to side.
At the same time he holds his wings out from
the body and. beats them at the rate of five
times per second, raising and depressing the tail
and the whole hind end of the body synchron-
ously. The beating of the wings makes a low-
frequency whirring sound, easily distinguishable
from the flight whirr. Apparently due to the re-
traction of the head between the “shoulders,”
as the wings are raised the feathers of the hind-
neck are pushed forwards and upwards so that
they stick out as two flickering white puffs be-
hind the bird’s head at either side (Text-fig.
10). This display is most often directed at a
female which has approached a male’s court.
Several seconds of fanning, during which the
male may repeatedly shift his position so as to
present himself to the female, are usually fol-
lowed by his flying to the court and executing
a series of rapid snap-jumps. The display is also
occasionally performed when no female is pres-
ent.

The upright posture

Displaying birds sometimes assume an upright

/ 2 3

Text-fig. 9. The “slide down the pole.” (1) the beginning of the slide, just after landing from
the “grunt-jump.” (2) half way down. (3) near the end of the slide, the body becoming more
horizontal and the tail fanned. (Drawn from movie film; no attempt has been made to show the
feet, which move too fast to appear on the film.)

1962]

Snow: Field Study of the Black and White Manakin

75

/ 2 3

Text-fig. 10. “Fanning.” Left, (1) and (2): wings closed and wings raised. Right (3): wings raised,
seen from the front. (Drawn from movie film.)

posture when perched, with the head pointing
upwards and beard extended. This posture does
not seem to play an integral part in any of the
main display movements described above,
though it sometimes follows a bout of snap-
jumps. The posture may be held for several
seconds.

Aggressive displays between males

When two males are perched close together,
the dominant one may extend his beard, turn
about frequently on his perch, give little jumps
along the perch, hold his head low and turn it
from side to side, or raise and vibrate his wings,
all these being incomplete forms of normal sex-
ual displays. Sometimes he may fan his tail and
turn it towards the other bird. The subordinate
bird sits quietly, with throat and body plumage
sleeked; when threatened he may flick his wings
and suddenly fan and close his tail, probably
flight-intention movements. Aggressive behavior
between males may culminate in prolonged flight
chases round and round the display ground and
in fights on the ground.

The Daily Rhythm of Display

Apart from the weeks when he is moulting,
the adult male manakin is at all times closely at-
tached to the display ground. Almost as soon as
it is light enough to see, the first males appear
at their courts. Display soon begins and reaches
a peak, then after an hour falls off to a low point
between 0900 and 1100 hours. In late morning
the intensity of display increases again and
reaches a second peak around 1400; it then
falls off again gradually and by 1700 has prac-
tically ceased. The males gradually move away,
a few late ones remaining until about 1800. The
precise times of the beginning and end of display
vary somewhat, depending on the times of sun-
rise and sunset and on the variations in light
intensity due to weather and the forest canopy.

Text -figure 11 shows the results of a continu-
ous watch on a single ringed male throughout
the daylight hours of January 27, 1960, a day

of moderately intense display. Between 0625,
when he arrived, and 1747, when he finally left,
he was present and under observation for 90%
of the time. He was probably present for some-
what longer than this, as at times he perched
a few yards away from his court in a place where
he could not easily be seen, and thus he was re-
corded as absent for a few short periods when
he was almost certainly present. His 24 recorded
absences nearly all lasted less than 5 minutes.
During one of the longest absences, of 6 minutes
from 1659 to 1705, he certainly bathed, as he
returned looking dishevelled and preened vigor-
ously ( and this is the time when observations by
forest streams showed that bathing activity is at
its height). There was no reason to suppose that
this bird’s activity on this day was in any way
unusual. Less complete observations on many
other birds showed a similar degree of attach-
ment to the display ground throughout the day.

Females visit the males at their courts mainly
during the two periods of intense display (0630*
0800 and 1330-1500). Intensity of display by
the males and the presence of the females are
of course interrelated, in so far as the presence
of a female stimulates the males, and intense
display by the males attracts the females. But
apart from this, the males’ daily cycle of dis-
play is to a great extent independent of the
females, as it is equally marked at seasons when
the females are visiting them hardly or not at
all. On the day when a continuous watch was
kept, only one visit by a female was recorded
in the afternoon. This was unusual; typically as
many females are seen visiting the males in the
afternoon as in the early morning period.

Text-figure 11 shows that the male under ob-
servation left the display ground, presumably
to feed, mainly at times when display was slack,
and that his first absence took place soon after
0800, when the early period of intense display
was over. Trapping and observation away from
the display grounds confirmed that these are the
main feeding times. In particular, trapping at
a feeding area (T in Text-fig. 3) regularly

76

Zoologica: New York Zoological Society

[47: 8

VISITS BY -

FEMALES

ABSENCES _ _ ,

OF S*tS

I I I | I I I I I I I I I I
06 07 08 09 10 II 11 13 ft 15 16 17 18 19

TIME OF DAY

Text-fig. 11. The activity at display ground A. January 27, 1960. Intensity of display measured by
number of snaps in 5-minute periods (snaps in excess of 200 omitted). Female visits: thin line, one female
present; thick line, several females present.

showed a period of great activity in the hour be-
tween 0800 and 0900. Text-figure 12 shows the
numbers of males, known to have courts, that
were trapped in this area in the different hours
of the day.

It was also found that, at least occasionally,
males may feed intensively in the short period
between first light and their arrival at the dis-
play ground. This may depend on the availa-
bility of fruiting trees in the forest canopy or in
open places along the forest edge, where the
light allows feeding at such an early hour.

Displays Between Male and Female

Males display vigorously and perform all the
display movements described above when no
female is present, as already mentioned. But the
arrival of a female leads to especially intense
and sustained display. Indeed the observer is

usually first made aware of the presence of the
cryptically colored female by the upsurge of
display and calling.

Her arrival is usually greeted by an outburst
of rolled snaps, as males fly to their courts to
display, by an intensification of the single snaps,
as males that are at their courts perform rapid
snap-jumps, and by a change in the calling from
an occasional “peerr” to a chorus of excited
“chwee” calls. If the female approaches a male’s
court, he may fly to her and, landing beside her,
give a prolonged display of “fanning,” after
which he will usually fly down to his court and
perform a series of rapid snap-jumps.

There is always a strong tendency for the
males to fly towards the female as well as to go
to their courts and display. Sometimes several
males will fly towards her while she is still at
the edge of the display ground, and though

1962]

Snow: Field Study of the Black and White Manakin

77

quantitative observations on this point would be
difficult and were not made, the tendency
seemed to be strongest in the first weeks after
the moult, and in those males that were less
settled at their courts. The tendency for males
to fly towards a female is also strong when, as
sometimes happens, she flies in high and perches
well above the courts. Several males then often
fly up to her and call and snap excitedly, but in
such cases the female appears unwilling to go
down to the courts. Females that are ready to
join the males at their courts fly in only a few
feet above the ground.

Usually males do not approach a female until
she is within a few yards of their court, and they
then perform the fanning display at her. Fan-
ning is a clear indication that the female is with-
in the male’s sphere of influence, even though
she may still be several feet from his court. Two
males were never seen fanning simultaneously
at the same female.

The behavior of the females when they first
arrive at the display ground is quiet and almost
secretive. They flit from perch to perch, within
a few feet of the ground, and show signs of ini-
tial nervousness. Early in the season, before
breeding has begun, they often arrive in groups
of up to five together; later they more often
come singly.

Early in the season they not only arrive to-
gether but also go down to the courts in groups,
and a rather confused display develops. The
male displays hard, but at the same time seems
somewhat “taken aback” by the number of fe-
males present. They perch on the surrounding
saplings and on his court, frequently flitting from
one perch to another and changing places with
each other, keeping clear of the male and yet
plainly attracted to his court.

But when a single female visits a male at his
court a well coordinated display or dance may
take place, which may culminate in copulation.

10 r— i

These dances become commoner as the breed-
ing season approaches. As would be expected,
all transitions may be seen between the confused
situation with several females present at a court
and the highly coordinated dance between the
male and a single female.

A full sequence of the precopulatory display
is as follows. The female approaches the court;
the male flies to her and “fans” for several sec-
onds; the male then flies down to his court and
executes a rapid series of snap-jumps; the female
soon follows, landing on one of the upright
saplings; a mutual dance then takes place, with
the male doing repeated snap-jumps across the
court and the female crossing him in mid-air
and landing at the place which he has just left;
after a series of these jumps, the female lands
on the “main upright;” the male jumps and lands
below her on it, then immediately executes a
“grunt-jump,” going down to the ground and
back up onto the main upright above the female;
he then “slides down the pole” onto her as she
is perched crosswise on it, and copulation takes
place; during copulation the male has one foot
on the perch and one on the female’s back.

More often, only a part of this sequence is
seen. In particular, the preliminary fanning,
which serves to attract the female to the court,
is usually given only to hesitant females, and
such birds, when they do go down to the court,
usually do no more than a few jumps with the
male. On the other hand, females that are ready
to copulate often fly straight to the main upright
and the male at once follows with the sequence
“grunt-jump”— “slide down the pole”— mounting.
It was sometimes observed, and is probably usu-
ally the case, that these birds had recently visited
the male and had danced with him several times.

When females first dance with a male they are
nervous, and the males in their turn sometimes
behave aggressively towards them. The female
tends, when flying to the perch that he has just

NUMBER

OF

ADULT
MALES 5 '

TRAPPED

u

— ltlT

A— i

06 07 08 09

10 II 12

13 14 15 16 17

18

TIME OF DAY

Text-fig. 12. Numbers of adult males with courts trapped at a feeding area throughout the day.

78

Zoologica: New York Zoological Society

[47: 8

left, to keep a little higher. Thus the two birds
cross in mid-air with the female a little above.
This tendency to keep above the male is un-
doubtedly an indication that the female is not
ready to mate, to do which, as described above,
she must allow the male to land above her on
the same upright. If the male, perched below her
on the main upright, does a “grunt-jump” to land
above her, the female, if not ready to mate, at
once flies to another perch.

The females that visit the same display ground
probably have a mutually stimulating effect upon
one another. As already mentioned, they tend
to come to the display ground in small groups,
especially at the beginning of the season, and at
all times the outburst of display that accom-
panies the visit of one female to the display
ground must help to attract other females. As
a result, there is a tendency for the nests near
any one display ground to be better synchro-
nized with each other than with nests near other
display grounds (Table I).

Table I. Synchronization of Nestings of
Females Along the Same Stretches
of Stream, 1961

Number of nests started along
different stretches of stream
(see map, Text-fig. 3)

Tripp

Stream

Arima

River

St. Pat’s
Stream

Feb. 26-Mar. 1

3

1

Mar. 2-6

2

1

—

Mar. 7-11

—

—

—

Mar. 12-16

1

4

2

Mar. 17-21

1

—

1

Mar. 22-26

_

—

—

Mar. 27-3 1

1

—

1

Apr. 1-5

2

—

—

Apr. 6-10

3

—

2

Apr. 11-15

1

—

1

Apr. 16-20

2

2

—

Apr. 21-25

1

—

—

Apr. 26-30

—

—

—

May 1-5

—

—

1

May 6-20

—

—

—

May 21-25

—

—

1

May 26-30

—

—

1

May 31-June 4

—

—

2

June 5-9

4

—

3

June 10-14

5

—

1

June 15-19

1

4

1

June 20-24

—

1

1

June 25-29

1

—

—

Notes. A small number of nests started after the end
of June are omitted.

The “Arima River” nests were all situated along the
600-foot stretch west and south of display ground F, at
the lower edge of Text-fig. 3.

Choice of Display Partner by Visiting
Females

At the main display ground, 39 color-ringed
females were seen visiting the courts and danc-
ing with males, most of which were themselves
color-ringed. Three of the females were seen
visiting the display ground in three successive
years, and eight in two successive years.

There was no evidence for any kind of pair-
formation at the display ground. Without excep-
tion, all those females that were seen visiting
more than once (21 out of the 39) visited and
danced with more than one male. Usually in the
course of a single visit females went from one
male to another. They all undoubtedly went to
many more males than they were seen to visit,
since they are difficult to follow as they flit silent-
ly from court to court in the undergrowth.

Certain males were outstandingly successful
in attracting females, especially 8 45, whose
court was close to the hide. Twenty of the 39
color-ringed females visited him, as well as many
unringed females. The large number seen visit-
ing him was only partly due to the fact that his
court was easy to observe, as only four color-
ringed females were seen to visit an adjoining
court, and at two other close courts, also easy to
observe, only 9 and 12 were seen. In the early
part of the season 8 45 was regularly visited by
four or five females together, while the males
displayed persistently but vainly at neighboring
courts. Within the area under detailed observa-
tion, two other males were also very successful,
but because their courts were farther from the
hide and the undergrowth round one of them
was rather thick, only 17 and 4 color-ringed
females were identified visiting them.

There was no obvious reason for $ 45’s suc-
cess. He was an old bird, established when
trapped in 1958. He was constantly at his court,
but not more so than many other males, and his
court was not obviously more suitable than many
others. He displayed vigorously, but so did most
other males, and the display movements are so
stereotyped that no differences could be seen
between his display and that of his neighbors.
His success remained high in all three years:
7, 7 and 1 1 color-ringed females were seen vis-
iting him in the years 1959-61, the higher num-
ber in 1961 being probably due to the greater
number of females in the population that were
ringed by then.

Copulation is seen much less frequently than
the courtship dances. As mentioned above, it
often takes place rather suddenly, with little
preliminary display. Only for 8 45 was any
worthwhile record obtained. He was seen copu-

1962]

Snow: Field Study of the Black and White Manakin

79

lating with two different color-ringed females
and 15 times with unringed females. At the least
he must have copulated with three different birds
in a single season, and undoubtedly with many
more, since observations were usually made on
only one morning a week.

Relations Between Males and the
Ownership of Courts

During most of the year there is little overt
aggressive behavior between the males. With the
ownership of their courts uncontested, neighbor-
ing males perch and display close to each other
without hostility. Nevertheless competition for
courts is strong, and when one falls vacant it is
usually soon taken over by another bird, either
an unestablished bird or one that has been oc-
cupying a less suitable court.

Some males remained unestablished for
months. Such birds often display for periods at
one of the outlying courts of a display ground,
but they do not occupy them permanently. They
shift frequently, and when an opportunity arises
move to a more central court. It is because males
do not remain satisfied with outlying courts that
the courts in the central nucleus of the display
ground remain much the same year after year,
while the number and positions of the outlying
courts change.

Unestablished males sometimes hang about
for days round the edges of an occupied court,
coming down to it and displaying when the own-
er is away and vacating it as soon as he returns.
There were also several cases of what appeared
to be joint ownership of a single court. This
situation was usually short-lived, except at one
court where it continued for 7 months. The two
males would sit together within a few inches
of each other on the court. However, one (a
ringed bird) was dominant and it alone dis-
played when both were present, the other bird
only when it was alone, and then only hesitantly.
The two birds had a strong tendency to leave
the court together and return together.

In such cases, as long as one bird is clearlv
dominant and the other subordinate, there is no
open hostility between them, aggressive behavior
being limited to the posturing described above
(p. 75). But it occasionally happens that an in-
truding male actively tries to oust the owner and
does not assume a subordinate position. Then
prolonged flight-chases may take place, the own-
er pursuing the intruder, who does not leave
the display ground but flies round and round.
Or the two birds may come to grips and roll to-
gether interlocked on the ground. The main
display ground under observation was on such
a steep slope that when this happened the two

birds would roll helplessly downhill over their
neighbors’ courts.

During the moulting season temporary
changes of ownership are frequent as old birds
abandon their courts, and newcomers, in many
cases young males who have completed their
moult earlier (p. 85), take them over. But when
the moult is over and the old birds have re-
turned, the courts normally revert to their origi-
nal ownership and the position becomes stabil-
ized. It is then rather rare for an established
bird to be ousted by another. Only one such
case occurred in the part of the display ground,
comprising some 16 courts, that was under the
most detailed observation, and the ousting was
only temporary. The aggressor, $ 109, though
an old bird, was of unsettled habits. From
March, 1959, when he was trapped, to May,
1960, he had held one court, abandoned it and
disappeared for three months, and had then oc-
cupied another court. At the end of May he
abandoned this court too and began to hang
about round the edges of the court of $ 45, al-
ready mentioned as an outstandingly successful
old bird who had been in possession of his court
for at least two years. On May 3 1 these two were
apparently involved in a fight, which unfortu-
nately was not seen. $ 109 was seen in the after-
noon at one of the bathing places, with the side
of his head bloody and his plumage dishevelled.
On the next morning he was in possession of
$ 45 ’s court and $ 45 was not present. But a
week later $ 45 was back in possession. $ 109
was again hanging about round the edges of the
court and later he shifted back to his second
court, which he retained for over a year until
observations ceased.

Once a male is well established at a court, his
tendency to keep to it still needs constant rein-
forcing by the presence of his neighbors at the
surrounding courts. This is particularly clearly
seen in the early part of the moulting season.
The display ground is then largely abandoned
except for some of the later males which have
not yet begun to moult. These birds move about
and display freely at courts at which they have
never been seen displaying before. A little later,
when the young males that have completed their
moult come to the display ground and try to
establish themselves, they too move from court
to court and seem unable to settle at one court
while others remain unoccupied.

Unestablished males occasionally shift from
one display ground to another. Ten color-ringed
males were seen frequenting two different display
grounds (A and B, and in one case A and C;
Text-fig. 3). Seven were known to be young
birds that had not yet settled, three of them

80

Zoologica: New York Zoological Society

[47: 8

being still in juvenile plumage. Two others were
adults, but were unestablished and had probably
only recently moulted into adult plumage. The
tenth was an old male (adult when ringed a little
over a year earlier) , which, having tried unsuc-
cessfully for a year to establish himself at display
ground A, shifted to another (C), where he
acquired a court.

Though they are aggressive towards each
other, males have a strong tendency to sit within
a few inches of each other on neutral ground
between their courts, especially between bouts
of display. Sometimes, but by no means always,
slight movements and postures show that one is
mildly dominant to the other. The same is also
seen when one male is hanging about round the
edge of another bird’s court; the two spend much
time sitting together when display is slack. Also
in the case of apparent joint ownership de-
scribed above, the two males would sit for long
periods close together. As a result of this ten-
dency, at times when display is slack the own-
ers of courts that are close together spend much
more time at the display ground than the own-
ers of the more isolated courts.

Obviously, this strong social tendency is es-
sential to the maintenance of a communal dis-
play ground. Aggressiveness and sociability are
so balanced as to result in a compact group of
constantly maintained, individual territories.

Mutual stimulation is also of great importance
in raising the intensity of display and so more
effectively advertising the display ground.
Though display is inhibited when two males are
too close together through competition for the
same court, in the normal situation display by
the owner of one court conspicuously stimu-
lates the neighbors to display. In the absence of
females display thus tends to occur in bursts,
one bird starting and setting off the others. The
chorus of rolled snaps that introduces these
bursts is one of the characteristic sounds of a
display ground.

The Settling Down of Young Males
at the Display Grounds

Juvenile males moult into adult plumage in
their second summer, between June and Sep-
tember. Before this they are indistinguishable in
color from the females, but their behavior dif-
fers and it is not difficult to tell the sex of a
female-plumaged bird when it has been watched
a few times in the field.

From as early as twelve weeks after leaving
the nest they begin to make incipient display
movements. They become highly social and
move about in small groups which may be seen
performing uncoordinated display, often in com-

pany with one or two males in adult plumage.
Often they display well away from a display
ground, but they also regularly visit the display
grounds and hang about round the periphery,
where they are often joined by the unestablished
adult males attached to the display ground.
They perform all the display movements of the
adult males, but as their wing-feathers are un-
modified the mechanical sounds are softer, the
“snap” being reduced to a “snip.” The displays
are also more confused. When displaying to-
gether, they have a strong tendency to fly to a
perch that another bird has just left; thus they
are continually changing places. At times they
utter a distinctive call heard only in this context,
a plaintive “pu” (p. 71).

In the study area there was a small tract of
secondary forest where juvenile males and unes-
tablished older males displayed so regularly that
several small courts were created, 10 to 20 yards
apart, some of which remained more or less
clear for months, while one remained for nearly
three years. Altogether, 1 8 different color-ringed
males were seen displaying at these courts, as
well as many unringed birds. Nearly all were
juveniles or birds known to have only recently
moulted into adult plumage; the two exceptions
were adult males from a neighboring display
ground. In addition, four females were seen dis-
playing with the males; three of these were
known to have nesting territories very close at
hand. The use of this “practice display ground,”
as it may be called, became a tradition in the
local population. During the period of observa-
tion none of the courts was ever occupied con-
tinuously for any length of time by a single
male. No other practice display grounds were
found, but their intermittent use would make
them difficult to find except in areas which are
under regular observation.

While moulting into adult plumage the young
males stay away from the display grounds and,
like the moulting adults, become inconspicuous.
When their adult plumage is complete they re-
appear and become bolder, temporarily occu-
pying courts vacated by moulting adults. But
they cannot maintain their ownership when the
adults return, and there begins a usually pro-
longed unsettled period, during which they try
to establish themselves at a court.

Fairly complete information was obtained on
1 1 color-ringed birds during this period, and less
complete information on several others. All
these were watched first as juveniles, and then
for at least a year after their moult into adult
plumage. All moulted into adult plumage in the
months June-October.

There was a strong tendency for all these

1962]

Snow: Field Study of the Black and White Manakin

81

birds to occupy courts temporarily and shift for
no obvious reason, and it was never possible
to be sure if a court, once occupied, was going
to prove permanent or not. Of the 1 1 birds, one
was established by the following April, one by
the following June, and one not until the second
February after its moult. Another held a court
in the February following its moult, but in De-
cember moved to another court, which at the
time observations ceased appeared to be perma-
nent. Another was definitely not established in
the second February after its moult but estab-
lished itself two months later. Thus these five
birds established themselves approximately 8,
10, 16, 18 and 20 months after their moult to
adult plumage. The remaining six birds were
less successful; all were still unestablished after
periods ranging from 12 to 24 months.

The history of $ 114, the bird that established
itself in the second February after its moult, was
known in some detail. It hatched in August and
left the nest on September 2, 1958. It was re-
trapped in April, 1959, and color rings were
added to the numbered ring that it had been
given as a nestling. It was subsequently seen
three times in June, once displaying at the prac-
tice display ground with another probably juve-
nile male, and once at one of the outlying courts
of display ground A with several other probably
juvenile males. It was last seen in juvenile plum-
age on June 24, being seen again in full adult
plumage on October 31, when it was again dis-
playing at the practice display ground. From
November onwards, though it was seen display-
ing five more times at the practice display

ground, it frequented display ground A more
and more. Usually it appeared nervous and hung
about round a group of central courts. In March,
1960, it tried unsuccessfully to clear a court in
a rather unsuitable, but central, place in the
display ground. From the end of June to mid-
October it was not seen and was undoubtedly
moulting. It reappeared on October 1 8 and from
then onwards was constantly present, sometimes
displaying at unoccupied courts and sometimes
hanging about. In February, 1961, at 2 Vi years
old, and IV2 years after moulting into adult
plumage, it began to clear a new court a few
yards from the central nucleus of courts, and
though the site was not very suitable it success-
fully cleared it and held it until the second week
of September, when it again disappeared for the
moult.

THE ANNUAL CYCLE
The Annual Cycle of Display

As already mentioned, except when they are
moulting the adult males are normally present
at the display ground throughout the year. From
the beginning of August, 1958, to the end of
July, 1959, an early-morning watch was made
each week at the main display ground under ob-
servation, and a quantitative record was kept of
the intensity of display by counting the snaps
accompanying each display jump. Text-figure
13 shows the intensity of display as thus record-
ed through a complete year, based on the aver-
age number of snaps per 5-minute period re-
corded in the hour of most intense display. The
figure also shows the numbers of females record-
ed as visiting the display ground during the

wo

•a OF

TRAPPED so
BIRDS
IN MOULT

200

INTENSITY

OF

DISPLAY

IOO

to

/VESTS
S STARTED
IN S-DAY
PERIODS

Text-fig. 13. The annual cycle of display, breeding and moult, August 1958 to July 1959.

82

[47: 8

Zoologica: New York Zoological Society

early-morning watches, the numbers of nests
found in the neighboring forest, and the percent-
age of adults trapped in the vicinity that were
undergoing wing-moult, and thus provides a
conspectus of the annual cycle of the species
for this one year.

It will be seen that display did not cease com-
pletely at the season of moult, but was merely
much reduced. This was because individual
males moulted at rather different times, so that
by the time that the latest ones had abandoned
their courts some of the early ones were back
again. More striking was the almost complete
cessation of display in the second half of March,
probably due to a temporary food shortage, as
discussed later. It did not occur in other years,
but in 1960/61 there was a more prolonged
cessation of display in December and early Jan-
uary, the cause of which was not clear and which
also was not recorded in other years (p. 87).
Thus the normally continuous cycle of display
may show occasional and irregular interruptions.

Text-figure 13 also shows that small numbers
of females visited the display ground from the
end of the moulting period onwards, but that,
as would be expected, the greatest numbers were
recorded at and shortly before the start of
breeding.

As shown in more detail later, the breeding
season did not start at the same time in each
year. In the five seasons of observation, the be-
ginning of breeding varied from one to six
months after the ending of the moult. Probably
in every year the males are ready to mate over a
very long period, from shortly after the end of
the moult until shortly before the beginning of
the following moult, and the timing of the breed-
ing season depends on the females, whose readi-
ness to nest is affected by environmental factors
that vary from year to year (p. 87).

To settle this point, it would be necessary to
make an examination of the males’ gonads
throughout the year. This was not a main part
of the present study, but specimens were exam-
ined when available. A series of eight males,
taken in the first half of December, i.e., very
soon after the end of the moult, had their testes
well developed (mean dimensions 3.6 X 2.4
mm.; largest 5 X 3.5, smallest 3X2), but not
so large as those of a series of nine collected
in February (mean 5.2 X 3.4 mm.; largest 7 X
5, smallest 3.5 X 2.5). These birds were col-
lected from a display ground about 12 miles
east of the Arima Valley.

The Breeding Season

Text-figure 14 shows how greatly the breed-
ing season varied in the five years. In 1957, apart

from a single nest in March, breeding was first
recorded at the end of May and quickly reached
a peak in early June. It then continued steadily
until early September. In both 1958 and 1959
breeding began abruptly in mid- April. In 1960
the start was in early January, with a single nest
at the end of December, 1959, and in 1961
breeding began in the latter half of February.
In all years, the start of breeding was well-de-
fined, many nests being built and getting their
eggs within a fortnight, while observations at
the display ground at and shortly before this
period showed intense activity resulting from
the visits of numerous females, and more copu-
lations were seen then that at any other time.
There was a little evidence, as can be seen from
Text-fig. 14, that nesting ended rather later in
1957 and 1958, when it began late, than in the
early breeding seasons of 1960 and 1961. The
breeding season lasted from 4 to 8 months in
the different years.

The relation of the breeding season to weather
and other environmental factors is considered
in a later section (p. 85).

The Moult
Sequence of moult

The full post-breeding moult follows a se-
quence similar to that of other passerine birds.
Replacement of the wing-feathers follows a con-
sistent pattern. It begins with the dropping of
the secondary major coverts. Typically the outer
ones drop a little before the inner ones, so that
as the new feathers grow they are in a graded
order of length from outside inwards. While the
coverts are growing, the innermost primaries
are moulted, the first two or three being dropped
at almost the same time. The moult of the ten
primaries proceeds regularly towards the end of
the wing, one feather being dropped when its
neighbor on the inside is about half grown. The
speed of growth of each feather is such that
three adjacent feathers are generally in various
stages of growth, four or only two being less
usual. Like the innermost ones, the outermost
two or three primaries drop at almost the same
time. The moult of the nine secondaries typically
begins when the primary moult has reached the
4th or 5th primary (numbered from the inside) .
Its sequence is less regular than that of the pri-
maries. Secondary moult begins from the two
ends and works inwards. Two or three of the
innermost secondaries drop at the same time,
but at the outer end of the row the second feath-
er does not drop until several days after the first.
Growth and replacement proceed more quickly
from the outer than from the inner end, so that
although two or three of the innermost second-

1962]

Snow: Field Study of the Black and White Manakin

83

10

NESTS

5 ■

1951

T T 1 1 7T

(w>

w

T~ — — 1

S

1958

5

1959

5

I960

5

1961

1 ! 1

4.4

, ■ , — ,1

li f

i 1 i l •'t' i •'t* 1

(w) w

1 ■— ■ i i

I , 5 s r r

•

H h.ILIlb 1

, i

i 'P 1 1 » « 8 8 •

W

iLi*i Mu m ,i

w

% IN
MOULT

Text-fig. 14. The breeding seasons, 1957-1961. Histograms: nests started in 5-day periods. Dotted lines:
percentages of trapped birds undergoing wing-moult. W: beginning of wet season. (W) : rainfall, followed
by further dry weather.

aries drop together, the last feather to grow is
usually the 5th or 6th (numbered from the out-
side), less usually the 4th. The moult of the tail
is also less regular than that of the primaries.
Typically it begins when the primary moult has
reached about the 5th primary; usually all the
feathers of the tail drop together and the new
ones as they grow are of much the same length.
The moult of the head and body is variable in
its sequence, but its beginning and ending coin-
cide closely with the beginning and ending of
the primary moult and it is usually heaviest in
the middle of this period.

The moult of the primaries, besides being the
most regular feature, spans very nearly the en-
tire period of the moult. The main stages of the
moult have been given numbers based on the
replacement of the primaries, following a meth-
od used by Miller (1961), and these numbers
have been used in Text-fig. 15. The stages recog-
nized are as follows:

Stage 1. Secondary major coverts being re-
placed; not primaries.

2. Primaries 1 and 2 being replaced;
3-10 old.

3. Primaries 4-10 old.

4. Primaries 5-10 old.

5. Primaries 6-10 old.

6. Primaries 7-10 old.

7. Primaries 8-10 old.

8. Primaries 9-10 old.

9. 3 or 4 outer primaries not yet full-
grown.

10. 1 or 2 outer primaries not yet full-
grown, or with traces of sheath at
base.

The season of moult

In contrast to the breeding season, the season
of moult in all probability varies very little. Text-
fig. 14 shows the percentages of trapped birds

84

Zoologica: New York Zoological Society

[47: 8

1959

1961

l

6

3

3

(o

3

h

8

ffl

4-

3

if

9

9

9

\6

(o

5

9

10

10

9

T17 9

©

o

>5

22.

39

3/

i ... i j

13 II O

1

5

3

4*

6

A/

0 03S.

Mr

7

mi\3

7

9

7

B ■ 1 a s

O O 10 II

1/ VI VII Vlll

Zb

IX

1 — “ T " 1 1 “1

X XI XII 1 II

Text-fig. 15. Moulting seasons, 1958-1961. Each square represents one individual trapped at the stage
of moult indicated by the number. Dotted squares: juvenile males moulting into adult plumage. Figures
below the line: percentages of the total number trapped fliat were undergoing wing-moult.

1962]

Snow: Field Study of the Black and White Manakin

85

undergoing wing moult in the four years for
which data were collected, and in Text-fig. 15
these moulting individuals are shown according
to their stage of moult. It will be seen that the
moult extends over the six months July-Becem-
ber, with most birds in moult in August-October.
Only one individual was found to be moulting in
June and one in January, and none in the months
February-May. In the four years, the moult be-
gan and ended at approximately the same time.

In calculating these percentages, all trapped
birds have been included. However, as will be
shown later, the moult of juveniles into adult
plumage, in their second autumn, takes place
on average a few weeks earlier than the moult
of adults. It has not been possible to separate the
age-groups in Text-fig. 15, because juvenile fe-
males moulting in their second autumn cannot
be distinguished from older females, but the five
juvenile males are indicated.

Length of time taken to moult

The trapping of five individuals twice, and two
three times, in the course of a single moult,
showed that, except for the innermost and outer-
most primaries, two or three of which drop more
or less together, the primaries drop at intervals
of 8 to 10 days, and that the complete replace-
ment of the primaries takes about 80 days. The
total period from the dropping of the secondary
major coverts to the completion of the growth
of the last wing-feather must take a few days
longer.

Observation at the display ground provided
an independent assessment of the length of time
taken to moult. It has already been mentioned
that the males are usually present at all times
except when moulting. Their disappearance be-
fore moulting and reappearance afterwards are
quite sudden, and in nine cases the period of ab-
sence was known accurately to within a few
days. All were between 76 and 85 days.

The moults of juveniles

Mainly in their first autumn and winter, juve-
niles undergo an apparently prolonged moult
which is not heavy at any time, involving the
head, body, lesser coverts and, at least some-
times, some of the inner secondary major cov-
erts. All but 4 of the 35 records of this moult
were in the months August to January.

In their second autumn the moult to adult
plumage takes place. Unless the past history of
the individual is known, this moult is recogniz-
able only in the male, since the female and
juvenile plumages are the same. What follows
therefore concerns only the males.

The moult to adult plumage takes place rath-
er earlier than the subsequent moults, being usu-

ally complete by the end of October and often
by the beginning of September. The most ex-
tensive information conies from field observa-
tions of 20 young males which were seen when
in juvenile plumage, most of them for the last
time in June, and then again when they had com-
pleted their moult into adult plumage, at which
time they are rather conspicuous as they display
persistently and try to establish themselves at
display grounds. Fifteen of these birds had com-
pleted their moult by the end of October, four
of them as early as mid-September. It was not
of course certain that they were seen as soon as
they had completed their moult, so these dates
are the latest possible. More exact but less
abundant evidence comes from the trapping of
five males during their moult into adult plum-
age, as shown in Text-fig. 15.

Annual variations in the moulting season

There were no obvious annual differences in
the moulting season when the population was
considered as a whole, but individual birds dif-
fered markedly in the times at which they moult-
ed in different years. Eight birds were trapped
while moulting in two different years (succes-
sive years, except in two cases), and one bird
in three successive years. These showed con-
siderable differences, of up to over two months,
in the timing of their moults. Certainly in one
case, and perhaps in two, this was due to the
fact that in the first year the bird was moulting
into adult plumage, but in other cases this was
not so.

Observations at the display ground showed the
same thing. The moulting periods of five adult
males were found in two successive years, one
in three successive years, and one in four suc-
cessive years. These showed considerable varia-
tions from year to year which could not be at-
tributed to the age of the bird (Text-fig. 16).
The cause of these variations is considered in the
next section.

Environmental Factors and the
Annual Cycle

In view of the annual variation in the breeding
season and the relative fixity of the moult, it
seems likely that the manakins’ annual physio-
logical cycle is, as it were, anchored to the
yearly cycle of the seasons by their response
to environmental factors initiating the moult.
Consideration of the juveniles strengthens this
supposition. Young birds may leave the nest in
any month from January to October, yet all, as
far as known, moult into adult plumage in the
months June-September of the year following
their year of birth.

In seeking the external regulator which year

86

Zoologica: New York Zoological Society

[47: 8

#45

56

59

60

61

VIII IX X XI XII I

672

59
6 o
61

681 59

6110 59

6109 60

61

6133 60

61

6190 60

61

Text-fig. 16. Moulting periods of seven males in successive years. Broken lines indicate that the begin-
ning or ending of a period was not exactly known.

after year maintains the timing of the birds’
physiological cycle, keeping all the individuals
more or less synchronized, we cannot consider
the Black and White Manakin in isolation. Al-
most all other Trinidad birds, so far as known,
sea birds, swamp birds and land birds, moult at
approximately the same season as the Black and
White Manakin (Snow & Snow, in prepara-
tion. 2). and this is of course the moulting season
qf northern hemisphere birds generally. It is most
reasonable to suppose that the same external
regulator is operating for all of them, and that
this regulator is day-length, even though in
Trinidad, at 10° N., the annual variation in day-
length is only 74 minutes.

If day-length is the main regulator, other more
variable environmental factors may still exer-
cise important modifying or secondary effects
on the timing of the different phases of the an-
nual cycle. The data are not inconsistent with
the hypothesis that the ending of the breeding

season and the start of the moult are in fact
partly controlled by the onset of the wet season,
though they do not follow for several weeks.
The wet season began on the following dates in
the four years for which there was information
on the moult (parentheses indicating heavy rain-
fall followed by a further dry spell) :

1958— April 29

1959— (April 9) , May 22

1960— April 20

1961 — May 22

The trapping data do not show any appreci-
able differences in the time of moult in these
four years (Text-fig. 15), but the samples are
not large and they are composed mainly of dif-
ferent individuals. The data on moulting dates
of the same individuals over successive years
show a more consistent pattern, though here
again, as would be expected, there are incon-
sistencies. The date of moult was known for four
individuals in both 1958 and 1959: all moulted

1962]

Snow: Field Study of the Black and White Manakin

87

later in 1959. The dates were known for eight
individuals in both 1959 and 1960: six of them
moulted earlier in 1960 than in 1959. The dates
were known for five in 1960 and 1961: three
of them were later in 1961 than in 1960. (One
individual, whose first moult was from juvenile
to adult plumage, has been omitted) . Thus of 17
annual differences in date of moult, 13 agreed
with the differences in the date of onset of the
wet season. The data for $ 45, the only bird
whose moulting dates were known in all four
years, agree exactly with the weather data (Text-
fig. 16). Clearly more records would be needed,
for more individuals and more years, for an ade-
quate test, but these data are at least suggestive.

The environmental factors concerned in the
onset of breeding are still obscure. Of most ob-
vious apparent importance is the weather. The
period from January to late May, during which
breeding may begin, is generally dry except at
the end, when heavy but irregular precipitations
may occur. Towards the end of this period, in
April and May, especially if there has been
little recent rainfall, the ground becomes ex-
tremely parched, even in the forest, and the
leaves of many herbs, shrubs and even large
trees wilt. In such a case, the first heavy rain-
falls stimulate many birds, including manakins,
to build nests and lay eggs (as also in northern
Venezuela, a short distance to the west of Trini-
dad; Gilliard, 1959). This occurred in 1959,
when heavy rain on April 9-10 was followed
by an outburst of nest-building and the first
manakin eggs were recorded on April 20. But
in 1957 heavy rain at the end of April had no
effect, while in 1958 breeding began at the same
time as in 1959 although there had been no
heavy rain for several weeks. The early start in
1961 might be attributed to the weather in Janu-
ary and February, which was wetter than usual,
but the very early start in 1960 took place before
the dry season had properly begun and cannot
be easily attributed to weather.

Once the breeding season has started, fluctua-
tions in activity may be pronounced and some of
these were certainly due to weather. Thus in
1961, after an early start, with steady nesting
from February to the end of April, there was an
almost complete cessation in May, followed by
a great outburst in early June. The second half
of April and most of May were exceptionally
dry in this year, and it was not until May 22 that
heavy rain fell.

Clearly other factors besides weather must
contribute to the timing of the breeding season.
Some observations suggest that fluctuations in
the availability of food may play a part in some
years. As will be shown later (p. 93), fruit grad-

ually becomes more plentiful from January to
April; in the first three months of the year the
number of species of Melastomaceae in fruit is
at its lowest, and annual variations in the fruit-
ing seasons of the various species may result in
temporary gaps, when no species is in fruit. It
seems probable that such annual variations in
the food supply during this period may help to
determine whether the manakins start breeding
early or late. The only marked food shortage
was recorded in March, 1959. In mid-March
there was a sudden decrease in activity at all
the display grounds visited in the Arima Valley.
The courts were deserted and became covered
with leaves. No Melastomaceae were found in
fruit— an exceptional situation, to judge from
later observations— and manakins visited the
guava trees in our garden and even fed on the
fallen fruit. This situation lasted two or three
weeks, after which display was resumed and re-
turned to normal. As mentioned above, there
was rain on April 9-10 and nesting began soon
after.

In the winter 1960/61 there was an even
longer period of cessation of display. Activity
declined at all display grounds in December and
at the main display ground under observation
several males which had returned after finish-
ing their moult disappeared again. The cessation
of display by Pipra erythrocephala was even
more complete. Display was gradually resumed
during January and was normal by the end of
the month. The data from the Arima Valley
alone suggested a correlation between this cessa-
tion of display and food shortage. The great
crop of Matchwood ( Didymopanax ) fruit,
which is an important part of the manakins’
food in November and December, was nearly
at an end, while two important melastomes,
Miconia kappleri and M. myriantha, had also
recently finished, after a good crop. Another
melastome of importance at this season, M. gui-
anensis, was not yet ripe. There was in fact a
gap in the sequence of the main fruits. How-
ever, during the same period Pipra ceased dis-
playing equally completely in a locality 6 miles
away on the north side of the Northern Range,
in an area of forest where food was abundant.
These observations emphasized the danger of
generalizing from the situation in one locality
only.

The problem of the start of breeding in the
Black and White Manakin becomes clearer
when it is considered in a wider context. There
is a common pattern to which the annual cycles
of most Trinidad land birds conform (Snow &
Snow, in preparation. 2). As already mentioned,
the months from July to October or November
are occupied by the moult. After the moult,

88

Zoologica: New York Zoological Society

[47: 8

there may be a minor peak of breeding, followed
by a decline in December and then a gradual
increase in breeding activity to a peak in April-
June. Seen against this common pattern, the
situation in the Black and White Manakin may
be better understood. Immediately after the
moult there is no actual breeding, though males
display hard and females pay preliminary visits
to the display grounds. From the end of Decem-
ber onwards the readiness to breed gradually in-
creases. Early in this period of increasing readi-
ness, especially good conditions of weather,
food, or perhaps other environmental factors
not obvious to the observer, are necessary to
stimulate breeding. Later, nesting may be stimu-
lated by quite slight environmental changes as
long as the weather is not too dry. Towards the
end of the dry season, if the drought has been
severe, the onset of nesting, or its resumption
after an earlier start, depends upon and follows
quickly after the first heavy rainfall. Whatever
the initiating factors, the start is abrupt, involv-
ing many birds at the same time, almost cer-
tainly due in part to the intense social stimula-
tion to which manakins are exposed both at and
away from the display grounds.

It must be emphasized that the details of the
breeding seasons, as shown in Text-fig. 14, apply
only to the Arima Valley. Visits to an area of
more humid forest 10 miles to the east showed
in each of three years that breeding began earlier
than in the Arima Valley, in one year several
weeks earlier. The difference was probably due
to the different climate, and suggests a reason
for the variable breeding season in the Arima
Valley. Rainfall decreases steadily from east to
west along the Northern Range. Concomitantly,
towards the west, with an increasingly severe dry
season there is an increasing tendency for birds
to delay breeding until the wet season begins.
In the Arima Valley, half way along the range,
it appears that environmental factors favoring
early or late starts are nicely balanced; hence the
great variations in different years.

NESTING

The Female and her Territory

The age at which the females start to breed
was not determined. Juvenile females were seen
visiting display grounds when they were not
more than five months old, but no information
was obtained on the first nesting of a color-
ringed bird. The only female of known age
found nesting was nearly three years old.

Breeding females occupy small territories
along the streams and gullies where most of the
nests are found. Soon after the moult is over
they advertise their presence by persistently call-

ing the monosyllabic “peerr”; this behavior then
declines, to be renewed again as the breeding
season approaches. A few observations of ag-
gressive behavior between two females at what
appeared to be territory boundaries suggested
that the territories are defended against neigh-
boring females, but competition for a nesting
territory does not seem to be intense. Territories
are small and probably highly compressible, as
occupied nests may be found within a few yards
of each other. As mentioned earlier (p. 78),
nests along the same stretches of stream tend
to be more closely synchronized with each other
than with nests along other streams, probably
because the females visit the same display
ground in company.

Although the females’ wing-feathers are un-
modified, in aggressive encounters with other
females and in the presence of intruders near
the nest they may give typical male displays,
but with softer sounds. The beard is protruded,
though the feathers are not very long, and little
jumps may be made from perch to perch, each
accompanied by a soft “snip.” Occasionally they
may make a soft “rolled snap.”

Observations were made on 14 color-ringed
females nesting along the section of stream that
was most frequently watched, and on two nest-
ing further afield. In general these birds re-
mained, and so far as known nested, within the
same small area during the period of observa-
tion. One bird remained in the same territory
for 3 Vi years, and others for shorter periods.
There were two known cases of change of terri-
tory. One of these birds was found nesting, with-
in an area of about 40 by 20 yards, once in 1958,
probably in 1959 (behavior suggesting nesting,
but nest not found), and twice in 1960. Half
way through the breeding season of 1961
(whether she had already attempted to nest or
not was not known), she shifted 320 yards up-
stream, where she was found nesting for the
fourth time. The other bird also shifted in the
middle of the breeding season, in this case only
150 yards.

The Nest

The nest is built within a few feet of the
ground, usually at a height of between 1 Vi and
5 feet. It is of typical manakin type, a rather
thinly woven shallow cup, slung between hori-
zontal supports. A large number of nests (38%
of the total) were slung between two side frond-
lets of a fern frond, and most of the others were
slung in forks between the side twigs of small
saplings or shrubs. Over 95% of all the nests
found were on or near the banks of streams or
gullies, many of them overhanging the water.

1962]

Snow: Field Study of the Black and White Manakin

89

The outer cup of the nest is made of rootlets
and black fungal hyphae (probably Marasmius
sp.; see Sick, 1957), with occasionally a few
dead leaves. Some of the rootlets used are very
long (29 inches in one case), and being wound
round and round, inside and outside the sup-
porting twigs, they bind the nest firmly to its
support. A little cobweb is also used to secure the
nest to its supporting twigs.

The lining is of different material. The great
majority of all nests examined were lined with
the branching panicles of Nepsera aquatica, a
small herbaceous member of the Melastom-
aceae. The rich brown stalks of these panicles
are very fine, smooth and shiny; each of the
branches into which they are subdivided ends in
a small fruit capsule. Due to the branching and
to the terminal capsules, they adhere to each
other when interwoven so that if pulled out the
lining usually comes away from the rest of the
nest intact.

Nepsera aquatica grows along forest edges
and road-sides, not in the forest itself. In spite
of this, it was the usual nest-lining of nests found
well within the forest as well as those nearer the
edges. At the main trapping place on the edge of
secondary forest, six females were trapped car-
rying nest-material in the 1961 breeding season,
and in every case it was Nepsera aquatica. In
spite of regular searching in the neighboring
secondary forest towards which they were flying
when trapped, none of the nests of these six
birds was found, which suggested that most of
them at least were flying some distance with the
nest-material, as indeed it is clear that manv
birds must which nest far from a forest edge.
No birds were trapped carrying nest-material in
other years, suggesting that Nepsera was unusu-
ally scarce elsewhere in 1961.

Unfortunately the identity of the nest-lining
was not discovered until late in the study. In-
complete observations made subsequently sug-
gested that Nepsera flowers mainly in the dry
season, from December to April, and that the
dead or fruiting panicles are mainly available
from March to August. In support of this sug-
gestion, it was several times noted that early
nests were lined with other less suitable material,
sometimes finely branched dead grass-heads. But
observations on this point were inadequate to
show whether the availability of Nepsera could
be a factor in the timing of the breeding season,
and the point requires further study.

For three nests an interval of four days was
recorded between the beginning of building (a
few strands only in place) and nest completion,
and for seven others intervals of 5-7 days were
recorded. Nests sometimes remain half com-

pleted for days, apparently because the female
is not ready to lay.

Incubation and Fledging

Detailed observations from a hide were made
at only one nest. These are the basis of the ac-
count of behavior at the nest given here. The
rest of the data in this section were obtained in
the course of repeated visits to many nests.

The clutch, normally of two eggs, is usually
laid soon after the nest is complete. In every
case in which there was information on the
point, the second egg was laid two days after the
first. Though visits were not usually frequent
enough to give an exact time of laying, the time
of laying of eleven eggs was known within a few
hours and all were laid round the middle of the
day, mostly between 1000 and 1400 hours, none
being earlier than 0750 or later than 1545.

Apparently complete clutches of one egg are
not rare. The clutch was regarded as complete
when the number of eggs present remained the
same for three days or more and the bird was
known to be incubating. Of the 244 complete
clutches thus recorded, 22 consisted of only one

egg-

The female does not usually sit until the sec-
ond egg is laid. Thereafter she is regularly on,
except for periodic absences during which she
must feed. At the nest under observation from a
hide, during three hours’ watching in the morn-
ing in fine weather, three days before the eggs
hatched, the female sat for 71% of the time.
Two completed periods on the nest lasted 46 and
53 minutes, and two recesses lasted 14 and 26
minutes. As incubation proceeds, the female sits
more and more tightly. Early in incubation she
usually flies from the nest when the observer is
still several yards away; in the last two or three
days before hatching she usually remains on the
nest until the observer is very close and then
flutters away close to the ground giving distrac-
tion display.

The incubation period (from the laying of the
second egg to the hatching of the second young)
was ascertained at seven nests to be 19 ± Vi
days and at two nests to be 1 8 ± Vi days.

The young hatch with a thin covering of
down. At the nest under observation, when they
were one day old they were brooded for 34%
of the time during a watch of nearly two hours
in the morning. When they were five days old
they were brooded for 41% of the time during
a watch of one hour 40 minutes, but the next
day they were not seen to be brooded at all.
They are generally very silent. Small nestlings
utter soft cheeping calls audible for only a few
feet; no calls were heard from large nestlings.

90

Zoologica: New York Zoological Society

[47: 8

The young are fed by regurgitation, mainly on
fruit with a small proportion of insect food. The
bird under observation from a hide normally
arrived with no food in the beak, perched on the
edge of the nest, and produced a succession of
fruits which she would give to the two young
alternately. The feeding rate was recorded on
five days and showed a steady increase, from one
feed every 28 minutes when the young were one
day old to one feed every 18 minutes at 13 days.

Like the adults, the nestlings are capable of
swallowing surprisingly large fruits. At the nest
under observation, some of the fruits brought to
the young were not only very large but con-
sisted mainly of a large seed only partially en-
closed by a small aril. After swallowing the
fruit the young regurgitate the larger seeds, and
nests with young are often found to have one or
two seeds lying in the nest-cup. At the nest
under observation, the female picked up these
seeds and usually swallowed them, but occasion-
ally carried them away. She also swallowed the
faeces of the young up to the tenth day; on the
thirteenth day she sometimes swallowed them,
and sometimes the young defaecated over the
nest-edge.

Thus as the young grow larger an accumula-
tion of regurgitated and defaecated seeds, to-
gether with a few insect hard-parts, spatters the
ground and leaves below the nest. Examination
of many of these piles of remains, as well as
observations from the hide, showed that the
young are fed on the same fruits as the adults
themselves eat, except for some of the largest
kinds.

The young remain in the nest for 13-15 days.
For six nests with two young the periods were:
13 and 14; 14; 14; 14; at least 14; 14 and 15
days. For two nests with one young the periods
were 13 and 14 days.

After they have left the nest the young are
extremely difficult to see. They perch in low
vegetation, move little, and either do not call or
call very seldom. Later they accompany the
female but are still not easy to locate because

they are so silent. It was remarkable that the
juvenile’s begging call, a loud and distinctive
“weeee-e-e,” rather plaintive and slurred at the
end, was only heard on one occasion, from two
full-grown juveniles accompanying a female.
Thus little information was obtained on the
post-fledging period by direct observation, but
data on the interval between the fledging of
young and the start of the next clutch (Table
V) suggest that the young are attended by the
female for three or four weeks.

Breeding Success and Reproductive Rate

In calculating nesting success, only those
nests have been used that were found before the
clutch was complete; most were in fact found
while they were being built. (For a discussion
of the bias introduced if nests found at a later
stage are included, see Snow, 1955) . In this way
we can obtain figures for the percentage of all
nests started ( i.e in which eggs were laid) that
reached the hatching and fledging stages, and for
the number of young produced for each nest
started. Table II summarizes the results of the
five seasons’ observations.

Like other tropical forest birds (Skutch,
1945), M. manacus has a very high rate of nest
failure. Combining all the years, 40% of all
nests started reached the hatching stage, and
19% produced fledged young; each nest started
produced an average of 0.33 fledglings. There
were considerable differences between the years,
but the significance of these is uncertain. An-
alysis by months shows that late nests were con-
siderably more successful than early nests
(Table III).

Most nests fail early, apparently through
predation (Table IV). Predation also falls heav-
ily on the 40% that reach the nestling stage. In
all, 86% of all nest losses were attributed to
predation. In nearly every case predation can
only be presumed, as there was no evidence ex-
cept a clean, empty nest. Indirect evidence sug-
gests that the chief predators are snakes, of
which there are several known or potential egg-

Table II. Breeding Success

Year

Number of
nests

Reached

hatching

Reached

fledging

% reaching
fledging

Number of
young fledged

1957

29

18 + 3?

12

41

22

1958

19

3

3

16

5

1959

30

7 + 2?

4

13

8

1960

79

39 + 11?

18

23

28

1961

70

13 + 6?

7

10

11

All years

227

80-102

44

19

74

Note. Queries in the hatching column indicate that the nest failed just before or just after hatching, the exact
time not being known.

1962]

Snow: Field Study of the Black and White Manakin

91

Table III.

Analysis of Breeding Success by Months

Number

Number of

Number

Reached

of young

young

Month

of nests

fledging

fledged

per nest

Jan.

6

0

0

-0.14

Feb.

16

2

3

Mar.

14

2

4

1

0.13

Apr.

34

2

2

J

May

43

10

17

•0.35

June

63

11

20

J

July

25

9

15

1

0.55

Aug.

26

8

13

J

eating species frequenting low growth in the
forest. The fact that no traces of chewed egg
shell were ever found in or near a nest seems to
rule out small mammals as important predators.
Neither have birds been implicated. The toucan
Ramphastos vitellinus, the only Trinidad, rep-
resentative of a family known to be nest-robbers
(Skutch, 1944), rarely comes within 20 feet of
the ground. The large cacique Psarocolius decu-
manus and the larger flycatchers may occasion-
ally rob nests, but there was no evidence that
any of these could be important nest-predators
near ground-level in the forest.

Frequently nests are found not only empty
but partly dismantled only a day or two after
they had been occupied. This is not however
due to a predator. The nest-material is much in
demand, not only by other manakins but by the
flycatcher Pipromorpha oleaginea, so that aban-
doned nests may soon be reduced to a mere
framework.

Inadequate construction or support of the
nest caused 6% of all the failures. In particular,
nests are sometimes slung between side-fronds
of a fern which after a time wither, so that the

Table IV. Causes of Nest Failure

Cause of failure

Number of nests

Eggs predated

105

Eggs deserted

3

Eggs lost through collapse of nest

11

Eggs or small young predated

24

Young predated

29

Young lost through collapse of nest

1

Nest flooded

7

Tree fell on nest

1

Human disturbance

2

nest tilts and the eggs or young fall out. Natural
catastrophes were the only other regular cause
of failure (4% of the total); in several cases
nests placed low over streams were flooded or
swept away after heavy rain, and once a nest
was destroyed by a natural tree-fall. Loss by
flooding was the only cause of failure which
showed seasonal variation; all cases occurred
in the wet months June-August.

Reproductive rate

Individual females nest more than once in the
course of the long breeding season, but it is not
easy to find out the average number of nesting
attempts made by each bird. Observations at the
display ground showed that known females vis-
ited and displayed with the males at intervals
throughout the heeding season, but were too few
to throw any certain light on how many times a
single female attempted to nest in the season.
More satisfactory information comes from ob-
servations on the repeated use of the same nests.

As already mentioned, the females usually
occupy fixed territories throughout the breeding
season. Thus when a nest is used more than
once, or when a new nest is built within a few
feet of the site of a previous one, it is likely that
the same female is involved. Also, in such cases
the intervals between the ending of one nesting
attempt and the laying of the next clutch show
certain consistencies (Table V), which further
supports the belief that the same bird is involved.
Fortunately, repeated use of the same nest or
nest-site is rather common.

Table V. Intervals Between Broods

Number of days between end of
one breeding attempt and laying
of first egg of next clutch

0-10

11-20

21-30

31-40

41 +

After loss of

eggs or young

4

10

5

5

7

After success-

ful fledging

—

—

3

4

1

In 1957 and 1958, when the breeding seasons
were rather short, 11 nests or nest-sites were
used twice. In 1960, with a very long breeding
season, 12 nests or sites were used twice, 3 three
times, 2 four times, and there was a single case
of a succession of five nestings in two nests, the
second nest, which was used twice, being built
within a few feet of the first, which was used
three times. In 1961, with a fairly long breeding
season, 11 nests were used twice and 3 three
times.

92

Zoologica: New York Zoological Society

[47: 8

These observations suggest that two to four
nesting attempts during the breeding season are
usual, two being commoner in a short season
and three or four occurring more often in the
longer breeding seasons. If three is provisionally
accepted as the average number, then, with each
nest started producing on average 0.33 fledg-
lings, each female will on average rear approxi-
mately one fledgling a year.

In the years 1957-60 15 nestlings were ringed
along the section of stream where intensive
trapping was carried out, and all, as far as
known, left the nest successfully. Of these, four
were later trapped and became fully adult (i.e.,
reached their second autumn) and a fifth was
seen several times, up to nearly a year after it
had left the nest. The other ten were never
trapped nor seen. Though a small sample, this
suggests that at least one-third of the birds which
leave the nest survive to become adult. Thus if
each breeding female rears on average one fledg-
ling a year, she will contribute on average 0.33
to the next year’s adult population.

FOOD

Feeding Habits

The Black and White Manakin is predomi-
nantly a fruit-eater, taking a great variety of
fruits from trees, shrubs and even low herbs.
Typically the bird makes a rapid sally from its
perch, plucks a fruit in flight, lands with it in
its beak, and then swallows it whole. But fruits
are sometimes plucked from a perched position
if they are accessible. Manakins do not hop
about to feed, nor do they cling to the bunches
of fruit on which they are feeding, as do the
tanagers and honeycreepers which are often
seen feeding with them.

Almost all the fruits eaten are small berries,
but a few exceptions were noted. Sometimes
pieces are plucked from the fruiting catkins of
Cecropia peltata, sections are taken from the
compound fruit of the introduced rubber tree
Castilloa elastica, and during the period of food
shortage mentioned earlier manakins came
down onto the ground and fed on fallen fruits
of the cultivated guava ( Psidium guajava ) .

They have a wide gape and can swallow very
large fruits for their size. They often have to
“juggle” large fruits in the beak for some time
before they can swallow them; probably the
fruit is softened during the process and so made
easier to swallow. With the largest fruits of all,
this juggling may go on for a minute or more,
while it seems to the observer that the bird can-
not possibly succeed in getting the fruit down.
The three largest fruits that manakins were seen
eating, after much juggling, were the following:

Coussarea paniculata (Rubiaceae), length 19,
diameter 16 mm.; Protium guianense (Burser-
aceae), length 15, diameter 11 mm.; and Cordia
lockhartii (Boraginaceae) , length 15, diameter
10 mm. Undoubtedly the wide gape is of great
importance to the species in making available a
greater range of fruits than could normally be
eaten by such small birds. Thus Black and White
Manakins can feed on fruits that cannot be eaten
by considerably larger birds such as the tanagers
Tan gar a mexicana and T. gyrola.

A small quantity of insect food is taken, also
in flight. Not uncommonly manakins follow
army ants and take the insects disturbed by
them, plucking them in flight from a leaf or tree
trunk. Occasionally a manakin is seen, when
feeding on berries, to pluck a small object, prob-
ably a resting insect or spider, from the under-
side of a leaf. When termites swarm after the
heavy rains which bring the dry season to an
end, manakins hawk for them from exposed
tree-top perches, as do many other birds.

A greater proportion of insect food is fed to
the young than is eaten by the adults themselves.
Thus insect remains were found in eight of 15
food samples collected from below nests with
young, but in only four of 93 samples col-
lected from display grounds. Nearly all the in-
sect remains consisted of hard parts of the
imagines of small Coleoptera and Diptera; re-
mains of a small damsel-fly (Odonata, Zygop-
tera) were also found in three of the samples.
The fact that insect larvae were never found
in the samples may not be significant in itself,
as they would probably not be distinguishable
among the food debris. However, the birds’ be-
havior when feeding suggests that they would
rarely be taken.

Black and White Manakins drink regularly
from the water that collects in the large colored
bracts of the abundant banana-like plant Heli-
conia cf. wagneriana. Less often, they drink
from streams.

Composition of the Food

Altogether, Black and White Manakins were
seen feeding on the fruits of 66 species of plants
(40 trees, 13 shrubs, 7 vines, 6 others), of which
4 (3 trees, 1 vine) remained unidentified. Col-
lections of seeds from display grounds and be-
low nests brought the total to 105 species, of
which 32 remained unidentified. Most of the
unidentified species were represented by only a
few seeds, and comprised a very small fraction
of the total food. The identified fruits, and the
months in which they were found to be eaten,
are listed in Appendix 2.

As already noted, one family of plants, the

1962]

Snow: Field Study of the Black and White Manakin

93

Melastomaceae, is of especial importance. 47 %
of all records of manakins observed feeding
were from Melastomaceae. All except nine of
the 108 samples collected from display grounds
and below nests contained melastome seeds,
which often far outnumbered everything else;
seven samples contained only melastome seeds.
Records were obtained of manakins feeding on
at least 17 species of melastomes, nearly all of
them trees and shrubs of the genus Miconia.
Several other less common species were prob-
ably fed on. The berries of these melastomes
contain large numbers of very small seeds which
show little difference between the species. Ex-
cept for one species with unusually large seeds,
they were not distinguished in the samples col-
lected from display grounds and below nests.

The family Rubiaceae was easily second in
importance to the Melastomaceae. At least 15
species of this family were found to be eaten,
and 14% of all observations of birds feeding
were from them. The remainder of the identified
food plants belonged to a variety of families,
with few species in each. The Euphorbiaceae
and the Moraceae, both with 4 species, were the
most important tree families in number of spe-
cies. The greatest number of records for single
tree species were from Didymopanax moroto-
toni (Araliaceae) and Ficus clusiifolia (Mor-
aceae) .

Records were obtained from a variety of
smaller plants, the only apparent requirement
being that they should bear fleshy berries of the
right size. Among these were aroids of the
genera Monstera and Philodendron, Costus
spiralis (Zingiberaceae) , Stromanthe tonckat
(Marantaceae), two species of Heliconia (Mus-
aceae) and the grass Lasiacis sorghoidea.
There were no records of manakins feeding on
the fruits of aroids of the genus Anthurium, the
epiphytic cactus Rhipsalus, or any of the mistle-
toes. They are all common, and were fed on by
some other fruit-eating birds. They are translu-
cent berries, with a rather tough skin enclosing
a sticky pulp, and perhaps are not easily digest-
ible by manakins.

The Melastomaceae— Trees and shrubs of the
genus Miconia are a conspicuous feature of the
secondary vegetation of the Arima Valley. They
also occur, but not so abundantly, in primary
forest. They range in size from shrubs a few
feet high to 70-foot trees, but apart from their
size they are all rather similar in appearance
and all bear, in conspicuous terminal panicles,
roundish berries which are from 3 to 10 mm. in
diameter and contain numerous very small
seeds embedded in pulp. Three other genera of
melastomaceous shrubs, Clidemia, Conostegia

and Platycentrum, have similar fruits, but as
food for manakins, at least in the study area,
they were of minor importance. There was also
one record of a manakin picking pieces from the
much larger fruit of a melastomaceous tree of
the genus Henriettea.

In the Arima Valley trees of the genus
Miconia produce a constant succession of fruit
throughout the year (Text-fig. 17). A greater
number of species are usually in fruit in the wet
than in the dry season (Table VI), so that the
food supply then is more assured. Two of the
commonest species, M. guianensis and M. multi-
spicata, fruit in the dry season, and one or two
others of lesser importance, but annual varia-
tions in their fruiting seasons can produce a tem-
porary gap in the succession. As already
mentioned, this apparently happened in March,
1959, but unfortunately the different species had
not been distinguished by then and it was only
recorded that no melastomes could be found in
fruit. There was also a gap in January, 1961,
when M. kappleri and M. myriantha had just
finished fruiting and M. guianensis had not yet
ripened.

The shrubs of the genus Miconia, as well as
those of the related genera mentioned above,
are of far less importance, only 4% of all the
records from melastomes being from them.
They tend to fruit less prolifically than the trees
and to spread their fruiting over a longer pe-
riod. Thus few berries are available on one plant
at one time, and manakins do not pay much
attention to them when the more abundantly
fruiting trees are available. They fruit more in
the wet season than the dry (Table VI).

The Rubiaceae— In the forest a large number
of the small trees and shrubs belong to this fam-
ily. The identification of some of them is not
easy, and a few more than the 15 recorded spe-
cies may have been involved. Like the mela-
stomes, many of them bear fruit conspicuously
in terminal panicles. Their fruits are more di-
verse in appearance, many of them being rather
larger than those of the melastomes. They con-
tain in most species one or two, in a few spe-
cies several, seeds embedded in pulp; some spe-
cies have arillate fruit. Like the melastomes, the
tree species have mainly well-defined fruiting
seasons, while the shrubs fruit over a long pe-
riod but mainly in the wet season. Thus they also
provide a more or less continuous food supply
throughout the year, but far less abundantly
than the Melastomaceae.

Availability of Food Throughout the Year

Text-figure 18 shows that there were well-
marked seasonal changes in the numbers of

94

Zoologica: New York Zoological Society

[47: 8

i a in iv v vi vii vni ix x xi xn

i i i s i i i s r“~p i ~r~

At multispicata — —

At splend&ns ■

M. tomentosa.

M. prasina
M. amplexans
M. chrysophylla
At affinis

At kappleri
At punctata
At my riant ha
At hoLoserice a

At yuianensis

Text-fig. 17. Fruiting seasons of Miconia tree species, Arima Valley, October 1959 to September 1961.
(Since most of the species did not fruit at exactly the same time in the different years, the fruiting periods
for any one year were mostly a little shorter than shown.)

kinds of fruit recorded as taken in the different
months, with a marked peak in April and a
minor one in November-December. But the
high figures in the period February-July are
partly caused by the inclusion of samples col-
lected from below nests, which included various
small seeds not found at the display grounds
and perhaps taken because the small berries con-
taining them were specially suitable as food for
nestlings. A truer picture of the variety of food
available is probably presented by Text-fig. 19,
which shows the average number of different
kinds of fruit in the collections made at display
ground A, and thus eliminates variation due to
locality and the requirements of nestlings. Again

the figure shows that there was a steady increase
in variety from January onwards, with the great-
est variety between March and June and the
least from August to October.

It has already been shown that the date at
which breeding starts is variable, ranging from
early January to late May. In relation to the food
supply, it starts at some time in the period when
the variety of food available is steadily increas-
ing. Food conditions are by no means the same
in each year, and it seems reasonable to suppose
that their variation may affect the time at which
breeding starts.

Though no quantitative assessment of insect
food has been made, it appears almost certain

Table VI. Number of Species of Melastomaceae in Fruit in Different Months

Jan.

Feb.

Mar.

Apr.

May

June

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

Trees 3 3 3 3 5 8 7

Shrubs — — — — 2 4 4

5 3 4 6 6

5 5 5 3 1

Note. Especially in the shrubs, there is a little out-of-season fruiting. This has been omitted; figures indicate
regular and abundant fruiting.

1962]

Snow: Field Study of the Black and White Manakin

95

Uj

K

<c

Ui

a:

K

s

c£

O

CC

Uj

cQ

XII I II III IV V VI VII VIII IX X XI XII

MONTHS

Text-fig. 18. Total numbers of fruit species found to be eaten throughout the year.

that insects are available in greatest numbers in
the early part of the wet season. It is then that
termites swarm, mosquitoes increase, and most
Lepidoptera are breeding. Thus the early part
of the manakins’ breeding season probably coin-
cides with the period of greatest abundance of

insects as well as fruit. By the time that breed-
ing ceases, in August and September, the variety
of fruit is decreasing and approaching its mini-
mum.

Although temporary, perhaps local, food
shortages may occur, it seems that for the Black

A /UMBER
OF FRUIT
SPECIES
IN SAMPLES

MONTHS

Text-fig. 19. Number of fruit species in food samples collected from display ground A, showing the range
(broken line) and the mean.

96

Zoologica: New York Zoological Society

[47: 8

qtn. j 9

18 ■

17

16 ■

IS

I r r — — »

i-ii iii-iv v-vi vn-vui ix-x

XI-XII

Text-fig. 20. Mean weights of adult males (crosses), juvenile males (open circles), and females (dots),
grouped in two-monthly periods. (From the data given in Appendix 3.)

and White Manakin there is no regularly recur-
ring period of food shortage such as plays such
an important part in the ecology of northern
birds. This is of course what might be expected;
the generalization has often been made by
writers on tropical forest birds, without the sup-
port of detailed data. For the Black and White
Manakin, the evidence from an analysis of the
food, given above, is further supported by ob-
servations at display grounds and from the
weights of trapped birds.

As already mentioned, adult males can feed
themselves in less than 10% of the daylight
hours. And the food collected in this short pe-
riod suffices for regular bouts of intense activity
at the court. This applies to all months of the
year, since even at the height of the moult some
early or late individuals are usually present at
the display grounds.

Analysis of weights lends further support. All
known birds which undergo annual periods of
food shortage lay down in advance reserves of
fat to meet the shortage (Lack, 1954). But
adult male Black and White Manakins showed
no significant variation in weight throughout the
year (Text-fig. 20). Adult females showed a
seasonal change in weight that was certainly
due to enlargement of the ovary and develop-
ment of eggs, and juvenile males showed a sea-
sonal increase in weight that was attributable to
the gradual development of the muscles in-
volved in display, bringing them up from the
juvenile to the adult male weight. The full data
are given in Appendix 3.

When weights of adult males and of females
from all months are combined (Text-fig. 21),
both show a similar frequency distribution, with
a long “tail” towards the upper end of the scale
and very few individuals at the lower end. This
suggests that during the period of the study few
individuals in the population could have been

undernourished or subnormal. It is significant
that none of the 38 color-ringed males that were
under observation at the display grounds died
during the two periods of temporary food
shortage.

THE ANNUAL SURVIVAL OF ADULTS

Because adult males normally retain their
courts indefinitely, once they have settled down,
it is comparatively easy, by observing color-
ringed males at display grounds over a long
enough period, to find the annual mortality. The
figure so obtained must of course be a maxi-
mum, since if a bird shifts its court to another
display ground it may not be found again and
will be recorded as dead. In fact this source of
error is almost certainly not serious in the pres-
ent case. Only one adult male was known to
have shifted to another display ground and it
had spent a year without establishing itself at
the first display ground (p. 80). Of the estab-
lished males which disappeared from display
grounds, none was ever seen or trapped again,
so it was probably correct to assume that they
were dead.

The annual survival of each male was as-
sessed from the date on which it was first found
to be established at a court. If it survived to the
same date on the following year, its further sur-
vival could again be assessed. The only male
ringed early enough to give evidence of survival
over three years did in fact survive. Several
males survived for the two years for which their
survival could be recorded. On this basis, the
survival of 38 color-ringed birds was available
for analysis, giving a total of 56 records of sur-
vival from one year to the next. The results,
given in Table VII, show that the annual sur-
vival of these males was 89%.

The annual survival of adult females could
not be assessed in the same way. Though they

1962]

Snow: Field Study of the Black and White Manakin

97

ADULT MALES FEMALES

gm. 21

11

20 ’
19 [
IQ [

n *

is [

IS ]
/¥ \
13 ]

10 20 30 90 io 20 30 90 SO SO

Text. -fig. 21. Combined weights of all adult males and females, all months.

hold breeding territories, these are shifted more
often than the males’ courts, and the disappear-
ance of a female from a small area could not be
taken as evidence of death. However, the trap-
ping data showed a high rate of survival. Thus
of the first eight females trapped, in June, 1958,
one was never retrapped nor seen again, but the
other seven were known to have survived for an
average of over 2Vi years, five of them being
alive at or shortly before the time when obser-
vations ceased, in September, 1961.

The trapping data also showed that the sex
ratio of the adults was equal or nearly so, which
implies that the females’ survival rate must be
about as good as the males’. Table VIII gives an
analysis of the first 100 birds trapped at a feed-
ing area away from the display grounds. Forty-
two of them were males, 40 of them females,

and the sex of 18 birds in female or juvenile
plumage was not determined. This last cate-
gory must have consisted of juvenile males, and
adult and juvenile females, the females prepon-
derating as they included two age-classes. The
sex-ratio of this sample must therefore have
been near 50:50. The figures for the second
hundred are similar, except that the undeter-
mined category was greater as there was not so
much time for subsequent observation and re-
trapping (known males 40, known females 35,
undetermined 25).

An annual survival of 89% is far higher than
that of any other small bird whose survival rate
is known (Lack, 1954). But very high survival
rates may well be general in tropical forest birds.
If, as seems usual, their reproductive rate is very
low, it follows that their annual survival must

Table VII. Annual Survival of Adult Males

Ringed early enough for survival over one year to be recorded: 38

Ringed early enough for survival over 2nd year to be recorded: 17

Ringed early enough for survival over 3rd year to be recorded: 1

All observations: 56

Number surviving: 35
» « 14

“ “ 1

“ “ 50

98

Zoologica: New York Zoological Society

[47: 8

Table VIII. Analysis of Sexes of Trapped Birds
(First hundred individuals trapped at feeding and
bathing places away from display grounds)

Trapped as adult males 28

Trapped in female/juvenile plumage:

later known to have been males 14

later known to have been females 40

sex not determined 18

be high. For the Black and White Manakin the
figures for reproductive rate and survival bal-
ance as well as could be expected. With 89%
annual survival, if the population is to remain
stable each female must contribute annually an
average of 0.22 adults to the next year’s popula-
tion. This figure is reasonably close to the 0.33
independently estimated for the reproductive
rate (p. 92).

DISCUSSION: THE EVOLUTION
OF COMMUNAL DISPLAYS

Advertisement and courtship displays of
males have evolved to an extreme degree in
many species of tropical forest birds— especially
hummingbirds, cotingas, manakins, bower-
birds and birds of paradise. In several species of
manakins, a few hummingbirds and cotingas,
and at least one bird of paradise, communal or
“lek” displays have been evolved. Such behavior
is of course known in birds of other habitats
(e.g., Ruff Philomachus pugnax and species of
grouse), yet there can be no doubt that the
tropical forest environment is especially favor-
able to the evolution of elaborate group displays.

The present study throws light on the ecolog-
ical factors favoring communal display in the
Black and White Manakin. To apply the con-
clusions reached here beyond this species and
other small manakins would be rash; neverthe-
less it is probable that similar ecological circum-
stances have been important in the evolution of
the sexual behavior of the other tropical forest
families mentioned above.

In the Black and White Manakin we find the
following. The food supply is rather uniform
throughout the year, with no regular period of
food shortage. Normally an individual can sat-
isfy its food requirements in a very short time
in each day (at most, 10% of the daylight
hours). The breeding season is long, and vari-
able in its time of onset. Clutch-size, and con-
sequently family-size, is low, and it seems
unlikely that a female normally finds it difficult
to supply the needs of one or two young ones.
Nesting success is very low, most of the losses
being due to predation. The expectation of life

of the adult is very high compared with that of
other small birds whose mortality rate is known.

The manakins’ low clutch-size cannot be re-
garded as in any way consequent on its special
way of life. Almost all other species building
open nests in the same habitat also lay a clutch
of two eggs (see also Skutch, 1949). I agree
with Skutch in attributing the low clutch-size of
birds in tropical forest mainly to the high rate
of predation, which gives a selective advantage
to nests which need visiting very seldom and are
as inconspicuous as possible. The problem is
essentially a quantitative one, whether on aver-
age more, or fewer, young would be reared if
the clutch were larger. As yet, neither for man-
akins nor for any other tropical forest bird are
there critical data enabling the hypothesis to be
tested. For the present argument, however, the
importance of the low family-size is that it is one
of the ecological prerequisites for the evolution
of the manakins’ communal display.

Where nest predation is heavy and clutch-
size low, the emancipation of the male from at-
tendance at the nest presents obvious advan-
tages, especially if he is more brightly colored
than the female. This is possible if the female
does not need much time to find food for herself
and her family. Thus her absences from the nest
during incubation need not be long, and when
the eggs hatch she will have no difficulty in feed-
ing two nestlings as well as herself. Selection
will thus favor the progressive dissociation of the
male from the care of the nest and young.

Once the male is free of nesting duties, the
pair bond can be broken and he becomes free to
mate with as many females as he can attract.
Thus sexual competition between the males will
intensify, and selection will promote the devel-
opment of all structures and behavior enhancing
the effectiveness of the male’s display. The ability
to find food quickly will enable the male to de-
vote a great part of his time to attracting pros-
pective mates. If the female’s breeding season
is a long one, natural selection will ensure that
the male’s period of display is at least as long.

Probably the fundamental requirement is a
food supply which enables the individual to get
its nourishment in a short time in each day. Thus
the manakins, and those cotingas and birds of
paradise with similar behavior, are all primarily
fruit-eaters. No primarily insectivorous groups
have evolved along these lines. Of the huge
family of New World flycatchers (Tyrannidae),
which includes many species inhabiting tropical
forest, one species, Pipromorpha oleaginea, is
known to have evolved this kind of behavior
(Skutch, 1960; personal obs.) : the males spend
most of their time calling and displaying in fixed

1962]

Snow: Field Study of the Black and White Manakin

99

places in the forest and take no part in the nest-
ing. Of all the small flycatchers observed in the
Arima Valley, this was the only species found
to eat fruit regularly; Skutch has recorded that
fruits are also fed to the young. The mainly
nectarivorous hummingbirds, whose communal
displays are highly developed, can also feed
themselves in a very short time in each day
(Snow & Snow, in preparation. 1 ) . Once this re-
quirement is met, the emancipation of the male
from the nest becomes possible and all further
developments can follow.

Whether or not evolution leads to communal
or “lek” displays depends on other factors. Thus
the bellbirds are sexually highly dimorphic and
the males spend much of their time in display,
but they display in scattered groups within ear-
shot but not within sight of each other (B. K!
Snow, 1960), and in some manakins the situa-
tion is similar. But the bellbirds have extremely
loud calls and can advertise their presence to
females a long way off. For a small bird in thick
forest, traditional display grounds, which the
females know and can visit when they are ready
to mate, are an obvious advantage. Almost cer-
tainly, however, this is not the main advantage
of a lek over a solitarily displaying male. A single
manakin displaying at a solitary court in the
forest would soon become known to the local
females and they would have no difficulty in
finding him. Other species of manakins have
display perches well apart from their fellows.
The crucial point is that a group of males dis-
playing together must have a much greater
attraction for the female and a greater stimulat-
ing effect on her, once she has arrived, than a
solitary male. Otherwise it would seem advan-
tageous for males to display at some distance
from their fellows, where competition for the
attention of the female would be less intense.

If the conspicuousness of males at communal
display grounds made them more liable to pre-
dation, we should have an important selective
factor working against the evolution of commu-
nal displays. But the evidence for the Black and
White Manakin is against such a hypothesis, as
the annual survival rate of males with courts
is extremely high and probably nearly the same
as that of the females, which are far less con-
spicous.

Strong sexual selection will lead to the extreme
development of display structures and exagger-
ated display movements, and as Sibley (1957)
has pointed out, the more distinct the displays
and structures are in related species, the more
effective they will be as isolating mechanisms.
But the extreme realization of the display
movements probably has another important

function. In a species in which no pair bond is
formed, the female has no opportunity to be-
come acquainted with her mate over a long
period. Ethological studies have shown the im-
portance of the agonistic elements as well as the
sexual in relationships between the sexes, espe-
cially in the early stages. When pairs are formed,
the hostile responses are gradually overcome;
in such species copulation is not usually pre-
ceded by very elaborate displays. But when pairs
are not formed and the two sexes meet only
briefly for mating, the coordination between
male and female necessary for successful copu-
lation needs to be brought about in some other
way. Hence the importance of the male’s pre-
cisely ritualized display movements, to which
the female responds with synchronized move-
ments. Thus a highly coordinated mutual dance
achieves in a few seconds what in a paired bird
needs a prolonged period of mutual adjustment.

Summary

The Black and White Manakin, Manacus ma-
nuals, a small, sexually dimorphic, mainly fru-
givorous passerine bird, was studied for 4Vi
years in an area of tropical forest in Trinidad.

There was an estimated adult population of
some 500 birds in 450 acres of forest. It is con-
sidered that this density is considerably greater
than is usual on the mainland of South America.

Males display at communal display grounds,
where each bird clears a “court” on the forest
floor. Each court has two or more saplings
around its edges, which are used by the male in
his display. Six display movements are described,
and the accompanying mechanical sounds made
by the specialized wing-feathers. Display goes on
all through the day, with a marked peak soon
after dawn and another peak in early afternoon.

Females visit the display grounds and join the
males in a highly coordinated dance over the
court. To mate, the male “slides down” one of
the saplings onto the back of the female.

Observations on color-ringed birds showed
that no pairs are formed. The males are polyg-
amous and the females to a great extent promis-
cuous.

The communal display grounds result from
the balanced social and aggressive tendencies
of the male. There is keen competition for
courts, and those near the center of the display
ground are favored. Young males take many
months to obtain a court.

Display continues all year, but is much re-
duced during the moult. The start of breeding
varied in the five seasons by up to five months,
but the season of moult did not vary appreciably.

100

Zoologica: New York Zoological Society

[47: 8

The moult lasts for about 80 days for each
bird. Juveniles moult into adult plumage in their
second year, a little earlier than the adults.

The start of breeding is probably affected by
food supply and weather, and its ending ap-
peared to be correlated with the time of onset
of the wet season.

The nest is attended by the female alone. The
incubation period was 18-19 days, the fledging
period 13-15 days. Only 19% of nests started
produced fledged young. Females usually start
2-4 nests per season. Each female is estimated
to rear on average one young per year.

Berry-bearing trees of the families Melasto-
maceae and Rubiaceae are of especial impor-
tance in the manakins’ diet, producing a succes-
sion of fruit throughout the year. There was
evidence of two periods of temporary food short-
age. The breeding season coincides with the
period of greatest availability of food.

The annual survival of adult males was 89%,
and survival of females was also very high.

The evolution of communal displays is dis-
cussed.

Literature Cited

Beard, J. S.

1946. The Natural Vegetation of Trinidad. Clar-
endon Press, Oxford.

Beebe, W.

1952. Introduction to the ecology of the Arima
Valley, Trinidad, B.W.I. Zoologica, 37:
157-184.

Chapman, F. M.

1894. On the birds of the Island of Trinidad.
Bull. Amer. Mus. Nat. Hist., 6: 1-86.

1935. The courtship of Gould’s Manakin ( Man -
acits vitellinus vitellinus ) on Barro Colo-
rado Island, Canal Zone. Bull. Amer. Mus.
Nat. Hist., 68: 471-525.

Darnton, I.

1958. The display of the manakin M. manacus.
Ibis, 100: 52-58.

Gilliard, E. T.

1959. Notes on some birds of northern Vene-
zuela. Amer. Mus. Novit., no. 1927.

Lack, D.

1954. The Natural Regulation of Animal Num-
bers. Clarendon Press, Oxford.

Lowe, P. R.

1942. The anatomy of Gould’s Manakin ( Man-
acus vitellinus) in relation to its display.
Ibis, (14) 6: 50-83.

Miller, A. H.

1961. Molt cycles in equatorial Andean spar-
rows. Condor, 63: 143-161.

SCHAUENSEE, R. M. DE

1950. The birds of the Republic of Colombia.
Caldasia, 5: 645-871.

Sibley, C. G.

1957. The evolutionary and taxonomic signifi-
cance of sexual dimorphism and hybrid-
ization in birds. Condor, 59: 166-191.

Sick, H.

1957. Rosshaarpilze als Nestbau-Material bra-
silianischer Vogel. J. Om., 98: 421-431.

1959. Die Balz der Schmuckvogel (Pipridae). J.
Orn., 100: 269-302.

Skutch, A. F.

1944. Life history of the Blue-throated Touca-
net. Wilson Bull., 56: 133-151.

1945. Incubation and nestling periods of Central
American birds. Auk, 62: 8-37.

1949. Do tropical birds rear as many young as
they can nourish? Ibis, 91: 430-455.

1960. Life Histories of Central American Birds,

II. Pacific Coast Avifauna, no. 34. Berke-
ley, California.

Snow, B. K.

1961. Notes on the behavior of three Cotingidae.
Auk, 78: 150-161.

Snow, D. W.

1955. The breeding of Blackbird, Song Thrush
and Mistle Thrush in Great Britain. Part

III. Nesting success. Bird Study, 2: 169-
178.

1956. The dance of the manakins. Animal King-
dom, 59: 86-91.

Snow, D. W., & B. K. Snow

(In preparation) 1. The time spent in feeding by
some tropical forest birds.

(In preparation) 2. The breeding seasons of Trin-
idad birds.

APPENDIX 1

The Display of Gould’s Manakin
(M. vitellinus)

In March, 1958, the display of Gould’s Manakin
was studied on Barro Colorado Island and neighbor-
ing areas in Panama. Most of the observations were
mad? at the display ground studied by Chapman in
1932 and 1935. As far as could be seen by field ob-
servations and by examination of motion picture
film, its display was in general identical with that
of M. manacus, with the following differences.

(1) “Fanning” was never seen, nor does Chap-
man mention it.

(2) The rolled snap, followed by “chee-poo,”
though the same as in M. manacus, appeared to be
a more stereotyped and commoner display pattern.
Chapman also remarked that it was a well-defined
display.

1962]

Snow: Field Study of the Black and White Manakin

101

(3) The “slide down the pole,” following the
“grunt-jump,” was much more commonly per-
formed by the individuals under observation than is
normal in M. manacus. Twice a male was seen to
slide down onto a female and copulate with her,
after a preliminary dance similar to that of M. man-
acus; the sequence was filmed. Analysis of motion
picture film shows that in one case, at the end of the
slide, the bird remained head downward near the
bottom of the perch for about a second, fanning its
wings and vibrating its body somewhat as in the
fanning display of M. manacus. The male that cop-
ulated also vibrated its body in the same way.

These observations provide a clue to the very puz-
zling “dirigible pose” described by Chapman. He
described this display, which he saw performed by
only one individual, as follows: “With bill touching
the end of a slender, broken sapling the size of a
pencil and about eighteen inches high, it fluttered
its wings while holding a horizontal pose; then, with
bill still pressed to the sapling, it slid down to the
court and, with bill now touching a root, wings still
fluttering, seemed to be standing on its head.” Aero-
dynamically it is difficult to see how a manakin
could flutter its wings and remain motionless in the
air head downwards, since in such a position the
force generated by its wings would have no down-

ward component. It seems almost certain that what
Chapman saw was the “slide down the pole” (during
which the bird’s feet grip the perch, though this is
not easy to see), followed by wing-fluttering at the
bottom of the perch. In the poor light of the forest
floor it is often difficult to see the details of these
very rapid display movements.

(4) During low-intensity to-and-fro jumps across
the court, without the initial snap and with normal
whirring flight, a low “flup,” clearly a wing noise,
was often made on landing. The same sound was
heard on two occasions when a bird landed after
“snap-jumps,” and was also occasionally heard
made in mid-flight. It was never noted in M. man-
acus.

(5) Juvenile males, engaged in uncoordinated
display behavior in the trees around the display
ground, were several times seen to perform a display
that was not recorded in M. manacus. Perched on a
horizontal twig, the bird would look downwards,
then with an initial little jump upward would jump
down to a lower perch, uttering a weak plaintive
“pu” on jumping. The same “pu” call was occa-
sionally heard from juvenile males of M. manacus
(p. 71), but it was not seen to be associated with
a jump.

APPENDIX 2

Fruits Eaten by the Black and White Manakin

This list includes all the identified fruits found to
be eaten by Black and White Manakins, by direct
observation and by collection of regurgitated seeds
from below nests, from display grounds, and in a
few cases from trapped birds. The numbers follow-
ing each plant name indicate the months in which
the fruit was found to be eaten. These figures are
used in Text-fig. 18, with the addition of the follow-

ing numbers of unidentified fruits in each month:
January, 2; February, 5; March, 5; April, 16; May,
2; June, 4; July, 3; August, 1; September, 1; Novem-
ber, 3; December, 3.

I am much indebted to Mr. N. Y. Sandwith and
Dr. J. J. Wurdack for help with many of these
determinations.

Podocarpaceae

Podocarpus sp., 12

Gramineae

Lasiacis sorghoidea (Desr.) Hitchc. & Chase, 1

Araceae

Monstera pertusa (L.) de Vriese, 11
Philodendron sp., 2

Liliaceae

Smilax sp., 4

Musaceae

Heliconia cf. wagneriana O. G. Peters, 11
Heliconia hirsuta L.f., 4, 6

Zingiberaceae

Costus spiralis Rose., 12

Marantaceae

Stromanthe tonckat (Aubl.) Eichl., 12

Moraceae

Castilloa elastica Cerv., 6, 7

Cecropia peltata L., 6, 7, 9

Ficus citrifolia Mill., 2

Ficus clusiifolia Schott ex Spr., 2-7, 10, 11

Urticaceae

Trema micrantha (L.) Blume, 1, 5

Nyctaginaceae

Pisonia eggersiana Heimerl., 7, 8
Pisonia sp., 7, 8

Lauraceae

Ocotea canaliculata (Rich.) Mez, 3-5
Ocotea oblonga (Meissn.) Mez, 3, 5, 6
Phoebe elongata (Vahl) Nees, 4, 5

Connaraceae

Rourea surinamensis Miq., 3-7, 1 1

Burseraceae

Protium heptaphyllum (Aubl.) March., 7, 8

102

Zoologica: New York Zoological Society

[47: 8

Euphorbiaceae

Aquifeliaceae

Sapindaceae

Tiliaceae

Dilleniaceae

Flacourtiaceae

Myrtaceae

Melastomaceae

Araliaceae

Boraginaceae

Verbenaceae

Solanaceae

Rubiaceae

Compositae

Alchornea triplinervia (Spr.) Mull. Arg., 12
Hieronyma caribaea Urb., 6-8, 10, 11
Maprounea guianensis Aubl., 1 1
Richeria grandis Vahl, 3, 4
Ilex sp., 12

Cupania rubiginosa (Poir.) Radik., 3-5
P aullinia fuscescens H. B. K., 10, 11
Sloanea laurifolia (Bth.) Bth., 3
Sloanea stipitata Spruce ex Bth., 5
Doliocarpus dentatus (Aubl.) Standi., 4-8
Pinzona calineoides Eichl., 2-5
Laetia procera (Poepp. & Endl.) Eichl., 2
Myrcia leptoclada DC, 4, 5
Clidemia sp., 8, 12
Henriettea sp., 8

Miconia acinodendron (L.) Sweet, 11

Miconia affinis DC, 6-9

Miconia amplexans (Crueg.) Cogn., 7

Miconia chrysophylla (Rich.) Urb., 5

Miconia guianensis (Aubl.) Cogn., 1, 2, 5-7, 12

Miconia kappleri Naud., 11, 12

Miconia mucronata (Desr.) Naud., 1, 2

Miconia multispicata Naud., 1-6

Miconia myriantha Benth., 10-12

Miconia nervosa (Sm.) Tr., 5, 7

Miconia prasina (Sw.) DC, 4-10

Miconia punctata (Desr.) D. Don, 10-12

Miconia racemosa (Aubl.) DC, 2

Miconia splendens (Sw.) Griseb., 4

Miconia tomentosa (Rich.) D. Don, 5, 6

Didymopanax morototoni (Aubl.) Dene. & Planch., 1-3, 10-12

Cordia bicolor A. DC, 7, 8

Cordia curassavica (Jacq.) R. & S., 2, 4, 5, 12

Aegiphila integrifolia (Jacq.) Jacks., 12

Lantana camara L., 2, 5

Cestrum latifolium Lam., 2, 4, 5

Amaioua corymbosa H. B. K., 2, 4-8

Cephaelis muscosa (Jacq.) Sw., 1, 5, 6, 8

Cephaelis tomentosa (Aubl.) Vahl, 2, 5, 6, 8, 9

Cephaelis sp., 1

Chiococca alba (L.) Hitchc., 11, 12
Coussarea paniculata (Vahl) Standi., 1
Gonzalagunia spicata (Lam.) Gomez, 1, 12
Lacistema aggregatum (Berg) Rusby, 5-7
Malanea macrophylla Bartl., 2, 4
Palicourea crocea (Sw.) DC, 1, 12
Psychotria cuspidata (Bredem. ex Willd.), 4-6
Psychotria marginata Sw., 4, 11, 12
Psychotria trinitensis Urb., 4, 5, 11
Psychotria undulata, 2, 4, 6, 8-11
Rudgea freemanii Sprague & Williams, 1 1
Wulffia baccata (L. fil.) Kze., 9

APPENDIX 3

Weights and Measurements

Weights

All trapped birds, unless they were wet, were
placed in a cloth bag and weighed immediately
after being caught. A spring balance, accurate to 0.5
gm., was used and was kept regularly calibrated.
Altogether, 774 weights were obtained.

Eighteen birds were trapped and weighed more
than once on the same day, two of them three times

and the remainder twice. Analysis of their weights
showed that individuals increase their weight slight-
ly in the course of the day, by approximately 0.75
gm. from 0700 to 1700 hours. For the analysis of
seasonal weight-changes, weights taken from 1100
onwards need to be reduced by 0.5 gm. to make
them comparable with those taken earlier. This has
been done in the tables that follow. In the case of
birds trapped more than once on the same day, the
first weight has been used.

1962] Snow: Field Study of the Black and White Manakin 103

The mean of 185 weights of adult males was female plumage have not been used, as it was not

18.20 gm. (range 15.5-22.5 gm.) The mean of 310 known whether they were females or juvenile

weights of females was 16.45 gm. (range 13.5-21.5 males.)
gm.). Juvenile males were intermediate, the mean of

59 weights being 17.02 gm. (range 13.5-19 gm.). Seasonal changes in weight have been discussed

(A considerable number of weights of birds in in the body of the paper (p. 96).

Weights of adult males

gm.

Jan.-Feb.

Mar. -Apr.

May-Jun.

Jul.-Aug.

Sep.-Oct.

Nov. -Dec.

22.5

1

22

1

1

21.5

1

1

21

1

2

20.5

2

2

20

2

1

2

4

19.5

1

1

1

1

5

19

5

4

4

4

2

18.5

5

5

9

4

5

4

18

4

4

10

6

4

5

17.5

2

4

11

4

2

6

17

4

1

5

5

6

2

16.5

6

5

1

1

4

16

1

15.5

1

Weights of juvenile males

gm.

Jan.-Feb.

Mar. -Apr.

May-Jun.

Jul.-Aug.

Sep.-Oct.

Nov.-Dee.

19

3

3

18.5

1

3

1

18

1

1

17.5

2

7

2

17

1

2

2

1

7

16.5

3

4

16

4

1

2

15.5

1

1

1

1

15 3

14.5
14

13.5 1

Weights of females

gm.

Jan.-Feb.

Mar.-Apr.

May-Jun.

Jul.-Aug.

Sep.-Oct.

Nov. -Dec.

21.5

1

21

2

1

20.5

1

20

1

1

2

19.5

3

19

2

2

5

1

1

18.5

1

2

6

2

1

18

1

1

6

2

17.5

4

2

9

7

3

2

17

3

5

11

13

4

5

16.5

3

4

9

14

10

11

16

5

2

12

13

7

11

15.5

3

3

8

11

8

8

15

4

2

8

8

5

5

14.5

2

1

1

2

5

6

14

3

2

13.5

1

104

Zoologica: New York Zoological Society

[47: 8: 1962]

Measurements

The wings of nearly all the birds trapped were
measured (the wing being held in the naturally
closed position and not flattened). In addition the
width of the outermost (10th) primary was meas-
ured 5 mm. from the tip in a smaller number of

Wing-lengths

mm.

Adult

males

Juvenile

males

Females

57

2

2

56

10

15

55

3

16

27

54

4

8

28

53

20

3

6

52

17

2

51

11

individuals, in order to find out if juvenile males are
distinguishable from females by the degree of modi-
fication of the primaries (p. 70). The results, tabu-
lated below, show that the wing-length of adult
males averaged 2 mm. less than that of juvenile
males and females, which did not differ appreciably
from each other. Juvenile males, however, had
slightly wider (less specialized) outer primaries
than females.

Width of outermost primary

mm.

Adult

males

Juvenile

males

Females

3.5

2

3

9

13

2.5

3

14

2

17

3

1.5

5

9

Longevity of Fishes in Captivity,
as of September, 19561

Sam Hinton

Aquarium-Museum, Scripps Institution of Oceanography,
University of California, San Diego

FISHES are often credited with the attain-
ment of tremendous age. Pliny, for exam-
ple, attributed a normal life-span of 90
years to “Esox,” probably the freshwater pike.
Gesner, in his leones, told of a pike 267 years
old, and Baldinger (ca. 1802) mentioned one
that died in its 277th year. It is well known that
we can place little credence in figures such as
these. As Frank Buckland wrote in 1880, “From
the days of Gesner downwards, more lies— to put
it in very plain language— have been told about
the pike than any other fish in the world; and the
greater the improbability of the story, the more
particularly is it sure to be quoted” (Trautman
& Hubbs, 1936). The skeleton of a supposedly
276-year-old pike, which had long resided in a
museum, was found to consist of bones assem-
bled from several individual fish (Breland, 1952),
which, to say the least, throws doubt upon its
authenticity. Furthermore, the abode of this pike
had been a natural lake, in an area where this
species is abundant, and it is not likely that posi-
tive identification of the same individual could
have been made throughout the years. The same
may be said of the various long-lived carp de-
scribed by several authors ( e.g . Suffield, 1874);
it is probable that each of these records repre-
sents not a single fish, but a succession of indivi-
duals. Carp have been credited with life-spans
ranging up to 375 years (Mehwald, 1872).

Nevertheless, figures for extreme longevity in
fishes have been widely accepted, and they have
helped to bring about the belief that fishes do
not suffer senescence, but instead continue in
the growing vigor of youth until their lives are
cut short by accident (Brown, 1957). There is,
of course, abundant evidence to the contrary.

Contribution from the Scripps Institution of Ocean-
ography, New Series.

Rasquin & Hafter (1951) and Gerking (1959),
for example, showed that senility changes in at
least some teleosts conform to the usual verte-
brate pattern. Comfort (1956) discussed symp-
toms of “old age” in Carassius auratus and
Abramis braina. Many small fishes, of the kinds
usually kept in the home aquarium, become se-
nile in a few months or years. Certain large
fishes, on the other hand, such as the true hali-
buts ( Hippoglossus hippoglossus and H. steno-
lepis) , appear to reach an equilibrium in which
they keep growing indefinitely at a low but nearly
uniform increment per annum (Carl L. Hubbs,
personal communication).

A number of techniques have been evolved by
means of which the age of a fish may be deter-
mined, but most of these are applicable only to
the first few years of life and have little value in
the determination of the maximum age of a
species. Length-frequency methods work very
well until the age of maximum growth rate has
passed, after which year classes are scarcely dis-
tinguishable. Counts of scale rings are accurate
for a greater portion of the life span, but the
rings tend to become obscured as the growth rate
slows down. Gandolfi-Hornyold (1935) exam-
ined the scales of a European eel, Anguilla an-
guilla, that had lived for 24 years in the Aquar-
ium at the Hydrobiological Institute of Toulouse;
he could count only 11 growth rings. This dis-
crepancy might be explained as the result of the
specimen’s having lived in a stable environment,
without normal seasonal changes of temperature
and food supply, but other observations indicate
that eels may stop growing and stop forming
growth marks on the scales. American eels, An-
guilla rostrata, known to have been confined for
a half-century in certain lakes of southern Michi-
gan, were determined by Carl L. Hubbs, who was
then Director of the Institute for Fisheries Re-

105

106

Zoologica: New York Zoological Society

[47 : 9

search at the University of Michigan, to be of
normal adult size and to have only about 11
year marks on their scales (Hubbs, personal
communciation). The scales of a European eel
that lived in captivity for 57 years according to
Macleod ( 1949) , were apparently not studied.

The otolith method is more accurate than the
scale method for some cases— especially when
thin otolith sections are examined microscopi-
cally — but this method, too, is most useful in
studying the important early growth years.

The method of direct observation, though
fraught with inaccuracy, and limited severely in
its applicability to animals under natural condi-
tions, seems to be the least equivocal man-
ner of determining the maximum age attained
by fishes. A number of lists of such observations
have been published, some of them setting forth
the records of longevity as observed in a particu-
lar aquarium. Notable among these are the arti-
cles by Director Charles Haskins Townsend, Sec-
retary Ida M. Mellen (1918) and Pathologist
Ross F. Nigrelli, all of the New York Aquarium.
Flower (1925, 1935) listed the records available
to him; a few were the result of scale ring studies
or other methods of estimating age, but the ma-
jority were based on direct observation. Some of
the latter concern fishes that were observed in
their natural environments, but most refer to
aquarium specimens. Bourliere (1946) used
Flower’s data, plus a few additional figures, in a
discussion of average age as compared to maxi-
mum age in 56 species of fishes.

In spite of these excellent lists, conversations
and correspondence with aquarium curators led
the author of the present paper to the realization
that a number of unpublished records were in
existence, and accordingly a program was under-
taken to gather as many as possible for conveni-
ent reference. It is hoped that this list of records
may be kept up to date, and revised at appropri-
ate intervals— much in the manner of the annual
list of records of snake longevity begun by Per-
kins (1947) and continued by Shaw (1958).

It was decided to limit records to those obser-
vations made in aquariums, excluding data from
observations made in natural bodies of water,
even though some of the latter are probably per-
fectly valid (see, for example, Gandolfi-Horn-
yold, 1935) . It was also decided to exclude those
species having a captive life span of less than
five years. Information on the shorter-lived spe-
cies would without doubt be extremely interest-
ing, but its compilation would have been a tre-
mendous task. Questionnaires were sent to 56
public aquariums in many parts of the world,
requesting information on the longest-lived spe-

cimen of every fish known to have lived in cap-
tivity for five or more years. Nineteen replies
were received, containing information on 238
species. To these have been added appropriate
non-duplicated items from Flower (1925, 1935),
Simpson (1957) and Nigrelli (1959).

The Aquarium of the Royal Zoological Soci-
ety, Amsterdam, Holland, listed the largest num-
ber of species: 95. The same aquarium also re-
ported the longest-lived fish— a sturgeon, Aci-
penser ruthenus, whose life span fell just four
months short of 70 years.

There were more reports for cyprinids than for
any other family— 36 species in all. The oldest
cyprinid is a carp, reported by the Frankfurt-am-
Main Aquarium as having lived there for 38
years. Carp were listed by ten aquariums, more
than for any other species, and the average max-
imum age was well over 19 years. Next most
numerously represented is the Family Characi-
dae, with 35 species. The oldest characin is a
piranha, Serrassalmus niger, still living in Chi-
cago’s John G. Shedd Aquarium after 21 years
and 4 months. Third is the Family Serranidae,
with 25 or 26 species (one listing is simply Epine-
phalus sp.). The oldest individuals are the 30-
year-old Dicentrarchus labrax of the Amster-
dam Aquarium and several groupers of the
genus Epinephalus, reported by A. W. C. Herre
(in a letter to Earl S. Herald) as having lived
in the Manila Aquarium for at least 30 years.

Several families are remarkable for the long
lives of their members. Fifteen specimens of
acipenserids are reported, for example, with an
average captive life of 19 years and. 6 months;
this includes the figures for Acipenser sturio,
which has apparently been relatively short-lived
in captivity, with a record of less than 7 years.
Acipenser ruthenus, on the other hand, was re-
ported 6 times, and its captive life has averaged
31 years and 6 months.

The lepisosteids comprise another hardy
group; the report lists 17 individuals, whose ages
average 17 years and 7 months. The average cap-
tive life span of the 34 individuals of the Serrani-
dae is 8 years and 10 months; only 6 of these,
however, lived for longer than 10 years, and the
median age of the records for the family lies
between 7 and 8 years. This, incidentally, was
one of the reasons for deciding upon the 5-year
minimum for this report, instead of reducing the
list by drawing the line at 10 years, as Perkins
and Shaw have done with their snake records; a
10-year minimum would have resulted in the
virtual exclusion of the serranids.

Muraenids have done well in captivity; 6
specimens of more than 20 years’ residence were

1962]

Hinton: Longevity of Fishes in Captivity

107

reported by various aquariums. This does not
compare favorably, however, with the legends
telling of morays kept by Roman emperors
through several generations. Perhaps such a long
life presupposes a steady diet of Christians.

It will be noted that members of the family
Esocidae are reported only three times and that
the maximum age reported for any pike is only
10 years, the average being 8 years and 6 months.
This makes it even more nearly certain that the
old stories of bicentennarian pike are, to put it
charitably, exaggerated.

It should be pointed out that some of the non-
anadromous salmonids appear to be quite long-
lived, and that there are hatchery records that
far exceed the ones from aquariums listed here.
For example, Leach (1924) states that female
rainbow trout have been kept and regularly
stripped for as long as 14 years, and there is a
record of two lake trout that lived for 23 and 24
years in the Shasta hatchery operated by the
California Division of Fish and Game (Anon.,
1949).

Mullets (Mugilidae) are represented on the
list by 5 species, one of which ( Liza chelo) re-
sided for 23 years in the aquarium of the Marine
Biological Association of the United Kingdom
in Plymouth. This was a surprise to the author,
whose experience with the one species of Mugil
occurring in southern California has been that it
is a delicate fish, rarely living for more than a
few months. California’s single species of mullet
is referred to the supposedly cosmopolitan Mugil
cephalus, which is reported as having lived for
9 years in the aquarium at the Hellenic Hydro-
biological Institute at the Station of Rhodes in
the Dodecanese Islands. This difference may, of
course, result from better equipment or better
aquariological technique, or it may reflect some
basic physiological difference between stocks
from different regions.

Many fishes are known to adapt well to cap-
tivity, in that they show a low initial mortality
after capture, but appear nonetheless to have a
naturally short life expectancy. Aquarists find,
for example, that as a rule most members of
the family Blenniidae adapt easily to the condi-
tions of captivity, and yet the family is repre-
sented on this list by a single report— a specimen
of Blennius pholis that lived for 5 years in the
Danmarks Akvarium in Charlottenlund, Den-
mark. Such regular aquarium inhabitants as the
cyprinodonts, gobiids, clinids, and cottids are
not reported at all, and it is tempting to infer
that their natural life spans fall short of five
years.

Also obviously absent from the list are most

of the fast-moving pelagic fishes of many fam-
ilies, including the Coryphaenidae, Thunnidae,
Istiophoridae and Xiphiidae. This is undoubted-
ly due, in some groups, to an inability to adapt
to captivity, but in others— perhaps in most— it
is again a question of a short span of life. Sail-
fish, for instance, very rarely reach the age of 5
years in the natural state (DeSylva, 1957), and

F. G. Wood, Jr., of the Marine Studios, Marine-
land, Florida, has found that the life span of the
dolphin ( Coryphaena hippurus ) may be even
shorter (personal communication). Several in-
vestigators are finding the tunas to be short-lived.

Elasmobranch fishes are reported only 20
times. The oldest of these is a nurse shark ( Gin -
glymostoma) reported as still living in the John

G. Shedd Aquarium after 24 years.

The original request for information did not
ask specifically for data concerning the manner
of death, although this question will be included
in future polls. Where such information was vol-
unteered, it appeared that accident plays an
important role even within the sheltered confines
of a well-managed aquarium, and that every
aquarist has suffered his share of catastrophes.
Few of us, however, have had the misfortune to
face wartime disasters such as those that brought
about 100% fatalities in Munich, Bremerhaven
and Manila. Douglas P. Wilson of Plymouth
wrote that his old Dicentrarchus labrax died
when vandals crept in through the aquarium’s
bomb-shattered windows and drained several of
the tanks.

In the following fist only the oldest record for
each species is noted. The second column (A)
shows the number of aquariums in which each
species is reported to have lived for five or more
years, while the third column (B) gives an ab-
breviation representing the home of the oldest
listed individual. The last column (C) tells the
age of this specimen in years and months; an
asterisk (*) indicates that the specimen was
still living when the report was first received by
the author (in 1956). The sequence of families
is that of Berg (1940).

Acknowledgments

Thanks are due the aquarists all over the world
who took time from their busy schedules to fur-
nish this information. Dr. James A. Bohlke of
the Academy of Natural Sciences of Philadelphia
helped by supplying information in the taxo-
nomy of the Characidae, Cichlidae, and Chan-
nidae. Special thanks are due Dr. Carl L. Hubbs
of the Scripps Institution of Oceanography, Uni-
versity of California, San Diego, who undertook
to bring the nomenclature into a consistent

108

Zoologica: New York Zoological Society

[47: 9

scheme and to re-group synonymies. Unfortu-
nately, circumstances beyond his control con-
spired to keep Dr. Hubbs from going over the
list as thoroughly as would have been desirable,
and any errors that remain must be attributed to
the author.

Abbreviations

Abbreviations used in the Table below are as
follows :

AMS— Aquarium of the Royal Zoological Soci-
ety, Amsterdam, Holland.

BREM — Tiergrotten und Nordsee-Aquarium,
Bremerhaven, Germany.

CHAR— Danmarks Akvarium, Charlottenlund,
Denmark.

CHI— John G. Shedd Aquarium, Chicago, Illi-
nois.

DAL— Dallas Aquarium, Dallas, Texas.

(F) — Following the abbreviation of an aquar-
ium, refers to a report made by Flower (1925
or 1935).

FLOW — Refers to Flower (1925 or 1935),
where the location of the specimen is not cer-
tain.

FRANK— Frankfurt-am-Main Zoological Gar-
den, Frankfurt, Germany.

GIZ— Giza Zoological Gardens, Giza, Egypt.

HERM — Ocean Aquarium, Hermosa Beach,
California.

LON— The London Aquarium, London, Eng-
land.

MON— Institut Oceanographique, Monaco.

MUN — Tierpark Hellabrunn, Munich, Ger-
many.

NY — New York Aquarium, New York, New
York.

NZ — Wellington Aquarium, Wellington, New
Zealand.

PHIL— Fair mount Park Aquarium, Philadelphia,
Pennsylvania.

PLY — Marine Biological Association of the
United Kingdom, Plymouth Laboratories,
Plymouth, England.

RHO— Athens Academy, Hellenic Hydrobiolog-
ical Institute, Station of Rhodes, Dodecanese
Islands, Greece.

SF— Steinhart Aquarium, San Francisco, Cali-
fornia.

SIO— T. Wayland Vaughan Aquarium-Museum,
Scripps Institution of Oceanography, Univer-
sity of California, San Diego.

SHA— Shasta Trout Hatchery, Mt. Shasta, Cali-
fornia.

TOL— Toledo Zoo Aquarium, Toledo, Ohio.

WUP — Zoologischer Garten, Wuppertal, Ger-
many.

Longevity of Fishes in Captivity as of September, 1956
(Limited to reported life of five years or more.)

Column A— Number of aquariums reporting given species.

Column B— Aquarium reporting the oldest specimen.

Column C— Age of the oldest specimen (years/months).

* Indicates that the specimen was living as of September, 1956.

Family and Species

Heterodontidae— Horned Sharks
Heterodontus francisci (Girard)
Orolectolobidae— Nurse Sharks
Ginglymostoma cirratum (Bonnaterre)
Carchariidae— Sand Sharks
Carcharias taurus Rafinesque
Scylliorhinidae— Cat Sharks
Scylliorhinus canicula (Linnaeus)

S. stellaris (Linnaeus)

Triakidae— Smooth Dogfishes
Triakis semifasciata Girard
Pristidae— Sawfishes

Pristis pectinatus Latham
Rajidae— Skates

Raja clavata Linnaeus
R. maculata Montagu
R. punctata Risso

A

B

C

1

SIO

12/3

1

CHI

24*

2

NY

9

4

LON(F)

8/4

2

AMS

19/4

1

HERM

6

1

CHI

8/1

4

CHAR

14/7

1

LON(F)

5/7

1

RHO

6/10

1962]

Hinton: Longevity of Fishes in Captivity

109

Longevity of Fishes in Captivity as of September, 1956 (Continued)
(Limited to reported life of five years or more.)

Family and Species

A

B

C

Dasyatidae— Stingrays

Dasyatis pastinaca (Linnaeus)

2

LON(F)

21

Torpedinidae— Electric Rays

Torpedo marmorata (Linnaeus)

1

RHO

6/10

Ceratodontidae— Australian Lungfish

Neoceratodus forsteri (Gunther)

3

CHI

22/4*

Lepidosirenidae— South American Lungfish

Lepidosiren paradoxus Fitzinger

1

AMS

9/8

Protopteridae— African Lungfishes

Protopterus aethiopicus Heckel

2

LON(F)

6/10

P. annectens Owen

4

NY

23*

P. dolloi Boulenger

1

LON(F)

8/8

Polypteridae— Bichirs

Polypterus senegalis Cuvier

1

GIZ(F)

34

Acipenseridae— Sturgeons

Acipenser brevirostrum Le Sueur

1

NY

7

A. fulvescens Rafinesque

3

CHI

23/3

A. ruthenus Linnaeus

6

AMS

69/8

A . sturio Linnaeus

4

AMS

9/11

Scaphyrhynchus platorhynchus (Rafinesque)

1

DAL

13*

Polyodontidae— Paddlefishes

Polyodon spathula (Walbaum)

1

CHI

5/2

Amiidae— Bowfins

Amia calva Linnaeus

3

NY

30

LEPISOSTEIDAE-GarS

Lepisosteus osseus (Linnaeus)

6

NY

30

L. platostomus Rafinesque

5

NY

20

L. productus (Cope)
L. spatula Lacepede

1

TOL

8/3

5

CHI

23/4*

Megalopidae— Tarpons

Tarpon atlanticus (Valenciennes)

2

CHI

19/10*

Salmonidae— T routs

Coregonus clupeaformis (Mitchill)

2

NY

12

Hucho hucho (Linnaeus)

1

MUN

10

Salmo gairdnerii Richardson

1

NY

5

S. trutta fario Linnaeus

1

NY

5

S. trutta lacustris Linnaeus

1

LON(F)

10/4

Salvelinus fontinalis (Mitchill)

1

NY

5

5. malma (Walbaum)

1

SHA

19

Esocidae— Pikes

Esox lucius Linnaeus

2

LON(F)

10

E. masquinongy Mitchill

1

NY

10

N otopteridae-N otopterids

Notopterus notopterus (Pallas)

1

AMS

9/4

Osteoglossidae— Osteoglossids

Osteoglossum bicirrhosum Vandelli

1

CHI

6/3

Pantodontidae— Pantodon

Pantodon bucholzi Peters

2

AMS

6/9

Mormyridae— Mormyrids

Marcusenius isidori Valenciennes

1

FLOW

28/11

Mormyrus kannume Forskal

1

CHAR

16/3

Characidae— Characins

Astyanax bimaculatus (Linnaeus)

1

AMS

18

A. fasciatus mexicanus (Filippi)

1

LON(F)

6/9

Chalceus macrolepidotus Cuvier

2

MUN

19*

Colossoma nigripinnis (Cope)

1

MUN

6*

Copeina guttata (Steindachner)

1

AMS

6/6

Corynopoma riisei Gill

1

AMS

13

Ctenobrycon spilurus (Valenciennes)

2

AMS

14/7

Exodon paradoxus Muller & Troschel

1

CHAR

6/8

110

Zoologica: New York Zoological Society

[47: 9

Longevity of Fishes in Captivity as of September, 1956 (Continued)
(Limited to reported life of five years or more.)

Family and Species

Gymnocorymbus ternetzi (Boulenger)
Hemigrammus caudovittatus Ahl
H. ocellifer (Steindachner)

H. uniliniatus Gill
Hemiodus semitaeniatus Kner
Hepsetus odeo (Bloch)

Hyphessobrycon bifasciatus Ellis
Leporinus fasciatus (Bloch)

L. friderici Bloch

L. megalepis Gunther
Metynnis maculatus (Kner)

M. roosevelti Eigenmann
M. schreitmuelleri Ahl
Moenkhausia oligolepis (Gunther)

Myleus asterias (Muller & Troschel)

M. gurupyensis Steindachner

M. schomburgkii (Jardine)

Mylossoma duriventre (Cuvier)

Nannaethiops unitaeniatus Giinther
Pristella riddlei (Meek)

Prochilodus insignis Schomburgk
Serrasalmus nattereri (Kner)

Serrasalmus niger (Schomburgk)

S. rhombeus (Linnaeus)

S. spilopleura Kner
Thoracocharax stellatus (Kner)
Electrophoridae— Electric Eel
Electrophorus electricus (Linnaeus)
Catostomidae— Suckers

Ictiobus bubalus (Rafinesque)

Ictiobus cyprinella (Valenciennes)
Cyprinidae— Minnows

Abramis brama (Linnaeus)

Aphyocypris pooni Lin
Balantocheilus melanopterus Bleeker
Blicca bjoerkna (Linnaeus)

Brachydanio albolineatus (Blyth)

B. nigrofasciatus (Day)

B. rerio (Hamilton-Buchanan)

Carassius auratus (Linnaeus)

Cyclocheilichthys apogon (Valenciennes)
Cyprinus carpio Linnaeus
Danio malabaricus (Jerdon)

Epalzeorhynchus kalopterus (Bleeker)

Leuciscus cephalus (Linnaeus)

L. idus (Linnaeus)

N otemigonus crysoleucas crysoleucas (Mitchill)
Osteocheilus hasseltii (Valenciennes)

O. vittatus (Valenciennes)

Puntius chola (Hamilton-Buchanan)

P. cumin gi (Gunther)

P. dunkeri (Ahl)

P. everetti (Boulenger)

P. lateristriga (Valenciennes)

P. nigrofasciatus (Gunther)

P. oligolepis (Bleeker)

P. schwanfeldi (Bleeker)

P. semifasciolatus (Gunther)

P. ticto (Hamilton-Buchanan)

P. titteya Deraniyagala

A

B

C

2

AMS

8/2

1

AMS

12/2

1

NZ

5

1

LON(F)

5/2

2

CHI

16/1

1

MUN

6*

1

AMS

9

3

MUN

18

3

MUN

18

2

CHAR

18/1

1

LON(F)

11/11

3

MUN

17/6*

1

WUP

6/4*

1

AMS

14/6

3

MUN

18*

1

MUN

18*

1

MUN

18*

2

MUN

18*

1

AMS

6/2

1

AMS

9/2

2

WUP

6/10*

2

MUN

17

1

CHI

21/4*

1

TOL

19

1

CHAR

7/9

1

CHAR

9/5

3

AMS

12/1

1

TOL

16/4

1

TOL

16/4

2

AMS

15/9

1

AMS

7/8

1

AMS

11/11

1

AMS

16/10

1

AMS

7/6

1

AMS

7/1

1

AMS

8

7

NY

10

1

AMS

6/2

10

FRANK

38

1

AMS

7/10

1

CHAR

12

2

AMS

7

3

AMS

21/7

1

NY

7

1

MUN

18

1

MUN

6

1

AMS

11/4

1

AMS

6/5

1

MUN

11

2

MUN

11

2

AMS

18/5

1

AMS

7

1

AMS

8/1

1

AMS

11/3

1

AMS

18/8

1

AMS

12/10

1

AMS

6/7

1962]

Hinton: Longevity of Fishes in Captivity

111

Longevity of Fishes in Captivity as of September, 1956 (Continued)
(Limited to reported life of five years or more.)

Family and Species

A

B

C

Rasbora einthovenii (Bleeker)

1

AMS

10/11

R. heteromorpha Duncker

2

LON(F)

6/9

R. trilineata Steindachner

1

AMS

6/10

Rhodeus sericeus (Pallas)

2

AMS

10

Rhinichthys atratulus atratulus (Hermann)

1

NY

5

Rutilus rutilus (Linnaeus)

1

AMS

9/3

Scardinius erythrophthalmus (Linnaeus)

5

CHI

20/3*

Tinea tinea (Linnaeus)

1

CHAR

7/11

Cobitidae— Loaches

Botia hymenophysa (Bleeker)

1

AMS

6

Cobitis taenia Linnaeus

2

FLOW

10/1

Nemacheilus barbatulus (Linnaeus)

1

AMS

8/3

Ariidae— Sea Catfishes

Galeichthys felis (Linnaeus)

2

PHIL

15/3*

Doradidae— Doradid Catfishes

Doras sp.

1

AMS

26/5

Seluridae— Sheatfishes

Silurus glanis Linnaeus

3

CHAR

11/6

Ictaluridae— North American Catfishes

lctalurus nebulosus (Le Sueur)

1

AMS

13/6

I. punctatus (Rafinesque)

2

TOL

8/5

Pylodictis olivaris (Rafinesque)

2

TOL

16/1

Chacidae— Chacid Catfishes

Chaca chaca (Hamilton-Buchanan)

1

AMS

6/4

SACCOBRANCHiDAE-Airbreathing Catfish

Saccobranchus fossilis (Bloch)

4

AMS

12/11

Clariidae— Clariid Catfishes

Clarius batrachus Valenciennes

2

AMS

13/2

Synodontidae— Upside-down Catfishes

Synodontis schal (Bloch & Schneider)

2

FLOW

31

Malapteruridae— Electric Catfish

Malapterurus electricus (Linnaeus)

5

CHI

8/1

Pimelodidae— South American Catfishes

Phractocephalus hemeliotropus (Bloch & Schneider)

1

CHI

12/2

Pimelodella gracilis (Valenciennes)

2

AMS

11/11

Pimelodus clarias (Bloch)

1

AMS

27/1

Rhamdia sebae (Valenciennes)

2

AMS

30/8

Callichthyidae— Mailed Catfishes

Corydoras aeneus Gill

2

MUN

12

C. julii Steindachner

1

WUP

6/5

Loricariidae— Armored Catfishes

Loricaria microlepidogaster Regan

1

AMS

7/7

Plecostomus commersonii (Valenciennes)

3

MUN

19/7

P. punctatus (Valenciennes)

2

MUN

18

P. rachovii Regan

1

MUN

18

Anguillidae— Freshwater Eels

Anguilla anguilla (Linnaeus)

6

MUN

19

A. rostrata (Le Sueur)

2

DAL

17*

Muraenidae— Morays

Gymnothorax funebris Ranzani

2

CHI

20/4*

G. mordax Ayres

3

SIO

26

Muraena clepsydra Gilbert

1

SIO

6*

M. helma Linnaeus

4

AMS

20/3

Congridae— Congers

Conger conger (Linnaeus)

2

RHO

9*

Gadidae— Cods

Lota lota lacustris (Walbaum)

1

NY

5

Pollachius pollachius (Linnaeus)

1

PLY

5

Cyprinodontidae— Killifishes

Aplocheilus lineatus (Valenciennes)

1

AMS

7/7

112

Zoologica: New York Zoological Society

[47: 9

Longevity of Fishes in Captivity as of September, 1956 (Continued)
(Limited to reported life of live years or more.)

Family and Species

A

B

C

Epiplatys chaperi (Sauvage)

1

AMS

9/10

P oeciliidae — Livebearers

Limia nigrofasciata (Regan)

1

AMS

5/7

Zeidae— Dories

Zeus faber Linnaeus

1

PLY

5

Mugilidae— Mullets

Liza auratus Risso

1

RHO

9

L. capito Cuvier

2

AMS

14

L. chelo Cuvier

2

PLY

23

Mugil cephalus Linnaeus

2

RHO

9

M. rammelsbergii Tschudi

1

AMS

14

Atherinidae— Silversides

Melanotaenia maccullochi Ogilby

1

AMS

7/7

M. ni grans (Richardson)

1

AMS

17/7

Channidae— Snakeheads

Channa striata Bloch

1

AMS

5/10

Ophicephalus obscurus GUnther

1

CHAR

11/7

Latidae— Nile Perches

Lates niloticus (Gmelin)

1

FLOW

12

Serranidae— Basses

Centropristis striatus (Linnaeus)

1

PHIL

5

Dicentrarchus labrax (Linnaeus)

4

AMS

30

Epinephalus aeneus (Geoffroy)

1

RHO

8/5*

E. adscensionis (Osbeck)

1

PHIL

6/11

E. analogus Gill

1

SIO

5*

E. chrysotaenia Doderlein

1

RHO

8/4

E. gigas (Briinnich)

2

MON

29/2

E. guaza (Lacepede)

1

RHO

9/1*

E. guttatus (Linnaeus)

3

PHIL

14/10*

E. itajara (Lichtenstein)

1

NY

12

E. labriformis (Jenyns)

1

SIO

5*

E. morio (Valenciennes)

1

NY

7

E. sp.

1

MON

7/1

E. striatus (Bloch)

3

PHIL

9/1

E. summana Forskal

1

CHAR

8

Morone americana (Gmelin)

1

NY

7

Mycteroperca bonaci (Poey)

1

PHIL

7/11

M. tigris (Valenciennes)

1

NY

5

M. venenosa (Linnaeus)

1

NY

5

Paralabrax clathratus (Girard)

2

SIO

12

P. maculatofasciatus (Steindachner)

1

SIO

8*

P. nebulifer (Girard)

1

SIO

8*

Roccus chrysops (Rafinesque)

1

DAL

6

R. saxatilis (Walbaum)

3

NY

24

Serranus cabrilla (Linnaeus)

1

RHO

7/5

S. scriba (Linnaeus)

1

RHO

7/5

Theraponidae— Theraponids

Therapon jarbua (Forskal)

1

LON(F)

7/7

Kuhliidae— Kuhliids

Kuhlia taeniura (Cuvier)

1

AMS

8/9

K. sandvicensis (Steindachner)

1

SF

7*

Centrarchidae— Sunfishes

Ambloplites rupestris rupestris (Rafinesque)

2

NY

18

Centrarchus macropterus (Lacepede)

2

AMS

16/9

Enneacanthus chaetodon (Baird)

1

AMS

7/1

E. gloriosus (Holbrook)

1

AMS

11

Lepomis gibbosus (Linnaeus)

3

AMS

16

L. megalotis (Rafinesque)

1

LON(F)

6/4

Micropterus dolomieui Lacepede

2

NY

11

M. salmoides salmoides (Lacepede)

3

TOL

13/5

1962]

Hinton: Longevity of Fishes in Captivity

113

Longevity of Fishes in Captivity as of September, 1956 (Continued)
(Limited to reported life of five years or more.)

Family and Species

A

B

I C

Pomoxis nigromaculatus (Le Sueur)

1

NY

12

Percidae— Perches

Perea ftavescens (Mitchill)

1

NY

12

P. fluviatilis Linnaeus

2

LON(F)

10/8

Stizostedion lucioperca (Linnaeus)

2

LON(F)

9/1

Carangidae— J acks

Caranx crysos (Mitchill)

1

PHIL

5/11

C. hippos (Linnaeus)

1

NY

5

Selene vomer (Linnaeus)

1

NY

5

Lutjanidae— Snappers

Lutjanus jocu (Bloch & Schneider)

2

NY

14/7

L. griseus (Linnaeus)

1

NY

7

L. synagris (Linnaeus)

1

NY

7

L. argentiventris (Peters)

1

SF

11*

Pomadasyidae— Grunts

Anisotremus davidsonii (Steindachner)

2

SF

9*

Haemulon sciurus (Shaw)

1

NY

5

Orthopristis chrysopterus (Linnaeus)

1

PHIL

5/6

Sciaenidae— Croakers

Aplodonotus grunniens Rafinesque

1

TOL

16/5

Corvina nigra (Bloch)

1

MON

11/11

Cy noscion regalis (Bloch & Schneider)

1

NY

5

C. xanthulus Jordan & Gilbert

1

SIO

5/2

Micropogon undulatus Linnaeus

1

PHIL

6/2

Pogonias cromis (Linnaeus)

2

PHIL

10/9

Roncador stearnsi (Steindachner)

2

HERM

6

Umbrina roncador Jordan & Gilbert

2

SF

6

Sciaenops ocellata (Linnaeus)

2

PHIL

12/1

Sparidae— Porgies

Archosargus probatocephalus (Walbaum)

1

NY

6

Dentex dentex (Linnaeus)

2

RHO

8

Diplodus annularis (Linnaeus)

2

RHO

9

D. sargus (Linnaeus)

2

RHO

9

Oblada melanura (Linnaeus)

1

RHO

9

Pagellus centrodontus (de la Roche)

1

PLY

7

P. mormyrus (Linnaeus)

1

RHO

9

P. pagrus (Linnaeus)

2

RHO

7

Puntazzo puntazzo (Cetti)

1

RHO

9

Sparus auratus Linnaeus

2

RHO

8

Spondyliosoma cantharus (Linnaeus)

3

MON

8

Monodactylidae— Monodactylids

Monodactylus argenteus (Linnaeus)

2

MUN

5/6

Toxotidae— Archerfishes

Toxotes jaculatrix (Pallas)

2

SF

6/5*

Scorpidae— Halfmoons

Medialuna californiensis (Steindachner)

1

HERM

6

Girellidae— N ibblers

Girella nigricans (Ayres)

2

HERM

7

Ephippidae— Spadefishes

Chaetodipterus faber (Broussonet)

1

NY

5

Platax orbicularis (Forskal)

2

MON

7

P. teira (Forskal)

1

AMS

6/5

SCATOPHAGIDAE-ScatS

Scatophagus argus (Linnaeus)

3

CHAR

9/3

Chaetodontidae— Butterflyfishes

Holacanthus isabelita (Jordan & Rutter)

1

NY

5

Nandidae— Nandids

Badis badis (Buchanan-Hamilton)

1

AMS

7/6

Pristolepis grooti (Bleeker)

1

AMS

8/1

114

Zoologica: New York Zoological Society

[47: 9

Longevity of Fishes in Captivity as of September, 1956 (Continued)
(Limited to reported life of five years or more.)

Family and Species

A

B

C.

Cichlidae— Cichlids

Astronotus ocellatus (Cuvier)

4

WUP

17/3

Aequidens pulcher (Gill)

1

AMS

5/3

Cichlasoma bimaculatum (Linnaeus)

1

AMS

11/10*

C. biocellatum Regan

1

AMS

6/10

C. face turn (Jenyns)

1

AMS

10/4

C. hellabrunni Ladiges

2

WUP

9/4

C. maculicauda Regan

1

AMS

8/11

C. nigrofasciatum (Gunther)

1

LON(F)

9

C. severum (Heckel)

3

AMS

10/8

Hemichromis bimaculatus Gill

1

AMS

14/6

Herichthys cyanoguttatiis carpintis (Jordan & Snyder)

1

LON(F)

5/6

Pterophyllum eimekei Ahl

2

AMS

5/4

Symphysodon discus Heckel

1

WUP

5/4*

Tilapia macrocephala (Bleeker)

1

AMS

8/6

T. mossambica (Peters)

1

NY

6

T. natalertsis (Weber)

1

AMS

12/7

T. nilotica (Linnaeus)

1

AMS

11/11

T. zilli (Gervais)

1

CHAR

5/1

Pomacentridae— Damselfishes

Amphiprion ephippium (Bloch)

1

CHAR

16/6

Chromis chromis (Linnaeus)

1

RHO

6/6

Dascyllus aruanus (Linnaeus)

1

CHAR

5/11

Hypsipops rubicunda (Girard)

1

SIO

7

Premnas biaculeatus (Bloch)

1

CHAR

8/11

Labridae— Wrasses

Coris julis (Linnaeus)

1

RHO

7

Pimelometopon pulchrum (Ayres)

3

SF

23

Labrus bergylta Ascanius

1

PLY

5

L. festivus Risso

1

RHO

8

Tautoga onitis (Linnaeus)

2

NY

8

Scaridae— Parrotfishes

Scarus cretensis (Bloch)

1

RHO

7

Blenniidae— Blennies

Blennius pholis Linnaeus

1

CHAR

5

Siganidae— Siganids

Siganus chrysospilos (Bleeker)

1

AMS

7/5

Anabantidae— Climbing Perches

Anabas testudineus (Bloch)

5

DAL

16

Belontia signata (Gunther)

2

LON(F)

10/7

Colisa lalia Hamilton-Buchanan

1

AMS

6/11

Ctenops vittatus (Valenciennes)

1

AMS

6/11

Osphrenemus goramy Lacepede

1

AMS

8/11

Polyacanthus hasselti Valenciennes

1

AMS

10/10

Trichogaster leerii (Bleeker)

1

AMS

5/8

Eleotridae— Sleepers

Dormitator maculatus (Bloch)

1

AMS

14/2

Eleotris vittata Dumeril

1

CHAR

11/10

Oxyeleotris marmorata (Bleeker)

1

AMS

20/5*

Scorpaenidae— Scorpionfishes

Pterois volitans (Linnaeus)

1

MON

10/4

Scorpaena guttata guttata Girard

1

SIO

10

S. porcus Linnaeus

2

RHO

9

S. scrofa Linnaeus

1

RHO

8

Triglidae— Gurnards

Trigla lucerna Linnaeus

3

LON(F)

6/9

Scophthalmidae— Turbots

Psetta maxima (Linnaeus)

2

AMS

10/1

Bothidae— Lefteyed Flounders

Zeugopterus punctatus (Bloch)

1

CHAR

5/5

1962]

Hinton: Longevity of Fishes in Captivity

115

Longevity of Fishes in Captivity as of September, 1956 (Continued)
(Limited to reported life of five years or more)

Family and Species

A

B

C.

Pleuronectidae— Righteyed Flounders
Platichthys flesus (Linnaeus)

1

RHO

5/2

Soleidae— Soles
So lea sole a (Linnaeus)

2

RHO

5/5

Mastacembelidae— Mastacembelids
Mastacembelus erythrotaenia Bleeker

1

AMS

10/11

Rhynchobdella aculeata (Bloch)

1

AMS

9

Balistidae— Triggerfishes
Batistes capriscus Gmelin

4

RHO

9

B. vetula Linnaeus

1

PHIL

5/1

Lagocephalidae— Puffers

Arothron reticularis (Bloch & Schneider)

1

MON

7/1

T ETRAODONTIDAE— Puff ers

Amblyrhynchotes honckenii (Bloch)

1

LON(F)

5/5

Tetraodon lineatus (Linnaeus)

2

CHAR

9/3

T. fluviatilis (Hamilton-Buchanan)

3

CHAR

9/3

Canthigasteridae— Puffers

Canthigaster margaritatus Ruppell

1

LON(F)

6/11

Diodontidae— Porcupinefishes
Diodon hystrix (Linnaeus)

1

PHIL

7/4

B atrachoidae— T oadfishes
Opsanus tau (Linnaeus)

3

CHI

13/11

Literature Cited

Anonymous

1949. Venerable trout dies. Aquarium Jour.
[San Francisco], 20 (4): 107.

Baird, Spencer F.

1872. On the death of an aged carp. Ann. Rec.
Sci., 1872-73: 262.

Baldinger, E. G.

1872(7). Sur l’age d’un brochet. N. Med. Physik
Journal, Marburg, 5 (1): 29.

Berg, Leo S.

1940. Classification of fishes, both recent and
fossil. Trav. Inst. Zool. Acad. Sci. USSR,
V: 2, pp 87-517. (Reprinted by J. W.
Edwards, Ann Arbor, Michigan, 1947. A
posthumus, revised edition issued in 1955,
as vol. XX in the same series, maintains
the identical sequence of families).

Bertin, Leon.

1956. “Eels: a biological study.” Cleaver-Hume
Press, Ltd., London.

Bourliere, F.

1946. Longevite moyene et longevite maximum
chez les vertebres. L’Annee Biologique, 3
ser., 22: 248-270. (Poissons, pp. 253-254).

Breder, Charles M., Jr.

1936. Long-lived fishes in the aquarium. Bull.
N. Y. Zool. Soc., 39 (3): 116-117.

Breland, Osmond P.

1952. How long do they live? Natural History,
61 (1): 131-134, 141-142.

Brown, Margaret E.

1957. Experimental studies on growth. Chapter
IX in “The physiology of fishes.” Vol. I:
361-400. Academic Press, Inc., New York.

Comfort, Alex

1956. “The biology of senescence.” Routledge
and Kegan Paul, London.

DeSylva, Donald P.

1957. Studies on the age and growth of the
Atlantic sailfish, Istiophorus americanus
Cuvier, using length-frequency curves.
Bull. Marine Sci. Gulf and Caribbean, 7
(1): 1-20. (Contribution 120, Marine
Laboratories, University of Miami).

Flower, Stanley S.

1925. Contributions to the knowledge of the
duration of life in vertebrate animals. I.
Fishes. Proc. Zool. Soc. London, 1925:
247-268.

1935. Further notes on duration of life in ver-
tebrate animals. I., Fishes. Proc. Zool. Soc.
London, 1935: 265-304.

Gandolfi-Hornyold, A.

1932. Observations sur l’anguille du Caumasee.
Bull. Suisse Peche Pise., 23. (Not seen:
quoted by Bertin, 1956).

1935. La longevite de l’anguille en liberte et en
captivite. Bull. Soc. Nat. Acclim., 76.

116

Zoologica: New York Zoological Society

[47: 9: 1962]

Gerking, Shelby D.

1959. Physiological changes accompanying age-
ing in fishes. Ciba Foundation Colloquia
on Ageing, 5: 181-208.

Leach, Glen C.

1924. Artificial propagation of brook trout and
rainbow trout, with notes on three other
species. Appendix VI, Rpt. U. S. Comm.
Fish., 1923. (Doc. No. 955): 74 pp.

Macleod, A.

1949. The oldest eel in the world. Scottish Zoo
and Wildlife, 2 (1): 33-35.

Mehwald.

1873. Ueber einen 375 Jahre alten Karpfen.
Sitzber. Natrw. Ges. Isis, Dresden, 1872:
109. (Not seen: quoted by Baird, 1872).

Mellen, Ida M.

1918. Complete list of exhibits at the New York
Aquarium. 23rd Annual Report N. Y.
Zool. Soc. for 1917: 94-117.

Nigrelli, Ross F.

1959. Longevity of fishes in captivity, with spe-
cial reference to those kept in the New
York Aquarium. Ciba Foundation Collo-
quia on Ageing. 5: 212-226.

Perkins, C. B.

1947. A note of longevity of amphibians and
reptiles in captivity. Copeia, 1947 (2):
144.

Rasquin, Priscilla, & Ethel Hafter

1951. Age changes in the testis of the teleost,
Astyanax mexicanus. Jour. Morph., 89
(3): 397-408.

Shaw, Charles E.

1958. Longevity of snakes in the United States
as of January 1, 1958. Copeia, 1958 (3):
221.

Simpson, Donald A.

1957. When you live in Steinhart. Aquarium
Jour. [San Francisco], 28 (8) : 294-296.

Suffield, Robert R.

1874. Longevity of the carp. Nature, 10: 147,
165.

Townsend, Charles H.

1904- Longevity of fishes at the aquarium. (Un-

1913. signed articles) Bull. N. Y. Zool. Soc.:
163, 416, 971-972, 1049.

1928a. The public aquarium. Appendix VII, Rpt.
U. S. Comm. Fish., 1928 (Doc. No.
1045): 249-337.

1928b. Notes on aquarium equipment, methods
and problems. Bull. N. Y. Zool. Soc., 31
(1): 11-25.

Trautman, Milton B., & Carl L. Hubbs

1936. When do pike shed their teeth? Trans.
Amer. Fish. Soc., 65: 261-266.

10

Hybridization Experiments in Acheilognathine Fishes (Cyprinidae,
Teleostei). A Comparison of the Intergeneric Hybrids
between Tanakia tanago and Rhodeus spinalis and
Rhodeus ocellatus from Korea and Japan

J. J. Duyvene de Wit

Zoology Department, University of the Orange Free State,

Bloemfontein,
(Plates I & II;

IN a recent paper the present author
(Duyvene de Wit, in press) has summar-
ized the results of his investigations into the
presence of gametic compatibility between a
number of acheilognathine1 species.

From Text-fig. 1, in which the ability of these
species to cross is represented diagramatically,
it appears that gametic compatibility is not lim-
ited to the subspecific and specific levels, but also
manifests itself on the generic level. In the ma-
jority of the intergeneric crosses, the Fi offspring
consisted of phenotypical males only, but the
intergeneric hybrids between the allopatric spe-
cies Tanakia tanago (Tanaka) from Japan
and Acheilognathus himantegus (Gunther)
from Taiwan consisted of fertile males and fe-
males. These fishes were allowed to interbreed
freely in the presence of freshwater mussels and
this experimental interbreeding population has
now been carried through its fourth generation.
The phenotypic character of all four generations
has remained stable.

In another paper (Duyvene de Wit, 1961),
we have expressed the view that Rhodeus spin-
alis (Oshima) from Taiwan and the two strains
of Rhodeus ocellatus (Kner), which have their
habitats in Japan and Korea respectively, repre-
sent three allopatric geographic subspecies be-
longing to a single species. As will be seen from
Text-fig. 1, the hybrids produced from these
strains consisted of fertile males and females.

following Chu (1935) and Hubbs & Kuronuma
(1943), the designation Acheilognathinae instead of
Rhodeinae is used, because Acheilognathus (Bleeker,
1863) has priority over Rhodeus (Gunther, 1868).

South Africa
Text-figure 1)

When comparing the phenotypes of these
three strains, slight differences between both the
Korean and the Japanese varieties of R. ocellatus
are apparent, while the Taiwan species, R. spin-
alis, is fairly similar to both strains of R. ocell-
atus. In order to investigate whether the genomes
of the three supposed subspecies would express
themselves differently in the Fi generations
when crossed with a single species of bitterling
belonging to another genus, we mated them with
the Japanese species Tanakia tanago. In Table
1, the six possible breeding combinations have
been designated 1 to 6.

Table 1.

| T. tanago $

T. tanago $

R. ocellatus (Japan) 2

1

R. ocellatus (Japan) $ 2

R. ocellatus (Korea) $

3

R. ocellatus (Korea) $ 4

R. spinalis (Taiwan) 2

5

R. spinalis (Taiwan) $ 6

Material and Method

All crossings were performed by means of arti-
ficial insemination. The number of larvae pro-
duced by the six crosses ranged from 8 to 20, but
some died in the course of development as a
result of infectious diseases. The following re-
sults are based on the hybrids that reached the
adult stage.

117

118

Zoologica: New York Zoological Society

[47: 10

Text-fig. 1. Diagram showing the gametic affinities between a number of acheilognathine species from
W. Europe, Korea, Japan and Taiwan, on the subspecies, species and generic levels, according to Duyvene
de Wit [in press].

Results

1. R. ocellatus (Japan) 2 X Tanakia tanago $
This combination yielded eight hybrids of

different body size and shape. Two of them
were large and fairly similar to the maternal
species, R. ocellatus (Japan), four were small
and showed greater similarity to the paternal
species, T. tanago, while two were intermediate.
All hybrids showed a male phenotype. During
the spawning season they displayed full nuptial
colors. Tubercles were present on the top of the
snout. No milt production could be detected by
stripping, however.

Representative specimens from the large and
the small form of hybrid are illustrated in Figs.
3 and 4 respectively. The maternal species, R.
ocellatus from Japan, is illustrated in Fig. 1 and
the paternal species, T. tanago, in Fig. 2.

2. T. tanago 2 X R. ocellatus (Japan) $

From this, the reciprocal combination, five

hybrids were obtained. They also were not uni-
form in appearance. Three of them were large
and in body size and shape tended to resemble
the paternal species, R. ocellatus (Japan). The
remaining two hybrids were medium-sized. One
of them tended to resemble the maternal species.

T. tanago, while the other was more similar to
the paternal species, R. ocellatus (Japan). All
hybrids showed a male phenotype. Nuptial
colors and tubercles on top of the snout were
present during the spawning season, but no milt
could be obtained by stripping.

One of the large hybrids is illustrated in Fig. 7.
A specimen of the maternal species, T. tanago,
is illustrated in Fig. 5, and one of the paternal
species, R. ocellatus from Japan, in Fig. 6.

It should be noted that the hybrids of the
reciprocal combinations are different in general
appearance.

3. R. ocellatus (Korea) 2 XT. tanago $

From this combination six hybrids were ob-
tained. They were relatively small and fairly
uniform in size and body shape. In general ap-
pearance and with respect to their small size
they were much more similar to the paternal
species, T. tanakia, than to the maternal species,
R. ocellatus (Korea) . All hybrids showed a male
phenotype. During the spawing season they dis-
played bright nuptial colors and tubercles were
present on the top of the snout. Milt production
could not be detected by means of stripping.

A representative hybrid specimen is illustrated

1962]

De Wit: Hybridization Experiments in Acheilognathine Fishes

119

in Fig. 9. The maternal species, R. ocellatus from
Korea, is illustrated in Fig. 8.

4. T. tanago $ X R. ocellatus (Korea) $

From this cross five hybrids were obtained.

Four of them were fairly large. With respect to
body size and shape they tended to resemble the
paternal species, R. ocellatus (Korea). One of
these showed a coin-like body shape, an abnor-
mality that is also sometimes encountered in the
pure-bred strain of R. ocellatus. The remaining
hybrid was much smaller than its brood mates.
All hybrids showed, a male phenotype. They dis-
played full nuptial colors during the spawning
season and tubercles were present on the top
of the snout. No milt production could be de-
tected by stripping.

A representative hybrid specimen is illustra-
ted in Fig. 11. The paternal species, R. ocellatus
from Korea, is illustrated in Fig. 10.

In their general appearance the hybrids of the
combinations 3 and 4 differ considerably from
each other.

5. R. spinalis $ X T. tanago 9

From this combination fourteen hybrids were
obtained. With respect to body size they showed
a gradual transition from large to very small.
The largest specimens tended to resemble the
maternal species, although the dorsal part of the
body was less curved than in R. spinalis. The
smaller ones were more similar to the paternal
species, T. tanago. Except the three smallest
specimens, which probably were neuters, the
hybrids showed a male phenotype. During the
spawning season they displayed full nuptial
colors, and tubercles were present on the top
of the snout. No milt could be obtained by
stripping.

Representative specimens of the large and
the small hybrids are illustrated in Fig. 13. The
maternal species, R. spinalis, is illustrated in
Fig. 12.

6. T. tanago 9 'X R. spinalis $

From this combination, five hybrids were ob-
tained. They were fairly uniform in size and
body shape and tended to resemble the paternal
species in that the dorsal part of the body was
curved as in R. spinalis. All of them showed a
male phenotype. During the spawning season
they displayed full nuptial colors. Tubercles were
present on the top of the snout, but milt could
not be obtained by stripping.

A representative specimen of the present hy-
brids is illustrated in Fig. 15. The paternal spe-
cies, R. spinalis, is illustrated in Fig. 14.

The hybrids of combinations 5 and 6 show
considerable difference in general appearance.

Discussion

The question whether the genomes of the
three supposed geographical subspecies of R.
ocellatus under investigation express themselves
differently in the six kinds of hybrids that could
be produced by crossing them with a single test
species, T. tanago, must be answered in the
affirmative.

In their general appearance the hybrids of the
combinations 1 + 2, 3 + 4 and 5 + 6 appeared
to differ considerably from one another. More-
over, the hybrids obtained from three hybrid
combinations (1, 3 and 5) were different from
their respective reciprocal ones (2, 4 and 6).
Finally, the six separate Fi generations showed
no uniformity among their brood mates. These
findings are in sharp contrast to the fact that the
progenies obtained from R. ocellatus (Japan)
X R. ocellatus (Korea), R. ocellatus (Japan)
X R. spinalis (Taiwan) and R. ocellatus
(Korea) X R. spinalis (Taiwan) were largely
uniform in body form and size.

On the other hand, all present hybrid genera-
tions correspond in showing the male pheno-
type, by displaying nuptial colors during the
spawning season and in the absence of milt pro-
duction.

The lack of uniformity in the six Fi genera-
tions probably points to incomplete genetic com-
patibility and presents an interesting problem for
future karyological research.

The taxonomic status of the six kinds of hy-
brids under discussion here will be published
separately.

Summary

1. Intergeneric hybrids have been obtained
from the combinations Rhodeus ocellatus (Jap-
anese strain) X Tanakia tanago, Rhodeus ocell-
atus (Korean strain) X Tanakia tanago and
Rhodeus spinalis X Tanakia tanago.

2. All hybrids showed the male phenotype.
During the spawning season they displayed full
nuptial colors, but milt production could not be
detected by means of stripping.

3. The six kinds of hybrids obtained from the
three combinations and their reciprocals differed
considerably from one another, and the varia-
bility among the individuals of the separate Fi
generations was conspicuous. This is interpreted
as an expression of incomplete genetic compati-
bility between the respective male and female
gametes.

Acknowledgements

The author is greatly indebted to Prof.
Tokiharu Abe and Dr. Yoshitsugu Hirosaki for
supplying the Japanese acheilognathine species,

120

Zoologica: New York Zoological Society

[47: 10: 1962]

to Prof. Moon-ki Chyung and Mr. Kyun Hyun
Kim for supplying the Korean acheilognathine
species, to Professor Fah-hsuen for supplying
the Taiwan acheilognathine species, to Mr. F. G.
de Jardin and Mr. R. A. Cragg for providing
photographs of the illustrated specimens, and
to Mr. C. A. van Ee for the drawing of the dia-
gram. The financial assistance of the Board of
the University of the Orange Free State and of
the Council for Scientific and Industrial Re-
search of the Republic of South Africa is grate-
fully acknowledged.

References

Chu, Y. T.

1935. Comparative studies on the scales and on
the pharyngeals and their teeth in Chinese
cyprinids with particular reference to

taxonomy and evolution. Biol. Bull. St.
John’s Univ., 2, 1-223.

Duyvene de Wit, J. J.

1961. Hybridization experiments in rhodeine
fishes (Cyprinidae, Teleostei). The intra-
specific hybrids of Rhodens ocellatus from
Japan and Korea, and Rhodeus spinalis
from Taiwan. Can. J. Zool., 39, 487-490.

[In press] Hybridization experiments in acheilog-
nathine fishes (Cyprinidae, Teleostei).
The hybrids between female Acheilogna-
thus lanceolatus and Rhodeus amarus,
and male Acheilognathus himantegus.
Can. J. Zool.

Hubbs, C. L., & K. Kuronuma

1943. Egg and ovipositor characters in two
acheilognathine fishes from Japan. Copeia,
3, 183-186.

EXPLANATION OF THE PLATES
Plate I

Fig. 1. Female specimen of Rhodeus ocellatus
from Japan. Standard length 44 mm.

Fig. 2. Male specimen of Tanakia tanago. Stand-
ard length 42 mm.

Fig. 3. Specimen of a large hybrid obtained from
the combination female Rhodeus ocellatus
(Japan) X male Tanakia tanago. Stand-
ard length 61 mm.

Fig. 4. Specimen of a small hybrid obtained from
the combination as indicated in the legend
of Fig. 3. Standard length 37 mm.

Fig. 5. Female specimen of Tanakia tanago.
Standard length 35 mm.

Fig. 6. Male specimen of Rhodeus ocellatus from
Japan. Standard length 69 mm.

Fig. 7. Specimen of a hybrid obtained from the
combination female Tanakia tanago X
male Rhodeus ocellatus (Japan). Stand-
ard length 60 mm.

Fig. 8. Female specimen of Rhodeus ocellatus
from Korea. Standard length 44 mm.

Fig. 9. Specimen of a hybrid obtained from the
combination female Rhodeus ocellatus
(Korea) X male Tanakia tanago. Stand-
ard length 37 mm.

Plate II

Fig. 10. Male specimen of Rhodeus ocellatus from
Korea. Standard length 61 mm.

Fig. 11. Specimen of a hybrid obtained from the
combination female Tanakia tanago X
male Rhodeus ocellatus (Korea). Stand-
ard length 51 mm.

Fig. 12. Female specimen of Rhodeus spinalis.
Standard length 39 mm.

Fig. 13. Specimen of a large hybrid obtained from
the combination female Rhodeus spinalis
X male Tanakia tanago. Standard length
54 mm. Left corner: specimen of the very
small hybrid form. Standard length 25 mm.

Fig. 14. Male specimen of Rhodeus spinalis. Stand-
ard length 46 mm.

Fig. 15. Specimen of a hybrid obtained from the
combination female Tanakia tanago X
male Rhodeus spinalis. Standard length
45 mm.

DE WIT

PLATE I

FIG. I

FIG. 3

FIG. 6

FIG. 7

FIG. 9

FIG. 2

FIG. 4

FIG. 5

FIG. 8

INTERGENERIC HYBRIDS BETWEEN TANAKIA TANAGO AND RHODEUS SPINALIS
AND RHODEUS OCELIATUS FROM KOREA AND JAPAN

DE WIT

PLATE II

FIG. 12

INTERGENERIC HYBRIDS BETWEEN TANAKIA TANAGO AND RHODEUS SPINALIS
AND RHODEUS OCELIATUS FROM KOREA AND JAPAN

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

630 Fifth Avenue, New York 20, N. Y.

PUBLICATION OFFICE

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

OFFICERS

PRESIDENT VICE-PRESIDENT SECRETARY TREASURER

Fairfield Osborn Laurance S. Rockefeller George W. Merck David H. McAlpin

SCIENTIFIC STAFF:

William G. Conway. .Director,

Zoological Park

Christopher W. Coates. .Director, Aquarium

John Tee-Van General Director

Emeritus

ZOOLOGICAL PARK

Joseph A. Davis, Jr. . . Curator, Mammals

Grace Davall Assistant Curator,

Mammals and Birds
William G. Conway . . Curator, Birds
Herndon G. Dowling . Curator, Reptiles
Charles P. Gandal. . . Veterinarian

Lee S. Crandall General Curator

Emeritus

Roland Lindemann .... Consultant in Mam-
mal Management

AQUARIUM

James W. Atz Curator

Carleton Ray Associate Curator

Ross F. Nigrelli Pathologist & Chair-

man of Department
of Marine Biochem-
istry & Ecology

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . . Research Associate

in Histology

Sophie Jakowska Research Associate

in Experimental
Biology

Klaus D. Kallman. . . . Research Associate

in Genetics

Louis Mowbray Research Associate

in Field Biology

GENERAL

William Bridges . . Editor & Curator,

Publications

Dorothy Reville . . Editorial Assistant

Sam Dunton Photographer

Henry M. Lester. .Photographic Consultant

DEPARTMENT OF TROPICAL
RESEARCH

Jocelyn Crane Directing Curator

John Tee- Van Associate

William K. Gregory .... Associate

AFFILIATE

L. Floyd Clarke Director,

Jackson Hole Biological Research Station

EDITORIAL COMMITTEE

Fairfield Osborn, Chairman
James W. Atz Lee S. Crandall

William Bridges Herndon G, Dowling

Christopher W. Coates John Tee-Van
William G. Conway

S*i6'S73

ZOOLOGICAL

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 47 • PART 3 • FALL, 1962

PUBLISHED BY THE SOCIETY
The ZOOLOGICAL PARK, New York

Contents

PAGE

11. Observations of the Sound Production Capabilities of the Bottlenose
Porpoise: A Study of Whistles and Clicks. By William E. Evans & John

H. Prescott. Plates I-IV; Text-figures 1-6 121

12. Notes on the Biology of Some Trinidad Swifts. By D. W. Snow. Text-

figure 1 129

13. The Genetics of Some Polymorphic Forms of the Butterflies Heliconius

melpomene Linnaeus and H. erato Linnaeus. I. Major Genes. By J. R. G.
Turner & Jocelyn Crane. Plate I; Text-figure 1 141

Zoologica is published quarterly by the New York Zoological Society at the New York
Zoological Park, Bronx Park, Bronx 60, N. Y., and manuscripts, subscriptions, orders for back
issues and changes of address should be sent to that address. Subscription rates: $6.00 per year;
single numbers. $1.50, unless otherwise stated in the Society’s catalog of publications. Application
to mail at second-class postage rates is pending at Bronx, N. Y.

Publication date of Part 3 will be printed here, in Part 4.

11

Observations of the Sound Production Capabilities of the
Bottlenose Porpoise: A Study of Whistles and Clicks1

William E. Evans

Lockheed California Company , Burbank, Calif.,

& John H. Prescott

Marineland of the Pacific Biological Laboratory,

Palos Verdes Estates, Calif.

(Plates I-IV; Text-figures 1-6)

Introduction

FOR centuries man has been aware that of
all the creatures of the sea the whales and
porpoises are among the most vociferous.
Recent research (Norris et al., 1961; Kellogg,
1958; Kellogg, et al., 1953) has demonstrated
that one functional aspect of this vocal behavior
is its use in navigation and food-finding by the
process of echo-location. This form of behavior
has proved to be particularly interesting to lay-
men and the military as well as psychologists
and zoologists, and it is not surprising that the
other vocalizations produced by these animals
have been relatively neglected. That they have
the ability to emit several sounds in addition to
those used for echo-location raises the possibility
that other functions, e.g. communication, are
involved.

The best known cetacean is the Atlantic bot-
tlenose porpoise, Tursiops truncatus (Montagu),
but before a total and accurate description of the
complex acoustic behavior displayed by this
species can be drafted, a great many more data
are needed. At present, information is available
on the frequency range, duration and direction-
ality of the signals it uses for echo-location.
Some aspects of the swimming behavior asso-
ciated with this activity have also been described
(Norris et al., 1961). On the other hand, little is
known about the mechanisms used by the por-
poise to produce its wide array of vocal signals,
and the range and variability of the animal’s

Contribution Number 15, Marineland of the Pacific
Biological Laboratory.

non-echo-location acoustic output has not been
specified.

This paper describes some preliminary efforts
to investigate the sound production apparatus
and signal repertoire of the Atlantic bottlenose
porpoise. The animals and facilities were made
available through the cooperation of Marineland
of the Pacific, where all observations were made.

We would like to express our gratitude to Mr.
Wm. F. Monahan, Marineland of the Pacific, to
Dr. John Dreher and Mr. William Sutherland,
Lockheed Aircraft, and to Dr. Carl George,
American University, Beirut, Lebanon, for their
able assistance. We also wish to thank Mrs.
Muriel Johnson for help in the preparation of
the manuscript.

Collection of Porpoise Vocalization Data

To initiate this investigation it was necessary
to acquire an adequate sample of the vocal be-
havior of Tursiops. This was achieved by re-
cording the sound production of a group of five
to eight specimens which were confined in a
concrete holding tank, 30 feet in diameter and
5 feet deep. Recordings from a single animal
confined in a plastic-lined holding tank, 30 feet
in diameter and 4 feet deep, were also obtained.
In all, some six hours of taped data were col-
lected during several recording sessions. A block
diagram of the instrumentation used is presented
in Text-fig. 1. The over-all response of this re-
cording system was from 100 cps to 17 kcps,
plus or minus two decibels, and the sensitivity
was down less than 12 db at 20 kcps, with a tape
speed of IVi inches per second. The equivalent

121

122

Zoologica: New York Zoological Society

[47: 11

Text-fig. 1. Block dia-
gram showing instrumen-
tation used to record por-
poise vocalizations.

frequency response of the system at the reduced
playback speed used for analysis, which was 3.75
inches per second, did not suffer, but was ap-
proximately one decibel better than that of the
recording system.

The system used for the data analysis con-
sisted of the tape recorder and a Sona-Graph
sound spectrograph (Kay Electric Co., Pine-
brook, N. J.). This instrument analyzes a com-
plex signal as a function of time and frequency.
The resultant portrayal, known as a sonagram,
displays frequency along the vertical axis, time
along the horizontal axis and intensity by the
darkness of the pattern. The use of this type of
equipment for the analysis of biological sounds
has been thoroughly reviewed in Lanyon &
Tavolga (1960).

Analyses of the recorded data indicated that
all of the sounds produced by the animals fell
into one of the following three categories.

(a) Clicks or plosive sounds— echo-location
pulses with a time duration of .001 to .01 second
and a frequency range 100 cps to 35 kcps and
beyond. Although the recording equipment used
in these studies has an upper frequency response
limit of 20-24 kcps, it is adequate for studying
certain aspects of the echo-location clicks. The
echo-location signal may have its greatest in-
tensity in the 20-35 kcps range; however, it ap-
pears in several regions of frequency reinforce-
ment with considerable energy from 3 kcps to
20 kcps, which is within our recording range. We
were thus able to determine and track pulse rate,
frequency shift and other physical data of the
echo-location clicks by utilizing the portions of
the signal of lower frequency (below 22 kcps).

(b) Whistles— frequency modulated signals,
very narrow band with one or two overtones;
frequency range 4,000 to 20,000 cps.

(c) Barks— sounds rich in overtones which

must be produced in a system of complex reso-
nators, frequency range 200 to 16,000 cps. Sona-
grams showing a frequency versus time display
for each of these types of sound emissions are
presented in Plate I.

Sounds associated with echo-ranging were the
most frequently observed acoustic phenomena.
The animals were constantly echo-ranging the
tank, checking for possible environmental
change, e.g., the addition of food. Repetition
rates of this type of signal varied from one or
two pulses per second to rates as high as 525
pulses per second.

The next most frequently observed sounds
were whistles. These whistles were usually as-
sociated with movement of one or more animals
about the tank. The lowest frequency observed
was 3.7 kcps and the highest about 20-22 kcps,
with the signal durations varying from .10 to 3.6
seconds. The shape of the whistle contours were
quite variable. In all, a total of 18 basically dif-
ferent whistles were observed. Stylized fre-
quency vs. time plots of some of the most fre-
quently observed whistles are presented in Text-
fig. 2.

Barks were the most infrequently observed
sound. The fundamental frequency of these sig-
nals varied between 200 and 300 cps with har-
monics as high as 16 kcps. Signal duration was
usually about 0.10 second. These barks usually
occurred when the animals were engaged in
some stationary activity rather than swimming.

Mechanisms for Sound Production

Anatomical examination of the nasal and
laryngeal tracts of the delphinids Tursiops trun-
catus and Stenella graffmani reveals structures
that appear to be suited for the production of
each of the sounds previously mentioned.

For the production of clicks (plosive echo-

1962]

Evans & Prescott: Sound Production Capabilities of Bottlenose Porpoise

123

Text-fig. 2. Composite diagrams showing frequency vs. time plots of six of the most frequently occurring
whistles produced by Tursiops truncatus.

location sounds), the complex nasal sac system
described by Lawrence & Schevill (1956) ap-
pears to be the most likely source. This system
represents an interconnecting pneumatic tract
with several cm/ de sacs, valves and lips.

The nasal sacs comprise the distal part of the
nasal airway and lie within the soft tissues of the
head, dorsad of the bony nares and premaxillary

bones of the skull (Text-fig. 3). They are lo-
cated slightly below the fatty and connective
tissue of the melon, or forehead, surrounded by
their own muscle groups. All of them are paired.
The premaxillary sacs are the largest and lie
immediately on top of the smooth premaxillary
bones, immediately anterior to the distal open-
ings of the bony nares. The premaxillary sacs

A

B

Text-fig. 3. Sketch of a bottlenose porpoise head with a cut-away exposing the naso-laryngeal tract, indi-
cating the approximate locations of the nasal sacs, fatty melon and bony nares. A. Cutaway of whole
head. B. Expanded view of sac system.

124

Zoologica: New York Zoological Society

[47: 11

communicate with the bony nares, their soft
posterior wall continuing as an extension to
become the nasal plugs. The vestibular sacs are
the outermost pair and lie posterolateral to, and
just below, the lips of the blowhole. As Lawrence
& Schevill (1956) have pointed out, “There is
a considerable mass of tissue lying between the
vestibular area and the dorsal surface of the
plugs and it is here that the last two pairs of sacs,
the tubular and the connecting, lie embedded.
These are both true sacs in the sense that they
are distinct from the main air passage and con-
nect with it only via rather small, well marked
openings.” The tubular sacs are U-shaped with
a slit opening in the lateral portion connecting
to the main airway. They lie nearly horizontal
in the head, almost surrounding the main air-
way. The connecting sacs are appendices to the
tubular sacs and provide the only direct passage-
way for air from the tubular into the premaxil-
lary sacs and then the main airway. These con-
nections are along the distal edges of the nasal
plugs near the base where the premaxillary sacs
join the airway and open into the bony nares.
Near the top part of the nasal plugs, at a point
opposite the slit openings of the tubular sacs,
there are small tongue-like nodes. Each of these
appendages is inserted into the opening of the
tubular sac (Text-fig.3). The intrinsic muscles
appear to function as mechanisms to distend the
tubular sacs, and transverse ligaments beneath
the posterior limbs of the tubular sacs help
stiffen the posterior walls of the passage (Law-
rence & Schevill, 1956). These interconnections
and other structures provide a passage for com-
munication between the main air passages and
the tubular sac. Such an arrangement links these
sacs with the nasal airway without disturbing the
position of the nasal plug and breaking the
water- and gas-tight seal of the blowhole. It
seems from examination of the musculature, the
surrounding connective tissue and the lips of the
tubular sacs that these mechanisms are capable
either of maintaining or accepting the air pres-
sures necessary to generate high frequency echo-
location clicks. When air flows past the nodes of
the nasal plugs projecting into the tubular sac,
these nodes could vibrate, acting as a type of
“raspberry” generator. This arrangement is not
unlike the human tongue, inserted between the
lips, used in producing a “Bronx cheer.”

If the lens-shaped fatty body, or melon, char-
acteristically found in the “forehead” of toothed
whales could function as a variable focus
“acoustic lens,” then the complex nasal sacs and
the accompanying pneumatic valve system
would be ideally positioned. In light of Reysen-
bach de Haan’s (1956) observations that blub-
ber and fatty tissue are essentially acoustically

transparent due to their excellent impedence
match for water, this arrangement would allow
little energy loss of the air-produced sounds
when transmitted into the water. Further evi-
dence for this system exists in the somewhat
modified asymmetric parabolic reflector formed
by the dense bone of the premaxillae. Thus an
acoustic lens, the melon, backed-up with the
specially modified skull, could act to “focus” the
animal’s echo-location signals. The observations
that when completing echo-location searches, in
order to find targets or pick up food, Tursiops
bends its head in a rapid circular scanning mo-
tion and appears to point its melon in the direc-
tion of search indicate that the sound pulse used
is generated in the nasal sac system (Kellogg,
1 960) . Norris et al. (1961) noticed that a blind-
folded Tursiops had difficulty in echo-locating
objects positioned on a line below the melon.

Although the literature suggests that the
Odontoceti have no vocal cords, as they exist in
man for example, their larynx is a complex
structure. The larynges of Tursiops and Stenella
are shown in Plate II. Structurally, these
larynges differ from those of terrestrial mam-
mals in the elongated arytenoid cartilages and
the accompanying epiglottic cartilage which
form the aryteno-epiglottid tube. From the
standpoint of structure alone, this smooth-sur-
faced, hard backed and constricted tube does
form an ideal sound generator. For almost one
hundred years investigators have speculated how
the toothed whales could use this unique larynx
to produce sound (Murie, 1871, and Turner,
1868). Vibration of the tissue folds and of the
elongated arytenoid cartilages have both been
considered. To the best of our knowledge, how-
ever, experimental data either to support or re-
fute these hypotheses have never been published.

Sound Simulation Tests

A series of experiments using the excised
larynx and two severed heads of Stenella and
the head of a fresh specimen of Tursiops was
carried out. The basic plan was to deliver a con-
trolled flow of air at various pressures through
the larynx and the heads to determine whether
sound could be produced, and if so, to analyze
its characteristics and measure the pattern with
which it was radiated from the head. A block
diagram of the instrumentation used during
these tests is presented in Text-fig. 4.

To test the excised larynx of Stenella, a glass
connector was inserted into the trachea and air
delivered through it at a pressure of 1 to 2 psi
and a flow of 10 to 15 liters per minute. The
lips of the arytenoids were approximated by
manipulation with forceps. This procedure re-

1962]

Evans & Prescott: Sound Production Capabilities of Bottlenose Porpoise

125

Text-fig.

suited in the production of a wide range of
sounds depending on the amount of pressure ap-
plied and the proximity of the lips to each other.

Tests with the whole head required the same
instrumentation as those with the excised larynx.
To permit free access around the heads, they
were suspended from a cross-arm with a glass
connector inserted into the trachea. In this series
of tests, no manipulation was used except to
make certain that the aryteno-epiglottid tube of
the larynx was well seated in the ventral open-
ings of the bony nares. Sound was produced at
pressures and flow rates of the same order of
magnitude as those in the previous test with the
larynx alone.

The sounds produced during both types of
test were quite similar. Depending on the pres-
sure applied and position of the larynx, sounds
ranging from barks, rich in harmonics, to
whistles were produced (Plate III, A & B). In

addition to these sounds, short duration pulse-
type signals (Plate III, C), very similar to those
employed in echo-location, were generated by
air flowing through the nasal sac system of one
specimen of Stenella. As air was forced into the
excised heads, the blowhole became closed and
the pressure was allowed to build up to two
pounds per square inch. With the blowhole effec-
tively sealed, the air sacs, especially the vesti-
bular ones, would become inflated. Pressures
inside the sacs, vestibular and premaxillary,
were checked by inserting a spinal tap needle
through the tissues into the sac. The core of the
needle was then removed and a manometer at-
tached. In the case of sounds similar to echo-
location clicks, it was necessary for the sacs to
be filled with air before any clicks were heard. If
pressure was applied to the melon, for example
by a hand pressed across the forehead, the sacs
could be collapsed. Although the exact rnechan-

126

Zoologica: New York Zoological Society

[47: 11

Stenella Graffrnani MLP 60-1! ^69

Ref: OASPL 93db re: .0002 dyn®s/cm^ (Odb)
0.2Sln.from heed /

Airflow: 10 iiters/rnin
1.04 lb /in2 7

-is <n>

Text-fig. 5. Over-all sound pressure levels meas-
ured at various locations .25 inches from the head
of a specimen of Stenella graffrnani during sound
simulation tests.

ism that produced these signals was not deter-
mined, deflation of the upper or distal nasal tract
stopped the echo-location clicks. These sounds
were achieved with air pressures of 1 psi within
the specimens of Stenella and 1.5 psi within the
Tursiops. Air flow was 10 liters per minute for
both.

In addition to the sonagraphic analyses made
of the clicks produced with the heads, measure-
ments were also made of the over-all sound pres-
sure level Va of an inch and 15 inches from the
head at every 60° in the horizontal and ver-
tical planes. The measures of over-all sound
pressure level near the surface of the head
at both distances are shown by the polar
plots presented in Text-figs. 5 & 6. Most note-
worthy is the asymmetrical radiation, with the
stronger signal output on the right. This asym-
metry immediately calls to mind the general
asymmetry of cranial structures in toothed

whales (Odontoceti) noted by anatomists. Ho-
sokawa (1950) has reported this in the muscu-
lature of the larynx of the sperm whale; Law-
rence & Schevill (1956) have noticed this in
the nasal sac system and skulls of Tursiops trun-
catus and Stenella plagiodon; and we have found
that the nasal sacs in Tursiops truncatus and
Stenella graffrnani are larger on the right side.
In fact, all the delphinids examined by the jun-
ior author have exhibited an asymmetrical devel-
opment of the cranium and nasal sac system.
These include Delphinus bairdi, Tursiops gilli,
Lagenorhynchus obliquidens, Stennella longi-
rostus and Globicephala scammoni.

The indirect evidence presented so far indi-
cates that porpoises do have structures that can
produce a variety of different types of sounds.
It does not, however, add any support to the
contention that echo-location clicks and whist-
ling sounds are produced by separate mechan-
isms. The following observations, however, add
credence to the supposition of specialized sound
producers.

All of the initial observations of sound pro-
duction using specimens of Tursiops were ac-
complished with groups of five or more animals.
In order to simplify the sound analyses and ob-
servations on behavior, one animal was trans-
ferred to the smaller plastic-lined tank one day
before testing. In this tank the porpoise was not
restrained and was given food immediately.
Using the previously described equipment, re-
corded Tursiops sounds, whistles and clicks were
broadcast into the tank. Until this test, not a
great deal of success had been achieved in elici-
ting whistles from solitary animals. During this
test and for a period after stimulation, however,
several long whistles were produced by the ani-
mal. The most interesting aspect of this behavior
was that in all cases, the solitary animal pro-
duced echo-location signals and whistles simul-
taneously. Lilly & Miller (1961) report similar
behavior from a physically restrained, solitary
Tursiops. The time duration of the whistle por-
tion of the simultaneous emissions that they
observed ranged from .1 to .4 seconds. We have
observed whistle durations from .45 to 1.5 sec-
onds, the majority of them being longer than
.6 second. Several sonagrams of these simultan-
eous signals are shown in Plate IV in which it
can be seen that the whistles are overlaid on
the pulsed signal. The whistles are not highly
modulated and appear normal, that is, as they do
in Plate I, B. If only one mechanism for sound
production was involved, the whistle could be
expected to show some of the modulations of
the echo-location signal.

1962]

Evans & Prescott: Sound Production Capabilities of Bottlenose Porpoise

127

Text-fig. 6. Over-all sound pressure levels measured every 60°, 15 inches from the head of a specimen of
Stenella graffmani in both A vertical plane and B horizontal planes.

Conclusions and Summary

1 . All of the natural vocalizations of a group
of Tursiops and a single individual Tursiops
can be placed in one of the three following
categories: Plosive or pulse signals, whistles
and barks.

2. Both Tursiops truncatus and Stennella graff-
mani have anatomical structures capable of
producing a variety of sounds under condi-
tions of artificial stimulation. These sounds
are not unlike the ones produced naturally by
these animals. Sounds artificially produced in
the head of a specimen of Stenella showed an
asymmetrical sound radiation pattern corres-
ponding with the asymmetry of the cranial
structures.

3. Tursiops is capable of producing echo-loca-
tion pulses and whistles simultaneously.

4. Anatomical and behavioral evidence, as well
as sound pressure measurements, indicate that
the echo-location clicks are produced in the
nasal-sac system of porpoises. Within the sac-
system, the tubular sacs combined with the
nasal plug nodes appear to be the site of
sound production. It is only through the
tubular and connecting sacs that we find a
continuous air passage, the vestibular and pre-
maxillary sacs being cul de sacs.

Bibliography

Hosokawa, Hiroshi

1950. On the cetacean larynx, with special re-
marks on the laryngeal sacs of the sei
whale and the aryteno-epiglottideal tube
of the sperm whale. Sci. Rep. of the
Whales Res. Inst., 3: 42-43.

Kellogg, W. N.

1958. Echo-ranging in the porpoise. Science,
128: 982-988.

1960. Auditory scanning in the dolphin. Psych.
Rec., 10(1): 25-27.

Kellogg, W. N., R. Kohler & H. N. Morris

1953. Porpoise sounds as sonar signals. Science
117: 239-243.

Lanyon, W. E. & W. N. Tavolga (Editors)

1960. “Animal Sounds and Communication.”
443 p. A.I.B.S., Washington, D. C.

Lawrence, B., & W. E. Schevill

1956. The functional anatomy of the delphinid
nose. Bull. Mus. Comp. Zool., Harvard,
114(4): 103-151.

Lilly, I. C, & A. M. Miller

1961. Sounds emitted by the bottlenose dolphin.
Science, 133: 1689-1693.

128

Zoologica: New York Zoological Society

[47: 11: 1962]

Murie, J.

1871. On Risso’s Grampus, G. rissoanus. Jour.
Anat. Physiol., 5: 118-138.

Norris, K. S., J. H. Prescott, P. Asa-Dorian &

P. Perkins

1961. An experimental demonstration of echo-
location behavior in the porpoise Tursiops
truncatus (Montagu). Biol. Bull., 120(2):
163-176.

EXPLANATION

Plate I

Fig. 1. Sonagrams of examples of the three types
of sound produced underwater by Tursiops
truncatus. A. Clicks or pulses. B. Whistle.
C. Bark.

Plate II

Fig. 2. A. Larynx from Stenella graffmani. B.
Larynx from Tursiops truncatus.

Plate III

Fig. 3. A. Sonagrams of sound produced by pass-
ing air through the excised larynx of Sten-

Reysenbach de Haan, F. W.

1956. Hearing in whales. Acta Oto-Laryngolo-
gica, Supp. 134.

Turner, W.

1868. A contribution to the anatomy of the
pilot whale ( Globicephala Svineval, (La-
cepede). Jour. Anat. Physiol., 2: 66-79.

F THE PLATES

ella graffmani. B. Sonagrams of sound pro-
duced by passing air through a severed
head from Stenella graffmani. C. Sonagram
of short-duration, repetitive, pulse sounds
generated by air flow through the nasal sac
system of one specimen of Stenella graff-
mani.

Plate IV

Fig. 4. Sonagrams of four instances of simultan-
eous production of echo-location pulses and
whistles by a single isolated Tursiops trun-
catus.

EVANS a PRESCOTT

PLATE 1

B

C

OBSERVATIONS OF THE SOUND PRODUCTION CAPABILITIES OF THE
BOTTLENOSE PORPOISE: A STUDY OF WHISTLES AND CLICKS

EVANS & PRESCOTT

A

B

TRACHEA

ARYTENOID CARTILAGES
ADDITUS LARY

EPIGLOTTIS

EXCISED LARYNX FROM EXCISED LARYNX FROM

STENELLA GRAFFMAN! TURSIOPS TRUNCATUS

OBSERVATIONS OF THE SOUND PRODUCTION CAPABILITIES OF THE
BOTTLENOSE PORPOISE: A STUDY OF WHISTLES AND CLICKS

FREQUENCY IN KILOCYCLES

EVANS & PRESCOTT

PLATE 111

TIME

OBSERVATIONS OF THE SOUND PRODUCTION CAPABILITIES OF THE
BOTTLENOSE PORPOISE: A STUDY OF WHISTLES AND CLICKS

REQUENCY (KILOCYCLES) FREQUENCY (KILOCYCLES) FREQUENCY (KILOCYCLES) FREQUENCY (KILOCYCLES)

EVANS a PRESCOTT

PLATE IV

16

14

12

10

8

6

4

2

0

16

14

12

10

8

6

4

2

0

16

14

12

10

8

6

4

2

0

16

14

12

10

8

6

4

2

0

OBSERVATIONS OF THE SOUND PRODUCTION CAPABILITIES OF THE
BOTTLENOSE PORPOISE: A STUDY OF WHISTLES AND CLICKS

12

Notes on the Biology of Some Trinidad Swifts1

D. W. Snow

Department of Tropical Research,

New York Zoological Society, New York 60, N. Y.

(Text-figure 1)

[This paper is one of a series emanating from the
Tropical Field Station of the New York Zoological
Society, at Simla, Arima Valley, Trinidad, West
Indies. This station was founded in 1950 by the
Zoological Society’s Department of Tropical Re-
search, 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 preserves. 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, and the annual rainfall
is 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, 37 (13): 157-184.]

Contents Page

Distribution, Status and Feeding Habits 130

The Breeding of Chaetura brachyura 131

Breeding, Moulting and Roosting of Other

Chaetura Species 1 34

The Moulting Sequence in Chaetura 134

The Breeding of Cypseloides rutilus 135

Hovering Flight in Cypseloides and Chaetura 137

Summary 138

Literature Cited 138

Appendix 139

THE biology of the neotropical swifts is
not well known. For the genera Chaetura
and Cypseloides, with which this paper is
concerned, only Beebe (1949), Haverschmidt
(1958) and Sick (especially 1948, 1951, 1958,
1959) appear to have made significant recent
contributions. They are difficult birds to watch
and many are difficult to identify in the field.
Several species have never been found nesting.

Compared with the mainland of South Amer-
ica, where, to judge from published data, it is

iContribution No. 1018, Department of Tropical Re-
search, New York Zoological Society.

unusual for several species to be common in the
same area, Trinidad is especially suitable for
field studies of swifts. Here, on an island meas-
uring some 50 by 30 miles, seven species are
resident: five of them are known to breed and
all seven probably do so. Only one species can
be called rare. Probably nowhere else in the
neotropical region can so many breeding species
of swifts be found in so small an area.

During 4Vi years’ residence in Trinidad, 1
made such observations on the swifts as oppor-
tunity permitted. They were not a main subject
of study, but their abundance brought them con-
stantly to notice. Nests of two species were
found in sufficient numbers for a limited analy-
sis of breeding season, number of broods, nest-
ing success and other aspects of breeding biol-
ogy. In addition, five species were caught in
mist-nets, two of them in good numbers, and
a sixth was twice caught by hand. This paper
presents the information thus collected.

I have followed Lack (1956) in his wide
definition of the genus Cypseloides. For vernac-
ular names, I have followed Eisenmann (1955)
except that I have preferred to use the more
evocative name Cloud Swift for Cypseloides
(Streptoprocne) zonaris.

I am grateful to my wife for much help with
the field work, to Dr. Wilbur G. Downs for pho-
tographic assistance, and to Charles T. Collins,
J. Dunston and R. P. ffrench for visiting nests
at times when I was unable to do so. C. T.
Collins discovered one of the Chestnut-collared
Swift nests and J. Dunston three of the Short-
tailed Swift nests. This study, part of a wider
program of field studies on the biology of neo-
tropical birds, has been generously supported
by National Science Foundation grants G 4385
and G 21007.

129

130 Zoologica: New York Zoological Society [47: 12

Distribution, Status and Feeding Habits The Cypseloides Species

J he Chaetura Species Chestnut-collared Swifts (C. rutilus) are

The Short-tailed Swift (C. brachyura) , Band- common in the Northern Range, but they feed

rumped Swift (C. spinicauda) and Gray-rumped much higher than the Chaetura species, and pre-

Swift (C. cinereiventris) are all common spe- senting only a black silhouette are often less easy

cies, but they have different local distributions. to identify with certainty. Because they fly high,

The Short-tailed Swift is the most widespread, they were the only species resident in the Arima

occurring all over the island; it is the only spe- Valley never to be caught in a mist-net. They are

cies regularly occurring over open country and usually seen in ones or twos, or small parties,
towns. It may be seen throughout the Northern Only once were many seen together; on Septem-
Range (the range of forested mountains, up to ber 3, 1958, large numbers were seen flying
3,000 feet high, running along the north side around the summit of El Tucuche (3,068 ft.),
of Trinidad), but is less common there than the second highest peak in Trinidad, and with
the other two species. The Band-rumped Swift them only a very small number of four other
occurs generally over forested country and species. This was a post-breeding aggregation;
wooded savanna, in both lowlands and hills. The many were seen to be moulting their wing-
Gray-rumped Swift is common in the Northern feathers.

Range; I did not see it elsewhere. The Cloud Swift (C. zonaris ) , the largest and

The fourth species, Chapman’s Swift (C. most majestic of the New World swifts, is a
chapmani ) , is very little known. Apart from the migrant to Trinidad. The first birds arrive in
original series collected by Chapman in 1894, July (first dates: 1957, end of July; 1958, July
only a few specimens have been collected, in 27; 1959, July 14; 1960, mid-July; 1961, July
widely scattered localities from Panama to cen- 8). August and September are the months when
tral Brazil (Wetmore, 1957). 1 several times the greatest numbers are present. By October
thought that I saw Chapman's Swift among their numbers have begun to decline, but a few
mixed parties of swifts in the Northern Range, are seen irregularly until February. This prob-
but the first positive evidence was obtained on ably represents a post-breeding dispersal from
November 27, 1960, when one was caught in a breeding areas in the Andes of Venezuela to the
mist-net at 1,800 feet at the head of the Arima west of Trinidad. A number of specimens col-
Valley in the center of the Northern Range, lected by Roberts (1934) on August 28 were all
Having had the opportunity to examine a living completing their moult. I obtained three speci-
bird in the hand and to note its field characters mens, in late July and early August; all appeared
as it flew away on release, I later was confident to be juveniles and were not moulting,
that I saw Chapman’s Swift on two occasions. Cloud Swifts are erratic in their appearance,
once in the same place and once in the lower being present in great numbers one day and ab-
part of the Arima Valley. But hundreds of the sent the next. They feed at all heights, and range
three other Chaetura species were examined rapidly and widely over the island. It is likely
before these were seen. From these few records, that at least on occasions they spend the night
and from Chapman’s specimens collected at in the air, as the European swift A pus apus has
Valencia at the foot of the Northern Range been found to do. On August 11, 1957, I
about six miles east of the Arima Valley, we watched a large flock at dusk in the Arima Val-
may conclude that Chapman’s Swift is prob- ley, in the center of the Northern Range. They
ably resident in the Northern Range in small were spiralling upwards, gradually drifting out
numbers. of sight behind some hills. Some at least were

I obtained no evidence for any differences in still circling when poor light prevented further
feeding habits of the three common Chaetura observation. Cloud Swifts roost on cliffs, in
species in places where they occurred together, clefts and behind waterfalls (Sutton. 1951).
though quantitative observations might eventu- There are few suitable cliffs in Tunidad, except
ally show fine differences. All feed regularly the sea cliffs of the north coast, and it seemed
from ground level to several hundred feet up. unlikely that the large flock I was watching
They commonly skim close to the ground in could all have found roosting places for that
sheltered clearings in the hills, probably because night.

of the abundance of flying insects in such places. Belcher & Smooker (1936), on rather unsat-
and when termites swarm after wet weather isfactory evidence, record another migrant swift
they all come low to feed on them. In wet wea- for Trinidad, the Black Swift (C. ( Nephoecetes )
ther great flocks of them move rapidly up and niger ) . This species would be expected to occur
down the valleys and across the watersheds, occasionally on passage, and further observation
avoiding the rain-storms. may confirm that it does so.

1962]

Snow: Biology of Some Trinidad Swifts

131

Other Species

Panyptila cayennensis is widespread in small
numbers, breeding from near sea level to at
least 1,000 feet in the Northern Range. My
records of its nests are all in the northern and
eastern parts of Trinidad, but the other parts of
the island were visited much less often. Panyp-
tila feeds high above the ground and is usually
only to be identified with binoculars.

Reinarda squamata is confined to savanna and
swamp forest in the east of Trinidad, following
the distribution of the palm Mauritia setigera,
on which it nests. It commonly feeds close to the
ground.

The Breeding of Chaetura brachyura
Nest-site

Belcher & Smooker give the only previous rec-
ords of the nesting of the Short-tailed Swift in
Trinidad. They reported one nest in a chimney,
with 3 eggs on June 20, and one nest in a sea
cave, with 3 eggs on June 3. There are no breed-
ing records from elsewhere, but observations by
Bond (1928, and in litt.) suggest that Short-
tailed Swifts nest in chimneys in St. Vincent.

I discovered a small breeding colony of Short-
tailed Swifts by chance on May 7, 1957, when,
watching swifts flying low over bushy savanna
near Valencia, I saw a bird suddenly drop out of
sight in the grass. I found that it had entered a
concrete man-hole, part of an abandoned drain-
age system constructed during the war when
this savanna was the site of a U.S. Air Force
camp. A search made in the immediate area
revealed eight other man-holes which swifts
could enter, and three more were eventually
found on other parts of the savanna. These holes
were kept under regular observation until Sep-
tember, 1961, giving complete records for five
breeding seasons.

Some other breeding sites were found. Short-
tailed Swifts were seen during the breeding sea-
son entering and leaving sea caves in the rocky
north coast of Trinidad, and in Huevos Island
off the northwest corner of Trinidad, but no ac-
cessible nests were found. One pair was found
nesting in a chimney in Arima. Finally, a pair
nested for two seasons in a nest-box which I
erected 30 feet up against the trunk of a large
Chataigne tree ( Pachira insignis ), 500 feet
above sea level in the Arima Valley.

The use of this nest-box was unexpected, as
all previous records of their nests had been in
caves, man-holes or chimneys, and the species
is less characteristic of forest than the other
Chaetura species. The history of its occupation
was as follows. On July 9, 1959, a pair of Band-
rumped Swifts had been observed entering a

natural cavity some 30 feet up in the trunk of
this tree, and from observations made on July
1 1 it was clear that they were feeding young. A
ladder was put up to the hole, but the nest could
not be seen. The entrance was narrow, and the
hole was irregular and descended several feet.
Since Band-rumped Swifts had never been found
nesting, I subsequently blocked up the natural
hole and fixed a nest-box near by, 5 feet long
and 1 Vi feet square in section, with a slit-shaped
hole near the top about 5 inches wide (about
one inch wider than the tree-hole). Occupation
of the box was first noted on June 22, 1960,
when a pair of Short-tailed Swifts were found to
be feeding five half-grown young. In 1961 a
pair of Short-tailed Swifts again nested in the
box. In both years a pair of Band-rumped
Swifts, presumed to be the former owners of the
tree-hole, were several times seen about the
tree. It seems that, in making the hole of the
box a little wider than the tree-hole, I inad-
vertently made it big enough for Short-tailed
Swifts to enter and they were thus able to dis-
possess the smaller Band-rumped Swifts (the
mean wing-length of the two species differs by
about 17 mm.). It is probably because of their
larger size that Short-tailed Swifts normally nest
in caves and man-made cavities, while the
smaller Chaetura species, as far as known, nest
in tree-holes.

It was of interest that when the Short-tailed
Swifts left the box after feeding the young, they
used to fly off down the valley in the direction
of the savanna three miles away. They would
return from the same direction.

Breeding Season

When I first found the nesting colony in the
man-holes in early May, 1957, nests were being
built and none had eggs until a week later. Table
I shows the monthly distribution of the 86 nest-
ings recorded at this colony in the five years.
Breeding begins in April or early May and con-
tinues until August or early September (the
three breeding dates obtained at other sites, and
Belcher & Smooker’s two records, all fall near
the middle of this period). Numbers were too
few to reveal slight annual variations, but there
were no marked differences in breeding season
in the five years.

There was usually a regular succession of
nestings in each hole in the course of the breed-
ing season, either in the same nest or, less often,
in a succession of nests. No hole ever contained
more than one occupied nest at the same time.
The intervals between the nestings suggested, as
would be expected, that these were normally
successive layings by the same female. Com-
bining all the years, two holes were used four

132

Zoologica: New York Zoological Society

[47: 12

Table I. Breeding Season of Chaetura brachyura

Number of nests started in
periods

aalf-monthly

1957

1958

1959

1960

1961

all

years

April 1

1

1

2

1

1

2

May 1

5

4

1

3

1

14

2

4

3

2

3

2

14

June 1

1

4

3

4

12

2

1

1

5

1

2

10

July 1

3

2

1

3

9

2

1

4

2

1

1

9

Aug. 1

3

1

2

3

9

2

3

3

Sept. 1

1

1

1

3

Totals

19

16

16

14

21

86

times in a season, six holes three times, 22 holes
twice, and 13 holes only once. These nests suf-
fered a very high rate of predation, which un-
doubtedly raised the number of nesting attempts
above what would be usual in a more success-
fully breeding population. If we consider only
the nine cases where the first nesting attempt
was successful, five were followed by a second
laying and four were not. The breeding season
is hardly long enough for three broods to be
reared successfully.

Clutch-size

Chaetura species are known to have, for
swifts, very large clutches, and the Short-tailed
Swift is no exception. The mean number of
eggs in 41 clutches was 3.6 (Table II). Mean
clutch-size probably decreases in the course of
the breeding season, since the only clutch of six
eggs and all the clutches of five were laid in
May.

Table II. Clutch-size of Chaetura brachyura

Number of nests with clutch-size of

1

2

3

4

5

6

April 2

1

May 1

2

2

2

1

2

3

3

2

June 1

2

2

1

2 1

2

1

July 1

2

3

2

2

2

Aug. 1

3

2

2

Sept. 1

1

1

All months 1

2

17

16

4

1

It was not possible to obtain exact data on
the intervals between the laying of successive
eggs, as my visits were not frequent enough, but
incomplete data for ten nests showed that the
normal interval is probably two or three days
and longer intervals are not uncommon.

A clutch of nine eggs has been omitted from
Table II, as it was certainly laid by more than
one female. Apart from the abnormal size of
the clutch, it could hardly have been laid by one
bird as all nine eggs were laid within 14 days.
It was of interest that the eggs were incubated
successfully, though they were two or three deep
in the nest; all nine hatched at about the same
time. There was another instance in which the
circumstantial evidence was strong that two fe-
males were laying in the same nest. Three eggs
were laid between 0830 hours on June 1 and the
afternoon of June 3, an abnormally quick rate
of laying for a single bird. All three eggs were
found broken below the nest on June 3, sug-
gesting that the two birds were in conflict. For
Chaetura andrei, Sick (1959) has reported a
similar case of more than one female laying in
the same nest. Such cases are presumably at-
tributable to shortage of nest-holes.

Incubation and Fledging Periods

Because the colony was as a rule visited
weekly, it was not possible to obtain incubation
periods for most of the nests. But more frequent
visits were made over limited periods, with the
result that for three nests the incubation periods
(from the laying of the last egg to the hatching
of the last young) were found to be 17 ±1,
17-18 and 18 ±1 days. None of the less exact
records was inconsistent with a 17-18 day in-
cubation period.

Fledging periods were difficult to ascertain for
the additional reason that the inspection of nest-
holes near the fledging time was liable to cause
the young to leave prematurely. Also it was
found that the young birds return to their nest-
hole by day after their first flight. Thus only
very careful, repeated observations would show
with certainty when undisturbed young first fly.

With these reservations, partial information
was obtained on the fledging periods at eleven
nests. At five of them, the young flew out on
inspection at the ages of approximately 28, 29,
29, 32 and 34 days. One of those that flew when
29 days old was caught as it struggled to rise
clear of the long savanna grass and was found
to have a wing-length of 1 12 mm., 8 mm. shorter
than the mean adult wing-length. At the six
nests where there was no evidence of disturb-
ance, the young left as follows:

1962]

Snow: Biology of Some Trinidad Swifts

133

(1)

one of the two at 29-32 days,

the

other

at 32-36 days.

(2)

the first young at 32-36 days,

the

other

two at 36-40 days.

(3)

one of the two at 34-38 days,

the

other

after 38 days.

(4)

all four young before 35 days.

(5)

both young at 35-42 days.

(6)

the first two young before 36

day

s, the

third after 36 days.

To sum up these observations, the young can
fly if disturbed as early as 28 days after hatch-
ing; if undisturbed, they do not usually leave
until they are 30-40 days old.

Like Chimney Swifts ( Chaetura pelagica),
young Short-tailed Swifts climb out of the nest
while they are still unfeathered and cling to
the wall near the nest. At one nest, one of the
four nestlings was found clinging to the out-
side of the nest when 15 days old, the other three
still being in the nest-cup. Two days later, at the
age of 16-17 days, all four were clinging to the
wall a few inches below the nest. At other nests
the young were not recorded leaving the nest
until three or four days later than this.

Intervals between Broods

The intervals between the ending of one nest-
ing attempt and the laying of the first egg of the
next clutch could not usually be ascertained
exactly, but 27 intervals were known to within
4 days. These were as follows:

after successful fledging of previous brood:
4,12,13,18,24.

after loss of eggs: 9, 10, 10, 15, 17, 19, 24,
24, 27, 27, 31, 32, 32, 32, 35, 51, 74.

after loss of young: 17, 19, 29, 51.

Over half of the intervals are between 10 and
30 days. Some of the very long intervals may be
false, due to the fact that an intervening clutch
had been started and lost between my visits. Be-
cause of the variability of the intervals, the num-
ber of records is rather small for any certain
conclusions, but there is a suggestion that re-
laying usually follows more quickly after the
successful fledging of a brood than after a
failure.

Nesting Success

This population of Short-tailed Swifts nesting
in underground man-holes was singularly un-
successful in its breeding; only 15 (17%) of the
86 recorded nestings resulted in fledged young.
Usually eggs vanished soon after they were laid,
but some losses occurred at all stages. However,
hardly any information was obtained on the
causes of failure. The possibility was not ex-

cluded that snakes and lizards could get into the
holes and attack the nests, but there was no
evidence that they did so. The only certain pred-
ator was a large spider (My gale sp.). When I
accidentally dropped two newly hatched nest-
lings on the floor of the hole, one of these
spiders rushed out of a crevice, seized one of
the nestlings before I could pick it up, and re-
treated back to its lair. The squeaks of the
nestling ceased the moment it was seized and it
appeared to be dead within a second. As this
large spider can climb vertical walls it could be
a regular predator of small nestlings.

Three nest-holes regularly filled with water
after heavy rain, and several nests, as well as at
least three adults, were lost in this way.

These observations can, however, have little
significance in the ecology of the species as a
whole. Underground man-holes must be such
an unusual nest-site for the Short-tailed Swift
that neither the high rate of nest failure nor the
causes of failure are likely to be typical.

The Moult, and Roosting Habits

Two birds trapped on August 14, one caught
on September 16, eight on September 18, and
one on September 25 were all in various stages
of moult. Only three other adults were caught,
one in March and two in July, and none of these
was moulting. Details of the sequence of moult
are given in a later section.

At 1000 hours on September 16, 1957, a fine
sunny day, one of the nest-holes which had not
been in use for some weeks was found to con-
tain about 50 adult swifts. On September 25,
50 birds were again found in this hole at 1020,
and ten were found in another hole containing a
nest with eggs, apparently deserted. The birds
caught on these two days were in moult, as has
been mentioned above. Subsequently it was
found that during the moult the adults regularly
spent part of the day in two or three of the holes,
either clumped, as at night, or clinging scattered
on the wall.

I visited the colony at night several times and
found the swifts roosting in the holes all the
year round, but in the months November-March
their presence was irregular. Outside the breed-
ing season, they tended to congregate in two or
three holes, the same ones that they were found
in by day during the moult, roosting in clumps
of 30 or more birds together. As the breeding
season approached their numbers declined, the
roosting aggregations broke up, and each hole
was finally occupied at night by a pair or three
birds. As late as February, up to 50 birds were
found roosting in the nine holes inspected; by
early May the numbers had fallen to 21.

134

Zoologica: New York Zoological Society

[47: 12

When disturbed at night, Short-tailed Swifts
give the same wing-clattering display as has been
described for the Chimney Swift (Fisher, 1958).
The birds slowly raise their wings above the
back, then spring away from the wall with a
clattering sound, to land back in the same spot
or a few inches away. The clattering appears to
be made by the rapid clapping together of the
wing-tips. Several birds in a clump will raise
their wings together and clatter at the same time,
so that the noise is quite startling to the intruder.
There seems no doubt that it is an anti-predator
display analogous to the hissing of tits and some
other hole-nesting birds. After a few bursts of
clattering a roosting clump becomes broken up,
with the birds more evenly scattered over the
sides of the hole.

Breeding, Moulting and Roosting of Other
Chaetura Species

Extremely scanty information was obtained
on the other Chaetura species. For Chapman’s
Swift nothing was found out except that the in-
dividual caught on November 27, 1960, was
undergoing slight body moult.

For the Band-rumped Swift there were only
two observations of note. On July 9 and 11,
1959, as already mentioned, a pair was found
to be feeding young in a hole about 30 feet up
in the trunk of a Chataigne tree ( Pachira in-
signis) . This is the first indication of the breed-
ing site of the species. A flash photograph of
Short-tailed Swifts roosting in one of the man-
holes on February 27, 1958, showed the pres-
ence among them of a single Band-rumped
Swift. This appears to be the first record of
roosting for the species.

More information came from the trapping
program. Twenty-one Band-rumped Swifts
were caught in mist-nets in the Arima Valley,
6 in January, 1 in February, 1 in April, 2 in
August, 8 in October and 3 in November. One
caught on August 3 was finishing its wing-moult,
the two outer primaries not being full-grown.
Another, caught on October 23, was probably
just finishing its wing-moult as some of the
primary coverts were growing but not the pri-
maries themselves, which were fresh. These
were the only examples of wing-moult. Eleven
of the 17 birds trapped in the months October-
January were undergoing body-moult. The
weights of these trapped birds are given in the
Appendix.

For the Gray-rumped Swift even less infor-
mation was obtained from field observation. I
once almost certainly saw a pair enter a cleft
in an Immortelle tree ( Erythrina micropteryx )
about 1,000 feet above sea level in the Arima

Valley at dusk, but the light was so bad that
when the birds, which I had had under observa-
tion for half an hour as they circled the tree,
finally entered, or appeared to enter, the hole,
I could not be certain of what I had seen.

Forty-three Gray-rumped Swifts were caught
in mist-nets, in several months of the year. An
analysis of the moults of these birds is given in
Table III. It will be seen that 9 of the 13 birds
trapped between June 14 and November 6 were
undergoing wing-moult, but none of the 29 birds
trapped from January 5 to May 26. The bird
trapped on May 26 had an incubation patch and
weighed 16 gm., the highest weight recorded
(see Appendix), and so was clearly laying eggs.
This evidence suggests that the Gray-rumped
Swift has much the same breeding season as the
Short-tailed Swift.

Table III. Moult Data for Trapped
Chaetura cinereiventris

Number of individuals showing
Body-moult

Wing-moult only No moult

Jan.

14

2

Feb.

2

April

4

6

May

1

June

1

1

July

1

Aug.

1

Oct.

3

Nov.

4

2

The Moulting Sequence in Chaetura

Nine moulting Short-tailed Swifts were ex-
amined, nine moulting Gray-rumped Swifts,
and one moulting Band-rumped Swift. The
wing-moult follows the pattern that is general in
many birds; the primaries moult in sequence
from inside outwards and the secondaries cen-
tripetally from the two ends. For the primary
moult, which spans practically the entire period
of the moult, the following stages are recognized,
using a method of notation adapted from Miller
(1961) : Stage 1, 1st (innermost) primary grow-
ing, 2-10 old; Stage 2, 2nd primary growing, 3-
10 old; and so on to Stage 9, 9th primary grow-
ing, 10th old; Stage 10, 10th primary growing.

The secondaries are much shorter than the
primaries and their replacement is much more
rapid. They do not begin to be moulted until the
primary moult is well advanced, and they finish
before the primaries finish. No difference was
noted between the Short-tailed and Gray-
rumped Swifts in this respect. Three birds were
examined whose primary moult was at Stage 6;

1962]

Snow: Biology of Some Trinidad Swifts

135

all had old secondaries. Two were examined at
Stage 7; in one the secondaries had begun to
moult, but not in the other. Three at Stage 8
were all replacing their secondaries. Two at
Stage 9 and two at Stage 10 had all finished re-
placing them.

The tail moults centripetally. It begins after
the primary moult has begun and ends at about
the time that the primary moult ends. In the

Short-tailed Swift it seems, by extrapolation
from the specimens examined, that the tail be-
gins to moult at Stage 4 or 5 in the primary
moult, but no birds were examined at these
stages. In the Gray-rumped Swift the tail moult
begins later, at Stage 7 or 8. This difference
between the two species is apparent from the
following comparison of birds at the same stages
of primary moult:

Short-tailed Swift

Gray-rumped Swift

Stage

6.

Two birds, both half way
through tail-moult

One bird, tail not moulting

Stage

7.

Two birds, about three-
quarters through tail-
moult

Stage

8.

One bird; tail-moult
nearly complete

Two birds; one just starting
tail-moult, the other half
way through

Stage

9.

One bird; tail-moult
complete

One bird; tail-moult nearly
complete

Stage

10.

—

Two birds ; tail-moult nearly
complete

Besides being less regular in its sequence than

which small streams flow under the main road

the wing-moult, the tail-moult is also sometimes
not complete Two birds that were not moulting
showed incomplete replacement of the tail-feath-
ers: in one the outermost pair of feathers, and
in the other the central pair, were old and worn,
the rest of the tail-feathers being new.

The Breeding of Cypseloides rutilus

Belcher & Smooker give records of three
clutches, from two nests, and Orton (1871)
mentions the nest-site. These appear to be the
only breeding records published for the Chest-
nut-collared Swift. During the present study ten
occupied nests were found, and two other unoc-
cupied nests. Nests are usually repaired and used
year after year, and the ten occupied nests gave
records of 33 nestings.

that runs up the Arima Valley.

The remaining nest was in a small sea cave on
the north coast. Owing to the collapse of its roof,
this cave had become a hole with a low tunnel
leading through to the sea. The nest was sit-
uated 7 feet up, under an overhang above the
landward end of the tunnel. Chestnut-collared
Swifts were also seen during the breeding sea-
son entering and leaving La Vache cave on the
north coast, a large sea cave occupied by Oil-
birds ( Steatornis caripensis) , and flying out of
a sea cave on the north side of Huevos Island,
off the northwest corner of Trinidad. There are
many caves along the north coast of Trinidad
and the off-lying islands, and doubtless Chestnut-
collared Swifts breed in them in considerable
numbers.

Nest and Nest-site

The nest is a substantial bracket, semicircular
in horizontal section, with a shallow depression
for the eggs. It is made of various plant fibres,
usually including some moss, and is fixed, pre-
sumably with saliva, onto a smooth rock-face
or wall, a few feet above water. The natural nest-
sites in the Northern Range of Trinidad are
vertical or overhanging rock-faces at the sides
of mountain streams or gorges; eight nests were
found in such sites. In addition, one nest was
found under a bridge, about 16 feet above the
water, and two were found in culverts, through

Breeding Season and Breeding Statistics

The breeding season is much the same as that
of the Short-tailed Swift; the first eggs were laid
in April or May in each year, and the last in
August. The date of laying of 32 clutches is
shown in Table IV (one nest could not be dated
accurately enough for inclusion). Is is curious
that two of Belcher & Smooker’s three clutches
fall outside these limits. One of them was accu-
rately dated August 31 -September 2, only a little
later than the latest of my nests; the other was
found on November 10. The possibility that the

136

Zoologica: New York Zoological Society

[47: 12

Table IV.

Breeding Season of Cypseloides rutilus

Number of nests started
in half monthly periods

April 2

1

May 1

3

May 2

3

June 1

4

June 2

4

July 1

2

July 2

6

Aug. 1

5

Aug. 2

4

latter was an old deserted clutch is not men-
tioned.

In all cases except one, the clutch was two
eggs. The interval between the laying of the two
eggs was not accurately determined, but at one
nest it was at least two days and the other rec-
ords are all consistent with a two-day interval
between eggs.

The single clutch of one egg was laid in cir-
cumstances that suggested a bird breeding for
the first time. The nest was begun in June, 1961,
in a site that was known not to have been used
for at least four years, as the place had been
under fairly frequent observation. Building was
slow and was hindered by heavy rain which
caused part of the nest to fall away. The nest
was finally completed by the third week of
August and the single egg was laid a few days
later.

The incubation period (from the laying of the
second egg to the hatching of the second young)
was ascertained at three nests to be 22, 22-23

and 24 days. Fledging periods were ascertained
to within a few days at four nests, and were
found to be 37-44, 38-40, 39-40 and 39-41 days.
At another nest the date of hatching was not
known, but the period from the laying of the
second egg to the flying of the two young was
found to be 60-63 days.

Two of the nests that were followed through
a complete breeding season were used only once,
ten were used twice and one was used three
times, the first two attempts being unsuccessful
and the third successful. The season would not
be long enough for three broods to be reared.

The most complete record was obtained for
one nest-site that was occupied for five successive
years, though eggs were not laid in it in the
fifth year. Two broods were reared in 1957,
two in 1958 and one in 1959. In 1960 the first
brood was reared and a second clutch was laid
but was lost soon afterwards. In the off-season
1960/61 the nest, which was only about four
feet above the water of a rock-bound pool in a
mountain stream, was washed away. Another
nest was built in April, 1961, a few feet away,
but even nearer the water; it remained empty for
several weeks, then it too was washed away in
July. The bird returned to the old site and built
a new nest in late July and August; but probably
by then the season was too far advanced and
no eggs were laid in it. The history of this nest-
site in shown graphically in Text-fig 1.

Intervals between broods (from the flying of
the young to the laying of the first egg of the
next clutch) were rather short. Those that were
known to within a range of six days were as
follows: 4-8, 5-11, 6-11, 9-11, 10-13 days. In

MAY JUNE JULY AUG. SEPT. OCT.

1957

1959

1980

1961

Text-fig. 1. Occupation of a Chestnut-collared Swift nest-site over five breeding seasons. Each nesting is
indicated by a heavy line, and the hatching date by a short bar. Dots: nest-building. For further details see
text.

1962]

Snow: Biology of Some Trinidad Swifts

137

addition there was one instance of the first egg
of the second clutch being laid 3-5 days before
the single nestling of the previous brood left the
nest. The two intervals following loss of eggs
were longer than those following successful
fledging: 13-17 and 21-23 days.

Chestnut-collared Swifts’ nests seem to be safe
from predators, and breeding success was fairly
high: 15 (63%) of the 24 nestings which could
be used for analysis resulted in fledged young.
Two clutches were soaked by water running
down the rock-face after heavy rain and were
abandoned. The eggs or young disappeared from
six nests, and one was deserted for no known
reason. Although the hatching rate was high in
the successful nests (27 out of the 29 eggs that
survived to the expected date of hatching), there
were a few losses of eggs during incubation, and
some losses of young at all stages, and the mean
number of young leaving 18 successful nests
(all of which had clutches of 2 eggs) was 1.3.

An estimation of the reproductive rate can be
made from these figures, which there is no rea-
son to think are untypical. If the mean number
of nesting attempts per year is taken to be two,
nesting success is 63%, and the number of
young fledging from successful nests is on
average 1.3, each breeding pair will rear on
average 1.6 young per year.

Adaptations to Cliff-nesting

There is suggestive evidence that birds which
nest in safe sites tend to have longer incubation
and fledging periods than related species nesting
in sites more exposed to predation, presumably
because the selective pressure for quick develop-
ment is relaxed (Lack, 1 954). This may be part
of the reason for the long incubation and fledg-
ing periods of the Chestnut-collared Swift com-
pared with Chaetura species, which are almost
certainly tree-hole nesters by origin.

If clutch-size is limited by the number of
young that can be fed, a reduction in the rate
of development of the young will enable clutch-
size to be increased. As Lack points out, there
is good evidence that this has happened: passer-
ines with safe nest-sites tend to have not only
slower rates of development but also larger
clutches than passerines with exposed nest-sites.
The small clutch of the Chestnut-collared Swift
and of other Cypseloides species (Lack, 1956),
compared with Chaetura species, is contrary to
this rule. Probably it is due to the nature of the
nest-site. Chaetura nestlings climb out of the nest
when half-grown, so that large families can be
reared from their small nests, but it would be
impossible for half-grown Chestnut-collared
Swifts to cling safely to the smooth, often over-
hanging rock-faces to which the nest is fixed,

and two full-grown young are as many as can be
accommodated on the flatfish top of the nest.
Towards the end of their nestling period they
exercise their wings while hanging to the outer
rim of the nest, a method also recorded for the
Black Swift (Bent, 1940).

An alternative explanation for the small
clutch-size in Cypseloides is that their method
of feeding is such that they are unable to find
food for more than one or two nestlings. As
mentioned above, Chestnut-collared Swifts feed
higher than the Chaetura species, and the density
of flying insects must in general decrease with
altitude. The point cannot be ruled out, yet it
seems unlikely that in many different habitats,
and over a wide range of latitude, Chaetura
species are always about three times as efficient
at collecting food as Cypseloides species.

Lack (1956) has stated that among swifts
only the Hemiprocninae and Cypsiurus develop
nestling down, presumably for warmth and cam-
ouflage, as they nest in exposed sites. But downy
young have been recorded for the Black Swift
(Bent, 1940), and in the Chestnut-collared
Swift, too, the nestling develops thick down.
Like the Black Swift, the Chestnut-collared
Swift is hatched naked and grows its down after
a few days; Table V gives details of the develop-
ment of a nestling whose exact age was known.
Presumably in Cypseloides the nestling down is
primarily an adaptation for heat conservation.
The shady and damp places where they nest
are several degrees colder than the surroundings.

Hovering Flight in Cypseloides
and Chaetura

It does not seem to have been recorded that
some swifts on occasion adopt a hovering
method of flight in which the wing action must
be more like that of a hummingbird than of a
swift in normal flight. During night visits to the
Oilbird colony in the Arima gorge, roosting
Chestnut-collared Swifts were sometimes dis-
turbed by my torchlight and would fly slowly
about the cave with a hovering flight, the body
being rather upright, until they came in contact
with a rock-face to which they could cling.
Twice I easily caught a bird as it hovered slowly
towards me and clung to my clothing. In the
same gorge in the day-time Chestnut-collared
Swifts, when disturbed from their nests, would
fly off with the usual rapid flight, as do those
disturbed from nests in more open places along
streams. Limited observations suggested that
they normally also fly to their nests with rapid
flight. Under undisturbed conditions, the hover-
ing flight is probably used by the swifts when
going to roost in caves and dark gullies at dusk.

138

Zoologica: New York Zoological Society

[47: 12

Table V. Development of Nestling Chestnut-collared Swift

Date

Age j

Description

3 June

Just hatched

Naked; skin pink; down rudiments visible as
dark spots on upper surface.

6 June

3

days

Down visible as blackish streaks, not yet
through skin.

9 June

6

days

Down beginning to sprout on upper surface.
Rhythmic clutching movements with feet
when removed from nest.

13 June

10

days

Eyes just open; gray woolly down all over
upper surface, much sparser on under-
side.

17 June

14

days

Thick down above and below. Primary and
secondary coverts up to 1 mm. out of
sheath; primaries and secondaries as long
pin-feathers.

22 June

19

days

Thick down all over body; feathers in sheath
beneath down; front half of head feath-
ered, feathers of fore-crown sooty with
rusty fringes; wing-feathers 10-15 mm.,
tail-feathers c. 5 mm. beyond sheaths.

27 June

24

days

Head well feathered; wing- and tail-feathers
well grown.

8 July

35

days

Nearly ready to fly; collar dull rufous.

11 July

38

days

Still in nest in late afternoon.

14 July

41

days

Nest empty.

and perhaps going to nests in clefts where a
straight run-in is not possible. Its value in allow-
ing safe flight in dark and confined places is ob-
vious.

Short-tailed Swifts were seen to use the same
hovering flight when rising vertically out of the
man-holes where they nested. As they rose clear
of the hole they at once changed to the usual
rapid flight and flew off low, skimming the grass.
Chaetura species must always fly thus when they
leave narrow chimneys or other cavities with a
top hole.

Summary

The breeding and other aspects of the biology
of swifts of the genera Chaetura and Cypseloides
were studied in Trinidad, West Indies. The status
of the five resident and one migrant species is
summarized.

A small breeding colony of Chaetura brachy-
ura was kept under observation for five breeding
seasons. The breeding season extended from
April to early September. Most birds made from
one to three nesting attempts in each season.
Mean clutch-size was 3.6 eggs, the incubation
period was 17-18 days, the fledging period 30-
40 days. Breeding success was low, but was con-
sidered not to be typical, owing to the unusual
nest-sites (underground sewer manholes). After
the breeding season, moulting birds often spend

part of the day clumped together in the nest-
holes.

Data are presented on weights and measure-
ments of Chaetura brachyura, C. cinereiventris
and C. spinicauda, and on the sequence of moult
in C. brachyura and C. cinereiventris.

Data are given on 33 nestings of Cypseloides
rutilus, nesting on overhanging cliffs above
streams, under bridges and culverts, and in sea
caves. The species is single or double brooded,
and repairs and re-uses its old nests year after
year. The clutch normally consists of 2 eggs.
Incubation lasts 22-24 days, and the fledging
period is about 40 days. The young is hatched
naked, but develops a thick covering of down.
The small clutch, the long development period
and the nestling down are all considered to be
correlated with the specialized nest-site.

Literature Cited

Beebe, W.

1949. The swifts of Rancho Grande, north-cen-
tral Venezuela, with special reference to
migration. Zoologica, 34: 53-62.

Belcher, C., & G. D. Smooker

1936. Birds of the Colony of Trinidad and To-
bago. Part 3. Ibis, (13) 6: 1-35.

1962]

Snow: Biology of Some Trinidad Swifts

139

Bent, A. C.

1940. Life histories of North American cuckoos,
goatsuckers, hummingbirds and their
allies. Smiths. Inst. Bull. no. 176.

Bond, J.

1928. On the birds of Dominica, St. Lucia, St.
Vincent, and Barbados, B.W.I. Proc.
Acad. Nat. Sci. Phila., 80: 523-545.

Eisenmann, E.

1955. The species of Middle American birds.
Trans. Linn. Soc. N. Y., 7.

Fisher, R. B.

1958. The breeding biology of the Chimney
Swift. New York State Mus. Bull. no. 368.

Haverschmidt, F.

1958. Schornsteine als Massenschlafplatz von
Chaetura brachyura in Surinam. J. Orn.,
99: 89-91.

Lack, D.

1954. The natural regulation of animal num-
bers. Clarendon Press, Oxford.

1956. A review of the genera and nesting habits
of swifts. Auk, 73: 1-32.

Miller, A. H.

1961. Molt cycles in equatorial Andean spar-
rows. Condor, 63: 143-161.

Orton, J.

1871. Notes on some birds in the museum of
Vassar College. Amer. Nat., 4: 713.

Roberts, H. R.

1934. List of Trinidad birds with field notes.
Tropical Agriculture, 11: 87-99.

Sick, H.

1948. The nesting of Chaetura andrei meridion-
als. Auk, 65: 515-520.

1951. Umstellung der Nistweise beim Stachel-
schwanzsegler Chaetura andrei. J. Orn.,
93: 38-41.

1958. Geselligkeit, Schornstein-Benutzung und
Ueberwinterung beim brasilianischen
Stachelschwanzsegler Chaetura andrei.
Vogelwarte, 19: 248-253.

1959. Notes on the biology of two Brazilian
swifts, Chaetura andrei and Chaetura cin-
ereiventris. Auk, 76: 471-477.

Sutton, G. M.

1951. Mexican birds: first impressions. Univer-
sity of Oklahoma Press, Norman.

Wetmore, A.

1957. Species limitation in certain groups of the
swift genus Chaetura. Auk, 74: 383-385.

Appendix

Wing-Lengths and Weights

The figures in parentheses give the number meas-
ured, and are followed by the range and the mean.

Weights were recorded immediately after cap-
ture, except where specified. Wings were measured
in the normally closed position, the primaries re-
taining their natural curvature.

Chaetura brachyura

Wing (10) 116-130, 119.9 mm.

Weight (11) 17-30. 19.8 gm.

The large weight-range is due to a single bird of
30 gm.; all the others were between 17 and 22 gm.
This very heavy bird, a female, was caught on March
22, about six weeks before breeding began; it was
very fat, the ovary being still quite small. The other
birds were all caught in August and September.

Chaetura cinereiventris

Wing (37) 98-105, 101. 5 mm.

Weight (43) 12.5-16. 13.8 gm.

Chaetura spinicauda

Wing (19) 98-107, 102.6 mm.

Weight (21) 13-18, 14.2 gm.

Chaetura chapmani

Wing (1) 120 mm.

Weight (1) 25.5 gm.

Cypseloides rutilus
Weight (2) 20, 22 gm.

The lighter of the two birds was captured in the
evening and weighed early next morning.

Cypseloides zonaris

Weight (3) 60, 63.5, 74 gm.

The two lighter birds had been knocked down
with sling-shots from a low-flying flock and had
been several hours without food when weighed.

13

The Genetics of Some Polymorphic Forms of the Butterflies
Heliconius melpomene Linnaeus and H. erato Linnaeus. I. Major Genes. 1,2

J. R. G. Turner3

Sub-Department of Genetics, The University of Liverpool

&

Jocelyn Crane
Department of Tropical Research,

New York Zoological Society, New York 60, N. Y.

(Plate I; Text-figure 1)

[This paper is one of a series emanating from the
tropical Field Station of the New York Zoological
Society at Simla, Arima Valley, Trinidad, West
Indies. The Station was founded in 1950 by the Zoo-
logical Society’s Department of Tropical Research,
under the direction of Dr. William Beebe. It com-
prises 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 principally for research in tropi-
cal 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.” by William Beebe.
Zoologica, 1952, Vol. 37, No. 13, pp. 157-184.]

Contents pAGE

I. Introduction 141

II. The Genetics of H. melpomene 142

A. Materials and Methods 142

B. Results 144

1. Inheritance of the Ray pattern. ... 144

2. Inheritance of the Dennis pattern . . 145

3. Inheritance of the width of the

bands 145

iContribution No. 1019, Department of Tropical Re-
search, New York Zoological Society.

^This study has been supported by the National Sci-
ence Foundation (G6376). A grant-in-aid from the
National Geographic Society is also gratefully acknowl-
edged. Much of the field work was made possible
through the co-operation of the Alcoa Steamship Com-
pany and the Suriname Aluminum Company.

4. Linkage of Ray, Dennis and Band

factors 145

III. The Genetics of H. erato 149

A. Materials and Methods 149

B. Results 149

IV. Discussion 150

V. Summary 151

VI. References 151

I. Introduction

THE neotropical butterflies Heliconius
melpomene and H. erato (Lepidoptera,
Nymphalidae, Heliconiinae) show com-
plex parallel polymorphism in their wing mark-
ings (Oberthiir, 1902; Eltringham, 1916; Joicey
& Kaye, 1916). In connection with research into
the biology of the heliconiine butterflies of Trini-
dad4, the genetics of several of the polymorphic
forms of H. melpomene has been investigated.
Crosses were made between several forms bred
from eggs obtained in the wild in Surinam (Dutch
Guiana) and between the Surinam stock and
insects from Trinidad, West Indies, where the
species is monornorphic. The results show that
most of the differences between the major poly-
morphic forms studied are produced by genes
in the same linkage group; analysis of previously
described broods of H. erato (Beebe, 1955)

3Present address: Genetics Laboratory, Zoology De-
partment, University of Oxford, Great Britain.

4ReIated in different degrees to the present contribu-
tion are the following: Beebe, 1955; Crane, 1955, 1957;
Fleming, 1960; Beebe, Crane & Fleming, 1960; Alex-
ander. 1961.1, 1961.2.

141

142

Zoologica: New York Zoological Society

[47: 13

shows that the mode of inheritance of analogous
patterns in this species is very similar to that in
melpomene.

This paper deals with differences produced
by major genes; a second paper will deal with
quantitative variation (Turner, in press).

The preliminary work involved in assembling
the material was long and exacting. The difficul-
ties included irregular seasonal scarcities, the
intermittent prevalence of disease, the apparent
impossibility of hand-pairing the images, the
usually rapid deterioration of the stock with in-
breeding and, finally, the fact that the larvae,
being not only non-gregarious but incompatible,
had to be reared singly. We would like, there-
fore, to express particular thanks to the people
responsible, through parts of two years at the
Trinidad Field Station, for the collection of the
stock and the rearing of the resultant broods.
Included are Mr. Henry Fleming, who collected
the breeding stock on field trips to Surinam and
who designed the breeding cages, and the fol-
lowing laboratory assistants: Mesdames Susan
Allan, Kathleen Campbell, Frances W. Gibson
and Jane S. Kinne, and Misses Constance Carter
and Diana Jeffrey.

Our appreciation for help in later stages of
the work goes to Dr. P. M. Sheppard of the
University of Liverpool for reading the draft of
the paper, to Professor K. Mather, F.R.S., for
giving us his valuable opinion on the crossover
values, and to Prof. R. J. Pumphrey, F.R.S., for
reading the script.

Finally, we express our indebtedness to the
late Dr. William Beebe for his continued and
constructive interest in the general study of the
biology of the heliconiines and for his helpful
suggestions in the course of the present section
of the work.

II. The Genetics of H. melpomene
A. Materials and Methods

1. Methods of Collection. Most of the mate-
rial for the present contribution was collected
in Surinam in December, 1958, and during the
same month in 1959. Rearing and breeding were
carried out at the Trinidad Field Station. Crosses
between Surinam material and melpomene from

Trinidad were produced by mating stock living
near the field station with broods resulting from
the eggs and larvae collected in Surinam.

In both years the Surinam collections were
made during pronounced dry seasons in the
vicinity of Moengo, a bauxite mining commu-
nity in the northeastern part of the country.
Some were taken near the wharf and others
close to the Moengo end of a 30-mile road join-
ing Moengo and Albina. The latter town borders
the Maroyne (=Maroni) River which separates
Surinam from the region of French Guiana from
which came the collection described by Joicey
& Kaye (1916). The Moengo area has become
so well known to us over a period of years that
individual food plants, located in previous sea-
sons in an area of some 20 square miles, could
be revisited daily.

Eggs and larvae of H. melpomene were ex-
ceedingly scarce during the two collecting peri-
ods under consideration, each of which lasted
one week. The total catch numbered 13 in 1958
and 17 in 1959; each year the specimens were
found in only two small areas; 12 of those taken
in 1958 came in fact from a single vine. During
each year only two or three imagos were seen
flying in the entire area of search. It therefore
seems likely that the resultant broods came from
not more than several females each season.

2. Methods of Breeding. A summary of the
material is given in Table I. As indicated, the
broods resulting from the eggs collected in De-
cember, 1958, were exceedingly subject to dis-
ease and/or genetic weakness. In the following
year therefore it was thought wise to breed pairs
in which the males and females in the genera-
tion came from localities separated by approxi-
mately 10 miles. Abnormalities, particularly
crumpled wings and the inability to emerge fully
from the pupal case, as well as a number of
forms of disease and/or abnormalities in the
larvae, frequently appear in the second inbred
generation.

Even when inbreeding is avoided, however,
as in our nongenetical studies, it is not possible
ever to plan definitely to rear particular species
of heliconiines, including H. melpomene, in any
given season, because of the frequent occurrence

Table I.

Year

]

1

Eggs Laid

p

Eggs

Hatched

%

Hatched

Larvae

Successfully

Reared

% Larvae
Successfully
Reared of
Eggs Laid

% Larvae
Successfully
Reared of
Eggs Hatched

1959

248

227

91.6%

25

10.1%

11%

I960

2,605

1,832

70.4%

777

29.8%

44.7%

1962]

Turner & Crane: Genetics of Some Polymorphic Forms of Butterflies

143

of disease. Sometimes a year of abundance is
followed by two poor seasons as in the material
under consideration; sometimes good and poor
years alternate; rarely do two good years follow
each other. Evidence is accumulating that poor
seasons occur simultaneously in Surinam and
Trinidad, and that both the annual amount of
rain and the condition of the food plants are
involved.

The eggs and larvae collected in Surinam
were brought to Trinidad by air and reared in
the laboratory by the same general method pre-
viously described (Crane, 1955, pp. 168-170;
Beebe, Crane & Fleming, 1960, pp. 113-115).
Briefly, this consists of rearing each caterpillar
in an individual dish, because of the cannibalistic
habits of the early instars and the frequent ag-
gressive behavior of larger larvae. Low slender
dishes, 60 X 28 mm. in size and with ground
glass lids, are ideal containers during the first
three to four instars. When adequate space and
glassware become a problem the early instars
are also reared in 3 X 1-inch vials, with the
uncorked mouths pushed into a tray of damp
sand. Humidity, a most important factor in the
rearing of heliconiines, is difficult to control
when the vials are used. In the stender dishes
and in the 4 x 4 X 2.5-inch glass refrigerator
dishes used for the last one or two instars,
humidity is furnished by small sections of cotton
dental wads, which are saturated and squeezed
partly dry. Larvae also drink from drops of
water left on the sprinkled leaves.

The normal foodplant of the species both in
Surinam and in Trinidad is Passiflora laurifolia
L. Eggs have been collected also in the field and
reared in the laboratory on three other species
of the same genus, P. auriculata HBK, tuberosa
Jacquin and lonchophora L.

Before the prepupal wandering phase, larvae
for genetical study are furnished with a piece
of firm wire netting laid beneath the glass cover
of the dish. Healthy larvae climb on the netting
and usually spin the pad toward the middle of
the screen, pupation normally taking place the
following morning. Pupae are kept undisturbed
for 24 hours, after which each screen with its
label is transferred to the top of a jar or tumbler
above a well-soaked cotton wad.

The approximate usual durations of the devel-
opmental stages are as follows: Egg, 4 days;
larval instars, 15; pupal, 9 to 10. Further details
are given in Beebe, Crane & Fleming, loc. cit.

The night before emergence, if the imago is
certainly not to be bred, the screen with the pupa
is transferred to the top of a cylinder made of
soft wire netting and set in a dish of damp sand.
After emergence the butterfly is allowed to dry

the newly expanded wings for several hours be-
fore being placed, wings folded, in a transparent
plastic envelope, and chloroformed. About 24
hours later, when rigor mortis has passed, the
butterfly is removed, the wings opened out flat
with forceps, and the insect returned to the en-
velope. In the absence of a drying cabinet the
envelopes may be stored in a cardboard carton
and, without uncovering, exposed daily to hot
sun or the top of a warm oven. Abundant and
constant paradichlorobenzine is a tropical neces-
sity, not only to ward off pests but to prevent
mold. The prompt spreading of the wings and
storage of the insect in fully transparent en-
velopes eliminates both pinning and traditional
spreading techniques; considerable time is thus
saved, while shipping, sorting and study are all
facilitated.

Butterflies intended for breeding are treated
as follows. Males are released, as soon as the
wings dry, into isolation cages of wire mesh
measuring at least six feet on a side. When more
than one male is kept temporarily in the same
cage, the specimens are marked with water-
proof, quick-drying lacquer. The males are al-
lowed to mature to the requisite age for mating
before being placed in a cage with a female,
since courtship is then more likely to take place,
particularly between individuals of low vitality.
No male mates before the second morning after
emergence, at the imaginal age of about 48
hours; 72 hours is the normal age in melpomene
while 96 hours is occasional. Males are placed
singly in breeding cages that measure at least
9x9x8 feet, preferably on the night before
breeding is expected to take place, so that any
shock of handling is minimized and time allowed
for recovery. The insect should never be picked
up between bare fingers or chased with a net;
a strip of wax paper will insulate wings from the
fingers or a small net may be used after the
insect has come to rest.

Females to be bred are placed either in small
isolation cages or directly in the breeding cage.
Mating can occur in this species at any imaginal
age from about 1 Vi hours to about 12 days.
Females are most attractive on the first and sec-
ond days. Males may be bred at least four times
with healthy offspring resulting. They are not
known to mate on successive days, although
mating is often repeated when one day has
elapsed between the pairings. Successful breed-
ing has resulted in males more than two months
old.

It has not been found possible to pair any
heliconiine by hand5, and they have not been

l'Dione juno, which rarely mates or oviposits in cap-
tivity, can be hand mated and will readily lay when

144

Zoologica: New York Zoological Society

[47: 13

mated successfully in cages much smaller than
the size specified above. Females will however
lay eggs in normal numbers in cages measuring
about six feet in each dimension. Single individ-
uals of either sex may be kept when necessary
for two or, rarely, three days in cages about
three feet on a side; however, they must be fed,
beginning on the second morning, with flowers
placed on a shelf near the cage top in the bright-
est corner. If the butterflies are kept longer in
such cramped quarters, their vitality proves to
be impaired when eventually they are moved to
the breeding cages.

Details of the maintenance of butterfly in-
sectaries will be found in Crane & Fleming,
1953, and Crane, 1955, 1957. The principal
factors ensuring success, regardless of cage size,
are: abundant natural flower food; abundant
humidity, maintained through sprinkling when
necessary; foliage, or at least growing grass,
inside the cage; and ample sun and shade.

No female has ever been found to lay eggs on
any plant not belonging to the genus Passiflora.
The maximum number of eggs laid by a single
H. melpomene was 196 over a period of fifty-
five days. One to four eggs are usually laid per
day, rarely more, beginning on the eighth to fif-
teenth day after emergence. The major egg-
laying period is finished after about two weeks;
later in the female’s life eggless days may be-
come frequent. Eggs are collected every after-
noon to avoid predation by ants.

It is doubtless unnecessary to emphasize the
importance for care and detail in labelling at
every step of the study, from the gathering of
eggs, whether wild or cage-laid, to the final
labels on the envelopes holding the imagos.
Self-adhesive labels are used on the rearing
dishes while metal garden markers work well
on breeding and egg-laying cages out-of-doors.

3. Methods of Study. The wing pattern of mel-
pomene is made up of a number of elements
which vary independently and which occur in
most if not all possible combinations, each one
of which has a separate taxonomic name (Seitz,
1913; Joicey & Kaye, 1916). Rather than be-
come involved in nomenclatural difficulties, we
have adopted the procedure recommended by
Camp and Gilmour (Gilmour, 1958) in another
context, of keeping the genetic and taxonomic

confined in a black silk-organza sleeve. Heliconius nu-
mata and H. erato have been hand mated and the latter
has laid in a sleeve. However, both species are refractory
in these respects and the techniques are not recom-
mended unless the normal methods have failed.

names completely separate; this system also has
the advantage that the same notation may be
used for H. erato. The various pattern elements
we studied are therefore designated by English
names, with appropriate symbolic letters, as
follows:—

Ray (R) : the presence of four to six
red radiate marks on the discal part of
the upperside of the hindwing (PI. I,

Fig. 1);

nonray (r) : the absence of such red
rays (PI. I, Figs. 2-5);

Dennis (D) : the presence of extensive
red areas at the base of the forewings
on the upper and under sides, and on
the base of the hindwings on the up-
perside (PI. I, Figs. 1-3);
nondennis (d) : the absence of such
red areas (PL I, Figs. 4-5);

Wide-band (B) : the presence of a
broad red band extending from the
costal edge of the forewing towards
the inner angle (PI. I, Figs. 1, 2 & 4);
narrow-band (b) : the presence of a
narrow red band extending from the
costal edge of the forewing toward the
inner angle (PI. I, Figs. 3 & 5).

In addition there was much quantitative vari-
ation in the width and shape of the forewing
bands and in the amount of yellow pigment in
the discal area of the forewing, which will be
discussed in the second paper (Turner, 1962).

In the course of the present study, Crane has
been responsible for the Trinidad aspects, in-
cluding methodology, pair selection and super-
vision of rearing and recording. The work of
analysis, on the other hand, has been accom-
plished altogether by Turner at the University
of Liverpool.

B. Results

Data for all the broods reared will be found
in Tables II and III. In the light of our findings
we have deduced the genotypes of most of the
parent butterflies, although in some instances,
marked with an asterisk, it has been possible
only to indicate the most likely genotype, as-
suming that crossing-over has not occurred. Note
that as the Trinidad population is monomorphic
all Trinidad butterflies must be homozygous at
the major loci.

1. Inheritance of the Ray Pattern. The only
brood in which this pattern occurred (Table XI,
1 ) shows that it is probably inherited as a single
factor, one of the parents of the brood being
a recessive homozygote, the other a heterozy-
gote, but it is not possible to tell from the data

1962]

Turner & Crane: Genetics of Some Polymorphic Forms of Butterflies

145

Table II. Breeding Data for H. melpomene (First Series). (1959)

$ Parent

9 Parent

Brood

Brood No.

Brood or
Origin

Phenotype

Origin

Phenotype

BDR

Bdr

bdr

Total

8

0

1

0

1

1

Surinam

Bdr

Surinam

BDR

$

3

4

0

7

3

5

0

8

$

0

2

6

8

2

Surinam

Bdr

Surinam

bdr

2

0

3

2

5

0

5

8

13

3

Surinam

Bdr

Surinam

Bdr

8

0

2

0

2

4

3

Bdr

Trinidad

Bdr

8

0

1

0

1

5

Surinam

Bdr

Trinidad

Bdr

8

0

1

0

1

whether the gene is dominant or recessive; the
use of the capital R for the Ray pattern is there-
fore provisional. For reasons explained under
Section 4 below, it is apparent that the locus
is not sex-linked.

2. Inheritance of the Dennis Pattern. A large
number of broods show that this pattern is pro-
duced by a single dominant gene which is not
sex-linked. For example, in Table III, broods

3, 6 and 27, all of which are D x D, give a
satisfactory approximation to the ratio 3 D:1 d.
The dominance of the D gene is confirmed by
brood 23, and broods 14, 22, 24, 26 and 32
among others conform to the 1:1 backcross
ratio, so confirming that a single gene is in-
volved. Broods 22, 24 and 26 also show that the
gene is autosomal, for if it were sex-linked all
males would be nondennis and all females
Dennis. The anomalous individuals in brood 7
are assumed to be the result of an error.

3. Inheritance of the Width of the Bands.
Similarly the difference between a Wide-band
and a narrow-band is produced by a single auto-
somal locus, the factor for Wide-band being
dominant. In Table III, among others, broods

4, 9, 10 and 22 are seen to be F2 generations,
corresponding to the 3:1 ratio, broods 13, 23
and 29 are backcross generations and brood 13
demonstrates the absence of sex-linkage. Again
the anomalous individual (brood 28) is assumed
to result from an error.

4. Linkage of Ray, Dennis and Band Factors.
The three loci controlling the patterns Ray and
Dennis and the width of the band are all in the
same linkage group. Thus in Table II, brood 1
is apparently a backcross for the Ray and Dennis
factors and departs significantly from independ-
ent assortment (P < .01 by Fisher’s exact test) ;

there are no crossovers. The maximum cross-
over value, estimated on the assumption that if
one more butterfly had emerged it would have
been a crossover, is 11%, with a standard error
of ± 11%. It is therefore highly unlikely that
the COV is more than 30% and it may be much
lower. We stated earlier that the Ray locus was
not sex-linked; the reason for this conclusion
is now obvious.

The data themselves do not exclude the pos-
sibility that the Ray locus is on a separate
chromosome from the Dennis locus, but that
the R gene can only express itself in the presence
of the D gene; however we think this unlikely,
as there is a variety of H. melpomene in which
the Ray pattern occurs independently of the
Dennis pattern (var. contiguus, see Eltringham,
1916, plate XII, 26). As this variety apparently
occurs only in Ecuador (Joicey & Kaye, 1916:
p. 420), it is possible that the COV between
D and R is very low.

Similarly the D and B loci are linked, as is
shown by broods 22 and 24 (P for independent
assortment less than .0003). If crossovers had
occurred they could have been detected by in-
spection of the phenotypes in broods 4, 15, 17,
22, 24, 27 and 30 (Table III); in these broods
no butterflies have appeared in the crossover
class (bd) and as at least one, and in some
broods probably both parents in each cross are
repulsion heterozygotes, it is not possible to
obtain a reliable estimate of the crossover value.
A total of 11 butterflies from broods 2, 14, 17
and 32 (Table III) has been crossed in such a
way that their genotypes can be determined
from their progeny; of these, 3 are carrying
crossover chromosomes, two of them bd and
one BD (the original chromosomes in the crosses

146

Zoologica: New York Zoological Society

[47: 13

EQ

J

3

H

o

SO

Os

K

M

CO

D ^
Z ^

o w
y >

ul CO
CO H

a

i «

o. £
g 2

si

BS

O

(L,

<

S< £

Q .3
o s
5 <

a

a

CQ

.a

.3

CO

O

H

x

x>

x

05

Q

05

O

c

<u

O

X

Oh

°.S

o.£P

o

£ °

CQ

o

a

<L>

a

o

a

a>

-C

Ok

o a

■g:§>

£ °

05

v-> ho m ©\ mh N r*o

( h-n h-i ,-1 fvi ts h-i hr

rf c<i
H-l fS

© © © © © © © © © © © © © © © o © © o o ©

© h-t *-! (S m

© © © © © © ON N © © © © © © © © © © © © © © © © © i©

r') f*> h© oo © 00 -in N © © © HH(fi W n N

O N |N £0 VO ^ OKI kl -* VO ho © O jO\ © *- Of © © I© © © © © © I©

Of <OOf «0 Of foot «OOf *0 Of «0 Of foos foot

<0 Of

x

CO

x

05

s x

S® CQ

§°s

Q

00

Q

03

x

CQ

Q

CQ

Q

CQ

•a

CQ

\

X

cq

x

03

x

CQ

XI

cc

X

as

X

CQ

X

CQ

CQ .3

9

ia

B

. CQ

UQ .3

X

05

D

X

a

CO

X

03

Q

XI

Q

CQ

x

03

©

Q

CQ

X

CO

Q

a

Q

CQ

X

CQ

X

CQ

X

«

co .3 Q .3

cl

m

(A

9

IA

X

©

X

CQ

s i
S-B
' £

•O

x>

T3

00

A

■§

OO

“M” Bd Bd/bd “H” Bd Bd/bd

Surinam Surinam

1962]

Turner & Crane: Genetics of Some Polymorphic Forms of Butterflies

147

3

.5

c

o

O

c

no

ON

u

2

u

C/3

Q

Z

O

u

BJ

co

ir^ r-

loo *■

_ _4

^ ON jo -

-> loo vo

*0 |*-H

rr, r-i

!■’*■ ■n t-~ l<

! m C4

1*0

1

ro

ro r-

<N m

r 1’

73

o

H

x jcoo!of«N^-j'4-oo|ooo oo|oooojoooo!ooojooo|c

Q

©

03

Q

m

o © © © oo|o © © © o o

O >o >o o o o lo © © © I© 'O NO

© ©©j© ©©I© >— rr |i/-i ® — ’ r-ir*N|Tj- <-i © © I© © ©Oj© oooojNO vi |r~

<o <oa <o o+ «o ch- «o o *ooi ot <oo> o *o c* *o o *o o

rt'

£
o
a

d>
1-1
ctf

<*> o

c 3

- 3

o V5

^ 13
^ ■*“*
S .9

^ i2

(J

K CO

o a

s «

Q

o

Z

03

CL,

o

a

o

a

<1)

O

G

4>

XI

CU

O G

■gs*

CO

IS *0"

« 3

T3

00 X

"D

©

«

©

CQ

© ©

©

03

©

03

©

00

« i

f4 -C

3

CO

© ©

go as

Q D

X) X

©

03

Q

X)

Q Ci Q

00 03 03

s

D ©'

•° 03

0 2
os «

T3 TJ

03 03

*0 '■a

03 CQ

-o "O

03 03

Tj

e°

D

X

D

03

Q

x

Q

x

C

x

^ Tf

Bi

J

CD

H

«>

£

O

G

O

o

i

o

G

a>

X

Oh

°.9

■gf

25

co

•o

03

2

©

03

s «

©

CQ

* •*
-a T3
03 CD

■*• *

’O *0
CQ 02

•O

X

\

’O

X

T3

CQ

T3

X

Q

X

Q ©

x “

Q

X

TD rO

CQ 03

\ \

"O

GC3 CQ

T?

03

T3

03

T5

2

Q

Q ©

Q

-a *u

T3 "O

-O

X)

CQ

00

DO oa

00

03 CQ

03 CQ

03

CD

<NI ^3

\n &

|

**•«

5-. ctf

2. -2
s.s

On

b ■£

- w-

0

x -H

" s.s

3 u

• G
C/5

o *=*

<M fM

(M

n

148

Zoologica: New York Zoological Society

[47: 13

•g

ce

■8

G

a

a

o

U

o

VD

Os

B)

GO 2

Q S

o O

U 4>

to “>
w <u

5

o

ft.

a:

ec

0

a.

<

H

•<

Q

o

z

5

Etf

eg

o

H

•a

x

Q

X

~o

CQ

Q

cq

On v> Kc io h \o N r) ht o\ n n tt n© © oooooo oo m b— i ct m 1/-1 nc n

■'O' On Cl fN pt ih h— 1 »-i (v| — 1 m c*n rn CJ >0

O O [O OOO cefli h- OOO

C4 m «-i On OOO O © |© OOjO

ro C4 >0 •Nt O ItJ- OOO t N Kc <-1 o |*i cn f- © © <M |cs N m k. re ~ ,-a-

| | | |^H _ I —I ffN ^ U")

CS r<-. — • *- tM OOO ^ O >0 O © jo r-l ■'t r' no O NO O O j© n « |Nf

♦00 +0 Ot <0 ot «OOt <fo Of <00 «0 Ctt- ♦OO *OCK

o

a

o

o

o

G

<u

X3

o<

Q

Xi

•a

x

Q

x

Q

x

G

x

Q

x

•a

Q

x

CP

G

CP

■o

x

0

x

Q

x

-a

x

Q

x

Q

x

-a

Q

x

CP

0

CQ

-o

CQ

— .

■o

CQ

•o

CQ

° S
o.g1
o i;

£ 0

CQ

s

8* - 1

S o r ^0

•§ 2 ^ S

3 CO C

CO f-

©

CQ

■o

CQ

•v

03

”D

ca

C T3

^;a

o

e

<u

O

X

CQ

^.a

O .£P
0 Js
£ O

CQ

■O

aa

■o

CQ

T>

CQ

•o

X

•o

X

•o

X

”0

Q

X

CQ

Q

03

G

x

G

x

■a

x

■o

CQ

•V

CQ

•a

a

X

CQ

Q

CQ

-a

CQ

T3

CQ

■O

CQ

•a

*c

Q

x

Q

CQ

E

r ca

Q.S

J 1m

D

00

T3

I1

♦Indicates that the genotype is not known for certain. The genotype given is the most probable, assuming that crossing over has not occurred.
Words and letters within quotation marks are the designations of wild-caught butterflies.

Discrepancies in the total numbers of offspring have been produced by insects which could not be scored for sex.

1962]

Turner & Crane: Genetics of Some Polymorphic Forms of Butterflies

149

being hD and Bd). Again the amount of infor-
mation is so small that no reliable estimate of
the crossover value can be obtained.

From Tables II and III it will be seen that the
following chromosomes, genotypes and pheno-
types have been observed in our broods (not
considering the R locus) :

Chromosomes: BD, bD, Bd and bd;
Genotypes: BD/Bd, Bd/Bd, bD/bD, Bd/bd,
bD/bd, bD/Bd and bd/bd;

Phenotypes: BD, bD, Bd and bd.

That is, all possible phenotypes and chromo-
somes occurred, and all but three of the possible
combinations of those chromosomes. Taking the
Ray pattern into account, the following pheno-
types were found:

BDR, BDr, bDr, Bdr, bdr.

Other phenotypes have been found in the wild
(loicey & Kaye, 1916).

III. The Genetics of H. erato
A. Materials and Methods
Data presented by Beebe (1955) of broods
reared from wild individuals of H. erato give
information about the inheritance of various
elements of the pattern in this species. The fol-
lowing symbols have been used in tabulating his
results:—

B: a wide forewing band, as in mel-
pomene;

Bb: a complicated, broken forewing
band, different from anything in the
melpomene material;

D: red pigmentation of the base of the
forewing, similar to that of Dennis in
melpomene, but lacking the basal red
on the hindwing;

d: absence of this red pigmentation;

R: Ray patterns, similar to those in
melpomene, but extending into the
basal area of the hindwing;

r: the absence of such red rays;

(y) : indicates that the feature with
which the symbol is bracketed is yel-
low instead of red; thus if B is a red
forewing band, (By) is a yellow one;
w: presence of prominent white spots
in the forewing band.

The difference between the Dennis and Ray
patterns in the two species is worthy of special
note and is illustrated in Text-fig. 1; for illus-
trations of the other patterns see the plates in
Beebe’s paper.

B. Results

Beebe’s results are summarized in Table IV
and permit the following conclusions:—

1. The female parent of Brood B, being from
Trinidad where the species is monomorphic, must
have been homozygous for all major wing-pat-
tern loci. This means that in this brood the fac-
tors for the D pattern, R pattern and Bb pattern
(Broken-band) were all dominant and the male
parent was homozygous. Unless complex factor
interaction is involved, the gene for yellow col-
oring of the forewing band is recessive to the
gene for red coloring.

2. It is therefore likely that brood D is a back-
cross for the D and R factors, the unknown male
parent having been a heterozygote (P > .5).
D and R could be recessive, as the fit to an F2
ratio is also good (P > .2), but in view of the
result for brood B this seems unlikely. The
brood gives a strong suggestion of linkage be-
tween the D and R loci. Assuming that the
brood is a double backcross, P for independent
assortment is less than .001, and the maximum
value for the crossover rate is 8% ±7%. The
data do not exclude the possibilities that (a)
Dennis and Ray are pleiotropes of one gene, or
(b) D and R are on separate chromosomes but
that one can only express itself in the presence of
the other; though both these alternatives seem
unlikely because of the existence of specimens

Text-fig. 1. Difference between Ray and Dennis elements in H. melpomene
(left) and H. erato (right). Ray element white, Dennis element stippled.

150

Zoologica: New York Zoological Society

[47: 13

Table IV. Breeding Data for H. erato.
(Beebe’s Data).

Brood

Letter

$ Parent
Phenotype
and Origin

9 Parent
Phenotype
or Genotype
and Origin

Brood

Bb DR

Bb DRw

(Bby) Dr

BDR

Bdr

1

Total

A

?

BbDr

9

0

0

1

0

0

1

Surinam

Surinam

B

(Bby)DR

Bdr/Bdr

13

0

0

0

0

13

Surinam

Trinidad

C

?

Bdr

$

1

0

0

0

0

1

Surinam

Surinam

9

0

1

0

0

2

3

T

T

~o

~o

T

4

D

?

Bdr

$

0

0

0

3

5

8

Surinam

Surinam

9

0

0

0

2

3

5

~o

"o

~o

T

8

13

E

?

Bdr

$

0

0

0

0

2

2

Surinam

Surinam

9

0

0

0

0

2

2

"o

~o

~o

~o

“4

~4

Insects designated “?” were the wild mates of the females and were never seen.

in which Dennis occurs without Ray (vars.
dry ope and cybelina; see Seitz, 1913, PI. 78) and
in which Ray occurs without Dennis (var. vesta\
see Seitz, 1913, p. 393).

3. Brood C therefore also provides evidence
of linkage between the D and R loci, and be-
tween these two loci and the factor, apparently
Bb, which is affecting band-width in this brood;
it is not possible to prove or estimate the linkage
( P for independent assortment greater than .1).
It is unlikely that the Bb and D patterns are
pleiotropes of the same gene, or appear linked
because of factor interaction, as they can occur
independently (vars. callicopis and erythraea;
Seitz, 1913, PI. 78).

4. Nothing can be said about the inheritance
of white coloration.

5. There is no evidence for or against sex-
linkage.

6. One reservation must be made : the appar-
ent linkage of the various loci could have been
produced even if the loci were not linked, if
the females had mated twice in the wild; double
matings have been suggested by Ford (1936) to
explain anomalous broods reared from wild
Papilio females.

IV. Discussion

The polymorphisms of Heliconius erato and
H. melpomene are of exceptional interest from
a number of points of view. The species are
undoubtedly aposematic (Crane and refs., 1955,
preliminary observations; and L. P. Brower,
J. V. Z. Brower & C. T. Collins, in prep., experi-

mental study). Almost certainly the species are
also Mullerian mimics and are involved in mim-
etic relationships of amazing complexity with
numerous other Lepidoptera (Eltringham,
1916). Again, the hue red, which in the most
usual form of both species is present as a fore-
wing band, has high value in releasing courtship
behavior; in fact an all-red model of about the
size of a normal butterfly acts as a supernormal
stimulus (Crane, 1955, on H. erato and in prep,
on H. melpomene.) Finally, evidence is emerg-
ing (Beebe, Crane & Fleming, 1960; Alexander,
1960.1; and Crane et al. unpubl.) that H. mel-
pomene and H. erato are closely related.

In those features studied so far, both poly-
morphisms are apparently controlled by loci in
the same linkage group and are therefore poly-
morphisms of the type described by Sheppard
(1953). At least two of the elements resulting
in more extensive red markings than usual are
dominant in both species, while that responsible
for much reduced forewing bands is recessive.

We have emerging, therefore, in the study of
these two species a number of related factors
to be considered in our further examination of
the biology and evolution of these and other
heliconiines. First, the probable value of red as a
warning hue associated with aposematism. Sec-
ond, the definite value of the same hue in court-
ship behavior, additional amounts of red having,
up to a certain size, extra stimulating value in
experimental situations. Third, the dominance,
as shown in the present contribution, of certain
elements responsible for red in excess of the
usual forewing band. Fourth, the apparent oc-

1962]

Turner & Crane: Genetics of Some Polymorphic Forms of Butterflies

151

currence of polymorphic forms exhibiting more
than a single red forewing area of moderate size
in few geographic localities, compared with the
wide distribution of the species. Fifth, our field
observations (unpublished and incomplete) in
Surinam, indicating that individuals with the
larger areas of red markings are uncommon to
rare in the field. Sixth, mimicry. Seventh, multi-
locus polymorphism with linked genes.

V. Summary

1 . The butterflies Heliconius melpomene and
H. erato are highly polymorphic; this paper, the
first of two, describes differences produced by
major genes.

2. In H. melpomene three major loci, all
linked on the same autosome, have been dis-
covered. They are:

B; affecting the width of the band on
the forewing;

D: producing a red suffusion on the
base of fore- and hindwing; and
R: producing red rays on the hind-
wing.

One of the crossover values appears to be
low (less than 30%); the other has not been
determined, although crossovers have occurred.

3. In H. erato the evidence is less certain but
it seems that there are three loci, all linked,
which produce patterns analogous to those pro-
duced by the three linked loci in melpomene.
There are slight but important differences in the
effects of D and R in the two species, and the B
locus gene which narrows the band is dominant
in erato but recessive in melpomene.

4. In erato there appears to be a single re-
cessive factor which changes the color of the
forewing band from red to yellow.

5. The interest of the polymorphisms in con-
nection with courtship behavior, aposematism
and linkage are noted.

VI. References
Alexander, A. J.

1961.1. A study of the biology and behavior of
the caterpillars, pupae and emerging but-
terflies of the subfamily Heliconiinae in
Trinidad, West Indies. Part I. Some as-
pects of larval behavior. Zoologica, 46:
1-24.

1961.2 A study of the biology and behavior of
the caterpillars, pupae and emerging but-
terflies of the subfamily Heliconiinae in
Trinidad, West Indies. Part II. Molting,
and the behavior of pupae and emerg-
ing adults. Zoologica, 46:105-123.

Beebe, W.

1952. Introduction to the ecology of the Arima
Valley, Trinidad, B. W. I. Zoologica, 37:
157-184.

1955. Polymorphism in reared broods of Heli-
conius butterflies from Surinam and Trin-
idad. Zoologica, 40: 139-143.

Beebe, W., J. Crane & H. Fleming

1960. A comparison of eggs, larvae and pupae
in fourteen species of heliconiine butter-
flies from Trinidad, West Indies. Zoo-
logica, 45: 111-154.

Crane, J.. & H. Fleming

1953. Construction and operation of butterfly
insectaries in the tropics. Zoologica, 38:
161-172.

Crane, J.

1955. Imaginal behavior of a Trinidad butter-
fly, Heliconius erato hydara Hewitson,
with special reference to the social use of
color. Zoologica, 40: 167-196.

1957. Imaginal behavior in butterflies of the
family Heliconiidae: Changing social pat-
terns and irrelevant actions. Zoologica,
42: 135-145.

Eltringham, H.

1916. On specific and mimetic relationships in
the genus Heliconius. Trans. Ent. Soc.
Lond., 1916: 104-148.

Fleming, H.

I960. The first instar larvae of the Heliconiinae
(butterflies) of Trinidad, West Indies.
Zoologica, 45: 91-110.

Ford, E. B.

1936. The genetics of Papilio dardanus Brown
(Lep.). Trans. R. Ent. Soc. Lond., 85:
435-466.

Gilmour, J. S. L.

1958. The species: yesterday and tomorrow.
Nature, 181: 379-380.

Ioicey, J. J., & W. J. Kaye

1916. On a collection of Heliconine forms
from French Guiana. Trans. Ent. Soc.
Lond., 1916: 412-431.

Oberthur, C.

1902. Observations sur la variation des Heli-
conia vesta et thelxiope. Etudes d’En-
tomologie, 21. 26 pp., 11 pis.

Seitz, A.

1913. Heliconiinae. In Seitz, A. (ed.), 1924. The
Macrolepidoptera of the World, 5. Stutt-
gart.

Sheppard, P. M.

1953. Polymorphism, linkage and the blood
groups. Amer. Nat., 87: 283-294.

152

Zoologica: New York Zoological Society

[47: 13: 1962]

Turner, J. R. G.

1962. The genetics of some polymorphic forms
of the butterflies Heliconius melpomene

Linnaeus and H. erato Linnaeus. Part
II. “Continuous” variation. Zoologica
(in press).

EXPLANATION OF THE PLATE

Plate I

Specimens of Heliconius melpomene showing the
major pattern elements discussed in this paper. All
the butterflies are red and black, except that the dif-
fuse C-shaped or triangular mark in the forewing
cell of Figs. 3 and 5, and two small areas near the
costal edge of the forewing of Fig. 3, are yellow;
these yellow marks will be discussed in the second
paper (Turner, 1962).

Fig. 1.
Fig. 2.
Fig. 3.
Fig. 4.
Fig. 5.

Wide-band, Dennis, Ray (BDR).
Wide-band, Dennis, nonray (BDr).
Narrow-band, Dennis, nonray (bDr).
Wide-band, nondennis, nonray (Bdr).
Narrow-band, nondennis, nonray (bdr).

TURNER a CRANE

PLATE I

FIG. 1

FIG. 2

FIG 3

FIG 4

FIG. 5

THE GENETICS OF SOME POLYMORPHIC FORMS OF THE BUTTERFLIES HELICONIUS
MELPOMENE LINNAEUS AND H. ERATO LINNAEUS

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

630 Fifth Avenue, New York 20, N. Y.

PUBLICATION OFFICE

The Zoological Park, Bronx 60, N. Y.

OFFICERS

PRESIDENT VICE-PRESIDENT
Fairfield Osborn Laurance S. Rockefeller

SCIENTIFIC STAFF:

William G. Conway . . Director,

Zoological Park

Christopher W. Coates . . Director, Aquarium
John Tee-Van ...... General Director

Emeritus

ZOOLOGICAL PARK

Joseph A. Davis, Jr. . . Curator, Mammals

Grace Davall Assistant Curator,

Mammals and Birds
William G. Conway. . Curator, Birds

Peter Ames Assistant Curator,

Birds

Herndon G. Dowling . Curator, Reptiles
Charles P. Gandal . . . Veterinarian
Lee S. Crandall ..... General Curator

Emeritus

Roland Lindemann .... Consultant in Mam-
mal Management

AQUARIUM

James W. Atz Curator

Carleton Ray Associate Curator

Ross F. Nigrelli Pathologist & Chair-

man of Department
of Marine Biochem-
istry & Ecology

C. M. Breder, Jr Research Associate

in Ichthyology

Harry A. Charipper. . . Research Associate

in Histology

SECRETARY TREASURER

George W. Merck David H. McAlpin

Sophie Jakowska Research Associate

in Experimental
Biology

Klaus D. Kallman. . . . Research Associate

in Genetics

Louis Mowbray Research Associate

in Field Biology

GENERAL

William Bridges . . Editor & Curator,

Publications

Dorothy Reville . . Editorial Assistant

Sam Dunton ..... Photographer

Henry M. Lester. . Photographic Consultant

DEPARTMENT OF TROPICAL
RESEARCH

Jocelyn Crane Directing Curator

John Tee-Van Associate

William K. Gregory .... Associate

AFFILIATE

L. Floyd Clarke Director,

Jackson Hole Biological Research Station

EDITORIAL COMMITTEE
Fairfield Osborn, Chairman
James W. Atz Lee S. Crandall

William Bridges Herndon G. Dowling

Christopher W. Coates John Tee-Van
William G. Conway

f] ! I’' -j 1 1

, ■ ' •. ■ •, •• " ■. • ?v.,„ i i •

7. > - . <\. 1 n ’ **t & h i

; : : •% ?tir ■:■■■ .v>-> d ’ rft-f • • 1 \U»4^HM||pi4

S^o s'! 3

ZOOLOGICA

SCIENTIFIC CONTRIBUTIONS OF THE
NEW YORK ZOOLOGICAL SOCIETY

VOLUME 47 • QUARTERLY PART 4 • WINTER, 1962

PUBLISHED BY THE SOCIETY
The ZOOLOGICAL PARK, New York

Contents

PAGE

14. Effects of Hybridization on Pigmentation in Fishes of the Genus Xipho-

phorus. By James W. Atz. Plates I-VIII 153

15. A Field Study of the Golden-headed Manakin, Pipra erythrocephala, in

Trinidad. By D. W. Snow. Text-figures 1-6 183

16. The Natural History of the Oilbird, Steatornis caripensis, in Trinidad. W.I.

Part 2. Population, Breeding Ecology and Food. By D. W. Snow. Plates
I-IV; Text-figures 1-4 199

Index to Volume 47 223

Zoologica is published quarterly by the New York Zoological Society at the New York
Zoological Park, Bronx Park, Bronx 60, N. Y., and manuscripts, subscriptions, orders for back
issues and changes of address should be sent to that address. Subscription rates: $6.00 per year;
single numbers. $1.50, unless otherwise stated in the Society’s catalog of publications. Application
to mail at second-class postage rates is pending at Bronx, N. Y.

Volume 47, Part 3 (Fall, 1962), was published on November 29, 1962.

14

Effects of Hybridization on Pigmentation
in Fishes of the Genus Xiphophorus

James W. Atz

New York Aquarium, New York Zoological Society 1
and New York University 2

(Plates I-VIII)

Contents page

I. Introduction 153

II. Materials and Methods 154

III. Results 156

1. Inheritance of Monomorphic Pigmen-
tary Patterns in Hybrids 156

2. Inheritance of Polymorphic Pigmentary

Patterns in Hybrids 158

a. Micromelanophore Polymorphic Pat-
terns 158

b. Macromelanophore Polymorphic

Patterns 161

IV. Discussion 169

1. Genetics of Micromelanophore Pigmen-

tary Patterns as Revealed by Their
Appearance in Hybrids 169

2. Genetics of Melanosis and Melanoma in

Hybrids 170

a. Capacity for Atypical Growth and

Specificity of Macromelanophore
Genes 170

b. Evidence for the Evolution of Macro-

melanophore Genes and Their Poly-
genic Modifiers 172

c. Atypical Pigmentation Associated

with the Sc Gene 174

3. Further Aspects of Hybridization 175

V. Summary 176

VI. References 177

I. Introduction

This paper is dedicated to Myron Gordon.
The dedication is particularly appropri-
ate for a number of reasons. Most of the
hybrids treated herein were produced under

1 From the Genetics Laboratory of the New York
Aquarium, New York Zoological Society. This Labora-
tory is supported by grants from the National Cancer
Institute, National Institutes of Health, U.S. Public
Health Service, and is located at the American Museum
of Natural History, New York, N. Y.

2 From a dissertation submitted to the Graduate
School of Arts and Science, New York University, New
York, N. Y., in partial fulfillment of the requirements
for the degree of Doctor of Philosophy.

Dr. Gordon’s direction, and publication of this
work was one of many projects left unfinished by
his untimely death in 1959. Dr. Gordon had
caught the parental stock from which the crosses
were made, founded the unique Genetics Labora-
tory of the New York Aquarium in which they
were made and, most important of all, provided
the rationale under which their existence and
study assumed significance. It is especially fitting
that the paper appears in the scientific journal of
the New York Zoological Society; Dr. Gordon
was intimately associated with the Society for
the major part of his career and numerous scien-
tific reports by him and his associates were pub-
lished in Zoologica over the years. Finally, deep
personal feelings add another dimension to the
dedication. Myron Gordon was a dear friend as
well as mentor. It was he who suggested the pig-
mentation of hybrids as a subject for a doctoral
thesis and made available the fishes and facilities
of the Genetics Laboratory.

Myron Gordon first recognized the remark-
able combination of features that makes the
fishes of the genus Xiphophorus so worthy of
study (Atz & Rosen, 1959). Not the least of
these is their ability to hybridize with one an-
other. In fact, it was the melanotic hybrids of
X. maculatus—X. hellerii that first brought these
fishes to the attention of biologists and medical
men (Atz, 1941). Because they have been more
readily available for experimentation than the
other species of Xiphophorus, because their hy-
brids often develop melanoma and because X.
maculatus shows an unusual but clear-cut sex-
linkage, these two species have been the subject
of many more investigations than any of their
congeners. Dr. Gordon’s broad approach to the
problems of comparative oncolgy, however, in-
cluded a study of all the then known species of

153

154

Zoologica: New York Zoological Society

[47: 14

Xiphophorus and their hybrids.3 As can be seen
in Table I, nearly two-thirds of the 28 crosses
recorded up to the present were first made by
Dr. Gordon (see the papers by Gordon, Gordon
et al. and Rosen).

The author wishes to thank Dr. Donn E.
Rosen for invaluable help received during many
long and spirited discussions. Dr. Klaus D. Kali-
man, Research Associate in Genetics of the
New York Aquarium, who succeeded Dr. Gor-
don in direction of the Genetics Laboratory, also
provided valuable assistance, including a critical
reading of the manuscript, and this is gratefully
acknowledged. Special thanks are due the Amer-
ican Museum of Natural History, in particular
the Department of Birds, for their most generous
provision of space and facilities.

II. Materials and Methods

The hybrid fishes upon which this study is
based have been produced over a period of more
than 25 years, but the great majority of them are
the result of crosses set up since 1939 in the
Genetics Laboratory of the New York Aquari-
um. Almost all of the stocks of fishes that have
been used were derived from specimens col-
lected alive in their native Mexico or British
Honduras, thus insuring purity of ancestry, for
domesticated fish, obtained from pet stores or
commercial breeders, almost invariably have a
hybrid somewhere among their progenitors.

The following is a list of the strains used in
the present studies, the geographical area from
which they were taken, and the expedition re-
sponsible for collecting the foundation speci-
mens:

Xiphophorus couchianus (Girard, 1 859) 4

Rio Santa Catarina, Nuevo Leon (1939) —
Gordon, Atz, Evelyn Gordon; (1958)— Gordon,
Evelyn Gordon.

Xiphophorus variants xiphidium
(Gordon, 1932)

Rio Purificacion, Tamaulipas (1939)— Gor-
don, Atz, Evelyn Gordon.

3 In accord with this approach, the recent discovery
of two new species, X. clemenciae and X. milled, and four
new subspecies (Rosen, 1960) opens up a wide field of
comparative genetics.

4 With the exception of the spelling of hellerii, these
scientific names are the same as the ones used by Rosen
(1960) in his comprehensive revision of the teleost genus
Xiphophorus (Family Poeciliidae, Order Cyprinodonti-
formes). The orthographic change is required by the
International Code of Zoological Nomenclature adopted
by the XV International Congress of Zoology and pub-
lished in 1961.

Rio Santa Engracia, Tamaulipas (1958) —
Gordon, Evelyn Gordon.

Xiphophorus variatus variants (Meek, 1904)
Rio Axtla, San Luis Potosi (1939)— Gordon,
Atz, Evelyn Gordon.

Xiphophorus variatus evelynae Rosen, 1960
Rio Necaxa, Puebla( 1957)— Rosen, Malcolm
Gordon, Gordon.

Xiphophorus montezumae montezumae
Jordan & Snyder, 1900
Rio Salto, San Luis Potosi (1957)— Rosen,
Malcolm Gordon, Gordon.

Xiphophorus montezumae cortezi Rosen, 1960
Rio Axtla, San Luis Potosi (1939) — Gordon,
Atz, Evelyn Gordon.

Xiphophorus pygmaeus pygmaeus
Hubbs & Gordon, 1943
Rio Axtla, San Luis Potosi (1940) — New
York Aquarium Expedition to La Cueva Chica,
C. M. Breder, Jr., leader.

Xiphophorus maculatus (Guenther, 1866)
8A— Domesticated, white, spotted strain (Au-
gust 7, 1939)— Matsuno, New York, N. Y.

23 & 30 — Rio Jamapa, Veracruz ( 1939) —
Gordon, Atz, Evelyn Gordon.

Gp— Rio Grijalva, Tabasco (1952)— Gordon.
Bp— Belize River, British Honduras ( 1949) —
Gordon, Fairweather.

Xiphophorus hellerii strigatus Regan, 1907
3B — Arroyo Zacatispan, Oaxaca ( 1939) —
Gordon, Atz, Evelyn Gordon.

124— Domesticated strain.

Xiphophorus hellerii guentheri
Jordon & Evermann, 1896
Gx— Rio Grijalva, Tabasco (1952)— Gordon.
Bx— Belize River, British Honduras (1949) —
Gordon, Fairweather.

Xiphophorus hellerii
hx— Domesticated, albino strain.

A description of the laboratory in which these
viviparous, tropical, freshwater fishes have been
maintained and the care that is given them may
be found in Gordon (1950c).

A total of 3,000 hybrids was involved in the
present study. Most of the specimens were fixed
and stored in formalin, some in ethyl alcohol. In
addition to the hybrids, preserved examples of
the strains from which they had been derived
were examined, as well as numerous fish caught
and preserved in the wild. The latter are now
part of the collections of the Museum of Zool-

Table I. Earliest References to Hybridization of Fishes of the Genus Xiphophorus 1

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

155

a a

<3 Q
5 g

Co Co

ft. Cl

ft

i'R

a a

a a
« :§
£

St

3

S3

•S

-5

a

B

<L>

Ph

o

p<

a

ft

•S

-5

,W)

£

CO

CO

O

G

O

73

l-l

o

O

cj .2

c

o

73

u

O

O

c

o

73

«->

o

O

Os

^ c# w ^

G ? | C

0> u, T5 0>

■ « O o

o

VO

as

o

o

VO

Os

o

&

a .2

i=4 33

Tf IT)

os as

g a
o o

73 73

o o

oo

Os 2

c c
o o

73 73
u i_

o o
OO

vo .tS

B

oo
wc3

o' C ^
I ^ oo

•3 o2
fflOC

c.£Pc^

° % o oo
*2 5 ■g S
ooo2
o^oc

-

m C

52 o ^ g

Os 73 Os C/3

— w

g°.

73 73

*- a>

°a®

• S O co

<C m

Sg"0'

O <u ez O

Op^jO

'Op

&

co

co

O

G

O

73

O

O

JD

cd X
^ »n vo .tS
Jn ^ ^ G

S 2 2 *

c .SP ^ B X

O & E O 00

^ £ i2 ’S ^

o o 13 o 2

_ a JS

G

O

73

Ih

O

O

c

o

73

u

O

O

as

LZ

2 S

o

o

VO

OS

o

P4

a

5

2S

tv •-*
^ tv

■*>4 o>

S3 £

5 rs

C

o

73

u

O

O

a a

S 5

Co Co
?s

ft. ft.

X co
X ^

O Tf

o

T-H C

w o

+* 73

CO u

3 O

& o

o

VO

os

o

■8 8
g 8

3 D

c +-*

% e

■a 2
c '*'1

<D CO

<D 73
X *G

W X)

o >r*

c ^

o co
co

3 73 |h

O w %

• ~ co Cl,
J3 3 cd
£3 p

> co ^
C3 co

•SS2

■55 s r

« b o

> O VM

O ^ 73

g "

•5 43 -a

:- S 2

-fJ

sJ C 1)

7 !U >

v*. 73 cd

o •" -c

A+-<

co O /^s

.Si >,•

Cl, G q

3g (3
CO 73 _

o 5

o 2 2

^ 3 s

^52
<d X G

h t; g
cd .5
cux

2 No record or other evidence of this cross could be found in the Genetics Laboratory after the death
of Dr. Myron Gordon.

3 According to the historical review presented by Gordon (1931b).

156

Zoologica: New York Zoological Society

[47: 14

ogy of the University of Michigan, Ann Arbor,
Michigan (Rosen, 1960).

For practical reasons, we have had to confine
our attention to patterns visible on the surface
of the fish and also to melanic pigmentation, be-
cause the latter resists fading in ethyl alcohol,
formaldehyde or glycerine — in contrast to the
red patterns that occur in these fishes, for exam-
ple. Two special techniques were used to facili-
tate the study of the morphology of pigment
cells and patterns. Selected specimens were
dehydrated in absolute alcohol and cleared in
methyl salicylate (synthetic oil of wintergreen),
as described by Gordon (1931a). The skins of
others were dehydrated, cleared in xylene and
mounted on slides in Permount.

The photographs of living and preserved fish
were taken by Sam Dunton, Photographer of the
New York Zoological Society, while those of
cleared specimens and skins were made by Dr.
Ross F. Nigrelli, Pathologist of the New York
Aquarium.

III. Results

From both a genetic and morphological point
of view, the pigmentary patterns of the fishes of
the genus Xiphophorus may be divided into (1)
those shared by all adult members of the form,
or sometimes all adult members of one sex of the
form, and (2) those found in some individ-
uals but not in others. The latter comprise the
polymorphic elements of the pigmentary system.
They are made up of either micromelanophores
or macromelanophores and are genetically con-
trolled in most instances by single major genes.
In contrast, the monomorphic or non-polymor-
phic patterns are composed of micromelano-
phores and scale melanophores and, as far as
known, are controlled only by multiple genetic
factors. Both types of pattern show variations
that may be correlated to a greater or lesser
extent with environment, age and sex.

The pigmentation of Xiphophorus maculatus
has been described in detail by Gordon (1931a),
and much of what he recorded also holds for
the other members of the genus. Rosen (1960)
has described the outstanding pigmentary fea-
tures of all the species and subspecies, and a
more detailed account of the forms under cur-
rent discussion may be found in Atz (1959a).
In X. maculatus, the patterns of macromelano-
phores are governed by major genes that are
dominant and sex-linked, while the major genes
that influence the polymorphic micromelano-
phore patterns are dominant and autosomal. Al-
though the distinction between the two types of
melanophores is definite, genetically speaking,

no diagnostic morphological features have ever
been described. Macromelanophores are consid-
erably larger than micromelanophores and may
attain a diameter of 500 microns (Gordon,
1959), but the two types overlap in size. In prac-
tice, however, it is usually easy to distinguish
between them because of the greater size and
denser pigmentation of the macromelanophores.

1. Inheritance of Monomorphic Pigmentary
Patterns in Hybrids.

That these pigment patterns are monomorphic
precludes the use of most intraspecific crosses to
analyze their genetic basis. Only mutants such as
i for albinism or st for xanthism are available
for genetic analysis. Interspecific or intraspecific
hybridization provides, however, another means
of revealing the hereditary foundation of spe-
cies- or subspecies-specific characteristics. The
individual genetic factors can seldom be identi-
fied, but the behavior of phenotypic characters or
character-complexes may be studied, and from
this conclusions may be drawn as to the nature of
the genetic elements at work.

The many crosses between species and sub-
species of Xiphophorus that were available made
this type of analysis feasible. A series of pairs
of opposing categories was set up (Tables II,
III and IV) that represented pigmentary char-
acters in which the parental species differed suf-
ficiently to make the assignment of a hybrid to
one or the other category, or to an “intermedi-
ate” one, not too arbitrary a procedure. Most
of these fishes are partially covered by a pig-
mentary pattern that consists of parallel lines
enclosing rhombic or hexagonal areas of lighter
pigmentation. Superficially, this pattern appears
to outline each scale, but its anatomical basis
is formed by the scale pockets, each edge of
which is bordered by a band of micromelano-
phores. This network has been appropriately
called the reticulum by Rosen (1960), and it
provides a natural basis for describing the pat-
terns on the body since most of them, e.g. the
mid-lateral stripe, can be considered as modifi-
cations of the reticulation (see Figs. 13, 18).
Vertical barring, however, is clearly separable
from the reticulum, lying deeper in the skin, as
was pointed out by Gordon (1931a).5 Back-
ground pigmentation and inter-radial pigmen-
tation of the caudal fin are distinctive in X. p.
pygmaeus, in which scale melanophores and skin

5 Gordon considered this pattern to consist of parr
marks, but any resemblance or relationship it may have
to the well-known salmonid pattern is problematical
For one thing, unlike the latter, it is frequently retained
as the fish matures. See Fig. 18.

1962] Atz: Effects of Hybridization on Pigmentation in Xiphophorus 157

Table II. Summary of “Dominance” Relationships among Non-polymorphic
Pigmentary Patterns in Fi Hybrid Xiphophorus

Number of Broods:1

Total No.
of Broods

Inter-

mediate

Resembling

Female

Parent

Resembling

Male

Parent

Reticulum

Type

22

9

5

9

Extent

18

13

4

1

Mid-lateral

Stripe

24

12

12

5

Background

Pigmentation

4

0

1

3

Vertical Barring

22

18

5

3

Dorsal Fin Pattern

15

10

2

4

Caudal Fin Patterns
Inter-radial

Pigmentation

4

4

0

0

Caudal Edging

3

2

1

0

Ventral Edging

11

6

6

3

Dorsal Edging

of Sword

3

1

0

3

Anal Fin, Caudal

Edging

12

6

3

5

Mid-ventral Stripe

14

10

2

4

Deep-lying Spots

10

0

8

2

Totals

162

91

49

42

1 Excess (20) above Total Number of Broods results from some broods having intermediate individuals
in addition to those resembling either male or female parent, and therefore being recorded in two columns.

melanophores (neither of which is associated
with the reticulum and thus may be considered
to belong to the background) are by far the
most poorly developed, and in which pigmenta-
tion is lacking between the finrays of the tail.
A thin, dark border along the caudal edge of the
caudal fin is found in some male X. p. pygmaeus,
and a somewhat similar pattern along the caudal
edge of the anal fin in female X. maculatus. In
the swordtails, there is a dark band of pigmen-
tation that runs along the caudal peduncle as
the mid-ventral stripe which may or may not
extend onto the ventral edge of the caudal fin.
The dorsal edge of the sword of male swordtails
may also be edged in black, and the way in
which this is accomplished differs in different
species. Deep-lying spots are a unique pattern of
X. couchianus and consist of groups of deep-
lying pigment cells, apparently associated closely
with blood vessels, some of which are neverthe-
less visible along the posterior flanks of the fish.

The following observations have been made
from a study of the data from Atz (1959a),
summarized in Tables II, III and IV:

(1) The Fis are not always uniform, either
within broods or among similar crosses—
even when judged by the relatively coarse
standards employed.

(2) Among the Fis, about as many characters
resemble those of either parental form as
are intermediate.

(3) Although characters in certain Fis may
resemble the female parent, there is, in
the aggregate, no sign of maternal influ-
ence, since Fi characters as frequently re-
semble the male parent.

(4) The F2S are more variable than the Fis,
but not so in the expression of all, or even
a majority of the characters.

(5) Backcrossing tends to result in characters
that resemble the parental species with
which the hybrid was backcrossed, but
this is not invariably the case.

(6) Nevertheless, the second backcross never
fails to produce at least some individuals
that resemble the backcross species in the
character in question, and usually the ma-

158 Zoologica: New York Zoological Society [47: 14

Table III. Comparison of Variability in Non-polymorphic Pigmentary
Patterns between F2 and Fi Hybrid Xiphophorus

Number of F2 Broods Exhibiting:

No

Increase

Toward

Increase

Toward

Increase

One

Both

in

Parental

Parental

Variability

Form

Forms

Reticulation

Type

2

6

0

Extent

2

5

1

Mid-lateral Stripe

1

8

0

Background Pigmentation

1

1

0

Vertical Barring

3

5

0

Dorsal Fin Pattern

1

2

1

Caudal Fin Patterns

Inter-radial Pigmentation

0

1

1

Caudal Edging

0

0

1

Ventral Edging

2

0

0

Dorsal Edging of Sword

2

0

0

Anal Fin, Caudal Edging

0

5

0

Mid-ventral Stripe

2

1

1

Deep-lying Spots

5

2

0

Totals

21

36

5

jority of the fish are practically indistin-
guishable from that species.

(7) A few characters, most notably ventral
edging of the caudal fin, persist in some
individuals after two backcrosses to the
opposite parental species.

(8) Reciprocal crosses do not always produce
similar offspring ( X . v. xiphidium and X.
maculatus; X. m. cortezi and X. macula-
tus; X. h. guentheri and X. maculatus ) .

(9) It seems impossible to predict, on any
morphological basis, whether the expres-
sion of a given character will be “domin-
ant,” “recessive,” or intermediate in the
Fi.

( 10) A character is occasionally “dominant” in
one interspecific combination and “reces-
sive” in another.

(11) Occasionally the expression of a character
in the Fi is noticeably different from that
in either parent, sometimes resembling
that of another species (Vertical bars
like X. v. xiphidium in X. couchianus X
X. maculatus and in X. couchianus X X.
v. variatus; mid-lateral stripe like X. v.
variatus in X. h. strigatus X X. couch-
ianus) .

2. Inheritance of Polymorphic Pigmentary

Patterns in Hybrids.

The polymorphic pigmentary patterns of
Xiphophorus maculatus have been extensively
studied by Gordon and his collaborators, who
have described them (Gordon, 1931a, 1948,
1951b; Gordon & Fraser, 1931), determined
their mode of inheritance (Gordon, 1931b,
1937, 1947a, 1948, 1950a, 1956b), recorded
and analyzed their frequencies in nature (Gor-
don, 1947a; Gordon & Gordon, 1957), and
studied their development and physiology both
in health and disease (Gordon, 1948, 1950a,
1951a,b, 1957, 1958a, 1959; Gordon & Smith,
1938; Nigrelli, Jakowska & Gordon, 1951).
Similar, but not as extensive, studies have been
made with five other species belonging to the
genus Xiphophorus. Four of these are polymor-
phic, but only one, X. variatus, approaches the
remarkable diversity of pigment patterns shown
by X. maculatus.

a. Micromelanophore Polymorphic Patterns.

Many specimens of X. maculatus, X. v. varia-
tus and X. v. xiphidium exhibit a distinctive
arrangement of micromelanophores located on
either side of the caudal peduncle at the base of
the caudal fin and extending onto it in varying
degrees. There are seven basic patterns in X.
maculatus and four in X. variatus, some of which
closely resemble each other. No fish ever carries
more than two patterns, and genetic experiments

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

159

Table IV. Effects of Backcrossing on Non-polymorphic Pigmentary
Patterns of Hybrid Xiphophorus

First Backcross

Second Backcross

Third Backcross

Total Number Broods

Broods Showing Change
Toward B. C. Parent

Broods Resembling B. C.
Parent1

Total Number Broods

Broods Showing Change
Toward B. C. Parent

Broods Resembling B. C.
Parent1

Total Number Broods

Broods Resembling B. C.
Parent

Reticulation

Type

12

8

10

10

1

10

2

2

Extent

10

9

7(3)

6

1

6

Mid-lateral Stripe

12

10

9(2)

10

3

8(2) 2

2

Background

Pigmentation

1

1

Vertical Barring

7

4

3(2)

4

2

2(2) 2

2

Dorsal Fin Pattern

7

6

5(1)

7

1

7

2

2

Caudal Fin Patterns

Inter-radial Pigmentation

1

1

1

Caudal Edging

1

1

1

Ventral Edging

6

6

2(4)

7

2

4(3) 2

2

Dorsal Edging of Sword

1

4

1

4

2

2

Anal Fin, Caudal Edging

7

5

K5)

5

1

3(2)

Mid-ventral Stripe

7

6

(5)

9

1

4(4) 2

2

Deep-lying Spots2

3

3

(3)

1 Numbers in parentheses represent broods in which not all individuals resemble the backcross parent,
and these are not included in the other figures in the same column. The number of broods in this column
often exceeds the number of broods that show a change toward the backcross parent, because some Fi
hybrids already resemble the backcross parent in certain characters and many first backcross fish do.

2 In one second backcross to X. couchianus (X. maculatus X X. couchianus), no deep-lying spots
appeared, even though some of the first backcross fish had exhibited them.

have shown that each pattern is controlled by a
member of a series of dominant, autosomal mul-
tiple alleles. Hybridization strongly indicates
that the two series of alleles occupy the same
locus in both species (Atz, 1959a). Two domi-
nant, autosomal gene modifers that alter the ap-
pearance of the tail patterns in X. maculatus
have been identified. One called extensor (E),
changes the comet (Co) pattern, in which micro-
melanophores form a thin, dark border along the
dorsal and ventral edges of the caudal fin, into
the wagtail complex, in which all the fins and
certain other extremities are pigmented (Gor-
don, 1946). The other modifier (Cg) changes
the twin-spot pattern (T) into a reversed,
C-shaped pattern called the Guatemala crescent
(Gordon, 1956) . Gordon found that these modi-
fier genes are present in at least some X. hellerii,
and he indicated that this species was perhaps

their only source, the factors having entered the
genome of various domesticated strains of X.
maculatus by introgressive hybridization.

Evidence for the presence of an extensor- like
gene in X. v. xiphidium was obtained in a brief
series of crosses (Table V). A female X. macu-
latus, carrying the comet (Co) tail pattern, was
mated to a male X. v. xiphidium, both parents
being descended from wild-caught fish, and 51
of the 99 offspring showed comet. Of these, 14
had this pattern modified in the direction of the
wagtail with an intensification and slight spread
of the pattern itself, a darkening of part of each
of the dorsal finrays and the appearance of heavy
pigmentation on the upper and lower lips, the
latter being a typical part of the wagtail complex
(Fig. 1). When one of the hybrid females with
this modified pattern was backcrossed to X. v.
xiphidium, 30 of the offspring showed comet and

160

Zoologica: New York Zoological Society

[47: 14

Table V. Influence of Hybridization on Micromelanophore Tail
Patterns of Fishes of the Genus Xiphophorus

Factor

Parent

with

Pattern

Parent

without

Pattern

Modification
of Pattern
in Offspring

Cross1

Co

maculatus $

xiphidium $

Extension (slight
wagtail effect)
(see Fig. 1)

h20

Co

h20 9

xiphidium 8

Enhanced wagtail
effect (See Fig. 2)

h30

Co

h20 $

xiphidium X variatus $

Extension (slight
wagtail effect)

h38

T

maculatus $

guentheri $

Extension ( Guatemala
crescent)

324

T

maculatus 8

cortezi 9

Extension (pseudo-
crescent)

103B

O

maculatus $

cortezi 8

Enhancement

80

Ct

xiphidium $

variatus $

Extension (pseudo-
crescent) (see Fig. 7)

h4

1 Number of cross, as designated in the records of the Genetics Laboratory of the New York Aquarium.

the 7 largest of these (at the time they were
sacrificed) showed a more strongly expressed
wagtail, with the caudal and dorsal finrays dark-
ened for most of their length, the pectoral and
anal finrays darkened to a lesser extent, and the
lips heavily pigmented (Fig. 2). In contrast,
when a hybrid male, showing a slightly devel-
oped wagtail pattern, was mated to a female
intraspecific hybrid X. v. xiphidium X X. v.
variatus, 57 of the offspring showed comet of
which the 8 largest (at the time they were sac-
rificed) exhibited a wagtail pattern, better ex-
pressed than it had been among the members
of the original hybrid cross, but not as strongly
as it was among the backcross offspring just
described. Gordon (1946) indicated that the
wagtail pattern (CoE) was not apparent in
young fish, but developed as they grew. This,
however, cannot account for the absence of the
wagtail effect in many of the Co fish, since some
of these were adults when sacrificed and pre-
served. That the effect was enhanced by back-
crossing to X. v. xiphidium indicates that more
than one modifying gene was involved.

In a cross between a twin-spot X. maculatus
and a swordtail from British Honduras, X. hel-
lerii guentheri, all the hybrids that inherited T
exhibited the Guatemala crescent. This pattern
was compared, partly by means of cleared speci-
mens, with some of the Guatemala crescent fish
studied by Gordon (1956b). As far as the tail
pattern was concerned, the fish appeared identi-
cal. In addition to the tail pattern, however,

there is a modification of pigmentation near the
mouth of Guatemala crescent fish (Gordon,
1956b). Most prominent are two spots of heavy
pigmentation at each mandibular junction. These
were lacking in the present specimens; instead
they showed a crescent of dense pigmentation
immediately behind the upper lip, overlying the
ethmoid region posteriorly and the heads of the
premaxillae and maxillae anteriorly. The pig-
ment was located in the dermis. An examination
of Guatemala crescent fish in the collection of
the Genetics Laboratory (which could not in-
clude all the specimens or crosses studied by
Gordon) revealed that in only one other cross
(250) did the TCg fish exhibit the crescent-
shaped pigmented area behind the upper lip. The
ancestry of the latter cross involved X. macula-
tus from both the Rio Jamapa and domestic
sources and X. hellerii strigatus from the Rio
Papaloapon. It thus revealed no obvious genetic
relationship to the present cross.

Another modification of the twin-spot pattern
occurred in hybrids of X. m. cortezi and X.
maculatus, in which a shadowy, crescentic pat-
tern was formed by micromelanophores more or
less connecting the upper and lower spots. In
the hybrids from another cross involving the
same species, the one-spot tail pattern was larger
than it ever appears in the parental species, X.
maculatus.

The offspring of an intraspecific cross, involv-
ing a female X. v. xiphidium, and a male X. v.
variatus, exhibited a complete range of tail pat-

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

161

terns from cut-crescent to crescent. (Fig. 7), al-
though the cut-crescent of the female parent was
entirely normal in appearance, and no crescent-
bearing fish is known among the ancestors of the
fish.

None of the offspring from the above crosses
was bred, and so no information exists on the
genetic behavior of their modified tail patterns;
nor is there any other cross involving a fish car-
rying genes for the same tail patterns, which
might indicate how widespread the supposed
genetic modifiers may be.

b. Macromelanophore Polymorphic Patterns.

The patterns composed of macromelano-
phores are all polymorphic, that is, none of
them occurs in every individual of the species.
They can most simply be described as spotted,
and the spots may be considered to vary from
the size of single macromelanophores to broad
bands of black pigment. Because of the rela-
tively large size of macromelanophores and the
concentration of melanin granules within them,
these patterns appear darker and more sharply
demarcated from their background than do the
patterns composed of micromelanophores. Var-
iation in location, size and number of spots is
therefore readily apparent, and this circum-
stance may contribute significantly to the im-
pression that the variability of the macromelano-
phore patterns is considerably greater than the
variability of the micromelanophore ones. There
is, however, no question about the greater range
of expressivity of the major genes controlling
macromelanophores in hybrid genomes, in which
this may be increased to pathological melanosis
on the one hand or reduced to no penetrance at
all on the other (Gordon, 1951a).

Five basic macromelanophore patterns found
in X. maculatus have been described by Gordon
(1951b, pp. 175-179) and Gordon & Gordon
(1957, pp. 3-6):

Spotted (Sp) — irregular spotting on sides.

Nigra (N)— irregular blotches or bands on
sides.

Striped (Sr)— discrete rows of spots on sides.

Spotted dorsal (Sd) — irregular spotting on
dorsal fin.

Spotted belly (Sb)— heavy spotting on belly
and ventral and anal fins.

With rare exceptions, no more than two of
these patterns occur in a fish, and there is good
evidence that they are controlled by dominant,
sex-linked alleles (Gordon, 1948). Crossing-
over has occurred, however, so that two macro-
melanophore genes have become located on a
single chromosome (MacIntyre, 1961). The

series might therefore better be designated as
pseudoallelic.6 7

Gordon (1943) recognized a single macro-
melanophore pattern in X. variatus and in X.
xiphidium (at that time considered to be sepa-
rate species rather than subspecies as they are
today), viz., spotted (Sp) but subsequent, un-
published, analysis revealed the presence of a
second pattern, spotted caudal. Rosen (1960, p.
81), however, lists four macromelanophore pat-
terns from X. variatus, viz., spotted (which he
calls blotched), spotted caudal, speckled and
black-banded. The latter three occur only in fish
from the Rio Cazones, from which no living
specimens have yet been collected for genetic
studies. The spotted (Sp) pattern is inherited as
a sex-linked dominant and, according to Atz
( 1959a), appears to be an allele of the Sp of X.
maculatus ?

The spotted pattern in X. variatus xiphidium
typically consists of numerous, somewhat diffuse
spots located in the region of the mid-lateral line
and immediately above it.8 In size they approach,
but never equal, the area of the exposed portion
of a scale, but the great majority are considerably
smaller, that is, less than half as large. As few as
eight spots have been found on one side of a fish,
but the number is usually very much more. A
common variant of this pattern (in certain pop-
ulations) is one in which the spots are so numer-
ous that they form, at a distance, a band of pig-
mentation along the side of the fish. Closer
examination reveals that this is composed of
numerous closely grouped macromelanophores,
many of them “touching” one another. Aside
from a tendency to follow the reticular pattern
(see below) and be concentrated near the mid-
lateral line, however, no pattern can be dis-
cerned. Sometimes the distribution of macro-
melanophores is so generalized that the whole
body of the fish above the region of the mid-
lateral line appears flecked with pigment. In such
fish, individual spots are hard to distinguish.

The majority of the larger spots in X. v. xiphi-
dium are roundish, but the smaller ones take on
less regular, more elongate forms. They usually
appear more diffuse and less clearly defined than
do the spots found on X. montezumae cortezi,

6 It was evidence and considerations of this nature that
led Rosen (1960, pp. 76-77) to state that the macro-
melanophore genes “are not all members of a single
allelic series.”

7 See Figs. 4-6 which show the Pi and some of the Fis
of one of the crosses that revealed this relationship.

8 See the Pi male in Fig. 10 as an example of the ap-
pearance of Sp in this subspecies.

162

Zoologica: New York Zoological Society

[47: 14

X. hellerii and, in some instances, X. maculatus.
This may be the result of either or both of two
factors: (1) as far as can be determined, the
number of pigment cells per unit area of a spot
is definitely less in X. v. xiphidium than in either
X. hellerii or X. montezumae, and (2) there may
be a lower concentration of pigment per cell in
X. v. xiphidium. This could account for the dif-
ferences observed between it and some strains of
X. maculatus, where the number of cells appears
to be substantially the same. A third possibility,
of course, is that the macromelanophores are of
different sizes, but those from the two swordtails
appeared smaller, if anything, than those in X. v.
xiphidium. This possibility could only be settled
by treating live specimens with adrenaline to
concentrate as uniformly as possible the pigment
in the melanophores, fixing the fish in that state,
and then making counts and measurements.

Although the spotted pattern of X. v. xiphi-
dium is characterized by its variability and lack
of definition, one feature always marks its dis-
tribution. This is its close spatial relationship to
the reticulum; it is very rare that a spot or in-
dividual macromelanophore is found that does
not seem to be touching or lying astride the retic-
ulum — or is not occupying a place that would
have shown reticulum, had it been present. In
those cases, described above, where macromel-
anophores are so numerous that they practically
form a horizontal band near the mid-lateral line,
numerous cells must fall in areas between retic-
ular elements, but this occurs only where the
cells are so numerous that they “touch” or “over-
lap” one another. Even in these cases, the cells at
the edges of the pigmented areas follow the retic-
ulum.

The spotted (Sp) pattern of X. variatus varia-
tus resembles that seen in X. v. xiphidium but
differs in the following ways.9 Typically the pat-
tern is confined to the region of the mid-lateral
line, but the spots are not concentrated in the lat-
ter region as frequently as they are in X. v. xiphi-
dium. There are more large spots and no fish
without large spots, except those from certain
regions. There are numerous roundish spots, but
because the spots sometimes follow the reticu-
lum very closely, V- and Y-shaped ones are not
rare. In some specimens from one region (Rio
Tempoal), the spots so closely follow the retic-
ulum that a pattern strongly reminiscent of the
striped (Sr) pattern of X. maculatus is pro-
duced. Although the number of pigment cells per

9 See the spotted hybrids in Figs. 8 and 9 for unmodi-
fied examples of this pattern. Fig. 18 shows an enlarge-
ment of some macromelanophore spots from this sub-
species.

unit area within a spot appears comparable to
that in X. v. xiphidium, the intensity and sharp-
ness of the spots is in general greater than in the
latter subspecies. The range of variability of the
spotted pattern is, however, greater in X. v.
variatus.

Two macromelanophore patterns, spotted
(Sp) and spotted caudal (Sc), are known in X.
montezumae cortezi and each appears to be con-
trolled by a single dominant, autosomal gene
(Atz, 1959a). These are not alleles nor is Sc an
allele of the Sp factor of X. hellerii guentheri,
but no direct evidence exists concerning their
relationship to the macromelanophore factors in
other species. In X. m. cortezi, the spotted (Sp)
pattern typically consists of prominent, deeply
pigmented, roundish spots mostly confined to
the mid- and post-dorsal regions, above the mid-
lateral stripe, and to the dorsal and caudal fins.
In size, individual spots approach, but rarely if
ever exceed, the area of the exposed portion of a
scale. Although there may be as few as five spots
on one side of an adult-sized fish, there are usu-
ally many more. Analysis indicates that the num-
ber of spots increases with size, and presumably
age, and that males exhibit more spots than fe-
males (Atz, 1959a). On the average, the closer
one approaches the mid-dorsal line, the greater
the density of spots (Fig. 13). The macromel-
anophores that comprise this pattern appear to
be associated in some way with the reticulum,
which is especially well developed in this species
(Fig. 13). No spot was ever found that was not
"touching” some part of the reticular pattern,
that is, no spot was located entirely within the
hexagonal or trapezoidal areas formed by the
reticular elements. Usually the spots appeared
to be directly astride the reticular bars and some-
times a halo effect was noticeable. In connection
with the latter phenomenon, it should be noted
that macromelanophores are located at or very
near the same level in the dermis as are the mi-
cromelanophores. Another characteristic of the
spots of X. m. cortezi is that they not infrequent-
ly coalesce, forming irregularly shaped blotches.
On the dorsal fin, the spots are found in the in-
terradial membranes, although they may extend
over the finrays. In heavily spotted males, they
tend to form two rows, one near the base of the
fin, the other about halfway between base and
distal edge. The spots on the dorsal fin are usu-
ally larger than those on the body. In females or
immature fish, the spotted pattern rarely extends
onto the dorsal fin. In some spotted males and a
few females, there may be spots on the caudal
fin as well. These are spindle-shaped or oval
and generally small. They lie between or along
the caudal finrays and have been seen to occur

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

163

between any two finrays except those involved
in the sword. They seldom occur under the su-
perficial caudal musculature at the base of the
fin, but rather in the middle half of the exposed
portion of the fin itself.

The tendency for macromelanophores to con-
gregate in the caudal fin of X. m. cortezi is most
strikingly seen in the spotted caudal (Sc) pat-
tern. This consists of one or more irregular,
elongate patches of heavy pigmentation, com-
mencing close to the base of the caudal finrays
and extending toward the rear for roughly one-
third of the fin’s length and ending in a variable
number of irregular, tapering extensions of pig-
mentation (Fig. 15). These extensions usually
are apposed to a finray; in fact, the whole pat-
tern appears to develop in close relation to the
caudal finrays. The macromelanophores invade
the perimysia of the superficial muscle of the
caudal fin as well as enveloping the finrays. In
adult fish, the pattern may be only a sliver of
pigmentation or it may be a blotch covering the
second through the twelfth caudal finrays. Such
large blotches are rare, however. Although the
most frequent number of pigmentary patches
comprising this pattern is one, two are not un-
common, and as many as four may be seen.

Macromelanophore spotting is known to oc-
cur in a small proportion of the individuals of
two subspecies of X. hellerii (Rosen, 1960, pp.
120 and 126). In the form presently available,
X. h. guentheri, genetic data indicate that the
spotted (Sp) pattern is inherited as an autosomal
dominant. Aberrant ratios have been noted,
however, which might be explained by incom-
plete penetrance of Sp. It is noteworthy that
true-breeding spotted X. h. guentheri were estab-
lished in our Genetics Laboratory only after in-
dividuals exhibiting spots had been selected reg-
ularly to carry on the line for more than seven
generations (Kallman, personal communica-
tion). Crosses with X. maculatus bearing Sr or
Sd showed that these two factors are not alleles
of the Sp from X. h. guentheri (Gordon, 1958b;
Atz, 1959a). This pattern is notable for its rela-
tively small number of large, intensely dark, ir-
regularly shaped spots (Fig. 19). The irregu-
larity is partly, but not wholly, the result of the
coalescence of adjacent spots as a result of in-
crease in size. Because of the discreteness of the
spots, it is easy to study their morphology indi-
vidually, and a series of stages beginning with a
single macromelanophore and extending
through intermediate stages (of, say, 30 macro-
melanophores) to large spots composed of un-
countable numbers of pigment cells (in the order
of hundreds but probably less than one thousand
cells) could be distinguished. The most simple

explanation is that the spots increase in size
through the appearance of more pigment cells.
There is good evidence that the number of spots
increases with the size of fish and therefore pre-
sumably with age (Atz, 1959a) . The usual maxi-
mum size of a single spot, i.e., one not involved
in any coalescence with other spots, approaches
the area of the exposed portion of a scale. Al-
though the poor development of the reticulum
makes the determination of relationship diffi-
cult, in all instances where such a determination
could be made, the spots were seen to be closely
associated with reticular elements. Frequently a
halo effect is in evidence. The location of the
spots on the body follows no discernible pattern
save that they are much more frequently found
on or above the mid-lateral stripe and somewhat
more often in the pre- and mid-dorsal regions.

The effects of hybridization on macromelano-
phore patterns in 108 crosses are outlined in
Table VI. The changes in the phenotypic expres-
sion of eight major genes belonging to five dif-
ferent species and subspecies vary from no dis-
cernible effect to loss of penetrance, on the one
hand, and severe melanosis with the production
of melanotic overgrowths, on the other. These
results may be summarized by stating that they
confirm and extend the conclusions reached by
Gordon (1951b). The following are the obser-
vations resulting from an analysis of Table VI:

(1) Genes vary in their ability to respond to
genetic influence from foreign genotypes.
The gene for spotting (Sp) in X. maculatus
is unquestionably the most potent in this
respect; in no other species has it failed to
produce well developed melanosis and, at
least occasionally, neoplastic overgrowths
in the form of melanomas (see items nos.
1-21 of Table VI). At the other extreme
stand the genes for spotting in X. hellerii
guentheri and X. montezumae cortezi
whose expressivity changes to only a limited
extent (nos. 82-91, 117-143) except in one
hybrid combination (nos. 92-95).

(2) Species vary in their tendency to influence
foreign genes controlling pigmentation. In
X. maculatus, the species whose pigmentary
genes are most capable of showing excessive
growth in foreign genotypes, no enhance-
ment of pigmentation belonging to patterns
from other species has ever occurred (nos.
66-68, 78-80, 84, 98, 117-140). If any
species could be assigned the role of being
most likely to show melanosis and mel-
anoma as a result of the introduction of
foreign pigmentary genes into its genotype,
X. hellerii would be the choice, based on
its reactions to Sp, Sb, Sd and N (Gordon,

164

Zoologica: New York Zoological Society

[47: 14

Table VI. Influence of Hybridization on Macromelanophore Patterns
of Fishes of the Genus Xiphophorus

No.

Parent with
Pattern

Parent without
Pattern

Offspring with
Pattern1

Cross2

Sp — X. maculatus

1

maculatus $

couchianus $

Fi: Melanosis

h7

2

h7

h7

F2: Severe melanosis to enhancement

h72

3

maculatus $

couchianus $

Fi: Melanosis

hl5

4

hl5 9

couchianus $

BC: Melanosis (5) overgrowths (1)

91

5

hl5 53^

couchianus $$$

BC: Melanosis (4), severe

melanosis (7), overgrowths (2)

92, 94,
99

6

99 3

couchianus $

2nd BC: Melanosis (6 out of 8)

100

7

91 9

couchianus $

2nd BC: Melanosis

101

8

maculatus $

couchianus $

Fi: Melanosis

845

9

845$

845 3

Fo: Severe melanosis (2 out of 13)
to moderate enhancement (6 out
of 13 ) ; overgrowths ( 1 out of 13 )

878

10

845 3

845 $

F2: Severe mleanosis (2 out of 11)
to moderate enhancement
(2 out of 11)

880

11

845$

couchianus $

BC: Severe melanosis

881

12

881 $

couchianus $

2nd BC: Severe melanosis

945

13

881 $

couchianus $

2nd BC: Severe melanosis

946

14

maculatus $

couchianus $

Fi: Severe melanosis

851

15

851 3

851 $

F2: Severe melanosis (8)
to melanosis (2)

893

16

8513

couchianus $

BC: Severe melanosis (2);
overgrowth (1)

934

17

maculatus X
variants $

variants $

BC: Melanosis (3 out of 33);
enhancement (see Figs. 4-6)

h61

18

maculatus 3

cortezi $

Fj: Strong enhancement (9);
melanosis (4); overgrowths (7)

103

19

maculatus $

cortezi $

Fi: Strong enhancement

103B

20

103

cortezi

BC: Melanosis (5); overgrowths (5)

103BC

21

h7 $

h3 3 Enhancement

Sd — X. maculatus

h29

22

maculatus $

couchianus $

Fi: Reduced penetrance

h7

23

maculatus $

couchianus $

Fi: Mild melanosis

325

24

maculatus $

couchianus $

Fj: No penetrance (14)

845

25

maculatus $

couchianus $

Fi: Melanosis

895

26

maculatus $

xiphidium $

Fi:Reduced penetrance (see Fig. 1)

h20

27

maculatus $

xiphidium $

Ft: No penetrance (see Fig. 3)

h31

28

maculatus $

320 3

BC: Mild melanosis; overgrowth (1)

350

29

maculatus X
strigatus $

guentheri $

BC: Melanosis; overgrowths
(2 out of 8)3

479

30

479 3

guentheri $

2nd BC: Severe melanosis (2);
melanosis (6) ; enhancement (1);
melanomas (4)3

671

31

479$

strigatus $

2nd BC: Severe melanosis (1);
melanosis (3); melanomas (2) 3

672

32

h7 5

h2 $ No penetrance

Sr — X. maculatus

h28

33

maculatus $

couchianus $

Fj: No effect

68

34

68

68

F2. Reduced penetrance and
expressivity

682

1962] Atz: Effects of Hybridization on Pigmentation in Xiphophorus 165

Table VI. Influence of Hybridization on Macromelanophore Patterns
of Fishes of the Genus Xiphophorus ( Continued )

Parent with

Parent without

Offspring with

No.

Pattern

Pattern

Pattern1

Cross2

35

maculatus $

couchianus 8

Fi: Enhancement

895

36

895 X 895

Fo: Mild melanosis (1), enhancement

8952

(6), reduced expressivity (3),

strongly reduced expressivity (4)

37

maculatus $

xiphidium 2

Fi: Slight enhancement (see Fig. 3)

h31

38

maculatus $

h2 2

Reduced expressivity

h22

39

maculatus 2

cortezi $

Fi: Slight suppression

h80

40

maculatus $

320 2

BC: Enhancement to nearly complete

384

suppression (see Fig. 21)

41

maculatus 2

2nd BC: Enhancement to nearly

419

X 384 $

complete suppression

42

384 2

maculatus $

2nd BC: Reduced expressivity

486

43

486 2 X 486 $

No effect

523

44

486 2 X 486 $

No effect

524

45

486 2

486 $

No effect

525

46

maculatus $

320 2

BC: Slight enhancement

385

47

385 2

maculatus $

2nd BC: Reduced expressivity

421

48

421 $

guentheri 2

Enhancement

458

49

458 2 X458 $

Enhancement to nearly

521

complete suppression

50

maculatus 2

320 $

BC: Enhancement

386

51

maculatus 2 X

2nd BC: Reduced expressivity

420

386 $

52

420 2 X 420 $

No effect to nearly complete

522

suppression (see Fig. 22)

Sp — X. variatus variatus

53

variatus $

couchianus 2

Fi: No effect (see Fig. 8)

h23

54

variatus 2

couchianus $

Fi: No effect

h 1 3

55

h 1 3 $

h 1 3 2

F2: No effect

hl32

56

maculatus X

variatus 2

BC: No effect (see Figs. 4-6)

h61

variatus $

57

variatus $

cortezi 2

Fi: No effect (see Fig. 9)

h8

58

variatus 2

pygmaeus $

Fi: Enhancement (1 out of 27)

hi

(see Fig. 11)

59

variatus $

pygmaeus 2

Fi: No effect

hll

60

hll

hll

F2: No effect

hll2

61

variatus 2

xiphidium $

Fi: No effect

h2

62

h2

h2

Fo: No effect

h22

63

variatus $

xiphidium 2

Fi: No effect

h3

64

h3

h3

Fo: No effect

h32

65

variatus $

xiphidium 2

Fi: No effect (see Fig. 7)

h4

66

h2 2

maculatus $

No effect

h22

67

h4 2

h20 $

No effect

h38

68

h2 2

hi $

No effect

h28

Sp — X. variatus xiphidium

69

x iphidium $

cortezi 2

Fi: Strong enhancement

903

(see Fig. 10)

70

xiphidium $

cortezi 2

Fi: Strong enhancement

914

71

cortezi X ( cortezi

cortezi 2

2nd BC: Severe melanosis with

941

X xiphidium) $4

overgrowth (3)

166 Zoologica: New York Zoological Society [47: 14

Table VI. Influence of Hybridization on Macromelanophore Patterns
of Fishes of the Genus Xiphophorus ( Continued )

No.

Parent with

Parent without

Offspring with

Pattern

Pattern

Pattern1

Cross2

72

xiphidium $

pygmaeus 2

Fi:Mild melanosis (27 out of 28)
(see Fig. 12)

h66

73

h66 2

pygmaeus $

BC: Enhanced melanosis (3 out of 9);
reduced melanosis (2 out of 9)

106

74

xiphidium $

evelynae 2

Fi: No effect

913

75

xiphidium $

variatus 2

Fi No effect

h2

76

h2

h2

F2: No effect

h22

77

xiphidium 2

variatus $

Fi: Enhancement (see Fig. 7)

h4

78

xiphidium 2

maculatus $

Fi: No effect (see Fig. 3)

h31

79

xiphidium $

maculatus 2

Fi: Slight suppression to
no effect (see Fig. 1)

h20

80

h20 2 X
xiphidium $

Fi & BC: No effect (see Fig. 2)

h30

81

h2 2

h7 $ Slight suppression

Sp — X. montezumae cortezi

h28

82

cortezi 2

couchianus $

Fi: Reduced expressivity

849

83

cortezi 2

cortezi X ( cortezi
X xiphidium) $4

2nd BC: Slightly reduced expressivity

941

84

cortezi $

maculatus 2

Ft: No effect

h80

85

cortezi $

pygmaeus 2

Ft: No effect

hl2

86

hl2

hl2

F2: No effect

hl22

87

cortezi 2

strigatus $

Ft: No effect

hlO

88

cortezi 2

strigatus $

Ft: Reduced expressivity and
penetrance (see Fig. 13)

h9

89

h9£

strigatus 2

BC: Slight suppression (see Fig. 14)

h27

90

h27 2 X h27 $

No enhancement

h272

91

h272 2

strigatus $

Slight suppression

h39

92

cortezi $

montezumae 2

Ft: No effect

900a

93

cortezi $

montezumae 2

Ft: Enhancement (8 out of 11 $$,
1 out of 12 29)

900b

94

cortezi $

montezumae 2

Fj: No effect

900c

95

cortezi $

montezumae 2 Fi: Strong enhancement (3 out of

9 22), enhancement (1 out of 9 22)
(no Sp $$ present)

Sc — X. montezumae cortezi

900d

96

cortezi 2

couchianus $

Ft: No penetrance (3 adult,
16 immature)

849

97

cortezi 2

variatus $

Ft: No penetrance (see Fig. 9)

h8

98

cortezi $

maculatus 2

Fj: No penetrance

h80

99

cortezi 2

strigatus $

Ft: Enhancement (see Fig. 15)

h26

100

h26 2

cortezi $

BC: Enhancement (2);
no enhancement ( 1 )

h40

101

h26 2

strigatus $

BC: Melanosis (1); enhancement
(greater than in h40) (17)

h41

102

h41 2, h41 $

strigatus $, 2

2nd BC: Enhancement (3);
no enhancement (4)

h44, h45

103

h41 2

strigatus $

2nd BC: Severe melanosis (3);
overgrowth ( 1 ) ; enhancement
( 16) (see Fig. 16) 3

h42

104

h42

h42

Severe melanosis (2); overgrowths
(5 ) ; melanosis to enhancement (31)

h422

1962] Atz: Effects of Hybridization on Pigmentation in Xiphophorus 167

Table VI. Influence of Hybridization on Macromelanophore Patterns
of Fishes of the Genus Xiphophorus ( Continued )

No.

Parent with
Pattern

Parent without
Pattern

Offspring with
Pattern1

Cross2

105

h42$

hellerii (albino) $

3rd BC: Mild melanosis (3);

h47

106

h42 2, h42 $

strigatus $ 2

overgrowths (2); enhancement (6)
3rd BC: Severe melanosis (3);

h50

overgrowth ( 1 ) ; mild melanosis
(2); enhancement (10);
no effect (2)

107

h50 2 X h50 5

Severe melanosis (2); overgrowths
(5); enhancement (5);
no enhancement (5)

353

108

h41 2 X h41 5

Enhancement

h43

109

h43 2 X h43 $

Melanosis (5); overgrowths (2;)
enhancement (5)

h46

110

h46

h46

Severe melanosis ( 1 ) ;

melanosis (5); overgrowth (1);
enhancement (2)

h462

111

h462

h462

Severe melanosis ( 1 ) ; enhancement
(6); no enhancement (3)

h463

112

h422 2

hellerii ( wagtail ) 5

Overgrowth ( 1 out of 3 )

314

113

cortezi 5

montezumae 2

Fi: Enhancement (6 out of 9 55,
1 out of 4 22)

900a

114

cortezi $

montezumae 2

Fi: Enhancement (3 out of 11 55,
1 out of 8 22)

900b

115

cortezi 5

montezumae 2

Fi: Enhancement (3 out of 5 55,
2 out of 3 22), melanosis (2
out of 5 $$)

900c

116

cortezi $

montezumae 2 Fi: Strong enhancement (3 out of

11 22), enhancement (7 out of
1 1 22) (1 Sc 5 present)

Sp — X. hellerii guentheri

900d

117

strigatus X
guentheri $

maculatus 2

Fj : No effect

322

118

guentheri $

maculatus 2

Fj : No effect (1)

324

119

guentheri 2

maculatus $

Fi: No effect (see Fig. 20)

320

120

320

320

F2: No effect

3202

121

3202

3202

F3: Slight suppression

3203

122

3203

3203

F4: Slight enhancement

3204

123

320 5

maculatus 2

BC: Reduced penetrance and
expressivity

350

124

320 5

maculatus 2

BC: No effect

386

125

386 5

maculatus 2

2nd BC: No effect

420

126

420 2

420 5

Slight enhancement (3 out of 35)
(see Fig. 22)

522

127

320 2

maculatus $

BC: No effect (see Fig. 21 )

384

128

384 5

maculatus 2

2nd BC: Slight suppression

419

129

384 2

maculatus 5

2nd BC: No effect

486

130

486 2

486 5

No effect

523

131

486 5

486 2

No effect

524

132

486 2 X 486 5

No effect

525

133

320 2

maculatus $

BC: No effect

385

134

385 2

maculatus $

2nd BC: Reduced expressivity

421

135

421 2

guentheri $

No effect

485

136

421 5

guentheri 2

No effect

458

168 Zoologica: New York Zoological Society [47: 14

Table VI. Influence of Hybridization on Macromelanophore Patterns
of Fishes of the Genus Xiphophorus ( Continued )

No.

Parent with
Pattern

Parent without
Pattern

Offspring with
Pattern1

Cross2

137

458 $

458 $

No effect

521

138

guentheri $

maculatus X

BC: Slight enhancement

479

139

479 $

strigatus $
guentheri $

( 1 out of 8 ) 5
2nd BC: No effect 5

671

140

479$

strigatus $

2nd BC: Slight suppression 5

672

141

guentheri $

strigatus $

No effect

321

142

guentheri X (B.C.

strigatus $

No effect

751

143

and inbred,
ScCoE, hellerii-
cortezi hybrid $
751 5

strigatus $

BC: No effect

887

1 Described by means of a somewhat arbitrary series of classes, ranging from complete absence of
macromelanophore pattern to melanoma, all but the first two categories being concerned with expressivity:
No penetrance Enhancement

Reduced penetrance Strong enhancement

Nearly complete suppression Mild melanosis

Reduced expressivity Melanosis

Slight suppression Severe melanosis

No effect (z'.e., normal) Overgrowths (melanoma)

Slight enhancement

Numbers in parentheses indicate the number of fish

2 Number of cross as designated in the records of the Genetics Laboratory of the New York Aquarium.
When more than one macromelanophore pattern was involved, the cross has been listed under each
genetic factor.

3 Photographs of other melanotic members of h42 may be found in Gordon (1951b, p. 199) and as
Fig. 1 of Marcus & Gordon (Zoologica, 39: 123-131, 1954). The caption of the latter erroneously im-
plies that the fish belong to h50.

4 Fish received in 1958 from Dr. Curt Kosswig (see Kosswig (1959) for a general account of the
genetic background of this fish).

5 In designating backcrosses (BC), X. hellerii strigatus and X. hellerii guentheri have been equated.

1948, and nos. 28-31) and Sc (nos. 99,
101-112).

( 3 ) The same gene may vary both positively and
negatively under the influence of different
genomes. Sd may show reduced penetrance
or none at all (Gordon, 1951a and nos. 22,
24, 26, 27, 32) or melanosis with melanoma
(Gordon, 1951a and nos. 23, 25, 28-31).
Similarly Sr exhibits both reduced and en-
hanced expressivity (Gordon, 1948 and
compare nos. 33-34, 38, 39 with 35, 36,
37), while Sc varies in expression from no
penetrance to severe melanosis and mela-
noma (compare nos. 96-98 with 99-112).

(4) Crosses involving the same pigment pat-
terns and species sometimes give different
results (compare nos. 1, 3 and 8 with 14; 40
with 46 and 50; 41 with 42; 87 with 88;
92 and 94 with 93 and 95; 102 with 103;
123 with 124, 127 and 133). This probably
indicates that several genes and their alleles
are involved in the modification of the pig-

mentary patterns and that different com-
binations of these are present in different
individuals.10

(5) Backcrossing to the parental form from
which the macromelanophore pattern orig-
inated lessens the enhancing or reducing
effect of hybridization on its expression
(compare nos. 40 with 42, 46 with 47, 50
with 51, 99 with 100 and 101). Gordon &
Smith (1938) backcrossed a melanotic Fi
hybrid, X. maculatus X X. variants xiphi-
dium, to X. maculatus and obtained fish

10 The age at which the fish were sacrificed is a vari-
able that was controlled only in a general way, and this
undoubtedly is the reason for some of the varied results.
In the examples given above, however, the broods were
of roughly the same age or the differences were so strik-
ing that ontogenetic stage could not account for them.
At any rate, these crosses serve as a salutary warning
against dogmatic assertions unless they are backed up
by a replication involving several of the same crosses
made with fish from different genetic strains.

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

169

with no melanosis, and Kosswig (1948) re-
ported results of a similar nature.

(6) Backcrossing to the parental form that did
not carry the macromelanophore pattern
increases the enhancing or reducing effect
of hybridization on its expression (compare
nos. 4 and 5 with 3, 11 with 8, 73 with 72,
101 with 99). A plateau is soon reached,
however, and further backcrossing may
even result in a diminution of the effect
(compare nos. 6 with 5, 7 with 4, 12 and 13
with 11, 20 with 18, 102 with 101, 105 and
106 with 102 and 103).

(7) Inbreeding frequently increases the range
of phenotypic expression of macromelano-
phore patterns (compare 2 with 1, 9 and 10
with 8, 34 with 33, 36 with 35, 49 with 48).
Two series of successive inbreedings showed
variability in the average expression of the
pigmentary pattern, the mean or mode of
expression not necessarily being the same
for successive generations (nos. 108-111
and 120-122). Gordon & Smith (1938) also
obtained more variable Fo offspring in two
crosses of X. maculatus X X. couchianus
and one of X. maculatus X X. variatus
xiphidium.

(8) The same individuals may show enhance-
ment of one macromelanophore pattern
and no change or reduction in the expres-
sivity of another (compare nos. 28 and 123,
30 and 139, 31 and 140, 26 and 79, 40 and
127, 48 and 136, 49 and 137, 50 and 124,
57 and 97, 71 and 83, 92 and 113, 94 and
115). 11

IV. Discussion

1. Genetics of Micromelanophore Pigmentary
Patterns as Revealed by their Appearance in
Hybrids.

The behavior of the reticulum and other non-
polymorphic micromelanophore patterns in in-
terspecific crosses within the genus Xiphophorus
strongly indicates that their inheritance is con-
trolled by polygenes, that is, by relatively nu-
merous factors, each of which has a small
phenotypic effect. This is in accord with the con-
clusions reached by Kosswig (1948) and with
analyses indicating the type and number of genes
involved in the species differences of other kinds
of animals (Dobzhansky, 1937). The intermedi-
ate appearance of the Fi, the more variable Fa,
and the return in appearance toward successive
backcross parents are all classical indications of

11 Gordon (1958b) briefly reported on the first three
crosses listed here.

polygenic inheritance. In none of these ways,
however, did the present hybrids perform entire-
ly in the classical manner; in fact, unorthodox
responses were sometimes as much in evidence
as classical ones. For example, approximately
the same number of characters in the Fi resem-
ble either parent as are intermediate (see Table
II and discussion on p. 176). Although the clas-
sical genetic behavior of the present hybrids
would be hard to explain except on the basis of
polygenes, the failure to conform to the above
criteria does not indicate the converse. The ap-
pearance of unstable developmental systems as
a result of the mixing of two foreign, mutually
unadapted genotypes (Schmalhausen, 1949;
Lerner, 1954) could very well bring about seem-
ingly inconsistent results. Such a reduction in
genetic homeostasis, as Lerner has designated
it, could also account for the striking lack of
uniformity occasionally seen in Fi broods in
which these hybrids were as variable as the F2
broods arising from them, or even more so. Gor-
don & Rosen (1951) also found high variability
in two gonopodial characters of Fi hybrids be-
tween X. maculatus and X. hellerii. It is of in-
terest to note that Hubbs & Strawn (1957) have
warned that high variability in a population of
fish is not a safe criterion for hybrid fertility,
that is, for the presence of F«, F3 . . . Fn genera-
tions, because high variability is sometimes ex-
hibited by the Fi— which may be sterile. Hubbs
(1956) attributed this variability to the com-
bination of dissimilar developmental rates of the
parental forms, and this may be considered a
special case of the more general phenomenon of
genetic homeostasis mentioned above.

Genetic modification of the polymorphic tail
patterns has been an accepted but meagerly doc-
umented part of the concept of hereditary influ-
ences on pigmentation in fishes of the genus
Xiphophorus. Hybridization has served to reveal
additional examples of this phenomenon. Al-
though the presence or absence of each poly-
morphic pigmentary pattern is typically con-
trolled by a single gene, numerous other genes
undoubtedly influence its phenotypic expression.
The problem of identifying genic modifiers, the
vast majority of which have small plus or minus
effects, is at present insurmountable. The specific
effects of two modifications of the tail patterns
of X. maculatus have, however, been described
by Gordon and attributed to two dominant, non-
allelic genes. Similar modifications have ap-
peared in the tail patterns of a few of the pres-
ent interspecific hybrids (Table V).

On one occasion the wagtail effect was ob-
served when a female X. maculatus carrying the

170

Zoologica: New York Zoological Society

[47: 14

comet (Co) pattern was crossed with X. v. xiphi-
dium. Gordon (1946) showed that this extensor
effect, in which supplementary pigmentation
appears around almost all of the finrays, the
mouth and the operculi, results from the pres-
ence of a single dominant autosomal gene (E)
that does not belong to the series of multiple
alleles affecting tail patterns of which comet is
a member. He also showed that this factor was
present in X. hellerii. Whether the same factor
is present in X. v. xiphidium is problematical,
however, since the wagtail pattern is much more
variably and also less clearly expressed in the
maculatus— xiphidium hybrids than in the inter-
specific crosses in which Gordon was able to
demonstrate the presence of E. Moreover, back-
crossing to X. v. xiphidium heightened the ex-
pressivity of the wagtail pattern, an effect not
seen in Gordon’s fish, since the pattern appeared
in the Fi full-blown, so to speak. At the very
least, however, we may conclude that in some
individuals of X. v. xiphidium, there are genes
capable of modifying the comet pattern of X.
maculatus.

Gordon (1956b) described the modification
of the twin-spot (T) pattern of X. maculatus to
form the Guatemala crescent, and he presented
genetic evidence that a dominant, autosomal,
independently segregating gene is responsible,
which he designated Cg. He indicated that this
gene was probably present in an appreciable
number of X. hellerii. This pattern appeared,
but in a slightly modified form, when a twin-spot
X. maculatus was crossed with X. hellerii guen-
theri. This modification may be the result of
additional modifying factors or may represent
the phenotypic expression of an allele of Cg.
Another modification of the twin-spot pattern
of X. maculatus appeared in a cross with X.
montezumae cortezi. The extension of pigment
was, however, less definite than in the Guate-
mala crescent and was not associated with any
extra pigmentation around the mouth. With re-
gard to other possible modifications of tail pat-
terns in hybrids involving X. montezumae cor-
tezi, it might be mentioned that in a cross with
a one-spot (O) X. maculatus, this pattern was
exceedingly intense. The puzzling relationship
of crescent (C) to cut-crescent (Ct) shown in
one cross between two subspecies of X. variatus
could only be elucidated by further crosses. It
does serve to emphasize, however, the similarity
of the developmental processes involved in the
two patterns.

2. Genetics of Melanosis and Melanoma in
Hybrids.

The striking exaggeration of expression that

certain macromelanophore patterns exhibit un-
der the influence of foreign genotypes has at-
tracted more attention than any other feature of
the genetics of Xiphophorus. Gordon (1951b,
1957) has reviewed the extensive experiments
performed with X. maculatus, X. hellerii and
their hybrids and summarized the concepts that
have been developed from them. Crosses involv-
ing these and other members of the genus in new
combinations have yielded results that are most
easily interpreted in the same way. They there-
fore add to the confidence with which we may
accept the concepts developed by Gordon over
the years.

a. Capacity for Atypical Growth and Speci-
ficity of Macromelanophore Genes.

Gordon (1948) pointed out that the different
macromelanophore patterns of X. maculatus are
enhanced to different degrees in hybrids with
X. hellerii, and Gordon & Smith (1938) and
Gordon (1951b) showed that the same pattern,
viz. the spotted pattern of X. maculatus, is en-
hanced to different degrees in hybrids with dif-
ferent species. The same relationships hold for
the macromelanophore patterns of species other
than X. maculatus (Table VII). On the basis of
present work and that of Gordon (1948, 1951b),
the allelic series of sex-linked dominants may be
arranged according to their potency for melan-
osis and melanoma production in hybrid com-
binations:

Sp (spotted) X. maculatus
Sb (spotted belly) X. maculatus
N (nigra or black-sided) X. maculatus
Sd (spotted dorsal) X. maculatus
Sr (striped) X. maculatus
Sp (spotted) X. variatus xiphidium
Sp (spotted) X. variatus variatus

The specificity that these genes exhibit in their
morphological manifestations must, of course,
be the product of equally specific physiological
processes, and these genetically controlled events
maintain a definite measure of specificity under
the abnormal conditions imposed by hybridiza-
tion, even though the limits of their phenotypic
variability may be considerably increased. It is
to be noted, for example, that certain hybrids in
which two major macromelanophore genes are
present show one pigmentary pattern in an en-
hanced form while the other remains within nor-
mal limits or even suffers some loss in expres-
sivity (Table VI, and item 8 on p. 169). One of
the reasons for the series of backcrosses and in-
breedings conducted with the offspring of the
cross of a spotted X. hellerii guentheri with X.
maculatus (Table VI, nos. 1 19, 133-137) was

Table VII. Macromelanophore Genes of Fishes of the Genus Xiphophorus
Ranked According to their Capacity for Atypical Growth in Hybrids1

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

171

O U

+-‘ a

C M

o o
Js <3

u S3

*5

«w y

0 9

D 2

•o -c
0 3

-3 _
W rt

.a g

73 _
<D rt

4*d 9
G 2

a *9

n3 *.
S a

9 C3

s-S

73 _
<D -G

m 9
G 2
3 -9

*3 j.

B a
0 a

£ a

O *G

*3 u_,

a a

s s

O *G

73 ^
H G
73 03

;§ .a

s-g

^ £

x 9

>< s

>< 9

9 £

3 1

X 1

t-4

CD O
00 73

CD O
CO 73

cu O
03 T3

CD O
CO 73

<1 T3

< -a

cd O
CO 73

S3

o g

^ Ui

'a ^

gj

o

q G

S

S 2

O Oh
O CO

hJ 4-.

M o

CD

a,

co

<

g

o

Z

G

o

£

rH %

^ ^ O

g 03 "S
° n-S

” -e

.9 CO 8

3.1 8

a_l CO

Is

Ip

cd G o

O 5

O rO

73

a

03

>>

73

O

P3

[L, .g

<D rG

E §

O O

2 3
E -S
q y

<D ’’G

s §

c/d a

o -s

o .2
44 CO
O O

2 a

S3 C3

T3 "3 ,
c E

(N

g

C/D (D

S 9

id O

*G CO

a a

<u o

o

a

o

.9

co n
G .2

s s

§ -9

<d G

S 8

CD 73 O

C G G

?ES

73
G

CN bA*-*

<d g 2

c g.g

O 2 4P

s

o
o

5 ffS «

2 a H
JlSfc
1 G o

CJ ,*_>

•S 3
3 1
§1*
J’| I

<l> 9 ■*-’
o

% 8.9

a a
2 o
9 'g

<D CC J

o ca
§ £
•g 6

G O
<D (j

bOr^

G

s «

£ G
CO o

CD

G

O

co

.5 g
^ .0

§ g

§ -9
0 2
G 5

J3

G

W

c§

o

Q

rG

G

73

G

a

>>

73

o

G3

G

2 co

°<S

73

G

73

G

G

>>

73

o

P3

<G

*G

73

G

G

o

rG 45

*13 -

. G *

>A U

'gG -5

0 G

73

G

w S w

a .3

a 3

S3 -ft
a .ft.
^ £

^ ts?

ft

o t:
g o
^ o

><

G

O

Z

<d

G

o

Z

rG

G

73

G

G

>Y

TD

O

CQ

a

^5 r

°£!

S S3

cu

S’B

a

.2 2

’53

-ft

■**»

ft

0 t:

S 8

**••

§ ft

-ft

'U

a

Oc

<^v

s

73

CO

0

CO

'w'

*c3

CO

/**s

*G

73

/*s

/*■* \

a.

CO

(H

0

Q

u

CO

Oh

CO

G

a

O

CL

03

'w'

Oh

CO

Oh

CO

•w

73

CD

73

<L>

73

(D

73

<D

73

CD

73

CD

73

<D

73

<D

O

O

.Oh

O

O

O

O

O

Cu,

Ph

l-i

Oh

Oh

Oh

Oh

Oh

CO

CO

55

CO

CO

CO

CO

CO

n G G O *d
° (D CL) •*-* T2
r< ^

5 C G

.Sa'O
GT* _r > 0

a 5 «

9 S-s

X) o ^

Gh

.2

55

o oj;

C'2i E « a o c

8 S 3 "“IN'S

tt co G & g •G

^ ^ .5P.S2 .S 'f -a g

o a es . § - °

M-i G V-I G
O G ^ O 3
„ >,43 ^

G cfl G D-O O
O -S O « ~ W
G o g G o o .Q
wo o G G .SS ^ O

Is 2 is > o ■s

ce a-o-g'Sho g>
^ 9 « 2 -5 >- -S -S
X G 3 G « 2 *1 t*

« y -s s

_r W >^42) t/mn tn o

s^J-slU

<L> CD

* ^

G
W) w
-* 00

03 rl

.G .9 ^

g o fl §

73 o 0

a S'S is ^2

*5 rG G G

a 53«-F«a

■9 <u O ok A 45*

O D

‘ CL,

tn *r

2 ‘H 3
^ CO W g 2,

§Ie^ 3

o „

. *. O C

” 3'^ 5 fo-ga-c y

9 a u “ x „ ,

° &,1" ft o+'O c y

^W^k-nCUDC

<£ * < * g “ s .&

O o ^ CL

<D^ r: G ^ ‘-'G
1— 1 a ■ d_ o •»

00 w S ^ ^ S< G
G r> . <d -- H

' " J-S 6

44 -a

■ rS -2
1 2 = 45 X

I C cj

o a -

CS M

pa oa

^ v c: r “

f^£l23l'

3;x.i ^

40 ^ 9 i* "

If I i IS gS
s * s 3 •'g
g-s.g

- ^ -s 8 » JS

G

O

00

o

a

o

43

G

O

G

O

G

5

o

5

U2 G

G O

<D jG
00 C/D

M S

• <5d 43

43 W

C3 td h

w : “-S

; O 9 ? H

G

O

Z

2 ^ »9 ^ 5 csJ

«5 ^ G 00. S ^

172

Zoologica: New York Zoological Society

[47: 14

to test what might for convenience be called the
integrity of the Sp gene. The rationale was to in-
troduce the gene into the X. maculatus genotype
by the appropriate cross, followed by several
backcrosses to X. maculatus. Reintroducing the
gene into X. h. guentheri would reveal a modi-
fication of it (through the acquisition of gene
modifiers or some change in the gene itself ) in
the form of enhanced expressivity. Inbreeding
at various stages in the process would serve to
increase the variability of genetic combinations
and bring out latent factor interaction. The an-
ticipated negative results were obtained in full
measure; no change could be detected.

On the other hand, the fact that X. maculatus.
the species in which the most potent melanosis-
and melanoma-producing genes occur, has never
been known to give rise to hybrids exhibiting any
enhancement of patterns from other species
(Table VI, nos. 79, 84, 98, 119) points to
some similarity in physiological process under-
lying the actions of all the macromelanophore
genes (Atz, 1959b). According to Gordon’s hy-
pothesis of the integration of macromelanophore
genes into the genotype of X. maculatus (see p.
173), this species possesses several modifying
genes that control the growth of its own macro-
melanophores. Presumably this control is gen-
eral enough to exercise restraint on macromel-
anophore genes from other species, and this in
turn presupposes some similarity in the biochem-
ical processes of the genes being regulated. The
existence of species like X. hellerii, the hybrids
of which seem to be especially likely to exhibit
enhanced pigmentary patterns, similarity pro-
vides support for the idea of biochemical char-
acteristics common to all macromelanophore
genes; but the fact that the hybrids of X. couch-
ianus and X. p. pygmaeus, both of which lack
macromelanophore patterns, do not show as
great a degree of atypical macromelanophore
growth as do those of X. hellerii and X. m. cor-
tezi, which do have such patterns, indicates im-
portant differences between the macromelano-
phores of X. maculatus and X. variants on the
one hand and those of X. hellerii and X. m. cor-
tezi on the other. Evidently the ability of the
latter two species to control the growth of their
own macromelanophore patterns confers little
or no ability to control the growth of others. Of
less critical value is the observation that the
manifestations of atypical growth shown by the
macromelanophores from different species ap-
pear to be basically the same. In all known
hybrid combinations, melanosis necessarily pre-
cedes melanoma. Reed, Gordon & Lansing
(1933) and Gordon & Smith (1938) have de-
scribed several characteristics in which the mel-

anoses and melanomas of maculatus-couchianus
hybrids resemble those of the more completely
described maculatus-hellerii ones, and Gordon
& Nigrelli (1949) stated that the melanomas
found in cortezi-hellerii hybrids are histologi-
cally not unlike those found in other combin-
ations.

Anders et al. (1961, 1962) have reported that
the amount of free amino acids in various spe-
cies of Xiphophorus varies inversely with the
ability of their macromelanophores to produce
tumors in hybrids: X. maculatus has the smallest
quantity, X. variants xiphidium a greater
amount, X. montezumae cortezi still more and
X. hellerii guentheri the greatest. This is the
order in which the authors rank the species ac-
cording to the potency of their macromelano-
phores for atypical growth in various hybrid
combinations, X. maculatus having the most
potent genetic factors and X. hellerii the least.
This arrangement agrees substantially with the
one in Table VII, which is based on independ-
ently acquired data. The significance of the
apparent relationship between tumor produc-
tion or susceptibility and amino acids has yet
to be explained.

b. Evidence for the Evolution of Macromel-
anophore Genes and Their Polygenic Modifiers.

The tendency for increased variability in the
F2, the return toward the parental type in one
backcross and the exaggeration of the modify-
ing effect in the other are all most easily ac-
counted for by assuming the presence of several
modifying factors, as Kosswig (1931) did in
order to explain similar findings involving X.
hellerii and X. maculatus. Kosswig and Gordon
(1937) explained their results with these two
species on the basis of a few, perhaps only two,
pairs of modifiers. The similarity of results be-
tween the present crosses and theirs, particu-
larly the rapidity with which a plateau of en-
hancement is reached with repeated backcrosses
to the Pi carrying the modifying genes, indicates
that approximately the same number of modi-
fiers is involved in the presently studied species
interactions.

Gordon (1950b, 1951b, 1958a), and Gordon
& Gordon (1957) considered the production of
melanosis and melanoma in Xiphophorus hy-
brids from an evolutionary point of view. They
described how the combination of a potentially
harmful gene, like Sp, and a series of modifying
genes that render it harmless might have arisen.
When the mutation to Sp first occurred, it is
supposed to have been detrimental in a manner
similar to the way Sp leads to melanosis and
melanoma when introduced into foreign gen-

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

173

omes today.12 Fisher (1928) and others have
shown that such a gene, whose over-all effect
is deleterious, will accumulate genic modifiers
that tend to make the expression of its harmful
effects recessive and, sometimes, eventually
mask them completely— providing that the locus
in question continues to yield mutant genes of
the type upon which natural selection can oper-
ate. If this has indeed been the history of the
integration of macromelanophore genes into
the genotype of X. maculatus, the genotypes of
the other species belonging to the genus Xipho-
phorus might better be considered to lack the
genes or alleles that control the expression of the
macromelanophore gene, rather than to possess
genes that enhance its expression, as postulated
by Kosswig ( 1 93 1 ) , or to remove a specific inhib-
itor that controls its normal growth and multi-
plication, as suggested by Gordon & Smith
(1938). This clarifying view, which represents
more than just a change in the terminology used
to describe the interacting genomes, was sug-
gested in 1958 by Bonn E. Rosen, who was then
a member of the staff of the Genetics Labora-
tory of the New York Aquarium.

Gordon (1951a, b, 1957) showed that the
same kinds of abnormalities in the development
of pigment patterns occur in intraspecific hy-
brids as do in interspecific ones, although they
are not as strongly expressed. He considered this
to be good evidence that subspecies and other
geographically isolated populations of Xipho-
phorus are proceeding along paths of genetic
differentiation similar to those already taken by
the species. In a small series of crosses between
the two subspecies of X. montezumae, two mac-
romelanophore patterns from one of them were
enhanced to an abnormal degree (Table VI,
nos. 93, 95, 113-116). This provides another
example of incipient melanosis, presumably the
result of the mixing of different constellations
of modifying genes that have evolved at the sub-
specific level of genetic differentiation.

In evaluating the significance of the pigment
cell abnormalities of hybrids as a criterion of
taxonomic relationship, Rosen (1960) pointed
out that the degree of abnormality could act as
a measure of the degree of genetic difference and
should, therefore, enter into the determination of

12 A difficulty of this hypothesis is that it makes no
provision for the development of the multiple allelism
often associated with polymorphism. It therefore im-
plies that despite the apparent similarity of the poly-
morphism of X. maculatus and X. variatus to that found
in other kinds of animals (most notably insects), its
origin has not been the same (see, for example, Ford,
1957).

systematic relationship. He indicated, however,
that the correlation between the relative tax-
onomic position (as determined by a whole
array of pertinent data) of any two non-inter-
breeding groups (species, subspecies or merely
river-system populations) and the hybrid pig-
mentary abnormality exhibited by their hybrids
is far from perfect. The present data support
this view. For example, intraspecific crosses, be-
tween subspecies, may or may not result in
offspring with atypical macromelanophore pat-
terns (Table VI, nos. 61-65, 74, 75-77, 92-95,
113-116, 141). In interspecific crosses, the same
macromelanophore gene may lead to an en-
hanced pattern in some combinations or a nor-
mal pattern, or none at all (zero penetrance) in
others (Table VI, compare 69, 70, 71 with 78
and 79; 99 with 97, 98). Moreover, in the same
interspecific cross, one pattern may be enhanced
and another unaffected or reduced in expressiv-
ity (Table VI, compare 18 and 19 with 98; 99
with 87 and 88). It is nevertheless true that
pigmentary abnormalities in interspecific hy-
brids are more severe and more frequently en-
countered than in intraspecific ones, as Gordon
(1951b) and Rosen (1960) have indicated. Yet
even here the distinction is not absolute, for one
case is known in which the hybridization of sub-
species produced a more atypical pattern than
did crosses with four other species (Table VI,
compare 93, 95 with 82, 84-88; 113-116 with
96-98).

In his illuminating discussion of homology,
de Beer (1958) cited a case of pigment pattern
modification in a moth that closely resembles
the situation in two geographically isolated pop-
ulations of Xiphophorus maculatus described by
Gordon ( 1951a) . When a platyfish from the Rio
Coatzacoalcos that carries the spotted dorsal
gene (Sd) is crossed with a platyfish from the
Rio Jamapa, the expressivity of the gene is en-
hanced, the pattern in the hybrid covering a
considerably greater area of dorsal fin and ad-
jacent body than the spotted dorsal of either
pure race, both of which are very similar in
appearance. Similarly, when Ford (1953)
crossed individuals belonging to a dark sub-
species of Triphaena comes from two widely
separated islands, the expression of the pattern
was modified. Since other tests had shown that
the same principal gene was responsible for the
color pattern on both islands, the change in its
expressivity in the hybrid must have resulted
from the interaction of two different systems of
genetic modifiers, each of which had neverthe-
less been responsible for producing an identical
phenotype. De Beer (1958) concluded that al-
though the two pigment patterns were pheno-

174

Zoologica: New York Zoological Society

[47: 14

typically alike and were controlled by the same
principal gene, they were not homologous since
they were the result of parallel evolution from a
common ancestor that undoubtedly did not ex-
hibit the dark pattern, at least not in the form
in which it exists at present. The same conclu-
sions could be reached with regard to the sit-
uation Gordon found in X. maculatus.

A somewhat different situation has been
found in two of the subspecies of X. variatus.
Each of these has at least one spotted (Sp) mac-
romelanophore pattern, and the two studied at
this time are morphologically distinguishable
(see p. 162). In only one of a number of crosses
between X. v. variatus and X. v. xiphidium did
any sign of atypical expression of their spotted
patterns occur (Table VI, compare 77 with 61-
65, 75, 76). Evidently these two alleles are sim-
ilar enough to be controlled by the same set
or sets of modifying factors, and evidently the
constellation of modifying genes is the same
or nearly the same in both subspecies. Never-
theless, when associated with hybrid genomes
involving two other species, X. p. pygmaeus and
X. montezumae cortezi, these alleles react quite
differently (Table VI, compare 58-60 with 72,
73; 57 with 69-71). Presumably the relatively
slight difference between them could become
manifest in a genetic environment where the
major gene and approximately half of its poly-
genic modifiers were not mutually adapted and
therefore allowed less “margin for error” dur-
ing ontogenesis, that is, maintained a lower
level of genetic homeostasis.

c. Atypical Pigmentation Associated with the
Sc Gene.

Kosswig (1936) was the first to describe the
enhancement of the spotted caudal pattern (Sc)
of X. montezumae cortezi in hybrids with X.
hellerii. In a preliminary report, Gordon
(1947b) pointed out that melanomas were ob-
tained “not in the first generation hybrids, but
in some of the backcrosses of the hybrids to
X. hellerii and in some of the second inbred
generation.” According to the Laboratory data
gathered together in Table VI, melanotic over-
growths were obtained among the second back-
cross hybrids and in the second inbred genera-
tion of the first backcross hybrids (nos. 103 and
109). The specimens available at the present
time show the following general relationships
between type of cross and severity of pigmen-
tary abnormality:

Fi— enhancement

1st backcross— melanosis
Inbreeding, 1st generation— enhancement
Inbreeding, 2nd generation— melanoma

Inbreeding, 3rd generation— melanoma
Inbreeding, 4th generation— melanosis

2nd backcross— melanoma
Inbreeding— melanoma

3rd backcross— melanoma
Inbreeding— melanoma

The percentage of melanomatous individuals
in any one brood varied from, roughly, 5% to
30%. The variability of the extent to which
melanoses and melanomas were exhibited by
various backcross broods may indicate that the
factors favoring enhancement of pigment cell
growth are not uniformly distributed in X. hel-
lerii. There is, however, a complicating factor
that makes the results appear distinctly less uni-
form than those from comparable crosses be-
tween X. hellerii and macromelanophore-carry-
ing X. maculatus, namely, the slow development
of melanosis and the late appearance of melan-
oma among these fish. For example, in our
present strain of melanomatous hellerii-cortezi
hybrids (RJ), which is maintained by repeated
backcrossing to X. hellerii, overgrowths do not
appear until the fish are nearly two years old
(Klaus D. Kallman, personal communication).
Unless all specimens are maintained for uni-
formly long periods of time, which was not the
case with those discussed above, an uncontrolled
variable of major proportions will have been
introduced.

Gordon & Nigrelli (1949) briefly described
the histology of the overgrowths associated with
the Sc pattern. They found that these melano-
mas somewhat resemble the others that occur
in the tail region of various hybrids— presum-
ably associated with Sp, Sb and Sd of X. macu-
latus since these factors also lead to overgrowths
on the caudal peduncle. As in all of the melan-
omas, melanosis always precedes the appearance
of any Sc overgrowth. This sequence has been
witnessed several times, and individuals with
overgrowths always had well developed melan-
oses while the reverse was frequently not the
case. The extensive invasion of tissues by black
pigmentation begins near the juncture of caudal
fin and peduncle, that is, the general location
of the Sc pattern. The gradual encroachment
of the melanosis up the caudal peduncle and,
subsequently, the body has been observed, but
all excessive pigmentation does not arise from
this single spreading concentration. In many of
the Sc hybrids, clusters of macromelanophores
were seen in areas quite far anterior to the mel-
anotic region on the peduncle (Figs. 16, 17).
Slashes of intensely black pigmentation also
sometimes appeared in the dorsal fin and, less
frequently, the pectorals. The latter, it should be

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

175

noted, are never pigmented in normal fish of
either of the parental species. Whether these
islands of macromelanophores hasten the for-
ward advance of melanosis or contribute mate-
rially to its development cannot be determined
from preserved material, but it seems likely. In
general, the number of clusters of macromelano-
phores could be roughly correlated with the
extent of pigmentation of the hybrid brood in
question, but this was not always the case. For
example, clusters of macromelanophores were
seen in seven fish that showed no sign of the
Sc pattern (4 from no. 107, 1 from 109 and 2
from 112). The smallest specimen that exhibited
macromelanophore clusters was 9 mm. in stand-
ard length.

The cases of extreme melanosis, involving at
least half of the body in addition to the tail, and
those of melanotic overgrowths, were not dis-
tributed equally between the sexes of Sc hybrids :

No. males with extreme melanosis 6

No. males with melanoma 0

No. females with extreme

melanosis as well as melanoma 3

No. females with extreme

melanosis but without melanoma 1

No. females with melanoma but
without extreme melanosis 11

The melanoses of the males tended to be
greater, even when compared with the most ex-
tremely melanotic females. One second back-
cross male (h42) was completely melanotic
except for the eyes and tip of the gonopodium.
No connection between sex and melanosis or
melanoma has ever been established in the hy-
brids of X. hellerii-X. maculatus, and Berg &
Gordon (1953) found melanomas in five hy-
brids that had no detectable gonads. Neverthe-
less, the possibility that sex hormones or some
genetic factor associated with sex may influence
melanoma development in hybrids among the
species of Xiphophorus may again be broached,
on the basis of present observations.

Gordon (1956a) presented a brief, popular
account of the origin of the Red Jet (RJ) strain
in which he called attention to two noteworthy
genetic interactions responsible for the striking
red and black coloration of these fish. The
strain originated with the crossing of a wagtail
swordtail with an Sc-bearing member of the
inbred offspring of a second backcross to X.
hellerii of hellerii-cortezi hybrids (Table VI, no.
112). The wagtail fish must have acquired its
pattern through hybridization with X. maculatus
carrying the comet (Co) factor (Gordon,
1946), and therefore a modicum of genes from
that species must be present in the Red Jet

strain. Gordon (1956a) reported that in this
strain, the macromelanophore spotted caudal
(Sc) pattern is more enhanced when the micro-
melanophore wagtail (CoE) pattern is present.

This factor interaction is similar to the one
between the micromelanophore stippled (St)
and macromelanophore spotted (Sp) patterns
in X. maculatus which Gordon (1928) also de-
scribed. That this interaction takes place be-
tween cells or elements of tissue is indicated by
the very localized enhancement of macromelan-
ophore patterns that is occasionally seen at
places on the bodies of hybrids where concen-
trations of micromelanophores occur. For ex-
ample, an F2 fish from the cross of a striped,
one-spot (SrO) X. maculatus with X. couch-
ianus (Table VI, no. 36) exhibited a large,
melanotic spot on either side of the caudal
peduncle, where the one-spot pattern is located.
The Sr pattern in this fish was barely perceptible,
that is, it showed reduced expressivity. At least
one other sib, and perhaps three, showed an
early stage of a similar macromelanophore spot.
In the latter three fish, the Sr pattern was fairly
well developed but not enhanced in any way—
except in the immediate vicinity of the one-
spot.™

Gordon (1956a) indicated that the red col-
oration of the Red Jet strain was the result of
an enhancement of the three or four horizontal
red stripes characteristic of the subspecies X.
hellerii guentheri. Gordon (1948, 1950a) had
described how the red-colored patterns of X.
maculatus are enhanced in hellerii-maculatus
hybrids and how erythrophoromas appear, in
rare instances, among such hybrids. Similar
interactions can occur in other kinds of hybrids,
and we have seen X. maculatus-X. couchiamts
hybrids in which the red dorsal (Dr) pattern of
the common platyfish parent was considerably
enhanced. Kosswig (1937, 1948, 1959) has in-
dicated that the same pattern is enhanced in
hybrids with X. v. variatus and that similar in-
tensification and spread of red pigment patterns
occur in hybrids with X. variatus xiphidium.

3. Further Aspects of Hybridization.

In his critical review of what is known about
hybridization among North American fishes,
Hubbs (1955) states that “It has proved to be an
almost universally valid rule that natural inter-

13 Another indication of localized interaction between
pigment cells is the halo effect in which a macromelano-
phore spot is surrounded by an area that appears to be
free of all melanophores. (See Fig. 13.) Enzymatic com-
petition for a limited amount of substrate might be the
explanation, and a search for unpigmented pigment cells
within the halo should be made. (See p. 162 for addi-
tional comment on the halo effect.)

176

Zoologica: New York Zoological Society

[47: 14

specific hybrids are intermediate between their
parental species in all characters in which those
species differ, whether they be external or in-
ternal, of shape, color, form, structure, or
numbers of parts (vertebrae, gillrakers, finrays,
teeth)— except for some features that reflect hy-
brid vigor.” There is no question that the over-
all appearance of the vast majority of fish
hybrids is intermediate between the parental
forms, but when a detailed, character-by-char-
acter comparison is made, an appreciable num-
ber of characters may be found to resemble one
parent only. The present hybrids illustrate this
very well (Table II), and numerous other ex-
amples have been reported in the literature (Atz,
1959a). Such instances of “dominance” do not,
of course, indicate that the character in question
is inherited in a simple Mendelian fashion, as
Newman (1914) recognized. The genotypes of
even closely related species differ in many genes
(Dobzhansky, 1937), and polygenic inheritance
is to be expected in the vast majority of cases.
There may exist, however, some regular associ-
ation between the relative amounts of “domin-
ance” and intermediateness that are exhibited
by an Fi hybrid and the phylogenetic relation-
ship of the parental forms. The more closely re-
lated the parents are, the less loci are involved in
genetic differences, not only in the aggregate
but presumably in the control of individual
characters as well. The smaller the number of
different alleles involved in determining a char-
acter, the greater the chance of those from one
parent being completely dominant over those
from the other. From this it may be deduced
that, in general, the greater the proportion of
Fi hybrid characters resembling one of the par-
ental forms rather than being intermediate, the
closer to each other are the parents phylogenet-
ically. This may be the reason why a relatively
large proportion of characters in the present
hybrids show “dominance” of one parental
character or the other, while many of the hy-
brids described by Hubbs and others are inter-
mediate to a significantly greater extent. More-
over, there are instances among the present hy-
brids in which a character unlike that in either
parental form appeared, or in which a character
resembling that found in a third species ap-
peared (see item 11 on p. 158).

Results remarkably similar to the ones ob-
tained with the present hybrids were described
by Newman (1914) for the pigmentation of
Fi hybrids among Fundulus heteroclitus, F.
majalis and F. diaphanus.u

Although the present specimens were not ex-

14 This paper was not seen until after the present data
had been gathered and analyzed.

amined for signs of hybrid incompatibility other
than pigmentary ones, this phenomenon was
manifest in abnormal sex ratios and the presence
of large, sexually undifferentiated fish. In gen-
eral there was a shortage of males, which is in
accord with Haldane’s Rule that the heterogam-
etic sex is more likely to be sterile, rare or absent
in animal hybrids. As far as is known, the male
Xiphophorus is heterogametic save for the platy-
fish that inhabit British Honduras (Gordon,
1957). These and other manifestations of hy-
brid incompatibility have been discussed in
detail by Rosen (1960) . We need only point out
that in no species combination where a con-
certed effort has been made has it proved im-
possible to obtain Fi, Fa, F3 or backcross gen-
erations.15 Hybrid sterility may in fact exist
between some species of Xiphophorus, since re-
duced fertility is frequently encountered, but
this has not yet been demonstrated.

It must be clear from the preceding dis-
cussion (see especially pp. 172-173) that atyp-
ical pigment cell growth in Xiphophorus cannot
be considered an isolating mechanism, as Crew
(1940) and Stebbins (1958) have done. That
some isolating mechanism must exist is evident,
however, from the following observations: (1)
X. hellerii and X. maculatus are found side-by-
side in several river systems in southern Mexico,
Guatemala and British Honduras; X. hellerii
and X. v. variatus both inhabit at least one trib-
utary of the Rio Nautla; X. v. variatus, X. p.
pygmaeus and X. montezumae cortezi exist close
to one another in the Rio Axtla and all three
species have been caught there in a single pool
(by Gordon, Atz and F. G. Wood, Jr. in 1948) ;
(2) All of these particular species combinations
have produced hybrids in the laboratory (Table
I) ;( 3) Among thousands of specimens of Xipho-
phorus collected in nature, not a single hybrid
has ever been discovered (Rosen, 1960). Lab-
oratory studies (Clark, Aronson & Gordon,
1954) and data collected in the field (Rosen,
1960) indicate strongly that a complex of
factors, is responsible for keeping the sympatric
species of Xiphophorus reproductively apart.
The relative importance of the several factors
and details of how they operate have yet to be
determined.

V. Summary

1. Previous genetic studies of fishes of the
genus Xiphophorus have concentrated on X.
maculatus, X. hellerii and their hybrids. The

15 Tri-hybrids have been produced involving X. macu-
latus, X. variatus and X. couchianus (Table VI, nos. 21,
32, 68, 81) and X. maculatus, X variatus and X. hellerii
(Kosswig, 1935a,b). The former were never bred, but
Kosswig found that the latter were fertile.

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

177

present studies are principally concerned with
five additional species and subspecies ( X . couch-
ianus, X. variatus xiphidium, X. v. variatus, X.
montezumae cortezi, X. pygmaeus pygmaeus)
and may be summarized by stating that they con-
firm and extend the work of Myron Gordon on
this group of poeciliid fishes.

2. The pigmentary patterns of these fishes
may be separated into monomorphic and poly-
morphic components and the latter into patterns
composed of micromelanophores or macromel-
ophores. The behavior of the monomorphic pat-
terns in hybrid crosses shows that they are con-
trolled by polygenes and that, in some cases the
number of genes involved is not large.

3. The changed appearance of polymorphic
pigmentary patterns in various hybrid combina-
tions also reveals probable methods of genetic
control; both micro- and macromelanophore
patterns are undoubtedly influenced by constel-
lations of modifying genes. A few new examples
of modified tail patterns (micromelanophore)
and several of spotted patterns (macromelano-
phore), including the spotted caudal pattern,
have been studied. In hybrids, tail patterns re-
sembling the wagtail and Guatemala crescent
and spotted patterns ranging from zero pene-
trance to melanosis and melanoma production
have been observed.

4. The major pigment cell genes vary in their
ability to produce melanosis and melanoma in
hybrids, and the species vary in their ability to
control the growth of pigmentary patterns from
other forms. X. maculatus, which possesses the
genes most potent in producing atypical pigment
cell growth in hybrids, is also the only species
in whose hybrids no enhancement of macromel-
anophore patterns from other species ever oc-
curs.

5. This indicates a similarity of biochemical
processes among all the major genes for macro-
melanophore patterns, but a highly developed
specificity also exists, because in the same hybrid
individual, the expressivity of one macromelano-
phore pattern may be increased while that of an-
other is reduced or remains unchanged.

6. The concept of genetic homostasis, as de-
veloped by Schmalhausen and by Lerner, has
proved useful in explaining departures from the
results expected in typical polygenic inheritance
and also the differences in expressivity in hybrids
of genetic factors that act practically the same
when associated with genomes more like their
normal ones.

7. The appearance, in various hybrids, of the
spotted patterns of X. variatus xiphidium and X.
v. variatus, and crosses between X. montezumae

cortezi and X. m. montezumae, indicate that the
same kinds of genetic differentiation occur in
subspecies as in species.

8. The frequency with which the characters
of hybrids resemble those of either parent, rather
than being intermediate, may be a measure of
the phylogenetic relationship of the parental
forms. The more closely these are related, the
greater the proportion of characters that are not
intermediate.

9. No absolute hybrid sterility between any
of the fishes of the genus Xiphophorus has ever
been demonstrated. Although the abnormal ex-
pressivity of pigment genes in hybrids leads to
lethal melanoma in some crosses, it seems never
to have served as an isolating mechanism.

VI. References

Anders, Fritz, Frederic Vester, Kate Klinke &
Hans Schumacher

1961. Genetisch bedingte Tumoren und der Ge-
halt an freien Aminosauren bei lebendge-
barenden Zahnkarpfen ( Poeciliidae ).
Experientia, 17 (12): 549-551.

1962. Genetische und biochemische Untersuch-
ungen iiber die Bedeutung der freien
Aminosauren fur die Tumorgenese bei
Art- bzw. Gattungsbastarden lebendgeba-
render Zahnkarpfen (Poeciliidae). Biol.
Zentralblatt, 81 (1/2): 45-65.

Atz, James W.

1941. From Quetzalcoatl to cancer. Bull. N. Y.
Zool. Soc., 44 (3): 76-83.

1959a. Morphological and genetic studies on the
pigmentary patterns of xiphophorin fishes
and their hybrids. PhD Thesis. Dept. Biol-
ogy, New York Univ.

1959b. Genetically controlled atypical pigment
cell growth among the hybrids of six spe-
cies of xiphophorin fishes. Genetics, 44
(4): 499. Abstract. [Rec. Genetics Soc.
Amer., (28): 59].

Atz, James W. & Donn E. Rosen

1959. Myron Gordon 1899-1959. Copeia, 1959
(4): 352-354.

de Beer, Gavin R.

1958. “Embryos and Ancestors.” Third ed. Ox-
ford. xii+197 pp.

Bellamy, A. W.

1936. Inter-specific hybrids in Platypoecilus: one
species ZZ-WZ; the other XY-XX. Proc.
Nat. Acad. Sci. 22 (9): 531-535.

Berg, Olga & Myron Gordon

1953. Relationship of atypical pigment cell
growth to gonadal development in hybrid
fishes. Pp. 43-72 of “Pigment Cell
Growth.” Academic Press, N.Y.

178

Zoologica: New York Zoological Society

[47: 14

Clark, Eugenie, Lester R. Aronson & Myron

Gordon

1954. Mating behavior patterns in two sympatric
species of xiphophorin fishes: their inherit-
ance and significance in sexual isolation.
Bull. Amer. Mus. Nat. Hist., 103 (2):
135-226.

Crew, F. A. E.

1940. Artificial insemination: its uses in the ex-
perimental study of the species problem.
Boll. Soc. Ital. Biol. Sperim., 15 (1):
75-80.

Dobzhansky, Theodosius

1937. “Genetics and the Origin of Species.” Co-
lumbia Univ. Pres, N.Y. xvi+364 pp.

Fisher, R. A.

1928. The possible modification of the response
of the wild type to recurrent mutations.
Amer. Nat., 62 (679): 115-126.

Ford, E. B.

1953. The genetics of polymorphism in the Lepi-
doptera. Advances Gen., 5: 43-87.

1957. Polymorphism in plants, animals and
man. Nature, 180 (4598): 1315-1319.

Gordon, Hugh & Myron Gordon

1957. Maintenance of polymorphism by poten-
tially injurious genes in eight natural pop-
ulations of the platyfish, Xiphophorus
maculatus. Jour. Genetics, 55 (1): 1-44.

Gordon, Myron

1928. Pigment inheritance in the Mexican killi-
fish. Jour. Hered., 19 (12): 551-556.

1931a. Morphology of the heritable color pat-
terns in the Mexican killifish, Platypoe-
cilus. Amer. Jour. Cancer, 15 (2): 732-
787.

1931b. Hereditary basis of melanosis in hybrid
fishes. Amer. Jour. Cancer, 15 (3): 1495-
1523.

1933. Sex-linked factors in crosses between vari-
eties, species and genera of Mexican fishes.
Amer. Nat., 67 (708): 75-76. Abstract.

1937. The production of spontaneous neoplasms
in fishes by selective matings. II. Neo-
plasms with macromelanphores only. III.
Neoplasms in day old fishes. Amer. Jour.
Cancer, 30 (2): 362-375.

1941. Hybridization in seven species of Mexican
fishes following removal of the geographi-
cal factors of isolation. Anat. Rec., 81 (4,
Suppl.): 112-113. Abstract.

1943. Genetic studies of speciation in the sword-
tail-platyfish group and of the experimen-
tally produced hybrids. Trans. N.Y. Acad.
Sci., Ser. II, 5 (4): 63-71.

1946. Introgressive hybridization in domesti-
cated fishes. The behavior of comet, a
Platypoecilus maculatus gene, in Xipho-

phorus hellerii. Zoologica [N.Y.], 31 (2):
77-88.

1947a. Speciation in fishes. Distribution in time
and space of seven dominant multiple
alleles in Platypoecilus maculatus. Ad-
vances Gen. 1: 95-132.

1947b. Melanomas in the hybrid offspring of two
species of swordtails, Xiphophorus monte-
zumae and Xiphophorus hellerii. Anat.
Rec., 99 (4): 613. Abstract.

1948. Effects of five primary genes on the site
of melanoma in fishes and the influence of
two color genes on their pigmentation.
Pp. 216-268 of “The Biology of Melan-
omas,” Spec. Pub. N. Y. Acad. Sci., 4.

1950a. Heredity of pigmented tumours in fish.
Endeavour, 9 (33): 26-34.

1950b. The origin of modifying genes that influ-
ence the normal and atypical growth of
pigment cells in fishes. Zoologica [N.Y.],
35 ( 1 ) : 19-20. Abstract.

1950c. Fishes as laboratory animals. Pp. 345-449
of “The Care and Breeding of Laboratory
Animals.” Wiley, N. Y.

1951a. The variable expressivity of a pigment
cell gene from zero effect to melanotic
tumor induction. Cancer Res., 11 (9):
676-686.

1951b. Genetic and correlated studies of normal
and atypical pigment cell growth. Growth,
Symposium, 10: 153-219.

1953. The ecological niche of the pigmy sword-
tail, Xiphophorus pygmaeus, in the Rio
Axtla, Mexico. Copeia, 1953 (3): 148-
150.

1956a. The red jet swordtail. Tropical Fish Hob-
byist [Jersey City], 5 (1) : 6, 8, 38-39, 42-
43, 46-47. [Extract from: “Swordtails. The
Care and Breeding of Swordtails.” By
Myron Gordon. 24 pp. T.F.H. Publica-
tions, Inc., Jersey City, N. J.].

1956b. An intricate genetic system that controls
nine pigment cell patterns in the platyfish.
Zoologica [N. Y.], 41 (4): 153-162.

1957. Physiological genetics of fishes. Pp. 431-
501 of Vol. II of “The Physiology of Fish-
es.” Academic Press, N. Y.

1958a. A genetic concept for the origin of mel-
anomas. Ann. N. Y. Acad. Sci., 71 (6):
1213-1222.

1958b. Genetic and developmental differences be-
tween two morphologically similar pig-
ment cells with reference to melanoma.
Anat. Rec., 132 (3): 446. Abstract.

1959. The melanoma cell as an incompletely
differentiated pigment cell. Pp. 215-239 of
“Pigment Cell Biology.” Academic Press,
N. Y.

Gordon, Myron & Allan C. Fraser

1931. Pattern genes in the platyfish. Jour.
Hered., 22 (6): 168-185.

1962]

A tz: Effects of Hybridization on Pigmentation in Xiphophorus

179

Gordon, Myron & Ross F. Nigrelli

1949. The genetics and pathology of the melan-
omas in the hybrid offspring of two species
of swordtails, Xiphophorus montezumae
and Xiphophorus hellerii Cancer Res., 9
(9): 553. Abstract.

Gordon, Myron & Donn E. Rosen

1951. Genetics of species differences in the
morphology of the male genitalia of xipho-
phorin fishes. Bull. Amer. Mus. Nat. Hist.,
95 (7): 409-464.

Gordon, Myron & George M. Smith

1938. The production of a melanotic neoplastic
disease in fishes by selective matings. IV.
Genetics of geographical species hybrids.
Amer. Jour. Cancer, 34 (4): 543-565.

Hubbs, Carl L.

1955. Hybridization between fish species in na-
ture. Syst. Zool., 4(1): 1-20.

Hubbs, Clark

1956. Relative variability of hybrids between the
minnows, Notropis lepidus and N. pros-
erpinus. Texas Jour. Sci., 8 (4) 463-469.

Hubbs, Clark & Kirk Strawn

1957. Relative variability of hybrids between the
darters, Etheostoma spectabile and Per-
cina caprodes. Evolution, 11 (1): 1-10.

Kosswig, Curt

1931. Uber Geschwulstbildungen bei Fischbas-
tarden. 3. Mitteilung. Zeitschr. Indukt.
Abstamm. Vererbungsl., 59 (1): 61-76.

1935a. Die Kreuzung zweier XX- bzw. AF-Ge-
schlechter miteinander und der Ersatz
eines F-Chromosoms einer Art durch das
A-Chromosom einer anderen. Der Zuch-
ter, 7 (2): 40-48.

1935b. Genotypische und phanotypische Gesch-
lechtsbestimmung bei Zahnkarpfen. V.
Ein X (Z)-Chromosom als Y-Chromosom
in fremdem Erbgut. Roux’ Archiv Entwik-
lungs. Organ. 133 (1/2): 118-139.

1936. Kleinere Mitteilungen fiber Art- und Gat-
tungbastarde von Zahnkarpfen. Zool.
Anz., 114 (7/8): 195-206.

1937. Uber die veranderte Wirkung von Farb-
genen in fremden Genotypen. Biologia
Generalis, 13 (1): 276-293.

1948. Homologe und analoge Gene, parallele
Evolution und Konvergenz. Comm. Fac.
Sci. Univ. Ankara, 1: 126-177.

1959. Beitrage zur genetischen Analyse xiphoph-
oriner Zahnkarpfen. Biol. Zentralblatt, 78
(5): 711-718.

Lerner, I. Michael

1954. “Genetic Homeostasis.” Wiley, N. Y. vii
+ 134 pp.

MacIntyre, Pamela A.

1961. Crossing over within the macromelano-
phore gene in the platyfish, Xiphophorus
maculatus. Amer. Nat., 95 (884): 323-
324.

Newman, H. H.

1914. Modes of inheritance in teleost hybrids.
Jour. Exp. Zool., 16 (4): 447-499.

Nigrelli, Ross F. & Myron Gordon

1951. Spontaneous neoplasms in fishes. V. Aci-
nar adenocarcinoma of the pancreas in a
hybrid platyfish. Zoologica [N. Y.], 36
(2): 121-125.

Nigrelli, Ross F., Sophie Jakowska & Myron

Gordon

1951. The invasion and cell replacement of one
pigmented neoplastic growth by a second,
and more malignant type in expermiental
fishes. British Jour. Cancer, 5 (1): 54-68.

Reed, H. B., Myron Gordon & Winifred Lansing

1933. Study of the post-embryonic growth of a
hereditary melanotic neoplasm in hybrids
between two fish-species. Anat. Rec., 57
(4, Suppl.): 107. Abstract.

Rosen, Donn E.

1960. Middle-American poeciliid fishes of the
genus Xiphophorus. Bull. Florida State
Mus., Biol. Sci., 5 (4): 57-242.

Rust, Waldemar

1941. Genetische Untersuchungen fiber die
Geschlechtsbestimmungstypen bei Zahn-
karpfen unter besonderer Berficksichti-
gung von Artkreuzungen mit Platypoecilus
variatus. Zeitschr. Indukt. Abstamm. Ver-
erbungsl., 79 (3): 336-395.

Schmalhausen, I. I.

1949. Factors of Evolution. The Theory of Stab-
ilizing Selection.” Blakiston, Philadelphia,
xiv + 327 pp.

Stebbins, G. Ledyard

1958. The inviability, weakness, and sterility of
interspecific hybrids. Advances Gen., 9:
147-215.

180

Zoologica: New York Zoological Society

[47: 14

EXPLANATION OF THE PLATES

Living and preserved fishes photographed by Sam
Dunton, Photographer, New York Zoological So-
ciety. Prepared skins photographed by Dr. Ross F.
Nigrelli, Pathologist, New York Aquarium, New
York Zoological Society.

Dimension of fishes is Standard Length and is
approximate.

Plate I

Fig. 1. Hybrid X. maculatus X X. variatus xiphi-
dium. Male on the right shows wagtail-
like modification of comet (Co) pattern
of maculatus and slightly suppressed
spotted pattern of variatus. Both males
show the crescent (C) tail pattern. See
Table V and Table VI, nos. 26, 79.

Fig. 2. Backcross offspring, males ( X . maculatus
X X. v. xipliidium ) X X. v. x iphidium
Upper fish exhibits wagtail- like modifica-
tion of comet (Co) pattern. See Table V
and VI, no. 80.

Fig. 3. Hybrid X. v. xiphidium X X. maculatus.

Female (32 mm.) on right, exhibits a
slightly enhanced striped (Sr) pattern
from maculatus. Male (30 mm.), on left,
exhibits spotted (Sp) pattern from xiphid-
ium. See Table VI, nos. 27, 37, 78.

Plate II

Fig. 4. Xiphophorus v. variatus, female (40 mm.),
on left. Hybrid, X. maculatus X X. v. var-
iatus, male (35 mm.), on right. Male ex-
hibits spotted (Sp) pattern from variatus,
the spots being small and located in front
of and under the dorsal fin. Male also ex-
hibits spotted (Sp) pattern from macula-
tus, the spots being large and located
under the dorsal fin and behind it. (Pi of
h61).

Fig. 5. Backcross offspring, females (31-35 mm.),
showing different degrees of enhancement
of the spotted pattern from maculatus.
(h61).

Fig. 6. Backcross offspring, males. Two upper
fish exhibit the spotted pattern from vari-
atus', lowest, exceptional fish the spotted
pattern from maculatus. Three exceptional
(crossover?) males and one female ap-
peared among 41 females and 53 males.
No fish exhibited both spotted patterns,
thus, indicating that the two Sp factors
are alleles. (h61). See Table VI, nos. 17,
56.

Plate III

Fig. 7. Hybrid X. v. xiphidium X X. v. variatus.

Both fish exhibit an enhanced spotted (Sp)
pattern from xiphidium. Female, on right,
also exhibits the spotted (Sp) pattern from
variatus, the spots of which are much
smaller and lighter. Male, on left, exhibits
a modified cut-crescent tail pattern; female
(35 mm.) a typical upper cut-crescent pat-
tern. See Table V, Table VI, nos. 65, 77.

Fig. 8. Hybrid X. couchianus X X. v. variatus.

Female, above, exhibits spotted (Sp) pat-
tern from variatus. See Table VI, no. 53.

Plate IV

Fig. 9. Hybrid X. montezumae cortezi X X. v.

variatus. Female, on right, exhibits spotted
(Sp) pattern from variatus. See Table VI,
nos. 57, 97.

Fig. 10. X. m. cortezi X X. v. xiphidium. Pi male
(25 mm.), upper right, exhibits the spot-
ted (Sp) pattern. Pi female (35 mm.),
upper left, exhibits no macromelanophore
pattern. Fi, lower left, exhibits strongly
enhanced spotted pattern from xiphidium;
Fi lower right, exhibits no macromelano-
phore pattern. See Table VI, no. 69.

Plate V

Fig. 11. Hybrid X. v. variatus X X. p. pygmaeus.

Male (25 mm.), below, exhibits spotted
(Sp) pattern from variatus. As this fish
grew older, the spots increased several
times in size and new ones appeared in ab-
normal locations. See Table VI, no. 58.

Fig. 12. Hybrid X. p. pygmaeus X X. v. xiphidium.

Female, on left above, exhibits strongly
enhanced spotted (Sp) pattern from
xiphidium. See Table VI, no. 72.

Plate VI

Fig. 13. X. m. cortezi X X. hellerii strigatus. Pi
female, on left below, exhibits spotted
(Sp) pattern. Fi male, on right above,
exhibits same cortezi pattern but with re-
duced expressivity. See Table VI, no. 88.
Note prominent reticulum of cortezi. (h9).

Fig. 14. Backcross offspring of h9 to female X. h.

strigatus. Female (40 mm.), on right
above, exhibits spotted (Sp) pattern from
cortezi with reduced expressivity. See
Table VI, no. 89.

1962]

Atz: Effects of Hybridization on Pigmentation in Xiphophorus

181

Plate VII

Fig. 15. X. m. cortezi X X. h. strigatus. Pi female,
on right, middle, exhibits spotted caudal
(Sc) pattern, which Fi female (40 mm.),
above, exhibits in an enhanced form. Pi
male, on left, below. See Table VI, no. 99.
(h26).

Fig. 16. Second backcross offspring of h26 to male
X. h. strigatus. Female (50 mm.) exhibits
melanotic caudal fin and peduncle with an
overgrowth. Note slashes of pigment in
dorsal fin and several melanophore clusters
on body, at least one of which is located
anterior to the dorsal fin. See Table VI,
no. 103.

Fig. 17. Cleared skin from side of cortezi-hellerii
Sc hybrid (353) showing macromelano-
phore cluster and faint mid-lateral stripe.
Anterior to the right. About 151A X-

Fig. 18. Cleared skin from side of wild male X. v.

variants showing reticulum, spotted pat-
tern and vertical bar (at center). Anterior
to the right. About 14% X.

Plate VIII

Fig. 19. X. h. guentheri from the Belize River in
British Honduras, female above, male be-
low. Both fish exhibit the spotted (Sp)
pattern.

Fig. 20. Hybrid X. h. guentheri X X. maculatus.

Fish on the right exhibits spotted (Sp)
pattern from guentheri. Fish (50 mm.),
on the left, exhibits a peculiar, undiag-
nosed atypical growth not associated with
Sp. See Table VI, no. 119. (320).

Fig. 21. Backcross offspring of 320 to male X.

maculatus. Male exhibits spotted (Sp) pat-
tern from guentheri and nearly completely
suppressed striped (Sr) pattern from
maculatus. See Table VI, nos. 40, 127.

Fig. 22. Inbred, second backcross offspring of 320
to female X. maculatus. Female (28 mm.),
exhibits spotted (Sp) pattern from guen-
theri and striped (Sr) pattern from macu-
latus. See Table VI, nos. 52, 126.

ATZ

PLATE I

FIG. 1

FIG. 3

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

ATZ

PLATE II

FIG. 5

FIG. 4

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

FIG. 6

ATZ

PLATE III

FIG. 7

FIG. 8

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

ATZ

PLATE IV

FIG. 9

FIG. 10

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

ATZ

PLATE V

FIG. 11

FIG. 12

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

ATZ

PLATE VI

FIG. 14

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

ATZ

PLATE VII

FIG. 17

FIG. 18

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

ATZ

PLATE VIII

FIG. 19

FIG. 20

FIG. 21

FIG. 22

EFFECTS OF HYBRIDIZATION ON PIGMENTATION IN FISHES
OF THE GENUS XIPHOPHORUS

15

A Field Study of the Golden-headed Manakin, Pipra
erythrocephala, in Trinidad, W. I.1'2

D. W. Snow

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

(Text-figures 1-6)

[This paper is one of a series emanating from the
Tropical Field Station of the New York Zoological
Society, at Simla, Arima Valley, Trinidad, West
Indies. This station was founded in 1950 by the Zoo-
logical Society’s Department of Tropical Research,
under the direction of Dr. William Beebe. It com-
prises 200 acres in the middle of the Northern
Range, which includes large stretches of undisturbed
government forest preserves. 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, and the annual rainfall is
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, 37 (13): 157-184.]

Introduction

THIS paper is to some extent complemen-
tary to a previous paper on the Black and
White Manakin, Manacus manacus
(Snow, 1962). Both species were studied at the
same time, largely by the same methods, and in
the same locality, an area of primary and sec-
ondary forest in the center of the Northern
Range of Trinidad.

About 3 Vi inches long, the Golden-headed is
a smaller and lighter bird than the Black and
White Manakin. Its bodily proportions are dif-
ferent; its wings are relatively and absolutely
longer, and its legs and tail shorter. Correspond-
ingly, its flight is more rapid and agile, and its
displays depend more on flight and wing actions

aContribution No. 1020, Department of Tropical Re-
search, New York Zoological Society.

2This study has been supported by National Science
Foundation Grants G 4385 and G 21007.

and less on jumping. The male is all black, with
a golden-orange cap and red and white thigh
feathers; the female is olive-green. The eye is
white in the male and the beak pale brown or
straw-colored; in the female the eye is variable
in color, but darker, and the beak is darker
brown. Like other manakins, Golden-headed
Manakins feed largely on fruit, which they take
on the wing, but insects are also taken. They
bathe in streams, and drink from water collected
in the leaves of trees.

For a field study, the Golden-headed is less
suitable than the Black and White Manakin. Al-
though it is the more abundant species, at least
in Trinidad, most of its activities are carried out
higher above the ground. It feeds on average
higher, it displays up in the trees, at a height of
20 feet or more, and its nests are placed on aver-
age much higher above the ground. Also, its very
abundance made it difficult to band more than
a fraction of the local population. Over 600 in-
dividuals were banded in the study area, but
when trapping ceased new birds were still being
caught far more often than previously-banded
birds. Thus much less detailed information was
obtained for the Golden-headed than for the
Black and White Manakin, especially as regards
nesting and the histories of individual birds.

For details of the environment, the reader is
referred to the earlier paper. Here it need only be
said that the vegetation is transitional between
seasonal forest and lower montane rain forest
(Beard, 1946); the temperature varies little
throughout the year; annual variation in day-
length is 74 minutes; the rainfall averages about
100 inches per year; there is a main dry season
from January to May, and the wet season lasts
for the rest of the year, being usually interrupted

183

184

Zoologica: New York Zoological Society

[47: 15

by a short spell of dry weather in September or
October.

Apart from Skutch’s study of P. mentalis
(Skutch, 1949), no detailed study has been made
of any species of Pipra. A preliminary account
of the present work, dealing only with display,
has already been published (Snow, 1956). Sick
(1959) has described displays of P. erythroceph-
ala and P. fasciicauda. Niethammer (1956) has
briefly described displays of P. chloromeros; and
Snow has described displays of P. pipra and P.
aureola (Snow, 1961 and in prep.). (I have re-
ferred to “displays” and not “the displays” ad-
visedly, as these studies are certainly not com-
plete; I have always found that manakin species,
when watched intensively over a long period, re-
veal a greater repertory of distinct displays than
is apparent from more cursory observation).
From these studies it is known that the males of
some species of Pipra display communally, and
some solitarily. Of the known species, P. ery-
throcephala is one of those in which communal
display is most highly developed.

Since many comparisons with the Black and
White Manakin are made in the course of this
paper, for brevity I refer to my previous paper
simply as ‘‘Manacus,’’ followed by the relevant
page-number in that paper.

As in the previous paper, I am indebted to
several persons for help with the field work;
especially to my wife, for much help in the mist-
netting and in finding nests; to Dr. W. G. Downs
and Dr. T. FI. G. Aitken, for making available
to me the birds trapped in the course of the work
of the Trinidad Regional Virus Laboratory; and
to Mr. N. Y. Sandwith and Dr. J. J. Wurdack
for many plant identifications. The whole work
was generously supported by National Science
Foundation Grants G 4385 and G 21007.

The Population

The Golden-headed Manakin is probably the
most abundant forest bird in Trinidad. At the
main trapping place, on the edge of secondary
forest, and at two subsidiary trapping places two
and four hundred yards inside the forest, 625
different individuals were caught over a period
of 314 years. When trapping ceased, as already
mentioned, more new individuals were still be-
ing caught than previously-banded birds. In the
last four months of trapping, 89 new birds and
32 previously-banded birds were trapped. Up to
that time, 529 individuals had been caught. Al-
lowing an annual mortality of 10% (see later in
this section), about 440 of these would have been
expected to be still alive. With the proportion of
recaptures to total captures at 32/121 or 26%,

the total population from which the captured
birds were drawn should have been of the order
of 1,700.

This calculation is clearly subject to consid-
erable error, as the ringed and unringed birds
were certainly not randomly mixed within the
population, and it is not known how far individ-
uals range in the course of their daily activities.
Probably most of those caught lived within the
steep-sided forested valley, of about 200 acres,
formed by the stream along which the trapping
places were situated. It may reasonably be sup-
posed that the density of Golden-headed Mana-
kins in this area was certainly more than one
per acre, and probably several times as great.

The Black and White Manakin, probably the
next commonest species of forest bird in the
study area, was found to have an adult density
of just over one bird per acre.

The annual mortality of adults can be esti-
mated approximately from the proportion of
adult males in the population. Of the first 100
birds caught at the main trapping places (feed-
ing and bathing places away from the display
grounds), 45 were adult males. Of the second 100
caught, 44 were adult males. Thereafter the pro-
portion of adult males declined to around 35%
and remained at this figure. This decline was
probably due to the fact that adult males are
more sedentary than females and young birds,
being closely attached to their display grounds,
and many of those whose display grounds were
nearest the trapping places had by then been
caught and were tending to avoid the trapping
area.

Male Golden-headed Manakins moult into
adult plumage at the beginning of their second
year of life. A population in which the sex ratio
is equal (as it is in Manacus and probably is also
in the Golden-headed Manakin) and 44% or
45% of the individuals are adult males, should
have an annual adult mortality of about 10%, a
figure that is very close to that obtained for the
Black and White Manakin by another method
{Manacus, p. 96).

The Display Grounds

Male Golden-headed Manakins display com-
munally, but in much more diffuse groups than
Black and White Manakins. Each male occu-
pies a particular perch, 20 to 40 feet above the
ground. The perch selected is usually a more or
less horizontal twig under the canopy of a lower
story tree, but sometimes a horizontal stretch of
vine is used. To be suitable, a display perch must
be roughly straight and without side branches,
and not too thick (5-10 mm.). It must have a

1962]

Snow: Field Study of Golden-headed Manalcin in Trinidad

185

clear approach from one side at least, to allow
the displaying bird to make the display-flight

(p. 186).

The males sometimes peck at excrescences on
their display perches, and if there are leaves or
tendrils of vine hanging near them they habitu-
ally pull at them while hovering or hang onto
them. As a result the leaves near a display perch
are often tattered. Thus Golden-headed Mana-
kins show in some degree the tendency to clear
the display area which is so marked in the Black
and White Manakins.

The display perches of different males may be
several feet apart in the same tree, or several
yards apart in neighboring trees. The number
of males at a display ground is variable; those
that were studied usually had from six to twelve
males.

As in Manacus (p. 69), display grounds were
found to persist year after year in the same
places, and the same display perches were used
in successive years. Because it is much less ob-
vious that individual perches are in constant
use than it is in Manacus, there being no equiva-
lent of the court, the persistent nature of the dis-
play ground is not apparent to the observer un-
less he is able to watch an area over a long
period. Probably for this reason, Sick (1959)
gives the impression that in Brazil the Golden-
headed Manakins display randomly in the for-
est. In Trinidad, random display could indeed
be seen at times all over the forest, mainly from
juvenile males, but these birds doubtless even-
tually settled at one of the permanent display
grounds.

Elements of the Display

The Golden-headed Manakin’s display move-
ments and postures are highly stereotyped and
very diverse. Some are much more often seen
than others, and it was not until I had watched
the birds for over a year that I saw some of the
less common displays. After years of inter-
mittent observation, with regular weekly watches
for the last 13 months of this period, I was con-
fident that I had seen the full range of normal
display given by the Trinidad population of the
species. As mentioned later (p. 187), there is
some evidence that the display differs markedly
in the Brazilian population.

As already mentioned, compared with the
Black and White Manakin the Golden-headed
has relatively long, pointed wings, and short legs.
It is extremely agile in flight, and specialized
flights are important in its display. It does not
hop or jump, but moves sideways or backwards
along the perch with very short quick steps that

give the bird the appearance of sliding. In rest-
ing posture, it sits hunched, with body-feathers
fluffed and legs concealed. When it begins to dis-
play its appearance changes markedly; it sleeks
its plumage and stretches its legs, so that the red
and white thigh feathers appear.

In view of the male’s extreme agility in flight,
it is of interest that he averages a little lighter and
smaller than the female (Appendix). In Manacus,
in which display depends on specialized wing-
feathers and hypertrophy of the muscles of the
shoulder region and thighs, the males are heavier
than the females and their wings are shorter.

In the following sections, I have for conveni-
ence used the same names as Skutch (1949) used,
for display movements that are clearly homolog-
ous with those of Pipra mentalis.

Advertising Calls

Males sitting inactive on their display perches
frequently utter advertising calls, which can be
arranged in a series of increasing excitement
leading up to active display. At lowest intensity
they utter an occasional clear “pu.” When ex-
citement is a little increased this is followed by
a trill and usually a sharp final note: “pu prrrr-
rrr-pt.” With mounting excitement the final
note is repeated two or three times, the initial
“pu” is replaced by one or two sharper notes,
and the trill may be lengthened, so that the call
becomes “pir pir prrrrrrrrr-pt-pt.” The first note
of the call may be accompanied by an upward
flick of the wings. At this stage the bird is likely
to break into more active display.

The other calls are characteristically associ-
ated with particular display movements, and will
be described in the following sections.

Darting Back and Forth

With the legs stretched so that the colored
thigh feathers show, and the body held rather
horizontally, the bird makes rapid flights to and
fro between its main perch and an adjacent perch
usually 3-5 feet away. When it has landed it at
once turns rapidly about to face the way it came
from. In flight between the two perches, the
wings make a brisk humming sound audible for
about 20 yards. Between bouts of darting back
and forth, a very short, sharp “zit, zit” is often
uttered.

The About-face

Rapid about-faces, such as are made between
darts back and forth, may be repeated while the
bird remains perched in the same place. Each
turn may be accompanied by an upward flick
of the wings.

The Backward-slide

With the legs stretched, the head held low and

186

Zoologica: New York Zoological Society

the tail elevated, the bird “slides” backwards
along its perch for several inches (Text-fig. 1, b
and d). The line of the body is usually at an angle
of about 45° to the line of the perch, but this
varies; sometimes the body is held almost in line
with the perch, and occasionally nearly at right
angles to it. At the end of the slide the tail is
suddenly depressed' and fanned, and the wings
raised and held vertically above the back (Text-
fig. 1, a and e). Half way through the slide, the
wings may be suddenly spread horizontally
and then closed again (Text-fig. 1, c and f); this
movement is often omitted. Very seldom, after
sliding one way the bird may turn and slide back
in the other direction. This display is silent.

The Display-flight

The male leaves his perch and flies with nor-
mal flight to a somewhat higher perch about
20-30 yards away. He perches and faces his dis-
play perch, utters usually two or three sharp
calls, “kew, kew,” then takes off and returns to
the display perch with extremely rapid flight,
uttering a succession of “kew” calls which speed
up and become sharper as he nears it. Usually a
total of 7-9 “kew” calls are uttered (including
those made before taking off). The bird ap-

[47: 15

proaches his display perch with a rapid down-
ward, then upward, then downward swoop, so
that its trajectory is a shallow S-curve. As it
lands it utters a sharp buzzing call. Thus the
listener hears an accelerating series of “kew”
calls followed by a sharp buzz, a sequence of
sounds that is diagnostic of this display. In this
way the display-flights of a group of males can
easily be counted even while the observer is
watching only one or two of them, as was done
during the all-day watch at a display ground
described later (p. 190).

Immediately on landing, the bird usually exe-
cutes a backward-slide. At lesser intensities the
slide may not be made, and sometimes the bird
lands without uttering the buzzing call. These
incomplete displays usually occur if for some
reason the bird is put off at the last moment, or
apparently misjudges its landing on the perch.

The Frenzied Flutter

This term is used to cover some rather varia-
ble movements, having in common that the bird
jumps, or makes as if to jump, upwards, and
utters an excited “zeek” or “zeek-eek,” reminis-
cent of the buzz made on landing from the dis-
play-flight but higher pitched. The bird may

e f

Text-fig. 1. The backward-slide. Arrows show the direction of the bird’s movement, b and d: posture
while sliding, c and f: wings spread half way through slide, b and e: wings raised and tail fanned at end
of slide, ct-d shows normal orientation with respect to perch; e and f show less usual orientation trans-
versely to perch. (Drawn from movie film).

1962]

Snow: Field Study of Golden-headed Manakin in Trinidad

187

jump up or slightly backwards a few inches, with
wings conspicuously fluttering, and land back on
the same spot; or it may jump up and land a foot
or more away with a rapid upward and down-
ward trajectory and a sudden alteration of course
in mid-flight; or it may flutter along the perch,
moving forwards (not backwards, as when slid-
ing) ; or it may jump up and flutter, rapidly twiz-
zling from side to side before it lands; or it may
simply crouch and flutter its wings violently with-
out leaving the perch. While fluttering it tends to
make downward pecking movements which ap-
pear to be directed at the perch.

It will be apparent from the foregoing descrip-
tion that this display has much in common with
copulation. It is given at times of great excite-
ment and may be thought of as copulation in
vacuo.

The Upright Posture

When another bird lands on or near a male’s
display perch, the owner often flies and lands
close beside it in a stiff, almost statuesque up-
right posture (Text-fig. 2). He may not fly
straight to the intruding bird, but may make a
short circling flight of a few yards away and
back, a flight which is sufficiently stereotyped to
be regarded as a preliminary part of this display.
Just after landing he utters a very sharp, short
“zick.” The upright posture may be held for
several seconds, the bird clinging to the side of
the perch rather than perching on it, with the
head pointing upwards and the beak often slight-
ly opened. At close quarters it is possible to see
that the pupil is greatly contracted while this
posture is held. In this display the aggressive
element is probably more dominant than in the
displays described above. It is often the initial
response to an intruder of either sex, and in sev-
eral cases when it was seen to be made to birds
in female plumage there was reason to think
that they were young males.

Wing-flicking

General excitement may be shown by quick
upward flicks of the wings. Wing-flicks may be
interpolated between pivoting, frenzied flutters,
and other display movements, or they may some-
times be repeated rapidly by a perched bird and
accompanied by the call “pir pir prrrrrrr-pt” or
the more excited “zeek-eek.” Repeated wing-
flicking, with calling, appears to constitute an
independent display, but one that is less ritual-
ized than those described above.

Sick’s Observations in Brazil

Sick’s account of the displays of Pipra erythro-
cephala in Brazil (Sick, 1959: 275-277) differs
in several respects from that given here. The

Brazilian form (subspecies rubrocapilla ) is very
distinct from the Trinidad form, which is close to
the nominate subspecies: the head is red, not
golden-yellow, and there is a conspicuous white
patch on the under wing-coverts, which is lacking
in the Trinidad bird. For the Brazilian bird, Sick
does not mention “darting back and forth,” the
“frenzied flutter” or the “upright posture”; nor
does he describe the full “backward-slide” fol-
lowing the “display-flight”, which is perhaps the
most conspicuous of all the display sequences of
the Trinidad birds. He mentions the backward-
slide briefly, as being accompanied by quivering
of the slightly raised wings, but not the spreading
of the wings half way through the slide nor the
raising of the wings and fanning of the tail at
the end of the slide. He describes the display-
flight as silent, whereas in Trinidad, although
the flight is unaccompanied by mechanical
sounds, the characteristic calls associated with
it make it far from silent. He says that wing-
flicking on the perch is sometimes accompanied
by a snapping sound, which I never heard. Fin-
ally, he describes how the white patch on the
under-side of the wing is exhibited in display,
by a posture to which there is no parallel in the
Trinidad bird.

Text-fig. 2. The upright posture. (Drawn from
movie film).

Further observations may show that not all
these differences are real ones; even so, it seems
clear that we have in this species an outstanding
example of subspecific differences in display
which would repay detailed study. It would be
desirable to have a motion picture record of the
displays of the Brazilian birds for comparison
with the Trinidad population.

Relationships between Males

As in Manacus (p. 79), neighboring males
usually show little overt aggressiveness towards
each other, but they are in fact potentially ag-
gressive. Hostility increases when a female visits
the display ground, and it was on such occasions
that the few flight chases between adult males
were seen. Each male defends its perch against

188

Zoologica: New York Zoological Society

[47: 15

intruders mainly by means of displays (espe-
cially landing beside the intruder in the upright
posture, backward-slides, and wing-raising) , but
fights were never observed, as they were sev-
eral times in Manacus, nor was there any evi-
dence of one male trying to usurp another’s dis-
play perch. Presumably Golden-headed Mana-
kins can find new display perches more easily
than Black and White Manakins can clear new
courts, so that serious competition for display
perches will be rare.

Neighboring males spend much time sitting
quietly side by side on some perch usually be-
tween their two display perches. The most
striking feature of this behavior is that both
birds constantly face away from each other.
Sometimes one bird shows its aggressiveness by
an occasional backward-slide towards the other,
or by suddenly raising its wings and fanning its
tail (as at the end of the full backward-slide),
and in such cases it may be seen that of the two
birds the subordinate one has its head more com-
pletely averted than the displaying bird.

As in Manacus (p. 80), which has similar be-
havior, birds sitting with a neighbor spend more
time at the display ground than those that are
alone. This is especially noticeable when dis-
play is slack. Frequently they leave the display
ground together and return together, which sug-
gests that they associate together even when
feeding. During an all-day watch on April 29,
1961 (p. 190), the male that was being watched
continuously (Male A) sat with his nearest
neighbor (Male B, whose display perch was only
a few feet away in the same tree) for a total
of 3 hours 44 minutes, in 48 spells, most of which
were from 2 to 6 minutes long. Usually they
sat together between their display perches, but
sometimes on another perch several yards away
in another tree. Male A also several times sat
with another male several yards away, and Male
B sat with two other males.

More than two males never sit together at
the same time. This is a matter of common ob-
servation, but it was once seen especially clearly,
when Male B was sitting with another male on
a perch where he sometimes sat with Male A.
Male A was sitting on a lower perch by himself
about 30 feet away. The male with B flew off, and
Male A, who appeared to have been waiting for
this to happen, at once flew up and sat by Male
B.

During the moult, when some males are ab-
sent, those that remain visit the unoccupied dis-
play perches and keep less strictly to their own,
exactly as was found in Manacus (p. 79). Thus
although there seem to be numerous suitable

display perches in the area of the display ground,
those that are actually “owned” by a male have
a special significance for other males. It seems
most likely that it is the mere fact that they have
been occupied by another male that gives them
this significance, rather than that they are es-
pecially suitable in some way not apparent to
the human observer.

At the display ground that was under regular
observation for a year, the display perches of six
males were in view from the hide. Of the males
that eventually occupied them after the moult,
one was unbanded, one was banded with a num-
bered and a colored ring, and four were banded
with numbered rings only and so were not in-
dividually distinguishable in the field. (The main
trapping area, where over 600 individuals were
caught, was about 200 yards away from this
display ground.) Except for the unlikely possi-
bility that the similarly banded males changed
places with each other without any disturbance
in their behavior being noted, these individuals
kept to the same display perches throughout the
greater part of the period of observation. The
only observed changes took place as the time
of moult approached; in June the color-banded
male moved and displaced one of the banded
males from a neighboring display perch, with-
out however completely abandoning his former
perch. The displaced male was not seen again,
and may in fact have begun to moult (the earli-
est birds begin in July, Text-fig. 6). In July
two more males disappeared and presumably
began to moult. The color-banded male began
to display occasionally at the perch of one of
them, so that he now displayed at three perches.
The display perch of the other male that had
disappeared (Male A) was occupied by Male B,
his close neighbor. The three males that re-
mained were present throughout August, but in
early September they all disappeared.

Text-figure 5 shows that a little display was
recorded in September at this display ground.
It was due entirely to a few newly arrived birds,
probably recently moulted young males (p. 194) .
They did not use the display perches of the old
males, but displayed on various perches round
the periphery of the display ground.

Behavior of Juvenile Males

On many occasions groups of two or three
birds in female plumage were seen away from
display grounds performing uncoordinated dis-
play in a similar way to that described for juv-
enile male Manacus (p. 80). By analogy with
Manacus they were assumed to be juvenile
males, but this was not proved. Occasionally
they were accompanied by adult males. All the

1962]

Snow: Field Study of Golden-headed Manakin in Trinidad

189

Text-fig. 3. Diurnal rhythm of display, April 29, 1961. Intensity of display measured by number of
display-flights in each 5-minute period (vertical lines).

display movements were made, but the display-
flight was not usually followed by the backward-
slide; displays tended to be incomplete and un-
coordinated. In these groups presumed young
male Golden-headed Manakins were occasion-
ally seen associating with female-plumaged
Black and White Manakins, and on one occasion
a Black and White Manakin known from its
later history to have been a juvenile male was
seen displaying quite persistently to a Golden-
headed Manakin. Parkes (1961) has discussed
these observations in relation to the occurrence
of occasional hybrids in the Pipridae.

Courtship and Copulation

I was usually unable to be certain of the sex
of the female-plumaged birds that I saw visiting
the display grounds. The behavior of these birds
seemed usually to be more ambivalent than that
of the female-plumaged Manacus (p. 77),
which were normally easy to sex after they had
been watched for a little in the field. Eye-color
sometimes provided a clue (p. 196), but was
not usually observable accurately enough from
a distance of several yards. Few color-banded
birds were seen visiting the display grounds.
Information on the behavior of males towards
females is thus far less satisfactory than for
Manacus.

I was unable to find many distinctions be-
tween displays given to birds presumed to be
females not ready for copulation and those given
to birds presumed to be juvenile males. In both
cases the owning male and the visiting bird are
both aggressive, and the situation is usually con-
fused. The owning male will land beside the vis-
itor in the upright posture, and may then slide
backwards towards it. The other may fly away
and return, also landing in the upright posture.
Both birds may dart back and forth, perform

about-faces, flick the wings, or turn away from
the other and raise the wings and fan the tail.

But when the visitor was a (presumed) recep-
tive female, there tended to be a more orderly
sequence of display. Typically the males would
start by repeated darting back and forth, utter-
ing the thin “zit, zit,” and would then change
to display-flights. If the female went to a
male’s display perch, he would continue his
display-flights, landing close beside her, inter-
spersing them with the upright posture, back-
ward-slides, and other elements of his display.

Copulation was seen only twice, and at-
tempted copulation once. On April 26 a female
came to the display perch of the male that was
under observation and perched there in normal
posture. He flew away and came back in dis-
play-flight, landing immediately beside her. She
remained still and he mounted. A little later the
female returned to the same perch. The male
flew in again in display-flight and landed beside
her, but she then flew away. On April 29, during
the all-day watch, a banded female flew to Male
A’s perch. He flew towards her from about two
feet away, landed beside her in the upright pos-
ture, and mounted. On August 29 Male B, in
the course of confused mutual displays of the
sort described earlier in this section, attempted
to mount a female-plumaged bird that was very
aggressive and was thought to be a juvenile male.
On this occasion mounting was not preceded by
a display-flight.

Very many times males were seen to fly in
with display-flight and land close beside a fe-
male, or presumed female, that had come to
their display perches. Normally the female at
once flew away to another perch. Flights to the
female from a short distance, followed by the
upright posture, were less often seen. Probably
therefore the copulation sequence seen on April
26 is the most usual.

190

Zoologica: New York Zoological Society

[47: 15

In Pipra mentalis, Skutch (1949) has de-
scribed how the male lands straight on the back
of the female after performing a circling flight.
Although it is not easy to be sure of the details
of the male Golden-headed Manakin’s extreme-
ly rapid display movements, I had no evidence
that they attempted to land straight on the fe-
male in this way.

The Daily Cycle of Display

An all-day watch was carried out on April
29, 1961, at the display ground that was under
regular observation. All display-flights were
counted from dawn to the time that display
ended. From Text-fig. 3, which gives the results
of this watch, it will be seen that after an early-
morning outburst of display, activity fluctuated
rather irregularly until 1545, when display
ceased for the day. There was an especially in-
tense bout of display between 1400 and 1500.
The early-morning watches that were carried
out weekly for over a year at this display ground
showed that the time when the early outburst
of display occurred was not constant. On some
mornings there was little display between 0600
and 0700, and the main outburst was between
0700 and 0800. Later in the day, too, the times
when display was most active were variable. In
this respect the Golden-headed Manakin shows
a less well-marked daily rhythm than Manacus
(P- 75).

Text-fig. 3 also shows the times when females,
or presumed females, visited the display ground.
As would be expected, the presence of a female
coincided with outbursts of display. Because
of the difficulty of seeing the cryptically colored
females unless they move, it was not easy to tell
whether it was the arrival of a female that stim-
ulated the males to display, or an outburst of
display that attracted the females. It seems, how-
ever, that it is usually the arrival of a female
that stimulates the males. Nevertheless outbursts
of display certainly occur without any females
being present, because, as in Manacus, the males
at a display ground have a mutually stimulating
effect on one another, so that when one starts
to display others often follow and a general out-
burst results.

During the all-day watch, a record was kept
of the periods of presence and absence of a male
that had a display perch near the hide (Male A,
mentioned above). This bird had occupied the
same perch for at least 8 months; his habits and
alternative perches were well known. His three
nearest neighbors were all banded, but he was
not. Between 0548, when he arrived, and 1700,
when he was still present but display had long
since ceased and the light was too bad for further

observation, he was present on or near his dis-
play perch for 88% of the time. His 43 absences
were mostly very short, 29 of them being from
Vi to l Vi minutes. It is probable that some of
them were due to his merely flying for a short
time to a perch that was out of sight. During
one of the longer absences, of 5 minutes from
1612 to 1617, he bathed, as he returned with
wet plumage and preened on a perch near his
display perch. Excluding this absence, the time
when he was away and could have been feeding
was only 11% of the daylight hours.

An attempt was made to watch another male
at the same time (Male B). His absences were
of the same sort of length as Male A’s, and they
were often synchronized, as they occupied
neighboring display perches, but it was not possi-
ble to obtain a complete record for this bird for
the whole day. Less prolonged watching on
many occasions gave no reason for supposing
that the behavior of the males under observa-
tion on this day was in any way unusual.

The Annual Cycle of Display

Male Golden-headed Manakins display at all
times of the year, except when they are moult-
ing, but the intensity of display fluctuates some-
what throughout the season. During the regular
early-morning watches at the main display
ground under observation from August, I960,
to September, 1961, the intensity of display
was measured by counting all the display-flights,
a procedure which is rather easy as the display-
flight is accompanied by a quite characteristic
series of calls (p. 186). Text-figure 4 shows the
intensity of display throughout a year, as meas-
ured during these watches. This figure also shows
the season of moult in the two years 1960 and
1961, and the breeding season as far as it was
ascertained, and thus provides a conspectus of
the annual cycle for just over a complete year.
It will be seen that display stopped during the
moult in 1960, but not in 1961, when it was
merely reduced. As mentioned above, the dis-
play in September, 1961, was not from the old
established males, who had all begun to moult
by then, but from newly arrived, probably re-
cently moulted young males.

A more marked break in display occurred in
December, 1960, and January, 1961, at exactly
the same time as a break occurred in the display
of Manacus (p. 82). Though there was evidence
that food was temporarily short in the Arima
Valley, where the observations were made,
Golden -headed Manakins ceased displaying
equally completely for the same period in an
area of forest several miles away on the north
side of the Northern Range, where food was

1962]

Snow: Field Study of Golden-headed Manakin in Trinidad

191

I960 1961

Text-fig. 4. Annual cycle of display, breeding and moult, August, 1960, to September, 1961. Uppei
graph: intensity of display during early morning watches, based on number of display-flights per hour,
and breeding season (horizontal line). Lower graph: moulting seasons, showing percentage of trapped
adults undergoing wing-moult in each month.

abundant. The reason for the cessation of dis-
play was therefore not clear. Regular observa-
tions were not made all through December and
January in other years, so breaks in display dur-
ing that period could have gone undetected, but
regular observations at a display ground of Man-
acus showed no comparable break in display in
other years.

Breeding

The Golden-headed Manakin ’s nest is typical
of the family, a small, shallow cup slung in the
fork between two horizontal twigs of a sapling
or shrub, or the lower branches of a tree. It is
thinly woven of brownish fibers and rootlets, and
often has a few dead leaves bound into the bot-
tom of the cup or hanging from the underside.
The cup is usually so thin that the contents are
partly visible from below. Those that were seen
lacked the lining, made of the branching pani-
cles of the melastomaceous herb Nepsera aqua-
tica, that was found to be so characteristic of the
nest of Manacus (p. 89).

Eight of the 15 nests that were found were
between 4 and 8 feet above the ground. The
other seven were from 10 to 35 feet. Probably

a much higher proportion of the nests are high
up than these figures indicate, since high nests
are much less easy to find. Since the low nests
are not much less easy to see than the nests of
Manacus, there seems no other way to account
for the fact that only 15 were found compared
with over 300 of Manacus, although the Golden-
headed Manakin is the more abundant of the two
species.

For 11 of the 15 nests, which contained eggs
or young, the months of laying were as follows:
March, 1; June, 4; July, 5; August, 1. The
months of laying of the two nests recorded by
Belcher & Smooker (1937) were January, and
late March or early April. Four nests were
found while being built, and so far as known
no eggs were laid in them or, if laid, they were
lost soon after; all these were in April and May.
In addition, birds with incubation patches were
trapped in May and July, and an egg-laying bird
in March. It seems then that the breeding season
extends from January to August, and that most
clutches are laid in the latter half of this period
—a breeding season substantially the same as that
of Manacus (p. 82).

192

Zoologica: New York Zoological Society

[47: 15

Few details of breeding were obtained, due to
the high rate of nest failure. Four of the com-
pleted clutches consisted of two eggs, and one
of one egg. The interval between the laying of
the eggs was ascertained at two nests to be two
days. One egg was found to have been laid be-
tween 1145 and 1545 hours. At one nest the
incubation period was found to be 16 or 17 days,
the fledging period was not ascertained, since
none of the seven nests found before the time
of hatching survived to the time of fledging.

Feeding Habits

Golden-headed Manakins feed in much the
same way as Black and White Manakins; they
pluck fruit on the wing, swallowing it whole on
landing, and pick insects and spiders in flight
from the twigs and foliage. The most obvious
difference between the two species is that on
average Golden-headed Manakins feed higher
above the ground. Another probable difference
is that they take relatively more insect food.
Thus male Golden-headed Manakins at their
display grounds regularly dart off for short
sallies and pick food from the underside of
leaves and from twigs, but Black and White
Manakin males were never seen to do this. Prob-
ably there is more suitable insect food in the
trees at the level at which Golden-headed Mana-
kins display than near the ground where the
Black and White Manakins have their courts,
and probably also Golden-headed Manakins are
the more fitted of the two species for this meth-
od of feeding, as they are far more agile on the
wing.

Stomach contents of several birds were ex-
amined, but as most of them were trapped near
places where fruit was abundant they yielded

no valid information on this point. The stomach
of a bird accidentally killed away from a trap-
ping place contained remains of a spider 13
mm. long, a small beetle, a dipteran, an uniden-
tified winged insect, an insect larva, and a few
seeds and the skin of the berry of a melastom-
aceous tree.

Golden-headed Manakins are a little smaller
than Black and White Manakins, and it seems
that they cannot take such large fruits. Of the
three largest fruits found to be eaten by Maria -
cus (p. 92), 15-19 mm. long and 10-16 mm. in
diameter, Golden-headed Manakins were seen
to take only one ( Protium heptaphyllum, 15 by
11 mm.).

Altogether, 445 records were obtained of
Golden-headed Manakins feeding on the fruits
of 43 species of plants (Table I). All but two
of the plants were identified, at least down to
generic level. All the fruits taken were small
berries which were swallowed whole, with one
exception: a bird was once seen picking pieces
in flight from the large fruit of a species of
Henriettea (Melastomaceae) .

As for Manacus (p. 93), the Melastomaceae
was found to be by far the most important fam-
ily of food plants: 11 species were seen to be
fed on, accounting for 63% of all the records.
In number of species the Rubiaceae, with 6
species, came next, but very few records were
obtained for each, and their total was only 2%
of the whole. For Manacus, 14% of all records
were from Rubiaceae. The difference is prob-
ably due in part to the higher level at which
Golden-headed Manakins feed; the Rubiaceae
with small fruit, likely to be most attractive to
Golden-headed Manakins, (especially Palicou-
rea, Cephaelis, Gonzalagunia and Psychotria),

Table I. Fruits Eaten by Golden-headed Manakins in the Different Months

Numbers of records in different months

I 1 II j III

IV

V

VI

VII

VIII I IX | X

XI | XII

Moraceae
Ficus citrifolia
Ficus clusiifolia

1

4 5

3

5

1

3

7

Urticaceae
Trema micrantha

Nyctaginaceae
Pisonia eggersiana

Lauraceae
Ocotea canaliculata
Phoebe elongata

Burseraceae
Protium heptaphyllum

1962]

Snow: Field Study of Golden-headed Manakin in Trinidad

193

Table I. Fruits Eaten by Golden-headed Man akins in the Different Months (continued)

Species

Numbers of records in different months

II

in

IV

V

VI

VII

VIII

IX

X

X! [

XII

Euphorbiaceae
Alchornea triplinervia
Hieronyma caribaea
Maprounea guianensis
Sapium aucuparium

Aquifeliaceae
Ilex sp.

Tiliaceae
Sloanea laurifolia
Sloanea stipitata

Dilleniaceae
Davilla aspera
Doliocarpus dentatus
Pinzona calineoides

Flacourtiaceae
Laetia procera

Myrtaceae
Myrcia leptoclada

Melastomaceae
Henriettea sp.

Miconia affinis
Miconia chrysophylla
Miconia guianensis
Miconia kappleri
Miconia multispicata
Miconia myriantha
Miconia prasina
Miconia punctata
Miconia splendens
Miconia tomentosa
Miconia sp.

Araliaceae

Didymopanax morototoni

Verbenaceae
Aegiphila integrifolia

Phytolaccaceae
Phytolacca icosandra

Solanaceae
Cestrum baenitzii

Rubiaceae
Cephaelis muscosa
Isertia parviflora
Malanea macrophylla
Palicourea crocea
Psychotria inundata
Psychotria marginata

COMPOSITAE
Wulffia baccata
Unidentified vine ( 1 )
Unidentified vine (2)

2

4

4

2

11

5

1

5

3

1

5 4

1

5 1

2

2

18 8 1
4

2 10

3

10 7 12 1

6 1

4

7 3

1

1 20 4 25 11

2

2 4

1

1 8
4 23 36

3 24

4

3 6

13 32

5

1

2

1

1

1

3

1

1

1

2

Number of species

4569 12 7553698

194

Zoologica: New York Zoological Society

[47: 15

are shrubs or low trees, while the larger rubia-
ceous trees and vines which are at the level
where Golden-headed Manakins mainly feed
( Amaioua , Coussarea, Lacistema, Malanea)
have larger fruits which are at or near the upper
limit of size for Golden-headed Manakins.

Of the other families, the Euphorbiaceae,
Moraceae and Araliaceae were the most im-
portant. Four species of Euphorbiaceae con-
tributed 6% or all the records, and two species of
Ficus (Moraceae) 5%. Apart from some of
the melastomes, the greatest number of records
from a single tree species came from Didymo-
panax morototoni (Araliaceae), which contrib-
uted 10% of the total.

The number of fruit species found to be eaten
in each month varied from 3 in September to 12
in May. Most different kinds of fruit were found
to be eaten in the months April-June and Nov-
ember-December. Though the species involved
are partly different and the total numbers small-
er, this is the same picture as was found for
Manacus (p. 93) and suggests that the availa-

bility of fruit shows a double peak in the year,
of which the first, major peak comes in the mid-
dle of the breeding season.

Plumages, Moults and Color of Soft Parts
Moults

The moults follow the same course as in
Manacus (p. 82). There is a slow post-juvenal
moult in the first year of life, involving the head
and body, lesser wing-coverts and some inner
secondary major coverts. Most of the birds
undergoing this post-juvenal moult were
trapped in the months August-January. The
next moult, to adult plumage, is complete and
takes place in the second summer and autumn,
Thereafter there is an annual moult in the
months August-November.

Fifteen birds were trapped while moulting
from juvenile to adult male plumage. As was
found in Manacus (p. 85), these birds on aver-
age moulted earlier than the adult males (Text-
fig. 5).

Four birds were trapped twice in the course

JULY AUG. SEPT. OCT. NOV.

Text-fig. 5. Moulting seasons of old and young males. Solid line: fully adult males trapped in moult.
Broken line: males trapped while moulting from juvenile to adult plumage.

1962]

Snow: Field Study of Golden-headed Manakin in Trinidad

195

of a single moult. The amount that the moult
of the wing-feathers had progressed between the
two captures showed that the complete moult
must last about 90 days, a figure similar to that
obtained for Manacus (p. 85). Four birds
trapped while moulting in two or three different
years showed slight annual variations in the tim-
ing of their moult. In the population as a whole,
the season of moult did not vary much in the
four years for which data were obtained (Text-
fig. 6), except that it ended earlier in 1958 than
in 1959 and 1960.

Plumages

Female Golden-headed Manakins regularly
have a few male-type feathers in the head and
body plumage, and in those that were trapped
more than once the number and distribution of
these feathers remained much the same from
one year to the next. Juvenile males also often
have a few adult male feathers in the head and
body, so that on the basis of plumage they can-

not be separated from adult females. A small
proportion of the adult males had some female-
type feathers in the plumage. Usually only a few
were scattered here and there, but two adult
males had the plumage of the underparts almost
entirely female, and many female-type feathers
elsewhere in the plumage; and one had all the
feathers of the head and body of partly male and
partly female color. A fourth had a completely
mixed plumage, with male and female-type
feathers intermingled in the head, body and
wing. This bird was trapped four times over a
period of just over two years, and its plumage
remained exactly the same.

Five birds in female plumage and one adult
male had partially albino plumage. In all of
them, the dark pigment was lacking in some of
the flight-feathers, which were white or pale
yellow.

In addition to these variations, the general
body color of birds in female plumage varied

% IN
MOULT

1958

1959

I960

1961

VI VII VIII IX X XI XII

Text-fig. 6. Moulting seasons, 1958-1961. Each graph shows the percentage of trapped adults under-
going wing-moult in each month, based on the following totals: 1958, 122; 1959, 104; 1960, 254; 1961,
124. Gaps in 1959 and 1961 indicate periods when no observations were made. (A single bird moulting in
March, 1961. is not shown).

196

Zoologica: New York Zoological Society

[47: 15

considerably, from the usual olive-green to a
paler, yellower green. Several birds had a pale
crown-patch of the same shape as the adult
male’s golden cap.

Variation in plumage is much greater in the
Golden-headed Manakin than in Manacus, in
which only three slight abnormalities of plum-
age were found in the 275 individuals caught
in the study area. The reason for the difference
is quite obscure, but it may be noted that it
applies also to the color of the iris, beak and
legs, as described in the following sections.
Eye-color

The iris of adult males is white. Only one
exception was found to this rule: the male with
mixed plumage, mentioned above, had its iris
mainly white but bordered with gray-brown
around the pupil. In birds of female plumage
the iris is variable, ranging from gray-brown or
seal-brown (as in juveniles) to white through
various intermediates in which it is blotched,
dotted, mottled or clouded with varying amounts
of white.

It has previously been assumed (e.g., Con-
way, 1959) that birds in female plumage with
some white in the eye are juvenile males. How-
ever, analysis of the eye-color of the birds that
were trapped more than once showed that the
situation is far from simple, and in particular,
that birds with intermediate eye-colors are more
often females than juvenile males.

Twenty-seven known females were retrapped
aften an interval of 8 months or more. In 19 of
them, the iris was dark when they were first
trapped. In 13 of these, the iris remained dark
after intervals of 8 months to 2 Vi years; in the
other 6, the iris was spotted or flecked with
white when they were next trapped, after inter-
vals of 8 months, 1 year, 1 year, 2 years, 3 years
and 3 years. Eight known females had varying
amounts of white in the iris when they were
first trapped. In four of them the eye appeared
unchanged when they were next trapped, after
intervals of 9 months to 1 year 8 months; in
two the white spots had enlarged, 1 year 8
months and 2 years later; and in two the white
spots later disappeared. One of these two,
trapped four times, had small white spots which
remained for a year, then disappeared four
months later. The other, trapped three times,
had large white spots when first trapped; five
months later it had small spots, eight months
after that the eye was dark, and 2 years and 4
months after that it again had a partially white
iris. On this last occasion it also had 2 white
flight-feathers in one wing and one white feather
in the other wing; previously its plumage had
been normal.

There must be a greater tendency for white
spots to appear and increase in the course of a
female’s life than to decrease, as they all start
life with dark eyes; but the amount of white
does not usually increase indefinitely. This is
shown not only by the evidence given above, but
also by the fact that no bird that was definitely
known to be female had an all-white eye.
Twenty birds were trapped in female plumage,
with an all-white eye. Three of them were later
found to be males, and eight others probably
were, as they were trapped at the season when
the eye-color of juvenile males changes to white
(May- July). Thus not more than 17, and prob-
ably less than half that number, of the approxi-
mately 300 females that were trapped could
have had white eyes.

Nine males were trapped in juvenile plumage
and then retrapped after moulting to adult male
plumage. Four of them, first trapped in their
first November, December and January (2), all
had dark eyes (as had two juvenile males col-
lected in October). Of the other five, one was
trapped in the April before its moult and had
a light gray, slightly mottled iris; one in May
had the iris gray, clouded with white; and three
in June all had the iris white, one of them being
still rather dull. In addition to these birds that
were trapped before and after their moult to
adult plumage, fifteen young males were caught
while moulting to adult plumage. In 13 of them
the eye was fully white, and in the other two
it was white, faintly clouded with gray.

Thus the eye-color of these males changed to
white in the last few weeks before they moulted,
the intermediate stages being uniform gray, or
clouded or faintly mottled with white. None had
the iris dark with white spots, as in females.

Color of Beak and Legs

The upper mandible is straw-yellow in adult
males and horn-brown in females and young
birds; the lower mandible is yellowish in all
birds, paler in males than females. Within the
sexes there is some individual variation, and in
adult males seasonal variation too was found,
the color of the upper mandible becoming dark-
er at the time of the moult.

The color of the legs varies, from the usual
flesh-pink to dull grayish pink and even occa-
sionally blue-gray. It seemed probable that the
birds with the darkest, grayest legs were mainly
young birds, but this was not proved.

Summary

The Golden-headed Manakin, Pipra erythro-
cephala, was studied in an area of primary and
secondary forest in the Northern Range of

1962]

Snow: Field Study of Golden-headed Manakin in Trinidad

197

Trinidad. On the study area it was the most
abundant forest bird, with an adult density in
excess of one bird per acre. The annual mor-
tality of adults was estimated to be about 10%.

The males display at communal display
ground?, where each occupies a display perch
20-40 feet up in a tree. Display grounds persist
year after year in the same places.

The various display movements are described
and figured. The most conspicuous display con-
sists of a rapid swooping flight to the perch, fol-
lowed by a backward-slide, with associated wing
and tail movements. Copulation follows imme-
diately after a display-flight.

Neighboring males are potentially hostile to
each other, but little overt aggressiveness is
shown. Pairs of males spend much time sitting
side by side mid-way between their display
perches.

The males spend about 90% of the daylight
hours at their display grounds, and are normally
present at them all through the year except when
they are moulting.

The breeding season extends from January
to August. The nest, a thinly woven cup slung
in the fork between two horizontal twigs, is
placed from 4 to 35 or more feet above the
ground. None of the nests found survived to
fledging.

The food consists of berries and small inver-
tebrates, taken on the wing. Observations
showed seasonal variations in the amount of
fruit available, with peaks in April-June and
N ovember-December.

Moults and plumages are described. There is
a slow post-juvenal moult in the first year,
followed by a complete moult in June-Septem-
ber. Subsequent annual moults occur a little
later, in July-November.

Literature Cited

Beard, J. S.

1946. The natural vegetation of Trinidad. Clar-
endon Press, Oxford.

Belcher, C., & G. D. Smooker

1937. Birds of the Colony of Trinidad and
Tobago. PartV. Ibis, (14) 1: 225-249.

Conway, W. G.

1959. The Mossy-throated Bellbird— an ornitho-
logical mystery. Animal Kingdom, 62: 66-
76.

Niethammer, G.

1956. Zur Vogelwelt Boliviens (Teil II:
Passeres). Bonn. Zool. Beitr., 7: 84-150.

Parkes, K. C.

1961. Intergeneric hybrids in the family Pip-
ridae. Condor, 63: 345-350.

Sick, H.

1959. Die Balz der Schmuckvogel. J. Om., 100:
269-302.

Skutch, A. F.

1949. Life history of the Yellow-thighed Mana-
kin. Auk, 66: 1-24.

Snow, D. W.

1956. The dance of the manakins. Animal King-
dom, 59: 86-91.

1961. The displays of the manakins Pipra pipra
and Tyranneutes virescens. Ibis, 103a:
110-113.

1962. A field study of the Black and White
Manakin, Manacus manacus, in Trinidad.
Zoologica, 47 (8): 65-104.

APPENDIX

Weights and Measurements
As for Manacus (p. 102), birds were placed in a

cloth bag and weighed immediately after being
trapped. A spring balance accurate to 0.5 gm. was
used.

Weights have been used only for birds of known
sex. A very large number of weights of birds in
female plumage, which were never retrapped and
so could not be sexed, have been discarded.

Similarly wing-lengths
of known sex.

are given only for birds

Wing-lengths

Wing-length

(mm.)

Adult

males

Juvenile

males

Females

62

1

61

4

60

2

5

15

59

13

4

28

58

42

7

14

57

79

4

56

64

1

1

55

30

54

5

1

53

1

52

1

Totals

237

17

68

Means

56.7

58.7

58.9

198 Zoologica: New York Zoological Society [47: 15: 1962]

Weights of Adult Males

Weight

Months

All

(gm.)

I-II

III-IV

V-VI

vii-vm

IX-X

XI-XII

months

17

1

1

16

1

1

15

3

3

2

8

14.5

2

2

4

8

14

1

1

1

5

11

6

25

13.5

4

3

3

11

12

8

41

13

4

5

16

11

13

7

56

12.5

15

10

9

10

9

11

64

12

10

6

11

7

8

9

51

11.5

5

4

6

5

2

7

29

11

2

2

2

6

10.5

1

1

Totals

42

31

48

54

63

53

291

Means

12.5

12.6

12.5

13.0

13.2

12.8

12.79

Weights of Juvenile Males
Weight (gm.)

All months

13.5

2

13

2

12.5

3

12

2

Total

9

Mean

12.7

Weights of Females

Weight

Months

All

(gm.)

I-II

III-IV

V-VI

VII-VIII

IX-X

XI-XII

months

16.5

1

1

16

3

5

3

11

15.5

1

3

8

2

1

15

15

1

5

13

5

5

29

14.5

3

5

3

11

13

3

38

14

9

5

4

16

10

44

13.5

9

3

3

5

9

5

34

13

6

1

3

5

6

5

26

12.5

2

1

2

1

2

4

12

12

1

1

2

4

Totals

30

13

27

53

55

36

214

Means

13.5

14.3

14.4

14.6

13.9

14.0

14.12

16

The Natural History of the Oilbird, Steatornis caripensis,
in Trinidad, W. I. Part 2. Population,

Breeding Ecology and Food1,2,3

D. W. Snow

Department of Tropical Research,

New York Zoological Society, Bronx 60, N. Y.

(Plates I-IV; Text-figures 1-4)

[This paper is one of a series emanating from the
Tropical Field Station of the New York Zoological
Society, at Simla, Arima Valley, Trinidad, West
Indies. This station was founded in 1950 by the
Zoological Society’s Department of Tropical Re-
search, under the direction of Dr. William Beebe.
It comprises 200 acres in the middle of the North-
ern Range, which includes large stretches of undis-
turbed government forest preserves. 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, and the annual
rainfall is more than 100 inches.

[For further ecological details of meteorology and
biotic zones, see “Introduction of the Ecology of the
Arima Valley, Trinidad, B.W.I.,” William Beebe.
Zoologica. 1952, 37 (13: 157-184.]

The Trinidad Population

THERE are at present eight known Oil-
bird colonies in Trinidad, in caves rang-
ing from sea level to about 2,500 feet
(Text-fig. 1, Table I). The four best known
have been the subject of numerous accounts,
ornithological and popular ( e.g ., Kingsley, 1871
(Huevos); Chapman, 1894 (Huevos); Cherrie,
1907 (Aripo “Main Cave”); Roosevelt, 1917
(Oropouche); Williams, 1922 (Arima, Oro-
pouche); Hollister, 1926 (Arima, Huevos);
Myers, 1935 (Arima); Sanderson, 1940 (Aripo
“Main Cave”). I have found no published ac-
count of La Vache cave and the smaller Aripo

Contribution No. 1021, Department of Tropical Re-
search, New York Zoological Society.

2This study has been supported by National Science
Foundation Grants G 4385 and G 21007.

3Part 1 of this paper was published in Zoologica, 46 (3) •
27-48 (April 28, 1961).

caves. In addition, at least five small colonies,
four of them in sea caves, have been extermi-
nated in recent years as the result of raids by
fishermen and others. These also are listed in
Table I. It is rather unlikely that any important
colony remains undiscovered, but not impos-
sible that in the limestone of the Aripo massif
there may be small undiscovered colonies. Some
of this country, which is honeycombed with
small caves, is difficult to penetrate and not
much visited.

I was able to visit all the known colonies,
some of them several times, and made an at-
tempt to assess the present adult population
(Table I). Direct counting of all the adults is
impossible in most of the caves; instead, it is
necessary to count all the nests which appear to
be occupied, count all the birds that can be
seen perched on nests and ledges, estimate the
number of birds flying about, and from these
figures assess the number of birds present. I
think it unlikely that the estimated total of 1,460
adults is out by more than 500 either way.

Such a small population of a very specialized
bird is highly vulnerable, and indeed the extinc-
tion of five colonies is evidence that the Trinidad
population has been reduced in recent times.
None of these exterminated colonies can, how-
ever, have been very large. The larger Trinidad
colonies, like those in Venezuela, have persisted
for years in spite of constant exploitation by
man. Though their numbers have probably been
reduced, they are certainly under no immediate
threat. As long as the forest remains, which pro-
vides the Oilbirds’ food, the steady harvesting
of a proportion of the young birds, though de-
plorable, is no danger to the survival of the

199

200

Zoologica: New York Zoological Society

[47: 16

Text-fig. 1. Trinidad Oilbird colonies. 1: Chacachacare (extinct). 2: Huevos. 3: L’Ance Pawa (extinct).
4: Saut d’Eau (extinct). 5: La Vache. 6: Maracas Bay (extinct). 7: Acono (extinct). 8: Arima gorge. 9:
Aripo main cave. 1 0: Aripo well cave. 1 1 : Aripo middle cave. 1 2: Aripo small cave. 1 3: Oropouche. For
further details see Table I.

species. A more real danger may come from the
mounting pressure of the human population on
the surviving areas of forest. It is to be hoped
that future governments will both conserve the
forests and protect the birds.4

The Arima gorge colony is known to have
fluctuated greatly. During the present study
there have usually been 25-30 adults present,
and the number of breeding pairs in the years
1957-62 has been as follows: 11, 13, 9, 9, 8, 11.
The breeding population cannot in fact rise
much above the 1958 figure, as there are not
more than 15 or 16 suitable nest-sites. In Sep-
tember, 1959, the colony was raided by a
poacher; at least eight adults were killed and for
some weeks the cave was almost deserted of
birds. Gradually the survivors returned, until by
December, when breeding began again, there
were 14 birds, and a little later 21 birds were
present. At various times in the past the num-
bers reported have been very low, probably as a
temporary result of the same cause. Clearly this
colony could easily be exterminated by steady
persecution.

The Annual Cycle of Breeding and Moult

There has hitherto been no exact information

4Oilbirds are protected by law in Trinidad, but the law
is not enforced and would be difficult to enforce. The
Aripo, La Vache and Huevos caves are still regularly
raided. The Oropouche cave is on private land and the
owners are trying to protect it, but it is not difficult for
poachers to enter unobserved and they apparently still
do so. The Arima gorge colony is also on private land
and is protected by the owner.

on the Oilbird’s breeding season, for either
Venezuela or Trinidad. Most naturalists have
visited the colonies in the first half of the year,
especially from March to May when the birds
are known to be nesting. There is a widespread
belief in Trinidad that May is the month to col-
lect young birds; luckily for the birds, this is
only partly true. May is a good month, but
other months may be equally good.

Table II shows that in the five years 1957-
1961 clutches were started in the Arima colony
in every month except October and November,
with 87% of the total in the six months Decem-
ber-May. There is a distinct annual breeding
season, with a very variable start. In three years
the main period of laying started in December,
in one year in January (1962, for which records
were obtained up to the end of May), and in
two years in March. In each year several nests
started within a week or two of each other, and
the later nests in the following few weeks. There
seems no doubt that some kind of mutual stimu-
lation is involved in this synchrony.

Limited information on other colonies has in-
dicated breeding seasons similar to that at the
Arima gorge colony. But there are not many
caves where it is at all easy to inspect the nests,
and information is very incomplete. Also, these
other caves are regularly disturbed by poachers,
so that the normal breeding season may be ob-
scured. Thus visits to La Vache cave in 1958,
the Huevos cave in 1959 and the Aripo Well
cave in 1960 and 1961 all indicated a breeding
season similar to that at the Arima gorge in the

1962]

Snow: Population, Breeding Ecology and Food of the Oilbird

201

same year; but on another visit to the Aripo Well
cave in March, 1959, when breeding was just
beginning in the Arima gorge (Table II), six
of the ten nests whose contents could be seen
contained young of a size that indicated egg-
laying in December and January.

Intervals between Broods and Sequence of
Broods.— The limited evidence suggests that un-
disturbed pairs retain their nests year after year.
One banded bird was seen on the same nest in
three successive years, and another in two suc-
cessive years. Unfortunately the raid on the cave
in September, 1959, when a number of adults
were killed, probably including two banded
ones, put an end to these observations, but they
suggest that successive clutches in the same nest
are normally laid by the same female, and this
has been assumed unless there was evidence to
the contrary.

On this assumption, an analysis has been
made of the intervals between the ending of one

nesting attempt and the laying of the first egg of
the next clutch, excluding intervals spanning the
time of the raid on the cave. Table III gives the
47 intervals that are available for analysis. It
will be seen that nesting attempts that ended in
failure in the months December-May were
nearly all followed by a new clutch after inter-
vals ranging from 19 to 75 days (mostly 19-33
days) . Those that ended in failure in the months
June-October were nearly all followed by in-
tervals of several months, during which the pair
concerned probably moulted (see next section),
leading to re-laying in the following breeding
season. The intervals following successful nest-
ing attempts were less consistent. Those that
ended in the months April- August were followed
either by intervals of less than 34 days, leading
to a further attempt in the same breeding season,
or by long intervals of several months during
which the pair probably moulted, leading to
re-laying in the next breeding season. Among

Table I. Present Colonies of Oilbirds in
Trinidad and Their Populations

Colony

Population

Oropouche cave

200

Aripo caves

“Main cave”

400

“Small cave”

10

“Middle cave”

140

“Well cave”

80

Arima gorge (Spring Hill cave)

30

La Vache cave

300

Huevos cave

300

1460

Notes on Colonies
Present colonies:

Oropouche cave. Extensive limestone cave
formed by stream flowing out of hillside through
narrow entrance. Nests from half-light near en-
trance to total darkness of furthest large chamber.
Population based on count of birds leaving the
cave in the evening (Part 1, p. 33).

Aripo main cave. Large limestone cave with
wide entrance and stream flowing down into cave.
Most nests in subdued light near entrance. Popu-
lation estimate based on adults seen in cave and
occupied nests.

Aripo small cave. Small limestone cave with
dry floor; nests in subdued light. Counts on three
visits: 12, 12 and 10 birds.

Aripo well cave. Dry limestone cave with top
entrance. Some nests in almost full daylight, pos-
sibly even receiving a little sunlight. Population

assessment based on counts of birds on nests and
ledges.

Arima gorge. Partially covered-over gorge with
stream flowing through. Adult population usually
25-30.

La Vache cave. Sea cave. Single large vaulted
chamber about 75 feet high, with semicircular
entrance 20 feet high. Assessment based on esti-
mates of flying and perched birds on three visits.

Huevos cave. Sea cave. Single long chamber
about 100 feet high, with entrance 15 feet high.
Assessment based on direct count of 230 perched
birds, with allowance for some flying and others
perched out of sight.

Extinct colonies:

Chacachacare. Small sea cave on west side of
island. Remains of nests and a few dry seeds
found by Dr. T. H. G. Aitken, November, 1958.
No further details known.

L’Ance Pawa. Sea cave about quarter of a mile
to seaward of L’Ance Pawa Bay. Contained about
a dozen or 20 birds in 1918 (Williams, 1922).
Deserted in 1960, and no sign of recent occupa-
tion.

Saut d’Eau. Open sea cave on northeast side of
Saut d’Eau Island, with a short tunnel through
the cliffs. According to local information, occu-
pied ten or twenty years ago but found deserted
in 1960 and apparently had been for several years.

Maracas Bay. Small sea cave a short distance
east of Maracas Bay. Deserted in 1957; some
remains of old nests found. According to local
information, had been occupied some 15 years
previously.

Acono cave. Gorge in Acono Valley. Occupied
in nineteenth century (Wall & Sawkins, 1860);
deserted in 1918 (Williams, 1922).

202 Zoologica: New York Zoological Society [47: 16

Table II. Breeding Seasons of Oilbirds in the Arima Gorge Colony

Numbers of clutches started

in each month

Jan.

Feb.

Mar.

Apr.

May

June

July

Aug-

Sept.

Oct.

Nov.

Dec.

19571

4

5

2

2

1

1

4

1958

2

1

8

1

4

2

—

—

_

—

—

—

1959

—

—

3

3

3

_

1

2

_

_

—

1

1960

3

4

—

1

1

—

—

_

3

1961

2

2

1

1

—

_

19622

Totals

1957-61

5

1

4

3

1

7

7

16

10

11

2

3

3

1

-

-

8

1Visits did not start until March 20, 1957. There were then no nests that could have started in January or
February, but it is possible that some had started and already lost their contents by March 20.

2The 1962 records were obtained by Mr. J. Dunston, who inspected the cave regularly from the end of September,
1961, to June, 1962, after my own observations had ceased. These extra 14 nests are not included in the analyses
of clutch-size, breeding success, etc., given elsewhere in this paper.

the short intervals there were two cases of the
first egg of a new clutch being laid, in one case
about 13 days before, and in the other case at
about the same time as, the departure (from the
same nest) of the last young of the previous
brood. Those nesting attempts that ended suc-
cessfully in the months October- January were
followed mainly by intervals of intermediate
length, leading to re-laying early in the following
breeding season.

A complete nesting cycle, from the laying of
the first egg to the fledging of the last young,
lasts about five months (Part 1, p. 39). Thus
with a short interval between broods it is just
possible to rear two broods in a year, and this
was in fact achieved by three pairs during the
present study. One of these reared two broods
between early December and the following
October, having also reared a brood between
April and September of the previous year. Thus
they reared three broods in eighteen months.
After the last they had an interval of 13 months
before the next clutch was laid. Another pair,
having completed the rearing of two broods be-
tween December and October, laid a new clutch
in December, after an interval of 49 days, cer-
tainly not long enough for the moult. This brood
was also successful and fledged in the following
May, so that three broods were again reared in
18 months. There was then an interval of eight
months before the next clutch was laid in the
following January. The third pair reared two
broods between March and the following Jan-
uary; they then had a gap of five months before
the next clutch was laid, in late May but this
failed and the sequence was broken. The data
are insufficient for a proper analysis, but they
suggest that pairs breeding successfully can rear
two broods in a year only every second or third

year. Discussion cannot be taken any further
without considering the relation between breed-
ing and the moult.

The Moult.— At every visit to the colony, all
moulted flight-and tail-feathers were collected.
Most of the moulted feathers are dropped out-
side the cave, presumably during feeding, but
enough are dropped inside the cave to give an in-
dication of the amount of moulting taking place
in the colony. Nearly all were found on the
slopes below the nests and could not be alloca-
ted to known nests.

Moulting takes place in every month of the
year, but there is a well-marked seasonal varia-
tion. There is most moulting in the six months
June-November (72% of all moulted feathers
being found in these months) and least moulting
in the six months December-May. Thus there
is an inverse correlation between egg-laying and
moulting, but nevertheless enough moulting
takes place in the months December-May to
make it very likely that some birds at least
undergo moult while breeding. Examination of
specimens suggests the same. Thirty-five dated
Trinidad specimens have been available for ex-
amination (13 in the American Museum of Na-
tural History, 1 1 in the Eritish Museum and 1 1
trapped or collected during this study). Twenty-
six were collected in the months January-May,
the main egg-laying period, and 16 of them were
undergoing moult of the primaries. The remain-
ing nine were collected in July-October, and
seven of them were moulting their primaries.

The sequence of moult is variable and some-
times irregular. Of 29 Trinidad specimens which
were in wing-moult (those mentioned above
and some undated specimens), 14 appeared to
be undergoing a normal replacement of the
primaries from the inner end of the row out-

1962]

Snow: Population, Breeding Ecology and Food of the Oilbird

203

Table III. Intervals between Broods

(Length of time between ending of one nesting attempt and laying of first egg of next clutch)

Month in which
nesting attempt
ended

Intervals of less
than three months
(in days)

Intervals of more
than three months
(in months)

January

—

5

February

23, 25

—

March

24

—

April

-

SVi

May

0, 19, 19, 20,

23, 33, 47, 65,
75

9

June

44

6, 6, 6V2, 1, 1, 7, 9

July

—

6I/2, 7, 8, 8V2, 91/2

August

-3, 25

5, 5, 6V2, 1, 7V2, 8

September

76, 78, 84

3

October

49

4, 5V2, 1, 15

December

24, 72

—

Note: Figures in ordinary type: previous nesting attempt successful. Figures in italic type: previous
nesting attempt ending in loss of eggs. Figures in heavy type: previous nesting attempt ending in loss
of young.

wards. Four other birds showed a similar con-
dition, except that the moult of one wing was
one or two feathers in advance of the other.
The remaining 1 1 birds showed more complex
or irregular conditions. In six, one wing was
moulting but not the other, and in one of these
moult was occurring in two separate places in
the row. The remaining five all showed moult in
two separate places in the row in at least one
wing, and in one bird three feathers in the
same wing (primaries 1, 6 and 9) were all in
various stages of growth.

The most reasonable explanation for the
presence of growing feathers at two separate
places in the row seems to be that a second
moult of the primaries had started before the
previous one was complete. The clearest case
was that shown in Text-fig. 2a, where a second
moult had apparently started when the previous
one was about half completed. Another bird was
similar, except that the two wings were at
slightly different stages. A third case, shown in
Text-fig. 2b, was the bird mentioned later in this
section, which was caught twice while moulting,
and another showed an exactly similar condi-
tion, including the asymmetry between the two
wings, except that the moult was at a slightly
later stage. Alternatively, it may be that the
moult sometimes starts simultaneously at two
“foci” and proceeds outwards from these two
points. Parallels for both conditions can be
found in other birds; e.g., overlapping successive
moults in boobies (Dorward, 1962), multiple

“foci” in birds of prey (V. & E. Stresemann,
1960).

The length of time needed for a complete
moult must be several months. Examination of
specimens showed that altogether there were
only seven instances of two adjacent primary
feathers growing, as against 51 instances of
a single feather growing, in many cases nearly
full-grown. Thus (unless moult starts at two
foci) replacement of the whole series of ten
primaries must on average take nearly ten times
the length of time needed for the growth of a
single feather. One individual of the Arima
gorge colony was caught twice during the same
moult, at an interval of exactly four weeks. Of
the three feathers which were one-quarter grown
on April 21, one was full-grown and the other
two nearly full-grown on May 19 (Text-fig. 2b) .
If this bird was typical, more than ten months
would be needed for a complete moult of the
primaries, and even with moult starting from
two foci not less than six months would be
required.

Table III shows that several of the long inter-
vals between nesting attempts were of less than
six months, so that for these birds the moult
must surely have overlapped the previous or
subsequent nesting cycle, and since six months
is probably an underestimate for the period of
the moult, others may have done so as well.

The possibility must be considered that the
rate of moult is variable; in particular, that it is
faster during the off-season and s’ows down

204

Zoologica: New York Zoological Society

[47: 16

A

B

Text-fig. 2. Stages of primary moult in two Oilbirds. The ten primaries are shown diagrammatically from
above, as follows: solid lines, complete or growing feathers, not recognizably old; broken lines, recogniz-
ably old feathers; arrows, amount grown between first and second captures (see text).

when the bird is breeding. Even without any
connection between the timing of the onset of
the moult and the breeding cycle, such a varia-
tion in rate of moult could account for the fact
that moulted feathers were found in all months,
but nearly three times as numerously in the
months July-November as in the months Be-
cember-May.

To summarize, the data show that the moult
mainly occurs in the six months June-November,
coinciding with the long intervals following nest-
ing attempts that end in those months, and the
evidence suggests that the moult of each indi-
vidual lasts for several months and may over-
lap the nesting cycles. The possibility of varia-
tion in rate of moult according to the state of
breeding must also be borne in mind. More
definite conclusions are not warranted on the
present evidence. Far more captures of birds
with known nesting histories would be neces-
sary to establish the exact relationship between
breeding and the moult. Since it was found that
the capture of adults had a disturbing effect on
the whole colony, it was not practicable to ob-
tain more than these fragmentary data.

Clutch-size

Clutches numbered from one to four eggs. Of
the four one-egg clutches, three were laid by the

same bird, perhaps a young bird breeding for the
first time (see next section). The mean size of
the 59 completed clutches was 2.7 eggs (Table
IV).

Mean clutch-size decreased in the course of
the breeding season from just over three to two
eggs, 10 of the 11 clutches of four eggs being
laid in the early part of the breeding season (De-
cember-March) . It will be shown later that most
food is available in the months March-June
(Text-fig. 4) , the period when most of the young
from the early clutches are in the nest. It is
reasonable to suppose that the seasonal variation
in clutch-size may be adapted to the seasonal
change in the food supply, as has been found for
many north-temperate birds (e.g., Lack, 1954).

Breeding Success

In the five years of observation, nearly half
of the nesting attempts produced flying young.
There were 68 nesting attempts (in which at
least one egg was laid) ; 45 of these reached the
hatching stage (and four more may have done
so, the eggs or young disappearing around the
hatching time); and of these, 31 reached the
fledging stage, producing a total of 60 young.
The earlier nests were more successful than the
later ones, 21 out of the 38 nests starting in the
months December-March producing fledged

1962]

Snow: Population, Breeding Ecology and Food of the Oilhird

205

Table IV. Clutch-size

Number of

eggs in clutch

Mean

clutch-

size

1

2

3

4

December

1

2

3)

'X 1

lanuary

1

2

1

3 (

J.l

February

1

4

2)

2 7

March

1

7

6

2 f

April

4

5

l

2 4

May

1

5

2

1 \

June

2

1

July

1

1

1-

(2.0)

August

1

J

Total (59)

4

23

21

11

2.7

young, compared with 10 out of the 30 starting
in April-September. These figures are too small
to be convincing of themselves, but the differ-
ence is probably real and due to the nesting be-
havior becoming less efficient as the season
advances. Thus there were seven instances of
eggs being deserted or lost soon after laying in
the months April-September, and only two in
the months December-March. There was an un-
broken run of successes at only one nest, prob-
ably kept by the same pair throughout the
period. At this nest there were five successful
nestings, producing nine flying young. (This was
one of the nests mentioned earlier, in which
three broods were reared in 18 months) . No nest
had an unbroken run of failures, though some
nearly did. In some cases a sequence of failures

probably represented the first nesting attempts
of a young pair. Thus at one nest four failures
were followed by two successes, the clutch-sizes
for this sequence being 1, 1, 1, 2, 4, 3.

Table V shows the fate of the 171 eggs laid.
For nine clutches there was uncertainty as to the
number of eggs laid, due to the disappearance of
the clutch during laying. In these cases the
greatest number known to have been laid has
been counted. For others there is uncertainty as
to the number of young hatched, owing to the
disappearance of an egg (or young) at about
the time that hatching was due. As unhatched
eggs were known in several cases to remain in
the nest long after the others had hatched, and
several young were known to have died soon
after hatching, it is likely that in most of these
uncertain cases the eggs did in fact hatch.

The causes of egg failure were various. Some-
times an adult may kick off an egg as it flies
from the nest. This was seen to occur once and
it probably accounted for several other egg
losses. Three times nests were flooded by water
seeping through cracks in the rock during very
wet weather and the eggs were chilled and even-
tually thrown out by the owners. Many other
egg losses were unaccounted for; a few of these
may have been caused by the crabs (Pseudo-
thelphusia gannani ) which occasionally walked
over the nests and probably attacked the young.

Nearly 60% of the young known to have
hatched reached the flying stage. As with eggs,
the causes of failure were mostly unknown.
Most died in the early stages, and it is probable

Table V. Hatching and Fledging Success

Eggs /young present

Losses

Eggs laid 171

Eggs remaining at hatching time 129

Eggs hatched 103

Young flying 60

Deserted without being incubated 2

Flooded; thrown or washed out 7

Kicked off by adult 1

Used for egg-white sample 1

Disappeared before possible hatching date

(cause unknown) 31

Infertile 7

Deserted 1

Flooded and chilled; embryo killed 1

Died during hatching 1

Disappeared at hatching time; not known if
hatched or not 16

Lost at ages 0-20 days 21

Lost at ages 21-40 days 15

Lost at ages 41-60 days 2

Lost at ages 61-80 days 4

Lost at 80+ days 1

206

Zoologica: New York Zoological Society

[47: 16

that competition for food with their older nest-
mates was an occasional cause of death at this
period (Part 1, p. 41); there was no evidence of
starvation in the later stages. Three young were
almost certainly killed by crabs, the corpses of
two of them being found on ledges near by,
partly eaten. (Large crabs had been seen fre-
quenting this and other nests, and no other po-
tential predators could be found. When the holes
in which the crabs lurked were blocked up, no
other deaths of this kind occurred). Four young
fell from their nests when part of the structure
collapsed. In addition one fell from an intact
nest and it is almost certain that others did.
There was no evidence of predation of the young
by man during the period of the study.

Food: General Remarks

There have been no detailed studies of the Oil-
bird’s food. Most authors have stated that they
feed mainly on the fruits of palm trees, while
some have remarked that the family Lauraceae
is also important. McAtee (1922) lists some
seeds found in a Trinidad colony and refers to
earlier accounts, pointing out that these may be
of doubtful value owning to uncertainty of iden-
tification. For Venezuela, Pietri (1957) lists sev-
eral species of palms, two species of Lauraceae
and one of Burseraceae, and says that probably
other kinds of fruit are eaten.

The feeding habits have been described in
Part 1, pp. 31-32. Here it may be recalled that
the birds pluck fruit in flight, at night, swallow-
ing them whole. Nearly all the fruits taken have
a firm pericarp enclosing a single rather large
seed. Only the pericarp is digested, the seeds
being later regurgitated intact.

In the course of the present study, regurgi-
tated seeds were collected regularly at the Arima
gorge colony from April, 1957, to September,
1961, samples of several hundreds, sometimes
several thousands, being taken in each month.
Most were collected from catching trays slung
on the slopes below the nests, and others from
the nests themselves. Care was taken to collect
only freshly regurgitated seeds. Altogether, over
112,000 seeds were collected and examined.
Samples were also collected from other caves
as opportunity arose.

The identification of seeds in an area of rich
tropical forest presents some difficulty. A few
kinds, especially the distinctive seeds of some
common palm species, were soon identified.
Others remained unidentified for a long time,
but in nearly every case the tree was eventually
found in the forest. Two species were identified
from herbarium material on the basis of whole

fruits and seeds collected in the cave. In the end
only one important food tree remained uniden-
tified. For the study of Oilbirds’ food it is unfor-
tunate that the Lauraceae is taxonomically a
difficult family. There are many species with
rather slight differences between them, and even
in the important national collections there is a
shortage of good material with mature fruits.

An incidental outcome of this study was the
discovery of some species of Lauraceae new to
Trinidad. Beilschmiedia sulcata represents a
genus and species new to the island; Licaria
guianensis and Ocotea caracasana were previ-
ously unrecorded from Trinidad; and specimens
of an undescribed species, since named Ocotea
trinidadensis Kostermans, were collected in the
course of examining Lauraceae in the forest,
though its fruit was not recorded in the Oil-
bird’s food.

Composition of the Food

Table VI gives the totals of all the seeds col-
lected from the Arima gorge colony throughout
the period of the study. Eighteen species were
important (over 100 seeds collected): 7 palms,
8 species of Lauraceae, 2 species of Burseraceae,
and one species of Araliaceae5. Numerically, the
Burseraceae outnumbered the Lauraceae in spite
of the small number of species. Text-fig. 3 gives
outline drawings of the more important seeds,
with descriptive notes, and six kinds are illus-
trated in Plates I-III.

Most of the fruits eaten were regularly sea-
sonal. This was especially true of the Lauraceae,
which with only one exception showed clearly
defined fruiting seasons varying little from year
to year.

Palms— The two important palms Euterpe
langloisii and Jessenia oligocarpa were the two
striking exceptions to the general rule that the
food trees are seasonal in their fruiting. Seeds
of Euterpe were present in every sample, and in
many samples they outnumbered all other seeds
(Table Vila). Jessenia, a much larger fruit, was
always present in much smaller numbers, and
there was one long period, from November,
1959, to July, 1960, when it was very scarce and
failed entirely for three months (Table Vllb).
During this period, no trees in the forest could
be found with ripe fruit.

In these two palms, each tree bears inflores-
censes and bunches of fruit in various stages of
growth, and field observations showed that the
fruit ripens very slowly. In addition, the ripe
fruit has a hard and dry pericarp and probably

5Wrongly described in Part 1, p. 42, as probably myr-
taceous.

1962]

Snow : Population, Breeding Ecology and Food of the Oilbird

207

remains in an edible condition on the tree for
many weeks. Thus these two palms normally
produce a rather constant supply of ripe fruit
throughout the year. The cause of the general
failure of Jessenia in 1959-1960 was not dis-
covered; probably unsuitable weather some time
previously had led to a gap in its flowering.

Euterpe is abundant in the forests of the
Arima Valley, but Jessenia is rare, having its
center of abundance in more low-lying swampy
forests five or ten miles to the southeast. This
is probably the main reason why the seeds of
Jessenia were always less numerous than those
of Euterpe.

Bactris cuesa, a small forest palm of the
under-story, is much more seasonal in its fruit-
ing than Euterpe and Jessenia (Table Vile).
Its season was even more sharply defined than is
suggested by the table, as most of the fruits col-
lected in the samples before May were unripe
and regurgitated intact.

Roystonea, which was collected regularly in
small numbers in the samples, does not grow in
the Arima Valley. Its natural habitat in Trinidad
is the low-lying swampy forest near the east
coast, especially the Nariva Swamp, but it has
been much planted in gardens and parks and the
Arima gorge birds probably obtained its fruit
from the town of Arima, five miles away. Table
Vlld shows that it was found in the samples in
all months of the year.

Livistona chinensis is an introduced palm. It
probably does not grow nearer to the cave than
Arima, where it has been planted in small num-
bers. October to April are the main months for
ripe fruit. Seeds of Livistona were found in mod-
erate numbers in the samples from 1957 to
April, 1959, but not thereafter. Probably this re-
flects a temporary habit of a few birds, or pos-
sibly the individuals visiting the Livistona trees
were among those killed in the raid on the cave
in September, 1959; the trees fruited abundantly
in Arima in the two subsequent years.

Geonoma vaga is a small palm of the forest
undergrowth, growing to about 12 feet. Its fruits
are very small (about 4 mm. in diameter), the
hard seed being covered by a thin pericarp. Like
Livistona, it too was found in the samples only
up to 1959, in the months January-July. It is
remarkable that Oilbirds should be able to take
its fruit at all, as it grows chiefly along streams
and gullies, in rather dark parts of the forests,
and surprising that they should have found it
worth while to do so. Possibly one or two trees
growing near the cave mouth were occasionally
visited by the birds as they entered or left the
gorge.

Unripe fruits of species of Bactris type were
found in the samples somewhat irregularly but
in all months of the year, occasionally amount-
ing to an important fraction of the total food.
The fruits were small and soft, without a seed,
and many were regurgitated whole or only
slightly broken up. Identification of most of
them was uncertain, but probably the great ma-
jority were immature fruits of Bactris gasipaes
(locally known as Pewa), a palm that is culti-
vated in small numbers for its edible fruits. In
addition to the mature fruits, which contain a

Table VI. Total Numbers of Fruits
Represented in Collections from the
Arima Gorge Colony, 1957-1961

Palmae

Euterpe langloisii

45,520

Bactris cuesa

6,315

Jessenia oligocarpa

4,669

Unripe Bactris sp.

663

Livistona chinensis

489

Roystonea oleracea

327

Geonoma vaga

165

Bactris sp.

17

Desmoncus sp.

6

Aiphanes sp.

3

Bactris sp.

2

Total

58,176

Lauraceae

Ocotea wachenheimii

9,880

Cinnamomum elongation

5,831

Ocotea caracasana

3,853

Nectandra martinicensis

2,846

Ocotea oblonga

2,370

Nectandra kaburiensis

290

Aniba firmula

267

Nectandra membranacea

149

Aniba trinitatis

70

Aiouea schomburgkii

68

Licaria guianensis

12

Nectandra sp.

4

Beilschmiedia sulcata

3

sp. indet.

2

Ocotea canaliculata

1

Total

25,646

Burseraceae

Dacryodes sp.

21,895

T rattinickia rhoifolia

6,731

Protiwn sp.

1

Total

28,627

Araliaceae

sp. indet.

148

Oleaceae

Linociera caribaea

50

Myristicaceae

Virola surinamensis

10

Anacardiaceae

Tapirira guianensis

7

Sapotaceae

Pouteria minutiflora

3

Boraginaceae

Cordia bicolor

1

Malpighiaceae

Byrsonima spicata

1

Unidentified

Several species

48

Total

268

Grand Total 112,717

208

Zoologica: New York Zoological Society

[47: lb

Text-fig. 3. Seeds of the chief fruits eaten by Trinidad Oilbirds. (All natural size except where indicated.)

1. Jessenia oligocarpa (palm). Hard and fibrous, streaked dark and pale brown (see also Plate I).

2. Euterpe langloisii (palm). Hard and fibrous, pale yellow-brown. When dry, fibers curl off giving a hairy
appearance (see also Plate I).

3. Badris cuesa (palm). Hard and blackish, sometimes with pale fibers adhering; three pits (germ-pores)
on the flattened end.

1962]

Snow: Population, Breeding Ecology and Food of the Oil bird

209

4. Roystonea oleracea (palm). Distinctively boat-shaped; pale fibrous surface, the fibers not coming away.

5. Livistona chinensis (palm). Smooth, hard, whitish, non-fibrous surface, with slight points at both ends,
and slightly ridged down opposite sides (see also Plate II).

6. Geonoma vaga (palm). Hard and smooth; dark brown, with a pale apical spot and a pale longitudinal
streak all the way round. (Left, natural size; right, enlarged.)

7. Unidentified Bactris sp. (palm). Hard and blackish, with three germ-pores, two with ridges above.

8. Dacryodes sp. (Burseraceae). Whitish, soon becoming stained, with demarcated “panel” becoming de-
tached when the seed dries. When fruit is mature, seed contains seedling coiled up within.

9. Trattinickia rhoifolia (Burseraceae). Hard, brown; with two points at one end and one at the other,
and complex ridges and depressions (see also Plate II).

10. Beilschmiedia sulcata (Lauraceae). Smoothish, pale brown; rather variable in shape and size, but
typically very slightly curved. Soon splits into two halves (and germinates readily).

11. Ocotea wachenheimii (Lauraceae). Smooth, very slightly ridged; pale brown when fresh (see also
Plate III).

12. Ocotea oblonga (Lauraceae). Dark brown, pitted with pale longitudinal streaks; fairly prominently
ridged down opposite sides.

13. Ocotea caracasana (Lauraceae). Smooth, brown; regularly spindle-shaped.

14. Cinnamomum elongatum (Lauraceae). Smooth, pale to dark brown depending on maturity; somewhat
variable in shape (two extremes shown); rather soft.

15. Aniba firmula (Lauraceae). Smooth, brown with blackish streaks; variable in size and shape, but
with point at one end and rounded at the other, and usually broadest at pointed end (see also Plate III).

16. Aniba trinitatis (Lauraceae). Smooth, pale brown, unstreaked or faintly streaked; variable in shape,
but characteristically widest near pointed end and often with slight ridges on opposite sides.

17. Nectandra kaburiensis (Lauraceae). Smooth, pale brown when fresh, nearly spherical; thin-skinned,
with pinkish-purple endocarp.

18. Nectandra membranacea (Lauraceae). Brown, with slight points at both ends, and regular “fluting,"
as shown in transverse section.

19. Nectandra martinicensis (Lauraceae). Smoky brown, with black showing through in irregular longi-
tudinal streaks. Slightly asymmetrical, one side more curved than the other, and apex with lip-like hilum
(right, enlarged).

20. Licaria guianensis (Lauraceae). Smooth, dark brown, with point at narrow end; slightly ridged
longitudinally; variable in size.

21. Aiouea schomburgkii (Lauraceae). Smooth, brown, mottled with dark markings; almost spherical.

22. Araliaceae sp. Seeds flattish, roughly half-moon shaped; embedded radially in berry as shown.

23. Linociera caribaea (Oleaceae). Pale brown, with a network of raised veining; bluntly oval.

large seed and are probably too large for Oilbirds
to swallow, they produce small fruits that do
not develop to maturity. One owner of a small
estate told me that he had shot an Oilbird that
was visiting his Pewa trees.

Besides these seven main kinds, there was a
very small number of palm seeds of other spe-
cies. Several seeds were close to Bactris cuesa
but differed consistently in shape and probably
represented other species. The taxonomy of this
section of the genus is difficult, and there has
been some fine splitting of species (Bailey,
1947). Aiphanes sp., ? minima, of which three
seeds were found, is a cultivated species with
edible fruits (locally known as “edible gri-gri”).
Six seeds were referred to Desmoncus, a genus
of spiny climbing palms, but they were not the
common species D. major.

Lauraceae— Table VIII shows the percentage

occurrence of the ten most important species of
Lauraceae in the food. With one exception
( Ocotea wachenheimii) , they all had well-de-
fined fruiting seasons which were much the same
in all years; in all but two of them the season fell
mainly in the months March-June. One of the
exceptions, Aniba firmula, fruited rather earlier,
in December-April, while the other, Nectandra
martinicensis, was the only species with a fruit-
ing season at the end of the year, from Sept-
ember to November. Table X summarizes the
fruiting seasons of these ten regularly occur-
ring species, and shows clearly the importance of
the March-June peak.

In number of seeds found in the samples,
Ocotea wachenheimii was the most important
species, and, as mentioned above, it was the
only species with an irregular fruit season. In
all years a main fruit season began in November

210

Zoologica: New York Zoological Society

[47: 16

Table VII.

Seasonal Occurrence of

Palms in the Food Samples

Percentage of the total in the different months

Jan. |

Feb.

Mar.

Apr.

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

(a)

1957

20

28

51

64

74

58

43

75

64

Euterpe

1958

26

33

31

28

31

18

16

37

46

53

45

4

langloisii

1959

9

44

54

49

26

28

6

2

36

32

41

82

1960

40

15

37

58

52

21

15

16

30

54

78

86

1961

62

54

73

56

45

44

30

6

21

(b)

1957

17

7

3

7

8

12

13

15

10

Jessenia

1958

5

6

2

1

1

2

4

11

12

7

8

2

oligocarpa

1959

6

14

25

20

5

3

3

1

2

1

+

+

1960

+

+

+

+

1

9

12

11

9

1961

3

1

1

1

2

6

5

2

3

(c)

1957

1

4

9

10

11

3

1

+

Bactris

1958

+

+

+

+

1

3

17

33

18

15

25

2

cuesa

1959

1

+

4

17

55

74

17

10

+

1960

+

+

+

2

5

4

7

11

1

1961

+

+

+

5

23

29

16

(d)

1957

+

+

+

1

1

1

2

+

Roystonea

1958

+

+

+

+

1

1

+

+

+

+

oleracea

1959

+

+

+

+

+

1960

+

+

1

2

+

+

1961

+

+

+

Note (applying also to Tables VIII and IX) : +, less than one percent. Dots no sample collected.

Where the year is omitted, no seeds were found in that year.

or December, which lasted from four to six
months. In 1958 and 1960, and less markedly in
1961, there was a secondary minor fruit season
in June or July. The fruiting period April-July,
1957, might have been either the tail-end of a
long and prolific fruiting beginning at the end of
the previous year, or it might have been a com-
bination of such a fruiting with a secondary
fruiting in June-July.

A single tree of 0. wachenheimii, growing
near the cave, was kept under observation from
January, 1958, to September, 1961. It had ripe
fruit regularly from late November or Decem-
ber to February, and had none in June or July.
Another tree on an exposed ridge at a higher
altitude, which was visited less regularly, had
nearly ripe fruit in June in two years. Three out
of the four samples of seeds taken from caves
at the extreme western end of Trinidad in July
and August had good numbers of seeds of 0.
wachenheimii, including one collected in 1959
when none were being taken by the birds in the
Arima gorge. It seems probable that individual
trees of this species have one main fruiting sea-
son per year, which may be either November-
February or June-July, and it may be that the
June-July fruiting season is more general in the
drier parts of the island.

Beilschmiedia sulcata is not included in the

ten main species of Lauraceae, as only three
seeds were found in the Arima gorge (in March,
May and June), but it is one of the most impor-
tant fruits taken by the Oilbirds in the caves to
the east of the Arima Valley. Fresh seeds were
found in nearly every month of the year in these
caves, but their presence was irregular. Thus
there were good numbers in May and Decem-
ber, 1957, but none in December, 1959, or May,
1960. Beilschmiedia grows at high altitudes on
the Aripo massif, where the rainfall is extremely
high and probably less seasonal in its distribu-
tion than at lower altitudes, and the fruiting of
trees may be less regularly seasonal.

Burseraceae — The most important species,
Dacryodes sp., fruited in alternate years during
the period under study (1958, 1960 and 1962),
very small numbers of seeds being found in the
intervening years. It is a common tree above
1,500 feet in the Northern Range and when it
was in fruit its seeds often outnumbered every-
thing else in the samples. Individual trees re-
main in fruit for many weeks, as the fruits,
borne in large bunches, ripen slowly and irregu-
larly. In the intervening years, some trees that
were examined produced fruit which grew to
full size and looked normal but were in fact
hollow. There was no evidence that these were
taken by Oilbirds.

1962]

Snow: Population, Breeding Ecology and Food of the Oilbird

211

Table VIII. Seasonal Occurrence of Lauraceae in the Food Samples

Percentage of the total in the different months

Feb.

Mar.

Apr.

May

June

July

Aug.

Sept.

Oct. | Nov.

O cotea 1957

28

28

20

11

2

+

+

caracasana 1958

+

3

1

1

1959

+

1

3

2

+

1960

+

2

1

+

1961

1

4

4

4

2

Ocotea 1958

3

13

1

+

oblonga 1959

21

17

1

1960

3

2

1961

3

7

4

Ocotea 1957

7

7

3

1

+

23

wachen- 1958

65

39

1

7

11

1

12

90

heimii 1959

56

1

1

1960

42

28

3

+

1

1

+

1

2

1961

34

44

21

4

+

+

Cinna- 1957

16

18

10

+

momum 1958

3

14

5

elongatum 1959

3

41

15

1960

+

4*

1

1

1961

+

27

37

9

Nectandra 1957

1

11

28

1

martini- 1958

20

19

3

censis 1959

26

50

56

2

1960

+

+

11

22

5

1961

+

+

Nectandra 1959

+

1

mem- 1961

3

1

branacea

Nectandra 1957

1

1

+

kaburiensis 1958

1

+

1959

+

1

1960

+

1961

2

1

Aniba 1958

2

+

firmula 1959

+

1960

+

3

4

+

Aniba 1958

+

+

trinitatis 1960

+

1

+

1961

+

1

+

Aiouea 1958

1

1

1

2

schom- 1961

+

burgkii

Note: The following species never accounted for as much as 1% of the monthly sample: Licaria guian-
ensis (found in May, 1958, and March-May, 1961); Beilschmiedia sulcata (found in March, May and June
in three different years).

Trattinickia seeds were found in the food
samples in nearly every month, but comprised
an important fraction of the total only in the
months June-September. A single tree kept
under observation for over 2Vi years fruited in
1959 and 1961, but not in 1960. After the first

lot of fruit was gone most of the leaves fell and
new flowers appeared seven months after the
growth of the new leaves; the fruits then took
ten months to ripen. The complete cycle took
two years. Three other trees that were less regu-
larly observed fruited at the same time as this

212

Zoologica: New York Zoological Society

[47: 16

Table IX. Seasonal Occurrence of Burseraceae and Araliaceae in the Food Samples

Percentage of the total monthly sample

Jan.

Feb.

Mar.

Apr.

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

Dacryodes

1957

+

+

sp.

1958

+

18

57

56

62

60

47

5

+

1

1959

+

3

1960

12

53

54

37

39

71

68

35

+

1961

+

Trattinickia

1957

4

7

6

6

4

5

6

3

1

rhoifolia

1958

1

2

1

1

1

1

2

1

1

4

4

1

1959

5

3

1

2

6

34

36

23

17

6

1

6

1960

4

1

+

+

+

+

11

41

37

4

1

+

1961

+

+

+

1

28

42

59

47

Araliaceae

1957

+

1

+

+

sp.

1958

+

+

+

1

+

1959

+

2

1

1960

+

+

2

+

1961

2

7

tree. But the continuous presence of Trattinickia
seeds in the samples shows that other trees must
have fruited in the intervening years. Like
D aery odes, Trattinickia trees remain in fruit for
several weeks.

Araliaceae— The only regularly eaten fruit
that did not conform to the usual type (firm
pericarp surrounding a single seed) remained
unidentified. The tree was never found, and
since the seeds were not found in any cave ex-
cept the Arima gorge, though five samples were
collected at other colonies at times when the
seeds were regular in the Arima gorge samples,
it is hard to avoid the conclusion that either the
tree is very local or its exploitation is an idio-
syncrasy of the Arima gorge birds. Occasional
regurgitation of whole fruits enabled the family
to be determined with reasonable certainty (see
also Appendix) .

Other Species.— Byrsonima spicata, Virola
surinamensis, Pouteria minutiflora, Tapirira
guianensis and Cordia bicolor are all common
trees in the Arima Valley, and their fruits are
much eaten by other kinds of birds. Their great
rarity in the food samples indicates that Oil-
birds in general avoid them. Byrsonima, Cordia
and Pouteria all have more or less succulent,
acid fruits, while Virola, the wild nutmeg, has a
net-like aril enclosing the seed. All these are
very unlike the sort of fruit usually eaten by
Oilbirds. Tapirira, also a common tree, has a
fruit that is superficially like that of Dacryodes
and is much eaten by pigeons. The reason for
its avoidance by Oilbirds is discussed in a later
section (p. 217).

Only one tree of Linociera caribaea (Olea-
ceae) was found in the Arima Valley, and two
others in a valley a few miles to the east. Thus

Table X. Fruiting Seasons of the Ten Most Important Species of Lauraceae

Jan.

Feb.

Mar. Apr. May June July Aug. Sept.

Oct.

Nov.

Dec.

Aiouea schomburgkii

X

X

X

X

Aniba firmula

X

X

X

Aniba trinitatis

X

X

Cinnamomum elongation

X

X

X

X

Nectandra martinicensis

X

X

X

Nectandra membranacea

X

X

Nectandra kaburiensis

X

X

Ocotea caracasana

X

X

X

X

Ocotea oblonga

X

X

X

Ocotea wachenheimii

X

X

X

X

X

X

X

X

X

X

Number in fruit in

each month

2

2

5

7

7

6

2

1

1

1

2

1

1962]

Snow: Population, Breeding Ecology and Food of the Oil bird

213

it is not a common tree, and the presence of its
seeds, in moderate numbers, in the food sam-
ples in one year only may indicate that a few
birds found a tree but did not revisit it in sub-
sequent years.

There was evidence of only one kind of fruit
being eaten besides those found in the food sam-
ples. A reliable observer reported an Oilbird
feeding on unripe fruits of the Tonka Bean
{Dipteryx odorata ) in an Arima garden. The
mature fruit, about three or four inches long, is
far too large to be taken. It is perhaps significant
that the Tonka Bean has a pervasive scent.

The Food at Other Colonies

Each time one of the other Trinidad colonies
was visited, a food sample was collected from
below the nests, and sometimes from the nest
themselves when they were accessible. On each
occasion, hundreds of seeds were examined and
an attempt was made to collect a sample show-
ing the proportions in which the different kinds
were present. Care was taken to collect only
fresh seeds. Altogether, 18 samples were col-
lected.

In general, the main kinds of fruit were the
same as were being eaten by the birds at the
Arima gorge colony in the same period, and
they were in roughly the same proportions. Not
a single seed was found at these other colonies
that was not at some time found at the Arima
gorge colony. Thus there is little doubt that the
food of the Arima colony is typical of the Trini-
dad population as a whole. There were, how-
ever, some differences between the samples
from these various colonies, which may be
briefly summarized.

Oropouche Cave. — The most easterly cave,
situated in an area of high rainfall, but at a low
altitude. Eight samples were collected in seven
different months. The composition was much
the same as at the Arima colony, but Beil-
schmiedia was regular, sometimes in large num-
bers, and Trattinickia was never found. Two
samples contained a smaller variety than was
found at the Arima colony at the same time, the
other six approximately the same variety.

Aripo Caves.— These are the highest caves, in
montane forest. Six samples were collected in
four different months. The composition was
much the same as the Oropouche cave (again,
Trattinickia was never found), but the variety
was regularly less than at the Arima colony. For
example in December, 1957, only Jessenia and
Beilschmiedia could be found, although eight
kinds of fruit were being taken at the Arima

colony. The reason may lie partly in the poorer
floristic composition of the montane forest
(Beard, 1946).

La Vache Cave— Two samples, both in Au-
gust. The composition and variety was almost
exactly the same as at the Arima colony at the
same time.

Huevos Cave — Two samples, in July and Au-
gust. The composition was similar to that at the
Arima colony, but Ocotea caracasana was abun-
dant in the August sample (not present at the
Arima colony; Table VIII), and Livistona was
present in both samples. The latter was almost
certainly obtained from Port-of-Spain gardens
and parks.

It is noteworthy that Trattinickia, which was
taken regularly by the Arima birds, was also
present in three of the four La Vache and
Huevos samples, but was never found in the
samples from the Oropouche and Aripo caves.
This strongly suggests that the tree is rare or
absent east of the Arima Valley.

The Foraging Range

Extravagant claims have been made for the
distances travelled by foraging Oilbirds in Vene-
zuela ( e.g ., 80 leagues— Funck, 1844) , but with-
out convincing evidence. For Trinidad the
evidence, incomplete though it is, suggests that
they can and do fly to places 15, or occasionally
even 30 miles distant from the cave in search
of food, but that foraging distances are usually
much shorter.

Oilbirds have several times been seen feed-
ing on the palm trees round the Queen’s Park
Savanna in Port-of-Spain, 14 miles from the
Huevos cave and 8 miles from La Vache cave.
The presence of numerous Livistona seeds in
the Huevos samples, but not in the La Vache
samples, suggests that the Huevos birds are in-
volved. The La Vache birds would in any case
have to cross hills of 1500 feet or more to reach
Port-of-Spain, while the Huevos birds need not
rise much above sea level. This is the longest
foraging distance for which there is direct evi-
dence, but almost certainly the Huevos birds at
times fly farther afield. The forests within five
miles of the Huevos cave are of dry monsoon
type, far poorer than those to the east; so that
the birds may often have to fly several miles
to suitable feeding grounds. Jessenia, which was
present in the Huevos samples, probably does
not grow within 15 miles, and Beilschmiedia, of
which one seed was found, has so far been
found only on the Aripo massif, 30 miles to
the east. Hence the Huevos birds must often
fly considerable distances to feed, and they must

214

Zoologica: New York Zoological Society

[47: 16

be at an ecological disadvantage compared with
the birds in the caves further east. It is not
known whether they visit Venezuela, of which
the nearest point is only nine miles away to the
west.

The Arima gorge birds visit Arima, five miles
away, where they take Livistona fruits, and
probably regularly visit the swampy forests be-
tween Arima and Valencia, eight or nine miles
away, since Jessenia is not commonly found
nearer than this. But they cannot regularly visit
the montane forest of the Aripo massif, only
five or six miles away to the east, since in over
112,000 seeds there were only three of Beil-
schmiedia, which forms an important part of the
food of the birds of the Oropouche and Aripo
caves. Conversely, Trattinickia was an impor-
tant element in the food of the Arima gorge
birds, but was never found in the samples from
Oropouche and Aripo, showing that the birds
from these caves do not normally visit the
drier forests only a few miles to the west.

Availability of Food

If a fruit-eating bird is found to be taking a
greater variety of fruits at some seasons than
others, it could mean that highly acceptable food
of several kinds is more abundant at that season
than at others, or it could mean that its pre-
ferred food is scarce and it is having to seek
other kinds. Conversely, few kinds of fruit in
the diet might mean either that food is scarce,
or that it is abundant and the bird is concentra-
ting on the kinds that it most prefers. A bare
analysis of food items in the diet may not enable
one to decide between these two opposite possi-
bilities. However, for the Oilbird we may safely
conclude that when it is taking a large number

of different kinds of fruit its food is especially
abundant. The different fruits which they take
and which make up the normal variety (the
exceptional ones being left out of account) all
recur regularly in their diet year after year, so
cannot be regarded as foods to which they turn
in times of shortage. Further, observations
showed that they were taking the fruits as they
became available in the forest, the periods when
they were found in the food samples corres-
ponding closely to the periods when the trees
were observed to have ripe fruits. This again
suggests a regular and normal exploitation of
a seasonally varying food supply.

On this criterion of available variety, the
Oilbird’s food supply shows well-defined season-
al changes in abundance (Text-fig. 4). The
greatest variety of food is available in the
months April-June. It has already been shown
that the breeding season is very long, and young
birds may be in the nest in any month of the
year. But there is a marked peak of laying in
the months December-May, which results in
more young birds being in the nests in the
months March-June than at any other time
(Text-fig. 4). This suggests that the breeding
season is timed so that most young are in the
nest when food is most abundant, though with
such a long breeding season and extremely long
nestling periods the adjustment cannot be very
exact.

For the Venezuelan population, living in a
more seasonal environment with a more marked
dry season, the seasonal availability of food may
show more marked fluctuations than in Trini-
dad, and the adaptation of breeding season to
the fluctuating food supply may be more im-
portant. But precise data from Venezuela are

MONTHS

NUMbER OF
SPECIES OF
LAURACEAE
IN FRUIT

Text-fig. 4. Relationship between number of young in the nests in each year, all years combined (dots),
and number of species of Lauraceae in fruit in each month (open circles).

1962]

Snow: Population, Breeding Ecology and Food of the Oilbird

215

lacking, both on the Oilbirds’ breeding season
and on the fruiting seasons of the food trees.

Regurgitation of Whole Fruits

As already mentioned in Part 1 , the seed sam-
ples regularly contained a proportion of whole
fruits which had been regurgitated intact, and
others from which only part of the pericarp had
been removed. A few of these were unripe fruits
that had been picked by mistake, or in apparent
anticipation of the time when they would be
ripe (especially Bactris cuesa in March and
April), but many were ripe and there was no
obvious reason why they should not have been
digested. In some months 10% or more of all
fruits were intact but usually the proportion was
between 1 and 5%. The percentage regurgit-
ated whole showed no correlation with the
availability of food, as measured by the number
of food trees that had ripe fruit in that month,
but was related to the state of breeding in the
colony. In each year, the highest percentages of
whole fruits were found towards the end of the
main period when young were in the nests
(July-September in 1958, April-May in 1960
and June-July in 1961 ) . In 1959, when very few
young were reared, the percentage of undigested
fruits never rose above 3%.

The way in which the pericarps are digested
and the seeds regurgitated is not properly under-
stood. Young birds taken from the nest and
kept under observation would regurgitate clean
seeds and unaltered whole fruits in the same
batch; no seeds were regurgitated with the
pericarp in a half-digested or softened state,
but some were regurgitated with it partially
stripped off. The few stomach contents ex-
amined also showed a mixture of whole fruits
and stripped seeds. Thus one must conclude that
no digestion takes place in the stomach, but
that the pericarps are gradually stripped off by
the muscular action in the stomach and passed
backward into the intestine, where digestion
takes place. Once the pericarp has begun to
come away from a seed, it will usually soon be
stripped off, while another fruit eaten at the
same time may still be intact.

Probably regurgitation begins to take place
a certain length of time after the food has been
eaten, and is largely under involuntary control
except when an adult is feeding its young (Part
1, p. 42). By the time it begins to take place,
most fruits will have been stripped of their
pericarps, but a few may have remained intact;
hence the regular occurrence of small numbers
of whole fruits in the food samples at all seasons.

Nestlings taken out of the nest, and kept all

day before being returned to the nest in the
evening, regurgitated a high proportion of whole
fruits: 118 (28%) of the total of 428. Regurgi-
tation may perhaps have been abnormal, owing
to the disturbance of being removed from the
nest, but it seems likely that at any rate the
larger nestlings, which receive very ample feeds,
may normally regurgitate a rather high propor-
tion of intact fruits. In addition the feeding
process itself involves much strenuous activity
by both adult and nestling (Part 1, p. 41), and
some fruits may be dropped as the food is
passed from adult to young. It is thus not sur-
prising that late in the nestling period a rather
large number of whole fruits are found on the
nests and the slopes below the nests. From the
ecological viewpoint, it suggests an ample food
supply, since if the supply of food were often
critical it is unlikely that such wasteful behav-
ior could persist.

Food Value of the Fruits Eaten,
and Its Ecological Implications

Oilbirds range widely over the forest at night,
and of the many fruits that are available they
take only a few kinds, avoiding many others
that are eaten by diurnal fruit-eating birds. Al-
most without exception, as already mentioned,
the fruits which they regularly take are kinds
that have a firm, in some cases hard, non-
succulent pericarp, enclosing a single seed.
From this it seems probable that succulent fruits
have insufficient food value for their weight to
be an economical food for Oilbirds, which not
only feed on the wing but must fly considerable
distances to their food trees.

To test the point further, the food values of
some of the more important kinds of fruit
eaten by Oilbirds were determined, through the
courtesy of Professor J. K. Loosli of the New
York State College of Agriculture. They are
given in Table XI, which also lists for com-
parison the equivalent values for some common
succulent fruits eaten by birds and men. It will
be seen that the fruits taken by Oilbirds have
an extremely high fat content and on average
a markedly higher protein content than the
succulent fruits; the fat: protein ratio is in fact
much like that of a nut kernel. (In fat content
they are exceeded by two well-known fruits,
the avocado pear (59%) and the olive (68%),
but the fat content of these may well have been
increased by artificial selection. They belong to
families which are represented in the Oilbird’s
diet, the Lauraceae and Oleaceae).

The food value of the fruits taken by Oilbirds
in comparison with succulent fruits is even

216

Zoologica: New York Zoological Society

[47: 16

Table XI. Analysis of Dried Pericarps of Fruits Eaten by Oilbirds,
and Some Other Fruits

Percentage composition

Protein

Fat

Carbohydrates,
ash and crude
fiber

Fruits regular
in Oilbird’s

Lauraceae

Cinnamomum

elongatum

9

44

47

diet

Ocotea

oblonga

11

19

70

Ocotea

wachenheimii

14

34

52

Burseraceae
Dacryodes sp.

11

24

65

Palmae

Bactris cuesa

13

39

48

Jessenia

oligocarpa

5

26

69

Fruits

occasionally

Lauraceae

Ocotea

canaliculata

8

34

58

taken but
generally
avoided

Anacardiaceae

Tapirira

guianensis

5

7

88

Some

Apple

2

Tr.

98

succulent

Fig

5

Tr.

95

fruits

Gooseberry

4

Tr.

96

Grape

4

Tr.

96

Red currant

6

Tr.

94

Note: Data for the first seven supplied by Professor J. K. Loosli, Cornell University, in lift.; percentages
for Tapirira by Mr. J. S. Leahy, Huntingdon Research Centre, England (in litt .); percentages for last five
reworked from data given in McCance & Widdowson (1946).

greater than is suggested by this table, as their
water content is lower. In most succulent fruits
the water content of the pericarp is between 75
and 85%, while that of the palms eaten by
Oilbirds is probably less than 50% (exact
figures not available), that of Dacryodes 60%,
and that of Ocotea wachenheimii about 75%.
Thus there seems little doubt that Oilbirds select
the kinds of fruit that they do primarily because
their food value is, for fruits, unusually high.

It may be noted that, although the palm
fruits have drier pericarps than the Lauraceae,
their seeds are relatively larger, weighing 54-
69% of the total weight of the fruit in the three
most important species, as against 33-42% in
the three species of Lauraceae for which there
are data (Table XII). Thus weight for weight
they yield approximately the same amount of
nutritive matter.

It was mentioned in Part 1 that many of the
trees on which Oilbirds feed are spicy or aro-
matic, and it was suggested that the olfactory

sense may be important in enabling them to
locate the food trees (see also Bang, 1960, for
anatomical evidence supporting this suggest-
ion). The situation in the Arima Valley in
March-May, 1961, was particularly illumina-
ting in this respect. Three common species of
Lauraceae were in fruit at the same time, Cinna-
momum elongatum, Ocotea oblonga and O.
canaliculata. The first two had fruited regularly
in previous years (Table VIII), but O. cana-
liculata had not been observed in fruit in the
lower part of the Arima Valley, and certainly
cannot have fruited prolifically. (Higher in the
valley, one tree under observation had produced
a little fruit in 1958 and 1959). The Oilbirds,
as usual, took large quantities of Cinnamomum
and O. oblonga fruit, but only one seed of O.
canaliculata was found among the thousands
examined. Samples of dried pericarp of all three
were analyzed (Table XI). In fat content O.
canaliculata was intermediate between the other
two, being exactly the same as O. wachenheimii,

1962]

Snow: Population, Breeding Ecology and Food of the Oilbird

217

Table XII. Weights of Fruits Eaten by Oilbirds

Number weighed

Mean weight
(gm.)

Mean weight of
pericarp (gm.)

Lauraceae

Cinnamomum elongation

5

0.6

0.4

Ocotea caracasana

2

9.5

5.0

Ocotea wachenheimii

16

3.3

1.9

Burseraceae

Dacryodes sp.

10

2.8

1.3

Trattinickia rhoi folia

2

1.5

1.0

Palmae

Badris cuesa

10

4.1

1.9

Euterpe langloisii

20

1.3

0.4

Geonoma vaga

15

0.17

0.10

Jessenia oligocarpa

6

16.8

6.2

Livistona chinensis

8

2.0

0.9

which is much eaten. In protein content it was
only 1% lower than Cinnamomum. O. canali-
culate. would thus seem from its nutritional
value to be a perfectly suitable food for Oilbirds.
There was, however, a striking difference be-
tween the three samples: the pericarp of Cinna-
momum was highly aromatic, that of O. oblonga
noticeably so, but that of O. canaliculata abso-
lutely non-aromatic. These differences were
noted without exception by several persons who
where presented with the samples in random
order and asked to comment on them.

A further comparison was made between the
three “gommiers,” Dacryodes and Trattinickia
which are much eaten, and Tapirira guianensis
which is avoided, though locally it is the com-
monest of the three. The local name reflects a
general similarity between these trees, all of
which produce gum from the bark, have rather
similar compound leaves and bear bunches of
fruit of much the same size and type. The Oil-
birds’ selection of the first two species and avoid-
ance of the third may have the same cause as
in the three Lauraceae: Dacryodes and Trattin-
ickia are aromatic, but Tapirira is not. How-
ever, analysis of Tapirira fruits (Table XI)
showed that their fat and protein content is
markedly low compared with the fruits that are
regularly eaten, so that it is probably not an
economical fruit for Oilbirds to take.

The palms on which Oilbirds feed are not
aromatic. It seems likely that they are located
visually, a method which would be easier for
palms, with their characteristic shape, than for
dicotyledonous trees.

These facts strongly suggest that, while the
food value of the fruit is of primary importance,
in searching for suitable food trees other than

palms Oilbirds are guided to an important de-
gree by the olfactory sense. This is in the main
an efficient method, since most of the trees on
which they feed, in the families Lauraceae and
Burseraceae, are aromatic; but it results in their
missing some species which from the nutritional
standpoint are suitable.

In Part 1, p. 42, it was shown that the three
young Oilbirds which were investigated ate
about one-quarter to one-third of their body-
weight in the course of a night. Using these fig-
ures, and the known food values of the fruits
eaten, we can make a rough assessment of the
total amount of protein, fat and carbohydrate
eaten by a nestling Oilbird in the course of its
growth. We may consider only the first 70 days
of its nestling period, when it is increasing in
weight.

If we assume that it eats one-third of its
weight every night (though this figure may well
be a little too high for the last 20 days), and
take an average growth-curve from 14 gm. at
hatching to 600 gm. at 70 days (Part 1, Table
III), we find that in the course of its first 70
days a nestling Oilbird will eat about 7,600
gm. of fruit. Using the mean figure for the
Lauraceae (Table XII), this corresponds to
4,560 gm. of pericarp, and to 1,370 gm. of
dried pericarp (using figures for Ocotea wach-
enheimii ) . Analysis of the percentage composi-
tion of dried pericarp showed that protein con-
tent was around 10% for most species, fat
content about 30%, and carbohydrate content
about 35%. Thus 1,370 gm. of dried pericarp
will give about 137 gm. of protein, 410 gm. of
fat, and 480 gm. of carbohydrate.

The total nitrogen in proteins is about 16%
by weight. Thus 137 gm. will contain about 22

218

Zoologica: New York Zoological Society

[47: 16

gm. of nitrogen. It is not possible to make more
than a rough assessment of the total nitrogen
in a 70-day nestling Oilbird, but the following
figures may be used provisionally. Probably
some 250 gm. of the total weight consists of
fat, and the remaining 350 gm. represents the
nitrogen-containing fraction, as the weight of
the adult, which has little fat, is on average
415 gm. and the body dimensions, muscular
development and feathering of the 70-day nest-
ling are somewhat less than in the adult. No
figures have been found for the total protein
content of a bird, but using the figures for
fresh meat (around 20%) and allowing a re-
duction for blood and bone, we may take 1 5 %
as a reasonable figure. Thus we may estimate
that a 600 gm. nestling Oilbird will contain
roughly 8.5 gm. of nitrogen (350 gm. X 15%
X 16%). This figure agrees well with an in-
dependent estimate from data given by Rubner
(quoted in Brody, 1945). According to Rubner,
one kg. of normal body substance (i.e., in-
cluding all organs in due proportion) contains
about 30 gm. of nitrogen. Consequently 350
gm. will contain about 10.5 gm.

In the absence of data on the nutritional value
and amino-acid composition of the proteins in
the nestling Oilbird’s diet, we can do no more
than guess whether an intake of 22 gm. of
nitrogen would be adequate to produce 9 or 10
gm. of nitrogen in the body. But the efficiency
of utilization of dietary nitrogen derived wholly
from fruit pericarps, over a period of 70 days,
could hardly be expected to exceed 50%, which
suggests that the young Oilbird is growing as fast
as is possible on such a diet.

Turning to the energy requirements of the
growing Oilbird, we may make another rough
calculation. The total intake of food, as esti-
mated above, yields 6,343 calories (on the basis
of 4.1 cal. per gm. of carbohydrate, 9.3 cal.
per gm. of fat, 4.1 cal. per gm. of protein;
Best & Taylor, 1949). The calorific value of a
70-day Oilbird weighing 600 gm. is not exactly
known, but it may be estimated at 3,025 cal.,
on the basis of 9.3 cal. per gm. of fat and 2
cal. per gm. of the rest of body-substance. This
gives a gross efficiency of growth, in respect
of consumed metabolic energy, of about 48%.
Brody (1945) quotes equivalent figures for rats
and chickens: for rats, gross efficiencies up to
13.6%; for slow-growing chickens, gross effi-
ciencies up to about 20%. The apparently very
high gross efficiency of the nestling Oilbird is
probably made possible by the development of
the thick insulating layer of fat, the growth of
a thick covering of down and relative inactivity.

These calculations, rough though they are,
probably have enough validity to justify the con-
clusion that on a diet of pericarps of palm
and lauraceous fruits young Oilbirds could not
obtain enough nourishment to develop any
faster than they do. Thus the very long nestling
period may be regarded primarily as an adapta-
tion to the very specialized diet, as provisionally
suggested in Part 1, p. 44. This adaptation could
not have been acquired unless the birds had
available to them nest-sites safe from most
natural predators.

This conclusion rests on the assumptions that
the parents could not provide more food for
the nestlings, or that if they could, the nestlings
would not be able to deal with a greater quantity
of bulky food per day. Neither of these assump-
tions can be tested on present evidence, but the
facts that Oilbirds have often to fly several miles
for food, that their flight-speed in level flight
is not high, and that they spend up to four hours
apparently feeding themselves before they re-
turn with food for the young (Part 1, p. 41),
all suggest that the three or four feeds in the
night that were recorded are as many as are
normally possible under the circumstances.

Acknowledgements

I have received much help in many ways in
the course of this investigation. In particular I
would like to thank the following: my wife,
for help with various aspects of the field work,
especially in finding some of the Oilbird’s food
trees in the forest; Mrs. H. Newcome Wright,
the owner of Spring Hill Estate where the Arima
gorge colony is situated, for her help and hos-
pitality throughout the whole period of the
study; Mr. John Dunston and Mr. R. P. ffrench,
for visiting the colony at times when I was
unable to do so, and to the former especially
for keeping a record of breeding activity for
eight months after my own observations had
ceased; Mr. N. Y. Sandwith, for invaluable and
long-continued help in identifying the food
trees; Dr. A. Kostermans, Dr. H. E. Moore, and
Dr. A. C. Smith, for help with special points of
identification and nomenclature; Professor J. K.
Loosli and Mr. J. S. Leahy, for analyzing sam-
ples of the Oilbird’s food; Dr. W. G. Downs and
Russ Kinne for photographic assistance; Dr.
T. H. G. Aitken, for information on the extinct
Chacachacare colony; Dr. A. N. Worden and Dr.
K. W. G. Shillam for advice on the section on
growth and nutrition: Dr. D. Lack for working
facilities at the Edward Grey Institute, Oxford,
during the writing of the paper; and Mr. R. E.
Moreau, for helpful criticism of the paper in
draft.

1962]

Snow: Population, Breeding Ecology and Food of the Oilbird

219

The work was made possible by generous
grants from the National Science Foundation
(G 4385 and G 21007), for which grateful
acknowledgement is made.

Summary

The present adult population of Oilbirds in
Trinidad, in the eight known colonies, is esti-
mated to be 1,460. Five small colonies have
been exterminated in recent years. The popula-
tion is considered to be in no immediate danger,
but is potentially vulnerable.

The breeding season is very long, extending
from December to September, but nearly all
clutches are laid before June. The moult takes
place mainly in June-November, but it lasts so
long that in some individuals it almost certainly
overlaps the beginning and/or end of the breed-
ing cycle. There is evidence that birds may begin
a second wing-moult before the first is complete.

Clutch-size varies from 1 to 4 eggs, the mean
being 2.7. It was found to decrease in the course
of the breeding season, probably in correlation
with decreasing abundance of food.

Nearly 50% of observed nestings were suc-
cessful, earlier nestings being more successful
than later ones. The causes of failure were var-
ious; there was no evidence of starvation of nest-
lings except perhaps occasionally in the first
few days.

An analysis is given of 1 1 2,000 seeds col-
lected from the Arima gorge colony. The im-
portant fruits comprised 7 species of palms, 8
of Lauraceae, 2 of Burseraceae, and 1 of Ara-
liaceae. Two important palms provided a more
or less steady supply of food throughout the
year: the Lauraceae were nearly all regularly
seasonal. Evidence is given that Oilbirds may
fly up to 30 miles for their food, but foraging
distances are usually much shorter.

The greatest variety of food is available in
the months April-June, corresponding to the
period when most young are in the nest. In all
months of the year, a proportion of the fruits
are regurgitated whole, rising to 10% or more
in the latter part of the breeding season.

The food values of some of the chief foods
are analyzed, and it is concluded that on a diet
of fruit pericarps young Oilbirds could not de-
velop much faster than they do. The extremely
long fledgling period is thus primarily an adapta-
tion to the specialized diet.

Literature Cited

Bailey, L. H.

1947. Indigenous palms of Trinidad and Tobago.

Gentes Plantarum (Cornell).

Bang, B. G.

1960. Anatomical evidence for olfactory func-
tion in some species of birds. Nature, 188:
547-549.

Beard, J. S.

1946. The natural vegetation of Trinidad. Clar-
endon Press, Oxford.

Best, C. H., & N. B. Taylor

1949. The living body. Chapman & Hall, Lon-
don.

Brody, S.

1945. Bioenergetics and Growth. Reinhold Publ.
Co., New York.

Chapman, F. M.

1894. On the birds of the island of Trinidad.
Bull. Amer. Mus. Nat. Hist., 6: 1-86.

Cherrie, G. K.

1907. The Museum News (Brooklyn Inst. Mus.),
3: 14.

Dorward, D. F.

1962. Comparative biology of the White Booby
and the Brown Booby Sula spp. at Ascen-
sion. Ibis, 103 b: 174-220.

Funck, N.

1844. Notice sur le Steatornis caripensis (Gua-
charo incolarum). Bull. Acad. Roy. Sci.
Bruxelles, 11(2): 371-377.

Hollister, G. E.

1926. The Guacharo or Oil Bird of the Arima
Gorge. Bull. N. Y. Zool. Soc„ 29: 139-145.

Kingsley, C.

1871. At last: a Christmas in the West Indies.
London.

Lack, D.

1954. The natural regulation of animal num-
bers. Clarendon Press, Oxford.

McAtee, W. L.

1922. Notes on the food of the Guacharo ( Stea-
tornis caripensis) . Auk, 39: 108-109.

McCance, R. A., & E. M. Widdowson

1946. The chemical composition of foods. Lon-
don (H. M. Stationery Office).

Myers, J. G.

1935. The Guacharo or Diablotin ( Steatornis
caripensis Humbt.). Proc. Zool. Soc. Lon-
don, 1934 (1935): 717-721.

Pietri, E. de B.

1957. El Guacharo (monografia). Bol. Soc.
Venezolana Cienc. Nat., 18: 3-41.

Roosevelt, T.

1917. [Account of Oropouche cave]. Scribner’s
Magazine, 1917.

220

Zoologica: New York Zoological Society

[47: 16

Sanderson, I. T.

1940. Caribbean Treasure. London.
Stresemann, V. & E.

1960. Die Handschwingenmauser der Tagraub-
vogel. J. Orn., 101: 373-403.

APPENDIX

Systematic List of the Oilbird’s Food Trees

The order and nomenclature follow Beard
(1946), with alterations and additions as noted.

Lauraceae

Aiouea schomburgkii Meissn.

Aniba firmula (Nees) Mez

Aniba panurensis Mez, listed in Beard, is a
synonym.

Aniba trinit atis (Meissn.) Mez
Beilschmiedia sulcata (R. & P.) Mez
Genus and species new to Trinidad.

Licaria guianensis Aubl.

New to Trinidad.

Nectandra martinicensis (Jacq.) Mez
Nectandra membranacea (Sw.) Griseb.
Nectandra kaburiensis Kosterm.

Listed in Beard as N. surinamensis Mez.
Ocotea canaliculata (Rich.) Mez
O cotea caracasana (Nees) Mez
New to Trinidad.

Ocotea oblonga (Meissn.) Mez
Ocotea wachenheimii R. Benoist

Ocotea arenaensis R. L. Brooks, listed in Beard,
is a synonym.

Cinnamomum elongatum (Vahl) Kosterm.
Listed as Phoebe elongata (Vahl) Nees in
Beard. Phoebe is synonymous with Cinnamo-
mum, fide Kostermans.

Myristicaceae

Virola surinamensis (Rol.) Warb.

Burseraceae
D aery odes sp.

Specific determination depends on specimens
with flowers, which have not yet been obtained.
Previously known from Trinidad on the basis of one
old specimen, but locally well-known as “mountain
incense” or “gommier montagne,” and referred to
under the latter name by Beard.

Wall, G. P., & J. G. Sawkins

1860. Report on the geology of Trinidad. Lon-
don.

Williams, C. B.

1922. Notes on the food and habits of some
Trinidad birds. Bull. Dept. Agric. Trini-
dad and Tobago, 20: 123-185.

Trattinickia rhoifolia Willd.

Malpighiaceae

Byrsonima spicata (Cav.) Rich.

Anacardiaceae

Tapirira guianensis Aubl.

Araliaceae

The atypical fruit which remained unidentified
(p. 212) is referred to this family with reasonable
certainty by Mr. N. Y. Sandwith (Kew) and Dr. A.
C. Smith (U.S. National Museum), who were sent
detailed drawings of the mature fruit. The latter
suggests the genera Dendropanax or Schefflera as
being the most likely.

Sapotaceae

Pouteria minutiflora (Britt.) Sandwith
Oleaceae

Linociera caribaea (Jacq.) Knobl.

Boraginaceae

Cordia bicolor DC.

Cordia lockhartii Kuntze, listed in Beard, is
a synonym.

Palmae

Bactris cuesa Crueg.

Euterpe langloisii Mart.

No consistent differences could be found be-
tween the many Euterpe seeds and fruits in the food
samples. All appeared to come from the common
Euterpe of the forests of the Northern Range, for
which the name langloisii may provisionally be
used. The species have been very finely split (Bailey,
1947).

Geonoma vaga Griseb. & Wendl.

Jessenia oligoarpa Griseb. & Wendl.

Livistona chinensis R. Br.

Roystonea oleracea (Jacq.) Cook

1962]

Snow: Population, Breeding Ecology and Food of the Oilbird

221

EXPLANATION OF THE PLATES

Plate I

Fig. 1. Seeds and whole fruit (upper right) of
Jessenia oligocarpa. (Millimeter scale is
shown in this and the next five figures.)
Fig. 2. Seeds of Euterpe langloisii.

Plate II

Fig. 3. Seeds of Livistona chinensis.

Fig. 4. Seeds of Trattinickia rhoifolia.

Plate III

Fig. 5. Seeds of Ocotea wachenheimii.

Fig. 6. Seeds of Aniba firmula.

Plate IV

Fig. 7. Oilbird nests in the Arima gorge, showing
eggs and small young, and seeds of Ocotea
caracasana (smooth, shiny and spindle-
shaped) and Jessenia oligocarpa (larger,
streaked) round the edges of the nests.

SNOW

PLATE I

1' 2\ . 3i" ,4! . 5| 6[ , 7 . 8| , 9 10 11

1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 1 1 ! 1 1 1 1 1 1 1 1 1 1 1 1 1 1 !' m i ! 1 1 : i ! 1 1 1 : mill! Iiiiiliiiilinilii!:!im' ■

FIG. 1

FIG. 2

THE NATURAL HISTORY OF THE OILBIRD,
STEATORNIS CARIPENSIS, IN TRINIDAD, W.l.

SNOW

PLATE II

- iZ-'Atutaw ■

1

2

3

4

5

imlmi

iiiilim

INI

III!

INI

INI

Mi!

III!

FIG. 4

THE NATURAL HISTORY OF THE OILBIRD,
STEATORNIS CARIPENSIS, IN TRINIDAD. W.l.

SNOW

PLATE III

1 . 1

2

1 3

, 4

1 5

1 6

1

Sim hni

1 1 1 1 1 1 1 1 1

lllllllll

lllllllll

lllllllll

lllllllll

mil

FIG. 5

iriwitjmm.K--.-~ ■MaamUttaUlB^

i

2

, 3

1 4

5I

l 6I

Inn

lllllllll

lllllllll

lllllllll

INI

llllllll

lllllllllll

FIG. 6

THE NATURAL HISTORY OF THE OILBIRD,
STEATORNIS CARIPENSIS, IN TRINIDAD. W.l.

SNOW

PLATE IV

FIG. 7

THE NATURAL HISTORY OF THE OILBIRD.
STEATORNIS CARIPENSIS, IN TRINIDAD, W.l.

C1962J

Zoologica: Index to Volume 47

223

Names in bold lace indicate new
genera, species or subspecies; num-
bers in bold face indicate illustra-
tions; numbers in parentheses are
the series numbers of papers con-
taining the plates listed immediately
following.

A

Abramis brama, 110
Acipenser brevirostrum, 109
fulvescens, 109
ruthenus, 109
sturio, 109

Admetus barbadensis, 25, 27, 30, 31
Aequidens pulcher, 114
Ambloplites rupestris rupestris, 112
Amblyrhynchotes honckenii, 115
Amia calva, 109
Amphiprion ephippium, 114
Anabas testudineus, 114
Anguilla anguilla, 111
rostrata. 111

Anisotremus davidsonii, 113
Aphyocypris pooni, 110
Aplocheilus lineatus, 111
Aplodonotus grunniens, 113
Archosargus probatocephalus, 113
Arothron reticularis, 115
Astronotus ocellatus, 114
Astyanax bimaculatus, 109
fasciatus mexicanus, 109

B

Badis badis, 113

Balantocheilus melanopterus, 110
Balistes capriscus, 115
vetula, 115
Belontia signata, 114
Blennius pholis, 114
Blicca bjoerkna, 110
Botia hymenophysa, 111
Brachydanio albolineatus, 110
nigrofasciatus, 110
rerio, 110

Bufo melanostictus, 21, 23, (4) PI. I

C

Cacicus cela, 39
Canthigaster margaritatus, 115
Caranx crysos, 113
hippos, 113

Carassius auratus, 110
Carcharias taurus, 108
Carcinoscorpius rotundicauda, 1,

3, 5, (1) PI. I

Centrarchus macropterus, 112
Centropristes striatus, 112
Chaca chaca, 111
Chaetodipterus faber, 113
Chaetura brachyura, 130
chapmani, 130
cinereiventris, 130
spinicauda, 130

INDEX

Chalceus macrolepidotus, 109
Channa striata, 112
Chromis chromis, 114
Cichlasoma bimaculatum, 114
biocellatum, 114
facetum, 114
hellabrunni, 114
maculicauda, 114
nigrofasciatum, 114
severum, 114
Clarius batrachus, 111
Cobitis taenia, 111
Colisa lalia, 114
Colossoma nigripinnis, 109
Conger conger, 111
Copeina guttata, 109
Coregonus clupeaformis, 109
Coris julis, 114
Corvina nigra, 113
Corydoras aeneus, 111
julii, 111

Corynopoma riisei, 109
Ctenobrycon spilurus, 109
Ctenops vittatus, 114
Cyclocheilichthys apogon, 110
Cynoscion regalis, 113
xanthulus, 113
Cyprinus carpio, 110
Cypseloides rutilus, 130
zonaris, 130

D

Damon variegatus, 25, 28, 30, 32
Danio malabaricus, 110
Dascyllus aruanus, 114
Dasyatis pastinaca, 109
Dentex dentex, 113
Desmognathus monticola
monticola, 10

ochrophaeus carolinensis, 10
Dicentrarchus labrax, 112
Diemictylus viridescens
viridescens, 10
Diodon hystrix, 115
Diplodus annularis, 113
sargus, 113
Doras sp., Ill
Dormitator maculatus, 114

E

Electrophorus electricus, 110
Eleotris vittata, 114
Enneacanthus chaetodon, 112
gloriosus, 112

Epalzeorhynchus kalopterus, 110
Epinephalus adscensionis, 112
aeneus, 112
analogus, 112
chrysotaenia, 112
gigas, 112
guaza, 112
guttatus, 112
itajara, 112
labriformis, 112
morio, 112
striatus, 112
summana, 112

Epiplatys chaperi, 112
Esox lucius, 109
masquinongy, 109
Eurycea bislineata wilderae, 10
Exodon paradoxus, 109

G

Galeichthys felis, 111
Ginglymostoma cirratum, 108
Girella nigricans, 113
Globicephala scammonii, 59, (7)

Pis. I & II

Gymnocorymbus temetzi, 110
Gymnostinops montezuma, 39
Gymnothorax funebris, 111
mordax, 111

H

Haemulon sciurus, 113
Heliconius erato, 141, 149
melpomene, 141, 149. (13) PI. I
Hemichromis bimaculatus, 114
Hemigrammus caudovittatus, 110
ocellifer, 110
uniliniatus, 110
Hemiodus semitaeniatus, 110
Hepsetus odeo, 110
Herichthys cyanoguttatus
carpintis, 114

Heterodontus francisci, 108
Holacanthus isabelita, 113
Hucho hucho, 109
Hyphessobrycon bifasciatus, 110
Hypsipops rubicunda, 114

I

Ictalurus nebulosus, 111
punctatus, 111
Ictiobus bubalus, 110
cyprinella, 110

K

Kuhlia sandvicensis, 112
taeniura, 112

L

Labrus beraylta, 114
festivus, 114
Lates niloticus, 112
Lepidosiren paradoxus, 109
Lepisosteus osseus, 109
platostomus, 109
productus, 109
spatula, 109
Lepomis gibbosus, 112
megalotis, 112
Leporinus fasciatus, 110
friderici, 110
megalepis, 110
Leuciscus cephalus, 110
idus, 110

Limia nigrofasciata, 112
Limulus polyphemus, 1, 3, 5, (1)

PI. I

224

Zoologica: Index to Volume 47

[1962j

Liza auratus, 112
capito, 112
chelo, 112

Loricaria microlepidogaster, 111
Lota lota lacustris, 111
Lutjanus argentiventris, 113
griseus, 113
jocu, 113
synagris, 113

M

Malapterurus electricus. 111
Manacus manacus, 65, 66-68, 70,
72-77, 81. 83, 84. 86, 94-97
Marcusenius isidori, 109
Mastacembelus erythrotaenia, 115
Medialuna californiensis, 113
Melanotaenia maccullochi, 112
nigricans, 112
Metynnis maculatus, 110
roosevelti, 110
schreitmuelleri, 110
Micropogon undulatus, 113
Micropterus dolomieui, 112
salmoides salmoides, 112
Moenkhausia oligolepis, 110
Monodactylus argenteus, 113
Mormyrus kannume, 109
Morone americana, 112
Mugil cephalus, 112
rammelsbergii, 112
Muraena clepsydra, 111
helma, 111

Mycteroperca bonaci, 112
tigris, 1 12
venenosa, 112
Myleus asterias, 110
gurupyensis, 110
schomburgkii, 110
Mylossoma duriventre, 110

N

Nannaethiops unitaeniatus, 110
Nemacheilus barbatulus, 111
Neoceratodus forsteri, 109
Notemigonus crysoleucas
crysoleucas, 110
Notopterus notopterus, 109

O

Oblada melanura, 113
Ophicephalus obscurus, 112
Opsanus tau, 1 15
Orthopristis chrysopterus, 113
Osphrenemus goramy, 114
Ostinops decumanus, 39,

40 , 41, 48. 49

Osteocheilus hasseltii, 110
vittatus, 110

Osteoglossum bicirrhosum, 109
Oxyeleotris marmorata, 114

P

Pagellus centrodontus, 113
mormyrus, 113
pagrus, 113
Pantodon bucholzi, 109
Panyptila cayennensis, 131
Paralabrax clathratus, 112
maculatofasciatus, 112
nebulifer, 112

Perea flavescens, 113
fluviatilis, 113

Phrrictocephalus hemeliotropus, 111
Pimelodella gracilis, 111
Pimelodus clarias. 111
Pimelometapon pulchrum, 114
Pipra erythrocephala, 183, 186, 187.

189. 191. 194, 195
Platax orbicularis, 113
teira, 113

Platichthys flesus, 115
Plecostomus commersonii, 111
punctatus, 111
rachovii, 111

Plethodon cinereus cinereus, 10
cinereus glutinosus, 10, 12
jordani metcalfi, 10
yohnalossee, 9, 12
Pogonias cromis, 113
Pollachius pollachius, 111
Polyacanthus hasselti, 114
Polyodon spathula, 109
Polypterus senegalis, 109
Pomoxis nigromaculatus, 113
Premnas biaculeatus, 114
Pristella riddlei, 110
Pristis pectinatus, 108
Pristolepis grooti, 113
Prochilodus insignis, 110
Protopterus aethiopicus, 109
annectens, 109
dolloi, 109

Psarocolius angustifrons, 39
Psetta maxima, 114
Pseudotriton ruber, 11
Pterois volitans, 114
Pterophyllum eimekei, 114
Puntazzo puntazzo, 113
Puntius chola, 1 10
cummingi, 110
dunkeri, 110
everetti, 110
lateristriga, 110
mgrofasciatus, 110
oligolepis, 110
schwanfeldi, 110
semifasciolatus, 110
ticto, 110
titteya, 110

Pylodictis olivaris, 111

R

Raja clavata, 108
maculata, 108
punctata, 108
Rasbora einthovenii, 111
heteromorpha, 111
trilineata, 111
Reinarda squamata, 131
Rhamdia sebae, 111
Rhinichthys atratulus atratulus, 111
Rhodeus ocellatus, 117, (10)

Pis. I & II

X Tanakia tanago, 118, (10) PI. I
sericeus, 111
spinalis, 117, (10) PI. II
X Tanakia tanago, 119, (10) PI. II

Rhynchobdella aculeata, 115
Roccus chrysops, 112
saxatilis, 112
Roncador stearnsi, 113
Rutilus rutilus, 11

S

Saccobranchus fossilis. 111
Salmo gairdnerii, 109
trutta fario, 109
lacustris, 109
Salvelinus fontinalis, 109
malma, 109

Scaphyrhynchus platorhynchus, 109
Scardinius erythrophthalmus, 111
Scarus cretensis, 114
Scatophagus argus, 113
Sciaenops ocellata, 113
Scorpaena guttata guttata, 114
porcus, 114
scrota, 114

Scylliorhinus canicula, 108
stellaris, 108
Selena vomer, 113
Serranus cabrilla, 112
scriba, 112

Serrasalmus nattereri, 110
niger, 110
rhombeus, 110
spilopleura, 110
Siganus chrysospilos, 114
Silurus glanis, 111
Solea solea, 115
Sparus auratus, 113
Spondyliosoma cantharus, 113
Steatornis caripensis, 199, 200. 204.

208. 214, (16) Pis. I-IV
Stenella graffmani, 122, 126. 127,
(11) Pis. II & III
Stizostedion lucioperca, 113
Symphysodon discus, 114
Synodontis schal, 111

T

Tachypleus gigas, 1, 3, 5, (1) PI. I
Tanakia tanago, 117, (10) PI. I
X Rhodeus ocellatus, 119, (10)

Pis. I & II

X Rhodeus spinalis, 119, (10)

PI. II

Tarpon atlanticus, 109
Tautoga onitis, 114
Tetraodon fluviatilis, 115
lineatus, 115
Therapon jarbua, 112
Thoracocharax stellatus, 110
Tilapia macrocephala, 114
mossambica, 114
natalensis, 114
nilotica, 114
zilli, 114 <,

Tinea tinea, 111
Torpedo marmorata, 109
Toxotes jaculatrix, 113
Triakis semifasciata, 108
Trichogaster leerii, 114
Trigla lucerna, 114
Tursiops truncatus, 121, 123, (11)
Pis. I, II, IV

U

Uca pugilator, 15, 16, (3) PI. I
Umbrina roncador, 113

Z

Zarhynchus wagleri, 39
Zeugopterus punctatus, 114
Zeus faber, 112

NEW YORK ZOOLOGICAL SOCIETY

GENERAL OFFICE

630 Fifth Avenue, New York 20, N. Y.

PUBLICATION OFFICE

The Zoological Park, Bronx 60, N. Y.

OFFICERS

PRESIDENT VICE-PRESIDENT
Fairfield Osborn Laurance S. Rockefeller

SCIENTIFIC STAFF:

William G. Conway. . Director,

Zoological Park

Christopher W. Coates. .Director, Aquarium

John Tee-Van General Director

Emeritus

ZOOLOGICAL PARK

Joseph A. Davis, Jr. . . Curator, Mammals

Grace Davall Assistant Curator,

Mammals and Birds

William G. Conway . . Curator, Birds

Peter Ames Assistant Curator,

Birds

Herndon G. Dowling. Curator, Reptiles
Charles P. Gandal. . . Veterinarian

Lee S. Crandall General Curator

Emeritus

Roland Lindemann . . . .Consultant in Mam-
mal Management

AQUARIUM

James W. Atz Curator

Carleton Ray Associate Curator

Ross F. Nigrelli Pathologist & Chair-

man of Department
of Marine Biochem-
istry & Ecology

Klaus D. Kallman .... Geneticist

C. M. Breder, Jr Research Associate

in Ichthyology

SECRETARY TREASURER

George W. Merck David H. McAlpin

Harry A. Charipper. . . Research Associate

in Histology

Sophie Jakowska Research Associate

in Experimental
Biology

Louis Mowbray ..... Research Associate

in Field Biology

GENERAL

William Bridges . . Editor & Curator,

Publications

Dorothy Reville . . Editorial Assistant

Sam Dunton Photographer

Henry M. Lester . . Photographic Consultant

DEPARTMENT OF TROPICAL
RESEARCH

Jocelyn Crane Directing Curator

John Tee- Van Associate

William K. Gregory .... Associate

AFFILIATE

L. Floyd Clarke Director,

Jackson Hole Biological Research Station

EDITORIAL COMMITTEE
Fairfield Osborn, Chairman
James W. Atz Lee S. Crandall

William Bridges Herndon G. Dowling

Christopher W. Coates John Tee-Van
William G. Conway

'Q